¹ The source for Thrasyllus and his arrangement is Diogenes Laertius iii 56–61.

² By Sparshott, F. E. in his splendidly knowledgeable ‘Introduction’ to Mill's dealings with classical topics written for Mill's Essays on Philosophy and the Classics, ed. Robson, J. M., Toronto and London, 1978Google Scholar, Collected Works of John Stuart Mill, xi. xix, n. 54; by the editorial footnote to the passage quoted above from Autobiography and Literary Essays, ed. Robson, John M. and Stillinger, Jack, Toronto and London, 1981Google Scholar, CW, i. 9; and in Appendix B to CW, i, ‘Mill's Early Reading, 1809–22’, 553 f. But see n. 4 below.

³ CW, i. 553.

⁴ This is indeed what Prof. Sparshott did. He writes (private communication) that, rather than rely on Burnet's OCT, he looked among the Plato editions he had to hand for the one closest in date to the writing of the first draft of the Autobiography (1853). This was Stallbaum's single volume Tauchnitz of 1850 (see Appendix 1 below). Although it does not follow the Thrasyllan arrangement, the table of comparative pagination at the end gives pride of place to the Aldine, which does follow the Thrasyllan order. The Thrasyllan order is in any case the only arrangement to survive into the modern world with a semblance of ancient authority.

⁵ ‘In my eighth year’: CW, i. 9, 12.

⁶ The libraries where I did find a few extras were Columbia University (the richest of all in editions of Plato), Harvard, Princeton, Pittsburgh, and Trinity College Dublin. I also checked the published catalogues of University College London, of the old Scottish universities (Aberdeen, Edinburgh, Glasgow, St Andrews), of the French Bibliotheque Nationale, and the Union Catalogue of libraries in the USA. It soon dawned on me that bibliography is not an exact science. Dates of publication for what looks like the same edition sometimes seem to vary from library to library; an editor like Stallbaum, who edited and re-edited Plato over many years, can be hard to keep track of. That said, my two Appendices are as accurate as I can make them.

⁷ Thus Evans, Frank B. III, ‘Platonic Scholarship in Eighteenth-Century England’, Modern Philology, xli (1943–4), 103–10CrossRefGoogle Scholar, researched the most pertinent period from partially different sources and identified nothing that my methods missed. For Ficinus's translation up to 16001 have relied on Hankins, James, Plato in the Italian Renaissance, 2 vols., Leiden, New York, Copenhagen and Cologne, 1990, ii. 738–85Google Scholar. He lists all European printed editions with a Latin translation up to 1600 and beyond, so he omits Greek only editions and translations into modern languages; nor does he report on my central concern, the order of dialogues. Nonetheless, he gave me useful information as well as a check on some of my findings.

⁸ Hermann's Preface does not explain why he chose to resurrect the Thrasyllan order. In his Geschichte und System derplatonischen Philosophie, Heidelberg, 1839, p. 358Google Scholar, he remarks that no ancient order has more justification than another, and that it may be merely chance that Thrasyllus influenced most of the MSS and early editions (cf. n. 20 below).

⁹ CW, i. 25; cf. 568.

¹⁰ Cousin's French translation of 1822–40 (referred to by Mill, J. S. at CW, xi. 42, n.Google Scholar, as so much better than Thomas Taylor's English translation) meets the requirement, with Theaetetus in fifth place; but it is French only, not Greek, and it postdates the year 1813.

¹¹ As Serranus explains in the introduction he wrote for the ‘verae solidaeque philosophiae studioso lectori’. Then, after some intervening matter, he proudly displays his table of contents under the heading ‘Catalogus dialogorum Platonis, iuxta novae huius distributionis seriem’. ‘Novae’ contrasts with the then prevailing Thrasyllan arrangement, and perhaps also with Ficinus's order; for discussion of whether there might be more to Ficinus's arrangement than the order in which he did the translating, see Hankins, i. 308–11.

¹² A new printing of the Lyons-Geneva edition of 1590, but with the Serranus-Stephanus order instead of Ficinus's. Yale University has a copy once owned by Bishop Berkeley: Hankins, ii. 786.

¹³ See Appendices, where the presence of Stephanus numbers is marked by ‘S’. The Bipont included the Stephanus letters (marking sections a to e) as well as the page numbers, but this took longer to become standard practice (cf. Stallbaum's 1850 edition, pp. iv–v). Hermann, p. 562, n. 2, is clear that it was not intrinsic merit, but the authority of Stephanus (plus the support of the Bipont) that promoted Serranus's arrangement, which now ‘allgemein bekannt geworden ist’; this last is practically German for ‘It has become the common arrangement’.

¹⁴ For the evidence, see Notopoulos, James A., The Platonism of Shelley: A Study of Platonism and the Poetic Mind, Durham, N. Carolina, 1949, pp. 41 f., n. 52Google Scholar.

¹⁵ Readers can verify this, without making a special trip to the library, by looking at the Thrasyllus-Stephanus concordance printed at the beginning of each volume of Burnet's OCT.

¹⁶ Both now generally considered inauthentic, Theages with more reason than Erastai: see Taylor, A. E., Plato: The Man and his Work, London 1926, pp. 529–34Google Scholar. They are first marked ‘Incerti Auctoris’ in Stallbaum's edition of 1821–5. An up-to-date English translation may now be found in Cooper, John M. ed., Plato: Complete Works, Indianapolis, 1997Google Scholar.

¹⁷ The Earlier Letters of John Stuart Mill (1812–1848), ed. Mineka, Francis E., Toronto and London 1963, CW, xii. 7Google Scholar; Mill's italics.

¹⁸ Hermann, , p. 359, notes that Fischer has gone back to ‘die alte Tetralogien’Google Scholar.

¹⁹ And in the appendix to his first volume, ‘Defensio locorum quorundam platonicorum ab emendandi libidine Henr. Stephani aliorumque’ (pp. 591–639).

²⁰ If so, he was not far wrong. According to Alline, Henri, Histoire du texte de Platon, Paris, 1915, pp. 124, 176–8Google Scholar, the archetype of our medieval MSS did follow a tetralogical order, though with some variation from Thrasyllus.

²¹ Here I transcribe, with gratitude, a private communication from Christopher Stray: ‘It depends on the editions, the frontiers and the period! Approach to Latin (USA, early 1930s) was adapted for the UK and was a market leader in the early 1930s and 1940s here. But that was very unusual indeed. Most schoolbooks are nationally specific; and some in the earlier 19th century were school-specific (a matter of pride) [for an example, see n. 77 below – MFB]. But in 1813, the continental blockade presented an additional barricade to imports; especially for books in the humanities, more likely to offend officialdom as prone to carry radical ideas. Booksellers' lists of 1816 were full of the German classical editions of the previous decade, suddenly available.’ Again, ‘Kennedy's, B. H.Revised Latin Primer (1888)Google Scholar never had a US edition because he refused to allow changes to be made in syntactical terminology to conform to US practice. Other-wise his new order of cases might have spread to American schools’.

²² The Works of Plato, 5 vols., London, 1804, i. cviiiGoogle Scholar. For other editions he mentions, see n. 48 below.

²³ Moreover, the volumes it has were clearly not bought at the time of publication, for they are all bound together in two sets of covers (one with a non-Cambridge coat of arms), each combining publications from different years; they could have been accessed any time prior to 1978.

²⁴ Catalogus librorum in bibliotheca universitatis Andreanae secundum literarum ordinem dispositus, Andreapoli, 1826Google Scholar; Catalogue of the Printed Books in the Library of the University of Edinburgh, 3 vols., Edinburgh, 1918–1923Google Scholar; Catalogue of Books in the General Library and in the South Library at University College London, London, 1879Google Scholar; Catalogus impressorum librorum in bibliotheca universitatis glasguensis secundum literarum ordinem dispositus … labore et studio Archibaldi Arthur, Glasgow, 1791Google Scholar; Catalogue of the General Library of the University of Aberdeen, Aberdeen, 1874Google Scholar.

²⁵ No record of Fischer either in Librorum impressorum qui in Museo Britannico adservantur catalogus, 2 vols., London, 1787Google Scholar, or in Librorum impressorum qui in Museo Britannico adservantur, 7 vols., London, 1813–1819Google Scholar; the entry for Plato is in vol. v., dated 1817. Both the third volume of Fischer's series, and the 1783 re-edited version of the first, came with the King's Library in 1823 (see below). The 1813–19 catalogue shows a number of additions to the Plato collection, including the Dublin collection of 1702, Forster (1745), Routh (1784) – all introduced below – and volumes 2 and 4 of Heindorf s series (listed in Appendix B); the absence of Fischer is not due to sluggish buying.

²⁶ For this information I am indebted to Ian McAuslan, Janet Taylor, James Lawson, Jonathan Katz, and Stephen Anderson respectively.

²⁷ Outside Britain, I found Fischer's first volume at Princeton and his second at Harvard; the National Union Catalogue lists other US libraries where these are found, but none with volumes 3, 4 or 5. None at all at Columbia or Trinity College Dublin. But the Bibliotheque Nationale in Paris has them all.

²⁸ Grote, George, Plato and the Other Companions of Sokrates, 3 vols., 3rd edn., London, 1875, i. chs. 4–5Google Scholar. The sole occurrence of Fischer's name is in a note on p. 132, where Grote quotes Wyttenbach, writing in 1776, to illustrate the general lack of interest in and study of Plato before the 19th century. The quotation is cited as coming from ‘Wyttenbach, Bibliotheca Critica, vol. i, p. 28. Review of Fischer's edition of Plato's Philêbus and Symposion’. Grote will also have seen Fischer referred to in Hermann's book (n. 18 above).

²⁹ Thus Euthyphro, Apology, Crito are in ii. (with Parmenides between the first two!), Theaetetus in iii., Phaedo in v., Theages and Erastai in vi. There is, however, a BekkerStephanus concordance at i. ix–x.

³⁰ See CW, xi. 37–238; the virtual certainty is due to Sparshott's ‘Introduction’ (n. 2 above), xvii–xx.

³¹ See the full title in Appendix 1. Who made the selection? The Preface ‘Lectori’ speaks of Bekker in the third person, but it is unsigned, and I can find no hint anywhere else in the work. However, the British Library Catalogue seems to know: ‘Edited by G.Burges’. Burges contributed to the Bohn's Classical Library English translation of 1848–54 (see Appendix 1).

³² So says the label inside the front cover of vol. i of Grote's personal copy of Bekker, now in the Bodleian Library at Oxford.

³³ CW, xi. 39. The bookseller, as well as the publisher and printer, was A. J. Valpy, also responsible for the Delphin Classics and The Classical Journal. The frontispiece to the Bekker edition says ‘Sumptibus Priestley’, which indicates that Priestley put up the money, which Valpy was presumably unable to repay. Eleven years later, in 1837, Valpy sold his firm and stock to Longmans: more on that below, n. 80.

³⁴ On the end pages of each Bekker volume Mill lists (with occasional comments of his own) the pages where he has underlined a passage that struck him as worthy of note. A likely occasion for this rather systematic annotation is when he re-read the whole of Plato as preparation for his 1866 review of Grote's Plato (cf. CW, xi. xxxviii). But why, in that case, do the pages of the Phaedo in vol. v. of his Bekker remain uncut? (No puzzle about the uncut state of the Spuria in ix., and the Phaedo was not among his 1834–5 translations.) On 10 November 1865 he writes to Grote that he is ‘reading once more quite through some of the most important of the dialogues I read last spring: Phaedon, Parmenides, Theaetetus, Sophistes, Politikos, etc’ (The Later Letters 1849–1873, ed. Mineka, Francis E. and Lindley, Dwight N., Toronto 1972, CW, xvi. 1116)Google Scholar. So he read the Phaedo, but not in Bekker. I guess that he took advantage of the fuller and more upto-date notes in Geddes' recently published commentary: The Phaedo of Plato, ed. with introd. and notes by Geddes, W. D., London and Edinburgh, 1863Google Scholar.

³⁵ For these details, more precise than those given by Bain, Alexander, James Mill: A Biography, London, 1882, pp. 18 f.CrossRefGoogle Scholar, I am indebted to David Robinson, who kindly looked at the entries for me and sent me a copy of the relevant page of the Catalogue of the Theological Library in the University of Edinburgh, Edinburgh, 1829Google Scholar.

³⁶ Bain's phrase, p. 19.

³⁷ The information about watermarks comes from Fenn, Robert Anthony, James Mill's Political Thought, diss., 2 vols., London, 1972, ii. 335, 368, 379Google Scholar. A revised version of this thesis was published (its second, bibliographical volume replaced by a more concise Appendix) as James Mill's Political Thought, New York and London, 1987Google Scholar, the Preface to which announces an edition of the Common Place Books to be published in 1987. It seems that the author's death intervened and the edition has not appeared. I am grateful to the London Library for allowing me to consult the originals.

³⁸ Thus Penn is premature when he speaks (in the 1987 book, p. 5, n. 15) of the Bipont as ‘the edition Mill used’. A one-volume Platonis opera was formerly among the books from John Stuart Mill's library presented to Somerville College by Helen Taylor: it is recorded on the removal firm's list of the books they delivered to the College on 6 September 1905, though not on a librarian's list dated 1944. I imagine this was his father's 1602 Frankfurt edition. The alternative is to suppose that John Stuart acquired the 1839 edition of Baiter, Orelli and Winckelmann or Stallbaum's 1850 Tauchnitz. Neither is as useful as the Bekker he already owned, but there are occasions when it is convenient to have all Plato in a single large volume.

³⁹ Taylor, i. ciii.

⁴⁰ Bain, p. 464.

⁴¹ ‘Taylor's Translation of Plato’, Literary Journal, iii (1804), 449–61Google Scholar and 577–89; ‘Taylor's Plato’, Edinburgh Review, xiv (1809), 187–211Google Scholar. [Both are to be reprinted in Apeiron, xxxiv (2001), with an introduction by myself.] The earlier review is signed ‘M.’; the second is anonymous. Since Mill was the editor of the Literary Journal throughout its existence and a regular contributor, and since the two reviews are plainly from the same author, it is safe to ascribe both to his pen. The second review was first credited to him in an elegant note by Levinson, Ronald B., ‘Concerning James Mill’, Modern Language Notes, Xl (1925, 379 fCrossRefGoogle Scholar. Fenn's survey of Mill's many anonymous writings (n. 37 above) confirmed his authorship of both; in the same year the same result was reached, through a somewhat different approach, by Lazenby, A. L., James Mill: The Formation of a Scottish Emigre, diss., University of Sussex, 1972Google Scholar. More recently, John Glucker has detected a minute misquotation from a key passage of Cicero in both reviews, which shows that when writing the second review Mill had the first in front of him and copied the Cicero from it: Glucker, John, ‘The Two Plato's of Victorian Britain’, Polyhistor: Studies in the History and Historiography of Ancient Philosophy, ed. Algra, K. A., van der Horst, P. W. and Runia, D. T., Leiden, 1996, 396–400 with n. 37Google Scholar.

⁴² ‘Taylor's Plato’, 207.

⁴³ CW, i. 25.

⁴⁴ The introduction to his Plato translations (n. 30 above) contains several echoes of his father's first review of Taylor. For an extra twist, consider the possibility that by 1853 Grote might have made Mill aware of Hermann's return to Thrasyllus in the Teubner. Then a phrase like ‘the common arrangement’ would also serve to exclude the Thrasyllan order for the benefit of Mill's more learned readers.

⁴⁵ Borrowing from the Main Library had been permitted since 1737, on payment (if you were a student) of a deposit of two shillings and six pence: Cuthbertson, David, The Edinburgh University Library: an account of its origin with a description of its rarer books and manuscripts, Edinburgh, 1910, p. 11Google Scholar.

⁴⁶ For a helpful introduction to this bewildering practice, see Ingram, William H., ‘The Ligatures of Early Printed Greek’, Greek, Roman and Byzantine Studies, vii (1966), 371–89Google Scholar. For illustration, see Appendix 3 below.

⁴⁷ As he explains in his own preface ‘lectori φιλοπλτωνι’, printed after Serranus's address to Queen Elizabeth ‘of England, France, and Ireland, etc’ and before Serranus's Introduction.

⁴⁸ Forster's selection contained five dialogues because he conceived Erastai as a prologue to the first tetralogy. The misspelling of his name could be a slip of memory by Mill rather than a misprint. Routh's work is again singled out for praise by Mill, J. S. at CW, xi. 39Google Scholar. Did James Mill come to own a copy, or was the son (in this as in other points) echoing the father's review? James Mill was not reading the Apology in Forster's ‘very good stile’ at the time he wrote the excerpt in CPB I. Both Forster and Routh are included in Thomas Taylor's acknowledgements (loc. cit., n. 22 above), along with Fischer's second volume, Massey's Republic, and some other items.

⁴⁹ The review appeared in no. 8 of the Literary Journal, dated 1 May 1804, and no. 10, dated 1 June 1804; 1804 is the year Taylor's translation was published.

⁵⁰ Fenn, diss., ii. 384, concludes from various indications that the material in CPB V ranges from 1800 to the last years of James Mill's life.

⁵¹ So Hankins, p. 786. He does not explain how he knows, but I take his word for it. Apart from the ordering of dialogues (n. 12 above), the sole difference between the 1590 Lyons-Geneva edition and the 1602 Frankfurt is that the former has an extra page (not recorded by Hankins): after the title page, identical in both, comes one headed ‘Typographus candido lectori salutem’, in which the writer, without revealing his name, enthuses about the greatness of Plato. Stephanus was certainly a typographus as well as a scholar.

⁵² Independently evidenced by his son's report (Autobiography, CW, i. 125) that James Mill was a great collector of ‘manuals on the school logic’.

⁵³ The ‘common readings’ at issue here come from the MSS now designated BT: ὃ μή. As Mill points out, this makes philosophical nonsense: ‘“scribendo ἔστιν ὃ τι,” says he [sc. Stephanus], “vel ἔστιν ἥ”’. In today's OCT Burnet prints ἔστιν ὃπη (‘corr. Coisl.’), which is a trivial variant on Stephanus' second alternative. The Bipont prints μή, yet the ‘Variae lectiones’ section at the end (iii. 358) quotes the relevant part of Stephanus's note. In the 1826 London edition of Bekker, Stephanus's note is transcribed more fully, and supplemented with a note by Heindorf which attributes πη to Schleiermacher. In Schleiermacher's first edition, i. 397, this is an emendation ‘eben so leicht als unentbehrlich’; the second edition, i. 408, is able to assign the correction to two of the MSS collated by Bekker.

⁵⁴ In 1818 James Mill thought about applying for the Chair of Greek at Glasgow: Bain, pp. 166–8. In the end, he preferred the India Office.

⁵⁵ Mill, J. S. remarks in his Autobiography, ‘in those days [1828] there was no public or subscription library from which works of reference could be taken home’ (CW, i. 134)Google Scholar; Mill, originally wrote ‘there was no institution like the London Library’ – that opened in 1841)Google Scholar.

⁵⁶ Miller, Edward, That Noble Cabinet: A History of the British Museum, London, 1973, 61–6Google Scholar.

⁵⁷ My evidence is the two catalogues cited n. 25 above.

⁵⁸ Bibliotheca Smithiana, pars altera: a catalogue of the remaining part of the curious and valuable library of Smith, Joseph esq., His Majesty's Consul at Venice, lately deceased, and of many other collections lately purchased, London, 1773Google Scholar. Bembo is item 2152: ‘Opere di Platone de Bembo, 3 torn, gilt, £1. Is.’ The bookseller was James Robson.

⁵⁹ Barnard, Frederick Augusta, Bibliothecae Regiae Catalogus, 10 vols., London, 1820Google Scholar.

⁶⁰ Paine, Elaine M., The King's Library, British Library publication, London, 1989, pp. 6 fGoogle Scholar.

⁶¹ Paine, p. 12.

⁶² Of the other translations mentioned by Mill, the King's Library had Ficinus and Serranus, but not Grou (1770). Nor did the British Museum Library yet possess Grou (n. 57 above), as it does today. Perhaps Mill owned or borrowed a copy.

⁶³ The Greek quotations in the review match the Frankfurt text, but diverge occasionally in their punctuation. Given Mill's remark (above, p. 14) about ‘pointing’, I would attribute the divergences to his editorial intervention: e.g., it is unnecessary, almost amateurish, to place commas both before and after προσανατριψμενος at Theaet. 169c2 (p. 586).

⁶⁴ Mill, James, Essay on Education, in James Mill: Political Writings, ed. Ball, Terence, Cambridge, 1992, p. 141Google Scholar, with the word ‘not’ omitted.

⁶⁵ From the various lists cited n. 68 below.

⁶⁶ Contracted ον is a feature of Mill's own handwritten Greek in the commonplace books.

⁶⁷ So says Appendix B (except that ‘Planudes’ is misprinted ‘Pladunes’), but it is unclear on what basis, for the book is bound without its first twenty pages (including the title page). For what follows I have checked, and where necessary corrected, the Appendix's description of the editions it cites.

⁶⁸ Appendix B notes that (a) is on a librarian's list of SC books made in the 1930s. (It has proved impossible to track down, even in Glasgow University Library. But no Foulis edition listed below merits a classification worse than ‘light’, and already their Xenophontis de Socrate commentarii; item Socratis Apologia of 1761 is ‘light’.) I found (b) on a similar list dated 1944 (n. 38 above), headed ‘List of books from the library of John Stuart Mill presented to Somerville College by Miss Helen Taylor, 1905’. The removal firm's inventory (also n. 38 above) mentions a 10-volume Xenophon, which is probably (b) with a couple of volumes missing. There is no trace today of the 1930s list cited in Appendix B. But this is the place for me to thank the Librarian of Somerville College and her staff for their patience and help during my many visits to view the collection.

⁶⁹ Appendix B identifies the edition (not even formerly SC) from Mill's, (only sometimes accurate) page references in ‘The History of Rome, by John Stuart Mill, aged 6 and a half’, printed in CW, i, Appendix A ‘Juvenilia’, 541–8Google Scholar.

⁷⁰ Who became a favourite with him later in a different edition: see Appendix B, no. 160.

⁷¹ References at Appendix B, nos. 7, 8, 25. Mill read Dionysius of Halicarnassus when he was 6 and a half, before Plato; Lucian and Isocrates before May 1813, hence possibly (but not certainly) before Plato. See CW, i. 542, 553–5.

⁷² References at Appendix B, no. 10.

⁷³ CW, i. 9.

⁷⁴ For doubts about its identity, see n. 67 above.

⁷⁵ Unfortunately, it does not explain the reason for the annotation. The sentence underlined gives Aesop's clever interpretation of a grave inscription consisting of nothing but the letters ΑΒΔΟΕΘΧ, namely, ‘Give back to king Dionysus the treasure you found here’. Nothing relevant to this appears on p.l of Rollin or on the first page of the final part of his work, where the history of ‘the Grecians’ starts – at least not in the several editions I have looked at. But Rollin went through many editions, with numerous engraved illustrations, especially of monuments. My guess is that the annotation refers to a picture of a grave with inscription.

⁷⁶ Appendix B, p. 552.

⁷⁷ Example: Aesopi Fabulae Graeco-Latinae … ad usum Juventutis Regiae Scholae Etonensis accommodata, Eton, 1796, which has contracted ou only. This is one of a series of editions of Aesop put out by Eton College between 1741 and 1831: the Fables was a book regularly given to boys who won a prize for merit. That virtually guarantees the availability of second-hand copies in the bookshops.

⁷⁸ In theory, Beck's ligature-free Tauchnitz edition would do instead. Vol. 1, containing the first seven dialogues in the Stephanus order, came out in 1813. But it is unlikely to have reached Britain in time, during the blockade (n. 21 above). Horsfall, Nicholas, ‘Classical Studies in England 1810–1825’, Greek, Roman, and Byzantine Studies, xv (1974), 449–71Google Scholar, quotes (469–71) letters by various hands to the Rev. Peter Elmsley about the difficulty of getting books into and out of Germany in the period 1813–14.

⁷⁹ Massey is as moderate with ligatures as Forster and Routh, but not so clearly printed. Heindorf is ligature-free.

⁸⁰ Frances Miller kindly looked at the Valpy papers in the Longman archive at Reading University, and sent me copies of relevant documents. The list of agreements and deeds relating to Messrs Longman's purchase of Valpy's stock in 1837 mentions a Plato, but it is not Bekker. It is

Plato's Four Dialogues: the Crito, Hippias, Alcibiades and Sisyphus. With English notes and examination questions … Heindorfs valuable notes are subjoined in English. Sold by Longman & Co. … and other booksellers, London. Printed by A. J. Valpy. No editor or date named, but the Cambridge University Library catalogue knows both: G. Burges, 1831.

This seems to be good evidence that by 1837 Valpy's stock of Bekker had gone. I shall give evidence below that it had gone by 1835 or even earlier. But the Valpy archive is silent on the question.

⁸¹ In the Office of India Correspondence, where by 1826 he was Assistant Examiner (second in the Office); in 1830 he became Examiner, chief of the Office (Bain, p. 185).

⁸² ‘i.s.’ means ‘in situ’ according to Farrington, Anthony, The Records of the East India College, Haileybury, and Other Institutions, London, 1976, pp. 104 f.Google Scholar, the Rev. James Amiraux Jeremie was Professor of Classical and General Literature 1830–1850, and the Rev. Joseph Hallett Batten, F.R.S. was Principal of the College 1815–1837.

⁸³ See n. 30 above.

⁸⁴ Packe, Michael St. John, The Life of John Stuart Mill, London, 1954, pp. 104 f., 348–57Google Scholar.

⁸⁵ The first evidently went with John Stuart Mill when he married Harriet Taylor and set up house on his own. That is how it came to be included in the collection of Mill's books given to Somerville College in 1905 (not 1906, as stated in Appendix B, 552, n.) by Harriet's daughter, Helen Taylor (nn. 38 and 68 above).

⁸⁶ Bain, p. 61n.

⁸⁷ Since the Library at University College does not normally record the day and month of a donation as well as the year, the date on the label, ‘15 Augt: 1862’, will be the date of his death, not of his sister's donation.

⁸⁸ This paper began as a footnote to a wider study of the role of Plato in the lives of the Mills, , Grote, George, and Jowett, Benjamin: ‘The Past in the Present: Plato as Educator of Nineteenth Century Britain’, in Philosophers on Education, ed. Rorty, Amélie Oksenberg, London, 1998, 353–73Google Scholar. I am indebted to many people in different countries for information, advice, help, and suggestions. Besides those thanked in the appropriate note, I should mention David Blank, Abigail Burnyeat, William M.Calder III, Walter Cavini, Stephen Cretney, Richard Davies, Kyriakos Demetriou, Nicholas Denyer, S. J. Eliot, Robert Franklin, John Glucker, Peter Lewis, Wolfgang Mann, F. E. Sparshott, Robert B. Todd, and above all the indefatigable Christopher Stray. I am grateful also to the Codrington Library at All Souls College, which allowed me to borrow and keep in my rooms, not only as many volumes as I wished of The Collected Works of John Stuart Mill, but also its fine collection of sixteenth and seventeenth century editions of Plato.