20 



Mr. D. Thoday. Experimental Researches on [June 11^ 



apparent : it is, in fact, uncertain whether either intensity of light or leaf tem- 

 perature were limiting factors to assimilation in these experiments of Broocks. 



Again, the conditions under which dry-weight assimilation experiments 

 have usually been conducted would tend to exaggerate errors from shrinkage. 

 In order- to ensure that the leaves are starch -free to begin with, they are 

 covered with tinfoil overnight, or else the experiment is commenced soon after 

 sunrise. In either case the leaf is likely to be in a condition of maximum 

 turgidity at the beginning of the experiment. On the one hand, close con- 

 finement prevents the escape of moisture, often to such an extent that water 

 vapour condenses on the inner side of the tinfoil; while on the other hand 

 the relatively low temperature and high degree of humidity prevailing at 

 night and in the early morning are also just the conditions to retard evapo- 

 ration. Thus the area of the first half-leaf is measured, or pieces cut from it, 

 immediately after treatment calculated to produce full turgidity. The other 

 half is exposed to direct sunlight, or, even if the conditions are not so extreme 

 as this, to conditions favourable to considerable evaporation. Hence the 

 turgor might reasonably be expected to diminish and with it the area, and a 

 positive error to be thus introduced. 



Moreover, the frequent mention by previous workers of the difficulty of 

 avoiding the wilting of detached leaves suggests that the much greater 

 increase in dry weight shown in experiments with detached leaves might, in 

 part at least, be only apparent. Even the results of translocation experiments 

 would be too great, for the conditions under which these experiments are 

 carried out are favourable to increase in turgidity, which would result in an 

 apparent decrease in dry weight per unit of area and an over-estimate of the 

 amount translocated. In addition to this, the leaves may still be slowly 

 growing, and tlieir growth expansion would introduce a further error in the 

 same direction. 



Saclis was aware of the danger of comparing turgid with llaccid leaves, and 

 in his experiment with detached leaves of Ifeliantkvs annvus he floated the 

 experimental halves on water for lialf an hour al the end of the experiment 

 to make them turgescent. ^'et neither Saclis, nor any subsequent experi- 

 menter, until JJrown and Kscombe pul)lislied their results, gave any evidence 

 of having tested whetlicr shrinkiige affected liis own experiments appre- 

 cialtly ; nor, it may l)e added, of liax ing ascertained dial the leaves used had 

 ceased growing. They assumed tliat, so long as the con(Htion of a leaf 

 appeared to the ey(! to be a])])r()xiinat(;ly the same, such changes coidd not Ije 

 great eiiongli to allect their results to an appreciable extent. 



The following c-xiierinuint shows how luijustifiable this assumption was: — 

 A li'aT of IfiiiiiiifliiiH I H.hrrnsKs was allowed to dry nj) slowly, and at 



