visible, besides the characteristic hydrogen-lines; at two places in the green a bright glow was to be seen; and also between Hẞ and Hy the two previously mentioned bright bands were visible. On further rarefaction to 3 or 2 millims. the characteristic lines retained the same brightness, and everything else disappeared almost entirely from the spectrum; yet, with a simultaneous enfeeblement of the bright lines, part of the continuous spectrum reappeared in the green, in the form of about five bright fields, when the gas was rarefied to fractions of a millimetre pressure. These observations show that the hydrogen-spectrum described belongs, in fact, to a lower temperature than that consisting of the three lines; for with increasing density of the gas in the spectrum-tube the temperature must become lower, since the induction-current experiences a greater resistance in the denser gas, and a larger quantity of the gas has to be heated. But, just as a great density of the gas does not permit the full intensity of the current to be developed, so the current is also enfeebled by great rarefaction; for by adequate rarefaction the current in a spectrum-tube may be completely stopped. The occurrence, therefore, of the continuous spectrum both with greater density and with greater tenuity of the gas proves that it belongs to a lower temperature. 4. By the experiments communicated in the preceding, the proof has been furnished that the continuous hydrogen-spectrum belongs to a lower temperature; and the question arises how it happens that Geissler's spectrum-tubes, which contain hydrogen under a pressure of from 5 to 10 millims. (a pressure favourable, therefore, for producing the line-spectrum), yet after some time, after lengthened use, may yield the continuous spectrum. The observation that this spectrum is especially seen when, after lengthened use, the tube is exposed to the action of a feebler current, led to the supposition that (possibly owing to a superficial melting of the electrodes) greater resistance was offered to the passage of the induction-current. That such a fusion of the electrodes has an influence of that kind was established by a series of experiments. In the extreme degrees of exhaustion mentioned above, and which are to be subsequently discussed, the resistance in the spectrum-tubes was so great that the whole of the positive electrode became incandescent. It thereby became quite bent, and after some time was partially fused, so that it looked like a series of small pearls on a thread. After this deformation of the electrodes had occurred, the tube, on gradual exhaustion of the hydrogen, always exhibited the continuous spectrum, even under pressures at which the green light was otherwise scarcely visible. The continuous spectrum was extremely bril liant under a pressure of 30 millims., 21 millims., 16 millims. ; and even under a pressure of 8 millims. it was doubtful whether the continuous part was really fainter, or whether it only appeared so in comparison with the dazzling lustre of the line HB. Several observations which we made on a phosphorus-tube and a sulphur-tube, prepared by Dr.Geissler, favour, I think, the view that the nature of the electrodes exerts an influence on the production of the continuous spectrum. These tubes also contained hydrogen. If the induction-current was allowed to pass without heating the tubes to the melting-point of phosphorus or of sulphur, they exhibited a beautiful continuous spectrum exactly as has been previously described; the spectrum of phosphorus or sulphur only appeared on stronger heating. In these tubes the electrodes are always more or less covered with phosphorus or sulphur, by which, since these substances do not conduct, the resistance to be overcome is greater, and therefore the intensity of the current must be less. 5. It has already been mentioned, in § 3, that when the pressure of the gas in the tube only amounted to fractions of a millimetre the continuous spectrum of hydrogen was again formed, particularly in the green. If the tube was then still further exhausted by means of the Sprengel's pump, the light in the tube first became feebler and its colour paler; and in the spectrum, while all the rest was weakened, the green part stood out still more beautifully. It appears in the form of six beautifully shaded bright bands, which are connected with each other by less bright intermediate spaces. In the brightest parts of the blue in the continuous spectrum bright fields also appear, of which that in the middle between Hẞ and Hy is seen as columnar grouped lines. On further pumping, the light in the tube suddenly becomes of a splendid green like the light of a thallium-flame, and the spectrum is quite changed; the red line Ha can scarcely be seen; the reddish-yellow part of the spectrum has completely disappeared; and in the green six splendid groups of lines appear on an almost black ground. Repeated measurements gave for the least deflection of these groups the following values*: (1) Middle bright line of the first group, consist ing of three bright lines, this being the brightest 62 47 40 (2) Middle line of the second group, consisting63 10 15 of three The flint-glass prism used for these measurements has a refractive angle of 60° 2' 00"; its refractive indices are, for Ha 1743355, for HB-1-772210, for Hy=1·790564: the measurements were made with a Meyerstein's spectrometer with a circle divided to 10". (3) Second brightest line of the third group, con 63 29 20 sisting of two bright lines (4) First bright line of the fourth group, consisting of two very closely adjacent lines (5) Middle line of the fifth group, consisting of three bright lines: this line is the brightest, and is more than a slit in breadth 63 46 25 64 22 20 (6) Middle bright line of a group of at least six64 38 40 individual lines (7) Hẞ still faintly visible 64 51 10 These groups, as is evident, corresponding to the green colour of light, all lie in the green part of the spectrum; in the red and yellow part there is nothing to be seen. Besides these measured groups, at the boundary of the green towards the yellow a feebly bright part was seen, between the first and second groups two feebly bright lines, and between the fourth and fifth groups about three faint lines. About as far to the right of Hẞ as the sixth group is to the left of B, there is a feebly bright line, too obscure, however, to be measured. Then at about 65° 20' there is a faint blue field bounded on both sides by two beautifully shaded bright bands; and behind this, after a perfectly dark space about half as broad as the field just mentioned, there is a faint field of considerable breadth; at times there is in the neighbourhood of H y, at 67° 10′, a faint lustre. This spectrum occurs whenever the gas in the tube has attained the extreme degree of rarefaction attainable with a Sprengel's pump. The resistance in the tube is here so great that the positive electrode becomes quite incandescent, bends, and appears to consist of a series of fused globules. That this deformation at the same time seriously hinders the passage of the current, as before mentioned, follows from the fact that on its entrance the current no longer started from the point of the electrode of aluminium wire, but from the part of it which lay against the platinum wire melted into the glass of the tube, where such a fusion could not be perceived. If the extreme rarefaction which furnishes the spectrum just described is maintained for some time with closed stopcocks, the light of the tube again assumes a white colour and again shows the continuous spectrum, the reddish-yellow part is again seen, the six groups of lines again disappear, and the green appears once more. But the density of the gas in the tube is not changed; for if the stopcock be opened which connects the tube with the air-pump, the position of the mercury remains quite unchanged. Notwithstanding this, renewed pumping again produces the linespectrum. Another means of again evoking the line-spectrum is the si multaneous interposition of a Leyden jar in the circuit of the induction-current. A brilliant bright-green light is then obtained in the tube, and the six groups of lines become truly splendid. At the same time the parts lying between the groups of lines become brighter, yet not to such an extent as to change the character of the spectrum. The tube then afterwards exhibits the same spectrum even without a Leyden jar. This spectrum is also obtained with the extremest degree of rarefaction if the spark of Holtz's machine with a condenser be caused to pass, or if a small Leyden jar with a short striking-distance be discharged through it. Care must at the same time be taken not to have the striking-distance too great; otherwise the calcium-spectrum, or the continuous spectrum of ignited glass with the dark line D, occurs. With a tube once arranged I have frequently observed this spectrum for fourteen days together and compared it with others; so accurately closed were Dr. Geissler's glass stopcocks. 6. The phenomenon of a hydrogen-tube at its extreme exhaustion furnishing a third spectrum essentially different from the earlier ones observed is so surprising, that the question must be considered whether this spectrum belongs in fact to pure hydrogen, or is not due to other elements standing in connexion with the tube. It might be believed that it was a spectrum of aluminium, of which the electrodes consisted, or of mercury, of which, perhaps, vapours had distilled over into the tubes, or of phosphorus or sulphur, as the gas was dried with phosphoric and sulphuric acids, or, finally, that some of the fat with which the stopcocks were slightly coated had evaporated and had entered the tubes, and thus that the lines belonged to the spectrum of carbon. As regards carbon, the description which Plücker gives of the spectra of this element* shows that none of the groups of lines of the kind described occur in them; and in an investigation of the spectrum in a tube filled with carbonic acid, we found it of the same character as Plücker describes. In reference to the sulphuric-acid spectrum, Plücker† states that it is one of the most beautiful spectra and at the same time most rich in colour, it consists of bright luminous bands on a black ground; and he then counts three red, one orange, one yellow, four green, nine blue and violet bands. He mentions at the same time that, using anhydrous sulphuric acid, this spectrum can only be obtained with the large induction coil. All this proves that the spectrum described cannot be due to sulphuric * Philosophical Transactions for 1865. acid, even if we were to assume that along with the hydrogen some sulphuric-acid vapour had entered the tube. Our spectrum cannot be confounded with that of phosphorus or mercury, as follows from Plücker's descriptions, and as we have convinced ourselves by experiments. Phosphorus exhibits in the green only one group of bright lines, about 8' broad, the most refrangible of which has the minimal deviation of 63° 38'. The mercury-spectrum consists of a series of bands, of which a yellow one is particularly characteristic; moreover, in a tube containing mercury, this spectrum is only formed when the tube is considerably heated. To compare the spectrum in question with that of aluminium, wires of this metal were fastened to the ends of the wires leading to the induction-apparatus, and the induction-spark was allowed to strike between these, which were placed in front of the slit of the spectrometer. It was then observed that, according to the distance of the electrodes, the aluminium-vapour might have two essentially different spectra. When the distance of the electrodes was only about two millims., the spectrum consisted of four green splendidly shaded fields. The fields are brightest on the more refrangible side, and gradually diminish towards the less refrangible side; at distances of 5 minutes these fields are traversed by sharp bright lines, and these give quite the appearance of fluting to the fields. The right limit of these fields was found, in the minimum deflection, at : The measured distances of the fields (that is, of their right boundaries) are so nearly equal that their differences may be regarded as errors of observation, since the adjustment is not perfectly accurate; the breadth of the fields is about 30'; so that this spectrum of aluminium consists of four equidistant almost equally broad groups of flutings. When the distance of the two wires between which the sparks passed was increased to 10 millims. and more, a totally different spectrum was obtained, both with the use of the small Ruhmkorff and with that of Holtz's machine with superposed condenser; the four fluted fields disappeared, and instead of them a number of bright lines and groups of lines started out upon a feebly illuminated background. With the minimum deflection the positions of these lines were :— |