unique feature was its situation between two clouds. Below us to the west was the coast cloud, which reached a varying distance inland according to the topography. Up the Rimac Valley it flowed like a great river; occasionally it filled not only the valley, but the barren ravines that branched from it north and south, rising till it covered even the ridges at our feet and, flowing around us, formed islands of Mount Harvard and the other more lofty points. The upper surface of this cloud was very sharply defined, but of wave-like form, so that its resemblance to water was at times so perfect that we could with difficulty persuade ourselves that far beneath its under surface all the varied activities of ordinary life were going on cheerfully. If from this view of 'clouds wrong side up' we turned our eyes upward, we at times saw another cloud system far above us; so that frequently we were between two clouds in a wide but shallow world, ourselves, perhaps, the sole inhabitants.
Fortunately for our work, the clouds above us appeared but seldom during the first months of our residence on Mount Harvard. During this time the extension of the Harvard photometry to the southern sky was begun. This is a determination of the brightness of all stars visible to the naked eye, a work begun by Professor Pickering in 1879. Photographic work was also carried on. The climatic conditions from April to September were excellent, but later clouds became troublesome. This condition of the sky growing worse as the cloudy season approached, it was decided to devote the following months to a meteorological study of different localities along the coast, and incidentally to extend the work of the meridian photometer in some region more free from clouds. Of the clearness and steadiness of the atmosphere in these different localities there was no certain knowledge, and the only way was to investigate it for ourselves. We left the Mount Harvard Station early in November in charge of a Peruvian assistant. During the next four months a personal examination was made by Dr. M. H. Bailey and myself of what appeared to be the most desirable localities along the coast, including Arequipa, the region about Lake Titicaca, both in Peru and in Bolivia, the Desert of Atacama, Valparaiso, Santiago and various other places in Chili. Perhaps no spot in all America offers a clearer sky than the Desert of Atacama. More than a month was passed at Pampa Central, near the center of this desert, and a study was made of the meteorological conditions, while the work of the meridian photometer was rapidly extended. There is a striking difference during the Peruvian cloudy season, between the cloudiness at Mount Harvard or at Arequipa, where the conditions are similar, and that on the Desert of Atacama. This is well shown in the following brief table, where A represents a perfectly clear sky during the whole night; B, a clear sky for a portion of the night; C, sky partially cloudy all night, and D, sky completely cloudy all night.
