the Beds of Running Streams, 341 



in the place where it is found. I had occasion to submit the 

 methods of accounting for ground ice by Dr. Farquharson and 

 M. Arago to my friend Mr. James Elliot, then resident on the 

 banks of the river Teviot ; he there tested the theories of the 

 formation of ground ice by the facts presented to him through 

 a careful examination of the bed of the river_, and came to the 

 conclusion that they did not agree well with one another. After 

 a lapse of some time I submitted to him a new explanation of 

 the phsenomenon of the appearance of ground ice, which was 

 based on details given in Mr. Knight and Mr. Eisdale's published 

 papers on the subject*, and also from the valuable evidence sup- 

 plied by Mr. Elliot against the theories in question. The theory 

 I gave of the formation of ground ice was, that it was all formed 

 at the surface^ being the coldest pai^t of a running stream, sub- 

 merged by the motion of the current, then carried along until 

 entangled by some plant or asperity in the bed of the stream. 

 An icy nucleus once begun, the trite illustration of a snow-ball 

 rolling down a hill describes what follows in a river at 32° tem- 

 perature, having minute particles of ice mingled with the water, 

 the process of accumulation, however, being inverted ; for in the 

 stream, the nucleus which corresponds to the snow-ball remains 

 stationary, while the icy particles corresponding to the snow are 

 rolled over it by the motion of the river. When the ice is lodged 

 in a plant moored to the ground by a slender stem, the accumu- 

 lation assumes a spherical form, from the twisting and turning 

 of the plant in the current ; but where the asperities are fast to 

 the bottom, the ice spreads over the bed of the stream after the 

 manner of a covering of snow on rough ground. The velocity 

 of the water has an influence on the nature of the particles of 

 ice which form the accumulation ; in slow running streams, like 

 those in the vicinity of Liverpool, it consists of thin scales of 

 nearly one-quarter of an inch superficies; pieces of this size 

 interlocked with one another have under water a transparent 

 appearance, so that to a casual observer they are easily overlooked 

 even when in large quantities in the bed of the river -, where the 

 water runs swiftly, the particles of ice become much smaller, and 

 when collected together assume the appearance described by 

 Mr. Knight, of the most silver whiteness. 



In the theories offered to account for the formation of ground 

 ice in the part of the stream where it is found, the motion of the 

 water is stated to prevent the formation of ice at the surface, and 

 the particles on the ground being nearly in a state of rest the water 

 freezes there. This view proceeds on the assumption that motion 

 in water is a preventive to freezing, which is only true in a 



* Abstracts from these will be found under article Ground Ice, Penny 

 Cyclopsedia. 



