(chi } 
The insect was killed instantaneously ; without time to with 
draw its ovipositor, or make any movement to speak of. 
My object was not mere murder, but the desire to secure 
a pictorial record of the phenomenon as it actually takes 
place. A photograph which I took of it immediately after- 
wards (see Plate I, Fig. 2) shows clearly enough for my 
purpose—on a slightly larger scale—(1) the attitude of the 
insect itself at this stage, (2) the extended saw quite visible 
beneath the cuticle of the stem, (3) the long fissure which 
serves as a single mouth for all the pockets, and (4) even 
the two eggs deposited already resting each in its separate 
corner or nidus. (The original object is now in the Natural 
History Museum at South Kensington.) 
Since I made these observations, I have often carefully 
thought them over—weighing in my own mind, or trying 
to do so, the real import and importance of this or that detail 
in the process, and asking myself what on the whole was the 
most reasonable view to take of the ovipositor considered as 
a tool. The result is that I consider it certainly not the 
equivalent or ideal of an ordinary saw. It is a tool of that 
class which do their work largely by acting as wedges—in 
a word by splitting. But it also mangles and scratches the 
substances worked on, and this doubtless has an advantageous 
result in loosening their natural cohesions and liberating 
fluids which the egg will require as it grows. (N.B.—The 
Sawfly’s eggs grow considerably while in situ, and meantime 
much contraction and drying up of tissues occurs around 
them (see Plate I, Figs. 3 and 4). Both in form and in 
some of its movements the organ reminds me of certain 
features characterising various human tools—but rather per- 
haps those of the surgeon than those of the carpenter. And 
I should say the same of the sort of results which are 
achieved by it. Actual sawing, producing anything equiva- 
lent to sawdust or to the groove made by a normal saw, 
does not take place. Cutting or carving does enter into the 
process, but rather as an accessory than as an essential. The 
same may be said of the rasp-like features of certain ovi- 
positors. They assist the work, but it could be done without 
them. What appears to me to be essential and indispensable 
