demonstrated—no such deflection of the Atlantic current would be likely to occur. This possibility is fairly confronted by Dr. Croll when he admits that the gravitation theory "militates strongly" against his theory of secular changes of climate. But there is another self-evident proposition which seems to have escaped his attention. Assuming that the results detailed above follow in the order laid down when the eccentricity is at a high value, it is perfectly manifest that the deflection of the Gulf-Stream feeder is an effect of glaciation, and, if a cause at all, only a secondary one. Hence, if the purely physical agencies alone are capable of causing glaciation, Dr. Croll's theory of secular changes in climate will stand, whether or not the wind theory of oceanic circulation is correct; but, if they are not capable of producing a glacial period alone, the theory will fall, even if the wind theory of oceanic circulation be correct. This can not be too strongly emphasized. Yet, not only in "Climate and Time," but in his other publications on the subject, Dr. Croll dwells upon the deflection-of ocean-currents as the principal telluric element in his theory!
The only positive evidence adduced to prove that the Gulf Stream was deflected during the glacial period is the fact, pointed out by Mr. Crosskey, that there is more difference between the glacial and recent shells of Scotland than between the glacial and recent shells of Canada. But this only proves that the present temperature of that part of Canada is lower than that of Scotland; for the temperature of the waters at the edge of the ice-sheet must have been approximately the same, whatever the latitude to which it extended. This evidence is, therefore, utterly valueless. Furthermore, we have pretty reliable positive evidence that the Gulf Stream was not stopped during the glacial epoch, in the more northerly limits of continental glaciation in those parts of Europe so greatly affected by the Gulf Stream to-day, any more than in the United States, where its influence is comparatively unfelt. Moreover, Professor Dana has shown in his "Journal" that the distribution of ice during the glacial period coincided in a general way with the present distribution of rainfall in the same latitudes. Now, the greater part of the moisture of the United States, especially of the great Mississippi Basin, is derived from the Gulf of Mexico; and the stoppage of the South Atlantic feeder of the Gulf Stream would cool the waters of the Gulf so considerably as to materially diminish our vapor-supply, and at the same time the distribution would be altered. Similarly, the stoppage of that current during the glacial period would have so altered the distribution that its relation to the present precipitation would not be recognizable, even if it did not so completely cut off the vapor-supply as to prevent glaciation. The mass of evidence is therefore against the hypothesis of the shifting of this important marine current during the glacial period; and, as this would, as Dr. Croll points out, be likely to be the first out of all of the ocean-currents to be deflected, it may reasonably be doubted whether
