125 
Primitive A ngio sperms. 
ovaries ; stalked stigmas and sessile ones. But all these forms occur in- 
differently among Monocotyledons and Dicotyledons. The natural infer- 
ence from the facts is that carpels were derived by inheritance from an 
ancestral race common to both classes. 
The improbability of a multiple origin is of course very greatly 
increased if Angiosperms are derived from more than two non-Angio- 
spermous races ; that is, if we suppose carpels to have been evolved in three 
or more parallel lines of descent. 
A similar argument applies to two other characters, which are most 
conveniently treated together, the germination of the embryo-sac, and the 
formation of the endosperm in Angiosperms. The strength of this argu- 
ment appears very great when it is considered (i) that the first nuclear 
divisions within the embryo-sac are highly characteristic of Angiosperms. 
In other words, they differ in a marked way from the first steps in the 
germination of any other megaspore. (2) That the origin of the endo- 
sperm from a fusion of three nuclei is still more characteristic, and that its 
morphological value is still unknown. (3) That both processes, though 
complicated and therefore offering many possibilities of variation, are 
remarkably uniform throughout the group. The very few exceptions 
recorded belong for the most part to parthenogenetic or apogamous species, 
or to forms clearly degraded by parasitism or similar causes. Such forms 
can hardly be regarded as primitive. 
These considerations point to a very long period during which the 
Angiosperms as we know them now were evolved — a period long enough to 
stamp such well-defined characters on the race, and also to allow of the 
disappearance of the numerous intermediate forms which must have 
existed once. 
The discovery of the origin of the endosperm from a fusion of three 
nuclei is quite recent, and as the literature of the subject is scattered 
we may briefly discuss the strength of the evidence. Up to 1898 the 
primary endosperm nucleus was believed to arise from the fusion of the two 
polar nuclei in the embryo-sac (Strasburger, 84 ). In that year Nawaschin 
( 61 ) showed that a third nucleus takes part in this fusion. He was working 
on Lilium Mar tag on and Fritillaria tenella, and in these plants he identified 
the third nucleus as one of the two generative nuclei derived from the 
pollen-tube. The other, of course, had fertilized the ovum. It was known 
that two generative cells entered the embryo-sac, but it had been previously 
believed that when the nucleus of one had fused with the nucleus of the 
ovum the second generative cell broke down like the synergids. 
Nawaschin’s discovery was soon confirmed by the publication of 
Guignard’s independent researches ( 32 ). Since 1899 a literature has grown 
up on this subject. The most recent resumes are those of Guerin ( 30 ) 
