88 PROCEEDINGS OF THE [Oct., 1875. 



the line immediately above, under which they are respectively written, and also 

 of course with the corresponding members of the Iron group on the line yet 

 above. The average difference between corresponding members of the Iron 

 group, and of the Platinum group with In as now written in the table, is about 

 45, but if we write Zr, Ce, La, and D, under Cr, Mn, Fe, and Co respectively, 

 also E,u, E.O, Pd under Mu, Fe, Co, the " parallelism " of these three lines will 

 be as well preserved as before, the difference of Iron and Platiniun groups will 

 then be about 47, In will accord better in ' ' parallelism " than in the table, but 

 Y not so well as there ; the difference 47 agrees closely with the average differ- 

 ence between Series C and E, and between E and G in the upper portion of the 

 table. The arrangement adopted in the table however shews best the corres- 

 pondence of the pairs of elements which have coincident atomic weights in the 

 same group, namely, Co, Ni, and Ce, La, and Ru, Ro ; thesfe with Pt, Ir, are 

 the only cases known, and nothing like it is seen in the other or upper portion 

 of the Table. If in connection with this remark, we take the following : that 

 there is no positive pentad, that there is but one positive triad, Bi, (in the mid- 

 dle third of table II,) that therefore all positives are artiads, except the Kalium 

 Group and Bi, and that all these artiads, except the Calcium Group are con- 

 tained in the lower division of the Table : it will be seen that this lower part is 

 in marked contrast to the upper, and that the scheme of arrangement adopted 

 for the upper is wholly unsuited to the lower. 



There now remain only the monad H=l, and the dyad Hg=200, both con- 

 tained in the middle third of table 11, for which I can find no suitable place in 

 either of these schemes of arrangement, and I place them below the last 

 group, and separate them by a line drawn across the table. The first of these 

 is our standard for chemical comparisons having the lowest of atomic weights, 

 which is therefore used as the unit ; the other has a high atomic weight, ex- 

 ceeded by those of three elements only, and at ordinary temperatures is the 

 sole liquid metal, and the only liquid element besides Br. 



To the Groups in the upper portion of the Table, 1 have applied names drawn 

 from that member of each group which is contained in series C ; several, if not 

 all, are in common use. Under each of these groups we may distinguish two 

 sub-groups, and this remark may induce some to accept the arrangement of the 

 Table, who would otherwise be disposed to object to certain portions of it. 

 The Series C, E, G, which contain the greater number of Dumas' "triads" 

 will in every case (except perhaps line— 4) furnish the most prominent sub- 

 group, to each of which sub-groups except in line —1, may be added an ex- 

 treme member lying in Series K. The members of the other or secondary sub- 

 group, for lines —3 and —2, lie in Series D, F, and column marked H, I ; for 

 line — 1, they are wholly wanting or missing, but for line +2, they are to be 

 found in Series B, D, F, and for line +1, we may say they lie in the same 

 series, B, D, F, the member in D being missing at present. In Series A line 

 +2, Gl probably belongs to the secondary sub-group, while in line +1, Li 

 belongs to the primary sub-group, and if Tl properly belong to this sub-group, 

 it will be the largest in the Table, containing five members. The primary sub- 

 group would retain the name of the Group as written in the Table, the second- 



