48 Mr. D. Thoday. E:tperimental Researches on [June 11, 



Section VIII. — General Considerations on the Degree of Accuracy and 

 THE Utility of the Method. 

 An attempt will be made in this section to estimate the degree of accuracy 

 which can be obtained with the dry-weight method, and its probable utility in 

 the future. 



Since our knowledge of the composition of the true photosynthetic increase 

 of dry weight is at present scanty, attention will be confined to the question 

 liow accurately the gain of dri/ weight itself can he measured bi/ the half-leaf 

 'method. For this purpose it is only necessary to consider the errors due to 

 asymmetry and change of area, as errors of technique are reducible within very 

 narrow limits and are always included in measurements of asymmetry. 



It will be interesting to begin by supposing that in an experiment lasting 

 5 hours it is desired to measure the increase of dry weight per hour per square 

 decimetre coiTect to the nearest milligramme, a degree of accuracy which 

 would be sufficient to allow a great deal of work to be done by the method ;* 

 and to consider the means necessary to secure this degree of accuracy. 



For results correct to the nearest milligramme the total error in the 

 increase per hour must not exceed +0"5^ milligramme. If the experiment 

 last 5 hours, the error in the total increase of dry weight during that time 

 must be within + 2"5 milligrammes per square decimetre. 



The percentage error to which this limit corresponds will vary with the 

 average dry weight per square decimetre of the leaf. This is widely different 

 for different plants, as the following figures show : 



Average Dry Weight of 1 sq. decimetre of various Leaves. 



AUiaria officinalis 0 "1 8 fjniiniiio.t 



llelianlhus ainiHus 0 '33 ,, f 



Rumex sp 0 '41 ,, f 



Calalpa hignonioides 0 '45 ,, 



I'au/oiviiia imperiO'fis 0 '6 „ 



Clierry Laurel 1 "0 „ 



It may be instructive to calculate tlie desired limits of the i)robabIe 

 percentage erroi's for Jfrli(ui//ivs and the Cherry Laurel, as examples of two 



* Blown and Esuonibe {loi: c/t., \). (JI) infer from tlicir own re.-ulta witli Catalpa 

 hignonioides that a total error from asymmetry and shrinkage of +2 percent, of the total 

 dry weight is easily possible, which, in a Tj-hour experiment, would mean an error in 

 tlie increase per s(iuare decimetre per hour of ±2 milligrammes; this is equal to the 

 average rate of assimilation found liy them for leaves of this plant by measuiing the 

 actual intake of carlfon dioxide from ordinary air. 



Ah a representative illustration this is unsatisfactory. Catalpa leaves are loss symme- 

 trical tlian the average, and thei'e can be no doubt that a rate of assimilation of only 

 2 milligi'ammes per hour is greatly exceeded by many leaves in tlie oj)on air (c/. 

 Tiioday, J). 'Brit. Assoc. Ileport, Dublin,' ]!)08). 



+ MUllor, he. cil., p. 474. 



