Omission of repeated words or syllables, and of words interposed between repeated words or syllables: see on § 13 omnium omnium, § 26 in infinito, *omnino in eo, § 25 *aquae adjunxit-aquam adjunxit, § 66 *hamata, § 71 quasi corpus aut quasi, § 78 formica formicae, § 81 apud eos-apud nos, § 82 *alia nobis, § 71 in ceris diceretur, § 58 anteferret et, § 103 oportet et, § 49 * neque eadem ad, § 98 nisi in eo-nisi in eo, § 103 *superior aeri aetheriis, § 98 *moribus paribus, § 2 *natura trahimur.

Construction altered through the influence of a nearer word: see on § 2 continet—in primis changed to continet est—in primisque, § 25 mentem changed to et mente, § 36 vi divina, changed to ut divinam, § 49 docet changed to doceat after ut, § 70 esse changed to esset after utrum, fieret changed to fieri to suit following esse, § 71 quam for quod after mirabilius, § 73 inanes for inane to suit imagines, § 104 rationis for ratione to suit mentis.

Substitution of synonyms: see on § 68 quia for quod, so igitur for ergo constantly in C.

Interpolation,

(1) by unintentional repetition: see on § 63 *aut before Neptuni suggested by aut Carbo, § 35 *immittendique after minuendi.

(2) to complete construction: see on § 86 id esse mortale added after si quid sit, § 107 *quam inserted after minus probari possit.

(3) owing to explanatory gloss: see on § 33 * Platone added to explain magistro, § 13 *Terentius, &c. added before Synephebis, § 34 *tum to explain modo, § 58 *L. Crasso to explain familiari, § 112 *nectar ambrosiamque to explaiu epulas, § 1 *id est principium philosophiae to explain the allusion to the Academics, § 28 *praeterea added to correct omne, § 25 *alia added to escape apparent inconsistency.

(4) owing to controversial gloss: see on § 19 *animi added as an answer to the question quibus oculis, on § 21 *quod ne-tempus esset possibly an answer to the preceding intellegi potest.

I proceed now to discuss the question of spelling. This has caused me some difficulty, as I am aware that my own feeling, or perhaps I should rather call it my prejudice, is opposed to the theory and practice of the most eminent both amongst our own and foreign scholars. I think however it is not mere obstinacy which prompts me to follow my own course in this matter, even

against the advice of friends for whose judgement I have the highest respect, and who have studied the subject far more deeply than I can pretend to have done.

It appears to me that this apparently unimportant question is not obscurely connected with the larger question whether the Classics are still to form the staple of higher education amongst us. If their claim to do so is to be allowed, they must show good reasons for it, and they must at the same time leave room for other more immediately pressing studies. I believe that this claim will be allowed in so far as the study of the Classics supplies the necessary instrument for entering into the life and thought of the ancient world, and one of the best instruments for learning the laws which regulate the expression of thought. But the Universities will have to see to it that this is done far more thoroughly than it has yet been done; and for this purpose it will be necessary to drop some of the impedimenta which now occupy the time of the learner without tending, in any corresponding degree, to discipline and feed the mind. Yet, of late years, it seems to me that the burden of the impedimenta has been added to rather than reduced by the new importance which has been given to questions of etymology and orthography. No doubt a wonderful advance has been made in these departments, and, as special subjects for investigation, they naturally and rightly attract to themselves the attention of leisured scholars, but I cannot think they should be made so prominent as they have been in College and University examinations. Viewed in relation to the main ends of a classical education, I hold that spelling is simply a necessary evil, and that, for practical purposes, the best spelling is that which obtrudes itself least, and least diverts the attention of the reader from the thought of the writer. In books therefore which are printed for ordinary reading, we should not seek to reproduce the spelling of a particular age or of a particular author, except where, as in Chaucer, it may be needed to show the scansion of a line, but we should endeavour to give the normal spelling of the language after it assumed a fixed and stationary form; just as we do not in our common Shakespeares reproduce the inconsistent spelling of the early folios and quartos, though for the purpose of studying the history of the language we rightly print facsimiles of these'. In Latin it is generally agreed that the

1 See on this subject the very sensible remarks of Ritschl, Opusc. 11 pp. 722 foll. and 728. I can but echo his final words, spoken with reference to the

language attained its highest formal development in the period which may be named after Quintilian, between Nero and Hadrian, according to Brambach (Hülfsbüchlein f. Lat. Rechtschreibung, p. VII), between the death of Augustus and that of Trajan, according to L. Meyer (Orthographiae Latinae Summarium p. 5). The latter lays down the following rules for our modern spelling of Latin: ne inaequalitate scribendi aut oculi offendantur legentium aut in errorem inducantur animi, scriptura nostra reddi oportet ad certae usum ac morem aetatis, et quidem ejus, qua ipsa lingua scriptorum pariter ingeniis et studiis grammaticorum ad summam est adducta perfectionem; and in p. 6, praeterea ut in sermone, ita in scriptura tamquam scopulum nos fugere oportet quaevis inusitata.

Adopting these rules, it will follow first, that we need not trouble ourselves to frame a conjectural text, such as Cicero might have written, but should use the undoubted spelling of the latter half of the first century A.D.; and secondly, that where this spelling itself was variable, as in the u or i of the superlative terminations, and the i ore of the accusative plural of i-nouns, we should select one mode and adhere steadily to that. In making the selection I should myself wish to apply to our own case the principle suggested by Meyer's second rule, that, of two allowable spellings, that should be preferred which is usitatius, least of a novelty to ordinary English readers.

Turning now to Müller's text I find there several examples of inconsistent, and some of unusual and, as I believe, incorrect spelling. This is the more to be wondered at, because in his excellent review of Baiter and Halm's ed. in the Jahrb. f. Cl. Philol. for 1864, vol. 89, p. 261 foll, he condemns a similar inconsistency in them.

The following are the points in which the spelling in my edition will be found to differ from that in Müller's:

(1) I have always given the superlative termination in -imus; Müller at times has the form in -umus. Thus we find facillume § 9, but facillimum § 61; turpissume § 29, but turpissime § 93; simillumus § 49, but simillimus § 98; praestantissumus § 47, but praestantissimus § 96; also levissumus § 13, vaferrumus § 39.

attempt to expel the old German forms 'genitiv', 'Virgil': möge doch nicht deutscher Pedantismus einen Schatten auf deutsche Wissenschaft werfen, der gegen diese selbst den Spott des weitern Kreises der Gebildeten herausfordern muss!

(2) I have always written u after v; Müller generally does so, e. g. vult in § 13, 33, 34, 69; vultis in §§ 89, 103, 107; Vulcanus § 81; but volt in § 41; voltis § 93; Volcanus $$ 83 and 84.

(3) I have always written es in the Acc. Pl. of the i- declension; Müller usually has is, but we find utiles, salutares, § 38, inmortales § 45, leves § 59, similes §§ 90, 91, venerantes § 85, noctes § 54 though partis comes just before. Baiter consistently gives -is in all cases.

(4) I have regularly assimilated, where it was allowable, because there is no doubt that assimilation was the tendency of the Latin language, and was practised in speaking even in the exceptional cases where it was necessary to preserve the spelling unaltered for the purpose of distinctness or to show the etymology, as in adsum (Roby, Vol. 1. p. 49 n.); Müller as a rule does not assimilate, but we find exceptions, as in regard to the assimilation of in before labials: before b; imbecillus § 122, but imbecillus § 45.

p; inpurus § 63, inpudenter § 69 &c., but impius § 63, impendeo § 45. (Baiter in both cases keeps in.). So conprehensio § 94, but comprehendo § 30.

m; inmensus § 22, and inmortalis frequently.

in before linguals:

1. inlustris § 12; so conligatus § 9, but colligo § 4, comparo § 16 (where Baiter has conligo and conparo).

r. inrigo § 120, inrideo § 101.

Assimilation of d:

before p. adpeto § 104, but appeto immediately after; so adpulsus, adprehendo, but appareo § 37, apparatus § 20, appello § 36.

f. adfluo § 49, but affluo § 114 and affluentia § 51, adfectus § 36,

but afficio § 19; so adfero, adfirmo &c.

r. adrideo § 17 and § 97, but arripio § 77.

l. allicio § 116.

s. adsentior § 12, adsequor § 23, adsidue § 114.

t. attinet § 84.

n. adnuo § 113.

c. accurate § 15.

g. adgredior § 57.

(5) I have always preserved an 8 following a in composition; Müller varies, giving exsistunt § 97, existat § 49, exstitit § 12 and § 21, extitit § 55 and § 91, extingui § 29.

(6) In regard to nouns borrowed from the Greek I have followed Madvig's rule (Gram. § 33 obs. 3), 'Where both forms are in use, it is better to adhere to the Latin', in accordance with the principles laid down by Quintilian 1 5 § 63 and Cicero Att. vI 9, (see the quotations in Roby SS 471, 482). Thus I have always used the termination -em for the Acc. of Greek nouns in -es, whereas Müller writes, at one time, Socraten (1 93), Timocraten (1 93), Simoniden (1 60), Nausiphanen (1 93), Cleanthen (111 5), but more generally Socratem (131), Timocratem (1 113), Simonidem (Div. 1 56), Empedoclem, Aristotelem, Ganymedem, Archimedem, Euphratem, Xenophanem &c. So I have written ibim, Apim in 1 82, where Müller has Apim but ibin: I have uniformly written Zeno, but in § 70 Müller gives Zenon.

(7) I have always written di in the Nom. and dis in the Abl.; Müller uses di or dei, dis or deis indiscriminately.

(8) I have written Lyceo in § 72 where Müller has Lycio, but in Div. I. 8 and 22 he gives the spellings Lyceum, Lyceo. writes oportune § 15, oportunitas § 92', benivolentia § 58, Xerses § 115, Argia § 82, I have written with Baiter opportune, opportunitas, benevolentia, Xerxes, Argiva. In one instance, incoho, I have preferred the less usual spelling to the ordinary inchoo (which Müller keeps) not merely on the ground that it has most authority in its favour, but because it is the more rational, as showing better the etymology and probably also the pronunciation.

Thus far I have not departed much from the prevalent usage in the latest editions. I have now to plead guilty to two heresies. The first is that I have used the character J for the consonantal I. My reasons for doing so are as follows: (1) the use of J, to distinguish the consonant from the vowel I, seems to me to stand on the same footing with the use of V to distinguish the consonant from the vowel U. Neither use was known to the ancients, but convenience has led most editors to preserve the distinctive V, indeed Madvig, who had dropped it in his first edition of the De Finibus, returned to it in the later editions; and all who write on the phonetics of Latin are compelled to mend the unscientific orthography of the Romans by treating the J and V as distinct letters known by distinctive characters. (2) It might perhaps be

1 See his own remarks on the untrustworthiness of MSS in their spelling of double letters, 1. c. p. 138.