lietween ,r = 0,i() and ./• r= 0,53 the agreement is decidedly bad; 

 at , lower temperatures slightly better. It is striking that the value 

 we lind for a is much too large, namely 0,355; for tin-mercury we 

 found ft)r « only the value 0,0453. 



We will further investigate whether the value of which is 

 calculated fr(»ni the initial straight i)art of the melting-point-line, 

 namely 6':=0,805, is in agreement witii the latent lieat of solidification 

 of pure silver. 



As 



and as Pkrson has found <i, = U)l ,My<^ '21,01 = '2Tfi iU-. kal., 

 we should have for 0: 



2X1232 



0—^ = 1,084. 



2274 



We liave, however, found the much .whtJ/rr value 0,805. This 



indicates the occurrence of mixed crystals already in the initial part 



of the melting-point curve, unless we assume, either that the value 



of Pkksox is about 1,35 times too small, or that the association of 



the lead, contained in the silver, is 1,35. 



III. Let us discuss in the second place the melting-point curve 

 of silrer-tin. We conclude at once from the figure of Heycock and 

 Neville, that complications, mixed crystals for instance, must occur. 

 For though the melting-point curve from 30 atom-procents tin 

 upwards shows the normal typical course, the initial part, instead 

 of being nearly straight, is strongly concave towards the side of 

 siUer, so that two inflection points occur, quite contrary to the course 

 indicated by formula (1) or (2). 



It is accordingly impossible to determine the value of ^ from the 

 initial part of the curve. If we calculate 0, a and ?■ from three 

 observations, for instance ,' = 0,43, ,c = 0,61 and .r = 0,86, then 



we get with : 



T„ = 961,5 ') + 278,2 = 1234,7 



the following values (conip. the table of Heycock and Neville, p. 40 



and 41): 



= 1,491 ; « = 0,718 ; r=- 0,277. 



1) The value 959" .2 given by Heycock and Neville has been augmented to 961°.5 

 on account of the accurate observations of Holborn and Day (quoted in Z. f. Ph. 

 Gh. 35, p. 490—491), from which appeared that pure silver, the air being excluded, 

 so that no oxygen can be absorbed, has a higher melting point (961 °.5) than silver 

 containing oxygen (955°). 



