368 The Spermatogenesis of Bufo Lentiginosus 
chromosomes, although Rickert considers that the paired arrangement 
of the chromosomes may be due to “eine eigentumliche Langsspaltung 
der Chromosomen im Dyaster der letzten Teilung des Ureies.” More 
conclusive evidence is given by Calkin’s (8) work on Lumbricus. In 
this form the spireme first splits longitudinally and then segments into 
the somatic number of segments, 32. ‘These double segments then 
conjugate in pairs forming 16 tetrad groups. he first maturation 
division is a reduction division, separating the pairs of chromosomes 
that conjugated to form the tetrads. More recently Steven’s (55) in- 
vestigations on the germ-cells of Aphids have shown that there is in these 
forms undoubtedly a pairing of the chromosomes previous to the first 
maturation division. If a more extended investigation of the spermato- 
cytes of the Urodela should make it seem probable that Janssen’s inter- 
pretation of synapsis in Batracoseps can be extended to other species, 
then univalent chromosomes will be separated in the first mitosis in all 
such cases, and the maturation divisions in this group of amphibians 
will readily fall in line with those of all other accurately known forms. 
In studying the spermatogenesis of Bufo I have been fortunate enough 
to find practically all stages in the development of the primary sperm- 
atocytes from the time that the cell is formed (Fig. 10) until the com- 
pletion of the first maturation division (Fig. 53); and I have traced, as 
earefully as possible, the complex changes taking place in the chromatin 
at this time. In the young spermatocyte shown in Fig. 10, the chro- 
mosomes are distinct and connected by fine linin fibres. At this stage 
the cells are very small and many of the chromosomes are crowded 
against the nuclear wall so that I have not been able to make out their 
number satisfactorily. As there are certainly many more than 12 
chromosomes in the nucleus at this time, I am very sure that pseudo- 
reduction has not yet taken place and that the somatic number 
of chromosomes, (24) is present. As shown in Fig. 10, the chromo- 
somes in many cases appear to be connected end to end, especially is 
this noticeable in the chromosomes that lie against the nuclear wall. 
In the later development of the spermatocytes the chromatin substance 
becomes distributed on the nuclear reticulum and all traces of the in- 
dividual chromosomes is lost. If, however, an end to end union of the 
chromosomes is established at or before the stage of Fig. 10, it is safe 
to assume that this connection is not broken during the later growth 
stages of the spermatocytes. 
I have never found any evidence of the longitudinal splitting of the 
spireme in the young spermatocytes that has been found at the same 
