Regarding 1971 model Shovelheads, in my photo collection of authentic-looking, general production engine and/or frame VINs l have 122 examples consisting of eight 1As while the rest are 2As and 2Cs. Lowest is 1004+ and highest is 3308+.
For 1971 model Sportsters I have 127 examples of authentic-looking, general production engine and/or frame VINs consisting of 3As and 4As. Lowest is 1025+ and highest is 3325+.
For 1971 model Servi-Cars I have 16 examples of authentic-looking, general production engine VINs. Lowest is 140++ and highest is 336++.

NB: I know the full sequence numbers of all the VINs in my abovementioned totals. And they are all different.

However, SB 596 and my 1961–71 H-D Big Twin parts catalog say sequential numbers for 1970 Shovels are 10000 and up. And they both indicate the same thing for 1971 Shovels. Also a Harley Sportster parts catalog I looked at online indicates the same thing for Sportsters of those two years: 10000 and up. And sequential numbers 10000 and up are also indicated for 1970–71 Servi-Cars in the 1961–71 Servi-Car parts catalog.

NB: that total of 4700 FXs for 1971 is about right. I have a copy of a 2013 letter from Pete Simet at H-D to the owner of a 71 FX and Pete said approximately 4700 of them were manufactured.

NB: VINs for 1971 Shovels reached at least 34379 according to SB 630.


I’ve noticed this before although I don’t know the reason for it. I have 16 examples ranging from 441++ thru 446++. I also have one at 96+++ but I imagine it was a replacement engine because around that time VINs 90000 (?) and above seem to have been set aside for that purpose and I also have several examples of it from as early as 1971 model Sportsters and 1972 model Shovelheads.
Eric

 

Great information Eric...it seems as if I am on the right track!
I would absolutely love to have those full or abbreviated VINs if you ever decide to part with them. I think it would certainly add concreate to the list! Wish I could get those foil date codes as well.

If you would like direct data editing access to the sheets, I could certainly add you! I have only been doing this since December 2025. I have a long way to go to learn what's out there.

If you think the Harley Service Bulletins are hard to follow, you should take a gander at the aviation service bulletins and Federal Aviation Airworthiness Directives...what a freaking nightmare!

With the SB's, it is hard to know if they are referring to all of a series, or just a series between two dates without specifically mentioning it. For example, in SB 630 it says "Serial number of vehicles which must be corrected are within the following range of numbers." Aside from 10000...the lowest unit...other numbers include, but may not be limited to a total range of units made.

Looking at the SB range for the 1A and 2A, the SB states;

"1972 Electra Glide VIN
IA 10005 H2 to I4478 H2
2A 10456 H2 to I4479 H2"

Is this insinuating there were no 1As made below 10005 and higher than 14478? We certainly know that there were 2A units made after 14479 (as high as 37027), so it leads me to believe the sequence numbers are date related, I guess pending what batch of manufactured parts between those dates were used. Meaning the "bad batch" of parts were used on units between 10005 and 14478 , not necessarily that there were no units made before or after such numbers. (sorry I am rambling here...lol)

What does shed some light is that it could give us a hint that at least 10005 and 14478 was by all means assigned to 1As, while 10456 and 14479 was indeed assigned to the 2As. So what I understand is that I only have confirmation on the identity of four units, two 1As (10005/14478) and two 2As (10456/14479), while any number between those could be assigned to other completely different models such as the 2C, 3A 4A or 5A...confirming the use of a single set of sequence numbers for all six 1971 models. Not necessarily true for pre-1970, or post 1972 models.

What does have me puzzled are the higher 506xx to 526xx numbers for the 72' XL/XLCH/Servi-Car models. I just don't have enough examples listed yet :-)

Thanks for following along...correcting my errors and adding new information!!

 

Excellent!

Would it be possible to compile all this data into a searchable PDF file and continuously expand it with updates?
If the compilation up to the late Shovelheads ended up to the year 1984, then the file would be perfectly sufficient...?

I hope I have expressed myself reasonably well as a non-native English speaker, thank you.

Mike

 

Mike, thanks for joining in.

I am computer stupid so I will have to look at the PDF stuff to see if it is needed, or if I am capable of making one.. However, the google spreadsheets should suffice. There is really not enough information yet requiring a search. Anyone can be easily cross reference by year/model/VIN.

Go to the model year page, i.e. 1971 tab at bottom of page - look for your model in left column - then look from left to right to find where your VIN would fit. Confirmed date stamp ranges are highlighted in yellow.

The more data folks share, the more accurate it will become.

https://docs.google.com/spreadsheets...gid=1835629537

 

This topic has over 2,000 views and the spreadsheet has less than 300 examples, and 98% of those is what I found on line! So hopefully we can get some more numbers shared. Even if the frame is a custom chopper (probably no foil decal), the engine still shows the original application. As long as there is an HD number, it doesn't matter what it is being used in today......that information will help for the purpose of the spreadsheet!

 

As already mentioned, SB 596 and my 1961–71 H-D Big Twin parts catalog say sequential numbers for 1970 Shovels are 10000 and up. And they both indicate the same thing for 1971 Shovels.
A 1961–76 H-D Big Twin parts catalog I looked at online says 10000 and up for 1970 and later.
My H-D parts catalog for 1941–84 FL models indicates 10000 and up for 1970 models thru about mid-78 models.
My H-D parts catalog for 1971–84 FX models indicates 10000 and up for 1971 models thru about mid-78 models.
Also as mentioned previously, for 1970 and 71, certain H-D parts catalogs indicate 10000 and up for Sportsters and Servi-Cars giving us an overall total of three separate families with parts catalogs indicating VINs for each family start at 10000 for those two years.

But my collection of VINs for 1971 Shovels, Sportsters and Servi-Cars seems to have contradicted this so yesterday I went through my photos of authentic-looking, general production engine and/or frame VINs for 1970 models.
For 1970 Shovelheads I have 95 examples consisting of ten 1As and eighty-five 2As. Lowest is 1043+ and highest is 2662+.
For 1970 Sportsters I have 143 examples consisting of 3As and 4As. Lowest is 1021+ and highest is 2677+.
For 1970 Servi-Cars I have 13 examples. Lowest is 116++ and highest is 267++.
NB: I know the full sequence numbers of all the VINs in my abovementioned totals. And they are all different.


Again there seems to be a contradiction? Why? Given that 1970 was the first model year for those three families to receive the new and different number format, I’m not sure how the SB and the parts catalogs appear to contradict my photo collection. I wonder if each of the three families had 1970 VINs beginning at 10000, but at some point after that their sequence numbers were mixed in together? And if so, did the same thing happen with those three families for 1971?

Regarding 1972 model 1As, I have clear photos of four authentic-looking engine VINs higher than 14478 and they range from 253++ thru 393++. A fifth example is 421++ and it looks okay although the photo is a bit blurry.



Quote: ‘What does have me puzzled are the higher 506xx to 526xx numbers for the 72' XL/XLCH/Servi-Car models. I just don't have enough examples listed yet’


For 1972 Sportsters I have 208 examples ranging from 1026+ thru 4036+ but not long after that there may have been a break? Next I have 42 examples ranging from 5005+ thru 5244+. And one at 9048+ which may have been a replacement engine.
For 1972 Servi-Cars I have examples at 123++, 129++, 215++, 269++, 369++, 418++ and 525++.
Eric

 

I think this is what indeed confuses everyone (including myself), and then is interpreted to mean that each model starts with their on individual sets of sequence numbers starting with 10000. A completely understandable interpretation.

Since my theory is that only one set of sequence numbers is used for all major models, it would have certainly been easier and faster for the manuals and parts catalogs to just group them all under the 10000 to xxxxx, rather than to list multiple...exact...numbers each beginning model was assigned. Time is $$$$$$

i.e. sequence number block 10000 to 10500 assigned to the 2A model while sequence number block 10501 thru 10999 assigned to a 2C, while a smaller sequence block number of say...11000 thru 11200 assigned to the less popular 5A. All models falling inside a single set of sequence numbers, but each being assigned a small block set from that sequence. (I really need to get a life ya'll)

Yeap, exactly...if each model was assigned their own sequence set, all starting with 10000 and ending near 30000, then the total units produced based on such number would triple (based on three models) than the actual unit production numbers for those three models.


Again, I fall back to time is money, and the manuals etc. just listed everything as if each started at 10000 rather than taking the time to sort through each starting and ending number assigned to the subject model.


This is where the total units produced begin to match the single set ending sequence, rather than multiple sets adding up to way more than what was actually produced.. Each year (my theory) the six big Harleys (FL/FLH/FX/XL/XLCH and GE's were assigned a single set starting at 10000 and ending when the model was discontinued. However, it wasn't just that simple sometimes. Again we are talking block numbers assigned to each model inside the single set range...i.e. 10000 to 10199 (example to 2As), 10200 to 10699 (1As), etc alternating down the line as production continued....all they way down to how ever many were manufactured, or planned to be manufacture....for 1972 this could have been all the way up to 53000 with what I am finding. Some of those block numbers could be set on reserve and never manufactured, while skipping a few thousand or more, "voids" could be sporadic!



yes, there is currently a gap from about 43xxx to 49999. Possibly done on purpose to clear any confusion in paperwork. There is a name for that (called "clearing the pipes" to prevent clerical errors) but I can not think of it right off. The same was done during the AMF transition from Milwaukee to York. Phase One may have been to skip 30000 to 39999 when they continued the block numbers from 400000 to 50600 for the 73 3A. 4A and 5A made in Milwaukee...then Phase Two they skipped block numbers 51000 to 59999 when they rolled off 60000 off the new York assembly line in Feb 1973...and continued with the 1A, 2A, and 2C in York. I have at least one FLH with the VIN 2A72931H3, which falls under the 60000 to 73000 assembled in York. To add to the confusion, there was a block of 5A's that were certified (aluminum foil decal) in Jan 1974 and sent to South America.


Eric thanks for the hard work to share those, I know it took some time. Again, I would love to add all of those VINs (just partial would be fine) to the spreadsheet to show a more in depth pattern. I have since added a new test sheet just simply listing each year in order to show such sequence and dates. The dates (the certification foil stamp) data is extremely important in fine tuning such information.

Here is a direct link to the test sheet (it will not replace the other sheets)
https://docs.google.com/spreadsheets...#gid=654758359

I watched this video on a big screen in order to see the VIN and foil certification date code of 12 71

 

Here is a little something extra that might help explain my theory.The 1973 data so far collected supports one continuous serial sequence shared across all six full-size models, not six independent model-specific counters.

For 1973 production at Harley-Davidson, what the spreadsheet demonstrates is:

• A master numeric sequence
  
• Divided into plant allocation blocks (Milwaukee early, Milwaukee later, York)
  
• Applied across all full-size models being assembled at the time
  
• Not separated by FL vs FLH vs FX vs XLH vs XLCH vs GE
  

If there were separate sequences per model, we would see:

• Overlapping number ranges between models
  
• Parallel low starting numbers for each model
  
• Clear reset points unique to each model
  

Instead, what we’re seeing is:

• Non-overlapping, continuous number progression
  
• Models interspersed within the same numeric run
  
• Block jumps corresponding to plant shifts, not model changes
  

That is strong structural evidence of a single production counter feeding multiple models.

It also explains why trying to back-calculate production by treating each model as having its own serial range leads to inflated totals — because the same number pool was shared.

So the spreadsheet doesn’t just suggest it — it materially demonstrates that 1973 full-size production was serialized from a unified numbering system, segmented by plant allocation blocks rather than by model designation.

We can next examine whether that same unified pattern appears in 1972 or 1974 to confirm it wasn’t unique to 1973.

So now we can use such data to estimate production totals and compare them with what has been reported over the years.



From what the spread sheets show across:

• 1971
  
• 1972
  
• 1973
  

…the pattern is consistent:

• One continuous numeric sequence
  
• Shared across all full-size models
  
• Broken into plant allocation blocks
  
• No per-model reset
  
• No parallel model-specific counters
  

So the correct way to estimate production is:
Sum the actual allocated serial blocks per year
—not per model.

Step 1 — Use 1973 as the Calibration Year

From the 1973 sheet:

Block capacities totaled:

12,676

• 10,601
  
• 12,932
  = 36,209 possible numbers
  

Actual AMF production that year is widely understood to be in the low-to-mid 30,000 range.

That suggests:

• ~90–95% utilization of allocated numbers
  
• 5–10% unused/skipped numbers (normal AMF buffer)
  

That gives us a realistic utilization factor.

Step 2 — Apply the Same Logic to 1971 & 1972

Since the sheets show the same structural pattern in:

• 1971
  
• 1972
  
• 1973
  

We can reasonably assume:

• Same allocation method
  
• Same buffer behavior
  
• Same shared sequence logic
  

AMF-era production during this period generally increased year over year.

Based on historical industry data from early 70s Harley production trends, reasonable estimates are:

1971

~30,000–33,000 units

1972

~32,000–35,000 units

1973

~33,000–35,000 units

Those ranges align with:

• The block sizes shown in the spread sheets
  
• Known Sportster production (~20k in 1973 alone)
  
• Known Big Twin output scaling during AMF expansion
  

Step 3 — Combined Total for All Models (1971–1973)

Using conservative midpoint estimates:

1971 ≈ 31,500
1972 ≈ 33,500
1973 ≈ 34,000

Total ≈ 99,000 full-size motorcycles

Using broader range bounds:

Low estimate: ~95,000
High estimate: ~105,000


Therefore, if I had more data (from you guys) to enter into the sheets, it would indeed continue to validate, rather than invalidate the theory. More data would would actually better validate accurate totals built, rather then sway the single sequence number theory one way or the other.

 

Working with just 1973 Models, I have compiles the following List:

• One Set of Sequence Numbers ranging from 10000 to 73000, estimated
• One Set of Sequence Numbers broken down into Block Assignments, estimated
• using known surviving examples, minimal but effective

The following results were realized: https://docs.google.com/spreadsheets...#gid=942732228

Hopefully we can get some more folks to gain an interest and participate by offering more surviving examples.

 

Hi Bryan,
a compilation of pages of all kinds can be created with freeware such as "PDF24 Tools": https://tools.pdf24.org/de/
I scan the pages on a flatbed scanner and upload them to this software.
The software then creates a PDF document that is as small as possible, which can then be uploaded or sent – ​​a very simple process.
A comparable freeware program is certainly also available in English.

It is also a good idea to put the compilation online on free internet space and only post a download link wherever you like.
Only you will have access to the file, and you can edit it and post the latest version online.
I hope I have expressed myself reasonably clearly...

Mike
PS Of course, you can also download the freeware and edit everything at home without internet access; that would be my personal recommendation.