In order better to comprehend what this minimum may be under extreme conditions, it will be useful to take note of the effects it actually produces on the earth in places where the conditions are nearest to those existing on the moon or on Mars, though never quite equalling, or even approaching very near them. It is in our great desert regions, and especially on high plateaux, that extreme aridity prevails, and it is in such districts that the differences between day and night temperatures reach their maximum. It is stated by geographers that in parts of the Great Sahara the surface temperature is sometimes 150° F., while during the night it falls nearly or quite to the freezing point--a difference of 118 degrees in little more than 12 hours.[10] In the high desert plains of Central Asia the extremes are said to be even greater.[11] Again, in his _Universal Geography_, Reclus states that in the Armenian Highlands the thermometer oscillates between 13° F. and 112°F. We may therefore, without any fear of exaggeration, take it as proved that a fall of 100° F. in twelve or fifteen hours not infrequently occurs where there is a very dry and clear atmosphere permitting continuous insolation by day and rapid radiation by night. [Footnote 10: Keith Johnston's 'Africa' in _Stanford's Compendium._] [Footnote 11: _Chambers's Encyclopaedia_, Art. 'Deserts.'] Now, as it is admitted that our dense atmosphere, however dry and clear, absorbs and reflects some considerable portion of the solar heat, we shall certainly underestimate the radiation from the moon's surface during its long night if we take as the basis of our calculation a lowering of temperature amounting to 100° F. during twelve hours, as not unfrequently occurs with us. Using these data--with Stefan's law of decrease of radiation as the 4th power of the temperature--a mathematical friend finds that the temperature of the moon's surface would be reduced during the lunar night to nearly 200° F. absolute (equal to-258° F.). _More Rapid Loss of Heat by the Moon._ Although such a calculation as the above may afford us a good approximation to the rate of loss of heat by Mars with its very scanty atmosphere, we have now good evidence that in the case of the moon the loss is much more rapid. Two independent workers have investigated this subject with very accordant results--Dr. Boeddicker, with Lord Rosse's 3-foot reflector and a Thermopile to measure the heat, and Mr. Frank Very, with a glass reflector of 12 inches diameter and the Bolometer invented by Mr. Langley. The very striking and unexpected fact in which these observers agree is the sudden disappearance of much of the stored-up heat during the comparatively short duration of a total eclipse of the moon--less than two hours of complete darkness, and about twice that period of partial obscuration.