?^f 7 
 
 17/ 
 
 CORNELL 
 
 UNIVERSITY 
 
 LIBRARY 
 
Z 8697.L74""'" """"""y "■'""'y 
 
 "■"^e ancient editions of Plautus. 
 
 3 1924 008 359 64? 
 
 Date Due 
 
 „j^j |^2r9-B9&^ _ 
 
 im 
 
 "8-M 
 
 mmmmm 
 
 23 233 
 
 »^™— ^sasra 
 
The original of tliis book is in 
 tine Cornell University Library. 
 
 There are no known copyright restrictions in 
 the United States on the use of the text. 
 
 http://www.archive.org/details/cu31924008359642 
 
THE ANCIENT EDITIONS 
 OF PLAUTUS. 
 
 BY 
 
 W. M. LINDSAY, Hon. Ph.D. (Heidelb.) 
 
 Professor of Humanity in the University of St. Andrews. 
 
 OXFORD : 
 lant^a barker anir ©0. 
 
 27 BROAD STREET ; AND 3 1 BEDFORD STREET, STRAND, LONDON. 
 
 1904. 
 
 t 
 
 ^7 
 
^7. 
 
 L7H 
 
 
CONTENTS. 
 
 PAGE 
 
 § I. Preliminary - - - - . x 
 
 § 2. Plautus-citations in Varro - - - 2 
 
 § 3. Plautus-citations in Verrius Flaccus - - 13 
 
 § 4. Plautus-citations in Nonius Marcellus - - 23 
 
 § 5. Other evidence - - - - 27 
 
 § 6. The Ambrosian and Palatine Archetypes - - 35 
 (I.) Lines inserted or omitted or substituted in 
 
 the ' Revival ' Text - - - 37 
 (II.) Phrases or words altered in the ' Revival ' 
 
 Text - - - - 57 
 (III.) External Form of the two Editions, Colo- 
 
 metry, Order of Plays, &c. - - 78 
 
 (IV.) The Arguments - - - 86 
 
 (V.) The Didascaliae - - - 88 
 
 (VI.) The Scenes and Scene-headings - - ib. 
 
 (VII.) Errors common to the two Recensions - 104 
 
 (VIII.) Hiatus - - - - 118 
 
 (IX.) Orthography - - - 136 
 
 ' § 7. Conclusion ----- 142 
 
THE ANCIENT EDITIONS OF PLAUTUS. 
 
 § I. Preliminary. 
 
 It is not until about the fourth century a.d. that we get a clear 
 view of the text-tradition of Plautus. But we can tap the pre- 
 vious flow of tradition at two points, at the time of Varro and at 
 the time of Verrius Flaccus, and get at least a glimpse, obscured 
 unfortunately by the fragmentary condition in which the writings 
 of these authors have survived. It will thus be convenient to 
 assign two separate sections in this monograph, one to the Plau- 
 tine citations of Varro, and another to those of Verrius Flaccus, 
 before proceeding to that period to which our extant MSS. can be 
 traced back, which is also, roughly speaking, the period of the 
 grammarian Nonius Marcellus. 
 
 For the earliest stage of all, the interval between Plautus and 
 Varro, the information that can be gleaned is so scanty and at 
 the same time (thanks to the work of Ritschl) so well known that 
 it is unnecessary to give it anything more than a brief notice here. 
 We know that the original manuscript text of a play would pass into 
 the hands of a theatrical manager, and that until the plays became 
 available for reading and for grammatical research, they would 
 exist only in the form of stage copies ; we know that there was 
 a revival of Plautine Comedy in the last century of the Republic, 
 and that the plays were re-staged * in a more or less altered form 
 
 " The frequently quoted prologue of the ' Casina ' was written for a revival 
 performance, which took place about a generation or more after the poet's 
 time, as we see from vv. II sqq. : — 
 
 nos postquam populi rumore intelleximus 
 studiose expetere vos Plautinas fabulas, 
 anticuam eius edimus comoediam, 
 quam vos probastis qui estis in senioribus ; 
 nam iuniorum qui sunt non norunt, scio. 
 The play was re-named ' Sortientes ' (vv. 30 sqq.) :— 
 comoediai nomen dare vobis volo. 
 Clerumenoe vocatur haec comoedia 
 graece, latine Sortientes. 
 Similarly the Mostellaria seems to have been re-named ' Phasma,' and so 
 
 B 
 
2 Tfie Ancient Editions of Plautus. 
 
 and often under a new title ; we know that a number of gram- 
 marians and antiquarians (e.g. Aurelius Opilius, Ael. Stilo, Ser. 
 Clodius) applied themselves to the discussion of the difficulties 
 which the text offered to readers and provided the plays with glos- 
 saries of rare and obsolete words. The most enthusiastic student 
 of the poet was Varro's teacher, L. Aelius Stilo of Praeneste, and 
 it was probably from him that a good deal of the Plautine lore 
 published by Varro originally came. 
 
 § 2. Plautus-citations in Varro. 
 
 Of Varro's large treatise (lexicographical, grammatical, etymo- 
 logical), the 'De Lingua Latina,' only Books V — X have been 
 preserved. In compiling them he seems to have largely availed 
 himself of the ' glossarum scriptores,' and their interpretations of 
 the difficult words in Plautus and the other early authors. The 
 interpretations offered by Ser. Clodius, the son-in-law of Aelius 
 Stilo, have a prominent place, also those of Aurelius Opilius and 
 others. For example, Varro mentions the puzzling line (Plaut. 
 frag. Nervolaria 96) : 
 
 scrattae, scrupedae (?), strittibillae, sordidae, 
 
 and tells us that one grammarian wrote the second word as scauri- 
 pedae, another (whom Varro himself seems to follow) as scrupipedae, 
 a third authority differently ''. Here are Varro's words (L.L. VII 
 65) : scrupipedam Aurelius {i.e. Aur. Opilius) scribit, a scauripeda; 
 luventius comicus dicebat a vermiculo piloso, qui solet esse in 
 fronde cum multis pedibus ; Valerius a pede ac scrupea. The 
 
 on (cf. Ritschl, Parerga I 180 sqq.). Another evidently ' Revival ' prologue is 
 that of the Pseudolus, found in both A and P in this brief form : — 
 Exporgi meliust lumbos atque exsurgier : 
 Plautina longa fabula in scaenam venit. 
 
 Editors suppose this couplet to be a mere fragment, the rest having been 
 lost. I do not see why it should not be the whole prologue. The opening 
 scene, one of the cleverest and liveliest in Plautus, explains the whole situa- 
 tion clearly enough, and Plautus probably never wrote a prologue for the 
 play. The ' Revival ' stage-manager would wish to announce to his audience 
 as briefly as possible that an old favourite was being re-staged. The expression 
 ' Plautina longa fabula ' is noticeable in view of the frequent curtailment of 
 scenes in the ' Revival ' text. 
 
 b Was it as scripipedae, from * scrips (Greek okvI^) and pes 1 
 
The Ancient Editions of Plautns. 3 
 
 result of the grammarians' researches had apparently been to pro- 
 vide the line with three variant readings. 
 
 Another example of the disagreement of authorities, but whether 
 accompanied or not by divergence of reading is not clear, we 
 find in VII 106, where Varro speaks of the word delicitum in 
 Cas. 206 : 
 
 quando tibi domi nihil delicuum est. 
 
 Varro remarks : dictum ab eo quod deliquandum non sunt (? leg. 
 est), ut turbida quae sunt deliquantur ut liquida fiant, Aurelius 
 scribit delicuum esse ab liquido ; Claudius " ab eliquato. Siquis 
 alterutrum sequi malet, habebit auctorem. 
 
 The same two rival authorities are quoted again in connexion 
 with the vfoxA fraefica (VII 70) : in Truculento : 
 
 sine virtute argutum civem mihi habeam pro praefica ; 
 
 dicta, ut Aurelius scribit, mulier, ab luco {scil. Veneris Libitinae) 
 quae conduceretur quae ante domum mortui laudes eius caneret. 
 Hoc factitatum Aristoteles, scribit in libro qui inscribitur vofufia 
 fiapPaptKa, quibus testimonium est quod t fretum (? kg. tritum) est 
 Naevii : 
 
 haec quidem hercle opinor praefica est ; nam mortuum collaudat. 
 
 Claudius scribit : ' quae praeficeretur ancillis, quemadmodum la- 
 mentaretur, praefica est dicta.' Utrumque ostendit a praefectione 
 praeficam dictam. 
 
 In VII 107, Varro mentions the practice of appending inter- 
 linear (or marginal) glosses to the texts of early writers. Speaking 
 of the vioxd persibus, found in the Demetrius, a comedy of Naevius, 
 he says : sub hoc glossema " callide " subscribunt. And in VII 
 64 he seems to give us a string of such glosses taken from a text 
 of the Cistellaria and bearing on v. 407 : 
 
 diobolares, schoenicolae, miraculae. 
 
 Diobolares a binis obolis. Schoenicolae ab schoeno, nugatorio 
 unguento. Miraculae a miris, id est monstris, a quo Accius ait 
 personas distortis oribus deformis ' miriones.' 
 
 « This is Ser. Clodius, son-in-law of Varro's teacher, Aelius Stilo. Did 
 he then write elicuum for delicuum ? 
 
 B 2 
 
4 The Ancient Editions of Plautus. 
 
 We can easily imagine how marginal notes of this kind on such 
 a line as that previously quoted : 
 
 scrattae, scrupedae, strittibillae, sordidae, 
 would produce in subsequent copies a variety of reading in the 
 second word of the line. One scribe would bring from the margin 
 into the text one of the variants mentioned, while others would 
 give preference to a different variant. 
 
 Varro's quotations of Plautus in the ' De Lingua Latina ' give 
 us apparently the text on which previous commentators or gloss- 
 writers based their glosses. We can therefore rely on the accuracy 
 of these quotations, since they are taken either from actual texts 
 edited by these scholars or, at least, from commentaries closely con- 
 nected with actual texts*. In VII 8i, Varro« cites Pseud. 955 
 in this form : 
 
 ut transvorsus, non provorsus cedit quasi cancer solet ! 
 
 whereas our traditional reading (in A apparently as well as in P)\%: 
 
 non prorsus verum ex transverso (-vor- ?) cedit quasi cancer solet. 
 
 Editors have, I think rightly, given preference to Varro's setting 
 of the line, judging that the change of the antique provorsus to 
 prorsus was a change made at the time of the Plautine Revival, 
 when obsolete words and phrases would naturally be re-cast in a 
 form more intelligible to the audience of the day. 
 
 It will be well to give here a full list of all the passages of Plautus 
 cited by Varro and a collation of them with the Ambrosian Pa- 
 limpsest (^) and the Palatine archetype (/*). It will shew that 
 with the exception of the line just quoted, in which the ' Revival ' 
 version has found its way into A as well as /*, the text used by 
 Varro (or rather by the ' glossographi ' whom he quotes) was practi- 
 
 ^ it might be urged that in any case Varro's reputation as a scholar forbids 
 us to attribute to him carelessness in quotation. Still we must remember that 
 the present standard of accuracy is hardly applicable to a time when books 
 were far scarcer and less accessible, and above all had not that convenient 
 division, of page and paragraph and lines which makes reference to a required 
 passage so easy now-a-days. 
 
 ^ He goes on to say : dicitur ab eo qui in id quo <[it> est versus, et ideo 
 qui exit in vestibulum, quod est ante domum, prodire et procedere ; quod 
 cum leno non faceret, sed secundum parietem transversus iret, dixit ' ut trans- 
 versus cedit quasi cancer, non proversus ut homo.' This sounds like the lan- 
 guage of a person who has seen the play acted. 
 
The Ancient Editions of Plautus. 5 
 
 cally identical with the text oi A P, even (so far as can be gathered) 
 in respect of Hiatus and of Colometry. 
 
 Amph. 27s (L.L. VII 50, s.vv. lugula, Vesperugo ; VI 6 s.v. 
 Vesperugo) neque lugula neque Vesperugo f neque Vergiliae occi- 
 dunt {iugulae P, A n.l.). 
 
 That the second half, at least, of the line (for the two opening 
 words are not quoted in VI 6) shews the precise form in which it 
 appeared in the text or commentary used by Varro is absolutely 
 certain. Varro therefore recognized the hiatus Vergiliae \ occidunt. 
 
 ? [Amph. 488 (ap. Non. s 57. 18, s.v. Enixae) uno ut labore {vel 
 uno labore) exsolueret (/<§". -re ?) aerumnas duas (uno ut labore ab- 
 soluat P, A n.l.)]. 
 
 Asin. 685 (L.L. VII 79, cf. VI ^ 7, s.v. Conticinnum) uidebitur, 
 factum uolo redito conticinno (r. hue c. P, A n.l.). 
 
 It is difficult to decide which is the right reading. Hue, if written 
 in its older form hoc might easily drop out in this context. On the 
 other hand the addition (as well as the omission) of these small 
 Adverbs and Conjunctions is a besetting sin of scribes of Plautus. 
 
 Aul. 191 (L.L. V 14, s.v. Locatum ' coUocatum') 
 
 uirginem habeo grandera dote cassa {leg. -am) atque inlocabili 
 
 {leg. -lem), 
 neque earn queo locare cuiquam (P, A n.l.). 
 
 Aul. 446 (L.L. VII 103, s.v. Pipulo 'convicio') pipulo te' dif- 
 feram ante aedis {p. hie. d. P, A n.l.). 
 
 ' Varro says (L.L. VII 50) : Vesperugo Stella quae vespere oritur, a quo 
 earn Opilius scribit Vesperum. This can hardly mean that Aurelius Opilius 
 proposed to read Vesperum for Vesperugo in this line. Varro says elsewhere 
 (L.L. VI 6) : eum Graeci vocant eVirepov, nostri Vesperuginem. 
 
 B That Nonius has taken the quotation from Varro is doubtful. See my 
 ' Nonius Marcellus,' pp. 16, 17. 
 
 ■^ The manuscript tradition of L.L. V n8— VI 61 is much weaker than that 
 of the rest of the work. We may therefore ignore the variants in VI 7 
 videbimus, conticinnio (-mio). 
 
 ' Nonius quotes the line rightly : piptdo te hie differam. Varro, who does 
 not give the whole line, seems to be quoting carelessly, being concerned more 
 with the phrase pipulo differre aliquem than with the actual words of Plautus' 
 verse. In P, or at least, our extant MSS. of this family, the omission of te 
 must be a mere scribal error, for the pronoun is required by the sense. 
 
6 The Ancient Editions of Plauttis. 
 
 Aul. 526 (L.L. V 181, s.v. Aes militare) cedit miles, aes petit 
 (P, A n.l.). 
 
 Cas. 206 (L.L. VII 106, s.v. Delicuum) 
 
 sineamesses sine {leg. sine amet, sine) quod lubet id facias 
 
 {leg. -at) 
 quando tibi domi nihil delicuum est {nihil domi P, A n.l.). 
 
 See above, p. 3. 
 
 Cas. 267 (L.L. VII 104) 
 
 Maccius in Casina a fringuilla : 
 
 quid fringuttis {leg. frig-) ? quid istuc tarn cupide cupis ? (P, 
 A n.l.). 
 
 Cist. I (L.L. VII 98, s.v. Cernere) quia ego antehac te amavi <et 
 mihi amicam esse creui)' {cum ego P, A n.l ). 
 
 Priscian quotes the line with cum, not quia. If the majuscule 
 archetype of our MS. of Varro had qvm, this might easily be 
 mistaken for qvia. 
 
 Cist. 6, 8, lo-ii. (L.L. VII 99, s.v. Frequens) mi frequentem 
 operam dedistis (P, A n.l.). 
 
 Valet assiduam . . . Itaque illud quod eaedem mulierculae dicunt : 
 <[pol isto> quidem nos pretio ptanti {leg. tanti) est frequentare, 
 
 {facile est P, A n.l.) 
 ita in prandio nos lepide ac nitide Accepisti (P, A n.l.), 
 Apparet dicere : facile est curare ut adsimus, cum tam bene nos 
 accipias. 
 
 This passage of Varro raises several questions. In our MSS. of 
 Plautus the words omitted by Varro, tibi utilisque habere, form a 
 separate Une (v. 9). Varro's omission suggests that the text from 
 which his quotation comes had the same Colometry. Again, his 
 use of the word facile in his explanation of v. 8 brings his reading 
 of the line into connexion with the reading of the Palatine MSS. 
 But how precisely ? Is ptanti a mere scribal error for facile ? That 
 seems unlikely. It can hardly be a corruption of anything else 
 than tanti, the preceding having suggested to a scribe optanti. 
 Rather it would seem that tanti est was glossed hy facile est, so that 
 the Palatine MSS. have substituted the gloss for the original phrase. 
 
The Ancient Editions of Flauttts. 7 
 
 On the other hand A phrase like tanti est would better suit a neuter 
 use oi frequentare (not the active use, ' curare ut adsimus '•) and the 
 Passive haberi (which is actually the reading of that corrector of the 
 Codex Vetus of Plautus [Bs] whose readings are generally right). 
 Again it is difficult to scan the line (supposing it to end with 
 frequentare) either with tanti est or with facile est. The requisite 
 metre is one which will serve as transition from the preceding 
 trochaic passage to the following iambic, anapaestic, and bacchiac 
 lines. All these difficulties can, I think, be solved if we suppose 
 that Plautus wrote a trochaic Dimeter acatalectic ending in a Colon 
 Reizianum : 
 
 p61 isto quidem nos prdtio tanti est {sc. ' adesse ') ; facile dst 
 frequentare {i.e. ' curare ut adsimus '), 
 
 and that the proximity of the two phrases tanti est and facile est 
 led to the omission of the one or the other by ancient or me- 
 diaeval scribes. 
 
 Cist. 14-15 (L.L. V 72) Venelia a veniendo ac vento illo quem 
 Plautus dicit : 
 
 quod ibi {leg. ille) dixit qui secundo uento uectus est tranquillo 
 
 mari, 
 uentum gaudeo (P, A n.l.). 
 
 Cist. 405, 407 (L.L. VII 64) In Cistellaria : 
 
 non quasi nunc haec sunt hie limaces liuidae (A ut vid., 
 P n.l.). 
 Limax ab limo, quod ibi vivit. 
 
 diobolares, schoenicolae, miraculae {miracuia A, P n.l.). 
 Diobolares a binis obolis, etc. (see above, p. 3). 
 Here again the words omitted by Varro form a separate line. 
 
 Cure. 236 (L.L. VII 60, s.v. Dividia, ' distractio doloris ') 
 
 Sed quid tibist? Lien enecat, renes dolent, {necat [leg. enecai\ 
 
 P, A n.l.) 
 pulmones distrahuntur. 
 
 Cure. 393 (L.L. VII 71, S.V. Codes) 
 
 de Coclitum (Cocul-) prosapia esse arbitror, {fros. te e. P, A n.l.) 
 
 nam hi sunt unoculi. 
 The faulty omission of te is more prob.ably due to the error 
 
8 The Ancient Editions of Flautus. 
 
 of some scribe (mediaeval or ancient) than to Varro's careless 
 quotation. 
 
 Cure. 474 (L.L. V 146) Secundum Tiberim ad lunium (lanuni 
 7w/-««^.) Forum Piscarium vocant, ideo ait Plautus.: 
 apud piscarium (a.^r^w/. P, A n.l.) 
 
 This passage comes from the well known 'entr' acte,' the 
 genuineness of which is doubtful. 
 
 Cure. 566-8 (ap. Fest. 375 M., s.v. Vapula Papiria). The words 
 of Festus' epitome of Verrius come apparently from Sinnius Capito's 
 treatise on Romari proverbs. Varro, he says, explained vapulare in 
 this phrase to have the sense oiperire, and quoted in support of this 
 explanation a line of Terence and this passage of Plautus : 
 
 reddin an non mulierem (virginem P, A n.l.) 
 priusquam te huic meae machaerae dbicio, mastigia ? 
 uapula ergo {leg. uapulare ego) te vehementer iubeo, ne me 
 territes. 
 
 Epid. 231 (L.L. V 131, s.v. Intusium) intusiatam, patagiatam, 
 caltulum, {leg. -am) ac crocotulam {aut cr. A P). 
 
 Since this quotation comes from that part of the 'Lingua Latina ' 
 for which we have the evidence only of Renaissance MSS., it is 
 quite likely that Varro actually wrote aut and not ac. 
 
 Men. 127 (L.L. VII 93, s.v. Euax) euax, iurgio uxorem tandem 
 abegi a ianua {itc. hercle tand. ux. P. A n.l.) 
 
 The omission of hercle, and the transposition of uxorem and 
 tandem may be mere errors of a scribe. However, those who 
 believe in the possibility of the occasional use in Plautus of the old 
 final d of the Abl. Sing, can appeal to Varro's quotation of this 
 trochaic Septenarius, and of the iambic Senarius in Cure. 393. To 
 me it seems that the evidence of inscriptions, &c., shews that the 
 use of this archaism would be as impossible to Plautus as the 
 disyllabic pronunciation of -tion in 'nation,' 'consummation,' &c., 
 &c. would be to an English comedian to-day. 
 
 Men. 183 (L.L. VII 56, s.v. Ascriptivi) idem istuc aliis ascriptivis 
 fieri ad legionem solet (A P). 
 
The Ancient Editions of Plautus. 9 
 
 Men. .89 (R.R. II, i, 20, s.v. Sacres) quanti\ ^ ,^^^ ^ -^ 
 
 sunt porci sacres ? I ^ . 
 
 (R.R. II IV, 16, S.V. Sacres) ^ sacres^ K^ 
 quanti hic porci sunt sacres ? F 
 
 The double occurrence of the quotation in divergent form is 
 fortunate, for it leaves no doubt that Varro is quoting carelessly 
 from memory. 
 
 Men. 352 (J-/.L. VII 12, s.v. Tu domi videbis, ' domum curabis') 
 intus para, cura, uide quod opus {leg. -ust) fiat. (P, A n.l.). 
 
 Here again Varro's colometry seems to agree with that in our 
 MSS. 
 
 Men. 797 (L.L. VII 54, s.v. Carere) inter ancillas sedere iubeas, 
 lanam carere (carpere [leg. carere\ P, A n.l). 
 
 Men. 1047 (^P- Aul. Gell. XVIII ix 4, s.v. Insecenda) Varronem 
 quoque versum hunc Plauti de Menaechmis : nihilo minus (Jeg. 
 mihi ?) esse videtur {leg. -entur ?) '^ sectius quam somnia {n.e. 
 mihi videntur setius \_-dus\ P, A n.l.), sic enarrare ' nihilo magis 
 narranda esse quam si essent somnia.' 
 
 What is the meaning of Gellius' statement? Does it imply that 
 Varro favoured the spelling sectius, or merely that Varro, in a freak 
 of etymology, chose to connect the insece of Ennius with the Adverb 
 secius ? Probably the latter. 
 
 Merc. 619 (L.L VII 60, s.v. Dividia)' non tibi istuc magis 
 diuidiaest quam mihi hodie fuit {nee [leg. non?'\ t.i.ra. diuidiae est 
 [leg. diuidiaest^ P, A n.l). 
 
 Here again we find thai Varro recognises Hiatus. For unless we 
 discard the patently genuine diuidiaest for diuidiae est, we must scan 
 the second hemistich ^udm mihi hodie fuit. 
 
 Mil. 24 (L.L. VII 86, s.vv. Epityrum, Insane) si {leg. nisi) unum, 
 epityra estuer {leg. estur) insane bene {nisi unum, epityrum [-ram ? 
 -ra ?] estur insanum bene A : nisi unum, epityra ut aput ilia estur 
 insane btne P ut vid.). 
 
 '' Unless Varro read hoc and not haec. 
 
 ' Varro adds : hoc idem est in Corollaria Naevii. So Plautus seems to have 
 taken the line from Naevius, just as Terence has taken lines from Plautus 
 (see below, p. 27). 
 
10 The Ancient Editions of Flautus. 
 
 Again the Hiatus (after the pause following nisi unum, 'but 
 one thing I must say') is recognized by Varro. The insanum 
 of A is Plautine, not the insane of Varro and F. The corrupt 
 setting of the line in P seems due to the insertion of the variant 
 (introduced by vel, misread as ut) aput ilia for epityra. I think the 
 construction epityra (Ace. PI.) estur is genuine (cf. 'Lat. Lang.' 
 viii, 63). 
 
 Most. 245 (L.L. IX 54, s.v. Nihili) uideo enim te jiihili pendere 
 prae Philolacho {Jeg. -che?) omnis homines (enim om. P, A n.l.). 
 
 Whether enim has been wrongly inserted by Varro (or a 
 mediaeval scribe) or wrongly omitted in the Palatine archetype, 
 cannot be determined, for the line scans equally well with or 
 without the word. I am inclined to believe enim to be genuine, 
 and its omission to be due to someone who did not know that 
 Plautus used the word in the sense of enimvero and not of namque. 
 
 Pers. 89 (L.L. VII 55, s.v. Congerro) iam pol ille hie aderit 
 credo congerro meus (P, A n.l.). 
 
 Poen. 530 (L.L. VII 69, s.v. Gralator) vinceretis circum curso 
 (? leg. cervum cursu) vel gralatorem gradum {leg. -du) (P, A n.l.). 
 Or is circum genuine and not a scribe's error for cervum ? It is 
 unfortunate that we have not the Palimpsest for this passage, so 
 as to ascertain whether it read circum or cervum. 
 
 Poen. 1034 (L.L. V 68, s.v. Proserpere) quasi proserpens bestia 
 (AP). 
 
 Pseud. 741 (ap. Non. 551, s.v. Murrina) Varro Anthropopoli : 
 non modo^vinum dare, sed etiam, ut Plautus ait : murrinam, 
 passum, defritum {leg. -ru-?) (AP). 
 
 Pseud. 955 (L.L. VII 81) ut transuersus, non proversus cedit 
 quasi cancer solet ! {non prorsus verum ex transversa cedit quasi , 
 cancer solely P A ut vid.). | 
 
 Varro's version seems to be what Plautus wrote (see above, p. 4). 
 
 Trin. 455-6 (L.L. VII 57, s.v. Ferentarius) nam ilium tibi {cetera 
 interciderunt) {ferentarium A cum Varrone : ferentaneum P). 
 
The Ancient Editions of Plautus. 1 1 
 
 Trin. 886 (L.L. VII 78, s.v. Concubium) concubium sit noctis 
 priusquam postremum perveneris {pr. ad.pos. P, A n.l.). 
 
 Whether the omission of ad is to be ascribed to Plautus or to 
 Varro or to a scribe is not clear. 
 
 True. 22 (L.L. VI 11, s.v. Aetas) non omnis aetas ad 
 perdiscendum est satis {sat est P, A n.l.). 
 
 This quotation occurs in that part of the ' Lingua Latina ' for 
 which we have the evidence only of Renaissance MSS. 
 
 True. 322-3 (L.L. IX 106, s.v. Lavo et Lavor) 
 
 piscis ego credo, qui usque dura vivunt lavant, 
 diu minus lavari quam haec lavat Phronesium (lauere 
 P, A n.l.). 
 
 Varro points out the incongruity of the Deponent lavari with the 
 Active lavant, and hints at the possibility of lavari being a scribe's 
 error for lavare. He selects this line, I fancy, because it was 
 a common topic of grammarians and ' glossographi.' To them the 
 other reading lavere may be due. 
 
 True. 495 (L.L. VII 70, s.v. Praefica) sine uirtute argutum ciuem 
 mihi habeam pro praefica (P, A n.l). 
 
 See above, p. 3. 
 
 What order of the plays was observed by these gloss-writers 
 in their Plautus citations is not distinctly shewn in Varro's excerpts 
 from their glosses. But a batch of glosses from Naevius points 
 to alphabetical"" grouping (L.L. VII 107) : 
 
 Multa apud poetas reliqua esse verba quorum origines possint 
 dici non dubito, ut apud Naevium in Aesiona mucro gladii 
 ' Hngula ' a Hngua ; in Clastidio ' vitulantes ' a vitula ; in Dolo 
 ' caperrata fronte ' a caprae fronte ; in Demetrio ' persibus ' a 
 perite; itaque sub hoc glossema 'callide' subiungunt ; in Lam- 
 padione 'protinam' a protinus, continuitatera significans; in 
 Nagidone ' clucidatus ' suavis, tametsi a magistris accepimus 
 mansuetum ; in Romulo ' sponsus ' contra sponsum rogatus ; in 
 Stigmatia ' praebia ' a praebendo ut sit tutus, quod sint remedia 
 
 '" Not quite what we should call in alphabetical order, but rather in con- 
 secutive letter-sections. 
 
12 The Ancient Editions of Plautus. 
 
 in coUo pueris ; in Technico ' confictant ' a conficto convenire 
 dictum ; in Tarentilla ' lucidum ' a luce illustre ; in Tunicularia 
 ' exbolas aulas quassant,' quae eiciuntur, a Graeco verbo eV^oX^ 
 dictum ; in Bello Punico ' nee satis sardare ' ab serare dictum, 
 id est aperire ; hinc etiam sera qua remota fores panduntur. 
 
 About the text used by Varro himself we can learn nothing, since 
 his Plautus citations occur mainly in the ' de Lingua Latina,' 
 and are there taken from the gloss-writers who drew from the whole 
 field of early literature, Naevius and the other dramatists, as well as 
 Plautus and of Plautus the spurious "' as well as the twenty-one 
 plays of our traditional text. But there is a statement of Aiilus 
 Gellius from which we might draw the inference ° that if a 
 collected edition (or editions) of Plautus was in circulation at 
 Varro's time, it would include only these twenty-one. Gellius tells 
 us that the name ' fabulae Varronianae ' was given to these (N.A. 
 Ill iii. 3) : quas idcirco a ceteris segregavit quoniam dubiosae non 
 erant sed consensu omnium Plauti esse censebantur. Aelius Stilo, 
 Varro's teacher, seems to have favoured a slightly larger list of 
 twenty-five plays, and we hear of other ' indices ' drawn up by 
 various Plautine scholars, Sedigitus, Ser. Claudius, Aurelius Opilius, 
 etc. (Gell. I.e.), the most notable being the ' index ' of Accius, from 
 which an extract p, quoted by Varro in his lost work ' de comoediis 
 Plautinis,' has been preserved. Possibly the twenty-one plays, the 
 accepted plays of Varro's time, were merely those which were 
 common to all these 'indices.' Varro himself, as we learn from 
 Gellius, was inclined to add one or two other plays to the accepted 
 twenty-one. It may be that in this matter, as in so many others, he 
 followed Aelius Stilo. But whether he would have ventured to 
 include these others, if he were asked by a bookseller of the time 
 to furnish an edition of Plautus, we cannot say. There are plenty 
 
 ° Vatro is merely reproducing the diction of these ' glossographi ' in phrases 
 like 'in Astraba Plautina' (L.L. VI 73), ' apud Plautum in Parasite Pigro ' 
 (VII 77) ; so it is not right to infer from his quotation of Cure. 474 (see above, 
 p. 8) that he regarded the ' entr' acte ' in the Curculio as genuine. 
 
 ° For an entirely different inference drawn from the same statement see Leo, 
 ' Plant. Forsch.' p. i8. 
 
 P The true interpretation of the extract has been furnished by Leo. Accius 
 rejected the plays whose title-pages shewed the words: (l) Plauti Gemini 
 Lenones ; Plauti Condalium ; Plauti Anus ; (2) Bis Compressa ; Boeotia ; 
 (3) Titi Macci Agroecus ; Titi Macci Commorientes. 
 
The Ancient Editions of Plautus. 13 
 
 of Shakespearian scholars to-day with a predilection for this or that 
 play outside of the select list included in popular editions, who 
 would not venture to indulge their personal feelings in print. 
 
 § 3. Plautus-citations in Verrius Flaccus. 
 
 Like the ' de Lingua Latina ' of Varro, the 'de Verborum Sig- 
 nificatu ' of Verrius Flaccus, the leading grammarian of the Augustan 
 age and the tutor of Augustus' grandsons, drew its materials (at 
 least in part) from the ' glossarum scriptores.' We may then look 
 to Verrius' encyclopaedia as well as to Varro's linguistic treatise 
 for a sight of the Plautine text of the editors and glossary-writers 
 of the last century of the Republic. In the ' de Lingua Latina ' 
 we found Ser. Clodius, the son-in-law of Varro's teacher, Aelius 
 Stilo, figuring (along with Aurelius Opilius and others) as a source 
 of information about the diction of Plautus. In the ' de Verborum 
 Significatu ' the name of Aurelius Opilius is similarly prominent. 
 
 But the composition of Verrius' encyclopaedia is too complicated 
 and the extant relics too meagre to allow us to follow the same 
 plan as with the ' de Lingua Latina.' We cannot safely collect 
 all the passages of Plautus cited by Verrius and accept them as 
 a fragment of the Plautine text of Republican early editions. In 
 the first place, most of the ' de Verborum Significatu ' is known 
 to us only in the epitome compiled by Paulus Diaconus at the time 
 of Charlemagne ; and it is notorious that the Carolingian editor 
 has woefully mangled the quotations. His epitome was not ex- 
 tracted from Verrius' work itself, but from an epitome made in 
 the reign of Tiberius by Pompeius Festus. Of Festus' epitome 
 we have only a fragment preserved. In respect of arrangement, 
 the encyclopaedia contrasts unfavourably (that is, for our present 
 purpose) with Varro's linguistic treatise. While in the ' de I^ingua 
 Latina ' we have paragraphs which give us a complete, if concise, 
 discussion of a difficult line, we have in our meagre relics of the 
 'de Verborum Significatu ' the same information supplied piecemeal 
 in different parts of the work. Take, as example, the paragraph 
 of Varro (quoted on p. 3) dealing with Cist. 407 : — 
 
 Diobolares, schoenicolae, miraculae. 
 
 "Diobolares a binis obolis. Schoenicolae ab schoeno, nugatorio 
 unguento. Miraculae a miris, id est monstris, a quo Accius ait 
 
14 The Ancient Editions of Plautus. 
 
 personas distortis oribus deformis 'miriones.'" Verrius makes us 
 seek out the same information in different parts of his encyclo- 
 paedia : 
 
 (i) in the D-section (on p. 74 of Mueller's edition) (Paul.') 
 " Diobolares meretrices dicuntur, quae duobus obolis ducuntur." 
 
 (2) in the M-section (p. 123 M.) (Paul.) "Miracula, quae nunc 
 digna admiratione dicimus, antiqui in rebus turpibus utebantur." 
 
 (3) in the S-section (p. 329 M.) (Fest.) " Schoeniculas app 
 <ellare videtw mere> triculas Plautus propter usum un<guenti 
 schoeni> quod est pessimi generis. Itaque (^dixit : Diobolares 
 schoeniculae OT/>raculae, Cum extertis <talis cum todillis crusculis). 
 
 Idem : prosedas pistorum <amicas, yg^J^ak '^^i'^^'^^'^^^i Miseras 
 schoeno &^\Qutas, servilicolas, sordid^zs," 
 
 For the line of the Poenulus (266) just quoted we get our 
 information from "Verrius not in one handy paragraph, as Varro 
 would give it us, but under two different letter-sections ^: 
 
 (i) (p. 226 M.) (Paul.) Prosedas meretrices Plautus appellat, 
 quae ante stabula sedeant. Eaedem et prostibula. 
 
 (2) (p. 7 M.) (Paul.) AHcariae meretrices dicebantur in Cam- 
 pania solitae ante pistrina alicariorum versari quaestus gratia, sicut 
 hae, quae ante stabula sedebant, dicebantur prostibula. 
 
 Similarly, while we have in the 'de Lingua Latina' different 
 versions of a line of Plautus discussed together, e.g. (see above, 
 p. 2), in one and the same paragraph the three variants scauri- 
 pedae of Aurelius Opilius, scripipedae (?) of Juventius, scrupipedae 
 of Valerius, the arrangement in the ' de Verborum Significatu ' is 
 alphabetical and not logical. Here are some examples : 
 
 1 i.e. this part is preserved only in the Carolingian epitome. So we have 
 presumably merely a particle of what Verrius actually wrote. 
 
 "■ When the gaps in our MS. of Festus cannot be supplied with certainty 
 by the help of Paulus Diaconus, I print the conjectural supplement in italics. 
 All that Paulus conserves of this paragraph is : Schoeniculae appellantur mere- 
 trices propter usum unguenti schoeni quod est pessimi generis. 
 
 3 A rare exception is the paragraph compiled from notes on Plaut. Aul. 510 : 
 Flammarii, carinarii, violarii. 
 (p. 89 M.) (Paul.) "Flammeari, infectores flammei coloris. Violari violacii 
 dicuntur. 
 
 It seems to have escaped the process of sorting out the materials for the 
 several letter-sections, although possibly the second item was also entered in 
 the V-section of Verrius and even of Festus. 
 
The Ancient Editions of Plautus. 15 
 
 True. 271 Quia tibi suaso infecisti propudiosa pallulam (with 
 variants suaso and insuaso). 
 
 (r) (p. 302 M.) (Fest. *) "Suasum colos appellatur, qui fit ex 
 stillicidio fumoso in vestimento albo. Plautus : quia tibi suaso 
 infecisti propudiosa pallulam. Quidam autem legunt 'insuaso.' 
 Nee desunt qui dicant omnem colorem qui fiat inficiendo suasum 
 vocari, quod quasi persuadetur in alium ex albo transire." 
 
 (2) (p. Ill M.) (Paul.) "Insuasum appellabant colorem, similem 
 luteo, qui fiebat ex fumoso stillicidio." 
 
 Cist. 408 cum extritis talis, cum todillis crusculis (with variants 
 todillis or -ellis, from iodiis, the name of some thin-legged bird, and 
 crotillis, connected with succrotilla vox, ' a shrill voice '). 
 
 (i) (P- 352 M.) (Fest.) "Todi sunt •(genus avium parvarum) 
 quarum meminit Plautus in Sy<(ra : cum extor)>tis talis cum todillis 
 crus<^culis)> '' 
 
 (2) (p. 301 M.) (Fest.) "Succrotilla tenuis diceba<tur et alta vox>. 
 Titinnius in ... : <(feminina)> fabulare succro^tilla vocula. Afra)'nius 
 in Epistola : . . . <]succro)>tilla voce serio. <Plautus in de^scribendis 
 mulie<(rum cruribus)' gracilibus in Syr<(a : cum extortis)' talis cum 
 sodellis " cr<^usculis. Belle, quae non pedibus valerent." 
 
 (3) (P- 52 M.) (Paul.) "Crotillum, valde exile. Plautus : extortis 
 talis cum crotillis crusculis." 
 
 Epid. 620 sed quis haec est muliercula et ille gravastellus qui 
 uenit ? (with variants gravastellus from ypaCr (?) and ravistelliis from 
 ravus ' grey '). 
 
 (i) (p. 96 M.) (Paul.) "Gravastellus, senior. Plautus : qui est 
 gravastellus qui advenit ? [Ut puto, gravastellus a gravitate dictus]." 
 
 (2) (p. 273 M.) (Fest.) "<(Ravi colons)- appellantur <^qui sunt inter 
 flavos et caesios) . . . ravo sub con . . . quod genus hominum 
 ■(Plautus appellat ravistellos)> : sed quis haec est <^mulier et ille 
 ravistellus, qui)> uenit?" 
 
 ' Festus gives us only an epitome of Verrius, but contrast his fulness with 
 the meagreness of Paulus Diaconus : Suasum colos appellatur qui fit ex stilli- 
 cidio fumoso in vestimento albo. Plautus : ' suaso infecisti propudiosa pal- 
 lulam.' 
 
 " Read todelKs. But it looks as if there had been an omission through 
 Homoeoteleuton, e.g. ^um crotillh crusculis ; quanquam alii dicuniy cum 
 todellis crusculi."!, etc. 
 
1 6 The Ancient Editions of Plautus. 
 
 Mil, 1 1 80 id conexum in umero laeuo, exfafiUato bracchio (with 
 variants exfafiUato and expapillato). 
 
 (i) (p. 79 M.) (Paul.) "Expapillato brachio, exerto, quod quum 
 fit, papilla nudatur." 
 
 (z) (p. 83 M.) (Paul.) " Eflfafilatum, exertum, quod scilicet 
 omnes exerto brachio sint exfilati, id est extra vestimentum filo 
 contextum." 
 
 The Ambrosian Palimpsest seems to supply us with a third 
 variant, expalliolato. In Varro's hands the information would 
 probably have been supplied in the same shape as his account 
 of the three variants in another line (see above, p. 2), scauripedae, 
 scripipedae (?), and scrupipedae. He would have compressed it 
 somewhat in this fashion : Expapillato A. scribit, a papilla ; B. 
 expalliolato, a palliolo; C. exfafiUato, a filo." 
 
 The identity of the materials used by Varro and Verrius is 
 indicated by some of the Plautus-lemmas mentioned above. It 
 is even more striking in some un-Plautine lemmas. Compare, for 
 example, Varro L.L. VII 35 "Apud Ennium : — 
 
 Subulo quondam marinas propter astabat plagas. 
 
 Subulo dictus quod ita dicunt tibicines Tusci ; quocirca radices 
 eius in Etruria non Latio quaerundae," with this lemma in Festus' 
 epitome of Verrius (p. 309 M.) " Subulo Tusce tibicen dicitur. 
 Itaque Ennius : 
 
 Subulo quondam marinas propter adstabat plagas." 
 
 Both Varro and Verrius take both word and citation from the 
 same ' glossographi.' 
 
 The Plautine citations however in the ' de Verborum Significatu' 
 do not all come from these gloss-writers. Not a few can be traced 
 directly to Sinnius Capito's collection of Roman proverbs, a book 
 from which Verrius drew a large amount of material for his encyclo- 
 paedia, e.g., (i) Cas. 524 (p. 310 M.) (Fest.) " SM(trium Quasi Eant 
 M^)>ique in proverbium <!^biit ex hac causdy. Gallico tw'Cfnultu quori) 
 dam edictum est, legiones Sutrii ut praesto essent cum cibo suo. 
 Quod usurpari coeptum est in is, qui suis rebus opibusque offici 
 quid praestarent, quibus deberent. Plautus : sed facitodum, merula 
 per uersus quod cantat colas. Cum suo cuique facito ueniant, quasi 
 eant Sutrium." (2) Epid. 592 (p. 230 M.) (Fest.) Plaustrum 
 
The Ancient Editions of Plautus. 17 
 
 Perculi, etc. Others are patently incidental references to Plautus 
 made by grammarians in the course of their discussion of a difficult 
 line in some other author, or of a difficult or interesting Latin 
 word. 
 
 Now we have already seen in the case of Varro how careless 
 and inaccurate these chance citations may be. The Plautus 
 citations in the ' de Lingua Latina,' taken from glossaries and 
 commentaries immediately connected with a Plautine text, are 
 reliable''; but when Varro, careful scholar though he was, in his 
 work on Husbandry makes casual reference in two places (R.R. II 
 i. 20 and II iv. 16) to a line of Plautus (Men. 289 : quibus hie 
 pretiis porci ueneunt Sacres ?), he is anything but accurate, giving 
 two different, but equally wrong, versions : quanti sunt porci sacres 1 
 and quanti hie porci sunt sacres ? It would be ludicrously absurd 
 for us to take these two versions seriously, as if Varro had 
 transferred them directly from the pages of a copy or copies of 
 Plautus. He is obviously quoting carelessly from memory. But 
 there will be later, in the section dealing with Nonius, a more 
 convenient opportunity of remarking the distinction between 
 quotations taken from a text or from a commentary based on a 
 text on the one hand, and mere casual references to a line of 
 an alien author on the other. Here I need merely say that it is 
 of the utmost importance for us to discriminate, if we can, these 
 two kinds of Plautus-citation in Verrius. The casual citations by 
 memory must not be allowed to suggest to our minds a variety 
 of text which had no real existence. 
 
 Unfortunately it is not so easy in the case of Verrius as in 
 the case of Nonius to decide which Plautus-citations are reliable 
 and which are not. We can without hesitation assign to careless 
 quotation from memory a citation like that of Aul. 555, Argus qui 
 oculeus totus fuit (I quote the whole paragraph) : 
 
 (p. 178 M.) (Fest.) Oculissimum, <^carissi>mum, ostium amicae 
 <^Plautus dixity ; item <(oculissi>me dixit, s\gn\?i(cans carissime, idem 
 iri} Pseudolo oculatum pro prae^senti posuit, cum dixit ; emi)>to die 
 caeca hercule. . . . Idem alibi oculatum Kx^gum dixity. 
 
 . . . Idem : pluris est ocu<latus testis unus quam auriti 
 decern^, id est qui se vidisse di<(cat)>. 
 
 ^ Even this cannot be said without reserve. In V 7 clamide clupeat bracchium 
 is apparently a curtailed version of Pacuvius' line (ap. Non. 87 M. ) : 
 currum liquit, clamide contorta astu clupeat bracchium, 
 C 
 
1 8 The Ancient Editions of Plautus. 
 
 Here it is clear that oculatus, unless it be a scribe's corruption of 
 oculeus, is a mere hazy reminiscence suggested by the other 
 occasions on which Plautus uses this form (cf. p. 21 below). 
 
 Again we may safely attribute to Paulus' inaccurate curtailment 
 the forms assumed by (i) Trin. 721 (video caculam militarem 
 me futurum baud longius), (2) Men. prol. 12 (non atticissat verum 
 sicilicissitat), (3) Rud. 535-6 (Quid si aliquo ad ludos me pro 
 manduco locem ? Quapropter ? Quia pol clare crepito dentibus), 
 (4) Rud. 576 (tegillum eccillud, mihi unum id aret ; id si vis 
 dabo) : — 
 
 (i) (p. 45 M.) (Paul.) Cacula, servus miUtis. Plautus : video 
 caculam militarem '^. 
 
 (2) (p. 28 M.) (Paul.) Atticissat, Attice loquitur. Plautus : ' non 
 atticissat, sed sicilicissitat (-issat MSS.) ; ' id est Sicule loquitur. 
 
 (3) (p. 128 M.) (Paul.) Manduci efifigies in pompa antiquorum 
 inter ceteras ridiculas formidolosasque ire solebat magnis malis 
 ac late dehiscens et ingentem sonitum dentibus faciens, de qua 
 Plautus ait : Quid si ad ludos me pro manduco locem ? Qua- 
 propter ? Clare crepito dentibus. 
 
 (4) (p. 366 M.) (Paul.) Tegillum, cucuUiunculum ex scirpo 
 factum. Plautus : tegillum mihi aret ; id, si vis, dabo. 
 
 These instances will make us hesitate before seriously attributing 
 to an actual divergence of text the following quotations : — 
 
 Aul. 162-4 (s. Postumus, p. 238 M.) (Fest.) : 
 
 Post mediam aetatem qui media ducit uxorem domum, 
 Si earn senex anum praegnatem fortuito fecerit, 
 Quid dubitas quin sit paratum his nomen pueris Postume? {p. 
 nomen puero P, An.l.). 
 
 Aul. 354-5 {s. Temetum, p. 364 M.) (Fest.) : 
 
 Cererin, Strobile, has sunt facturi nuptias ? 
 
 Qui? Quiatemeti nihil allatum uideo ^ (intellego P, A n.L). 
 
 ^ As bad an example of the neglect of the punctuation of the sentence is the 
 citation of Stich. 352 {s. nassiterna, p. 169 M.) (Fest.) : ecquis hue effert nassi- 
 ternam cum aqua sine suffragio ? The senseless curtailment of the quotation 
 should be ascribed to Festus rather than to Verrius Flaccus. 
 
 " Macrobius (Sat. Ill 77) also has video. He may have borrowed from 
 
The Ancient Editions of Plautus, 19 
 
 Mil 213 {s. Comoedice, p. 61 M.) (Paul;) : 
 
 Euge, euscheme adstetisti et dulice et comoedice {etisch. hercle 
 astitit P, A n.l.). 
 
 (On Cure 566-8, see above, p. 8.) 
 
 In nearly all * the other cases the Plautus citations in our relics 
 of Verrius' great work agree with the text of our extant MSS. of 
 Plautus. 
 
 Where Festus' epitome is extant we can often make a guess re- 
 garding the source from which a Plautus-citation has come. Take 
 for example the paragraph on the word naucum (p. 166 M.) : 
 
 Naucura ait Ateius Philologus poni pro nugis. Cincius, quod 
 in oleae . . . nucisque intus sit. Aelius Stilo, omnium rerum 
 putamen. Glossematorum autem scriptores, fabae grani quod 
 haereat in fabulo. Quidam ex Graeco, quod sit vai koX oixi, levem 
 hominem significari. Quidam nucis iugulandis, quam Verrius 
 iugulandam vocat, medium velut dissaepimentum. Plautus in 
 Parasito Pigro : ambo magna laude lauti, postremo ambo sumus 
 non nauci. Item in Mostellaria : quod id esse dicam verbum, 
 nauci, nescio. Et in Truculento : amas hominem non nauci. Et 
 Naevius in Tunicularia : eius noctem nauco ducere. Et Ennius : 
 illic est nugator, nihili, non nauci homo. 
 
 Verrius, although indeed Mueller in his edition of Festus doubts whether the 
 word actually occurred in the Festus MS., of which we have for this part 
 only a Renaissance apograph. 
 
 "■ The chief exceptions are : 
 
 Bacch. 888 (s. Naenia, p. 161 M.) (Fest.) reddam u<bi> (faciam si P, A n.l.). 
 
 Cas. 443 {s. Nepa, p. 164 M.) (Paul. ; cf. Fest.) dabo me ad parietem, 
 imitabor nepam [cedam ad P, A n.l.), and the other lines already mentioned 
 on pp. 17 sqq. 
 
 The quotation of Amph. 275 is interesting' (j. Vesperugo, p. 368 M.) (Paul.) 
 nee Vesperugo nee Vergiliae occidunt, for it exhibits the line with the same 
 hiatus Vergiliae \ occidunt as appears in Varro's citation (p. 5) and in our extant 
 MSS. So is the quotation of Cas. 443, a double setting of which is suggested 
 by a reference to Nonius, p. 145. 
 
 (i) recessim dabo me ad parietem, imitabor nepam (the Plautine version). 
 
 (2) retrorsum cedam ad parietem, imitabor nepam (the ' Revival ' version). 
 
 Nonius has retrorsum cedam ad, but the Palatine MSS. of Plautus (A n.l.) 
 recessim cedam ad, which seems to be a scribe's perversion of : 
 recessim 
 
 retrorsum cedam ad, 
 the suprascript variant having ousted the reading in the text. 
 
 C 2 
 
20 The Ancient Editions of Plautus. 
 
 Sinnius Capito is mentioned in the paragraph on the O. Lat. use 
 of nee (p. 162 M.) in which a quotation from the Mostellaria (here 
 called the ' Phasma ') occurs '' : 
 
 Nee coniunctionem grammatici fere dicunt esse disiunctivam, 
 ut ' nee legit nee scribit,' cum, si diligentius inspiciatur, ut fecit 
 Sinnius Capito, intellegi possit, earn positam esse ab antiquis pro 
 non, ut in XII {sc. Tab.) est: ast ei custos nee escit. Item: si 
 adorat furto, quod nee manifestum erit. Et apud Plautum in 
 Phasmate ( = Most. 240): nee recte si illi dixeris. Et Turpilium 
 in Demetrio : nee recte dici mihi qui iam dudum audio. 
 
 Aurelius Opilius is apparently the source of the quotation of 
 
 Mil. 94 (P- 375 M.) : 
 
 Valgos Opilius Aurelius aliique complures aiunt dici, qui diversas 
 suras habeant. Plautus in Milite Glorioso : maiorem partem 
 uideas ualgis sauiis. Et in Sitellitergo : si ea mihi insignitos 
 pueros pariat postea, Aut uarum aut ualgum aut compernem 
 aut paetum aut bocchum filium. 
 
 Whether we shall ever be able to bring the analysis of Verrius' 
 encyclopaedia to the same degree of certainty as the analysis of 
 Nonius Marcellus' Compendiosa Doctrina, and assign each para- 
 graph and each quotation to a definite source, remains to be 
 seen. For the present we must content ourselves with Reitzen- 
 stein's careful and judicious sketch ■= of its composition. We may 
 
 •> Was it from the same grammatical treatise of Sinnius Capito that the 
 paragraph on the O. Lat. use of super (p. 305 M.) was taken? There too 
 we find a quotation from the Mostellaria referred to ' Phasma.' 
 
 <= Contained in ' Verrianische Forschungen ' (Breslauer Philologische Abhand- 
 lungen I iv), Breslau, 1887. The main points are these. The letter-sections 
 (the A-section, the B-section, &c.), consist nearly always of two parts, of which 
 the first is arranged alphabetically (by A B — or A B C— ), while in the second 
 the various items stand in the order in which they were entered by Verrius 
 from the source from which he took them. The authorities used in the com- 
 pilation of Parts I were apparently : Aelius Gallus ' de Significatione Verborum 
 quae ad lus Civile pertinent ' (a glossary of legal terms), Aelius Stilo's glossary 
 to the Carmina Saliorum, etc. , Santra ' de Verborum Antiquitate ' (a glossary 
 of early Latin), Cincius ' de Verbis Priscis' (a glossaiy of early Latin), Cor- 
 nificius ' de Etymis Deorum ' (an etymological glossary of mythology), Varro, 
 Aurelius Opilius, Ateius Philologus ' Liber Glossematorum ' (a glossary of • 
 early Latin literature), and perhaps Cloatius ' Verba a Graecis tracta ' (a list | 
 of Greek loan-words in Latin). For Part II the authorities were : Ateius Capito I 
 and Antistius Labeo (authors of several legal text -books), Veranius ' de Verbis 
 Pontificalibus ' (the priests' ' vade mecum '), Varro ' Antiquitates Rerum Hu- 
 
The Ancient Editions of Plaufus. i i 
 
 regard as certain one detail of this sketch which is of special 
 importance to us in our investigation. The earlier paragraphs 
 of Parts II often contain series of words which seem to come 
 directly from a Plautine text (or commentary), and which therefore 
 probably follow the order in which they occurred in the plays'". 
 Unfortunately these Plautus-series appear in those portions of 
 the encyclopaedia for which our MS. of Festus' epitome is 
 defective. We have to fall back upon the epitome of Paulus 
 Diaconus, which suppresses, as a rule, the citations of lines, and 
 so reduces and alters the paragraphs that we cannot be sure of their 
 original import. However, it will be worth while to give a list 
 of the Plautus-citations which come from these Plautus-series, 
 and indicate their agreement or disagreement with our extant 
 MSS. : 
 
 manarum ' and ' de Vita Popull Roraani ' (text-books of public and private 
 Roman Antiquities), and perhaps Messalla ' Explanatio Auguriorum ' (an account 
 of the technical terms used in augury). The Parts II begin with lists of words 
 taken from certain early Roman authors, (l) Ennius, (2) C. Gracchus, (3) Cato, 
 (4) Plautus, (s) Naevius. 
 
 The most probable explanation (given by Goetz in Berl. Phil. Woch. 1887, 
 p. 1152) of the different arrangement of Parts I and Parts II is that Verrius 
 drew his materials for Parts I (in which alphabetical arrangement holds) from 
 glossaries which were themselves alphabetically arranged, but for Parts II from 
 (l) texts of early authors (or commentaries upon texts), {2) treatises on law, 
 antiquities, grammar, &c. It should be added that the two parts must not 
 be too rigorously dissociated, for there are traces of items, which properly 
 belong to Parts II, having been worked into Parts I. Some letter-sections 
 have no Parts II. 
 
 ^ The clearest example is in the A-section (p. 28 M.) (Paul.) : 
 (= Aul. 50) " Adaxint, adegerint. 
 
 {= Aul. 555) Argus oculeus. Argus nomen est hominis qui fuisse fingitur 
 
 oculis plenus. 
 (? = Cas. looi) Amasso, amavero. 
 
 (?? = Cist. 115) Amiculum, genus vestimenti, a circumiectu dictum. 
 (? = Mil. 1004) Adlivescit (leg. lib- ?), livere [leg. lib- ?) incipit. [hoc est 
 
 lividum fierit.] 
 (= Men. prol. 12) Atticissat, Attica loquitur. Plautus : non atticissat, sed 
 
 sicilicissitat. Id est, Sicule loquitur. 
 (= Poen. 1290) Atritas (leg. -tus), atri coloris. 
 (= Poen. 1291) Aegyptinos, Aethiopas. 
 (?= Rud. 525) Advelitatio iactatio quaedam verborum figurata ab hastis 
 
 velitaribus. Velites dicuntur expediti milites, quasi vo- 
 
 lantes." 
 
22 The Ancient Editions of Flautus. 
 
 Bacch. 476 (5. Creduas, p. 60 M.) (Paul.) ipsus nee amat nee 
 tu creduas (P, A n.l.)- 
 
 Baech. 1088 {s. Blenni, p. 35 M.) (Paul.) stulti, stolidi, fatui, 
 fungi, bardi, blenni, bueeones (P, A n.l.). 
 
 Cas. 837 [s. Corculum, p. 61 M.) (Paul.) ego sum liber, Meura 
 eorculum, mellieulum, uereulum (P, A n.l.). 
 
 Men. prol. 12 {s. Attieissat, p. 28 M.) (Paul.) non atticissat 
 sed sieilicissitat {uerum P, A n.l.). 
 
 Mil. 213 {s. Comoedice, p. 61 M.) (Paul.) euge euseheme 
 adstetisti et duliee et eomoediee {eus. herde astitit P, A n.l.). 
 
 The two variants are probably due to Paulus, and eannot safely 
 be ascribed to Verrius Flaeeus himself. 
 
 These Plautus-series contain as many quotations from the 
 spurious plays as from the twenty one 'fabulae Varronianae.' That 
 these last were arranged in alphabetical groups (as in our extant 
 MSS.) there is some evidence. Thus in the specimen given above 
 (p. 21) the order apparently is: Aul., Cas., (?) Cist., Mil., Men., 
 Poen., Rud. But what the order was within each alphabetical 
 group, and how the spurious plays were arranged relatively to the 
 others we cannot say. Indeed there are strong indications of 
 a double Plautus-series in Parts II, due, probably, to Verrius' use 
 of two distinct texts or commentaries. But the evidence is far 
 too defective to enable us to pronounce any certain judgment ^ 
 
 Nor, so far as I can see, can we glean any information from 
 the Plautus-citations of Verrius (whether within these Plautus-series 
 
 ^ Reitzenstein for once seems to step outside of his usual limits of caution 
 in his treatment of these Plautus-series. For example he refers the lemma 
 Cavitio (p. 6i M. Paul. : Cavitionem dicebant, quam mode dicimus cautionem) 
 to Bacch. 597, in which line (as in several others) the word cautio occurs, 
 and infers that this lemma should stand after the following lemma Consuetioncm 
 (Consuetionem Plautus pro consuetudine dixit), because the latter clearly belongs 
 to Amph. 490 : et clandestina ut celetur consuetio. But why should not the 
 lemma cavitio have come from a note on, let us say, the form fmiitores (for 
 fautores) used in the prologue of the Amphitruo? And what grounds are there 
 for recognizing three Plautus-series, rather than two, in both of which the 
 spurious plays are represented after the Varronian ? From the form (in Paulus' 
 epitome) of the lemma-word Advelitatio, he infers (p. 62 n.) that Verrius re- 
 garded ad velitationem as one word in Rud. 525 : 
 
 equidem me ad velitationem exerceo. 
 
 But may not Verrius' lemma have been headed Ad velitationem exercere, and 
 may not its form in Paulus' epitome be one of the numerous cases in which 
 the Carolingian monk has mangled the lemma-words of his original ? 
 
The Ancient Editions of Plaiitus. 23 
 
 of Parts II, or in other portions of the book) regarding the arrange- 
 ment of the Cantica in the texts from which the citations come. 
 One instance of a difference in the nomenclature of the plays 
 has already been mentioned. In two paragraphs ^ the Mostellaria 
 is referred to as the ' Phasma,' although in another instance e 
 (p. 166 M.), quoted on p. 19, above, the usual name is given. 
 
 The two references {s. Osculana Pugna, Osculum, p. 197 M., 
 Fest, and s. Nassiterna, p. 169 M., Fest.) to the Stichus as the 
 ' Nervolaria ' have not been satisfactorily explained. But the two 
 references to Cist. 408 (see above, p. 15) by the words Plautus 
 in Syr(ay suggest that ' Syra ' (the name of a ' lena ' in the play?) 
 was actually a second title, just as ' Phasma ' was the title adopted 
 at the revival of the Mostellaria. The words might also, but 
 less naturally, be understood as ' Plautus speaking in the character 
 of Syra.' Cf. Fulgentius Exp. Serm. (s.v. Istega) : sicut Plautus 
 ait in Crisalo {i.e., Bacch. 278, spoken by Chrysalus). 
 
 § 4. Plautus-citations in Nonius Marcellus. 
 
 Nonius Marcellus, who lived in the small town of Thubursicum 
 in North Africa, perhaps in the fourth century, compiled a large 
 Dictionary of Republican Latin, the Compendiosa Doctrina, in 
 20 books, with copious quotations from the older writers. By 
 a singular piece of good fortune this compilation has been left 
 in such a state that we are able to analyse it with minute thorough- 
 ness into its component parts. We can tell with a quite extra- 
 ordinary degree of precision what works of what authors were 
 read and excerpted by Nonius himself, and what quotations he 
 took from marginal commentaries on these authors, or from pre- 
 vious lexicographical writers ; and we can trace almost every quota- 
 tion in his Dictionary to its actual source •". The Republican authors 
 
 1 Both taken from Sinnius Capito ? (see above, p. 20). 
 
 8 The play is cited only four times in all, once without a title. There is 
 hardly any significance in the fact that it is the Mostellaria which is (with the 
 Vidularia) the play least in evidence in Nonius Marcellus' dictionary. Other 
 plays (e.g. Amph., Asin.) seem also to have been very meagrely represented 
 in the encyclopaedia of Verrius Flaccus, although Nonius makes copious use 
 of them (see p. 25 , below). 
 
 •" The details need not be given here. A full analysis will be found in No. I 
 of this series of publications. 
 
24 The Ancient Editions of Plautus. 
 
 used by Nonius himself were presumably those whose works (ii 
 whole or in part) were available for him in the library o 
 Thubursicum or on his own book-shelves. The lexicographica 
 writers from whom he borrowed included certainly Aulus Gellius 
 and possibly Verrius Flaccus (or Festus) ; the names of th( 
 others are unknown to us, for Nonius, unlike Verrius, nevei 
 mentions his authorities. 
 
 This newly attained insight into the construction of the Compen 
 diosa Doctrina has removed a number of misconceptions. I 
 used to be argued that, because the language of Nonius in th( 
 discussion of this or that word resembled the language used bj 
 some other grammarian, both must have drawn from a commoi 
 source. Or again that the absence or paucity of quotations fron 
 this or that author (or a particular portion of his works) impliec 
 that his works (or that particular portion of them) had disappearec 
 from the book-market. In the case of Plautus the great pre 
 ponderance of quotations from the first three plays (Amph. 
 Asin., Aul.) was used as evidence in a way we now see to hav( 
 been unwarranted. Now that we can almost claim to know th( 
 exact source from which Nonius has drawn each item of informa 
 tion, we can decide whether the resemblance between his languagi 
 and that of other grammarians is accidental or not. We can sei 
 that the paucity of quotations from, let us say, certain books o 
 Sisenna's History or of Varro's Treatise on Husbandry was reall; 
 due to the mere accident that Nonius' library (or the library c 
 his town) contained only Sisenna Hist. III-IV and Varro R.R. I 
 And not only are these misconceptions removed in the case o 
 Nonius. We can now gauge the likelihood and unhkelihood c 
 similar inferences that have been made regarding any othe 
 lexicographical and grammatical writer (e.g. Charisius, Diomede 
 Priscian), the construction of whose materials still defies analysis 
 I am afraid that the result must be to expose the weakness c 
 the premises on which inferences regarding these Grammarian 
 have been based by recent writers. Now that we know th 
 facts about Nonius, we can see how erroneous similar inference 
 would be in his case. 
 
 As regards Plautus, Nonius had two copies available '. On 
 
 ' We have therefore no right to assume that the non-Varronian plays ha 
 disappeared from the book-market by Nonius' time. All we can say is thi 
 
The Ancient Editions of Plautus. 25 
 
 contained all the twenty-one ' Varronian ' plays, the other merely the 
 first three. He excerpted his smaller copy (text and commentary) 
 with more thoroughness than the other, and this is the sole and 
 simple reason for the preponderance of quotations from the 
 Amphitruo, Asinaria and Aulularia. There is no evidence (rather 
 the contrary) of any difference of text in the two copies. The 
 number of excerpts that he made in his reading of these plays 
 was so considerable that we get a very satisfactory view of the 
 text used in N. Africa in the fourth (?) century. I have pub- 
 lished elsewhere ^ a full list of them, from which it will be seen 
 that this text, while not identical with either of the two texts in 
 our extant MSS., was by no means radically different, and that the 
 theory of a ' third recension ' receives no support from the Com- 
 pendiosa Doctrina. What gave rise to this theory was the frequent 
 divergence from our MSS. in a number of the Plautus-citations 
 in Nonius. But our analysis reveals this significant fact, that these 
 discrepant citations do not, as a rule, come from Nonius' copies 
 of Plautus, but were taken by Nonius from alien sources, from 
 the marginal commentary on this or that author or from one of 
 of his lexicographical authorities. They are therefore likely to 
 be mere chance citations that never were taken directly from 
 any text, and to have practically no authority. An example is 
 Men. 94 : 
 
 Ita istaec nimis lenta vincla sunt escaria (P, etiam A ut vid.), 
 
 which appears in Nonius (p. 108 M.) as : 
 Ea enim fere lenta vincla sunt escaria. 
 
 He has taken the quotation here from the marginal commentary 
 in his copy of Varro's Treatise on Education. Aul. 116 is quoted 
 rightly (in agreement with our MSS.) at p. 476 M. : 
 
 adsunt, consistunt, copulantur dexteras, 
 
 where we know it to come from Nonius' own copy, but in another 
 shape at 479 M., where it comes from one of his lexicographical 
 
 at Thubursicum Nonius could not (or at least did not) use a copy of them 
 for personal inspection. 
 ^ In ' Philologus ' of this year. 
 
26 The Ancient Editions of Plautus. 
 
 authorities. Cas. 267 is at p. 7 M. cited from Nonius' own copy, 
 and has the same form as it has in our MSS., but the line 
 appears in a different form at p. 308, where Nonius took it from the 
 marginal commentary in his copy of Afranius, to be precise, from a 
 note on Afran. Privign. 245 R. The list need not be extended, 
 Nothing that the newly-found analysis of Nonius' Dictionary has 
 taught us is more instructive than this discrimination between 
 the accurate quotations, agreeing with our extant MSS., which 
 come from an actual text of Plautus, and the inaccurate quotations 
 which come from commentators or careless grammarians. 
 
 That the text used by Nonius in N. Africa in the fourth (?; 
 century was at least a different bookseller's issue from its two 
 contemporaries (?), the Ambrosian Palimpsest {A), which was a N, 
 Italian codex, and the majuscule archetype (^F) of the other MSS. 
 (apparently a codex of Gaul), we may infer from the different ordei 
 of the plays (see below, g 6). For Nonius' copy had this arrange- 
 ment (some details are doubtful) : i Amph., 2 Asin., 3 Aul, 4 
 Bacch., s Cist, 6 Cas., 7 Capt., 8 Cure, 9 Epid., 10-13 the M- 
 plays (Mil. preceding Men.), 14 Pers., 15 Pseud., 16 Poen., 17 Rud., 
 18 Stich., 19 Trin., 20 True, 21 (?) Vid. \ 
 
 But the divergences of text which we are allowed to see arc 
 far too few to support any hypothesis that it belonged to a differenl 
 (a third) recension. Of the first three plays in Nonius' text 
 between two and three hundred lines have been preserved foi 
 us. The cases of real divergence between the ' Nonius ' texi 
 and the Pa recension (or rather the P-text) in all this numbei 
 are so few that they can almost be counted on the fingers of on( 
 hand. The most striking is Aul. 399 : 
 
 Gongrum, murenam, exdorsua quantum potest {exossaia fa, 
 stent P). 
 
 Unluckily, these three plays are not found in the fragmentar 
 Palimpsest, so that we cannot demonstrate conclusively that tin 
 Nonius' text was actually descended from the P* recension a. 
 opposed to the A* recension. But it is certainly unlikely tha 
 the Aa recension would in so large a portion of text be identica 
 with its rival. 
 
 In the remaining plays the more noticeable divergences are : 
 
 1 See Rhein. Mus. LVII. 196. 
 
The Ancient Editions of Plautus. 27 
 
 Epid. 1 88 lam ego me convortam in hirundinem atque eorum 
 exorbebo sanguinem (exsugebo P, A n.l.). 
 
 Mil. EI 80 (see above, p. 16). Nonius quotes expafillato. 
 Pseud. 864 (see below, § 6 (II) ). 
 
 Trin. 252 vestispici (? leg. -ca) {vestispica A, vesiiplicae P). 
 Trin. 839 Cum hisce aerumnis deluctavi {Quibus aer, AP). 
 
 If we consider how soon a variant entered by some owner in the 
 margin would be transferred by a transcriber into the text, we 
 shall see that no stress can be laid on so trifling a number of 
 divergences. We may believe that the text of Nonius was 
 practically identical with the text of the P'^ recension. 
 
 § 5. Other evidence. 
 
 Before taking up the actually extant texts of Plautus, it will 
 be well to gather together the stray references in other authorities 
 than the three we have treated, Varro, Verrius, and Nonius. 
 
 In the Repubhcan period the first writer to be mentioned is 
 Terence ™, who has borrowed two lines of Plautus : 
 
 (i) Eun. 8oi faciam ut huius loci dieique meique semper memi- 
 neris. 
 
 = Capt. 800 faciam ut huius did locique meique semper 
 meminerit (/", A n.l.). 
 (2) Phorm. 976 malum quod isti di deaeque omnes duint. 
 
 = Most. 655 malum quod isti di deaeque omnes duint 
 {AP). 
 
 From Charisius and from Rufinus ' in metra Terentii ' we get 
 a few scraps from Sisenna's commentary on Plautus ; but whether 
 this Sisenna was the historian (cf. Klotz Grimdziige, p. 562) or 
 a' grammarian of the Early Empire (cf. Froehde in Philologus 
 Suppl. XVIII 596) is uncertain. Charisius used Sisenna in com- 
 piling his list of Adverbs, and expressly" ascribes the following 
 information to Sisenna : 
 
 ■" Those two imitations of Plautus are so early that they are worth quoting. 
 But since there is no guarantee that imitations exactly reproduce the 'ipsa 
 verba' of the original, I refrain from quoting those in Lucilius, Varro, &c. 
 
 n Notice that the mention of Sisenna's name is confined by Charisius to these 
 quotations from the Amphitruo; the first of the plays. This implies, I think, 
 
28 The Ancient Editions of Plautus. 
 
 Amph. 165 hoc luci] Quaecumque nomina e littera ablativo 
 singulari terminantur, i littera finita adverbia fiunt, ut 'mani.' 
 
 „ 313 ilium tractim tangam] Pro lente. 
 
 „ 554 tuatim] Ut 'nostratim.' Significat autem tuo more. 
 
 „ 843 examussim est optima] Pro examinato. Amussis 
 autem est tabula rubricata quae dimittitur examintindi operis 
 gratia an rectum opus surgat. 
 
 The notes on Metre and Prosody that have been rescued for 
 us from Sisenna's commentary are provokingly few. His remarks 
 (if correctly given by Rufinus, VI 561 K.) on the O. Lat. genitive 
 malai and the O. Lat. scansion fUi, ficit shew that his views on 
 Plautine scansion were somewhat crude : 
 
 " Malai dialpicns metri causa." 
 
 (ad Rud. 217 °) " Fuit extendit primam syllabam metri gratia." 
 
 But the remark of his subsequently quoted by Rufinus is fairly 
 true, for Plautus loves to present women's excited utterances in 
 Cretic, Bacchiac, and other lyric metres : 
 
 ad Rud. I iv (? v) " Habiliore metro usus est, ut solet in 
 mulierum oratione." 
 
 His observation on Poen. 930 shews us that the Carthaginian 
 passage (vv. 930 sqq.) was written as iambic Senarii, like its Latin 
 version (vv. 950 sqq.), and that a metrical licence was taken with 
 the scansion of the Carthaginian word for ' gods ' : 
 
 " Alonim (haloniura MSS.) Poeni dicunt deum et producenda 
 syllaba metri gratia <]sicut> exigit iambus." 
 
 What he refers to in the other notes is uncertain p, viz. : 
 
 that Sisenna is used, without express acknowledgment, in other Plautine refer- 
 ences in Charisius' (Julius Romanus') list of Adverbs. Certainly the Plautus- 
 citations of Charisius in this list shew a marked contrast to such careless 
 citations by Grammarians as are discussed on p. 34 below. We have then 
 some right to ascribe to Sisenna Charisius' remark about the spuriousness of 
 Bacch. 545, ' in quibusdam non ferunt ' (see below, p. 38), which occurs in 
 this section dealing with Adverbs (s.v. meditate, p. 205 K.). 
 
 ° I take it that v. 217 and not v. 1105 is referred to. See Schoell in his 
 Appendix, p. 149. 
 
 P I fancy that the whole passage in the archetype of our MSS. of Rufinus 
 has suffered from omissions caused by Homoeoteleuton, so that various cita- 
 tions have become attached to the wrong plays. The second extract may really 
 
The Ancient Editions of Plautw;. 29 
 
 (on some scene in the Aulularia) : Haec scaena anapaesto 
 metro est, sed concisa sunt, ut non intellegas. 
 
 (on some line of the Captivi) : hie ' ornatu ' (-tur, -tus) s 
 litteram metri causa amisit. 
 
 (on some line of the Rudens) : 'latronem' producit metri 
 causa. 
 
 Evidently Sisenna did not understand that Plautine scansion 
 represents the pronunciation of Latin at Plautus' time. In the 
 time of Plautus fui, which had hitherto had a long u, was coming 
 to be pronounced with a short u. That is, of course, the reason 
 why Plautus scans, now fui, now fui. Sisenna looks upon the 
 scansion ftii as a mere metrical licence, possible only in ancient 
 times before exact rules of scansion had been imposed. 
 
 We find a similar misapprehension of Plautus' language by 
 the author of the earliest Roman treatise on Rhetoric, generally 
 known as the ' Rhetorica ad Herennium.' The author (his name 
 is unknown) wrote in the first part of the first century B.C. (see 
 Marx in the Preface to the large Teubner edition). 
 
 Speaking of Trin. 23-26 : 
 
 Amicum castigare ob meritam noxiam 
 Immoenest facinus, verum in aetate utile 
 Et conducibile. Nam ego amicum hodie meum 
 Concastigabo pro coramerita noxia, 
 
 he calls this a vitious syllogism, mistaking the Plautine use of nain 
 'for instance' for the causal sense of the Conjunction in his own 
 time : (II 3S) Vitiosa ratio est quae ad expositionem non est ac- 
 commodata vel propter infirmitatem vel propter vanitatem. Infirma 
 ratio est quae non necessario ostendit ita esse quemadmodum 
 expositum est ; velut apud Plautum ' amicum castigare . . . et 
 conducibile.' Haec expositio est; videamus quae ratio adferatur: 
 ' nam ego amicum . . . noxia.' 
 
 Cicero in his youthful work on Rhetoric has borrowed this 
 passage (de Invent. I 95) ; Si non ad id quod instituitur accom- 
 modabitur aliqua pars argumentationis, horum aUquo in vitio 
 
 be: /«V(i.e. 'in this line') ornatus r litteram, etc., the line being some such 
 line asphsttme Smdtus eS. The loss of r in a consonant-group in Early Latin 
 scansion is a favourite topic of the Grammarians (cf. Journ., Phil. XXII 7). 
 
30 The A ncient Editions fff Plautus. 
 
 reperietur . . . aut si ratio alicuius rei reddetur falsa . . . aut si 
 infirma, ut Plautus ' amicum . . . noxia.' 
 
 That Cicero was a reader of Plautus (as of Terence) is clear from 
 such quotations as : 
 
 Trin. 419 Ratio quidem hercle apparet, argentum oixerai. (in 
 
 Pis. 61), 
 Aul. 178 Praesagibat <nni> animus frustra me ire cum exiremi 
 domo (de Div. I 65). 
 
 More are taken from the Trinummus than from any other play. 
 There is no indication of his familiarity with any other than the 
 ' Varronian ' plays ; and even of these only a few, we may guess, 
 were seen on the stage during his lifetime. 
 
 Passing to Imperial times, we may recall Horace's allusions 
 to Plautus as at least a proof that the old Dramatists had still 
 a circle of readers and admirers. And we may infer the same 
 from the contemptuous lines of Persius (I 77) : 
 
 Sunt quos Pacuviusque et verrucosa moretur 
 Antiope, aerumnis cor ' luctificabile ' fulta. 
 
 and Martial (XI 90) : 
 
 Carmina nulla probas molli quae limite currunt, 
 
 Sed quae per salebras altaque saxa cadunt, 
 Et tibi Maeonio quoque carmine maius habetur 
 
 ' Lucili columella hie situst Metrophanes,' 
 Attonitusque legis 'terrai frugiferai,' 
 
 Accius et quidquid Pacuviusque vomunt. 
 
 Quintilian had a somewhat better appreciation of the Early 
 Dramatists, and recommends students of Rhetoric to read them 
 (I viii. 8) : multum veteres etiam Latini conferunt, quamquam 
 plerique plus ingenio quam arte valuerunt, imprimis copiam ver- 
 borum, quorum in tragoediis gravitas, in comoediis elegantia inveniri 
 potest, (cf. X i. 88 — 100). 
 
 But Seneca clearly regards th^m as fit study for the Grammarians 
 of the time but for no one else : 
 
 1 Is this adaptation of the mood to the requirements of classical grammar 
 the work of Cicero or of a scribe ? Nonius (at least the archetype of our MSS. 
 of Nonius) cites Epid. 138 with cum miiierem instead oi aim mititbam. 
 
The Ancient Editions of Plautics. 31 
 
 (cf. de Ira III 37,5; Epist. 108 ; and see Gellius' censure of 
 him for his inappreciative attitude towards the writers of the . 
 Republic, Noct. Att. XII 2.) 
 
 The Grammatical works of the first century a.d. have unfor- 
 tunately been lost to us, and the actual date of some Grammarians 
 who are usually ascribed to this century, and of whose Plautine 
 lore scraps have been preserved (Caper, Velius Longus, Caesellius 
 ' Vindex '^) is far from certain. 
 
 Pliny in his Natural History mentions a great controversy of the 
 learned world over the genuineness of Aul. 400 : " Artoptam 
 Plautus appellat in fabula quam Aululariam scripsit, magna ob 
 id concertatione eruditorum an is versus poetae sit illius." He- 
 quotes Plautus in evidence regarding 'murrina vina' (N.H. XIV 
 13, 92) : lautissima apud priscos vina erant murrae odore condita, 
 ut apparet in Plauti fabulis, quanquam in ea quae Persa inscri- 
 bitur et calamum addi iubet, etc. (the reference is to Pers. 88 : 
 calamum inice). He goes on to mention a different theory based 
 on other lines in Plautus by Fabius Dossenus. And he seems 
 to have cited (with reference to Scaevola, Aelius Stilo, and Ateius 
 Capito) Pseud. 740 in this form : 
 
 Quod si opus est ut dulce promat idem, ecquid habeat ? 
 Rogas ? {Quid si opus sit . . , indidem ecquid habet AP). 
 
 His nephew, the younger Pliny, speaking of a piece of good 
 Latin diction says (Epist. I xvi. 6) : Plautum vel Terentium metro 
 solutum legi credidi. 
 
 In the next century Probus (of Berytus) is the outstanding figure. 
 His editorial labours on the older Republican writers are well 
 known ^ and although his works are lost, we have one or two 
 
 '' I do not know whether we may believe on the authority of Caesellius 
 Vindex that it was a stage-tradition for Pseudolus at Pseud. 235, to imitate 
 the action of a trumpeter. Here is all the evidence (Charisius, p. 239 K.) : 
 ' bat ' sonus ex ore cornicinis lituum eximentis, ut Caesellius Vindex libro B 
 litterae scribit (without any allusion to Plautus' line). The word is also found 
 in Epid. 95. 
 
 s Suetonius tells us how it was at Berytus that he acquired the taste for the 
 neglected Republican literature (de Gramm. 24) : legerat in provincia quosdam 
 veteres libellos apud grammatistam, durante ibi antiquorum memoria necdum 
 omnino abolita sicut Romae. Hos cum diligentius repeteret atque alios deinceps 
 cognoscere ciiperet, quamvis omnes contemni magisque opprobrio legentibus 
 
32 The Ancient Editions of Plautus. 
 
 anecdotes about his teaching related by Aulus Gellius in the Noctes 
 Atticae. Thus at VI vii we learn that the poet Annianus told 
 Gellius that Probus in his presence had read a passage of the 
 Cistellaria with a peculiar accentuation of the word adfatim ; at 
 XIII xxi Gellius gives us on the authority of a friend of Probus 
 that grammarian's pronouncement on the proper spelling of the 
 Ace. Plur. urbis or urbes in this and that passage of Virgil and 
 his appeal to an old copy of Georg. I corrected by Virgil's own 
 hand ; at IV vii Probus' letter to Marcellus is quoted regarding 
 the Old Latin pronunciation Hannibdlis, etc. Everything that 
 we can learn of Probus gives us the impression of the extraordinary 
 pains which he took to recover the 'ipsa verba' of the old writers 
 even in minute details of orthography and pronunciation. 
 
 Gellius is the first of the Grammarians of the Empire whose 
 works have been preserved. From his gossiping chapters we 
 get some informition about Plautus and an interesting glimpse 
 at the careful treatment of the text of the ancient writers by the 
 scholars and publishers of the time. A bookseller is willing to 
 stake any sum on the accuracy of a text of Fabius Pictor's Annales 
 in his shop (V iv. 2 grammaticus autem quispiam de nobilioribus, 
 ab emptore ad spectandos Hbros adhibitus, repperisse se unum 
 in libro mendum dicebat ; sed contra Hbrarius in quodvis pignus 
 vocabat si in una uspiam littera delictum esset). Old copies are 
 hunted up in the libraries of Rome and elsewhere as evidence 
 for the true diction and orthography of early authors. Thus a copy 
 of Claudius Quadrigarius in the library of Tibur is found to have 
 fades (Gen.) in the text (IX xiv) ; a 'liber verae vetustatis ' (Livius 
 Andronicus' Odyssey) in the library of Patrae witnesses to the 
 spdhng insece in the first line (XVIII ix. 5) ; a scholar shews 
 Gellius a copy of the second book of the Aeneid ' mirandae vetus- 
 tatis . . . quem ipsius Vergili fuisse credebatur ' with the spelling 
 aena corrected to ahena (II iii) ; Gellius goes in quest of a copy 
 of a work of Aelius Stilo to the library in the temple of Pax 
 (XVI viii) ; he finds the Dative of dies spelt die in an old copy 
 of Sallust's Jugurtha (summae fidei et reverendae vetustatis libro, 
 IX xiv. zb); an old text of Ennius' Annals, ' liber summae atque 
 
 quam gloriae et fructui esse animadverteret, nihilominus in proposito mansit ; 
 multaque exemplaria contracta emendare ac distinguere et adnotare curavit soli 
 huic nee ulli praeterea grammaticae parti deditus. 
 
Ttie Ancient Editions of Plautus. 33 
 
 reverendae vetustatis,' said to have been corrected by Lampadio 
 himself, is quoted for the true reading quadrupes eques against 
 the corruption quadrupes ecus (XVIII v), and so on. Everything 
 indicates that a microscopically accurate reproduction of the an- 
 cient authors was aimed at, and that anything like re- writing 
 or over-bold editing would not be tolerated. The quotations from 
 Plautus are not numerous enough to give us a clear picture 
 of the text used by Gellius, nor is it always possible to determine 
 whether Gellius is quoting from his own text or is reproducing 
 a quotation as he found it in the pages of some grammatical 
 authority. The spurious plays (of which he says in III iii. 1 1 : 
 feruntur sub Plauti nomine circiter centum atque triginta) are 
 laid under contribution as well as the Vafronian . 
 
 The Noctes Atticae was one of the books used by Nonius in com- 
 piling his Compendiosa Doctrina, .and some of Nonius' Plautus-cita- 
 tions come directly from Gellius and not from Plautus himself; e.g. 
 Poen. 365, with attestation of delicia Sing., a form obscured by 
 the scribes (probably independently of each other) of our two 
 archetypes A and P. 
 
 The other Grammarians whose works remain are either con- 
 temporaries of Nonius or else one, two, or three centuries later, 
 and do not add much to the fuller knowledge which Nonius 
 gives us of the Plautus-text of the time. 
 
 Charisius is an important witness, for, as we have seen, he 
 has preserved certainly a little (like Rufinus) and possibly a 
 good deal of Sisenna's commentary. Of the commentary on the 
 Pseudolus (and other plays ?) by Terentius Scaurus, the famous 
 scholar of Hadrian's time, one item has been preserved by Rufinus 
 (VI 561 K.): 'nunciam': 'iam' divisit in duas syllabas metri causa'. 
 But as a rule these Grammarians of the later Empire are of use 
 only as attesting word-forms or spellings which have so often been 
 obscured in our extant MSS. through the fault of the scribes. Thus 
 
 ' Of course the truth is that Plautus regularly scanned nunciam (like quoniam) 
 as a trisyllable, and perhaps did not recognize the scansion mmc iam (? Epid. 
 135). The confusion of se and sese and transposition are too common to allow 
 us to refer with certainty to ignorance of this usage the ..4 -reading in Poen. 746 : 
 
 Suspendant omnes nunciam sese haruspices [se P), 
 or the P-reading in Mil. 357 : 
 
 Age iam nunc insiste in dolos [nunciam A). 
 D 
 
34 The Ancient Editions of Plautus. 
 
 in Cure, i ted is attested by Charisius and Diomede ; in Aul. 566 
 pellucet by Priscian ; in Aul. 667 Fide (dative) by Charisius. 
 
 The warning given in § 4 against reliance on casual quotations of 
 Plautus by Grammarians and commentators is strikingly exemplified 
 by the quotations of Pseud. 178 : 
 
 Nam nisi mihi penus annuus hodie convenit, eras populo 
 prostituam vos. 
 
 (1) in Servius' commentary on Virgil (ad Aen. I 703): 
 Nisi mihi annuus penus datur, 
 
 (2) by Priscian V p. 170 : 
 
 Nisi mihi annuus penus hie ab amatoribus congeratur, eras 
 populo prostituam vos, o puellae, 
 
 (3) again by Priscian VI p 260 : 
 
 Nisi mihi annuus penus ab amatoribus congeretur ; 
 
 and of Pseud. 1172 : 
 
 An etiam ille umquam expugnavit carcerem, patriam tuam ? 
 
 carelessly cited by Servius (ad. Aen. I 140) in this form : 
 
 Nisi forte carcerem aliquando efifregistis vestram domum. 
 
 Most. 1 103-4 : 
 
 Sic tamen hinc consilium dedero : nimio plus sapio sedens 
 Turn consilia firmiora sunt de divinis loeis 
 
 is quoted carelessly or rather paraphrased by Servius (ad Aen. 
 1X4): 
 
 Sine iuxta aram sedeam, et dabo meliora consilia. 
 
 It is, no doubt, necessary that such chance citations should 
 be admitted to the critical apparatus of the larger editions of 
 Plautus. And yet their presence cannot but be misleading, unless 
 the reader is warned of their worthlessness. It is for the phantom 
 brood of these citations in Nonius, Servius, &e., that a phantom 
 parent has been imagined, the ' third recension.' 
 
 The Grammarians and commentators are most of them as liable 
 as we found Nonius to misapprehend Plautus' meaning. Thus 
 Servius misunderstands the word tenus in Bacch. 793 (ita intendi 
 tenus), and explains it as 'extrema pars arcus.' And they are 
 
The Ancient Editions of Plautiis. 35 
 
 sometimes misled by corruptions in the text which they used. 
 Priscian twice (i p. 97, p. 264 K.) attests veteriorem in Bacch. 
 1 1 50, where Plautus clearly wrote ulteriorem ; 
 
 . Senem ilium tibi dedo ulteriorem, lepide ut lenitum reddas. 
 
 Priscian's copy had VE- instead of VL-. Lactantius quotes Trin. 
 340 with producit instead of prodit. This corruption of the text, 
 due to the substitution of a gloss for the actual word which it 
 was written to explain, appears also in the ' Palatine ' family of 
 MSS., while the Ambrosian Palimpsest has the true reading (which 
 is also attested by Servius) : 
 
 Nam et illud quod dat perdit et illi prodit vitam ad miseriam. 
 
 But the remarks on Metre and Colometry in the treatise of 
 Priscian 'de Metris Terentii' are worth our attention, for the recent 
 discoveries in Egypt are opening our eyes to the folly of dis- 
 regarding even the less important of the ancient authorities on 
 Metre. For example, in the first Canticum of the Amphitruo, 
 where (owing to the want of the Ambrosian Palimpsest) we are 
 quite at fault about the disposition of vv. 161-4, it is not wise 
 to disregard Priscian's colometry : 
 
 Ita peregre adveniens 
 Hospitio publicitus accipiar. 
 (?) Haec eri immodestia coegit me, 
 Qui hoc noctis a portu 
 Ingratiis excitavit, 
 
 even though we can hardly accept his metrical account of the lines. 
 
 His Colometry at True. 120 agrees with the Palimpsest's : 
 
 Pessuma, mane. 
 Optume, odio es. 
 
 § 6. The Ambrosian and Palatine Archetypes. 
 
 We have seen the course taken by the text of Plautus in its his- 
 tory from the lifetime of the author down to the fourth or fifth 
 century of the Empire. There were two main divergent channels, 
 the one the direct tradition of the genuine ' ipsa verba ' of Plautus, 
 the other the transmission of that altered text served up by stage- 
 managers to a later generation of spectators at the time of the 
 
 D 3 
 
36 The Ancient Editions of Plautus 
 
 Plautine Revival. Further, the labour of Grammarians and 
 Antiquarians on the text of the old dramatist, useful as it was 
 in restoring the true ancient forms and phrases (for it is to them 
 that we must ascribe the preservation of what we may call the 
 ' genuine ' text), often resulted in the creation of new variants. 
 Thus the net result of their researches into fr. loo (see above, 
 p. 2) seems to have been the production of three rival readings, 
 scrupipedae, scauripedae, scripipedae (?) ; and apparently in the 
 same way in Mil. 1180, exfafiUato, expapillato and expalliolato. 
 
 The conditions under which written, as opposed to printed, 
 literature is offered to the public and transmitted by successive 
 copyings would make it impossible for an edition of Plautus to 
 preserve for any considerable time its original features intact. An 
 edition which originally embodied, let us say, the direct tradition of 
 the ' ipsa verba ' of Plautus would sooner or later be contaminated 
 with readings from the ' Revival ' text. These would be entered in 
 the margin, possibly by some owner, possibly by the bookseller, 
 and in course of successive copyings might find their way from 
 the margin into the text. The same thing would happen to 
 such variants as had been proposed by this or that Grammarian 
 or Commentator. However pure a text might be at the outset, 
 it could not fail in course of time to become more or less what 
 is called a ' mixed ' text. 
 
 We could not, therefore, expect to find in one of our two 
 archetypes an exact presentation of the ' genuine ' text, and in the 
 other an exact presentation of the ' Revival ' text. That would 
 be beyond the bounds of possibility. But we may hope to find 
 (and I think we do find) that the one archetype has in the main 
 conserved the one kind of text, while the other archetype retains 
 a great part of the characteristics of the rival text. In examining 
 the texts of our archetypes we must of course be careful to 
 remove, as mere accidental accretions, the corrupt readings which 
 are purely and solely due to the clerical errors of scribes, and which 
 may have attached themselves to our extant MSS. at any period 
 of the text's transmission. Some of them, no doubt, may be 
 very ancient. We have seen that Priscian's copy of Plautus (at 
 least, the copy from which he takes the quotation of this Une) 
 had in Bacch. 1150 VETERIOREM- instead of VLTERIOREM, 
 a mistake due to the common confusion of L with E. But a 
 mistake like this, though it might disfigure the very earliest issue of 
 
The Ancient Editions of Plautus. 37 
 
 his or that edition, does not belong to the text as constituted 
 ly the editor. 
 
 Besides difference of text there is possible a difference of 
 "olometry, i.e., of the division of lines in the Cantica ; also 
 liflFerence in point of the order of the plays and in other points 
 if outward form, such as the Didascaliae, the Arguments, the 
 Scene-headings, the indication of the speakers throughout a 
 icene, &c., &c. 
 
 I propose in this section to enumerate the divergences of the 
 ^.mbrosian and Palatine recensions" (to be styled A* and P'^) in 
 ill these particulars, with the view of determining what kind of 
 mcient editions they were, and above all of ascertaining which 
 )f the two has best preserved the ' ipsa verba ' of Plautus. The 
 numeration will help us to decide some questions that have 
 ately been mooted by scholars, such as (i) whether the two 
 iditions have all along followed two different channels or whether 
 )oth come from one and the same source, namely some ' variorum ' 
 :dition (cf. Seyffert Berl. Phil. Woch. 16, 252), (2) whether our 
 radition of the text has been unbroken from Republican times, 
 )r whether the furthest point that it can reach is an edition 
 nade in the early Empire. 
 
 I may as well indicate at the outset the result to which the 
 oUowing array of evidence seems to me to point. I think that 
 he AA-text is shewn to be mainly a representative of the genuine 
 ipsa verba' of the plays, the P^-text to hold in nucleus most of the 
 Revival ' adaptations, and that the two rivals (or at least the first) 
 ire possibly to be traced back in an unbroken line of tradition to 
 Republican times. 
 
 (I.) Lines inserted or omitted or substituted in the ' Revival' Text. 
 
 Of all the varieties characteristic to rival recensions, insertions 
 ire the most likely to produce 'mixture' of text. For neither 
 I bookseller nor an owner would be likely to resist the tempta- 
 ion of adding in the margin a passage found in another text, 
 .nd so saving his copy from the reproach of being imperfect. As 
 , rule we should expect to find the ' Revival ' text characterized 
 
 " So it will not be necessary to mention any mere orthographical or other 
 linor peculiarities of A (the Ambrosian Palimpsest) and the other MSS. By 
 ' I indicate the archetype of these. 
 
38 The Ancient Editions of Plautus. 
 
 rather by curtailment of the ' Plautinae longae fabulae.' But 
 here and there a manager would probably find that the audience 
 required some lines to be inserted for explaining matters more 
 clearly, or for stimulating the interest at a weak part of the play. 
 
 The following seem to me the more certain examples : 
 
 Asin. 23 sqq. : 
 
 Per Dium Fidium quaeris : iurato mihi 
 Video necesse esse eloqui, quidquid roges. 
 Ita me opstinate adgressu's ut non audeam 
 Profecto percontanti quin promim omnia (P, A n.l.). 
 
 Leo argues convincingly (see his note) that the second couplet 
 was substituted for the first in the ' Revival ' text. 
 
 Asin. 32 sqq.: 
 
 DEM. Quid istuc est ? aut ubi istuc est terrarum loci ? 
 
 LIB. Apud fustitudinas, ferricrepinas insulas, 
 
 Ubi vivos homines mortui incursant boves. 
 
 DEM. Modo pol percepi, Libane, quid istuc sit loci : 
 
 Ubi fit polenta, te fortasse dicere, etc. 
 
 All this passage down to v. 46 was apparently omitted in one 
 text (undoubtedly the ' Revival ' text)^ and for it was substituted 
 this couplet, which appears in P (the evidence of A is wanting) 
 after v. 47 : 
 
 DEM. Quid istuc est ? aut ubi istuc est ? nequeo noscere. 
 LIB. Ubi flent nequam homines qui polentam pinsitant. 
 
 Again (as in Asin. 23 sqq.) /'combines the two versions. 
 
 Similarly in Bacch. 377-8 and 379-81 (P, A n.l.) and possibly 382 
 and 383 (P, A n.l.), and, I think, 392-3 and 394-404 (P, A n.l.). 
 
 Bacch. 540-551 cm. A* : hab. Y\ Charisius in his account of 
 Plautine adverbs has occasion to quote a line (v. 545) of this 
 passage, and mentions the fact of omission : ' in quibusdam '^ non 
 
 ^ Similarly the scholiast of Terence remarks (ad Andr. 601) : et sane hi 
 versus (vv. 602-9) desunt, quos multa exemplaria non habent. Wessner (in 
 Berl. Phil. Woch. XXIII 222) shews that the passage contains an un-Terentian 
 construction oi fungi. 
 
The Ancient Editions of Plautus. 39 
 
 ferunt (-tur edd).' We have seen (p. 28) that he (or Julius 
 Romanus, his source) got the statement probably from Sisenna. 
 
 Capt. 1016-1022 am A* : hab. Pa : 
 
 TY. Quid tu ais ? adduxtin ilium huius captivom filium ? 
 PH. Quin, inquam, intus hie est. TY. Fecisti edepol et recte 
 
 et bene. 
 PH. Nunc tibi pater hie est : hie fur est tuos, qui parvom hinc 
 
 te abstulit. 
 TY. At ego hunc grandis grandem natu ob furtum ad 
 
 carnificem dabo. 
 PH. Meritus est. TY. Ergo edepol <merito> meritam 
 
 mercedem dabo. 
 Sed die, oroj pater meus tune es ? HE. Ego sum, gnate 
 
 mi. 
 
 TY. Nunc demum in memoriam redeo, quom mecum recogito. 
 
 Instead of this passage we find in A* a single line (v. J 023), 
 which appears also (at the end of the passage) in P : 
 
 TY. Nunc edepol demum in memoriam regredior, audisse me ; 
 
 and this line is certainly the ' ipsa verba ' of Plautus. The scansion 
 regredior proves it. So A* has preserved the ' genuine ' text, while 
 Pa (or at least P) presents both versions. 
 
 Cist. 126 sqq. is a clear instance of an alternative version, 
 omitted in the Palimpsest, having stood in the margin of the other 
 recension. Here is the whole passage (the quasi-prologue spoken 
 by the Lena) as it appears va. P : 
 
 LE. Idem mihi magnae quod parti est vitium mulierum 
 Quae hunc quaestum facimus : quae ubi saburratae sumus, 
 Largiloquae extemplo sumus, plus loquimur quam sat est. 
 (123) Nam ego illanc olim, quae hinc flens abiit, parvolam 
 Puellam proiectam ex-angiportu sustuli. 
 Adulescens quidam hie est adprime nobilis 
 [Quin ego nunc quia sum onusta mea ex sententia 
 Quiaque adeo me complevi flore Liberi, 
 Magis libera uti lingua conlibitum est mihi, 
 Tacere nequeo misera quod tacito usus est.] 
 Sicyone, summo genere ; ei vivit pater. 
 
40 The Ancient Editions of Plautus. 
 
 Is amore misere banc deperit mulierculam, 
 
 Quae hitic modo flens abiit. Contra amore eum haec 
 
 deperit. 
 Earn meae ego amicae dono huic meretrici dedi, etc. 
 
 The lines enclosed in brackets are omitted by A, and clearly 
 stood in the margin of the other recension, for they have found 
 their way into the text at the wrong place. They ought to precede 
 V. 123. There is no cogent argument for or against their genuine- 
 ness ; but they certainly are quite in the Plautine manner y, and 
 if the play was twice acted in the poet's lifetime we may assign 
 both versions to his pen. 
 
 Epid. 164-5. -^t the end of a Scene in trochaic Septenarii come 
 two lines which in A^ are iambic Octonarii : 
 
 Ibo intro atque adulescenti dicam nostro erili filio 
 Ne hinc foras exambulet neve obviam veniat seni, 
 
 but in Pa are trochaic Septenarii like the rest of the Scene : 
 I, i, abi intro atque adulescenti die iam eriU filio 
 Ne hinc foras ambulet neve usquam obviam veniat seni. 
 
 The assimilation of the metre and the removal of the hiatus 
 {tie I hinc) betray the ' Revival ' text. 
 
 Merc 269 sq. 
 
 Verum hercle simia ilia atque haedus mihi malum 
 Adportant atque eos esse quos dicam hau scio. 
 
 For these two lines of A^ there is a single line substituted in P* : 
 Sed simia ilia atque haedus timeo quid velint. 
 
 I suspect the last of being a stage-copy variety. There is a 
 Plautine ring about the A* version with its eos esse quos dicam hau 
 scio. 
 
 y The line preceding and the three lines following the passage in question 
 interrupt, the narrative. It has been suggested that, although found both in 
 A and P, they are a Plautine insertion designed as a substitute for the second 
 quasi-prologue, spoken by the god Auxilium, in which the story of the young 
 Sicyonian is given. But I do not know that they are not in keeping with the 
 rambhng talk of the 'miiltiloqua et multibiba anus:' 'I picked her out of the 
 gutter — (for all that there's a fine young gentleman head over ears in love with 
 her)— and gave her to a friend of mine to bring up.' 
 
The Ancient Editions of Plautus. 41 
 
 Again at v. 276 we have two versions : 
 
 (i) Ac metuo ne illaec simiae partis ferat (A^}, 
 (2) Atque illius haec nunc simiae partis ferat (P*). 
 
 Merc. 547 
 
 Breve iam relicuom vitae spatiumst : quin ego 
 Voluptate, vino et amore delectavero (A^). 
 
 r^^ offers for the first Hne : 
 
 Decurso spatio breve quod vitae reHcuomst. 
 
 I would assign to Plautus the A^-version with its loose gram- 
 matical structure. 
 
 Merc, 555 Nunc tamen interea ad med hue invisam domum. 
 
 This line is found both in A and P. But in P it is followed 
 by a variant : 
 
 Interea tamen hue intro ad me invisam doraum. 
 
 Can the variant be due to the avoidance of the archaism med ? 
 (See below, on Cas. 137.) 
 
 Mil. 997 ^ <(Er^a mea cuius propter amorem cor nunc miser<[a)>e 
 co<(ntremit)>. 
 
 This line, omitted by P, is added in the top margin of A by 
 the corrector {A =). I see no cogent reason for believing A^ to 
 have drawn on any different (or additional) original than that used 
 by the scribe {A ^) (see the Preface to Studemund's Apograph, 
 p. xxviii). So I take it that this line stood in the margin of the 
 original from which A was copied. 
 
 Most. 409 sqq. (Tranio's soliloquy pending the arrival of the 
 house-key) : 
 
 Homini cui (? leg. qui) nulla in pectore est audacia — 
 Nam cuivis homini vel optumo vel pessunio 
 Quamvis desubito facile est facere nequiter. 
 Verum id videndum est, id viri doctist opus, etc. 
 
 After the third line the Palatine Archetype {A n.l.) offers v. 425 
 which recurs in its proper place. This looks like an indication 
 
42 The Ancient Editions of Plautus. 
 
 of an alternative shortening of the passage^ by omission of vv. 
 412 — 424. 
 
 Most. 817 — 847. The possible omission of the whole passage 
 and the substitution of these two lines : 
 
 Vin qui perductet ? Apage istum perductorem ; non placet. 
 Quidquid est errabo potius quam perductet quispiam, 
 
 were indicated in P* recension or in some archetype of F by 
 the adscription of the couplet in the margin at v. 817. Hence 
 their intrusion into the text of F {A n.l.) at this point. 
 
 Most. 939 sqq. 
 
 Heus vos, pueri,. quid istic agitis ? quid istas aedis frangitis ? 
 
 (940) Heus senex, quid tu percontare ad te quod nihil attinet ? 
 Nihil ad me attinet ? Nisi forte factu's praefectus novos, 
 Qui res alienas procures, quaeras, videas, audias. 
 Non sunt istae aedes ubi statis. Quid ais ? an iam vendidit 
 Aedis Philolaches ? aut quidem iste nos defrustratur senex, 
 
 (945) Vera dico ; sed quid vobis est negoti hie ? Eloquar : 
 
 Erus hie noster potat. Erus hie voster potat ? Ita loquor. 
 
 Vv. 940-5 in this passage (which stands complete in A*) were 
 omitted in P^, clearly a device of the ' Revival ' text for the purpose 
 of shortening the Scene. No subsequent owner in the series of 
 tradition from P^ to F has in this case supplied the defect by 
 a marginal entry. 
 
 Pers. 498. The bacchiac Heptameter : 
 
 Tabellas tene has pelligd. Ha^ quid ad m^? Immo ad te 
 attinent et tua refert, 
 
 is followed in A* by this form of line : 
 
 Nam ex Persia sunt istaec adiatae mi a meo ero. Quando? 
 Hau dudum, 
 
 2 The first three lines may be accepted if we suppose Tratlio to break off 
 abruptly at the end of the first line and re-cast the sentence, or if we change 
 cui into qui ( = qui fit ut). The common theory is that the second line comes 
 from a parallel passage written in the margin of some early archetype. But 
 suspicion attaches to the theory of adscript quotations. Editors have circulated 
 it for far more than it is worth. 
 
The Ancient Editions of Plautus. 43 
 
 but in P by this form : 
 
 Nam e Persia ad me<d> adlatae modo sunt istae<c> a meo 
 domino (? leg. ero), Quijndo? Hau dudum. 
 
 Since the next line is anapaestic : 
 
 Quid istae (-aec ?) narrant ? Percontare ex ipsis : ipsae tibi 
 narrabunt, 
 
 we must, I think, refer the divergence to different metrical settings, 
 probably due to rival commentators. In A'^ the Hne is treated 
 as anapaestic (like the following line), in P^ as a second bacchiac 
 Heptameter. So the Aa editor (or rather the commentator whom 
 he followed) scanned tua as a long monosyllable (less probably two 
 short syllables) ; the P* editor gave tua its usual iambic scansion, 
 and made two words of refert'-^, beginning the second line with 
 fert. 
 
 Poen. 121 sqq. The prologue to the Poenulus had two alter- 
 native endings : 
 
 (i) Is hodie hue veniet reperietque hie filias 
 
 Et hunc sui fratris filium, ut quidem didici ego. 
 Ego ibo, ornabor ; vos aequo animo noscite. 
 
 (2) Hie qui hodie veniet reperiet suas filias 
 Et hunc sui fratris filium. dehinc ceterum 
 Valete, adeste. ibo, alius nunc fieri volo : 
 Quod restat, restant alii qui faciant palam. 
 Valete atque adiuvate ut vos servet Salus. 
 
 The two are written continuously (in this order) m F {A n.L). 
 The second seems to me Plautine. 
 
 Poen. 706 sqq. A possible shortening of this prolix passage 
 seems to have been indicated in the margin both of A^ and of P^. 
 Apparently the shortened version passed from v. 706 to v. 720, and 
 thence directly to v. 730. At least this is what is suggested by 
 the condition of the Ambrosian PaHmpsest, in which vv. 720 and 
 
 ■'' Plautus' phrase (Capt. 296) tua re feceris, 'You will benefit yourself,' 
 suggests that in Early Latin the Ablative (or rather Instrumental) of res was 
 capable of the sense expressed at a later time with the help of the preposition 
 ex. So that refert was two words to Plautus, like magna opere. 
 
44 The Ancient Editions of Plautus. 
 
 730 immediately follow v. 707, while in the Palatine MSS. v. 720 
 (without V. 730) immediately follows v. 706. The Ambrosian 
 fragment unfortunately breaks off at v. 708, and does not resume 
 till V. 746, but there is no reason to doubt that v. 720 and also v. 
 730 recurred in their proper place, as they do in the Palatine MSS. 
 At its proper position v. 720 has in the Palatine MSS. this form : 
 
 LYC. Quin *> sequere me ergo. COLL. Abduc intro : addic- 
 tum tenes ; 
 
 but at its previous occurrence it is unmetrical, both in them : 
 
 LYC. Quin sequere me intro. COLL. Due me ergo intro : 
 addictum tenes, 
 
 and also apparently in the Ambrosian : 
 
 (LYC.) Quin sequere me intro. (COLL.) Due me ergo intro. 
 
 Poen. 930 sqq. The Carthaginian passage has a different form 
 in Aa and in P*. But in the latter the A^-version was also present 
 (in the margin apparently, to judge from its marred and curtailed 
 form in our MSS). 
 
 Poen. 1042 sqq. In the margin of A^ there seems to have been 
 indicated an alternative shortening of all this passage by the sub- 
 stitution of vv. 1042-3 for V. 1044, and the omission of vv. 1045 — 
 1052. Hence in the Palimpsest vv. 1042-3 follow v. 1048, and 
 have appended to them v. 1053, which recurs in its proper place. 
 
 Poen. 1333-5 are omitted in A, which however appends to 
 v. 1332 the opening words of v. 1333, a sign that the PA-variant was 
 added in the margin of A*. It must, however, be allowed that 
 there is a possibility .of the omission being a scribal error due 
 to the Homoeoarcton AGOR. These three lines are identical 
 with three in the ' alter exitus ' (see below), viz. vv. 1382-4. 
 
 Poen. fin. The Poenulus has two endings (1322 — 1371, 1372 — 
 fin.), both of which were exhibited by A and P. The second 
 
 '° The Plautine mannerism of beginning a remark and a rejoinder with tlie 
 same word quin suggests that the shortened version of the passage may be 
 the version of Plautus. For v. 706 begins with quin : 
 
 COLL. Quin hercle accipere tu non mavis quam ego dare. 
 
The Ancient Editions of Plautus 45 
 
 begins (at v. 1372) without any indication in A, but the other MSS. 
 leave a space as for the heading of a new Scene °. Since Plautus 
 as a rule has a trochaic ending for his plays (e.g. Amph.), the first 
 version should possibly be ascribed to the ' Revival ' text. 
 
 Pseud. 372. There were apparently two versions of this line : 
 
 (i) Verum quamquam multa malaque dicta dixistis mihi, 
 
 (2) Sed quamquam multa malaque in me dicta dixistis, tamen, 
 
 which have been ' mixed ' both in A and in P : 
 
 Verum quamquam multa malaque in me dicta dixistis 
 
 tamen (A), 
 Sed quamquam multa malaque in me dicta dixistis -mihi (/"). 
 
 Pseud. 392. This line also (it is preceded by the words tih' nunc 
 dilectum para) appears to have had two versions : 
 
 (i) Ex multis ; ex illis paucis unum qui certust cedo, 
 (2) Ex multis ; exquire ex ilhs unum qui certus siet (P^ ; but 
 P and our MSS. omit ex). 
 
 They have been ' mixed ' in A : 
 
 Ex multis [atque] exquire ex illis paucis unum qui certust 
 cedo, 
 
 where the intrusive at^ue may be either a scribe's inadvertent 
 addition or a wrong expansion of the symbol AL. (denoting a 
 variant reading), unless, indeed, the construction at^ue exquire illis 
 is genuine. 
 
 Pseud. 433. Here, too, is apparent trace of two versions : 
 
 (i) Sed si sint (? sunt) ea vera, ut nunc mos est maxume, 
 (2) Sed si ea vera sunt (P'sint), quae tibi renuntiant, 
 
 which have been ' mixed ' both in A and va. P : 
 
 Sed si sint ea vera ut nunc mos est maxume tibi renuntiant 
 
 Sed si ea vera sunt ut nunc mos est maxume {P). 
 
 " The Palimpsest fragment ceases with v. 1381, so that it is impossible to 
 say wliether it shared a peculiarity of the other MSS. which seems to indicate 
 a possible shortening of the 'alter exitus fabulae. ' In these v. 1377 recurs 
 after v. 1381, apparently by way of indication that vv. 1378-81 might be dis- 
 pensed with. 
 
46 The Ancient Editions of Plautus. 
 
 Some ten lines further on Callipho repeats his appeal : 
 Sintne ilia necne sint quae tibi renuntiant, 
 
 and so some editors claim that this parallel passage was adscript 
 in the original of A and produced the hypermetrical appendix tibi 
 renuntiant. However, the theory of adscript citations, always 
 a doubtful theory, is in this case definitely disproved by the 
 fortunate retention by P of the first half of the alternative version. 
 
 Pseud. 523. In the Palatine MSS. (the Ambrosian is defective in 
 this part) we have two equivalent lines : 
 
 Studeo hercle audire, nam ted ausculto lubens. 
 Age'dum, nam satis libenter te ausculto loqui. 
 
 The second line is proved to be post-Plautine by Abraham 
 (Studia Plautina, p. 182). Can the archaism ted have been its 
 'raison d'etre' (cf Cas. 137, below)? Or is it merely the freak 
 of some stage-copy? 
 
 Stich. 48 sqq. At the close of the lyric passage with which 
 the Stichus begins there are found in P*, but not in A*, ten lines of 
 ordinary iambic metre, which contain a summary of what precedes. 
 There can be little doubt that they were intended as a substi- 
 tute, a dialogue pas sage instead of a version which required musical 
 accompaniment. The simplification of the metre betrays the 
 ' Revival ' text. 
 
 Stich. 156 sq. have in A* this form : 
 
 Neque quisquam melius referet matri gratiam 
 Quam egd meae matri rdfero — invitissimus. 
 
 But there was another version of the second line, known to 
 Charisius (? Julius Romanus), who attests the form /ami. 
 
 Neque rettuht quam ego refero meae matri Fami. 
 
 And the presence of the two versions in some archetype has 
 resulted in the banishment of v. 156 from our Palatine MSS., which 
 present the couplet in this form : 
 
 Quam ego meae matri refero invitissimus. 
 Neque rettulit quam ego refero meae matri Fami. 
 
The Ancient Editions of Plautus. 47 
 
 The first version of the line, with its rude metre and (I think) 
 hiatus ' in pausa,' appears to be the older. 
 
 Trin. 362-8. A possible omission of this passage is indicated by 
 the occurrence in ^ P of v. 369 (in the Palatine MSS. followed by 
 V. 368) immediately after v. 361, instead of at its proper place. 
 The passage bears marks of old Latin (nevolt, apiscatur Passive), 
 and may be ascribed to Plautus. The omission then was a feature 
 of the ' Revival ' text. 
 
 Trin. 788. The two versions are both iricorporated in the text of 
 P {A n.l.) : 
 
 (i) Sed epistulas quando opsignatas adferet, 
 (2) Sed quom opsignatas attulerit epistulas. 
 
 The prosody of the second is abnormal. 
 
 True. 246 sqq. were in ^^ trochaic Septenarii : 
 
 Velut hie est adulescens qui habitat (? /eg. habet) hie agrestis 
 
 rusticus, 
 Nimis mortalis lepidus nimisque probus amator (leg. dator): 
 
 sed is clam patrem 
 Etiam hac nocte ilia (-ac ?) per hortum transiit ** ad nos. eum 
 
 volo 
 Convenire, sed est huic unus servos violentissumus. 
 
 The Une-division of Cantica in our extant minuscule MSS. has of 
 course no traditional authority, but their wording of the passage 
 suggests two bacchiac Tetrameters preceded and followed by Iambic 
 metre : 
 
 Velut hie agrestis est adulescens qui hie habet, 
 Nimis pol mortalis lepidus nimisque probus dator 
 
 Sed is clam patrem etiam hac nocte iliac (-a ?) per hortum 
 Transi[l]vit ad nos. eum volo convenire, 
 Sed est huic unus servos violentissumus. 
 
 The following four lines (which end this polymetric Scene) 
 are iambic Septenarii. At vv. 224-7 there are also traces, not 
 so clear as here, of a different metrical arrangement in A^ and P*. 
 
 d Tte dactyl in the first foot of the hemistich is quite regular. 
 
48 The Ancient Editions of Plautus. 
 
 The inventive faculties of students having been for some time 
 directed to the discovery of alternative versions (dittographiae) 
 in our traditional text, it will be readily understood that a huge 
 number of supposed instances have been unearthed by their united 
 labours. For the purposes of this enquiry it will be well to put in 
 a second list such of these as have the best claims on our con- 
 sideration, and to ignore the rest. 
 
 Asin. 480 sqq. : 
 
 In ius voco te. Non eo. Non is ? memento. Memini. 
 
 Dabitur pol supplicium mihi de tergo vostro. Vae te ! 
 
 Tibi quidem supplicium, carnufex, de nobis detur? (} kg. 
 
 datur ?) Atque etiam 
 Pro dictis uostris maledicis poenae pendentur mi hodie 
 
 (/', A n.L). 
 
 Editors are agreed in regarding these lines as an insertion by 
 some ' Revival ' stage-manager. ' Ut a sententia, metro, sermone 
 alienos damnat Ussingius' is the comment in the large Teubner 
 edition. The ' plebeian ' Latin construction " vae te (cf. Seneca 
 Apocoloc. iv. 3 : vae me ! puto, concacavi me) can hardly be as- 
 cribed to Plautus. 
 
 Asin. Vi opens in P {A n.L) with two iambic Senarii (IV u is 
 composed in this metre), vv. 828-9 : 
 
 Age, decumbamus sis, pater. Ut iusseris, 
 Mi gnate, ita fiet. Pueri, mensam apponite, 
 
 and proceeds in iambic Octonarii to the end of the Scene (v. 850). 
 It has been suggested that these two Senarii are the substitute 
 in the ' Revival ' text for the longer version of the Scene. However 
 we have other examples of a new Scene beginning with a few 
 
 How far Plautus suits his language, his metre, and perhaps his prosody 
 to his characters is a subject that would reward investigation. As a starting 
 point might be taken remarks like those of Sacerdos (p. 433 K.) on True. 259 : 
 quamvis Plautus in Truculento posuerit ' non salveo,' sed inridenter ; nam de 
 persona rustica dixit, and of Priscian de Metris Terentii (p. 425 K.) : Terentius 
 trochaico niixto vel confuso cum iambico ulilur in sermone personarum quibus 
 maxime imperitior hie convenit, quem, puto, ut imitetur, hanc confusionem 
 rythmorum facit (with quotation of Ter. Andr. 481 sqq., Plaut. True. 95 sqq. 
 and a passa:ge of Turpilius). Cf. Tschernjaew ' de serm. Terentii,' 1900. 
 
The Ancient Editions of Plauius. 49 
 
 ntroductory lines in the metre of the previous Scene, before taking 
 ip a new metre of its own. 
 
 Asin. 894 sqq. In the ioth(?) century original of our minuscule 
 VISS. which contain this play, there is a curious disarrangement 
 )f these lines. Between v. 898 and v. 900 stands v. 894 ; between 
 ^ 895 and v. 896 stands v. 899 ; after v. 906 we find v. 895 and 
 ^ 899 recurring in the margin. If all this goes back to ancient 
 imes, it may indicate at least one possible shortening of the Scene. 
 
 Aul. 592-8. The omission of these lines does not injure the 
 ense and relieves the monologue of a certain amount of repetition. 
 But the MSS. give us no clue. The Arabrosian Palimpsest does 
 lot contain this play. 
 
 Bacch. 65 sqq. In the Palatine MSS. {A n.l.) v. 73 occurs after 
 '. 64 as well as at its proper place. It seems to me unlikely 
 hat the passage vv. 65-73 would be either omitted or composed by 
 1 stage-manager. The natural inference from the transposition is 
 hat in the Palatine archetype a page began with v. 65, and the 
 op margin of the page contained a corrected version of v. 73. 
 
 Cas. 185 sqq. seem to me a case in point, although the frag- 
 uentary condition of the Palimpsest at this part obscures the truth. 
 )ne version was, I think : 
 
 CL. Pessumis me modis despicatur domi. 
 
 MY. Hem, quid est ? die idem (nam pol hau satis meo 
 
 Corde accepi querellas tuas) opsecro. 
 
 CL. Vir me habet pessumis despicatam modis 
 
 Nee mihi ius meum optinendi optio est. 
 
 In the other text the first three lines were omitted. Unfortu- 
 lately the fragment preserved by A breaks off after the third line, 
 n F the first line is not found, but in its place stand the fourth and 
 fth ; then follow the second and third, a transposition which I 
 rould ascribe to a marginal indication in the original edition that 
 le first three lines might be omitted. The omission I would 
 scribe to the ' Revival ' text, and would explain it as due to the 
 bsolete construction of the accusative after the passive verb ' {me 
 
 ' In the Palimpsest the unmetrical insertion of vir before pessumis looks 
 ke a scribe's attempt to construe the sentence. 
 
 E 
 
50 TTie Ancient Editions of Plautus. 
 
 despicatur, ' contempt is shewn to me '; cf. Lat. Lang. ch. viii. 
 §63). 
 
 Bacch. 5 1 2-4 (spoken by Mnesilochus) : 
 
 Varum quam ilia umquam de mea pecunia 
 
 Ramenta fiat plumea propensior 
 
 Mendicum malim mendicando vincere. 
 
 This form of the passage is found both in A and P. But in our 
 minuscule MSS., towards the close (after v. 519) of Mnesilochus' 
 tirade, we find a variant form : 
 
 Sed autem quam ilia umquam meis opulentiis 
 
 Ramenta fiat gravior aut propensior 
 
 Mori me malim (? leg. Moriri malim med) excruciatum inopia. 
 
 The repetition (in slightly altered language) of his protestation 
 would not be wholly out of keeping with the circumstances. But 
 the theory of an alternative version (entered, we may suppose, in 
 the bottom margin of F or of some archetype) is very probable. 
 The exaggeration ramenta plumea sounds Plautine, while gravior 
 aut is a weaker, but more strictly logical, substitute s. 
 
 Cist. 510-11, two lines in the conversation of Alcesimarchus with 
 Melaenis, are omitted va. A. It may be a mere error of a scribe. 
 It may also be that they were substituted in the 'Revival' text 
 for vv. 512 — 518 (or even for a longer passage preceding v. 519) 
 with the view of shortening the Scene. The longer version seems 
 to me characteristically Plautine. 
 
 Cure. 455 sqq. The ending of this Scene {P, A n.l.) is so 
 awkwardly abrupt as to suggest that some Revival stage-manager 
 cut it down to its present form. 
 
 Men. 360 stands between a Versus Reizianus and an anapaestic 
 Dimeter. 
 
 The A* recension seems to have offered an iambic line : 
 Nunc eum adibo atque ultro adloquar, 
 but the Pa recension an anapaestic Dimeter : 
 Nunc eum adibo, adloquar ultro, 
 
 e Nonius' citation of the line offers apparently another attempt at amelioration, 
 ilumbea for plumea. 
 
The Ancient Editions of Plautus. 51 
 
 with a scansion adloquar wliich (unlike adloquor) I suspect of being 
 un-Plautine. At v. 364 also there seem to have been variant 
 versions, but A is barely legible. 
 
 Men. 639* is practically identical with v. 645 in P{An.l.). This 
 may indicate a possible shortening of this prohx dialogue by the 
 omission of vv. 640-4, and the adscript verse may have ousted 
 a verse of the text. 
 
 Men. 1028 sqq. This passage is very puzzling indeed. Cer- 
 tainty seems unattainable, and I will merely say in a few words 
 what seems to me the most likely account. In P two versions 
 of the line following v. 1040 stood in the text (i) etiam . . . emisi 
 manu (the version found in A), (2) vel ille . . . emisi manu. Some 
 later scribe accidentally omitted (through Homoeoteleuton) the 
 second version, but a fortunate accident has preserved it for 
 us. That accident was that the scribe of his original had omitted 
 (also through Homoeoteleuton) vv. 1037 — 1043, which were then 
 added by the ' corrector ' in the top margin along with the 
 following line (v. 1044) This last straggled over into the side 
 margin and soon found a place between vv. 1036-7. 
 
 Merc. 619 — 624. This passage can be omitted without injury 
 to the sense. But the same may be said of a good many passages 
 in Plautus. 
 
 Mil. 1273. At the end of this line in P stand the words (found 
 below at the end of v. 1278) : 
 
 iube domum ire. 
 
 Is this a mere accidental transposition or does it indicate a 
 possible shortening of this prolix Scene by the omission of vv. 
 1272-8 ? 
 
 Mil. 1287-9 are omitted in A, perhaps by a mere accident. 
 
 Most. 93 sqq. run thus in P {A n.l.) : 
 
 Atque hoc baud videtur veri simile vobis, 
 
 At ego id faciam esse ita ut credatis. 
 Profecto esse ita ut praedico vera vincam. 
 
 E 2 
 
52 The Ancient Editions of Plautus. 
 
 Atque hoc vosmet ipsi, scio, proinde uti nunc 
 Ego esse autumo, quando dicta audietis 
 [Mea], haud aliter id dicatis. 
 
 Is tliis merely an instance of the prohxity, the 'non astrictus 
 soccus' of Plautus? Or are the first three lines meant as a sub- 
 stitute for the last three ? 
 
 Most. I coo sqq. 
 
 Numquid processit ad forum | hodie novi ? 
 Etiam. Quid tandem ? Vidi efferri mortuom. Hem ! 
 Novom unum : vidi mortuom efiferri foras. 
 Modo eum vixisse aiebant. Vae capiti tuo ! 
 
 The third line is generally believed to be a substitute for the 
 second. But it is not out of keeping with Simo's tantalizing eva- 
 siveness that he should repeat the remark to the impatient Tranio. 
 
 Pers. 666 sq. 
 
 DOR. Toxile, quid ego ? TOX. Di deaeque te agitant irati, 
 
 scelus, 
 Qui banc non properes destinare. DOR. Habeto. TOX. 
 
 Eu praedatu's probe. 
 
 After the last words P offers : 
 
 abi, argentum ecfer hue. 
 
 They may be an alternative ending, substituted with a view to 
 shorten the Scene by the omission of vv. 668 — 72. But it is 
 also possible that the eye of a scribe wandered (as in vv. 691-2) 
 from this line to the following. The illegible state of A at this 
 part does not enable us to decide whether a verse (ending in g,bi 
 argentum ecfer hue) has actually been lost after v. 666. 
 
 Poen. 217 sqq. 
 
 Nam nos usque ab aurora ad hoc quod diei est, 
 Postquam aurora inluxit numquam concessamus, 
 Ex industria ambae numquam concessamus 
 Lavari aut fricari aut tergeri aut ornari. 
 
 The first and third lines seem to offer an alternative couplet 
 to the awkward second line, awkward because it does not contain 
 
The Ancient Editions of Plauius. 53 
 
 a single bacchius. Is the couplet a later improvement upon Plautus 
 (or possibly a 'second thought' of the poet himself), or is the 
 couplet Plautine, and the single line an unskilful piece of shorten- 
 ing by a ' Revival ' stage-manager or (more likely) his faulty sup- 
 plement of a lacuna in his stage-copy ? Or has the scribe of some 
 ' Palatine ' archetype allowed his eye to wander from the true 
 ending (now lost) of v. 2 to the ending of the following hne? 
 It is unfortunate that we have not the evidence of the Palimpsest 
 to help us to a decision. 
 
 Poen. 300 sqq. 
 
 Invidia in me numquam innatast neque malitia, mea soror. 
 Bono med esse ingenio ornatam quam auro multo mavolo : 
 Aurura, id fortuna invenitur, natura ingenium bonum. 
 [Bonam ego quam beatam me esse nimio dici mavolo. 
 Meretricem pudorem gerere magis decet quam purpuram :] 
 (3°S) ^lagisque id meretricem, pudorem quam aurura gerere con- 
 decet. 
 Pulchrum ornatum turpes mores peius caeno conlinunt, 
 Lepidi mores turpem ornatum facile factis comprobant. 
 
 Early Roman Dramatic poets liked to ' ring the changes ' on 
 an idea or a play on words in a way that is more in sympathy with 
 ' Euphuistic ' passages in Shakespeare than with the modern drama* 
 so that we are in danger of suspecting interpolation where only 
 the poet's verbosity is at fault. Still we can hardly avoid regarding 
 some of these lines as alternative versions of others. I am inclined 
 to regard the two which I have bracketed as a later improvement 
 of the original passage. Both the Palimpsest and P contain all 
 the lines, but v. 304 is put in the PaUmpsest after and in P before 
 V. 305, which suggests an intrusion from the margin of A into 
 the text. 
 
 Poen. 390 Huius colustra, huius dulciculus caseus, mastigia 
 is omitted by the scribe of A but inserted (incompletely) by the 
 corrector, who has written above Imius savium of the following 
 line the words dulciculus caseus. This has been interpreted to 
 mean that the corrector {A') used a P-'^text; but there is no 
 satisfactory evidence of A'^ having used any other text than the 
 original of ^^ More probably the line was present (in imperfect 
 form) on the margin of the original. Or again the original may 
 
54 The Ancient Editions of Plautus. 
 
 have had it in the text but in a miswritten shape (due to confusion 
 with V. 391): huius cor, huius studium, huius d. c. m., so that 
 A'- had some justification for omitting it. But the most natural 
 supposition of all is that the line stood in correct form in the text 
 of the original, that its omission is a mere clerical error (due to 
 Homoearcton), and that the corrector did not shew proper care 
 in making his correction. 
 
 Poen. 45 7*"'' are omitted in A. Their omission does not inter, 
 fere with the sense of the passage. 
 
 Poen. Ill i. This Scene (for which the evidence of ^ is lacking 
 until V. 572) seems to drag unreasonably. Leo cleverly suggests 
 that a 'Revival' shortened substitute (vv. 543-6, 567-77) has been 
 incorporated with the original Scene of Plautus. 
 
 Poen. IV ii. fin. The soliloquy 01 Milphio which ends this 
 Scene runs in ^ ^ as follows (vv. 917 sqq.) : 
 
 Illic hinc abiit, di immortales meum erum servatum volunt 
 Et hunc disperditum lenonem : tantum eum instat exiti. 
 Satine prius quam unumst iniectum telum, iam instat alterum? 
 Ibo intro haec ut meo ero memorem, nam hue si ante aedis 
 
 evocem, 
 Quae audivistis modo, nunc si eadem hie iterum iterem 
 
 inscitiast. 
 (922) Ero uni potius inlus ero odio quam hie sim vobis omnibus. 
 Di immortales, quanta clades, quanta adventat calamitas 
 Hodie ad hunc lenonem ! sed ego nunc est quom me com- 
 
 moror. 
 Ita negotium institutumst, non datur cessatio ; 
 Nam et hoc docte consulendum quod modo concreditumst 
 Et illud autera inserviendumst consilium vernaculum. 
 Remora si sit, qui malam rem mihi det merito fecerit. 
 Nunc intro ibo : dum erus adveniat a foro, opperiar domi. 
 
 It is hardly conceivable that Milphio should repeat twice the 
 same ideas in somewhat similar language. Besides, the change 
 to iambic metre in v. 922 seems to mark the conclusion of a Scene. 
 Therefore vv. 917—922 make one version, and the following lines 
 another. In the first version Agorastocles is represented as being 
 
Tlie Ancient Editions of Plaiitus. 55 
 
 in the house ; in the second he is said to be in the forum. From 
 vv. 805 sqq. we see that he was really in the house. The jingle 
 in V. 922 en? , . . era seems characteristically Plautine. 
 
 Stich. 161 sqq. The appearance in A of w. 165-6 between vv. 
 i6o-r, instead of at their proper place, if not a mere accidental 
 transposition, indicates an optional shortening of the parasite's 
 garrulity by the omission of w. 161-4. 
 
 Stich. 174 sqq. 
 
 Gelasimo nomen mi indidit parvo puer 
 (Propter pauperiem hoc adeo nomen repperi) 
 Quia inde iam a pausillo puero ridiculus fui, 
 Eo quia paupertas fecit ridiculus forem. 
 
 Editors prune the redundance of this passage, which exhibits 
 the same form both in A and P, by assigning the second line to 
 the ' Revival ' text. 
 
 Stich. 208* (in the famous Auction-scene) is omitted in A, its 
 place being taken by vv. 232-3, which recur in their proper place. 
 Was this an indication of the possibility of shortening the Scene 
 by the omission of vv. 208" — 231 ? Or was it merely that the scribe 
 of some archetype of A, having miswritten the two lines, entered 
 them in correct form in the top margin of the page, with the result 
 that the next transcriber substituted them for the top line ? 
 
 Stich. 253-4 appears in P in this form : 
 
 Quid igitur me volt ? Tritici modios decem 
 Rogare, opinor, te volt Mene ut ab sese petam ? 
 
 TOth an unmetrical second line. In A they are jumbled up : 
 
 Quid igitur me volt, mene ut ab sese petam 
 Tritici modios decem rogare opinor te volt. 
 
 Were there two versions of v. 254 ? 
 
 (i) Rogare, opinor. Mene ut ab sese petam ? 
 (2) Rogare, opinor, te volt. Ut ab sese petam ? 
 
56 The Ancient Editions of Plautus. 
 
 Stich. 483-5 
 
 Quando quidem tu ad me non vis promittere — 
 Sed— quoniam nil processi sat, ego hac ivero, 
 Apertiore magis via ; ita plane loquor. 
 
 Since the last two lines are not found in our Palatine MSS., 
 the theory has been stated that the first line is the ' Revival ' 
 version of the two following. I see no reason for denying the 
 omission to be a mere accident. The unconnected, repetitive style 
 of the passage is exactly in keeping with the occasion ; for the 
 parasite is desperately beating about for a chance of an invita- 
 tion to dinner. 
 
 Stich. 688 may have had two settings : 
 
 (i) Nam hinc quidem hodie poUuctura praeter nos datur 
 nemini (A^ ut vid.), 
 
 (2) Nam hinc quidem hodie praeter nos iactura dabitur ne- 
 mini, 
 
 for the Palatine MSS. offer (2) in a 'mixed' form: 
 
 Nam hinc quidem hodie pollectura praeter nos iactura 
 dabitur nemini. 
 
 But possibly the word iactura is due to some gloss or to some 
 confusion o{ poUuctura with pol iactura or the like. 
 
 Trin. 72 sqq. 
 
 Nam si in te aegrotant artes antiquae tuae, 
 Sin immutare vis ingenium moribus. 
 Aut si demutant mores ingenium tuom, 
 Neque eos antiquos servas ast captas novos, 
 Omnibus amicis morbum tu incuties gravem. 
 
 Editors suppose the second line to be an 'adscript parallel 
 passage,' and the third and fourth to be an alternative version of 
 the first. 
 
 Trin. 901 (in the long dialogue between Charmides and the 
 Sycophant) : 
 
 Ubi ipse erat ? Bene rem gerebat. Ergo ubi ? In Seleucia 
 
The Ancient Editions of Plautus. 57 
 
 The question is repeated in v. 928 (with a different answer) : 
 
 Sed ipsa ubi est ? Pol ilium reliqui ad Rhadamantem in 
 Cercopio. 
 
 Were it not that this is a long Scene, Plautus (or rather Char- 
 mides) might be allowed to repeat himself unchallenged, and the 
 inconsistency of the lying Sycophant would pass for a humorous 
 touch. But editors seize the opportunity of discovering an indi- 
 cation of a ' Revival' curtailment of the Scene, and declare v. 901 
 to be a substitute for v. 928. The only curtailment, however, that 
 could be indicated by this position of the two alternative versions 
 would be the omission of the intervening lines, which, unfortunately 
 for the theory, can hardly be dispensed with. On the other hand 
 there is a third feature of the MSS. in this Scene that must be 
 taken into consideration. They offer after v. 937 three lines which 
 editors put after v. 888, supposing them to have been omitted in 
 the text and entered in the margin of some archetype or other. 
 This archetype must have contained more than the 20 (19, 21) 
 lines indicated by various holes in the pages (see below, p. 79). 
 
 It will be noticed how prominently one or two plays (notably 
 the Poenulus) figure in the above hsts. Of course in the case of 
 plays and portions of plays which have been lost from the extant 
 Palimpsest fragment we have not the same facilities for detecting 
 alternative versions, and that is partly the reason of the prominence 
 of the Poenulus. But it stands to reason that the vicissitudes of 
 each play must have been different. Some might never be re-staged 
 at all. Others, found by ' Revival ' managers to be popular favour- 
 ites, would be brought out again and again. Some would be offered 
 to die public with little "^ or no alteration of their original setting, 
 while a few {e.g. the Poenulus) might he greatiy transformed. 
 
 (II.) Phrases or words altered m the ' RevivaP Text. 
 The following seem to me the more certain examples : — 
 Bacch. 506 Ego faxo hau dicet nactam quem deluderet (A : 
 derideat P). 
 
 ^ So we need not have the uneasy feeling that, since the Poenulus-text con- 
 tains so many ' dittographiae,' there must be a number of others, if we could 
 only detect them, in the other plays too. 
 
58 The Ancient Editions of Plautus. 
 
 This loose construction is, I think, quite in keeping with Plautus' 
 conversational style. The substitution of the more correct derideat 
 I would ascribe to an editor or to a stage manager, who was pos- 
 sibly guided by a later line of the same play (v. 864) : Faxo se 
 haud dicat nactam quem derideat. 
 
 Bacch. 519 Quam si ad sepulcrum mortuo narres logos (Aa : 
 dicat iocum P*). 
 
 The ' ipsa verba ' of Plautus are unmistakably preserved in A*. 
 
 In the preceding line the exact nature of the divergence of A and 
 of P respectively from what Plautus wrote is not clear. 
 
 Turn quom nihilo pluris mihi blandiri refert {leg.-xt.'C) (A), 
 Turn quom mihi nihilo pluris blandiri referet {P). 
 
 Perhaps the change required is from pluris to plure scanned 
 ■blur' (a monosyllable). 
 
 Cas. 137 Sine vero amari te, meus festus dies (A*: Sine a. tea 
 amari Pa). 
 
 The Palatine version undoubtedly improves the line. But if it 
 were the original version, why should any stage-manager of the 
 ' Revival ' period think of altering it to the A^-form ? Can the 
 archaism Z'^^have been the cause? (cf. Pseud. 523 above). 
 
 Cas. 769 Illaec autem armigerum ilico exornant duae (A* : 
 I. a. in cubiculo a. ornant d. P). 
 
 The ^-reading may be due to the mere substitution of a gloss 
 in cubiculo for the O. Lat. ilico ' in (eo) loco.' 
 
 Cist. 84 Opsecutast, gessit moreni oranti morigerae mihi (A*: 
 O. de ea re, g. m. m. m. P). 
 
 Again I suspect the intrusion of a gloss de ea re. 
 
 Cist. 132 contra amore eum haec deperit (A* : perdita est Pa). 
 
 Editors print deperit. But the Palatine version is perhaps 
 Plautine, the other being due either to a wish to remove the 
 obsolete construction or to the intrusion of a gloss. (See, however 
 Berl. Phil. Woch. XVI 285.) 
 
 Epid. 515 There seem to have been two beginnings of this line 
 (i) propera sis, (2) proper a igitur. The latter is the Pa. version 
 
The Ancient Editions of Plautus. 59 
 
 while the Ambrosian Palimpsest offers propera sis igitur, a ' mix- 
 ture ' that points to an original propera " ■ ' The use of sis in 
 peremptory commands is Plautine (cf infra ad Pers. 574). 
 
 Epid. 606 Si inuenio, exitiabilem ego faciam ut <(illic)> hie fiat 
 dies (Aa : /////. hunc ut fiat diem Pa). 
 
 The Palatine version seems Plautine ; and on the other hand 
 the Ambrosian in a similar case of discrepancy in Epid. 235 : 
 
 Haec vocabula auctiones subigunt ut faciant viros {viri P). 
 
 Possibly in one or both instances the divergence does not go 
 back to the ancient recensions, but has arisen in the mere process 
 of transcription. 
 
 Epid. 620 Sed quis haec est muliercula at ille ravistellus qui venit 
 (Aa : gravastellus Pa). 
 
 The variants, as we have seen (p. 15), can be traced to the early 
 commentators and gloss-writers. 
 
 Epid. 733 (the concluding line of the play) : 
 
 Plaudite et valete : lumbos porgite atque exsurgite (Aa : surgite 
 atque extollite Pa). 
 
 i Clearly the AA-version is what Plautus wrote; the other is the 
 
 adaptation in the ' Revival ' text, with avoidance of the obsolete 
 verb porgo (ior porrigo). 
 
 Men. 572 
 
 Ut hoc utimur maxime more moro 
 
 Molesto atque multo (A : molestoque multum P). 
 
 The use of multus in the sense of ' tedious,' ' a bore,' is a feature 
 J of colloquial Latin, that cannot be said to have been obsolete at the 
 time of the Plautine Revival {e.g. Catull. 112, i : Multus homo es). 
 ^- If the Ambrosian gives (as I fancy) the ' ipsa verba,' the other 
 ■ version, if not due to a mere scribal error {e.g. molestoque multo), 
 ^' must come from some later editor or corrector. 
 
 Men. 587 Aut ad populum aut in iure aut apud aedilem rest 
 ;^ (Aa : ad iudicem rest Pa ut vid.). 
 
6o The Ancient Editions of Plauttis. 
 
 The stricter laws of bacchiac metre require a pure bacchius in the 
 last foot when the line ends in a monosyllable and are satisfied by 
 iudicem but not by aedilem. But there are difficulties in the passage. 
 The contracted form rest for res est {res est is the reading of all 
 our MSS.) is not above suspicion. And is it certain that Plautus 
 did not end with a Colon Reizianum {aedilem r'es est) as he has 
 done in the preceding line and in several lines in this Canticum ? 
 
 Merc. 251 Ego enim lugere atque abductam illam aegre pati 
 (A* : atque illam abductam conqueri P*). 
 
 The reason of the discrepancy is not obvious, unless it be the 
 mere result of a transposition of abductam and illam in a stage-copy 
 or elsewhere. 
 
 Merc. 256 Postquam id quod volui transegi, ibi ego conspicor 
 Navem (Aa : atque e. P). 
 
 This use of atque, ' all at once,' ' of a sudden,' belongs to Old 
 Latin (cf. Cell. X 29). Is ibi a gloss ? 
 
 Merc. 271 Sed conticiscam, nam eccum it vicinus foras (Aa; 
 s.c. vicinum eccum exit foras Pa). 
 
 The Palatine version is shewn by Niemeyer to have an un- 
 Plautine construction of eccum. 
 
 Merc. 301 Sed ausimne ego tibi eloqui fideliter (Aa; el. siguid 
 velim Pa). 
 
 Here too the reason is obscure. To me it looks as though some 
 stage-manager had supplied for himself the missing ending of the 
 line in some mutilated copy. 
 
 Merc. 314 
 
 Nam meo quidem animo plane decrepitus senex (A : vetulus d. 
 Pa) 
 
 Tantidemst quasi sit signum pictum in pariete. 
 
 In Epid. 666 we have the phrase vetulus decrepitus (Satine illic 
 homo ludibrio nos vetulos decrepitos duos Habet ?) To me this 
 furnishes an argument that Plautus used the same phrase here. 
 Others use it as proof that the phrase here is an editor's invention 
 or the result of an adscript parallel passage. Is plane a gloss on 
 tantidemst ? 
 
The Ancient Editions of Plautus. 6i 
 
 Merc. 371-2 express the same idea as 389. Leo cleverly suggests 
 that they were a substitute in the ' Revival ' text for vv. 373—389. 
 The evidence of A is wanting. 
 
 Merc. 767 Quid mihi molestu's, Quia novisse me negas (A* : 
 quia me non novisse ais P*). 
 
 The choice here is difficult. 
 
 Mil. 209 Ecce autem aedificat : columnam mento suffigit suo 
 (Aa; suffulsitV). 
 
 ^il- 359 Credo ego istoc exemplo tibi esse pereundum extra 
 portam (A*). 
 
 The reading of the Palatine MSS. appears to be ' mixed ' : tibi 
 esse eundum actutitmst extra portam, which points to an original like 
 this : 
 
 eundum actutumst 
 tibi esse pereundum extra portam. 
 
 Plautus may have used pereo as a translation oiippa in his Greek 
 original, and the suprascript eundum adutimist may have been an 
 early gloss. 
 
 Mil. 472 Quid iam? Quia hanc attingere ausu's mulierem hinc 
 ex proxumo (A* : hicin pr. P). 
 
 The Aa reading is undoubtedly the older. 
 
 Mil. 1 142 Ut lepide deruncinavit militem ! At etiam rogas i 
 A*: parum P*). 
 
 Editors usually give the preference to the P* reading. 
 
 Mil. 1 180 Id conexum in umero laevo, expalliolato bracchio (A* 
 ut vid.: exfafillato P*). 
 
 The Pa reading is undoubtedly what Plautus wrote. The A* 
 reading is a grammarian's emendation. 
 
 Another of the kind, as we saw before (p. 16), was expapillato. 
 
 Pers. 597 Ne temere hanc te emisse dicas me inpulsore aut 
 inlice (A* : suasu atque inpulsu meo Pa). 
 
62 The Ancient Editions of Plautus. 
 
 Which is Plautus' version ? If the play was twice performed 
 in his lifetime, both may be his. 
 
 Poen. 342-3 
 
 Proba mers facile emptorem reperit, tam etsi in occulto 
 
 sitast (A : in abstruse P). 
 Quid ais tu ? quando illi apud me mecum palpas et lallas 
 (A : mec. caput et corpus copulas P*). 
 
 In the first line the occulto oi A may be a gloss. In the second, 
 if we may make the /4-reading metrical by spelling lalas, it is far 
 more suitable to the character of the speaker than the /"-reading. 
 The latter, perhaps, is also unmetrical. At any rate the excision of 
 mecum would greatly improve the metre. (Cf Berl. Phil. Woch. 
 XVI 285.) 
 
 Poen. r332 Bonum virum eccum video redeuntem domum (A*: 
 se recipit d. P*). 
 
 Pseud. 85 
 
 Actum est de me hodie ; sed potes tu mutuam (A* : actwn 
 
 - hodie iam de me est sed potes nunc mu. P) 
 Drachumam dare unam mihi ? 
 
 It is conceivable that a miswriting turn mu. would lead to an 
 emendation nunc mu. 
 
 Pseud. 223 
 
 Reprehendam ego cuncta hercle una opera, nisi quidem tu haec 
 
 omnia (A* : qu. hodie tu omnia Pa) 
 Facis ecfecta quae loquor (A* :facis scelesta haec ut loquor P*). 
 
 To my ear the assonance heix-, haec, ec- sounds Plautine. 
 
 Pseud. 298 Postilla omnes cautiores sunt ne credant alteri (A*; 
 Ab alienis cau. Pa). 
 
 The PA-reading is clearly that of the ' Revival ' text, avoiding the 
 obsolete postilla. 
 
 Pseud 315. Di meliora faxint (Aa : di melius Jaciant V^). 
 
 Faciant may be a gloss (cf. Corp. Gloss. Lat. IV 73, 30 j V 295, 
 25) in Merc. 285, where Pa has faxint, while A ha.s facient (leg! 
 -ant) ■ but not here, for the metre is consulted in the variant. 
 
The Ancient Editions of Plautus. 63 
 
 Pseud. 389. The line seems to have ended in two ways : 
 
 (i) either adduc hominem celeriter, or propera hominem 
 
 ad due celeriter, 
 (2) propera, adduc hominem cito, 
 
 of which the second is the Pa version, while A shews ' mixture ' : 
 propera adduc ho7ninetn celeriter. It is equally unlikely that celerite? 
 is a mere scribe's error for cito, and that it is a gloss. Cito 
 might more readily be a Late Latin gloss on celeriter, for neither 
 celer nor the" Adverb survived in Vulgar Latin. 
 
 Pseud. 397 Quoi neque paratumst quicquam certi consili (A : 
 parata gutta P). 
 
 Alliteration favours the ^-reading, but would still more favour 
 a possible ciccum. 
 
 Pseud. 432 Fors fuat an istaec dicta sint mendacia (A^ : forsitan 
 ea tibi d. P*). 
 
 The Pa version is clearly that of the ' Revival ' text, due to 
 the avoidance of the obsolete fors fuat an. For it is designedly 
 adapted to the metre and so cannot be a gloss. In Plautus' 
 time forsitan was unknown. 
 
 Pseud. 669 Namque ipsa Opportunitas non potuit mi oppor- 
 tunius (Aa : ]\Jam ipsa mi Op. Pa). 
 
 The metre is equally preserved in both versions, so that this is no 
 mere accidental case of transposition by a scribe. The re-writing 
 of the line must be ascribed to a ' Revival ' stage-manager (or to an 
 editor). 
 
 Pseud. 700 nimium est mortalis scitus (A : graphicus P). 
 
 Is the .(^-reading a gloss? 
 
 Pseud. 864 
 
 Si iste ibit, ito, stabit, astato simul ; 
 Si conquiniscet iste, conquiniscito. 
 
 This is the reading of A^ and, I think, of Plautus. But another 
 version ends the second line, not with conquiniscito, but with 
 ceveto simul, an ending awkwardly similar to the ending of the 
 previous line. Nonius quotes this second version from his copy 
 of Plautus. In some original of P the Aa reading had apparently 
 
64 The Ancient Editions of Plmitus. 
 
 been written above ceveto simul, so that we find in P the 
 ' mixed ' reading : 
 
 Si conquiniscet iste, conquiniscito simul, 
 
 which makes the line unmetrical. 
 
 Pseud. 901 Eum promisisse fortiter dixit sibi (A : firmiter P). 
 Is the i'-reading a gloss ? 
 
 Pseud. 997 Propera pellegere epistulam ergo. Id ago ; tacitus 
 sis modo (A* : ag. si taceas modo P*). 
 
 Pseud. 1 1 75 Strenue mehercle ivisti. Quam velis pernix homost 
 (Aa : quamvis pernix hie est homo P). 
 
 Can it be that the mere miswriting of velis as veis might 
 lead to an editor's (or stage-manager's) re-casting the line in the 
 form offered by Z' ? Or was quam velis deliberately avoided ? 
 
 Pseud. 1220 Magnis pedibus. Perdidisti, ut nominavisti pedes 
 (A* : perd. postquam dixisti pedes. P*). 
 
 The P^-reading is more alliterative. 
 
 Pseud. 1294 Di te ament, Pseudole. Hae! I in malam crucera 
 (A: Pseudole ; pfui t (tyin m. c. P). 
 
 Possibly the original difference was merely in the spelling of 
 the interjection {i)fae, {2)foe'^. 
 
 Pseud. 1295 Cur ego adflictor? (-er) ? Quid tu, malum, in os 
 igitur mi ebrius inructas? (A* : mal. ergo in os mi Pa). 
 
 Pseud. 1299 : 
 
 Quae istaec audaciast te sic interdius 
 
 Cum corolla ebrium ingrediri ? Lubet (A : incedere P). 
 
 It is conceivable that incedere is a mere gloss. But the two 
 variants present two perfectly legitimate varieties of the Cretic line. 
 
 ' In Cas. 727 where the Palatine MSS. offer ey ey (for FY FY of their 
 majuscule archetype), the Ambrosian shews something like pol (first and third 
 letters doubtful) preceded by three indecipherable letters. Editors print edepol. 
 %mAyfoefoe is much more likely. Tliis Interjection/^* has the same diphthong 
 as the sethfoeteo (not ' feteo ') ; cf. foedus,foetidus. 
 
The Ancient Editions of Plant us. '65 
 
 Ingrediri makes the line conform to the preceding line ; incedere 
 makes it of the same type as v. 1301 : 
 
 Suavis ructus mihi est : sic sine, Simo. 
 
 So the /'-reading is quite likely to be the ' Revival ' text substitute 
 for the obsolete ingrediri. 
 
 Rud. 1 126 had apparently two endings : 
 
 (i) sine me pro parti loqui, 
 
 (2) sine me pro re mea loqui. 
 For the Palatine MSS. offer a ' mixed ' version : sine vie pro 
 re mea parte loqui. The Palimpsest fragments do not contain 
 this part of the play. 
 
 Stich. 76 
 
 Utrum ego perplexim lacessam oratione ad hune modum 
 Quasi numquam quicquam adeo adsimulem an quasi quid 
 indaudiverim (A* : q. n. q. in eas simulem, quasi nihil ind. 
 
 Eas in se meruisse culpam, an potius temptem leniter 
 An minaciter ? 
 
 One is tempted to ascribe the Tmesis in eas simulem (for 
 
 insimtilem eas) to Plautus, but the PA-reading conflicts with "the 
 
 context. The miswriting adeassitnulem might easily become in 
 eas simulem. 
 
 Stich. 166 Sed matrem parere nequeo, nee quid agam scio 
 (Aa : neq. nescio quomodo P^). 
 
 The reason of alteration is not apparent. Probably some 
 stage-manager, who had a mutilated or defective copy to work 
 on, supplied the missing ending of this line by his own conjecture. 
 
 Stich. 237 Adibo ad hominem. Quis haecst quae mihi advorsum 
 venit ? (Aa : quis haec est quae advorsum it mihi V^). 
 
 If I anv right in making haecst (of. hicst Poen. 1333) the reading 
 of Aa (the Palimpsest offers haec est), is it possible that this 
 unfamiliar contraction was the reason for re-writing the line ? 
 
 Stich. 255 Immo ut a vobis mutuom nobis dares (Aa : al)s te(^ey 
 P-). . 
 
 F 
 
66 The Ancient Editions of Plautus. 
 
 Stich. 262 Malum quidem si otV (A* : tibi di dent P*). 
 
 Or has there been an omission (through Homoeoarcton) of 
 different parts of a couplet in the two MSS.? Or is tibi di dent 
 a gloss? 
 
 Stich. 373 Venit inquam. Tutin ipsus ipsum vidisti? Lubens 
 
 (Aa : tun euin ipsum mdistil Ita ego lubens P*). 
 
 The AA-reading might be ascribed to avoidance of the hiatus 
 vidisti I ita. The P^-reading might have originated from conjectural 
 emendation after the loss (by Haplography) of ipsus, just as two 
 lines below the loss of nimium may have caused the divergence : 
 
 Argenti aurique advexit nimium. Nimis factum bene (A: 
 multum P). • ; 
 
 Of course the conjectural emendations to supply the gap may be 
 the work of a, ' Revival ' stage-manager. But they might also be 
 ascribed to a scribe (or editor) at any subsequent period. The 
 unmetrical re-setting of Pseud. 307 (with detque usque for det det 
 usque) clearly reveals the mediaeval or Late Latin scribe. 
 
 Stich . 390-1 Vidistin virum sororis Pamphilum ? Non. Non 
 adest ? (A* : Pamphilippum. Non. Adest ? P). 
 
 The A^-reading (with Painphilus, a pet-name form of Pamphilip- 
 pus) is certainly Plautine. The P-reading may have originated in 
 consequence of the loss of one non by Haplography. 
 
 Immo venisse eum simitu aiebat ille : ego hue citus (A*": immo \ 
 
 aiebant venisse eum simul sed ego h. c. P). 
 Praecucurri. 
 
 The substitution of simul for sitnitu may be the result of a gloss; 
 and the order of the words in P may be a mere case of trans- 
 position by a scribe. 
 
 Stich. 586 Valuistin bene? Sustentatumst sedulo. Edepol 
 gaudeo (A* : sustentavi P). 
 
 Plautus certainly wrote sustentatumst (as in v. 467). Cf. belOW, 
 True. 369. 
 
 Stich. 631 lamne abierunt ? Gelasime vide, nunc consilio capto 
 opust (Aa : iamne abiisti 1 G. v. quid es capturus consili Pa). 
 
The Ancient Editions of Plautiis. 67 
 
 Which is right ? And what could be the motive for the change ? 
 Or are both Plautine ? 
 
 Trin. 70 Nemost. Quid tu igitur rogitas tene obiurigem (A* : 
 obiurgitem Pa). -• 
 
 The AA-reading is of course Plautine (cf. Ritschl Opusc. II 
 p. 428). Nonius (190 M.) quotes (but not apparently from his 
 own copy) Cist. frag, xii y/ith. purgitans instead oi purigans. 
 
 Trin. 238" Postulat se in plagas conicere : eos petit, eos sectatur 
 
 (A : eos cupit eos consectaiur Pa). 
 
 The sectatur of A may be a mere scribal error (or even a 
 gloss) for consectaiur. But it is possible that there is a difference of 
 metre underlying the discrepancy. The ^-reading perhaps exhibits 
 an anapaestic Dimeter, which ends (as the iambic Dimeter in 
 V. 236 seems to end) with a Colon Reizianum. The PA-reading 
 makes the line purely anapaestic. 
 
 Trin. 255 Haec ego quom cum animo meo reputo (Aa : Quoin 
 ago cum meo an. et recolo Pa). 
 
 The PA-reading involves the total elision of meo, 
 
 Trin. 328 Bene volo <^ego^ illi facere nisi tu non vis. Nempe de 
 tuo ? (A : ego il. fa. si tu non nevis P). 
 
 The_ ^-reading requires an impossible scansion (as a trochee) 
 of nemfe. But whether the ^-reading should be ascribed to the Aa 
 recension is doubtful. It may have arisen from the gloss "^ nonvis 
 for nevis. Si tu non nonvis may have become si tu non vis and 
 ultimately nisi tu non vis. 
 
 Trin. 1064 Si bonus es, obnoxius sum ; sin secus es, faciam 
 ut (uti?) mones {A^ : uti iubes V^. 
 
 It is noticeable how often these apparently causeless equivalent 
 variants occur at the end of a line, the part which is liable to 
 
 ^ The change of nevis, etc., to nonvis, etc., presumably the substitution 
 of the suprascript gloss for the word explained (see below, VIII), is ex- 
 tremely common at all stages of the text tradition. Thus at Epid. 586 nevolt 
 (-ult) is retained by A and by B, but was changed to nonvult in the original 
 olEVJ. 
 
 F 2 
 
68 The Ancient Editions of Plautus. 
 
 te lost either through a defect in the page or through the 
 practice of writing the ' overflow ' portion of a long line in the 
 nearest available blank space. I suspect the agency of an editor 
 who had a defective copy of Plautus to work upon, or rather of 
 a stage-manager with a tattered stage-copy. 
 
 True. 197 Vise illani ; atque opperimlno : iam exibit; nam 
 lavabat (A* : atqiie opperire ihi iam Pa). 
 
 The P.A- reading is that of the ' Revival ' text, a substitute for the 
 obsolete verbal form. 
 
 True. 238 Nosque esse avaras. qui sumus ? quid est quod male 
 agimus tandem ? (A* : qui su. qui {? quae, .? quam) male nos agimus t. 
 
 Pa). 
 
 True. 24s Qui de thensauris integris demum oggerunt (A*: 
 deiiius danunt Pa). 
 
 The older phrase seems to be preserved in Pa. 
 
 True. 289-90 : ^ 
 
 Quia ad fores nostras unguentis uncta es ausa accedere 
 Quiaque bucculas tam belle purpurissatas habes (Aa : quiaque 
 is/as buccas Pa). 
 
 In the first line I think P had uncta es audes accedere, a ' mixture ' 
 of (i) es ausa, (2) audes. 
 
 True. 369 Benene ambulatumst (Aa : ambulasti Pa). 
 Cf. above, Stich. 586. 
 
 True. 374 Plus pollicere quam abs te posco aut (?hau) postulo 
 (Aa : quam ego abs te postulo Pa), 
 
 _ The construction ab aliquo^poscere seems to be unPlautine. The 
 hiatus in the Pa version, ego \ abs, is of doubtful legitimateness. Can 
 the origmal hne have been : 
 
 Plus polliceri quam te posco hau postulo (quam ego posco 
 hau postulo) ? \T 6 i' 
 
The Ancient Editions of Plautus. 69 
 
 True. 375 
 
 Utinam item a principio rei pepercisses (? leg. repersisses) 
 
 meae (A : ut. a pr. rei item parsisses m. P). 
 Ut nunc repercis saviis. 
 
 Possibly a mere scribal error has been the cause of the dis- 
 crepancy. 
 
 The following are more capable of being classed as mere scribal ' 
 errors : 
 Bacch 496. 
 
 (i) Melius multo me quoque una si cum hoc reliqueris {A). 
 
 (2) Melius esset me quoque una si cum illo relinqueres (P). 
 The first seems the Plautine version with Hiatus cum \ hoc. 
 
 Bacch. 948 Is Helenam avexit (A : abduxit P). 
 
 Cas. 636 Face ventum, amabo, pallio. Timeo hoc negoti quid 
 siet (A* : negotium quid est P). 
 
 The bracketed variant violates Plautus' laws of metre. 
 
 Cas. 747 Sed lepide nitideque volo, nil moror barbarico bliteo 
 (Aa: rituY). 
 
 Plautus certainly wrote bliteo. 
 
 Cas. 749 Gladium Casinam intus habere ait qui me atque te 
 interimat (A* : invitat P). 
 
 Of course interimat is the right reading. 
 
 Epid. 621 Hie est danista, haec ilia est autem quam [ego] 
 emi de praeda . Haecinest? (A: quam emi ex pr. P). 
 
 Were there two versions: {i)quam emi de, (2) quatn ego emi 
 ex ? The Plautine phrase seems to be emi de (see Abraham ' Studia 
 Plautina,' p. 201). 
 
 ' The use of the symbol al. (i.e. aUi, aliter, etc.) by a scribe or corrector 
 is not always to be taken literally to imply that a ' varia lectio ' was found in 
 the original or was taken from another text. A mere blunder of a transcriber 
 is often corrected in this fashion, e.g. Bacch. 344, where B has in the text 
 an obvious miswriting, id niihi aut utrum uerim, but in the margin by the 
 same hand al. haud utrum tulim licere (the true form of the sentence). 
 
Q The A ncient Editions of Plautiis. 
 
 Epid. 624 Estne consimilis quasi cum signum pictum pulchre 
 aspexeris (A* : videris P). 
 
 It hardly seems as if a common word like aspexeris would 
 require a gloss videris, or an editor would replace the one word by 
 the other. I would rather ascribe the variety to a stage-copy's 
 inaccuracy or the conjecture of an editor or stage-manager who had 
 a defective Hne to restore. 
 
 Merc. 319 Humanum amarest atque id vi optingit deum. 
 
 This is the A-io'cva. of the line, but in P the second hemistich is : 
 humanum autem ignoscere est. 
 
 It is possible that the passage was originally the same in both 
 recensions, and that its form both in A and in P is due to omission 
 through Homoeoteleuton and Homoeoarcton : 
 
 Humanum amarest, humanum autem ignoscere est, 
 Humanum <(? ego patior) atque id vi optingit deum. 
 
 Merc. 475 Tuos amicus et sodalis et vicinus proxumus (A : simul 
 
 V. Pa). 
 
 Is the ^-reading due to a suprascript gloss ? 
 
 Merc. 757 Scitam hercle opinor concubinam hanc ! Quin abis 
 (A : nan abis P). 
 
 V. 754 ends with non taces? V. 779 ends with quin abis? Is 
 either ending the cause of the discrepancy ? Or is it due to 
 
 a gloss gm non (for quin) so written, . . The gloss cur non 
 
 i 
 
 may have ousted quin both in A and in P at Pseud. 501 (see 
 below, VIII) : 
 
 Cum ea mussitabas. Scibo. Quin dictum est mihi? {cur 
 non d. AP). 
 
 Merc. 765 Non, non te odisse aiebat, sed uxorem suam (A*: 
 nxorejn verum suam P). 
 
 Is the T'-reading due to the Haplography non for non, non ? 
 Mil. 176 quis is homo est.? (A*) 
 
 the^l'reTZf "^h' f' "/'^ '''""''^ ''^"^^ '° be a mixture of 
 the A-readmg and of a. variant quis is erit? 
 
The Ancient Editions 0/ Plautus. 71 
 
 Mil. 599 Unde inimicus ne quis nostri spolia capiat consili. 
 
 This line occurs both in A and P. But in the latter it is followed 
 by a slightly different setting : 
 
 Unde inimicus nequis nostra spolia capiat auribus, 
 a decided deterioration, which must be referred to a ' Revival ' text 
 or an incorrect stage-copy, and not to Plautus himself But it 
 is conceivable that the mere presence of a gloss auribus may 
 have originated the whole of this supposed rival version of the 
 line. 
 
 Mil. 716 Nimis bona ratione nimiumque ad te et tuam vitam 
 sapis (A*: ut vid.: vi. vides P). 
 
 The actual reading of A is not sapis but habes. There is a 
 possibility that this is what Plautus wrote, and that the vides of 
 P is a gloss or a conjectural emendation. 
 
 Mil. 762 Were there two versions (the second written in the 
 bottom margin, the other in the text of F) ? 
 
 (i) Sed procellunt se et procumbunt dimidiati dum appetunt, 
 (2) Sed procumbunt {leg. procellunt sese) in mensam dimi- 
 diati <dum ap^petunt. 
 
 Or is not the whole divergence rather due to the intrusion of 
 a gloss, procumbunt for procellunt sese ? 
 
 Mil. Ill iii ends with the lines (vv. 945-6) : 
 
 Ut accurate et commode hoc quod agendumst exsequamur, 
 Ne quid, ubi miles venerit, titubetur. Tu morare. 
 Our MSS. (A n.l.) exhibit them in the reverse order. Is this a 
 case of two variant Scene-endings or a mere case of transposition ? 
 
 Most 605 Datur faenus mihi. Faenus illie, faenus hie (A : date 
 mihi faenus P). 
 
 Most. 681 Videndumst primum utrum eae velint aut non 
 velint (A : velintne an non v. P). 
 
 Most. 687 
 
 Eugae ! optume eccum aedium dominus foras 
 Simo progreditur intus (A : ipsus P). 
 
72 The Ancie?it Editions of Plautus, 
 
 Most. 714 Tempus nunc est senem hunc adloqui roihi 
 seems to have stood in the margin of some archetype of A in 
 this form : 
 
 Tempus nunc est mihi hunc adloqui senem. 
 
 Most. 761 
 
 Nam sibi laudavisse hasce ait architectonem 
 Nescioquem exaedificatas insanum bene (A* : ex. has sane bene 
 P?). 
 Our existing Palatine MSS. have has sane instead of the Plautine 
 insamim, but whether their archetype P had not INSANE which 
 was misread as lASANE and emended to has sane it is impossible 
 to say. 
 
 Most. 928 faciam ut iubes (A : voles P). 
 
 The miswriting of iicbes as iules would lead to the substitution of 
 voles. Or any scribe might in a moment of absent-mindedness 
 write the one word instead of the other. 
 
 Pers. 20 Mihi quidem tu iam eras mortuos, quia non te visitabam 
 (A : visitavi P). 
 
 Pers. 171 Me quidem iam satis tibi spectatam censebam gsse 
 et meos mores (A : censiieram P). 
 
 Pers. 178 
 
 Amas pol misera : id tuos scatit animus. 
 
 Ego istuc pelagus tibi ut sit faciam (A : placidum P). 
 
 Is the substitution due to a gloss? And which is the Plautine 
 word ? 
 
 Pers. 269 Verberibus caedi iusserit, compedis impingi? vapulet! 
 (A^ : iniponi P). 
 
 Plautus certainly wrote ifuptngi, but the other may be a mere 
 scribe's error, and not the reading of the Pa recension. In Nonius' 
 Compendiosa Doctrina we find a citation of Trin. 720, which 
 Nonius took from some Grammarian or Commentator, and not 
 from his own copy of Plautus. It shews supponi instead of 
 suppingi. 
 
The Ancient Editions of Plautus. 73 
 
 Pers. 321 Sequere hac sis: argentum hie inest, quod mecum 
 dudum orasti (Aa : quod me dudum rogasti P). 
 
 I suspect the agency of a gloss. Rogo is the stock-gloss on oro 
 in dictionaries of the Empire. 
 
 Pers. 436 
 
 Ubi quid credideris citius extemplo a foro 
 
 Fugiunt quam ex porta ludis emissus lepus ( A : cum emissust I. 
 
 n 
 
 Pers. 515-6 lucrifera.-.lucrifera (A: lucrifica...lucrifica P). 
 
 Pers. 574 I sis [in] malum cruciatum. I sane tu — hanc eme, 
 ausculta mihi (A : sis eX tu om. Y) (cf. supra ad. Epid. 515). 
 
 Pers. 709 animus <^iam> in navist mihi (A* : mens P). 
 Cf. below, Pseud. 637. 
 
 Poen. 163 sq. 
 
 locare. Vin tu illam hodie sine damno et dtspendio 
 Tuo tuam libertam facere ? Cupio, Milphio. 
 
 Does this unmetrical setting of v. 163 in P (A n.l.) point to 
 an older : sine dispendio \ Tuo tuam and a later (with avoidance of 
 the proceleusmatic) : sine damno tuo \ Tuam ? 
 
 Poen. 648 Canes compellunt in plagas lepide lupum (A : lycum 
 [PXuKav, ? Lycum] P). 
 
 The /'-reading is probably a gloss explaining the allusion. 
 
 Poen. 1249 Quid si eloquamur ? Censeo hercle, patrue. Misera 
 timeo (Aa : Quin eloqunr \leg. -quamur] P). 
 
 Poen. 1252 Ne indigna indignis di darent, id ego evenisset 
 vellem (A : evenire P). 
 
 Pseud. 91 Quis mi igitur drachumam reddet, si dedero tibi? (A : 
 dederitn P). 
 
 Plautus may have written dederim (but cf. below, Pseud. 376). 
 
74 The Ancient Editions of Plautus. 
 
 Pseud. 140 officium A : opus P. 
 
 It is by no means certain that the one word is a mere gloss 
 on the other (see Baier ' de Plauti recensionibus,' p. 10). It 
 is possible that A and F offer two arrangements of the same 
 metrical line, a trochaic Septenarius with a Colon Reizianum, 
 viz. : 
 (i) Eorum officium . . quam hos domi 
 
 linquere custodes (^) ; (2) Est eorum opus . . linquere 
 quam hos domi custodes (F). 
 
 Pseud. 321 Quid nunc vis? Ut opperiare hos sex dies aliquos 
 modo (A* : saltern mo. P). 
 
 Is saltern a gloss ? 
 
 Pseud. 376 Quid id est? Si tu argentum attuleris, cum illo 
 perdidero fidem (A : perdiderim P). 
 
 Cf. above, Pseud. 91. 
 
 Pseud. 383 Nunc, Calidore, te mihi operam dare volo. Ecquid 
 imperas? (A* : Sed nunc Ca. op. mihi te dare volo P). 
 
 Pseud. 378 Sine argento frustra es qui me tui misereri postu- 
 las (Aa : frustra est qicod P). 
 
 The Plautine construction frustra es qui might be unintelligible 
 to a later generation. 
 
 Pseud. 385 Ad earn rem usust hominem astutum, doctura, 
 cautum et callidum (A* : d. scitum et call. P). 
 
 The alliterative version is likely to be PJautine. 
 
 Pseud. 421 hoc alii mihi renuntiant ; 
 
 Atque hoc iara pridem sensi et subolebat mihi (A : aiqueidY). 
 
 Pseud. 573 '^ Tibicen vos interibi hie delectaverit (A* : interea P). 
 
 Is this a ' Revival ' text removal of an archaic form ? Or is 
 it a mere gloss ? Cf. below, Stich. 371. 
 
 Pseud. 631 This form of the line was certainly present both 
 in Pa and in Aa : 
 
 Vae tibi ! tu inventu's vero meam qui furcilles fidem. 
 
The Andent Editions of Plautus. 75 
 
 But whether the intrusive solus in the Hne as offered by the 
 Ambrosian Palimpsest {inv. solus ve.) points to a variant solus 
 (in place of vero) or is a mere scribal error is uncertain. 
 
 Pseud. 637 id est nomen meum (A : mihi P). 
 Cf. above, Pers. 709. 
 
 Pseud. 657 Scin quid te oro ? (A* : orem P). 
 
 Pseud. 713 Si quid opus est (A : quicquid 0. e. P). 
 
 Pseud. 718 et quinque argenti minas (A : cum q. a. minis P). 
 
 Pseud. 723 Egone? Tu istic ipsus, inquam (A* : tu(e ergo ip. P*^ 
 
 Conceivably a gloss. 
 
 Pseud. 889 The Aa recension had : 
 
 Molestus ne sis. nimium iam tinnis. tace. 
 
 If there was another version : nimium tinnis. non taces 1 
 then we can explain the unmetrical /'-reading as ' mixed ' : 
 
 nimium iam tinnis. non taces ? 
 
 We might explain the other version as a conjectural emenda- 
 tion (by an editor or stage-manager) of a defective copy in 
 which iam had dropped out after nimium. But there is also this 
 possibility to be reckoned with, that both A and P are guilty 
 of aberration from this original ending of the line : 
 nimium iam tinnis. taces ? 
 
 Rud. 577 ubi fluit (A: si pi. P). 
 
 Rud. 787 non hercle equidem censeo (A : egomet P). 
 
 Stich. 94 PAN. Mane pulvinum. ANT. Bene procuras mihi 
 satis sic fultumst. PAMPH. Sede (A : m. sa. sicf. mihi P). 
 
 Stich, 140 Hostis est uxor, invita quae viro nuptum datur 
 (Aa : ad virum P^-). The P^ construction may be Plautine (e.g. 
 Epid. 38). 
 
 Stich. 371 Interibi Epignomum conspicio tuom virum et servom 
 Stichum (A* : interim P). 
 Cf. above, Pseud. 573*. 
 
 Stich. 451 Ibo opsonatum atque eadem referam opsonium (A: ea 
 ibo opsondtum, eadem ref, ops. P*). 
 
^6 The Anciefit Editions of Plautus. 
 
 The presence of atque in A may be due to a common scribal 
 error. 
 
 Trin. 6i Namque enim tu, credo, me inprudentem obrepseris 
 
 (A*: miki inprudentiV). 
 
 Trin. 88 Sed istuc negoti cupio scire quid siet ( A : scire cupio 
 quicquid estY). 
 
 Was the archaism siet the cause of the change ? See Leo's 
 note on the line and cf. Seyffert in Berl. Phil. Woch, XVI, 285. 
 
 Trin. 186 Were there two versions of the line? 
 
 (i) Hasc' mihi propter res maledicas famas ferunt {malas A), 
 (2) Hascine mi propter res malas famas ferunt ? (P*). 
 
 And is the malas of A due to a marginal variant or interlinear 
 gloss ? 
 
 Trin. 2 14 Bonis qui hunc adulescentem evortisset omnibus (A : 
 ev. suis P.) 
 
 To my ear bonis qui . . suis gives the truer Plautine rhythm. 
 
 Trin. 251 Vestispica, unctor, auri custos, flabelliferae, sandalige- 
 rulae (A : vestiplicae P). 
 
 Vesti{s)pica (cf. Nonius 12 M.) is Plautine, vestiplim late. (See 
 Leo in Melanges Boissier). 
 
 Trin. 335 praemandatum A : praedicatum P. 
 Is praedicatum a gloss ? 
 
 Trin. 448-9 
 
 Neque te derisum advenio neque dignum puto. (A* : veni P). 
 Verum hoc quod dixi (A^ ; ut d. P). 
 
 Trin. 452 Cum vestra nostra non est aequa factio (A: vesfris P). 
 
 Trin. 456 fertntaxinva {M: ferentaneumV). 
 
 The AA-reading is confirmed by Varro. Was the other a plebeian 
 form? (cf. above, p. 48 Subitarius became in Later Lat. 
 subitaiieus (cf. French ' soudain ' from Vulg. Lat. subitanus). 
 
 Trin. 537 Were there two endings of this line? 
 
 (j) apage istunc agrum, 
 
 (2) apage a me istum agrum. 
 
The Ancient Editions of Plautus. 11 
 
 Trin. 560 Lepide hercle agro ego hoc hunc senem deterrui (A : 
 /. h. de agro ego hunc s. d. P). 
 
 Trin. 743 Incolumem sistere illi (A : columem te s. i. Pa). 
 The PA-reading seems Plautine. 
 
 Trin. 747 Eadem omnia istaec veniunt in mentem mihi (A : Nam 
 hercle 0. Pa). 
 
 Is eadem a gloss ? 
 
 Trin. 839 dehinc iam certumst otio dare me (Aa : deinde hinc 
 ce. P). 
 
 The /'-reading may be due to a mere suprascript correction of 
 deinde (presumably a scribal error) into dehinc. 
 
 True. 73 
 
 (?) (i) Neque equidem id factura neque tu ut facias consilium 
 dabo (Aa), 
 
 (?) (2) Neque ego factura id sum neque, etc (Pa). 
 
 True. 189 spero (A) : credo (P). 
 
 True. 202 Turn pol istic est puero pater Babyloniensis miles 
 (Aa: isti puero quidemst pa. Pa). 
 
 True. 259 Salve. Sat mi est tuae salutis : nil moror : sat salveo 
 (A : nan salveo P), 
 
 If non be, as some think, a conjectural stop-gap after the loss 
 of sat (before ja/-), then the conjecture is of early date. For 
 the grammarian Marius Plotius Sacerdos twice quotes the line with 
 this version, non salveo. On the other hand sat may have been 
 a miswriting (left uncorrected) of the first syllable of salveo and may 
 have ousted non from its place. 
 
 True. 281 Sed quid ad nostras negoti, muher, est aedis tibi? (A ; 
 apud P). 
 
 A gloss? 
 
78 The Ancient Editions of Plautus. 
 
 True. 304-5 : 
 
 Quid maceria ilia ait in horto quae est, quae in noctes singulas 
 Latere fit minor qua is aput vos damni permensust viam ? (A : 
 
 qtm\ isii\ P). 
 Nil mirum — vetus est maceria — lateres si veteres ruont (A : non 
 
 m. P). 
 
 The MSS. of Priscian offer quani is hoc ad vos. Editors print 
 
 c]ua is to ad vos or qua is toe ad vos. 
 
 True. 378 DIN. lam lauta es? PHRON. Lauta mihi quidem 
 atque oculis meis (A : es? PHRON. lamfolm. Pa). 
 
 Is iam pot a stop-gap after the loss of the repeated lauta 1 Or 
 is the repeated word a gloss ? 
 
 True. 381 Cum inter nos sordebamus alter de altero (Aa ; alter 
 a Iter i Pa). 
 
 (For other examples, especially of ' mixed ' readings, see Seyffert 
 in Berl. Phil. Woeh. XVI 284 sqq.) 
 
 We have already had instances of a Grammarian's citation agree- 
 ing with Aa against Pa. 
 
 In the greater portion of Plautus' writings the evidence of A 
 is unprocurable. It will then be well to supplement this list by 
 a few typical cases of such divergent citations by Grammarians 
 as seem to have some claim to be regarded as the Aa versions : 
 
 Cure. 424 Clupeatus elephantum ubi machaera dissicit (diligit 
 Gramm.). 
 
 Men. 854 Barbatum, tremulum Titanum Cycno prognatura patre 
 {qui duet Cygno patre Gramm.). 
 
 Pers. 463 Tiara ornatum lepide condecorat tuum (lepida con- 
 decorai schema Gramm.). 
 
 (III.) External Fortn of the two Editions, Colofnetry, 
 Order of Plays, etc. 
 
 The form of the Aa text we infer from the form of the ancient 
 copy that has survived in a fragmentary and often illegible state to 
 
The Ancient Editions of Plautus. 79 
 
 our own time, the Ambrosian Palimpsest written in Rustic Capital 
 script of the 3rd or 4th century. To reconstruct the form of that 
 ancient copy of the Pa edition which was the archetype at once 
 of our existing minuscule MSS. {BCDEVJO) and of the lost 
 minuscule codex Turnebi {T) is not easy, but the recent dis- 
 covery of a collation of parts (Bacch. 35-80, 570-650, 810-900, 
 Pers., Poen., Pseud. 730-fin., Rud. 1-790) of 7" has greatly facili- 
 tated the problem. We may believe it to have been a codex 
 of the same script and therefore of about the same age as A. 
 While A has 19 lines to the page, it appears"" to have had 19 
 in some plays, 20 in others, 21 in others. But the number of 
 lines assigned to a page is a trifling matter. More important is the 
 form in which the lines appeared. The Palatine archetype seems 
 to have followed the same method as the Palimpsest for distinguish- 
 ing lines of different metre, the Alexandrian " method of Colometry, 
 by which the longest lines begin at the extreme left-hand of the 
 page (fV eKdecrei), the shortest nearer the middle of the page 
 (cV ela6ea-€i). Whether these curious opening KmMpia, which seem in 
 the Palimpsest to indicate a metrical ' system ' or run, were 
 also a feature of the Palatine archetype we cannot say °, e.g. 
 Epid. 9 sqq. : 
 
 " Since the number of lines in the Palatine archetype is inferred from the 
 gaps resulting from holes in leaves, there is always a possibility that all these 
 flaws may not really have belonged to one and the same MS. Some gaps 
 may go back to the original from which the archetype was itself copied, others 
 to the parent of that original and so on. The anticipation of Pseud. 1205-7 
 (which recur in their proper place) after 1 161 in T as well as in the minuscule 
 archetype points to a page of 2i lines in that original of which the majuscule 
 archetype was a copy. Another thing to be borne in mind in this connexion 
 is that a copy often retained on its pages the exact number of lines of its 
 original, so that the displacement of the Mostellaria leaves and the lacunae at 
 vv. 784, 802 may belong to different MSS. (See also p. 57 above.) 
 
 ^ From discoveries in Egypt and elsewhere we learn the ancient practice 
 of writing lyric passages continuously without breaking them up into lines. 
 Prof. Leo, who can speak with more authority than almost any other person 
 on the subject, ascribes the jjolometrical arrangement of Plautus' Cantica to 
 the time of Accius. 
 
 o But we may safely make a feature of both recensions the irregular dis- 
 tribution of lines in Synaphea, e.g. Capt. 524-5 omnis \ res for om- \ -nis res, 
 Epid. 173-4 credidi \ uxorem for credidi ux- \ -orem (of. Amph. 1067-8). It 
 is on the same principle, I fancy, that Men. 200 subcingulnm haud \ Hercules 
 aeque magna umquam abstulit periculo, appears in AP z.^ stibcingulum \ haud 
 Hercules, &c. , 
 
8o The Ancient Editions of Plautus. 
 
 QVID 
 
 TV AGIS VT VALES 
 
 EXEMPLVM ADESSE INTELLEGO EVGAE 
 
 Unfortunately the original line-division in our extant minuscule 
 MSS. has been so much abandoned in Song-Cantica for the sake of 
 saving space that we hardly dare to reconstruct it from the line- 
 division that now appears in these MSS. But in this matter too 
 the newly discovered collation of 7' is most helpful. Discrepancy in 
 the line-division of the two rival recensions can scarcely be looked 
 for outside of the Song-Cantica, for the ordinary Iambic and 
 Trochaic metres do not admit of variety in analysis. So that 
 our field for comparison of A* and Pa in this respect is narrowed to 
 the half-dozen-Song Cantica (or rather portions of them) for which 
 the evidence at once of T'and of ^ is available. Even within these 
 narrow limits we find sufficient trace of discrepancy to enable us to 
 infer, by induction from ' the known to the unknown,' that the two 
 recensions differed in this feature too, a difference which can hardly 
 (like mere difference of text) be referred to marginal variants. We 
 cannot suppose that an alternative scheme of a Canticum was 
 exhibited in the margin of this or that recension in such a form 
 that a subsequent copyist was likely to embody in his transcript the 
 marginal arrangement instead of the arrangement actually exhibited 
 in the text. 
 
 Pseud. 1329-30 appear^ in A* as a long bacchiac series : 
 Quid nunc ? numquid iratus es aut mihi aut filio pro- 
 -pter has res, Simo ? Nil profecto. 
 I hac : te sequor. Quin vocas spectatores simiil ? Her- 
 -cle me isti hau solent, etc. 
 Whether vocare . . . istos was in A* bacchiac (a second paeon 
 with a bacchius or a bacchius with a fourth paeon) or iambic (with 
 ergo) is not quite clear. 
 
 In Pa a bacchiac Trimeter Catalectic was followed by a long 
 cretic series : 
 
 P It is true that in A (in which the practice is to begin a new play on a new 
 leaf) the play ends on this page, the second page of the leaf, so that the scribe 
 might be tempted to combine short lines into long for the sake of saving space. 
 But he can hardly have hit upon a legitimate alternative metrical arrangement 
 by mere accident, and of course he cannot be credited with any knowledge 
 of bacchiac metre. 
 
The Ancient Editions of Plautus. 8i 
 
 Quid nunc? numquid iratus es 
 Aut mihi aut filio propter has res, Simo ? 
 Nil profectd. I Mc : te sequor. Quin vocas 
 Spectatores simul ? Hercle me isti hau solent. 
 
 Pseud. 131 5 sqq. The anapaestic run or 'system' was in A* 
 written in Tetrameters, but in P^ in Dimeters. 
 
 Similarly perhaps at Pseud. 908, 939, more certainly at 1 320-1 
 and 1 13 1. 
 
 Undoubtedly in some Cantica (e.g. Trin. II i) for which we have 
 not the evidence of T, we are almost entitled to infer from their 
 appearance'' in the Codex Vetus {B) that a different arrangement 
 prevailed in P* from that followed in A*. But the evidence even of 
 the Codex Vetus in point of Colometry is notoriously unsafe ; so 
 I prefer, in the absence of T, to ignore all metrical discrepancies 
 except such as are unmistakably ^ indicated by an alteration of the 
 language, e.g., Epid. 164-5 (in A^ iambic Octonarii, in P* trochaic 
 Septenarii). These have been already enumerated in Part I of this 
 section and need not be repeated here. But the strong argument 
 should not be forgotten which they furnish for the independence 
 of the two recensions. They almost suffice of themselves to 
 confute the commonly accepted theory that P^ and A^ are both 
 derived from one and the same ' variorum ' edition. 
 
 An interesting question is how far stage-directions were present 
 in the two recensions. In the Palimpsest there is now no trace 
 of any, though it is of course conceivable that they were origi- 
 nally written in red pigment and were washed out when the 
 vellum came to be prepared for re-writing in the 7th or 8th 
 century. In our extant Palatine MSS. there are one or two, 
 but it is not easy to say whether they are relics of the P'*^ recension, 
 or interpolations by some mediaeval scribe. 
 
 At Aul. 60j for example, where the dialogue between Euclio 
 and Staphyla ceases, and Euclio, after a remark to Staphyla, 
 goes on in a monologue, his monologue had in the archetype (at 
 
 1 The opening lines of Men. II iii are quite illegible in A. But the colo- 
 metry of B at least for v. 352 /nius .... fiat seems to have agreed with 
 that of Varro, who quotes this line (see p. 9 above). 
 
 r Most of Baier's instances of editorial re-casting of lines in dialogue-metres 
 seem to me so uncertain that I do not venture to use them for purposes of 
 argument. See above, for example, on Trin. 328 (p. 67). 
 
 G 
 
82 The Ancient Editions of Plautus. 
 
 least in the original of our extant MSS.) the heading' hoc secum 
 loquitur. The Menander fragments recently discovered in Egypt 
 have stage-directions here and there; e.g., in the mplmlpo^kv^ 
 fragment we find in the margin f^epx^'"'" ^<»P"' between the lines 
 noXeVo../ rfcreio-i, etc. And in the Pa family of MSS. we have 
 clear traces of the use of the Greek word Travres at the end of 
 the play when all the actors stepped forward 'to the footlights' 
 and asked for the applause of the audience (see below, p. 83). 
 
 Of the 'notae criticae' used by Alexandrian editors and their 
 Roman imitators* we find only this doubtful trace in our extant 
 MSS. Varro (Serm. Lat. iv 74 W.) mentions a ' nota transversa ' 
 (apparently like our symbol for hiatus) used to indicate ' syllaba 
 anceps ' (and hiatus ?) at the end of a line (or hemistich ? and in 
 pausa ?), as in Terence's line : 
 
 amicus summus meus et popularis Geta | . 
 
 It is possible that traces of this symbol survive in our MSS. of 
 Plautus. Schoell in his introduction to the Truculentus (p. xxxiv) 
 refers to this origin a curious habit of the Palatine scribes of append- 
 ing either the letter t or / or / (originally this ' nota transversa ') 
 to words ending a short colon (e.g. v. 120 mani^Vi^; 123 tu\t\, which 
 has become tuli) or the remarks of one speaker where there is 
 a change of speakers within the line (e.g. 152 probe\{\ Em, which 
 has caused Idem to be written for Em ; 203 expetit\f^, and the 
 instances which he cites do give some support to his theory". 
 In the same play at v. 209 Astaphium, who has previously been 
 speaking with Diniarchus, is left alone on the stage. The 
 
 " If Staphyla had left the stage, there would probably have been a new 
 Scene and Scene-heading here in Pa (see below, p. 88). This suggests the 
 possibility of some Scene-divisions in P, which are not found in A, having 
 arisen from a mere stage-direction (see p. 89). 
 
 ' The ' anecdotum Parisinum,' edited by Reifferscheid among the Suetonius 
 fragments, gives a list of the notae quae versibus apponi consuerunt, and adds : 
 his soils in adnotatlonibus Ennii, Lucilii et historicorum usi sunt f uarroshen- 
 nius t Aelius (i.e. Ael. Stilo) aeque et postremo Probus qui illas in Virgilio 
 et Horatio et Lucretio apposuit ut Homero Aristarchus. The corruption has 
 been emended variously (l) Varro, Ennius (if there was a grammarian of 
 the same name as the poet), (2) Vargunteius, (3) Varro, Sisenna. 
 
 >i It may be objected that our minuscule MSS. evidently had a very obscure 
 original to transcribe in the Truculentus. But its very obscurity may well have 
 made them more exactly reproduce what they found in its pages. When 
 a scribe understands what he is writing, it is then that he is least trustworthy. 
 
The Ancient Editions of Flauhis. 83 
 
 Palimpsest makes no break, but in CD there is a blank \mh, 
 while B indicates that a Scene-heading stood in the Archetype, 
 viz., the name astaphivm preceded by its ' nota personae ' and 
 followed by some symbol or other. This symbol was interpreted 
 by Dziatzko as a stichoraetrical number VL, i.e. 45, the number 
 of lines in the Scene, an interpretation which seems very doubt- 
 ful. The other supposed stichoraetrical number LX at the heading 
 of Trin. II ii, is still more doubtful ; it rather suggests that the 
 scribe was going to write lvsiteles before philto, but recog- 
 nised his error after he had transcribed two letters. The 
 H-like symbol in the margin of a later addition in A at 
 Stich. 623 is said not to be the symbol ' nota ' (like our " N.B."), 
 but merely a 'nota personae." And so is the X-mark which 
 preceded the Carthaginian passage in the P^ recension. 
 
 On the ' notae personarum ' used to indicate the speakers" 
 throughout a Scene, see below, pp. 91 sqq. Here it may be men- 
 tioned that Greek letters were normally used. A, B, r, A (often 
 confused in our MSS. with A), E, z, etc., while the last letter of 
 the Greek alphabet, a, was employed to indicate that all the 
 actors who were on the stage at the time joined in the speech. 
 We find it probably at Cas. 800, where Lysidamus and Olympio 
 join in the marriage-song : 
 
 OL. Suavi cantu concelebra omnem hanc plateam hymenaeo 
 mi (this is said to the tibicen). 
 
 CO Hymen hymenaee o hymen, 
 
 certainly at the end of one or two plays, Pers,, Poen., Trin. In the 
 Persa it was accompanied (glossed ?) by the stage-direction wdvTis, 
 which has led to the corruption pantio in our extant MSS. The 
 same ' nota ' or the same stage-direction in contracted form 
 has been, I think, the origin of the curious ite of our minuscule 
 MSS. at Bacch. 1207, and it is pos.sible that the grex and caterva 
 offered by the same MSS. at the ends of various plays are mere 
 interpretations of the same symbol. At Epid. 732 B has poeta, 
 / has GREX. (Can plaudite plaudite at the end of the Curculio 
 represent plaudite with a marginal Trdvrcs ?) 
 
 Ritschl's explanation of « as cantor (cf. Hor. Epp. II iii, 155) 
 is disproved by the newly-found collation of the Codex Turnebi. 
 
 In regard to the arrangement of the plays there is one point 
 for consideration to which sufficient attention has not, I think, been 
 
 G 2 
 
84 The Ancient Editions of Plautus. 
 
 paid. The earlier material for books was papyrus. The Hercu- 
 lanean papyri (not to go so far afield as Egypt), which formed 
 the library of the Greek philosopher, Philodemus, shew us what the 
 papyrus copies of the Early Empire were like. The several 
 ' volumina ' (i.e. books) were bound by a cord into a ' corpus ' 
 ij.e. volume), and a 'capsa' would contain one or (if there was 
 room) more than one volume. Now there would be no difficulty in 
 securing the right arrangement of the several books of a Greek 
 philosophical work, Book I, Book II, Book III, &c. But each play 
 of Plautus would be contained in a separate papyrus-roll. These 
 rolls might be grouped in a number of ' corpora ' and the whole 
 collection assigned to one 'capsa' (if it was large enough) ; but how 
 could the order of the several plays in each group be properly 
 determined ? That the arrangement in the time of Verrius Flaccus, 
 and (to judge from the parallel case of Naevius) in the time of 
 Varro too, was in alphabetical groups (an A-group, a C-group, 
 and so on), we have already seen. The place of the Amphitruo 
 at the beginning and of the Vidularia at the end of a collection of 
 the twenty-one plays would no doubt be indicated by some 
 notice that the collection began here and ended there. But there 
 could be no means of determining the true order of the other 
 two plays of the A-group, the Asinaria and Aulularia, or of the four 
 of the C-group, Captivi, Casina, Cistellaria, Curculio. We now- 
 a-days should not feel the difficulty, for we have carried alphabetical 
 arrangement to a point further than the Romans of the Early 
 Empire, and take account, as a matter of course, of the second 
 letter, and, if need be, of the third and fourth too, for the purpose 
 of arrangement. But the Romans at this time had not the idea 
 of this minute subdivision^. They lumped together anything 
 that began with A, anything that began with C, and so on. So 
 when the time came for the papyrus ' volumina ' of the plays to 
 
 T In the ' Parts I ' of Festus' Epitome of Verrius Flaccus (see above, p. 20) 
 there is indeed grouping by AB- and occasionally ABC-. But the proper order 
 of the second (or third) letter is not taken into account. Thus Part I of the 
 O-section is arranged so: Oc-, Or-, Ob- and Op-, 01-, Off-, Oe-, Ou-, Os-. 
 Our analysis of Nonius would be helped if we could fix a ' terminus a quo ' for 
 the modern alphabetical arrangement. For one of Nonius' lexicographical sources 
 contained a list of Verbs and another of Adverbs in which the modern arrange- 
 ment is followed (see No. I of this series), and we might be able to assign 
 a date to it and find out something about the earlier authorities from which 
 it is borrowed. 
 
The Ancient Editions of Plautus. 
 
 85 
 
 be written off on one (or more) parchment volume, the arrange- 
 ment of the plays within each alphabetical group would be left 
 to the caprice of the publisher. Unless therefore we are to assign 
 to the two recensions, Aa and Pa a later date than the papyrus 
 period, we have no right to use the different order of the plays 
 as evidence of anything more than different booksellers' issues. 
 The order of the Archetypes of our two families of MSS. and of 
 Nonius' copy (see above, p. 26) seems ^ to have been : 
 
 Ambrosian 
 
 Palatine 
 
 Nonius 
 
 Amph . — Asin . — Aul. 
 
 Amph. 
 
 Amph. 
 
 
 Asin. 
 
 Asin. 
 
 
 Aul. 
 
 Aul. 
 
 Bacch. 
 
 Bacch. 
 
 Bacch. 
 
 Capt. 
 
 Capt. 
 
 Cist. 
 
 Cure. 
 
 Cure. 
 
 Cas. 
 
 Cas. 
 
 Cas. 
 
 Capt. 
 
 Cist. 
 
 Cist. 
 
 Cure. 
 
 Epid, 
 
 Epid. 
 
 Epid. 
 
 Merc. 
 
 Most. 
 
 J 
 
 Most. 
 Mil. 
 
 Men. 
 Mil. 
 
 ,^ ■ >Merc. — Most, 
 Men. 1 
 
 Men. 
 
 Merc. 
 
 ) 
 
 Poen. 
 
 Pseud. 
 
 Pers. 
 
 Pers. 
 
 Poen. 
 
 Pseud. 
 
 Pseud. 
 
 Pers. 
 
 Poen. 
 
 Rud. 
 
 Rud. 
 
 Rud. 
 
 Stich. 
 
 Stich. 
 
 Stich. 
 
 Trin. 
 
 Trin. 
 
 Trin. 
 
 True. 
 
 True. 
 
 True. 
 
 Vid. 
 
 Vid. 
 
 Vid. 
 
 St I take it to be a mere accident of transcription that Trin. True, follow 
 Men. in A. The transposition of Bacch. after Epid. was effected in the Pala- 
 tine text in mediaeval times. The gap at the end of Aul. and the beginning 
 of Bacch. points to a large loss at this part in P. Presumably the remaining 
 leaves of the Bacchides were loose and had dropped out of their place. The 
 Carolingian abbot (if he it was) who gave them a position after the Epidicus 
 was no doubt influenced by the allusion to this play in Bacch. 214. If the 
 division of the plays into two volumes belonged to his time, then it was at 
 the beginning of Vol. II that he inserted the loose leaves containing the 
 Bacchides. 
 
86 The Ancient Editions of Plautus. 
 
 Punctuation in our extant MSS. of Latin authors generally dates 
 from the Carolingian period. That the minuscule archetype of the 
 Palatine MSS. retained some traces at least of ancient punctuation 
 is clear from the Carthaginian passage in the Poenulus, where 
 the words were separated by dots^ (preserved in a line or two 
 by B). And at Amph. 523 the corrector {B^) of the Codex Vetus, 
 whom I believe to have used for his corrections the minuscule 
 archetype, has supplied the punctuation-mark omitted by the scribe 
 of B after abii : _ 
 
 Clanculum abii : a legione operam banc subripui tibi. 
 
 This was also the punctuation of the immediate original of B and 
 D, for it is reproduced by D. And both the scansion abii \ a in 
 hiatus and the construction of subripio prove it to be right (cf. 
 Journ. Phil. XXVI, 292). But whether this punctuation belonged 
 to the actual Pa recension or was introduced by some later cor- 
 rector we cannot say. The Codex Bembinus of Terence has 
 been punctuated by a sixth-century corrector, whose name was 
 Joviales ^ His punctuation sometimes agrees, sometimes disagrees 
 with that recommended by Donatus in his commentary on Terence. 
 
 The manuscript tradition of Terence shews us the existence of an 
 illustrated ancient edition. We have no trace » of this in the case of 
 Plautus. 
 
 (IV.) The Arguments. 
 These are post-Plautine, and are of two kinds, Acrostic and non- 
 
 y Is there a trace of the same thing in A at v. 942 ? 
 
 ^ That he was some learned friend of the owner of the codex may perhaps 
 be inferred from the way in which he has signalized the completion of each 
 instalment of his task by the entry : hue usque loviales. The similar ser- 
 vices of Asterius bestowed on a friend's copy of Virgil are well known from 
 the commemorative lines which have been preserved in the Codex Mediceus : 
 Distincxi emendans gratum mihi_munus amici 
 Suscipiens ; operi sedulus incubui, etc. 
 
 " In Harv. Stud. IX 108 n. I have suggested that the curious titles CHLAMY- 
 DATvs, 'the man in the travelling-cloak,' and lvrchio (?), 'the tippler,' 
 assigned m the Scene-heading in our Palatine MSS. to the 'mercator' in the 
 Asmaria and the ' puer ebrius ' in the Miles, may conceivably have come from 
 an accompanying picture with these words written beside the figures. But 
 ftis IS, of course, a mere guess. A more likely account will be found below, 
 
The Ancient Editions of Plautus. 87 
 
 Acrostic. In both the archaic style is affected, not always with 
 success (e.g. itdque Cist. arg. 10) ; in the Acrostic hiatus is a 
 recognised feature of the verse, and is pushed far beyond the Hmits 
 imposed by Plautus. The Pahmpsest had originally no Arguments, 
 but a later owner (5th cent.) has added non-Acrostic Arguments in 
 Uncial script wherever a blank space was available (Pers., Pseud., 
 Stich.). Everything indicates that they did not come from the 
 original of which ^ is a copy, nor is there any reason for think- 
 ing that they came from any MS. embodying the A* recension. In 
 the Palatine Archetype the Acrostic Arguments were used. There 
 are also here and there (Amph., Aul, Merc, Mil.) non-Acrostic 
 Arguments. Since these latter sometimes (Amph., Aul.) precede, 
 sometimes follow (Merc, Mil.) the Acrostic, we may perhaps 
 infer that they were marginal additions which found their way 
 into the text. The present state of the conclusion of the Mercator 
 non-Acrostic Argument may point to the same thing. 
 
 So far then as the evidence goes, we may believe that the A* 
 edition had no Arguments, while the P^ edition was provided 
 with Acrostic Arguments, and these must be earlier than the time of 
 Donatus, if that commentator in his note on Andr. Ill v 4 
 (610) really quotes a phrase of the Argument to the Asinaria, 
 strangely ascribing it to Plautus : sic Plautus locutus est ' pretium 
 ob asinos ' pro asinorum pretium. The non-Acrostic Arguments, 
 each consisting of 12 lines, prefixed to the plays of Terence 
 were the work of Sulpicius Apollinaris, the teacher of Aul. Gellius, 
 in the first half of the second century ; but of the date and author- 
 ship of those which ultimately found their way here and there 
 into the text of the Palimpsest and of the Palatine archetype (or 
 its original) we know nothing. Since they are not identical in 
 the number ^ of their lines, it is improbable that they came from one 
 and the same bookseller's issue (much less ' third recension ') 
 of the plays. At any time an owner with a poetical gift, or with 
 a friend so gifted, might embelUsh his copy with one or more 
 Arguments, which would be transmitted in all subsequent transcrip- 
 tions. 
 
 *> As examples of identity in the number of lines of metrical Arguments may 
 be mentioned the Arguments in a Virgil-edition with 6 lines each, in an Ovid- 
 edition with 10 lines each. The size of these Arguments in ^ is : Pers. (14), 
 Pseud. (IS), Stich. (9); in P : Amph. (10), Aul. and Mil. (15), Merc. (16 
 or IS). 
 
88 The Ancient Editions of Plautus. 
 
 (V.) The Didascaliae. 
 
 The Palimpsest had Didascaliae for the Pseudolus and Stichus, 
 possibly for the Rudens (but the leaf is lost), also (but the page 
 is quite illegible) for the Vidularia. It had apparently none for the 
 Mercator, Trinuminus, Persa. The extant fragments do not allow 
 us to pronounce on the other plays. The Didascalia for the Stichus 
 is best preserved. Apparently, like the Didascaliae in the Codex 
 Bembinus of Terence, it was written partly in black, partly in 
 red. The red lines have been washed out at the time when the 
 manuscript became palimpsest, leaving the Didascalia in this form : 
 
 (blank space) 
 
 GRAECA ADELPHOE MENANDRV 
 
 ACTA LVDIS PLEBEIS 
 
 CN. BAEBIO C. TERENTIO AED. PL. 
 
 (blank line) 
 
 T. PVBLILIVS PELLIO 
 
 (blank line) 
 
 MARCIPOR OPPII 
 TIBIIS SARRANIS TOTAM 
 
 (blank line) 
 
 C. SVLPICIO C. AVRELIO COS. 
 
 (blank space) 
 
 So far as the evidence goes, we may conclude that the P* edition 
 had no Didascaliae, for there is absolutely no trace of them in 
 our MSS. They would probably be the product of the earliest 
 application of study to the plays; but the antiquarians of the 
 Republic were, we may suppose, unable to furnish them in the 
 case of the Mercator, Persa, Trinummus and probably other plays. 
 
 (VI.) The Scenes and Scene-headings. 
 
 In the differences which appear in the Palimpsest and in 
 the Palatine MSS. in respect of Scene-division, there is clear 
 trace of a difference of principle, which must be ascribed to the 
 two rival recensions. Where, for instance, one speaker remains 
 on the stage after the others have left it, the monologue is often 
 not made into a new Scene in the A* recension, unlike the P* 
 
The Aticieiit Editions of Plant us. 89 
 
 recension. Examples are : Pseud. 667 (Pseudolus' monologue), 
 1238 (Simo's monologue); True. 209 (Astaphium's monologue). 
 Similarly in the discrepancies at Pseud. 230, 1063, Pers. 711, 
 Bacch. 530, Most. 783, where new Scenes begin in Pa but not 
 in Aa, there is a clear difference of principle "= involved ; also at 
 Epid. 475, Pers. 7, where new Scenes begin in A* but not in P*. 
 
 As regards the form of the Scene-headings the question is 
 obscure, owing to the accidents that have interfered with their 
 transmission, particularly in the MSS. of the Pa edition. In dis- 
 cussing these it will be necessary to enter into some detail; for 
 the true history of the Palatine Scene-headings has not, I think, 
 been apprehended by scholars, and the question is by no means 
 unimportant. The Scene-headings must have been an essential 
 part of any ancient edition of Plautus and cannot be assigned (as 
 the titles of Horace's and Catullus' odes are usually assigned) to 
 a mere ' librarius ' in the employment of some ancient bookseller. 
 They may have come down from quite early times and have been 
 in the main identical in the two recensions *. 
 
 The normal form in ancient editions of Plautus, Terence, &c., 
 was, to judge from the Codex Bembinus of Terence, a two-line 
 heading, containing in the first line the name of the speaker, 
 and in the second the character of his part, the role he had to 
 play. Thus at Stich. II i, for example, we should have : 
 
 PINACIVM GELASIMVS 
 
 PVER PARASITVS. 
 
 <= Of course the failure to indicate a new Scene may often be the mere result 
 of a copyist having failed to notice the blank line (left blank by the remissness 
 of the lubricator) which had been reserved for the Scene-heading in his original. 
 This he would be especially prone to do when the blank line stood at the 
 top or bottom of a page. The ignoring of Pers. IV i.x in A may be due to 
 an error of this kind. But the confusion in our Palatine MSS. at Most. 1063 
 is the result of the disarrangement of the leaves of this play in some early 
 archetype. At Aul. 328 I would ascribe the erroneous Scene-division in 
 these MSS. to the lacuna in v. 328 having been mistaken for a space left 
 blank for a heading. We have not the Palimpsest to afford a contrast of 
 the A* Scene-division with the curious breaks in the Palatine MSS. at 
 Aul. 449, 731. 
 
 ■• The alternative ending of the Poenulus (see above, p. 44) had no heading 
 in A* P ^. The Palatine MSS. leave an interval of one line; the Palimpsest 
 goes on without a break. The two recensions also agree in not making a 
 new Scene at Mil. 1427. 
 
go The Ancient Editions of Plautus. 
 
 The order in which the speakers are enumerated follows ° the 
 order in which they appear in the Scene. In this Scene of the 
 Stichiis Pinacium is the first speaker, Gelasimus speaks after 
 him. But where two speakers had similar roles (two old men, let 
 us suppose, or two slaves), they would be enumerated together 
 for sake of convenience, e.g. Pseud. I v : 
 
 SIMO CALLIPHO PSEVDOLVS 
 
 SENES • II • SERVVS. 
 
 Here Simo is the first speaker, Pseudolus the second and Callipho 
 the third ; and the departure from the true order is obviously 
 necessitated by the convenience of having senes • ii • in the second 
 line ^. The right order of the two ' senes ' as compared with each 
 other is adhered to (cf. Most. Ill ii). A variety of this normal 
 form arose when some of the speakers were the same as had 
 figured in the previous Scene. They might s in that case 
 be designated by the single word eidem, e.g. Poen. V v : 
 
 ANTAMOENIDES EIDEM 
 
 MILES 
 
 8 The occasional exceptions, e.g. Vid. 69 in ^ : 
 
 DINIA NICODEMVS CACISTVS 
 
 (instead of Cac, Nic), are, no doubt, due to a scribe's error (cf. Epid. I i 
 in A, Mil. IV viii, Cas. Ill vi, Cas. IV ii in P). # 
 
 f It seems to me hyper-critical to object to the presence of the numeral duo 
 and to argue that it could only appear when the names were suppressed or 
 omitted. Certainly SIMO CALLIPHO SENES is sufficient in itself and does not 
 need the addition of duo. But the Roman editors may quite well have followed 
 the practice of adding the numeral. 
 
 E This formula seems to have been singularly conducive to divergence of 
 transcription. Possibly the best explanation is that scribes were tempted to 
 expand eidem and substitute for it the actual names. Thus at Poen. V iii 
 where A offers idem and P gives the names, the difference may not really 
 be attributable to the two recensions. Whether the type found occasionally 
 in Pa, e.g. at Cist. I ii, was also found in Aa, we cannot say : LENA RESTITIT. 
 Resto, be it noticed, is Plautus' own expression for remaining on the stage, 
 e.g. Trin. 718 Abiit ille quidem . . . Hie quoque hinc abiit. Stasime, restas 
 solus. There are a large number of one-line Scene-headings in the Palimpsest, 
 now blank, but presumably at some time filled with writing. In fact until 
 the Casina the one-line heading would seem to have been normal in A, also 
 throughout the Menaechmi, and often elsewhere. (See, however, the remarks 
 on this matter in the Preface to Studemund's Apograph.) At Pers. IV vi 
 I think both Aa and Pa had : LEND [ ? ac] EIDEM. 
 
The Ancient Editions of Plaiitus. 91 
 
 Another feature of Scene-headings was the appending of the 
 symbols ^ -Q.- (i.e. Canticum) and • DV • (i.e. Deverbium or Diver- 
 bium) to indicate the presence or absence of musical accompani- 
 ment. Another was the presence of those ' notae personarum ' 
 which throughout the Scene itself were used to indicate the dif- 
 ferent speakers. 
 
 What then are the accidents which have interfered with the trans- 
 mission of these Scene-headings of the two rival recensions ? In the 
 Palimpsest, where, as we have seen, two lines have usually been 
 reserved for a Scene-heading, the first line (containing the names) 
 remains, the second line is blank. Thus the Scene-heading at 
 Stich. II i, mentioned above, appeais thus in A : 
 
 PINACIVM GELASIMVS 
 
 (blank line) 
 
 We may infer that (as we found to have probably been tlie 
 case with the Didascaliae) the blank line had been filled in by 
 the ' rubricator.' The red pigment had been washed off when 
 in mediaeval times the IMS. was prepared to receive another 
 piece of writing. Where a one-line Scene-heading occurs in 
 A, the Hne is as a rule blank ; it had been, we may suppose, 
 written in red letters. The result of this accident to the trans- 
 mission is that we are left ignorant ' of the nomenclature of the 
 roles in the Palimpsest. All that has been handed down to 
 us of the Scene-headings of the A-^ recension is the proper names. 
 We are of course entitled to infer that the second lines, now blank, 
 contained the roles, and we may perhaps also infer that the Symbols 
 ■ C • and • DV ■ were likewise present in the Palimpsest, being written 
 (as such symbols naturally might be) in red pigment. There are 
 
 •> Leo assigns these to the second century a.d. on the ground that the 
 contraction of Deverbium would at an earlier time have taken the form, not 
 of DV, but of DEV. But this reasoning is not conclusive. Surely any scribe 
 or editor, who found in an older copy the old-fashioned type of contraction, 
 would alter it to the type current in his own day. I gather from Donatus' 
 Preface to the Adelphi that these symbols stood after the names in the firet 
 line of the heading. 
 
 i By some freak, I suppose, of the scribe we find in A at Mil. Ill i and 
 Rud. II vii a two-line heading with the roles written in black in the firet 
 line, and with die second line blank, thus reversing tlie usual order of things. 
 In Stich. V iv, where A offers merely tibicin'a, this is probably to be explained 
 by the ancient practice (see below) of not assigning names to ' mutae personae ' 
 and quite subordinate personages (but cf. I ii, II ii of the same play). 
 
g2 The Ancient Editions of Plauius. 
 
 blank spaces for ' notae personarum ' throughout the Scenes; so 
 presumably they were in red too. And they may have been entered 
 (in red) beside the names in the Scene-headings ; for the Codex 
 Bembinus of Terence, which has the proper names in black in the 
 first line of a heading and the roles in red in the second line, shews 
 the ' notae personarum ' in red standing in the first line, each before 
 the name it represents. We cannot say how far they were identi- 
 cal ^ with the Pa symbols ' or followed a lettering of their own. 
 
 Far more obscure is the history of the Palatine recension in 
 respect to the transmission of the Scene-headings. Any reader 
 of the best extant MS. {B) of the Palatine family is at once struck 
 by the number of what we may style single-line headings (cf. 
 Berl. Phil. Woch. XII 253). These contain, as a rule, the names 
 of the roles, not of the speakers. At Stich. II i, to take our previous 
 example, the best extant copy of the Palatine recension has merely : 
 
 PVER PARASITVS. 
 
 k The use of the identical ' nota ' E for the ' leno ' in both recensions at 
 Poen. 474 is suggested by the error found in both A (A^) and P, evolaHcorum 
 for vol- (cf- below), perhaps also, A for Stasimus in Trin. 495. Errors like this 
 in A (cf. Mil, 173, Cist. 518, &c.) suggest that in /4's original or archetype 
 these 'notae' were occasionally in black ink like the text. The same kind of 
 error is very frequent in the Palatine text (e.g. Mil. 276, 1 138, Most. 719, 750, 
 Aul. 829, Stich. 542, Trin. 931, True. 690; cf. Poen. 140; Men. 151). That 
 scribes at an early time had the habit of expanding these contracted symbols 
 and substituting the actual name, may be inferred from a common corruption 
 in our MSS., viz. the omission of a name (in the vocative case usually) which 
 is immediately followed by a remark of a speaker bearing this name. Also 
 from an error like that of the Palatine MSS. at Pseud. 954 : 
 
 Illicinest ? Illic'st. Mala mercest, Pseudole ; illuc sis vide, 
 where our MSS. shew instead of Pseudole the ' nota personae.' 
 
 1 In the Palatine MSS. the speakers are so often erroneously indicated that 
 an editor may ignore their testimony with but little scruple. In particular the 
 error of assigning the opening words, instead of the middle or end, of a line 
 to a new speaker is so frequent as to rouse the suspicion that the paragraplios- 
 symbol stood in the margin of some early archetype. In the recently found 
 Greek Dramatic papyri sometimes this symbol, sometimes the name of the 
 speaker is used. In the Ambrosian Palimpsest (5th cent. ?) of Seneca's Tra- 
 gedies there is a curious practice of assigning a new line to the second part 
 of a verse which is divided between two speakers. This practice has led to 
 the loss of a half-line at Oed. 399 (see Studemund, p. xix of Leo's preface 
 to Seneca). I do not know that there is any trace of this practice in the 
 Palatine tradition. 
 
The Ancient Editions of Plautus. 93 
 
 This feature might be passed over as devoid of significance, since 
 our extant MSS. are separated from the Palatine recension by a 
 number of centuries, were it not that there are unmistakable traces 
 of some defect in the transmission of the proper names in all this 
 family of MSS. The name of a leading character in the Casina, the 
 old reprobate, Lysidamus, does not occur in our Palatine Scene- 
 headings ; we learn it from the Palimpsest. In the Menaechmi 
 we have not the Palimpsest to help us at the requisite parts of 
 the play, and so we are left in ignorance of the name of the 
 wife of the Epidamnian twin-brother. We find nothing in the 
 Palatine headings but the name of the role, mvlier. The name 
 Lampadio appears in the Palatine headings in the Cistellaria as 
 LAMPADiscvs, and so on. How are we to account for this ? 
 Some scholars suppose that Plautus did not assign a name to 
 some of his characters, and that ancient editors, or even ' librarii,' 
 might supply any name that suited their fancy. No name, they 
 say, was given to the old reprobate in the Casina. The editor 
 of the A* recension chose to call him Lysidamus. The editor 
 of the Pa recension chose to leave him unnamed. This seems to 
 me extremely unlikely in the case of the leading characters of 
 a play, like Lysidamus in the Casina and Menaechmus' wife in 
 the Menaechmi, although it is conceivable for minor personages, 
 such as the ' senex ' in Terence Eun. V v "". Far more likely is the 
 supposition that by some accident the names of Lysidamus in the 
 Casina and of the matron in the Menaechmi had dropped out 
 of. all the Scene-headings of some archetype in these plays. Why 
 then were they not restored by the next scribe who took a copy 
 of this archetype ? The answer is clear. Because it so happens that 
 these two names do not occur at any part of the text. 
 
 Further consideration shews us " that all the names which are 
 given correctly in the headings are such as could easily be recovered 
 from the text itself, so that what has certainly happened in the 
 Casina, Menaechmi, &c., we may suppose to have happened in 
 all the other plays too. Also that there are positive indications 
 
 "■ He appears only in this Scene, a short Scene of some thirty lines, and 
 nowhere else in the play. Donatus says of him : huius senex nomen apud 
 Terentium non est ; apud Menandrum Simon dicitur. The Bembinus has 
 DEMEA, the other MSS. laches. 
 
 1 For .details, I refer the reader to the article by my old pupil, Mr. H. 
 Prescott, in Harvard Studies, Vol. IX, pp. 102 sqq. 
 
94 The Ancient Editions of Plautus. 
 
 that this did actually happen throughout the volume, for certain 
 errors in the presentment of names in headings can be traced 
 to peculiarities of the text °. The name Lampadio, we have seen, 
 has in the Cistellaria headings of Palatine MSS. the form Lam- 
 padiscus both at its first occurrence (II ii) and subsequently 
 (IVi, ii). Why? Evidently because in the text the slave is first 
 mentioned (v. 544) under the pet-name: 
 
 Audire vocem visa sum ante aedis modo 
 Mei Lampadisci servi. 
 
 For the old man of the Casina the Palatine MSS. offer in some 
 Scene-headings an impossible substitute for the true name Ly- 
 sidamus, viz. stalicio. This is exactly what a blunderer might 
 elicit from the corrupt reading of this family of MSS. in v. 960, 
 where Chalinus calls out to Lysidamus : 
 heus ! sta ilico, 
 
 of which the Palatine reading is : 
 
 heus stalicio. 
 
 The blunder was confirmed by the change of the puzzling word 
 tittibilicio in v. 347 into tibi Stalitio (-do) : 
 
 non ego istuc verbum empsim tittibiUcio (tit tibi stalicio P), 
 
 a change which would seem to have been made after, rather than 
 before, the acceptance of the fictitious name. 
 
 If the reader will take the trouble p of going through the Scene- 
 headings, he will find that each and every one squares with this 
 theory. Of course the vast majority of the characters are mentioned 
 by name in the text itself, and so the actual number of instances 
 
 We have no such indications in the Palimpsest. So that while we are 
 entitled to infer that the names in the Palatine Scene-headings are later in- 
 sertions from the text, we are not entitled to make this inference regarding 
 the Ambrosian Scene-headings. We are not entitled, for example, to declare 
 boldly that the name Lysidamus in the Casina may have been found by the 
 scribe of some original of ^ in a part now lost of the text and thereupon inserted 
 in the Scene-headings. Such a theory is wholly baseless. There is absolutely 
 no indication of anything of the kind. 
 
 P Readers, I know, will not take the trouble, and so I am not confident 
 of this theory being accepted until someone writes a dissertation on the Scene- 
 headings of Plautus and gives a full list of them all, or at least the more notice- 
 able, one after another. 
 
The Ancient Editions of Plautus. 95 
 
 relevant to our argument are few. But it is not right to turn this 
 paucity of crucial instances into an objection against the theory. 
 If the vast majority of instances are capable of being explained 
 equally well on either hypothesis, either as directly transmitted 
 or as recovered from the text, then it is on the few remaining 
 instances that the proof of the true hypothesis must of necessity 
 depend 1. A more reasonable objection has been made on this 
 score, that these (according to our theory) re-written Scene-headings 
 shew as a rule the correct order of the names, the order which 
 corresponds to the actual order of the speakers in the ensuing 
 dialogue. This objection must be narrowed down to its true 
 limit. Since our theory posits the direct transmission of the roles, 
 e.g. at Stich. II i : 
 
 PVER PARASITVS, 
 
 the direct transmission of the order of the names (though not of 
 the names themselves) is in most cases involved. The only cases 
 which the objection can refer to are the few headings where the 
 speakers bear the same role, e.g. senes • 11 •. But surely any scribe 
 or ' corrector ' who filled in the names of the ' duo senes ' with 
 the help of the text would naturally place them in the order in 
 which they appear in the following lines. I for my part can hardly 
 imagine him to have deliberately adopted any other order, and am 
 inclined to ascribe the exceptions (e.g. Merc. II ii) to a mere trans- 
 position ' by a scribe, of the same sort as the transposition of two 
 words in a line. The scribe of the 12 th century MS.y (or rather, 
 I fancy, of the common parent of y and 0) has placed the ' senes 
 duo ' of Epid. V ii in their right order. He certainly found nothing 
 in his original except senes • 11 • (so in BE). But a more telling 
 example, one which, in my opinion, suffices of itself to destroy 
 the whole objection, is that in the Mostellaria (v. 858), where 
 the transposition of leaves in some Palatine archetype had oc- 
 
 "J I admit, however, that the slightness of the material will always leave room 
 for suspicion regarding the extent to which the original names in the Scene- 
 headings had disappeared. In A we see occasionally the roles preserved instead 
 of the names (cf. above). A similar freak may have disturbed the uniformity 
 of the Palatine recension. Where this recension offered one-line headings I 
 believe them to have survived in their original form (cf. above). 
 
 '' Another ' causa erroris ' may have been the mis-writing of the ' notae 
 personarum ' in the Scene-headings (see below). 
 
96 The Ancient Editions of Plautus. 
 
 casioned the re-writing of the traditional Scene-headings, the re- 
 writer presented precisely the order required by the accident to 
 the text : 
 
 ADVERSITORES TRANIO THEVROPIDES SIMO DANISTA. 
 
 For the disarranged state of the lines exhibits this arrangement : 
 858-883, 802-841, 601-646. The first speaker is Phaniscus, one 
 of the ' adversitores,' the next (at v. 803) Tranio, then Theopropides, 
 then Simo, finally (at v. 603) the ' danista.' 
 
 These are, so far as I know, the only objections that have been 
 offered to this theory of the accidental loss of the proper names 
 in the Palatine Scene-headings and of their subsequent restoration 
 with the help of the text. The theory so exactly suits the facts 
 that it must, I think, hold the field until it is definitely disproved. 
 It follows that the names in the Scene-headings of our Palatine MSS. 
 are to be regarded as devoid of all traditional authority. Accident 
 therefore has deprived us on the one hand of the roles in the 
 A* Scene-headings, and on the other of the names in the P* 
 headings, and we have not the means of making a comparison 
 between the Scene-headings of the two recensions ; although we 
 may fairly re-construct the ancient type of headings by a com- 
 bination of the valid materials ^ preserved by the Palimpsest and 
 by the Palatine MSS. We cannot, for example, infer that in the 
 Scene-headings of the Poenulus the Carthaginian was designated 
 differently in A* and P*. In A* indeed (at least in A) he certainly 
 appears merely as poenvs ; but we no right to assign to P* the 
 fuller heading indicated by our Palatine MSS. : 
 
 HANNO 
 
 POENVS. 
 
 Pa may perfectly well have had the same heading as Aa ; for the 
 HANNO of our MSS. has, our theory shews, been transferred from 
 
 s I see no reason to doubt that the r61e of Harpax in the Pseudolus was 
 styled in ancient times cacvla (P, A nj.), and of Simia in the same play, 
 SYCOPHANTA {F, A n.l.). Possibly even the designation of Pseudolus in 
 the concluding scene of the play as servvs ebrivs (/", A n.l.) may have been 
 the accepted designation in ancient editions. It is possible that mercator 
 CHLAMYDATVS, 'the merchant in the travelling-cloak,' was a similar r61e-title 
 in the Asinaria, and even PVER LVRCio(-CHO), 'the tippling slave,' in the 
 Miles (see above, p. 86). Whether the name Truculentus comes from a r61e- 
 designation SERVVS TRVCVLENTVS is discussed below. 
 
The Ancient Editions of Plaiitus. 97 
 
 e text into the headings. Nor can we assert that a custom of 
 *^ was alien to the Pa headings, viz. of not giving the names 
 
 ' mutae personae ' or quite subordinate characters, and appeal 
 
 Aul. II iv, where our Palatine MSS. offer phrvsivm exelevsivm 
 BiciNAE • II • . These names have been taken from a line (v. 333) 
 
 the same Scene. 
 
 Until the recent discovery of the partial collation of the Codex 
 
 irnebi we had no means of dating this accident that has occurred 
 
 the Palatine Scene-headings. That codex however now appears 
 
 have had substantially the same headings as the parent of our 
 her minuscule MSS. These defective and wrongly supplied names 
 rALicio, LAMPADiscvs, etc.) vvc may refer to the common parent 
 
 T and the other minuscule codices, in other words to the ancient 
 ajuscule Archetype. The latest possible date for the arbitrary 
 -writing of the first lines of the Palatine Scene-headings is thus 
 ished back to some early century (3rd — 5th). We cannot ascribe 
 to the Carolingian abbot who first found the ancient codex in 
 e library of some monastery (shall we guess Fleury ?) and gave 
 to his monks to transcribe in minuscule characters ; for names 
 
 the fragmentary plays, phanostrata in the Cistellaria, phedria 
 d FiTODicvs in the Aulularia, have been recovered for the Scene- 
 adings by help of portions of the text which were undoubtedly 
 issing at that time as they are now. The date of the accident to 
 £ transmission of the Palatine Scene-headings must therefore be 
 ry early. 
 But it may be asked (and here, to my mind, lies the strongest line 
 
 attack upon this theory), ' How can we be sure that all these 
 writings of the traditional Scene-headings were made at one and 
 ; same time ? ' The possibility of successive scribes and cor- 
 ;tors having each contributed their quota of alterations cannot 
 
 denied. For example, the mediaeval editor*, who prepared 
 It ' doctored ' version of the first eight plays from which the MSS. 
 and O, and the corrections (F==) in , the MS. Twere derived, 
 doubtedly ' doctored ' the Scene-headings " as well as the text 
 
 The little poem which he prefixed to his edition is well known : 
 
 Exemplar mendum tandem me compulit ipsum, etc. 
 Sometimes successfully. At Asin. IV ii his original had adolescens 
 JYRIPPVS ET (?) PARASITVS • DV ■ which I take to have been an early re- 
 ting of : 
 
 (blank spaces interspersed by 'notae pers.' and followed by -DV ■) 
 
 adolescens PARASITVS. 
 H 
 
g8 The Ancient Editions of Plautus. 
 
 (e.g. Curc.l iii, IV ii). For these eight plays the nearest view that 
 we can get of the Archetype's readings is provided by the cor- 
 rections (^3) in the Codex Vetus {B). I beUeve them to come 
 straight from the minuscule archetype of our extant MSS. But 
 this corrector has also interspersed corrections of the text on his 
 own responsibility. Now the Scene-heading of Amph. IV ii in B 
 shews the name blefaro introduced by him, where the other MSS. 
 point to an original mercvrivs et idem. Similarly he changes 
 PROLOGVS at Cist. I iii into prologvs avxilii dei (cf. v. 154 nam 
 mihi Auxilio est nomen). Are these emendations of his own, or 
 part of that general re-writing of the Scene-headings which our 
 theory would ascribe to an early date? If the latter, then we shall 
 have to refer the gap in the Amphitruo to this early time; for 
 the addition of the name Blepharo at IV ii betokens ignorance 
 of the fact that Blepharo does not really appear till a later Scene. 
 The transposition of the leaves in the Mostellaria already mentioned 
 must be equally early if the Scene-heading of Most. IV i was (as 
 I certainly am inclined to believe) re-written at the same time as 
 the others. Again as evidence that the scribe of D had introduced 
 modifications of his own the Scene-heading of Trin. Ill iii might be 
 quoted, where the common original oi B C D had undoubtedly 
 no names but only the role senes ■ 11 •, but where the scribe of D 
 has written Philto Callicles senes • II' . Here however his mistake, 
 Philto instead of Me^aronides, gives us a clue to the real expla- 
 nation. In the common original (the minuscule archetype) of 
 the three MSS the entry senes • 11 • was accompanied not only by 
 the symbol D V (found in B and C) but also by the ' notae per- 
 sonarum.' The ' nota ' used for Megaronides in this Scene is iden- 
 tical with that used in previous Scenes for Philto. What the scribe 
 of £> has done was not, strictly speaking, to add by his own con- 
 tribution two proper names, but merely to write in full these ' notae ' 
 which he found in his original ; in short he has done nothing more 
 than any scribe who writes pater in full instead of the contracted 
 pr of his original. Nevertheless there is no denying of the pos- 
 sibility of successive copyists and correctors having contributed, 
 here a little and there a little, to the alteration of these Scene- 
 
 The re-writer had not noticed that the ' adolescens ' in this Scene is not 
 the same ' adolescens ' as in the previous Scenes, viz. ' Argyrippus,' but a new 
 ' adolescens,' viz. Diabolus. y offers ' Diabolus ' instead of ' Argyrippus.' 
 
The Ancient Editions of Flautus. 99 
 
 headings, although, if the commonly accepted view ^ of the genea- 
 logy of our MSS. be correct, the successive stages of transcription 
 are extremely few. And here is the weak point of our theory. 
 How can we prove that this or that re-writing of a Scene-heading 
 actually goes back to the early date we assign? May not the early 
 Archetype (in Rustic Capital Script) of T B C D have had the 
 Scene-headings in the Casina and Menaechnii in full and correct 
 form? May not the omission of the names of Lysidamus and of 
 Menaechmus' wife be due to the scribe of the minuscule archetype 
 oi B C D and the interpolation of the fictitious stahcio to some 
 subsequent transcriber? What makes our theory weak against an 
 attack of this kind is the fact that we have not the e\idence of 
 the Codex Turnebi {T) for these plays y, but only for the Persa, 
 Poenulus and parts of the Bacchides, Pseudolus and Rudens. Now 
 it is the consensus of 7" with BCD which takes us back to the 
 early majuscule archetype, while the consensus o[ B C D (without 
 T) takes us only to the Carolingian minuscule archetype. We can 
 assert positively that the majuscule archetype had, e.g. in Rud. 
 liv. : 
 
 5 That the minuscule archetype of BCDEVJO was immediately transcribed 
 from the early majuscule archetype and was the immediate source of the B^- 
 corrections in the iii-st eight pUiys and tlie immediate original of B in the 
 last twelve. Will not some one who has leisure examine the readings of 
 BCD in the last twelve plays and settle once for all the question whether 
 there is any indication of a second minuscule MS. having intervened between 
 B and the minuscule archetype? The close relationship of i? to Z> in the 
 first eight plays is not, I think, affected by Zangemeister's pronouncement 
 that B is later than D. For the difference in age cannot be considerable. 
 The theory suggested by certain blanks left by ^' (e.g. at Cure 534, 53S), 
 vii., that B was written at a time when certain passages (transcribed in D 
 and the original of VEJ) had been lost or obscured in tlie common original 
 of BDVEJ, is hardly warranted. These blanks are the mere result of the 
 habit of the scribe of B to omit anytliing that puzzled him, and to leave it 
 for the corjector (£3) to supply. What puzzled him in these two lines of 
 the Curculio was the forms eafsi and cj/iv (cf. Schoell's note on Cas. 347). 
 
 y At the same time it might be inferred from the silence of Turnebus that 
 he did not find at least the name Lysid.amus in the Casina. And since 
 Gas. 347, through the Umid h.abit of the scribe of the Codex Vetus, \4-as only 
 written by him as far as the fifth (not fourth) word, the puzzUng terminaUon 
 of the line being left to the corrector to supply, the fictitious name is attested 
 by .55, that U (if our theory of the 'stemma codicum ' be right) by the min- 
 uscule archetype itself. 
 
 H 2 
 
loo The Ancient Editions of Plautus. 
 
 PALAESTRA AMPELISCA • C • 
 MVLIERES • II • 
 
 and at Pseud. IV ii (where the Palimpsest shews ballio psevdolvs 
 siMiA followed by a blank line which presumably contained the 
 roles) : 
 
 BALLIO IDEM • C • 
 
 and at Pseud. IV vii : 
 
 CACVLA LEND SENEX, 
 
 and at Bacch. IV i : 
 
 PARASITVS PVER ADOLESCENS, 
 
 and so on. For all or most of the Scene-headings in the parts 
 for which the newly found collation T helps us, we can confidently 
 reject the possibility of a later origin of this or that modification. 
 We can recover the ' ipsa verba ' of the majuscule archetype ^ 
 without any fretting uncertainty. Now it is the practical uniformity 
 of the 'ipsa verba' in these headings with the best tradition^ of 
 the MSS. B C D m these parts of Plautus which encourages us 
 to believe that the same uniformity prevailed in other parts too. 
 On the other hand, if any objector requires absolute proof in each 
 and every instance, we can only admit that the proof is not procur- 
 able and that it will remain so until some lucky chance gives us 
 the rest of the collation of the Codex Turnebi. But we may in our 
 
 2 We can therefore state as a fact that the name SIMIA never appeared in 
 the Pseudolus Scene-headings but only the role sycophanta. How well 
 this suits our theory ! Since our modern distinction of proper names by the 
 use of capital initial letters was unknown in MSS., this name, though it 
 occurs here and there in the text, might easily be taken for the noun simia 
 (later simmia) 'a monkey.' In the Argument Simiae (Simmiae) shews in all 
 our MSS. the corruption si nimiae, a corruption which appears to go back 
 to a period of majuscule script simmiae. 
 
 " Sometimes B and C fail us and we have to fall back on D alone, a much 
 weaker testimony. For example, in Rud. Ill iv where D has daemones 
 SENEX and where the rubricators of B C have omitted the Scene-heading, we 
 find in T what I believe to have been the ' ipsa verba ' of the Archetype, 
 viv. SENEX preceded by the ' nota personae' A. In Aul. II vi the form of 
 the Scene-heading in our extant Palatine MSS. enables us at least to refer 
 the heading to one stage earlier of transmission. Both B and V offer STRO- 
 BILVS SERVVS STAPHILA ANVS cocvs HYL. Here the last item was clearly 
 a marginal correction in some archetype of the spelling STAPHILA. 
 
The Ancient Editions of Plauius. lot 
 
 turn demand from the objector (and really this is all that is 
 required for our present purpose) that he for his part shall 
 cease to speak of the names in the Palatine Scene-headings as 
 if they certainly came by direct transmission from the Palatine 
 recension itself ^ 
 
 The exact nature of the accident is not quite certain ; but the 
 simplest and most natural way in which the first line of a Scene- 
 heading would become blank would be through the remissness 
 of a rubricator. The scribe himself would supply the black line 
 (in this case presumably the second line) of the heading, and would 
 leave the other line for the rubricator to fill in. If the PA-headings 
 had a form identical with that exhibited by the Codex Bembinus 
 of Terence and possibly exhibited, as we have seen, by the Plautus 
 Pahmpsest, and if the only difference was that the PA-heading 
 shewed red where the Bembinus and A shewed black and vice 
 versa, then the parts of the P^-headings which the scribe himself 
 would supply would be (i) the ' notae personarum ' in the first 
 hne, (2) the roles in the second line, (3) the symbols D V and 
 C (presumably in the first line); the part which he would leave 
 for the rubricator to supply would be (4) the proper names. Un- 
 doubtedly the nearest glimpse that we can get of the Palatine 
 archetype, by the help of the more accurate of the extant MSS. 
 in their most accurate portions^ shews some such form of heading 
 
 * On the strength of the presence of the name Euclio in the heading of the 
 last Scene of the fragmentary Aulularia in our Palatine MSS. : 
 
 STROBILVS LICONIDES EVCLIO 
 
 SERVVS ADOLESCENS SENEX, 
 
 it is argued that Euclio must have figured in the lost part (the play breaks off 
 in our MSS. at v. 831) of this Scene and that therefore this Scene provided 
 the denouement of the play. Our theory shews that all that really was 
 directly transmitted from the Palatine recension was : 
 
 SERVVS ADOLESCENS SENEX. 
 
 Those objectors who insist on assigning to a minuscule archetype or even 
 to extant MSS. the interpolation of stalicio in the Casina and lampadiscvs 
 in the Cistellaria will, if they are consistent, insist on the possibility of senex 
 having really referred to Megadorus here and of evclio being a mere guess 
 of some scribe. 
 
 Our theory, while it admits the possibility of EVCLIO being right, requires, 
 however, at least one more Scene than this one for the fifth Act of the play, 
 a Scene namely (or Scenes) from whose text the names phedria (fe-) and 
 FiTODicvs were taken by the re- writer of the headings of IV vii and II vii. 
 
1 62 The Ancient Editions of Flautus. 
 
 in which only the red portion would be lacking, and confirms 
 the possibility of a rubricator's remissness having been the sole 
 and efficient cause °. 
 
 The Scene-headings of Plautus are not devoid of importance <■ 
 for the editor of the text. It will then be convenient to devote 
 the last part of this; section to an examination of some problems 
 by the new light of our hypothesis : 
 
 (i) The names of the sisters in the Stichus : — 
 I take it that the A* recension or some copy offered for the wife 
 of Epignomus the alternative names Panegyris and Philumena. Its 
 first Scene-heading (possibly also II ii, where the Palimpsest has 
 panegyris) had this form : 
 
 SeU PHILVMENA 
 PANEGYRIS PAMPHILA 
 
 MVLIERES • II • 
 
 Panegyris is the name adopted in the text (vv. 247, 331) of both 
 recensions. The alternative Philumena I would ascribe to the 
 'Revival' text, for it is patently an improvement of Panegyris, 
 suggested by the name of the sister Pamphila and her husband 
 Pamphilippus. The three names are admirably suitable to the 
 loving family that is so charmingly pourtrayed in the play. But 
 if the play was twice presented in Plautus' life-time, the new name 
 may be the poet's own 'second thought.' Whether P* offered 
 an alternative like A* or confined itself to the mention of one name 
 for the wife of Epignomus cannot be stated. Our theory of the 
 names in the Scene-headings of the Palatine MSS. shews that 
 the PANEGYRIS which appears in them has been taken from the 
 
 <= Another possibility, the origin of Scene-headings from miniatures, is 
 maintained for MSS. of Terence by Mr. J. Watson in an article (in Harvard 
 §tudies, Vol. XIV) which has just appeared. 
 
 •1 For example, we have seen that their conventional order follows the order 
 of the speakers' appearance in the Scene. In Mil. IV \s A P agree in putting 
 PLEVSICLES ADVLESCENS before PALAESTRIO SERVVS. So the words el ego 
 vos in V. 1 138 are by some assigned to Pleusicles, not Palaestrio (but cf. 1132 
 sqq.). After Epid. 189 something may have been lost. For the opening words 
 of the next line, continuo ut maritus fiat, are clearly the words of Apoecides, 
 whereas the Scene-heading in the Palimpsest, epidicvs periphanes apoecides, 
 if not a mere transposition (as in I i), shews that Periphanes took part in 
 the dialogue before Apoecides. The Palatine MSS. have the other order. 
 By our theory all that was actually transmitted in their qase was senes • II ■. 
 
Tin Ancient Edilions of Plautus. to3 
 
 text, tlie name that originally stood (whatever it may have been) 
 in the Scene-heading having been lost at an early period. The name 
 of the second sister does not occur in the text. The A''^ recension 
 offers PAMPHiLA without any alternative. The re-writer of the 
 Palatine Scene-headings thought that he had discovered the name 
 in the text at v. 284 : 
 
 Proinde ut decet amat virum suom ; nunc expedi Pinacium 
 
 (so the Palatine MSS. for expetit, nunc Pinaciiini), where the slave 
 Pinacium is of course the person mentioned. Hence the pinacivm 
 or (with confusion of D for P) dinacivm in our extant MSS. If we 
 bear in mind the true hypothesis of the transmission of the names 
 in the Palatine Scene-headings, we shall see the incorrectness of 
 the common view, viz. that the P^ recension named the sisters 
 PANEGYRis and DINACIVM, while in the A-\ recension they figured 
 as PHiLVMENA and pamphila. 
 
 (::) The names of the slaves in the Truculentus. 
 
 Renaissance MSS. give the name of the surly slave as 
 STRATiLAX ; and it is strange that their worthless authority 
 should have caused the acceptance of this name in the large 
 Teubner edition. It is patently a corruption of strabax (possibly 
 written strauax like uallio for ballio Pseud. IV ii in T) in some 
 Scene-heading of the Palatine MSS. Strabax is of course the 
 ' adulescens rusticus.' The Palimpsest, as we have seen, gives 
 the proper names, not the roles, in its Scene-headings ; and since it 
 offers at II ii 
 
 TRVCVLENTVS ASTAPin\-JI 
 
 followed by a blank line, in which we may suppose to have been 
 written in red pigment 
 
 SERWS ANCILLA, 
 
 I have little doubt that Truculentus was the actual name, and 
 that Astaphium's remark in v. 265 : 
 
 Nimis quidem hie truculentust, 
 
 was in the way of a pun, like the pun on the name Epidicus in 
 Epid. 25 or Pinacium in Stich. 271. 
 
 The other slave was called Cuamus, as appears from the Palatine 
 text (which in this play is in a woefully corrupt state) in w. 
 
104 "^^ Ancient Editions of Plautus. 
 
 583, 586, 702. The re-writer of the Palatine Scene-headings 
 believed it to be geta on the strength of v. 577 : 
 
 lubeo vos salvere. Et nos te, Geta ; quid agis ? ut vales ? 
 
 where all that is meant is ' a Getan slave.' 
 
 (On the real names of the tippling slave in the Miles and the man 
 in the travelling cloak (chlamydatus) in the Asinaria, see above, 
 p. 96). 
 
 (VII.) Errors common to the two Recensions. 
 
 The two copies, A and P, of two rival recensions, A* and P*, are 
 only ' blurred ' copies. One cause that has confused their outlines 
 is the inevitable tendency of scribes to make mistakes. The 
 immediate original of A has, we may be s,ure, by no means been 
 faithfully transcribed in A itself, and the remote archetype of A 
 is still less faithfully reproduced. The case oi P\% even worse. In 
 the parts for which we have not the evidence of T, all that we 
 can appeal to is the testimony of a minuscule archetype ; and 
 who can say how many errors have been made by the mediaeval 
 German monk (or monks) who penned it? Could we recover in 
 these parts the majuscule archetype, we should certainly find that 
 in scores of passages it had identically the same correct text as 
 
 A, where our MSS. {B, C, D, etc.) all have gone wrong, 
 through the error of this German monk-scribe. Here are some 
 examples which the newly found collation of T has revealed to 
 us: Pers. 536: mihi AP (pm.) ; 629 eveniant AP {conveniant) ; 
 Poen. 310 quia A P {qui) ; 472 quom AP{quo); 860 dignus qui 
 siet A P (pm.) 3977 punicast guggast homo A P (pm.) ; 1019 tu aliud 
 sapis A P (tua) ; 1036 (tu) om. A P ; 1204 addunt A P {am.). 
 
 And on the other hand a great deal of the apparent harmony of 
 A with our minuscule MSS. is equally specious. In Pseud. 1326 
 the mistake of reddi for redi is found in A. It did not appear 
 in the majuscule nor yet in the minuscule archetype, but it intruded 
 upon that late transcript which was the original of our MSS. C and 
 
 B. In Trin. 530 the same mistake, reddit for redit, is found in 
 A and the minuscule archetype. But how can we be sure that 
 it was also found in the majuscule archetype and did not first 
 intrude at the Carolingian period? Errors of this kind are at 
 all times a temptation to a scribe, and there is every possibility that 
 
The Ancient Editions of Plautus. 105 
 
 the scribe of A and the scribe of some copy of the other recen- 
 sion fell into them independently. We have therefore no right 
 to take for granted, as is generally done, that A* and P* exhibited 
 a consensus in such errors as Trin. 773 gererem for gerere rem. 
 Pseud. 98 libellae for libellai, Poen. 876 resistam for res sistam, 669 
 accurres for accures. It is extraordinary how many treatises on the 
 two recensions of Plautus have assumed that, because natural mis- 
 writings like these are found in our extant minuscule MSS., they 
 must have been present in the majuscule archetype, and even, 
 a still more dangerous inference, that they existed in some 
 imaginary original from which both A* and P^ were derived. A 
 much less natural miswriting, hamum for hamulum, has been made 
 in Stich. 289 independently by the scribe of the original of C 
 and D and by the scribe of A (or the original of A). The reading 
 of both the majuscule and the minuscule archetype (as of B) was 
 hamulum. Had B not retained the true form, we should have 
 imagined that hamum was the reading of P (cf. Pers. 572 anidum 
 for anellum). The discovery of the collation of T has revealed to us 
 a number of errors introduced into the text for the first time by 
 the scribe of the minuscule archetype. Great care therefore is 
 necessary in compiling a list of the passages in which A and this 
 archetype exhibit either on the one hand a divergence of read- 
 ing, or on the other a consensus in error. And even when we 
 have clear evidence for the reading of A and F. we have still 
 to assure ourselves whether A and P in this respect offer a faithful 
 or a blurred reflection of the two rival recensions from which they 
 have originated. 
 
 Only a little study of the critical apparatus of the large Teubner 
 edition of Plautus is needed to convince us how inevitable are 
 the corruptions which are so recklessly offered and so thoughtlessly 
 accepted as ' Errors common to the two Recensions.' I mean 
 corruptions like eveniat for evenat (Trin. 41), iit for iiii (Poen. 
 rz89 and passim), opinor for opino (Bacch. 487 and passim), illi 
 (dat.) for illic (Mil. 351, etc.), as well as illic (adv.) for illi, 
 ilium for illunc (Poen. 1302, etc.), possim for possiem (Pers. 319) 
 and sim for siem,possu7n iox potis {pote) sum (Pseud. 355), -ae {gen.) 
 for ai (Pseud. 98, etc.), and other misspellings like habeas for abeas 
 (Pseud. 393), scimus for simus (Pseud. 683), honestam for onustam 
 (Pseud. 1306), hostium for ostium (Most. 768). The newly found 
 evidence of T shews us how often such errors originated in the 
 
io6 The Ancient Editions of Plautus. 
 
 minuscule archetype and were not found in the majuscule, even 
 when they appear in A, e.g. Pers. 442 qum P, quin A B C D. 
 Similarly the evidence of B shews us when they are to be referred 
 to the scribe of the original of C and D and not to the minuscule 
 archetype, e.g. Trin. 371 tolerabilis A C D, tolerabis Pj Mil. 374 
 mihipossunt h.C'D, possunt mihi P; Men. 11^2 postquatn A CD, 
 post quern B. No argument whatever regarding the ancient recen- 
 sions Aa and Pa can be based on consensus in errors of this 
 description, even though such consensus could be established for A 
 and P. In Poen. 365 we have the express testimony of Gellius 
 (reproduced by Nonius) that Plautus wrote 7nea delicia. This 0. 
 Lat. unfamiliar form appears in the famihar guise meae deliciae in A 
 and P ; but it would be rash to assert that itieae deliciae was the 
 deliberate reading of the editor of one or other (or both) of the 
 rival recensions, and not a mere mistake committed separately 
 by transcribers of the text ^. 
 
 The full information which the new Corpus Glossariorum Latin- 
 orum now provides regarding the glossaries or ancient dictionaries 
 has facilitated the detection of apparent divergences in A and P 
 which are rather errors of transcription, due to the substitution 
 of a suprascript gloss for the actual word of the text, than varieties 
 of reading. Rogo, for example, is the stock interpretation of 0. Lat. 
 ore in the dictionaries of the Empire ; and so rogas, the reading of 
 A in Most. 682 (cf. P in Pers. 321, p. 73 above) : 
 
 bonum aequomque oras, 
 
 is not to be atttibuted to the editor of the recension which A 
 embodies, but merely to the error of a scribe who found in his 
 original : 
 
 rogas 
 bonum aequomque oras 
 
 ^ In the case of these O. Lat. forms which were systematically changed 
 by scribes at all periods, ancient and mediaeval, alike, e.g. illi (adv.), illic 
 (dat.), opino, etc., it seems to me to be the duty of an editor of Plautus to 
 warn his readers that although the familiar form (e.g. opinor) is offered by 
 the extant MSS., there is no real evidence that the unfamiliar form (e.g. opino) 
 did not stand in the immediate original of the MSS. In fact we may be said 
 to be more certain that Plautus used opino (examples are in Langen ' Beitraege,'. 
 p. 64) than that he used opinor (e.g. Bacch. 155). Of course opinaius sum 
 no more attests opinor than solitus sum attests ' soleor.' Cf. Most. 832 ludi- 
 feat (-tur A P) and perhaps Stich. 165 oboriunt (-tur A P). 
 
The Ancient Editions of Plautus. icti 
 
 and miscopied it as 
 
 bonum aequomque rogas. 
 
 Totus is similarly the stock explanation of O. Lat. perpes; and 
 so totam was in some original^ ol A written above perpete/n {-im) in 
 True. 278 : 
 
 noctem in stramentis pernoctare perpetim. 
 
 The transcriber mistook the suprascript word, not for a correction 
 (as in the line of the Mostellaria just quoted), but for an omission, 
 producing in A the unraetrical line : 
 
 noctem in stramentis pernoctare perpetim totam. 
 
 Similarly the Palatine archetype shewed Trin. 350 in this form : 
 
 Minus, pater ; sed civi inmuni [inmunifico] scis quid cantari 
 solet ? 
 
 Some examples of apparent divergence in A and F, which may be 
 really due to this cause, have been already mentioned in Part II 
 (pp. 58, 60, 63, 70 sqq.) Among other passages that may be named 
 in this connexion are : Merc. 300 benest A, bonum est P ; Pers. 408 
 periure A, iniure P ; Pseud. 43 impertit A, mittit P; 232 nihil 
 curassis A, bene curassis (if mis written for 7ie curassis) P; 417 
 anteveniat A, antecedat P; 901 fortiter A, firniiter P; 1142 ipsus 
 ipsuni A, ipsus coram P; Stich. 455 logis A, meis P; 523 ubi A, 
 si Pj Trin. E071 hie A, ipsus P; True. 261 in nostra domo h, 
 nostrae domi P; 363 puer A, mihi P, and perhaps Stich. 167 
 auditavi A, audivi saepe P (cf. Paul. Test. 28 M. auditavi, saepe 
 audivi). 
 
 Now precisely the same practice of adding suprascript glosses may 
 have occasionally led to a consensus oi AF'm error. In Poen. 1317, 
 for example, cur {qur) non, the reading of A and F, may not be the 
 original reading of either recension, but may have found its way into 
 A and F through the suprascription of a gloss over the word in the 
 text, producing the apparently unmetrical line : 
 
 qur non adhibuisti, dum istaec loquere, tympanum ? 
 
 ' A glance at the apograph of the Palimpsest will shew how inevitable such 
 errors must have been. There is nothing to indicate whether a suprascript 
 
 A 
 
 letter is meant as a substitute (e.g. copiam Stich. 351) or an addition (e.g. 
 c u 
 
 SIRE Mil. 477, CATELLA Mil. 603). 
 
jog The Ancient Editions of Plautus. 
 
 Bothe supposed the original reading to have been qurm, in 
 support of which might be quoted the numerous lines m which 
 «g is glossed s by nan ; Geppert made it quin, and this reading 
 might be supported by Pseud. 501, where the Palatine MSS. {An.l.) 
 offer nir nan but editors print quin, and by the Bembine scholiast's 
 explanation of quintii as cur nan in Ter. Eun. 328. 
 
 When in Paulus' epitome of Festus' dictionary (51. 5 M.) we 
 read that custodela was the O. Lat. form of custodia, we under- 
 stand bow it is that our Palatine MSS. offer custodia instead of 
 custodela in the four lines in which the word occurs (Merc. 233, 
 Most. 406, Rud. 625, 696). None of these lines has been preserved 
 in our Palimpsest fragment ; but there would be nothing remarkable 
 in a consensus oi A F in this form ^. In a number of the cases 
 mentioned above, where unfamiliar early forms have been modern- 
 ized in our MSS., e.g. eveniat for evenat, a suprascript gloss may 
 have been the real agency at work. Cas. 702 is an instructive 
 example of how glosses marred the two texts : 
 
 Ut nubat mihi — illud quidem volebam, 
 Nostro vilico ; 
 
 for the peculiar phrase illud quidem volebam, 'I meant to say,' 
 has brought glosses, but fortunately different glosses, into A (dicere 
 volebam) and P {volebam sed non). There was an even chance here 
 of the same gloss being foisted upon the text in both MSS. 
 
 Other possibilities of specious, not real, consensus in error are 
 more difficult to determine v — 
 
 In True. 383 : 
 
 Quod tu hie me absente novi negoti gesseris ? 
 both scribes have made the same mistake of transposition, but 
 
 6 E.g. nevis by nonvis, nevull by nonvuli (Nonius 144 M. ; nevult pro non- 
 vult). Thus at Trin. 361 nevoli (F) has become nonvolt in A. Priscian attests 
 ne (? ne) viderit for Mil. 149 : ut quod viderit ne viderit, but both A and 
 the extant minuscule MSS. have non viderit. 
 
 '^ The similarity too of L to I in Capital Script might possibly cause custodela 
 to be mistaken for a supposed custodeia and altered in the usual fashion (see 
 part IX, below) to custodia. I do not know of any sufficient grounds for believing 
 custodia to have been an O. Lat. form. It is true that the word aio is glossed 
 by dico (cf. the Thesaurus Glossarum s.v.) ; but the substitution of, e.g., dice- 
 bant for aiebant at Men. 1046 in our minuscule MSS. is perhaps more easily 
 explained by the similarity of a to if in early minuscule script. From diebant 
 to dicebant was but a step. 
 
The Ancient Editions of Plautus. 109 
 
 fortunately their deviation from their original has taken different 
 directions, ^'s transposition is me hie absente, while F's is hie 
 absente me. This suggests the possibility of an occasional coinci- 
 dence' in transposition. In Mil. 727-9 : 
 
 Sicut merci pretium statuit qui est probus agoranomus : 
 Quae probast mars, pretium ei statuit, pro virtute ut veneat, 
 Quae improbast, pro mercis vitio dominum pretio pauperat, 
 
 the similarity of the clauses led to omission (but fortunately not the 
 same omission) in P {am. qui est — statuit) and in A (pm. mers — 
 improbast). Another example will be found in Poen. 389 {om. C D) 
 and 390 (om. A^). But the most irresistible of all temptations to 
 a scribe was the temptation to Haplography, to write a repeated 
 word or syllable once instead of twice. It would hardly be rash to 
 assert that there is not a repeated word or syllable in a line of 
 Plautus which in some MS. or other has not come to suffer hap- 
 lography. If Plautus wrote : 
 
 Pseud. 443 'q Zev, Zev, quam pauci estis homines commodi ! 
 Stich. 384 lam, iam non facio auctionem : mi obtigit hereditas, 
 Poen. 1272 '^ Cur, cur numero estis mortui, hoc exemplo ut 
 pingeretis ? 
 
 Poen. 969 Cretast, cretast profecto horum hominum oratio, 
 Cas. prol. 72 Et hie in nostra terra in terra Apulia, 
 
 we have no right to ascribe the haplography in A and P to 
 a common original of the two MSS. Such a mistake would with 
 the utmost ease be made independently by different scribes. 
 
 Of late there has been a tendency to minimize the indications of 
 different origin of A* and P*. Cases of ' mixture ' of text have been 
 put forward as a proof that both recensions came from some original 
 
 i Men. 201 is not a case in point. See above, p. 79. In Most. 1043 
 ut filium suum arcesserem (A P) instead of filium ut might almost be called 
 an ' inevitable ' transposition. 
 
 ^ This line is quoted by Festus in the same form (at least in our single 
 MS. of Festus) as in the Plautus MSS. Whether Festus deliberately dis- 
 carded the first word of the line is not clear. In any case this evidence for 
 the retention of final d in the Abl. numerod (beside exemplo 1) seems to me 
 very weak indeed. In Cas. 646 our MS. of Festus has (like our minuscule 
 MSS. of Plautus but unlike the Palimpsest) omitted me through Haplography 
 in the phrase mea me ancilla, a phrase which would be a pitfall for any 
 scribe at any time. 
 
no The Ancient Editions of Plautus. 
 
 'variorum ' edition of the collected plays, ah edition' crammed with 
 variant readings; and the divergence Of the "two recensions has 
 been referred to the choice by transcribers, now of the accepted 
 reading, now of the variant. It seems to me that in most cases 
 a much more natural and likely explanation of the ' mixture ' is that 
 the reading of one recension came in course of time to be entered 
 in the margin of a copy of the other recension, and from there 
 found its way into the text. In fact we can trace the same process 
 still going on in A and in P themselves (or their originals). In 
 Pseud. 975 impiwn, the reading of P*, may have been entered 
 in the margin ' of the original of A, whose reading is impuruiu, just 
 as in Pseud. 880 what was a marginal (or interlinear) variant in the 
 original has retained a place, but not its right place, in A (iu 
 illos Pa and A-text, tuos A-margin). In Pseud. 1207 abduceret, the 
 reading of A*, is entered in the margin of P, whose reading is 
 arcesseret, and so on. And yet some advocates of unity of 
 origin for the two recensions go so far as to ascribe certain 
 apparent instances of consensus of A P 'va. error to the existence 
 of holes in the pages of this supposed original and to estimate 
 the number of lines which each iniaginary page must have con- 
 tained. This is surely to forget that A and P are two out of a vast 
 number of ancient copies of Plautus, belonging to different parts of 
 the Roman world, with as little likehhood of being closely related 
 to each other as two copies of Shakespeare, published, let us say, at 
 the interval of a century or half a century, the one at Glasgow 
 and the other at Melbourne. The great argument used by the 
 supporters of such theories is the consensus in error of the two 
 ancient codices. By going through the apparatus criticus of an 
 edition of Plautus and taking note of each and every occurrence 
 where a reading marked '^/" has not been printed in the text, 
 they are able to confront us with an imposing list of lines in which 
 the reading of A and of P is the same and apparently erroneous. 
 The greater part of the examples consists of trivial miswritings, such 
 as we have indicated in the preceding pages. These may be struck 
 off the lists at once, as proving nothing. As regards the remaining 
 examples, we find that year by year the list grows smaller; for, 
 
 ' The P* version is : legerupam impium peiiurum atque improbum. A has : 
 legirupum inpurum peiiurum atque impium, the marginal variant having ousted 
 improbum. But impium may have been a marginal correction of inpurum. 
 
The Ancient Editions of Plauius. in 
 
 as our knowledge of Plautine diction and prosody grows, we 
 recognise the correctness of this or that reading supported by 
 the consensus of A and P. Before 1892, when Prof. Skutsch 
 published the first volume of his Forschungen, with its interesting 
 discovery of the suppression (before an initial consonant) of final 
 e in ille, nempe, inde, proinde, etc., in Plautus' verse, just as in all 
 literature in atque (ac), neqiie (nee), neueineu'), lines like Stich. 175: 
 
 Quia inde iam a pausillo piiero ridiculils fui 
 
 used to form a considerable part of these lists. Men. 167 will, I 
 presume, be omitted from them, now that the same scholar has 
 shewn us that olfactare has its old pronunciation olefadare : 
 
 Silmmum olefact^re oportet. 
 
 The whole history of Plautine textual criticism in the last half- 
 century has taught us that truth lies, if anywhere, in the con- 
 sensus of A and P, and that the danger in tampering with a reading 
 supported by A alone or P alone is not nearly so great as the 
 danger of discarding the united testimony of the ' two witnesses.' 
 No judge will arrive at a correct verdict who does not weigh the 
 evidence. The evidence of A P must outweigh the single evidence 
 of P. The practice of emending lines of Plautus without stating 
 whether the reading which is impugned rests on the authority of 
 P only or of A only, or of A and P combined, obscures the 
 conditions of the problem to the reader and encourages the writer 
 to reject genuine readings too hastily. 
 
 The whole weight of tradition supports the reading penitus'^ 
 (in its original sense of 'from within') in Pseud. 132 : 
 
 atque ipse egreditur p&nitus {intus edd.), periuri caput. 
 
 Are we as much justified in substituting intus in this line as we 
 might be in a line for which we had no better evidence than the 
 Carolingian archetype? In Stich. 704, does not the consensus 
 of AP in the reading in lecticis rather point to some Plautine 
 coinage like inledtce (adv.) of the type of accubuo (True. 422) ? 
 
 Stich. Nimum lepide in mdntem uenit : pdtius quam in subsdlio 
 
 ^ Pmilus is normally an Adj. in Plautus, 'inner;' its adverbial sense, 
 'within,' is not Plautine. But Comparative Philology shews us that the 
 original sense of the Adverb must have been ' from within.' 
 
112 The Ancient Editions of Plautus. 
 
 Cynice hie accipimur quam inlectice {in lectis edd.) Sag. Immo 
 enim nitnio hie dulcius. 
 
 Must we not retain their reading stultitiis in Trin. S09. and give de 
 the sense of ' after' or ' in consequence of ' (as in Cas. 415, etc.)? 
 nam is (sc. ager) de stultitiis {diuitiis edd.) meis 
 soMs superfit pradter uitam r^licuos. 
 
 Should we disregard their testimony to the old trisyllabic form 
 of er£;o (as iurtgo of iurgo, purigo oi purgo) in Poen. 1051 ? 
 patrituser <(e> go hdspes Antidamds fuit ? 
 
 Should we ignore their indication of an O. Lat. fortasse est like 
 necesse est in Poen. 1004-5 '' 
 
 Mil. Fortasse medicos n6s esse arbitrarier. 
 Agor. Si ^st (Si ita est edd^, ,^nega esse : nolo ego errare h6s- 
 pitem. 
 
 And is the phrase in ius uos uolo so impossible that we must 
 suppose both A and P to be in error in Poen. 1225 ? 
 
 quid istic ? quod faciilndumst cur non agimus ? in ius uos uolo 
 (uoco edd.). 
 
 Certainly, if we consider the number of lines supported by the 
 consensus of A and P, whose reading has been justified through 
 advance in our knowledge of Plautus, we shall be inclined to 
 predict that nearly every line so supported will prove to be free 
 from error, unless there be an error into which A and P have 
 fallen independently, like the ' modernizing ' of an archaic form, 
 e.g. ridiculosissimos for ridiculissimos (Stich. 389), Haplography, 
 or some other equally obvious miswriting, such as illorim for 
 Iliorum (Bacch. 951), atque euoca for atque uoca (Poen. 11 16), 
 optumi maxumi for opt. maxume (Men. 573), festiua mulier for 
 festiuam mulier (Mil. 591). But to argue on the other side is 
 much more easy, for one has ready to hand all the apparent 
 instances of consensus in error which have not yet received 
 their explanation ; and, although the number available is dimin- 
 ishing steadily, there still remains a sufificient quantity to provide 
 a respectable case. A large list of instances is furnished by lines 
 which shew Hiatus. These, however, lose their force, if we are 
 to believe (and I do not see how we can venture to disbelieve) 
 
The Ancient Editions of Plautus. t 1 3 
 
 Cicero's express statement that the early poets made extensive 
 use of this license. A discussion of the limits within which we 
 may suppose Plautus to have used it I reserve for Part VIII. 
 
 Even if real cases of consensus in error, these lines with 
 Hiatus would hardly justify the theory of so close a relation 
 between A^ and P^ as is assumed. Cicero's contemporaries 
 regarded Hiatus as a feature of Plautine verse, and a ' versus 
 hians ' would be accepted without question by all editors of 
 Plautus in Cicero's time as well as later. 
 
 And certainly on the face of it it seems likely enough that there 
 would be other mistakes found in all of even the earliest editions 
 (whether of single plays or a collection), like the misprints that were 
 handed down through successive editions of Shakespeare" or any 
 other poet. But it must be confessed that it is extremely difficult 
 to get completely satisfactory instances. Take, for example, Epid. 
 568 (I include the following couplet and the Scene-heading) : 
 
 PERiPHANES. lube Telestidem hue prodire filiam ante aedis 
 
 meam, 
 Ut suam videat matrem. philippa. Remigrat animus nunc 
 
 demum mihi. 
 
 (Act IV Sc. ii) ACROPOLisiis periphanes philippa 
 
 FIDICINA SENEX MVLIER 
 
 ACROPOLiSTis. Quid est, pater, quod me excivisti ante aedis ? 
 
 In this play the ' fidicina ' Acropolistis has passed as Periphanes' 
 daughter, under the name Telestis. Her father of course speaks 
 of her as Telestis, while in the Scene-heading and throughout the 
 dialogues she naturally figures as Acropolistis. Both A and P offer 
 erroneously Acropolistidem in v. 568. And yet surely any scribe 
 or corrector would suspect Telestidem in this line to be an error, 
 since it is followed by acropolistis in the Scene-heading and 
 ACROPOLISTIS (in some contracted form) through the ensuing dia- 
 logue. We can hardly venture to assert that Acropolistidem stood 
 in the immediate original of A or in the Carolingian archetype 
 of the minuscule MSS., much less that it stood in the recensions A* 
 and Pa. 
 
 Or take Epid. 455 with centones farcias for centones sarcias. In 
 minuscule script s and / are so persistently confused that one 
 transcript at least is fairly sure to offer /(?rj {oxftiit or fit oifalsus) 
 
 1 
 
1 1 4 The Ancient Editions of Platttus. 
 
 wliere the original has sors (or suit or sit or salsus). We cannot 
 with confidence ascribe /am«i even to the minuscule archetype. 
 As regards A there would not be the same ' inevitableness ' of 
 error. But the confusion of similar words, e.g. capiam and copiam, 
 (see above, p. 107), is so common as to prevent us from referring 
 with any confidence the error to the A* recension. 
 
 Or take Merc. 598, where it appears as though two lines had 
 been inserted in A as well as P (but only the beginnings of the 
 lines are legible in .4) which belong to another part of the play, 
 vv. 842-3 (but the leaves of A containing this part have been 
 lost) : 
 
 Divom atque hominum quae speratrix atque era eadem es 
 
 hominibus, 
 Spem speratam quom obtulisti banc mihi, tibi grates ago. 
 
 It is impossible to say that these lines are out of keeping with 
 the passage at 598 sqq., for Charinus, the speaker, is represented 
 as in a feverish mood, passing at a moment from hope to despair : 
 
 Sed isne est quem currentem video ? ipsus est, ibo obviam. 
 Divom atque hominum quae speratrix atque era eadem es 
 
 hominibus, 
 Spem (spes A') speratam quom obtulisti banc mihi, tibi grates 
 
 ago. 
 Nunc, quod restat, ei ! disperii : voltus neutiquam huius 
 
 placet : 
 Tristis incedit, — pectus ardet, haereo, — quassat caput. 
 Eutyche ! Eu, Charine Priusquam recipias anhelitum, 
 Uno verbo eloquere : ubi ego sum ? hicine an apud mortuos ? 
 If Plautus did not put the couplet at both places p, then its 
 
 P The recurrence of z^ sentiment in a drama may be, like the recurrence of 
 a ' leitmotif,' in a piece of music, a touch of the composer's art. How effective 
 the exit of Stichus at the end of Act III Scene i is made by the repetition 
 of the same sentiment as had probably brought him a round of applause earlier 
 in the Scene (v. 436) ! Editors of Plautus seem to me to tolerate it when 
 the words are not identical (e.g. Bacch. 498 and 380 ; Men. 810 and 777, 
 IQ31 and 1 148 ; cf. Langen's Plautinische Studien, and Kellermann's recent dis- 
 sertation), but to condemn it when the line is repeated unchanged. The 
 Pseudolus curiously offers four examples of identical recurrence (116=1073, 
 381 = 600, 409 = 788, 485 = 527), but we have not the evidence of .^ in the 
 case of the last two (as indeed we have not in Merc. 842) to shew us whether 
 the recurrence was a feature of A as well as of P? We must not forget 
 Horace's sneer at the ' non astrictus soccus ' of Plautus. 
 
The Ancient Editions of Plautus. irj" 
 
 insertion (here or later) was probably a feature of the ' Revival 
 text. 
 
 Or take Poen. 474 beginning volaticorum. The ' nota personae ' 
 prefixed to the line in both texts was E {i.e. Lycus). What was 
 more natural than that Evolaticorum should be suggested ? This is 
 the reading of the Palatine MSS. in which this form of error is very 
 frequent indeed (see the instances quoted on p. 92 above). 
 
 It was not in the original of A apparently, but the scribe of A 
 momentarily fell into the trap and wrote evol — only to correct 
 himself the next moment by expunging the E. The best example 1 
 that I can find of an error shared by A and P like the misprints 
 shared in early editions of Shakespeare is Poen. 670 : 
 
 Immo ut ipse nobis dixit, quo accures magis, 
 Trecentos nummos Philippos ^oxizX praes ibi\prae sibi), 
 
 for one can imagine praes ibi to be continued without suspicion 
 in successive transcriptions, and can accept the conjecture praesidi 
 with some confidence. And yet caution rather than confidence 
 in departing from the MSS. is the lesson of Plautine textual 
 criticism during the last fifty years. With what confidence did 
 editors of the last generation reject CMo (adj.) in Cure. 78 and 
 cor in Poen. 388 ! But it only needed a full investigation ot 
 Plautus' scansion of such final syllables as originally ended in 
 two consonants, e.g. corid), lac{t), miles{s), dives{s), to establish 
 once for all the correctness of the manuscript reading. And 
 a single paper by Buecheler in the Rheinisches Museum was 
 enough to confirm'' Chto in Cure. 78, Poen. 699, Pellaeo in Asin. 
 333. etc. 
 
 1 Those who believe in ' adscript parallel passages ' will find a better example 
 in Stich. 282 (cf.-Leo ad loc. But see Seyffert in Berl, Phil, Woch. xvi 284.) - 
 
 ■■ This field of enquiry is by no means exhausted. A full discussion of the 
 subject would be of great service. The instances adduced by Buecheler of 
 ' vocalis anceps ante vocalem ' in Early Latin like Classical ilttus may prob- 
 ably be multiplied. Fius, pietas seem likely to be Plautine as well as pms, 
 pletas. Since froprius is a compound of pro and privus, the old scansion 
 must have been proprlus, and this scansion has strong claims in Capt. 862, &c. 
 Editors who reject _/?«/ (Men. 352? Mil. 1054?) and substitute fuat can appeal 
 to the Glossaries, which regularly gloss the older verb by the other, and to 
 passages like Pseud. 1029 where fuant, retained in the minuscule archetype, 
 became _/?a»; in the original of C and D. 
 
 I 2 
 
1 1 6 The Ancient Editions of Plautus. 
 
 Since the publication of Leo's treatise on the Song-Cantica of 
 Plautus no editor exhibits 'confidence' in departing from. the 
 reading of the MSS. in these portions of the plays. A com- 
 parison of the Cantica in Leo's edition and in the large Teubner 
 edition will shew the reason of this change of mood. 
 
 In fact the whole line of advance in our knowledge of Plautus' 
 text has been steadily in the one direction of the re-establishment 
 of the readings common to A and P. It may be that criticism 
 has gone as far in this direction as it is destined to go, and that 
 another movement is now necessary ; but I must confess that I see 
 no indication of this. Our knowledge of Plautine Accidence and 
 Syntax and Prosody and Metre is still progressive, and success lies 
 in the same direction now as ever. An editor therefore should 
 exercise great caution in departing from the united reading of 
 A F, unless there is clear possibility of the two scribes having fallen 
 independently into the same error. For no thoroughly satisfactory 
 evidence has as yet been produced of errors common to the two 
 ancient recensions. 
 
 For the sake of completeness I add the more noticeable 
 examples adduced in the most recent list : 
 
 Cas. 625 yi\^ fadis lox factu looks like an example to be set 
 beside Poen. 670. 
 
 Poen. 1 168 
 Sed eccas video ipsas. Haecine meae sunt filiae ? 
 Quantae e quantillis iam sunt factae ? Scin quid est ? 
 Thraecae sunt : in celonem {sunt celumne P) sustolli solenl. 
 
 There seems no reason for re-writing the last line. It is pre- 
 served correctly by A. The preposition has the sense of 'like,' 
 ' after the manner of,' 
 
 Mil. 1419 Di tibi bene faciant semper, quom advocatus mihi 
 bene's (mihi benest A, bene mihi es P). 
 
 This reading, presumably common to the two recensions, should 
 not be altered (see Seyflfert in Berl. Phil. Woch., xvi, 234). 
 
 Stich. 620 Tantillum loculi ubi catellus cubet, id mi sat e rest 
 loci (res est uivid A). 
 
 The same remark applies to this line (cf. Seyffert ibid.). 
 
The Ancient Editions of Plautus. 117 
 
 Poen. 331 
 
 primum prima salva sis, 
 Et secunda tu insecundo salve in pretio ; tertia 
 Salve extra pretium. 
 
 Insecundo (cf. Auct. ad Herenn. IV. 56) may be Gerundive of 
 insequor as secundo of sequor. 
 
 Cas. 571 Rogitare oportet prius et contarier. 
 
 Prlus, the original scansion of the word, seems to be found 
 elsewhere in Plautus (Cas. 839, Bacch. 932, etc.). Percontor pre- 
 supposes the simple verb cantor, lit. ' to use a punt-pole,' hence ' to 
 make enquiry.' 
 
 Stich. 223 (in the parasite's imaginary auction). 
 
 Adeste sultis, praeda erit praesentium, 
 Logos ridiculos vendo,'age licemini. 
 Qui cena poscit ? ecqui poscit prandio ? 
 (Hercules te amabit) — prandio, cena tibi. 
 Ehem, adnuistin ? etc. 
 
 Hercules te amabit is a most natural parenthetical exclamation 
 to a supposed bidder. Of course the dactyl in the first foot of 
 the line is perfectly normal. 
 
 Stich. 243 The exclamation eu ecastor seems to be ' extra 
 metrum,' like attat in Cas. 619, &c., &c. 
 
 Pseud. 306 Non est iustus quisquam amator nisi qui perpetual 
 data. 
 
 Justus amator, ' a normal lover,' ' a lover worthy of the name,' 
 seems perfectly correct. 
 
 Pseud. 442 Idne tu mirare, si patrissat filius ? 
 Idne tu, whether it should be scanned idne tH or idn' tii, need 
 not be altered. 
 
 Mil. 254 vera ut esse credat quae mentibitur. 
 
 The corrector of B writes mentibimur. But I think we have here 
 an O. Lat. construction, the Accusative with the Impersonal 
 Passive (see above, pp. 10, 49). 
 
ii8 The Ancient Editions of Pl'aittus. 
 
 Of the rest some readily admit of emendation by the removal of 
 one ' inevitable ' error, e.g. Most. 586 : 
 
 Reddeturne igitur faenus? Reddet (-/«?■ A P), nunc abi, 
 
 while in others there is a choice between two possible errors of this 
 type: 
 
 e.g. Most. 984, where it is doubtful whether quaesitum has 
 become quaestum or pote siet has become potest; or Stich. 357, 
 where some print nisi . , venturi si sunt (si om. A P), others lectos 
 insternite {st. A P). 
 
 Others again still await satisfactory emendation : 
 e.g. Pseud. 627 (a trochaic Septenarius) : 
 Ballionis euro, argentum accepto, expense et quoi debet date. 
 
 Should a ' pet-form ' (e.g. Balh') replace Ballionis, as in Most. 560 
 Tranium is employed for the usual Tranionem ? Or has there been 
 ' mixture ' of two versions (i) accepto et quoi debet dato, (2) acceptor 
 expenso et dato? If so, I would conjecture that the second is 
 the ' genuine ' version and was the reading of Aa, while the first is 
 the ' Revival ' version. 
 
 (VIII.) Hiatus^ 
 
 Before the time of Ritschl most editors of Plautus felt themselves 
 bound to accept almost every instance of Hiatus that was forced 
 upon them by the consensus of the then available MSS. In all the 
 MSS., for example, Trin. 18, an Iambic Senarius, appeared in this 
 form : 
 
 huic nomen Graece est Thensauro fabulae, 
 
 with hiatus between Graece and est ; the second hemistich of Pseud. 
 375, a trochaic Septenarius, in this form : 
 
 facere oiiScium meum, 
 
 s A full collection of all conceivable instances in Plautus will be found in 
 Maurenbrecher 'Hiatus und Verschleifung, ' a book, however," whidi must be 
 called ' something between a hindrance and a help : ' a help, because the col- 
 lection is full, a hindrance, because it is indiscriminate. Why should a mere 
 scribal error, which would be recognised as such in any other line, be. added 
 to the list of ' Instances of Hiatus presented by the Traditional Text,' if it 
 occurs in a line in which one word ends in a vowel or m and the next begins 
 with a vowel or ^ ? . 
 
The Ancient Edttiotis of Plautus, 119 
 
 with hiatus between facere and officium. And the acceptance of 
 these ' versus hiantes ' was justified by an appeal to Cicero's words 
 in the Orator (XLV 152): sed Graeci viderint ; nobis, ne si 
 cupiamus quidem, distrahere voces conceditur. Indicant orationes 
 illae ipsae horridulae Catonis, indicant omnes poetae praeter eos 
 qui, ut versum facerent (i.e. ' through metrical exigencies '), saepe 
 hiabant, ut Naevius : 
 
 uos, qui i accolitis Histrutn fluuium atque algidam, 
 
 et ibidem : 
 
 quam niimquam uobis Graii | atque barbari, 
 
 at Ennius semel (saepe edd.) : 
 
 Scipio I inuicte, 
 
 et quidem nos : 
 
 hoc motu radiantis Etesiae | in vada ponti. 
 
 Hoc idem nostri saepius non tulissent, quod Graeci laudare 
 etiam solent. 
 
 Ritschl's production of tlie evidence of the Ambrosian Palim- 
 psest (A) changed the aspect of the case. It was found that the 
 consensus of the MSS. in these two lines, and in others of the kind, 
 was merely the result of their derivation from a common archetype 
 (/*), a MS. probably of Charlemagne's time or later. The scribe of 
 this minuscule archetype (or its original) had transposed the words 
 nomen and Graece in Trin. 18 and had omitted the Pronoun me 
 in Pseud. 375. In the fourth-century Palimpsest, which precedes, 
 P in age by many centuries, these lines appear in their true form : 
 
 huic Gradce nomen est Thensauro fibulae, 
 
 and 
 
 facere me officium meum. 
 
 Now transposition of words is one of the commonest errors of 
 scribes ; and the omission of small words, especially such as are 
 not necessary to the sense of the sentence, is an error to which 
 scribes of Plautus' Comedies are peculiarly liable, for this comedian 
 delights in the otiose use of Pronouns {ego, tu, hie, etc.). Particles 
 (/(?/, qu% etc.), and the like. How then is an editor to decide 
 in the numerous passages for which the evidence of the fragmentary 
 Palimpsest is not available, whether a ' versus hians ' retains the 
 
I20 
 
 The Ancient Editions of Flautusr 
 
 correct version of Plautus or owes its abnormal form to a mere 
 scribal error ? This is the most difficult of all the problems which 
 an editor of Plautus has to face ; and it is one of constant occur- 
 rence, for in the Palimpsest the whole of the Amphitruo, Asinaria, 
 Aulularia and Curculio is missing, nearly the whole of the Captivi 
 (and Vidularia), and a great part of the other plays; the best 
 preserved are the Stichus, Persa, Poenulus, Pseudolus and Trinum- 
 mus. For the greater part of the plays we have only the evidence 
 of the minuscule Palatine archetype to appeal to, save that lines 
 here and there are preserved in quotations by ancient Gram- 
 marians, such as Nonius, Festus, Charisius, Priscian. 
 
 That Hiatus was a feature of the older poetry is a fact which 
 cannot be denied. Cicero's * statement {ut versum facerent, saepe 
 hiabant) is conclusive on this point; and it is supported by the 
 evidence (so far as that goes) of Saturnian Verse, in which any 
 final long vowel or final syllable in -m seems normally to remain 
 unelided before an initial vowel. Thus uirginem oraret and aut ibi 
 ommentans seem to be six-syllabled hemistichs of the same type 
 as Naeuio poetae (for details see Amer. Journ. Phil, xiv 309). And 
 we have Hiatus expressly attested by Priscian in a couplet of 
 Ennius (Ann. 332 V., 354 M.): 
 
 insignita fere turn milia militiim | octo 
 duxit delectos, bellum superare potentes. 
 
 Saturnian versification transmitted to the early Latin adaptations of 
 Greek Metre not merely its use of Alliteration and something of its 
 regard for Accent, but also, in greater or less degree, its toler- 
 ance of Hiatus. Ritschl's uncompromising attitude of hostility 
 to Hiatus is now given up by all editors of Plautus. No one now 
 believes that Plautus, while readily admitting Elision at the end 
 of the first hemistich of long lines, did not also readily admit 
 Hiatus. Lines like : 
 
 Men. 778 n^scio quid uos udlitati | dstis inter uds duos, 
 Amph. 208 rediicturum, abiturds agro | Argiuos, pacem atque 
 
 <5tium, 
 Mil. 1226 namque ddepol uix fuit cdpia | adeiindi atque 
 
 impetrandi, 
 
 ' Presumably not originated by Cicero, but rather the common verdict of 
 literary antiquarians of that time. 
 
The Ancient Editions of Plaictus. 121 
 
 are no longer tampered with by editors ; for it is acknowledged 
 that the pause that followed the utterance of the first half of these 
 lines justifies Hiatus, in the same way that it justifies the lengthen- 
 ing of a short syllable in lines like Mil. 1226 (just quoted), 
 
 Asin. 634 quas h6die adulescens Diabolus ipsi daturus dixit 
 
 At the same time no two editors are agreed upon the exact limits 
 observed by Plautus in his tolerance of Hiatus, upon the precise 
 extent to which Cicero's statement, ut versum faceret, hiabat, applies 
 to this early poet, the earliest whose works have, in any measure of 
 completeness, been preserved. 
 
 It is clear that in the investigation of this subject we must be 
 careful in the selection of our material. We must confine ourselves, 
 for the first at least, to lines whose text depends on something 
 better than the evidence of one archetype. It is useless to compile 
 lists of instances of Hiatus which have no stronger evidence than 
 the Carolingian archetype of the Palatine MSS. ; what assurance 
 have we that they are not all of the same type as the examples 
 quoted above, Graece \ est and facere \ officium, and that' the 
 Ambrosian Palimpsest (^), if we could discover or decipher its 
 version of the lines, would not present them in a different and 
 more correct form " ? Lines which are supported by the evidence 
 of both P and A, or of P and some ancient Grammarian, have 
 far stronger claims to our credence. They are likely to be either 
 the 'ipsa verba' of Plautus, or at least the version that passed 
 current as such in the early centuries of our era. 
 
 The method therefore that I propose to follow is this, to use 
 as material only those ' versus hiantes ' whose text is strongly 
 estabhshed and to examine how many of the types of Hiatus which 
 they exhibit are justified from later poetry, whether by its occa- 
 sional use of the same licence or by its patent avoidance of Elision 
 in such cases. For I take it that Latin Poetry flowed in a con- 
 tinuous stream from Livius Andronicus to Virgil, and that the 
 prosody of one generation was never wholly alien from that of 
 the generation that preceded it. When we find Catullus (xcvii i) 
 
 M For example, one of the lines used as evidence that Hiatus was permis- 
 sible before initial h is Rud. 1 1 : 
 
 qui facta | hominum, mores, pietatem et fidem (/" A n.L), 
 where, it now appears, the Codex Turnebi had moresque. 
 
J 2 2 The Ancient Editions of Plautus. 
 
 and Virgil {Eel. viii io8) admitting ita me di ament and an qui 
 amant, we cannot disconnect these scansions from the forms in 
 which these phrases normally (not occasionally) appear in Plautus, 
 e.g.: 
 
 Trin. 242 nam qui amat quodamat, etc., 
 
 Cist. 280 nam quf amant stulte atque inmodeste atque inprobe, 
 
 Merc. 744 nam qui amat quod amat, etc. 
 
 Cure. 142 (anapaestic) Palindre. Edepol qui amat, si eget, etc. 
 
 Pseud. 943 (anapaestic) Ita m^ di ament — Ita ndn facient, etc. 
 
 With Plautus' lines before us we can no longer regard the hiatus of 
 di, qui in these phrases in Catullus' and Virgil's lines as a mere 
 artificial imitation of Greek metrical licences, like Actaeo \ 
 Aracyntho. Clearly the phrases were pronounced by Roman lips 
 in this way ; and the Early Dramatists, who aim at the repro- 
 duction of the language of actual, everyday life, felt no scruples 
 in giving them this scansion' in their verse. 
 
 The Prosodic Hiatus of monosyllables ending in a long vowel or 
 m persists so determinedly in Republican poetry and even in 
 the more colloquial part of Horace's writings (the Satires and 
 Epistles'), that Ritschl himself was forced to allow it a place in 
 Plautus. Its exact limits in Plautine Verse are not easy to define ^ 
 On the one hand we see a clear tendency to avoid by this means 
 the total absorption by Elision of an emphatic monosyllable, e.g. 
 
 Tii I erus es, tu sdruom quaere, tii salueto, tii uale. 
 
 On the other, we see Enclitics or subordinate words joined with 
 a neighbouring word into a word-group and thus avoiding elision, 
 e.g. qui-amant, qui-homo like deamatit, etc. More questioned is the 
 Prosodic Hiatus with Iambic and Cretic words, as in Virgil's vale, 
 vale, I inquit (cf. Ovid Met. iii 501), and insulae \ lonio in magno, 
 Lucretius' remigi \ oblitae pennarum, Catullus' uno in lectulo, \ erudi- 
 tuli ambo (v.l. lecticulo), Ennius' Scipio \ invicte, and milia militiim \ 
 octo (see above) ; for many scholars believe these to be imitations of 
 Greek scansions like 'iatj^rai ^/iap (pronounced, according to Blass, 
 
 ^ That the unelided monosyllable occasionally constitutes one 'mora,' e.g. 
 de hSrdeo (Asin. 706, P, Nonius) like quanMrem, dehSrtor, i kdc (Pseud. 1331, 
 A P), is now generally admitted ; but the matter has not yet been thoroughly 
 investigated. The change of me, te in Hiatus of this kind to med, ted may 
 not be invariably a necessity. 
 
The Ancient Editions of Plautus. 123 
 
 t<r(reTa-y-^liap). At the same time it is well known that the elision 
 of the final vowel of an Iambic or Cretic word is avoided in Latin 
 Poetry; and this fact suggests that there was something in the 
 actual Latin pronunciation that operated against the suppression 
 of the final long vowel or (the equivalent of a long vowel) the vowel 
 followed by -m in such a collocation of syllables. The rarity of 
 elision of the last syllable qI quidem, for example (see Munro's note 
 on Lucr. iii 904), lends credence to the traditional text of Ennius 
 Ann. 494 V. (32 2 M.) : 
 
 dum quidem | unus homo Eomanu' toga superescit. 
 
 A very strongly attested instance of this kind in Plautus, attested 
 by both P and A and by an ancient Grammarian also, is : 
 
 Pseud. 319 una opera alligem fugitiuam cdnem | agninis lactibus 
 {A P, Nonius), 
 
 and similarly : 
 Stich. 152 siquae forte ex Asia nauis heri | aut hodie uenerit 
 
 Pseud. 317 aiit terra aut mar'f | alicunde, etc. (A P), 
 
 Pers. 537 mea quidem istuc nil refert, | tui ego hoc facio gratia 
 
 Poen. 497 Certiim. Turn tu igitur die bono | Aphrodi'siis {A P), 
 Bacch. 51 ddae | unum expetitis palumbem, etc. {P, Nonius), 
 Merc. 257 nauem ex Rhodo qua est heri | aduectus filius {A P). 
 
 while for Cretic words we have lines like : 
 
 Most. 675 atque euoca | aliquem intus ad te, Tranio {A P), 
 Pseud. 1 121 (anapaestic) atque aliquem euocem | hinc intus 
 
 (^ ^)- 
 Bacch. 134 ibidem dgo meam operam perdidi, | ubi tii tuam 
 
 {P, Charisius), 
 
 Poen. 988 pro di immortales, plurirai | ad ilium modum {A P), 
 Pseud. 346 quindecim | habeo minas {A P). 
 
 The incidence of the ictus, be it remarked, cannot have had 
 anything to do with the scansion. If Plautus pronounced the 
 phrase as evoca aliquem (or hinc) intus, he would scan it after this 
 fashion, whether the ictus or beat of the verse fell on the first 
 or second or third syllable of the word evoca. I can see no 
 justification for the widely prevalent opinion that Plautine Prosody 
 
T24 The Ancient Editions of Plautus. 
 
 is in great measure controlled by the incidence of ictus, by the fall, 
 in other words, of the baton of a conductor marking the time 
 {poinds ictus, Horace). Surely the prosody of a word, the quantity 
 and articulation of a syllable, must have been controlled by the 
 pronunciation of the word in the phrase or sentence, and not 
 by anything else whatever. If the phrase vale ait was pronounced 
 by Roman lips va-le-a-it, Plautus could employ it as an opening for 
 an Iambic line, with ictus vale ait, as well as for a Trochaic line, 
 with ictus vdle ait. The exact conditions, apart from mere metrical 
 exigencies, under which Plautus uses Hiatus with Iambic and 
 Cretic words (and word-endings) are difficult to determine. It 
 seems to me that an emphatic disyllabic, like tuo in such a line as 
 Asin. 147 : 
 
 tuo facit iussu, tiio | imperio piret, etc., 
 
 is treated like emphatic tu in the line quoted above (Men. 1076) : 
 tujerus es, tu sdruom quaere, etc. 
 
 But since a full collection of the instances of this type of Hiatus 
 can be found in Klotz Altromische Metrik, pp. 119 sqq., it is 
 unnecessary to say more about it here. Nor need I do more 
 than mention another kind of Hiatus, viz. Hiatus with Interjec- 
 tions (including eccum) ; for this remained as fully in use in 
 classical, as in Plautine verse. And other two types of Hiatus, 
 viz. (i) at the Diaeresis of long lines, (2) at a change of speaker, 
 are now admitted by all editors to be unmistakable Plautine usages. 
 Ritschl's argument that the frequency of Elision under the same 
 circumstances is incompatible with Hiatus is now universally 
 abandoned. Without further delay on familiar ground let us 
 push our investigations further a-field, using the clue already 
 indicated, the appearances of Hiatus in post-Plautine poetry. 
 An unmistakable occasion of Hiatus in the classical poetry is 
 the pause between the two clauses of an Antithesis, as in Virgil 
 
 qui cultus habendo 
 sit pecori, | apibus quanta experientia parcis. 
 
 Under this heading we may bring the following well-attested line of 
 Plautus : 
 
The Ancient Editions of Plautiis. 125 
 
 Men. 882 lumbi sedendo, | 6culi spectando dolent (P, Auso- 
 nius). 
 
 We have a similar pause, accompanied by Hiatus in classical 
 poetry, between other separate clauses of a sentence, e.g. Virgil 
 Eel. ii S3, Aen. i 405, i 16 : 
 
 addam cerea pruna ; | honos erit huic quoque pomo, 
 et uera incessu patuit dea. | Ille ubi matrem, 
 posthabita coluisse Samo ; | hie illius arma, 
 
 to which we may perhaps add some instances of Hiatus in the 
 penthemimeral Caesura like Eel. x 13, viii 41 : 
 
 ilium etiam lauri, | etiam fleuere myricae, 
 ut uidi, ut perii, | ut me malus abstulit error. 
 
 Similarly in Plautus : 
 
 Stich. 270 sed ^ccum Pinacium, ^ius puerum. | hoc uide (AP), 
 Poen. 1009 quid in hinc uenistis lirbem ? | aut quid quadritis ? 
 
 Most. 976 sft profectus pdregre, perpotasse assiduo. | ic simul 
 {A P : — avisse Skutsch), 
 Merc. 259 inscendo in lembum, | dtque ad nauem d^uehor 
 
 Poen. 1 1 13 specie uenusta, | ore atque oculis pernigris {AP 
 
 Gellius), 
 Men. 476 prandi, potaui, scdrtum accubui, | abstuli {A P), 
 Pseud. 673 hic argentum, | hic amica amanti erili filio {A P), 
 Pers. 413 tene sis argentum: | ^tiam tu argentiim tenes? {A P) 
 
 (tene sis is unlikely), 
 Poen. 685 blande hominem compellabo. | hospes hospitem 
 
 salutat : saluom te aduenire gaiideo i^A P), 
 Pseud. 890 em illic ego habito. | Intro abi et cenam coque 
 
 Bacch. 946 milds Menelaust, ego Agamemno, | idem Ulixes 
 Lartius {A P) (agamennon B), 
 Merc. 538 Etiam cum uxore non cubet ? Amabo, | an maritust ? 
 
 Stich. 221 logos ridicules udndo. | age licemim {A P), 
 Trin. 185 em mda malefacta, | €m meam auaritiam tibi {A P) 
 (meam ? cf. also Philologus lix 500). 
 
126 The Ancient Editions of Plautus. 
 
 Trin. 48 o amice salue, | dtque aequalis. lit uales ? {-A P), 
 Most. 583 immo abi domum, uerum hdrcle dico, | abi modo (4 P) 
 (abf modo is unlikely), 
 
 Cas. 782 nam n6uom maritum, | €t nouam nuptdm uolo {A P), 
 Pers. 696 eum ego lit requiram, | atque uti redimim uolo {A.P), 
 Bacch. 495 Sdrua tibi sod^lem, | et mihi filiura. Factiim uolo 
 
 Merc. 530 ego \.6 redemi : | ille mecum orduit, etc. {A P). 
 
 ■t 
 A short vowel is allowed to stand in this hiatus in classical poetry, 
 
 and a long vowel to retain its length. We may therefore accept 
 Cas. 550 Propter operam illi'us hirci, fmprobi, | eddntuli (AP). 
 
 The justification of all these examples of Hiatus is to be found in 
 the pause that ensued after the word whose final vowel is left 
 unelided. Just as the inevitable pause at the end of a line prevents 
 elision of a final vowel when the following line begins with a vowel, 
 so in a minor degree the pause after redemi in the line just quoted : 
 
 ego td redemi : | ille mecum or^uit, etc. ; 
 
 and we may say that the hiatus is designed by the poet to indicate 
 that these lines are to be pronounced with a pause after these 
 words y. We seem to have the same justification of Hiatus in this 
 line of Virgil (Aen. iii 606) : 
 
 Si pereo, | hominum manibus periisse iuvabit, 
 
 where the emphatic articulation of the strongly stressed first syllable 
 of hominum would require a momentary rest after the word pereo 
 (cf. Aen. iv 235). With this line of Virgil I would associate these 
 well-attested examples of Hiatus in Plautus : 
 
 Poen. 89 Praes^nti argento | h6mini, si leno dst homo (A P, 
 
 Priscian), 
 
 Poen. 474 Volaticorum | hominum ? Ita dico quidem {A P). 
 
 y The recently discovered fragment of Menander's U^piKHpon^vr, has hiatus 
 at a pause : 
 
The Ancient Editions of Plaiittis. 1 27 
 
 (Cf. Most. 1032, a line for which we have the evidence of the 
 Palatine MSS. only : 
 
 Turbauit? Immo | dxturbauit omnia.) 
 
 Klotz in his account of Early Roman Metre (Altromische Metrik, 
 pp. 108 sqq.) claims that the utterance of any Proper Name, at 
 least any unfamiliar name, would be attended by a pause of this 
 kind and accepts as cases of legitimate Hiatus lines like : 
 
 Bacch. 354 sendx in Ephesum | ibit aurum arc^ssere (P, A n.l. ; 
 <^hinc^ ibit Camerarius). 
 
 He appeals to Horace (Epod. v. 100) : 
 
 et Esquilinae | alites, 
 
 where others find a mere imitation of the Greek Prosodic Hiatus 
 with -at (see above). 
 
 I find it difficult to beheve in a pause of this sort, unless 
 the name were specially emphasized, e.g.: 
 
 Poen. 443 
 nam isti quidem hercle orationi | Oedipo 
 opust coniectore (^ P, Priscian) ; 
 
 although an equally strongly attested case is : 
 
 Amph. 275 Ndc lugulae neque Vesperugo risque Vergiliae | 
 dccidunt {F, Varro, Festus). 
 
 A pause would be natural in some lines with the Vocative of a 
 Proper Name or the like, e.g. : 
 
 Poen. 1 127 O mi ere, salue, I Hanno insperatissime (A P), 
 Pers. 617 Virgo, | hie homo pr6bus est. Credo. Ndn diu apud 
 hunc sdruies {A P). 
 
 It is clear that the admission of this principle that any kind 
 of pause in the utterance of a line justified Hiatus opens the way 
 to a great deal of absurdity. One editor "^ may retain Hiatus in 
 
 2 In Rud. 526 sqq. Seyffert has ingeniously explained (or rather removed) 
 Hiatus by supposing the speakers, who have just come out of the water, to 
 split up their syllables in chattering fashion : 
 
 Nam omnia corusca prae tremore fabulor. 
 
 Edepol, Neptune, es balineator frigidus : 
 
128 The Ancient Editions of Plautus. 
 
 one line, another in another line, by managing to persuade them- 
 selves or their readers that some pause or other would be found 
 in the actor's articulation of the line. Almost any example of 
 Hiatus might be defended by more or less fantastic reasoning 
 of this kind. And yet, if we beheve that Hiatus was as much 
 the rule in Saturnian Poetry as it is the exception in classical verse, 
 we must suppose it to have played a considerable part in the Early 
 Drama. I think the balance of probability is in favour of the 
 correctness of well-attested lines like : 
 
 Stich. 216 Cons6nui, paene stim fame | em6rtuos {^A P), 
 
 Stich. 489 Scis til med esse | linisubselli uirum I^A P), 
 
 Merc. 312 Lysimache, auctor sum ut mdd amando | Unices 
 
 Pseud. 44 Lacrumans, titubanti | inimo, corde et pdctore (A P), 
 
 Pseud. 349 I gladium adfer : Quid opus gladio ? Qui hunc 
 occidam | atque me {A P), 
 
 Poen. 694 
 
 Ego id quaere hospitium, dbi ego curer mdllius 
 Quam r^gi Antiocho | dculi curari solent {A P), 
 
 Poen. 1290 Ita replebo, | Xtritate atn'tior multo dt siet {A P), 
 
 Pseud. 424 
 quo in cdmmeatum udlui | argentarium 
 proficisci, ibi nunc oppido opsaeptast uia (A P), 
 
 Aul. 703 nam jstos reges ceteros 
 
 memorare nolo, | hdminum mendicabula (P, Nonius), 
 
 Stich. 180 propt&ea credo mine esurio I icrius (A P) (adesurio 
 Ritschl, <|eo)> a. Mueller), 
 
 Pers. 556 quarta inuidia, quinta ambitio, s6xta | obtrectatio 
 
 Cum vestimentis, postquam abs te abii, al-algeo (ted ab. al. edd.) 
 Ne thermipolium quidem uUum in-instruit, etc. (nullum edd.). 
 In Asin. 755 sqq. the Parasite adds a clause to a contract and reads it 
 aloud : 
 
 Pa. Addone ? Dl. Adde, et scribas vide plane et probe. 
 
 Pa. ' Alienum | hominem | intro niittat neminera. 
 
 Quod ilia aut amicum | aut patronum nominet,' etc. 
 
 That the Hiatus here is not merely a possible but a necessary consequence 
 
 of writing down the additional words in the presence of the audience, will, 
 
 I fancy, be denied only by those who have never seen a play acted. Neither 
 
 for the Rudens-passage nor for this one have we the evidence of the Palimpsest. 
 
The Ancient Editions of Plautiis. 129 
 
 where in an English printed play a dash would indicate the pause 
 before the bizarre or recherchd expression. For I believe that 
 the law of Elision did not press so heavily on the Early Dramatists 
 as on the classical poets, and that the former thought only of avoid- 
 ing any scansion inconsistent with the actual pronunciation. The 
 actual pronunciation of a word like voluptatem was as near vol- 
 uptatem as volUptatem. Plautus accordingly allows either scansion, 
 whereas the classical poets follow the law of ' length by position ' 
 and scan only voluptatem. Similarly the actual pronunciation of 
 Vergiliae occidunt was probably as near Vergiliae occidunt as 
 Vergili{ae) occidunt, and Plautus accordingly tolerates Prosodic 
 Hiatus in such a phrase along with Elision. Where a phrase had 
 one and only one pronunciation in current utterance, Plautus allows 
 (in his dialogue metres, at least) only the scansion that corresponds 
 to this. Thus while he admits voluptatem, volo, he recognises only 
 voluptas-mea, volo-scire. He similarly restricts himself to Hiatus 
 in the phrase flagitium-hominis, clearly because that scansion was 
 postulated by current pronunciation, while he allows Elision or 
 Hiatus in intro-ibo, circum-imus and the like. Spelling often 
 indicates to us the course taken by Latin Pronunciation with 
 regard to Elision and Hiatus. We find Elision in anim{um)adverio, 
 magn(p)opere, tant{o')opere, circitor, but Prosodic Hiatus in circu(in)it, 
 factu{ni)iri, etc. The pun on domum-itionem and Domitium (Auct. 
 Herenn. iii 21) indicates Elision ; so does the spelling domusio (for 
 domi-usio, 'home use,' Petron.). Latin pronunciation thus appears 
 to have recognised now Elision, now Prosodic Hiatus with long 
 vowels or syllables in -m, while short vowels are invariably elided, 
 e.g. suav{e)olens, sesqu{i)opus, sem{i)esus (on triennium, etc., see 
 Brugraann's Griindriss). We are accordingly prepared to find 
 a corresponding variety of treatment * in Plautine versification. 
 But whether we can or should hope to determine in each case 
 the conditions of Hiatus and of Elision is a matter of doubt. 
 Cicero's words qm, ut versum facerent, saepe hiabant, would rather 
 
 » ' Where are you ? ' is in Plautus either ilbi tu^s ? or ilii tu is ? (cf. Leo 
 ad Cas. 963). Contrast, e.g., n(am) id est Ter. Hec. 696 with nam \ ego hue 
 (anapaest) Plant. True. 113 ; in{j) homo Pers. 620 with the usual mi\ homo 
 (anap.); ir{em) agis Men. 685, 898 with the usual rem\agis (tribr.) Also 
 manu \ emitto, e.g. Cure. 497 Alienos manu emittitis, with man{u) emitto, e.g. 
 
 Pers. 483 Di dent quae velis. Eho ! an iam manu emisisti mulierem? 
 (unless we should here read eviisti). 
 
 K 
 
130 The Aficieni Editions of Plautus. 
 
 lead us to regard Hiatus, at least in its less familiar types, as an 
 occasional, irregular licence, resorted to merely through metrical 
 necessity. 
 
 Naevius, unless we are to throw undeserved discredit on Cicero's 
 express statement, left gtii in Hiatus (not Prosodic Hiatus) in 
 the phrase vos qui. But this is certainly not the normal treatment 
 of the phrase. It is a licence of which the poet avails himself 
 in this particular line and would inevitably be ' emended ' by a 
 modern editor, unless Cicero's authority stood in the way. Unfortu- 
 nately we cannot attach the same weight to the consensus of P and 
 A as to an express declaration by an ancient author like Cicero, 
 for it is patent that these two authorities occasionally have fallen 
 into one and the same error. A very common error in texts of 
 Plautus is the ' modernizing ' of unfamiliar, archaic forms, the 
 substitution of ut for uti, istum, ilium for isiunc, illunc, and so 
 on. The scribe of A and the scribe of P, or some archetype of P, 
 have both committed the mistake of ' modernizing ' uti in Stich. 
 234, Pars. 685, &c.: 
 
 uti decimam partem Hdrculi poUiiceam, 
 
 cruminam banc emere aut fdcere uti remigret domum, 
 
 istunc, illunc m Poen. 651, 1302, Pers. 738, &c. : 
 
 atque fstunc e naui ^xeuntem onerdria, 
 
 idm hercle ego illunc dxcruciandum tdtum carnifici dabo, 
 
 nisi 6go illunc hominem p^rdo, perii, atque dptume 
 
 (so Hit for illic, Cas. 666, ? True. 200, posse for potesse, Pseud. 26, 
 sit for siet, ? Men. 519). 
 
 Mistakes like these and the others mentioned in the last para- 
 graph (part VII) afford no evidence whatsoever of relationship 
 between MSS. that exhibit them. They belong to the class of 
 ' inevitable ' mistakes, into which any scribe at any moment is 
 likely to fall. The appearance of Hiatus which they impose on the 
 line is illusory. 
 
 Sies has become sis (Men. no) in P, in the MSS. of Servius and 
 in the MSS. of Donatus ; and the MS. of Festus, the MSS. of 
 Nonius and the Ambrosian Palimpsest have, each of them, altered 
 expurigabo to expurgabo in Cist. 304 : 
 
 expiirigabo hercle 6mnia ad raucdm rauim {P n. I.). 
 Moreover it is quite possible that A and P perpetuate some errors 
 
The Ancient Editions of Plautus. 131 
 
 which had crept into some very early recension of Plautus, from 
 which they both are ultimately derived, although it is not likely 
 that these errors would be very numerous (see above, p. 113). 
 
 In reading the list, which I now furnish, of the remaining ' versus 
 hiantes' supported by the consensus of A P, it must be remembered 
 that there are four possibilities for each instance : (i) the text may 
 be erroneous, the error belonging either to the ancient ' accepted 
 text' of Plautus, or (2) having been inserted separately by the 
 scribe of A and the scribe of P or of some archetype of P, (3) the 
 text may be correct but the hiatus may be apparent and not 
 real, (4) the hiatus may be legitimate. 
 
 Bacch. 558 nequam | hominis ego parui pendo gratiam (? nequam- 
 homo, a word-group \\ke flagitium-hominis), 
 
 Bacch. 530 rdddidi patri | omne aurum. ndnc ego illam m^ 
 uelim {pm. ego A ; reddidit or reddidie A), 
 
 Cas. 126 post adtem ruri, nfsi tu acervom | ^deris (perhaps 
 aceruom, 4 syll.), 
 
 Cas. 564 homin^m | amatorem uUum ad forum procddere 
 (? hominem-amatorem, a word-group), 
 
 Cas. 1004 MYRRHiNA. C^nseo ecastor ueniam hanc dandam. 
 CLEOSTRATA. Faciam lit iubes (perhaps dandam, Cleostrata. 
 CLEOSTRATA Faciam), 
 
 Epid. 214 dbuiam ornatae 6ccurrebant sUis quaeque | araatoribus 
 (perhaps quaequae am.), 
 
 Men. 223 nam parasitus dcto | hominum miinus facile fUngitur 
 (so also Nonius) {(unus} munus Mueller), 
 
 Men. 1151 quoniam haec euendrunt, frater, nostra | ex sent^ntia. 
 (nostra, frater Gruterus), 
 
 Mil. 4 praestringat oculorum aciem | in acie hdstibus. 
 
 Mil. 604 quippe | hi si rdsciuere inimici consilidm tuom {om. 
 hi P : perhaps quippe qui), 
 
 Mil. 1 136 una exeuntis uideo | hinc e proximo {uid. ex. 
 Acidalius : Mm (hue} Mueller : perhaps uideon), 
 
 Pers. 262 n^m hoc argentum | alibi abutar: b6ues, quos emerem, 
 non erant (abutar al. Guyet. Perhaps aliuU ab.), 
 
 Poen. 1 130 GiDD. cognoscin Gidden^nem, | ancillam tuam ? 
 (perhaps (mey anc), 
 
 Poen. 862 Quid agis ? Facio qu6d manufesti moechi | hau ferme 
 solent {moechi (hie) Bothe), 
 
 K 3 
 
J 32 The Ancient Editions of Plautus. 
 
 Poen. 969 (see above, p. 109). 
 
 Poen. 328 n^mque edepol lucriim | amare milium amatorem 
 dddecet, 
 
 Poen. 1295 propemodum | hoc opsonare prandium potero mihi, 
 
 Poen. 1246 quoqu^ modo | huius filias apud uos habeatis sdruas 
 (the normal scansion is apud uos), 
 
 Poen. 1272 (see above, p. 109). 
 
 Poen. 982 adibo | hosce atque appellabo Punice (perhaps 
 adibon), 
 
 ■ Poen. 1327 siqufd lenoni | obtigit magni mali (siquidem P ; 
 siquidem quid Camerarius), 
 
 Poen. 448 me oboddientem | ^sse seruo li'berum (ob. me Bothe), 
 
 Poen. 782 idque in istoc adeo | aiirum inest marsiippio, 
 
 Poen. 105 1 patritus ergo | hospes Antidamds fuit {patri tuus ut 
 vid. P; perhaps erego, the old form of the conjunction). 
 
 Pseud. 151 nempe ita animati | ^stis uos : uincitis duritia hoc 
 dtque me {uin. hoc dur. ergo a. m. P), 
 
 Pseud. 443 (see above, p. 109), 
 
 Pseud. 897 pat^r Calidori, | opere edixit maxumo {fecit P), 
 
 Pseud. 410 erum eccum uideo | hiic Simonem una simul. 
 
 Pseud. 153 hue adhibete auris quae ego loquar, plagigera 
 genera | hdminum (plagigerula Bothe), 
 
 Stich. 171 nunc si ridiculum | hominem quaerat quispiam {} ri- 
 diculum-hominem, a vior6i-gro\\.^\\\:.Qflagitium-hominis), 
 
 Stich. 235 ecdstor auctionem | baud magni preti, 
 
 Stich. 477 Nesci'o quid uero | hibeo in mundo. f modo, 
 
 Stich. 344 iamdudum | ego fstura patior dicere iniustd mihi 
 (perhaps iamdudumne), 
 
 Stich. 384 iam non facio | adctionem : ini obtigit herdditas 
 (perhaps iam (ianiy), 
 
 Stich. 374 argenti | aurique aduexit nimium. Nimis factiim bene 
 (adv. multum P), 
 
 Trin. 539 nam fdlguritae sdnt alternae | arbores (alternas, 
 alternis edd.), 
 
 Trin. 540 su^s moriuntur angina | acdrrume {macerrumae 
 Onions). 
 
 To these may be added this instance of consensus of P 
 with a Grammarian in a Trochaic Septenarius : 
 
The Andent Editions of Plautus. 133 
 
 Pseud. 762 aui sinistra (-tera?), auspicio liquido atque ex 
 sententia (P, Nonius), 
 
 The following instances look suspiciously like errors inherited both 
 by P and by A from a common original, the ancient ' accepted 
 text ' of Plautus : 
 
 Poen. 453 sqq. (the ' leno ' is relating his experiences) 
 
 sex immolaui | agnos, nee potui tamen 
 
 propi'tiam Venerem facere uti | esset mihi (ut Ai) 
 
 quoniam litare nequeo, abii illim ilico (abi A P) 
 
 iratus, uotui | dxta prosicarier, 
 Stich. 459 sqq. (the parasite's relation) 
 
 auspicio | hodie | optumo exiui foras : 
 
 mustdla murera | abstulit praeter pedes ; 
 
 cum strena | obscaeuauit ; spectatum hoc mihist. 
 Poen. 485 sqq. (the soldier's relation) 
 
 tam crdbri ad terram | accidebant quam pira. 
 
 ut quisque acciderat, eum necabam | ilico 
 
 per cerebrum pinna sua sibi quasi tdrturem ; 
 
 but it is certainly remarkable that all three are narrative passages of 
 the same type. 
 
 Quintilian's account of tolerable and intolerable Hiatus in prose 
 Oratory "^ can hardly throw much light on the conditions of Hiatus 
 in Plautine verse. Dr. Maurenbrecher has arranged his lists on 
 the theory that Plautus' acceptance and avoidance of Hiatus de- 
 pended on the nature of the final syllable left unehded. An 
 examination of these lists will, I think, convince us that Plautus 
 makes no distinction between one final long vowel and another, 
 or between a final long vowel on the one hand and a final syllable 
 ending in -m on the other. He leaves virum in hiatus as 
 readily as m'ro, and viro as readily as viri. The theory that final 
 
 •> Turn vocalium concursus, qui cum accidit, hiat et intersistit et quasi laborat 
 oratio. Pessime longae, quae easdem inter se litteras committunt, sonabunt. 
 Praecipuus tamen erit hiatus earum, quae cavo aut patulo maxime ore efferun- 
 tur; « planior littera est, i angustior est, ideoque obscurius in his vitium. 
 Minus peccabit qui longis breves subiciet et adhuc qui praeponet longae 
 brevem, minima est in duabus brevibus ofifensio. Atque cum aliae subiun- 
 guntur aliis, proinde asperiores erunt, prout oris habitu simili aut diverso 
 pronuntiabuntur (Inst. IX iv 33). 
 
134 The Ancient Editions of Plautus. 
 
 -m had a different pronunciation in the time of Plautus and in the 
 time of Cicero cannot stand. No more can Prof. Birt's theory 
 that initial h was more resistive of elision in the pronunciation of 
 Plautus' time; for Plautus scans gm amat just as readily as qui 
 homo. Very unlikely is the view that the old Ablative suffix in 
 ■d was occasionally used by Plautus, as he occasionally uses the 
 old Subjunctive siet, the old Verb-form iurigo, &c. (See below, 
 p. 141.) 
 
 Like the i Decl. Gen. Sing, ending -as, the by-forms homonem 
 etc. (beside hominem etc.), qtiamde (beside quam), hoce die (beside 
 hodie), it appears to have been obsolete in the current usage of 
 Plautus' time, except in the monosyllables (like hand) med, 
 ted (not 'sed'; but possibly red), and would be as unsuitable 
 for his Comedies " as, let us say, the old-fashioned disyllabic pro- 
 nunciation of the ending -tion would be in an English Comedy 
 to-day. On the other hand the early i Decl. Genitive ending 
 -ai, and the old Pronominal Dative quoi'i {cmi) seem still to have 
 lingered on, like uti beside ut. Inf. -arier, etc., beside -ari, etc., 
 and should often be restored to lines which have the appearance 
 of Hiatus. It is conceivable that archaisms that were not used 
 by Plautus may appear in post-Plautine prologues (e.g. anticuus 
 in the prologue to the Casina, vv. 7, 13), and un-Plautine Hiatus 
 may find a place there too (vv. 47 — 48, 79 ?), just as it appears 
 in the Acrostic Arguments. An editor should therefore not be too 
 hasty in ruling out Hiatus in a line of a Prologue (e.g. Aul. s). 
 In the plays themselves he must balance the probability of the 
 Hiatus being genuine against the probability of the reading being 
 corrupt. Where there is consensus of PA or of /" with an 
 ancient Grammarian the latter probability will be greatly reduced. 
 His chief liability to error will be in those abnormal cases of Hiatus 
 like the uos qui \ accolitis of Naevius (cited by Cicero), which occur 
 in parts for which we have no other evidence than the Palatine 
 
 ■= In Amph. 191 sqq. there is a parody of the grandiloquent style : 
 
 Id vi et virtute militum victum atque expugnatura oppidum est 
 Imperio atque auspicio mei eri Amphitruonis maxume. 
 Praeda atque agro | adoreaque adfecit popularis suos (/*, A n./.). 
 Some editors accordingly feel themselves justified in admitting the form agrod. 
 I prefer to admit another usage of elevated diction, the triple gui, and to read 
 pyaedaque, agroque, adoreaque. Once that a scribe had miscopied praedaqite 
 BSjiraeda alque, the whole corrupt version would naturally ensue. 
 
The Ancient Editions of Piautus. 135 
 
 MSS. The temptation to ' emend ' these is invincible. The 
 only safe criterion of such Hiatus will be the presence of metrical 
 exigency, the necessity of getting certain words in a certain order 
 into a line. Where such necessity plainly exists and plainly calls 
 for Hiatus of an abnormal type, the editor should acquiesce 
 in the traditional version of the line, even though he can produce 
 no parallel example. 
 
 But to close our ears to Cicero's unimpeachable testimony to the 
 prevalence of Hiatus in Early Poetry is surely irrational, especially 
 when his testimony is confirmed by what we know of Saturnian 
 Verse and of the phonetic characteristics of the Latin language. 
 The rude versification of plebeian epitaphs and the like shews 
 us that ordinary, unconventional diction, when not trammelled 
 by artificial laws of Metre, acquiesced in Hiatus between words, 
 just as literary diction itself acquiesced in Hiatus between the 
 components of Word-groups or Compounds like quam-obrem, tdni- 
 eisi,pr{a)ehendo, dehortor. Hiatus is therefore not alien to the nature 
 of the Latin language ^ \ and this being the case, we should expect 
 to find it playing a part in the early Drama, whose verses de- 
 liberately reproduce the actual form of everyday utterance. The 
 reaction begun by Ritschl against the old indiscriminate admission 
 of Hiatus into Piautus' verses was a good thing, but it has been 
 carried too far. It is contrary to all laws of textual criticism when 
 editors continue to exercise their ingenuity in ' emending ' lines 
 whose text rests on the firmest possible basis of evidence, and 
 treat the united testimony of the Palatine MSS., the Ambrosian 
 Palimpsest and the citation by ancient Grammarians, in as cavalier 
 a fashion as the single testimony of one of these three witnesses. 
 Leo's edition in too violent reaction from this uncalled-for patch- 
 ing and tinkering of Piautus' lines exhibits almost every ' versus • 
 hians' in the form in which the MSS. present it. His theory 
 is that, although nine-tenths of these instances are un-Plautine, 
 still the lines may have had this form in the earliest collected 
 edition of the plays, since the belief was current in the Early 
 Empire that Hiatus was a feature of the older poetry. This 
 treatment of the MSS. seems to me to err in the other direction, 
 
 ^ The theory of the rhythmical formation of Latin Prose has not yet been 
 fully formulated. But the recent investigations of the rhythmical sentence- 
 endings affected by Cicero indicate that in his time a long final vowel was 
 normally left in ' prosodic ' hiatus before an initial vowel or h. 
 
136 The Ancient Editions of Plauttis. 
 
 in exaggerating the authority of the Palatine MSS. Their con- 
 sensus does not give us the reading of an ancient recension ; 
 far from it. It gives us merely the reading of a single Carolingian 
 codex, a codex abounding in transpositions, omissions and mis- 
 guided corrections such as characterize the work of every mediaeval 
 scribe. We cannot treat apparently erroneous readings of this 
 authority with the same respect as we treat the readings of an 
 ancient authority like the Ambrosian Palimpsest. Infinitely 
 greater respect is due to the readings supported by the consensus 
 of^and^. 
 
 (IX.) Orthography. 
 
 A full discussion of the orthography of the Plautine text would 
 require a monograph to itself. All that I aim at doing in this section 
 is to make clear the actual facts which condition the traditional 
 orthography of the several plays and to shew the weakness of 
 certain arguments which ignore these. A comparison of the two 
 recensions in point of orthography is, it will be shewn, very 
 difficult, if not quite impossible. 
 
 The text of the plays rests, where we have not the Ambrosian 
 Palimpsest to help us, on the mere authority of a minuscule arche- 
 type of the 8th or more likely 9th century. In certain portions, 
 those for which we have a collation of the Codex Turnebi (T), 
 it rests on the authority of a majuscule archetype, of about the 
 same age as the Palimpsest. But in the matter of Orthography 
 we are even worse off. For that corrector (.53) of the Codex 
 Vetus (^), who in the first eight plays seems to have entered the 
 readings of the minuscule archetype, where it differed from the 
 text of this codex, did not much concern himself with the ortho- 
 graphy of the older MS. The Palatine orthography of these eight 
 plays rests in reality merely on the immediate original of -5 and D. 
 In the remaining twelve, if the usual view be correct, that B and 
 the original of C and D are immediate transcripts of the minuscule 
 archetype, it rests upon that archetype's authority"; while in 
 the portions for which the collation of T has been preserved, it 
 can claim, at least in a degree (for a collation is unfortunately 
 
 ^ Fortunately a fairly large stafif of scribes was employed on the latter part 
 of B (from the P-plays onwards). One scribe has preserved some features 
 of the orthography of the original, another has preserved others, and so on. 
 
The Ancient Editions of Plautus. 137 
 
 not a text), to go back to the majuscule archetype itself. It therefore 
 stands on different levels in different portions of Plautus, and any one 
 might, if he thought it worth while, succeed in disentangling the various 
 threads brought into the present texture by the scribes or correctors 
 (i) of the original of B and D in the first eight plays, (2) of the 
 minuscule archetype, (3) of the majuscule archetype. Seldom 
 do our MSS. of Amph. — Epid. (not the Bacchides, which now 
 begins the second part of the Palatine archetype) retain trace 
 of ancient spelling, unless through some misapprehension. For ex- 
 ample in Capt. 887 they offer ^uo iusserat msteSid of ^uoius erai ; and 
 at Asin. 589, 593, where the corrector seems to have omitted 
 to expunge or stroke out the older form, leaving the modern 
 form suprascript, the supposed variant has been perpetuated in 
 
 cui 
 our MSS., ^uoi. He does not seem to have had the same hostility 
 to the old spelling vo in voltus, parvos^ etc., although the substi- 
 tution of the more familiar form seldom fails to be made by 
 one or other of the subsequent transcribers (e.g. in Capt. 169 
 captivom has become captivutn in VE). For every mediaeval 
 copyist was liable to correct what he conceived to be misspellings 
 in his original '. And so even at the highest stage of the Palatine 
 orthography we cannot hope to find the archaic spelling of the 
 majuscule archetype reproduced with any thoroughness. At Pseud. 
 1334 the collation (which fortunately pays some heed to ortho- 
 graphical details) of the Codex Turnebi shews sei where the other 
 MSS. {BCD') all offer sed. The scribe of the minuscule arche- 
 type presumably misread I as T and wrote down set; the corrector 
 then changed this to sed. Curiously enough at v. 1324 of the 
 same play it is BCD which retain sei, while in T xX. appears as 
 sed. So we can infer that the majuscule archetype offered the same 
 spelling in other passages s where our extant minuscule MSS. shew 
 sed instead of si. 
 
 Seeing that the Palatine orthography might degenerate '^ at 
 
 ' A deliberate change of ei to i is seen in readings like True. 559 perditum 
 sit {ox perditum se it. Pseud. 2^2 placidis for placide is. 
 
 s In Aul. 159 the substitution of sei es for sed est (P, A n.l.) seems almost 
 required by the sense. 
 
 •> A good example is the old spelling tostrina which has been systematically 
 banished. The sentence of banishment we may assign to the scribe (or cor- 
 rector) of the original of .5 and Z>. For the word occurs only in the first eight 
 
i^S The Ancient Editions of Plautus. 
 
 each successive transcription we can never infer that modern 
 spelUngs in this or that MS. stood in its original, nor even that 
 modern spellings (e.g. sed for set) found in a group of MSS. 
 were so written by the scribe of the common original or archetype. 
 For they may be due to the interference of the corrector. Nor is 
 this uncertainty confined to mediaeval transcription. In ancient, as 
 in mediaeval, times a scribe was always liable to alter an unfamiliar 
 form under the idea that it was a misspeUing. Varro in his 
 Lingua Latina (VIII 51) felt himself compelled to limit his 
 examples of the correct case-forms demanded by Analogy ' quod 
 librarios haec spinosiora indiligentius elaturos putavi.' Aul. , 
 Gellius makes it abundantly clear how much care had to be taken 
 to weed out such miswritings, in copies of early authors, as quadrupes 
 ecus for quadrupes eques (N. A. XVIII v), {hanc rem) futuram for 
 futurum (I vii), diei (Gen.) for dies (IX xiv. d), faciei (Gen.) ior fades 
 (IX xiv. 2). We cannot therefore infer with any.confidence that even 
 the spellings found in the Ambrosian Palimpsest {A) have not been 
 tampered with. We dare not so much as ascribe them to the im- 
 mediate original of A, much less to the A* recension itself. An 
 unfamiliar' spelling like terrai for terrae would (we might almost 
 say, inevitably) be altered to terrae by most scribes or correctors 
 at that time, if they understood that the Genitive of terra was 
 meant '^. Different scribes and correctors would have different 
 degrees of repugnance to the various archaisms found in Plautus' 
 text; so we must not think of inferring from even so ancient 
 testimony as that of the Palimpsest that the prominence of this 
 play over that play in respect of the retention of archaic spellings 
 implies that the earlier tradition of the text had been different 
 for each play. It impHes most likely nothing more than that 
 a new scribe or a new corrector intervened. Indeed I am not 
 sure that within the limits of the Palimpsest's text itself we do 
 not detect something of this kind. At Trin. 181 tessurae is 
 
 plays, six times in all. Five times it is written in our MSS. with -nst-, but 
 in the sixth occurrence, Capt 266, in tonstrina, the word in, which in early 
 minuscule is often hardly to be distinguished from hi or (sometimes) bi, was 
 misread as bi. The ' ghost- word ' bitostrina neither scribe nor corrector ven- 
 tured to touch. 
 
 ' Martial chooses terrai frugiferai as one of the most striking examples of 
 the uncouthness of obsolete forms. (See above, p. 30.) 
 
 I' Fortunately they often mistake it for the Nominative and write terra, 
 leaving us a trace of the archaism. 
 
The Ancient Editions of Plautiis. 139 
 
 corrected to usurae, and I fancy that from this point in the play 
 one can detect a change in the orthography. Certainly the re- 
 tention of archaic forms is greater in some parts of A than in 
 others, but I doubt whether we may substitute for the word 
 ' parts ' the word ' plays.' 
 
 From the presence of a modern spelling in Plautus MSS. or 
 archetypes we can rarely venture to make any inference regarding 
 its antiquity. It is the unfamiliar spellings, the spellings which 
 scribes would be tempted to alter but have by some good fortune 
 not altered, which we can refer to the old text. The frequency of 
 the spelling periuro (even where the metre requires the short penult) 
 in the Palimpsest as well as the oth.er MSS. is not so significant 
 as the preservation of the ancient orthography with -er- instead 
 of -ur- in True. 30 and Asin. 293 {perierat perhaps confused with 
 Plup. Ind. of pereo). On the other hand we must not be too 
 rash in generalizing from even a respectable number of instances. 
 There is no doubt that the editors of the large Teubner edition 
 made a mistake in extending to all Superlatives the form -umiis. 
 It has lately been shewn by Dr. Brock how admirably the arche- 
 types agree with the evidence of inscriptions in spelling minimus, 
 not ' minumus ' like optumus. To force upon minimus the u of 
 optumus is like forcing on incipio, recipio the u of occupo. 
 
 That the orthography adopted by ancient texts of Plautus had 
 a genuine claim to antiquity and, we may perhaps add, was 
 not invariably uniform, is suggested by the Grammarians' frequent 
 ' attestation of this or that spelling in this or that line. Sisenna 
 attested liui (not 'luce') for Amph. 165; Priscian attests the 
 spelling pe/Iucet for Aul. 566, also -//- in the compound of per and 
 iego at Asin. 747 ; Charisius attests Mde (dative) for Aul. 667. 
 A reference to our existing MSS. shews, as is only natural, how 
 woefully scribes have sinned in the transcription of these un- 
 familiar forms which the Grammarians attest; but it is pleasing 
 to find that here and there the archaism has managed to survive, 
 e.g. ted attested ' for Cure, i by Charisius and Diomede. And 
 a comparison of the spelling of A or F with the spelling of Plautus- 
 citations in our MSS. of Nonius (derived, unfortunately all of 
 them, from a minuscule archetype of perhaps the 8th century) 
 
 1 As a rule iect and meii survive in our minuscule MSS. only before esse, 
 the phrase being wrongly understood as ie desse (te deesse). 
 
140 The Ancient Editions of Plautus. 
 
 reveals here and there a coincidence which cannot be accidental. 
 Thus in Amph. 979 monosyllabic sis is spelt both in Nonius and 
 P as sies. Similarly both A and P agree very often in the 
 spelling quom (sometimes in qum, cf. Pers. 442); very often in 
 the spelling qur^ but never, I think, in quor, so that the occasional 
 appearance of the last may be nothing but a scribe's alteration 
 of the qur of his original; in the spelling aiio at Cas. 71, and 
 so on. A combination of the exol- of A and the exsul- of P at 
 Most. 597 leads us to the old spelling exsolatum. In fact although 
 a huge amount of details of archaic spelling must have been 
 modernized by scribes as early as the transcription of A and 
 P (e.g. eiram has become iram in True 264 in both)^ nevertheless, 
 owing to scribes' misunderstandings and other causes, so large 
 a number of traces have survived that I doubt whether there was 
 any type of archaic spelling in the ancient recensions of which 
 all trace whatsoever has been effaced. Against -os scribes or cor- 
 rectors of both texts seem at one time or other to have exercised 
 hostility (witness eius for ei os at Pseud. 719 in A, cuius for cui 
 OS at Trin. 558 in P) ; still a kind fortune has preserved opos sit 
 (perhaps mistaken for possit f) at Stich. 573 as late as B {CD 
 have possii). That -ce was also banned we learn from Pseud. 
 685 hoc venit (P) for hoc evenit (misread as hoce venit), Men. 349 
 hinc creditur (P) for hinc egreditur, and so on. But in spite of 
 all that we have lost, I do not know that it is too optimistic to 
 say that by collecting and appreciating all the traces that have 
 survived we can represent to our imagination a fairly true picture of 
 the general style of orthography in the best ancient texts of Plautus. 
 And this orthography has every mark of genuine transmission"" 
 from an early time, the time apparently when learned labour was 
 first applied to the plays ". How much it retained of the actual 
 
 "' It therefore seems to me not impossible that the traces of the spelling 
 tnoenio in the girl's remarks in Pers. 553 sqq. take us back to the ancient 
 text: 
 
 Sag. Quid id quod vidisti ? ut munitum muro tibi visum oppidumst ? 
 ViR. Si incolae bene sunt moratae, id pulchre moenitum arbitror 
 . . . . ea urbs moenita muro ^^leg. moero) sat erit simplici. 
 " If an editor in the earlier centuries of the Empire attempted to re-write 
 the orthography of Plautus, he would be sure to introduce exaggerated and 
 false archaisms. But of such we find, I think, no trace in our traditional text. 
 There is no sure trace, for example, of a final -d in Ablatives. The spellings 
 
The Ancient Editions of Plautus. 141 
 
 orthography of Plautus is another matter. Our text tradition is 
 too uncertain and the evidence of inscriptions too meagre and 
 unsatisfactory ° to give us much confidence in attempting to 
 restore that. Nor dare we venture to affirm this or that archaism, 
 found in one archetype, to have been absent from the other p. For 
 
 just mentioned, sies for the monosyllabic form, are not of this kind. We htive 
 the evidence of actual ancient inscriptions to shew that the older spelling was 
 often retained under circumstances which called for the later forms, and there- 
 fore Loewe's proposal to read seis may be rejected ; certainly the present 
 habit of editors of abandoning the disyllabic spelling in lines like Amph. 106 
 cannot be commended. Similarly quom for cum, the Preposition, is a frequent 
 spelling not only in our texts of Plautus but on early inscriptions. Besides 
 we may say that the use of siet where the scansion requires sit is quite of 
 a piece with the use of atque where the scansion requires ac, e.g. Trin. 935, 
 or of quamsi where we should print quasi, e.g. Pseud. 641. Cf. surriipuit 
 for trisyllabic surptiit in Capt. 8 and perhaps si voltis for suUis Men. 350. 
 The -eis (for -is') of A in curabeis Merc. 526, ibeis Cas. 92 is more suspicious, 
 but even of this substitution of ei for { there are examples in old inscriptions ; 
 nor should we forget the possibility that a scribe, who had been lately writing 
 a number of archaic «'-forms might unconsciously introduce the spelling, just 
 as the scribe of V at Capt. 755 has changed the offerre [-ere?) nahim of his 
 onpxisX to offer e gnaitim. 
 
 ° A common misconception in this matter must be corrected. We are not 
 entitled to infer that the orthography of the S. C. de Bacchanalibus was also 
 the orthography of the Comedies. In a state document obsolete diction and 
 archaic spelling were deliberately practised which would be quite out of place 
 in a Comedy. Our legal style, with its e.g. ' witnesseth, ' 'judgement' (instead 
 of 'judgment'), affords plenty of examples. Nearer to Plautus' lauguage, 
 because rather less formal, is the inscription with the Decree of Aem. Paulus 
 (189 B.C.), found in Spain, vi-hich, though slightly earlier than the S. C. Bacch., 
 admits spellings of the day, e.g. in turri Lascutana, essent, possidere, along 
 with the archaisms appropriate to legal documents, e.g. posedisent. The spelling 
 Troiad in a line of Naevius' Saturnian poem is indicated by our MSS. of 
 Servius' commentary on Virgil : 
 
 Noctu Troiad exibant capitibus opertis {troia de MSS.), 
 
 ' shrouding their heads they passed in the night-time from Troy.' An archaism 
 like this is as suitable to the elevated tone of Saturnian Epic as the archaisms 
 introduced into their Hexameter verses by Ennius, Lucretius, or Virgil. It 
 should, however, be noticed that only the spelling, not the pronunciation, 
 Troiad is here implied, for Saturnian verse was accentual and not quanti- 
 tative. 
 
 P Whether it might be possible to demonstrate a difference between the two 
 recensions in respect of the admission of vulgar or plebeian diction (e.g. lolarius, 
 meletrix, glarator, etc., in Pa) is a question worth consideration but for 
 which there is no room in this monograph. 
 
142 The Ancient Editions of Plautus. 
 
 example, until the recent discovery of the collation of T with 
 its caussa at Rud. 145, we should have guessed, from the entire 
 absence of this form (which is fairly well retained by A) from 
 our minuscule MSS., that the older spelling was peculiar to the 
 A* recension. The evidence of T shews us that it was merely 
 the scribe (or corrector) of the minuscule archetype who banished 
 the form from the Palatine text, and suggests that the same in- 
 terference is responsible for the similar absence oi essurio, etc. ^ 
 
 § 7. Conclusion. 
 
 We have seen how the Revival of the Plautine Drama a genera- 
 tion or more after the poet's death caused a parting of the ways 
 of text-tradition. Henceforth there were roughly speaking two 
 rival texts of the plays, the one adhering to the genuine 'ipsa 
 verba ' of the poet, the other exhibiting all the alterations, curtail- 
 ments or amplifications introduced by the stage-managers of the 
 
 q An editor who sets himself to repjroduce exactly the orthography of the 
 ancient text of Plautus has a difficult course to steer between Scylla and 
 Charybdis. Leo follows the MSS., or rather the collation printed in the 
 apparatus criticus of the large Teubner edition (which in the earlier plays 
 sometimes fails to notice a spelling like capthiom, etc.), and so prints in the 
 first eight plays the modernized orthography of a mediaeval scriptorium. The 
 Teubner editors, recognizing the tendency of scribes and correctors to make 
 a wholesale removal of certain unusual types of spelling, e.g. -list for -tis est, 
 disregard in these cases the evidence of the MSS. and print -ust wherever the 
 metre does not require -us est. Now -ust was the form of rapid, -us est of 
 more deliberate utterance, and it is quite possible that the nuance of expression 
 has been correctly retained by the MSS. at Men. 433 : 
 
 Mess. Quid eo opust ? Men. Opus est — Mess. Scio, ut ne dicas. 
 Men. Tanto nequior ! 
 It would be a useful piece of work if some student who has leisure would 
 collect all the indications of the actual orthography of -ust, -us est, whether 
 supplied by MSS. or metre, also those of the orthography, or at least pro- 
 nunciation, of Plautus which are afforded by puns or (more frequently) by 
 assonance and alliteration. The difficult question of assimilation or non- 
 assimilation in Compounds would be greatly facilitated by a collection of 
 instances like Amph. prol. 13 ; haec ut me voltis adprobare, adnitier ; Amph. ' 
 993 adsto, admoneo ; Asin. 657 coUoca cruminam in coUo plane. Bacch. 276 
 Quin lu audi. Immo ingenium au(i)di baud pemoram hospitis supports haud 
 before a consonant; Merc. 68 rus rusum confestim exigi solitum a patre, 
 supports rusum, &c., &c. 
 
The Ancient Editions of Plautus. 143 
 
 Revival time in order to make the performance pleasing to the 
 audience of the day. Further in the case of single lines here 
 and there, where a puzzling word or phrase occurred, the labours 
 of Grammarians and Antiquarians often resulted in the production 
 of two or three variants. We have traced the history of the 
 rival texts down through the last century of the Republic, the 
 Augustan Age and the Early Empire until the period of the 
 Grammarians whose works have survived. Nonius, Charisius, etc. 
 This period is also, roughly speaking, the period of the Ambrosian 
 Palimpsest, which appears to preserve the main features of the 
 'genuine' text, and of that majuscule archetype, from which the 
 other family of MSS. is sprung. This archetype (/") exhibits the 
 ' Revival ' modifications often with ' mixture,' usually with addition ^ 
 of the 'genuine' version. How far the latter features are on 
 the one hand due to marginal insertions by owners of successive 
 copies, or on the other were actually present in the PA-recension 
 (which would then have pjesented something of the appearance of 
 a ' variorum ' edition) it is hard to say. Although the distinctive 
 features of both recensions have undoubtedly been a good deal 
 blurred by ' mixture,' we have seen that a study of the divergent 
 readings of A and P leaves the impression rather of two different 
 editions which had in many passages been assimilated through 
 the adoption by one of some readings of the other, than of two 
 copies of the same edition which were beginning to exhibit points 
 of dissimilarity. But intertwisted and intertwined as the threads of 
 our text-tradition appear, they can always be resolved into these 
 two, (i) the 'genuine' text, (2) the 'Revival' text. There is 
 nothing that indicates the re-writing of the verses by late editors 
 (from the second century a.d. onwards) in the fashion that they 
 were re-written by the Italian editors of the Renaissance. Where 
 we have not the 'ipsa verba' of Plautus, we seem to have the 
 adapted version of the ' Revival ' period, a period not later than 
 a generation or two after the poet's time. 
 
 In the latter part of last century the view steadily gained ground 
 
 ■^ A very fortunate practice for us. Otherwise we should have lost the 
 ' genuine ' version in parts for vifhich the evidence of Aa v^as wanting. We 
 should have lost, e.g., .\sin. 23 — 24, and have had in their place the mere 
 ' Revival ' substitute, vv. 25 — 26, without that interesting Old Latin allusion 
 to the custom of coming out of the house into the open air for an oath by 
 Dius Fidius. 
 
144 ^'^^ Ancient Editions of Plauhis. 
 
 that the errors (or apparent errors) common to A and P prove them 
 to be two divergent copies of one and the same edition. This view, 
 which I have tried to combat on pp. 104 — 118 above, has recently 
 culminated in the theory stated by Profr. Leo in the first chapter of 
 his ' Plautinische Forschungen,' viz., that the writings of Plautus 
 disappeared in the Early Empire for something like half a century, 
 and were recovered in the provinces by the Grammarian Valerius 
 Probus, and that with the help of the copies which he brought from 
 the provinces to Rome the first collected edition of the plays 
 was made, an edition limited to the 21 'fabulae Varronianae.' 
 Both A and P are, he says, divergent copies of this edition. 
 
 Before a startling theory like this can be accepted, definite proof 
 of its truth is needed. Let us see how Leo arrives at his con- 
 clusions. In the first place he takes quite literally the words of 
 Suetonius (quoted above, p. 31) regarding the decay of the study 
 of the earlier writers, and endeavours by a reference to the con- 
 flagrations in Nero's and in Titus' reigns to account for the absence 
 of so important an author as Plautus from the public libraries in 
 Rome. Next he makes up as imposing a list as possible of the 
 errors or apparent errors that are common to A P, pressing into 
 service such miswritings as these mentioned above on p. 105, and 
 even modernized orthography like -ae for -at in the Gen. Sing, of 
 the First Declension. He declares the common source of the 
 Aa and P* recensions to have been a ' variorum ' edition in which 
 doubtful lines, interpolations, &c., were all incorporated in the text 
 and merely indicated by diacritical symbols. ^ But he lays chief 
 stress on the extent to which Hiatus is present in our traditional 
 text. An editor in the second century a.d. would, he says, readily 
 tolerate ' Hiatus, because it was believed at the time (witness 
 
 3 This explanation would suit the appearance of A and P in some cases, 
 e.g. in the two versions of the Poenulus ending, which both stand in the text of 
 A and P in the same order. But in most cases their appearance gives indi- 
 cation of an intrusion into the text from the margin. To me this suggests 
 rather ' mixture ' of text by successive owners. 
 
 * His line of argument here is not clear. He mentions Cicero's statement 
 that the early poets ' saepe hiabant.' The instances quoted by Cicero, e.g. 
 Naevius' vos qui \ accolitis and Graii \ atque l/arbari, etc., are certainly not 
 less striking than the instances Leo quotes from our traditional text of Plautus. 
 If then Cicero and his contemporaries tolerated these, regarding them as 
 peculiarities of early poetry, why should the toleration of them in an edition 
 of Plautus assign the edition rather to the age of Probus than of Cicero ? 
 
The Ancient Editions of Plautus. 145 
 
 the Acrostic Arguments and contemporary archaistic (?) epitaphs) 
 that Hiatus was a feature of Archaic poetry. And he makes an 
 interesting, but too brief examination of the different plays in 
 respect to the amount of 'illegitimate ' Hiatus (i.e. such as the text 
 of Terence does not exhibit), making this a test of the merit of 
 the text-tradition in the several plays. These are, I think, the main 
 props on which the theory rests ; but the details are filled in 
 with such fulness and skill and the whole theory is set forth 
 with such learning and ingenuity and with so great a wealth of 
 illustration from Greek literature, that I do not think any un- 
 prejudiced person can read with attention the whole of the first 
 chapter of Leo's book and pronounce the theory to be utterly 
 impossible '°-. All that can be said is that so revolutionary a theory 
 requires very strong evidence to support it, and that the avail- 
 able evidence is not strorig enough. Whether the theory ultimately 
 hold the field or not, there is no doubt of the permanent benefit 
 conferred on Plautine studies by this brilliant chapter, and this 
 chapter is by no means the best in Leo's epoch-making volume. 
 
 After all there is a certain amount of common ground for 
 rival theories. Everyone must admit that the revival of the 
 older literature at the time of Probus would likely be marked 
 by a re-editing of the old dramatists, Plautus and Terence ". The 
 individual copies i^A and F) of two booksellers' issues, derived 
 
 " The point which to me seems most difficult to believe is that Probus 
 could find no materials for an edition in the libraries of Rome or other towns 
 of Italy. Gellius speaks of a copy of Claudius Quadrigarius preserved in the 
 library of Tibur, and of his having found the genuine old spelling ^cifj (Gen.) 
 in its text. (See above, p. 32.) Why may not the same library (cf. N.A. 
 XIX V. 4) have preserved an old copy of Plautus ? From Suetonius' words 
 (see p. 31) we gather that Probus while at Berytus had read (legerai) some 
 old literature which first directed his inclination to this field of study, and 
 that afterwards he got together a number of copies (multa exemplaria con- 
 traxif) of the older authors as materials for editions of their writings. Surely 
 Probus and his friends would hunt through the libraries of Italy as zealously 
 as the Renaissance scholars did. However this is perhaps a minor point. It 
 matters little where Probus got his materials for an edition of Plautus. The 
 main point is whether the materials were sufficient for turning out a good 
 edition. ■ None of the examples which Leo cites of the tinkering of corrupt 
 and defective passages by a second century editor seem to me at all con- 
 clusive. 
 
 » Leo does not make out Terence to have been lost and resuscitated like 
 Plautus. 
 
 L 
 
146 The Ancient Editions of Plautus. 
 
 mediately or immediately from two recensions, A* and P*, have 
 left us no clue like the ' Calliopius recensui ' of Terence MSS. On 
 the strength of A being a codex of a N. Italian library and the 
 majuscule archetype P apparently of a French library, we are 
 induced to locate the two booksellers' issues in Italy and Gaul 
 respectively. The date may be any time not later (at least in 
 the case of the ^-issue) than the fourth century. They would 
 be likely (one or both of them) to have closer ties of connexion 
 with a second century edition than with any belonging to the 
 Republican age, just as any edition, let us say, of Lucretius pub- 
 lished after Probus' time would be likely to retain what had been 
 won for the text by that scholar's labours. If we suppose Probus to 
 have procured a sufficient number of copies, embodying, either 
 in pure or in ' mixed ' form, the two divergent types of text, the 
 ' genuine ' and the ' Revival ' text, and some scholar to have con- 
 structed out of these materials, more or less blended, an edition 
 or editions which in time led to fresh booksellers' issues, there 
 is nothing in all this supposition which cannot commend itself 
 to all parties. Opposition becomes necessary only when a theory 
 is asserted with a confidence of which the evidence at our disposal 
 does not allow, whether the theory be that the first collected edition 
 of the plays was made (i) in the time of Varro, or (2) in the time 
 of Probus, or else (the prevailing theory, I am afraid) that both our 
 recensions. A* and P*, come from one and the same ' variorum ' 
 edition (whether that belonged ^ to Varro's or Probus' time or to 
 the Augustan age). The evidence is not, in my opinion, sufficient 
 either to establish or to destroy any of these hypotheses ". All that 
 
 y The acceptance of the prevalent view regarding the errors common to 
 A and P seems to me almost to preclude the possibility of a date earlier than 
 Probus. Consider for a moment the reading common to A and to our minus- 
 cule MSS. at Epid. 455, centones farcias instead of c. sarcias. This error I 
 believe (see p. 113) to have been made by the scribes of the two texts indepen- 
 dently, in the one case owing to the close resemblance of s to / in minuscule 
 script. But the prevalent view ascribes it to the (supposed) common original of 
 A* and P*. How then can a patent miswriting like this have persisted in ancient 
 texts (of course it would not be patent to mediaeval scribes) for any length 
 of time ? The presence of such obvious miswritings in the Codex Vetus (5) 
 as kaec, koc for haec, hoc, due to misreading of the Rustic Capital form of H, 
 is admitted evidence for the close proximity of B to the majuscule archetype 
 P. The same reasoning must hold good for corruptions like centones farcias. 
 
 ' Nor yet this other, which has more in its favour, viz. that Pa represents 
 
The Ancient Editions of Plautus. 147 
 
 can be said is that the balance of probability inclines against the 
 third. 
 
 Nor, so far as I can see, should it alter any of the approved 
 methods of Plautine textual criticism, if the theory were to be 
 provisionally accepted (though unproved) that Plautus' plays were 
 actually lost to the world for some 50 years and were actually 
 recovered in provincial book-stalls and finally published in a 
 ' variorum ' edition (containing careful record of the two divergent 
 forms of text) by Probus himself or by a pupil or friend of his. 
 Why then, I may be asked, do I think it necessary to emphasize 
 the uncertainty of this theory ? Because I am convinced that the 
 unquestioned acceptance of it is likely to lead to mischievous 
 results. Sooner or lat er the feeling would arise that our traditional 
 text of Plautus is nothing, after all, but a concoction of the second 
 century a.d., and differs from the real text as much as we may 
 suppose that a second century restoration of an inscription of 
 Plautus' time would differ from the real inscription. It is true that 
 Leo points out that Probus and his school would shew great 
 carefulness, and fidelity in combining the newly discovered materials 
 into an edition. For all that, he himself in the matter of Hiatus 
 seems to despair of arriving by the help of the traditional text at or 
 near to Plautus' own usage. He adheres to the old idea that 
 Plautus still pronounced the final d of Ablatives and final m (not, 
 I think, initial h) in such a way that auspicio \ hodie j optumo and 
 murem \ abstulit produced no hiatus in the verse ; and he supposes 
 that cases like these would open the way to the intrusion of a host 
 of un-Plautine types of Hiatus into the ancient traditional text. So 
 he assigns to the edition of Probus' time the cases of Hiatus 
 which are attested merely by the Carolingian ^ archetype of the 
 Palatine MSS. as readily as the cases attested by the consensus 
 of A P, indicating all in the same way by the apex-symbol. For 
 example, in Amph. 263 he refuses the obvious emendation of//// to 
 iih'c and prints : ibo ego illi 6bviam. All this is, I fear, likely 
 
 a 'variorum' edition made in the second century A.D., while Aa represents 
 the edition of some Plautine scholar in the last century of the Republic or 
 the first of the Empire. 
 
 a An example will illustrate how readily ' Hiatus ' may step in at each new 
 transcription. Cure. 229 appeared in the original of £ fy as : Quis hie est qui 
 loquitur^ Quoiam vocem \ audio ? But B has preserved the reading of the 
 common original of BEVJ : quoiam vocem ego audio ? 
 
1 48 The Ancient Editions of Plautus, 
 
 to have something of a paralysing effect upon students and to 
 retard rather than stimulate new discoveries in Plautine Prosody 
 and Plautine Accidence. And it is mainly with the view of 
 exposing the weakness of much of the evidence on which the 
 new theory rests that I have written this monograph ^ 
 
 In time we shall probably get clear light thrown on these dark 
 corners of the history of the Plautine text. Perhaps the discoveries 
 in Egypt of one or more of Plautus' Greek originals will change the 
 situation. But even if Egypt brings us nothing, there are still 
 plenty of things " in which patient investigation will win certainty 
 for -us. We are yet far from the ultimate stage of research. We 
 have yet to learn the full details of the Metres used by Plautus and 
 of the restrictions under which he used them. We do not yet fully 
 know Plautine Accidence ; for example, whether Plautus used 
 the form eapsa as well as ipsa and eapse, or how far by-forms'^ 
 
 •> I wished to state the case, the ' conservative ' or ' optimist ' case, clearly 
 and connectedly, rather than to put it in the form of a polemic against this 
 or that theory. So the various items of evidence have been left to follow 
 their own natural order without express reference to the position they hold 
 in other theories. For instance, Varro's quotation (ap. Fest. 375 M.) of 
 Cure. 568, which is often adduced in evidence that the text of Plautus in 
 Varro's time contained errors which were removed by the editors of Probus' 
 time, appears in my monograph (in the Varro-section, p. 8) merely in this 
 form : 
 
 " Vapula ergo {leg. Vapulare ego) te vehementer iubeo, ne me 
 territes (P, A n.l.)," 
 because I could not bring myself to believe that vapula ergo was anything 
 more than a mere error of the scribe of our MS. of Festus, influenced possibly 
 by the vapula in Varro's preceding quotation of Terence. Varro is not 
 attesting the Imperative vapula : he is explaining the proverb vapula Pafiria 
 by an appeal to the postulated O. Lat. use of vapulare in the sense oiferire. 
 
 " It is disappointing to find that Semitic scholars have done nothing with 
 the improvements furnished by the new collation of the Codex Turnebi in the 
 Carthaginian passage in the Poenulus (published in the Classical Review, 
 XII 361, five years ago). It would greatly help us if the two versions could 
 be satisfactorily dated and discriminated. 
 
 '' The theory that Plautus allowed elision of final i (after a short vowel) 
 before initial vowels (or h) indiscriminately, can hardly, I imagine, be accepted 
 by any one. It conflicts too violently with the text of the plays. It is im- 
 possible, for example, to believe that in such a line as Rud. 888 Plautus 
 would scan collus haud multo post erit in the same way as collum haud 
 mullo post erit. It has always seemed to me that, before this theory can be 
 considered worthy of discussion, it must state its own limits and conditions. 
 
The Ancient Editions of Plautus. 149 
 
 in -e (e.g. sedeate Poen. 5?) appeared beside forms in -is. Nor 
 yet Plautine Prosody; e.g. whether Plautus recognized (i) hodie, 
 nesei (cf. Brock), &c. (like siquidem beside szquidem, qna(ni)si beside 
 quasi)., (2) hulus (trisyll., hke illms ; cf. Sommer ' Handbuch,' 
 P- 454 n^, (3) siquis, etc. (Bacch. 976 suggests that this scansion 
 was unknown to Plautus), (4) nihilum {-i, etc.). Nor yet Plautine 
 Grammar; e.g. whether the construction of an Ace. Case with 
 an Impersonal Passive Verb is Plautine (see above, p. 117). And 
 Leo's suggestion, that the traditional text in some plays® is (and 
 was at the time of Probus) less sound than in others, should be 
 fully followed up. Before it can be established, a full record must 
 be compiled of the extent to which all kinds of questionable usages 
 are present, not merely unusual types of Hiatus, but also such 
 things as (i)the shortening under the Breves Breviantes Law of 
 naturally long syllables or of syllables beginning with qu or with 
 a mute and a liquid, (2) the total absorption by elision of meo, 
 mi, mei, etc. (rei seems certain), (3) contractions like haecst, illicst, 
 rest (Merc. 857, etc.) for res est. 
 One by one these and other moot points will be decided, and 
 
 Meanwhile there is undoubtedly something attractive in a combination like 
 sejueris and sequere, potis and pole, tnagis and mage, satis and sai(e), which 
 suggests extension of the same phonetic variety to pluris {-re) in Pers. 353, 
 panis {-ue) in Cure. 367, aetatis [,-te) in Trin. 1090 (cf. Trin. 1153). Potest 
 (ior potis 'st, as well as for pote 'st) too now appears in a new light, and the 
 point of Cist. 573 Servate di med obsecro. At me perditis. But even in 
 this particular matter, the recognition by Roman Orthography of the weak 
 pronunciation of j' in the final syllable -is (before a consonant ?), there is still 
 uncertainty. For example, did sahuis est, nimis est, become in rapid utterance 
 salutest, nimest, or salutist, nimist ? And should we print magis {magP) or 
 mage, where the word has pyrrich scansion, e.g. Cure. 171 ? (Cf. ipsii' Cure. 
 1 70 and ipse). 
 
 * Leo's assumption that the corrupt state of the text of the Truculentus in 
 our (minuscule) MSS. must date from ancient times is certainly unwarranted. 
 The little that we have of the Palimpsest for this play exhibits a good enough 
 text. In the American Journal of Philology (XVII 442) I have tried to shew 
 that the curious change from a good to a bad text in ^ C Z> at the beginning 
 of the Truculentus is exactly what would happen if there was at this point 
 in the minuscule archetype a change of scribe, like the change that we 
 often see in 8th — 9th century MSS., the concluding portions of which have 
 been penned by a monk trained in a different school of penmanship, and 
 using unfamiliar abbreviations and ligatures (e.g. True. 2 deUeris for de 
 vesiris). 
 
1 5 o The Ancient Editions of Plautus. 
 
 the text thereby be so well established as perhaps to allow an 
 unqualified answer to the problems discussed in this monograph. 
 For the present, on the strength of the evidence collected in 
 the preceding pages, the duties of an editor of Plautus I take 
 to be : 
 
 (i) To adhere to the consensus of A and P, unless there is 
 evidence of scribes having fallen independently into the same error. 
 The consensus gives us at the worst the text as estabHshed by 
 the labours of Probus and his school (and that is a great deal), 
 at the best the traditional text of the last century of the Republic. 
 
 (2) To give little or no credence to the chance Plautus-citations 
 by a Grammarian, where they differ from our MSS., unless it 
 appear that he took them directly from a text of the plays. On 
 the other hand, 
 
 (3) To give all possible credence to Grammarians' attestations 
 of unusual words, forms or spellings ; for there was every chance 
 of these being altered to the normal shape by copyists at some 
 period or other of the text's transmission. 
 
 ADDENDUM (to pp. 26 and 85). 
 
 Nonius' copy of Plautus may have been not a single volume, 
 but a collection of separate papyrus rolls ; so that, in excerpting 
 the plays composing the several letter-groups, he may have fol- 
 lowed an order of his own and not the order assigned by a 
 publisher. In excerpting Lucilius xxvi — xxx he begins from the 
 end (xxx, xxix, xxviii, xxvii, xxvi). 
 
INDEX OF LINES DISCUSSED. 
 
 Amph. io6, p. 141 ; 193, p. 134 ; 208 
 p. 120; 27s, p. 5, p. 127; 488: 
 p. S ; 523, p. 86 ; 1067-8, p. 79- 
 
 Asin. 23-6, p. 38; 32 sqq., p. 38 ; 47 
 8, p. 38 ; 480 sqq., p. 48; 685, p 
 S ; 706, p. 122 ; 755 sqq., p. 128 
 828-9, p. 48 ; 894 sqq., p. 49. 
 
 Aul. 5, p. 134; 116, p. 25; 159, p 
 137 ; 161-4, p. 35, p. 18 ; 178, p 
 30 ; 328, p. 89 ; 354-5, p. 18 : 399 
 p. 26 ; 446, p. 5 ; 555, p. 17 ; 592- 
 8, p. 49 ; 703, p. 128 ; 829, p. 92. 
 
 Bacch. 51, p. 123; 65 sqq., p. 49 
 134, p. 123 ; 377-8, p. 38 ; 379-81 
 p. 38 ; 382-3, p. 38 : 392 sqq., p 
 38 ; 487, p. ics ; 495, p. 126 ; 496. 
 p. 69; 506, p. 57; 512-4, p. 50 
 519, p. 58; 530, p. 131 : 540—551 
 p- 38; 558. P- 131 ; 932, p. 117 
 
 946, p. 125 ; 948, p. 69 ; 1 150, p. 
 
 35 ; 1207, p. 83. 
 Capt. 524-5, p. 79 ; 1016-23, P- 39- 
 Cas. 7, p. 134; 13, P- 134; 47-8> 
 
 p. 134; 72, p. 109; 79, p. 134; 
 
 126, p. 131 ; 137, p. 58; 185 
 
 sqq., p. 49 ; 206, p. 3 ; 267, p. 
 
 26 ; 347. P- 94 ; 443. P- '9 ; 55°. 
 
 p. 126; 564, p. 131; 571, p. 117; 
 
 625, p. 116; 636, p. 69; 646, p. 
 
 109; 666, p. 130; 727, p. 64; 
 
 747, p. 69 ; 749, p. 69 ; 769, p. 58 ; 
 
 782, p. 126; 800, p. 83; 839, p. 
 
 117 ; 960, p. 94; 1004, p. 131. 
 Cist. Arg. 10, p. 87 ; l, p. 6 ; 8, p. 6; 
 
 84, p. 58; J26 sqq., p. 39; 132, 
 
 p. 58; 408, p. 15; 510-11, p. 50; 
 
 518, p. 92; 573. P- >49- 
 Cure. 367, p. 149 ; 393, p. 7 sq. ; 424, 
 
 p. 78 ; 455 sqq.. p- 5° ; 568, p. 148 ; 
 
 fin., p. 83. 
 Epid. 9 sqq., p. 79; '64-5- P- 4°; 
 
 173-4. P- 79; 187, p. 27; 189-90, 
 
 p. 102; 214, p. 131 ; 231, p. 8; 
 
 235. P-59; 455. P- 113; 515. P- 58; 
 
 568, p. 113; 606, p. 59 ; 620, p. 
 
 IS. P- 59 ; 621. P- 69 ; 624, p. 70 ; 
 
 732, p. 83 ; 733. P- 59- 
 Men. 12, p. 22 ; 94, p. 25; 127, p. 
 
 8 ; 151, p. 92; 200-1, p. 79, p. 
 
 109 ; 223, p. 131 ; 289, p. 17 ; 349. 
 
 p. 140; 352, p. 81, p. IIS; 360, 
 
 P- 5° i 433. P- «42 i 476) P- I2S ; 
 
 519, p. 130; 572, p. 59 J 573, p. 
 112; 587, p. 59; 639^p. 51; 854, 
 p. 78; 882, p. 125; 1028 sqq., p. 
 51; 1047, p. 9; 1 151, p. 131. 
 Merc. 251, p. 60; 256, p. 60; 257, 
 p. 123 ; 259, p. 125; 269 sq., p. 
 40; 271, p. 60; 276, p. 41 ; 285, 
 p. 62; 300, p. 107; 301, p. 60; 
 312, p. 128; 314, p. 60; 319, p. 
 70; 371-2, p. 61 ; 475. P- 70; 53°. 
 p. 126; 538, p. 125; 547, p. 41 ; 
 555. P- 41 ; 598, p. 114; 619, p. 
 
 9; 619-24, p. 51 ; 757, p. 70; 765, 
 p. 70; 767, p. 61 ; 842-3, p. 114. 
 
 Mil. 4, p. 131 ; 24, p. 9; 149, p. 108; 
 173, p. 92 ; 176, p. 70 ; 209, p. 61 ; 
 213, p. 22 ; 254, p. 117 ; 276, p. 92; 
 351, p. 105 ; 357, p. 33; 359, p. 61 ; 
 472, p. 61 ; 591, p. 112 ; 599, p. 71 ; 
 604, p. 131 ; 716, p. 71 ; 727-9, p. 
 109; 762, p. 71 ; 945-6, p. 71 ; 
 997^ P- 41 ; 'OS4, p. 115 ; 1136, 
 p. 131 ; 1138, p. 92, p. 102 ; 1142, 
 p. 61 ; 1180, p. 16, p. 61 ; 1273, 
 p. 51 ; 1287-9, P- 5« ; 1419, P- "6. 
 
 Most. 93 sqq., p. 51 ; 245, p. 10; 409 
 sqq., p. 41; 580, p. 118; 583, p. 
 126 ; 605, p. 71 ; 675, p. 123 ; 681, 
 p. 71 ; 682, p. 106 ; 688, p. 71 ; 
 714, p. 72 ; 719, p. 92 ; 750, p. 
 92; 762, p. 72 ; 768, p. 105 ; 817- 
 47, p. 42 ; 832, p. 106 ; 928, p. 72 ; 
 939 sqq., p. 42 : 97^. P- 125 ; 984. 
 p. 118; 1000 sqq., p. 52; 1032, 
 p. 127; 1043, p. 109; 1103-4, p. 34- 
 
 Pars. 20, p. 72 ; 171, p. 72 ; 179, 
 p. 72; 262, p. 131 ; 269, p. 72; 
 319. P- 105 ; 321. p. 73; 353. P- 
 149 ; 408, p. 107 ; 413, p. 125 ; 
 
 437, P- 73; 463. P- 78; 483. P- 
 129; 499, p. 42; 515-6, p. 73 ; 
 
 537, P- 123 ; 553 sqq.. p. 140 ; 556, 
 p. 128 ; 572, p. 105 ; 574, p. 73 ; 
 597, p. 61 ; 617, p. 127 ; 666 sq., 
 p. 52 ; 685, p. 130 ; 696, p. 126 ; 
 
 709, p. 73 ; 738. p. 130- 
 
 Poen. 5, p. 149 ; 89, p. 126 ; 121 sqq., 
 p. 43 ; 140, p. 92 ; 163-4, p. 73 ; 
 217 sqq., p. 52; 300 sqq., p. 53; 
 328, p. 132; 331, p. 117; 342-3, 
 p. 62; 365, p. 33, p. 106; 390, 
 
 P- S3; 443. P- »27; 448, p. 132; 
 
152 
 
 Index of Lines Discussed. 
 
 453 sqq., p. 133; 457»-% P- 541 
 474, p. 92, p. 115, p. 126 ; 405 
 sqq., p. 133 , 497. P- 123 ; S30, p. 
 10; III i (vv. S43-6, 567-77), P- 
 54; 648, p. 73 ; 651, p. 130; 669, 
 p. 105 ; 670, p. IIS ; 68s, P- i^S ; 
 694, p. 128; 706 sqq., p. 44 ; 746, 
 p- 33; 782, p. 132 ; 862, p. 131 ; 
 876, p. los ; IV ii fin. (vv. 917 
 sqq.), p. S4; 930> P- 28 ; 930 sqq., 
 
 . p. 44; 969, p. 109; 982, p. 132; 
 988, p. 123; 1005, p. 112; 1009, 
 p. I2S ; 1042 sqq., p. 44; 1051, 
 p. 112; III3, p. 125; II16, p. 
 112; 1127, p. 127; I130, p. 131, 
 1168, p. 116; 1225, p. 112 ; 1246, 
 p. 132 ; 1249, p. 73 ; 1252, p. 73 ; 
 1272, p. 109; 129s, p. 132; 1302, 
 p. los ; 1317, p. 107 ; 1327. P- 
 132; 1332, p. 62; 1333-5. P- 44; 
 fin., p. 44. 
 
 Pseud. 1-2, p. 2 ; 26, p. 130 ; 43, p. 
 107 ; 44, p. 128 ; 8s, p. 62 ; 91, 
 P- 73; 98, p. los ; 116, p. 114; 
 132, p. in; 140, p. 74; isi, p. 
 132; 153, P- 132; 178, p. 34; 223, 
 p. 62 ; 232, p. 107 ; 23s, p. 31 ; 
 298, p. 62; 306, p. 117; 307, p. 
 66 ; 31S, p. 62 ; 317, p. 123 ; 319, 
 p. 123; 321, p. 74; 346, p. 123; 
 349, p. 128 ; 355, p. 105 ; 372, p. 
 
 45 ; 376, p- 74 ; 378, p. 74 ; 381, 
 
 p. 114; 383, P- 74; 385. P- 74; 
 389, P- 63; 392, p. 45; 393, p. 
 105 ; 397, P- 63 ; 409, p- 114; 410, 
 p. 132; 417, p. 107; 421, p. 74; 
 424, p. 128 ; 432, p. 63 ; 433, p. 
 45 ; 442, p. 117 ; 443, p. 109 ; 485, 
 p. 114; 523-4, p. 46; 527, p. 114; 
 573^ P- 74; 600, p. 114; 627, p. 
 118; 631, p. 74; 637, p. 75 ; 641, 
 p. 141 ; 657, P- 75 ■> 669, p. 63 ; 
 673, p. 125 ; 683, p. 105 ; 685, p. 
 140 ; 700, p. 63 ; 713, p. 75 ; 718, 
 
 p- 75 ; 723, p. 75 ; 740, p. 31 ; 
 762, p. 133; 788, p. 114; 864, p. 
 
 63 ; 880, p. no; 889, p. 75 ; 890, 
 p. 125 ; 897, p. 132; 901, p. 64; 
 908, p. 81 ; 939, p. 81 ; 954, p. 92 ; 
 
 955, P- 4 ; 975, P- "° ; 997. P- 64 ; 
 1073, p. 114; 1121, p. 123; 1131, 
 p. 81; 1142, p. 107; 1172, p. 34; 
 1175, p. 64 ; 1205-7, p. 79 ; 1207, 
 p. no ; 1220, p. 64 ; 1294, p. 64 ; 
 
 129s, p. 64 ; 1299, p. 64 ; 1306, 
 p. 105; 1315, sqq., p. 81 ; 1320-1, 
 p. 81; 1329-30, p. 80; 1331, p. 
 122. 
 Rud. II, p. 121 ; 525, p. 22; 526 
 sqq., p. 127 ; 577, p. 75 ; 787, p. 
 
 75 ; 888, p. 148 ; 1226, p. 65. 
 Stich. 48 sqq., p. 46 ; 77, p. 65 ; 94, 
 
 p. 75 ; 140, p. 75 ; 152, p. 123 ; 
 157, p. 46; 161 sqq., p. 55; 165, 
 p. 106 ; 166, p. 65 ; 167, p. 107 ; 
 171, p. 132; 174 sqq., p. 55; !l8o, 
 p. 128 ; 208% p. 55 ; 216, p. 128 ; 
 221, p. 125 ; 223, p. 117; 234, p. 
 130 ; 235, p. 132 ; 237, p. 65 ; 243, 
 p. 117; 253-4, p. 55; 255, p. 65; 
 262, p. 66 ; 270, p. 125 ; 282, p. 
 115; 344, p. 132; 357, p. u8; 
 371. p. 75; 373- P- 66; 374. p. 132; 
 384, p. 109; 389, p. 112; 390-1, 
 p. 66 ; 451, p. 75 ; 455, p. 107 ; 
 
 459 sqq-, p. 133 ; 477, p- 132 ; 483- 
 
 5, p. 56 ; 489, p. 128 ; 523, p. 107 ; 
 542, p. 92 ; 586, p. 66 ; 620, p. 
 116 ; 631, p. 66 ; 688, p. 56 ; 704, 
 p. III. 
 Trin. 41, p. 105 ; 48, p. 126; 61, p. 
 
 76 ; 70, p. 67 ; 72 sqq., p. 56 ; 88, 
 p. 76 ; 185, p. 125 ; 186, p. 76 ; 
 214, p. 76 ; 238a, p. 67 ; 250, p. 
 27; 251, p. 76; 255, p. 67; II ii., 
 p. 83; 328, p. 67; 335, p. 76; 340, 
 p. 35 ; 362-8, p. 47 ; 448-9, p. 76 ; 
 452, p. 76 ; 456, p. 76 ; 495, p. 
 92; 509, p. 112; 530, p. 104; 
 537, P- 76; 539-40, p. 132; 560, 
 P- 77; 720, p. 72; 743, p. 77; 
 747, P- 77 ; 773. P- I05 ; 788, p. 
 47; 839. p. 27, p. 77; 886, p. 11; 
 901, p. 56 ; 931, P- 92 ; 935, P- 
 141; 1064, p. 67; 1071, p. 107; 
 1090, p. 149; 1153, p. 149- 
 
 True. 22, p. II ; 73, p. 77 ; 120, p. 
 82 ; 123, p. 82 ; 152, p. 82 ; 189, 
 P- 77; 197, P- 68; 200, p. 130; 
 202, p. 77 ; 203, p. 82 ; 209, p. 83 ; 
 238, p. 68 ; 245, p. 68 ; 246 sqq., 
 p. 47; 259, p. 77; 261, p. 107; 
 271, p. 15; 281, p. 77; 289-90, 
 p. 68; 304-5, p. 78; 323, p. II ; 
 363, p. 107 ; 369, p. 68 ; 374, P- 
 68 ; 375, p. 69 ; 378, p. 78 ; 381, 
 P- 78 ; 577, P- 104 ; 690, p. 92- 
 
 iPrfnUft bs 3ame« patfter aiiS Co., (ttown ffiavb, ©rtort.