by Ben Russell
Working brass and glass into precision instruments led to innovation in three key areas. The first machine that instrument makers needed was the lathe, designed to hold a piece of material and rotate it, so that different cutting tools could cut and shape it as it turned.34 The lathe ranged in size from the tiny watchmaker’s ‘turns’, small enough to be held in a vice and hand-spun to make watch shafts and spindles, to the ‘great lathe’, a bigger workshop machine powered by an assistant turning a hand capstan or working a treadle. The precision engineer Richard Roberts employed his wife for this duty, working in a bedroom while his wife drove the lathe from the basement.35 Over the course of the century, the lathe went from being largely constructed of timber, and used to work wood, horn or ivory, to being a more robust iron-framed machine capable of standing up to the rigours of working metal. It has been described as the first engineer’s machine – both because it was the first machine that could make the parts of other machines and because of its versatility: instrument components could be turned, bored and finished on it; glass lenses could be polished – and screw threads could be made.36
Screw threads – cylinders with an indented spiral running around the outside – were essential to successful precision instruments. They did not just hold components together, but were a valuable means of measuring: if a screw is an inch long and consists of twenty threads, rotating it ten times will move it axially by half an inch. Knowing the number of threads on a screw, and making them as fine as possible, meant screws presented new means of accurately measuring tiny angles and distances. At first they required highly skilled hand work. Exactly measured diagonal lines called transversals were drawn on a strip of paper, their length matching the circumference of the thread needed. Then the paper was glued onto the rod and, if done carefully, the ends of the diagonal lines linked up to form a spiral along its length. Using the spiral line for guidance, a narrow, pointed hand file started the thread. This was deepened with a larger triangular-section file, and then it was finished with a chaser, a tool with three or more teeth of the profile needed for the completed thread, which was pressed into the thread as it rotated.37 London instrument maker Jesse Ramsden later built a number of special screw-cutting lathes, and their output was used to make fine adjustments in all sorts of instruments – some later threads were employed in measuring devices claiming an accuracy of within one-millionth of an inch.
Another use of screw threads was to make, or ‘divide’, graduated scales. Navigation instruments needed to be smaller and lighter, but the barrier preventing this was being able to produce a small enough scale by hand. This entailed securing the scale on a ‘dividing plate’, calculating the position of each graduation by geometrical means using compasses and then marking each with a steel blade before polishing to produce a clean finish. The work was so delicate that some London makers kept their tools and the scales at a near-constant temperature to avoid distortions caused by thermal expansion, only working in early mornings during the spring and autumn.38 And, as he had devised ways of cutting screw threads, between 1768 and 1774 Ramsden also cracked the problem of making scales by machine. He designed and constructed a ‘dividing engine’ (see page 59): the scale to be graduated was fixed to a large horizontal wheel with teeth cut around its circumference, which engaged with a finely cut screw thread that, when rotated, turned the wheel through minutely controlled angles. The engine was straightforward to operate, and Ramsden’s apprentices, and sometimes his wife, used it.39 A complete, immensely accurate scale could now be made by relatively unskilled labour, not in a matter of days, but in half an hour.40
Making the best use of these new materials and techniques determined the organization of the instrument trade.41 It was structured into two broad groups: mathematical and optical instrument makers.42 A company might position itself in either one of these areas and specialize in making a particular type of instrument. For example, James Short of London made reflecting telescopes only and John Cuff made only microscopes. Many confined themselves to making only the most critical parts of each instrument, putting the rest out to subcontractors. As Campbell noted, the optical instrument maker ‘himself executes very little of the Work, except the grinding the Glasses . . . The Cases and Machinery of his Instruments are made by different Workmen, according to their Nature, and he adjusts the Glasses to them.’43 Some makers – ‘out of doors’ workmen, ‘Chamber masters’ and ‘Garret masters’ in London – worked in their own homes with a small number of employees, selling products to the public retail outlets of Martin, Adams, Dollond, Ramsden and others.44 In Sheffield ‘Little Mesters’ rented space in a factory and sold their output to its owner.45 And finally, within each company each individual might have a particular specialism: grinding lenses, making instrument frames or finishing and lacquering.46
A consequence of this complex trade structure was that the name engraved on a finished instrument was a ‘brand’ name – that of a businessman and designer in overall charge of the manufacturing process but responsible for physically making few of the components. Such a man was required ‘to have a pretty good Education, and a penetrating Judgment . . . and must be a thorough Judge of such Work as he employs others to execute’.47 What was the ‘Watt’ brand? And how did Watt organize and equip his business to make it?
Far from being a solitary craftsman, Watt assembled a team of employees. In February 1758 he was looking for ‘any lad that can file tolerably well’ and thereafter the workforce expanded to a maximum of fourteen trained workmen and apprentices.48 Only one of them, John Gardner, was with Watt for the long haul – in fact, he became Watt’s right-hand man and took over the instrument making business in about 1770. The others worked as required on a variety of different projects. The men had their own areas of specialization: Alex Gardner worked on lenses and James Couper mainly made quadrants, for example – but the range of output demanded meant they often turned their hands to anything required. Robert Allan, for example, was employed making parts for quadrants in 1760, but in the following two years was making violin bows and guitars.49
Having developed his team, Watt also remained closely linked to the instrument making trade in London. He was well acquainted with Jesse Ramsden: he had his business correspondence sent to Ramsden’s shop and they also corresponded about a telescope Watt developed, which survives in his workshop, that could be used to measure distances. The close links between Ramsden and Watt endured: Watt’s later colleague John Southern was present as Ramsden was taken to be buried in December 1800, and he wrote, ‘My feelings were considerably hurt this morning at the sight of poor Mr Ramsden’s coffin, in which he was descending the stairs for the last time, just as I called in.’50 But in technical terms there was a gulf between the world of making instruments in London and Glasgow, and explaining this requires consideration of the materials and techniques Watt used in instrument making.
Jesse Ramsden alongside his dividing engine, J. Jones after Robert Home, 1790.
Both brass and glass can be found in quantities in Watt’s workshop, and references to many other materials can be traced in his ledgers. The deep brown of cocoa-wood from the West Indies made it popular for flutes.51 Black ebony from India or Mauritius, and mahogany from the Caribbean, tended to warp and twist less than other woods, and were often used for the frames of scientific instruments, with graduated scales to take measurements added in brass. Later brass replaced wood entirely. Green ebony and boxwood could be made into rulers or turned in a lathe.52 Plane wood (called ‘plaintree’ by Watt) was used in musical instruments, while John Williams of Newcastle provided Watt with ‘tubes and cylinders of glass’.53 Steel was used in tiny quantities for pinions and pivots, which had to withstand considerable loads – and Watt squirrelled away broken table-knives as a useful source of tool steel.
To work these materials, Watt used hand tools extensively and drawer upon drawer of them remain in his workshop. He would have been familiar with them from his training in London, and a list
he sent his father prior to travelling from London to Scotland in June 1756 gives the things which he deemed ‘absolutely necessary’.54 It includes among much else a ‘standing vice’ – probably a leg vice to fit on the side of a workbench – hammers and anvils in different sizes, files for shaping metal, cast and plate brass, screw plates for cutting threads on metal rods, dividers and compasses, an iron brace and bits for drilling holes, a selection of hardwoods, tortoiseshell and mother-of-pearl for instrument making and even ‘an alphabet of letters’ – letter stamps for marking wood or metal. By October 1759 the tool list had expanded considerably to include a ‘great turning lathe’, as well as draw plates for making metal wire of different diameters, and ‘patterns for casting brass’.55 Later the noted toolmaker John Wyke of Liverpool supplied a clockmaker’s vices and files, screw plates, dividers and more, and the workshop devoured a huge range of other consumables, from half a hundred-weight of emery powder for polishing to cat gut, copper wire and varnishing brushes.56
Many of Watt’s tools can be matched against contemporary descriptions of workshop equipment. Some, for example, can be identified in Joseph Moxon’s Mechanick Exercises, the grandfather of all modern DIY books.57 Others can be spotted in the details of the sale arranged when the watchmaker Larcum Kendall died in 1790.58 Many of the basic tools are similar, and these are a counterbalance to the emphasis on mechanization, lathes and dividing engines in instrument making. But among them are Kendall’s ‘curious’ tools – ‘a box containing curious models’, ‘a curious deepening tool’, ‘a curious engine for cutting spirals’ and even ‘a very curious unique engine for cutting horizontal wheels’. These are the special-purpose tools designed to give Kendall a technical edge over his competitors. What were Watt’s equivalent ‘curious’ tools, and how successful were they?
Prominent among Watt’s early specialized tools were printing plates for making the scales on quadrants and other instruments. Two plates for printing the scales for barometers on paper still exist in Watt’s workshop. One has two scales depending on the temperature: ‘Dry/Very Dry; Set Fair/Set Frost; Fair/Frost; Changeable; Rain/Snow; Much Rain/Much Snow; Stormy’, which is suggestive of the ferocious Glaswegian weather, about which Watt could describe to a friend the ‘100 people I see just now running by wett to the skin, no doubt cursing god in their hearts. I believe the [rain] drops are an inch in diameter.’59 Alongside these plates there is also a beam compass with interchangeable heads, for making scales on instruments. One of the heads consists of a small ‘comb’ with eleven sharp points on it that, pressed against a brass sheet and dragged in a radius across it, would scribe eleven lines at once – much quicker than making each individually. That Watt was making barometers with printed paper scales, rather than engraved brass ones, and speeding up production by scribing multiple lines at once, suggests that he was making a cheaper product for the lower end of the market.
An instrument makers’ workshop, 1849. Although of a later date, much of what can be seen would have been very recognizable to Watt.
It may also be that these more ingenious techniques are not as novel as they appear at first sight. For instance, in Moxon’s Mechanick Exercises there are details of a method of ‘laying moldings either upon metal, or wood, without fitting the work in a lathe’.60 This called for the use of a ‘strong iron bar’ – a beam compass – with a centre-pin on one end, and fitted with ‘a tooth of steel with such roundings and hollows in the bottom of it, as I intended to have hollows and roundings upon my work’ on the other. With the point placed in the centre of the workpiece, he
employ’d a labourer, directing him in his left hand to hold the head of the center-pin, and with his right hand to draw about the beam and tooth, which (according to the strength) he us’d, cut and tore away great flakes of the metal, till it receiv’d the whole and perfect form the tooth would make; which was as compleat a molding as any skillful turner could have laid upon it.
Watt’s scriber for making multiple lines at once is a simpler form of what Moxon describes here.
Historian of mechanism Michael Wright has even suggested that one of Watt’s projects appears to have gone subtly wrong. Resting on one of the shelves in Watt’s workshop sits a wheel. It is cast brass and 15 inches in diameter.61 Shallow teeth have been cut around the circumference, indicating that it is a dividing plate, a small version of those made by Ramsden in London and used to make the scales for instruments. But it hasn’t been used much, and some detective work shows us why. The teeth around the outside would have been cut by placing the edge hard up against a worm screw whose leading edge was nicked to make it cut into the metal. As the screw was rotated it would have turned the plate through 360 degrees, cutting teeth all the way round as it progressed – three or four complete rotations would have been needed to make the teeth deep enough. Because the accuracy of any instruments divided on the plate depended on the positioning of the teeth and hence on the accuracy with which it was rotated, a careful instrument maker would painstakingly check the position of the teeth, using a magnifying lens, against marks determined by geometrical means – and there are some marks made with compasses, showing some attempt at carefully laying the whole thing out. But in making the plate, Watt didn’t check against these marks, and though there are supposed to be 360 around the circumference – one per degree of rotation – there are only 359. Because of this carelessness, any instrument Watt made using the plate would have been correspondingly inaccurate.
If Watt’s attempt at constructing a dividing engine went awry, he had more success making screw threads. He wrote to his friend William Small in 1773 that ‘My dividing screw can divide an inch into 1,000 tolerably equal and distinct parts on glass’ – meaning he was using it to mark the scales on glass thermometer tubes.62 A year later, he wrote again to Small, ‘I had occasion to use my last dividing screw for the first time the other day . . . It did not err the 1/200th of an inch in the whole 9 inches.’63 Watt could do fine work if he put his mind to it, but a question mark rests over whether the bulk of his work actually required him to do so.
The London instrument makers were an acknowledged elite and their capabilities do not necessarily reflect those elsewhere. This is not to denigrate those makers based outside London, like Watt. But precision was subjective: what might be an impressively accurate work of precision to a shopkeeper buying an instrument to amuse his family might be sniffed at by a dilettante used to the output of the very best London workshops. The standards of precision the London makers achieved were something for the others, including Watt, to aspire to. Watt certainly followed the trend for theoretical learning to underpin practical instrument making. He took a copy of Nicholas Bion’s The Construction and Principal Uses of Mathematical Instruments, translated into English by Edmund Stone, with him when he returned from London to Scotland. Also in his library were Gregory’s The Elements of Astronomy, Physical and Geometrical, Desaguliers’ A Course of Experimental Philosophy and Smith’s Compleat System of Opticks.64 And he borrowed Gravensande’s Elements of Natural Philosophy from his father.
But whether Watt’s business demanded that he pursued such a theoretically grounded, precise approach is doubtful. Instrument making on a tight budget required a more pragmatic approach – like copying the work of others, for instance. Watt wrote to his father, ‘If you can get a good well shaped Davis quadrant that has been approved by some experienced seaman send it up. I shall return it in a week at any rate send the best you can get.’65 He also wrote, ‘I wish you would send up a pair of each sort [of brass dividers] . . . as I am getting a boy who I intend to set making them . . .’.66 Three months later, he wrote again, ‘If you’ll please send up a Binacle [sic] compass & the exact measure of all the different sizes, I shall get you some made.’67 As Watt copied his peers’ work, others did the same to him. Watt sold 80 of his perspective drawing machines in total.68 But George Adams, the prominent London instrument maker, pirated its design as his own invention ‘for taking the
true perspective of any landscape, building, gardens, &c.’, selling it for the considerable sum of six guineas.69 And rather than the latest precision methods, other ruses were frequently employed. John Rabone later told how the straight rules ‘in general use by mechanics’ were made
by placing the article to be divided and the original pattern side by side, then passing a straight-edge with a shoulder fixed at right angles to serve as a guide, along the original, and pausing at each division; then a corresponding line is made on the copy by the dividing knife . . . This method, in skilful hands, admits of much accuracy.70
So for some instrument makers, ‘good enough’ was preferable to ‘the best’, and satisfied their customers.
Concentrating on making instruments that were just good enough did not preclude innovation and awareness of the latest new theories in other areas, however. Brief mention of Adam Smith has been made above, and we return to him now: Watt’s workshop at Glasgow College was a ‘favourite resort’ for Smith during his tenure at Glasgow College.71 Watt and the older professor established a sufficiently strong rapport for Watt, in retirement after 1800, to manufacture a portrait bust of him, a privilege most often retained for Watt’s closer friends and acquaintances.72 Of more immediate import is how Smith’s theories of how the economy worked relate to practices inside Watt’s workshop. In his most famous book, The Wealth of Nations, published in 1776, Smith wrote how labour, not land, was the source of wealth and prosperity, and organizing labour as efficiently as possible by having a high degree of ‘division of labour’ would enable wealth to be maximized. This meant taking workers from making an entire, finished product from scratch and giving them specialized jobs, each based on a single part of the overall process. The example Smith chose was of a factory making pins, where ten specialized workers operating together might make 48,000 pins per day, compared to 200 if they had each to undertake the entire process from a length of wire to the finished pin.73