MINDELEFF] OUTLOOKS OR FARMING SHELTERS 143 
The separation of this class of ruins from the preceding viiage ruins, 
while clear and definite enough in the main, is far from absolute. The 
sole criterion we have is the presence or absence of the kiva, as the 
sites occupied are essentially the same; but this test isin a general way 
sufficient. It is possible that in certain cases the kiva is so far obliter- 
ated as to be no longer distinguishable, but the number of cases in 
which this might have oceurred is comparatively small. The kivas, as 
arule, were more solidly constructed than the other rooms, and, as the 
preceding ground plans show, sometimes survived when the rectangu- 
lar rooms connected with them have entirely disappeared. 

Fic. 46—Ground plan of cliff outlook No. 35. 
Figure 46 is the plan of an outlook in the same cove as the last 
example of village ruin illustrated, and only 200 or 300 yards south of 
it. It may have been connected with that ruin, but could not in itself 
have been a village, as there are no traces of a kiva on the site, and 
hardly room enough for one on the bench proper. At the extreme 
northern end there are traces of walls on the rocks at a lower level. 

Fic. 47—Plan of a cliff outlook. 
The walls which were at right angles to the cliff were not carried 
back to it after the usual manner, but stopped about 3 feet from it, and 
the rooms were closed by a back wall running parallel to the cliff, and 
about 3 feet from it. This wall rises to a height of about 4 feet before - 
it meets the overhanging cliff, and consequently there is a long narrow 
passageway, about 3 feet high and 3 feet wide on the bottom, between 
it and the cliff. A small man might wriggle through, but with difficulty. 
