126 ERNEST HILL AND L. (J. HAYDON. 



fifth on the opposite, but the larva behaved in all respects, 

 and assumed a similar attitude to others, in which the hairs 

 were complete. 



Relative length, in the average, of filament to leaflet is of 

 general value in determination of species, although the rela- 

 tion difi^ers in individual leaflets and individual specimens, 

 l)ut in some species, as fun est a, it is more inconstant than 

 in others. This species, which exhibits tlie least constancy in 

 larval features, is also the most variable in the wing pattern 

 of imago. 



In our limited experience we find that, although larvge — 

 and some species more than others — are subject to much 

 individual variation in some essential points, yet the sum total 

 of the characters is sufficiently pronounced to enable the 

 identity of any one of the nine species which we present to 

 be readily recognised, even in the living larva. 



We have not followed up larvre systematically from youth 

 to maturity, observing any changes of character in im])ortant 

 features, but from a limited number of measurements it 

 appears that in very young larviB the frontal and palmate 

 hairs, though relatively large, are absolutely small, whereas, 

 when about half-growth is attained, both are in absolute 

 measurement about the size found in the mature larva. In 

 very young larvte the characteristic shape of the palmate 

 leaflets, as a rule, is not defined, but in larv« which have 

 passed through but a quarter of their career the features are 

 quite distinct. 



Duration of life. — The length of time occupied in each 

 stage varies with the temperature. In w^arm summer Aveather, 

 mean daily temperature about 73° F., very tiny larva? will 

 arrive at pupation in twelve to fourteen days, the imago 

 emerging in a further period of two ; but in autumn and 

 winter, mean daily temperature about 56° F., the larval stage 

 is prolonged to seven weeks, and the pupal existence to five 

 days. We have observed larvae certainly not one third of 

 average mature length as long as six weeks after collection, 

 when certainlv no insect has liad access to the water. 



