Purpose of recording
I record serial numbers as I record prices, issuance and cancellation conditions, and contributors. allows the ability to:
- determine relative availability of varieties
- estimate the frequency of appearances of varieties
Serial number order
As mentioned elsewhere on this website, I have never discovered any inviolable rule concerning stocks and bonds. That is why I use words like "generally" and "usually" so often. For instance, companies generally numbered their stock certificates in ascending order starting at #1. In doing so, serial numbers within any given series should increase as dates of issuance progress.
If companies issued multiple types of stock certificates and multiple denominations, then there might be several streams of serial numbers ticking ever-upward, each increasing in date order. For instance, a company might have issued two types of stock during the same time period: common and preferred. If each type were represented by <100-share, 100-share, and odd-share denominations, then each denomination would have been numbered separately. Such a company might have had six entirely separate sets of serial numbers, all starting at #1 and ascending at different rates.
Bonds were numbered similarly. Each series of bonds generally had their own sets of serial numbers starting at #1 and ascending to the maximum number of bonds authorized. Unlike stock certificates, most coupon bonds were numbered from #1 upward through all denominations. For instance, a company might have issued:
- two hundred $1,000 bonds numbered from #1 to #200
- one hundred $500 bonds numbered from #201 to #300
- one hundred $100 bonds numbered from #301 to #400.
Larger companies often issued several series of bonds consisting of both coupon bonds and registered bonds. In those cases, coupon bonds were usually issued in a single $1,000 denomination starting at #1. Serial numbering among the various denominations of registered bonds usually started at #1, with unique letter prefixes for each denomination. For instance, $1,000 bonds would have been numbered starting at #M1, $5,000s at #V1, $10,000s at #X1, etc. Registered bonds often had a series of "odd" denomination bonds, issuable in units of $1,000 and numbered with some other letter prefix
More elaborate prefixes were necessary among the very largest companies that sold hundreds of thousands of stock certificates and hundreds of thousands of bonds. Serial number prefixes sometimes correlated with place of issuance such as "P" for Philadelphia, "N" for New York and "C" for Chicago. And of course, there are several cases where serial number prefixes remain entirely mysterious.
Odd serial number order
Not surprisingly, there are examples of serial numbering whose logic is baffling.
It appears that some companies, especially early companies, recycled serial numbers. For instance, stock certificate #23 might have been sold in 1840 and another #23 was sold out of the same stock certificate series in 1844. In many cases, it appears some companies issued new certificates with old serial numbers whenever shares were sold and stock certificate were transferred to new investors.
I have seen some series of stock certificates where number-only serial numbering appeared to ascend in ordinary date order, but with occasional prefix or suffix letters thrown in. For instance, I might have records of certificates arrayed in date order something like #4, #7, #8, #10, #A3, #14, #A9. What did the "A"s mean? Were they recycled numbers? Were they issued in different places? Were they special in some way?
There are also cases where serial numbers are unique but issued out of date order. In other words, records might show that #3, #5, and #9 were issued in May, 1850 and #4 in August, 1850. Why was #4 issued after #9? Perhaps number #4 had been sold but not paid for by some important friend of the company vacationing on Long Island. Or maybe it was simply a recycled serial number that had been issued and transferred. Matters would probably be clearer if all issued serial numbers were known.
Serial numbers that span denominations
Companies that used multiple pre-printed denominations on their stock certificates normally had engravers print lagre numbers of <100-share and 100-share denominations and limited numbers of intermediate denominations such as 1-share, 5-share, 10-share, 25-share, and perhaps 50-share certificates. In general, each denomination had its own set of pre-printed serial numbers, usually with unique prefixes. If prefixes were not used, then each denomination normally had its own unique range of serial numbers.
- 5-share from 2000 to 3999
- 10-share from 10000 to 14000, etc.
The Lehigh Valley Railroad Co. is unusual in that it used a series of certificates printed with nineteen different denominations in three colors, all with the same general design. Each denomination had one of three colored borders, but each had sparate custom-engraved borders that contained small engraved denominations in different repeating patterns. This series of Lehigh Valley Railroad certificates is made even more unusual in that serial numbers were not pre-printed.
Instead, the company printed serial numbers as it issued certificates. If all certificates of this series existed, and were known, serial numbers would form a continuous range of about 73,000 numbers that would parallel dates but span denominations. Only two other companies are known that share this type of numbering: the West Jersey & Seashore Railroad Co and the Pennsylvania Rail Road Co. Speculation about the West Jersey and Seashore is unconfirmed because there are not a sufficient number of certificate images available for analysis.
Analysis positively supports the notion that the Pennsylvania Railroad did the same thing as the Lehigh Valley, except that it issued many more than a mere nineteen variations.
For three decades, I had assumed American Bank Note Company had printed a series of certificates (shown at right) with pre-printed serial numbers as normal. You will notice a green "10" is visible in the share medalion at the top right along with a bold "TEN" printed on the front and a "NET" printed in reverse on the back. I had further assumed that the railroad had overprinted the "10, "TEN," and "NET" overprinted denominations at the time of issuance. (Image courtesy of Robert Warchocki.)
In October, 2021, I was recording offerings and sale prices from auction 120 of the German auction house Freunde Historischer Wertpapiere (FHW) when I noticed it offered 33 lots of these certificates, all separately. It had offered the series as if each of twenty denominations was different.While only four lots had sold, prices were abnormally strong compared to U.S. prices ($116 to $173 US). I found that curious.
I did not own any those certificates at the time, so examined high-resolution images sent by one of my long-time contributors. I contacted him and asked him to re-examine his certificates. He confirmed that the green denominations had been printed BEFORE the black text, vignette and border. Not only that, but the bold verbal values ("TEN" and the reversed "NET" on the back were printed in perfect registration, front and back. That would have been impossible had ABNco used separate plates. I don't know ABN's process, but I do know the front and back had to have been printed at the same time. There was absolutely no way the values had been overprinted.
Since then, I have acquired several certificates and can absolutely confirm the black plate was printed after the green denominations. Furthermore, my records of serial number clearly confirm that serial numbers parallel dates and cross denominations.
This series of Pennsylvania Rail Road certificates can be found with three pre-printed date variations: '18--', '190-' and '19--.' That means there are at least 300 (!) possible sub-varieties of this same design. I do not expect anyone to try to collect all 300, so I have not create separate variety numbers. Moreover, these certificates are NOT at all common.
But wait...that's not all!
A later series of Pennsylvania Railroad stocks certificates (at right) were printed and issued in the same manner with serial numbers crossing denominations. I solicit help in fleshing out this series.
Low serial numbers
For the last forty years, experienced collectors have sought out certificates with serial #1 and have routinely paid premiums of 25% to 75% for their efforts.
Amateur sellers on eBay frequently promote "low" serial numbers (often in the hundreds) hoping to bump up the price for dreadfully common certificates. I hate to break it to them, but they are wasting their time.
I have occasionally recorded smaller premiums paid for serials #2 and #3, but not for higher. I have encountered a couple collectors who wanted specific serial numbers, but if they were paying premiums, their amounts were insignificant when compared to the highly flexible range of prices paid for certificates in the various markets.
As mentioned before, every issue of bearer bonds had (at least to my knowledge) only one number #1 certificate. If a company sold registered bonds, then there were usually several #1s with different serial prefixes. Unfortunately, I know some collectors are looking for #1s among varieties that never had any #1s. Take for instance the nineteen Lehigh Railroad varieties of stock certificates mentioned above. The lowest serial number ever issued for ANY denomination in that series was probably around #85,000 because numbering had continued unbroken from earlier certificate designs. Even if companies issued several different variations of stock certificates, there often was only one #1 stock certificate per company, or per certificate series.
Accuracy of recorded numbers
The online database is available to all users and shows every serial number recorded up to the time of the last update, usually no more than 15 days before. Numbers are as accurate as possible, but mistakes are inevitable:
- handwritten serial numbers can be terribly hard to read
- handwritten serial numbers can be mis-interpreted
- many serial numbers have been reported by collectors and dealers without images for proof
- some sources are notorious for reporting serial numbers and dates that cannot possibly fit into existing and provable series of serial numbers, so both dates and serial numbers become unreliable
- many sellers use stock images for illustrations and therefore render reported serial numbers unprovable
I BEG collectors to contribute send scans or photographs of certificates to help fill out the database further, no matter how common their certificates might be.