I've found some great information in the book "Arsenal of Freedom", page 75, from the fiscal year 1896 report. Six carbines were used to test sights, the serial numbers listed in the results table are: 11803, 11804, 15787, 15903, 15969, and 21451. When I inquired through the Springfield Research Service for information on my carbine with serial number 16720, they responded that they had no information on my particular serial number, the closest they had records for was a carbine with serial number 16968. I've read that the Springfield Armory would have stockpiles of parts, which they would draw from to assemble rifles & carbines, so that would explain how the serial numbers could jump around. The 1898 cartouche would make sense for when they upgraded it with the 1898 bolt.

I don't mind at all if you add my carbine serial number to the KCA data base, thank you.

It's been a while since I've checked-in. I'm hitting the books again and compiling as much detail as I can to, at the very least, prove that it's possible that this carbine is authentic, and not a parts gun. If along the way, I wind up proving that it is a parts gun, I'll have learned a great deal from the research effort.

Dick Hosmer wrote on Mar 14^th, 2019 at 2:22pm:

  When you find your copy of Mallory 2, check page 78, right column under topic of Major Variations:  "Most Model 1896 carbines will be found with a Model 1896 rifle sight, rather than a carbine sight; and although there is a good supply of Krag spare parts, Model 1896 carbine rear sights are practically unobtainable.  No reason has been found in the Ordnance records for this situation."   Then check Brophy, page 54: "It is likely, though, that during the late years of the Krag, the Model 1898 and Model 1901 carbine rear sights were used as replacements.  Records do not indicate that rifle sights were ever used on carbines."  This would agree with you and some of us as well.

I don't have access to Mr. Farmer's work, but I would be interested to see if the above two statements are referencing the same record or if there is yet another possible explanation.