)- 
 
 > ^. 
 
 
 ■! %■ " 
 
THE COLOURS OF FLOWERS. 
 
NATURE SERIES. 
 
 . * ♦ 
 
 THE 
 
 COLOURS OF FLOWERS 
 
 AS 
 ILLUSTRATED IN THE BRITISH FLORA 
 
 BY 
 
 GRANT ALLEN 
 
 WITH ILLUSTRATIONS 
 
 ITonboit 
 MACMILLAN AND CO. 
 1882 
 
 T^t Right of Translation aitd Reproduction is Reserveii. 
 
9/z 
 
 bonbon : 
 R. Clay, Sons, and Taylor, 
 
 BREAD STREET HILL, E.C. 
 
PREFACE. 
 
 The first germ of the theory contained in this 
 little volume was originally set forth as an article 
 in the Conihill Magazine, and I have to thank the 
 courtesy of the Editor and Publishers of that 
 periodical for their kind permission to expand it 
 into its present far fuller form. 
 
 I have been encouraged thus to give it shape in 
 a more permanent dress by the friendly appre- 
 ciation of the late Mr. Darwin, who wrote to me 
 as follows with regard to the central idea of my 
 original paper — the derivation of petals from flat- 
 tened and abortive stamens : — " Many years ago 
 I thought it highly probable that petals were in 
 all cases transformed stamens. I forget (except- 
 ing the water-lily) what made me think so ; but 
 I am sure that your evolutionary argument never 
 occurred to me, as it is too striking and too appa- 
 rently valid ever to be forgotten." It appeared to 
 me that if the idea so commended itself to Mr. 
 Darwin, it might also commend itself to other 
 
vi PREFACE. 
 
 evolutionary biologists ; and I have ventured ac- 
 cordingly to work it out here to its furthest 
 legitimate conclusions. 
 
 My acknowledgments are due in the highest 
 degree to Sir John Lubbock's admirable little work 
 on British Wild Flowers in Relation to Insects, 
 and to Mr. Bentham's British Flora. I also owe 
 much to Sir Joseph Hooker's Student's Flora, 
 to Professor Sachs's Botany, and to other books 
 too numerous to mention. Personally, I have to 
 thank my friend Mr. F. T. Richards, Tutor of 
 Trinity College, Oxford, for many valuable sug- 
 gestions and corrections of which I have gladly 
 availed myself. 
 
 G. A. 
 
CONTENTS. 
 
 CHAPTER I. 
 
 PAr^E 
 THE ORIGIN OF PETALS j 
 
 CHAPTER II. 
 
 GENERAL LAW OF FROGRF.SSIVE COLOURATION 17 
 
 CHAPTER III. 
 
 VARIEGATION 
 
 CHAPTER IV. 
 
 RELAPSE AND RETROGRESSION . . . 
 
 61 
 
 71 
 
 CHAPTER V. 
 
 DEGENERATION 
 
 91 
 
 CHAPTER VI. 
 
 MISCELLANEOUS .... 
 
 IIO 
 
THE COLOURS OF FLOWERS. 
 
 CHAPTER I. 
 
 THE ORIGIN OF PETALS. 
 
 Everybody knows that flowers are rendered beautiful 
 to us by their shapes, by their perfumes, and above 
 all by their brilliant and varied colours. All people 
 who have paid any attention to botany further know 
 that not every flower is thus bright and conspicuous ; 
 as a general rule, only those blossoms which depend 
 for their fertilisation upon the visits of insects are 
 provided with special attractions of honey, scent, and 
 vivid hues. An immense number of flowering plants, 
 perhaps even the majority among them, produce 
 only small and unnoticeablc inflorescences, like those 
 of grasses, oaks, conifers, and many other field weeds 
 or forest trees. The flowers that most people observe 
 and recognise as such, are the few highly developed 
 forms which possess large expanded coloured surfaces 
 to allure the eyes of their insect fertilisers. It is with 
 flowers in this more popular and ordinary sense that 
 we shall have to deal mainly in the present little 
 » B 
 
THE COLO URS OF FLO WERS. 
 
 treatise ; and our object must be to determine, not 
 why they are all as a group brightly coloured, but 
 why this, that, or the other particular blossom should 
 possess this, that, or the other particular hue. Why 
 is the buttercup yellow, while the stitchwort is white, 
 the dog-rose pink, and the harebell blue ? Why is 
 the purple foxglove dappled inside with lurid red spots ? 
 Why are the central florets of the daisy yellow, while 
 the ray-florets are pinky-white ? Why does sky-blue 
 prevail amongst all the veronicas, while yellow pre- 
 dominates in the St. John's worts, and white in the 
 umbellates .'' These are the sort of questions which we 
 must endeavour briefly to answer, by the light of 
 modern evolutionary biology, from the point of view 
 of the function which each colour specially subserves 
 in the economy of the particular plant which 
 displays it. 
 
 The brilliant pigments of flowers usually reside in 
 the specialised organs known as petals, though they 
 are sometimes also found in the sepals and bracts, or 
 more rarely in the stamens and even in the pistil. 
 For the sake of those readers who happen to be 
 imperfectly acquainted with the subject at large (and 
 also to bring the botanical reader definitely into the 
 required point of view), it may be well to begin with 
 a very brief description of the organs which go to 
 make up a typical flower, together with a short 
 account of the part played by colour in general in 
 the fertilising function. 
 
 The essential elements in the flower are not at all 
 the showy and brilliant leaves which we usually 
 associate most with the name, but a set of compara- 
 tively small and unnoti eable organs occupying the 
 
THE ORIGIN OF PETALS. 
 
 centre of most ordinary blossoms. The simplest 
 type of flower consists of two such organs only, a 
 pistil or ovary containing an embryo seed, and a 
 stamen which produces the pollen necessary to im- 
 pregnate it. The production of seed is in fact the 
 sole function of the flower; every additional part is 
 only useful in so far as it conduces to this practical 
 end. In the most simple (though not the most 
 primitive) blossoms, fertilisation is eflected by a grain 
 of pollen from a stamen falling upon the stigma or 
 sensitive surface of the pistil, and thence sending 
 forth a pollen-tube, which penetrates the ovule or 
 embryo seed, and so impregnates it. 
 
 As a rule, however, it is not desirable that a flower 
 should be fertilised by pollen from its own stamens. 
 Mr. Darwin has shown in many cases that when a 
 pistil is fecundated by pollen from a neighbouring 
 blossom, or still better from a difl"erent plant, it sets 
 more and sounder seeds, or produces heartier and 
 stronger young seedlings. To attain the benefits of 
 such cross-fertilisation, many plants have acquired 
 special peculiarities of structure : or, to put it more 
 correctly, those plants which have spontaneously 
 varied in certain favourable directions have been 
 oftenest cross-fertilised, and have thus on the average 
 produced more and stouter offspring. The advantage 
 thus gained in the struggle for existence has enabled 
 them to live down their less adapted compeers, and 
 to hand on their own useful peculiarities to a large 
 number of descendants. There are two ways in 
 which plants have ensured such a benefit ; the one is 
 by adapting themselves to fertilisation by means of 
 the wind, the other is by adapting themselves to 
 
 B 2 
 
THE COLOURS OF FLOWERS. 
 
 fertilisatioa by means of insects. The first class arc 
 said to possess anemophilous, the second class 
 entomophiloas, flowers.^ It is with the latter alone 
 that we have here mainly to deal. 
 
 Entomophilous or insect-fertilised flowers are those 
 in which the pollen is habitually carried from the 
 stamens of one blossom to the pistil of the next on 
 the head or legs of butterflies, bees, beetles, or other 
 flying arthropods. In order to allure these insects, 
 and to induce them to visit one flower after another 
 of the same kind, the plants have often developed 
 small quantities of honey in the neighbourhood of 
 the essential organs, as well as specially coloured 
 floral leaves known as petals. Accordingly, a fully 
 evolved entomophilous blossom usually consists of 
 the four following whorls, or sets of parts, beginning 
 from within outward. Firsts in the very centre, the 
 pistil, or carpellary whorl, consisting of one or more 
 carpels or ovaries, each containing one or more embryo 
 seeds. Secondly, outside the pistil, the staminal 
 whorl, consisting of one or more pollen-bearing 
 stamens, usually three or six in the great class of 
 Monocotyledons, and five or ten in the great class 
 of Dicotyledons. Thirdly, outside the stamens, the 
 corolla or petaline whorl, consisting of several separate 
 or united petals, usually three in the Monocotyledons 
 and five in the Dicotyledons. Fourthly, outside the 
 corolla, the calyx or sepaline whorl, consisting of 
 several separate or united sepals, usually the same in 
 number as the petals. The position and arrangement 
 
 • For further particulars .see Sir Jolm Lubbock's work on British 
 Wild Flowers in Relation to Insects, in the Nature Series. 
 
THE ORIGIN OF PETALS. 
 
 of these parts is shown in the accompanying diagrams 
 (Figs. I and 2). 
 
 As regards function, the pistil produces the seeds 
 and grows later into the fruit. The stamens produce 
 pollen, to impregnate the pistil. The petals attract 
 the fertilising insects by their bright colour, and 
 advertise the honey, if any. The calyx covers up 
 the flower in the bud, and often serves to protect it 
 
 Fig. t. — Diagram of typ.cal primitive m inncotyledonoiis flower, rt, rarpels; /', 
 inner whnrl of staiiiens ; c. outer whorl of stamens; d, petals; e, sepals. Each 
 whorl consists of three members. 
 
 from the attacks of useless creeping or honey-eating 
 insects. 
 
 One more preliminary explanation is necessary 
 before we enter upon the consideration of our main 
 subject. Flowering plants, at a very early date in 
 their history, split up into two great divisions. One 
 of these, the GvMNOsrERMS, to which all the oldest 
 fossil plants belong, as well as our own modern 
 conifers and cycads, possessed and still possess 
 flowers and fruits of a very simple character. Each 
 
TH2 COLOURS OF FLOWERS. 
 
 blossom consists as a rule of' a single stamen or a 
 single naked ovule, growing on a scale or an altered 
 leaf: and they display some remarkable analogies 
 with ferns, club-mosses, and odier flowerless plants. 
 Of course, they never have petals or coloured organs. 
 The existing Gymnosperms may be regarded as 
 living survivors of a great class, once dominant, but 
 now nearly extinct ; and their flowers probably still 
 preserve for us the original type of all blossoms, very 
 
 Fiu. 2. — Diagram of typical primitive dicotyledon .us fl wer. a. carpels; b, 
 stamens ; c, petals ; </, sepals. I'.ach whorl consists of five members. 
 
 slightly altered by time and circumstances. This is 
 especially the case with the cycads, small tropical 
 trees of palm-like appearance, whose inflorescence is 
 the very simplest of all known flowering plants. The 
 other great division, that of the Angiosperms, has 
 always more fully developed blossoms, often entomo- 
 philous, and possessed of brilliant colours. It split up 
 again at an early period of its development into two 
 secondary large classes, those of the Monocotyledons 
 
THE ORIGIN OF PETALS, 
 
 and the Dicotyledons. These important classes are 
 distinguished from one another by several special 
 points of structure ; but the most striking one, so far 
 as regards our present subject, consists in the fact 
 that the Monocotyledons are generally arranged in 
 whorls of threes, while the Dicotyledons are generally 
 ananged in whorls of fives. At the present day, 
 the two modes of arrangement are often obscured in 
 various ways; for example, among some Monocotyle- 
 dons the stamens are doubled, or consist of two 
 separate whorls, while the petals and sepals are 
 coloured alike and are otherwise almost indistinguish- 
 able, so that the flowers seem at first sight to be 
 arranged by sixes rather than by threes ; and again, 
 among some Dicotyledons, one or more petals are 
 suppressed, or added, or all the petals are united into 
 a single circular or tubular corolla, so that the arrange- 
 ment seems at first sight to be by fours, or by eights, 
 or by threes rather than by fives. Originally, how- 
 ever, all Monocotyledons had a trinary arrangement, 
 while all Dicotyledons had a quinary arrangement ; 
 and these fundamental plans can still distinctly be 
 traced in by far the larger number of existing 
 species. 
 
 So much by way of introduction. No\v^, since the 
 bright pigments of flowers usually reside in the petals, 
 and since petals have for their main if not for their 
 only function the display of such pigments as an 
 attraction for the fertilising insects, it is clear that we 
 must begin our inquiry by asking, — What was the 
 original colour of the organs from which petals were 
 developed } For we may take it for granted that the 
 primordial petals would at first follow the hue of the 
 
8 THE COLOURS OF FLOWERS. 
 
 part out of which they were originally evolved. For 
 example, if we ought to regard them as altered and 
 modified leaves, then we may fairly conclude that the 
 earliest petals were green ; while, if we ought to 
 regard them as altered and modified stamens, then 
 we may fairly conclude that they were yellow. 
 Which of these two alternatives is the most likely 
 to be true ? Apparently, the latter. 
 
 Petals are in all probability originally enlarged and 
 flattened stamens, which have been set apart for the 
 special work of attracting insects. It seems likely 
 that all flowers at first consisted of the central organs 
 alone— that is to say, of a pistil, which contains the 
 ovary with its embryo seeds ; and of a few stamens, 
 which produce the pollen, whose co-operation is 
 necessary in order to fertilise these same embryo 
 ovules and to make the pistil mature into the ripe fruit. 
 But in those plants which took to fertilisation by 
 means of insects — or, one ought rather to say, in 
 those plants which insects took to visiting for the 
 sake of their honey or pollen, and so unconsciously 
 fertilising — the flowers soon began to produce an 
 outer row of barren and specialised stamens, adapted 
 by their size and colour for attracting the fertilising 
 insects ; and these barren and specialised stamens 
 are what we commonly call petals. Any flowers 
 which thus presented brilliant masses of colour to 
 allure the eyes of the beetles, the bees, and the 
 butterflies, would naturally receive the greatest 
 number of visits from their insect friends, and would 
 therefore stand the best chance of setting their seeds, 
 as well as of producing healthy and vigorous offspring 
 as the result of a proper cross. In this way, as we 
 
THE ORIGIN OF PETALS. 
 
 have seen, they would gain an advantage in the 
 struggle for life over their less fortunate compeers, 
 and would hand down their own peculiarities to their 
 descendants after them. 
 
 But as the stamens of almost all flowers, certainly 
 of all the oldest and simplest flowers, are yellow, it 
 would seem naturally to follow that the earliest petals 
 would be yellow too. When the stamens of the 
 outer row were flattened and broadened into petals, 
 there would be no particular reason why they should 
 change their colour ; and, in the absence of any good 
 reason, they doubtless retained it as before. Indeed, 
 we shall see, a little later on, that the earliest and 
 simplest types of existing flowers are almost always 
 yellow, seldom white, and never blue; and this in 
 itself would be a sufficient ground for believing that 
 yellow was the original colour of all petals. But as 
 it is somewhat heretical to believe, contrary to the 
 general run of existing scientific opinion, that petals 
 are derived from flattened stamens, instead of from 
 simplified and attenuated leaves, it may be well to 
 detail here the reasons for this belief, because it 
 seems of capital importance in connection with our 
 present subject ; for if the petals were originally a 
 row of altered stamens set apart for the special 
 function of attracting insects, it would be natural 
 and obvious why they should begin by being yellow ; 
 but if they were originally a set of leaves, which 
 became thinner and more brightly coloured for the 
 same purpose, it would be difficult to see why they 
 should first have assumed any one colour rather than 
 another. 
 
 The accepted doctrine as to the nature of petals is 
 
lo THE COLOURS OF FLOWERS. 
 
 tliat discovered by Wolf and subsequently redis- 
 covered by Goethe, after whose name it is usually 
 called ; for of course, as in all such cases, the gi cater 
 man's fame has swallowed up the fame of the lesser. 
 Goethe held that all the parts of the flower were 
 really modified leaves, and that a gradual transition 
 could be traced between them, from the ordinary 
 leaf, through the stem-leaf and the bract, to the 
 sepal, the petal, the stamen, and the ovary or carpel. 
 Now, if we look at most modern flowers, such a 
 transition can undoubtedly be observed ; and somc- 
 t'mes it is very delicately graduated, so that you can 
 hardly say where each sort of leaf merges into the 
 next. But, unfortunately for the truth of the theory 
 as ordinarily understood, we now know that in the 
 earliest flowers there were no petals or sepals, but 
 that primitive flowering plants had simply leaves on 
 the one hand, and stamens and ovules on the other. 
 The oldest types of flowers at present surviving are 
 certain Gymnosperms, such as the cycads, of which 
 the well-known Zaniias of our conservatories may be 
 regarded as good examples. These have only naked 
 ovules on the one hand and clusters of stamens in a 
 sort of cone on the other. The Gymnosperms arc 
 geologically earlier than any other flowering plants. 
 But, if petals and sepals are later in origin (as we 
 know them to be) than stamens and carpels, w^e can 
 hardly say that they mark the transition from one 
 form to the other, any more than we can say that 
 Gothic architecture marks the transition from the 
 Egyptian style to the classical Greek. It is not 
 denied, indeed, that the stamen and the ovary are 
 themselves by origin modified leaves — that part of 
 
THE ORIGIN OF PETALS. ir 
 
 the Wolfian theory is absolutely irrefutable — but with 
 the light shed upon the subject by the modern 
 doctrine of evolution, we can no longer regard petals 
 and sepals as intermediate stages between the two. 
 The earliest flowering plants had true leaves on the 
 one hand, and specialised pollen-bearing or ovule- 
 bearing leaves on the other hand, which latter are 
 what in their developed forms we call stamens and 
 carpels ; but they certainly had no petals at all, and 
 the petals of modern flowers have been produced at 
 some later period. It is probable, too, that they have 
 been produced by a modification of certain external 
 stamens, not by a modification of true leaves. Instead 
 of being leaves arrested on their way towards 
 becoming stamens, they are really stamens which 
 have partially reverted towards the condition of 
 leaves. They differ from true leaves, however, in 
 their thin, spongy texture, and usually in the bright 
 pigments with which they are adorned. 
 
 All stamens show a great tendency easily to become 
 petaloid, as it is technically called ; that is to say, 
 to flatten out their filament or stalk, and finally to 
 lose their pollen-bearing sacs or anthers. In the 
 waterlilies — which are in certain ways one of the 
 oldest and simplest types of flowers we now possess 
 still preserving many antique points of structure 
 unchanged — we can trace a regular gradation from 
 the perfect stamen to the perfect petal. Take for 
 example our common English white waterlily, 
 Nymph(Ea alba (Fig. 3). In the centre of the flower 
 we find stamens of the ordinary sort, with rounded 
 stalks or filaments, and long yellow anthers full of 
 pollen at the end of each ; then, as we move outward, 
 
12 
 
 THE COLOURS OF FLOWERS. 
 
 we find the filaments growing flatter and broader, and 
 the pollen-sacs less and less perfect ; next we find a 
 few stamens which look exactly like petals, only that 
 they have two abortive anthers stuck awkwardly on to 
 their summits ; and, finally, we find true petals, broad 
 and flat, and without any trace of the anthers at all. 
 Here in this very ancient though largely modified 
 flower we have stereotyped for us, as it were, the 
 mode in which stamens first developed into petals, 
 under stress of insect selection.^ 
 
 fiG 3.— Trsr.sition fr:in stamens into petals in white waterlily {Myin/>/uea nlhi). 
 
 " But how do we know," it may be asked, " that the 
 transition was not in the opposite direction ? How 
 do we know that the waterlily had not petals alone to 
 start with, and that these did not afterwards develop, 
 as the Wolfian hypothesis would have us believe, into 
 stamens ? " For a very simple reason. The theory 
 of Wolf and Goethe is quite incompatible with the 
 doctrine of development, at least if accepted as a 
 
 ^ The waterlilies belong to a very ancient type, in some respects 
 partially intermediate between jVIonocotyledons and Dicotyledons ; but 
 the comparative unification of their pistil shows them to have under- 
 jione considerable modification. 
 
THE ORIGIN OF PETALS. 13 
 
 historical explanation (which Wolf and Goethe of 
 course never meant it to be). Flowers can and do 
 exist without petals, which are no essential part of 
 the organism, but a mere set of attractive coloured 
 advertisements for alluring insects ; but no flower can 
 possibly exist without stamens, which are one of the 
 two essential reproductive organs in the plant. With- 
 out pollen, no flower can set its seeds. A parallel 
 from the animal world will make this immediately 
 obvious. Hive-bees consist of three kinds — the 
 queens or fertile females, the drones or males, and 
 the workers or neuters. Now it would be absurd to 
 ask whether the queens were developed from an 
 original class of neuters, or the neuters from an 
 original class of fertile females. Neuters left to them- 
 selves would die out in a single generation : they are 
 really sterilised females, set apart for a special function 
 on behalf of the hive. It is just the same with petals : 
 they are sterilised stam'ins, set apart for the special 
 function of attracting insects on behalf of the entire 
 flower. But to ask which came first, the petals or the 
 stamens, is as absurd as to ask which came first, the 
 male and female bees or the neuters. 
 
 Indeed, if we examine closely the waterlily petals, 
 it is really quite impossible to conceive of the transi- 
 tion as taking place from petals to stamens instead of 
 from stamens to petals. It is quite easy to under- 
 stand how the filament of an active stamen may 
 become gradually flattened, and the anthers (or pollen- 
 sacs) progressively void and functionless : but it is 
 very difficult to understand how or why a petal should 
 first begin to develop an abortive anther, and then a 
 partially effective anther, and at last a perfect stamen. 
 
14 THE COLOURS OF FLOWERS. 
 
 The one change is comprehensible and reasonable, 
 the other change is meaningless and absurd. Of 
 course, it is not intended to deny the truth of Wolf's 
 great generalisation in the way in which he meant it 
 — the existence of a homology between the leaf and 
 all the floral organs : but the conception certainly 
 requires to be modified a little by the light of later 
 evolutionary discoveries. The starting-point consists 
 of a plant having three kinds of organs, true foliage 
 leaves, staminal leaves, and carpellary leaves : the 
 petals and sepals are apparently later intermediate 
 modifications, produced in special connection with the 
 acquired habit of insect fertilisation. 
 
 In many other cases besides the waterlily, we know 
 that stamens often turn into petals. Thus the 
 numerous coloured rays of the MesembryantJiemiims 
 or ice-plant family are acknowledged by many 
 botanists to be flattened stamens. In Cmma, w^here 
 one anther-cell is abortive, the filament of the solitary 
 stamen becomes petaloid. In the Ginger order, one 
 outer whorl of stamens resembles the tubular corolla, 
 so that the perianth seems to consist of nine lobes 
 instead of six. In orchids, according to Mr. Darwin, 
 the lip consists of one petal and two petaloid stamens 
 of the outer whorl. In double roses (Fig. 4) and almost 
 all other double flowers the extra petals are produced 
 front the stamens of the interior. In short, stamens 
 generally can be readily converted into petals, especi- 
 ally in rich and fertile soils or under cultivation. The 
 change is extremely common in the families of Ran- 
 iinctdacecB, Papaveraccce^ Magnoliacea^, Malvacece, and 
 RosacecB, all very simple types. Even where stamens 
 always retain their pollen-sacs, they have often broad, 
 
THE ORIGIN OF PETALS. 
 
 15 
 
 flattened petaloid filaments, as in the starof Rethleheiri 
 and many other flowers. The curious scales on the 
 petals of Parnassia palustris are now known to be 
 altered stamens. Looking at the question as a whole, 
 we can see how petals might easily have taken theii 
 origin from stamens, while it is difficult to understand 
 how they could have taken their origin from ordinary 
 leaves — a process of which, if it ever took place, no 
 hint now remains to us. We shall see hereafter that 
 the manner in which certain outer florets in the 
 compound flower-heads of the daisy or the aster 
 have been sterilised and specialised for the work of 
 
 \J i, ^J ^0 1 
 
 Fig. 4.— Transition from stamen («) to petal {U) and sepal (c) in flower of doubl 
 
 e rose. 
 
 attraction, affords an exact analogy to the manner in 
 which it is here suggested that certain stamens may 
 at an earlier date have been sterilised and specialised 
 for the same purpose, thus giving rise to what we 
 know as petals. 
 
 In a few rare instances, petals even now show a 
 slight tendency to revert to the condition of 
 fertile stamens. In Monaudra fistulosa the lower lip 
 is sometimes prolonged into a filament bearing an 
 anther: and the petals of shepherd s-purse {Capsella 
 Imrsa-pastoris) have been observed anthcriferous. 
 
 The hypothesis upon which we shall hereafter 
 
i6 THE COLOURS OF FLOWERS. 
 
 proceed, therefore, will be that petals are really de- 
 rived from altered stamens. We shall return at a later 
 point to the proofs of this position, and examine a 
 few arguments which may be brought against it. For 
 the present, it will be better to put forward the 
 remainder of our general theory at once, without 
 interrupting the exposition by any alien controversial 
 matter. For the most part, it must find its evidence 
 in its perfect congruity with all the established facts 
 of the science. 
 
CHAPTER II. 
 
 GENERAL LAW OF PROGRESSIVE COLOURATION. 
 
 If the earliest petals were derived from flattened 
 stamens, it would naturally follow that they would be 
 for the most part yellow in colour, like the stamens 
 from which they took their origin. How, then, did 
 some of them afterwards come to be white, orange, 
 red, purple, lilac, or blue ? A few years ago, when the 
 problem of the connection between flowers and insects 
 still remained much in the state where Sprengel left 
 it at the end of the last century, it would have seemed 
 quite impossible to answer this question. But now- 
 adays, after the full researches of Darwin, Wallace, 
 Lubbock, and Hermann M tiller into the subject, we 
 can give a very satisfactory solution indeed. We now 
 know, not only that the colours of flowers as a whole 
 are intended to attract insects in general, but that 
 certain colours are definitely intended to attract 
 certain special kinds of insects. Thus, to take a few 
 examples only out of hundreds that might be c"ted, 
 the flowers which lay themselves out for fertilisation 
 by miscellaneous small flies are almost always white ; 
 those which depend upon the beetles are frequently 
 yellow ; while those which specially bid for the favour 
 
 C 
 
1 8 THE COLOURS OF FLOWERS. 
 
 of bees and butterflies are usually red, purple, lilac, or 
 blue. Certain insects always visit one species of 
 flower alone ; and others pass from blossom to 
 blossom of one kind only on a single day, though 
 they may vary a little from kind to kind as the season 
 advances, and one species replaces another. Mijller, 
 the most statistical of naturalists, has noticed that 
 while bees form seventy-five per cent, of the insects 
 visiting the very developed composites, they form 
 only fourteen per cent, of those visiting umbelliferous 
 plants, which have, as a rule, open but by no means 
 showy white flowers. Certain blossoms which lay 
 themselves out to attract wasps are, as he quaintly 
 puts it, '* obviously adapted to a less aesthetically 
 cultivated circle of visitors." And some livid red 
 flowers actually resemble in their colour and odour 
 decaying raw meat, thus inducing bluebottle flies to 
 visit them and so carry their pollen from head to 
 head. 
 
 Down to the minutest distinctions between species, 
 this correlation of flowers to the tastes of their par- 
 ticular guests seems to hold good. Hermann Miiller 
 notes that the common Galium of our heaths and 
 hedges {G. inollugo) is white, and therefore visited by 
 small flies ; while the lady's bedstraw, its near relative 
 {G. venivi)^ is yellow, and owes its fertilisation to little 
 beetles. Mr. H. O. Forbes counted on one occasion 
 the visits he saw paid to the flowers on a single bank ; 
 and he found that a particular bumble-bee sucked the 
 honey of thirty purple dead nettles in succession, 
 passing over without notice all the other plants in 
 the neighbourhood ; two other species of bumble-bee 
 and a cabbage-butterfly also patronised the same dead- 
 
LA W OF PROGRESSIVE COLOUR A TION. 19 
 
 nettles exclusively. Fritz Miiller noticed a Lantana in 
 South America which changes colour as its flc -ering 
 advances ; and he observed that each kind of butterfly 
 which visited it stuck rigidly to its own favourite colour, 
 waiting to pay its addresses until that colour appeared. 
 Mr. Darwin cut off the petals of a lobelia and found that 
 the hive-bees never went near it, though they were 
 very busy with the surrounding flowers. But perhaps 
 Sir John Lubbock's latest experiments on bees arc 
 the most conclusive of all. He had long ago con- 
 vinced himself, by trials with honey placed on slips of 
 glass above yellow, pink, or blue paper, that bees 
 could discriminate the different colours ; and he has now 
 shown in the same way that they display a marked 
 preference for blue over all other hues. The fact is, 
 blue flowers are, as a rule, specialised for fertilisation 
 by bees, and bees therefore prefer this colour ; while 
 conversely the flowers have at the same time become 
 blue because that was the colour which the bees 
 prefer. As in most other cases, the adaptation must 
 have gone on pari passu on both sides. As the bee- 
 flowers grew bluer, the bees must have grown fonder 
 and fonder of blue ; and as they grew fonder of blue, 
 they must have more and more constantly preferred 
 the bluest flowers. 
 
 We thus see how the special tastes of insects may 
 have become the selective agency for developing 
 white, pink, red, purple, and blue petals from the 
 original yellow ones. But before they could exercise 
 such a selective action, the petals must themselves 
 have shown some tendency to vary in certain fixed 
 directions. How could such an original tendency 
 arise ? For, of course, if the insects never saw any 
 
 C 2 
 
20 THE COLOURS OF FLOWERS. 
 
 pink, purple, or blue petals, they could not specially 
 favour and select them ; so that we are as yet hardly 
 nearer the solution of the problem than ever. 
 
 Here Mr. Sorby, who has chemically studied the 
 colouring matter of leaves and flowers far more deeply 
 than any other invest'gator, supplies us with a useful 
 hint. He tells us that the various pigments of bright 
 petals are already contained in the ordinary tissues of 
 the plant, whose juices only need to be slightly 
 modified in chemical constitution in order to make 
 them into the blues, pinks, and purples with which we 
 are so familiar. "The coloured substances in the 
 petals," he says, " are in many cases exactly the same 
 as those in the foliage fron which chlorophyll has 
 disappeared ; so that the petals are often exactly like 
 leaves which have turned yellow and red in autumn, 
 or the very yellow or red leaves of early spring." " The 
 colour of many crimson, pink, and red flowers is due 
 to the development of substances belonging to the 
 erythrophyll group, and not unfrequently to exactly 
 the same kind as that so often found in leaves. The 
 facts seem to indicate that these various substances 
 may be due to an alteration of the normal constituents 
 of leaves. So far as I have been able to ascertain, 
 their development seems as if related to extra oxidi- 
 sation modified by light and other varying conditions 
 not yet understood.'* 
 
 The different hues assumed by petals are all thus, 
 as it were, laid up beforehand in the tissues of the 
 plant, ready to be brougVt out at a moment's notice. 
 And all flowers, as we know, easily sport a little in 
 colour. But the question is, do their changes tend to 
 follow any regular and definite order } Is there any 
 
LA IV OF PROGRESSIVE COLOUR A TION. 2 1 
 
 reason to believe that the modification runs from any- 
 one colour toward any other ? Apparently there is. 
 The general conclusion to be set forth in this work is 
 tlie statement of such a tendency. All flowers, it 
 would seem, were in their earliest form yellow ; then 
 some of them became white ; after that, a few of them 
 grew to be red or purple ; and finally, a comparatively 
 small number acquired various shades of lilac, mauve, 
 violet, or blue. So that, if this principle be true, such 
 I flower as the harebell will represent one of the 
 most highly-developed lines of descent ; and its 
 ancestors will have passed successively through all the 
 intermediate stages. Let us see what grounds can be 
 given for such a belief. 
 
 Some hints of a progressive law in the direction of 
 a colour-change from yellow to blue are sometimes 
 afforded us even by the successive stages of a single 
 flower. For example, one of our common little 
 English forget-me-nots, Myosotis versicolor, is pale 
 yellow when it first opens ; but as it grows older, it 
 becomes faintly pinkish, and ends by being blue like 
 the others of its race. Now, this sort of colour-change 
 is by no means uncommon ; and in almost all known 
 cases it is always in the same direction, from j'cllow 
 or white, through pink, orange, or red, to purple or 
 blue. For example, one of the wall-flowers, CJiciran- 
 tJiiis cliaincelco, has at first a whitish flower, then a 
 citron-yellow, and finally emerges into red or violet. 
 The petals of Stylidiuni fruticosuDi are pale yellow to 
 begin with, and afterwards become light rose-coloured. 
 An evening primrose, Oenothera tctraptera, has white 
 flowers in its first stage and red ones at a later period 
 of development. Cobcca scandcns goes from white to 
 
22 THE COL O URS OF FL WERS. 
 
 violet ; Hibiscus miitabilis from white, through flesh- 
 coloured, to red. The common Virginia stock of our 
 gardens {Malcobnia) often opens of a pale yellowish 
 green ; then becomes faintly pink ; afterwards deepens 
 into bright red ; and fades away at last into mauve or 
 blue. Fritz M tiller's Laiitana is yellow on its first 
 day, orange on the second, and purple on the third. 
 The whole family of Boraginacece begin by being pink 
 and end with being blue. The garden convolvulus 
 opens a blushing white and passes into full purple. 
 In all these and many other cases the general direc- 
 tion of the changes is the same. They are usually 
 set down as due to varying degrees of oxidation in 
 the pigmentary matter. 
 
 If this be so, there is a good reason why bees should 
 be specially fond of blue, and whyblue flowers should 
 be specially adapted for fertilisation by their aid. For 
 Mr. A. R. Wallace has shown that colour is most apt 
 to appear or to vary in those parts of plants or animals 
 which have undergone the highest amount of modifi- 
 cation. The markings of the peacock and the argus 
 pheasant come out upon their immensely developed 
 secondary tail-feathers or wing-plumes ; the metallic 
 hues of sun-birds and humming-birds show themselves 
 upon their highly-specialised crests, gorgets, or lap- 
 pets. It is the same with the hackles of fowls, the 
 head-ornaments of fruit pigeons, and the bills of 
 toucans. The most exquisite colours in the insect 
 world are those which are developed on the greatly 
 expanded and delicately-feathered wings of butter- 
 flies ; and the eyt-spots which adorn a few species are 
 usually found on their very highly modified swallow- 
 tail appendages. So, too, with flowers ; those which 
 
LAPV OF PROGRESSIVE COLOURATION. 23 
 
 have undergone most modification have their colours 
 most profoundly altered. In this way, we may put 
 it down as a general rule (to be tested hereafter) 
 that the least developed flowers are usually yellow 
 or white ; those which have undergone a little 
 more modification are usually pink or red ; and 
 those which have been most highly specialised of 
 any are usually purple, lilac, or blue. Absolute 
 deep ultramarine probably marks the highest level 
 of all. 
 
 On the other hand, Mr. Wallace's principle also 
 explains why the bees and butterflies should prefer 
 these specialised colours to all others, and should 
 therefore select the flowers Vv'hich display them by 
 preference over any less developed types. For bees 
 and butterflies are the most highly adapted of all 
 insects to honey-seeking and flower-feeding. They 
 have themselves on their side undergone the largest 
 amount of specialisation for that particular function. 
 And if the more specialised and modified flowers, 
 which gradually fitted their forms and the position of 
 their honey-glands to the forms of the bees or 
 butterflies, showed a natural tendency to pass from 
 yellow through pink and red to purple and blue, it 
 would follow that theinsects which were being evolved 
 side by side with them, and which were aiding at the 
 same time in their evolution, would grow to recognise 
 these developed colours as the visible symbols of 
 those flowers from which they could obtain the largest 
 amount of honey with the least possible trouble. 
 Thus it would finally result that the ordinary un- 
 specialised flowers, which depended upon small insect 
 riff-raff, would be mostly left yellow or white ; those 
 
24 THE COLOURS OF FLOWERS, 
 
 which appealed to rather higher insects would become 
 pink or red ; and those which laid themselves out for 
 bees and butterflies, the aristocrats of the arthropo- 
 dous world, would grow for the most part to be purple 
 or blue. 
 
 Now, this is very much what we actually find to be 
 the case in nature. The simplest and earliest flowers 
 are those with regular, symmetrical, open cups, like 
 t\\i^. Ranuncidus genus, the Potcutillas, and the Alshiece^ 
 or chickweeds, which can be visited by any insects 
 whatsoever; and these are in large part yellow or 
 white. A little higher are flowers like the campions 
 or Sileneco^ and the stocks {Matthio/a), with more or 
 less closed cups, whose honey can only be reached by 
 more specialised insects ; and these are oftener pink 
 or reddish. More profoundly modified are those 
 irregular one-sided flowers, like the violets, peas, and 
 orchids, which have assumed special shapes to accom- 
 modate bees or other specific honey-seekers ; and 
 these are often purple and not unfrequently blue. 
 Highly specialised in another way are the flowers like 
 harebells {Campanula), scabious {Dipsace(E), and heaths 
 {Ericaceae), whose petals have all coalesced into a 
 tubular corolla ; and these might almost be said to be 
 usually purple or blue. And, finally, highest of all 
 are the flowers like labiates (rosemary, Salvia, &c.) 
 and speedwells {Veronica), whose tubular corolla has 
 been turned to one side, thus combining the united 
 petals with the irregular shape ; and these are almost 
 invariably purple or blue. 
 
 We shall proceed to give a few selected examples 
 from the families best represented in the British 
 flora. 
 
LA IV OF PROGRESSIVE COLOURATION. 
 
 = 5 
 
 The very earliest types of angiospermous flowers 
 x\o\w remaining are those in which the carpels still 
 exist in a separate form, instead of being united into 
 a single compound ovary. Among Dicotyledons, the 
 families, some of whose members best represent this 
 primitive stage, are the Rosacccs and Raminculaccce ; 
 among Monocotyledons, the Alismacece. We may 
 conveniently begin with the first group. 
 
 Fig. 5 — Flower of cinquefoil yPotcntilUi). Prijiiitive yelLw, 
 
 The roses form a most instructive family. As a 
 whole they are not very highly developed flowers, 
 since all of them have simple, open, symmetrical 
 blossoms, generally with five distinct petals. But of 
 all the rose tribe, the Potentillece, or cinquefoil group, in- 
 cluding our common English silver-weed, seem to make 
 up the most central, simple, and primitive members 
 (^'S- 5)- They are chiefly low, creeping weeds, and 
 their flowers are of the earliest symmetrical pattern, 
 
26 THE COLOURS OF FLOWERS. 
 
 v/ithout any specialisation of form, or any peculiar 
 adaptation to insect visitors. Now, among the poten- 
 tilla group, nearly all the blossoms have yellow petals, 
 and also the filaments of the stamens yellow, as is 
 likewise the case with the other early allied forms, 
 such as agrimony {Agriinonia Eiipatorid) and herb- 
 bennet {Getim tirbafuiift). Among our common yellow 
 species are Potentilla reptans (cinquefoil), P. tonnai- 
 tilia, P. argentea, P. verna, P. fi'uticosa, and P. anse- 
 rina. Almost the only white potentillas in England 
 are the barren ;strawberry {P . fragariastriivi) and the 
 true strawberry [Fragaria vesca), which have, in many 
 ways, diverged more than any other species from the 
 norma of the race. Water-avens {Geiun rivale), how- 
 ever, a close relative of herb-ben net, has a dusky 
 purplish tinge; and Sir John Lubbock notes that it 
 secretes honey, and is far oftener visited by insects 
 than its kinsman. The bramble tribe (Rtibece), including 
 the blackberry (Fig. 6), raspberry, and dewberry, have 
 much larger flowers than the potentillas, and are very 
 greatly frequented by winged visitors. Their petals 
 are usually pure white, often with a pinky tinge, 
 especially on big, well-grown blossoms. But there is 
 one low, little-developed member of the blackberry 
 group, the Rubus saxatiiis, or stone-bramble, with 
 narrow, inconspicuous petals of a greenish-yellow, 
 merging into dirty white ; and this humble form seems 
 to preserve for us the transitional stage from the yellow 
 potentilla to the true white brambles. One step 
 higher, the cherries and apples (though genetically 
 unconnected), have very large and expanded petals, 
 white toward the centre, but blushing at the edges into 
 rosy pink or bright red (Fig. 7). We shall see hereafter 
 
LAW OF PROGRESSIVE COLOURATION. 27 
 
 that new colours always make their appearance at the 
 outer side of the petal, while the base usually retains 
 its primitive colouration. For the present, this prin- 
 
 FiG. 6 — Vertical section of bramble-flower {Rubus). White. 
 
 ciple must be accepted on trust. Finally, the true roses 
 (Fig- 8), whose flowers are the most developed of all, 
 have usually broad pink petals (like those of our own 
 
 Fig. 7.— Vertical section of apple-blossom {Pyrus vialus). Pinky white. 
 
 dog-rose, Rosa canina, R. villosa, R. riibigmosa, &c.), 
 which, in some still bigger exotic species, become 
 crimson or damask of the deepest dye. They are 
 
2^ 
 
 THE COLOURS OF FLOWERS. 
 
 more sought after by insects than any others of their 
 family. 
 
 Now, if we look closely at these facts, we see that 
 they have several interesting implications. The yellow 
 potentillas have the very simplest arrangement of the 
 carpels in the whole family, and their fruit is of the 
 most primitive character, consisting only of little dry 
 separate nuts. They have altered very little from 
 their primitive type. Accordingly, almost all the 
 genus is yellow ; a very few members only are white ; 
 and these in their habits so far vary from the rest that 
 
 Ki ;. 8.— Vertical section of dog-ro>e (^Rosa). Pink 
 
 they have very erect flowers, and three leaflets instead 
 of five or more to each leaf. One of them, the straw- 
 berry, shows still further marks of special differentia- 
 tion, in that it has acquired a soft, pulpy, red fruit, pro- 
 duced by the swelling of the receptacle, and adapted 
 to a safer mode of dispersal by the aid of birds. This 
 group, however, including Geum, cannot claim to be 
 considered the earliest ancestral form of the roses, 
 because of its double calyx, which is not shared by 
 other members of the family, as it would be if it had 
 belonged to the actual, common ancestor. In that 
 
LA VV OF PROGRESSIVE COLOURA TIOX. 29 
 
 respect, agrimony more nearly represents the primitive 
 form, though its tall habit and large spikes of flowers 
 show that it also has undergone a good deal of modi- 
 fication. Nevertheless, the yellow members of the 
 potentilla group, in their low creeping habit, their 
 want of woodiness, and their simple fruit, certainly 
 remain very nearly at the primitive ancestral stage, 
 and may be regarded as very early types of flowers 
 indeed. It is only among handsome and showy 
 exotic forms, which have undergone a good deal more 
 modification, that we get brilliant red-flowered species 
 like the East Indian P. ncpalensis and P. atropiirpiirea. 
 But as soon as the plants rise a little in the scale, 
 and the flowers grow larger, we get a general tendency 
 towards white and pink blossoms. Thus the Prunecs 
 have diverged from the central stock of the rose 
 family in one direction, and the Potnecs and Rosece in 
 another ; but both alike begin at once to assume white 
 petals ; and as they lay themselves out more and more 
 distinctly for insect aid, the white passes gradually 
 into pink and rose-colour. To trace the gradations 
 throughout, we see that the Riibccs^ or brambles, are 
 for the most part woody shrubs instead of being mere 
 green herbs, and they have almost all whitish blossoms 
 instead of yellow ones ; but their carpels still remain 
 quite distinct, and they seldom rise to the third stage 
 of pinkiness ; when they do it is generally just as 
 they are fading, and we may lay it down as a 
 common principle that the fading colours of less 
 developed petals often answer to the normal colours 
 of more developed. In the Prunece^ aga'n, the de- 
 velopment has gone much further, for here most of 
 the species are trees or hard shrubs, and the number 
 
30 THE COLOURS OF FLOWERS. 
 
 of carpels is reduced to one. They have a succulent 
 fruit — a drupe, the highest type — and though the 
 flower contains two ovules, the ripe plum has only 
 one seed, the other having become abortive. All 
 these are marks of high evolution : indeed, in most 
 respects the Priinecs stand at the very head of the rose 
 family ; but the petals are seldom very expanded, and 
 so, though they are usually deeply tinged with pink 
 in the cherry {Prunus cerasiis)^ and still more so in 
 larger exotic blossoms, like the almond, the peach, 
 and the nectarine, they seldom reach the stage of red. 
 Our own sloe (/*. communis) has smallish white 
 flowers, as has also the Por (gal laurel {P. lusi- 
 tanicus). In these plants, in fact, higher development 
 has not largely taken the direction of increased attrac- 
 tion for insect fertilizers; it has mainly concentrated 
 itself upon the fruit, and the devices for its dispersal 
 by birds or m-ammals. In the Rosecs, on the other 
 hand, though the fruit is less highly modified, the 
 methods for ensuring insect fertilisation are carried 
 much further. There are several carpels, but they 
 are inclosed within the tube of the calyx, and the 
 petals are very much enlarged indeed, while in some 
 species the styles are united in a column. As regards 
 insect-attraction, indeed, the roses are the most ad- 
 vanced members of the family, and it is here, accord- 
 ingly, that we get the highest types of colouration. 
 Most of them are at least pink, and many are deep 
 red or crimson. Among the Poinece, we find an inter- 
 mediate type (as regards the flowers alone) between 
 RosecB and Prunecc ; the petals are usually bigger and 
 pinker than those of the plums ; not so big or so pink 
 as those of the true roses. This interesting series 
 
LAW OF PROGRESSIVE COL O URA TION. 3 1 
 
 exhibits very beautifully the importance as regards 
 colouration of mere expansion in the petals. Taking 
 them as a whole we may say that the smallest petals 
 in the rose family are generally yellow; the next in 
 size are generally white ; the third in order are gene- 
 rally pink ; and the largest are generally rose-coloured 
 or crimson. 
 
 At the same time, the roses as a whole, being a 
 relatively simple family, with regular symmetrical 
 flowers of the separate or polypetalous type, have 
 never risen to the stage of producing blue petals. 
 That, probably, is why our florists cannot turn out a 
 blue rose. It is easy enougii to make roses or any 
 other blossoms vary within their own natural limits, 
 revert to any earlier form or colour through which 
 they have previously passed ; but it is difficult or 
 impossible to make them take a step which they have 
 never yet naturally taken. Hence florists generally 
 find the most developed flowers are also the most 
 variable and plastic in colour ; and hence, too, we can 
 get red, pink, white, straw-coloured, or yellow roses, 
 but not blue ones. This would seem to be the 
 historical truth underlying De Candolle's division of 
 flowers into a xanthic and cyanic series. Of course, 
 there is nothing to prevent florists from developing a 
 blue rose in the same way as the insects would do it, 
 by gradually selecting and preserving the most bluish 
 or slate-coloured among their pink or crimson kinds. 
 Hut it would appear from the comparative rarity of 
 blue flowers in nature that the spontaneous variations 
 which make towards blue are far less frequent than 
 those which make towards pink, red, purple, or 
 orange. 
 
32 THE COLOURS OF FLOWERS. 
 
 There is one small set of rosaceous plants which 
 exhibit green flowers, such as the genera Alchemilla 
 and Poterium. When we come to consider the sub- 
 ject of degeneration, however, we shall see that these 
 are not really primitive blossoms, but very degraded 
 and altered types. For the present, it must suffice to 
 point out that they have lost some of their sepals, all 
 the'r petals, and many of their carpels ; and that they 
 cannot therefore be regarded in any way as represen- 
 tatives of the central primordial stock from which the 
 roses are originally derived. This place certainly 
 belongs rather to Fotentilla^ Agrimonia^ and some 
 allied exotic types, with simole regular yellow 
 blossoms. 
 
 Even more primitive in type than the RosacecE are 
 the lowest members of the Rammculacece, or buttercup 
 family, which perhaps best of all preserve for us the 
 original features of the early dicotyledonous flowers. 
 The family is also more interesting than that of the roses 
 because it contains greater diversities of development, 
 and accordingly covers a wider range of colour, its 
 petals varying from yellow to every shade of crimson, 
 purple, and blue. The simplest and least differentiated 
 members of the group are the common meadow butter- 
 cups (Fig. 9), forming the genus Ranunadus, which, as 
 everybody knows, have five open petals of a brilliant 
 golden hue. Nowhere else is the exact accordance in 
 tint between stamens and petals more noticeable 
 than in these flowers. The colour of the filaments 
 is exactly the same as that of the petals ; and the 
 latter are simply the former a little expanded and 
 deprived of their anthers. We have several English 
 meadow species, all with separate carpels, and all very 
 
LA IV OF PROGRESSIVE COLOURATION. 
 
 33 
 
 primitive in organization, such as R. acris (the central 
 form), R. bulbos7is, R. reputs, R.Jlaminula, R. sceierattis, 
 R. auricomus, R. philonotiSy &c. In the lesser celan- 
 dine or pilewort, R. ficaria, there is a slight divergence 
 from the ordinary habit of the genus, in that the 
 petals, instead of being five in number, are eight or 
 nine, while the sepals are only three; and this 
 divergence is accompanied by two slight variations in 
 colour : the outside of the petals tends to become 
 slightly reddish or purplish, and the flowers fade 
 
 / .^-* 
 
 J 
 
 
 Fig. 9. — Vertical section of buttercup (Kanunculun acris) ; primitive yellow. 
 
 white, much more distinctly than in most other 
 species of the genus. 
 
 There aro two kinds of buttercup in England, how- 
 ever, which show us the transition from yellow to white 
 actually taking place under our very eyes. These 
 are the water-crowfoot, R. aqiiatilis (Fig. 9^), and its 
 close ally the ivy-leaved crowfoot, R. hederaceiis, whose 
 petals are still faintly yellow toward the centre, but 
 fade away into primrose and white as they approach 
 the edge. We have already noticed that new colours 
 usually appear at the outside, while the claw or base 
 
 D 
 
34 
 
 THE COLOURS OF FLOWERS. 
 
 of the petal retains its original hue ; and this law is 
 strikingly illustrated in these two crowfoots. It is 
 remarkable, too, that in this respect they closely 
 resemble the half-faded flowers of the lesser celandine, 
 which become white from the edge inward as they 
 die. The petals also similarly vary in number, 
 though to a less extent. White flowers of the same 
 type as those of water-crowfoot are very common 
 among aquatic plants of like habit, and they seem to 
 be especially adapted to water-side insects. 
 
 Fig. 9*. — Flower of water crowfoot {Rnntmcultts aquatilis); white, with yellow 
 
 claws. 
 
 In many RanunculacecB there is a great tendency 
 for the sepals to become petaloid, and this peculiarity 
 is very marked in Caltha pahistris , the marsh-marigold, 
 which has no petals, but bright yellow sepals, so 
 that it looks at first sight exactly like a very large 
 buttercup. 
 
 The clematis and anemone, which are more highly 
 developed, have white sepals (for the petals here also 
 are suppressed), even in our English species ; and 
 exotic kinds varying from pink to purple are culti- 
 vated in our flower-gardens. 
 
LAW OF PROGRESSIVE COLOURATION. 
 
 35 
 
 It is among the higher ranunculaceous plants, how- 
 ever, that we get the fullest and richest colouration. 
 Columbines [Aquilegia) are very specialised forms of 
 the buttercup type (Fig. lo). Both sepals and petals 
 
 Fig. io.— Flower of columbine iAquiUgia vulgaris), with petals produced uito 
 honey-bearing spurs ; purple or blue. * — " 
 
 are brightly coloured, while the former organs are pro- 
 duced above into long, bow-shaped spurs (i^ig. 1 1), each 
 of which se :retes a drop of honey. The carpels are also 
 
 D 2 
 
36 THE COLOURS OF FLOWERS. 
 
 reduced to five, the regularity of number being itself 
 a common mark of advance in organization. Various 
 columbines accordingly range from red to purple and 
 dark blue. Our English species, A. vulgaris, is blue 
 or dull purple, though it readily reverts to white or 
 red in cultivated varieties. Even the columbine, 
 however, though so highly specialised, is not bila- 
 terally but circularly symmetrical. This last and 
 highest mode of adaptation to insect visits is found in 
 larkspur {Delphinium ajacis), and still more developed 
 in the curious monkshood [Aconitiun napellus, Fig. 12). 
 
 Fig. II. — Petal of columbine produced into a honey-bearing spur. 
 
 Now larkspur is usually blue, though white or red 
 blossoms sometimes occur by reversion ; while monks- 
 hood is one of the deepest blue flowers we possess. 
 Both show very high marks of special adaptation ; 
 for besides their bilateral form, Delphinium has the 
 number of carpels reduced to one, the calyx coloured 
 and deeply spurred, and three of the petals abortive ; 
 while Aconitum has the carpels reduced to three and 
 partially united into a compound ovary, the upper 
 sepals altered into a curious coloured hood or helmet, 
 and the petals considerably modified (Fig. 13). All 
 these very complex arrangements are defintely corre- 
 
LAW OF PROG RE SSI VE COL O URA TION. 37 
 
 lated with the visits of insects, for the two highly 
 abnormal petals under the helmet of the monkshood 
 produce honey, as do also the two long petals within the 
 spur of the larkspur. Both flowers are also specially 
 adapted to the very highest class of insect visitors. 
 
 Fig. 12. — Flowers of monkshood {Aconitum «rt/^//«i), dark blue. 
 
 Aconitiim is chiefly fertilized by bees ; and Sir John 
 Lubbock observes that '^ Anthophora pilipes and Born- 
 bus hortorum are the only two North European 
 insects which have a proboscis long enough to reach 
 to the end of the spur of Delphiniiun elatum. A, 
 
38 
 
 THE COLOURS OF FLOWERS. 
 
 pilipes, however, is a spring insect, and has already 
 disappeared before the Delphinium comes into flower, 
 so that it appears to depend for its fertilisation 
 entirely on Bombus hortoriimy 
 
 Thus within the limits of the Ranunculacece we get 
 every gradation in colouring, from the very simple, 
 open, yellow buttercups, through the white water 
 
 Fig. 13- Petals of monkshood, concealed by sepals, and produced into honey- 
 bearing sacs. 
 
 crowfoots, the red adonis, the scarlet paeony, and the 
 purple columbine, to the very irregular blue larkspur, 
 and the extremely complex ultramarine monkshood. 
 In this family it may be noted, too, that increase of 
 adaptation to insect visits is shown rather by peculi- 
 arities of shape and arrangement than by mere 
 increased size of petals, as among the roses. 
 
 Observe also that every advance either in insect 
 
LAW OF PROGRESSIVE COLOURATION. 39 
 
 fertilization or in special adaptation for dispersal of 
 seeds results in a lessening of the number of carpels 
 or of seeds. The plant does not need to produce so 
 many when all are fairly sure of arriving at maturity 
 and being dispersed. 
 
 Flowers in which the carpels have arranged them- 
 selves in a circle around a common axis, like the 
 Geraniacece and Malvacece, thereby show themselves 
 to be more highly modified than flowers in which all 
 the carpels are quite separate and scattered, like the 
 simpler Rosacecs and Rammculacece. Still more do 
 families such as the Caryophyllacecs^ or pinks, in which 
 all the five primitive carpels have completely coalesced 
 into a single five-celled ovary. Accordingly, it is not 
 remarkable that the pinks should never be yellow. 
 On the other hand, this family has no very specialised 
 members, like the larkspur and the monkshood, and 
 therefore it very rarely produces bluish or purplish 
 flowers. Pinks, in fact, do not display so wide a 
 range in either direction as Ranunculace<s. They 
 begin as high up as white, and hardly get any higher 
 than red or carnation. They are divisible into two 
 sub-families, Alsinece and Silenece. The Alsinecs have 
 the sepals free, the blossoms widely expanded, and no 
 special adaptations for insect fertilization (Fig. 14). 
 They include all the small undeveloped field species, 
 such as the chickweeds [Stellaria media, Arenaria trin- 
 erviSf Cerastiiim vu/gattun, &c.), stitchworts {Stellaria 
 holostea, &c.), and cornspurries {Spergula arvensis), 
 which have open flowers of a very primitive character ; 
 and almost all of them are white. These are ferti- 
 lized by miscellaneous small flies. The Silenece, on the 
 other hand, including the large- flowered types such as 
 
40 
 
 THE COLOURS OF FLOWERS. 
 
 the campions (Fig. 15) and true pinks, have a tubular 
 calyx, formed by the coalescence of the five sepals, 
 and the expanded petals are raised on long claws ; 
 which makes their honey, inclosed in the tube, acces- 
 sible only to the higher insects. Most of them also 
 display special adaptations for a better class of insect 
 fertilization in the way of fringes or crowns on the 
 petals. These more profoundly modified kinds are 
 
 Fig. 14.— Flowers of chickweed {StcUaria) ; white. 
 
 generally pink or red. For example, in the most 
 advanced British genus, Dianthus, which has usually 
 vandyked edges to the petals, our four English species 
 are all brightly coloured, D. armeria, the Deptford 
 pink, being red with dark spots, D. prolifer purplish 
 red, D. deltoides, the maiden pink, rosy spotted with 
 white, and D. cmius, the Cheddar pink, bright rose- 
 coloured. It is much the same with the allied genus 
 Lychnis. 
 
LA W OF PROGRESSIVE COLO URA TION. 4 r 
 
 Our own beautiful purple English corn-cockle 
 (Z. githagd), is a highly developed campion, so 
 specialised that only butterflies can reach its honey 
 with their long tongues, as the nectaries are situated 
 at the bottom of the tube. Two other species of 
 campion, however, show us interestingly the way 
 in which variations of colour may occur in a retro- 
 grade direction even among highly evolved forms. 
 
 Fig. 15.— Flower of night campion (^Lychnis vespertina) ; white. 
 
 One of them, the day lychnis, {L. diiirna), has red, 
 scentless flowers, opening in the morning, and it is 
 chiefly fertilized by diurnal butterflies. But its de- 
 scendant, the night lychnis (Z. vespertina), has taken 
 to fertilization by means of moths ; and as moths can 
 only see white flowers it has become white (Fig. 1 5), and 
 has acquired a faint perfume as an extra attraction. 
 Still, the change has not yet become fully organised 
 
42 THE COLOURS OF FLOWERS. 
 
 in the species, for one may often find a night lychnis 
 at the present time which is only pale pink, instead of 
 being pure white. 
 
 Sir John Lubbock remarks of the CaryophyllacecB 
 that " the order presents us with an interesting series 
 commencing with open-flowered species, the honey of 
 which is accessible even to beetles, and short-tongued 
 flies, through those which are adapted to certain flies 
 (Rhingia) and Bees ; to the species of Dianthus, 
 Saponariay and Lychnis Githago, the honey of which 
 is accessible to Lepidoptera only." It is a curious 
 fact that in just the same progression the flowers pass 
 gradually from small white inconspicuous petals to 
 large and deeply coloured red or purple ones. 
 
 The Cistacece are another family of simple flowers, 
 with the carpels united, but otherwise very primitive 
 in form. Their petals usually spread around the ovary 
 in a regular discoid form ; and the earliest members 
 only differ essentially from Potentilla in the union of 
 their carpels into a single imperfectly three-celled 
 capsule. Our English genus, Helianthemuni or rock- 
 rose, comprises some of the smallest and simplest 
 forms, with yellow petals, and very like Potentilla in 
 general appearance. One species, however, H.poli- 
 folium (a mere slight variation on the yellow H. 
 vtilgare), has white flowers. The larger South Euro- 
 pean forms, which make up the genus Cistus, have 
 much more expanded petals, and these are usually 
 white, pinkish, or rose-coloured. One Mediterranean 
 species has a yellow centre with white edges : another 
 closely allied to it, has a white centre with pink edges. 
 Here, as in the roses, mere increase in size (coupled 
 of course with special insect selection) seems to have 
 
LA IV OF PROGRESSIVE COLOUR A TION. 43 
 
 effected the progressive modification of the pig- 
 mentary matter. 
 
 The PapaveracecB or poppies show us somewhat 
 similar results. Many of the less developed forms, 
 with small petals, are yellow. For example, the 
 celandine, Chelidonitun majus, has a very simple ovary 
 and comparatively inconspicuous flowers : and its 
 petals are of just the same primitive golden yellow as 
 those of the buttercups, the potentillas, and the rock- 
 roses. Glaticium luteum, another little-developed 
 form, is also yellow. So are the Californian Esch- 
 scholtzias of our gardens. But in the field poppies, 
 Papaver rkceas, P. dubium, P. hybridum, &c., which 
 have extremely large and expanded petals, together 
 with a highly modified compound ovary, bright scarlet 
 is the prevailing colour, though pale red and white 
 also occur. The still larger garden poppy, P. somni- 
 ferumy is bluish white, with a purple base to the 
 petals. 
 
 The CriicifercE are a family which display a good 
 deal of variety of colouration, though not so largely 
 within the limits of our British species. The most 
 primitive and simple forms have yellow flowers, as in 
 the case of the cabbage genus {Brassicd)'mQ\M^\n^ char- 
 lock, mustard, and turnip ; the rockets {Barbarea and 
 Sisymbrium) ; and the gold-of-pleasure {Camelina 
 sativd). Most of these are dry-field weeds, and they 
 have open little-developed blossoms. Their petals 
 usually fade white. In the genus Nasturtium or 
 watercress we have four species, three of which are 
 yellow, while one is white. In treacle-mustard {Eiy- 
 simum), the yellow is very pale, and the petals often 
 become almost white. Just above these earliest forms 
 
44 THE COLOURS OF FLOWERS, 
 
 come the common small white crucifers like Cardamhie 
 hirsuta, Cochlearia officinalis^ and Capsella bursa- 
 pastoris. Many of these are little if at all superior in 
 organization to the yellow species, and some of them (as 
 we shall see hereafter) are evidently degenerate weeds 
 of cultivation. But such flowers as Alyssum mariti- 
 mum^ with its sweet scent, its abundant honey, its 
 reduced number of seeds, and its conspicuous, 
 spreading milk-white petals, are certainly more 
 developed than small yellow species like Alysstim daly- 
 cimim. Even more remarkably is this the case in the 
 genus Iberis or candytuft, which has become slightly 
 irregular, by the two adjoining exterior petals growing 
 larger than the interior ones. They thus form very 
 conspicuous heads of bloom, obviously adapted to 
 higher insect fertilisation. Accordingly, they are 
 usually white, like our British species, /. aniara ; while 
 some of the larger exotic species are a pretty pink in 
 hue. The genus Cardamine supplies us with like 
 instances. Here the smaller species have white 
 flowers, and so has the large C. amara. But in 
 C. pratensis, the cuckoo-flower, they are usually tinged 
 with a pinkish purple, which often fades deep mauve ; 
 and in some showy exotic species the flowers are a 
 rich pink. So vj'xth. Arabis : our small English kinds 
 are white; A. petrcea, with larger petals, is often 
 slightly purplish, and some handsome exotics are a 
 vivid purple. In Hesperis we get a further degree of 
 modification in that the petals are raised on rather 
 long claws ; and the flowers (represented in England 
 by H. matronalis^ the dame's-violet) are a fine purple 
 and possess a powerful perfume. Closely allied is 
 the Virginia stock of our gardens, {Malcolmia) which 
 
LA W OF PROGRESSIVE COLOUR A TION. 45 
 
 varies from pale pink to mauve : its calyx has become 
 tubular. Yellow blossoms occasionally occur in this 
 genus. But the highest of all our crucifers are con- 
 tained in the genera Matthiola and Cheiranthus^ which 
 have large spreading petals on long erect claws, 
 besides often being sweet scented. The common 
 stock {M. incana) is purple, reddish, or even violet ; 
 our other British species, M. sinuata, is pale lilac ; and 
 no member of the genus is ever yellow. The wall- 
 flower {Cheiranthus cheirt) is rich orange or red, some- 
 times yellow : its colour, however, differs widely from 
 the primitive golden yellow of the charlocks or 
 buttercups ; and it will receive further attention here- 
 after. 
 
 There is one special (perhaps artificial) tribe of 
 crucifers, the LomentoscB^ which display specially high 
 modification in the pod or fruit ; and these deserve 
 separate treatment. Yellow flowers are here very 
 rare ; but one English species, Isatis tinctorial the 
 dyer's woad, has small yellow petals. Raphanus 
 raphanistrum, the wild radish, has usually in its sea- 
 coast form pale primrose blossoms, much larger than 
 woad ; but inland they are oftener white with coloured 
 veins, and sometimes even lilac. Crambe maritima, 
 the seakale, a somewhat more developed type, is 
 always white, never yellow ; and Cakile maritima, a 
 still higher plant of the same tribe, has purple 
 blossoms, much like those of a stock. 
 
 So much by way of illustration of the families with 
 usually regular polypetalous flowers and free superior 
 ovaries. The other families of this type not noticed 
 here will receive attention in a later chapter. We 
 may next pass on to the families of polypetalous 
 
46 THE COLOURS OF FLOWERS. 
 
 flowers with usually irregular corollas, which represent 
 of course a higher stage of development in adaptation 
 to insect visits. 
 
 Of these, three good illustrative cases are included 
 in the British flora. They are the Polygalacece, the 
 Violacece, and the PapilionacecB. 
 
 Polygala vulgaris^ or milkwort, our only British 
 representative of the first named family, is an ex- 
 tremely irregular flower, very minutely and remark- 
 ably modified for special insect fertilisation. It is 
 usually a bright blue in colour, but it often reverts to 
 pink, and not infrequently cv^en to white. Two of 
 the sepals are coloured, as well as the petals. 
 
 The ViolacecB or violets are a whole family of 
 bilateral flowers, highly adapted to fertilisation by 
 insects ; and as a rule they are a deep blue in colour. 
 This is the case with four of our British species, Viola 
 odorata, V. canmay V. hirta, and V. palustris. Here, 
 too, however, white varieties easily arise by reversion ; 
 while one member of the group, the common pansy, 
 V. tricolor, is perhaps the most variable flower in all 
 nature. This case, again, will receive further attention 
 when we come to consider the subject of variegation 
 and of reversion or retrogression. 
 
 The P apilionacece or peaflowers are closely related 
 to the roses, but differ conspicuously in their bilateral 
 form. The lower and smaller species, such as the 
 medick, lotus, and lady's fingers, are usually yellow, 
 though often reddish outside. So also are broom 
 and gorse. Among the more specialised clovers, 
 some of which are fertilised by bees alone, white, 
 red, and purple predominate. Even with the smaller 
 and earlier types, the most developed species, like 
 
L/^JV OF PROGRESSIVE COLOURATION. 47 
 
 lucerne, are likewise purple or blue. But in the 
 largest and most advanced types, the peas, beans, 
 vetches, and scarlet runners, we get much brighter and 
 deeper colours, often with more or less tinge of blue. 
 In the sweet-peas and many others, the standard 
 frequently differs in hue from the keel or the wings — 
 a still further advance in heterogeneity of colouration. 
 Lupines, sainfoin, everlasting pea, and wistaria are 
 highly evolved members of the same family, in which 
 purple, lilac, mauve, or blue tints become distinctly 
 pronounced. The colouration of the Papilionace^, 
 however, does not as a whole illustrate the general law 
 so well as that of many other groups. 
 
 When we pass on to the Corolliflorce, or flowers in 
 which the originally separate petals have coalesced 
 into a single united tube, we meet with much more 
 striking results. 
 
 Here, where the very shape at once betokens high 
 modification, yellow is a comparatively rare colour 
 (especially as a ground-tone, though it often comes out 
 in spots or patches), while purple and blue, so rare 
 elsewhere, become almost the rule. 
 
 The family of CampanulacecB^ or bluebells, forms an 
 excellent example. Its flowers are usually blue or 
 white, and the greater number of them, like the hare- 
 bell (Campanula rotimdifolia) and the Canterbury bell 
 (C. media),^.rQ deep blue (Fig. 16). We have nine British 
 species of the genus, varying from pale sky-blue to ultra- 
 marine and purplish cobalt, with an occasional relapse 
 to white. Rampion and sheep's bit, also blue, arc 
 clustered heads of similar blossoms. The little blue 
 lobelia of our borders, which is bilateral as well as 
 tubular, belongs to a closely-related tribe. One of 
 
48 
 
 THE COLOURS OF FLOWERS. 
 
 our British species, Lobelia Dortmanna, is sky-blue ; 
 the other, L. urens, is a dingy purple. Not far from 
 them are the Dipsacecs, including the lilac scabious, 
 the blue devil's bit, and the mauve teasel. Amongst 
 
 Fig. i6,— Flower of harebell (Campanula); bright blue. 
 
 all these very highly-evolved groups blue distinctly 
 forms the prevalent colour. 
 
 In the great family of the Ericacecu, or heaths, which 
 is highly adapted to insect fertilisation, more particu- 
 larly by bees, purple and rose are the prevailing tints, 
 
LAW OF PROGRESSIVE COLOURATION. 49 
 
 so much so that, as we all have noticed a hundred 
 times over, they often colour whole tracts of hillside to- 
 gether. In all probability there are no really yellow 
 heaths. The bell-shaped blossoms mark at once the 
 position of the heaths with reference to insects ; and 
 the order, according to Mr. Bentham, supplies us 
 with more ornamental plants than any other in the 
 whole world. Among our British species, in the less 
 developed forms, like Vaccininm^ Arbutus, and Andro- 
 meda^ the flowers are usually white, flesh-coloured, 
 pinkish, or reddish. The highly developed Ericce, on 
 the other hand, are mostly purple or deep red. E. 
 vulgaris has the calyx as well as the corolla coloured 
 with a mauve variety of pink. Menziesia ccsrulea is a 
 deep purplish blue. Monotropa alone, a very degraded 
 leafless saprophyte form, has greenish-yellow or pale 
 brown free petals. 
 
 The BoraginacecB are another very advanced family 
 of CorolliflorcB, and they are blue almost without 
 exception. They have usually highly - modified 
 flowers, with a tube below and spreading lobes above ; 
 in addition to which most of the species possess 
 remarkable and strongly-developed appendages to the 
 corolla, in the way of teeth, crowns, hairs, scales, 
 parapets or valves. Of the common British species 
 alone, the forget-me-nots {Myosotis) are clear sky- 
 blue with a yellow eye ; the viper's bugloss {Echium 
 vulgare) is at first reddish purple, and afterwards a 
 deep blue ; the lungwort {Pulmonaria officinalis) is 
 also dark blue ; and so are the two alkanets {Anchusa)^ 
 the true bugloss {Lycopsis)^ the mad wort {Asperugo), 
 and the familiar borage {Borago officinalis), used to 
 flavour claret-cup ; though all of them by reversion 
 
 £ 
 
so Tf{E COLOURS OF FLOWERS. 
 
 occasion ally prod uce purple or white flowers. H ounds- 
 tongue {Cynoglossiim officinale) \<, purple-red, and most 
 of the other species vary between purple and blue ; 
 indeed, throughout the family most flowers are red at 
 first and blue as they mature. The purplish-red of 
 the less advanced species exactly answers to the 
 immature colouring of the more advanced, which are 
 blue in their full-blown stage. Of these, borage at 
 least is habitually fertilised, by bees, and the same is 
 partially true of many of the other species. All of 
 them are adapted to a high class of insect visitors. 
 
 Other families of regular Corolliflorcs must be 
 glanced at more briefly. Among the Gentianacecey 
 the less advanced types, like the simple Chlora per- 
 foliata and Lininantheinum nyiup/iceoides^ are yellow, 
 perhaps by reversion ; but Menyanthes trifoliata, a 
 slightly more developed ally of Limnanthenmm^ has 
 white blossoms, tinged outside with red ; ErytJircsa 
 centauriuviy with a divided calyx and the cells of the 
 ovary imperfectly united, is red ; and the true gentians, 
 Gejitimia verjia, G. campestris, G. 7iivalis, Sic, with a 
 tubular calyx, long throat, and sometimes fringed 
 hairs to the tube, are bright blue. In Apocynacecs, we 
 have the highly developed periwinkles, Vinca major 
 and V. minor, normally blue, though pink and white 
 varieties or species are also cultivated. In Plumbaginece 
 we have the bluish purple sea-lavender {Statice Li- 
 moniiini) and the pink thrift [Arineria vulgaris). Other 
 families with special peculiarities will receive notice 
 later on. 
 
 It is necessary, however, here briefly to refer to the 
 great family of Compositce, some of whose peculiarities 
 can only properly be considered when we come to 
 
LAW OF PROGRESSIVE COLOURATION. 51 
 
 inquire into the phenomena of relapse and retrogres- 
 sion. Nevertheless, even at the present stage of our 
 inquiry, the composites afford some excellent evidence. 
 In certain ways they may be regarded as the very 
 highest race of flowering plants. Not only are their 
 petals united into a tubular corolla, but their blossoms 
 are compounded into large bunches or groups of a 
 very showy and attractive sort. Each flower-head 
 here consists of a number of small florets, crowded 
 together so as to resemble a single blossom. So far 
 as our present purpose is concerned, they fall natur- 
 ally into three groups — Jussieu's old-fashioned sub- 
 orders of Ligidatcs, CynaroidecB, and Cotyinbiferce, 
 which are quite sufficient for all ordinary objects of 
 botanical study. The first, or ligulate tribe, is that 
 of the dandelions or hawkweeds, with open florets, 
 fertilised, as a rule, by very small insects ; and these 
 are generally yellow, with only a very few divergent 
 species. They will receive further notice hereafter. 
 The second, or cynaroid tribe, is that of the thistle- 
 heads, visited by an immense number of insects, in- 
 cluding the bees ; and these are almost all purple, 
 while some highly-evolved species, like the corn-flower 
 or blue-bottle {Centaurea cyamis, Fig. 17) and the true 
 artichoke {Cynara scolynius), are bright blue. The 
 third, or corymbiferous tribe, is that of the daisies and 
 asters, with tubular central florets and long flattened 
 outer rays ; and these demand a closer examination 
 here. 
 
 The central florets of the daisy tribe, as a rule, are 
 bright golden ; a fact which shows pretty certainly 
 that they are descended from a common ancestor 
 who was also yellow. Moreover, these yellow florets 
 
 E 2 
 
52 
 
 THE COLOURS OF FLOWERS. 
 
 are bell-shaped, and each contains a pistil and fi\'e 
 stanaens, like any other perfect flower. But the outer 
 florets are generally sterile ; and instead of beino- 
 
 Mi(/^:,. 
 
 Fig. 17.— Flower of corn bluebottle {s:entaurea cyanus); highest type of cynaroid 
 
 composite ; bright blue. 
 
 bell-shaped they are split down one side and unrolled, 
 so as to form a long ray ; while their corolla is at 
 the same time much larger than that of the central 
 
LAW OF PROGRESSIVE COLOURATION. 53 
 
 blossoms (Fig. 18). In short, they are sterilised mem- 
 bers of the compound flower-head, specially set apart 
 for the work of display ; and thus they stand to the 
 entire flower-head in the same relation as petals do to 
 the simple original flower. The analogy between the 
 two is complete. Just as the petal is a specialised 
 and sterilised stamen told off to do duty as an 
 allurer of insects for the benefit of the whole flower, 
 so the ray-floret is a specialised and sterilised blos- 
 som told off to do the self-same duty for the benefit 
 
 Fig. iS.— Vertical section cf head of daisy {Bellis perenms); central florets, yellow ; 
 
 ray florets, white, tipped with pink. 
 
 of the group of tiny flowers which make up the 
 composite flower-head. 
 
 Now, the earliest ray-florets would naturally be 
 bright-yellow, like the tubular blossoms of the central 
 disk from which they sprang. And to this day the 
 ray-florets of the simplest corymbifcrous types, such 
 as the corn-marigold {ChrysantJieimun segetum), the 
 sun-flower {HcliantJms annuus), and the ragwort 
 {Senecio jacoba^d)y are yellow like the central flowers. 
 In the camomile, however, the ox-eye daisy, and the 
 may-weed {Antheinis cotula, Chrysanthemum leucan- 
 t/iemtim, &c.), the rays have become white ; and this, 
 
54 THE COLOURS OF FLOWERS. 
 
 I think, fairly establishes the fact that white is a 
 higher development of colour than yellow; for the 
 change must surely have been made in order to 
 attract special insects. Certainly, such a differentia- 
 tion of the flowers in a single head cannot be without 
 a good purpose. In the true daisy, again {Bdlis 
 perennis), the white rays become tipped with pink (Fig. 
 1 8) which sometimes rises almost to rose-colour ; and 
 this stage is exactly analogous to that of apple- 
 blossom, which similarly halts on the way from white 
 petals to red. In our own asters {^A. tripolium^ &c.) 
 and the Michaelmas daisies of America, we get a 
 further advance to purple, lilac, and mauve, while 
 both in these and in the chrysanthemums true shades 
 of blue not infrequently appear. The Cinerarias of 
 our gardeners are similar forms of highly-developed 
 groundsels from the Canary Islands, /r 
 
 Tubular flowers with an irregular corolla are ob- 
 viously higher in their mode of adaptation to insect 
 visits than tubular flowers of the ordinary symmetrical 
 type. Amongst them, the first place must be as- 
 signed to the Labiates — perhaps the most specialised 
 of any so far as regards insect fertilisation. Not 
 only are they deeply tubular, but they are very 
 bilateral and irregular indeed, displaying more modi- 
 fication of form than almost any other flowers except 
 the orchids. They mostly secrete abundant honey, 
 and often possess highly aromatic perfumes. Mere- 
 over, they form geologically one of the latest families 
 of flowering plants, specially developed in adaptation 
 to bees and other highly-evolved honey-suckers. 
 Almost all of them are purple or blue. Among the 
 best known English species are thyme, mint, marjoram. 
 
LAW OF P/WGRESSIVE COLOURATION, 
 
 55 
 
 sage (Fig. 19), and basil, which it need hardly be said 
 are great favourites with bees. Ground-ivy [Nepeta <^le- 
 chomd) is bright blue ; catmint {Nepeta cataria), pale 
 blue ; Prunella, violet-purple ; and common bugle 
 {Ajtiga reptans), blue or flesh-colour. Many of the 
 others are purple or purplish. It must be added that 
 m this family the flowers are very liable to vary 
 within the limit of the same species ; and red. white, 
 or purple specimens are not uncommon in many of 
 the normally blue kinds. 
 
 Fig. 19 — Flower of sage {Salvia) visited by b^e ; blue. 
 
 The Scrophularmecs^n^ other allied irregular tubular 
 families are mostly spotted, and so belong to a later 
 stage of our inquiry ; but even amongst this group, 
 the Veronica genus has almost always pure blue 
 flowers ; foxglove {Digitalis purpurea) is purple; and 
 most of the Kroomrapes {Orobanchacece) are more or 
 less bluish. Blue and lilac also appear abundantly 
 in spots or stripes in many species of Linaria., in 
 Euphrasia^ and in other genera. 
 
 We have given so much consideration to the Dico- 
 tyledons that the relatively simple and homogeneous 
 
56 THE COLOURS OF FLOWERS. 
 
 Monocotyledons need not detain us long. Their 
 colouration is as a whole both less complicated and 
 less instructive. As a rule, sepals and petals are here 
 petaloid and often indistinguishable. 
 
 The Alismacece answer very closely to the Ranun- 
 culacece 2l?, being in all probability the earliest sur- 
 viving type of entomophilous Monocotyledons. Their 
 arrangement is of course trinary, but they have 
 similarly separate carpels, often numerous, surrounded 
 by one, two, three, or many rows of stamens, and 
 then by one row of three petals and one row of three 
 sepals. All our English species, however, are white 
 or rosy, instead of yellow. As they are marsh plants, 
 they seem to have reached or passed the stage of 
 Raminailiis aqiiatilis. One species, Alisma plantago, 
 the water-plantain, however, still retains a yellow 
 claw to the petals, though the limb is white or pale 
 pink. So also does Damasoniiim stellaium. These 
 two interesting plants present a remarkable analogy 
 to the water-crowfoot. 
 
 Among monocotyledonous families with a united 
 ovary, the Liliacecs are probably the most primitive. 
 Their simplest type in England is Gagealiitea (Fig. 20), 
 a yellow lily looking extremely like a bunch of Ranun- 
 culus Ficaria. In L^loydia serotina, a closely allied 
 but more developed form, the petals are white, with 
 a yellow base, and three reddish lines. The wild 
 tulip is likewise yellow. Allium ursinum, a some- 
 what higher type, is pure white. The fritillary 
 {Fritillaria Meleagris), a large, handsome, bell-shaped 
 flower, with separate petals converging into a campanu- 
 late form, and with a nectariferous cavity at their base 
 (Fig. 21;, is purple or red, chequered with lurid marks ; 
 
LAW OF PROGRESSIVE COLOURATION. 
 
 57 
 
 but it often reverts to white or even to a faint yellow. 
 In Scilla, however, including our common wild hya- 
 cinth {S. nutans), the deep tubular flowers, composed 
 of perianth pieces with long claws, are usually blue, 
 rarely pink or white ; while in Hyacinthus and Mus- 
 cari, which have a united bell-shaped or globular 
 blossom, formed by the coalescence of the sepals and 
 petals, dark-blue and ultramarine are the prevalent 
 
 KiG. 2o.-Flowers of si„,plest typical lily iGnge.^ lut.a); primitive yellow. 
 
 tones. Meadow saffron {Colchicuvi autumnale), which 
 has also a united tube and very deep underground 
 ovary, is a fine reddish purple : its stamens secrete 
 honey. 
 
 The IridecE and Ainaryllide(B are more advanced 
 than the lilies, in that they possess inferior ovaries— 
 in other words, their perianth tube has coalesced with 
 the walls of the inclosed carpels. In many cases 
 
58 
 
 THE COLOURS OF FLOWERS. 
 
 especially in the more highly-developed species, their 
 flowers are red, blue, or purple. TricJionema Buibo- 
 codium is purplish-blue with a yellow centre. Our 
 two native crocuses {C. vermis and C. niidijiorus) are 
 also purple. SisyrincJiium Berumdianum is a delicate 
 blue. Gladiolus comiminis is brilliant crimson. Iris 
 fcetidissiina (Fig. 22) is violet. Our own Amaryllids 
 are white or primrose, but brilliant reds and purples, 
 
 Fir.. 21. — FI )v,cr of fritillary {Fritiliaria Mcleag^rs); \n\r\Ac, spotted with whitt;. 
 
 as well as highly-developed spotted types, are common 
 amongst the cultivated exotics. 
 
 The Orchidacece stand at the head of the cntomo- 
 philous Monocotyledons by virtue of their inferior 
 ovary, their irregular flowers, and their extraordinary 
 adaptations to insect fertilization. Purples are the 
 prevailing ground-tones ; but in the commonness of 
 variegation and of specialized lines or spots of colour 
 the Orchids answer closely to the ScrophularineiC 
 
LA W OF PROGRESSIVE COLOURATION. 
 
 59 
 
 among Dicotyledons, and may therefore best be con- 
 sidered in the succeeding chapter. 
 
 This brief review of the chief famih'es of Engh'sh 
 cntomophilous flowers will probably have made clear 
 the general truth of the law of progressive coloura- 
 tion here laid down. There are many exceptions and 
 special peculiarities, some of which will be explained 
 
 I'll; 22.— Flower of common flag (Irisfa-tidisshiUi) ; vi.,lft i)!iie. 
 
 or accounted for in the sequel ; but on the whole we 
 
 may consider the following facts fairly proved : 
 
 (I.) Most of the very simplest flowers arc yellow. 
 (2.) Many of the simpler flowers in each family 
 
 (except the highest) are apt to be yellow. 
 (3.; The more advanced members of very simple 
 families are usually white or pink. 
 
6o THE COLOURS OF FLOWERS. 
 
 (4.) The simpler members of slightly advanced 
 
 families are usually white or pink. 
 (5.) The most advanced members of all families 
 
 are usually red, purple, or blue. 
 (6.) Almost all the members of the most advanced 
 
 families are purple or blue. 
 (7.) The most advanced members of the most 
 
 advanced families are almost always blue, 
 
 unless spotted or variegated. 
 
CHAPTER III. 
 
 VARIEGATION. 
 
 So far we have spoken for the most part only of 
 what ladies would call self-colour, as though every 
 flower were of one unvaried hue throughout. We 
 must now add a few considerations on the subject of 
 the spots and lines which so often variegate the 
 petals in certain species. In this connection, again, 
 Mr. Wallace's hint is full of meaning. Everywhere 
 in nature, he points out, spots and eyes of colour 
 appear on the most highly-modified parts, and this 
 rule applies most noticeably to the case of petals. 
 Simple regular flowers, like the buttercups and roses, 
 hardly ever have any spots or lines ; but in very 
 modified forms like the labiates and the orchids they 
 are extremely common. The scrophularineous family, 
 to which the snap-dragon belongs, is one most specially 
 adapted to insects, and even more irregular than that 
 of the labiates ; and here we find the most singular 
 cfl"ects produced by dappling and mixture of colours. 
 The simple yellow mullein, it is true, has few such 
 spots or lines, nor have even many of the much higher 
 blue veronicas ; but in the snapdragons, the foxglove, 
 the toadflax, the ivy-linaria, the eyebright, and the cal- 
 
62 THE COLOURS OF FLOWERS. 
 
 ceolarias, the intimate mixture of colours is very notice- 
 able. In the allied tropical Bigno7iias and Gloxinias 
 we see much the same distribution of hues. Many 
 of the family are cultivated in gardens on account of 
 their bizarre and fantastic shapes and colours. As to 
 the orchids, it is hardly necessary to say anything 
 about their wonderfully spotted and variegated flowers. 
 Even in our small English kinds the dappling is 
 extremely marked, especially upon the expanded and 
 profoundly modified lower lip ; but in the larger 
 tropical varieties the patterns are often quaint and 
 even startling in their extraordinary richness of fancy 
 and apparent capriciousness of design. Mr. Darwin 
 has shown that their adaptations to insects are more 
 intimate and more marvellous than those of any other 
 flowers whatsoever. 
 
 Structurally speaking, the spots and lines on petals 
 seem to be the direct result of high modification ; but 
 functionally, as Sprengel long ago pointed out, they 
 act as honey-guides, and for this purpose they have 
 no doubt undergone special selection by the proper 
 insects. The case is jujt analogous to that of the 
 peacock's plumes or the wings of butterflies. Tn 
 either instance, the spots and eye-marks tend to 
 appear on the most highly-modified surfaces ; but 
 they are perpetuated and intensified by special selec- 
 tive action. Among birds and insects, sexual selection 
 performs the work of fixing the colours ; among 
 flowers, the visits of bees and butterflies effect the 
 same end. Lines are comparatively rare on regular 
 flowers, but they tend to appear as soon as the flower 
 becomes even slightly bilateral, and they point directly 
 towards the nectaries. Hence they cannot be mere 
 
VARIEGATION. 63 
 
 purposeless products of special modification ; they 
 clearly subserve a function in the economy of the 
 plant, and that function is the direction of the 
 insect towards the proper place for effecting the 
 fertilization of the ovary. In the common rhodo- 
 dendron, the connection can be most readily observed 
 with the naked eye, and the honey tested by the 
 tongue. In this case, one lobe of the corolla secretes 
 a very large drop of nectar in a fold near its base, 
 and the lines of dark spots appear on this lobe alone, 
 pointing directly towards the nectariferous surface. 
 
 The GcraniacecB afford an excellent illustration of 
 the general principle. They are on the whole a com- 
 paratively high family of polypetals, for their ovary 
 tends to become compound and very complicated, 
 and they have many advanced devices for the disper- 
 sion of their seeds. Oxalis cornicidata, our simplest 
 English form, is pale yellow : O. acetosclla is white, 
 with a yellow base, and its vi ;s are delicately tinged 
 with lilac. The flowers of Ei odium and Geranhnn, 
 which arc much more advanced, are generally pink 
 or purplish, often marked with paler or darker lines. 
 For the most part, however, these regular forms are 
 fairly uniform in hue ; but many of the South African 
 Felargoniiuns^ cultivated in gardens and hot-houses, 
 are slightly bilateral, the two upper petals standing 
 off from the three lower ones ; and these two become 
 at once marked with dark lines, which are in some 
 cases scarcely visible, and in others fairly pronounced. 
 From this simple beginning one can trace a gradual 
 progress in heterogeneity of colouring, till at last the 
 most developed bilateral forms have the two upper 
 petals of quite a different hue from the three lower 
 
64 THE COLOURS OF FLOWERS. 
 
 ones, besides being deeply marked with belts and spots 
 of dappled colour. In the allied Tropcsolum (Fig. 23) or 
 Indian cress (the so-called nasturtium of old-fashioned 
 gardens — though the plant is really no more related 
 to the water-cress and other true nasturtiums than we 
 ourselves are to the great kangaroo) this tendency is 
 carried still further. Here, the calyx is prolonged 
 into a deep spur, containing the honey, inaccessible to 
 any but a few large insects ; and towards this spur all 
 the lines on the petals converge. Sir John Lubbock 
 
 Tig. 23. — Flower of Indian cress {Tropceoluttt waJMs). with one sepal prolonged into 
 a honey-bearing spur ; orange and yellow. 
 
 observes that without such conventional marks to 
 guide them, bees would waste a great deal of time in 
 bungling about the mouths of flowers ; for they are 
 helpless, blundering things at an emergency, and 
 never know their way twice to the same place if any 
 change has been made in the disposition of the 
 familiar surroundings. 
 
 The readiness with which the Geraniacece pass into 
 irregular forms in Pelargoniiim and in the balsam 
 genus {Impatiens), in itself shows that they are a fairly 
 advanced family, and explains the common appear- 
 
VARIEGATION. 
 
 65 
 
 ance of pink and purple petals among flowers which 
 at a casual glance seem scarcely so much modified as 
 the pinks or the crucifers. 
 
 The Malvacece are another family in which lines 
 and stripes commonly occur ; and they are shown to 
 be of a comparatively advanced type by their pecu- 
 liarly modified pistil, and by the union of their 
 stamens into a tube, to which the petals adhere at the 
 base. Lavatera arborea is purplish red \ of our three 
 
 ■^^^^HfM ^i^. 
 
 Fig. 24— Flower of •^\x^{Pianthus)\ pink, with lighter spots. 
 
 Malvce, one, M. rotiindifolia^ is pale blue ; M, syl- 
 vestris is purple ; and M. moschata, rose-coloured, or, 
 rarely, white. All have strongly-marked lines, and, 
 in addition to this, the vivid green calyx, seen through 
 the interstices of the petals, acts apparently as a 
 supplementary honey-guide. The same peculiarities 
 distinguish the genus Alt/um, of which we have two 
 species, A. officinalis, pale rose-colour, and A. Jiirsuta, 
 purplish blue. 
 
66 THE COLOURS OF FLOWERS 
 
 Among the very regular Caryophyllacece^ the small, 
 open Alsinece are never spotted or variegated ; and 
 even in the more developed Silenece, with their tubular 
 calyces, only the highest British genus, DiantJiiis 
 (Fig. 24), ever has marks on its petals. 
 
 In most regular flowers, the lines are mere intensi- 
 fications (or diminutions) of the general colouration 
 along the veins or ribs of the corolla ; and they 
 point towards the base or claw of the petal, where 
 the honey is usually secreted. But in irregular 
 flowers, we often get a higher modification of colour, 
 so that one region of the petal is yellow or white, while 
 another is pink or blue ; and these regions often run 
 transversely, not longitudinally. Such modifications 
 usually affect the most highly altered parts of the 
 irregular flower. 
 
 The common wild pansy, Viola tricolor^ affords a 
 good example of complex variegation. Its flowers 
 are purple, white, or yellow ; or have these pigments 
 variously intermixed. The two upper pairs of petals 
 are usually the most coloured ; the lower one is 
 broadest, and generally yellow at the base, with dark 
 lines leading toward the spur. Viola palustris ex- 
 hibits the same tendency in a less degree ; it is pale 
 blue, with purple streaks. V. canina has hardly more 
 than a few darker or lighter lines. The whole family 
 is immensely interesting from the present point of 
 view, and should be closely observed by the student 
 at first hand. Its changes and varieties will be found 
 full of instructive suggestions as to the origin and 
 nature of colour modifications. 
 
 InPolygala vulgaris,t\iQ two coloured petaloid sepals, 
 commonly called wings, are also delicately veined. 
 
VA RIEGA TION. 67 
 
 Among the PapilionacecB^ variegation and lines of 
 colour are common in the higher genera, which are 
 more strictly adapted to bees and other specialised 
 honey-suckers, such as Vicia, Lathyms, Onobrychisy &c. 
 The standard is usually the most highly coloured ; 
 the wings and keel are generally paler, or one degree 
 lower in the scale of progressive colouration. In 
 Lathyms hirsjitus, L. silvestris, Vicia Bithyjiica, and 
 Astragalus alpijius, this peculiarity is well marked. 
 The cultivated sweet-pea, a Sicilian Lathyms, illus- 
 trates the general principle even better than any of 
 our native kinds. 
 
 Among regular CorolliflorcBy variegation is not very 
 common, though it occurs much oftener than in the 
 polypetalous classes, especially at the throat of the 
 tube, as in the forget-me-nots {Myosotis) ; but in irre- 
 gular Corolliflovce it is exceedingly frequent. The 
 LentibtdacecB and other small families afford several 
 examples. In the great order of Labiatce, the highly 
 modified lower lip is very often spotted, especially 
 where it is most developed. This is the case in 
 Stachys silvatica, Lamium piirpurewUy Galcopsis tetra- 
 hit, Calamintha acinos, Nepeta cataria, N. glechoma, 
 Ajtiga reptanSy Scutellaria galericiilata, and many other 
 species. Several exotic kinds show the same tendency 
 in a more marked degree. 
 
 The ScropJiularinecB, however, form perhaps the 
 best example of any. It was noticed above that 
 comparatively few of these are a^ blue or as purple 
 as might be expected from their high organisation. 
 The explanation is that they have mostly got beyond 
 the monochromatic stage altogether, and reached the 
 level of intense variegation. They are, in fact, a 
 
 F 2 
 
68 
 
 THE COLOURS OF FLOWERS. 
 
 family with profoundly modified flowers (Fig. 25), most 
 of which are very specially adapted to very exceptional 
 modes of insect fertilisation. The Veronicas alone 
 among our English genera are simply blue, with 
 white or pink lines ; the others are mostly spotted or 
 dappled. AntirrJiimim majus is purple, sometimes 
 crimson or white, with the curiously closed throat a 
 bright yellow. Linaria cymbalaria is blue or lilac, 
 with white patches, and the palate a delicate primrose. 
 Z. spuria is yellow, with a purple throat. L. minor is 
 purple, with a white lower lip and yellow palate. 
 
 Fig. 25.— Flower of toad flax (^Linaria vulgaris), with corolla prolonged into a 
 honey-bearing spur ; yellow, with orange palate. 
 
 The very strange flowers of Scrophiilaria have a 
 curious, indescribable mixture of brown, green, dingy 
 purple, and buff. Sibthorpia is pink, with the two 
 smaller lobes of the corolla yellow. Digitalis pur- 
 picrea, the foxglove, is purple, spotted with red and 
 white. Euphrasia, eye-bright, is white or lilac, with 
 purple veins, and the middle lobe of the lower lip 
 yellow. Melampynmi arvense is red, with pink lips 
 and a purple throat. Description, indeed, is quite in- 
 adequate to convey any sufficient notion of the inti- 
 mate intermixture of hues in most scrophularineous 
 
VARIEGATION. 
 
 69 
 
 plants. As a rule, the spots or patches of intrusive 
 colour are developed transversely near the palate or 
 around the throat. Purple, red, or blue appear to be 
 the prevalent ground-tones, with white and yellow 
 introduced as contrasted tints to heighten the effect 
 of the principal constructive parts. 
 
 Among Monocotyledons, such plants as the highly 
 modified Iris genus show similar results. Our own 
 I.foetidisshna has blue sepals, with yellow petals and 
 
 r-N 
 
 , -^r^^ 
 
 
 f'lG 25.— Flower of spotted orchid (Prchi^ ma.cul.ita) ; purple, dappled with pink 
 
 and white. 
 
 spathulate stigmas, all much veined. The OrcJiidacecs 
 exhibit the same tendency far more markedly. 
 Orchis mascula, O. maciilata^ O. laxiflora, and many 
 other British species have the lip spotted (Fig. 26). 
 In 0. militaris and 0. hircina, the variegation is 
 even more conspicuous. In 0. ustulata, the spots on 
 the lip are raised. The problematical bee-orchid, 
 Opiirys apifera, is singularly dappled on the lip and 
 disk, and has the sepals different in colour from the 
 
70 THE COLOURS OF FLOWERS 
 
 rest of the flower. Aceras anthropophora, the man- 
 orchid, has green sepals and petals, edged with red, 
 and a yellow lip, pink fringed. Cypripedhmi calceolus, 
 the lady's slipper, CephalantJiera grandifloi-a, white 
 helleborine, and most other British species, are simi- 
 larly very diversified in colour. As to the exotic 
 species, some of them are more peculiarly tinted and 
 blended with half a dozen different hues than any 
 other forms of flowers in the whole world. 
 
 On the other hand, primitive yellow flowers of the 
 earliest type never have any lines or spots whatsoever. 
 We may therefore sum up the facts about variegation 
 as follows : — 
 
 (i.) Very simple and primitive flowers are usually 
 of one colour throughout. 
 
 (2.) Regular flowers of higher types are often 
 marked by lines of a different colour or shade, which 
 generally correspond with the venation of the petals. 
 
 (3.) Irregular flowers are often marked with distinct 
 patches of different colour (sometimes transverse), 
 which seem to act as honey-guides. 
 
 (4.) The most divergent colours usually appear on 
 the most highly modified parts of the flower. 
 
CHAPTER IV. 
 
 RELArSE AND RETROGRESSION. 
 
 Flowers which have reached a given stage in the 
 progressive scale of colouration often show a tendency 
 to fall back to a lower stage. When this tendency is 
 of the nature of a mere temporary reversion (that is 
 to say, when it affects only a few individuals, or a 
 casual variety), it may conveniently be described as 
 Relapse. When, however, it affects a whole species 
 and becomes fixed in the species by a new and pre- 
 sumably lower adaptation, it may best be styled Retro- 
 gression. The difference between these two forms of 
 reversion will become clearer after we have examined 
 a few cases of each in detail. 
 
 Primary yellow flovvers, like the buttercups, poten- 
 tillas, and St. John's worts, show little or no ten- 
 dency to vary in colour in a state of nature. They 
 have never passed through any earlier stage to which 
 they can revert ; and they are not likely to strike out 
 a new hue for themselves except through the action 
 of some special differentiating circumstances, such as 
 those ensured by cultivation. 
 
 Some v/hite flowers, on the other hand, show a 
 decided tendency occasionally to revert to yellow. 
 
72 THE COLOURS OF FLOWERS. 
 
 especially in the simpler orders. Erysimum orientale 
 varies from white to pale primrose. RapJianus ra- 
 phanistnim, as already noted, is usually even lilac, 
 often white, and on the sea-shore yellow. The 
 white cistuses often revert to a pale sallow tinge. In 
 some roses, the throat becomes yellow in certain 
 specimens. Many umbcllates vary from white to a 
 faint yellowish green. In several other cases, stray 
 yellow specimens of normally white species are not 
 uncommon. 
 
 Pink and red flowers almost invariably revert in 
 many individuals to white. Indeed, there is probably 
 not a single blossom of these colours in England of 
 which white specimens may not occasionally be 
 gathered. A few typical instances must suffice. All 
 the British roses are reddish pink or white. So are 
 Saponaria officinalis^ and many pinks. Malva moscJiata 
 runs from rose-coloured to white ; M. rotundifolia from 
 pale lilac-pink to whitish, Erodiiim cicutarinin has 
 rosy or white petals ; all the geraniums occasionally 
 produce very pale flowers. White varieties of heaths 
 are frequent in the wild state. Where the red or 
 purple is very deeply engrained, however, as in 
 labiates, reversion to white occurs less commonly. 
 But almost all pink or red flowers become white with 
 the greatest readiness under cultivation. 
 
 Blue flowers in nearly every case produce abundant 
 red, pink, and white varieties in a state of nature. It 
 would seem, indeed, as though this highest develop- 
 ment of colour had not yet had time thoroughly to 
 fix itself in the constitution of most species. In an 
 immense number of cases, it still appears as a late 
 modification of red, the bud or young petals being 
 
RELAPSE AND RETROGRESSION. 73 
 
 still of tliat colour, and only deepening into blue as 
 the flower opens. Hence individual reversion is here 
 almost universal as an occasional incident in every 
 species. The columbine {Aquilegia vulgaris) is blue 
 or dull purple, sometimes red or white. The larkspur 
 {^Delphinium ajacis) often declines from blue to pink 
 or white. The monkshood {Aconitum napellus) is an 
 extremely deep blue, very rarely white. White violets 
 everybody knows well. The rampions (P/iytcuma) 
 vary from blue to white ; so do many of the cam- 
 panulas. Gentiana campestris is sometimes white. In 
 most Boraginecs — for example, in borage, viper's bu- 
 gloss, and forget-me-not — pink and white varieties arc 
 common. Pink and white Veronicas also occur in 
 abundance among normally blue species. Prunella 
 vulgaris occasionally produces rosy or white blossoms. 
 White wild hyacinths are often gathered. Many 
 other cases will suggest themselves to every practical 
 botanist. 
 
 Blue flowers, however, very seldom revert to yellow, 
 though this change takes place in some cultivated 
 hyacinths, and somewhat differently in the pansy. As 
 a rule, the blue goes back only as far as those shades 
 from which it has more recently been developed. 
 This is, perhaps, the true rationale of De Candollc's 
 law of xanthic and cyanic types. 
 
 Sometimes, indeec, wo may say that the new colour 
 has not yet begun to fix itself in the species, b^t that 
 the hue still varies under our very eyes. Of this the 
 little milkwort {Polygala vulgaris) affords an excel- 
 lent example, for it is occasionally white, usually pink, 
 and frequently blue. Here we may fairly regard 
 the pink as the normal hue, while the white is doubt- 
 
74 THE COLOURS OF FLOWERS. 
 
 less due to reversion, and the blue to progressive 
 modification, not yet fully selected by insects ; so 
 that in all probability it is now actually in course of 
 acquiring a new colour. Much the same thing hap- 
 pens with the common pimpernel {Anagallis arvensis), 
 Its ancestral form is probably the' woodland loose- 
 strife {LysimacJiia nemoriini)^ for though the capsule 
 of the pimpernel now opens transversely, it still retains 
 the five dark lines which mark the primitive dehis- 
 cence ; and in other respects it most closely resembles 
 Lysijuachia, which is a bright yellow. But pimpernel 
 itself is usually orange-red, while a blue variety is 
 frequent on the continent, and sometimes appears in 
 England as well. Every botanist can add half a 
 dozen equally good instances from his own memory. 
 
 Highly variegated flowers show the greatest ten- 
 dency of any to such occasional reversions, or, as it 
 is usually put, are extremely variable. The pansy 
 ( Viola tricolor) is an excellent example. The snap- 
 dragons, orchids, and irises are also cases in point. 
 
 Indeed, the extent to which flowers are modifiable 
 in the hands of gardeners largely depends upon the 
 amount of modification which they have already 
 undergone in the natural condition. Very highly 
 developed plants have on the one hand acquired a 
 great constitutional plasticity of nature, and on the 
 other hand have a large number of previous stages to 
 fall back upon. Hence gardeners can do almost any- 
 thing they like with Dahlias^ Cinerarias, Asters, 
 CJirysaiithemunts, and other advanced corymbiferous 
 composites; with Calceolarias, Aiitirrhinnnis, and 
 other ScropJinlarinccc; with pansics, pea- flowers, heaths, 
 and lilies ; with exotic Gloxinias, Dignonias, Teconias, 
 
RELAPSE AND RETROGRESSION. 75 
 
 and Gesnerias.. Un the other hand, variegated or 
 largely altered flowers of the simpler types rarely 
 occur even in cultivation. Hence we may, perhaps, 
 reasonably infer that great readiness to assume new 
 colours affords in itself a certain slight presumption 
 in favour of some previous colour modification. We 
 shall apply this hypothetical principle in the sequel to 
 sundry cases of yellow colouration in apparently high 
 families, as one among several tests by which we may 
 be aided in distinguishing retrograde from primitive 
 yellowness. 
 
 This seems also the proper place to consider the 
 proofs of the position already advanced, that new 
 colours make their appearance at the edge of the 
 petal, and gradually work their way inward. Four 
 such proofs may be advanced. In the first place, 
 purely adventitious individual colours almost always 
 so appear. For example, the reddish tinge occa- 
 sionally observ^ed on many yellow flowers is usually at 
 the tip : so is the lilac tinge on certain white anemones, 
 and the pink tinge on many crucifers and umbellates. 
 In the second place, the slight blush which occurs 
 normally on flowers like the daisy, the apple-blossom, 
 and the blackberry, and which appears to be as yet 
 comparatively uninfluenced by insect selection, seeing 
 that it is deepest on the back of the petals, generally 
 occurs near the tip. The same peculiarity may be 
 observed also in several small CaryopJiyUaccce, Papi- 
 /ionaccic, and Ericacccc. In the third place, flowers 
 which open pink, like so many BoraginacciCf and then 
 become blue, remain always red at the base, and only 
 acquire the nev/ hue in the ex[)andcd limb of the corolla. 
 In tulips, Hydrangea^ Ric/uirdia, &c, like facts occur. 
 
76 THE COLOURS OF FLOWERS. 
 
 And in the fourth place, white varieties of blue iElowers 
 usually have the centre bluish and the edge white; 
 pink varieties have the centre bluish and the edge 
 pink, and so forth. Here, we know what the normal 
 colour is like, and can see that the new hue appears 
 first at the periphery. For example, white violets, 
 a variety of Viola odorata, have the spur and lower 
 part of the petals blue or bluish ; the whiteness only 
 extends to the broad part of the petals. In a large 
 number of varieties examined by the writer, the same 
 law holds good. Hence we are justified to some 
 extent in assuming that when a plant exhibits a 
 different colour at the base and at the tip of the 
 petal, the basal colour is probably more primitive 
 than the peripheral one. 
 
 If we turn from the white violet, with its blue spur, 
 to the very variegated pansy, we may perhaps ask 
 ourselves which is the earlier of its colours, the purple, 
 the white, or the yellow. But if we observe that the 
 spur, unseen at the back of the flower, is usually deep 
 violet blue, as are also the bases of the petals, while 
 the yellow is usually found on the most expanded 
 and modified part of the corolla, the lowest petal, and 
 in its xHost nodal or functionally attractive part, just 
 in front of the honey-cavity, we can hardly resist the 
 inference t^" ''ustupon us by analogy — that the pansy was 
 once all blue, and that the yellow has been developed 
 here, as in the snapdragon and the ivy-linaria, to guide 
 the bees to the proper place for securing the nectar 
 and effecting cross fertilisation. 
 
 It is an interesting fact in this connection that an 
 immense number of the very simplest flowers, when 
 not themsel/es yellow, have yellow spots or patches 
 
E ELAPSE AND RETROGRESSION. 77 
 
 at the bases of their petals. The reader is recom- 
 mended to notice this point for himself in the 
 commoner white or pink polypetalous flowers. 
 
 With the light thus cast upon the question to guide 
 us, we may pass on to the general consideration of 
 Retrogression in colours. Certain species of advanced 
 families have apparently found it advantageous in 
 certain circumstances regularly and consistently to 
 revert to colours lower in the scale than the normal hue 
 of their congeners. The reasons for such Retrogression 
 are often easy enough to understand. 
 
 We may take the evening campion {LycJmis vesper- 
 tind) as a good example (Fig. 15). This white flower, 
 as we saw, is evidently descended from the red day 
 campion [LycJinis diuriia), because it is still often pale 
 pink, especially towards the centre, verging into white 
 at the edge. But it has found it convenient to attract 
 moths and be icrtiliscd by them ; and so it has lost its 
 pinkness, because white is naturally the colour best 
 seen by crepuscular insects in the dusky light of 
 evening. It is scented at nightfall, and readily allures 
 many moths by the combined attraction of sight and 
 smell. Sir John Lubbock notes that such evening flowers 
 never have any spots or lines as honey-guides on the 
 petals, because such marks could not be seen at night, 
 and would therefore be useless. All the other British 
 species oi Lychnis are red, pink, or purple. 
 
 The evening primrose {Oenothera bicnuis), now 
 naturalised in England, is another excellent instance 
 of the same sort. It belongs to the family of the 
 Ofiagraccce, which are highly evolved polypetalous 
 plants, with the petals reduced to four or two in 
 number, and placed above instead of below the ovary. 
 
78 THE COLOURS OF FLOWERS. 
 
 We should thus naturally expect them to be pink or 
 lilac, and this is actually the case with most of our 
 native species, the genus Epilobmm having usually 
 purple or red flov/ers, rarely white ; while the smaller 
 CirccF.as are pink or whitish. Why, then, is the even- 
 ing primrose yellow ? Because it is a night-flowering 
 plant, fragrant in the evening, and its pale yellow 
 colour makes it easily recognisable by moths. In this 
 case, however, two points mark it off at once from the 
 really primitive yellow flowers. In the first place, it 
 has not the bright golden petals of the buttercup, but 
 is rather more of a primrose tint ; and this is a 
 common distinguishing trait of the later acquired 
 yellows. In the second place, it belongs to a genus 
 in whicli red and purple flowers are common, whereas 
 the buttercups are almost all yellow or whitey-yellow^ 
 and the potcntillas mostly yellow or white. In short, 
 primitive yellow flowers are usually golden, and belong 
 to mainly yellow groups : reverted yellow flowers are 
 often primrose, orange, or dull buff, and occur 
 sporadically among blue, red, or purple groups. 
 
 There are other cases less immediatelyapparent than 
 these. For instance, Laniium galeobdolon, a common 
 English labiate, belonging to a usually purple or blue 
 family, is bright yellow. But we can form some idea 
 of how such changes take place if we look at the 
 pansy, which wc have seen reason to believe is 
 normally violet-purple, but which usually has a yellow 
 patch on the lowest petal. In the pansy's var. lutea, 
 the yellow extends over the whole flower, no doubt 
 because this incipient form has succeeded in attract- 
 ing some special insect, or else grows in situations 
 where yellow proves more conspicuous to bees than 
 
RELAPSE AND RETROGRESSION. 7() 
 
 blue or purple. So, again, another English labiate, 
 Galeopsis tetraJdt^ the hemp-nettle, has a pale purple or 
 white corolla, sometimes with a tinge of yellow in the 
 throat : and in the var. versicolor y the yellow spreads 
 over all the flower, except a purple patch on the 
 lower lip. In G. ocJiroleiica, the whole corolla has 
 become pure yellow. In this way, one can under- 
 stand the occurrence of such a flower as Lainitim 
 galeobdolon, especially since an allied species, L. a/dum, 
 is white, and all the genus is extremely variable in 
 colour. Indeed, it is to be noted that the yellow 
 labiates do not commonly occur among the less 
 developed thymes, mints, and marjorams, but among 
 the extremely specialised StacJLydcce, which have very 
 modified flowers, and usually variegated or spotted 
 lips. They seem to be essentially reversionary forms 
 from purple or blue species, spotted with yellow. 
 Nay, the lower lip of L. galeobdolon itself still shows 
 marks of dark orange variegation; exactly answering 
 to that of several purple Laiiihims : and the base of 
 the corolla tube is still pink or purplish. 
 
 Another hint of Retrogression is given us by flowers 
 like our English balsams, Iiupaticus noli-inc-tangere 
 and /. fulva, in the fact that their yellow is generally 
 dappled with numerous spots of deeper colour. The 
 balsams are highly modified irregular GeraJiiaccce, 
 sepals and petals being both coloured : and at first 
 sight it seems curious that our species should be 
 yellow, while the simpler GeraniiDiis and Erodiiuns 
 are pink or red. But the genus as a whole contains 
 many red and variegated species, and alters in colour 
 with much plasticity under the hands of gardeners. 
 /. noli-nie-taugcre is pale yellow, spotted with red : 
 
8o THE COLOURS OF FLOWERS, 
 
 L fulva is orange, dappled with deep brown. Both 
 are almost certainly products of retrogressive 
 selection. 
 
 Something of the same sort is seen in CJieiranthus 
 cheiri, the wallflower. This large and highly de- 
 veloped stock-like crucifer is a peculiar yellowish 
 brown in the wild state, frequently even primrose or 
 primitive yellow. But it varies readily, often be- 
 coming red at the edges ; and under cultivation it 
 assumes numerous shades of red, purple, and brown. 
 It appears to be a product of retrogressive selection 
 from an original form like the European stock. This 
 flower, combined with some others like Adonis 
 aiitiunnalisy Ranunailus Ficaria^ and Lotus cornicidattiSy 
 seems to suggest the idea that yellow may sometimes 
 merge directly into red, without passing through the 
 intermediate stages of white and pink. The other order 
 would appear, however, to be the more regular and 
 usual gradation. 
 
 In the PrinndacecB, we find similar instances. 
 liottonia palustris, a less developed form, is rosy 
 lilac. Cyclamen europceiim is white or rose-coloured. 
 Trientalis europa^a is white or pale pink, with a yellow 
 ring. From such a stage as this, it is easy to get at 
 our primroses, cowslips, and oxlips, which have pale 
 yellow corollas, with orange spots at the throat. 
 Indeed, one English species, Prbmda farinosa, is 
 pale-lilac, with a yellow centre : and this might easily, 
 under special circumstances, become pale primrose all 
 over. The cultivated varieties of the cowslip, called 
 Polyanthuses, readily assume various tints of orange, 
 red, and pink, always at the edge, the deep yellow 
 of the throat remaining unchanged. On the other 
 
RELAPSE AND RETROGRESSION. «f 
 
 hand, the yellow of the allied LysimacJiias certainly 
 appears to be primitive. 
 
 Among heaths, our only yellowish sort is the 
 very degraded Monotropa, a leafless saprophyte of 
 the lowest type, obviously a product of extreme 
 Retrogression. 
 
 The colours of many ScropJmlarinece may be ex- 
 plained in the same way. Perhaps the yellow of the 
 mulleins is primitive ; but as some species are white or 
 purple, it is just as likely to be retrogressive. In 
 Liiiaria, we may almost be sure that retrogression has 
 taken place ; for we can trace a regular gradation from 
 lilac flowers with a yellow palate, like L. cymhalaria, 
 to pale yellow flowers, like L. vulgaris (Fig. 25), which 
 has the mass of the corolla primrose, and the palate 
 orange. Miimilns hitciis is also yellow, but it is usually 
 marked inside with small purple spots, and sometimes 
 has a large pink or red patch upon each lobe. In 
 Melampyrmn cristatiivi, the yellow corolla is variegated 
 with purple : in M. pratense, it has the lip deeper in 
 hue. All these genera include many purple and 
 variegated species ; and the yellow members almost 
 always bear evident marks of being descended from 
 polychromatic ancestors. 
 
 Among Lentibulacece, our debased Utriailaricc have 
 pale yellow flowers ; but Piugiiicnla vulgaris is bluish 
 purple, and Pingiiiaila liisitanica yellow, tinged with 
 lilac : so that here, too, we may suspect Retrogression. 
 
 The case of the yellow Compositor, especially the 
 Ligiilata;, is more difficult to decide. It would seem 
 as though these plants, which have all their florets ligu- 
 late (Fig. 27), must be more highly developed than the 
 Corymbifcrcc, which have only the ray-florets ligulate, 
 
 G 
 
82 
 
 THE COLOURS OF FLOWERS. 
 
 or than the Cynaroidecu, which have no hgulate florets 
 at all. Hence we should naturally pcct them to be 
 blue or purple, whereas they are loi the most part 
 yellow of a very primitive golden type, while the 
 ray- florets of the Corymbifers are usually white or 
 pink, and all the florets of the Cynaroids are usually 
 purple. It is, of course, quite possible that a flower 
 might have progressed as far as the CoroUifloral stage 
 — might have joined all its petals into a united corolla 
 under the influence of insect selection — and yet might 
 
 Fig. 27. — Vertical section of head of typical Hgulate ; yellow. 
 
 never have got beyond its primitive golden-yellow 
 colouration. Something of the sort may be the 
 case with Chlora perfoliata diVnong the GentianacecE ; it 
 may even be the case with the LysimacJiia genus 
 among the Priviulacece ; and therefore it may also have 
 been the case with the primordial ancestor of the 
 Composites. But there are great difficulties in the way 
 of this explanation. In the first place, it would be 
 hard on that supposition to understand why the less- 
 developed Cynaroids should have outstripped the 
 more-developed Corymbifers and the most-developed 
 
RELAPSE AND RETROGRESSION. 83 
 
 Ligulates. In the second place, it would be equally- 
 hard to understand why the most primitive 
 Corymbifcrs, such as Eupatoriiiniy should have 
 purple or white florets, while the more developed 
 genera, such as Aster and ChrysantJieniiini, should 
 have most of them yellow. The following hypothetical 
 explanation is suggested as a possible way out of this 
 
 difficulty. Y 
 
 The primitive ancestral composite had reached the 
 stage of blue or purple flowers while it was still at a 
 level of development corresponding to that of the 
 scabious or the Jasione. The universality of such 
 colours among the closely allied Dipsacecu, Valeraincce, 
 Lobeliacece^ and Caiiipamilaccce, adds strength to tliis 
 supposition. The central and most primitive group 
 of composites, the Cynaroids, has kept up the 
 original colouration to the present day ; it includes 
 most of the largest forms, such as the artichoke, and 
 it depends most of any for fertilisation upon the 
 higher insects. Very few of its members have very 
 small florets. All our British species (except the 
 degenerate Carlind) are purple, sometimes reverting 
 to pale pink or white, while Centaurca cyauus, our 
 most advanced representative of the tribe, rises even 
 to bright blue. 
 
 Next to the Cynaroids in order of development 
 come the Corymbifers, some of which have begun to 
 develop outer ligulate rays. ?Iere the least evolved 
 type, Eupatoriuin, with few and relatively large 
 florets, is usually purple or white, never yellow. But 
 as the florets grew smaller, and began to bid for the 
 favour of many miscellaneous small insects, reversion 
 to yellow became general. In a few cases here and 
 
 G 2 
 
84 THE COLOURS OF FLOWERS. 
 
 there we still find purple or white central florets, as in 
 Petasites vulgaris, the butter-bur ; but even then we 
 get closely related forms, like Titssilago farfara, colts- 
 foot, which have declined to yellow. The smallest and 
 most debased species, such as Solidago virga-aiLvea, 
 golden rod, Tanacetitm vtilgare^ tansy, and Senecio 
 vulgaris^ groundsel, have all their florets yellow and 
 similar ; unless, indeed, like GnapJialiwn and Filago, 
 cud-weed, Artemisia absinthium, wormwood, or Xa7i- 
 thiimi stnimaritim, burweed, they have declined as 
 far as colourless or green florets, in which case they 
 must be considered under our next head, that of 
 Degeneration. On the other hand, the larger and 
 better types of Corymbifers began a fresh progressive 
 development of their own. In many Senecios, Imilas, 
 Chrysanthemums, they produced yellow ray florets, 
 similar in colour to those of the disk. In Chrysau- 
 themum leucanthejnum, AntJiemis cotula, Matricaria 
 inodora, &c., these rays, under the influence of a 
 different type of insect selection, became while. In 
 the daisy they begin to show signs of pink, and 
 in \}iiQ Astei's, Cinerarias, &c., they have become lilac, 
 purple, and blue. Complicated as these changes 
 seem, they must yet have taken place two or three 
 times separately in various groups of Corymbifers, 
 for example in the Asteroidece, the AntJiemidece, and 
 the Scnecionido!. 
 
 The Ligulates were again developed from yellow- 
 rayed Corymbifers by the conversion of all the disk 
 florets into rays. Appealing for the most part to 
 very large and varied classes of miscellaneous insects, 
 they have usually kept their yellow colour ; but in a 
 few cases a fresh progressive development has been 
 
RELAPSE AND RETROGRESSION. 85 
 
 set up, producing the violet-blue or purple florets cf 
 the salsify {Tragopogon porrifolius), the deep blue 
 Sonchus alpimis, and the bright mauve succory, 
 CicJiorium intybus. As a whole, however, the Ligu- 
 lates are characterised by what seems a primitive 
 golden yellow, only occasionally rising to orange-red 
 or primrose in a few hawk weeds. 
 
 That this hypothetical explanation may be the 
 true one seems more probable when we examine the 
 somewhat similar case of the StellatcB. Here it seems 
 pretty clear that mere dwarfing of the flowers, by 
 throwing them back upon earlier types of insect 
 fertilisation, has a tendency to produce retrogression 
 in colour. Even in the more closely allied DipsacecEy 
 ValeriaiiecBy and CanipanulacecBy we see a step taken 
 in the same direction, for while the large-flowered 
 Campanulas and Scabiosas are bright blue, the smaller 
 flowered teasel {Dipsaais silvestris) is pale lilac, the 
 Valerianas are almost white, and the Valcriandlas 
 are often all but colourless. In the StcllatcCy the 
 same tendency is carried even further. As a whole, 
 these small creeping weeds of the temperate regions 
 form a divergent group of the tropical Rubiacece (in- 
 cluding CiucJioniaccce)y from which they are clearly 
 derived as a degraded or dwarfed sub-order. Their 
 square stems, their leaf-like interpetiolar stipules, and 
 their usually lessened number of corolla-lobes, all 
 point them out as derivative forms, not as survivals 
 from an early ancestral type. Now, the tropical 
 Rubiacece have tubular blossoms with long throats, 
 and as x rule with five lobes to the corolla ; but many 
 of the Stellates have lost the tube and one corolla 
 lobe. Sherardia arvensis, which has departed least 
 
86 THE COLOURS OF FLOWERS. 
 
 of our British species from the norma of the race, has 
 a distinct tube to the corolla, and is blue or pink. 
 Asperula, which approaches nearer to the retrograde 
 Galiums, has one pale lilac species and one white. 
 The Galiums have no corolla-tube at all, and most of 
 them are white ; but two British species, G. veriim and 
 G. criiciata, are yellow, and one of these has become 
 practically almost bi-sexual — a common mark of 
 Retrogression. Rubia peregrina is even green. This 
 clearly marked instance of Retrogression from blue 
 through lilac and white to yellow makes the case of 
 the Composites easier to understand. No doubt the 
 dwarfed northern Stellates have found that they sue 
 ceeded better by adapting themselves to the numerous 
 small insects of the fields and hedgerows, and there- 
 fore have fallen back upon the nf^utral colours, white 
 and yellow. 
 
 Something the same explanation may be hazarded 
 in the case of the Umbellifera;, a difficult family to 
 deal with satisfactorily. It would be rash to assert 
 that these curious flowers are descended from mono- 
 petalous ancestors, yet, with the analogy of Galium 
 before our eyes, a suggestion to that effect may at 
 least be entertained as possible. The relations of the 
 Umbdliferce with the Araliacece and Cornece are very 
 close ; and it is difficult to believe that the last named 
 family at least are not truly monopetalous. From 
 the Rtibiacecc, indeed, a regular line of affinities leads 
 us on through Caprifoliacea; and Cornea; to Uuibclli- 
 fercu ; and if we allow that Galium is descended from 
 a tubular form like Sherardia it is hard to draw a line 
 at such types as the elder, the wayfaring tree, the 
 dogwood, or the ivy, till we arrive at the true 
 
RELAPSE AND RETROGRESSION. 87 
 
 iimbellates themselves. Be this as it may, the family 
 is one which clearly lays itself out to attract a large 
 number of miscellaneous insects, as Miiller has shown ; 
 and its prevailing colours are white and pale yellowish 
 green. The flowers are all adapted to such small 
 visitors as prefer these hues. 
 
 " The position of the honey on a flat disk," says 
 Sir John Lubbock, " which renders it accessible to 
 most insects, has the opposite result as regards the 
 Lepidoptera, which therefore, as might naturally be 
 expected, are but rare visitors of the Umbellifevce. 
 I have sometimes wondered whether the neutral tints 
 of these flowers have any connection with the number 
 of species by which they are frequented." This 
 pregnant hint is full of meaning for the student of 
 floral colouration. 
 
 After so many instances of more or less probable 
 Retrogression, it will not surprise the reader to learn 
 that in an immense number of other cases there is 
 good reason to suspect some small amount of dwarf- 
 ing or even Degeneration. These cases might perhaps 
 be properly treated in the next chapter ; but their con- 
 nection with our present subject is so close that they 
 fall into place more naturally here. It may have struck 
 the reader, for example, when we were dealing with the 
 Crucifers, that many of the smaller white forms were 
 apparently lower in type than large and brilliant 
 yellow flowers like the charlocks. That is quite 
 true ; but then, many of these small types are de- 
 monstrably dwarfed and slightly degraded, as, for 
 example, Cardamine /lirsuta, which has usually only 
 four stamens instead of six, thus losing the most 
 characteristic mark of its family. In Soicbieva didyma, 
 
88 THE COLOURS OF FLOWERS. 
 
 the petals have generally become quite obsolete ; in 
 some species of Lepidium, Arabis, Draba, &c., they are 
 inconspicuous and often wanting. So in the smaller 
 AlsinecB there are many signs of Degeneration. The 
 normal forms of CaryopJiyllaccce have two whorls of 
 five stamens each ; but these little creeping or weedy 
 forms have often only one whorl, as in Holostewn, 
 some Cerastiums, the smaller Stellarias, Spergula, 
 Polycarpoit, &c. In Sagina, Cherieria, and other very 
 small types, the petals are often or always wanting. 
 Indeed, most botanists will probably allow that nearly 
 all our minute- flowered species, such as Montia 
 fontana^ Claytonia perfoliata, Elatine hexandray Radiola 
 Millegrana, Circcea bttetiana, Liidzuigia pains trisy 
 Peplis Porttda, Tillcsa inuscosa, MyriopJiylliun spicattim, 
 Hipptiris vidgaris, Centuncidiis ininivius^ and Cicendia 
 pusillay are distinctly degenerate forms. Though ob- 
 viously descended from petaliferous ancestors, and 
 closely allied with petaliferous genera or species, 
 many of them have lost their petals altogether, while 
 others have them extremely reduced in size. In 
 several cases, too, the number of sepals, petnls, or 
 stamens has been lessened, and the plant as a whole 
 has suffered structural degradations. Most of these 
 dwarfed and degenerate flowers, if they have petals 
 at all, have them white or very pale pink. 
 
 Readers of Sir John Lubbock's admirable little 
 book on British Wild-Flowers in Relation to Insects 
 will readily understand the reason for this change. 
 They will remember that white flowers, as a rule, 
 appeal to an exceptionally large circle of insect 
 visitors, mostl}'- of small and low grades. Hence, 
 some among these very small flowers may often 
 
RELAPSE AND RETROGRESSION. 89 
 
 succeed, in certain positions, better than larger ones. 
 Moreover, they will recollect that in numerous in- 
 stances the larger blossoms of each family are so 
 exclusively adapted to insect fertilisation that they 
 cannot fertilise themselves ; while among the smaller 
 blossoms alternative devices for self-fertilisation com- 
 monly come into play after the flower has been open 
 for some time, if it has not first been cross-fertilised. 
 Structural considerations show us that in most such 
 instances the larger and purely entomophilous flowers 
 are the more primitive, while the smaller and occa- 
 sionally self-fertilising flowers are derivative and 
 degraded, having usually lost some of their parts. 
 Hence, in tracing the progressive law of colouration 
 in the families generally, it is necessary, for the most 
 part, to consider only the larger and more typical 
 species, setting aside most of the smaller as products 
 of degeneration. 
 
 Moreover, as Mr. Henslow has shown, a large 
 number of minute species have fallen back almost 
 entirely upon self-fertilisation ; and these, we must 
 presume, when they retain their petals at all, retain 
 them either by mere hereditary habit, as functionless 
 relics, or else use them to effect an occasional cross 
 at long intervals. Such degenerate and dwarfed 
 species survive exactly as mites and other degraded 
 forms survive in the animal world — because they fill 
 certain holes and corners of the organic system better 
 than more highly developed forms could do. The 
 advantages of cross-fertilisation are seen in the large 
 stature and vigorous constitution of the truly en- 
 tomophilous or anemophilous plants, which have 
 
90 THE COLOURS OF FLOWERS. 
 
 usurped all the best and richest places in nature, 
 like dominant races that they are ; the small de- 
 graded types are the species that have fallen behind 
 in the struggle, and manage merely to eke out a 
 precarious existence in the back slums of nature's 
 economy. 
 
CHAPTER V. 
 
 DEGENERATION. 
 
 The cases detailed in the last chapter lead us 
 gradually up to the consideration of those very 
 degenerate flowers whose structure has become com- 
 pletely debased, and especially of those which have 
 green perianths instead of coloured corollas. As a 
 rule, evolutionists have taken it for granted that 
 green flowers were the earliest of any, and that 
 from them the coloured types have been derived 
 by insect selection. But if the principles laid down 
 so far in this little treatise be correct, then it is 
 obvious that, since all petals were originally yellow, 
 green petals must be degraded, or at least altered 
 types. Of course, the flowers of gymnosperms (in 
 their blossoming stage) are mostly composed of green 
 scales or leaves ; and so it no doubt remains true 
 that all flowers are ultimately descended from green, 
 or greenish, ancestors. But if petals are by origin 
 modified stamens, it will follow that all corollas at 
 least were once coloured ; and we shall probably see 
 reason in the sequel to extend the principle to all 
 perianths whatsoever. Without insisting upon the 
 rule too dogmatically, so as to embrace every kind 
 
92 THE COLOURS OF FLOWERS. 
 
 of angiosperm, we may, with some confidence, assert 
 that wherever a flower possesses a rudiment of a 
 perianth in any form, it is descended from coloured 
 and entomophilous ancestors. 
 
 Those who have read Professor Ray Lankester's 
 able little work on Degeneration will not be surprised 
 to find that this retrograde agency has played as 
 large a part in the vegetable as in the animal world. 
 We will begin by examining some of the most certain 
 cases, and then will proceed to those in which the 
 evidence is more remote, and the traces of the original 
 petaliferous structure more completely obliterated. 
 
 The Composites are, perhaps, in some respects, the 
 very highest family of entomophilous flowers now 
 existing on the earth. Their very structure implies 
 the long and active co-operation of insect fertilisers. 
 They could not otherwise have acquired the tubular 
 form, the united corolla, the sheathed anthers, the 
 compound heads of many-clustered florets. That 
 originally green flowers could attain to this stage of 
 development, and yet remain green, is simply incon- 
 ceivable. But the Composites contain also some of 
 the most degraded flower types in all nature. Be- 
 ginning with such forms as the common groundsel 
 {Senecio imlgaris), which has an inconspicuous yellow 
 rayless head, specially adapted to self-fertilisation, 
 we go on to plants like the Artemisias, with small 
 greenish florets, which have taken, or are taking, to 
 wind-fertilisation. Still more degraded are the Gna- 
 phalmmSf Filagos, and Antennarias, whose mode of 
 fertilisation is problematical. And at the very bottom 
 of the scale we get the little green XantJiium ; so 
 degenerate a form that its connection with the other 
 
DEGENERA TION. 93 
 
 Composites can only be traced by means of several 
 intermediate exotics, in every stage of progressive 
 degradation. Such conclusive examples clearly show 
 us that green flowers may occur as products of degra- 
 dation even in the most advanced families. 
 
 Adoxa moschatellina is another excellent specimen 
 of a green corollifloral blossom. This pretty little 
 plant is closely allied to the honeysuckles and ivies ; 
 but it has somehow acquired a light green corolla, in 
 place of a white or pink one. It is still entomophi- 
 lous, and scantily secretes honey, so that the reason 
 of the change cannot be immediately pointed out. 
 Perhaps its very inconspicuousness saves it from the 
 obtrusive visits of undesirable insect guests. The 
 flowers of Hedera helix, common ivy, are also yel- 
 lowish green. In the allied family of Umbelliferce 
 many flowers have declined to similar greenish tints ; 
 but this can hardly be their primitive colour, as they 
 have an inferior ovary, which marks high develop- 
 ment, even if they are not, as suggested, degenerate 
 gamopetalous forms. Sinyrnmm olusatrum in this 
 family, and Chrysosplenmm among the Saxifragace<2, 
 exhibit very well the steps by which green corollas or 
 perianths may be produced from originally white or 
 yellow flowers. Rtibia peregrina (wild madder) has 
 also greenish monopctalous blossoms. AK these are 
 entomophilous. Their high structural development 
 obviously negatives the notion that they ar-e primitive 
 green flowers ; and we must necessarily conclude that 
 they have become green for some special functional 
 purpose of their own. 
 
 Among the highly developed Encacece, greenish 
 flowers occur in Pyrola, and to a less extent in 
 
94 THE COLOURS OF FLOWERS. 
 
 Arbutus and Vaccmiuin. Pyrola, at least, is a 
 clearly degraded type. 
 
 The Orchids themselves, that most specialised of 
 entomophilous types, show us other examples of 
 flowers which have become more or less green ; such 
 as Malaxis pahcdosa, which has a yellowish tinge ; 
 Liparis loeselii, also yellowish ; Epipactis latifolia^ 
 greenish brown ; Listera ovata, grass-green ; Habe- 
 naria viridis, yellowish green, and Herminium 
 monorchism pale greenish yellow. Why these highly- 
 developed entomophilous blossoms should have found 
 green suit them better than white, pink, or purple, it 
 would be hard to say ; but the fact remains indis- 
 putable ; and it would be almost inconceivable that 
 flowers of so high a type should have remained green 
 all through the various stages of their long previous 
 development. We may confidently set them down 
 as products of incipient degeneration. 
 
 Among polypetalous flowers we get some equally 
 interesting facts. Green appears as a variegation- 
 colour on the highly-developed pea-flower, Lathyrus 
 silvestris, and some others. Helleborus viridis, a 
 doubtfully English ranunculaceous plant, has small 
 green petals, employed as nectaries, and concealed 
 by the large green sepals. It is entomophilous, and 
 much visited by insects. Instead of being one of 
 the least-developed Ranuntulacecey however, it is one 
 of the most advanced and highly differentiated types. 
 In the lily family, again, the onion genus {Allium) is 
 a small, and often degraded, group, whose more retro- 
 grade members produce green in place of purple or 
 white flowers. In Allium vineale, and some others, 
 the flowers often degenerate so far as to become 
 
DEGENERA TION. 95 
 
 small caducous bulbs. Here, degeneration is the 
 only possible solution of the problem presented by 
 the facts. 
 
 More frequently, however, reversion to wind-ferti- 
 lisation (probably the primitive habit of all flowering 
 plants) has produced green blossoms among angio- 
 sperms. This may result in two or three distinct 
 ways. Either the corolla may become dwarfed and 
 inconspicuous, or it may coalesce with the sepals or 
 calyx-tube, or it may cease to be produced alto- 
 gether. We may take the plaintains {Plantago) as 
 a good example of the first-named case. Here we 
 have tubular florets with four corolla-lobes, apparently 
 descended from some form not unlike Vcj'onica (though 
 with four cells to the ovary) but immensely degraded. 
 The corolla is thin and scarious, and its lobes are 
 tucked away at the sides, so as not to interfere with 
 the stamens and style. These, again, as in most 
 wind-fertilised plants, hang out freely to the breeze ; 
 so that the whole spike when flowering shows no 
 signs of a corolla from without, but seems to consist 
 entirely of scales, stamens, and styles, just like a 
 sedge or grass-plant. The expensive display of 
 petals is no longer useful to the plant, which, there- 
 fore, economises the material that would otherwise 
 be employed to allure the insects. It is impossible, 
 however, to examine the functionless corolla without 
 coming to the conclusion that Plantago must be de- 
 scended from an entomophilous ancestor. Indeed, 
 P. media still to some extent lays itself out to attract 
 small flies, by which it is even now often visited and 
 fertilised. 
 
 More degraded still is the allied Littorella, which 
 
96 THE COLOURS OF FLOWERS. 
 
 leads on to the very degenerate types of water- 
 plants. 
 
 The Rosacea; offer some good examples of green 
 flowers in which the petals have become quite extinct 
 Some of them are entomophilous, and some anemo- 
 philous. Alchemilla vulgaris (lady's mantle, Fig. 28) 
 is one of the former class. It is a degraded represen- 
 tative of the same group as agrimony ; but it has lost 
 its petals altogether. That it is a late, not a primitive 
 
 Fig. 28. — Flower of lady's mantle i^Alchcmilhi), with double calyx, but no petals; 
 
 green. 
 
 form, is shown by its very reduced carpels, and its small 
 number of stamens. AlcJicmilla arvensis (parsley- 
 piert) is an extremely debased moss-like descendant 
 of some similar ancestor. It has tiny green petalless 
 axillary flowers, self-fertilised, but occasionally visited 
 by minute insects. Not far from these may be placed 
 Potcriiim sanguisorha (Fig. 29), anotiicr degraded 
 type, which has become anemophilous. This flower, 
 too, is green, and has no petals ; it usually possesses 
 but one carpel, and it is altogether a clearly 
 
DEGENERA TION. 
 
 97 
 
 debased bisexual form. Its stamens are numerous, 
 and they hang out to the wind, as do also the 
 feathery stigmas in the female flowers, to catch the 
 pollen from neighbouring heads. But the closely- 
 allied Sarigitisorba officinalis (Fig. 30) is evidently 
 an entomophilous variation on the same ancestral 
 form ; for it resembles Poteriiun in every respect 
 except in its flowers, which have very few stamens, 
 enclosed in the purple calyx- tube. This interesting 
 case shows us that when a flower has once lost its 
 
 Fig. 29. — Flower of salad b>tmet 
 {Poienuin sans;iiisorbd)\ green 
 and anemophilous. 
 
 Fig. 30. — Flower of great hiirnet (Saiigut 
 so7-ba offici uiiii); purple and entomophi- 
 lous. 
 
 petals and become anemophilous, it cannot rc-dcvelop 
 them if it reverts to insect fertilisation, but must 
 acquire a coloured calyx instead. The same lesson 
 is perhaps elsewhere enforced by Glaiix maritima 
 among the Primnlacccc^ and by Clematis among the 
 Ranitnailaccce. 
 
 Mr. Darwin remarks that anemophilous flowers 
 never possess a gaily-coloured corolla. The reason 
 is clear. Such an adjunct could only result in the 
 attraction of stray insects, which would uselessly eat 
 up the pollen, and so do harm to the plant. Hence 
 
 II 
 
98 
 
 THE COLOURS OF FLOWERS. 
 
 when flowers revert to wind-fertilisation, both disuse 
 and natural selection cause them to lose their petals, 
 and become simply green. 1/ 
 
 In practice, however, it is often hard to distinguish 
 between the casually entomophilous, the self-fertilised, 
 and the really anemophilous species ; and they are 
 so intermixed that it may perhaps be best to consider 
 them together. For example, the common ash {Fraxi- 
 nus excelsior, Fig. 31) belongs to a gamopetalous 
 family, the Oleacece, and is closely related to the white 
 privet {Ligiistrinn vidgare), which has conspicuous 
 
 Fig. 31. — Hermaphrodite, male and female flowers of English ash {Fra.vinus 
 excelsior); purplish— no petals. 
 
 white flowers. But many large trees, owing, perhaps, 
 to their long life, and consequent less necessity for 
 producing many seeds, tend to lose their petals ; and 
 this is remarkably the case among the olive group. 
 The shrubby species have usually flowers with a four- 
 lobed corolla ; and so have many of the southern ar- 
 boreal forms (Fig. 32) : but the northern trees, like our 
 ash, have lost both calyx and corolla altogether, each 
 naked flower consisting only of two stamens, with a 
 single ovary between theni. In appearance their 
 blossoms seem oi much the same sort as the wind- 
 
DEGENERATION, 
 
 99 
 
 fertilised catkins and oak-kinds. Nevertheless, they 
 are entomophilous, for their pollen, their arrangement 
 in large masses, and their dark purple colour, suffi- 
 ciently serve to entice numerous insects. 
 
 The spurges {Euphorbiacece) are a very interesting 
 family of the same sort, exhibiting every gradation 
 from perfect corolliferous blossoms to the most de- 
 graded flowers in all nature. Our English species have 
 no true petals ; but some exotic forms are truly 
 
 Fij. 33. — Hermaphrodite perfect flower of South European flowering ash {Fraxinns 
 omits) ; white, with four-lobed corolla. 
 
 dichlamydeous ; and from them we can trace a gradual 
 decline, through plants like dog's mercury (Figs. 33 and 
 34, McrcunaHs percnnis)^ which has a green calyx, but 
 no corolla, to very degenerate green blossoms like our 
 own spurges {Euphorbia), which consist of several ex- 
 tremely simplified flowers, collected together in a 
 common involucre (Fig. 35). Plach separate male 
 floret is here reduced to a single stamen, raised on a 
 short jicdunclc, and with a distinct joint at the spot 
 where the petals once stood. It is worthy of notice, 
 
 H 2 
 
lOO 
 
 THE COLOURS OF FLOWERS. 
 
 too, that when these degenerate, but still entomophilous, 
 green flowers have found it desirable to attract insects 
 
 Fig. 33. — Male flower of dog's mercury 
 (Merctirialis) ; green. 
 
 Fig. 34. — Female flower ot dog's 
 mercury {Mercurt'alis)', green. 
 
 by developing new coloured surfaces in place of the 
 lost corolla, they have not done so by producing a 
 fresh set of petals, but have acquired coloured bracts 
 
 Fig. 35.— Inflorescence of spurge {Euphorbia), the male flowers reduced to a single 
 stalked stamen, the female fljwers to a naked ovary ; green. 
 
 or involucres instead, as in the well-known latrophas 
 and Poinsettias of our hot-houses. This instance is 
 
DEGENERA TION. loi 
 
 exactly analogous to that of the Smiguisorba, con- 
 sidered above. It tends to show that petals are not 
 developed from bracts, but from altered stamens. 
 
 From cases like these, we go down insensibly 
 through all the ranks of the dicotyledonous Mono- 
 cJilamydcc. In the ParonycJiiacece, for example, we 
 get an order closely allied to the CaryopJiyllacece 
 (especially to Polycarpon) ; and in one genus {Cor- 
 rigiolii) the flowers have small white petals, which 
 certainly aid in attracting insects. But in Herniaria 
 the flowers are quite green, and the petals are re- 
 duced to five small filaments, thus partially reverting 
 to their presumed original character as stamens. In 
 Sclerantlius the filaments are often wanting, and in 
 some exotic species altogether so. ^\\^ Amarantacece^ 
 unrepresented in Britain, approach the last-named 
 family very nearly, but have the petals altogether 
 obsolete ; and in many cases, such as Prince's feather 
 {Amaranthus hypocJiondriacus) and Love-lies-bleeding 
 {A. caudatus), the calyx becomes scarious and brightly 
 coloured. The Chenopodiacece are other near relations, 
 in which also the petals are quite obsolete ; and in most 
 of them the perianth (or calyx) is green. In Salicornia 
 it has become so embedded in the succulent leafless 
 stem as to be almost indistinguishable. The Poly- 
 gonacece^ on the other hand, are a group of plants, 
 allied to C/ieiiopodiacccc, but with a row of degraded 
 petals, and a strong tendency to produce coloured 
 perianths, analogous to that which we observed in 
 Sa/igiiisorba. The flowers of Riuncx, the docks, are 
 sometimes green, sometimes red ; those oi Polygoniuu 
 are pale-green, white, or pink. Rumex is sometimes, 
 Polygcvium constantly, fertilised by insects. 
 
1 02 THE COL O VRS OF FLO WERS. 
 
 Submerged or floating plants especially tend as a rule 
 to become green-flowered, and to grow very degraded 
 in structure. As instances, we may take Myrio- 
 phyllum, Ceratophylhun, Elodea^ Leinnay Callitriche, 
 Potamogeton, Ruppia, and Hippuris. In most of 
 these groups the proofs of great degeneration are 
 too obvious to need insisting upon. 
 
 There remain doubtful, then, among green Dico- 
 tyledons, only the highly anemophilous families, like 
 the nettles {Urticacece), and the catkin-bearing trees 
 {AmentifercB). The former have a well-developed 
 calyx, at least to the male flowers ; and it is difficult 
 to see how any one who com.pares them with Sclermi- 
 tlms ox Mercurialis, known descendants of pctaliferous 
 forms, can doubt that they too are degenerate types. 
 Indeed, the mere fact that the stamens are opposite to 
 the lobes of the calyx (Fig. 36), instead of alternate 
 with them, in itself shows that a petal-whorl has been 
 suppressed ; as is likewise the case in the goose-foots 
 and many other doubtful instances. Moreover, the 
 nettles are closely allied to the elms ( tZ/wrt-r^^), which 
 are obviously degenerate, and have acquired a coloured 
 perianth, side by side with their resumption of the 
 entomophilous habit. 
 
 As to the AmentifercB, Cnpuliferce, and other catkin- 
 bearers, at first sight we might suppose them to be 
 primitive green anemophilous orders. But on closer 
 consideration, we may see grounds for believing that 
 they are really degenerate descendants of entomo- 
 philous plants. In the alder {Almis) the male catkins 
 consist of clustered flowers, three together under a 
 bract, each containing a four-lobed perianth, with four 
 stamens within (Fig. 37). These little florets exactly 
 
DEGENERATION, 
 
 103 
 
 resemble, on a smaller scale, those of the nettle; 
 and the stamens here, again, are opposite to the' 
 calyx-lobes, which of course implies the suppression 
 
 Fig. 36.-Male rtower of netile {^Urtica atoica); yreui, with stamens opposite 
 
 the sepals. 
 
 of a corolla. In the beech {Betidd) the three florets 
 under each bract are loosely and irregularly arranged ; 
 and in the male hornbeam {Carpinus) and hazel {Cory- 
 his) the perianth is wholly obsolete. All these are 
 
 Fig. 37.— Flowtrs of alder {Alnus); jrreen, with stamens 
 
 opposite the sepals. 
 
 probably quite anemophilous. The willows {Salix), 
 on the other hand, though included by Sir John 
 Lubbock in the same category (doubtless through 
 
104 
 
 THE COLOURS OF FLOWERS. 
 
 inadvertence) have really become once more ento- 
 mophilous ; and they are much visited by bees, 
 which obtain honey from the small glands between 
 the florets and the axis (Figs. 38 and 39). Degenerate 
 as these last-named species undoubtedly are, they 
 may be connected by a regular line of illustrative 
 examples (not genetically) through the beech, aider* 
 
 Fig. 38.— Male flower of willow 
 {Salix) ; greenish. 
 
 Fig. 39 —Female flower of willow 
 {Sa/ix); greenish. 
 
 nettle, goosefoot, SclerantJuis^ Hemiaria, and Corri- 
 giola, with such perfect petaliferous types as the pinks, 
 and ultimately the buttercups. 
 
 Among Monocotj'Iedons, the very degraded little 
 entomophilous flowers of the Arum (Fig. 40), enclosed 
 in their green spathe, are often spoken of as though 
 they represented a primitive type. In reality, how- 
 ever, they are degenerate dichlamydeous blossoms, 
 linked to the lilies by Acorns (Fig. 41), which has 
 numerous hermaphrodite flowers, each with a perianth 
 of six scales, two rows of stamens, and a two-celled 
 
DEQENEEA TION. 
 
 JOS 
 
 or three-celled ovary. Here, again, the green flower 
 is obviously of late date. 
 
 What, then, are we to say about the ancmophilous 
 Monocotyledons, the great families of the sedges and 
 grasses ? Surely these, at least, are primitive green 
 wind-fertilised flowers. Dogmatically to assert the 
 contrary would, indeed, be rash with our existing 
 
 Fig. 40. — Spike or spadix of cuckoo-piiu 
 {A rum tnacitlatum), consisting of nu- 
 merous naked male.female, and neuter 
 flowers, in separate clusters ; purplish 
 green. 
 
 Fn;. 41. — Single floret of sweet sedge 
 (Acorus), wiih six perianth pieces. 
 six stamens, and an ovary ; greenish 
 yellow. 
 
 knowledge ; yet we may sec some reason for be- 
 lieving that even these highly ancmophilous types 
 are degenerate descendants of showy petalifcrous 
 blossoms. For, if the origin here assigned to petals 
 be correct, it becomes clear that the Juncacece, or 
 rushes, are only Liliacecu in which the perianth has 
 become dry and scarious ; for the absolute homology 
 
io6 
 
 THE COLOURS OF FLOWERS. 
 
 of parts in the two orders cannot, of course, be 
 denied. Some rushes, such as Ltisula, approach 
 very closely in general character to the grasses ; 
 and they also show themselves to be higher types 
 by the further development of the ovary, and the 
 decreased number of seeds. Eiiocaulon and the 
 RestiacecB give us a further step towards the grass- 
 like or sedge-like character. Some of the Cyperacece 
 show apparent relics of a perianth in the bristles which 
 surround the ovary, especially in Scirpiis (Fig. 42) ; 
 
 Fig. 42. — Flower of a sedge (^Scirpvs). with six hypogynous bristles, representing the 
 
 calyx and corolla. 
 
 and perhaps the perigynium of Ca7'ex may represent 
 a tubular perianth, though this is far more doubtful. 
 In the grasses {Grajuinecu) the perianth is either 
 altogether obsolete, or else is reduced to the palese 
 with the hypogynous scales or lodicules (Fig. 43). 
 According to the most probable view (Fig. 44), the 
 two paleae represent the calyx (for the inner palea 
 exhibits rudiments of two sepals, thus making up, 
 with the outer palea, a single trinary whorl) ; while 
 
DEGENERATION 
 
 107 
 
 the lodicules represent two of the petals, the third 
 (the inner one) being usually obsolete. It is fully 
 developed, however, in the bamboo. The connection 
 IS here less clearly traceable than in the Amentifercc 
 but It IS still quite distinct enough to suggest at 
 least the possibility that even grasses and sedges are 
 ultimately derived from entomophilous flow^ s. 
 
 There is, however, another and more powerful arjru- 
 ment against the idea that any of these existing green 
 
 flowers are really primitive. For what are the known 
 marks of the most primitive existing flowers ? Nu- 
 merous simple superior carpels ; distinct flowers on 
 separate peduncles ; no specialised bracts, no heads 
 no complications of any sort. And what are the 
 known marks of late and more developed or degraded 
 flowers } Unification of the pistil by union or sup- 
 pression of the carpels ; grouping of flowers in heads ■ 
 separation of sexes; multiplication of accessory parts' 
 
io8 
 
 THE COLOURS OF FLOWERS. 
 
 involucres, bracts, glumes, glands, awns, and so forth. 
 Now, to which of these classes do the yellow flowers 
 ordinarily belong ? Clearly to the first. To which 
 do the green flowers ordinarily belong } Clearly to 
 the second. The organisation of the catkins, the 
 sedges, and the grasses is exactly analogous to that 
 of the spurges, which we know by an unbroken line 
 of intermediate links to be descended from petali- 
 ferous ancestors. The inference is almost irresistible 
 
 a.i. 
 
 Fig. 44 —Diagram of flower of grass, a, sepals ; a i, outer sepal, flowering glume. 
 or outer palea ; a 2 and a 3. inner sepals, combined into a single inner palea ; 
 b, petals ; b 1 and /5 2, the lodicuks ; b 3, suppressed ; c, stamens, all present \ d, 
 styles or stigmas; d\ and ^2, present ; ^3, suppressed. The whole inner side 
 of the flower is thus abortive. 
 
 that so highly complicated a flower as that of the 
 grasses, with its one-celled, one-ovuled ovary, its two 
 styles, and its advanced paraphernalia of lodicules, 
 paleae, and glumes, arranged in long and subdivided 
 spikes, must be a very specialised or degenerate, not 
 a primitive or early type. The more closely we 
 examine green flowers, the more do we see that 
 they form the opposite pole to such simple and truly 
 
DEGENERA TION. 1 09 
 
 primitive forms as the buttercups, the potentillas, the 
 Alismacece^ and the simpler HHcs of the Gagea type. 
 
 Thus we are led, at last, to the somewhat unex- 
 p^ected conclusion that anemophilous angiosperms 
 are later in development than entomophilous angio- 
 sperms, and are derived from them. Though the 
 earliest flowering plants — the pines, cycads, and other 
 gymnosperms — were undoubtedly anemophilous from 
 the first, yet the probability seems to be that all 
 angiosperms were originally entomophilous, and that 
 certain degenerate types have taken later on either to 
 self-fertilisation, or to fertilisation by means of the 
 wind. Why this apparently retrograde change has 
 proved beneficial to them it would be impossible pro- 
 perly to inquire at the close of a work devoted to the 
 simple question of the colours of flowers. We must 
 content ourselves with noting that such degraded 
 green flowers fall for the most part under one or 
 other of four heads: (i) dwarfed or weedy forms; 
 (2) submerged or aquatic forms ; (3) forest trees ; 
 (4) grass-like or plaintain-like plants of the open 
 wind-swept plains. That there are no primitive 
 families of green or anemophilous angiosperms, it 
 might perhaps be rash and premature to assert ; 
 but at least we may assume as very probable the 
 principle that wherever green flowers possess any 
 perianth, or the relic or rudiment of any perianth, 
 or are genetically connected with perianth-bearing 
 allies, they have once possessed coloured insect- 
 attracting corollas. In short, green flowers seem 
 always (except in gymnosperms) to be the degene- 
 rate descendants of blue, yellow, white, or red ones. 
 
CHAPTER VI. 
 
 MISCELLANEOUS. 
 
 A FEW general hints upon various side questions 
 may here be conveniently thrown together in con- 
 cluding our hasty survey. They must be accepted 
 in most cases merely as suggestions for observation 
 on the reader's own part. The subject is still a new 
 one, and only vague ideas can as yet in certain 
 directions be formulated upon it. 
 
 We have seen in several cases already that flowers 
 which have lost their corolla often tend to re-develop 
 brilliant colours in their calyx, as in Sanguisorba] while 
 flowers which have lost both corolla and calyx often 
 tend to re-develop such colours in bracts, involucres, 
 or leaves, as in latropJia and Poinsettia. Among our 
 British MonocJdauiydce there are comparatively few 
 instances of such coloured calyxes, Glaiix, Daphne 
 mesereiim, and Ulmiis, being our best examples ; but 
 in many well-known exotic species, such as Mirabilis 
 dicliotoma, marvel of Peru, and the BcgL>. 'as, the 
 calyx is quite as beautifully coloured as any corolla. 
 In BougainviUcay three lilac bracts form the attractive 
 organ. In Aristolochiay the tubular calyx simulates 
 an irregular corolla, and in A. cordata it is large and 
 
MIS CELL A NEC US. 1 1 1 
 
 brilliantly scarlet. In RicJiardia africana, the so- 
 called Ethiopian lily, the spathe, surrounding a 
 spike of very degraded achlamydcous flowers, is 
 pure white, and very attractive. AmJierstia nobilis, 
 Brornelia piiigiiin, several species of Salvia, and 
 many other exotics, have handsome bracts, which 
 add greatly to their beauty. 
 
 The fact that in such cases flowers do not develop 
 new petals from bracts or leaves, but acquire instead 
 coloured calyxes or involucres, goes to prove the 
 validity of the view with which we set out, that 
 petals are really altered stamens, not altered leaves 
 or sepals. For if they could once be developed from 
 leaves, there would be no reason why they should 
 not be developed from them here again. But if they 
 were developed from stamens, and then lost in these 
 instances, we could easily understand why the plant 
 could not afford to waste any more of its now 
 diminished number of stamens for purely attractive 
 purposes, and so was forced to pour the necessary 
 pigment for alluring insects into the other surround- 
 ing organs. In other words, on the Wolfian principle, 
 there would be no reason why flowers with petals 
 should not appear sporadically among monochlamy- 
 deous families ; on the principle here advocated, it is 
 quite clear why stray entomophilous species, de- 
 veloped from these degraded types, should have 
 coloured calyxes, instead of coloured petals. 
 
 Among highly-developed, or succulent plants, the 
 calyx and bracts often tend to assume colours like 
 those of the petals, as do also the peduncles and the 
 stem. Cases occur in Ajitga rcptans, Echiuvi, Scdiim, 
 and Rinncx^ among British plants ; and more notice- 
 
1 1 2 THE COL O URS OF FL WERS. 
 
 ably in Pcperomia, EcJievei'iay EpipJiylbun, and other 
 exotics. The calyx and the expanded stipules which 
 cover the young flower- heads in red clover {Trifolitiin 
 pratense) and many other clovers, are delicately pink 
 or purple. In T. ai'vense the sepals are pale lilac, 
 and in T. ijicarfiatum pale yellow. The whole upper 
 portion of the flowering stem in Clirysosplenium is 
 bright golden. 
 
 Where the calyx is largely exposed to view, as in 
 the globe-flower {Trollhis), the columbme, the helle- 
 bores, and the monkshood, it is apt to become quite 
 as brilliant as the petals. In such cases its coloura- 
 tion usually follows the same law of progressive 
 development as the corolla. Sometimes, under these 
 circumstances, the now almost useless petals are 
 suppressed altogether, as in CaltJia, a near relative 
 of Trollius, as well as in the Anemones and Clematis. 
 At other times they arc utilised as nectaries, as 
 in columbine, hellebore, and monkshood. In the 
 meadow-rues {TJialictrnni) the petals are suppressed, 
 and the sepals very small, so that the flower depends 
 for attractiveness almost entirely upon its clustered 
 yellow stamens. In Impafiens, Polvf^ala, and some 
 other British genera, sepals and petals share almost 
 equally in the attractive display. 
 
 Where the petals have become much dwarfed, the 
 calyx is apt to aid them, if brilliant colouration again 
 becomes necessary. For example, our own wild 
 gooseberry, wild currant, and most other members 
 of the Ribcs genus, have very inconspicuous petals ; 
 but in the North American scarlet Ribes of our 
 gardens {R. sani^uineus), the flower has re-assumed a 
 brilliant colouring, and it has done so by making its 
 
MISCELLANEO US. 113 
 
 calyx bright red, instead of by increasing the size 
 and deepening the hue of its small white petals. In 
 the Fuchsia, the Hydrangea, and many other well- 
 known exotics, we get exactly similar devices. 
 
 Parasites and saprophytes do not as a rule require 
 to produce green leaves ; hence, most of them, like 
 Cuscuta, Orobanche, Lathrcsa, and Monotropa, have 
 the stem and leaves (or scales) coloured like the 
 flowers. Imperfect parasites which contain chloro- 
 phyll, however, have the leaves more or less green, 
 as in Viscum, Bartsia, RJiinantJiiis, and Melampyrum. 
 
 The outer florets of compound heads are apt to 
 produce larger petals than the inner ones, as in many 
 Umbellates (like cow-parsnip), the guelder-rose, the 
 Hydrangea, and the rayed forms of Composites. These 
 are obviously intended to increase the total attractive 
 effect. In the Umbellates and in candytuft the outer 
 petals of the individual flowers grow longer than the 
 inner ones. 
 
 Petals have perhaps been independently developed 
 from .stamens at least twice over, once in the Dico- 
 tyledons, and once in the Monocotyledons. Insects, 
 having once learnt to visit coloured surfaces in search 
 of pollen and honey, would naturally tend to visit all 
 such surfaces in future, and thus to select for fertili- 
 sation any coloured flowers that offered them any 
 attraction in the way of food, of whatsoever sort. 
 Apparently, at last one species of gymnosperm, the 
 larch {Pinus larix), has thus become entomophilous, 
 its fertile scales being interspersed with bright pink 
 or red empty bracts, which seem to subserve an 
 attractive function. They are certainly visited by 
 insects, perhaps in search of some secretion from 
 
 I 
 
114 THE COLOURS OF FLOWERS. 
 
 the bracts. This case may be looked upon as 
 analogous to those of the ripe cones in the juniper 
 and yew, which have similarly assumed the guise of 
 attractive fruits, eaten by birds, who disperse their 
 seeds. Such gymnospernis may be said metaphori- 
 cally to have taken a hint from the angiosperms 
 about them, and have acted upon it for their own 
 advantacje. 
 
 It has been assumed throughout that the progres- 
 sive modification of the colours of petals is due in 
 the main to increased oxidation of their contents. It 
 may be added here that the thin edges of the petals 
 where the newer colours usually first make their ap- 
 pearance, are the parts where oxidation would most 
 naturally take place. Hence, probably, the distinct 
 analogy between fading colours and progressive 
 colours. In most cases, colours appear most vividly 
 on the outside of the petals, where they were ex- 
 posed in the bud, and where oxidation would most 
 readily occur. The red tinge on the outside of 
 daisies, apple-blossom, &c., is here once more very 
 significant. In Convolvulus arvensis the mass of the 
 corolla is white ; but the lines exposed in the bud 
 are deep pink, evidently from oxidation ; and at the 
 same time they form excellent honey-guides of the 
 ordinary simpler sort. In many others of the same 
 genus a similar result may be observed. The under 
 side of the petals in St. John's worts, and the back 
 of the standard in Lotus, are frequently ruddy red. 
 The outer glumes of grasses are often purplish ; the 
 fruiting perianths of Rumex grow red as they ripen. 
 Put beside the rosy cheeks of peaches, apples, and 
 many oilier fruits, and the obvious oxidation colours 
 
MISCELLANEOUS. 1 1 5 
 
 of injured parts and autumn leaves, these facts are 
 full of functional significance. 
 
 In a single flower, the common pink Phlox, a 
 change apparently takes place in the reverse order 
 to that laid down in this treatise as the general law, 
 for it presents early in the morning a light blue tint, 
 and retrogrades to pink as the sun advances in the 
 sky. But it has been suggested (quite apart from 
 our present theory) that the blue colour is due to the 
 presence of some substance which becomes blue by 
 non-elimination of oxygen during the night ; and as 
 the oxygen is given out during the day, the blue 
 colour disappears. If this theory be well founded, 
 the apparent exception really confirms our rule. 
 
 It has been objected by two or three authoritative 
 critics that the original petals need not necessarily 
 have been yellow, because they represent the flat- 
 tened filament, not the anthers ; and it is the pollen 
 that gives the yellow colour to most stamens. But it 
 may be answered that in the primitive yellow flowers 
 (for example, the buttercups) the filaments are usually 
 of the same golden yellow as the petals ; and in many 
 other flowers they retain more or less of a yellowish 
 tinge. In white flowers they show a strong tendency 
 to become white ; but in pink and blue ones, pink 
 or blue filaments are comparatively rare. Sometimes, 
 indeed, the filaments become brightly coloured, so as 
 to share in the attractive display ; as a rule, however, 
 they are yellow in the yellow flowers, white or greenish- 
 yellow, with more or less of a pinky tinge, in almost 
 all others. The subject is certainly one which requires 
 further investigation. 
 
 According to Sachs, the yellow pigment of the 
 
1 1 6 THE COL URS OF FL WERS. 
 
 flowers here described as primitive is usually identi- 
 cal in composition with the ordinary yellow chloro- 
 phyll of leaves, while the orange, pink, red, and blue 
 pigments are of more elaborate kinds. 
 
 If the botanical reader will provisionally accept the 
 principles laid down in this little book, and will then 
 test their validity by applying them to the flowers 
 which he meets in his daily walks, he wi'l find that 
 many other confirmatory examples occur to him at 
 every step, most of which are too numerous to insert 
 here. He will also often find that close inspection 
 reveals some unexpected answer to a superficial 
 difficulty, some solution for the problem of an ap- 
 parent exception, which can only be obtained by 
 personal examination of the specimen with that 
 particular object held definitely in view. For ex- 
 ample, the case of the dead-nettle {Lainiiun album) 
 was cited above as one of a labiate grown white by 
 reversion (Fig. 45). This may have seemed at the 
 time a purely gratuitous and arbitrary supposition. 
 Why should not the white form be primitive, and the 
 purple or pink ones be derived from it } But if 
 the flower of a dead-nettle be carefully examined, 
 it will be found in most cases to be not purely white, 
 but to have some dusky lines and markings on its 
 lower lip, of a pale brown or dim grey-black, which 
 exactly answer to those on the lip of L. macidatum, 
 and in a less degree of L. piirpureutn. Now, such 
 markings do not occur among original white flowers 
 like the crucifers and Caryophyllacecs ; but they are 
 common on the lower lip of purple labiates. More- 
 over, we know that Lavtium maailatiim is very closely 
 allied to L. aibum^ and that it is purple-red instead 
 
MfSCELLANEO US. 
 
 tiy 
 
 of white. Both have the leaves occasionally marked 
 in the centre with a white line or spot, which is a 
 symptom of very close relationship. Indeed, Mr. 
 Bentham formerly united the two in a single species ; 
 and even now he is doubtful whether they should be 
 regarded as more than mere varieties. When we 
 consider that all purple labiates are liable to be 
 spotted with white ; that the purple and white forms 
 are here closely allied ; that in Galeopsis tetrahit we 
 
 Fi(;, 45.- -Vertical section of dead-nettle {Lainiuin album) \ white, with dark lines. 
 
 have a regular gradation from pure purple flowers to 
 almost pure yellow ones ; that in Lamium galeobdolon 
 we have a related yellow form similarly spotted ; that 
 all the Lamiums show a tendency to variegation ; and 
 that the white flower has itself an inconspicuous and 
 probably function less variegation, where the purple 
 one has conspicuous and useful honey-guides, the 
 inference is almost irresistible that the common white 
 form, L. album, represents a retrogressive modification 
 
1 1 8 THE COLO VRS OF FLO WERS. 
 
 of the rarer and obsolescent purple form, L. 
 maculatuni. 
 
 It would, of course, be impossible to treat every 
 similar instance at equal length without swelling this 
 volume to an unreasonable extent. But if the reader 
 will carefully examine, at first hand, all cases of what 
 seem to him adverse examples, he will usually find 
 some such hint of the true relation, surviving in the 
 flower itself Excellent studies may thus be made of 
 Teucriiim scorodonia, compared with our three other 
 British Teiicriums (where the calyx suggests the rela- 
 tive stages of development) ; of Ajuga cJiaincepitys 
 with A. reptans ; of our three Melanipyriinis \ of 
 RliinantJms and PediculariSy and of the various 
 Linarias. The OrobancJies are also full of instruc- 
 tivencss, as are likewise Pingiiicula and Utricularia. 
 The descriptions given in Floras and other botanical 
 works, and even the best coloured plates, supply very 
 inadequate ideas of the minute observation involved 
 in the study of this subject from the evolutionary 
 point of view. Dried specimens are of course almost 
 useless. The investigation must be conducted upon 
 the living corolla in all stages of its development. 
 Those who will take the trouble thus to watch the 
 actual growing flowers for themselves will soon learn 
 to recognise many other little marks of relative pro- 
 gress or retrogression which cannot all be set down 
 definitely in black and white without unnecessary 
 and tedious prolixity. 
 
 If the general principle here put forward is true, 
 the special colours of difl"erent flowers are due to no 
 mere spontaneous accident, nay, even to no mean- 
 ingless caprice of the fertilising insects. They are 
 
^TTSCELLANEOUS. i , 9 
 
 due in their inception to a regular law of progressive 
 modification ; and they have been fixed and stereo- 
 typed in each species by the selective action of the 
 proper beetles, bees, moths, or butterflies. Not only 
 can we say why such a colour, once happening to 
 appear, has been favoured in the struggle for exist- 
 ence, but also why that colour should ever make its 
 appearance in the first place, which is a condition 
 precedent to its being favoured or selected at all. 
 For example, blue pigments are often found in the 
 most highly-developed flowers, because blue pigments 
 are apparently a natural product of high modification 
 —a simple chemical outcome of certain extremely 
 complex biological changes. On the other hand 
 bees show a marked taste for blue, because blue is 
 the colour of the most advanced flowers ; and by 
 always selecting such, where possible, they both keep 
 up and sharpen their own taste, and at the same 
 tmie give additional opportunities to the blue flowers, 
 which thus ensure proper fertilisation. May we not 
 say that it ought always to be the object of naturalists 
 m this manner to show not only why such and such 
 a "spontaneous" variation should have been favoured 
 whenever it occurred, but also to show why and how 
 It could ever have occurred at all ? 
 
 THE END. 
 
fonbon : 
 R. Ci.AY, Sons, and Taylor, 
 
 BREAD STREET HILL. E.C. 
 
NATURE SERIES. 
 
 THE SPECTROSCOPE AND ITS APPI irATIOMO 
 
 '^"sEn?^Fl'^ ,;^,JiP, METAMORPHOSES OF IN- 
 THE TRANSIT OF VENUS. By G Forbes P A 
 
 ■^"^RS^S^'.^^^r ''^O^- Ey St. George Mivart, 
 
 t.R.S. Illustrated. Crown 8vo, 3s. 6d. ^^^y^i\±, 
 
 Sva^'aV'^sT''"' °' ''^ ^°^^^ ^°"^'y- ^""^^•■^''•i- Second £di^io^Cro^ 
 
 ^^o.??^'^^^^^ ^^ILD FLOWERS CONSIDERED IN 
 THE SCIENCE OF WEIGHING AND MEASURING 
 
 B>' H. W. CHISHOLM, Warden of the Standard. Illulated^ cLTsro! 
 
 HOW TO DRAW A STRAIGHT LINE : A Lecture on 
 
 L.nkages. By A. B. KEMPE, B.A. Illustrated. Crown 8v^ xfS 
 
 LIGHT: A Series of Simple, Entertaining and Useful 
 
 SOUND : A Series of Simple, Entertaining and Inexpensive 
 
 TfchnoJ«y. &^c^- ^^L^^nier^Tl^^LloJ^^rrowVj^^^^^ ^^ 
 
 SEEING AND THINKING. By W. K. Clifford F R S 
 
 0/Aers tofoUoxv. 
 
 MACMILLAN AND CO., LONDON. 
 
Published every Thursday, price 6d. ; Monthly Parts, 2S. and 
 2s. 6d.y Half' Yearly Volumes, 15^. 
 
 NATURE: 
 
 AN ILLUSTRATED JOURNAL OF SCIENCE. 
 
 Nature expounds in a popular and yet authentic manner, 
 the Grand Results of Scientific Research, discussing 
 the most recent scientific discoveries, and pointing out 
 the bearing of Science upon civilisation and progress, and 
 its claims to a more general recognition, as well as to a 
 higher place in the educational system of the country. 
 
 It contains original articles on all subjects within the 
 domain of Science; Reviews setting forth the nature and 
 value of recent Scientific Works ; Correspondence Columns, 
 forming a medium of Scientific discussion and of intercom- 
 munication among the most distinguished men of Science ; 
 Serial Columns, giving the gist of the most important 
 papers appearing in Scientific Journals, both Home and 
 Foreign ; Transactions of the principal Scientific Societies 
 and Academies of the World, Notes, &c. 
 
 In Schools where Science is included in the regular 
 course of studies, this paper will be most acceptable, as 
 it tells what is doing in Science all over the world, is 
 popular without lowering the Standard of Science, and by 
 it a vast amount of information is brought within a small 
 compass, and students are directed to the best sources for 
 what they need. The various questions connected with 
 Science teaching in schools are also fully discussed, and the 
 best methods of teaching are indicated. 
 
 MACMILLAN AND CO., LONDON.