Spermatogenesis in spiders. 551 



three very small bodies lag-giug behind in the spindle (Fig. 101). 

 The ultimate distribution of these bodies would seem to be a matter 

 of chance. 



Number of chromosomes. 



Flat equatorial plates of both the first and second division 

 spindles were counted (Figs. 99 and 100) and nine is the haploid 

 number of autosomes. The chromosome constitution of the specimen 

 studied would be then: 



Autosomes 18 



Accessory chromosome 2 



Total 20 not including the planosomes. 



Spermatogenesis of Agalena naevla. 



No description will be given of the course of the chromosomes 

 in this species, as Wallace has already done this in her admirable 

 paper. It seems well to call attention to one feature of Wallace's 

 paper: namely, her estimate of the number of chromosomes. She 

 estimated that the haploid number was at least twenty-five, although 

 in some cases she found more. I have not been able to make accu- 

 rate counts in this form but from the preparations examined, it 

 seems certain that Wallace over-estimated the number. It seems 

 probable that it is more nearly fifteen. 



Spermatogenesis of Dolomedes fontanus. 



The Pisauridae are represented in my preparations by Dolomedes 

 fontanus and, in many respects, this form has proved the most in- 

 structive of any of the spiders studied. One is able here to follow 

 the course of the autosomes after the synaptic contraction with 

 almost schematic clearness. Observations have been made on only 

 one specimen, which was preserved in Peteunkevitch's sublimate 

 and stained in iron-haematoxylin. 



The synaptic stages do not prove illuminating in this form, so 

 it is not possible to say what goes on at this time. Before the 

 first contraction we have the thin leptotene-threads and afterwards 

 the thick pachytene-threads. There has also been an obvious reduc- 

 tion in the number of the threads, so that there can be little doubt 

 that a pairing has taken place. But whether this pairing was end 



