6 The Philippine Journal of Science ms 



It will be observed in Table I that type indefinitus has the 

 narrower range and, in general, the less variability. The amount 

 of overlapping of the two types is such as to indicate the im- 

 possibility of differentiating female adults of these types by pal- 

 pal ratios alone, unless the first black band should have a length 

 near the maximum of type Giles or the minimum of type inde- 

 finitus. No other adult character has been thus far noted suffi- 

 ciently constant to differentiate these forms. Type Giles of 

 brackish-water origin seems to have less variability than the 

 fresh-water forms. However, all of the brackish-water type 

 Giles were collected in one locality, and collections from a variety 

 of brackish-water habitats might show a greater amount of 

 variability. To sum up, the data given in this table indicate 

 that it is difficult, if not impossible, to describe adults of certain 

 closely related members of the rossi group without recourse to 

 the statistical study of many specimens. Larval characteristics 

 must be taken into account in differentiation, and possibly breed- 

 ing experiments will be necessary to furnish the final data for 

 the classification of some more closely related types. 



Habitat. — The larva? of type indefinitus of Malaya are most 

 commonly found in muddy pools exposed to the sun, and the 

 greater part of my material was collected in such places. Ano- 

 pheles rossi Giles of India is reported to frequent similar places. 

 In Malaya, however, so far as my observation has gone, larvse 

 of type Giles are never found in such pools, but frequent ponds 

 or large pools relatively clear and supplies with grass or other 

 vegetation. They are commonly found associated with fidi- 

 ginosus, barbirostris, and sine7isis. In the collection of many 

 hundreds of larvae in small muddy pools, I have never once found 

 this type. Type indefinitus, while commonest in muddy pools, 

 seems to be less selective in its habitat and is found in a variety 

 of places. It is sometimes associated with type Giles in ponds. 

 On a number of occasions I have found small muddy pools con- 

 taining only type indefinitus, while a few feet away ponds 

 contained type Giles in abundance. 



The relatively restricted habitat of type Giles, especially its 

 absence from small muddy pools, would be evidence that it is phy- 

 logenically distinct from type indefinitus and would raise the 

 question as to whether it may not differ, biologically at least, 

 from A. rossi Giles of India. 



No microscopical character was noted that was of much use 

 in distinguishing the two types of Malaya in collecting. Type 

 indefinitus of muddy pools seems slightly larger than type Giles, 



