¹ ‘The streak colour of a mineral is the colour of its powder obtained by scraping or rubbing the specimen on a streak plate made of unglazed china, or other material harder than the mineral to be tested. The streak of a mineral may be quite different from its colour’, Rosenfeld 1965 p. 38. This is a technical definition of something familiar: graphite is a soft mineral with a grey streak colour - so soft that it leaves a streak on paper. Hence graphite is used to make pencils.

² There is a lacuna in the text, long enough for a word of about 6 or 7 letters, which was probably the name of the stone. There is insufficient detail here for a positive identification of this stone, but Augite would fit (a common pyroxene which is black to a dull dark green and has a white to murky green streak colour).

³ See n. 1. Rosenfeld’s general discussion of colour in rocks makes an interesting comparison with [Aristotle]’s quoted in the text: ‘When the colour of a mineral depends on its chemical composition it is said to be inherent. Many minerals have no inherent colour in the pure state, they are colourless or white. Such minerals may be coloured by minute inclusions of an impurity in chemical combination, or disseminated as small particles in the mineral. This type of colouration is called exotic. The exotic colours of minerals vary quite widely according to the type and amount of colouring material. The opaque white colour of many colourless minerals such as quartz, gypsum, and talc is usually due to reflection of light from countless minute cracks or bubbles of fluid in the mineral, and rarely to the inclusion of white impurity, such as kaolin in some feldspars’, p. 37 f.

⁴ Belopoiika 71. He closes the discussion shortly after this ‘in case we inadvertently digress too far and enter deeper into physical arguments’ §72.3-4.

⁵ For this subject and its interconnections with much else in ancient physical thinking, see Berryman 1998. Smith 1996 has recently translated and commented upon Ptolemy’s Optics.

⁶ Note the difference with modern Physics, which is defined in the OED as being ‘the science, or sciences, treating of the properties of matter and energy, or of the action of the different forms of energy and matter in general (excluding Chemistry and Biology)’. Since the discovery in the last decade of life-forms on earth, in deep caves and in the abyss of the ocean, which are based on direct chemical synthesis (not photosynthesis), the nature of life itself is open to question, and many of the relevant definitions will need reconsideration.

⁷ Thorndike’s volumes stand as a massive and valuable exception: volume 1 of a History of Magic and Experimental Science 1923 starts with Pliny, and fills nearly 300 pages with the Roman period alone.

⁸ Düring tried to overcome this problem by entitling his commentary on Meteorology 4 as Aristotle’s Chemical Treatise 1944. But like the ancients, most moderns discussing this subject tackle it in works with Physics, rather than Chemistry, in the title. Ancient chemical ideas may also be discussed in works with alchemy in the title, e.g. Keyser 1990a.

⁹ ‘Distilling, sublimation and the Four Elements: the aims and achievements of the earliest Greek chemists’ in Tuplin and Fox 1999, to appear. My thanks to the editors for allowing pre- publication access to the MS.

¹⁰ Hero’s Automatopoietikes (Automaton-making) has recently been translated into English by Murphy 1995. It gives a wonderful impression of the backstage activities which went on during the performance of dramas.

¹¹ See de Solla Price 1974.

¹² See Lewis 1992.

¹³ The role of applied geometry in such things is noted by Proclus, Comm. on Euclid’s Elements Book 1, Prologue Part 2, 63-4. His late antique [C5 A.D.] view on pure geometry is well summarized in §§49-56.

¹⁴ For many reasons, the most important of which is the general methodological principle that it is revolution, not the absence of revolution, which needs explaining. For a discussion of related points see White 1993 and Greene 1994.

¹⁵ See Lennox’s remarks to the same effect with respect to Aristotle’s zoological works, 1994. Occasional and brief periods of financial support, such as in early Ptolemaic Alexandria, are the exceptions which prove the rule, and were appreciated as such at the time. See, for example, Philo’s comment upon the great leap forward in mechanical technology, in Belopoiika 50.24-6: ‘Alexandrian craftsmen achieved this first, being heavily subsidized because they had ambitious kings who fostered craftsmanship’, Marsden trans, pp. 108-9.

¹⁶ On Diesel see Nitske and Wilson 1965. There is perhaps an analogous situation in the modern scholarly habit of tracking back a scientific idea to the earliest source one can find, crediting that person with the ‘discovery’ or ‘creation’ of the idea (however hazily it might then have been expressed), and playing down its development and transformation in later authors. The tendency to create heroes (see chapter 1 §5) plays a part here too.

¹⁷ These are distinguished from terrestrial earth, air, fire and water in being utterly pure, whereas terrestrial earth, air, fire and water are always subject to mixture with other elements.

¹⁸ In the sense of juxtaposition rather than chemical synthesis.

¹⁹ Reported in the preface to Hero’s Pneumatics. Gottschalk 1965 concludes that this preface is ‘a jumbled but otherwise faithful version of an extract from a book by Strato, almost certainly the περì τοΰ κε;νοΰ [On the Void] included in Diogenes’ catalogue of his works (5.59). Strato’s most important proofs of the existence of discrete void are given substantially as he wrote them; of the other material in Strato’s book, which was of less interest to the engineers, something survives in the digression on pp. 110.3-112.6, but this part has evidently suffered much more severely from condensation and rearrangement. No doubt a good deal has also been omitted . . . but except for the paragraph pp. 114.14-29 nothing of importance has been added from any other source’.

²⁰ That part of his book concerned with private building was epitomized in the late empire by Faventinus (c. A.D. 300) and then Palladius. For Faventinus’ text with commentary see Plommer 1973, who thinks that they are better than the average epitomes produced in this period (pp. 2-3).

²¹ Which was the general practice for all but the simplest and cheapest items, such as domestic pottery, or the most expensive imported regional specialities, such as cloth made of a particular fabric or colour.

²² The translation of the postulate is from his little book Archimedes, 1920, which reads more easily than his translation of the same in The Works of Archimedes, 1912. The rest of the translation is from the latter work.

²³ After the proof of this proposition follows the second and last postulate for the entire work; as Thomas says of this in the Loeb (GMW 2 p. 245 n.a.) ‘if the object of mathematics be to base the conclusions on the fewest and most self-evident axioms, Archimedes’ treatise On Floating Bodies must indeed be ranked highly’. There follow two more propositions in book 1, then book 2, which is described by Heath as a geometrical tour de force.

²⁴ Such formulae are given whenever this subject is discussed, e.g. in Thomas GMW2 pp. 38-9 and 250-1, Heath 1912 pp. 260-1; given the constraints of space here, it seems superfluous to repeat them.

²⁵ And removal of the mass from the water without displacing more fluid in the process would demand further equipment and considerable care of execution.

²⁶ Lloyd has an extended discussion of Aristotle’s views on pepsis (concoction) with reference to the biological works in 1996a chapter 4.

²⁷ Galileo knew and cited Philoponos’ works and, like many of his contemporaries, was clearly influenced by him. See Wolff 1987.

²⁸ On which see Marsden 1971.

²⁹ See e.g. Philo Belopoiika §50.26-9: ‘The fact that everything cannot be accomplished by the theoretical methods of mechanics, but that much is to be found by experiment, is proved especially by what I am going to say’ (Marsden trans.).

³⁰ Known in antiquity as the Ktesibian pump, and the sort of pump used in Roman fire-engines. See Oleson 1984 for full discussion of this and other pumps.

³¹ Drachmann 1948 focuses on their work on pneumatics. Also useful is Drachmann 1963, which contains a translation from the Arabic of much of Hero’s Mechanics with relevant passages from other authors. On Archimedes’ machines see Sleeswyk 1990 and Simms 1995.

³² See e.g. Green 1986, who uses this phrase of Ktesibios’ water-organ.

³³ Aetna 294-9, and 328 respectively. In a completely different time and context the Christian author Tertullian referred to God playing the water-organ, De baptismo 8.

³⁴ Even the real Aristotle occasionally ventured out of his depth, for example on astronomy. Lloyd takes him to task for this (1996a chapter 8), perhaps a little unjustly; Aristotle was no mathematician. See Heath 1949.

³⁵ The 15,000 pages of Greek texts by the late (c. A.D. 200-600) commentators on Aristotle are the subject of a large programme of English translations (over 60 volumes planned) under the general editorship of R. Sorabji, published by Duckworth.

³⁶ On implicit translation see Chapter 3.1 below.