34 



more closely. It will then be seen that it is really of a very peculiar kind. As 

 briefly mentioned above and often pointed out by earlier authors, it must firstly 

 be remembered that the specific gravity of the pupae in contradistinction to that of 

 the larvae is smaller than that of the surrounding medium. They are always super- 

 compensated, more especially as they approach the last ecdysis; this is due to secre- 

 tion of air beneath the pupa skin (silvery gloss). By means of this supercompensa- 

 tion the pupae are pressed against the surface, this being their natural plane 

 of support, just as the bottom of the ponds is the support of all undercompensated 

 animals. 



The supercompensation is due to a large air globule, already mentioned by 

 MEINERT as "Flydekugle" (1886 p. 389) but more thoroughly described by HURST 

 (1890). The air-globule comes from the air in the tracheal system of the larva. 

 In a manner not hitherto thoroughly studied, it is carried down between the wing- 

 sheaths and legs of the pupa, being in connection with the trachealsystem of the 

 pupa through two large spiracles. As mentioned by HURST the air globule has 

 probably mainly a hydrostatic function (p. 7), but also serves to keep the pupa 

 afloat in a particular position, with the thorax uppermost and the apertures of 

 the trumpets at the surface of the water; further, to enable it to ascend to the surface 

 passively in a vertical direction and with the broad cephalo-thorax lying horizontally. 



The supercompensation is however so great that it would force the pupa out 

 of the water, on to the surface, if it were not counterbalanced by other forces. If 

 the pupa, after one of its somersault-movements downwards, again passively ascends 

 to the surface, the motion always takes place along a vertical line; the animal 

 is always directed in such a w r ay that the broad cephalo-thorax turns upwards; 

 reaching the surface the pupa will always touch it with four points lying above the 

 rest of the body, viz. the edges of the openings of the trumpets, and the two 

 stellate hairs on the first abdominal segment. By .means of these four points the 

 body is fastened to the surface and prevented by the surface film, from being forced 

 out of the water. 



The pupa never voluntarily leaves its place below the surface unless forced 

 either by light reflexes or by currents in or upon the waters, produced by enemies. 

 More especially in the last hours before ecdysis, the pupa is almost glued to the 

 surface film, being supercompensated to so high a degree that its power of loco- 

 motion is too slight to take it downwards. 



Otherwise, if variations in the surrounding medium force the pupa away from 

 its resting place, active motions set in. These are really very peculiar. They are 

 restricted to a series of downward-directed somersaults, but whether these somer- 

 saults are to conduct them to the left, to the right, or vertically downwards, the 

 pupae are not able to determine; they cannot direct their motions, being un- 

 able to produce any steering motion in any direction; they are only able to dart 

 away from the point to which they were fastened. The moment the activity ceases, 

 the pupae rise following vertical lines as do air-bubbles to the surface. 



