1888. J 47 



invariably dragged down towards the posterior end. I mention this, 

 as after microscopical preparation it would not be otherwise always 

 easy to be sure that the real insect was under examination. 



The forms of the scales have been made the basis of the generic 

 characters, and although, as I have mentioned, they vary considerably, 

 there appears to be to some extent an evolutionary principle, which 

 may guide us a little in tabulating the different genera. 



I think it will be found — ■ 

 (i). That the male scales always have a tendency to the linear form, even although 



the females of the same species are circular (e. g., Aspidiotus). 

 (ii) . That this tendency to a linear form on the part of the male has, in some species, 



developed, and become a sexual constant character, the female remaining 



circular (e. g., DiaspisJ. 

 (iii). That in some species this linear tendency appears to have extended to both 



sexes, but the scale has assumed a compromise between the circular and 



linear, i. e., an elongated circular or mussel-shape (e. g., Mytilaspis). 

 (iv). That in some species the linear tendency has developed and become common 



to both sexes, but whilst constant in the male, the female is liable to 



variation {e.g., Parlatoria). 

 (v). That in some species the same character has become still more nearly constant 



in both sexes, the female, however, frequently showing a tendency to a 



compromise between the circular and linear form (e. g., Chionaspis). 

 (vi). Lastly, that the linear form has been definitely assumed by both sexes {e.g., 



VhleriaJ. 



Putting these observations in a tabulated form, we have — 

 Aspidiotus, $ & $ circular ; S tends to linear form. 

 Diaspis, <j> circular, $ linear. 



Mytilaspis, S & ? elongated-circular, or mussel-shaped. 



Parlatoria, ? variable, sometimes circular, sometimes elongate j $ 

 linear. 



Chionaspis, $ linear and ? elongate, but $ sometimes much 

 rounded. 

 Uhleria, <J & 9 linear. 



I have given above only those six genera which appear to be ac- 

 cepted as definite, and which depend for their determination upon the 

 formation of the scales. The two genera, Parlatoria and Chionaspis, 

 appear to be very closely allied, but I have not yet had an opportunity 

 of examining specimens of either, and I have, therefore, taken their 

 characteristics from Prof. Comstock's description. Although the 

 characters on which the several genera have been established seem to 

 be sufficiently constant for the purpose, yet, as might be expected, 

 some species of the same genus show a greater inclination than others 

 to assume the characters of another genus. For example, the male 

 scale of Aspidiotus ostreceformis, which, when fully adult, possesses a 

 scale very much more developed posteriorly than is shown in fig. 5 



