In addition to stabilizing buoyant radio-lobes against shredding, we
have discussed how viscosity can dramatically slow the evolution of
the bubble. This suggests a natural solution to the ``shock problem''
noted in the introduction. Under the action of viscosity, the AGN
jets may inflate their ICM bubbles subsonically before the buoyancy
effects become important. Consequently, the ICM can be displaced into
a cavity-bounding shell without the need of a strong shock. The
thermal structure of this gas will be determined by the action of
compression (due to its displacement from the cavity), decompression
(due to the fact that this material will typically be lifted to a
higher level in the ICM atmosphere), viscous dissipation, and
radiative cooling. One can envisage a scenario in which the rather
slow evolution of the radio-lobe in a viscous ICM allows the shell of
displaced ICM to radiatively cool by an appreciable amount. Further
simulations of the inflation stage of these bubbles are required to
assess whether the rather cool cavity-bounding shells observed in
Per-A can be reproduced within the context of this model. Given the
strong temperature dependence of the coefficient of viscosity
(eqn. 1), we might expect significant changes in the effective
viscosity between the various ICM structures defining the system
(i.e., the ambient material, the cool rims and the shocks, if any).
Thus, it will be important to move beyond the constant 
assumption in these next generation of simulations.