ICE-EROSION m THE CUILLIN HILLS, SKYE. 237 



cuspate septum left by the excavation of two opposed cirques, so each culminating 

 peak is in general the triangular pyramid left in the midst of three such cirques. 

 The form is not strictly pyramidal, for the cirques are concave in horizontal as well 

 its in vertical section, so that the ground plan of each peak comes to have the outline 

 of a tricuspate curve which is highly characteristic. This is well shown by the 

 3000 feet contour-line on the map given below. Bidein Druim nan Ramh and 

 Sgiirr a' Ghreadaidh, standing each in the midst of four corries, are four-cusped 

 instead of three -cusped. The former of these mountains and the subsidiary spurs 

 of the latter do not reach 3000 feet, but their shapes are partly indicated on the map 

 by the 2000-feet line. 



(viii.) Longitudinal Profile of Valleys : Lake-Basins. 



We have next to examine the form of the valleys in longitudinal section along 

 tlie actual main drainage- line or 'Thalweg'; and we observe at the outset that 

 this line is by no means always concave upward, nor does its declivity show anything 

 like a steady diminution from the head to the outlet of the valley. If we may take 

 the Cuillin district as a type, it appears that ice-erosion does not, like water-erosion, 

 work constantly towards the establishment of an even gradient along a valley in which 

 it operates. It tends, not to reduce, but to exaggerate within certain limits the more 

 marked inequalities of the longitudinal profile ; and in some circumstances it may 

 set up a negative gradient in a certain portion of the valley. 



The bed of a river which has attained a mature state maintains a steady 

 gradient so long as the volume of water is unchanged, and the gradient diminishes 

 down-stream in a manner proportioned inversely to the increasing volume. This law 

 is a consequence of the relations which necessarily subsist between declivity, velocity, 

 volume, and load. Perhaps the clearest presentation of the argument is that in 

 Gilbert's Geology of the Henry Mountains ; and a glance over his treatment of the 

 subject is sufficient to show that these fundamental principles in the case of water 

 have no counterpart in the case of ice. On the other hand, it appears that in 

 ice-erosion certain other principles will come into operation which are peculiar to this 

 agency. Thus we may expect that, other conditions being the same, erosion will be 

 most efficient where the pressure below the ice is greatest, i.e., where the thickness is 

 greatest. If the longitudinal profile of a valley be of irregular form, while the upper 

 surface of the ice declines steadily, the thickness will be greater over parts which have 

 re-entrant or concave forms than over adjacent parts which have salient or convex 

 forms ; and, on the principle laid down, differential erosion will operate so as to 

 exaggerate the inequalities. The original inequalities postulated must have a certain 

 magnitude. On the other hand, the condition that the upper surface of the ice declines 

 steadily implies that they must not be too great in proportion to the thickness of the 

 ice. Within these limits it appears that the steady gradient, which for water-erosion 



VOL. XL. PART II. (NO. 12). 2 N 



