November 19, 1908] 



NATURE 



negatives showing streamers two solar ''iameters in 

 length. These long streamers are rather evenly dis- 

 tributed around the sun, but the strong inner corona 

 is much more intense on the east and west than at 

 the poles. 



Prof. Campbell and Dr. .\lbrecht speak very highly 

 of the simplicity and the performance of the 40-feet 

 " tower camera " of 5 inches aperture, designed by 

 Prof. Schaeberle for the 1893 eclipse. Fig. i Is repro- 

 duced from an illustration accompanying the former's 

 report in the Publications of the Astronomical Society 

 of the Pacific, and shows the instrument, in position, 

 at Flint Island. Many advantages are claimed for 

 this form of "direct" instrument; among them the 

 removal of the lens and a greater part of the tube 

 from the ground region of intense atmospheric tremor, 

 and the elimination of the possible distortion of a 

 coL'lostat mirror are not the least. Six exposures, 

 varying from 2 to 64 seconds, were made, and the 

 negatives show a wealth of coronal details. The 

 chromospheric stratum is shown on the first and last 

 photographs, and there is a large eruptive prominence 

 in position-angle 214°. About fifteen streamers ex- 

 tending to i^ lunar diameters and a similar number 

 extending to i lunar diameter are shown on the 

 longer exposures. An unusuallj' straight and slender 

 streamer starts near the south point and extends, not 

 radially, but in the direction of 190°, for about 

 li diameters; during the last seven-ninths of its length 

 this streamer is accompanied by a fainter branch 

 which makes an angle of 4° with its axis. Compari- 

 sons of the photographic brightness of the corona, 

 effected by impressing standard squares on some of 

 the plates, show that almost all the effective photo- 

 graphic light came from the inner corona situated 

 within i' or 2' of the moon's edge. 



A plate exposed, in a spectrograph with the plate 

 continuously moving, for fifteen seconds on either side 

 of the end of totality, shows a spectrum in excellent 

 focus from A 3800 to \ 5100 the linear scale being such 

 that from a 3-00 to a 5300 is 13 inches. Hundreds of 

 bright lines, showing the depths and locations of the 

 corresponding vapours, are recorded, and should fur- 

 nish a wealth of information as to the sun's higher 

 atmosphere. 



A spectrograph fitted with quartz lenses, and 

 prisms, of 9*2 cm. aperture and i metre focal 

 length, was employed by Prof. Lewis to obtain photo- 

 graphs of the spectra of the corona and the " flash." 

 A sliding diaphragm in front of the slit permitted 

 comparison solar spectra to be obtained on the same 

 plate, and an exposure of 3:5 minutes on the corona, 

 with the slit tangential to the east limb, and a 

 shorter exposure on the chromosphere were made 

 immediately before, and at, third contact. The un- 

 expected curtailment of totality caused the latter 10 

 be over-exposed, but some eighty chromospheric lines 

 are seen projecting beyond the solar spectrum ; the 

 resulting spectra are some 14 cm. long from A 3000 to 

 D. Both the continuous and the line radiations of 

 the corona are shown, and both apparently emanated 

 from the inner corona within a region of less than a 

 quarter of a radius from the photosphere. Fifteen 

 certain, and ten doubtful, lines are recorded, and of 

 the former those at AA 5304, 3602*3 and 3390*7 are the 

 strongest, the respective intensities being 10, 15 and 

 30. That neither Huff, in 1900, nor Dyson, in 1900, 

 1901 and 1905, recorded the strong line at A 3602*3, and 

 that the latter obtained lines not shown on his own 

 spectra, is taken by Prof. Lewis as evidence 

 supporting the idea of a variable corona. 



In discussing the relative temperature of the corona. 

 Prof. Lewis employs the principle enunciated bv 

 .Stokes in 1876, and used by Lockyer in his researches 



NO. 2038, VOL. 79] 



on stellar temperatures. This principle depends upon 

 a comparison of the relative amounts of the ultra- 

 violet and the red radiations in the light-sources 

 examined, predominance of the former denoting 

 higher, and of the latter lower temperature. In 

 Lockver's stellar work it was found that by the 

 assumption of this law the previous results depending 

 upon the chemical classification of the stars were 

 plenarilv confirmed, and Prof. Lewis's conclusions 

 are no less regular. Comparing the coronal with the 

 solar spectrum, he finds that the latter is, relatively, 

 much richer in violet light, and says, " hence it may 

 be inferred that the corona is considerably colder than 

 the sun." Subsequent comparative tests with a 

 standard candle, allowing for the atmospheric absorp- 

 tion of the ultra-violet radiations, fix the lower limit 

 of coronal temperature at considerably more than 

 2000° absolute. 



Fig. I. — The 40-ft. Coronagraph. 



Some eighty-two chromospheric lines, between 

 AA 3i99'8 and 48630, with identifications from Row- 

 land's tables and Dyson's eclipse paper, are given, 

 but, with only the tips of the lines projecting from the 

 dense, o\*er-exposed spectrum, the wave-lengths are, 

 naturally, only approximate. 



For the examination of polarisation effects in the 

 corona, four exposures, varying from two to six 

 seconds each, were made with each of the four 

 cameras arranged for this purpose. Each photograph 

 shows marked polarisation in all parts of the corona, 

 and a comparative study of the series should, incident- 

 ally, produce valuable information concerning the 

 effect of clouds on such photographs, for during one 

 exposure a cloud passed over the southern half of the 

 corona, leaving the northern half cloud-free. 



