Vol. 51.] FLOW OF A VISCOUS FLUID. 367 



Theoretical explanations founded on conjectures as to the manner 

 in which ice-sheets might possibly rise in ascending currents have 

 been met by theoretical objections ; the experiment just described 

 plainly shows, however, that under certain conditions such an 

 ascending movement not only may, but undoubtedly must take place. 

 The slope of the upper surface of the pitch in my experiment is 

 naturally much greater than will generally be presented by the 

 surface of an ice-sheet, and the viscosity of ice is much greater 

 than that of pitch ; but these differences will not affect the final 

 result. The viscous nature of glacier-ice once fairly admitted, it 

 necessarily follows that, given time enough, this ice is bound to 

 find its own level, in precisely the same sense that water is said 

 to do so. Given sufficient time, and under tbe vertical pressure to 

 which it is subject, an iceberg will (theoretically) flatten out into 

 an infinitely thin film, and the reason that polar ice-sheets termi- 

 nate in steep cliffs is that the ice breaks off faster than it flows out. 

 There seems, then, no reason to doubt that a mass of ice occurring 

 under conditions similar to those of our experiment would move in 

 the same way : and without necessarily committing myself to the 

 hypothesis that stones have actually been carried upwards from 

 their source by ascending currents of ice, I am compelled to admit 

 that, given the necessary conditions, its possibility cannot be denied. 

 If we turn to the observations in the Isle of Man, first made by 

 Mr. Cumming, and lately confirmed and extended by Mr. P. F. 

 Kendall, we find that fragments have been carried from Granite 

 Mountain (fig. 5, p. 366), which does not attain a greater height than 

 679 feet, on to the flanks of South Barrule, up to 100 feet below 

 its summit, which is 1588 feet above the sea-level. Wow if the 

 Isle of Man were, as extreme glacialists suppose, hemmed around 

 by confluent glaciers of considerable thickness, not necessarily over- 

 whelming Snaefell, it would appear well within the range of possi- 

 bility that the snowfall on Snaefell and the mountains to the south 

 and west of it may have produced an ice-flow with an upper surface 

 somewhat as in the diagram on the opposite page. Between South 

 Barrule and Creelea Mountain there would have been situated a great 

 lake of ice under a higher pressure at its northern than its southern 

 end, and the natural result would have been an ascending forward 

 movement on the south : this we may suppose, without being 

 guilty of an absurdity, capable of transporting boulders from 

 Granite Mountain into the position that they now occupy on South 

 Barrule. 



Those who treat the absence of evidence as evidence of absence 

 have asserted that by no possible process can stones be introduced 

 from the floor of a glacier into its substance, and it is possible 

 that they might until quite recently have objected that, while ice 

 might ascend South Barrule, it could not pick up fragments from 

 Granite Mountain ; fortunately we are now in possession of Mr. 

 Chamberlin's observations on the glaciers of Greenland, 1 which not 



1 Bull. Geol. Soc. America, vol. vi. (1895) pp. 199-220. 

 Q. J. G. S. No. 203. 2 d 



