440 



TYLEB AND PREISENDORFEB 



[chap. 8 



collector must be provided. This usually consists of a diffusing glass or plastic 

 disc which has been prepared or mounted so as to collect flux in accordance 

 with the above cosine law. However, diffusing materials differ widely in their 

 properties, and do not in general behave the same way when submerged as they 

 do in air. Consequently design specifications are of limited value. The best 

 procedure is to conduct a practical underwater test of the collecting properties 

 of a plate as a function of the angle of incidence, using an axis tangent to the 

 front surface of the plate. The source of light should be a uniform field of 



PHOTOCELL 

 EDGE STOP 



IRRAOIANCE COLLECTOR 

 OPTICAL FILTER 



Fig. 21. Suggested design for an irradiance collector using a barrier-layer-type photo- 

 detector. 



collimated light which more than floods the surface being tested. In this way 

 the plate's properties can be brought to any desired degree of perfection. 



Fig. 21 illustrates a flat-plate collector which exhibited a total error of 2% 

 in irradiance when used to measure a typical underwater light field. 



7. Diffuse Attentuation Function and Reflectance Function 



The measurement of these two properties is intimately associated with the 

 measurement of irradiance [see equations (38) and (34)]. Photoelectric instru- 

 ments for measuring the downwelling flux and the upwelling flux have been in 

 use at least since 1922 (Shelford and Gail, 1922). However, measurements since 

 that time have not all been made using a diffusing collector, and the "extinction 

 coefficients" obtained can in no sense be substituted for the diffuse attenuation 

 coefficient as defined herein. Atkins et al. (1938) recommended the use of 

 diffusing glass in measurements of "submarine illumination." This important 

 recommendation makes much of the data published since that time of value 

 in approximating the value of K at various locations. Some of the diffusing 

 plates used since 1938 may not have collected exactly according to the cosine 

 law (this would mean that the value of K obtained was a function of the 

 directional distribution of the flux in the field as well as of the properties 

 of the diffusing plate), but the diffuse attenuation function, it will be 

 remembered, is obtained from the ratio of two values obtained at two depths. 

 The errors due to an imperfect plate thus tend to cancel. If a diffusing plate of 



