Hi Don
@DaytonandElgin
Always good to discuss these things, sometimes you get new ideas from a different person's point of view.
Here's what I think I know (which may, or may not, be correct, but I've tried to give the evidence where possible).
I have compiled a survivor's list of Huffmans, from the start of production, through to around mid 1947, when the new Huffman dating code system was introduced..
It started as just a list of Military Daytons, but after a while I realised I needed to know what else was going on during this early phase in order to complete the picture. My list is currently around 210 bicycles.
It soon became obvious that trying to accurately date prewar Huffmans by serial number is not easy, since it does seem that they made the forks and frames, then used whatever was next on the shelf to build the required model. So over time there is progression to higher serial numbers and newer forks, but it is not a constant rate.
Incidentally, the earliest fork date I have recorded is 58, May 1938, though there could be earlier ones.
Wartime changed the random assembly, and as mentioned, the frame serials and forks generally correspond in date. This is confirmed by the Morrow hub dates, and also because, helpfully, the military bikes have the year stamped on them too. (-:
Military production stopped at the beginning of 1944, and Victory Lightweight production effectively stopped around August 1944. After that Huffman went back to normal civilian production.
Considering the chart below, and comparing it to known survivors, several errors are evident.
Firstly, though I don't know the exact dates, there was a period between when the serial number was on the crank, to when it was on the rear toe plate. In 1951 at least, it was being stamped on the saddle stem collar.
For 1938, 1939, 1940 and 1941 only a very limited number of frames had serial numbers starting H, mainly made 1938, some 1939. The rest had prefixes of A, B, W, D, N, HW, perhaps others. Also serial numbers followed by B, C, D, FAH, FBH, HAW, HBW, SAH, perhaps others. And also serial numbers with the following letters underneath D, FBH, FCH, FSH, HCW, again, perhaps others.
The letters likely refer to the company Huffman are contracting for, S being Firestone, for example.
At no time between 1938 and 1941 did serial numbers reach 400000. (or 800000!). The highest number recorded by me is N102578. Low numbers are in the 100's, I have noted 104FAH.
In 1942 the prefix D was used from the start of the year, (first recorded surviving serial is D1477, with the serial changing to H (for Huffman) at the request of the government in September 1942 (see Federal Register below).
The first change to H from D I have recorded is H93937 which is a Victory Lightweight with a September 1942 fork, so corresponds well.
The first 1943 dated bicycle I have recorded is H102177, a USMC Lightweight with a January 43 fork, and Q1 43 Morrow hub.
The last 1943 dated bicycle I have recorded is H157495, a G519 Dayton with a November 1943 fork and Q3 43 Morrow hub.
The first 1944 dated bicycle I have recorded is H158382, a G519 Dayton with a March 1944 fork and Q2 43 Morrow hub.
The last 1944 dated bicycle was around H220000 (going on fork dates).
The last 1945 dated bicycle was around H295000 (going on fork dates).
From then on H was the prefix until 1947 when the year dating system starting 7H began. The highest serial number I have recorded is H810302, which, according to the list above is almost the last of the H numbers, and made in 1947.
From 1942 to 1944 serial numbers never approached H800000 to H999999, and only just exceeded H800000 in 1947.
After 1947, the 7H, 8H etc. makes dating a bit easier, but the number sequence repeats every 10 years, so there has to be some knowledge of bicycle models to identify which decade the bike belongs to.
I hope that makes some sense (-:
Best Regards,
Adrian