316 Mr. Poulton's/M.r^/ier notes upon the 



When the larva is not at rest the ridge bearing the 3rd 

 pair of thoracic legs is bent upwards, so as to be nearly 

 parallel with the head and two anterior thoracic seg- 

 ments, presenting a remarkable appearance. The 

 colouring of the whole, the swollen 4th and 5th abdo- 

 minal segments with their tubercles, and the manner in 

 which the claspers are applied to the branch, — all 

 assist in forming a very perfect imitation of a twig. 

 Just before spinning up the larva becomes very stout 

 and short. The fourth stage began twenty-three days 

 after hatching in the case of most of the larvae, but 

 press of work prevented me from ascertaining the length 

 of this stage and the one succeeding it. The duration 

 of the stages is, however, extremely variable. Thus 

 out of seventeen larvae hatched in the same twenty-four 

 hours, one was changing the first skin, several the 

 second, while many had already entered the third stage. 

 Four larvae (out of the seventeen) spun slight cocoons 

 just thirty days after hatching. In these instances the 

 whole period was shorter than usual, and hence each of 

 the stages mentioned above (as far as possible average) 

 was also abbreviated. The larvae of this and other 

 batches pupated after variable periods all greater than 

 thirty days, but most of them less than forty. In some 

 few instances, however, the two last stages (especially) 

 were immensely prolonged, so that larvae which hatched 

 about May 5th had not spun up at the end of July. 

 Although these last larvae died before pupation there can 

 be little doubt that in this greatly prolonged larval 

 period, in a few cases, and in the extreme irregularity 

 altogether, there are indications of an older monogon- 

 eutic condition. This is all the more interesting in a 

 species which exhibits seasonal dimorphism to such a 

 marked degree as S. illunaria (in the perfect state). 

 This dimorphism must, of course, have arisen gradually, 

 long after a digoneutic condition had been established ; 

 and the very exceptional degree which is shown by the 

 former implies immense antiquity for the latter. Further- 

 more, the species has been polygoneutic in the present 

 year (1884), but the indications of the ancient monogo- 

 neutic condition are less remarkable in relation to 

 polygoneutism than to such extreme and exceptional 

 seasonal dimorphism, for well-marked instances of the 

 former relationship have been already adduced. It 

 must also be remembered that the dimorphism of this 



