VARRO’S LIFE AND WORKS
MARCUS TEREXTIUS VARRO was born in 116 B.C., probably at Reate in the Sabine country, where his family, which was of equestrian rank, possessed large estates. He was a student under L. Aelius Stilo Praeconinus, a scholar of the equestrian order, widely versed in Greek and Latin literature and especially interested in the history and antiquities of the Roman people. He studied philosophy at Athens, with Antiochus of Ascalon. With his tastes thus formed for scholarship, he none the less took part in public life, and was in the campaign against the rebel Sertorius in Spain, in 76. He was an officer with Pompey in the war with the Cilician pirates in 67, and presumably also in Pompey’s campaign against Mithradates. In the Civil War he was on Pompey’s side, first in Spain and then in Epirus and Thessaly.
He was pardoned by Caesar, and lived quietly at Rome, being appointed librarian of the great collection of Greek and Latin books which Caesar planned to make. After Caesar’s assassination, he was proscribed by Antony, and his villa at Casinum, with his personal library, was destroyed. But he himself escaped death by the devotion of friends, who concealed him, and he secured the protection of Octavian.
He lived the remainder of his life in peace and quiet, devoted to his writings, and died in 27 B.C., in his eighty-ninth year.
Throughout his life he wrote assiduously. His works number seventy-four, amounting to about six hundred and twenty books; they cover virtually all fields of human thought: agriculture, grammar, the history and antiquities of Rome, geography, law, rhetoric, philosophy, mathematics and astronomy, education, the history of literature and the drama, satires, poems, orations, letters.
Of all these only one, his De Re Rustica or Treatise on Agriculture, in three books, has reached us complete. His De Lingua Latina or On the Latin Language, in twenty-five books, has come down to us as a torso, only Books V. to X. are extant, and there are serious gaps in these. The other works are represented by scattered fragments only.
VARRO’S GRAMMATICAL WORKS
The grammatical works of Varro, so far as we know them, were the following:
De Lingua Latina, in twenty-five books, a fuller account of which is given below.
De Antiquitate Litterarum, in two books, addressed to the tragic poet L. Accius, who died about 86 B.C.; it was therefore one of Varro’s earliest writings.
De Origine Linguae Latinae, in three books, addressed to Pompey.
Peri Χαρακτήρων, in at least three books, on the formation of words.
Quaestmies Plautinae, in five books, containing viii interpretations of rare words found in the comedies of Plautus.
De Similitudine Verborum, in three books, on regularity in forms and words.
De Utilitafe Sermonis, in at least four books, in which he dealt with the principle of anomaly or irregularity.
De Sermone LciHno, in five books or more, addressed to Marcellus, which treats of orthography and the metres of poetry.
Disciplinae, an encyclopaedia on the liberal arts, in nine books, of which the first dealt with Grammatica.
The extant fragments of these works, apart from those of the De Lingua Latina, may be found in the Goetz and Schoell edition of the De Lingua Latina, pages 199-212; in the collection of Wilmanns, pages 170-223; and in that of Funaioli, pages 179-371 (see the Bibliography).
VARRO’S DE LINGUA LATIN A
Varro’s treatise On the Latin Language was a work in twenty-five books, composed in 47 to 45 B.C., and published before the death of Cicero in 43.
The first book was an introduction, containing at the outset a dedication of the entire work to Cicero. The remainder seems to have been divided into four sections of six books each, each section being by its subject matter further divisible into two halves of three books each.
Books II.-VII. dealt with the impositio vocabulorum, or how words were originated and applied to things and ideas. Of this portion, Books II.-IV. were probably an earlier smaller work entitled De Etymologia or the like; it was separately dedicated to one Septumius or Septimius, who had at some time, which we cannot now identify, served Varro as quaestor. Book II. presented the arguments which were advanced against Etymology as a branch of learning; Book III. presented those in its favour as a branch of learning, and useful; Book IV. discussed its nature.
Books V.-VII. start with a new dedication to Cicero. They treat of the origin of words, the sources from which they come, and the manner in which new words develop. Book V. is devoted to words which are the names of places, and to the objects which are in the places under discussion; VI. treats words denoting time-ideas, and those which contain some time-idea, notably verbs; VII. explains rare and difficult words which are met in the waitings of the poets.
Books VIII.-XIII. dealt with derivation of words from other words, including stem-derivation, declension of nouns, and conjugation of verbs. The first three treated especially the conflict between the principle of Anomaly, or Irregularity, based on consuetudo ‘popular usage,’ and that of Analogy, or Regularity of a proportional character, based on ratio ‘relation’ of form to form. VIII. gives the arguments against the existence of Analogy, IX. those in favour of its existence, X. Varro’s own solution of the conflicting views, with his decision in favour of its existence. XI.-XIII. discussed Analogy in derivation, in the wide sense given above: probably XI. dealt with nouns of place and associated terms, XII. with time, ideas, notably verbs, XIII. with poetic words.
Books XIV.-XIX. treated of syntax. Books XX.- XXV. seem to have continued the same theme, but probably with special attention to stylistic and rhetorical embellishments.
Of these twenty-five books, we have to-day, apart from a few brief fragments, only Books V. to X., and in these there are several extensive gaps where the manuscript tradition fails.
The fragments of the De Lingua Latina, that is, those quotations or paraphrases in other authors which do not correspond to the extant text of Books V.-X., are not numerous nor long. The most considerable of them are passages in the Noctes Atticae of Aulus Gellius ii. 25 and xvi. 8. They may be found in the edition of Goetz and Schoell, pages 3, 146, 192-198, and in the collections of Wilmanns and Funaioli (see the Bibliography).
It is hardly possible to discuss here even summarily Varro’s linguistic theories, the sources upon which he drew, and his degree of independence of thought and procedure. He owed much to his teacher Aelius Stilo, to whom he refers frequently, and he draws heavily upon Greek predecessors, of course, but his practice has much to commend it: he followed neither the Anomalists nor the Analogists to the extreme of their theories, and he preferred to derive Latin words from Latin sources, rather than to refer practically all to Greek origins. On such topics reference may be made to the works of Barwick, Kowalski, Dam, Dahlmann, Kriegshammer, and Frederik Muller, and to the articles of Wolfflin in the eighth volume of the Archiv fur lateiniscke Lexikographie, all listed in our Bibliography.
THE MANUSCRIPTS OF THE BE LINGUA LATINA
The text of the extant books of the De Lingua Latina is believed by most scholars to rest on the manuscript here first fisted, from which (except for our No. 4) all other known manuscripts have been copied, directly or indirectly.
1. Codex Làurentianus li. 10, folios 2 to 34, parchment, written in Langobardic characters in the eleventh century, and now in the Laurentian Library at Florence. It is known as F.
F was examined by Petrus Victorius and Iacobus Diacetius in 1521 (see the next paragraph); by Hieronymus Lagomarsini in 1740; by Heinrich Keil in 1851; by Adolf Groth in 1877; by Georg Schoell in 1906. Little doubt can remain as to its actual readings.
2. In 1521, Petrus Victorius and Iacobus Diacetius collated F with a copy of the editio princeps of the De Lingua Latina, in which they entered the differences which they observed. Their copy is preserved in Munich, and despite demonstrable errors in other portions, it has the value of a manuscript for v. 119 to vi. 61, where a quaternion has since their time been lost in F. For this portion, their recorded readings are known as Fv; and the readings of the editio princeps, where they have recorded no variation, are known as (Fv).
3. The Fragmentant Cassinense (called also Excerptum and Epitome), one folio of Codex Cassinensis 361, parchment, containing v. 41 Capitolium dictum to the end of v. 56; of the eleventh century. It was probably copied direct from F soon after F was written, but may possibly have been copied from the archetype of F. It is still at Monte Cassino, and was transcribed by Keil in 1848. It was published in facsimile as an appendix to Sexti Iulii Frontini de aquaeductu Urbis Romae, a phototyped reproduction of the entire manuscript, Monte Cassino, 1930.
4. The grammarian Priscian, who flourished about A.D. 500, transcribed into his De Figuris Numerorum Varro’s passage on coined money, beginning with multa, last word of v. 168, and ending with Nummi denarii decuma libella, at the beginning of v. 174. The passage is given in H. Keil’s Grammatici Latini iii. 410-411. There are many manuscripts, the oldest and most important being Codex Pansinus 7496, of the ninth century.
5. Codex Laurentianus li. 5, written at Florence in 1427, where it still remains; it was examined by Keil. It is known asf.
6. Codex Ilavniensis, of the fifteenth century; on paper, small quarto, 108 folia; now at Copenhagen. It was examined by B. G. Niebuhr for Koeler, and his records came into the hands of L. Spengel. It is known as II.
7. Codex Gothanus, parchment, of the sixteenth century, now at Gotha; it was examined by Regel for K. O. Mueller, who published its important variants in his edition, pages 270-298. It is known as G.
8. Codex Pansinus 7189, paper, of the fifteenth century, now at Paris; this and the next two were examined by Donndorf for L. Spengel, who gives their different readings in his edition, pages 661-718. It is known as a.
9. Codex Parisinus 6142, paper, of the fifteenth century; it goes only to viii. 7 declinarentur. It is known as 6.
10. Codex Pari sinus 7535, paper, of the sixteenth century; it contains only v. 1-122, ending with dictae. It is known as c.
11. Codex Vindobonensis lxiii., of the fifteenth century, at Vienna; it was examined by L. Spengel in 1835, and its important variants are recorded in the apparatus of A. Spengel’s edition. It is known as V.
12. Codex Basiliensis F iv. 13, at Basel; examined by L. Spengel in 1838. It is known as p.
13. Codex Guelferbytanus 896, of the sixteenth century, at Wolfenbiittel; examined by Schneidewin for K. O. Mueller, and afterwards by L. Spengel. It is known as M.
14. Codex B, probably of the fifteenth century, now not identifiable; its variants were noted by Petrus Victorius in a copy of the Editio Gryphiana, and either it or a very similar manuscript was used by Antonius Augustinus in preparing the so-called Editio Vulgata.
These are the manuscripts to which reference is made in our critical notes; there are many others, some of greater authority than those placed at the end of our list, but their readings are mostly not available. In any case, as F alone has prime value, the variants of other than the first four in our list can be only the attempted improvements made by their copyists, and have accordingly the same value as that which attaches to the emendations of editors of printed editions.
Fuller information with regard to the manuscripts may be found in the following:
Leonhard Spengel, edition of the De Lingua Latina (1826), pages v-xviii.
K. O. Mueller, edition (1833), pages xii-xxxi. Andreas Spengel, edition (1885), pages ii-xxviii. Giulio Antonibon, Supplemento di Lezioni Varianti ai libri de lingua Latina (1899), pages 10-23.
G. Goetz et F. Schoell, edition (1910), pages xi-xxxv.
THE LAURENTIAN MANUSCRIPT F
Manuscript F contains all the extant continuous text of the De Lingua Latina, except v. 119 trua quod to vi. 61 dicendofinit; this was contained in the second quaternion, now lost, but still in place when the other manuscripts were copied from it, and when Victorius and Diacetius collated it in 1521. There are a number of important lacunae, apart from omitted lines or single words; these are due to losses in its archetype.
Leonhard Spengel, from the notations in the manuscript and the amount of text between the gaps, calculated that the archetype of F consisted of 16 quaternions, with these losses:
Quaternion 4 lacked folios 4 and 5, the gap after v. — 162.
Quaternion 7 lacked folio 2, the end of vi and the beginning of vii., and folio 7, the gap after vii. 23. Quaternion 11 was missing entire, the end of viii and the beginning of ix.
Quaternion 15 lacked folios 1 to 3, the gap after x. 23, and folios 6 to 8, the gap after x. 34.
The amount of text lost at each point can be calculated from the fact that one folio of the archetype held about 50 lines of our text.
There is a serious transposition in F, in the text of Book V. In § 23, near the end, after qui ad humum, there follows ut Sabini, now in § 32, and so on to Septimontium, now in § 41; then comes demissior, now in § 23 after humum, and so on to ab hominibus, now in § 32, after which comes nominatum of § 41. Mueller, who identified the transposition and restored the text to its true order in his edition, showed that the alteration was due to the wrong folding of folios 4 and 5 in the first quaternion of an archetype of F; though this was not the immediate archetype of F, since the amount of text on each page was different.
This transposition is now always rectified in our printed texts; but there is probably another in the later part of Book V., which has not been remedied because the breaks do not fall inside the sentences, thus making the text unintelligible. The sequence of topics indicates that v. 115-128 should stand between v. 140 and v. 141 ; there is then the division by topics:
General Heading v. — 105
De Victu — v. — 105-112
De Vestitu — v. — 113-114, — 129-133
De Instrumenta — v. — 134-140, — 115-128, 141-183
Then also vi. 49 and vi. 45 may have changed places, but I have not introduced this into the present text; I have however adopted the transfer of x. 18 from its manuscript position after x. 20, to the position before x. 19, which the continuity of the thought clearly demands.
The text of F is unfortunately very corrupt, and while there are corrections both by the first hand and by a second hand, it is not always certain that the corrections are to be justified.
THE ORTHOGRAPHY OF THE DE LINGUA LATINA
The orthography of F contains not merely many corrupted spellings which must be corrected, but also many variant spellings which are within the range of recognized Latin orthography, and these must mostly be retained in any edition. For there are many points on which we are uncertain of Varro’s own practice, and he even speaks of certain permissible variations: if we were to standardize his orthography, we should do constant violence to the best manuscript tradition, without any assurance that we were in all respects restoring Varro’s own spelling. Moreover, as this work is on language, Varro has intentionally varied some spellings to suit his etymological argument; any extensive normalization might, and probably would, do him injustice in some passages. Further, Varro quotes from earlier authors who used an older orthography; we do not know whether Varro, in quoting from them, tried to use their original orthography, or merely used the orthography which was his own habitual practice.
I have therefore retained for the most part the spellings of F, or of the best authorities when F fails, replacing only a few of the more misleading spellings by the familiar ones, and allowing other variations to remain. These variations mostly fall within the following categories:
1. El: Varro wrote El for the long vowel I in the nom pi of Decl. II (ix. 80); but he was probably not consistent in writing El everywhere. The manuscript testifies to its use in the following: plebei (gen.; cf plebis vi. 91, in a quotation) v. 40, 81, 158, vi. 87; eidem (nom sing.) vii. 17 (eadem F), x. 10; scirpeis vii. “; Terentiei (nom.), vireis Terentieis (masc.), Terentieis (fern.) viii. 36; infeineiteis viii. 50 (changed to infiniteis in our text, cf. {in)finitam viii. 52); i(e)is viii. 51 (his F), ix. 5; iei (nom.) ix. 2, 35; hei re{e)i fer(re)ei de(e)i viii. 70; hinnulei ix. 28; utrei (nom pi.) ix. 65 (utre.I. F; cf utri ix. 65); (B)a(e)biei, B{a)ebieis x. 50 (alongside Caelii, Celiis).
2. AE and E: Varro, as a countryman, may in some words have used E where residents of the city of Rome used AE (cf v. 97); but the standard orthography has been introduced in our text, except that E has been retained in seculum and sepio (and its compounds: v. 141, 150, 157, 162, vii. 7, 13), which always appear in this form.
3. OE and U: The writing OE is kept where it appears in the manuscript or is supported by the context: moems and derivatives v. 50, 141 his, 143, vi. — 87; moenere, moenitius v. 141; Poenicum v. 113, viii. 65 his; poeniendo v. 177. OE in other words is the standard orthography.
4. VO UO and VU UU: Varro certainly wrote only VO or UO, but the manuscript rarely shows VO. or UO in inflectional syllables. The examples are novorn ix. 20 (corrected from nouum in F); nominatuom ix. 95, x. 30 (both -tiuom F); obliquom x. 50; loquontur vi. 1, ix. 85; sequontur x. 71; clivos v. 158; perhaps amburvom v. 127 (impurro Fv). In initial syllables VO is almost regular: volt vi. 47, etc.; volpes v. 101; volgus v. 58, etc., but vulgo viii. 66; Volcanus v. 70,’ etc.; voleillis ix. 33. Examples of the opposite practice are aequum vi. 71; duum x. 11; antiquus vi. 68; sequuntur viii. 25; conjluunt x. 50. Our text preserves the manuscript readings.
5. UV before a vowel: Varro probably wrote U and not UV before a vowel, except initially, where his practice may have been the other way. The examples are: Pacuius v. 60, vi. 6 (Catulus (Fv)), 94, vii. 18, 76, and Pacwius v. 17, 24, vii. 59; gen. Pacui v. 7, vi. 6, vii. — 22; Pacuium vii. 87, 88, 91, 102; compluium, impluium v. 161, and pluvia v. 161, compluvium v. 125; simpuium v. 124 his (sinipulum codd.); cf panuvellium v. 114. Initially: uvidus v. 24; uvae, uvore v. 104; uvidum v. 109.
6. U and I: Varro shows in medial syllables a variation between U and I, before P or B or F or M plus a vowel. The orthography of the manuscript has been retained in our text, though it is likely that Varro regularly used U in these types:
The superlative and similar words: ulbissumum viii. — 75; fnigalissumus viii. 77; c(a)esi(s)sumus viii. 76; intumus v. 154; maritumae v. 113; melisxumum viii. — 76; optumum vii. 51; pauperrumus viii. 77; proxuma etc v. 36, 93, ix. 115, x. 4, 26; septuma etc. ix. — 30, x. 46 ter; Septumio v. 1, vii. 109; superrumo vii. — 51; decuma vi. 54. Cf proximo optima maxima v. — 102, minimum vii. 101, and many in viii. 75-78.
Compounds of -fex and derivatives: pontufex v. 83, pontufces v. 83 (F for pontifces); artufces ix. 12; sacruficiis v. 98, 124. Cf pontifces v. 23, vi. 54, etc.; artifex v. 98, ix. Ill, etc.; sacrifcium vii. 88, etc.
Miscellaneous words: — monumentum v. 148, but monimentum etc v. 41, vi. 49 his; mancupis v. 40, but mancipium etc v. 163, vi. 74, 85; quadrupes v. 34, but quadripedem etc vii. 39 his, quadriplex etc x. 46 etc., quadripertita etc v. 12 etc.
7. LUBET and LIBET: Varro probably wrote lubet, lubido, etc., but the orthography varies, and thç manuscript tradition is kept in our text: — lubere lubendo vi. 47, lubenter vii. 89, lubitum ix. 34, lubidine x. 56; and libido vi. 47, x. 60, libidinosus Libentina Libitina vi. 47, libidine x. 61.
8. H: Whether Varro used the initial H according to the standard practice at Rome, is uncertain. In the country it was likely to be dropped in pronunciation; and the manuscript shows variation in its use. We have restored the H in our text according to the usual orthography, except that irpices, v. 136 his, has been left because of the attendant text. Examples of its omission are Arpocrates v. 57; Ypsicrates v. 88; aedus ircus v. 97; olus olera v. 108, x. 50; olitorium v. 146; olitores vi. 20; ortis v. 103, ortorum v. 146 his, orti vi. 20; aruspex vii. 88. These are normalized in our text, along with certain other related spellings: sepulchrum vii. 24 is made to conform to the usual sepulcrum, and the almost invariable nichil and nichili have been changed to nihil and nihili.
9. X and CS: There are traces of a writing CS for X, which has in these instances been kept in the text: arcs vii. “ (ares F); acsitiosae (ac sitiose F), acsitinsa (ac sitio a- F) vi. 66; ducs (duces F) x. 57.
10. DOUBLED CONSONANTS: Varro’s practice in this matter is uncertain, in some words. F regularly has littera (only Uteris v. 3 has one T), but obliterata (ix. 16, -atae ix. 21, -amt v. 52), and these spellings are kept in our text. Communis has been made regular, though F usually has one M; casus is invariable, except for de cassu in cassum viii. 39, which has been retained as probably coming from Varro himself. Iupiter, with one P, is retained, because invariable in F; the only exception is Iuppitri viii. 33 (iuppiti F), which has also been kept. Numo vi. 61, for nummo, has been kept as perhaps an archaic spelling. Decusis ix. 81 has for the same reason been kept in the citation from Lucilius. In a few words the normal orthography has been introduced in the text: grallator vii. 69 his for gralator, grabatis viii. 32 for gr a battis. For combinations resulting from prefixes see the next paragraph.
11. CONSONANTS OF PREFIXES: Varro’s usage here is quite uncertain, whether he kept the unassimilated consonants in the compounds. Apparently in some groups he made the assimilations, in others he did not. The evidence is as follows, the variant orthography being retained in our text:
Ad-c-: always acc-, except possibly adeensos vii. 58 (F, for acensos F).
Ad-f-: always aff-, except adfuerit vi. 40.
Ad-l-: always all-, except adlocutum vi. 57, adlucet vi. 79, adlatis (ablatis F) ix. 21.
Ad-m-: always adm-, except ammonendum v. 6, amministrat vi. 78, amminicula vii. 2, amminister vii. 34 (F, for adm- F).
Ad-s-: regularly ass-, but also adserere vi. 64, adsiet vi. 92, adsimus vii. 99, adsequi viii. 8, x. 95 od- significare often (always except assignificant vii. 80), adsumi viii. 69, adsumat ix. 42, adsumere x. 58.
Ad-sc-, ad-sp-, ad-st-: always with loss of the D, as in ascendere, ascribere, ascriptos (vii. 57), ascriptivi (vii. 56), aspicere, aspectus, astans.
Ad-t-: — always ait-, except adtributa v. 48, and possibly adtinuit (F, but att- F) ix. 59-
Con-l-, con-b-, con-m-, con-r-: — always coll-, comb-, comm,-, corr-.
Con-p-: always comp-, except conpernis ix. 10.
Ex-f-: always eff-, except exfluit v. 29.
Ex-s-: exsolveret v. 176, exsuperet vi. 50, but exuperantum vii. 18 (normalized in our text to ex su per ant uni).
Ex-sc-: exculpserant v. 143.
Ex-sp-: always expecto etc vi. 82, x. 40, etc.
Ex-sq-: regularly Esquiliis; but Exquilias v. 25, Exquiliis v. 159 (Fv), normalized to Esq- in our text.
Ex-st: extat v. 3, vi. 78; but exstat v. 3, normalized to extat in our text.
In-l-: usually ill-, but indicium vi. 88 his, 93 (illici- turn F), 94, 95, inliceret vi. 90, inliciatur vi. 94; the variation is kept in our text:
In-m-: always imm-, except in {in)mutatis vi. 38, where the restored addition is unassimilated to indicate the negative prefix and not the local in.
In-p-: always imp-, except inpos v. 4 his (once ineos F), inpotem v. 4 (inpoientem F), inplorat vi. 68.
Ob-c-, ob-f-, ob-p-: always occ-, off-, opp-.
Ob-t-: always opt-, as in optineo etc vii. 17, 91 J x. 19, optemperare ix. 6.
Per-l-: pellexit vi. 94, but perlucent v. 140. xxii Sub-c-, sub-f-, sub-p-: always suce-, suff-, supp-, except subeidit v. 116.
Sub-s- and subs- + consonant: regularly sus- + consonant, except subscribunt vii. 107.
Sub-t-: only in suptilius x. 40.
Trans-l-: in tralutum vi. 77, vii. 23, 103, x. 71; tralaticio vi. 55 (trmilatio Fv) and trajislaticio v. 32, vi. — 64 (traiislatio F, tranlatio Fv), translaticiis vi. 78.
Trans-v-: in travolat v. 118, and transversus vii. 81, x. 22, 23, 43.
Trans-d-: in traducere.
12. DE and DI: The manuscript has been followed in the orthography of the following: directo vii. 15, dirigi viii. 26, derecti x. 22 his, derigimtur derectorum x. 22, derecta directis x. 43, directas x. 44, derigitur x. 74; deiunctum x. 45, deiunctae x. 47.
13. SECOND DECLENSION: Nom sing, and acc sing, in -uom and -uum, see 5.
Gen sing, of nouns in -lus: Varro used the form ending in a single I (cf viii. 36), and a few such forms stand in the manuscript: Muci v. 5 (tnuti F); Pacui v. 7, vi. 6, vii. 22; Mani vi. 90; Quinti vi. 92, Ephesi viii. — 22 (ephesis F), Plauti et Marci viii. 36, dispendi ix. — 54 (quoted, metrical; alongside dispendii ix. 54). The gen in II is much commoner; both forms are kept in our text.
Norn, pi., written by Varro with El (cf ix. 80); examples are given in 1, above.
Gen pi.: The older form in -um for certain words (denarium, centumvirum, etc.) is upheld viii. 71, ix. 82, 85, and occurs occasionally elsewhere:
Velabrum v. 44, Querquetulanum v. 49, Sabinum v. 74, etc.
Dat.-abl pi., written by Varro with EIS (cf ix. 80); examples are given in 1, above, but the manuscript regularly has IS.
Dat.-abl pi of nouns ending in -ius, -ia, -ium, are almost always written IIS; there are a few for which the manuscript has IS, which we have normalized to IIS: Gabis v. 33, (Es)quilis v. 50, hostis v. 98, Publicis v. — 158, Faleris v. 162, praeverbis vi. 82 (cf praeverbiis vi. — 38 his), my si er is vii. 31 (cf mysteriis vii. 19), miliaris ix. 85 (militari s F).
Deus shows the following variations: Norn pi de(e)i viii. 70, dei v. 57, 58 his, 66, 71, vii. 36, ix. 59, dii v. 58, 144’, vii. 16; dat.-abl pi dels v. 122, vii. 45, diis v. 69, 71, 182, vi. 24, 34, vii. 34.
14. THIRD DECLENSION: — The abl sing, varies between E and I: supellectile viii. 30, 32, ix. 46, and supelleciili ix. 20 {-lis F); cf also vesperi (uespert- F) and vespere ix. 73.
Norn, pi., where ending in IS in the manuscript, is altered to ES; the examples are mediocris v. 5; partis v. — 21, 56; ambonis v. 115; urbis v. 143; aedis v. 160; compluris vi. 15; Novendialis vi. 26; auris vi. 83; dis- parilis viii. 67; lends ix. 31; ornnis ix. 81; dissimtlis ix. 92.
Gen pi in UM and IUM, see viii. 67. In view of dentum viii. 67, expressly championed by Varro, Veientum v. 30 (uenientum F), caelestum vi. 53, Quiritum vi. — 68 have been kept in our text.
Acc pi in ES and IS, see viii. 67. Varro’s distribution of the two endings seems to have been purely empirical and arbitrary, and the manuscript readings have been retained in our text.
15. FOURTH DECLENSION: — Gen sing.: — Gellius, Nodes Atticae iv. 16. 1, tells us that Varro always used UIS in this form. Nonius Marcellus 483-494 M. cites eleven such forms from Varro, but also sumpti. The De Lingua Latina gives the following partial examples of this ending: usuis ix. 4 (suis F), x. 73 (usui F), casuis x. 50 (casuum F), x. 62 (casus his F). Examples of this form ending in US are kept in our text: fructus v. — 34, 134, senatus v. 87, exercitus v. 88, panus v. 105, domus v. 162, census v. 181, motus vi. 3, sonitus vi. 67 sensus vi. 80, uSUS viii. 28, 30 his, — ix. 76, manus ix. 80.
Gen pi.: For the variation between UUM and UOM see 4, above. The form with one U is found in tribum v. 56, ortum v. 66, manum vi. 64 (manu F), magistratum viii. 83 (-tus F), declinatum x. 54; these have been normalized in our text to UUM (except manum, in an archaic formula). Note the following forms in the manuscript: cornuum v. 117, declinatuum vi. — 36 (-tiuum Fu), x. 31, 32, 54, sensuum vi. 80; tribuum vi.86; fructuum ix. 27; casuum ix.77, x. 14,23, manuum ix. 80, nominatuom (-tiuom F) ix. 95, x. 30, nomination x. 19.
16. HETEROCLITES: There are the following: gen sing, plebei v. 40, 81,158, vi. 87, and plebis vi. 91; nom sing, elephans and acc pi elepkantos vii. 39; abl sing. Titano vii. 16; abl pi vasis v. 121, poematis vii. 2, 36, viii. 14, and poematibus vii. 34.
17. GREEK FORMS: There are the following: acc sing, analogian ix. 1, 26, 33, 34, 15, 49, 76, 79, 105, 113, 114, but also analogiam ix. 90, 100, no, x. 2, and analogia(m) ix. 95, 111. Acc sing. Aeihiopa viii. 38 (ethiopam F). Norn pi. Aeolis v. 25, 101, 102, 175, Athenaiis viii. 35.
18. FORMS OF IS AND IDEM: The forms in the manuscript are kept in our text; there are the following to be noted:
Nom sing masc.: idem often; also eidem vii. 17 (eadem F), x. 10.
Nom pl.: ii v. 26, ix. 2; iei ix. 2, 35; idem ix. 19.
Dat.-abl pl.: eis vi. 18, vii. 102, ix. 4, x. 8; ieis viii. — 51 (his F, but assured by context), ix. 5; is vii. 5 (dis F); iisdem vi. 38; isdem vii. 8 (hisdem F), viii. 35 his (hisdem F).
19. QUOM and CUM etc.: Varro wrote quom, quor, quoins, quoi, and not cum, cur, cuius, cm, though the latter spellings are much commoner in the manuscripts, the readings of which are kept in our text. Quom is not infrequent, being found vi. 42, 56, vii. 4, 105, viii. 1, x. 6, and in other passages where slight emendation is necessary. Quor is found only corrected to cur, viii. 68, 71, and hidden under quorum corrected to quod, viii. 78. Quoins is written viii. 44, ix. — 43, x. 3, and in other passages where emendation is necessary. Quoi nowhere appears, unless it should be read for qui vi. 72, and quoique for quoque ix. 34, adopted in our text.
Both qui and quo are used for the abl sing, of the relative, and quis and quibus for the dat.-abl pl., and similar forms for quidam, hi quo is used with a plural antecedent of any gender: v. 108, vi. 2, 55, 82, vii. 26, viii. 83, ix. 1, x. 8, 41.
20. ALTER and NEUTER: Gen alii ix. 67 is found as well as allerius ix. 91 5 neutri ix. 62, neutraie) x. — 73, as well as neutrius ix. 1; dat fern, aliae x. 15.
21. CONTRACTED PERFECTS: Only the contracted perfects arc found, such as appellarunt v. 22 etc., declinarit v. 7, aberraro v. 13, appelassent ix. 69, curasse vii. — 38, consuerunt consuessent ix. 68, consuerit ix. 14 his; exceptions, norissent vi. 60, auspicaverit vi. 86 (quoted), nuncupavero vii. 8 (quoted), vitaverunt x. 9-
INTRODUCTION
Similarly, the V is omitted after I, as in praeteru ix. 7, prodierunt v. 13, expediero viii. 24, etc.; exception, quivero v. 5 (F, for quiero F).
22. PONO in Perfect: The text always has posui and its forms, except twice, which we have standardized: imposiverunt viii. 8, imposierint ix. 34.
23. GERUNDIVES: Varro used the old form of the gerundive and gerund with UND in the third and fourth conjugations, but the forms have mostly been replaced by those with END. The remaining examples of the older form areferundo v. \0%, ferundum vi. 9Q,faciundo vii. 9, qnaerundae vii. 35, reprehendiindi ix. 12, reprehendundus ix. 93.
24. VERSUS: The older forms vorto, vorti, tor sus are not found in the manuscript. The adverbial compounds of versus have (with one exception) been retained in our text as they appear in the manuscript: susus versus v. 158, susum versus ix. 65; deorsum, susum v. 161; rursus vi. 46, 49, ix. 86; deosum versus ix. 86; prosus and rusus (rostts F) x. 52.