180 Performance of Some Timekeepers. 



centre, but to the squares of these distances, so that while 

 graphically the effect on the spring would be represented by 

 a straight line, that on the balance would be a curve, which, 

 near the point of contact with the straight line, would 

 nearly coincide with it for a moderate distance, or would cut 

 it in two points with a moderate deviation half-way between 

 them. It is in the latter way that the chronometer is 

 generally adjusted, so that it goes right at 55° and 85°, and 

 it then gains about six-tenths of a second a day at 70°, the 

 mean temperature. Various methods have been proposed 

 for getting rid of this secondary error. The most successful 

 seems to have been Kulberg's, which was exhibited in his 

 marine chronometers at our late Exhibition. Another plan, 

 invented by the superintendent of the Waltham factory, 

 was used in some of their watches tested at the Observatory. 

 The watches, however, were in very bad order, so that its 

 efficiency was not made apparent. Watches that are carried 

 in the pocket by day and placed under the pillow at night 

 are in a great measure guarded from extreme changes of 

 temperature. Carriage and similarly governed clocks are- 

 generally, however, exposed to great and sudden variations, 

 so that, if not well compensated, their going must be greatly 

 inferior to that of a common pendulum clock. From 

 records kept at the Observatory I find that the maximum 

 temperature of my room there, which has thick walls and a 

 room above it, during the last ten years has been 90°, and 

 the minimum 47°. This would cause a difference of four 

 and a-half minutes a day in an uncompensated watch. In 

 a weatherboard-lined room at my quarters, without a fire- 

 place, the maximum has been 102°, and the minimum 31'5°. 

 I have also carried a delicate thermometer in my watch 

 pocket, to ascertain the temperature a watch is ordinarily 

 submitted to. Sitting in my room where the thermometer 

 indicated 59"3°, the temperature of the pocket was 82 # 4°. 

 Another time the room was o8*0° ; pocket, SO"!) . Open air, 

 55*5° ; pocket, 70'4°. Air, 57'5° ; pocket, with back to the 

 sun, 720°; facing the sun, 81"0 o . Bedroom, 63 0° ; under 

 pillow, 71-5°. Air 91-0°; pocket, 851. Koom, 74-0°; pocket, 

 85-0°. Bedroom, 77-0° ; under pillow, 81-8°. Air, 101 -0°; 

 pocket, in shade, 96"4°. Bedroom, 52*0° ; under p>illow, 590. 

 Air, 50'0°; pocket, 7l'0°. According to the above observa- 

 tions, the watch had been kept in temperatures varying 

 from 59*0° under the pillow to 9 6 '4° in the watch pocket 

 while walking in the open air. This would correspond to a 



