282 J. W. SCOTT MACFIE. 



endeavouring to pierce it. In doing this they keep the valves tightly closed, and 

 failing in their attempts, they wriggle down into the water again. In this way I 

 believe they may sometimes avoid drawing oil into their tracheae, and so escape the 

 direct action of the kerosene. If, however, oil does enter the siphon, as probably 

 '"happens eventually in the case of fully developed laiTae confined under a deep layer 

 of kerosene, it acts in a manner similar to that described in the case of C. fafigans, 

 namely, as a direct poison. In one experiment six fully developed larvae (fourth phase) 

 from the same sample were placed in couples in three tubes each containing 8 cc. of 

 the fluid in which they had been living. One tube was then closed with a layer of 

 crude kerosene one cm, deep, one with a similar layer of olive oil, and the third with 

 paraffin. The lai-vae were found dead after H and 1| hours in the first tube, after 

 2 and 2| hours in the second, and after 2 and 2|- hours in the third. The kerosene, 

 then, seemed to cause death more quickly than did the paraffin or the olive oil, which 

 merely deprived the larvae of access to air. It is interesting that the olive oil seemed 

 "to have the same action as the paraffin, for it was of course fluid and just as likely as 

 kerosene to enter the tracheae of the siphon and block them. The fact that olive oil 

 acted differently from kerosene is additional proof that the action of the latter is not 

 due to a mechanical blocking of the siphon tracheae. 



The difference between the survival time under kerosene and paraffin was, however, 

 sometimes so inconsiderable as to suggest that the larvae had escaped the poisonous 

 action and had succumbed mainly as the result of being prevented from breathing 

 the external air, and it was found that the larvae in dift'erent experiments survived 

 very different lengths of time, and this led to a series of obser%^ations being made on 

 the oxygen requirements of submerged mosquito larv' ae, a short account of which is 

 giv6n in the following pages. 



The Survival of Submerged Mosquito Larvae. 



In order to determine the length of time mosquito larvae can survive when sub- 

 merged, and the factors influencing this sui-vival, it was necessary to decide on one 

 standard method of carrying out the experiments. Da Costa Lima (1914) and Sen 

 (1914) have recorded the results of experiments on '' drowning " mosquito larvae. 

 The former confined the larvae in glass flasks filled with well-aerated water and covered 

 or stoppered, submerged in larger vessels of water, and the latter used test-tubes 

 ca,pped with gauze and immersed in basins of water. Both these methods in my 

 hands proved to be unsatisfactory, because now and then, especially when tap- water 

 was being used, minute bubbles of air separated on the sides of the submerged vessels. 

 These bubbles were sometimes so small as only to be clearly visible with the aid of a 

 lens, but as they appeared to influence the course of the experiments, all confidence 

 was lost in the technique. There was an additional disadvantage in the case of Sen's 

 inethod, for not only were air bubbles apt to form on the glass tubes and the gauze 

 covers, but also it was practically impossible to be sure that a film of air, or a small 

 bubble, did not adhere to the mesh of the gau/e. 



After trying a great many different methods — such as the use of specific-gravity 

 bottles, test-tubes simply inverted in a basin of water, etc. — it w^as decided that for 

 the present purpose it was most satisfactory to confine the larvae singly in small tubes 

 holding about 8 cc. of fluid, and to effect their submergence by covering the surface 



