1924] Systematic Position of the Family Termitaphididce 271 
abdominal segments). There are thus fewer head lobules than 
in any other species of the family and more than the average of 
abdominal lobules. The flagella seem to me in most cases longer 
than figured by Wasmann. Both in the distinct separation of 
the marginal lobules and in the length and flagellate appearance 
of their appendages Termitaphis is clearly intermediate between 
the Aradid nymph described above and Termitaradus. The 
swollen form of the body is however very un-Aradoid, and may 
be best explained as an instance of terniitophilic physogastry. 
THE SPECIES OF TERMITARADUS. 
Eight species, including the new one described below ma}^ 
be referred to this genus. Of these, one, namely T. australiensis 
(Mjob.) is quite inadequately described and its relationships at 
present obscure. 
The important characters in the genus appear to be the 
form and average number of the flabella. It is therefore un- 
fortunate that these peculiar structures are so easily detached. 
In their absence however, their number can be ascertained by 
counting the lobules in which they arise. The number of the 
lobes, at least in those species which show 13 or 14 on each side 
seems less reliable. One would of course have been inclined to 
regard the presence or absence of an additional lobe as a charac- 
ter at least of specific importance, but the following -considerations 
have led the writer to reject it as such.— 
It is a single meristic character, and such are known to vary 
intraspecifically. 
Nymphs in species in which the adult female is 13-lobed {T. 
annandalei and the normal 13-lobed Panama form) show the 
extra lobe clearly developed. 
Specimens taken together in the case of three separate lots 
include a mixture of 13-lobe and 14-lobe examples. 
The specimens in these lots agree exactly in all other char- 
acters. 
The males taken with these lots are all identical and all 
show 12 lobes, as does also the male of the typically 13-lobed 
T. annandalei. 
The 14-lobed specimens have the lobes distributed as fol- 
lows in the manner indicated by Morrison. — 2 to the head, one 
to the prothorax, 2 to the mesothorax, one to the metathorax and 
