114 VISCOUS THEORY OF GLACIER MOTION. [1845. 



the uniformity of the triple curves exhibited in fig. 2 is surpris- 

 ing, considering the local errors to which the fixation of the 

 pins was liable, and the smallness of the quantities sought. 

 The first two curves, those for the 21st and 23d August, are 

 indeed as perfectly regular as it would be possible to expect from 

 this kind of observation, even much more so than I had ever 

 hoped to attain ; but on the 26th the holes containing the 

 pins were more degraded, and some manifest errors have arisen 

 from this cause, and evidently affect only single marks, such as 

 the twelfth and twenty-fifth mark, which singly have inclined 

 forward or backward, by the fusion of the ice. With these pre- 

 liminaries as to the reasons why the irregularities of these 

 curves should be judged [of] with indulgence, I will state briefly 

 the most apparent general results. 



First. The flexure of the ice is proportional, almost exactly 

 to the time elapsed in the intervals of the observations ; and it 

 is also graduated from point to point, not staccato, as would 

 inevitably have been the case had the relative motion been due to 

 the sliding of finite portions past one another, as in Plate II. 

 fig. 3. ' We perceive nothing of the kind. 



Secondly. A singular peculiarity strikes the eye, which at 

 first puzzled me, but when the cause was explained, confirms in 

 no slight degree the accuracy of the methods employed, and their 

 fitness to reveal the minutest motion of the glacier. The curves 

 of Plate IX. fig. 2 [of the Phil. Trans.], cut the axis, not all 

 exactly at the same point ; but on an average, this point may be 

 fixed with tolerable accuracy between the third and fourth mark, 

 or seven feet from the fundamental station Q. The first and 

 second marks moved evidently slower than the point Q, or the zero. 

 and the progress of the third and fourth is dubious or irregular. 

 The cause of this peculiarity was clearly ascertained on the spot 

 to be the existence of two crevasses belonging to the lateral 

 system of crevasses, between which, at their thinning out, the 

 station Q had been placed, under the idea that their distance 

 was sufficiently great not to affect the motion. The position of 

 these crevasses is shown in Plate III. fig. 3, by which it will 



