317 



rine clock, with jewelled movement, to secure its action in 

 cold weather. In practice, it will be necessary to have clock 

 movements constructed with special reference to the number 

 of records it is desired to obtain. A common thirty-hour 

 clock, in order to have the necessary power to move the 

 hammer^ requires to have each alternate pin removed from 

 the wheel that actuates the hammer lever ; the lever requires, 

 also, to be modified so as to accommodate the change. This 

 change, if quarter-hourly records be required, will make it 

 necessary to re-wind the clock too often. 



My apparatus has the striking part of a clock constructed 

 to move both the time and striking trains, as the equivalent 

 of the striking 'part of one clock ; the whole service of the 

 spring being turned upon the striking train — the time train 

 being removed, its space being occupied by the necessary 

 levers. A marine clock, with time movement only, imparts 

 the time to the striking part, very much as in any ordinary 

 clock, except that the shaft of the minute wheel has/owr pro- 

 jections instead of one, so that I can obtain records four times 

 each hour. A peculiar arrangement, difficult of description, 

 enables me to change the rates from quarter-hour to half- 

 hour, or hour movements of the hammer. 



In describing the Thermometer, I omitted to mention that 

 the lower end of the central brass wire in the bundle of com- 

 pensated rods, was furnished with an adjusting screw, by 

 means of which the index and registering point may be ad- 

 justed to any desirable point. 



My fillet of paper is 2 j inches wide. This enables me to 

 obtain a range of nearly 50° without shifting the position of 

 the registering point. In a trial of one week I have not 

 found any occasion to readjust the index and registering point. 

 I have much difficulty, however, in the scale of my apparatus, 

 and this arises from using a mercurial thermometer as a 

 standard of comparison, the tube of which is probably unequal 

 in its dimensions in various parts, and the scale not corrected 

 to correspond. At first, I constructed my scale from tem- 

 peratures —10° to -1-80°, about 164° on my dial. Subse- 

 quent comparisons show that between 30° and 40° of the 



