In the appropriate size range it can be argued that the absolute value of a
negative curling nucleation field in a prolate spheroid is a good approxim
ation for the switching field in any regularly shaped ferromagnetic body, w
hich has the same volume, provided the correct demagnetizing field is used.
An example is given for the case of a square prism. The discrepancy betwee
n this theory and some experimental results is claimed to be due to volume
and surface anisotropies in the experimental particles. (C) 1999 American I
nstitute of Physics. [S0021-8979(99)08414-5].