RIVERS TEVIOT, NITH, AND CLYDE. 4g3 



of the miller on the Clyde, already read, who states that it frequently becomes so 

 thick as to fill up the interstice between the bottom of his sluice and the channel 

 of his dam. I believe that this is the origin also of that kind of ice having much 

 the appearance of melted snow, which often floats down our rivers in great abun- 

 dance, and which, in Dumfriesshire and Roxburghshire, is known by the name 

 of " grue." It is not retained long enough at the bottom to be formed into a 

 solid cake of ice, but by its lighter specific gravity, is able to separate itself from 

 the weeds or mud at the bottom, and rise to the surface of the current. 



But to return from this digression, let me repeat, that the formation of this 

 ground-ice cannot, by obstructing the flow of the stream, have the effect of lower- 

 ing its level, and far less of drying it up. The very opposite effect would follow. 



3. It must be obvious from these remarks, that no explanation can be the 

 true one, which assumes that the quantity of water in the rivers on this occasion 

 remained the same, and was only stopped or arrested in its flow. We must seek 

 for some explanation which involves, as a principal element, the actual abstrac- 

 tion or withdrawal of the current from the bed or basin of the river. Now, this 

 condition would be accomplished, if, whilst the supply of water at their sources 

 was stopped, the channels in the lower parts of the river remained open, so as to 

 permit the main current to run off", and thus drain the bed of the river. If the 

 frost acted in such a way, as to seal up and stop the flow of the springs at the 

 fountain-head, and yet allow the stream in the deeper channels to flow as before, 

 then it will not be difficult to see how the phenomenon in question occurred. It 

 is proper, therefore, to inquire whether this theory is at all supported by observa- 

 tion, and whether it can be justified on known principles. 



That this proposed explanation is, to a certain extent at least, warranted by 

 observation, is manifest from the letters above quoted ; in all of which it is stated, 

 that the rivulets, and even the springs, were frozen up and arrested in their flow 

 on the night of the 26th and morning of the 27th November. 



That this should have been the case, when the thermometer sunk in the 

 course of that night to 26° or 27° Fahr., is quite intelligible. The only apparent 

 difficulty is to discover in what way the water of the river could be frozen at the 

 sources, whilst it continued to flow in the deeper channels. It has been shewn 

 that the thermometer at 3 p. m. on the 26th stood at 32°, that at 1 p. m. it sank 

 to 26°, and that by 10 a. m. on the 27th November it had returned to 32°. It 

 may be assumed then, that, on the night of the 26th, there was no frost of any 

 degree of severity which lasted longer than eight or ten hours. If, then, the 

 stoppage of the supplies to the rivers in question was affected by the frost, was 

 there any thing which enabled it to produce, during the short period of its con- 

 tinuance, a greater effect on the small streamlets at the main sources of the rivers, 

 than on the main body of the current in the larger channels ? It is not difficult to 



