188 HOWARD E. PULLING 



lamps can be stated with a degree of accuracy sufficient for very 

 good work of at least a preliminary nature without resorting 

 to measurements of the radiation emitted by them after they 

 are installed. It is of course necessary that the current upon 

 which they are operated be sufficiently constant for the needs 

 of the particular case in hand. New lamps fluctuate in emissive 

 power for a time and then assume a very nearly constant state 

 which persists during almost the entire remaining life of the 

 lamp (with the vacuum tungsten lamps, until blackening of 

 the glass is serious, with the globular gas-filled lamps properly 

 placed, practically until the filament breaks) especially if they 

 are operated below their maximum brilliancy. 



Turning from this digression to the main topic — -the methods 

 for the measurement of the sun's radiation — it is necessary to 

 note that before selecting an instrument it is essential that 

 the investigator decide just what feature of the incident radiation 

 he desires to measure. The time has gone by when the choice 

 of instruments is simply made ; for unless we know what features 

 of the sun's radiation we wish to measure and the applicability 

 of the method selected, the acquirement of useful results is 

 purely accidental, regardless of the care exercised in using the 

 method. Often very rough measurements from the physicist's 

 standpoint yield valuable results in biology and very sensitive 

 instruments may give data that are wholly erroneous. Scien- 

 tific data are, of course, no more accurate than the net accuracy 

 of the entire method, and to measure radiation with a precision 

 of 1% while the corresponding observations on the plants vary 

 to 30% or more not only entails sheer waste of time but is mis- 

 leading and hurtful to scientific progress. On the other hand, 

 the idea that any instrument is as good as any other for rough 

 preliminary research (as all botanical radiation work will be 

 in the near future) is also wrong because the instrument may 

 not measure at all the component of radiation that is effective 

 in the biological process under consideration. The photo- 

 graphic actinometer, when used to obtain a measure of the 

 amount of energy available to plants for photosynthesis, is a 

 case in point. 



