26 RESULT OF METEOROLOGICAL OBSERVATIONS 



Diagram K. — Brunswick, Me. 



36 

 34 



32 

 30 

 28 h 



..y\-/\/ 



/\ 



- /" 



8 9 10111213141516171819 202122 23 24 25 

 March. 



The irregularities in the Marietta curve are quite considerable, but as they are 

 not supported by corresponding irregularities in the Brunswick observations^ they 

 must be regarded as accidental. There is a slight corresponding depression in the 

 two curves on the 22d. 



I For the proper estimation of the irregularities in the Marietta curve of the daily 

 means derived from 32 years of observations we require to know the probable error 



of each mean value ; this value (q- = 0.845 LJ) in February when it is near its 



maximum, is ± 1°.3, and in August, when near its minimum, is ± 0°.8 very 

 nearly. Within the limits ±1.3 \/2"and ±0.8 \/2, therefore, the means vary from 

 day to day without indicating the presence of any unusual cause of deviation. 

 I Table IV contains the monthly means for the observing hours 6 or 7, 2, 9 ; the 

 morning observations being made at various times a column has been added to each 

 month indicating the hour as noted in the journal or as inferred. In Mr. Wood's 

 observations the evening record refers to sunset. 



' The daily means given for Brunswick are simply the means of the three observations on each 

 day, uncorrected for diurnal irregularity, and consequently correspond to other means given in Table 

 I of my results of meteorological observations at that station. (Smithsonian Contributions to Know- 

 ledge— 204— Dec. 1866.) 



