bcariiifj on the orifjin of Insects. 147 



the female of the common Ct/clops of oui- frcsliwatcrs, or 

 as an ordinary Zoea, in both of which latter, as in the 

 larvfB of the former, the antennre similarly consist of a 

 proximal or basal two-jointed portion, the protopodite or 

 pednncle, terminated by two branches, the endopoditc and 

 exopodite. 



As I had good reasons for believing that the ancient, 

 cosmopolitan, and little-modified group of the cockroaches 

 is directly descended from some extinct form of which 

 the Lepismatidoi are the only existing representatives 

 that we know of, I thought that, by carefully tracing the 

 development of the antennae in some species of it, I should 

 probably find a vestige of a second antennary branch 

 occupying the place of the rudiments in Machilis and 

 Lepisma. Nor have I been disappointed. In ripe em- 

 bryos of Blatta i^Panestliia) Javanica, which are still 

 invested in a larval skin, the one, probably, that is cast 

 by the young cockroaches at the moment of quitting the 

 marsupium of the mother, the antenn;:© in all essential 

 particulars resemble those of Machilis, consisting of a 

 many-jointed flagellum borne upon a two-jointed peduncle, 

 from the apical joint of which arises a relatively huge 

 somewhat compressed conical process, in the precise place 

 of the papilla-bearing tubercle in Machilis. This process 

 seems quite as entitled to be considered a distinct part as 

 is the simple one-jointed antennary (III) endopodite of 

 many Zoeas. It is probably cast off with the skin the 

 young insect sheds on leaving the body of the mother, 

 and to which it appears to belong ; be this as it may, no 

 vestige of it is to be detected in the smallest active 

 *larvjB' of the same species yet examined by me. 



There seems to me little doubt that we have here to do 

 with an ancestral phase of the antennae, a phase in which 

 those appendages each consisted, to use the terminology 

 employed in carcinology, of a proximal two-jointed proto- 

 podite, carrying at its extremity a long many-jointed 

 exopodite, and a short and simple rudimentary represen- 

 tative of an endopodite. The ancient condition of things, 

 of which we thus get a passing glimpse in the embryonic 

 history of this cockroach, may be presumed to have been 

 inhei-ited from some extinct form closely resembling the 

 much less modified Lepismatidce, in which also an endo- 

 podite is present, though it is reduced to the merest rudi- 

 ment. In short, in the lowly Pauropus we have antennae 

 with two fully-developed branches; in the higher Lepisma- 

 tidce the inner of these two branches is reduced to a mere 



