POOR, THE FIGURE OF THE SUN 409 



These results are fairly accordant, and indicate that at this time the 

 radius passing through latitude 26° exceeded that passing through latitude 

 64° by nearly one second of arc. The plates, however, were few in number; 

 but ihey show, at least, the value of the method, and confirm to some slight 

 degree the non-spherical shape of the sun. 



The photographic tests made at the Yerkes Observatory established 

 the value of the photographic method and of the photo-he liometer; but 

 they showed conclusively that, in order to obtain satisfactory plates, two 

 essentials must be observed. 



1. The photographic lens must be carried on a separate and specially 

 designed equatorial mounting, so constructed that the shutter is independent 

 of the mounting. If attached to the mounting, the shutter will almost 

 invariably cause jars and consequent distortions of the image. 



2. Wet plates should be used. On the modern rapid dry plates the 

 edges of the image are not sufiiciently sharp for accurate measurement. 



Part III. Discussion of Observations and Results. 



THE FIGURE OF THE SUN. 



A New Tabulation of the German Heliometer Measures. — In 

 the investigation of the general subject of the figure and the possible varia- 

 tions in the size and shape of the sun, the writer was led to a re-discussion, 

 or rather to a re-tabulation, of the results obtained by Auwers from the 

 transit of Venus observations. 



In forming his means from which the heretofore given result was obtained, 

 Auwers kept together all observations made with a single instrument, and 

 thus observations of different years were grouped together. As arranged 

 by Auwers, these observations do not afford any indication of a change of 

 the relative diameters with the time. In order to investigate this point, I 

 re-arranged the series of observations as given by Auwers, arranging them 

 in order of the time without regard to the observer or the instrument. When 

 thus arranged, the observations fall into two series: one extending from 

 September, 1873, to January, 1875; the other, from May, 1880, to June, 

 1883. There is an isolated observation in July, 1877, another isolated one 

 in March, 1884, and tjvo short series in the latter part of 1884 and the be- 

 ginning of 1885. 



There is an uncertainty of some days in assigning a date to each de- 

 termination of the ratio of the solar diameters; for the value of the difference 

 between the polar and equatorial diameters (P.-E.), as given by Auwers for 

 each observer, is found by him as the mean result of a number of observa- 

 tions, extending in many instances over a period of a month or more. In 

 very few cases did an observer measure the polar and equatorial diameters 



