Booth Tarkington's Penrod (Doubleday, Page &
Co., New
York, 1914) has sold close to a million copies and has been rated
as one of the outstanding American books of the first half of this
century. Moreover, it has proved an interesting 'guinea-pig' for
twentieth-century bibliographers.
The first to approach the problem was Barton Currie, who in a
1932 Colophon article (vol. III, pt. 9) entitled "Hints to
Tarkingtonians" discussed the then known points in an attempt to
establish priority for interested collectors: namely, that the
first "issue" was bound in light blue mesh cloth and measured
one-eighth of an inch more across the top of the leaves than later
"issues," which were bound in greenish-blue grained cloth and
printed in blacker ink. Later in the same year appeared Currie's
Booth Tarkington, A Bibliography, which included the
Colophon article as an introduction and added the further
important information that 20,000 copies were printed before its
publication on March 26, 1914. In 1938 Jacob Blanck published some
new points[1]: the earliest copies
had the numeral 'viii' at the foot of that page, which was
"partially or wholly removed" in later copies,
and the first 'state' could be distinguished by the spelling
'sence' instead of the later state's 'sense' in line 23 on page
19.
After considerable scholarly research D. R. Russo and T. L.
Sullivan produced in 1949 their comprehensive Bibliography of
Booth Tarkington 1869-1946, which considers the above points
and two others which were inadvertently attributed to Currie
through an error in the revision of the final manuscript.[2] The second of these latter two
points
was: "p. 344, line 6, 'l' present in 'perfectly', later dropped
out." The difficulties of this account, together with a
dissatisfaction with the bibliography's use of 'state'
to avoid the much abused 'issue' where sufficient information was
not at hand to indicate an impression, provided the incentive for
the research which has led to this present paper.
[3]
In the first place, it did not seem possible for the missing 'l'
on page 344 to have dropped out after printing started, for the
sheets seemed to have been printed from plates and thus the 'l'
ought to have been lost before the plates were made. If so, a type
could not have dropped unless the setting had been made by
monotype. Mr. Powers, an employee at Doubleday Press, who had been
with the Country Life Press in 1914, confirmed these assumptions by
stating that at the time Penrod was printed plates were used
and monotype machines set the type. Further evidence that the 'l'
was added to the plate came to hand when it was seen that the
letter in question is shaky and malformed and that the preceding
't' and the following 'y' are not properly formed. This line of
investigation established that the sheet for gathering [22], a
normal quire of eight leaves in which page 344 is sig.
[22]7
v, was first printed with the 'l' missing.
Comparison then demonstrated that both states of gathering [22]
were indiscriminately distributed among the fourteen copies
examined for evidence.
The following table lists the variant copies in three groups for
convenience of future reference:
- I (a) Page viii numbered; p. 19 'sence'; 'l' missing p.
344.
- (b) " " " ; " " ; 'l' present p. 344.
- (c) Page viii unnumbered; " " ; 'l' missing p.
344.
- (d) " " " ; " " ; 'l' present p. 344.
- II (a) " " " ; p. 19 'sense'; 'l' missing p. 344.
- (b) " " " ; " " ; 'l' present p. 344.
- III (a) " " " ; " " ; 'l' missing p. 344.
- (b) " " " ; " " ; 'l' present p. 344.
- Note: I & II bound in sky-blue mesh cloth. III bound in
sky-blue ribbed cloth.[4]
With the exception of the state of 'presently' on page 344 and
the elimination of the page number viii, the nine copies of Group
I are sufficiently uniform to indicate that they came from the same
impression. The "Ross" copy, classed as I(b), which is in the Smith
Library of the Indiana Historical Society, contains a presentation
inscription from the author dated March 29, 1914, to Morris Ross,
then managing editor of The Indianapolis News. The presence
of the second state of gathering [22] in this copy, which was
inscribed only three days after publication, is of interest as
evidence that the dropped 'l' was inserted in the plate of page 344
before any copies were bound. It is also worth notice that the
correction must have been made late in the first press run or
between the first and second printings noted below, owing to the
number of sheets printed without the 'l', as evidenced by the first
state of the sheet in the "Wing" copy at Yale, III(a).
A comparison of the three copies in Group II with the nine in
Group I produced conclusive evidence that the plates had been
handled and a new imposition made before they were printed. The
correction of 'sence' to 'sense' on page 19 is so perfect that it
almost certainly indicates a resetting and a new plate. Even if we
disregard many damaged page numbers and much damage to the type,
which could conceivably have been caused by linting or press
batter, the damage to three particular plates is worthy of
consideration as evidence of handling. The right margin of page
153, lines 5 to 15, was quite obviously damaged by the ratchet key
slipping when the plate was being locked to the bed for a new
impression. Considerable damage to the type on page 167, running
diagonally down the page from the heading to the lower left,
appears to have been caused by handling the plate;
the same is true of the damage to lines 2 and 3 from the bottom of
page 182, which is the conjugate of page 167. Definite proof of a
new imposition, however, is provided by the fact that the plates of
the two half-titles, pages [i] and [1] (sigs. [1]1
r and
[1]6
r), are reversed. In the first half-title in Group I,
which is printed as the second half-title in Group II, there is a
diagonal white line running across the shadow on the girl's
forehead which does not appear in the second; and the short curved
line at the lower left of the boy's coat shows as an elongated "S"
in the second half-title.
Whereas the evidence in support of a separate impression for
Group III is not so conclusive as that for Group II, it appears
likely that there was indeed a third impression. There is new, if
slight, margin batter which points to ratchet damage having
occurred during a new imposition: see especially page 161, second
line from the bottom, right margin; page 165, line 14, left margin;
and page 232, right margin of lines 2 and 3 from the bottom. The
damage to pages 167 and 182 in Group II has been
partially repaired by reshaping some of the type faces and by
morticing in new type. The latter process was undoubtedly used on
the second line from the bottom of page 182, where through an error
an inverted comma was inserted in place of a semi-colon which had
been battered above the line. (This damage could have been repaired
in press, but as corroborative evidence and as an interesting
example of repair work it is included here.) The sheet bulk in
Group III averages 3/32 of an inch less than that in Group II; and
the binding is ribbed cloth instead of the mesh of the earlier
copies.
The last two points may assume new significance when considered
in connection with the following supposition: That the publisher's
estimate had to be expanded by two additional printings owing to
the popular interest created by the magazine serialization which
overlapped the publication date for the book. This was not a novel,
but a collection of stories about a boy and his dog and his
friends. Therefore, it seems reasonable that a printing of 10,000
copies was first planned, with two additional printings of possibly
5,000 each called for prior to publication. Unfortunately, no
information was obtainable from the printer or publisher regarding
the printings of this book.
There can be little doubt that Group III also represents a
pre-publication printing on the evidence of the presence of the
first state of gathering [22] in the "Wing" copy at Yale, III(a).
It is unreasonable to suppose that any of those sheets remained
unused or undestroyed during the interval of more than two months
between the March publication and the next 3,750 copies which were
brought out in June (Currie, Bibliography, p. 63). I regret
that I have been unable to find any copies of this later printing
for comparison with those already examined; no doubt some
interesting and instructive variations would be found.
It may be of more general interest to conclude by listing a few
points which developed from the study of Penrod and which
represent printing evidence too often overlooked or misinterpreted
by bibliographers of twentieth-century books. I hope to develop
these on a future occasion.
- 1. It is almost safe to assume that plates were made before
the first printing, but it is possible that some books were first
printed from standing type in small impressions until such time as
a demand for copies warranted the expense of plating.
- 2. It is well to determine if the type was set by monotype or
by linotype, and whether it was plated before printing
started.
- 3. If the type was plated, a dropped letter would indicate
(a) that it was dropped prior to plating, and (b) that the type was
probably set by monotype. A type set by linotype can appear to be
dropped only if the compositor has inadvertently hit the spacer, an
unusual accident.
- 4. Broken type does not, per se, denote a later state or
printing.[5]
- 5. Broken type can be repaired and plates corrected to some
extent in press by morticing; i.e., by cutting out faulty
letters and soldering in new type. (Mr. Richmond of Doubleday Press
is of the opinion that in adding the 'l' on p. 344 the 'y' was also
morticed out to provide more working space, which would account for
the malformation of both letters in the later state.)
- 6. Damage to upper and lower page numbers and the headlines
tends to indicate handling of the plates between impositions, but
is not conclusive evidence.
- 7. However, ratchet batter to the margins of the plates, as
evidenced by distortion or bending of the type face, is fairly
definite proof of a new imposition. Such damage is the result of
carelessness in permitting the ratchet key to slip while locking
the plate to the bed.
- 8. In the case of bad damage, the press run on the sheet is
stopped until a new plate can be made, which may take anywhere from
a few hours to a few days before printing of that sheet can be
resumed.
- 9. Sheet bulk does not vary to any great extent within an
impression; however, any variation should be substantial before
considering it as an indication of a separate printing.
- 10. Variations in the blackness of the inking can occur
within the impression; one of the most difficult problems in book
printing is holding even color from forme to forme.
- 11. Variant sheets can become mixed during the folding before
any copies are bound, as the skids on which they are stacked are
not necessarily delivered to the folding machine in any
pre-arranged sequence. (During the war shortage when different
papers had to be used, serious problems developed in binding
together only gatherings of the same paper stock.)
