3/O THOMAS H. MONTGOMKRY. 



nuclei in one cell body, indicating nuclear di\ IMOII without .cyto- 

 plasmic division of antepenultimate spermatogonia: in .ill MX 

 of these cases only one rod was present, and an example i-> -ho\\ n 

 in Fig. 9. There can thus be no doubt that hall" tin- penultimate 

 spermatogonia contain rods and half do not. 



Characteristic appearances of the rods are illustrated in Figs. 

 9-11, PI. I., 14^16, PI. II. They differ on the average from those 

 of the preceding generation in being usually larger, straighter, 

 and more deeply-staining with haematoxyline, which indicates 

 they have been undergoing growth changes. 1 Not infrequently 

 they are curved around the nuclear surface (Fig. 14) and tin- 

 length of a rod may equal the diameter of a nucleus. Conse- 

 quently they are in this stage very prominent constituents of the 

 cell bodies and easily differentiated by safranine or haema- 

 toxyline. 



Mitoses of these cells were not frequent, but two clear cases 

 (PI. II., Figs. 12, 13) were found, showing that the rod (R.) 

 passes undivided into one of the daughter cells, and this is fully 

 borne out by a study of their distribution in cells of the following 

 generation. Fig. 18 shows the end result of such a mitosis in a 

 case where the cell body had not divided and here there is but 

 one rod. \Yhat is the nature of the scattered globules shown in 

 Figs. 12 and 13 is doubtful; they may be discharged nucleolar 

 material. 



4. Tiii': ULTIMATE SPERMATOGONIA (Fios. 17-24, PL. II.). 



These are the smallest of the spermatogonia and tin- most 

 numerous in the testis studied. One quarter of them contain 

 each a rod; three quarters lack rods. One hundred and forty- 

 two of these cells were studied, at stages before any of them had 

 enlarged into Sertoli cells, the precaution being taken to include 

 only cells lying wholly within the section; of these twenty-fix r 

 showed each one rod, and one hundred and seventeen showed no 

 rods. This ratio is somewhat less than I : 3, which is readily 

 explained on the ground that some of the spermatogonia \\ith 

 rods had already become Sertoli cells and then-lore were not 

 included in the count. A very important and clear case i- that 



'The condition of the pair of rodlcts (r. 2) in Fig. 10 will l><- .\plainrd luti-r. 



