86 P.E.WEATHERLEY 



and because the flux of water is very possibly different in the detopped 

 plant from that in the intact transpiring plant. Sabinin (1925) showed that 

 the exuding root behaved hke an osmometer and the thorough study by 

 Arisz et al. (195 1) confirmed and greatly extended this conclusion. This 

 implies that the flux is through semi-permeable membranes. Yet De 

 Lavison (19 10) showed that some solutes could penetrate the cortex rapidly 

 as far as the endodermis, indicating a rapid movement in the cell walls 

 stopped by the casparian strips. Other solutes which could enter the 

 cytoplasm could pass the endodermis. Later Scott & Priestley (1928) 

 emphasised the endodermis as a barrier to free movement in the cell walls 

 of the cortex and in more recent years the apparent free space concept is 

 clearly in accord with rapid physical movement up to the endodermis 

 (Bernstein and Nieman, i960). 



The osmotic barrier indicated by Arisz et al. would presumably therefore 

 reside in the endodermis. Thus between the external medium and the 

 xylem of the root there will be a catena of at least three Hnks: (a) free 

 diffusion or probably mass flow up to the endodermis, (b) osmosis across 

 the endodermis and (c) probable mass flow beyond the endodermis. As is 

 well known, the flux through such resistances in series will be largely 

 controlled by the Unk with the highest resistance, provided its resistance is 

 much higher than that of the other links. Osmosis across the endodermis 

 is hkely to be such a high resistance Hnk, so that even if the pathway across 

 the cortex and within the stele is by mass flow, the flux will be doirdnated by 

 the endodermis, and the root after all will behave Hke an osmometer. But 

 can we exclude the possibihty under transpiring conditions, of a mass flow 

 through the cytoplasm of the endodermal cells, through passage cells or 

 even via the cell walls of the endodermis in regions of the root where the 

 casparian strips are not suberised; In this event we are faced with the 

 endodermal link itself presenting two pathways in parallel: one diflu- 

 sional, the other permitting mass flow. 



Theoretically it ought to be fairly easy to detect and even to measure 

 the relative resistances of two such parallel pathways. Movement in 

 response to an osmotic potential gradient should be less than that in 

 response to an equal hydrostatic pressure gradient, since the latter woidd 

 not only cause an osmotic flux equal to the former, but would also draw 

 water through the mass flow pathway in addition. Such a difference in the 

 case of a hving cellular barrier was demonstrated by Kramer (1932) using 

 a hollow petiole. The work described in the second part of this paper 

 represents an attempt to apply this approach to the root system (Mees and 

 Weatherley, 1957). 



