Prof. T. Carnelley on the Periodic Law. 21 



weight of Be. From his determination of its specific heat, 

 Emerson Reynolds (Phil. Mag. [5] iii. p. 38) concludes that 

 its atomic weight is 9*2, and that it is a dyad; whilst Nilson 

 and Pettersson (JDeut. chem. Ges. Ber. xi. p. 381), also Hum- 

 pidge (Phil. Trans.), from their determination of its specific 

 heat, give to it the atomic weight 13*8, in which case it would 

 be a triad. Now according to calculation the melting-points 

 of its chloride, bromide, and iodide ought to be (820-870), 

 (802-820), and (766-777) respectively if it has the atomic 

 weight 9*2; and although these numbers are somewhat uncer- 

 tain, from their having been calculated for an element of the 

 second group (see Table IX.) occupying the position generally 

 assigned to Be, yet they are sufficiently near in order to de- 

 termine whether Be is really a dyad or a triad. If it be a 

 dyad, the melting-points of its halogen compounds must be 

 considerably higher than those of boron (see Conclusion V. 

 p. 13, also Plate); whereas if it be trivalent, these melting- 

 points must be comparatively low. If the metal is a dyad, 

 the melting-point of its chloride should also be higher than 

 that of the bromide, and the latter higher than that of the 

 iodide; whereas if it be a triad, the reverse should be the case. 



To throw light on this point, I determined as carefully 

 as possible the melting-points of the chloride and bromide of 

 beryllium, and found that the chloride fuses between the 

 melting-points of Ag 4 P 2 7 (melting-point = 858°) and NaP0 3 

 (melting-point = 890°), thus agreeing with the number (820- 

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

 between 858-890, and almost exactly at the same tempera- 

 ture as the chloride, but, if any thing, slightly higher, the 

 calculated number for BeBr 2 being 802-820. The rather 

 high number obtained for the bromide is probably due to the 

 substance being so readily volatile below its melting-point, 

 that the heat absorbed during its conversion into vapour 

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

 so soon as it otherwise would do. BeBr 2 , however, is much 

 more easily volatile than BeCl 2 . The chloride and bro- 

 mide used for the determination of the melting-points were 

 the pure compounds obtained by repeatedly distilling the crude 

 substances, produced by passing chlorine and bromine respec- 

 tively over a red-hot mixture of the oxide and charcoal. 



An analysis of the chloride gave: — 



Found. Calculated. 



Be 11-93 11-47 



Cl 2 88-90 88-53 



100-83 100-00 



