166 
Journal of Agricultural Research 
Vol. XXXI, No. 2 
Hine (7) reports observations on the effect of freezing insect larvae. 
Those of Bellura obliqua were subjected to temperatures below 0° F. 
and revived. He records that larvae were frozen solid and could be 
snapped in two, but the pieces were alive when thawed out. He also 
records observations on tomato hawk moth larvae. Those that had 
pupuated before the first frosts came carried through normally; im¬ 
mature forms were frozen as the plant was frozen, those on the tips 
being killed with the first touch of frost while those on the lower parts 
of the plant were not killed until frost struck that part of the plant. 
Sir John Ross’s account of his Arctic expedition is quoted by Hine. 
Ross records that Larria rossi withstood successive freezings of— 40° 
F. Further references along the same lines are found in Houlbert 
(8, p. 255) . This writer, quoting Justi, a worker in 1753, reports that 
larvae could be frozen so hard that they could be snapped in two like 
ice. If the frozen larvae were permitted to return to normal tempera¬ 
tures, then development proceeded normally. 
Duval and Portier (4) found the temperature limits of Cossus cossus 
to be —22° C. They report larvae so hard as to be easily broken in 
two without water appearing at the point of section. They incline to 
accept Bachmetjew’s conclusion that only the intercellular water 
freezes and the rest supercools. 
The latest paper available to the writer is that of Knight (9) on the 
nature of the color pattern of Perillus bioculatus Fabr., which contains 
a table of data on the supercooling of that insect considered from the 
standpoints of season (involving the colloidal state of the lymph) and 
repetitions of freezing. 
EXPERIMENTAL WORK ON SUPERCOOLING AND FREEZING 
METHOD 
The electrothermal method of taking individual temperatures was 
used. This method is described by Taylor (17), Harvey (6), and 
White (18). A portion of the apparatus is figured by Pirsch (18), but 
in the present study a mirror galvanometer was substituted. This 
provided an admirable indicator of high sensitivity. A modification 
of the method for taking individual temperatures will be described 
later. 
In each case of individual freezing tabulated in Table I, a finely 
pointed thermocouple of constantan and copper was thrust into the 
insect at approximately the ventral middle. The temperature read¬ 
ings were taken at intervals while the insect was being cooled. 
It was found best to avoid using thermocouples made of heavy 
wire, the best and most uniform results being obtained by the use of 
No. 36 wire. The only trouble encountered with this wire was short- 
circuiting due to the rapid wearing of the silk insulation. Erratic 
jumping of the galvanometer soon gave notice of this trouble. When 
heavy wire was used, even though it could be tapered to a very fine 
point by means of acid, considerable difference was found to persist 
between the temperature registered by the instrument and a control 
thermometer. 
All the specimens used for individual freezing were taken from 
exactly the same environmental conditions, so that variables due to 
changes in environment were eliminated. The variables due to 
