Colt Forum banner

1 - 20 of 517 Posts

·
Registered
Joined
·
2,412 Posts
Discussion Starter #1
With the help of some of you, I have been accumulating data on serial (crane recess) vs. service (butt) numbers for Colt 1917 revolvers. While I will continue to do so, I thought is was a good time to show you what I have so far, at least in a general sense. As you know, there is no absolute correspondence between the two numbers, but there is a general correlation. I was interested to know, among other things, how much difference there was between the correlation and the actual numbers. The data are shown below:



There is a fair amount of data displacement from the regression curve, and it seems more pronounced during the middle of production. The experimental regression curve itself is very close to the theoretical one, the small difference likely due to the finite data set (about 70 values). The maximum deviation is about ±12000, although much of the data is within ±5000 or less. The standard deviation of the absolute deviations would be a better measure of scatter, but I haven't gotten around to a better statistical analysis of the data yet.

Anyway, that where the research stands so far. Please keep the numbers coming. :)

Buck
 

·
Registered
Joined
·
935 Posts
"There is a fair amount of data displacement from the regression curve, and it seems more pronounced during the middle of production. The experimental regression curve itself is very close to the theoretical one, the small difference likely due to the finite data set (about 70 values). The maximum deviation is about ±12000, although much of the data is within ±5000 or less. The standard deviation of the absolute deviations would be a better measure of scatter, but I haven't gotten around to a better statistical analysis of the data yet."

Uh, yeah, sure it is. :(

In other words, they come purty close, right?

Me thinks you have been eating your user name again. :D
 

·
Premium Member
Joined
·
555 Posts
That's interesting. I have long believed that statistical analysis of serial numbers can open doors for our understanding of particular models in the larger context of a company's overall production.

Keep it up!
 

·
Registered
Joined
·
1,532 Posts
I agree. Excellent analysis. We need more people doing such things and advancing the knowledge base.

Thanks,
Kevin Williams
 

·
Registered
Joined
·
2 Posts
I have a Colt M-1917 and I am trying to figure the date of manufacture, but because this gun has several different numbers and stamps on it, I am getting conflicting info.

There is a matching 3-digit (52X) on the crane, opposite on the interior frame, and the front of the cylinder. It also has a matching 4-digit number (559X) on the barrel, the crane, opposite on the interior frame, and the cylinder base (under the extractor). And a matching 6-digit (189,2XX) number on the crane, opposite on the interior frame, and under the side-plate. ... At the Colt Web Page at Colt Firearm Serial Number Lookup, the 6-digit number matches to the 1917 model with a manufacture date of 1919. (It is my assumption that the 3-digit and 4-digit numbers were added during Arsenal or Colt Factory refurbishings.)


There is a 5-digit number on the grip-frame butt that is the military serial number, and this gun also has the inspector mark of Colonel John M. Gilbert (JMG), which I have read elsewhere appears only from about serial number 29,700 (1910) to about 64,000 (May, 1918). Based upon the Gilbert stamp, this guns 40,0XX military number would match an acceptance during WW-1 and before May 1918.


The gun was also parkerized, presumably by Colt when it was refurbished shortly after World War II, - when they added a VP or Verified Proof mark just above the trigger guard. This gun also has Colt Factory Bakelite grips with silver Colt medallions, - I presume that these too were added during Post WW-2 refurbishing.

So I am trying to figure out whether it was produced 1917, 1918, or 1919?
 

·
Registered
Joined
·
2,412 Posts
Discussion Starter #7
Welcome to the forum.

Send me the full service (butt) number (40xxx), and I can give you the week it was delivered to the Government and how many were delivered that week. I would also like the full Colt serial number (189,2xx) for my M1917 database, if you agree. Nothing in the database ties a specific gun to its owner. The numbers 52x and 559x are likely arsenal rebuild numbers (some guns were refurbished twice) between the world wars, and 559x was likely done second since that number is on the barrel. The 52x barrel was probably replaced during the second rebuild. In addition to the places you mentioned, the rebuild numbers were supposed to be on the ejector rod shaft and internally on the hammer and trigger.

The verified proof is interesting. Like you, I would suspect that it went to Colt after WWII under one of their refurbishing contracts. I am not sure where Colt is getting their dates - perhaps from Wilson's research which is in error for New Services in the WWI era. The same inaccuracies plague Proofhouse's dates from the same era, primarily because it is based on Wilson's data as well. There were very few M1917's produced in 1919, maybe about 5000 before the end of February. Yours is the first half of 1918. Again, I can get much more specific with the service number.

Buck
 

·
Registered
Joined
·
2 Posts
Welcome to the forum.

Send me the full service (butt) number (40xxx), and I can give you the week it was delivered to the Government and how many were delivered that week. I would also like the full Colt serial number (189,2xx) for my M1917 database, if you agree. Nothing in the database ties a specific gun to its owner. The numbers 52x and 559x are likely arsenal rebuild numbers (some guns were refurbished twice) between the world wars, and 559x was likely done second since that number is on the barrel. The 52x barrel was probably replaced during the second rebuild. In addition to the places you mentioned, the rebuild numbers were supposed to be on the ejector rod shaft and internally on the hammer and trigger.

The verified proof is interesting. Like you, I would suspect that it went to Colt after WWII under one of their refurbishing contracts. I am not sure where Colt is getting their dates - perhaps from Wilson's research which is in error for New Services in the WWI era. The same inaccuracies plague Proofhouse's dates from the same era, primarily because it is based on Wilson's data as well. There were very few M1917's produced in 1919, maybe about 5000 before the end of February. Yours is the first half of 1918. Again, I can get much more specific with the service number.

Buck
The full service (butt) number is 40011, - the full Colt serial number is 189205.

Thank you Sir !

Rodger (AbnCavScout)
 

·
Premium Member
Joined
·
2,316 Posts
US Army 1-993; Colt 152364


If you already have this please excuse the duplication.
Thank you for providing the analysis.
rayb
 

·
Registered
Joined
·
2,412 Posts
Discussion Starter #10
Rodger, your Colt M1917 was delivered to the Government during the week of March 2, 1918 as part of a weekly shipment of 2200 guns.

rayb, your very early M1917 was delivered during the month of October, 1917 as part of the monthly shipment of 3800 guns. I don't have a weekly breakdown for the first month of production. BTW, I already do have your data which you likely had given me earlier.

Thanks much to both of you for enriching my database.

Buck
 

·
Registered
Joined
·
1,374 Posts
Here/s another: 240565/92111 It fits your regression equation fairly well. If you can provide precise dating information on these numbers, please do so.

Statistical analysis is a valuable tool. I have read stories about how, during WWII, allied statistical analysis of even a small number of serial numbers on captured or destroyed German equipment (tanks, planes, artillery, etc.) provided very valuable intellegence information for estimating production rate and the number produced, which was later shown to have been very accurate.
 

·
Registered
Joined
·
2,412 Posts
Discussion Starter #13
Thanks, guys.

M1917 service number 69785 was delivered to the Government during the week of May 25, 1918 as part of a weekly shipment of 2600 guns.

M1917 service number 92111 was delivered to the Government during the week of July 27, 1918 as part of a weekly shipment of 2700 guns.

Buck
 

·
Registered
Joined
·
1,374 Posts
Exactly what is the Service Number supposed to represent, and why did the Army use both a Service Number and a Serial Number on the M1917? I don't know of any other US weapons using that dual numbering system. I've long been aware that both numbers were used, I just don't know the reason for it. Maybe something about separating identification of Colt and S&W M1917s?

I did a little more research and didn't get any good results. What I did find (correctness not assured) was that Colt 1917s were made in the factory SN range of 170500-259999. No idea of the Colt serial number when the service numbering system started, or at what service number (1? 10? 100?). What is the lowest service number known?

Another Colt 1917 number pair as a data point was also found mentioned on another forum - 188974/40134. You may already have it.

Using only the 5 data pairs found in this posting, I calculated that the slope is 0.9991 and the y-intercept is 150037, with R2=0.999. I doubt that any additional points beyond those you have already accumulated will improve the regression analysis materially. All indications are that the service numbering started at about serial number 150,000 and the service numbering tracked the serial numbering very closely thereafter. Human error in applying service numbers incorrectly probably caused the slope to deviate from being 1.0000.
 

·
Registered
Joined
·
2,412 Posts
Discussion Starter #15
DWalt,

I've interspersed some answers among your various questions below.

Exactly what is the Service Number supposed to represent, and why did the Army use both a Service Number and a Serial Number on the M1917?
The Army required its number on the butt as done on previous purchases. Colt considered the M1917 to be just another New Service and numbered the crane cutout in sequential order in the same manner that all other New Services were done. S&W just created a new model called the M1917, based on its large hand ejector, and gave it its own serial number sequence marked on the butt. That also served as the Army service number. Which number on the Colt was correct depended on who you talked to.

I did a little more research and didn't get any good results. What I did find (correctness not assured) was that Colt 1917s were made in the factory SN range of 170500-259999. No idea of the Colt serial number when the service numbering system started, or at what service number (1? 10? 100?). What is the lowest service number known?
Colt serial numbers ran from about 150,000 to about 305,000. Army service numbers started at #1 (sent to Maxim Munition Corp., Watertown, NY, for ammunition testing) and there were at least 154,802 guns built. There are more service numbers than guns, but the exact highest number is not known for sure. The lowest service number in my database is #7, owned by a forum member. The highest service number in the database is 154,658, also owned by a forum member.

Another Colt 1917 number pair as a data point was also found mentioned on another forum - 188974/40134. You may already have it.
I do, thanks.

Using only the 5 data pairs found in this posting, I calculated that the slope is 0.9991 and the y-intercept is 150037, with R2=0.999. I doubt that any additional points beyond those you have already accumulated will improve the regression analysis materially. All indications are that the service numbering started at about serial number 150,000 and the service numbering tracked the serial numbering very closely thereafter. Human error in applying service numbers incorrectly probably caused the slope to deviate from being 1.0000.
My objective has never been to establish the general correlation between serial and service numbers - that has been known for a long time. I am much more interested in other things that have or may come from the data:

  • There are a few number pairs that differ greatly from the regression line by as much as ±20,000 to ±60,000. These are numbers that I have visually confirmed. Some can be explained by Colt holding on to a particular gun for a long time and then releasing it to the Government. There are a few on the other side of the regression line that are not so easily explained. BTW, these "outside" points are not shown on the graph shown earlier in this thread. I intend to release an updated graph showing these points soon.
  • There are number of guns that regularly show service numbers higher than they should be by 2000-4000 numbers. I think that this is because Colt held them for testing for a week or two before releasing them to the Army.
  • I have found no strong evidence of a gap in service numbers. However, the database consists of only about 150 guns out of 150,000, and there are some places where data are still sparse.
I hope this answers what you asked. If you are really interested in New Services, I suggest you get a copy of Bob Murphy's Monograph, "Colt New Service Revolvers", which is both inexpensive and still in print.

Regards,

Buck
 

·
Registered
Joined
·
1,374 Posts
So who applied the Service Number, and when? Was it done by Colt or by the Army? Logic would dictate that the Colt SN and the Service Number would be applied at the same time, but logic does not always work. If the Army service number was applied at a different time and place than the Colt serial number, there are many possible reasons that the outliers noted exist.
 

·
Registered
Joined
·
2,412 Posts
Discussion Starter #17
Both sets of numbers were applied at Colt, but different departments handled it. Firearms manufacturing placed the serial numbers on the guns after they were done with them, then they went to a holding area, and after that the guns had the Army service numbers applied. No attempt was made to maintain a one-for-one correlation between the numbers because there was no need for it. Colt's number satisfied its internal requirements, and the Army's requirement was met by theirs. Army service numbers were applied just before delivery in sequential order. Any concern for consistency or the effect on collectors a century later was non-existent. The only objective was to get sidearms to the military as quickly as possible while maintaining acceptable quality.

It is easy to postulate several causes for guns having service numbers higher than we would expect from their serial numbers - the gun was made and something delayed its progress to the application of the service number. Service numbers lower than expected is a lot harder to explain other than the general "somebody screwed up" reason. Small differences in expected numbers in either direction is explainable by the random selection of guns for service number marking. But that reason can't handle lower expected service numbers by less than -4000, approximately 1-2 weeks production. When it's -20,000, I am stumped at a reason beyond human error. Any suggestions for systemic causes will be greatly appreciated.

Buck
 

·
Registered
Joined
·
1,374 Posts
I have no idea how numbers were applied or what diligence was used to prevent mistakes, but one possibility could be that the guy doing the service numbering reversed two digits - such as 35843 when it should have been 53843. But I would expect that type of error would have been caught somewhere along the line.
 

·
Registered
Joined
·
16 Posts
My M1917 Colt Serial Number 150742, US Army No. 7 was shipped to Winchester Repeating Arms on Oct. 24 1917 as one of two shipped for 45ACP testing.
Beverly Haynes, Historian Archives Services gave me the date of Sept. 15, 1917 as the date of manufacture by Colt Manufaturing Co.. The information stated here is in Bob Murphy's book "Colt's New Service Revolver" on page 34 and confirmed in letter from Colt Archive Services. The M1917 shipped with this revolver to Winchester Repeating Arms had US Army No. 8, per Murphy's book.

The US Army No. --- , were applied by Colt when a revolver was accepted by the Army Inspector. The US Army No. was a requirement of the US Army in their contract with Colt.
 

·
Registered
Joined
·
10,287 Posts
I certainly have little knowledge of the Colt Model 1917, but is the information on the pistols having the service number applied only after they had been through inspection and acceptance found in Murphy's book? This would be a change in the normal procedure of Ordnance acceptance, so any information would be appreciated.
 
1 - 20 of 517 Posts
Top