ἀρροιαν, ἀγώμενος, πιμπλεῦσαι, χρύσεα δῶ for δώματα, ξεις for εἷς, γέντο (for ἐγένετο), &c., and misuse of words, e.g. apoiтos of a person, yλavký for 'the sea,' xóvios meaning merely 'terrible.' Another evidence is the dative plural ois and as for ouσi(v) and awi(v), e.g. Ovηtôis péy' ovelap. These forms, though very old in the Dorian (and Cyprian), were strange to the Aeolian, and did not belong to the older Ionian dialect. Where they occur in the genuine portions, they are alterable; e.g. Op. 167 for Toîs dè may be substituted the Aeolic accusative roìs (governed by karévaσoe), which the Ionic reviser would misunderstand. On the other hand, we can explain from the Delphian dialect Toîs μèv-Toîs dè, Theogon. 538, 40, in the sense which the context requires, viz. 'on the one side'-'on the other side'; for in Delphic inscriptions ois frequently occurs with the meaning 'whither' (e.g. ois ka Oén), an interpretation which the other reading τῷ μὲν τῷ δὲ will not bear. Instead of rejecting v. 638, as he formerly did (Bezzenb xii. p. 34), Fick now explains ovderépois similarly as an adverb ('to neither side'). His hypothesis, of course, excludes forms exclusively Aeolic from the Theogony, e.g. the infinitive But these are removable, e.g. by reading v. 306 μiyêμev évì, and v. 480 μèv трáþev 48 åтɩтáλλŋν (compare pépev for pépe in Delphian inscriptions). The dative plural - εσσι in other than to stems belongs also to the Delphic (e.g. ἐμ πάντεσσι τοῖς ἀγώνεσσι); also θεν = θησαν (e.g. κατεστάθεν, διελέγεν); the shortened as representing original avs (πâσăs, "Apπviăs, Bovλas) is inconsistent with Aeolic aus, but probably belonged with similar forms (es for Evs, &c.) to the Thessalian dialect, and was borrowed by the poet himself in the 'Works,' compare vv. 663, 675 with Theogon. 60; accusatives plural rós and aúrós occur in Delphian inscriptions. The Delphic word apuara (noticed by Plutarch) meaning 'ties' or tokens of friendship,' gives the sense required in Theogon. 639, and is restored by Fick in place of appeva, so getting rid of the next line with its misplaced allusion to 'nectar and ambrosia.'

-μεναι.

Fick's reconstruction no longer rests on the linguistic evidence obtained by the experiment of retranslation. He has now formulated a second theory which, if verified, must certainly go far, as he claims, to establish the first-viz. that those portions both of the Homeric and Hesiodic poetry, which his analysis indicated as the genuine products of the earlier schools, were constructed symmetrically in regularly proportioned sections

and (in the case of the oldest poems) subdivided in strophes; while on the other hand this systematic arrangement was nearly always disturbed or neglected by the Ionian translators. The application of this theory requires a considerable reduction of the 'Menis,' and a slight reduction of the other Homeric poems as at first reconstructed. The arrangement is much more obvious in the Hesiodic poetry (where indeed it was conjectured by Gruppe), and the analysis of the latter could alone have given the clue, which Fick now employs so boldly and ingeniously. He has not investigated the Homeric Hymns from this point of view. Gemoll, the last editor, notices the recurrence of short strophes in the Hymn to Apollo, and the three successive speeches of the god (247 seqq.) measure 7 lines each. The recurrence of 5-line strophes in the Hymn to Aphrodite must be regarded as hardly more than accidental, unless the text has been largely altered. The 'Menis' with its extension, and the first 'Nostos' were constructed, according to Fick, on a somewhat complex scheme involving two units (9, 11), the 'Oitos' on a simpler system, in which each section merely numbers some multiple of 5 or 10, the 'Tisis' on a similar principle (10 being here multiplied by 11 to form the larger unit), the unmutilated portion of the original 'Telemachia' corresponds with the 'Tisis' in its dimensions, and the second 'Nostos' has a definite, though less formal structure. The arrangement of the genuine Theogony is strictly stichometric; the three books are of equal length and similarly subdivided, the unit being (as he now holds) 18. component parts of the Works,' if the Ionic additions are eliminated, disclose the same symmetry. The 'Song of the Five Ages' is exactly half the length of one book of the Theogony, viz. 72 lines, which are divided in accordance with the five-fold partition of the subject and are arranged in strophe and antistrophe. The 'Reproof' likewise measured 72 lines in 9 equal strophes. The genuine "Epya consisted of the introductory description of the twofold Eris (24 lines in short strophes), instructions to farmers in 3 divisions making 72 lines, and instructions to sailors in 48 lines; the mode of subdivision as well as the total (144) corresponding to the division of the Theogony, and going to prove so far that both poems were the work of Hesiod. The nucleus of the Hesiodic poetry that is, the actual production of Hesiod thus consists of 720 lines (Theo

The

gony 3 x 144, Reproof 72, Five Ages 72, Works 144). The earlier of the subsequent additions, in the 'mixed' dialect, but still retaining a superficial symmetry, amount to 138 lines. The remainder (the production of Cercops) is quite devoid of symmetry and likewise of poetical genius.

In determining the original dialect of the principal Homeric Hymns, Fick applies the same criteria combined with internal evidence as to the locality and audience, in accordance with his leading doctrine, that all the older Epic poetry was composed in a native, living dialect. He argues that the Hymns (iv. vi. x.) to Aphrodite were designed for a Cyprian festival, and probably to serve as an introduction to the recitations of the Cypria at Salamis, from the frequent reference to that city and to Paphos, especially in the opening and concluding lines of the principal hymn; from the mention of warchariots (oariva) as still in use (iv. 13, cf. Herod. v. 113); and from the praise of the Aeneadae, which he interprets as intended indirectly for Teucer's Salamis. He draws the same inference from the text, which, he maintains, originally preserved the digamma, apparent exceptions being due to alteration, e.g. kai eiuara, where kaì is a substitute for the peculiar Cyprian idé (cf. v. 164). The text lends itself to retranslation in the Cyprian dialect by a uniform preference for open ea, cos, aos (which forms are connected in that dialect with the retention of yod, e.g. ξεπι]α, τέρχνι]α, ἀτέλια). Dissimilar vowels (where the text is not corrupt) remain uncontracted, while ce, oo, aa, coalesce (e.g. ὁρέουσα, but ὁρῶν). The genitives in -ou are probably altered from and "Epμew, 'Ayxíoew from 'Epuav, 'Ayxíoav (a Cyprian and Arcadian form), as in the Cypria (Kinkel, 9, 4) Τανταλίδαυ has been changed to Τανταλίδου : the infinitives in -v represent Cyprian -ŋv. As regards the well-known description of the adorning of Aphrodite (iv. 55 seqq.), he supposes that the Ionian composer of the Homeric lay of Ares and Aphrodite borrowed it (Od. viii. 362-6), as it is out of place there. Gemoll argues the other way, pointing out rightly that ès Пápov (v. 59 of the Hymn) must have been copied from the Homeric line (363), otherwise it would have been ἐν Πάφῳ; and that the closing of the doors (v. 60) is an obvious addition; the phrase itself (θύρας επέθηκε φαεινάς) is one of the Homeric tags, which are very frequent in this hymn. Hence he concludes that the author of the Hymn must have known the Odyssey in its present form. I cannot, however, agree with his opinion that the com

-W,

poser of the Cypria copied his description from the same source. The only evidence he adduces is that, in the extant fragment, the Horae are added to the Charites, and the details are much more elaborated. The contrary is, I believe, the true inference. The minute and rich description in the Cypria (the missing portion of which was, no doubt, equally elaborate) must have been the original, which was poorly abridged by the composers both of the Hymn and the Lay.

As regards the Hymns to Hermes and Demeter, Fick holds that they were composed as they stand in a dialect which neglected the digamma. In both Ionic contractions are found. The name of the god, in the former, is declined Ἑρμῆς, Ἑρμέω, Epμv. But the regular use of the fuller forms οισι, ηισι for oις, αις, shows that this hymn belongs to the older Ionian school. He maintains that the latter forms, in Ionian poetry, are not older than the end of the sixth century, though they had previously been used by 'Ionising' poets, who adopted the Ionic dialect for their elegies, such as Tyrtaeus and Theognis, to whom they were familar in the Doric dialect. He accordingly rejects the fragment ascribed to the Samian Asios (Kinkel, p. 206) which contains καλοῖς and δεσμοις, and alters θεοῖς to be in a fragment of Evenus (of Paros) wrongly attributed to Theognis (Bergk, fr. 490), which in the old Parian writing would be liable to be miscopied (EOIXIIENAEIX). On the other hand, the frequent occurrence of ous and as in the Hymn to Demeter implies a later date for that hymn. Fick conjectures that it was composed (for the Eleusinian festival) by a Parian poet, from the special reference to Iambe and to Paros (in the last lines); he might have noticed, as another indication, the Ionic γλήχων for βλήχων (ν. 209). He argues more decidedly for the Ionian authorship of the Hymn to Hermes from its general comic tone, resembling that of the Homeric lay just mentioned. From the

scene of the main incident being laid at Pylos, he infers that it may have been composed for the festival of the Clarian Apollo at Colophon (a colony of Pylos), being (as he thinks) intended in honour of Apollo rather than Hermes himself. The occurrence of 'Hesiodic' phrases is not, of course, inconsistent with an Ionian origin from his point of view.

Fick follows the majority of editors in dividing the Hymn to Apollo into two distinct hymns, the first addressed to the

Delian, the second to the Pythian, Apollo. He explains the correspondences between them on the supposition that the author of the latter (composed for the Delphian festival) borrowed from the composer of the Delian hymn (Cynaethus). Now that he accepts the traditional date (Ol. 69) for this author, abandoning a conjecture which placed him a century earlier, he would presumably invert that hypothesis. But Gemoll is right, I think, in spite of so many authorities, in refusing to bisect the Hymn. All the ancient references are to one hymn. The other theory rests only on a very late writer (Aristides), who in requoting (as he evidently does) from Thucydides (iii. 104) refers the lines 169 seqq. by mistake to the end of the hymn (pooiμtov) instead of the end of the Taivos. To me it appears most probable that Cynaethus himself added to his Delian hymn some older material borrowed from Delphi, which he transposed, with additions and repetitions, into the same dialect for the benefit of his Ionian public. Thus it is clear why the digamma can be restored in most of the Pythian, but not in the Delian section; and this theory alone, I believe, explains the awkwardness which marks the conclusion of the first and the commencement of the second portion. Such a clumsy compilation is just what we might expect from the composer who 'rhapsodised the Homeric verse.'

Fick's general theory of an Ionian translation of the Epic poetry is certainly strengthened by the cumulative evidence now collected; and so is his argument in favour of the Aeolian dialect as the original vehicle of the Homeric poetry, if he has succeeded in establishing the Aeolic element in the sister (Hesiodic) epic. Few, however, will agree with him in ascribing so large a performance to one translator, even if he is justified in interpreting the statement of Hippostratus, which I have just quoted, as meaning that Cynaethus rhapsodised' the whole Iliad and Odyssey. The alternative view which I have suggested 1-that the Ionicising of the Epic poetry resulted from its reduction to writing for Ionian readers by Ionian experts in that art-seems to be confirmed by the discovery that the earlier poems (which I hold to have been composed memorially) were constructed in regular sections and strophes. For such a framework, far from being a hindrance, would be a great, if not indispensable, aid to memorial 1 See Nos. 2 and 3 of this review: element in the Iliad and Odyssey.'

The Aeolic

composition, as well as an assistance to the singer or reciter, and a partial safeguard against alteration. It is, I suppose, an article of faith in the conservative cultus of Homer, that the entire complex of the Iliad and Odyssey was held together in the mind of one poet by mere power of memory. But, in fact, even such relatively small poems as Fick's analysis supposes could hardly have been framed so consistently without some such system. On the other hand the revisers might well dispense with it, if they availed themselves of writing. And without writing the amplification and dove-tailing of the earlier materials would be totally impossible. Our Iliad and Odyssey are strictly books of highly artificial construction.

If

To

That the Ionians took the lead in extending the use of writing need not be demonstrated here. This might be inferred from their early and constant intercourse with the Phoenicians, which perhaps explains the origin of their name (if Brugsch is right in deriving it from the Semitic), and likewise the fact that their name alone reached the Hebrew geographer (Gen. x.), while there is no Oriental record of the Achaeans (excluding the Akaiusha of the Egyptian inscriptions), or their immediate descendants and namesakes the Aeolians (Axaioλeîs). this be so, it is difficult to believe that the Ionicising of Homer was not begun till so much other written poetry had appeared, and literary prose had been introduced by the Ionian philosophers and historians. pass over other arguments, the 'Epic' dialect must have become the established vehicle of written hexameter verse when it was adopted for the Delphic oracles. Unfortunately, indeed, the extant oracles, which purport to be the oldest, are not genuine, e.g. that given to Glaucus (Herod. vi. 86), which was compiled after the event, partly from Hesiod (Op. 282) and partly from current proverbs. But it may be reasonably assumed that they were uniform in dialect with the oldest of the genuine texts which we possess. The Greek of the oracles was so familiar that they appear never to have been translated into local dialects. Hendess (Oracula Graeca) rightly considers that the two quasi-Doric oracles cited in Herod. iv. 157, 159, were invented in Cyrene, but partly altered in transcription, and that the so-called oracle so often quoted, & piλoxpr ματία Σπάρταν ὀλεῖ, ἄλλο δὲ οὐδέν, is merely a Doric proverb.

G. C. WARR.

THIS is the seventh half-volume of Dr. Iwan Müller's Handbook of Classical Antiquity, the general plan of which will be already known to readers of this Review. It will only be necessary to point out how far this instalment supplies the material we look for in a manual intended for advanced students.

Its two divisions (they are practically two, as the military antiquities are inserted rather as an appendix to the first part) are handled very differently; each has great merit, but no attempt whatever has been made by the editor to enforce any kind of uniformity of treatment. Even the German of the two is as unlike as can well be in two German writers: Dr. Schiller's part being clear and readable, Dr. Voigt's awkward, tangled, and in every way embarrassing to the foreign reader.

The chief merits of Dr. Schiller's work are, first, that it presents in a compendious form the main results of Mommsen's Staatsrecht, following closely his method of dealing with the constitution, and rarely dissenting from his conclusions even in detail; only now and then are Herzog or Hirschfeld preferred to the great master, as, e.g., in the interpretation of the Lex de imperio Vespasiani as bearing on the tribunicia potestas of the early principes. Unluckily the volume was published before the appearance of Mommsen's volume on the ' Bürgerschaft,' and will doubtless have to be re-edited in order to cover the results of that work together with the forthcoming one on the Senate. Secondly, Dr. Schiller has conferred a great boon on the student by prefixing to his work a concise and extremely lucid survey of the history of research in this department of antiquity, in which the position and method of Mommsen is well explained; and beside this, at the end of each section will be found a tolerably complete bibliography of the subject, including both French and Italian works, and 'rari in gurgite vasto,' here and there an English name. But no English writer on Roman law is quoted, either here or by Dr. Voigt; Mr. Strachan-Davidson's articles are unknown,

and so is Prof. Nettleship's contribution to the history of 'jus gentium,' which Mommsen incorporated in his last-published volume. The Journal of Philology seems to be known only through Bursian's Jahresbericht, the editors of which, by the way, never appear to have read the articles in it, which they notice by title only. It is, however, both surprising and refreshing to find mention of an article on Roman fortifications in Britain by so great an antiquary as Mr. G. T. Clark.

On the whole, both in respect of bibliography and in presentation of the subject, Dr. Schiller's work is likely to be very useful, though it contains little or nothing that is new. Dr. Voigt's contribution is of a very different kind. It was doubtless a wise thought to entrust to so eminent an authority on Roman law the treatment of the private life and culture of the Roman people, seeing that the one great legacy which that people left to posterity was a legal system which did not owe its development to any one great legal mind, but arose spontaneously out of the genius of the Roman character for order and government, first in the family, then in the state. To understand how this legacy came to be left us, it is absolutely necessary to know the Roman in his every-day life within the family, in his habits of thought, in his domestic as well as in his public economy. Marquardt's admirable later volumes summed up for us what had been done earlier by himself, W. A. Becker, and others, but without directing the reader's attention to the character and spirit which is implied in this side of Roman life, and which is really the true key to the undying influence of Roman civilisation. This need has been supplied by Dr. Voigt, and in spite of its unattractive style, his work is eminently worth a most attentive reading. It falls into three parts, corresponding to the periods of the Monarchy and early Republic, the late Republic-the age of Hellenistic influence-and the Empire to Diocletian. Each period is appropriately divided into sections; but here all resemblance to a 'Handbook' ceases. As a work of reference it will be almost useless, owing to the cumbrousness of paragraph and sentence alike; even the references are given in a method which in a manual is both provoking and bewildering. Here is an example: 'an Kleidungsstücken nicht weniger als siebzehn den Griechen entlehnte (§ 23 a..

w. 00. pp. xx. zz. yy. ee. ll. 13. 22. 26. 28. 40. 41).'

In spite of such drawbacks as are indicated by this appalling parenthesis, the genuine student of Roman civilisation will find his view sensibly widened and his

insight sharpened by this attempt by the author of the Jus naturale to trace the development and decay of the Roman genius. But he must not expect to be able to use it as a handbook. W. W. FOWLER.

O-STEMS:

SING. NOM. -Ŏ8 masc., -ŏm neut. (nom. and acc.) (Celt. -Ŏ8, om). IO-Stems encroach on the I-Declension.

In Old Latin Inscriptions -os, -om are regularly found till the end of the first Punic War, when they are supplanted by -us and -um, although after u the old spelling continues till Quintilian's time, e.g. servos, mortuos. On early inscriptions we often find -o for -om, e.g. pocolo (= poculum) C.I.L. I 45, and sometimes for -08, e.g. Fourio (= Furius) C.I.L. 1 63. The faint sound of the final -m (in the acc. and elsewhere) shows itself in Elision in Latin Poetry of all ages, whereas the dropping of final -8 is found only in the earlier poetry, and in Cicero's time was regarded as 'subrusticum ' (Cic. Orat. 48, 161). The Old Latin damnas (= damnatus) in the legal phrase dare damnas esto has suffered syncope of o, while in RO-Stems commonly, and in one LO-Stem, the final -os is dropped altogether, e.g. ager, caper, famul. So with final -um in nihil(um), non (earlier noenum), IO-Stems appear on inscriptions either with the ending -ios (e.g. Manios on the Praenestine fibula) later-ius, or with is (from an earlier ies) e.g. Clodis, C.I.L. 856, Caecilis 842. The latter form sometimes loses the final -8, e.g. Claudi, Valeri, Minuci on the S. C. de Bacchanalibus. Similarly we have alis for alius in Catullus 66, 28 and on inscriptions.

In Oscan the masc. O-Stems regularly suffer syncope, e.g. horz (= hortus) Zv. 87 B, Bantins (= Bantinus) 231. In famel (quoted as Oscan by Paulus Diac. p. 87) the final syllable has been entirely dropped. Neuter O-Stems retain -om, eg. sakaraklom (=* sacraculum, sacellum) 136, terom (=territorium) 136, dunum (= donum) 107.

Masc. IO-Stems have -iis, e.g. Hosidiis 83, Viinikiis 143, Aadiriis 161, or -is, e.g. Ohtavis (=Octavius) 83, Σmedis 247; neuters have probably -im, but the single instance medicim (= magisterium?) 231 is a doubtful one.

Umbrian goes as far, or rather farther, in the same direction. Masc. O-Stems: e.g. katel (=catulus), Tab. Jg. II A 43, ager Inscr. 1, pihaz (=piatus) Tab. Ig. I B 7, Ikuvins (= Iguvinus) Inser. 7: Neut. O-Stems: e.g. esonom (= divinum), esono (with loss of m) Tab. Ig. Masc. IO-Stems: Atiersir (=* Attiedius) VII B 3, Trutitis (= Truttidius) Inscr. 3: Neut. IO-Stems: e.g. tertim (= tertium) VI B 64, terti (with loss of -m) II A 28.

In Faliscan -os appears as -o, the final -8 having been dropped, e.g. Maxomo ( = Maximus) Zv. 60, Tito 64, 76; -ios, as -io, e.g. Cauio Vetulio 75, Voltilio 76; -om as -um or -u, e.g. vootum 70, sacru 70.

So in Vestine, Vetio (= Vettius) Zv. 11, duno (donum) 11; in Marsic, Cumnios (= Cominius) 43, and on the same inscr. Pacuies (= Pacuvius), unless this be a gen., pesco (nom.sg. neut. = sacrificium) 39, dunom (= donum) 43; in Sabine, dunom 10, and perhaps hiretum (chosen ?) 10; in Pelignian, faber 14, Arghillus 25, Anaes (= Annaeus) 14, Obelies 19, Ofturies 19, Helevis Rustix (Helvius Rusticus) 15, 17, uus pritrom-e (= in venturum annum ?), (hanustu 13 and pracim-e 13 are doubtful); in Marrucine, Alies (= Allius) 9; in Volscian, pihom (= pium) 47, statom (= statum) 47, esaristrom = sacrificium ?) 47.

=

GEN. (1) eis, (2) -ī. (Celtic -i).

The first of these is the Oscan and Umbrian suffix, for O-Steins, as well as for (Consonantand) I-Stems, the second the Latin and Faliscan.

(1) Thus in Oscan we have, e.g. Lovfreis (= Liberi) 82, sakarakleis (= *sacraculi, sacelli) 136, while 10-Stems take -ies, e.g. Luvcies Cnaiviies sum (= Lucii Gnaei sum) 135. In Umbrian we get -es, by Rhotacism -er, and, with loss of final -8, e, e.g. katles Tab. Ig. II A 22, popler VI A 19, agre V B 9, and similarly for IO-Stems, e.g. Marties, Martier, Fisier, Fisie Tab. Ig.

(2) In Latin - is the earliest form of the