_ 
( exli ») 
less highly developed and elaborated in accordance with any 
obvious peculiarity in its actual work. And it is certain—a 
fact which for many reasons I find exceedingly puzzling—that 
some ovipositors, apparently highly adapted for employment 
as excavating tools, are in reality never so used at all; but are 
possessed by insects which either make no visible perforation 
whatever, or at most just prick or scratch slightly the surfaces 
on which they oviposit, and leave the eggs not embedded in 
their resting-place, but merely adhering to it. 
I come now to another matter which very much interests 
me, but with which I can now deal only in a most cursory and 
inadequate way, viz. the actual relation of those ten “ pieces ” 
to the original abdominal-segments, out of which we must 
suppose them to have been developed by adaptation to their 
present functions. Briefly, I regard the abdomen of a Sawfly 
as consisting theoretically of ten segments, each having two 
distinct components, one ventral, the other partly dorsal and 
partly lateral (but with no visible differentiation of the sides 
and the back—the lateral regions, however, being indicated in 
all the segments except the 10th, by bearing each a spiracle). 
Of these segments, or (as Packard calls them) “ uromeres,” the 
Ist or basal one (=the “ propodeum ” or “ median segment ”’) is 
practically incorporated during pupation into the complex struc- 
ture called the thorax, and is often ignored in enumerations of 
the abdominal segments. The 10th and 9th are to a certain 
extent fused into one, which bears one pair only of spiracles, 
but, at its extreme apex, a pair of lateral palpiform appendages, 
called the cerct. This fusion occurs, I believe, not during 
pupation but in the embryonic stage. So at least I understand 
the statements which I find in books which I have consulted 
on this matter. Uromeres 2 to 7 can be recognised without 
difficulty as complete rings or annuli (each with its two plates 
and its pair of spiracles) in the imago of every 2 Sawfly. The 
dorsal, or dorso-lateral, plate of uromere 8 is also normal ; and 
the dorsal plate following this and forming the apex of the 
entire dorsum is also normal in its general appearance, except 
as to bearing cerci—though, as aforesaid, I take it to be 
really compounded of two uromeres, the 9th and the 10th, 
The ventral plates of uromeres 10, 9, and 8 are represented, I 
