Losses in West. Aug Through September 1944

[Art]

For example, those figures do not actually agree with those found in NARA RG 331, SHAEF General Staff, G-1 Administrative Section Decimal File 383.6/1, Daily POW Reports. It breaks down the PW reports by capturing nation as (UK/US/Total):
June 12,683/17,017/29,700
July 451/52,369/52,820
August 42,105/106,898/149,003

Is it possible that POWs captured in operation "Dragoon" were somehow retrospectively included in the August count?

It seems that in both cases prisoners captured at Cherbourg (or at least a large part of them) were included in July figures (the time lag mentioned above). I would also suppose that daily numbers were not necessary fully accurate and were subject to corrections later.

[Gooner1]

21 Army Group claimed about 67,000 prisoners captured beginning from 25 August until the end of September (from "Notes on operations of the 21 Army Group"). That includes some 16,000 in the operation "Market Garden" (the same source) and also the garrisons of Le Havre, Boulogne and Calais (about 30,000 combined). Add to that the Mons pocket (the usual figure is 25,000) and more than 30,000 POWs at Brest and that would make SHAEF stats look reasonable enough.

About 6,000 prisoners in Antwerp. The Market Garden figure looks too high.

[Carl Schwamberger]

Yes it would.

[Art]

About 6,000 prisoners in Antwerp. The Market Garden figure looks too high.

Probably, but that what "Notes.." say. Quite possibly that might actually include all prisoners captured by the 2 British Army during the period in question.

[Richard Anderson]

For example, those figures do not actually agree with those found in NARA RG 331, SHAEF General Staff, G-1 Administrative Section Decimal File 383.6/1, Daily POW Reports. It breaks down the PW reports by capturing nation as (UK/US/Total):
June 12,683/17,017/29,700
July 451/52,369/52,820
August 42,105/106,898/149,003

Is it possible that POWs captured in operation "Dragoon" were somehow retrospectively included in the August count?

That may indeed account for the apparent discrepancy in the September figures. SHAEF began including cumulative totals from 6th Army Group as of 20 September 1944. As of 1800 30 September, they totaled 87,705, since 15 August.

Nevertheless, 21 Army Group reported 108,582 PW for September, 12th Army Group 141,545. Thus, the total for September (including 6th AG since 15 August) was 337,832.

It seems that in both cases prisoners captured at Cherbourg (or at least a large part of them) were included in July figures (the time lag mentioned above). I would also suppose that daily numbers were not necessary fully accurate and were subject to corrections later.

Indeed yes there was a time lag as well as other issues. For example, US VII Corps reported 2,836 PW captured on 1 July, which were of course actually the final counts of the major surrenders at Cherbourg. In the same vein, while the SHAEF daily reports accounted for 17,017 PW captured by American forces in June, the First US Army recorded 33,321 captured of which 25,208 were evacuated (i.e. transported out of the beachhead to England). However, the reports by the PESBG and 1st ESB for PW evacuations from OMAHA and UTAH for June actually total 30,180, not the 25,208 recorded by FUSA.

The culprit seems to be differing reporting times, lags in getting reports in under deadline, and incomplete definitions, categorization, or summaries of cumulative figures. For example, there is strong evidence that wounded PW were evacuated and reported through medical channels and were not always included in counts of ambulatory PW. Also, these counts of PW include all types of Wehrmacht...Feldheer, Ersatzheer, Luftwaffe, Kriegsmarine, RAD, NSKK, OT, HiWi, Freiwillige, Beamte, and so on, as well as captured "foreign enemy nationals" such as Italians and even some others. I suspect a large proportion of the PW captured in September were actually assigned to the Militärbefehlshabern and were both military and "uniformed civilians".

[Carl Schwamberger]

Sorting though these numbers and the possible reasons for discrepancies reminds me of the relatively poor documentation among the Marines during Desert Sheild/Storm. I saw entire company & battalion size units formed that never existed on paper. Or worse where the documentation was grossly contradictory. If some historian ever tries to reconcile where all those men fit in tidy TO/E he's in for a frustrating time.

[Richard Anderson]

Sorting though these numbers and the possible reasons for discrepancies reminds me of the relatively poor documentation among the Marines during Desert Sheild/Storm. I saw entire company & battalion size units formed that never existed on paper. Or worse where the documentation was grossly contradictory. If some historian ever tries to reconcile where all those men fit in tidy TO/E he's in for a frustrating time.

Its worse than that Carl. When we did the Army Record Keeping Study in 2000, we found a steady decline in the quality and completeness of Army records compared to previous eras. There was a precipitous decline from the quality of World War II records.

What sparked the study was the inability of the Army personnel office in Alexandria to confirm the location of soldiers possibly exposed to chemical weapons and other hazards...in fact, they could not even confirm if individual soldiers were even in the war zone. One big reason I tell soldiers I run into now to keep their own file of their Army records - do not expect the Army to keep those records for you, which may be critical when it comes to retirement or working with the VA.

However, unit records were no better kept than personnel records. We found one instance where a unit (BCT) was told to "retire" their records at the end of the campaign and rather than sort, pack, and ship them to their home station, they instead dumped them in an empty Conex, doused them with JP and threw a couple of flares inside. End of problem.

So yeah, this particular "some historian" found you were correct, just nine years after the fact.

[Carl Schwamberger]

Not surprised at all Rich. Back in 1979 I took a course on 'how to do history' Professor Flanigan reviewed how evidence could be divided into five classes, ranging from anecdotal up to 'official record'. He then deconstructed how each actually is created and showed the trap of assuming a nice looking document with official stamps is likely to be anything resembling accurate. Who, wrote this, Why, When, Where, were questions he said were necessary for context and analysis of the evidence. I found that useful a few years later doing lost equipment investigations.

Working with veterans in trouble I've seen a fair amount of missing documents. Some times deliberate. What I also learned is that other individuals or sections are busy making excess documents. Duplicates, and extras. We've been able to prove claims cause documentation that never should have been made turned up.