June i, 1882] 



NA TURE 



therefore probably best represented in prominence-spec- 

 trum. B and C, layers further from the sun, and there- 

 fore cooler, and therefore probably best represented in 

 spot-spectrum. 



If this be so, when'we can see the lines of these layers 

 we shall see something like this — 



The lines of A — the hottest layer — will be brightest 

 and shortest. 



The lines of B— the next cooler layer — will be less bright 

 and longer, and will also go down to the sun, on account 

 of the part of the layer at U, although it is unrepresented 

 at A, along the section X — V. 

 And so on with C. 



In an eclipse we have a condition in which the atmo- 

 spheric light is gradually withdrawn. The lines should 

 appear, therefore, in the order of their lengths ; that is, 

 the line which turns out to be longest should be the last 

 to appear, and this is a magnificent proof that the sub- 

 stance which produces the line does not extend down to 

 the sun, for if it did it should be brightest at bottom, and 

 should at first appear as a short line. 



Now what were the facts ? Dealing with the region 

 between F and b, and not all of that, and especially with 

 the three iron lines I have so often mentioned, this was 

 the order of appearance — 



May 17, 8.18 a.m., saw F and T x very short. 

 (Tj meaning the single iron line of the three w/j.9233 

 so constantly seen by Tacchini in prominences). 



8.19 ... F -f T l 4- 4933 short. 



8.20 ... F + T, + 4933 -+ b long. 



8.23 -. F + T, + 4933 + b + T 2 short. 



(T 2 meaning a high temperature iron line at Tf/50176, 

 constantly seen by Tacchini with 49233). 



At this time the darkness sensibly decreased, and then 

 for the first time several long thin lines suddenly burst 

 out. 



8.23.30 : Single iron line at 49565, and double at 4918 

 and 49195 and line at 49325, the last three being the 

 longest. Other long lines made their appearance, but 

 their positions were not absolutely determined. 



Totality was announced at 8h. 25m. 42s., and it was 

 arranged that I should then change my instrument. I 

 fancy the signal was given a little too soon, for when I 

 went to my 3J telescope to study the structure of the 

 corona the cross wires were still some distance from the 

 point at which the sun disappeared ; but be this as it 

 may, I missed the flash, but this was unimportant, the 

 real work was done. Still this is a point so crucial that 

 we ought not to be satisfied till all these changes, even 

 including the flash, have been photographed on a moving 

 photographic plate, an idea which struck me too late for 

 utilisation during the present eclipse. 



Next, as to the structure of the corona. Again the 



predictions were fulfilled ; we were in presence of a repe- 

 tition of the eclipse of 1871 ; everything special to that 

 of 1878 had disappeared. There was absolutely no struc- 

 ture near either pole. I was using the same telescope as 

 in 1878, when this feature was so marked, so there can 

 be no mistake on this point. The filamentous character 

 of the streamers, a marked feature in 1871, was however 

 not so decided. 



As with the structure so with the ring spectrum. The 

 rings so bright in 1871, so conspicuously absent in 1878, 

 were again visible, but with a Rutherfurd grating they 

 were not so obvious as I at all events expected to find 

 them. As seen at mid-eclipse, 1474 was the faintest ring, 

 and C the brightest. 



With regard to the spectrum of the corona as seen with 

 an ordinary tele-spectroscope, arranged to give as much 

 light as possible, I have not so much to say as I had hoped, 

 for the reason that the totality lasted longer than we 

 counted upon. The result of all the preliminary pour- 

 parlers had been to fix upon sixty-five seconds as the most 

 probable duration of totality, or rather as the duration to 

 be provided for especially from the photographic point of 

 view, since a photograph exposed during totality would be 

 ruined if the sun reappeared before the cap of the camera 

 had been replaced. Sixty-five seconds having elapsed 

 from the announced commencement of totality, I went 

 to the corona spectroscope which I should have gone to 

 ten seconds earlier (but the comet had taken five seconds, 

 and the grating observation had been more uncertain 

 than I had expected) At the moment I made the obser- 

 vation the eclipse was over, but still I noted F, and 1474, 

 and C, bright, and extending right across the field, and a 

 banded spectrum, that is to say, not a continuous spectrum 

 certainly, but into maxima and minima, though the 

 minima gave no signs of dark lines. The observation 

 however, was almost instantaneous, and too much im- 

 portance must not be attached to it. 



Here my notes must close for the present ; 104 in the 

 shade is not conducive to writing, even if camels were 

 not growling, and flies teasing, as they can tease in 

 Egypt. J. Norman Lockyer 



Siout, May 21 



(To be continued.) 



BIOLOGY AND AGRICULTURE 

 DECENT advances in our knowledge of the lowest 

 -"-V forms of life have tended to bring into prominence 

 not only their relation to disease but to the ever-increasing 

 importance of the part which they play in our arts and 

 industries. Probably in none of the industrial arts, save 

 those concerned with fermentation, commonly so called, 

 has the progress of this branch of biology shown such 

 remarkable development as in its bearing on the art of 

 agriculture. 



It has even been suggested that a bacterium is at the 

 bottom of the present state of agricultural depression, 

 and there is a considerable amount of force in this 

 suggestion. The loss of nitrogen from the soil in the 

 form of nitrate is one of the most serious difficulties with 

 which the farmer has to contend ; and, as this loss takes 

 place by the washing out of nitrates in the drainage and 

 its diffusion into the subsoil below the reach of the 



