274 SURVEY OF THE COAST 



Standard unit of length measure, and that the standard unit 

 employed in it must be well authenticated. The means 

 which I used, and the authenticity of my standards, are de- 

 tailed in another place, to which I must refer for informa- 

 tion on this subject. I shall only observe that I had pecu- 

 liarly authentic and well adapted means to obtain a multi- 

 pie of the metre lately determined from the measurements 

 of twelve degrees of the meridian in Kurope, and to deter- 

 mine the length so formed in English measure. 1 was there- 

 fore lead to give this multiple the preference, and accord- 

 ingly I formed bars of eight metres in length, which I con- 

 sidered as the longest that would be well manageable in the 

 actual measurement of a base line. 



Considering the principles on which the measurement of 

 a base line must depend, it is evident that the problem re- 

 quires : 



1. To determine absolutely in space the extreme points 

 of the unit employed in the measurement. 



2. To make this line begin at any given point. 



3. To give it a certain determined direction. 



4. To ascertain its position with respect to the horizon. 

 To satisfy tiie first condition, the theory of mathematics 



applies most generally three rectangular ordinates, and it is 

 easily conceived that in this case something similar must be 

 mechanically executed by some means or other. 



In all the methods hitherto used, the meclianical contact 

 of the bars or chain with the point from which eitlier a pre- 

 vious bar or a line perpendicular to the direction of the base 

 is measured, has been aimed at, and as the moving of the 

 bars in their perpendicular direction presents much difficul- 

 ty, on account of their weight and friction, this has been ob- 

 viated l)y small sliders measuring the intervals between the 

 bars laid near each other in the direction, occasioning of 

 course a vernier reading which required much care and at- 

 tention to small quantities. 



I considf'red an optical contact of the ends of the bar with 

 the determined point preferable to any other, both for ac- 



