414 ANNALS NEW YORK ACADEMY OF SCIENCES 



Conclusions. — While very little weight can be attached to the meridian 

 observations, yet every series shows a measurable departure from the spheri- 

 cal form. The observations of the Von Lindenau series were afterwards 

 included in the more complete and thorough discussion of Auwers, who 

 showed that the great difference, 5", found by Lindenau was not warranted 

 by the observations. Meridian observations are unsuitable for the investi- 

 gation, and if there be any departure of the sun from sphericity, it is below 

 the limit of such measures. 



The heliometer measures confirm the result that the departure from 

 sphericity is extremely small, but they hardly warrant the assumption that 

 the sun is a sphere. Auwers, Ambronn, Seeliger and Wellmann have 

 shown that the measured diameters depend upon the color and thickness 

 of the shade-glasses used in the observations; Auwers found instrumental 

 peculiarities, different instruments giving different results; and Ambronn 

 found that the introduction of a reflecting prism diminished the measured 

 diameter by 0''.4. The series of observations used by Auwers were ex- 

 tremely heterogeneous, twenty-three observers using five heliometers. The 

 individual series were short, extending over periods of a few months 

 only. The series of Ambronn, on the other hand, are perfectly homo- 

 geneous, and they furnish the very best evidence yet obtained. They 

 indicate that the departure of the sun from spherical form is extremely 

 minute, and at the very limit of possible measurement by these means. 

 Considering this series by itself, if no weight be placed upon possible varia- 

 tions in the solar diameter, then the conclusion must be that the sun is 

 sensibly a sphere. 



Unfortunately for a thorough test of the photographic method, no long 

 series of plates Avere available. While several observatories have for many 

 years photographed the sun on each clear day, yet these photographs are 

 not suitable for the present investigation.^ They were mostly made with 

 horizontal instruments and no attention paid to sharpness of edge. The 

 mirror of such instruments introduces errors and makes the image uns}Tn- 

 metrical. Only such plates as have been made with an equatorially mounted 

 objective of relatively long focus can be used, and there are no long series 

 of such plates. The early plates of Rutherfurd give quite consistent results, 

 the general mean of all the plates agreeing closely with that of Auwers as 

 found from the heliometer measures in the same years. These plates of 

 Rutherfurd were made on collodion fikns and give the sharpest and best 

 images of any plates measured. 



» This applies to American observatories. The photographs taken at the Royal Observa- 

 tory at Greenwich are probably well adapted for this investigation. 



