The influence of air and water in the j^ro- Chemical duction of chemical phenomena is another of theory of air the most perfect doctrines of chemistry as it ^^^^ water, stands. The importance of the action of air and water in the terrestrial economy has induced some Grerman philosophers irrationally to set up the system of these two fluids into a sort of third reign, between the inorganic and the organic: but abstract chemistry has nothing to do with natural history, and regards the study of air and water from a dilferent point of view, while aware of its fundamental importance.
All chemical phenomena take place in the presence of air; and they almost invariably require the intervention of water: it is clear therefore that before we study any chemical reaction, we must be able to analyse the participation of these two fluids. Thus the chemical theory of air and water is a sort of necessary introduction to the system of chemistry, properly so called, as belonging more to method than to doctrine, and as immediately followingthe study of simple bodies. It is an historical fact that the double analysis of air and water marked the first great advance in modern chemistry...The influence of the air, not less important than that of water in chemical phenomena, was less difficult to characterize: for the air is simply a mixtui'e, and its chemical action is merely that of the gases which compose it, each of which acts as it were isolated, allowing for the diminution of intensity from its diffusion, and for the very few cases in which the accomplishment of the proposed phenomenon determines the combination of the gases in an accessory way. Chemistry has only to analyse it, leaving all other study of it to the cleimrtment of natural history. This analysis was effected in the early days of modern chemistry, except that there is still some uncertainty about the proportion of carbonic acid gas, and perhaps of some other more considerable principles, as, for instance, hydrogen, the existence of which begins to be generally suspected. Though no aj^preciable change in the comijosition of the atmosphere has taken place within half a century, it is impossible to conceive that some alteration must not happen, in some direction, in course of time, among the many perturbing infltiences which act upon the mixture. Their antagonism, and that of vegetable and animal action, partly neutralizes them: but the equilibrium cannot be precise and continuous. G-eological considerations and botanical fossils lead us to suppose that at some remote periods the composition of the air must have been sensibly different: and chemists themselves have actiially established some slight periodical variations, dependent on the proportion of carbonic acid at different seasons. Our analytical resources are, however, very imperfect with regard to the accessory principles of the atmosphere; for chemists can ascertain nothing of the distinctions which are proved to exist in the best-marked localities, by their influences on living beings. The study of these variations, all-important in its way, — even as possibly indicating the limits of human life in a remote future, — belongs to natural history; and that is probably the reason why chemists trouble themselves so little about it: and if there is neglect, it should WATEK. 347 be charged upon the naturahsts. It is true that a preparation is required for their order of study, like all otliei's, — a provision of knowledge, rising from physiology to astx'onomy itseK: but the research is not especially a matter of chemical duty.
The study of water requires much more ex- „.
tended and complex researches than that of the air; and it is indispensable to the general system of chemical science: for water being a real combination, and perhaps the most perfect known to us, ma.y exercise chemical effects proper to itself, independently of those attributable to its elements, and apart from its imj^ortance as a solvent, — to say nothing of it as a simple mixture. Thus there are three aspects under which water must be considered by chemists, all distinct and all essential; and the appreciation of them has been slow and difficult, if even we may say that this fundamental examination is yet complete.
The analysis of water, represented by a quantity of hydrogen double in volume that of oxygen, and unquestionably confirmed by synthesis, is the finest of the early discoveries of modern chemistry, not only from the light it casts upon the whole of chemical phenomena and the general economy of nature, but also from its conquest of prodigious difficulties. In regard to the first view, chemical science leaves nothing to desire. Yet, a notion has arisen, in recent times, of the existence of a new and more highly oxygenated combination between the two elements of water, which may raise some interesting questions, not about the irreversible composition of water, but about the kind of chemical influence which is taken for granted in its decomposition and recomposition in a multitude of phenomena; and especially, about the true mode of union of oxygen and hydrogen in all substances, and above all in liquids, which cannot be obtained without water. Some doubts have lately been proposed about this, which seem to me to deserve mature examination.
The dissolving action of water has been the subject of a long series of laborious researches, much less difficult, and not far from complete. Yet more attention ought to be paid than is paid to the first experiment of Vauquelin, in which it is shown that watei*, saturated with one salt, remains capable of receiving another, and even acquires by that the singular property of dissolving a new quantity of the first. This experimeiit, which has been in a manner desi^ised, seems to me of the first order in its way, and a fit basis of a series of interesting researches about the apparently capricious laws of solubility, the study of which is yet essentially empirical.
Chemists were long in conceiving that water, besides being a solvent, might act in a really chemical manner, otherwise than by its elements. It seemed as if a combination so eminently neuter must be inoffensive, and inoperative, except by its decomposition. It was Proust who thought that this neutrality itself afi^orded a presumption of certain chemical affections, independently of its composition. This was the rational consideration which led him to create the important study of the hydrates, regarded as a sort of new salts, in which water plays the part, with regard to the alkalies, of a kind of hydric acid.
The examination of these combinations, and of all others that water can form with any substances without being decomposed, constitutes the thii'd and last jDart of the fundamental study of water, regarded as an indispensable preliminary to the general system of chemical studies.
DOCTRINE OF DEFINITE PROPORTIONS.
THERE are two general doctrines in chemistry, as it now exists, which present a systematic appearance, and invest the science with such rationality as it has attained. The first of these is the important doctrine of Definite Proportions.
Even if this doctrine were complete, it could exert only a secondary influence on the joctHne solution of the great problem of the science, — the study of the laws of the phenomena of composition and decomposition. The essential question is, what separations and new combinations must take place under determinate circumstances; and the theory of definite proportions affords no assistance to this kind of prevision. It proceeds, indeed, on the supposition that the question is already solved; and that it is to be taken as the point of departure for the estimate of each of the new products, — of their quantity and the j^roportion of their elements. Thus, the theory of definite proportions presents the singular scientific character of rendering rational, in its numerical details, a solution which usually remains empirical in its most important aspect.
It was natural that the founders of modern chemistry should have attended to the laws of composition and decomposition, in preference to a study which they regarded as subordinate; and it was natural also that, as the advance of science disclosed to them the vast difficulties of the main problem, they should attend more and more to the secondary study, which promised an easier and more speedy success. But the most important office of this subordinate theory, — that of supplying the defect of immediate experiment, — can be but very imperfectly fulfilled, while it is regarded apart from the principal theory; and thus, the doctrine of definite proportions will never acquire its iull scientific value till it is connected with an unquestionable basis of chemical laws, of which it will be the indispensable numerical complement.
Meanwhile, however, it affords a real, though secondary assistance to chemists, in renderino- their analyses more easy and more precise. Moreover, it restricts the number of cases of combination logically possible, by exhibiting the very small number of distinct proportions; and by thus diminishing the uncertainty in cases of chemical action, it is, in fact, a natural preliminary to the establishment of those chemical laws to which it will be, under another view, a necessary supplement.
.,, ^. In regard to doctrine, this theory offers a perfect type of the precise kind of rationality which must hereafter belong to Chemistry as a whole. In .,, regard to method, the inquirers who have devoted themselves to establish the theory have advanced chemical science while appearing to diverge from it; simplifying the vast problem which their successors will solve, and preparing for the disclosure of the great laws of composition and decomposition, which would be undiscoverable amidst the infinity of products, if substances could combine, within certain limits, in all imaginable proportions. Such are the claims of this theory, as to doctrine and to method. J-,. It assumed its existence and present form during the first quarter of this century: and it arose from a phenomenon discovered by Richter, and a speculative discussion established by Berthollet. — During the latter half of the last century, several chemists had observed that, in the mutual decomposition of two neutral salts, the two new salts thus formed are always equally neuter. Bergmann, among others, had steadily and specially dwelt upon this. Yet the fact was neglected or underrated till Richter, at the end of the century, generalized the observation, saw what it imported, and derived from it the fundamental law which bears his name. The J,..,, law is this: that the ponderable quantities of the different alkalies requisite to neutralize a given weight of any acid are always proportionate to HISTORY OF THE THEORY. 351 those required for the neutralization of the same weight of every other acid. This is, in fact, evidently the immediate consequence of the maintenance of neutrality after the double decomposition. Such a transformation would appear almost spontaneous if it related to a simpler and more developed science than Chemistry; but amidst its complications and the imperfection of our intellectual habits, the closest deductions are difficult if they have anv character of generality, and therefore of abstraction; and this achievement of Eichter's is, in consequence, eminently meritorious, on other grounds than its high utility.- — His law, with the complements it has since received, is the original basis of the general doctrine of definite proportions. It exhibited, in the case of a considerable number of compounds, the great end of this doctrine; viz. the assignment to every substance of a certain chemical coefficient, invariable and specific, indicating the proportions in which it can combine with each of those that have been similarly characterized. When it had been determined, by a double series of trials, what was the numerical composition of all the salts that may be formed by any one acid with the different alkalies, and any one alkali with the different acids, Richter's law enabled us to deduce immediately the proportions relating to all the compounds that can result from the binary combination of these two orders of substances. Richter himself brought his discovery up to this result, and prepared (but on a basis of experiment too narrow and imperfect) the first table of what were afterwards called chemical equivalents.
These neuti-al salts constituted a particular case, which could hardly have led on to a extenSo?'^ general theory of definite proportions. The idea of perfect neutralization must probably, at all times, have suggested to chemists that of a single proportion, on either side of which the neutrality must be destroyed; and thus the neutral salts were a natural first stage of the general theory; but they could not in themselves involve such a theory. It was Berthollet who extended the consideration of proportions to the whole of chemical phenomena. Some years after Richter's discovery he established as a fundamental principle, in his " Chemical Statics," the necessary existence of definite proportions for certain compounds of all orders; and he assigned the essential conditions of this characteristic property, which he attributed to all causes which can release the product of chemical reaction, as it forms, from the ulterior influence of the primitive agents. He thus added to Richter's restricted case the idea of a great number of cases subjected to the same principle, and able to lead on to its entire generalization. It is assigning much too little honour to Berthollet to recognize only the influence of his controversy with Proust, eminent as was the service rendered by Proust in that conflict, in establishing directly the general principle of determinate and invariable proportions.
Such was the double origin, experimental and speculative, of numerical chemistry. The next development had also a double character, arising fi'om the extension ^^ ^^^ harmony between the conception of Dr.
Dalton and the ex2:)eri mental reseai'ches of Berzelius, Gay-Lussac, and Wollaston. The inquiry was in a nascent state when Dalton's philosophic mind discerned its i^ossible generality. He proposed the great . ^. ^, Atomic theory, under which the doctrine of dennite proportions was developed to the whole extent that it has reached, and which serves as the basis of its daily application. The general jjrinciple of the theory is this: all elementary bodies are conceived of as formed of individual atoms, the different species of which unite, generally by twos, in a small number of groups, constituting compound atoms of the first order, always mechanically indivisible, but thenceforth chemically divisible, and, in their turn, constituting all the other orders of composition by a series of analogous combinations. The principle is in such harmony with scientific conceptions in all departments, that it appeared like a hapi)y generalization of the most familiar ideas of scientific men in every province of natural philosophy; and its universal and immediate admission took place as a matter of course.
It was observed by Berzelius that the deduction of the existence of definite proportions from this principle would be illusory if the combinations were not restricted to a very small number of atoms: for otherwise, — if the number THE ATOMIC THEORY. 353 was, though limited, very great, — the binary assemblages would be so multiplied that we might as well have combinations in any proportions whatever; and then the atomic theory might almost equally well represent the opposite doctrines of definite and indefinite proportions. Dalton was well aware of this; and the restrictions that he enunciated were presently declared too narrow by his successors, who found that they would not comprehend all existing combinations. His assertion was, that, in every combination, one of the immediate principles always enters for a single atom, and the other generally for a single atom also, and always for a very small number, rarely exceeding six. Taken with the exj^ansion proposed by his successors, the atomic conception evidently represents the entire doctrine of definite proportions. But it is the theory of successive multi])les, derived from the primary doctrine, which especially distinguishes Dr. Dalton's influence upon numerical chemistry. From the ground of his doctrine he easily saw that if two substances can combine in various distinct proportions, the ponderable quantities of the one which correspond, in the diffei-ent compounds, to the same weight in the other, must naturally follow the series of whole numbers, since these compounds will have resulted from the union of one atom of the second substance with one, two, three, etc., of the first: and this constitutes a principal element, then perceived for the first time, of the theory of chemical proportions.
Berzelius followed, with his vast experimental study of the whole of the important jjerzriius" ^^ points concerned in numerical chemistry, the diiferent parts of which he has done more than any other chemist to develope and systematize. He first perfected Richter's law, so as to connect it closely with the atomic theory; by which it became susceptible of the extension given to it by Berzelius himself, to all compounds of the second order. But the most important new knowledge has arisen from his numerical study of compounds of the first order. By comparing the composition of the metallic sulphurets and that of the corresponding oxides, he discovered a law, analogous to Richter's in regard to the salts. This law, — that the quantity of sulphur of the first is always proportionate to the quantity of oxygen combined with a Kke weight of the base in the second, — is now regarded, by induction, as appKcable to all the compounds of the first order to which the same degree of chemical neutrality is assignable. And again, the luminous series of the analyses of Berzelius have precisely verified in another direction the law of successive multiples discovered by Dalton in pursuance of his atomic theory.
Gay-Lussac followed, with the valuable Gav-Lussac numerical analyses he eftected by having recourse to gaseous combinations, considered, not as to weight, but to volume. He thus not only verified, in a special manner, the general principle of definite proportions, but j^resented it under a new aspect, which, by a wise induction, comprehends all possible cases, — showing that all bodies in a gaseous state combine in invariable and simple numerical relations of volume. An accessory advantage of this achievement was that the specific gravity of the gases might be obtained with a precision often comparable to that of experimental estimate. It is necessary however to warn inquirers not to be led away, in their application of the theory of volumes to substances which have never been vaporized, from the point of view which in Gay-Lussac's aj^jilication is equivalent to Dalton's, as adopted by Berzelius.
, The labours of Wollaston bore a great verilication^ P^'^"^ ^^^ establishing the doctrine of definite proportions. I do not refer chiefly to his ti-ansformation of the atomic theory into that of chemical equivalents, though it has a more positive character, and tends to restrain the student from wandering after inaccessible objects, to which the first might tempt him, if not judiciously directed. The substitution would be of high value, lao doubt, if it were not less a change of conception than an artifice of language. Nor have I in view the ingenious expedients by which Wollaston popularized numerical chemistry by rendering its use more clear and convenient. A greater service, in our present view, was his furnishing us with the indispensable complement of E/ichter's discoveiy, by establishing the theory in regard to the acid salts, since extended by analogy to the alkaline THE APPLICATIONS OF NUMERICAL CHEMISTRY. 355 salts. The case of the acid salts was j^erhaps the most unfavourable possible for the ascertainment of the principle of invariable proportions. Wollaston effected the proof in the most satisfactory manner; and this sjjecial confirmation of the principle is considered, from its nature, the most decisive of all.
Such has been the logical and historical progress of the researches which have constituted numerical chemistry as it is now. We can represent by an invariable number, appropriated to each of the different elementary bodies, their fundamental relations of chemical equivalence, whence, by very simple formulas, immediately expressing the laws just indicated, we easily pass to the numerical composition proper to each combination. No further evidence of the truth of the doctrine is needed, than the fact of so many illustrious inquirers having attained the same view by ways which each one opened for himself, and all agreeing as to its positive application to all cases of importance, differing only as to the mode of expression of the results, in as far as the atomic theory left it indeterminate, and therefore optional. But we must glance at the difficulties thrown in the way of its application by a consideration of the aggregate of chemical jihenomena, in order to form a clear idea of the final improvement of which this doctrine yet stands in need.
Among the points which are beyond dis- Scope of applij)ute, it is, first, evident, and no chemist has cation of Nuever doubted it, that substances differ as merical Chemuch in the proportion as in the nature of ""^^ry. their constituent principles. It is an axiom of chemical philosophy that an}^ change whatever in the numerical composition causes a change in the whole of the specific properties, in a more marked degree as the alteration is greater. Varied and gradual above all others as are the proportions produced by the chemical phenomena proper to living bodies, they afford a striking confirmation to this universal maxim. Therefore, in the lowest stages of chemical analysis, chemists have always endeavoured to assign, as a characteristic property, the proportion of the elements of each substance, as far as was possible: and when this was omitted, it was on the understanding that the proposed combination admitted of only a certain proportion; as in the case of the neutral salts.
Again, it has long been acknowledged that there always exists, between any two substances, a certain minimum and maximum of recijiroeal saturation, beyond or short of which all combination becomes impossible. At the utmost, certain variations, themselves restricted, have been supposed procurable. Berthollet established, more directly than any one else, the general and necessary existence of these limits of combination, — one of the principal characters which distinguish it from simple mixture. It is clear that the two extreme degrees of all combination must be subject to special and invariable proportions: and, as all agree in this, all argument about the opposite doctrines of indefinite and definite proportions is reduced to the question whether the passage from the minimum to the maximum of saturation can be effected gradually and almost imperceptibly, or whether it takes place always abruptly, through a small number of well-marked degrees.
Thirdly, the possibility and actual existence of intermediary definite proportions ai'e admitted by all chemists, who can have no other dispute than about the greater or smaller generality of such a property. We have seen that the idea of neutrality must, sooner or later, bring after it that of a determinate and unchangeable proportion; and the gradual development of chemical knowledge has extended this character to more and more varied cases. Berthollet disclosed several other causes of definite proportions, which were entirely misconceived before his time, and which may meet in almost all combinations, modifying certain circumstances of the phenomenon. The precise question now is, therefore, whether, besides these determinate compounds, subject to fixed proportions, within the two limits of possible combination, there does or does not exist, in general, a continuous series of other intermediate compounds of a less marked chai'acter; in a word, whether definite proportion constitutes the rule, as is now generally supposed, or, as Berthollet endeavoured to establish, the exception. This is now the only dispute. It is no derogation from the interest of the doctrine of definite proportions to say, as some preceding considerations compel us to do.
VALUE OF THE THEORY, 357 that the decision of this disputed point is not of the importance commonly supposed. The doctrine has tended to simj)lify the genei'al problem of chemistry; but it must not be supposed that the solution would have been impossible without this aid: — it would have been simply more difficult and less precise. The eminent chemists who concurred in establishing the doctrine were naturally engrossed by that labour; but their successors, who find numerical chemistry constituted to their hand, must beware of losing sight in it of the true scientific aim of chemistry. They must not linger in this vestibule of the science, to the neglect of the direct construction of Chemistry itself, — an enterprise scarcely begun, and to which it is high time that attention should be once more fully directed.
If we inquire, as we must do, how far the doctrine of definite jDi-oportions is irrevocably established, we shall bear in mind that the founders of numerical chemistry have accomjjlished that chief part which depends on an investigation of all known compounds, leaving only the question whether the doctrine is compatible with certain chemical phenomena, neglected during its formation, and remaining to be since referred to it.
The first general objection relates to the.
important phenomenon of dissolution, evi- dissolution^ dently 2:)ossible in an infinity of different proportions. It must be actnowledged that the distinctions between the state of dissolution and that of combination, by which the difficulty has been met, afford little satisfaction. In my opinion the only eifectual reply must consist in the extension of the principle of definite proi:)ortions to the phenomena of dissolution; and, difficult as it may be to do it, it does not seem to me impossible. The way is by the use of an hypothesis already proposed for other cases in which it miglit appear less admissible. All the successive degrees of concentration of the liquid must be regarded as simple mixtures of the small number of definite dissolutions which shall have been established, either between themselves or with the dissolvent, in the manner of habitual mixtures of water with alcohol, with svilphuric acid, etc. In any case, the positive verification of this hypothesis must be extremely delicate. Furthermore, to render the study of dissolutions fully rational, in this point of view, it is necessary to combine with it that of other analogous chemical phenojnena, relating to the absorption of gases by liquids or by porous solids. All these different modes of molecular union are often energetic enough to resist influences able to destroy certain combinations, properly so called: why should they not be, like them, subject to the rule of definite proportions, if that rule is truly a fundamental law of nature?
The next case, that of various metallic allovs alloys, is very extensive, though more particular. The difficulty lies in the question whether these are cases of combination or of mixture. The state of combination has been taken for gi'anted in the case of alloys; whereas the general application of the principle of numerical chemistry reqidres that they should be mixtures; while, again, it is diflicult to conceive of such a mixture of solids as could resist perturbing influences which would appear to be necessarily destructive; as great changes of temperature, the influence of crystallization, etc. The question can be decided only by a series of special experiments, devised to find the general limits of the permanence of unquestionable mixtures; and the results might be extended to other questions of numerical chemistry, as of certain oxides, on which explanations have been hazarded too freely. When a true chemical theory of mixtures is established on a proper l)asis of experiment, and we leave off referring to an hypothesis of mixture all cases in which combination seems susceptible of an indeterminate proportion, in order to bring them under the law of definite propoi'tions, we shall get rid of a formidable objection to the principle of numerical chemistry.