192 



Mr. J. W. L. Glaisher. 



[June 19, 



+ + 4 0, 5 C, or 0"8 per cent., and for all known melting points +15, 

 or 3 per cent. The calculated values for a large number of unknown 

 melting and boiling points are given. 



It is also shown that a knowledge of the melting or boiling points 

 of the halogen compounds of an element may serve for the determina- 

 tion of the atomic weight when the application of the methods of 

 specific heat and vapour density are inadmissible, or the results 

 obtained by these contradictory. 



An application of this is made in the case of beryllium, concern- 

 ing the atomic weight of which there has recently been some dispute. 

 From his determination of the specific heat of the metal, Emmerson 

 Reynolds concludes that it is a dyad with atomic weight 9*2, whereas 

 Nilson and Petterssen, from their determination of the specific heat, 

 assign to it the atomic weight 13*8, in which case it would be trivalent. 

 i^ow, according to calculation, the melting points of the chloride, 

 bromide, and iodide of beryllium ought to be (820-870°)* (802-820°) 

 and (766-777°) respectively. The boiling points, also, ought to 

 diminish from the chloride to the iodide, if beryllium is a dyad with 

 atomic weight 9 '2. If it be a triad with atomic weight 13 "8 these 

 melting points ought to be about 500° lower, and the melting and 

 boiling points ought to increase from the chloride to the iodide. With 

 the object of settling this point the melting points of the chloride 

 and bromide have been carefully determined, and it was found that 

 the chloride melts at 858-890°, thus agreeing with the limits (820-870°) 

 calculated for BeCl 2 (Be = 9'2). The bromide also fuses between 

 858 and 890°, and at almost exactly the same temperature as the chlo- 

 ride, but if anything slightly higher, the calculated number for BeBr 2 

 being 802-820°. The rather high number found for the bromide is 

 probably due to the substance being so readily volatile below its 

 melting point that the heat absorbed during conversion into vapour 

 cools the remainder of the solid, and thus prevents it melting so soon 

 as it otherwise would do. These results confirm the view that beryl- 

 lium is a dyad, with atomic weight 9"2 ; for though the melting point 

 of the bromide is apparently slightly higher than that of the chloride, 

 for the reason already given, yet BeBr 2 is far more volatile than BeCL. 



XIII. " Separate Enumerations of Primes of the Form 4ti + 1 and 

 of the Form 4n + 3 " By J. W. L. Glaisher, M.A., F.R.S.. 

 Fellow of Trinity College, Cambridge. Received June 11, 

 3879. 



The present note contains the results of separate enumerations of 

 primes of the forms 4.^ + 1 and 4w + 3 for seven groups each of 100,000 

 * Beckoned from absolute zero —273. 



