6 BULLETIN 273, U. S. DEPARTMENT OF AGRICULTURE. 
portion of the vesicles of the hairs had buckled or partly collapsed* 
Larvae still living at this time presented the same appearance to a 
slight extent. As a check on this experiment, larvae were killed with 
chloroform and after 24 hours showed a smaller proportion of buckled 
or collapsed vesicles, indicating that sulphuric acid took up moisture 
from the porous hairs. 
Sections were made of the larvae and cells at the base of the 
hairs studied, but there could be found no indication of the presence 
of glandular cells. A large trichogen, however, is present at the base 
of each hair. This would signify that the liquid inclosed in the 
hairs is not a toxin, as Cholodkovsky suggests, but a mere colorless 
mobile liquid secreted during the formation of. the hairs. The 
phenolphthalein and litmus-solution tests were also tried, with nega- 
tive results, further indicating a neutral liquid. The peculiar shape 
of the vesicular hairs, however, suggests that they may at one time 
have had a poisonous function but that it has been lost in the present 
generations. 
The results of the experiments and observations indicate that both 
the normal vesicular and acuminate hairs are filled with a colorless 
liquid; that the hairs are hollow throughout, and that the vesicles 
contain the same media as the remaining portion of that type. A 
few globules of air were seen in hairs of living larvae which were no ted 
as exceptions. Air partially replaces the liquid after death, following 
which period the vesicles collapse. It is therefore probable, as Ige- 
nitzky and Shcherbakov have indicated, that the acuminate hairs 
play the greater role in making the larvae more buoyant, as these 
are from four to six times as long as the vesicular hairs. 
SCOPE OF INVESTIGATIONS ON WIND DISPERSION. 
In that the main purpose of the following investigations was to 
secure data on the maximum distance and the extent to which small 
larvae are borne by the wind, it was necessary to find conditions 
where an abundance of larvae were present in close proximity to 
treeless areas. These conditions were best afforded along the beaches 
in Massachusetts and New Hampshire, where there are stretches of 
marshland from 1 to 2\ miles wide and many miles long. These 
marshlands are occasionally flooded with salt water and do not con- 
tain vegetation favorable to the development of gipsy-moth larvae. 
The areas selected for the experiments were to the east of the heavy 
infestations, thereby getting all the advantages previously known to 
accrue from the northwest, west, and southwest winds. The Isles 
of Shoals, off the coast of New Hampshire, afforded ideal conditions 
in so far as their remoteness from the mainland was concerned and 
the fact that the country opposite has been infested since 1905. 
Locations were also selected in the hills of New Hampshire near the 
