PHANEROGAMIA : FLOWERS. 423 



ripening of the seed, because I have not been able to pursue the 

 development in those plants to which undivided embryos are 

 usually ascribed (Bertholetia and Lecythis). Wherever I have 

 been able to follow this out, I have found the formation of the 

 cotyledons which I am about to describe ; the genus Cuscuta 

 alone forming an exception to this. In these plants the embryonal 

 globule grows into a longish stem, without any trace of foliar 

 organs except in the (single ?) case of Cuscuta monogyna. In all 

 the remaining cases which I have hitherto been able to investigate, 

 there are formed on the embryonal globule, sometimes leaving free 

 a considerable part of the point in a papillary form, sometimes only 

 a small part of it extending to a few cells, but never occupying 

 the extreme point, two leaves, at first as little lateral papilke, which 

 gradually extend on both sides, embracing between them, with 

 their bases, the point of the embryo as a free bud. This is also 

 considerably developed, and produces sometimes more, sometimes 

 fewer, or occasionally no leaves at all in the embryonal condition. 

 Here, again, the varieties of the perfect embryo depend upon the 

 varied ulterior developments of the individual parts thus formed in 

 a rudimentary condition. 



Sometimes the radical end is disproportionately developed, as in the 

 Pekea and Rhizophora ; sometimes the cotyledons; occasionally, but not 

 frequently, one cotyledon alone is greatly increased, whilst the other 

 makes no advance in growth ; this is the case with Trapa natans, in 

 which I have observed, in an early condition, a large papilliform 

 terminal bud, and at the sides two equal-sized cotyledons (?) ; but I 

 have not been able hitherto, notwithstanding every pains, to discover the 

 intermediate stages between this state and that of the ripe seed. 



Of a large series of interesting conditions, which were in great part 

 observed by Bernhardi (Linnsea, Bd. vii. 572.) in germinating plants, 

 we have unfortunately no history of the development of the embryo. 

 All that is commonly said on that subject is merely idle, useless specula- 

 tion, more calculated to mislead than to enlighten. The cotyledons may 

 be blended, as often occurs; and cotyledons originally equal may be sub- 

 sequently unequally developed. Future minute researches can alone 

 solve the problem here. 



C. DEVELOPMENT OF THE GERMEN AND SEED-BUD TO FRUIT 

 AND SEED. 



170. During the development of the embryo, cellular tissue, if 

 not already existing, is always formed in the embryo-sac, always 

 growing from the walls as well as from the surface of the nascent 

 embryo into the cavity : this is called endosperm. How far this 

 new cellular formation may be carried, how soon and to what 

 extent it may be again displaced by growth of the embryo, vary 

 extremely ; but they are usually constant in each particular family. 

 Thus a considerable portion of this endosperm is still to be recog- 

 nised in the ripe seed in the Liliacea, Palms, Grammcce^ and Cypc- 



F. 4 



