mr u.-:

W. I

titJJJJ

l»'t

rt:i I >.»

* S

*■ -*

i i^

t »'f..».

%' « i

^'r t

t i I

?!.

r' i * # S

^ .?• ^ f

rt

*■ fr" *' (»" f * * '^- ^.'^^**.''

".. ^-. <r, ''a '^ A ' « i ft & f

1- r r r r i ?... « ^JJJJJJJJJ '

PROCEEDINGS

OF THE

AMERICAIST ACADEMY

ARTS AND SCIENCES.

Vol. XLVI.

FROM MAY 1910, TO MAY 1911.

BOSTON:

PUBLISHED BY THE ACADEMY.

1912.

■^(s 3

r^^

SEntticrsttg '^xz«s\ John Wilson and Son, Cambridge, U.S.A.

%5i'i-

CONTENTS.

Page I. A Sttidy of the Greek Epigram before 300 B. C. By F. A. Gragg 1

11. The Reactions of Earthoorms to Acids. By S. H. IIurwitz . 65

III. The Magnitude of an Error tohich sometimes affects the Results of

Magnetic Tests upon Iron and Steel Rings. By B. O. Peirce 83

lY. Note on Kirchhoff's Laio. By G. C. Evans 95

V. The Electromotive Eorce produced in Solutions by Centrifugal

Action. By R. C. Tolman 107

VI. Feeding Reactions of the Rose Coral (Isophyllia). By F. W.

Carpenter 147

VII. On Four-Dimensional Vector Analysis, and its Application in

Electrical Theory. By G. N. Lewis 103

VIII. The Resistivity of Hardened Cast Iron as a Measure of its Temper and of its Fitness for use in Permanent Magnets. By B. O. Peirce 183

IX. The Magnetic Permeabilities at Low Excitations of Two Kinds of

Very Pure Soft Iron. By B. O. Peirce 205

X. A Revision of the Atomic Weight of Neodymium. First Paper The Analysis of Neodymium Chloride. By G. P. Baxter and H. C. CHApm 213

XI. (LIV.) On the Equilibrium of the System consisting of Calcium Carbide, Calcium Cyanamide, Carbon, and Nitrogen. By M. D. Thompson and R. H. Lombard 245

XII. The Nature of Some Supposed Algal Coals. By E. C. Jef- frey 271

IV CONTENTS.

Page

XIII. Theory of Coupled Circuits, under the Action of an Impressed

Electromotive Force, toith Applications to Radiotelegraphy . By

G. W. Pierce 291

XIV. The Action of Mercury on Steel at High Pressures. By P. W.

Bridgmax 323

XV. Infini'.esimal Properties of Lines in S^ with Applications to Circles

in S3. By C. L. E. Moore 343

XVI. The Indeterminate Product. By II. B. Phillips 363

XVII. Vector-Diagrams of Oscillating-Currenl Circuits. By A. E.

Kennelly 371

XVIII. Division of Labor among Ants. By E. N. Buckingham . . 423

XIX. yi Method for Determining Heat of Evaporation as Applied to

Water. By T. W. Richards and J. H. Mathews . . . 509

XX. The Effects of Sudden Changes in the Inductances of Electric Circuits as Illustrative of the Absetice of Magnetic Lag and of the von Waltenhofen Phenomenon in Finely Divided Cores. Certain Mechanical Analogies of the Electrical Problems. By B. O. Peirce 539

XXI. The Internal Resistance of the Lead Accumulator. By II. W.

Morse and L. W. Sargent 587

XXII. Notes on the Electrical Conductivity of Argentic Sulphide. By

II. V. Hayes 613

XXIII. On the Electromagnetic and the Thennomagnetic Transverse and

Longitudinal Effects in Soft Iron. By E. H. Hall and

L. L. Campbell 623

XXIV. On the Opacity of Certain Glasses for the Ultra-Violet. By

Louis Bkll 609

XXV. Records of Meetings 683

Ru.MFORD Premium 748

Index 749

Proceedings of the American Academy of Arts and Sciences. Vol. XLVI. No. 1. September, 1910.

A STUDY OF THE GREEK EPIGRAM BEFORE 300 b. c.

By Florence Alden Gragg.

A STUDY OF THE GREEK EPIGRAM BEFORE 300 b. c.i

By Florence Alden Gragg.

Presented by H. Weir Smyth, February 9, 1910. Received May 10, 1910.

Although there exist several collections of Greek epigrams and many treatises on individual epigrammatists, such as Anacreon, Simonides, and Plato, no one has as yet collected and arranged in chronological order the epigrams which have been preserved both on stone and in MSS. and examined them in detail with the purpose of throwing light

* Selected Bibliography together with a List of Abbreviations by which the works are cited :

AEMO = Archaeologische epigraphische Mittheilnngen aiis Oesterreich. AM = Mittheilungen des Deutscheii Archaeologischen Instituts, athenische Abtheil-

ung. AP = Anthologia Palatina.

I-IX, 563, ed. H. Stadtmiiller. Leipzig, 1894-99.

IX-end, with ap])endix, ed. F. Jacobs. Leipzig, 1814. A PI = Anthologia Planudea, ed. F. Jacobs. Leipzig, 1884.

Allen = F. D. Allen, On Greek Versification in Inscriptions. Papers of the American School of Classical Studies at Athens, vol. 4, pp. 37-204. Boston, 1888. B, Bergk, or PLG = Th. Bergk, Poetae Lyrici Graeci*. Leipzig, 1892. BCH = Bulletin de Correspondence Helleni(|ue. Boas = M. Boas, De Epigrammatis Simonideis. Groningen, 1905. CIA = Corpus Inscriptionum Atticarum.

Vol. I, ed. Kirchhoff. Berlin, 1873.

Vol. II, ed. Koehler. Berlin, 1877-88.

Vol. Ill, ed. Dittenberger. Berlin, 1878-82.

Vol. IV, ed. Koehler. Berlin, 1877-91. CIG = Corpus Inscriptionum Graecarum.

Vol. 1, ed. Boeckh. Berlin, 1828.

Vol. 11, ed. Boeckh. Berlin, 1843.

VoL III, ed. Franz. Berlin, 1853.

Vol. IV, ed. Curtius & Kirchhoff. Berlin, 1877. CRAI = Comptes Rendus de I'Academie d'Inscriptions et de Belles- Lettres. Fava = D. Fava, Gli Epigrammi di Platone. Milan, 1901. Haenel, J. De Epigrammatis Graeci Historia. Breslan, 1852. II = E. Hoffmann, Sylloge Ei)igrammatum Graecorum quae ante medium .saeculum

a. Chr. n. tertium incisa ad nos pervenerunt. Halle, 1893. Hauvette = A. Hauvette, Del'Authenticite desEpigrammesdeSimonide. Paris, 1896. Henkenrath, R. Studien zu den griechischeu Grabschriften. Feldkirch, 1896.

4 PROCEEDINGS OF THE AMERICAN ACADEMY.

on the development of this branch of Greek literature. Historians of Greek letters ^ have touched but lightly on the epigram. Haenel's book, published in 1852,^ is now out of date and, even at the time of its publication, did not do justice to its subject. Mackail, in a work as useful as it is charming,* has included only such epigrams as are

Hiller = E. Hiller, Zu den Simonideischen Epigrainmen, Philologus, 48, 229-247.

lA = Inscriptiones Argolidis, ed. M. Fraenkel. Berlin, 1902.

IGA = Inscriptiones Graecae Antiquissiniae, ed. Roehl. Berlin, 1882.

IGS = Inscriptiones Graeciae Septentrionalis, ed. Dittenberger. Berlin, 1892.

IIS = Inscriptiones Italiae et Siciliae, ed. Kaibel. Berlin, 1890.

IP = Die Inschriften von Pergamon, ed. M. Fraenkel. Berlin, 1890-95.

JHS = Journal of Hellenic Studies.

JOAI = Jaliresheft des Oesterreichischen Archaeologischen Instituts zu Wien.

Junghahn, A. A. De Simonidis Cei Epigrammatis Quaestiones. Berlin, 1869.

K = G. Kaibel, Epigrammata Graeca ex Lapidibus Conlecta. Berlin, 1878.

K(RM) = G. Kaibel, Supplementum Epigrammatum Graecoruni, Rheinisches Mu- seum, 34 (1879), 181 ff.

Kaibel, G. Quaestiones Simonideae, R. M. 28 (1873), 436 ff.

Kirchhoff, A. Zur Gescluchte der attischen Epigrammen, Hermes 5 (1871), 48 ff.

Mackail, J. W. Select Epigrams from the Greek Anthology. London and New York, 1906.

Von Mess, A. Quaestiones de Epigraramate Attico et Tragoedia Antiquiore Dialec- ticae. Bonn, 1898.

Ol3'mpia = Die Inschriften von Olympia, ed. Dittenberger & Purgold. Berlin, 1896.

Pr = Th. Preger, Inscriptiones Graecae Metricae ex Scriptoribus praeter Antholo- giam CoUectae. Leipzig, 1891.

REG = Revue des l&tudes Grecques.

RM= Rheinisches Museum fiir Philologie.

Reitzenstein = R. Reitzenstein, Epigramm und Skolion. Giessen, 1893.

Reitzenstein, R., in Pauly-Wissowa, Real-Encyclopaedie, s. v. Epigramm.

Roberts = E. S. Roberts, An Introduction to Greek Ejiigraphy, vol. I. Cambridge, 1887.

Smyth, H. W. Sounds and Inflections of the Greek Dialects ; Ionic. Oxford, 1894.

"Wagner, R. Quaestiones de Epigrammatis Graecis, Leipzig, 1883.

Weber = L. Weber, Anacreoutea. Gottingen, 1895.

Wilamowitz-Moellendorff, Ein Altattisches Epigramm, Hermes, 20 (1886), 62 ff.

Simonides der Epigramniatiker, Gottinger Nachrichten, 1897, 306 ff.

Wilhelm, A. Simonideische Gedichte, JOAI 2 (1899), 221 ff.

The lyric poets are cited by Bergk's numbers ; Aeschylus by the edition of Sidgwick

(Oxford, 1902) ; Sophocles by that of Jebb (Cambridge, 1906) ; Euripides by that of

Prinz-Wecklt'in (Leipzig, 1898-1902) ; the fragments of the tragic poets by the num- bers of Nauck, Tragicorum Graecorum Fragmenta', Leijizig, 1889.

2 I have consulted the works of Hergk, Berlin, 1872-87; Bernhardy, Halle, 1880;

K. 0. Muller(cd. Heitz), Stuttgart, 1882; Sittl, Munich, 1884-7 ; Bender, Leipzig,

1886; Croiset, Paris, 1887-99; Christ, Munich, 1905; Flach, Ges. d. Or. Lyrik

nach d. Quellen dargestellt, Tiibingen, 1883-4.

De P^jngrammatis Graeci Historia, Breslau, 1852.

* Select Epigrams from the Greek Anthology, London and New York, 1906.

GRAGG. THE GREEK EPIGRAM BEFORE 300 B. C. 5

remarkable for beauty of thought or execution. Consequently he has admitted few early epigrams and his arrangement by subject precludes any attempt at chronological order. Reitzenstein, in " Epigramm und Skolion," discussed brilliantly the nature and history of the epigram,* but his interest was chiefly for the work of the Alexandrian period and he treated the early verses only as they could be made to support his original theory about those of later date. A recent article by the same scholar ® is by far the most satisfactory presentation of the subject known to me, but the necessity of discussing the whole history of the epigram in a few pages has prevented him from giving much space to the early period or considering individual epigrams to any extent. Moreover, since the publication of the collections of Kaibel, Allen, Preger, and Hoffmann many new epigrams have come to light. These, together with numerous suggestions of various scholars, are scattered in footnotes and separate dissertations, where they easily escape the notice of the general reader.

For these reasons it seemed to me a profitable task to collect the early ^ epigrams and, so far as I could, to trace the changes which gradually took place in the nature of the epigram and its relation to other branches of literature.

Appended to this paper is a list of epigrams earlier than 300 b. c.® To the material already at hand in the various thesauri I have added such epigrams as I have my self gathered from the chief classical journals published since the appearance of Hoffmann's book in 1893. I trust that no epigram has been omitted for lack of care or diligence on my part, but, even so, I cannot hope that the list will seem complete to every reader. We possess, in the Palatine Anthology and elsewhere, epigrams which give us absolutely no clue to their age, though certain of them may seem to individual scholars to be early.^ If any of these are missed, it is because I did not feel warranted in inserting in a list from which historical conclusions were to be drawn, any epigrams which are assigned to an early date merely by the " literary feeling " of this or that scholar. On purely literary grounds it is often possible to say with comparative certainty that an epigram is later than the fifth century ; it is practically never possible to say that it is earlier than the third, for the distinctive characteristics of the epigrams composed after 400 B. c. do not make their appearance earlier, while the austere and

6 pp. 87 ff.

' Pauly-Wissowa, Real-Encyclopaedie, s. v. Epij:;ranim. ' I. e., those composed before 300 B. c. * pp. 45 flF.

» E. g. Pr 64, 65, 206. Cf. PLG 2. 377 ff.

6 PROCEEDINGS OF THE AMERICAN ACADEMY.

simple epigram belongs in greater or less degree to all periods. I have also thought it safer to omit epigrams which, though they are ex- tant only in inscriptions of a late period, are thought by some to reproduce early inscriptions.**'

At the end of the list** I have enumerated the epigrams which, in my opinion, are wrongly attributed to early poets.

When Haenel said that we ought to call no poem an epigram unless we know when it was composed *2 he meant to draw a distinction between epigram, as we commonly use the word, and fniypamia as used by the Greeks of the sixth and fifth centuries b. c. A Greek epigram ** is to most of us any short poem irrespective of the sentiment ex- pressed — complete in itself and composed in the elegiac metre. Such a poem, Haenel says, would not have been an epigram at all to the Greeks of the sixth century b. c. In this he is of course quite right, for it is clear that the early Greeks would have assented to the definition

we find in SuidaS TrdvTa ra iinypa(^6fj.eva ricri, kuv fxr) iv fitrpon flprj^e'va,

iinypap.fiaTa'XeyeTat..^^ This Continued to be the meaning of the word for a long time, for there is no proof that in Herodotus the word is " on the point of acquiring its literary sense," *^ if by " literary sense " is meant any sort of poem as distinguished from prose, and that Demos- thenes could still apply the term to prose is abundantly evident from his orations.*^ It is not until 94 A. D. that we have actual proof that the word signified a poem in elegiacs. In an inscription of that date discovered near Rome *^ we find the word fmypapp-aTa prefixed to elegiacs to distinguish them from the hexameters which precede. Even in the Palatine Anthology the word appears but twice *8 and the two verses in question both very late merely prove that the authors understood epigrams to be poems ; they are not in themselves positive evidence that the term included poems which were not inscribed.

Still, that the idea of epigram actually had changed long before 94 A. D., all agree. Collectors of imypdppaTa in the third century b. c.

" E. g. CIG 1050, 1051. See PLG 2. 238. " pp. 55 <r. " p. 18.

*' We are not concerned hero with the somewliat dilTerent coloring of "epigram " as applied to poems in Latin and other languages.

** S. V. eiriypajxixa. ^^ Mackail, p. 1.

" E. g. Or. 22. 72 ; 2t. 180. " K CIS.

^' AP 9. 342 : <I>7;^ti woKvcrTixh" iTriypdufiaTOi oi"' Kara Moi5(raj clvai /j.'ri ^-qTeir' iv aradltp SJXiX"*'.

AP 0. 360 : JldyKa\6i> iar' iirlypafi/xa rb ditTTixo" '^v hi irapfKO^t Toiis Tpe7s, ^aif/udets kovk iirlypa/xfia X^yeis.

GRAGG. THE GREEK EPIGRAM BEFORE 300 B. C. 7

included in their lists many verses that were never meant to appear on stones. Philochorus ^^ and Polemo,^^ indeed, seem to have started with the intention of gathering only inscriptions, but Polemo himself ad- mitted at least one poem of a different sort ; ^^ and for his contempo- raries " tTnypcififiara " were no longcr "inscriptiones." We learn this from Atheniiius, for when he quotes eVtypa/i/iara it is probable that he quotes them by the titles which their authors gave them. In Iledylus, 22 Nicaenetus,23 Posidippus 24 we find the name given to convivial poems, and the meaning which the word had assumed in the time of Athenasus himself is clear from many passages.25 In short, among the Greeks epigram came to have an even broader meaning than it has with us.

Under these circumstances, if we should try to trace the history of the form of literature which the Greeks from age to age called epigram, we should be met by almost insurmountable difficulties, since neither the times or the causes of the changes in meaning can be determined with any degree of accuracy. Epigram, then, in this paper wnll have its later meaning, a short, complete elegiac poem. For if we kept the earliest meaning, we should have to exclude from our consideration all verses except those on stone. This would be most unfortunate, for we are searching for the origin of a particular kind of poem, not of a name, and it is the purpose of this investigation to learn whence the later epigram had its source rather than to discover what finally developed out of the early inscription.

Therefore, for purposes of literary history, it is absurd to deny to the following verses of Theognis ^6

A(f)pov€i avdpatnoi Koi vrjirioi oi re davovras Kkaiovcr , oC8' rjjSrjs av6os airohXvfxevov,

the name which we give to these lines from the Anthology ^t

Tovf KaTakeiyj^avras yXvKepov (jidos ovKeri 6pT]va), Toiis 8' eVi npocrdoKir] ^avTus del Oavdrov.

merely because the former date from the sixth or fifth century while the latter are some centuries younger. The same is the case with many other distichs, e. g.

" Suidas, s. V. ^i\6xopos. 20 Athen. 10. 436 d.

2^ Athen. 10. 442e:*HXts Kal /xedvet. Kal i/zej^SeToi olos eKacrTov

oIkos, TOiaiJTr] Kal ffvvdiracra ir6\is. 22 Athen. 11. 473 a. 23 Id. 15. 673 b,

24 Id. 10. 415 b.

«» E. g. Athen. 2. 39 c ; 3. 125 c ; 4. 162 a ; 13. 604 f. 1069, 1070. 27 AP 11. 282.

O PROCEEDINGS OF THE AMERICAN ACADEMY.

Hj3a fioi, (f)iXt 6vfie tu^' tiv rivts hXAoi eaovrai

or

Ai yhp arep vovcrcop re Koi apyaXicov fj.f\f8a>voiiV f^rjKom-afTT] poljia Ki;^ot davarov.^^

All these we should call epigrams, if only we could be sure that they are complete in themselves.

By far the greater number of the extant early epigrams were inscribed so that it becomes necessary to examine all early metrical inscriptions, whether elegiac or not all elegiacs, whether inscribed or not that we may learn as accurately as possible what causes w^ent to the making of the later epigram.

It is easy to distinguish the inscribed epigrams (if I may be allowed the apparent pleonasm) of the fifth century and earlier, since the forms of the letters testify to their age. It is harder to be sure of those belonging to the fourth century,^© but usually the style or the content comes to help out any doubtful epigraphical evidence.^^ When, how- ever, we come to the epigrams which are preserved only in MSS. the case is quite different. In the first place the works of the Greek lyric poets are extant in so fi'agmentary a condition that very often we are unable to say whether a given distich is a complete poem i. e. an epi- gram — or a shred torn from a longer elegy. The difficulty will be obvious if we compare the following verses. Soph. 0. C. 1224-8.

fiT] cf)vvai fjLiv anavra vi- Ko. Xoyov TO 8 end 4"^*'V

^rjflU K('l6fl>, 66tVT7fp tjKfl

iToXi) 8evTfpui> (OS Td)(i(TTa

Theog. 425-428

irduTdu n^ (f)vvai iirixBovioimv apurrov pTj8 eai8fli> aCyas o^eoy rjfXiov

28 Theog. 877, 878. 29 .Mi„HUTnuis, ti.

^<* In the period before 300 R. c. no evidence as to tlie date is furnished by the arrangcnu'nt of verses in inscriptions. As early as the sixth century each verse may begin a new line and throughout the period we find epigrams where the verses are not so separated. The latter method is more common in tlie sixth century, the former in the fifth and fourth. There is no examjjle of an indented i^ntameter, un- less it be 224, v. 2, which in CIA, H, 3. 2339 and in Kumanudes, I use. Sepulch. Att. 858, is represented as indented one letter. I have not seen the stone.

" From the collcetions of Kaibel and Allen I have taken no epigrams which those editors do not definitely assign to a date earlier tlian 300 u. c.

GRAGG. THE GREEK EPIGRAM BEFORE 300 B. C.

(fivvTa 8 ono)s ojKiaTu nvXns 'At'Sao Trfptjcrai Ka\ Kfiadai TroXXiji* yriv eVa/xjjo-d/xei'oj',

AP 9. 355).

Iloirju Tis ^lOTOio Tanj] rpi^ov ; . .

Hi' a()a Toivbe bvoiv ivos atpeais, Jj ro yeve<T$ai HrjdeTTOT tj to 6aveiv avriKa TiKTOfievov.

Upon examining these passages we can say without hesitation that the first is not an epigram and that the third is, about the second we are quite at a loss. Sometimes the presence of a particle, as 8e or ydp, prevents our including such fragments among epigrams, for in the whole Anthology such a particle introduces only one epigram which is not manifestly corrupt or lacking its original beginning.32 In many cases, however, we are left in doubt, although, even so, it is only in name that they differ from true epigrams. If we should discover for a certainty that they were parts of longer elegies, they would still be of use for historical purposes, since the epigram itself is but a species of elegy.

Epigrams which have been handed down to us in MSS. seem, at first sight, to furnish three kinds of evidence by which we may determine their age. In some famous persons or events are celebrated ; some are attributed to known poets ; in others we have only the diction or the sentiment to help us. This testimony is not, however, so valuable as it at first appears. It is easy to see how uncertain the first test is, for it merely supplies us with a terminus post quern and, as a matter of fact, of the numerous epitaphs purporting to be those of men of the sixth century not one can with any probability be assigned to a date earlier than the third century. 33 The second test is somewhat surer. Still, when an epigram attributed to Anacreon 3* is found inscribed in letters considerably later than the age of that poet, and when poems in praise of the works of M)Ton 35 are assigned to the same author, it is easy to see how blind is the trail we follow. More than this, recent discoveries have proved that certain epigrams of four verses 36 preserved

^* AP 9. 547, where 5' is evidently inserted for the sake of incbiding all the let- ters of the alphabet in one and the same verse.

2' The epitaphs Pr 238-47, which celebrate early sages, are an excellent example of a set of epigrams composed probably by one author at a late period. Cf. also Pi-ejffr's note, p. 199.

** 124. (Arabic numerals refer to epigrams on pp. 46 if.)

8B Anac. 11.5, 116.

" 83 (= Sim. 96) and 125 ( = Sim. 150).

10 PROCEEDINGS OF THE AMERICAN ACADEMY.

to US in MSS. originally (i. e. in the fifth century) consisted of only two verses. The joinings had escaped the notice of critics for centuries and it is impossible to say how many more such pieces of patchwork there may be in the Anthology and elsewhere.

And yet to admit all this is not necessarily to believe with many scholars 37 that no confidence is to be placed in those MSS. which assign epigrams to definite authors. For, though the tests of author- ship which we can apply are most uncertain, still, unless we can bring forward at least highly probable arguments to the contrary, we are bound to give the benefit of the doubt to the only evidence we have. However weak may be the authority of the Palatine Anthology, it is not for us to make it actually testify against itself Therefore Reitzen- stein seems to go too far when he says, 38 " Es ist meines Erachtens unmethodisch bei dieser Art Pseudo-tradition auch nur den Beweis der Unechtheit zu verlangen."

There are two reasons why scholars incline to reject the testimony of the Anthology. In the first place they are reluctant to believe that the early poets wrote epigrams at all a reluctance which has no evi- dence to support it. When this art was so generally cultivated were the famous poets the ones to neglect it ? Because very few epigrams of these poets have come down to us, are we to reject even those that we have 1 This is to let individual conjecture weigh against probability and, indeed, against some actual evidence. In the second place many seem convinced that the scribes of the Anthology were possessed by a desire to assign every poem to too early a date. This is certainly, however, not the case, for the discovery of a number of epigrams inscribed on stones 39 has proved that, even if they are not the work of the particular poets whose names they bear in the Anthology, these names point at least to the approximate dates. In some cases epigrams are actually assigned to too late a period. For example '217 found inscribed in letters of the fourth century bears in the Anthology the name of Gaetulicus, a poet, indeed, little known to us, but, if we may judge from the other epigrams attributed to him, much later than the fourth century. Again 128, a poem certainly inscribed in the fifth

37 Wilamowitz (fioett. Nachr. 1897, p. 320) : Fiir uns ist .lie Cnwi'qwm. nnvor- nKU.Uich, (lass wir die Autoritiit selbst der alcxaiidrinisehen SaniniUinj; da audi sel.r niedrif; schiitzon, wo die Gedichte solbst keinen unmittclbaren Aiistoss geben.

KaTbel (RM 28, p. 441) : Cavendnni utiiiue igitur est lie citius uiupiain huic testi (AP) credaiiius ; meiidax est ac fallax lit iioti iiingis alter.

A. A. .Tiiiighalin, De Simonidis Cei Epigraniiiiatis Quaestioiies (Berlin, 1869), p. 30 : Paeiie nulla in his rebus tides est Antliologiae.

38 Pauly-Wissowa, p. 80. "' K- g- "5 and 125.

GRAGG. THE GREEK EPIGRAM BEFORE 300 B.C. 11

century, is assigned to no poet at all, so that we may conclude that famous names were not sprinkled over the contents of the Anthology (juite so profusely and indiscriminately as some would have us think. Indeed, in all probability, mistakes in .authorship come not so much from the perversity of scribes as from confusion and changes in arrange- ment. Finally, we ought always to bear in mind what gaps there are in our knowledge of Greek literature gaps which are nowhere wider than in our knowledge of the lyric poets. Under these conditions we ought to give the MSS. at least an unprejudiced hearing.

The third test that of style helps us less than one might at first expect. Since the approximate date of the inscribed epigrams can usually be determined with certainty, we should naturally look to them for the standard by which to judge the epigrams preserved only in MSS. But the standard they set is hardly adequate, for while the inscriptions are the work of men widely different in rank, education, and ability, many of the epigrams preserved in MSS. may be the work of famous poets and it would be unfair to deny to a great master the author- ship of a given epigram merely because it exhibits more charming sentiment, more graceful diction, more brilliant genius than do the. inscriptions composed by ordinary men. Some assistance is given us by the fact that certain formulas seem to leap into favor at certain periods, but any such evidence must be used with caution, for it may be that the original use of a phrase by a great poet gave that phrase its popularity with a later generation. *o

For our purpose it will be sufficient to determine the age of the epi- grams without considering their authorship, but from what has been said above it is evident that even this is difiicult enough. It is, there- fore, with diffidence that I have approached the task, especially when I remember that certain epigrams which recent discoveries have shown to belong to the fifth century were pronounced late by very excellent scholars, ^i If 217 and 224 had come down to us only in MSS. I ven- ture to think there would be no lack of critics to assign them to a far

" E. g. 19 (= Sappho, 119). Cf. Table III.

" Some ha%-e even denied that tliey were inscribed at all. So Kaibel (EM 28, p. 45.')): Ante omnia Simonidi abiudicanda cum Bergkio aliis [sic] epp. 95 et 96, manifesto demonstrativa.

Hauvette, p. 94 (Sim. 108) : A notre avis I'inscription ne pent etre ni de Simonide ni meme d'lin poete dii 5^ siecle ipii I'aurait composee dans les premieres annees de la guerre da Peloponnese. Cf. Kaibel, EM 28, p. 456.

Hauvette, p. 1.33 (Sim. 150) : Par sa forme, par les idees qu'elle exprime, et par .son style cette piece ne saurait . . . passer pour une inscription reelle, gravoe au debut du 5^ sieclo sur une statue dans I'academie . . . une telle formule ne convieiit (^u'i une pitce composee apres coup h I'Dccasion d'une olfrande.

12 PROCEEDINGS OF THE AMERICAN ACADEMY.

later date than the fourth century to which they actually belong. I have therefore tried to err on the side of accepting rather than of reject- ing too much. In a case where, so far as internal evidence goes, an epigram might be late or early, the burden of proof rests with those who would assign it to a date later than that indicated by ancient tradition. It is not enough for them to show that it may be late ; they must also show that in all probability it is not early. I have not rejected without specific reasons any epigram which any ancient authority assigns to a date earlier than 300 b. 0.*^

Because of the greater number of hexameter inscriptions of the sixth century, it is necessary to devote more attention to them than to later hexameters. *2 The great majority, consisting as they do of a single verse, do not exhibit that difference in feeling which distinguishes the later series of hexameters from the elegy.

In the be'ginning the Greeks used inscriptions merely as a means of informing the reader as briefly and easily as possible of the reason for setting up the stones on which they were inscribed. Iliad H. 17** is familiar evidence that some sort of epitaph was usual. Doubtless in earliest times merely the names of the dedicator and the divinity or of the dead man and his father were cut upon the stones a practice which survived in combination with the later custom. Cf IGA

149 (= H 58) KaWia AlyL6(&)oio Tti 8' eu irpaa-i^r at) Trapodwra.

Thus, although among the Greeks poetry precedes prose as a literary form, it must have been itself preceded by a ruder form of expression. The use of metre testifies to a certain degree of conscious art and there- fore we cannot wonder if in the earliest epigrams which we possess some attempt is made to adorn the bare record of facts. The earliest Greek metrical inscriptions ever composed must have represented, not the first attempts to convey certain information in writing, but the first attempts to convey that information in artistic form. Without doubt long before the time fi^om which our earliest inscriptions date poets had composed songs in memory of the dead and had celebrated offerings to the gods.*^ When the custom arose of inscribing such poems upon stone, those who could not or would not employ the services of professional poets, turned

« See pp. 55 ff. « See p. 8.

** Kal wori Tis etiTTTiffi Kal 6\l/iy6i>uv avdpihirwv

&i>Spbs fikv rdSe ffrj/xa TrdXai KarareOvtOiTOi 6v wot' ipiffTttjovra Kar^Krave <palSifj.os"EKTwp. The vocabnl.ary of this passage is echoed by the early inscriptions. Cf. 8, 88, 89. *" It is worth while to rememher that our earliest Attic inscription is metrical. See AM 6 (1881), p. 107.

GRAGG. THE GREEK EPIGRAM BEFORE 300 B.C. 13

poets themselves they wrote indocti doctique and the results are what might be expected. Hence the crudeness of some of the verses, which is to be attributed to the particular author, not to antiquity in general. Afterward, when they had become more accustomed to com- posing and had more models before their eyes, even ordinary men with no greater inspiration than their predecessors acquired greater ease of style and produced fewer rude epigrams. Again the extreme simplicity of many verses is the result of the restraint, not of the lack of skill of the authors, since the Greek of early times felt that only simplicity could be in place in approaching his gods or his dead'. So it happen^ that an epigram very probably written by Anacreon shows the same characteristics as the epigrams on the stones of the Dipylon.*® In the simplest epigrams a few common and familiar words fill out the metre, often merely forming a complete sentence of words that in the earliest times had been disconnected. So

AvcTfa (v6a8e crTJfia Trarqp 'Srjfjicou imdrjKev ^"^

adds nothing to the meaning of the earlier form Ava-ea irjficovos. In

^rdXa Sfvrapeot rov Mkei^ioi (ifi eni rvfji^co ^^

only the words enl rv^^a are added to the primitive formula. The same is the case with dedicatory inscriptions, e. g.

'AjXKtjSiof avidrjKev KidapwSos vrjaidjTrjs.^^

IGA 410 (=K 1098) perhaps shows most clearly the metrical inscrip- tion in the making.

'AX^rjvci>\^p fjwoirjaev 6 Na^toj dXX' e(ri8e(r[^df.

It is the desire to conform to the fashion of the time that has led the artist to this naive expression of pride in his work. Epithets of the gods, too, suggest a convenient method of filling out the verse, especially since the poet found them adapted to dactylic measure and ready to his hand. Examples are

Ajfti'ayo[p]fys p aveBrjKfv eKij^oXa 'ATToXXtui't.'*^ 'Adrjuuia TToXioii)(co[i ^^

*6 Cf. 7 and 20. « CIA I. 468 ( = K 5).

" IGA 344 ( = K 181). « CIA I. 957 ( = Roberts 48),

" CIA IV. 373"8 p. 86 ( = H 240). Cf. IGA 466 (= H 286) and CIA I. 344 ( = H 216).

" IGA 408 (= H 300). 62 ciA IV. 373«^ p. 89.

14 PROCEEDINGS OF THE AMERICAN ACADEMY.

iKT]^6\(!> l(T)(eaiprj 5* IIoTjetSaFcor't Fiivukti ^^

From the very nature of the metre early elegiac inscriptions tend to be more diffuse than hexameters, but in them, too, the addition to the original bare formula may be only the name of the father or some word describing the gift (e. g. ayaXfia) or some phrase which had come to be a familiar part of the pentameter, e. g. fv$a^lel>os SeKar?;i/ (;-5'2, 36), iraibl xapi^ofxem) (15), iraiKi Aios fxeyaXov (31, 32, 36). Epp. 26 and 32 are two of the many concise distichs in which every word except the epithet gives the reader some definite information.

26. IlaXXjaSt fx iypefia)(a Aiofvcrto[? rdjS nynKpa arrjae KoXoiov nais [€v^a\ixfuos dtKaTrju.^^

In epigrams such as this the art appears in the adoption of a metrical form not, barring the epithet, in the choice of words or the method of expressing the sentiment. Thus 8 is a poem rough and without charm, written with more effort than success,

8. 2^|U.]a t68 (^f)y(y)vs oSoii Qeoarjuov [arrjaa- . . . dy^dpos ffiov (y)f (pi\ov Kayadoii ducplortpov.

On the other hand 5, though scarcely more ornate, shows a style some- what easier and freer.

5. T^jSe (f)i[\T]v aXo)(OP 6 df7va] KartdrjKe Bavovaav Afa/iTrtjTcu al8oiT]v, yrjs dno TvaTpooiqs.

Often we find the prayer, which was indeed always in the mind of a Greek, that the divinity may be graciously pleased to return an equiva- lent for the gift offered. ^^ Now it is good report that is desired

42. 56j Se P Iv di/dpooTTOis 86^av fXft" dyad(ci)u,^'

now gain

38. TTOTvia TOiv dyaBcov rat crv 86s dcpdoviav,

now we find expressed that craving to be remembered among men,58 to which, indeed, the very existence of the stones bears witness. Of a different sort are epp. 1, 2, and 11.

2. "Avdpuinf oi ((T^^T(l)(([i]s Ka6^ 68ov (fypacrw «X(X)a jjLfvoivoiv crTrjOi Ktil o'lKTipov aijpa Opacroivoi I86iv.

11. 2TJij.a 'PpaaiKXfUis. Kovprj KfKSrja-opai alfi

dvTi ■yd^ov irapa ^fcoi' tovto Xn^^oitr ovop.a.

" IGA 407 (= H 299). " IGA 20, 12 { = H 293).

" Cf. 10, 22, 40.

6* In a liile epigniin (AT C. 42) we read even 56s irXiov uu' <'\o/3es. »T Cf. 27, 43. " 24.

GRAGG. THE GREEK E;PIGRAM BEFORE 300 B.C. 15

It is not only that these verses are far more charming than any yet quoted, but we can see in them the beginnings of that {)rincii)le whicli characterizes the elegy in contrast to the epic. For the epic, with all its simplicity and directness of construction, depends for much of its effect on sonorous and splendid words, while the charm of the elegy is in familiar, even intimate sentiment never overshadowed by mere magnificence of vocabulary. Even in these poems, bare and brief as they are, it is the personal feeling of the writer that is expressed, and expressed with pathos all the more touching because of the simple means emplo3'ed. In hexameters, even those which express grief, the writer is telling a story, he is objective ; in the elegiacs he is subjective. Cf. the following poems.

Upa^iTfXfi Todf fivafxa F'ktccv iroiFrjm 6avo\vTi t\ovto d' eriiipoi (Tajxa x^'"' t^"[p]f'^ <TTfva)(ovTfi, Fepy(x>v avT ay\a\6uiv, KrjTrafKpov f^€Te\e(T((r)av.^^

nat[Sor dTrn](f)6ifxevoio K[Xfot]rou tov Mfveaaixfjiov fiv^fi' icropa>i> niKTip' ws kuXos a)u (6ave,^^

Passion as well as pathos is expressed in elegiacs and the author of 38 went so far as to threaten the enemies of the dedicator with human or divine anger, for the general sense is plain, though the last verse is mutilated.

ot T6 Xey[ov]crt \oyovs ab'iK\ovs\ '^ev8as Ka[T ] «(ic[€iVou.

In 8 the writer even comes forward in the first person.

These examples are enough to show that as early as the sixth century men entrusted to the stones their thoughts and griefs and desires. ®i Compressed and restrained though most of the epigrams are, there is in them the personal element, the lyric quality, which comes out more freely in the work of the fourth century and later.

The first traces of poetic color come less, perhaps, from deliberate art than from almost unconscious imitation. In 6 ou 6avaro[s 8aKpv]6fts Ka6[f]xei and in IGA 15 (= K 463a. add.), rov&Xfae ttoi/tos 0^4517^, the well-known epic vocabulary shows the absence of originality in the writer. The words or phrases were ready to his hand as familiar to his readers as to himself, and he is a poet because he chose to use them in his verses, not because he made them or used them in any

" lA 800. «o 7. Cf. 1 and 2.

^^ Cf. also CIA IV. p. 118 (= K 19), the only hexameter inscription which belongs to this class.

'Ec^dS' dvqp w/j.o(a)ev to. fidpKia traiSos €pa[<T]6[e'\ls

veiKea (n/y/xe[i]cr[7et«'] iroXefiSf 6' d/xa SaKpvofVTa, k. t. X.

16 PROCEEDINGS OF THE AMERICAN ACADEMY.

way peculiarly his own. On the other hand, 11, 17, 23, 43, 46 differ from the verses just quoted in that the writers have somehow managed to make their own the familiar expressions. 43 will illustrate

Ae^o Fdva^ KpoviSa Zev '0\vvTTie Kokov aya\fj.a IXtjFco 6vfi<ji) Tco AaKedaLfjiovia).

But in spite of the simplicity of these early epigrams, their variety is remarkable. The same ideas are expressed, the same words used, in a number of constructions, e. g. the name of the dead and of the divinity appear each in four cases,®^ that of the dedicator in three. Now it is the tomb or statue that speaks now the buried man or the dedicator ; now the god, now the passer-by is addressed.

We possess a few early epigrams which show greater poetic power, poems where art and elegance seemed to the authors as important as utility. An example is the well-known 25

Edfea Botcorwii kcu XaX/c/Secoi/ hnfiacravres TTal8(i ' Adrjvaioiv, ipyjiaaiv iv itoK(jj.ov,

roiv tTTTTOvs BeKUTTjv IlaXXtiSt rdcrS e'deaau

an epigram which approaches more nearly those of the next century,

because phrases like fdvea 8ap.d(TavTei, tpyfiaa-iv iv TToXffjiOV, ecr^faav v^piv

give to the whole poem a poetic coloring.

Appended to this paper are tables showing the elements which appear more or less constantly in the inscriptional epigrams. It is remarkable liow definitely they speak, how consistently they keep the reason for their existence before our eyes. In the sepulchral inscriptions we find always the name of the dead (but it is in the verses themselves, never extra metrum) ;*3 always some word meaning "tomb," except in the

«2 See Tables I, II.

*' Koehler thinks that in 10 the name of the dead was omitted in the verses and inscribed above them. His restoration, however, is by no means ceitaiu and it seems to me more reasonable to suppose it faulty than to accej)t on conjeeture a reading which would make the inscription an exception, not only in the sixth but in the fifth century, at least as far as we can tell fiom the eviilence at our command. Tlie only inscription that could possibly support Koehler's view is 12.

Tjv 'y'\ap a.irA(Tr)%

vovv re koL avol^pilfiv f^oxos rjXiKlas 'Eir]i(TT'^/i(ov t65' fTToet 'l7r(7r)o(r[rpdr]ou ffTJfia.

This is, however, not a parallel, for the name almost certainly ajipeared in the itiissiii<,' part of the hexameter and it njipears below, not as part of the epitaph, re- jieating extra mrtriun information already ^iven in verse, but as pait of a second and separate inscription with (piite a dill'erent function.

GRAGG. THE GREEK EPIGRAM BEFORE 300 B.C. 17

very early epigram 1, where, however, the language indicates plainly that the verses were inscribed on a tond). Almost always wo lind some word meaning " dead " (e. g. 6avui>, <pdifi€vos). In dcdicatijry inscrip- tions we may expect to find the name of the dedicator,** the name of a divinity, a verb of dedication. In 24 the last clement is lacking, but it must have been sufiiciently evident from the place where the stone was set up that it was a dedicatory oflering.

'AaTu>i> ^nX(X lofTfi)!', rro'Kir'jn^e ttotvi 'Addva, ^fjLiKpov Kill naidoiV fJ-vfifj. ('x'"' ^^^ TTuXis.

For these reasons, then, we are justified in refusing to assign to this early period any epigrams preserved in ]\ISS. only, which would require for the explanation or completion of their meaning any words on the stone e:rtra metrnm. In the case of dedicatory epigrams the informa- tion given in the verses may be supplemented hy inferences drawn from the places where the stones were set up. So we sometimes miss the verb of dedicating, as in 24. This is especially likely to be the case when the dedicatory offering takes the form of an honorary statue.*^ The epigrams of the fifth century show that the verb of dedicating was regularly omitted in inscriptions for such statues.

Of the epigrams preserved only in MSS. the great majority were intended to be inscribed. We observe in them the same stages of development as in the inscriptions, although in neither case does fuller development necessarily indicate later date.** 53 is as severely simple as any verse carved on stone.

Upa^ayopas riiSe 8a>pa Beo'n dvtd>]Ke AvKalov vlos fnoiijcrfv 8 epyov Avn^ay6piis.

19 expresses with greater elaboration, but with no greater charm, the same sentiment as 11. The very fact that of the epigrams attributed to illustrious poets some are as brief and severe as the inscriptions, while others are more elaborate, may serve as an indication at least that they are correctly attributed. 49, Archilochus's iinicojiore, nato dl due petall soli, could hardly be simpler.

** In 42 the words