These four plates from Philosophical Transactions are among the finest and most detailed precision drawings published in the eighteenth century. They accompany the first of three long papers by William Roy dealing with his trigonometric survey of southern England. This was the British contribution to a project to link the observatories of Paris and Greenwich by triangulation, and it would expand into the national institution of the Ordnance Survey. Running to 96 pages, the first paper, of 1785, was easily the shortest: in 1787 he published 282 pages and 504 in 1790. The subject of the first paper was also foundational to the entire project, since the initial and fundamental requirement of such a survey is to establish as accurate a baseline as possible. The paper was ‘An account of the measurement of a base on Hounslow-Heath’, Phil. Trans. R. Soc. 75, 285–480 (1785).
The most obvious characteristic of Roy's paper is its painstaking thoroughness—not only in technical details, but also in its narrative. Roy tells a historical story that begins with the Jacobite rebellion in Scotland in 1745, and he points to the general truth that large-scale survey will often depend on military necessity: ‘if a country has not actually been surveyed, or is but little known, a state of warfare generally produces the first improvements in its geography’ (ibid., p. 385). Having played a leading role in the survey of Scotland that followed the battle of Culloden, Roy gives an account of his subsequent promotion of a survey for the whole island, the impetus that came from the proposal by Cassini de Thury to connect the two national observatories, and the choice of Hounslow Heath for the baseline. The first three plates illustrate three methods for measuring such a line.
‘One of the first instruments, which that able artist Mr. Ramsden had orders to prepare, was a steel chain, one hundred feet in length, the best that he could make’ (ibid., p. 394). Jesse Ramsden was Britain's finest maker of precision instruments, soon himself to become a Fellow of the Royal Society. Although not intended for the final measurement, this was a chain constructed ‘on the principles of that of a watch’ (ibid., p. 394). It is described in great detail and its parts are illustrated exhaustively and at their actual sizes in the accompanying plate (plate 1). A surveyor's chain was normally one of the least regarded of everyday instruments, but it was precisely Roy's intention to elevate mundane measurement, through obsessive care over every detail. Surveying was to be raised from a routine practical art to a national project and even the chain could be an object of aesthetic admiration: a flick from one end was communicated to the other ‘in a few seconds... in a beautiful vertical serpentine line’ (ibid., p. 397).
But this was only the prelude: the chain was not intended for the final measurement. So Roy describes in even greater detail his first attempt to measure the line with maximum precision, using reinforced or ‘trussed’ deal rods capped with bell-metal end-pieces. Three of these rods, each 20 ft long, were made and could be used alternatively by abutting them end-to-end, or by overlapping them and aligning marks close to their ends. The plate illustrates these alternative methods (plate 2). The descriptive account is exhaustive, but even so, Roy thinks that the illustrations have the explanatory edge: ‘Their [the rods’] general construction will be better conceived from the plan and elevation, and other representations of their principal parts, in tab. XVIII. than by any description, however particular, conveyed in words' (ibid., p. 399). The exhaustive account is particularly remarkable, because it was eventually decided that the deal rods were not sufficiently stable and that method was abandoned.
Wooden rods were replaced by three glass tubes (plate 3). The magnificent plate shows the cases that contained them, set on their stands, the central case open to reveal the tube, the covers of the cases to either side showing the horizontal stems of thermometers whose bulbs hung inside, close to the tubes. The final plate (plate 4) is an outstanding visual account of the pyrometer designed by Ramsden to determine the temperature corrections to be applied to measurements with the steel chain and glass tubes. The engraver of the plates is identified as the leading practitioner of the time, James Basire, who was engraver to the Royal Society, but there is no attribution on the plates of the original draughtsman; perhaps we are to assume that it was Roy himself.
Why all this obsessive detail? Roy explained that he had ‘given a very minute account of the actual operations in the field, that the Public, being informed of every circumstance, might be better enabled to judge of the accuracy of the result’ (ibid., p. 461). Clearly they were meant to judge the result sufficiently accurate to bear a national survey, and the plates were to play their part in this persuasion. Their quality and detail had a rhetorical as well as an informative purpose. They were part of the assurance that every detail had been carefully considered and every source of uncertainty minimized. In a sense the paper, with its impressive plates, was to play a role analogous to the baseline itself: its integrity and the confidence it could command had to be sufficient to sustain the entire project—not just linking the observatories but also Roy's greater ambitions for a national survey. The baseline measurement was indeed a public event. The King paid a visit, and the President of the Royal Society, Sir Joseph Banks, set up what we might call a hospitality tent, ‘where his immediate guests, and the numerous visitors whom curiosity drew to the spot, met with the most hospitable supply of every necessary and even elegant refreshment’ (ibid., pp. 425–426).
It was appropriate that Roy's conclusion was simply a number, but one towards which all the resources of his project and his paper had been directed. The final calculated figure for the length of the baseline, reduced to sea level, was 27 404.7219 feet.
It will doubtless be allowed, that infinite pains have been taken in the field and otherwise, throughout the whole of this operation, to obtain a just conclusion; but as the most accurate measurement imaginable is still more liable to err in excess than in defect, we will throw away some useless decimals, and establish the ultimate length of the base at 27 404 feet and seven-tenths. (ibid., p. 478)
- © 2006 The Royal Society