OF ARTS AND SCIENCES. 143 



like our volcanic pumice rocks, a withdrawal of the solar rays might 

 be attended by so sudden a fall of temjierature as to explain the eclipse 

 observations. Somewhere between the two the explanation probably 

 lies : but at present we can do little better than conjecture. 



Atmospheric Transmission Curve. 



An attempt has been made to construct a curve which would show 

 the transmission of the atmosphere for lunar rays at every altitude of 

 the moon. Observations for this purpose have been made, at every 

 opportunity, for nearly a year, and already amount to some thousands 

 in number. It has been customary to begin the observations at moon- 

 rise, or, if the moon were already in the sky, as soon as the evening 

 temperature had become sufficiently stationary, and to continue them 

 until culmination. 



The working of the apparatus is attended with considerable fatigue, 

 and it is not possible for one person to operate it uninterruptedly. 

 The readings, therefore, are made in sets of ten or twenty with a short 

 rest between the sets. Sometimes a third person has measured the 

 altitudes with a portable transit, while two gave their attention to 

 making the readings ; but usually the same person has directed the 

 thermograph-box and taken altitudes at the beginning and the end of 

 each set, the mean reading of each set being regarded as the deflection 

 corresponding to the mean altitude. 



Below is given the work of the night of February 28, 1888. Con- 

 siderable dilhculty has been experienced in obtaining the deflections for 

 the high altitudes, as it very frequently happens that, late in the even- 

 ing, clouds arise from the ocean near at hand and obscure the sky. On 

 this particular evening the sky was very serene and beautiful until 

 10 h. 40 m., when the clouds arose. Because of this unusual seren- 

 ity, this series of observations shows more than ordinary regularity, 

 although less than the customary number of observations was made 

 to a set. 



