Insects. 0751 



particles of excrement of these larvae were always of two colours; 

 I have, however, since observed the same thing to occur in other 

 species, and in various colours. In the case of the species under con- 

 sideration, the particles were of an oblong shape, rounded off at each 

 end, being somewhat narrower in the middle and always having a clear 

 white elliptical spot on one side. 



The larvae were tardy in their movements, keeping the abdomen 

 somewhat curved round underneath, and, when observed, were gene- 

 rally feeding, so that their presence on my rose-bushes was easily 

 recognizable. They devoured the leaves right up to the midrib. 



On the 26th of July the number of the larvaj in the garden began to 

 diminish, and those which I had kept in ajar began to crawl about in 

 an unsettled manner, and appeared, as I thought, to be looking about 

 for a suitable place in which to pupize. I put some hollow stems of 

 Deutschia scabra into the jar, as some sawfly larvae have a custom of 

 creeping into the hollow stems of plants for the purpose of under- 

 going their change ; but, after crawling about for some time, my larvje, 

 totally disregarding the hollow stems, spun up among the withered 

 rose leaves, which were lying on the sand at the bottom of the jar. 



The cocoons of these insects were very curious; they were double, 

 the outer case being very pale yellowish brown, extremely tough, and 

 having large meshes like a miniature cocoon of Cimbex Amerinse. 

 Some of them were entirely covered with grains of sand, but as these 

 easily came off on being touched, it appeared they were not spun 

 together by a web, but had merely been accidentally included in the 

 meshes. On cutting open the first cocoon a second was discovered, 

 consisting of a very fine and apparently membranous tissue, which 

 seemed to be quite free from the external cocoon, as it could be readily 

 taken out of it. Fig. 8 represents the cocoon of the natural form 

 and colour, and fig. 9 a cocoon of which the outer coat has been 

 partly torn off, leaving the inner part projecting beyond it. 



In September I found the contained larvae had not yet changed to 

 pupae ; on the other hand, during the first week of that month, a con- 

 siderable number of young larvae of another brood had established 

 themselves on the rose-trees in my garden. I had not observed any 

 images during the month of August; I nevertheless concluded that 

 the eggs which had produced these young larvae had been laid during 

 that month, because at this time I observed on the twigs of my rose- 

 bushes an appearance I had not noticed before, namely, that all the 

 twigs bearing leaves on which young larvae were found were very 

 much bent round, and that at the inside of the curve there was a long 



