344 Mr. G. H. Knibbs : Mathematical Analysis of 



Owing to the anomalous character of the first two pairs, 

 it is preferable to combine them. This gives the results 

 shown above in brackets. 



In computing the loss per degree we see that on the 

 whole it increases with the temperature and with the relative 

 or the absolute dryness, the relation, however, being non- 

 linear in either case. 



On referring to the preceding table it will be seen that 

 the saturation deficit or " absolute dr}mess " (see Obsns. 33, 

 34 and 38, 39) may be an increasing quantity, though the 

 relative dryness is a decreasing one. Hence absolute dryness 

 is consequently in all probability the preferable quantity to 

 employ as argument *. To test this a second combination 

 of observations appeared to be desirable in which the ranges 

 of temperature were as large as possible for the same 

 " absolute dryness." 



Table VI. 



No. of 

 Observation. 



3, 5,8 

 11, 15 



Temp. C. 



Relative 

 Dryness. 



Absolute 

 Dryness. 



Loss 

 Weight. 



Loss 

 perl°C. 



Mean 

 Temp. 



Mean 



Loss. 



18-8 

 21-8 

 Incr. 3 



•379 



•312 



Mn. -352 



6-18 



607 



Mn. 6-13 



58 



62 



Incr. 4 



1-33 



20-0 



60 



17, 18, 20, 21 

 22, 23, 24, 25 



24-57 



27*15 



Incr. 2-58 



•644 



•558 



Mn. -601 



14-74 



14-75 



Mn. 14-74 



83 



108 



Incr. 25 



9-68 



25-9 



955 



27, 29, 32, 38 

 26, 36, 37 



3302 

 34 

 Incr. -98 



•785 



•724 



Mn. -759 



29-58 i 198 



29-07 221 



Mn. 29-36 Incr. 23 



23-44 



33-4 



208 



These results give a rate of increase of loss per degree of 

 temperature which, arbitrarily assuming that it may be 

 referred wholly to the saturation deficit, amounts very closely 

 to the following, viz. : — 



Loss in grammes per 1° C. = —4*5 -I- 0'955 J, 



this formula giving 



1-34, 9-57, and 23'53 instead of 1'33, 9-68, and 23'44. 



* The saturation-deficit measures the capacity of a given volume of 

 air mixed with aqueous vapour to absorb further aqueous vapour ; the 

 temperature measures the activity of the molecular motion leading to 

 absorption. 



