OF ARTS AND SCIENCES. 



93 



at all heights. It is true that the level should he constantly adjusted, 

 but it is also true that an instrument cau be made where such an ad- 

 justment is not necessary. And where time is an element in the 

 accuracy, such an instrument should be used. In the present case it 

 was absolutely necessary to read as quickly as possible, so as not to 

 leave time for the column to change. In the first place the round 

 column, when made, was turned in a lathe to nearly its final dimen- 

 sions. The line joining the centres of the sections must then have 

 been very accurately straight. In the subsequent fitting some slight 

 irregularities must have been introduced, but they could not have, been 

 great with good workmanship.* The upright column was fixed, and 

 the telescope moved around it by a sleeve on the other sleeve. Where 



the objects to be measured are not situated at a very wide anjrle from 

 each other, this is a good arrangement, and has the advantage that any 

 side of the column can be turned toward the object, and so, even if it 

 were crooked, we could yet turn it into such a position as to nearly 

 eliminate error. 



It was used at a distance of about 110 cra- from the object, and 

 no difficulty was found after practice in setting it on the column to 

 5^ mm - at least. The cross hairs made an angle of 45° with the 

 horizontal, as this was found to be the most sensitive arrangement. 



The scale was carefully calibrated, and the relative errors f for the 



* The change of level along the portion generally used (lid not amount to 

 more than .1 of a division, or about .01 n,n '- at the mercury column, as this is 

 about the smallest quantity which could be observed on the level. 



t These amounted to less than .016 mm - at any part. 



