University of Virginia Library

NEARLY a quarter of a century ago it was noted that a study of variant spellings in the First Folio Macbeth suggested the possibility that two compositors had set type for the play.[1] Later investigation supported the hypothesis and revealed the work of these same men in other plays of the Folio.[2] More exacting principles for a spelling test to detect compositors have since been laid down.[3] The test has been applied to other plays,[4] and in at least one instance the results have been verified by still another technique.[5] Although its accuracy seems validated by these studies, the test has thus far been employed in comparatively few cases. Therefore, it seems desirable to extend its use both to substantiate conclusions

previously drawn and to add to our knowledge of early printing house procedure.

Application of the spelling test to the 1593 quarto of Peele's Edward 1, printed by Abel Jeffes for William Barley, reveals the work of two compositors. However, they did not divide their material alternately throughout the play as seems to have been the case in other quartos studied. Compositor A set from the beginning at least through D, and Compositor B set at least from E4 to the end of the play. Although it seems likely that they are to be attributed to Compositor B, the remaining pages (E1-E3v) present some difficulties which will be discussed later. The evidence upon which these conclusions are based is given in the table below.

Most remarkable is the fact that through D, the word friar appears 60 times, and only once is it spelled frier. From E4 to the end of the play, the word appears 121 times, and 118 times it is spelled frier. The remarkable consistency of these variant forms indicates clearly the two-compositor pattern, but since the word does not appear on every page, it is necessary to seek out supplementary evidence.

Such evidence is ample. Compositor A wavers inconclusively between go and goe (7: 6), but his fellow shows a decided preference for goe (23 times), using, in his undisputed work, go only once, and that in a full line.[6] Much the same holds true for do.[7] While Compositor A uses doe only twice to fifteen instances of do, Compositor B has do only four times (all in full lines) in 33 appearances of the word. This suggests that wherever either go or do appears outside a full line the work should be attributed to Compositor A. However, this facile rule will not apply, since on E2v and E3 go occurs four times, only once in a full line, and other evidence indicates that the work is Compositor B's.

In speech ascriptions Compositor A tries several forms before he settles on Long. (for Longshankes), but he also uses Longsh. ten times. Compositor B much prefers Longsh., but three times he also uses Long. Compositor A generally shortens Lluellen to Lluel. (three times to Llu.), but four times he tell, the full name. Compositor B consistently uses the complete name, abbreviating once in a full line to Llue., a form never used by Compositor A. In speech ascriptions and in the text Compositor A uses Glocester (abbreviating Glocest. and Gloce.) in his work Gloster occurs twice, no doubt through the influence of copy. Compositor B uses Gloster (abbreviating occasionally to Glost.) exclusively.

In two categories of words the practises of the two compositors diverge. For words like well, Compositor A ordinarily has the normal spelling with some -I variants; Compositor B uses the -I variant far more frequently than does his fellow.[8] For words like say Compositor A again likes the normal spelling, but

he also employs an -aie variant. Compositor B prefers the -aie variant.[9] While, as the main table will show, this evidence is not trustworthy for any individual page, the following tabulation indicates its significance for the mass of work.      

-ll	-l	-ay	-aie
A	180 (91.4%)	17 ( 8.6%)	58 (78.4%)	16 (21.6%)
B	187 (61.3%)	118 (38.7%)	47 (32.9%)	96 (67.1%)

Even though the variants in these two groups represent a long and a short form, it is unlikely that line length exerted much influence over the choice of spelling. In the first place, Compositor B most often uses the long -aie form, but he also uses the short -l form with greater frequency than does Compositor A. Furthermore, investigation of some pages of high concentration of a particular form demonstrates that line length had little to do with spelling. For example, F2, a page of doggerel with short lines, has a high number of the long -aie forms, coupled with heavy use of short -l forms, the characteristics of Compositor B. B3v has Compositor A's highest concentration of the shorter -ay form, but on that page occur only two full lines, neither of which has an -ay spelling in it. B4, a page of prose, uses the long -ll and short -ay forms exclusively, the characteristics of Compositor A. It seems obvious that habit, not line length, dictated the spellings to be used.

Thus far, there seems little doubt about the evidence, but the problem of E1-E3v is more complex. The friar/frier spellings are ambiguous; on E3v, the only questionable page on which the word appears, friar occurs twice, and frier five times. Go occurs four times on E2v and E3, suggesting strongly that those pages were set by A, but contradictory evidence appears in the three doe forms (E1v, E2v, E3), which provide as distinct a vote for Compositor B as the go forms do against him. Long., a rare form for B, appears three times on E2v, seven times on E3; but on one of those pages (E2v) Longsh., a B form, occurs four times, more often than on any single page of A's undisputed work, although A does have this spelling more frequently than B has Long. On E3v Lluellen, a B form (which A occasionally employs) occurs three times. Gloster spellings, a very strong evidence of B, occur eleven times on E1v-E3 (seven on E2). Compositor B's -l variant is used twenty times to 32 -ll spellings in those pages. His -aie variants appear on only two pages, and not in significant number.

Excluding the -ll/-l and -ay/-aye categories, in which there is the most frequent mixture of A and B forms, the spellings may be tabulated thus:

     

	EI	EIv	E2	E2v	E3	E3v	Total (EIv-E3v)
A	5	6	8	2	16
B	3	7	6	2	1O	28

E1 seems certainly to be Compositor A's work, since it has only A spellings, including one instance of Glocester and two of Lluel., spellings never used by B. For E1v-E3v Compositor B appears to have a slight advantage. The only real difficulty in assigning these pages to Compositor B is the extremely high number (10) of Long. abbreviations on E2v and E3. Furthermore, the high number of Gloster forms would be just as surprising for Compositor A. Therefore, it seems best to make the line of demarcation between E1 and E1v.

This division is borne out in part by two additional spelling variants which have not been included in the tables. Compositor A consistently spells eye(s), and Compositor B spells eie(s).[10] Eie and eies appear on E2. Compositor A keeps to neere, and Compositor B has neare.[11] Neare occurs on E2. This seems definite proof that E2 is to be assigned to Compositor B; therefore, presumably E2v-E3v are also his, since it is unlikely that compositors would divide their work by single pages. On E1v is the ambiguous form nere, evidently influenced by a rhyme with here, but E1v also has two Gloster spellings, more than any single page of Compositor A's undisputed work.

Departures from characteristic habits on these pages, as well as elsewhere, are doubtless to be explained in part by the influence of copy. Apparently the copy had friar, from which Compositor A once inadvertently departed, and which Compositor B followed on five occasions in spite of his habit of spelling frier. The copy probably read Gloster throughout, which led to Compositor A's twice using that form although he was normally a Glocester speller. There is no assurance that the copy was any more regular than the printed text in abbreviating the speech ascriptions for Lluellen and Longshankes, but assuming that it generally had the full names seems the likeliest explanation for the compositorial divergences. Compositor B would, then, have followed his copy for Lluellen; Compositor A normally departed from it, but on four occasions was led to use the full name. So it is also with Compositor A's treatment of Longshankes, which he occasionally has in full and which is abbreviated Lon., Long., and Longsh.; his fellow was more consistent. Go and do in the copy might explain the departures from his normal spelling in B's work, but assuming that the copy had a norm in these words seems risky. Similarly, in the -lland -ay groups the copy no doubt showed much the same variety as the printed text, and it would be futile to attempt to ascertain any particular spot in which copy influenced the compositor. Generally, it appears, Compositor B was the steadier speller.

The difficulty in assigning E1v-E3v points up an important lesson to be drawn from a study of compositorial habits: one must not be overscrupulous in demanding consistency from his compositors. Obviously, whichever compositor set up those pages departed from his normal procedure. As a further illustration of

this point, from about I3v through the end of the play the use of the -I variant drops noticeably. However, on I3v, I4, K2v-L1, and L2 appears the trier variant, which is the hallmark of Compositor B. Throughout most of this section one also finds Compositor B's -aie forms. Inconsistencies occur on individual pages throughout. HI is surely the work of Compositor B, but he uses the -l variant only once in eleven possibilities. Conversely, Compositor A quite suddenly uses the -l form five times out of eleven on A4v. Examples can be multiplied to demonstrate that for no apparent reason a compositor may within a given page show characteristics more like those of his fellow than his own.

Whatever conclusion one may reach about E1v-E3v, the major proposition seems proved beyond any doubt: two compositors did work on the play, dividing it not quite evenly. There is no evidence to suggest that they worked simultaneously. In fact, since two sets of running titles can be traced throughout the book, composition and printing evidently proceeded seriatim.[12]