*.

I
I
I
I
I
I
I
I
I

I,

Marine Biological Laboratory Library

Woods Hole, Mass.

Presented by

Dr. P. W. Whiting
Aug . 6, 1959

X

I

I
I
I
I
I
I
I

tr
tr

O

m
o

THE ANTHOCYANIN PIGMENTS

OF PLANTS

CAMBRIDGE UNIVERSITY PRESS

C. F. CLAY, MANAGER
ILonttOtt: FETTER LANE, E.G.

100 PRINCES STREET

lork: G. P. PUTNAM'S SONS
Botnbao, Calcutta ant) fHatorag: MACMILI.AN AND CO., LTD

Toronto: J. M. DENT AND SONS, LTD.
ffokp.o: THE MARUZEN-KABUSHIKI-KAISHA

All rights reserved

r^/'

__ r

"

THE

ANTHOCYANIN PIGMENTS

OF PLANTS

MURIEL

Fellow of Newnham College, Cambridge, and formerly Research Student at the
John Innes Horticultural Institution, Merton, Surrey

Cambridge :

at the University Press
1916

PREFACE

the various investigations which have been made upon the
anthocyanin pigments along botanical, chemical and genetical
lines, no complete account has yet been written. It is the object of
this book to provide such an account of the work which has been done.
Although it is only within recent years that any very notable researches
have been made upon these pigments, I feel that consideration is due
to the many workers who, in the course of the last century, have
paved the way for their successors. This I offer as my excuse for
dwelling in the following pages upon some researches which are now
almost entirely superseded.

I do not pretend to claim that anthocyanins will ever have a great
significance from the strictly botanical point of view r . Even when the
obscurity which surrounds their physiological functions is elucidated,
it can scarcely be expected that they will have a significance in the
least comparable, for instance, to that of chlorophyll. From the
strictly chemical standpoint, as chemical compounds, they have a
certain interest. But I believe it to be in connection with problems
of inheritance that they will provide a great and interesting field for
research.

We have now ample evidence that the development in plants of
many and various anthocyanin pigments affords an almost unlimited
supply of material for the study of inheritance. It must also be patent
to those who have been working on the subject of Genetics that
a proper conception of the inter-relationships and inheritance of the
manifold characters of animals and plants will be greatly facilitated
by a knowledge of the chemical substances and reactions of which
these characters are largely the outward expression. Herein lies the

03

vi PREFACE

interest connected with anthocyanin pigments. For we have now, on
the one hand, satisfactory methods for the isolation, analyses and
determination of the constitutional formulae of these pigments. On
the other hand, we have the Mendelian methods for determining the
laws of their inheritance. By a combination of the two methods, we
are within reasonable distance of being able to express some of the
phenomena of inheritance in terms of chemical composition and
structure. There can be little doubt that exact information of this
kind must be at least helpful for the true understanding of the vital
and important subject of Heredity.

In the preparation of this book I gratefully acknowledge the help
afforded to me by many of my friends, and I am especially indebted
to Mrs E. A. Newell Arber for kindly correcting my proofs.

To Professor Bateson, F.R.S., my sincerest thanks are due for the
great interest he has taken in much of my work which is included in
this volume, and for his many valuable suggestions and criticisms.
I wish also to record my thanks to Dr F. F. Blackman, F.R.S., for
criticisms and assistance with the manuscript.

I regret that some of the most recent and important work on the
subject has not been altogether successfully incorporated in the book,
owing to the difficulty I have experienced in learning, at the earliest
opportunity, of the results obtained by scientists in other countries
during this and the preceding year.

M. W.

CAMBRIDGE,
May, 1916.

CONTENTS

PART I

GENERAL ACCOUNT

PAGE

CHAPTER I. INTRODUCTORY ... 1

Definition of Anthocyanin Pigments Early Observations of Boyle
and Nehemiah Grew Main Lines of Investigation Distribution-
Influence of Outside Factors Functions Chemical Nature Connec-
tion with Genetics Factors for Anthocyanin Formation in Antirrhinum,
Latliyrus and Matthiola Origin of Anthocyanins from Flavones
Hypothesis as to Mode of Formation Formation by Oxidation-
Formation by Reduction Chemical Interpretation of Mendelian
Factors.

CHAPTER II. THE MORPHOLOGICAL DISTRIBUTION OF ANTHO-
CYANINS 17

Wideness of Distribution in Higher Plants Presence in Lower
Plants Distribution in Leaves In Stems and Petioles In Young
Leaves and Shoots In Young Leaves in Tropics Autumnal Colora-
tion Development in Injured Organs Formation in Winter In
Drought In High Alpine Plants In Halophytes In Red-leaved
Varieties In Petals, Perianths and Inflorescences In Fruits In
Seeds In Roots In Parasites and Saprophytes.

CHAPTER III. THE HISTOLOGICAL DISTRIBUTION OF ANTHO-
CYANINS 29

Condition of Pigment in Cell-sap In Crystalline and Solid Form
in Cells List of Species in which Solid and Crystalline Anthocyanin
has been noted Artificial Formation of Anthocyanin Bodies Various
Deposits Coloured with Anthocyanin Classification of Distribution
of Anthocyanin in Leaf-tissues In Petals Combination of Pigments
in Petals Localisation of Anthocyanin in Various Natural Orders.

via CONTENTS

PAGE

CHAPTER IV. THE PROPERTIES AND REACTIONS or ANTHO-

CYANINS 44

Question as to Number of Pigments Included in the Group Crystal-
line Form Solubilities Qualitative Reactions With Alkalies With
Acids With Lead Acetate Reactions of Accompanying Substances
Weigert's Classification Willstatter's Interpretation of Colour Reactions
Reactions with Iron Salts With Sodium Bisulphite Action of
Nascent Hydrogen Compounds with Acids Spectroscopic Exami-
nation Willstatter's Groups of Anthocyanins.

CHAPTER V. THE ISOLATION AND CONSTITUTION OF ANTHO-
CYANINS 58

General Views Special Cases of Isolation and Analyses Flower-pig-
ment of Centaurea by Morot Pigment of Wine by Glenard Pigment
of Coleus Leaves by Church Flower-pigment of Rosa by Senier Pig-
ment of Wine by Gautier Pigment of Grapes by Heise Pigment of
Red Vine Leaves by Gautier Flower-pigment of Althaea by Glan
Fruit-pigment of Bilberry by Heise Flower-pigment of Pelargonium by
Griffiths Flower-pigment of Althaea and Pelargonium by Grafe
Flower-pigment of Antirrhinum by Wheldale Flower-pigment of Cen-
laurea, Delphinium, Malva, Pelargonium and Rosa and Fruit-pigment
of Vaccinium by Willstatter Constitution of Anthocyanins.

CHAPTER VI. PHYSIOLOGICAL CONDITIONS _AND FACTORS IN-
FLUENCING THE FORMATION OF ANTHO-
CYANINS 81

Connection with Photosynthesis Connection with Accumulation of
Synthetic Products Formation as Result of Mechanical Injury and
Injury by Insects and Fungi As Result of Pecortication Formation
in Alpine Plants In Parasites Connection with Nutrition Effect
of Temperature Effect of Light Varying Results in Flowers, Stems,
Leaves and Roots Connection with Presence of Oxygen Analyses of
Gaseous Exchange in Red Leaves Effect of Drought Effect of Sugar-
feeding Experiments of Overton, Katie and Gertz Summary of
Results.

CHAPTER VII. REACTIONS INVOLVED IN THE FORMATION OF

ANTHOCYANINS 104

Early Views as to Origin of Anthocyanin from Chlorophyll and Tannins
Palladin's Respiration Pigments Oxidase Reactions in Plants Hypo-
thesis as to Origin from Flavones Criticism of Hypothesis Evidence
for Hypothesis from Results of Cross-breeding From Analogous
Reactions From Connection with Photosvnthesis From Gaseous

*/

Exchange in Red Leaves Work of Keeble, Armstrong and Jones on
Oxidases and Anthocyanins Formation of Artificial Anthocyanin
Results of Combes and Willstatter.

CONTENTS ix

PAGE

CHAPTER VIII. THE SIGNIFICANCE or ANTHOCYANINS . . 126

Biological Significance Physiological Significance Light-screen
Hypothesis Pringsheim's and Reinke's Results on Bleaching of Chloro-
phyll Kerner's Suggestions as to Uses of Anthocyanin Evidence of
Engelmann and Reinke against Light-screen Hypothesis Pick's Hypo-
thesis of Protection of Diastase Stahl's Hypothesis of Conversion of
Light Rays into Heat Ewart's Criticism of Stahl Investigation of
Internal Temperature of Red Leaves by Smith.

PART II

ANTHOCYANINS AND GENETICS
CLASSES OF VARIATION 145

Kinds of Colour- variation Different Series in Different Genera
Variation to Albinism Red Varieties Soluble- and Plastid- Yellow
Varieties The Cream Variety The Blue Variety Deep and Pale
Varieties Coloured Varieties from Inhibited Types Variation in Nature.

DETAILS OF CASES OF MENDELIAN INHERITANCE IN COLOUR-
VARIETIES 155

Details of Colour-inheritance in more Important Cases In Antir-
rhinum, Lathyrus, Matthiola, Mirabilis, Phaseolus, Pisum, Primula,
Solanum, Zea, and many others.

CONNECTION OF FLOWER-COLOUR WITH THE PRESENCE OF ANTHO-
CYANIN IN VEGETATIVE ORGANS, FRUITS AND SEEDS 180

Examples from Cases Investigated.

HETEROZYGOUS FORMS 183

Examples from Cases Investigated.

COLOUR FACTORS IN REDUPLICATION SERIES .... 185

Phenomenon of Coupling and Repulsion in Lalhyrus and Primula
Colour-factors involved.

PATTERN IN COLOUR VARIATION 188

Examples from Cases Investigated Differentiation of Kinds of
Pattern.

x CONTENTS

PAGE

STRIPED VARIETIES AND BUD-VARIATION 190

Occurrence of Striping in Various Genera De Vries' Eversporting
Varieties Inheritance of Striping in Antirrhinum and Zea Emerson's
Explanation of Results Connection between Striping and Bud-variation
Bud-variation due either to Mendelian Segregation or Somatic Muta-
tion.

THE EFFECT OF OUTSIDE FACTORS ON COLOUR-VARIATION . 200

Experiments of Molisch on Hydrangea Cultures with Iron and Alu-
minium salts Effect of Insolation on Flower-colour.

CONNECTION BETWEEN COLOUR AND OTHER PLANT CHARACTERS 204

Colour and Vigour Colour and Flavour Colour and Hoariness in
Matthiola Colour and Hooded Structure in Lathyrus.

THE CHEMICAL INTERPRETATION OF FACTORS FOR FLOWER-
COLOUR 207

Chemical Interpretation of Factors in Antirrhinum Reversible
Reactions Chemistry of Different Kinds of Factors Interpretation
of Reddening, Blueing, Diluting and Deepening Factors.

APPENDIX 218

Further Details of Researches of Willstatter on Flower- and Fruit-
pigments Analyses of Pigments of Colour- varieties Bearing on
Mendelian Problems.

BIBLIOGRAPHY 227

INDEX 305

PART I

GENERAL ACCOUNT

Idea 27. And first, their Colours; where, with respect to several Plants and
Parts, they are more Changeable; as Red, in Flowers; or Constant, as Green in
Leaves, Which, with respect to several Ages of one Part, are more fading, as Green
in Fruits; or durable, as Yellow in Flowers. In what Parts more Single, as always
in the Seed; or more Compounded, as in the Flower; and in what Plants more
especially, as in Fancy. Which proper to Plants that have such a !Fas<e or Smell,
as both, in White Flowers, are usually less strong. To Plants that flower in such
a Season, as a Yellow Flower, I think, chiefly, to Spring Plants. And to Plants
that are natural to such a Soil or Seat, as to Water-plants, more usually, a white
Flower. What, amongst all Colours, more Common to Plants, as Green; or more
Rare, as Black. And what all these Varieties of Colours are upon Cultivation, but
chiefly, in their natural Soil. To observe also with their superficial Colours, those
within: so the Hoots of Docks, are Yellow; of Bistort, Red; of .4ve%s, Purple; but
of most, White. Where the Inward, and Superficial Colours agree; as in Leaves;
or vary, as in the other Parts frequently. And in what manner they are Situated ;
some universally spreading, others running only along with the Vessels, as in the
Leaves of Red Dock, and the Flowers of Wood-Sorrel.

From Nehemiah Grew's Anatomy of Plants, with an Idea of a Philosophical
History of Plants. 1682.

PART I

GENERAL ACCOUNT
CHAPTER I

INTRODUCTORY

The subject of the anthocyanin pigments of plants has received
considerable attention, yet the results are comparatively slight. Such
colouring matters have sometimes been spoken of as soluble pigments,
since they are in a state of solution in the cell-sap, as contrasted with
those which are in some way bound up with the structure of organised
protoplasmic bodies known as plastids. The innumerable shades of
blue, purple, violet, mauve and magenta, and nearly all the reds, which
appear in flowers, fruits, leaves and stems are due to anthocyanin pig-
ments. On the other hand, green, and the large majority of orange
and yellow colours, are in the form of plastid pigments. Other colours,
again, notably some scarlets and orange-reds, browns and even black,
are produced by the presence of both plastid and anthocyanin pig-
ments together in the tissues.

For almost a hundred years botanists have employed one term to
denote the soluble pigments anthocyanin 1 (avOos, /cvavos), a word
first coined by Marquart (5) 2 for these substances in 1835 and retained
in the same sense to the present day ; other rival terms, now obsolete,
such as erythrophyll, cyanophyll and cyanin have also been used
from time to time.

Vegetable pigments served as matter for investigation at a very
early date, and the property which first attracted attention was their
behaviour towards acids and bases, that is, the reddening by acids
and the formation of green coloured products with a base. Thus, in
1661, in the Experiments and Considerations touching Colours of Robert

1 In the present volume anthocyanin is largely used in the collective sense, that is
as a term which we know to include a number of substances ; sometimes however, when
the context demands it, the plural form is used.

2 Numbers in brackets refer to papers, etc., in the Bibliography at the end of the book,
and such references are either entirely concerned with anthocyanins, or have some direct
bearing upon them. References, on the contrary, in the foot-notes are not directly
concerned with anthocyanins.

W. P 1

2 INTRODUCTORY [CH.

Boyle (107), we find the following directions: 'Take good Syrrup of
Violets, Impregnated with the Tincture of the flowers, drop a little
of it upon a White Paper... and on this Liquor let fall two or three drops
of Spirit either of Salt or Vinegar, or almost any other eminently Acid
Liquor, & upon the Mixture of these you shall find the Syrrup imme-
diately turn'd Red,... But to improve the Experiment, let me add
what has not... been hitherto observ'd,... namely, that if instead of
Spirit of Salt, or that of Vinegar, you drop upon the Syrrup of Violets
a little Oyl of Tartar per Deliquium, or the like quantity of Solution
of Potashes, & rubb them together with your finger, you shall find the
Blew Colour of the Syrrup turn'd in a moment into a perfect Green."

And again in The Anatomy of Plants of 1682, Nehemiah Grew (1)
tells us that " ...Spirit of Harts Horn droped upon a Tincture of the
Flower of Lark-heel & Borage turn them to a verdegreese Green. 7 '

" S'pirit of Sulphur on a Tincture of Violets turns it from Bleiv to
a true Lacke, or midle Crimson."

"Spirit of Sulphur upon a Tincture of Clove- July -Flowers makes
a bright blood Red. Into the like Colour, it hightens a Tincture of
Red Roses."

It is not clear from the writings of Boyle and Nehemiah Grew
whether these authors considered the great variety of shades to be the
expression of one or of many substances, though it is clear that Nehemiah
Grew differentiated between red, blue and green pigments as regards
their solubilities. For he says, "...The Liquors I made use of for this
purpose, were three, sc. Oyl of Olives, Water, & Spirit of Wine. The
Water I used was from the Thames, because I could not procure any
clear Rain Water, & had not leasure at present to distill any. But
next to this, that yields as little Salt, as any." The oil he found dis-
solved the green colour, " ...But there is no Vegetable yet known which
gives a true Red to Oyl, except Alkanet Root." This root "which
immediately tinctures Oyl with a deeper Red, will not colour Water
in the least. Next it is observable, That Water will take all the Colours
of Plants in Infusion except a Green. So that as no Plant will by
Infusion give a perfect Blew to Oyl; so their is none, that I know of,
which, by Infusion will give a perfect Green to Water 1 ." Boyle, on the
other hand, was evidently impressed by the uniformity of the acid

1 This is entirely in accordance with our present-day knowledge of the pigments.
Chlorophyll, like all plastid pigments, is insoluble in water, but soluble iu oil. The red,
blue and purple anthocyanin pigments are soluble in water, but insoluble in oil. The
red pigment of Alkanet root is not an anthocyanin and has quite different solubilities.

i] INTRODUCTORY 3

and alkali reactions: "...it is somewhat surprizing to see,. ..how
Differingly-colour'd Flowers, or Blossoms,... how remote soever their
Colours be from Green, would in a moment pass into a deep Degree of
that Colour, upon the Touch of an Alcalizate Liquor." The gradual
evolution of the idea of the multiplicity of pigments included under
anthocyanin can only be realised in a comprehensive survey of the
subject.

After the preliminary and somewhat diffuse observations of the
earliest scientists, there follows a period during which certain definite
lines of investigation emerge, and it is practically to these lines that
the chapters in this book correspond. It may not be out of place,
however, to give a general account of the different phases and kinds of
investigation showing how they have developed, and how they are
related to each other in the complete history.

The above lines of investigation may be enumerated thus: the
morphological and histological distribution of the pigments, the factors
controlling their formation, their function, their chemical composition,
their mode of origin, and lastly, the part they have played in heredity.

First let us consider their distribution. Contributions to this portion
of the subject naturally formed a great part, though by no means the
whole, of the earlier work on pigments, since it merely involved general
observation and microscopical examination of petals and leaves.
Various writers published full accounts of the pigments of flowers,
fruits, leaves, etc., showing how some colours are due to plastid, others
to soluble pigments, and others again are the result of the combination
of both in the cells. As would be expected, this form of investigation
has tended to diminish in later years for, the histological basis once
laid down, the physiological, chemical and biochemical aspects have
come to the front. Mention must be made, however, of the publications
of Buscalioni & Pollacci (17) in 1903 on the histological distribution
of anthocyanin, and of a paper on similar lines read by Parkin (77)
at the meeting of the British Association of the same year. Still
later work of this kind is that of Gertz (19) which appeared in 1906;
this author made a most thorough and systematic investigation of the
occurrence of anthocyanin in representative genera of all natural
orders, but the publication of his work in Swedish unfortunately
restricts the circulation of his results.

The question of factors controlling anthocyanin formation is the
next line of enquiry most convenient for consideration. The main
factors concerned are light, temperature and nutrition.

12

4 INTRODUCTORY [CH.

The matter of fundamental significance in all questions of factor
control, and one which it will be well to understand before turning
to the factors in detail, is the dependence of anthocyanin formation
upon the supply of an initial product or chromogen, as it is called,
from which the pigment arises. This substance, like all others in plant
metabolism, ultimately depends for its existence on a supply of carbo-
hydrates (sugars), the first products of synthesis in the plant. The
relationship between anthocyanin formation and the presence of sugars
in the tissues has provided an important subject for research. Overtoil
(333, 334), in 1899, first drew attention to its significance 1 ; this
botanist had noticed, while carrying out some experiments on osmosis,
that culture of Hydrocharis Morsus-ranae in sugar solutions leads to
greater production of anthocyanin in the leaves. Further experiments
showed the phenomenon to be constant for quite a number of species
when isolated leaves and twigs were fed on solutions of cane sugar,
dextrose, laevulose and maltose. Repetition of experiments on these
lines at later dates by Katie (354), Gertz (386) and others gave full
confirmation to Overtoil's results. This discovery led Overton to the
fairly obvious inference that possibly, in the normal plant, reddening
of leaves, etc., is correlated with excess of sugar in the tissues, and he
states that by tests upon red autumnal leaves he could detect more
sugar in red than in green leaves.

More elaborate and conclusive work in this direction was commenced
by Combes in 1909 and carried on in the following years. Combes
(374, 385) had noted that decortication in some plants brings about
a considerable development of anthocyanin in the leaves above the
point of decorticatiou. Analyses were made by him, not only of leaves
reddened through this cause, but also of autumnal and other red
leaves. In all cases he (207) claims to have shown that the red leaves
contain greater quantities of sugars and glucosides than green ones
from the same plant. From other, more general, phenomena, in
addition to Combes's results, it may be safely inferred that an accumu-
lation of such synthetic products as are manufactured in the leaves
leads to production of anthocyanin. For example, we frequently find
abnormal reddening of a single leaf on a plant otherwise in full vigour,
and investigation almost invariably shows the reddening to be accom-
panied by injury, and the injury, whether it be due to mechanical
cutting or breaking, or to the attacks of insects, will be found to

1 The connection between pigmentation and the presence of sugars was also pointed
out in 1899 by Mirande (332) as a result of his investigations on the genus Cuscuta.

i] INTRODUCTORY 5

affect those tissues which conduct away the synthetic products of
the leaf.

There is little doubt that the chromogen of anthocyanin, in the
form of a glucoside, is also manufactured in the leaves, and the whole
trend of results goes to show that an enforced accumulation of this
chromogen, together with sugars, by stoppage of the translocation
current, leads to formation of anthocyanin in the leaf; and a similar
result may arise from artificial feeding with sugars. To the more
precise relationships between chromogen, anthocyanin and sugars
attention will be given in a later chapter.

Thus it will be seen that in all problems connected with effects
of temperature, light, etc., on anthocyanin formation, we are confronted
with two distinct questions, i.e. the direct effect of light and temperature
on the actual production of pigment, and the indirect effect of these
factors on the supply of organic compounds from which the chromogen
of anthocyanin is synthesised.

The relationship between pigment formation and light constitutes
a problem to which there is no very satisfactory solution. Sachs
(269, 271), Askenasy (282) and others have tried the obvious methods
of growing plants in the dark with controls in the light, of darkening
leaves while leaving inflorescences uncovered and so forth. The
outcome of these researches, as well as of several others, has been to
show that in many cases, for example, in flowers of Tulij)a, Hyacinthus,
Iris and Crocus, anthocyanin develops equally well in the dark ; in other
cases, such as Pulmonaria, Antirrhinum, and Prunella, the development
is feeble or absent. A general survey of anthocyanin distribution leaves
us in no doubt that, as far as organs where anthocyanin may be expected
to develop are concerned, the greatest production takes place in the
most illuminated parts. But we have on the other hand not a few
examples, of which the root of Beta is a good illustration, of development
of pigment in total darkness. In the absence of fuller evidence, the
most reasonable point of view is that the actual process of pigment
formation may in itself be entirely independent of light, should the
tissues contain sufficient reserve materials to supply the chromogen.
But if there is a shortage of reserve materials, such as would arise from
diminished photosynthesis, the anthocyanin may fail to appear from
lack of chromogen.

The problem of the effect of temperature offers similar difficulties.
Does the temperature influence the actual formation of pigment, or
is it again an indirect cause, making itself felt only through its effects

6 INTRODUCTORY [CH.

upon the supply of materials from which the pigment is synthesised?
That low temperature favours pigment formation would seem to be
demonstrated by autumnal coloration, and the winter reddening of
leaves of Hedera, Ligustnim, Mahonia and other evergreens. Conversely,
Overton (333) found in Hydrocharis, the higher the temperature, the
less anthocyanin. Klebs (360) also notes that flowers of Campanula
and Primula may be almost white in a hot-house, but the same indivi-
duals kept in the cold will bear coloured flowers. The consideration
of temperature is perhaps more difficult than that of light; for low
temperature, on the one hand, retards photosynthesis by which sugars
are formed, but, on the other hand, it also retards growth, starch forma-
tion and probably translocation, thereby tending to raise the sugar
contents of the tissues. High temperature, on the contrary, accelerates
growth and respiration, and consequently prevents the accumulation
of any excess of synthetic products.

An interesting application of these views upon light and tempera-
ture effects can be made in the case of Alpine flower coloration. The
subject has been extensively studied by Gaston Bonnier (288, 289,
294, 307, 328), Flahault (288, 289, 290, 292) and others, and has had
a great vogue with the writers on flower coloration in connection with
insect pollination. The special features of the case are intensity of
flower-colour and the formation of anthocyanin in the vegetative parts.
Gaston Bonnier & Flahault have compared individuals grown at heights
of 2300 metres with individuals grown in the plains, and have found
that the latter produce paler flowers and less anthocyanin in the leaves
and stems. It seems most reasonable to suppose that these phenomena
form a natural demonstration of some of the relationships which we
have just been considering between colour and factors. High Alpine
plants are stunted in growth, i.e. little material is expended vegetatively ;
they are exposed by day to intense insolation while the night tempera-
ture is low. One may therefore suppose photosynthesis to be very
active, whereas translocation and growth are retarded, these being
conditions which favour high sugar and chromogen concentrations in
the tissues and resultant abundance of pigment.

The so-called functions of anthocyanin have provided material for
another main line of research. Two essentially different types of
function are readily distinguished, the biological and the physiological.
The biological function is a subject which, in itself, needs extensive
treatment, and does not lie within the scope of this book. It is solely
connected with the attractive value of the coloured floral organs for

i] INTRODUCTORY 7

pollination by insects, and the subsidiary question of the attractive
value of ripe pigniented fruits for dispersal by birds. The relationship
between flower-colour and entomophily has received great attention
from botanists, and the whole matter is dealt with most thoroughly
in Knuth's Handbook of Flower Pollination (426), which includes an
excellent bibliography.

The physiological function is a very difficult and far less satisfactory
matter. Several different functions of a physiological nature have
been attributed to anthocyanin. One of the most famous is the screen
theory, the idea of which was first based on work published in 1880 by
Pringsheim, who showed that chlorophyll was bleached by intense light,
but not if protected artificially by a red screen. Thus the view arose
that anthocyanin might be protective in function, but experimental
evidence does not altogether favour this hypothesis. For, in 1885,
Reinke pointed out that it is those rays absorbed by chlorophyll which
have the greatest destructive effect on chlorophyll, and Engelmann
(394), in 1887, demonstrated that the absorption spectrum of antho-
cyanin is on the whole complementary to that of chlorophyll. Hence
anthocyanin absorbs those rays which are least harmful to chlorophyll,
and cannot therefore be said to provide an effective screen. A second
suggestion, brought forward by Stahl (405) in 1896, and largely supported
by him, is that anthocyanin absorbs certain of the sun's rays, and by
converting them into heat, raises the temperature of the leaf, and this
may serve to accelerate transpiration in difficult circumstances, as in
damp regions of the tropics, or may protect leaves from low temperature
as in Alpine regions. The chief points in favour of this hypothesis are
the distribution of anthocyanin in leaves of shade-loving plants, and the
fact, observed also by Stahl, and confirmed in 1909 by Smith (420), that
the internal temperature of red leaves is greater than that of green.

The next line of investigation, the chemical composition of the
pigments, is also difficult, and though spasmodically attacked from
time to time, met with no very serious consideration till 1906. So
intimately connected with its chemical composition that it can scarcely
be considered separately, is the question of the mode of formation of
anthocyanin, that is, the chemical reactions involved in the process.
Closely connected also, though in a lesser degree, is the part played by
anthocyanin pigments in heredity. It is proposed therefore to deal
with these three lines of research more or less together.

As already pointed out, the reactions with acids and alkalies are
the most obvious and striking chemical properties of anthocyanins,

8 INTRODUCTOKY [OH.

and they have helped to draw the attention of chemists to the subject;
for much of the earlier chemical work on anthocyanin, notably of a
group of French chemists (1800-1825), Braconnot (110), Payen &
Chevallier (113, 114, 115) and Roux (116), centred round these reactions,
especially in some cases round their role as indicators. But the idea
of anthocyanin as an indicator was fully conceived long before 1807
by Robert Boyle (107). '' When," he says, " we have a mind to examine,
whether or no the Salt predominant in a Liquor or other Body, wherein
'tis Loose and Abundant, belong to the Tribe of Acid Salts or not...
if such a Body turn the Syrrup of a Red or Reddish Purple Colour,
it does for the most part argue the Body (especially if it be a distill'd
Liquor) to abound with Acid Salt. But if the Syrrup be made Green,
that argues the Predominant Salt to be of a Nature repugnant to that
of the Tribe of Acids." After the reactions of anthocyanin with acids
and alkalies, other reactions were noticed with iron salts and various
reagents, many of which modify the colour as it is modified in nature.
These reactions gave rise to views among some chemists, Fremy & Cloez
(126) and Wigand (136), that natural blue, purple and red pigments
are modifications of the same substance, brought about by the presence
of other compounds in the cell-sap. But as analyses and investigations
proceeded, the view of a certain multiplicity of pigments gained the
ascendency.

One of the first actual analyses of anthocyanin was carried out
in 1849 by Morot (122) who isolated the blue pigment of the Corn-
flower, and found it to contain carbon, hydrogen and oxygen, with
nitrogen as impurity. Ten years later Glenard (129, 130) isolated the
pigment of wine, found it also to contain carbon, hydrogen and oxygen,
and gave it a percentage formula.

An early suggestion as to the chemical nature of anthocyanin and
its mode of formation was that of Wigand (136) in 1862. This author
suggested that anthocyanin arises by the oxidation of a colourless
tannin-like chromogen, a substance widely distributed in plants and
giving a green reaction with iron salts and a yellow reaction with alkalies.
The same substance obviously had been noted at an earlier date by
Filhol (125, 132) who observed it to be widely distributed, and main-
tained that the green coloration of anthocyanin with alkalies was due
to a mixture of a blue anthocyanin reaction plus the yellow reaction
of these accompanying substances; a view also held by Wiesner (135).
The idea of the formation of anthocyanin from a tannin-like chromogen
by oxidation was, from this time onward, generally accepted by botanists

i] INTRODUCTORY 9

and chemists. It was, for instance, again (1878) suggested by Gautier
(149) on the strength of isolation and analyses of the pigments of various
wines. Later, further analyses of Vine pigments were made by Gautier
(175) and other investigators, notably Heise (167), but their results
were in no way concordant except in so far as they agreed that the
pigment contained carbon, hydrogen and oxygen only. These were
followed again by analyses of the pigment of Bilberry fruits by Heise
(178).

In 1895 Weigert (179) made the first attempt at a classification of
anthocyanins based upon their chemical reactions, and he differentiated
two groups, the 'Weinroth' and the 'Riibenroth,' according to the
behaviour of the pigment to acids, alkalies and lead acetate solution.
The Riibenroth group is however a small group of anthocyanin pigments
of the orders Phytolaccaceae, Amarantaceae and Portulacaceae of the
Centrospermae ; the Weinroth group, including the greater number
of anthocyanins, still remained undifferentiated.

In 1906 Grafe (197) published the results of some careful analyses
of the pigment of the Hollyhock (Althaea rosea), followed by further
analyses (222) of the pigment of scarlet Pelargonium flowers in 1911.
In this publication Grafe clearly expresses the view that anthocyanin
must be regarded as a general term for a number of substances, having
similar properties and reactions, but differing somewhat from each
other in chemical constitution.

Meanwhile from the year 1900 and onwards, the study of anthocyanin
pigments gained a new impetus through their connection with the
inheritance of colour in plants. For many experiments on the cross-
breeding of colour- varieties were commenced about this time by Bateson,
Punnett, Saunders (475, 487) and others working on the lines of Mendelian
inheritance. Two of the first plants to be used for this purpose were the
Stock (Matthiola) and the Sweet Pea (Lathyrus). Both these plants
have a number of varieties, and although it is doubtful whether we know
the character of the original type in Matthiola, it is fairly certain that
the original wild Sweet Pea resembled the form known as 'Purple
Invincible' which has a chocolate standard and purplish-blue alae.
We now know from the teachings of heredity that all the coloured
varieties of Sweet Pea, and probably many of those of Stocks, have
arisen through the loss of certain factors from the type, and by recrossing
selected varieties, the type can be obtained again, the process being
known as the phenomenon of 'reversion to type.' One of the most
striking results (496) of the experiments on Sweet Peas and Stocks

10 INTRODUCTORY [CH.

was the phenomenon of the production of a plant bearing purple flowers
(in the Sweet Pea the original type) by crossing two white-flowered
strains. From such results the hypothesis arose, that red colour
(anthocyanin) is due to the presence of two factors, C and R, in the
plant, the loss of either factor resulting in an albino. A third factor,
B, by its presence modifies the red colour to a blue or purple, but
unless C and R are also both present in the plant, B is without effect.
Hence white plants of known parentage can be selected which are
carrying C, R, or B only, or CB, or RB.

Such a discovery is not only highly important in its connection
with heredity, but it also provides us with well-defined material for
the solution of the problem as to what chemical processes are involved
in anthocyanin formation. Conversely, these processes once discovered,
we should also be provided with a chemical interpretation of the
Mendelian factors for flower-colour.

The problem was first attacked by the author in 1909. The plant
selected for investigation was Antirrhinum, majus, since the inheritance
of the flower-colour of this species had already been worked out (514,
535) during the years 1903-9, and confirmed independently by Baur
(536). Antirrhinum majus presents a case of singular interest. In
this species (as in Matthiola and Lathyrus) two white varieties, or more
strictly two varieties which are albinos as regards anthocyanin, when
crossed together, produce the magenta pigment (anthocyanin) character-
istic of the type; but, whereas in Lathyrus and Matthiola the two
whites producing a purple are identical in appearance, the two albinos
in Antirrhinum can be distinguished at sight. One, which is known
as ivory, is ivory-white in colour, the other, the true white, is dead
white. These two varieties must obviously, between them, contain
the chemical substances, or the power to produce the chemical sub-
stances, essential to the formation of anthocyanin.

The first clue to the solution was provided by the difference in the
action of ammonia vapour on the two flowers. The ivory exposed to
ammonia turns bright yellow whereas the white is practically unaltered.
It has been mentioned previously that Filhol (125), and in fact several
of the earlier workers on plant pigments, had noted the occurrence of
colourless substances in the plant which gave a bright yellow reaction
with alkalies. Moreover, a tentative suggestion was made by Bidgood (18)
some years ago (1905) in the Journal of the Royal Horticultural Society,
that these substances were the colouring matters known as flavones.
Reference to the work of A. G. Perkin on the flavones provided the

i] INTRODUCTORY 11

information that these substances have a general distribution in nature,
and had attracted the notice of chemists on account of their value
as vegetable dyes. Some flavones have been found in many species,
others in but one, or a few, though this appearance of restriction may
be due to lack of information. All members of the group are yellow
substances differing slightly from each other in their chemical composi-
tion. In the plant they frequently occur as glucosides, in which
condition their colour is much paler, and their solubility greater, than
in the free state. All give intense yellow or orange-yellow coloration
with alkalies, and green or brown coloration with iron salts. This
information led the author to conclude that the widely distributed
substances, occurring in all plant tissues which give the yellow reaction
to ammonia vapour, are flavones, and that the pigment of the ivory
variety of Antirrhinum is of the same nature. The white variety of
Antirrhinum is without the flavone, and must contain some substance
capable of action on the flavone to form anthocyanin.

At about this period a series of papers were published by Palladin
(203, 210) setting forth what is known as the theory of 'Atmungs-
pigmente.' Broadly speaking, Palladin supposes plants to contain
chromogens which are capable of being oxidised by enzymes to pigments
which in turn pass on oxygen to respirable substances in the plant.
Among the ' Atmungspigmente ' Palladin includes anthocyanin.

Palladin's conception, together with the evidence provided by
Antirrhinum, led the author (212, 217, 226) to bring forward the hypo-
thesis that anthocyanin is formed from a flavone by the action of an
oxidising enzyme or oxidase. Thus the formation of magenta pigment
in Antirrhinum would be brought about by the action of an oxidase
produced by the white variety upon the flavone present in the ivory.
Likewise in every plant anthocyanin would be formed by the action
of an oxidase upon a flavone, and just as the flavones are a class of
substances having similar properties, though showing some variety in
structure, the anthocyanins also would have properties in common as
a class but would, many of them, differ somewhat in chemical structure.

The oxidase hypothesis of anthocyanin formation has been taken up
by Keeble & Armstrong (230, 231). These authors invented a micro-
chemical method for testing for oxidases in plant tissues, and applied
the method chiefly to colour varieties of the Chinese Primrose (Primula
sinensis), a plant largely used for Mendelian experiments. By means
of their test they discovered that oxidases in Primula are chiefly
confined to the epidermis and the bundle sheaths, both regions where

12 INTKODUCTOKY [OH.

anthocyanin also is localised. A further point of interest was discovered
in connection with certain varieties of Primula known as dominant
whites. These varieties have pigment (anthocyanin) in stems and
leaves, though the flowers are white, and they are regarded, from their
behaviour in crossing, as coloured forms in which the flower-colour
is inhibited by some factor present in the plant. Keeble & Armstrong
were able to show that flowers of these varieties gave ordinarily no
oxidase reaction, but after treatment with certain reagents, the inhibitor
was removed, and the oxidase reaction appeared in the petals. On
the other hand, they found that the true albinos in Primula gave the
oxidase reactions as well as the coloured varieties. Hence to make
the hypothesis fit the case of Primula, they were obliged to assume the
lack of colour in the true albino to be due to lack of chromogen, an
assumption which has never been verified.

Meanwhile the author's investigations on Antirrhinum proceeded on
the lines of isolation and analysis (244) of the pigments of the colour-
varieties. These pigments were prepared on as large a scale as possible
by making water extracts of the flowers, precipitating the pigment as
a lead salt by lead acetate, and again decomposing the salt with sulphuric
acid. Filtered from lead sulphate, the pigment was obtained in dilute
acid solution. It had been previously ascertained that anthocyanins
are largely present in the plant in combination with sugar as glucosides.
On boiling with acid, the glucoside is hydrolysed and the pigment,
which is less soluble than the glucoside, separates out. This method
was adopted for obtaining the pigment from the acid solution. The
pigment of the ivory variety of Antirrhinum was identified by Bassett
and the author with apigenin, a flavone of known constitution occur-
ring in Parsley (Apium Petroselinum). Of the anthocyanin-containing
varieties of Antirrhinum there are two, red and magenta, analogous
to the red and purple varieties of Lathyrus and Matthiola. Analyses
were made (254) of the red and magenta pigments prepared separately
from several different varieties, and the results were concordant. In
the case of both pigments the percentage of oxygen was greater than
that in the flavone apigenin. Determination of the molecular weights
of the red and magenta pigments, though not obtained to a high degree
of accuracy, indicated that the molecules of anthocyanin are consider-
ably larger than those of apigenin.

The conclusions drawn by Bassett and the author (254) were that
in Antirrhinum the anthocyanins are derived from the flavone apigenin
by oxidation, accompanied by condensation, possibly of two flavone

i] INTRODUCTORY 13

molecules, or possibly of a flavone molecule with other aromatic sub-
stances present in the plant.

The problem of anthocyanin formation has quite recently received
a most important series of contributions by Willstatter and a number
of his co-workers. In 1913 Willstatter (245) first published some
results on anthocyanins in general and the anthocyanin of the Corn-
flower (Centaurea Cyanus] in particular. Willstatter states that all
natural anthocyanins are present in the plant in the condition of gluco-
sides, and that many of them, moreover, including the anthocyanin of
Centaurea, are very unstable in water solution, and readily change in
these circumstances to a colourless isomer. The change can be prevented
by adding certain neutral salts, and also acids, to the pigment solution.
The explanation, according to Willstatter, of these phenomena lies in
the fact that anthocyanin is an oxonium compound having a quinonoid
structure and containing tetravalent oxygen. The quinonoid structure is
rendered stable by the formation of oxonium salts with acids or with
neutral salts, such as sodium nitrate and sodium chloride. The pig-
ment itself, in the neutral state, is purple in colour, and has the structure
of an inner oxonium salt. With acids it forms red oxonium salts, and
with alkalies blue salts, the position of the metal being undetermined.
More recently Willstatter (256, 257) has published the results of the
isolation and analyses of a number of other anthocyanins in addition
to that of Centaurea, i.e. from the flowers of Delphinium, scarlet Pelar-
gonium, Mallow and Hollyhock; from Grapes also, and the fruits of
the Bilberry and Cranberry. In every case crystalline salts were
obtained with hydrochloric acid, and from the analyses of these salts
Willstatter obtained a series of percentage formulae for his antho-
cyanins. Compared with the percentage formulae of the known
flavones, he found that each anthocyanin could be derived theoretically
from a flavone by loss of an atom of oxygen, i.e. by reduction.

The view of anthocyanin formation by reduction from a flavone
has received support from a series of phenomena of a very curious
and interesting nature, which were first observed by Keeble & Arm-
strong (240), and of which an account was published in June, 1913.
These observations were concerned with the effect of reducing plant
extracts with nascent hydrogen. A number of flowers, among which
were the yellow Wallflower, Daffodil, Crocus and Polyanthus, w r ere
extracted with alcohol, and the extracts acidified and treated with
zinc dust. Filtered from the zinc, and exposed to air, the solution
develops a red colour which changes to green on addition of alkali.

U INTRODUCTORY [CH.

Thus Keeble & Armstrong were led to believe that they had made
artificial anthocyanin, and that preliminary reduction and subsequent
oxidation are the essential processes of anthocyanin formation. In
November of the same year, a note was published by Combes (234)
in the Comptes Rendus, giving a record of practically the same observa-
tion though in a much more complete form. By treating with sodium
amalgam the acid alcohol solution of a yellowish crystalline substance
which he had extracted from Ampelopsis leaves, he obtained a fine
purple pigment which crystallised in needles. This pigment, he claims,
on the basis of chemical and physical properties, to be identical with
crystals of natural anthocyanin from the same plant; but no analyses
are included. In addition, he found (235) that natural anthocyanin,
after treatment with hydrogen peroxide, gives a yellow product identical
with the natural yellow substance from which he started.

Combes believes that these results entirely revolutionise our ideas on
anthocyanin formation. Thus he says: "La production experimentale
d'une antliocyane pent done etre consideree comme realisee. Ce resultat
permet d'entrevoir comme tres proche la solution du probleme de la
formation des pigments anthocyaniques pose depuis plus de 120 ans
et qui fut aborde par de nombreux physiologistes. On salt que dans
toutes les hypotheses relatives a cette question, formulees depuis 1825,
la pigmentation a toujours ete consideree comme un phenomene d'oxy-
dation; cette opinion ne peut plus etre soutenue, puisqu'il apparait
que 1'anthocyane des feuilles rouges prend naissance lorsque le compose
correspondant contenu dans les feuilles vertes est soumis a 1'hydrogene
naissant, c'est-a-dire dans un milieu qui est au contraire reducteur."

The essential difference between Keeble & Armstrong's results and
those of Combes can be explained in the following way. The purple
pigment, as indeed the natural anthocyanin, will become colourless
on powerful reduction, though it regains its colour in air. Thus
Keeble & Armstrong's reduction was carried beyond the preliminary
formation of pigment to the colourless stage, the colour returning
again on exposure to air, whereas Combes's product was not reduced
so far.

The production of a purple colour on reduction of flavone with
sodium amalgam had been known to chemists some time before the
work of Keeble & Armstrong. The almost universal distribution of
flavones naturally explains the appearance of the colour on reduction
of many plant extracts. The crystalline yellow product used by
Combes is undoubtedly a flavone also, as will be seen on referring to

i] INTRODUCTORY 15

his original communications. The question at issue is whether these
purple pigments are anthocyanins.

The problem of the artificial formation of anthocyanin has been
attacked by Willstatter (257) also in 1914. By treating quercetin,
according to Combes's method, Willstatter in the same way obtains
a purple pigment, the bulk of which he finds is unstable in acid solution
on boiling, though a certain amount of colour remains after this process.
The product unstable to heat is, in Willstatter's opinion, not a true
anthocyanin, but he maintains that the small amount of stable product
left after heating is, when isolated, identical with the anthocyanin
(cyanidin) of the Cornflower. Thus Willstatter regards this reaction
as giving further confirmation to his views of the reduction hypothesis
of pigment formation, which we have seen to be based upon analyses
of the anthocyanins he has isolated, and to be practically a proof of
his suggested constitution of anthocyanin. To quote from his latest
paper: "Die Bildung von Cyanidin aus Quercetin hat zweifache
Bedeutung. Es ist dadurch eine Synthese von Cyanidin ausgefiihrt,
da das Quercetin selbst vor zehn Jahren von St. von Kostanecki...
synthetisch dargestellt worden ist. Ferner wird durch diese Umwand-
lung die Konstitutionsformel des Cyanidins bewiesen."

The author has been so far unable to bring the results provided by
the case of Antirrhinum into line with Willstatter's views. In Antir-
rhinum the chromogen is, without doubt, apigenin ; when treated with
nascent hydrogen this flavone produces a purple pigment, yet the latter
does not resemble the natural anthocyanin of Antirrhinum, either in
composition or properties (255).

Thus we have arrived at the last stage in the subject as far as inves-
tigation has gone. There is little doubt that anthocyauins are derived
from flavones, but further evidence is needed of the actual chemical
connection between these two classes of substances. Neither of Will-
statter's two lines of proof is beyond criticism. Actual comparison
of the percentage formula of any anthocyanin with any flavone,
though significant, does not prove their relationship in nature, by
reduction or otherwise, unless we know that they occur together. For
since the known flavones can be arranged in a series, each member
of which differs from the others by an atom of oxygen, any antho-
cyanin might be either an oxidation or a reduction product of a flavone.
What is really necessary is to know which flavone accompanies the
anthocyanin in a number of plants. At present Antirrhinum alone
supplies us with information, for as far as it is possible to judge,

16 INTRODUCTORY [OH. i

Willstatter did not in any case extract the flavone accompanying the
anthocyanin he isolated. The second line of evidence, i.e. the formation
of a very small quantity of a purple pigment, as a by-product, during
the reduction of a large quantity of quercetin, might quite well involve
reactions other than reduction.

Willstatter's views 1 provide a very different interpretation of Men-
delian factors for coloration from that of Keeble & Armstrong and the
author. For we must suppose, should Willstatter be correct, that the
factors for colour are the chromogen (flavone) and the power to reduce
the chromogen with a complete change of structure from the pyrone
grouping of the flavone to the quinonoid structure of anthocyanin.
A purple flower would contain the pigment in its neutral form; a
reddening factor would be the power to produce acid cell-sap ; con-
versely, a blue variety must have alkaline sap. Thus we have a reversion
to the ideas of the earliest writers, and it is less easy to correlate such
a view with the results of cross-breeding. But there is no doubt that
Willstatter's researches, whether his interpretation prove to be the
right one or not, have brought the whole problem of soluble pigments
within measurable distance of explanation. Such a solution, moreover,
will provide what is perhaps of greater importance, namely, a chemical
basis for one series, at any rate, of Mendelian phenomena.

1 See also Appendix.

CHAPTER II

THE MORPHOLOGICAL DISTRIBUTION OF ANTHOCYANINS

It is a striking fact that anthocyanins are almost universally distri-
buted. This is not fully realised until a systematic investigation is made,
and then- it is seen how few plants are entirely without these pigments.
Many plants may appear to the casual observer to be free from antho-
cyanin, but a closer examination generally reveals its presence, often
only in minute quantity, in such places as the base of the stem, the
petioles, bud-scales, bracts or even in some unexpected organ, as for
instance, the anthers and stigma. Again, an abnormal condition of
drought, or fungal attack, may cause its appearance when otherwise
the healthy plant is green. A number of plants, moreover, only form
anthocyanin in their young leaves at the beginning, or in their old
leaves at the close, of the vegetative season, or pigment may develop
shortly before the death of the whole plant. Reviewing the flowering
plants as a whole, two orders present an anomaly with respect to the
distribution of anthocyanin. The power to form anthocyanin seems
to be absent, first, from the Cucurbitaceae, as far as it has been possible
to ascertain, and secondly, from certain genera of the Amaryllidaceae
(Leuoojum, Galanthus, Ornithogalum and Narcissus). This observation
is confirmed by Gertz (19), who also adds to the list, Herniaria (Caryo-
phyllaceae) and Chrt/sosplenium (Saxifragaceae), some species of
Potamogeton, Eleocharis (Cyperaceae), Kniphqfia, Aloe, Reseda, i.e. R.
odorala and R. lutea, and Buxus. To the statement as regards R. odorata
and R. lutea exception must be taken, since both these species form
anthocyanin in their leaves under adverse conditions, such as are
brought about by drought or injury.

The distribution of anthocyanin in the lower groups of plants has
been investigated by Gertz (19). It appears in the young fronds of
Osmunda regalis, of many species of Adiantum, of species of Doodia,
Pellaea and Davallia. Gertz also notes it in Blechnum and Azolla,
and among the Gymnosperms, in the cones of Picea, Pinus and Larix.

w. P. 2

18 THE MORPHOLOGICAL DISTRIBUTION [CH.

As regards the Augiosperms, the present chapter contains an enu-
meration of the organs in which anthocyanin is formed in the varying
circumstances of plant life. Some authors have attempted to classify
the appearance of anthocyanin into permanent and transitory or
periodical, normal and abnormal, etc., etc., but the variety and com-
plexity of the underlying physiological causes in each case render
such classifications of little value. No distinctions of this kind are
attempted in the following classification.

1. The leaves of the majority of plants are green; nevertheless,
in many genera and species, anthocyanin is always present as a normal
development, and a complete series might be found ranging from plants
whose leaves show a faint trace of anthocyanin to those whose leaves
are more or less heavily pigmented. Some writers, notably Kerner
(398), have pointed out that plants in which there is a considerable
development of anthocyanin have often a special habitat. It would
seem on the whole that this view is justified, though in many instances
red pigmentation appears to be a specific character bearing no special
relation to environment, as for example in Geranium Robertianum.

Kerner's views can be best expressed in his own words: "...antho-
cyanin frequently occurs only on the under side of foliage-leaves.
This is observed especially among plants in the depths of shady forests,
which, although belonging to widely-differing families, agree in a remark-
able manner in this one point. One group of these plants has thick,
almost leathery, evergreen leaves lying on the ground, which arise from
subterranean tubers, or root-stocks, or from procumbent stems. The
widely-distributed Cyclamen europ&um may serve as a type of this
group. . . .Amongst other species belonging to this group may be mentioned
Cyclamen repandum and C. hederifolium, Cardamine trifolia, Soldanella
montana, Hepatica triloba and Saxifraga Geum and cuneifolia. Growing
in habitats similar to these are to be met biennial, occasionally perennial,
plants which in autumn form a rosette of leaves on their erect stems
which survive the winter ; these are always coloured violet on the side
turned towards the ground, while the leaves which develop in the
following warm summer on the elongated flower-stalks usually appear
green below. To this group belong, especially, numerous Cruciferae
(e.g. Peltaria alliacea, Turritis glabra, Arabis brassicceformis) ; species
of spurge (e.g. Euphorbia amj/gdaloides), bell-flowers (e.g. Campanula
persicifolia), and hawkweeds (e.g. Hieracium tenuifolium). Finally,
deciduous shrubs are to be found in the depths and on the margins of
forests whose leaves do not survive the winter, but which produce

u] OF ANTHOCYANINS 19

on the stems developing in the summer flat leaves whose under side
contains abundant anthocyanin, as, for example, Senecio nemorensis
and nebrodensis, Valeriana montana and tripteris, Epilobium montanunt,
Lactucu muralis, and many others." And again, "That which occurs
in plants of the forest shade occurs similarly in those marsh plants
whose leaf-like stems or flat, disc-like leaves float on the surface of the
water. The green discs of duckweeds (e.g. Lemna polyrrhiza), of the
Frogbit (Hydrocliaris morsus-ranw), of the Yillarsia (Villarsia nym-
phuidcs), of water lilies (Nymphcca Lotus and thernialis), and of the
magnificent Victoria regia, are strikingly bi-coloured, being light-green
above and deep violet below." Occasionally anthocyanin is limited
to definite areas in the leaf, and in this form is the cause of the dark
brown or black spots on leaves (Orchis maculata, Arum maculatum 1 ,
Medicago maculata, Polygonum Persicaria 2 ), the dark colour being the
result of the mixture of purple pigment and green chloroplastids.

Most of the above examples have been taken from flora of the
temperate regions, but when we consider the tropical and sub-tropical
regions, cases of red pigmentation are not only more numerous but are
also much more striking. It is almost impossible to enumerate the
specie which have wonderfully variegated leaves, and constitute the
class of 'beautiful-leaved plants' of cultivation. Examples are quoted
by Hassack (393) in a paper on variegated leaves, and good plates -have
been published in a book by Lowe & Howard (29). A few of the most
remarkable are as follows : Calathea, Maranta (Marantaceae), Alocasia,
Caladium, Xanthosoma violaceum (Araceae), Tradescantia discolor
(Commelinaceae), many Bromeliaceae (Vriesia splendens, Nidularium
spp., Billbergia gigantea, Tittandsia), Dracaena spp. (Liliaceae), species
of Croton, Acalypha and Codiaeum (Euphorbiaceae), many Gesneriaceae
(Sinningia atropurpurea, Episcia sp., Aeschynanthus sp. and Gesneria
cinaberina), Coleus Verschaffeltii (Labiatae), Iresine sp., Alternanthera
versicolor (Arnarantaceae), Fittonia sp. (Acanthaceae) and numerous
Begonias. Hassack points out in a most interesting way how the colour
in such leaves may vary from purple or crimson, through brown-reds

1 In the case of Arum maculatum the leaves are not always spotted ; in fact, according
to the evidence of Colgan (73) and of Pethybridge (78), the spotted form is much less
common, both in Great Britain and Ireland, than the unspotted. These two authors
give some interesting data with regard to the distribution of both forms, as well as obser-
vations upon the structure and inheritance of the spot.

2 Garjeanne (74) mentions the discovery in Polygonum Persicaria of a variegated
form in which the chlorophyll is absent from large areas of the leaf, and when this is the
case, the anthocyanin patch shows bright red.

22

20 THE MORPHOLOGICAL DISTRIBUTION [en.

to browns and bronze-greens, according to the amount of anthocyanin
present, and its situation in the leaf, whether epidermal or deep-seated.

Stahl (405) lays great emphasis on the fact that many of these
variegated-leaved plants, and those with purple and red under-surfaces,
inhabit in the tropics, as in the temperate zones, the damp and shady
forest regions; he quotes examples of such in support of his views
upon the physiological significance of anthocyanin, to which we shall
return in Chapter vm. He notes that the orders specially represented
are Araceae, Marantaceae, Piperaceae, Begoniaceae, Melastomaceae,
Orchidaceae (Pogonia crispa) and Vitaceae (Cissus discolor) ; these
probably include many of the plants already mentioned above. Some
of these shade-loving plants are, in addition, characterised by a velvety
surface due to epidermal papillae, which emphasises the richness of
the variegation colour. Spotting and flecking with 'anthocyanin are
also common in tropical plants, and occur on a much larger scale than
in the temperate zones; instances are given by Hassack (393) and
Stahl (405), such as Gesneria cinaberina, Tradescantia zebrina, Musa
zebrina Van Houtte and Costus zebrinus.

Although the observations, which led Stahl to conclude that antho-
cyanin development is a characteristic of many shade-loving plants,
are undoubtedly correct, yet it is also true that when a plant is capable
of foaming anthocyanin, those individuals which inhabit dry and sunny
situations develop more anthocyanin than other individuals in moist
and shady positions. That is, a plant may be green-leaved in the
shade, and more or less red-leaved when exposed to the sun. This
fact is at once patent to any one who gives attention to the matter,
as the observations of F. Grace Smith (412) have shown. This author
examined plants from regions in Massachusetts and Maine, and of 285
plants showing anthocyanin, 150 were from dry sunny places, 61 from
dry shady places, 40 from wet sunny places and 34 from wet shady
places. It seems likely that the phenomenon of pigmentation in shade,
especially in the tropics, is a special adaptation, and of a different nature
from the more general reddening of vegetation in sunny situations,
and that some explanation may be found for the apparent contradiction
between the two phenomena. Though F. G. Smith's observations do
not give sufficient information to be conclusive, there is one significant
fact among them, namely, that in wet shady places, the percentage of
cases of red colour in the leaves is higher than that in the stems, whereas
among the plants as a whole, anthocyanin was more frequently found
in steins and petioles than in leaves. This may indicate that reddening

n] OF ANTHOCYANINS 21

in the shade is a peculiar function of the lamina, and is not connected
with general reddening of the plant.

In petioles, anthocyanin is more frequent than in leaves, and in
stems one might say that it is almost universal; the stem may be
entirely red, or only red at the basal internodes : or the nodes may be
red and the internodes green, or vice versa. A spotted effect produced
by the local distribution of anthocyanin is seen in some stems, for
example in Chaerophyllum temulum and Conium maculatum (Umbelli-
ferae).

Red pigment is also very characteristic of the bud-scales of trees
and shrubs; in some cases these scales represent leaf-bases (Aesculus
Hippocastanum, Acer pseudoplatanus, A. campestris}., in other cases,
stipules (Ulmus campestris, Tilia europaea). In addition the stipules
of herbaceous plants are often red, and the same statement holds good
for enlarged leaf -bases (Umbelliferae).

2. Anthocyanin is characteristic, though by no means universally
so, of young developing leaves and shoots. It is interesting to note
how in some species one finds both reddening of young leaves and
autumnal colouring, whereas in others this is not the case. There
are, in fact, four possibilities:

(a) Both anthocyanin in young leaves and autumnal reddening.
Species of Acer, Rosa, Crataegus and Rubus.

(6) Anthocyanin in young leaves but no autumnal reddening.
Corylus Avellana, Juglans regia, Fraxinus excelsior and Quercus Robur.

(c) No anthocyanin in young leaves but some autumnal reddening.
Aesculus Hippocastanum. This is a much rarer combination than (a)
or (fe).

(d) No anthocyanin in either young or old leaves. Fagus sylvatica 1 .
Anthocyanin is especially abundant in the leaves of certain tropical

trees (Cinnamomum, Haematoxylon, etc.) ; in addition to being coloured
red, the young leaves often hang vertically downwards and straighten
up again when mature (Bauhinia, Dryobalanops, Cinnamomum) 2 .
In other species, especially among the Caesalpineae group of the
Leguminosae, the whole young shoot has a vertical position, and in
this respect, together with its intense red colour, forms a striking
contrast to the older parts of the tree with green leaves in the normal
horizontal position. As instances, certain species are mentioned by
Stahl (405) and Keeble (403) : Amherstia nobilis with young leaves of

1 Pick (391) reports the presence of anthocyanin in young Beech leaves.
Willis, J. C., The Flowering Plants and Ferns t Cambridge, 1904, p. 157

22 THE MORPHOLOGICAL DISTRIBUTION [CH.

a bright red tint ; a variety of Cynometra cauliflora, bright rose-colour ;
Jonesia reclinata, golden-green to bright red; Brownea hybrida and
B. grandiceps, green spotted with red. Keeble further remarks that
the coloration of young foliage in low latitudes is of such general
occurrence that at the time of leaf-renewal, a tropiqal forest rivals in
its tints the autumnal forests of the temperate regions ; he also quotes
in confirmation the following extract from a paper by Johow (392),
writing of the Lesser Antilles: "...all at once a red tint, due to the
young foliage of the trees, appears in the landscape." Further details
of such cases are given by Stahl (62, 405), Ewart (406), Keeble (403),
Smith (420) and Weevers (421), and the physiological significance of
anthocyanin formation in young leaves is discussed in Chapter vm.

3. The older leaves of many herbaceous plants often acquire a
considerable formation of anthocyanin by the end of the vegetative
(not necessarily autumnal) season (Lilium candidum, Digitalis purpurea).
Sometimes the whole plant becomes red in a striking way (Ckaerophyllum
sylvestre, C. temulum, Galium Aparine).

4. The autumnal coloration characteristic of temperate climates
is one of the most pronounced cases of anthocyanin development.
With the onset of the cooler weather of autumn, anthocyanin appears
in quantity in the leaves of a number of trees, shrubs and climbing
plants (species of Acer, Rims, Euonymus, Crataegus, Viburnum, Cornus,
Vitis, Ampelopsis). In the Pflanzenleben, descriptions are given by
Kerner (398) of the beauty of the autumnal foliage in the forests along
the Rhine and the Danube, on the shores of the Canadian lakes of
North America and among the vegetation of the Alpine slopes. Kerner's
descriptions give such vivid pictures of autumnal colouring that this
opportunity is taken of quoting one of the passages from his account.
" The heights along the middle course of the Danube, for example, the
region known as the Wachan, below the town of Melk, shows wide
expanses of forest, in which beeches, hornbeams, evergreen oaks, common
and Norway maples, birches, wild cherries and pears, mountain ashes
and wild service-trees, aspens, limes, spruces, pines and firs take a share
in the greatest variety. Bushes of Barberry (Berber is vulgaris), Dogwood
(Cornus sanguined), Cornel (Cornus mas), Spindle Tree (Euonymus
Europceus and verrucosus), Dwarf cherry (Prunus Chamcecerasus), Sloe
(Primus spinosa), Juniper (Juniperus communis), and many other low
shrubs arise as undergrowth, and spring up on the margins of the forests.
The mountain slopes abutting on the valleys are planted with vines,
and near by grow peach and apricot trees in great abundance. In

n] OF ANTHOCYANINS 23

the meadows on the shore, and on the islands of the Danube, rise huge
abeles and black poplars, elms, willows, alders, and also an abundant
sprinkling of trees of the bird cherry (Primus Padus)....The first frosts
are the signal for the beginning of the vintage ; all is busy in the vine-
planted districts, and the call of the vine-dresser resounds from hill
to hill. But it is also the signal for the forests on the mountain slopes
and in the meadows to change their hues. What an abundance of
colour is then unfolded ! The crowns of the pines bluish-green, the
slender summits of the firs dark green, the foliage of hornbeams, maples,
and white-stemmed birches pale yellow, the oaks brownish-yellow, the
broad tracts of forest stocked with beeches in all gradations from
yellowish to brownish-red, the mountain ashes, cherries and barberry
bushes scarlet, the bird cherry and wild service trees purple, the cornel
and spindle-tree violet, aspens orange, abeles and silver willows white
and grey, and alders a dull brownish-green. And all these colours
are distributed in the most varied and charming manner. Here are
dark patches traversed by broad light bands and narrow twisted stripes ;
there the forest is symmetrically patterned; there again the Chinese
fire of an isolated cherry-tree or the summit of a single birch, with its
lustrous gold springing up among the pines, illuminates the green back-
ground. To be sure this splendour of colour lasts but a short time.
At the end of October the first frosts set in, and when the north wind
rages over the mountain tops, all the red, violet, yellow, and brown
foliage is shaken from the branches, tossed in a gay whirl to the ground,
and drifted together along the banks and hedges. After a few days
the mantle of foliage on the ground takes on a uniform brown tint,
and in a few more days is buried under the winter coat of snow." The
extreme beauty of the Alpine vegetation in autumn is also described
by Overton (333), and led him to investigate the cause of the coloration.

Detailed accounts of the physiological processes producing autumnal
coloration will be given in Chapter vi.

As in the case of reddening of young developing leaves, autumnal
coloration is by no means universal, since, as we have seen, there are
many trees and shrubs of which the leaves turn either yellow or brown
and show no trace of pigment (Fagus sylvatica, Carpinus Betulus,
Quercus Robur, Betula alba, Alnus glutinosa, Populus alba, P. nigra,
P. tremula). There is some indication of a general tendency among
genera of certain orders to develop red autumnal coloration, whereas
genera of other orders are without it. Thus, for instance, the trees
and shrubs of the Rosaceae (species of Pyrus, Primus, Rosa, Crataegus)

24 THE MORPHOLOGICAL DISTRIBUTION [CH.

very readily produce anthocyanin. On the other hand, the majority
of the trees in the Amentales do not form red pigment in the autumn
(species of Salix, Populus, Betula, Alnus, Juglans, Morus, Carpinus,
Corylus, Castanea, Fagus and Quercus). To the above, Quercus rubra
and Q. coccinea form an exception. On the whole, however, little
emphasis can be laid on the connection between autumnal coloration
and relationship, as is illustrated by the Viburnums V. Opulus and
V. Lantana ; the former rapidly reddens in autumn whereas the latter
forms little or no anthocyanin under similar conditions.

5. In a number of species, mechanical injury frequently leads to
formation of anthocyanin in the portion of leaf, stem, or shoot, distal
to the point of injury. Partial severing of a portion of the leaf blade
may induce reddening in the severed portion. Crushing, or partial
breaking of a stem or petiole, leads to the same result, as was shown
by Gautier (175) in the Vine, or ringing of a stem will cause the leaves
above to turn red, as Combes (374, 385) has demonstrated in Spiraea.
The same kind of injury, with similar results, may be brought about
by the attacks of insects. Under this heading must be included, as a
result of injury, the copious formation of anthocyanin in galls, such as
those which occur on species of Rosa, Salix, Quercus and many other
plants. Also the local appearance of pigment in tissues infected with
Fungi : as an example one may quote the purple blotches produced
on leaves of Tussilago Farfara and the purple streaks on Wheat stems
when these plants are infected with Puccinia. The physiological causes
involved in the production of pigment on injury are dealt with in
Chapter vi. Some plants show a greater tendency to redden on injury
than others (Oenothera, Rumex, Rheum) ; also the reddening on injury
is more rapid and frequent in autumn when the vegetative season is
about at an end.

6. Low temperature may induce anthocyanin formation apart from
autumnal coloration, for, in the development of the latter, factors
other than temperature play an important part. Many evergreen
shrubs show a reddening 1 in their leaves during winter (Ligustrum
vulgare, Mahonia spp.). The same statement holds good for certain
herbaceous plants which retain their leaves in winter (Saxifraga umbrosa).

7. Plants exposed to drought often develop anthocyanin. This
may be seen sometimes in individuals in pots (Pelargonium] which have

1 This is of course quite different from the reddening shown by some evergreens in
winter, notably species of the Gymnosperms, and which is due to the formation of a red
product from chlorophyll in the chloroplastids.

n] OF ANTHOCYANINS 25

been insufficiently watered. Miyoshi (375) has observed that leaves
of trees in the East Indies, Ceylon and Java redden during the
dry period in the same way as autumnal leaves in the temperate

regions.

8. In High Alpine plants there is undoubtedly, on the whole, a
much stronger development of anthocyanin in all the parts stem,
leaves, bracts and flowers, than in plants growing in lowland regions.
Gaston Bonnier (294, 307) has made observations upon this point, and
has found that the flowers of many species are more highly coloured,
the greater the altitude. Kerner (398) also remarks : ' The leaflets and
stem of the Alpine Sedum atratum, those of Bartsia alpina, and, above
all, numerous species of Pedicularis (e.g. Pedicularis incarnata, rostrata,
recutita) are coloured wholly purple or dark violet. ...It is also a very
striking phenomenon that widely-distributed grasses (e.g. Aira cccspi-
tosa, Briza media, Festuca nigrescens, Milium effusum, Poa annua and
nemoralis), which in the valley possess pale-green glumes, develop
anthocyanin in them on lofty mountains, so that there the spikes
and panicles exhibit a deep violet tint, and on this account the regions
in which grasses of this kind grow in great quantities receive a peculiar
dark colouring.... The same occurs in the numerous sedges and rushes
growing in the Alps, which have dark- violet, almost black, scales covering
the flowers (e.g. Carex nigra, atrata, aterrima, Juncus Jacquinii, trifidus,
castaneus)....It is known that the floral-leaves of many plants growing
on lofty mountains, and in the far north, are coloured blue or red by
anthocyanin, whilst in the same species, growing in warm lowlands
and in southern districts, they appear white. Particularly noticeable
in this respect are the Gypsophyllas (Gypsophylla repens), the Carline
Thistle (Carlina acaulis), the large-flowered Bitter-cress (Cardamine
amara), the Milfoil (Achillea Millefolium), and many of those Umbelli-
ferae which have a very wide distribution, and occur all the way from
the lowlands up to a height of 2500 metres in the Alps, such as Pim-
pinella magna, Libanofis montana, Chaerophyllum Cicutaria, and Laser-
pitium latifolium." And again: 'The flowers of species grown in the
Alpine garden on the Blaser at a height of 2195 metres above the sea
exhibited, as a rule, brilliant floral tints, and some were decidedly
darker than the flowers grown in the Vienna Botanic Garden. Agro-
stemma GitJiago, Campanula pusilla, Dianthus inodorus (sylvestris),
Gypsophila repens, Lotus corniculatus, Saponaria ocymoides, Satureja
hortensis, Taraxacum officinale, Vicia Cracca, and Vicia sepium are
good examples of this. Several species, which produced pure white

26 THE MOKPHOLOGICAL DISTRIBUTION [CH.

petals in the Vienna gardens, e.g. Libanotis montana, had petals coloured
reddish- violet by anthocyanin on their under sides in the Alpine garden.
The glumes of all the Grasses which were green, or only just tinged with
violet at a low level became a dark brownish- violet in the Alpine garden.
The abundant formation of anthocyanin in the green tissue of the foliage-
leaves and sepals, and in the stem, was particularly apparent. The
leaves of the Stonecrops, Sedum acre, album, and sexangulare became
purple-red, those of Dracocephalum Ruyschianum and Leucanthemum
vulgare violet, those of Lychnis Viscaria and Satureja hortensis a brownish-
red, and the foliage-leaves of Bergenia crassifolia and PotentiUa Tiro-
liensis, even in August, had the scarlet-red colour which they usually
assume in sunny spots in the valley in late autumn."

As regards Arctic vegetation, Gertz (19), on the authority of Wulff
(71), maintains that abundant production of anthocyanin is a dis-
tinguishing feature ; and so much so, that the periodicity of autumnal
coloration is very little marked. The causes of the intense colour in
Alpine and Arctic vegetation are dealt' with in Chapter vi.

9. Production of anthocyanin in vegetative organs is characteristic
of certain halophytes (Salicornia, Suaeda, Atriplex).

10. The red-leaved varieties of green-leaved types are due to the
production of anthocyanin, which is either absent from the type, or
present to only a slight extent (red-leaved varieties of Fagus, Berberis,
Brassica and Prunus).

11. The petals and perianth are essentially the organs producing
anthocyanin. To these must be added the perianth-like calyx (Anemone,
Delphinium, Aconitum, Begonia and others). Also the bracts of the
inflorescence may develop anthocyanin to a large extent, and may either
assist (Salvia), or even take the place of, the corolla as organs of attrac-
tion (Bougainvillaea, Euphorbia, Poinsettia, spathes of Araceae). In
general, we may say that the stems and pedicels of most inflorescences
show an abundant formation of anthocyanin which no doubt increases
the conspicuousness of these parts ; yet, at the same time, the under-
lying cause is probably that the physiological conditions in all tissues
of reproductive organs enhance the formation of pigment.

Anthocyanin is not confined to the perianth, but is frequently present
in the carpels and style : of its occurrence in the stigma we have many
examples among the Amentales (Myrica, Alnus, Betula, Corylus, Car-
pinus, Salix, Populus), as well as in Rumex and Ricinus.

In the stamens, too, it is found, especially in many anemophilous
plants (Populus, Fraxinus) and in the Graminaceae (Molinia, Phleum,

n] OF ANTHOCYANINS 27

Alopecurus, Cynosurus, Dactylis). Pollen grains may also be coloured
with anthocyanin (Campanula).

12. Many fruits form abundant anthocyanin. The pigment may
be temporary in some stages of development of capsules which eventually
become dry and dead (many of the Leguminosae, Hypericum, Sedum)
and schizocarps (Umbelliferae). The more obvious cases are the drupes
and berries, and of these, two types stand out, i.e. those in which the
pigment is limited to the epicarp (Prunus. Vitis), and those in which
it is present in both epicarp and mesocarp (Rubus, Ribes, Ligustrum,
V actinium, Atropa, Solanum, Sambucus, Viburnum, Lonicera). It may be
formed also in the tissues of 'false fruits' (Fragaria, Pyrus). From other
fruits (Cucurbita, Solanum, Lycopersicum, Citrus 1 ) it is entirely absent.

13. Anthocyanin is sometimes developed in the cells of the testa
of seeds (Abrus, Phaseolus, Pisum). Further instances are the seeds
of Adenanthera, Erythrina, Lepeirosia (Potonie, 56), Trifolium pratense
(Preyer, 66) and many Podalyrieae (Lindinger, 76). More frequently
pigmentation in the seed-coat is due to impregnation of the cell-wall
by pigments which cannot be classified as anthocyanins, and of which
very little is known.

14. Roots and underground stems often form anthocyanin in
considerable quantity. Three cases of this type of pigmentation may
be differentiated :

(a) The pigment develops under ground apart from the influence
of any outside factors. Well-known examples for roots are Beta
vulgaris and Raphanus sativus, and for underground stems, Solanum
tuberosum; in the latter species the pigment may be confined to the
epidermal layers or may be present in the inner tissues (Salaman, 544).
To the above may be added other cases such as :

Root-tips of species of Saxifragaceae and Crassulaceae (Irmisch, 27).

Roots of Parietaria diffusa and Gesneria sp. (Zopf, 48).

Roots of species of Pontederiaceae, Haemodoraceae and Cyperaceae
in the ground (Ascherson, 43).

Roots of Wachendorfia in the ground (Hildebrand, 44).

(6) The pigment develops in roots normally exposed.

Aerial roots of Ficus indica (Mobius, 68).

Floating roots of Pontederiaceae (Hildebrand, 44 ; Ascherson, 43).

Aerial roots of Aroideae (van Tieghem, 33; Lierau, 51).

Aerial roots of Orchidaceae (Leitgeb) 2 .

1 Except in the red-fleshed variety, the so-called 'Blood Orange.'

2 Leitgeb, H., 'Die Luftwurzeln der Orchideen,' Denkschr. Ak. Wiss., Wien, 1865,
xxrv, p. 179 (p. 204).

28 MOKPHOLOGICAL DISTRIBUTION OF ANTHOCYANINS [CH. n

(c) The pigment develops in roots abnormally exposed.

Exposed roots of Alnus (Mobius, 68).

Roots of Salix grown in water in glasses (Schell, 287).

Adventitious roots of Echeveria metallica (Pirotta, 42).

Roots of Zea Mays growing in water or in moss and exposed to
light (Dufour, 305; Devaux, 308).

Exposed roots of Saccharum qfficinarum (Benecke, 312).

Adventitious roots above ground of Poa nemoralis (Beyerinck, 303).

To Kerner (398) we owe the observation that rhizomes of Dentaria
bulbifera, Lathraea Squamaria and of species of Cardamine and Viola
become violet when exposed in water to sunlight.

Gertz (19) lays stress upon the fact that numbers of roots and under-
ground stems form anthocyanin when, for some reason, they have been
exposed, as for instance when they grow out from banks of streams,
or in shallow water, or when exposed during autumnal ploughing.
He appends the following list of such cases including Tradescantia
zebrina, Eriophorum angustifolium, Phragmites communis, Rumex
Acetosella, Geranium molle, Potentilla anserina, Gunnera scabra, Stachys
palustris, Plantago major., Anthemis arvensis and Artemisia vulgaris.

15. A development of anthocyanin appears to be characteristic
of many parasites and saprophytes, though possibly the absence of
chlorophyll has the effect of making the red pigment more conspicuous
(species of Cuscuta, Orobanche, Lathraea). Mirande (332) gives an
interesting account of the development of anthocyanin in the genus
Cuscuta.

CHAPTER III

THE HISTOLOGICAL DISTRIBUTION OF ANTHOCYANINS

When tissues containing anthocyanin are examined microscopically,
in the majority of cases the pigment is found to be in solution in the
cell-sap, which, of course, occupies the vacuole or vacuoles of the cell.
We know, as Gertz (19) points out, from the researches of de Vries (141)
and Pfeffer 1 that in living cells the protoplasm is impermeable to antho-
cyanin in the cell-sap. But when the protoplasm is dead and semi-
permeability ceases, anthocyanin exosmoses out of the cell, and at the
same time it is often taken up, like a stain, by the protoplasm and
nucleus 2 .

From the condition of solution in the cell-sap, two deviations are
possible : first, in tissues which gradually die or become woody, the
anthocyanin may be absorbed into the substance of the cell- wall, thereby
causing the latter structure to become coloured ; secondly, the concen-
tration of anthocyanin in living cells may become so great that the

1 Pfeffer, W., Osmolische Untersuchungen, Leipzig, 1877 (pp. 134, 135).

2 Staining of the nucleus under these conditions has been noted by Hartig, de Vries
and Molisch.

Hartig, Th., 'Chlorogen,' Bot. Ztg., Berlin, 1854, xn, pp. 553-55(5 (p. 555).

Hartig, Th., 'Ueber das Verfahren bei Behandlung des Zellenkerns mit Farbstoffen,'
Bot. Ztg., Berlin, 1854, xn, pp. 877-881 (p. 879).

Hartig, Th., 'Ueber die Bewegung des Saftes in den Holzpflanzen.' Bot. Ztg., Leipzig,
1862, xx, pp. 89-94 (p. 93).

Vries, H. de, ' Plasraolytische Studien iiber die Wand der Vacuolen,' Jdhrb. wiss.
Bot., Berlin, 1885, xvi, pp. 465-593.

Molisch, H., Untersuchungen iiber das Erfrieren der Pflanzen, Jena, 1897 (pp. 28, 29).

On account of this property, anthocyanin has been used as a stain for liistological
purposes. The red anthocyanin of Cabbage has been employed in this way by Tait,
Gierke, Flesch (159), and Claudius (184); Lavdowsky (160) also recommends anthocyanin
from Myrtillus berries.

Tait, L., 'On the Freezing Process for Section-Cutting,' J. Anat. Physiol., London,
1875, ser. 2, vin, pp. 249-258.

Gierke, H., 'Farberei zu mikroskopischen Zwecken,' Zs. wiss. Mikrosk., Braunschweig,
1884, i, pp. 62-100, 372-408, 497-557 (p. 99).

30 THE HISTOLOGICAL DISTRIBUTION [CH.

pigment separates out either in crystalline or amorphous form. In
this solid state, anthocyanin may be pure, or may merely be absorbed
by substances in the cell, such as proteins, tannins.

The following account is taken from Gertz (19) who has particularly
investigated this portion of the subject. Several cases of pigmented
cell-walls are recorded by Gertz among Monocotyledons, but no very
definite statements are made as to whether the colour is actually due
to anthocyanin. The cases quoted are in Pliormium lenax (Engelmann),
Eriophorum angustifolium (Wulft'), Sorghum (Kraus), Tillandsia dian-
toidea (Tassi), Pontederia crassipes (Hildebrand), Angraecum superbum
and Macroplectrum sesquipedale (Hering) ; Gertz also includes examples
which he has observed himself. Among Dicotyledons coloured cell-
walls are more rare. Red bast (Oxalis Ortgiesii) and red epidermis
(Epilobium montanum) are quoted ; in Jacquinia smaragdina ( Vesque 1 )
the cuticle is crimson. According to Gertz, the coloured cell-walls
described in Homalomena rubrum (Nageli & Schwendener 2 ) and in
Geranium Robertianum, Begonia maculata, Quercus palustris, and Rumex
Acetosella (Wigand, 12) are artificially produced ; such artifacts can be
brought about by putting anthocyanin-containing cells in a medium
absorbing water, but in which anthocyanin is insoluble, such as ether.
The phenomenon can be demonstrated in leaves of Begonia spp. In
addition to the above cases, red pigment is found in walls of testa
cells in Abrus precatorius (Nageli & Schwendener 2 ) though Mobius (68)
found cell-sap too. Sclerotic cells of testa of Erythrina and Goodia
have blue, violet and red cell-walls (Tschirch 3 ) ; also in seeds of
Trifolium pratense (Preyer, 66).

As regards the presence of solid and crystalline anthocyanin in
cells the following list of cases compiled by Gertz (19) is extremely
useful :

Tillandsia ainoena, Hildebrand (30).
Aechmea sp., Dennert (14).
Allium Schoenoprasiim, Courchet (97).
Lilium Martagon, Overton (333).
Dracaena Jonghi, Hassack (393).
Convallaria majalis, Weiss (91).
Iris pumila, Dennert (14).

1 Vesque, J., 'Memo-ire sur 1'anatomie cornparee de 1'ecorce,' Ann. sci. nat. (Bot.),
Paris, 1875, ser". 6, n, p. 82.

2 Nageli, C. L., und Schwendener, S., Das Mikroskop, Theorie undAnwendung desselben.
Leipzig, 1877.

3 Tschirch, A., Angewandte Pflanzenanatomie, Wien und Leipzig, 1889, Bd. i, pp. 62, 66.

m] OF ANTHOCYANINS 31

Strelitzia Reginae, Mohl (84), Hildebrand (30).

Orchis mascula, Nageli (85).

Epidendrum cochleatum, Schlockow 1 .

Cattleya quadricolor, Malte 2 .

Laelia Perrinii, Malte 2 .

Oncidium sphacelatum, Schlockow 1 .

Dianthus Caryophyllus, Molisch (104).

Aquilegia atrata, Molisch (104).

Delphinium elatum, Weiss (91), Molisch (104).

D. formosum, Zimiuerniann 3 .

D. tricolor, Fritsch (94).

D. sp., Molisch (104).

Glaucium fulvum, Schimper (93).

Papaver sp., Hildebrand (30).

Brassica oleracea, Molisch (104).

Hydrangea hortensis, Ichimura (192).

Rub-iis caesius,

R. corylifolius,

R. (jlandulosus ,

Trecul (89).

R. laciniatus,

R. lasiocarpus,

Geum sp., Weiss (91).

Rosa sp., Molisch (104).

Pyrus communis, Sorauer (304) .

Neptunia oleracea, Rosanoff (92).

Lathyrus heterophyllus,

L. silvestris,

Cytisus Laburnum,

C. Alschingeri,

C. scoparius,

Molisch (104).

Medicago saliva,

Hedysarum coronarium,

Ononis natrix,

Baptisia australis,

Amorplia fruticosa, Hildebrand (30).

Erodium Manescari, Molisch (104).

Pelargonium zonale, Molisch (104).

P. Odier hortormn, Molisch (104).

Vitis sp., Molisch (104).

Ampelopsis quinquefolia, Mer (284).

1 Schlockow, A., Zur Anatomic der braunen Bliiten, Inaugural-Dissertation zu Heidel-
berg, Berlin, 1903.

3 Malte, M. 0., ' Untersuchungen iiber eigenartige Inhaltskb'rper bei den Orchideen,'
Bihang till K. Svenska Vet.-Akaderniens Handlingar, Bd. xxvn, No. 15, p. 32.

3 Ziinmerinann, A., 'Die Morphologic und Physiologic der Pflanzenzelle ' (Schenck's
Handbuch der Botanik, Breslau, 1887, Bd. in (2), pp. 104, 105).

32 THE HISTOLOGICAL DISTRIBUTION [CH.

Ampelopsis humulifolia, \

j . f JjC'IlLTC'l -n-C-ll

4. murahs,

Viola, tricolor, Nageli (85).

Passiflora acerifolia, Bohm (88), Weiss (91).

P. alata, Weiss (91).

P. laciniata, Unger (86).

P. limbata, Bohm (88).

P. suberosa, Unger (86), Bohm (88), Weiss (91).

Begonia discolor, Unger (86).

B. maculata, Molisch (104).

Centradenia floribunda, Buscalioni & Pollacci (17).

Primula sinensis, Bokorny (99).

Anagallis arvensis, Courchet (97).

A. arvensis var ciliata, )

- Molisch (104).
A. arvensis var coerulea, )

Acantholimon sp., \

Swertia verennis,

. ,. . ' Y Miiller (61).

Amsoma saiicijolia,

Hoya carnosa, }

Gilia tricolor, Hildebrand (30).
Nemophila sp., Molisch (104).
Salvia splendens, Mohl (84).
Airopa Belladonna, Trecul (89).
Hyoscyamus niger, Miiller (61).
Solanum americanurn, Nageli (90).

5. guineense, Trecul (89).
S. Melongena, Weiss (91).

S. nigrum, Hartig (83), Trecul (89), Weiss (91).

Antirrhinum inajus, Molisch (104).

Gesneria caracasarta, Dennert (14).

Columnea Schiedeana, Weiss (91).

Thunbergia alata, Fritsch (94).

Justicia speciosa, Pirn (95).

Coffea arabica, Tschirch (100), Kroemer (102).

Viburnum Tinus, Fritsch (94).

To the above list Gertz adds a number of plants in which he himself
has observed solid anthocyanin as well as anthocyanin in solution:

Melica nutans, Restrepia sp.,

Tradescanlia discolor, Laelia Perrinii,

Hyacinthus orientalis, Dendrobium chrysanthum,

Dracaena Jonghi, Maxillaria Henchmanni,

Crocus vernus, M. tenuifolia,

Iris hybrida, Oncidium sp.,

1 Lengerken, A. von, 'Die Bildung der Haftballen an den Ranken einiger Arten der
Gattung Ampelopsis,' Bol. Ztg., Leipzig, 1885, XLIII, pp. 376, 378.

Ill]

OF ANTHOCYANINS

33

Oncidium sphacelatum,

Myrica Gale,
Juglans macrocarpa,
J. regia,
Carya glabra,
Corylus tubulosa,
Castama vesca,
Quercus macrocarpa,
Q. r libra,
Ulmus montana,
Rumex crispus,
R. domesticus,
Rheum rugosum,
Pol ijgo 1 1 urn Persicaria,
Silene tunetana,
Dianthus sp.,
Helleborus multifidus,
Epimedium alpinum,
Drosera rotundifolia,
Saxifraga granula.ta,
S. tridactylites,
Ribes petraeuni,
Rub us Idaeus,
PotentiUa thuringiaca,
Baplisia minor,
Robinia viscosa,
Geranium sangiiineum,
Pelargonium zonale,
Ailanthus glandulosa,
Rhus typhina,
Serjania grammatophora,

I ntpatiens glanduligera,
Ceanolhus azureus,
Malva rotundifolia,
Garcinia Mangoslana,
Trapa bicornis,
T. natans,
Fuchsia triphylla,
Primula sinensis,
Androsace lacti flora,
Gentiana Oliveri,
G. acaulis,
Amsonia latifolia,
Hot/a carnosa,
Calyslegia sepium,
Cobaea scandens,
Collomia grandiflora,
Scopolia orientalis,
Solarium Dulcamara,
Paulownia imperialis,
Scrophularia nodosa,
Columnea picta,
Episcia cupreata,
Alloplectus speciosus,
A. vittatus,
Isoloma hirsuta,
Sciadocalyx Luciani,
Dircaea macrantha,
Gesneria fulgens,
Sambucus racemosa,
Viburnum Tinus.

In the above plants with a few exceptions, according to Gertz,
the solid form of anthocyanin could not be obtained by chemical reagents,
but in the following plants he was able to obtain ' anthocyanin bodies ' in
the cells by treatment with chemical reagents such as alcohol, alkalies,
sulphuric and hydrochloric acids.

Adiantum macrophyllum alcohol.
Nidularium Cervantesi alkalies, alcohol.
Tradescantia discolor sulphuric acid.
Phrynium cylindricum caustic potash.
Rheum rugosum hydrochloric acid.
Viscaria viscosa sulphuric acid.
Nepenthes Hookeriana alkalies ?

Sempervivum dichotomum caustic potash, Eau cle Javelle.
Baptisia minor sulphuric acid,
w. P. 3

34 THE HISTOLOGICAL DISTRIBUTION [CH.

Lupinus grandiflorus \

L. perennis hydrochloric acid.

L. polyphyllus

Anthyllis Vulneraria caustic potash.

Begonia acanthostigma alkalies ?

Gentiana alba hydrochloric acid.

Scopolia orientalis sulphuric acid.

Sciadocalyx I/nciani sulphuric acid.

Strobilanlhes Diirianus alcohol, sulphuric acid.

S. Sabinianus- alcohol.

Justicia formosa ?

Filtonia Pearcei alcohol.

There has been much doubt in some cases as to whether the coloured
deposits are of pigment only, or of other substances such as proteins,
tannins, etc., impregnated with pigment. Very beautiful crystals are
figured by Molisch (104) from cells of red Cabbage leaves and leaves of
Begonia maculata. Molisch notes that a low temperature induces the
formation of the crystals which may disappear at higher temperatures.
In Brassica and Begonia, appearances are certainly in favour of the
crystals being pigment. In other plants where the solid deposit is in
the form of coloured spheroids or globules, the latter often being viscous
in consistency, it seems more probable that the matrix of the bodies
is of a tannin or protein nature, and has become impregnated with
pigment.

Gertz maintains several possibilities for the natural occurrence of
anthocyanin in cells. (1) In crystals as mentioned above either of
pure anthocyanin or of colourless substances infiltrated with anthocyanin.
(2) In very young cells containing several vacuoles, anthocyanin in
the vacuoles may appear as red or blue globules having a superficial
likeness to anthocyanin bodies. (3) Anthocyanin may occur in amor-
phous or globular form, or bound up with amorphous or globular bodies
which are of a different nature from the pigment.

To demonstrate how solid anthocyanin bodies may arise, Gertz
describes two examples of what may happen artificially in the cell.
If tissues of Slrobilanthes Diirianus be placed in alcohol, coloured drops
appear which afterwards disappear again. The explanation offered is
that the alcohol first causes the cells to plasmolyse, and substances in
the cell may be precipitated and take up anthocyanin. Later, when
semipermeability ceases, the alcohol redissolves out the pigment.
Secondly, in Scopolia orientalis, on addition of sulphuric acid, there
is a precipitation of coloured bodies which does not disappear. In this

in] OF ANTHOCYANINS 35

case Gertz supposes that the precipitated substances are changed by
the reagent, or are insoluble in the new medium, and so remain. As
already stated, the above examples are artificial, but Gertz suggests
that in a similar way in the normal cell, substances may be precipitated
and absorb anthocyanin. This precipitation may be brought about
by increased concentration, or by lowering of temperature. The latter
was found by Molisch to be the cause of crystallisation of anthocyanin
in Brassica.

As to the chemical nature of the substances which absorb antho-
cyanin in the cell, according to Gertz, little can be said. Often they
appear to be tannin-like compounds, impregnated with anthocyanin,
in the form of solid bodies (in young leaves of Juglans, Quercus, Rubus,
Ribes and others), or globules as described in the Bromeliaceae (Wallin,
101), Cissus (D'Arbaumont 1 ), Euphorbia (Gaucher 2 ) and Dicentra
(Zopf, 48). Possibly in some other cases the ground substance is
protein; coloured aleurone grains have been recorded by Hartig (87),
which are red in Laurus nobilis and blue in Cheiranthus annum, and
blue grains have been noted by Spiess (103) in Zea. There is also the
possibility, even when the anthocyanin deposits are crystalline, of the
stroma being of some other substance. Nageli (90) has suggested that
coloured protein crystalloids occur in Solanum americanum.

We may next consider the distribution of anthocyanins in the tissues
of leaves, stems, flowers, fruits and seeds.

As regards leaves, reference may be made to the work of Morren
(26), Pick (391), Hassack (393), Engelmann (394), Stahl (62, 405),
Berthold (64), Griffon (331), and Buscalioni & Pollacci (17). No very
general statement can be made about the histological distribution in
leaves beyond that the pigment is often present in the epidermis, quite
frequently in the subepidermal tissues also, and in many cases in the
more internal tissues as well. Neither can it be predicted with certainty
for any plant where the localisation will be, and it may vary considerably
in the same plant according to whether we select for investigation leaves
in the young, autumnal or winter condition, or leaves during the normal
vegetative period; or again leaves under the influence of attacks of
insects or Fungi. Accounts of the histological distribution in leaves
have been made by various authors (Hassack, Engelmann, Pick, Stahl,

1 D'Arbaumont, 'La tige des Ampelidees,' Ann. sci. nat. (Bot.), Paris, 1881, ser. 6,
xi, pp. 186-253 (pp. 241, 242).

2 Gaucher, L., 'Rechercb.es anatomiques sur les Euphorbiacees,' Ann. sci. nat. (Bot.),
Paris, 1902, ser. 8, xv, pp. 161-309 (p. 179).

32

36 THE HISTOLOGICAL DISTKIBUTION [CH.

Massart), and a very detailed classification is recorded by Buscalioni
& Pollacci.

A less detailed classification due to Gertz is quoted as follows :

A. Permanent anthocyanin.

1. In the epidermis. Ex. Orchis, Canna, Maranta, atropurpurea-
forms of Fagus, Corylus and Acer, and also of Beta vulgaris, Atriplex
hortensis and Croton sp.

2. In peripheral layer of the ground tissue. Ex. Dracaena, Eucomis
punctata, Erythronimn Dens-Canis, Arum maculatum, Arisarum vulgare,
Berberis vulgaris atropurpurea.

3. In both epidermis and ground tissue. Ex. Aerua sanguinolenta,
Aeschynanthus atropurpureus, Iresine Herbstii, Coleus sp.

4. In median ground tissue. Ex. Higginsia refulgens, Sinmngia
purpurea, Gesneria Donkelaari, Pellionia Daveauana.

B. Periodic anthocyanin.

(a) In young leaves and spring leaves.

1. In the epidermis. Ex. Rubus, Rosa, Rhus, Silene, Malva.
Veronica, Odontites, Ajuga, Salvia.

2. In the ground tissue. Ex. Salix, Populus, Fagus, Acer, Cydonia,
Vicia, Stellaria, Cerastium, Campanula, Taraxacum.

3. In hairs. Ex. Quercus rubra, Q. macrocarpa, Castanea vesca,
Chenopodium album, Mallotus japonicus, Vitis alexandrina, V. Labrusca.

(b) In older and autumnal leaves.

1. In the epidermis. Ex. Philadelphus, Deutzia, Euonymus
europaeus, and some Oenotheraceae.

2. In ground parenchyma. Ex. Populus, Salix, Quercus, Rhus,
Acer, Vitis, Ampelopsis, Prunus, Cornus, Viburnum.

(c) In winter leaves.

1. In the epidermis. Ex. Silene, Oenothera, Gaura, Veronica,
Lamium.

2. In the ground tissue. Ex. Secale, Hedera, Calluna, Empetrum.

One or two further observations of interest are mentioned by Gertz.
This author notes that in some leaves there is an indication of division
of labour as regards anthocyanin. When the pigment appears princi-
pally in the epidermis, it is often found in addition in isolated mesophyll
cells which are poor in chlorophyll (Spiraea, Rosa and Rubus).

Although, as we have seen, from the observations of the above-
mentioned authors, anthocyanin formation is by no means limited to

in] OF ANTHOCYANINS 37

any particular tissue of the leaf, yet Parkin (77), in a recent investigation
on the localisation of anthocyanin, points out the prevalency of the
idea among botanists that anthocyanin is usually formed in the epidermis.
Parkin's researches are interesting because they give an indication as
to the particular localisation of the pigment when formed under varying
circumstances and under the influence of different factors. He examined
400 cases and classified his results as follows :

1. Transitory anthocyanin of young leaves. This is a feature of
the young foliage of many plants in the tropics and temperate regions.
The number of species examined was 235, and the distribution was
classified as:

Epidermis Mesophyll Epidermis and mesophyll

20 % 64 % 16 %

2. Autumnal anthocyanin. The number of species examined was
81 and the distribution :

Epidermis Mesophyll Epidermis and mesophyll

11% 78% 11%

3. The permanent anthocyanin of mature leaves. In this case
the pigment appears as the leaf matures, and persists throughout the
life of the leaf as a normal character. This class includes (a) leaves
with uniformly red lower surfaces; (6) leaves with definite pigmented
areas in the form of spots, blotches or zones ; and (c) leaves of horti-
cultural varieties with coloured foliage. The number of species examined
was 54, and the distribution:

Epidermis Mesophyll Epidermis and mesophyll

70% 17% . 13%

4. The accidental anthocyanin of mature leaves. This is not
normally present in mature leaves but arises only under exceptional
conditions such as: (a) excessive insolation, followed by cool nights,
as seen in Alpine plants, and in evergreens during winter; (6) the
result of injury, when a reddish zone often appears round a wound in a
leaf; (c) through the accidental exposure of lower surface to the full
rays of the sun. The number of cases examined was 30, and in the
majority the anthocyanin was confined to the mesophyll. Parkin
sums up by saying that anthocyanin of young and autumnal leaves is
usually confined to the mesophyll ; of mature leaves, when a normal
feature, to the epidermis; and, when exceptional, to the mesophyll.

38 THE HISTOLOGICAL DISTRIBUTION [CH.

The mesophyll, he considers, to be the usual, and, perhaps, more primitive
seat for anthocyanin.

As regards the distribution in steins, we have the work of Berthold
(64) and Buscalioni & Pollacci (17). Two types of distribution can
be recognised:

1. Anthocyanin in the epidermis. Ex. Gentiana, some Labiatae
(Berthold, 64), some Solidago and Aster spp.

2. Anthocyanin in subepidermal assimilating cells. Ex. Alsinaceae,
Papilionaceae and Convolvulaceae.

The histological distribution in petals has been investigated by
Hildebrand (30), Koschewnikow (40), Mliller (61) and Buscalioni &
Pollacci (17), and in anthers by Chatin (34). In the corolla, anthocyanin
is chiefly located in the epidermis, either upper or under, or both, though
its appearance in the subepidermal and inner tissues is by no means
uncommon. The following table, selected from Buscalioni & Pollacci's
(17) records, gives an idea of the relative frequency with which pigments
are formed in the various regions of the corolla and perianth.

6 I I.

S 4 'a
s ?

^ M '&,

3 P FM
1

2 Lamium purpureum + + ... ... + +

3
4
5

6 Tropaeolum ... ... ... +

7 Magnolia + ... ... +

8 Dianthus + ... ... +

9 Dielytra spectabilis + ... ... +

10 Hibiscus + ... ... +

11

12 Tulipa + +

13 Camellia ... + ... ... +

14

15 Clerodendron + ... ... +

16 Corydalis bulbosa + ... ... +

17 Azalea indica + ... + ... + ... +

18 Muscari comosum + ... + ... 4 r ... +

19 Dendrobium tirsiflorum ... + ... + + + ... +
20

21
22
23
24
25
26

15

CD
_0

of

Name of plant K & 1
Ciipripedium sp +

a g

| 6

3 ^ >> a ^2

1* >i^c >j j_,

| |1 1 1

O Oa3 * e8

^ ^-o OH >

Parenchyma

Upper hypo-

1 ,_1 In-*

+

Epacris

,

Erica

4,

Cineraria

Tropaeolum

Magnolia

,

Dianthus

,

Dielytra spectabilis

,

Hibiscus

,

+

Tulipa

,

Camellia

,

.

Clerodendron

,

Corydalis bulbosa

,

Azalea indica

+ +

4-

Muscari comosum

, ,

Dendrobium tirsiflorum
Helleborus niger

+-..++

+

...

Hyacinthus orientalis
Dalecliampia

4 r

4_

Echeveria grandiflora

+

Begonia floribunda

+ + 4-

4-

4-

Vanda suavis

' -(-

4-

Rhododendron hybridum

Total.., 2

17 5 5 2

6

^

in] OF ANTHOCYANINS 39

Chatin (28) notes that anthocyanin occurs in deep-seated tissues
in thicker petals, such as Ulloa and Asclepias, whereas in thinner petals
it is usually confined to the epidermis.

The most interesting feature in connection with the histology of
coloured petals is the combination of colour effects produced by the
simultaneous presence of two, or even more, pigments in the cells.
Strictly speaking, the plastid pigments, i.e. the yellow, orange-yellow
and orange colouring matters which are bound up with special proto-
plasmic structures the plastids, have no place in this account; and
the same may be said for the soluble yellow pigments (mostly flavones).
But both these classes occur so often with anthocyanin, and so frequently
modify its colour, that some mention of them at this point will not be
out of place. References can be made, in addition, to other authors;
Hildebrand (30) wrote an early account of flower pigmentation, including
combinations of plastid and sap colours; there is also an interesting
paper by Bidgood (18) on flower colours, and many detailed observa-
tions by Dennert (14) and by the author (211).

One of the most frequent combinations of pigments is purple,
purplish-red, or red anthocyanin and yellow plastids. The resultant
colour may be brown (Cheiranthus Cheiri, Tagetes signata), crimson
(Zinnia), scarlet (Geum coccineum), or orange-red (Tropaeolum majus);
there are of course a great many other cases of combination of these
two pigments, and a correspondingly large number of shades of brown,
crimson, or scarlet, as the case may be. A less frequent combination is
dark brown, brownish-black, or black resulting from purple anthocyanin
and chloroplastids. This effect is produced in some Cypripedium
flowers, and we have already noted it in the case of leaves (see p. 19).
The brown or black effect is due to the fact that the two pigments are
complementary as regards the rays they absorb ; those which are not
absorbed by chlorophyll are absorbed by anthocyanin, and so the
result is negative as regards colour. But black or brown is not always
due to this combination; the black spots on Adonis and Papaver
flowers owe their appearance to deep blue cell-sap, and in the dark
markings on some varieties of Bean (Phaseolus) seeds the cells contain
purple anthocyanin. There are, in addition, true brown and black
pigments which appear in some flowers as in the spots on the alae of
Vicia Faba 1 .

1 Mobius, M., 'Das Anthophaei'n, der braune Bliithenfarbstoff,' Ber. D. hot. Ges.,
Berlin, 1900, XVHI, pp. 341-347. Schlockow, A., Zur Analomie der braunen Bliiten,
Inaugural-Dissertation zu Heidelberg, Berlin, 1903.

40 THE HISTOLOGICAL DISTRIBUTION [CH.

Another class of combinations is the outcome of the mixture of
anthocyanin and a soluble yellow pigment in the same cell. Such a
combination occurs in the crimson flowers of Mirabilis Jalapa, though,
on the whole, it is usually found to be characteristic of varieties which
have arisen under cultivation, as for instance in the crimson varieties
of Antirrhinum majus, Dahlia variabilis and Portulaca grandiflora. In
these species the flower of the original wild type had some shade of
magenta anthocyanin only. Three variations are then characteristic-
variation to ivory-white, to yellow and to crimson, the latter being a
mixture of magenta anthocyanin and soluble yellow pigment, such a
combination as would not normally occur in nature.

It is interesting to note that colour of type and variation in
the anthocyanin-soluble-yellow series is reversed in the antho-
cyanin-plastid series. In the latter, the original type is either
crimson (Zinnia), yellow striped with brown (Cheiranthus), or orange-
red (Tropaeolum) ; a variety characteristic of this series is one in which
the yellow pigment almost disappears from the plastids, or is replaced
by a much paler yellow substance. When anthocyanin is present
with these pale yellow pigments, or the yellow pigment is altogether
absent, a purple or magenta variety results, as in the purple Cheiranfhus
and magenta Zinnia. When the anthocyanin is more red, as in the
orange-red Tropaeolum majus, a carmine variety arises when these
pale plastids only are present. Thus, in the anthocyanin-soluble-yellow
series we have a magenta (or purple) type with crimson, ivory-white
and yellow varieties, whereas in the anthocyanin-plastid series we have
a crimson type, with magenta (or purple), ivory and yellow varieties.

Bidgood (18) mentions one or two unusual colour combinations:
the lurid colour of some Delphiniums he attributes to the presence of
cells containing red anthocyanin side by side with cells containing blue.
Crocus aureus, also, on the outer side of the perianth leaves, shows
green stripes which are due to a combination of blue sap colour on the
outer side of the perianth leaves and yellow soluble pigment on the
inner side. Dennert (14) gives many examples of the different ways
in which plastid and sap pigments occur arranged in the tissues, and
reference should be made to his paper if details are required. Generally
speaking, the anthocyanin pigments occur in the epidermis of the corolla,
and the chromoplastid either in the inner tissues, or in the epidermal,
or both, and the great variety of such combinations accounts to a large
extent for the numerous shades and tints. Dennert also observed that
when both plastid and soluble pigments occur in the same cell,

in] OF ANTHOCYANINS 41

especially in the papillae of the epidermis, the plastids tend to occupy
the base of the cell, while the soluble pigment is uniformly diffused.

Investigations on the distribution of anthocyanins in fruits have
been undertaken by Borbas (38) ; and in seeds by Sempolowski (36),
Preyer (66), Brandza, Lindinger (76) and Coupin (82).

Finally it may be as well to quote the Results which Gertz (19) has
given of a comparative examination of the localisation of anthocyanin
in members of many of the Natural Orders, and from which he finds
evidence of a certain amount of agreement between systematic relation-
ship and distribution of pigment. His results are shortly summarised
as follows :

Helobiae. In Alisma, Butomus and Hydrocharis subepidermal, in
Vallisneria epidermal.

Gramineae. Epidermal in Panicum, Oplismenus, Pennisetum, Cala-
magrostis, Set aria, Holcus, Weingaertneria, Catabrosa, Melica, Briza;
subepidermal in Pharus, Phleum, Alopecurus, Baldingera, Bromus,
Secede, Triticum, Avena, Aira.

Aroideae. Subepidermal all through and similarly for Lemnaceae.

Bromeliaceae. Chiefly in hypodermis but in some forms also in
epidermis.

Commelinaceae. Epidermal.

Juncaceae, Liliaceae and Amaryllidaceae. Subepidermal. In outer-
most scales of bulb of Allium often epidermal.

Scitamineae. Epidermal.

Orchidaceae. Epidermal in Cypripedium, Orchis, Epipactis, Limo-
dorum, Oncidium; subepidermal in Haemaria, Pogonia, Goodyera,
Microstylis, Reslrepia, Cattleya, Laelia, Dendrobium, Phalaenopsis.

Piperaceae. Anthocyanin in subepidermal water tissue or in spongy
parenchyma.

Salicaceae. Subepidermal all through.

Betulaceae. Periodic anthocyanin in ground tissue. Permanent in
epidermis.

Fagaceae. In general subepidermal. In certain species of Quercus
in hairs. In leaves of Copper Beech in epidermis.

Moraceae. Subepidermal in Ficus.

Aristolochiaceae. Epidermal.

Polygonaceae. In young leaves chiefly epidermal. In older, sub-
epidermal, but types somewhat mixed so that localisation rather
indefinite, and still more indefinite in Amarantaceae.

Nyctaginaceae. Aizoaceae and Portulacaceae. Epidermal.

42 THE HISTOLOGICAL DISTRIBUTION [CH.

Caryophyllaceae. In Alsinoideae always subepidermal ; in Silen-
oideae usually epidermal, at least as regards origin. Exceptions are
species of Dianthus, and, according to Wulff, Silene acaulis. Also
subepidermal in stem of Saponaria.

Ranunculaceae. Subepidermal nearly all through. Epidermal in
Paeonia coriacea, Anemone Hepatica, A. japonica and A. Pulsatilla.

Berber! daceae and Papaveraceae. Subepidermal.

Cruciferae. Subepidermal except in Camelina silvestris, Bray a,
Draba verna, Alyssum, Farsetia, Malcomia.

Crassulaceae. Anthocyanin epidermal ; in epidermal idioblasts and
in parenchymatous sheaths of vascular bundles.

Saxifragaceae. Localisation more indefinite in Saxifraga, but
appears to start from the epidermis, at least in young leaves. In
spring leaves of Ribes, epidermal; in autumnal, subepidermal.

Rosaceae. In spring leaves in general, anthocyanin in epidermis;
in autumnal leaves, anthocyanin in the ground tissue. Subepidermal
in spring leaves of Cotoneaster, Cydonia, Pyrus, Photinia, Amelanchier.

Leguminosae. In general subepidermal but found in the epidermis
in Acacia, Mimosa, Cercis, Gymnocladus, Gleditschia, Lupinus, Medicago,
Trifolium, Indigofera, Glycyrrhiza, Lourea, Onobrychis.

Geraniaceae. In young leaves epidermal, in older, subepidermal.

Tropaeolaceae, Linaceae, Polygalaceae. Subepidermal.

Euphorbiaceae. Generally subepidermal, epidermal in Croton,
Ricinus and some Euphorbia spp.

Anacardiaceae. In Pistacia and Rhus, anthocyanin epidermal in
spring leaves, subepidermal in autumnal leaves.

Aceraceae. Subepidermal in Acer, but epidermal in the red-leaved
variety.

Vitaceae. Spring leaves epidermal, autumnal subepidermal.

Tiliaceae. Subepidermal.

Malvaceae. Epidermal.

Theaceae, Hypericaceae and Violaceae. Ground tissue.

Begoniaceae. Epidermal in leaves, and sometimes in certain
spongy parenchyma cells.

Lythraceae and Myrtaceae. Subepidermal.

Oenotheraceae. In Oenotkera, Epilobium, Godetia, Gaura, Circaea,
usually epidermal. Subepidermal in autumnal leaves of Jussieua,
Chamaenerium, Fuchsia.

Umbelliferae. Subepidermal, but epidermal in Eryngium amethys-
tinum, and in stem of Chaerophyllum temulum, and Conium maculatum.

m] OF ANTHOCYANINS 43

Cornaceae, Ericaceae and Epacridaceae subepidermal. In Primu-
laceae in ground tissue only except in Androsace and Cyclamen where
it is epidermal.

Oleaceae. Usually subepidermal but in spring leaves of Forsythia,
Syringa, Chionanthus, Jasminum in epidermis.

Gentianaceae. Epidermal except in Menyanthes.

Apocynaceae, Asclepiadaceae and Convolvulaceae. Subepidermal,
but in Cuscuta often epidermal.

Boraginaceae. Altogether in ground tissue except in Sympliyium.

Labiatae. Marked epidermal localisation. Subepidermal in certain
Coleus varieties, and in autumnal leaves of Prunella vulgaris, Ballota
nigra, Betonica officinalis, Lycopus europaeus.

Solanaceae. Subepidermal.

Scrophulariaceae. Subepidermal in Verbascum, Linaria, Nemesia,
Limosella, or epidermal in Pentstemon, Veronica, Digitalis, Euphrasia,
Odontites, Rhinanfhus, Pedicularis, Melampynun.

Bignoniaceae and Orobanchaceae. Epidermal.

Gesneriaceae. Chiefly epidermal but often also hypodermal and in
the spongy parenchyma.

Lentibulariaceae. Epidermal in Pinguicula.

Acanthaceae. Varying localisation, usually epidermal.

Plantaginaceae. Epidermal.

Rubiaceae. Mesophyll in most cases.

Caprifoliaceae. Subepidermal except in Sambucus racemosa and
S. Ebulus.

Valerianaceae, Dipsaceae and Campanulaceae. In peripheral ground
parenchyma.

Compositae. Both epidermal and subepidermal localisation. In
older leaves nearly always the latter. In younger, subepidermal in
Solidago, Achillea, Matricaria, Centaurea, Leontodon, Taraxacum, Trago-
pogon and Scorzonera. Anthocyanin in hairs in Hebeclinium janthinum,
Eupatorium atrombens and Gynura aurantiaca.

CHAPTER IV

THE PROPERTIES AND REACTIONS OF ANTHOCYANINS

From time to time the question has been under discussion as to
whether all the varieties of red, purple and blue plant pigments- are
merely one and the same compound, the different shades being due
to the presence of other substances in the cell-sap, i.e. acids, alkalies,
etc., or whether the term anthocyanin includes many different members
of a great group. The earliest expression of opinion on this point is
possibly that made by James Smithson (112) in the Proceedings of the
Royal Society in 1818; here he remarks, on the slightest experimental
basis: "The colouring matter of the violet exists in the petals of red
clover, the red tips of those of the common daisy of the fields, of the
blue hyacinth, the holly hock, lavender, in the inner leaves of the
artichoke, and in numerous other flowers. It likewise, made red by
an acid, colours the skin of several plumbs, and, I think, of the scarlet
geranium, and of the pomegranate tree. The red cabbage, and the
rind of the long radish are also coloured by this principle. It is remark-
able that these, on being merely bruised, become blue; and give a
blue infusion with water. It is probable that the reddening acid in
these cases is the carbonic; and which, on the rupture of the vessels
which enclose it, escapes into the atmosphere." In the same way
Marquart (5), and Fremy & Cloez (126) originally recognised only
one blue pigment, from which the red pigments were believed to be
derived by action of acids. Of a similar opinion was Wigand (136),
who writes : " Die rothe und blaue Farbe der Bliithen sind, wie sich theils
aus den Uebergangserscheinungen, theils aus dem Auftreten beider
Farben als homogene Farbimg der Zellenfllissigkeit, theils aua dem
iibereinstimmenden Verhalten beider gegen chemische Reagentien
ergiebt, unwesentlich verschiedene Zustande eines und desselben Stoffes,
des Antliocyans." Hansen (11), also, believed most red flower colours
to be due to one substance.

CH. iv] PROPERTIES AND REACTIONS OF ANTHOCYANINS 45

Later, N. J. C. Miiller (169), on the basis of spectroscopic examina-
tion, announced that the red and blue pigments were of various kinds,
but it is doubtful whether his materials were pure. Wiesner (135) also
appeared to be uncertain as to whether all anthocyanins are alike,
whereas Weigert (179) definitely distinguishes two groups of anthocyanin
of which more will be said later. From this time onward, as investi-
gations increased, there seemed to be little doubt that a number of
substances are responsible for the different colours. Overton (333)
held this point of view, and we may quote the words of Molisch (104)
to the same effect: "dass der Begriff Anthokyan, wie er bisher in der
Literatur gefasst wurde, kein einheitliches chemisches Individuum
darstellt, sondern eine Gruppe von mehreren verschiedenen, wahr-
scheinlich verwandten Verbindungen." The more recent suggestion
of Grafe (209) comes nearer the truth, for he says we must regard
anthocyanin as a term to be used for a whole series of pigments, which
may have a similar fundamental nucleus, but which differ in the com-
plexes attached to the nucleus. Colour and other chemical reactions
would then depend on a particular grouping common to all or most of
the pigments. This view has been supported by the recent researches
of Willstatter (245, 256, 257) ; he has isolated anthocyanin from the
flowers or fruits of ten (or more) plants, and has shown that, in all
probability, they have the same fundamental structure. Some of the
number, though derived from plants quite unrelated, appear to be
identical; others differ in the number of their hydroxyl groups; others
again, he suggests, have their hydroxyls replaced by different radicals.
Thus we may now correctly consider the word anthocyanin to stand,
as a collective term, for a class of substances comparable to the sugars,
tannins, fats, proteins, etc.

Hence, in giving an account of the reactions of anthocyanin, one is
always dealing with a large group of substances of which the properties
may differ considerably. Therefore to make any general statement
in some cases is difficult, and there is always the further consideration
that the substances examined have rarely been obtained pure.

The appearance of anthocyanin as crystals in the living tissues
has been discussed in Chapter in. Outside the cell anthocyanin has
also been obtained in undoubted crystalline form. Molisch (104)
prepared crystals very readily from the petals of the scarlet Geranium,
Pelargonium zonale, by placing a petal in distilled water under a cover-
slip on a slide. The pigment diffuses out, and on slow evaporation
deposits groups of beautiful needle-shaped crystals. By a similar

46 THE PROPERTIES AND [CH.

method, using acetic acid instead of water, Molisch (104) obtained good
crystals from petals of garden roses and of Anemone fulgens. Several
other investigators have crystallised anthocyanin from extracts ; the
cases may be mentioned here though the methods of preparation will
be given in the next chapter. From the red pigment of Vine leaves,
Gautier (175) isolated two substances termed by him a- and /3-ampelo-
chroi'c acids. From a solution in hot water, the ct-acid was obtained
on cooling as a red crystalline powder. The ^8-acid also was deposited
from water in red crystals on slow evaporation. Griffiths (191) prepared
crystals from pigment of 'Geranium' (Pelargonium) flowers. From
flowers of Althaea rosea, Grafe (197) isolated a deep red pigment which
separated out from alcohol in minute crystalline plates. Portheim &
Scholl (204) also succeeded in crystallising the anthocyanin from the
testa of seeds of Phaseolus multiflorus. Later, Pelargonium flowers
were again employed by Grafe (222) as material for the purification
and analysis of anthocyanin, and like Molisch, Grafe found that this
pigment very readily crystallised. The facility with which Pelargonium
pigment may be made to crystallise, as compared with other antho-
cyanins, is no doubt due to the difference in its chemical nature. Not
only as regards the scarlet colour, but also in its reactions towards
reagents, the Pelargonium pigment differs from the more universally
distributed purples and purplish reds.

Before describing the solubilities of anthocyanin, it should be
mentioned that the pigment usually exists in the plant in the form
of a glucoside. In this form it has been isolated from fruits of the
Bilberry (Heise, 178), and from flowers of Althaea, Pelargonium (Grafe,
197, 222) and Centaurea (Willstatter, 245). There is little doubt, as
will be shown in later chapters, that anthocyanins are aromatic sub-
stances containing hydroxyl groups, and, as is well known, hydroxyl
groups in plant products are frequently replaced by sugars. As a
glucoside, anthocyanin is readily soluble in water, and since it is in
this form that the pigments chiefly occur in the cell, they can be
extracted with water. After hydrolysis, in the non-glucosidal state,
the pigment is far less soluble in water, and in some cases almost or
quite insoluble, i.e. Antirrhinum (Wheldale & Bassett, 254) and Bilberry
(Heise, 178). To the consideration of these glucosides we shall return
again later in the chapter.

In ether, anthocyanin is insoluble, as also in benzene, carbon bisul-
phide, chloroform and similar solvents in which plastid pigments are
soluble. In alcohol, the greater proportion of anthocyanins are soluble ;

iv] REACTIONS OF ANTHOCYANINS 47

there are definite exceptions, such as those of the Amarantaceae,
Chenopodiaceae and Phytolaccaceae which are entirely insoluble in
this solvent. To these may be added the blue pigment of Centaurea
(Willstatter, 245), the glucosidal pigment of Althaea (Grafe, 197) and
probably others. In many flowers it is difficult to extract the petals
completely with alcohol. This is possibly due in some cases to the
presence of several pigments, certain of which are insoluble in alcohol ;
or there may also be retention of the pigment to some extent by the
coagulated cell-contents. A curious phenomenon is connected with
the alcohol solutions of most anthocyanins ; such solutions, though
at first coloured red or purple as the case may be, somewhat rapidly
lose their colour and eventually become quite colourless ; the same
effect is produced by immersing petals in strong alcohol. The colour
returns on evaporation of the alcohol or, in many cases if the solution
is sufficiently strong, on adding water. A few drops of acid, also,
restore the colour completely ; similarly a few drops of alkali produce
the green (or yellow) reaction characteristic of anthocyanin. This
phenomenon was first remarked upon by Nehemiah Grew (1): "Again
though no Blew Flowers, that I know of, will give a Blew Tincture to
Spirit of Wine : yet having been for some days infused in the; said
Spirit, and the Spirit still remaining in a manner Limpid, and void of
the least Ray of Blew, if you drop into it a little Spirit of Sulphur,
it is somewhat surprizing to see, that it immediately strikes it into a
full Red, as if it had been Blew before: and so, if you drop Spirit of
Sal Armoniac or other Alkaly upon it, it presently strikes it Green....
It is likewise to be noted, That both Yellow and Red Flowers give a
stronger and fuller Tincture to Water, than to Spirit of Wine; as in
the Tinctures of Cowslip, Poppys, Clove- July-Flowers and Roses, made
both 'in Water and Spirit of Wine, and compared together, is easily
seen." Loss of colour in alcohol \vas also mentioned by Morot (122),
Filhol (125) and Fremy & Cloez (126). In 1884, Hansen (11) commented
on it, and suggested that anthocyanin, in absolute alcohol, forms a
colourless anhydride. Keeble & Armstrong (239) have recently offered
another, though unsatisfactory, explanation of this, decolorisation (see
p. 120). Willstatter (245) considers the loss of colour to be due to
the formation of a colourless isomer (see p. 72).

As regards the appearance, colour, melting point, crystalline form,
etc., of solid anthocyanin, the accounts of the few workers who have
prepared the pigment are so varied that information is best obtained
by reference to individual cases quoted in the next chapter.

48 THE PROPERTIES AND [CH.

Qualitative Reactions.

With respect to qualitative reactions, the necessity, already men-
tioned in connection with the properties of anthocyanin, for a guarantee
of the purity of the pigment used is of paramount importance. Crude
extracts invariably contain other substances which may modify, or
completely alter, the reactions of the pigment itself. In the next
chapter, accounts are given of special methods for purification and
analyses of authocyauin, and on studying these, the futility of applying
tests to any but pure material will be recognised at once. It "is only
after careful extraction, and purification by means of analyses, that
the reactions of any pigment can be determined with certainty, and
qualitative tests on impure extracts are to a large extent worthless.
Nevertheless, numbers of observations have been made on more or
less impure material, and the following account deals with the more
important results.

With alkalies. When an aqueous or alcoholic extract of anthocyanin
is treated with alkali, the pigment turns green 1 and often finally
yellow. Sometimes a blue colour precedes the green ; with very dilute
alkali, or with solutions of salts having a weakly alkaline reaction, a
blue colour only may appear. Similarly red, purple and blue flowers
placed in ammonia vapour as a rule turn green.

With acids. Anthocyanins almost invariably turn bright red with
acids, though the shade may vary in different cases.

With lead acetate. Anthocyanin extracts are generally precipitated
by lead acetate, and the colour of the precipitates is usually some shade
of green or blue; occasionally it is red.

The reactions of anthocyanins with acids and alkalies have formed
a subject for discussion from time to time. The whole matter is so
bound up with the views of those who have worked on the pigments
that something of the nature of a historical summary must be included.

The question first to be considered is whether the green coloration
given when tissues and crude extracts containing anthocyanin are
treated with alkalies is a reaction of anthocyanin alone, or is the com-
bined result of reactions with anthocyanin plus reactions with other
substances present in the cell or solution, and this can only be determined
satisfactorily by testing pure pigments. In most cases, when white
flowers are treated with alkalies, or exposed to ammonia vapour, a bright
yellow colour is developed, indicating a reaction of alkalies with a class

1 In some cases such solutions are slightly dichrolc, green and red.

iv] REACTIONS OF ANTHOCYANINS 49

of substances known as flavones which are almost universally distri-
buted in plants. The flavone pigments in bulk are yellow, but exist
in the cell-sap in such small quantities as to be inconspicuous except
when treated with alkali, when an intense yellow colour is developed;
with ferric chloride solution they give a green or brown coloration.
They, moreover, occur in the plant largely as glucosides, in which form
they are readily soluble in alcohol and water, and hence they are present
in all crude aqueous and alcoholic extracts of anthocyanin. Extracts
of white flowers, or in fact of any non-anthocvanic parts, give as a
rule yellow or orange-yellow precipitates with lead acetate, which are
insoluble flavone salts of lead.

That there is some substance in white flowers which turns yellow
with ammonia was noticed by Boyle (107): "we thought fit to make
Trial upon the Flowers of Jasmin, they being both White as to Colour,
and esteem'd to be of a more Oyly nature than other Flowers. Where-
upon having taken the White parts only of the Flowers, and rubb'd
them somewhat hard with my Finger upon a piece of clean Paper,...
a strong Alcalizate Solution, did immediately turn the almost Colour-
less Paper moisten'd by the Juice of the Jasmin, ...a, Deep, though
somewhat Greenish Yellow,... when we try'd the Experiment with the
Leaves of those purely White Flowers that appear about the end of
Winter, and are commonly call'd Snow drops, the event, was not much
unlike that, which, we have been newly mentioning." Later in the
Comptes Rendus of 1854 and 1860, Filhol (125, 132) published papers
of considerable interest in connection with this point. Filhol found
that when white flowers of Viburnum Opulus, Philadelphus coronarius
and other plants were exposed to ammonia, they turned yellow ; as
yellow, he says, as the flowers of Laburnum. The same results he
observed in leaves in the parts free from chlorophyll. When the
flowers, after treatment with alkali, were placed in acidified water,
they became white again. The substance which gives the yellow
colour was found to be soluble in water and alcohol and slightly so in
ether, and Filhol terms it xanthogene.' When coloured, red or purple,
flowers were treated with ammonia, they turned green as a rule, but
in some cases blue (Papaver, Pelargonium, Salvia splendens}. In one
particular experiment when he added aluminium hydroxide to an
extract of Vervain flowers, the aluminium hydroxide became yellow
but the supernatant liquid retained the purple colour. Thus he comes
to the conclusion that the green coloration with alkalies in Viburnum
and other flowers is due to a mixture of a blue colour given by

w. P. 1

50 THE PROPERTIES AND [CH.

anthocyanin plus a yellow colour given by xanthogene. But in the
case of the flowers of Pelargonium, Papaver and Salvia the xanthogene is
absent and so the anthocyanin becomes blue or violet with alkalies. In the
later paper (132) he remarks on the resemblance of xanthogene to luteolin
(which we now know to be a flavone), but he was unable to establish
the identity. Thus Filhol's investigations brought him very near to
the truth. Similar views were advanced by Wiesner (135) in 1862.
From a series of reactions given by various flower pigments he concludes
that colourless sap contains, as a rule, a tannin giving a green reaction
with iron salts, and a yellow colour with alkalies. Like Filhol he
believes that anthocyanin itself gives a blue, never a green, reaction
with alkalies, and the green colour is due to admixture with yellow given
by the tannin. In plants free from tannins giving the yellow reaction,
anthocyanin turns blue with alkali. Wiesner's tannins are probably
for the most part flavones, since true tannins are rare in flowers. These
views on the reactions of anthocyanin gave rise to a certain amount
of controversy, for Wigand (136) and Nageli & Schwendener (138)
held the opinion that the green coloration is given by anthocyanin
itself, and may appear when the iron-greening tannins are absent.
Some of the arguments involved in the discussion are given by Wiesner
in a later paper (142). The alkali reaction of anthocyanin is also
mentioned by Overton (333), who considers the blue colour to be due
to the formation of an acid salt, the green colour to a neutral salt of
the pigment, anthocyanin itself being a dibasic acid. This view is
accepted by Grafe (197) as being in accordance with the reactions of
Althaea pigment which he prepared in a pure state. Anthocyanin
from Antirrhinum (Wheldale & Bassett, 254), purified from accompanying
substances, still gives the green alkali reaction. Willstatter (245), as
far as can be determined from his publications, considers the reactions
of anthocyanin to be as follows. The pure blue anthocyanin from the
Cornflower is not altered by alkali, i.e. pure anthocyanin gives a blue
colour with alkalies, but if a solution of the blue pigment has stood for
a time, the colour reaction with alkalies is green. This is due to the
fact that from the pigment a colourless isomer (see pp. 53, 72, 75, 78)
has been formed, and this gives a yellow colour with alkalies; hence
the blue plus yellow results in a green reaction. Crude extracts from
the flowers, he says, also give a green colour owing to the presence of
yellow pigments obviously flavones. (Die reine blaue Farbstoff-
losung zeigt auf Zusatz von wenig Soda zunachst keine Farbanderung,
eine gestandene Losung wird hingegen griinblau oder blaugriin weil

iv] REACTIONS OF ANTHOCYANINS 51

die Losung nun auch die farblose Modifikation enthalt. Dies ist das
Verhalten eines wassrigen Bliitenauszugs, worin sich iiberdies noch
gelbe FarbstofTe befinden, deren Alkalisalze intensiv gelb sind.) The
blue colour given by alkalies with the pure anthocyanin pigment will,
Willstattei found, become green or yellow by decomposition on standing,
or with excess of alkali. Further consideration will be given below
(pp. 52-54) to these reactions. With anthocyanins from other plants
Willstatter notes various reactions with alkalies (see p. 56). Hence
we have at present the following suggestions. Pure anthocyanin from
Centaurea gives a blue colour with alkalies; the green colour given in
solutions and crude extracts is due to mixture with the yellow colour
produced by the colourless isomer or accompanying flavones, or both.
Pure anthocyanin from Antirrhinum gives a green colour with alkalies,
and this cannot be due to admixture with flavones, as the latter are
removed by purification ; nor can we suppose it due to admixture with
a colourless isomer, since this is not formed in a strongly acid solution
such as that from which the pigment separates out in preparation.
Thus, whether the green or blue reaction is given by pure anthocyanin,
or whether it is green in some cases, and blue in others, remains undeter-
mined until anthocyanins from many other species have been purified
and tested. In many cases the green reaction either rapidly or slowly
changes to yellow, and the original colour does not return on neutralisa-
tion, so that evidently some anthocyanins are completely destroyed by
alkalies.

The same difficulty arises with regard to the precipitates with lead
acetate, for the accompanying flavones produce yellow or orange-
yellow precipitates with lead acetate, and hence the actual lead salt
of anthocyanin is not identifiable except from the pure pigment with
which the results are the same as with alkalies. The colour of the
lead precipitate varies very considerably in crude solutions. In extracts
from white flowers containing little anthocyanin, it is greenish-yellow ;
from flowers containing much anthocyanin, bright green or blue-green,
whereas from red varieties (Wheldale, 211), it is practically red with a
greenish tinge. The latter must be distinguished from the red preci-
pitates given by the special pigments of the Amarantaceae and Phyto-
laccaceae which are mentioned below.

On the basis of qualitative reactions, Weigert (179), in 1895,
attempted a classification of anthocyanins into two groups:

The 'Weinroth' group, of which the pigments give blue-grey or
blue-green precipitates with basic lead acetate; give the Erdmanu

42

52 THE PROPERTIES AND [CH.

reaction (see below) ; are precipitated and give a bright red colour
with concentrated hydrochloric acid; and turn green on addition of
alkalies. Ex. Pigments from Vitis, Ampelopsis quinquefolia, Rhus
typhina, Cornus sanguined, etc.

The ' Riibenroth' group, of which the pigments give a red precipitate
with basic lead acetate, do not give the Erdmann reaction ; turn dark
violet with concentrated hydrochloric acid; violet with ammonia, but
with other bases yellow. Ex. Pigments from Beta, Iresine Lindeni,
Achyranthes Verscliaffeltii, Amaranthus, Atriplex hortensis, Pliytolacca
decandra. According to Gertz (19), the following should be added to
Weigert's ' Riibenroth ' group : several Chenopodiaceae (Blitum virgatum,
Atriplex litoralis, Corispermum canescens), Amarantaceae (except
Mogiphanes brasiliensis), Nyctaginaceae (Oxybaphus nyctagineus),
Phytolaccaceae (Phytolacca decandra, leaves), Aizoaceae (Tetragonia
crystallina, Mesembryanthemum nodiflorum), Portulacaceae (Portulaca
grandiflora) and Basellaceae (Basella rubra). The remaining families
of the group Centrospermae, i.e. Polygonaceae and Caryophyllaceae,
seem to be distinguished by anthocyanin of the 'Weinroth' group
as also the greater number of the Chenopodiaceae.

The reaction of Erdmann (618) was originally employed in order
to detect true wine pigment and may be described as follows : The
pigment solution (wine) is diluted with four times its volume of water,
eight drops of concentrated hydrochloric acid are added, and the
mixture shaken up with 16 c.c. of amyl alcohol. The amyl alcohol
separates out with a fine violet-red colour, the underlying acid solution
being yellow or cherry-red. If the amyl alcohol is separated off, and
an equal volume of water added, together with two drops of concentrated
ammonia, the amyl alcohol decolorises and the underlying solution
becomes bright green. If the original acid solution below the amyl
alcohol is placed in a porcelain dish and carefully neutralised with
dilute ammonia, at neutralisation point an indigo-blue colour is produced
which afterwards becomes green.

The reactions of anthocyanins with acids and alkalies, the Erdmann
reaction, etc., have received a new interpretation through the recent
researches of Willstatter (245). These views are based upon work
devoted in particular to the anthocyanin of Centaurea but also to
anthocyanins in general. As far as can be gathered from a preliminary
publication, the following represents, in the main, the views of Will-
statter (see also Chapter v).

1. Red, purple and blue pigments occur in the plant entirely as

iv] REACTIONS OF ANTHOCYANINS 53

glucosides (anthocyanins), which can be hydrolysed artificially with
the formation of sugar and non-glucosidal pigments (anthocyanidins).
From a dilute acid solution anthocyanidins are completely removed
by amyl alcohol, whereas anthocyanins are not taken up at all by this
solvent. This may be demonstrated experimentally by extracting
fresh material containing anthocyanin with dilute sulphuric acid.
After nitration and addition of amyl alcohol, no pigment is taken up
by the alcohol. But if the solutions are heated for one half to three
quarters of an hour on a water-bath, the anthocyanins are hydrolysed,
and on addition of amyl alcohol, the anthocyanidins are quantitatively
removed.

2. Anthocyanin is itself an acid and in the free state is purple.

3. There is a blue modification which is the potassium salt of the
purple.

4. There is a red modification, which is the oxonium salt of antho-
cyanin; the pigment may be combined with either organic or inorganic
acids.

5. In Centaurea, as well as in some other plants, all three forms
of anthocyanin readily change to a colourless isomer; with the red
form this only occurs in absence of excess of acid. The change to a
colourless isomer can be prevented by adding neutral salts to the
anthocyanin solution ; the anthocyanin forms additive compounds
with these substances thereby preventing the isomeric change.

As regards the colour reactions with alkalies, Willstatter, as we
have already seen, gives the following explanation. With alkalies
a blue salt is formed, which may become green owing to mixture with
flavones or the colourless isomer (see pp. 50, 72), if these are present
in the extract. The blue salt is also unstable, and with excess of
alkali passes to a greenish or even yellow decomposition product. But
if a neutral mineral salt is present, the blue salt is rendered stable.
This may be brought about by either (1) acidifying the anthocyanin
and then neutralising or (2) by adding the neutral salt (NaCl, NaN0 3 ).
Thus, for example, if pigment of Bilberries or Grapes (which presumably
contain little flavone) are treated with alkali, it gives a greenish colour.
But if treated first with some salt, it gives a blue colour, and the same
result may be brought about by acidification and subsequent neutralisa-
tion.

It is now possible to explain the Erdmann reaction, which, according
to Willstatter, has hitherto been misunderstood. The following account
is given more or less in Willstatter's words. The Erdmann reaction

54 THE PROPEKTIES AND [CH.

is based on the fact that new, or fairly new, wine gives a green colour
on neutralisation with ammonia, but after treatment with hydrochloric
acid and neutralisation, a dark greenish-blue colour. This behaviour
was incorrectly explained by Erdmann as being due to the splitting
of the wine pigment into two pigments by the hydrochloric acid.
Erdmann separated the above two pigments by shaking with amyl
alcohol ; the violet-red amyl alcohol layer gives with dilute ammonia
first a bright green, then a brownish-green colour, while the acidified
water solution becomes indigo-blue on neutralisation. As a matter
of fact no breaking up is brought about by the acid, but during fermen-
tation of the grape juice, a portion of the anthocyanin has been hydro-
lysed to anthocyanidin. The effect of ammonia on the wine pigment
is to give a blue coloration rapidly passing to a green decomposition
product. But, as explained by Willstatter, if the anthocyanin is first
acidified and then made alkaline, the blue colour is more stable. On
shaking up the acidified wine with amyl alcohol, the latter takes up
the small portion of anthocyanidin as an oxonium salt. If the wine
is heated with hydrochloric acid, or if the hydrochloric layer of the
Erdmann reaction is heated, the whole is hydrolysed, and the antho-
cyanidin can be extracted quantitatively with amyl alcohol. But the
hydrolysis does not happen in the cold. Willstatter further points out
a source of error in the reaction. Grape juice from fresh berries gives,
after acidification with hydrochloric acid, a little pigment in the amyl
alcohol. If the amyl alcohol is then washed with dilute sulphuric
acid, it is nearly decolorised. The small amount of pigment in the
amyl alcohol is not hydrolysed, but is due to the fact that hydrochloric-
amyl alcohol takes up a little of the anthocyanin oxonium salt, whereas
sulphuric-amyl alcohol takes practically none. Hence it is advisable
to use sulphuric acid for the test. The statement by Weigert, that the
'Weinroth' group gives the Erdmann reaction, is regarded by Will-
statter as erroneous, for the latter maintains that, apart from the small
amount of glucoside salt which can be washed out again from the
amyl alcohol, the pigment of flowers, berries and leaves remains com-
pletely in the water-acid layer.

The sensitiveness of anthocyanin to acids and alkalies has suggested
its use as an indicator. In fact these were the first reactions to attract
the attention of chemists (see p. 8). Its use in this way has been
revived from time to time by Pellagri (146). Sacher (215) and others
but without any permanent success.

Reactions with iron salts. Here again the actual colour reaction

iv] REACTIONS OF ANTHOCYANINS 55

is difficult to estimate unless the pure pigment is tested. The flavones
give green or brownish-green colorations with iron salts, the tannins
green or blue. Though anthocyanins on the whole appear to give
green-blue, green, or sometimes brown reactions, if tannin or flavone
be present it is obvious that no reliance can be placed on the results.
In a few cases, however, where anthocyanin pigment has been obtained
approximately pure, one of the above colour reactions was given.
Such a result indicates that anthocyanin has probably hydroxyl groups
of the phenol type.

Reactions with sodium bisulphite. An interesting reaction of antho-
cyanins is that given with sulphur dioxide and bisulphites It was
well known at a very early date that flowers containing anthocyanin,
or extracts of the pigment, are bleached by sulphur dioxide gas, and
that the colour is again restored by stronger acids. Boyle (107) writes :
'That Roses held over the Fume of Sulphur, may quickly by it be
depriv'd of their Colour, and have as much of their Leaves, as the Fume
works upon, burn'd pale, is an Experiment, that divers others have
tried, as well as I. But (Pyrophilus) it may seem somewhat strange
...That, whereas the Fume of Sulphur will,... Whiten the Leaves of
Roses; That Liquor, which is commonly call'd Oyl of Sulphur... does
powerfully heighten the Tincture of Red Roses." Further observations
on this bleaching action were made by Kuhlniann (118), Hiinefeld (120)
and Schonbein (123). Solutions of anthocyanin decolorised by sulphur
dioxide have been employed by Kastle to test the relative 'strengths'
of acids. Kastle (196) is of the opinion that the decolorisation is not
caused by reduction, and the same view is held by Grafe (197, 222),
Avho prepared bisulphite derivatives from the anthocyanins of both
Althaea and Pelargonium by addition of sodium bisulphite. Both
products were colourless, but the red colour returned on addition of
a trace of a stronger acid. Grafe concludes that the anthocyanins
contain aldehyde colour-producing groups, which form additive com-
pounds with bisulphites, whereby the linkiugs in the molecule and
the resultant colour are changed.

Action of nascent hydrogen. A reaction which would appear to be
one of reduction is that produced by treating acid solutions of antho-
cyanin with zinc dust. The colour rapidly disappears and the solution
remains colourless if air be excluded. On exposure to air, if the reducing
action is not very violent, the colour returns, the surface of the liquid
becoming coloured before the deeper layers. Kastle (196) does not
consider the reaction to be of the nature of reduction, since the colour

56 THE PROPERTIES AND [CH

did not return on treatment with oxidising enzymes. Grafe (222)
holds the view that the loss of colour is due to changes brought about
in the aldehyde groups (which he postulates) by the action of nascent
hydrogen. The fact that the return of colour in air is not equally
great with all acids (Wheldale & Bassett, 621) may indicate that the
reaction is not a simple reduction process.

Compounds with acids. In some cases anthocyanins appear to form
definite compounds with acids (Willstatter's oxonium salts), since such
compounds occur in crystalline form. Grafe (222) obtained the antho-
cyanin from Pelargonium in combination with two molecules of acetic
acid as a crystalline substance. Willstatter (245) also obtained both
anthocyanin, and the corresponding anthocyanidin. in combination
with hydrochloric acid as crystalline compounds.

Spectrum of anthocyanin. A considerable amount of attention
has been devoted to the spectroscopic examination of flower and leaf
pigments. Sorby (139, 144), Miiller (169), Engelmann (394), Lepel (151)
and Formanek (186) may be mentioned as workers on these lines;
but the results are of little value for identification, or otherwise, on
account of the impurity of the products employed, that is the doubt
as to the number of pigments present, etc.

Willstatter (245) has distinguished various groups of anthocyanins
by their different behaviour to reagents, though the observations do
not pretend to include any kind of systematic classification. The
following represent some of the classes :

1. Red in acid solution, blue with soda and a blue precipitate
with lead acetate ; pigment readily isomerises to a colourless modifica-
tion (Centaurea, Rosa, Lathyrus).

2. Red in acid solution, blue with soda and a blue precipitate with
lead acetate; pigment decolorises less readily or not at all (Grapes,
Bilberries, flowers of Delphinium).

3. Yellowish-red in acid solution, blue with soda, red precipitate
with lead acetate (Radish).

4. Yellowish-red in acid solution (Pelargonium) and blue-red
(Papaver) ; both violet with soda and decolorised by isomerisation.

5. Red in acid solution ; with soda red in dense layers, blue-green
in thin layers (Pinks) or red-violet to red-brown (Aster).

6. Violet in acid solution, red with soda, red precipitate with lead
acetate (Beet-root, Atriplex).

The author (211, 212) has also made observations on crude extracts
of anthocyanins from a very large number of flowers, using the colour

iv] REACTIONS OF ANTHOCYANINS 57

reactions given by various chemical reagents, i.e. sulphuric, hydro-
chloric and oxalic acids, ammonia, caustic potash, lime water, ferric
chloride, ferrous sulphate, potassium ferrocyanide, uranium, lead,
copper and sodium acetates, stannous chloride and others. Although
a certain amount of differentiation was possible on this basis of quali-
tative reactions, it was soon found that there were too many aberrant
and peculiar forms of pigment to arrive at any satisfactory classification.

Before we close the chapter, there is yet another extract which may
well be quoted from the writings of Boyle (107), since it dealt two
hundred and fifty years ago with some of the phenomena which have
formed the basis of Willstatter's constitutional formulae for the cyanidin
of the Cornflower, i.e. the reactions of anthocyanin with acids and
alkalies, and its instability in water solution.

Boyle writes : ' There is a Weed, more known to Plowmen than
belov'd by them, whose Flowers from their Colour are commonly call'd
Blew-bottles, and Corn-weed from their Growing among Corn. These
Flowers some Ladies do, upon the account of their Lovely Colour, think
worth the being Candied, which when they are, they will long retain
so fair a Colour, as makes them a very fine Sallad in the Winter. But
I have try'd, that when they are freshly gather'd, they will afford a
Juice, which when newly express'd, (for in some cases 'twill soon enough
degenerate) affords a very deep and pleasant Blew. Now, (to draw this
to our present Scope) by dropping on this fresh Juice, a little Spirit of
Salt, (that being the Acid Spirit I had then at hand) it immediately
turn'd (as I predicted) into a Red. And if instead of the Sowr Spirit
I mingled with it a little strong Solution of an Alcalizate Salt, it did
presently disclose a lovely Green;... And I remember, that finding
this Blew Liquor, when freshly made, to be capable of serving in a
Pen for an Ink of that Colour, I attempted by moistning one part of
a piece of White Paper with the Spirit of Salt I have been mentioning,
and another with some Alcalizate or Volatile Liquor, to draw a Line
on the leisurely dry'd Paper, that should, e'vn before the Ink was dry,
appear partly Blew, partly Red, and partly Green."

CHAPTER V

THE ISOLATION AND CONSTITUTION OF ANTHOCYANINS

Several general and rather vague views have been held as to the
constitution of anthocyanin, without any particular experimental
evidence. Thus Wigand (136) believed these pigments to arise by
oxidation from a colourless tannin-like chromogen; Overton (333)
considered them to be tannin-like substances combined with sugar,
and Katie (354), too, found that anthocyanin gave the reactions of a
tannin. Palladia (203), again recognised in anthocyanin a respiratory
pigment, and yet other suggestions have been advanced by Filhol,
Mirande and Combes.

But in the following cases, definite isolation of the pigments has
been attempted, and analyses have been made; the methods and
results, however, are so varied that a separate account is essential
in each case. We may enumerate the cases thus :

1849. Morot (122). The flower-pigment of Cento-urea Cyanus.

1858. Glenard (129, 130). The colouring matter of wine.

1877. Church (147). The pigment from leaves of Coleus.

1877. Senier (148). The flower-pigment of Rosa gaUica.

1878. Gautier (149). The colouring matter of wine.
1889. Heise (167). The colouring matter of grapes.
1892. Gautier (175). The pigment from red Vine leaves.
1892. Glan (176). The flower-pigment of Althaea rosea.
1894. Heise (178). The pigment from fruits of the Bilberry.
1903. Griffiths (191). The flower- pigment of Pelargonium.

1906 and 1909. Grafe (197, 209). The flower-pigment of Althaea rosea.

1911. Grafe (222). The flower-pigment of Pelargonium.

1913 and 1914. Wheldale (244, 254). The flower-pigment of Antirrhinum majus.

1913. Willstatter (245). The flower-pigment of Centaurea Cyanus.

1914. Willstatter (256, 257). The flower-pigment of Delphinium.

Hollyhock (Althaea rosea).

Mallow.

Pelargonium.

Rosa gaUica.

The pigment from fruits of the Bilberry.

,, ,, ,, Cranberry.

from grapes.

CH. v] ISOLATION AND CONSTITUTION OF ANTHOCYANINS 59

Morot (122), 1849. The flower-pigment of Centaurea Cyanus.
The pigment was prepared by pouring a water extract of the flowers
into alcohol. The blue flakes of the precipitate, when collected and dried,
formed a blue powder, which, on analysis, gave the following results :

C

H

(i)

36-62 %

5-04 %

58-54 %

(ii)

38-76 %

5-30 %

55-94 %

(iii)

37-00 %

5-24 %

57-76 %

(iv) 37-03 % 5-26 % 57-71 %

Glenard (129, 130), 1858. The colouring matter of wine.

The pigment was prepared by first adding lead acetate to wine.
The blue precipitate of the lead salt was then treated with ether con-
taining hydrochloric acid gas, which set free the pigment. The latter
was then extracted with rectified spirit, the solution evaporated, and
water added which precipitated the pigment in red flakes, since it is
scarcely soluble in water, The product was not crystalline ; on analysis
the results were :

C H

From anthocyanin ... 57-02% 4-89% 37-89%
Calculated for C 2 HO ... 57-1% 4-8% 38-1%

The pure lead salt was also prepared by adding lead acetate to
a dilute alcoholic solution of the pigment. The product, washed and

dried at 120, gave:

C H O

Lead salt of pigment ... 59-67 % 4-49 % 35-84 %
Calculated for C 20 H 9 9 ... 59-71% 4-47% 35-82%
Hence Glenard gave to the pigment, which he termed oenoline,
the formula C 20 H 9 9 . OH, and to its salt, the formula C 20 H 8 9 . PbO.

Church (147), 1877. Pigment from the leaves of Coleus.
The pigment was extracted with cold alcohol acidified with sulphuric
acid ; the acid was then precipitated with baryta, and the solution
concentrated. The product was further purified by extraction with
alcohol and precipitation of the alcohol solution with ether or water.
Church termed the pigment cole'in and gave it the composition C 10 H 10 5 ;
its lead salt he expressed as C 20 H 18 Pb0 10 .

Senier (148), 1877. The flower-pigment of Eosa gallica.
The dried petals were first digested with ether, and the pigment
then extracted with alcohol and precipitated with lead acetate. The

60 THE ISOLATION AND [CH.

lead salt was decomposed either by sulphuretted hydrogen or sulphuric
acid. Microscopic crystals of the sodium, potassium, and ammonio-
sodium and potassium salts were obtained by evaporating the pigment
with alkali in a drop of alcohol. Senier gave to the lead salt, as a result
of several analyses, the composition Pb 2 C 21 H 29 30 .

Gautier (149), 1878. The colouring matter of wine.

Gautier was of the opinion that the colouring matters are formed
from tannins which become red on oxidation. He also suggested that
homologous series of colouring matters exist, such as C 20 H 20 10 , C 21 H 20 10 ,
etc., and that each variety of Vine contains one or more members of
the series. The method he employed for isolation was to decompose
the lead salts of the pigments with hydrochloric acid ether, and then
to take up with alcohol and to precipitate with water. It has been
shown later by Heise that Gautier's pigments were mixtures.

Heise (167), 1889. The colouring matter of grapes.

Heise maintains that there are two pigments present in the skins
of purple grapes. One, soluble in alcohol, which he termed B, is
chiefly responsible for the colour ; the other, almost insoluble in alcohol,
he termed A. In wine, the soluble pigment B is said to be converted
by oxidation into A, which is then precipitated from the wine.

The method described for the isolation of the pigments is to preci-
pitate the extract of fresh skins with lead acetate. The mixture of
the lead salts is then decomposed with hydrochloric acid ether, washed
with ether, taken up with methyl alcohol and precipitated from the
alcohol solution by adding ether. The product is again taken up in
alcohol and poured into water. The precipitate so formed is a mixture
of the two pigments which can then be separated by digesting the dried
product with absolute alcohol in the cold. Heise devises two methods
for separating the pigments. (1) The mixture of the pigments already
isolated is treated as above with absolute alcohol ; B goes into solution,
whereas A is insoluble. (2) The lead salts of the isolated pigments
are treated with acetic acid ; the lead salt of B is soluble in that solvent ;
the salt of A insoluble.

Pigment A is a brown-black substance, insoluble in absolute alcohol,
ether, water and acetic acid ; but in alcohol containing a trace of acid
it is soluble to a red solution.

Pigment B is soluble in alcohol to a brownish solution which, on
addition of acid, becomes red with a tinge of violet.

v] CONSTITUTION OF ANTHOCYANINS 61

Heise is, apparently, not able to suggest any formulae for the pig-
ments on the basis of analysis.

Gautier (175), 1892. The pigment from red Vine leaves.

The extract of the pigment from the Vine leaves was precipitated
with neutral lead acetate until the colour of the precipitate becomes
blue. This first precipitate was then separated off, and precipitation
continued with the formation of a dark-green precipitate. The latter
precipitate was then decomposed with sulphuretted hydrogen, treated
with ether and taken up in 95 % alcohol. A red product was obtained
consisting of two substances and having, it is said, the characteristic
properties of a tannin. These substances are termed by Gautier
ampelochroic acids ; one was found to be insoluble in cold water
(a-ampelochroic acid) : the other soluble (/3-ampelochroiic acid).

a-ampelochroic acid. After washing away the -/3-pigment with cold
water, the insoluble residue was taken up in hot water from which, on
cooling, the pigment separated out as a red crystalline powder. On
analysis the results were :

C H

56-43 % 3-96 % 39-61 %

from which the formula C 19 H 16 10 is derived.

According to Gautier, the a-acid is dibasic, and the normal lead
salt is dark green. The pigment is soluble in hot water and cold alcohol,
but is insoluble in ether. With ferric salts it gives a green-black
coloration.

jS-ampelochroi'c acid. This is the portion soluble in cold water;
on slow evaporation, red crystals are deposited. The product is soluble
in water; the solution precipitates gelatine, and gives a violet-black
coloration with ferric salts. On analysis the results were:

C HO

(i) 53-89% 4-34% 41-77%
(li) 53-96 % 4-29 %

with which the two following formulae agree most closely:

c i7Hi 6 10 or C 26 H 24 15

After determination of the molecular weight by means of the neutral
zinc salt, Gautier decides in favour of the formula C 26 H 24 15 .

From the first precipitate originally obtained with lead acetate,
Gautier extracted by similar methods a third pigment y-ampelo-
chroic acid ; to this he gave the formula C 17 H 18 10 . It is described as

62 THE ISOLATION AND [CH.

a brown powder, very soluble in water, and having the properties of
a tannin. Gautier, in fact, considers all three acids to be coloured
tannins.

Glan (176), 1892. The flower- pigment of Althaea rosea.

The petals, after treatment with petrol ether, were extracted with
alcohol ; the alcohol was distilled oft', and the residue taken up in water
and precipitated with lead acetate. The lead salt was then decomposed
with sulphuretted hydrogen, and after concentration, the residue was
extracted with alcohol and the solution poured into ether, from which
the pigment separated out in red flakes. In properties it is a dark red
powder, insoluble in ether, chloroform, etc., and soluble in water and
alcohol. It gives a green colour with ammonia and alkaline carbonates,
a blue pigment with normal, and a green precipitate with basic lead
acetate. When the pigment is heated with dilute sulphuric acid,
sugar is split off, and the sugar-free product is insoluble in water.

The composition of the pigment before heating with sulphuric acid is :

C H

48-43% 6-18% 45-39%

and after heating with acid:

C H

54-46 % 5-81 % 39-37 %

When the solution of the pigment in sulphuric acid is neutralised
with potash, a neutral potassium compound of the pigment separates
out in blue flakes. The potassium salt has the following composition:

C H O K

48-32% 5-62% 39-73% 6-32%

Heise (178), 1894. The pigment from fruits of the Bilberry.

The pigment was first precipitated with lead acetate, and the lead
salt was then decomposed with hydrochloric acid ether. After drying,
the residue was taken up with methyl alcohol and precipitated with
ether. This product was found to consist of two substances, one of
which occurs in excess, and is separated from the other by its solubility
in acidified water. There are two methods of separating the two
pigments (termed A and B). (1) The mixture is warmed with water
acidified with hydrochloric acid; B goes into solution, A does not.
The pigment B can be purified again by precipitating with lead acetate,
decomposing with hydrochloric acid ether, taking up with methyl

v] CONSTITUTION OF ANTHOCYANINS 63

alcohol and precipitating with ether. (2) The lead precipitates are
heated with acetic acid, in which the lead salt- of B is soluble, whereas
that of A is insoluble. Heise gave to B the formula C 20 H 24 12 . By
heating B with dilute hydrochloric or sulphuric acid, A was formed
together with sugar. Hence Heise concluded that B is a glucoside.
The decomposition of the glucoside is represented as follows :

C 20 H 24 12 + H 2 = C 14 H 14 7 + C 6 H 12 6

The pigment B is described as a reddish-violet powder, soluble in
water, alcohol and acetic acid, insoluble in ether, benzene, chloroform
and carbon bisulphide. It reduces Fehling's solution : gives, a bluish-
green precipitate with lead acetate, and an intense red colour with acids.

The pigment A is a dark brown powder : it is soluble in 60 %
alcohol to a red-brown solution, or in acid alcohol to a red solution ;
it is insoluble in cold water (either neutral or acid), in absolute methyl
or ethyl alcohols, chloroform, ether and carbon bisulphide. The re-
actions of A varied, but were on the whole as follows : dirty bluish-green
colour with ammonia, dirty green precipitate with lead acetate and a
black precipitate with ferric acetate. On fusion with caustic potash,
protocatechuic acid was identified as a product of decomposition.

Griffiths (191), 1903. The flower-pigment of Pelargonium.

The pigment was extracted with alcohol and the solution gave,
on evaporation to dryness, a crystalline substance. The analyses of
the results were:

C H

From anthocyanin 62-85 % 3-53 % 33-62 %

Calculated for C 15 H 10 6 ... 62-93% 3-49% 33-58%

An acetyl derivative was obtained by heating the pigment with
acetic anhydride and sodium acetate; it crystallised from methyl
alcohol in red needles melting at 125 C. If potassium acetate is added
to a hot alcoholic solution of the pigment, orange prisms are obtained.
An analysis of the potassium salt gave the following result:

K 21-50%

Calculated from C 15 H 8 6 K 2 :

K 21-54%

Grafe (197, 209), 1906, 1909. The flower- picjment of Althaea rosea.

Preliminary experiments showed that the extraction of the petals
with water or dilute acid is impracticable, since so much mucilage is

64 THE ISOLATION AND [CH.

present in these solutions. Hence the pigment was extracted from the
dried petals with absolute alcohol ; the greater part of the alcohol
was then distilled off. an equal volume of water added, and the solution
precipitated with basic lead acetate. The green lead salt was suspended
in water, acidified w r ith hydrochloric acid, decomposed with sulphuretted
hydrogen, and filtered from lead sulphide. The deep red nitrate was
neutralised and left to evaporate : a dark red metallic powder remained
which was taken up in acetone and poured into a large quantity of ether,
whereupon the pigment separated out in granular plates. The product
was then treated with absolute alcohol, in which only a part was soluble,
the remainder being soluble in water.

To the alcohol solution enough ether was then added to precipitate
the pigment as a brown precipitate, which was filtered off and taken
up again in absolute alcohol. After evaporation, the alcohol-soluble
pigment was deposited in tiny crystalline plates of a deep red colour.

The water solution was evaporated in vacua and the residue taken up
again in a little acidified water, from which it was deposited as a granular
amorphous mass (the water-soluble pigment) which would not crystallise,
but gave concordant combustion results.

The results of analysis of the two pigments were as follows :

C H O

The portion soluble in alcohol 60-10 % 5-62 % 34-28 %

Calculated for C 14 H 16 6 ... 60-00 % 5-72 % 34-28 %

The portion soluble in water 50-11 % 6-48 % 43-41 %

Calculated for C 20 H 30 13 ... 50-21 % 6-27 % 43-52 %

Determinations of the molecular weights were made by the lowering
of the freezing point of phenol. The results were :

Molecular weight of the pigment soluble in alcohol ... 274

Calculated for C 14 H 16 6 280

Molecular weight of the pigment soluble in water ... 460

Calculated for C 20 H 30 13 ... ... ... ... ... 478

Grafe considers that the two pigments are related in the following
way:

^20^30^13 + H 2 = C 14 H 18 8 + C 6 H 12 6
2C 14 H 18 8 2H 2 2 =-- 2C 14 H 16 6

since the portion soluble in water reduced Fehling's solution, whereas
the portion soluble in alcohol did not; also dextrose could be split off
from the larger molecule by heating the substance in acid solution.

v] CONSTITUTION OF ANTHOCYANINS 65

Further investigations were first made upon the water-soluble
portion. The potassium salt was prepared by acidifying the water
solution of the pigment with sulphuric acid, and bringing to the neutrali-
sation point with dilute potash. On standing, the potassium compound
of the pigment was obtained as a blue-green precipitate. It was found
to contain 13-28 % of potassium which, according to Grafe, agrees
fairly well with 14-08 % of potassium calculated on the hypothesis that
the pigment is a dibasic acid of the formula C 20 H 30 13 .

The pigment, as in the case of other anthocyanins, was found to
be decolorised on the addition of either a solution of sulphur dioxide
in water, or a solution of sodium bisulphite. Grafe does not consider
this to be a reduction process, since the red colour does not return on
standing in air, but on addition of a stronger acid. He considers it
rather to be due to the formation of an additive compound with the
aldehyde groups present in the anthocyanin molecule. In the formation
of these compounds, certain double linkings are destroyed, thereby
depriving the substance of colour.

The sodium bisulphite compound of the water-soluble pigment was
prepared by adding the salt to an alcoholic solution of the pigment.
The alcohol was then driven off, and the residue, after purification
from sodium bisulphite, was distilled in vacuo. At about 150-160 C.
there came over a colourless oily substance distilling with difficulty,
which showed a violet-red coloration with a trace of acid. The number
of aldehyde groups in the anthocyanin was determined by estimating
the sodium bisulphite as sulphate, after oxidation with bromine. The
results showed that one molecule of sodium bisulphite combines with
one molecule of anthocyanin indicating that one aldehyde group is
present in the molecule of the water-soluble pigment.

In order to find out whether the sodium bisulphite causes any change
in the anthocyanin molecule, some of the bisulphite compound was
decomposed with acid, the solution neutralised, evaporated to dryness,
and taken up with acidified alcohol; on evaporation, a red granular
mass was deposited which gave the same combustion results as the
original substance.

The acetyl derivative of the portion soluble in alcohol was made
by treating the pigment with acetic anhydride and anhydrous sodium
acetate, and pouring into water in which the derivative is insoluble.
The acetyl compound came down in red crystals from methyl alcohol.
The results of hydrolysing, and estimating the acetic acid, indicated the
presence of two hydroxyl groups in the pigment molecule.

w. P. 5

66 THE ISOLATION AND [CH.

In an alkali melt of the alcohol-soluble pigment, pyrocatechin was
detected as one of the products of decomposition.

Grafe (222), 1911. The flower- pigment of Pelargonium.

Molisch (104) had already shown that when Pelargonium petals
are mounted in glacial acetic acid on a slide, and covered with a cover-
slip, very fine crystals are formed both in the cells and in the solution
on the slide. After several preliminary trials had been made, the
following was considered by Grafe to be the best method for dealing
with material on a large scale. The juice is squeezed out of the petals
by means of a press ; the dry residue is treated with glacial acetic acid
for several days, and then filtered off. Both juice and filtrate are shaken
up with ether, and the yellowish ethereal layer separated. Glacial
acetic acid is added to the juice, and it is mixed with the filtrate, and
the mixture again filtered. The filtrate is then dialysed, and this
process separates the anthocyanin into two components ; the dialysate
is deep yellow-red in colour, and deposits groups of crystals on evapora-
tion, while the liquid within the membrane is dark red in colour, and
does not crystallise. The separation, however, can be made in another
way. By adding ether to the filtered glacial acetic extract, a brown
flaky precipitate is deposited. The filtrate is yellowish-red and gives
crystalline anthocyanin; the precipitate is readily soluble in water,
to which a little alcohol has been added, to a brown-violet fluid which
will not crystallise. If, also, to the glacial acetic extract, lead acetate
solution is added, a dense violet precipitate is formed, and the deep
red filtrate gives no further precipitate with lead acetate, neither will
it crystallise. The lead precipitate can be decomposed either with
sulphuric acid, or sulphuretted hydrogen, and the solution of pigment
therefrom crystallises in characteristic rosettes of needles.

If the glacial acetic extract is dialysed, and the acid removed from
the dialysate by diminished pressure, a white crystalline precipitate
is formed, while the colour of the fluid becomes lighter. A microscopic
examination of the fluid shows a number of colourless crystalline plates,
together with a few rosettes of anthocyanin needles.

The crystalline anthocyanin is also unstable and liable to give rise
to the amorphous form.

The crystalline product was first investigated. It is soluble with
difficulty in absolute alcohol, but readily soluble in acetone. It
crystallises from acetic acid in good rosettes of crystals which are
very hygroscopic. If warmed, they give rise to the colourless crystalline

v] CONSTITUTION OF ANTHOCYANINS 67

product. The pigment itself decomposes and melts at 270 C. It
gives a deep red colour with acids, even with a trace; with alkalies
and ammonia, a greenish-red fluorescent solution : with ferric chloride,
a blue-violet coloration. The red colour of the pigment is changed to
yellow on standing, or by boiling with hydrogen peroxide.
An analysis of the substance, dried in air, gave:

C H O

From anthocyanin 46-34% 5-87% 47-79%

Calculated for C 18 H 26 13

+ 2CH 3 COOH 46-31 % 5-96 % 47-73 %

After drying in vacuo over caustic potash:

C H

From anthocyanin 48-69% 5-10% 46-21%

Calculated for C 18 H 26 13 ... 48-00% 5-78% 46-22%

The molecular weight, found by the lowering of the freezing point
of phenol, was 437, and calculated for C 18 H 26 13 it would be 450.

The number of hydroxyl groups was determined by means of the
acetyl derivative which was prepared by boiling the pigment with acetic
anhydride. The acetyl derivative crystallises in brown crystalline
plates from ethyl acetate. To ascertain the number of hydroxyl
groups, the derivative was hydrolysed with barium hydroxide, the acetic
acid distilled off, and estimated by means of barium hydroxide. For
the molecular weight assumed, the result corresponded with the presence
of two hydroxyl groups. The number of carboxyl groups was found
by making the potassium salt of anthocyanin by exact neutralisation,
and estimating the potassium. For the molecular weight assumed,
the result corresponded to three carboxyl groups.

A sodium bisulphite compound was obtained by shaking an alcoholic
solution of the pigment with a concentrated solution of sodium bisulphite.
The product formed a pale yellow, extremely unstable, crystalline
mass, which was soluble in water, and insoluble in ether, chloroform,
carbon bisulphide and aniyl alcohol. The product is regarded by Grafe
as an additive compound with an aldehyde group in the pigment, and
the anthocyanin colour returns on addition of a trace of a stronger
acid. The aldehyde groups were determined by allowing a known
amount of sodium bisulphite solution to react with the pigment, and
then estimating the excess of bisulphite used. The result was also
confirmed by oxidising the anthocyanin bisulphite, and estimating the

52

68 THE ISOLATION AND [CH.

sulphate formed as barium sulphate. Both methods gave numbers
indicating the presence of two aldehyde groups. In the opinion of
Grafe, it is the aldehyde groups which determine the colour in the
anthocyanin. An alkali melt was made of the pigment, and pyro-
catechin was identified as one of the products of decomposition. The
formula for anthocyanin is then represented by Grafe as

C 13 H 19 3 (OH) 2 (COOH) 3 (COH) 2 .

An analysis was also made of the colourless crystalline products
which are formed when acetic acid is driven off from the acetic acid
solution of anthocyanin. The colourless crystals are soluble in water,
alcohol and ether. The substance was identified on analysis, etc.,
as protocatechuic acid. Some doubt is expressed by Grafe as to whether
this substance is a decomposition product, or whether it is present as
impurity.

The amorphous anthocyanin was next investigated. This product
forms the principal part of the total pigment extracted. After drying,
it is soluble with difficulty in water, but readily soluble in dilute alcohol.
With acids it gives a less bright colour than the crystalline form : with
alkalies, a greenish-brown colour, while sodium bisulphite removes the
colour.

The analyses gave:

C H O

From anthocyanin ...... 43-78% 6-22% 50-0%

Calculated from C^H^Oso ... 44-17% 6-75% 49-08%

The molecular weight was determined in phenol, and found to be
663 as compared with 652 calculated for C^H^ao.

This latter product was found to be a glucoside, and dextrose was
identified as one of the products of hydrolysis with acid. Grafe suggests
that the relationship between the two pigments may be represented
thus:

(amorphous)

H 2 == C G H 12 6

- 4H 2 + 2 = C 18 H 26 13

(crystalline)

He is also of the opinion that the amorphous form arises by changes
induced in the crystalline, rather than vice versa, since the crystalline
is never produced from the amorphous, but the latter may arise from
the former.

v] CONSTITUTION OF ANTHOCYANINS 69

Wheldale (244, 254), 1913, 1914. The flower-pigment of

Antirrhinum ma] us.

The anthocyanin pigment, as in most flowers, is only present in
the epidermis, while the inner tissues contain a flavone, from which,
we have reason to believe, the anthocyanin is derived. Pigment was
prepared separately from the following varieties (see p. 160) : magenta
(various shades together), ivory tinged with magenta, crimson, rose
dore (various shades) and bronze (various shades). In the magenta
and rose dore series, apigenin is present in addition to anthocyanin ;
in the crimson and bronze, both apigenin and a second pigment, luteolin
(see p. 114). All the flowers have, in addition, a patch of deep yellow
pigment on the palate. The latter pigment can be eliminated, if desired,
by tearing away the lower half of the flower, and using the upper half
only for extraction ; this device was adopted in the preparation of some
samples of pigment. Thus, in any method of extraction, we have to
deal, not only with an anthocyanin pigment, but also with one or more
accompanying flavones. Two anthocyanins were found to be respon-
sible for the colour varieties, viz. a true red anthocyanin in the rose
dore and bronze series, and a magenta (blue-red) anthocyanin in the
magenta and crimson series. The following method was employed
for obtaining the pigment in quantity.

The flowers are boiled with water in saucepans, and the water
extract filtered through large filters into lixiviating jars. The antho-
cyanins and flavones are then precipitated as lead salts by adding solid
lead acetate to the hot solution until no more precipitate is formed.
(The colour of the precipitates varies according to the flowers used;
it is blue-green for full-coloured magenta, yellow-green for tinged ivory,
dirty red for rose dore, and so forth. The colour of the lead salt of the
anthocyanin is obviously modified by the amount and colour of the
lead salts of the accompanying flavones.) The lead precipitate is
filtered off, a vacuum pump being used for filtration ; the solid cake of
lead salt is then decomposed with 5-10 % sulphuric acid. The lead
sulphate is filtered off, and a bright red solution of the pigments is
obtained. This solution contains all the pigments in the flower used,
both anthocyanin and flavones, in the form of glucosides in dilute acid
solution. The solution is now boiled in a large Jena flask, fitted with a
reflux condenser, for several hours. On cooling, the anthocyanin and
flavones, now less soluble and no longer in the form of glucosides,
separate out as a dark purplish- or brownish-red powder, according

70 THE ISOLATION AND [CH.

to the flower-colour used. The crude pigment is filtered off through
as small a funnel as possible by means of a vacuum pump, washed,
and dried over calcium chloride. The well-dried pigment is then
finely ground, and placed in a Soxhlet thimble. The thimble is sus-
pended just above the surface of ether contained in a wide-necked
Erlenmeyer flask fitted with a condenser, and the ether kept boiling
upon an electric heater. This process is continued until the ether
ceases to extract any yellow colour. In this way the anthocyanins are
obtained practically free from flavones, since the latter are soluble
in ether. The anthocyanin residue in the thimble is then taken up in
absolute alcohol, and filtered, and is, in this way, freed from a quantity
of brown substance, which is insoluble in alcohol, and which is probably
formed during the hydrolysis of the glucoside with sulphuric acid.
The absolute alcohol solution, evaporated to its minimum bulk, is
then poured into a large volume of ether, and the anthocyanin is
precipitated, but any flavone present as impurity is retained in solution.
The dried precipitate of anthocyanin is again extracted with ether
to remove traces of flavone. The method of precipitation gives better
results, as regards the purity of anthocyanin, than crystallisation, for,
on crystallising a mixture of anthocyanin and flavone, both substances
crystallise out together, and one is unaware of the presence of flavone
in the product obtained.

The two forms of anthocyanin, red and magenta, were extracted
and purified in this way from the flowers of different varieties of
Antirrhinum mentioned above.

Pure red anthocyanin is an indian-red powder. It is readily
soluble in absolute alcohol, almost insoluble in water, and slightly
soluble in dilute acids and ethyl acetate; insoluble in ether, chloro-
form and benzene. In concentrated sulphuric acid it forms a reddish
solution with a slight green fluorescence. It is soluble in alkalies
to a greenish-yellow solution. With ferric chloride it gives a
brownish-green coloration. With lead acetate, a brownish-yellow
precipitate.

Pure magenta anthocyanin is a magenta-red powder with similar
properties and solubilities to the red. In concentrated sulphuric acid
it gives a red solution with a slight greenish fluorescence. It is soluble
to a green solution in alkalies. With ferric chloride solution it gives
a brownish-green coloration. With lead acetate it forms a greenish-
black precipitate of a lead salt. In alkaline solutions it is strongly
fluorescent, green by transmitted, red by reflected light.

v] CONSTITUTION OF ANTHOCYANINS 71

The results of combustion of the pure anthocyanin were:

C H O

Eed anthocyanin : from rose dore 51-93 % 5-02 % 43-05 %

bronze ... 51-37% 5-05% 43-58%

^9.12 / 4-Q7 / 42-Q1 /

,, ,, ,, ,, ... Ol il / Q Vi /Q tA Ul /Q

Magenta magenta 50-26 % 4-89 % 44-85 %

ivory tinged

with magenta 50-68% 5-54% 43-78%

from crimson ... 50-56% 4-90% 44-54%

Attempts were made to determine the molecular weight by depression
of freezing point, using phenol as a solvent, but the results, though
consistent for a series of experiments, were obviously far too low. '
Acetic acid, and various other solvents, did not dissolve enough of
the pigment to give measurable depressions. Attempts to determine
the molecular weight by elevation of the boiling point in absolute
alcohol gave mean values of 572 for the red, and 717 for the magenta.
The elevation of the boiling point was so slight that the error in the
value obtained may be very considerable.

The combustion results give, as simplest formulae, C 15 H 18 10 for the
magenta, and C 8 H 9 5 for the red. The boiling point determination
of the molecular weight would appear to indicate that the molecule
is 2 (C 15 H 18 10 ), i.e. C 30 H 36 0. 20 , which has a molecular weight of 716
for the magenta, and 3 (C 8 H 9 5 ), i.e. C 24 H 27 15 , which has a molecular
weight of 555 for the red.

An attempt was made to estimate the number of hydroxyl groups
present in the anthocyanin molecule by means of ZerewitinofFs modifi-
cation of Hibbert & Sudborough's method. This consists in dissolving
the substance in thoroughly dried pyridine, treating it in a suitable
apparatus with a considerable excess of methyl magnesium iodide, and
collecting the gas evolved. Each hydroxyl group causes the evolution
of a molecule of methane. It should be noticed that 'hydroxyl groups,'
as determined by this method, include those forming part of the carboxyl
groups, and also such ketone groups as can give rise to hydroxyl by
tautomeric change. The values obtained indicate that the red antho-
cyanin, taking the formula as C^H^Ojg, contains twelve hydroxyl groups
as defined above, while the magenta, taking the formula as C 30 H 36 20 ,
contains fifteen hydroxyl groups.

72 THE ISOLATION AND [CH.

Willstatter (245), 1913. The flower-pigment of Centaurea Cyanus.

In Centaurea flowers, according to Willstatter, there are three
modifications of one anthocyanin pigment: a purple pigment (cyanin),
which is itself a free acid ; a blue pigment, which is the potassium salt
of the purple, and which constitutes the greater part of the colouring
matter of the flower ; a red pigment, which is the oxonium salt of the
purple with some organic acid (other oxonium salts can be obtained
artificially with inorganic acids, such as hydrochloric acid). Cyanin,
moreover, isomerises to a colourless form which is an acid too, and
forms colourless alkali salts 1 ; there is also a colourless isomer of the
blue pigment. When the anthocyanin of the flower is extracted with
water, the deep blue solution rapidly loses its colour; this is due to
the above isomerisation, and the colourless solution, on addition of
acid, will become as red as a solution of the original blue pigment would
on acidification. The red modification also loses colour in absence of
acid, i.e. if sufficiently diluted with water or alcohol. On concentration,
a colourless solution will return to its original colour, blue, red or violet,
as the case may be.

For preparation of the blue pigment on a large scale, dried flowers
ground to a fine powder were employed. The blue pigment can be
extracted rapidly with water or very dilute alcohol, but change to the
colourless isomer tends to take place. This change can be prevented,
however, by addition of much sodium nitrate or chloride to the pigment
solution. The cyanin salt can then be precipitated from the water
solution with alcohol in which it is insoluble. It is further purified
by fractional precipitation with alcohol from water solution.

Cyanin is a glucoside, but on hydrolysis it gives the free pigment,
cyanidin, and sugar. The pigment, cyanidin, like the glucoside, cyanin,
forms crystalline oxonium salts with hydrochloric acid.

The method of isolation of the blue pigment, in greater detail, is as
follows. The powder of ground flowers is mixed with sand, extracted
with water or 20 % alcohol, and filtered, and finely powdered, sodium
nitrate is added. The deep blue solution is then mixed with 96 %
alcohol (2-5 vols. alcohol : 1 vol. extract), and the pigment is preci-
pitated out in blue flakes, which are separated by a centrifuge. The
pigment is further purified by taking up in water, and precipitating

1 " Das Cyanin isomcrisiert sich zu einer farblosen Modifikation, wclche glcichfalls
sauer 1st und farblose Alkalisalze bildet." The alkali salts of the isomer are, however,
definitely stated to be yellow later in the paper.

v] CONSTITUTION OF ANTHOCYANINS 73

with alcohol. For the exact sequence of operations, the original paper
should be consulted. The nitrate from the precipitates contained
much of the colourless isomer. The pigment was also extracted by
another method in which the use of sodium chloride or nitrate was
eliminated. This is necessary if one wishes to obtain the naturally-
occurring cyanin salt, for if the method first described is employed
the sodium replaces the potassium originally present. In the second
method, the flower powder is mixed with sand, and extracted with dilute
alcohol (80 vols. water : 20 vols. of 96 % alcohol). It is then filtered
and precipitated with alcohol (3 vols. of filtrate : 5 vols. 96 % alcohol),
and the precipitate separated by a centrifuge. All the operations should
be carried out as rapidly as possible in order to avoid isonieric change.
The crude pigment was again purified by reprecipitation. It was
found that the product obtained by the first method contained sodium
nitrate as impurity. This was removed by extracting with 75 %
alcohol, in which the impurity is soluble, though not the pigment. The
product was, however, still further purified by precipitation, and in
the end contained both sodium and potassium, the former as a result of .
the use of sodium nitrate. The product obtained by the second method
(i.e. when the use of sodium salts for protection against isomerisation
was avoided), after purification, contained no sodium.

By dialysis of the cyanin salt in a 20 % sodium chloride solution,
dark blue crystals were obtained. These Willstatter regards as an
addition product of the cyanin salt with sodium chloride.

With regard to properties, the blue cyanin salt is insoluble in alcohol,
but soluble in water ; a concentrated water solution shows loss of colour
only after a day or two, but a dilute solution decolorises in an hour or
so. The isomerisation, as already mentioned, is best prevented by
sodium chloride or nitrate, but potassium nitrate or chloride has
little or no effect. The pure blue product shows practically no change
in colour on addition of a little sodium hydroxide solution, but a solution
of pigment which has stood becomes blue-green, or green-blue, on
account of the presence of the colourless modification.

The next operation was the preparation of a crystalline salt of
cyanin with hydrochloric acid. The cyanin alkali salt, obtained by
the methods described, is dissolved in 20 % hydrochloric acid. From
this solution, some accompanying carbohydrates (pentosans), which are
present as impurity, are precipitated by addition of absolute alcohol.
After filtration, the pigment chloride is precipitated by ether. The
crude product is then taken up in absolute alcohol, which frees it from

74 THE ISOLATION AND [CH.

colloidal substance and other impurities. The purified alcoholic solu-
tion is then acidified with strong hydrochloric acid, and concentrated in
a vacuum desiccator. Cyanin chloride separates out as an amorphous
product which is soluble in both alcohol and water to a bright red
solution ; the solution becomes rapidly paler, but in absolute alcohol,
the loss of colour is much less rapid than in the presence of water. To
obtain the crystalline form, the amorphous product is dissolved in
absolute alcohol, filtered, and mixed with a third of its volume of a
7 % solution of hydrochloric acid in water, and set to crystallise. The
crystals are deep blue rhomboidal plates with a golden lustre: in a
powdered form, the colour is brown-red. The chloride crystallises out
with three molecules of water of crystallisation, and the formula obtained
by elementary analysis is CagHggO^Cl . 3H 2 0. The water-free product
on analysis gave C 28 H 33 17 C1. The crystalline cyanin chloride is almost
insoluble in water, soluble with difficulty in cold alcohol, acetone and
chloroform; insoluble in benzene; slightly soluble in dilute hydro-
chloric and sulphuric acids. It is stable in acid solution : in water
solution it decolorises, with the formation of the isomer, especially
if dilute, but the colour returns on acidification. The colourless isomer
is also formed by warming with absolute alcohol ; such a solution then
gives with lead acetate a green precipitate, on account of the mixture
of the blue salt of the pigment with the yellow alkali salt of its isomer.
Pure cyanin chloride gives with calcium carbonate a violet solution;
with sodium hydroxide, a blue solution ; with lead acetate, a blue
precipitate, which gives a green lead salt on standing ; and with sodium
carbonate, a violet colour which eventually becomes yellow. The
chloride is reduced with zinc and hydrochloric acid, and also decolorised
with sodium bisulphite, the colour in the latter case returning on
addition of an acid. As a glucoside it reduces Fehling's solution.

The glucoside cyanin is rapidly hydrolysed with 20 % hydrochloric
acid, cyanidin being formed which separates out as the chloride from
the hot solution. The combined sugar was identified as glucose. The
crystals, in the form of needles, of cyanidin chloride are brown-red
under the microscope, give a violet streak, and a brown-red powder.
They have no water of crystallisation, and an elementary analysis gave
the formula C 16 H 13 7 C1. The chloride is soluble in alcohol with a fine
violet-red colour ; it is also soluble with difficulty in dilute hydrochloric
acid. It crystallises from a mixture of alcohol and dilute hydrochloric
acid. It is also soluble in amyl alcohol, and when such a solution is
shaken with caustic soda solution, the pigment goes into the watery

CONSTITUTION OF ANTHOCYANINS

75

layer with a blue colour. With sodium carbonate it gives a violet or
blue colour (green when the colourless isomer is also present) ; with
lead acetate, it gives a blue precipitate.

The violet modification of anthocyanin was obtained apparently
by precipitating a solution of cyanin chloride with lead acetate, and
decomposing the lead salt with excess of sulphuretted hydrogen. On
filtration and evaporation, the violet form of the pigment is obtained.

The colourless modification of the pigment is soluble in ether;
on evaporation of the ether it remains in the form of colourless crystals,
which give an intense red colour on heating with hydrochloric acid.
With alkalies it gives a yellow colour, but is apparently not a flavone
(see p. 78).

As regards the constitution of anthocyanin, Willstatter, in this
first paper, is of the opinion that the pigment of Centaurea contains a
nucleus of the following type, colour being due to the presence of the
quinonoid and tetravalent oxygen :

=

the above would represent the violet pigment as an inner oxonium salt.
Under certain conditions, as we have seen, a change readily takes place
by tautomerism to a colourless isomer:

HO-

HO-

which is really a flavone with an extra hydroxyl in position 2.

The blue pigment in the plant is the potassium salt of the
violet pigment, the position of the potassium being uncertain.
If the cell-sap is alkaline, the blue salt predominates, or is alone
present.

76 THE ISOLATION AND [CH.

If there is excess of acid in the cell-sap, a red oxonium salt with an
organic acid is formed. The crystalline salts with hydrochloric acid
are regarded as artificial compounds of this kind :

oci
^\ >

HO-f

o =

N^

OH

In the presence of excess of water, both the red and the blue forms
will pass to the colourless isomer, but in the case of the blue the tauto-
meric change can be prevented by adding a neutral salt which forms an
additive compound :

ONO 3

S\ y.
NaO,

The chief argument against the above hypothesis is that flavones
of the above constitution are at present unknown. The analyses so
far made only point to the existence of a definite hydrochloric acid
compound, and indicate nothing as regards the constitution, the above
suggestions being purely hypothetical. Variations in this hypothesis,
based upon later work, will be considered in the following paragraphs.

Willstatter, in conjunction with Bolton, Nolan, Mallison, Martin, Mieg
and Zollinger (256, 257), 1914. The flower -pigments of Centaurea,
Delphinium, Malva, Pelargonium and Kosa gallica; the fruit-
pigments of Vaccinium Myrtillus, Cranberry and Vitis vinifera.

Flower-pigment of Centaurea.

The statement made previously that the cyanin of Centaurea is
decomposed on hydrolysis into one molecule of cyanidin and two
molecules of glucose is confirmed. But the formula C 16 H 13 7 C1, originally
given for cyanidin chloride, is found to be incorrect owing to the fact
that the product was insufficiently dried. It is found that reliable
numbers can only be obtained after long drying in a high vacuum at

v] CONSTITUTION OF ANTHOCYANINS 77

105 C. The formula is then found to be C 15 H n 6 Cl from the following
analyses :

C H Cl

From cyanidin chloride ... (i) 5540% 3-62% 10-75%

(ii) 55-77% 3-60%
Calculated for C 15 H U 6 C1 55-81% 3-41% 11-01%

Flower-pigment of Rosa gallica.
This pigment is found to be a diglucoside of cyanidin.

Pigment from fruit of Cranberry.

This pigment idaem is found to be a galactoside of cyanidin,
formed from one molecule of cyanidin and one molecule of galactose.

Calculated for C 31 H 21 O n Cl Found

Galactose 37-2 ... 33-4

Cyanidin chloride ... ... ... 66-6 ... 67-5

Pigment (Oenin) from grapes.

The method employed for preparation was as follows. The skins
of dark blue grapes are extracted in the cold with glacial acetic, and
the dark red nitrate precipitated with ether. A sticky precipitate is
obtained which, after washing with ether, is put into an excess of water
picric acid solution, and warmed for a short time. On cooling, the
picrate crystallises out from the solution in long prisms of a fine red
colour. By changing the solution to methyl-alcohol-hydrochloric acid,
it yields the solution of the pigment chloride, which is precipitated with
ether-petrol-ether and crystallised from water-alcohol-hydrochloric acid,
in the form of hard beetle-green prisms. On hydrolysis, oenin decom-
poses into oenidin C 17 H 14 7 and one molecule of glucose.

Pigment (Myrtilliri) from the Bilberry (Vaccinium Myrtillus).

The skins of the berries are used after being dried and ground.
The pigment is extracted rapidly by warming with ethyl alcohol, which
contains a small percentage of hydrochloric acid, and the solution is
precipitated with ether. The precipitate, mixed with a large quantity
of a colourless product, is separated, by taking up in water, from many
of its impurities. By addition, with cooling, of a double weight of
concentrated hydrochloric acid, the chloride is precipitated almost
pure, and quite pure by repetition of the operation. For crystallisation,
a third of its volume of 9 % hydrochloric acid is added to the solution

78

THE ISOLATION AND

[CH.

of the pigment in wood spirit ; on slow evaporation the pigment separates
out in beautiful flat prisms. Myrtillin gives, on hydrolysis, one molecule
of glucose and one molecule of myrtillidin C 16 H 12 7 . Myrtillin was
also isolated from flowers of the Hollyhock (Althaea rosea).

The following anthocyanins were also isolated, but no details are
given of the methods.

The flower-pigment of Delphinium delphinin which gives, on
hydrolysis, two molecules of glucose, two molecules of p-oxybenzoic
acid and one molecule of delphinidin C 15 H 10 7 .

The flower-pigment of Pelargonium pelargonin which gives, on
hydrolysis, two molecules of glucose, and one molecule of pelargonidin

""^JsHjijOg

The flower-pigment of Malva -malvin which gives on hydrolysis
malvidin C 17 H 14 7 .

The chlorides of pelargonidin, oenidin and delphinidin are described
as crystalline salts, insoluble in water, but soluble in alcohol. In the
glucosidal form, they give, with the exception of delphinidin, colourless
isomers in water solution. The anthocyanidins isomerise less readily.
The isomerisation is said to take place according to the following
equation :

C 15 H 10 7 HC1 + H 2 = HC1 + C 15 H 12 8

In alkali melts of the pigments, the following decompositions were
found to take place:

Cyanidin gave phloroglucin and protocatechuic acid.

Pelargonidin gave phloroglucin and ^-oxybenzoic acid.

Delphinidin gave phloroglucin and gallic acid (not isolated in pure
state).

As a result of these further researches Willstatter then suggests
that for the formula of cyanidin chloride the choice lies between :

OCl

OC1

I.

HO

OH

OH

vj

CONSTITUTION OF ANTHOCYANINS

79

Formula II was discarded because, if such a formula were correct,
it should be possible to obtain a substituted benzophenone, for example,
maclurin 1 , as a product from cyanidin, and this was not found to be
the case.

Willstatter then suggests that the anthocyanidins and flavones are
related in the following way :

Flavones Anthocyanidins

Luteolin, karnpherol and fisetin C 15 H 10 6 -* pelargonidin C 15 H 10 5
Quercetin C 15 H 10 7 -* cyanidin C 15 H 10 6

Myricetin C 15 H 10 8 -* delphinidin C 15 H 10 7

that is, that each anthocyanin is derived from a flavone by reduction.
The formulae eventually suggested for the chlorides of cyanidin,
pelargonidin and delphinidin are as follows:

OC1

OH

OC1

OC1

HO

\

HO C
H

Cyanidin chloride

HQ

Pelargonidin chloride

Delphinidin chloride

whereas myr-tillidin, oenidin and malvidin are represented as methyl
derivatives of delphinidin :

OCl

OCl

OH

OCl

HO/V

CHO

OH

HO
Myrtillidin chloride

HO

Oenidin chloride

HO
Malvidin chloride

Willstatter also brings forward, as additional evidence in favour
of these constitutional formulae, the preparation, artificially, of cyanidin
from quercetin (see p. 124).

1 Maclurin is a pentaoxybenzophenone occurring in Morus tinctoria;
is represented as :

HO OH

>OH

constitution

HO CO

80 ISOLATION AND CONSTITUTION OF ANTHOCYANINS [OH. v

A short summary of the formulae 1 given by various authors for antho-
cyanins may be stated as follows 2 :

Flower-pigment of Althaea rosea

(glucoside and oxida-
tion product)

Antirrhinum (magenta

(red) ...
Centaur 'ea Cyanus ...

99 99

Delphinium

Malva

Pelargonium

(glucoside and oxida-
tion product)

,, ,, Rosa gallica

99 99 99

Fruit- pigment of V actinium Myrlillus

99 99 99 99

Vitis-idaea

,, Vitis vinifera (in wine)

Red leaf -pigment of Coleus

,, Vitis vinifera

z(C 4 H 5 O 2 )
C 14 H 16 6

(Glan).
(Grafe).

C 16 H 12 O 7 (Willstattcr).
z(C 15 H 18 O 10 ) (Wheldale).
z(C 8 H 9 5 )

o;(C 2 H 3 O 2 ) (Morot).

C 15 H 10 O 6 (Willstatter).
. Ci 5 H 10 O 7

C 17 H 14 O 7 ,,

C 15 H 10 6 (Griffiths).

C 18 H 26 13 (Grafe).

C 24 H 44 O 20

C 15 H 10 O 5 (Willstatter).

C 7 H U O ]0 (Senier).

C 15 H 10 O 6 (Willstatter).

C 14 H M 7 (Heise).

C 16 H 12 O 7 (Willstatter).

C 20 H 10 O 10 (Glenard).

C 20 H 20 O 10 (Gautier).

^21^20^10 99

C 17 H 14 O 7 (Willstatter).

C 10 H 10 5 (Church).

rC 19 H 1(i 10 (Gautier).
A C 26 H 24 O 15 ,,

! C 17 H 18 O 10 ,,

1 The formulae are so arranged rather as a matter of interest than for comparison,
since the earlier workers laboured under disadvantages which would obviously detract
from the value of their numbers.

2 See also Appendix.

CHAPTER VI

PHYSIOLOGICAL CONDITIONS AND FACTORS INFLUENCING
THE FORMATION OF ANTHOCYANINS

Connection with photosynthesis.

An examination of the relative distribution of anthocyanin and
chlorophyll at once suggests that these pigments are more or less
complementary as regards their appearance in the plant tissues. In
leaves, the chief seat of chlorophyll, anthocyanin is found to a less extent
than in any other organ, and under normal circumstances the develop-
ment is most frequently confined to the epidermis (see p. 37), or to
a few sub-epidermal layers, often only where they overlie the midrib
or main veins. Moreover, when red pigment is present in the epidermis,
the guard cells of the stomata,. which contain chlorophyll, are generally
free from pigment. In petioles and stems also, anthocyanin is on the
whole limited to the epidermis, or to a few sub-epidermal layers. Bracts
of all kinds, on both the inflorescence and other parts of the plant,
frequently contain anthocyanin and correspondingly little chlorophyll ;
although flowers in the bud stage and unripe fruits have fairly abundant
chlorophyll, as the flowers and fruits mature the chlorophyll disappears
and anthocyanin develops.

Since chloroplasts are invariably concerned with photosynthesis,
one would naturally conclude that the latter process and the formation
of anthocyanin are to some extent mutually exclusive. The existence
of some such alternation is further emphasised by the appearance of
anthocyanin which accompanies lessened photosynthetic activity, as
in plants towards the end of their vegetative season, in autumnal
reddening, in leaves in an unhealthy condition and in evergreens during
winter. Since all metabolic activity ultimately depends on photo-
synthesis, it is not a convincing argument that a decline in general
metabolism (apart from photosynthesis) may directly bring about
the formation of pigment. There is without doubt good evidence for

w. p. 6

82 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

believing that anthocyanin is not readily produced where carbon
assimilation is most active, and that decreased photosynthesis from any
outside cause is favourable to its formation.

Hence in any consideration of the direct bearing of outside factors,
such as temperature and light, on anthocyanin formation, recognition
should be first given to possible indirect effects produced by these factors
through the medium of photosynthesis.

Connection with accumulation of synthetic products.

Even at the height of summer when the vegetative organs are in
a condition of maximum activity, quite a number of individual plants
may be found having isolated leaves or shoots which are either entirely
red or have developed patches or blotches of anthocyanin. In the
majority of cases, one will find on careful examination that there has
been some injury to the leaf, petiole or stem, as the case may be, and it
is to the distal side of the injured spot that the reddening occurs. Such
injuries may be classified as: (1) mechanical, caused by chance cutting
or breaking; (2) attacks of insects, including gall insects and cater-
pillars; (3) infection by Fungi.

(1) Let us deal first with mechanical injury. Frequently leaves
may be found in which the lamina is partially severed transversely, and
the severed portion has reddened. Or the petiole or stem is partially
broken, and the leaf or leaves above the point of injury have turned
red. In Rumex, Oenothera, Pelargonium, Plantago and many other
plants, it is easy to bring about such reddening artificially by pinching
the lamina or petiole, and in other cases by decortication. Or sometimes
isolated leaves, as for instance those of Rheum, left lying on the ground
in a damp place will eventually redden. In other genera and species, it
is difficult, or impossible probably, to induce reddening by such means.
Reference to these phenomena has often been made by various authors :
Gautier (175), Kraus (311), Berthold (64), Linsbauer (341), Kiister
(350), Daniel (337) and finally Combes (374, 385). Further Combes has
made a series of experiments on decortication of stems of many species
with a view to investigating the phenomena more fully. By these means
he has distinguished three types of results :

(a) Those in which anthocyanin appeared more or less rapidly
in the branches, petioles and above all in the leaves. Ex. Spiraea spp.,
Mahonia aquifolium, Prunus Pissardi.

(6) Those in which anthocyanin appeared more or less rapidly in

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 83

the stems and petioles but not in the leaves. Ex. Ceanothus azureus,
Catalpa bignonioides.

(c) Those in which pigment never appeared as a result of decorti-
cation. Ex. Rhodotypos kerrioides, Robinia Pseudacacia, Pinus excelsa.

The time elapsing before the appearance of pigment also varied
considerably. In addition it was noted, as would be expected, that no
pigment appeared on decortication in albino varieties, although the
coloured type might produce abundant pigment under similar treat-
ment.

(2) Injuries brought about by insects present no points of special
interest but may be regarded rather as cases of (1). Frequently the
midrib or petiole is partly eaten away, and reddening occurs on the
distal side ; or holes are eaten in the lamina, and red blotches are formed
in their vicinity. Mirande (362) has observed that excursions of leaf-
boring larvae in leaves of Galeopsis Tetrahit result in the production of
anthocyanin. Under this heading also may be included the develop-
ment of anthocyanin in or near galls. For details and examples the
works of Hieronymus (314), Kiistenmacher (326), Kiister (350) and
Guttenburg (353) may be consulted.

(3) It is frequently found that the pathological conditions called
forth by the attacks of Fungi are accompanied by abnormal develop-
ment of authocyanin. In leaves of Tussilago, for instance, infected by
Puccinia a circular band of anthocyanin often appears surrounding the
aecidium spots. Other references to this matter may be looked for in
the works of Sorauer (304), Tubeuf (329), Liidi (342) and Rostrup (345).

There is little doubt that, in the above cases of injury and decorti-
cation, the formation of pigment is directly connected with an inter-
ference with the translocation current. Injury to the living tissues of
the conducting system of the veins, midrib or petiole of the leaf, or
of corresponding tissue in the stem, leads to an accumulation of synthetic
products in the leaves. Of such products several authors Mirande
(365), Combes (207) have maintained that it is the carbohydrates and
glucosides which most influence the formation of anthocyanin, and
Combes has shown by analyses that leaves reddened by decortication
contain a higher percentage of sugars and glucosides. It seems likely
also that parasitic growths may interfere with the progress of the
translocation current through the small veins of the leaf, thereby causing
congested areas to arise in which the sugar contents are above normal.
But it is conceivable that the pathological condition resultant on fungal
attacks may be the direct cause, in some way, of pigment formation.

62

84 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

The more intimate connection between anthocyanin and sugars will
be discussed in a later chapter.

The experiments of decortication, etc., lead also to the conclusion
that the chromogen 1 of anthocyanin is synthesised in the leaves. For
in cases where leaves and shoots have reddened owing to the blocking
of the translocation current, less development of pigment has often been
noticed in flowers and fruits. Gautier (175) made various experiments
on vines in order to illustrate his view that the chromogen of the grape
pigment is synthesised in the leaves, and is oxidised after passing into
the fruit. Vine branches were deprived of their leaves, and this was
shown to prevent a development of pigment in the fruit. In another
experiment, the petioles of the leaves were ligatured with the result
that the fruits remained green and the leaves themselves reddened.
Ravaz (380), on the other hand, grafted a vine with purple grapes on
to a white-fruited variety, and found that although the white variety
produced no pigment in the leaves, the fruit of the graft was coloured
every year. Hence Ravaz concludes that the pigment is synthesised
in the fruit itself, though the latter may be nourished by the leaves.
A direct connection between leaves and flower-colour may be demon-
strated by removing a developing inflorescence from a plant, such as
Digitalis purpurea, when the leaves will generally turn red. Gertz (19)
observed the same result in a plant of Geum rivale from which the flowers
had been cut off.

There is also reason to believe that special richness in nutriment,
or synthetic products, is connected with anthocyanin formation. An
experiment is quoted by Berthold as illustrating this point. From
two- or three-year old individuals of Acer pseudoplatanus all buds were
removed but the terminal one, which, as a result, received an excessive
amount of nutriment and on development was strongly reddened.
Some such explanation, as Gertz suggests, may account for the excessive
reddening of leaves of adventitious shoots arising at the base of felled
trees (Populus, Tilia, Acer).

Bonnier (294, 307, 328) and Heckel (300) have noted that a greater
intensity of flower-colour is produced in many species when grown at
high altitudes, as compared with the colour of the flowers in lowland
regions. It is conceivable that this increase in intensity of colour

1 The chromogen, as we shall see later (p. 109), is in all probability a flavone occurring
in the form of a glucoside. There is reason to believe that, not only does the synthesis
of glucoside from flavone and sugar take place in the leaf, but also the synthesis of the
flavone itself from sugars.

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 85

is brought about by increase of synthetic products, due to greater
insolation by day accompanied by low night temperature, and to
stunted growth, rather than to any direct action of light on pigment
formation, but at present there is no conclusive evidence.

In the same way lack of synthetic products due to poor conditions
of the plant reduces pigment formation. The experiments of Sachs
(271), Askenasy (282), Vochtin-g (325) and Klebs (360), in which leaves
of plants were kept in darkness while the flowers were exposed to
light, are not so conclusive as those previously mentioned. The method
is so drastic, and may influence the whole nutrition of the plant to such
an extent, that the non-development of flower-colour cannot be regarded
as having any great significance.

A phenomenon of considerable interest in connection with nutrition
and formation of red pigment is that pointed out by Mirande (332)
as occurring in the genus Cuscuta, in which anthocyanin is widely
produced. From observations made upon the development of many
different species of Cuscuta on various hosts, Mirande concludes that
the amount of pigment varies not only in different species but also in
each species according to the host on which it grows. For instance,
the same species growing on Sanibucus nigra (poor) and Forsythia
viridissima (rich in sugar) becomes green on the former but very red on
the latter. Hence Mirande correctly deduces the fact that not only
good development of the parasite but also the formation of red colour
is correlated with good nutrition. In nature, species of Cuscuta
passing from one host to another are seen to show different amounts
of pigmentation. Chemical tests made by Mirande on extracts from
the host plants showed that the greatest production of colour was
found when the host plants were capable of producing most sugar.

The conclusions which may be drawn from all the instances quoted
above are that an unnatural accumulation of synthetic products may
cause colour to be developed in organs not normally coloured. At the
same time a good supply of nutrition will intensify colour in parts of
plants which are normally coloured. There is reason to believe that,
of the accumulated substances, the most potent in bringing about
colour production are sugars and glucosides, and this will be found to
be borne out by observations connected with the effect of other factors
considered in this chapter. Also an increase of the chromogens from
which anthocyanins are produced takes place under the conditions men-
tioned above. The manner in which glucosides, sugars, chromogens
and pigment may be connected is dealt with in the next chapter.

86 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

The relationship between anthocyanin development and sugar-
feeding is reserved for a later paragraph (see p. 93).

Effect of temperature.

Most general observations bear out the conclusion that increase of
anthocyanin is correlated with lowering of temperature. The most
obvious demonstrations are the autumnal coloration of leaves and, to
a lesser extent, the reddening of evergreen leaves in winter (Ligustrum,
Hedera, Mahonia). The question of the relationship between low
temperature and anthocyanin formation has been specially considered
by Overtoil (333). This author had observed that Hydrocharis plants
grown in cane sugar solution became very strongly reddened, and the
idea occurred to him that an excess of sugar in the cell-contents might
similarly be the cause of autumnal and winter coloration of leaves. In
view of this suggestion, it is difficult to estimate any direct effect of low
temperature on anthocyanin formation, because of the indirect effect
produced by the same conditions on (1) photosynthesis, (2) formation
of starch from sugar, (3) growth in general and probably (4) transloca-
tion. A decrease of activity of (1) leads to a decrease of sugar contents
in the cell ; but a decrease of (2) and (4) has the opposite effect. The
process of removal of synthetic products from the leaves is, according
to Sachs, greatly retarded by low temperature. Hence similar condi-
tions of clogging to those brought about by injury, which were mentioned
in the previous section, might arise and there would be a resultant
production of pigment. The synthesis of starch from sugar is also a
process which is retarded by low temperature. Thus Miiller-Thurgau 1
has shown that at temperatures below 5 C. quite a considerable portion
of the starch contents of the potato is changed to sugar, and w r ith a
rise of temperature the greater portion of starch is again regenerated.
According to Lidforss 2 , evergreen leaves in winter are also completely
starch-free but contain very considerable quantities of glucose, which
is again to a large extent changed back to starch if the leaves are
artificially warmed. Overton himself examined the sugar content of
autumnal leaves and found considerable quantities present, considerably
more, at any rate, than in the same species at midsummer.

From the above statements it will be seen that low temperature

1 Miiller-Thurgau, H., 'Ueber Zuckeranhaufung in Pflanzentheilen in Folge niederer
Tempera tur,' Landw. Jahrb., Berlin, 1882, xi, pp. 751-828.

2 Lidforss, B., 'Zur Physiologic und Biologic der wintergriinen Flora,' Bot. Centralbl.,
Cassel, 1896, Lxvin, pp. 33-44.

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 87

may greatly affect the sugar contents of the tissues, and hence may
in this way cause the reddening, apart from any more direct effect.

Overton made some observations on the effect of temperature on
reddening of Hydrocharis leaves, and found that the higher the tempera-
ture the less anthocyanin is formed and vice versa, but obviously in
this case it is impossible to eliminate the effect of temperature on the
photosynthetic activity of the leaves, and on growth and respiration
in general with the resultant employment of synthetic materials.
Klebs also (360) gives an account of the effect of temperature on the
colour of the flowers of Campanula Trachelium. From cultures at
various times of the year in green-houses, etc., kept at different tempera-
tures, he obtained a variation in flower-colour from white (in heat)
through pale blue to deep blue (in cold). He also observed that Primula
sinensis produces tinged flowers in a hot-house and full-coloured flowers
in the cold. Klebs is of the opinion that the colour changes induced
by changes of temperature are not directly due to the effect of tempera-
ture on pigment formation but indirectly to the effect of temperature
on metabolism. At high temperatures, growth is so rapid that the
substances used in pigment formation are not present in sufficient
quantity.

Effect of light.

As regards the effect of light on anthocyanin formation, there have
been numerous observations, some more or less conflicting. As with
temperature, the main question at issue is again, whether light directly
affects anthocyanin formation, or whether its influence is only indirect,
in so far as it affects photosynthesis and the accumulation of the products
of this process, among which are the chromogens from which pigment
is formed 1 .

The effect of the absence of light on the development of pigment
in flowers may first be considered. As early as 1799, Senebier (2)
noted that the Crocus and Tulip develop coloured flowers in the dark.
The same observations were recorded by Marquart (5) in 1835 for
Crocus sativus. Later Sachs made definite experiments on various
plants either by growing them entirely in the dark, or by enclosing

1 I am indebted to Dr F. F. Blackman for drawing my attention to a fact which is
interesting in this connection, i.e., that the development of chlorophyll may be affected
by nutrition. The statement is founded on the observation that certain Algae develop
chlorophyll in the dark when provided artificially with protein; when supplied with
nitrates under the same conditions, however, the pigment does not appear.

88 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

in a dark chamber certain shoots or branches only, the other parts of
the plant being in the light. Of plants grown entirely in the dark,
Sachs (269) was able to distinguish two classes : (a) flowers which develop
colour normally in the dark without being previously exposed to light
(Tulipa, Iris, Hyacinthus, Crocus); (b) flowers which only develop
colour in the dark if the buds have been fully exposed to light until
just before opening (Brassica, Tropaeolum, Papaver, Cucurbita). Of
Tulipa Gesneriana Sachs specially remarks : " Die schon gefarbten und
normal entfalteten Bliithen auf den etiolirten Pflanzen machten einen
hochst sonderbaren Eindruck." And of Iris pumila he says: "der
zart hellblauliche Grundton der Perigonzipfel, die dunkelviolette
Aderung, welche gegen den Grand der Zipfel bin in das blaulich Pur-
purescirende iibergeht, das Orangegelb der Barte, das schon warme
Blau der Narben und die himmelblaue Farbung des Pollens, alle diese
Farbungen waren bei der Bliithe der etiolirten Pflanze eher glanzender
und gesattigter als bei den am Lichte entfalteten." In the case of
Tropaeolum, Papaver, etc., if the plants were darkened just before the
buds unfolded, normally coloured flowers were produced, but later
buds develop flowers with decreasing amounts of pigment. In the
cases where shoots only were darkened, the flowers borne upon them
were normal as regards size, but the coloration was less intense. Further
experiments were made by Askenasy (282) who confirmed Sachs'
results for Tulipa and Crocus, though he found Hyacinthus flowers
rather less coloured in the dark. Other plants (Pulmonaria, Antirrhinum
Silene, Prunella) showed less production of anthocyanin in the dark
if the buds had not been previously exposed to light; in Prunella,
almost white flowers were formed. Sorby (144), Beulaygue (339), Gertz
(19) and others have confirmed these results for the flowers of various
species, showing that considerably less, very little, or no colouring
matter at all, is formed in the dark.

As regards coloration of fruits, observations do not entirely agree.
Senebier notes that apples do not redden unless exposed directly to
light. However, as pointed out by Gertz, the reddening of apples is
more or less accessory, but in fruits of which anthocyanin production
is a distinguishing feature (Crataegus, Rosa, Sambucus), Askenasy (282)
found by partial darkening when green, the pigment developed equally
well in both illuminated and darkened parts. Both Laurent (315)
and Miiller-Thurgau (298) agree that the coloration of grapes can take
place in the dark.

In purely vegetative organs observations are more conflicting. Of

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 89

anthocyanin formed in the dark, the following examples may be
quoted :

Leaves of Beta (Morren).

Seedlings of Phalaris and Secale (Hallier, 272).

Intense rose-red coleoptile in Secale (Gertz, 19).

Red coloration in shoots of Opuntia robusta and 0. leucotricha.

Spots on leaf of Orchis latifolia (Gertz, 19).

Red- veined leaves of Crepis paludosa (Gertz, 19).

Faint reddening in stolons of Solanum tuberosum (Gertz, 19; Sachs,
269).

On the other hand innumerable cases may be quoted in which light
appears necessary for the formation of the pigment:

Reddening of seedlings is entirely absent in the dark in Polygonum
tartaricum, Celosia, Beta (Weretennikow, 273). This has been confirmed
by Schell (285) for seedlings of Polygonum, Rumex, Rheum and Amaran-
thus and by Batalin (293) for Fagopyrum. The most casual observation
will also afford instances of cases where anthocyanin is developed on
the sides of stems, twigs and petioles which are exposed to the sun,
the opposite side remaining green 1 . Such phenomena are specially
mentioned in stems of Cornus sanguinea, C. sibirica, species of Tilia,
Rosa and Rubus (Gertz, 19), of Cuscuta (Mirande, 332), of Helianthus,
Crataegus and stolons of Fragaria (Dufour, 305).

The development of autumnal coloration often only takes place
in the parts of leaves and stems exposed to light, as was noted long
ago in Viburnum Lantana (Voigt, 264) and Rhus Coriaria (Macaire
Princep, 3). Gertz (19) also points out that, in Viburnum Opulus,
V. Lantana, Cornus sanguinea, C. sibirica and Prunus Padus, the
autumnal leaves may show quite clear natural photographs of the
leaves covering them, because anthocyanin is absent from the covered
surface. Similar lines and spots may be observed in winter-reddened
leaves (Silene, Viscaria, Armeria, Hieracium, Pilosella).

Development of pigment in roots exposed to light has been observed
in Salix (Hallier, 272 ; Schell, 287) and Zea (Dufour 305 ; Devaux, 308).

It is difficult to draw conclusions from the rather conflicting state-
ments given above. One fact stands out definitely, namely, that
absence of light itself is no bar to the formation of anthocyanin in many
cases, such as the root of Beta and the flowers of Tulipa and 7m. But
in the majority of such cases there is obviously a plentiful supply of

1 It has been noted by Parkin (77) that the under surface of leaves is more sensitive
to light and reddens more easily than the upper.

90 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

synthetic products in the storage tissues of the bulbs, corms, seeds and
fruits concerned, from which the chromogen for the pigment can be
synthesised, if there is not also already a storage of this substance.
But where the supply of chromogen is directly dependent on the photo-
synthetic activity of the leaves, darkening of the flower-buds, unless
they are already practically mature and supplied with chromogen,
prevents or diminishes the formation of anthocyanin. The reddening
of leaves and stems where exposed to light may in some way be due to
local accumulation of synthetic products, though the direct effect of
light is superficially more probable.

The effect of light in autumnal coloration is even less well explained
on this accumulation hypothesis, except that in the last stages of the
leaf's existence, photosynthesis must be best carried out in the parts
most exposed to light. There is also a general tendency to accumulation
of synthetic products owing to low night temperature.

The formation of anthocyanin in normally uncoloured roots when
exposed to light appears to be the most convincing evidence at hand
for the production of. anthocyanin due to the direct action of light.
Until more evidence has been collected from a number of carefully
devised experiments, no definite inference can be drawn.

In conclusion we may mention some work on more systematic lines
which was published by Linsbauer (371) in 1907. He endeavoured
to find out the more precise relationships between light and the forma-
tion of anthocyanin. For this purpose he used seedlings of Fagopyrum
esculenlum which had been grown in the dark and were quite etiolated.
Such seedlings were then exposed to light (lamp) of different intensities
and for varying lengths of time. From his results Linsbauer concluded
that the photo-chemical process of anthocyanin production in light is
a typical stimulus reaction, and is dependent upon both the intensity
and duration of light. He investigated also the relationship between
the times of reaction and presentation, and found it analogous in many
respects to that in other stimulus processes, i.e. geotropism, for instance.
Whether, however, the appearance of anthocyanin in these seedlings
is due to the direct action of light, or to the products of photosynthesis
induced by light, cannot be readily ascertained.

Connection ivith presence of oxygen.

That oxidation plays a part in the formation of anthocyanin has
frequently been suggested. The production of red pigment through
the oxidation of a chromogen was the hypothesis brought forward

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 91

by Wigand (136) as early as 1862, and the same idea has been revived
successively by Palladin (203), Wheldale (211, 212) and Combes (379).
That the process is controlled by a specific oxidase has been postulated
by Buscalioni & Pollacci (17), Mirande (365) and Wheldale (211, 212).
The actual dependence of the process on the presence of oxygen is
illustrated by the experiments of Mer (284), who mentions the fact that
leaves of Cissus do not redden under water, and in 1899 Emery (309)
noted that no colour is produced in submerged flowers.

More definite experiments were performed by Katie (354), inciden-
tally, among a series of investigations primarily concerned with the effect
of culture solutions on the formation of anthocyanin. Leaves in certain
culture solutions of sugars, and other substances, were found to produce
anthocyanin, but if enclosed in vessels, from which all oxygen had been
removed by alkaline pyrogallol, no trace of colour was observed. Katie
also found that colour was less rapidly developed in air under reduced
pressure than in normal atmosphere. Also in certain culture solutions
an increased pressure of oxygen produced greater development of
colour.

Still more elaborate investigations were made by Combes (379).
The experiments consisted in analysing the gaseous exchange of red
and green leaves under similar conditions. Leaves were employed
in which reddening had taken place and was proceeding from different
causes, as for instance: leaves of Ampelopsis hederacea reddened by
exposure to light; of Rumex crispus and Oenothera LamarcJciana by
attacks of parasites ; of Spiraea prunifolia and Mahonia aquifolium
by decortication ; of Rubus fruticosus with autumnal coloration ; and
finally of young leaves of Ailanthus glandulosa in which reddening,
on the contrary, was disappearing.

From observations upon gaseous exchange in the above leaves,
which are exemplified in the accompanying table, Combes drew the
following conclusions :

Oxygen fixed (+) or lost ( ) during one hour, consisting of half-
hour of day and half-hour of night.

Red leaves Green leaves

Ampelopsis - -0872 ... -329

Rumex + -0963 ... -8067

Oenothera + -2442 ... + -1622

Spiraea + -1357 ... -3226

Mahonia + -0829 ... -1565

Rubus + -2011 ... -0435

Ailanthus - 1-959 - 2-589

92 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

The appearance of anthocyanin is accompanied by an accumulation
of oxygen in the tissues ; the disappearance of the pigment is, on the
contrary, accompanied by a considerable loss of oxygen. The variations
detected in the gaseous exchange during the formation of red pigment
are concerned with assimilation ; it therefore appears that the produc-
tion of these pigments is intimately bound up with the phenomenon
of assimilation. This accumulation of oxygen on anthocyanin formation
can be explained by the diminution in intensity of assimilation, and
'in the modification produced in the ratio of gaseous exchange in assimi-
lation. Further, the accumulation in the cells of soluble carbohydrates
which accompanies anthocyanin production accelerates oxidation, and
the gaseous exchange is fundamentally modified.

Effect of drought.

Here again the direct effect is difficult to estimate owing to the
simultaneous effect on photosynthesis, but whether direct or indirect,
there is ample evidence that drought increases anthocyanin formation.
Molisch (316) found that leaves of Peireskia aculeata, Tradescantia,
Panicum, variegatum and Fuchsia reddened strongly if only watered a
little. The author has made the same observations for pot plants of
Pelargonium. Eberhardt (347) also found an increase of anthocyanin
in leaves of Coleus Blumei and Achyranthes angustifolia when grown
in a very dry atmosphere. According to Warming (327, 346) plants
such as Tillaea aquatica, Peplis Portula and Elatine are green when
growing in water, though individuals on land may be strongly red.

The physiological drought of salt marshes may similarly explain
the development of anthocyanin in halophytes (Salicornia, Suaeda).

An interesting case of the connection between reddening and drought
has been observed by Miyoshi (375). This author noticed that the
leaves of certain tropical trees, especially Terminalia Cattapa, in the
East Indies, Ceylon and Java, take on a beautiful red colour before the
leaf-fall. The reddening is described as affecting at first a few leaves
only, but later the number increases. Of the early stages the author
says: "Vor der Feme betrachtet erschienen die gefarbten Blatter wie
rote Bliiten in voller Pracht." Since the phenomenon takes place at
the dry period of the year, Miyoshi suggests the term 'Trockenrote,'
and considers the causes of reddening to be drought coupled with strong
insolation.

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 93

Effect of sugar-feeding.

The term sugar-feeding means that the plant is supplied artificially
with extra amounts of sugar. In the case of floating or submerged
water-plants, the whole plant can be immersed for experiment in dilute
sugar solution ; in the case of land plants, the stems of leafy branches,
or the petioles of isolated leaves, can be put in the solution; or the
leaves can be cut into pieces and floated in the liquid.

The first investigations of the results of this process were carried
out by Overton (333). While conducting some experiments on osmosis
with Hydrocharis Morsus-ranae, Overton noted that the leaves of this
plant tended to become red when the plants were grown in 5 % cane
sugar solution. Later, the idea occurred to him that autumnal colora-
tion might be due to excess of sugar in the leaf tissues, and he claims to
have shown that autumnal leaves contain more sugar than green leaves.
On the basis of this idea, he then commenced some systematic investi-
gations on sugar-feeding with a view to gaining more knowledge of the
whole phenomenon.

Experiments were first made with Hydrocharis plants grown in
various solutions, and the results were as follows :

Hydrocharis Morsus-ranae. Plants in 2 % invert sugar showed
excess of anthocyanin over the control in four days. This was true
of pigment in all parts leaves, petioles, stolons and roots, and the
intensity of pigmentation increased with time. In 2 % cane sugar
the results were similar.

Plants in both the above solutions flowered rather earlier than control
plants, but the flowers were unaffected by the cultures.

In 2 % glucose there was the usual reddening ; also in 2 % laevulose.
In 2 % galactose there was no reddening, and in 5 % lactose, the
reddening was very slight. In 2 %, 4 % and 10 % glycerine, no colour
developed ; this was also the case in solutions of potassium nitrate,
sodium chloride, sodium sulphate and of other salts.

In 3 % invert sugar in a dark room there was no trace of reddening.

Microscopically it was found that colour was produced in the meso-
phyll cells, and never in either the upper or under epidermis.

The following species were also used:

Elodea canadensis. In 2-3 % invert sugar, a reddish colour developed
in the younger leaves.

Vallisneria spiralis. In sugar cultures there was an increase of red
colour which was located in the epidermis as well as in the inner tissues.

94 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

Potamogeton perfoliatus, P. pectinatus. In 2-3 % invert sugar there
was no result. In P. pusillus, a reddish colour appeared.

In Najus major there was no effect but in N. minor there was a slight
result.

Lemna minor, L. trisulca in various kinds of sugar solutions gave
negative results. The same was true for Pistia Stratiotes.

Utricularia Bremii. In 2 % invert sugar reddening in the bladders
appeared in two to three days. Finally the leaves and bladders became
quite red. In 2 % cane sugar the result was similar. In Utricularia
minor, reddening appeared in -5 % cane sugar; also in glucose, invert
and cane sugars (-5 %-5 %). U. vulgaris behaved similarly to U. Bremii
and U. minor. In 2-5 % lactose, there was no colour in two weeks,
but after four weeks a slight colour (due to hydrolysis probably).
There was no colour in galactose; slight colour only in glycerine and
none in salt solutions.

In Utricularia, reddening in sugar cultures developed just as little
in complete darkness as in Hydrocharis.

Ceratophyllum demersum. In 2-3 % invert sugar there was some
reddening in the cells of subepidermal and deeper tissues, though the
epidermis was uncoloured.

Experiments were next made with land plants, using either leafy
twigs or isolated leaves :

Lilium Martagon. A leafy stem was placed in 2 % invert sugar.
This and control stems in distilled water were placed in a south-east
window. After about seven days, the plant in sugar solution showed
reddening in the leaf-tips, which afterwards spread, day by day, over
the leaf, while the control showed no reddening. The pigment was
found to be located in the inner leaf tissues, the upper and under epi-
dermal cells being free. Experiments were made with about 20 other
specimens, and in each case the results were the same. In 2 % glucose
solution reddening could be detected in four days, becoming more intense
in the course of time. In 2 % fructose, there was a similar result. In
2 % cane sugar the red coloration came later and was less intense.
In lactose, galactose and glycerine solutions there was no reddening,
and the same was the case with various salt solutions. It was found
also that ethyl and amyl alcohols, ketones and ether, in solution, caused
development of red pigment, but Overtoil considers the result to be one
of injury. He is inclined to believe that the alcohols, ketones, etc.,
acted as narcotics and so prevented translocation, rather than that
they served as material for building up the pigment.

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 95

Fritillaria imperialis was found to produce no colour in sugar solution.

Ilex Aquifolium. A twig of Ilex was put in 3 % invert sugar. In
two days some reddening appeared whereas control plants showed
no colour. The pigment was found in the palisade, and to some extent
in the spongy parenchyma, but none was present in the epidermis.
In glucose and fructose there was also considerable reddening.

Hedera Helix. In 2-3 % invert sugar, red colour appeared in a
few days, but it was not intense nor was it uniformly distributed.

Mahonia Aquifolium. A twig of this plant gave a negative result
in 2 % invert sugar.

Ligustrum vulgare. In 2 % invert sugar, red pigment appeared
after eight days. The pigment was localised in the palisade cells.

Ampelopsis hederacea. In 2 % invert sugar reddening commenced,
but the experiment could not be continued as the leaves tended to fall
from the leaf -stalks. In autumn, good results of artificial reddening
were obtained, green leaves only, of course, being used.

Saxifraga crassifolia. In 3 % invert sugar the leaves reddened
in a few days.

Aquilegia vulgaris. Young leaves with petioles in 2 % invert sugar
showed distinct reddening in four days.

Taraxacum ojficinale. When the base of the leaves was placed in
a 2 % solution of invert sugar, a very fine red colour developed in two
or more days over the whole leaf, and the pigment was located in the
inner tissues and not in the epidermis. Leaves, however, in distilled
water may redden in time. Reddening of leaves was found to be
characteristic of many Compositae, and as this result was often obtained
to some extent in distilled water in a good light, they did not afford
very suitable material. Leaves of Eupatorium Cannabinum and Pre-
nanthes pur p urea which did not redden to any extent in distilled water,
became very red in sugar solution.

Epilobium spp. Leafy stems of E. parviflorum in 2 % invert sugar
gave a good red colour.

Negative results in 2-3 % invert sugar were obtained with
Anthriscus silvestris', also with Rubus species.

From the above researches Overtoil draws the conclusion that in
many species of Monocotyledons and Dicotyledons, both water and
land plants, sugar-feeding will bring about anthocyanin formation.
There is also this further correlation, that, with the exception of sub-
merged water plants, there is a negative result with sugar-feeding
when the plant in normal circumstances produces anthocyanin in the

96 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

epidermal cells. But if in normal circumstances the pigment is formed
in the inner tissues, then sugar-feeding (especially with fructose and
glucose) produces red pigment in a high percentage of cases, and this
pigment is localised in the mesophyll and not in the epidermis.

Overton also tried the effect of putting the inflorescence of white-
flowered varieties into sugar solution. Thus, for instance, the
inflorescence stalk of white Pelargonium zonale was placed in 3 %
solution of invert sugar, but no trace of red colour was formed in the
flowers though the stalk showed reddening. Negative results were
also obtained with Anemone japonica. Overton concludes that some
other factor, apart from presence of sugar, is necessary in these cases.

Further researches on sugar-feeding were made some years later
by Katie (354), and of these the experiments on Hydrilla verticillata
(Hydrocharitaceae) are given in the greatest detail as follows:

Inorganic culture media containing various salts of potassium, sodium,
calcium, magnesium, ammonium, iron, aluminium and lithium had
practically no effect on the formation of pigment. In glucose (-05-3 %)
solutions, red pigment was formed in the light and in the dark, and in
isolated leaves more quickly than in pieces of leafy stem. In laevulose
(1-5 %) and cane sugar also pigment developed both in the light and
in the dark, but the cane sugar (of concentration -5-25 %) was most
favourable to reddening. Only a slight coloration appeared with maltose
(1-3 %) both in the light and in the dark. In lactose (1-5 %), raffinose
(1-10 %), inulin (1-3 %) and glycerine (4-5 %) pigment was formed
only in the light. Some colour was developed in (1-4 %) ethyl alcohol
and in mannite (1-2 %), but none in galactose (up to 5 %), arabinose,
formol (-001 %), dextrin, salicin or asparagin.

The effect of mixed solutions of carbohydrates and various salts,
such as those mentioned above, was tried. It was found that potassium
nitrate and mono-potassium phosphate quickened the formation of
pigment in sugar solution. Various other potassium salts (except
potassium bichromate) had the same effect in a less and varying degree.
Katie is of the opinion that the effect of potassium salts is chemical
and not osmotic. Sodium salts, on the whole, were found to have
little effect; magnesium sulphate and nitrate some positive effect;
calcium salts (except CaH 4 (P0 4 ) 2 ), aluminium sulphate, ferric chloride
and some ammonium salts a preventative effect.

Various alkalies, sodium carbonate, potassium hydroxide, calcium
hydroxide and magnesium oxide were tried with sugar solutions; the
result was to quicken the formation of anthocyanin. Acids, as for

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 97

instance, tartaric, salicylic, citric and oxalic, on the contrary, appeared
to slow down the formation ; to this tannic acid formed an exception,
since the reddening appeared more quickly in its presence.

Although anthocyanin was produced in the cultures in the dark,
Katie found that the development was always earlier in the light. As
regards temperature, it was found that 25 C. was the optimum for the
appearance of the pigment in the light and 28 C. in the dark.

The absence of carbon dioxide had no effect on the formation of
anthocyaniu in sugar solutions, except in the case of glycerine, where
the development was weaker. Oxygen, on the contrary, was found to
be necessary for the reddening.

Katie made further observations with a number of other plants,
and although his experiments are described in great detail in his disser-
tation, no more than a short statement will be given here as they were
largely on the same lines as those with Hydrilla.

Elodea canadensis. In -26-1 % cane sugar there was a slight colora-
tion in the light, but none in the dark ; in 1 % grape sugar there was
colour both in the light and in the dark. In the case of cane sugar the
development was increased by addition of potassium nitrate and calcium
sulphate.

Hydroclmris Morsus-ranae. In cane sugar and laevulose some
pigment appeared in the dark; in grape sugar and maltose, there was
none in the dark.

Sagittaria natans. In 5 % cane sugar there was a considerable
development of colour, but only in the light.

Allium Cepa. The colourless scales of a variety which normally
contained some anthocyanin turned red in sugar solution ; a white
variety, however, produced no colour under any conditions.

Canna indica. Etiolated leaves in 5-15 % cane sugar formed pig-
ment.

Veronica Chamaedrys. In sugar cultures colour only developed in
the light and was located in the epidermal cells.

Rosa 'Marechal Niel.' Green leaves, or pieces of leaves, in 15%
cane sugar formed colour only in the light. It was found in both
epidermis and spongy tissue.

Saxifraga cordifolia. Pieces of a leaf in 5 % cane sugar only gave
a good development of anthocyanin in the light.

Pittosporum (undulatumt). Green leaves and pieces of leaves in
5-15 % cane sugar formed pigment only in the light, and it was found
to be located in the inner tissues and not in the epidermis.

w. P. 7

98 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

Bellis perennis. In 15 % sugar solution green leaves developed
a weak red coloration only in the light, and it was confined to the
epidermal cells.

Thus we see that Katie's results differ from those of Overtoil in two
points. First, anthocyanin may develop in sugar cultures in the dark,
and secondly, it is not necessarily confined to the epidermis.

The question of the results of sugar-feeding has been taken up more
recently (1912) by Gertz (386). The following account is taken entirely
from his paper. He deals with a point in Overtoil's results we have just
mentioned, i.e. the statement that by sugar-feeding anthocyanin can be
induced to form in the mesophyll, but not in the epidermal cells. Thus it
would appear possible that the chloroplastids might have some influence
on the formation of pigment, and the failure of petals to produce antho-
cyanin in sugar culture might be considered a corroboration of this
view. In Vallisneria spiralis and Elodea, Overtoil found epidermal
anthocyanin on sugar culture, but this is no proof, since the epidermis
in those plants contains chlorophyll. General evidence is, moreover,
against the view of the connection with chloroplasts, as the epidermis
often contains anthocyanin, and this pigment is also developed in
chlorophyll-free saprophytes and parasites and in perianth leaves.

In order to investigate this point, Gertz made some sugar culture
experiments with parts of albino leaves which were free from chloro-
plastids. For this purpose he used first the leaves of Oplismenus
imbecillis (Graminaceae), and found that anthocyanin may be induced
to form in the complete absence of chlorophyll. Therefore the question
of the activity of the chloroplastids is definitely solved, but as only this
one species had been used, further experiments were made with other
plants. A modification of Overton's method was employed; portions,
20 x 20 mm. square, were cut out from the leaves and floated upside
down on the solutions in glass dishes, and in this way free transpiration
and respiration could go on through the stomata. The solutions were
also changed from time to time, and the edges of the leaves freshly
cut. Cane sugar only was used of 5-10 % concentration. The following
results were obtained.

Oplismenus imbecillis. The leaves are variegated red, green and
white. In the white portions there are no chloroplastids, not even
in guard cells of the stomata. When white portions were used, antho-
cyanin appeared in quantity in the epidermis after four days. In the
dark, however, there was only slight, though definite, coloration.
Hence Gertz concluded that, though not absolutely necessary, light is

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 99

a powerful agent in assisting the formation. Cultures in distilled water,
both in the light and in the dark, produced no anthocyanin. To obtain
absolutely comparable results, parts of the same leaf, A, B, C and D
were treated as follows. A and B were placed in 5 % sugar solution,
C and D in distilled water ; A and C were exposed to the light, B and
D kept in the dark. The results were as before. Both A and B formed
anthocyanin, but it was insignificant in B; C and D were free from
anthocyanin.

Tradescantia Loekensis. Again the leaves are variegated red,
white and green. No chlorophyll occurs in the white parts except in
the guard cells. In cane sugar the result was negative and no antho-
cyanin was formed. Gertz has no suggestions to offer except that
possibly some unknown factor is essential to the formation of pigment.

Beta vulgaris. Only negative results were obtained.

Rumex domesticus. A variegated green and white form was used,
of which the white parts, including the guard cells of the stomata,
are entirely free from chlorophyll. In less than one day in 5 % sugar
solution, anthocyanin appeared, and was found to be located in the
lower as well as in the upper epidermis. In the dark, less pigment was
formed. Experiments were made with ordinary Rumex, and were in
complete agreement, since anthocyanin was formed chiefly in the
epidermis (except stomata). Similar observations on R. Patentia have
also been made by Palladin (203),

Cornus florida. The leaves have green and white parts, the latter
being free from chlorophyll. After a week in 10 % cane sugar they
showed only traces of anthocyanin which was located in the spongy
parenchyma.

Euonymus radicans. In 10 % sugar, the albino parts showed a
faint reddening in the lower epidermis with the exception of the stomata.

Lonicera brackypoda. The leaves have white reticulations on a
green ground. After a week in 10 % cane sugar solution, anthocyanin
was found in the palisade parenchyma.

Other experiments were made with green leaves as follows :

Plantago major. In 10 % sugar solution, the pieces of leaves red-
dened strongly, and anthocyanin was located entirely in the lower
epidermis except the stomata. Attempts were made to induce antho-
cyanin formation in sugar cultures in isolated epidermal layers, but they
were unsuccessful.

Sium latifolium. In 10 % sugar solution, anthocyanin appeared
readily in pieces of leaves, as well as in entire leaves and shoots, and

72

100 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

was found to be localised in cells of the mesophyll. With Cerefolium
sylvestre negative results were obtained.

Epilobium parviflorum. In 5 % sugar solution, anthocyanin
developed plentifully in the lower epidermis and occasionally in the
upper.

Euphorbia Cyparissias. In bracts in 10 % sugar solution, antho-
cyanin was formed in both the upper and lower epidermis. In Phyto-
lacca decandra it appeared in the epidermis (except stomata).

Thus Overtoil's view, which is expressed as follows "...mit
Ausnahme der untergetauchten Wasserpflanzen scheinen solche Versuche
fast durchweg bei denjenigen Pflanzen negativ auszuf alien, deren
natiirliche Rothfarbung...der Gegenwart von rothem Zellsaft in den
Epidermiszellen zu verdanken 1st," is, as Gertz points out, not correct;
for colour due to anthocyanin appears entirely in the epidermis in
Oplismenus imbecillis, Euonymus radicans, Plantago major, Euphorbia
Cyparissias and Phytolacca decandra; both in the epidermis and meso-
phyll in Rumex domesticus, Tussilago Farfara and Epilobium parviflorum.
On the other hand in Cornus florida, Lonicera brachypoda and Sium
latifolium it appears in the ground parenchyma. These results are
also in accordance with those of Katie.

Thus it would appear to be definitely settled that chloroplastids
are not essential to anthocyanin formation from the results with leaves
of Oplismenus imbecillis , Rumex domesticus, Cornus florida, Euonymus
radicans and Lonicera brachypoda, though the results with Tradescantia
and Beta are negative. Yet Gertz does not .seem to be quite assured
on the point, partly because, as he points out, the epidermal leucoplastids
are closely related to chloroplastids, and partly because the epidermis
is connected with chloroplast-containing cells.

As regards the effect of light on the results of sugar-feeding, Overton
maintained that colour was not produced in the dark. This was not
found by Gertz to be the case with Oplismenus and Rumex, and Katie,
moreover, demonstrated that Hydrilla, Hydrocliaris, Allium and Pha-
laris develop pigment in sugar cultures in the dark. Gertz is of
the opinion that under natural conditions the appearance of antho-
cyanin may not be very largely affected by illumination, as a whole
constellation of factors may take part in its formation in the kind of
way we have tried to indicate in the previous sections.

Gertz finallv considers the formation of anthocyanin in petals
resulting from sugar-feeding. As we have seen, Overton failed to get
any result with Pelargonium and Anemone. Gertz considers that such

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 101

a result, if it were possible, would give an even more important proof
that anthocyanin production is independent of chloroplastid activity.
A striking observation in this connection has been made by Goiran
(310) on Cyclamen persicum giganteum of which the flower is white
except for a patch of red anthocyanin at the base of the petal. Goiran
made transverse cuts through part of the petal, and though the petal
tips remained fresh, after a time they developed red pigment. This
result is akin to the formation of anthocyanin in leaves when the
conducting system is injured by cutting, or by insects, Fungi, etc.
There is also an observation by Gertz on white petals of Saxifraga
in which anthocyanin appeared as a result of insect attacks. Gertz,
however, failed to get any result from sugar cultures of Bellis perennis,
Anemone japonica, Magnolia acuminata, Tropaeolum majus, Deutzia
gracilis, Begonia sp., and Pisum sativum. But eventually a positive
result w T as obtained with white petals of Viburnum Opulus (the cultivated
form with neuter flowers). These petals in sugar solution formed
anthocyanin in either one or both epidermal layers. Thus Overton's
prophecy that petals might be found which would redden in sugar
culture appears to be justified in this case.

It is not an easy task to set forth any general explanation of all
the facts recorded in the previous sections, but a short review, as far
as possible, of the situation may be useful at this stage. Thus, we
may say that two points emerge from the preceding considerations :
(1) that the chromogen for anthocyanin is formed in the leaf; (2) that
an accumulation of carbohydrates (sugars) and glucosides (including
the glucoside of the chromogen) leads to the formation of anthocyanin.

One explanation offered for the reddening phenomenon has been
advanced by various authors, Palladin (203), Mirande (365) and Combes
(379), that presence of excess of carbohydrates increases oxidation
processes, and hence, if anthocyanin is an oxidation product, the forma-
tion of this pigment. Such an explanation would meet practically
all cases of anthocyanin formation, but there is no special evidence for
its justification.

Another suggestion is that made by the author (226) ; it is based
on certain reversible reactions involving pigment, chromogen, glucosides
and sugars, and will be considered in detail in the next chapter. For
the moment we need only say that the hypothesis assumes that the
chromogen is formed from sugars in the leaf, and that increase in amount
of sugar leads to increased formation of chromogen with the resultant
production of anthocyanin, unless the chromogen be removed.

102 PHYSIOLOGICAL CONDITIONS AND FACTORS [CH.

Let us now consider how this explanation fits various cases.

(1) In a normal plant with green leaves, coloured flowers and tinged
stems and petioles, the chromogen is synthesised from sugars in the
leaves and translocated away when formed. If the plant is kept in
the dark or shade, photosynthesis stops or is lessened, the supply of
chromogen falls below normal, and the flower-colour may be pale and
the stems and petioles green. Conversely, great photosynthetic
activity produces a plentiful supply of chromogen which results in
rich flower-colour and appearance of pigment in the vegetative organs.
(In addition, light itself may directly increase pigment formation.)
The intense colours in the flowers and the development of anthocyanin
in the vegetative parts of High Alpine plants may be explained by
strong insolation, stunted growth employing little material, and slow
translocation due to low night temperature 1 . The power of some

1 Exception will probably be taken to this statement on the ground that many High
Alpine plants can be grown in lowland regions and their flowers do not then show any
perceptible loss of colour, whereas if the statement were true, we might expect to find
considerable diminution of colour under these conditions. As a matter of actual fact,
on the basis of observation, this criticism is not altogether valid, for, as we have previously
stated, Gaston Bonnier (328). Kerner (398) and others have shown that in many cases the
flowers of plants growing at high altitudes and in high latitudes have a more intense colour
than those of individuals of the same species grown either in the plains or in lower latitudes.
There is also a more general aspect of the question, which may be outlined as follows.
Every plant is the expression of a chemical (or, fundamentally, physical) entity, and this
expression can only fluctuate within limits on account of the defmiteness of the chemical
(or physical) constitution underlying it. Broadly speaking these chemical (or physical)
entities are adapted to their habitats, that is, they are only able to exist under those con-
ditions in which the chemical reactions (or physical processes) essential to their existence
can take place. Sometimes a plant, for example species of Opuntia, will live and flourish
to a certain extent in a habitat to which it is not adapted, such as the temperate zones.
The conditions for its metabolic processes do not in this case lie beyond the limits of such a
climate. But the expression of its entity, of necessity, remains the same. We may assume
that, for Opuntia, variation in the direction of the characters of plants of the temperate zones
does not occur, since such changes would involve chemical reactions (or physical processes)
which are outside the sphere of its constitution. Hence it remains a typical species of
Opuntia. Other examples no doubt could be selected where there were greater fluctuations
on change of habitat, but these again would depend either on the wideness of the sphere
of chemical (or physical) activity of the particular plant, or on the relation of the particular
fluctuation to the environment. The same line of argument we have applied to Opuntia
can be applied to the members of any plant formation, and among them, to High Alpine
plants. These are on the whole adapted to their environment, and intense flower-colour
is one of the expressions of the entity of this particular type of plant. Transferred to
lowland regions, such a plant is still able to grow and flourish, and of necessity it retains
its entity. Yet fluctuations in colour intensity of the flowers and in the amount of
pigment in the vegetative organs are, within limits, the kind of variation we should suppose
possible, though not necessarily inevitable ; for such fluctuations depend merely on the

vi] INFLUENCING THE FORMATION OF ANTHOCYANINS 103

plants to produce normally coloured flowers and fruits in the dark is
due to a plentiful supply of reserve material (flowers of Tulipa).

(2) If there is more than a plentiful supply of chromogen owing
to the flowering parts being removed or to a large supply of carbo-
hydrates,, a certain amount of abnormal reddening may take place,
ex., leaves of plants deprived of the inflorescence, adventitious shoots
from tree trunks ; possibly this may explain reddening in sugar cultures.

(3) A still further increase of carbohydrates and chromogen such
as is caused by blocking the translocation current leads to still more
abnormal reddening, ex. injured leaves, decorticated stems, etc.

(4) In autumnal and winter leaves, there may be an accumulation
of sugar and chromogen owing to the slowing down of the translocation
current, and the lack of starch formation at low temperatures. The
effect of low temperature may also retard growth in general and the
using up of synthetic products.

(5) In young leaves it is possible that the mechanism for trans-
location is not developed so early as the powers of synthesis of sugar
and chromogen. Hence again there is accumulation of these products.

direction, or amount, of certain chemical reactions which are, in any case, part of the
essential metabolism of the plant. If we are to believe the evidence of Gaston Bonnier
and others, it is just these variations which may occur.

CHAPTER VII

REACTIONS INVOLVED IN THE FORMATION OF ANTHOCYANINS

Nehemiah Grew (1) evolved, as a result of his observations, an
elaborate scheme to explain the origin of the various pigments in plants.
It is expressed in the scientific language of the period and is difficult to
translate into modern conceptions, but it is clear that he believed the
presence of air to play an important part in the process. Speaking
of colours in roots, he says these organs show less variety in this respect
than leaves and flowers 'The Cause hereof being, for that they are
kept, by the Earth, from a free & open Aer; which concurreth with
the Juyces of the several Parts, to the Production of their several
Colours. And therefore the upper parts of Roots, when they happen
to stand naked above the Ground, are often deyed with several Colours."

A publication by Schiibler & Franck (117) in 1825 perhaps contains
the first definite hypothesis as to the origin of red and blue pigments.
These investigators treated extracts from coloured flowers and leaves
with acids and alkalies, and the results led them to the view that green
pigment of leaves occupies a mean position, from which an oxidised
yellow-red series is formed, on the one hand, by the action of acids,
and a deoxidised blue-violet series, on the other hand, by the action
of alkalies. The hypothesis was further reinforced in 1828 by Macaire-
Princep (3), who maintained that chlorophyll, when treated with acids,
becomes oxidised first to yellow, then to red and orange pigments, and
this oxidised chlorophyll can be turned green again by alkalies. The
red plant pigments are, in his opinion, oxidised chlorophyll, and the
blue, a mixture of red chlorophyll with a vegetable alkali. All colours
may thus be looked upon as simple modifications of chlorophyll.

These erroneous ideas were accepted, without any real experimental
basis, by plant physiologists of the day as Von Mohl (7, 8) very clearly
points out in 1838. "Les physiologistes s'occupaient plutot generale-
ment de faire des speculations sur les couleurs des plantes, de les

CH. vii] KEACTIONS IN FORMATION OF ANTHOCYANINS 105

rapporter aux couleurs du prisme, que de chercher a connaitre la nature
des matieres colorantes elles-memes. Comme dans le prisme, le vert
tient le milieu et se trouve borde, d'un cote, par le jaune et le rouge,
et de 1'autre cote,, par le bleu et le violet, on croyait que le vert des
plantes etait de meme le point indifferentiel entre la serie de couleur
rouge-jaime et celle du bleu, et c'est par 1'oxygenation et la desoxy-
genation de la couleur verte, qu'on cherchait a expiiquer 1'origine de
ces couleurs, en se fondant sur des experiences chimiques incertaines,
sur des idees fausses d'oxygenation et de desoxygenation, sur Faction
des alcalis et des acides." A more correct point of view was reached
too by Leopold Gmelin (137) in 1828 1 ; he refused to accept the fact
that chlorophyll is reddened by acids, and that the products formed
from chlorophyll by acids, or naturally in autumnal leaves, again
become green with alkalies. He notes, also, that red autumnal leaves
contain both yellow chlorophyll, and a blue colouring matter which
is reddened by acids.

Gmelin's criticisms, as well as others of the same kind, appear
to have been unconvincing; for again, in 1835, Clamor Marquart (5)
returns to the origin of anthocyanin from a metamorphosis of chloro-
phyll. Though objecting to the oxidation and deoxidation hypothesis,
he retains the idea of the formation of other pigments from chlorophyll,
this time, however, by addition and subtraction of water. By the
action of strong sulphuric acid, i.e. by subtraction of water from chloro-
phyll, a blue colouring matter 2 (anthocyanin) is obtained which turns
red with acids and green with alkalies. This blue substance, he says,
forms the basis of all blue, violet and red flower pigments.

Marquart's evidence was thought to be insufficient by Von Mohl
(7, 8) ; for, with regard to the sulphuric acid reaction, the latter very
naturally remarks: "Si, dans ce cas, la couleur bleue doit annoncer
la formation artificielle de 1'anthocyane par la chlorophylle, il est
impossible de concevoir pourquoi, malgre la presence de 1'acide sulfurique
libre, la chlorophylle reste bleue et ne devient pas rouge." Also, with
respect to Marquart's evidence that cells which originally contained
chlorophyll, later contain anthocyanin, as for instance petals which
are primarily green, and then become blue or red. Von Mohl points
out, among other evidence, first, that anthocyanin is characteristic
of the epidermis while chlorophyll is found in the inner tissues ; secondly,
that chlorophyll may be found, apparently in no lessened quantity,

1 Earlier edition than that in the Bibliography.

2 This is of course, as we now know, a reaction given by plastid pigments.

106 REACTIONS INVOLVED IN [CH.

in cells which have become reddened with anthocyanin. In the course
of time, as investigations proceeded, it became clear that anthocyanin
is not derived from chlorophyll; Macaire-Princep and Marquart had
been led astray in their views by the close connection between the
disappearance of chlorophyll, and the appearance of anthocyanin,
and vice versa. Yet, in the light of the additional knowledge which
we now possess, it can still be stated with truth that there is a certain
relationship, though an indirect one, between the two pigments, or,
more strictly perhaps, between the spheres of their chemical activity.
The hypothesis of Wigand (136) in 1862 came very near to those
held at the present time. Wigand maintained that anthocyanin arises
by oxidation from a colourless substance which occurs in solution in
the cell, and which changes to red under certain circumstances, and
may, after a time, become colourless again. He considers it to be a
tannin on the grounds that:

1. The red colouring of spring and autumn appears only in tannin-
containing plants, and though it is not always found in these, yet it
never develops in plants free from tannin.

2. Only in those tissues or cells (especially the epidermal cells and
veins) in which the tannin was previously present does the colouring
matter develop later.

3. The red sap, like the tannin, is turned green or blue with iron
salts, and yellow with potash or ammonia.

The tannins of Wigand were very probably, in many cases, flavones,
from which it is now believed that the anthocyanins may arise.

From analyses of pigments of wine and grapes, Gautier (149, 175), in
1893, also formed the opinion that anthocyanin is an oxidised product
of tannic acids, but no suggestions are made as to the particular reactions
involved. At a later date, in 1897, Overton (333) concluded, on the
basis of many observations and experiments, that red pigment is formed
when there is an accumulation of sugars, but beyond stating that antho-
cyanin is probably a tannin -like substance existing, in combination
with sugar, as a glucoside, Overton makes no definite statement as to
the mode of formation. Mirande (365) in 1907 suggested that the
appearance of anthocyanin when leaves are injured by insects is due
to an accumulation in the tissues of tannins and glucose, accompanied
by the presence of an oxidase. Laborde (199, 200, 201) in 1908 also
came to the conclusion that there is a relationship between tannins
and anthocyanins. To quote Combes: "1'auteur (i.e. Laborde)
assimile le phenomene du rougissement a une action diastasique qui

vn] THE FORMATION OF ANTHOCYANINS 107

donnerait naissance a une matiere colorante rouge aux depens d'un
noyau chromogene de nature phenolique que possederaient tous les
tannins." Laborde was able to obtain red pigments from tannins by
means of certain chemical reagents and other treatment.

The authors mentioned are only a selection of those who have believed
in the origin of anthocyanin from tannins, and the hypothesis held the
field for many years; we find it accepted in most text-books even of
fairly recent date (Pfeffer) 1 , though the evidence for its acceptance is
far from satisfactory.

The next hypothesis of importance is that of Palladin (203, 210) 2 ;
he considers anthocyanin to be a member of a class of pigments which
he himself terms 'respiration pigments.' Although this hypothesis is
connected primarily with the function of anthocyanin, nevertheless
certain reactions are involved which justify its consideration in this
chapter. Since Palladin's views are largely based on the action, of
oxidising enzymes, some preliminary account of these substances may
not be out of place at this point.

Certain organic compounds, such as guaiacum tincture, a-naphthol,
paraphenylene-diamine. benzidine, etc., are used as tests for oxidases
since they become oxidised to coloured products when treated with
oxidising enzymes under certain conditions. When the juice or water
extract of some plants is added for instance to guaiacum tincture, a
blue colour is immediately developed, and the plant is said to contain
a direct oxidase. Of other plants the juice or extract gives no colour
until hydrogen peroxide is added, and the plant is said to contain ah
indirect oxidase. A direct oxidase has been considered by many
authors to consist of a system, peroxide-peroxidase ; in the case of the
indirect oxidase, the peroxide is missing, and has to be supplied in the
form of hydrogen peroxide. If a systematic examination be made of
all natural orders as to their oxidase content, it will be found that the
direct oxidase reaction is characteristic on the whole of certain orders
or genera (Compositae, Labiatae, Umbelliferae, etc.), and the indirect
oxidase reaction of other orders (Cruciferae, Ericaceae, Crassulaceae,
etc.). It may be noted, in addition, that the plants giving the direct

1 The Physiology of Plants, translated by A. J. Ewart, Oxford, 1900.

2 Also Palladin, W., 'Das Blut der Pflanzen,' Ber. D. hot. Ges., Berlin, 1908, xxvia,
pp. 125-132. 'Die Verbreitung der Atmungschromogene bei den Pflanzen,' ibid.
pp. 378-389. 'Ueber Prochromogene der pflanzlichen Atmungschromogene,' ibid.
1909, xxvn, pp. 101-106. 'Ueber die Bedeutung der Atmungspigmente in den Oxyda-
tionsprozessen der Pflanzen,' ibid. 1912, xxx, pp. 104-107. 'Die Atmungspigmente
der Pflanzen,' Zschr. physiol. Chem., Strassburg, 1908, LV, pp. 207-222.

108 REACTIONS INVOLVED IN TCH.

oxidase reaction turn brown on injury, or on exposure to chloroform
vapour, or often when placed in absolute alcohol. The juices and
extracts of such plants also turn brown or reddish-brown on exposure
to air.

The results of investigations made by the author 1 , considered in
conjunction with those obtained by other workers, led to the suggestion
that the direct oxidase reaction is given only when, in the plant meta-
bolism, certain substances are formed which can (after the death of the
plant) autoxidise in presence of air, and which then in the state of
organic peroxide form a system capable of oxidising certain artificial
acceptors, such as guaiacum, etc. There is evidence 2 that such a
substance is either pyrocatechin or some compound containing the
pyrocatechin nucleus. Hence, if this supposition be true, it is obvious
that the classification into direct and indirect oxidase, on the basis
of the blueing of guaiacum, etc., is a purely artificial one. The important
element is the peroxidase which is practically universally distributed
in plants. There is no direct evidence that the peroxide-peroxidase
system exists in the living cell, though the presence of some such system
is extremely probable. The formation of brown or reddish-brown
pigments in extracts or tissues of those plants which contain direct
oxidase may then be regarded as the outcome of the oxidation of a
mixture consisting of a peroxidase and a number of aromatic substances,
of which one at least is capable of acting as a peroxide by autoxidation.

It is upon reactions of the kind just mentioned that Palladin's
hypothesis of ' respiration pigments ' is based. Palladin makes extracts
from a number of plants throughout the vegetable kingdom, and after
boiling to destroy any enzyme in the extracts themselves, adds peroxidase
solution (obtained from Horse-radish root) and hydrogen peroxide. In
all cases, red, reddish-brown, brown or purple pigments are produced in
the extracts ; these pigments, moreover, are formed most readily and in
greatest quantity in extracts of plants we know to contain, previous
to heating, a direct oxidase, i.e. an organic peroxide. From his results
Palladin deduces the fact that chromogens of an aromatic nature are
universally distributed, and may be oxidised in the presence of oxidising
enzymes and a peroxide (though these, let it be noted, he always adds
to the extract). The whole series chromogen, enzyme, peroxide and
pigment form a system for transferring oxygen to respirable materials,

1 Wheldale, M., 'On the Direct Guaiacum Reaction given by Plant Extracts,' Proc.
R. Soc., London, 1911, LXXXIV B, pp. 121-124.

2 Wheldale, loc. cit.

vn] THE FORMATION OF ANTHOCYANINS 109

and hence the term 'respiration pigments.' The chromogens, moreover,
are considered by Palladin to be present in the living plant as prochro-
mogens of the nature of glucosides, the hydrolysis and synthesis of which
are controlled by glucoside-splitting enzymes, and chromogen is only
produced as required for oxidation. After the death of the plant the
hydrolysis of the glucosides is rapidly increased, and the chromogen
becomes entirely oxidised:

prochromogen (glucoside) + water -^- chromogen + sugar
chromogen + oxygen -*- ' respiration pigment '

Palladin includes anthocyanin 1 among the respiration pigments,
and explains its appearance in leaves fed on sugar, in young leaves
and autumnal leaves, as due to excess of carbohydrates: "Diese
Tatsache kann in der Weise gedeutet werden, dass durch Zuckerzugabe
die Atmungsenergie so gesteigert wird, dass ein Teil des oxydierten
Chromogens nicht momentan wieder reduziert werden kann."

It was first suggested in 1909 by the present writer (212) that many
anthocyanins may be derived from the flavones or possibly xanthones 2 .
The flavones are a group of natural colouring matters, some of which
have been artificially synthesised by Kostanecki 3 . As a group they
are widely distributed, and the greater number have been isolated by
Perkin 4 from plants used commercially for dyeing. They may be
regarded as oxy-derivatives of /3-phenyl-benzo-y-pyrone :

CO

The flavones differ from each other in the number and position of
their hydroxyl groups. They are all substances coloured yellow,

1 It must be clearly understood, however, that anthocyanins, apart from the fact that
they are aromatic substances, have very little in common with the respiration pigments.
The latter are formed only after death (unless we believe, with Palladin, that they are
reduced immediately in the living plant), the former only in living plants. It is possible
that the reactions taking place in the formation of the two sets of pigments are upon
similar lines, but there is no reason for thinking that they arise from the same chromogens.

2 Later work, however, has not yet confirmed the origin of any anthocyanin from the
xanthones.

3 See Abderhalden, E., Biochemi&ches Handlexikon, Berlin, 1911, Bd. vi.

4 Perkin, A. G., various papers in Chem. Soc. Trans., 1895-1904.

110

REACTIONS INVOLVED IN

[OH.

Table giving properties and characteristics of the commoner fiavones.

Flavone
Quercetin

Myricetin

Constitutional Formula

Products of
Decomposition
Melting Point from Alkali Melt

Distribution

HO CO

HO CO

Sublimes

- OH above 250
OH

357 C

Phloroglucin Free and as various gluco-

and

protocatechuic
acid

sides in Quercus (bark),
Rhamnus (berries), flowers
of Cheiranthus, Crataegus,
Viola, Prunus, Hibiscus,
leaves of Ailanthus, Rhus,
Arctostaphylos, Calluna, Eu-
calyptus and many others

Phloroglucin Myrica (bark), leaves of Rhus,
and Haematoxylon, Arctostaphy-

gallic acid los

Fisetin

Chrvsin

Above 360 C

275

347

OH

HO CO

Luteolin O

/\
HO

HO CO

Kampherol

OH
OH

327 C

276

OH

Resorcinol

and

protocatechuic
acid

Phloroglucin

and
benzoic acid

Rhus (wood)

Populus (buds)

Phloroglucin Leaves of Apimn, Reseda

and
p-oxybenzoie

acid

Phloroglucin Leaves of Reseda, Genista,

and Digitalis

protocatechuic
acid

Phloroglucin In flowers of Prunus, Del-

and phinium, leaves of Poly-

?;-oxybenzoic gonum, Indigofera, Robinia
acid

HO CO

vii] THE FORMATION OF ANTHOCYANINS 111

and according to Witt the colour is due to the chromophore
group :

o
/\

-c c-

c

c

II

o

in combination with the auxochrome (hydroxyl), -OH, and the intensity
of coloration is said to depend on the position of the hydroxyl groups.

The flavones, as a class, are yellow crystalline substances with high
melting points (see accompanying table). They give either a deeper
yellow, or an orange-yellow, coloration with alkalies, correspondingly
coloured precipitates with lead acetate, and a green or brown coloration
with iron salts. That their distribution in plants is practically universal
can be readily demonstrated by the colour reaction with alkalies.
This reaction is best shown by colourless parts of plants, such as white
flowers. Placed in ammonia vapour, almost any white flower will turn
bright yellow (with the exception of certain albinos of Antirrhinum and
Phlox Drummondii, see pp. 158, 209). The same yellow reaction is given
by green organs, though it can only be detected microscopically on
account of the presence of chlorophyll. On reference to the accom-
panying table, it will be seen that some flavones occur in genera of many
natural orders, while others are limited to one or a few ; this apparent
limitation is probably only due to lack of knowledge of their occurrence
in a number of plants. Further investigation will no doubt show a
very much wider field of distribution for all the flavones.

There is little doubt that the flavones, as well as many other aromatic
substances, are synthesised in the leaves from sugar. The actual
steps of the process would be very difficult to demonstrate. On fusion
with alkalies, the flavones split, as a rule, into phloroglucin and an
oxybenzoic acid. Conversely, they are probably synthesised from oxy-
benzoic acids, or their derivatives, and phloroglucin. As we know,
somewhere in the plant and at some stage of plant metabolism, some
aromatic nucleus must be synthesised, that is an aromatic substance such
as phloroglucin must be synthesised from an aliphatic substance such
as glucose. Putting together what physiological and chemical evidence

112 REACTIONS INVOLVED IN [CH.

there is to hand, it seems most likely that the leaf is the organ where
such a synthesis takes place.

The flavones. moreover, are usually present as glucosides in the plant,
one or more hydroxyl groups being replaced by sugar ; hence the crude
alcohol or water extracts of many plants are only pale yellow (the
auxochrome groups being replaced). On hydrolysis with dilute acids,
the sugar is split off, and the colour deepens ; at the same time a deposit
of flavone is formed, as the free pigment is less soluble than the glucoside.
According to Witt, the capacity for dyeing of the flavones depends
on the presence of free hydroxyl groups. Thus, it comes about that
aqueous or alcoholic extracts of most plants dye but slightly when
boiled with mordanted cloth, but after hydrolysis of the glucoside with
dilute acid, and neutralisation, the same extract dyes more deeply.

It has been suggested by the author (226) that the flavones may,
in many cases, be the chromogens from which anthocyanins are derived.
The reactions involved would then be expressed in very general terms
as follows:

glucoside + water "*~^ chromogen (flavone) + sugar
x (chromogen) + oxygen -* anthocyanin

It is suggested that the first reaction is controlled by a glucoside-
splitting enzyme or enzymes, and the second reaction by an oxidising
enzyme. Also that several of the hydroxyl groups of the flavones, as
they actually exist in the cell-sap, are replaced by sugar. After hydro-
lysis of one or more, but not necessarily all, of these hydroxyl groups,
oxidation of the flavone molecule, accompanied by condensation of
either two flavone molecules, or of a flavone with other aromatic
substances, may take place at these points. Hence the final product
anthocyanin would be itself a glucoside, and the reacting substances
would at all times be in the glucosidal state.

Exception has been taken by Everest (248) to the above hypothesis,
which will in future be referred to as the glucoside hypothesis, and
it is advisable to consider his criticisms at this point, since they do not
affect the general evidence for the hypothesis to be considered later.

Everest makes the following statements which are to some extent
the outcome of his investigations :

1 . No known glucoside of a flavone has more than two hydroxyl groups
replaced by sugar ; most of the glucosides contain only one sugar molecule.

2. All anthocyanin pigments present in the natural state in plants
are glucosides. No free anthocyanidins (non-glucosidal) have been
detected (Willstatter & Everest, 245).

vn] THE FORMATION OF ANTHOCYANINS 113

3. Artificial pigments identical with natural products can be
prepared from flavones. From the flavone glucosides (except in the
case of quercitrin, a glucoside of quercetin) anthocyanins are formed,
and from the flavones themselves, anthocyanidins.

4. When anthocyanins are prepared artificially from flavone gluco-
sides, no anthocyanidin stage is passed through. This may be shown
by conducting the experiment under amyl alcohol, in which the antho-
cyanidins are soluble, should they be formed.

5. When flavone glucosides are hydrolysed, and the flavone
converted artificially into anthocyanidins without removal of sugars,
the latter do not again combine with anthocyanidins to form antho-
cyanins.

On the basis of the above statements Everest maintains that the
glucoside hypothesis is untenable. For (1) precludes the possibility of
the reacting substances being, as a rule, in the glucosidal state throughout
the reaction ; (2) makes it essential for the final product to be a gluco-
side, whereas (5) appears to make it impossible for such a recombination,
i.e. between anthocyanidin and sugar, to take place. Finally (3) and

(4) render hydrolysis unnecessary.

Let us now consider these points. As regards (1), it is by no means
proved that flavones in the living plant have at most two hydroxyls
replaced by sugar. It is quite conceivable that in the living cell more
hydroxyls are replaced, and that only stable glucosides containing one
or two molecules of sugar have been isolated, and this by virtue of their
stability. No definite investigations have been made as to how many
hydroxyls are replaced in the flavone in the plant, and in absence of
further evidence, no conclusive statements can be made. With (2)
the author is in agreement (Wheldale, 244). Again, it is not yet deter-
mined whether the artificial products mentioned in (3) and (4) are
anthocyanins, and the matter is discussed later in the chapter. But,
for the moment, grant them to be so. Then, with regard to statement
(3), that it is not essential for hydrolysis to precede the formation of
anthocyanin, it may be pointed out that nothing is known of the reactions
giving rise to the artificial pigment ; nor is there any reason for supposing
the artificial and natural reactions to be the same. As regards (4),
the glucoside hypothesis is still in agreement with a mode of formation
which does not involve complete hydrolysis at any moment. Criticism

(5) is of no value since it is well known that glucosides are not synthesised
in vitro in the absence of an enzyme. Moreover, in the presence of
an enzyme, synthesis only takes place to a small extent under special

w. p. 8

114 REACTIONS INVOLVED IN [CH.

conditions which are certainly not realised in Everest's experiments.
In this respect, the results obtained in vitro have no bearing upon the
reactions taking place in the plant.

Hence the original hypothesis is not affected by Everest's criticisms.
First, it is quite conceivable that some of the hydroxyls of the chromogen
(flavone) remain unhydrolysed, so that all the products are in the
glucosidal state. Secondly, should this be disproved, there still remains
the alternative, suggested by Everest himself and not disproved by
his experiments in vitro, that the anthocyanidins formed in the plant
immediately recombine with sugar. There is no reason to believe
that the artificial reactions carried out by Everest at all reproduce the
course of events in the living plant.

Apart from the similarity in distribution and reactions of flavones
and anthocyanins, evidence in favour of the glucoside hypothesis may
be collected and arranged under the following headings:

1 . Evidence from results obtained in the cross-breeding of Antirrhinum.
These results will be discussed in greater detail in Part II, but a short
statement is necessary at this point, since the facts recorded are of
value in this connection. The original type of Antirrhinum has magenta
flowers, the colour being due to anthocyanin. During cultivation, two
varieties, among others, have arisen as sports, viz. an ivory, and a white,
both incapable of producing anthocyanin. Ivory, as the name suggests,
is ivory-white in colour, and has on the palate a spot of yellow which
is common to all varieties of Antirrhinum, except white. White is
dead white, without the yellow spot on the palate. When a white 1
is crossed with an ivory, a plant having magenta flowers like the type
is produced, and in these flowers, magenta anthocyanin is present
in the epidermis of the corolla. Hence the original ivory and white
varieties must between them contain the materials for the formation
of anthocyanin. It has been shown by Wheldale & Bassett 2 that
the pigment in the ivory variety is the flavone, apigenin:

HO

CO
which occurs in the plant as a glucoside.

1 Of certain ancestry. See p. 161.

2 Wheldale, M., & Bassett, H. LI., 'The Flower Pigments of Antirrhinum ma jus.
n. The Pale Yellow or Ivory Pigment,' Biochem. Journ., Cambridge, 1913, vn, pp. 441-444.

vn] THE FORMATION OF ANTHOCYANINS 115

No flavone is present in the white variety, but it must nevertheless
contain a factor which, in some way, acts upon the chromogen, apigenin,
with the production of magenta anthocyanin. The magenta flowers
contain apigenin in the inner tissues of the corolla, and anthocyanin
in the epidermis. This anthocyanin, isolated according to methods
given elsewhere and purified from apigenin, gave on combustion the
following percentages as compared with apigenin 1 :

C H O

From anthocyanin 50-50% 5-11% 44-39%

From apigenin 66-66 % 3-70 % 29-64 %

from which it will be seen that anthocyaniu in Antirrhinum is a more
highly oxidised substance than apigenin. Determinations of the
molecular weight of this anthocyanin gave results of the order of 700
showing that the red pigment has a much larger molecule than apigenin,
the molecular weight of the latter being 270. Hence it is possible that
in the formation of anthocyanin, either two or three molecules of flavone
condense with oxidation, or the flavone condenses with some other
aromatic substance present in the plant.

2. Evidence from analogous reactions. It has been suggested above
that sugar is first split off from certain hydroxyl groups of the flavones
(which we know to occur as glucosides), and only then can changes,
such as oxidation and condensation, take place at these points the
hydroxyl groups. Co-ordinated reactions of this kind are known to
be common in plant metabolism, the most familiar case being that of
indigo which may be represented:

C 14 H 17 6 N (indican) + H 2 - C 8 H 7 ON (indoxyl) + C 6 H 12 6
2C 8 H 7 ON + 2 = C 16 H 10 2 N 2 (indigo) + 2H 2

The first reaction is brought about by a glucoside-splitting enzyme,
indimulsin, which hydrolyses the glucoside indican ; the second by an
oxidase which oxidises the colourless indoxyl to the pigment indigo.

The development of a bright red pigment which rapidly appears
when flowers and leaves of Schenckia blumenaviana are placed in chloro-
form vapour has led Molisch 2 to form the opinion that the reactions

1 It should be noted that in preparation, the glucosides of both the anthocyanin and
apigenin are split up, and the pigments obtained for analysis free from sugar. Any
sugar present in the molecule would naturally raise the percentage of oxygen to a con-
siderable extent.

2 Molisch, H., 'Ueber ein neues, einen carminrothen Farbstoff erzeugendes Chromogen
bei Schenckia blumenaviana^ Ber. D. hot. Ges., Berlin, 1901, xix, pp. 149-152.

82

116 REACTIONS INVOLVED IN [CH.

involved are the hydrolysis of a glucoside and subsequent oxidation.
Quite similar pigments have been described by Parkin, Bartlett, Tammes
and others 1 . The reactions taking place in the formation of the
respiration pigments of Palladin are probably of the same nature.
An analogous case too is the development of blue pigment in the white
flowers of an Orchid (Phajus sp.) on the death of the tissues. The change
of colour has been used to determine the death-point of the protoplasm
when subjected to freezing 2 .

3. Evidence from the connection of anthocyanin formation ivith
photosynthesis. It has already been stated in Chapter vi, that a very
definite connection exists between photosynthesis and the appearance
of anthocyanin. In organs, such as green leaves, in which photo-
synthesis takes place to the greatest extent, there is least production
of anthocyanin. In organs, on the other hand, such as flowers, which
are unable to carry on photosynthesis, there is the greatest formation
of pigment. All intermediate conditions may be found between these
two extremes and the same relationship will be found to hold. From
these data, the deduction can be made that the greatest concentration
of sugar arises in tissues where photosynthesis is most active. Thus,
in leaves, there is abundant sugar to combine with the chromogen,
and provided translocation of all synthetic products away from the
leaf is unhindered, the direction of the reaction suggested in the hypo-
thesis will be :

chromogen + sugar - glucoside + water

Under these conditions there would be no tendency for free chromogen
to exist, and consequently no anthocyanin would be formed.

4. Evidence from the results of the blocking of the translocation
current. It has also been stated that injury to the tissues, through

1 Parkin, J., 'On a brilliant Pigment appearing after Injury in species of Jacobinia,'
Ann. Bot., Oxford, 1905, xix, pp. 167-168.

Bartlott, H. H., 'The Purpling Chromogen of a Hawaiian Dioscorea,' U.S. Dept. of
Agriculture, Bureau of Plant Industry, Bull. 264, 1913.

Tammes, T., 'Dipsacan und Dipsacotin, cin neues Chromogen und ein neuer Farbstoff
der Dipsaceae,' Bee. Trav. Bot. Neerl., Nijmegen, 1908, v, pp. 51-90.

Palladin, W., 'Die Bildung roten Pigments an Wundstellen bei Amaryllis vittata,'
Ber. D. hot. Ges., Berlin, 1911, xxix, pp. 132-137.

2 Miiller-Thurgau, H., 'Ueber das Gefrieren und Erfrieren der Pflanzen,' Landw. Jahrb.,
Berlin, 1880, ix, pp. 133-189. Also:

Prillieux, Ed., 'Coloration en bleu des fleurs de quelques orchidees sous 1'influence de
la gelee,' Bui. soc. hot., Paris, 1872, xix, pp. 152-157.

Bommer, J. E., ']tude sur le bleuissement des fleurs du Phajus maculatus Lindl.,'
Butt. soc. hot., Paris, 1873, xx, pp. xxvii-xxxiii.

vn] THE FORMATION OF ANTHOCYANINS 117

which the products of synthesis pass from the leaves, may cause red-
dening in the portion distal to the point of injury. Analyses, made
by Combes (207) of leaves of Spiraea which had turned red owing to
decortication, show that red leaves contain greater amounts of both
glucosides and sugars to the following extent:

Glucosides Sugars

Green leaves ... ... 1-64 2-21

Red leaves 6-15 4-26

and that the increase of glucoside in red leaves is proportionally greater
than the increase of sugar. When we assume also that the chromogen
(flavone) is being continually synthesised in the leaves from sugar (see
p. Ill), then it follows that in red leaves the hydrolysis reaction, owing
to the increased concentration of the glucosides, may take place in
the reverse direction :

glucoside + water -*- chromogen + sugar

Hence more free chromogen and consequently more pigment will be
formed.

5. Evidence from data as to the absorption of oxygen in gaseous exchange
in red and green leaves. By making analyses of gaseous exchange in
red and green leaves respectively, Combes (379) has shown that the
red absorb more oxygen during the reddening process than the normal
leaves. Katie (354) also demonstrated that leaves kept in culture
(sugar) solutions in the absence of oxygen, failed, although they remained
alive, to form any anthocyanin in contrast to similar leaves when the
experiments were conducted in air. Molliard (376) also showed that
oxygen is necessary for the formation of anthocyanin in radishes by
totally submerging the roots in sugar solution, under which conditions
the pigment failed to appear.

That anthocyanin is produced by the action of an oxidase on a
chromogen forms the basis of the hypothesis of Keeble, Armstrong
& Jones (230, 231). It would appear to be evident from the results
obtained by these authors that there is an intimate relationship between
the distribution of oxidase and the development of anthocyanin.
There is every reason to believe, on their evidence, that oxidation
processes would readily take place in those tissues in which anthocyanin
is found. These authors consider that the formation of anthocyanin
is represented by the following equation:

glucoside + enzyme -*- sugar + chromogen
chromogen + enzyme -+- peroxide -- pigment

118 REACTIONS INVOLVED IN [CH.

Their evidence may be summed up under two headings : viz. evidence
from (1) presence of oxidising enzyme, (2) presence of chromogen.

1. Presence of oxidising enzyme.

In order to demonstrate the existence of oxidising enzymes in the
tissues, Keeble & Armstrong have adopted an excellent microchemical
method. The mode of procedure is to place the tissue to be examined
in either a 1 % solution of benzidine in dilute alcohol, or in an equally
dilute solution of a-naphthol and incubate at 37 C. 1 If no direct oxidase
reaction results, the material is removed from the tube and washed
with a dilute solution of hydrogen peroxide. The above method was
employed with great success on petals of Primula sinensis. It was
found that a-naphthol gave a delicate lilac-blue colour with oxidases,
but only detected them in the bundle sheath of the veins, whereas
benzidine gave a brown coloration and detected oxidases both in the
epidermis and in the bundle sheath. The results as regards Primula
sinensis may be summed up as follows :

Flowers from all coloured varieties and recessive whites (see p. 177)
gave a benzidine reaction in the epidermis and an a-naphthol reaction
in the bundle sheath. These tissues being the chief seat of the pigment,
it may be said that the distribution of enzyme and pigment in coloured
varieties is practically coincident. Flowers of dominant white varieties
gave no oxidase reaction. In the case of a blue variety with inhibited
white patches on the corolla segments, there was a more or less corre-
sponding lack of oxidase reaction in the inhibited patches. It was found
that the inhibitor could be removed by treatment with hydrocyanic acid
and other methods, and the above varieties then gave the usual benzidine
and a-naphthol reactions. Flowers of certain flaked (magenta and
white) varieties showed no oxidase reaction in the white parts.

Other genera were also investigated by the same method. In
Sweet William (Dianthus barbatus) flowers, it was found that the
oxidase reaction was entirely proportional to the amount of pigmenta-
tion. Of white varieties, some were found to contain oxidase; white
parts of other varieties were found to contain no oxidase. In Geranium
sanguineum the purple type contains epidermal oxidase, but the white
variety gives no such reaction. In the Sweet Pea (Lathyrus] and
Garden Pea (Pisum), no whites were found which did not contain

1 Benzidine and naphthol are artificial acceptors. Whether the reaction obtained is
direct or indirect depends on the presence of a natural peroxide suitable for acceptors
used. Absence of direct action does not preclude a system peroxide-peroxidase in the
plants.

vn] THE FORMATION OF ANTHOCYANINS 119

oxidase. Hence Keeble & Armstrong hold the view that in those
cases where albinos give the oxidase reaction, albinism is due to lack of
chromogen, but in the case of Geranium sanguineum, Dianthus and
flaked Primula, lack of oxidase causes lack of pigmentation.

The assumption that albinism in Primula, Lathyrus and Pisum is
due to loss of chromogen must thus remain for the present until we have
further experimental data either for or against it. In the author's
experience the only instances of absence of chromogen (as determined
by the flavone reaction) are the albinos of Antirrhinum and Phlox
Drummondii. Flavones can be detected by the intense colour given with
alkalies, and this reaction was given by all albinos of Lathyrus and
Pisum examined. Nevertheless it must be admitted that this is no
direct proof, since there may be several flavones from only one of which
anthocyanin is derived. Keeble & Armstrong's views cannot however
explain the case of Antirrhinum, for we must suppose the ivory variety
to be an albino through lack of enzyme, since it contains the chromogen
(apigenin). Yet all ivory individuals tested by the author have con-
tained peroxidase. A possible explanation for this universal presence
of enzyme in ivories is that several oxidases are involved in pigment
formation. Thus only individuals free from all factors (and these are
extremely rare, as for instance 1 in 1024 when five factors are concerned)
would show no enzyme action. Yet at present there is no evidence
that more than one enzyme is involved. Another possibility is that in
Antirrhinum some third factor of a different nature is essential, con-
ceivably for condensation (see pp. 71, 211, 212).

Consequently, as methods for detection of enzymes and chromogens,
one qualitative reaction is of no more value than another. Tissues
may give the flavone reaction, and yet we cannot be sure that the actual
chromogen of anthocyanin is present. Similarly, we have no absolute
proof from qualitative reactions for peroxidase that we are dealing
with a factor for colour. We may be localising areas where any
oxidative process is carried on (anthocyanin formation among others,
if it is oxidative). Apart, then, from the case of Antirrhinum of which
the chromogen has been isolated, and there is good evidence that only
one exists, little can be gained from speculations as to the presence or
absence of chromogens and oxidising enzymes, without more definite
isolation and analysis. Atkins (233, 246, 258), too, working with Iris
flowers has found that there is no satisfactory evidence for the co-
incidence of anthocyanin formation with the presence of oxidases.
Keeble, Armstrong & Jones (239) have also maintained that the

120 REACTIONS INVOLVED IN [CH.

behaviour of anthocyanin in alcohol further confirms the view that
the pigment is formed by oxidation from a colourless chromogen.
According to these authors when coloured petals of stocks (Matthiola)
are placed in absolute alcohol, some colour is extracted by the alcohol,
but both petals and alcohol become colourless in the course of an hour
or so. If now the colourless petals are removed from alcohol and placed
in water, the colour returns, and more rapidly if the water be hot.
Keeble, Armstrong & Jones explain these phenomena in the following
way. Anthocyanin is formed from a colourless chromogen by oxidation,
the agent being an oxidase which can only act in the presence of water ;
there is present also a reducing agent (which is not an enzyme), and
this reduces anthocyanin to its colourless chromogen when the action
of the oxidase is inhibited by absolute alcohol. The validity of this
explanation has been questioned both by Wheldale & Bassett (621) and
Tswett (252). The two former authors note that an extract of antho-
cyanin made in boiling absolute alcohol (which cannot according to
Keeble & Armstrong contain oxidase) loses its colour on standing.
The colour can be brought back by dilution with water. Hence
obviously no oxidase is necessary. Moreover colour can be restored
to the alcohol solution by dry hydriodic acid gas, a powerful reducing
agent. Colour is also restored to petals, decolorised in alcohol, when
they are placed in boiled water through which a stream of hydrogen
has passed for some time and still continues to pass ; a condition under
which any leuco-compound should be stable. Similar criticisms to
these are also offered by Tswett.

We now turn to the second class of evidence. (2) Presence of
chromogen.

Evidence for the presence or absence of chromogens is given by
Jones (237). According to this author four classes of white flowers
may be distinguished, (a) Those, such as Lychnis coronaria, Anemone
japonica, Chrysanthemum sp. which produce a brown or brownish-red
pigment when subjected to alcohol, chloroform, etc. Such flowers
contain a direct oxidase ; (b) those, such as varieties of Dianthus Caryo-
phyllus and D. barbatus, which give similar pigments only on addition
of hydrogen peroxide and contain peroxidase only; (c) those, such as
white varieties of Plumbago capensis and Swainsonia Tacsonia, which
do not produce a brown pigment even after addition of hydrogen peroxide
but contain a peroxidase, and (d) those, such as varieties of Sweet
William mentioned above under (1) which have no peroxidase and
hence no oxidase. Reference to the account previously given of

vii] THE FORMATION OF ANTHOCYANINS 121

Palladin's 'respiration pigments' at once makes it clear that Jones'
chromogens are identical with these pigments. Class (a) contains some
substance which can act as an organic peroxide, class (6) no such sub-
stance, so that hydrogen peroxide must be added in order to give a
pigment. Class (c) apparently contains no substance which can act
as a respiration pigment. There is no evidence whatever that the
chromogens of the respiration pigments are in any way the chromogens
of anthocyanin, and hence their absence has no bearing on albinism
with regard to anthocyanin (see footnote on p. 109).

More recently Combes (234, 235, 620) has brought forward evidence
which he considers to be in complete contradiction to the oxidation
hypothesis of the formation of anthocyanin. Combes' work may be
summed up under three headings :

1. The isolation of two pigments from leaves of Ampelopsis
hederacea : namely, a brownish-yellow pigment crystallising in rosettes
of needles and having the properties of a flavone, and a purple pigment
also crystallising in rosettes of needles and having the properties of
an anthocyanin.

2. The transformation of the flavone into a purple pigment identical
with anthocyanin by reduction.

3. The transformation of anthocyanin into a flavone by oxidation.
The method employed by Combes of obtaining anthocyanin from

flavone (from Ampelopsis) is to dissolve the flavone in alcohol, acidify
with hydrochloric acid and add sodium amalgam. The solution thus
treated with nascent hydrogen becomes violet-red and increases in
intensity of coloration. After neutralisation and filtration, the solution
obtained gives a purple substance on evaporation. The latter crystal-
lises in needles grouped in rosettes like the natural anthocyanin and
has the same melting point and properties as the natural product.
Combes also obtained crystalline anthocyanin from a crystalline flavone
isolated from leaves of the Privet (Ligustrum vulgare) and from a similar
substance extracted from a variety of Vine which does not redden in
autumn. From these results he concludes that the phenomena observed
for Ampelopsis are not confined to that plant. And, moreover, though
the variety of Vine (Chasselas dore) does not redden in autumn, yet
its leaves produce a substance capable of being reduced to form an
anthocyanin. In addition, he employed flowers of Narcissus incom-
parabilis. The genus Narcissus, as is well known, contains a readily
crystallisable flavone, and from this substance he also prepared an
anthocyanin-like product by reduction.

122 REACTIONS INVOLVED IN [cm.

As a converse to the above experiments, Combes oxidised antho-
cyanin and obtained a flavone. The method employed consisted in
taking Ampelopsis anthocyanin and purifying it by crystallisation,
first from, alcohol, and secondly from water. The pure product is
dissolved in alcohol, and an equal volume of hydrogen peroxide is
added. The purple colour of the anthocyanin gradually disappears,
and a yellow solution is left which deposits crystals of a flavone identical
in properties and melting point with the natural product derived from
the plant. Hence he concludes that the anthocyanin has been oxidised
with the formation of flavone.

A criticism of this last experiment and the deduction therefrom may
be made at this point. It is not mentioned in Combes's paper whether
the crystalline anthocyanin had been tested in order to ascertain if
it were free from flavone. It is the experience of the author (Wheldale,
254) that when a mixture of flavone and anthocyanin obtained from
Antirrhinum is allowed to crystallise from alcohol, both pigments
crystallise out together in plates which, in spite of the fact that they
may consist largely of flavone, are deep red in colour, and very careful
purification is necessary before anthocyanin can be obtained entirely
free from flavone. When pure anthocyanin from Antirrhinum is treated
with hydrogen peroxide, the pigment is destroyed but no flavone is
found. Hence it is possible that the flavone derived from Combes's
anthocyanin may have been present as impurity. If, however, this
were not the case, it still may be possible that the action of the hydrogen
peroxide is not necessarily one of oxidation.

Before making any criticism upon the question of the action of
sodium amalgam, the work of Keeble, Armstrong & Jones, Tswett
and Everest on similar lines, must be considered.

Keeble, Armstrong & Jones (240) have obtained red pigments,
giving, in some cases, the anthocyanin reaction by treating alcoholic
extracts of a number of plants with zinc dust and hydrochloric acid.
Under these conditions, extracts from the following plants gave red
pigments: pale yellow 'primrose' variety of Cheiranthus, yellow
Daffodil, yellow Crocus, cream Polyanthus, and the dominant white
variety, but not recessive white variety, of flowers of the Chinese
Primrose (Primula sinensis). The red pigment which is first formed
will, after continued treatment with nascent hydrogen, become colourless,
the colour returning on exposure to air. This phenomenon led the
authors to conclude that a preliminary reduction, followed by oxidation,
is the sequence of events.

vn] THE FORMATION OF ANTHOCYANINS 123

Tswett (243) also found that an 'artificial anthocyanin ' could be
prepared from apple juice by the action of strong mineral acids in the
presence of formaldehyde or acetic aldehyde. The artificial anthocyanin
had properties and reactions very like the natural pigment, but was
soluble in ether, whereas natural anthocyanins are insoluble in this
solvent. Tswett produced similar pigments, though he did not study
them in detail, from extracts of white grapes, bananas, flesh of purple
grapes, white petals of Rosa and Cyclamen. He failed however to get
any formation of pigment on treatment of the following: leaves of
white Cabbage, mesophyll of red Cabbage, leaves of Pelargonium, Orange
skins, petals of white Pinks, white petals of buds of red Pinks, flowers of
Lily of the Valley, Carrots, Potatoes, Kohlrabi and Barley seedlings.

Artificial anthocyanin resembles natural anthocyanin in the following
respects. It is soluble in alcohol, gives a green reaction with alkalies,
red with acids and a green precipitate with lead acetate ; it is decolorised
by sodium bisulphite, but the colour returns again on acidifying with
sulphuric acid. It differs from the natural product in its solubility in
ether and also its ready solubility in water, the natural pigments being
only slightly soluble in water, except in the condition of glucosides.
The work of Everest (248, 249) in this direction has been more
elaborate. Mention has been previously made of the fact that Everest
has prepared from the flavones, substances which give some of the
qualitative reactions of anthocyanin. Everest's methods are as follows :
He reduces an alcoholic solution of flavones, to which is added 2 N acid
solutions, by means of nascent hydrogen formed by (1) addition of
granulated zinc, (2) sodium amalgam, (3) electrolysis, using sulphuric
acid and lead electrodes, and (4) magnesium ribbon. The materials
employed were cjuercetin, quercitrin and extracts from various flowers
and tissues. In many cases a red pigment is obtained as reduction
proceeds. When quercetin was employed, the product could be
extracted quantitatively from the acid solution by amyl alcohol showing
it to be an anthocyanidin. When quercitrin, a monoglucoside of quer-
cetin, was used, anthocyanidin, contrary to expectation, was also
obtained. Extracts, however, from flowers and other tissues, in which
the flavones are presumably in the glucosidal state, gave anthocyanins
in the cold and anthocyanidins on hydrolysis.

Everest also lays stress upon the fact first noted by Keeble, Arm-
strong & Jones (240), that the artificial anthocyanin is reduced to a
colourless substance by vigorous reduction with nascent hydrogen.
This is also the case with the natural pigment (see p. 55). This similarity

124

REACTIONS INVOLVED IN

[CH.

in behaviour, together with the solubility in amyl alcohol and the
green reaction with alkalies, he considers to be a complete proof of the
identity of the two pigments.

The author, in conjunction with Bassett (255), has prepared red
products from both quercetin and apigenin. In the case of apigenin,
the substance was analysed and found to be a reduction product, but
its reaction with alkalies in no way resembled that of anthocyanin. On
the other hand, there is every reason to believe that the anthocyanin
of Antirrhinum may be derived from apigenin by oxidation and con-
densation. Hence, in the only instance where both artificial and
natural anthocyanins have been obtained from a flavone and have
been analysed, they appear to be by no means identical.

More recently Willstatter (257) has investigated the problem of the
artificial formation of anthocyanin, and maintains that two substances
are produced when quercetin is reduced with nascent hydrogen. In
his experiments, an alcoholic solution of quercetin acidified with hydro-
chloric acid is reduced with sodium amalgam or magnesium. A purplish-
red product is obtained, and the bulk of this substance he terms allo-
cyanidin ; on the basis of analysis (by combustion) of the product he
suggests the following constitution and reactions:

O OH OH/H

HO/Y\ <^ ^QH HOf"

V/X/OH

H <r +2 H 8 ^

O

Quercetin

Allocyanidin forms a crystalline compound with hydrochloric acid,
but is unstable on heating with the dilute acid. When, however, the
crude solution obtained from the reduction of quercetin is diluted and
heated, it does not entirely lose colour owing to the presence, in Will-
statter's opinion, of a small amount of a second product. Isolated and
analysed, this latter product was found to be identical with cyanidin
chloride, and it is suggested that the reaction takes place as follows :

OCl

Allocyanidin

Quercetin

Cyanidin chloride

viz] THE FORMATION OF ANTHOCYANINS 125

From 33 gnis. of quercetin, 0-165 gm. of the product was obtained.
It seems to be open to criticism to assume that this product is formed
by reduction ; it is also doubtful whether the analysis (by combustion)
forms an efficient proof of its identity with cyanidin, though its resem-
blance in chemical and physical properties certainly lends support to
the statement. Granted, however, that cyanidin is formed from quer-
cetin by reduction, we still have no proof that this is the process employed
by the living plant.

The contents of the present chapter may be summarised as follows.
There is strong evidence for the belief that anthocyanins are formed from
the flavones, and that these substances are the chromogens of the
Mendelian writers. But the actual processes involved are still debatable.
There is evidence, from Keeble & Armstrong's work, that the presence
of anthocyanin is often connected with the presence of oxidases, but there
is no proof that oxidases are the agents in transforming flavones into
anthocyanins. From the analyses of pigments of Antirrhinum, it appears
that the anthocyanins of that genus are more oxidised than the flavone,
apigenin, which is present in the plant, and the artificial product
formed by reduction from apigenin is moreover not identical with a
natural anthocyanin. On the other hand, there is the evidence from
Willstatter's experiments that quercetin on reduction will produce, as
a by-product, a substance identical with a natural anthocyanin. To
obtain a satisfactory solution to the problem, further experimental
work is necessary, especially on the lines of isolation and analysis of
other plant anthocyanins and the flavones which accompany them 1 .

1 See also Appendix.

CHAPTER VIII

THE SIGNIFICANCE OF ANTHOCYANINS

Biological significance.

The function of anthocyanins in rendering the corolla, perianth,
bracts and other parts of the inflorescence attractive to insects may be
regarded as their biological significance ; so also the function of making
ripe fruits conspicuous to birds. Neither of these subjects is dealt with
in the present volume, but references are given to the work of Knuth
(426), Focke (423) and others, from which further information can be
obtained.

Physiological significance.

Several physiological functions have been assigned to anthocyanins
but, broadly speaking, they fall into three classes. Though each hypo-
thesis has been closely criticised by the various investigators of the
problem, the final issue is far from being complete and satisfactory.
The three main ideas are :

1. That of shielding the chloroplastids from too intense insolation;
this is known as the 'light-screen' hypothesis. In slightly different
forms it has been advocated by Kerner, Kny, Wiesner, Keeble and
Ewart.

2. That of assisting the action of diastase by screening it from
deleterious rays, and thereby facilitating the hydrolysis of starch and
subsequently translocation. This view has been mainly supported by
Pick, and Koning & Heinsius.

3. That of absorbing certain light rays and converting them into
heat. Though the conception of anthocyanin as a medium for raising
the temperature of tissues was due to earlier writers, notably Comes
and Kerner, yet it was Stahl who first regarded this property as a valuable
asset in assisting transpiration; he also considered other processes of

CH. vni] THE SIGNIFICANCE OF ANTHOCYANINS 127

general metabolism, including translocation and fertilisation, to be
furthered by the warming effect of the pigment.

The 'light-screen' hypothesis had no doubt a basis in the work of
Pringsheim. About 1880 Pringsheim 1 published the results of various
experiments on carbon assimilation, and among these, the effect of
light of different colours and intensities on chlorophyll. He found that
white light of a high intensity decolorised chlorophyll, but in red light
the effect was suppressed. In yellow, green, or blue light, on the
other hand, it is easy to decolorise and kill the cells of many Algae,
Characeae, Musci, Filices and Phanerogams. Thus in white light,
the result was attained in two or three minutes, in green and blue in
five minutes, whereas in red light of the same intensity and after twice
or four times as long exposure, no changes took place. Hence he sup-
poses the red rays to be photochemically inactive, or only very slightly
active, on plant cells. That the injurious effects were not due to heat
is shown by the fact that in green light the temperature did not rise
sufficiently to be harmful to plant cells. For instance even in red light,
the temperature of the water drop in which the tissue was placed,
was raised to over 45 C. within five minutes, and yet after 15-20
minutes' exposure, there was neither decolorisation nor death of the
cell. On the other hand, in green and blue light, cells were decolorised
and killed in five minutes, although after 15-20 minutes' exposure,
the water scarcely reached 35-36, a temperature quite harmless to the
cells. The solutions used were for red light, iodine in carbon bisulphide,
for yellow, potassium bichromate, for green, copper chloride, and for
blue, ammoniacal copper sulphate.

It would seem, however, that Pringsheim may have overrated the
destroying effect of sunlight on chlorophyll, for Reinke 2 , in a paper
published in 1883, mentions some results he obtained by exposing both
the green and red (anthocyanin) parts of plants to light intensified by
a lens, the heat rays being cut off by a screen of alum solution. He
found that the chlorophyll in leaves of Elodea and Impatiens and the
red pigment in the petals of Papaver and Rosa were not decolorised
unless the light intensity became 800-1000 times greater than normal

1 Pringsheim, N., 'Ueber Lichtwirkung und Chlorophyll! unction in der Pflanze,'
Jahrb. wiss. Bot., Leipzig, 1879-1881, xn; 1882, xm. Also ' Pringsheim's Researches
on Chlorophyll,' translated and condensed by Bayley Balfour, Q. /. Hicrosc. Sci., London,
1882, xxn, pp. 76-112, 113-135.

2 Reinke, J., ' Untersuchungen iiber die Einwirkung des Lichtes auf die Sauerstoff-
ausscheidung der Pflanzen,' Bot. Ztg., Leipzig, 1883, XLI, pp. 697-707, 713-723, 732-738.

128 THE SIGNIFICANCE OF ANTHOCYANINS [CH.

sunlight. When in light of only 200 times the intensity, there was no
bleaching of Elodea even after two hours' exposure.

Kerner (398) appears to be one of the first, if not the first, to make
suggestions as to the significance of anthocyanin, though in his Pflanzen-
leben (1885) 1 , he by no means confines himself to one use of antho-
cyanin to explain its appearance in different places under various
conditions. Thus, for instance, in stems and petioles, he says, it may
be of use in keeping back light rays which would decompose the travelling
materials. (This view may be original or may be taken from Pick,
see p. 131.) In other cases it is produced temporarily for the same
purpose when there is a transmission of metabolic substances on a
large scale as in the seedlings of certain starchy seeds (polygonums,
oraches, palms and grasses). Again, in spring, it appears in young leaves
and shoots when supplies are travelling to them from their place of
storage in the stem. In autumnal leaves, all materials of use to the
plant are passing out for storage ; hence the value of autumnal colora-
tion. Anthocyanin developed on the under surfaces of leaves, is, on
the contrary, not protective but absorbs light and changes it into heat
which is serviceable for growth, metabolism and translocation. When,
in shrubs and herbs, anthocyanin appears on the under leaf-surfaces,
this is only the case in the lowest leaves near the ground ; the upper
leaves remain green and so do not prevent light from passing through
to the leaves beneath them. The development on the under surface
of leaves of marsh plants and floating plants has similar uses. Thus,
there is not only retention by anthocyanin of rays injurious to meta-
bolism, but there is, at the same time, a useful transformation into heat.
When, he continues, the surrounding temperature is low, plants are
often entirely reddened, as in some small annuals developing early
in the spring (Saxifraga, Hutchinsia, Androsace) ; also in seedlings
germinating at low temperatures. It is, for the same reason, common
in the vegetative parts of many High Alpine plants. Of similar value
is its distribution on the under surface of petals and perianth leaves
of flowers which close by night.

Two further observations of interest are mentioned by Kerner.
One, that plants which have leaves covered with a felt or wool of hairs
rarely develop anthocyanin, since such leaves do not need so much
protection from light.

The second observation is that made on Satureja hortensis, a plant

1 The Pflanzenleben appears to have been issued in parts several years before the
6rst edition in 1888.

vin] THE SIGNIFICANCE OF ANTHOCYANINS 129

growing wild in the Mediterranean region and in cultivation in gardens
where it is known as the Summer Savory. In the shade it is green;
exposed to the sun its stems and leaves are dark violet. Kerner sowed
the seeds of this plant in his Alpine garden, 2195 metres above sea-level,
in the Tyrol. There, as a result, it is supposed, of the intensity of
the sun's rays, it produced anthocyanin in extraordinary abundance.
Such an adaptation, as he points out, can only take place in plants which
are able to form the pigment. Seeds of Linum usitatissimum were
sown next to the Savory ; the seedlings turned yellow and failed to
survive owing to the fact, in Kerner's opinion, that they are unable
to form anthocyanin to protect the chlorophyll and are also without
any development of hairs.

In 1885 Reinke 1 published further results which, if accepted, make
it difficult to uphold the 'light-screen' hypothesis. The problem he set
out to investigate was the effect of the different parts of the spectrum
on the destruction of chlorophyll, and he claims to have shown that
the red rays have the maximum destructive effect, and that this power
decreases in the other parts in the following order orange, violet,
yellow, blue, dark red and green : when this sequence is compared
with the absorption spectrum of chlorophyll, it will be seen that the
rays which are most absorbed by chlorophyll have the greatest destruc-
tive effect upon it, and those which are least absorbed, do not have
any effect upon it.

Further support was given by Hassack (393)' in 1886 to the screen
hypothesis, though his work is chiefly confined to the histological
distribution of anthocyanin in leaves.

In 1887 a paper appeared by Engelmann (394) which has considerable
bearing on the problem. By means of a microspectral photometer,
Engelmann investigated the spectra of the pigments in red leaves,
and found that the absorption of red pigment is, on the whole, comple-
mentary to that of chlorophyll. Hence we are confronted with this
dilemma. Anthocyanin absorbs those rays which are not absorbed
by chlorophyll; the rays least absorbed by chlorophyll are, according
to Reinke, the least harmful to chlorophyll. Therefore anthocyanin
absorbs the rays which are least, and not most, harmful to chlorophyll.
How then can it be a protective screen ? If we accept Reinke's results,
a green screen would be the best protection as it would absorb the rays

1 Reinke, J., 'Die Zerstorung von Chlorophylllosungen durch das Licht und eine
neue Methode zur Erzeugung des Normalspectrums,' Bot. Ztg., Leipzig, 1885, XLUI,
pp. 65-70, 81-89, 97-101, 113-117, 129-137.

w P. 9

130 THE SIGNIFICANCE OF ANTHOCYANINS [CH.

which are most absorbed by chlorophyll and are, at the same time,
most injurious. If, however, we accept Pringsheim's results that
the green portion of the spectrum is most harmful, then it is possible
to consider the question as to whether anthocyanin is a screen
or not.

The matter was again brought up by Kny (397) in 1892. To test
the efficiency of anthocyanin as a light screen, Kny placed a solution
of chlorophyll behind a parallel-walled glass vessel which was filled in
one case with an extract of red, in another with an extract of white,
Beet-root. Behind the red extract, the chlorophyll retained its colour
much longer than behind the white. This experiment is of little value
since chlorophyll in a dull light (such as would be the case behind the
red solution) invariably retains its colour much longer than in a bright
light.

Again, in 1894, Wiesner (400) supported the screen theory for the
protection of chlorophyll, especially in young leaves.

In 1895 Keeble (403) published a paper of some significance on the
hanging foliage of tropical trees. The characteristic pendent position
and red coloration of young leaves of some of these trees has been
mentioned in Chapter u. Keeble is of the opinion that, though the
development of red pigment is not universal among them, yet it may
serve as a protection against strong insolation, not only as a screen
but also by protecting the leaf from too great heating. This view is
based on an experiment carried out with red and green leaflets of
Amherstia nobilis. These were laid side by side in the sun, and the
temperature taken by thermometers placed upon the upper surfaces
of the leaves; the thermometers were then placed under the leaves
and the temperature again taken. His results and conclusions are
expressed as follows: "...when young thin red and older tougher
green leaves of Amherstia are exposed side by side to the direct rays of
the sun, the temperature, as registered by the thermometer, is higher
to the extent of 1 C. at the upper surface of the red leaves : conversely,
of the temperatures registered at the lower surfaces of the leaves (i.e.
behind them), that beneath the green is higher than that beneath the
red, by a similar amount.

"Put in general terms, the surface-layers of the red leaf reflect
more heat than those of the green leaf. The green leaf, conversely,
absorbs more of the sun's thermal rays than the red. Now, the two
sets of leaves differ in two respects : first, in colour ; then in that the
green has a thicker, more developed cuticle (and mesophyll) than the

vm] THE SIGNIFICANCE OF ANTHOCYANINS 131

red. This might, therefore, be expected to oppose a more solid resistance
to the heat-rays than the thin immature cuticle of the red leaf. That
such is not the case must be regarded as due to the fact that the different
colouring-matters have different powers of reflection and absorption
of heat, and that this difference is of such a nature that the red colouring-
matter more effectually cuts off heat-rays from the body of the leaf
than does the green. That is to say, the red colouring-matter
acts as a screen by which the thermal effects of the sun's rays are
moderated." In conclusion Keeble says: "In addition, then, to its
value as a screen, the red or reddish coloured saps of such trees as
Amherstia nobilis have the capability of affording to the young leaf,
if necessary, a protection against too great heating effects of the rays
of a tropical sun."

Later, Keeble's conclusions have been adversely criticised by both
Stahl (405) and Smith (420) : the latter remarks, alluding to Keeble :
"From his somewhat rough experiments he drew the novel conclusion,
which... is exactly the opposite of the conclusion logically to be drawn
from his observations, that the red colour is a protection against too
great heating up of the leaf."

In 1895 a long paper appeared by Ewart (401) on ' Assimilatory
Inhibition in Plants.' Ewart maintains that leaves, especially of shade
plants, when exposed for a long time to strong illumination, suffer from
inhibition of photosynthetic power. He considers the development
of authocyanin, on exposure to light, to be an adaptation for protection
against this light rigor, rather than a protection against the destruction
of chlorophyll. We shall return later to further criticisms by
Ewart.

We will now pass on to another suggestion as to the uses of antho-
cyanin, though it was the first in point of time. It is that made, in
1883, by Pick (391), who maintained that red light increases the rate
of hydrolysis and trauslocation of starch, while it allows photosynthesis
to go on unhindered. Hence its frequent occurrence in sterns and
petioles ; also in young leaves where there is active metabolism and in
autumnal leaves from which there is a migration of substances which
may be useful for storage. Pick brings forward certain evidence,
namely that in red leaves the palisade parenchyma contains less starch
than the spongy, on account of the increased translocation under the
influence of the red light. He, moreover, gives the results of an experi-
ment with a view to proving the same point. Of the lobes of a large
leaf of Ricinus, one was illuminated by light coming through orange

92

132 THE SIGNIFICANCE OF ANTHOCYANINS [CH.

glass, a second by light through ruby glass, a third by light passing
through anthocyanin solution (of beet-root), while a fourth remained
in bright insolation. All the lobes were illuminated thus for four hours.
He then found that the insolated lobes had more starch in the palisade
parenchyma, though there was a considerable amount in the spongy
tissue too, whereas the lobes covered by red glass and anthocyanin
solution had more starch in the spongy parenchyma ; of the lobe
covered by orange glass, there was nothing in particular to note. From
these results Pick concludes that the lobes illuminated by ruby glass
and anthocyanin solution had assimilated as much as the directly
insolated lobe, but translocation had been more rapid.

Pick's conclusions were adversely criticised by Wortmann (619)
who maintained that the observations on starch distribution in red
leaves are valueless because, first, it was not noted whether green
leaves behave differently under similar circumstances, and secondly,
the effect of red light on diastase activity outside the cell should
be investigated also. As regards the experiment on the Ricinus leaf,
he points out that, according to Stahl, the palisade cells and chloro-
plastids are adapted to starch formation in bright light, while those
of the spongy parenchyma are adapted to diffuse light. Hence the
distribution of starch in the lobes exposed to red light may be explained
on the ground that they receive less light than the directly insolated
leaf. Since Wortmann's criticisms have not been refuted, the question
still remains unsolved. Later, Ewart (406) also criticises Pick's
deductions. Pick has stated at one point that young red leaves have
little starch because the translocation is furthered, but as the leaves
mature, and the pigment disappears, more starch develops. Ewart
makes what appears to be a more natural suggestion, namely that
photosynthetic activity increases as the leaf matures and this is so
even if the leaf should retain its red colour.

More recently Koning & Heinsius (416) have again reopened the
question as to whether anthocyanin acts as a screen for diastatic activity.
These authors quote the results of Brown & Morris to the effect that
the diastase content of leaves decreases after a period of exposure to
bright light, and that the enzyme is chiefly destroyed by the violet
and ultra violet rays. Koning & Heinsius claim to have shown experi-
mentally that the above-mentioned rays are absorbed, not only by a
water extract of anthocyanin, but also by the pigment in the living
leaves; also, by placing branches of Quercus rubra and other species
under double-walled vessels filled with anthocyanin solution, that it

vm] THE SIGNIFICANCE OF ANTHOCYANINS 133

is these same rays which cause the formation of anthocyanin. Finally,
it was shown that red leaves always contain more diastase than green.
From these results the authors conclude that anthocyanin has a
protective action on diastase.

The third theory which has been so largely supported by Stahl
(405) is based on the power shown by anthocyanin of converting light
rays into heat. Kerner, as previously mentioned, had the idea of this
function of anthocyanin, but Stahl laid stress on a special aspect, viz.
that of accelerating transpiration in red leaves.

Stahl, however, was by no means the first to suggest a connection
between light absorption and transpiration. As early as 1879-80,
Comes (388, 389, 390) published the results of work on the effect of
light on transpiration. He used pot plants which were enclosed in a
zinc case with a glass front, and the loss of water was determined by
weighing. The plants were also exposed to differently coloured light
obtained by using potassium bichromate, ammoniacal copper, and
alcoholic chlorophyll solutions. The transpiration of various coloured
corollas was observed, the absorption bands of the pigments being also
ascertained. From his experiments Comes arrives at the following
conclusions. Transpiration is affected by light as well as by physical
conditions; other things being equal, a plant transpires more in the
light than in the dark, and the effect of light is proportional to its
intensity. Light favours transpiration only in so far as it is absorbed
by the colouring matters, and thus, that organ transpires most which
is most highly coloured. (This point was demonstrated by the coloured
corollas; those of which the pigment showed absorption bands in the
greatest number, breadth and intensity, transpired most strongly.)
Moreover, the rays which are most favourable for transpiration in a
coloured organ are those which are most absorbed by it. Hence the
transpiration of an organ is slightest under the influence of light which
is of the same colour as the organ itself, and strongest when under the
rays of the complementary colour. We have here then in Comes'
results the basis of Stahl's hypothesis, that the red pigment of leaves
brings about additional absorption and heating of the leaf, and conse-
quently greater transpiration.

In addition, Kny (397) claims to have shown, by means of a simple
experiment, that red leaves do attain a higher temperature than green.
Kny filled tw T o parallel-sided vessels with green and red leaves respec-
tively of the same plant, and then filled up the vessels with water.
When these were exposed to light, the heat rays being cut off by alum

134 THE SIGNIFICANCE OF ANTHOCYANINS [CH.

solution, the temperature in the vessel containing red leaves rose higher
than that in the one containing green.

Stahl's (405) experiments and his criticism of the work previously
done in this direction are all included in his paper 'Ueber bunte Laub-
blatter.' An account is first given of his experimental work on the
method of finding the respective temperatures of red and green leaves.
The apparatus used was a thermo- junction with spathulate electrodes
which could be buried in the substance of the leaf. The source of light
was a gas burner, and sometimes a Leslie cube was used by which dark
heat only was obtained. Owing to the size of the electrodes, thick
and fleshy leaves were most suitable for use. One of the species
employed was an epiphytic orchid, Sarcanthus rostratus. On taking
the temperature of both green and red leaves, it was found that the
red had a temperature of 1-5 above the green in one experiment, in
another of 1-82. Similar results were obtained with the red and green
parts of one and the same leaf of Sempervivum tectorum. Though more
difficult to manipulate, some thin-leaved plants were used with similar
results; for example, a difference of temperature of 1-35 for Begonia
heracleifolia var. nigricans, 0-22 for Pelargonium peltatum and 0-14 for
Tulipa Greigi. A rise in temperature of red leaves over green was
also obtained when the Leslie cube was used. In addition, other
investigations were made using coco-butter instead of the thermopile.
The melting-point of the butter was raised by mixture with beeswax,
and was then spread in a liquid condition as thinly as possible on the
under surface of the leaf. When the butter was set, the upper leaf-
surface was exposed either to the sun's rays or to the heat of the Leslie
cube. The butter was found to melt more rapidly, and to a greater
degree, on red leaves, or parts of leaves, than on green.

Stahl then proceeded to apply these results to cases which have
been discussed by previous investigators. With regard to Kerner's
experiments on Linum and Satureja in the Tyrol, which we have already
noted, Stahl suggests that it is not from lack of protection of chlorophyll
that the flax suffers, but from the low night temperature, against which
it would be protected were it able to form anthocyanin with the resultant
raising of temperature. Stahl proposes that the additional experiment
should be conducted of covering up the Linum by night. If then the
plant is still unable to survive, Kerner's view is more feasible. In
general, also, with Alpine vegetation, Stahl is of the opinion that the
red leaves and stems take a higher temperature than the green, and
thus in a wider and more general sense than Pick he says "in dem

vm] THE SIGNIFICANCE OF ANTHOCYANINS 135

warmeabsorbierenden Blattrot besitzt die Pflanze ein Mittel, die Stoff-
und Kraftwechselprocesse zu beschleunigen." Stahl suggests that at such
low night temperature, Linum is unable to translocate starch and thus
the plant becomes 'starch-sick/ and synthetic reactions cannot proceed ;
but he admits that the experiments require further attention. He has
himself made observations upon cultures of Linum and Satureja at
Pontresina and found, after a night with temperature about C., in
the morning the Linum leaves were still full of starch, whereas Satureja
leaves were starch-free, although exposed during the previous day to
intense sunlight. Stahl's view on this function of anthocyanin was
also applied to autumnal leaves.

Again, the reddening of stigmas of anemophilous flowers is considered
by Stahl to be another illustration of the value of the heating properties.
As examples of trees and plants possessing such a characteristic, he
quotes species of Populus, Salix, Platanus, Ulmus, Ostrya, Carpinus,
Corylus, Alnus, Acer, Fraxinus, Poterium sanguisorba and Rumex
scutatus, and of these, the trees, he points out, flower early in the year.
Hence the presence of anthocyanin is a special adaptation for raising
the temperature and thereby furthering the growth of the pollen-tube.

Stahl next deals with his own suggestion as to a special significance
of anthocyanin which it possesses by virtue of its temperature-raising
properties. He maintains that anthocyanin is frequently found in
leaves of water-plants inhabiting marshy places ; also to a considerable
extent in the leaves of shade-loving plants in damp tropical regions.
The presence of anthocyanin in these leaves leads to a rise of tempera-
ture, and thereby accelerates transpiration which is rendered difficult
by the conditions of such habitats. He quotes as examples the red
under surfaces of leaves of Nymphaea, Villarsia and the frond of
Lemna ; of marsh plants, Orchis maculata, 0. latifolia, Ranunculus acris ;
of shady wood plants, Arum maculatum, Phyteuma spicatum, Hypochaeris
maculata. These are, however, insignificant as compared with tropical
plants of which he quotes many examples from Borneo, Java and Mexico,
of the orders Begoniaceae, Orchidaceae, Acanthaceae, Gesneriaceae,
Marantaceae, Araceae and Melastomaceae ; Stahl even goes so far as to
say that in the case of leaves which develop blotches of anthocyanin,
the lower leaves near the moist ground are more strongly marked than
the upper leaves which may not be marked at all (PolygonumPersicaria}.
Also in Java he has noted that shaded pitchers of Nepenthes lying on
the ground amongst ferns and mosses were deep red, while those borne
high up above the vegetation had only a slight reddening.

136 THE SIGNIFICANCE OF ANTHOCYANINS [CH.

One item of special interest he also notes, namely, that red blotches
on leaves give off less water (tested by cobalt paper method) than
green ; this he found to be due to the fact that the red parts contained
fewer stomata. Stahl, however,, seems to think this compatible with
his views, as too greatly increased transpiration might otherwise occur.
Branches of the red-leaved Beech and Hazel he finds to transpire more
strongly than branches of the ordinary species when both were placed
in the shade, but in the sun and a dry atmosphere, the opposite was the
case.

As regards the coloration of young leaves, Stahl considers that
reddening in temperate regions assists metabolic processes at low
temperatures. In the tropics, on the other hand, the most intense
colouring, he says, frequently occurs in the densest shade forests, and is
then an adaptation for promoting transpiration by heating up of the leaf ;
a view completely opposed to that held by Keeble (403).

Ewart's (406) work and criticisms are so much connected with both
the screen hypothesis and that of Stahl, that it will be more convenient
to consider the arguments in the order which he follows in his papers.
He first concludes, from a number of observations and experiments,
that strong insolation may bring about inhibition of photosynthesis,
and hence too bright light may be injurious in this way rather than
in the destruction of chlorophyll, and this is especially so in the case of
shade plants. He is of the opinion, moreover, that anthocyanin does
act as a protection in such cases, for he apparently trusts Pringsheim's
observations, in preference to Reinke's, that concentrated blue or green
sunlight will kill and bleach chlorophyll grains in five minutes, whereas
concentrated red rays (the heat rays being eliminated) will only
cause bleaching after twenty minutes. Anthocyanin absorbs 70-90 %
of the green and 50 % of the blue, that is just those rays which are most
harmful to chloroplast activity, while rejecting those useful for photo-
synthesis. If, according to Stahl's view, its function were tha.t of a
heat absorber it would, instead of absorbing the green and yellow rays
(of which the heating effect is comparatively slight), show a marked
absorption of the dark heat rays, and this is not the case. But he
admits that Stahl has obtained a rise in temperature of one or two
degrees in red leaves by exposure to dark heat rays. This higher tem-
perature, he considers, to be a disadvantage, and since the red leaves
might thus be liable to transpire too much, excess of transpiration is
prevented by development of fewer stomata. Stahl has, as we have
seen, shown that red leaves, or areas, have fewer stomata than green,

viii] THE SIGNIFICANCE OF ANTHOCYANINS 137

and that green leaves also give off more vapour in sunlight and dry air
than red ones.

The development of anthocyanin in young leaves and many tropical
shade-loving plants is not considered by Ewart to be an adaptation
for promoting transpiration. Its purpose, on the other hand, is to
guard against assimilatory inhibition to which shade plants may be
especially liable if accidentally exposed to intense light. He points out
that in trees and shrubs with young red foliage, individuals growing
in shady positions are less red than those in the sun, whereas, according
to Stahl, it should be the reverse. Further, though it is true that
certain plants growing in the shade form a considerable amount of
pigment, yet of these, again, individuals growing in the more exposed
positions have more colour than those in the deepest shade. Ewart
infers that such plants are extremely sensitive to, and are injured by,
light of marked intensity. He further points out that in many plants
which have anthocyanin on the under surfaces, the ventral surface of
the young leaves is exposed to light ; in one variety of Musa, for instance,
anthocyanin develops on the under surface when the young leaves are
vertical and rolled up, but it disappears as the leaf unrolls and expands
horizontally. In Uncaria sclerophylla the young leaves are so folded
that the ventral surface is most exposed and this develops anthocyanin
which disappears as the leaf expands. Several other instances are
also quoted, one of the most interesting being Mimosa pudica, in which
red coloration is developed on the parts of the under surfaces of the
leaflets which are exposed when folded. Ewart, however, does admit
that, in the case of certain Begonias which have horizontally expanded
leaves with red under surfaces, Stahl's view of the function of antho-
cyanin may be the correct, one, and there are other cases where it is
difficult, he says, to find any use at all for the pigment.

That the distribution of anthocyanin does not conform to Stahl's
hypothesis is emphasised in the following passages from E wart's paper.
"If the primary function of the red dye in the tropics were to increase
the amount of transpiration, then it would be only natural to expect
that it would be formed in greatest abundance where the temperature is
lower and the air more nearly saturated with water-vapour. The very
opposite is however the case. Thus at the foot and sides of the volcanic
mountain of Gedeh, and in the valleys around, very many plants
have a reddish colour, especially in the young leaves. As one ascends
this becomes less marked, until at Tjibodas and in the forests above it
(4500 ft. to 6500 ft.), the number of plants showing a red colouration,

138 THE SIGNIFICANCE OF ANTHOCYANINS [CH.

and the intensity of the latter when present, reach a minimum.
The vegetation at this elevation is almost entirely green, a few plants
only,, especially if growing in open clefts or glades in the forest, having
more or less reddish young leaves. Yet it is just here where a power
of stimulating transpiration is apparently most needed; for at this
elevation the air is, during the greater part of the time, at or near
saturation-point. On the other hand if the red pigment acts as a pro-
tection against sunlight, it is easy to understand why here, where the
sun rarely shines for more than a few hours daily and then generally
through a haze of clouds, the protective red pigment should almost
entirely disappear; for it is just the more refrangible photochemical
rays which the air saturated with water-vapour absorbs in greatest
amount."

And again, " In Java at the commencement of the wet S.W. monsoon
and in Ceylon at the rainy commencement of both monsoons, the
vegetation acquires a more marked reddish tinge than the dry periods
between the monsoons. This is, however, simply due to the fact that
the young foliage, which in most tropical plants is more or less tinged
with red, is very much more abundantly formed at this period than
during the dry season. Even during the wet season in West Java,
there is almost always bright sunlight until mid-day, lasting often till
3 or 4 p.m., and occasionally all day; so that the young foliage which
the rain has caused to be produced in such abundance is exposed for
six hours on the average to very bright illumination, the sunlight from
9-12 being the brightest of the day. Hence the protective red coloura-
tion is perhaps quite as necessary during the wet season as during the
dry."

The question as to the significance of the non-development of pig-
ment in the stomata in leaves with red epidermal cells is also considered
by Ewart. Stahl's view is that the stomata by this means transpire
less, and so are able to keep open longer for purposes of transpiration
and photosynthesis. Ewart, on the contrary, regards the absence of
anthocyanin as being due to the fact that the stomata are organs which
react to light, and it is important that they should be exposed to the
same intensity of light as that falling on the rest of the leaf even at the
risk of injury.

Again, Stahl, as we mentioned previously, looks upon the develop-
ment of anthocyanin in stigmas of anemophilous plants as an adaptation
for increasing the temperature and aiding the growth of the pollen-grain.
Ewart, however, notes that it has been shown that the pollen-tube

vni] THE SIGNIFICANCE OF ANTHOCYANINS 139

growth is retarded by light and hence is protected by the pigment.
Similarly, Stahl considers the reddening of young shoots in temperate
regions in the spring to be protective against cold, whereas Ewart
considers the function to be protective for the chloroplast, since the
latter is more sensitive to light when the temperature is low. Reddening
of submerged plants in sugar-cultures Ewart explains as being due to
an unhealthy condition of the plant in which state it requires more
protection against light.

Ewart concludes thus: 'There can be little doubt that, both in
the tropics and in temperate climes, the main and primary function
of the red dye, when present in exposed parts, is to act as a protection
against light of too great intensity; though in all cases its presence
at the same time confers upon the plant a slightly increased power of
absorbing heat. For calling attention to this latter possibility Stahl
deserves full credit from both the physiologist and the biologist: in
a few cases, such as in the horizontal leaves of shade plants having the
red colouration present on the under surfaces only, the relatively slight
heat-absorbing power of the dye may, by secondary adaptation, have
become its most important function."

The latest work on the physiological significance of anthocyanin,
and undoubtedly the most accurate, is that published by Smith (420)
in 1909. This author determined the temperature of leaves in tropical
insolation in Ceylon, using a thermo-electric apparatus. The latter
was of the improved pattern which had been employed by Blackman
& Matthaei, and had the advantage that it could be used even in the
lamina of thin leaves and also for the internal temperature of leaves
in natural illumination.

In the first of a series of experiments connected with this point
young leaves of Amherstia nobilis and Saraca indica were compared.
Both trees have young foliage of the pendent type we have previously
described. The leaves of Amherstia are of a deep brownish-red colour
due to the presence of anthocyanin, and after this pigment is removed
by a solvent, there is seen to be but a slight development of chlorophyll.
The leaves of Saraca are greenish-white and have, in the same way,
a small development of chlorophyll. Both leaves are also thin and
flaccid and hence form good objects for comparison. On exposing the
leaves and testing the temperature, it was found that the coloured
leaf of Amherstia reached a temperature of 2 C. higher than the leaf
of Saraca. Another experiment was conducted with a young red leaf
of Mesua ferrea as compared with a young leaf of Saraca indica. The

140 THE SIGNIFICANCE OF ANTHOCYANINS [CH.

red Mesua leaf registered a temperature of 2-8 C. higher than the
Saraca leaf.

It was next found by comparing green and white leaves, that a green
leaf also reaches a higher temperature than a white one. Hence it
seems clear that any pigment, either chlorophyll or anthocyanin, raises
the temperature of the leaf, and that the simultaneous presence of
anthocyanin and chlorophyll in the leaf will raise the temperature
considerably above that of a leaf without either pigment. The latter
point was demonstrated in a comparison between a green and white
leaf of Caladium sp. and a red and green leaf of the same genus, the
electrodes being placed in the green and white leaf where there was
little chlorophyll. The red and green leaf then showed a most striking
rise of 3-9 C. above the green and white leaf temperature.

Attention was finally given to the differences between the temperature
of young flaccid and coloured leaves as compared with older green
leaves from the same tree. With Saraca indica, a mature green leaf
showed a difference of 4 C. above a young colourless leaf. When,
however, a young leaf of Theobroma Cacao which is red, was compared
with a mature green leaf, the young leaf showed a difference of 3-5 C.
above the mature leaf, and this one would suppose to be due to the
presence of the anthocyanin. On the other hand, comparisons of young
red leaves with mature green leaves of Amherstia nobilis, resulted in
a decidedly higher temperature in the mature leaf.

"Thus," Smith concludes, "it seems, that we have a series beginning
with Saraca indica, in which anthocyan is almost absent, and in which
the mature leaf is always higher in temperature than the young leaf.
Then comes Amherstia nobilis with a brownish-red colour, in which the
mature leaf is, as a rule, only slightly higher in temperature than the
young leaf. Lastly, we have Theobroma Cacao with the young leaf an
intense pinkish-red and the mature leaf lower in temperature than
the young leaf.

"No doubt the relative temperatures of mature and young leaves
are to be correlated with the amount of anthocyan in the young leaf.
The young leaves, without this pigment, would be always cooler than
the mature leaf, as is the case in Saraca indica. The presence of more
or less anthocyan produces a temperature in the young leaf, which
almost reaches (Amherstia} or exceeds (Theobroma) the temperature of
the mature leaf. Thus the general tendency of these results is to confirm
and extend Stahl's conclusion that the presence of anthocyan tends to
raise the internal temperature of the leaf. What biological advantage.

vin] THE SIGNIFICANCE OF ANTHOCYANINS 141

if any, is gained by the plant in this way is quite another question,
but it is well to have this physical effect definitely established."

When we review the work on the subj ect of the uses of anthocyanin,
we find it singularly unconvincing in any direction. In favour of the
light-screen hypothesis, there is undoubtedly a greater development of
anthocyanin in sun-exposed leaves than in leaves of the same plant in
the shade; and, conversely, in red-leaved plants like the Blood Hazel
and Copper Beech, a lack of colour in leaves which happen to be
shaded. But, on the other hand, Kerner's experiments on Satureja
and Linum are not very convincing, since so many factors might account
-for the unhealthiness of the Flax plants. And again, even if Prings-
heim's red screen were protective to chlorophyll, it consisted of iodine
in carbon disulphide which has a different absorption spectrum from
anthocyanin. Moreover, Engelmann's analyses seem to prove that the
anthocyanin only absorbs those rays which pass through chlorophyll,
and an effective screen would be one with a similar absorption spectrum
to chlorophyll itself. As regards Stahl's hypothesis, it is certainly
supported by a definite piece of evidence, viz. that the temperature of
red leaves is higher than that of green. Also, there are without doubt
a number of shade plants which do possess anthocyanin to a considerable
extent. Yet it is difficult to find a hypothesis which would fit all cases
of anthocyanin distribution without reduction to absurdity. The
pigment is produced, of necessity, in tissues where the conditions are
such that the chemical reactions leading to anthocyanin formation
are bound to take place. For the time being we may safely say that
it has not been satisfactorily determined in any one case whether its
development is either an advantage or a disadvantage to the plant.

PART II

ANTHOCYANINS AND GENETICS

PART II

ANTHOCYANINS AND GENETICS

CLASSES OF VARIATION.

As pointed out in Chapter n, practically all flowering plants produce
anthocyanin ; moreover, when this pigment is developed in the flower,
we are able to see that each specific type forms anthocyanin of a certain
characteristic colour. In nature to some extent, but under cultivation
very commonly, colour- varieties arise from the type. The underlying
cause of these variations is still a matter for conjecture, though such
suggestions as can be offered will be given later. Colour- varieties
have afforded plentiful material for Mendelian research, and it is to
the cases of inheritance involving anthocyanin as a character that the
following pages are devoted.

Among the colour-varieties of different genera and species, there
is a certain correspondence in the series of varieties produced, that is,
we may find one series in a number of plants not necessarily related,
and another series in a number of other plants, and so forth. The
kinds of variation which may occur we are able to classify as follows:

1. The loss of power to form anthocyanin pigments which results
in albinism. The albino may be white or yellow ; if yellow, the pigment
may be either plastid or soluble.

2. The loss of power to produce blueness. The type has blue, or
purple, anthocyanin : the variety has red anthocyanin .

3. The loss of power to produce redness. The type has red antho-
cyanin : the variety blue, or purple, anthocyanin.

4. The loss of power to augment the formation of anthocyanin
and hence to intensify its colour. The type is fully coloured: the
variety tinged only.

5. The loss of power to inhibit the formation of anthocyanin and
hence to diminish its intensity. The type may be pale in colour and
the variety deep: or the type may be tinged only and the variety
fully coloured.

The following variations are independent of variation in the

w. P. 10

146 ANTHOCYANINS AND GENETICS

anthocyanin pigments, and may (with a certain exception in 9) take
place simultaneously with variation in the anthocyanins.

6. The loss of power to produce yellow pigment in the plastids.
The type has yellow plastids: the variety colourless plastids.

7. The loss of power to inhibit the formation of pigment in the
plastids. The type has colourless plastids : the variety yellow plastids.

8. The loss of power to inhibit the formation of yellow soluble
pigments. There is no yellow in the type: the variety is yellow.

9. The loss of power to produce yellow soluble pigment. This
variation is of rare occurrence. The type may contain both yellow
soluble pigment and anthocyanin, and the variety may be without one
or both : or the type may be without yellow colour, but may produce
a soluble yellow variety, which, in turn, may lose its power of forming
yellow pigment and is then also unable apparently to form anthocyanin.

It should be understood that the above is only a classification on
the broadest basis ; when we come to the details of variation it is almost
necessary to consider each case separately. It is difficult, for various
reasons, to make any kind of comprehensive classification, apart
from the fact that no two cases are exactly alike; in many species,
for instance, the inheritance of colour has not been systematically
worked out, and relationships between variations cannot be correctly
judged; new 'breaks/ moreover, are continually occurring in horti-
cultural plants; there are also the complexities introduced by species-
crossing, and in these circumstances it is often difficult to identify
the type from which the variety has arisen. In the following paragraphs
an attempt is made to indicate some of the main series or ranges of
variation. It has only been possible to include a selection from the
great mass of material provided by observations on plants under
cultivation, but it is hoped that the lines suggested may provide a basis
for further classification, as our knowledge of the inter-relationships of
varieties increases.

A point to be emphasised is that one cannot judge correctly of the
inter-relationships of varieties from appearances. For true knowledge
it is not only necessary, as we have said, to find out the behaviour
of the pigments in heredity, but we must examine their properties,
and above all we should know their chemical composition. The colour
series in Dahlia and Tropaeolum, for instance, are similar to the eye,
but are really fundamentally different. Hence, in the following series,
from lack of knowledge, instances may be grouped together which are
not in reality of the same nature.

ANTHOCYANINS AND GENETICS 147

We may now consider the ranges of variation in greater detail.
The simplest series is that which involves variation to albinism only,
and, in the case of some genera grown as horticultural plants, this is
the only important colour- variation known : the type produces antho-
cyanin, and the albino is without that pigment. As examples we may
quote : Datura Stramonium (Thorn-apple), Dictamnus Fraxinella (Rue),
Geranium sanguineum, Epilobium angustifolium, Lavatera trimestris,
Linaria Cymbalaria, Malope trifida, Malva moschata and Polemonium
caeruleum (Jacob's Ladder); in some of these species there may also
be different intensities of type colour due to heterozygous forms (see
p. 183).

If it should happen that the type produces both anthocyanin and
plastid pigment, then loss of anthocyanin will not give a white variety but
a yellow. Examples of such a case are the yellow varieties of Abutilon
spp., Fritillaria imperialis and the variety lutea of Atropa Belladonna;
in all these species the type is yellow suffused with anthocyanin.

Variation to redness, in addition to variation to albinism, is charac-
teristic of another series. Variation to redness is a more complex
phenomenon than albinism, and the series requires analysis. In the
first place the type may be one of two kinds ; it may be either some shade
of magenta, purple, or purplish-red, or it may be blue. As examples
of the first group we may suggest Anemone Pulsatilla which has a variety
rubra; Clarkia elegans (Bateson, 524) which has a red (pink) variety,
the type being magenta. Linaria alpina (Saunders, 586) and Pisum
sativum (Garden Pea) (Lock, 518) have definite red varieties; Salvia
Horminum (Saunders, 487) is violet in type with a red variety, and of
Viola odorata there is a variety redder than the type. With the
exception of the last, all the above mentioned also give true albinos.
Examples of the second group (type blue) are Centaurea Cyanus (Corn-
flower) of which purplish-red and pink varieties are known; Lobelia
Erinus and Vinca minor (Periwinkle) produce a purplish-red variety
and Myosotis sylvatica (Forget-me-not) a pink; Delphinium Ajacis
(Larkspur) and Campanula medium have both mauve and pink varieties ;
in Aquilegia vulgaris (Columbine) the type is violet-blue, and there are
several purplish-red and pink varieties. This second group also varies
to white, though it is doubtful in some cases whether there is complete
albinism. In other species there is more than one grade of variation
to redness. In the garden Stock (Matthiola) (Saunders, 487) it is believed
that the type was some shade of purple; there is a crimson (blue-red)
variety and also a true red, ' Terra-cotta.' Of Dianthus barbatus (Sweet

102

148 ANTHOCYANINS AND GENETICS

William) the type was doubtless of a magenta shade and there is varia-
tion to crimson and also to a true red, 'Scarlet.' Similarly Primula
sinensis (Chinese Primrose) had in all probability a magenta type
(Hill, 577) and varies to crimson and true red, 'Orange King.' In
Antirrhinum majus (Wheldale, 535, 548) the type is magenta and the
red variety, 'Rose Dore': the case of Antirrhinum (Snapdragon) is
further complicated by the existence of a yellow variety (see below).
Variation to redness also occurs in Cheiranthus (Wall-flower), Hyacinthus
and others, but these are dealt with later in connection with more
complex series.

Another variation-series is that which includes a yellow variety.
Colour in the yellow variety may be due either to plastid or soluble
pigment, and the two series have very different characteristics. It
should be emphasised that the soluble-yellow series is essentially
different from the plastid-yellow series. In the latter, as we have
already pointed out, the yellow is really the albino as regards antho-
cyanin pigment, whereas a soluble yellow variety is formed, as a rule,
by the loss of a factor from an albino. In the case of the plastid-
yellow series, further complexities, in addition to the simple loss of
anthocyanin, are introduced by variation in the plastid colours, or by
total loss of these pigments.

To deal first with the soluble-yellow series: Antirrhinum majus
forms a typical example. The type is magenta; loss of anthocyanin
gives an ivory-white variety. From the ivory a soluble-yellow variety
is derived; mixture of yellow with the magenta anthocyanin of the
type, i.e. simultaneous presence of both pigments, produces another
variety, crimson. A red variety, 'Rose Dore,' has arisen from the
magenta, and a mixture again of rose dore with yellow results in another
variety, bronze. Finally there is a white variety which contains neither
anthocyanin nor yellow pigment. Hence the series can be expressed:
magenta, crimson, rose dore, bronze, ivory 1 , yellow and white. Though
we have no exact knowledge of the colour relationships in Althaea
rosea (Hollyhock), Dahlia variabilis and Dianthus Caryophyllus (Carna-
tion), yet these species in the main exhibit a similar series to Antirrhinum,
for they produce magenta, crimson, yellow and ivory-white varieties;
D. Caryophyllus, however, is characterised by many other shades. It

1 In the soluble-yellow series ivory or ivory-white is used for the albino, which is
without anthocyanin, in contrast to the true white which is without both anthocyanin
and soluble yellow pigment. At present the only two species in which both these
varieties are known are Antirrhinum majus and Phlox Drummondii.

ANTHOCYANINS AND GENETICS 149

cannot at present be stated whether Dahlia and Althaea have a red
corresponding to rose dore. In many respects Phlox Drummondii
shows the same series as Antirrhinum, though violet is included in
addition. It is interesting to compare the series given by Primula
sinensis (Gregory, 557) and Dianthus barbatus with Antirrhinum. In
the two former there is a crimson similar in appearance to the crimson
of Antirrhinum, but it is not due to mixture with yellow, since a soluble
yellow variety is unknown in Primula and D. barbatus. But the
'Scarlet' of D. barbatus and the 'Orange King' of P. sinensis appear
to be truly comparable to the rose dore of Antirrhinum.

The Marvel of Peru, Mirabilis Jalapa (Marryat, 533), produces a
varietv with soluble yellow pigment, and although this species must
obviously be included in the present series, yet it is, in a sense, funda-
mentally different from Antirrhinum. From our knowledge of the
colour-inheritance of Mirabilia, one is led to believe that the original
type was crimson, and this apparently contains a mixture of magenta
anthocyanin and soluble yellow pigment. Loss of anthocyanin gives
a yellow variety, and further loss of yellow pigment, a white variety.
Anthocyanin may also be present on a pale yellow ground and then we
have a magenta variety. Thus the series is an inversion, so to speak,
of Antirrhinum and runs: crimson, magenta, yellow and white. It is
further complicated by the existence of heterozygous forms, but these
will be considered later (see p. 170).

Cheiranthus Cheiri (Wall-flower), on the other hand, is typical of
the plastid-yellow series. The original wild type has deep yellow
flowers tinged with brown, from which has arisen in cultivation the
ordinary brown variety (see p. 153). The brown colour is due to the
simultaneous presence of purple anthocyanin and deep yellow plastids.
Loss of anthocyanin from the brown gives a yellow variety, which,
strictly speaking, is the albino as regards anthocyanin. Some loss
from, or change in, the deep yellow plastids results in a lemon yellow
variety. When the purple anthocyanin is present with pale yellow
plastids, the latter are masked, and we get the purple variety now
commonly grown. During recent years further varieties have appeared,
some of which are practically cream in colour. The purple anthocyanin,
too, has produced a red variation comparable to the red group previously
mentioned. This red, on a background of lemon or pale yellow plastids,
gives us the varieties ' Eastern Queen ' and ' Ruby ' : on a background
of deep yellow plastids, the new 'Scarlet.' Thus the Cheiranthus
series runs: crimson (brown), purple, scarlet, ruby, yellow, lenlon

150 ANTHOCYANINS AND GENETICS

and cream. In main outline a similar series, brown or crimson, purple
or magenta, deep yellow and pale yellow is shown by Zinnia elegans,
and also by the new varieties of the Sunflower, Helianthus annuus
(Cockerell, 602, 603, 611). Variation in the Garden Nasturtium
(Tropaeolum majus) is on the same lines as in Cheiranthus, though it
differs in one respect, namely that in Cheiranthus the type anthocyanin
is purple and gives rise to a red variety, while in Tropaeolum the type
anthocyanin is scarlet, or carmine red, and gives rise to a purple variety.
The colour representing the original type of Tropaeolum is the orange-
red due to carmine anthocyanin on yellow plastids. A deep yellow
variety is the albino after the loss of anthocyanin and, as in the case
of Cheiranthus, it produces several pale yellow varieties approaching to
cream as a result of the loss of some factor from the plastids. On this
pale yellow ground the red anthocyanin appears as carmine or ruby. The
red anthocyanin also varies to a blue-red or purple, which, on the deep
yellow plastids, appears as maroon, and on the pale yellows as purple.
There is additional variation caused by anthocyanin blotches at the base
of the petals ; these may be carmine or purple, and may be retained when
the main part of the flower is free from anthocyanin ; they may, however,
be entirely lost, so that the flower is wholly yellow. Variation in Salpi-
glossis sinuata is also in all probability on similar lines to Cheiranthus.

We may next consider an additional range in the plastid-yellow
series due to complete loss of colour from the plastid. In this way a
white arises and it may exist as such, or anthocyanin may be present
on the white ground. Typical of this series is Chrysanthemum ; the type
was in this case yellow, probably slightly tinged with anthocyanin.
Loss of an inhibiting factor (see p. 152) would give rise to a crimson
variety, i.e. purple anthocyanin on yellow plastids. Loss of yellow
plastids, without loss of anthocyanin, gives purple, purplish-red or
magenta ; loss of anthocyanin gives either yellow or white. The series
is then : crimson, magenta, yellow and white. Helianthemum vulgare
(Rockrose), Viola tricolor (Pansy) and the garden Tulips probably
come into the Cheiranthus-Chrysanthemum class, but our knowledge of
these species is not very systematic. Viola tricolor possibly contains
species-crosses. We can readily distinguish two groups of the garden
Pansy : one includes white, reddish-purple, purple and purple-blue ;
the other yellow, brown and crimson. But we have no evidence as
to how the groups are related.

We can thus differentiate two groups or series, soluble-yellow and
plastid-yellow : in the former the type is generally magenta and gives

ANTHOCYANINS AND GENETICS 151

the series crimson, red, orange, yellow and ivory-white: in the latter
the type is, as a rule, crimson and gives the series magenta, orange,
red, yellow, pale yellow and sometimes cream or white.

There is another series we may best consider at this point, and that
is one in which a particular variety of plastid pigmentation occurs
known as 'cream.' The plastids contain an orange-yellow pigment,
but only in sufficient quantity to give the petals a cream appearance,
and the variety is further characterised by being recessive to white
containing colourless plastids. Such a ' cream ' variety is found in the
Sweet Pea, Lathyrus odoratus (Bateson, 524) and Matthiola (Saunders,
475); in Matthiola the series runs: purple, crimson, terra-cotta, white
and cream. The cream may underlie any of the anthocyanin pigments,
but it is too pale to affect the resultant colour, and such individuals
are indistinguishable from those having anthocyanin on a white ground,
except at the 'eye' of the corolla where the cream or white ground,
as the case may be, is shown. In Lathyms there is a very great range
of colour, and the exact inter-relationships of many varieties are still
unknown; the series includes blue, purple, mauve, crimson, pink,
salmon, white and cream. In some varieties (Thoday, 547) the under-
lying cream modifies the effect of the anthocyanin. Cream plastids
are also found in Hyacinthus orientalis in which they produce a fairly
deep yellow ; in this species the range of colour is also very great, since
it includes several shades of blue, purple, magenta, pink, as well as
white arid cream, but, as there have been no systematic breeding experi-
ments, their relationships to each other cannot be stated. It is not
known whether the cream plastids in varieties of Rosa are of the same
nature as those of Lathyrus and Matthiola.

Finally there is the series which includes a blue variety. Variation
to redness when the type is blue or purple is, as we have seen, a common
phenomenon, but variation to blueness is much less frequent. In
Lathyrus odoratus varieties have appeared which are bluer than the
type, 'Purple Invincible,' and the same is true for Phlox Drummondii.
In Primula sinensis, true blue- flowered varieties, 'Cambridge' and
'Oxford' blues, occur, the type in all probability having had pale
magenta flowers (Hill, 577); so also from the original Cineraria with
flowers of a pale magenta, blue-flowered varieties have arisen, though
probably under the influence of species-crossing. A striking instance
of a blue variation from a scarlet is the variety, coerulea, of Anagallis
arcensis (Scarlet Pimpernel) 1 .

1 Kajanus (579) states that blue is recessive to red in Trifolium (Clover).

152 ANTHOCYANINS AND GENETICS

We may next consider those variations which affect the intensity
of colour. If a colour series such as we have described for Lathy rus
or Antirrhinum be examined, a number of definite varieties will be
found of a paler or deeper shade of the same colour. Without Mendelian
analyses it is not as a rule possible to arrive at the underlying significance
of the shade differences. Thus, a paler shade than the type may signify
either loss of a factor for full-colour, or the existence of a heterozygous
form between type and albino. Hence, in dealing with this kind of
variation, we can only quote as satisfactory instances those cases of
which we have knowledge from experimental breeding. The main
expressions of intensity-variation are tinged and deep varieties. The
tinged variety forms a case in which a deepening, or full-colour factor,
affecting anthocyanin formation has disappeared from the type, leaving
the flower flushed or tinged with colour only. Examples are the
'tinged ivory' variety of Antirrhinum (Wheldale, 535) and the 'tinged
white' and 'picotee' of Laihyrus (Bateson & Punnett, 500). The
variation of tingeing is common to both red and blue classes.

The deep variety can be defined as one having a deeper shade of
anthocyanin, i.e. more pigment, than the type from which it is derived,
as the result of the loss of an inhibiting factor. Since it is often difficult,
and sometimes impossible, to ascertain the original type in many horti-
cultural plants, this definition cannot be rigidly applied. Like tingeing,
the variation of deepening is common to both red and blue classes.
Examples of such varieties are the deep shades of magenta, crimson,
rose dore and bronze of Antirrhinum (Wheldale, 535, 548), purple-
winged 'Purple Invincible' and 'Miss Hunt' varieties of Laihyrus
odoratus (Bateson & Punnett, 500) ; deep shades of crimson and magenta
in Primula sinensis (Gregory, 557) and deep purple and crimson in
Matthiola (Saunders, 487) ; also deep varieties of Cyclamen. The
following species, among many others, have varieties obviously deeper
than the type : Althaea rosea, Dahlia variabilis, Dianthus Caryophyllus,
D. barbatus, Hyacinthus orientalis and Phlox Drummondii.

In the instances quoted above, the type, from which the deeper
variety is derived, is itself fully pigmented. There are, however, other
cases in which the type is either unpigmented, or only slightly pigmented,
and, on the loss of an inhibiting factor, a coloured variety is produced.
As examples may be quoted Anemone japonica in which the type has
a white petal oid calyx tinged with anthocyanin on the under surface,
whereas the variety is purple-flowered ; the wild Bellis perennis (Daisy)
has white ray florets tinged with anthocyanin and the garden variety

ANTHOCYANINS AND GENETICS 153

is red-flowered. Cheiranthus Cheiri, also, in the wild state is only tinged
with brown (anthocyanin), and from it in cultivation have appeared
varieties fully coloured with anthocyanin of deep shades of brown
and purple. In Crataegus Oxyacaniha (Hawthorn) the white-flowered
type has produced a red-flowered variety. In Cyclamen persicum the'
type has white flowers with a basal magenta spot on the petals and has
given rise to varieties with fully-coloured magenta and crimson flowers.
In Helianthemum vulgare, Primula acaulis (Primrose) and P. veris
(Cowslip), the types are yellow and the varieties both crimson and
magenta. Further examples are Freesia and Acliillea MiUefolium (de
Vries, 498) with coloured flowers. De Vries (498) also notes this phe-
nomenon. He says: "a pink-flowered variety of the ' Silverchain' or
' Bastard- Acacia ' (Robinia Pseud-Acacia) is not rarely cultivated. The
'Crown' variety of rice, oats and barley are also to be considered as
positive color-variations, the black being due in the latter cases to a
very great amount of the red pigment." And again : "The best known
instance is that of the ever-flowering begonia, Begonia semper florens,
which has green leaves and white flowers, but which has produced
garden varieties with a brown foliage and pink flowers." Cockerell (602,
603, 611) has also drawn attention to a red-flowered variety of Helianthus
of which the type is yellow. Again, the red colour may be developed in
the carpels as in the varieties of Primula with red stigma (Gregory, 557)
and the purple-podded forms of Pisum and Phaaeolus. The variety
appears in other fruits too (the ' Blood Orange ' and the red Banana) ;
also in the Gooseberry, Ribes Grossularia (de Vries, 498).

Another phenomenon which sometimes appears in variation is that
which may be termed partial albinism ; the type has coloured flowers
but the variety has white flowers, though all the vegetative organs may
still form anthocyanin. This is the case in the blue Vinca minor with
its white-flowered variety of which the corolla is slightly tinged with
purple. Of Geranium Robertianum (Herb-Kobert), also, there is a
white-flowered variety with red stems and leaves. A similar variation
is sometimes found in the native flora, i.e. Polygala vulgaris, Jasione
montana, and others, but it is not really possible, in the absence of
Mendelian evidence, to distinguish these varieties from the tinged
varieties already mentioned. Partial albinos, moreover, are similar
in appearance to types in which anthocyanin is inhibited; but an
important distinction lies in the fact that partial albinos cannot give
rise to coloured varieties though inhibited types may.

A case difficult to place is that of the dominant or inhibited white

154 ANTHOCYANINS AND GENETICS

of Primula sinensis (Gregory, 557). This variety has apparently the
power to form pigment in the flower but colour is prevented from
appearing by an inhibiting substance. Hence the case most closely
resembles the inhibited types of Anemone, Crataegus, Primula veris, etc.,
mentioned above, and must, for the present, be included in this class.

Up to this point we have been solely concerned with variation in
flower colour. But it should be borne in mind that the colour-varieties
already described may be characteristic of the fruit and seed and also
of the vegetative organs ; in fact, the whole plant may either be a total
albino, or pigment may be lost from various organs or parts indepen-
dently of each other. These phenomena will be dealt with later in
greater detail (see p. 180). A few instances of colour-variation in
different organs may however be mentioned. Of albinism in fruits
we have the white-fruited varieties of Atropa Belladonna, Daphne
Mezereum, Fragaria vesca, Ribes rubrum, Rubus Idaeus, Solanum nigrum,
Vitis vim/era and many others. De Vries (565) also notes a red-berried
variety of Empetrum nigrum. In seeds there are white-seeded varieties
of Pisum and Phaseolus. Of coloured roots also, white varieties may
occur as in Beta vulgaris. Variation to redness may also affect the
vegetative organs as for instance in the red-flowered varieties, rose
dore of Antirrhinum and ' Orange King ' of Primula sinensis, in which
the leaves, stems and petioles develop red anthocyanin instead of purple.
In the tubers, moreover, of the Potato, Solanum tuberosum (Salaman,
544), one may have both red and purple varieties (pigment due to
anthocyanin) as well as dominant and recessive white tubers.

Loss of an inhibiting factor which in the case of the flower gives
a coloured from a white variety, is quite common in the vegetative
organs. This coloration of vegetative organs may be connected with
deepening of flower-colour, for example, deep varieties of Antirrhinum
and Primula sinensis which have pigment in leaves and petioles. Or
it may occur more obviously as a red-leaved variety, as in the following :
Atriplex hortensis, Berberis vulgaris, Beta vulgaris, Brassica oleracea,
Canna indica, Corylus Avellana, Fagus sylvatica, Lactuca sativa, Perilla
nankinensis, Plantago major; de Vries (498) also notes Tetragonia
expansa and the brown-leaved Trifolium 1 . Lastly a very interesting

1 With regard to red-leaved varieties, de Vries (498, 565) notes that they may produce
green-leaved branches (red Corylus, Fagus, Betula). Also red bananas have produced
a green variety with yellow fruits. Whether this phenomenon is bud variation, or whether
it may be in some instances due to the production of the true albino and not the inhibited
type, is uncertain. An analogous example in flower-coloration would be Cheiranthus

ANTHOCYANINS AND GENETICS 155

case of the occurrence of a red variety in roots is recorded by Wittmack
(495), that is the variety of the Carrot, Daucus Carota Boissieri
Schweinfurth, which has anthocyanin in the root, in addition to the
orange plastid pigment (carotin).

As to variation in nature, that of albinism is most frequent. It is
found in the wild state among the British flora in Armeria vulgaris,
Calluna vulgaris, Campanula rotundifolia, Carduus nutans, C. palustris,
Digitalis purpurea, Erica cinerea, Lamium purpureum, Lychnis Flos-
cuculi, Ononis arvensis, Pedicularis sylvatica, Symphytum qfficinale
and Scilla nutans. Bentham & Hooker 1 also mention varieties of
Ajuga reptans, Campanula latifolia, Cenlranthus ruber, Delphinium Ajacis,
Iris foetidissima and Malva moschata without anthocyanin in the
flowers, but it cannot be deduced from the text in each of the above
cases whether there is complete absence of the pigment from the plant,
or loss from the flower only. As regards variation to redness, among
native species Poly gala vulgaris has a red-flowered variety. Bentham
& Hooker 2 record in addition red-flowered varieties of Delphinium
Ajacis, Veronica spicata, V. officinalis, V. Beccabunga and V. Chamaedrys.
Of species in which the flowers are normally white though they may
produce red varieties, or red pigmentation under certain conditions,
Gillot (457) gives a considerable list. Guinier (456) also notes a variety
of strawberry (Fragaria) with purple flowers, and Chabert (461) records
red flowers of certain species of Galium which are normally white-
flowered. A variation, extremely rare in nature, if not altogether
unknown, is that of white from a type with plastid yellow pigment,
though as a species differential character it is quite common, for
instance in the Ranunculaceae and Compositae.

DETAILS OF CASES OF MENDELIAN INHERITANCE IN
COLOUR-VARIETIES.

Many of the varieties mentioned in the last section have provided
material for cross-breeding work on Mendelian lines. Since most of
the species investigated differ more or less in their colour series and
behaviour, a separate account will be given of the more important
cases in turn. The genera and species which have been employed in
these investigations are the following:

Cheiri ; the type is yellow tinged with anthocyanin : this varies to deep brown and this
again gives a yellow (the albino) free from anthocyanin.

1 Handbook of the British Flora, London, 1896.

2 Loc. cit.

156 ANTHOCYANINS AND GENETICS

Agrostemma Oithago (Corn Cockle), de Vries (474, 498).

Amaranthus caudatus, de Vries (498).

Anagallis arvensis (Scarlet Pimpernel), Heribert-Nilsson (576), Weiss (566).

Antirrhinum majus (Snapdragon), Bateson (480), Baur (517, 536), Wheldale (514,

535, 548), de Vries (565).
Arum maculatum, Colgan (552).
Aster Tripolium, de Vries (474, 498).

Atropa Belladonna (Deadly Nightshade), Saunders (475).
Beta (Beet-root), Kajanus (559, 580, 598).
Brassica (Cabbage), HaUqvist (631), Kajanus (580, 598).
Canavalia ensiformis, Lock (504).
Canna indica, Honing (608, 633).
Cattleya, Hurst (502, 531, 540, 596).
Clarkia elegans, Bateson & Punnett (524), de Vries (498).
Corchorus capsularis (Jute), Finlow & Burkill (572).
Coreopsis tinctoria, de Vries (474).

Cypripedium (Paphiopedilum), Hurst (502, 531, 540, 596).
Datura Stramonium (Thorn-apple), Saunders (475, 487), de Vries (474).
Dendrobium, Hurst (596).
Digitalis purpurea (Foxglove), Jones (578), Keeble, Pellew & Jones (542),

Saunders (563).
Geum (Avens), de Vries (498).

Gossypium (Cotton), Balls (515, 523), Fletcher (508), Leake (561).
Helianthus (Sunflower), Cockerell (602, 603, 611), Shull (520).
Hepatica (Anemone), Hildebrand (473).
Hieracium (Hawkweed), Ostenfeld (505, 543).
Hordeum (Barley), Biffen (501).

Hyoscyamus (Henbane), Correns (490), de Vries (474).
Lathyrus odoratus (Sweet Pea), Bateson (499), Bateson & Punnett (487, 496,

500, 516), Thoday (547).
Linaria alpina, Saunders (586).
Linum usitatissimum (Flax), Tammes (564, 610).
Lychnis dioica, Correns (485), Saunders (475), Shull (520, 546, 588), de Vries

(474).
MattUola (Stock), Correns (472), Saunders (475, 487, 496, 500, 506, 562),

Tschermak (494, 590).
Mirabilis Jalapa (Marvel of Peru), Baur (517), Correns (482, 497, 537), Marryat

(533).

Nicotiana (Tobacco), Lock (532), Haig Thomas (594).
Oenothera (Evening Primrose), Davis (592, 604), Gates (539, 555, 573, 606),

Heribert-Nilsson (575, 632), Shull (609).
Oxalis, Nohara (635).

Papaver somniferum (Opium Poppy), Hurst (502), de Vries (474, 498).
P. Rhoeas (Field Poppy), Shull (588).
Phaseolus multiilorus (Scarlet-runner), P. vulgaris (French Bean), Emerson (483,

492, 528, 529), Lock (504), Shull (513, 521), Tschermak (479, 494, 590).
Phyteuma (Rampion), Correns (486).

ANTHOCYANINS AND GENETICS 157

Pisum sativum (Common Pea), Bateson (487), Lock (512, 518), Tschermak (479,

494, 590).

Polemonium (Jacob's Ladder), Correns (486), de Vries (498).
Primula sinensis (Chinese Primrose), Gregory (557, 607), Keeble (582), Keeble

& Pellew (541).

Reseda (Mignonette), Compton (569).
Salvia Horminum, Saunders (487).
Senecio vulgaris (Groundsel), Trow (589).
Silene Armeria (Lobel's Catchfly), de Vries (474, 498).
Solatium nigrum, de Vries (474).

S. tuberosum (Potato), East (538), Salaman (544, 585).
Trifolium (Clover), Kajanus (579).
Tropaeolutn (Nasturtium), Weiss (534).
Veronica longifolia, de Vries (474, 498).
Viola cornuta, de Vries (474, 498).
Zea Mays (Maize), Burtt-Davy (568), Correns (476), East (570), East & Hayes

(553), Emerson (554, 571, 605, 629), Lock (493, 504, 583).

Agrostemma Githago. Pigmented type dominant to albino. De
Vries (474, 498).

Amaranthus caudatus. Red-leaved type dominant to green-leaved
variety. De Vries (498).

Anagallis arvensis. Weiss (566) crossed the scarlet-flowered type
with the blue-flowered variety and obtained the red type in F x . In F 2
there was segregation into red and blue in the proportion 3:1. Hence
the blue arises from the red by the loss of a single dominant factor.

Heribert-Nilsson (576) crossed a variety which had pink, almost
white, flowers, which bred true, with the red type. F r was red like
the type. In F 2 there was segregation into red and pink in the propor-
tion 3 : 1.

Antirrhinum majus. De Vries (565) made the first experiments on
this species using the following varieties:

Rot tube and lips of the corolla red, the lips deeper.

Fleischfarbig tube and lips pale red.

Delila tube pale or white, lips fairly deep red.

Weiss white, often with a distinct very pale red tinge.

De Vries regarded red as made up of fleischfarbig, F. and delila, D,
two dominant characters, and white as carrying two characters, w and
w', recessive to F and D respectively. Then F x from red x white
would be FDww', and F 2 would give red, flesh, delila and white in the
proportion 9:3:3:1, which de Vries maintains agrees with his experi-
mental results. Various F x reds, delilas, and flesh colours were selfed,
and the F 2 was in accordance with the above scheme.

158 ANTHOCYANINS AND GENETICS

Bateson (480) suggested that a better interpretation of de Vries'
results would be to suppose that the Fj plants produced, in each sex,
in equal numbers, gametes having the characters, R, F, D and W,
and these, on fortuitous mating, would give 9R : 3F : 3D : 1W. In
view of more recent work, both views have to be revised.

In 1907 the author (514) published results of work on a number of
colour- varieties ; those used were

1. White lips and tube of the corolla pure white.

2. Yellow lips yellow; tube ivory.

3. Ivory -lips and tube ivory.

4. Crimson lips crimson; tube magenta.

5. Magenta lips and tube magenta.

With the exception of white, all varieties have a constant orange-
yellow palate. Further varieties occur in which the lips are magenta
or crimson, and the tube is ivory; the term 'delila' was used by de
Vries for these varieties and has been retained.

The inheritance of colour in the varieties employed was represented
by the following factors :

Y. A factor representing yellow colour in the lips associated with
ivory tube-colour.

I. A factor representing ivory colour in the lips.

L. A factor representing magenta colour in the lips.

T. A factor representing magenta colour in the tube.

It was found also that certain relationships existed between the
factors, i.e.

1. All zygotes, from which Y is absent, are white, though they
may contain any of the factors I, L and T.

2. The effects of the factor T are not manifested unless L is also
present in the zygote ; that is, no magenta colour appears in the tube
unless there is magenta colour in the lips.

3. All zygotes containing Y are coloured, the actual colour being
modified and determined by the presence of one or more of the remaining
factors. A zygote containing Y only, or Y and T is yellow.

4. Ivory is dominant to yellow; a zygote containing Y and I, or
Y, I and T is ivory.

5. Since magenta superposed upon yellow gives crimson, a zygote
containing Y, L and T is crimson, Y and L only, crimson delila.

6. Magenta upon ivory gives magenta. A zygote containing
Y, I, L and T is magenta, Y, I and L only, magenta delila.

A further variety termed ' rose ' was employed, but full details of

ANTHOCYANINS AND GENETICS 159

the results of crossings with this variety are not given, though it was
shown to be recessive to magenta.

Rose, which has the lips of the corolla tinged with magenta and a
pale magenta tube, was identified with de Vries' fleischfarbig ; magenta
with rot, delila with magenta delila, and weiss with ivory.

Hence de Vries' results may be expressed:

magenta ivory

YYIIR1R1LLTT { YYIIrlrllltt (where Rl = rose colour in
Fj magenta the lips).

I

magenta
magenta delila

F

2-

rose

rose delila I De Vries' 'whites.'

ivory

Since magenta is dominant to rose, and the tube of the corolla
always takes the same depth of colour as the lips, rose x magenta
delila would give magenta which would also corroborate de Vries'
suggestion that rot consists of fleischfarbig and delila.

In 1908 Baur (517) published some results of his work on the same
species. He worked with magenta (rot), ivory (elfenbein), yellow and
white, and his results are entirely in agreement with those published
by the author in 1907.

In 1909 the author (535) published further work and certain additional
results of which the most important are as follows:

'Rose' is now termed 'tinged ivory,' and the factors L and T repre-
sent tingeing in lip and tube respectively.

D is used to denote a deepening, or full-colour, factor which converts
tinged ivory into magenta.

It was found that zygotes heterozygous for L in the presence of D
are pale magenta in colour, whereas those homozygous for L are inter-
mediate magenta. The colour remains unaltered whether the plant
be homo- or heterozygous for D.

Sometimes the deeper colour occurs only in streaks and blotchings
when S may be substituted for D. When the zygote is heterozygous
for L, the streaks are pale magenta; when homozygous for L, inter-
mediate magenta. In non-striped forms, as mentioned above, the
flower-colour is the same whether the zygote is homo- or heterozygous
in D. But striped forms, heterozygous in S, have magenta stripings

160 ANTHOCYANINS AND GENETICS

on a ground tinged with magenta, whereas, when they are homozygous
in S, the ground colour between the stripings is ivory. The four
zygotic types may be represented as follows:

1. YYIILLTTSS Ivory striped with I magenta.

2. YYIIL1TTSS Ivory striped with P magenta.

I. YYIILLTTSs T. ivory striped with I magenta.

A. YYIILlTTSs T. ivory striped with P magenta.

In 1910 Baur (536) published a full account of his work on Antir-
rhinum ; his results are in complete agreement with those of the author.
A comparison of the two sets of results was published by the author
(548) in 1910, together with further observations on the crossing of
certain additional, true red, varieties, rose dore (on ivory) and bronze
(rose dore on yellow). Rose dore is shown to be recessive to blue-
red or magenta ; similarly bronze is recessive to crimson. Both rose
dore and bronze occur in the tinged, pale and intermediate states.

The work, published in 1913 and 1914 by Bassett and the author 1
on the pigments of Antirrhinum, has shown yellow to be identical with
the known substance luteolin, and ivory with the known substance
apigenin. The pigment in rose dore was shown to be a red anthocyanin,
and the same pigment mixed with yellow gives the colour, bronze;
similarly magenta contains a magenta anthocyanin, and this mixed
with yellow gives crimson colour.

All the varieties may now be expressed in terms of the following
factors :

Y. A factor leading to the production of luteolin in the lips and
apigenin in the tube.

T. A factor leading to the suppression of luteolin in the lips,
apigenin being formed instead.

L. A factor leading to the production of a tingeing of red pigment
in the lips.

T. A factor leading to the production of a tingeing of red pigment
in the tube.

D. A factor leading to the production of more anthocyanin pigment,
i.e. a deepening, or full-colour, factor.

B. A factor converting red into magenta anthocyanin.

1 Wheldale, M., & Bassett, H. LI., 'The Flower Pigments of Antirrhinum majus.
II. The Pale Yellow or Ivory Pigment,' Biochem. Journ., Cambridge, 1913, vn, pp.
441-444. 'The Chemical Interpretation of some Mendelian Factors for Flower-colour,'
Proc. E. Soc., London, 1914, LXXXVII B, pp. 300-311.

ANTHOCYANINS AND GENETICS 161

The different varieties may be represented as:

yyI(i)I(i)L(l)L(l)T(t)T(t)D(d)D(d)B(b)B(b) white.
YY(y)iillT(t)T(t)D(d)D(d)B(b)B(b) yellow.
YY(y)II(i)llT(t)T(t)D(d)D(d)B(b)B(b) ivory.
YY(y)iiLL(l)ttddbb yellow tinged bronze delila.
YY(y)iiLL(l)TT(t)ddbb yellow tinged bronze.
YY(y)II(i)LL(l)ttddbb -ivory tinged rose dore delila.
YY(y)II(i)LL(l)TT(t)ddbb ivory tinged rose dore.
YY(y)iiLL(l)ttDD(d)bb bronze delila,
YY(y)iiLL(l)TT(t)DD(d)bb-bronze.
YY(y)II(i)LL(l)ttDD(d)bb rose dore delila.
YY(y)II(i)LL(l)TT(t)DD(d)bb rose dore.
YY(y)iiLL(l)ttDD(d)BB(b) crimson delila.
YY(y)iiLL(l)TT(t)DD(d)BB(b) crimson.
YY(y)II(i)LL(l)ttDD(d)BB(b) magenta delila.
YY(y)II(i)LL(l)TT(t)DD(d)BB(b) magenta,

Certain varieties deeper .than intermediate magenta are known
to exist; they are recessive to intermediate magenta, but otherwise
they have been very little worked with. The heterozygous forms
have already been mentioned above. The original wild type is
thought to be intermediate magenta and would have the constitution
YYIILLTTDDBB. As would be gathered from the above statements,
ivory x white will give magenta ; hence two albinos, as regards antho-
cyanin, will give colour. Whether or no this case is strictly comparable
to the results of crossing two whites in Lathyrus and Matthiola will be
discussed later.

Arum maculatum. A. maculatum occurs in nature in two forms;
one has black (anthocyanin) spots on the leaves, the other is without
spots. Colgan (552) obtained seeds from a spotted plant, and found,
out of 11 seedlings, 5 were spotted and 6 unspotted. Colgan suggests
the female parent was heterozygous for spotted character and was
fertilised by the unspotted form, but there was no evidence in support of
the suggestion beyond the fact of equality of forms among the offspring.

Aster Tripolium. De Vries (498) notes that pigmented type is
dominant to albino.

Atropa Belladonna. Saunders (475) made crosses between A. Bella-
donna typica and A. Belladonna var. lute.a. The type has brown
(anthocyanin on plastid) flowers, black (anthocyanin) fruits, and stem
tinged with anthocyanin: the variety, which is without anthocyanin,

w. p 11

162 ANTHOCYANINS AND GENETICS

has yellow flowers and fruits, and a green stem. F x was found to have
the fruits of the type, but the intensity of colour in flowers and stems
was sometimes diminished. Further results clearly indicated that
there was normal Mendelian segregation in F 2 .

Beta. Kajanus (559, 580) has experimented with a large number
of varieties. As far as anthocyanin pigmentation in the root is con-
cerned three classes can be differentiated:

(a) Red )

;,' > With anthocyanin.

(b) Rose j

(r) White, yellow Without anthocyanin.

The results from crossing are at present rather obscure. Red was
obtained in F x from crossing :

Red x red, rose, white and yellow.
Rose x yellow.
White x white and yellow.
Rose was obtained in F x from crossing:

Rose x rose.
White x yellow.

In F 2 red gave red, rose, white and yellow, and also every selection
and combination of these varieties except red only.

Rose, on the other hand, gave red, rose, white and yellow, and again
every combination and selection except red only, and red, white and
yellow.

/

White gave:

Red, rose, white and yellow.
Red, rose and white.
Rose and white.

Kajanus suggests a number of factors, but in the absence of further
data, very little light is thrown on the results.

Kajanus notes that the pigment is not always confined to the skin
of the root. In the salad beet, the flesh is completely violet-red; in
the common red beet it is red, reddish or colourless; in the rose and
in the white it is colourless.

The leaves on the whole are green, but in red beets they are sometimes
red. Quite red leaves only occur in the case of red-fleshed beets. In
red-fleshed beets there are also varieties in which the petioles and larger
veins only are red, the leaf tissue being green. The results of crossing
red- by green-leaved varieties seems to show that in some cases the
srreen colour is due to inhibition of red, in other cases, not.

ANTHOCYANINS AND GENETICS 163

Brassica. Kajanus (580) has worked with two species B. napus
and B. rapa.

B. napus. As regards the colour of the root, distinction must be
made between upper and lower parts. Anthocyanin is only found in
the upper part, which may be violet-red, intermediate, or green. When
the root is intense violet, the neck (the lower basal portion of the stem)
is also violet-red. If the root-colour is reddish only, the neck is usually
green. Hence there are three classes :

1. Red with red neck.

2. Red with green neck.

3. Green with green neck.

It has been deduced from crossing the varieties that there are
two factors involved in anthocyanin pigmentation, i.e. P x which gives
pale violet-red, and P 2 which gives deep violet-red colour. P 2 is
dominant to P A . When both Pj and P 2 are absent, the root is green.
The differentiation between the classes is often not sharp.

B. rapa. Here also anthocyanin is only found in the upper part of the
root, which may be violet-red (deeper or paler), or if pigment be absent,
green. The violet-red colour may be continuous or blotched. It has
been deduced, from crossing varieties, that anthocyanin is due to the
presence of one dominant factor, P, and when this is absent, the root
is green.

Canavalia cnsiformis (Leguminosae). Lock (504) showed pink
colour in the flower to be dominant, or nearly so, to white, the latter
reappearing in F 2 . Absence of red pigment from the testa (the author
does not. state whether this is anthocyanin) is dominant to red. In F 2
red reappeared, but in nothing like its former intensity. Some of the
plants of F 2 bore mottled grains, but in these also the pigmented patches
were of a very faint reddish colour. In F 2 , plants with no red pigment
in the testa were more numerous, probably three times so, than those
with reddish and mottled testas taken together.

Canna indica. Inheritance of anthocyanin when red- are crossed
with green-leaved plants. Honing (608).

Cattleya. Hurst (502, 531, 540, 596) notes that purple pigment is
dominant to its absence in the flower.

Clarkia eleyans. Bateson (524) states that the magenta-flowered
type is dominant to the salmon-pink variety.

Corchorus capsularis. Finlow & Burkill (572) investigated the
inheritance of anthocyanin in the Indian Jute plant. As regards pig-
mentation 33 races were broadly classified into the following types:

112

164 ANTHOCYANINS AND GENETICS

(a) Deep red stem, petioles and fruits: the teeth of the leaves
also tipped with red.

(&) Brownish red stems, petioles and fruits, with no distinct red
borders to the leaves.

(c) Green stems with reddish petioles and fruits.

(d) Pure green stems, petioles and fruits.

The red pigment is found:

1. Chiefly in the parenchyma cells which lie immediately under
the epidermis of the stems and petioles.

2. In the parenchyma of the petioles, sporadically, even as deep
as the phloem.

3. In sub-epidermal cells near the margin of the leaf.

4. In small multicellular hairs on the leaf and on the stipules.

The intensity of coloration is due to the general distribution of the
pigment. Conversely, the fewer the pigment cells, the less red in the
stem. The authors point out that, in fact, there are only two real
colour types, red and green, since the classes (a), (6) and (c) are without
definite boundary lines.

The results of crossing red races with green were as follows. When
pure green is crossed with a fixed red, red is dominant. In Fj the
hybrids appear to consist entirely of plants of one tint of redness,
which is less dense than the colour of the red parent. The red plants
of the F 2 generation vary widely in the amount of red colour they
contain. F 3 ,, from red F 2 , though fixed as regards red colour, shows the
same variation in intensity as F 2 . As a result of the experiments
examples were produced, either fixed or unfixed, of all intermediate
colour types of jute hitherto met with, including a pure fixed culture
of one of the commonest of these.

Coreopsis tinctoria. De Vries (474) has shown that the yellow type
is dominant to the variety, 'brunnea,' in which the brown colour is
due to the development of anthocyanin. The type has evidently an
inhibitor of anthocyanin.

Ci/pripedium (Paphiopedilum). Pigmented (anthocyanin) flower is
dominant to albino. Hurst (502, 531, 540, 596).

Datura Stramonium. Saunders (475, 487) used two types, D. Tatula
having reddish stems and violet flowers, and D. Stramonium with green
stems and white flowers. The F t had red stems and violet flowers,
though the intensity of colour varied. In F 2 there was evidence of
typical Mendelian segregation.

ANTHOCYANINS AND GENETICS 165

Dendrobium. Pigmented (anthocyanin) flower dominant to albino.
Hurst (596).

Digitalis purpurea. Keeble, Pellew & Jones (542) made certain
observations on the inheritance of anthocyanin in flowers of this
species. The plants used were:

1. White with yellow spots (mmddww).

2. White with red spots (MmddWw).

3. White with red spots (MMddWw).

4. Purple with red spots (MmDdww).

5. White with purple flush and red spots (???).
Colour is due to the factor, M, producing magenta sap. Absence

of M gives a recessive white. A deepening factor, D, is dominant to
M and changes it to purple. The colour may be inhibited by a dominant
factor, W, so that the corolla is white except for red spots. The spots
on the corolla are always present. In recessive whites they are brown
or yellow: in magenta and dominant whites they are red. They
depend on the presence of the factor M and they are not inhibited by W.

Saunders (563) confirms the results of Keeble, Pellew & Jones
for the inheritance of spot colour. It is stated that white-flowered
plants with red spots may either breed true, or give a mixture of whites
with red spots (dominants) and white with greenish-yellow spots
(recessives) according as they are of pure-bred or cross-bred parentage.
Coloured flowers may vary from deep purplish-red to white with a faint
flush. The white-flowered plants with red spots frequently become
tinged as they get older.

Geum. De Vries (498) mentions the inheritance of anthocyanin in
hybrid from the cross of type (yellow plastids plus anthocyanin) by
yellow (plastids) variety.

Gossypium. Balls (515, 523) working on Egyptian cotton recognised
the following pairs of Mendelian characters which are connected with
anthocyanin :

Full red spot on the leaf and faint spot.

Large purple spot on the petal and no spot.

The red spot on the leaf is due to the development of anthocyanin
in the epidermal and sub-epidermal cells of the petiole at the point
where it divides into the leaf-veins. Crosses of spotted with spotless
give spotted F 1? but the intensity of colour in the spot is less than in
the spotted parent. In F 2 the heterozygote spot can be distinguished
from the homozygote. In the case of flower spot, the homozygote
has a large purple spot, the heterozygote a small purple spot.

166 ANTHOCYANINS AND GENETICS

Leake (561) worked with Indian cotton. Four varieties connected
with anthocyanin pigmentation were used, i.e.

1. A variety in which the petal is entirely red having a darker
spot at the .base.

2. A variety in which the petal is yellow with a deep red spot
at the base.

3. A variety in which the petal is pale yellow with a basal spot.

4. A variety in which the petal is white with a basal spot.
Leake identifies the following factors. A factor for yellow which is

dominant to white, and a factor for red which is dominant to white and
yellow. Individuals heterozygous for the reddening factor have petals
only partially coloured ; this is very obvious in red on yellow, but less so
when on white. A scheme of the factors may be represented as :

YRr red on yellow.

YRR red.

RR(r) red on white.

Y yellow.

white.

In some of the strains used there was also anthocyanin in the vege-
tative parts, that is in the young leaves, and in the ribs and veins of
the mature leaf; in other strains, these were quite green. In the Fj
from a cross between these red-foliaged and green-foliaged strains, the
red colour was dominant, though diminished in amount. In F 2 there
was segregation into red and green in the proportion 3:1. Among
the individuals with red colour there was a considerable range in
intensity, though the DD individuals had more pigment in the leaf
than the DR, and by this means they could be separated with a fair
degree of certainty. There is an association also between anthocyanin
in the vegetative parts and the complete redness of the flowers.

Helianthus. Shull (520) has made some experiments with this
genus. The wild Helianthus annuus of the Prairie region has a purple
disk, the colour being found in the tips of the paleae which are a deep
metallic purple, the margin of the corolla which is brownish-purple
and the style and stigmas which are reddish-purple. The 'Russian
Sunflower' (Helianthus annuus, var.) has the tips of the paleae yellowish-
green, the corolla a clear lemon yellow, and the styles and stigmas
usually have the same colour as the corolla. Shull concluded, on the
results of crossing, that the purple disk is a strict Mendelian character
and is dominant to the yellow disk.

The red sunflower mentioned by Cockerell (602, 603, 611) appears

ANTHOCYANINS AND GENETICS 167

to be dominant to the yellow, though the F x may vary in the amount
of anthocyanin it produces. The red variety is no doubt formed
owing to the loss of an inhibiting factor, and the F a plants would only
receive the inhibitor from one parent. When anthocyanin appears
in the primrose-coloured variety of Helianthus, the result is purple
and not chestnut red.

Hordeum. BifEen (501) mentions that in Hordeum the paleae may
be white, black, brown or purple. The grain also may be white, bluish-
grey or purple. Apparently the purple colour is due to authocyanin.
According to BifEen, purple paleae as contrasted with white, and dark
grain with light grain, form pairs of Mendelian characters.

Laihyrus odoratus. Bateson & Punnett (487, 496, 500, 516) have

carried out extensive work on the colour inheritance in the Sweet Pea.

The original wild type is probably most nearly represented by

the variety now T known as 'Purple Invincible' with chocolate standard

and bluish-purple wings.

Loss of a diluting factor produces a variety with deeper wings,
' Purple- winged Purple Invincible.' Loss of a full-colour factor gives
a tinged variety, 'Picotee.'

When the blueing factor is absent, a series of red varieties appears
comparable to the above: 'Painted Lady/ with a deeper variety,
'Miss Hunt,' and a tinged variety, 'Tinged White.'

It was shown early in the experiments with Sweet Peas that two
white varieties, indistinguishable except that one has long, the other
short pollen, gave a ' Purple Invincible ' hybrid, and from this result the
fact was deduced that colour production is dependent on two factors,
or that the two factors taking part in its formation can be inherited
independently. As in Matthiola, we must suppose that two factors
produce the most hypostatic colour, i.e. 'Tinged White,' and that 'Purple
Invincible ' resulted in the original cross because the white varieties
employed carried both B and a full-colour factor.

These facts may be represented in tabular form as follows (De == full-
colour factor, Di == diluting factor) :

CRBDeDi Purple Invincible.

CRBDe Purple- winged P.I.

CRBDi Picotee (see Bateson & Punnett, 498).

ORB Picotee.

CRDeDi Painted Lady.

CRDe Miss Hunt.

CRDi Tinged White.

CR Tinged White.

168 ANTHOCYANINS AND GENETICS

Any plant without C or R is white. Hence whites can carry any
of the other factors in any combination or arrangement, but never
C and R simultaneously. There is also a certain connection between
colour and the hooded form of the standard. In hooded varieties the
standard always approaches in colour to the wings, and has never the
bicolor appearance of the varieties with erect standard.

M. G. and D. Thoday (547) have given an account of experiments
with certain varieties of Lathyrus, and the crosses form a very complex
series in F 2 . The chief point of interest is that they find a scarlet
anthocyanin, distinct from, and recessive to. the bluish-pink of the
' Painted Lady ' variety.

Linaria alpina. Saunders (586) made crosses between the type and
a variety. The type has purplish- blue flowers and the variety pink
flowers. Both pigments are anthocyanins. It was found that blue
is dominant to pink.

Linum usitatissimum. Varieties of this species have been investi-
gated by Tammes (564). Among other species, four varieties of
L. usitatissimum were used, i.e. common flax, Egyptian flax (deep
blue flowers), and two other varieties, one with pale blue, the other
with white, flowers. Among other crosses are the following which are
fairly typical.

From Egyptian flax x white-flowered variety of common flax, the
F! had paler flowers than Egyptian flax. In F 2 there were 8 pale
blue, 3 pure white and 3 like Egyptian flax. This was regarded
as an ordinary Mendelian ratio of 1 : 2 : 1. From the white-flowered
variety x pale blue-flowered variety of common flax, the F x was fairly
constant, and still paler than the blue parent. In F 2 , out of 39 plants,
11 were pure white; the remainder were a mixture of pale blues like
the parent and grandparent, but were difficult to separate. The
author however believes the ratio to be again 1:2:1.

In a later paper (610) Tammes shows that in the second and following
generations from a cross between the Egyptian and the white-flowered
varieties of common flax there is a deficiency of white-flowering plants.
This is found to be due to two causes, i.e. first, a deficiency of seed
formed from the mating of gametes without the factor for forming
colour, and secondly, an inferior germinating power of the seed of the
white-flowered individuals.

Lychnis dioica. Saunders (475) used this species in certain crosses.
L. dioica was crossed with a glabrous form of L. vespertina, and a
glabrous variety of L. dioica was crossed with L. vespertina, F x and
their offspring showed pink, the depth of colour varying according

ANTHOCYANINS AND GENETICS 169

as the wild hairy species or de Vries' glabrous strains (see original
paper) were employed in the cross.

Shull (520) first mentions the results of crosses between white- and
purple-flowered Lychnis dioica, the purple being dominant.

Later, Shull (546) published further results, in which he states that
two varieties of coloured flowers can be detected in Lychnis, i.e. a
reddish- and a bluish-purple. The former is dominant to the latter,
this being the reverse of what is found in other plants. Two factors
are necessary for formation of the bluish-purple, and a third factor
modifies this to reddish-purple.

In a still later paper Shull (588) again states that the type has reddish-
purple flowers, and that there is a bluish-purple variety recessive to
the type. The albino is without anthocyanin. No albinos mated
together produced colour. Two new German strains were introduced
Melandrium album and M. rubrum. A certain individual of M. album
x white dioica gave reddish-purple offspring. M. rubrum x M. album
gave a mixture of both purple- and white-flowered offspring in the
proportion of 4 : 23.

Matthiola. Our knowledge as regards the inheritance of colour
in this genus is due to the work of Saunders (475, 487, 496, 500, 506,
562).

It is not known what variety represents most nearly the original
type, but if it be assumed that each variety arises by loss of some
factor from the type, then the latter would be represented by a plant
with pale purple flowers. Loss of a diluting factor gives rise to deep
purple; loss of another factor from the deep purple gives a duller shade
of purple termed 'plum.'

Loss of the blueing factor B -from each of the above varieties
gives rise to the corresponding blue-red series; rose, a dilute variety;
carmine and crimson, deep varieties; and 'copper,' a dull red variety
represented by plum in the bluer series.

Loss of a further factor from the blue-red class reveals a true, less
blue, red class containing a dilute variety, 'flesh,' and a recessive
deeper variety, 'terra-cotta.'

Early in the experiments with Matthiola it was ascertained that
two factors are necessary for the production of colour, and that certain
white varieties crossed together produce coloured offspring, purple in
the original experiment, since one at least of the original whites used
contained the blueing factor, B.

The varieties may be represented in the following scheme. C and

170 ANTHOCYANINS AND GENETICS

R are the factors for colour, B modifies the blue-red class to the purple
class, D causes dilution in colour and X the difference between the
pure and dull, or impure, colour:

CRBDX pale purple.

CRBX deep purple.

ORB plum.

CRDX pale red (rose).

CRX deep red (carmine, crimson).

CR copper.

C white.

R white.

In some families, other varieties, such as terra-cotta, flesh and
lilac, appear in addition, but, apart from the fact that they are recessive
to crimson and purple, their relationship to each other and to the
other colours is not clear at present. As we see from the table, indivi-
duals without either C or R are white; thus white can carry every
factor and combination of factors, except C and R simultaneously.
Many of the above results have been confirmed by Tschermak (590).
Mirabilis Jalapa. Correns (482, 497, 537) published the first work
on the colour-inheritance in this genus. He used several varieties,
white, yellow and 'red.' In F x he obtained a 'rose' by crossing white
x yellow, but as he did not realise the existence of heterozygous forms
(see below), he was unable to solve successfully his results in the F 2
generation.

Baur (517) made crosses with two varieties, and obtained a hetero-
zygous form in Fj, but his results are similar to those of Marryat (533)
to whom we owe the bulk of our knowledge on the inheritance of flower-
colour in Mirabilis. The work of the latter author leads to the following
conclusions :

We must suppose that the original type had crimson flowers. Loss
of a factor for anthocyanin production from crimson gives a variety
with yellow flowers free from anthocyanin. Further loss of a factor
produces white, an albino both as regards red and yellow pigment.
The peculiar interest of Mirabilis (among other points) is centred in
the occurrence of heterozygous forms which may be best represented
by the following scheme:

CCMM crimson.

CcMM magenta.

CcMm magenta rose.

CCMm orange red.

ANTHOCYANINS AND GENETICS 171

CCmm yellow.

Ccinin pale yellow.

ccMM white.

ccMm white.

ccmm white.

It was also found that a certain two white individuals crossed
together gave coloured (anthocyanin) F x , as in Matthiola and Lathyrus.
Hence the factors for anthocyanin production can be separated into two
components, of which one is M, and the other is not represented in the
above scheme. Absence of colour may be due to loss of either of these
components, or to loss of the yellow pigment.

Nicotiana. Inheritance of anthocyanin in species-crosses. Lock
(532), Haig Thomas (594).

Oenothera. Since Oenothera is a plant which forms anthocyanin
in its stems, petioles, buds, etc., the complex inheritance of characters
among the numerous strains which have been employed experimentally
involves also the inheritance of this pigment. Mention will only be
made of one or two cases in which anthocyanin pigmentation has been
considered an important character.

The first case to be dealt with is that introduced by Gates (539,
555, 573), and which concerns the appearance of 0. rubricalyx. This
variant was found among the offspring of self-fertilised rubrinervis
plants, the latter being characterised by having the calyx of the buds
streaked with anthocyanin, whereas 0. rubricalyx has a completely
red calyx. The segregation of the offspring from self-fertilised rubri-
calyx plants in two generations, into rubricalyx and rubrinervis, led
Gates to consider the case purely Mendelian. The original rubricalyx
plant was regarded as a heterozygote which has acquired a dominant
Mendelian character, the character being purely quantitative, i.e. as
causing an increased formation of anthocyanin.

In a later paper (606) Gates again states that the red pigmentation
character, R (which originated by a mutation and distinguishes rubri-
calyx from rubrinervis), is more or less completely dominant in F, from
the cross grandiflora x rubricalyx and its reciprocal. In F 2 , 3 : 1 ratios
were obtained, and also ratios of 5 : 1 and 10 : 1 as well as 3 : 1. Gates
has no satisfactory explanation of these facts.

Shull (609) makes further experiments on rubricalyx by selfing
this strain and crossing it with rubrinervis and with Lamarckiana ;
he obtains what he calls a series of negative correlations in the distri-
bution of the red pigment, the pigmented buds of rubricalyx being

172 ANTHOCYANINS AND GENETICS

invariably associated with a low degree of pigmentation in stems and
rosettes. Moreover, the segregation of the rubricalyx character was not
found to be a simple Mendelian case. Shull maintains also that certain
of Gates' conclusions are erroneous, viz. that the rubricalyx character
represents a quantitative difference, and that it can be expressed by one
factor.

The other case of interest in connection with anthocyanin is that
considered by Davis (592, 604) in connection with the stem coloration
(the formation of papillae or glands coloured with anthocyanin at the
base of long hairs) in parents and hybrids of crosses, 0. grandiflora
x biennis, 0. franciscana x biennis and their reciprocals. Davis con-
siders this character to be dominant to the green stem, but it has
not been shown to segregate in a Mendelian way.

Oxalis. Nohara (635) worked with varieties of the so-called Oxalis
corniculata L. which differed from each other in the presence or absence
of purple (anthocyanin) in the eye of the corolla and in the leaves.
The presence of anthocyanin was found to be dominant to its absence,
and the intensity of pigmentation in Fj from pigmented by unpigmented
was found to be intermediate. The purple colour in eye and leaf is
due to one factor so that eye- and leaf-purples are associated, but the
leaf-purple can appear without the eye-purple.

Papaver somniferum. De Vries (474) and Hurst (502) have shown
that the basal patch on the petals is dominant to its absence. Hurst
(502) also states that colour in the rest of the petal is dominant to
albinism, and that purple is dominant to red.

Papaver Rhoeas. Shull has published work (588) on this genus.
It was found that varieties with a white margin were dominant to
varieties without the margin, i.e. the type. Certain whites crossed with
some reds gave white or striated offspring, but the same whites were
found to be recessive to pink or orange. Some red-flowered varieties
crossed together gave whitish offspring. Two suggestions are made:
(a) that only spectrum red is inhibited ; (6) that two factors are necessary
for inhibition.

Phaseolus multiflorus. The characters which have received most
attention in this genus are those concerned with the pigmentation of
the seed-coat.

Lock (504) made some preliminary experiments by crossing a dark
purple-seeded bean with a dark yellow-seeded bean. The F x was dark'
purple. The F 2 could be subdivided into two groups: (A] containing
beans of various shades of purple; (B) of various shades of yellow.

ANTHOCYANINS AND GENETICS 173

It was found that members of (.4) might throw (B) but not vice
versa.

Shull (513) gives an account of the results obtained by crossing
several varieties, i.e.

'Prolific black wax' purple-black seeds (anthocyanin in testa).

'Ne plus ultra' yellow-brown seeds.

'Long yellow six- weeks' light greenish-yellow seeds.

'White flageolet' seed coats white.

The results in Fj were as follows:

Purple x yellow-brown = purple.
Purple x yellow purple.
White x purple

All gave similar F, with testa mottled
White x yellow-brown \ .,

with purple.
White x yellow )

Shull then postulates the following factors:

P = pigment.

B = modifier which changes pigment to purple.

M = mottling factor.

Then the constitution of the different beans is
Brown and yellow Pbm.
Black beans PBm.
White pBM.

Shull concludes that the mottling factor is carried by the white
bean, whereas Bateson and Tschermak had regarded the mottling factor
as latent in the pigmented bean.

Shull in a later paper (521) gives the proportions of the varieties
in F 2 from the above cross. They were found to be

Purple mottled 18.
Purple self-colour 18.
Brown or yellow mottled 6.
Brown or yellow self-colour 6.
White 16.

Shull's explanation for this result is that beans containing PB and
heterozygous in the M factor are mottled, whereas those beans
homozygous in M are self-coloured. Hence it is possible for purple to
carry the mottling factor, and Tschermak, Bateson and Lock are also
correct. Mottled beans of the above constitution are heterozygous
and should never breed true, and this was found to be the case. Shull

174 ANTHOCYANINS AND GENETICS

points out that there are however other strains of mottled beans which
do breed true.

Emerson (528) publishes a long paper on pigmentation in bean seeds.
He gives first a list of the crosses made by Tschermak, Shull and himself.
He points out, as Shull has done, that there are two kinds of mottled
beans, viz., strains which breed true, and heterozygous forms not
breeding true. Emerson then suggests a scheme to explain the existence
of two sorts of mottling, namely by postulating two factors for mottling :
M the sort of mottling which breeds true, and X, the sort which is visible
only in the heterozygous condition. The results he obtained experi-
mentally can be explained on the basis of this scheme. In .a later
paper (529) Emerson gives further results of the inheritance of total
and eyed (round hilum) pigmentation, but since the kind of pigment
is not described, it is not clear how far it concerns anthocyanin.
Emerson also mentions another hypothesis suggested to him by Spillman
to explain the two kinds of mottling mentioned above. Spillman sup-
poses that the mottled races which breed true have in them two corre-
lated factors, and that there are three types of non-mottled beans
resulting from the 16ss of one, or the other, or of both of the correlated
factors. On Spillman's plan, just as on Emerson's two factor hypothesis,
a definite formula can be assigned to all the races used in crossing, and
in this way all the results, with one or two exceptions, can be accounted
for. Emerson himself inclines to the coupled-factor, rather than the
independent-factor, hypothesis.

Pliyteuma. Inheritance of anthocyauin in the cross Ph. Halleri
x Ph. spicatum. Correns (486).

Pisum sativum. Here again as regards pigmentation, the colour
of the testa has received a great deal of attention. Bateson & Killby
(487) noted that greys and browns in the seed-coat are associated with
coloured flowers, and crossed with whites they occasionally give 'rever-
sionary' Fj with purple (anthocyanin) spots, though spots were absent
from the parents.

Lock (512, 518) sums up many of the results on pea colour. He
states that the albino variety has no anthocyanin. The presence of
a factor C produces grey (chromogen) in the testa, and red anthocyanin
in the leaf axils and flowers. The presence of S produces spotting of
a reddish shade on the testa. The presence of P modifies red anthocyanin
of the axils, flowers and spots on the testa to purple.

Polemonium. Inheritance of blue anthocyanin in F t from cross of
P. caeruleum by P. flavum. Correns (486).

ANTHOCYANINS AND GENETICS 175

Primula sinensis. The bulk of the work on this genus is due to
Gregory & Bateson. The following is a summary of the results which
have been published by Gregory (557).

All colour in P sinensis is due to anthocyanin except the yellow
of the eye which is plastid pigment.

Varieties of flower-colour.

Albinos: The flowers contain no anthocyanin; they are differen-
tiated into dominant and recessive whites, but the flowers of both
look alike.

Full colours:

Salmon-pink.

True red ' Orange King.'

Blue reds Light (dilute) and deep shades of magenta and crimson.
Crimson is less blue than magenta but is more blue than 'Orange King.'

Blues various shades.

Pale colours : Shades of pink, of which ' Reading Pink ' is the deepest ;
of these apparently some correspond to magentas, and others to crimsons,
but the difference cannot be detected except by crossing.

Distribution or pattern of colour.

'Sirdar.' The pigment of the petals occurs in separate minute dots,
and the edges of the petals are white. The pigment itself may be either
magenta, crimson or blue.

'Duchess.' The pigment occurs as a flush round the eye of the
corolla, and may be either magenta or crimson.

Colour in the spot external to the eye.

In certain varieties, there are spots of deep colour on the petals
just external to the eye. Deep spots are not fully developed unless
the stigma is coloured ; nor even if the stigma is coloured, are they
developed in plants which have no yellow (plastid) eye. Spots are
deeply coloured only in deeply coloured flowers; on light flowers
they are similar to those in the flowers with a green stigma. They
also depend on the base colour, since they are not visible in pale
coloured flowers, nor in flaked flowers unless the stripe occurs on the
area occupied by the spot.

Colour in the ovary, style and stigma.

By loss of an inhibiting factor, colour may be formed in these organs.

176 ANTHOCYANINS AND GENETICS

Varieties of stem-colour.

Anthocyanin may be entirely absent when the stem is green. This
condition is associated with recessive white flower-colour (occasionally
dominant white, i.e. 'Pearl'), or with pale colours of which the deepest
is 'Reading Pink,' and no flower-colour deeper than 'Reading Pink'
is ever found on plants devoid of anthocyanin in the stem.

There may be faint colour in young leaves and petioles associated
with pale flower-colour.

There may be pigment at the bases only of petioles and pedicels,
associated with 'Sirdar' flower-colour.

The stem may be fully coloured with purplish-red sap, either light
(dilute), or deep, associated with full flower-colours, i.e. salmon-pink,
crimsons and magentas; but whereas light magentas and crimsons
can be borne on both light- and deep-stemmed plants, the deepest
magentas and crimsons can only be borne on deep-stemmed plants.

The stem may be coloured with pure red sap associated with ' Orange
King' flowers.

The stem may be coloured with blue sap associated with blue flowers.

Hence the direct relationship between stem and flower-colour may
be summed up as

Recessive whites and green stem.

Pale flower-colours and green, or faintly coloured, stem.

Full flower-colours and full-coloured stem (and of these deep flower-
colour and deep stem).

True red ('Orange King') flower-colour and true red stem.

Blue flower-colour and blue stem.

Factors (stems).

Colour in both flowers and stems can be produced by two or more
complementary factors. Keeble & Pellew (541) obtained an Fj with
magenta flowers from two whites. As regards stem, there is no case
of colour from mating two greens, but plants heterozygous for colour
in the stem have given 9 coloured : 7 green stems.

Slight colour in the stem is due to the factor Q, which is dominant
to complete absence of colour.

The difference between full colour and faint colour is due to a
dominant factor, R.

The difference between 'Sirdar' and full colour is due to a factor,

ANTHOCYANINS AND GENETICS 177

F, which regulates distribution of colour. Thus an individual hetero-
zygous for all three factors will give

9 FRQ...full colour (magenta or crimson).
3 RQ ...Sirdar (magenta or crimson).

3 Q faint colour (pinks).

1 - - green (whites).

Purplish-red in the stem is dominant to the true red of 'Orange
King,' and one factor only is necessary to produce the change.
Blue in the stem is recessive to all colours.

No case is known of inhibition in the stem, that is, there is no domi-
nant green stem.

Light shades in the stem colour (flowers light) are dominant to
deep shades (flowers light or deep). Probably one factor is concerned
with heterozygous forms, but there may be more.

In some cases there is one factor between crimson and magenta;
other cases indicate two factors.

It is doubtful whether there exists one or more diluting factors.
There is one factor diluting stem colour and another flower-colour,
and these are inherited independently.

Factors (flowers}.

Colour in flowers may be produced by two or more complementary
factors.

The difference of one factor regulates the distribution of colour
between ' Sirdar ' and full colour, and also between pale and full colour
as in stems.

Full colours are dominant to pale colours.

Magenta is dominant to crimson.

Magenta and crimson are dominant to full red ('Orange King').

Blue is recessive to all magentas and reds.

Whites may be dominant or recessive to colours.

Dominant whites are generally red-stemmed. A green-stemmed
white variety, 'Pearl,' is a dominant white (Keeble & Pellew, 541);
the same authors also report that a full red-stemmed white, 'Snow
King,' may be recessive. The flower is only white in dominant whites
if homozygous for the inhibiting factor. It has also been suggested
that suppression of colour is due to two inhibiting factors, central and
peripheral, one of which is represented in ' Duchess.'

Keeble & Pellew (541) provide further results on dominant whites.
Hitherto all whites with red stems have been regarded as dominant

w. p. 12

178 ANTHOCYANINS AND GENETICS

whites. Keeble & Pellew give evidence for regarding ' Snow King,'
which is a red-stemmed variety with white flowers, as a recessive white.
Crosses between various plants of ' Snow King ' and fully coloured
varieties showed that ' Snow King ' may be homo- or heterozygous for
the dominant white factor, or may be entirely without it.

'Snow King' x 'Snow Drift' gave magenta Fj. Hence these two
white-flowered varieties can give colour.

Reseda odorata. Compton (569) has shown that orange-red colour
(anthocyanin) in the pollen is dominant to bright yellow ; self-fertilised
heterozygotes throw about three reds to one yellow.

Salvia Horminum. Saunders (487) has worked with the type and
varieties of this species. The type has violet flowers, and the bracts
of the inflorescence are also coloured violet. A red variety was used
in which the flowers and bracts were red. Both pigments are antho-
cyanins. Loss of pigment gives an albino from w y hich anthocyanin
is completely absent.

The inheritance can be represented by the following scheme:

BR .purple.

R red.

B white.

white.

Senecio vulgaris. Trow (589) has published results of crossing a
number of elementary species. It was found that red colour, antho-
cyanin, in the stem is dominant in some elementary species. It is
suggested that one factor is involved, or possibly two.

Silene Armeria. Colour in flower dominant to albinism (de Vries,
498).

Sclanum tuberosum. East (538) has published some results. He
states that presence of anthocyaniu in the stem is dominant to its
absence, and segregates on Mendelian lines. Purple in the flower is
probably dominant to its absence. The colour of the tubers is either red
or purple, and purple is dominant to red.

Salaman (544) has published more extensive results. Black tubers
of the variety, 'Congo,' have purple anthocyanin in the skin. The
red-tubered variety has red anthocyanin. White tubers have no antho-
cyanin. Two factors are required for colour (red), and a third dominant
factor gives purple. S. eiulerosum has a dominant white factor inhi-
biting the purple.

In the flowers all pigment is anthocyanin. Heliotrope colour is
due to two factors, and purple to a third factor. In varieties of

ANTHOCYANINS AND GENETICS 179

tuberosum, colour is confined to the upper surface. In S. etuberosum,
pigment is developed on the lower, and is inhibited on the upper surface.
In S. verrucosum both surfaces are probably free from inhibitors.

Trifolium. Inheritance of anthocyanin in flowers and seeds. Red
in flower-colour is dominant to both white and blue (Raj anus, 579).

Veronica longifolia. Colour (anthocyanin) dominant to its absence
(de Vries, 498).

Viola cornuta. Colour (anthocyanin) dominant to its absence
(de Vries, 498).

Zea Mays. Correns (476) made the first investigations on this
species. As regards characters connected with anthocyanin he states
that two such characters behave in a Mendelian way, i.e. colour in the
pericarp (red or absence of red), and colour in the aleurone layer (blue
or white). The degree of dominance was found to vary.

Lock (493, 504) has published further results. He states that blue,
black or purple pigments are confined to cells of the aleurone layer.
Red is confined to the pericarp. The pigment is situated in the cells,
and to some extent also in the cell walls of the pericarp. He points
out that in cross-bred cobs, blue will occur mixed since the character
is not maternal; the pericarp colour, on the other hand, appears in
either all or none of the grains, as it is a maternal character. The
work includes a large number of results in connection with the cross,
blue x white and the reciprocal. The results show that there is
irregularity in the dominance of blue. Red colour in the pericarp
is dominant to its absence, and forms a simple Mendelian case.

East & Hayes (553) have published extensive results on Maize.
They show that Lock's irregular results were due to the fact that the
whites he employed in crossing were carrying different factors. They
themselves worked with a variety having red (anthocyanin) pigment
in the aleurone layer, and the formation of this pigment was found to
be due to two factors, C and R. A third factor, P, modifies the red to
purple. There is also a fourth factor, I, which inhibits the red and
purple colour. Red pigment in the pericarp, cobs and silks, is dominant
to its absence. It may be present in each of these parts separately
and independently of the others. No plant has been obtained which
has red glumes and yet shows no red colour in other parts of the plant.
One, however, has been found that is pure for red glumes, and shows
no red in other parts with the exception of the silks.

122

180 ANTHOCYANINS AND GENETICS

CONNECTION OF FLOWER-COLOUR WITH THE PRESENCE OF ANTHO-
CYANIN IN VEGETATIVE ORGANS, FRUITS AND SEEDS.

Such cases are illustrated among the following genera and species:

Antirrhinum ma jus, Helianthus,

Atropa Belladonna, Lathyrus odoratus,

Beta, Oenothera,

Corchorus capsularis, Pisum sativum,

Datura, Primula sinensis,

Digitalis purpurea, Salvia Horminum,

Gossypium, Zea Mays.

There appear to be several possibilities in the inheritance of colour
in such cases as these we are considering. Colour (anthocyanin) in
the plant may be represented by one factor which causes the flowers,
as well as other organs, to develop pigment. When this factor is
absent, pigment disappears from the whole plant including the flower ;
this is perhaps the most common case. Colour, however, may disappear
from the flower, or any other organ, independently of the rest of the
plant. One of the most striking examples of independent connections
of this kind we find in Maize ; the plant may have pigment in the sheath,
cob, pericarp, silks, etc., and practically in all or any of these parts.
In such a case, if colour in each part is represented by a separate factor,
we should expect by crossing suitable plants to obtain all the Mendelian
combinations. This, however, is not always so, as for instance in Maize
(Emerson, 554), and a number of other plants. The lack of some
combinations may be due either to reduplication phenomena (see
p. 185), or to some form of genetic correlation, and to decide between
these alternatives often involves further data than we have at present.
In the following genera indication is given of such colour relationships
as we know to exist:

Antirrhinum majus. The white ; ivory and yellow-flowered varieties
never produce anthocyanin in any part of the plant. The author
has noticed that the stems and leaves of deep magenta- and crimson-
flowered varieties produce anthocyanin to a considerable extent,
whereas in intermediate and pale varieties these organs are practically
free from anthocyanin.

Atropa Belladonna. The type has anthocyanin in the flowers
(brown), fruits (black) and stems (tinged with red). The albino has
yellow flowers and fruits, and a green stem (Saunders, 475).

ANTHOCYANINS AND GENETICS 181

Beta. Some varieties have leaves coloured quite red with antho-
cyanin. This only happens in red-fleshed varieties, i.e. when antho-
cyanin is formed in the inner tissues, as well as in the skin, of the root.
In varieties having colourless flesh, the leaves are green (Kajanus, 559).

Corchorus capsularis. Varieties with deep red (anthocyanin) stems,
petioles and fruits have the teeth of the leaves also tipped with red.
Other strains with less pigment in the stem have no colour in the leaf
teeth (Finlow & Burkill, 572).

Datura. D. Tatula has violet flowers and red stems (anthocyanin).
D. Stramonium has white flowers and green stems (Saunders, 475, 487).

Digitalis purpurea. The presence of a factor, M, for magenta pig-
ment always brings with it colour in the spots on the lower inner surface
of the corolla. Even in the presence of M the general colour in the
corolla may be inhibited by another factor I, but the spots remain
red (var. white spotted with magenta). If M is absent, the spots are
colourless (Keeble, Pellew & Jones, 542; Saunders, 563).

Gossypium. In the Indian cotton, strains having yellow flowers,
or white flowers with only a basal spot on the petals, have green foliage.
In strains having red pigment, the varieties having red flowers (forms
homozygous for the reddening factor, see p. 166) have the veins and
lamina of their leaves suffused with red. In leaves of the plants which
have flowers red on yellow, or red on white (forms heterozygous for the
reddening factor) the anthocyanin is confined to the veins (Leake, 561).

Helianthus. In the wild H. annuus the disk is purple, the colour
(anthocyanin) being in the tips of the paleae, the margin of the corolla
and the styles and stigmas. The 'Russian Sunflower' has the tips of
the paleae yellowish-green, the corolla, styles and stigmas yellow
(Shull, 520).

The chestnut-red-flowered Sunflower, which has anthocyanin on
yellow in the ray florets, has dark reddish-purple stems. The purple-
flowered variety, anthocyanin on primrose, has no reddish-purple in
the stems (Cockerell, 611).

Lathyrus odoratus. The albinos always have green leaf-axils.
When a certain factor for production of colour in the axils is present
and the flowers are coloured, there is a reddish-purple spot in the axil ;
when the axil colour factor is absent, the axils are green (Bateson, 524).
Red tendrils are (?) always associated with red in the axils of leaves
(Bateson, 499).

Oenothera. In the variant Oenothera rubricalyx red buds and
hypanthia are said to be associated with red stems (Gates, 555). In

182 ANTHOCYANINS AND GENETICS

the Fj, however, from reciprocal crosses between 0. rubricalyx x 0.
Lamarckiana, plants appeared in which red buds were associated
w r ith a low degree of red pigmentation in stems and rosettes, whereas
pale red buds and green hypanthia were associated with brilliant red
stems, and buds entirely free from anthocyanin with dark stems
(Shull, 609).

Pisum sativum. White-flowered plants have no colour in the leaf-
axils and no red or purple (anthocyanin) colour in the testas. Red-
flowered plants have red colour in the leaf-axils and red spots on the
testa. Purple-flowered plants have purple colour in the leaf-axils
and purple spots on the testa (Lock, 518).

Primula sinensis. Recessive white-flowered varieties may have
green or red stems (Keeble & Pellew, 541). Pale-flowered varieties
may be on green or faintly coloured stems; deep-flowered varieties
appear on deep red stems only; whereas light-flowered varieties may
be borne on either deep- or light-stemmed plants. The 'Sirdar' variety,
in which the pigment is in minute dots and the edges of the petal are
white, has stems with a red 'collar,' i.e. base of stem. If the flower-
colour is crimson or magenta, the stem colour is purplish-red ; if the
flower is true red ('Orange King') the stem pigment is also true red.
Blue flowers have blue stem pigment.

The deep spots of colour external to the eye are not fully developed
unless the stigma is coloured. They are deeply-coloured only in
deeply-coloured flowers : in light flowers they are similar to those in
flowers with a green stigma. They also depend on base colour as they
are not present in pale flowers, nor in striped flowers, unless the stripe
occurs in the area occupied by the spot (Gregory, 557).

Salvia Horminum. The bracts of the inflorescence are purple,
red, or green according to whether the flowers are purple, red, or white
(Saunders, 487).

Zea Mays. Red pigment may appear in the pericarp, cobs and silks
(stigmas) separately and independently in each part. If the glumes
are red, red is present in some other part of the plant, though it may
only be in the silks (East & Hayes, 553; Emerson, 554).

The various relationships between flower-colour and the presence
of anthocyanin in the vegetative organs may therefore be classified
as follows:

1. Simple case. Loss of colour in the flower accompanied by total
loss from vegetative organs. Ex. : true albinos of Antirrhinum, Atropa,
Lathynts, Pisum, Primula, etc.

ANTHOCYANINS AND GENETICS 183

2 (A). Loss of the B factor from the flower will give red antho-
cyanin. Accompanying this there is a corresponding change in the
pigment of the stem and leaves. Ex. Antirrhinum: rose dore and
bronze varieties ; the same red pigment is developed in the stem and
leaf. Primula sinensis: 'Orange King'; the same pigment appears
in the stem and petioles. Salria Horminum : red-flowered plants have
red inflorescence bracts. Pisum : red flowers are accompanied by red
spots on the testa and red leaf-axils.

2 (B). Blue-flowered varieties of Primula sinensis develop blue
anthocyanin in the stem.

3. In pale varieties (flower-colour), there is less anthocyanin in
the stems and leaves. Ex. Primula sinensis : pale varieties and faintly
coloured stems; light-flowered varieties and light stems; 'Sirdar'
flower-colour and the pigment in the collar of the stem. Gossypium
(Indian): plants with petals with a basal spot have green leaves;
plants with red petals have leaves suffused with red.

4. Deep varieties. Ex. Antirrhinum: the deepest flower-colours
have red leaves and stems. Primula sinensis: deep flower-colours
have deep red stems.

Then follow the more complex cases in which pigmentation in
different organs appears to be inherited independently, either partially
or completely. Ex. Helianthus: flower-colour apart from stem colour
in the purple variety. Lathyrus odoratus : coloured flower and coloured
leaf-axil; yet at the same time there may be coloured flowers and a
colourless leaf-axil. Oenothera: red bud pigmentation with green
stems; green buds with red stems. Primula sinensis: white flowers
and red stems. Zea Mays : independent inheritance of colour in cobs,
pericarp, silks, etc.

HETEROZYGOUS FORMS.

Just as in the case of factors for other characters, so also in the
case of factors concerned with pigmentation we find heterozygous
forms. The one of most frequent occurrence is that due to the inheri-
tance in the zygote, from one parent only, of the factor which actually
produces colour (anthocyanin). We find it, for instance, in the following
genera :

In Atropa Belladonna: when the red-stemmed, brown-flowered
type is crossed with the green-stemmed, yellow-flowered variety, the
colour in stems and flowers of F x is less intense than the parent (Saunders,
475).

184 ANTHOCYANINS AND GENETICS

In Corchorus capsularis : when deep red-stemmed varieties are
crossed with green-stemmed varieties, the colour in stems of F t is less
intense (Finlow & Burkill, 572).

In the cross Datura Tatula, with red stems and purple flowers, by
D. Stramonium, with green stems and white flowers, the colour in
flowers and stems in Fj is less intense (Saunders, 475).

In Egyptian cotton, Gossypium : when the variety with a spot on
the leaf is crossed with the non-spotted variety, the spot in F a is paler
in colour (Balls, 515, 523).

In Indian cotton : when strains with anthocyanin in the vegetative
parts are crossed with green-leaved strains, the colour in leaves and
stems in F x is paler (Leake, 561).

In Linum usitatissimum : a cross between a blue-flowered and a
white-flowered variety gave an F x with paler blue flowers than the
parent (Tammes, 564).

On the other hand, in many cases where there are factors producing
colour (anthocyanin) no heterozygous forms at all exist, as in Lathyrus
and Matthiola.

Cases, fundamentally different, though giving a similar result to
those above, are those in which the type has anthocyanin inhibition
by an inhibiting factor, and the variety has lost the inhibitor and
produces anthocyanin. When the type is crossed with variety, the Y l
is heterozygous for the inhibitor and is less intensely coloured than the
parent. Examples of such are the F x from Coreopsis tinctoria with
yellow (plastid pigment) flowers and the variety brunnea (anthocyanin
on plastids) (de Vries, 474) ; the Fj has paler flowers than brunnea ; also
the F x , probably, from the yellow Sunflower by its chestnut-red variety.

Somewhat similar, though not strictly comparable, is the case in
Primula sinensis where whites tinged with anthocyanin are obtained
in F x from the cross 'dominant white x fully-coloured variety (Gregory,
557).

In connection with the class mentioned above which is heterozygous
for the factor for colour, Antirrhinum offers an interesting illustration.
Colour, i.e. anthocyanin, in Antirrhinum is produced by the action of
one factor, L, which gives ivory tinged with magenta, but there is no
apparent difference in the flower-colour between individuals of the
composition LL and LI. In the presence of an additional factor D,
a deepening factor, the zygote develops more pigment and is a deeper
colour, but the ultimate colour depends on whether the zygote is homo-
zygous or heterozygous for L, its condition as regards D having no

ANTHOCYANINS AND GENETICS 185

effect; thus LlDD(d) is pale magenta, and LLDD(d) is intermediate
magenta.

A second heterozygcms form in Antirrhinum is also interesting.
Individuals heterozygous for the striping factor have magenta stripes
on a tinged ground, individuals homozygous for the striping factor,
magenta stripes on an ivory ground (Wheldale, 535).

Mirabilis Jalapa is of special interest in this connection since it is
a species- in which individuals heterozygous for any of the factors so
far identified have a heterozygous form, and this applies, not only to
the factor converting yellow chromogen into anthocyanin, but also
to the factor for the presence of chromogen (Marryat, 533).

In Indian cotton, individuals heterozygous for the factor which
produces anthocyanin in the flower have this pigment diffused through
part only of the petal, whereas homozygotes have anthocyanin through-
out the flower (Leake, 561).

In some strains of Phaseolus, plants which are heterozygous for
mottling with anthocyanin in the testa have mottled seeds, whereas
homozygotes in this factor have self-coloured purple seeds (Shull, 521).
It is at the same time well known that homozygous true-breeding
mottled races do exist, and Emerson has attempted to explain this
anomaly by postulating two factors concerned in mottling.

COLOUR FACTORS IN REDUPLICATION SERIES

It was first observed by Bateson & Pimnett (549, 550) that in the
case of certain factors, which for the moment we may call A and B,
the nature of the F 2 generation varies according to whether A and B
are carried into the Fj heterozygote by the same gamete or by different
gametes. Thus if the heterozygote AaBb is formed from the gametes
AB and ab, then the gametes of the heterozygote are not formed in
the proportion

1AB : lAb : laB : lab,

but may be formed in the following proportions:

SAB : Ab : aB

TAB : Ab : aB

15AB : Ab : aB

3ab
Tab
15ab, etc.

If. on the other hand, the heterozygote is formed from the gametes
Ab and aB, the gametes from the heterozygote so derived are produced
in some one of the following proportions :

186 ANTHOCYANINS AND GENETICS

AB : 3Ab : 3aB : ab
AB : 7Ab : 7aB : ab .
AB : 15Ab : 15aB : ab

'Coupling' was the term first employed for the phenomenon observed
among the offspring for the heterozygote AB.ab, 'repulsion' for the
phenomenon from the heterozygote Ab.aB. The terms are now rejected
in favour of the expression 'reduplication.'

From evidence first obtained, it was thought that the repulsion
was complete, but later evidence is strongly in favour of partial
repulsion such as is expressed above.

The relation between the gametic and zygotic series for a few terms
is explained in the following table:

Gametic series

Number Number
of of

Nature of zygotic series

'-

AB

Ab

aB

ab

g*U4*UV*>a fJJ gWVUO

in series formed

AB

Ab

aB

ab

1 (n-l)

(n-l) I

2n

4n 2

2?1 2 +1

n z -l

n 2 -l

]

I

31

31

1

64

4096

2049

1023

1023

1

I

15

15

1

32

1024

513

255

255

1

I

7

7

1

16

256

129

63

63

1

I

3

3

1

8

64

33

15

15

1

Partial

repulsion

from

zygote of form

AbxaB

3

1

1

3

8

64

41

7

7

9

7

1

.1

7

16

258

177

15

15

49

15

1

1

15

32

1024

737

31

31

225

31

1

1

31

64

4096

3009

63

63

961

63

1

1

63

128 16384

12161

127

127

3969

(n-l)

1

1

(n-l)

2n

4w 2

3n z -(2n-l)

2n-l

2n-l

n z -(2n-

Partial coupling from zygote of form AB x ab

Only those cases connected with anthocyanin pigmentation are
quoted in this account. As it happens, the first example noticed of
these phenomena was in the Sweet Pea (Lathyms) from F x plants
heterozygous for blue and red colour, and for long and round pollen.
The blues were, for the most part, long-pollened, and the reds, for the
most part, round-pollened. The partial coupling was shown to be:

7BL : 1B1 : IbL : 7bl

where B is blue colour, and L is long pollen.

The second example was that of F a plants heterozygous for dark
and light axils, i.e. pigmented (anthocyanin) and unpigmented axils,
and fertile and sterile anthers. The coupling was shown to be

15DF : IDf : IdF : 16df

where D is dark axil, and F is fertile anthers.

The first example of repulsion was noted in the F 2 from a plant
heterozygous for blue and red flowers, and for erect and hooded

ANTHOCYANINS AND GENETICS 187

standards. No red hooded individual was found. Also in the cross
between a dark-axilled plant with sterile anthers (Df) and a light-
axilled plant having normal anthers (dF), i.e. the converse of the case
quoted above, no sterile light-axilled plants were found.

As previously mentioned, the repulsion cases were originally thought
to be total, and, on consulting the above table, it will be seen that,
unless the repulsion is of small intensity, the form aabb is very rare.
There is, however, some evidence that the case of blue colour and long
pollen is of the 1 : 7 order. This case is the reverse of that quoted
above and occurs in the F 2 from the cross Bl x bL. One red round-
pollened plant was found in 419 individuals. The coupling, as we have
seen, is on the 7:1:1:7 system. The coupling between blue and erect
standard, i.e. from BE x be, is of the 127 : 1 : 1 : 127 system, so that
if there should be the same intensity in the coupling and repulsion
systems, the number of plants required to give 1 hooded red would be
65,536, which would lead to the impression first gained, i.e. that there
is total repulsion.

Gregory (558) has also discovered a case of coupling in Primula
sinensis between magenta colour and short style, which are dominant
respectively to crimson colour and long style ; the case is on the 7 : 1
system. There is also evidence of partial coupling with a third factor
which suppresses the development of pigment in the stigma, giving
rise to the dominant green stigma. Yet another instance of complete
repulsion is that between the above-mentioned factor suppressing
colour in the stigma and a factor suppressing colour in the stem.
Thus a heterozygote from L stem r stigma x g stem g stigma gave
no plants with red stem and red stigma.

It is not proposed here to enter further into the complexities of
the question of reduplication though one more point may just be
mentioned. In a recent paper 1 , Trow has suggested a distinction
between what he calls primary and secondary reduplication series.
Thus, he says, if we take a case of three factors, A, B and C, where
there is reduplication between A and C of the form n : 1 : 1 : n, and
between A and B of the form m : 1 : 1 : m, then the secondary form of
reduplication derived from these primary ones, and denoting the relation
between B and C, has theoretically the form

nm + l:m + n:m + n: nm + 1.

Experimentally this hypothesis appears to fit the above-mentioned

case occurring in Primula sinensis (Gregory, 558) between the three

1 Trow, A. H., 'Forms of Reduplication,' J. Genetics, Cambridge, 1913, n, pp. 313-324.

188 ANTHOCYANINS AND GENETICS

factors M (magenta, dominant to crimson, colour). S (short, dominant
to long, style) and G (green, dominant to red, stigma). The MG
reduplication is of the form 2:1, the MS of the form 7:1; the secondary
reduplication, SG, calculated on the above hypothesis, should be
5:3:3:5, and this is fairly near the experimental result. Further
confirmation of Trow's hypothesis is also given by results published
by Punnett (599) for the Sweet Pea crosses, BEL x bel and BeL x bEl,
where B is blue, dominant to red, colour, E erect, dominant to hooded,
standard, and L, long, dominant to round, pollen.

PATTERN IN COLOUR VARIATION

Most of the following cases have been mentioned in the previous
accounts, but they are enumerated again here for the convenience of
reference. Pattern generally implies the localisation of pigment . in
definite areas.

Antirrhinum majus. The 'delila' variety may perhaps be regarded
as an instance of pattern. For every variety with anthocyanin there
is a corresponding ' delila form,' that is one in which the lips are coloured
but the tube is ivory; there is always a sharp line of demarcation at
the point of union of the tube and lips (Wheldale, 535).

Arum maculatum. The black spots (anthocyanin) on the leaf are
probably due to a Mendelian factor which is dominant to the lack
of spots (Colgan, 552).

Cypripedium (Paphiopedilum). Patterns of spots or stripes (antho-
cyanin) may be present. On crossing, there is segregation into striped
and spotted; the former appears to be dominant (Hurst, 502).

Digitalis purpurea. The presence of spots on the inner lower surface
of the corolla is bound up with the presence of pigment in the corolla
generally. The ground colour of the corolla can be inhibited by an
inhibiting factor, but not the spot colour (Keeble, Pellew & Jones, 542).

Erodium cicutarium. A variety occurs without patches at the base
of the petals (de Vries, 565).

Gentiana punctata. There is a variety in which the dark patches
in the flower are absent (de Vries, 565).

Gossypium. The red anthocyanin 'spot' on the leaf is a Mendelian
character, and is less intense in the heterozygous forms. In Indian
cotton a red spot on the flower is characteristic of the heterozygote,
but an almost completely red flower is characteristic of the homozygote
(Balls, 515, 523; Leake, 561).

Mimulus quinquevulnerus . " Here the dark brown spots vary between

ANTHOCYANINS AND GENETICS 189

nearly complete deficiency up to such predominancy as almost to hide
the pale yellow ground-color " (de Vries, 498).

Papaver orientale. The dark patches of anthocyanin at the base
of the petals are absent in some varieties (de Vries, 565).

P. Rhoeas. In some varieties there is a white margin to the petals
due to the presence of a dominant inhibiting factor (Shull, 588).

P. somniferum. Some varieties have dark basal patches of pigment
(anthocyanin) on the petals. From others it is absent. On crossing
dark "'Mephisto' with the white-hearted 'Danebrog' the hybrid
shows the dark pattern" (de Vries, 498).

Phaseolus. In some strains the mottling of anthocyanin in the
testa is due to a factor M, but the pattern only shows when the plant
is heterozygous for M. In other strains the mottling occurs in both
heterozygous and homozygous forms (Emerson, 528, 529; Shull, 513,
521 ; Tschermak, 479, 590).

Pisum. The mottling of anthocyanin in the testa is due to a definite
Mendelian factor (Lock, 518).

Primula sinensis. The patches of anthocyanin outside the yellow
eye are a definite Mendelian character. They do not develop to their
fullest intensity unless the stigma is red, or the flower is deep-coloured ;
the spots are pale in pale-coloured flowers and in flowers with a green
stigma.

The 'Duchess' distribution of pigment, i.e. a ring of colour round
the eye, spreading to a flush, may be regarded as pattern. So also
the 'Sirdar' variety, where the colour is present in minute dots giving
the flower a dusty appearance, the margins of the petals being white
(Gregory, 557).

Solanum tiiberosum. The flowers of several white varieties have
'tongues' of colour radiating out from the throat to the tips of the
corolla segments. The 'tongues' show different degrees of intensity
of coloration, governed, probably, by different factors (Salamau,
unpublished work).

Tropaeolum majus. There is at present no published work on the
flower colour of this genus but, as far as observations go, there is a
definite inheritance of the blotch at the base of the petals. The original
type has an orange-red (anthocyanin and yellow plastids) flower, rather
deeper in colour at the base of the petals, and with dark ' honey-guides '
running into the 'claw.' There are varieties in which the anthocyanin
has disappeared from the main part of the flower so that the flower
is yellow with orange-red blotches at the base of the petals. If the

190 ANTHOCYANINS AND GENETICS

flower-colour varies to pale yellow (see p. 150) the orange-red blotch
then appears as carmine. The red anthocyanin may give rise to a purple
variation and there are correspondingly coloured blotches, maroon on
a deep yellow and purple on a pale yellow ground.

Reviewing the cases of pattern set out above we are able to distinguish
several different types. One type, for instance, includes all those spots,
lines and streaks which form a normal part of flower-coloration and
which, in the opinion of biologists, have a significance as signals or guides
to insect visitors. This group would include the markings in Cypri-
pedium, Digitalis, Erodium, Mimulus, Papaver, Primula and Tropaeolum.
As far as can be judged from the evidence at hand, some of these patterns
(Papaver, Tropaeolum) are inherited independently of the ground colour
of the flower ; the factors for others are intimately associated in various
ways with the ground-colour factors (Digitalis, Primula). A second
type of pattern is that which is not a feature of the normal flower-
coloration, but is only revealed in varieties from which some of the
colour factors are absent, for example the ' delila' variety of Antirrhinum
and the 'Duchess' and 'Sirdar' varieties of Primula. A third type of
pattern, for instance the spots on Arum leaves, the leaf-spot of Gossypium
and the mottling of seeds of certain strains of Pisum and Phaseolus,
forms a normal attribute of the plant to which we. can assign no special
significance.

As regards the nature of the factor which produces mosaics or
mottling in seed-coats we cannot at present offer any interpretation.
It apparently limits the distribution of colour to certain areas ; whether
this is due to inhibition of pigment, or whether it is due to the fact
that the pigment formed in the plant is less, and hence insufficient to
give self-colour, cannot be decided. But it is evident that the mottling
factor is one apart from colour and can be carried by albinos. Bateson
(524) draws an interesting comparison between colour-pattern and
dilution. In both cases, he points out, less pigment is present in the
zygote, but in one case the area is diminished though the intensity
remains the same, in the other the intensity is diminished, but the
area of coloration is constant.

STRIPED VARIETIES AND BUD- VARIATION.

A common form of variation in many flowers is striping, i.e. the
arrangement of colour in bands or stripes; these markings may vary
in thickness from the narrowest hair-like streaks to broad bands, or
elongated patches, which may then occupy almost the whole of the

ANTHOCYANINS AND GENETICS 191

flower. It is difficult to define the limits of striping for, on the one hand,
among striped varieties we frequently find sectorial variations in which
colour is definitely and symmetrically confined to a half, a third, or
some other fraction of the flower. On the other hand, striping may
pass into spotting or blotching, and it i questionable whether spotted
and blotched flowers- should be placed in the same category, though
their genetical behaviour may be similar. Some of the genera and
species in which striping occurs are the following:

Antirrhinum ma jus, Papaver,

Cheiranthus Cheiri, Petunia,

Dahlia variabilis, Primula sinensis,

Dianthus Caryophyllus, Tagetes,

D. barbatus, Tulipa,

Mirabilis Jalapa, Zinnia elegans.

As further examples of striping and sectorial variation, we may
add those mentioned by de Vries (565), i.e. Celosia variegata cristala
(inflorescence), Centaurea Cyanus, Clarkia pulchella, Convolvulus tricolor,
Cyclamen persicum, Delphinium Ajacis, D. Consolida, Geranium
pratense, Helichrysum bracteatum, Hesperis matronalis, Impatiens
Balsamina, Nemopliila insignis, Papaver nudicaule, Portulaca grandi-
flora and Verbena.

Striping may be regarded as a variation under normal conditions 1 ;
there is, however, a kindred phenomenon called flaking which appears
in flowers, otherwise self-coloured, towards the end of the vegetative
season, or when the plant is in an unhealthy condition. Ex. Dahlia
(Hildebrand, 458, 466), Matthiola.

Striping is apparently not a phenomenon which occurs in plants
in the wild state, and, according to the theory of Louis Vilmorin, as
expounded by de Vries (565), it only appears in coloured plants which
have already produced a white or yellow variety, that is a variety free
from anthocyanin pigment. It is certainly true that striping is most
usual in connection with anthocyanin; as a rule it is not shown by
soluble yellow pigments (Antirrhinum, Althaea), though in Papaver
nudicaule, according to de Vries (565), one finds a yellow variety with
dark orange stripes. In Mirabilis, also, yellow pigment may be found
striped upon a white ground. Flowers with yellow plastids rarely,
if ever, occur in the striped condition ; such yellow flowers may, however,
occasionally show light segments containing paler derivative plastid

1 De Vries (565) notes, however, that striping in Camellia japonica may depend on
the time of flowering.

192 ANTHOCYANINS AND GENETICS

pigments. Tlie common form of striping is that of anthocyanin on an
albino ground, either white (Primula sinensis] or yellow ; the yellow
may be a soluble pigment (Antirrhinum majus var. yellow striped
crimson), or a plastid pigment (Cheiranthus Cheiri, Tagetes, Tulipa spp.).
Less commonly one finds deep anthocyanin stripes on a pale antho-
cyanin ground of the same colour (Antirrhinum,, see p. 160), or stripes
of one coloured anthocyanin on the ground of another colour, as in the
case mentioned by de Vries (565) of Delphinium Consolida striatum
plenum in which the flowers may be red striped with blue 1 . Mirabilis
Jalapa presents a remarkable series of striped varieties : there may
be stripes of magenta (anthocyanin) on a white ground, or of soluble
yellow pigment on a white ground, or both magenta and yellow stripes
on white (tricolour). Moreover Mirabilis is characterised by a number
of heterozygous forms, and we may find a homozygous stripe-colour
on a heterozygous ground-colour, for instance, deep yellow stripes on
a pale yellow, or orange-red stripes on a magenta-rose ground (Marryat,
533).

Another curious phenomenon is the limitation of striping in most
species to one colour, for instance in Antirrhinum the anthocyanin of
rose dore and bronze varieties does not occur in stripes, and the
same is true of the pale 'tingeing' of anthocyanin. It is only the
factor for full colour that is affected by the striped condition.

With regard to sectorially coloured flowers de Vries (565) notes
that they appear to manifest a tendency towards a simple proportion
between the two parts. "Frequently exactly half of the flower is
atavistic, sometimes a quarter or three quarters. I observed the
proportion f in white and red striped tulips and in partially dark blue
and partially pale blue flowers of Iris xiphioides, etc."

The inheritance of striping has received a certain amount of attention.
De Vries (565) has investigated the problem in Antirrhinum and other
striped flowers. He noted that the striped variety of the larkspur
(Delphinium Ajacis and D. Consolida) produces self-coloured flowers as
well as striped ones; the self-coloured, moreover, may appear on the
same racemes as the striped ones, or on different branches, or some
seedlings from the same parent may be self-coloured while others are
striped. Seeds, on the other hand, of these blue sports give rise again
to the striped variety. Such a variety as this Larkspur, de Vries called
' eversporting,' for as he says : " Here the variability is a thing of absolute
constancy, while the constancy consists in eternal changes." His more

1 There is also a tricolour of red, white and blue.

ANTHOCYANINS AND GENETICS 193

definite cultures were made with Antirrhinum majus luteum rubro-
striatum. In Antirrhinum (see p. 160) the flower may be ivory striped
with magenta anthocyanin, or yellow striped with crimson, this latter
colour being merely magenta superposed upon yellow ; the yellow
pigment (luteolin) does not show striping. In the offspring of de Vries'
plant, which was yellow striped with crimson, there were no pure
yellows, though all grades of striping were found, ranging from yellow,
with a few of the finest stripes, through intermediates, to those with
the broadest stripes, or even whole sections of red ; and with these
occurred a certain number of pure red flowers. The red flowers will
appear suddenly as a sport (though without graduated links with striped
forms) among the striped individuals, and conversely, the striped
flowers will appear among the offspring of the self-form. Thus red
can be obtained from striped, and striped from red, and both variations
can be obtained by seed ; but de Vries noted bud-variations and varia-
tions within the spike, only in striped individuals, though he is of the
opinion that they might be found on red plants if sufficiently large
numbers of individuals were dealt with.

The following table 1 represents the successive families obtained
by de Vries from one striped plant:

Striped plant ... ... ... ... P!

F a

Striped

Striped
plants
98%

i
Striped plants

90%

i

1
Red plants

10%

...

n

Striped plants Red plants

QS / 9 /

y /o /o

i i
Striped plants Red plants

24% 76%

1 1
branches Bed branches
i i

n r
Red Striped plants

plants 29 %

9 o/ |
& /o

i
Red plants ...

71%

...

1 1
Striped Red

plants plants

OK O/ K O/

yo /o /o

Striped Red
plants plants
16% 84%

So far then de Vries expresses the facts of inheritance of striping
thus:

1. The striped race is an inconstant one and consists of striped and
red-flowered plants.

2. Striping itself varies continuously, but there is discontinuity
between the striped and red forms.

3. The intensity of inheritance of finely striped plants is about

1 See also Emerson (605)

w. P. 13

194 ANTHOCYANINS AND GENETICS

95-98 %, and they give rise to the red type either by seed or bud
variation.

4. The broadly-striped individuals produce more reds than the
narrow-striped, the average being about 39 %.

5. The red individuals thrown by the striped resemble the red
type, but differ in again throwing striped. The intensity of inheritance
of the red character is about 70-85 %.

6. The yellow variety (without anthocyanin) does not arise from
the striped race.

De Vries draws attention to the fact that striped varieties do not
occur in nature. Other examples he gives of striping are in Stocks,
Liver-leaf (Hepatica), Dame's Violet (Hesperis}, Sweet William (Dianthus
barbcttus) and Periwinkle ( Vinca minor) ; also in Hyacinthus, Cyclamen,
Azalea, Camellia, and in the Meadow Crane's-bill (Geranium pratense)
when cultivated. Also these, he says, are known to come true to
striping when seed is taken from striped individuals, and from time
to time to throw self-coloured individuals. He made pedigree cultures
of the Dame's Violet (Hesperis matronalis) for five years, and of Clarkia
pulchella for four years, and they both behaved in exactly the same way
as Antirrhinum. He further points out that other parts of plants may
be striped, as, for instance, the red and white striped roots of Radishes,
and the inflorescence of the Cockscomb (Celosia cristata).

From these results one would gather that. the inheritance of striping
is non-Mendelian. More recently Emerson (605) has suggested that
striping, on the basis of a certain hypothesis, may conform to the
Mendelian scheme of inheritance. It appears that ears of certain
varieties of Maize show striping in the red pigment developed in the
pericarp. Emerson says : " Plants in which this pigment has a variegated
pattern may show any amount of red pericarp, including wholly self-
red ears, large or small patches of self-red grains, scattered self-red
grains, grains with a single stripe of red covering from perhaps nine-
tenths to one-tenth of the surface, grains with several prominent stripes
and those with a single minute streak, ears with most of the grains
prominently striped and ears that are non-colored except for a single
partly colored grain, and probably also plants with wholly self-red
and others with wholly colorless ears." A number of selfmgs were
made for several generations of both homozygous and heterozygous
variegated cobs ; homozygous and heterozygous variegated cobs were
also crossed with true-breeding white male plants. From his experi-
ments Emerson deduces the following results:

ANTHOCYANINS AND GENETICS 195

1. That the amount of pigment in the pericarp of a variegated
grain bears a definite relationship to the amount of pigment in the
grains of the plant which grows from it. The relationship is such
that the more red pigment there is in the grains planted, the more
likely are the plants which come from them to produce self-coloured
red ears, and the less likely to produce variegated ears.

2. That when F x red ears produced by selfing a homozygous
variegated-eared plant are selfed and sown they give rise to red-eared
and variegated- eared plants in Mendelian proportions; in the same
way when crossed with white-eared races, they behave as if they had been
produced by a cross between red-eared and variegated-eared races.

3. That the Fj red ears arising from selfed heterozygous variegated-
eared plants behave in some cases as if they were hybrids between
red-eared and variegated- eared races, and in other cases as if they were
hybrids between red-eared and white-eared races.

4. That the F a reds arising from crosses between both homozygous
and heterozygous variegated-eared plants and white-eared races behave
as if they were hybrids between red-eared and white-eared races.

Thus, says Emerson, any interpretation of the above results must
take into account these facts: (1) that the more reel there is in the
pericarp, the more frequently do red ears occur in the progeny ; (2) that
such red ears behave just as if they were F x hybrids between red and
variegated, or red and white races.

To explain these phenomena Emerson suggests the following hypo-
" thesis. The zygotic formula for a plant homozygous for variegated
pericarp may be regarded as VV ; heterozygous for variegated pericarp
as V . If in any somatic cell VV, from some unknown cause, a
V factor were transformed into a factor for self-colour, S, that cell
would then be represented as VS. Any pericarp cells descended from
such a cell would be red, and if all the pericarp cells of a grain were
thus descended the grain would be self-red, just as if the plant bearing
it were a hybrid between pure red and variegated races. Of the gametes,
moreover, arising from such somatic cells, one-half would carry V, and
one-half S, again just as if the plant were a hybrid between red and
variegated races. If both the V factors were changed to S, the grain
would be red as before, but all, instead of half, of the gametes would
carry S. If the modification from VV to VS should occur very early
in the life of the plant, or even in the embryo, then all the ears of the
plant might be self-red, and one-half of all the gametes, both male and
female, might carry S, and the other half V as in an ordinary hybrid.

132

196 ANTHOCYANINS AND GENETICS

If the modification should occur much later, for instance when the ear
is beginning to form, there might be then only a patch of red grains
on a variegated ear, and only those gametes arising from these red
masses of tissues would carry half S and half V. Finally if the modifi-
cation should occur after the grains begin to form, the latter would
have broad or narrow stripes according to the amount of pericarp
directly descended from the modified cell, and the larger the amount
of modified tissue the greater the chance that the gametes concerned
would carry S. Similarly in any cell of a heterozygous, variegated-
eared plant, V -- , he assumes that the V factor may be changed to S.
The effect on the pericarp colour would be the same as before, and of
the gametes arising from the modified tissue one-half, and never more,
would carry S, the other half would carry no factor, and would be
represented by .

Emerson then applies his hypothesis to results obtained in the F 2
and F 3 generations. As stated above Fj red-eared plants which had
arisen from selfed homozygous variegated ears gave in F 2 only red-eared
and variegated-eared offspring. On the hypothesis quoted, the con-
stitution of the Fj red-eared plants would be either VS or SS, the former
being more frequent than the latter on account of the rarity of S in the
cJ gametes. The Fj red ears tested were evidently of the composition
VS. Of two F 2 reds from selfed F^s, one gave reds and variegated,
the other bred true to red. Hence the hypothesis is in accordance with
the results. Again F x red-eared plants which had arisen from hetero-
zygous variegated selfed plants, as we have seen, behaved in some cases
like hybrids of red and variegated races, in other cases like hybrids
of red and white races. On the hypothesis that variegated-eared plants
were V - and their red grains S , the Fj plants would be SV, SS,
S - , V , or - - . Of the F x reds tested some were evidently SV,
and others S . Of the F 2 reds, one bred true in F 3 , and others segre-
gated into reds and variegated.

Finally when Fj red-eared plants arose from either homo- or hetero-
zygous variegated ears that had been crossed with whites they gave
only red-eared and white-eared offspring, never variegated. By hypo-
thesis the parent variegated-eared plants were V - - and VV, and the
red grains S - and S V (or SS possibly) and the male parents were

- . The F x plants therefore would be S , V - and - -, and only
S - would be red-eared. The red-eared F x plants tested gave red-
and white-eared in Mendelian ratios. Of the F 2 red-eared, one bred true
in F 3 , and the others segregated into reds and whites.

ANTHOCYANINS AND GENETICS 197

Emerson further maintains that de Vries' results for striping in
Antirrhinum indicate that this case is of a similar nature to Zea. Also

the results obtained by Correns (537) in striped-flowered plants of
Mirabilis; these results show that plants with self-coloured flowers
behave as if they had occurred in an F 2 from a cross of striped by self-
coloured plants. But flowers from self-coloured branches on striped
plants produce few, if any, more self-coloured plants than flowers on
branches with striped flowers. As an explanation of this anomaly
Emerson suggests that in the case of seed sports the factors for variega-
tion are affected, whereas in somatic variations there is no corresponding
change in the Mendelian factors.

A curious phenomenon in connection with striping in Antirrhinum
is one noted by the author and previously mentioned (see p. 159).
If the factor for tingeing with anthocyanin be denoted by L, and the
factor for full colour by D, then LLDD(d) is magenta and LlDD(d)
pale magenta. If in the striped variety we represent D by S, then
LLSS is ivory striped with magenta, L1SS, ivory striped with pale
magenta, LLSs is tinged ivory striped with magenta and LISs tinged
ivory striped with pale magenta. No explanation of this interesting
result can be offered at present.

In Primula sinensis (Gregory, 557) striping appears to belong to a
different category from that in Antirrhinum, Mirabilis, etc., for it
behaves as a simple recessive to a self-coloured form.

Bearing in mind the facts just recorded one cannot fail to realise
that the common occurrence of reversion to self-colour among striped
varieties is but one expression of the much more widely distributed
phenomenon of bud-variation. Of the latter there are a number of
cases known, many of them involving pigmentation. Some are con-
cerned with the distribution of chlorophyll, and these it is not necessary
to consider here. Of those connected with anthocyanin there are
several recorded by de Vries (498, 565), for instance in Ribes sanguineum,
Veronica longifolia and others. Ribes sanguineum has red racemes of
flowers and a certain amount of anthocyanin in its twigs and petioles,
and there is a variety which has white flowers tinged slightly with red,
while the vegetative parts lack pigmentation. Of the variation de
Vries says: "Occasionally this white-flowered currant reverts back
to the original red type and the reversion takes place in the bud. One
or two buds on a shrub bearing perhaps a thousand bunches of white
flowers produces twigs and leaves in which the red pigment is noticeable
and the flowers of which become brightly colored. If such a twig is

198 ANTHOCYANINS AND GENETICS

left on the shrub, it may grow further, ramify and evolve into a larger
group of branches. All of them keep true to the old type." Another
case is that of the hybrid from Veronica longifolia and its variety alba ;
the blue hybrid occasionally produces white flowers. We may also
include the production of green-leaved branches by the purple-leaved
Beech, Hazel, etc. Other cases are mentioned by Bateson (591) : for
example, Azalea, of which there are white varieties streaked with red
giving rise to self-red sports. Also Pelargonium', here the variety
altum which is normally red may produce very occasionally one or two
magenta petals, and, conversely, there is a variety 'Don Juan' which
may bear trusses or branches of red flowers, though the normal colour
is magenta. Another case of exceptional interest is that in which an
individual of the purple- winged 'Purple Invincible' variety of Lathyrus
developed a flower of the variety 'Miss Hunt' which is lacking in the
blue factor (Bateson, 524).

With regard to bud-variation there are several fundamental points
to be borne in mind, and these have been well expressed by Bateson
(524, 591) from whom the following quotations are taken. First:
" when a bud-sport occurs on a plant, the difference between the sport
and the plant which produced it may be exactly that which in the case
of a seminal variety is proved to depend on allelomorphism." This
is exemplified in the cases given above, i.e. in the presence or absence
of factors controlling pigment formation (Antirrhinum, Ribes, Veronica,
Fagus, etc.). Secondly, on consideration of these cases, we may arrive
at the conclusion that the segregation of the allelomorphs which control
the production of colour must have taken place at some somatic division,
and "we are thus obliged to admit that it is not solely the reduction-
divisions which have the power of effecting segregation." The third
point is that: "The distribution of colour in this case (bud- variation)
lies outside the scheme of symmetry of the plant." For "though the
parts included in the sports show all the geometrical peculiarities proper
to the sport-variety, yet the sporting-buds themselves are not related
to each other according to any geometrical plan," just as striping itself
in the Carnation or Antirrhinum is not under geometrical control. And
it is precisely the plants with this disorderly arrangement of striping
which so often give rise to bud-sports 1

1 Bateson (591) notes a most interesting case of bud-variation in the Azalea Vervaeana.
The flowers have symmetrical markings of one shade of red, and red streaks of another
shade. When self-red sports arise they are of the shade of the red streaks, not of the
symmetrical markings.

ANTHOCYANINS AND C4ENETICS 199

At the present moment it is difficult to form any general conception
of the mechanism of the change which underlies bud-variation. In a
recent paper Emerson (629) describes the occurrence of anomalous
seeds of Maize (Zea), two of which are concerned with pigmentation.
In one, the seed was half colourless and half purple : in the other, half
purple and half red. He admits that the occurrence of this phenomenon
could be explained on the hypothesis that, subsequent to normal endo-
sperm fertilisation, there occurs a vegetative segregation of genetic
factors. But such a segregation, he maintains, cannot be typically
Mendelian because in neither of the cases quoted are all the genetic
factors (in a heterozygous condition) involved. He prefers to interpret
the phenomenon as a somatic mutation, that is as a change in genetic
constitution rather than a segregation of genetic factors. He proceeds
to apply this conception to the case of bud-sports in general. His
hypothesis in fact has already been mentioned in connection with striping
in Zea and Antirrhinum. In his opinion the somatic mutation may be
a gain of at least one new factor, the loss of a factor, or the permanent
modification of a factor. Against the segregation hypothesis he brings
the following considerations. In material homozygous with respect to
the Mendelian factors concerned it is not possible for bud-sports to arise
by segregation. Nor is it possible if a new character previously unknown
to the species, should arise ; nor if a dominant character appears as a
bud-sport in material known to be homozygous in a recessive character
which is allelomorphic to the dominant character in question (ex.
striping in Zea). If however the bud-sport be due to the loss of a
character, and the material be also heterozygous for the character,
then the case can be interpreted equally well as a segregation or a
mutation.

Bateson, on the other hand, inclines to the view that bud-sports
are the result of somatic segregation, the sporting branches being the
outgrowth from cells containing one only of the segregating factors.
He also suggests that the phenomenon of striping may be due to the
fact that there is an insufficient amount of the colour factor in the
gametes to make the zygote self-coloured. And this may also apply
to some cases of pattern (see p. 190). In these cases the more intimate
the mixing, the more likely are the offspring to be striped, and this,
as we have seen, is borne out by observations upon Zea and Antirrhinum.

200 ANTHOCYANINS AND GENETICS

THE EFFECT OF OUTSIDE FACTORS ON COLOUR- VARIATION.

It has very often been stated that the colours of flowers are affected
by the soil in which the plants are growing, and the belief is still prevalent.
In any discussion upon the subject we generally meet with one or other
of two illustrations of classical interest. One is the change of colour
in Hydrangea (or Hortensia) flowers when the roots are supplied with
some form of iron or aluminium salts; the other is the 'zinc Violet'
which is a blue-flowered form of Viola lulea. This variety grows in
a soil containing as much as 20 % zinc oxide, and in the ash of the
plant may be found 1 % zinc oxide. Hence the idea arose that the
'zinc Violet' might derive its blue colour from the presence of zinc in
the tissues. The effect of soils on Hydrangea must have been noticed
at a very early date, for Schiibler & Lachenmeyer (429) in 1834 gave an
account of analyses of soils which change the colour of the flowers.
Trials have been made from time to time with other chemical substances
upon other plants, but no striking results have been obtained. In
view, however, of the value attached to such results, it may be well
to give an account of some of the experiments.

As regards the 'zinc Violet,' Hoffmann was led to conclude that
the presence of zinc had no influence on the colour, because, w T hen
transferred to another soil which did not contain zinc, the blue colour
remained ; the colour, moreover, may vary even when the soil contains
zinc. Though practically only confirming the fact that colour-changes
in Hydrangea are brought about by iron and aluminium salts, the
work of Molisch (467) in this direction is important, since he treated
the subject on an experimental basis. His method was to grow
Hydrangea plants in pots and to mix with the soil the substances of
which he wished to test the effect. The following table will give a general
idea of his results:

Substance added to soil Colour of flowers Remarks

Normal soil ... ... Red

Aluminium sulphate ... Strongly blue ... Leaves became brown in

some cases and died
Common alum ... ... Blue, sometimes bluish

A1 2 (S'O 4 ) 3 + K 2 SO 4 + 24H 2 O sometimes pink with

blue filaments

Ferrous sulphate Pink, bluish, or blue... In most cases, after a

few days, the leaves
turned brown and com-
menced to fall off

ANTHOCYANINS AND GENETICS

201

Substance added to soil

Iron oxide (Hammerschlag)

Fe 3 4
Moor soil (from Wittingau

in Bohemia)
Heath soil (from Cibulka

near Prague)

Peat

Aluminium oxide, A1 2 O 3 ...

Iron filings

Iron shavings
Iron tacks ...

Ochre

Granulated zinc ...
Tin foil (clippings)
Charcoal
Emery powder
Powdered slate

Powdered sulphur
Soda

Lime

Ferric chloride

Potassium sulphate
Ammonium sulphate
Manganese sulphate
Copper sulphate ...

Colour of flowers

Two plants bluish, the

rest red
Blue

Red, with blue filaments
All red

All red

Nickel sulphate ...

Cobalt sulphate ...
Zinc sulphate
Potassium carbonate

Remarks

Plants strong and flowers
large

Some tendency to become
blue in filaments and
fruiting flowers

Leaves died in most cases
after a week: plants
weakly

Used in too large quan-
tity

Leaves tended to fall off

Leaves became brown
Only one plant: red
flowers

Substance very poison-
ous: plants died

Molisch found the filaments of the stamens very sensitive to the
substance added, and notes that if one is in doubt as to whether any
effect has taken place, one should examine these organs. In the case of
alum, it is evidently the aluminium salt which is effective, since potas-
sium sulphate produces no blue colour by itself. Aluminium sulphate
alone, moreover, produced the most intense blue colour. Aluminium
oxide has no effect on account of its insolubility, whereas in the slate
there must be some slightly soluble iron or aluminium compound. From
his researches Molisch draws the conclusion " dass Alaun, schwefelsaure

202 ANTHOCYANINS AND GENETICS

Thonerde und Eiseuvitriol die rothe Farbe der Hortensienbliithe in
die blaue umzuwandeln vermoge." He points out that the pigment
of the flowers is anthocyanin, and in sections of the petals the colour is
similarly turned blue on addition of solutions of iron and aluminium
salts. He further emphasises the fact that all plant anthocyanins do
not behave in the same way, and hence this may be one of the reasons
why it has not been possible to change the flower-colour of other plants
by adding salts to the soil.

In 1906, Kraemer (503) gives an account of a number of experi-
ments which he made to test whether flower-colours can be modified by
treating the soil with chemical substances. The compounds he used
were : acetic acid, citric acid, malic acid, phosphoric and other acids ;
various iron salts, such as acetate, citrate, chloride and sulphate;
certain aluminium salts, such as sulphate, phosphate, and the double
salts of aluminium and potassium sulphate, and aluminium and am-
monium sulphate; ammonia water, potassium hydrate, ammonium
nitrate, potassium nitrate, potassium iodide, iodine and potassium
cyanide. These were supplied in solution through the soil, beginning
with 1 : 10,000 of water, and the strength gradually increased until
1 : 1000 of water was reached. Most of the substances could be supplied
in this strength every five days for some months without injury to the
plants. There were no marked effects. In yellow roses supplied with
aluminium and potassium sulphate, the leaves and stems became
slightly reddish. ' La France ' roses treated with iron citrate and citric
acid had uniformly pink petals. Some rather insignificant changes
were noted in other plants. The scarlet carnation showed a tendency
to form white streaks with iron and ammonium sulphate, aluminium
phosphate, iron citrate and citric acid ; also a maroon carnation treated
with ferrous sulphate. The petals of a white carnation when fed with
potassium and aluminium sulphate showed a tendency to form red
streaks.

Later, in 1908, Vouk (522) repeated a number of experiments on
the same lines as Molisch, using plants of Hortensia (Hydrangea hortensis).
Vouk remarks, as a preliminary, that Miyoshi (? 187) has succeeded in
changing the lilac colour of Callistephus chinensis and Campanula
alliariaefolia into blue, and the red of Lycoris radiata into .lilac by
artificial treatment. Of Molisch's experiments he observes that the
leaves of the plants used were frequently injured, and this may be due
to too large quantities of salts added. He therefore set out to deter-
mine (1) how the change to blue colour is affected by different quantities

ANTHOCYANINS AND GENETICS 203

of aluminium salts (potassium alum and aluminium sulphate) : (2) what
amount of salt is sufficient to bring out the blue colour without injury
to the plant. His methods were to water plants with -5 %, 1 %, and
3 % solutions of potassium alum and aluminium sulphate respectively.
His results show that cultures watered with 3 % potassium alum
produced the finest blue colour, though the leaves were affected with
brown spots. The 1 % cultures he regarded as best, since the plants
were quite sound, and there was at the same time almost complete
blueing of the flowers. This was true for the same concentrations of
aluminium sulphate, but the blue coloration was weaker. Hence Vouk
does not agree with Molisch that aluminium sulphate has a better
effect on the blueing than alum, but he maintains that the latter salt
gives the most intense results. His experiments, moreover, show that
the change of flower-colour depends, not only on the quality, but also
the quantity of the salt used. Other experiments of a similar nature
were tried with Phlox decussata but with no result.

Thus we may conclude that, with the exception of Hydrangea, there
is no experimental evidence for the belief that chemical substances
in the soil can materially affect flower-colour. There are accounts of
various experiments where dyes have been employed, by means of
root absorption, for colouring flowers, but the results are purely tem-
porary and negligible as far as scientific interest is concerned. We may
in this respect well agree with Nehemiah Grew (1) when he remarks:
"From what hath been said, we may likewise be confirmed in the
use of the already known Rules, and directed unto others yet unknown,
in order to the variation of the Colours of Flowers in their Growth.
The effecting of this, by putting the Colour desired in the Flower, into
the Body or Root of the Plant, is vainly talked of by some : being such
a piece of cunning, as for the obtaining a painted face, to eat good
store of white and Red Lead."

Another set of experiments concerned with changes in flower-colour
though involving a different external factor, viz. insolation, are those
of Rawson (519). These experiments have been conducted upon
Tropaeolum plants growing in South Africa and the principle followed
was "to shade off with a perfectly opaque screen all direct rays of the
sun for certain intervals of daylight." Otherwise there was no other
special treatment. By this means Rawson claims to have changed
ordinary scarlet and orange varieties into a new mauve variety, and
"no known instance occurred of its reverting to the original. Experi-
ment so far goes to show that seeds from it after the second year may be

204 ANTHOCYANINS AND GENETICS

planted in any aspect, and will come true even if sown in such different
climates as those of York and Pretoria. The crimson variety similarly
treated in York and in Pretoria gave the same flowers of a bronze old-
gold colour in both places, and the seed of this latter variety brought
from Pretoria and sown in York gave the same curious colour, in spite
of the great difference of altitude between the two localities. Cuttings
of the mauve variety could be grown in any aspect at Pretoria, without
any change in the mauve colour of the flowers. In addition to the
mauve and bronze old-gold colours, varieties of rose-salmon and of
sallow flesh-colour have been obtained, and no difficulty has been
experienced in changing any of the known orange, yellow, or scarlet
flowers into these curious colours." In a later paper, Rawson (600)
states that still further varieties have been obtained in England by
adopting the same methods. Pending confirmation it is difficult to
make anything of the significance of these facts.

CONNECTION BETWEEN COLOUR AND OTHER PLANT CHARACTERS.

There is little doubt that albinism, the most common form of
variation, brings with it in many cases a general weakness of constitution.
The true albino (see p. 158) of Antirrhinum majus, for instance, is more
stunted in growth than the coloured forms; it is also less resistant to
cold, drought and other adverse conditions. In connection with this
point there is a paragraph which may be quoted from the Flora Anomoia
(1817) by Hopkirk (428). This author evidently believed that lack of
robustness may cause variation for he says: "That these varieties of
colour are produced very frequently by weakness, is evident from the
circumstance of the variety being often much more tender than the
original species. The white-flowered variety of the Virginian Spider-
wort (Tradescantia virginica) is with difficulty preserved in the open
air during winter, whilst the blue is perfectly hardy ; in like manner,
the white Snapdragon 1 is much more tender than the common
Antirrhinum majus." We now know that the white variety of
Antirrhinum is lacking in the flavone, apigenin, as well as in anthocyanin ;
both these substances are of the aromatic group, and no doubt play
a certain part in general metabolism. It is also possible that other,
more fundamental compounds, such as oxybenzoic acids, the progenitors
of the flavones, are in addition absent from the white variety, and so
its metabolism, as regards aromatic substances, may in a sense be

1 This may be the ivory and not the true white. It is of course impossible to know.

ANTHOCYANINS AND GENETICS 205

pathological, and may quite well be responsible for the general weakness
of the variety.

As to the connection between colour and other characters such as
flavour and odour, information is rather scanty. Some interesting
suggestions in this direction have been made by Goff (444). He main-
tains the truth of the statement that white varieties of fruits and
vegetables have a milder and more delicate flavour than coloured
forms, and he quotes the following instances. White varieties of
onions, he says, are less strong in flavour than red ones, the blood-red
variety being the strongest flavoured; white currants are less acid
than red or black; white and yellow tomatoes are sweeter than the
scarlet ; white raspberries have a more delicate flavour than the coloured
type. The same idea is involved in the blanching of celery, endive and
seakale, and in the use of the inner leaves only of lettuce and cabbage
for eating. Similarly, 'sun-burned' potatoes, that is tubers which have
been exposed to light and sun, have a strong and bitter flavour ; shoots
of seakale also, if allowed to come above the earth, develop purple
pigment and become strong flavoured. Red cabbage, when cooked,
is less mild and tender than the green varieties. In the sugar-pea,
too, the purple-flowered form has seeds spotted with brown which are
strongly flavoured when cooked. He also points out that the percentage
of sugar in the Beet increases as the percentage of colouring matter
decreases. Further data would be necessary before we could definitely
prove the truth of these suggestions, but what evidence we have is
certainly in their favour. For odour, flavour, astringency, bitterness,
etc., are essentially connected with aromatic compounds, as for instance
bitterness and astringency with the tannins, and we know that absence
of colour is in some cases, as in Antirrhinum, accompanied by the
absence of certain aromatic complexes. The subject is well worth
investigation, and now that cases of Mendelian segregation offer such
well-defined material, it should not be beyond the scope of the plant
chemist. Also, the possibility of a connection between colour and
sweetness is materially strengthened by the evidence, given in previous
chapters, of a relationship between pigmentation and sugar concen-
tration. Goff states, for instance, that the red-fleshed Peach is of little
value for eating purposes, and cases of this kind should offer suitable
material for research.

There are certain relationships of a more precise nature between
colour and other plant characters which have come to light in Mendelian
investigations. One group of such relationships includes the phenomena

206 ANTHOCYANINS AND GENETICS

(in which a certain number of colour factors are known to be involved)
termed reduplication (see p. 185). Another kind of relationship is
that existing in Matthiola (Saunders, 587) between the factors for hoari-
ness of the leaves and flower-colour In Stocks, as we have alreadv

i/

seen, colour (anthocyanin) is due to the presence in the zygote of two
factors (C and R), and if either of these be absent, the plant is an albino
as regards anthocyanin. In certain strains of Stocks, the hoariness
of the leaves has been found to depend also on the presence of two
factors (H and K). Between these two pairs of factors there is a
certain relationship, viz. that the hoariness due to H and K is only
manifested when C and R are both present. Hence an albino may
contain both H and K, and may yet be glabrous because it cannot
contain at the same time both C and R. An anthocyanin form, on the
other hand, which is glabrous carries of course C and R, but can only
contain either H or K, and not both ; when it carries C and R, as well
as H and K, it is hoary and coloured. Thus a heterozygote in all
four factors would give on selfing :

81 CRHK Red hoary.

27 CRH Red glabrous

27 CRK Red glabrous.

27 RHK White glabrous.

27 CHK White glabrous.

9 CR ... ... Red glabrous.

9 CH White glabrous.

9 CK

9 RH

9 RK ... ... ,,

9 KH

o p

L) V ... ... , ,

3T>
JL V * ^ }

3TT
-1-L *

3T7-
I *- y )

1

... ... ,,

Finally a relationship between shape and colour occurs in the
Sweet Pea (Bateson, 524). All those varieties with an erect standard
are 'bicolor,' that is, within the variety, the standard differs more or
less in colour from the wings; ex. 'Purple Invincible' and 'Painted
Lady.' Varieties, on the other hand, with a hooded standard have
wings and standard more of a uniform tint; ex. 'Duke of Westminster'
and 'Duke of Sutherland.'

ANTHOCYANINS AND GENETICS 207

THE CHEMICAL INTERPRETATION OF FACTORS FOR FLOWER-COLOUR.

If we write out the factorial composition of the type of any species
which has been investigated on Mendelian lines, as for instance.
Antirrhinum (see p. 161):

YYIILLTTDDBB
or Lathyrus (see p. 167) :

CCRRBBDiDiDeDe

we see that a number of definite factors go to build up the colour both
the amount and kind of colour of the type, and each of these factors
must represent a power in the plant to control a chemical reaction;
the sum-total of these reactions is the production of the pigment of
the plant. To know with what kind of process each of the factors
corresponds is of very great importance for the understanding of
genetical problems. It may be stated without hesitation that we can
only find out w r hat these processes are by means of exact chemical
analvses.

ti

Such chemical analyses have been attempted in the case of Antir-
rhinum, and we may first consider the kind of deductions which can
be drawn from the results. The factors which have been most investi-
gated are the yellow (Y) and ivory (I). The tissues (except those of
the palate) of the flowers of the ivory variety contain a very pale yellow
soluble pigment, the flavone, apigenin, which is present in all the cells,
both epidermal and mesophyll. The pigment is not recognisable in
the petal, for, as we have pointed out in earlier chapters, the flavones,
though practically ubiquitous, are too slightly coloured to affect the
colour of the cell-sap. The presence of apigenin in the ivory Antirrhinum
can be demonstrated by placing the corolla in ammonia vapour, when
the flower turns bright yellow owing to the formation of a more intensely
coloured salt of the pigment. If, on the other hand, we examine sections
of the corolla of the yellow variety, we find the epidermal cells filled
with a soluble yellow pigment, the flavone. luteolin, while the inner
tissues appear colourless. On treatment with ammonia vapour, the
yellow epidermis turns orange, and the inner tissues yellow, showing
that the latter contain apigenin. These observations are confirmed by
analyses of the extracted pigments ; apigenin, only, is found in the ivory
variety: both apigenin and luteolin can be extracted and identified
from the yellow. From the true white variety no apigenin could be
extracted, nor do the flowers turn vellow when treated with ammonia

208

ANTHOCYANINS AND GENETICS

vapour. Hence we may assume that the white variety contains no
flavone.

Now let us see how this information can be connected with the
facts of Mendelian inheritance. It has been assumed that the yellow
plant contains a factor, Y, and when this is absent, the flower is white,
whatever other factors may be present. The ivory contains an addi-
tional factor, I, which is dominant to yellow. Thus ivory is YY(y)II(i)
and may, if heterozygous, throw either yellow or white, or both. Yellow
is YY(y)ii and may throw white. Hence the factor which has been
termed Y must represent the formation of luteolin in the epidermis,
and apigenin in the inner tissues of the corolla, while the factor, I,
represents the inhibition of the formation of luteolin. A better under-
standing may be reached, perhaps, by considering the constitution of
the flavones, apigenin and luteolin :

HO CO

Apigenin

HO CO

Luteolin

It will be seen that luteolin contains one more hydroxyl group than
apigenin, and this number and special grouping of hydroxyls accounts,
it is said, for the greater intensity of colour. The behaviour also of
the flavones, on heating with caustic alkali, indicates that each flavone
can be considered to be built up of phloroglucin and an oxybenzoic
acid; thus apigenin decomposes into phloroglucin and p-oxybenzoic
acid:

COOH
and luteolin into phloroglucin and protocatechuic acid

OH OH

/\

iOH

COOH

ANTHOCYANINS AND GENETICS 209

It is probable that the converse represents the synthesis in the
plant ; phloroglucin and the oxybenzoic acids and their derivatives are
widely distributed, and it is from these, no doubt, that the flavones are
synthesised. Hence the kind of oxybenzoic acid present affects the
kind of navone formed. Thus the ivory variety might be supposed
to form only one type of oxybenzoic acid, whereas the yellow forms two.
As yellow arises from ivory by the loss of a factor, we must look upon
this factor as the power of inhibiting the formation of luteolin, or the
oxybenzoic acid from which it is synthesised. This result is interesting
as a demonstration of the suggestion made by Bateson in the Presidential
Address to the British Association in 1914. He says : " I feel no reason-
able doubt that though we may have to forgo a claim to variations by
.addition of factors, yet variation both by loss of factors and by fractiona-
tion of factors is a genuine phenomenon of contemporary nature. If
then we have to dispense, as seems likely, with any addition from
without we must begin seriously to consider whether the course of
Evolution can at all reasonably be represented as an unpacking of an
original complex which contained within itself the whole range of
diversity which living things present."

As regards the true white variety 1 , it lacks either the constituents,
one of them, or both, which go to make up the flavone, or the power
of synthesising the constituents. Many interesting lines for investi-
gation are suggested by these points, both in Antirrhinum and in
allied cases. The true whites might be tested for phloroglucin and
oxybenzoic acids, and the chemical composition of white individuals
bearing the I factor might be compared with those which are derived
from yellow, and which therefore do not carry I 2 . One point which
may be emphasised with regard to the Y and I factors is that their
power is not affected in the heterozygote ; a plant may be homo- or
heterozygous for Y or I, and there