On the subject of the possible Fermi surface
topological effect involving necks that Norman and
Morr were postulating, it's difficult to see how this
can work when the maximal cross-section of the
(spherical) Fermi surface is nearly independent of
orientation. In the absence of any significant
orientation depedence to the Fermi wave-vector, there
would have to be an enormous deviation from parabolic
behaviour in order to give a mass enhancement at the
`hot spots' that is more than a factor of 10 greater
than everywhere else on the sphere. 

It seems that very little attention has been given to
the possibility of other fluctuations modes of the
antiferromagnetic phase. I wonder if the correlation
length (v_F/T (Fermi velocity/temperature)) might be
something to consider (which determines the lenthscale
over which the order of a SDW or CDW is correlated).
Although the AFM phase of CeIn3 isn't a SDW, the
effect of Bragg reflection from the the ordered local
moments induces something analogous to a virtual SDW
for the conduction electrons acording to Morr. From a
naive perspective, the correlationn length ought to
give an imaginery component to the periodicity vector
proportional to T/v_F that will ultimately follow
through to changes in the Onsager phase of the
quasiparticles at points of Bragg reflection. Its
temperature dependence might have the effect of
leading to an effective mass enhancement? although I
have no idea how the correlation length might change
on the approach to a QCP. Although the hot spots don't
coincide precisely with points of Bragg reflection,
the Fermi surface is modified in a qualitatively
similar manner as what would happen due to Bragg
reflection (i.e. reconstruction of the Fermi surface
leads to a local modification of the Onsager phase
within the AFM phase). This latter point pertains to
experimental observations,