Mev. E. Hill — On Evaporation and Eccentricity. 483 



supply, by means of their effect on radiation, lower the average 

 temperature. May not this neutralize the increase of evaporation 

 which would otherwise result? 



This diiBculty may perhaps be removed by aid of an illustration. 

 The temperature of the earth's surface as heat is being poured on to 

 it depends on the loss of heat from radiation and evaporation, and 

 whatever other causes of loss exist : the temperature rises till the 

 expenditure equals the supply. This is very analogous to the depth 

 of water in a cistern into which there is a continual flow, but from 

 which there is also a continual leakage. If we suppose the leaks to 

 be cracks in the sides, which widen rapidly upwards, we get an 

 analogy to the rapid increase of radiation and evaporation with rising 

 tempei-ature. The water deepens in the cistern till the efflux equals 

 the supply. At that point a large increase of supply may not raise 

 the level much, since the leakage increases so fast as the water rises. 

 But a decrease of supply may lower the level a good deal if the 

 cracks are rather narrow below. If this be the case, an inter- 

 mittent supply will give a lower average level than a steady supply 

 of the same total amount, which answers to fluctuations in the supply 

 of heat lowering the mean annual temperature.' Again, if one of the 

 cracks widen upwards much more rapidly than the other, it will have 

 a greater share in producing this effect, and the loss through it will 

 be increased. Now as between radiation and evaporation it may be 

 shown that the latter has by far the fastest increase. The latter, 

 therefore, has the principal share in causing the extra loss which 

 follows from varying heat-supply. 



It may perhaps be asked, if the heat-loss from evaporation be 

 thus increased while the heat-supply is unchanged, will not the earth 

 be losing more than it receives ? Does not this conclusion conflict 

 with the equilibrium of emission and supply ? If we export more 

 than we import, must we not be growing poorer in heat? Those 

 who can see this difficulty will probably see the answer, that under 

 such circumstances less is lost by means of radiation : more flows 

 out by one leak, and less by the other. 



These reasonings seem to me to prove that increased eccentricity 

 must produce an increased amount of snow. But when we try to 

 calculate the possible amount, we are, as usual, pulled up by an 

 impossibility. This arises in the present case rather from the 

 uncertainty of the data than from any difficulty in the operation. 

 We may, however, form some rough notion of what is possible. 

 Basing a calculation on the assumption of evaporation proportional 

 to capacity of dry air for vapour, we find that a temperature fluctu- 

 ating 10 F. on each side of 80 F., would give about one-fortieth 

 more precipitation than would a steady 80 F. An equal fluctuation 

 on each side of freezing would give much the same fractional increase. 

 One-fortieth may seem small, but it must not be despised. If the 

 present annual Arctic snowfall were 20 inches, all by supposition 

 just melted, and if an uncompensated increase commence and 

 gradually grow during the 12,000 years which will interchange the 

 seasons and bring the summer of the Northern Hemisphere into 



