EXPERIENCES IN SUMATRA. 295 



The principal work of preparation was accomplished nearly a week 

 before the eclipse. The remaining time was devoted to arranging the 

 final details and to training the assistants. As soon as it was known 

 that help would be needed to make the observations, it was tendered 

 in abundance by the Dutch residents. Fifteen, including several of 

 the army officers, and the general manager of the government railways, 

 were invited to assist. The observations with each instrument were 

 very minutely planned. Every motion to be made by each observer 

 was carefully considered and arranged beforehand, and numerous re- 

 hearsals enabled the operations of taking the photographs to be per- 

 formed with certainty and great rapidity. 



Although light clouds covered the sun during totality, the resulting 

 observations were very successful and yielded much important infor- 

 mation. Photographs were secured with three sets of cameras espe- 

 cially designed to record the inner, middle and outer corona. The 

 large-scale photographs of the inner corona showed a peculiarly dis- 

 turbed region, which later was found to have been at the time of the 

 eclipse above a group of sun spots. The inference that there was a 

 close connection was irresistible. 



This eclipse was particularly favorable on account of the great 

 duration of totality — six and one-half minutes — for the search for any 

 planets between Mercury and the sun. Four cameras of long focus 

 were especially designed for the search. They were capable of record- 

 ing any such objects as faint as the ninth magnitude. The photo- 

 graphs were taken in duplicate to guard against defects. They cov- 

 ered a region 6° wide and 30° long, in the direction of the sun's 

 equator, as the most probable orbit of such a planet would lie in that 

 region. Although the search was not so complete as desired, owing 

 to the clouds, our knowledge was considerably extended. In two 

 thirds of this area, stars fainter than the eighth magnitude were 

 photographed and in the remaining third, stars of sixth magnitude 

 and brighter. All the objects on the plates were identified as known 

 stars. It is practically certain from the observations, therefore, that 

 there is no such planet as bright as fifth magnitude, and little prob- 

 ability of there being any as bright as eighth magnitude. A planet 

 thirty miles in diameter in the region searched would appear as a star 

 of the eighth magnitude. To account for the observed deviations of 

 Mercury from its computed orbit would require over half a million 

 bodies 30 miles in diameter and as dense as Mercury. 



Another investigation undertaken concerned the nature of the 

 coronal light. The bright coronal lines observed at previous eclipses 

 showed that a small portion, at least, of the light was due to in- 

 candescent gases. But the great portion of the light gave a spectrum 

 which appeared to be perfectly continuous. Some observers had 



