STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world byJSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.istor.org/participate-istor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. Classical Philology Vol. V January, igio No. i CEKTAIN NUMEKALS IN THE GKEEK DKAMATIC HYPOTHESES Br Roy C Flickinger It is well known that the ancients designated the productions of the great Greek dramatists by numbers. The remains of this system, however, are scanty, being confined to the following items: arg. Soph. Antigone: XeXeKrai Be to hpafia tovto TpiaKoarov hevTe- pov; arg. Eurip. Alcestis: to Spafia hroiriBri i^; arg. I Aristoph. Aves: ea-Ti Be Xe;' and Grenfell and Hunt, Oxyrhynchus Papyri IV, p. 71: AiovvofoAe^avSpos V KpaT[etvou. Inasmuch as these numerals have become known one at a time, dis- cussion of the subject has been perennial, and ought, one would think, to have yielded a definite conclusion ere now. But such has not been the case. The meaning of these figures and the principle underlying the system are still far from being matters concerning which general agreement has been secured. On the 1 Omitted in E and E ; G reads Io-ti Xe. I have liindly been permitted to examine in proof this portion of Professor White's edition of the Aves scholia, where it is pointed out that the current statement (Dindorf, Dubner, Blaydes, Hall and Geldart, Van Leeuwen) that this phrase is not found in V, is incorrect. Blaydes (IV, p. 3) would delete the words ; but this is wrong method, since they are the lectio difficilior. In other words, it is much easier to understand how a scribe might accidentally, or because he did not understand it, even purposely omit the item than how it crept in, if it did not belong there. Fritzsche ( Quaest. Aristoph. I, p. 178) believes it a corruption of eis Si A^Kata ( !). [Classical Philology V, January, 1910] 1 2 KoY C. Flickingeb contrary, we are in a period of skepticism on the subject. Wila- mowitz has rejected all interpretations alike/ and it is significant that the authors of the articles on Aristophanes and Euripides in Pauly-Wissowa did not deem it worth their while even to mention these numbers. The discussion of the matter has greatly suffered from being conducted (for the most part) by means of brief notices in the periodicals or of such notes as the editors of the three (extant) plays in question could find room for. And the latter have usu- ally contented themselves with suggesting the most plausible inter- pretation of the text before them without considering the other instances. Of course, it is conceivable that the numbers were not uniformly derived or that their arrangement was a purely arbitrary one; and in that case its recovery would now be neither possible nor, if possible, particularly valuable. However, the recent addi- tion of the Dionysalexandros numeral to those previously known has seemed to me to warrant another attempt to discover whether they are not capable of a uniform explanation. So far only one solution of the Dionysalexandros numeral has been proposed. Korte would refer it to an alphabetical arrange- ment.^ Ranke' had made this suggestion for the Aves numeral long ago, and Wex* brought it forward again after the publication of the Alcestis numeral in 1834. This last instance is the most plausible one for the theory, since among Euripides' extant titles ^Einleitung in die griech, TragOdie (Heracles, Vol. I, Isted.), p. 150, note 55: "Zwei notizen scheinen darauf zu fuhren, dass die tragodien auch eine laufende num- mer ftlhrten, in der hypothesis der Alkestis .... und in der der Antigone .... aber Bie haben sich bisher jeder deutung entzogen." Analecta Euripidea, p. 133: "Opi- nantur sane vulgo, propter iiroiiiBri illud potissimum, numeris illis quota f abula a poeta conscripta sit indicari, et quoniam uterque numerus opinioni repugnat, corrigunt eum donee congruat aut abiciunt . . . . et Aristophanes (so. Byzantius) .... ordinis cuiusdam, cuius rationem ignoramus, numerum etiam in Alcestide indicaverit." Ibid,, p. 143 : " Obvia coniectura est, hunc ordinem ab eo ( sc. Aristophane Byzantio) indicari, sed ad liquidum perduci nequit." Van Leeuwen in his edition of the Birds (p. 3, n. 4) accepts this conclusion. Of course, Wilamowltz was not the first to advocate it, cf. Schneider De veterum in Aristophanem scholiorum fontibus (1838), p. 53 : " Immo cou- fitendum est in posterum quoque tempus idem incognitum mansurum esse, donee plura et certiora testimonia fuerint reperta." Bergk (Meineke II, p. 1000) cites Schneider with approval. •iCf. Hermes XXXIX (1904), pp. 484 f. 'Cf. De vita Aristophanis, p. cccxiv (prefixed to Thiersch's edition, 1830). <0f. Rhein. Museum II (1843), pp. 147 f. Numerals in Greek Dramatic Hypotheses 3 fifteen begin with alpha and it is by no means impossible that the same letter introduced enough of the lost titles so that, by coming at the very end of the alphas, the Alcestis might have been seven- teenth. Wex thought to strengthen his case by citing the fact that such an alphabetical catalogue appears on the back of the Villa Albani statue of Euripides.' This, however, was a most unfortunate move on his part. The Villa Albani list is not com- plete, containing only thirty-six plays; nevertheless thirteen of them begin with alpha and the Alcestis stands in the first place! Nor does the Dionysalexandros instance lend itself easily to this explanation. Although Cratinus was credited with only twenty -one plays in antiquity, Meineke and Kock admitted frag- ments from twenty-six plays into their collections under his name. Accepting this whole list, although five of them must be rejected and at least one or two of the five are a priori likely to come from the first third and although both Meineke and Kock did in fact challenge the Bousiris, we yet find only seven names beginning with alpha, beta, gamma, and delta. In order to get an eighth play Kdrte has to count the AtoVuo-ot, although that is probably a mere Verschreiben for LiowaaXe^avhpov.'' The Antigone numeral is still less probable. In our almost complete list of Sophocles' titles only twenty- three (two of which are spurious) have alpha as their initial letter. For the Antigone to be thirty-second, practically every lost title would have to begin with alpha and the Antigone stand at the very end of the plays listed under that letter. But neither supposition is likely, though, in favor of the second, it must be granted that the ancients did not arrange alphabetical lists with such strictness as do moderns and, in particular, seem commonly to have disregarded all letters except the initial. But the fact that in all three instances so far considered this theory necessitates arbitrarily placing the plays in question at the very end of those beginning with the same initial adds cumulative weight to the objection. The remaining number (that of the Aves) is just as unfavor- able. All the titles of Aristophanes' plays are known, so that ' Published by Welcker Die griech. TragOdien, pp. 444 f. 2Cf. Koch Rhein. Museum XL VIII (1893), p. 238. 4 Roy C. Fliokingee there can be no doubt on that score. But in an alphabetical list still extant' the Birds is numbered thirty-one^ or, if the spurious titles be omitted, twenty-eight. And in either case, thirty-five is entirely out of the question.' Bergk, therefore, proposed to read X' or \a instead of Xe'. Both conjectures (especially the latter) are open to objection paleographically, and the former does not satisfy the demands of an alphabetical list. If the other numerals favored the alphabetical interpretation, we might be justified in accepting Bergk's second proposal,* but under the circumstances the theory must fall to the ground." The oldest and most popular explanation is that these numbers are based upon a chronological enumeration,^ and is likewise open to grave objection. Though in my opinion this point is not one of much consequence, it has seemed to some that this theory allows too few plays for the earlier years of Sophocles and Euripides.' iCf. Novati Hermes XIV (1879), pp. 461 ft., and Pauly-Wissowa II, 972 f. 2 Counting the second Peace, which was inadvertently omitted from the Ust. This omission explains Bergk's first conjecture. 'Nevertheless, Christ seems to accept it, cf . Geschichte der griech. Literafur, p. 306, n. 1 (4th ed.). Ranke De vita Aristoph., p. cccxii£E., succeeds in making the Birds 35th by counting the 'AxSpo/i^So, rXaOxos, AaiSa\os, AijWa, EpexSeis, and 'M.TfTpo'p&v among Aristophanes' titles ! rs and Acting, pp. 46 f. (about 433). ■t Of . RA. Museum XXI ( 1866) , p. 471. See likewise Hadley's A Icestis, p. 43, Earle's, pp. 4f., and Hayley's, p. xlx. «Cf. Gr. lAteraturgeschichte III, p. 493, n. 89, and p. 414, n. 161. His reasons (in the latter case) are as follows : it is generally accepted that the Antigone was presented in the spring of 441 b. c, but the Marmor Parium informs us that Euripides won his first victory in that year ; consequently the Antigone must have received the second Numerals in Gbeek Dramatic Hypotheses 7 arbitrary changes, however, have deservedly failed to gain the approval of scholars. Leaving this problem, then, unsolved for a moment, let us proceed to see how well this chronological hypothesis fits the other instances. For the Dionysalexandros Grenfell and Hunt seem to have been inclined to accept a chronological interpretation, but were deterred by the improbability involved in their supposing the comedy to have been produced in 430-429 b. o. This date is due to the sentence KWfi^heirai S' ev rqi Spdfiari Il€/ot«\?j? fidXa TnOav&i Si' ifufxicreco'i w? eTrayetoj^ft)? rot? 'AdfjvaioK tov TroKe/xov, which occurs at the end of the hypothesis whence this numeral is derived. It should be noted that what war Pericles is thought of as having brought upon the Athenians is not specified, doubtless because other parts of the hypothesis, in which the year of production and other germane facts were given, rendered this unnecessary. Grenfell and Hunt' assume the Peloponnesian War to be intended and consequently assign the play to the years above indicated. To decide this question, we must consider the facts of Oratinus' life as we know them. He died shortly after 423 B. c. at the age of ninety-seven.'' He wrote twenty-one plays and won nine vic- tories,' six at the City Dionysia and three at the Lenaea.* Hieronymus and Eusebius mention him under 453 b. o. and his name occurs in the City Victors' List in a position corresponding prize. That the words of the argument eiSoKipL'^aavTa iv tJ SiSaffKaKlf rijs ' A.my6viis are not Inconsistent with this interpretation may be seen from the similar language of Aristophanes concerning his Banqueters (cf. Nitbes 529 and schol. ad loe.), which was likewise awarded the second prize. Furthermore, the number (32) cannot be correct, since the Antigone would then be the fourth member of its group (i. e., a satyric play), while with the proposed reading it would occupy the second place. This somewhat attractive suggestion is, however, open to the following objections: (a) we need some independent reason for believing the Antigone to be number thirty, (6) we must remember that some think the Antigone to have been produced in 442 b. c. ; cf. Christ Geschichte d. gr. Lit., p. 310, n. 7 (5th ed.) and (c) the criticism of the numeral is needless, cf. p. 13, below. I may safely ignore the numerous conjectures for the Alcestis number (t^', «', etc.). 'Followed by K6rte, Wilamowitz, Oroiset, Rutherford, Thieme, etc., cf. Thieme's Quaestionum comicarum ad Periclem pertinentium capita tria (1908), pp. 27 ff. 2Cf. Lucian Macrob. 25, and Capps Harv. Stud. XV (1904), pp. 61 f., who places his death "soon after 422/1." sCf. Smdass.i). Kpartms. *Ct. CIA II, 977 d and i, and Capps A. J. P. XX (1899), pp. 390-96. 8 KoY C. Fliokingee to about that date. The arrangement of the list would seem to indicate that he won several other victories soon after, and, finally, he produced plays in 425, 424, and 423, when he closed his career with a victory.' Thus, his eighth play might well have come as early as 446 b. o. That Cratinus had already at this general period directed his satire against Pericles is proved by Meineke II, p. 218, where he is charged with dilatoriness in completing the middle wall and ibid., p. 61, where there is an unmistakable allusion to his escape from ostracism in 442 b. o. These were troublous times for Athens. In 447 the battle of Chaeronea cost her Boeotia, and the revolt of Euboea and Megara quickly followed. Moreover, as soon as the five years' truce expired (446), the Spartans invaded Attica. The responsibility for these disasters was naturally placed at Pericles' door by his opponents. The attempt of the goddesses to bribe their judge in the Dionysalexandros might well have called forth an allusion to Pericles' alleged bribery of the Spartan king to withdraw his army, and the fact that the country people were now for the first time couped up within the Long Walls might have been referred to in that part of the play where Helen is concealed in a bird-cage (11. 29 ff. : Kal ttjv jxev ''^X'qvrjv eh rdXapov &crTr[^€p opvivy Kpvy^a<:). In my opinion, however, it is possible to estab- lish a still closer and broader connection between the plot and contemporaneous events. Dionysus, of course, represents Pericles himself and Alexander the conservative leader Thucydides, whose prerogatives Pericles (Dionysus) has usurped. The immense resources at his command would have enabled Pericles to establish his own personal authority over his people [rvpawiho'; ukiv/jtov, Hera's offer in the bribery scene) or by wise military expenditures to guarantee the stability of the Athenian empire through success in war (evrvxia.'S Kara nvoXefiov, Athena's offer). Recent occur- rences might well have seemed to justify Pericles' enemies in believing that he had lost these two chances forever by deciding in favor of Aphrodite. The hypothesis tells us that Aphrodite "Cf. argg. Acham., Equit., and Nubes, and Luclan Macrob. 25. • So Korte. Grenfell and Hunt supply rvpiv. Thleme loc. cit., p. 16, finds addi- tional support for Korte'e reading in Meineke II, p. 42, fr. xii : xvo^o"!"'^ /3oi//t6Xoi. Numerals in Geeek Dramatic Hypotheses 9 promised that Dionysus should be "most beautiful and most beloved" {kcIX\i(7t6v re Kal eirepauTov avrbv vnrdp'xeiv) ;' and as a result he succeeded in wooing Helen away from Sparta. Of course, the myth hampered Cratinus and the symbolism is less exact and clear than if he had been dealing with a theme of his own inven- tion throughout, but there ,can be little doubt as to the meaning of Aphrodite's offer — it referred to Pericles' plans for the beautifi- cation of Athens. This is not the only place where Athens is com- pared to a woman. In a passage which Sauppe has correctly recognized as derived from a comic poet Plutarch describes the city at this very period as a wanton decking herself out with costly marbles" and thousand-talent temples.^ If this interpreta- tion is correct, we can understand the meaning of one of the extant fragments (Meineke II, p. 42, fr. ix) : TrapaoraSas kol irpodvpa /SouAci iroiKiXa. Dionysus (i. e., Pericles with his ambitious building schemes) evidently expressed surprise at the simplicity and rudeness of Alexander's mountain quarters and was answered with this sar- castic comment. In the denouement Alexanderhandsover Dionysus to the Achaeans' and keeps Helen for himself; in other words, Cratinus advises the people to accept Pericles' plans for the city's adornment, since it is now too late to turn back, but to disavow their author and intrust their administration to safer hands. If this explanation seems to attribute too bold an utterance to the poet, it must be remembered that Pericles' name was not yet sur- rounded with such a glamour as at a later period, that at this particular hour the course of eventst seemed against him, and, in particular, that the aristocratic party was so confident of success that soon afterward (442 b. o.) they dared to challenge him to the test of ostracism. If the outcome disappointed their fond ' Can the reading be KiWuTTOv .... airrf iicdpx^^>'1 2 Cf . Plutarch Pericles xii and Sauppe Ausgewahlfe Schriften, p. 500, n. 2. There may be an allusion also to Aspasia, whose connection with Pericles began at about this period, cf. Pauly-Wissowa II. 1716 f. 5 Wilamowitz sees in this a reference to the Spartan demand for his surrender in 432 B. c. We ought not to suppose, however, that this was the first time Pericles' enemies had thought of his descent from the accursed Alcmaeonidae. On the contrary, this step was doubtless merely the finale of long years of discussion both at home and abroad. 10 KoY C. Fliokingeb expectations, it was neither the first nor last instance of such mis- placed optimism. Furthermore, such freedom of attack is consis- tent with the known character of Cratinus' comedy, and the fact that, as soon as Pericles' position was assured (after 442), restrictive legislation was enacted' proves earlier excesses. I therefore believe the Dionysalexandros to have been brought out in the spring of 445 and think that it may have been Cratinus' eighth play. Some one may object that this date is too early for such "Mythenparodie" as this play affords. And it is true that this motif is more characteristic of the Middle Comedy, and that this is one of the reasons for assigning the Nemesis to the close of the century and to the younger Cratinus.^ But no one would put the Dionysalexandros much later than 430 — a date twenty or twenty- five years earlier than the new date for the Nemesis — and there is no reason to suppose that the Dionysalexandros was the first play of this sort. The fact is that owing to the dearth of infor- mation before the Acharnians negative generalizations are quite unsafe. The Aves numeral has always been a stumbling-block in the way of every interpretation, and it is no exception in this case. It is evident at once that the Birds cannot be the thirty-fifth piece Aristophanes brought out.' So that the only recourse left adherents of this explanation is to follow the path that Dindorf pointed out long ago: "Ut si quid absurdi edicere videantur, emendanda potius quam eicienda esse pateat." If any solution that would fit the case (even without regard to its suitability in other instances) had ever been propounded, this procedure would be less justifiable; but under the circumstances it is only a ques- tion of finding an intelligible and paleographically possible conjecture. As we have seen, advocates of the alphabetical inter- pretation have made use of this method here, but without satis- factory results. Dindorf, however, proceeded to employ his own 1 Cf . schol. Achar. 67 : t6 ^i}0ur;ua r6 ttepl rov nil KW/ufSttv, ypa^'ii' iirl MopvxlSov (440/39). 2Cf. Harvard Studies XV (1904), pp. 61 fE. ^Neverthelees, some hare blindly adopted this explanation, cf . Bockh CIQ I, p. 3506 and O. K. MilUer (Donaldson) Hist, of Lit. of Anc. Greece I, p. 449, n. 1. Numerals in Gkeek Dramatic Hypotheses 11 advice and in a most convincing manner ; cf . his Oxford edition of Aristophanes II, pp. 524 f.: "Est igitur qui in libris traditur numerus Xe aperte corruptus et, si quid litterarum ductibus tribu- endum, fortasse restituto te' corrigendus. Ita ad novem, quas Avibus antiquiores esse fabulas ex didascaliis constat, quinque aliae de ceteris accedere debebunt. Eae autem fortasse hae sunt, Teapyoi, Ti]pa<;, Filpijvr) Bevrepa, 'OXwaSe?, TayTjviaTai. Ceterum \e' ne tum quidem verum erit, si fabulas secundum litterarum ordinem numeraverit grammaticus: quod non est verisimile." Between 427 B. c. and about 388 b. o. Aristophanes wrote forty plays, or one a year. Inasmuch as the Birds was produced in the fourteenth year of his career, a priori there is no difficulty involved in supposing it to have been fifteenth in chronological order. For his earlier years we have the following details: Banqueters 427, Lenaea Babylonians 426, City Dionysia Acharnians 425, Lenaea Knights 424, Lenaea Clouds A ' 423, City Dionysia Wasps 422, Lenaea Peace A' 421, City Dionysia Amphiaraus 414, Lenaea Birds 414, City Dionysia To these must be added the unknown piece with which it seems likely that he won a City victory in 425.^ To the list of plays suggested by Dindorf as belonging to this period must be added the Apafiara rj K.e'vravpo^J' For some of these plays there is room between 421 [Pax A') and 414 [Amphiaraus), for, although Aristophanes seems to have then been less productive than earlier, there is no reason to believe that this was a period of complete inactivity.' 1 Of. Capps A. J. p. XXVIII (1907), p. 198. 2 So Wilamowitz. The literature concerning this and the other plays in question is cited in Pauly-Wissowa II, 976 fE. In arg. Vespae there is a confused notice con- cerning Philonides and a Tlpoaytiv, which in my opinion ought not to be identified with Aristophanes' play of the same name; cf. Capps A. J. P. XXVIII (1907), p. 199. sCf. Bergk (Meineke II, p. 995) : " At vero ad illud tempus, quod inter Vespas et Aves intercedit, fere nuUae tabulae ref eruntur, neque tamen per tam longum temporis spatium prorsus conticuisse Aristophanem consentaneum est;" and Pauly-Wissowa II. 979 : " Sicher scheint nur, dass A. zwischen dem Frieden und dem nUchst datierbaren Stack Amphiaraos nicht vollig pausiert hat, wenn wir auch die Lttcke nicht ausffillen k6nnen." 12 KoY C. Flickingee We have thus seen that the chronological interpretation falls short mainly at two points, viz., the Antigone and Alcestis nume- rals are somewhat smaller than we might expect (not a serious objection) and do not seem to conform to the tetralogic system used at the City Dionysia. Perhaps a study of the purpose of these numbers would assist us. BOckh' pointed out long ago that they were not official but due to private research, but further than this we have not been able to go. We are therefore fortunate in that the Dionysalexandros instance has at length made the pur- pose clear beyond reasonable doubt. This number is taken from an hypothesis. It was not, however, incorporated within it, but stood at the top of the second (last) column as a sort of heading and had doubtless appeared also at the beginning of the hypothesis (now lost). In my opinion, this was the original form of such a notice and shows why in the fuller form of statement found else- where a different verb is employed in each case — iiroirjOr], \e\eK- rai, ea-Ti. When Aristophanes of Byzantium, or whoever was responsible for the change, transferred these items from the head- ing and made them integral parts of the hypothesis, finding no verb in the original statement before him and resting under the necessity of now using one, he did not deem it essential to para- phrase the information always in the same way, but, as was natural, employed now one expression and now another.'' If it be true that the original function of the numerals was as we find it in this uTTo'^ecrt?, only one explanation is possible — it is a device for the convenience of some library, probably that at Alexandria. If so, every play in the library would bear a number, and these numbers 1 Cf . CIG I, p. 351a. 2 Consequently, It is inadvisable to press the use of a particular word in any one case, cf. Wex Antigone, p. 35 ; "Puerunt adeo qui .... bis et tricies earn (so. Anti- gonam) in scaenam productam esse opinarentur, decepti scilicet male intellectaaliqua Aristophanis grammatici notatione in argumento fabulae, ubi dicit ille : \4\eKTai di t6 Spafia X|3', quod compendium legere illi volebant dvoKaiTptaKorrdKis;" Wilamowitz Analecta Euripidea, p. 133: " At \4\eKTai. illud lucem accipit ab Andronici de Platone verbis oBrtvos rk Spd/juiTa rdSe X^yofTat, i. e. in catalogis recensentur ; " and Ranke De vita Aristoph., p. clxvii : '• Ipsa vox XAsktoi satis ostendit, non esse de doctis, sed de servatis lectisque fabulis dictum." This last is in embryo the theory which I have developed in the text ; but it is here a mere guess unsupported by evidence, and is further hampered by being yoked with the alphabetical explanation, cf. ibid., pp. cccxii ff . Of course, 4woi-fi6rj lends itself most easily to a chronological interpretation ; but the question must be decided upon broader grounds than that. Numerals in Geeek Dramatic Hypotheses 13 would run consecutively for each author. In other words, if any play were not represented in the collection, that fact would not be indicated by a gap being left for it in the enumeration. Of course, it is conceivable that the basis of arrangement was purely arbitrary, but it is more probable, until the contrary be proved, that some rational system (alphabetical, chronological, etc.) was employed. Now which system was actually chosen becomes clear when we note that the above statement of the numerals' purpose and use obviates the two objections to the chronological interpretation. For the extant fragments prove that of Euripides' ninety-two plays only about seventy-two were known to the Alexandrians. This test can always be accepted, when the conclusion is positive, and usually when it is negative. Thus, the Roman Didascaliae report that the Bacchae of Lysippus was the only play of his pre- served and sure enough it is the only one from which we have quotations.' Accordingly, we are not surprised to learn from Suidas that seventy-eight plays of Euripides (four of them spuri- ous) were preserved.^ If the Alcestis was seventeenth among this number and retained the same relative position as in the original list, it must have been about the twentieth play Euripides brought out, which number (being a multiple of four) would be suitable for the last play of a tetralogy. Similarly, the hundred Sophoclean plays, which are preserved in whole or part make it seem likely that the Antigone was about the thirty-ninth play that he produced. These figures are, of course, mere estimates, but they have the merit of assigning a slightly larger number of plays to the earlier years of these poets and of affording a reasonable solution for these two misfit numerals. In the case of Cratinus and Aristophanes the fragments indicate that all their comedies were still extant in ancient times, and consequently these two numerals are not altered by the proposed interpretation.' Of 1 Of. IG XIV. 1097, and Capps Class. Phil. I (1906), p. 219, 1. 9 and note. 2Cf. Pauly-Wissowa VI, 1247. 3 If the date I have assigned the Dionysalexandros should prove to be incorrect, that fact vrould not necessarily invalidate my interpretation of its numeral, for, (1) as only about a dozen of the twenty-six titles which Meineke accepted under the elder Cratinus' name are indisputably his, it is possible that several should be assigned to the younger poet, that not all the elder Cratinus' twenty-one plays were known to the 14 Roy C. Fliokinger course, to render this hypothesis absolutely certain, one ought to arrange in chronological sequence such plays of these poets as the Alexandrians knew and to prove that the Dionysalexandros was eighth in Cratinus' series of twenty-one, the Aves fifteenth in Aristophanes' series of forty-four, the Antigone thirty-second in Sophocles' series of about one hundred, and the Alcestis seven- teenth in Euripides' series of seventy-eight.' But our meager didascalic information puts such a procedure entirely out of the question, and we must be content to have shown that such a theory satisfies the general probabilities of the case. The question at once arises where second versions would be placed in such an enumeration as we have been considering. If the second version had actually been produced, its number would naturally correspond to its chronological position. But if it had merely been published and were included in the library collection, it would still have to have a number, and I am of the opinion that, no matter how long an interval had intervened between the pro- duction of the first version and the publication of the second, the latter's library number would immediately follow that of the former. Such a supposition would explain the source of the mis- information in arg. V Nubes. The extant fragments and the fact that the differences between the two versions are so clearly dis- tinguished in arg. VI prove that both were preserved in Alexan- drian times. The anonymous author of arg. V correctly tells us that the first Clouds was produced at the City Dionysia in 423 and was awarded the third place, and then incorrectly adds that Aristophanes brought out the second edition in the following year and was still less successful [dTroTvx^v ttoXv fjidWov). Anony- mous probably reasoned as follows: Alexandrians, and that the Dionys. was eighth (chronologically) among those pre- served, or (2) it is possible that Cratinus was much more productive than is commonly supposed, that only twenty-one of his plays were known to the Alexandrians, and that thirteen preserved plays belonged to the period after 430-429. 1 Extant fragments prove that the Alexandrians were acquainted with Euripides' Pleiades (with which he made his maiden appearance), and Cressae, Alcmeon in Psophis, and Telephus (which came out with the Alcestis). These plays, then, would be numbers 1, 14, 15, and 16, respectively. He is known to have won his first victory in 441, and there is no doubt that these plays (or some of them) are represented in our fragments, if we only knew how to pick them out. The satyric plays would probably be the first to disappear, cf . arg. Medea. Numerals in Greek Dramatic Hypotheses 15 My edition of Ne<^eAai A' bears such and such a number and the records show that it was third in 423. My edition of Net^eXot B' bears a number larger by one than the other, but the Didascalia^ say nothing about it. The numeral indicates that the second edition came out the year after the first. I explain the silence of the records by supposing that, although the practice of having five poets compete at a time obtained,' it was at this time the custom to record the names of only the three highest contestants. Therefore, I conclude [we may represent Anonymous as saying] that in his second attempt Aristophanes fared still worse than before and was awarded fourth or fifth place. We thus obtain an adequate and plausible explanation for an error that has caused scholars much trouble. It remains to point out that such an arrangement as I have suggested would not have seemed an unnatural thing to the ancients. Terence's plays, for example, are not only arranged in that way in our MSS but are given numerals on that basis in the Didascaliae. It is interesting to note that there was some uncer- tainty as to what should be done in the case of a play that failed and was repeated, like the Hecyra. The MSS arrangement ignores the first production of that play (after the Andria), when it was scarcely more than begun, and gives it a position corre- sponding to the second exhibition (after the Phormio) ; doubtless because on this second occasion a considerable portion of the play was presented, although it was not produced in toto until the third attempt (after the Adelphoe). Nevertheless, the didascalic numeral assigned the Heauton is III, as if the first exhibition of the Hecyra were to be counted. The Eunuchus, however, is likewise given the number III, and the other numerals follow its lead. The Terence numbers, of course, trace back the system only to the Romans and to about the time of Varro. Aeschines' speeches, however, are also arranged in this way, so that there is no doubt the Alexandrian Greeks were familiar with such arrange- ments. Wilamowitz's objection,^ "At haud probabile aut certe non probatum est apud scaenicos Graecos umquam eandem dis- ponendi rationem usu venisse. catalogi enim, unicus, unde aliquid 1 Both before and after the Peloponneslan War the number of competitors seems to have been five, but for a time in the interim only tliree, cf. Capps Class. Phil. I (1906) p. 219, n. on 1. 5. ''Analecta Euripidea, p. 132. 16 Rot C. Flickinger sciri potest, fons nullum ordinem norunt praeter eum, qui ex lit- teris principalibus titulorum fabulas digerit (/cara aroixeiov^,'''' overrides the very evidence we have been considering. To summarize: If we follow Dindorf in reading te' for \e' in arg. Aves, the numbers are capable of a uniform interpretation. They were a library device and were assigned the plays repre- sented in the Alexandrian collection according to the date of their production. A second version of a play, if only published and not actually produced, was given a number immediately following that of the first version — a practice which explains the error of Anonymous arg. V Nuhes concerning the second Clouds. At least nineteen plays of Euripides preceded the Alcestis. Cratinus' Dionysalexandros was probably brought out in 445 B. c. APPENDIX There are still two other numerals which used to be considered in this connection: and although they are now commonly disregarded and I have made no use of them in the preceding discussion, I do not feel free in a paper of such length to neglect them entirely. 1. Bekker Anecdota Graeca, p. 430. 15: d5roXoy«racr0at koI airoXoylliuv TO €irii(X0€LV iKacrra. ' Apiavr]i iv t<5 6 ^rjpa- iyC> S' diroKoylt^eiv re, ki^t'' iir'' &,v6p6.K6.V7)i Iv t XeTrifeix. 2 It Is still cited as such by Christ (Schmidt) Oeschichie d. griech. Literatur, p. 403, n. 2 (5th ed.). Cf. Suvem Vber Aristophanis Tf/pas, p. 24. Numerals in Geeek Dramatic Hypotheses 17 Stas opviOai iyKaOiv&uv crvixfiijiriKev, Apurro^avrp Xiya, mnrep Kal KptfiAcrTpav iv TaZi Net^cAflis. Bothe' seems to have been the first to bring this pas- sage into the discussion and from the comparison with the Pollux passage drew the very natural inference that e referred to the Clouds in a chron- ological arrangement — Banqueters 427, Babylonians 426, Acharnians 425, Knights 424, and Clouds 423. The difficulty is that, while KpepMTTpa occurs in the Clouds in the form KpenajSpa (1. 218), irinvpov apparently does not; and it has accordingly been cited among the fragmenta fabu- larum incertarum by Dindorf II, p. 696; Meineke (Bergk) II, p. 1218; and Kock I, p. 582. Two problems confront us and it would be well to consider them separately: (a) does irerevpov occur in the Clouds and, if so, (6) does iv T' o!> ai AXfKTpvoplSes Kot/iiovTai (Rutherford) and the Pollux passage in the text. Note also iieriwpov ^i\ov in Photius. 18 EoY 0. Flickingek Does €v ria I refer to the number of the Clouds or to the spelling of irtTcvpoi'? Opinions differ. Bergk thought I a numeral but that the Clouds was not the play referred to.' There is a pronounced tendency, however, to believe that the phrase has reference to the spelling.^ This view is certainly possible, although a cursory examination of a portion of Photius has not yielded me a precise parallel. Hilberg cites o-iva/ttopos- Iv T(p d, oix' o-ivofttopos. But of course this is infinitely more explicit than the instance under consideration, and the same objection can be lodged against the nearest parallels my hasty search discovered. Cf.' aSa^a-ai- TO KVTJdax, ovK Iv Tio <^o^ oSa^rjiTai .... ' KpuTTo4>avrfi iv 'OXkoxtw, and dA.oa- (xavTa- (V Tm a, ^ipexparrp Iwvw .... dXo^aai iv rm rj. On the other hand, parallels to justify interpreting e as a numeral are little more convincing. Plays are regularly cited in the dative case either with or without the preposition €v. Sometimes the two construc- tions occur side by side, cf. aiMvySoXrj .... 'Ep/tiirTros ^opfw6pois koI *i\^/A(ov iv MwTTiSi Koi Ai<^iA.os TeXetna. But, of course, no numerals are used in this connection. They are employed, however, in referring to the books of histories, dialogues, etc., cf. dyZva- ©ovxvStSTjs i* "irpoiXOmv as Tov ayiova" ktX. [Thuc. v. 50]; avadoXuxrai- dvarapct^ai, Ndjiwov ^ XlXaruiv [^Laws 8246J; advixm- 'HpdSoros iv rm d avrov Xdyoi Tr]V driip^iav Xeyei [Herod, i. 37]; aijuao-ia .... SrjXoi Se 'HpdSoTos iv Trj a. [Herod, i. 180]. These parallels, though not entirely satisfactory, are yet close enough to prove that e might possibly be a numeral. Further than this it is neither pos- sible nor necessary to go. The Clouds may have been Aristophanes' fifth piece, though, if so, we must reject the City victory in 425. The important point is that the chronological interpretation can readily accommodate itself to the possibility of the Clouds being Aristophanes' fifth play. To the supporters of the alphabetical explanation, however, such a possibility is disconcerting, since in an alphabetical list the Clouds would be number twenty-seven.* Northwestern University iMeineke 11, p. 1218; contrast p. 1000. 2Cf. Dindorf II, p. 696; Hilberg Zeit.f. Sster. Gym. XXX (1879), p. 906; Kook I, p. 582. 'AH the Photius quotations that follow are taken from the new text published by Reitzenstein Der Anfang des Lexicons der Photios. * Professor Capps suggests the following : the original text was ENTOINE^EAAIS- APAMATI, which was shortened to BNTfilNE and then corrupted to ENTfflE.