Their contents follow a uniform arrangement, consisting of a series of short sections, which are made up of three elements: (1) a word or phrase, which is given prominence by projection into the left margin (cf. e. g. P. Oxy. 1801) and is divided by means of a pair of double dots from any further letters in the same line, (2) a group of, normally, four words written one below another, and (3) in the margin to the left of these and immediately below the initial word or phrase, a clearly-drawn symbol. Presumably the symbol is the tachygraphic compendium of the preceding word or phrase, and this natural supposition is confirmed by e. g. II, 28, where the symbol which accompanies opws our is composed of two parts corresponding to known tachygraphic equivalents of ws and ovv (cf. WESSELY, Ein System der altgr. Tachygraphie, No. 8, 7, No. 2, II, 1). Some surprise may indeed be felt that collocations of the length of τῇ βουλῇ καὶ τῷ δήμῳ χαίρειν οι κομισάμενός σου τήν ἐπιστολήν (Ι, 1, 6) should be compressed into a single sign, and certainly the use of such more or less arbitrary symbols must have added greatly to the difficulties of the system; an analogy may however be found in another of Wessely's texts (op. cit., No. 1), where apparently the tachygraphic equivalents of the phrases τύχῃ ἀγαθῆ, τύχῃ τῇ ἀγαθῇ, σὺν ἀγαθῇ Túyn are prefixed to a shorthand syllabary.

A correspondence between the initial phrase or word and the accompanying symbol may, then, be assumed without further demur; but what is their relation to the group of four words with which they are regularly associated? Of these words four kinds occur: (1) proper names, personal or geographical, (2) common names, mostly, but not always, in the nominative, (3) adjectives, (4) verbs, usually in the third person singular present indicative (? an aorist in I, 27). Sometimes a connexion can be readily traced throughout a group, e. g. I, 7-10, Τιτάν, πυραυγεῖ, φαεσφόρος, σελασφόρος, 35-38, σιβύνη, λάρναξ, κοῖλον, ὕπτιον, ΙΙ, 5-8, Κροῖσος, Γύγης, Κόττος, Βριάρεως, and such sequences warrant theinfer

ence that a similar logical nexus lurks in those groups where it is not so easily recognized. Mr. E. Lobel, to whom I owe some useful suggestions on this text, recalls ARISTOTLE Περί μνήμης καὶ ἀναμνήσεως (p. 452 a 14): ταχὺ ἀπ ̓ ἄλλου ἐπ ̓ ἄλλο ἔρχονται, οἷον ἀπὸ γάλακτος ἐπὶ λευκόν, ἀπὸ λευκοῦ δ ̓ ἐπ ̓ ἀέρα, καὶ ἀπὸ τούτου ἐφ ̓ ὑγρόν, ἀφ' οὗ ἐμνήσθη μετοπώρου; cf. Themist., ad loc., who adds a rather more recondite chain : Αθῆναι, Λύκειον, οἶκος Πλάτωνος, νεομηνία, συμπόσιον, Σωκράτης, τὸ τυφθῆναι ὑπὸ Σωκράτους, λύρα. The purpose of our groups, then, is no doubt mnemonic; but the mode of their application to the tachygraphic symbols and expanded equivalents remains entirely obscure. In a few instances the recurrence of a syllable is noticeable, e. g. II, 1, un-xe-t and e-xe-λos, II, 9, ax-pt and vε6-pi-s (cf. I, 35, 40, III, 32); more often however there is no such common element, and the key to the puzzle must be sought elsewhere.

Another question to which I have found no satisfactory answer is, on what principle were the symbols arranged? Here and there an intelligible sequence is found; e. g., μέχρις is succeeded by ἄχρι and ὕπερθεν by ἄπωθεν, but for the most part the order appears to be hazardous and quite unsuited to a tachygraphic dictionary. Yet a further problem arises in col. II, where below 1. 16 there is an elaborate coronis, as if to mark the end of a book, with a title consisting of the letters EP enclosed in an ornamental border (cf. e. g. P. Oxy. 850, 21; 1011, 90-91). The succeeding groups provide no obvious explanation of this singular heading. In view of these various difficulties, it may perhaps be suggested that the papyrus is a school exercise-book in which more or less fortuitous collocations would not be surprising. But that hypothesis is excluded by the literary character of the hand, which is evidently that of a practised, if not very accurate, scribe; and the scale of the text negatives the supposition that a practised scribe was merely amusing himself,

In the following transcript words have been divided

and capital initials added in the proper names, otherwise the text is printed as it stands; the symbols have not been reproduced, but the forms of these can be studied in the accompanying facsimile (pl. XVI). Letters with dots beneath them are to be regarded as doubtful. Some details of reading and other minor points are considered in the commentary.

I, 2-5. The idea of endurance is common to the first three words, but Νεοπτόλεμος is less evidently apposite. Possibly he merely forms a contrast of youth with age.

12. Perhaps Ταντάλιον, οι τανταλ. may-ταλαντ. : but the reading is not at all certain.

16. As the text stands the group consists of three words instead of the usual four. Papéτpa however is not a term which would be expected in the phrase to be treated tachygraphically, and seems moreover incongruous with what remains of that phrase. It may therefore be suggested that the pair of dots is displaced, and that Capérpa should have been the first word of the group. Cf. III, 1, where two words have been written in a single line instead of being placed one below the other, and III, 10-12, where there may be a similar dislocation or omission.

20. Most of the letters are clear and those which are

in doubt may be read with probability. Perhaps ouapaye, which is similar in sense to xe, was intended. Kóμbos looks incongruous, but may well have been inserted owing to its phonetic nearness to βόμβος.

24-25. L. ȧpxαipeoiά[?]. From the combination of this word, which though not synonymous with could easily suggest αίρεσιαρχεῖ, with Διόσκορος, it might at first sight be inferred that the well-known bishop of Alexandria who presided at the council of Ephesus in 449 is meant; but that hypothesis is at once excluded by the date of the papyrus (cf. introd.).

27. Perhaps Jo6eos, but the letters need not belong to a single word.

35. Owing to the length of this line the second vertical stroke of the N in σúvn is made to do duty as the first vertical of the H, as happens not infrequently in texts of a rather later date. This is more likely than that obvλn was written in error.

38. For the association of noiλos and πtios, cf. Arist., Meteor., I, 13, 12 (p. 350 a 10), xoiλnv nai iπτíav... περιφέρειαν.