192 CELESTIAL MEASURINGS AND WEIGHINGS. 



part of its polar semi-axis : that is to say, if we disregard 

 so insignificant an error as a furlong upon 8000 miles, 

 or one part in 64,000. 



(15.) It follows, moreover (as may be verified by any 

 one who will make the calculation), that if we consent 

 to disregard so trifling an error as one part in 8000 : one 

 cubic geoineti'ical foot of distilled water at our standard 

 temperature weighs exactly 1000 of our actual imperial 

 ounces, and is exactly filled by 100 of our actual impe- 

 rial half-pints.* 



(16.) Having thus exhibited the connexion between 

 our ordinary measures of len^ h, weight, and capacity, 

 and the dimensions of the globe we inhabit (a connexion 

 of singular felicity, when we consider the simplicity of 

 the numerical relations), we are prepared to take a fur- 

 ther step, and, by using the diameter of the earth itself 

 as a base-line, carry on the same j^rinciple of triangu- 

 lation into our solar and planetary system. In this, the 

 natural unit that to which astronomers have a<2;reed 

 with one accord to refer all its dimensions is the mean 

 or average distance of the earth from the sun, or the 

 semi-axis of the ellipse which it describes about that 

 luminary. 



(17.) The way in which a knowledge of this distance 

 is obtained being very fully described in our Lectures on 

 " The Sun" and on '' Comets," f it is unnecessary to re- 



* The deviation of the achial French litre and gramme from 

 their true theoretical values, is more than three times as great, be- 

 ing one part in 2730. 



t A very unfortunate enatuni exists in one of the numbers in p. 



