This book consists of a metrical concordance to Ovid’s Metamorphoses. It is meticulously prepared and accurately produced, which is a cardinal virtue in such a volume. In my opinion a digital format with suitable search programs would have provided a much more flexible instrument. The information contained is confined to the use of dactyls and spondees, and hence pays no attention to caesurae or the other important features used so subtly by the great poets to vary the aesthetic effect of the hexameter. One must, however, be grateful for the tedious labor that has produced this reliable index. This reviewer has some difficulty in envisaging the variety of uses to which it may be put by its readers.
There is a clear introduction setting out the author’s purpose (“to offer a … picture of the hexameter patterns” (p. ix)) and method. The bulk of the book consists of two parts. Part I sets out a complete text of the Metamorphoses, with sets of marginal marks. Both numerical codes and the letters ‘l’ and ‘s’ (see below) on the left margin indicate the pattern of dactyls and spondees in each line, and indications are given when the pattern of scansion is repeated in successive lines. The right margin gives the book and line reference, and shows whether the verse contains any word in reported speech.1 On the right hand also there are used on occasion (i) the numeric symbol (two obliquely crossing pairs of parallel lines), which indicates 56 passages where variants in the source editions give differing scansions, and (ii) the black spot, which is the mark of verses bracketed by Tarrant’s alone among the five modern source editions.2 Verses “bracketed in most if not all of the source editions as an interpolation or some form of redundant duplicate” are marked in the text with square brackets. The obelus is used in the usual way to indicate corrupt transmission in six passages only. Part II of the work classifies the verses by the possible patterns or ‘schemata’ of scansion. In this part the verses are arranged in “whole-word alphabetic-character” order.3 As in Part I, the patterns of scansion are numerically and graphically coded in the left margin, and on the right are found book and line references, and the other right hand symbols used in Part I.
We are told that the text and scansions were created manually, and that we should allow for human error. I proof read a section of 600 lines and checked the scansions: there was found only the trivial mistake at 8.283, which has “habet Sicula” for “habent Sicula” (an error reported by Magnus from Tarrant’s L, s. xi-xii). For reasons of copyright, it seems, Dee’s publisher required him to construct a text of his own.4 As a sample, I list his deviations from Tarrant’s text in Met. 8.100-199: 117: ‘exponimur orbae’; 121 ‘-que tigres’; 131 ‘vero’; 150 (no obelus); 157 ‘thalamo’; 190 (no brackets), ‘sequenti’.
The possible patterns of dactyls and spondees are indicated numerically by numbers from 1 to 32, and graphically by ‘lss’ for dactyls and ‘ll’ for spondees. Thus the first four lines of the First Book are represented by
Dee thinks that this system gives a good impression of the texture of dactyls and spondees in the versification,5 and claims that in the traditional notation of macrons and breves the symbols “are usually too small as printed and not easily distinguished at a glance” (p. xi). I find his system less attractive, if only because ‘l’ is a horizontally shorter letter than ‘s’, and the visual impression is given that the dactyl (‘lss’) is longer in duration than the spondee (‘ll’), and the code for a holodactylic line when printed is much longer than that for a holospondaic line. It surely is not beyond the powers of type designers to make the traditional system serve Dee’s purpose just as well.
Dee states that his book presents a clear picture of the hexameter patterns by giving their linear sequence from the beginning to the end of the poem, and the set of individual lines showing each pattern. “The novelty of such an approach and the obvious appeal of being able to see and make extended comparisons between Vergilian and Ovidian phenomena will, I hope, constitute sufficient justification for issuing this volume” (p. ix). Can one believe that it will be easy to see comparisons between Virgil and Ovid? If one opens a page of Dee’s Repertorium for each author, is it likely that the eye will easily comprehend clear resemblances or differences? It seems to me that a scholar will have to take out pencil and paper and compile figures for anything beyond a few verses. To extract those figures from this book could be laborious. Similarly, one might wish to discover whether the various books of the Metamorphoses show variations in the proportions of the different metrical schemata used. The results could possibly lead to the detection of chronological development or stylistic variation.6 (Thus Dee believes that “unanticipated discoveries” of a “metrical-aesthetic” nature may become available from the material gathered in Part II (p. x).) A digital version with a suitable search program would greatly increase the flexibility of use.
A cheaper (though less convenient) book might have been produced by printing no text, but simply supplying the numerical codes and schemata, and assuming the use of Tarrant’s text. The author has already produced two parallel volumes: Repertorium Homericae Poiesis [sic] Hexametricum, 2 vols. (Hildesheim, 2004),7 and Repertorium Vergilianae Poesis Hexametricum (Hildesheim, 2005). Now that Olms have started the series, do they envisage further expensive volumes for Apollonius, Lucan, Valerius Flaccus and others? The utility of the volumes already published would increase as indexes to other authors were added, but they would be an expensive purchase. For the Ovidian hexameter the scholarly world still awaits something like Maurice Platnauer’s exemplary Latin Elegiac Verse: a Study of the Metrical Usages of Tibullus, Propertius & Ovid (Cambridge, 1951). S. E. Winbolt’s Latin Hexameter Verse: an Aid to Composition (London, 1903) indicates the complexity of the task.
1. The use of quotation marks would have misled computers in sorting the text.
2. H. Magnus, Berlin 1914 (‘1915’ on p. xviii is a slip); R. Ehwald, Leipzig 1915; M. Haupt-O. Korn-R. Ehwald-M. von Albrecht, 2 vols., Zürich-Dublin 1966; W. S. Anderson, Leipzig 1977; R. J. Tarrant, Oxford 2004.
3. I believe that an arrangement gathering together lines with similar word-lengths would be far superior. Then “admonuisse, iubet deponere, Quique fuisti (15.543)” would be grouped with “immemoresque mei disceditis, inquit, Achivi (13.445),” “occubuisse neci; mirabere, vixque probabo (15.499),” “pallidiora gerens exhorruit aequoris instar (4.135),” “vaticinata vias: Ismenides, ite frequentes (6.159)” (though the pauses are different), rather than “adicit. accipimus sacra data pocula dextra (14.276).” (The computer sorting sent all bracketed verses to the end of each section, and on p. 277 placed ‘Quos’ (Met. 9.431) before ‘QVEM’ (Met. 2.328).)
4. I can see why an editor may guard his apparatus, but, if he believes his text is that written by the author, how can he claim copyright in the words (passing over the punctuation and other details of format) that he prints? Dee describes his text, based on the editions listed in note 2 above, as derivative and eclectic.
5. “… one can skim down the column of Schemata and see … without having to perform any cerebral calculations, the … always changing and never tediously repetitive metrical flow of the narration …” (p. xi). This should be more directly perceived by simply reading the verses aloud. However, when comparing one poet with another one cannot usefully hear both at the same time: perhaps an eye may secure a useful conspectus.
6. I amused myself by arranging the occurrences of Schema 8 (lllllllsslsslx). The results were I – 7; II – 12; III – 12; IV – 15; V – 8; VI – 12; VII – 9; VIII – 12; IX – 6; X – 7; XI – 15; XII – 16; XIII – 17; XIV – 9; XV – 11. No indisputable pattern is apparent from this result, but in Book XIII there are only two occurrences in the first 380 verses (the Debate of Ulysses and Ajax), fifteen in next 404 verses, and none in the final 183 verses. This appears to mark a stylistic difference.
7. There is an interesting discursive review by Donald Lateiner (BMCR 2004.11.24). He complains about the use of that inferior style of binding known by the misnomer “perfect”. Underneath its meretricious headbands my copy has the same kind of binding. It is just not good enough for reference books.