TIIE EMBRYO-SAC IN ASTGIOSPERMS. 539 



But as we ascend from the lower Cryptogams to the Coniferae, 

 it is remarkable that the male prothallus always suffers the 

 greatest reduction ; and even in the Coniferae we find a few cells 

 in the pollen-grain, while a relatively large female prothallus (en- 

 dosperm) forms and bears archegonia (corpuscula &c.) in the em 

 bryo-sac. This being so, it is a fair inference that in Angiosperms 

 we ought to find the macrospore, if it exists, producing a relatively 

 larger structure than the microspore (pollen-grain). But ifc is 

 now well known that two or more nuclei occur in many Angio- 

 spermous pollen-grains as first shown by Strasburger*; and if, as 

 is highly probable, we have here a rudiment of division in tho 

 microspore only differing in degree from the formation of a pro- 

 thallus in others, we may expect to find not fewer, and probably 

 more, divisions in the macrospore. 



These considerations would urge us to reject the idea that the 

 nuclei in the embryo-sac are homologous with pollen-grains, and 

 are therefore spores, while they might encourage the theory that 

 we have in the repeated bipartition of the nuclei in the embryo- 

 sac a rudimentary prothallus-formation, as Strasburger believes. 



And there is another, to my mind, considerable objection against 

 the acceptance of the spore nature of the nuclei. If we regard 

 them as together representing a prothallial structure, we may look 

 upon the " egg-cell " as the equivalent of the " oosphere " of 

 vascular Cryptogams ; but if not, we must imagine a process of 

 reduction carried past the point where one might suppose every 

 thing had been removed but the essentials — not only the pro- 

 thallus reduced to an " oosphere/' but even beyond. This appears 

 by no means easy to conceive, and, taken into consideration along 

 with what has been said above, carries some weight. 



Let us see what follows by assuming, as a point de depart, the 

 homology of the " egg-cell " with the " oosphere " of Cryptogams 

 (the H corpusculum " of Coniferae). The sister nuclei must then be 

 cells of the prothallus ; and, from the position of the * synergidse," 

 it is perhaps not impossible to regard them as two neck-cells of 

 the archegonium, which have persisted in virtue of their acquired 

 use in fertilization. The fourth nucleus falls into the sac, pass- 

 ing along the protoplasm to fuse with a similar one from 

 below in typical cases. The " antipodal " cells and the ascend- 

 ing nucleus must be regarded as cells of a prothallus; and it 

 depends upon the nature of the vacuole-like space which sepa- 



* Strasburger, ' Befruchtung und Zclltheilung ;' and Elfving, 'Die Pollen- 

 korner/ &c. . 



