recapitulation seldom occurs in plants, the development of the whole and of its organs being usually direct. When it does occur, it is generally limited to characters coming from comparatively recent ancestors and not even in the most transitory form entering into the formation of the fundamental parts of the plant.1

1 The rarity of recapitulation among vegetables is the result partly of their fixed condition in the soil, and partly of the more rigid nature of their cells.

The immobility of a plant forces the adult to live in the same place as the embryo. Among animals, on the other hand, it frequently happens that the young pursue a manner of life different from that of the adult and resembling that of the ancestor. Young Cirrepedes are vagrant and have the same needs and use the same organs as other vagrant Crustacea; larval frogs inhabit the water like their fish-like ancestors. In plants there is nothing similar; all the aquatic flowering plants are derived from terrestrial ancestors, but if at the beginning of their existence these aquatic plants were to bear leaves adapted to aerial life they would ensure their own destruction. The exceedingly rare ancestral traits to be found in a few species are naturally of a kind not to incommode their possessors. It is improbable that these are a legacy from distant ancestors; they would not have been spared by natural selection had they not come from ancestors of very much the same habit. The absence of locomotion in plants has also produced a greater adaptability than among animals. Animals, when conditions are unfavourable can remove in search of more suitable localities, plants being fixed in the soil must become modified or perish. Plants, therefore, offer numerous cases of individual adaptation. We do not know if these adaptations are transmitted by heredity, but natural selection has at least secured the widest range of plasticity. Thus plants rapidly rid themselves of ancestral legacies which have become useless.

The transitory organs of animals are employed for the service

N

In consequence, vegetable embryology is of little use for investigation of the supposed backward path of degeneration, for the rudimentary or reduced organs of plants do not generally represent ancestral stages.

The seedling of Lathyrus tenuifolius (fig. 67), a vetch, possesses rudimentary organs which cannot be ancestral stages as their development is direct. In this plant a whole series of leaves are formed between those arising at germination and the adult leaves. This intermediate series displays many arrests of development.

The adult leaf has a pair of stipules, foliage leaflets, and tendrils (fig. 67, J). The leaves just before these, have a pair of stipules (fig. 67, 1), which are absent in the leaves next before (fig. 67, H). Still earlier leaves are produced with fewer leaflets and tendrils (fig. 67, D-G), leaves without leaflets and with a single tendril (fig. 67, c), and leaves entirely without tendrils (fig. 67, B). Lastly, at germination very rudimentary leaves are proof the whole body, the branchial arches of mammals are employed in the formation of important parts of the head and neck. The tail of the tadpole is reabsorbed by phagocytes and its substance used for the nutrition of the body. In the case of plants, such occurrences are rare and limited; the cells are enclosed in a rigid wall which resists displacement or alteration; the protoplasmic contents may be absorbed and used as nutritive material by another part, but the cellulose cell-wall remains. A useless organ can be eliminated only at the expense of loss of material. J. Massart, La Récapitulation et l'Innovation en embryologie végétale (Bull. Soc. Roy. Bot. Belg., t. xxxiii., p. 150, 1894).

A, B, seedings in two stages of growth. C to J, different forms of successive leaves.

duced, the sole function of which is to protect the axillary bud (fig. 67, B, 1, 2). We have now to show that this series is by no means a retracing

of ancestral stages.

All, or nearly all,

W

6

FIG. 68.-Seedling Vicia of monanthos.

the Papilionaceous plants have stipulate leaves; shows that

this

the exstipulate leaves of L. tenuifolius do not re

present an ancestral stage. Moreover, before the acquisition of tendrils, vetches had a terminal leaflet (see Vicia Pyrenaica, fig. 44); none of the reduced leaves in L. tenuifolius reproduce this stage; moreover

the simple leaves without leaflets do not represent an ancestral condition; the winged petiole is not an ancestral character.

The primary leaves of another vetch Vicia monanthos confirm our conclusion. In this case

the first few leaves formed in the seedling remain rudimentary, aud serve only to protect the axillary buds (fig. 68, leaves 1 and 2). Contrasted with the condition in L. tenuifolius, although all the vetches had probably a common ancestor, these primary leaves have three little projections, the two lateral of which are reduced stipules.

We do not know of any vegetable example of recapitulation in the case of an organ reduced by arrest of development. Plants exist, however, which after having produced leaves of typical structure begin to produce leaves the development of which remains arrested. Such a plant is the Acacia which bears phyllodes. At first the leaves are like those of other Acacias; next it bears leaves, the blades of which are rudimentary; and finally leaves with normal stipules, but with no trace of lateral leaflets on the petiole.

In this case the ancestral conditions are known and are quite different.1

2. Organogeny of flowers.-What we have said. about leaves applies to the organogeny of flowers. Here also in the cases of atrophy produced by arrest of development there is no indication of

1 Many other reduced plant organs might be instanced, such as the stipules of Sambucus or the teeth of the calyx, in many Compositæ and Umbelliferæ. After their formation such organs grow very slowly and exhibit no trace of recapitulation. This happens with the leaves of Sempervivum (fig. 49). Without doubt these leaves are derived from leaves normally divided into hypopodium and epipodium, but they show no trace of this division.