Decemper 13, 1918] 
havior of other animals is also thus controlled. 
This resulted in the anthropomorphic expla- 
nations of reactions current at the time. For 
example, it was maintained that organisms 
which are photo-positive go toward the light 
because they hate darkness or love light, that 
the moth flies toward the candle-flame to sat- 
isfy its curiosity, etc. These explanations have 
justly been severely criticized, and yet the 
method is not necessarily at fault. 
If human conduct is dependent upon sub- 
jective states, and if other animals have such 
states, is it not altogether probable that their 
reactions are also dependent upon subjective 
states? If this is true it is possible to explain 
in a certain sense reactions in animals on the 
basis of psychic phenomena. It is maintained, 
however, that this is putting the cart before 
the horse, that it consists in attempting to ex- 
plain the unknown in terms of something still 
more unknown. With this contention I do not 
agree, for I hold that every individual knows 
his subjective sensations better than anything 
else. The question, then, resolves itself pri- 
marily into this. Does conduct depend upon 
subjective sensations? If it does then it is 
evident that in the study of behavior it is of 
the greatest importance to ascertain the dis- 
tribution of such sensations. But whether 
eonduct is dependent upon subjective phe- 
nomena or not, knowledge regarding the dis- 
tribution of such phenomena is fundamental; 
for it seems to be the only knowledge that bears 
upon the problems of the origin and evolution 
of consciousness. 
We judge as to the presence and nature of 
such phenomena in others almost wholly by 
comparing their behavior with ours. We 
know that conscious states in ourselves are ac- 
companied by certain reactions and when we 
see these reactions in others we conclude that 
their subjective experience is the same as ours, 
and by comparing the conclusions thus reached 
regarding subjective experience throughout 
the animate kingdom, we formulate conclu- 
sions as to the origin and evolution of these 
phenomena. I realize full well that conclu- 
sions based upon such evidence are precarious, 
but this method is the only method available 
SCIENCE 
583 
in the investigations of subjective states in 
others, and precarious as the conclusions may 
be they are far more likely to be correct than 
those formulated without such investigations. 
We must consequently either abandon this 
profound problem altogether or proceed along 
the line indicated. 
Aside from its bearing on consciousness the 
method of comparative behavior has an im- 
portant bearing on the problems concerning 
the evolution of reactions themselves and their 
interrelation, their. sequence. In its bearing 
on this problem comparative behavior is simi- 
lar to comparative morphology. As compara- 
tive morphology yields results concerning the 
relation between structures in different or- 
ganisms, so comparative behavior yields results 
concerning the relation between reactions. It 
is not primarily concerned in the relation be- 
tween the environment and the reactions. Its 
primary interest lies in the relation between 
the reactions themselves as manifested in va- 
rious organisms. 
In regard to the evolution of reactions, the 
comparative method in behavior must, however, 
give way to genetics just as the comparative 
method in morphology has. In this field we 
have as yet scarcely made a beginning. It is 
a virgin field of great promise. I should like 
to refer to Yerkes’s work on mice and Mc- 
Ewen’s on Drosophila. 
I have pointed out a number of important 
problems which are dependent for their solu- 
tion upon the relations between reactions and 
not primarily, if at all, upon the nature or the 
mechanics of the reactions. There are many 
other problems which can be greatly illumin- 
ated by a study of such relations. I shall re- 
fer to but one of these, modifiability in be- 
havior including habit formation and learning 
in general. 
Much of the recent work on the behavior of 
the higher animals centers about this problem, 
the work of Thorndike, Morgan, Yerkes, Wat- 
son, Carr and others. The results of this 
work have been of inestimable value, practical 
as well as theoretical, and yet it is based al- 
most entirely upon the relation between re- 
actions. Practically nothing is known regard- 
