STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. Classical Philology Vol. VII January, igi2 No. i ON CATULLUS By J. P. POSTGATE The following suggestions have occurred to me while engaged in reading the text of Catullus which Mr. F. W. Cornish is preparing for the Loeb series of classical texts and translations. ii. 5ff.: cum desiderio meo nitenti carum nescio quid lubet iocari. All lovers of Catullus are much indebted to Professor J. S. Philli- more for the service which he has done the poet here. I had long felt a vague dissatisfaction with the current rendering of vs. 5: ov fiav i^ra^a fidrav ek avBpa yevei&v. desiderio meo nitenti is really shocking Latin for "the bright lady of my longing love" (Munro), not only upon the grounds on which Mr. Phillimore assails it, 1 but for the added offense in nitenti. I do not wish here to under- take the somewhat ungracious office of criticizing his ingenious and not inelegant reconstruction of the poem as a whole. I confine my- self to a single detail in which amendment is possible without dis- turbing his edifice. He would write tenetur for nitenti, though apparently preferring mouetur, were that as near to the MSS. I should prefer it too, and for the following reason. One of the charms of the poem to my mind is the lightness of its touch. Catullus seems to play with his passion as Lesbia plays with her bird. Now the permanence suggested to the mind in tenetur, "is possessed by," i Classical Philology V, 217. [Classical Philolog? VII, January, 1912] 1 2 J. P. POSTGATE jars upon this fancy, if fancy it be; and I should like instead a verb with more fugitive associations, incidente is the sort of word that would express the fits of longing for her lover which Lesbia sought to beguile by this apparently aimless toying, ti might easily fall out after in=ni and the other changes would be easy. It may be noticed that line 3 ends with -enti. The appropriateness of desiderium incidit, epto? eymhrrei, scarcely needs substantiating, and elsewhere Catullus uses the verb of casual happenings, x. 5. xxxix. 10 ff. : si urbanus esses aut Sabinus aut Tiburs aut parens Vmber aut obesus Etruscus aut Lanuuinus ater atque dentatus. The amplest and up to a certain point the most satisfactory of all recent discussions of this well-known crux is that on pp. 202, 203 of G. Friedrich's Kommentar zu Catullus. First he disposes of the pretensions of parcus OG. For (1) Catullus is speaking of physical characteristics only; (2) Ellis' parallels for the application of parcus, "thrifty," to the Umbrian are illusory. For in Mart. xii. 81 "Vmber" is a personal name. Athenaeus (see below) gives the Umbrians a very different character. The right sense is that pre- served in the Vatican gloss, Mai CI. Auct. vii. 574, "pinguis grossus. nam obesus plus est quam pinguis. Catullus ait Aut pinguis ubera aut obesus et prossus." 1 Starting from the letters of parcus he would find some adjective in sense equivalent to pinguis. Rejecting fartus, which cannot be shown to be used thus, he accepted the pastus of Voss. pastus might be corrupted to parcus through partus as at Sil. It. 9. 603. It may be used adjectivally. For, unlike fartus, it is glossed by adjectives, evrpocpos, mfieXijs, Xnrapos, as well as by participles, fioo-KriOek, Tpacpefc. And it is used in literature of a "permanent condition" and hence approximates to an adjective; Mart. 12. 75. 3, "pastas g 1 a n d e nates habet Secundus"; 13. 41. 1, "lacte mero pastum"; Spect. 3.4, "uenit et epoto Sarmata pastus equo"; Ov. Am. 3. 8. 10, "sanguine pastus eques." Here the case breaks down. For the addition to pastus of the abla- tives which are printed in spaced type shows that it has its full 1 The gloss is corrupt; but why he corrupts it further by altering et prossus to et grossus I don't know. On Catullus 3 participial force, and thus the use of pastus for pinguis remains unsup- ported for the literature. In the Journal of Philology XVII, 235 I defended porcus, the correction of Scaliger, anticipated by the Itali, "h et Brixiani," Ellis ed. minor. In all respects but one this may claim to be equal or superior to pastus. But it involves a certain jolt in the expression (more apparent, however, in English than in Latin), inasmuch as Vmber porcus (adjective, noun) would have to correspond to Etruscus obesus (noun, adjective). What then are we to do? Perhaps this. We may concede to Herr Friedrich that pastus underlies parous, without however attrib- uting pastus to Catullus; or we may suppose that parcus is an unlucky attempt to fill up a gap in the text from the passage of Silius upon which Professor Ellis vainly builds "his [Vmbris] Sassina diues | lactis et haud parci Martem coluisse Tudertes," 1 8. 461. In neither case will the tradition help us toward restoring the actual word which Catullus wrote, although pastus and pinguis will corroborate each other in regard to its sense. The outward connection between a rich country and a well-fed and self-indulgent nation is attested for Umbria by Athenaeus 12. 526, to tw '0/i/3pt/ewi/ npav re e^eiv ayadrjv odev irpoeKdelv eh evSai/ioviav. And of this natural fertility the native poet Propertius (himself no Spartan) writes i. 22, fin., "Vmbria .... | me genuit terris fertilis uberibus." The adjective here applied to the land was one suitable also to the inhabitants. Sabinus Masurius, quoted by Aulus Gellius iv. 20, "equum nimis strigosum et male habitum sed equitem eius uberrimum et habitissimum uiderunt," and in Plautus Merc. 519 we have "subtemen uberius" a "stouter" thread, opposed to "sub- temen tenue." Catullus himself uses the adjective with ager, xlvi. 5, where it is corrupted to ruber. So it seems not unlikely that Catullus wrote "uber Vmber," "the plump Umbrian." If this is so, he wrote what, once in the form uber uber, could hardly have survived through so many centuries of copying, except by a miracle. pastus and pinguis will then be either explanatory glosses which have ousted the original word or metrical supplements of a defective line. 'It is clear that Silius is holding a brief for his "ruricolae Vmbri." Sassina is diues, but lactis. The Tudertes are haud parci, but in Marte colendo. 4 J. P. POSTGATE lxi.21ff.: floridis uelut enitens myrtus Asia ramulis quos Hamadryades deae ludicrum sibi roscido nutriunt umore. The trouble here is purely metrical. For, as Riese's parallel (Pliny N.H. ix. 38) shows, the collocation roscidus (rosidus) umor is unexceptionable. But it is a serious trouble none the less. The repetition in the pherecratean of the dactylic rhythm in the glyconics must be regarded as of the essence of the stanza, and the appearance of the spondee here is almost as shocking as it would be in the fourth line of the Sapphic. 1 For such contractions of short syllables in other meters of Catullus reasons are never difficult to find. The 55th poem, addressed to Camerius, which is in a five-footed measure, shows a variant line in which the dactyl of the hendecasyllable is replaced by a spondee, and L. Mueller de re metrica p. 183 (2d ed.) acutely observes "multiplicatis spondeis bene exprimitur quantis molestiis aerumnisque conflictatus sit poeta in quaerendo Camerio." In one single instance the still longer Galliambic line has all its shorts con- tracted into longs. But no one who reads aloud the final verse of Attis' lament can fail to perceive that the heavy rhythm of lxiii. 73, "iam iam dolet quod egi iam iamque paenitet," is fully justified, we may almost say demanded, by the functions which it has to perform. But what can we say here except in the words of L. Mueller, that nulla apparet causa 2 cur semel 2 in hymenaeo priore admiserit phe- recrateum talem." 3 It is not strange then that many efforts have been made to correct the line. The following are known to me: "nutriunt in odore" Guarinus; "nutriunt in agello" Pleitner; "nutriuniwr honore" Maehly; "nutriwere cruore" Weber; "nutriere liquore" H. Richards loc. tit. Perhaps the least improbable from the diplomatical point of view is Maehly 's; for ho might fall out before no and the unusual middle form might very well have its ending detached. But the phrase it offers is artificial and so inferior to 1 1 don't say "quite as shocking" because the longer the line the less obtrusive the license. 2 My italics. » Compare Mr. H. Richards' remarks in the Classical Review IX (1895) 305a. On Catullus 5 what it would supplant. Mr. Richards' suggestion is not exposed to this objection, but by the way of letter corruption there is no road from nutriere liquore to nutriunt (h)umore, nor can we make one by assuming that liquor would be glossed by umor. When a MS lection is unmeaning as well as unmetrical, the critic will attack the letters; when unmetrical only, the phrase. The verb, at least in some form or other, is admittedly unassailable. Not so the noun. We observe that the idea of moisture is already present in the epithet, and that therefore, though it may, it need not be given again in the substantive. Let us then seek for a phrase on which umore would be a reasonable gloss. We shall find it, I think, in "roscido . . . . alimento." This expression is not a direct one, and so not unsuitable for a gloss; it contains a rarity (the singular of alimenta; see the Thesaurus article), and so it is suitable again, alimentum is often applied to plants: Thesaurus, p. 1586, from which I take Seneca N.Q. iii. 11.3 "[aquas] in alimentum suum nemora ducebat"; ibid. 4, "si aquas arbusta siccarent quibus alimentum ex proximo est"; Columella iv. 2. 24; Columella ii. 2. 24 "inferiores [radices] .... facilius capient umoris alimentum"; Pliny N.H. xvii. 252, "cum alimento nimis abundare sentiuntur [ulmi]." With nutriunt also is alimentum appropriate. The Hamadryades are nutrices of the tree and in Ov. M. 10. 391 f. the nutrix "inania nudans | ubera per cunas alimentaque prima precatur"; Macrobius Sat. vii. 12. 3 conjoins the words "dum semper nouo spirandi nutriuntur alimento." Ibid. 176 fit.: illi non minus ac tibi pectore uritur intimo flamma sed penite magis. So far as I know, Baehrens (Commentary) and Professor Giri are the only two scholars of recent date who are dissatisfied with this. The former says, "flamma urit pectus, igni uritur homo" (Ovid Met. i. 495) "sic deus in flammas abiit sic pectore toto uritur"; Apul. Met. ii. 7 "ureris intime"; "nee hercle Graeco illo 'irvp Saierai' 1 (h.e. incenditur) defenditur insolens utique in sermone latino flamma 1 This is in answer to Riese, who had quoted Soph. Track. 765. 6 J. P. POSTGATE uritur," and I have nothing to add except to note the irrelevance of Ellis' citation of colores uruntur, inuruntur. Baehrens' correction was "ilia non minus hie tibi | pectws uritur intimwra | flamma," that is, "Here (where you can see your husband) your heart burns with a flame as hot as his." The conjecture changes too much, and assumes, unwarrantably, that Catullus must here be speaking of the bridegroom's feelings "which reason and the context show are more openly expressed than the bride's." As to "reason," Baehrens has not seen that Catullus is putting in an apol- ogy for a certain masculine shyness, which, as "best men" know, is apt to be in evidence on these fateful occasions; so further on the bridegroom is commended for overcoming it, "bonum | non abscondis amorem." And I can see nothing in 173 "totus immineat tibi," or in "u. 139" to justify the statement "flagrare maritum u. 139 dictum est." The commentators have quoted (Tibullus) iii. 11. 17, "optat idem nobis iuuenis sed tectius optat," put in the mouth of Sulpicia, at which Ovid glances in his Ars (where are other allusions to the third book of the Tibullus collection) 1 "uir male dissimulat: tectius ilia cupit." But the situation is not the same as here. For Sulpicia did the courting, not Cerinthus. Giri's proposal is to write ardet in for uritur. It is true that in might easily drop away before "intimo." But why should ardet have been tampered with? The hypothesis of metrical alteration would have no support in the facts; cf., e.g., the MS reading of 191, "ad maritum tamen iuuenem." Professor Giri would doubtless have conjectured urit in if uro could be used without an object. Till there is proof of this we must all observe a similar restraint. The fifteenth-century correction was "ille non minus atque tu," which avoids all objections except two: (1) we should be glad of an epithet for flamma; (2) there seems small reason for the corruption of atque tu. I propose to meet these by accepting "ille" but changing "(h)actibi . . . . flamma" to "acri- bus . . . . flammis." I take "acribus" from xlv. 15 f. (which, we may observe against Baehrens, is put in the mouth of a lady), "ut multo mihi maior acriorque \ ignis mollibus ardet in medullis." flammis (which perhaps should be rather "flammeis," as the spelling -eis is indicated by MSS corruptions elsewhere ("acsuleis" See my Selections from Tibullus Introd. xlvi-xlix. On Catullus 7 xvii. 3, "indigneis" lxvi. 86), and that a and ei might be confused is shown by " V'"s "basiei" at vii. 9), lost s before the initial of "sed." In the previous line wrong distribution of letters into words, almost the commonest source of error in Catullus, separated ac (then wrongly aspirated; cf. lxiv. 366, 233) from rib; which became tibi by misread- ing of the compendium for us, and confusion of r with t, for which cf . xii. 1, "Mafrucine" V, "Marrucine" Catullus, xxxvi. 19, "furis" V, "raris" Catullus, 1 lxiv. 392, "Juentes" V, "ruentes" Catullus, lxxxiv. 11, "ariius" V ("artius" G, "arcius"0), "arrius" Catullus. lxiii. 14: aliena quae petentes uelut exules loca celeri. The current correction of this line is to chop off the last word, though why it got there no one condescends to explain. The rhythm of the result is less than admirable: by the contraction of two shorts in the long final of exules it halts when it ought to gallop. A better remedy, which I accepted formerly, was to omit uelut and shift celeri, or rather celere (Baehrens), in front of exules, assuming that it had been misplaced by accident as was maestast in 49. But celeri seems to be not only the right word but in the right place. It means "breui tempore" as in Catullus' contemporary Calvus fragm. 12 "frigida iam celeri superatur Bistonis ora" and is to be taken with the verbs in the following line, loca must, of course, precede it with ah obvious improvement in the rhythm. We are thus left with uelut exules, whose true character is now revealed. It can only be a gloss, and this upon nothing but "prof ugae," a much less com- mon word. The line will run aliena quae petentes profugae loca, celeri. i The twentieth century sees this corruption defended (Proc. Camb. Phil. Soc. [1910], p. 3). Here in spite of the parallel in a fragment of Greek verse preserved by Hesychius d 7 pov ir\4ws • iypouctas n-\i}pi)s, which one might have thought was enough of itself to settle the matter, the annates Volusi, pleni turis et inficetiarum, translated "full of frankincense and frigidity," are "regarded as having gone to the grocer's only to return with frankincense inside; and bag and contents together make an acceptable offering to the god of fire." But why let our mixing up of literal and metaphorical stop here ? The very next words are cacata charta. So why not assume that the annates had come back from the jakes? What to call this attempt to ruin one of the sprightliest poems ever written on a paper-spoiler, I hardly know; but its author calls it "a jest of Catullus." 8 J. P. POSTGATE Our examination of the present passage, as also that of xxxix. 11, has raised the question of the character of "V," the presumptive ancestor of all extant MSS. That the text which "V" presented had already deteriorated through omissions and additions (neither however necessarily wilful) it seems impossible to gainsay. Compare Class. Rev. XIII (1899), 438 f., "On the Veronese Codex of Catullus," where the existence of variants in "V" is dealt with. Some of the unintelligible lections in our tradition are readily explicable thus. The mysterious "guioclero" at lxvi. 6 is simply a conflation of "guro" text and "circlo" gloss. 1 A second example is lxi. 5 "iletos acuto" for "ilei acuto," where "as" the error and "ac" the correc- tion were both united in "V." The miswriting of c by s, of which there are other traces in our MSS (Ixiv. 229 "has" "V" for "ac"), may be seen in another conflation "arsinoes elocridicos," lxvi. 54, in which the Greek and Latin forms of the genitive of two declen- (a)e ) is ) sions have been fused, arsinoes ) and locridos ) . This will help us at Ixiv. 243, a place over which scholars have been long and not unreasonably divided, "cum primum inflati conspexit lintea ueli"; "iv&ati" has come from infecti, with at', i.e., atri (cf. Schwabe's critical note at lxiii. 5 matf) written over it. As Professor Vahlen, in his Opuscula Academica II, 246 (1908) de emendandis Catulli carminibus, republishes his attack on the bona fides of O's blanda in Ixiv. 139 and builds on the insecure foundation of G's nobis, I may add a word to what I have said elsewhere. We need not accuse G of interpolation here as its nobis may be, as already said, a misreading of no h' ( = non haec, a correction of nee haec in the next line) as nob\ But the genuineness of O's blanda is all but proved by the fact that the whole of the disorder in 139 is explicable as an effect of homoio- graphon, blanda having been accidentally omitted after quonda, and O putting the omitted word in its place, while G filled the gap by means of a correction intended for the next line. 7&w*.31ff.: f uribunda simul anhelans uaga uadit animam agens comitata tympano Attis per opaca nemora dux. i Rather than "cyclo (which Catullus is translating)," Tucker Classical Quarterly IV, 6. In the same article at xlv. 8 Professor Tucker, by some uncorrected oversight, On Catullus 9 The defenders of comitata tympano apparently regard it as a synonym for cum tympano. But it is not. Even in English, which permits much that is inadmissible in Latin, we could not speak of a policeman being "attended" by his truncheon or his rattle, even if he were engaged in nourishing these adjuncts of his calling, And before we can imagine the tympanum as "thought of personally" we want an answer to the question why ? Why should the tympanum be assigned any personality at all, and why especially should it be assigned this here, where it will draw our attention off the main figure, the frantic devotee Attis? Baehrens proposed "animwm agens | comitwm ante tympano" taking ante with uadit and under- standing the rest to mean "stirring his followers by the noise of his tympanum" ; the phrase he illustrated from Hor. A .P. 120 "animum auditoris agunto." Mr. Garrod, Journal of Philology XXXI, 65, suggested "comitum alta" to go with "nemora." But I believe the remedy to be very different. Comparing Lucretius 2. 618 "tympana tenta tonant palmis et cymbala circum | concaua" and Varro ap. Non. 49 "tibi tympanon inanis sonitus matri' deum | tonimus — tibi nunc semiuiri," I propose tonitante, in which the frequentative form and the alliteration 1 have full significant value. The word, like tonimus, tonescit, retonent (vs. 82 of our poem) will occur in but one passage of our Latin literature; but it has plenty of analogues, crepito, strepito, mussito, and it is by no means the only aira.% Xeyo/ievov in the Attis. With tonitante the passage gains in coher- ence. The loudly sounding tympanum is needed to guide the Galli through the opaca nemora. Ibid. 63: ego mulier, ego adulescens, ego ephebus, ego puer I see no reason to retract the objections I made long ago to the vulgate interpretation of ego mulier (compare also Classical Philol- ogy III, 258, on a similar attempt to force the Latin in lxvi. 77) or my suggestion that uir is required. But it would be better to read is made to conjecture "sinister amanti." No doubt it should have been "sinistra amanti" which would coincide with Mr. Macnaughten's conjecture {Journal of Philology XXV, 146). 1 Compare also Catullus' own alliterative lines, briv. 261 {., "plangebant aliae proceris tympana palmis | aut tereti tenues tinnitus aere ciebant." 10 J. P. POSTGATE "ego nam uir" rather than "ego enim uir" since enim occurs in the previous line. For the place of nam compare xxiii. 7, xxxvii. 11, lxiv. 301. lxiv. 116 ff.: Bed quid ego a primo digressus carmine plura commemorem ut linquens genitoris filia uultum, ut consanguineae complexum, ut denique matris, quae misera in gnata deperdita laeta omnibus his Thesei dulcem praeoptauit amorem. It is hard to see why the editors with scarcely an exception have assumed that the loss in vs. 119 occurred at the end. This has led them to introduce a spondaic rhythm, which, though this is no doubt a matter of opinion, is not particularly suitable here, and to efface the contrast between the daughter's and the mother's feelings which laeta expresses so appropriately "pulcrior fingi nequit oppositio quam matris deperditae et laetae filiae," as Baehrens says (Comm.). The "tabet" after "gnata" seems not to be used of mental distress and its effects, as tabesco is; and so we must cast about for another word: pendet offers itself. The likeness between its letters and those following in "deperdita" accounts for its loss and the word is appro- priate in both its literal and its metaphorical senses. The mother would have her arms round her daughter's neck, cf. Prop. V. i. 43, Ov. Met. i. 652, Fast. ii. 760, while her state of mind would be similar to that described in 69 ff. of this poem, "toto ex te pectore, Theseu, | toto animo, tota pendebat perdita mente." This passage must be considered in connection with a note that will form an excellent illustration of the procedures in classical research upon which I have had recent occasion to comment. It is taken from Riese's commen- tary, and runs, in English, as follows: 70. mente occurs 11, pectore 10, and corde 6 times in lxiv. In the frag- ments of the Annals of Ennius the same ablatives occur 1, 5, and 9 times and in Lucretius 17, 22, and 7 times respectively. They are therefore originally peculiar to the epic language, mens, pectus, and cor besides occur 5, 4, and 1 times in lxiv and elsewhere in Catullus 21, 8, and 4 times. Now lxiv contains 408 lines, the fragments of Ennius some 550 lines (I have made a liberal deduction for fragmentary lines from the total of 628) and Lucretius de natura about 7,400 lines. It would On Catullus 11 follow then that the " Marriage of Peleus and Thetis " has 27 instances of the "epic peculiarity" referred to in the sentence which I have italicized in about 400 lines, Ennius 15 in 550, and Lucretius 46 in 7,400; in other words, that the "Marriage" of Catullus is, in respect of this epic feature, more than twice as epic as the Annals of Ennius, and more than nine times as epic as the de natura of Lucretius. This is surprising. But more is to come. On examination it will be found that 10 instances of mente, 8 of pectore, and 4 of corde (22 out of 27) are contained in the section of the poem which runs from 52 to 253 (202 lines) or, in other words, that, in the episode of Theseus and Ariadne, Catullus is four times as epic as in the rest of the poem! The absurdity of these conclusions makes us suspect that the noteworthy phenomena have nothing to do with epos or ablatives; and our suspicions are confirmed. For we observe that other forms of mens occur 5 times in this section and only once outside it, of cor (cordi) once in and not at all outside it, and that the two words of cognate meaning, animus and sensus, occur respectively 4 and 2 times within it and not at all outside it. The final result is that this group of words is represented by 37 occurrences in the section of 202 lines and by only 7 in the remaining 206 lines of the poem. The reader will naturally find the explanation of this in the subject of the episode, and presume that in the portraying of passion the feelings and their seat will be frequently mentioned. This is true, but not all the truth. In the Attis, of a length equal to about 85 hexameters, there are 9 occurrences, which would correspond to about 22 in our section. There is therefore a considerable difference to be accounted for. The difference must be due to frequent and unusual repetition of the words. And with a writer like Catullus we must assume that such repetition has a motive. The episode of Theseus and Ariadne is in fact a subtly artistic study of feeling — passionate, forgotten, resenting, and avenged — and these repetitions are the means which the poet uses to bring into connection and into relief its various phases. I will take two examples. 145 "dum aliquid cupiens animus praegestit apisci" is answered by "simul ac cupidae mentis satiata libido est," 200 "sed quali solam Theseus me mente reliquit" is followed by "tali mente, deae, funestet seque suosque." The same device is used to keep before the reader cardinal points or significant 12 J. P. POSTGATE features in the narrative. Such are the double forgetfulness of Theseus, culpable and retributive; hence immemor 58, immemor 135, mente immemori 248; the gaze of the deserted Ariadne over the sea, prospedans 52, 1 prospicit 61 f., prospectum .... petebat (of the doomed Aegeus) 241; the black sail, infeda 1 225, infedi 1 271; the distress of Ariadne, anxia 203 and of Aegeus anxia 242. The study of such correspondences shows us that in two other passages, 212 dassi, cf. 53 dasse, and 148 metuere, cf . 146 metuurd, changes in the MSS reading have been too hastily accepted, and that in the present passage they favour an emendation which helps to bring before the imagination the despair of the mother when abandoned by her daughter and the despair of the daughter when deserted by her lover. ]xviii.57ff.: 2 qualis in aerii pellucens uertice montis riuos muscoso desilit e lapide qui cum de prona praeceps est ualle uolutus 60 per medium densi transit iter populi, dulce uiatorum salso in sudore leuamen cum grauis exustos aestus hiulcat agros. This is a simile; and in similes we are in an ideal world. And provided that all appearance of unreality is avoided, a poet is free, or rather bound, to draw the imagined figure or scene such in every respect as it must be if it is to fit the function for which he has designed it. Let us apply this principle to the present passage, not forgetting that we are dealing with the work not merely of a great poetical mind but a direct and clear-sighted one. The subject is a simple one: water is welcome to the thirsty. But Catullus has not just sketched his theme in outline; he has filled it in with all the circumstances which in real life were likely to heighten the intensity of the craving and enhance the pleasure of its satisfaction. First for the sufferer. His thirst is raging ; it is the height of torrid summer (62). He is on the high road, a pedestrian, and in a sweat. 3 He i The bearing of these correspondences on the reading of the passages has been noted by Miss G. M. Hirst in the Classical Review (1911) 108; also cedentem 53 and 249. * I have printed these lines as they stand in Riese's edition, not because I believe his version of 61 to be certain, but because it gives a reasonable sense and because none of the alternatives offered by the editors, uiatori; lasso, basso, etc., will affect any of the questions to be raised in my argument. ' Riese's salso, if we accept it, is clearly to the point. On Catullus 13 comes upon water. Where? The MSS and the editors in their train say, where this water "densi populi iter transit," that is, where it crosses the path of a thick crowd (or population). What is this crowd ? It is the crowd of the village street, answers one commen- tator. Then why does not our uiator try the village inn with the attractions depicted by mine hostess in the Copa, "quid iuuat aestiuo defessum puluere abesse?" (vs. 5). We know, of course, that in real life great thirst will make men drink anything. But there is no reason for Catullus to mar his effects by making a stream of his imagination flow "durch die Landstrasse mit ihren dichten Menschenscharen" (Riese; italics are mine). The commen- tator betrays his uneasiness by adding that "this last somewhat exaggerating expression forms a suitable contrast to the loneliness in the mountains," that is, he excuses his author for making his stream more populous, and presumably less potable, than he should, on the ground that he wished to introduce a wholly irrelevant contrast between the abodes of men and uninhabited mountains. In the latest attempt to save densi it is understood to mean "follow- ing in thick succession," "dicht hinter einander" (Friedrich; Juvenal i. 120 and Claudian Cons. Stil. i. 138 are cited). But, waiving the question whether densus populus can stand for "densa populi [i.e., hominum] series" (and I am convinced that it cannot unless this sense is imposed by the context), what is the stream of people doing here ? The situation is very different in Prop. iii. 16. 25, where the poet prays that he may not be buried on a public high road "qua facit adsiduo tramite uulgus iter." But it may be urged that if the densus populus has no business in this context populus itself has none either. Not at all. It gives a touch of reality which helps on the simile. Like populi loca Lucr. vi. 1267, it means the high road; and the stream is placed there by the poet because on such a day as is described the only walking that a man would do would be for busi- ness, and along a business route. To return to our thirsty friend. He is trudging along the highway, thinking we may suppose of cool clear springs, when suddenly his eye is caught by a glitter high up on the hillside. It is no mirage; the rocks are mossy green about it. "It is water," he cries, "water from the hills. Two hundred paces more and I shall drink!" Be not so hasty, my good friend. 14 J. P. POSTGATE You have to reckon with Haupt and his clientele. These have decreed that the streamlet's indecorous bubblings, its tumbling crystal are not for you. You must wait till a good stretch of level country has subdued its transports and raised its temperature, until, it flows past your path not cursim but sensim, "nur oder kaum merk- lich!" We will not tantalize you thus. We will allow the mountain stream to redeem the promise of its first appearance. And first we will call in the aid of another poet whose words have a noticeable similarity to these, Vergil G. 3. 522, "non qui per saxa uolutus | purior electro campum petit amnis." Vergil's stream is still making for the plain, "campum petit," not flowing through it, as Baehrens openly 1 and others tacitly assume for Catullus' streamlet. But the parallel helps us positively as well as negatively, suggesting that we need an attribute for the riuus instead of a useless and unmeaning epithet with populi. It has indeed been said (Tartara with the approval of Ellis) that "in this part of lxviii each noun has its adjec- tive, aerii montis, prona ualle, medium iter, dulce leuamen, etc." The observation is inexact. It cannot be applied to uertice nor to both uiatorum (-ori) and sudore. It is also superficial. If the lines 57-60 are examined it will be seen that the place of the riuus and its appearance in that place are described in every one in corresponding terms: (a) in aerii uertice montis (6) e lapide muscoso (c) de prona ualle (d) per medium iter populi pellucet prosilit uolutus est, praeceps transit, I have small doubt that this led Palmer, who had a fine sense for such niceties, unconsciously to conjecture that densi concealed a participle in the nominative. Our task is narrowed now. If we strip from densi the ending which it owes to popuK, we have dens to work with. Palmer offered "dens" or "dens." But neither is satisfactory, for neither helps the picture. " dens" on the other hand would do very well. It would agree excellently with Vergil's purior electro; for splendor seems to have been specially applied to bright clear water: Fron- 1 He reads per campum for per medium. On Catullus 15 tinus de aquis Romae 89 "in primis Marcia et Claudia ac reliquae quarum splendor a capite integer nihil aut minimum pluuia inquinatur si putea extructa et obtecta sint." Compare the splendor aquai of Lucr. 4. 211 and the "fons Bandusiae splendidior uitro" of Horace C. iii. 13. The cause of the corruption would be the homoiographon en in its two syllables bringing about the loss of the first. But though splendens dazzled me to begin with, I now think another participle in -dens would be equally pertinent, paleographically easier, and perhaps even more in the manner of Catullus. This is dens. For the "laughing water," bright and rippling, so refreshing a sight to our dusty traveler, we may compare the usage of Lucretius ii. 559, "subdola cum ridet placidi pellacia ponti," and i. 8 "tibi rident aequora ponti" (where Munro says, not quite adequately, that it means the same as "nitet diffuso lumine," vs. 9). Also Horace C. iv. 11. 6 (in a stanza which some have deemed spurious) "ridet argento domus." Catullus, as is well known, uses the word in his address to the wavelets of his beloved lake (xxxi. 13) and in lxiv. 284 he has a metaphorical use of it which is bolder still, "quo permulsa domus iucundo risit odore." The cause of the corruption would be the same as before, ri easily falling out after li. Ibid. 143 ff.: nee tamen ilia mihi dextra deducta paterna fragrantem Assyrio uenit odore domum. Dextra paterna with deducta is a patent falsehood as the bride's father had no part in the deductio. Riese has seen part of the truth, comparing Ixii. 60 "pater cui tradidit ipse." But the Latin cannot mean "aus der Hand des Vaters zugefuhrt." The correction is easy, "dextra de ducta paterna." For the position of de, see Lucr. vi. 855, "supera de reddere parte" and v. 703 "certa de surgere parte" (Journal of Philology XIX, 289) and Munro on Lucr. i. 841. There is no ambiguity in the position here and the reason for it (euphony) is obvious. No Roman could have failed to understand Catullus. Ixxi: si quoi iure bono + sacratorum obstitit hircus aut si quern merito tarda podagra secat, 16 J. P. POSTGATE aemulus iste tuus qui uestrum exercet amorem mirifice est a te nactus utrumque malum. nam quotiens futuit totiens ulciscitur ambos; illam affligit odore, ipse pent podagra. I have printed the piece with several of the MSS corruptions uncorrected because what I desire to draw attention to is not affected by any of those so left standing. The difficulties of the earlier part seem to have been the cause why line 4 has so long gone unamended. In the first place a te should surely be a s e. It is admirable poetic justice that your rival should "punish himself." There seems to be nothing in the objection (Friedrich ad loc.) that nancisci is not found with ab; see Enn. F r. Sc. 358. In the second place utrumque should be utrimque. The point is not that he has got two troubles but that he has got trouble on two sides, as explained in the following couplet. Note. — I take the opportunity to state that my repunctuation of Catullus briv. 383 f . (Classical Philology III, 446) had, I find, been already given by the late Professor Earle in the Revue de Phihlogie. University op Liverpool September, 1911