He tells us that the larvae showed signs of changing into pupae in June. The change was made without the larvae leaving the water, and they underwent all their transformations in less than a fortnight. At the approach of the change the larvae became whiter in colour, but less transparent. Then they cast their skin, leaving therein the air-vessels, or rather their external covering. After this last moult was over, he was surprised to find that the tube which formed the tail of the larva, and by which it took in a supply of air, though it serves the same purpose in the pupa, is attached to the thorax, near the top of the head[68]. Lyonnet appears to have overlooked the fact that there was a second and shorter tube given off from the thorax, which most observers consider to be functionless (Fig. 90).

Strange as is the larva, the pupa is stranger still, and seems even better adapted for existence in the mud. The hinder part of each segment of the abdomen is thickened and studded with chitinous projections. This thickening is more marked, and of greater extent in each succeeding segment, reaching its greatest development in the last segment, which is armed with hooks. The body part of each segment bears rows of smaller spines, so that this pupa should have little or no difficulty in moving through pretty thick mud.

Several of these Crane-flies have passed through all their stages with me, and in nearly every case the transformation from pupa to perfect insect was made in water—in a tube three inches long, with a diameter of about an inch. The larvae were taken about the middle of September. My notes show that the first pupated on November 17, and the first fly came out on November 25.

The long breathing-tube of the pupa was several times purposely displaced from its position on the surface. The creature was evidently incommoded, and twisted itself into strong curves; the head was thrown from side to side till part of the breathing-tube was raised above the surface and adhered to the side of the glass tube. Nor did the efforts cease till a considerable portion of the tube was in free communication with the air. This would seem to show that the air-supply is taken through the bladders which appear at irregular intervals in the breathing-tube, and not at the tip, where the keenest observers have failed to find an opening.