wn. In view of the doubt it is better to omit it
from consideration.

Directing our attention now to the six sure towns of Driver, we find in
his paper some very pertinent data with regard to their size and
demographic characteristics. The sizes and house numbers given on page
183 (Driver, 1936) are:

Koticomota: "large town"; 2 sweathouses.

Netstul: "large town"; 40 houses; 1 sweathouse.

Owotelpete: 40 houses; 1 sweathouse.

Pipoholma: 40 houses; 1 sweathouse.

Unutsawaholmanoma: 1 sweathouse; 11 houses in 1870; 17 houses
"formerly."

Osoyukeju: "small town"; 1 sweathouse.

Before discussing the house numbers in detail we should call attention
to Driver's analysis of the village Unutsawa-holma-noma. This analysis
(1936, pp. 201 ff.) he says is based upon "concrete genealogical census
data of about the year 1870." There can therefore be no argument
concerning the validity of the figures he presents. He found, in brief,
that in this village there were 11 houses, containing 21 families and
92 persons. The occupants per house ranged from 4 to 21 with an average
of 9, the families from 1 to 6 per house with an average of 2 (actually
1.91), and the persons per family averaged 4.5 (actually 4.38).

When we examined Gifford's figures for the Clear Lake Pomo village of
Cigom we found 5.0 persons per family, 2.35 families per house, and
11.75 persons per house. The similarity between the two sets of values,
derived by different investigators independently, is clearly
significant. Moreover, the slightly smaller numbers discovered by
Driver at Unutsawaholma are explicable on the basis of the later date
(1870) taken as the base line. At any rate there can be no doubt that
the two villages were remarkably alike in composition of population.

In computing aboriginal population at Cigom and the surrounding country
it was pointed out that Gifford actually was dealing with a _declining_
population and that, if the aboriginal state were to be conceived
properly, his figures would have to be increased. For this reason the
family number was set at 7 instead of 5, which raised the number of
persons per house to 16.45. Because of other evidence the latter value
was reduced to 14.0.

For Driver's village the same considerations must apply. However, since
the family number was found to be 4.38, rather than 5.0 the aboriginal
value may be put at 6 instead of 7. Then, if the number of families per
hous