Robert Neale's Steam Press

This article was published in the GB Journal Vol61 6/7,  November 2023 - January 2024. No formal response has been received at the time of writing (July 2026)

Introduction

In 1959, Bertrand Russell said that ‘when dealing with intellectual matters, one should ask oneself what are the facts and what is the truth that the facts bear out. Avoid what you wish to believe’ ref1.  I know what you are thinking; ‘End of article then…’. However there are a surprising number of facts available to us, though they are inconveniently scattered. As far as I know, Neale’s steam press was first mentioned in Wright and Creeke (W&C) in 1899 ref2, and E.D. Bacon later made reference to it ref3. More recently there have been several articles published by Bill Barrell ref4, followed by W. de L. M. Messenger in the GB Journal (GBJ) ref5 with numerous follow-up articles and by David Rowse in the London Philatelist (LP) ref6. Many other writers have also made reference.

The aim of this article is to pull together the facts of this story, some of which are unpublished or at least unrecognised and explore the influence that the steam press project had on the stamps that we have in our collections. I also want to point to possible avenues of future research. As is my usual wont, I will be ignoring the wisdom of the wise and will make some suppositions, though I aim to give fair warning when I am doing so.

On 9 January 1855, Neale registered his US patent (Figs.2 and 3). It is interesting to compare the drawing attached to the original 1853 UK patent with the new drawing showing the developments of the design. The machine is clearly longer, with the plates aligned along the direction of travel. There is also a method of twisting the plates within their frame to an oblique angle which may have helped when wiping the plates. Printing plates were very heavy, a rough and ready estimate being around 40 lb (20 kg) each, so reducing the number of plates from three to two would have made the mechanism significantly less bulky.

The main trials undertaken can be split into two by reference to the plates used on the machines. The early trials used die 2 plates 16 onwards and later trials used die 2 plates 39, 41, 50 and 51. In reality these trials were a continuous cycle of test – adjust – test again – followed by a more significant rebuild, then repeat. It is the sheer amount of work and printing that took place over nearly three years that is often not appreciated. This was not just an experiment but was a full-blown commercial development program. PB hoped to gain a technological advantage that could nullify any damaging price reductions imposed by Government, as well as frustrate any potential competition.

On 2 April 1855 PB wrote to the IR the following letter 

“Having secured a patent and prepared the machinery for a new method of copper or steel plate printing from engraved plates, by means of mechanical instead of hand power and which we think will be applicable to, and useful in, printing Postage stamps, we should feel greatly obliged if the Honble. the Commissioners of Inland Revenue will give permission to have it tested by printing one of the Postage plates before registering, in one of the rooms of the postage department of our premises. We propose to test it with a paper altogether different in appearance from that which is in actual use, and all the impressions taken from the plate can be kept in charge of, or destroyed by, your officer. As several days may be consumed in bringing new and complicated machinery to accomplish that which has never yet been done and the machine is in the drying room over that in which your officer is stationed perhaps you would be better satisfied to appoint a special officer to superintend the trial, in which case we are quite ready to bear the expense.

...

P.S.—We expect to be ready for trial on Thursday morning next.” Ref7 

There are several points of interest here, but one that is often overlooked is the fact that the operation is now based at Fleet Street, almost certainly for security reasons. Having a boiler in the drying room may also have helped heat the room. It is also likely that for convenience a secure strongbox to house the plates in use was installed, along with a large cupboard to house the paper. Trials were probably ongoing during 1855 from early April through to at least August, as we have the well-known Chase National Bank sheet from plate 22 being dated 15 August 1855 as an example (Fig.4).

On 31 October 1855 PB cancelled the order for machines 4 and 5 before the first main trial had even finished. Clearly they had had enough experience with the new machines to realise that all was not well. Matters would not have been helped when Neale asked to show off the machine to the agent of a third party, Harper &Co ref8.

On 28 November 1855 Neale wrote to PB asking for ink & men for 3 days. This was followed by a letter on 7 December 1855 from PB stating that the trial had been quite unsatisfactory, and asking how he was going to complete the machine. It was at this point that I believe that JBB drew an imaginary line under Neale and decided to go on without him. It is likely that soon after this all plates and the stock of paper were cleared out from the dual control cupboard and transferred to the main print room store ref8.

On 18 February 1856 PB wrote to the IR 

“We have the honor to acknowledge the receipt of your letter of the I5th inst. stating that as the experiments upon the new machine for printing the Postage Stamps had been brought to a conclusion, you are directed by the Board to call upon us to pay to the Receiver-General the sum of £33 14s. 0d. being for 674 hours of time your officer was in attendance in connection with that business.

We beg to state that we have paid to the Receiver-General the sum of £33 14s. 0d. as above for which we have his receipt. You are under an erroneous impression however, as to the experiments having been concluded we having informed your officer Mr. Peacock, that such was not the case when he put the question to us about three weeks back. It is true, however, that the inventor after having made great progress has broken down in his attempt to perfect the machine, but as we have such a large sum upon the invention, and a little additional time and expense will enable us to ascertain whether we cannot take it up where he has left it and thus bring it to perfection, we shall feel greatly obliged if your Honble. Board will allow us to have an officer again in a few days and for a few days in order to test the improvements it is now undergoing. The expense of his attendance to be paid by us as before. Thanking you for this favor and requesting that permission may be continued for a short time longer.” ref9. 

Emphasis mine, as 674 hours at 4 sheets per minute equates to over 160,000 sheets, although I do not doubt that a good deal of this time was taken up extracting mangled sheets from the mechanism. Rowse estimates that the bill of £33 amounts to just over 84 days’ work spread over a nine-month period. PB were working on this project full time. This letter heralds the start of the late trials which took place approximately between February and July 1856.

On 22 August 1856 JBB writes to Neale ref10 formally withdrawing from their agreement. There followed litigation the result of which we do not presently know. The project had been taken as far as it could without practical success, which must have been a huge disappointment. David Rowse’s comment “... his [Neale’s] solution was fundamentally flawed. He sought to move a heavy plate to the ink & paper rather than the other way round.” is undoubtedly correct.

Were the stamps printed by the steam press issued?

Some commentators (particularly Messenger) have perhaps understandably assumed stamps from the steam press were issued to the public, but I have serious doubts that this was the case. All of the known examples that we have are on large crown watermark, which was introduced around the end of the first stage of trials. All of these show marked blurs where there should not be ink, and bald patches where there should. The quality is just not good enough. Even the souvenir sheet from plate 22 is not free from defects, and surely this sheet would have been picked out because it was one of the better ones. I do not believe that SH would accept steam press-printed against issued warrants, since why would they pay for a demonstrably inferior product? It was far simpler to keep the product of the steam press and its accounting separate, at least until such time as the quality came up to that already being obtained by using hand presses.

Many of the plates that were used on Neale’s machines can be said to be uncommon to scarce, which could indicate that their working lives had been significantly reduced by their use in the programme. Plate 22 is estimated to have only produced 7,000 sheets that were issued to the public. Even taking into account the poor layout of this plate, I think there would be many more sheets issued if the product of the steam press went into normal stock.

No examples are known from the later main 1856 trials, as earliest known uses from the four plates used in those trials are not found until 1861 on un-blued paper, well after the steam press experiments had ended. If any examples from the later trials were to come to light, they would be red-brown on blue paper and in the case of plate 51, in state 1.

Colour

It may be tempting to point to the famous Chase National Bank sheet from plate 22 and suggest that steam press stamps all have the same deep brown-orange colour. However this would be to ignore the fact that the trials took place over such a long time period. It is reasonable to assume that stocks of ink currently in use at Fleet Street would have been used for the steam press and as a result would have varied over time as we see in the standard stamp issue. The stamp lettered PH from plate 18 in Fig.5 shows a deeper red-brown colour, with only a hint of orange. 

There is an additional complication however that needs to be considered and that is the speed at which the ink dries.  On the hand presses, the printers relied on two factors to ensure quick drying; direct heat and ink composition. The hand presses in use at Fleet Street had a hole in the bed of the press immediately under the printing plate through which a gas burner heated the plate, which started the drying process on that part of the ink that was soon to be exposed to the air once the paper had been removed from the plate. Neal's presses did not have this feature as the plates were fixed in a frame and travelling at speed around the machine. 

The drying factors dependant on ink composition are twofold. Firstly, normally the ink was spread onto the plate as a thick paste. Too much oil would make spreading easier but would extend the drying time significantly. A very thick mix is preferable, but that presents problems when you want to mechanise the process. A viscous ink would require more mechanical force to ink and then wipe the plates correctly, which in turn could have been a major factor leading to plate wear. 

The second factor was the addition to the ink of a drying agent. In the case of the issued stamps this took the form of Prussiate of Potash (PoP). None of the known examples from the steam press show any bluing, so PB must have used a different ingredient as a dryer. Red lead was a widely available pigment that also acted as a drying agent, and being relatively cheap would have been an attractive alternative. The boiled linseed oil that was used to dilute the ink is itself a natural dryer, but other resin- or metal-based additives could have been used to speed up the process. 

It is perhaps more than a coincidence that soon after the later steam press trials we start to see changes in colour found in the issued C8 stamps (‘C’ numbers are from the SG Specialised Catalogue), which, with the absence of PoP as a bluing agent turns into the C9 range of shades. How much of the knowledge learnt at the steam press found its way into the issued stamps? 

Whilst writing this, it has occurred to me that the issue of drying agents is probably a general factor in the shade of the issued stamps. Stamps of this period printed during the winter months are often found in dark red-brown shades with lots of bluing. This may indicate that more drying agent (PoP) was added when the weather was cold to help the ink dry quickly. If and when the spectroscopic study of ink components restarts, it would be interesting to see whether the mix of ingredients changed over the months as well as how the steam press examples relate to the standard issued stamps.

Paper

As we have seen, the main trials started “with a paper altogether different in appearance from that which is in actual use”. Possibly coloured, certainly without a crown watermark. However, very soon the call would have gone out that in order to have a proper feasibility trial, the actual paper then in use would need to be tried. There is no written document to prove that this happened but we do know that all known examples are on large crown watermarked paper. Moving the operation to Fleet Street would have solved any security concerns, as the IR officer was on hand to account for the plates and sheets of paper. As PB were already paying for the officer’s time it is reasonable to assume that paying for the paper at cost price would have been an acceptable addition to the costs of the project. There is an interesting letter in Bacon (appendix C 229) that discusses paper storage facilities just after the fire. A tall cupboard about five feet wide and two feet deep could have held 40,000 sheets.

I believe that W&C are underrated as a source of information. Bacon, in particular tends to question the veracity of the information found in W&C, often stating that he does not know the source of the information they quote. I have found that much of the information found in W&C ties in well with that found in IR79/79 ref11. It would seem that the person who prepared that register was either their original contact, or was drawing on the same source for this information. Bacon, who for obvious reasons had excellent contacts within the printing firm, would not necessarily have had a contact within the IR and so may have been unaware of the source. IR79/79 itself seems to be an attempt to pull together a record of dies, plates, rollers and the suchlike and probably dates from just post-1857, the timing of which would make sense. The source of the information found in this register would have come from within the IR itself, that is from the official ledgers. Ledgers found within any Government department are likely to be painfully accurate, so we ignore IR79/79 and by extension W&C at our peril. 

Plates 16-20 were registered on 12 May 1855. I believe that the put-to-press date of 15 May 1855 for plates 16 and 17 as recorded by W&C is the date that Neale’s machine, equipped with those plates, went live with small crown watermarked paper. If plain paper had been used there would have been no reason to record the fact within the IR ledgers. One does not account for printer’s waste on plain paper, but one does account most carefully for official watermarked paper that incurs a £5-per-sheet fine if found to be missing.

Paper movement was controlled by Somerset House (SH). They worked on a ‘First In First Out’ (FIFO) basis, and drip fed paper to PB when needed using the warrant system. If SH was the only source of paper, then there should have been a sharp cut-off seen in the issued stamps between small and large crown paper. The exact change-over date is not known but is believed to have occurred early-to-mid-June 1855, with the earliest known use of the large crown paper being 7 July 1855 ref12. After the warrant issued with the last of the small crown paper there would have been only large crown paper in stock. Except for that stock of small crown paper in the steam press paper cupboard. As time went on and more trials took place, large crown paper started to be supplied and would have been placed on top of the older paper stock. It was not a large strongroom with space to arrange the stock, and it would have been a large cupboard. The only consideration was that there was enough stock of paper to allow work to carry on without pause. We as philatelists place great importance on the difference between small and large crown watermarks. To the staff then at work, that difference was totally immaterial. This stock of paper with small crown watermark had effectively fallen off the FIFO wagon.

As has been described, after the unsuccessful trials which finished at the start of December 1855 the  strongbox and paper cupboard would have been cleared out. The IR official would have had a record of the sheets which were available for use with the next warrant or two issued, with just the balance of sheets needed to come from SH. I believe that this is probably the source of the batch of small crown paper that gave us the later C3s and all of the C7s at the end of 1855. The latest (2020) SG Specialised Vol 1, Part 1 (pp.259 and 279) gives estimates of sheets printed for C3 plates 16-21 and the C7 issues which total about 20,000 sheets. Which at first glance seems a lot. However, based on the 2,000 sheets printed on 15 October, this amounts only to 10 days’ printing on the machine(s?).

I have not conducted a study accounting for paper against warrants for this period as Alan Druce has for the early 1840s. Alan told me that the way the details were recorded changed after the period he was looking at; however, there may be some further information to be found in the archive somewhere which could provide a fruitful avenue for future research.

What Plates were used for the trials?

Dates are important here, so I have pulled together some for ease of reference (Table 1). The source for these dates is Allan Oliver’s excellent ‘Lifespan’ ref13, as well as the engraving book as transcribed by De Worms ref14.

Plates - Early main trials

According to W&C, the early main trials covered plates 16, 17, 18 and 22. (Fig.6) Messenger pointed out in his GB Journal article ref5 that plate 21 was also a possible candidate due to the long delay between its put-to-press date and its earliest known use. He also mentioned plate 25 as a possible as it was ‘said to have been used’ probably due to examples that match stamps known to have been printed on the steam press.

Plate 16 is unusual in that it was initially only partially engraved and lacked the surrounding legend. This was added a month later and I assume that the plate was hardened only at that later point. W&C state that this plate with plate 17 was put to the steam press on 15 May 1855, that is three days after the imprimatur sheet was signed. This seems to contradict PB’s letter of 2 April which stated “...give permission to have it tested by printing one of the Postage plates before registering,…” ref7 However, plate 16 was substantially complete on 4th April and was available for testing perhaps using plain unwatermarked paper, but there is a problem. It is only one plate, so I do wonder if the machine could have been run at anywhere near full speed, as the whole set up would have been out of balance. Plate 17 was completed on 18 April 1855 which would have allowed for a balanced run. Between these two dates, I would suggest that there was plenty to keep the engineers occupied testing things like clearances, checking that each part of the apparatus worked as it should when hand cranked and suchlike. No examples from these two plates are known to me.

Plate 18 was completed just after Plate 17 and joined the two other plates, presumably as a stand-by. It is recorded as being put to steam press on 1 June 1855, although we do not know which plate came off. Bill Barrell’s block from this plate (Fig.7) clearly shows some of the problems of Neale’s machine. Any slight depression on the plate collects ink which is not removed by the mechanical wiping rollers. There is a marked tendency for ink to be left in the vertical margins as well as across “ONE P” of the value. To some extent these features are common to all known steam press examples. This shows just how difficult it was to try to replicate the palm of a skilled intaglio printer. Bill’s block is well inked, and there does not appear to be any un-inked areas showing. There is damage to the paper, but it is not clear if this occurred during printing or after it. It should have been destroyed but clearly was not. 

At least there can be no doubt that Plate 22 was one of the plates used on the steam press, as we have the Chase National Bank sheet dated 5 August 1855. According to the handwritten legend it was one of 2000 sheets printed on that day, although it does not tell us if there were one or two machines at work. At 4 sheets per minute, that would equate to a little over eight hours running time for one machine, or four hours for two. This sheet, not being printer’s waste but a gift to visitors would have had to have been paid for, presumably for the then substantial sum of £1. 

I am of the opinion that plate 25 was used on the steam press. The reason for my certainty is the existence of an imperforate half sheet (now split) from this plate that is clearly printer’s waste. This half sheet displays the normal features we would expect from the steam press. Stamps are without gum or bluing and exhibit blurs in the vertical margins and across ‘ONE P’, and areas that lack ink. On this particular half sheet there are also vertical wrinkles that caused unprinted areas and are also a probable cause of the sheet tearing as can be seen in figs 10 and 11. All of which points to the fact that the sheet was printed by machine. A printer working by hand would have a real challenge to replicate such damage! There is a fine block in the Phillips collection lettered IA-JD (Fig.9) which illustrates what I am trying to describe. Phillips thought that this block was from the transitional period due to the orange-brown colour and the fact that there is no bluing of the paper.  Fig.10 shows the adjoining pair KA-KB which beautifully illustrates the practical problems associated with the steam press. The K row from this half sheet is the lowest known with no examples known from the rows below. The back of the K row is illustrated (Fig.11) which shows the top half of the hand written word ‘Specimen’, which probably goes some way to explain its existence. 

Were plates 19-21, 23-24, and 26-27 used on the steam press? All these plates were available for use in the early trials if needed, though there is no evidence that I am aware of that would indicate that they were. You could argue that, for example, plate 26 had a low print run which may indicate some use on the steam press, especially as the number of sheets printed almost match those printed from Plate 25, but this is no more than circumstantial evidence.

Plates – Later main trials

The plates used for the later main trials are far easier to pin down as we have evidence within the various archives. There are four plates that appear to form two natural pairs of plates which corresponds to there being two machines in use. Stamps from these plates are not known used before 1861 and seem to have been brought into use on the hand presses to extend the period that PB were printing penny stars. They wanted to get as much out of all existing plates that they could before they were forced into printing the penny plate series. These four plates did not all escape the ravages of being on a steam press unscathed as some repairs were needed before going to the hand presses.

In the register of plates IR79/79 ref11 is a note next to plates 39 and 41, “For use at steam press”. (Fig.12). Plates 37-44 were registered on 2 April 1856, with these two plates perhaps being picked because they were of a similar weight. The trial using these plates began at the start of April 1856 and continued for some weeks, as at this time we see in the archive Neale kicking up a fuss and asking for the return of his first machine from PB, but a short time later offering to work at £4 per week for three months. Examples from these two plates do not show any significant wear and I am unaware of any second states, though my knowledge of C10s is limited. It could be that these plates were used on the steam press until plates 50 and 51 were available for use, though this assumes that only one machine of the two available was in use at any one time.

PB sent the following letter to Ormond Hill on 23 June 1856 

“We have now 7 Postage Stamp plates ready for registering and as 2 of them are of a new size adapted to printing by machine and the machine is now ready for what we hope will be a conclusive trial; we shall feel greatly obliged if your Honble Board will appoint as early a time as convenient for the registration in order that we may have it tested.” ref15

The two plates evidently of a new size were plates 50 and 51. These two plates are interesting for multiple reasons. Size of the plates is an obvious point but looking at the plate on the flat, there clearly cannot be more than a shaving from the overall size of that plate. They had to have been of a different thickness. A standard plate started its life at 5/8 inch thick. Successive defacements, grinding smooth and re-use would have brought this down slightly over time, but not appreciably. At a later date PB made the halfpenny plates only 3/16 inch thick due to the larger size of the plates (Bacon p.196), but these were slightly flexible and had to be screwed down to the press bed when in use. If they had made plates 50 and 51 say 3/8 inch thick this would have made a significant difference in weight, yet still retain rigidity, though this is little more than guesswork.

Plates 50 and 51 went to the steam press near the end of June 1856. When plate 51 eventually came to be used on the hand press in 1861, many impressions show re-cut sides and baselines which do not appear on the imprimatur. It was suggested by Dr Osborne in a letter to Dr Wiggins in1955 ref16 that these repairs were necessary and probably due to corrosion which occurred during storage. However, this ignores the fact that these plates had been used on the steam press, which at the time the letter was written may not have been appreciated. Mike Williams and others have echoed Dr Osborne’s suggestion ref17 that corrosion was a direct result of the fire of 1857, though I do not agree with this viewpoint. Fire damage could take the form of adjusting the temper of the case hardening of a plate, depending on whether it were doused with water completely or left to cool slowly, but that would not cause corrosion. It is feasible that water could have puddled on the cool plate surface just after the fire, but it would not have been there for long enough to cause significant damage to the metal. I believe a more probable reason for the necessity for repair was the harsh mechanical wear caused by the plates being used on the steam press over a significant time. 

Postscript

I cannot with a clear conscience call this a conclusion as I hope that this story will continue to develop. There has to be more facts that will emerge to fill in some of the blanks and as Bertrand Russell mentions, more truths that will be borne out from those facts revealed.

I hope that this article has highlighted certain points which are not currently widely acknowledged. Firstly, the scope of this trial was massive. It was much more than an experiment but was a significant production development project that spanned many years. A significant amount of time and money was expended in trying to perfect this method of printing. Secondly, these trials were the source of the batch of sheets printed on small crown paper that was produced by PB at the end of 1855. They may also have had an influence on the colour of the issued stamps.  Thirdly, plates used were subjected to significant wear which is evidenced by issued stamps and which reduced their working life. 

Supposition alert! How much of a coincidence is it that the 1857 Fleet Street fire badly affected the area for postage stamp production, including the drying room which happened to house the two (useless) steam press machines. Had the (valuable) boiler already been disposed of? Were those two machines (which still had a significant book value) covered by the insurance policy? I am not even going to mention Prussiate of Potash. Fact: did you know that rags soaked in boiled linseed oil can spontaneously combust when exposed to air. Nasty stuff, one needs to be careful. 

And what of Neale? Robert Neale continued with his idea long after the unsuccessful attempt with PB. In 1881 he took out a patent in America (No.245,970) for a flatbed cylinder power press for printing banknotes. I wonder how far he got with it.