June 1, 1898.1 



KNOWLEDGE 



133 



The long exposures made near mid-totality show the 

 distribution of " coronium " in the corona. The well- 

 known radiation of this gas, 1 171 K, is, in these, shown 

 to correspond more or less with the general structure of 

 the corona ; but it is very much brighter on the east side 

 than on the west, where it is hard to trace it at all. 



Six out of the ten photographs taken are reproduced in 

 the plate, enlarged about two and one-third times. The 

 series number is given beside each spectrum, iluch ot 

 the delicate detail shown on the original negatives is 

 inevitably lost in the reproduction, although the main 

 features are well brought out. In referring again to the 

 flash spectrum, as seen in Nos. 3 and 7, I would call 

 attention to the beautiful sequence of the hydrogen lines 

 (the positions of these referring to No. 3 are given at the 

 top of the plate). I do not know that these lines have ever 

 before been photographed in such completeness in the 

 chromosphere spectrum. In the original negative twenty- 

 six or twenty-seven lines can be counted, starting with the 

 line a at the red end. In the ultra-violet they become so 

 closely crowded together that it is not easy to say exactly 

 how many there are and where the series ends. According 

 to the well-kuowu empirical formula of Balmer, which 

 expresses with such extraordinary accuracy a series of 

 numbers given by nature, the limit should be at >. 3(ji7, a 

 position in the spectrum which is indicated by an arrow 

 at the top of the plate. In the photograph, however, 

 there is a beautifully regular gradation in the intensity 

 of the lines, which become fainter and fainter as this limit 

 is approached, so that line No. 27 (A SCUM) is so exceed- 

 ingly faint as to be barely distinguishable, and it apparently 

 forms the termination of the series. 



Another point which is well shown in No. 3 is that, if 

 we except the ordinary chromosphere lines (those of 

 calcium, hydrogen, and helium), all the fainter lines due to 

 the flash spectrum proper are of the same length and form, 

 a well-defined band of even width running from end to end 

 of the spectrum. This shows that the low-lying gases at 

 the base of the chromosphere form a well-defined layer 

 pretty definitely bounded, and not fading by insensible 

 gradations into the higher portions of the chromosphere. 



An estimate based on the width of this band of bright 

 lines gives for the depth of the layer about one and a 

 half seconds of arc — or say seven hundred miles — the total 

 depth of the chromosphere itself being some eight seconds 

 of arc, or three thousand six hundred miles. 



The limits of this article preclude my entering upon 

 the discussion as to the relation between the flash spectrum 

 lines and the Fraunhofer spectrum, beyond saying that, 

 from a careful consideration of the facts so far brought to 

 light, I am inclined to believe that the flash 

 spectrum does in fact represent the upper 

 portion of the layer which by its absorption 

 gives us the ordinary dark- line spectrum, as 

 was held by Prof. Young, who first discovered 

 the so-called "reversing layer" at the eclipse 

 of 1870. 



I think that too much stress has been 

 laid upon the fact that, while the flash 

 spectrum lines correspond in position with 

 the Fraunhofer lines, yet in relative intensity 

 there are marked differences. This, it seems 

 to me, is only what we should expect to be the case when we 

 consider that in the one case we are looking tangentially 

 through the higher parts of the layer (which, it is to be 

 remembered, is some seven hundred miles in depth, with 

 a probable enormous increase of density at the base), while 

 in the other the line of sight passes entirely through the 

 layer from top to bottom. 



OCCULTATION OF 26 ARIETIS OBSERVED 

 PHOTOGRAPHICALLY. 



THE disappearance of a bright star when occulted 

 by the moon is always a striking phenomenon. 

 There is no celestial event whose time is susceptible 

 of more precise determination. For many years 

 various plans have been suggested, both here and 

 elsewhere, by which this time could be determined with 

 greater accuracy than by ordinary visual observation. In 

 fact, the apparatus for photographing the eclipses of 

 Jupiter's satellites, used here for several years, was devised 

 in part for this purpose. 



On February 2.5th, 1898, Mr. Edward S. King for the first 

 time succeeded in satisfactorily photographing the occul- 

 tation of a star. The apparatus used was an improved 

 form of that constructed for photographing the eclipses of 

 Jupiter's satellites, and described in the Astrnpln/^ical 

 .Journal, Vol. I., p. 146. The plate was moved automatically 

 every second by means of an electro-magnet. A motion 

 of about 303 cm. was given to the plate whenever the 

 circuit was closed, and of an equal amount when it was 

 opened. Connecting the apparatus with the standard 

 clock, Frodsham 132'7, two images alternately faint and 

 bright were obtained every second. As the faint images 

 are three magnitudes fainter than the bright images, the 

 ratio of the durations was about one to sixteen, so that 

 the absolute durations were O-OGs. and 0'.»ls. It is here 

 assumed that, as the times of exposure were very short, 

 the chemical action was proportional to the time. This 

 assumption is verified by actual measurement. 



Considering only the images taken during the minute 

 following 6h. 35m. Os., the bright images of 26 Arietis, as 

 shown, are equally intense, including that having an 

 exposure lasting from oQ-OGs. to SlOOs. Since this image 

 appears to be as bright as the others, the light of the star 

 could not have begun to diminish much before the time 

 51003. If the star had disappeared suddenlyat 50-!»3. the 

 last image would be at least 012 of a magnitude fainter 

 than the others, an amount readily measurable. The next 

 image is apparently invisible. Had the disappearance 

 taken place at 51-06s. the image would appear, and would 

 be as bright as the other faint images. A slight darkening 

 of the film is perceptible near the position the next image 

 would have had, with an intensity nearly equal to that of 

 the fainter images. If this were due to the star it would 

 denote that the latter suddenly disappeared at about 5ri2s. 

 The absence of the preceding image would indicate a more 

 gradual disappearance. In any case, the time is fixed at 



Occultatioa of 20 Arietis. (Kiilargod 10 times.) 



51-ls., to within one-tenth of a second. As the clock was 

 2m. 19-ls. fast, not including armature time, the corre- 

 sponding Greenwich mean time is 12h. 51m. 26"5s. By 

 using shorter exposures the uncertainty in the time of dis- 

 appearance can doubtless be greatly reduced, especially in 

 the case of the brighter stars. Since satisfactory images 

 of 26 Arietis, magnitude 01, were obtained in OGs., it is 



