[ 53 ] 



point within the ohject-gia(s about one third of its 

 thicknefs from that furface which is towards the 

 wires, if the glafs be, as ufua], equally convex on 

 both fides ; but if the focus of the object- giafs 

 is pretty long and its thicknefs not great, the error 

 that can ariie by meafuring from any part of the 

 object-glafs will become inieniible as to the altera- 

 tion in the angle. 



The table for the micrometer may likewife be 

 made by fetting up two marks at a diftance on the 

 ground, and obferving with the micrometer the 

 revolutions, &c. which they fubtend when ken. 

 through the telefcope, and then computing the 

 angles thofe objects fubtend at the object-glafs, by 

 meafuring their diflance from each other and from 

 the obj eft- glafs. The like may aifo be done by 

 opening the threads to any number of revolutions, 

 and then making a itar move exactly upon the per- 

 pendicular thread, and noting the time it is paffing 

 from one parallel thread to the other ; for that time 

 turned into minutes and feconds of a degree, by 

 allowing for the fear's declination and going of the 

 clock, &c. will be the angle anfwering to the num- 

 ber of- revolutions ; from which the whole table 

 may be made. This method perhaps might be 

 moft advantageoufly practifed in ftars near the pole, 

 where the apparent motion being flow a fecond in 

 time will aniwer to a much fmaller angle than 

 towards the equator. But I believe, upon trial, 

 the firft method will be found mod eafy and prac^ 

 ticable, efpeeially if the fcale made ufe of be well 

 divided* 



VII. A 



