Alc. fr. 44 unde putevσns). (3) Numquam. In Plaut. Capt. 149 numquam istuc dixis is ruled out from the prohibitives (323; 46), but numquam ... quisquam . . . dixerit in Plaut. Rud. 790 (Studies, VI 26) is counted in. (4) Nihil. Nihil ignoveris, Cic. Mur. 31, 65 (322; 45), and other examples are ruled out, but nil fuerit, Hor. Sat. 1, 2, 57 (Studies, VI 26), is apparently included in the prohibitives (and is so regarded by most editors). (5) Nemo. Elmer objects to the instance of nemo which I cite, neminem riseris, Cato, Coll. 1, 31; but in Studies, VI 26-7 he gives dederit nemini, Cato, Agr. 5; nemo habessit, Cic. Leg. 2, 8, 19; moratus sit nemo, Liv. 9, 11, 13; and nemo quemquam deceperit, Liv. 9, II, 4; of which last example he remarks: "this is from a very impassioned speech at the time of a grave military crisis." (6) Nec. Elmer rules out all my examples for Silver Latin. Yet he himself (Studies, VI 26-7) has included two examples from early and classical Latin in his list—namely, nec temptaris, Hor. Carm. 1, 18, 2, and nec me ille sirit Iuppiter (for sinit of the MSS), Plaut. Curc. 27.1

Passing to his consideration of my examples of ne and cave with the perfect, I wish to restate a principle which he endeavors to use against me. I maintained that in prohibitions addressed to an indefinite second person (general precepts), be they perfect or present, there is no means of determining with certainty the presence or absence of emotion in a given case, for the simple reason that they are general. They certainly can not be counted for Elmer's theory; and the fairest course to pursue is to leave all of them out of consideration. Supposing, however, for the sake of argument, that one could determine the presence or absence of emotion, the presents, in fact, far exceed the perfects in number, so that in advancing this view I was aiding Elmer rather than myself. Excluding these cases, I gave three instances of non-emotional perfects. Of one of these, Phaedr. App. 26, 5, Elmer says: "ne timueris is merely one of several conjectures and has not the slightest authority of any kind whatever" (the italics are mine). The case would perhaps seem to call for the strong language which it evoked. But Elmer must share with me the rebuke; for in his original paper (326; 49) he himself gives the example with the same reading and without mention of its being a conjecture. Evidently we both used the Teubner text, which in its enumeration of the more important conjectures adopted makes no mention of this passage. That Tac. Ann. 6, 8 ne cogitaveritis, a passage whose context I carefully considered at the time, is emotional, I am not yet convinced. On the third instance, Mart. 2, 68, 3 ne dixeris, one should read Elmer's later comments on Auson. 296, 83 and Sen. Troad. 553, to see how he treats perfects which he wishes to retain and presents which he desires to exclude. Although Elmer (137; 5) cites Liv. 22, 39, 2

1 Sirit here is certainly an optative, but if nec can be used with the optative, it is absurd to say that it can not be used with the volitive.

sis, neque... desis, neque... des, remarking: "Livy and later writers frequently use neque for neve," he admits no instance of the perfect (pp. 156-7) or present (162–3) in my paper of a similar character.

In his closing words on the perfect he evidently fails to grasp what are the two things requisite to establish the validity of his theory, and without which the theory is untenable. Not merely must the majority of the perfects be emotional, but the largest part of the presents must be the reverse. In Silver Latin the proportion of emotional perfects is much larger than in Plautus, as readers of Bennett's critique (Studies, IX) will recall. In Cicero's Letters Elmer himself does not claim "great earnestness, either real or assumed," for all the perfects (150; 18). Supposing all the examples I cited were clearly emotional (and even Elmer does not claim this), the theory would not be proved, if there were any considerable number of emotional presents. That this last is the case in the period discussed I am convinced, despite Elmer's efforts to remove the examples.

In my treatment of the present I made several mistakes in classification and interpretation. These are frankly to be admitted, though I am glad to say they are far less numerous than Elmer would have his readers believe. Whatever their cause, they were not due to hasty work; for each subjunctive passage was carefully considered at least six times, sometimes after intervals of weeks or even months.

On pages 161 and 163 I cited five instances of nec possis as prohibitions. The interpretation is doubtless incorrect, but when Elmer asks: "Who ever heard of such a prohibition as 'And do not be able'?", I would refer him to Giles' note on 'Latin Negatives and Their Use in Prohibitions' (Cambridge Philological Society's Proceedings, 1901, pp. 12-13), which Professor Gildersleeve very courteously brought to my attention. There an Oscan prohibition is given, the Latin translation of which reads as follows: nec dicere nec fari possit. (Giles points out that nep, the Oscan equivalent for neque, is used only in prohibitions.) In Ov. Art. Am. 1,668 and Ex Ponto 2, 4, 31 cave ne possit occurs. Nec adsequare, Tac. Ann. 6, 8, is not a prohibition. "Neque enim," continues Elmer (I quoted two instances with the perfect, Ps.-Quint. 22, 3 and 50, 6), "is not used with a prohibitive subjunctive for the same reason that it is not used with the imperative mood." The fact that enim can be used with the imperative (Ter. Eun. 751 and Cic. de Sen. 19, 69 are examples) and that nec can be used with the imperative makes it impossible to rule out neque enim with the volitive.

"Another evidence of inexcusable carelessness will be noticed in Clement's free intermixture of subordinate clauses (oro ne facias, etc.) with prohibitions proper." As Elmer has been a serious offender in this regard, though in simpler clauses (Bennett, Studies, IX, pp. 51, 52, 58-60), it is interesting to get such an

unbiased opinion of his own work. I shall be obliged to refer to the original article on the prohibitive and an instance or two from Studies, VI to show Elmer's theory and practice. It will be most convenient to consider each verb separately: (1) Obsecro. Plaut. Amph. 924 te, Alcumena, oro, obsecro te, da mi hanc veniam, irata ne sies; Mil. 862 ne dixeritis, obsecro; Most. 1083. On page 140 (8), Elmer says: "Many of these are accompanied by expressions which betray the speaker's earnestness"; and he includes the word obsecro. This shows clearly that in effect he recognizes the construction with obsecro as prohibitive. Yet he omits Ter. H. T. 292 Syre mi, obsecro, ne me ... conicias, and H. T. 1028, 1029 and Phorm. 945,' without a word of explanation. Again (135; 3): "the orations of Cicero alone contain 81 prohibitions (or probably twice that number, if we include such expressions as quaeso ne facias, obsecro ne, etc.)." His statement shows that he recognizes the feeling to be prohibitive in all these constructions. Why does he include some and exclude others? (2) Obtestor. Studies, VI 27, he cites Plaut. Capt. 320 te obtestor, ne faxint as" perhaps " dependent. On the other hand, he does not mention Ter. And. 291 te oro, . . . te obtestor, ne... segreges neu deseras,1 and rejects II. Lat. 724 vos... obtestor, ne ... velitis. And. 291 and Amph. 924 (which he accepts) are almost identical, obtestor in the first being represented by obsecro in the second. Elmer certainly would not advance the theory that clauses with obsecro are independent, but are not with obtestor. (3) Quaeso. Plaut. Mil. 1333 ne interveneris, quaeso (141; 9) is a prohibition; Cic. ad Att. 14, 1, 2 quaeso, ne pigrere (151; 19) "might well be explained as" among "instances of the same use" (i. e. prohibitions), and de Rep. 6, 12, 12 (136; 4), but Curt. Ruf. 9, 2, 28 oro quaesoque, ne... deseratis is not a prohibition. (4) Dico. Plaut. Trin. 501 dico, ne... siris is independent (the clause is plainly substantive, as I believe one of Elmer's pupils has shown [Durham, Substantive Clauses in Plautus, p. 18]), while Ter. And. 205 sed dico tibi: ne temere facias; neque haud dices tibi non praedictum; cave is omitted. Here the indicative and imperative point to the independent character of ne facias. (5) Oro. Elmer accepts an instance in Plaut. Amph. 924, cited above, and rejects instances like Expectes oro neve interimas me, Incert. 3, 273, 15, which one of our two greatest authorities on Latin syntax pronounces independent. (6) Peto, rogo, etc. "Next to noli (149; 17) the most common form of prohibition in Cicero is, I should say, some circumlocution like peto, rogo, oro, etc., followed by ne with the subjunctive, but I have made no attempt to collect the examples." Elmer cites Cic. ad Fam. 16, 9, 4 petam, ne... naviges as "probably independent" (it seems to me almost parallel with Sil. Ital. 17. 367, which he rejects), and excludes Apul. 19, 3 ne spernas, peto. What principles Elmer follows I can not make out. He rejects some instances where the verb precedes ne and the subjunctive,

1 These passages are given in full, C. R. XV 158.

accepts others, and pursues the same inconsistent course when the verb follows or is thrown in parenthetically. It can not be a question of position or verb or tense, for there is no consistency in his use of any of the three. It will be seen that I have tried to follow him as faithfully as the tangle of contradictions would permit, only to be censured for my carelessness and failure to read and profit by his words.

It is clear that the grammarians are far from harmonious in their treatment of independent and dependent clauses. One phenomenon deserves more consideration than it appears to have received. From Plautus on there are numerous instances where an imperative occurs with various verbs, in the same position as subjunctive clauses with ne. In some instances an imperative and a ne prohibition are used in exact parallelism (e. g. Amph. 924). Is it not possible that all such clauses were felt as paratactic?

Curt. Ruf. 6, 3, 12 and Dracont. 5, 148, 276 are plainly subordinate. Apul. 19, 3 is a proviso (I recognized this too late to prevent the appearance of the example). All other examples of alleged subordination have, I think, been discussed.

Let us consider some of the examples of the present with ne and cave, which I regarded as emotional and where disaster would follow disregard of the prohibition. I will take up only a few representative cases as illustrative of the whole.-Pers. 3, 96. It is true that the friend's help may save the invalid's life. Many invalids, however, regard any interference with them as a distinct injury to their feelings or interests, and often express themselves vigorously. Stat. Theb. 3, 243. Elmer objects to my reading pugnare for pugnate. It does not change in any way the prohibition ne certetis.-Stat. Silv. 4, 9, 55; Mart. 11, 102, 7, and Apul. 30, 6 (the last two with cave) are in a spirit of fun. One can employ vigorous expressions or a vigorous tone even in jest. -Vespa 62. The loss of a debate is often regarded and felt as a disaster.-Apul. 146, 3. Not to be recognized, as a noted robber would doubtless be a serious shock to a bandit chief's feelings. -Curt. Ruf. 4, 1, 22. Should a poor man, suddenly raised to power, forget his humble origin, his reign would in all probability be tyrannical.-Curt. Ruf. 4, 10, 26. Darius urges a messenger not to spare him. Failure to comply with the prohibition would at first sight seem a kindness, but he shows it would not be by adding: "it is often a solace in calamity to know your fate."Stat. Theb. II, III is certainly emotional. (Elmer, having cited the only instance in Plautus (Men. 994) of cave with the third person of the perfect, seems to shun all other instances of the first and third persons with cave as religiously as those of nec. They certainly should receive the same consideration.) There are various other examples, but as Elmer has not attacked them, it is not necessary to defend them. It has been shown conclusively, I think, that, confining the investigation to the lines Elmer would insist on (ne with the subjunctive and cave with the subjunctive), that there is a large proportion of emotional presents.

The most trifling objection Elmer urges against my treatment is to be found in his criticism of the examples given on page 165 of my paper of the disasters resulting from non-compliance with the prohibitions : one has to search for them, as citations are omitted." All the passages referred to were given on the three pages preceding, properly labelled. One appreciates the full value of the criticism when he discovers that, owing to omissions and the absence of citations, he must read over 8000 pages of Teubner text to secure the examples of cave in the period Elmer claims to cover.

If it be true, as Elmer admits, that "the types nec feceris and nec facias are freely used in Silver Latin in prohibitions," why he should be unwilling to consider examples of such usage is a mystery. It is interesting in connection with his statement: "It is beyond all dispute that neque (nec) had in Silver Latin come to be regarded often as an exact equivalent of neve (neu)” to read Giles' note, in which he shows that, in Oscan nep, the equivalent of Latin neque, is used only in prohibitions, and "the form with -que, therefore, is not an usurper in the territory of neve; neve itself is the usurper." That being the case, what becomes of the subjunctive of obligation or propriety with nec?

I am perfectly willing to admit that there are instances in Silver Latin where the perfect or the present with nec can not be translated as a prohibition. The instances which Elmer cites I treated as he did. On the other hand, I do not see why, when subjunctives with nec make perfectly good sense as prohibitions, they should be regarded as anything else, especially since a number of them follow a subjunctive with ne or an imperative.

To lists previously given where the present and perfect occur side by side should be added Plaut. Trin. 101, 1012 Cave ne crepent; ne destiteris; Cic. Att. 10, 13, 1 (150; 18) ne demiseris: pertimescas cave, and Prop. 1, 10, 20, 23, 24 Cave ne capias, neu negaris, neu cadant.

To Elmer's "complete" list of perfects in Cicero's Letters (150; 18) should be added Quint. Frat. 2, 5, 3 ne omiseris.

It will be admitted, without citations from Elmer's original article, that the use of dependent prohibitions introduced by cave was one of the two parts of the theory which he aimed to establish for the period prior to Livy. It will also be admitted by all scholars that it is absolutely essential for the demonstration of any theory's validity that all the examples of the usage within. the period covered be collected, that the citations be given, and that the instances be properly classified so that any one desiring to test the theory for himself can do so with comparative ease. Elmer's frequent references to cave led me to compare the statistics, he gives with my own collections. The following table will best present the results: