June 27, 1889] 



NATURE 



213 



hold at a pressure of i atmosphere. Differences in the sources of 

 lij^ht, in the spectroscope, and the observers, would, however, 

 cuunt for a good deal in observations of this kind. 



In order to try the influence of temperature on the absorption, 

 the shorter of the experimental tubes, 165 cm. long, was sur- 

 rounded by a jacket filled with a mixture of .'■olid carbonic 

 anhydride and ether, which was rapidly evaporated by means of 

 a large air-pump. By this means the temperature would be re- 

 ihiced to - 100^. The absorption of oxygen at several different 



I ssures up to 104 atmospheres was observed through the cooled 

 e. The authors were not, however, able to detect any in- 

 wcase of intensity, or other change, in the absorptions which 

 could be ascribed to the cooling. To try the effect of an increase 

 of temperature, the 18-metre tuhe was surrounded by a jacket 

 and heated up to ico" by steam. Heating appeared to render 

 ihe diffuse bands rather more diffuse and less distinct. On the 

 whole the influence of a change of temperature of 100° either 

 way is slight. 



The authors have observed repeatedly the absorption of liquid 

 oxygen in thicknesses of 8 and 12 mm. Their observations con- 

 firm those of Olszewski. 8 mm. of liquid oxygen gives plainly the 

 three diffuse bands above C, D, and F, respectively. With a 

 thickness of 12 mm. the authors were not able to detect any 

 more. 



The authors observed the absorption produced by liquid 

 oxygen on the one hand when it was cooled by its own evapora- 

 tion until the tension of its vapour was only equal to that of the 

 atmosphere — that is, to a temperature of - 181°, according to 

 r)lszewski — and also when the temperature of the liquid was 

 allowed to rise under pressure up to nearly the critical tempera- 

 ture. There appeared to be no appreciable difference in the 

 absorption under these different circumstances when the oxygen 

 was completely liquid, when it was near its critical temperature, 

 and when it was completely gaseous ; so far at least as concerns 

 the three principal bands, which were all that could be seen in 

 the light transmitted by the liquid in a thickness of 12 mm. 



It will be observed that taking the density of oxygen at 

 - i8i°'4 to be I "124, as given by Olszewski, 12 mm. of the 

 liquid would be equivalent to 9'37 metres of the gas at atmo- 



heric pressure —hardly more than half the thickness required 

 make A visible. The experiments, therefore, point to the 

 c inclusion that gaseous and liquid oxygen have the same 

 absorption-spectrum. This is a very noteworthy conclusion. 

 1 or, considering that no compound of oxygen, so far as is 

 known, gives the absorptions of oxygen, the persistency of the 

 absorptions of oxygen through the stages of condensation to the 

 state of complete liquidity implies a persistency of molecular 

 constitution which we should hardly have expected, 

 i In order to compare the absorption of ozone with that of 

 oxygen, the authors employed a tube 12 feet long, made of tin- 

 plate fitted with glass ends and coated with paraffin on the inside. 

 Ozonized oxygen was passed into the tin tube for some time, 

 while the ozonizer and the tube itself were cooled with ice and 

 salt. The lime-light, viewed through the tube, looked very 

 blue, and also the spot of light thrown from the tube on to a 

 sheet of white paper was equally blue, indicating a considerable 

 absorption of the less refrangible part of the spectrum. The 

 absorption, so far as the visible rays are concerned, appeared to 

 be of a general character, for the spectroscope revealed only four 

 extremely faint absorption-bands. The centres of these bands 

 were at about the wave-numbers 1662, 1752, 1880, and 1990, 

 and their positions with reference to the bands of oxygen are in- 

 dicated in the diagram. They were so faint as to be seen only with 

 difficulty. When the hot carbon of an arc lamp was substituted 

 for the lime-light they were rather more distinct, but the positions 

 of the edges were undefinable. The light of a gas-lamp was 

 insufficient to show them, and they were no better seen with a 

 single-prism spectroscope of low dispersive power than with the 

 spectroscope employed for observing the oxygen. Only one of 

 these bands is nearly coincident with an oxygen-band — namely, 

 that near E, the faintest of the oxygen-bands. That at wave- 

 number 1752 overlaps the strongest oxygen-band, but not at its 

 strongest part, and has none of the peculiar character of its 

 shading, abruptly increasing on the less refrangible side and 

 slowly decreasing on the other side. Photographs of the 

 spectrum (taken through a tube with quartz ends) showed that 

 the ozone absorbed all the rays above the wave-number 3086 — 

 a point between Q and R — while partial absorption extended 

 below Q. It may be said, therefore, that no identity can be 

 traced between the absorptions of ozone and those of ordinary 



oxygen. There is no mere displacement of tl>€ bands, such as 

 sometimes occurs when a coloured substance is dissolved in 

 different menstrua, nor any such resemblance as subsists between 

 the absorption- Lands of the different cobaltous salts derived from 

 different acids. 



The four bands which are seen to be produced by ozonized 

 oxygen correspond fairly with the second, third, fifth, and sixth 

 of the, bands described by Chappuis as due to ozone {A nnales 

 de V Ecole Nor male, 2nd sen, vol. xi. , May 1882). These four 

 bands, he says, are the first to be seen. The authors have failed 

 to perceive any others with the 3'66 m. tube, though the oxygen 

 was highly ozonized, and maintained at a low temperature. 

 None of the bands were of sufficient intensity to make them- 

 selves visible on photographic plates. 



It will be noted that the absorption by ozone extends far 

 below the limit of the solar spectrum. By diminishing the pro- 

 portion of ozone to oxygen in the tube the limit of the transmit- 

 ted light was continually advanced, as already described by 

 Hartley. The limit of the solar spectrum may, therefore, very 

 well be determined by the average amount of ozone in the atmo- 

 sphere, as Hartley supposes. The known variations in the limit 

 of the solar spectrum may be taken as confirmatory of this hypo- 

 thesis, although the comparatively small amount of those varia- 

 tions is certainly less than we should have expected if they 

 measure the changes in the proportion of ozone in the 

 atmosphere. 



The absorptions of the class to which A and B belong must be 

 those which are most easily assumed by the diatomic molecules 

 (O2) of ordinary oxygen. Whether oxygen in more complex 

 molecules, as in ozone (0:j), may be capable of taking up the 

 corresponding vibrations cannot easily be determined, because we 

 cannot isolate ozone ; but since none of the compounds of oxy- 

 gen with nitrogen, hydrogen, or carbon, or, so far as known, 

 with a,ny other element, exhibit these absorptions, it is very 

 probable that they are peculiar to the molecule Oj. From this 

 point of view it will be interestingtodetermine whether liquefied 

 oxygen, which we suppose to have more complex molecules, 

 produces these absorptions The corresponding spectrum of 

 emission has not as yet been observed, probably because the 

 agency employed to render the gas luminous breaks up the 

 molecules into single atoms of oxygen. 



As for the other class of absorption, the diffuse bands', since 

 they appear to have intensities proportional to the square of the 

 density of the gas, they must depend on a change produced by 

 compression. This may either be the formation of more com- 

 plex molecules, as for example O4, corresponding to the devia- 

 tion from Boyle's law exhibited by oxygen gas, or it may be 

 the constraint to which the molecules are subject during their 

 encounters with one another. Increase of temperature would 

 affect the former, tending to diminish the number of complex 

 molecules formed at a given pressure, but would have no effect 

 on the latter, for though the number of encounters of the mole- 

 cules in a given interval of time would be greater the higher the 

 temperature, yet so long as the volume was unaltered the ratio of 

 the duration of an encounter to that of free motion would be 

 sensibly unaltered. So far as any change due to temperature 

 has been observed, it is that a rise of temperature slightly 

 weakens the diffuse absorptions. 



Reverting to the compounds of oxygen, none of them show 

 the absorptions of oxygen, not even the general absorption of 

 the ultra-violet rays. Some of them, such as water and carbon 

 dioxide, appear quite transparent to ultra-violet rays, while in 

 others, such as nitrous oxide, which show a general absorption, 

 of the ultra-violet rays, the limit of transparency is different from, 

 that of oxygen. In other respects we may say that there is no 

 resemblance between the absorptions of the compounds of oxy- 

 gen and those due to oxygen itself. Some of the former have 

 very definite and characteristic absorptions, such as the well- 

 known spectra of the peroxides of nitrogen and chlorine, and we- 

 must regard these as indicating the rates of vibration which the 

 molecules NO2 and ClOj respectively are capableof easily taking 

 up. The absence of the absorptions due to oxygen from all com- 

 pounds of oxygen seems to indicate either that chemical combina- 

 tion is not, as has been supposed by some chemists, a temporary 

 relation in which the molecular groupings are continually break- 

 ing up, to be formed anew with ever-changing elementary 

 atoms ; or, that the periods of dissociation are very small com- 

 pared with the periods of association. For otherwise we should 

 expect that such compounds of oxygen as COj and NO2 must 

 always have amongst their molecules some molecules identical 



