Avogadro,the Chemists,and Historians of Chemistry: Part 1

In many respects the information provided by Avogadro's hypothesis with respect to molecules is logically prior to Dalton's discussion of atoms. The acknowledgement that chemistry is the science of molecules rather than of atoms came late in the nineteenth century. This change of view is the subject of Meldrum's Andrew N. Avogadro and Dalton: The standing in chemistry of their hypotheses (Aberdeen, 1904); note the order of names. 1 shall return to this when discussing the eventual acceptance of Avogadro's hypothesis at the end of this paper.

Russell C. A. , The history of valency (Leicester, 1971), 38.

Greenaway F. , “Avogadro, Amedeo”, in A biographical dictionary of scientists , ed. Williams T. I. (London, 1969), 23. In citing Russell and Greenaway I am certainly not trying to pillory the work of two fine historians of chemistry. Plenty of other examples of such judgments could have been chosen from among historians of chemistry or historians of science as a whole. They were commonplace a few years ago. Here and below I quote representative comments by historians. Since virtually every historian who has discussed the chemistry of the nineteenth century during the last hundred years has mentioned Avogadro's hypothesis and its fate, a complete bibliography would be a complete bibliography of the history of nineteenth century chemistry.

Leicester H. M. , The historical background of chemistry (New York, 1956), 158–9; Pauling L. , “Amedeo Avogadro”, Science, cxxiv (1956), 710–13; Berry A. J. , Modern chemistry (Cambridge, 1946), 1–6.

Partington J. R. , “Amedeo Avogadro (1776–1856)”, Nature, clxxviii (1956), 8–9; idem, A history of chemistry , iv (London, 1964), 213–18; Ihde A. J. , “The Karlsruhe Congress: A centennial retrospect”, Journal of chemical education, xxxviii (1961), 83–84; idem, The development of modern chemistry (New York, 1964), 120–2.

Graebe C. , “Der Entwicklungsgang der Avogadroschen Theorie”, Journal für praktische Chemie, cxcv (1913), 145–208; Meldrum , op. cit. (ref. 1); Freund I. , The study of chemical composition: An account of its method and historical development (Cambridge, 1904; repr. New York, 1968), 317–59; Muir M. M. Pattison , A history of chemical theories and laws (New York, 1907), chs iv-v.

Freund , op. cit. (ref. 6), 322–9.

This physical argument in support of the hypothesis was first used by Prout William , Chemistry, meteorology, and the function of digestion (8th Bridgewater Treatise, London, 1834), 62–64; followed (almost certainly independently) by Dumas J. B. A. , Leçons sur la philosophie chimique professées…en 1836 (Brussels, 1839 edn, repr. Brussels, 1972), 224–5.

Muir , op. cit. (ref. 6), 146.

Cannizzaro S. , “Sunto di un corso di filosofia chimica”, Il nuovo cimento, vii (1858), 321–66; English trans, in Alembic Club reprints, xviii (reissued Edinburgh, 1966; hereafter ACR); see also Cannizzaro's remarks at the Karlsruhe Conference of 1860, as recorded by A. Wurtz and published by R. Anschütz as an appendix to his August Kekulé, i (Berlin, 1929), 671–88.

Wurtz A. , A history of chemical theory from the age of Lavoisier to the present time , trans, by Watts H. (London, 1969), 41–45; Ladenburg A. , Lectures on the history of the development of chemistry since the time of Lavoisier , trans, by Dobbin L. (Edinburgh, 1900, from the 2nd (1887) German edn). 61–64.

Tilden W. A. , “Cannizzaro memorial lecture”, Journal of the Chemical Society, ci (1912), 1677–93.

Kopp H. , Die Entwickelung der Chemie in der neueren Zeit (Munich, 1837), 349–54.

Knight D. M. , Atoms and elements: A study of theories of matter in England in the nineteenth century (London, 1967), 83–104; idem. The transcendental part of chemistry (Folkestone, 1978); Mauskopf S. M. , “The atomic structural theories of Ampère and Gaudin: Molecular speculation and Avogadro's hypothesis”, Isis, lx (1969), 61–74; idem, “Crystals and compounds: Molecular structure and composition in nineteenth-century French science”, Transactions of the American Philosophical Society, lxvi (1976), pt 3. Cole T. M. has extended Mauskopf's discussion of Gaudin, one of Avogadro's few early followers, in “Early atomic speculations of Marc Antoine Gaudin: Avogadro's hypothesis and the periodic system”, Isis, lxvi (1975), 334–60.

Rocke A. J. , “Atoms and equivalents: The early development of the chemical atomic theory”, Historical studies in the physical sciences, ix (1978), 225–63. See also idem, “Gay-Lussac and Dumas: Adherents of the Avogadro-Ampère hypothesis?”, Isis, lxix (1978), 595–600.

Crosland M. P. , “Avogadro, Amedeo”, in Dictionary of scientific biography, ed. Gillispie C. C. , i (New York, 1970), 343–50.

Bonner J. K. , “Amedeo Avogadro: A reassessment of his research and its place in early nineteenth century science” (Ph.D. dissertation, Johns Hopkins University, 1974). It is unfortunate that this study is available only in dissertation form (University Microfilms 74–29, 018).

Many surveys of the history of chemistry have pursued this line: cf. ref. 4 above. See also general histories of sciences such as Singer C. , A short history of scientific ideas to 1900 (Oxford, 1959), 343–5; or Wightman W. P. D. , The growth of scientific ideas (New Haven, 1951), 229–43. Mason S. F. , in A history of the sciences (revised edn, New York, 1962), 449–67, is much more sensitive to the complexities of nineteenth century chemists. Note also the prevalence of the ‘straight development’ picture in modern textbooks, of which Holton G. and Roller's D. H. D. Foundations of modern physical science (Reading, Mass., 1965), is one of the more historically aware.

Frické M. , “The rejection of Avogadro's hypothesis”, in Method and appraisal in the physical sciences, ed. Howson C. (Cambridge, 1976), 277–308, p. 300. The most prominent name mentioned by Frické is that of Thomas Thomson!.

Brooke J. H. , “Avogadro's hypothesis and its fate: A case study in the failure of case-studies”, History of science, xix (1981), 235–73, pp. 246–9.

It would be nice to know which these were.

Frické , op. cit. (ref. 19), 278, 284, 299ff.

There is a good concise account in Ihde, Development (ref. 4), 109–11. When only one compound was formed between two elements, the atomic combination was assumed to be one to one as in HO, water. Where two elements formed more than one compound—most notably oxygen and nitrogen—more complicated combinations were assumed. There was a spatial rationale for Dalton's rules, but this was generally ignored by his contemporaries.

Frické , op. cit. (ref. 19), 285–6.

Crosland M. , Gay-Lussac, scientist and bourgeois (Cambridge, 1978), 93. Some of the ascriptions of ‘pieces’ could be questioned.

Dalton J. , A new system of chemical philosophy , i, part II (London, 1810), 556–9.

The extremely strong sense of personal property in chemistry in the nineteenth century is at variance with the sociologists' norm of ‘communality’ ( Merton R. K. , “The normative structure of science (1942)”, in his The sociology of science (Chicago, 1973), 267–78, pp. 273–5 —and throughout his works). At least in those days. publication was not equivalent to sharing freely with the scientific community.

For elements, as opposed to compounds, the ‘formula’ corresponds to the ‘atomicity’: O₂ and P₄ are ‘diatomic’ and ‘tetratomic’, respectively.

For a careful distinction between combining proportions (what I have called combining weight) determined empirically, and equivalent defined as atomic weight divided by valence, see Rocke , “Atoms and equivalents” (ref. 15), 226–30, esp. p. 227, n.7. Valency was of course a late nineteenth century concept. See Russell C. A. , op. cit. (ref. 2).

Baudrimont A. E. , Introduction à l'étude de la chimie par la théorie atomique (Paris, 1933), 50–51. Now that we have a whole range of volatile compounds it is possible to break out of the chemical circle through a consistent application of Avogadro's hypothesis to the physical data of vapour densities.

Laurent was the first to point out that the empiricism of these ‘equivalents’ was illusory: Chemical method (London, 1854), 1–16. This criticism has been pursued by Rocke , “Atoms and equivalents” (ref. 15), following Laurent , and also by Ladenburg , who was particularly scathing about Wollaston's empirical claims, op. cit. (ref. 11), 64–66. Combining weights are empirical, of course, but as soon as you fix on one value of the combining weight and call this the equivalent', you have made as much an assumption as any of Dalton's. The foundation of Wollaston's “Synoptic scale of chemical equivalents”, Philosophical transactions, civ (1814), 1–22, was the reaction of carbon with oxygen to give “carbonic acid” and “carbonic oxide”. In deciding that the former contained two equivalents of oxygen, and the latter one, Wollaston was hardly empirical. But the assumption was made with less show than when Dalton explicitly used assumptions of simplicity almost with sleight of hand, indeed and went apparently undetected until Laurent made a polemical point of it to emphasize the need for his own reforms. What is important in this story is that most chemists, misguidedly perhaps, believed that equivalents were at least more empirical (and indeed more chemical) than atoms, and exercised a clear choice for equivalents.

Avogadro A. , “Essai d'une manière de déterminer les masses relatives des molécules élémentaires des corps, et les proportions selon lesquelles elles entrent dans ces combinaisons”, Journal de physique, lxxiii (1811), 58–76, p. 62. The mentions of Dalton were presumably for ‘sales’ purposes. Many ‘inductivist’ historians have been misled by Avogadro, and so, ironically, has Frické; see Brooke , op. cit. (ref. 20). Interestingly, as Morselli has pointed out, Avogadro's draft for the “Essai” does not contain the remark in the final section (p.76) about using Gay-Lussac's results to add precision to Dalton: Morselli M. A. , “The manuscript of Avogadro's ‘Essai’ (1811)”, Ambix, xxvii (1980), 147–72, p. 152.

The most thorough discussion of the early nineteenth century French usage of ‘molécule’ is in Bonner, op. cit. (ref. 17), 186–214. He ascribes its introduction into chemistry to Macquer (1777) and Guyton (1786). Its general use seems to have been rather earlier. Both the OED and Robert's Dictionnaire (1959 edn) trace it to the mechanical philosophy. As the OED says, “The word seems to have arisen in the seventeenth century in the discussions initiated by the physical speculations of Descartes”. In the organic realm Buffon's ‘molécules organiques’ attracted considerable attention. We can also add to Bonner's account that since the word, if not the concept of, ‘molécule’ was common both to the Foucroy-Thenard school of chemistry and to the Berthollet-Gay-Lussac tradition, Avogadro's use of the term does not of itself locate him in the latter, as Bonner claims. Dalton's atomic theory arrived in France via Jean Riffault's 1809 translation of the 3rd edn (1807) of Thomas Thomson's System of chemistry. Berthollet sponsored this translation, and wrote an ‘introduction’ spelling out his reservations about atoms; see Crosland M. P. , “The first reception of Dalton's atomic theory in France”, in John Dalton and the progress of science , ed. Cardwell D. S. L. (Manchester, 1968), 274–87. This translation of Thomson was Avogadro's source for his knowledge of Dalton; see Bonner , ibid., 45, 125–9, 179ff.

ibid., 152–66. See also Crosland M. P. , “The origins of Gay-Lussac's law of combining volumes of gases”, Annals of science, xvii (1961), 1–26.

Ibid ., 55, 125, 291–301. 311–12. In order to establish the isolation of Avogadro's position with respect to chemistry, Bonner argues that the atomic theory made rapid strides in France in the second and third decades of the century ( ibid ., 55, 197–8): This is considerable exaggeration. For the fall of Berthollet, see Fox R. “The rise and fall of Laplacian physics”, Historical studies in the physical sciences, iv (1974), 89–136. See also Crosland M. P. , The Society of Arcueil, a view of French science at the time of Napoleon I (London, 1967), viii.

Fox R. , The caloric theory of gases from Lavoisier to Regnault (Oxford, 1971). 196.

Crosland Contrast , op. cit. (ref. 16) and Frické , op. cit. (ref. 19) with Pauling , op. cit. (ref. 4).

(i) op. cit. (ref. 32); (ii) “Mémoire sur les masses relatives des molécules des corps simples”, Journal de physique, lxxviii (1814), 131–56; (iii) “Nouvelles considérations sur la théorie des proportions déterminées dans les combinaisons et sur la détermination des masses des molécules des corps”, Memorie della Reale Accademia delle Scienze di Torino, xxvi (1821). 1–162; (iv) “Mémoire sur la manière de ramener les composés organiques aux lois ordinaires des proportions déterminées”, ibid., 440–506.

Op. cit., (ref. 32). 58. Avogadro's immediate debt to Gay-Lussac was made explicit in the opening sentence of this paper.

Ibid ., 58, 60–61. The corollary was the necessary consequence of the hypothesis that equal volumes contain equal molecules, and the empirical observation that one volume of oxygen forms two volumes of steam: The oxygen molecule must be divisible into two parts at least (one for each steam molecule). For Avogadro's conviction in the simplicity of Nature, see op. cit. (ref. 38, ii), 132.

See ref. 10 above.

Op. cit. (ref. 32), 71–72, and throughout the four papers. Coley N. G. , “The physico-chemical studies of Amedeo Avogadro”, Annals of science, xx (1964), 195–210, points out (p. 198) that Berzelius used much the same approach when in 1813 he introduced the elementary symbols we use today. Initially these referred to ‘volumes’ rather than atoms, even of the non-volatile elements, where the ‘volume’ had to be inferred from chemical data. As with the non-empirical nature of Wollaston's equivalents (see ref. 31 above), this ‘danger’ was not generally perceived.

See Baudrimont (ref. 30) on the inevitable circularity of arguments confined within chemistry. Avogadro may have invented a way out of the circle by applying physical evidence to chemical weights, but when he got his evidence from chemistry, and reasoned from chemical weight to chemical weight, he was no better off than anyone else. Note also that in relying on the combining weights found by other chemists largely in solid or in liquid conditions, Avogadro was doing violence to the prescriptions of Berthollet's chemistry which Bonner has identified as his major guiding influence: Only in gaseous conditions did ‘true’ neutralization occur; elsewhere, chemical proportions should vary; cf. refs. 33 and 34, above, and Brooke, op. cit. (ref. 20), ref. 76.

Partington , Nature (ref. 5); Pauling , op. cit. (ref. 4). Hagiography is not perhaps inappropriate in centennial celebrations.

Meyer L. , Die modernen Theorien der Chemie , 5th edn (Breslau, 1884), 25; Wurtz , op. cit. (ref. 11), 42; Debus H. , Ueber einige Fundamentalsätze der Chemie, inbesondere das Dalton-Avogadro'sche Gesetz (Cassel, 1894), 2.

Ibid.

Cf. Greenaway , op. cit. (ref. 3).

Cannizzaro also needed a lot more information than was provided by the application of Avogadro's hypothesis to reach his system of atomic weights. See below.

Gaudin M. A. A. , “Recherches sur la structure intime des corps inorganiques définis…”, Annales de chimie, lii (1833), 113–33, p. 115; Baudrimont A. E. Traité de chimie générale et expérimentale , i (Paris, 1844), 8–10; Laurent A. , “Recherches sur les combinaisons azotées”, Annales de chimie, xviii (1846), 266–98, p. 296. As the first part of Laurent's paper makes clear, he was led to his ideas by Gerhardt's analogous arguments for the reform of molecular weights: Water (for example) was never eliminated from an organic reaction in a unit less than H⁴ O². This, the smallest quantity of water found in practice, should be assigned the conventional formula of water, H²O. Most organic formulae should be halved to correspond to this. Gerhardt never accepted Avogadro's hypothesis, but operational reasoning about ‘the smallest quantity yet found’ was a commonplace among his several followers in the 1850. who contributed to the eventual acceptance of the hypothesis. See below.

In an address to the Società Italiana por il Progresso delle Scienze in 1875, quoted by Graebe , op. cit. (ref. 6), 172, from Gazzetta chimica italiana, v (1875), 354. Cannizzaro said that the idea of his system arose from his reading one of Gaudin's papers while considering Gerhardt's system in general. Frické is clearly wrong to claim (as he must if he is to maintain that the Berzelian programme was the most progressive before Karlsruhe) that “Cannizzaro produced a completely new and original solution to the independence problem”, op. cit. (ref. 19), 295.

Avogadro , op. cit. (ref. 38, iii), 7–10. Besides, like Lavoisier's operational definition of element which it so strongly resembles, this definition of atom is a chemist's definition. Reductionist chemists like Berthollet and Davy and Avogadro found this sort of working definition particularly objectionable, and were always hoping that Lavoisier's elements would be broken down, and/or found to be deducible from fewer and higher principles; see Knight , Transcendental chemistry (ref. 14), passim.

Bonner , op. cit. (ref. 17), 4, 201–14. Crosland's op. cit. (ref. 16), 344, is the main object of Bonner's criticism, but the mistaken claim is as old as Cannizzaro , ACR (ref. 10), 2.

Avogadro , op. cit. (ref. 32), 60–61; op. cit. (ref. 38, iii), 14. Mundy B. W. , “Avogadro on the degree of submolecularity of molecules”, Chymia, xii (1967), 151–5, has identified two rather ambiguous passages in Avogadro, op. cit. (ref. 38, ii), in which Avogadro seems to suggest that in different reactions the oxygen molecule may divide into four and five parts—which would require the fundamental unit of oxygen to be at most one-twentieth of the molecule. But this only goes to underline my point that Avogadro was not really interested in units smaller than the ‘molecule’. As Frické says, the corollary to the equal numbers hypothesis is entirely ad hoc: “There is no experimental outcome that Avogadro's hypotheses could not ‘explain’. For instance, if six volumes of steam had been produced, Avogadro would have described the reaction as: For a new reaction, Avogadro's hypotheses made no predictions either about the degree of submolecularity or the volume of the resulting gas…”; op. cit. (ref. 19), 290.

Bonner calls the degree of submolecularity the “division-number”. Having gone carefully through Avogadro's argument that it is not fixed, he finally admits that “One of his most serious errors was to suppose that the sub-division number was two in practically all cases”; op. cit. (ref. 17), 242. Bonner knows Avogadro's work more intimately than I do, but I know of no exceptions, and Bonner does not cite any. In the manuscript of his “Essai”, Avogadro referred to the possibility of triple division, but this vanished from the published version, never to reappear; see Morselli , op. cit. (ref. 32), 157.

Divers E. , “The atomic theory without hypothesis”, BAAS report, lxxii (1902), 557–75. It is ironic that in practice Berzelius, the pure chemist, paid much more attention to physical information than Avogadro, usually credited with first bringing physics to bear on chemistry.

Ihde A. J. , “Editorially speaking”, Journal of chemical education, xxxviii (1961), 483, makes the salient and sympathetic point that “A discipline with growing pains is sure to ignore useful concepts which appear before their significance is obvious”, citing the fate of Semmelweiss and Mendel as other examples. Knight , Atoms and elements (ref. 14), 90, makes similar remarks.

Brooke , op. cit. (ref. 20).

Crosland , op. cit. (ref. 16), 345–6; Frické , op. cit. (ref. 19); Brooke , op. cit. (ref. 20), 255–6. Lothar Meyer in 1885 claimed that “No valid and important objections have ever been made against this hypothesis. Whenever any discussion has taken place, this has only been about whether the adoption of the hypothesis would be convenient. It could never be shown to be demonstrably in error”; op. cit. (ref. 45), 37. Equally, it could not be shown to be demonstrably correct before about 1870.

For exact citations, see Partington , History (ref. 5), passim.

Partington , Nature (ref. 5), 8.

Henry W. C. , Philosophical magazine, v (1834), 33–39.

Gmelin L. , Handbook of chemistry , i, trans. Watts H. (London, 1848), 143, 147. For Gmelin's notice of Gaudin, see Cole , op. cit. (ref. 14), 359. For the strength of Gmelin's influence, see Brock W. H. , “The society for the perpetuation of Gmelin: The Cavendish society, 1846–1872”, Annals of science, xxxv (1978), 599–617.

Dumas J. B. A. , Traité de chimie appliquée aux arts , i (Paris, 1828), xxv ff. (I shall return below to Dumas's renunciation of Avogadro in 1836); Daubeny C. G. B. , An introduction to the atomic theory , 1st edn (Oxford & London, 1831), 46–47; Baudrimont , op. cit. (ref. 49), 150 and passim.

Freund , op. cit. (ref. 6), 331.

Cannizzaro , ACR (ref. 10), 2–3. It should be noted that combination between like atoms was also forbidden by Dalton's original gas model of mutual repulsion between like atoms. This was the basis of Henry's criticism of Prout mentioned in the text at ref. 61 above.

For a comparative table of Berzelius's values of 1814, 1818, and 1826 (with modern values) see Ihde , Development (ref. 5), 142–3. For the difficulties Berzelius had in the first place in reconciling his own chemical proportions with Dalton's physical atoms and Gay-Lussac's reacting volumes, see Anders Lundgren, Berzelius och den kemiska atomteorin (with a summary in English: “Berzelius and the chemical atomic theory”, Uppsala, 1979). For the inconclusiveness of any single method of atomic weight determination, cf. Brooke , op. cit. (ref. 20), 242.

By this time, too, Berzelius had another preoccupation which stood in the way of his acceptance of Avogadro's hypothesis: Its acceptance would have destroyed his whole system of formulation of organic compounds, particularly the acids: Acetic acid was written C₄H₆O₃ + H₂O, by analogy with sulphuric, SO₃ + H₂O. If the formula of acetic acid were halved, as required when Avogadro's hypothesis is consistently applied in organic chemistry, then the analogy would vanish. John Brooke has argued this point frequently, most notably in “Chlorine substitution and the future of organic chemistry: Methodological issues in the Laurent-Berzelius correspondence (1843–44)”, Studies in the history and philosophy of science, iv (1974), 47–94. Cf. also idem, op. cit. (ref. 20).

Thomson T. , An attempt to establish the first principles of chemistry by experiment (2 vols, London, 1825). For a brief account of Thomson's problems, and Berzelius's scathing response, see Partington, History (ref. 5), 225–6: For a fuller account see Knight , Transcendental chemistry (ref. 14), 161–77. For the eclipse of Thomson's reputation (for this and other reasons) see Morrell J. B. , “The chemist breeders: The research schools of Liebig and Thomas Thomson”, Ambix, xix (1972), 1–46.

Dumas J. B. A. , “Mémoire sur quelques points de la théorie atomistique”, Annales de chimie, xxxiii (1826), 337–91 + plate (p. 337). For a discussion of Dumas's adoption of Avogadro, and a claim that Gay-Lussac also held the hypothesis at this time, see Rocke , “Gay-Lussac and Dumas” (ref. 15).

Gaudin , op. cit. (ref. 49). Dumas discussed Gaudin's work in “Considérations générales sur la composition théorique des matières organiques”, 3^e partie, Journal de pharmacie , xx (1834), 261–94, and in his 1836 Leçons he gave a careful exposition of the lack of a clear relationship between vapour densities and ‘chemical atoms’, op. cit. (ref. 8), 225–8.

Ibid ., 228–9. Cf. his notorious “banishment of the word atom from chemistry”, at p. 246. Dumas's flirtation with atomism between 1826 and 1836 is the focus of Buchdahl G. , “Sources of scepticism in atomic theory”, British journal for the philosophy of science, x (1959), 120–34, which contrasts Dumas and Berzelius. In view of my claim above that Berzelius would have had no use for Avogadro, it is worth remarking that Berzelius greeted Dumas's earliest efforts at determinacy with interest, but no acceptance: Jahresbericht über die Fortschritte der Chemie , vii (1827), 79–91.

K. J. S. Boughey's claims are altogether too great in this respect: “Studies in the role of positivism in nineteenth-century French chemistry” (Ph.D. thesis, University of Leeds, 1972)—though much of his analysis of French attitudes to atoms and molecules is of great value.

Wollaston , op. cit. (ref. 31), 1.

Letter from Liebig to the French chemist Jules Pelouze of 14 October 1838, printed as Appendix to Fox R. , op. cit. (ref. 36), 319–20. The next year Liebig made this campaign public: He published Berzelius's “Ueber einige Fragen des Tages in der organischen Chemie”, Annalen, xxxi (1839), 1–35, and turned an editorial apology for not being able to print Berzelius's barred symbols into an attack on Berzelius's atomic weights themselves: “Bemerkungen zu vorstehender Abhandlung”, ibid ., 35–8. He even tried to recruit Berzelius to the campaign in a personal letter of 5 September 1839, but Berzelius not unnaturally declined. This exchange is discussed by Rocke A. J. , “Origins of the structural theory in organic chemistry” (Ph.D. dissertation, University of Wisconsin – Madison, 1975), 482–5.

Though in Britain, at least, equivalent weights and ‘normal solutions’ have taken a long time to die.

Cf. ref. 29, above.

Kekulé , one of the founders of valency theory, seldom strayed outside the bounds of organic chemistry. Throughout his life he maintained that valency was a constant numerical property of elements, which involved ad hoc adjustments to his theory in order to save the phenomena which others accounted for with variable valency. See Russell , op. cit. (ref. 2). ch. x.

One of the conclusions of Morrell's excellent contrast of Liebig and Thomas Thomson as teachers of chemistry, op. cit. (ref. 68), is that the former was more successful because the techniques of systematic organic analysis were less complex than the techniques of mineral chemistry on which Thomson tried to build. It should also be noted that in organic chemistry the problems of variable combining weights are almost non-existent, making systematic teaching on the basis of Liebig's equivalents possible. Thomson's struggles with the atomic theory have been noted above.

The debate over equivalents versus atoms is the theme of Knight , Atoms and elements (ref. 14). The origins of valency theory and its bearing on this problem are discussed by Russell, op. cit. (ref. 2). See also Freund , op. cit. (ref. 6). ch. xvii.

The use of analogical argument from inorganic to organic chemistry and vice-versa is the main focus of Brooke J. H. , “The rôle of analogical argument in the development of organic chemistry” (Ph.D. thesis, University of Cambridge. 1969). For one part of the story, concentrating on Berzelius and Laurent, see idem. op. cit. (ref. 67). Cf. also idem. op. cit. (ref. 20), ref. 55.

For British atomic debates, see Brock W. H. and Knight D. M. “The atomic debates” in The atomic debates , ed. Brock W. H. (Leicester, 1967), 1–30; and Knight , opera cit. (ref. 14), passim. For a useful review of the issues in the French debates, see Duhem P. , “Notation atomique et hypothèses atomistiques”, Revue des questions scientifiques, xxxi (1892), 391–454 (though Duhem's discussion is more philosophical than strictly historical). I shall return to these debates briefly below. For the philosophical positions of Mach and Ostwald, and their opposition to atomism, see Hiebert E. N. , “Ernst Mach”, in Gillispie C. C. (ed.), op. cit. (ref. 16), viii (New York, 1973), 595–607, pp. 596, 598; idem and Korber H-G. , “Ostwald, Wilhelm”, in ibid ., xv (New York, 1978), 455–69, pp. 462–4; see also Alexander P. on Mach, and M. Capek on Ostwald, in The encyclopedia of philosophy , ed. Edwards P. (New York, 1967), v, 115–19; vi, 5–7, respectively. For a general survey, see also Nye M. J. , “The nineteenth-century atomic debates and the dilemma of an ‘indifferent hypothesis’”, Studies in the history and philosophy of science, vii (1976), 245–68.

Meyer L. , “Anmerkung 1”, in the German translation of Cannizzaro's op. cit. (ref. 10), “Abriss eines Lehrganges…”, Ostwalds Klassiker, xxx (Leipzig, 1891), 50–60. In this long note giving the historical background to Cannizzaro, Meyer mentioned no German names, except that of Gmelin, whom he cited to contrast his equivalent weights with Gerhardt's molecular weights.

Brooke , op. cit. (ref. 80), quotes passim many chemists in agreement on this point.

The age-old gap between theory (alchemical, Paracelsian, Stahlian) and practice (chemical cookery) had hardly begun to narrow in the 1830s and '40s.

Fisher N. W. , “Organic classification before Kekulé”, Ambix, xx (1973), 106–31, 209–33. The point that organic ‘units’ were compound was neatly and famously made by Dumas and Liebig in 1837: “In inorganic chemistry the radicals are simple; in organic chemistry they are compound. That is the difference”; Comptes rendus, v (1837), 567–72, p. 567. This “only difference” in principle made a considerable difference in practice.

The sterility of such arguments in the 1840s called forth Russell's exasperated comment quoted at ref. 2, above. For a discussion of ether, see Liebig, op. cit. (ref. 98). See also Brooke , op. cit. (ref. 20), ref. 133.

Fisher N. W. , “Kekulé and organic classification”, Ambix, xxi (1974), 29–52. The point that every atom in the organic molecule was significant, and that division into hypothetical groups distorted this significance, had been made by Baudrimont as early as 1833 ( Journal de chimie médicale , ix (1833), 39–41, pp. 40–41), in a note to the Académie about Gaudin's op. cit. (ref. 49). (Significantly, Baudrimont was an early adherent of Avogadro's hypothesis.) But in 1833 chemists lacked the analytical ability to explore every atom in the molecule, so Baudrimont treated organic molecules as wholes, as Dumas and Gerhardt were to do in the 1840s, as the only way to avoid this distortion. As John Brooke has pointed out to me privately, Laurent was one chemist who was certainly concerned with the exact number of atoms combining with his ‘nuclei’. It was largely this that led him to accept Avogadro's hypothesis in 1846. But Laurent was no normal chemist.

For some introductions to what ‘physics’ might have meant in this period, see Fox , op. cit. (ref. 35); Smith C. , “‘Mechanical philosophy’ and the emergence of physics in Britain, 1800–1850”, Annals of science, xxxiii (1976), 3–29, is a useful introduction, not least in stressing that natural philosophers of the British (especially Scottish) school were as concerned to demarcate their field of interest from that of the chemists as were the chemists to remain independent of the physicists. See also Crosland M. & Smith C. , “The transmission of physics from France to Britain: 1800–1840”, Historical studies in the physical sciences, ix (1978), 1–61; Caneva Kenneth L. , “From galvanism to electrodynamics: The transformation of German physics and its social context”, ibid ., 63–159. It is important to understand the difference between ‘chemical’ and ‘physical’ properties, since arguments over the acceptability in chemistry of vapour density data hinged on the acceptability of physical data more generally. ‘Chemical properties’ were those properties manifested on the occasion of chemical change; the properties of unreacting substances (optical, electrical, thermal, etc.) were generally taken to be physical. There were many discussions of the distinction in the nineteenth century literature. One of the neatest encapsulations is that of Muir M. M. P. , The story of the chemical elements (London, 1897), 164–5. Another common definition of chemistry and physics is also apposite to the fate of Avogadro's hypothesis: “We may define chemistry as the science of atoms, and physics as the science of molecules”: Kekulé A. , “The scientific aims and achievements of chemistry”, Nature, xviii (1878), 210–13, p. 211. Kekulé went on to say that theoretical chemistry was dependent on theoretical physics, but that since physicists had shown little interest in molecules “it becomes clear why for the present theoretical-chemical investigation has principally turned its attention to those questions which are more or less independent of physics”.

G. Fownes's A manual of elementary chemistry, theoretical and practical is typical. In the sixth of its fourteen editions (1856), edited after the author's death in 1849 by Jones H. Bence and Hofmann A. W. , 103 of 715 pages are devoted to physics, but there is no integration with the chemistry. The 11th edn (1873), ed. Watts H. , has an almost identical 104 pages of physics, while the extensive revision of the chemistry has swollen the whole to 1,026pp. Neither edition makes any mention of Avogadro, with whom Watts, at least, must have been familiar since 1869 when he had translated Wurtz's op. cit. (ref. 11). For scepticism towards atomism in what Brock and Knight call “the British textbook tradition”, in which they locate Fownes, see Brock and Knight , op. cit. (ref. 81), 10.

This is far too broad a topic to document here. I am working on a study of the ideology of early nineteenth century chemistry and the chemists' sense of identity. On the need to maintain the purity of chemistry unadulterated by physics, cf. Brooke , op. cit. (ref. 20), ref. 134.

Gerhardt C. F. , “Exposé sommaire des travaux de chimie parus en 1840 en France et à l'étranger”, Revue scientifique, iv (1841), 145–219, p. 209. Gerhardt's remarks are quoted, and the reaction of Biot briefly discussed, in Fisher N. W. , “Wislicenus and lactic acid: The chemical background to van't Hoff's hypothesis”, in van't Hoff-Le Bel centennial, ed. Ramsay O. B. (Washington, DC, 1975). 33–54, pp. 33–34.

Gerhardt used H. Kopp's data on the boiling points of methyl and ethyl compounds as one of the bases of his homologous series; the idea of series came first, however. See Fisher , op. cit. (ref. 85, part 2), 212. Kopp was arguably the first physical chemist, in the sense that (unlike Avogadro) he worked within the discipline of chemistry and on the physical properties of chemical substances. As a German chemist he ignored Avogadro's hypothesis (cf. ref. 82, above), and investigated such physical properties as boiling points, melting points and molecular heats which, being governed by additive rules, are as easily interpretable in equivalent as in atomic terms. Though it is probable from his discussion in op. cit. (ref. 13) that he had known about Avogadro's hypothesis for some time, he adopted it only in the late 1860s, after the Karlsruhe congress, in which he played a fairly central part; see Anschütz , op. cit. (ref. 10).

Dumas , op. cit. (ref. 8), 351.

Ibid ., 352–3. There was of course some doubt and debate in the early part of the century whether current electricity was ‘physics’ or ‘chemistry’. The success of electrolysis in the years immediately following Volta's discovery of the pile located the most exciting electrical discoveries in the realm of chemistry. The discovery and exploitation of electromagnetism (which had no implications for chemistry) by Oersted, Ampère, and Faraday seems to have swung electricity back towards physics, and in the passage quoted Dumas seems to have no doubt of this. In the passage immediately following (ibid., 356–9) he welcomed Faraday's newly-discovered electrochemical equivalents, which he thought might be as important to chemists in their search for determinate weights as were Dulong and Petit's specific heats. Significantly, no chemists seem to have followed this suggestion, and further investigation of the laws of electrolysis was confined to physicists; see Partington , History (ref. 5), 122 ff. For a review of some of Ampère's electrical ideas, which were the object of Dumas's criticism, see Caneva Kenneth L. , “Ampère, the etherians, and the Oersted connexion”, The British journal for the history of science, xiii (1980), 121–38.

Avogadro , op. cit. (ref. 32), 76.

In my broader study noted in ref. 90 above, I am looking at inductivist rhetoric, which seems to have held sway at least until Liebig's attack on Bacon in Ueber Francis Bacon von Verulam, und die Methode der Naturforschung (Munich, 1863). For the resistance of chemists to deductive arguments see below.

Gaudin , op. cit. (ref. 49); Berzelius , Jahresbericht über die Fortschritte der physikalischen Wissenschaften, xiv (1834), 84–86; deduction, 84; seduction, 86. The untypical mildness of Berzelius's comments can perhaps be ascribed to the fact that from Gaudin he turned directly to Baudrimont's related ideas, which he subjected to ridicule.

Liebig J. , “Ueber die Constitution des Aethers und seiner Verbindungen”, Annalen, ix (1834), 1–39, p. 16. The value of the specific gravity to Liebig lay solely in checking the accuracy of gravimetric analysis, the very basis of his success as a trainer of organic chemists: “In the analysis of a volatile substance, the determination of the specific gravity of vapour is a most valuable means of control over the analysis by combustion”, idem, Instructions for the chemical analysis of organic bodies , trans, by Gregory W. (Glasgow, 1839), 50. When there was any conflict, though, between vapour density and chemical evidence, the chemical evidence was to be preferred: Idem, Introduction to the first elements of chemistry, trans. by Richardson T. (London, 1837), 81–82.