62 H. F. Blanford — Experiments with the Psychrometer. [No. 2, 



Table III. 



Method. 



Sets obs. 



Mean 

 Error. 



Highest. 



+ 



Lowest. 



Sums of squares 

 of errors. 





in excess. 



in defect. 







+ 





By August's form., 

 „ Apjohn's „ 

 „ Glaisher's facs. 



13 



18 

 16 



7 

 2 

 4 



+ 0-46 

 + 370 

 + 3-00 



4-9 

 8-0 



8-7 



6-1 

 0-2 

 1-6 



86-13 

 406-98 

 327*80 



71-19 



0-05 

 5-62 



If we take those series only that were made in the thermometer- sheds, 

 with a mean difference of 42-7° (extremes 354 and 49 - 6) between the air- 

 temperature and the dew-point, the errors of the several methods are as 

 follow : 



Table IV. 



Method. 



Series. 



Mean 

 Error. 



Highest. 



+ 



Lowest. 



Sums of squares 

 of errors. 





in excess. 



in def. 







+ 



— 



By August's form., 

 „ Apjohn's „ 

 „ Glaisher's facs., 



8 

 12 

 12 



4 

 

 



+ 0-68 

 + 4-40 

 + 3-62 



4-9 

 8-0 

 6-7 



4-6 



60-95 

 292-76 

 200-85 



33-73 



The first general conclusion to be drawn from this discussion is that 

 while the results shew a considerable range of error (which may be in part 

 due to a faulty use of the Regnault), on the mean of all the observations, the 

 dew-point computed by August's formula from observations of the psychro- 

 meter made under an open shed, comes very near to that observed with 

 Eegnault's hygrometer, even when the dew-point is more than 40° below 

 the temperature of the air. Both Apjohn's formula and Glaisher's factors 

 appear to give too high a result. 



With regard to the less complete exposure afforded by a verandah, or 

 any place through which the air is not moving freely, the two Allahabad 

 series (16 and 17) and the first Lahore series (19) shew that the psychro- 

 meter, in such a situation, gives too high a humidity by all the methods of 

 reduction ; and the Secunderabad series (1 to 4) though made under a shed, 

 tend to support the conclusion that a still atmosphere is prejudicial to all 

 the psychrometric method. These observations were taken on the day of 



