1910.] Vegetable Assimilation and Respiration. 



423 



on water at the end of the experiment ; it is highly probable that he thereby 

 approximately eliminated errors from shrinkage. 



In the following experiments, it will be observed, leaves which have 

 remained turgid have given results which entirely confirm Sachs' high value. 

 Others have shown smaller rates of increase in proportion to their departure 

 from this ideal condition. Observations with the horn hygroscope* have 

 indicated that the stomatal aperture is at any rate the principal factor in 

 determining the reduction of assimilation. 



The experiments were all made with cut leaves. The method of supplying 

 them with water calls for some remark, as much difficulty was met with in 

 keeping them turgid. 



Tht Treatment of Detached Leaves. — Leaves which are really robust, when detached 

 in the early morning, can stand insolation without wilting when the cut leaf-stalk is 

 simply immersed in air-free water. This procedure sufficed to keep fully turgid three of 

 the five leaves used in Expt. 1 in 1908. But in 1909 it was found necessary to 

 supply the less robust leaves that the poor summer produced with water under pressure 

 during insolation. This was done by fastening the leaves, after removing the greater 

 part of their petioles, with rubber tubing to the short arms of J -tubes. The joint was 

 made by wrapping the stalk with cotton wool which had been dipped in melted soft wax- 

 mixture, and tightly tying the rubber tube round this with string. In this way iujurious 

 local pressure of the string was avoided. In Expts. 4 and 5 several leaves were attached 

 to the same resei-voir of air-free water, 1^ metre above them. 



In the first experiment pieces were cut from the control and experimental 

 half-leaves by means of templets. The experimental half-leaves were 

 marked and their change of dimensions during the experiment measured, 

 and a corresponding correction was made for the change of area which had 

 taken place.f 



Having thus eliminated the error from change of area there still remains 

 an error from want of identity in respect of dry weight between similar 

 pieces from either side of the midrib of a leaf (the " asymmetry error "). 

 This is an error inherent in the method. Its effect on the result varies with 

 the weight of the leaf and the duration of the experiment, and must, there- 

 fore, be estimated for each experiment.^ The average and maximum errors 

 to be expected from this source are given above the table. 



Experiment 1. — August 11, 1908. Five leaves cut 6 a.m., and put in a dull 

 light. Experimental half-leaves set up in the open at 10 a.m. ; transferred 

 to the greenhouse at noon because of a storm of wind and rain. Morning, 

 sunshine intermittent ; afternoon, sunny ; maximum sun temperature,^ 



* See p. 429. 



+ Thoday, loc. cit., p. 37. 



X Jbid., p. 55 ; Summary, 5, etc. 



§ Given by a bright, bulb mercury thermometer. 



2 L 2 



