13i Suggestion of a Nexs: Method of Trial 



chloride with peroxide, and by some other processes. He adds, 

 however, that it has never been analysed : — from the last-men- 

 tioned circumstance, as well as some general views offered in 

 my communication, I have not thought proper to suppress it, 

 although it contains less novelty than I had imagined. 



XIX. Suggestion of a New Method of Trial for the Chronome- 

 ters at Green-doich. By A Correspondent. 



To the Editors of the Philosophical Magazi7ie and Annals. 



Gentlemen, 



HAVING perused in your valuable Journal the remarks 

 of Caleb Mainspring on the System of Prize Chronome- 

 ters at Greenwich, and having also noticed the Proposal of 

 another trial number suggested by J. L. T. as published in 

 the last Number of your Journal, I am induced as a chrono- 

 meter-maker to address this letter to you upon the subject; 

 hoping that my observations may prove useful to those who 

 have any interest in the matter. The Admiralty, by awarding 

 premiums for chronometers, has proved how much might be 

 effected by the encouragement of science, as will be seen by 

 comparing the rates of the chronometers for the first year's 

 trial with those of the subsequent ones. But when the Ad- 

 miralty, in their wisdom, offered premiums for the best chro- 

 nometers after one year's trial at the Royal Observatory, they 

 of course did not intend that the inferior chronometer should 

 hold the place of its superior ; but such, however, is the case 

 by the present system of deciding the prizes, as the remarks 

 of Caleb Mainspring clearly show. The perfection of a chro- 

 nometer, in my opinion, consists in maintaining its i-ate with 

 uniformity, and not irregularity ; and the more steadily a chro- 

 nometer goes, the greater is the proof of its being perfect in the 

 different parts of its construction : and I consider that a chro- 

 nometer which varies but little from its rate, but which goes 

 at an irregular pace, is not equal with one that varies a little 

 more, but which goes more regularly. By the present method of 

 deciding the prizes, the chronometer may be the reverse of what 

 I have stated, as will be seen by the example given by Caleb 

 Mainspring, published in your Journal for December last. But 

 now, whilst I find fault with the present system, I consider 

 myself called upon to suggest another, such as would remedy 

 the evil complained of; and therefore I propose that the varia- 

 tion from the mean monthly rate be not doubled, as is now the 

 case, but that it may be left single, and added to the mean 

 extreme variation, which would give the true error of the chro- 

 nometer; 



