44 Mr. D. Thoday. ExperimenLtal Researches on [June 11, 



avoided by careful selection of the leaves to be used for experiment, and are, therefore, 

 not inherent in the method. 



It may be remarked that by cutting small areas, as with the rotating punch described 

 above, the error from slight bulging would be considerably less than when a large 

 templet is used. 



The Stamping Metliod. 



The new method, which has been devised to eliminate shrinkage errors, 

 depends upon marking out a known area on each half-leaf before experiment. 



Preliminary tests have been carried out with an inked rubber stamp, 

 making an impression of a rectangle measuring 5 cm. by 2 cm. The areas so 

 marked out were cut from the leaf with scissors. To test the accuracy with 

 which this can be done, a number of impressions were made on a piece of 

 paper, cut out, and measured by means of a scale with a vernier reading to 

 O'l mm. 



In cutting, the outside of the line was followed, and hence a positive 

 difference from the actual area of the rectangle marked out was always found, 

 but the degree of uniformity which can be assured by a constant procedure 

 is tlie question of most interest and importance. 



The results were as follows :— 



Area marked out. Areas found after cutting out. 



10 -94 sq. cui. 11 -09 sq. cm. 



11 -07 „ 

 11 09 „ 

 Jl-U „ 

 11 -04 „ 



Averaf;e 11-08 „ 



The difference between the extreme values is here 0 0 per cent, of the 

 total area. Considering the sniallness of the urea, this coni])ares very 

 favourably with the templet errors as estimated above.* 



Similar tests have not been applied directly, using leaves, as in doing so 

 the same dilliculties present themselves as were met with in testing the 

 templet method. Experience with Clierry Laurel, Iiowever, has shown that 

 to follow a line with scissors is easier with a soft leaf than with tougher 

 paper ; and it is believed that, especially when larger areas are used, this 

 metliod will l)e quite as accurate as, if not more accurate than, the templet 

 method. It will l»e eni])loyed in some ])rojeeted experiments with the dry- 

 weight methofl by which the solution of some of the nidic iuiportant 

 problems of assimilation in tlie 0])en iiir will be attempted. 



* I'p. 11. 



