10 BULLETIN 1080, U. S. DEPARTMENT OF AGRICULTUEE. 
chusetts. The are.a over which this modifying effect protected the 
eggs was wide in Rhode Island and Massachusetts, but it gradually 
narrowed northward until at Portsmouth, X. H., the onl}' modifica- 
tion was right on the coast, and even there some of the eggs were 
killed. Farther north the cold was too great to be overcome suffi- 
ciently to allow the egg clusters to hatch completely, and only a par- 
tial hatch was recorded. 
The Weather Bureau records show that the foregoing hatch rec- 
ords are just what would be expected if low temperature was the 
cause of nonhatch. 
Bodies of fresh water apparently had no influence on the hatch. 
Elevation made no difference except in sections right on the border 
line of killing cold, that is. sections which as a whole had tempera- 
tures not quite low enough to kill. In these sections clusters in low 
areas were killed while those on the higher levels escaped, a differ- 
ence caused b}' cold draining into the low land and settling there. 
The object upon which the clusters were placed had no effect, as 
no matter what it was, if they were above snow ])rotection in a local- 
ity Avhich had severe cold, they were killed. 
Considerable observation on the possible influence of the food 
plants failed to show any difference. Records taken in any locality 
were uniform no matter what species of tree the eggs were laid upon, 
and irrespective of the abundance or scarcity of favored food. Of 
course this latter factor had some bearing on the abundance of clus- 
ters, but it made no difference in the nonhatch. 
At only one place, a location right on the coast in Rye. X. H.. was 
it possible to make a direct comparison between egg clusters com- 
j^letely exposed to the prevailing winds and others well protected 
from them. In a little grove about a quarter of a mile from the sea, 
exposed on its ocean side to the full sweep of northeast storms, the 
chisters on the windward side had the hairy covering entirely 
weathered away, leaving the eggs uncovered. These did not hatch. 
On the other side of the same grove other clusters were found which 
had not been deprived of their hairy covering. These hatched al- 
most perfecth\ 
Following the mild winter of 1918-19. the same localities were vis- 
ited again. Ever}^ cluster found, no matter what its position on the 
tree or other object, hatched perfectly. In fact, there was not a 
single nonhatch cluster found during the progress of the summer 
survey: neither was there one reported by men engaged in other 
branches of the gipsy-moth investigations. 
The winter of 1919-20 was a complete contrast to its predecessor but 
was almost exactly the same in temperature as that of 191T-18. 
There was a considerably greater fall of snoAV in some localities, 
which gave protection to a greater number of clusters. Observa- 
tions after this winter gave the same results as those after the other 
cold one, except that in some places which had more snow than dur- 
ing the other cold winter a greater number of clusters had hatched. 
All of the results of the summer survey observations, compared 
with records of temperature obtained by the Weather Bureau, serve 
to add conclusiveness to the fact that low temperature is the cause 
of nonhatch. 
