132 



PSYCHE. 



[June-Sept. 1S94. 



The bodies and feet were now of a pale 

 green. On the first five segments was an 

 additional row of tubercles below the others. 

 Of the two middle rows, all were yellow 

 except those on the second and third seg- 

 ments, which were a bright coral red. and 

 those on the first, which wei'e light blue. 

 All the other tubercles were light bine. The 

 head was greenish yellow with a black stripe 

 down each side. The average length of the 

 caterpillars at this time was one and one- 

 eighth inches, the largest measuring an inch 

 and a half. 



They moulted for the third time on the 

 eighteenth day, having eaten nothing for 

 the preceding forty-eight hours. The bodies 

 were of a beautiful, clear, light blue on the 

 back, changing to a pale green at the sides 

 and under part. The coloring of the tuber- 

 cles was the same as before. The black 

 spots were fewer, smaller, and fainter. The 

 head and the feet were pale green, the former 

 marked with black at the sides. The great- 

 est length was two and one-eighth inches, 

 the average, an inch and a half. 



On the twenty-seventh day the fourth and 

 last moult took place. The caterpillars had 

 not eaten for two days and a half. The skin 

 was now entirely of a light green color, the 

 black spots having disappeared. The two 

 middle tubercles on the second and third 

 segments were of a dull red with seven or 

 eight very short black bristles; those on 

 the fourth segment were orange-vellow, 

 nearly as large as the red, and bore six 

 bristles. On each of these tubercles was a 

 row of black warts under the bristles. 



When full-grown the caterpillars measured 

 three inches and a half and were as thick as 

 the forefinger. Two completed their growth 

 in five weeks and a half; and on September 

 first all the cocoons but one were either 

 finished or partly made. 



During the first eight days the caterpillars 

 fed mostly at night and ate very little. 

 Through the day they would lie motionless 

 on the exposed surfaces of the leaves, in 



groups, for hours at a time. Just before 

 the second moult they developed vigorous 

 appetites which constantly increased in 

 voracity, both day and night, until the end. 



In eating, they always began at the edge 

 of the leaf, sometimes cutting away the part 

 between two veins till the midrib was 

 reached, sometimes eating directly across 

 the veins. When half of the blade had 

 been eaten they would often devour the 

 entire midrib before beginning on the other 

 half. They drank greedily whenever wet 

 leaves were given to them. 



In long periods of rest the head was drawn 

 under the second segment, but in short 

 periods it was merely withdrawn into the 

 first segment, while the thoi-acic feet still 

 retained their hold of the edge of the leaf, 

 so that feeding could be resumed without 

 change of position. 



There was great unanimity in their move- 

 ments. It was rare to find even one feeding 

 while the others were resting; and after 

 lying motionless for some minutes they 

 would begin simultaneously to eat again, 

 as suddenly as at the touch of an electric 

 button. 



They were social in their habits from the 

 beginning, and even when full-grown as 

 many as nine would lie close together on 

 one branch, the head of one perhaps rest- 

 ing on the body of another. Two would 

 sometimes feed from the same part of a 

 leaf, their heads touching every time they 

 returned to the starting point. They never 

 showed any of that irritation at contact with 

 each other so common with some caterpil- 

 lars, nor did they pay any attention to the 

 touch of a finger either on the skin or the 

 tubercles; but if an aphis, introduced into 

 the box on a leaf, crawled over one of them, 

 the caterpillar tried by writhings and twist- 

 ings to rid itself of its visitor. 



Touching the larvae with the finger was 

 an experiment made when they were well 

 grown. Otherwise they were not handled 

 at all during their entire development. 



