METHODS OF INVESTIGATION 177 



severe maltreatment of the plants may be advisable in the 

 first experiments, in order to provide an unmistakable 

 effect as a base-line. The result is that such material 

 provides not one experiment, but a series of some sixty 

 experiments in one, with consequent economy of time 

 and labour in handling the material. 



Continuity of Record. Such treatment involves a prin- 

 ciple of considerable moment to biological research, and 

 especially in agricultural matters namely, the utility 

 of continuity in data. Every additional point in a curve 

 increases the definition of the curve, and a continuous pro- 

 jection is, moreover, much closer akin to the requirements 

 of the practical man than the exact definition of a few 

 isolated points. Much of the agricultural investigation 

 of the past has been unconsciously an effort to draw com- 

 plicated curves from the knowledge of only one or two 

 points along their course, so that divergencies of opinion 

 have naturally arisen and have given a proverbially 

 bad name to agricultural experts. 



Continuous data are so much more easily interpreted, 

 especially in field records. We are all familiar with the 

 recording barograph, and find it much easier to observe 

 the changes of the weather from the rise and fall of its 

 inked trace than in the days when the old game of tapping 

 the barometer prevailed in our households ; the barometer 

 might have fallen J inch and returned again during the 

 night, but we were none the wiser. The physical labora- 

 tories think in continuous projection, but in many fields 

 of biological work even the effort to obtain continuous 

 data has not yet begun, especially in tropical agriculture. 



12 



