### 4. DISCUSSION

The first five problems on our list are the most generic and intractable in DM
models. Problems 1, 2 & 4 do not merely arise from mismatches between theory
and observation, but require fine tuning in any DM scenario. Any physical
process to account for them must somehow couple the dark and luminous
components. If the only interaction between the baryons and DM is
gravitational, it is hard to imagine how such a mechanism could arise.

Problems 7-10 are well-rehearsed difficulties of CDM. The possibility that 7
& 8 could be solved by a more realistic treatment of baryonic processes alone
is remote since they relate to the total mass distribution which, in the
simulations, is dominated by DM. Even if the baryonic mass fraction is
significant, it is unclear what process could loosen tight DM concentrations;
for example,
Debattista & Sellwood
(2000)
show that an intolerable degree of
dynamical friction has a very mild effect on the halo density profile.
Problems
9 & 10 could conceivably be solved by some kind of complex baryonic physics
(i.e. star formation), but this solution to 10 would require more fine tuning.

Another approach to problems 7-10 is to modify the assumed properties of
the DM particles. Warm DM (e.g.
Colombi, Dodelson &
Widrow 1996;
Sommer-Larsen & Dolgov
1999;
Hogan 1999)
is designed to aid problems 7 & 9 in two ways: WDM
particles stream freely in the early universe suppressing small-scale power.
They also have larger velocity spreads in halos of greater volume density,
because of Liouville's theorem, thereby precluding strong density gradients.
However, fermionic WDM falls foul of problem 6, while bosonic warm DM,
which has
a small fraction of very high phase space density material (e.g.
Madsen 2000),
would probably again form cusped halos as in CDM. Self-interacting dark
matter
(Spergel & Steinhardt
1999)
can in principle solve problems 7 & 9
(Burkert 2000;
Davé et al. 2000)
but only if the cross section is tuned; it may also
have difficulties with its prediction of spherical halos. We emphasize that
hypothetical modifications to the properties of DM particles have been
motivated
mostly by problems 7 & 9, and any collisional effect removes 6, but none
address the much more difficult problems 1-5.