372 University of California Publications in Zoology [Vol. 16 



(1907) changed his opinion and stated that there were sixteen. Kirk- 

 ham (1907a, b) counted twelve. Long and Mark (1911) determined 

 the number to be twenty, thus agreeing with Tafani. The above 

 numbers were all obtained in the stiidy of the egg in maturation, and 

 thus represent the haploid number of chromosomes in the female. No 

 one has counted the chromosomes in the male, with the exception of 

 Loukianow (1898), who stated the number to be twelve in the primary 

 spermatocyte, with six double chromosomes in the secondary spermato- 

 cj^te. His figures, however, are small and poorly represent the shape 

 of the chromosomes, so that they would lead one to think that his 

 technique was deficient and his results inaccurate. 



The primary object in beginning this study of spermatogenesis in 

 the mouse was the determination of the number of chromasomes in 

 the spermatocytes and their correlation with those found in the egg. 

 This correlation in size, shape, and number of chromosomes, although 

 found in insects, has not been heretofore described for mammals. 

 Such a relation is verj' important, since our ideas of the mechanism of 

 heredity and the determination of sex are based largely upon evidence 

 obtained from studies of the chromosomes of germ cells. This paper 

 covers only a small part of the process of spermatogenesis, being 

 confined mainly to the chromosomes of the primary and secondary 

 spermatocytes. 



This work was undertaken under the direction of Dr. J. A. Long, 

 to -sx'liom my thanks are due for his many helpful suggestions and 

 criticisms. 



METHODS 



The mice used were active, healthy individuals. They were of 

 different ages, some being adult and some three, four, and five weeks 

 old. The testes were taken from freshly chloroformed animals, cut 

 into two or three pieces and put into the killing fluids. Several killing 

 agents were employed: namely, Zenker's, Carnoy's, and Fleming's 

 strong solution, and Benda's modification of the latter. While any 

 of the above gave fairly good results, Carnoy's and Flemming's were 

 found most satisfactory. 



Sections six to twelve microns thick were stained in various ways. 

 A few were treated in Heidenhain's iron-alum haematoxylin, others 

 with Mallory's phosphotungstic-acid haematoxylin, wliile the majority 

 were stained with an alcoholic iron-alum haematin with or without a 

 counter-stain. 



