case we find, at least in the critical editions of Plautus and Terence, that the first word is preserved entire, and that est loses its vowel. The texts, therefore, are written and pronounced temulenta 'st mulier, homo 'st, molestum 'st. The same thing has been found here and there in very ancient MSS. containing fragments of Cicero's works, e. g., una notio 'st, difficile 'st, and in the oration for Milo: quae illa barbaria 'st. (See Niebuhr's note on the fragment pro Fonteio, p. 60.)* In like manner, we find est joined with a preceding word terminating in us, e. g., opust and dictust; but in this case it remains doubtful as to whether the s of opus is thrown out, or whether est has lost its first two letters. Something similar, though more rarely. occurs in the termination is, e. g., quali 'st. Whether the second person es was likewise joined with a preceding word terminating in us is uncertain. (See Schneider, Elementarlehre, p. 162, foll.)

[ 11.] Note 3.-The hiatus which occurs within a word is generally not removed, and for this reason we did not notice it above. It should, however, be observed that two vowels of the same sound are frequently united (contracted) into one long vowel, and the poets always make dero and desse out of deero and deesse. This explains the forms nil for nihil, and deprendo for deprehendo, which arise from the elision of the aspirate. The contraction of two equal or unequal vowels in the perfect of verbs, after the elision of the v, is still more frequent; e. g., audisti for audivisti, audiisti; deleram for deleveram, norunt for noverunt, concerning which see § 160. It also not unfrequently happens in verse that two different vowels are united, by a rapid pronunciation, into a diphthong; in which, however, both vowels are audible. This is called by a grammatical term synaeresis, and occurs when the two vowels of the words dein, deinde, proin, proinde, huic and cui, are united into diphthongs which are otherwise foreign to the Latin language. In this way alone it is possible to make use of the word fortuitus in the dactylic hexameter; and it is for the same purpose that in nouns terminating in eus, when this ending is preceded by a long syllable, we must contract into a diphthong not only the e in the genitive singular, and ĕis in the ablative plural, but also ea and eō; for example, alvei, aurēt, Nerei, aureis (also antes, from the verb anteco), Eurystheo, cerea, just as a gynæresis sometimes occurs in the Greek words θεός, Νεοπτόλεμος, and ta. Some harsher kinds of synæresis, such as quia, via, vietis, and quoad, are found in the comic poets and in Lucretius.

[§ 12.] 5. There is no necessity for giving any special rules about the orthography in Latin, since there is absolutely nothing arbitrary in the spelling of words that requires to be learned; but there are a great many separate words of which neither the pronunciation nor the spelling is established, and with regard to which the ancients themselves were uncertain even in the best times of their literature, as we see from the monuments still extant. We shall here notice a few things which have not been mentioned in our previous observations. We spell and pronounce anulus, sucus, paulum, belua, litus better with one consonant than with two; whereas immo, num

*[This species of elision, as Niebuhr remarks, was previously supposed to be peculiar to the comic writers only. The same writer observes that he has found no mention made of it in the ancient grammarians, even after diligent search, save perhaps in a mutilated passage of Velius Longus, p. 2238. Niebuhr also cites a somewhat similar usage in the modern Tuscan of Florence; as, for example, lo 'mperadore, la 'neisa.1-Am. Ed.

mres, sollemnis, sollers, sollicitus, Juppiter, and quattuor are more correctly spelled with two consonants than one. It is not certain whether we ought to write litera or littera, though in most MSS. the t is doubled. The authority of the ancient grammarians and the best MSS. teach us to spell the singular mille with a double, and the plural milia with a single 7. The forms narus and navus are not customary now, though they appear to be better than gnarus and gnavus.* Artus (narrow) is certainly better established than arctus; auctor and auctumnus, on the other hand, are justly preferred to autor and autumnus.† The insertion of a p between m and t, e. g., in emptus, sumpsi, rather facilitates the pronunciation than otherwise; and the verb temptare is decidedly preferable to the form tentare, which is now commonly used, the former being found in the best MSS. The forms conjunx, quotiens, and totiens are demanded by most of the ancient grammarians, and are found in good MSS., instead of conjux, quoties, and toties. The words caecus, maereo, are more correctly spelled with the diphthong ae than oe, and saeculum, saepire, and taeter are better with the diphthong than with the simple vowel e; whereas in heres, fetus, femina, and fecundus, and therefore probably in fenus, fenoris also (which are of the same root), the simple vowel is better than the diphthong. But it is very doubtful whether we ought to write scena or scaena, and obscenus or obscaenus, or obscoenus. We do not notice any other points here, because the orthography now commonly adopted is the correct one. Compare Cellarius, Orthographia Latina, ed. Harles, Altenburg, 1768, 8vo; and Schneider, Elementarlehre, Berlin, 1819, 8vo.

[§ 13.] 6. The Romans had no other point than the full stop, and our whole artificial system of punctuation was unknown to them; but, to facilitate the understanding of their works, we now use in Latin the same signs which have become established in our own language. The peculiarities, however, in the formation of Latin sentences,

* [The forms gnarus and gnavus are the original ones, and were softened down, in course of time, to narus and navus. So gnatus is older than natus. There is an evident connexion between gnatus and yiyvouaɩ, and a probasle affinity between gnavus and κνάω, οι γνάπτω.]-Am. Ed.

[There is here an evident inconsistency. If artus be better than arctus, on what principle can autor be inferior to auctor? Compare Journal of Education, vol. i., p. 93.]—Am. Ed.

the many complications of their parts, and the attraction of the relative pronouns, demand great caution in applying the signs of punctuation, in order that we may not by the use of too many signs separate those parts of a sentence which belong to one another.

7. With regard to the use of capital and small letters, it must be observed that the Romans, generally speaking, wrote only in capital letters (litterae unciales), until in the latest period of antiquity the small letters came into use, which are now always employed in writing Latin.* Capital initials are at present used: (a) at the beginning of a verse, or at least of a strophe; (b) at the beginning of a new sentence, both in prose and in verse, after a full stop, and after a colon when a person's own words are quoted; (c) in proper names, and in adjectives and ad verbs which are derived from them, e. g., Latium, sermo Latinus, Latine loqui; (d) in words which express a title or office, such as Consul, Tribunus, and Senatus, but not in their derivatives.

8. The diaeresis (puncta diaereseos) is a sign to facilitate reading; it is put upon a vowel which is to be pronounced separately, and which is not to be combined with the preceding one into a diphthong, as in aër, aëris, aërius, poëta; and also in auraï, vitaï, since ai is only an ancient form for ae. In cases where the diphthong would. be foreign to the Latin language, the diaeresis is unnecessary, as in diei, Persei, because there can be no fear of any one pronouncing the ei as a diphthong; ferreus, too, does not require it, since in a Latin word no one will regard eu as a diphthong. But we must write Gaïus and silüa, when the consonants j and v are to be pronounced as vowels. The signs to indicate the length or shortness of a vowel or a syllable (and) were sometimes used by the ancients themselves.

*[The cursive character arose from a principle of rapidity, by which the letters are made to run on in continuous succession. Such modes of writing were no doubt common in very early times; and, as regards the Romans, we are not left to mere conjecture, as the British Museum contains an inscription of the kind on papyrus, which is referred to the secondor third century. The statement in the text, therefore, requires correction. (Key, Alphabet, p. 36.)]-Am. Ed.

[§ 14.] 1. A vowEL or a diphthong may by itself form a syllable, as in u-va, me-o; all other syllables arise from a combination of consonants and vowels. The Latin language allows only two consonants to stand at the end of a syllable, and three only in those cases where the last is S. At the beginning of a syllable, also, there can be no more than two consonants, except where the first is a c, p, or s, followed by muta cum liquida; and at the beginning of a word there never are three consonants, except in the case of sc, sp, and st being followed by an r or Ì; for example, do-ctrina, Ba-ctra, corru-ptrix, sce-ptrum, castra, magi-stri, I-sthmus; spretus, strenuus, scriba, splendor.

2. It often appears doubtful as to how a word is to be divided into syllables, and where the division is to be made at the end of a line, when the space does not suffice. The following rules, however, which are founded on the structure of the language, should be observed. 1. A consonant which stands between two vowels belongs to the latter, as in ma-ter. 2. Those consonants, which, in Latin or Greek, may together begin a word, go together in the division of syllables; e. g., pa-tris, and not patris, as tr occur at the beginning of tres.* In like manner, li-bri (brevis), i-gnis (gnomon), o-mnis, da-mnum (μváoμai), a-ctus, pun-ctum (ктñμa), ra-ptus, scri-ptus, pro-pter (Ptolemaeus), Ca-dmus (duwes), re-gnum (yvovs), va-fre (fretus), a-thleta (θλίβω), i-pse, scri-psi (ψαύω), Le-sbos (σβέννυμι), e-sca, po-sco (scando), a-sper, ho-spes (spes), pa-stor, faustus, i-ste (stare). The cases in which three consonants begin a syllable have been mentioned above. Whenever there occurs any combination of consonants which cannot stand at the beginning of words, they are treated according to the analogy of the rest. All combinations of muta cum liquida, for instance, go together, as most of them may commence a word; and we must therefore divide ara-chne, a-gmen, fra-gmentum, Da-phne, Pha-tnae, rhy-thmus, smara-gdus, and Lu-gdunum, since gd is to be

* [This mode of dividing is well intended, but perhaps too methodical; it occasions difficulty to learners, and has little use, but rather betrays some affectation. (Scheller, L. G., vol. i., 31, Walker's transl.)]-Am. Ed

B

treated like ct. 3. In compound words, the division must be made so as to keep the parts distinct, as inter-eram (not inte-reram), because the word is compounded of inter and eram. So, also, ab-utor, ab-rado, abs-condo, abs-temius (from temetum), sus-cipio (from the form subs), dis-quiro, et-iam, ob-latum; and red-eo, red-undo, prod-eo, and sed-itio, for the d, here inserted to prevent hiatus, must go with the preceding vowel, because, if added to the second, it would obscure the elements of the compound word. But when the component parts of a word are doubtful, or when the first word has dropped its termination to prevent hiatus, the syllables are divided as if the word were not a compound; e. g., po-tes (from pote or potis es), ani-madverto, and not anim-adverto, ve-neo (from venum eo), ma-gnanimus, am-bages, and lon-gaevus.*

CHAPTER III.

OF THE LENGTH AND SHORTNESS OF SYLLABLES.

[§ 15.] SYLLABLES are long or short, either by the nature of the vowel they contain, or they become long by their short vowel being followed by two or more consonants, that is, by their position. We shall first speak of the natural length and shortness of vowels.

1. All Diphthongs are long, and also all those single vowels which have arisen from the contraction of two into one, such as cōgo (from coago), mālo (from măvălo), tibicen (from tibiicen and tibia, but tubicen from tuba), bīgae (from bijugae), būbus and bōbus (from bõvibus), and so, also, dis for diis, gratis for gratiis, and nil for nihil.

Note. The preposition prae is commonly made short when compounded with a word which begins with a vowel, e. g., Ovid, Metam., vii., 131 : Quos ubi viderunt praeacutae cuspidis hastas. The reason for this peculiarity is explained in the rule following; but there is no other instance in the Latin language of a diphthong standing before a vowel. It occurs only in Greek proper names, in which, however, the diphthong remains long, as Aeolides Sisyphus, and Aeeta relictus, for the examples which are adduced as proofs of the diphthong being shortened (Ovid, Heroid., vi., 103, and Trist., iii., 12, 2) are not decisive.

* [The carrying out of this system would lead, it is apprehended, to some ludicrous results; as, for example, in such cases as fra-gmentum, a-gmen, &c. (Compare Journal of Education, vol. i., p. 94.)]-Am. Ed.

[The syllable pre being originally praï or praë, the latter of the two vowels is tacitly elided. Consult Anthon's Latin Prosody, ed. 1842, p. 25, not.]-Am. Ed.