UC-NRLF 
 
 B M D3b ^^l 
 
A STUDY 
 
 OF 
 
 CONJUNCTIONAL TEMPORAL CLAUSES 
 IN THUKYDIDE8 
 
 A DISSERTATION 
 
 PRESENTED TO THE FACULTY OF BRYN MAWR COLLEGE 
 FOR THE DEGREE OF DOCTOR OF PHILOSOPHY 
 
 WINIFRED WARREN 
 
 BERLIN 
 
 PRINTED BY UNGER BROTHERS 
 1897 
 
A STUDY 
 
 OF 
 
 CONJUNCTIONAL TEMPORAL CLAUSES 
 IN THUKYDIDES 
 
 A DISSEKTATION 
 
 PEESENTED TO THE FACULTY OF BRYN MAWR COLLECxE 
 FOR THE DEGREE OF DOCTOR OF PHILOSOPHY 
 
 BY 
 
 WINIFRED WARREN 
 
 BERLIN 
 
 FEINTED BY UNGER BROTHERS 
 1897 
 
OP TTTF 
 
 UNIVERSITY 
 
 INTRODUCTION. 
 
 By a temporal clause two actions^) are Vrougiit into, 
 temporal relation. / . '« ^ ^ i i ' 
 
 An action may be regarded as extensive or as aoristic^). 
 
 The present stem shows an action extended, the aorist 
 an action concentrated. Certain aorist stems appear in verbal 
 forms of present time; but, though the possibility of the re- 
 presentation of present action under the aoristic conception is 
 not to be denied, it is doubtful how far aoristic force w^as felt 
 in these present forms derived from aorist stems. Further, 
 morphological connection between the aorist and the future, 
 and the coexistence of future forms from the aorist stem and 
 from the present stem, point to a distinction of aspect in the 
 field of the future. The question of aspect in the perfect 
 stem will be considered later. 
 
 The temporal relation of one action to another may be 
 that of antecedence, or of contemporaneity, or of subsequence. 
 
 The relations of antecedence and contemporaneity, and 
 of subsequence and contemporaneity may be combined in over- 
 
 1) Action is here used broadly of that which is affirmed, though 
 predication covers not only activity in a strict sense, hut also state. 
 
 2) The original force of the aorist is still a matter of controversy, 
 and the terminology of the question is not fixed. The word positioe is so 
 close a counterpart of the word extensive, since the two stand in the 
 geometrical relation of point and line, and since positive answers to the 
 function whereby the aorist simply posits an action, that it is to be 
 regretted that other uses have impaired the availability of the term. In 
 the present paper the word aoristic will bo employed to denote this 
 affirming value, graphically represented by a point. — For a historical 
 sketch of the distinction of aspect in action (Aktionsart), see Herbig Aklioiix- 
 art und Zeitstufe, I. F. VI (1896) p. 171 sgq., and the literature there cited. 
 
 1* 
 
 72m^ 8 
 
lapi)ing actions. For if one of two actions brought into 
 mutual rolation is prior to the other and is viewed as extensive, 
 it is evident that there is a possibility of ])artial coincidence. 
 In the sentence, Thuk. I 13 b^): tjiyihu'i re oVEkXriveg iJ,d)lov 
 enkcolov, Tag vavg xTi]odfj.n'oi to hjOTiy.bv xa&t'jQovv, the action 
 of tlie subordinate verb is still continuing when the main 
 action takes place. (But compare the aorist IJQ^ai'To in I 5 1 : 
 Of ydg "EkXifveg . . . ijieid)) iJQ^avro jukTaXov JisQaiovo&ai ravoiv 
 i"i' aX.iijXovg, tcQUTiovTo tiqoq X}]oreiav.^ So in the sentence YUl 
 45 1: iv de rovro) xai hi ngoregov, Jiolv eg t)]v 'Podov avrovg 
 dvaorfivai, rdde ejigdooero, the principal action is not ended at 
 the time of the subordinate action. Since in the relation of 
 antecedence the prior action is that of the dependent clause, 
 it is to be concluded that for the unambiguous expression of 
 pure antecedence, the action of the dependent clause must be 
 regarded as aoristic. Similarly, for the expression of the idea 
 of subsequence exclusively, an aoristic view of the main action 
 is to be expected. 
 
 Yet Hultsch'') in the treatment of the aorist finds no 
 ground of differentiation between subordinate clauses and 
 main clauses. He says: "So oft im Yorhergehenden vom 
 Aorist die Rede war, hat sich nirgends ein Anlass gefunden 
 den Gebrauch in Nebensatzen von dem in Hauptsatzen zu 
 unterscheiden, und in der That ist kein Grund denkbar, wes- 
 halb der Schriftsteller die Form des abschliessenden Berichtes 
 in Relativsatzen, nach temporalen und auderen Conjuuctionen 
 anders hatte verwenden sollen als in Hauptsatzen." Clearly, 
 the possibility that an action represented as extensive in a 
 temporal clause of antecedence may be antecedent in part 
 only, makes the aoristic conception necessary if there is to 
 be unequivocal expression of pure priority. Hultsch says 
 further: "Auch in Nebensatzen, die durch enei eingeleitet 
 werden, ist die durch den Aorist gegebene abgeschlossene 
 
 1) Citations from Thiikydides are made according to the chapters and 
 sections of Poppo-Stahl, books I-II, 3d edition, books III— VIII, '2d edition. 
 
 2) Die erzdhlenden Zeitforinen bei Polybios, Abhandl. d. k. sacks. Ge- 
 sellsch. d. Wissemch., ■phiL-hist. Classe XIII (1893) p. 454. 
 
Haudlung zuineist audi der des Hauptsatzes vorangegangen."^) 
 But if Euei is a conjunction of antecedence, aoristic expression 
 of the subordinate action necessitates the relation of complete 
 priority. 
 
 Sturm ^), after his study of jiqiv, formulates a general rule 
 in which he says: "Der Infinitiv steht in iiberwiegender An- 
 zahl nach affirmativem Yordersatze, um das Praedicat desselben 
 zeitlicli naher zu bestimmen; Tempus und Modus des letzteren 
 bleiben vollstandig unberiicksichtigt." But theoretically, it 
 is evident that if the prior action is regarded as extensive, 
 there is nothing to prevent its overlapping upon the later 
 action; and examination of the passages cited by Sturm bears 
 out the view that when nqiv with the infinitive denotes pure 
 subsequence, forms from the aorist stem rather than from the 
 present are to be expected in the main clause^). 
 
 In the case of contemporaneity, two relations are possible, 
 coincidence and insertion. 
 
 In the relation of coincidence, both actions must be 
 regarded either as extensive or as aoristic, but there is oppor- 
 tunity for choice between the two modes of conception. The 
 following sentences may serve to illustrate. 
 Hom. t 325-6: vvv 6)) jieq /hev cixovoov, ejiel jidgog ov nor 
 axovoag 
 gaiojuEvov, ote jx EggaiE xXvrog Ivvooiyatog. 
 Hom. A 07 1-2: d>g onor 'HXeioioi xal rjfuv vEixog hvyd^r] 
 djuqji /iorjlaoh], or Eycb xrdvov 'Irvjuovija. 
 In the former example, the actions of QaiojuEvov and Eggaie 
 are coincident, and the aspect is extensive; in the latter, the 
 
 1) C/. Miller A. J. P. XVI (1895) ;;. 179. 
 
 2) Geschichtliche Entwickelung der Constructionen mit JUPIN. Schanz's 
 Beitrage. Heft 3. 1882. p. 150. 
 
 3) The few exceptions admit of explanation; as, for example, Arist. 
 Ran. 885: sv^eods Stj xai 09 a> Tt, jtqiv to-t?; ?Jy£ir. Here there is a notion 
 of distributive action, for Aischylos and Euripides are called upon to pray 
 in turn. But not infrequently when jiqiv with the infinitive depends upon 
 a verb of the present stem, the idea is that of overlapping action (it may 
 be, with the conception of limit). Thus, Plat. Soph. 217 B: xal fisr 8)] 
 xaia Tv^rjv ys, d> ^wxQarsg, loycov ijTs?.a^ov 7iaQa:T?^7]aio)v cbv y.al jiqiv >//<d? 
 8evq eldsiv disQOOToJvxe? avrov izvyxuvo/iiev. Hdt. VI 119: (^aadsvg 8e AaQsTog, 
 
actions of t-rvyj^rj and xrdvov are coincident, and the aspect 
 is aoristic. In the rehvtion of insertion, one action must be 
 viewed as extensive, the other may be viewed as extensive 
 or as aoristic. 
 
 Hom. /? 171-3: xal ydo y.eivqy cptjjul rekevrr]'&i]vai obiavra 
 o3s oi i.fxv&eoixi'jv, otf "Ihov e\oavE(icuvov 
 WQyeToi. 
 
 Hom. Ilbdl -S: Tov juev aga FXavxog orfj'&og jueoov ovraoe Sovqi, 
 OTQEcp&eig i^amrtjg, ore juiv xarejuaQJiTs dimxtov. 
 
 Four forms are thus possible: 
 
 Coincidence — (1) extensive aspect, 
 
 (2) aoristic aspect, 
 Insertion — (3) extensive aspect within extensive aspect, 
 
 (4) aoristic aspect within extensive aspect. 
 
 It follows that, as regards aspect, the forms of the con 
 ception of two actions in temporal relation are these: 
 
 Main verb Subordinate verb 
 
 Antecedent action 
 
 Overlapping antecedent action 
 
 Contemporaneous action 
 Coincidence 
 
 Extensive aspect 
 Aoristic aspect 
 Insertion 
 
 Extensive aspect within 
 
 extensive aspect 
 
 Aoristic aspect within 
 
 extensive aspect 
 
 I extensive 
 \ aoristic 
 extensive 
 aoristic 
 
 extensive 
 aoristic 
 
 extensive 
 
 aoristic 
 extensive 
 
 [[extensive] ^1 
 
 \ aoristic 
 extensive 
 
 extensive 
 aoristic 
 
 extensive 
 
 extensive 
 aoristic 
 
 [ 
 
 JiQiv (.lev alyjiaXcoTOvg yevio&ai Tovg 'EgezQisag, ivei/s aqn deivov yolov, oTa 
 ag^avTcov adixhjg 7tqozsq(ov twv 'Eostoiewv' ejieite di side atpsag d.-rax&srTag 
 :raQ ewvzoy xai ecovzai vjioyeiQiovg iorzag, i:;ioit]as xaxov a.)J.o oi'dsv, 
 1) Cf. p. 24. 
 
Main verb Subordinate verb 
 
 I6XtPTlSlVG I 
 aoristic I 
 
 Subsequent action { 
 
 aoristic 
 
 I [extensive] M j extensive 
 
 aoristic 
 
 The temporal relation of two actions may be that of 
 delimitation of one by the other. The one action may be 
 coextensive with the other and mark its limits; the sentence 
 is then an expression of the relation of coincidence. For 
 example, 
 
 Hom. jLi 327 - 8 : oT d' ei'cog juev oaov tyov xal olvov EQvd^Qov, 
 r6(pQa ^oajv ankyovTO Xilaiofxevot (iioxoio. 
 
 Or one action may define the bounds within which the other 
 action falls; this is an expression of insertion. Thus, in the 
 passage, 
 
 Hom. ^ 15-6: elog 6 rav&' ajg/xaive xaxa (pqeva xal xatd ^vjuov, 
 TOcpQa ol iyyv&ev fjk&ev dyavov Neorooog vlog, 
 
 the leading action is represented as occurring within the 
 course of the subordinate action. Or one action may define 
 the limit of the progress of the other. In this case the action 
 to be defined, viewed as extensive, must in part precede the 
 defining action. If the defining action forms the limit, it is 
 viewed as aoristic; if it embraces the limit, it is viewed as 
 extensive. The following sentences exemplify the two forms. 
 Hom. £ 428-9: djLi(poT8Qf]ai dh y^Qolv eneoovfievog M^e nejQrjg, 
 
 rfjg e'xsTo orevdj^cov, ei'cog jusya xvjua TcagrjX'&e. 
 Soph. ^w^. 415-7: yoovov rdd' rjv rooovxov, eg r ev di&eQi 
 
 jueocp xaxeorr] ^.ajuJiQog r)Xiov xvxXog 
 
 xal xavil e&a}\jie. 
 
 This relation of partial coincidence of a defining action 
 with a prior action that it defines is found only in combined 
 subsequence and contemporaneity. Hence it is to this field 
 that the expression of limit in one direction belongs^). 
 
 1) Cf. p. 36. 
 
 2) H. Lattmann {Gebrauch der Tempora im Latetmschen. Gottingen 
 1890. p. 90), classing these clauses of limit under contemporaneity, says: 
 
8 
 
 In the application of tlio goiieral princ'ij)los that govern 
 the use of stems in connection with temporal clauses, certain 
 influences are to be taken into account. Tlie system of 
 Greek verbal forms is far from perfect, and a full set of 
 stems is not made from every root. Again, some verbs by 
 virtue of their meaning have a tendency in one direction or 
 the other ^). Further, negation combined with the two modes 
 of conception has a distinct force with each. A negation 
 with the aoristic conception denies the action simply and 
 sharply. The denial of an action in its extent points toward 
 lack of power or will^). Hence it is evident that in temporal 
 correlation of actions in a negative sentence the aoristic view 
 may replace the extensive view of the affirmative; but the 
 extensive can never be substituted for the aoristic. 
 
 The temporal relation that exists between an independent 
 verb and the verb of a temporal clause is denoted by a 
 conjunction^). This statement, assuming distinct lines of 
 demarcation, is subject to certain limitations. It is possible 
 with broad stroke to synchronize actions that a finer touch 
 would distinguish as in sequent relation, and hence there 
 arises a loose employment of conjunctions of contemporaneity. 
 
 "Satze mit donee, quoad auch duin enthalten aber auch Handlungen, mit 
 deren Eiutritt die Handlung des Hauptsatzes ihr Ende erreicht. Im 
 Deutschen leiten wir solche Satze mit 'so lange bis, bis dahin dass' ein. 
 Hier die Beziehung der Nachzeitigkeit auszudriicken, entspriiche nicht 
 dem natiirlichen Verlialtnisse. Denn dies wiirde den Sinn haben: nach 
 Abschluss der Handlung des Hauptsatzes trat die des Nebensatzes ein. 
 Im Deutschen sind wir vielmehr geneigt, die Handlung des Nebensatzes 
 als vorzeitig aufzufassen, z. B.: 'Es regnete bis ich nach Hause zuriick- 
 gekehrt war', d. h. erst nach der Riickkehr horte der Regen auf." The 
 example given is simply a case in which the limit falls within an action 
 of which the beginning is antecedent to a part of the limited action. 
 
 1) Cf. Hultsch op. fit. pp. 2G, 69 el al. Koch Ncue Jahrbucher II 
 (1892) p. 439 sy. 
 
 2) Cf. Gildersleeve A. J. P. II (1881) p. 466 aq. 
 
 3) A full consideration of participles as expressing temporal relation 
 cannot be attempted here, and the discussion is therefore restricted to 
 "conjunctional temporal clauses". Only now and then reference is made 
 to participial construction. 
 
9 
 
 Further, certain conjunctions of limit may be used either of 
 contemporaneous action or of subsequent action overlapping. 
 With action of the latter kind, again, conjunctions of sub- 
 sequence may under certain circumstances be used to express 
 limit. 
 
 Temporal clauses fall into the two classes — prior and 
 posterioi'^) It is clear that clauses of antecedence and of 
 combined antecedence and contemporaneity are prior, while 
 those that denote either combined subsequence and con- 
 temporaneity or subsequence are posterior. As regards clauses 
 of contemporaneity, examination shows that when the two 
 actions stand in the relation of coincidence, the clause is 
 prior ^); when the relation is that of insertion, the clause is 
 prior if the main action is inserted, posterior if the action 
 of the subordinate member of the sentence is inserted^). In 
 accordance with their nature, prior temporal clauses normally 
 precede the main clause, posterior follow. 
 
 There are three spheres of time, the past, the present, 
 and the future. Sphere is, so far as the indicative mood is 
 concerned, usually shown by tense; but apart from a few 
 special usages (as, for example, the historical present), the 
 distinction of imperfect, aorist, and pluperfect, of present and 
 perfect, of future and future perfect, is not such as to aifect 
 the question of S])here. Desire to represent action in its 
 relation to the present leads occasionally to a free combination 
 of spheres. For example, an action in the past sphere may 
 be temporally defined by reference to an action antecedent 
 to it, and yet this antecedent action, viewed as extensive 
 and as holding true for the present, be represented by a form 
 of present time 
 
 The origin and significance of the moods is a vexed 
 
 1) For the distinction of prior and posterior clauses see Delbriick 
 Sijnt. Forsc/i. I. Halle 1871. p. 35. 
 
 2) Cf. Delbriick op. cit. p. 67 sqq. 
 
 3) The posterior clause of contemporaneity seems to be the real basis 
 of the so-called inverse construction. The sense of inversion arises from 
 the fact that an action regarded as extensive cannot properly be temporally 
 delimited by an action regarded as aoristic. 
 
10 
 
 question, and its discussion is boyond the scope of the present 
 paper. One or two mutters, however, appropriate to the 
 consideration of temporal clauses, may be noticed. 
 
 Wlien the repetition of two actions in temporal relation 
 is to be affirmed, the form of ])redication, if there is to be 
 no ambiguity, must distiuguisli between the correlation of two 
 iterative series as wholes and the correlation of individual 
 acts repeated in like temporal relation. Distribution, in the 
 latter case, may be effected through words other than the 
 verb, that is, through an indefinite relative conjunction, a 
 distributive pronoun, and the like. The indicative mood 
 may then be used in the temporal clause. Generally, however, 
 the subordinate verb appears in the subjunctive or the 
 optative, according to the sphere of time^). The ground of 
 this seems to lie in the function of these moods whereby 
 they denote ideality. A single occurrence of one action is 
 predicated as ideal, to serve, in proper relation, for each 
 occurrence of the other action. 
 
 It has for some time been recognized that the future 
 indicative has modal force. The future indicative in temporal 
 clauses is rare; for, as has been pointed out^), greater exactness 
 is gained by the use of forms that reveal the mode of con- 
 ception of the action (as extensive or aoristic), and in future 
 forms this distinction is not wholly clear. 
 
 1) It is not necessary, with Flag^ {Outline of the Temporal and Modal 
 Principles of Attic Prose. Univ. of California 1893. pp. ix, 75), to make 
 a distinct class of clauses introduced by words meaning 'until'. Such a 
 clause with subjunctive or optative frequently, it is true, implies finality, 
 but not universally. Without limitation by means of other words to 
 effect the distribution, the Greeks were not at liberty to use the indicative 
 in such designation of general action. 
 
 2) Gildersleeve {Trans. A. P. A. 1876 p.Q) applies to temporal clauses 
 the same reasoning that he employs in regard to conditions: "In future 
 relations, si with the future indicative may be dissected into ear with the 
 present subjunctive and Mv with the aorist subjunctive; and hence, whenever 
 it* is important to distinguish continued from concentrated action, whenever 
 it is important to distinguish overlapping from priority, Mv with the sub- 
 junctive is preferred, not only in general sentences but in particular 
 sentences." (See, further, Clapp Trans. A. P. A. XXII (1891) p. 86 sqq. 
 Gildersleeve Johns Hopkins Univ. Circulars XI (June 1892) p. 102 sqg. 
 A. J. P. XIII (1892) p. 502.) 
 
11 
 
 From the foregoing discussion it is apparent that a study 
 of conjunctional temporal clauses must be based upon a con- 
 sideration of moods, tenses, stem-systems, conjunctions, and 
 the relative position of main and subordinate clauses. 
 Accordingly, the conjunctional temporal clauses of Thuky- 
 dides will be considered under these heads. 
 
CHAPTER L 
 
 § 1. MOOD. 
 
 A. Past and present spheres. 
 
 In the past and ])resent s\i\ieves,' particular action in accord 
 ivith fact is regularly expressed by the indicative in both 
 clauses. 
 
 The indicative of the main clause is not infrequently re- 
 presented by infinitive or participle. The place of the in- 
 dependent verb is taken by a verbal noun in VIII 87 4: 
 ijuol tLihtoi doy.d oaqreorrxTOv fb'cu dtaxQi^fjg evexa xcA di'oxcoyfjg 
 rcbv ' EA/a]V(xojv to ravTtxdv ovx ayayeh', (p'&ogag juev , Iv ooco 
 Tiagijei Ixeioe xal SifjueAAev, dviowoeojg de, onoog f^ajdsTSQovg 
 Tiooo&e^uevog loyvQOJEOovg 7iou)orj. Here (p&OQdg involves a verb 
 expressing action that is coextensive with that of TragfjEi and 
 diEfielAEv. Occasionally there is ellipsis of the independent 
 verb. This is regular in the case of comparisons where a 
 temporal clause is employed. The occurrences are as follows: 
 ore (with Moneq) IV 34 1. V 16 3. 41 2; (with //') YI 
 31 1; orav (with if) HI 5rt 7; ttq'lv (with h rco Sfioko xai) 
 VI 11 1. An interesting structure is that in which a temporal 
 clause is used with an expression of the lapse of time ck 
 with an ordinal; for example, I 18 1: It)] ydg eon fidXioxa 
 revQaxooia xai o/uyqj TiXeio} eg rijv TeAevzijv rovde tov noXefxov , dqf 
 ov Aaxedrujuovioi rfj avrfj noXixeia yQCJovrai, and HI 68 5: xal rd 
 fiev xaxd nkdraiav erei rgircp xal evevrjxoorcp ejieid)] 'A&rjvaioiv 
 ^viJLfxayoi eyevovxo omcog eteXevT7]oev. There is a similar con- 
 struction of a TiQiv clause in III 64 3: rtjv relevxaiav xe tiqiv 
 jieQixecyiCeoOat nQoxlrjoiv. The force of ngiv here is not far 
 
13 
 
 from prepositional, as may be seen by comparison with I 14 
 2: ravra ydg reXevjala jtqo rfjg Seq^ov OTQaTsiag vavrixa a^ioXoya 
 iv rfj 'EXXddi xazeoji]. Another noteworthy usage is that by 
 which the temporal clause does not define the time of the 
 main action, but gives the content of an act of memory. 
 Thukydides has two examples. EL 21 1: ixtixvYnxhoi xal 
 nXeioTodvaxra rov Uavoavlov Aaxedaijuovicov ^aoiXm, ore io- 
 ^aXwv Tfjg 'AxTixfjg eg 'EXevolva . . . dve^^MQtjoe ndXiv. Prolepsis 
 of the subject of the subordinate clause with this construction 
 is striking. II 54 4: jiiv)]jH7] da iyevexo xal rov Aaxedaijuovicov 
 XQyjoirjQtov roTg eiSooiv, ore IjzeQCorcbaiv avroTg rov debv el '^gi] 
 TioXejiieTv dveiXe xard xodrog noXefxovoi vixrjv eoeo&at. Here the 
 clause introduced by ore is practically an explanatory ap- 
 positive of rov XQV^''^VQ^^^- Compare, further, III 59 2. 
 
 As regards the form of the verb in the subordinate clause, 
 there are certain limitations imposed by the apparatus of the 
 language, notably in connection with the conjunction jiqIv. In 
 the expression of simple subsequence jiqIv governs the in- 
 finitive. So the infinitive follows Jigoxegov ij, and once in 
 Thukydides (YI 4 2), on the same principle, voregov fj. It 
 is noteworthy that this clause with voregov fj immediately 
 follows jiQiv with an infinitive. In the circumstances now 
 under consideration, there is but one other instance of a 
 temporal clause that does not offer the indicative. In the 
 case of this exception, the verb upon which the temporal 
 clause depends is an infinitive, and the subordinate verb, even 
 though it precedes, is assimilated, 11 102 5: Xeyerai de xal 
 AXxjuecovi rep AjucpidQeco, ore di] dXdo^ai avrov jxerd rov q)6vov 
 rfjg fxrirgog, rov AjioXXo) ravrrjv rijv yf]v iQijoai oixelv. Although 
 among temporal clauses this stands without a parallel, there 
 are several instances of the infinitive in clauses introduced 
 by other relative forms ^). 
 
 "Eoriv ore in Thukydides is probably completely crystallized 
 as an adverb. In three (I 25 4. Y 102. YII 21 3) of 
 the eight occurrences, it is impossible to refer any finite verb 
 
 1) See commentators on I 91 5. 
 
14 
 
 to uTf, and the formula must be an adverbial modifier of a 
 participle. In three other passages (III 43 5^). 45 6"). 
 \I 38 3) the order of words militates against the consideration 
 of eoT(v ozE otherwise than as phraseological. In II 81 3: 
 dier/^ov dk ttoXv im dLK)J]X(ov xal eoriv me ovde ecooonTo, although 
 a finite verb follows the expression immediately, it seems im- 
 possible to accept such an explanation as Hofmann's: "Et hie 
 et in similibus locis non facile est ad dijudicandum, utrum 
 Eoriv ore prorsus ])articulae vim habeat necne. Yidetur autem 
 hie usus praesentis soji sequente praeterito inde ortus esse, 
 quod qui rem factam narrat, tanquam praesentem earn menti 
 suae pro])onere potest, I'oti verbo utens. turn vero verbum, 
 quo res ipsa declaratur, quia praeterita est, praeterito tempore 
 eloquitur." The only remaining passage is II 39 1: j))v te 
 yoLQ TTohv y.oiv)]v jiaoE/ojuev, xal ovx eon}' ore ^evrjXaoiaig 
 ajTEigyojue}' Tiva // /la&jjjuaTog 7] ^ed/iarog. This shows nothing 
 to compel the interpretation of eonv ote as an adverb; but in 
 view of the balanced structure of the speech, it is hardly to 
 be doubted that te and xai are intended to correlate jiagexojUEv 
 and (hrEioyo^uEv. ovx eoriv ote is equivalent to ovdejxoTe. 
 
 Iterative or generic action in the leading clause Thuky- 
 dides regularly expresses by the indicative, in the temporal 
 clause by the subjunctive for the present sphere, the optative 
 for the past. Two passages claim attention. 
 
 II 35 2: ioIetiov yaQ to juergkog eItieTv ev m /.lohg xal fj' 
 
 1) Kriiger makes eotiv ote modify CtjfuovTf, Classen the whole sentence. 
 But it is better to join it •with the participle expressing the condition 
 upon which the general statement of the apodosis, 'QrjfuovxE, depends. So 
 Hofmann {De verhorum Graec. quae in formulas ahierunt vsu et natura, 
 Berlin I860, p. bb.) Possibly Kriiger's suggestion of eotiv o ri is to be 
 accepted, along with the emendation of Abresch for the reading of the 
 manuscripts in II 94 3. 
 
 2) Hofmann holds that if a finite verb, whether past or present, follows, 
 Eoxiv oxE is not to be regarded as phraseological. The passages on which 
 he bases this conclusion do not bear inspection. Thus, in Hdt. II 120 
 the coiTelation by i-iiv and 8e and the position of avxov IlQidfiov render 
 the adverbial sense necessary: as in Thuk. I 65 3 the construction with 
 jTo/.ta/taxa, and II 13 7 the trajection of y.ai show that the combination of 
 the substantive verb with the relative has nominal force; and III 24 2 is 
 poor proof. 
 
15 
 
 d6x7]oig rrjg dkr]i)eiag fie/^aiovrar o re yaq ^vveidojg xal evvovg 
 dxQoaTi]g Tuy' civ ri IvdeeoreQcog nQog a (iovXerai tt xal imoiarai 
 vojuloeie dr]lovo&ai, 6 re uneiQog eoriv a xal nXeovd'Qeod^ai^ did 
 (f&ovov, el' ri vjikg t)]v lavrov (pvoiv dxovoi. Here ev co is 
 regarded either as local -temporal, 'under circumstances in 
 which' ^), or as meaning ev co /nerQicog emeTv'^). If h' co is 
 circumstantial, the explanatory sentence beginning o re yc'xQ 
 should give the ground of the apodosis, whereas this sentence 
 evidently explains the ev co clause. Even Herbst, though he 
 tries to fuse the two clauses in such a way as to make the 
 causal sentence explain yaleixov, says, when off his guard, 
 that "die rede keine tiberzeugung hervorbringt, weil die 
 einzelnen zuhorer ilire eigene, teils giinstigere teils geringere 
 meinung von den thaten haben und behalten". It cannot be 
 supposed that the thought of the writer was of the definite 
 circumstances under which the orator was speaking, for some 
 definite limitation would be necessary, as in VII 51 1: dX)^ 
 avrov cog rdytoTa xal ev co ocpioi ^vjucpegei dvayxdoai avrovg 
 vavjLiayeiv. The relative, then, may be referred to to juergkog 
 einelv, and the passage excluded from a collection of temporal 
 clauses. 
 
 Y 103 1 : rolg 6' eg djiav to vjidgyov dvaQQinrovoi (ddnavog 
 ydg cpvoei) d/urj. re yiyvcaoxerai ocpaXevroov xal ev orco en cpvXd^erai 
 rig avrijv yvcoQioO^eioav ovx eXXe'mei. Few now hold the opinion 
 that ev oTcp here means 'while' ^). Scholefield translates: 
 "neque destituit, quamdiu ab ea cognita cavere poterit aliquis''. 
 But if ev orcp is equivalent to quamdiu, since the speaker 
 has just said that the nature of hope is not known until 
 failure comes, yvooQiodeloav must be an unfulfilled protasis of 
 which (fvXd^erai is the apodosis, and this is clearly impossible. 
 eXXeiJiei is to be regarded as transitive, governing the antecedent 
 of orco. 
 
 1) Krahner Pbilol. X (1855) p. 441, Kriiger, Herbst PIdlol. XXIV 
 (1866) p. 713, Classen, Kraz Die drei Rtden des Ferikles bei Thuk. Nord- 
 lingen 1880, Stahl, Stcup, Marchant (though he translates otherwise), et al. 
 
 '2) Poppo, Boehme-Widmann. 
 
 3) So Scholefield, Poppo, Jowett (with an alternative translation). 
 See also Graves. 
 
16 
 
 A gt'iicric t(Mn|»oriil clause cannot be subordinate to a 
 main verb of definite action. There is no temporal correlation 
 in such a sentence as I 41 3: ir xaiools roiovroig iyevero olg 
 {.idhcnn avdQOinoi eJi eyOgohg rubg Offeregovg lotTeg r(bv ndvjcov 
 d:Teo('o:TToi dot :jand to vixnv, and the dependent clause is not 
 to be classed as temporal in passages like YII 47 2: rrjg 
 xe <x)Qag tov h'lavxov ovoi]g h> [/ do&evovoty dvdocojioi jud?uoTa. 
 But iu III 56 5: dixaiov fjf.icbv Tijg vvv djuaQTiag, el ciga 
 f]fidgT)]Tni^ diTi&eTvat t)]v tore Txgo&vjxiav^ xal jueiCoj re ngog 
 t)Aooco evg>)oej£ y.ni h' yjugoTg olg ojidviov )]v rcbv 'EkXrjVWV iiva 
 dgerip' jfj Zm^ov dvvdjiiei dvrird^ao&ai, the subordinate action 
 is not generic, but particular. 
 
 Of action contrary to fact in connection with temporal 
 clauses there are two instances. One occurs in a chapter 
 probably s])urious, III 84 2: ov ydg dv tov ts ooiov ro 
 TijLicogeio&ai jigovTi&eoav tov re jui] ddixelv ro y.egdaiveiv, iv d) /ui] 
 ^Xdjixovoav loyvv elx£ to <p&oveiv. The relative clause with its 
 verb in a past tense of the indicative contains the unfulfilled 
 condition. In the other passage, the action of the verb of 
 the temporal clause is not itself contrary to fact. The verb 
 is an infinitive introduced by Jigi'r. VII 28 3: eg qpiXovixiav 
 Ha'&eoTaoav ToiauT}]v )}v^) jrglv yeveo&ai )]JTioT)]oer uv rig dxovoag. 
 
 B. Future sphere. 
 
 In the future sphere, the main verb assumes, according 
 to thought and context, any of the forms used of future time, 
 namely, future indicative (e. g., II 72 3), subjunctive (I 91 3), 
 optative (IV 111 1), imperative (V 47 7), or substitutes for 
 these. 
 
 The subordinate verb, unless it is under the influence of 
 a past indicative, is regularly subjunctive. The future 
 indicative is found but twice. I 39 3: ovg XQ*~P'> ^t£ docpake- 
 oxaxoL Tjoav, xoxe ngooievai, xal jui] ev co fjixelg juev fjdixiiued^a, 
 ovxoi de xivdvvevovoi, ju7]d' ev co vjueig xfjg re dvvdjuecog avxcov 
 rore ov i^iexaXa^ovreg rTjg dxpeXiag vvv jueTadcoaere, xal rcbv 
 
 1) On the subordination through the relative compare Sturm op. cit. 
 p. 115. 
 
17 
 
 djuaQTrjfxdzcov dnoyevojuevoi rfjg acp fjjucov ahiag to loov e^ere. 
 VI 86 5: f]v el rep vjiotito) r/ ajiQaxrov idoete dneXd^eiv fj xai 
 ocpaXeToav, en ^ovh]oeo&e xal noXXoorov fxoQiov avrrjg IdsTv, ore 
 ovdkv m jiegavet jiagayevojuevov v/luv. In the former passage, 
 jueradcooere is clearly modal, and vvv hints the present sphere. 
 In the latter, the future in the temporal clause is the apodosis 
 of the conditional participle naQayevo/xevov, equivalent to edv 
 naoayevrjrai. There is no instance of the future indicative 
 in a clause of antecedence or subsequence, a position in 
 which distinction of stems is of greater moment than in 
 clauses of contemporaneity. Once (III 39 8) the present 
 indicative del is used with future force; and in YIII 109 2, 
 nX}]Qovrai is a virtual future, but it is quite improbable that 
 this passage is from the hand of Thukydides. Even when 
 the governing verb is a past indicative or is under the 
 influence of a past indicative, Thukydides, following his 
 inclination to give repraesentatio wide range, generally uses 
 the subjunctive in the subordinate clause. The optative 
 appears in the following cases: — 
 
 a) The temporal clause depends upon an optative in a 
 final clause (four times : eneidiq — YII 80 5 ; onore — IV 111 1 . 
 VI 97 5; TTQiv — m 22 8). 
 
 b) The temporal clause depends upon a subjunctive in 
 a final clause, which, in turn, depends upon a secondary tense 
 of the indicative (once: onore — I 91 3; the optative follows 
 the indicative and precedes the subjunctive). 
 
 c) The temporal clause depends, directly or indirectly, 
 upon an infinitive with dv (three times: Icog — III 95 1; 
 ^nore — VII 48 1; nqiv — IV 117 1). 
 
 d) The temporal clause depends upon a past future 
 expressed by e/ts^/loi' with the infinitive (three times: onore — 
 in 22 3. IV 77 1. VI 66 1). 
 
 e) The temporal clause depends upon an indicative 
 (actual or virtual) of past time, but expressing action so 
 directed toward the future that there is partial obliquity 
 (twice: emz — HI 102 7. V 35 4). 
 
 A single example of the aorist indicative in a clause of 
 
18 
 
 the future sphere is found. V 18 5: ojrAa dk jui] i^eorco 
 fmq^iQeiv'Ai))]vaiovi; jurjde rovg ivju/udxovg im y.axM, djTodidovrayv 
 rov (fOQOv, hieidr} at ojiovdai iyevovro. The treaty is written 
 in consistent regard of the sphere as future with the exception 
 of the expressions ooag de noXeii; jiaQedooav and inetdi] at 
 ojtovdal Fyh'Oi'To. In defense of these, editors can but refer 
 from one to the other, for Stahl's citation of VI 10 5 is irre- 
 levant. Yet it is generally acknowledged that the sentence 
 opening with ooag de noleig bears evidence of corruption, 
 and a change of reading or a lacuna befere these words is 
 accepted. The temporal clause is referred by Betant, Jowett, 
 and Classen to emcpEQeiv, but by Stahl, Steup^), and others, 
 with varying interpretations, to dnodidovrcov. Steup argues 
 that if there had been no express statement of the time at 
 which the payment of tribute should begin, it would have 
 been possible for the Athenians to claim arrears. But though 
 BTzeidi] al onovdai iyevovro might be a comment on d7ioSLd6in:(jov, 
 it would be natural to take for granted that all provisions 
 of the treaty should go into effect at the same time, and to 
 make no specification on this one point. Surely the passage 
 is not certain enough to compel the acceptance of the unique 
 aorist indicative of antecedent action in the future sphere. 
 
 Use of dv. 
 The modal particle dv in temporal clauses Thukydides 
 uses for the most part in accordance with the principles of 
 standard Attic prose. Certain exceptions, however, appear in 
 the manuscripts, and the question arises whether these are 
 genuine cases of omission or are to be attributed to the 
 carelessness of scribes. The instances are as follows: — 
 With juexQi — I 137 2: jurjSeva ex^iivai kx rrjg ved)g juexQi 
 TiXovg yevrjrai. With juexQi ov — III 28 2: Ildyrig 5' 
 dvaoxrioag avrovg wore juij ddixfjoai, xararid^erai eg Tevedov juexQi 
 ov ToTg 'A'&'i]vaioig ri do^rj. IV 16 2: eoneTodai de avrdg juexQi 
 ov enaveX'&ojoiv oi ex rcbv 'Ad^r^vwv Aaxedaijuovicov Jigeo^eig. 
 41 1 : e^ovXevoav deofioTg juev avrovg (pvXdooeiv juexQi ov ti 
 
 1) Tlmkydideische Studien. Freiburg i. B. 1881. I p. 37. 
 
19 
 
 ivju^Moiv. 46 3: xal avzovg i.g ri]v vfjoov oi oxQazrjyol t)]v 
 Tlrvxiav eg (pv?.axrjv Siexojuioav vnoonovdovg , i^diQi ov 'A^^vaCe 
 7iejuq)^a)oiv. With ov — IV 17 2: bjiixcoqiov ov •^/uIv ov /uev 
 ^ga^eTg dgxcboi jui] noXXoTg ygijo&at, nXsioai Ss ev co av xaiqbg fj 
 diddoxovTag ri rc7)v Trgovgyov Xoyoig to deov TiQaooeiv. With 
 jTQiv — VI 10 5: iQi] . . . fu) . . . oiQifjg dXXrjg OQeyeod'ai tiqIv 
 i)v exofisi' l^e/iauooojjiie&a. 29 2: xal on oaxpQOVSorsQov eirj /ut] 
 jiierd ToiavTrjg ahiag, tiqIv diayvojoi, jiejujieiv avrov em tooovtu) 
 oTQarevaari. 38 2: r/^ieZ? de xaxol, jiqIv ev rep na'&eXv wjuev, 
 TiQocpvXd^ao'&ai. VIII 9 1: ov jTQov^vjuijdrjaav ^vjujiXeTv, jiqIv 
 rd "lo&juia, d Tore fjv, dieoQrdocooiv. 3: to re JiXfj&og ov f^ovXo- 
 juevoi JTCo jToXefuov e'xeiv, tiq'lv ri xal loyvgov Xdficooi. With 
 71 Q ore gov }j — VII 63 1: jui] jrgoreQOv d^iovv dnoXveo&ai I] rovg 
 djio rov TioXefxiov xaraoTQajjuarog ojiXirag djiaQa^rjre. 
 
 The absence of dv with tiqiv is discussed by Sturm ^), 
 who ascribes the failure of the particle to the negligence of 
 the copyist, and points out that palaeographically the errors 
 are not improbable. But it is better to look at the matter 
 as a whole and in its historical relations. A few conditional 
 and relative clauses that show the subjunctive without dv 
 take their place beside the temporal clauses^). Stahl^) defends 
 the omission of dv in Thukydides on the ground of its omission 
 in the tragedians and of the affinity of the historian's style 
 with theirs. Diener*), after a comparative study of prose, 
 accepts the omission. The evidence thus brought together 
 is weighty. Certainly, unless there is stronger argument for 
 dv than has yet been put forward, the ratio ^) between the 
 
 1) Op. cit. p. 94 sq. 
 
 2) See Lange De coniunctivi et optativi usu Thucydideo. Cassel 1886. 
 p. 17 sqg. Dessoulavy De la particule av dans Thucydide. Neuchatel 1895. 
 p. 22 sqg. Dessoulavy accepts "without argument the omission of civ in 
 relative and temporal clauses in Thukydides. He says: "Quoi qu'il en 
 soit, la question ne nous parait pas absolument resolue pour les propositions 
 conditiounelles, comme elle I'est pour les propositions relatives et tempo- 
 relies, oil Fomissiou de civ, plus frequente, est hors de doute." 
 
 3) Quaestiones yrainmaticae. Lipsiae 1886. p. 26 sq. 
 
 4) De sermone Time, quatenus cum Herod, congruens differat a scrip- 
 toribus Atticis. Lipsiae 1889. p. 67 sqg. 
 
 5) With av — l^exQi 1, fisxQi ov 1, ov 1, nQiv 6; without av — f^eXQ'' 
 
 1, flixQl ov 4, oi 1, TTQIV 5. 
 
 2* 
 
20 
 
 number of times that the particle occurs in temporal clauses 
 in Thukydides and the number of times that it is missing is 
 not such as to demand the insertion of the word against the 
 manuscript tradition. 
 
 § 2. TENSE. 
 
 This division will deal with tense only as indicating the 
 sphere of time. Normal usage needs no remark. 
 
 The historical present is a matter for separate treatment, 
 and it must suffice here to state the usage of Thukydides in 
 connection with temporal clauses^). 
 
 In the main clause, the conditions under which the 
 historical present appears do not differ from those under 
 which it occurs in sentences of other sort. It is by no means 
 infrequent, appears alone or coordinate with imperfect or 
 aorist, and is found whether the temporal clause denotes 
 antecedent, contemporaneous, subsequent, or overlapping 
 action. The following passages show the historical present 
 in the principal clause*): eTxsi — YII26 1. meid/] — 124 6. 
 46 3. 58 1. II 70 1. 815. 82. 83 3. 1011; 5. Ill 20 1. 
 34 3. IV 113 2. Y 17 2. 43 3. 72 3. YI 101 3. YH 32 1. 
 43 3. 82 1. YIII 80 1. 94 1. votsqov ^' — YI 4 2. cbg — 
 1611; 3. 65 1. 126 10. II 79 5. 84 3. IH 7 5. 8. 27 1. 
 69 1. 108 1. 112 5. lY 25 9. 70 1. 73 4 75 1. 89 1. 93 2. 
 110 1; 2. 116 1. 1283. Y 6 1. 8 1. 10 5. 36 1. YI 60 2. 
 YII 73 3 83 2. 84 3. YIII 17 1. 28 2. 38 4. 51 1. 60 3. 
 74 3. rjvixa — YII 73 3. ore — YI 74 1. juexQi ov — IE 28 2. 
 TiQiv — II 12 2. 67 3. 82. IH 29 1. YI 4 2. 97 2. 
 
 In the subordinate clause the examples of the historical 
 present are few. It occurs with cog when the main verb is 
 
 1) For Thukydides the historical present has been discussed by Eode- 
 meyer {Das Praesens historicuin bei llerodot und Thukydides. Basel 1889), 
 on the basis of a theory of his own. 
 
 2) In a few instances the historical present occurs two or more times, 
 in others it is followed by the imperfect or the aorist. Earely an aorist 
 participle intervenes between the subordinate clause and the main verb. 
 It has seemed unnecessary to include a few passages in which a temporal 
 clause depends immediately upon an imperfect or an aorist with which 
 an historical present appears in coordination. 
 
21 
 
 itself an historical present, V 10 5: xdv roinco Bgaoidng wg 
 Sga rov xaigov xat ro orQarevjua tcov 'Ad^p'aicov xivovjuevov, Xeyei 
 roTg jued' iavrov, and VII 84 3: cog Sk yiyvovxm tn avrcp, 
 EOJibiTovoiv ovdevl xoajuq) hi, and when the main verb is aorist, 
 I 63 1 : £7iava^(OQCov dk 6 'AqiOTEvg and trig dixh^ecog, (bg ogq 
 ro alio oTQarevjua ^oorj/xevov, tjjfOQr^oe. It occurs with m)g 
 when the main verb is imperfect and an imperfect is 
 coordinate with the historical present of the temporal clause, 
 YU 81 4: ov nQOv^MQei jiaXXov i] ig j-idirjv ^vvexdooexo, t'cog 
 ivdiazQi^cov xvxAovrai re vji avrcbv xal iv noXXco doqvficp avrog 
 re xal oi juet' avrov [A&f]vo.Toi] rjoav, and with jiqIv di] (= ecog) 
 when the main verb is imperfect, VII 39 2: eju noXh difjyov 
 rrjg tjjuEQag jieiQCOjuevoi dXX)jXojv, jtqIv Si] 'AqIotow . . . jtel&ei 
 Tovg ocpETEQovg rov vavrixov uo)rovTag, and aorist, I 132 5: dXX' 
 ovd" ojg . . . rj^icoaav vecoteqov ri ttoieIv ig avrov, . . . jtqiv ys 
 Sr] avroTg . . . dvtjQ 'AgyiXiog . . . jui]vvri]g yiyvErai . . . xal . . . 
 XvEi rag imoroXdg. In every case the relative position of the 
 principal and subordinate verbs is normal, that is to say, 
 prior clauses precede the main verb, posterior clauses 
 follow. 
 
 As regards the attitude of Thukydides toward the formal 
 present with future force in connection with temporal clauses, 
 two or three passages call for notice. Ill 39 8: rh> yqovov 
 roTg vvv xa&Eor)]x6oi SeT ij^dgoig dvMaraodai, rolg olxELOtg 
 $vjujudyoig jioXE/LirjoojUEv. The force of the expression Sei 
 dv&ioraodai is close to that of the future indicative as used 
 rarely in temporal clauses. VI 77 2: // ^) juevo^uev Eoyg av 
 ExaoroL xard noXEig Xrjcp'&cbjUEv, EtSoxEg on ravrt] juovov dXanoi 
 EOjUEV, Two interpretations are given for jhevojliev. First, it is 
 synonymous with the present subjunctive or the future indica- 
 tive, but throws greater emphasis upon the thoughf*). Second, 
 it is a true present, and ^ juevo/aei' means "aw ideo coniunctis 
 viribus iis non resistimus, quod exspectamus'^). Either view 
 
 1) The junction through »/ is awkward. Should not >/ be read? Cf. 
 Hom. A 666 — 7: >/ f^iivei elg o xe 8t) rfjsg &oai ayxc daXaoa)]g 
 
 'AQyei'wv asxt]ii Jivgog dtjioio deowvzai; 
 
 2) So, among others, Kriiger, Classen, Boehme-Widmann. 
 
 3) Stahl. 
 
22 
 
 must t'lnbrace uot only fuvofiev but the parallel o'ioaeda. In 
 the thought of the passage there is nothing against the present 
 sphere; and if /ah'o/iiev can be regarded as a present, the 
 fact that the license required for the former interpretation is 
 then unnecessary, is sufficient ground for the rejection of that 
 interpretation. It may be concluded, then, that this sentence 
 does not offer a case of the present with future force. VIK 
 109 2 is doubtless s])urious. 
 
 In the subordinate construction introduced by a verb of 
 "wishing that follows a secondary tense, the verb of a temporal 
 clause may be of the ])ast sphere, as IV 67 4: sOeot' dgojuco 
 ix Ti]g evedgag, ^ovkojuevoi cpduoai tiqIv ^vyxh]o&fjvai ticDuv rag 
 mdag y.al eojg hi fj ctjua^a iv avralg rjv. Also V 16 1, The 
 writer looks beyond the conception as it stood in the original 
 thought, and makes the predication according to the realization 
 that he sees in fact. 
 
 By a free combination of spheres the representation of 
 an action in its relation to the jiresent may lead to certain 
 irregularities in the form of expression Thus, of antecedent 
 action that overlaps upon action in the past sphere and is 
 going on in the present, Thukydides has two examples. 
 I 68 2: ov TiQiv Tidoyeiv, u'/X tTTeid)) ii' T(p egyco lo^iev, xovg 
 ^vjiijudyovg Tovods Tiagexakeoare. II 8 3: hi dh Avjlog exn'/j&rj 
 o/Jyov Tigo tovtcov , ngoTegov ovjico oeio&Eloa acp ov "EXXip'eg 
 lu€juv)jt'Tai. Of future action inserted in action that is 
 represented as true for the present, there are several in- 
 stances; for example, YI 17 1: eojg iyoj re hi dx^idCco juer 
 avT}~]g xai 6 Nixiag evrvyjjg doxei eivai, djTo/gijoaode t/) exaregov 
 fi/xibv (hcpeXiq. Further, with eojg — I 78 -4 (Ioti omitted). 
 Ill 70 6 (the principal clause has j^ieXXeiv . . . dvajreioeiv). 
 rV 67 4 (imperfect for present). V 9 6. YI 49 1; 2. 
 ATI 47 3 (Ion omitted). YIII 40 3 (eon omitted); with 
 iv w — Y 16 1 (imperfect for present). Of overlapping 
 subsequent action that extends into the present, there is a 
 single case. I 76 2: ede^djueda . . . a^ioi re ajua vojui^ovxeg 
 eivai xai vluv SoxomTeg, jueygi ov rd ^v fxcpEgovxa XoyiCojuevoi rep 
 dixaUp Xoycp vvv ygfjo'&e. The present ygijods is reinforced by 
 
23 
 
 the adverb vvv. In all these passages it is the dependent 
 verb that shows the irregularity. But in VI 77 2 (quoted 
 above on page 21) the main verb, not the subordinate, is of the 
 present sphere. So I 137 4: og xaxd juev jiXeToza "EXXrjvaiv 
 EiQyaojuai rov vjuezeQov oTxov, ooov j^govov tbv obv jiarsQa ejiiovra 
 e/jLol dvayxTj rjjjLvvojiA-rjv, noXv d' ezi TiXeio) dyad^d, ijzeidi] ev rep 
 docpaXei juev ijnoi, ixeivco dk ev ejitxivdvvco Jidhv t] djzoxojLiidi] 
 iyiyvETO. 
 
 The substantive verb of the present sometimes appears 
 with an expression of time when a past fact stated in the 
 dependent clause is dated. I 13 3: ht] ^' iorl judhoxa 
 TQiaxooia ig r)]v rekevrrjv rovde zov noXe/Aov, ore 'A/bieivoxh~]g 
 2!ajuioig r^Xd^ev. Ov jioXvg XQovog eneidr) is almost phraseological, 
 'not long since' in I 6 3: xal ol jtQeof^vreQoi amolg twv 
 evdaifiovMv did xb df^Qodiauov ov noXvg ^Qovog ejteiS)] /ixcovdg 
 re Xivovg enavoavxo cpoQovvxEg. Some difficulty is presented by 
 another passage, I 14 3: dips, xe dqf ov 'A§)]vaiovg OejuioroxXrjg 
 eneioev Alyivr]xaig TioXe/uovvxag, xal djua xov ^aQ^dgoi' ngoodoxi- 
 juov ovxog, xdg vavg Jioiijoao&ai. Classen, Stahl, and others 
 regard oipe xe dcp' ov . . . eneioEv as arising from a fusion of 
 two points of view. Yet it is not easy to admit this when 
 the words that indicate the two positions are in immediate 
 connection. Shilleto suggests that £xexr)]vxo may be carried 
 on from the preceding, but against this Stahl argues that the 
 later sentence refers to the Athenians alone. Kriiger and 
 van Herwerden cut the knot by deletion of dcp ov. The passage 
 'has hardly as yet found a satisfactory explanation. 
 
 A fact thus dated by a numerical phrase with the sub- 
 stantive verb of the present may itself be expressed in its 
 relation to the present. So I 6 5: ov noXkd ett] ETxeidrj 
 jiEJiavxai, and 18 1 : eti] ydg eoxl judXioxa xExgaxooia xal oXiycp 
 jiXeiw ig TTjv xeXevxtiv xovde rov jioXe/uov, dqf ov Aaxedatjuovioi 
 rfj avxfj JToXiTEia ^Qojvxai. 
 
 § 3. STEM -SYSTEM. 
 
 It will be convenient to consider the subject under the 
 classification of the temporal relation of the actions. 
 
24 
 
 A. Antecedent action. 
 
 ^VluMl the action of the subordinate clause is antecedent, 
 completely or partially, no restriction arising from the tem- 
 poral clause affects the aspect of the main action. Hence 
 consideration of the leading verb is unnecessary here. In 
 the dependent clause, the aoristic view is used of action 
 purely antecedent, while the extensive view may be used 
 either of such action, or of action overlapping upon that of 
 the main clause and therefore at once antecedent and con- 
 temporaneous. On account of possible ambiguity, the exten- 
 sive view is practically almost confined to the expression of 
 overla])ping action. 
 
 A good illustration of the imperfect used by Thukydides 
 to represent overlapping action is found in Y 72 3: eTieidi] 
 yoLQ ev '/Fooh' eyiyvovro rdig evavrioig, to j^ikv JMaxTirecov ds^tov 
 TQETCEi avxMv Tovg ^yuoiTaq not Tovg Bgaoideiovg. One would 
 perhaps expect the aorist iyevovTo as a parallel to the aorist 
 participle that occurs in YII 5 2: y.ai iv y/Qol yevo/iisvoi i^udyovro. 
 (Compare III 108 1.) But in the progress of the narrative, 
 it appears that the action of the imperfect eyiyvovro was not 
 over at the time when the main action took place, Y 72 4: 
 xai xavTY] /uev fjoocbvxo ol AaxEdaijuovioi * tc5 d' akXo) orgarojtedq), 
 xal judXiora rco jueocp, . . . eigeyjav ovde ig xeTgag Tovg jioXXovg 
 vjiojLieivavzag, aXX' cog ijifjoav ol Aaxedaiuovioi, eviDvg evdovrag. 
 
 At first sight it might seem that the distinction between 
 the aorist stem as denoting action wholly antecedent and the 
 present stem as denoting overlapping action could not be 
 maintained for the many examples of verbs of perception 
 (actual or intellectual) and thought; for these verbs appear 
 sometimes to occur in forms from the present stem and from 
 the aorist indifferently. This seeming indifference arises from 
 the fact that for the most part the choice of the manner of 
 representation is in the writer's power. Cases where there 
 is no option test the principle, and these are clearest when 
 the verb expresses perception by the senses. Thus, I 72 1 : 
 roiavja juev ol Kogiv^ioi ehrov. . . . xal d>g fja^ovro xcbv Xoyoiv, 
 edo^ev avroig TcaQcrrjrea eg jovg Aaxedacjuoviovg eivai. Further 
 
25 
 
 illustration is furnished by a comparison of II 94 2 — 3 : 
 ^orjd^YjoavTeg de (ifl r'jjueQa 7iavd7]juel oi 'A&rjvaioi Ig rov UeiQaia. 
 vavg re xa&eTXxov. ... oi de Ihlonovvrjoioi, cog fjo&ovro t})v 
 ^oij&eiav, . . . xaTO. xajog inl rfjg Nioaiag h'nleov, with VII 18 1: 
 jiaQeaxEvdCovTO de xal t>)j' eg ti]v 'Amxtjv eo^oXijv ol AaxeSai- 
 juovioi, cooneq xe JtQoededoxjo amoTg xal rcbv 2vQaxooio)v xal 
 KoQtv&iMv erayoiTcov, ejieid)] enwddvovio Ttjv axed tmv 'Adijvaiojv 
 fio/j&eiav eg rijv 2!ixeXiav, oncog dij eopoXrjg yevojuevijg diaxcjoXudfj. 
 
 The imperfect is rarely used when the conception of 
 overlapping action seems excluded. II 18 3: eneid)) re ^vve- 
 Xeyero 6 OTgarog, ij re ev tm lo&fACO ejii/xorij yevo/iih')] xal xmd 
 rrjv dXXijv noQeiav fj oxoXai6T)]g diej^aXev avrov. The reference 
 in fj emfiov/] is doubtless to the incidents narrated in II 10 3 
 — 12 5 and these, it is expressly stated, occurred ejieiSt] 
 ndv TO OTQareviLia ^vveiXey/nerov t)v. In Y 17 2: ejreidij ex rcbv 
 ivi'oScov cifia TioXXdg dixaicooeig jigoeveyxovrcov dXXi'jXoig ^vve- 
 XMQeiTO coore a exdiegoi jioXejuco eoiov djicjSovrag Ttjv eiQ/jvtjv 
 Ttoielo&at, . . . noiovvxai xijv ^vjufiaoiv, unless the force of the im- 
 personal ovyxoiQel is attributed to the neuter passive ^vveyoiQeho, 
 the subordinate action is wholly antecedent to the main action. 
 It would not be difficult, in either passage, to assume textual 
 corruption in the subordinate verb. 
 
 Antecedence expressed as temporal may pass into causal 
 antecedence. How close together the two conceptions lie 
 may be seen by examination of two or three passages. 
 II 3 1 : xovg Xoyovg de^djuevoi -^ovyo-Cov, ciXXcog xe xal ejieidi] 
 eg oudeva ovdev evecoxeQiCov. YII 69 2: 6 de Nixiag vno xcbv 
 jiaQovrojv ex7ienX)]y [^levog xal oQcbv olog 6 xivSwog xal cbg eyyvg 
 ijS)], ejieid)] xal ooov ovx ejiieXXov dvdyeo&ai, . . . av^ig rcbv 
 xQujQdgxcov eva exaaxov dvexdXei. 80 1 : xco Nixicx xal Jrjjuo- 
 o&evei idoxEi, ejieidij xaxcbg ocpioi ro orgdrev/ua eiye rcbv re ejiirrj- 
 deictiv Jidvxcov dnoQiq. y'jd)] xal xararerQavjuariojuevoi fjoav noXXol 
 ev TioXXalg ngoofioXalg rcbv jioXe/uicov yeyevrjfievaig, nvgd xavoavxag 
 cbg nXeloxa dndyeiv xtp' oxgartdv. 
 
 Failure to appreciate causal force in the imperfect in 
 clauses of antecedence has brought failure to grasp the 
 
26 
 
 thought of Thukydides in sonio passages, and has consequently 
 led to conjectural emendation. 
 
 II 51 5: rhe .-TQooioiev, diecpdeigoiTO, xal jurVuora ol agexfjg 
 Ti /neruTToiovjun'oi' aio/vvrj ydo i)q)eiSovv o(f>a)v avxcov loiovreg 
 TiaQcx Tore (pt?Mv^, Ijtu xal Tag oXoqnvQoeig tojv djroyiyvojuevwv 
 relEvroivreq xal oi oixeToi iiexa/uvov vjto tov Tiokkov xaxov 
 vixojfievot. Of tlie many explanations proposed for this passage, 
 those that involve textual change^) may be disregarded here, 
 and the others fall into three classes. 
 
 a) tjiei = whereas. This is the interpretation given by 
 Steup in the fourth edition of Classen. He supports this 
 signification of met by two other passages in Thukydides, 
 VI 79 2 and VII 30 2. In both these places, however, enei 
 means 'for'. In VI 79 2 it gives the ground for the sentence 
 that precedes; and in VII 30 2, the support for the previous 
 statement, aTioxTeivovoiv avjcov iv rfj eofidoei Tovg rikeiorovg. In 
 defense of the meaning 'whereas' in the former place, Stahl 
 cites two passages from Plato, Prot. 333 C and 335 C; but in 
 both, strictly, enei is causal^). 
 
 b) EJTei = since. Attempts to refer to words beyond 
 aiaxvvf] are vain. The argument by which the causal sentence 
 is brought into relation with the words immediately preceding 
 is stated thus by Stahl ^): "Haec enuntiatio cum superiore 
 ita cohaeret ut amici ad curam aegrotautibus adhibendam eo 
 impulsi esse dicantur quod postremo etiam doraestici ita 
 defatigati fuerint ut ne lamentari quidem morientes sustine- 
 rent, nedum opem aegrotautibus ferre possent." 
 
 c) ejiei = after. No editor advocates this view, and the 
 only mention of it is its summary dismissal by Steup : "Ohne 
 Frage haben ja die dgeryg ti fieTajroiovjueroi sich um ihre er- 
 krankten Verwandten und Freunde nicht erst gekiimmert, 
 nachdem die Zeit eiugetreten war, in welcher xal rdg dXocpvg- 
 
 1) See Torstrik Philol. XXXI (1872) p. 39 s(]q. Steup Rh. Mus. XXVI 
 (1871) p. 473 sqq. 
 
 2) Cf. Zycha Wiener Stud. VII (1885) p. 96 sq. 
 
 3) For other explanations, open to greater objection, see Classen ad 
 he, and Fritzsche Philol. XXXII (1873) /;. 147 sqg. 
 
27 
 
 oeig rcbv djioyiyj'Ofievrov xai oi olxeXot i^exafn'ov.^^ This is in 
 harmony with Steup's opinion that there can be no causal 
 connection; but the interpretation oi ajrei as causal presupposes 
 antecedence in time. Steup's argument from probability has 
 no weight. So long as the strength of kinsmen in the close 
 relations of the household sufficed for the care of the sick 
 and for lamentation, there was no reason why even ol aQexfjg 
 XI juerajioiovfiEvoi should take the place of those to whom 
 these duties by a right almost sacred peculiarly belong. 
 
 Against each of the interpretations given above, it may 
 be objected that there is no good ground for the imperfect 
 iiexajuvov with jehvicovrEg, 'at last' (compare II 47 4. YI 53 3 
 et al.)^), and that the accusative after exxdjuva) is unparalleled. 
 Is all exegesis impossible? One other explanation may be 
 ventured, d^ocpvgoeig is the object of TEXevTC~)VTe<;, which means 
 'performing, fulfilling'. This obviates not only the combi- 
 nation of reXevTcovreg with the imperfect, but also the accu- 
 sative after exxd/uvco, and substitutes therefor the ordinary 
 participial construction. The imperfect with mei then denotes 
 overlapping action with causal coloring, and the meaning of 
 the passage is, 'after they saw that merely in making the 
 lamentations over the dying the relatives themselves were 
 exhausted'. 
 
 A participle of a verb of perception is sometimes found 
 in coordination with a clause of antecedence with causal 
 coloring. 
 
 The manuscripts read in Y 55 4: xal ^A&y^vaiwv avrolg 
 yihoi E^ori{h]oav ojiXirai xal 'Alxi^iddi^g OTQarrjyog' 7iu&6fJ,evoi 
 de Tovg Aaxedaifioviovg i^eoTQarevod^ai, xal cog ovSh en avrcbv 
 edei, dm~jXdov. A difficulty presents itself in that, obviously, 
 the fact that the Lakedaimonians had undertaken an expedition 
 could not be a reason for the departure of the Athenians. 
 Portus' proposal to omit Se and unite jivdujuevoi . . . i^sotQa- 
 revo&ai with tiio)jOi]oav has met with favor, and Classen, 
 Stahl, van Herwerden, and Boehme-Widmann, among others, 
 adopt this reading. But it is against probability that a scribe 
 
 1) See also Hultsch op. cit. p. 446 sqq. 
 
28 
 
 sliould, iu the face of the o)^ clause introduced by y.ai., attempt 
 to draw the participle over to this sentence. Moreover, the 
 aid must have been sent by order of the Athenians, and 
 Thukydides would hanlly ascribe to the military body the 
 perception that led to the expedition of the forces. Some 
 critics have sought either in ti - or in the perfect the sense 
 of completion, and have interpreted, 'had ended their ex- 
 pedition'. For such meaning of this compound there is no 
 support; and even if it were possible to find, iu the pre- 
 position or the stem, the meaning assigned, there would be 
 an irony quite out of harmony with the surroundings in the 
 statement that the Lakedaimonians had completed their ex- 
 pedition, when in fact they had been unable to leave their 
 own territory. May not the explanation be found in a scribal 
 blunder that would have been extremely easy, namely, the 
 omission of ovy. before i^eorgarevodai? This reading gives 
 point to the middle voice as contrasted with the active of 
 § 8; and the perfect with the negative conveys, in regard to 
 an action past with reference to jrvdojuevoi, the notion of 
 inability that would be expressed, were the actions contem- 
 poraneous, by the present stem with the negative. 
 
 V 58 1 : 'Agyeloi ds ngoaiodojuevoi to re Txoonov Ti)v Jiagaoxevrjv 
 Tcbv Aaxedaifiovion', xal ijretd)] eg tov ^/.e(ovtTa j^ovl6f.iEvoi To7g 
 aXXoig jiQOOjim^ai i^ojgovv, tots Si] e^EOTgdrevoav xal avroL In this 
 passage the coordination that is effected by t£ , . . xai has 
 seemed to some critics to be inappropriate and inexact. 
 Classen reads tot£ Tigonov and regards xai as an interpolation. 
 Miiller-Striibing ^) wi'ites to re. but thinks that some such 
 words as jiagexdXeoav xovg ivjLifmxovg should be inserted after 
 rcbv AaxedaijuovtMv. Even Stahl, though he defends the 
 reading of the manuscripts, speaks of the trajection of re: 
 "Particula re autem ita traiecta est, quasi sententia ea deinde 
 ratione conformetur, quam interpretando expressit schol.: 
 'Agyeioi 7igoaio§6jjLevoi r/p' re ngcon^v rcov Aaxedaijuovicov 
 Jiagaoxevi]v xai av&ig jTgoo/ajgouvrag rovg Aaxedaijuoviovg em 
 ^Xiovvrog vjieg xov rdig tdioig ov/ujul^ai ovjujud^oigJ'^ But eneidrj 
 
 1) Thukydideische Forschungen. Wien 1881. p. 99 sqq. 
 
29 
 
 with the imperfect often implies perception, and Thukydides 
 can say to t£ jiQoinov, since he has in mind also something 
 perceived later. Compare V 44 1: ol de 'AQyeioi axovoavxeg 
 Tfjg T£ dyyeUag xal eneidr] eyvcooav ov juet' 'A&ijvai(ov Jigaxi^EToav 
 X7]v T(bv Boiojtcov ivjLifiaxiav, . . . rcov juh iv Aaxedai/uovi jiqI- 
 o^ECOv . . . y]jueiovv. Also IV 116 1. 
 
 Yn 60 5: 6 de Nixiag, meidi] ra noXXd hdljua r}v, oqcov 
 xal rovg oxQaxiwxag xco xe Jiagd x6 eicod'og noXv xaig vavoi 
 XQaxy^rjvai d&vjuovvxag xal did xi]v xcbv ejitxi]deiajv ondviv cog 
 xd^ioxa (iovlofiEVOvg ^) diaxivdvveveiv, ^vyxaleoag dnavxag Tiage- 
 xeXsvoaxo xe nqwxov xal ele^e xoidde. Of recent editors Classen 
 alone speaks in defense of xai before xovg oxgaxiMxag. He 
 reasons that the preceding consultation had been held only 
 with the strategi and taxiarchs, while here the effect that the 
 measures adopted had had upon the whole body of troops is 
 in question, and is assigned as the occasion of Nikias' 
 summons of all to an assembly. But justification of the xai 
 exists rather in the connection of the participle SqcTjv with the 
 preceding clause. Nikias sees that preparations are almost 
 complete and that there is need to encourage the dispirited 
 soldiers before the engagement. 
 
 The bearing of Thukydides' use of the reflexive pronoun 
 upon this point is not to be overlooked. Dyroff^) explains 
 the appearance of oq:dw in temporal sentences where the use 
 
 1) This is the usual reading. Van Herwerden {Mnem. I N. S. p. 178 
 sq.) transposes zs, and writing jioido/iisi'og, gives: ogaJr zs zoik ozgaziMzag 
 zcp naqa to eIxo? noli) zaZg vavai KQazrjdfivai d&v/itovvza?, xal 8ia zijv zu>v 
 sjiizrjdeicov andviv (bg za^ioza ^ovkofxevog diaxivdwEvsiv, For, as he justly 
 remarks, "pugnat ilia militum a^vfiia cum hac primo quoque tempore de- 
 cernendi cupidine, nee quicquam est in Mciae concione, unde iste pugnandi 
 ardor confirmari posse videatur." It seems better to follow the manuscripts 
 more closely and write oqcov xal rovg azgazicozag zqi ze jiagd z6 eioidog nolv 
 xaig vaval xoaTTjdijvai ddi'iuovvTag xal did rijv zcov EJiiTt]8eio)V ojtavtv, a>g 
 zd^ioza (iovl6f.ievog diaxivdwevsiv. For /]ovl6/iievog cf. II 86 6. There need 
 be no objection in the array of participles, for oqwv is closely united with 
 the ETCELdri clause in the expression of the external grounds, ^oidofiEvog 
 gives the motive, and ^vyxaUoag begins the description of the action. 
 Cf. Stahl on III 3 5. 
 
 2) Geschichte des Pronomen Refkxivum. Schanz's Beitrage. Heft 9. 
 1892. II p. 11. 
 
30 
 
 might be called anaphoric, by ''subjectivity" of the pronoun. 
 He considers the clause purely tem})oral in A^ 73 3: xal 
 yerofih'ov toi'tov ol /luv 'AdrivaToi Iv tovto), tog jiaorjXdT. xai 
 e^ey.hvFV ujto oqcov to OTQarevjua, xad' fjovyjm' ioo')d)]aav, and 
 he does not observe that the shift of pronouns corresponds 
 to a shift of tenses in I 58 1: ijieidrj ex re 'A{^i]vaia)v ex 
 jToXkov TTQaooovreq ovdev tfvQoiTO t7iir))deiov, alT al vfjeg em 
 Maxedovldv xal em acfdg ofiokog mleov, xnl t« reli] rcbv 
 Aaxedaiftoi'icov vTieoyeJO nvrdig, yv em Tloreidmav I'vooiv 'A&}]vdioi^ 
 eg r)]v 'Arzixiji' eofiaXelv, rare dt] xard rov xaiQov tovtov dqyiorajTai. 
 This use of the reflexive pronoun as almost anaphoric in a 
 temporal clause of antecedence occurs in the case of over- 
 lapping action when there is causal coloring. The instances 
 are as follows: — I 30 3. 58 1. Ill 108 3. V 65 5. 
 73 3. YI 63 2. YII 80 1. A single passage shows o<f)cm> 
 with the aorist, YIII 90 1 : JTQoregov re, enel rdyiora xareoryoav 
 xal ejreid)] rd ev if] 2^djna) ocpMv eg dyj^uoxgaTiav aTxeorrj, ifQeofieig 
 Te djxeoreXXov ocpoyv eg t)]v Aaxedaffiova, . . . noXXw le judXXov en, 
 ijietdrj xal ol ex Trjg 2!d/uov Jigeofieig oq^aJv ijX&ov, OQonneg xovg 
 T£ TioXXohg xal ocpdiv rovg doxouvrag tiqoteqov morovg elvai 
 juera^aXXo/Lievovg. The recurrence of oq)(bv excites suspicion, 
 and van Herwerden ^) proposes to delete the word after 
 dneoxeXXov, on the ground that savrcov or o(pa)v amdjv w^ould 
 be necessary under the circumstances. But compare Dyroff 
 op. cit. p. 5. oq)d)}' preceding di]juoxQaTiav can hardly be 
 correct, for the democratic party was in power at Samos 
 when the Four Hundred established themselves at Athens 
 (YHI 21), the attempts at oligarchy had been unsuccessful 
 (73 6), and the Four Hundred never had a sufficient hold 
 at Samos to justify the expression ret ev ifj .Zdjuco ocpwv djieorrj. 
 Possibly oacpcag was written by Thukydides, having regard 
 to YHI 75 2: /xexd de rovro XajujTQcbg ijd'i] eg d7]^uoxQaTiav 
 ^ovXojuevot jueraoTTJoai rd ev rfj ^dfxw. The o(pd)v before r^Xd^ov 
 looks like a gloss; for it is not probable that the ten men 
 sent to Samos were of the number of the Four Hundred, 
 and as no envoys had been sent from Samos, there would 
 
 1) Studio Thucydidea. Traiecti ad Rheu. 1869. y. 109. 
 
31 
 
 be no need to define jigeoj^eig as those sent by the Four 
 Hundred. 
 
 Similar domination of the agent of the main clause over 
 the subordinate is seen in such a passage as II 81 7 — 8: 
 Tcbi' de 'EX/i}]rix<x)v GXQarojzedcov ovderegov fjo&sro rrjg j^^dxV^ ^'■^ 
 TO jioXv jTQoeX&eh' amovg xal oxQarojiedov oh^&fjvai xaxaXrj- 
 yjOjUEVovg ijieiyeodai. enel de ivexeivTO cpevyovxeg ol ^aQ^aqoi, 
 aveXdf.i[iavuv re avrovg xal ^vi'ayayovTeg to. oxQarojieda fpv](^a^ov. 
 Although Of /^dgfiaoot is formally the subject in the temporal 
 clause, the failure to indicate the subject of dveXdju^avov 
 shows that the continuity of thought is almost as unbroken 
 as if the subordinate clause read ejrel Se fp&dvovxo xovg jjcxq- 
 ^dqovg iyxeijuevovg. 
 
 Clauses of antecedence offer three instances of the 
 historical present. Which point of view does this represent? 
 Two of the examples show verbs of perception, and the 
 most that can be said is that they are not necessarily of 
 either extensive or aoristic aspect. I 63 1 : ejiavaycDQcoy de 6 
 'AQioxevg djio xrjg didb^ecog, cog oqu x6 dkXo oxgdrev/ua f]oo7]juevov, 
 fjjioQrjoe jiiev SnoxsQcooe diaxivSwevo)] icoQ}]oag. (Compare III 
 108 3. IV 33 1.) Y 10 5: xdv xovxco Bgaoldag cog Sga xov 
 xaiQov xal xb Gxgdxevjua xcbv "A&rp'aicov xivovfievov, Xeyei xoTg 
 fxex^' eavxov. (Compare 11 90 3. Y 8 1.) The third case is 
 as follows, YII 84 3: cog de yiyvovxai en avxco, eonmxovoLv 
 ovdevl xoojucp exi. The action of the subordinate clause is 
 wholly antecedent, and the point of view accordingly aoristic. 
 (Compare YII 35 2.) 
 
 The position of the perfect stem in this connection remains 
 to be considered. It has been seen above that, in the tempo- 
 ral correlation of two actions in strict sequence, an extensive 
 view of the prior action leaves room for ambiguity, precision 
 demands the aoristic view. An attempt to compensate for 
 the loss of the extensive by substitution of the intensive was 
 natural, and investigation shows that Thukydides uses the 
 pluperfect in a temporal clause of pure antecedence when he 
 wishes to compensate for the extensive view lost by the use 
 of the aorist. The pluperfect thus corresponds to the im- 
 
32 
 
 perfect in its power to express duration, description, causal 
 color, and, with a negative, failure of ])0wer or will. Demand 
 for the extensive view may be revealed in various ways. 
 There is an adverb of extent of time in III 98 1 : ineidr] de 
 Tov Tf To^doxov ano'davovTog ovrot dieox€ddo§t]oav xal avroi 
 Ixexfo'jy.Foar xal ejil noXv rc5 avrco novco ^vveyoixevoi, oi re 
 AhdiXoi irexeoTO xal eotixovn^ov , ovrco Si] TQajio/nevoi eqpevyov. 
 Often the action of the temporal clause has already been de- 
 scribed by an imperfect or by a narrative. This is especially 
 marked when pluperfect is combined with aorist, as in I 29 
 3 — 4: ol KeQxvgaToi xiqQvxd re jigoeiiejuyjav avroig ev dxarico 
 OJieQOvvra jui] nlelv em ocpdg^ xal rag vavg d/ua enlrjQovv, t,ev^avreg 
 re rag nakaidg wore JiXoifiovg eivai xal rag dXXag mioxevdoa^nceg. 
 cog de 6 xfJQV^ re djirjyyeiXev ovdev eiQtjvaiov naod ran' Kogivdiwv 
 xal ai vijeg avroig eneJiXrjQOivro ovoai oydorjxovra, . . . ivavjudy7]oav. 
 Causal force and indication of the spectator are not in- 
 frequently found, as in YII 47 3: rco ovv Arjjiioo&evei ovx 
 edoxei eri yofjvai jueveiv, dXX' djieg xal diavoi]&elg eg rag 'EjiinoXdg 
 diexivdvvevoev, ejietSi] e'oq)aXro, djiievai eipr}g?iCexo. On this prin- 
 ciple it is possible to explain the apparent anacoluthon in 
 III 96 3: rovg de AlrojXohg ovx eXdvd^avev avri] )] nagaoxev)] 
 ovre ore ro nqcbrov bie^ovXevero, ejieidrj re 6 orgarog eoe[ie(ih)xei, 
 TioXXf] yeiQi ene^ori&ovv ndvreg. All that would be stated by a 
 second clause modifying ovx eXdv^avev and parallel with ore . . . 
 ejie^ovXevero is given in ejieidi) . . . eo£^e^X)]xei if this has the 
 force, 'after they saw that the army had entered', and Thuky- 
 dides, with his characteristic rdyog, utilizes the possibility of 
 proceeding thus to the next thought. 
 
 B. Contemporaneous action. 
 
 The different phases of contemporaneous action are marked 
 by the forms assumed by both the independent and the sub- 
 ordinate verb. When distinction is possible between the 
 present and aorist stems, if both verbs are of the present 
 stem, the two actions may be coincident in extension — 
 necessarily, if the conception of limit is involved — or one 
 may be inserted in the other; if both verbs are of aoristic 
 
33 
 
 form, the two actions are in aoristic coincidence; if one verb 
 is of the present stem and the other of the aorist, one action 
 is inserted in the other. 
 
 With regard to these distinctions notes may be made on 
 a few passages. 
 
 I 138 1 : 6 d' ev rco y^qovco ov IneoyE rfjg IJeooiSog yXd)ao}]g 
 ooa edvvruo xaTEv6}]oe. Thougli the aorist ejieoxe is used of 
 action within which the action of the main clause is inserted, 
 the accusative or makes clear the extent of time, and the 
 preposition ev marks the relation of insertion. Compare III 
 81 4. Elsewhere, with or following the prepositions xard, 
 jTEQi, vjio, in similar expressions, the subordinate verb is im- 
 perfect. 
 
 YI 02 4: TO re cpilojioh ovx ev m ddixovftai eyoo, (OX Iv 
 (h aocpalGyg ejioXixev&tp'. In the first member of this sentence 
 the forms do not determine whether the conception is that of 
 coincident extension or of aoristic coincidence. In the second, 
 however, ejioXirev'&rjv shows the aoristic view. Hence, while 
 the supplement of Kriiger and Classen, elyov, is in harmony 
 with the scholiast's variation, eTtohTevofiip', consistency would 
 require here k'o/ov. 
 
 The form of insertion whereby the action of the temporal 
 clause falls within the action of the main clause is used by 
 Thukydides only to a limited extent. The imperfect of the 
 substantive verb appears in the principal sentence five times 
 (II 78 4: TooovToi. Ijoav ol ^vjimavreg ore eg rip' jrohogxiav 
 xadioravTO, xal aXkog ovdelg fjv iv rco reiyei ovre dovXog ovre 
 eXev&£Qog. 99 6. 102 5. IE 29 1. IV 91), and in two in- 
 stances (I 25 4. II 13 9) imrJQyov is found. A second verb 
 of equivalent force is joined with fp' in 11 102 5: jtqIv av 
 evQCOV iv rami] rfj yoiqa xaroixiorirm, ijrig ore exreive r}]v jui]TeQa 
 fXYjTico vjtd fjXiov ecoQdro jui]de yrj fjv. One case of the inverse 
 construction with the perfect stem occurs, Y 49 2: Aaxedai- 
 jiwvioi de TiQeof^eig jiefiyjavreg dvreXeyov fi)j dixakog oqmv xcira- 
 deSixdo&ai, Xeyovreg jLii] in)]yye}c&ai nco ig AaxeSatfiova rag ojtov- 
 ddg, or' ioejiejuyai' rovg ojr/Jrag. 
 
 In consideration of the demand for the extensive view 
 
 8 
 
34 
 
 under the conception of limit, comment upon some passages 
 is necessary^). 
 
 II 65 5: ooor re yno yooroy ttoovot)] Tfjg jzohcog ev rfj 
 eiot'iv)), /leroicog eitjyaTO, xal aorprdcog diecpvka^ev avT)p> xal iyevero 
 £.-7' iy.eivov fieyioT)j. The aorist in the temporal clause is owing 
 to the limitations of the word employed. dieq)vXa^ev is not a 
 second member of the sentence modified by ooov ygovov jiqovoti], 
 and if a comma is to be used, it should be placed after 
 i^)jyfTTo ratlier than after avj/p: 
 
 III 28 1: iv oocp d' av JidXiv tXdtooi, Udxjpa jld'jte dijoai 
 MvTih]vaUov i.i')]dha firjre dvdQanodioai. Here by confusion of 
 ideas, a particle denoting limit in both directions is combined 
 with a verb suited to a particle expressing limit in one 
 direction. So III 52 3. 
 
 IV 39 1 : ygovoq de 6 ^ujUJiag iyevero ooov 01 avSgeg 01 ev 
 rf] v)'jO(p ejrohoQxt]&7]oav , cuio Trjg vavjuax,io.g fxe^Qt T)~]g ev Tf/ 
 v)']ao) f(uyj]g, e/^dojiiYjyMVTa fjjLieQai xal dvo. This is the only 
 instance in which Thukydides chooses the conception of aoristic 
 coincidence in connection with a particle denoting limited 
 extent, and here it is justified by the character of the thought 
 in the main clause. 
 
 In the two examples (II 21 1. 54 4) of a ore clause 
 after a verb of remembering, Thukydides uses the aorist. 
 
 ^\\e historical present m connection with temporal sentences 
 of contemporaneity is used of the main action twice. YI 74 1 : 
 'A?,xil3idS)]g yoLQ me dTirjei ex Trjg dQy)~jg ydij jneraTiejuTcrog, emoid- 
 jiievog on cpev^ono, jui]vvei roTg tmv ZvQaxooiwv cp'doig rolg ev tj) 
 Meoorjvr] $vveiSd)g to jue?dov. VII 73 3: Jiefmei tcov eraigwr 
 Tivdg rcbv eainov juerd inneon' ngog to rwv 'Adrp'aiwr OTgarojiedov 
 fjVLxa ^vveoxoTa'Qev. In both cases the principal action is 
 inserted in that of the subordinate clause, and there is 
 accordingly no restriction on the choice of conception as 
 extensive or aoristic. In the subordinate clause of contem- 
 poraneity the historical present does not occur. 
 
 The perfect stem gives the extensive aspect in the 
 
 1) On I 90 3 see ^p. 38 sq., 43 sy. 
 
35 
 
 representation of a state or condition. Many of the passages 
 need no comment. 
 
 I 39 3: ovg XQW> ^^^ aocpaXeoraroi ipav, tote nqooievai, xal 
 jui] £v (p fjjuEig fxev tjdixijfie&a, ovtoi de xivdvvsvovoi, i.ii]b' iv co 
 vjUEig TTJg TE SvvdjUECog avTcov tote ov f^iETaXajiovTEg rZ/g dxpEMag 
 
 VVV jUETaScOOETE. 
 
 I 137 4: 0E,LuoTOxh~]g i]xco jiaga. oe, og xaxa fiEV TiXeloTa 
 EXXrivoiv Eigyaojiiai tov v/j,exeqov oixov, ooov yQovov tov oov 
 TiaTEQu EJiiona i/uol dvayxf] rjfivvo/urjv, noXv b' exi jiXeioj dya&d, 
 
 ETlEldi] EV T(p do(paXEl jUEV EjLlOl, EKELVCp 6e EV ETllXlvdvVO) TldXdV f] 
 
 djioxofuSt] iyiyvETO. Here Eigyao/im denotes a state resulting 
 from a past action to which the temporal clauses ooov 
 XQovov . . . fjinvv6jiu]7' and ejieiS)) . . . EyiyvETo refer. A similar 
 relation between the leading verb and the temporal clause 
 will be seen in other instances. 
 
 II 11 6: )iQ)} xal Jidvv eXjiH^eh' did fidyjjg lEvai avTovg, ei 
 jui] xal VVV ojQjLUjVTai ir co ovjico jrdQEOjUEv. 95 1 - 2: ZiTdXx}]g 
 6 TriQEOi 'OdQvoi]g, Ogqxcov [iaoiXEvg, EOTQdTEVOEV . . . Svo 
 vjiooxEOEig T}]v jUEV ^ovXojuevog dvajiqa^ai, Tip' Se avrog djioSovvai. 
 o TE ydg IlEgSixxag airccp vjiooyo/xEvog ... a vjiEdE^aTo ovx 
 ettexeXei, Toig te "A&tp'aioig amog d)jiwXoy)]XEt, ote T)p' ^vfipayjav 
 EjTOieTto, tov ml 0Qdx)]g XaXxiSixov jioX^Ef-iov xaTaXiVOEiv. IV 85 2: 
 vvv ydg, ote jiagEO/Ev, dcpiy/iiEvoi xal //era v^umv jTEigaoojuE&a 
 xaTEQydCeo&ai ainovg. 
 
 IV 133 3: ETi] Se fj XQvolg tov tioXJjiov tovSe EJiiXaj^Ev 
 oxTU) xal EvaTov ex fiEoov, OTE EJiEcpEvyEi. It seems impossible 
 to account for the pluperfect in this sentence. Emendations 
 propose, on the one hand, the imperfect or the aorist for the 
 subordinate verb, and on the other, the deletion of the 
 temporal clause^). Since the clause is by no means superfluous, 
 it is better to accept a change of tense. 
 
 V 49 2: dvTEXEyov /u)) Sixaicog oq)cbv xaTadEdixdo&ai, XEyovTEg 
 ju)] EJirjyyEXdai tio) ig AaxEdaijuora Tag onovddg, 6V EOEJiEjuipav 
 Tovg OTiXiTag. VI 49 1 : Ecpy] /Qfp'ai tiXeIv im ZvQaxovoag xal 
 jiQog Tjj tioXel (bg TdyioTa t)ji' f-idyjp' noiETodai, Eojg eti djiagdoxEvoi 
 
 1) Cf. Herwerdeu ad loc. and Stadia p. 67. Rutherford. Hude 
 Commentarii critici ad Thucydideiii pertinentes. Hauniae 1888. p. 145. 
 
 3* 
 
36 
 
 Tf Ftoi y.al finXima ey.7xe:rXi]yfievoi. VII '2 4: ervys de xara rovro 
 Tor y.aioov fA»9(bi' iv w tTna jnh y oxko oiadicov fjdi] djierereAeoTO 
 ToTg 'Ai'hji'w'oi^ eg rov jneyav hfievn dinXovv relyog nXi]v yMia 
 ^Qdyi' Ti TO .-700C Tijy Ihujinndv' tch'To (Y hi coxodo/aovv. 
 
 C. Subsequent action. 
 
 The reverse of tlie relations of antecedence is seen in 
 the ease of subsequence. If the action of the main verb is 
 viewed as aoristic, the relation is that of pure subsequence; 
 if it is viewed as extensive, there may be pure subsequence, 
 or the actions may overlap. When the actions overlaj), the 
 concej)tion may also take the form of limit, and this is the 
 only case in which there is restriction upon the aspect of 
 the subordinate verb by reason of the temporal relations. If 
 the verb in a clause of limit is aoristic, the action forms the 
 limit, if the action is viewed as extensive, the limit falls 
 within this action. 
 
 1. Main clause. 
 
 When discrimination between the present and aorist 
 stems is possible, the verb upon which a clause with the 
 infinitive after Jioh depends is as a rule aoristic. Some 
 instances of the present stem, when the actions do not 
 overlap, call for notice. Expressions of customary or generic 
 action are not in point. 
 
 II 56 1 : eri d' avrcbv iv tco Tredico bvrcov, jtoiv eg T?)r 
 naoaXiav yfjv e/idelv^^, ey.cnov vecbv eTiiTi/Mvv rfj IleAOTiovvrjocp 
 jraoeoy.evdCero. The ^oiv clause is added for accuracy of 
 temporal definition, though a phrase with m already serves 
 to designate the period within which the main action is 
 inserted. So 11 13 1. IV 14 1. Compare the similar usage 
 with an adverb in V 8 4: e/^ovXevero emyeioelv aicfviSkog, ttqIv 
 (hxe/Melv xovg 'Adipaiovg. 
 
 IV 78 5: 6 de y.eX.evoincov tojv dycoycbv, ttqIv ti TiXeov 
 ^voTfjvai TO xcoXmoov, eycooei ovdev bnoy/ov Sqojuco. Here the 
 
 1) Cobet (Hyperidis Orationes Duae. Lugd. Bat. 1877. pp. 59, 71) 
 without good reason regards the nqlv clause here and in II 13 1 as an 
 interpolation. Van Henverden also brackets .toh' . . . e'/MeTv in II 56 1. 
 
37 
 
 guides' order that Brasidas should proceed before any further 
 hindrance should arise, and the narration of his compliance 
 with the order are condensed into one statement. 
 
 On the other hand, forms from the present stem and 
 not the aorist appear in the principal clause when the 
 underlying conception is that of overlapping action. The 
 notion of limit may enter, and accordingly clauses introduced 
 by eojg, jnexQi, and fiexQi^ ov are to be added to those with 
 JiQiv and jiQorsQov fj. 
 
 In the first place, there are a few occurrences of jiqiv and 
 jTQOTeQOi' fj with the infinitive when the prior action has not 
 ceased at the time when the later action begins. This re- 
 lation is very clearly marked by other temporal expressions 
 in VIII 45 1 : iv de tovtco xai en jtqotsqov, jiqIv ig rijv 'PoSov 
 avTovg dvaorrjvai, rdds modooero. Without such modifiers 
 is VIII 45 5: rdq ^' dXXag noleig ecf)] ddty.dv, at eg 'A&t]- 
 vaiovg nQOTEQOv y dnooTrjvai dvii]Xovv, el //// y.al vvv Tooavra 
 xnl en jjleim vneg ocpoiv edeh'povoiv eo(peQeiv. The payments 
 continued up to the time of the revolt. VI 29 1: o (Y ev re 
 x(p JiaooiTi JTQog nj. pip'VfiaTa dneloyeTTO xal hoTfiog }]v ttqIv 
 exjikeTv xQtveoi'^ai There is ellipsis of yv in VI 11 1 : dv6)]Toy 
 (5' em TOiOVTOvg levai d>v xqaxyoag re /.«) xazaoxtjoei ng xal juy 
 xaroQ&cooag jui] ev rep S/uokp xal jiqIv eTzixeigfjoai eorai. Com- 
 pare IV 85 4: fifxeXg jiiev ydg ol Aaxednifiovioi olo/^ievoi re Jia^d 
 ^vjujudxovg, xal jiqIv eoyco dcpixeodai, rij yovv yvo'jpf] Ij^eir xal 
 (iovlof^ievoig eoeo&ai, x.r.L In the case of V 65 3: o de... ndhv 
 ro OTQdrevjLia xard rdyog jtoIv ^v/n/nei^ai dnijye, the succeeding 
 narrative (QQ 1) shows that the army had not yet on the 
 return march reached the camp, when an engagement took 
 place. So in VIII 42 1: enejilei ovv coojieo elye nobg rip' 2!v- 
 jiup' o 'Aorvoyog tcqIv exjivorog yeveod^ai, the context shows that 
 discovery was made during the approach of the ships. Compare 
 III 30 1. Further, in I 141.1: auro&ev d)] diavoipJijre // 
 vjianoveiv ttqiv n (ilal^ijvai fj, el jiole^atjoofiev, x. r. X., imaxoveiv 
 designates the alternative that involves no decisive action, but 
 rather a continuation of the present condition. The imperfect 
 in the isolated example of measure of space, with idea of 
 
38 
 
 limit, is especially noteworthy, VI U7 3: orudtoi dt jiqIv tiqoo- 
 jiieTiat ty. tov /.eificbvog iylyvovro avzoTg ovx eXaooov 1} nevxe 
 y.al n'y.oot. 
 
 In the next place, passages in which a finite verb follows 
 Tigiy or ,-7ooT£i)0)' //, and in which clauses of posterior limit 
 are introduced by eiog, fisygi, or ixexqi ov are to be examined. 
 With .TOO' and n[(j6Tfoor ij the main sentence is actually or virtu- 
 ally negative^) save in two cases (VII 39 2, 71 5), and here the 
 verbs are in the imperfect. With the other conjunctions the 
 leading verb is from the present stem as opposed to the 
 aorist, except in the following instances. The clause depends 
 upon an aorist participle in VII 2{) d: 6 de XaQiyJS]g TTSQijueivag 
 Iwg TO ycooi'ov e^erer/joE. There is ellipsis in a few passages. 
 I 28 5: eroTjuot. de elvcu xal More djuqioTegovg jueveiv xard yoiqav, 
 oziovbdg be. noirjoao&ai eojg av 1) Sixij yevi]rai^). IV 46 3: xal 
 avrovg ig ti]v vvjoov 01 orQari]yol ri]v IlTvyiav eg cpvXayJiv die- 
 xofuoav vnooJTOvdovg fie^Qi ov ^Adrjvat,e jiejuqp^cboiv. VII 16 1: 
 01 Se 'A^}]va7oi dxovoavreg avrfjg tov jiih' Nixiav ov jragelvoav Tj/g 
 dQXfjg, d?:A.' avTcp, ecog av eregoi ^wdgyoi^Teg aigedevreg dtplxcovTat, 
 Tcbv avTOv exel Svo TXQOoe'iXovTO. 
 
 The determination of the form as present or aorist is a 
 
 1) See p. 8. 
 
 2) This is the reading- of the manuscripts. There are tliree leading inter- 
 pretations: (1) fiEVEiv and jion)oaadai depend upon sroT/uot elvai, and wars 
 is pleonastic. To this it is objected that the redundant Sots is not found 
 with hoTfiog elsewhere. Herbst objects, further, that with this construction 
 there is no expression of what the Corcyraeans assent to in their alter- 
 native proposition, namely, dtxdCEadai. But Ecog ur ?) dUi] yEvtjTcu may 
 have final coloring. (2) dixukEodai is to be supplied after etoT^ioi slvai, 
 and wazE introduces the two following infinitives. This is difficult, and 
 gives no satisfactory result. (3) ds after a^rorddg is to be expunged, and 
 the MOTS clause made to depend upon jioi^aaadai. But mgzs should then 
 introduce the object of the treaty, whereas the recall of forces or their 
 maintenance in position appears throughout the chapter merely as con- 
 ditional. The trouble centres about coars, but it would be relieved by an 
 easy emendation to cog. Then the phi'ase p<cu w? has here, as elsewhere, 
 the force of affirmation even imder circumstances contrary to those just 
 described. Cf. Stahl on VIII 51 2. Moreover, this reading is confii'med 
 by the scholium (the entire note is probably made up not of two comments, 
 according to Stahl's arrangement, but of three): sroT/iwi dk Etrail dvTi tov 
 szoifiiog k'j(ovoiv, El firj exeTvo j]ov?MVTai, ajTorddg jioir]aao&ai. 
 
39 
 
 matter of accent in I 90 8: iavrdv S" ixekevev anooreXleiv (bg 
 idytora 6 Oe/uioroxXrjg ig Ti]v Aaxsdai/uova, riXXovg dk jigog eavrco 
 ilojuevovg TiQeo^eig /ld) svdvg ixjie^ujisiv, all' enioieiv i~i^XQ^ 
 TOGOVTOV ecog av to rer/og Ixavov aiQwoiv^) coore d7T0jua'/,^adfu ex 
 Tov dvayxaiordrov vipovg. mioieiv has been written, and tlie 
 following critical annotations are found. Bekker: "Nescio an 
 rectius scribatur emoxsiv.^^ Poppo: "Bekkero imoxsiv magis 
 placeret. At xaraoxelv habes capite proximo." Liddell and Scott 
 quote the passage under Ijiexoi (absolute); and again under 
 ETiioym (intransitive), but with the perispomeuon accent. The 
 structure of the sentence, however, is against the rendering 
 aliquantum subsistere (Betant), and calls for transitive force, 
 that TTQsofietg may be the object of this verb as well as of 
 ixjiE/xjieiv. Compare I 91 4: oT re ovv "x-idipaloi xohg TiQeojhig 
 coojieg ejieoTdX)] xatelyov. xaxaoyelv (I 91 3), to which Poppo 
 refers, is used of the act of seizure; not xaraoxsTv but /,«) 
 d(pelvai answers to emoyeiv here, and the aorist in that case 
 is justified by the negative. Since Thukydides uses emoxco 
 elsewhere (III -IS 4, compare I 129 H), as well as ejieyoi^ it 
 is certainly fair, if there is nothing to turn the scale, to write 
 emoxEiV' 
 
 The manuscripts read in Y 26 1: yeygacpe Se xal rama 
 6 avTog 0ovxvdid}]g "Ad^i/jvaiog e^rjg, d)g exaora eyevero, xard •&eQy] 
 xal ;f£<^coi'ag, /neyQi ov njv re dgyr]v xarejiavoav zcbv 'A{t}]vaicov 
 AaxeSaijLwvioi xal ol ^vjufiayoi. Classen substitutes eylyrsTo on 
 other ground: "eljjfg — ;^£</ia;i'ac: iibereinstimmend wie 2 1 
 [II 1 : yeygajirai dk e^tjg, cog exaoxa eyiyvexo, xaxd itegog xal 
 ysijucbva^, wesshalb ich auch hier eyiyvexo st. eyevexo geschrieben 
 habe, wie es der successiven Entwicklung der Begebenheiten 
 (eirjg) allein entspricht." This reading brings the passage 
 into harmony with the principle under consideration. 
 
 The only remaining case is II 31 3: eyevovxo 6e xal ulXai 
 voxegov ev xco TToXe/ucp xaxd exog exaoxov eoj^oXal A'&rjvalow ig 
 XYjV Meyagida xal InTreoiv xal navoxgaxta, i^ieygi ov Nioaia edXco 
 vTi Adrjvauav. In the other occurrences of xaxd exog exaoxov 
 
 1) See below p. 43 sq. 
 
40 
 
 in ThukydiJes, the verb is always of the present stem as ex- 
 pressing iterative action, and it can hardly be too bold to 
 write iylyrovTo here. 
 
 The historical present iu the main clause represents both 
 forms of concejjtion, for the aspect is clearly extensive in 
 HI ■_"•• 1: Tol'.; i(h' t-i tTj^ Tiolewg 'A^i]vaiovg Xav&dvovoi, jiolv 
 ())j Tfj A)jk(p eoyov, whereas it is aoristic in II 12 2: aiio- 
 Tiejiijiovoiv ovv avTov jtqIv dxovoai. 67 3: 6 de Jisio&elg nogevo- 
 jLih'Ovg avTovg did ri]g Oqax^iq em to jtAoiov co ejuekXov rov 'EDJj- 
 o.-T(»'Toy ntoauooeiv, jtoIv eoftaivEiv ^vXkaju^dvet. 82: dvaycogei 
 jia(/ ai'Torg txqIv Tijv ^v /.iftcnjd nav eWnv. An aorist participle 
 intervenes between the temporal clause and the historical 
 present in VI 4 2: jiqIv de dvaoTfjvai, ereoiv voiegov exarov Tj 
 avTovg olxtoai, Ildjiijiidov Jiefi^Kivieg ^eXivovvra xziCovoi. In YI 
 97 2 noir with the infinitive follows cp&dvei. dvafidg. There 
 is ellipsis in III 28 2: Ildyi^g 6' dvaoripag avrovg coore jiuj 
 ddixfjoai, xarajiderai eg TeveSov jueygi ov rdig \4.&i]vaioLg ti (5o|>;. 
 
 The perfect stem in a verb u])on which a clause of sub- 
 sequent action dei)ends is found twice. IV 16 2: eojreTodat 
 de avrdg fiexQi ov eTrnveldojoiv ol ex rojv "Aihp'cbv Aaxedcujiovkov 
 ngeo^etg. This is in harmony with the use of the perfect in 
 a clause of antecedent action, namely, to replace an aorist 
 when the extensive view is desired. (Compare VI.) IV 72 
 1 : d^ua de rfj eco ol Boionol TxaQrJGar, dia7'ei'0)]jiih>ot /Lih' xdl 
 jtqIv BQaoidav jTe/uii'm j'^ot]&eTr em to. Meyaga. The participle 
 must express time past with reference to the verb Tcagfjoav, 
 but the action of the main clause obviously overlaps upon 
 that of the subordinate. Compare \1 93 1. 
 
 Here for the first time a discussion of the value of future 
 forms as regards aspect is necessary. In clauses of antecedent 
 action no futures are found, and iu contemporaneous action 
 there is so much freedom of choice between the aoristic view 
 and the extensive that, after the influence of negatives is also 
 taken into account, nothing noteworthy remains. In connection 
 with clauses of subsequent action, however, a few occurrences 
 of future stems deserve attention. IV 30 4: xdl djua yevo- 
 fieroi Tienjiovai jiqcotov eg to ev rfj fjjreioo) OTQaTOJredov x/jQvxa 
 
41 
 
 7iQo>iaXov}XEVoi, el (iovloLVTO, fivev xivdvvov rovg iv rfj v/joco avdgag 
 ocpioi TO. re onka xal ofpdg avrovg xekeveiv Jiagadovvai, i(p oj 
 qwXaxfj jfj fiexQia xt]Qi)oovTai, swg (iv ri Jieoi rov Jikeovog ^v[.ifia&f]. 
 It has already been seeu that the extensive view is used for 
 a verb upon which a clause introduced by £co?, 'until', depends; 
 and that the form j}]ot)oovrai should give this view is in ac- 
 cordance with the theory upheld by Blass^). In IV 97 2 the 
 sentence is negative, and there is ellipsis in YII 83 2: i^ie^Qi 
 ^' ov av ra yQ))j.iaTa anododf], avdoag do'ioeiv 'Adijvaimv 6jiii]Qovg, 
 eva xara rdXavTov. With tiqvv^ 'before', the aoristic conception 
 in the main clause is expected, and two passages show an 
 aorist participle accompanying the future verb, Y 9 (>: Eyco 
 j-iev eywv rovg fier' i/iiavTov xal cf&daag, ijv dvvcoi^im, TToooneaou- 
 juai dgojiw) xard fxeoov ro aroa.Tevi.ia. YlII 12 1: Xeyoiv on, 
 (p&)]oovrai je JtXevoavreg tioIv Tijr rcbv vewv ^vfiq^ogdv Xiovg 
 aiodeo&ai. Finally, one sentence has a future form to which 
 aoristic force is attributed by Blass, II 53 3: ad)]/.ov voi^d'Qayv 
 el Jiolr ETi auTO eXd^elv diaqdaQ^'perai. 
 
 2. Subordinate clause. 
 As regards aspect, the verb in a clause of pure sub- 
 sequence is not restricted by the temporal relations. By 
 virtue of their negative character, however, noiv and TXQoxeQov 
 1] naturally take the aorist^). Thukydides shows sixteen 
 occurrences of the present, all with jto'lv. Of these, three 
 (lY 2 1. 67 3. 125 4) have ehai, and two others (I 39 1. 
 ^ 60 6) are of general action in the present sphere. Some 
 have the notion of will or effort, as the imperfect with a 
 negative has, while in a few cases the tendency of the verb 
 may be influential (II 67 3. Ill 24 2. 64 3. lY 14 1 {avd- 
 yeo&ai). V 84 3. YI 29 1 {exTikelv). 61 1 {ex:iXeiv). In Y 
 41 3: exeXevov d' ol Aaxedaifiovioi, jiqIv reXog ri am&v exeiv, 
 eg TO "Agyog tiqcotov ejtavaycoQj'joavrag avjovg dei^ai rco Tihj&ei, 
 xal ip' doeoxovra /;, rjxeiv eg rd "Yaxivdia rovg oqxovg Tioojoojue- 
 vovg, the character of the expression Ttlog eyeiv as a peri- 
 
 1) Rh. Mas. XLVII (1892) p. 269 sgq. 
 
 2) Gildersleeve .4. ./. P II (1881) p. 466 S(/'^ 
 
42 
 
 phrasis may account for the tense. In Y 10 3, a negative 
 in the principal clause gives to jtqiv the force of 'until', and 
 the limit falls within the course of the action expressed by 
 the infinitive. Twice a clause with ov ttoi'v occurs, and the 
 main action in reality falls within the action denoted by the 
 present infinitive with .loiv. I 39 2: ovtoi d' ov :roiv tto/uoo- 
 y.eTr to yowior, aXX" ejieidi] fjyrjomno fjtia^ ov neoi6ii>eo&ai, tote 
 y.al TO Fv.-Toe.-rh tz/s ^I'y.i]? migfoyoiTo. The offers of arbitration 
 by the Corcyraeans took place after the siege had been begun 
 and while it was still in progress. So I 68 "2. 
 
 "With .-7g<V and rr^oT^^oi- ij followed by a finite verb, and 
 with the other conjunctions of limit, the subordinate verb is 
 aoristic unless the limit of the prior action is within the 
 posterior. In the latter case, as in overlapping antecedent 
 action, the spectator may be indicated. The number of 
 instances of limit falling within an action is small. 
 
 a) AYith the imperfect indicative. I 30 3: Tovg twv Ko- 
 Qiv&kov ^vftjud/ovg ejitTt/Joi'Teg ecpdeigov, ^neygi ov Kogiv&ioi 
 TieQuox'Ti Tcp deqei Treuif^m'Teg rat'? xal otqcxtiulv, ijiel ocpibv ol 
 ^vjiijiiayoi enovovv, eoTQaTOTiedevovTO em "Axtico. Here eorgaTO- 
 jiedevovTo means not 'encamped', but 'were in camp', in 
 accordance with the constant usage of Thukydides. Com- 
 pare Stahl on TI 42 1. The aorist is the form employed for 
 the meaning 'to pitch a camp'; and despite Rutherford's 
 change of present and imperfect to perfect and pluperfect 
 when the force 'to be in camp' is required, the perfect has 
 a force of its own. It occurs only in the participle (lY 54 1. 
 YI 88 3), and is used with reference to a past stay in camp, 
 for in both passages the narrative immediately following shows 
 the camp broken up. 
 
 I 118 2: ol de Aaxedaijiiovioi alo&oj^ievoi ovte IxayXvov el jurj 
 em j^oayv, i)ovyat,6v Te to ttAeov tov ygovov, . . . txqIv Si] f] dvvajLug 
 Tcov 'Ad^i^vakov oacpcbg JjgeTo y.al Tijg ^vju/uayiag avTcov ^jjirovro. 
 
 Y 82 3: ol Se Aaxedaijuonoi, e'ojg jciev avTOvg ^iieTeTiejiijiovTO 
 ol (pikoi, ovy. ))/Mov ex Tileiovog, a.va^a?MjLievoi de Tag yvjuvo- 
 naidiag e^o)]dovv. The difficulty is to assign to ey. Trkeiovog 
 its usual force when used temporally by Thukydides, that is, 
 
43 
 
 'for a long time before', iatn dudum (IV 42 3. 103 4. 
 YIII 88. 91 1); and to obtain the needed temporal oppo- 
 sition for the jjiev and de clauses. Kriiger suspects the 
 genuineness of the tradition. Stahl and van Herwerden mark 
 a lacuna after ex nhiovog. Classen suggests that ix jtXslovog 
 may mean 'with great zeal'. The citation of the sentence in 
 the discussion of posterior limit has already suggested a way 
 of escape, ecog is not 'while' but 'until', ecog jliev marks the 
 limit of the extended action denoted by ovx fjldov ex jiMovog, 
 and dva/3fA6fievoi de introduces the action that follows. 
 
 YIII 24 5: xal ovd" amip' t)]v anooxaoiv, el rovro doxovoi 
 Jiaod TO doq^akeoxeQOv nqd^ai, Jigoregov hokfxtioav Jioirjoaod'm y 
 jnerd jiokXwv re xdya&Mv $vjLijLid/cov ejuekkoi' ^vyxivdwevoeiv xal 
 Tovg "A^^yjvalovg fjo&dvovro ovS' avrovg dvxileyovrag en fierd ri]v 
 ZixeXix-ip' ^vjiKpogdv c5? ov jidvv novi^Qa oq?c~jv (ie(iakog zd Tcgdy- 
 /Liara eh]. Classen adopts from the Yaticanus /Lie&'' cov after 
 ^vjujudxojv, and makes fjo&dvovro coordinate with ovd' erol- 
 juyoav. Apart from the question of historical basis, which 
 Stahl shows to be in favor of the former text, the objection 
 to fjoddvoino^) overlooks the nice correspondence, under 
 antecedence and subsequence, between this passage and such 
 a sentence as III 108 3: ejTavaycoQovvreg de, ojg koQcov to nkeov 
 vevixrji^ievov xal ol dXkoi 'Axagvdveg ocpioi nQooexeivTo , yaXenoig 
 dieocpt,ovTO eg rdg "'OXriag. 
 
 b) With the present optative. Ill 102 7: fjovya'Ce T(p 
 OTQaTcp Jiegl TOvg j^coQovg rovrovg, eojg TOig 'A/.ijTQaxid)Taig exoTga- 
 Tevoa^uevoig negl to "Agyog Seoi fiot]d^eTv. 
 
 One other passage showing the present stem is to be 
 considered. I 90 3: eavrov S' exeXevev djiooxeXXeiv cog Tdyiaxa 
 6 OejLUOTOxXfjg eg Ti]v Aaxedai/uova, ciXXovg de ngog eavrco eXo- 
 juevovg ngeofieig jli)] ev^g exnefxneiv, dX)i emoyeiv /JieyQi tooovtov 
 ecog dv to Teiyog Ixavov aigcooiv cooTe djiofxayeod^ai ex tov dvay- 
 xaioTcjLTov vif'ovg. The present cuocooiv is the reading of the 
 manuscripts It has been justly criticized, however, as in- 
 appropriate to the clause as an expression of posterior limit, 
 
 1) By a strange lapse Goodhart cites III 2 1 in support of this 
 imperfect with jtootfqov ?j. 
 
44 
 
 for the end is definitely stated to be the least possible height 
 for defense. Bekker's correction to aQcootv has therefore been 
 widely adopted. Shilleto, defending- the manuscript reading, 
 says : "it must first be noticed that whereas we have now a 
 marked distinction between while and until^ the Greeks and 
 Latins had the same words ecog, e'ore, jt^teygi or ^ueyoiJieg, dum, 
 donee, the tenses alone fixing the meaning of the particles." 
 He translates the sentence, "till we are raising our wall to 
 the barest possible height to fight from". Arnold, before 
 him, says: "It [eojg av aiQCJOiy/ may signify, 'all the time that 
 they were raising', but then iiityot tooovtov seems to be less 
 suitable than jiavTa ror yooror. I have therefore adopted 
 Bekker's correction."' But tooovtov is a word not of position 
 but of extent, and j-dyQi means 'during' as well as 'until' ^). 
 It seems possible to keep the manuscript reading, and the 
 meaning, then, is doubtless 'to detain them for so long a 
 time as they were raising the wall to a sufficient height'. 
 Either interpretation of hog requires the extensive view in 
 the verb upon which the clause depends^). 
 
 The historical present in clauses of posterior limit is found 
 three times, and the context shows that in each case the verb 
 must be interpreted as aoristic. I 132 5: dir ovd' cog ovde 
 TQJv Eilo'non' fop'VTaTg tioi Jiiorevoavreg y^icooav vecoTsgov ti 
 noieiv eg amov, . , . jiqiv ye dt] avrolg . . . dvijQ 'AqyiXiog . . . 
 jj.rjvvT'i'ig yiyvexai . . . xal . . . Ivei rag emoTokag. . . . tots di] ol ecpogoi 
 det^mnog alnov to. ygdju^uaTa judkkov jLiev emorevoav. YII 39 2: 
 ijil Tiokv diiiyov rfjg y/uegag Jieigcojuevoi dXkrjXcov, nglv Si] 'Agi- 
 GTOiv . . . nei&ei rovg ocperegovg lov vauuxov agyovrag. . . . xai ol 
 juev netod evreg e'jiejuyjav dyye?Mv. 81 4: ov ngovycoget ^dlXov i] 
 eg judyi]v ^vveidooerco, ecog cvdiargi^wv xvxAovTai re VJi axncbv 
 xal ev noXXcp dogvfiq) avrog re xal ol /.lex amov [A&ijvaToi] yoav. 
 dveiX}]&evTeg ydg eg ti yojgiov co xvxXo) fxev reiyiov Jiegifjv, . . . 
 e^dXXovTo Ttegioradov. 
 
 1) Cf. V 47 6. — Herodotos uses fie/^gi oaov of coextension, VIII 3: 
 ijiiazd/iiEvoi wv avio rovro ovx avxExtivov cdX eixov, l-i^XQ'' ooov y.aQxa 
 idiovTO avTOJv, w, dieds^ar. Cf. Hdt. VIII 84. 
 
 2) See p. 38 sq. 
 
45 
 
 The pei'fect stem does not occur in a clause of subsequent 
 limit or of pure subsequence. 
 
 § 4. CONJUNCTIONS. 
 A. Antecedence. 
 1) enei, eneid)], lm]v, ETceiddv. 
 
 That these conjunctions usually denote antecedence is 
 undisputed. In answer to the question whether Thukydides 
 uses them to express pure contemporaneity, the discussion of 
 one or two passages must suffice. 
 
 II 34 2—3 : rd jLiev dord jTQOTidevTcu rwv djioyevojuevwv jzqo- 
 TQira oxi]V})v JiODJoavreg, xal ijiiq)€Qei rc5 aurov e'yMOTog ijv ti 
 ^ovh]Tai. ETXEiddv Sk f] exqpoqd fj, Mgvaxag xvnaQiooivag dyovoiv 
 ajua^ai. Although it has already been seen that the combi- 
 nation of sjiei and its derivatives with the present stem de- 
 notes action partially, indeed, contemporaneous, but also 
 antecedent, nevertheless, since in the present passage the 
 predicates of the two clauses concern the same action, ejieiddv 
 might seem a precise equivalent of oTav. Marchant takes 
 this view, and writes: ^ejTeiSdv — used here like oTar. Con- 
 trast C below and c. 72 .3." But in the case of Thukydides, 
 it is not a vain inquiry to ask why this word was chosen 
 and not that; and accordingly, here it may be asked why 
 Thukydides should choose not 'when' but 'after. Has not 
 iTzeiddv some force that otuv would not convey? Slight though 
 it may be, ground seems to exist in an opposition. There 
 is a contrast between events on one side of a fixed point and 
 those on the other; events before are balanced with events 
 after. This opposition finds expression in the particles juev 
 and de, and the one force represented in jigoTQira must find 
 its counterweight in ETieiddv. 
 
 A parallel passage is found in YII 23 2: xat iy. fxh rov 
 71QCOTOV dkovrog yalzTioK ol dv&QCOJiot ... eg to OTQaxonedov 
 e^exo/Lil^ovTo' rcov ydg 2vQaxookov xaTg iv rep f^ieydlco Xif^iht vavol 
 XQUTOvvTCOv Tpj vavjua/Ja imo TQi/jQOvg /.udg xal ev Ji^eovorjg excs- 
 dicoKOVTO' ijiEidr] dk xd dvo jsixiojuara fiUoxexo, ev rovrw xal ol 
 2vQax6oioi hvy/avov i]Sr] vixcojuevoi, xal ol i^ avrcbv (pevyovreig 
 
46 
 
 ^aov TiaQETiXEvoav. Classen says: ^^inFid)) — i))doy.ero, genaue 
 Angal)e tier gleiclizeitigen (Imperf.) Yorgiinge an zwei ver- 
 scliiedeneu Puukten: 'zu der Zeit, da die beiden kleineren 
 Forts in die Hande der syrak. Landtruppen fielen, hatte sich 
 das Kriegsgliick bei der Flotte geandert'." But here again 
 tliere is opposition, the contrast between events following a 
 turn of fortune and events previous. 
 2) cog. 
 
 The Protean character of wg somewhat resists analysis. 
 Apart from uses foreign to the present subject, cog, beginning 
 as a modal ])article, assumes temporal and causal significance. 
 The beginnings of temporal force may be felt in a con- 
 struction occasionally found, in which cog comparative has xai 
 as correlative; for example, YIII 27 5: cbg dh. e'jieioe, xai 
 edQuoe rnmn. From such forms the development of a tem- 
 poral clause is easy. Compare the use of ut in Latin. This 
 temporal clause is naturally of antecedent action, and cJjg be- 
 eomes a less precise synonym of ensL Curschmann^) goes too 
 far when he says: "Ubi nihil refert diserte exprimi et efferri, 
 quod enuntiato temporali expressum sit aut eodem tempore 
 aut prius factum esse, ubi vinculo laxo enuntiata et temporale 
 et primarium coniungenda sunt, ibi cog Invenimus." 
 
 In order to maintain the position that Thukydides uses 
 <w? as a conjunction oi antecedence, i- few passages require 
 notice. Wisen^), after citing instances in which a phrase with 
 Ev serves as correlative antecedent to c6g, and sentences in 
 which ovTco d)'] and tote d}) serve as resumptive particles after 
 <wg, states that the actions of main and subordinate clause 
 take place at one and the same time. But a phrase with h 
 may be used with a clause of overlapping antecedent action, 
 and ovro) St] and tote d)'] are employed by Thukydides after 
 temporal clauses other than those introduced by cbg, only with 
 the conjunction EJiEidi)^). Beyond the passages designated by 
 
 1) Progr. Giesseu 1876. p. 3. 
 
 2) De vi et usu particulae cog apud Thucydidem. Hauniae 1862. p. 31. 
 
 3) Cf. also Stahl's emendation of IV 30 2. 
 
47 
 
 Wisen, there are three sentences in which cog has been re- 
 garded as denoting contemporaneity. 
 
 I 19: y.al lyevero avToTg ig Tovde tov noXet^wv i) Idia Tia- 
 gaoxev)) juelCajr i) cog to. xgdriord nore jiiercx axgaicpvovg rrjg 
 ^vfijimxiag ijvd)]oar. Here the question is complicated by the 
 uncertainty hovering about other parts of the sentence. Classen 
 translates, "und so bildete sich die Kriegsmacht eines jeden 
 der beiden Staaten (y Idia mxQaoy,.) zu einer grosseren aus, 
 als da sie mit der gesammten Bundesgenossenschaft vereint 
 auf der Hohe ihrer Macht gestanden hatten." Herbst^) justly 
 objects to this extravagant statement, and explains that "beide 
 sich jetzt ein jeder auf seine weise mit krankung seiner 
 bisherigen bundesgenossen, so zu sagen eine eigne hausmacht 
 beschafft haben. Diese hausmacht des einzelnen jetzt ist 
 grosser als die grosste macht, die der einzelne je friiher in 
 verbindung mit seiner noch ungeschadigten bundesgenossen- 
 schaft besessen hatte." The contrast always implicit in Idiog 
 seems here to be between the central power in its indepen- 
 dence at the opening of the war and in its union with allies 
 previously. These two ideas are compared by the words 
 jiieiCcoy fj — 'their own resources for the war were greater 
 than their own together with those of the allies at the most 
 flourishing period of the undisturbed alliance'. cog thus 
 is modal. 
 
 IV 117 2: Tovg ydg Si] ai'dgag jisqI nXeiovog etioiovvto ko- 
 jLuoao&ai, cog hi BgaoiSag '}']VTViei, xal e'jiielXov em jiieVQov 
 y^cJL>Qi]oavTog cimov xcd dvj'maXa xaraonjoavTog , tcov jusv ore- 
 geo&ai, roig S' ex tov I'oov djiivvojueroi xivdvveveiv, xal xQampeiv^). 
 The early exegesis of Herbst^ , which is adopted by Classen, 
 makes cog comparative; hence if this interpretation were 
 accepted, the passage would not bear upon the present question. 
 Herbst's later judgment*) is in favor of the reading given by 
 
 1) Fhilol. XXXVIII (1879) p. 535 sq. 
 
 2) An extended discussion of various interpretations may bo found in 
 Studien zu T/iucydides — Hampke. Progr. Lyck 1876. -p. 24 sqq. 
 
 3) Philol. XVI (1860) p. 313 sqq. 
 
 4) Zu Thuk. Erkldrungeii u. Wiederlierstellungen, Leipzig 1892. p. 102. 
 
48 
 
 the scholiast on Arist. Ran. 478: e'cog o re Boaoidag eiTv^^i. 
 The explanation of Golisch and Hampke makes (bg practi- 
 cally equivalent to k'(og; and Stahl, on the unstable support 
 of I 19 and Y '20 1, translates by quando. The whole passage, 
 however, would become clearer if cog were regarded as a 
 conjunction of antecedence and interpreted as elsewhere in 
 combination with the imperfect, 'after they saw that'. Com- 
 pare IV 79 2: ^y. yag T)~]g IJeXojiovv/jaov, cog to. row 'Adrjvaicov 
 )]VTvyei, dehfUTeg oi re em 0Quy.i]g uqeoraneg 'Adi]vaicov y.al 
 IJegdiy.y.ag e^ijyayov rov orgarov. The comparative force of 
 jTegl Txleovog is met by the comparative force of 'after'. The 
 following words, y.al e'jueU.ov . . . y.ai y.garrjoeiv, may be co- 
 ordinated with ijVTv/ei; the good fortune of Brasidas means 
 victory for the Lakedaimoniaus. Between the verb eiLteA/.oy 
 and its complement xganjoeiv is a long insertion, subordinate 
 in construction, rear aev is perhaps a copyist's correction of 
 rolg ith\ for the jiiev and Se clauses may well represent re- 
 spectively the Athenian and Lakedaimonian sides established 
 by Brasidas as dvrL-ia/M. The passage resembles lY 73 4. 
 The meaning, then, would be somewhat as follows : 'For they 
 were interesting themselves more in the recovery of their 
 men after they saw that Brasidas was still fortunate, and that 
 if his success increased and he balanced loss on the Athenian 
 side with defense risked without disadvantage on their own, 
 they were going to conquer.' 
 
 Y "20 1: avrai at ojiovdal eyevovro . . . avrodey.a ercbv biel- 
 ■dovrcov y.al y^uegcbv o/Jyow Jiageveyy.ovooyv i) cog ro Tigonov f} 
 eo(ioh) f] eg rrjv 'Arrixijv y.al fj dgyj] rov noXefiov rovde eyevero. 
 The explanation of Bauer, that the comparative particle ?;' 
 follows voregov latent in the general idea of the sentence is 
 unsatisfactory. There might, indeed, be some such irregularity 
 if the following words were consistent with the construction 
 to which the shift was made. But one would not then ex- 
 pect the word cbg. (Compare I 51 4. 60 3. YI 4 2.) It 
 seems better to reject /y'^), and take cog as in lY 90 3. 
 
 1) Cf. Morris on I 13 3. 
 
49 
 
 It may be considered, then, that cog, for Thukydides at 
 least, is a conjunction of antecedence. 
 
 3) VOTSQOV i]. 
 
 voregov ij is indisputably used of antecedent action, and 
 that only. 
 
 4) d(p' ov. 
 
 A preposition of motion from with the relative pronoun 
 may serve as a conjunction of antecedence. It is apparent 
 that this combination means 'after' only in the sense of 'from 
 the time when'. 
 
 5) rf] jTQcoTtj fjjLieQa fj. 
 
 ejietddv ruyjora finds a virtual equivalent in rfj jrocorfj 
 fifXEQU fi civ (YI 23 2). 
 
 6) cjodxtg. 
 
 oodxig in its single appearance is used of antecedent action, 
 where perhaps sjieidi) with the optative might have stood, 
 had not this conjunction already been used to introduce the 
 clause in which the Sody.ig clause is incorporated, YII 18 3: 
 ejisidi] .dk 01 lA&rjvaToi raig rQidxovTa vavolv e$ 'Agyovg ogjuco/iievoi 
 'EnidavQov re ri xal IlQaoicbv xal u}Jm id/jcooav xal ex IIvXov 
 ajLta ehjorevov xat, oodxig JieQi rov diacpogal yevoivro jcbv xmd 
 rdg onovbdg di^icpio(i}]Tovf,ievon>, eg Sixag jTooxa?Mi\uei'cov rcbv Aaxe- 
 daifiovicov ovx f'l&eloi' eTnxQeneLV, rore di], x. t. /I. 
 
 B. Contemporaneity. 
 
 1) a ore, ojiore, orav, onorav. 
 The actions correlated by these conjunctions are in the 
 main contemporaneous. Occasionally, however, the writer 
 unites under the general conception of contemporaneity actions 
 strictly sequent. So in a few passages fjl&ov and e^ijX&ov 
 occur where the action of the independent verb is really sub- 
 sequent to the actual coming or going out (J]kdov — I 73 4. 
 VI 46 3. VIII 73 2. i^i]l&ov — IV 5 1. VII 4 6). Verbs 
 of asking and answering are connected by cmore in I 90 5: 
 xcxl ojiore rig avrbv egoiro rcbv ev rekei ovrcov o ri ovx eneQierai 
 em ro xoivov, ecprj rovg ^vjUTrgeo^eig cxvrxueveiv. Other instances 
 are I 91 3: ecfofielro yuQ ju)) oi AuxedaijLiovioi ocpdg, (more 
 
 4 
 
50 
 
 oof/(Os ny.ovoFiav, orxhi nqxontv. IV 111 1: snarov de 7xe)aaaTdq 
 noont:fiJjei, o:i(o:;, oTiore jtv/ju Tiveg avoiy&elev xal to oijjuelov 
 nodeh] o ^vvexeiro, TiQcinoi eodgdjuoisv. VII 50 3: Jigodjiov d)g 
 edvvdt'To dthj/MTfaa fx.-rlovr h. tov mQaroTiedov Tidoi, xal Jiaga- 
 oxevdaanDat^), orav rig o)j/(ip')j. Oiice ote is said to be used for 
 'since', I 13 3: eDj (Y ion fidXima TQiaxdain eg rip' releimjv 
 rovde tov noUi^iov otf 'AfAeivoxh~]g ZajLuoig ijXOev. It is to be 
 supposed, however, that ore is still 'when' in the thought of 
 the writer, just as lavr)] in 13 4 shows position, not measure 
 from a i)oint 
 
 b. fp'ixa. 
 
 Contemporaneous actions in the relation of insertion are 
 In'ouo'ht together by this conjunction in the single passage 
 where it occurs (VII 73 3). 
 
 c. Forms of 6g. 
 
 It is evident that dative forms with or without h', the 
 accusative denoting extent, and the local-temporal genitive ov 
 express contemporaneity; but here, too, employment with 
 actions strictly sequent is possible. Compare VII 29 4: to 
 ydo yh'og tmv Ogaxcdr, oiiioTn Toig fidXaoTcx tov /^ag/^agixov, iv 
 CO dr dagoijor], fpovixdiTUTov eoTi. 
 
 2) Conjunctions of limit. 
 
 a. ecog. 
 
 In the expression of extent uog is a conjunction of pure 
 contemporaneity. It may be used of action within which the 
 main action is inserted — 'while'; or of action coincident in 
 extent with the main action — 'so long as'. 
 
 b. i^dygi, ix^y^gi ov. 
 
 f-ieygi is twice (III 10 4. 9S 1) used of coincident ex- 
 tension, l^'-^xgi' /*^^' oi- is given by some manuscripts, f^eygi 
 jLih ovv by B and F, in II 21 1: "A&^jvaioi de, /-leygi [^dev ov 
 jiegl 'EAEvoiva xal to Ogidoiov Jiedlov 6 OTguTog fjv, xai Tiva 
 
 1) The emeudatiou of Abresch, jTaoer>>isvdadai for Ttaoaaxsvaaaodai, 
 seems imuecessary, for there may be a "Ready!" before the "Go!" See, 
 among other discussions of the passage, Junghahn Studien zu Thukydides. 
 Seue Folge. Berlin 1886. p. 54 sqi/. 
 
51 
 
 ilmda elyov. If this reading is correct, the passage is the 
 only one in which fir/jM ov does not mean 'until'. 
 
 c. Forms of ooog. 
 
 sv oocp shares the meanings of ewg — 'while', of partial 
 coincidence (VIII 61 1); and 'so long as' (III 81 2. VIII 87 
 1; 4). By confusion of ideas it is used with constructions 
 implying the force 'until' in the conjunction (III 28 1. 52 3). 
 The accusative oaov needs no remark^). 
 
 €. Siibsequeuce. 
 
 1) TTQcr, jiooTegov fj. 
 
 The leading conjunction is jToty. tiqoxsoov ij of pure sub- 
 sequence appears four times (I 69 5. II 40 2. VI 58 1. 
 Vm 45 5). 
 
 2) Conjunctions of limit. 
 
 a. JTQIV. 
 
 With a finite verb jtqo' is used of subsequent action over- 
 lapping, in the meaning 'until'. 
 
 b. I'ojg. 
 
 C. ,ll£XQl, /^(^XQ'' ^^' 
 
 d. eg o. 
 
 eg 6 is read in V 66 2: judhora Si] Aaxedaijiiovioi eg o 
 ijnejuvrjiTO ev rovrcp rco xaiocp e^e7Tldyi]oav^). In such connection 
 it is amx^ eigij/iievor, and is probably an expression less of 
 time than of degree — not 'so far back as they remembered', 
 but 'so far as they remembered'. The parallel passage, II 
 8 3, shows d(p' ov. 
 
 § 5. RELATIVE POSITION OF CLAUSES. 
 
 In the arrangement of the period, the subordinate clause 
 may precede or follow the main clause, or may be inserted 
 between parts of the main clause. Often the principal member 
 of the sentence is represented by its subject, or some word 
 
 1) Cf. oaov ill II 23 3: yoovor s/tfieivarrsg sr t/j 'AxxiyS] oaov eiyov to. 
 ijnxrjdeia dvs/o'jgijoay. So ov in III 1 2. 
 
 2) The correctness of the passage has been questioned. See Classen, 
 Stahl, et al. 
 
 4* 
 
52 
 
 or phrase, placed in advance of a temporal clause that pre- 
 cedes the j)redicate upon which it depends. Sometimes a 
 colorless auxiliai*y verb of the main clause thus precedes, 
 while the complement follows the subordinate clause. When 
 two or more clauses are found in paratactic construction with 
 the same leading verb, this verb often takes an inclusive 
 position before or after the clauses With constructions more 
 or less elliptical in character, as with comparatives, akXoK re 
 xai, jTo/Mi) i^kVaXov hi, and the like, the orderly . march of 
 clauses is broken. In the temporal use of relative forms with 
 expressed antecedent, Thukydides shows the antecedent al- 
 ways in advance unless it is incorporated in the relative 
 clause. Otherwise, deviation from the normal arrangement 
 by which a prior clause precedes the main clause, a posterior 
 follows, is in general a mark of a developed style. 
 
 A. Prior clsiuses. 
 
 Clauses of antecedence that follow the independent pre- 
 dicate are chiefly of a few types easily discerned. 
 
 It is natural that when the clause marks an anterior 
 limit, 'since', 'after', the action to be limited should be ex- 
 pressed first. Accordingly the temporal clause follows that 
 which it modifies in III 70 1: ol yao KeQxvQoioi eoTaoiaCov 
 i:Teidi] ol ar/jidXcoroi fjAdov avxoig. In II 73 3: om sr tm ttoo 
 xov XQOVM . . . d(p' ov iv^u^uayoi iyevojue&a, 'A&ijvaioi qaoiv ev. 
 ovdevl v/uag JiQoeod^ai ddixovjuevovg, since the clause introduced 
 by d(p' ov determines a limit of ev tc5 ngb xov yQovo), it fol- 
 lows this immediately. Compare YIII 81 1 : OQaov^ovXog, del ye 
 rfjg avTrjg yvc6jui]g ixojuevog, ejieidt] jueTeoT)]os rd TigdyjLiaTa, coore 
 y.ardyeiv 'AXxifiiud}]v. So regularly with an expression of the 
 lapse of time, or with an ordinal numeral: ejieidri — 16 3; 
 5. m 68 5. VIII 68 4. ^ti^jv — V 47 6. d(f' ov — I 14 
 3. 18 1. n 8 3. 102 6. Ill 116 2. wg — lY 90 3. V 
 20 1^). The enel clause is pushed forward by the tiqiv clause 
 
 1) Cf. [J. 48. 
 
53 
 
 in YI 2 5 : xal to. xgdnoTn t//s yrjg cpxijoav exovreg, md dii- 
 /hjoar, krr] eyyvg rqiaxooia jtqIv "EXXip'ag eg Zixeliav U&m'^). 
 
 The temporal clause sometimes stands between a parti- 
 ciple and the finite verb of the sentence when ])resumably 
 the action of the participle follows that of the subordinate 
 clause, as III 24 3: ol d' Ix rfjg JioXewg UXaraii'jg . . . xi'jovxa 
 ixjisjiiyjavreg, ejiel fj/aega eyevero, eoTtevSovro dvaigeoiv roig vexQoTg, 
 and clearly in I 49 1 : ^v^ufiei^avTsg de, ejieidij to. oi]/u£ia exare- 
 Qoig fJQ&r], havjiidyovv. Compare Y 59 2. YII 34 4. The 
 clause is inserted between i^ieXlco and the infinitive in YII 50 
 4: xal fieXX6)TCOv avrcbv, eneidr] hoI/Lia tjv, anonXeXv fj oeX/jvJ] 
 ixXemei. Expressions like ijteid)) xal rovg 'Adrpaiovg eldov 
 regularly come last; and this is the case even when the 
 temporal element is brought out by the adverb ev&vg in YII 
 69 1 : xal ol juev ribv HvQcxxooioiv oTQan^yol xal rvXiTinog 
 TOiavra xal ainol roig ocpexsQOig oxQajidiTaig jiaQaxeXevodjuevoi 
 d}TejTX}]govv rdg ?'«?? £in%g mEiS)) xal rovg 'A&rjvaiovg fio&dvovro. 
 In other passages, elliptical construction, or parallel structure 
 in the temporal modification of the main verb, or epexegeti- 
 cal and causal character of the subordinate clause account for 
 an arrangement by which the temporal clause follows the verb 
 upon which it depends. ETrei — I 30 3. II 51 5^). Eneidi) — 
 I 137 4. II 3 1 86 4. UI 16 4. 31 2. 33 3. lY 73 4. 
 132 2. 133 1. Y 22 2. 29 1. 50 4. YII 18 1. 46. 69 2. 
 80 1. YHI 31 1. 52. 82 2. 90 1. 91 1. 105 3. ok — 
 
 I 61 1. ni 69 2. lY 96 5. 117 2«). Y 59 2. 61 5. 72 1. 
 [116 3^).] YI 102 2. 103 3. YII 23 1. YIII 25 4. 94 1. 
 100 4. 
 
 With clauses of contemporaneity, modification of position 
 is easier. Designation of the limits within which an action 
 falls or of the measure of its extent is not infrequently placed 
 after the verb expressing the action, eog — I 58 2. 90 3. 
 
 II 72 3. ni 82 2. rS^ 67 4. Y 60 3 75 5. YI 49 1; 2. 
 
 1) On V 18 5 see p. 17 sy. 
 
 2) See p. 26 sq. 
 
 3) See }>. 47 sy. 
 
 4) See p 65, footnote 12. 
 
54 
 
 Vn 47 3. 71 4. VIII 40 3. 78. ooov — I 137 4. (See 
 also IV 39 1.) Compare h ooco — IE 52 3. VIII 87 4. 
 Epexegetical clauses, clauses that follow (oorreo or ij, and many 
 other cases show postposition of the subordinate member of 
 tlie sentence. fiviyM — \I1 73 3. ore — I 8 2. 9 2. II 5)5 
 2. 98 2. lU 13 1. 34 1. 55 1; 3. 5G 4. 68 1. 96 3. 
 113 1. IV 34 1. 92 (;. 116 2. 133 3. Y 16 3. 30 2. 
 41 2. VI 31 1. 46 3. 86 5. VIII 73 2. 76 4. 78. 86 3. 
 108 4. 6.-T6re — II 13 7. 18 2. 51 4. Ill 22 3. IV 23 2. 
 26 7. V 16 1. VI 66 1. 97 5. VII 48 1. 70 3. char ~ 
 
 I 36 1. 141 6. 142 1. 144 2. H 11 6. Ill 39 7. 56 7. 
 IV 64 3. Y 47 7. 98. VI 79 1. YII 50 3. Smkav — 
 
 II 84 2. IV 21 2. Y 7 5. YIII 67 3. ir f5 — I 39 3; 3. 
 II 11 6. Ill 39 2. [84 2.] lY 17 2. 
 
 B. Posterior clauses. 
 
 In one case of inverse insertion, the temporal clause pre- 
 cedes the main verb, II 102 5: jtqIv av evgcbv ev ravif] rfj 
 yon)a y.aToiyJot]Tai, fjrig- ore exTeire tijv fUjTsga ju)] jtoj imb f]Xioir 
 koQCLTO jiiijde yv] tjv. 
 
 As regards subsequent action, jtqiv and jiqotsqov fj with 
 the infinitive seem not to be given the full rights of a clause, 
 for the phrase more often precedes than follows the principal 
 verb. Sometimes the position of the jTgti> clause is influenced 
 by the fact that it is parallel with another temporal modifier. 
 ov TTQiv precedes the leading verb, I 39 2. 68 2. The other 
 passages in which the verb that expresses the subsequent 
 action precedes are as follows. ttqIv — I 20 2, 39 1. 
 78 1. II 13 1. 53 3; 4. 56 1. 67 3. 93 1. Ill 46 6. 69 2. 
 94 4. lY 2 1. 14 1. 67 3. 69 1. 70 2. 78 5. 85 4. 
 1014. 125 1; 4. 135 1; 1. Y 9 6. 38 2. 413. 60 6. 
 65 3. 84 3. YI 4 2. 29 1. 61 1. 76 1. 97 3. YE 28 3. 
 36 1. 50 4. YIII 45 1. TiQoxeQov fj — YI 58 1. YIE 45 5. 
 
 A clause of posterior limit rarely precedes. «o? — 
 Y 82 3. YII 16 1 (ellipsis), jnexgi ov — Y 26 4 (the clause 
 is one of two temporal modifiers of the main verb). YII 83 2 
 (ellipsis). TTpiv — YI 29 2. Compare ev ooco — III 28 1. 
 
CHAPTER 11. 
 
 A cLassified collection of the temporal clauses in Thuky- 
 dides follows. It is difficult to draw a line between temporal 
 clauses and causal, and between temporal clauses and circum- 
 stantial; but it has seemed best to include in the following- 
 catalogue doubtful cases in which an interpretation as temporal 
 is possible. 
 
 Since neither the extensive nor the aoristic conception 
 can claim for itself alone the copulative verb, the occurrence 
 of this in forms of the imperfect indicative, the subjunctive, 
 and the optative, has been noted, ecpriv and cfah]v have been 
 classed as aoristic^), but the verb is specified when it occurs. 
 
 EJIEI. 
 
 Past sphere, a) Subordinate verb indicative imperfect: 
 — I 30 a. 49 1^). II 3 4^). 51 5*). 81 8. V (>9 1. VI 
 65 1^). YII 34 4. b) Subordinate verb indicative aorist: — 
 II 65 5"). Ill -24 3. IV 30 l>''). 44 "2 83 2«). VI 2 5. 
 50 5"). VII 1 1. 26 1. 44 7. VIII DO r°). c) Subordinate 
 
 1) Cf. Gildersleeve A. J. P. IV (1883) />. 161. 
 
 2) Twice. For the first verb the best manuscripts have iysrero, but 
 Classen rightly points out that iyiyveto is necessary The expression of 
 the critical moment when the change came does not, as is held by Boehme- 
 Widmann, require the aorist, but is found in the opposition of ejTEi Si to 
 TO /.lEv TiQonor. The imperfect is not inconsistent with kaujiQcog; cf. oatpcog 
 in I 118 2. On the reading e^eiSij see p. 68. 
 
 3) ■^v. 
 
 4) See p. 26 sy. 
 
 5) ^»' twice. 
 
 6) See p. 69. 
 
 7) ijrei is Stahl's conjectm-e for yai of the manuscripts. 
 
 8) For IV 93 1 see under l-reid/j with the indicative aorist, p. 57. 
 
 9) Twice. 
 
 10) ijrei laxiaxa. 
 
56 
 
 verb indicative pluperfect: — HI 23 3. 107 4. d) Sub- 
 ordinate verb optative present: — YIII 38 5^). 
 
 I. Past spliere. a) Subordinate verb indicative imperfect: 
 — I 13 5. 24 6. 2i> V). 58 1»). (53 2*). 74 2. 102 3. 
 131 V). 137 4. II 3 1. 12 3*). 18 3. 19 P). 23 1. 34 8. 
 
 50 r). 6(; 2. 70 1«). 81 5'). 83 3'). 90 1. 98 1'). 
 101 V)- :)'). Ill 3 1. 10 4. 18 1. 20 r»). 34 3'). 
 61 2. 70 ()«). 98 1"). r\^ 36 1'). 67 3. 91^'). 93 3. 
 106 2. Y17 2. 22 2^"). 29 1. 58 1. 63 2; 65 5'*). 72 3. 
 YI 2 6. 32 1'"^). 61 1. 63 2; 2*). 100 1*) YH 4 4. 
 6 1'). 18 1; 3"). 23 2. 27 3. 32 1*). 37 3^. 50 4'). 
 
 51 2'). 60 5'). 65 3'). 69 1; 2. 75 1. 80 1. 82 1. 
 YIII 8 2. 56 3. (57 2. 68 2'«). 73 6. 79 1. 80 1"). 105 3. 
 b) Subordinate verb indicative aorist: — 15 1. 6 3. 111. 
 18 1. 26 3. 28 1. 29 3. 30 2. 39 2. 46 3. 49 1. 
 50 2; 3. 54 2; 2. 58 !'«). 62 5"). 63 2"). 74 3. 79 1. 
 89 2«); 3. 102 4. 125 1. 126 5. 128 3. II 5 7. 12 3'^). 
 
 1) See p. 69. 
 
 2) >})• and naofjv, 
 
 3) Also the aorist twice. 
 
 4) Also the aorist. 
 
 5) Three times. 
 
 6) Twice. 
 
 7) ^v. 
 
 8) Twice: also the pluperfect three times. 
 
 9) .Tagijv. 
 
 10) Three times (once »)r). 
 
 11) Twice; also the aorist and the pluperfect. 
 
 12) Twice (once .Tapj/r). 
 
 13) On V 27 1 see p. 57, footnote 6. 
 
 14) Twice; also the aorist. 
 
 15) Twice (once »}i'). 
 
 16) The passage is con-upt. CGM have only the imperfect iy.axovTo, 
 ABEF have also /leTsoT)] and y.aTFarr]. 
 
 17) In Classen's change to avrartjydyovjo the point of the imperfect 
 with the negative is missed. 
 
 ly) Twice; also the imperfect. 
 19) Also the imperfect. 
 
57 
 
 15 2. 16 1. 17 1. 20 3. 212. 55 1. 65 6. 78 1. 80 4. 
 82. 83 3^. 86 4. 102 1. Ill 8. 13 1. 15 1. 16 4. 26 1. 
 27 3. 33 3. 47 3. 62 1; 5^). 68 5. 70 1. 98 V). 
 102 1*). 108 1. IV 45 2. 46 1. 48 4. 72 1. 74 3=^). 
 78 1'). 811. 93 l*^). 103 4'). 113 2. 130 7. Y 21 3. 
 27 r). 28 1. 44 1. 45 40.. 50 4. 63 1. 65 5«). [76 1^).] 
 VI 46 5. 51 3. 53 2. 61 6. 63 2^'»). 88 7. 100 V'). VII 
 5 2. 18 3"). 26 3. 32 1"). 35 1. 43 3. 44 8. 46. 55 2. 
 70 2. 74 P'). 78 3. 80 6. 84 1. VIII 11; 1. 10 1. 
 29 1. 311. 47 2. 48 3. 52. 63 3. 68 2^^); 3; 4. 69 1. 
 811. 86 9. 88. 89 1. 90 1; 1. 911. 92 2^*). 94 1*). 
 100 1. c) Subordinate verb indicative pluperfect: — I 46 1. 
 48 1. 132 3. II 10 3^'). 70 1''). 78 2. Ill 22 1. 31 2. 
 96 3. 98 1"). 102 r«). IV 67 1. 73 4. 132 2. 133 1. 
 
 1) Twice; also the imperfect. 
 
 2) Twice. 
 
 3j Also the imperfect twice and the pluperfect. 
 
 4) Also the pluperfect. 
 
 5) See p. 69. 
 
 6) If the manuscript reading is retained, the apodosis begins with 
 xal ai . . . jTQso/^sTai. But it is not unlikely that y.ai before oi /lev allot 
 is to be deleted (Campe Phliol. XI (1855) p. 52). This makes the im- 
 perfect avExoiQovv a second predicate in the temporal clause. 
 
 7) e(f}ir. 
 
 8) Also the impea-fect twice. 
 
 9) The clause is probably to be deleted as a gloss. 
 
 10) Also the imperfect. 
 
 11) Also the imperfect twice. — For VII 22 1 see under ejieidr] with 
 the pluperfect, p. 58. 
 
 12) Stahl writes yMl wg, and regards y.al sjieiSfj as a gloss. He says 
 that if Hcd ok is read, the meaning must be ijteiSi] xal Ijiiaxovxeg Ttjv 
 vvxza ovx svdvg 6jQ/j,r]aav, and then one would expect xal ndaav ttjv imovoav 
 rjfxsQav. But may not xal &g be referred to ovx omdxrjv slrai? 
 
 IF)) The passage is corrupt. CGM have only the imperfect ixaxovzo, 
 ABEF have also /ustsotij and yazeort]. 
 
 14) Tliree times. 
 
 15) Compound form. 
 
 16) Three times; also the imperfect twice. 
 
 17) Also the imperfect twice and the aorist. 
 
 18) Also the aorist. 
 
58 
 
 Y K; V). 4;> :i. TC -J. YI lOl a. YII-2-2 1-)- ^7:^. YIII 
 S-J -J. '.>4: 1^). (I) Subordinate verb optative present: — II 
 10 'J. e) Subordinate verb optative aorist: — I 4!) 3*). 
 A'll 44 8. 70 f). 
 
 II. Combination of past and present spheres. Subordinate 
 verb indicative present: — I G8 2. 
 
 III. Present sphere. a) Subordinate verb indicative 
 present: — I 32 5°). A'll 13 '2. b) Subordinate verb in- 
 dicative perfect: — I 5. 3l> 5«). \ll 13 2. 
 
 IV. Future sphere. Subordinate verb optative'^) aorist: — 
 YII 80 5. 
 
 Future sphere. Subordinate verb subjunctive aorist: — 
 
 Y 47 6. YIII 58 G; 7. 
 
 eneidrlr. 
 
 I. Present sphere, a) Subordinate verb subjunctive present: 
 
 — n 34 3^). b) Subordinate verb subjunctive aorist: — 
 n 34 G. Y 103 2«). YII 66 3. 
 
 II. Future sphere, a) Subordinate verb subjunctive present: 
 
 — YII G7 2^). b) Subordinate verb subjunctive aorist: — II 
 72 3. Y 6.3 4. 
 
 ecog. 
 A. 'So long as'. Future sphere. Subordinate verb sub- 
 junctive present: — I 58 2«). OO 3^°). II 72 3'). m 82 2"). 
 Ym 81 3. 
 
 1) Twice. 
 
 2) TiaQiay.evaaro is cited from D and two inferior manuscripts, ttcios- 
 oxevdaazo from the rest. Stahl jirefers the former reading, "nam classem 
 Gylippus non ipse comparavit, sed Syi-acusanos ut comparareut adhortatus 
 est". The pluperfect would be in accordance with the usage ali-eady ob- 
 served, whereby the pluperfect is used of purely antecedent action 
 previously described by an imperfect. 
 
 3) Compound form: also the aorist. 
 
 4) jigoajiahtin' is to be preferred to nooapd'/.loiev of most manuscripts, 
 not on the ground that it expresses the rapid shock (^Classen), but because 
 each onset is over before release can be considered. 
 
 5) Twice (once iari omitted); also the perfect. 
 
 6) Also the present twice (once kazi omitted). 
 
 7) On i.-Tscdt] with the indicative to express future action see p. 17 sq. 
 
 8) c5. 9) Eead ijidhcooiv with ABF. 10) See p. 43 sy. 
 
59 
 
 B. 'While'. I. Past sphere, a) Subordinate verb in- 
 dicative imperfect: — V 60 3'). 75 5. VII 63 4. 71 4. 
 VIII 78"). b) Subordinate verb indicative pluperfect: — 
 VIII 78'). 
 
 11. Combination of present and future spheres, a) Sub- 
 ordinate verb indicative present: - I 78 4*). Ill 70 6. 
 
 V 9 6"). YL 17 1^). 49 1; 2. VII 47 3*). VIII 40 3*). 
 b) Subordinate verb indicative imperfect: — IV 67 4"). 
 
 C. 'Untir. I. Past sphere, a) Subordinate verb indica- 
 tive historical present: — VII 81 4''). b) Subordinate verb 
 indicative imperfect: — V 82 3«). VII 81 4^). c) Sub- 
 ordinate verb indicative aorist: — II 19 2. 81 4. Ill 93 2. 
 
 VI 44 2 62 3. VII 19 5. 26 3. 35 2. 
 
 n. Combination of present and future spheres. Sub- 
 ordinate verb subjunctive aorist: — VI 77 2. 
 
 ni. Future sphere. a) Subordinate verb subjunctive 
 aorist: — I 28 5 29 5. II 7 2. Ill 97 1. IV .30 4. VII 
 16 1. VIII 11 2. 29 1. 84 5. b) Subordinate verb optative 
 present: — III 102 7. c) Subordinate verb optative aorist: 
 — Ill 95 1. V 35 4. 
 
 fjvixa. 
 
 Past sphere. Subordinate verb indicative imperfect: — 
 
 VII 73 3. 
 
 A. 'So long as'. Past sphere. Subordinate verb indica- 
 tive imperfect: — III 10 4. 98 T"). 
 
 B. 'Untir. I. Past sphere. Subordinate verb indicative 
 aorist: — IV 4 1. 
 
 1) fjv. 
 
 2) i)v: also the pluperfect sQQWfirjv. 
 ;3) iQQcofitjv: also the imperfect ijv. 
 
 4) ioTi omitted. 
 
 5) Twice. 
 
 6) ip'. On the imperfect tense see p. 22. 
 
 7) Also the imperfect tjv. 
 
 8) See p. 42 sy. 
 
 9) ^v; also the historical present. 
 10) Twice (once -^r). 
 
60 
 
 n. Future sphere. Subordinate verb subjunctive aorist: 
 
 — I 137 2. VIII 58 5. 
 
 jj-eyoi ov. 
 
 A. 'So long as'. Past sphere. Subordinate verb indica- 
 tive imperfect: — II "21 1^). 
 
 B. 'Until'. I. Past sphere, a) Subordinate verb indica- 
 tive imperfect: — I 30 3. YIII 42 3"). b) Subordinate verb 
 indicative aorist: — I 109 4. II 22 2. 31 3. V 20 P); 4. 
 YIII 42 3*). 
 
 II. Combination of past and present spheres. Subordinate 
 verb indicative present: — I 76 2. 
 
 ni. Future sphere. Subordinate verb subjunctive aorist: 
 
 — Ill 28 2. lY 16 2. 41 1. 46 3. YII 83 2. 
 
 Forms of og. 
 ov. Present sphere, a) Subordinate verb subjunctive 
 present: — IV 17 2. b) Subordinate verb subjunctive aorist: 
 
 — I 37 4. 
 
 aqf ov. I. Past sphere, a) Subordinate verb indicative 
 imperfect: — II 102 6. b) Subordinate verb indicative aorist: 
 
 — I 14 3. II 73 3. 
 
 II. Combination of past and present spheres, a) Sub- 
 ordinate verb indicative present: — I 18 1"). Ill 116 2. 
 b) Subordinate verb indicative perfect: — II 8 3®). 
 
 1) fjV. 
 
 2) Also the aorist. 
 
 3) Twice. 
 
 4) Also the imperfect. 
 
 5) A scholium is quoted: aqp ov] yQarpsrai xai a, y sig a v Eirj, ol 
 Aay.edaifiovioi' tovteotiv iv otg eteaiv. Van Herwerden (Stud. Time. p. 6) 
 says: "Non memoratur, quod visum est Popponi, duplex lectio a et ek a, 
 sed eig a esse lectionis a interpretamentum apparet ex additis vocabulis 
 v' ert]. Ceterum de hac varietate, cuius vestigium servarent cod. Italus 
 et Parisinus H. exhibentes a dcp' ov Aay.., diligenter videndum est, si quidem 
 quod hodie vulgatur 07' ov potest esse illius glossema, non rations inversa." 
 But doubtless the scholium is to be resolved into three notes: — rsTQay.oaia] 
 yQd<psTai xai a. i] slg a v err) (the error might easily arise by dittography 
 from ^laliaxa), xa&iozaaav'] ot Aaxsdai/Liovioi ("Das Subject geht aus dem 
 Zwischensatz {Aa>c. ygmviai) auch auf y.adiaxaaav uber." Classen), aqy' 
 ov] rovTEOTiv ir oTq hsoiv. 
 
 6) The subordinate verb is fiifivtjrzai. 
 
61 
 
 CO. Past sphere. Subordinate verb indicative imperfect: 
 
 — V 42 1. 
 
 Ev CO. I. Past sphere, a) Subordinate verb indicative 
 imperfect: — 11 86 1. IE 39 2^). [84 2^).] YI 61 2. YIII 
 87 1, b) Subordinate verb indicative aorist: — III 39 3. 
 VI 92 4. c) Subordinate verb indicative pluperfect: — VII 2 4. 
 
 II. Present sphere, a) Subordinate verb indicative present: 
 
 — I 39 3"). II 11 6. YI 92 4. b) Subordinate verb indica- 
 tive perfect: — I 39 3*). c) Subordinate verb subjunctive 
 present: — I 21 2. 37 4. 42 2. lY 17 2"). d) Subordinate 
 verb subjunctive aorist: — YII 29 4. 
 
 III. Combination of present and future spheres. Sub- 
 ordinate verb indicative imperfect: — Y 16 1®). 
 
 lY. Future sphere. Subordinate verb indicative future: 
 
 — I 39 3'). 
 
 /}. Future sphere. Subordinate verb subjunctive aorist: 
 
 — YI 23 2. 
 
 oJg. Past sphere. Subordinate verb indicative imperfect: 
 
 — Ill 56 5^). 
 
 alg. Past sphere. Subordinate verb indicative imperfect: 
 
 — lY 120 1. 
 
 h alg. Past sphere. Subordinate verb indicative im- 
 perfect: — lY 39 2«). 
 
 6v. I. Past sphere, a) Subordinate verb indicative im- 
 perfect: — III 17 Vy 18 V). 94 1. lY 46 V''). Y 55 1^). 
 YIII 73 1. b) Subordinate verb indicative aorist: — I 138 1. 
 
 1) t)v. 
 
 2) The chapter is probably not Thukydidean. 
 
 3) Also the perfect. 
 
 4) Also the present. 
 
 5) f5. 
 
 6) Twice (once ijv). On the imperfect see p. 22. 
 
 7) Twice. 
 
 8) O-T*))'. 
 
 9) The chapter is suspected as an interpolation. 
 
 10) While the fact that a formula is several times repeated does not exclude 
 the possibility of a variation of that formula, and while the paratactic 
 construction is plausible for this sentence, a comparison of the parallel 
 passages, especially III 18 1, 94 1, and VIII 73 1 shows the reading 6V 
 
62 
 
 11. Future sphere. Subordinate verb indicative present: 
 
 — Ill 39 8. 
 
 fig. Past sphere. Subordinate verb indicative aorist: — 
 III 81 4. 
 
 fs o. Past si)here. Subordinate verb indicative pluperfect: 
 
 — V 66 20. 
 
 Forms of oaog\ ooay.ig. 
 
 h oocp. I. Past sphere, a) Subordinate verb indicative 
 im])erfect: — III 81 2. YIII 61 1. 87 4^. b) Subordinate 
 verb indicative aorist: — III 52 3^. 
 
 n. Future sphere. a) Subordinate verb subjunctive 
 present: — YIII 87 1^. b) Subordinate verb subjunctive 
 aorist: — III 28 1^. 
 
 oaov. I. Past sphere, a) Subordinate verb indicative 
 imperfect: — II 57 1*). VIE 5 3'^). b) Subordinate verb 
 indicative aorist: — II 65 5. IV 39 1. c) Subordinate verb 
 optative present: — 11 49 6. YII 70 5. 
 
 II. Combination of past and present spheres. Subordinate 
 yerb indicative imperfect: — I 137 4. 
 
 oodxig. Past sphere. Subordinate verb optative aorist: 
 
 — Aai 18 3. 
 
 ore. 
 I. Past sphere. a) Subordinate verb indicative im- 
 perfect: — 18 2. 9 2. 25 4. 39 3^. 73 2. 74 3^. 91 5. 
 n 13 9^. 50 3. 78 4. 95 2. 99 6. Ill 34 1. 55 3. 
 56 4^; 4. 68 1«). 96 3. 113 1. lY 34 1. 116 2. Y 5 2. 
 
 to be extremely probable, y.m is hardly left, as Classen objects, "ohue 
 passende Beziehung", for the coupling of the two actions by y-ara. t6v aviov 
 XQovov renders easy the insertion of xal at the introduction of the second. 
 The passage (VIII 10 1) adduced by Stahl for the reading without 6V 
 does not mark the correlation in time of two disconnected events. 
 
 1) See p. 51. 
 
 2) Twice. 
 
 3) OJioj. 
 
 4) Twice (once »/i'). 
 
 5) fjv. 
 
 6) Explanations of the passage as it stands are uusatisfactory, and 
 the simplest remedy seems to be to delete a. 
 
63 
 
 13 2. 16 3. 30 2. YI 31 1. 74 1. YIII \)d'). b) Sub- 
 ordinate verb indicative aorist: — I 13 3. 73 4. II 21 1. 
 54 4. i)§ 2. 102 5. Ill 13 1. 29 2. 54 5. 55 1. IV 85 2. 
 91. 92 6. V 41 2. 49 2. VI 46 3. VIII 73 2. 76 4. 
 86 3^). 108 4. c) Subordinate verb indicative pluperfect: — 
 IV 133 3=*). VII 15 2*). d) Subordinate verb infinitive 
 present: — II 102 5, 
 
 II. Present sphere. Subordinate verb indicative present: 
 
 — VIII 78^). 
 
 III. Future sphere. Subordinate verb iudicative future: 
 
 — VI SQ 5. 
 
 OJIOTS. 
 
 I. Past sphere, a) Subordinate verb optative present: 
 
 — II 7(; 4 79 6. Ill 21 4«). 97 3^). IV 23 2«). 26 7«). 
 100 3®). VIII 73 5. b) Subordinate verb optative aorist: — 
 I 90 5. 99 3. II 13 7. 15 1. 18 2. 34 7. 43 1. 49 3. 
 51 4. 65 9. Ill 68 1'). V 16 1. VII 4 6. 44 6. 70 3; 
 3^). 75 3. VIII 53 :V). 
 
 II. Future sphere, a) Subordinate verb optative present: 
 
 — Ill 22 30- IV 77 V). VI 6Q 1. 97 5. VII 48 1. 
 b) Subordinate verb optative aorist: — I 91 3. IV 111 1^). 
 
 OTav. 
 
 I. Present sphere, a) Subordinate verb subjunctive pre- 
 sent: — I 36 1. 141 6«). Ill 56 7"). VI 3 1. b) Sub- 
 ordinate verb subjunctive aorist: — I 142 9. VI S6 3. 
 
 II. Future sphere, a) Subordinate verb subjunctive pre- 
 sent: — I 62 3. 121 1. 142 1. II 11 6. Ill 39 7^^). IV 
 77 2. 126 6''). V 9 7. 47 7^^). VI 79 1. 93 3"). VIII 86 8. 
 
 1) Hude brackets the temporal clause. 
 
 2) AEM have the imperfect FoiliaXlov. 
 3; See p. 35. 
 
 4) SQQOJ/iOjV. 
 
 5) Twice. 
 
 6) sirjv. 
 
 7) On the reading ojtots see p. 73. 
 
 8) (paltjr. 
 
 9) Twice. The indicative o.-iEV(iF(. given by AF is out of place. 
 
 10) The acceptance of Heilmann's conjecture, i'xovoi for k'xojoi, leaves 
 but one subjunctive. 
 
 11) c5. 
 
64 
 
 90 3. b) Subordinate verb subjunctive aorist: — I 1'21 4. 
 144 '2. II 43 1. 64 1. lY 5 1. 60 2. 64 3. V 98. VII 
 50 3. Vm 11 3. [KM) 2^).] 
 
 Future sphere. Subordinate verb subjunctive present: — 
 II 84 2. IT 21 2. Y 7 5. YIII 67 3. 
 
 A. AVith infinitive, a) Infinitive present: — I 39 1; 2. 
 6S 2. II Cu 3. Ill 24 2. 64 3. lY 2 1^). 14 1. 67 3^)-- 
 125 4^). Y 10 3. 41 3. 60 6. 84 3. YI 29 1. 61 1. 
 b) Infinitive aorist: — I 20 2. 78 1. 125 2"). 141 1. II 
 12 2. 13 1. 53 3; 4. 56 1. 82. 86 6. 93 1. Ill 30 1. 
 46 6. 69 2. 94 4. lY 4 3. 20 1. 67 4. 69 1. 70 2. 
 72 1. 78 5. 79 1. 83 6. 85 4. 101 4. 104 5. 125 1. 
 128 1. 135 1; 1. Y 8 4. 9 6. 38 2. 45 4. ' 65 3. YI 
 2 5. 4 2. 11 1. 49 2. 76 1. 97 2; 3. YII 28 3. 36 1. 
 50 4. YIII 12 1. 42 1. 45 1. 
 
 B. With finite verb. I. Past sphere, a) Subordinate 
 verb indicative historical present: — I 132 5^). A'TI 39 2. 
 b) Subordinate verb indicative imperfect: — I 118 2^). YII 
 71 5*). c) Subordinate verb indicative aorist: — I 51 2. 
 n 65 3. ni 29 1. 101 2. 104 6. Y 10 9. 61 1. YII 
 71 b^). Yin 105 2. 
 
 n. Present sphere. Subordinate verb subjunctive pre- 
 sent: — YI 38 2^). 
 
 III. Future sphere. a) Subordinate verb subjunctive 
 aorist: — I 91 3. 6 2. 84 1. 102 5. lY 97 2. YI 10 5. 
 29 2. 71 2'). Yin 9 1; 3. b) Subordinate verb optative 
 aorist: — in 22 8"). lY 117 1. 
 
 1) The passage is doubtless not Thukydidean. 
 
 2) eivai. 
 
 3) Twice. 
 
 4) Also the aorist. 
 
 5) Also the imperfect. 
 
 6) (o. 
 
 7) Six times. 
 
05 
 
 TiQoreoov Tj. 
 
 A. With infinitive aorist: — KU) 5. II 40 2. VI 58 1. 
 YIII 45 5. 
 
 B. With finite verh. I. Past sphere, a) Subordinate 
 verb indicative imperfect: — YIII 24 5^). b) Subordinate 
 verb indicative aorist: — 11 65 12, 
 
 II. Future sphere. Subordinate verb subjunctive aorist: 
 — YII 63 1. 
 
 voxegov //'. 
 With infinitive aorist: — YI 4 2^). 
 
 Past sphere, a) Subordinate verb indicative historical 
 present: — I 63 1. Y 10 5. YII 84 3. b) Subordinate verb 
 indicative imperfect: — I 26 5. 61 3. 63 2^). 64 3. 65 1 
 102 1. 103 1. 109 3*). 126 10*); 11. II IS 1. 21 3 
 33 2. 57 1*). 51) P). C>S '.). 77 1*). S4 3'). DOS. 02 3 
 III 4 1. 7 5. 16 2'). 26 4^). 27 1«). 33 3. 75 4. 77 2^) 
 81 3. 85 3. 88 4. 91 3. 103 1. 108 P); 3*). lY 3 1') 
 4 1. 15 2"). 44 3. 57 2. 71 3. 73 4'); 4. 75 1*). 79 2 
 89 1*). 96 5. 104 3. 110 1. Ill 2^). 117 2^°). 127 2") 
 Y 3 1*). 10 4. 22 2. 36 1. 40 V). 50 1; 2. 52 1 
 54 2. 55 4. 57 1"). 59 2. 61 5. 72 4. 73 3'). 84 2 
 114 1. 116 1; [3^=).] YI 31 1. 44 3. 50 1'); 3. 54 4 
 60 2'"). 61 7''). 62 2. 66 3. 74 2^*). 97 5. 98 3. 102 1 
 
 1) Twice; see p. 43. 
 
 2) Cf. I 51 4. 60 3. 
 
 3) Also the aorist. 
 
 4) Twice. 
 
 5) Also the pluperfect. 
 
 6) Six times 
 
 7) Twice (once .-raoijv). 
 
 8) Twice; also the pluperfect. 
 
 9) f]r. 
 
 10) Twice; see p. 47 sy. 
 
 11) Four times: also the aorist. 
 
 12) The words wc ravra syfyrezo have been condemned on account of 
 the imperfect tense for the aorist that might have been expected, and on 
 account of their position. Probably they are an interpolation 
 
 18) Three times (once /}»-). 
 14) Three times. 
 
 5 
 
66 
 
 103 3. 104 1. All 17 .T). -21 •>. m ■_>. A-1 3'). 43 1 -) 
 .-)() 3. 73 3'). VIII Ki 2; 3. 17 3. 'iO •>. -23 5. 25 4. 
 31 3. 32 1') 3.") 4. 3S 4. 40 3. (iO 3. 62 3^). 63 2. 
 
 70 1*). 77 7!) (;. *>4 1 i).') 4. <»!)'). 102 1^) 103 1. 
 c) Subordinate verb indicative aorist: — I *.) 2*^). 27 1. 
 29 4'). 47 1. 48 3. 611. 63 2«) 72 1. IKi 1. 134 1. 
 II 3 1. 4 1. 5 4. 12 4^). 13 1. 31 1. 74 2. 79 5. 81 ('>. 
 94 3. Ill .") 1. S. 22 4. 2S 2. 36 5. .51 4. 60. 68 1. 
 (;9 1. 9.') 2. 102 :>. 106 1. 108 1'). Ill 2. 112 5; 7. 
 113 .V). IV 3 V). 6 1. 13 3. 15 1; 2«) 25 9. 33 1. 
 35 2. 42 4^). 44 4; 4^. 47 1^). 48^). 68 3. 70 1; 2. 
 73 4^\ 74 3. 89 2 90 3. 93 2. 94 1. 101 3. 110 2. 
 116 1. 122 4. 123 2. 125 1; 2. 127 2^"). 128 3. 129 4. 
 130 4 Y 6 1. 8 1. 10 3. 20 1. 32 4^^). 37 3. 38 4. 
 40 V). 46 5. 55 3. 65 1. 72 1. 73 2; 3'). 80 3 83 1. 
 112 1. VI 7 2. 50 V). 51 2. 52 1. 57 1; 2. 65 2. 
 96 1. 102 2. YII 3 3. 4 3. 23 1. 33 3. 35 2^). 43 1"). 
 81 P); 2. 83 2. YIII 10 2^). 16 3. 17 1. 19 3. 23 4. 
 27 1 28 2. 33 4^'). 35 2. 40 2. 43 1. 51 1. 70 1"). 
 
 71 2^*). 74 3. 79 2. 83 2. 84 3. 92 6. 94 3. 96 1. 
 100 4'') 103 2. 104 2. 108 3. d) Subordinate verb in- 
 dicative pluperfect: - I 29 ^''). II 59 1"). Ill 23 V). 
 26 4"). 27 r^). 69 2. lY 90 4 1112'). Y 57 P'). 73 1. 
 YIII 99 ^''). 
 
 1) Twice. 2) Also the aorist. 3) ?)>•. 
 
 4) Twice; also the aorist. 
 
 5) Four times (once rjv): also the pluperfect. 
 
 6) On I 19 see p. 47. 
 
 7) Also the pluperfect. 
 
 8) Also the imperfect. 
 
 9) There does not seem to be sufficient reason for Stahl's change to 
 &? with the insertion of t?. 
 
 10) Also the imperfect four times. 
 
 11) f(;i?;)'. 
 
 12) In VIII 34, the reading of B presents no difficulty, but the testi- 
 mony of the other manuscripts points to a reading with owttsq. 
 
 13) Also the imperfect twice. 
 
 14) Four times. 
 
 15) B has w?, the other manuscripts me. 
 
 16) Also the imperfect four times (once i)v). 
 
CHAPTER III. 
 
 For convenience, statistics^) of moods and tenses with 
 the different conjunctions, and some notes on miscellaneous 
 points may be given here. 
 
 § 1. a) enel, ejieidi). Number of 
 
 occurrences 
 
 ejiei is followed by the indicative imperfect. 8 
 
 ijiei ,, 
 
 •)■> 
 
 „ aorist . . 
 
 10 
 
 55 
 
 55 55 • • 
 
 1 
 
 It 
 
 ,, pluperfect 
 
 2 
 
 t1 
 
 optative present . 
 
 1 
 
 55 
 
 indicative ,, 
 
 '2 
 
 15 
 
 „ imperfect . 
 
 (;3 
 
 15 
 
 „ aorist . . 
 
 106 
 
 55 
 
 „ perfect. . 
 
 -> 
 
 55 
 
 „ pluperfect. 
 
 19 
 
 55 
 
 „ present and 
 
 
 
 perfect .... 
 
 1 
 
 55 
 
 indicative imperfect and 
 
 
 
 aorist 
 
 10 
 
 51 
 
 indicative imperfect and 
 
 
 
 pluperfect . . . 
 
 1 
 
 J5 
 
 indicative imperfect, 
 
 
 
 aorist, and pluperfect 
 
 1 
 
 55 
 
 indicative aorist and 
 
 
 
 pluperfect . . . 
 
 2 
 
 55 
 
 optative present . . 
 
 1 
 
 55 
 
 „ aorist . . . 
 
 4 
 
 1) Totals are given in accordance with the classification of Chapter II. 
 Passages bracketed are not counted. 
 
 2) See p. 69. 
 
 5* 
 
68 
 
 f.Tf/ and i.-rfidi'i, temporal, may sometimes by virtue of 
 tlie context be translated by 'since". The real meaning of the 
 conjunctions is, however, unchanged. 
 
 The idea of antecedence runs easily into that of cause, 
 and the conjunctions i.-rei and ejietd)] sometimes introduce 
 causal clauses^). The interrelation of the main and sub- 
 ordinate actions is then not subject to the same limitations 
 as when the connection is temporal. iTrei has a range not 
 granted to iTTsidt], in that it is used occasionally to append an 
 explanatory sentence almost or quite independent. 
 
 Distinction between eTiei and ejieidyj is not to be too 
 sharply drawn, but in general e.-T€id)'j is the more exact and 
 vivid. Accordingly ejieidi] is used when there is contrast and 
 opposition, or when there is a shift of subject. Manuscript 
 variations in a few" passages demand notice. 
 
 I 41) 7: ol b" 'A&rjvaioi ogowres tovq KsQxvQaiovg jrie^o- 
 jiievovg fxukXov fjdt] d7ioo(paoiorcog EJiexovgovv, to fih> ttomtov 
 cme'/pfxevoi ojore fii] eju^d?J,eiv Tivr enel Se y tqoti)] eyiyvero 
 ?i.ajuJiQ(bg y.al evexeivro ol Koolvdioi, rore Sij eoyov :ndg ei'yero 
 i]d)] xal Siexey.oiTO ovSev en. C has ijisiS)]. Since this is the 
 only place where there is important manuscript authority in 
 favor of f.Tf/ when opposition is expressed by juev and Se 
 (except Vni 38 5), or when the resumptive tot£ d/j appears 
 in the apodosis, eTzetd)] may be the correct reading It is a 
 little strange that in the closest parallel to this passage (I 63 "J), 
 C has ijiei against ejieid)'] of the other codices. 
 
 I 132 3: rov ixhxoi Ilavoaviov ddixi]jua xat tot' edoxei. 
 eivm, xal ijisiS)) ev tovtco xa§eioT)]xet, tioXXco /.idAAov TiaQo^ioiov 
 TToa/J^fjvai E(paivezo t/) Tiagovo)] diaroiq. C and G read i^rei ye 
 d)]. Although ejieidi] under similar circumstances appears alone 
 in V 76 2, the fact that all the manuscripts have ijzsid/j ye 
 
 1) It is hard to draw a line between temporal and causal clauses. 
 The following, however, seem purely causal. in:et' — I 12 1. 41 3. 69 5. 
 II 89 3: 4. 93 3. Ill 45 3. 82 1. IV 78 3. 80 3. VI 16 4. 18 2. 
 86 4. 79 2; 3. 89 6. 92 1. VII 24 2. 30 2. 67 3. VIII 66 1. 87 4. 
 STieiSi] - II 35 3. 61 2. Ill 31 1. 60. IV 73 2. 91. V 27 2. 50 1. 
 85. 90. VI 13 2. 16 1. 18 1: 3. 20 1. 63 2. 88 1; 2. VH 15 1. 
 33 6. VIII 63 4. 
 
69 
 
 in YII 55 2 gives some support to ye here, whothor it is in- 
 serted or is to be transposed to follow di'j. 
 
 II 65 5: ooor re yaQ xqovov jtqovot)] tj]s tto/.fo)!; n' rfj 
 do/p'fj , jLieTQicjog e^rjyeiTO, xai aocpaXo)g diecpvka^ev avxip' y.al 
 iyevexo en exeivov /.leyioTi], ene'i re 6 JioXe/iOi; xareoTi], 6 de 
 q}aiverai xal ev rovrco Tigoyvovg rijv dvvajLuv. C and G give 
 ejieidi'j. Two periods of time, it is trne, are in comparison; 
 but since the relation of the two sentences is not adversative 
 but copulative, ejieid/] is not necessary. 
 
 IV 1)3 1 : xai xard rdyog avaorijoag f]ye rov otohtov (ij'^tj 
 ydo xal riyg fjjuegag dye fjv), xal meidi] JTQOoefiei^ev eyybg rov 
 OTQar.evjj.axog avrdn' . . . haooe. A B F Gr have xal ejiel de, C 
 has ejiel de. Probably xal ejieidt) is to be read. 
 
 YIII 38 5: at d' ex rfjg Zdfxov vfjeg avrdig enijxlovg /uh 
 inoiovvro laTg ev rfj AIi?Jjto), eTxel de jLii] dvravdyoiev, dvayrnQoinneg 
 Tidhv eg ti]v Zdf.iov ipvya'Cov. In B ejreidi) is found Classen 
 objects to this on the ground that eneib)) is very rarely used 
 with the optative, and Stahl on the ground that it is less 
 appropriate here where there is no progress in the narrative. 
 But in all the other occurrences of the optative in antecedent 
 action (I 40 3. VII 44 8. 70 5) Thukydides uses e7xeidyj\ 
 and it is hard to see that there is no progress in the narrative. 
 Further, the opposition of //£>' and de is present; and the 
 passage occurs in that part of the history where the authority 
 of the Vaticanus is greater than that of the other manuscripts. 
 
 b) enijv^ eJTeiddv. Number of 
 
 occurrences 
 
 tTTtp' is followed by the subjunctive aorist ... 3 
 
 ejieiddv „ „ „ „ „ present . . 2 
 
 „ „ „ „ ., ,, aorist ... 5 
 
 The three occurrences of ejiip' are in treaties (V 47 (!. 
 
 VIII 58 6; 7). StahP) seeks to explain the appearance of 
 
 ejT)'iv on the ground that it would have been understood more 
 
 readily than the Attic eTretddv; but as regards the first 
 
 treaty, he assumes that the copy set up at Olympia was in 
 
 1) Quaestt. Grnmin. p. 48 sqij. 
 
70 
 
 the Attic dialect, and so far as the second is concerned, the 
 selection of a single form from Ionic as a kind of xoivi) is 
 very doubtful. It may be that the occurrence of the form 
 only iu treaties is accidental, and that for Thukydides ejirjv 
 simply holds the same relation to IjiEtMv that EJtei has to 
 ijieidi]. In V 47 G: loTg de jio)]dovoiv i) Ji6?ug fj jiejujtovoa Jiage- 
 yhco fxey^Qi (mv zQiaxovra ijjueQMv oirov, Imp eXd^oioiv eg rijv nohv 
 T})v biayyeiXaoav j^otjd'eTv, the temporal clause is not emphatic. 
 It limits /(^ZQi rQiaxo^ia fjjue.Qwv for greater accuracy, but the 
 sentence could hardly be misunderstood without it. Compare 
 VI 2 5: y.al rd xodrioTa rfjg ypjg cpx}]oav exovreg, ejiel ddlh]oav, 
 hi] eyyvg XQiaxooia ttqIv "EkXip'ug eg ZixeXiav eXdelv, as against 
 III Q^ 5 : xal id /nev xard IDAjaiav erei tqitco xal evevt]xoor(p 
 eTreidij 'A&ijvaicov ^vft^ua/oi eyevovTO ovrcog EjeAevx)]oev . Previous 
 mention of the temporal limit renders the clauses unemphatic 
 in VIII 58 5-7: rQocp}]v be raig vavol laTg vvv naQovoaig 
 Tiooaffeovi]v jrageyeiv xard rd ^vyxeijueva j^eyQi av at vrjeg at 
 ^aoUewg I'Adcooi' . . . em]v de at ftaoiXecog vfjeg dcpixannai, ai le 
 Aaxedaijuorlcov vrjeg, x.r.A. Contrast the use of ejzeiddv to arrest 
 coextensive action in II 72 3: nvTol de jueTayoogijoare ojtoi 
 ^ovXeo&e, ecog av 6 jioXejLiog fj' eneiddv de JiaoeX&ij, djioSojoofiev 
 vjiuv a dv JiaQaXd/Hcojiiev. 
 
 § 2. e(og. Number of 
 
 occurrences 
 
 ecog, 'so long as', is followed by the subjunctive present . 5 
 
 „ 'while', ,, „ ,, „ indicative „ . 8 
 
 „ „ „ „ „ „ „ imperfect . 5 
 
 11 ;? II •)■> 11 11 >i 11 
 
 and pluperfect . . 1 
 „ 'until', „ „ ,, „ indicative imperfect . 1 
 11 11 11 }i 11 11 11 aorist . . o 
 
 11 11 ,1 11 11 1, 11 historical 
 
 present and imperfect 1 
 
 „ „ „ „ „ ,, subjunctive aorist . . 10 
 
 „ ,, „ „ ,, „ optative present . . 1 
 
 '1 11 11 11 11 11 11 aorist . . . z 
 
 ecog, 'so long as', has the correlative /neygi tooovtov in 
 
71 
 
 I 90 3, if the temporal clause in that passage is regarded as 
 one of coincident extension. With eoK, 'until', there is some- 
 times final force; and o(piai occurs in a clause with l'(og 
 followed by the optative in V 35 4. 
 
 Thukydides does not use k'wg with adverbial expressions 
 or with substantives, eg dye is to be read in III 108 3. 
 
 § 3. i]vixa. 
 In the single occurrence of fji'fy.a in Thukydides (VII 
 73 3), it is followed by the indicative imperfect. 
 
 § 4. fiexQi^), fiexQi ov. 
 
 Number ol' 
 occurrences 
 
 'so long as\ „ 
 
 }1 
 
 55 
 
 5? 
 
 'untir, „ 
 
 ■>1 
 
 JJ 
 
 55 
 
 •>1 •!■> 
 
 55 
 
 >5 
 
 55 
 
 fir/Q^, 'so long as', is followed by the indicative imperfect 
 
 „ 'until', „ „ „ „ „ aorist. . 
 
 ,5 55 55 5, subjunctive „ . . 
 
 jLiexQi ov, 'so long as', „ „ „ „ indicative imperfect 
 
 „ present . 
 „ imperfect 
 „ ,, ,5 55 55 55 55 aorist . . 
 
 55 „ „ imperfect 
 
 and aorist 1 
 „ ,, ,, „ „ „ subjunctive aorist . 5 
 
 In I 109 4 the duration of the action limited is stated, 
 eviavxbv xal e^ iifjrag. 
 
 ixeym (never i^ir/Qi ov) as a preposition is used freely 
 (seventy times) by Thukydides. 
 
 § 5. Forms of og and ooog; oodxig. 
 
 is followed by the subjunctive present. 
 
 aorist 
 
 a(f.> 
 
 ov 
 
 
 55 
 
 55 
 
 
 J5 
 
 55 
 
 
 55 
 
 55 
 
 
 7- 
 (p 
 
 
 
 
 55 
 
 55 
 
 ""^ 
 
 
 55 
 
 indicative 
 
 present . 
 
 
 J5 
 
 55 
 
 imperfect 
 
 
 55 
 
 55 
 
 aorist . 
 
 
 J5 
 
 5? 
 
 perfect . 
 
 
 55 
 
 55 
 
 imperfect 
 
 Number of 
 occurrences 
 
 1 
 
 1 
 
 1) Thukydides does not use the form fn/.Qt;. See Stahl QuaesU. 
 Graiiiin. p. 50. 
 
72 
 
 h' o> is followed by the indicative present. 
 
 „ imperfect 
 
 1} ?1 
 
 ?1 11 
 
 olg 
 aig 
 
 iv alg 
 6v 
 11 
 11 
 
 ag 
 eg 6 
 iv ooco 
 
 yi }i 
 
 11 11 
 
 11 11 
 
 11 11 
 
 11 11 
 
 11 1} 
 
 11 11 
 
 r-1 11 
 
 „ future . 
 
 ,, aorist . 
 
 „ pluperfect 
 
 „ present and 
 
 perfect 
 subjunctive present. 
 „ aorist . 
 
 11 ^1 
 
 indicative imperfect 
 
 present . 
 imperfect 
 aorist . 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 11 
 
 „ pluperfect 
 
 ,, imperfect 
 
 „ aorist . 
 
 subjunctive present . 
 „ aorist . 
 
 indicative imperfect 
 „ aorist . 
 
 optative present . 
 oodxig „ ,, „ ,, „ aorist . 
 
 Only a few forms of the relatives are used absolutely as 
 temporal conjunctions: — of', ucp ov, er co, eg o^), ev ooco. 
 It has seemed best, however, to notice instances in which 
 there is as antecedent a word denoting time f/Qorog, xaigog, 
 fjfteoa), with value but slightly different from that of the cor- 
 relative t6t£ sometimes used witli ots; and also instances in 
 which the forms used absolutely are not exclusively temporal. 
 ov occurs in clauses parallel with others introduced by iv co. 
 
 Number of 
 occurrences 
 
 1) See /;. 51. 
 
73 
 
 § G. a) ore, otiote. Number of 
 
 occurrences 
 ore is followed by the indicative present .... 1 
 „ „ „ „ „ „ imperfect ... 28 
 
 55 5J 55 55 55 55 lUtUre .... 1 
 
 55 55 55 55 55 55 aOriSt .... '10 
 
 55 55 55 55 55 55 plupOrfeCt ... '1 
 
 „ „ „ „ „ infinitive present .... 1 
 
 oTTore „ „ „ „ optative „ .... 13 
 
 ,5 5? 55 55 55 ,5 aorlst .... -20 
 
 ore is definite, otxoxe indefinite. The latter is used only 
 with the optative, the former only with the indicative; for 
 doubtless the two passages where "ne is found with the op- 
 tative are to be corrected. In III 97 3: xal ore fxhv ejtIoi to 
 TOJv 'A§)]vuio)v oroar6nebo}\ vjieycooovr, avayoiQOvai dk ETiexeirro, 
 there is no manuscript authority for 6:t6ts; but tliis is the 
 only passage in Thukydides in which ote with the optative 
 has full support, and the adoption of ore here is favored by 
 the parallel in II 79 (J : xal otiote /liev E7((oiev ol \-i&)]vaToi, h'E- 
 hidooar, avaycooovoi dk evexeh'to^). In YII 70 3: ol' te EJxifidTai 
 i&EodjTEVov, OTE TTgooTTeooi raT's r)ji, juij lEiTtEod^ai tu diro tov 
 xaTcxoTQCOjuaTog rijg aXXi]q TEyvijg, A C E F G have ote^ B M otiote, 
 which has found acceptance with recent editors. 
 
 OTE is the only conjunction with which the inverse con- 
 struction appears. 
 
 EOTLV OTE occurs eight times. 
 
 Other particles are rarely joined with these conjunctions. 
 OTE 7TEO is found in 18 2 and III 54 5; ote di) in 11 102 5. 
 In A^III 53 3 A B C E F read otiote Sij, G M otiote Se. Hude 
 conjectures otiote Se dt), but this seems doubtful. Compare 
 the parallel passage, III i^fi 1. 
 
 Correlative adverbial expressions in considerable variety 
 appear. With ote: — tote (I 39 3. Ill 56 4 et al), ttote 
 (V 41 2), Tidkm (III 13 1), fxi]... TTxo (V 49 2), vrr (Till 7S), 
 El' exEivo) Tcp y.aiQOj (III 56 4). With otiote: — (hi (V Ki 1). 
 
 ore jiiev...6T£ de is found once (YII 27 4). 
 
 1) Cf. Curschmann op. cit. p. 10 sy. 
 
74 
 
 o.ToTf is once causal (II 60 4), and three times interro- 
 gative (I ■_> ± YIII 95 4. 1)6 2). 
 
 b) oTUi', ojiikav. Number of 
 
 occurrences 
 
 oTftv is followed by the subjunctive present ... 17 
 
 ,, ,, „ ,, ,, „ aorist ... li 
 
 oTiorav „ „ ,, ,, „ present ... 4 
 
 Once (III 56 7) or<iv is used to introduce a definition. 
 
 a-Torar is used of the indefiniteness contingent upon will 
 and choice. There would be complete uniformity in this if 
 onoxav were read in VIII 90 3. 
 
 
 
 
 
 
 
 § 7. JTOIV 
 
 
 Number of 
 occurrences 
 
 Tioii 
 
 ' is 
 
 followed 
 
 by 
 
 the 
 
 infinitive 
 
 present. . . 
 
 . 16 
 
 •,1 
 
 35 
 
 55 
 
 
 55 
 
 35 
 
 35 
 
 aorist . . . 
 
 . 50 
 
 •>•) 
 
 55 
 
 55 
 
 
 55 
 
 55 
 
 indicative 
 
 historical present 2 
 
 f) 
 
 55 
 
 55 
 
 
 55 
 
 55 
 
 55 
 
 imperfect . . 
 
 1 
 
 ^1 
 
 55 
 
 35 
 
 
 33 
 
 55 
 
 55 
 
 aorist . . . 
 
 8 
 
 II 
 
 55 
 
 5' 
 
 
 53 
 
 55 
 
 55 
 
 imperfect and 
 aorist . . 
 
 1 
 
 55 
 
 ;5 
 
 55 
 
 
 55 
 
 35 
 
 subjunctive present . . . 
 
 1 
 
 5? 
 
 55 
 
 53 
 
 
 55 
 
 53 
 
 33 
 
 aorist . . . 
 
 . 10 
 
 15 
 
 55 
 
 55 
 
 
 55 
 
 55 
 
 optative 
 
 53 ... 
 
 2 
 
 Final force is felt in some jiassages offcov is found in 
 a TToir clause with the optative, III '22 <S. 
 
 The indicative follows ttqiv only twice (YII 89 2. 71 5) 
 in sentences not actually or virtually negative. In these two 
 passages the narrative is intense, and perhaps the writer 
 chooses the conjunction that does not admit of the construction 
 of contemporaneity, in order to put the reader at once upon 
 the right path and husband his attention. In the second 
 passage, indeed, soig of contemporaneous action has been used 
 in the preceding sentence. The subjunctive and the optative 
 with jioiv occur only after main clauses that are negative 
 
 di] accompanies jiqiv with the indicative four times (I 118 2. 
 Ill 29 1. 104 6. YII 39 2), ye dij twice (I 132 5. YII 71 5). 
 The manuscripts have ttqIv ij in Y 61 1: ngog xhv drjfiov ov 
 
<0 
 
 JiQOofjyov ^ovkoi-dvovg ;f^j;/.<aT/aa/, jto\v i] Mavrtvrjg y.al 'HXdoi . . . 
 xaT}]vdyxaoav deojuevoi. The emendation d/j has been proposed, 
 and has been accepted by some scholars. Stm-m^) would 
 delete //'. There does not seem, however, to be sufficient 
 ground for correction. 
 
 Thukydides uses jrotV adverbially twenty- eight times, 
 but never as a correlative to the conjunction. The pheno- 
 mena of correlative usage are interesting. Trooreoov precedes 
 jioiv with the indicative twice (11 ()."> 3. V 10 i)); nco the 
 subjunctive twice (A^I 71 2. VIII i) 3). "With the infinitive 
 there is wider range. jiQoreQov is found in YIII 45 1 ; compare 
 VII 50 4. cpddvo) occurs in the main sentence eight times 
 (III 69 "2. lY 67 4. 79 1. 104 5. Y 9 6. YI 97 2. YIT 
 36 1. Yin 12 1). Yerbs compounded with jtqo- appear in the 
 main clause in I 7(S 1, III 69 2. lY 125 1; compare III 
 46 6. YI 38 2. 
 
 § 8. Ttoozeoov ij. Number of 
 
 occurrences 
 
 TiQOTEQov f] is followed by the infinitive aorist . . 4 
 
 „ „ „ „ „ „ indicative imperfect . 1 
 
 „ „ „ „ „ ,, „ aorist . . i 
 
 „ „ „ „ „ subjunctive „ . . 1 
 
 § 9. voTsgov )]. 
 voTEQov ij is once (YI 4 2) used with the infinitive aorist on 
 the same principle as TTooTsgov fj. In I 51 4 vareoov is a 
 comparative adverb. There is a clause of antecedent action 
 following a numerical expression in I 60 3: y.al affixvovvrai 
 TeooagaxooTfj fjixegq voreoov Im Ooax^g j}^) UoTfidaia d:ifOT7], 
 but voreoov fj is probably to be resolved so that voreoov is 
 adverbial as in ^^II 24 1: xal roonaXov rom] rj^uegn voreoov 
 dianAevoavreg I'orrjoav. 
 
 § 10. cog. Number of 
 
 occurrences 
 
 cog is followed by the indicative historical present . 3 
 
 „ „ „ „ „ „ imperfect. . . . 97 
 
 1) Op. cit. p. 96. 
 
 2) Against the reading >/ compare StahPs note. 
 
76 
 
 Number of 
 occurrences 
 
 aorist 
 
 118 
 
 pluperfect.- . . . 
 
 4 
 
 imperfect and aorist 
 
 11 
 
 „ plu- 
 perfect . . . 
 aorist and pluperfect 
 
 () 
 1 
 
 The mood used by Thukydides with ok temporal is the 
 indicative. Wisen^) cites also the optative clause of IV 5G 1: 
 ToTg (T 'A&i]vaioig tote t)jv jTaqa^aXdooiov di]ovoi to. /uh' 7io?da. 
 fjov/aoav, fbg y.a&' eyAor^v (pgovgrn' yiyvoiro rig ajrofiaoig, but 
 the temporal force is accidental. 8o xaioog seems to give a 
 temporal idea to cog in VIII 1 3: oiTiveg jzeol rwv naoovTcov 
 cog av y.aiQog f] nQo^ovXevoovai^). 
 
 1) Op. cit. p. 31. 
 
 2) On cb; temporal with the moods in Attic see Gildersleeve A. J. P. 
 VII (1886) pp. 167, 543. 
 
INDEX OF SUBJECTS. 
 
 'Ar, 18. 
 
 Antecedent action, 4, 6, 24 syy. 
 
 Aorist: indicative, of future action, 
 
 17; stem, 3, 24, 82, 36, 41. 
 Aspect, 3. 
 
 Causal antecedence, 25, 32. 
 Coincidence, 5, 6, 32. 
 Combination of spheres, 22. 
 Conjunctions, 8, 45 sytj. 
 Contemporaneous action, 5, 6, d2sgg. 
 
 Ellipsis, 12, 88, 
 'Hvi'y.a, 50, 59, 71. 
 "Ecog, 50, 51, 58, 70. 
 'E.-TSI, 45, 55, 67. 
 'E.-ietddv, 45, 58, 69. 
 'EjTEidrj, 45, 56, 67. 
 'Ejiriv, 45, 58, 69. 
 'Eaxiv 0T£, 13. 
 
 Future: indicative, 10, 16; sphere, 
 16 sqg.; stem, 3, 40. 
 
 Generic action, 10, 14. 
 
 Historical present, 20, 31, 34, 40, 44. 
 
 Indicative in temporal clause, 12, 16. 
 Infinitive in temporal clause, 13, 41. 
 Insertion, 5, 6, 32. 
 Inverse construction, 9, 33. 
 Iterative action, 10, 14. 
 
 Limit, 7, 34, 42. 
 
 Me/gi, 50, 51, 59, 71. 
 Mood, 9, 12 sgq. 
 
 Negatives, influence of, 8, 28, 32. 
 
 'Ojiozav, 49, 64, 74. 
 'Omke, 49, 63, 73. 
 Optative in temporal clause, 14, 17,43. 
 "Og, forms of, 49, 50, 51, 60, 71. 
 'Qg, 46, 65, 75. 
 'Oody.ig, 49, 62, 71. 
 "Ooog, forms of, 51, 62, 71. 
 "Orav, 49, 63, 74. 
 "OiF, 49, 62, 73. 
 
 Overlapping action, 3, 6, 7, 24 sgg., 
 36 sqq. 
 
 Participles, 8, 27. 
 
 Past sphere, 12 sqq 
 
 Perfect stem, 31, 34, 40, 45. 
 
 Pluperfect, 31. 
 
 Position of clauses, relative, 51 sqq. 
 
 Posterior clauses, 9, 54 sqq. 
 
 Present: indicative with future force, 
 
 21; sphere, 12 sqq.; stem, 3, 24, 
 
 32, 36, 41. 
 ngir, 51, 64, 74. 
 Prior clauses, 9, 52 sqq. 
 IJooTsoov »)', 51, 65, 75. 
 
 Reflexive pronoun, 29. 
 Remembering, verbs of, 13, 34. 
 
 Sphere of time, 9. 
 Stem-system, 23 sqq. 
 Subjunctive in temporal clause, 14, 16. 
 Subsequent action, 4, 7, 36 sqq. 
 
 Tense, 20 sqq. 
 
 Unreal conditions, 16. 
 "Yoregov ij, 49, 65, 75. 
 
INDEX OF PASSAGES 
 DISCUSSED IN DETAIL OR EMENDED 
 
 II 
 
 III 
 
 IV 
 
 18 1, 
 
 page 
 
 60. 
 
 19, 
 
 5) 
 
 47. 
 
 28 5, 
 
 T> 
 
 38. 
 
 30 3, 
 
 ?5 
 
 42. 
 
 90 3, 
 
 ■>■) 
 
 39, 48 
 
 137 4, 
 
 5» 
 
 35. 
 
 138 1, 
 
 U 
 
 33. 
 
 31 3, 
 
 ri 
 
 39. 
 
 34 3, 
 
 ., 
 
 45. 
 
 35 2, 
 
 1? 
 
 14. 
 
 51 5, 
 
 5? 
 
 26. 
 
 65 5, 
 
 11 
 
 34. 
 
 28 1, 
 
 11 
 
 34. 
 
 96 3, 
 
 11 
 
 32. 
 
 39 1, 
 
 11 
 
 34. 
 
 117 2, 
 
 11 
 
 47. 
 
 IV 
 
 133 3, pa 
 
 ge 35. 
 
 V 
 
 18 5, , 
 
 , 18. 
 
 
 20 1, , 
 
 , 48. 
 
 
 26 1, , 
 
 , 39. 
 
 
 55 4, , 
 
 , 27. 
 
 
 58 1, , 
 
 , 28 
 
 
 36 2, , 
 
 , 51 
 
 
 72 3, , 
 
 , 24 
 
 
 82 3, , 
 
 , 42 
 
 
 103 1, , 
 
 , 15 
 
 VI 
 
 77 2, , 
 
 , 21 
 
 
 92 4, , 
 
 , 33 
 
 VII 
 
 23 2, , 
 
 , 45 
 
 
 60 5, , 
 
 , 29 
 
 VIII 
 
 24 5, , 
 
 , 43 
 
 
 90 1, , 
 
 , 30 
 
LIFE. 
 
 I was born iu Cambridge, Massachusetts, January 3, 1870. 
 After preparatory training in the public schools of Cambridge, 
 I entered the College of Liberal Arts, Boston University, in 
 the fall of 1887, and having completed a four years' course, 
 Avas promoted, in June, 1891, to the degree of Bachelor of 
 Arts. Later, I went as Fellow in Latin for the year 18!)3-94 
 to Bryn Mawr College, and in June, 1894, was promoted to 
 the degree of Master of Arts at Boston University. I pursued 
 my studies in Greek, Latin, and Sanskrit at Bryn Mawr, with 
 Professors Herbert Weir Smyth, Gonzalez Lodge, and Edward 
 Washburn Hopkins, until the summer of 1896. Then as 
 Mary E Garrett European Fellow for the year 189G-i)7, I went 
 to Germany. During the w^inter semester, I attended the 
 University of Munich, and heard lectures by Professors von 
 Christ, von Wolffiin, and Furtwangler; and daring the summer 
 semester, at the University of Berlin, I attended the lectures 
 of Professors Diels, KirchhofP, and von Wilamowitz-Moellen- 
 dorfP. 
 
 It is a pleasure to acknowledge my obligation to each 
 of these men, and to express my hearty thanks. To Professor 
 Smyth and Professor Lodge, who directed my graduate work, 
 I owe a peculiar debt; their unfailing kindness and assistance 
 merit my deep gratitude. 
 
UNIVERSITY OF CALIFORNIA LIBRARY 
 BERKELEY 
 
 Return to desk from which borrowed 
 This book is DUE on the last date stamped below. 
 
 ^4.<i4iS£T£i 
 
 lOScf'53VW 
 APR 19 19? 
 
 18Sep'57B I 
 REC'D LD 
 
 MAYE919S; 
 
 U FEB'6*PV 1 Due end of WINTER Qukrter „^„ j i 8 3 
 
 AUG 
 
 LD21-100m-9,'48(B399sl6)476 
 
Photomount 
 
 Pamphlet 
 
 Binder 
 
 Gaylord Bros. 
 
 Makers 
 Stockton, Calif. 
 
 PAT. JAN. 21. 1908 
 
 U.C.BERKELEY LIBRARIES 
 
 CDM7Da=lbDl 
 
 720948 
 
 UNIVERSITY OF CALIFORNIA LIBRARY