1909.] Vegetable Assimilation and Respiration. 37 



left attached ; but this i.s not always convenient, and is out of the question if it is desired 

 to measure the whole of each half-leaf, by the planinieter method. However, under 

 fomparable conditions of turgidity the tissue tensions would probably be similar on the 

 two sides, and if the midrib is removed from the experimental half-leaf before measuring- 

 it the error is likely to be reduced to a minimum. This was found to be the case in 

 experiments with Bipsacus ; the change in the longitudinal measurement of the control 

 half-leaf following its separation from the rest of the leaf was approximately the same as 

 the corresponding change in the exi)erimental half-leaf conseiiuent on the removal 

 of the midrib. 



To sum up, if care is taken to prevent the experimental half-leaf from 

 becoming very flaccid, soaking both half-leaves in water eliminates large 

 shrinkage errors ; but a more certain and accurate method is desirable. 



B. Correction for Slirinhugc. 

 An alternative plan, which avoids the uncertainty that attaches to 

 soaking, is to make a correction for shrinkage. 



(1) The method of estimating this correction, from changes in linear dimensions, has 

 been described already. Though only approximate it has been used successfully in 

 several experiments. The marks should be placed symmetrically on the two half-leaves, 

 and with reference to the portions to be used. 



Measurement between marks on the leaf with a millimetre scale can only 'be 

 applicable where the leaves used are large enough for the dimensions measured to 

 approach 10 cm. 



For small leaves similar methods coidd be adopted, based upon the use of a scale with 

 a vernier, or of a small portable microscope with a micrometer eye-piece. The use of 

 the microscope, which would involve the attachment of some form of index to the leaf, 

 might prove somewhat clumsy, and would take longer than measurement by eye. Any 

 form of measurement, indeed, must occupy a considerable time, and it would often be 

 quite impossible to measure the large number of leaves necessary to reduce the 

 asymmetry eiior within reasonable limits. On the othei' hand, for many purposes 

 a sufficiently accurate estimate of the aveiage change of dimensions could be obtained 

 from measurements of a few representative leaves. I hope to give some further 

 attention later to the question of the most convenient apparatus and technique for this 

 purpose. 



(2) The method adopted by Brown and Esconibe for measuring the shrinkage of 

 Catalpa leaves is also available as a general method for estimating the shrinkage 

 correction. They took prints of half-leaves still attached to the midrib and petiole 

 before and after exposing them to their experimental conditions, and measured these 

 prints with a planimeter. By following this plan a measurement of the degree of 

 alteration in aiea would be obtained directly, and, subject to the limits of accuracy of 

 measurement by planimeter, would be more satisfactory than the estimation of area 

 changes from changes in the linear dimensions of selected regions of the leaf. It would 

 be less suitable for attached than for detached leaves, since to obtain a print of a leaf 

 while still on the plant would be an awkward process. 



C. The Stamping 2[ethod. 

 The simplest and most satisfactory method of eliminating all possible 

 errors from change of area is to mark oat tlie area in some way at the very 



