134 LOUISE B. WALLACE. 



workers can be said to favor Montgomery's view that the primary 

 spermatocytic division results in a separation of maternal and 

 paternal chromosomes which have conjugated in the synapsis 

 stage. In my previous work I pointed out that Agalena offers 

 especially favorable material for the investigation of this point 

 and also that my results led to a full endorsement of Mont- 

 gomery's interpretation. Further study has convinced me that 

 in all probability that view is the correct one. Every one who 

 has tried to follow the behavior of chromosomes during synapsis 

 knows the difficulties in the way of reaching any certainty about 

 the matter. What we do know is that, as a rule, before synapsis 

 we find a certain number of chromosomes present and after syn- 

 apsis we find but half that number. Many facts can be advanced 

 in support of the theory of the conjugation of maternal and 

 paternal chromosomes resulting in numerical reduction and if 

 such a pairing of these nuclear elements does occur it may be 

 brought about either by an end to end union of homologous 

 chromosomes or by a union side by side. It is clearly of utmost 

 importance to determine which method of union obtains before 

 endeavoring to interpret what follows. 



In Agalena, in the early prophase of the primary spermatocytes, 

 the nucleus becomes contracted and its structure cannot be fully 

 made out, but the chromatin seems to be in the form of a dense 

 reticulum. At this time I believe the synapsis to occur both 

 because there is no later stage in which there is any sign of its 

 occurrence and also because the delicate spireme loops which 

 issue from this stage appear to be in the reduced number. 

 Strong evidence against the view that a side to side pairing of 

 the chromosomes occurs is the fact that the spireme at first shows 

 no trace of a split, then a barely perceptible one and finally quite 

 a wide one running the whole length of each loop. The space 

 intervening between two threads, therefore, must represent a 

 longitudinal split and not the space between two threads about 

 to unite side by side. 



As evidence in support of the view that the chromosomes unite 

 end to end, I would call attention to the noticeably denser struc- 

 ture of the chromatin at the bend of the loops which is quite 

 marked in some stages, the chromomeres appearing to be massed 



