302 
The comet is now as bright as it is expected to become 
according to computation, and moreover is rapidly moving 
southwards, so that it will soon be beyond the reach of 
observers in these latitudes. During the week it passes from a 
position near the 6th mag. star 17 Capricorni to the vicinity of 
the 4th mag. red star A Capricorni. 
STELLAR AND NEBULAR SPECTRA WITH CONCAVE GRAT- 
ING.—In the earlier part of 1898 Messrs. Poor and Mitchell 
described the results of their attempts to photograph stellar 
spectra witha Rowland concave grating (Astvo-Physecal Journal, 
8, p. 157). The grating used was a small one, having a ruled 
surface of only I x 2 inches with 15,000 lines to the inch, the 
radius of curvature being about 1 metre. Later a special 
grating was made with a ruled surface 2x 5? inches, having 
7219 lines to the inch. The radius of curvature of this was 
also I metre. The instrument was mounted on the 9°3-inch 
IIasting’s refractor as guiding telescope, and the results ob- 
tained were very promising, although the observatory is on the 
sixth floor of the Physical Laboratory at Baltimore. In 
November 1898, however, by the kindness of Prof. Hale, it be- 
came possible for Mr. Mitchell to mount the grating on the 12- 
inch Brashear refractor of the Yerkes Observatory (Astro 
Physical Journal, x. pp. 29-39, 1899). It will be remembered 
that the grating is used ‘‘ 27ect,” the concave surface bringing 
the diffracted beam from the star to focus on the plate, and that 
a considerable advantage obtains in that the spectra obtained 
are zormal, The grating was so oriented that the lines were 
parallel to the equator, so that irregularities in the driving-clock 
should have no effect on the definition. The astigmatism alone 
not being sufficient to give the spectrum sufficient width, this 
was effected by allowing the star to trail in right ascension. 
Photographs of the spectra of a large number of stars 
have been thus obtained, with exposures varying from 
5 to 60 minutes. These are given in a table in the 
article. Of special interest is the fact that these photo- 
graphs show the ultra-violet region remarkably well, as is 
to be understood when it is remembered that the light has to 
traverse neither lenses, prism trains nor slit. The photograph 
of Sirius showed about 75 lines between HB and Hy, and in the 
ultra-violet 21 lines of the series due to hydrogen were 
measured. 
In February two very interesting photographs of the spectrum 
of the Orion nebula were obtained with exposures of about 200 
minutes. Just as with an objective prism, these spectra consist 
of a series of images of the nebula, the measures of correspond- 
ing regions of which give the wave-lengths of the various lines 
they represent. 
With the grating used, the length of the photographic region 
in the first order was about 1} inches, using Seed’s gilt edge 
plates. In the second order the distance from Hf to Hy was 
0-6 inch, and from H& in the first order to HB in the second was 
2°8 inches. The photographic plate used, 1 x 5 inches, thus in- 
cluded both spectra, and their duplicate measurement afforded a 
definite control over the wave-lengths determined. 
Attention is directed to the fact that the spectra being zzorma/, 
absolute measurements of wave-length, and therefore of motion 
in line of sight, may be determined when larger instruments of 
this kind are available. A grating with ruled surface 10 x 15 
inches would probably be fully equal in performance to any 
spectroscopes in present use. 
THE REASON FOR THE HISSING OF THE 
ELECTRIC ARCH 
II. 
AND now we come to the most important of all the changes 
that take place when the arc begins to hiss, viz. the 
alteration in the shape of the fosz/2ve carbon. 
During the course of his 1889 experiments, Luggin (Ven 
Sitsungsberichte, 1889, vol. xcviii. p. 1192) observed that the 
arc hissed when the crater filled the whole of the end of the 
positive carbon. He was thus the first to call attention to the 
fact that there was a direct connection between hissing and the 
relation between the area of the crater and the cross-section of 
the tip of the positive carbon. My own observations in 1893 
1 Based on a paper read before the Institution of Electrical Engineers 
by Mrs. W. E. Ayrton. (Continued from page 286.) 
NO. 1552, VOL. 60] 
NATURE 
[Juty 27, 1899 
led to a conclusion somewhat similar to Luggin’s, but yet differ- 
ing in an important particular. It seemed to me that, with hiss- 
ing arcs, the crater always more than covered the end of the 
positive carbon—that it overflowed, as it were, along the side. 
How far this is true will be seen from an examination of Figs. 
4, 5, 6 and 7, which show the shaping of the carbons under 
various conditions with silent and hissing arcs. These figures 
have all been made from tracings of the images of actual normal 
arcs, burning between carbons of various sizes, and they were care- 
fully chosen with special reference to the shaping of the positive 
carbons. For, with normal arcs, the shape of the end of a 
positive carbon, even taken quite apart from that of the negative 
carbon and of the vaporous arc itself, is capable of revealing 
almost the whole of the conditions under which the are was 
burning when it was formed. It is possible, for instance, with 
a normal arc, to tell, from a mere drawing of the outline of the 
positive carbon and of its crater, whether the arc with which it 
was formed had been open or enclosed, short or long, silent or 
hissing, burning with a large or with a small current for the 
size of the carbon. 
Take, for example, Fig. 4 (see p. 285, July 20), and note the 
difference in the shape of the positive carbon with a current of 
3°5 amperes, as in (a), and with one of 34 amperes, as in (4). In 
the first case the tip of the positive carbon is rounded, so that the 
crater lies in its smallest cross-section ; in the second, the tip 
would be practically cylindrical for some distance, but that the 
HISSING. 
SILENT. 
+ 
Fic. 7.—Carbons :—(a) Positive, 18 mm. Cored. Negative, 15 
mm. Solid. (4) Positive,'g mm. Cored. Negative, 8 mm. 
Solid. LengthofArc,5 mm, Current, 25 amperes. 
crater has burnt away a part of the cylinder, making the tip look 
as if it had been sheared off obliquely. Comparing now the tips 
of the positive carbons when the arc is silent and when it is 
hissing in all the four figures, 4, 5, 6, 7, we find the same 
difference. With all the silent arcs the tip is more or less 
rounded, and the crater lies in its smallest cross-section, and 
consequently is less in area than any but the smallest cross- 
section. With all the hissing arcs, on the other hand, the tip 
of the positive carbon is practically cylindrical for a short dis- 
tance at least, or would be but that it is sheared away by the 
crater ; consequently the area of the crater is gveater than the 
smallest cross-section of the tip, or, indeed, than the cross- 
section of the tip for some little distance along its length, 
We have now arrived at the real, the crzcza/, distinction 
between a silent and a hissing arc. When the crater occupies 
the ed of the positive carbon only, the arc is sz/ent ; when it 
not only covers the end, but also extends up the sede, the arc 
hisses. Hence, the arc must be at the Aesszg port when the 
smallest increase in the area of the crater will make it begin 
to cover the szde of the positive carbon, and this can only be 
when the tip of that carbon has very nearly the same cross- 
section for some little distance from its end—in other words, 
when its sides are nearly vertical. 
I shall now proceed to show that the extension of the crater 
up the side of the positive carbon is not the effect but the cause 
of hissing; that, in fact, hissing zs produced by the crater be- 
