most active in the dry torrid zone. It may be said that those dry portions
are, for the time being (as the rainless portions of the earth are
continually), within the operation of the surface-trades, and that
therefore the evaporated moisture is carried away by them toward the
equatorial rainy belt. Precisely so; but why carried away? Why should it
not condense, occasionally, at least, and drop the rain as it passes
along, if a great supply of moisture from excessive evaporation could
furnish rain. Perhaps it may still be said it is going from a cold to a
warm section. This is not true, as we have shown.

But, it may be said that the rainless regions at any rate receive no
moisture, and therefore can not supply any by evaporation. This would not
meet the case, as it would still be true that when the rainy belt has left
a given spot, the dry weather sets in with excessive evaporation, and the
north-east trades in summer, blowing from the countries lying north of the
rainless regions, and which have been supplied during the interval by the
extra-tropical rains, and are loaded with evaporation, are passing over
the rainless regions on their way to enter the central belt. So blow the
N. E. trades from the Mediterranean, and the Barbary States _over the
Desert of Sahara_ and into the rainy belt south of it; but drop no
moisture on their way, because exposed to no magnetic currents of an
opposite polarity.

But it is not true that all the rainless regions are without evaporation.
Egypt is an exception. The annual freshets of the Nile saturate its
central valley, and vast reservoirs of water are saved from it and let out
over its surface, and it all evaporates, but produces no rain. And so are
large quantities turned aside and scattered over the bottom lands of
Northern Mexico, and other countries, during the dry season, and their
evaporation furnishes no rain. Hygrometers and dew points are of no
consequence there--nor are they of any, on either side of the rainy belt,
where six perpendicular feet of moisture is evaporated in six months.

Again we have alluded to a strip of coast on the Pacific west of the
mountain ranges of South America, lying partly in Peru, partly in Bolivia,
and partly in Northern Chili, which, although long and narrow, washed by
the broad Pacific Ocean, is without rain. South America has no other
_wholly_ rainless region, so far as is known. A part of this region would
lie between the equatorial belt