FROM THE USUAL STRUCTURE OF SEEDS. 305

certain relation to the central cavity where the embryo is afterwards formed, and which, filled with a glairy ilnid, is distinctly visible before the separation of the seed. It is a curious consequence of this tardy evolution of the embryo, which in some cases does not become visible unless the [no seed be placed in a situation favorable to germination, that very different directions may be given to its radicular extre- mity, according to circumstances which we have it in our power to regulate.

There is a fourth kind of anomaly in the structure of certain seeds, which, as I have formerly described it,^ 1 shall here notice in a few words. It is that wdiich takes place in certain Aroidcce, especially in some species of Caladiiim. In these, the nucleus of the seed is not pro- perly a monocotyledonous embryo, but has an appearance {uul economy more nearly resembling those of the tuber of a root ; for, instead of being distinguishable into a cotyledon, a plumula and radicula, and of germinating in a determi- nate manner and from a single point, it is composed of a mass wdiose internal structure is uniform, and on the sur- face of which frequently more than one germinating point is observable.

None of these anomalies appear to me materially to lessen the importance of the characters derived from the seeds of plants ; but they evidently render a minute atten- tion to every circumstance absolutely necessary in all attempts either to deduce affinities or establish genera from this source ; and they especially demonstrate the neces- sity of carefully ascertaining the state of the unimpreg- natcd ovarium; for, wdiile its structure remains unknown, that of the ripe fruit can never be thoroughly understood.

1 Prodr. Flor. Nov. Holl. p. 335.

�� �