1909.] LIFE-HISTORY OF THE AGRIONID DRAGONFLY. 207 
It has occurred to me that this failure of loose eggs is due to 
asphyxiation, as possibly the eggs buried in the plant-tissues 
obtain a certain amount of air from the intercellular spaces 
of the plant; but if this is the case, it is somewhat surprising 
that the decaying leaves of the frog-bit retamed sufficient air to 
supply a fair number of eggs. 
At the time when the nymphs appeared I did not observe the 
emergence of the nymph from the egg; but I have since had 
the opportunity of watching nymphs of /schnura elegans hatch, 
and I have no doubt, from the similarity in all the later stages, 
that the peculiar phenomena observed in the case of J. Asseuus 
oceur in all the Agrionide, and, from the fact that I have 
observed some of the same phenomena in one of the Aniso- 
pterids, Sympetrum striolatwm, they are evidently common to all 
the Odonata. 
The dragonfly, both as a nymph and as an imago, requires 
living animals as food, and at the commencement of the work the 
difficulty was to find food small enough for the nymphs, which 
were but little over 1 mm.in body length. An infusion of horse- 
dung in water solved this difficulty, as in a few days it swarmed 
with life, one jar in particular constantly producing, for five or 
six months, a rich supply of Paramecia, which the nymphs ate 
readily. As the nymphs grew the Paramecia were replaced by 
daphnids, and it was surprising what large specimens the nymphs 
caught and demolished. Daphnids served as food all through the 
nymph stage after the protozoa became insufficient. 
~ The nymphs proved exceedingly easy to rear. A number of 
them, as soon as they appeared, were sorted out and measured, 
and each was then placed in a tumbler of water by itself, a small 
piece of Anacharis being put in to give it foothold and also to 
keep the water fresh. Hach tumbler was then labelled, and 
a record kept of the life of its occupant. The water in the 
tumblers was only very occasionally changed, and only when 
the glass became so thickly coated with algz that I could not 
see the inmate. For the first two months of their hfe the 
nymphs were examined and measured under the microscope at 
least twice a week—for the first week or two, every other day ; 
but such frequent interference with their existence proved 
unnecessary, and latterly the examination and measurement 
was much less frequent, being almost entirely confined to the 
times of moulting. 
Although I ine only followed Agrion pulchellum and /schnura 
elegans right through from the egg to the imago, I have also 
kept through many stages nymphs of Pyrrhosoma nymphula 
and Hr; yihromnus naias, and from my observations I think there 
is no doubt that, so far as the general facts go, what is true for 
one of these species is true for all the others. 
In this paper I am treating the subject from two points of 
view. First, I have detailed the phenomena observed in relation 
to the stages in the life-history in which they occurred, the stages 
