316 



SCIENCE 



[Vol. LVI, No. 1446 



and vegetable oils, in the region of the near 

 infra-red. In this paper he shows that because 

 of the great similarity in the spectral curves 

 of these oils, a study by means of infra-red 

 absorption does not lead to a detection of adul- 

 teration of one oil with another. 



But the region of the spectrum between the 

 visible and 30,000 A is filled with character- 

 istic absorption bands in the ease of organic 

 compounds, and consequently makes a very 

 interesting part of the spectrum to study. The 

 writer is at present studying the absorption of 

 many carbon compounds in this region by 

 means of a speetro-thermograph designed by 

 Professor E. P. Lewis of the University of 

 California, under whose supervision this work 

 is being done. There are certain advantageous 

 features in the construction and use of this 

 instrument. 



The source of radiation is a 108 watt lamp 

 made with a helix of tungsten wire for a fila- 

 ment. It has been found that the intensity can 

 be increased by as much as forty per cent, bj' 

 placing a concave mirror of about a ten centi- 

 meter diameter behind the lamp and focussing 

 the rays on the slit throug'h the lamp itself. 



For dispersion two 30" flint glass prisms 

 were used. Now flint glass gives a greater dis- 

 persion in this region than does quartz which 

 is usually used in near infra-red work. Fur- 

 thermore, it is almost as transparent as quartz, 

 the limit of transmission for each being around 

 30,000 A. The back of the second prism was 

 silvered and the radiations caused to pass twice 

 through each prism, thus doubling the disper- 

 sion. This gives a dispersion of between two 

 and three times that obtainable with a single 

 60° quartz prism and between four and five 

 times that obtained from a similar rock salt 

 prism. The silvered prism was capable of rota- 

 tion about the first prism so that the angle 

 between these was equal to twice the angle of 

 incidence necessary for minimum deviation of 

 radiation falling upon the thermopile, which 

 was used in series with a sensitive Leeds- 

 Northrup galvanometer to detect the radiation. 

 This rotation was accomplished by means of a 

 small thumb-screw which projected through the 

 black box which enclosed the system of mirrors 

 and prisms. Thus it was possible to focus the 

 eye upon the galvanometer scale and, turning 



the prism, make an approximate analysis of a 

 given substance in a few iminutes. 



This rapidity of locating either absorption 

 bands or emission lines proved to be very help- 

 ful in the process of calibration of the instru- 

 ment. Since no indices of refraction were ob- 

 tainable for the glass prisms, the instrument 

 was calibrated by observing certain metallic 

 emission lines and solar absorption bands deter- 

 mined largely by Paschen with a grating. 

 Coblentz states that emission lines may be 

 used in calibrating up to 10,000 A, but it 

 was found that the following lines could 

 also be detected when the proper salts were 

 introduced into the carbon arc: Tl, 13,010 

 A; Na, 18,460 A; Ca, 19,800 A. In lo- . 

 eating the positions of the solar absorption 

 bands, sunlight was refiected in from the win- 

 dow and focussed upon the slit. As the prism 

 was rotated the galvanometer defiection would 

 decrease to a minimum in the neighborhood of 

 an absorption band. By approaching from 

 both the long and the short wave-length sides a 

 set of quite consistent values for the corre- 

 sponding dispersion could be obtained. The 

 highest calibration point was for the solar ab- 

 sorption band at 25,000 A. 



It has been mentioned in the earlier litera- 

 ture of the subject that all compounds that have 

 a carbon atom united directly to a hydrogen 

 atom have characteristic absorption bands in 

 the neighborhood of 17,000 A. In this present 

 work this same band is found, and two others, 

 at about 11,500 A and 13,800a, appear in every 

 compound in which there is a C-H group. 

 Twelve compounds have already been carefully 

 analyzed, and an approximate determination of 

 the positions of the absorption maxima has 

 been made for some fifteen more. There are 

 very prominent displacements of these absorp- 

 tion maxima in certain compounds. In some 

 cases these shifts are toward the shorter wave- 

 lengths and in other cases toward the longer 

 wave-lengths. 



The mapping of the spectral curves of cer- 

 tain groups of compounds is being continued, 

 and an attempt will be made to ascertain 

 whether the shifting of absorption maxima fol- 

 lows any definite laws. 



Joseph W. Ellis 



University of California 



