On the 2d of October, Tucker, Hinckley, Corporal Gamarra and I left Arequipa; Watkins followed a week later.  The first stage of the journey was by train from Arequipa to Vitor, a distance of thirty miles.  The arrieros sent the cargo along too.  In addition to the food-boxes we brought with us tents, ice axes, snowshoes, barometers, thermometers, transit, fiber cases, steel boxes, duffle bags, and a folding boat.  Our pack train was supposed to have started from Arequipa the day before.  We hoped it would reach Vitor about the same time that we did, but that was expecting too much of arrieros on the first day of their journey.  So we had an all-day wait near the primitive little railway station.

We amused ourselves wandering off over the neighboring pampa and studying the medanos, crescent-shaped sand dunes which are common in the great coastal desert.  One reads so much of the great tropical jungles of South America and of wellnigh impenetrable forests that it is difficult to realize that the West Coast from Ecuador, on the north, to the heart of Chile, on the south, is a great desert, broken at intervals by oases, or valleys whose rivers, coming from melting snows of the Andes, are here and there diverted for purposes of irrigation.  Lima, the capital of Peru, is in one of the largest of these oases.  Although frequently enveloped in a damp fog, the Peruvian coastal towns are almost never subjected to rain.  The causes of this phenomenon are easy to understand.  Winds coming from the east, laden with the moisture of the Atlantic Ocean and the steaming Amazon Basin, are rapidly cooled by the eastern slopes of the Andes and forced to deposit this moisture in the montana.  By the time the winds have crossed the mighty cordillera there is no rain left in them.  Conversely, the winds that come from the warm Pacific Ocean strike a cold area over the frigid Humboldt Current, which sweeps up along the west coast of South America.  This cold belt wrings the water out of the westerly winds, so that by the time they reach the warm land their relative humidity is low.  To be sure, there are months in some years when so much moisture falls on the slopes of the coast range that the hillsides are clothed with flowers, but this verdure lasts but a short time and does not seriously affect the great stretches of desert pampa in the midst of which we now were.  Like the other pampas of this region, the flat surface inclines toward the sea.  Over it the sand is rolled along by the wind and finally built into crescent-shaped dunes.  These medanos interested us greatly.

The prevailing wind on the desert at night is a relatively gentle breeze that comes down from the cool mountain slopes toward the ocean.  It tends to blow the lighter particles of sand along in a regular dune, rolling it over and over downhill, leaving the heavier particles behind.  This is reversed in the daytime.  As the heat increases toward noon, the wind comes rushing up from the ocean to fill the vacuum caused