VISIBLE AND NEAR-VISIBLE RADIATION 199 



we may evaluate the dosage, i.e., the total amount of energy over a speci- 

 fied time, which has contributed to the biological effect. Measurement 

 then becomes theoretically proportional to effectiveness. Such observa- 

 tions are not, however, true energy measurements, but weighted energy 

 measurements. If the selectivity curve of the detector departs from the 

 sensitivity curve of the phenomenon, the result will be erroneous. It is 

 impossible to emphasize too greatly the need of caution in the use of such 

 methods. Where such a method becomes thoroughly standardized, 

 names are generally given to weighted energy measurements, the units 

 being proportional to effectiveness. The most completely worked out 

 system of this character is the illuminating system referred to above. 

 The visibility curve of the standard observer (full-line curve Fx, Fig. 33, 

 upper section) is the standard weighting coefficient for the phenomenon 

 vision. Wherever the sensitivity curve or weighting is known, absolute 

 energy measurements may be converted to the special units of the system. 

 Too often, however, the fact that a weighting curve must be employed 

 is overlooked and the observations for a particular phenomenon used for 

 purposes to which they are not applicable. This has been responsible 

 for much of the chaotic condition of radiation measurements. 



Where one wishes to know the absolute sensitivity of a photocell, 

 one is generally given values in lumens. Yet, with few exceptions, the 

 sensitivity curves do not in the least correspond to the visibility curve. 

 Consequently, the response of such an instrument will vary widely with 

 the type of source used. Unless the source is specified, observations 

 indicated in illumination units are absolutely meaningless. In our efforts 

 to present quantitative data on photocells, we find satisfactory data 

 available in only two or three cases. Relative-sensitivity curves are 

 generally given, but the absolute magnitude of the response can be 

 arrived at in only a very approximate fashion. Consequently, the curves 

 presented must not be used for quantitative measurements but only for 

 guidance in selection of cells. It is therefore with some hesitation that 

 the application of selective detectors to a number of biological problems 

 will be illustrated. Thus, 



Vision. — The sensitivity curve of vision may be approximated as 

 indicated in Fig. 33 by combination of photronic cell with filter. Such a 

 detector will produce deflections proportional to illumination, provided, 

 (a) the sensitivity curve of the eye has been accurately duplicated by the 

 sensitivity curve of the combination; (6) the response is linearly pro- 

 portional to the illumination. Neither of these requirements is generally 

 satisfied. Thus, the dotted curve in the upper section of Fig. 33 is an 

 exceptionally good approximation to the standard observer. Never- 

 theless, applied to sources of selective emission, it might yield results 



o 



greatly in error if lines occurred in the region from 4000 to 5200 A. 



