Tio6 SURVEY OF THE COAST 



different portions with one another, for which coMiparison and 

 the subsequent division, he liad constructed a proper appa- 

 ratus. He thus formed a table of errors for his scale, in tlie 

 manner described in his method of dividing, (Phil. Trans, of 

 1809) and then laid off the new scale, correcting each point 

 according to the indication of the table. 



The French standards were all compared with this scale 

 of Mr. Troughton, by means of the apparatus described, and 

 by methods which will be mentioned hereafter. 



My particular method of standarding the double metre 

 jjars required another apparatus, which I had constructed 

 for the purpose. 



It may be said to be impossible to cut a bar perpendicu- 

 larly to its length, by hand, with the accuracy required. The 

 tool in question is intended for this purpose. The following 

 are its principal parts : — 



A plate of cast iron, about eight inches broad and two feet 

 and a half long, exactly even and smooth, is adjusted by screws 

 from below, upon a strong iron frame, at one end of which 

 two pieces direct, in a perpendicular slide, the socket piece, 

 which receives the axis of a circular file, of about tliree 

 inches in diameter, to which the above plate is adjusted, so 

 as to make the bar laid upon this plate, and presented to 

 the file, exactly perpendicular to it. A strong iron bar is 

 made to slide over the whole length of the plate, by means 

 of two horizontal screws. The bar to be filed being laid 

 upon the plate, and against this bar, a trial of filing, in two 

 inverted positions, will show any defect of adjustment dou- 

 ble. The adjustment must be made accordingly, and the 

 surface of the cut will be perfectly even. There is a change 

 of files of different fineness, and for the last, a turkey stone, 

 which will take off all the marks of the file, and grind the 

 surface smooth. 



When only one bar is to be standarded, this tool must be 

 used throughout with its different files, and the turkey stone 

 last. But as I had four equal bars to standard at once, and 

 could make the planes of their ends more perfect in conse- 



