144 MONTANA EXPERIMENT STATION. 



while constructing the new retreats. In building its ^cocoon the 

 larva usually selected a position along the mid-rib of the leaf and 

 began to weave a silken web back and forth, using mouth and 

 feet. After a considerable net-work had been spun this operation 

 was suspended and the worm proceeded to tap the mid-rib near 

 the petiole, or the petiole itself, in such a way as to cut off the food 

 suppl3^ The elastic threads draw the edges of the leaf together as 

 it wilts and, aided by the larva within, soon it is doubled, form- 

 ing a secure hiding place. As the leaf withers and droops it nat- 

 urally touches others of the cluster and is skilfully attached to them 

 by its busy inhabitant. This serves a double purpose since it pre- 

 vents the withered leaf from dropping and provides a very convenient 

 feeding ground for the worm. The hibernaculum is continually en- 

 larged and strengthened with .silk, frass and scraps of leaves. In a 

 few days it became evident that the larvae must emerge from these 

 withered leaves in order to obtain food. This indeed was found to 



I 



be the case, and each morning showed signs of renewed attacks up- 

 on the surrounding foliage. In the early morning, especially on 

 dark, cloudy days, a large percent of them were actively engaged 

 in these exposed places, retiring to the nest in the bright hours of 

 the day. As a rule these feeding grounds were entirely free from 

 the nest and upon exposed surfaces of the leaves since the larvi- in 

 this stage do not feed upon the epidermis of the under side. This 

 propensity to migrate to new quarters did not seem to be due to 

 the fact that the tree was so badly infested, as might be expected, 

 for even in this case there was seldom more than a single worm 

 in the cluster of leaves from one twig. It was rat! .-r due to thr fact 

 iliat the former quarters were not only unfit for tin: gi-owuig \vorjn 

 bat were also usually a considerable distance, c-nuparatiN ely, from 

 desirable food. In a few cases the larvae remained in the original 

 hi;)ernacula, but a careful estimate showed that about 75 per cent 

 o^ them did actually migrate and establish new homes. The gener- 

 al trend of the moving was from the center of the bole and larger 

 limbs where the winter cocoons were spun toward the outer and 

 younger growth. 



At no time during its larval stage does the insect eat as voracious- 

 ly and grow as rapidly as din-ing this interval between the first and 

 second spring molts, so that a considerable surface is actually repre- 



