DISCOVERY 



49 



drawn to scale : if it were, and se were taken as repre- 

 senting 10 inches, then the diameter of the sun would 

 be 1 Vth inch, and the distance sp, for a near star, would 

 be something like fifty miles. Then the result which 

 has been estabUshed is that p, L, e do not lie in a straight 

 line, the rays of light from the star being bent by the 

 gravitational action of the sun, so as to consist practi- 

 cally of two straight portions PL, le. The raj's coming 

 to the observer along the direction le cause the star to 

 appear at p' in el produced. The ray from the star 

 if it were not bent would come to E' instead of e, and 

 the distance ee' may be used as a measure of the 

 deflection of the light. On the scale which has been 

 mentioned above, in which se is lo inches, the displace- 

 ment ee* would be about the ten-thousandth part of 

 an inch. E.xpressed otherwise, the angular displace- 

 ment, which may be measured by either the angle 

 ele' or the angle plp', is 175 seconds of arc. A 

 second of arc is the angle which would be subtended 

 at the eye by a halfpenny at a distance of three 

 miles. 



These figures apply in the case of a star, the light 

 from which just grazes the edge of the sun. For 

 stars appearing further from the centre of the sun the 

 deflection is less, its magnitude varying in\'ersely with 

 the distance from the sun's centre. Thus, for instance, 

 if the apparent distance of the star from the sun's 

 limb or edge is equal to the sun's radius, the deflection 

 will be o''8y, i.e. I'^ys di^ded by two. If it is at twice 

 this distance from the limb (i.e. at a distance equal to 

 three radii from the centre), the deflection will be o'''62. 



Owing to the great brightness of the sun, it would of 

 course ordinarily be quite impossible to see stars so 

 near the edge of the sun. Moreover, the amount of the 

 deflection anticipated (and predicted by Einstein) is 

 so small that the only body with a large enough mass 

 to use for the purpose is the sun. It would require a 

 precision far exceeding the highest so far attained in 

 laboratory measurements before the effect could be 

 established bj' means of laboratory experiments. The 

 only remaining possibility is to take advantage of a 

 total eclipse of the sun, when the moon comes between 

 the earth and the sun, blotting out the light from the 

 latter. Wlien a total eclipse occurs, complete dark- 

 ness does not ensue, as there is then seen around the 

 sun a fringe of bluish-white hght called the solar corona 

 (see Plate) , which has a brightness somewhat exceed- 

 ing that of the full moon, and which cannot be seen at 

 any other time. The nature of the corona is not com- 

 pletely understood, as its shape varies considerably 

 from one eclipse to another. It is known, however, 

 to consist of gaseous matter in a state of great rare- 

 faction ; the principal element in it has not yet been 

 discovered on our earth, and has therefore been 

 termed coronium. 



Even when a solar eclipse occurs, the stars near the 

 sun will not in general be sufficiently bright to be 

 visible to the naked eye, and we must call to our aid 

 the photographic plate. By giWng an exposure of 

 sufficient length, the plate will accumulate the light 

 impressions which it receives, and so render even faint 

 stars visible. 



Total solar eclipses do not occur very frequently ; the 

 last occurred on May 29, 1919, and the next will occur 

 on September 21, 1922. Moreover, they are only visible 

 over a very limited portion of the earth's surface. The 

 apparent diameters of the sun and moon are so nearly 

 equal that the shadow of the moon cast by the sun at 

 an eclipse is in the form of a very elongated cone, whose 

 point or apex is near the earth's surface ; the section of 

 this cone by the earth is therefore a region not covering 

 very many square miles. As the moon and earth move 

 relativel}' to the sun, the shadow traces across the 

 earth's surface a narrow band, and it is only at points 

 on the earth's surface within this band that a total 

 eclipse of the sun is visible . 



The total eclipse of May 29, 1919, was a peculiarly 

 good one for the purpose of these observations, as the 

 sun then happened to be in a region of the heavens 

 containing an unusually large number of bright stars. 

 In fact, if the course of the ecliptic — i.e., the apparent 

 path of the sun in the heavens during each year — be 

 traced on a star map, it will be found that no more 

 favourable opportunity could have occurred. It will 

 not occur again for a few hundred years. The eclipse 

 track on this occasion crossed Brazil and the Atlantic 

 Ocean, passed over the Island of Principe, and crossed 

 Africa from Cape Palmas in Liberia to Lake Tanganyika. 

 It was arranged by the Joint Permanent Eclipse Com- 

 mittee of the Royal and Royal Astronomical Societies 

 that two expeditions should be sent out to secure the 

 necessary observations, and in order to reduce the risk 

 of bad weather spoiling the observations, it was 

 decided that one party should proceed to Sobral, in 

 the State of Ceara, Brazil, and another to Principe. 

 The conditions in Liberia appeared too unfavourable, 

 although, as events turned out, the conditions there 

 during the eclipse were much superior to those which 

 obtained in Principe. The observations at Principe 

 were spoiled by light cloud, which partially obscured 

 the sun throughout totality ; the plates taken were 

 useful, however, as serving to confirm the results 

 obtained by the Sobral observers. At Sobral, the 

 region of the sky around the sun was clear for the 

 greater part of totality, but for the remainder it was 

 veiled by thin cloud. 



Two series of photographs were obtained with 

 two different instruments. For the sake of brevity, 

 we will refer only to the better series of the two. A 

 photographic camera was used whose lens had an 



