Using the Fulton History Newspaper Site

Cliff Lamere     Feb-Mar 2014, last revised June 12, 2020

----------------------------------------------------

http://fultonhistory.com/Fulton.html

The Fulton History Newspaper Site is a truly amazing accomplishment for one man working alone.  Tom Tryniski, a retired engineer, has been working at it for 18 years, about 10 hours a day, seven days a week.  As of October 24, 2017, he had managed to put 40,743,000 newspaper pages on the internet.  By comparison, The Library of Congress has only 12,289,671 pages on their Chronicling America site for which they were awarded a $22 million federal grant years ago.

[On June 12, he had 48,156,000 newspaper pages online.] 

Tom Tryniski at his work (video)

Haper's Magazine published an article about Tom Tryniski on September 13, 2017. (link)

_______________________

Optical Character Recognition (OCR) software looks at the text in an image and converts it to digital text which can be searched, copied and edited.  On the Fulton History site, the OCR software must work with newspaper images which have been copied from microfilms.  Some of the newspapers are very old, and the text can be a bit difficult to read.  In addition, the Fulton History site uses microfilms which have been previously used and therefore can have scratches, smudges, and dust on them.  All of these can result in the OCR software misreading what the human eye sees easily.

The results from using the search engine on the Fulton History site are very different from the results you get when doing a Google search, because Google does not attempt to search the text that was on images.  Getting the best results from searches on the Fulton History site will require some considerable adjustments to the way you do searches.

Important Note:  This webpage describes the various kinds of searches that can be made on the Fulton History site.  You need to understand this information, but the related webpage about boolean searches and wildcard searches will show you how to make advanced, more effective searches when the searches on this webpage fail.

Exact Searches

With over 40 million newspaper pages available on the Fulton History website, searches often find hundreds or thousands of hits.  Searches using "the exact phrase" will make the hits fewer and thereby more manageable.

Searching for 'all of the words'

The meaning of this search is obvious, but there is a refinement of this search that should be mentioned here.  Let's say you want to search for:

     Joseph E. Brady Albany died

Using the option "all of the words", you would find every page that contained the name Joseph and the name Brady and the word died, etc. anywhere on the page.  What you really want is to find Joseph E. Brady on a page that also mentions the words Albany and died.  Here is how you do that:

     "Joseph E. Brady" Albany died

Searching with quotation marks will require that everything within the quotation marks is treated as just one word and therefore must be in every hit in exactly the right order.

Searching For A Person

When searching for the name of a person such as John Arthur "Jack" Smith, several exact searches can be made.  (Boolean searches and wildcard searches will give you the same or better results in fewer searches, thus saving you time.)

   John Arthur Smith

   John A. Smith

   John Smith

   Jack Arthur Smith

   Jack A. Smith

   Jack Smith

Unfortunately, many people who were born with a first name that they didn't like decided instead to use their middle name as a first name.  For example, John A. Smith might become Arthur J. Smith.  Keep this in mind if you are not certain.  My feeling is that official records like birth certificates and draft registrations should have the person's correct birth name.  A young child's name on a census (ages 0-4 or 5) also tend to be correct unless they are obvious nicknames.

A person giving information to a reporter for a news story may not have had accurate knowledge of the spelling of the name of the person named in the article.  Therefore, you may want to also use alternate spellings of a surname.  (The ? wildcard search mentioned on another webpage may simplify this search.)

Adjusting For OCR Errors

In newspapers, the small letter 'i' and the small letter 'l' are often difficult for an OCR program to tell apart.  Gardenier is the main spelling of the surname that most interests me.  Let's say that I are searching for the spelling variation Gardinier and wanted to find all of the times it was mentioned in a newspaper.  On this site, there are 5000 hits, the maximum the search engine will find (there are probably many more than that).

3183 hits occurred when I replaced the first i in Gardinier with an l (the lowercase L).

4139 hits occurred when I replaced the second i in Gardinier with an l.

5000 hits (the maximum that will be shown) occurred when I replaced both i's with an l.

In the three searches using altered spellings, over 12,000 hits resulted, showing that the substitution of i for l and l for i is an important search strategy.  Of course, many of the 12,000 articles will also contain the proper spelling of the surname, so searching for the normal spelling of the surname alone would not have missed those articles.  However, a significant number of the articles which contained the corrupted spelling would mention the surname only once.  That means that a large number of articles would have been missed if you had not searched for the corrupted spellings.  That is especially important when you are making an exact search for a person's first and last name, or first name/middle initial/last name.  Those combinations in an article may be available to the search engine only with the corrupted spelling.

Unfortunately, many letters can get changed during the OCR conversion to digital text, including those in first names.  How many variations of the search you try will depend on the importance of the information you seek.  The following is a list of some letter changes that I found on one day (plus rn suggested by Charles Dobie).  Some are more common than others.

   Fairly common

c becomes e  (Saranac became Saranae)

c becomes o  (Lucy became Luoy)

e becomes c

e becomes o  (the became tho)

G becomes O  (Gardinier became Oardinier, 1183 hits)

h becomes b  (his became bis)

i  becomes l

l  becomes i

o becomes c

rn becomes m  (Burnside became Bumside)

u becomes n

   Probably less common

a becomes e

a becomes u

e becomes a  (feet became faet, he became ha)

f becomes t

G becomes (i  (a parenthesis is not a search term, so G often is just an i.  Try ieorge for George.)

n becomes a

n becomes u

ni becomes m

s becomes a  (Parks became Parka)

s becomes e  (Syracuse became Syracuee)

S becomes 8

y becomes v  (Syracuse became Svracuse)

Sometimes The Name Cannot Possibly Be Found

Sometimes, the OCR text is spelled so badly that there is almost no chance that it could be found.  A "Lukins-Guardenier" wedding became "Laklas-Gnardenier" wedding.  You wouldn't have found either surname with a normal search, and the names were mentioned only once, so the wedding announcement would have been missed.  If you know the date of an event, the site will allow you to browse its newspaper pages by date.

Names Hyphenated By Newspapers

Searching for Gard-inier resulted in an additional 294 hits.  Searching for Gardin-ier resulted in another 229 hits.  If the name was only mentioned once in the article, you would have missed it if you didn't search for the hyphenated version.

Newspaper Image as a PDF or Shown on the Fulton History Site

When the newspaper image appears, it may be presented by your computer as a separate PDF page, or it may appear as part of the Fulton History site, in the part to the right of the list of hits.  I prefer the PDF pages because they occupy the entire screen.  There are Preferences or Options in your Internet browser which control what happens, although the right option is difficult to find, and that location will probably change if you upgrade to a new version of the browser.  There are so many browsers and versions available that it is not possible to give helpful instructions here.

Searches Done On The Newspaper Page After It Appears

Although Edit/Find can be used, it is quicker to use Ctrl-f to cause the search box to appear.  If the newspaper page appears as part of the Fulton History site, you may have to click on the newspaper image before trying to bring up the search box.  Otherwise, you may find yourself searching only through the hits on the left side of the page.

While searching a newspaper page for the article of interest, I usually type the surname in the search box. Immediately, I repeat the search as many times as necessary to get an error message that there are no more instances of that word on the page.  Sometimes, there will be both an article and a photograph of the person on the same page. You certainly don't want to miss the photograph, or a possible second article mentioning that person or a possible relative with the same surname.  

Finding the Name and Date of a Newspaper 

When an obituary or death notice says that a person died yesterday or Friday, you need to know the date of the newspaper.  Some newspaper pages do not have the name and date of the newspaper at the top of the image.  Good genealogical research requires that you record this information. To get it, you should do the following.

Copy the link on which you clicked in order to get the newspaper page to show.  Paste it into the search box at the top of the page.  There will be a four digit number just before the ending of .pdf.  You must change that number by reducing it by one, then make sure that you will be searching for "the exact phrase".  The new filename will appear as a link.  Click on it, and an image of the previous newspaper page will appear as a PDF file.  Check the top of that page for the desired information. If it is missing on this page, repeat the process, going backward one page at a time, until you find a page which contains the name of the newspaper and the date. If all else fails, eventually you will arrive at page 1. That always has the information that you are seeking.

In case you think it would be advisable to increase the number by one at the end of the .pdf filename rather than to decrease the number as I have suggested, increasing the number will take you toward the end of the newspaper, and then to the beginning of the next newspaper. Because these are old newspapers that may have survived many decades or perhaps even a century, you cannot assume that every issue still exists and was microfilmed. Therefore, while looking at the next newspaper, you have no way of knowing how many days, weeks, or months later that issue is.  It will give no definite clue to the date you are seeking.     

Copying Text From a Newspaper Page

Copying text is normally quite easy from the average Internet page, but often it is a very challenging process on the Fulton history site. Each page will consist of about six columns of text. Using the mouse cursor, you can highlight text in an article and try to copy it. However, what happens too often is that as you are running the cursor down the lines in one column, you may notice that text in other columns is also being highlighted.  There are often vertical black lines between columns. When these lines have breaks in them, the highlight may flow into the next column.  However, there is frequently no break, and there is no obvious reason why the highlighting should go outside the desired column, but it does.

Let's say that text in three columns was highlighted.  If you use the copy function, it will copy the highlighted text in all three columns.  The text that you actually want may show up on every third line, meaning that you must delete the unwelcome lines.  Rather than do that, I usually retry the highlighting of the text.  As you are moving the cursor down the column, you can notice at what point the highlighting bleeds over into other columns.  Move the cursor up one line and copy and paste that amount of text.  Then, start at the next line and highlight as much as you can without overflowing into other columns.  Repeat as often as necessary.

At times, it may be quicker to simply type some of the information (perhaps only a troublesome line or two, but sometimes I type the entire article).  Since I use the Fulton History site pretty often in my research, I invested in Dragon Naturally Speaking which allows me to read aloud the words in the article and have the computer type them for me.  Even that process is not flawless, but the few corrections that must be made when using that program are not numerous enough to discourage me from using it.

    Example of a Copied Paragraph

The following text was copied from a column in a newspaper.  Notice that a hard return (caused by hitting the Enter key) occurs at the end of every line in the column. To save space, and also have the text presented in a format that is easier to read, I go to the end of each line, hit the Delete key (which brings the next line of text up), then I hit the spacebar once to separate two words that have been temporarily joined together.

When the case was called before

Justice Borst this morning, there was

a large attendance .of out-of-towii attorneys

and public officials. The state

immediately opposed proceedLng_j.yiii\.

the trial, but the objections.wer» overruled.

Some problems in this transcription can be explained. A speck on line three is shown here as a period. The 'n' was interpreted as 'ii'. The jumble of letters at the end of the fifth line is the result of a microfilm scratch that touched the bottom of some of those letters, although the final word 'with' was very clear to a human eye.  In the sixth line, the final 'e' in 'were' was not completely printed in the newspaper image.  The words 'attorneys' and 'overruled' (lines 3 and 6) were hyphenated words that were separated, part of each word being on two lines.  The OCR program sometimes shows the parts separately, at other times joins them together into one word (sometimes without deleting the hyphen).  If the hyphen in not deleted, that word will not be found in a search (unless you search with the hyphen as part of the word).

If you can't read something as it appears on the image of the newspaper page, sometimes the OCR version is readable. 

At the end of the fourth line above is the word state. It certainly didn't look like that on the newspaper page, not even close.  I never could've guessed it, yet when I highlighted, then copied and pasted the text, it was very readable.  So, when I can't read something that is important, I copy it, then paste it somewhere to see what the OCR program thinks it is.  This has worked best with numbers, because I have sometimes been able to confirm a year of birth or death already in my records (I never accept an unreadable number without confirmation).

Recording the Source of Your Information

It is important to record the sources of your genealogical information so that at a later date you can return to the source to check something that seems doubtful, or to possibly get additional data from that source.  Also, if you pass your data to another person, it is nice to include the sources so that the other person can double-check it.  Some data just doesn't seem reasonable or accurate no matter who gave it to you, or who was quoted as the source.  Dates, for example, can be mistyped.  If the date was interpreted by an OCR program, it could have been misread.

For Internet sources, I always record the URL (address) so that I can return later to that exact page.  On the Fulton History site, the address does not change as new searches are made.  So, I record the unchanging address followed by the full link to the PDF page where the information was extracted.  For example,

http://fultonhistory.com/Fulton.html

Amsterdam NY Daily Democrat and Recorder 1925 Sep-Dec Grayscale - 0728.pdf

   The Newspaper Name

After recording the information on a newspaper page, I go to the top of the page to extract the name and date of the newspaper issue.  The PDF link on this site contains the name of the newspaper according to the webmaster's standard format which is often not an accurate name.  I prefer to use the name as it actually appears in that newspaper, which is sometimes different.  However, the name on a particular page may be shortened to save space.  The full name of the newspaper may only appear on the front page, but that is much too difficult to reach quickly.  If the article is on page 6, you can subtract 5 from the four digit number at the end of the PDF link and hope to arrive at the front page (do an exact search, not 'all of the words'). 

Over a period of time, a newspaper's name may have changed more than once.  I use the name at the top of the page or front page that I have examined.  For large city newspapers, in parentheses I usually add the state abbreviation.  For newspaper names (like Herald Standard) which do not include the name of the locality in which the newspaper was published, I add the locality and state within parentheses.

A Perpetual Calendar Can Be Very Useful

When the day of the month cannot be read at the top of the newspaper page, access to a perpetual calendar can be very helpful.  If you can read that the newspaper was issued on a Saturday, you can look up the year and month on the perpetual calendar, then check the dates for all of the Saturdays that month.  If you are still not sure which Saturday it was, then you must find the full date on another page.

A perpetual calendar can be found here or here.

Some Cautions

The text you find on the Fulton History site will sometimes be part of a section devoted to a certain geographical area.  You should search for the title of the newspaper article (it might be in the previous column to the left).  Sometimes your text may be a part of an article with a title like "10 years ago today" or 20 years ago, or even 50 years in the past.  If that is the case, and you don't see that title, you will record the wrong year for the event that interests you.

If an article appears to end at the bottom of a column, you should look at the top of the next column on the right to make sure that the article has not been continued there.  It does sometimes, and you won't want to miss that part.

Advice

See the webpage about boolean searches and wildcard searches for a description of better, more effective searching.

----------------------------------------------------

The webpage you are on is part of a series of webpages about the Fulton History newspaper site.  It is written to help you make more productive searches.

Using the Fulton History Newspaper Site    (the webpage you are on)

Boolean Searches and Wildcards on the Fulton History Newspaper Site

A Wildcard Search That Works Poorly on the Fulton History Site

Visitors since 21 Sep 2015