36 Mr. D. Thoday. Experimental Researches 07i [June 11, 



The use of air-free water is undoubtedly an advantage. In the sun the 

 leaf is heated to a high temperature, and, if well aerated water is used, air is 

 set free in the vessels. Air-free water, on the other hand, may even dissolve 

 any air bubbles that already exist in the water channels. 



The water supply to attached leaves may be improved by watering the ground 

 thoroughly. Even this might be inadmissible in necological investigations, and there are 

 obvious objections from the same point of view to shading the leaves from the full 

 intensity of the sun. Even in investigating some questions of pure physiology shading 

 is to be avoided ; but for many purposes this plan could with advantage be adopted as 

 an accessory means. 



It is also possible, with attached leave.s, to choose the time of the experiment in such 

 a way as to ensure approximately the same degree of turgidity at the beginning and end 

 of the experiment. This would often be difficult and uncertain, however, as under 

 difierent weather conditions the duration of shrinkage and lute of recovery vary 

 enormously. 



Elimination of Shrinkage Errors. 

 We will now turn to questions of more immediate importance, and discuss 

 methods of completely eliminating or correcting for change of area. 



A. Immersion m Water. 



The most obvious way of eliminating shrinkage errors is to bring the 

 experimental half-leaf to its original condition of turgidity before determining 

 the area. 



♦ 



The experiment with leaves of Jlelianthus anmim testing the effect of soaking them in 

 water, and other similar experiments, suggested that care in avoiding flaccidity, combined 

 with immersion of both control and experimental half-leaves for half an hour before 

 cutting or measuring the area, would suffice to eliminate shrinkage errors sufficiently 

 accurately for many purposes. 



Other experiments have shown, however, that this is not always the case, and have 

 revealed other difliculties. Even when both half-leaves are soaked, the experimental 

 half-leaf after exposure to light may expand too nmcli relative to the control half-leaf. 

 With a detached leaf of Dipsacuit sp. a difference of as much as 1"6 per cent, was 

 observed. In an aSsimilation experiment with detached leaves of Ileliaidhw aniinus tiie 

 excess varied between 0*9 and 1-2 per cent. 



Such changes as these, of course, introduce a negative error into the uncorrected 

 increase of dry weight, wliich, though small, is not negligible if the nietliod is to be of 

 much service in solving problems of interest to the pure jjhysiologiHt. 



Another difficulty lias i)resonte(l itself in my experience with Dipsacus loaves. The 

 longitudinal dimensions near the midrib may be appreciably dili'erent when the midrib 

 is Htili attached and after it has been removed. Tlie change is fairly, small as a rule, 

 but for iiccurate work it is necessary to take care that the conditions undei' which the 

 two half-leaves are measured are similar iu respect of the tissue tensions existing, 

 as w(!ll as of tlie state of turgidity. When the ti'inplel method is used it is sufficient, if 

 the midrib Ik; stout, to cut pieces from tlie contidl half-leaf before removing it, for the 

 rosistance of the niidril) itself will |>revent deformation of the experimental half-leaf 

 owing to the relaxing of the tension on one side. Otherwise, the control lialf may be 



