University of Virginia Library

The Servius of Cassel for Aeneid III-V
by
Arthur Frederick Stocker

In a previous volume of these Studies [1] it has been shown that there exists for the text of Servius Danielis on Aeneid III-V not only the one manuscript source (F) upon which Thilo placed virtually his sole reliance but also another (G) which, though curiously intermittent as a witness, nevertheless gives within its own limits a measure of independent testimony.

The purpose of this paper[2] is to establish the value of a third manuscript for Aeneid III-V, codex Cassellanus ms. poet. fol. 6 (C), a fine manuscript of the mid-ninth century[3] of which Thilo took some cognizance but in the appraisal of which[4] he was partly blinded by an outmoded view of Servius Danielis as Servius plus miscellaneous interpolations. The Cassellanus consists, in its present form, of 109 parchment folia. Savage states that it was composed originally of 23 quaternions, of which eight and some miscellaneous leaves are now missing. It begins with Servius Danielis on Aeneid I 1, and ends with the words numquam aliquis deorum simul colitur in a note which is common to Servius Danielis and the vulgate Servius on Aeneid VI 830.[5]

Within Aeneid I-II the Cassellanus is the principal manuscript for Servius Danielis. Thilo[6] and Savage[7] equate it with the basic codex Fuldensis, of which Daniel states in his Praefatio[8] that he received a collation just as his edition was going to press, too late for him to do more than publish the collation as an appendix keyed to the text which he had already constituted from other sources. So confident was Thilo in this equation that he employed the siglum C indiscriminately to designate readings taken either from the Cassellanus itself or from Daniel's appendix, to which recourse was and still is necessary for passages presumably once found in the now lost quaternions of the Cassellanus. The identity of these two texts has lately been challenged by Elder,[9] but surely the similarity between them was uncommonly close. For Aeneid III and following, however, the Cassellanus presents an entirely different and somewhat ambiguous aspect. It is almost completely devoid of the long passages foreign to the vulgate that most strikingly characterize Servius Danielis as represented by F G, and has superficially the appearance of a vulgate text. In notes that are substantially common to the two traditions, however, it manifests most of the textual peculiarities of Servius Danielis.

The situation will become clearer with a review of the essential nature of Servius Danielis. Following upon the studies of Barwick[10] and Rand,[11] there is general agreement with the proposition that Servius Danielis represents a fusion of the genuine commentary of Servius on Vergil, as embodied in well over a hundred extant manuscripts, with some other single ancient commentary, probably older than Servius' and in certain respects more valuable. Servius', however, as the more widely known and respected, became the dominant partner in this merger of texts and served as the basis for the contaminatio which resulted. The other (which we may call the D) commentary the compiler used as a mine of interesting and clearly ancient material with which Servius' might be enriched. Some of this was entirely extraneous to Servius, keyed to lemmata about which Servius had nothing to say. This could readily be introduced, with a new lemma, where the sequence of words in Vergil entitled it to a place. Some took the form of additional comment on words

or phrases already discussed by Servius. Material of this sort is loosely and often inelegantly grafted upon the Servian note, sometimes without so much as an effort to adjust conjunctions and transitional expressions to their new context. In a few cases the compiler appears to have found parallel exegesis in the two commentaries, with D offering a more expanded form of the same note that is found in Servius.[12] It is possible that D may have been Servius' direct source for these notes. In view, however, of the penchant of ancient scholiasts and encyclopaedists for mechanically transcribing the words of their sources, it is perhaps more likely that both D and Servius were drawing upon the same original. Where the compiler found notes that were parallel in this way, it was of course natural for him to adopt the fuller form. Finally, for considerably more than half of his notes, the compiler seems to have been quite satisfied with what he found in Servius. At any rate, whether because he discovered no new material in the D commentary or because he chose to reject it, he simply adopted the Servius as he found it in his Servius text.

Against this background, the character of the Cassellanus emerges more plainly. In Aeneid III ff., it consistently lacks the more extensive additions made by Servius Danielis to the vulgate text, not only those notes which are new in their entirety, as, for example, the note on TRISTE PER AVGVRIVM (Aen. V 7; Th. 589, 5-17), but also the supplements to notes occurring in Servius, for example, to the one on DARDANIVM ACESTEN (Aen. V 30; Th. 593, 3-20). When the Danieline additions are of shorter extent, however, they sometimes appear in the Cassellanus, e.g.,

• A. III 251 (Th. 386, 11) ne possint a se dicta contemni] Serv. ne possint (posint F) a se dicta quasi ab irata conficta contemni (contempni F G) F G C
• A. III 361 (Th. 406, 24-26) aves aut oscines sunt aut praepetes; oscines ore futura praedicunt, praepetes volatu significant] Serv. aves autem aut oscines (obscines P) sunt aut praepetes, oscines quae ore futura praedicunt, praepetes quae volatu augurium significant cum sunt prospera (cum sunt prospera om. C) F P C
• A. III 699 (Th. 455, 20-21) PROIECTA (proiectaque G) porrecta (prorecta F proiecta G), extenta, ut proiecto dum pede laevo] F G C non habet Servius.
• A. III 705 (Th. 457, 1-2) PALMOSA SELINVS civitas iuxta Lilybaeum, abundans palmis et apio] Serv. PALMOSA SELINVS (selinis F selynis P C) civitas

  ---

  96

  
  
  
  est iuxta Lilybaeum (lybeum F libeum P lilybacum C), abundans palmis quibus vescuntur et apio (palmis vel ab equis nobilibus P) F P C

and when the manuscripts of Servius Danielis exhibit textual idiosyncrasies embedded within notes that are common to the two traditions, the Cassellanus oftener than not shares them. The following instances are typical of many:

• A. III 305 (Th. 395, 28) quia] Serv. nam F G P C
• A. III 621 (Th. 445, 19) et mire] Serv. bene rem F C
• A. III 661 (Th. 449, 4) caecitatis solatium] Serv. solatium caecitatis F G C
• A. III 686 (Th. 452, 17) sequeremur] Serv. Sequi F G C
• A. III 689 (Th. 453, 17) incederet] Serv. flueret F G C
• A. III 701 (Th. 456, 9) carentes] Serv. sedata F G C
• A. III 707 (Th. 457, 8) non longe a monte Eryce] (erico plerique codd.) Serv. ante pedem montis (montes F C motis P) erycis F P C
• A. III 713 (Th. 457, 28) praedicere] Serv. nuntiare F C
• A. V 1 (Th. 587, 7) primum] Serv. om. F G C
• A. V 45 (Th. 595, 6) tacitam] Serv. tam F G C
• A. V 55 (Th. 597, 15) quo] Serv. ut F G C
• A. V 85 (Th. 603, 5) locus sine genio] (vel locus est sine genio) Serv. sine genio locus (locum F G) F G C
• A. V 95 (Th. 604, 10) apotheosin] divinitatis confirmationem add. F C
• A. V 233 (Th. 614, 22) esse] Serv. est et F G C
• A. V 823 (Th. 652, 2) vitam] Serv. animam F G C

On the other hand, the reverse situation (with the Cassellanus on the side of vulgate Servius) also occurs, e.g.,

• A. III 571 (Th. 439, 3) aquarum odor] Serv. C odor aquarum F G
• A. III 572 (Th. 439, 14) et est poetica descriptio] Serv. et est poetica discribtio C sane tota hec poetica discriptio F
• A. V 135 (Th. 610, 4) numquam tectum] Serv. C quod numquam tege (pro tegi?) solet F G

and on occasion the text of the Cassellanus appears to be eclectic, as on Aeneid III 297:

Servius Danielis

• PATRIO MARITO atqui Thebana fuit, de Thebis Phrygiis. sed aut provinciam pro patria posuit—et (aut F) probamus pro patria—ut Sallustius, etc.

Vulgate Servius

• PATRIO MARITO atqui Thebana fuit. sed aut provinciam pro patria posuit, ut (I 380) Italiam quaero provinciam, item Sallustius, etc.

Cassellanus

• PATRIO MARITO atqui Thebana fuit. aut probamus pro patria, ut Salustius, etc.

In this passage, it will be observed, the Cassellanus lacks the Danieline de Thebis Phrygiis and also (perhaps by a kind of homoeoarchy) the sed aut provinciam pro patria posuit which is common to the two traditions. But it concludes with Servius Danielis, aut probamus pro patria, ut Salustius.

Analysis shows that the text of Servius which the compiler used in constructing Servius Danielis did not belong to any one of the three families of Servius manuscripts distinguished by the Harvard editors. Servius Danielis, therefore, adumbrates a fourth variety of vulgate text, and an attractive hypothesis suggested by the suddenly reduced state of the Cassellanus in Aeneid III ff. is that it may represent the compiler's Servius, in which case the Danieline symptoms which it betrays would derive neither from the D commentary nor from any activity on the part of the compiler but rather from an otherwise now extinct form of Servius text which the compiler used. If this could be established, the Cassellanus would assume substantial importance in what must at best be a doubtful reconstruction of the lost D commentary.

The weight of probability is indeed on the side of the view[13] that some, at least, of the discrepancies between Servius Danielis and the extant vulgate do stem from the compiler's Servius. That the Cassellanus affords any clue to their identity, however, is by no means clear, and on the contrary becomes exceedingly dubious in the face of clear indications[14] that the Cassellanus, for Aeneid III ff. as well as for the books that precede, descends from a source which once embodied the full Servius Danielis and out of which the D supplements were systematically, if not entirely effectively, filtered.

On Aeneid III 244, for example, where Vergil says of the harpies,

Servius, elucidating vestigia, explains briefly: pedum signa. Servius Danielis subjoins an alternative: vel pedum signa vel indicium sordium quibus epulas discedentes polluebant. The Cassellanus, in its customary fashion, fails to show the Danieline addition, but offers the text, VESTIGIA vel pedum signa. In the restricted context, the vel becomes utterly otiose and finds its readiest explanation in the hypothesis that somewhere in the Cassellanus' line of descent the vel indicium sordium quibus epulas discedentes polluebant was marked for deletion, probably by someone whose purpose it was to purge the text of its D elements, but the vel before pedum was overlooked. A similar situation recurs a few lines later, at Aeneid III 249, where the two traditions diverge as follows:

The Cassellanus reads PATRIO REGNO ait Neptunio, omitting all that follows Neptunio in Servius Danielis. As the note on PATRIO REGNO, ait Neptunio is nonsense, and would seem to represent a clumsy emendation of an intermediate aut Neptunio, in which the aut, like the vel above, became impossible after the loss of Daniel's aut sicut alii dicunt . . . fuit. One more example may suffice. At Aeneid III 310, Vergil makes Andromache inquire of the unexpectedly appearing Aeneas:

Servius comments: et secundum muliebrem adfectum interrogat et bene suspicatur; nam et inferis sacrificat et in luco, in quo habitant manes. Servius Danielis carries the same note, with the addition of sicut supra dictum est after muliebrem adfectum, the allusion being to a foregoing observation found only in Servius Danielis: et opportunum principium feminae quam incertam timor fecerat. The Cassellanus, as might have been expected, omits the et opportunum principium feminae quam incertam timor fecerat, but retains the sicut supra dictum est which refers to it—a further instance of apparent oversight on the part of one whose efforts were directed toward excluding the D from Servius Danielis.

From these examples, which could be multiplied, it is apparent that the Cassellanus for Aeneid III ff. constitutes not a fourth form of the vulgate Servius but a mutilated specimen of Servius Danielis. Why this mutilation came about only in Aeneid III ff., while Aeneid I-II escaped, is a matter of conjecture. Perhaps, as has been suggested,[15] archetypal Servius Danielis was in several volumes, with a break between Aeneid II and Aeneid III. Perhaps the answer may be found in the mechanics of the filter process, as, for example, if the D supplements were marked in an ancestor of the Cassellanus (call it an Ur-C) with symbols which the copyist in Aeneid I-II either overlooked, did not understand, or had not been instructed to observe. Whatever may be the explanation, the conclusion is plain, that the existing Cassellanus represents a text that is in its essence Danieline throughout, and that, while for Aeneid III ff. the notes parallel in extent those of the vulgate Servius, the residual text is of the character of Servius Danielis.

That there are a significant number of instances in which the Cassellanus reads with the vulgate against Servius Danielis has already been remarked. A few of these are of considerable extent, as, for example, at Aeneid V 241, where the two traditions diverge as follows, and the Cassellanus reads wholly with Servius:

and at Aeneid III 351, where the Cassellanus reads with Servius except for the omission of ut in the third line:

In order, however, to eliminate the D from Ur-C, it was necessary to subject that text to rather careful collation with the text of some Servius manuscript. While it was undoubtedly the additions from D for which the collator was particularly on the watch, it would not have been at all unnatural, since his whole activity was predicated on the assumption of the superiority of the vulgate Servius to Servius Danielis, for him occasionally to have noted—marginally or interlinearly—variants or larger discrepancies on which his eye might have happened to light. Such instances, therefore, need not undermine our belief in the Cassellanus as fundamentally a Danieline text.

If the Cassellanus disappoints us in one hope, however, it at least partially fulfills another. As has been shown in my former article,[16] there is only one complete manuscript of Servius Danielis for Aeneid III-V, Thilo's F, and, while G seems not to have been copied from it, its relationship to F is so close that the two together comprise an exceedingly

narrow base for the text. A Paris manuscript (B. N. Lat. 1750)[17] serves as a check upon F-G in the passages which it contains, but, since its text is very much abridged, its testimony is often lacking. The Cassellanus, its authority as a manuscript of Servius Danielis vindicated, considerably extends the area in which a check is available upon the vagaries of F-G, and, although not, unfortunately, in the long added notes from D on which help is particularly needed, enables the editor to dismiss some of their divergences from the vulgate Servius as the discrepancies of a single manuscript line rather than of the tradition of Servius Danielis.

The writer would lay down the following principles to govern the use of the Cassellanus:

• a. When the Cassellanus supports F-G in a one-word variant, an inversion of word order, or the omission of an unessential word, the reading receives strong confirmation as authentic Servius Danielis.
• b. The shorter additions of F-G which occur in the Cassellanus are probably Danieline and were overlooked by the collator when he was endeavoring to "purge" the Ur-C. Cassellanus readings, therefore, merit almost if not quite the respect accorded to the readings of F-G.
• c. When the Cassellanus attests short bits of text that do not occur in F-G but are found in the vulgate, they probably belong to the tradition of Servius Danielis. Since the purge was not rigorous enough to eliminate the shorter additions of Servius Danielis which were found in the Ur-C, it is unlikely that the collation was performed with sufficient care to result in the supplying from Servius of omissions in the Ur-C of the same extent.
• d. In the comparatively few cases where the text of the Cassellanus diverges sharply from that of F-G or is substantially more extensive, the probability is that it is not Danieline but that it was introduced into the Ur-C from the text of Servius with which it was being collated.

Notes