136 VISCOUS THEORY OF GLACIER MOTION. [1846. 



precisely as in rivers of great magnitude and length of course, 

 and of moderate declivity, it yields sluggishly to impulsive or 

 retarding forces which are checked and opposed by the multi- 

 tude of sinuosities, the embaying of the ice in rock-bound 

 expansions of the channel, the struggle of its passage through 

 defiles, and the enormous friction of its lower surface. Yet, 

 lest we might attribute the irregularities of the torrential glacier 

 to causes quite local and uncertain, we find them reflected more 

 or less distinctly in the movements of the neighbouring one. 

 Thus the anomalous retardation in the end of March and be- 

 ginning of April appears in three stations out of four, as does 

 that in the first half of June, showing clearly that it is not an 

 error of observation. It appears that the thaw of the winter's 

 snow during the month of May, saturating the pores of the 

 glacier with water, produced (as we know that a thaw always 

 does) a sudden and violent march, especially of the more sus- 

 ceptible or torrential glacier. So completely had this sudden 

 move forced on the~glacier of Bossons, encumbered by the spring 

 avalanches, and loaded with all the fragments and snow masses 

 which had remained temporarily suspended during the winter 

 months, that the lower part of the glacier (as we read in the 

 memoranda to the register) advanced and widened greatly, to 

 an extent which it had not done for many years past, and 

 seemed to change its whole character ; and in February a 

 similar temporary increase of volume had taken place ; " on ne 

 s'y reconnait presque plus," writes Balm at ; thus accounting 

 for the particular accession of speed which appears in that 

 month. In both cases, after the rapid march in February and 

 in May, a reaction takes place ; the material is deficient, the 

 excessive pressure has been removed by the previous overflow, 

 and a lull occurs in March and in June. 



VII. These irregularities, such as they are, even should we 

 fail in entirely explaining them, are at least not to be attributed 

 entirely to errors of observation, since different observations 

 (which, it is to be recollected, were sent to England in so rough 

 a state that they required to be reduced and computed before 



