88 Mr. J. C. McConnel on the 



scattered eai-th and sky light. 1 shall therefore denote the 

 ratio 2B/(A— B) by the symbol s. In completely polarized 

 light 5 = 0, and in unpolarized light s = co. In the time of 

 Rubenson it seems to have been the established custom to 

 call the ratio (A — B)/(A + B) of the polarized light to the 

 whole light " the polarization.'^ This expression has the ad- 

 vantage that it increases with the strength of the polarization 

 and vanishes when the light is unpolarized. But it seems to 

 be for most purposes inconvenient, and I shall not use it 

 at all. 



Results deduced fro'm'Rv}d&nso\'i's> Work. 



Rubenson's observations were made at Rome, and may 

 therefore be taken as a fair sample of the polarization at the 

 sea-level. Rome is sufficiently far (some thirteen miles) from 

 the sea to prevent serious disturbance from the reflexion on 

 the water. He found the polarization strongest in the morn- 

 ing and evening, attaining a maximum about mid-day, some- 

 times before and sometimes afterwards. About mid-day the 

 rate of alteration is of course slow, and he noticed that slight 

 perturbations, i.e. sudden irregular variations, were most 

 frequent about that time. The increase of polarization be- 

 came more and more rapid towards sunrise or sunset. He 

 has collected into a table his readings on many diffisrent days 

 at noon and at sunrise or sunset, in summer and winter. 

 Between May 22nd and July 27th, 18G2, the strongest mid- 

 day polarization was (July 5th) ?' = '276 (s = '76), while there 

 were ten days with r less than '34 (s<l*03). With the sun 

 on the horizon the strongest polarization was r — "218 

 (^,s = *56), while there were seven days on which r fell as low 

 as '25 (5 = "67). In winter the polarization was much stronger. 

 Between Oct. 21st, 1861, and Jan. 8th, 1862, there was one 

 noon (Nov. 5th) at which r was '20 (5 = "50), while on ten 

 occasions r at noon was less than "25 (s<*67). With a 

 horizontal sun the strongest polarization was given by r=*156 

 (s = -37), and there were eleven days on which r was less than 

 •19 (s<*47). Not much difference is noticeable between 

 sunrise and sunset. 



I have only cited the days on which the polarization was 

 strongest, for these always have the most perfect skies, as any 

 observer must very soon convince himself. Any modifying 

 circumstances, such as haziness of the sky, smoke, the faintest 

 cloud in the field of view, the presence of large clouds 

 scattered about the sky, serve invariably to diminish the 

 polarization. The increase of s, as the sun mounts from the 

 horizon to its highest altitude, in summer from '56 to "76, 



