106 Mr. H. C. Sorby on the Freezing-point 



ture far below the freezing-point of water; and then quickly 

 placing it in a saturated solution of common salt previously re- 

 duced to a very low temperature, I was much surprised to find 

 that the liquid in the fluid-cavities had not been frozen, but the 

 enclosed bubbles still moved about like those in spirit-levels, 

 when the fragment was turned round, just as they did at first. I 

 again repeated this experiment along with Dr. Tyndall ; and we 

 both came to the conclusion that the fluid would not freeze at a 

 temperature nearly 20° C. below the freezing-point of water, or 

 even lower than that of strong saline solutions, and therefore 

 there was some doubt as to whether it really could be water or 

 some saline solution. I, however, had such confidence in the 

 value of other independent proofs, that I still adhered to the 

 idea of its being water, and that the pheeuomenon must be owing 

 to some peculiarity in the freezing of water in minute cavities. 

 I afterwards learnt that Dr. Percy had made some experiments 

 with water in capillary tubes, and had arrived at a similar con- 

 clusion. He has not, I believe, published an account of these, 

 but kindly told me all the particulars. In his experiments the 

 water apparently did not freeze at all : but when in such minute 

 tubes, it is difficult to ascertain for certain whether the water is 

 frozen or not, if it be merely examined with an ordinary magni- 

 fying glass ; and therefore, thinking there might still be some 

 doubt about this, I resolved to carefully investigate the whole 

 subject. 



Of course we all know very well that the temperature at which 

 water will freeze is not necessarily that at which it thaws. It 

 has long been known that the temperature may be reduced very 

 considerably below 0° C. if no ice be present and the water be 

 kept tranquil, freezing, however, when agitated. If the water 

 be in a glass tube about one-fourth of an inch in diameter, the 

 temperature can easily be lowered to —5° C. (23° F.), even 

 when the tube is shaken ; and if kept quiet, the temperature may 

 be reduced somewhat more. At —6°, however, it freezes at once, 

 even when kept perfectly quiet. The same takes place in tubes 

 ^Lth of an inch in diameter ; but when much less than that, the 

 result is very different. 



To ascertain beyond all doubt whether the water in very minute 

 tubes was liquid or in the state of ice, I had recourse to the aid of 

 the microscope and polarized light. In using this, some°care is 

 required. The analyser should be so arranged as to have a per- 

 fectly dark field ; and the plane of the polarization of the light 

 should be either in the line of the length of the tube or at right 

 angles to it ; for, if not, the light is depolarized by the curving 

 sides of the tube. "When all is properly arranged, and the focus 

 adjusted to the centre of the tube, if it contain liquid water the 



