Let us now discuss two physical limitations of our
determination of 
as a function of frequency:
1. A straightforward consequence of our fitting method is that it is
impossible
to determine very small values of (i.e., 
),
because
in this case the antenna response is a weak function of except
for a multiplicative factor independent of 
.
Owing to the shape of the dispersion curves (Figure B1), this
restriction corresponds to observing frequencies a little below the
gyroharmonics (in addition, in this case 
is not small compared
to 
so that (1) is no longer valid).
2. Since the antenna moves with respect to the plasma
(this motion is mainly the corotation of the IPT with Jupiter),
the received waves are Doppler shifted.
How far does that effect bias the determination of ?
This point was qualitatively analyzed by
[Meyer-Vernet, Hoang and Moncuquet, 1993] concerning the frequency shift, but
there is also a Doppler effect on the shape of the antenna
response function 
given by (2).
We further discuss that point in appendix A.
We find that our determination is unaffected
by the Doppler shift as long as 
, but its value
is in most cases underestimated for larger
. Again, this occurs for frequencies close to the gyroharmonics
but this time above them.