38o 



NA rURE 



[August 15, 1889 



inch, its temperature must be known to o°o6 F. if it is of brass, 

 or to o°'09 F. if it is of iron. To get thermometers that will 

 indicate their own temperature to that degree of accuracy is by 

 no means easy, but to determine the temperature of a bar from 

 their readings is far more difficult. Again, we imagine the 

 length of our standards to follow their temperature rigorously, 

 ■but what proof is there that such is the case ? If we determine 

 the freezing point of an old thermometer, then raise it to the 

 temperature of boiling water, and immediately thereafter again 

 determine its freezing-point, we invariably find that the freezing- 

 point has fallen a little ; and we explain this by saying that the 

 glass has taken a 'set, from which it requires time to recover. 

 Is it not probable that an effect similar in kind, although less in 

 degree, occurs in all solids when their temperature is varying ? 

 When we look at the highly polished terminals of an end- 

 standard, we are apt to regard them as mathematical surfaces, 

 separated by an interval which is perfectly definite, and which 

 •could be measured with infinite precision if we only had the 

 necessary instrumental appliances ; but is that a correct view ? 

 The atomic theory answers emphatically. No. According to it, 

 -all matter consists of atoms, or molecules, of a perfectly definite 

 size, and with definite intervals between them ; but even if 

 that is denied, the evidence is now overwhelming that matter 

 is not homogeneous, but .possesses a grain of some kind, regu- 

 larly repeated at intervals which cannot be greater than 

 1/2,000,000 nor less than 1/400,000,000 of art inch. Accordingly 

 we must picture our standard bar as a conglomeration of grains 

 of some kind or other, having magnitudes of the order specified, 

 and all in ceaseless motion, the amplitude of which depends 

 upon the temperature of the bar. To our mental vision the 

 polished terminals are therefore like the surface of a pot of 

 boiling v^ ater, and we recognize that there must be a limit to the 

 accuracy with which the interval between them can be measured. 

 As a basis for estimating how near this limit we have approached, 

 it will suffice to say that for fifty years past it has been customary 

 to state comparisons of standards of length to 1/100,000, of an 

 inch. Nevertheless, most authorities agree that, although 

 1/100,000 of an inch can be distinguished in the comparators, 

 1/25,000 of an inch is about the limit of accuracy attainable in 

 comparing standards. Possibly such a limit may be reached 

 under the most favourable circumstances, but in the case of the 

 yard and the metre, which are standard at different temperatures, 

 the following values of the metre by observers of the highest 

 repute render it doubtful if anything like that accuracy has yet 

 -been attained : — 



1818 Capt. Henry Kater 



1866 General A. R. Clarke 



1883 Prof. Wm. A. Rogers 



18815 General C. B. Comstock 



39 '37079 inches. 

 39 '37043 ,. 

 39-37027 „ 

 39 '36985 ,. 



The earliest standard of English weight of which we have any 

 -very definite knowledge is the Mint pound of the Tower of 

 London. It weighed 5400 troy grains, and the coinage was 

 regulated by it up to the year 1527, when it was abolished in 

 favour of the troy pound of 5760 grains. Contemporaneously 

 with the Tower pound there was also the merchant's pound, 

 whose exact weight is now involved in so much doubt that it is 

 impossible to decide whether it consisted of 6750 or of 7200 

 grains. The Tower pound and the troy pound were used for 

 weighing only gold, silver, and drugs, while all other commodi- 

 ties were weighed by the merchant's pound until the thirteenth 

 or fourteenth century, and after that by the avoirdupois pound. 

 It is not certainly known when the troy and avoirdupois pounds 

 were introduced into England, and there is no evidence of any 

 relation between them when they first became standards. The 

 present avoirdupois pound can be clearly proved to be of similar 

 weight to the standard avoirdupois pound of Edward III. 

 (a.d. 1327-77), and there is good reason for believing that no 

 substantial change has occurred either in its weight or in that of 

 the troy pound since their respective establishment as standards 

 • in England. 



The oldest standard weights now existing in the English 

 archives date from the reign of Queen Elizabeth, and consist of 

 a set of bell shaped avoirdupois weights of 56, 28, and 14 pounds, 

 made in 1582, and 7, 4, 2, and i pounds made in 1588 ; a set of 

 flat cii cular avoirdupois weights of 8, 4, 2, and i pounds, and 

 8, 4, 2, I, \, \, I, and ^-^ ounces, made in 1588 ; and a set of 

 •cjp-shaped troy weights, fitting one within the other, of 256, 

 128, 64, 32, 16, 8, 4, 2, I, i, \, \ (hollow) and \ (solid) ounces, 

 also made in 1588. All these standards were constructed by 



order of Queen Elizabeth, under the direction of a jury composed 

 of eighteen merchants and eleven goldsmiths of London, the 

 avoirdupois weights being adjusted according to an ancient 

 standard of 56 pounds, remaining in the Exchequer from the 

 time of Edward III. ; and the troy weight being adjusted 

 according to the ancient standard in Goldsmiths' Hall. 



In view of the fact that the weight mentioned in all the old 

 Acts of Parliament from the time of Edward I. (a.d. 1274-1307) 

 is universally admitted to be troy weight, the Parliamentary 

 Coii\mittee of 1758, appointed to inquire into the original 

 standards of weights and measures in England, recommended 

 that the troy pound should be made the unit or standard by 

 which the avoirdupois and other weights should be regulated ; 

 and by their order three several troy pounds of soft gun-metal 

 were very carefully adjusted under the direction of Mr. Joseph 

 Harris, who was then Assay Master of the Mint. To ascertain 

 the proper mass for these pounds, the Committee caused Messrs. 

 Harris and Gregory, of the Mint, to perform the following 

 operations in their presence : — 



(i) In the before-mentioned set of troy weights, made in 1588, 

 which were then the Exchequer standard, each weight, from 

 that of 4 ounces up to that of 256 ounces was compared suc- 

 cessively with the sum of all the smaller weights ; and by a 

 process for which no valid reason can be assigned, it was con- 

 cluded from these weighings that the troy pound composed of the 

 8 and 4 ounce weights was \\ grains too light. 



(2) The aforesaid 8 and 4 ounce weights of the Exchequer 

 were compared with five other authoritative troy pounds, four of 

 which belonged to the Mint and one to Mr. Freeman, who, like 

 his father before him, was scale-maker to the Mint, and from the 

 mean of these weighings it appeared that the sum of the 

 Exchequer 8 and 4 ounce weights was l grain too light. 



The Committee adopted the mean between the latter result 

 and that which they had deduced from the Exchequer weights 

 alone, and accordingly Mr. Harris made each of his three troy 

 pounds \\ grains heavier than the sum of the Exchequer 8 and 

 4 ounce weights ; but sixty-six years were destined to elapse 

 before Parliament took action respecting them. 



The Commissioners appointed in 1818 to establish a more 

 uniform system of weights and measures repeated the recom- 

 mendations of the Committee of 1758, and as the avoirdupois 

 pound which had long been used, although not legalized by any 

 Act of the Legislature, was very nearly 7000 troy grains, they 

 recommended that 7000 such troy grains be declared to consti- 

 tute a pound avoirdupois. These recommendations were embodied 

 in the Act of Parliament of June 17, 1824, and thus one of the 

 troy pounds made in 1758 became the Imperial standard. That 

 standard, like Bird's standard yard, was deposited in the Houses 

 of Parliament, and was burned up with them in October 1834. 



The present English standard pound was made in 1844-46 

 by Prof. W. H. Miller, who was one of the members of the 

 Commission appointed in 1843 to superintend the construction 

 of the new Parliamentary standards of length and weight 

 destined to replace those destroyed in 1834. A number of 

 weights had been very accurately compared with the lost 

 standard ; namely, in 1824 or 1825, by Capt. Kater, 5 troy 

 pounds of gun-metal, destined respectively for the use of the 

 Exchequer, the Royal Mint, and the cities of London, Edin- 

 burgh, and Dublin ; and in 1829, by Capt. von Nehus, two troy 

 pounds of brass and one of platinum, all in the custody of Prof.. 

 Schumacher, and a platinum troy pound belonging to the Royal 

 Society. The first step for recovering the mass of the lost 

 standard was manifestly to compare these weights among them- 

 selves, and upon so doing it was found that for the brass and 

 gun-metal weights the discrepancies between the weighings 

 made in 1824 and 1844 amounted to 0'0226 of a grain, while for 

 the two platinum weights the discrepancies between the weigh- 

 ings made in 1829 and 1845 was only 0-00019 of a grain. With 

 a single exception, all the new brass or gun-metal weights had 

 become heavier since their first comparison with the lost 

 standard, the change being probably due to oxidation of their 

 surfaces, and on that account the new standard was made to 

 depend solely upon the two platinum weights. For convenience 

 of reference these weights were designated respectively S/ (Schu- 

 macher's platinum, and RS (Royal Society). A provisional 

 platinum troy pound, T, intermediate in mass between S/ and 

 RS was next prepared, and from 286 comparisons made in 

 January, February, July, and August, 1845, i^ was found that in 

 a vacuum 



T = S/ -t- 0"ooio5 grain, 



