564 REPOET — 1886. 



inhabit, and may thus be said to separate out these elements from the chlorine with 

 which they are mingled. 



But if we examine these cases of elimination we see that they are limited in 

 their scope. They extend only to substances existing in solution, of which there is 

 a fresh supply always at hand, and which are capable of entering into the animal 

 or vegetable economy. Again, the elimination of iodine and bromine, effected as 

 just described, is of a very imperfect character, and, when such water-plants and 

 animals die and decay, their constituents will be again distributed in the water. 



We cannot well consider that nickel and cobalt have been deposited in ad- 

 mixture by organic agency, nor yet the groups iridium, osmium, and platinum 

 — ruthenium, rhodium, and palladium. 



Since the earthy metals to which I have referred — such as yttrium, samarium, 

 holmium, erbium, thulium, ytterbium, &c. — are very rare, the probability of their 

 ever having been brought together in some few uncommmon minerals discovered 

 only in a few localities must be regarded as trifling indeed, if we suppose that 

 these metals had at any time been widely diffused in a state of great dilution with 

 other matter. The features which we have just recognised in these earths seem to 

 point to their formation severally from some common material placed in conditions 

 m each case nearly identical. The case is strengthened by a consideration of the 

 other groups of elements, also similar in properties, having little affinity for each 

 other and occurring in admixture ; either all or at least some of the elements con- 

 cerned being moreover decidedly rare. Thus we have nickel and cobalt not 

 plentiful or widely distributed ; cobalt, perhaps, never found absolutely free from 

 nickel, and vice versa. AVe have also the two platinum groups, where very 

 similar features prevail. 



A weighty argument in favour of the compound nature of the elements 

 is that drawn from a consideration of the compound radicles, or, as they 

 might be called, pseudo-elements. Their similarity with the accepted elements 

 is perfectly familiar to all chemists. If, for example, we suppose that in some age 

 or in some country men of science were cognisant of the existence and of the 

 behaviour of cyanogen, but had not succeeded in resolving it into its constituents, 

 nothing, surely, would prevent their viewing it as an element, and assigning it a 

 place with the halogens. It may fairly be held that if a body which we know to 

 be compound can be found playing the part of an element, this fact lends a certain 

 plausibility to the supposition that the elements also are not absolutely simple. 

 This line of thought, or at least one closely approximating to it, was worked out 

 by Dr. Camelley in a paper read before this Association at its last meeting. From 

 a comparison of the physical properties of inorganic with those of organic com- 

 pounds, Dr. Carnelley concludes that ' the elements, as a whole, are analogous to the 

 hydrocarbon radicles.'' This conclusion, if true, he adds, should lead to the fm-ther 

 inference that the so-called elements are not truly elementary, being made up of 

 at least two absolute elements, named provisionally A and B. Hence, he argues, 

 it should be possible to build up a series of compounds of these two primary 

 elements which would correspond to what we now call elements. Such an 

 arrangement, to be admissible, would have to fiilfil certain conditions : — The 

 secondary elements thus generated from A and B must exhibit the phenomena of 

 periodicity, and the series would have to form octaves : the entire system is bound 

 to display some feature corresponding to the ' odd and even series ' of Mendeleeffs 

 classification ; the atomic weights must increase across the system from the first 

 to the seventh group ; that is, from the positive to the negative end of each series ; 

 the atomicity would ha\e to increase from the first to the middle group, and then 

 either increase or decrease to the seventh group ; some feature should appear cor- 

 responding to the eighth gi-oup ; and, lastly, the atomic weights in such a system 

 ought to agree with the atomic weights as experimentally determined. 



This last condition Dr. Carnelley rightly regards as the most crucial, and he 

 finds his arrangement gives atomic weights which in a majority of instances 

 coincide approximately with the actual atomic weights. Thus out of a total 

 of sixty -one elements whose atomic weights have been determined with at 



