22 Frederick Guthrie on Salt-Solutions 



Now, judging from available data, the normal chlorides of 

 all triad elements melt below 510° (BiCl 3 being the highest 

 known, melting-point = 503°), and their normal bromides 

 below 500° (BiBr 3 being the highest known, melting-point = 

 480°). These results therefore confirm the view that beryl- 

 lium is a dyad, with atomic weight 9*2. This is rendered still 

 more evident by an inspection of the Plate, showing the 

 curves of the melting-points of the halogen compounds of the 

 elements*. 



Fluoeides. 



As fluorine is an even, whilst CI, Br, and I are odd members 

 of the seventh group, the fluorides cannot strictly be compared 

 with chlorides, bromides, and iodides as regards their melting- 

 and boiling-points. The melting-points and boiling-points of 

 the fluorides, however, taken apart from those of the other 

 halogens, appear (if we may judge from the few melting- 

 points and boiling-points which are known) to obey laws 

 similar to those which govern the melting-points and boiling- 

 points of the chlorides, bromides, and iodides. 



II. On Salt-Solutions and Attached Water. 

 By Feedeeick Gutheie f. 



VIII. 



[The paragraphs are numbered in sequence with those of my memoir 

 on Eutexia, anted, vol. xvii. p. 462.] 



§ 232. TNTRODUCTION.— When, some ten years ago, 

 the first experiments in this subject were brought 

 before the Physical Society, it was pointed out that the key to 

 the temperature of any salt-ice freezing-mixture was to be 

 found in the temperature of fusion of the cryohydrate of that 

 salt. Since then further light has been thrown upon the 

 subject by the discoveries and observations of others. Thus, 

 Ditte has obtained a hydrate (probably a subcryohydrate) of 

 hydrochloric acid, and Wroblewski has obtained a similar 

 hydrate of carbonic acid. Pettersson, in his exhaustive study 

 of marine ice, made in the ' Vega ' expedition, has explicitly 

 admitted that the phenomena presented by freezing and frozen 

 sea-water can only be satisfactorily explained by the formation 

 and properties of the cryohydrates. Though this tacit corro- 



* Since writing the above, Messrs. Nilson and Pettersson {Deut. chem. 

 Ges. Ber. xvii. p. 987) have determined the vapour-density of beryllium 

 chloride, and find that it corresponds with the formula BeCl 2 , and conse- 

 quently that Be =9-2. 



t Communicated by the Physical Society. Eead May 24 ; 1884. 



