1909.4) LIFE-HISTORY OF THE AGRIONID DRAGONFLY. 271 
also one which passed through 12 stages, while all the other 
nymphs were in their 6th, 7th, or 8th stages at the beginning of 
December. Now all these nymphs were in the incubator at the 
same time and for 81 days, and during that time they moulted 
a number of times, varying from 3 to 7. If we consider the 
number of times that each of those nymphs moulted in relation 
to the total number of moults it passed through, we find that 
those introduced into the incubator in stage 4 passed through half 
(5 out of 10 and 7 out of 12) their total number of moults in the 
81 days, while those introduced at later stages passed through less 
than half their total number—a fact which supports what I have 
already shown by two other methods, namely, that the stages tend 
to lengthen out as the nymphs advance. 
Whether or not this lengthening out of successive stages is a 
phenomenon common to the larvee of all insects has still to be 
shown, and I can find no reference to the subject, but it is 
interesting to note that Newport (1839) mentions that the rate 
of the pulse in the larve of Sphinw ligustri gets slower in 
successive stages—a fact which might well be correlated with the 
lengthening of the stages. 
Leaving out for the present the question as to whether the 
winter warmth has curtailed the number of the stages, a point on 
which I have very little evidence, it is easy to show that the 
favourable conditions greatly curtailed the duration of the life of 
the nymph. 
All the imagines from the winter-warmed nymphs appeared 
between April and July of the year following that in which the 
eges were laid—that is, within 10 months. On the other hand, 
the imagines from the nymphs which were in ordinary laboratory 
temperatures during the winter, with one exception, appeared a 
year later than the others, more than 20 months after the hatch- 
ing of the eggs, and in spite of the incubation from April to July. 
In connection with these two extremes, the three cases of nymphs 
which for 21 days in February enjoyed the warmth of the 
incubator are most interesting. In all other respects they were 
treated exactly like the second batch of nymphs—that is, they 
endured ordinary laboratory temperatures until April 20th, and 
then were in summer temperature,—yet they all three hatched out 
with the first batch, one in June, one in July, and the third in 
August. It is certainly interesting that three weeks’ warmth in 
February, during which each nymph moulted twice, should have 
made a difference of more than 10 months to those three nymphs ; 
and since temperature has such a marked effect it may be possible 
to account for the periodic swarms of dragonflies on the ground 
that a specially mild winter in some region may bring ona batch 
of nymphs, so that the imagines appear a season earlier than under 
normal circumstances. In such a case the batches of 2 years 
would appear together, and might be a prelude to a migration. 
I cannot account for the single individual which, without any 
winter warmth, emerged with the first batch of nymphs. It is 
possible that it was in the incubator with the other three in 
