714
CHEMISTRY
future to characterize the materials used, either by means of spectroscopic examination or by methods of equivalent delicacy—if indeed such can be found; so-called chemical methods must often prove too coarse for such purposes. The difficulty—it may almost be said the impossibility— of separating the constituents of mixtures of isomorphous substances under limited conditions is well known, but it may well be that some of the materials hitherto regarded as uniform single substances are in reality mixtures of this kind. The need of studying the rare earths from such a point of view is beyond question; indeed, it is time that the opportunities which this group of elements presents of solving some of the most difficult but entrancing problems in chemistry were appreciated. But there can be little doubt that the main sources of error in estimating atomic weights occur in carrying out the operations, and that constant unperceived errors often affect determinations Avhich have been conducted with the most scrupulous care. The conception that a genetic relationship exists among the elements almost necessarily involves the idea that their Relation- atomic weights are in some way commensurable ships of' numbers. Prout first gave expression to this atomic view in his well - known hypothesis that the weights. atomic weights of the elements generally are exact multiples of that of hydrogen. Stas commenced his work fully persuaded of the validity of this hypothesis, but his results led him to an entire change in his opinion, and the tendency of all modern investigations has been to disprove it. Nevertheless the feeling remains that the question is still open. Thus Mallet, in his lecture delivered in memory of Stas (Trans. Chem. Soc. 1893, 35), after fully discussing the question, said :— But the fact that so many well - determined atomic weights, referred to hydrogen as unity, present numbers nearly approaching integers, is very striking, and calls for further investigation. The writer of this paper is in a position to quote Stas himself as admitting this much. Having enjoyed the pleasure of a conversation of considerable length with the eminent Belgian chemist at his home in Brussels in the summer of 1887, and having urged upon him the improbability of this near approach to integer values for so many atomic weights being due to chance alone, the writer £< was answered II faut croire qu'il y a quelque chose 1 by the remark, la-dessous.” Apart from Stas’s great work, which can now be studied without difficulty in the three volumes of his CEuvres completes published after his death, no more important or more interesting literature dealing with this subject can be consulted than that relating to the numerous and persistent attempts made during recent years to fix the ratio between hydrogen and oxygen. The problem has engaged 1 In view of this pronouncement, it may be well to direct attention to the opinion expressed by Mendeleeff in 1871 in his original essay on the periodic law [Liebig's Ann. sup. viii. p. 206): “Even if it be allowed that the matter of elements is entirely and uniformly the same, there is no reason to presuppose that n parts by weight of an element, or n atoms, in being changed to an atom of the second element, would produce the same n parts by weight; or that the atom of the second element will be n times heavier than in the first case. The law of the conservation of weight can be regarded as a special case of the law of the conservation of energy, or of motion. Weight is, of course, caused by a special kind of motion of matter, and no reason exists for denying the possibility of a transformation of these motions in the formation of elementary atoms into chemical energy-or one or other form of motion. Two of the phenomena that are to be observed in elements, viz., the constant weight of atoms and the impossibility of further decomposing them, have hitherto been in intimate, even historical connexion ; if then a known element should decompose or a new one be formed, these phenomena might be attended by an increase or reduction of weight. In this way, moreover, the difference in the chemical energy of various elements might to a certain extent be explained. But in giving expression to this theory, I intend nothing further than this, that there is a certain possibility of bringing the opinion tacitly entertained by chemists of the composite character of the elements into harmony with the non-acceptance of Prout s hypothesis.”
the attention of workers all over the civilized world, and we are much indebted to the American chemists, and to Morley in particular, who have spared no trouble to arrive at trustworthy results. Stated in very few words, the result uniformly arrived at is that far from being in the integral ratio 1:16, which was so long accepted, the two atomic weights stand to each other in the relation of 1 :15‘88. But the difficulty of attaining to finality in such work is well illustrated by the case of nickel and cobalt. These elements have over and bver again attracted attention on account of their atomic weights being so nearly alike. Becently Bichards, assisted by Cushman and Baxter, has again attacked the problem, and the accounts given of the work are such as to impress the reader with the greatest confidence in the results. A variety of independent methods was used, and the consistency of the results obtained is remarkable. The conclusion ultimately arrived at is that cobalt has a slightly higher atomic weight (59) than nickel (58‘7). Unfortunately, however, this does not settle the question, for, as Mallet observes in his Stas lecture—after pointing out that the general tendency of investigation up to that time had been to assign to cobalt a rather higher atomic weight than to nickel—on comparing the properties of the three sets of metals—Be, Co, Ni—Bu, Bh, Pd—Os, Ir, Pt— cobalt plainly occupies a position intermediate between iron and nickel, and therefore its atomic weight should be lower than that of nickel. No better example can be given of the severity of the struggle that is being waged between theory and practice, and of the difficulty with which the victory is to be won. Those desiring to study the subject of atomic weights more fully should consult F. W. Clarke’s Recalculation of the Atomic Weights, issued by the Smithsonian Institution in 1897. In this invaluable monograph reference is made to all that had been done in this department of chemistry. The subject is kept up to date in the annual reports published in the Journal of the American Chemical Society. Molecular Weights. The conception the chemist forms of the properties of an element is a very complex one, and is based not only on the study of the element as it occurs in a separate form, but of every available compound. The series of impressions thus obtained are ultimately combined into an ideal, which in the chemist’s mind constitutes the “character ” of the element, and it is such a series of ideals that he has before him when considering the periodic system. He is, in fact, to a great extent guided by instinct. But in the case of elements, as in that of individuals, the determination of character is often attended with very great difficulty, a true estimate being only slowly arrived at, and when at last such an estimate is formed, it can only be very partially expressed in words. It is therefore justifiable to say that imagination and even sentiment play an important part in chemistry, and that if too rigidly and narrowly interpreted, facts may become very misleading factors. One of the greatest difficulties in the way of forming a correct estimate of the properties of an element arises from the extraordinary manner in which properties are dependent on molecular composition; the task is equivalent to that of studying individuals through observations of crowds in which they are barely distinguishable. Familiar instances are afforded by oxygen and ozone, and by properiies ordinary yellow and red phosphorus. No two andmolecsubstances are more unlike in behaviour than u,ar comare oxygen and ozone, yet both are composed of P°slt,onthe same “stuff,” the difference being that the one has diatomic (0<2) and the other triatomic (03) molecules. Again, the inactivity of nitrogen, as we know it, is altogether
