HATCHIXG OF GIPSY-MOTH EGGS. 3 
nonhatch i^oints — that is. points at which considerable nonhatch had 
been noted each year since the start of the " observation-point " inves- 
tigations. The sixth set. collected as controls, was obtained from a 
point which had not shown nonhatch. All the clusters thus obtained 
were kept in a cool place at the laboratory from the time of collec- 
tion until just before normal hatchino- time in the spring. Each 
cluster was then placed in a glass tube and allowed to hatch nor- 
mally. All larvae thus produced were counted, and later, when 
hatching had ceased, the remains of the clusters were rubbed over a 
cheesecloth sifter to separate the unhatched eggs from the hair with 
which the egg cluster is made. 
As soon as the records from the six sets of collections could be 
studied it was seen that there appeared to be some connection between 
nonhatch and cold weather, for clusters obtained from the five points 
before our normally cold months hatched almost perfectly, while 
those obtained after the cold season began failed to produce larvae. 
This suggestion opened up a new line for investigation, and ac- 
cordingly three sets of weather instruments, each consisting of a 
recording thermometer and a recording hygrometer, were procured 
and placed in small instrument houses. As the instruments required 
considerable attention the three houses were placed at points in the 
towns of AVestford, Acton, and Dover, Mass., two being chronic non- 
hatch and the third no nonhatch, not very many miles from the 
Melrose Highlands Laboratory. 
The instruments were started September 1, about the end of the 
laying season, and were run continuously until about May 1, the nor- 
mal hatching time. They were tested for accuracy at frequent inter- 
vals to insure true records, a discussion of which will be given later. 
METHOD OF HANDLING ALL EGG CLUSTERS USED IN 
EXPERIMENTS. 
Egg clusters when brought in to the laboratory were placed 
where they would be kept cool but not exposed to severe weather. 
Just before normal hatching time, which is about Ma}^ 1 at Melrose 
Highlands, each cluster was placed in a glass tube large enough to 
contain it unbroken. This tube was closed with a cotton plug. As 
soon as hatching was well advanced, the larvae were removed from 
the tubes and counted. At the end of the hatching season, the re- 
mains of each cluster were rubbed over a cheesecloth sifter, made 
by stretching fine-meshed cheesecloth over a frame, to separate the 
unhatched eggs from the mass of hair with which all clusters are 
made. These unhatched eggs were then examined microscopically 
and divided into infertile and nonhatch, the records of each cluster 
being kept separately. 
MONTHLY COLLECTIONS. 
Some of the clusters collected during the season of 1915-16 proved 
to be old nonhatch. During the winter, after the new clusters have 
become weathered, it is difficult to distinguish between old and new. 
To make certain that only new clusters would be collected, a suffi- 
cient number of these were marked with lumber crayon at the end of 
each laying season at the same six points used for the collections 
