414 



NATURE 



{Sept. 1, 1 88 1 



tlieir causes, and lo cli^-covev the guiding principles which hnve 

 brought them about : he will try to derive, from a knowledge of 

 those guiding principles, a perception of the means by which 

 such progress can best be continued and extended — how it can 

 be most effectively directed to the benefit of his fellow-men. 



It is on this aspect of the question that I propose to address 

 yo'.i to-day. 



The process of scientific investi;TatiDn includes a great variety 

 of operations, which may be con'-idered under three headings, 

 mental, sensual, and physical. We think, we observe, and "C 

 work with our hands. In planning a new experiment we call to 

 mind what is known of the phenomena in question, and form an 

 opinion as to what is likely to happen under conditions somewhat 

 different from tho e which existed in previous experiments. We 

 regulate by careful observations the necessary manual operations, 

 so as to obtain with accuracy the desired conditions for the new 

 experiment, and we observe attentively the changes which take 

 place in the course of that experiment. The re-ult of such ob- 

 servations is somelimes in accordance with our anticipation, but 

 very frequently at variance with it. If it accords with our 

 anticipation, we put on record the extension which it has given 

 to the application of the general theory on which that anticipation 

 was founded. But if the result is not what we expected, we 

 carefully and critically revise the reasoning which had led ns to 

 expect a particular result, and often repeat the same experiment 

 with greater care, or some modification of it. 



Materiils for a new theory are gained when logically faultless 

 reasoning, checked by accurate observations, have led to re-ults 

 which could not have been foreseen by the aid of any previjus 

 theory. When a theory has thus gained a footing in science, it 

 serves a= a guide in further work. It guides us in arraiging 

 known facts. It guides us to the discovery of new facts. Some 

 ti lies it doesthee things for a short tine only, and is then super- 

 seded by some more general theory derived fro.n a wider and 

 more comprehensive view of the facts. 



There is, perhaps, nowhere so severe and rigorous a test of the 

 truth of an idea as that which is afforded by its u e in any accurate 

 department of experimental science ; and it is Avorth while, on 

 philosophical grounds, to consider briefly the conditions of 

 growth of the chief chemical theories which have withstood this 

 ordeal and proved them.selves to be trustworthy guides in experi- 

 mental science. 



Now as far as I know them, the general theories which have 

 played the chief pirt in the development of chemistry are mere 

 condensed statements of fact. 



Every thoughtful man of science has doubtless indulged in 

 speculations to find the cause of facts which are as yet unexplained ; 

 has imagined sojie fundamental condition or property of matter 

 which might can e it to produce effects such as are witnessed. It 

 is to be hoped that the ti'ne may be ivc distant when men of 

 science will confine their thoughts within the range of ideas which 

 are proved to lie true. But it is most important that they should 

 not confuse such hypothelical sreculatijn? with theories which 

 have received expCLimeiital verification, and that while enploy- 

 ing any theory they should not lose sight of the limits within 

 which it has been proved to be correct, beyond which it can only 

 be used as an hypothesi-. 



The foundation of the science of chemistry was laid by the 

 discovery of chemical elements ; those distinct varieties of 

 matter which we can neither produce nor destroy. Chemical 

 science treats of tho;e changes of property in matter which can 

 be represented as due to changes of combination of elementary 

 atoms. It knows nothing of the proiuction or destruction of 

 those elementary atoms. Sjieculatiins respecting their ultimate 

 form or structure will have fo.ind a place in the science as soon 

 as such speculati jns have helped to arrange the facts which are 

 known, and to oiscover new chemical facts. 



At the commencement o? oar epoch chemists had classified 

 elenentsacco.di ig to their electro-chemical properties. Chemical 

 analysis had e tabli^hed the fret that a good many compjunds 

 could be represented as consi ting of elementary atoms of two 

 kinds combmed in small number. Thus carbonic oxide and 

 carbonic acid had been found to possess respectively a com- 

 position which could be represented (adopting our present afo iiic 

 symbols) by the formuk-e CO and COj, water by the formula 

 HoO, marsh gas CHj, olefiant gas CH.,. The oxides and acids 

 of nitrogen were represented by f 'raiufe corre ponding empiri- 

 cally to those which we now adopt. So also ammonia and 

 hydric chloride had their present forujulse. Sulphuro.is and 

 sulphuric acid had the respective f jrmul^ SO^ and SOj. Phos- 



phorus and phosphoric acid had the formulae P^Oj and PjOj. 

 Baryta and the oxides of iron had the formula; BaO, FeO, 

 Fe„03. 



Such primary compounds were classified upon the same prin- 

 ciple which served for the claisification of the elements them- 

 selves, into electro-po4tive or basylous and electro-negative or 

 chlorous compounds, and the .smallest quantity of each of them, 

 which consistently with an atomic representation of the results of 

 analysis, v^ as deemed capable of existing, was called an atom of 

 that compound. 



Very simple compounds possessed of prominent characteristics 

 and distinct reactions had first been isolated and identified. 

 They were found to contain their con-tituent elements in propor 

 tions easily recognisable as multiples of atomic weights. But 

 such simple compounds are rare exceptions among mineral and 

 organic material-, and if the atomic theoi-y could have gone no 

 further than to guide us to an understanding of these few simple 

 compounds, it must soon have given place to some more funda- 

 mental conception. It is moreover worthy of notice that in this 

 its m ist elementary form the atomic theory was not the only con- 

 ceivable mterpretation of the proportions of combination between 

 elements. Those proporlions could be as consistently repre- 

 sented by fractions as by integral multiples. Thus, instead of 

 representing carbonic scid .as containing twice as much oxygen as 

 is contained in carbonic oxide, we might have represented it as 

 containing the same quantity of oxygen combined with half as 

 much carbon, and u-ing for the moment atomic symbols for a 

 non-atomic theory, we might have wri'ten carbonic acid thus 

 c 

 -O. Or w-e might represent them both by percentage numbers. 



2 



It was so simple and natural to adopt the atomic hypothesis, 

 and to represent compounds as built up of atoms, that chemists 

 seem to have paid little attention to any other mode of repre- 

 senting the proporlions of combination. They assumed that the 

 variable proportions of elements, which were ohseiTed in com- 

 pounds, were due to the various numbers of elementary atoms 

 respectively aggregated together in each compound. They per- 

 ceived that the existence of eleaientary atoms involved the 

 exi-tence of compound atoms, or molecules, as we now call 

 them, and accordingly they represented each known compound 

 of two elements by a molecular formula as simple as possible, 

 consi.stently with the view of its atomic constitution. Many of 

 these molecules, such as those of the acids, were four.d to be 

 capable of combining with others of the other class, forming 

 salts, and those combinations were found lo take place in poro- 

 portions corresponding to the \a eights of the respective molecules, 

 or to very simple multiples of those weights, and the secondary 

 compounds or salts thus formed combined (if at all) in propor- 

 tions corresponding to simple multiples of their molecular 

 weijhts. The dualistic representation of the constitution of 

 salts served to represent the results of their analysis consistently 

 with the atomic theory, and a vast number of fundamental facts 

 were collected and arranged by the aid of the dualistic theory of 

 combination. 



The actual numbers obtained by analysis of any particular 

 compound exhibited sometimes a very near approximation to 

 tho e required by an atomatic fonnula of its composition. Some- 

 times they differed con-iderably from tho e required by theory ; 

 but it was always found that the more pure in substance and 

 the more accurate the analytical oper.:tion.o, the more nearly 

 did the result agree with some atomic formula of the 

 substance. 



The compound atoms were units v\hich had grown out of the 

 atomic tlieory. Each of th^m was the smallest quantity of a 

 compound, which (consistently wilh the results of analysis) could 

 be represented as built dualistically of its constituent atoms. 



Chemical combination was viewed as a process of juxtaposi- 

 tion, of simple or compound atovs, little account being taken of 

 the disturbance of the previous arrangement of those compound 

 atoms. It was when a constitution, similar to that attributed to 

 salts, was imagined for other compounds not saline in their cha- 

 racter, th.t the dualistic theory broke down. Thus chlorocar- 

 boric acid was represented as a compound of carbonic acid with 

 carbonic chloride, and was accordingly designated ns carbonate of 

 carbonic chloride, while the formula w.is made to co-tain the 

 formula; of those bodies. Chloro ulphuric acid and chlorochromic 

 acid " ere in like manner represented as compounds of sulphuric 

 and chromic acid respectively with imaginary hexachlorides. 



Careful investigations of the reactions in which chlorocarbonic 

 acid takes part showed, however, that in each of them it behaves 



