424 ANNALS NEW YORK ACADEMY OF SCIENCES 



3. In addition to this long-period variation, the observations of Schur 

 and Ambronn v^ould seem to indicate a fluctuation in the measured value 

 of P.-E. having a period of about twenty-eight days. The observations 

 are so scattered that they do not permit of a thorough determination of the 

 reality of this fluctuation and of the exact length of its period. If real, 

 this fluctuation can be accounted for by a permanent or semi-permanent 

 deformation of the sun's equator. 



4. Questions as to the exact shape of the sun, and as to possible varia- 

 tions in its size and shape, can only be set at rest by a long series of homo- 

 geneous observations. The data at present available are not sufficient 

 for this purpose. Observations should be made on every clear day, and the 

 series should be extended over at least one solar cycle. 



5. A photographic heliometer would probably furnish the best results. 

 With such an instrument, the moments of good "seeing" can be utilized 

 and a number of plates taken within a short time. These plates can later 

 be measured and reduced by the ordinary staff of a computing bureau. 



Columbia University, 

 March, 1908. 



