The cause of this absolute dependence of terrestrial temperatures upon density of the air-envelope is seldom discussed in text-books either of geography or of physics, and there seems to be still some uncertainty about it.  Some impute it wholly to the thinner air being unable to absorb and retain so much heat as that which is more dense; but if this were the case the soil at great altitudes not having so much of its heat taken up by the air should be warmer than below, since it undoubtedly receives more heat owing to the greater transparency of the air above it; but it certainly does not become warmer.  The more correct view seems to be that the loss of heat by radiation is increased so much through the rarity of the air above it as to more than counterbalance the increased insolation, so that though the surface of the earth at a given altitude may receive 10 per cent. more direct sun-heat it loses by direct radiation, combined with diminished air and cloud-radiation, perhaps 20 or 25 per cent. more, whence there is a resultant cooling effect of 10 or 15 per cent.  This acts by day as well as by night, so that the greater heat received at high altitudes does not warm the soil so much as a less amount of heat with a denser atmosphere.

This effect is further intensified by the fact that a less dense cannot absorb and transmit so much heat as a more dense atmosphere.  Here then we have an absolute law of nature to be observed operating everywhere on the earth, and the mode of action of which is fairly well understood.  This law is, that reduced atmospheric pressure increases radiation, or loss of heat, more rapidly than it increases insolation or gain of heat, so that the result is always a considerable lowering of temperature.  What this lowering is can be seen in the universal fact, that even within the tropics perpetual snow covers the higher mountain summits, while on the high plains of the Andes, at 15,000 or 16,000 feet altitude, where there is very little or no snow, travellers are often frozen to death when delayed by storms; yet at this elevation the atmosphere has much more than double the density of that of Mars!

The error in Mr. Lowell’s argument is, that he claims for the scanty atmosphere of Mars that it allows more sun-heat to reach the surface; but he omits to take account of the enormously increased loss of heat by direct radiation, as well as by the diminution of air-radiation, which together necessarily produce a great reduction of temperature.

It is this great principle of the prepotency of radiation over absorption with a diminishing atmosphere that explains the excessively low temperature of the moon’s surface, a fact which also serves to indicate a very low temperature for Mars, as I have shown in Chapter VI.  These two independent arguments—­from alpine temperatures and from those of the moon—­support and enforce each other, and afford a conclusive proof (as against anything advanced by Mr. Lowell) that the temperature of Mars must be far too low to support animal life.