184 REPORT OF THE BOTANICAL DEPARTMENT OF THE 
tion of these late-planted fields were killed by frost on Septem- 
ber 22, before the plants had completed their growth. This was 
especially true of sprayed fields which were protected from 
blight. Had it not been for frost, several of the experiments 
reported in this bulletin would have shown considerably larger 
gains due to spraying. In our opinion, late potatoes in this 
State should be planted before May 25, in order that they may 
mature before frost. 
Late blight, Phytophthora infestans, was destructive over the 
greater part of the State. In western and central New York 
and on Long Island it caused enormous loss. It appeared in the 
latter part of July and during August became very virulent. In 
the majority of the fields the blight was followed by more or less 
rot. Through the northern part of the State late blight was not 
destructive until nearly the close of the season, but it was fol- 
lowed by a violent attack of rot. For example, in Franklin 
County losses of 25 to 50 per ct. of the crop were not uncom- 
mon. 
The localities suffering Jeast from blight and rot were in the 
eastern part of the State, in Columbia, Rensselaer and Washing- 
ton Counties and along Lake Champlain. 
Karly blight was very unimportant. It caused some damage 
on Long Island and also in a few fields in other parts of the 
State, but it was not common. About the same may be said of 
flea-beetles. On Long Island they were exceedingly numerous 
and destructive, but elsewhere in the State they were not 
troublesome except in a few localities. Colorado potato beetles 
or “bugs” as they are called were also less troublesome than 
usual. In many cases there were so few that it was unnecessary 
to use poison for them. 
We estimate that the loss from potato blight and rot in 1904 
was at least 60 bushels per acre on the average. In 14 farmers’ 
business experiments spraying increased the yield 6214 bushels 
per acre and the average gain from spraying in 41 other experi- 
ments by farmers was 58% bushels per acre. As these experi- 
ments were well distributed over the State they should represent 
the average conditions fairly well, Besides it should be taken 
into consideration that in many of these experiments the spray- 
ing was not thoroughly done so that the blight was only partially 
