^{page 50 note 1} Novum Organum, I, 73Google Scholar.

^{page 50 note 2} Op. cit. I, 63.

^{page 50 note 3} Aristotle, A Chapter from the History of Science (London, 1864), p. 49Google Scholar.

^{page 50 note 4} Playfair, J., ‘Dissertation … Exhibiting a general view of the progress of mathematical and physical science, since the revival of letters in Europe’, Encyclopaedia Britannica (7th ed., 1842) I, PP. 453 ffGoogle Scholar.

^{page 51 note 1} History of Ancient Geography (Cambridge, 1948), p. 94Google Scholar.

^{page 51 note 2} A number of authorities are cited by Thomson, op. cit. p. 94, n. 1 and addenda p. 401. More recently H. D. P. Lee, for example, in his introduction to the Loeb edition of Aristotle's Meteorologica (1952), p. xxvii,Google Scholar puts it that the experimental method eluded the Greeks: ‘They observed but they did not experiment, and between observation and experiment there is a fundamental difference, which it is essential to recognize if the history of Greek thought is to be understood.’ Compare the rather more cautious judgement of Sambursky, , The Physical World of the Greeks, trans. Dagut, M. (London, 1956), p. 2:Google Scholar ‘With very few exceptions, the Ancient Greeks throughout a period of eight hundred years made no attempt at systematic experimentation.”

^{page 51 note 3} Early Greek Philosophy (4th ed., London, 1948), p. 27Google Scholar.

^{page 51 note 4} Essays and Addresses (London, 1929), pp. 253fGoogle Scholar. (cf. Cornford, 's remarks in Principium Sapienae, Cambridge, 1952, p. 4)Google Scholar.

^{page 51 note 5} The Heroic Age of Science (Baltimore, 1933), pp. 78f.Google Scholar, and later in Hippocratic Medicine, its Spirit and Method (New York, 1941), pp. 96 ffGoogle Scholar.

^{page 51 note 6} In an article entitled ‘Did the Greeks perform experiments?’ in American Journal of Physics,XVII (1949), pp. 384 ff.Google Scholar On the topic as a whole cf. also Edelstein, L., ‘Recent Trends in the Interpretation of Ancient Science’, Journal of the History of Ideas, XIII (1952), pp. 573 ff.CrossRefGoogle Scholar; Farrington, B., ‘The Greeks and the Experimental Method’, Discovery, XVIII(1957), pp. 68 f.Google Scholar, and Zubov, V. P., ‘Beobachtung und Experiment in der antiken Wissenschaft’, das Altertum, V (1959), pp. 223 ffGoogle Scholar.

^{page 52 note 1} Cael. 292a, 7 ff.; Mete. 343b, 9 ff. and 28 ff., and cf. also Cael. 270b, 13 ff.

^{page 52 note 2} Mete. 372a, 28f. (referring to the rare occurrence of moon rainbows).

^{page 53 note 1} Cael. B, ch. 5, 287b, 22 ff. The problem in question is why the heavens revolve in one direction rather than the other, or, as we should put it, why the earth revolves on its axis in one direction rather than the other (so far as I know, the problem has still not been solved).

^{page 53 note 2} Cael. 287b, 28ff. There is a similar disclaimer at Cael. B, ch. 12, 291b, 24 ff., where Aristotle discusses the problem of why some of the heavenly bodies have ‘complex’, others ‘simple’ motions.

^{page 53 note 3} Sambursky, op. cit. pp. 50 ff., has drawn attention to the improvement of astronomical measurements that took place in the Greek period, instancing the increasing accuracy of the approximations which were made of the length of the solar year.

^{page 53 note 4} Aet. III, 3, 1–2; DK, 12A 23, 13A 17.

^{page 53 note 5} Mete. 369a, 29 ff.

^{page 54 note 1} Mete. 342a, 3 ff., cf. 344a, 25 ff. on comets.

^{page 54 note 2} Mete. 369b, 8ff. The expression that sight might be thought rather odd in view of the fact that elsewhere Aristotle flatly denies (against the opinion of Empedocles) that light travels at all (e.g. Sens. 446a, 25 ff., b, 27 ff.).

^{page 54 note 3} Mete. 374b, 3 ff. He refers, also, to the rainbows formed under certain conditions when an oar is raised out of the water (374a, 29 ff.).

^{page 54 note 4} This was first achieved by Newton, in his Opticks (1704)Google Scholar, following on the work of Descartes and Marci.

^{page 55 note 1} This is the conclusion of the most recent editor, Lejeune, A., L'Optique de Claude Ptolémeé (Louvain, 1956), Introduction, pp. 13–26Google Scholar.

^{page 55 note 2} With reference to the diagram where MR is the mirror, A the eye, B the object, B′ the image, O the point at which the visual ray strikes the mirror, and TO and BP perpendiculars to the mirror, these three principles are: (1) B′ lies on AO produced, (2) B′ lies on BP produced, (3) ∠TOA = ∠TOB. (After Cohen, and Drabkin, , A Source Book in Greek Science, New York, 1948, p. 269,Google Scholar n. 1.)

^{page 56 note 1} See Lejeune, op. cit. n. 9 to 225, 9.

^{page 56 note 2} It should be noted that Ptolemy considers the ray from the eye, not that from the object, the incident ray. For his results, which are expressed to within half a degree, compare the tables given by Brunet, and Mieli, , Histoire des sciences: Antiquité (Paris, 1935), pp. 825 ffGoogle Scholar.

^{page 57 note 1} See, for example, Guthrie, , A History of Greek Philosophy, vol. I (Cambridge, 1962), pp. 220 ff.Google Scholar, esp. 223 ff., who gives the Greek sources and comments on the various difficulties which these stories contain. Guthrie concludes that if the discovery of the relations between the intervals is indeed Pythagoras', it was no doubt on the monochord that he carried out his experiments.

^{page 57 note 2} Cf. the judgement of Zubov, op. cit. p. 224: ‘Allein die Tatsache, daß eine solche Legende in der Antike aufkam, zeugt von dem Verständnis des Wesens des Experimentes als einer rationalen Untersuchung der Wirklichkeit.’

^{page 58 note 1} A scholium on Plato, Phaedo 108 D (DK, 18 A 12) contains a report of Aristoxenus which refers to Hippasus' construction, and it seems possible that this may have been designed (as Burnet suggested in a note in his ed. of the Phaedo ad loc.) to provide a model to illustrate the harmony of the spheres.

^{page 58 note 2} Fr. 1. The significance of this fragment for our understanding of the role of experiment in Presocratic, particularly Pythagorean, philosophy, has been discussed by Senn, G., ‘Über Herkunft und Stil der Beschreibungen von Experimenten im Corpus Hippocraticum’, Arch. f. Gesch. d. Medizin, XXII (1929), pp. 271 ffGoogle Scholar.

^{page 58 note 3} On Diseases, IV, ch. 57, L. VII, 612, 6ff. (on which see Senn, op. cit. pp. 248 ff.) and ch. 51, 588, I7ff. Cf. also On the Nature of the Child, ch. 18, 502, 2 f., and ch. 25, 522, 20 ff. (on which see Senn, pp. 245 ff.).

^{page 58 note 4} On Diseases, IV, ch. 39, L. VII, 556, I7 ff., on which Senn commented (op. cit. p. 232): ‘wieder eine anschauliche Beschreibung eines einwandfrei durchgeführten physikalischen Experiments’.

^{page 59 note 1} Senn, op. cit. pp. 242 ff., concludes that the test described was probably in origin a ‘Sedimentierungsversuch’, which the Hippocratic author applied to support his theory of the action of the ‘breath’ causing like substances to come together in the developing embryo.

^{page 59 note 2} Perhaps the best known example in the Hippocratic Corpus is the test described in On Airs Waters, Places, ch. 8, CMG, I, 1, 63, 12ff.Google Scholar, in which a bowl of water is left out of doors to freeze and when the water is thawed it is found on being remeasured to be less than the original quantity (a test which the writer supposes supports his contention that freezing causes the ‘lightest and finest’ part of the water to be dried up and disappear). In Aristotle we find, for example, two references to the fact that if a vessel is heated and then inverted over water, as the air in the vessel cools it contracts and some of the water is drawn up into the vessel, though it may be thought unlikely that Aristotle himself originated this test (Cael. 312b, 13 ff.; GA, 739b,.10 ff., reading with Platt, for ).

^{page 60 note 1} Sext., Adv. math. VII, 117Google Scholar (DK, 68 B 164), cf. also Aet. IV, 19, 13 (DK, A 128). Plato uses a similar model to illustrate the separation of like to like in the Receptacle at Ti. 52E–53 A.

^{page 60 note 2} He also notes, e.g. at Cael. 313 a, 14 ff., that the speed of an object is influenced by its shape.

^{page 60 note 3} E.g. Cael. 273b, 30 ff.; 277b, 3 ff.; Ph. 215a, 25ff. (on natural morion) and Ph. 249b, 27ff. (on forced motion).

^{page 60 note 4} Ph. 230b, 24ff. and Cael. 277a, 27ff. At Cael. 288a, 19ff. it is suggested that in some cases the highest speed is attained not at the point from which flight begins nor where it ends but , but if this means what it appears to mean, namely that the highest speed is attained in the middle of the flight, it is hardly possible to interpret it consistently with Aristotle's known doctrines. See the notes of Stocks in the Oxford Translation and of Guthrie in the Loeb edition of the de Caelo.

^{page 60 note 5} See Cohen and Drabkin, op. cit. p. 203, n. 1.

^{page 61 note 1} Ph. Δ, ch. 8, esp. 215 a, 24ff.; 216a, 11ff.

^{page 61 note 2} A notable exception is the passage at Cael. 311b, 9 ff., where he says that each of the elements, except fire, has weight in its own ‘natural place’ and adduces as proof of this that a bladder weighs more when inflated than when empty. This experiment has been sharply criticized, e.g. by Ross, , Introduction to his ed. of the Physics, pp. 26 ff.Google Scholar, on the grounds (1) that the supposed experimental fact is not correct, and (2) that Aristotle's theory is not true to the experimental facts he thought he had at his command. But whether or not a bladder will weigh more when inflated will depend on, among other things, (1) whether it is inflated with atmospheric air or with breath (which contains a higher proportion of carbon dioxide), (2) whether it is inflated under pressure or not, and (3) the amount of water vapour in the gas with which it is filled. At least three different attempts were made to carry out this test in antiquity (by Aristotle, by Ptolemy and by Simplicius,see Simp, in Cael. 710, 24 ff. and cf. also Anon. Lond. XXXI, 33 ff.), but that three different results were obtained is, perhaps, hardly surprising, considering the lack of precise weighing instruments and the number of factors which might influence the outcome of the test.

^{page 61 note 3} The Greeks had no precise means of measuring time, but no more had Galileo. In the Discorsi he describes the method he employed for measuring time in his dynamical experiments, namely that of weighing the amounts of water which percolated through a thin jet, a simple adaptation of the principle of the water–clock.

^{page 62 note 1} E.g. the account of objects rebounding from a plane, and that of the figures described by certain solids, namely the cylinder and the cone, when these are revolved (Probl. XVI, 4 and 13,Google Scholar 913b, 6ff. and 915 b, 37 ff.). A passage in the Mechanics (858a, 13 ff.) is worth quoting as it illustrates the difficulty which the writer experienced in tackling the problem of ‘why objects which are hurled come to a standstill’. ‘Does it stop when the force which started it fails, or because the object is drawn in a contrary direction, or is it due to its downward tendency, which is stronger than the force which threw it? Or is it absurd to discuss such questions while the principle escapes us ()?’ (trans. E. S. Forster).

^{page 62 note 2} See Simplicus, in Ph. 916, 10 ff. (Wehrli, , Straton von Lampsakos, fr. 73)Google Scholar.

^{page 62 note 3} Philoponus, in Ph. 683, 16 ff.

^{page 63 note 1} See frr. 96 and 98 on the formula or composition of bone, blood and ‘other forms of flesh“.

^{page 63 note 2} Under Aristotle includes the of liquids such as wine and water, for here too an interaction may take place and the resultant compound may be (GC, A, ch. 10, esp. 328a, 26ff. and b, 3 ff.). On Aristotle's theory of different types of mixture and combination, see esp. Joachim, H. H., Journal of Philology, XXIX (1904), pp. 72 ffGoogle Scholar.

^{page 63 note 3} Thus in GC, B, chs. 1 ff. when he sets out to determine the nature and number of the elements, he argues (1) that coming to be and passing away are impossible without perceptible bodies (328b, 32 f.), which in turn cannot exist apart from contrarieties; for a body must be either heavy or light, either hot or cold (329a, 10 ff.), (2) that the principles of perceptible body will be tangible contrarieties (329b, 7 ff.), and (3) that the tangible contrarieties may be reduced to two pairs of opposites, hot and cold, and wet and dry, but these cannot be resolved any further (329b, 24 ff.; 330a, 24ff.).

^{page 63 note 4} For the purposes of this paper I shall include the fourth book of the Meteorologica as this is clearly evidence for the work of Aristotle's immediate school, even if some have denied that it is an authentic text of Aristotle himself (e.g. Hammer–Jensen, , Hermes, L (1915), pp. 113 ff.Google Scholar, and Gottschalk, , CQ, n.s; XI (1961), pp. 67 ff.CrossRefGoogle Scholar, and contrast Düring, Aristotle's Chemical Treatise, Meteorologica, Book IV (Göteborg, 1944), pp. 17 ff.Google Scholar and Lee, Introduction to the Loeb edition, pp. xiii ff. and Preface to the second edition, 1962, p. vii, who have argued that there are no good reasons for not accepting the book as a whole, or in the main, as the work of Aristotle).

^{page 64 note 1} See Mete. 383a, 32 ff. (with Lee's useful note) and cf. 3833, 24f. on the manufacture of pottery and 383b, 7ff. on that of millstones.

^{page 64 note 2} E.g. Diels, H., Hermes, XL (1905), pp. 310 ffGoogle Scholar.

^{page 64 note 3} Cf. HA, 590a, 24ff. Aristotle's statement was apparently accepted on trust by Pliny (XXXI, 37) and by Aelian (IX, 64).

^{page 65 note 1} Flat. ch. 8; CMG, I, 1, 96, 15 ffGoogle Scholar.

^{page 65 note 2} This method of distillation is described in Dioscorides, de Mat. Med. I, 72, 3.

^{page 65 note 3} It may be noted that elsewhere in Mete. Δ different sorts of wine are distinguished according to their reactions to being heated or frozen: new wine, for example, is said to thicken most under the influence of heat and to solidify least under the influence of cold (388a, 33 ff.; cf. also 384a, 4ff. and 387b, 9ff.)

^{page 65 note 4} Mete. Δ does, however, owe a good deal to Plato's Timaeus, particularly to the account of the varieties of water and of earth and of their compounds, 58 D–61 c.

^{page 65 note 5} In general, Aristotle considers most compounds to be composed of water and earth (those that are solidified by cold and melted by fire are said to have a greater proportion of water, those that are solidified by heat to contain more earth), but olive oil and semen, in particular, both present difficulties which he attempts to resolve by suggesting that these are compounds of water and air or (Mete. 383b, 20 ff.; GA, 735 a, 29 ff.).

^{page 66 note 1} The author of On Airs, Waters, Places (ch. 1, CMG, I, 1, 56, 7ff.Google Scholar) for example, notes that waters differ a good deal from one another both in taste () and in weight (), while in On Stones (chs. 22 and 39) Theophrastus refers to differences in ‘density’ () and weight () as methods of distinguishing between different types of ‘stones’ (including, e.g., ores).

^{page 66 note 2} I have dealt solely with the four-element theory. But it may be noted that its main rival in the fourth century, the atomistic theory, was equally comprehensive and vague, and equally incapable of being corroborated or falsified by means of practical tests.

^{page 66 note 3} The history of the use of dissection in antiquity has been described by Edelstein, L., ‘Die Geschichte der Sektion in der Antike’, Quell, u. Stud. z. Gesch. der Naturwissenschaften u. der Medizin, III (1933), pp. 100 ffGoogle Scholar.

^{page 66 note 4} At GA, 765a, 21 ff., Aristotle refers to those who held that the sex of the embryo is determined by whether the seed of the male comes from the right or the left testicle, and who apparently thought that if one of the testicles is tied up or excised, the offspring produced are all of the same sex—a passage which indicates that some earlier writers were aware of the possibility of putting this theory to the test, even if they evidently assumed the results of such a test to be a foregone conclusion.

^{page 67 note 1} This theory is advanced in On Fleshes, ch. 6, L. VIII, 592, 11 ff., for example, where it is supported by various arguments, e.g. that the new–born baby instinctively knows how to suck.

^{page 67 note 2} Ch. 8, L. IX, 86, 4 ff.

^{page 67 note 3} Ch. 10, 86, 13 – 88, 9. At 88, 6, the writer asserts that not even air can be pumped back through the aorta or pulmonary artery into the heart, but he qualifies this in ch. 12, 90, 14 – 92, 1 by suggesting that a little air can and does penetrate into the heart through the pulmonary artery.

^{page 67 note 4} It is notable that in the De motu cordis etsanguinis in animalibus (1628) Harvey refers frequently not only to Galen, but also to Aristotle and to Hippocrates (mentioning On the Heart in particular in ch. 17) and among the passages he cites is a text of Galen in which the latter observed that the three semilunar valves placed at the opening of the aorta prevent the return of the blood into the heart.

^{page 67 note 5} Aristotle, for instance, refers quite often to specimens of different species of animals which were, apparently, deliberately mutilated in order to investigate such questions as whether they can survive without certain organs, whether certain tissues will grow again when excised, or the method of locomotion of different species, e.g. Resp. 471b, 19 ff.; 479a, 3 ff.; HA, 519a, 27 ff.; IA, 708b, 4 ff.

^{page 68 note 1} Thus although the writer attacks those who based their pathological theories on the hot, the cold, the wet and the dry, he himself numbers such things as the salty, the bitter, the sweet, the acid and so on among the constituents of the body (ch. 14, CMG, I, 1, 45, 26ff.Google Scholar), and while he criticizes his opponents for oversimplifying the causal principles of diseases (ch. 1, 36, 2ff.) a similar criticism might also be levelled against his own pathological theories, in which he refers, for example, to the diseases which arise from ‘depletion’ () and from ‘repletion’ () (ch. 10, 42, 11ff.).

^{page 68 note 2} One notable instance where the Greek biologists might have carried out systematic experiments without great difficulty, but failed to do so, is in investigating hybrids (although it should be remarked that for all the interest in heredity in post-Renaissance times the simple experiments of crossing pea-plants which led Mendel to the discovery of the law of the segregation of characters were not performed until the mid-nineteenth century). Equally when the dogma of spontaneous generation was challenged by Redi in the seventeenth century, the experiments he undertook to show that the worms found in decaying meat derive directly from the droppings of flies and not from the putrefaction of the meat itself, were not technically impossible in antiquity.

^{page 69 note 1} Ch. 4, L. VIII, 588, 25 – 590, 4.

^{page 69 note 2} Ch. 7, esp. 50, 9ff. Elsewhere too this writer describes the effects of different drugs, e.g. ch. 6, 44, 11 ff.; 46, 3 ff.

^{page 69 note 3} Several examples from the treatises On Generation, On the Nature of the Child On Diseases IV are discussed by Senn, op. cit. pp. 219ff., and cf. also Regenbogen, O., ‘Eine Forschungsmethode antiker Naturwissenschaft’, Quell, u. Stud. z. Gesch. der Mathematik, Astronomie u. Physik, Abt. B, I (1931), pp. 131 ffGoogle Scholar.

^{page 69 note 4} Ch. 8, L. VIII, 594,9ff. He notes that so long as the blood is hot it does not coagulate, but then he also observes that it does not coagulate if it is beaten. Cf. also ch. 9, 596, 9ff., where he remarks that when the ‘skin’ is removed from blood which is left to clot, another ‘skin’ forms shortly afterwards (an observation wliich he uses to support his account of how the skin of the body itself is formed).

^{page 70 note 1} On Diseases IV, ch. 39, L. VII, 556,15 ff.Google Scholar, suggests that as all the vessels may be filled or emptied by filling or emptying any one of them so the reservoirs () of the humours in the body are filled or emptied by the stomach being filled or emptied.

^{page 70 note 2} On the Nature of the Child, ch. 17, L. VII, 496, 17 ff., argues that growth takes place when the breath, , in the body separates the different substances according to their kinds (the dense, the rare and so on), just as the substances put into the bladder (earth, sand, lead-filings) will be found to be sorted according to their kinds when the bladder is left to dry and opened.

^{page 70 note 3} Ch. 1, L. VIII, 12, 9 ff. It is worth remarking that this test, like that in On Airs, Waters, Places, ch. 8, mentioned above, p. 59, n. 2, involves the use of quantitative measurements.

^{page 71 note 1} I may mention especially the remarkable experiment, described in Anon. Lond. XXXIII, 43 ff., in which Erasistratus showed that there are invisible effluvia from animals by keeping a bird in a vessel without food for a given period of time and weighing the animal together with the visible excreta and comparing this with its original weight.

^{page 71 note 2} PA, A, ch. 5, 644b, 28 ff. Aristotle suggests that our meagre acquaintance with the heavenly bodies gives us greater joy, but that our knowledge of plants and animals is superior in that we can acquire more and better information about them.

^{page 71 note 3} E.g. 529 A, ff. and 531 A, ff.

^{page 72 note 1} History of the Inductive Sciences from the Earliest to the Present Times (London, 1837), vol.3 p. 344Google Scholar.

^{page 72 note 2} I must express my gratitude to Dr M. B. Hesse, who read and criticized an earlier draft of this paper, and to Dr H. B. F. Dixon and Mr F. H. Sandbach for their advice and comments on particular points.