OF TIIE HORIZONTAL FORCE. 7 



corresponding temperature, as given in volume I of the record, may be reduced to 

 the true mean by the interpolation of values, between December 31 and January 

 12. The daily mean (at 32), on December 31, was 842.3, and on January 12, 

 913.0, hence, omitting the readings for January 3d, and 10th, as Sundays, the com- 

 plete monthly mean should be 18.6 divisions less or equal 926.0. 



The observations were resumed on the 12th, and continued to February 8th at 

 22 1 ' 49 | m , when the wires were found to have been slightly deranged, two days 

 previously, February 6, 18 h 22 m (Philadelphia time), a great change in the position 

 was noticed ; on re-arranging the instrument it did not return to its former read- 

 ings. A correction of -f- 116 has been applied (in the record) to the previous mean 

 readings only in this month, and in consequence + H6 divisions should be added 

 to each individual reading from the commencement of the series ; but on account 

 of another disturbance of the instrument, on the 22d, at 16 h 22 m (Philadelphia time), 

 a further correction of -f- 92.8 scale divisions should be applied. The total correc- 

 tion is therefore + 208.8. Besides these corrections the readings on the 22d from 

 O u 22 m (Philadelphia time) to 10 h 22 m (Philadelphia time), inclusive, should be 

 increased by + 25.1 divisions, the alhidade of the instrument having been dis- 

 turbed. 1 



On the 2d of June, 1841, the suspension wires were struck accidentally, derang- 

 ing the instrument ; the readings were then near the end of the subsidiary scale, 

 and in rearranging the instrument the new readings were brought near the middle 

 of the scale. The total difference between the old and new scale readings, the 

 latter commencing with the first of the month, is 900 scale divisions. The means 

 between June 1st and 5th are already corrected in the record, but the individual 

 bi-hourly readings require a correction of + 21 3 2 scale divisions to produce these 

 means. It was thought best not to apply this correction of 900 divisions to the 

 observations between June, 1840, and June, 1841, but simply to state the quantity 

 since it can be applied easily to any result hereafter. 



At the close of 1842 the regular observations were discontinued for three months, 

 during January, February, and March, 1843; a daily reading was taken at 14 h 22 m 

 (Philadelphia time), in order to keep up a continuity in the series. By means of 

 the reduced readings in the same months in the other years, it was found that a 

 correction of -- 3 d . 4 3*. 7 and + l d .5 for January, February, and March, respect- 

 ively, was required to refer the mean at 14 h 22 m to the mean of a complete bi-hourly 

 daily series. Applying these corrections, the corrected monthly means become : 



Fur the first day only + 142, according to the mean in the record. 



