TEACHING AND LEARNING 15 



to record the reverse of the real facts. Readings of the volumes 

 of gases gauged by the surfaces of liquids are liable to several 

 errors. When the liquid communicates with the atmosphere, 

 error is introduced by changes of barometric pressure, or by 

 vapor tension, which alters greatly with temperature, or by 

 variations of temperature which affect the gas-volumes. Capil- 

 larity is another disturbing factor where liquids are concerned. 

 Some of these errors are so well known that they can be calcu- 

 lated from tables prepared for the purpose, of which those im- 

 portant to Plant Physiology will be found in Part III of this 

 book; but others can be met only by special precautions for 

 each case. In their relations to external conditions plants are 

 so complexly sensitive, that it is essential, in control experiment- 

 ing, to ensure that they shall be exposed to surroundings pre- 

 cisely the same even to minutiae. 



3. Errors of the Organism. The chief error from this source 

 arises from the innate variation between any two individuals, 

 no matter how closely related. For this reason, where two 

 plant parts are required, as in control experimenting, these 

 should be from the same stock, or as closely akin as possible. 

 But it is much better to use two shoots or roots of the same 

 plant, better yet to use two leaves, or other parts, of the same 

 age on one shoot, and best of all to use two corresponding parts 

 of the same leaf or other part, thus practically eliminating the 

 individual error. It is for this reason that so many of the experi- 

 ments recommended in this book are arranged, even at the 

 expense of considerable complexity, to use for experiment and 

 control two corresponding parts of the same structure. Where, 

 however, it is impossible to avoid using separate plants, it is 

 best to employ a considerable number, preferably at least ten, 

 when, by taking the mean of the results of all, the individual 

 error is minimized if not eliminated. Difference of age of parts 

 is especially liable to introduce serious error, but mechanical 

 injury, though serious, is, owing to the low degree of division of 

 labor among plant cells, less serious than is commonly thought. 



4. Errors of the Person. Of these I need hardly include 



