374 



NATURE 



[August 17, 1882 



and from them we can trace a gradual decline, through 

 plants like dog's mercury {Alercurialis perennis), which 

 has a green calyx, but no corolla (Figs. 31 and 32), to 

 very degenerate green blossoms like our own spurges 

 {Euphorbia), which consist of several extremely simplified 

 flowers, collected together in a common involucre. Each 

 separate male floret is here reduced to a single stamen, 

 raised on a short peduncle, and with a distinct joint at 

 the spot where the petals once stood (Fig. 33). It is 

 worthy of notice, too, that when these degenerate, but 

 still entomophilous, green flowers have found it desirable 

 to attract insects 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 or involucres instead, as in the well-known Jatro- 

 phas and Poinsettias of our hot- houses. This instance 

 is exactly analogous to that of the Sanguisorba. 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 Monochlamydce. In 

 the Paronychiacece, for example, we get an order closely 

 allied to the Caryophyllacece (especially to Polycarpon) ; 

 and in one genus (Corrigiola) the flowers have small 

 white petals, which certainly aid in attracting insects. 

 But in Herniaria the flowers are quite green, and the 

 petals are reduced to five small filaments, thus partially 

 reverting to their presumed original character as sta- 

 mens. In Sclercwthus the filaments are often wanting, 

 and in some exotic species altogether so. The Amaran- 

 tacea, unrepresented in Britain, approach the last-named 

 family very nearly, but have the petals altogether obso- 

 lete ; and in many cases, such as Prince's feather 

 {Amaranthus hypochondriasis) and Love-lies-bleeding 

 (A. caudal 'us), the calyx becomes scarious and brightly 

 coloured. The Chenopodiacea are other near relations, 

 in which also the petals are quite obsolete ; and in most 

 of them the perianth (or calyx) is green. In Sali- 

 cornia it has become so embedded in the succulent leaf- 

 less stem as to be almost indistinguishable. The Poly- 

 gonacece , on the other hand, are a group of plants, allied 

 to Chenopodiacea', but with a row of degraded petals, and 

 a strong tendency to produce coloured perianths, analo- 

 gous to that which we observed in Sanguisorba. The 

 flowers of Rumex, the docks, are sometimes green, some- 

 times red ; those of Polygonum are pale-green, white, or 

 pink. Rumex is sometimes, Polygonum constantly, fer- 

 tilised by insects. 



There remain doubtful, then, among green Dicoty- 

 ledons, only the highly anemophilous families, like the 

 nettles {Urticacea), and the catkin-bearing trees (Amen- 

 ti/era). The former have a well-developed calyx, at least to 

 the male flowers (Fig. 34), and it is difficult to see how any 

 one who compares them with Scleranthus or Mercurialis, 

 known descendants of petaliferous forms, can doubt that 

 they too are degenerate types. Indeed, the mere fact 

 that the stamens are opposite to the lobes of the calyx, 

 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. 



As to the Amentifera, Cupuliferce, and other catkin- 

 bearers, at first sight we might suppose them to be primi- 

 tive green anemophilous orders. But on closer considera- 

 tion, we may see grounds for believing that they are really 

 degenerate descendants of entomophilous plants. In the 

 alder {Alnus) the male catkins consist of clustered 

 flowers, three together under a bract, each contain- 

 ing a four-lobed perianth, with four stamens within. 

 These little florets exactly 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 of a corolla. In the beech {Betula) the three 

 florets under each bract are loosely and irregularly ar- 

 ranged ; and in the male hornbeam (Carpinus) and hazel 



(Cotylus) the perianth is wholly obsolete. All these are 

 probably quite anemophilous. The willows (Salix), on 

 the other hand, have become once more entomophilous 

 (Figs. 35 and 36); and they are much visited by bees, 

 which obtain honey from the small glands between the 

 florets and the axis. Degenerate as these last-named 

 species undoubtedly are, they may be connected by a 

 regular line of illustrative examples (not genetically) 

 through the beech, alder, nettle, goosefoot, Scleranthus, 

 Herniaria, and Corrigiola, with such perfect petaliferous 

 types as the pinks, and ultimately the buttercups. 



Among Monocotyledons, the very degraded little ento- 

 mophilous flowers of the Arum (Fig. 37), enclosed in 

 their green spat he, are often spoken of as though they re- 

 presented a primitive type. In reality, however, they are 

 degenerate dichlamydeous blossoms, linked to the lilies 

 by Acorus (Fig. 38), which has numerous hermaphrodite 

 flowers, each with a perianth of six scales, two rows of 

 stamens, and a two-celled or three-celled ovary. Here, 

 again, the green flower is obviously of late date. 



What, then, are we to say about the anemophilous 

 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 knowledge ; yet 

 we may see some reason for believing that even these 

 highly anemophilous types are degenerate descendants 

 of showy petaliferous blossoms. For, if the origin here 

 assigned to petals be correct, it becomes clear that the 

 fitneacca, or rushes, are only Liliacece in which the 

 perianth has become dry and scarious. Some rushes, 

 such as Luzula, approach very closely in general cha- 

 racter 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. Eriocaulon and the 

 Restiaceee give us a further step towards the grass-like or 

 sedge-like character. Some of the Cyperaeew show 

 apparent relics of a perianth in the bristles which sur- 

 round the ovary, especially in Scirpus (Fig. 39) ; and per- 

 haps the perigynium of Carex may represent a tubular 

 perianth, though this is far more doubtful. In the grasses 

 [Gramineee) the perianth is either altogether obsolete, or 

 else is reduced to the pales with the hypogynous scales 

 or lodicules (Fig. 40). According to the most probable 

 view, 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 the lodicules 

 represent two of the petals, the third (the inner one) being 

 usually obsolete (Fig. 41). It is fully developed, however, 

 in the bamboo. The connection is here less clearly trace- 

 able than in the Amentifera, but it is still quite distinct 

 enough to suggest at least the possibility that even 

 grasses and sedges are ultimately derived from ento- 

 mophilous flowers. 



Thus we are led, at last, to the somewhat unexpected 

 conclusion that anemophilous angiosperms are later in 

 development than entomophilous angiosperms, and are 

 derived from them. Though the earliest flowering plants 

 — the pines, cycads, and other gymnosperms — were un- 

 doubtedly anemophilous from the first, yet the probi- 

 bility 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 im- 

 possible properly to inquire here. We must content our- 

 selves with noting that such degraded green flowers fall 

 for the most part under one or other of four heads : (1) 

 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 



