July 27, 1882] 



NA TURE 



303 



dog-rose, Rosa canina, R. villosa, R. rubiginosa, &c), 

 which in some still bigger exotic species become crimson 

 01 damask of the deepest dye. They are 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 poten- 

 tillas have the very simplest arrangement of the carpels 

 in the whole family, and their fruit is of the most primi- 

 tive 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 they have very erect flowers, and 

 three leaflets instead of five or more to each leaf. One of 

 them, the strawberry, shows still further marks of special 

 differentiation, in that it has acquired a soft, pulpy, red 

 fruit, produced 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 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 modification. Never- 

 theless, 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 

 tike the East Indian P. nepalensis and P. atropurpurea. 



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 Prunea have diverged 

 from the central stock of the rose family in one direction, 

 and the Pomea and Rosea 1 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 Rubea 

 or brambles are for the most part woody shrubs instead 

 of being mere green herb?, and they have almost all 

 ■vhitish 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 deve- 

 loped. In the Prunea, again, the development has gone 

 much further, for here mo = t of the species are trees or 

 hard shrubs, and the number 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 Prunea? 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 (Primus cerasus), 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 (P. 

 communis) has smallish white fl iwers, as has also the 

 Portugal laurel (P. lusitanicus). In these phnts, in fact, 

 higher development has not largely taken the direction of 

 increased attraction for insect fertilisers ; it has mainly 

 concentrated itself upon the fruit, and the devices for its 

 its dispersal by birds or mammals. In the Rosea, on the 

 other hand, though the fruit is less highly modified, the 

 methods for insuring 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 advanced members of the 

 family, and it is here accordingly that we get the highest 

 types of coloration, Most of them are at least pink, and 

 many are deep red or crimson. Among the Pomea we 

 find an intermediate type (as regards the flowers alone) 

 between Rosea and Prunea; 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 exhibits 

 very beautifully the importance as regards coloration 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 generally pink ; and the largest are 

 generally rose-coloured or crimson. 



Even more primitive in type than the Rosacea are the 

 lowest members of the Ranunculacea, 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 

 var)ing from yellow to every shade of crimson, purple, and 

 blue. The simplest and least differentiated members of the 

 group are the common meadow buttercups, forming the 

 genus Ranunculus (Fig. 6), 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 

 primitive in organisation, such as R. acris (the central 

 form), R. bulbosus, R. rcpens, R. flammula, R. sceleratus, 

 R. auricomus, R. philonotis, &c. In the lesser celandine 

 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 coppery, and the flowers fade 

 white, much more distinctly than in most other species 

 of the genus. 



There are two kinds of buttercup in England, however, 

 which show us the transition from yellow to white actually 

 taking place under our very eyes. These are the water- 

 crowfoot, R. aquatilis, and its close ally, the ivy-leaved 

 crowfoot, R. hederaceus, whose petals are still faintly 

 yellow toward the centre, but fade away into primrose 

 and white as they approach the edge (Fig. 7). We have 

 already noticed that new colours usually appear at theout- 

 side, while the claw or base of the petal retains its original 

 hue ; and this law is strikingly illustrated in these two 

 crowfoots. 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. 



In many Ranunculacea there is a great tendency for 

 the sepals to become petaloid, and this peculiarity is very 

 marked in Caltha palusiris, 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 cultivated in our 

 flower-gardens. 



It is among the higher ranunculaceous plants, how- 

 ever, that we get the fullest and richest coloration. Colum- 

 bines (Aquile^ia), are very specialised forms of the but- 

 tercup type (Fig. S). Both sepals and petals are brightly 

 coloured, while the latter organs are produced above into 



