Message Formatting - Word Counting


I’ve done some research recently, and I’d like to share the results.  It deals with a method of formal telegraph message preparation.  My limited research results are meant to be information for use in deciding how to more fully develop a telegraph impression in our units.  


Several pieces need to fit together when preparing an “official” message using the rules that seemed to be in place during the war.  I initially chose the Western Union Rules of 1866 as a standard. I think others may agree that these were most likely used during the war especially by the UMTS.  Turns out, I found out that the CSA operators were mainly American Telegraph folks, and they used very similar methods to format and record formal message traffic.  That would make sense because they had to talk to each other over the wire prior to the war, so the systems and formats could not have been too different. The operators were after all from the same fraternity. 


As I mentioned, I initially convinced myself that the WU Rules were the single source of instructions used during the war.  Upon closer examination, it looks like the WU Rules were written around what people were actually doing in the field prior to 1866.  When I examined several pictures of telegrams, I discovered at least two anomalies with respect to the rules that aren’t very large, but worth noting.  Both differences between the telegrams and the Rules involve word counting and recording.  The reference documents I used were: “The Western Union Telegraph Company Rules, Regulations, and Instructions” dated 1866, “The Telegraph Manual: a Complete History and Description… by Tal. P. Shaffner dated 1859, “The Electric Telegraph” by Prescott dated 1958, “Lee’s Despatches” edited by Douglas Southhall Freeman, and a web site containing original pictures of old telegrams (http://Members.aol.com/cd102/).  


In 1866, the Rules seemed to indicate that additional charges for messages forwarded by WU would be levied. Those charges were the amount necessary to cover the word count in the preamble.  In all the messages I reviewed, I could not see where that additional charge was applied between 1855 and 1865.  Also, in all the telegrams I reviewed where a word count was evident it was based on the message text only.  The other difference was that for military telegrams, no charge seemed to be levied for anything both for USA and CSA.  Moreover, in many cases I could not find a word count on any military telegrams I examined.  It may be that I was just looking at the original telegram, not the copy that was sent.  If it proves to be the case that word counting was not done for military telegrams, it could mean that they were shooting messages off into space and they didn’t care too much about the accuracy.  That is difficult to believe, but may be accurate because there were errors, some deliberate.  If the word counts were not transmitted in military telegrams, they were certainly recorded.  There are numerous accounts where each quarter a summary of telegraphic activity was submitted to superiors.  That summary included message word counts, numbers of messages transmitted, and received.  The other observation is that the WU rules didn’t seem to use word counts for error detection as much as for tariff gathering.  On commercial lines, money was the name of the game.  Apart from those two anomalies, the Rules of 1866 generally seemed to be in play for both the North and the South.  In all cases what I saw including notes in the margins, the spelling of numbers in the text, and the word count for the text seem to be in tune with the WU Rules of 1866.  As it turns out, they were the same procedures alluded to in a book by Prescott written in 1860.  Therefore, I think I can say with some confidence that using the WU Rules for message preparation during the war is an accurate assumption.


I have divided a typical message into five pieces.  The first is the preamble, followed by the “to,” then by the text.  Next is the signature (“from”) and finally the message management indicators.  I should also note that once the word counting is understood, the rest of the Rules as printed seem very straight forward. 


The example that I use is from the WU Rules page 12.


The preamble is made up of the telegraph company/organization announcement along with the date and place the message is originated.  The banner at the top of the telegram blank provides the organization, such as the United States Military Telegraph.  This part of the telegram or message does not get transmitted or word counted.  The next part of the preamble does get sent.  Please note that the origination location of the telegram may not necessarily be the same place as the transmitting station. For example:


Buffalo, Oct. 10th, 1865.


This message originated in Buffalo on 10/10/1865 but was transmitted from Rochester.  The preamble counting is a bit strange, but I’ll show it here to provide an idea of what they sent and how they counted words in the preamble.  If I were following these Rules for an impression, I would not count the preamble.  According to the Rules, there are seven “words” in that preamble.  The only way that is possible is to count punctuation, an unusual occurrence.


  1. Buffalo
  2. ,
  3. Oct
  4. .
  5. 10th
  6. ,
  7. 1865.


As part of the preamble, the message number was sent, but not charged for.  That means some sort of a sequential numbering system was used.  From what I observed, the numbering was restarted each day.  Also, the last period was not charged for.  It, as I will discuss later, was a separator.


The “to” is not word counted, but sent.  In the example:


John Brown, Geneva, N.Y.


Note again the use of punctuation.  The commas are significant as is the period.  The Rules note that a period is always inserted after the last of the “to” prior to the text as a separator.  Today, we might insert BT or break rather than a period.  There were still a few offices using printing machines and registers, so the BT would not have been necessary and a period would have worked fine.  I’ll discuss the commas later.    


The text in this example is:


Meet me here next Monday, at 10 o’clock in the forenoon.


There are eleven words in the text.  They are:



  1. Meet
  2. me
  3. here
  4. next
  5. Monday
  6. at
  7. ten
  8. o’clock
  9. in
  10. the
  11. forenoon.


Note that punctuation is not counted as it was in the preamble.  Again, after the text is sent, they added a period for a break.


The signature is sent but not word counted.


In the example, the signature is H. Smith.


The last part sent is the word count, station origination, the tariff and other message management indicators.  Note there is a period inserted after Smith, remembering that a period is like a break or BT.


In this case:


18 Rh 64 pd if we are counting the preamble.  If we don’t count the preamble, the last line becomes 11 Rh 39 pd, where 39 would be the charge.  Since in our impressions we are doing “free telegrams,” unless we want to make a few bucks on the side, I would eliminate the tariff part, or just put free, as they might have done according to the Rules.  Using that premise, the last line would be: 11 Rh free.  It’s interesting again to note that today, when telegrams are sent in the American Radio Relay League National Traffic System, and as described by the Morse Telegraph Club people, the word count and tariff information was sent as part of the preamble.  From what I have observed looking at the telegrams, telegram blanks, and the WU Rules, it was sent at the end.  All that means eighteen words (or 11 words) transmitted from Rochester.  They paid 64 cents for the message.  Reviewing telegrams in the CSA arena, I found the notation for word counting and application of tariffs was a bit different.  Perhaps it’s what they did on the American Telegraph lines.  Using the example above, the notation would have been something like “18/64 col”, meaning there were 18 words and to collect 64 cents at the receiving end.  Typically, in the CSA example, the sending station was listed next. 


Once we understand how numbers and punctuation is counted, word counting is pretty straight forward.  The Rules explain what a word is and how numbers are counted.   I’ve mentioned this before, but a great deal of the message management is done in the margin of the message, and much of that management information is transcribed onto a blotter, or register.  Because message management is done right on the message, recording in the record books can be done later at a more convenient time.


I have a few of observations.  First, the old guys seem to spell out nearly everything.  They used few abbreviations.  In fact, they spelled out the numbers, as we might today.  They did use station call signs in the message management portion of the telegram.  Also, I found the uses of either the “92” code a couple of times or the WU reply request code.  It was the number “33” which means deliver only to the person addressed under the newer 92 code or wait for a ten word response under the WU reply code.  Either would have been applicable in the messages I read.  The “33” was coded in the margin, and would have been sent in the last line with the word count information.  Interestingly enough, the 92 code talked about in the Shaffner book written in 1859 describes the number 33 as “answer paid here.”  So, if we apply the WU rules, I think we are close to the mark for accuracy.  Another observation related to the time of day.  The time of day may have been recorded in the margin, but not sent.  Later (after 1900), I found examples where the time of day was sent. 


Speaking with a former Western Union operator, and reading some information from the Morse Telegraph Club, another symbol was added to help with the “to” part of the message.  After each line in the “to,” the symbol “a a” was added by an operator.  Today we still add the symbol “a a”.  It really isn’t “a a” however, it’s a comma in (American) Morse.  Hence we see the use of commas as separators in our examples.  If we were doing messages by sound, as most of us do, I would add the “a a” rather than a comma if using the Continental Code after each line in the preamble for separation.   


There may be some concern we will take too much time to send an official message using this WU method.  That may be true during an operator’s learning curve, but once you get the hang of message traffic using the Rules, things should move along quite fast.  According to Prescott, an experienced operator can pass messages at the rate of 25 words per minute, each message containing an average of ten words.  I don’t know if we are to that level yet, but I know we can pass traffic at 10 to 15 words per minute, because we are doing it.  I would be thrilled if we all could pass 15 messages per hour.  That is sending speeds of less than five words per minute.  Five Word per minute allows for beverage and comfort breaks and really isn’t too bad.


Although I did not find documents where operators exchanged information without it being “official” traffic, I found many references to the fact that they did talk to each other.  I don’t believe those conversations followed any specific format, but were generally tied to the Rules for calling another station and keeping some sort of order.  Thinking that the operator’s day was not necessarily packed with message traffic, there was plenty of time for chit-chat.  I would suspect that the 92 code was used a great deal for station-to-station traffic.  For example “73” was and is still used today.


So that is what I found.  I’d say these methods are easy to do in the 21st century, and are historically accurate.  I hope my limited research will prove useful to someone in carrying out their impression.