820 
Proceedings of the Royal Society 
within the hollow of the nutritive hlastophore (fig. F'). This cannot, 
indeed, he directly collated with any actual form of ovum segmen- 
tation, but rather with the inversion of such a form of gastrula as is 
seen ( e.g .) in Teleostean ova, where the yolk protrudes in hernioid 
fashion from the endoderm cells of the epibolic gastrula (fig. F). 
Starting from the uninvaginated form, we should thus have in these 
two types of sperm and ovum segmentation what might be regarded 
as the result of invagination from different sides, — in the former the 
nutritive, normally internal portion becoming the layer enveloping 
the formative cells, — -just the reverse of what occurs in the latter. 
Of this inverted gastrulation, possibly represented in Elasmobranchs, 
traces may be detected in such a case as that described by Renson 
in mammals, where the sperms produced alongside of the large nutri- 
tive cells yet find their way into them, and sinking in are again 
borne up and finally set free (fig. F", “obgastrula” type). 
In regard to such a comparison, which appears to us a possible 
method of reconciling, without discrediting, the discrepant observa- 
tions of competent authorities, and of rationalising the various 
methods of spermatogenesis, by comparing them with parallel 
processes in ovum-segmentation, the writers do not overlook that 
such a theory must wait for absolute verification till more data are 
available as to the behaviour of the nuclei in both cases, especially 
in spermatogenesis, for without this a real similarity of process can 
only be inferred from the likeness of the result. It must be noted 
also that the theory in no wmy falls with the failure of any parti- 
cular instance. Further, if it be true that the multitudinous 
details of spermatogenesis can be morphologically rationalised by 
collating them with the details of ovum-segmentation, the physio- 
logical problem remains of interpreting both in terms of that differ- 
ence in protoplasmic metabolism on which sex must finally depend. 
In a subsequent paper* by one of us, these sex differences are 
traced to a preponderance of anabolism in the female and kata- 
bolism in the male, and if this conception be applied to the pre- 
ceding, the physiological rationale of the morphological process may 
become no longer wholly unintelligible. 
* Geddes, “Theory of Growth, Reproduction, Sex, and Heredity,” Proc. 
Roy. Soc. Ed/in., 1886. 
