March 25, 1880] 



NATURE 



489 



they are always present, but only cast a visible shadow 

 when the solar light is confined to a narrow crescent, we 

 do not know. These bands were noticed a short time 

 ago, without an eclipse, in Lord Lindsay's observa- 

 tory, and were then produced by the sun disappearing 

 behind the mountains. The eleventh chapter treats of 

 the brushes of light apparently emanating from the cusps 

 of the sun during the partial phase. Prof. Stokes has 

 offered an explanation of this phenomenon, and the ex- 

 planation seems to agree very well with the drawings 

 given by some of the observers. It cannot, however, be 

 reconciled, we believe, with the description given by Mr. 

 Brett, who observed them through his telescope and 

 called them "exquisitely defined." 



The subjects of greatest interest and importance'of the 

 volume are contained in the three last chapters, which 

 form about two-thirds of the whole volume. They treat 

 of " Polariscopic Observations," " Spectroscopic Obser- 

 vations," and " Photographs and Drawings of the Corona." 

 A good part of the chapter on polarisation may serve as 

 an example of what the book might have been, had Mr. 

 Ranyard more frequently indulged in theoretical specula- 

 tions. We cannot of course go into any detailed criticism, 

 but shall give two examples in order to justify our remarks. 

 In the first place, we think that Mr. Ranyard has treated 

 the observations which relate to the polarisation of our 

 atmosphere in the parts of the sky occupied by the corona, 

 in a rather careless manner. It appears that one observer 

 makes the polarisation Vertical, one horizontal, and seven 

 leave it doubtful whether the polarisation was vertical or 

 horizontal. But on p. 255 all the observations which 

 leave the matter doubtful are given as proving vertical 

 polarisation, and about the one contradictory observation 

 we are told that " it must be borne in mind that a half 

 rotation of his instrument would be sufficient to leverse 

 the position of the colours." 



We are then informed that vertical polarisation is 

 exactly what " we should expect." But in a footnote on 

 P a ge 330 we are treated to some vague considerations, from 

 which it "evidently" appears that the polarisation ought 

 to be horizontal or exactly opposite to what we were first 

 led to "expect." And now we are informed that out of 

 eight observations which have been' adduced to prove the 

 first statement, seven prove nothing, and the observation 

 which we first had to set aside on account of the " half 

 rotation "is now treated as the .only good one. Of the 

 one observation which contradicts this new result we are 

 told that " it seems, perhaps, more natural to assume that 

 he commenced his observations with the polarimeter in a 

 vertical rather than in a horizontal position." 



Mr. Ranyard adds that this note must be taken as 

 overriding the opinion expressed on page 255. 



What wc should " expect " is not at all so " evident " as 

 Mr. Ranyard makes it out to be. If we confine ourselves 

 to clouds, as Mr. Ranyard does, or to the reflection of 

 sunlight from the surface of the illuminated parts of the 

 earth, the polarisation would altogether depend on the 

 distribution of the clouds, it would in all cases be very 

 small, and a small change in the distribution might change 

 the polarisation from one plane to the other. The polari- 

 sation due to reflection from unobserved parts of the 

 earth or clouds uniformly covering it, ought to be vertical 

 a s we might "expect," in Mr. Ranyard' s first statement. 



and not horizontal, as is " evident " in his second. If the 

 clouds are higher up a different result is possible. There 

 is, however, a second and much more powerful cause of 

 polarisation which is due not to clouds, but to particles 

 which scatter the light without polarising it, but are not 

 sufficiently dense to make the medium opaque or to form 

 a cloud. It is well known that in the neighbourhood of 

 the sun the light is always polarised in a horizontal 

 plane, and the cause we have mentioned seems to be 

 the only one which can account for the facts. Exact 

 observations during total eclipses would be of great use, 

 and might throw some more light on the subject, which 

 has by no means been fully cleared up. A great deal 

 more may be said on this question, but we have only 

 made the foregoing remarirs in order to show that the 

 subject cannot be treated in the short offhand way adopted 

 by Mr. Ranyard. 



Our second remark refers to a somewhat clumsy 

 mathematical investigation. Mr. Ranyard investigates 

 the polarisation of the light scattered from an atmosphere 

 of particles uniformly distributed within a spherical shell 

 surrounding the luminous point. We may mention 

 that no conclusions whatever can be drawn from this 

 problem with regard to the corona. One of the sup- 

 posed results, however, of the investigation is that the 

 light coming to us in the direction of the luminous point 

 itself is polarised, and on this result Mr. Ranyard 

 remarks — "At first sight it might appear that at the 

 central part of the corona the light should be entirely 

 unpolarised, but it must be remembered that the illumi- 

 nation of the particles adjacent to C will be very 

 great, and the polarisation of the light dispersed by 

 them will, it appears, overpower the non-polarisation 

 of the light dispersed by the higher parts of the 

 corona." In this passage the atmosphere of particles 

 referred to above is called the "corona," and C is 

 the luminous point. Had Mr. Ranyard ever a=ked 

 himself the question in what plane the polarisation in the 

 central part of the corona ought to be, the above passage 

 could never have been written, for, reasons of symmetry 

 would have been sufficient to show that no such polarisa- 

 tion is possible. Nor is the reason why the formula 

 breaks down in this particular case far to seek. 



We turn with pleasure to the chapter on spectroscopic 

 observations, not that we consider this chapter faultless, 

 but because in it we have brought before us a greater 

 array of facts and net so much theory. For any one who 

 in future eclipses intends to follow up the results achieved 

 with the spectroscope by previous observers, this chapter 

 will be invaluable. In no case is the comparison of 

 different observations of such importance as in spectro- 

 scopv, and we are glad to have before us all the evidence, 

 positive and negative, on a great many partly still 

 undecided questions. The last and longest chapter is in 

 some respects the most interesting. In it we have a 

 careful comparison of the more important drawings and 

 photographs of the corona, made during a long series of 

 eclipses. The symmetry of the corona about an axis, 

 nearly coincident with the solar axis, cannot be doubtful 

 any longer, but whether the axis of symmetry is really the 

 sun's axis or only approximately coincident, is one of the 

 important points which will have to be decided during 

 future eclipses. 



