ON THE EXISTENCE OF CERTAIN ELEMENTS 1 0f the spectrulll not lying between wave length 3600 TOG ETHER WITH THE DISCOVERY OF and wa�e leng�h !i000, that . o�r photographic work was 1 done chleflY'wlthm those hJllltS. It was, however, sup-PLATINUM, IN 'I'HE SUN.* plellleuted in many cases by eye observations in other portions of the spectrum. CONTRIBUTIONS FROM THE PHYSICAL LABORATORY We are convinced that there is much in the whole OF HARV ARD UNIVERSITY. matter of coincidences of metallic and

more »

... lar lines that By C. C. HUTCHINS and E. L. HOLDEN. needs re-examination; that sOlllething more than the mere coincidence of two or three Iines out of many is LA TE in the fall of 1886 it was decided by the writers, necessary to establish even the probability of the pres who were then at work in the physical laboratory of ence of a metal in the sun. With the best instruments Harvard University, to atteJupt a revision of so me of the violet portion of the solar spectrum is found to be the previous work in regard to the chemical constitu-so thickly set with fine lines that, if a metallic line were tion of the sun, as weil as to discover, if possible, new projected upon it at random, in many places the facts bearing ou the same subject, For the purpose of chances for a coincidence would be even, and coinci this investigation a magnificent diffraction grating, dences could not fail to occur in case of such met.als as made by Professor Rowland, of Baltimore, was kindly cerium and vanadium, which give hundreds of lines in placed at our dispOSld by Professor J ohn Trowbridge, the arc. under who�e super vision and directions the subseq uent Moreover, a high dispersion shows that very few lines work has been done. of llIetals are simple and short, but, on the contrary, After sorne delay, caused by the mounting of the winged and nebulous, and complicated by a great va grating and its attachments, work was begllll early in riety of reversal phenomena. A "line" is sometimes January, 1887, but, owing to bad weather and other half an inch wide on the photographic plate, or it may hinderances, was not regularly and systematically pro-be split into ten by reversals. secuted till somewhat later, At first we believed that these reversals were due to 'l'he grating used is of speculurn metal. with a ruled defects in the ruling of the grating, but we are con surface measuring 6 inches by 2, having 14,438 lines to vinced that they are true phenomena from the follow the inch, It is concave, its radius of curvature being ing experiments: 21 � feet" and is mounted according to Professor Row-1st. The wings continue when various portions of the land's method, Suffice it to say that the method is grating are covered. such that by Silllply rolling the camera along an iron 2d, �hey are the same in three successive orders of track it passes not only from one part of the spectrum spectra� to another, but also to the spectra of different orders, 3d. They are very different in different metals, and at the will of t,he operator. As the distances on this in some are not seen at all. track are prop0rtional to the relative wave lengths of 4th, We arranged a flat grating, with collimator and the lines tl?at fall successively on a /iiVfm point in the projecting lens, each of five feet focus, and found that carnera, it is easy, by means of a suitable scale of equal with this apparatus the same phenomena appeared, parts, placed beside the track, to set the cen ter of the On pages 87 and 88 of "The Sun," Professor Y oung photographic plate instantly within a single wave gives a list of elements in the sun according to the best length of any given line in the spectrurn. authorities, which is followed by a list of doubtful ele-And here let us parenthetically state that all our ments. Some,of these we have examined, wlth the fol wave lengths are those given by Professor Rowland's lowing results : photographic map of the solar spectrum. the position Cadmium.-The coincidence of the two lines given of every lilie refel'red to being carefully identified upon by Lockyer at wave lengths 4677 and 4799 is perfect. the map, and its absolute wave length thus determined, These are the only cadmium llnes near, and the sun Although sOllle of the uegatives contain many lines too lines in the vicinity are not numerous. faint to show on the map, yet we feel confident that Lead.-The evidence for lead, due to Lockyer, is our numbers correspond in all cases to those of the based upon three lines at 4019'7, 4058'2, and 4061'8. map within one-tenth of a wave lengt.h. We have photographed these lines with the sun Illany The light is brought into the room by means of a tim es, Thev are broad and nebulous. and often porte lumiere and then sent through the slit after total several times reversed. Lines in solar spectrum nu reflect.ion by a right-angled prism, Before striking the merous and faint. 4019'7 and 4058'2 certainly do not prism it passes throug:h a cylindrical lens, which con-coincide. 4061'8 is very difficult to pronounce upon; it denses it to a band of light about two inches long and may coincide. one-eighth inch wide, The jaws of the slit move equally Cerium, Molybdenttm, Uranium, and VanarZium. in opposite directions, so that, however widely they These foul' metals may be classed together. Lockyer Illay be opened, no lateral displacement of lines can re-finds four coincidences each for molybdenum and va sult frolll this cause, nadium, three for uraniuJll, and two for cerium, The Directly in front of the slit is placed a large tin lan-arc spectrum of each is characterized by great com tern containing an electric lamp. The image of the arc plexity and vast numbers of lines. So nUlllerous are the can be brought exactly upon the slit by means of an lines, in fact, that often on the photographs the total adjustable lens in the front of the lantern. In the lower space occupied by them is greater than the space not carbon of the lamp is made a cup-shaped cavity, which so occupied. A plate ten inches long may contain a is filled with the substance a spectrulll of which is de-thousand or so, Evidently coincidences between these sil'f�d, It is not at all necessary that this be in the form and solar lines cannot fail to occur as matters of of a metal, for any ordinary compound is at once re-chance, and therefore prove not hing. One can easily duced by the intense heat and the presence of carbon count a hundred or so such coincidences without the vapor to the metallic state. slightest conviction that the connection is other than The plan of working has been as folIows: The appa-fortuitous. Of course all this is nothing against the ratus being arranged as described, the sunlight is ad-probability of these metals being in the sun ; but at mitted and the desil'ed portion of solar spectrum pho-the same time those peculiarities of grouping, strength tographed upon the upper half of the plate; then the of lines, and other characteristics which occur in the sunlight is excluded by a shutter, and the image of the case of iron and other spectl'a, and which alone can electric arc containing the proper metal is allowed to serve as evidence in such cases, are conspicuously ab fall upon the slit, and its spectrum photographed on sent. the lower half of the plate. (Most of the plates used Among the metals whose existence in the solar at mo were those made by the M. A. Seed Co" and were cut sphere has seemed probable, we have exaUlined the fol to the size of eight inches by two. The most sensiti ve lowing: plates were obtained, and even then we found the Bismuth,-The line of the above metal at 4722'9, the required time of exposure fO!' some parts of the spec-only line of bismuth in the arc in that whole region, trum inconveniently long,) coincides perfectly with the more refrangible of a very In order to effect the exposure of either half of the faint pair of solar lines. plate at will, we placed directly in front of the cam-Tin.--The solitary tin line at 4525, thought by Lock era an opaque screen, in which was a rectangular yer to coincide, falls directly between two fine lines in opening one half the size of the plate. By turning a the solar spectrurn. handle, this screen is raised 01' lowered without the Silver,-Lockyer mentions a certain possibility of sil slightest disturbance of camera or plate, The metallic ver in the solar atnlOsphere frorn the apparent agree spectrulll, being thus photographed immediately be-ment of two of its nebulous lines with solar lines, One low the solar spectrum, can be cornpared with it at of these we have never been able to find in thc course leisure. of llIany photographs of thB region in which it is given These spectra are then examined with the aid of a by him. glass rnagnifying about ten diameters, and any coin-We find SE'ven lines of silver between 4000 and 4900. cidences between solar and metallic lines carefully Of these seven, three are what Thalen calls nebulous; noted according to their wave lengths. In order to so broad and hazy that their true positions cannot be eliminate any personal error, they are examined by determined with much accuracy, These lie at about both observers separately, and their results after·-4055'5, 4063'6, and 4212. A fourth line at 4023 is of the ward compared, same general character, but has a sharp reversal which To eliminate errors arising from suspected impurities ag re es with a solar line. The remaining three lines are of materials, as also frum the impurities known to ex ist represented in the sun, and are given by Thalen in the in the carbons employed, we took what we called "com-spark spectrum of the meta!. parison photographs," For these we pla'led in the car-4476'2. Very strong line ; nebulous on lower edge. bon cup a portion of the substancps known 01' suspected Snn line strong. (Thalen, 4475,) to be present as impurities in our metal, and then pho-4668'8. Strong, solitary line. (Thalen, 4666'5,) tographed the spectrulll thus given on the upper half, 4874'3. Fairly strong. (Thalen, 4874,) of the plate. A piece of the metal under experiment j 'l'hus, between the limits given above, every line of