With respect to house counts it is interesting to compare the six villages in Hupa Valley which occur, on the one hand, on the Yurok list of Kroeber or on the Gibbs map and, on the other hand, on Goddard's map. The former give a total count for these towns of 82 houses, whereas Goddard shows 78. The Yurok census and Gibbs's map were formulated in 1851 and 1852 immediately after the advent of the whites. Goddard presumably derived his data from informants in or about the year 1900. From the two sets of figures it is clear that Goddard's cannot be too high and therefore those he gives for villages not covered by the earlier sources must be reasonably reliable. Goddard's total for 11 sites is 128, or an average of 11.6 houses per settlement. In default of other information this value, rounded off to 11, may be applied to no. 8.
Passing to the second group, we find that the five villages above Hupa Valley on the main Trinity River are shown on the 1851 census list as having 23 houses. The map by Gibbs assigns house counts to three of these, nos. 15, 16, and 19 with 4, 3, and 6 houses respectively. The average from the census is 4.6 houses per village and that from Gibbs is 4.3. We may accept from these data the value 4.5 as representing the mean house count for villages outside Hupa Valley proper. This is notably lower than the mean for the valley itself but is consistent with the poorer, more remote terrain.
Using Goddard's counts and the 1851 census where possible and supplementing by the estimate given above for the other villages we get a total of 198 houses for the Hupa. At 7.5 persons per house the population would have been 1,485. This is considerably above Kroeber's "barely 1.000" (p. 130).
A further question presents itself at this point. Should we accept without reservation the Yurok value of 7.5 inhabitants per house? Two lines of evidence become pertinent here. Goddard in describing Hupa society makes the following statements (p. 58):
A typical family consisted of the man and his sons, the wife or wives of the man, the unmarried or half-married daughters, the wives of the sons, and the grandchildren. To these may be added unmarried or widowed brothers or sisters of the man and his wife.... All the children born in the same house called each other brothers and sisters, whether they were children of the same parents or not.
(Emphasis mine.) To this Kroeber adds (p. 132): "The ultimate basis of this life is obviously blood kinship, but the immediately controlling factor is the association of common residence; in a word, the house." Now the social family in the usual monogamous tribe included the father, mother, children, and occasional close relatives. This was the underlying assumption of Kroeber's estimate of 7.5 persons as the social family among the Yurok. Here, very clearly, the social family was far more extensive, perhaps in occasional instances as much as double the Yurok value. At any rate the value 7.5 seems definitely too low.
Another approach is through the data furnished by Kroeber on page 131 of the Handbook. Here he shows a population census taken from seven villages in the year 1870 (the last item "sawmill" may be deleted as impossible to place). The total is 601 persons. Goddard's data show for these same seven villages a house count of 92 for the years centering around 1850. The direct average number of persons per house would be 6.53. Meanwhile Kroeber points out the disparity between the sexes: 232 males and 359 females. This he attributes to warfare alone, a dubious conclusion. Regardless of cause, however, we may calculate that in the absence of this male mortality and with a normal sex ratio of approximately unity the population would have been twice the female number or 718. The average number per house under such conditions would then have been 7.80.
It must be borne in mind that the population count is of 1870 and the house count is of 1850 or earlier. Although Kroeber feels that there was no population decline, apart from the effect of warfare on the males just mentioned, I cannot agree with him. In the face of the overwhelming evidence for a tremendous decline subsequent to 1850 on the part of the Indian population throughout all California it is impossible to concede complete immunity to any one tribe no matter how well protected it might have been. Consequently, we must allow for a reduction from 1850 to 1870 even on the part of the females. It is impracticable to set any sure figure on the decline but a value of 20 per cent would be very conservative, particularly in comparison with all the northwestern tribes. This would mean a population for the seven villages of 879, or say 900 in 1850. On this basis the number of persons per house becomes 9.78.
I think therefore we are justified in ascribing 10 persons to each Hupa house. If so the population would have been 1,980, or approximately 2,000. It is entirely possible that even this is too conservative an estimate.
HUPA ... 2,000