at the opening of book I). If the common original was marked in some way like the Leyden MS, we can well understand how the two divergent groups arose. We thus obtain from our MS a valuable hint for the history of the transmission of the text of our author.

Another point in which L throws light on the nature of the original MS is the treatment of the lemmas, the method of indicating that a new word was subject of discussion in a new paragraph. If these head-words had been indicated by initials (rubricated or not) in the original, we can hardly imagine a calligraphic copy like L ignoring this treatment. But it is not till fol. 15 r. in L that the use of rubricated initials begins. In the earlier pages a horizontal stroke is drawn above the lemma-word by the corrector, e. g. fol. 5 r., above Inlicere 6 M. 15. Traces of the (at least occasional) absence of indication of the lemma in the archetype are seen in corruptions, shared by all MSS of Nonius, like cintinnire for tintinnire 40 M. 12, tibicidas for cibicidas 88 M. 8, for it is in their minuscule, not their majuscule or initial, form that the letters c and t are liable to confusion. There is one miswriting of a lemma which points to majuscule script-the corruption gladatores for glaratores (gralatores 'walkers on stilts'), 115 M. 18, with D for R. It may date from some proto-archetype whose whole text was in majuscules.1

The omission of lemmas in the Extract MSS is often due to the absence of an explanation of the word treated, e. g. adsestrix 73 M. 29, although sometimes, as we have seen, the compiler of the original of the Extract-group has added an explanation of his own; e. g. 167. 20 Reda [vehiculum]. But undoubtedly another cause lay in the absence in the archetype of any indication of the new lemma. Thus 33 M. 10 Pedetemtim has a small p and no indication of a new paragraph in L', and presumably this indication was lacking also in the archetype. The lemma is passed over in the Extract MSS.

(On traces in L of the use in the archetype of c or caput to indicate a new paragraph see Philologus, LV 167.)

Some peculiarities of the spelling of the archetype which are revealed to us by L' (and Gen.') have been already mentioned, such as the use of e for y. This barbarism is the cause of the erroneous reading rex for Eryx in 302 M. 33, where L' has erex,

1 The form glaratores may be the form of the word actually used by Nonius, a popular form like Phyrgio for Phrygio, etc.

and was probably already corrected in the archetype in 237 M. 8, where the quid dante tyranno of our MSS, instead of quiddam

ty

tyranno, seems due to a suprascript correction, teranno. The late Latin use of ui for y, from which our name for the letter is derived, we have already found in the archetype. It has led to the corruption virum for gyrum in 252. 18, where L' reproduces the spelling of the archetype, guirum.

That the script of the archetype was minuscule we see from the confusion of letters like c and d (e. g. 361. 6 hercle] haec de L', herde Gen.'), a and u (e. g. maliorum L' for mulierum).

We may safely assign to the archetype some peculiar contractions, which are reproduced by L' and Gen.' (e. g. supl with horizontal line above for suppliciis in references to Cicero's Verrine oration de suppliciis, e. g. 271 M. 25), or of which we find clear trace in these transcripts. For example, the curious reading of L1, sati for senati, 130 M. 10, which the corrector 'corrects' to satis, suggests that in the archetype the unusual contraction s with horizontal stroke above was used for sen, just as m with horizontal stroke above is the common contraction of men. And this suggestion is supported by the corruption in our MSS at 312 M. 38 sensu iacerent] sed subiacerent H', subiacerent LVH'. Similarly ostari for ostentari in our MSS at 539 M. 2 may be due to a like contraction of the syllable ten. In 269 M. 35 consentire appears in certain MSS as consistere; in 392 M. 29 evenit appears as evit. In an article in Philologus, already referred to, I have mentioned some other contractions which may with more or less probability be ascribed to the archetype (LV 168).

I will conclude this paper by pointing out a possible feature of the archetype of which we seem to find traces in L: I mean the indication of a word by its initial letter or its first syllable merely, in cases of repetition. At 353 M. 5 sqq. we have the verb niti exhibited in its various meanings: niti est conari... niti, fultum esse, etc. At the second occurrence of niti we find merely ni in Gen.', while the verb is omitted by L'. At 162 M. 1 we have the lemma Permittere, with two examples of the verb from Sisenna. In the second example: multi praemissis armis ex summo se permitterent, we find perm representing permitterent in L. Again at 93 M. 24 (lemma Continuari) L omits the verb continuatur in the example from Sisenna, I fancy, because c stood for continuatur in the archetype. In this way I would explain the corruption at 66 M. 4 Excordes concordesve (vae L') ex corde, where the

example from Cicero includes vecordes (vaecordes L) as well as excordes and concordes. The archetype had: excordes concordes vae (i. e. vaecordes). Similarly at 175. 25 sqq. the words, Subsicivum positum succedens succidaneum, had been misinterpreted as Subsicivum, positum. Succedens, succidaneum. In the appended example from Cicero subsicivis, written, presumably, s. or su. in the archetype, has become succedens in our MSS. The omission of sumet in the Lucilius example in the lemma Sumere 395. 31 sqq. may be accounted for in like fashion. If this brief indication of a repeated word was really a feature of the archetype, light is thrown on the corrupt readings of our MSS in 167. 6 and 229. 13. At 408. 37, where tangere (with Acc. of person, Abl. of thing) in the sense of circumvenire 'to cheat' is illustrated by an iambic trimeter passage of Turpilius (129 Ribb.):

hoc quaero; ignoscere

istic solentne eas minoris noxias,

terum si forte quasi alias res uini cauot,

the verb tangere does not appear in the example. Editors have found a place for it by changing cavo to tago, although this second aorist form of tango is certain only in the subjunctive mood (ne attigas, attigat, etc.).

It may be that the omission of the verb is due to its having been represented in the archetype by its initial letter merely, in which case cavo may be regarded as a corruption of cado. true reading may be

erum si forte, quasi alias, vini cado
tangam,

The

the word res being a gloss on alias, which, however, is really the adverb, on other occasions.'

UNIVERSITY OF ST. ANDREWS, SCOTLAND,

W. M. LINDSAY.

III.-THE IEPEIAI OF HELLANICUS AND THE BURNING

OF THE ARGIVE HERAEUM.'

The testimony of Pamphila in Aulus Gellius, XV 23, to the relative ages of Hellanicus, Herodotus, and Thucydides, even though based on Apollodorus, the pupil of Aristarchus and Panaetius, may be, and probably is, factitious in its exact figures. It may have been a mnemonic device of some helpfulness to have Hellanicus sixty-five, Herodotus fifty-three, and Thucydides forty years of age at the outbreak of the Peloponnesian war, giving those who stop to reckon the problem out the years 496, 484, and 471 respectively as the natal years of the three great historians; but the mnemonic device must not be made to serve, and probably was never intended to serve, as an exact chronological canon, especially when authentic literary remains of the historians (such, for instance, as those preserved for us in the Scholia on Aristophanes, Ranae, 694 and 720) give distinct and clear chronological evidence which is at least difficult, though not impossible, to bring into harmony with the exact figures of the canon. There can be no reasonable doubt that Hellanicus described with considerable detail the events of the year 407/6 B. C., when Antigenes was Archon Eponymous at Athens, and that he did this in his Atthis. If we cling to the date 496 as that of his birth, then we must be prepared to allow that he was productive as a historian when past his ninetieth year. This, to be sure, is no more incredible than that Isocrates should finish his Panathenaicus in his ninety-seventh year, and is by no means a fatal demand upon

1 LITERATURE.—Mueller, Fragmenta Hist. Graec. I, pp. xxiii-xxxiii, 45–69. 1876: Diels, Chronologische Untersuchungen über Apollodors Chronika, Rhein. Mus. XXXI, pp. 48-54. I, pp.

1876: Wilamowitz-Moellendorff, in criticism of the above, Hermes, XI, 291-4.

1888: Niese, Die Chronik des Hellanikos, Hermes, XXIII, pp. 81 ff. 1892: Eduard Meyer, Forschungen zur alten Geschichte, I, pp. 117–21. 1893: Gutschmid, Kleine Schriften, IV, pp. 316-26.

1893: Wilamowitz-Moellendorff, Aristoteles und Athen, I, pp. 260–90; II, Pp. 19 f.

1893: Busolt, Griechische Geschichte, 13, pp. 151 ff. 1895: Wachsmuth, Alte Geschichte, pp. 510 f., 555 f.

our credulity. But it is not at all necessary to fix upon the year 496 as the exact year of his birth. The testimony of Pamphila may be not exactly, but generally true; in the words of Aulus Gellius, "Hellanicus, Herodotus, Thucydides, historiae scriptores, in isdem temporibus laude ingenti floruerunt et non nimis longe distantibus fuerunt aetatibus."

Grant to Hellanicus, then, a length of days much less than that of Isocrates, and he may have been a slightly older contemporary of Herodotus, a much older contemporary of Thucydides, and may have survived even the latter, as he undoubtedly did the former. Dionysius of Halicarnassus (ad Pomp. 3; de Thuc. iud. 5), whom Diels calls "der genaueste Kenner der Logographie," and Plutarch (de mal. Herod. 36; Theseus 26) thought of him as preceding Herodotus; Wilamowitz-Moellendorff insists on ranking him after Herodotus. Both views may be in a measure right. Such a work as the Persica of Hellanicus may well have been composed before Herodotus had published his history; the Atthis of Hellanicus must have been published, at least in its ultimate form, long after the death of Herodotus. Thucydides certainly, and Herodotus probably, drew much material from prior works of Hellanicus, though both looked down upon his methods as far inferior to their own.

The multiplicity of the works of Hellanicus, even after subtitles have been merged as far as possible under main titles, bespeaks a literary career of extraordinary length; so does the great variety in form and method employed by this historian. He never attained the art of throwing mythical and historical material into progressive and climactic epic form, as Herodotus did; or into progressive and climactic dramatic and rhetorical form, as Thucydides did. But it is clear from the fragments of his works now before us that he passed through the horographical, chorographical, and genealogical methods of composing sectional history, up to the method of the general Hellenic chronicle and annal. Beyond the last method, in spite of the brilliant example of Herodotus, he never advanced.

The horographical Lesbiaca naturally precedes and merges into the chorographical Aeolica, and this into the chorographical and genealogical Troica. Of the ten larger works that are with certainty to be attributed to Hellanicus, none is wholly exclusive of the others either in method or material. It is clear that he worked over again much of his material as he passed from one predominating method of composition to another, or