1880.] On Actinometrical Observations, made in India. 165 



within the range of the scale. In an instrument so essentially differ- 

 ential as the actionometer, any adjustment involving a change of 

 circumstances cannot be too earnestly deprecated. On the other hand, 

 without excessive sensitiveness, the instrument can command only a 

 very limited range of utility. 



20. There remain to notice a few experiments made in view of very 

 short period fluctuations in the solar or residual radiation ;* these 

 are given in Table IX. On this occasion the two instruments were 

 set up at Dehra within three or four feet of one another ; both were 

 read by means of the same chronometer placed between them, and the 

 two observers capped and uncapped their tubes at the same times. 

 Thus, the readings were taken at the same instant, and under pre- 

 cisely identical circumstances, the observations being continuous from 

 11 h. 39 m. to h. 31 m. apparent time. Taking for each instrument 

 the difference between each result and the mean of all, we obtain the 

 values in columns dA and c£B. Now, the magnitudes of these changes 

 are undoubtedly in excess of fallibility in reading, for the eye can cer- 

 tainly read to one- fourth of a scale division (or less), i.e., to \ millim., 

 or about 0°'005 F., so that apart from calibration errors, which are 

 not likely to be excessive, and which in the two instruments are 

 almost certain to be dissimilar, the inference is that these large differ- 

 ences are at least in part due to solar or residual causes. This con- 

 clusion is numerically tested in the last columns of the table, where I 

 have placed every pair of dA or d~B, in which either member exceeds 

 ten in magnitude irrespective of size ; twelve such pairs occur, and it 

 may be worthy of notice that the two members of every pair without 

 exception are affected by the same sign ; so that so far as the evidence 

 goes, both instruments recorded a rise, or both recorded a fall. To 

 present a large argument of the fact, I take the sums, disregarding 

 signs, of each column, from which it appears that A recorded 151 units 

 of change, and that B recorded no less than 138 units, and in the same- 

 directions. The result, so far as it goes, confirms the conjecture that 

 the solar or residual radiation is remittent, and in very short periods 

 of unknown duration. 



[I cannot too earnestly express my hopes, that actinometrical observa- 

 tions will soon be begun on at least two of the several elevated and 

 eminently eligible sites which India offers. A series conducted sys- 

 tematically and skilfully for even one twelvemonth could not fail to 

 throw much light on a subject which, in its eventualities, concerns 

 every individual on this globe far more than is perhaps commonly 

 contemplated. As to the more immediate questions of periodicity, &c, 

 these cannot be more than approached by means of desultory observa- 

 tions, such as those here presented, undertaken under the pressure of 



* See " Proc. Roy. Soc," vol. 19, p. 228, Article 12. 



