Rodgers's recent OCT edition of Columella's De re rustica has deservedly been hailed as offering a radically improved text, with a hint that it may occasionally have gone a little too far in attempting to emend the paradosis ope ingenii.Footnote 1 While it is true that Rodgers's is the most readable text to date, and while some of his interventions may, nevertheless, be less successful than others, more than once he seems to have gone not far enough, leaving a number of textual issues untackled. Book 10, a didactic poem on horticulture, is the case in point: edited more often by specialists on agricultural writings than those on Latin poetry, the poem still arguably remains even more rustic than it was originally created.Footnote 2 In what follows I propose some thirty emendations in the text of the tenth book of the De re rustica, with varying levels of uncertainty.
LINES 9–10
Columella begins his account of horticulture by explaining what kind of soil is suitable for a vegetable garden (6–12):Footnote 3
The syntactic role of habilis natura (9) has been interpreted in two ways. One option is to take it as a second subject of praebeat (6, along with ager, 7): atque (9) recommends such a construal, but it makes the overall syntax rather heavy, and the sense is less than ideal (‘the nature of the soil’ cannot normally be said to provide a spot for a garden).Footnote 4 The alternative, clearly superior and followed by the majority of editors and translators, has been to take habilis (with implied est) as the predicate of natura. This I believe is the correct construal, but in the transmitted version of the text it faces two obstacles. First, atque ‘and’ normally links coordinate syntactic units, whereas here it has to mark the transition from a jussive clause (praebeat) to an indicative one (habilis [sc. est]), which may rather be seen as explaining or elaborating the preceding prescription.Footnote 5 Second, the lack of an expressed copula invites the reader to take habilis natura as a noun phrase, and it is only at the end of the following quae clause that he realizes that habilis must be interpreted as the predicate. These obstacles can be avoided by replacing atque with est (→ et → atque).Footnote 6 Before we move on, it may be worth briefly revisiting an old conjecture that has been ignored by the more recent editors. All modern editions unanimously print uuida at line 10, despite the fact that the proper term for moist soil would be umida, conjectured by Lipsius.Footnote 7 In fact, uuidus may not be an appropriate term to describe the quality of the soil suitable for garden plants, as it seems to imply an excessive amount of liquid: the soil should be moist, not dripping wet.Footnote 8
LINE 43
Columella describes the tasks that need to be performed around the time of the autumnal equinox (41–6):
As Boldrer discusses, line 43 has been interpreted in three ways.Footnote 9 First, pomis has been construed with satur (‘replete with fruits’), but the distance between the two words speaks strongly against this, whereas satur works well enough in isolation (and perhaps plays on the contrast with sitiens [41], likewise absolute). Second, pomis has been taken with sordidus, but again the enjambement is harsh, and the sense awkward (‘stained with fruits’). Third, scholars have interpreted the line to refer to a crown of fruits: this is clearly the sense we should expect, but the syntax cannot be construed to produce it.Footnote 10 In other words, we need a word to govern pomis that would convey the idea of Autumnus being crowned with fruits. One option might be to replace sua (which, with quassans, is rather superfluous) with a passive participle, but none I can think of produces the required sense.Footnote 11 The alternative is to replace quassans with cingens: ‘crowning his head with fruits’; for the phrasing, we may compare lines 256–7 iam uersicoloribus anni | fetibus alma parens cingi sua tempora gaudet.Footnote 12
LINE 53
Columella associates the beginning of the winter season with two astronomical events, the evening setting of the Corona Borealis and the morning setting of the Pleiades (50–4):
I fail to see how line 53 can possibly make sense. The first three words are intelligible: the constellation ‘disappears in the blue sea’; but what does uertice mundi mean? A number of unconvincing interpretations have been advanced,Footnote 13 but in any event the presence of two bare ablatives (aequore and uertice) is a clear indication that the text is corrupt (cf. below on lines 152 and 294). The Virgilian model is unfortunately of little help (G. 1.221–2 ante tibi Eoae Atlantides abscondantur | Cnosiaque ardentis decedat stella Coronae), but Columella's own references to the evening settings of other constellations in the next prose book (11.2.34 Vergiliae uespere celantur, 11.2.36 Suculae se uespere celant, 11.2.37 Canis se uespere celat) make it likely that uertice is an error for uespere. Although it might be marginally possible to make sense of uespere mundi (‘the evening setting upon the sky’), and although some might be happy to have a local (aequore) and a temporal (uespere) ablative in the same clause, I tentatively propose further changes. If we accept uespere, a shadow of doubt is cast over aequore caeruleo, which allows us to reinterpret mundi as in undi<s>. The next step is to restore, in place of aequore caeruleo, a genitive dependent on undis, and a conceivable option is aequoris occidui (ſoc → coe, cıd → rul): ‘in the evening the Corona disappears in the waves of the western sea’. I admit that, in view of the number of changes involved, this may not be an indisputable solution, but all the changes are palaeographically plausible, and they do restore a meaningful text.
LINE 59
Columella argues that the extant human race is not the one created by Prometheus from clay but a subsequent generation produced by Deucalion from stones (58–67):
Editors are unconcerned by the shape of line 59, but ista and altera cannot both be right at the same time, as only one is needed to provide a counterpart for altera at the beginning of the next line. Suspicion falls on altera, since ista is necessary to indicate that the line is speaking about the earth, referred to as matri … falsae in the preceding line. In fact, the two medieval manuscripts (SA) are defective at this point, as they omit cretae at the end of line 59 and altera at the beginning of line 60, and it is only the younger manuscripts (R) that supply the omission; while it is not impossible that the recentiores (R) preserve the paradosis, their reading may be no more than a conjecture (or worse). This suspicion is corroborated by the fact that the medieval manuscripts read Promethei rather than Prometheae, which may indicate that the original reading was limi (the term used by Horace at Carm. 1.16.13–15 fertur Prometheus addere principi | limo coactus particulam undique | desectam), while cretae could be an interpolated gloss. If altera (59) is corrupt, it probably replaces an epithet of genetrix. I have thought of aspera (‘unloving’?), but perhaps infera would be more pointed (almost ‘infernal’). In the lines that follow, Columella makes a point of demonstrating that the current race of humans cannot have been created by the earth, because during the flood it was completely submerged, including the underworld. We, Columella insists, have different ancestors: rocks taken from high mountains (montibus altis, 66). It would make sense if the opposition of the two races were not merely based on the difference of material (rock as opposed to earth) but also included a hierarchical dimension (low as opposed to high, implying closeness to the underworld as opposed to heaven).
LINES 151–2
Columella instructs his readers that, if the garden lacks an ample supply of irrigation water, seedlings should be grown on specially constructed dry beds (150–4):
One problem is nemoris (151): on the one hand, uertex nemoris is not a plausible turn of phrase; on the other, the point of dumosis is that nothing naturally grows on those hills or slopes other than thorns (a grove would require a source of irrigation). I have considered writing umoris (umoris … riui would be a variation on the standard idiom riuus aquae), but montis (mōtis → <ne>moris, interpolated from the preceding nec) seems altogether preferable in sense:Footnote 14 the garden is spread over a mountain's slope (collibus), but there is no stream running from the mountain's top.Footnote 15 Another problem is the accumulation of two ablatives absolute in line 152 (cf. above on line 53 and below on line 294).Footnote 16 The easiest way to solve the issue is by adding -que after cumulatis (cf. below on line 261), which will produce a reasonable description of the process of building this kind of plant bed: first a mound of rubble should be amassed (to serve as drainage?), then clods of earth should be piled on it.Footnote 17
LINE 158
Columella explains when seedlings should be transplanted (155–8):
A long-established problem is the use of reflective se with nubere at line 158.Footnote 18 The recentiores (R) eliminate it by writing denubere (which is problematic for other reasons), Holford-Strevens conjectures iam nubere.Footnote 19 Yet nubere itself is questionable too, since only a few lines later transplanted seedlings are metaphorically described as ‘adopted children’ (priuignasque rogat proles, 163). Accordingly, Richter prints sinere ubera plantis (‘den eingesetzten Pflänzchen die Brüste zu geben’), but the expression seems unparalleled.Footnote 20 I suggest we should write se iungere, for which an excellent parallel is provided by lines 194–5 dum cupit et cupidae quaerit se iungere matri | et mater facili mollissima subiacet aruo (in reference to planting lettuce).
LINE 165
Columella continues with his advice on seedlings (163–5):
At first he speaks about seedlings as a garland with which the earth should be wreathed (uiridi redimite parentem | progenie, 164–5), but then they become the earth's own hair that needs to be looked after (tu cinge comas, tu dissere crines, 165). The second imperative singular appears to produce a clear sense (‘arrange her locks’), but the point of cinge is less obvious: while cingere can be used of hair, it means ‘to crown’ and requires a specification of what hair is crowned with.Footnote 21 I suggest we should write finge ‘dress her hair’.Footnote 22 While the two imperative clauses might in principle be largely synonymous, it seems more likely that they refer to two different hairstyles. Note that crinis is properly ‘a lock’, and that the prefix of dissere may suggest a relatively free arrangement;Footnote 23 by contrast, finge could imply a more ‘fixed’ coiffure.Footnote 24
LINE 177
Columella instructs that it is now time to transplant flower seedlings (174–7):
It is awkward that the (implied) object of disponat (sc. colores, 177) should have two relative clauses, introduced by the same pronoun in the same form (quos at lines 176 and 177) but entirely different and unrelated in content. It will make better sense to write quas for the second quos and to take its clause to refer to plantis: ‘let the gardener plant those flowers with seedlings which he has grown from seed’.
LINE 184
Columella describes different varieties of lettuce (181–4):
At line 184 both haec and sua are superfluous: the anaphoric haec after such a short relative clause with the same subject is unnecessary, whereas the possessive pronoun sua is in conflict with the possessive genitive Cappadocae … gentis (contrast seruat flauae cognomina cerae, 417). It seems clear that haec sua conceals an adjective, though it is more difficult to decide which, as is often the case with epithets. I have thought of incluta (ınc → hıc → hec, lıta → ſua?), but hospita (hos → hoc → hec, pıta → ſua?) might be somewhat more pointed, stressing the contrast with the two former varieties named after a Roman.Footnote 25
LINE 203
Columella describes the effect of the spring on the sea (200–3):
I fail to see how reserat ‘opens’ (203) can possibly be right. It is true that reserare can be used in contexts speaking of childbirth in reference to ‘opening the womb’, but this cannot work with pontum.Footnote 26 In the present context the sense could only be ‘opens the sea for navigation (or swimming)’; while as such this is not an impossible idea, in both what precedes and what follows, Columella is speaking about new creatures being produced in the sea (natantibus refers to neither sailors nor bathers). I suggest we should write reserit ‘repopulates’: the verb can be used to refer to ‘resowing’ a field, and in a poem on agriculture this would not be an inappropriate metaphor.Footnote 27
LINES 218, 220 AND 224
Columella rejects the ambition of a didactic poet concerned with matters of a cosmic scale (217–24):
The first problem is that mouentem (218) fails to produce plausible sense with orgia and foedera (219). Foster and Heffner render it with ‘explore’, White with ‘evoking’, but neither is a valid translation.Footnote 28 With orgia, the verb could perhaps mean something like ‘to perform’, but this cannot work for foedera.Footnote 29 I think monentem would be an undeniable improvement: ‘teaching the sacred mysteries of nature and the hidden laws of the universe’. There is, however, another problem, concerning the articulation of the passage as a whole: the clause of inpulit (218) is disproportionately concise if not abrupt, whereas that of extimulat (220) is extremely convoluted. The ellipsis of inpulit ad rerum causas (‘has pushed towards the causes of things’) borders on being incomprehensible, and one also has a feeling that rerum causas and sacra … orgia naturae belong together. One option might be to substitute the gerund monendum, which would have causas, orgia and foedera as its direct objects; however, I cannot parallel ad with a gerund clause in poetry. I suggest we should rather write the gerundive monenda, likewise taking it with causas, orgia and foedera.Footnote 30 To avoid the clauses of inpulit and extimulat being connected asyndetically, we have then also to change the latter, either to exstimulans or, which seems to produce better style, to et stimulat. Finally, there are problems with the last line of the quoted passage. To begin with, the hiatus between Euhie and Euhie (224) is a cause for concern,Footnote 31 as is the exclamation itself: Euhie normally addresses Dionysus, whereas Paean at the end of the hexameter should refer to Apollo, as it clearly does before the caesura.Footnote 32 In fact, the manuscripts read et ehyie ehyie paen for the second half of the verse: Euhie may not as such be an implausible correction for ehyie, but it is far from certain. I suggest that a more appropriate exclamation would be ieie ieie, which can be paralleled in Varro Atacinus, fr. 127.2 Hollis hortantes ‘o Phoebe’ et ‘ieie’ conclamarunt (for which the witnesses read, instructively, loliscona clamarunt and locolicon aclamarunt). A conceivable objection could be that in Greek texts ἰήιε seems never to be repeated twice in a row, and that the final epsilon never seems to be elided (which may reflect the derivation from ἰὴ ἰή); but this need not hold true for Latin. If we do wish to conform to Greek usage, though, we could write ieie o ie Paean or ieie Delie Paean.Footnote 33 Once we have fixed the second half of the line, and established that there is no need for te (humanist) after et, we may wonder what te is doing in the first half. I fail to see how te can be part of the exclamation itself, and while it could in principle depend on frementem, such a construal is rather awkward. I suggest that we should write io, or perhaps ie (ἰή), instead (compare Ov. Ars am. 2.1 dicite ‘io Paean!’ et ‘io’ bis dicite ‘Paean!’). The verse 224 may thus originally have read along the lines of ‘Delie ie Paean’ et ‘ieie o ie Paean’, though unfortunately certainty seems unobtainable.
LINE 261
Columella lists a number of spring flowers (258–62):
At lines 260–1 ingenuo confusa rubore and uirgineas adaperta genas are coordinate, and the asyndeton is harsh and pointless; we should add -que after uirgineas (cf. above on line 152).
LINES 294 AND 296
Columella advises his readers to pluck daffodils and pomegranate flowers at either sunrise or sunset (294–7):
The passage hosts a number of problems. First, line 294 has two bare and uncorrelated ablatives with a temporal force (iubare exorto and nocte suprema), which, moreover, contradict each other: the sun does not rise in the last part of the night. As Stat. Theb. 3.683–4 iam nocte suprema | ante nouos ortus shows, the sense we need is ‘in the last part of the night, before the sunrise’, and it can be obtained by writing sub for uel, which will also eliminate one of the bare ablatives (ſub may have been omitted before ıubare owing to homoearchon, after which uel would have been supplied to fix the metre).Footnote 34 Second, sicubi ‘if at any place’ (296) fails to produce a plausible sense; translators usually take it to mean ‘wherever’ (which it does not), but the question remains why you should harvest daffodils and pomegranate flowers only where marjoram grows. I suggest that the line is an alternative reference to the evening: aut ubi, ‘or when the marjoram is casting a shade’.Footnote 35 This brings us to the third problem, the sense of praetexit (296): editors take it to mean ‘to extend’, and this is indeed the sense we expect, but the verb cannot convey it. We should write protendit, ‘the marjoram stretches forth its shade’.Footnote 36
LINE 310
Columella announces that it is time to sell roses and marigolds at the town market (306–10):
First, a minor issue of interpretation: White translates pressa (307) as ‘little’, but, even though pressus can mean ‘moderate, restrained’ (OLD s.v. pressus 1 6), a flower-seller will use a basket as large as he can carry; Boldrer offers ‘dilatato’, but it is doubtful that pressus can have this sense.Footnote 37 I think pressa must mean here ‘heavy with, weighed down by’.Footnote 38
Now the main problem, the exact point of grauis (310). It is clear that the line reworks Verg. Ecl. 1.35 non umquam grauis aere domum mihi dextra redibat, and it may appear natural to construe grauis with aere in our context as well; yet this will be strained in syntax as well as repetitive in sense, since aere properly belongs with plenos. Alternatively grauis can be taken absolutely, but this will produce the wrong sense: no matter how much money the flower-seller may have made, it can only be described as ‘heavy’ in relative terms, and certainly not in comparison with the basket he brought to the market. This objection is substantiated by the Virgilian parallel, where, moreover, grauis aere is specifically said of dextra, as well as by its imitation in Mor. 80 inde domum ceruice leuis, grauis aere redibat: a farmer returning from the market can be said to be loaded with money, but in absolute terms he is travelling light. This I suggest should likewise be the point in our context, grauis being an error for leuis: the flower-seller lightly carries his pockets full of money. The corruption may be due to scribal recollection of the Virgilian passage (or simply be a polar error); cf., for example, the opposite corruption leues for graues at Sen. Herc. f. 1117.Footnote 39
LINE 313
Columella advises which plants and vegetables should be harvested and sold around the time of the summer solstice (311–15):
Line 313 refers to the sun entering the constellation of Cancer, but the phrasing raises doubts: while it is true that haurire can be used of fire, a constellation can hardly be said to be ‘consumed’ by flames.Footnote 40 The idea we should expect is that of the sun scorching or setting aflame the constellation it is passing through, not destroying it.Footnote 41 This sense can be obtained by writing usserit.Footnote 42
LINES 333–6
Columella describes two kinds of pest caterpillar, one that damages vines and willows and another that infests garden plants (331–6):
The passage presents a number of minor and easily eliminated problems. First, the metaphorical reference to the damage infected by the second kind of caterpillar (exurit ‘scorches’) alludes to the folk etymology behind the alternative form of eruca (333): uruca (urica), which should be restored in the text.Footnote 43 Second, as is shown by lines 335–6, the uruca does not eat seeds but plants; it is true that semen can refer to ‘anything planted by way of propagation (a slip, cutting, etc.)’ (OLD s.v. 3), yet, apart from the potential ambiguity, it is doubtful that the uruca should be imagined to be so selective as only to damage seedlings; germina ‘shoots’ would produce a far more fitting sense.Footnote 44 Third, the pyramid of subordinate adjective clauses (e quibus, 332; quos, 334; quae, 335) is inelegant even by Columella's standards; in addition, relating super ingrediens (‘stepping upon’, 334) to hortos produces a less than ideal sense (contrast serpitque uruca per hortos ‘creeps through the garden’). The object of super ingrediens should be germina: this can be achieved by replacing quos with quae (note that in fact the two medieval manuscripts [SA], and most of the recentiores [R], read quo, not quos), while changing quae to haec at the beginning of the next line.Footnote 45 Fourth, though less certain, the standard epithet of uenenum is taetrum, not triste (336): perhaps we should restore taetro here as well.Footnote 46
LINE 368
Columella compares the effect of menstrual blood on pest caterpillars to that of Medea's magic on the serpent which guarded the Golden Fleece (367–8):
The obvious problem is that Iolcus, about a thousand miles away from Colchis, could not witness the serpent fall asleep. One option has been to take Iolcos (368) as a metonymy for the Argonauts, but there are a number of obstacles: first, the name of a city could stand for its actual inhabitants but hardly for people who left it; second, in fact far from all Argonauts came from Iolcus; third, Jason was the only Argonaut present when Medea charmed the serpent.Footnote 47 The alternative is to interpret Iolcos, or rather Iolcus as recently argued by Lucarini, as an adjective periphrastically referring to Jason (‘the Iolcian’).Footnote 48 It is, however, highly doubtful that Iolcus can be an adjective: the derivation of demonym adjective Ἰωλκός from toponym Ἰωλκός (fem.) is morphologically most implausible, and given the lack of classical attestations (in either Greek or Latin) one needs very strong reasons to accept it. Lucarini cites in support Dictys Cretensis (quid Medeam? ignoratisne a Colchis in Iolcorum fines transuectam? 2.26) and Servius (on Verg. Ecl. 4.34 socii uero Iasonis Minyae appellati sunt uel ab agro huius nominis Iolcorum [Vossius: Colchorum codd.]). Besides the fact that these are late texts, however, in Servius Iolcorum is only a conjecture, and in Dictys the best and oldest manuscript reads Iolchorum (no doubt under the influence of the preceding Colchis), which the rest further corrupt to Colchorum (or worse); since Iolchorum could as easily be a corruption for Iolciorum as for Iolcorum, it seems quite likely that the former should be restored in both cases.Footnote 49 I suggest we should read Iason; while the corruption is not implausible in terms of palaeography (a → ol, ſ → c), the fact that Iolcus is thematically relevant to the context may no doubt have also contributed.
LINE 371
Columella tells his readers that it is now time to harvest certain varieties of vegetable (369–71):
It is clear that line 371 should refer to the making of bunches of celery and leek, but cingere cannot be construed to yield such a sense.Footnote 50 Some translators take the line to refer to the binding of some other bunches with sprays of celery and leek, but such an interpretation is absurd; besides, cingere does not mean ‘to bind’ either.Footnote 51 The required sense is restored by writing iungere (note that MS A reads in fact cınıgere, which has the exact same number of strokes as ıungere).Footnote 52
LINE 401
Columella explains when it is time to harvest different tree fruits (400–3):
Line 401 has two problems. First, aperit ‘opens’ is not a plausible verb to refer to (presumably) the ripening of fruits.Footnote 53 Second, advising that certain kinds of tree fruit should be harvested ‘when tree fruits ripen’ is not very informative; in fact, Columella is quite clear that only some varieties (note, for instance, praecox ‘early’, 403) mature at this time. As the second sign—the harvesting of mulberries, or perhaps rather blackberries—suggests, the first one should likewise come from outside the realm of horticulture.Footnote 54 Sirius is often conceived of as scorching the fields and, more specifically, ripening the wheat;Footnote 55 without complete confidence, I propose triticeos urit ‘scorches the wheat crops’.Footnote 56
