642 R. J. GOLDACRE 



living membranes, though Hving membranes, using metabolic energy, usually 

 have it in their active transport mechanisms ; passive diffusion remains symmetrical, 

 otherwise thermodynamic laws would be violated, as Mitchell indicates. On the 

 other hand expansion-contraction cycles can in fact concentrate adsorbable 

 substances (i.e. change their free-energy state). This happens in the working 

 model I described [Internat. Rev. Cytology, I, 135 (1952)] — " the inversion tube " — 

 which by adsorbing substances on a large surface and subsequently desorbing 

 them by contracting the surface mechanically, in a small volume, can give you a 

 very large increase in concentration. The energy for the concentration comes from 

 the unequal energies of expansion and contraction of the surface — in contraction, 

 there are adsorbed molecules to squeeze off the surface, and work must be done 

 against the energy of adsorption. In the mechanical model, the energy provided 

 is mechanical ; in the living cell, the energy source of the contraction-expansion 

 cycle would be metabolic (probably ATP) as in muscle. 



Davis : May I direct a comment to Dr. Mitchell ? Let us consider a model in 

 which a site with a certain affinity for the permeant is also attached to something 

 contractile ; the contraction would distort the site and thereby decrease its affinity. 

 You are not getting work for nothing, as ATP energy would be expended in the 

 contraction. Are you sure that Maxwell's demon is invoked in this kind of model ? 



Mitchell : This is evidently a difficult matter to discuss, for it has been 

 chewed over many times and there is not yet general agreement about the con- 

 clusions — especially, I believe, since the concepts that we must depend upon are 

 still in process of formulation. I think 1 would look at it like this. When you 

 postulate a macroscopic model of the propulsion process in membrane transport 

 by considering something like a piece of elastic and how you can manipulate it as 

 a catapult, you are likely to run straight into a conceptual difficulty. Of course, 

 when you let go a piece of elastic it goes flip straight away and one does not think 

 of the thermal activation of this contraction process. If a protein molecule becomes 

 "stretched" or unrolled, the regaining of the configuration that it originally had 

 is, in fact, a diffusion process which must occur by the making and breaking of 

 residual valencies. Moreover, the "stretching" during the phase of the process 

 when it actually occurs must also be a diffusion (or down hill) process, since it 

 would not otherwise happen. In this sense a "contraction" and "expansion" 

 process can be associated with the movement of an ion or other particle through 

 a membrane; but if we are going to "take" the ion and "put it on" to such a 

 system we have to change the ion to "put it on ". It has to be attached, not by hand 

 (as a stone may be put in a catapult), but by a chemical bond. When the elastic 

 has got through the membrane (in the ion or molecule type of transport that we 

 are considering here, but not, of course, in group transport) the carried particle 

 has to be detached again — the bond must be broken. The actual work done on the 

 particle in the transport is determined by the difference in free energy in attaching 

 and detaching, and it has no necessary connection with the change of configuration 

 of the elastic which happened in between. You may say, as Dr. Davis has just 

 done, that the change of configuration is related in some way to the affinity between 

 the carrier and the particle, but this does not make the actual change of configura- 

 tion responsible for the work of transport. Although the difficulty with which we 

 are confronted is undoubtedly partly conceptual, it is not, as has sometimes been 



