114 ERNEST HILL AND L. G. HAYDON. 



various shapes are present on different parts of the thorax, 

 abdomen, and especially the wings of ' Anopheles,' and it is 

 a matter of great difficulty to decide in some instances what 

 form of scale predominates; nor does Mr. Theobald give us 

 aiiy indication of what portion of a wing should be examined 

 to decide this point. 



"(3) The terms 'lanceolate,' Mong and nai-row/ 'true 

 scales,' etc., are not sufficiently definite to permit of such 

 scales being easily distinguished from one another. ... It 

 is obvious that the distinction between ' hair-like curved 

 scales ' and ' narrow curved scales ' is not great, and also 

 that it would be difficult to decide whether the abdomen 

 is ' hairy,' or whether it is covered with ' hair-like scales,' 

 which apparently resemble hairs so closely that they cannot 

 be termed ' true scales.' As regards the wing again, it would 

 certainly be difficult to decide whether most of the scales 

 were 'lanceolate,' or whether they were 'mostly long and 

 narrow ' especially as the part of the wing to be examined is 

 not stated, but on this decision alone depends the distinction 

 between the genera Anopheles and Myzomyia. 



" (4) One of the objects of classification is to simplify the 

 identification of species, but the new classification does not 

 aid this in any way. In practice, it will be found much 

 easier to determine the specific name of any specimen of 

 ' Anopheles ' than its generic name, according to the new 

 system." 



With this criticism we are for the most part in accord, but 

 some genera, particularly Myzorhynchus and Cellia, 

 appear to be sufficiently distinct ; Ny ssorhynchus, how- 

 ever, touches closely on the latter, and both of the genera 

 are relatively uncommon. 



A genus should possess one or more characteristics which 

 are prominent, and by which it can be readily and certainly 

 distinguished from other genera, and this characteristic or 

 these characteristics should be, at any rate in broad terms, 

 common to all species in that genus, and shared by none in 

 any other. A classification by genera which fails in these 



