I would be interested to know people's opinion on the following matter.
Often, when the far-from equilibrium behaviour of a chemical reaction system is analysed mathematically (in papers or text books), the system is placed in a gradient between a fixed non-zero concentration of "resource" species, and a fixed concentration of "waste" species (which is often zero). The other species in the reaction system are designated as "intermediates", and it is their behaviour which is of prime interest.
For example, for the oscillating Lotka-Volterra reaction system:
$R + X \rightarrow 2X$
$X + Y \rightarrow 2Y$
$Y \rightarrow W$
Species $R$ and $W$ are the fixed concentration resource and waste species respectively, and species $X$ and $Y$ are the (oscillating) intermediates.
It appears as if these reaction systems are being studied in a physical scenario which could not be realised in the lab, a magical reaction tank if you like.
The two reasons most often stated as to why resource and waste are held at constant concentrations, whilst the intermediates can vary, are that:
1) The reaction takes place in a large reservoir, where resource is abundant and where any waste produced by the reaction quickly diffuses away into the vastness
2) The reaction has resource species quickly "pumped in" by some external agent, and likewise waste species quickly "pumped away".
My problems with these explanations are as follows:
If in explanation 1, a very large reservoir is keeping resource and waste species at constant concentration, then if follows that the intermediate species can never reach any appreciable concentration, for they too exist in the same large volume. In other words, if the waste diffuses away, then why don't the intermediates?
Conversely, in explanation 2, if a pumping process is artificially maintaining resource and waste concentrations, it is hard to see how this pumping process is being "micro selective" in completely removing the waste compounds, but leaving the intermediate compounds totally unaffected.
More physically realistic models of chemical reactions can be made by also explicitly including the reactor in the model - for example by considering the reaction as taking place in semi-permeable compartments or in bulk in a CSTR. However, additionally modelling the reactor often changes the mathematical properties for the "bare" reaction reported in the literature (for example, in a CSTR, the intermediates as well as the waste are removed).