44 



L. F. Nilson and Otto Pettersson. [Nov. 18, 



removed by tlie results above mentioned. Fixed at 13*65, according 

 fco our determinations, the value of its atomic heat becomes perfectly 

 harmonious with the law of Dulong and Petit. 



We will, in conclusion, say a few words upon a paper wherein the 

 results of our former researches have been criticised. In the " Pro- 

 ceedings of the Royal Society," 1879, Mr. Carnelley applies a new 

 method of calculating the fusion points of halogen compounds, and 

 applying it to those of beryllium, argues in the following manner : — 



Beryllium must be either a dyad or a triad, and must belong either 

 to the second or to the third group of Mendeleeff 's series ; if a dyad or 

 Be=9'2, its chloride, BeCl 2 , can be calculated to fuse at -f 547 — 

 + 600° C, which is confirmed by experiment; if a triad or Be= 13*85 

 the chloride, BeCl 3 or Be 3 Cl 6 , ought to fuse about 500 degrees lower, 



i.e., at + 50° |-100 o C, which it obviously does not; ergo, beryllium 



is a dyad, and Be=9*2. 



We presume that Mr. Carnelley's knowledge of the physical pro- 

 perties of the triads is, like our own, very limited. With the 

 exception of aluminium, we really know little or nothing of the 

 melting or boiling points of chlorides, bromides, and iodides belong- 

 ing to this group, and we think analogies taken only from one 

 member, aluminium, to be too narrow a base for a calculation 

 which is meant to apply to the whole group. There maybe chlorides, 

 bromides, and iodides which do not behave like A1 2 C1 6 , in regard 

 to boiling and fusion points. We will, in the following paper, give 

 reasons for our opinion that beryllium and aluminium are each 

 leading members of two different groups of trivalent metals. The 

 nearest relatives of beryllium amoug these are neither calcium and 

 magnesium, with which it has, in fact, little or nothing in common, 

 nor aluminium, with which it has very much more in common, but 

 rather the rare elements, scandium, yttrium, erbium, and ytterbium. 

 We believe that Mr. Carnelley ought to try his calculation on the 

 halogen compounds of the rare elements before asserting " that 

 Nilson's and Pettersson's determination of the specific heat of 

 beryllium must be incorrect." If the properties of the halogen com- 

 pounds of these elements should be found to agree with the calcula- 

 tion, then we confess that the matter becomes somewhat uncertain, 

 for then one will have to choose between the law of Dulong and 

 Petit and that of Carnelley.* 



Our above-mentioned experimental researches, confirmed still more 

 by the experiments, which will be quoted in a second paper, lead us 

 to the conclusion that the real atomic weight of beryllium is = 13*65. 

 But with this value the periodic law cannot admit this element among 

 the metals nearest related, and this fact obviously militates against its 



* " Ber. d. Deutscb. Chem. G-esellsch.," v, 303. 



