318 THE ANALYST. THE PROPOSED STANDARDIZATION OF METHODS OF CHEMICAL ANALYSIS." BY BERTRAM BLOUNT, F.I.C. THERE has been evident of late years a tendency to apply the principle of standardiza- tion to matters and in directions differing considerably from those in which the advisability of erecting and enforcing a standard is generally conceded. The commonest examples of standardization - namely, that of weights, of measures, and of money-have come into existence and are maintained by their manifest necessity. Various mechanical standards are so convenient as to have become practically necessities ; for instance, standard gauges for wire and sheet have been used very freely, and standard screw threads have also been found desirable. But, passing from these simpler cases where standardization is universally conceded to be convenient to a consideration of the advisability of extending this principle to more complex matters, we enter into a region of polemic ; the advantages of an artificial uniformity are less clear, and the disadvantages become more easily discernible and are ultimately prominent.As an instance of this may be mentioned the proposal to standardize sections of structural metals, sizes of test-pieces for mechanical tests, and methods of applying such tests. If the difficulty of accepting the principle of standardization is considerable in the matter of mechanical testing, it becomes still greater in the case of chemical an a1 y sis. I n mechanical testing, it is recognised that the values obtained for strength in tension and compression, for elongation of a ductile test-piece, for resistance to shock and the like, are appreciably influenced by the conditions of the test, and, in conse- quence, there is something to be said in favour of standard methods of applying the tests.But there is no such justification for the adoption of standard methods in chemical analysis where the object in view is the determination in a given material of a definite substance ; in such a case all.methods which are chemically sound must give the same results. But notwithstanding these obvious facts, a desire has arisen in some quarters for the formulation of standard methods of chemical analysis, and for their gradual imposition on the community of chemists; steps have been taken to secure this end, and it is necessary to decide whether this movement is laudable or not, lest judgment should go by default.A serious proposition in this sense was put before the Society of Chemical Industry by Professor Lunge in the year 1884. Discussions followed in which leading chemists, many of whom are still with us, took part, and in the face of an evident conflict of opinion it became necessary to appoint a committee to consider the whole question. I have not been able to find that any report by this committee was published, The wish to establish standard methods of analysis is no new thing. * Abridged froin a paper read before the British Association at their meeting in Belfast, 1902.THE ANALYST. 319 and from the situation to-day it is fairly certain that no definite decision was arrived at.About the same time, now nearly twenty years ago, the Society of Public Analysts occupied themselves with a similar subject. The methods of water analysis were discussed, and recommendations for a standard process were formulated. A like course was adopted for the analysis of milk, and some years later for the detection of butter in margarine by the Reichert-Wollny process. I t will be observed that in all these cases the substances to be examined are natural products of indeterminate composition, and that the methods for their examination are of the nature of empirical tests rather than of analysis proper. Such tests are often of the highest importance and utility for practical purposes, and from their nature are best performed under standard conditions, but they must not be confused with actual chemical analysis, which is under no such obligation to rule.At the meeting of the British Association held in Bath in 1888 a committee wa8 appointed to consider the best methods of establishing international standards for the analysis of iron and steel.” The initiation of this investigation appears to have been due to Professor J. W. Langley, who undertook the preparation of the necessary samples. Other national committees were invited to join in the work, and after some years certain of those committees reported. At an early stage of the proceedings the British committee defined the object of the inquiry as the provision and preservation of specimens of iron and steel of a composition most carefully ascertained.These were intended to serve as standards whereby the accuracy of any given process or any given worker might be checked. I t will be observed that there is here no word of a standard method of analysis; the members of the committee were to be free to use the method they thought proper, and the results alone were to be the standard. But the members of the American committee seem to have held a different object in view ; they applied themselves to the examination of methods, and especially of those for the determination of carbon. I t may be conjectured that their objective was the provision of a standard method, but this aini was not attained, and no subse- quent attempt in the domain of steel analysis has met with greater success. It is interesting to note that whilst no standard method of steel analysis has been arrived at, the improvement in existing processes and the device of new methods have continued without pause or set-back, and that this branch of analytical work has been brought to a high state of efficiency. C.B. Dudley, of the Pennsylvania R.R., in 1893 discussed the causes of discrepancy between the results of different chemists, and advocated the use of standard methods for avoiding these discrepancies. Quite recently (in 1901) a committee of the New York section of the Society of Chemical Industry approached the same subject in the same spirit, and has reported in favour of a standard method of analysis for cement and cement materials. I had the honour in May last of reading a paper before this body, in which I ventured to criticise the method in detail, and to dissent from the acceptance of a standard method. These criticisms and objections were very frankly and fairly received and discussed, and it may be hoped that the discussion will induce some of those who But the idea, although not realized, was by no means dead.320 THE ANALYST. were perhaps over ready to except standardization on its superficial merits to consider its real and weighty disadvantages, and to confine their activity to those branches of analytical work in which some form of standardization is admittedly useful.The introduction of standard methods for the analysis of coal, sugar, fertilizers, and feeding-stuffs and for water analysis has also been advocated in recent years, chiefly in the United States, and some of the proposed methods have been officially adopted by various associations.These instances will suffice to show that there is an increasing tendency to propose standardization of analytical methods, irrespective of the nature of the material to be analysed, and it remains to consider how this tendency has arisen and what effect it is likely to have if it continues to grow without opposition. In the first place, it will be seen that there are in this matter two schools of thought sharply differentiated. On the one hand, we have chemists so impressed with the necessity of avoiding discrepancies that they are anxious to set up standard methods, stated in such detail that a faithful observance of the prescrip- tions by two workers cannot fail to secure identical results.And there are others less confident of the success of this system. This conflict of opinion arises in great measure from confusion of thought. There are many operations performed by the chemist which are from their nature arbitrary. To take an extreme case, it is evidently impossible to carry out the Maumenh test for oils or to determine the flashing-point of petroleum, and to obtain concordant results without having recourse to a fixed procedure. In the examination of potable waters such arbitrary methods are in wide and general employment, and yield much useful information. Feedingstuffs and manures afford a similar case, Citrate-soluble phosphoric acid ” is a meaningless term unless the conditions of its extraction are laid down.The fractional distillation of crude benzol and the return of its computed contents of benzene is another instance. But this sort of codification, though useful and often necessary, has nothing to do with analysis. The object of the analyst is to determine with the best precision in his power the constituents of the substance which he is analysing, Sometimes he cannot do this, and is forced to have recourse to the determination of the properties of the substance; that is to say, he is compelled to apply to the material under examination various arbitrary tests. Standardization for these tests is legitimate enough ; standardization of analysis implies that analysis is an arbitrary procedure not based on ultimate chemical facts. This view is too grotesque for discussion.To turn to the much-debated question of discrepancies in analytical results, which is at the root of the whole matter. In trade it is necessary frequently to appeal to the analyst, and if the results given by one analyst differ from those of another, examining what is believed to be the same sample, much delay and expense may be incurred. The trader is usually unaware of the great difficulty which exists in procuring identical samples, and he is almost certainly ignorant of the slow and laborious character of all analytical processes having claim to be exact, From the former cause he is disposed to ascribe all differences to errors of the analyst, and from the latter he is inclined to urgeTHE ANALYST. 321 greater rapidity of work than is compatible with accuracy.Sometimes from bad sampling or from hurried or unskilful work discordant or erroneous results may arise. Then the trader seeks a remedy, and believes that it may be found in a standard method of analysis. It need hardly be urged before this assembly that such a conclusion is fallacious. The same imperfection in sampling, the same hurried and unskilful work will lead to the same erroneous results; the only difference will be that each Chemikant-for he cannot be called a Chemiker-will be entitled to say that he had followed the prescription as faithfully as he was able. It is not likely that chemists who practise as independent consultants will submit to dictation, even from a committee of their peers, as to the methods they are to employ. A member of this branch of the profession, asked to determine in some material a definite constituent, is either competent to choose a suitable method, and to execute it accurately, or he is not.If he is, a standard method is super- fluous ; if he is not, his early retirement is desirable, and will probably occur from the operation of natural causes. But there are some consultants who are more or less identified with particular branches of the chemical trade; khey are constantly called upon to carry out analyses or assays for the buyers and sellers of ores, metals minerals, salts and the like, and usually the results obtained by the buyer’s and the seller’s chemist are compared, and may be found to disagree. Now, here is a plausible case for the use of a standard method. I t is argued that if the two chemists concerned were to use the same process the likelihood of their disagree- ment would be diminished.There is even a limited truth in this; the likelihood of disagreement would be diminished to some extent because both operators would be exposed to identical chances of error, instead of different chances. But the object of analysis is not to obtain a false concord by repeating it process, errors and all, with Chinese fidelity; it is to arrive at the truth as closely as knowledge and skill will allow, and one of the recognised methods of doing this is to vary the conditions of experiment. How misleading are mere concordances is sufficiently shown by some of the earlier determinations of atomic weight vitiated by a systematic error, exposed at once on radically changing the method.If the atomic weight of nitrogen had depended on the determination of the density of the residuum obtained by eudiometric analysis, a considerable error would have been admitted in a fundamental figure for a period of many years. The need of the buyer’s and seller’s chemist for identical results will best be met by insuring that they are sufficiently well trained to refuse to employ a faulty method; their choice among several sound methods is obviously a matter of in- difference. It is only when the assay is from its nature arbitrary, and is, in fact, a conventional test, that an identity of method is necessary. A great part of the clamour for standardization arises from the discrepancies which often occur between the results of different works’ chemists, and it appears that the “ standard” methods which have from time to time been proposed in the United States have the works’ chemist as their particular objective.These dis- crepancies are highly inconvenient, not only immediately to those concerned, but ultimately as tending to discredit the value of analytical control of manufacturing processes. When last in America I took some pains to ascertain the opinion of382 THE ANALYST. chemists on this point, and found that those who advocated the adoption of standard methods, when confronted with the grave objections which can be urged against them, took refuge in the necessities of the works’ chemist, maintaining that only by the provision for him of standard methods could uniformity and accuracy be attained.I t cannot be denied that this view is defensible, but it is defensible on one ground alone, and that hardly likely to commend itself to the works’ chemist. His friends, in their anxiety to help him, infer that he is incompetent, and needs to be guarded and guided like a schoolboy. Now, as a matter of fact, although often overburdened with work and most inadequately paid, the works’ chemist has done much to advance analytical chemistry, originating some methods and improving many others. I t is neither just nor courteous to assert that he, above all others, is in need of direction and supervision from without, and that he is willing to barter his right of individual judgment, according to his training and experience, for a bundle of prescriptions. It is, however, unfortunately true, that while there is nothing in the nature of things why the works’ chemist should not be as highly trained and as capable of individual decision as his consulting colleague, yet it remains a, fact that some works’ chemists are not fully competent, and, conscious of their deficiency, might even welcome the promulgation of a standard method by which a given analytical process shall be carried out.Such cases are freely advanced by the advocates for standardiza- tion, and doubtless afford some show of argument in favour of the movement. But surely it is better and sounder to direct effort, not towards the provision of rules for half-trained chemists, but towards insuring that the chemists are fully trained.And here one is met by the money question. At present the average manufacturer is so little alive to the absolute necessity of obtaining proper scientific and technical aid in the conduct of his business that he will offer to the chemist a salary hardly in excess of the wages of a skilled mechanic. Is it to be wondered at that the man who is willing to accept such remuneration should not be fully trained, that his knowledge should be empirical, and that he shocld need the prop of a (‘ standard” method scientifically on a, par with a cookery recipe? If it is the desire of those who advocate the standardization of chemical methods to promote the multiplication of analytical machines turning out results of artificial uniformity, no better course could be adopted than the prescription in detail of processes which shall relieve their user from the troublesome process of thinking.I do not like to suppose that this is the object deliberately desired, but it is well at least to recognise that it is likely to be attained by the successful establishment of standard analytical methods. There are other results to be expected of wider and more general importance. Let us suppose that standard methods had been prepared and published with the consent and under the direction of some body of chemists of sufficient standing to make their pronouncements authoritative. The position of all chemists who ventured to disagree with the findings of this central body would become very difficult. I n all cases of dispute, in all legal proceedings, these dissentients would be placed at a disadvantage ; on them would be thrown the onus of proving, usually before a non-technical tribunal, that the method they preferred was as good as that of the central committee.Now, it is as certain as anything not actually accom-THE ANALYST. 323 plished can be, that the standard method at its best would be a compromise between the different processes in use by the various specialists in that particular branch of work, and would have the usual quality of a compromise-namely, the preference of expediency for principle. Consequently, no independent worker basing his process on what he conceived to be the soundest principles could avoid a conflict with a method which from its origin is a hybrid product.If the dissentient were not able to convince the lay tribunal that his process was reliable, a greve injustice might be done to him ; if, on the other hand, he did so succeed, the standard process would be discredited. The endeavour here made to trace to its logical outcome the effect of standardization is sufficient to show that such a scheme, evenif it could be regarded as desirable, would lead to such complications as to render it impracticable. If a standard process is adopted by a central authority, it will be regarded by many as final, and all incentive to original inquiry in that direction will be removed. Periodical revision by a standing committee will probably be carried out, but naturally on the lines of the existing process, and with the view of improving it in detail.The deviser of a method radically different from that which is official will be regarded as a sort of chemical heretic, and will receive the treatment usually accorded to the heterodox. Now, analytical research, although not the highest form of chemical investigation, is of great value both as a school of training and as a means of perfecting the chief instrument of the chemist, whose work, unless constantly controlled by its aid, mould be little better than a maze of ingenious speculation. Anything calculated to dis- courage the criticism of existing analytical methods and the device of new ones is, therefore, much to be deprecated. There is too gcod a case against the standardization of methods of chemical analysis to make elaboration of argument necesaary or wealth of instance and illustration desirable.The plain facts constitute so crushing an indictment, and appeal so strongly to every thinking chemist who is jealous of the dignity of his chosen profession, that their statement without ornament or rhetoric must suffice for conviction. But as there is, nevertheless, a respectable body of opinion in the contrary sense, it may not be superfluous to register a formal pronouncement that standardization is undesirable except in cases where arbitrary methods of examina- tion cannot be avoided. The spirit which is dissatisfied with the present status of the analyst, that does not hesitate to indicate the causes of errors and discrepancies, and is eager to incur serious and voluntary labour in order to diminish the frequencyof their occurrence, is so entirely admirable that if directed towards a better end than the construction of a cramping and ineffective code it could not fail to animate a larger and more liberal project with its own zeal.There are many branches of analysis (as distinct from arbitrary methods of chemical examination) in which there still exists some doubt whether certain of the recognised methods are actually as reliable as the users of them maintain. I t would be outside the scope of this paper to put forward a list of these discussable methods; every practising chemist can provide illustrations from his daily work, and the specialiste A whole vista of trouble and dispute can be foreseen. But there are still further drawbacks of a kind yet more serious. To turn to the construction of a positive policy.324 THE ANALYST. in any well-defined branch of analytical work would, if they agreed in nothing else, be thoroughly in accord with the idea that revision of many accepted processes, after proper experimental inquiry, might be usefully undertaken.To realize this idea it seem8 to me that there might be formed standing committees of chemists interested. in particular branches of analytical work ; that those committees should meet as often as is practicable, in order that their members might confer ; that each should draw up a programme of work to be carried out by selected members of the committee, the programme to be directed to the examination of existing methods of analysis, to the elimination of sources of error which might be discovered, and to the device of new and improved processes; that the results of the deliberation of these com- mittees should be published periodically, in order that chemists at large might be informed of the progress made and warned against the errors discovered. All these inquiries would be of vast value to professional chemists, who would be in a position to judge how far the methods indicated might usefully be received and adopted. I t will be seen that the programme suggested differs radically from that to which I have ventured to take exception. The latter involves the pronouncement of a defined method for the determination of a given substance; the former is directed to the discovery of litent errorg, and to the provision of processes free from these. I n the one case a set of rules is imposed from without ; in the other, advance and improvement will take place from within, unfettered by canon and unhampered by rule. I think it not too much to say that the whole spirit of science, which is instinct with the love of freedom and the defiance of mere authority, will be found favourable to the plan which I have put forward, and will unconditionally condemn any attempt to standardize methods of chemical analysis.