¹ Torstrik and Ross, in their editions (, Torstrik, Aristotelis De Anima, Berlin, 1862;Ross, Aristotle, De Anima, Oxford, 1961), excise 435b24–5, possibly correctly, for it is not relevant to the immediate context and could easily have been added by an editor who had read 416b16–22. There would have been reason for an editor to add something at this point if the papyrus MS he was working from was damaged (see n. 2, and below, pp. 379–81Google Scholar

² There is a complication at this point (435b22–4) in that Aristotle says that taste, too, is for the sake of well-being. But it is not a distance sense, and at 434b22–6 (which is referred to at 435b20) Aristotle has said that ‘;these senses[sc. touch and taste] are necessary for animals and it is clear that no animal can live without touching [at b21 taste is said to be a kind of touch] but the other senses are for the sake of well-being and need to belong not to just any kind of animal, but only to some, e.g. to those which can move.’ Again, the ‘other senses’, which at 435b19–21 are said to be for the sake of well-being, are marked off in the preceding discussion in distinction to touch because they do not by themselves involve contact with the object (435b7–13). So in III 12–13, taste is said to be necessary for an animal, and not one of the senses which are for the sake of well-being. Again, in De Sensu 436bl4–21, taste is said to be necessary for animals and is distinguished from smelling, hearing and seeing in that the latter belong only to animals which can move. Only at III 13 (435b22–4) is it said, against the view clearly expressed in four other places, that taste is a sense which exists for well-being. It is also strange that smell is not mentioned at 435b22, for smell is earlier said to be a distance sense which exists for the sake of well-being. I suggest that we should read for at 435b22; then both problems disappear. (According to the account which I give below, these lines were arguably at the very outside of a roll which was damaged by moisture and moths, and in which Apellicon is said to have ‘filled in the gaps incorrectly’, Strabo, Geography 609. See below, pp. 379–81.) But this crude solution to the issue does not affect the larger issue of whether III 12–13 is the background for II 8.

³ I do not mean to suggest that there were chapters already marked out in Aristotle's text which then got shuffled; the practice of dividing the De Anima into the present chapters goes back no further than the third edition of Erasmus' complete Aristotle (Basle, 1550).

⁴ Kenyon, J. G., Books and Readers in Ancient Greece and Rome² (Oxford, 1951), 111–12.Google Scholar

⁵ § 13 of the catalogue in Diogenes Laertius (5.22) and § 13 of the catalogue in the Vita Menagiana refer to which cannot refer to our De Anima; it is part of a series of published works, not one of his scientific works, and what Plutarch refers to as ‘Eudemus, or De Anima’ (Consolatio ad Apolloniam, 115B), which is a dialogue. Again, § 73 of the catalogue in Diogenes Laertius (5.24) refers to , which is also unlikely to be our De Anima, for that work does not consist of , unless possibly it is a separate edition of Book I of our De Anima, which does consist of doxographical reports and criticisms.

⁶ Andronicus knew the De Anima, for he rejected the De Interpretatione as spurious, on the grounds that it did not square with his reading of the De Anima; see Ammonius, In De Int. 5.27–6.4 Busse. The report is repeated by Philoponus, In De An. 27.21–7 Hayduck, and Alexander had earlier said that Andronicus rejected the De Interpretatione (In Anal. Prior. 160.31–161.1 Wallies). For the view that nobody between Theophrastus and Andronicus knew the De Anima we have only the argument from silence.

⁷ Since the De Anima was not prepared for publication by the earlier Peripatetics, it would not have been assembled into a single-roll edition, and it would have remained on the separate stationer's rolls on which it was written. These were never more than twenty sheets glued together, according to Pliny, HN 13.23 (confirmed byLewis, N., Papyrus in Classical Antiquity [Oxford, 1954], 54–5). At a maximum width of 7½12 inches per sheet, a roll would amount to 12½12 feet, and it is impossible to fit the 96,000 characters of the De Anima onto a roll of 12½ feet, even assuming (unrealistically) the largest number of characters per line (25), the largest number of lines per column (45), and the largest number of columns per foot (6). On the other hand, if it were written on two rolls, it would have been written at an ordinary density (all figures taken from Kenyon, op. cit. (n. 4), 51–60). The reason for the De Anima now being in three books is presumably that when it was published it was given large writing and large margins for a scholarly marketGoogle Scholar

⁸ Kenyon, op. cit. (n. 4), 61. Although this is not always true (I understand that it was not the case with the Herculaneum papyri), there is evidence contemporary with Aristotle in a marble statue of Sophocles. This marble is evidently a copy of the bronze which Lycurgus arranged to have set up in the theatre at Athens c. 336–324 B.C. (Ps.-Plutarch, Lives of the Ten Orators, 841F) and which was still standing in Pausanias' day (Description of Greece 1.21.1). At his right foot there is a capsa, or scroll-bucket, containing eight papyrus rolls, of which one is stored half-way open, one is stored with the beginning on the outside (or is upside-down), and six are stored with the end on the outside. The copy is now in the Vatican (Museo Gregoriano Profano, § 9973 = § 237 of the former Museo Profano Lateranense).

⁹ Here is the calculation (all figures taken from Kenyon, op. cit. (n. 4), 51–60): roll A has approximately 46,465 characters; it, being most of a stationer's roll, was between 10 and 12½12 feet in length; so it was written on at a density of 300–60 characters per inch of roll. III 12–13 has approximately 4190 characters, which therefore occupied 10–14 inches of roll, and since there were 2–3 inches per column, 4–7 columns. But since of modest quality would be from 5 to 7½12 inches, the text might have covered two which became detached because their glue perished in the moisture to which Strabo refers. So we can picture the detached scrap(s) as being, very roughly, five or six columns on two . (These figures are confirmed by working from the other direction; there were approximately 18–25 characters per line, so III 12–13 would have occupied 167–233 lines, and 25–45 lines per column, i.e. somewhere between 4 and 9 columns and between 8 and 27 inches.)