152 



KNOWLEDGE • 



[Ukc. 23, 1881. 



hand, we arc to tbink of the ultimate atoms as little hard 

 balls, all exactly alike, and the chemical (elementary) 

 molecules as aggregations of a gi-eatcr or less number of 

 these atoms, variously arranged, it would be easy to 

 inm"ine some such likeness in the molecules of allied sul>- 

 stances as would account not only for the spectroscopic 

 facts, but many others. 



Mr. Lockyer, however, after experimentally disposing 

 of the theory that these " basic " lines arc due to impuri- 

 ties, has preferred tlie hypothesis that what is common to 

 two elements which show certain identical lines in their 

 spectra, is some constituent substance, found in both, and 

 liberated in greater or less proportion with change of tem- 

 perature ; and he finds in the existence and beha^dour of 

 tliese lines one of his strongest arguments in favour of the 

 compound character of the chemical elements. 



But I fear that the foundations of this argument are 

 insecure, however it may bo with the other evidence upon 

 which he rests his case. 



My observations in 1872 were made with a spectroscope 

 inferior to very few then in use. It had a dispersive power 

 equivalent to that of a train of twelve flint-glass prisms 

 of 60°, with collimator and view-telescope of 10 in. focal 

 length. In optical perfection I have never seen, and do 

 not expect to see, its superior ; of course, it easily showed 

 every line laid down upon the maps of Kirchhoff and 

 Angstrom, and a multitude beside. 



But soon after this time Mr. Rutherfurd's ditt'raction 

 gratings V)egan to come into use ; and as they were made of 

 continually -increasing fineness and excellence, something 

 like a new world opened before observers in respect to the 

 details and structure of the solar spectrum. Up to 1877, 

 no gratings were made with a ruled surface exceeding about 

 one s()uare inch in area ; but in that year Mr. Chapman 

 (Mr. Rutherfiu-d's mechanician), at my request, and with 

 Mr. llutherfurd's hearty concurrence, altered the machine 

 so that it could draw a line 1 \ in. long, and ruled for me 

 three plates on speculum metal, with difiraction surfaces of 

 nearly four square inches. One of these new plates allows 

 the use of so much larger object-glasses, that, by giving a 

 suitable focal length to collimator and telescope, the avail- 

 able dispersion is increased four-fold over that obtainable 

 from the smaller plates, with the same apparent brightness 

 of the spectrum. Perhaps I may mention in an aside that 

 one of our most distinguished American physicists has now 

 in hand a machine, nearly completed, and bidding fair to bo 

 perfectly successful, which rules lines four inches in length. 

 With this he expects to make gratings ha^^ng a ruled 

 surface 4 in. by G in., and with 20,000 linos to the inch, or 

 even finer. Quod favstnm sit. One would be almost 

 ready to die after a good study of the solar spectrum with 

 such a grating and accessories to match. But with one of 

 Mr. Chapman's best gi-atings, l^in. by 2| in., containing 

 over 10,000 lines (17, .100 to the inch), combined with colli- 

 mator a!id telescope of 42 in. focus and 3 in. aperture, and 

 a magnifying power of 2.">0 (which is perfectly l)orne under 

 the Viest conditions, when the grating is in good humour, 

 perfectly flat and of uniform temperature throughout) — • 

 with this instrument, I say, one reaches about the limit of 

 present abilities. Hundreds of lines shown .as single in the 

 older maps of the solar spectrum turn out to be double, 

 triple, or multiple, and the vacant spaces of the spectrum 

 are filled with crowds of fine lines and details of shading 

 before undreamed of. 



When attacked with this instrument, nearly every one 

 of the " basic " lines of the solar spectrum is resolved ; out 

 of seventy such on Angstrom's map, only .seven withstand 

 its power, and three of these seven are probably on the 

 "basic" list by mistake, as the numerical tables of llialen 



are at variance with the map regarding them. With an 

 instrument almost identical with my own, ProfesaorB 

 Living and Dewar have recently investigated the elec- 

 tric-arc spectra of nearly all the metals involved. Their 

 results accord with mine in almost every instance. They 

 find that the apparent coincidences between the lines of 

 difterent sul>stances always lireak down under powerful 

 dispersion ; so that as matters now stand, I see no evidence 

 to be drawn from such coincidences in favour either of 

 Mr. Lockyer's view, or of the slightly-different hypothesis 

 which I advocated myself. 



The spectra of different metals, so far as we can now 

 make out, have absolutely no points of exoct agreement 



As to the remarkable and disproportionate number of 

 these apparently common lines which are found in the 

 catalogues of chromosphere and sun-spot .spiectra, may not 

 the explanation be substantially as follows : — A line which 

 is really composed of two or more belonging to different 

 elements is much more likely to catch the eye than others. 

 In the first place, this line will appear when either of the 

 metals vigorously reverses its lines, even though the other 

 does not ; and again, in cases where both metals reverse 

 their lines, but too feebly to be detected against the back- 

 ground of the atmospheric spectrum, this line, of double 

 brightness, will be clearly seen. Since, prol>ably, all, or 

 pretty nearly all, of tlie lines are actually reversed close to 

 the sun's limV), though only a few arc usually blight enough 

 to be caught by ordinarj* instruments under ordinary 

 atmospheric conditions, it is ea-sy to see that a slight ad- 

 vantage of tliG kind indicated above would give a composite 

 line a great lift in the scale of relative frequency and 

 brightness. 



I have not yet been able to apply the highest obtainable 

 dispersion to the examination of the spots and promi- 

 nences, as the great spectroscope is too large to be attached 

 to the eye-end of our 12 ft. equatorial I have, however, 

 used the same grating with collimator and telescope of 

 12 in. focus, giving about one-fourth the resolving power of 

 the large in.strunient. Even this is more than four times as 

 powerful as the piismatic spectroscope used at Shemian, 

 and with it, it becomes perfectly clear that the catalogues 

 of prominence and sun-spot lines contain many serious 

 errors, the correction of which will be likely to remove 

 some puzzling anomalies. In a nunilwr of cases, lines 

 which are noted in the catalogue as bright in the spectrum 

 of a prominence, or thickened in that of a spot, turn out 

 to be entirely unaffected, the real culprit being a neigh- 

 bouring line, so tine and so close to the larger one, as to be 

 absolutely indistinguishable from it with the old instru- 

 ment. 



^\^len our great equatorial (of 23 in. aperture, and 30 ft 

 focus) is mounted, as it will be next spring, it will be able 

 to carry the large spectroscope without difficulty, and I 

 shall then hope to review the catalogues with power enough 

 to settle most of the questions of this sort. Until such a 

 review is made, it seems to me that generalisations founded 

 ou our piesent data must be very cautiously liandled. 



WHAT IS A GRAPE ? 



Bv Gr.\st Allen. 



^I"^IIEY make n beautiful picture, these big English hot- 

 _l liouse Vilack 1 lambros, with their purple bloom and 

 their waxlike texture, clustered thickly together in rich 

 luxuriance on their slender and heavily-weighted branching 

 fruit stalks. Indeed, we have Jiow cultivated them to such 

 a pitch of excellence, that tlieir old wild ancestors would 

 hardly recognise them to-day for members of the same 



