206 Rev. H. Moseley on the Mechanical Possibility [Jan. 7, 



shear necessary to its descent by its weight only in the natural channel 

 must be less than that. 



A second possible difference between the case supposed and the actual 

 case lies in this, that the velocities of the surface-ice at different distances 

 from the edge, and at different heights from the bottom, are assumed to be 

 proportional to those distances and heights ; so that the mass of ice at any 

 time passing through a transverse section may be bounded by plane sur- 

 faces, and have a trapezoidal form. This may not strictly be the case. 

 All the measurements, however, show that if the surfaces be not plane, 

 they are convex downwards. In so far therefore as the quantity of ice 

 passing through a given section in a day is different from what it is sup- 

 posed to be, it is greater than it. A greater resistance (other than shear 

 ing) is thus opposed to each day's descent, and also a greater weight of 

 ice favours it ; but the disproportion is so great between the work of the 

 additional resistance to the descent, and that of the additional weight of 

 ice in favour of it, that it is certain that any such convexity of the trape- 

 zoidal surface would necessitate a further reduction of the unit of shear, to 

 make the weight of the actual glacier sufficient to cause it to descend. 



A third difference between the actual glacier and the imaginary one, to 

 the computation of whose unit of shear the following formulae are applied, 

 is this — that the formulae suppose the daily motion of the surface of the 

 glacier and the daily motion of its side to have been measured at the same 

 place, whereas there exist no measurements of the surface motion and the 

 side motion at the same place. The surface motion used has been that of 

 the Mer de Glace at Les Ponts, and the side motion that of the Glacier 

 du Ge'ant at the Tacul — both from the measurements of Prof. Tyndall. 

 This error again, however, tends to cause the unit of shear, deduced from 

 the case of the artificial glacier, to be greater than that in the actual one ; 

 for the Glacier du Ge'ant moves more slowly than the Mer de Glace. The 

 quantity of ice which actually passes through a section at Les Ponts is 

 therefore greater than it is assumed in the computation to be, whence it 

 follows, as in the last case, that the computed unit of shear is greater than 

 the actual unit of shear. 



To determine the actual value of fx (the unit of shear in the case of 

 ice) the following experiment was made. Two pieces of hard wood, each 

 three inches thick and of the same breadth, but of which one was con- 

 siderably longer than the other, were placed together, the surfaces of con- 

 tact being carefully smoothed, and a cylindrical hole, l^inch in diameter, 

 was pierced through the two. The longer piece was then screwed down 

 upon a frame which carried a pulley, over which a cord passed to the 

 middle of the shorter piece, which rested on the longer. There were 

 lateral guides to keep the shorter piece from deviating sideways when 

 moved on the longer. The hole in the upper piece being brought so as 

 accurately to coincide with tfcat in the lower, small pieces of ice were 



