Catalog numbers look like this:

NEW-530-S-50

This system indicates two kinds of information:

Read all about Numbering Companies

Read all about Numbering Certificate Varieties

Ideally, company names SHOULD be in alphabetical order.

Unfortunately, that is not always the case. Company numbering reflects the way the project started, and how new railroad companies are discovered.

How the project started

Back in the early 1990s, I sold collectible paper money as well as railroad stocks and bonds. Pricing in the paper money hobby was well-understood, but the pricing of certificates was hard to decipher.

As I researched prices, I discovered that prices varied wildly among dealers.

In order to better understand pricing, I began recording the prices of every railroad certificate I found in auction catalogs and price lists. At that time, there was no such thing as the internet or eBay and very few dealer price lists were illustrated.

Within two years, my certificate price database had grown to over two thousand varieties. Soon, I began sharing my information with collectors. The more information I shared, the more information collectors gave me. Since then, my little pricing project snowballed into the current database and books.

Devising a numbering system

Once I decided to make a catalog, I knew I needed to arrange information so collectors could find things quickly. Obviously I needed to arrange certificates by company name. The first major question, though, was whether to identify certificates by full company names, partial company names, or abbreviated company names.

Designing a catalog around full company names would have been impossibly cumbersome. And while a catalog could have been designed around railroad abbreviations (B&O, for example), I saw there would be hundreds of duplications.

I needed a more forgiving numbering system.

The paper money hobby has two numbering systems. Both the "Hessler" and "Friedberg" systems use an arbitrary, but fixed, numeric system. That makes sense because few new discoveries ever appear in that hobby. The paper money numbering system can be static.

In U.S. stamp hobby, collectors have depended on Scott's Standard Catalog and its alphanumeric system for decades. While varieties are occasionally added and subtracted, the Scott system is essentially fixed. By themselves, designations are uninformative, but once collectors get used to the Scott catalog, the system works very well.

I reasoned that if I started with a 100% complete list of railroad companies from North America, I could have designed an alphanumeric nomenclature that would make some sense.

There were two major obstacles to that approach

First, there was no complete list of railroads that had operated in North America. Secondly, there was no list of companies that issued stocks and bonds, but never laid rail. With no acceptable master list, I realized my system had to be immensely flexible in order to accommodate hundreds, maybe thousands, of new companies that would surely appear.

The difficulty of planning for growth

Designing a catalog numbering system must plan for growth.

With 52 letters in the English alphabet, I knew I had the possibility of referring to 140,608 different companies by using only three letters.

However, if I combined three letters with three numbers, I had over 140 million unique combinations.

Rightly or wrongly, I decided to use the first three letters from every company as my starting point. That way, as long as collectors knew a company name, they could quickly flip to the right part of the catalog to start searching for more information.

Starting with those letters, I added three numbers in a manner so a computer could sort all companies in alphabetical order.

By assigning numbers manually, I was able to force my computer to sort companies in a customized manner. I needed to control sorting so I could overcome peculiarities that would arise with strict alphanumeric sorting. For instance, I needed to make sure that companies with person's names sorted on the last name, not the first. I also needed the computer to ignore 'The' in official company names.

Even with this system, I knew numbering would still be problematic for companies with common names

In the numbering scheme, the first three letters come from company names, but the three numbers are arbitrary.

That is perfectly acceptable for companies whose names are unusual.

For instance, even today, I only know of three companies known whose names start with "YAZ." I could number those three companies in any manner I wanted, and collectors could still find them.

Numbering companies with very common names, however, gets very tricky.

For instance, there are almost 600 companies whose names start with "New." That is a very crowded group. Even in the beginning, numbering such a group was hard and I knew it would only get harder with time.

So, the first step was finding the names of all railroad companies

Before I created a single catalog number, I created a list of every railroad and railroad-related company I could find in North America.

I started with two sources of railroad names. William Edson's Railroad Names listed 6,932 companies. Joseph Gross' Railroads of North America listed 9,840 companies. As you can imagine, there was substantial overlap between those two books, so my beginning master list included 11,128 names.

As with every other endeavor of this magnitude, I knew that Gross and Edson had missed many companies simply because of the nature of their individual projects. For instance, Edson reported only those companies that had reported to the ICC.

Next, I searched every major reference to railroading that I could find. By about 1993, my catalog numbering system included 17,276 companies.

The next step was numbering companies

I tried to number companies in a manner that would leave as much room as possible for new discoveries. For instance, if there were twenty companies whose names began with the same three letters, I spaced them about 50 numbers apart. (KEN-050, KEN-100, KEN-150, etc.)

I encountered a dramatic problem when numbering railroads accessed Americas' two largest rail centers: Chicago and New York. By the time I started numbering companies, I already knew of several hundred companies whose names started with "CHI" and "NEW." In the case of companies beginning with "NEW", I initially numbered companies only three to four numbers apart. I knew that numbering problems could arise with the addition of as few as three new companies if those new companies appeared in the wrong places.

I also had to decide how to number companies whose names were only partially known

Many references, including Edson's monumental work, shortened titles by chopping off words like "Railroad" and "Railway". Even "Street Railway" or "Inter Urban Railroad" do not appear in all references.

Unknowingly, railroad references created an even worse problem by spelling company names very imprecisely. Think of the sorting and numbering problems that might arise among similarly-named, but potentially misspelled, companies whose names started with "West Chester" or "Westchester". I found a remarkable number of authors who were completely unconcerned with the differences between names such as "Ogdensburg" and "Ogdensburgh", "Painsville" and "Painesville."

Because of my sources, I could not initially concern myself with the differences between "Railway Co" and "Railroad Co"

It is amazing how many printed references use words like "Railway" and "Railroad" interchangeably. Seemingly, the authors never realized or cared that those words may have identified separate incorporations of the same company, or even completely different companies. Now that the World Wide Web exists, even more problems are being created by sloppy research.

When I started, I did not record "RR" or "Ry" in my database. But, when certificates finally appeared and spellings improved, odd numbering sequences ultimately appeared in the database:

In 1999 and 2000, I made a major, year-long effort to add qualifier words like railway, rail way, rail-way, railroad, rail road, and rail-road. Even today, many companies in the database still lack those words because I have not yet found reliable sources.

But the damage to the numbering system was already done

Even if it is understandable, the vestiges of my initial oversight remain. Fortunately, electronic database searches do not care. The database will search for whatever names you enter, and it will find them however they may be numbered.

It is a little harder in the printed version. In general, if you do not find a company listed in its strict alphabetical order, look higher in the list.

Punctuation

Punctuation in company names is more variable and unpredictable than spelling. Except for hyphenation, I do NOT include any punctuation in my database. As best I can determine, no state records commas or quotation marks in official company names. Some states record apostrophes, but the use is dramatically inconsistent.

The rules about "Saint" and "Mount" and other abbreviations

Every language has its own rules for abbreviations. Every language assumes that you already know what its abbreviations mean.

For example, if you have a certificate from the St. John Railway, you are supposed to know that "St." is an abbreviation for "Saint." Never mind that the word "Saint" is almost never spelled out, even in the official names of cities and towns.

Yes, many non-English speakers understand that "St" = "Saint." But, how many collectors know Spanish abbreviations such as "Cia" = Compañia or "FC" = Ferrocarril?

So I made a rule.

I only abbreviate ONE word for catalog numbering purposes: "ST" = Saint

Why the exception for "St?"

Space. Had I not allowed the abbreviation for "Saint," there would have been almost 800 companies listed under the "SAI-" category!

What about "MT" = Mount?

While "Mount" is often abbreviated, "Mount" tends to be spelled out more often than "Saint." In order to be consistent, and make companies with similar names sort near each other in print, I made a rule. I ALWAYS catalog companies whose names start with "Mount" or "Mt" as "MOU."

Even my flexible numbering system was not as flexible as I had hoped

Three developments forced minor alterations in my system.

I resisted as long as I could, but the tremendous growth of the database forced me to modify my original scheme.

Development one: raw growth in the number of companies known

Since the first edition of my catalog, I have added 9,000 more companies!

Currently, I list almost 600 companies whose names start with the letters "NEW." With only 999 possible numbers for companies in the "NEW" category, it was inevitable that the numbering system would fail.

When that happened, I added a decimal extension to the general numbering scheme. That gave me room to squeeze in new companies without changing existing numbering.

Here is an example of how I handled the crowding problem:

Development two: companies with the same general name.

Today, most US states prohibit corporations from using overly similar names. However, because they operated at different times in different states, many similar names have emerged. Take these companies for instance.

The discovery of so many name variants has further enlarged the database. Like the example above, I try to squeeze name variants in as best I can, but sometimes I must resort to decimal numbering.

In case you're worried, there are only 156 decimalized catalog numbers among the companies currently known.

Development three: companies with identical, or nearly identical, names

Unlike the previous two examples, there are numerous companies that were re-incorporated or re-named for countless reasons.

Take the Southern Pacific Railroad for instance. Depending on how you count, there are fourteen or fifteen companies with that name. A Texas company chartered in 1852 seems to have been the first. The more famous California giant did not use the Southern Pacific Railroad Co name until 1865. However, as it grew, it reincorporated under that same name in 1870, 1873, 1874, 1888, and 1898. Throw in separate incorporations in other states and references to "Southern Pacific" get very confusing.

A different sort of example occurs with the Baltimore & Ohio. I do not currently know of multiple incorporations of that company, but its certificates plainly show four major spelling variations. To further confuse modern collectors, certificates clearly show that name usages overlapped through time.

There are many other examples of less famous railroads whose certificates show minor spelling differences for no obvious reasons. There are several companies whose bonds spell company names one way and stock certificates spell another. Believe it or not, there are many examples where company names are spelled two different ways on the same certificate!

So, I decided to handle things a little differently in these situations. If company names are identical, or nearly identical, I label them as a subgroup with lower case letters. For instance, here is how I handle the B&O. Because of its divisional offerings, the BAL-662 category is inflated more than most other companies with known certificates.

Do I ever move numbers around?

Once certificates are described and numbered, I try as hard as possible NOT to change numbers.

Unfortunately, I had to renumber Mexican railroads whose names started with "Ferrocarril."

On the other hand, if certificates are not yet known, then I will occasionally re-number companies to give the database a little more room in key areas. Even though that numbering is not critical, I only change company numbers in extreme circumstances.

Back to TOP

How catalog numbers are assigned to certificates

Designing numbering systems for certificates is a balance between ideals, philosophy, and reality. Reasonable people will always disagree about the first two. Many people simply try to deny the third.

The ideals

• I would like catalog numbers would to be roughly correlative with dates.
• I would like the numbering to be easy to decipher.
• Varieties should be easy for beginners to decipher.

The philosophy

I want to avoid flyspeck scripophily. I do NOT want to number every tiny, microscopic variation. I want varieties to be easy to identify by practically every collector.

The reality

Certificates are discovered and cataloged randomly.

Varieties defined

In general, I define varieties by what was initially PRINTED on certificates, not by what was added later.

Certificates were created for thousands and thousands of different companies. Therefore, I cannot identify a single hard and fast rule about any aspect of this hobby. Consequently, I needed two exceptions to my rule about varieties.

Exception 1: Handwritten certificates and generic certificates with handwritten, typewritten, or rubber-stamped titles are considered legitimate varieties.

Exception 2: Handwritten, typewritten, or rubber-stamped name changes are considered as legitimate varieties.

The logic behind varieties

I think we can adopt a few critical observations from other hobbies:

• pricing follows popularity
• popularity is defined by the number of average collectors attracted to a hobby
• as prices rise, smaller and smaller distinctions become important

Currently, our hobby is very new. The hobby is so new, in fact, that all collectors can still express opinions about what they think constitutes important varietal differences – and experts will still listen!

The most important representations of opinions are prices. Prices are opinions that all dealers and auctioneers listen to. Prices make it perfectly clear that average collectors do NOT care about small, flyspeck varietal differences.

I fully recognize that at some time, collectors will look back on this period and wonder why we didn't notice or record more subtle varieties. At this time, I simply cannot ignore the fact that collectors will not pay for minor varietal differences when they have so many things to collect.

Features that create separate varieties

The whole idea of defining varieties is to help collectors identify certificates that might (now or in the future) have different rarities, and consequently different collectible values. While there are always exceptions to the rule, these features probably account for 99+% of the varieties identified so far.

• discrete company names
• discrete vignettes
• discrete denominations
• discrete colors
• discrete types of certificates (preferred vs common)
• discrete capitalization amounts
• discrete par values
• discrete loan types (1st mortgage vs 2nd)
• discrete loan percentages
• discrete series (series A vs series B)
• discrete printed dates
• discrete revenue imprints
• discrete engraving/printing companies

Ideally...

I want collectors to be able to identify separate varieties in average quality scans or photographs. The harder the differences are to see, the closer we are approaching "flyspeck scripophily."

Features that do NOT warrant separate variety numbers
or
LUMPERS vs SPLITTERS

Collectors are like biologists. There are lumpers and there are splitters.

At one extreme, lumpers consider micro variations tiresome. They lump sub-varieties into one classification. They don't even see – let alone care – about minor differences.

Splitters take the opposite view. They consider every tiny difference as a separate variety.

I try to take the middle ground between the extremes.

While there may be similar numbers of collectors at the extremes, splitters tend to contact me more often. I am affected by everyone's input. Splitters try to convince me to make more and more varieties. I give in more times than I care to admit and I tend to create more varieties than I would normally prefer.

On the other hand, I desperately try to avoid flyspeck scripophily.

Generally, I do not consider anything added after printing as a separate variety. Consequently, I do NOT separate certificates into separate varieties based upon:

• slight color differences
• issuance (issued vs unissued)
• cancellation (cancelled vs uncancelled)
• handwritten dates
• seal colors
• seal types (embossed vs. applied)
• serial numbers (colors, series, or styles)
• stock owner's names or autographs
• bond holders' names of autographs
• officers' names (even if printed)
• adhesive revenue stamp types
• silver overprinted name changes (silvering)
• trust company names

Currently...

If collectors are currently willing to pay more for one variation than another, than separating into varieties is definitely warranted.

And then, looking into the future...

If collectors might be willing to pay more for one variation in the future, than splitting into separate varieties might be warranted.

But...

If variations are so minor as to be completely meaningless to the vast majority of collectors, then separation is unwarranted. I try to keep notes in the database just in case whoever follows me has different opinions.

Certificate types

In my cataloging system, I assign every certificate to one of these certificate types:

• S = stock certificate
• B = bond certificate
• O = other certificate related to stocks or bonds

As I have already said, hard and fast rules do not exist. Every rule has an exception. There are always certificates that stretch the boundaries of one definition or another.

For instance, there are hundreds of certificates that look like stock certificates, but that specifically say that they are receipts for partial payment on stock. In many respects these certificates represented marketing efforts by weak companies. Often, it was impossible for companies to sell $100 shares at full price, so they offered shares for 5% or 10% down. Once an adequate number of shares were sold, companies then "assessed" holders for more money. Often, those certificates say they were "subject to assessment." Because such certificates functioned like stock, I catalog them as stock certificates.

Another major certificate problem involves "equipment trust certificates." These are certificates that look like bonds and even include coupons. Text on such certificates, though, usually says that the certificates represent one share of stock in equipment trusts at $1000/share. Instead of indicating payment of interest, equipment trust certificates refer to dividend payments. However, because these certificates functioned very much like bonds, and certainly look like bonds, I catalog them as bonds. If someone wants to split hairs, they actually functioned in a gray area between true stocks and true bonds.

Variety numbers

Going back to my "ideal," I prefer to number varieties in a datewise manner. Unfortunately, certificates are discovered randomly. A "clean", uniform number sequence is never possible among companies with large numbers of certificates.

Among, bond certificates, low variety numbers are assigned to earlier bonds. Sometimes, though, issue dates are not known at the time new varieties are described, so disruption in logical sequences are common.

Stock certificates normally fall into one of three main types

• capital stock
• common stock
• preferred stock

Capital and common stock are nearly synonymous terms. If companies issued only one class of stock, they usually labeled their certificates capital stock. However, if companies wanted two classes of stock, they usually issued common and preferred shares. (The main difference was that preferred stock received preferential treatment in receiving dividends.)

I never know when new stock certificates will appear, nor what types of shares they will represent. Preferred shares are less common than common or capital shares, so I usually reserve numbers above "S-65" for preferred stock.

As a rule, I only use variety numbers below "15" for transitional certificates. Transitional certificates are normally stock certificates that carry names of predecessor companies which were modified in some manner to new company names.

I rarely use variety numbers below "10". I reserve those variety numbers for handwritten certificates (stocks, bonds, or others), which by their very nature, were issued before companies received printed certificates.

Back to TOP