THE LBllTATIOXS OF KEROSENE AS A LARVICIDE. 283 



of the fluid with a layer of soft parafiiu one cm. deep. The paraffin was melted before 

 being poui-ed into the tubes and thus spread evenly over the surface of the water 

 leaving neither air iilm nor bubbles, and formed a smooth covering effectually cutting 

 off the larva from all contact with the air. Some difficulty was experienced in deter- 

 mining the most suitable hardness for the paraffin, for if it were too soft it did not 

 solidify sufficiently on coming into contact with the fluid, and larvae rising to the 

 surface might get caught by it, and if it were too hard it was apt to shrink away from 

 the sides of the tubes. Wlien once the right consistency had been ascertained, 

 however, the method worked well. 



The larvae were of course confined in a very small quantity of fluid (8 cc.) and in a 

 restricted space : but a large number of control experiments proved that the species 

 used in my experiments, namely Stegomyia fasciat'a and Culexfatigans, were not in 

 any way incommoded by these conditions and completed their life-cycles in such tubes 

 in a normal manner. I do not think therefore that pollution of the fluid need be 

 considered as a factor in these experiments, especially as they were seldom continued 

 for longer than a day, but as the amount of fluid was so small it could only hold in 

 solution a minute quantity of oxygen that might be available for the larvae when 

 access to the external air was cut off by the layer of paraffiii. Other experiments by 

 other methods were therefore necessary to ascertain the value in respiration of the 

 dissolved oxygen, but in determining the factors influencing the survival of submerged 

 larvae the standard method described above was emploved. 



Another factor that had to be considered was starvation, or the lack of suitable 

 food in the fluids used. This factor may also be neglected, I believe, because larvae 

 of the species employed were found to live perfectly well in tubes of distilled water 

 and eventually completed their development in them, if they were left with free sur- 

 faces exposed to the air. The period of the cycle was of course prolonged, as at first 

 the medium was devoid of all food materials, but the fact that the larvae did not die 

 shows I think that starvation would not affect the larvae appreciably in the relatively 

 short times occupied by the experiments. 



Factors Determining the Survival of Submerged Larvae. 



The Species of Mosquito. As might be expected, mosquito larvae of different species 

 showed very great differences in their ability to survive when cut off from access to the 

 air. 



In one experiment carried out in the standard method described above, five larvae 

 of Stegomyia fasciata and three of Culex fatigans were used. The fluid in this case 

 was a natural medium full of vegetable debris, etc. The tubes were examined everv 

 hour. One of the S. fasciata larvae was found to be dead after 6 hours, three after 

 7 hours, and the fifth after 8 hours ; that is, the average length of time the lai-vae 

 survived was about 7 hours. All the three C. fatigans larvae were found dead after 

 2 -hours. 



In another experiment carried out by Sen's method with five larvae each of 

 S. fasciata, C. fatigans and C. invidiosus, it was found that all the C. incidiosus larvae 

 and three of C. fatigans were dead after 30 minutes, that one C.faiigans lar^'a was 

 dead after |45 minutes, and the last after "one hour, but that all the S. fasciata larvae 

 were still alive after two hours. It is perhaps of some significance that the proportion 



(C325) f2 



