RIVERS TEVIOT, NITH, AND CLYDE. 465 



it being well known that the atmospheric temperature diminishes in a certain 

 ratio with the height.* In the next place, the temperature of the ground in the 

 higher and more exposed districts would, from the violence of the wind blowing 

 on them, approximate much more rapidly to that of the air than it would do in 

 the lower and more sheltered districts. So that, when the former had fallen to 

 27° or 28°, the latter would scarcely have reached the freezing point. From these 

 united causes, the water oozing through marshy ground, or trickling along open 

 drains, or in the channels of streamlets on the hills and muirs, would be suddenly 

 congealed and arrested, whilst the larger body of water in the bed of the rivers 

 would continue to flow on unobstructed, and thus effect an entire drainage of the 

 channel. 



It is the last of these causes to which, more particularly, I ascribe the pheno- 

 menon which has formed the subject of this paper. It is in itself sufficiently sim- 

 ple, and depends on the plainest principles. 



The nature of this cause, serves also to explain the unfrequency of the phe- 

 nomenon. The severe ft^osts in this country are seldom accompanied with gales 

 of wind. Our gales are generally from the S. or SW., bringing with them warm 

 vapours, which are exclusive of frost ; and even when we have easterly gales, 

 there is seldom sufficient dryness to admit of a very low temperature. But, what- 

 ever be the cause, it is an undoubted fact, that, in this country, during the pre- 

 valence of severe frost, the air is comparatively calm. Under these circumstances^ 

 although water at the sources of the rivers must always be more rapidly cooled 

 than in the lower parts of their course, the difference in their respective rates of 

 cooling cannot be nearly so great as in a gale of wind, which affects only the ele- 

 vated and unsheltered districts. In an ordinary frost, the streams, and especially 

 the springs at the sources of the rivers, are seldom frozen without there being ice 

 formed in the current of the main channel. When the frost continues a sufficient 

 length of time, ice will be formed on the surface of the current where it is deep 

 and sluggish, and at the bottom as well as at the top where it is shallow and 

 rapid. In this way the flowing of the current becomes obstructed ; so that, in 

 these circumstances, the river would actually appear more full than usual, were 

 it not that, in consequence of the freezing of the streams at or near the sources of 

 the river, much of the supply of water to it, is cut off". In the case of an ordinary 

 frost, therefore, that is, when it acts with pretty nearly equal effect, both on the 

 sources and on the main current, there will be no drainage of the channel. It is 

 only when the frost is accompanied with a high wind, that it is enabled to affect 

 the sources before it has had time to freeze, the larger bodies of water having a 

 rapid motion, flowing at a lower level, and in a sheltered situation. 



There is only one other point necessary to be adverted to, in order to com- 



* The temperature sinks 1° of Fahr. for about every 350 feet. 

 VOL. XIV. PART II. 4 D 



