130 CAROLINE BURLING THOMPSON 



observations upon the structure of the other nymphs, but he 

 states his behef that all termite castes are predetermined at the 

 time of hatching. I venture to assert that, even if the future 

 worker nymphs of Eutermes lacustris are not externally dis- 

 tinguishable from the reproductive nymphs at the time of hatch- 

 ing, the two forms have internal characters which are different. 



It is evident that the various castes manifest themselves at 

 different ages in different genera, and it is probable that there 

 are differences in this respect within a genus. For example, 

 Knower ('94) found that the soldier of Eutermes (rippertii?) 

 originates late in development by molting from an antecedent 

 worker-like form; whereas Bugnion ('12) could distinguish the 

 soldier of Eutermes lacustris at the time of hatching. 



In this connection it is interesting to recall the observation of 

 Bobe-Moreau ('41) regarding a tiny soldier of the European 

 L. lucifugus, which he described as just emerged from the egg. 

 It is probable that Bobe-Moreau mistook one of the small pale 

 newly molted soldier nymphs for a newly hatched form, since 

 Lespes observed the young soldier nymphs of L. lucifugus emerge 

 from a worker-like skin. But if the soldier nymph seen by Bobe- 

 Moreau was actually newly hatched, then the suggestion offered 

 in regard to the genus Eutermes will hold good also for Leuco- 

 termes, namely: that the castes may originate in different ways 

 within the species of a genus. 



My final conclusion is that all termite castes are predetermined 

 in the eggs; that some castes are distinguishable, either by ex- 

 ternal or internal characters, at the time of hatching, while 

 others may appear considerably later in the course of develop- 

 ment. In Leucotermes the more generalized castes are the 

 first to be manifested, the more specialized are later in their 

 appearance. 



It is a possibility that the castes of termites are mutations 

 from the ancestral parent stock; they would therefore, if fertile, 

 breed true. There is great need, however, of observational and 

 experimental proof to establish this question. 



There is also need of cytological study with termites. The 

 intrinsic cause of caste determination may be as readily determin- 



