190 MK. K. D. OLDHAM ON THE [May I9OI, 



draining a considerable area, and consequently of considerable 

 volume, has cut a deep valley and is then beheaded by another 

 stream which, favoured by a steeper gradient, or cutting back 

 along a band of soft rock, penetrates the drainage-area of the first 

 and directs its headwaters to a new course. In this case, after a 

 gradual ascent along an open valley, there is an abrupt descent 

 from the summit-level to a deep-cut valley. Such is not the case 

 in the Dunmail Eaise. 



Another form of wind-gap will not show an abrupt descent into 

 a larger valley at the head of the gap, but will descend from the 

 summit-level through (what Prof. W. M. Davis has called) an 

 obsequent valley to a main valley crossing the head of the gap : 

 an obsequent valley being one which has been cut back along the 

 beheaded valley, on account of the lesser resistance oifered by it, 

 as compared with the higher ground on either side. In this case 

 there will be an open valley on the one side of the summit-level 

 occupied by no stream at all, or by a stream which is too small for 

 it ; on the other side the obsequent valley, having been formed by 

 the stream which occupies it, will fit the stream that it carries and 

 will be narrower and steeper-sided than the beheaded vallej^ beyond 

 the summit-level. This again is not the case with the Dunmail 

 Raise. 



A third mode in which a gap may be formed by erosion is where 

 two streams opposite each other cut back into the same mass of high 

 ground. Here the two valleys cutting back against each other will 

 cause a notch to be cut in the ridge where they meet, but in such a 

 case the summit of this notch will be narrow, if not sharp, and the 

 slopes on either side steep. In an old land-surface which has 

 long been exposed to denudation, and where all the surface-forms are 

 rounded off by weathering, a notch with a cross-section not unlike 

 the Dunmail Eaise might originate in this way ; but then the pro- 

 file of the watershed would have a very different form, and in a 

 hilly country of steep slopes and crnggy summits the notch would 

 have a shape much more like Striding Edge than the Dunmail Eaise. 

 Moreover, the streams draining from both sides of the watershed 

 would fit their valleys, which is not the case in the Dunmail 

 Eaise. 



In short, wherever a wind-gap is formed by the recession of 

 watersheds, consequent on stream-action, the streams on one or 

 both sides of the watershed must fit their valleys ; and it is im- 

 possible in this way to form a gap in which the valleys on both 

 sides of the summit-level are too large for the streams which now 

 flow in them. 



Another mode of origin which might be, though I am not aware 

 that it has been, suggested, is that the gap was carved out by an 

 ice-sheet. Against this hypothesis may be placed, first of all, the 

 absence of any evidence of an ice-sheet having ever swept across 

 these mountains ; and secondly, a glacial origin seems precluded 

 by the fact that at both ends the gap is crossed by the lateral 



