DAVIS: SPERMATOGENESIS. 124 
others, which are also probably paired, although the difference in size 
is slight. In the gastropods we have the observations of Bonnevie 
(:05, :06), who found in the odgonia thirty-four autosomes, which 
can be divided according to their size into three groups, each group 
containing an equal number. She also observed slight variations in 
the size of the elements within each group. ‘Turning to the vertebrates, 
we find that similar results have been arrived at by Montgomery 
(:04) and by A. und K. E. Schreiner (:04, :05, :07). According to 
Montgomery the spermatogonial autosomes of several salamanders 
can be grouped in pairs of like volume and form, the components of 
each pair usually lying close together in the spindle. Similarly the 
Schreiners have shown that in the spermatogonia of Myxine and 
Spinax certain autosomes can be recognized by their size, and that 
in such cases there are always two of equal volume. ‘They also con- 
firm Montgomery’s observations that the components of each pair 
usually lie close together. . 
My own results in the Orthoptera are in perfect accord with those 
of the other observers cited. 
In the light of the results detailed above there would seem to be little 
room for doubting that in the germ cells of all animals which reproduce 
sexually, there is a double series of autosomes, one being of paternal 
the other of maternal origin; but such a series is recognizable only 
where there are considerable differences in the volumes of the elements. 
The further fact, first noted by Montgomery (:04), that the components 
of each pair are usually closely associated, is significant as indicat- 
ing a physiological relationship between the two. Montgomery has 
sought to explain this association on the supposition that during the 
prophases the components of each pair lie close together in a continu- 
ous linin spireme. ‘This is, of course, based on his assumption (:00) 
that the chromatin and linin form a single element of which the chro- 
mosomes are simply subdivisions. It is not my purpose to point out 
here the many, and as I believe fatal objections to this view, but to 
consider only its bearing on the association of homologous chromo- 
somes. If there is any such fixed organization of the nucleus as Mont- 
gomery imagines, it would seem that the elements of each pair ought 
always to lie close together, but this is certainly not the case. While 
it is undoubtedly true that such an association is the rule, yet excep- 
tions are common and the members may be widely separated on the 
spindle. ‘This is especially well shown in Steiroxys where the elements 
of the largest pair can always be recognized at a glance. In this 
case the large autosomes usually lie close together, but may be as far 
