June. 1894. NOTES ON GROUND-ICE. 451 



popularity, a result probably of its greater ease of comprehension, as 

 much as of the observations to which it was conjoined. There are, 

 indeed, certain facts which seem to support this theory, such as the 

 general consensus of observers that ground-ice is only produced in 

 clear water, and on cloudless nights, when radiation is most active ; 

 and Mr. Farquharson in his paper lays stress on the fact that it was 

 not found near the piers of a bridge, close to an embanking wall, or 

 under a grass-clad bank. This is, however, explicable in other ways 

 than by radiation having been checked by these objects, and as regards 

 the first part of the argument the coincidence may well be due to the 

 fact that sharp frosts seldom occur in temperate latitudes except in 

 cloudless weather, while the clearness of the water is as much a 

 necessity of the first-named theory as a support of the second. More- 

 over, if there were any truth in the assumption, that the clear water 

 of a stream flowing over a stony bottom is cooled to any important 

 extent by radiation from the stony bottom over which it flows, this 

 would be equally true in the case of standing water. But such we 

 know is not the case, and it is the fact that the loss of heat takes place 

 principally from the surface of the water, combined with its expansion 

 and consequent decrease of density as the freezing point is approached, 

 that makes our ponds freeze first on their upper surfaces. 



Against the other theory, that the pebbles on the bottom merely 

 act as nuclei on which the ice crystallises out from water cooled 

 down below freezing point, it might be urged that this cannot occur 

 with water such as is met with in nature, but only after it has been 

 distilled or otherwise cleared of all dissolved air and solid particles. 

 This, however, is not altogether the case. I have myself observed 

 the cooling of water below freezing point when no special precau- 

 tions had been taken. While camping in the desert of Western 

 Rajputana it was a common experience to find, after a clear and 

 cloudless night in January and February, that the water in my wash- 

 hand basin was frozen over, and frequently frozen solid. On one 

 occasion, after a cold night, I was surprised to find that the water 

 was unfrozen, and, thinking that perhaps the night had not been so 

 cold as I imagined — I had not then examined my thermometer — I 

 plunged my hands into the water, which immediately became con- 

 verted into a pasty mass of ice crystals and water. Here ordinary 

 well water, in a basin which was only clean in the common sense 

 of the word, had cooled below freezing point without solidifying ; 

 but the occurrence was an exceptional one, and it may well be 

 doubted if it could take place as an ordinary event in the waters of a 

 stream. 



There is yet a third theory, originally published in i8i6,^ which 

 appears to have been almost completely lost sight of since, that 

 ground-ice is due to the presence of numerous minute crystals of ice 

 in the ice-cold water ; these are carried to the bottom of the stream 



^ T. A. Knight, Phil. Trans., iai6, pp. 286-293. 



2 G 2 



