466 Dr. M. Wildermann on Real and Apparent 



Jones and Abegg* use pretty much the same amount of over- 

 cooling, viz. about 1°, Loomisf only o, 3; LoomisJ, in his new- 

 method, and others have an exceedingly small value of t — t 0B , 

 since they determine the freezing-temperature with ice, when 

 it is almost completely melted) ; the ice must be separated into 

 as fine needles as possible, and for this it is necessary that 

 the liquid be overcooled in a bath which is not too cold, that it 

 may not freeze of itself § (since in this case large flakes 

 are formed instead of fine films) ; the ice must not be allowed to 

 form a conglomerate mass, nor to settle on the sides or at the 

 bottom of the beaker. The ice, in a word, must always be 

 separated into fine films and in sufficient quantity throughout 

 the whole of the liquid. 



(c) tg—t' must be kept as small as possible ; %. e. the 

 convergence-temperature must be brought as near as possible 

 to the freezing-temperature. (The value of t g —t' can be made 

 = '1° or "2°, or even less. In some of the methods t ff — t' 

 amounts to several degrees.) For freezing-point depressions, 

 however, the more important point is that t g — t' and tg—t'' in 

 two liquids may be equal ; i. e. the ice-bath must be kept within 

 very narroio limits of temperature. I use the expression " as 

 much as possible," since we are practically limited in bringing 

 about the ideal conditions we require ; we are not able to 

 have an infinite quantity of ice, i. e. to make t — t ov = co, and 

 so also to make t q — t / = 0. Our object is therefore to arrange 

 the freezing-method so that t — t' is reduced to a value which, 

 under given circumstances, can be quite neglected. 



Attention has been already drawn, when the equation (A') 

 was under consideration, to the fact that if the convergence- 

 temperature is above the freezing-temperature, the temperature 

 t' of the liquid is at equilibrium higher than the real freezing- 

 temperature t (and below the convergence-temperature tg). 



According to our conception of perfect equilibrium, the 

 solid part (ice) cannot exist above the temperature t : 

 therefore the ice will not be warmed by conduction 

 to the higher temperature of the liquid, but will melt 

 aud remain at the freezing-temperature t . It has been 

 already empirically found (see paper mentioned above, 

 " Zur Bestimmung/' &c.) that if the ice-cap be well closed, 

 about the same values are obtained for the difference between 

 the read freezing-temperature with ice-cap and that without it, 

 and also || that in the case of ice-bands, network of ice, 



* Zeitschrift fur Physik. Chemie, xv. p. 682. 

 t Wied. Ann. li. % Wied. Ann. 1896. 



§ Measurements of the velocity of ice-nielting in all these cases will be. 

 given in the paper " On the Velocity," &c. 



H See Phil. Mag. July 1895; Zeitschr. f.phys. Chem. xv. 



