MB. G. D. H4.VILAND ON TERMITES. 361 



distinct and readily distinguishable, and in other cases to pass 

 indistinguishabJy into one another. In those groups in which 

 the species are easily distinguishable, I trust that my descriptions 

 will be found sufficient. In the groups in which the species are 

 not easily distinguishable, I have not attempted to outdo oature 

 in distinctness ; indeed, in this respect I am conscious of short- 

 comings. In every case I trust that more reliance will be placed 

 on my specimens than on my descriptions. 



Characters. 



The following brief remarks on the external characters of 

 Termites refer almost entirely to the species of the geuus Termes, 

 for on the other genera I have nothing new to advance. 



The enormous number of individuals in a nest, all of whom 

 maj be considered as the children of the same parents, provides 

 material for the study of normal variation and of specific limits 

 scarcely to be met with elsewhere. The great diflerence of 

 character in the different caster also introduces new conditions 

 in the classification of species, and in the study of heredity not 

 often to be met with. 



In the genus Termes the soldier is by far the best caste to 

 determine species from ; not only is the soldier easier to determine 

 than the male, but it is found in almost every nest, and usually 

 wherever the workers go. Though the imago was the caste on 

 which Hagen founded most of his species, though it is the form 

 found fossil in amber, though it is the form caught flying round 

 a lamp at night, yet it is generally absent from the nests, and is 

 often insufficient for the determination of species. I have nob 

 found the characters of the wings very useful or reliable. In 

 one case I have based species on differences in the imago, though 

 I could see no difference whatever in the soldier ; but as a rule 

 my species are based chiefly on apparent differences in the 

 soldiers. 



There are two external characters which are correlated in the 

 soldiers and the males of the genus Termes: the abdominal 

 papillae show a corresponding degree of development, and the 

 number of segments of the autennae is approximately in the 

 proportion of 8 to 9. The characters of the antennae are probably 

 more important than any others in the determination of the 

 species. It is easy enough with a Httle care to determine 

 whether the apical segments are present or, as often happens, 



