51 
When the nymphs are ready to leave the water to trans- 
form they cannot ascend very small stems, such as those of 
reeds, owing to the wide divergence and separation of the legs; 
but they can readily climb up on a broad surface like the 
piers and timbers of a bridge or the side of a floating barge, or 
on large rocks or tangled growths. They do not go far above 
the water, not more than a few feet at most, and here they may 
be seen transforming at any hour of the day or night, but 
most commonly about daybreak, leaving behind them their 
mud-incrusted shells, which are often seen in great numbers in 
the situations already mentioned. A study of the different 
sizes of larve indicates that the nymphs require two or three 
years to attain maturity. 
The imago gomphid is in most species a medium-sized 
clear-winged dragon-fly, the thorax rather prettily striped with 
green or yellow, the abdomen rather slender in comparison 
with the thorax but often dilated towards the tip, the eyes not 
touching as in the @schnide and remaining Anisoptera, but 
widely separated as in the Agrionide and Petaluridw. In wing 
venation they approach the Libellulide, the triangles being well 
differentiated, with few or no cross-veins, but the antecubital 
cross-veins are mostly non-coincident, as in the other Ani- 
soptera. The anal loop is small, inclosing from one to a few 
cells. As with the Libellulide, the ovipositor is not developed, 
and the more or less oval eggs are transferred free to the water 
by successive taps at its surface with the tip of the abdomen 
during flight, the females ovipositing alone.* The number of 
eggs deposited is very large. Copulation takes place while at 
rest on shore. The females frequent the trees or pathways back 
of the shore, while the males rest nearer the water. The latter 
may be quickly recognized by their curious terminal append- 
ages resembling four fingers about to grasp something, while 
those of the female are short, simple, and inconspicuous. The 
period of flight usually occurs rather early in the season, vary- 
*The eggs in a Mympfhea leaf figured by Lampert, '99, as eggs of Gomphus 
have been demonstrated by our breeding work at Havana to be those of Agrionide- 
