136. The Grcund Splrlt-Ltvel csmparfd 'with the Plumi'Llne. 



the habit of turning the iron pipes they hold in their hands, both the fame way, as they walk 

 backwards. This is the pradice which, by producing undulations and irregularities in the 

 tube, is fuppofed to render them unfit for fpirit-Icvels. I have remarked, that fuch tubes as 

 have veins, or extended bubbles, running lengthways very ftraigiit through the fubllance, have 

 afforded good levels. 



But the mod regular and accurate levels are obtained by grinding the infide of the tube. 

 I do not know that this operation requires any particular management capable of being de- 

 fcribed in words ; though there is no doubt but in this, as well as in every other mechanical 

 procefs, facility and (kill will be acquired by pradice. A cylindrical piece of wood is turned 

 fo as to go eafily through the portion of tube intended to be ground. It is then worked in 

 the tube with water and fine emery in the ufual way. As foon as the polifh has by this means 

 difappcared on one fide, the tube is cleaned, filled, and tried ; and accordingly as its figure 

 proves to be more or lefs ftraight or curved, the grinding is either repeated or difcontinued. 

 Some operators polilh the infide again after grinding; but this has not been found to in. 

 creafe their fenfibility. 



From the great delicacy of the fpirit-level, compared with the few obfervations here 

 prefented on the plumb-line, the former inftrument may appear greatly to deferve the 

 preference. Aftronomers are not however agreed on this point. When a fpirit-level is 

 adjufted by reverling, at a certain temperature, and both ends of the bubble markedjit 

 may be allowed that the inftrument may be fuccefsfully applied to ufe. But if the tem- 

 perature be raifed, the fpirit will expand, and of courfe the bubble will become fliorter. 

 "Whence it appears neceflary that a divifion and adjuHing-piece fhould be applied, from ex- 

 periment, to afcertain the true ftation of the bubble at different temperatures ; and even 

 this application feenis fcarcely adequate to fupply the place of repeated adjuftments. The 

 variation of the bubble will differ according to the quantity of fpirit contained in the tube. 

 In two good levels, of nearly the fame magnitude and figure, I found it amount to one- 

 fifth of an inch for every ten degrees of Fahrenheit. The bubble therefore may be one inch 

 longer in winter than in fummer, which in thefe individual levels amounts to near one- 

 third of the fummer length. The curvature of a fpirit-level will alfo vary from unequal 

 temperature •, fuch, for example, as may arife from one end of the tube being touched or 

 breathed upon, while the other end is left at the original temperature. The error from 

 each of thefe caufes may amount to feveral minutes, as is eafily fhewn by trial ; b\it I do 

 not find that the prefence or abfence of funfliine caufes any perceptible difference. It is 

 probable that the rays may not fpeedily alter the temperature, on account of the tranf- 

 parency. And with regard to thefe three lad fources of error, it muft be allowed that they 

 are eafy to be avoided, and indeed not likely to be prefent in the operations of accurate 

 cbCcrvers. 



MA Till- 



