CHROMOSOMES OF MOSQUITO 553 



suggestions of it in some of the figures in other works upon 

 Diptera, but it is nowhere else shown that the chromosomes 

 begin their division only after assuming this arrangement. In 

 diploid complexes of Drosophila ampelophila, Bridges ('16) has 

 shown a splitting at the ends of the chromosomes. The spindle 

 fiber is median, as in Culex, but the spermatogonial and ovogonial 

 divisions begin by a split at the ends of the chromosomes and 

 "the separation at the point of attachment does not take place 

 until the elongation of the cell." If we may judge from the few 

 figures given, the chromosomes which are splitting appear to 

 take on a parallel arrangement. In spermatogonial divisions of 

 Culex this split at the ends does not occur, but it does occur 

 occasionally in the somatic cells. The division of the somatic 

 chromosomes, however, seems to be from the middles, compar- 

 able with the divisions of spermatogonia and ovogonia. These 

 matters are to be discussed more extensively in a paper by Mr. 

 Hance. 



It has been noticed that, in well fixed material, the condensed 

 chromosomes act as if they repelled each other. In second 

 spermatocyte late prophases (figs. 82, 83, 84, and 85,) and 

 metaphases (fig. 86) the ends of the chromosomes are far apart, 

 while their middles appear as if tied together by the spindle 

 fibers. This condition may also be seen in the tendency of the 

 outer members of the pairs in diploid metaphases to bend away 

 from the inner members at their ends while their middles are 

 more approximated, as if pulled together or toward the center 

 of the metaphase plate by the spindle fibers. In spermatogonia 

 (figs. 11, 12, 13 and 14), this is striking, and is even more so in 

 the actual preparations, in which differences of vertical focus 

 are evident. 



As the chromosomes condense and stiffen, the tendency is for 

 the pairs to assume a parallel arrangement apparently under the 

 influence from the tension of the spindle fibers. Figure 14 looks 

 very much as if pair a would be pulled about beneath b and c 

 and come up between the two and parallel to them. This ap- 

 pears to be taking place in figure 15, in which I have lettered the 

 pairs in a comparable manner. The result would be a parallel 



