on colonr-relation. 225 



the smallest 4|- cm. The large larva measured was the 

 larfi;est I ever obtained of this species. 



Jul}' 23rd. — The first larva left off feeding. 



Aug 3rd. — One larva died. 



By Aug. 12th all the larvfe had pupated. 



Total results of the foregoing experiment : — 



Colour. No. of Iarva3. 



Light brown, with dark dorsal lines 5 



Slightly darker brown, dark dorsal lines ... 1 



Total 6 



All these were fed on Salix vitellina. 



The larvce with white surroundings. 



Cijlinder 11. — July 2nd.— I placed twelve larvae in the 

 second stage in cylinder 11, feeding them on Salix 

 vitellina, but mixing white sticks with the food-plant. 



July 9th. — I changed the food, giving the larvae 

 the silver-leaved Salix regalis instead of S. vitellina, 

 with a view to increasing the effect of the white sur- 

 roundings. 



July 13th. — The second ecdysis took place. The larvae 

 till now were of a uniform dusky brown, like all the rest. 

 After ecdysis they were all very slightly lighter. 



July 20th. — The length of the largest larva, fully 

 stretched, was 4h cm., the length of the smallest, 2|-cm. 

 Four were dark brown, three lighter brown, but with 

 distinct dorsal lines, and five were quite light. 



July 22nd. — The third ecdysis occurred. 



July 29th. — Two larva3 died, one dark, one light. I had 

 had more larvae than usual in this cylinder, being short 

 of cylinders ; but, as I was afraid they died from over- 

 crowding, I removed five of the ten left to cylinder 16, 

 still giving them white sticks and white willow. Five 

 larvfe remained in this cylinder. 



Aug. 1st. — The fourth ecdysis took place. One larva 

 died, one of the intermediate forms, between dark and 

 light. 



Aug. 6th. — Three larvae died, two light ones and an 

 intermediate. 



Aug. 20th. — The last larva died, an intermediate one. 



s2 



