788 SoLLAS — A Map to slioiv the Distribution of Eskers in Ireland. 



Finally the late Professor Carvill Lewis has designated as marginal kames* 

 certain eskers in Pennsylvania which are associated with the terminal moraine, 

 and which were formed by a subglacial drainage in the direction of the local 

 valleys, but in the opposite direction to the flow of the ice. 



The literature treating of the origin of eskers is very extensive ; but we shall 

 considerably shorten our account if we dispense, as we fairly may, with the treat- 

 ment of those theories which invoke the action of the sea as a probable cause. 

 Professor Carvill Lewis has, indeed, summarily dismissed such theories as 

 " antiquated," and they are certainly too well known to require fresh presenta- 

 tion. It may be further remarked that they account for scarcely any of the 

 features which are most characteristic of eskers, and they are inconsistent with 

 some known facts. Since explanations depending on the effects of glacial rivers 

 are free from these objections, there is every reason why we should restrict our 

 attention to them. 



The first to propose, or rather to adumbrate, a glacier theory of eskers was 

 Mr. N. H. Winchell.f His words are as follow : — " The effect of such a persistent 

 obstruction [i.e. of a ridge of Niagara Limestone] beneath the glacier must have 

 been to fracture the ice profoundly, those parts toward the east and toward the 

 west settling gently away from the uplifted centre. Into these crevasses the 

 drift would fall, and through them streams of water would flow. The result 

 would be an extraordinary accumulation of coarse and assorted drift materials, 

 which, after the complete withdrawal of the ice, would lie in irregular knolls and 

 short ridges in places where such streams formerly existed. This is the origin of 

 a great many short, but steep and narrow, gravel ridges in the North-western 

 States outside Ohio, which are known locally as ' Devils' Backs,' or ' Hogs' Backs.' " 

 ..." They generally rise from 25 to 40 feet, with slopes as steep as such material 

 can be piled into. On either side there is usually a low, swampy tract, from 

 three to ten rods wide, or sometimes of indefinite width and form, the ridge being 

 in some cases not more than 20 feet across on the summit. The largest boulders 

 are sometimes seen on the very top of the ridge." 



It is unfortunate that Professor Winchell does not identify these gravel ridges 



o 



with eskers, kames, or asar; the more so, as some pages further on (p. 181), he 



thence to Meadhraiglie, a peninsula extending into the Bay of Galway." See also the (earlier) Annals 

 of Clonmacnois, under Conn of the Hundred Battles, whose kingdom stretched southwards to the Esker 

 Eiada. The " phonic reasons " are certainly worth considering ; every English-speaking geologist knows 

 how to pronounce the word esker, hut few seem to he aware of the correct pronunciation of asar ; it is not 

 quite " osar " as seems to he generally supposed. The grammatical form is a still more serious objection 

 to tlie use of the Scandinavian term, for we do not readily form our plurals in ar. 



^Marginal Kames. — Proc. Acad. Nat. Sci., Philadelphia, 1885, p. 157. 



t The Surface Geology of North- Western Ohio. By N. H. Winchell. — Proc. Amer. Assoc. Adv. Sci., 

 vol. xsi., 1872 (1873), p. 165. 



