MUTATIONS IN CRUSTACEA OF THE FAMILY ATYID.E. 793 



coast (St. Denis), where the experiments were carried on, it 

 was impossible to keep the animals alive in small aquaria, but 

 after several failures Bordage succeeded in keeping living 

 specimens in a small tank of masonry through which a current 

 of water from the town supply was kept flowing. The inflow 

 and outflow were guarded by fine wire gauze covered with 

 musliu to prevent the escape of adults or larvte, or the acci- 

 dental introduction of additional specimens. A single 

 ovigerous female of the Ortmannia form was placed in the 

 tank, and in a few days numerous zoea larvjB were observed 

 in the water. Only seven individuals survived to assume the 

 perfect form a fortnight later, and these proved to be all, like 

 the parent, of the Ortmannia-type. A second experiment, 

 however, was more successful. Another ovigerous Ort- 

 mannia was placed in the tank (which had been emptied and 

 cleaned out between the experiments) and the larvae were 

 hatched in due course. When they were about to pass into 

 the final stage of their metamorphosis some weeks of torrential 

 rain rendered the water-supply muddy and opaque, so that the 

 young prawns were lost sight of. On cleaning out the tank, 

 however, sixteen specimens were discovered among the mud, 

 and of these ten were like the parent, while six were of the 

 Atya-type. Bordage assures us that the precautions he 

 took absolutely exclude the possibility of these young 

 prawns having come from any source other than the eggs 

 carried by the original female. In another experiment two 

 females of the Atya-type produced twenty-seven young, 

 all of which resembled the parents. Bordage states that he 

 was unable to obtain fecundation of Ortmannia females 

 by Atya males, while they bred readily with males of their 

 own type. 



These results are somewhat surprising, and can hardly be 

 accepted as final without a good deal more experimental 

 evidence. If the two forms do not interbreed, and if, as 

 Bordage considers probable, the Atya -form always breeds 

 true, it is evident that the Ortmannia-form would disappear 

 (in the absence of a selective death-rate operating in its 



