Figure 4
shows results obtained using IsoDen for Model 3, using
both statistical and evaporative methods to reject spurious halos.
In this particular case all of the halos found by IsoDen
turn out to be real, in terms of their internal velocities
(when examined individually).
This is the case for both methods of noise suppression.
Also, IsoDen distinguishes halos even in the high density central
region of the cluster -- the single large central halo found by
FOF is split into 9 smaller halos by IsoDen (or 10 halos using the
evaporative test).
All halos found by FOF using 
or evaporative
testing are also found by IsoDen.
The only difference between the IsoDen results using the
two noise suppression tests is that the evaporative method
identifies three extra halos which did not pass the significance
test with n=3.
Figure:
Results of the IsoDen method for Model 3, using the
same format as Figure 1.
The panels on the left show results using the statistical noise
suppression test; those on the right use the evaporative test.
For each halo only those particles with densities between 
,
the halo peak (central) density, and 
, the density of the first
overlap of that halo, are shown.