THE PRODUCTION OF SCIENTIFIC KNOWLEDGE 179 



conditions and supplied widi analytically controlled food 

 must be established, and these require much attention and 

 care if the experiments are not to be interrupted by acci- 

 dental losses from disease. It is necessary, in fact, for us to 

 pay more attention to the health of our experimental animals 

 than we do to our own health. Similar colonies are required 

 for the study of heredity. 



The mere accumulation of facts is being expedited very 

 much by improved apparatus. In the study of photography, 

 for instance, much of the fundamental information is ob- 

 tained in the form of a curve known as the characteristic 

 curve, which relates the density of a developed image to the 

 exposure given to the light-sensitive material. To obtain 

 these curves, the material is exposed to a series of light in- 

 tensities and developed, and then the densities resulting are 

 measured and the curve plotted. With a visual instrument, 

 the measurement of density is a very slow operation, and 

 much effort is required to produce twenty curves in a day. 

 Indeed, such a rate of production cannot be maintained; the 

 making of some four hundred photometric matches in a day 

 is very tiring. Today automatic instruments using photo- 

 electric cells measure the densities and draw the curves, and 

 it is well within the capacity of such an instrument to produce 

 over a thousand curves in a day when used by an unskilled 

 operator. More and more, scientific men are designing im- 

 proved methods of collecting and analyzing the data on which 

 they can base their studies. Thus they are again accelerating 

 our production of knowledge. 



A useful classification of research laboratories in general 

 is based on consideration of whether all the problems investi- 

 gated are connected with one common subject or are of many 

 kinds having no connecting bond of interest. The first type 

 of laboratory might be called unipurpose or convergent and 

 the second, multipurpose or divergent. 



In the convergent laboratories, although the actual investi- 

 gations may cover as great a range of science as those under- 

 taken in a divergent laboratory, all the investigations are 



