Prof. Challis on the Principles of Theoretical Physics. 313 



house was attached. Air slightly warmed, acting on the glass 

 surface from underneath, melted the snow in immediate contact 

 with the glass, and the layer in consequence slid slowly down 

 the glass roof. The inclination of the roof was very gentle, and 

 the motion correspondingly gradual. When the layer overshot 

 the edge of the roof it did not drop off, but bent like a flexible 

 body, and hung down over the edge for several inches. The 

 continuity of the layer was broken into rectangular spaces by the 

 inclined longitudinal sashes of the roof, and from local circum- 

 stances one side of the roof was warmed a little more than the 

 other; hence the subdivisions of the layer moved with different 

 velocities, and overhung the edge to different depths. The bent 

 and down-hanging layer of snow in some cases actually curled 

 up inwards. 



Faraday has shown that when small fragments of ice float on 

 water, if two of them touch each other they instantly cement 

 themselves at the point of contact ; and on causing a row of 

 fragments to touch, by laying hold of the terminal piece of the 

 row you can draw all the others after it. A similar cementing 

 must have taken place among the particles of snow now in 

 question, which were immersed in the water of liquefaction near 

 the surface of the glass. But Faraday has also shown that, when 

 two fragments of ice are thus united, a hinge-like motion sets in 

 when you try to separate the one from the other by a lateral push : 

 one fragment might in fact be caused to roll round another, like 

 a wheel, by incipient rupture and the re-establishment of rege- 

 lation. The power of motion thus experimentally demonstrated 

 rendered it an easy possibility for the snow in question to bend 

 itself in the manner observed. The lowermost granules, sub- 

 jected to pressure when the support of the roof had been with- 

 drawn, rolled over each other without a destruction of continuity, 

 and thus enabled the snow layer to bend as if it were viscous. The 

 curling up was evidently due to a contraction of the inner surface 

 of the layer, produced, no doubt, by the accommodation of the 

 granules to each other as they slowly diminished in size. 



Waverley Place, St. John's Wood, 

 March 21, 1862. 



XLIV. On the Principles of Theoretical Physics. By the Rev. J. 

 Challis, M.A., KR.S., F.R.A.S., Plumian Professor of 

 Astronomy in the University of Cambridge *. 



THE great progress that is being made at the present time 

 in experimental philosophy is remarkable in respect both 

 to the skill and ingenuity displayed in making the experiments, 

 * Communicated by the Author. 



Phil Mag. S. 4. Vol. 23. No. 151. April 18G2. Y 



