366 



Scientific Proceedings^ Royal Dublin Society. 



The total area of the lumina of the cells was only 3'1 per cent, of the whole 

 cross-section. That of the tracheae was 44-6 per cent., that of the walls 

 52'2 per cent. 



Table VI. 

 8alix habylonica. 



8aKx habylonica also has a verj' small proportion of cells, as may be 

 seen from inspection of Table VI. With Salix injection caused no increase 

 in the rate of percolation. This may probably be explained, not because 

 the wood was already full of water — for, in fact, it was dry, and floated — but 

 because the leak took place principally through the large vessels which 

 were full of water, while the air was confined to the tracheids, which in 

 any case transmit water comparatively slowly (11). The amount of leak 

 per sq. cm. was 0'122 g. per minute through the fresh, and 0'115 g. through 

 the injected stem. Protoplasmic streaming in tlie cells to overcome these 

 leaks would require velocities of 4'299 cm. and 4'032 cm. per minute 

 respectively. 



Table VII. 



Sijrvnga vulgaris. 



Syringa vulgaris formed our last subject for these measui-ements ; and 

 in Table VII the results are recorded. The wood was sappy, and sank in 

 water ; but, notwithstanding this, the injected specimen showed a considerably 

 larger leak than the fresh one. The leak for the fresh was 0-082 g. and that 

 for the iajeoted specimen was O'l 29 g. per minute per sq. cm. From these 

 figures Janse's hypothesis would demand a velocity of protoplasmic streaming 



