APPENDIX II: THE METRE OF THE ARGONAUTICA.

The most characteristic features of the metre of the Argonautica are the rarity of elision, the constant occurrence of hiatus, the prevalence of the bucolic diaeresis, the variation in pause and caesura, and the frequent introduction of a spondee in the fifth foot often in two lines consecutively. There are 32 possible forms of the hexameter, and of these Apollonius employs 26, while Callimachus has 21, and Theocritus 28.²⁰⁷ We observe the same fondness for the use of the dactyl as marks the hexameters of Callimachus. Of 139 hexameters in the Epigrams of Callimachus 45 have dactyls in every foot except the sixth; so too in the Argonautica about one line in every four has the same peculiarity. Other favourite combinations with Apollonius are dsddds (about one line in five), sdddds and dddsds (one in ten).

The following special points may be noticed:--

First four feet.

Out of the first thousand lines of the poem 402 begin with two dactyls, 313 with a dactyl followed by a spondee, 180 with a spondee followed by a dactyl, 105 with two spondees. There are about 65 instances altogether where we find the first and second feet [p. 412]contained in two separate words, e.g. I 760 βούπαις οὔπω, II 66 νήπιοι ὕστατα, III 257 ὑψοῦ χάρματι. Where we have two spondees thus divided in the first two feet the effect in some cases is to express with added emphasis the notions of solemnity, doubt, etc., e.g. III 714 ἴστω Κόλχων ὅρκος. When the first foot is a spondee the first word in the line is most frequently a molossus, e.g. I 4 χρύσειον μετὰ κῶας, the molossus being produced at times by the use of an enclitic, e.g. I 701 ἤδη κεν.²⁰⁸ In cases where the first word is a molossus and the second foot a spondee the second word is generally an antibacchius, e.g. I 158 Νηλῆος θείοιο. Lines like I 243, where the first word is a molossus and the second word is a spondee, are much rarer. When the first word forms a spondee it is usually followed by a word of the metrical value of a choriambus, e.g. I 129 δεσμοῖς ἰλλόμενον; a molossus, e.g. I 970 ἀνδρῶν ἡρώων; or either of these lengthened by one short syllable, e.g. II 283 τάων ἀκροτάτῃσιν.

The rule, which Meyer²⁰⁹ derives from Callimachus, that a dactyl (or the first two syllables thereof) in the second foot should not be formed by the ending of a word containing three or more syllables which is connected with the first foot, is subject to about seventy exceptions in the Argonautica, e.g. III 1123 τιμήεσσα γυναιξί. Many of these exceptions occur in three cases where Apollonius, according to Merkel, allows himself greater license: (1) with proper names, e.g. I 60 Κενταύροισιν ὀλέσθαι; (2) where there is a stop, e.g. I 875 ὧς νείκεσσεν ὅμιλον:; (3) where the poet is aiming at imitative effects, e.g. I 461 πορφύρεσκεν ἕκαστα.

Apollonius avoids the trochaic division of the second foot followed by a dissyllabic (iambic) word; there are only about ninety instances in which it is found, e.g. I 541 πόντου λάβρον ὕδωρ.

In the third foot dactyls are at least six times more numerous than spondees. We notice the same preference for the dactyl in the fourth foot. In the whole poem (5835 lines) there are only 898 instances of a spondee in this position. [p. 413]

Gerhard²¹⁰ was the first to point out the avoidance of the lengthening of the thesis of the fourth foot by position in Greek hexameters. This was also noticed a few years afterwards by Wernicke on Tryphiodorus, and the rule forbidding it has come to be known as 'Wernicke's Law.' It is stated in a qualified form by Platt (Class. Rev. X 432):--"A syllable naturally short cannot be lengthened at the end of the fourth foot by position unless it forms a monosyllabic word, and unless the consonant or consonants lengthening it are part of the same word." In this form the rule seems to hold good for Apollonius with the one exception of III 1084 ἐξερέω: μάλα γάρ με καὶ αὐτὸν θυμὸς ἀνώγει.²¹¹ The ending of this line is an echo of Il. X 389 ἦ σ' αὐτὸν θυμὸς ἀνῆκε; (v. Leaf, App. N).

Merkel lays down the rule that in the Argonautica where the fourth foot is a spondee the thesis must form part of a word of more than two syllables. It is a rule that is observed in the great majority of cases, but there are many exceptions to it. Merkel gets over most of the exceptions by saying that they are due to proper names or a stop within the line, or else by treating a great number of words as if they were enclitics or proclitics, and by assuming that elision makes two words one. There are, however, several instances that cannot be thus explained away, e.g. δύω υἷες Βορέαο (I 1399, etc.), ἵνα ζώων ἀκάχοιτο (II 191), πάλιν χρειὼ ἀλιτέσθαι (II 390), ἄφαρ βωμὸν τετύκοντο (II 694), ὃν αὐτὴ Γαῖ' ἀνέφυσεν (II 1209), κακῶν ἢ ἔνθα γένωμαι (III 771). In I 60, IV 49, 556, 1720 σφέας in the fourth foot is two short syllables (like πάις in I 67, etc.), though it is scanned as a monosyllable in IV 1008, 1308.

Fifth and sixth feet.

Apollonius is very fond of combining a dactylic fifth foot with the sixth in a single word, e.g. I 380 ἀμφοτέρωθεν, endings like this being found on an average in one line out of every ten or twelve.

The conditions under which a spondee is allowed in the fifth foot [p. 414]deserve attention:--(1) In 384 cases where we find a spondee in the fifth foot the fifth and sixth feet are contained in a single word, e.g. IV 1000 θωρήξεσθαι.²¹² An ending of this form is preceded by a dactyl almost invariably, but exceptions are found in I 186 ἀγαυοῦ Μιλήτοιο, 1297 πυρὸς ὣς ἰνδάλλοντο. In II 692, III 241, 508, IV 850 the fourth is the only foot not a spondee. A slight variation of the quadrisyllabic spondaic ending is afforded by lines like III 579 which ends βόες διαδηλήσωνται. There are about twelve other instances of this. (2) If the whole fifth foot, when a spondee, is not combined with the sixth in a single word, at least the thesis must be so combined, e.g. II 568 σπιλάδας τρηχείας. Other instances of this rare ending occur in I 66, 124; II 33, 296, 592, 675; IV 268, 1632, 1641. In this case too the fourth foot is usually a dactyl, the only exception being II 296 μετακλείουσ' ἄνθρωποι.

Two consecutive lines with spondees in the fifth foot are found in 34 passages, e.g. I 402-3. In IV 1191-3 three consecutive lines have this peculiarity; cf. Theocr. XIII 42-44, Cat. LXIV 78-80.

Caesura.

The principal caesuras in the hexameter are those in the third and fourth feet. A strong (masculine) caesura is a break after the ictus-syllable of the foot; a weak (feminine) caesura is between two syllables in the thesis (the two short syllables in a dactyl). We notice in the Argonautica a marked preference for that type of line which has the weak caesura in the third foot and none in the fourth. On examining the caesuras in the first book, which contains 1362 lines, we find weak caesura of the third foot alone in 559 instances; strong caesura of the third foot alone in 242; weak caesura of the third together with strong caesura of the fourth in 311; strong caesura of the third together with strong caesura of the fourth in 249.

In the Iliad and Odyssey there are over 250 instances in which there is no caesura in the third foot.²¹³ This was avoided by the Alexandrian poets. Apollonius has only two examples, both in proper names, I 176 Ἀστέριος δὲ καὶ Ἀμφίων, II 387 Ὀτρηρή τε καὶ Ἀντιόπη. Theocritus has three, VIII 61, XIII 41, XXII 72. [p. 415]

Trochaic caesura of the fourth foot, which is very rare in Homer, is not found in Apollonius. In the case of endings like I 132 Λέρνον γε μὲν ἴδμεν ἐόντα, 582 ἔδυνε δὲ Σηπιὰς ἄκρη, IV 1530 πύθεσκε γὰρ ἔνδοθι σάρκας, the monosyllables μέν, δέ, γάρ, which we accent in a conventional way, are in reality enclitics,²¹⁴ and the fourth foot forms a single rhythmic whole. Endings like the Homeric ἴθυσε μάχη πεδίοιο and πέντε κασιγνήτῃσιν are not found in the Argonautica.

A strong caesura in the fifth foot is regularly accompanied by a weak caesura in the third. Lines like I 216, καί μιν ἄγων ἕκαθεν, Σαρπηδονίην ὅθι πέτρην, are quite exceptional.

With regard to the quasi-caesura with elision at the arsis of the fifth foot there is the same conflict of views amongst editors of the Argonautica, with the same conflict of evidence in our mss., as in the case of the Homeric poems. For example, in Il. I 484 Aristarchus read σπλάγχνα πάσαντο at the end of the line, not σπλάγχν' ἐπάσαντο, which we know from the analogy of other passages was the reading of Aristophanes. Modern editors of the Iliad differ in the same way, e.g. in XIII 835 Monro and Allen read οὐδ' ἐλάθοντο, Leaf οὐδὲ λάθοντο. To judge from the consensus of L and G in a number of passages (e.g. I 234 πάντ' ἐτέτυκτο), Apollonius seems to have favoured the theory of Aristophanes, and I have followed Brunck, Merkel, and Seaton in making the slight change necessary to bring several passages into conformity with this view. Wellauer, on the other hand, was a strong advocate of the doctrine that Apollonius dispensed with the augment in all such cases rather than admit apostrophe.

Bucolic diaeresis.

In the Greek bucolic poets, Theocritus and others, we frequently find a break in the line produced by the fourth foot ending with a word. This is called the bucolic diaeresis, and it is a prominent characteristic of our poem. In the first book it occurs in 849 lines out of 1362. In such cases the general rule is that the fourth foot must be a dactyl, e.g. I 4 χρύσειον μετὰ κῶας ἐύζυγον ἤλασαν Ἀργώ. A spondee is rarely found, e.g. I 542 ἀφρῷ δ' ἔνθα καὶ ἔνθα κελαινὴ κήκιεν ἅλμη. In the 849 instances of bucolic diaeresis in the first book we have a dactyl in the fourth foot in 811. The rule that a [p. 416]molossus must not precede the bucolic diaeresis seems to be consistently observed; endings like ὑσμίνῃ δηιοτῆτος (Il. XX 245) are avoided.

Influence of the digamma.

In studying the Argonautica one is struck by the small number of instances in which the digamma seems to be a determining element in the metre. Apollonius appears to have understood little of its import, and to have cared little about it. The limited number of cases in which we may have evidence of an initial digamma are mainly taken directly and mechanically from the poems of Homer. In many of the common words which so often exhibit the f in Homer, e.g. οἶκος, οἶνος, ἔλπομαι, ἑλίσσω, ἕσπερος, ἅλις, we find no traces in our poem of the recognition of the original consonant. Furthermore, the evidence which at first sight would seem to prove the influence of the digamma on the metre is often inconclusive, inasmuch as the instances may be otherwise explained, either as cases in which the ictus preserves the length of a diphthong or vowel in hiatus, or else as cases in which hiatus is due to vowels which are not liable to elision such as the ο of πρό and the genitive endings -οιο, -ειο, -αο, the ι of τί, προτί, etc., and the -ι of the dative singular which is not elided, so far as I have observed, in the Argonautica, and very rarely in Homer. It is important to bear these facts in mind when considering the following words:--

I Words with initial f.

ἄγνυμι.

There is no trace of initial digamma in this verb where it occurs in the poem, though the aorist forms ἔαξε ἔfαξε, II 1109, and ἐάγη, III 954, presuppose it. In III 1251 we find ἀαγές ἀfαγές.

ἄναξ.

I 908 τοῖο ἄνακτος, III 1273 καταφθιμένοιο ἄνακτος, I 411 κλῦθι, ἄναξ, II 693 ἵληθι, ἄναξ (contrast IV 1411 ἵλατ', ἄνασσαι). There are seventeen instances of a contrary kind, e.g. I 422 λύσαιμι δ', ἄναξ, 968 αὐτὸς άναξ, and in ἄνασσα, ἀνάσσω, ἀνακτορίη we find no trace of f. [p. 417]

ἄστυ.

I 774 προτὶ ἄστυ, II 809 μετέπειτα πρὸ ἄστεος, IV 1179 ἰθείας ἀνὰ ἄστυ (but ἀνά τ' ἄστυ, II 1084). There are over twenty instances where the f is neglected, which is peculiar, as ἄστυ is a word in which the digamma is very consistently observed in Homer, exceptions being rare and doubtful.

εἴκω.

I 1049 ἄλλοι εἴξαντες, II 295 ὅρκῳ εἴξαντες, III 849 πυρὶ εἰκάθοι; cf. I 805, III 521, 797, IV 1658. In III 421 we find ἀνέρι εἶξαι, but in II 338 βέλτερον εἶξαι. We have also the compound forms ὑπόεικεν II 1266, ὑπόειξε IV 1676, ὑπόειξαν IV 41, as in Homer; but ὑπείξω IV 408, ὑπείξομεν II 23, after the Homeric ὑπείξομαι. Homer has also ὑποείξω and ὑποείξομαι. The root is fικ.

ἕκηλος.

IV 1778 ἀλλὰ ἕκηλοι. In I 303 and III 969 the digamma is not recognized. The form εὔκηλος ἐfέκηλος, ἔfκηλος often occurs.

ἕκητι.

I 116 Διωνύσοιο ἕκητι; cf. I 334, II 755, III 621, IV 1087. Contrast IV 1018 οὐ μὲν ἕκητι, I 773, III 266, 1060, IV 390. It is probable that the apparent instances of f in this word in the Argonautica are merely cases of hiatus with the genitival termination.

ἑκάς, Ἑκηβόλος, Ἑκάτη.

We find no traces of f in ἑκάς or Ἑκηβόλος. In IV 829 we have a short syllable lengthened before Ἑκάτη νυκτιπόλος Ἑκάτη; cf. the Homeric Ἀπόλλωνος ἑκάτοιο, Il. VII 83, etc.

ἕκαστος.

IV 1030 ἄνδρα ἕκαστον, I 339 τὰ ἕκαστα μέλοιτο. In more than twenty other verses we find hiatus before this word, but there are also over twenty contrary instances, e.g. IV 1291 δῆθεν ἕκαστος. In Homer the proportion of cases with hiatus is two to one.

ἔοικα.

I 461 κατηφιόωντι ἐοικώς; cf. I 739, 764, II 170. All these are probably only cases of the unelided -ι in the dative, as there are fifteen instances where there is no trace of f, e.g. III 1006 ἦ γὰρ ἔοικας. [p. 418]

εἴκελος, ἴκελος.

These adjectives, from the same root fικ as ἔοικα, are almost invariably found with a vowel in hiatus before them, e.g. I 544 φλογὶ εἴκελα, III 664 τῇ ἰκέλη. An exception occurs in II 600 ἡ δ' ἰκέλη.

ἔπος, εἰπεῖν.

I 705 δήμοιο ἔπος, III 81 ἤ ἔπος, 194 νέοι ἔπος, IV 1200 πυκινὸν φάσθαι ἔπος. There are twenty-four adverse instances, e.g. I 277 τοῖον ἔπος. There is one possible instance of f in εἰπεῖν, III 26 παιδὶ ἑῷ εἰπεῖν, and Apollonius freely uses ἔειπον ἔfειπον.

ἔργον, ἔοργα.

Hiatus with ἔργον is very common, as in Homer, e.g. I 662 μέγα ἔργον, III 229 θέσκελα ἔργα. There are about fifteen other instances. In eight places the digamma is neglected, e.g. I 721 Τριτωνίδος ἔργον. Apollonius uses the Homeric ταλαεργός in IV 1062, and on the analogy of it forms ὑποεργός, I 226. The only place where ἔοργα fεfοργα occurs is in the Homeric phrase οἷα ἔοργα, IV 380.

ἔργω.

The compound ἀποέργει, I 865, is for ἀποfέργει. In the simple verb we find in all cases forms from ἐέργω ἐfέργω, with the exception of ἐεργμένος, II 550, IV 1580, and εἶργε, IV 1639, which is not Homeric. In II 201 we have the Homeric imperfect ἔεργον.

ἔτος.

One instance of hiatus, I 688, τελλομένου ἔτεος. The f is shown by Lat. vetus.

ἕννυμι, εἷμα, ἑανός.

In the forms of the simple verb ἕννυμι fες the digamma is ignored, e.g. IV 1438 πελωρίου ἕστο λέοντος, yet from ἐπιέννυμι we find the Homeric ἐπιειμένος (III 45, IV 179), and apparently on the mistaken analogy of it Apollonius coined from ἵημι the strange forms διαειμένος (II 372) and καταειμένος (I 939, III 830). A proof of his inconsistency lies in the fact that he, in common with Theocritus and other Alexandrian poets, uses ἐφέσσεσθαι (I 691) and ἐφέσσατο (I 1326). In one place we find a diphthong in thesis kept long before εἷμα, III 329, ἅλις καὶ εἵματ' ἔδωκαν. There are three [p. 419]contrary instances, I 364, II 1168, IV 671. ἑανός occurs three times, and in two cases there is hiatus, IV 169 λεπταλέῳ ἑανῷ, 1155 τεινάμεναι ἑανούς.

ἰδεῖν, οἶδα.

There are only a few passages out of the many in which ἰδεῖν occurs where we may have traces of f. III 923 ἐσάντα ἰδεῖν (contrast IV 1712 νῆσος ἰδεῖν), IV 475 λοξῷ ἴδεν, 1480 ἢ ἴδεν. So too in the case of οἶδα a few instances favour f, e.g. I 508, νήπια εἰδώς, III 103 ἅλις εἰδυῖα, 243 θέσθαι Εἰδυῖαν ἄκοιτιν. The influence of the digamma is seen in the Homeric compound ἐπιίστωρ, which occurs in II 872, IV 16, 89, 1558.

ἶσος.

In Homer the f is inferred from about thirty instances of hiatus. Possible traces in our poem are I 774 ἀστέρι ἶσος, II 581 σκοπιῇ ἴσον, IV 1246 ἠέρι ἶσα, 1449 φορβάδι ἶσος. On the other hand we have many cases like IV 384 δέρος δέ τοι ἶσον ὀνείροις.

ἰάχω, ἰαχέω, ἠχήεις.

In Homer the initial f in fιfαχω is shown by many cases of hiatus with lengthening of a short vowel. In the Iliad we twice find ἴαχον ῐ in the phrase Ἀργεῖοι δὲ μέγα ἴαχον, and in twenty places ἴαχον ῑ which never has f, and the latter is the form which Apollonius uses, e.g. III 1370 Κόλχοι δὲ μέγ' ἴαχον. Before ἰάχησεν (from ἰαχέω, which is not found in Homer) we twice have hiatus, IV 592 Ἀργὼ ἰάχησεν, 640 Ἑρκυνίου ἰάχησεν. Apollonius does not use ἰαχή or ἠχή, and in every case before ἠχήεις we have a diphthong shortened, e.g. I 1308 κίνυται ἠχήεντος, whereas in Homer we find unelided vowels, e.g. Od. IV 72 δώματα ἠχήεντα.

Ιρις.

In II 286 Apollonius uses ὠκέα ̂Ἰρις, a combination which occurs nineteen times in Homer. The name may be connected with εἴρω fερ.

II Words with initial σf.

οἷ, ἕ, ὅς, etc.

Before the pronominal forms οἷ, ἕ (root sva, Lat. sui, etc.), we consistently find long vowels and diphthongs remaining long, short [p. 420]vowels unelided, and short syllables lengthened. So too we have οὖ ἕθεν, I 362, etc. In Homer we find at times short vowels lengthened in arsis before the possessive ὅς, e.g. πατέρι ᾧ, θυγατέρα ἥν, though there are about thirty passages which do not admit the f. There is no recognition of f in ὅς in the Argonaulica; the only case of unelided vowel before it is in the dat. sing., II 559. The forms ἑοῖο, ἑοῖ, etc., (σfε) show no trace of any initial consonant. In I 1176, πολλὰ ἑῇ, and III 591, χεῖρα ἑήν, the hiatus occurs in the weak caesura of the third foot, which is one of the positions in which our poet sometimes leaves vowels unelided, e.g. I 543.

άνδάνω.

There is one possible instance of f in the simple ἁνδάνω σfαδ, I 828 ναιετάειν ἐθέλοις, καί τοι ἅδοι (contrast III 350 ὥς κεν ἅδῃ). The f is presupposed in the perf. ἕαδα, I 867, etc., and in the aor. εὔαδον ἔfαδον, I 697, etc. In III 171 we find ἐπιανδάνει, which occurs once also in Homer. In all other places we have the usual Homeric ἐφανδάνει. In III 950 Apollonius uses ἐφήνδανε, while Homer uses both ἐφήνδανε and ἐπιήνδανε. The same root is probably contained in ἕδνα, I 977, II 239. Homer uses ἔεδνα as well as ἕδνα. The compound ἀνάεδνος, II 1149, implies f.

ἑκυρός, ἑκυρή.

As in the Homeric φίλε ἑκυρέ there may be a trace of the original σf (Lat. socer), so in IV 815 we find νυῷ ἑκυρή περ ἐοῦσα.

ἐσσείοντο, ἔσσυτο, νηοσσόος, ὑποσσαίνων.

In these forms Ahrens explains the double ς by original σf.

ἔτης.

We have hiatus with ἔτης (σfε-της 'one's own man') in III 1126 κασίγνητοί τε ἔται. τε, but not in I 305 ὁμαρτήσουσιν ἔται.

ὦλκες, etc.

κατὰ ὦλκας (κατὰ ὦλκα, Hom.) occurs in III 1054, 1333. Apollonius also uses αὔλακας αfλαξ, III 1347. We frequently find ὁλκός (which does not occur in Homer) with no trace of f. The Lat. sulcus presupposes original suelq (Darbishire).

II Words with initial δf.

The words δέος, δεῖσαι, etc., in which there was original δfι, often [p. 421]have a short vowel lengthened before them in Homer; so too in our poem we find instances like I 639 ἐπὶ δέος (cf. III 435 ὑποδδείσαις, 1293 ἔδδεισαν). In II 183 the last syllable of ἐπί is lengthened before δηναιόν. We have similar lengthenings in Homer before δήν δfαν, διfαν, δηρόν, and δηθά.

IV Words with initial fρ.

Apollonius follows Homeric precedent in the metrical value of initial fρ. Any short vowel may be lengthened before it. We regularly find a vowel lengthened before ῥήγνυμι and its derivatives. In the case of ῥέζω a vowel may be long or short before it, e.g. II 1022 ἐνὶ ῥέζουσιν ἀγυιαῖς, IV 1719 ἐρημαίῃ ἐνὶ ῥέζειν. The form ἔρρεξα occurs twice, ἔρεξα seven times. Twice a vowel is lengthened before ῥίζα (II 320, III 1401), once it remains short (III 857). In III 970 a vowel is lengthened before ῥιπή; in III 1020 before ῥοδέη; in IV 174 before ῥινός. In II 884 we have the compound ἀπορρίψαντες. In IV 1497 we find ἔπεφνεν ἐπὶ ῥήνεσσιν fαρν, fραν; cf. ἐύρρηνος III 1086, πολύρρηνες II 377.

Hiatus.

I Diphthongs and long vowels in hiatus.

(1) There are 219 instances in which αι, οι, ει, ευ, ου, ῳ, ῃ are shortened at the end of the first foot before a vowel or diphthong at the beginning of the following foot, e.g. II 137 νήπιοι, οὐδ' ἐνόησαν. In 76 of these instances the diphthong is a monosyllable, e.g. II 104 ἔνθα καὶ ἔνθ'. We find a long vowel shortened at the end of the first foot ten times, e.g. III 786 ἐρρέτω ἀγλαΐη. There are 101 instances of diphthongs shortened at the weak caesura of the first foot, e.g. II 203 οὐδοῦ ἐπ' αὐλείοιο; and 45 instances of a long vowel shortened in this position, e.g. II 290 δώσω ἐγών. Furthermore, there are 47 lines in which either an enclitic or other monosyllable is shortened in the first syllable of the thesis of a dactyl, e.g. II 613 ἥ οἱ ἐνέπνευσεν. On the other hand, there are 12 cases in which a diphthong is kept long in arsis before a vowel or diphthong at the beginning of the second word, e.g. II 1016 ᾗ ἔνι, this in some cases being due to the digamma, e.g. I 362 οὗ ἕθεν. Similarly a long vowel is kept long in arsis in 14 instances, e.g. II 279 ἢ αἶγας; at times, perhaps, with the help of the digamma, e.g. III 81 ἢ ἔπος. There are two instances in [p. 422]which a long vowel or diphthong is kept long in the thesis of the first foot, I 251 δειλὴ Ἀλκιμέδη, III 745 ναῦται εἰς.

(2) In 34 instances we have a diphthong shortened at the end of the second foot, e.g. II 492 ἂν δὲ Βορήιοι υἷες. Of these instances 22 are monosyllables, e.g. II 440 τηλόθεν, ὄφρα τοι υἷες. We find a long vowel so shortened in II 1115 αὐτίκα δ' ἐρράγη ὄμβρος. There are 9 instances of a diphthong shortened in the weak caesura, e.g. II 397 αὐτοὶ Κόλχοι ἔχονται, and 17 in which an enclitic or other monosyllable is shortened in the first syllable of the thesis of a dactyl, e.g. II 1165 ἀλλὰ τὰ μὲν καὶ ἐσαῦτις. On the other hand, there are 74 instances in which a diphthong in the arsis of the second foot is kept long, e.g. II 668 τοῖς ἴκελοι ἥρωες; the digamma at times being possibly a contributory cause, e.g. II 295 οἱ δ' ὅρκῳ εἴξαντες. A long vowel is kept long in arsis in 18 instances, e.g. I 163 υἷε δύω Ἀλεοῦ. Here too the influence of the digamma may be seen, e.g. I 141 ἤιε μή οἱ δῆμος. In II 160, if the reading proposed there is right, we have a diphthong kept long before a vowel in the thesis of the second foot.

(3) A diphthong is shortened at the end of the third foot in 128 instances, e.g. II 1155 τῷδε Κυτίσσωρος πέλει οὔνομα, and of these instances 120 are monosyllables, e.g. II 156 οὐταμένων ἀκέοντο, καὶ ἀθανάτοισι. We find 126 instances of a diphthong shortened in the weak caesura, e.g. II 73 ἱεμένου φορέεσθαι ἔσω; 7 instances of a long vowel so shortened, e.g. II 975 πεμπάζοι: μία δ' οἴη ἐτήτυμος; and 55 instances where an enclitic or other monosyllable is shortened in the first syllable of the thesis of a dactyl, e.g. II 24 αὐτὸς ἑκὼν ἤδη τοι ὑπίσχομαι. There are 47 instances in which a diphthong in the arsis of the third foot retains its length before a vowel or diphthong at the beginning of the next word, e.g. II 274 φάσγαν' ἐπισχόμενοι ὀπίσω θέον, and 11 similar instances in the case of a long vowel, e.g. II 735 κοίλη ὕπαιθα νάπη, ἵνα. In II 696 we have an example of a long vowel kept long in thesis, εἴ κέ τιν' ἢ κεμάδων, ἢ ἀγροτέρων ἐσίδοιεν.

(4) In 363 instances (27 being monosyllables) a diphthong is shortened at the bucolic diaeresis at the end of the fourth foot; most frequently αι (153 cases), e.g. II 11 ἴδμεναι ὔμμιν ἔοικεν, and οι (106 cases), e.g. II 37 ἐναλίγκιοι εἰσοράασθαι. There are 18 instances of η shortened, e.g. II 363 τετραμμένη αἰθέρι κύρει, and 19 of ω, e.g. II 425 περαιτέρω ἐξερέεσθε. In the first syllable of the thesis [p. 423]of the fourth foot we find καί shortened 27 times, e.g. II 369 μέγας καὶ ὑπείροχος ἀγκών. In 108 lines a diphthong in arsis is not shortened, e.g. II 55 πάλου ἄτερ ἐγγυαλίξω. At times this may be due to the digamma, e.g. IV 1658 ἐμοὶ εἴξειε δαμῆναι, IV 503 νέοι ἔπος Αἰακίδαο. The vowels η and ω are kept long in arsis 19 times, e.g. II 18 κρατερὴ ἐπιέψετ' ἀνάγκη, III 711 ῥέξω ἄκος, οἷ' ἀγορεύεις. There are further a few instances in which a diphthong or long vowel retains its length in thesis, I 72 Ἴρου Ἀκτορίδαο, 774 φαεινῷ ἀστέρι ἶσος, II 762 γενεὴν καὶ οὔνομ' ἑκάστου, III 329 ἅλις καὶ εἵματ' ἔδωκαν (where there is f), 771 κακῶν ἢ ἒνθα γένωμαι.

(5) There are 155 cases of a diphthong shortened at the end of the fifth foot, e.g. II 22 εὔχεαι εἶναι. In 94 of these the diphthong is a monosyllable, e.g. II 44 ἀλλά οἱ ἀλκή. Instances of a long vowel thus shortened are very rare, III 45 ἐπιειμένη ὤμοις, 785 ἐρρέτω αἰδώς. In the weak caesura of the fifth foot there are 78 instances of a shortened diphthong, e.g. II 6 πειρήσασθαι ἑοῖο, and 27 cases in which an enclitic or other monosyllable is shortened in the first syllable of the thesis, e.g. II 195 ὧν οἱ ἰόντων. The ictus keeps a diphthong long in 51 instances, e.g. II 40 οὐρανίῳ ἀτάλαντος. So too η is kept long ten times, e.g. II 1210 Τυφαονίη ὅθι πέτρη, and ω twice, II 390 χρειὼ ἀλιτέσθαι, 1132 ἄμφω ἱκέται τε.

(6) In I 1349 the apparent hiatus in the sixth foot μή οἱ is due to the digamma; cf. the ending γάρ οἱ in II 501, etc.

II Short vowels in hiatus.

We find, as in Homer, the vowels not liable to elision standing in hiatus, e.g. IV 127 αὐτὰρ ὁ ἀντικρύ, II 1268 πεδίον τὸ Ἀρήιον, I 332 νηὶ ἐφοπλίσσασθαι. In I 705, δήμοιο ἔπος, the hiatus would be justifiable without any assumption of f; so too in II 65 οὐδέ τι ᾔδειν, 809 μετέπειτα πρὸ ἄστεος. Hiatus with short vowels is also found, though rarely, (1) at the weak caesura of the third foot, e.g. I 543 δεινὸν μορμύρουσα ἐρισθενέων; cf. II 955, III 263, 492, 737; (2) at the bucolic diaeresis, e.g. IV 1283 κατὰ μυρία ἔκλυσεν ἔργα, I 778, II 660, IV 236, 546, 1502, 1637; (3) at the weak caesura of the fifth foot, e.g. II 779 χνοάοντα ἰουλούς, III 561 ἐρητύεσθε ἀέθλων, 1112 ἐκλελάθοιο ἐμεῖο; (4) at the end of the fifth foot, I 881 ἄλλοτε ἄλλον, III 1134 μήδετο Ἤρη.

With regard to hiatus Apollonius allowed himself a much greater [p. 424]license than Callimachus,²¹⁵ and his laxity in this respect must have been displeasing to his master. Where hiatus is found in the verses of Callimachus it is generally at the end of a dactyl with shortening of the syllable. Hiatus in the second syllable of a dactyl occurs but seldom.

Lengthening of short syllables.

In a previous section we have considered cases in which a short syllable may be lengthened before initial f, δf, σf, and fρ. Other circumstances under which Apollonius, following Homer, sometimes lengthens short syllables are--

(1) Before the spirant yod, e.g. I 1297 πυρὸς ὣς ἰνδάλλοντο (ὥς = γώς). (2) Before the spirant ς, e.g. II 415 παρὰ σέο, cf. Od. X 238 κατὰ συφεοῖσιν. The loss of the original ς in ἅλς justifies the lengthening of the first syllable in παρραλίης (IV 1560), and may also explain the lengthening in ἀλκυόνος ἁλίης (I 1096). The lengthening of the first syllable in συνεχές (for συν-σεχές), II 738, is similar. (3) Before the liquids and nasals λ, ρ, μ, ν.

λ: e.g. III 445 παρὰ λιπαρήν.

ρ: besides cases of fρ there are words like ῥέω and ῥώομαι where ῥ represents original σρ; hence such lengthenings as I 217 παρὰ ῥόον, IV 311 ποτὶ ῥόον, and the forms ἐρρώσαντο (I 385), ἐπερρώοντο (II 661, etc.), ἐπερρώσαντο (III 1258).

μ: e.g. IV 528 τρίποδα μέγαν.

ν: e.g. IV 620 τράπετο νόος.

In nearly every case such lengthenings take place in arsis; instances in thesis such as III 848 οὔτε ῥηκτός are very rare.

There are, moreover, several cases where, as in Homer, the lengthening is to be explained solely by ictus, e.g. I 289 πολέος ἐμέγηρε, 1198 ἠνορέῃ πισυνός: ἐν δέ (cf. IV 282), II 360 ἔστι δέ τις ἄκρη, IV 1398 χθόνιος ὄφις, 1422 λισσόμενος ἀδινῇ.

Synizesis.²¹⁶

In about fifty instances Apollonius employs the various kinds of synizesis common in epic poetry:--

εα: σφέας IV 1008, 1308.

εη: χρυσέην IV 729. [p. 425]

εῃ: χαλκέῃ III 218, etc., χρυσέῃ I 740, etc.

εο: χρύσεον IV 176, 1319.

εοι: χαλκέοις III 499, χρύσεοι II 676.

εω: in the gen. sing., e.g. Αἰακίδεω IV 853, ἀήτεω IV 1537; in the gen. pl., e.g. Αἰολιδέων III 339, ὑμέων IV 1031; and in two cases which are not terminations, χαλκεῶνα III 41, τεθνεώτων III 748.

εῳ: χρυσέῳ II 1271.

There are also two cases of synizesis of υω and υι in proper names, Ἠλεκτρύωνος (I 748) and Φόρκυι (IV 828), which are taken from Hesiod (Sc. 3, Th. 333).

The one exception to the rule that if the synizesis is in thesis it must be in the first or sixth foot is in III 748, where τεθνεώτων (so Stephanus, mss. τεθνειώτων) involves synizesis in the second thesis. To remove this anomaly Rzach would read τεθναότων, a form on the analogy of ἐφεσταότας (III 1276) beside ἑστηῶτας, assuming that Quintus Smyrnaeus, who uses τεθναότων, borrowed it from Apollonius, as he borrowed so much besides.

Crasis.

(1) With the article we have ὧλλοι in I 998, etc. In Homer we find τἆλλα, which Apollonius also uses (II 335). According to the Homeric scholia²¹⁷ Zenodotus was one of those who read ὦλλοι for ἄλλοι in Il. II 1, X 1, while others read ὧλλοι; in our schol. on I 1081 ὧλλοι is attributed to Zenodotus: ὧλλοι: ἡ τοιαύτη συναλοιφὴ τῆς νεωτέρας Ἰάδος (i.e. the Ionic of Herodotus) ἐστι: διὸ καὶ μέμφονται Ζηνοδότῳ εἰπόντι "ὧλλοι μέν ῥα θεοί τε καὶ ἀνέρες," οὐ κέχρηται γὰρ ταύτῃ Ὄμηρος. Our mss. differ as to the breathing, L having the rough breathing in every place, G the smooth everywhere except in II 874.

With the article we also have τἀμά, III 102. Homer has οὑμός, and Callimachus τοὐμόν.

(2) We find crasis with καί in κἀκεῖνος I 83 (where see note), 972, 996, IV 1441, and in κἀκεῖθεν IV 1731.

(3) Through crasis we get δἤπειτα (for δὴ ἔπειτα) in II 435, etc.

(4) The Homeric crasis οὕνεκα occurs frequently, e.g. I 616, as also τούνεκα, e.g. I 204. We also find the post-Homeric ὁθούνεκεν (for ὅτου ἕνεκεν) in III 933. [p. 426]

Apocope.

Apollonius follows Homer in the cutting off of final vowels before a consonant in the case of ἄρα (e.g. ἄρ κε II 1011), and the prepositions ἀνά (e.g. ἄνστησον IV 1325, ἄγκειμαι II 828, ἂμ μέγα I 127), κατά (e.g. κάτθετο III 867, κάββαλε II 34, κὰδ δ' ἄμυδις I 434), and παρά (e.g. παρσταίη III 1239, πὰρ δέ IV 223). He does not imitate the Homeric apocope of ὑπό (ὑββάλλειν = ὑποβάλλειν, Il. XIX 80) or ἀπό (ἀππέμψει = ἀποπέμψει, Od. XV 83).

Varying quantities.

Many of the cases of variation in quantity have been already noticed in the Commentary, but it is convenient to bring them together for purposes of reference.

ἀάσθην. The first α is long in arsis, IV 817, 1080; short in thesis, IV 412. In Homer it is always in thesis and short; the lengthening in arsis is found in h. Hom. Cer. 248.

ἀείδω. In this verb and its derivatives the α is short in thesis; we once find it long in arsis, IV 1399 (v.n.).

Ἄιδα, Ἀίδης. In Ἄιδα Αὔιδα, Αfιδα the initial vowel is long and in arsis, III 61, IV 1510. The forms from Ἀίδης, which are all in thesis, have it short, e.g. II 353.

ἀίω. In the pres. forms we find ᾰ; in the impf. ᾱ (I 124, II 1256). In Homer the α in the impf. is long or short. The root is αf (Curtius).

ἀλύω. ἀλύων with long penult. ends the line in III 866, but in IV 1289 ἤλυον has the penult. short.

ἀμάω. The first α is long in arsis; it is short in thesis except in III 859 (v.n.).

ἀνήρ. ᾱ in arsis, e.g. III 438, as always in the trisyll. cases ἀνέρος, etc.; ᾰ in thesis, e.g. I 182.

ἀνίη. The penult. is always long except in I 1216.

Ἀπόλλων. The first syll. is long in arsis, e.g. I 403; short in thesis, e.g. IV 612.

Ἄρης, ἄρης. As a rule the first syll. is long in arsis, e.g. III 1357, short in thesis, e.g. III 1227; but it is twice long in thesis, II 991, III 183.

Βέβρυκες. The υ is long in arsis, e.g. II 792, and once in thesis, II 70; it is short in thesis in II 98 as in Βεβρυκίη, II 136. [p. 427]

ἔνδιος. The penult. is long in thesis in I 603, short in IV 1312 (v.n.).

Θρήιξ, etc. The ι is long in arsis, e.g. I 24, 632, 1110; short in thesis, e.g. I 214, 637.

θυγάτηρ. The υ is long in arsis, e.g. II 947; it is short in thesis in the syncop. forms and in the voc. θύγατερ, III 11.

ἰαίνω. ἰάνθη (IV 24) and ἰαίνοντο (IV 1096) with ῑ are apparently augmented forms, while ἰάνθη (II 639, IV 1591) and ἰαίνετο (II 162, III 1019) with ῐ are unaugmented. In Od. XXII 59, however, the line begins with ἰανθῇ which is unaugmented.

ἴαχον, ἰάχησα. ἴαχον (e.g. II 573) has ῑ which is due to the augment; so too ἀνίαχον (II 270, III 253), ἀντίαχον (IV 76), ἐπίαχον (I 387). In the unaugmented ἰάχοντος (I 1240, 1260) we find ῐ.

In ἰάχησα (e.g. IV 592) the augment is omitted and the first syll. is short; so too ἀντιάχησα (II 828).

ἱερός. ῑ in arsis, e.g. II 1173; ῐ in thesis, e.g. II 609 (cf. ἱερῆες, II 526, etc.).

ἵημι. In the active the initial vowel is long in arsis; it is short in thesis except in II 356, IV 634. In the middle (passive) forms it is always in arsis and long.

ἱκέτης, ἱκεσίη, etc. By the side of ἱκέτης (III 987, etc.) ἱκέτις (IV 743) with ῐ Apollonius lengthens for metrical purposes the first syll. in ἱκεσίη (IV 709) and in the adj. Ἱκέσιος (II 215, etc.). The root, according to Curtius, is fικ, which helps to justify the lengthening.

ἱμάς. In arsis the ι is long (IV 890); in thesis it is once long (II 67), twice short (II 52, 63). Homer also has it long in arsis and either long or short in thesis.

ἶσος, ἴσος. ἴσος is the only form used in Homer; ἴσος may occur in Hesiod (Op. 752); Apollonius, like Callimachus and Theocritus, has ἶσος in arsis (e.g. I 774), ἴσος in thesis (II 581, cf. ἰσαζέμεν III 1045).

ὀίω. In the open form of the present the ι is always long and the word ends the line. Except at the end of the line Homer generally has it in thesis and short. For the new form ὠισάμην beside ὀίσσατο see on I 291. [p. 428]

πλημμυρίς. The υ is twice long in thesis, II 576, IV 1241, but short in thesis in IV 1269 (cf. Od. IX 486).

φᾶρος, φάρος. In Homer the α is long in both arsis and thesis; Apollonius follows Hesiod in shortening it in thesis, III 863, while in arsis he has it long, e.g. II 30.

ὕδωρ. The υ is long in arsis, e.g. I 940, II 791, and in the compound ἐφυδατίη, I 1229; it is short in thesis, e.g. II 590, 939.