March 22, 1906] 



NA TURE 



487 



spectrum. All these will be of great interest to other 

 observers of the same eclipse, as comparison of results 

 is of great importance in the interest of future eclipse 

 work. 



It would lead one too far, and, indeed, it is not 

 necessary, to enter into the very complete treatment 

 here published of the various branches of work so 

 well brought together, but perhaps a word or two may 

 be mentioned with regard to some expectations that 

 were not so successfully realised as was hoped. 



In the case of the io-feet concave grating erected 

 at Griffin, the plates turned out to be very much 

 under-exposed. In explanation of this, Prof. Crew 

 not only summarises the possible causes of failure, but 

 adds some useful suggestions for use on future 

 occasions. In the case of the former he mentions 

 seven possible causes, the first being that the intensity 

 of radiation of the eclipsed sun was underestimated, 

 and consequently only the very brightest lines of the 

 chromospheric spectrum were recorded. The second 

 was that the effect of astigmatism in the curved 

 grating was underestimated. 



the ultra-violet and of silver for light of longer wave- 

 lengths." 



With regard to the plane grating spectrograph 

 used at Pinehurst, Mr. Jewell states that the definition 

 of the lines was good from wave-length 3850 to 4100, 

 and " remarkably fine near the H and K lines." The 

 definition between 3750 and 3850, and from 4100 to 

 4200, is described as "fair," but "very poor" at 

 wave-lengths less than 3700 and greater than 4200. 

 A complete and long table of wave-lengths of the 

 chromospheric lines measured is given in the report. 



The coronal lines observed were six in number, the 

 mean wave-lengths of which were as follows : — 



3382-4. 3453-3- 3644-0. 3S°!-8. 3987-5. an d 53°4-i- 

 The objective prism spectrograph at the same station 

 also secured chromospheric and coronal spectra, and 

 although a long list of the wave-lengths of the lines 

 is published, great weight cannot be given to their 

 accuracy, since the definition on the negative is de- 

 scribed as "poor over the whole spectrogram and par- 

 ticularly poor for the violet end." 



Although no word has vet been said about other 



-The U.S. Naval Observatory Eclipse St; 



Pinehurst, North Carolina. The 40-foot Coronagraph i 



A third reason was that an exposure of two plates 

 instead of one during totality was more than was 

 justified by the dispersion and slit width. 



In Mr. Humphrey's report relative to the 21.5-feet 

 Rowland grating at Griffin, a very dissatisfied tone is 

 pronounced with regard to the instrument's efficiency. 

 He finally states : — " It does not seem probable, how- 

 ever, that a grating mounted according to the Row- 

 land method and used with a slit can, even under the 

 most favourable circumstances, vield nearly as many 

 flash lines as may be obtained with a prism or a 

 concave grating used directly without lens or slit. 

 ... It would be well in using a spectrograph of 

 this kind to avoid the chromatic aberration of lenses 

 entirely and to form the image on the slit by means 

 of suitable reflectors. It might also be best to use 

 silver reflectors and to avoid that part of the ultra- 

 violet of wave-length shorter than A 3600. Direct 

 grating or prism spectrographs should be used, if 

 possible without reflectors of any kind, but when a 

 reflector is used it should consist of magnalium for 



NO. f S 9 9 , VOL 7$] 



branches of work taken up, the reader must be re- 

 ferred to the volume itself for the numerous reports 

 on them. We reproduce in Fig. 1 (upper part) a pic- 

 ture of the corona as photographed with the 40-feet 

 camera at Barnesville, the exposure of which lasted 

 thirty-five seconds. It will be seen that it illustrates 

 well the " wind-vane " form typical at epochs of sun- 

 spot minimum. The lower part of this illustration 

 shows the corona of 1901, taken also with the 40-feet 

 instrument. Fig. 2 shows a general view of the 

 station selected at Pinehurst, the method of support- 

 ing the 40-feet coronagraph, and the houses for the 

 other instruments erected there. 



The second portion of this important volume deals 

 with the observations made during the solar eclipse 

 in Sumatra in May, 1901, the stations selected being 

 Solok, Fort de Kock, and Sawah Loento. At the 

 first station clouds reigned supreme during totality, 

 but the chromosphere was photographed at third con- 

 tact through clouds. At the second station the 

 weather is described as " perfect," while the observers. 



