14 Mr. D. Thoday. Experimental Researches on [June 11, 



In these determinations the area of the whole of each half-leaf was 

 measured by the planimeter method. A number of prominent veins were 

 therefore included, and careful examination of a number of Catalpa leaves, 

 even if chosen for their apparent symmetry, reveals the fact that the 

 venation is not symmetrical. It seemed unlikely that the mesophyll itself 

 would exhibit differences as great as those found by Brown and Escombe for 

 mesophyll phis veins. For this reason modifications in the methods of area 

 determination have been adopted* with the object of avoiding all out- 

 standing veins. The rotating punch (p. 42) is very useful in this respect, 

 owing to the small size of the discs it will accurately cut, although it is not so 

 rapid in use as templets. In certain cases small templets have also been used. 



The advantage of avoiding veins altogether is clearly shown by some of 

 the results given in the tables that follow. For instance, the average 

 difference for 11 leaves of Helianthus when the punch method was used 

 (Table VIII) was + 1"4 per cent., whereas four leaves for which the templet 

 method was used, although the main veins were avoided, showed the 

 considerably higher average difference of + 2 per cent.f 



The experiment with Pmdownia imperialis (Table XIII) gave results still 

 more striking. The large leaves of this tree are strongly veined, but 

 between the main veins are areas practically without projecting veins. 

 From these areas pieces can be cut with templets as small as 4 cm. by 2*5 cm., 

 with considerable accuracy. Towards the margin, especially near the base, 

 are other areas traversed by the slightly projecting ultimate branches of 

 main veins, but these are nevertheless not prominent enough to increase 

 appreciably the errors of cutting. The contrast between the differences 

 shown by pieces including veins, and by other pieces from the " veinless " 

 areas, is instructive. For the latter the average difference was + 1*4 per cent., 

 for the former + 5-9 per cent., more than four times as much. 



The results for CaUd-pa hUjnonioides (Table XIV), obtained by the templet 

 method without avoiding veins, correspond fairly well with Brown and 

 Escombe's results for the same plant.J On the whole, omitting leaf 5, 

 wliere the extreme dilference was due to the slight convexity of one half, 

 the differences are ratlier less than theirs. 



The other tables give the results of experiments with other leaves. 



The general jjlan of dealing with individual leaves lias been followed in order to 

 determine how far the different results given by individual leaves in the same assimila- 

 tion experiment were to be accounted for by asymmetry, and how far other sources of 



* See Section VJ, pp. 39—44. 



+ Errors involved in area determination by templets may account for part of this 

 difference. Cf. Section VI, pp. 39, etc. 

 I Q\i()ted on p. 13. 



