1909. ] LIFE-HISTORY OF THE AGRIONID DRAGONELY. 275 
of the nymph decide its destiny as to the number of stages. 
Incubation, then, had apparently no effect on the number of 
stages which had been determined before incubation was resorted 
to, but it seems probable that if the nymphs had been incubated 
from the moment they hatched, the effect would have been 
different. That there is not complete regularity in the Table is 
perhaps not to be wondered at. A nymph may “ stagnate” for 
a period owing to some upset of its metabolism, and that might 
cause it to fall behind for a stage or even longer, in which case 
the question arises whether a long delay would not possibly 
necessitate an extra moult? I will consider this point further 
when discussing the significance of the moult. 
THE ANTENNA. 
The segmentation of the antennz has already been described. 
It is to be noted that the segment immediately above the scape, 
i. e. the third from the base, is the one which divides three times 
to produce new segments for the antenna, and that these divisions 
are, with a very few exceptions, definitely connected with certain 
stages in the life of the nymph. The actual formation of the 
new segment is not, of course, an instantaneous process, but takes 
place by slow degrees within the “shell” of the antenna, only 
becoming revealed when that shell is cast off. 
It is interesting to note that in one of the Ephemeride, where 
development has been followed, Lord Avebury (1863 and 1865) 
found that the third segment of the antenna, the one next to the 
scape, is there also the centre of growth. 
As the nymph increases in length the antenne also grow, but 
these organs do not maintain a rate of growth proportionate to 
that of the body. Whereas in the first stage the antenne are 
23 per cent. of the body-length, the percentage gets steadily less 
as the nymph grows, until, in the full-grown individual, the 
antenne are less than 12 per cent. of the body-length. 
The growth in length of the antenna is not due to the gradual 
lengthening of each segment, but is chiefly due to growth of the 
region immediately above the scape, This is easily shown by 
taking the average length of the antenna at each successive 
stage and finding what percentage of that length is taken up 
by each segment. The results are shown in the accompanying 
chart (p. 276), from which it will be seen that, whereas segment 3 
gradually increases its percentage of the length of the antenna, 
the scape and segments 4 and 5 only maintain a rate of growth 
proportional to that of the antenna, while segment 6 grows much 
too slowly to maintain its proportion to the total length. 
Stages X., XI., XIT.,and XIII. may any of them be the last stage 
in the life of the nymph, and in finding the average length for each 
segment at each stage I had to exclude the “ 7-segmented ” 
wa eee in each of these stages, except XITI., where there were 
“¢§-seomented” nymphs. For comparison I also found the 
meee for the “ 7-segmented” condition in stages X., XI., 
and XII., and they are given as X.a, XI. a, and xan ce: and it 
will be seen that in each of them segment 3 is proportionately 
