484 PHILOSOPHICAL TRANSACTIONS. [aNNO 1754. 



steam being very repellent and active, the bar is liable to be sensibly affected in 

 its length, before the measure can be taken, both by heat and moisture, which 

 both tend to expand the bar; but as the quantity is small, and capable of being 

 nearly ascertained, a wooden bar, thus applied, will answer the same end as if it 

 was unalterable by heat or moisture. To know therefore the quantity of this 

 alteration, let the time elapsed between the first approach of the bar to the instru- 

 ment, and the taking of the measure, be observed by a second-watch, or other- 

 wise; after another equal interval of time, let a second measure be taken ; and 

 after a third interval, a third; and a fourth; the three differences of these four 

 measures will be found nearly to tally with three terms of a geometrical progres- 

 sion, from which the preceding term may be known, and will be the correction, 

 which, if applied to the measure first taken, reduces it to what it would have 

 been if the wooden bar had not expanded during the taking of it. From a few 

 observations of this kind, carefully repeated, the expansion of the basis may be 

 settled; and this once done, the making experiments on other bars will become 

 very easy and compendious. 



The basis of this instrument, as well as other parts of it, is brass. He chose 

 this substance, rather than any other whose expansion was greater or less; because 

 he found, from some gross experiments previously made, that the expansion of 

 brass was nearly a medium between those bodies which differ most in their ex- 

 pansion: a considerable convenience arises from this circumstance; because as 

 the measures, taken in common experiments, are their difference from brass, the 

 dependence on the thermometer will be less, as these differences are less. 



The bar of brass which compose the basis, is an inch broad by half an inch 

 thick, and stands edgewise upwards ; one end is continued of the same piece at 

 right angles, to the height of 3-i- inches, and makes a firm support for the end of 

 the bar to be experimented ; and the other end acts on the middle of a lever of 

 the second kind, whose fulcrum is in the basis; therefore the motion of the ex- 

 tremity of the lever is double the difference between the expansion of the bar, 

 and the basis. This upper part of the lever rises above the lid of the cistern, so 

 that it and the micrometer-screw are at all times clear of the water. The top of 

 the lever is furnished with an appendage which he calls the feeler: it is the extre- 

 mity of this piece which comes in contact with the micrometer-screw. It hence 

 appears, that having the length of the lever from its fulcrum to the point of 

 suspension of the feeler, the distance between the fulcrum and the point of con- 

 tact with the bar, the inches and parts that correspond to a certain number of 

 threads of the micrometer, and the number of divisions in the circumference of 

 the index-plate; the fraction of an inch expressed by one division of the plate 

 may be deduced; those measures are as follows. 



