34 BUREAU OF AMERICAN ETHNOLOGY [BULL. 50 
on account of the constant water supply in the creek and the patches 
of land in the valley that could be cultivated. This was a desirable 
place for their farms. Had there been no caves in the cliffs they 
would probably have built habitations in the open plain below. 
They may have been harassed by marauders, but it must be borne 
in mind that their enemies did not come in great numbers at any one 
time. Defense was not the primary motive that led the sedentary 
people of this canyon to utilize the caverns for shelter. Again, the 
inroads of enemies never led to the abandonment of these great cliff- 
houses, if we can impute valor in any appreciable degree to the 
inhabitants. Fancy, for instance, the difficulty, or rather improba- 
bility, of a number of nomadic warriors great enough to drive out the 
population of Kitsiel, making their way up Cataract canyon and 
besieging the pueblo. Such an approach would have been impos- 
sible. Marauders might have raided the Kitsiel cornfields, but they 
could not have dislodged the inhabitants. Even if they had suc- 
ceeded in capturing one house but little would have been gained, as 
it was a custom of the Pueblos to keep enough food in store to last more 
than a year. In this connection the question is pertinent, While hos- 
tiles were besieging Kitsiel how could they subsist during any length 
of time? Only with the utmost difficulty, even with aid of ropes 
and ladders, can one now gain access to some of these ruins. How 
could marauding parties have entered them if the inhabitants were 
hostile? The cliff-dwellings were constructed partly for defense, but 
mainly for the shelter afforded by the overhanging cliff, and the 
rause of their desertion was not due so much to predatory enemies as 
failure of crops or the disappearance of the water supply. 
The writer does not regard these ruins as of great antiquity; some 
of the evidence indicates that they are of later time. Features in 
their architecture show resemblances derived from other regions. 
The Navaho ascribe the buildings to ancient people and say that the 
ruined houses existed before their own advent in the country, but 
this was not necessarily long ago. Such evidence as has been 
gathered supports Hopi legends that the inhabitants were ancient 
Hopi belonging to the Flute, Horn, and Snake families. 
There is no evidence that cliff-house architecture developed in 
these canyons, and rude structures older than these have been 
found in this region. Whoever the builders of these structures were, 
they brought their craft with them. The adoption of the deflector 
in the rectangular ceremonial rooms called kihus implies the deriva- 
tion of these rooms from circular kivas, and all indications are that 
the ancient inhabitants came from higher up San Juan river. 
Many of the ruins in Canyon de Chelly situated east of Laguna 
creek show marked evidence of being modern, and they in turn are 
not so old as those of the Mesa Verde. If the ruins become older as 
