Naval experts, mathematicians, and/or codebreakers

[Tom Morgan]

Thanks, Pete. It has been some years since I saw a plaque close up, and of course I didn't even notice the numbers until I learned about them here.

Tom

[gem22]

Thanks Pete

Now you can see why I stick to making pots of tea rather than tea-pots.

Garth

[Myrtle]

This seems, to me, to suggest some kind of order.....??

Yes the order could be in the pattern numbering not in the casting.

Do you have any interior photographs of Acton which show equipment that was used and/or methods of storage ?

RT

Just reposting this in case you missed it.

[Guest Pete Wood]

Sorry, M. No photos, that I'm aware of, exist of the plaques being made. Some of the equipment survives, though, and I have details of other equipment/machinery used.

I don't know how things were stored, though there was a storekeeper - and items had to be signed for (such as smoothing files etc).

[Myrtle]

RT

The system of numbering the patterns, would have probably related to a pattern storage system and also to a ledger.

Myrtle

[Guest Pete Wood]

Myrtle, are you suggesting that the factories kept the plaster pattern for Joseph Sylvester Smith.....?? So 1,000,000 patterns were kept....?? Why...??

That doesn't make sense to me, so I am obviously missing your point.

Please give me an example of how you think the numbering/storage/ledger system would have worked.

I fully accept that I am as dense as a railway sleeper.

[Myrtle]

RT

I could be wrong but I was under the impression(Sorry about that ) that there would have been a main pattern for the plaque and then a pattern for the name alone layered on top. The patterns for the plaques would have been stored and the name patterns would have been discarded. I believe that with some items the layered patterns would have been glued but in this case I don't know.

M

[Clive Maier]

The sample of names is quite a small one, and it may not be sufficient to reveal any batch number pattern. With that proviso, I have put the data into a table and sorted it this way and that, and I can’t see any pattern. I am minded to suggest that there is no link between the batch number and any detail of the person commemorated. Given that the person’s name is already on the plaque, and that details can be linked to that name, a batch number performing the same function would seem to be redundant. Except of course, that names were not unique in many cases. But at least one of your names – Edgar William Grimaldi – was unique. According to CWGC, the only other Great War Grimaldi has the initial JM. So it seems unlikely that the numbers were intended to distinguish one William George Brown from another.

I don’t know too much about foundry casting as it was practised around 1918 but I do know about injection moulding. This can be regarded as pressure casting into a closed mould using plastics. The metals equivalent is diecasting. An injection mould may contain one or more cavities or impressions. A four-impression mould for example, would make four similar articles in each operating cycle. It is common practice to engrave a small identifying number on each mould cavity, so that it will appear as a raised number on a non-presentation surface of the moulded part. This is so that any quality problem with a moulded part can immediately be related to the responsible mould cavity. If the production requirement is very large, there may be a number of identical moulds. Now there are two possibilities for numbering the cavities. You can just carry on in sequence or you can use a composite number. For example, 24 could indicate cavity 4 in mould 2. There are obviously many alphanumeric variations on this theme.

There is an important different between injection moulding with a closed mould and gravity casting in an open mould. The injection mould is extremely durable and will be used for hundreds of thousands of operations. The foundry sand mould on the other hand is not at all durable and will only serve for one operation. So while the mould is the important identifier in injection moulding, it will be a precursor of the final sand mould that is important in gravity casting. So it makes sense that the number was stamped into the plaster intermediate positive that was used to make the foundry sand negative.

Is the batch number a composite number? It seems unlikely because of the presence of zeroes (50, 80). Numbering schemes rarely if ever started from zero in pre-binary days. The number could indicate a worker, or a team, or a production cell. It could, at least on the evidence of the sample, indicate a week number with 96 being almost two years from the start of production. It could be a metal batch identifier. I feel that the identifier is more likely to be there for quality audit purposes rather than piecework considerations. The latter would mean that book-keeping minutiae was being perpetuated for all time on the plaque. It seems unlikely.

Presumably, the plaster intermediate positive would itself be cast from a negative mould. If the negative could only produce a limited number of plaster positives before degrading, this could be a reason for serially numbering the positives, especially if you were experimenting with different negative materials in an attempt to increase durability. This would also explain why numbers recur.

Unfortunately none of this explains why Woolwich women’s plaques were numbered 11. And I have never understood the narrow H.

[Clive Maier]

What are the highest and lowest known batch numbers?

[Guest Pete Wood]

Interesting, Clive. Many of your theories/ideas are things I have thought about and investigated, but there are some new ideas, here, which should be considered.

For example, I have tried every formula I can think of to marry a time period (day/week/month) with the number on the plaque. But none of my formulae work.

The reason I thought, like you, it might be the week of the year is as follows (and also answers your last question).

Acton were producing plaques with the numbers 1 – 51, and all have a wide H. The number on the plaque is placed outside the lion’s leg. You do occasionally find the odd plaque with no number at all. These cover casualties from all periods (months and years) of the war but, because of their finish (quality), I have attributed them to Acton. Acton produced around 200,000 plaques in two years. I can find (almost) no evidence that Acton produced any plaques for Navy/Maritime casualties – nor were any plaques made for female casualties.

But the 51/52 weeks of the year, does not match up with any casualties (and I have used every variable possible, with week one starting in August or January etc, to no avail) or numbers on the plaques.

Woolwich did carry out some experimentation, prior to production, with stamping. But the results were not good, and the Acton ‘casting’ method was used. The detail finish of the Woolwich plaques is not as fine as that of the Acton plaques, but after consulting many people with foundry experience, this has been associated with inferior products (mainly the type of sand) and bad working practices (incorrect pouring temperature etc). This backs up statements made by Manning Pike, Acton’s foundry manager, comparing the factories’ quality control.

When production started at Woolwich, it obviously thought that the Acton numbering system should also be continued. Woolwich must, I am sure, have inherited the wide H, Acton moulds (or masters) – as there is no difference at all between the castings from either factory - and used the numbers 1-100 – with the numbers inside the lion’s leg (and in most cases stamped WA on the back).

Again, I have only come across a few naval/maritime casualties that were made at Woolwich with a wide H. However, at this stage it was realised that no provision had been made for female casualties. Carter Preston, the plaque’s designer, was paid to modify the master – replacing HE DIED with SHE DIED; the S & H were narrower than other letters, because of the confines of space.

This SHE DIED master was then adapted (by removing the S) that resulted in the narrow H, HE DIED master – that is used for almost all Naval/Maritime plaques, (though many army plaques were also produced with a narrow H). The narrow H plaques were also numbered 1-100.

I have come across many examples of army casualties (with unique names), from the same unit, and who died on the same day. In some cases, the plaque numbers are the same (just like the examples shown on this thread, for the Jutland casualties). But there are also casualties, from the same unit, who died on the same day, that have a different number. On a few occasions, I have found that, as well as having the same numbers, the plaques were produced with a narrow AND a wide H.

This has led me to believe, though I can’t be certain, that both the narrow and wide H masters were being used at the same time – or there was a cross-over period, at least. In theory, this means that Woolwich may have been using 200 (100 x narrow + 100 x wide H = 200) numbers on plaques, as opposed to the 51 by Acton.

Around 800,000 plaques were produced by Woolwich in the same time frame, as those produced by Acton.

The negative moulds that survive are made from metal. This does not surprise me. If the factories had used one-piece negative plaster moulds (which also wear out quite quickly) to produce positive plaster patterns, physics would have worked against them. If you pour liquid plaster (slip) into a plaster mould, the liquid expands as it dries. So the mould would have to be made in two parts. Having examined many plaques, I have never come across any evidence of flash (join lines), to suggest that a two piece mould was used. So I am (fairly) sure that the metal moulds are genuine and were the main source of producing the positive plaster patterns.

Like you, Clive, I am unable to associate the plaque numbers with a date. The closest I came was to take a unit, say 1 Hampshire (for example), and divide the number of dead by the number on the plaque. So the 248th soldier to die had the plaque number of 24. So I started to work on 10 casualties equalling 1 on the plaque; the 22nd casualty would have plaque number 2 etc. This seemed to work with the (relatively low number of) 1914 casualties, but not with 1915 casualties – and I realised I was in danger of trying to make the theory ‘fit’ the evidence.

Why were ALL the female plaques stamped with the number 11?? That I can’t answer, either. The same number, 11, was used for male casualties also….

It also makes little or no sense to record book-keeping details on the plaque. But, with respect to all the ideas/suggestions made by all, I haven’t (yet!!) heard something that makes me yell, “Eureka.”

So that’s when I decided to start this thread.

Please keep the suggestions/theories coming, though. I’m sure the answer is out there…..

[Clive Maier]

Blast! This is too interesting. My deadline is going to kingdom come and I have to go out this afternoon. Ah well, it’s back to night shifts again.

I tend to the view that the number is a quality or manufacturing identifier and has nothing to do with the person commemorated. So if the number is a week number, that would be a manufacturing week, not a week of death. I did not understand your point about 51/52 but if two numbers in the series are missing, that could be two weeks of annual shutdown.

I don’t know how the factories received their lists of names but these may not have been the nicely ordered lists we are used to today. Perfect sorting is a product of the computer age. Prior to that, sorting and ordering was a considerable hassle. And even if they received ordered lists, the plaques may not have been manufactured in strict list order. I imagine there was no undertaking to post the plaques to next of kin in any given order such as date of death, seniority of service or regiment, theatre of war or whatever.

The factory may have had a job card for each plaque. Their only concern would have been to move the card from a metaphorical in tray to an out tray where it would be joined by the named plaque. Provided the ‘in’ count matched the ‘out’ count, the factory would be happy. The lack of any unit or date-of-death relationship between one plaque and those made immediately before and after it would, it seems to me, have no significance for the factory.

[PhilB]

You mention, RT, that stamping was tried but wasn`t successful. I find it hard to think why stamping wouldn`t have been a better and cheaper method. It would probably have necessitated the name being stamped on separately, but that doesn`t seem to be a problem. They could also have had a designed reverse instead of the cast plaque`s rough, flat back. Any comments? Phil B

[Myrtle]

When moulds and patterns were not in use they would have been stored. Larger items on shelves, smaller ones in boxes or similar. To locate these a list would have been kept .To ensure that the correct item was being used, the mould/ pattern would have been marked in some way to ensure that the mould/pattern found in that particular place was correct. e.g. where there is more than one company using the factory, the mould would include the customer's initials. John Price Ltd moulds could be marked with JP.

Simplified eg. Mr John Price would realise that he needed 6 more items to be cast and would place his order. The foreman at the factory would look at his list to locate where the moulds were being stored. He would go to that shelf and picking up the mould/pattern, look .to check that it showed the raised or indented initials JP. He would then send it into production. Once 6 items were produced they would be sent to Mr. John Price. To be sent to Mr Price the order would have to be recovered to ensure the 6 items were sent to the correct address. It is more likely that the order information would have been kept in a ledger or on cards which would then be marked to show that the items had left the factory. If for any reason the 6 items did not reach their destination there would be recorded proof in the ledger/cards that the items had been produced and had left the factory premises.

If moulds were not marked in some way it would not have been possible to guage how long or how many times they had been used, therefore denying the possiblity of rotation. Also there would have been difficulty in stock taking and keeping tabs on their location.

If any of the plaques had not reached their destination then how would the factory know if that plaque had been sent or not, if they had not kept records of production?

[Guest Pete Wood]

Phil

Remember that, by now, Acton had completed around 20% of the plaques - whcih had already been sent out to the next of kin.

If Woolwich had been able to match - in an identical manner - the plaques that already existed, then this would indeed have been the easiest/cheapest option.

My understanding is that, because of the 'raised names on the name plate' stamping would have been almost impossible to achieve. If the original design had called for the name to be BELOW the name plate surface, then it would have been a cinch.

Remember, also, that the person who came up with the idea of a cast plaque, was Sir George Hill - the keeper of the coins and medals department at the British Museum. He was deeply fascinated by the history of casting techniques used by the ancients. Hill had a lot of influence on the whole plaque story, and he was heavily respected by everyone from the King downwards.

Woolwich would have had to show Hill their attempts and, because they weren't perfect, there was no way Hill would have accepted a second-rate item.

Clive

I apologise for the confusion. I was trying to say that Acton had 51 numbered plaques and some plaques were without numbers - so, in theory, 52 variations. But I agree, in those days, it was not unusual to only shut down a factory for a few days around the Christmas period.

No one knows how the factories received the list of names. Sadly, none of the paperwork survives. But I would like to explore further your thoughts on the sorting that may have been employed at Acton/Woolwich. While I agree that the factory did not have to carry out production in any particular order, can't help but feel that there must have been a system that worked in harmony with the military records offices.

The records offices would have sent a list (or lists) of names to the factory. Certain records offices were responsible for more than one regiment - eg Exeter was responsible for Hampshire, Dorset, DCLI, Somerset LI, and Devon regiments - so the lists could have been alphabetical, by battalion, regiment etc etc.

How do you think the operation would have been sorted?

Why would Woolwich want to carry on a numbering system, whatever it was for, that was used by Acton - and (roughly) double it?? Remember, Acton thought that it was always going to have the contract. So the 1-51 system was working fine for them. When Woolwich took over, it knew (almost) exactly how many plaques it had to make - and could be more.... organised. Woolwich was much more military minded than Acton.

My last 'hope,' of finding a link, if there is one, between the plaque numbers and casualties is to see if there is a common factor on the existing records in Canada.

In the Canadian records, I have recently discovered (thanks to the Canadians on this forum) that the production numbers are listed in the soldiers' records. So if it is found that Pte A Smith had plaque number 678567 and Pte B Smith has plaque number 678568, it might go someway to explaining how the factory decided on the order of production.

[Guest Pete Wood]

Myrtle, I need to think about this, to give a decent reply.

Both Acton and the Woolwich Arsenal Brass Foundry were making only memorial plaques. However, I agree that some system was very much required. I'll post something soon, which might take your idea further.

Anyway, some figures to make you think a little:

Acton was pumping out approximately 20,000 plaques per week for two years.

Woolwich was pumping out approximately 80,000 plaques per week for two years.

[Myrtle]

Acton was pumping out approximately 20,000 plaques per week for two years.

Woolwich was pumping out approximately 80,000 plaques per week for two years.

RT

Even more reason for there to be a system.

The system I described was a simplified description of a system used in the early 1970s at a foundry located in Acton and then at Shepherd's Bush.

Regarding my earlier post; from what you have seen do you think the names were applied to the master pattern ?

[Guest Pete Wood]

Regarding my earlier post; from what you have seen do you think the names were applied to the master pattern ?

No. I am sure that the names were always applied to the positive patterns - and not the master.

But let me check my notes and if I'm wrong, I'll let you know.

[Myrtle]

RT

I could be wrong but I was under the impression(Sorry about that  ) that there would have been a main pattern for the plaque and then a pattern for the name alone layered on top. The patterns for the plaques would have been stored and the name patterns would have been discarded. I believe that with some items the layered patterns would have been glued but in this case I don't know.

M

RT

So this could have applied ?

[Guest Pete Wood]

If you think that knowing the actual process will help, please contact me off forum.

[PhilB]

RT:-

Quote:-

"Acton was pumping out approximately 20,000 plaques per week for two years.

Woolwich was pumping out approximately 80,000 plaques per week for two years."

That makes around 10 million plaques! Phil B

[Guest Pete Wood]

Eeek.

Deduct one zero from each total. Sorry.

Acton 2000 plaques per week

Woolwich 8000 plaques per week.

Thanks for the correction

[PhilB]

You`re right, RT. If the name had to be in raised letters, it would virtually have to be cast. Why would they insist on that (after all, all the names on medals were stamped) when they must have known the costs involved? Is it possible that they wanted to keep casting firms in business when they no longer had armament orders? Or other motives? Phil B

[Clive Maier]

On the ordering of names.

Producing a million or so named plaques was a massive task in itself and I can’t see it being further complicated by imposing a particular order of production unless that conferred some real benefit. If there was no undertaking to issue the plaques to relatives in a particular order, I don’t see what that benefit might be. All the factory had to do was check that every name that came in, duly went out again with a plaque. Alphabetical order could help with that but would not be essential if a given individual dealt with small batches of say a hundred.

But let’s suppose again that the list as received by the factory was converted into a job card or works order for each plaque. These orders would be numbered serially and the numbers would provide the means of checking ‘outs’ against ‘ins’. It then would not matter at all that the distribution of names, units and dates was random.

I agree that there would be some sort of logic about the lists arriving at the factory, and that these are likely to be ordered by unit, name and date. These lists are likely to be imperfect for two reasons. First, there would be some ‘stragglers’ - names that only came in later. Second, the lists would probably have been typed – a significantly laborious process. If a clerk missed a name out and noticed it on checking, that name is likely to be added at the bottom rather than retype the whole page. Already order is beginning to break down.

Back at the factory, these imperfect lists are arriving for a variety of units. There are divided between a number of clerks for transcribing onto job cards. The clerks are producing separate piles of job cards, reversing the order if they pile them face up. A production boss gathers the piles into a bundle, and does that before the clerks are finished with a particular list if the factory floor is running out of work. The boss then takes a cut off the top of the bundle and hands it to foreman A, another for foreman B and so on.

This sample scenario shows there are a thousand and one ways in which an initial degree of ordering could be almost completely lost, simply because it was easier and because the ordering did not mean a thing to the factory or the next of kin.

[Myrtle]

On the IWM plaque site they mention:

"A number (possibly an operative's or Ministry of Munitions factory number) was impressed by the animal's right paw."

They also mention that the name was produced by an "ingenious way of casting".

I was wondering if this could be lost wax casting ?

RT

I will contact you off forum with thoughts on process and relevance.

[Clive Maier]

On stamping or coining.

It seems almost certain that stamping or coining must have been considered as a production option. I don’t know whether coining is possible with bronze anyway, but given the size of the plaque, it would certainly have required considerable force. This would demand very strong and costly dies, and probably rules out the option of incorporating the name during the coining operation. The name could have been added afterwards by engraving or stamping, but here we come to what I believe is a philosophical objection.

Such a plaque is a mass-produced item that is customised by engraving. The cast plaque on the other hand is a custom item in its own right. It is unique, a one-off. I sense that nothing less than that was seen as appropriate for the massive sacrifice of a young life. I feel that production difficulties would almost have been embraced as exemplifying this feeling.