American Type Founders

"Collective Specimen Book" (1895/6)

[A Bad Scan of the Public Domain Portions of the 1981 Garland Reprint]

image link-topic-sf0.jpg

1. Introduction

This was the first and most extensive specimen published by ATF during the period before its consolidation into a single Central Plant which presented a comprehensive showing of the offerings of its various manufacturing foundries. There were other "company-wide" ATF specimen books in the late 1890s, but by that time the process of reducing the number and diversity of faces had begun. This 1895/6 "Collective Specimen Book" is a magnificent example of a late high point of Victorian ornamented types. It is also particularly valuable to the student of 19th century types because in most cases it associates faces with their originating (or sometimes their manufacturing) foundries.

Stephen O. Saxe, who owns copies of the Minneapolis and New York versions of the original edition, notes that some of the pages in the front section of the original are in color. These are in black-and-white in the Garland reprint.

Unfortunately, it is also little-known and difficult to acquire. Many people well versed in the history of type have never even heard of it. The original edition is scarce and has not been digitized from the original. It was reprinted once, by Garland Publishing in 1981 in their "Nineteenth-Century Book Arts and Printing History" reprint series. The Garland reprint contains a valuable Bibliographical Note by John Bidwell and an Introduction by Alexander Lawson. But even the Garland reprint is now scarce and increasingly expensive (as I write this, a spot check reveals no copies for sale anywhere). If you can find and afford a physical copy of the Garland reprint, it is well worth having:

American Type Founders Company. Specimens of Type, Brass Rules and Dashes, Ornaments and Borders, Society Emblems, Check Lines, Cuts, Initials and Other Productions of the American Type Founders Co. Intro. by Alexander S. Lawson. Bibliographical Note by John Bidwell. NY: Garland Publishing, Inc., 1981. [No. 15 in the Garland series "Nineteenth-Century Book Arts and Printing History."] ISBN: 0-8240-3889-4.

Before I acquired my copy of the Garland reprint, I did a hasty scan of a library copy so as to have this material for my own reference. But this material is important to our understanding of the history of type and of type-making in America; it should be more widely available. This present Notebook reprints the portions of my scan of the Garland reprint which are in the public domain (the original 1895/6 material). Regrettably, it must omit the fine Bibliographical Note by John Bidwell and Introduction by Alexander Lawson; these are still in copyright.

AESTHETIC NOTE: Yes, this is an ugly scan. But it's better than nothing and I'm not going to unbind my copy to do a better one. (Additionally, the technical qualities of the Garland reprint and, possibly, the presswork of the original aren't as crisp as they might be.)

2. Reading the Specimen

Titles and Versions

In his Bibliographical Note, Bidwell calls this book "formidably inconsistent." It exists in several versions, each assembled to suit a particular ATF selling house. The earliest version may have appeared in October of 1895. At least nine versions were published through 1896, with substantial differences between versions for the eastern and western markets. Various title pages exist. The pagination is complex. The version presented here follows the page sequence of the 1981 Garland edition (which has an MSJ / Philadelphia branch title page).

The book is never called the "Collective Specimen Book" on its title page. That term appears in its preface, and is a useful name to apply to a book the copies of which exhibit so much individual variation.

Foundry Numbers

One of the valuable aspects of this specimen book is that it associates faces with the originating or manufacturing foundries within ATF. But some of this is done in code, and even when it isn't the history of ATF and the foundries it amalgamated has become garbled in late 20th century print and 21st century digital references.

At the beginning of 1892, there were at least 36 type foundries operating in the United States. The February 1892 formation of ATF merged 23 of them (only two-thirds of the total) into a single, but by no means yet unified, business entity. The foundries which merged to become ATF were listed in a trade note in The Inland Printer in November 1892 (Vol. 10, No. 2, pp. 150-151).

Note that it is necessary to understand the distinction between a foundry name (such as the Dickinson Type Foundery) and the name of the business which owned it (Phelps, Dalton & Co. for the Dickinson in 1892). Failure to understand this distinction has marred several accounts of the formation of ATF both in the print literature (e.g., Consuegra) and the digital (e.g., Wikipedia).

The following foundries which existed in early 1892 were not a part of the formation of ATF, although some of them later were acquired by ATF. Several of these foundries were among the most important of their day (e.g., Bruce's, Farmer, BB&S, or the Keystone).

(For a detailed discussion of the incomplete evidence concerning the early history and consolidations of ATF, see the Notebook on American Type Founders: Early History Through 1906.)

The 23 foundries were each assigned a number, but as they were consolidated rapidly these numbers did not long endure. At present, we do not even know what all of them were. They have been used in two situations.

First, in a very few cases these original foundry numbers were incorporated into the names of typefaces. The standard example of this is Caslon Oldstyle No. 471, which was face No. 71 of MacKellar, Smiths and Jordan (which became ATF Foundry No. 4). It is identified as No. 71 in this Collective Specimen Book. Later the MSJ foundry number, 4, was prepended to this. (But note that these numbers as a part of the names are not related to the ATF series numbers more familiar to later printers and to letterpress printers today. The series numbers were introduced in the 1930s; these designations are much older. "Caslon Oldstyle No. 471" is ATF series 50.)

Second, and of interest here, these original foundry numbers frequently were used in this Collective Specimen Book to identify the "originating" or the manufacturing foundry associated with a particular face, collection of ornaments, etc. So (to take the most common example) "Excelsior Music" on p. 17 has the code "B4". The 'B' is for "Branch" and the '4' is the foundry number of the former Central Type Foundry in St. Louis.

Our knowledge of these foundry numbers is tenuous. They appear to have been forgotten, and were never fully published. Stevens L. Watts, a salesman and manager at ATF whose near-subversive activities preserved an important part of our typographical history, wrote a letter dated July 23, 1965 to the typefounder and scholar Paul Hayden Duensing in which he identified at least nine (possibly more) of these foundry numbers. This letter has never been published, but Duensing provided a copy of it to John Bidwell and he, in turn, gave the foundry numbers which appear in this Collective Specimen Book:

(Actually, I haven't spotted where B23 appears in this book yet.)

Foundry Letters

Additionally, at some point during the consolidation of the initial companies into eight (later nine) manufacturing foundries, these foundries were assigned letters. These foundry letters were never used externally in literature or in designations visible to customers, but they are relevant to ATF type casting machine mold designations. They are also referred to by Bidwell in his Bibliographical Note. These foundry letters are:

Foundries Listed and Actually Existing

The "List of Foundries" which appears in this Collective Specimen should also be treated cautiously. For example, it lists both the Dickinson Type Foundry [sic, not Foundery] and the Boston Type Foundry. The date of the specimen reprinted by Garland is 1896. Yet we know from the ATF Annual Report for the fiscal year ending Aug. 31, 1895 that:

"At Boston the new foundry built for the Company at Wales' Wharf and taking under its roof the old plants of the Dickinson and Boston foundries has proved more than satisfactory."

So although they are both listed by name, neither the Dickinson nor the Boston Type Foundries had any legal or physical existance when this specimen book was published. You might have searched the streets of Boston all day and would never have found them. The same is true of the former North-Western Type Foundry of Benton, Waldo & Co. (merged into the ATF NY Foundry B), the former Franklin Type Foundry of Allison & Smith (merged into the ATF Cincinnati Foundry D), and the former Cleveland Type Foundry of H. H. Thorp. Mfg. Co. (merged into the ATF Chicago Foundry E).

By the August 1, 1895 ATF corporate Annual Report (and assuming that the Central T. F. and the St. Louis T. F. had been consolidated into a single plant), eight manufacturing foundries survived. These were the ones assigned letters. Here is the "List of Foundries" which appears in this Collective Specimen Book arranged against the manufacturing foundries which actually existed at the time the book was published:

List Reality
MacKellar, Smiths & Jordan Foundry ATF Foundry C, Philadelphia
Dickinson Type Foundry ATF Foundry A, Boston
Boston Type Foundry ATF Foundry A, Boston
Central Type Foundry ATF Foundry F, St. Louis
Marder, Luse & Co. Type Foundry ATF Foundry E, Chicago
Cincinnati Type Foundry ATF Foundry D, Cincinnati
Benton-Waldo Type Foundry ATF Foundry B, New York
Conner Type Foundry ATF Foundry B, New York
Allison & Smith Type Foundry ATF Foundry D, Cincinnati
Cleveland Type Foundry ATF Foundry E, Chicago
John Ryan Type Foundry ATF Foundry H, Baltimore

The fact that Robert Allison was the first president of ATF and L. B. Benton was on the board of directors perhaps accounts for the continued presence of their former foundries' names in this list.

The "List of Foundries" does not mention ATF Foundry G, San Francisco (formerly Palmer & Rey), but evidence of its actual operation during this period is scant and it was often omitted in ATF advertisements which listed their other foundries.

A number of types in the book are identified as "American Type Founders' Company, New York." It is not yet clear to me whether this indicates the ATF NY Foundry B, specifically, or whether it simply means the company as a whole (which was then based in NY).

Further Complexities

A few faces are coded with a compound number. For example, "Broadface Series," p. 60 (PDF 38 below) is coded as "B4-21" and "Book Series," p. 61 (PDF 39) is coded as "B4-22". But the "21" and "22" cannot be foundry numbers, as "Large-Face Series [No. 54]" is coded "B4-24" and there was no foundry 24. An examination of the faces indicate that the first number of the pair is the foundry number. It isn't obvious what the second number might be.

Other faces present further complexities of attribution. The Wilson Series, for example, is shown over three pages (68-70 / PDF 46). Sizes from 6 to 18 point are shown on the first two pages and are associated with foundry 1 (in two-part codes: B1-4, B1-5, B1-7, B1-17), which was the Dickinson Type Foundry. This makes sense, as "Wilson" was a face by Alexander Phemister for the Dickinson Type Foundery. But the third pages shows sizes from 22 to 48 point and says "Manufactured by Boston Type Foundry." On its own, this would seem to associate the face with the former Boston Type Foundry (competitor to the Dickinson); this cannot be the case. So in this instance "by Boston Type Foundry" might just mean "by ATF foundry A, the Boston branch."

As noted earlier, the ATF series numbers familiar to letterpress printers today were introduced decades after this specimen book was published. The series numbers given for various faces here have no necessary relation to them.

It is also useful to have a general knowledge of the pre-point-system type body sizes: agate, nonpareil, etc.

3. The Collective Specimen Book Piecemeal

My scan of this book comes to 758 pages, 18 of which are in copyright and cannot be reprinted. That still leaves 740 pages which, even compressed, comes to almost 10 Gigabytes. This exceeds the network capacities of many people and makes its presentation as a single PDF file impractical. (But see here if you really want it that way.) It is also a potentially confusing book which should be searchable by names (and OCR isn't really up to the task of 19th century ornamented typefaces yet).

Here I've divided the scans into several hundred PDF files, many of which are only a single page long. In the list below the first number or range of numbers (if present) is the pagination of the original (possibly inconsistent or with gaps). The next field is a description of the contents of the PDF. When a foundry association is given it is in parentheses afterwards. No distinction is made in the list between "originating" and "manufacturing" foundries. No attempt has been made to analyze these attributions. I've left the 'B' designations as presented, without attempting to translate them into foundry names. "MSJ" is MacKellar, Smiths and Jordan.

Pages noted as "missing" do not appear in the Garland reprint. They may or may not appear in other copies of this book.

If there are too many small PDFs here to suit your preferences, and if you have sufficent network bandwidth, see the section below containing The Collective Specimen Book In Big Chunks.

4. The Collective Specimen Book In Big Chunks

This is a series of PDF files dividing the Collective Specimen Book into 35 chunks ranging in size from 22 Megabytes to 559 Megabytes (mostly at the larger end of this range). They contain the same contents as the "piecemeal" version above, but may be easier to work with in some environments.

The divisions here are sometimes arbitrary and the descriptions of the individual PDF files are abbreviated.

If you have a good network connection and like single-file versions, you may wish to try The Collective Specimen Book In One File (below).

5. The Collective Specimen Book In One File

Here it is as one very large file (9.6 Gigabytes). Don't try to download or view this unless (a) you are sure your network bandwidth is up to it and (b) your receiving system or medium can handle a 9.6 Gig file.


(Note: This file opens in Okular, but mupdf can't handle it. I haven't tried commercial PDF viewers.)

6. Digital Production Details

This scan was done on an Epson flatbed office scanner capable of A4 size, which is slightly smaller than the Garland edition page size. Some cropping of the image area was therefore necessary. The scan was done at 600 dpi in RGB and page images were saved in PNG format. The total size of all of the PNG files is 45 Gigabytes.

I have retained these PNG images, but they are too large to present online at present. In any case, the other issues with this scan mean that there is no visually apparent loss in a single-generation conversion to JPG format. The resulting JPG files total 9.6 Gig.

These JPG image files have been integrated without further loss into PDF container files using Johannes Schauer's img2pdf utility.

7. Notes

1. There is actually some ambiguity in the list of foundry numbers as Bidwell presents it. Since there were 23 original foundries and since the highest number in Bidwell's list is 23 (formerly Palmer & Rey), it is reasonable to conclude that the 23 foundries which merged in 1892 each were assigned a distinct number. But Bidwell's list does not associate numbers with foundry names. Instead, he associates numbers with ATF manufacturing foundry letters. These were assigned at some point in the 1890s as the foundries were merged. So for example Bidwell actually says that No. 1 was "ATF Foundry A at Boston, formerly the Dickinson Type Foundery" and that No. 2 was "ATF Foundry A at Boston, formerly the Boston Type Foundry." In most cases there is no confusion; either Bidwell distinguishes between foundries (as in the Dickinson vs. Boston) or there is only one foundry mentioned. But in two cases there is an ambiguity.

Bidwell says that No. 3 was "ATF Foundry B at New Your City, formerly James Conner's Sons and Benton, Waldo & Co." Now, Foundry B did include those two foundries (and also the Manhattan Type Foundry and A. W. Lindsay). But only one of these was the original No. 3.

I have not yet been able to disambiguate this case by identifying types designated B3 and known to be from one foundry or the other.

Bidwell says that No. 12 was "ATF Foundry E at Chicago, formerly Marder, Luse & Co. and the Cleveland Type Foundry of H. H. Thorp. Mfg. Co." Foundry E included these two foundries as well as the Union Typefoundry. Only one of them could have been No. 12 in the original ATF amalgamation.

But "Series No. 57" is identified as "Manufactured by Marder, Luse & Co." on p. 54 / PDF 32 and as being from "B12-17" on p. 65 / pDF 43. Since the first number in a two-part code is the manufacturing foundry number, foundry 12 must be the former Marder, Luse & Co.