March 6, 1879] 



NATURE' 



413 



presence. This effect would be expressed graphically by 

 the dipping of the isothermal of 39"2, as at G, and the 

 generation of those of temperatures intermediate between 

 it and 32" F. This alteration in the temperature means 

 also an alteration in the density ; and if we consider a 

 vertical section through the ice at H K and through the 

 middle of the lake at L B we shall find the mean density 

 at L B greater than at H K, and the result would be the 

 production of convection currents. What takes place 

 at E would take place on the other side at A also, 

 and we should have a system of circulation which in 

 broad outUne would consist of surface currents from 

 the sides towards the middle, and under ciurents from 

 the middle towards the sides, somewhat after the 

 manner indicated in Fig. 2. The length of time that 

 these currents would continue to flow before the lake \s'as 



covered over with ice would depend on a number of local 

 circumstances. The shape, size, and position of the lake 

 would hare much influence, and also the severity of the 

 frost. However low the temperature of the air, the 

 strength of the currents would always diminish as the 

 water got cooled down, for the lowest available tempera- 

 ture is 32° and the highest 39"2<*. Hence the convection 

 currents would be the strongest at the first formation of 

 the ice, and would gradually, and at an increasing rate, 

 get weaKer as the general temperature of the water got 

 lower. When the whole of the water was cooled to 32" 

 they would stop altogether. 



It is therefore certain that in the water of a frozen lake 

 we must find a tolerably imiform temperature, and this 

 temperature must lie between 39*2° and 32^. In order 

 that either of these extreme temperatures should 



J,-S 



J2 



prevail, the weather would require to be of a very 

 exceptional character. Admitting that the water had 

 been cooled to a uniform temperature of 39*2°, this 

 temperature could be fixed only by a sudden frost of 

 extraordinary and probably unknown intensity. As the 

 convection currents become weaker the more nearly the 

 temperature of the water approaches 32^ it would require the 

 least possible frost during an infinite winter to reduce the 

 temperature of the water to 32° before it was covered 

 with ice. The uniform temperature of 32^ however could 

 be produced in another way, namely by the cooling of 

 he water after it was covered with ice. This condition 

 is probably not uncommon in shallow lakes in very cold , 

 climates. i 



In general, in climates such as our own, an inter- 

 mediate temperature would prevail. In the Balloch basin 

 of Loch Lomond this temperature was found to be about 



33"9° F. In Linlithgow Lo^h it was much higher, about 

 yf F. The lake ^vlll remain open with an ice-fringe 

 along the side as long as the water leaves the ice-edge 

 with sufficient velocity to mix with the warmer water off" 

 shore before freezing under the influence of surface 

 radiation. As soon however as this velocity is reduced 

 so far as to enable the frost to overtake the water as it 

 leaves the ice-edge before mixing with the warmer water 

 the propagation of ice from the fringe out into the middle 

 of the lake will take place with great rapidity, and a 

 single night will often be suflficient to cover a large lake. 



From the moment of the formation of the complete ice 

 covering, the water is subjected to a uniform climate, its 

 surface being everj'where in contact with ice ; and it is 

 only under these conditions that the whole of the lake can 

 be said to be exposed to an identical cUmate. 



Hence we see that, even admitting that a uniform tempe - 



rature of 392° had previously existed, the whole of the water 

 would be very considerably cooled down before it could be 

 covered with ice, and the extent of this cooling would de- 

 pend on local circumstances of position and chmate, so that 

 the final temperature of the water would be difterent for 

 different lakes in the same winter, and for the same lake 

 in different winters. The body of the water of a lake 

 would be cooled more when it has bsen frozen by a 

 moderate and comparatively long continued frost, than 

 when the ice has been frozen quickly by very severe frost. 

 For the more severe the frost the sooner will it be able to 

 overtake the water leaving the ice fringe ; in other words, 

 the stronger will be the current which it will be able to 

 arrest, and the greater the Jiead which it will be able to 

 stem. But the head which causes the current, is caused 

 by the higher temperature of the open water as compared 

 with that under the ice. Hence the more severe the.frost, 



the higher will be the temperature which it will be able to 

 fix. Let us now consider the causes which would tend to 

 alter the temperature fixed by the ice covering. 



In the first place there is alteration in the thickness of 

 the ice and conduction from it downwards. The effect so 

 produced would be concentrated in the surface layer, and 

 its nature can be easily imagined. 



In a large lake like Loch Lomond there is, even in 

 severe weather, a considerable amount of change of water 

 going on. The river Leven, which leaves it at Balloch, is 

 navigable for small vessels, and a number of not incon- 

 siderable streams fall in at different parts of the loch. 

 The ice extended from the lower end at Balloch up to 

 Luss, where it terminated in a miniature cliff which swept 

 in a cur\-e from the Dumbartonshire shore to the island 

 of Inchlonaig and thence in another cur\'e to the Stirling- 

 shire shore. It is easy to see that the streams which fall 



