University of Virginia Library

For many eighteenth-century books the printer supplied only enough copies to meet immediate needs and then, whenever he could hold the type, reimpressed as often as a continuing sale might require. Work in moderate demand might proceed through several printings, each issued at a certain interval and readily identified by differences in paper, text, typography, or press figures. Work achieving sudden notoriety, on the other hand, might run through numerous impressions, all issued, perhaps, only a day or so apart and bearing very few marks of identification. Occasionally, indeed, as with Erskine's View of the Present War with France (1797), the interval is so close as to be non-existent and the variation so slight as to be detected only in the changing 'edition' label on the title. The View, published 8 A.M. Saturday, 11 February, went through nine 'editions' by the end of the following week, fifteen by the end of the month, twenty-five by the end of March, and thirty-five shortly thereafter.

Whether this tract actually comprised so many 'editions' was then and is now a matter of some concern. In his Letter to Erskine John Gifford remarks that he was unable to peruse a copy "until it had been deemed expedient to imprint 'The Ninth Edition' in the title-page," and somewhat later plaintively inquires:

Sir, permit me to ask you, whether the artifices which have been employed to provoke a forced circulation of your work, are such as are consistent with the dignity of your public character? As if determined to prove the justice of my suspicion, that you wrote for the populace, and not for the people, a cheap edition has been announced, men have been hired to hawk it about at the coffee-houses, and in the streets, and not a retailer of seditious publications in the populous suburbs of the metropolis but proclaims the amplitude of his stock by a bill at his door, calculated, by its enormous size, to catch the eye of the passenger.—In short, the very same manoeuvres have been exerted to extend the sale of your production as were used to circulate the treasonable tracts of your client—that infamous incendiary Thomas Paine.
To this Gifford appends a revealing note:
The public will be less surprized at the number of editions through which your work has so rapidly passed, when they are informed, that each edition consisted only of a fourth part of the number which it


Page 190
is usual to print; so that the fourteen editions of Mr. Burke's pamphlet[1] would be equal to fifty-six of your own. The public taste, therefore, in this instance, is not so depraved as it appears to be.—The press, too, has been kept standing, in order that any number of editions might be printed, without any additional expence to the publisher.[2]

Printing from standing type, the procedure Gifford regards as exceptional, was, of course, by this time a common device, evident not only in the 35 'editions' of Erskine, but in the 14 of Burke and, as a matter of fact, in the 11 of Gifford's own Letter. Whatever the number printed, the total count is thus, for each of these, equally suspect. Among these sixty only six may be correctly described as editions; all others are of lesser significance and uncertain denomination, ranging, so far as one may judge by appearances, from a completely differentiated impression to a single alteration at press—the 'edition' label.

When it comes to defining these and other doubtful specimens even the most scrupulous bibliographer is soon reduced to an expedient no less reprehensible than the one Gifford deplores. Since in these instances he is usually unable either to distinguish separate from continuous printing or to determine the relationship among points confined to one or extending to several 'editions', the bibliographer must designate all such curiosities by the most inclusive and least exact term at his disposal. 'Issue,' always implying a certain interval of time, cannot apply to these uncertainties. 'State,' often implying an alteration at press, is likewise inapplicable, contrary to the explicit declaration on the title and inappropriate for variations which, though unseen, may yet be as real as those existing in stereotype. 'Impression,' however, being much less restrictive, offers the bibliographer a way around the difficulties already mentioned, allows the printer every right to his claim ('impression' being, for him, equivalent to 'edition'), and imposes upon the reader the task of proving otherwise. In this context all that the bibliographer implies is recurrent printing, a somewhat indefinable condition affecting an indeterminate number of pages and extending over an immeasurable period of time.

Unfortunately, if the bibliographer assumes even this much, some inquisitive reader may discover a point apparently limiting the condition to the title and thus degrade the supposed impression to a variant state. Such evidence probably will be found, not in the defects earlier printers endeavored to remedy—in mispagination, omitted signatures, false catchwords, errata lists, and the like—all of which are now generally left


Page 191
unattended, but rather in slight alterations of the text. If these occur either as cancels persisting through several 'editions' or as readings mixed in various sheets obviously the gatherings so affected derive from stock available upon original issue. But again, the evidence, though disallowing 'impression' as proper reference for the entire work, is itself so qualified that 'state' may not everywhere apply. A hybrid of this sort appears in specimen C5.

Amid all these uncertainties I now present various examples of what, for want of a better word, may be described as recurrent impressions. As these progress, often within the same book, through several gradations subject to more precise definition, they are listed most conveniently by author and the variant identified by one of the six letters cited below. After each letter I indicate the significance of the term as employed in these circumstances.

  • (a) Edition. All or an appreciable portion reset. Since resetting may extend from 1 to 99 per cent of the type, 'edition' here refers to any book reset in two or more consecutive gatherings, an apparent indication that earlier distribution was intentional and later composition an undertaking not previously contemplated.
  • (b) Issue. A much-abused word, but useful in this analysis for anything less than an edition but more than an impression.[2a]
  • (c) Impression. A reprint, with or without revisions, issued possibly with other printings, but separately prepared and distinguished by a full array of press figures or other series of points.
  • (d) Recurrent impression. Certainly in part and perhaps in whole a separate printing; without the usual sequence of points, but exhibiting at least one other than title.
  • (e) Recurrent impression (assumed). Differing it would seem only in title, but presumed to vary consistently in other respects. A provisional designation for variants which, in the absence of other evidence, cannot be classified as d or f.
  • (f) State. Variant occurring within an impression.

When these criteria are applied to the twenty books listed below and the results posted in the accompanying Table it is apparent that, even upon the most liberal construction of the term 'edition', the 158 so described by the printer actually comprise only 28, and that of the 130 remaining no less than 82 fall into limbo e, beyond definition. Yet these imponderables are not to be disregarded, even in a bibliography limited to 'the first edition', for by definition of that term all are essential to it.


Page 192

Tabulation of Variants

Book  Total 
Edition Titled  14  12[4]   35[5]   12  10  158 
(a) actual[3] Comprising  28 
(b) issue 
(c) imp[3]   37 
(d) r. imp  32 
(e) r. imp?  10  20  82 
(f) state