35 6 CAVE AND CLIFF DWELLERS. 
was some closer connection between them 
than that of mere resemblance. 
I had heard a dozen theories to ac¬ 
count for these terraces in the North, 
as for collecting water in dry seasons, for 
conducting water, as places for defense, 
etc., etc., but, with an actual case directly 
under observation, this seems to be a 
better explanation : In decades and 
centuries of rainy seasons of more or less 
violence, after the people had abandoned 
these northern houses, or had been killed 
by their enemies, all the retained loose 
earth would have been swept away, leav¬ 
ing only rude and dilapidated walls or 
terraces sweeping around the mountain 
sides, from which almost anything could 
be inferred, whether the most peaceful 
form or the most warlike fortification. 
Although our journey began at four 
