The boarded novels at Aberdeen, including the Minerva Press volumes, vary in physical condition: some are ragged and frayed, others fresh and intact. Most, however, are in a solid state of preservation, though not all arrived complete. Some lacked preliminary leaves. At least six of Newman’s works are defective in this regard. From 1819, the anonymous Families of Owen and De Montfort (3 vols) and Arthur Spenser’s Iskander (3 vols), lack half-titles and title pages in all of their volumes, while M. S[?mith]’s Frances (3 vols), lacks them in volumes I and III. [7] Sarah Green’s Gretna Green Marriages (3 vols, 1823), lacks preliminaries in volumes II and III, Mac-Erin O’Tara’s Thomas Fitz-Gerald (3 vols, 1825), in volume I. These defective volumes almost certainly arrived at King’s College in this state, since the Latin accession inscription (Lib Coll: Reg: Ab. St. Hall, or variants thereof, an abbreviated form of Liber Collegii Regii Aberdonensis. Stationers’ Hall) is written on page 1 of the narrative text, the first page available in each volume. Did the deposit status of these volumes make Newman or his binders less concerned about sending King’s College defective copies? Are these statistically unusual occurrences? It is hard to answer either question. What one can venture is that with production levels of the kind maintained by the Minerva Press, occasional imperfection is to be expected. Whatever the condition of the novels, they were not ignored. A few are unopened, but most have clearly been read. Occasionally readers have left their comments in margins or on blank pages. [8] Whoever ordered the works knew there was an audience for them among the university population. However many times they were handled or read, and to the extent that they were complete when they arrived at King’s College library, they come down to us with all of their original constituents (boards, spines, labels, textblocks) intact.

The level to which the boarded Minerva books in the collection overlap with the Press’s total output between 1814, the date of the collection’s earliest example, and 1834, the date of its last, can be suggested by the following selected data: the library contains, in original boards, nine out of eleven works by Selina Davenport; [9] four out of five by Miss C. D. Haynes (later Mrs Golland); [10] three out of sixteen by Sarah Green; [11] four out of six by Anne Raikes Harding; [12 ]four out of twelve titles by Jane Harvey; [13 ]eleven out of fourteen works by ‘Anne of Swansea’, Anne Julia Kemble Hatton, accounting for all except her three earliest; [14] the final eight novels of Francis Lathom’s total of nineteen; [15] the last five of eleven works by Henrietta Rouviere Mosse; [16] four of eleven by Regina Maria Roche; [17] all but two of the pseudonymous Rosalia St Clair’s twelve novels; [18 ]five out of sixteen works by Louisa Sidney Stanhope; [19 ]and four by Zara Wentworth. [20] That is seventy-one titles, or 263 volumes, and, as indicated, that is only a portion of the total number in the collection. [21]

Format, Sewing, Squaring

As in earlier years, the standard Minerva novel in boards in the second and third decades of the nineteenth century is a 12mo in format; that is, the individual gatherings that make up the textblock comprise twelve leaves or twenty-four pages each. The outside board measurements are about 7 1/2″ x 4 1/2″. The width, or thickness, of volumes is principally determined by the number of pages in any given volume. It is usually less than an inch. Volume thickness is also affected by the pattern and tightness of the sewing, and by the work’s handling over time: use tends to loosen the structure and expand the width of the textblock. The spine is solid (the spine paper covering glued to the back edge of the textblock) and, in a curve of varying sharpness, rounded. Sewing is invariably on two recessed supports (with two outer kettle-stitches). Sewing patterns, as we shall see, show considerable variation.

The textblock is usually slightly smaller than the board dimensions, and this creates the squaring, the projection of the boards, at top, front, and bottom edges, beyond the dimensions of the textblock, to afford protection to the page edges. The pages are untrimmed, leaving the irregular deckle edges at the front and bottom of the textblock. Since the pages are not trimmed to a uniform dimension, their exact size varies. Beyond these structural features are two further physical elements of significance, the one decorative (paper colour), the second substantive (printed labels). The most common colours of the paper used on Minerva and other novels of the period are various shades of blue-grey for the boards and cream for the spines. Almost as frequent are shades of brown paper, used for both boards and spines. The printed labels carried by the Minerva volumes typically provide four to five items of information: the title of the work; the author’s name (when not anonymous); the number of volumes in the set; the number of the volume in question; and the price. The label size changes over time, a development that is central to the evolution of the distinctive Minerva look. Looked at more closely, all of these physical features of the Minerva novel provide information about Newman’s production and retailing practices.

The regular 12mo format is an unassuming size of volume for an adult audience. Children’s works were even smaller, but that was part of their appeal. The relatively small physical size of 12mo novels was one of the reasons for the critical condescension aimed at them. We know that Scott sought to raise the dignity of the novel by insisting that Ivanhoe (which appeared on 20 December 1819) be published in the larger and more culturally prestigious octavo (8vo) format. [22] At about the same time, other publishers played with variations on the 8vo format. In March 1819, Taylor and Hessey published, in one volume, The Authoress (attributed in EN2, Item 1819: 67 to Jane Taylor). Though no larger than the standard 12mo (the boarded copy in the Aberdeen collection measures 7 1/4″ x 4 1/2″), the work was advertised as a ‘foolscap 8vo’. [23] The phrase is ambiguous, if not actually misleading, given the meaning of ‘foolscap’. Though the leaves might have been gathered and sewn in 8vo signatures (that is, with eight leaves or sixteen pages in each), the size of a foolscap sheet (17″ x 13 1/2″), on which a single gathering was printed, was smaller than that of a demy sheet (22 1/2″ x 17 1/2″), on which a full-sized 8vo was normally printed, and hence it yielded a textblock no larger than a standard 12mo. [24] What the advertised format offered in folding and sewing, it took away in paper size. It seems that if some form of 8vo could be claimed, it was worth doing so. Another example of this practice in the Aberdeen collection is Lady Caroline Lamb’s Ada Reis (3 vols, John Murray, 1823). It also was advertised as a foolscap 8vo, and it also is no larger than a 12mo publication. [25]

Newman did not use an 8vo format, whether foolscap or demy, but from 1813 onwards the size of his 12mo format became larger than anything the Minerva Press had employed hitherto. [26] Later, very occasionally, while retaining a 12mo sewing structure, he used a large page size. There are two examples in the Aberdeen collection, Mac-Erin O’Tara’s Thomas Fitz-Gerald the Lord of Offaley (3 vols, 1825) and The Stranger Chieftain (2 vols, 1834). In both instances, the overall dimensions of the volumes are in excess of 8 1/2″ x 5″. The setting of the type on the page, however, conforms to a 12mo scale, the result being unusually wide margins. Thomas Fitz-Gerald was advertised as an 8vo, but it is that only in page size, not in sewing structure. [27] Advertisements at the back of each volume of The Stranger Chieftain list only 8vo works, a designation with which The Stranger Chieftain is presumably to be associated, but again on the basis of page size, not sewing structure. These large-page 12mos represent a reversal of the foolscap 8vo strategy practised by others: Newman offered in paper what he saved on sewing. He might have advertised them as ‘Elegantly printed’. This was the phrase used by F. C. and J. Rivington, and T. Hookham in their advertisements for Lætitia-Matilda Hawkins’ Heraline (4 vols, 1821). [28] The work was a large-page 12mo, for which the phrase ‘Elegantly printed’ drew attention to the typographical luxury of wide margins—an extravagance befitting, we might assume, the novel’s royal dedicatee, the Duchess of Gloucester. But Newman did not follow suit. What he did do, with some frequency, to give his Minerva novels increased consumer allure, was to advertise them as being ‘large’.

In this context, the term ‘large’ refers not to the size of the boards or pages, but to the thickness of the textblock, as determined by the number of pages. A count of fifteen novels at Aberdeen that Newman published between 1815 and 1825, and advertised as comprising ‘large’ volumes, reveals that volumes earned the adjective if the average number of pages in each approached 300 (the word being applied not to the whole work but to its constituent volumes). [29] Each of the fifteen works save one (Jane Harvey’s Singularity [3 vols, 1822]) is a four- or five-volume work (and Anne Hatton is responsible for six of them). On average, each volume in a set designated ‘large’ contains 294 pages, with a high of 354 in the case of Hatton’s Secrets in Every Mansion (5 vols) and a low of 236 in the case of the anonymous Jessy (4 vols), both published in 1818. The disparity in the average number of pages in these two works would seem to suggest a problem with, at best, the consistency, at worst, the credibility of the word ‘large’. But Jessy is a singularly slim contender for the designation. Six of the works average over 300 pages per volume, a further three over 290 pages, and others come within a respectable distance of that number. In sum, there is a rough-and-ready truth in advertising. What, by contrast, is surprising are the odd occasions upon which Newman could have justifiably used the ‘large’ claim in advertisements but did not. Such is the case with three further works by Anne Hatton—Lovers and Friends (5 vols, 1821), whose total page count is 1,498, for a volume average of 300; Deeds of the Olden Time (5 vols, 1826), at a total page count of 1,615, for a volume average of 323; and Uncle Peregrine’s Heiress (5 vols, 1828), at a total of 1,722 pages for a volume average of no less than 344 pages. The first omission could simply have been an oversight; the second and third were published after the date in the mid-1820s when Newman seems to have dropped this particular promotional tool in his advertisements. While he was employing the designation ‘large’, Newman would doubtless have been aware of one exceptionally formidable challenge to his definition of the word. The anonymous writer of Hardenbrass and Haverill (4 vols, Sherwood, Neely, and Jones, 1817), in ‘the Author to his Book’, says to his work, ‘thy bulk is four fat volumes’ (I, ix). He was not exaggerating: these volumes, while standard 12mo in format, provide a total of 1,830 pages of narrative (I 422pp., II 606pp., III 382pp., IV 420pp.), for an average of 458 pages per volume.

However ‘large’, whatever its thickness, the extent to which a volume will remain intact relies principally on its sewing. Sewing patterns in the Minerva novels range from the rare 1-on, to the more common 2-on, to the even more common 3-on, and up from there. [30] Inconsistencies, in the form of mixed sewing patterns, occur both within volumes and between volumes comprising a single title. Amelia Beauclerc’s Montreithe (4 vols, 1814), can stand as a representative example. Volumes I and III utilise 1-, 2-, and 3-on patterns, in no particular order. Volumes II and IV use 1-on to commence and conclude the sewing, and 2-on, consistently and competently, for the bulk of the volumes. Two inferences suggest themselves. First, different sewers worked on different volumes, perhaps one on volumes I and III, another on volumes II and IV. Second, the first sewer was either not particularly skilled, or not particularly careful, or perhaps did not consider that he, or she, was being paid for a more than a passable level of work. Whatever the case, that indifferent sewer was more adept than the sewer or sewers of Mary Johnston’s The Lairds of Glenfern (2 vols, 1816). Here we find a confused mix of 1-, 3-, 4-, and 6-on sewing patterns. And even that mixture is restrained compared to the sewing of M. S[mith]’s Frances (3 vols, 1819), which offers the following eclectic combinations: in volume I, 4-on and 2-on; in II, 10-on; and in III, 8-on and 1-on. [30] This mix suggests the levels to which untrained or indifferent craftsmanship employed on these bindings could dip, and the minimal attention given to quality control. But the Minerva Press was not selling fine bindings, rather economic and serviceable ones, and whatever the lapses and inconsistencies of workmanship, by far the greater number of Minerva volumes at Aberdeen have survived any carelessness in their sewing.

If sewing ensures the cohesion of the textblock, squaring—the projection of the boards beyond the top, front, and bottom of the textblock—offers protection to the exposed page edges. The Minerva Press volumes reflect the increasing use of squaring in boarded books in the opening decades of the nineteenth century. [31] Its introduction and refinement during the period is illustrated by a quite different kind of work in the Aberdeen collection. William Frend’s Evening Amusements; or, the Beauty of the Heavens Displayed (19 vols, London: J. Mawman, 1804–22), was an annual publication offering descriptions of the night skies and written for a general audience. These nineteen 12mo volumes offer an object lesson in the gradual and steady adoption of squaring by binders of books in boards. Though there is some squaring on the first volume—presumably the publisher wished to make the opening number of the work as attractive as his budget would allow—there is effectively none on the next seven volumes, 1805–11. Squaring then begins to creep in with volumes IX–XIII (1812–16). By volumes XIV–XIX (1817–22), it is fully established. A comparable evolution is visible in the binding of Minerva novels, but now, nearly two hundred years after their production, it is often difficult to judge the depth of the original squaring with which the novels were provided on first being published. The novels’ solid spines would originally have been hammered into a rounded curve as part of their preparation. This convexity pulled the textblock up and back against its own spine, away from the edges of the board squaring. The boards would then have remained clearly projecting beyond the edges of the textblock and protectively functional. Over time, the textblock, especially if the sewing was poor and the volumes were ‘large’, would sag under its own weight, sinking down and forward to the edges of the boards and pulling the spine flat behind it, to the detriment of the squaring. This is the condition of many of the Minerva volumes, especially the heavier ones. But cause and consequence is less straightforward than it seems. There are a sufficient number of volumes from the 1820s that have both flat spines and good all-round squaring to suggest that their spines were only minimally rounded to begin with. The possibility arises that, by this decade, what was being spent on squaring was being saved on the hammering and rounding of the spine. In sum, while squaring is ever more present in Newman’s publications from the mid-teens to the late twenties, its execution and quality is unpredictable.

Binding and Wrappers

That Minerva novels of this period were invariably published in boards meant that they were retailed to circulating libraries ready-to-rent (to echo James Raven’s statement that ‘novels were published to rent out’), [32] and made to last at least as long as the novel was in fashion and being read. Other publishers were still issuing novels bound in a more fragile, old-fashioned manner—in wrappers. The Aberdeen collection suggests that the Hookhams did this often: T. Hookham, Jr and E. T. Hookham issued John Hamilton Roche’s A Suffolk Tale (2 vols, 1810) in blue-grey wrappers. T. Hookham published Abel Moysey’s The Confederates (3 vols, 1823) in blue wrappers together with, in a puzzlingly decorative detail for a case of the most ephemeral binding possible, pink labels. From the 1830s, we find a further three works from Thomas Hookham in wrappers, brown for Elizabeth Cullen Brown’s Passion and Reason (4 vols, 1832), bright navy-blue for the anonymous Marston (3 vols, 1835), a perishable binding all the more unexpected for being on an 8vo publication, and brown again for Frederic Reynolds’ The Parricide (2 vols, 1836). Beyond the Hookhams, we find wrappered novels at Aberdeen from Longmans (Robert Gillies’ The Confessions of Sir Henry Longueville [2 vols, 1814]), and Law and Whittaker (the anonymous Delusion [2 vols, 1818]). One can interpret the use of wrappers on novels during the 1810s to 1830s in various ways: it was a way of moving unsold stock as cheaply as possible; it was a way of keeping the publication less expensive by passing on the cost of a more durable binding to the purchaser; or it was, perhaps, simply a case of flimsy covers for fugitive literature. Whatever the motive for those who employed wrappers, Newman at the Minerva Press used only board bindings, and thereby furthered the binding style’s prevalence, increased its acceptability, and supplied his own and other circulating libraries, and individual customers, with a binding that blended economy and utility.

Very occasionally, we find a Minerva novel bound in some colour other than the ubiquitous and economic blues and browns. Coloured papers on boarded works in general were not uncommon by the 1810s, [33] but they were more expensive than standard blues and browns and were not usually employed on novels. In the Aberdeen collection, we find coloured papers employed on other literary genres: for instance, on poetry (in the musk-rose boards of Mrs Cowley’s The Siege of Acre [G. Wilkie and J. Robinson, 1810]), on print history (the olive-green boards and marbled spines of William Parr Greswell’s Annals of Parisian Typography [Cadell and Davies, et al, 1818]), and on drama (the all-over carmine covering to Richard Paul Jodrell’s two slim quarto volumes, The Persian Heroine and Illustrations of the Persian Heroine [Printed for the Au-thor, 1822]). They do also appear on novels, such as John Jones’ Hawthorn Cottage (2 vols, James Asperne, 1815), bound in beige-pink boards, a cream spine, and grey-green labels. Among the Minerva works, Amelia Beauclerc’s Montreithe (4 vols, 1814), Selina Davenport’s The Hypocrite (5 vols, 1814), the two earliest Minerva novels in boards in the collection, and Anne Hatton’s Secret Avengers (4 vols, 1815), each have pale- to buff-pink spines, and the anonymous Bandit Chief (4 vols, 1818), is bound in rose-red boards and a buff spine. But these are exceptions, the reasons for their more attractive colours open to guesswork, and probably rather banal (some spare coloured paper on hand at the binders, for example).

Labels

Just as he stuck to board bindings, so Newman almost invariably used printed labels. This was not the universal practice. Boarded novels without printed labels are to be found published by Walker, Baldwin, Longman, and Whittaker, and in two cases from the Minerva Press—Elizabeth Bennett’s Faith and Fiction (5 vols, 1816), and Charles Lucas’ Gwelygordd (3 vols, 1820). These last two works, as with most of the originally unlabelled works by other publishers at Aberdeen, do in fact carry labels, but these were supplied not by the publisher but apparently by the library (just when is difficult to say). Further, the individual volumes so labelled have been stamped with their respective numbers. The style of this library labelling is uniform. Its tell-tale sign are the ruled lines at the top and bottom of each label: they run over the edge of the label paper onto the spine paper. These were not pre-printed labels; they were added later to the spine and ruled in place.

While present, labels can be erroneously placed. Those, for example, on the two volumes of Mary Johnston’s The Lairds of Glenfern (whose muddled sewing has already been noted) are reversed. Misplacement also occurs on two Longman publications, Edward Harley’s The Veteran (3 vols, 1819), and the anonymous Arthur Seymour (2 vols, 1824). The labelling of Alexander Brodie’s The Prophetess (3 vols, Thomas Clark, and Longmans, 1826), turns into a three-card trick: the label for volume II is on volume I, III is on II, and I on III. Statistically, these errors of labelling are insignificant, but even on that scale they suggest, as with missing preliminaries, part of the cost in quality of low-budget production.

The size, specifically the length, of labels is determined by three factors: the amount of information they carry; how they are cut; and how they are designed. Save for a rare addition, labels carry the four or five pieces of information already noted (title, number of volumes, price, volume number, and, when given, the name of the author). The rare addition is the date of the work. The earliest example in the Aberdeen collection is on the unusual work already noticed, Lætitia-Matilda Hawkins’ ‘Elegantly printed’ Heraline—another sign, perhaps, of its over-production. The label, in addition to title, author, number of volumes, and volume number, though no price, carries the date, 1821. Newman did not add such data until the mid-1830s, on the anonymous Benson Powlett (2 vols, 1833) and Timothy Flint’s Francis Berrian (3 vols, 1834). The latter is exceptional in offering six pieces of information—title, author, number of volumes, price, volume number, and date—indeed, we could claim seven, if the subtitle, or, the Mexican Patriot, also supplied, is counted in its own right.

Fig. 1. The page of undivided labels from vol. III of Anne Hatton, Secrets in Everty Mansion (5 vols, 1818). Reproduced by permission of the University of Aberdeen.

The cutting of the labels, self-evidently if minimally, affects their size. Labels were issued printed and undivided on a single integral sheet forming part of the preliminaries of one of the volumes. Which volume, in a multi-volume work, can vary. The survival of these sheets of undivided labels in situ is rare. They are usually found in leather-bound volumes for which the binder had no need of the paper labels. In these cases, by intention or oversight, the binder omitted to cut out and discard the unwanted leaf. In the Aberdeen collection, there survives an example of the rarest label sheet of all, a page of undivided titles in a boarded book which, nevertheless, carries a set of printed labels on its spines. The work is Anne Hatton’s Secrets in Every Mansion, whose ‘large’ volumes we have already met. The recto of the final leaf, R6, in volume III, contains the five undivided labels printed laterally on the page in a vertical stack (Figure 1, above). The binder took the labels used on the spines either from another copy of the work (which was issued unlabelled?) or from an additional, unbound sheet of labels provided by the printer. [34] The surviving integral sheet, in its unfaded and fresh whiteness, is a vivid reminder of just how bright the labels on boarded books, and the rest of their paper coverings, would have been when they were first issued, before dimmed by time and use. [35] The labels to Secrets in Every Mansion are ruled at top and bottom, the ruled lines providing a guide to their cutting and dividing. The space between the top and bottom rule on each label is 1 3/4″. The labels on the spine are 1 7/8″ tall. Thus, the person who cut the labels on the spine found an extra margin of 1/8″, divided between the upper and lower edge. In the case of ruled labels, there is not much room for choosing the height of the label. When there are no ruled lines, there is more choice, a small measure of discretion. We can assume that labels were usually cut with a chopper, but sometimes they must have been cut more fastidiously with scissors. The irregular bulge on either side of the labels to Selina Davenport’s An Angel’s Form and a Devil’s Heart (4 vols, 1818), there to accommodate the length of the title’s lines of type, indicates careful hand cutting.

Fig. 2. Selina Davenport, The Hypocrite (5 vols, 1814); Anne Hatton, Woman’s a Riddle (4 vols, 1824). Reproduced by permission of the University of Aberdeen.

The principal determinant, however, in the size of Minerva labels was a matter of discretionary design: in the space of a few years the labels were increased in height. Between the mid-1810s and the mid-1820s, there is a general evolution from shorter, squarer labels to taller, rectangular ones, from labels 1 1/2″ in height to labels approaching double that measurement. Davenport’s The Hypocrite (5 vols, 1814), has labels that are 1 1/2″ (Figure 2, above). They carry four pieces of information: the novel’s title, the number of volumes, the price, and the volume number. With slight variations, this remains the height of Minerva labels until 1818, in which year an increasing number of the press’s labels approach 2″ or more. Typical is Miss C. D. Haynes’ The Foundling of Devonshire (5 vols, 1818) (Figure 3, below). The five labels vary in length from 1 3/4″ to 2″, there is some inconsistency in their cutting, they are affixed on the volumes at differing heights, and they carry five pieces of information—to the earlier four is now added the name of the author. This new length becomes the norm, until by the mid-1820s it stretches to 3″. The labels on Anne Hatton’s Woman’s a Riddle (4 vols, 1824) are 2 13/16″ (Figure 2), those on her Deeds of the Olden Time (5 vols, 1826), a full 3″.

Fig. 3. Miss Broderick, The Cumberland Cottager (3 vols, 1818); Miss C. D. Haynes, The Foundling of Devonshire (5 vols, 1818). Reproduced by permission of the University of Aberdeen.

Newman’s tall labels of the late 1810s and early 1820s are distinctive for prose fiction of their period. To make but one contrast: the labelling on boarded copies of Scott’s novels, whether published by Constable or by Cadell, and their respective co-publishers in London, hover around 2″. The taller dimensions of the Minerva labels have patent retailing motives. In the first place, on small 12mo volumes, they attract attention. What, in part, they draw attention to is a commercial communication quite different from that offered by the labelling of earlier leather-bound volumes and from many other contemporary books in boards. To make the point in the form of a double question: what leather label ever carried on it the price of the book to which it was affixed, or indicated the number of volumes in the set of which it was a part? [36] Other boarded novels at this period often omit any mention of price, but not those from Minerva. Certainly a volume’s title, author, and volume number are equally convenient for bookseller and purchaser. But, in general, the labelling of leather-bound volumes, and of boarded books without price information on their labels, points to a moment in the existence of work that comes after its purchase, to the condition of ownership, to the moment the book is shelved in the library. The labelling of books in boards with the price on them points to an earlier moment, to the activity of selling the product. The same is true of the label’s indication of the number of volumes in the set: it ensures that vendor and purchaser, at the moment of transaction, know how many volumes are meant to be changing hands. (Later on, it let the circulating-library client or reader know how many volumes there are to go.) The label then becomes as much a retailing device as a sign of bibliographic identification. One might speculate about the evolving impression over time that a price on a label might make on an observer, but in its original release into the public view we can probably be more certain about its messages: it indicated a fixed and guaranteed price; it enabled the purchaser to calculate the per volume cost; it marked the book as a consumer item; the title and, if supplied, the author, made it, the publisher trusted and the buyer would hope, worth the cost; and for those lucky enough to get a discounted price, the label would be a satisfying reminder of their savings. [37]

The basis of the gradual elongation of the Minerva label had been laid down by a structural innovation in leather-bound books of the second half of the eighteenth century—the introduction of recessed supports and smooth spines (or, more accurately, the reintroduction, since the technique had been used in the seventeenth century, dropped, and then revived again). The traditional placing of labels on book spines had been determined by the position on the spine of the raised bands covering the raised sewing supports beneath. Conventionally, on leather-bound books, the upper label occupied the compartment demarcated by the raised bands of the top two sewing supports. Such other labels as were used would then be placed in lower compartments. [38] Early books in boards with printed labels followed the practice of using the upper compartment, and when there was no printed label, and the title or volume number was handwritten onto the spine, there was a natural or, more properly stated, culturally determined tendency to write the information in the upper compartments. [39] The return of smooth spines on leather bindings in the later eighteenth century, effected by the sawing of grooves in the spine and the sinking therein of the sewing supports, might have changed all this, since the surface of the spine was no longer structurally divided into compartments by the protrusion of the sewing supports. But it did not. For the most part, smooth leather spines continued to be tooled and decorated, and even supplied with false raised bands, as if their surfaces were still divided by raised supports, and labels continued to occupy, and to be confined in size, by the memory of their traditional compartments. [40] It was likewise for printed labels on books in boards. But there was an early difference between leather labels and paper labels: with the former there is frequently more than one label (each placed in its own spine compartment); with the latter, that is never the case. [41] On books in boards, a single label carries all the information the publisher wants dis-closed; it made no economic sense to print, and to pay for the labour to attach, more than one. If there was to be but one label, the more information it carried, the taller it needed to be.

The development of the smooth spine opened up more vertical space on the surface of the spine in which labels might grow in length. Further encouragement was given to employing the spine space more fully by a reduction in the number of sewing supports used in the binding of books in boards. At the turn of the century, many 12mo books used four sewing supports; by the second decade of the century this has been reduced to two. (The same is true of many 8vo volumes.) If the two sewing supports are well spaced, this enlarges the distance between them in which a label can expand and be placed. On many books in boards, however, recessed supports are not always sunk deeply into the textblock, they are often awkwardly positioned, the spine covering is usually of thin paper, and the labels are not always placed between the supports but frequently placed across the upper one. The result is rubbing, erosion, and deterioration in the legibility of the label. The upper portion of many labels on 12mo novels shows stress of this kind.

Even if its motives were primarily commercial and pragmatic (the provision of retailing information), the evolution and design of the long label on Minerva novels also had subtle but effective aesthetic consequences. The tall labels on certain Minerva titles conduce to an elegant poise and slimness of line. This is most apparent when the spines retain their original, rounded convexity, and that most often occurs when the novels are not obviously of the ‘large’ category. Two works of 1818 illustrate this appearance. Miss Broderick’s The Cumberland Cottager (3 vols) is of standard 12mo dimensions, approximately 7 1/2″ by 4 1/2″ (Figure 3). Its boards are covered in blue-grey paper, its spine in green-grey. The labels are 2 3/8″ tall, that is, very close to being in a one-third proportion to the (7 1/2″) height of the spine. Because the page count is low—including preliminaries and advertisements, volume I is 250 pages, volume II 240, and volume III 265, for a volume average of 252 pages¬—the spines are relatively narrow, about 3/4″. They also remain well rounded, the convexity adding to their structural elegance. There is consistent squaring on each volume. Volumes II and III are almost entirely unopened, which brings their compactness close to what they enjoyed on the day they first left the bindery. In the intervening two hundred years, the surface of the volumes has become soiled, stained, inscribed, and library-labelled, but these blemishes notwithstanding, the proportions and design of the volumes still suggest a crisp, simple, and finished elegance. The same ingredients combine to produce comparable effects in Miss C. D. Haynes’ The Foundling of Devonshire, already noted for its taller labels (Figure 3). This was one of those titles whose volumes were advertised as being ‘large’, but at an average volume length (on this occasion including preliminaries in the count) of 260 pages, it is not a bulky work. Its overall dimensions are the same as those of The Cumberland Cottager. It is wholly covered in a not-so-common olive-green paper. Its labels are about 2″ in length—not as long as those on The Cumberland Cottager, but sufficient to create the effect of vertical extension on the spines. The spines are slim and rounded, the squaring even and functional. Here again, surface fraying and distress apart, are volumes of poised elegance, created out of the standard materials of books in boards, but assembled and harmonised to raise their appearance to a level of distinctive visual appeal.

Conclusion

Overall, the physical quality of Minerva novels suggests a rough-and-ready calculus of costs and economies in their construction and binding. Strength or quality in one area is offset by weakness or haste in another. Time on or investment in one task is taken away from another. Competent sewing and adequate squaring do not always come together. Thick volumes can come at the price of adequate squaring. There can be qualitative differences between volumes in the same set, particularly in the sewing patterns employed. All such variables are, we can surmise, the result of achieving rapid, bulk production while trying to contain the aggregate cost of the product. We can assume that publishers preferred to allot a fixed cost for binding. The separate ingredients may vary in quality and price, but the overall outlay remains constant. In economic terms, the isocost remains fixed, while the cost of the various parts may fluctuate.

Even given the variable quality and the limited structural and decorative aims of their bindings, there was a brief moment when Minerva novels achieved an elegant proportion of design and finish that stands out and is worth acknowledging. Few of the volumes produced in the 1820s or later, with flatter spines and heavier, block-like shapes, prolong the appealing design of the late 1810s. By the mid-1820s, all paper-covered books in boards were, in any case, being rendered if not obsolete—they persisted, in dwindling numbers, until the 1840s and beyond—then certainly outmoded by the introduction of cloth binding. Cloth-bound books were to achieve levels of design, appearance, and durability quite beyond the reach of the paper-covered books in boards. Many books in boards quickly adopted the cloth spine. Others did not, and the standard blue or brown, all-paper book in boards, failing to develop in the 1830s, became increasingly fusty and outmoded in appearance. What could and had been achieved, however, in this binding genre, as applied in the commercial and consumer context of low-budget, high-production publishing, is memorably illustrated in the assembly of Minerva novels at Aberdeen. [42]