summer would exactly neutralize each other’s effects, and. on the 
average of years, no accumulation could begin. Primd facie there- 
fore high eccentricity will not account for Glacial periods. 
Dr. Croll does not deny that this is primd facie true, but he 
alleges that secondary actions will prevent the summer from thus 
neutralizing the winter. Here come in the four causes which we 
have already quoted as the bases of his theory. It may be thought 
that what has been said renders consideration needless. Since to 
set them at work, an accumulation of snow must have commenced, 
and since so long as no accumulation can begin they cannot operate, 
at least to any greater degree than at present, it should follow that 
any discussion of their operation is superfluous. This appears to be 
Mr. 8. V. Wood’s view, when he objects to the reasoning as a 
reasoning in a circle. But the objection is not altogether well 
founded. I think it can hardly be doubted, that though the summer 
would melt more ice than is now melted, the winter would form 
more than is now formed. At midsummer there might be less 
than there now is, but at midwinter more. We have as yet no 
prima facie reason for yearly increasing accumulation, but we have 
prima facie reason to think that the extreme cold of winter would pro- 
duce quantities of snow in excess of what we now experience, even 
after allowing for the quantities melted by the extreme heat of the 
previous summer. And this greater extension of midwinter snow is 
exactly what Dr. Croll requires for his causes. It is the force neces- 
sary to set his machinery in motion. The question now is, will the 
machinery work ? 
The First alleged reason is the cold produced by masses of ice and 
snow. If this means that the snow will not be melted easily, 
because it cools the air, there is a self-contradiction. Snow can 
only cool the air by abstracting heat from it, and if snow takes up 
heat, the heat is spent in melting it. As for what is said about cold 
rendering air diathermanous, so much the better for the melting. 
If the sun’s rays can readily pass through the air, they will the 
more readily reach the snow which they have to melt. These para- 
graphs are not worthy of the book. 
The Second cause alleged is that snow crystals tend to reflect the 
heat which ought to spend itself in melting them. This is perfectly 
true, an entirely valid argument, a real cause acting in the required 
direction. The matter is not altogether obvious. Suppose that at 
present over a given region snow is formed in one half the year and 
just entirely melted in the other half. Then, if owing to some 
change of circumstances twice as much snow is formed in the first 
half, but twice as much heat received in the second, then the snow 
would be entirely melted as before, and no accumulation would 
begin. But the case we are dealing with is different. At present, 
a certain amount of heat is lost during the year, but just as much 
received. Part of that lost is radiated off into space by water or 
vapour freezing into ice and snow. An equal part of that received 
is used up in melting this snow, and a certain additional quantity, 
_a fixed fraction of the above, is wasted by reflexion from the snow 
Rev. E. Hill—Eccentricity and Glacial Epochs. 13 
