112 



must be inferred that they had been left behind in the roots." I can add that this 

 supposition of Dr. Edwards is quite right; in every case I found pierced into the 

 roots the pointed terminal portions of the breathing tubes; in the tubes there is a 

 conspicuous predestinated zone of weakness, where the tubes break off; then the 

 pupa? rise to the surface. Most likely their barbed structure, while facilitating their 

 insertion, would prevent their being withdrawn. (Se also Edwards 1919 p. 83). It 

 is only difficult to understand how without hairs the pupa? are able to be suffici- 

 ently supported by the surface film during the metamorphosis. 



It is of some interest to remember that the Tcrniorhynchus larvae are probably 

 not the only mosquito larva 1 which make use of the air in the intercellular spaces 

 of the plants. According to Howard, Dyar and Knar (1917 p. 894), the larvae of 

 the genus Aedeomyia (A. squamipennis and catasticta) are to be found in shallow 

 ponds covered with the water plant Pistia "from which they probably derive their 

 supply of air although their habits have not been exactly determined". They do not 

 come to the surface, but hide between the water plants. The sipho presents no 

 peculiar structure, being shaped as the sipho of a common mosquito larva; only 

 the tracheae of the tube are said to be rudimentary. On the other hand, the an- 

 tennae have a very peculiar appearance in comparison with those of the other mos- 

 quito larva, being strongly curved, inflated, hollow, with a stout spinose digit on 

 the tip. Howard, Dyar and Knar (1917 p. 898) suppose that these inflated antenna? 

 may play some part in the respiratory process. 



In 191(3 Dette has made the supposition probable that a Tipulid-larva, Tri- 

 chosticta fhwescens, too, makes use of the air in the intercellular spaces of the plants. 



How the peculiar method of respiration, especially with regard to the Tcenio- 

 rhynchu.s-la.rva, has arisen it is difficult to understand. Further studies will probably 

 acquaint us with larvae in which the modus of respiration as well as the shape 

 of the sipho represent the missing link. 



For a long time I thought that we should have to study the tropics if we 

 wanted to solve these questions, and that it would be almost unreasonable to ex- 

 pect that it would be possible to find the solution of the problem in a little pond 

 in North Seeland, far away from the centre of the geographical distribution of the 

 genus. Later observations from this last winter have somewhat altered my opinion 

 upon that point. 



In my aquaria I had some Myriophyllum, which were growing well. In the 

 same aquaria there were a great many C. morsitans larvae. One day in January I 

 removed one of the aquaria into the bright sunshine and now observed the well- 

 known fact that vertical series of air-bubbles from different spots of the plants rose 

 to the surface. The numerous C. morsitans larvae now fixed themselves to the plants. 

 The position was always the same; the two strong curved seta? on the flaps of the 

 sipho were pierced into the plant-tissue and the apex of the open air-tube was 

 pressed against its cutis. The larvae were hanging down from the plant, quite as a 

 Tcrniorhynchus larva. For more than twelve hours they did not rise to the surface 



