30 Mr. D. Thoday. Exiierimental Researches on [June 11, 



same day is given for comparison. The chief point of interest is the contrast between 

 Leaf 7 and Leaves 8 and 9 at 10.15 a.m. The shadow of an avenue of limes left the 

 latter about 9 a.m., so that by 10.15 they had been in full sunlight for more than an hour. 

 Leaf 7, on the other hand, belonged to a plant several yards nearer to the trees, and 

 was not insolated till 11 a.m. Consequently by 10.15 the latter had decreased in area 

 very little, only 0'5 per cent., while Leaves 8 and 9 show a decrease by the same time of 

 1*8 per cent. 



Sachs' Experiments and Shrinkage Errors. 



The extreme importance of shrinkage as a source of error in the dry- 

 weight method is fully demonstrated by these results. To show the relative 

 magnitude of the errors which are possible, it will be convenient to take the 

 experiments described by Sachs and estimate the errors that shrinkage may 

 have introduced into his results. 



There is much difficulty in selecting, from my measurements of shrinkage, 

 results which can fairly be compared with Sachs' experiments, for there is 

 no means of knowing to what extent he took notice of differences in the 

 appearance of the leaves he used. One hesitates to put a high value to the 

 personal equation of an experimenter of Sachs' experience and ability ; 

 although, on the other hand, after an interval of 10 hours, slight differences 

 are easily overlooked, and even half-consciously ignored when the extent of 

 their effect on the experiment in hand is not suspected. Wliat follows is 

 thei'efore not to be regarded as finally condemning Sachs' original results, 

 however mucli doubt it may throw on tlieir trustworthiness. 



He commenced his experiment with attached leaves of Heliaiithus annuus* 

 at 5 A.M. on August 13, 1883, by cutting seven lialf-leaves from one large 

 plant. The other halves he took at 3 p.m., after 10 hours of clear sunny 

 weather with blue sky. The day was warm ; the liighest temperature he 

 records was 25° C. at 3 j'.m. 



It is quite certain that a shrinkage of 2 or 3 per cent, during this experi- 

 ment would have been overlooked, while if Sachs used robust leaves, similar 

 to Leaf II described above, he might easily have allowed a change of .area of 

 4 or 5 per cent, to pass unnoticed. 



Perhaps, however, it will l)e better to take a general average from actual 

 observations. The following table contains my measurements made in the 

 summer of 1908 under weather conditions which were in several cases less 

 severe than those holding during Saclis' exjjeriment : — 



* Loc. CI I., )). 23. 



