378 * BOTANICAL GAZETTE [may 



The permeability of the same individual membrane was measured 

 under the different conditions studied, therefore, and the same set 

 of observations made with a number of other membranes of the 

 same species. In some cases it was possible to allow the same 

 membrane to remain in the osmometer during a whole series of 

 readings. When it was necessary to remove the membrane from 

 the osmometer, care was taken in replacing it that the same portion 

 of the membrane was used in the next reading. 



Constancy of semi permeability. — When a membrane gave con- 

 stant rates for i hour, readings being taken at intervals of 10 

 minutes, it was assumed that its permeability to salt or sugar had 

 not changed during the experiment, or at any rate that any change 

 in permeability that had occurred did not affect the readings taken. 

 No serious attempt was made to determine the completeness of 

 semipermeability. Preliminary experiments indicated that the 

 membranes were slightly permeable to sodium chloride, but a 

 passage of cane sugar through the membrane was not detected. 

 Conductivity measurements are to be made to determine the 

 permeability of these membranes to salts, and these results will 

 be reported in a later paper. 



Preparation of solutions.— C&ne sugar solutions were prepared 

 in accordance with tables given by Find lay (ii), rock candy 

 being used. Sodium chloride was made up on the volume molec- 

 ular basis and its osmotic pressure figured from the data given 

 by Renxer (25). 



Capillary tube errors. — Although according to the law of Poise- 

 uille the flow of water through capillary tubes is affected by 

 temperature, it is not believed that this was a factor in these 

 experiments. In Poiseuille's experiments the liquid was sub- 

 jected to a head of pressure and flowed through the capillary tube 

 with rapidity, whereas in these experiments there was no hydro- 

 static pressure applied and the rate of movement through the tube 

 was very slow. According to Barker (3), when water is the liquid 

 in question this law applies only to tubes with less than 0.5 mm. 

 diameter of capillary; while the capillary used in these experiments 

 was 1 mm. It is not believed that the results obtained are affected 

 by the influence of temperature on the flow of water through the 

 capillary tube of the apparatus. 



