1 8 PLANT PHYSIOLOGY. 



conditions that mechanical neatness and exactness are so essen- 

 tial in experimenting. 



In recording- results, in addition to careful tabulation, these 

 should, whenever possible, be plotted in curves or polygons. 

 Such curves not only have the great merit of expressing results 

 clearly to the eye at a glance, but they bring out facts and 

 relations altogether unsuspected from even the minutest inspec- 

 tion of tables of figures. Moreover their effect upon the 

 students is extremely good ; they encourage greater accuracy, 

 more critical study of topics, a better understanding of the 

 significance of results, and a greater general interest in the 

 work. The curves are, of course, especial!)' valuable where 

 more than two sets of conditions are to be compared, as, for 

 instance, where one wishes to compare the rate of transpiration 

 with the variations in temperature and in moisture, but they are 

 valuable in all cases where it is desired to express quantitative 

 relationships. Such curves are plotted on co-ordinate paper 

 ruled in faint colored cross-lines ; the abscissae or horizontal 

 lines are used for the degrees of the external condition which 

 alters steadily (as, for instance, ^temperature), and the ordinates 

 or vertical lines express the degrees of the effect upon the 

 plant; the joining of the tops of the ordinates gives the curve 

 or, more properly, polygon showing the relation between the 

 conditions expressed by ordinates and abscissae. A particu- 

 larly favorable opportunity for introducing the students to these 

 methods is offered by the study of the relations of protoplasmic 

 movement to temperature (page 54). It is well for students 

 not only to construct their own curves in this and other cases, 

 but also to unite and construct an average or class curve which 

 each student plots with his own for comparison. To allow of 

 this averaging, the different records must, of course, be made 

 upon the same basis. This class curve has also the advan- 

 tage already mentioned, that it offers a convenient, even though 

 not very exact, method of allowing the student to ascertain 

 approximately the probable error in his own individual results. 

 As an example of class and individual curves there is here 



