158 



Mr. J. B. N. Hennessey. 



[Dec. 16, 



attempt to assert ivhich prevails : so that, in the sense liere employed, 

 residual and intrinsic (or other equivalent phrase) are convertible 

 terms, as is intended throughout this paper. This admission need not, I 

 think, be received as a deterrent to the use of the actinometer under 

 proper conditions, for when we know more of V 1 the road to S may 

 become easier. Meanwhile, as a matter of fact, the actual radiation 

 by which we are normally affected, at least in sunny lands, is repre- 

 sented, perhaps, more nearly by V x than by S. 



It may also be stated here, once for all, that the conclusions 

 advanced in this paper, like most others of their kind, are not in- 

 tended to cover more ground than the observations themselves ; the 

 former go, of course, no further than the latter.] 



Of these the residual effect alone presents the real object of measure- 

 ment. Complete elimination of instrumental defects, with retention of 

 excessive sensitiveness, is still a desideratum, and local causes may not 

 only produce overwhelming effects,* but they are so completely beyond 

 estimation or control, that the only remaining alternative is never to 

 observe when they are visible or likely to be present, at any rate if 

 small quantities are the objects of search. As, however, the necessity 

 for placing the actinometer, so far as practicable, beyond local influences, 

 is now so fully recognised, I need only here dwell on one of the further 

 conditions necessary to be secured. This suggests itself if only on the 

 score of fallibility and the necessity to collect in general abundant 

 evidence. On these grounds alone a series of observations should be 

 prolonged as much as practicable, so as to yield numerous instead of 

 only a few results. Here, however, apart from instrumental inade- 

 quacy, we are met under even the most favourable local conditions, by 

 the varying absorptions of our own atmosphere at different zenith 

 distances, and it thus becomes a matter of primary importance to 

 establish a time-range, during which inconstancy of radiation from 

 this cause may be reckoned as certainly absent, at least so far as the 

 sensitiveness of the instrument can detect, f In fact, while the obser- 

 vations at varying zenith distances may eventually yield valuable 

 results for reduction to the local zenith, it appears wiser in the first 

 instance to set aside every complicating cause that can be avoided and 

 to establish a daily time-range of the kind indicated. 



* In course of the present observations I have arrived at the conclusion that 

 the presence of strati, even if distant from the sun, sensibly affect the radiation, 

 •while cumuli are comparatively innocuous. 



f The capability to measure minute changes, besides the power it confers, governs 

 the duration of a single exposure, which in turn is the less likely to afford means of 

 detecting short solar periodicities, the more numerous the latter are within the 

 duration. No doubt a contrivance for yielding a continuous curve woiild leave nothing 

 further to be desired in respect of periodicities, but I am not aware that this has as 

 yet been achieved, without sacrifice of the essential sensitiveness, accuracy, and 

 durability. 



