148 PLANT PHYSIOLOGY 



which does not occur in the case of the roots. Thus our experi- 

 mental membrane is permeable to both liquids, while the roots 

 are permeable to water but not sugar, in other words, are semi- 

 permeable. In pursuance of this matter we again turn to physics 

 and ask whether semi-permeable membranes are known, and 

 we find that they are; so here arises this problem: 



What are the osmotic phenomena where membranes are semi- 

 permeable? 



Experiment. Fill an upright test-tube or equivalent with a 5% 

 solution of potassium ferrocyanide (using care, for it is poisonous), 

 and into it quickly drop a small lump of copper chloride, which should 

 sink to the bottom. Observe the growth of the resulting membrane, 

 the physical conditions of which should be understood with a clear- 

 ness permitting their diagrammatic representation. (The structure 

 formed is sometimes called, from its discoverer, the Traube cell.) 



Experiment. Prepare an osmoscope as for the preceding experi- 

 ment, preferably the diffusion-shell form F. Immerse the well- 

 soaked cup, attached to its tube, in a 3% solution of copper sulphate, 

 and leave it to soak for some hours, preferably also exhausting the 

 air while it is immersed. Then empty the cup, rinse it in distilled 

 water, refill it to the glass with the potassium-ferrocyanide solution, 

 replace it in the copper-sulphate solution, bring the two liquids to 

 one level, and leave it twenty-four hours (to form the semi-permeable 

 membrane). Then pour out the liquid and replace it by molasses 

 or other selected solution; add, by a rubber-tube connection, a tube 

 of the diameter of the glass ring, immerse the cup in water and observe 

 the rise of the liquid until it stops; then test the water outside for 

 the presence of sugar. 



Semi-permeable Membranes. Only a few of these are known, includ- 

 ing those formed of calcium phosphate (made as recommended by Detmer, 

 147, and by Darwin and Acton, 124), of gelatin (or mucilage), and of tannin 

 (Pfefeer, 1, 106), but the best known and most useful is that of potassium 

 ferrocyanide, made as indicated by the experiments above, after the follow-/ 

 ing reaction: K 1 Fe(CN) 6 + 2CuS0 4 = Cu 2 Fe(CN) 6 + 2K 2 S0 4 . The mem- 

 brane forms only when the contact of the substances is gradual, for if they ' 

 are mixed in solution there results only a flocculent brown precipitate. The 

 membrane unfortunately is very fragile, and hence for experimental purposes 

 must be supported by some firm substance. For this purpose parchment 

 serves well so long as no pressure develops, but as soon as it is stretched by, 

 pressure the delicate membranes break. Hence for all experiments involv- 

 ing pressure it is needful to support them upon an unyielding substance, 

 for which porous porcelain, the finer in texture the better, is used. There 

 is also an electrolytic method, somewhat too complicated for our present ' 



