of indifference if the word-foot happened sometimes to jar against the theoretical verse-foot at these more or less stressless points, as was particularly likely to be the case in Latin with its more conspicuous and immovable word-accent; so, for example, in Catullus, 63. 1:

super alta vectus Attis celeri rate maria

where the theoretical ictus on the last syllable of rate was too weak to conflict seriously with the word-accent on the first, and hence the frequent admission of such forms side by side with the more natural effect, as in v. 3,

where there is no such conflict of word-foot and verse-foot.

To sum up: The history of ionic rhythms from Alkman, Alkaios, Sappho, and Anakreon, down the ages through the Greek drama to Plautus himself shows a clear tendency to maintain the unmistakable purity of the ionic foot. On the other hand, the name itself of the galliambic, the typical association of the ditrochaeus with every phase of it from Kallimachos to the citation of Diogenes Laertius, the testimony of Hephaistion and the Roman metricians, and the remarkable fact that with all the freedom of resolution in the first foot of each dimeter, the only undebatable ionic resolution ZUU is confined to the pure ionic place, the end of the first dimeter, — all these considerations point to the truth of the thesis, that the resolution of the final long of the ionic foot in a galliambic connection amounts to rhythmic anaklasis for all known remains of this latest and most artistic creation of Greek metric.

24. Notes on the Bucolic Diaeresis, by Professor Samuel E. Bassett, of the University of Vermont.

This contribution, which was read by Dr. Weller, of Yale University, will be found in the TRANSACTIONS.

On motion of Professor Elmer Truesdell Merrill, of Trinity College, it was

Voted, that the Secretary send to Professor Francis A. March, Sr., the greetings of the Association as follows:

The American Philological Association, assembled in its annual meeting, sends affectionate greetings to its absent ex-President, Professor Francis A. March, whose four-score years of life, recently completed, have not been years of labor and sorrow, but of labor sweetened by manifold successful achievement.

Serus in caelum redeas, diuque

Laetus intersis populo!

25. The Ablative of Association, by Professor Charles E. Bennett, of Cornell University.

This paper will also be found in the TRANSACTIONS.

The Committee on the Place of Meeting in 1906 reported by its chairman, Professor Rolfe, of the University of Pennsylvania. It was recommended that the Association accept the kind invitation of the George Washington University, Washington, D.C.

The report of the Committee was accepted and adopted.

The question of a change in the mode of publication was raised and discussed in detail by Messrs. Scott, Radford, Harrington, Smyth, Merrill, Sanders, and Hempl.

Voted, that the matter of a change in the method of publication be referred back to the Executive Committee, to consider, and, after conference with the American Oriental Society, and the Modern Language Association, to report at the beginning of the next meeting.

Voted, to refer to the Executive Committee with power the suggestion of Professor Radford that the next volume of the TRANSACTIONS be so published that the publishers keep on hand a sufficient number of off-prints of the several articles to meet possible demands of purchasers.

26. The Classification of Latin Conditional Sentences, by Professor Karl P. Harrington, of Wesleyan University.

It cannot be denied that there is sufficient cause to search for a new classification of Latin conditional sentences. The present confusion in terms and methods of classification is bewildering and frequently results in misstatement, unpractical diffuseness, or meaningless conciseness. Professor Rolfe's paper before the New York Latin Club about a year ago led me to attempt a new classification which should include all the common types of conditions without relegating any to the hopeless limbo of fine-print exceptions, yet should state the facts as they appear, without reading into them any theories as to their origin or development.

It is not safe, for example, to lump all Indicative conditions together in such a way as this: "Indicative conditions. Conditions in any tense, with nothing implied as to their fulfilment and expressed positively (or vividly)." For frequently, on the contrary, the actual fulfilment of the Indicative condition is very definitely implied. So, when Cicero (in Cat. i. 16), addressing the arch-conspirator, and taunting him with the fact that none of his friends gave him the customary greetings when shortly before he entered the senate house, adds, Si hoc post hominum memoriam contigit nemini, vocis exspectas contumeliam, cum sis gravissimo iudicio taciturnitatis oppressus? Catiline himself and every auditor in the temple understood perfectly that the condition was stated as a well-recognized fact, and actually as the reason for the ironical question which follows it.

Further, while, if the tense of the Indicative be future, the conceivable case may be felt as stated more vividly than if the mood (referring to the same time) were Subjunctive, we cannot always, if we can ever, speak of a present or past Indicative condition as being especially "vivid." For example, when Pliny (Ep. vi. 20), in describing his own experiences during the great eruption of Vesuvius, quotes the exhortation of his uncle's Spanish friend, si frater tuus,

tuus avunculus vivit, vult esse vos salvos; si periit, superstites voluit, it is difficult to conceive of any possibility of putting the dilemma in any other form. It is neither a more nor a less vivid statement; it is absolutely colorless so far as any implication about the facts is concerned.

Again, suppose we examine the proposed category, "Indeterminate conditions: (a) Conditions in any time with nothing implied as to their fulfilment, expressed positively (vividly) in the Indicative." Now, in the first place, of course the examples quoted above can be fairly cited as a reasonable ground of objection to this classification as a whole. But, besides this, may it not be doubted whether it is consistent to speak of a condition with "nothing implied as to its fulfilment " as being expressed "positively"? How can we speak "positively " and yet convey no hint of the truth or falsity of our words? In referring to future time, to be sure, one may have a choice of moods, and thus express or imply a feeling on his own part of a greater or less degree of probability that the condition will be fulfilled. In cases, however, where present or past time is expressed in the assumption no such variation in the degree of probability can be expressed by any variation of mood.

Now it would be highly satisfactory if we could make such a classification as this:

I. Probable conditions: Indicative mood.

II. Possible conditions: Subjunctive mood, primary tenses.

III. Impossible conditions: Subjunctive mood, secondary tenses.

But without multiplying objections, it is sufficient to say that (1) while a large proportion of Indicative conditions do imply probability, from the standpoint of the speaker, or of the person addressed, or of the world in general, that is not always the case; (2) sometimes primary tenses of the Subjunctive are used to imply non-fulfilment of a condition; and (3) secondary tenses of the Subjunctive do not always imply a supposition contrary to fact.

It seems, therefore, wiser to make a modal classification, with such subdivisions according to general signification as are warranted by the facts. The proposed classification is according to protases, which are the rational basis of such classification, and no attempt is made to include any abnormal types, but to give due recognition to all the regularly occurring types, as follows:

I. INDICATIVE CONDITIONS.

(a) Suppositions implying actual fulfilment. Si hoc post hominum memoriam contigit nemini, vocis exspectas contumeliam, cum sis gravissimo iudicio taciturnitatis oppressus? Cic. in Cat. i. 16.

(b) Suppositions implying probable fulfilment. Si damnatus eris, atque adeo cum damnatus eris (nam dubitatio damnationis, illis recuperatoribus, quae potest esse?) virgis te ad necem caede necesse erit. Cic. in Verr. II. iii. 70.

(c) Suppositions implying possible fulfilment (in future time). Si patriam prodere conabitur pater, silebitne filius? Cic. de Off. iii. 90.

(d) Suppositions implying nothing as to fulfilment. Si frater tuus, tuus avunculus vivit, vult esse vos salvos; si periit, superstites voluit. Pliny, Ep. vi.

20, 10.

II. SUBJUNCTIVE CONDITIONS.

1. Primary Tenses. — (a) Suppositions implying actual or probable fulfilment (in general conditions). Nam vita humana prope uti ferrum est: si exerceas, conteritur; si non exerceas, tamen robigo interficit. Cato, de Mor.

(b) Suppositions implying possible fulfilment in future time. Si, inquis, deus te interroget, . . quid respondeas. Cic. Ac. ii. 80.

Eos non

(c) Suppositions implying non-fulfilment (comparatively rare). curare opinor, quid agat humanum genus: Nam si curent, bene bonis sit, male malis, quod nunc abest. Ennius, Tel.

2. Secondary Tenses. (a) Suppositions implying customary fulfilment (past general conditions). Accusatores si facultas incideret, poenis adficiebantur. Tac. Ann. vi. 30.

(b) Suppositions implying non-fulfilment. Nam nisi Ilias illa exstitisset idem tumulus, qui corpus eius contexerat, nomen etiam obruisset. Cic. pro Arch.

24.

27. Types of Sentence Structure in Latin Prose Writers, by Professor Clarence Linton Meader, of the University of Michigan. This paper is printed in the TRANSACTIONS.

Adjourned at 12.50 P.M.

FOURTH SESSION.

The Association met in the usual place at 3 P.M.

Professor Thomas Day Seymour, of Yale University, reported the following list of officers for the year 1905-1906, as proposed by the Nominating Committee:

President, Professor Elmer Truesdell Merrill, Trinity College.
Vice-Presidents, Professor Edward D. Perry, Columbia University.

Professor Francis W. Kelsey, University of Michigan.

Secretary and Treasurer, Professor Frank Gardner Moore, Dartmouth College.
Assistant Secretary, Professor William Kelly Prentice, Princeton University.
Executive Committee, The above-named officers, and

Professor Charles E. Bennett, Cornell University.
Professor Edward B. Clapp, University of California.
Professor Thomas Fitz-Hugh, University of Virginia.
Professor John C. Rolfe, University of Pennsylvania.
Professor Paul Shorey, University of Chicago.

The Nominating Committee also presented the name of Professor Herbert Weir Smyth for the vacancy in its membership created by the expiration of Professor William Gardner Hale's term.

The report of the Nominating Committee having been accepted and adopted, the President declared the officers elected.

The present membership of the Standing Committee to Nominate Officers is as follows:

To serve for one year, Professor T. D. Seymour, Chairman.

To serve for two years, Professor Samuel Hart.

To serve for three years, Professor M. W. Humphreys.

To serve for four years, Professor M. L. D’Ooge.

To serve for five years, Professor H. W. Smyth.

28. Gemination in Terence, by Professor Eva Johnston, of the University of Missouri.

The term gemination is adopted to denote the repetition of a word without change in form or meaning. Terence's use of gemination corresponds to that of other writers in that he oftenest doubles vocatives, imperatives, interjections. He furnishes six examples of the gemination of a vocative, four of them. used in address, two in calling to some one. All of them are found in passages where intense feeling is shown.

There are five examples of the gemination of an imperative; excitement is regularly back of this repetition.

Age age with interjectional force is found five times. Terence regularly puts the words into the mouth of some one who has decided to pursue a certain course of action against his better judgment. Five times we have heus heus. In three of the examples the word is repeated in calling to some one, in two when knocking at a door. Au au is used once and denotes distress on the part of the speaker.

In the gemination of vocatives, imperatives, interjections, Terence's use stands close to that of Plautus, but it is to be noted that Plautus occasionally trebles such words, while Terence never does.

In one or two cases the repetition of a word indicates doubt and uncertainty on the part of the speaker, and occasionally rhetorical effect is gained by such repetition, but in most cases gemination is found in passages in which deep emotion, such as joy, sorrow, or anger, is shown.

The paper was discussed by Professor Meader, of the University of Michigan.

On motion of Professor Merrill the following minute was adopted: The American Philological Association desires to express its grateful appreciation of the hospitality extended to it during its session now drawing to a close. Its warmest thanks are extended to ex-President White, to President Schurman, and the other authorities of Cornell University, and to individual representatives thereof, for their exceptionally generous and thoughtful kindness; to Professor Elmer for arrangements that have left nothing undone that could be devised for the comfort and convenience of his fellow-members; and to the Town and Gown Club for courtesies that have added much to the enjoyment of the meeting. The Association will remember the present session among the most delightful that it has held.