340 



♦ KNOWLEDGE * 



[April 24, 1885, 



matter, a burnished edge, he is thinking of, not the sun 

 himself t^rought into view in the form of a ring all round 

 the earth, as is really the case. The fiue ring of light, 

 which he compares in area with the whole surface of the 

 eirth (more effectively illuminated, as he supposes) is in 

 reality not an illuminated ring at all, but self-lumiuous ; 

 its brilliancy is not the same as (or, as Mr. Williams 

 supposes, even much less than) the brightness of the sunlit 

 earth, but hundreds of thousands of times greater. This is 

 not a point open to doubt, and therefore requiring demon- 

 stration, though it may need (and indeed appears to need) 

 explanation. But if proof were needed it would be found 

 in the corresponding case of the arc of light seen round 

 Venus during transit in 1874, and partially during the 

 transit of 1882, when however the air of Venus would 

 seem to have been more cloudy and less of it therefore 

 available for the work of bending sun rays : for the arc of 

 light left its photographic record as quickly and sharply as 

 the light from the sun's disc, showing that it was not 

 reflected sunlight, but the sun's own light slightly deflected 

 only on its course. 



The actual condition of things during the time of total 

 lunar eclipse is, I find (from talking the matter over with 

 many), very little understood. For instance many speak 

 of a ring of light round the moon as if a total lunar 

 eclipse were always central, which is far from being the 

 case ; or as if at any rate some point of the sun's iace were 



directly beyond the earth's centre as supposed to be seen 

 from the moon, which again is not the case. And then 

 the way in which different layers of our atmosphere work 

 to raise the sun into view is not at all understood by 

 certainly 999 out of 1,000 who have read the usual explana- 

 tion of lunar eclipses. Fig. 1 represents the case of a 

 central total eclipse — central, however, only as supposed to 

 be seen from a single point on the earthward lunar face. 

 Fig. 2 represents the case of a non-central total eclipse. 

 In one case the sun is distorted into a uniform ring, in the 

 other into two meniscus arcs of unequal extent, and of 

 unequal breadth. 



But let us examiue the matter a little more closely. 



The refractive power of our atmosphere is such that 

 under favourable conditions the setting sun is raised about 

 35 minutes of arc or by more than his own diameter. 



Doubling this, we find the greatest amount of deflection of 

 solar rays in passing into our earth's atmosphere and out of 

 it, on its way towards the moon, to be about 70' of arc. 

 Now the earth as seen from the moon appears nearly two 

 degrees, or more exactly, about l-j",, deg. in width. So that 

 if the sun were a bright point just beyond the earth's centre, 

 as looked at from the moon, he would only require to have 

 his rays bent round .', "ths of a degree, or 57', to be seen as 

 a ring all round the earth. With the actual deflecting 

 power of the earth's atmosphere, he would be more than 

 sufficiently lifted into view, which means that instead of 

 appearing as a ring just touching the earth's edge, he would 

 appear as a ring slightly separated from the edge of the 

 earth's disc. 



Appearing really as a disc, 32 minutes of arc in diameter, 

 as seen from the moon's distance, the sun when centrally 

 behind the earth as in Fig. 1, would come within 41 minutes 

 (57 less 16) of the earth's edge, all round — that is the dis- 

 tances S s, S' s', would all be 41 min. The earth's air would 

 lift the sun then well into view. But not only so, parts 

 hei/ond the centre, so to speak, would be lifted into view. 

 A point as A, on C S' for example, would be lifted into 

 view not only towards s' but towards «. This would hajipen 

 close up to S', though S' itself, being 73 minutes (57 plus 

 16) from s, would not be lifted into view near s, — at least 

 not usually. When the earth is at her farthest from the 

 moon so as to look smallest, and the sun at hia nearest so as 



Fig. 2. 



to look largest (as at or about Jan. 1) S' would be brought 

 into view towards s as well as towards »•' ; only at s this 

 ))oint of the sun would be seen on the very verge of the earth's 

 disc, while at s' it would be raised above the earth's edge. 

 Usually, all the points along a line S C A of the sun's 

 face (A close to S') would be brought iuto view at s, in the 

 form of a very short line s a, extending from the earth's 

 edge at a to a height of perhaps about 2^1 miles. (The 

 height of the point s would be determined as that at which 

 the atmosphere would have power to deflect the sun's rays 

 41 minutes, or when the horizontal refraction would be 

 about 20.', minutes. I know of no experimentally-obtained 

 formula— and no other would be trustworthy — which would 

 show at what height above the earth's surface the sun on 

 the celestial horizon would be raised any given amount by 

 refraction.) A line as B S' on the sun's face would be seen 



