The CoO basically believed that simply a completed receiver was the 'completely prouduced rifle' whereas Springfield didn't consider a rifle complete until it was fully assembled and moved to their warehouse. Not only this, but what was considered a 'complete rifle' by the Chief of Ordnance and Springfield were completely different, so any semblance of records kept by Springfield storekeepers differ from what is reported by the Chief of Ordnance. However, even these reports are inconsistent. Most information we have on rifle production comes from quarterly reports between the Chief of Ordnance to the Secretary of War discussing the total number of rifles and carbines in inventory. So, to break it down, the short version is that neither the US Army Ordnance Department or Springfield kept proper records of the serial numbers for rifles/carbines produced. Click to expand.Just saw this! There is a huge discrepancy for Krag rifle serial numbers unfortunately, but this isn't so much the several authors faults as it is the US government.