CELLS IN DIVISION 



example of the tendency towards 'somatic pairing' in the Diptera 

 (Metz^'^^). In Gerris, homologous chromosomes are usually found in 

 the same part of the nucleus, but there is no tendency towards a closer 

 association (Geitler^'*). No further stages of mitosis are seen, and it is 

 therefore concluded that the reduplicated chromosomes must resume 

 their interphase condition after this modified prophase. Grafl^''^ has 

 described instances of endomitosis in plant tissues ; in mouse tumours, 

 BiESELE et alii^''^ have figured nuclei with an intact membrane, within 

 which are fully formed chromosomes and without nucleoli; such 

 nuclei they regard as endomitotic. Phase counting should reveal 

 whether this is a stage of some duration, as would be expected if the 

 view of these authors is correct, or whether it represents a brief and 

 variable epoch at the end of prophase. 



CooPER^'^' and Painter^'^ consider that the development of the 

 polytene condition of the saHvary gland chromosomes in larval Diptera 

 is to be regarded as an endomitotic process. During the larval period 

 in Culex, the nuclei of the ileum increase in size and their chromosomes 

 are reduplicated, yet remarkably enough, the resting condition is 

 maintained throughout (Berger;^'^ Grell^^'^). At metamorphosis the 

 polyploid condition of these nuclei is revealed; they then enter prophase 

 and undergo a series of 'reduction divisions' in which the number of 

 chromosomes is reduced successively from 48 or 96 to the normal 

 diploid number of six (Berger^^^ ^^^). In the prophase of these mitoses, 

 there is again synapsis of the sister chromonemata. They contract, and 

 separate at metaphase into pairs; there are no chiasmata to be seen 

 during prophase. Berger suggests that the implication of this observa- 

 tion is that 'chiasmata in germ cells are the result of previous crossing 

 over, rather than that crossing over results from chiasma formation'. 



Should it ever prove possible to follow these variants of mitosis in the 

 living nucleus, such observations would undoubtedly be of very great 

 interest. An adequate technique of cultivating insect tissues might 

 make it possible to study the factors which determine the onset of 

 reductional mitoses in the ileum of the larval mosquito at metamor- 

 phosis. It would not be easy to find many problems in experimental cell 

 biology of equal fascination and importance. 



Amitosis 



We pass now to consider the evidence for a form of nuclear division 

 in which the study of the living cell has not yielded so far a uniform 

 answer. Amitosis is known to occur in some cells, but many instances 

 where the 'direct' division of nuclei have been described have subse- 

 quently been doubted. In 1887, Arnold^^^ published a paper on the 

 amitotic division of living leucocytes of the frog. His technique was 

 an interesting one; with aseptic precautions, he inserted into the dorsal 



151 



