June 15, 1917] 



SCIENCE 



617 



now be obtained in stamped plates 6J X ^i 

 inches. 



A brief characterization of a few of these 

 is given below : 



Bed: High transmission in red removes all 

 light below .61 fx. 



Blue : Transmits only blue below .52 ju, and 

 may be made deeper to transmit only below 

 .50 fj,. 



Yellow : High transmission in red and infra- 

 red and through green to .48 /x giving about 

 75 per cent, of incident white light. All ultra- 

 violet absorbed. 



Uviol: Transparent to visible spectrum, 

 transmitting ultra-violet to .31 ^ in sheets J 

 inch thick, to 30 fj. through J inch thick. 



Heat-Ahsorhing : Absorbs most of infra- 

 red and 97 per cent, of heat of Nernst lamp — 

 gives a pyrheliometer reading about half that 

 of good window glass. Transmits 65 per cent, 

 of incident white light. 



Formulae of twenty other glasses are avail- 

 able by the use of which the regions of the 

 spectrum noted above may be modified or dif- 

 ferent separations made. Desirable effects 

 may also be obtained by combination of two 

 screens. Thus for instance light passing 

 through a yellow of the type described above 

 and the heat absorbing glass loses all the 

 spectrum except the yellow and a small part 

 of the red. 



The only thoroughly reliable measurements 

 of solar radiation available to the biologist 

 are those made with the Angstrom and Abbot 

 type of pyrheliometer which recorded the total 

 normal insolation in heat units. However, in 

 the blue-violet region of the spectrum, which 

 is of especial interest to the biologist, this 

 type of instrument is not sufficiently sensitive. 

 It is therefore proposed to use the photo-elec- 

 tric cell as developed by Elster and Geitel. 

 This instrument has the great advantages of 

 extreme sensitiveness in the blue-violet region 

 and ease of manipulation; it records immedi- 

 ate and directly proportional values, and can 

 be used for extensive ranges of intensities. 



A comparison of the results to be obtained 

 by the use of the two methods is afforded by 

 the data given below. Direct sunlight at the 



Desert Laboratory is taken as 100, and the 

 figures in both cases are percentages of this 

 total. The values from the pyrheliometer were 

 calculated in calories per sq. cm. per minute, 

 and those of the sodium photo-electric cell are 

 from readings of the high sensitvity galva- 

 nometer. 



Proportion of Direct Sunlight (Sun- 

 llgllt = 1.39 Calories per Sq. 

 niumlnation Cm. per Minute) 



Smltlisonian Pyrhe- Sodium Cell 

 liometer Values Values 



Direct sunlight at 100% 100% 



Transmission — 



Uviol glass 90.2% 86.6% 



Yellow glass 53.6% 5.1% 



Eed glass 42.4% 0.62% 



Heat-absorbing glass. 25.4% 63.2% 



Blue glass 10.5% 49.0% 



These results show a total value of normal 

 sunlight through uviol glass transmitting the 

 entire spectrum not being widely different by 

 the use of the two instruments, although the 

 pyrheliometer values are derived from the 

 longer wave-lengths and those of the sodium 

 cell from the shorter ones. 



It may be assumed that half of the total 

 energy registered by the pyrheliometer is 

 strictly within the red, which causes but little 

 action in the sodium cell. 



The pyrheliometer shows a total of nearly 

 54 per cent, of the energy of sunlight passing 

 through the yellow screen which transmits 

 from red to and including the blue-violet. 



Perhaps the most interesting results are 

 those which are obtained by measurement of 

 light passed through the so-called heat-absorb- 

 ing screen, which has been found to transmit 

 the visible spectriun except the longer red and 

 infra-red. 



The pyrheliometer reading of such a glass is 

 but 25 per cent, of clear sunlight, while the 

 sodium cell records 63.2 per cent, of the total. 

 A notable difference between the recording 

 action of the two instruments in the blue is 

 also evident. It is self-evident that the uni- 

 versal method of calibration of sunlight in- 

 tensities by the pyrheliometer does not give 

 results which are adequate or correct in all of 

 the various aspects of the physiological effects 

 of light. 



