Fellow confocalists,
 
In reply to my own question, further contemplation of the (apparent)
problem has made me realise the reason why the denominator wavelength is
centred upon the "second" peak of the spectral response curves, and not the
isosbestic point. From the way the curves typically cross over, the result
is that for any one ion concentration there is still a unique ratio. So all
one has to do is calibrate and everything is OK.  Actually by choosing the
second peak for the denominator wavelength, one gets the maximum spread of
ratios for differing ion concentrations, so it is  the best point to use.
 
Please accept apologies for my defamatory allegations of ovine behaviour
amongst my esteemed colleagues. I can only say that bleating my problems to
the world has at least caused a rush of enough adrenaline to my own brain
to allow it to understand the obvious.
 
Geoff "Silent Lamb" Hyde.