448 H. F. Blanford — The Calcutta Standard Barometer. [No. 4, 



the difference of level of the cisterns to be 2 foot, my barometer 

 being the lower. This represents a barometric difference of 

 + 0-0002 which must be deducted from the readings of my own 

 standard in its present situation. Of the 146 comparative readings, 

 73 were made at even hours, simultaneously with the readings of the 

 Surveyor General's Standard, and these give a mean difference of 

 No. 94 = No. 84 ( — .0249). The readings were made at intervals 

 between January 7th and August 23rd, the barometric range in this 

 period being from 30*141 to 29371. The error of No. 94 to No. 84 

 has therefore evidently remained constant, and may be taken with 

 sufficient approximation as — 0*0251. 



The following table gives the results obtained with the several 

 Casella's barometers. 



Results of a comparison of eight larometers ivith JVeivman's No. 94, 

 the Kew and Greenwich Standards. 





No. of 

 Casella's 



Error to No. 94. 



Error to 

 Kew. 



Error to 

 Green- 





barometer. 









wich. 







Rising. 



Falling. 



Mean. 











Inch. 



Inch. 



Inch. 



Inch. 



Inch. 





628 



+ .0132 



+ .0159 



+ .0145 



—.001 



+ .003 





632 



+ .0188 



+ .0158 



+ .0173 



+ .005 



+ .009 





637 



+ .0202 



+ .0211 



+ .0206 



+ .006 



+ .010 





627 



+ .0193 



+ .0188 



+ .0190 



+ .006 



+ .010 





634 



+ .0148 



+ .0171 



+ .0159 



+ .002 



+ .006 





635 



+ .0209 



+ .0235 



+ .0222 



+ .004 



+ .008 





630 



+ .0137 



+ .0163 



+ .0150 



+ .001 



+ .005 





631 



+ .0181 



+ .0203 



+ .0192 



+ .005 



+ .009 



The error of No. 94, (my office standard) to those of Kew and 

 Greenwich, is obtained in each case by changing the signs and 

 taking the difference of the figures in column 4 and those in columns 

 5 and 6 respectively. The results, with two exceptions, are, I think, 

 sufficiently accordant to afford an assurance of the general trust- 

 worthiness of the mean result. 



