This database had its origins in my research for a PhD at Exeter University on the diffusion of power technology in Britain. John Robey of Moorland Publishing, author and publisher of many books on mining and local history, had been compiling a card index of individual steam engines built by 1800, to meet the challenge set by John Harris in his tentative assessment published in History in 1967, which had suggested that a possible total of 1,330 might be conservative. We were put in touch and agreed to collaborate, since we both believed that it was possible to generate a statistical or factual base from which to assess the extent to which steam power was taken up by the 18th-century economy
John had other calls on his time so I took on this work, combining and extending John's card index with my own literature search and entering all the details on 80 character IBM punch cards. This was at the time the best information technology available, and it made feasible the cross-referencing, checking and verifying the data and analysing the numbers according to various traits and criteria. We were able to log over 2,000 engines over the century and publish the results in an article in Technology and Culture in 1980; I also included a copy of the full computer printout in my 1979 PhD.
There the matter stood for many years, as we both had other pressures on our time. John bowed out of further involvement, but I did try to keep track of the main publications which identified further additional engines or enabled double counting and other mistakes to be corrected, especially the work of Jennifer Tann on the Boulton and Watt archives, but also many monographs and articles with a more local focus.
Nowadays powerful database and spreadsheet software, beyond the wildest dreams of the 1970s, is commonplace, so it was possible to scan the prinouts of the 70s database into an Excel spreadsheet (for which my thanks to Alessandro Nuvolari), following which I turned the number pair codes into text so it was more easily analysed, and added detailed references, which had previously had to be written longhand onto the back of the cards because of the limits of the punch card system, and did some limited editing - correcting spellings, location details etc.
Once this was done, on my retirement it was possible to update and extend the listings, involving much studying of research published since the 1970s; additional archives research (principally on newspapers and insurance records); and using the power of the internet. In a relatively short time work it proved possible to complete work which would previously have taken years of research and miles of travelling around the country to achieve, and this has been added to in recent years.
The original data was logged under as many categories as could be captured on the cards - date, location, type of engine, county, industry, purpose (function), power, size and other useful information.
In many if not most cases, however, several components of this information were not available, and for many engines we have only a location and the industry in which it was used; sometimes even these are not known.
In recording these data, arbitrary decisions as to which of two conflicting pieces of information were to be used were sometimes necessary. In the main, however, the criteria involved in assigning dates, locations, and the like were self-dictating, governed by the nature of the sources. In over a third of the records (38%) the date given is the one when the engine was first reported, the starting date being unknown; for early colliery engines this date could often be many years after construction, for later factory engines a year or two is more usual. Moreover, in some 40 cases an approximate date is recorded in the absence of better data.
For the same reasons, while secondhand engines were in the original list for checking purposes, they are not included in this database. If every move of the materials of an engine were to be counted as a new installation, a very distorted picture would be drawn of certain counties, most notably Cornwall. On the other hand, of course, some of the engines recorded as being new in the absence of information to the contrary were doubtless secondhand, but this is inevitable when reliance must be placed on incomplete data.
A further difficulty arises in distinguishing between secondhand engines and major rebuilds. A new engine in an old house was always counted separately. In some cases, where substantial modifications were made to an existing engine, as for instance Smeaton's rebuild of the Wheal Busy engine, this has been included in the listing. Where there was detail to show that the installation was in essence a move, however, as for instance when the cylinder and other key components of an existing engine was used with new timberwork on a different site, this was generally regarded as a secondhand engine and was not counted, unless there were clear reasons to include. This approach is not without problems, since in the 18th century almost all engines were built up on site from materials assembled from many places, and it is usually impossible to tell their origin. When the timberwork of an old engine was used with a new cylinder on a different site, should this count as a new or a secondhand engine? In the final analysis each case had to be examined and a decision made. Though some of these may be open to objection, the policy adopted is probably the least unsatisfactory overall. Local knowledge is invaluable here, and has led to several deletions and corrections; if knowledgeable users are able to correct errors, this would be invaluable.
For location purposes the counties as recorded as they were before the local government reorganization of 1974, putting modern conurbations in their respective counties apart from London, where to assist analysis engines are located to their parish or locale rather than merely "Surrey" or "Middlesex". By and large there were, however, few problems in assigning engines to their correct counties, apart from a significant number of early Coalbrookdale cylinders and some later Carron ones, for which we have only the name of the purchaser, and some engines on the borders of counties, e.g. Leicester and South Derbyshire and the Black Country, but OS maps and local knowledge mostly resolved these issues.
Before 1776 it is simple to categorise the type of engine - apart from a small number of Savery pumps, all were Newcomen engines with a cylinder, beam and pumping apparatus. The entry of the Watt separate condenser meant it was necessary to add a new category, and from 1780 increasing numbers of both B&W and other makers' engines began to be applied to rotary motion, so thereafter engines are identified if known under whether they were rotary or pumping. Piracies of Watt's condenser patents and other designs which began to be used from this period, such as Hornblower's compound engine and Bull's beamless inverted ones, are also noted when known.
For most engines, recording the industry in which it was set to work and the job it performed are realtively straightforward, although some engines worked machinery in different industries at the same time (e.g. a textile and flour mill in differnent buildings at the same site; or an engine at a mine which both pumped water and would material from the mine. Where possible, explanatory information is included in the "comments" column of the secondary industry or purpose.
Recording the maker of an engine sets problems similar to those encountered with rebuilt and moved plant. Most Newcomen (and, indeed, Boulton and Watt) engines were not really built by any one person or works but were rather the product of several different concerns who supplied parts (cylinder, gearing, wooden parts, boilers, etc.) and an errector who assembled the parts, had them put toghter and made the whoe work (often not an easy task). Where the principal designer is known, that was recorded; otherwise data of the factory or erector, if that was relevant. For over two thirds of these early engines, however, we have no details of the designer, suppliers of parts or the erector.
Likewise, the power and size of the cylinder are not usually recorded, and we have details of dimensions of under half of engines other than Boulton and Watt's production. The horsepower is in any case of limited and contested value - few reliable records of the net power delivered to the pumps or machinery survive, except for some Watt engines and a scattering of others. For Watt engines and some others we have the nominal horsepower, but this is a calculation based on dimensions and declared rate of working rather than a measurement, and is affected by the intensity of the work done - for example, a pumping engine might work all day and night slowly, or at full capacity over restricted hours, depending on the needs of the site and the personnel available.
We do have much useful data on who made the cylinders, the most important part of all these engines. Early engines had small brass cylinders, but from the early 1720s until 1748 and for a scattering of later years and areas we have good information on the cast iron ones made by the Coalbrookdale Company, and good but not complete records of the Carron Company, Coalbrookdale's main competitors from the mid-1760s.
The biggest gap in all these information points is the huge number of small engines erected in textile and other factories from the late 1780s onwards. At this time there was a massive expansion of iron foundries building these small engines and apart from Bateman and Sherratt in Manchester we have little information on these unless they fell foul of Bouton and Watt's pursuit of "piracies" of their separate condenser and other patents.
All in all, however, partial and patchy information is better than no information, and I hope this database will be useful to researchers in all disciplines.