The following classification, we believe, will embrace nearly all that is contained in the work:

First, the more precise meaning of words often considered. as synonymous; secondly, the laws of construction, more especially in reference to the moods and tenses of the verb; thirdly, the order of words in the Latin language, as contrasted with the order in our own tongue; fourthly, English exercises for translation into Latin, together with a key to the same.

Between the more formal work of Dumesnil upon the subject of synonyms, and the incidental discussions occurring throughout the Gymnasium, there is this material difference: while the systematic character of the former, which, according to the preface, treats of nearly 7000 words, rendered the utmost conciseness necessary, Dr. Crombie has allowed himself a very considerable latitude in his remarks. Beginning frequently with the expositions of his predecessors, he examines them by the test of a number of passages presented to the eye of the reader, and thus discarding what is incorrect, gradually arrives by induction at a more precise definition. Here the pupil, having the evidence before him, may judge for himself as to the value of the decision; and if he be satisfied of this, he will fix the definition in his memory with much greater certainty than if it had been simply communicated to him on the mere assertion of any lexicographer. Yet even this is of inferior moment, compared with the advantage of habituating the mind to independent investigation. But if the utility of Dumesnil's work is diminished by the brevity essential to its character, Dr. Crombie has still more impaired the value of the Gymnasium by the unreasonable length to which he has carried many of his articles. Thus six octavo pages are employed in pointing out the distinction between reperire and invenire. As many are devoted to metuere and timere. Grandis has nearly two pages to itself. The four words peto, rogo, posco, postulo, are allowed five pages. Sal, dicacitas, facetic, constitute the title of an essay upon which above nine pages are consumed. With equal extravagance, nearly ten are taken up with a discussion upon quoad and quod ad, and six are assigned to candidus and albus; whilst, by way of compensation, ater and niger in a different part of the book devour almost three more. In other cases, where the articles are of a more moderate compass, a similar effect is produced by repetition; and it is scarcely enough to reply to this, that repetitions are necessary for the assistance of the young student; a simple reference in three words would fully answer this object.

The exact discrimination between words apparently equivalent involves so many niceties, that no two persons can advance far into such inquiries without disagreeing. Thus whilst we are ready to assent to a majority of the definitions given by Dr. Crombie, it will scarcely be thought_extraordinary, if on some points we think he has been less successful in his labours. In the discussion on the words connected with the notion of cutting, it is objected to Dr. Hill, who considers that cædere implies severity in the blow, that this observation is inconsistent with cædere ferula of Horace, and the phrases, cædere femur, pectus, frontem, in Quintilian. Our notions of the violent action used by the ancient orator will not allow us to admit the inference from the three last phrases, and as to the first, we may appeal to the authority of any school-boy, whether the use of the ferula is incompatible with severity in the blow. In the passage of Horace referred to, the emphasis lies not upon cædas but ferula :-Nam ut ferula cædas dignum majora subire verbera, non vereor.

The section upon sino, patior, permitto, appears to us deficient in reference to permitto, which signifies not merely 'to suffer,'' to permit,' 'to give leave,' but 'to leave a matter entirely at another's disposal.' The very passages quoted by Dr. Crombie confirm this, and the usual power of the preposition accounts for it. If, indeed, we examine the primitive meaning of the word, it signifies to let go entirely,' that is, 'to abandon all control over anything.' It is in this very way that, while cedo in its first sense signifies to go,' concedo obtains the meaning of 'to leave the field entirely,' 'to give it up.' If situs and sino be connected words, and there can scarcely be a doubt about it, the notion of 'to suffer' must have originated in a similar manner. Sino, to place,' or 'put down,' will easily become equivalent to abandon,' and that which is abandoned is of course left at the disposal of others. Thus originally in all these words the permission would be merely negative; but in the case of permitto, the intensive power of the preposition, and the usual addition of some particular person in the dative-hoc tibi permittoaffords a sufficient explanation of the change of meaning from negative to positive permission. In the phrase of Livy*. concitant equos, permittuntque in hostem-it would be a nice question to decide how far the entire abandonment of the rein is a negative or positive permission. It is certainly very positive in its results.

*Dumesnil quotes this passage to prove that the original meaning of permitto is 'to send through.' The words 'in hostem' disprove this.

One of the greatest faults in the Gymnasium, is the neglect of etymological formation, which though not always an unerring guide, is at least a necessary auxiliary in such inquiries. It will not be disputed that dux and duco are connected, notwithstanding the difference of quantity. Had Dr. Crombie considered this, he would have been more accurate in his distinction between imperator and dux, which he explains in these words:

'Imperator means "the commander in chief"; dux, the highest of the inferior officers, having himself an important command. "Præstate eandem nobis ducibus virtutem, quam sæpenumero imperatori præstitistis." In most cases they may be used indiscriminately, &c.'

The passage from Cæsar is quoted apparently on the supposition that ducibus and imperatori are opposed to one another, whereas in fact the antithesis lies between nobis and imperatori. Dux, in short, means simply a guide or leader, who of course must be present from the very meaning of the word. The only word opposed to imperator in the sense given by Dr. Crombie is legatus. The notion that dur and imperator may be used indiscriminately is, it appears to us, altogether incorrect, dux being a generic term, whilst the other title is altogether confined to one who holds the imperium. But this is not the only instance in which the aid of etymology has been neglected. The distinction between comes and the connected words, (p. 92.) would have been more evident to the student, and at the same time more firmly fixed in his memory, had he been informed of the primitive meaning of the word, one who goes with another; com-it-ium being a word of like origin, which stands in the same relation to com-ire that initium, exitium, do to inire, exire. Again, had jumentum been referred to juvare, to which it as strictly belongs as adjumentum to adjuvare, the pupil would have seen that the word 'help' could not well be limited to any particular kind of cattle; but must include all those animals whose labour man appropriates to his use. The word ingens is another instance of the advantage to be derived from this method of examination. The passage quoted from the Eunuch (p. 155), is a proof indeed that ingens rises in power above magnus; but it does not tend to fix the meaning so precisely as the formation of the word from in and gens, the latter of which is of course a derivative from gen, nature,' 'production.' Thus ingens would signify unnatural,' ' prodigious.' It may, perhaps, be opposed to this, that gens being a substantive, an adjective could not be in this manner formed from it. The objection would be equally valid against ex-pers,

in-ers, a-mens, de-mens, centi-manus, ex-sanguis, &c. &c. But the true way to consider the question, is to look at the element gen, which enters into the formation of ingens, and to leave the word gens, which contains the same element, to its own proper classification and explanation.

In the double series of pronouns given in p. 118 of the second volume*, some improvement would be made by pointing out the etymological connexion between the words, and still more by a better arrangement, and the addition of some other forms to make the list more complete. Thus to commence with the connexion between quis and uter, the student should be told that they have a common stem, the resemblance being destroyed solely by the loss of the guttural at the commencement of the second word, which still remains in the Herodotean form xórepos, and in ne-cuter †. The second pair in Dr. Crombie's series should be alius (alis) and alter, rather than unus and alter; so nequis rather than nullus should be opposed to neuter; and the list would be improved by the insertion of ali-quis, alter-uter, with the adverbs unde, utrinde; undique, utrimque; ubi, utrobi; ubique, utrobique, &c.

On the other hand, in the few passages where Dr. Crombie has availed himself of the aid of etymology, he has adhered to the awkward system of English etymologists, who can never conceive a root to exist, except as the nominative of a noun, or first person ind. pres. of a verb, or else the infinitive. Thus he deduces ingenium from ingenitus, as if the existence of this participle were in any way a necessary condition to the formation of ingenium. How much more simple is it to consider gen, divested of all suffix, as the significant syllable which expresses the notion of production,' 'birth,' nature,' &c. in the various forms, gen-itus, gen-ui, gen-itor, gen-etrix. The shape of in-gen-ium is then precisely analogous to in-cend-ium, ædi-fic-ium, od-ium, conjug-ium, con-nub-ium, &c. According to the ordinary system, we are not surprised to find Dr. Crombie, like so many before him, treating mactus as a compound of magis auctus. When magnus is submitted to the crucible of the German philologist, it is readily found that the basis of the word is mag (like the Greek μey), and the existence of a participial form (mag-tus) mac-tus presents no longer any difficulty. An English etymologist would altogether fail

*This reference is to the third edition. In his last edition Dr. Crombie has omitted the whole passage. Our remarks in page 318 were also made upon that edition. † Marini Iscriz. Albane n. 148.

in explaining the forms of the comparative and superlative ma(g)ior, mag-simus.

In the same section in which ingenium is derived from ingenitus, we find the following: indoles from inolescere (per epenthesin literæ d).' What the five last letters of inolescere have to do with indoles it would be difficult to say. But we object still more to what is contained within the brackets, particularly the use of the quadrisyllable epenthesis. If some foreign, not English phrase is necessary, inserto d has the advantage of brevity. But after all, what explanation is included in either of the phrases? The real solution of the difficulty is perhaps this that in the older Latin tongue many words had a final d, which, in the later language, disappeared; thus, med for me, prod for pro, red for re, extrud for extra, ind for in (like vdov). Upon this principle it is most probable that the disjunctive particle sed, and the disjunctive prefix se in sepono (sed-pono), separo (sed-paro), &c. are identical. The disappearance of the final d is nothing more than has since taken place in the case of the preposition ad in all the modern forms of the Latin.

:

In p. 322 we meet with a very extraordinary derivation :dubitare* (ex duo et obs. bitere "to go"). Surely Dr. Crombie must have forgotten the little word dubius, which seems to have some claims of consanguinity with dubitare. An attention to the forms of suffixes enables us in each case to separate dub as the stem, dub-ius corresponding to ex-imius, e-greg-ius, &c. and dub-itare to quær-itare ag-itare, im-itari, &c. That the two words are connected with duo we can readily admit. Indeed the b may naturally arise from the digamma sound, which so often inserts itself when an o is followed by a vowel, as octo (oydons) octavus; wov, ovum, &c.

Before we leave the subject of synonyms, or, to speak more correctly, those words which are apt to be confounded from an ignorance of their precise meaning and usage, it will be important to examine what Dr. Crombie has said upon the four little words, hic, iste, ille, is, which deserve the attention of one who wishes to be an accurate Latin scholar, more than any other words in the language, with the single exception of the relative.

As Dr. Crombie has written so ably on the use of the relative, we had expected to find some useful matter on the subject of these pronouns also; yet, although they form the subject of two distinct articles, besides incidental notices,

The difference of quantity (dubĭtāre, bītère) is a strong, but not alone a decisive objection.