48 
ing hypothesis that the similarities are due 
perhaps to convergent evolution, and that 
the two phenomena are in origin independ- 
ent. <A third poimt must also be consid- 
ered in this connection. In tracing cul- 
tural influences and relationships, espe- 
cially over wide areas, it is not to be ex- 
pected that the implement, custom or belief 
will remain throughout precisely the same. 
As we pass from tribe to tribe and from 
region to region modifications are almost 
certain to be made, either in accordance 
with the various tribal or national char- 
acteristics and inherited tendencies, or as 
a result of the varying environment under 
which the people live; the implement will 
vary as a result of new uses and materials, 
the custom or belief will change with the 
different habit of thought and attitude 
toward the world of different tribes. 
Of great importance again in drawing 
inferences from the observed similarity 
between implements or customs in differ- 
ent areas is the character and relative 
complexity of the things between which 
the similarity exists. Thus that two 
widely separated peoples both make use of 
a simple dug-out canoe is no necessary 
indication of historical or genetic rela- 
tionship between them, as this is one of the 
simplest forms of canoe possible, and one 
which any two peoples making use of wood 
as a material must almost inevitably hit 
upon. If, on the other hand, both peoples 
make use of a canoe of unusual shape or 
one which shows some peculiar technical 
features in its construction or ornamenta- 
tion, then the possibility or probability of 
relationship between the two peoples is es- 
tablished and should the peculiarity 
moreover be known only among these two 
peoples the evidence would become all the 
stronger. Further, if the peculiarities are 
in one case dependent on and in close rela- 
tion to the environment or necessitated by 
SCIENCE 
[N.S. Vou. XXXV. No. 889 
it, and in the other they show no such re- 
lation or are present only as useless or 
even detrimental features, the probability 
that the latter has in some way been de- 
rived from the former becomes great. The 
validity of the evidence for historical or 
genetic relationship thus is directly pro- 
portional to the unusualness or complex- 
ity of the things compared, features of 
very wide distribution or of very simple 
character being almost worthless for pur- 
poses of argument. 
Lastly, if an attempt is made not merely 
to show relationship between single imple- 
ments, customs or beliefs among different 
peoples, but to demonstrate a similar re- 
lation for an entire group of cultural ele- 
ments, consideration should be given both 
to the question of the relative perman- 
nence and resistance to change of the dif- 
ferent elements separately, and to the rela- 
tive importance in the respective areas of 
the group of elements on whose similari- 
ties the claim for general cultural rela- 
tionship is based. 
The historical method in anthropolog- 
ical investigation, then, if it is to lead to 
trustworthy results, requires on the part 
of the student not only most careful and 
minute investigation and comparison of 
the facts themselves, but also the constant 
consideration of these in the light of their 
relation in time and space, their continu- 
ity, their modifications, their individual 
character and their relationship to the 
sum total of the culture of the respective 
peoples. That enthusiastic adherents of 
or converts to this historical as opposed to 
the evolutional school should be led away 
by their enthusiasm and in so doing neg- 
lect to give due weight to these considera- 
tions is natural. It is also natural that the 
conclusions so arrived at should often be 
striking and almost revolutionary. But 
although the results may not receive gen- 
