Pollen Mother -cells of Lilium canadense % 237 
between the first appearance of double strands and that of a four-parted 
spirem a considerable time evidently, from Brauer’s figures and descriptions, 
elapses ; and it is at least conceivable that the occurrence of a double 
thread at the very beginning of synapsis is the same phenomenon that 
I have observed at the same stage in the lily, and that in Ascaris also these 
two threads actually fuse into one, which soon after is doubly split. Miss 
Sargant (’96, ’97) and Farmer and Moore (’95) also describe two rows of 
* chromatin dots ’ in the nuclear threads of the lily previous to synapsis ; 
but, as I have pointed out, there is some doubt as to the real nature of 
these bodies. 
Sabaschnikofif (’97) attempts to harmonize the occurrences in the 
oogenesis of Ascaris with Weismann’s notion of a reduction division. He 
thinks that, during synapsis and previous to the formation of a spirem, the 
chromatin granules become arranged into groups of four, not by a double 
fission of single granules, but by the approximation of originally separate 
bodies. The tetrad groups are connected with one another by fine linin 
fibres ; they approach each other and become arranged to form a spirem, 
which is thus composed from the start of four threads. Sabaschnikofif’s 
view is not very different from that of a conjugation of separate spirem 
threads, or at least of their constituent chromatin granules ; but it involves 
no notion of a fusion ; the grouping of the granules is simply preparatory 
to their ultimate distribution into different germ nuclei. 
Winiwarter (’00) seems to have been the first actually to observe 
a fusion of parallel threads in the prophases of the heterotypic division. 
He finds this process to occur during synapsis in the oogenesis of the 
rabbit and of man, and his figures are very similar to mine. The result 
of the fusion is a comparatively thick moniliform thread (whether con- 
tinuous or not at this period he does not determine), which later splits 
longitudinally. A complete fusion in pairs of all the chromomeres seems 
not to occur ; the double nature of the thread remains always apparent 
in places. Winiwarter considers the longitudinal splitting to involve 
merely a separation of the threads which previously fused. Schoenfeld (’01) 
finds in bovine spermatogenesis a similar fusion of slender threads in 
pairs. 
In a recent preliminary paper, A. and K. E. Schreiner (’04) describe 
a like series of processes in the spermatogenetic divisions of Myxine and 
Spinax. According to these authors, the material of the daughter chromo- 
somes of the last spermatogonial division becomes distributed upon fine 
linin fibres ; there is no continuous spirem, but in synapsis the strands 
aggregate in that half of the nuclear cavity nearest the centrosphere ; they 
lie parallel to each other, or converge slightly toward the pole. The 
strands now approach each other and gradually fuse in pairs. Somewhat 
later a longitudinal splitting appears, which the authors seem to consider 
