32 METAMORPHOSES OF BRAZILIAN LEPIDOPTERA. 



feeding. When the caterpillar is full-fed the tube is eleven 

 centimetres in length, the diameter at the end attached to 

 the bark is a little over a millimetre, and the free end 

 between two and three millimetres. The caterpillar itself is 

 only twelve millimetres in length when full-fed. It never 

 leaves the tube, and is very timid, darting with wonderful 

 rapidity into the chamber in the bark upon the least alarm. 

 The head and first segment of the caterpillar are shiny black, 

 the rest of the body being a dull semitransparent drab. 

 The free end of the tube is generally lightly secured to the 

 surface of the bark, so that when the caterpillar retreats into 

 the bark the end is not blown away from the surface. I 

 tried some experiments to see whether the caterpillar could 

 in any way reach the bark from the open end of the tube, in 

 case it should be blown loose. I raised the tubes about half- 

 way down by bits of paper, leaving the mouths from half an 

 inch to an inch from the surface. After a short time thos^ 

 that were only raised half an inch had attached themselves to 

 the bark. As the caterpillar emerged from the chamber in 

 the bark, and moved along the tube, the latter curved slightly 

 inwards, so that the caterpillar was able when he reached the 

 end to catch the surface with his legs, and draw the tube 

 down to the bark. The curvature of the tube was so distinct 

 that it suggested the idea of its being done intentionally, by 

 the caterpillar pulling together the threads of the side of the 

 tube next the bark ; but I am inclined to think it was in 

 reality only the effect of the weight of the caterpillar as he 

 advanced. In those cases in which the tubes were raised 

 three-quarters of an inch and an inch, the caterpillars failed 

 to reach the surface of the bark. It is probable that the end« 

 of a tube very seldom does become loosened from the bark, 

 and in case such an accident should happen, the caterpillar 

 waits till the wind blows it against the bark, where the 

 loosely woven fabric is very likely to catch. 



