Observations on Glaciers. 

 of the transverse section of the glacier, and that it is not un- 

 frequently entirely destroyed in one part of the glacier, to be 

 renewed in a totally different direction in another. A mole- 

 cule of ice is as passive and structureless a unit as a molecule 

 of water, so far as it has not that structure impressed by some- 

 thing external at the time. Like the water in the river, my- 

 riads succeed one another, and might be mistaken for the 



same. 



Few words will suffice to shew how intimately what I have 

 stated is connected with the first rudiments of a theory of gla- 

 cier motion, which I endeavoured to sketch m my last letter 

 and the truth of which all that I have since seen has tended 

 ereatly to confirm. The centre of the glacier stream is m-ged 

 onwaiis by pressure from above (how caused we shall im- 

 xuediately consider}, which is there resisted less than at the 

 sides and bottom, owing to the comparative absence of fric- 

 tion The lateral parts are dragged onwards by the mo- 

 tion'ofthe centre, and move also, but it is quite compatible 

 with this idea of semifluid motion, that the bottom of the 

 glacier should remain frozen to its bed, as some writers 

 have supposed to be the case, though I am far from as- 

 serting this to be the fact, or even supposing it probable. 

 Why then, are the fissures generally vertical, and also where 

 a glacier is most regular, simply transverse, and not con- 

 vex towards the lower extremity .^ The first of these ques- 

 tions had always till lately appeared to me a serious diffi- 

 culty Th&fai-t stated in the second, combined with the posi- 

 tive certainty that the centre €f a glacier moves faster than 

 its sides, in the ratio frequently of 5 to 3, shews that an an- 

 swer mmt be found, and, therefore, that it offers no msur- 

 mountable objection. The explanation is to be sought m the 

 continuallv varying condition of the glacier, the perpetual re- 

 newal of the crevasses, the action of water in tending to pre- 

 serve verticalitv, and the really small variation of velocity of 

 different parts of the ice towards the centre of a glacier of im- 

 mense depth. From these circumstances, it follows that a 

 crevasse is either renewed or altogether extirpated before its 

 verticality is sensibly effected. For the same reason, a stick 

 several feet long, inserted vertically in the ice, remains sensi- 



