THOMAS] DISCUSSION AND COMPARISONS 927 
exuunple is noted from Polynesia or from any of the Malayan dialects. 
So far the data seem to agree with Conant’s conclusion, but more 
detailed examination presents at least some exceptions. 
We see the Nahuatlan family divided into two groups in this 
respect, the Aztecan and part of the Sonoran branches using the vigesi- 
mal system, while the Shoshonean and other divisions of the Sonoran 
branch follow the decimal method. Among the multiplicity of small 
linguistic families in California and Oregon examples of the vigesimal 
system occur sporadically, so far as is indicated by the still incomplete 
data, even occurring in one or two small tribes of a family while other 
tribes of the same family use the decimal system. But it is necessary 
to bear in mind that here, as in the Sboshonean group, the lists have 
been obtained after there has been long intercourse with the whites, 
which may have materially modified original systems. These facts 
are sufficient to show that ethnic lines do not always govern the range 
of the system. 
That there is a very general agreement among students in the opin- 
ion that as a general rule the adoption of the vigesimal system results 
from bringing the toes as well as the fingers into the count is admitted, 
yet it is possible that there are more exceptions to the rule than is 
supposed. That every vigesimal as well as decimal system has 5 at the 
base, or in other words, started with the hand, may be safely assumed, 
and that whenever 20 is expressly or impliedly understood as the 
equivalent of ‘one man” the toes are considered in the count may, 
perhaps, also be assumed. However, there are reasons for believing 
that in some instances the hands alone were used in actual count, being 
doubled to make the whole man; yet in such cases the toes were prob- 
ably originally used. 
It is possible and even probable that in some cases where the 
numeral terms have no reference to the toes or mana change from 
the original name has taken place. Such a change seems to be shown 
in the name for 20 in the Mayan dialects. In the Huasteca, Pokonchi, 
Pokomam, Cakchiquel, Quiche, Uspanteca, [xil, Aguacateca, and Mam 
the name for 20 is **man,” while in the Maya, Tzotzil, Chanabal, Chol, 
and Kekchi other terms are used, but even in these (except the Maya 
and Chol) awz/n7k, or **man,” is introduced into the terms for the mul- 
tiples of 20. Even in the Mexican (Aztec), which Conant looks upon 
us an exception, cempoalli (=one 20), which signifies ‘*1 counting,” 
evidently refers to something so well known and so generally under- 
stood as to require no explanatory term. What else could this, the 
thing counted, have been than one man—the fingers and_ toes? 
Although it must be admitted that there are some systems which can 
not be explained in this way, yet the explanation may be accepted as 
generally, in fact, almost universally, xpplicable. Even among the 
Greenland Eskimo, where we would suppose Protessor McGee’s sug- 
