M-uicH 1, 1895.] 



KNOWLEDGE. 



59 



to be individually discerned, but producing in the aggregate 

 a certain definite effect. The formation or dissipation of 

 an amount of light ckrus cloud, which elsewhere would 

 be unnoticed, might easily, over regions like these, 

 change what appeared to be sea into what looked like land, 

 and vice versa, and this in a few mmutes of time. Indeed, 

 since eyes are differently sensitive to dehcate differences 

 of shade, the same reg"ion at the same instant might 

 appear to one observer as land, to another as sea. Except 

 on some such principle as this, it seems hard indeed to 

 account for the repetition, at long intervals of time, of 

 changes of light regions to dark and dark to light, not 

 mdiscriminately all over the planet, but in certain com- 

 paratively small areas of well-known and definite outline. 



Hcttcrs 



[The Editor does not hold himself responsible for the opinions or 

 statements of correspondents.] 



• 



To the Editor of Knowledge. 



Sir, — In an article called " Stars and Lights," by the 

 Rev. Charles Pritchard, in Good Words for May, 1869, 

 the writer states that Herschel, obser\'ing the star Mira. 

 found that it reached its greatest lustre on November 

 2nd, 1779, and on October 21st, 1790. He goes on to 

 say that the interval between these two dates gives eleven 

 periods of 331 days, but that interval exceeds the stated 

 periods by 365 days. The year 1790 being three times 

 mentioned in connection with the latter observation, a 

 student would not feel justified in correcting to 1789, while 

 a correction of the former observation date to 1780, a leap- 

 vear, would still throw the calculation out to the extent of 

 one day. 



Could you kindly say whether 1789 should be substituted 

 for 1790 •? Yours faithfully. 



Maisonette, Wallington. Thomas Henchmax. 



[The late Prof. Pritchard was a very accurate writer, 

 but even accurate writers occasionally make mistakes ; 

 moreover, it is always diiScult to be secure against 

 printers' errors in a popular periodical. 



In the present case it is evident that Prof. Pritchard 

 wrote, or intended to write, twelve (not eleven) periods 

 between the maxima of 1779 and 1790. W. Herschel's 

 observation of the former is interesting as being given in 

 the first paper he ever sent to the Royal Society, which 

 was communicated by Dr. Watson of Bath, and read on 

 May 11th, 1780. It will be noticed that he does not fix 

 the exact day of the maximum in the preceding year, but 

 says that on November 2nd the star's lustre was "still 

 increasing " ; on November 20th it was " as bright as 

 before, but no brighter '' ; and on November 30th it had 

 " considerably decreased." In the account given amongst 

 his " Miscellaneous Observations" (read on December 22nd, 

 1791), he says that he takes the epoch of October 21st, 

 1790, as " one of the best ascertained moderate appear- 

 ances " he could obtain, and that he believes it to be 

 " more proper for settling the period than that which 

 might be deduced from a brilliant blaze of the star, such 

 as took place in 1779, owing to causes that are not regular, 

 and therefore may be apprehended to disturb the general 

 order of the change." Sir \YUliam deduced what he 

 thought the most likely period from the observation of a 

 maximum by Fabricius, August 13th, 1596, and this 

 of October 21st, 1790, contending that its length 

 was 331 days 10 hoitrs 19 minutes, though others had 

 made it as much as 331 days. Sir John Herschel, we 



may add, puts the average period at 831 days 8 hours 

 4 minutes ; but it is undoubtedly subject to some amount 

 of irregularity. W. T. L. 



Blackheath. 



0\ THE ELECTRIC ORIGIN OP THE SOLAH 

 CHROMOSPHERE. 



To the Editor of Knowledge. 



Dear Sir, — As you have kindly communicated to me 

 Mr. Evershed's answer to my former letter, I now send 

 you some further remarks. 



1st. Sodium vapour. — The conclusions adopted by Mr. 

 Evershed, after his last explanations, appear to be more 

 probable ; but, though the experiments related have a 

 great value, his proofs, in my opinion, are not complete. 



Mr. Evershed relies on the difference in the width which 

 the yellow D line shows in the heated tube and in the flame ; 

 but is not the greater width in the tube due simply to the 

 greater proportion of sodium, which is slowly distilling 

 from a hot to a cold part of the tube, whereas in the flame 

 it is rapidly drawn out '? 



Mr. Evershed supposes that hydrogen gas has no 

 influence on the production of the characteristic line D ; 

 how then can he explain the familiar experiment of 

 Prof. Lockyer on sodium when heated and electrically 

 illuminated in a vacuum tube '? ( Proceedi)i;is of the Royal 

 Soriettj, 1879, p. 110) : " Hydrogen is given off in large 

 quantities, and at the end the vapour shows the red and 

 green lines, without any trace whatever of the yellow 

 one." 



These questions are very difficult and complex, even 

 with the apparently simple case of sodium. For my part, 

 I think it is almost impossible to decide, at very high 

 temperatures, that no chemical combinations or allotropic 

 changes are taking place. 



2nd. Hydrogen. — Application to the solar chromosphere. — 

 The experiments on sodium are not sufficient for applica- 

 tion to the case of the solar chromosphere, because sodium 

 is not the only constituent of the latter ; hydrogen gas, 

 which is mucli more imporant than sodium in the sun, 

 and which exists also in the stars and nebul:?, should be 

 also examined. As yet, however, it has not been found to 

 give its characteristic lines by heat or by chemical com- 

 bination after a great many experiments tmder different 

 conditions— an electric interference is necessary ; so that 

 we may conclude that the chromosphere is an electric 

 flame. 



Some observers, including Prof. Tacchini and Dr. 

 Huggins, have also set forward the same opinion or con- 

 clusion, but on other grounds. Dr. Huggins, in his 

 Bakerian lecture in 1889, concludes from a general and 

 remarkably able study of the solar corona that the 

 phenomenon is electrical, and, extending this result, he 

 adds, " We can scarcely doubt that electric disturbances of 

 exceptional magnitude accompany the formation of the 

 prominences ; indeed, these phenomena may themselves be, 

 in part at least, electric discharges analogous to terrestrial 

 aurorsB." . . 



From aU points of view, then, an electric origin is, in the 

 actual state of our knowledge, decidedly the most prob- 

 able. Moreover, the new photographic method, which 

 reveals the chromosphere on every point of the solar disc, 

 has permitted us to take a further step, and to compare 

 exactly the electric states of terrestrial and solar atmos- 

 pheres. Now the distribution of electricity appears exactly 

 the same in the two bodies, and I wrote in 1893 {Comptes 

 rendus de I'Academie des Sciences de Paris) : " The 

 phenomena of solar chromosphere and terrestrial atmos- 



