232 Lilian J. Gould's e.rpcriiiu'nts 



leaves and remained quiescent on the tioor. This set 

 had no earth in which to hury, hut they constantly 

 covered themselves as much as possible with their 

 excreta, which was brownish green in colour, and always 

 of great quantity ; so that it had to be continually 

 cleared out. 



July 21st. — The third ecdysis took place, and the 

 larvae changed to various shades of very dark greenish 

 brown or olive-green ; heads brown. Two were very 

 dark. I removed seven of the lightest to cylinder 3, 

 with light green leaves only. Seven were left hero. 



July 2Gth. — The fourth ecdysis occurred. No change 

 of colour, which had been quite normal ever since the 

 alteration at third ecdysis, 



Aug. 13th. — Five larvre pupated. 



Aug. 14th. — The two last larva) pupated. No change 

 of colour due to surroundings was perceptible in any of 

 them. 



Cylindc)- 3. -July 21st. — Seven of the lightest larvae 

 were placed here from cylinder 1, and were given light 

 green leaves and no earth. 



July 26tli. —The fourth ecdysis took place. The colour 

 had been quite normal since the change at third ecdysis, 

 and no change took place now. 



Aug. 12th. — Three larvte pupated. 



Aug. 13th.— The last four larva3 pupated. 



Total results of experiment : — 



Colour. No. of larvie. 



Various shades of olive-greeu to brown ... 29 

 Other colours ... ... ... ... ... 



Total 29 



It will be seen that the above results were entirely 

 negative. The shades of colour are difficult to describe 

 in this species ; but all my larv;e, variable as they were 

 in shade, were more brown than green, even when in 

 green surroundings, and this was the case with any 

 larvaB I captured and compared with them. The lack of 

 resemblance to surroundings in this species may be 

 partly due to the burying habit ; the greenish brown is 

 sufficiently protective while on the earth, and once 

 buried the colour would be of less importance. Mr. 

 Poulton made some previous experiments with M. bras- 

 sicce, the results of which are shortly to bj publisheil. 



