J. W. H. Harrison and L. Doncaster 239 



The second spermatocyte division resembles the first very closely, 

 except that the chromosomes are half the size ; 56 can easily be counted, 

 of which four are larger than the others (fig. 18). 



Some follicles of both zonaria and hirtaria have abnormal spermato- 

 cyte divisions, leading to spermatozoa without nuclei, as in Pygaera, 

 Abraxas, etc. 



(6) Oogenesis. My zonaria pupae were too old to give satisfactory 

 observations on the oogenesis. The ovaries were already large, with 

 eggs at the lower end of the tubes in which a considerable amount of 

 yolk had been deposited, so that the tubes were becoming moniliform. 

 I have found no diploid mitotic figures in which the chi-omosomes can 

 be counted accurately. The younger oocytes at the top of the tube had 

 already undergone synapsis, and were in the stage with the chromosomes 

 aiTanged under the nuclear membrane. Completely accurate counts of 

 this stage are scarcely possible when the chromosomes are numerous ; 

 it can only be .said that there are between .50 and 60 small double 

 chromosomes, and a composite chromatin-nucleolus of which the two 

 largest portions are almost always of recognisably unequal size. 



3. Chromosomes of the Hybrids. 



(t() Zonaria % x hirtaria ^. This cross gives only male offspring. 

 My material consists of testes of two lai'vae shortly before pupation, 

 and of one pupa about three weeks old. The larval testes contain no 

 divisions later than the spermatogonia ; the pupal testis has also first 

 and second spermatocyte divisions, and contains spermatids in an 

 advanced stage of development towards spermatozoa. 



The spermatogonia! equatorial plates show at a glance two kinds of 

 chromosomes — comparatively few large ones intermingled with a nnich 

 larger number of small ones. 



A count of a very good figure (fig. 19) gives 5.5 — 57 small and 14 

 large ; although this cannot be i-egarded as absolutely accurate, the 

 error certainly does not amount to more than two or three small 

 chromosomes at most. The theoretical expectation is 70, so that it 

 may be assumed that complete iiaploid sets of hirtaria and zonaria 

 chromosomes are present. 



The spermatogonial divisions are succeeded by a stage in which a 

 thin spireme is contracted to one side of the nucleus, after which 

 the thread thickens somewhat and becomes reticular, but the typical 

 "bouquet" stage (pachynema), which is found in pure hirtaria and less 



