OP ARTS AND SCIENCES. 301 



date, the Greenwich mean time, the angular distance of the point ob- 

 served from the sun, and the resulting polarization. The first 44 sets 

 were made by myself, the others by Miss N. A. Farrar. 



Nos. 1 to 11 were made on the zenith after sunset. The later ob- 

 servations were ditficult on account of the increasing darkness. On 

 reducing the observations, an unexpected result was attained. The 

 last column shows that the polarization increases regularly, becoming 

 twice as great at the end as at the beginning of the series. The 

 probable explanation of this phenomenon is, that after the sun has set 

 the light reaching us from the zenith comes from the upper portions 

 of the atmosphere. As this is presumably clearer, and more free from 

 dust and haze, we should anticipate a more complete polarization. It 

 affords a simple means of determining the effect of ascending to vari- 

 ous heights above the surface of the earth. From the distance of the 

 sun below the horizon these heights are readily calculated. Nos. 22 

 to 33 were made in the same way, and aiford a confirmation of these 

 results. The measures on August 1 are probably somewhat too small, 

 owing to an inclination of the bars to the plane of polarization. On 

 August 2, the sky was clear, but hazy around the sun. The observa- 

 tions were abandoned owing to the approach of cirrus clouds, which 

 probably caused the diminution in the amount of the polarization. 

 On September 13, the air was very clear. No satisfactory explanation 

 appears for the small amount of polarization on July 15. The record 

 states that clouds interfered with set 45, but that it was clear when 

 sets 46 and 47 were taken. The bars may not have been set parallel 

 to the plane of polarization, as it was the first time the observer had 

 used the instrument. 



The law determining the amount of polarization at different dis- 

 tances from the sun is indicated by the observations made on July 28 

 and on September 21. On the first date, the remark occurs, "Sky 

 hazy," and on the second, " Sky entirely clear." The greater polari- 

 zation on the latter date is doubtless due to this cause. Both sets 

 show the maximum polarization at 90°. They also indicate a greater 

 polarization at 15° than at 30°. This may be due to Babinet's neu- 

 tral point; but in that case we should anticipate a negative polai-i- 

 zation at 15°. Probably these sets are uncertain, on account of the 

 proximity of the sun. 



One of the most important applications of this polarimeter will be 

 as a measure of the haziness of the air. As the polarization becomes 

 insensible when the sky is entirely overcast, we apparently have a very 

 delicate test for the amount of haze. Simultaneous observations of a 



