384 
PHYSIOLOGY: LOEB AND NORTHROP 
rate of such life phenomena and of chemical reactions is the same and 
that such life phenomena are due to definite chemical reactions, e.g., 
a reaction resulting in the formation of a hormone. Such a tempera- 
ture coefficient was first observed by F. Lillie and Knowlton for the 
development of the egg of the frog and has since been found to exist 
generally. 
If we could show that in a species with a definite metamorphosis 
there exists not only a definite temperature coefiicient for the duration 
of life of the order of that of a chemical reaction, but that this coef- 
ficient is approximately identical with the temperature coefficient of 
the duration of the larval stage or stages, we should not have certainty 
but at least some probability that the duration of life is determined 
by causes similar to those which determine the duration of the larval 
stages. Our experiments were made on the fruit fly which has three 
definite stages, larval, pupa, and imago. We have already reported on 
some experiments in this direction^ but our new experiments were made 
on forms which had been freed from all microorganisms. This was 
necessary on account of Metchnikoff's suggestion that poisons formed 
by intestinal microorganisms play an important role in limiting the 
duration of life. 
Moreover, in our previous experiments only water or sugar solutions 
were offered to the flies and it was necessary to keep them on adequate 
food. It was found that on 2% glucose-agar the flies lived as long as 
on their natural food, yeast or yeast-agar. We preferred glucose-agar 
since it does not allow the larvae to hatch and we had to guard against 
this possibility since otherwise it would have been impossible to obtain 
correct figures for the death rate of the old flies. About 20 flies were 
kept in a large flask and about 8 to 10 different flasks with flies were 
used for the determination of life for any one temperature. The num- 
ber of flies which died were counted each day and the duration of life 
for each temperature is the average for all the cultures kept at this 
temperature. The experiments were carried on in thermostats whose 
temperature was constant within one- tenth of a degree. The tem- 
peratures selected were 30°, 25°, 20°, 15°, and 10°. Table 1 gives the 
duration of the three stages for the different temperatures with the 
probable error. 
These figures show, first, that the temperature coefficient for the 
larval stage, the pupa stage, and the life of the imago is of the order of 
2 or above for 10°C., which is characteristic for both chemical reactions 
and life phenomena in general. 
They show in addition, however, that the ratio of the duration of 
