Donnelley and Lee Library Archives and Special Collections at Lake Forest
Internet-Based Railroad Research
Historic, Narrowly-Focused U.S. Railroad Research on the Internet
David W. Mattoon ‘76, Special Collections Research Associate
Elliott Donnelley Railroad Collection, Donnelley and Lee Library
Lake Forest College, Lake Forest, Illinois
The Internet provides a wonderful opportunity for the armchair researcher with a computer to conduct remote research. The information available is literally expanding daily as Google—among others—add information to the Internet.
The techniques and sources I use to answer railroad questions are described below. This article is best used in an open browser window or browser tab so the hyperlinks can be used directly in your research. The embedded links are too long for me to convert to a PDF document.
Most of the techniques or sources can be applied to more topics than implied by the title of this article, though some are railroad-specific.
The “U.S.” is a reference to that fact that many articles and books are available for full download if the U.S. published books were copyrighted prior to 1923. Some later articles are also available for full download or preview. The public domain law for U.S. Copyrights is described in the Wikipedia article http://en.wikipedia.org/wiki/Public_domain_in_the_United_States and are quite complex.
“Narrowly-Focused” refers to the fact that these techniques are especially suitable when the topic being researched is limited in some respect such as time or geography. For example, a search from Google www.google.com will easily locate information on Grand Central Terminal. Many or most public and school libraries have books that include descriptions of Grand Central Terminal. Or the reference librarian can obtain those books through inter-library loan. However, detailed information on the building of Grand Central Terminal while Grand Central Station was still in use is more difficult to find from these sources. A technique shown below will show how searches can be limited to 1900-1913. Similarly, you may be interested in a small railroad station in some community. The techniques below will show how this can be researched as well. These techniques often work best when they are applied cyclically—searches from one method will provide keywords or ideas to try on other methods.
Step 1: PATENT?
An actual railroad artifact, such as a railroad lantern, a patent search may be useful. It can establish a rough can’t-be-built-before date though the object might be in use several years before the patent was granted. Historic searches used to require looking at the Patent Index—a large set of books at a major library with limited indexing. Patents can now best be searched for historic studies using Google’s Patent Search (http://en.wikipedia.org/wiki/Google_Patents ) at http://www.google.com/patents?hl=en or by starting from Google’s home page http://www.google.com/ , moving the mouse up to make the menu visible, selecting “More” from this Menu, “Even More” from the next menu, and the “Patent Search” link from the “Search” section of this page: http://www.google.com/intl/en/options/
The patents have been scanned into an OCR (optical character recognition http://en.wikipedia.org/wiki/Optical_character_recognition ) system but at least some have not been proofread by a human. For example, clicking the Advanced Patent Search link and entering 854803 in the Patent number field takes you to this patent http://www.google.com/patents/about?id=SFttAAAAEBAJ&dq=patent:854803. At the time this is written, Google has this summary information:
Filing date: Sep 20, 1808
Issue date: May 28, 1907
However, by clicking the “Read This Patent” key, you can see from the final line of the header that the correct filing date is clearly 1905. The wrong date also means that sorting by date is problematic.
To be fair, Google Patent searches are in Beta (http://en.wikipedia.org/wiki/Beta_software#Beta ). The Wikipedia article on Google Patents http://en.wikipedia.org/wiki/Google_Patents has links that leads to other articles that cover the limitations in more detail. In spite of the OCR misspellings, the keyword searching makes this a valuable service. (But because of the misspellings, among other reasons, it should not be used to determine “prior art” when filing a patent.)
One of George Westinghouse’s patents on air brakes is here http://www.google.com/patents?id=pepCAAAAEBAJ&zoom=4&pg=PA2#v=onepage&q&f=false. The fact that the title is “IMPROVEMENT IN STEAM-POWER-BRAKE DEVICES” shows the advantage of having a keyword search as it would likely not be spotted in a casual title search. (Though it does make sense since locomotives used to have steam-activated brakes.)
Seemingly the more logical choice is to go the U.S. Patent and Trademark Office (www.uspto.gov/patents/process/search). However, patents prior to 1976 cannot be searched by keyword here, only by patent number. The TIFF (http://en.wikipedia.org/wiki/Tagged_Image_File_Format) images require a browser plug-in as described on the USPTO page. Installing this software may require “administrator privileges”. If so, if would be difficult or impossible for a “guest” to install it on a public computer at a library.
Step 2: Railroads, Railways, and Trains
The Operating Department (of a U.S. railroad or railway) definition of a Train is: One or more locomotives with or without one or more cars displaying markers. A railroad or railway is a corporation where some employees operate trains, some employees maintain the track the trains run on, some employees work on signaling, and some perform human resource tasks unrelated to trains. Are you researching a railroad or railway or researching a train? Railroads and railways are basically synonyms. Railroads that went bankrupt often emerged as Railways. For example, the Union Pacific Railroad was reorganized as the Union Pacific Railway in 1880. The terms overlap, however, in railroad station and train station.
Step 3: Background Information – Wikipedia.
Wikipedia is a user-written encyclopedia which covers a broad spectrum of information. The main web address is http://www.en.wikipedia.org where “en” stands for the English language version of Wikipedia. The good description of Wikipedia is in Wikipedia at http://en.wikipedia.org/wiki/Wikipedia:About . Enter information in the upper-right search box. For example, entering “railroad watch” leads to http://en.wikipedia.org/wiki/Railroad_watch for the article on “Railroad chronometer.” (Wikipedia searching is “indexed” with the term most likely to be search for rather than the term that is most precise.)
Wikipedia is a good tool for finding the full name (expansion) of an acronym or initialism. Disambiguation pages will show the context of possible expansions and therefore led from AAR to Association of American Railroads or ICC to Interstate Commerce Commission or Illinois Commerce Commission.
Wikipedia will also lead from the informal name of a railroad (e.g., “North Shore Line”) to the formal (“Chicago, North Shore and Milwaukee Railroad.”)
The Wikipedia article may not answer the question. But its “bottom matter” will lead to other articles and sources. “See Also” articles are within Wikipedia whereas “References” and “External Links” are not.
If the research topic is related to a specific railroad, Wikipedia will often lead to historical societies that study that particular railroad. Entering “Milwaukee Railroad” leads to its informal name: Milwaukee Road; its formal name: Chicago, Milwaukee, St. Paul & Pacific; and the group that studies it: Milwaukee Road Historical Society. The text of the article shows that literature searches prior to 1928—especially those much earlier that that reorganization date—should refer to the Chicago, Milwaukee and St. Paul. Initials of railroads (e.g., CNW for Chicago and North Western) can often be determined just by entering them into Wikipedia although a comprehensive list is in one of the external links to the Reporting Marks article http://en.wikipedia.org/wiki/Reporting_marks .
Step 4: Is there a book, book chapter, or article on the topic?
Search for the topic in Google Books. This site is constantly being revised. http://books.google.com As of this date, the home page has two parts: “Researching a Topic” and “Google E-books.” I have found “Researching a Topic” to be the better choice for out-of-copyright works since the results seem to load faster. Entering “Grand Central Terminal” with the double quotes goes to a page that used to be the home page of Google Books. On the left below the Any Time option is a Custom Range option. Here in our example, “Grand Central Terminal” can be narrowed down from 1/1/1895 to 12/31/1918. The date range has been extended by five years on each side to allow for advance planning articles and articles submitted to journals after completion. (The custom date screen opens at a calendar. This is also useful for precisely dating references given in a newspaper article of something happening “last Tuesday.”)
For even more control click the Advance Search button below the Search button. As of the date of this paper, this clears the custom date entered from the “Custom Range” option and thus the date range needs to be reentered in the Publication Date section of this page.
Results are given in one of four forms: Full View, Preview, Snippet, and No Preview.
Full view allows the complete book, magazine, or article to be viewed on a screen and to be downloaded in PDF format which can be read with the free Adobe Reader. If you don’t have this software, use www.google.com and enter the three words: adobe reader download. Note that while the Google Books version on the web is searchable, the downloaded PDF is not. Therefore, it is important to note the citation information on the web version so its location in the PDF can be determined. For example, the third result of the custom search returns “King’s how to see New York: a complete trustworthy guide book - Page 124.” The view of this page on the web shows that it is printed page 124. However, the PDF file has page 1 with the Google disclaimer, page 2 as a scan of the outside cover, page 3 as a scan of the bookplate probably on the inside front cover, etc. Blank pages and roman number pages are all included in the PDF page count. So printed page 3 occurs on PDF page 10 and printed page 124 occurs on PDF page 131. I recommend that when you save the file you append “PDF page 131” to the filename.
Scanned journals offer a slightly more complicated problem. For journals which are renumbered from 1 for every issue in a year, you need to note not only the page number of the article; but also the year and month, week, or day it came out; as well as the year and month, week, or day of the first issue in the bound volume. (The first issue may not be in January. Your article may be in a later year.) Also note if the issue information is listed at the top or bottom of some of the pages. Then with both the Books Google page open and the downloaded PDF, try entering a guess for the PDF page number and enter another guess or search sequential pages until you locate the PDF number. Again, I recommend including the PDF page number as “PDF page X” as part of the saved filename. The download process may require that you enter a Captcha (http://en.wikipedia.org/wiki/CAPTCHA) to ensure that you are human.
Limited Preview is available for some copyrighted works. This allows you to view a certain number of pages. Sometimes it is an introduction only. Other times it is only part of specific chapters. Certain pages will be restricted—often those with full-length illustrations.
Snippet View will provide a small excerpt from the page. Rarely this will include the information you want. More likely it will only help to confirm the book is describing the correct subject. For example, if the keyword is Station intended for railroad station, the snippet may show the particular work is about the Great Lakes Naval Training Station and hence not appropriate. Occasionally the snippet given appears to have no bearing on the keywords entered.
No Preview is provided for some works. Here you have no confirmation of the subject.
You should bookmark successful searches, copy the URL to a text file, or copy a page with PrtSc and paste the image into a file such as MS Paint or another screen capture program. The last will also record the keywords you used to locate the article if you capture the Results page. Many times it is difficult to locate the same article again because slight variations in keywords or keyword order will not produce the same results.
For historic railroad research, a first approach may be to select the “Limited Preview and Full View” option in the Advance Search page and only click the All Books option if you are unsuccessful. Similarly, you can start with precise keywords and choose more general ones if unsuccessful.
Regardless of the preview option, clicking a result brings up a page with more information. At times you need information from both the result page and the detail page. For a journal article in a Snippet or No Preview mode, the result page may give the page number and the detail page may give the volume number.
Sometimes on the left and sometimes on the right is the “Get this book” and “Find in a Library” option. The “Get this book” gives several sources where the book or magazine may be purchased. Another possible source is Addall http://www.addall.com/ and click on the Used Books tab. The “Find in a Library” option is one entry into the OCLC database. Entering the location such as a zip code or a city and state will produce a list of nearby libraries that have the book. Clicking on a specific library will enter that library’s online catalog. Occasionally the OCLC number will be out of date and the title or author will have to be reentered. Also occasionally the library will have no record of the book—presumably it has been discarded and not updated in the OCLC system. If the book is marked as circulating or on microfilm, it can probably be ordered through interlibrary loan by your local library. If the book is marked as non-circulating and you cannot visit the library in person, you might still be able to get an image of the pages you need by contacting the reference librarian. Some libraries have overhead scanners which do not strain the bindings the way conventional copies do—a computer built into the scanner factors out the distortion caused by not photographing it flat. I have also had good results asking librarians to send digital photos of articles. I indicate to them that I am interested in the information and don’t care if the images are keystoned (taken at an angle to prevent the flash reflecting back into the camera) or unreadable at 100% if the images can be read when zoomed higher in a photo image package. Basic photo image software is included when a scanner is bought (and might be included with the software when the computer was purchased). Adobe Photoshop or a powerful free program IrfanView (http://en.wikipedia.org/wiki/IrfanView ) or GIMP (http://en.wikipedia.org/wiki/GIMP ) can also be used.
When Books Google returns a volume and page number but not the month or week of a bound magazine, there are a few additional things to try. Try entering the title in Books Google. Occasionally the same article is printed in several magazines and one in the other journal is offered in Full View. Entering the title article in another Books Google search might find a reference for a popular article in a scanned version of The Reader’s Guide to Periodical Literature or an entry for a technical article in Engineering Index. Not all these issues have been scanned but doing a Books Google search on these indices will indicate a library that has them. Again, if you cannot visit it in person, an email to a reference librarian asking them to look up the citation for a formal interlibrary loan request may work. If the bound copy of a journal has one or more indices, ask them to look it up in the indices. As a last resort, especially for a monthly magazine, ask them to inspect all “page 68’s.” Check the Books Google pages for both a volume number and a year and check the library listings for volume numbers. A volume may run from July to June and the year would eliminate half the issues to be searched. Some institutions offered a paid search service if more extensive searching is required.
The OCLC page that lists the libraries where the book can be found also has a Cite/Export menu item. Clicking it and one of the options such as MLA (Modern Language Association) presents a citation that can be cut-and-pasted into a report. Be careful for bound magazines where the year given might be incorrect and additional information related to the title of the article, volume, etc. will need to be added.
Another way to reach Worldcat is http://www.worldcat.org/ . I have found this occasionally returns results that Books Google (using Worldcat) does not. Libraries also pay a fee to subscribe to Worldcat. This professional version I have found occasionally returns additional results not available in the free version.
Step 5: Search the newspapers.
Proquest and Newspaper Archive are two of several companies that offer keyword searches of newspapers. Libraries tend to subscribe to a few major newspapers and many newspapers in their city or state. Thus ask the reference librarian in the state where the event occurred or the building is located to search their Proquest for a citation may work. The New York Times offers a search on their website (http://www.nytimes.com/ ) and offers recent and non-copyrighted works for free and others for a fee. (However, a Proquest search I did of the New York Times found more articles than their website search engine.)
Other online newspapers (http://en.wikipedia.org/wiki/Wikipedia:List_of_online_newspaper_archives ) are available for searching. I have found that the servers for some of them are not very robust. A request may return an error message. But repeating the request a few minutes later return results.
Step 6: Asking others for help.
Mention has already been made of emailing reference librarian for help and locating groups studying individual railroads. You can also try contacting museums near the locations you are researching. Use Google Maps by either selecting the Map menu from Google http://www.google.com/ or using http://maps.google.com directly. For example, for information on the driving of the Golden Spike at Promontory, Utah enter “Historic near Promontory, Utah”, “Museum near Promontory, Utah” and “Museums near Promontory, Utah” “Historic” will capture Historical Societies that may not have a museum, “Museum” appears to capture museums with “museum” in the title whereas “Museums” are more general. Local libraries may have special collections and searching for “Libraries near Promontory, Utah” would find them. These searches will also locate institutions in nearby larger cities or county and state collections. Searching “Libraries near Philadelphia, PA” or “Libraries near Eddystone, PA” or the corresponding “Historic” searches will likely find special collections related to the Baldwin Locomotive Works. Search Wikipedia for the towns of the companies.
The Linda Hall Library http://www.lindahall.org/ has the consolidation of several historic science and technology libraries http://www.lindahall.org/collections/ . Along with their online catalog, they offer a free 20-minute search and a fee-based search for more time-consuming searches. Check the websites of other libraries for research services.
Fee-searches will likely have a rapid turn-around whereas free requests may take a week or more at institutions without a formal search service. In the later case, the request may be handed off to a one-day-a-week volunteer. Bearing this in mind be as specific as you can be in your request. This inspires confidence in the researcher that they can successfully answer your question—yes-we-have-it or no-we-don’t-but-try-here. If you saw a reference to a specific article, provide a source so they can check it themselves. I have told researchers the specific keywords I used in Google; provided the URL of an online paper that referenced it; or the book information including page number that listed it in its bibliography.
Step 7: Does the National Archive have the answer?
Check the Frequently Asked Questions in Part II (http://www.rlhs.org/narip2.htm ) of the Records Relating to North American Railroads. Compiled by David Pfeiffer. Reference Information Paper 91. National Archives and Records Administration. Washington, DC (http://www.rlhs.org/research.htm ). Much of this information is from the Interstate Commerce Commission Valuation Documents which surveyed the line of major railroads in the 1920s. The valuation documents are described in Archivist’s Perspective, “Riding the Rails Up Paper Mountain: Researching Railroad Records in the National Archives,” by David A. Pfeiffer, Prologue Magazine, Spring 1997, Vol. 29, No. 1 Part I (http://www.archives.gov/publications/prologue/1997/spring/railroad-records-1.html ) and Parts II and III (which are linked from Part I).
Unless you live locally and can afford to spend the time, you can hire a private researcher. See http://www.archives.gov/research/hire-help/ for a list of them. I have personally used Do You Graphics and Nicom, Inc. (http://do-you.com/ and http://www.nicom.com/ ).
Suggestions for keyword searches
The searching methods for www.google.com and http://www.books.google.com appear to work differently. The former ignores punctuation within double quotes (except in special cases such as C++ and C# which are programming languages) while the latter does not.
When possible, don’t include Railroad or Railway in a company name. Searching for “Union Pacific” will find “Union Pacific Railroad”, “Union Pacific Railway”, “Union Pacific RR”, “Union Pacific Ry”, “Union Pacific R. R.” etc. However, sometimes the company name must be qualified. A search for “Indiana” would bury any “Indiana Railway” results. Make separate searches on the “&” form and the “and” form. Consider alternate spellings and abbreviations as well. Chicago & Northwestern was at times a Railway, a System, a Line, and a Transportation Company. Leaving off these qualifiers still require four searches: “Chicago & Northwestern”, “Chicago & North Western”, “Chicago and Northwestern”, and “Chicago and North Western”. Unless searching within a specific railroad journal, a search for “Northwestern” would likely have the non-railroad hits swamp out the railroad ones. Note that these searches still do not cover all the possible cases. Especially in brief items, the railroad might be referred to as “CNW”, “C and NW”, “C & NW”, “C&NW”, “C. & N.W.”, etc.
Use the Wikipedia articles to get the age-appropriate name of the railroad. Contemporary articles will not mention the “Chicago, Milwaukee, St. Paul and Pacific” Railroad before about 1928. References to the “Chicago, Milwaukee and St. Paul” may occur after that date, especially in the popular press. Its nickname, “Milwaukee Road” should be searched as well as variations with the “&”
Both “station” and “depot” should be used in searches. By using the advance search option both words can be entered in the “With at least one of the words” box. Both words have non-railroad connotations as well such as “supply depot” or “Home Depot” and “gas station.”
[This article reflects the experience and ideas of its writer; Lake Forest College, its library, and its personnel are not responsible for the efficacy of referrals and/or suggestions and examples included here.]