We study the cluster mass function in mixed dark matter (MDM) models, using
two COBE-normalized simulations with Omega(h) = 0.26, n = 1.2 and Omega(h)
= 0.14, n = 1.05, both with two massive neutrinos (models MDM1 and MDM2, r
espectively). For the sake of comparison, we also simulate a tilted cold da
rk matter model with spectral index n = 0.8 (TCDM), also COBE normalized. W
e argue that, in our nonhydrodynamical simulations, where cold dark matter
(CDM) particles describe both actual CDM and baryons, the galaxy distributi
on traces CDM particles. Therefore, we use them to define clusters and thei
r velocities to work out cluster masses. Since CDM particles are more clust
ered than hot dark matter (HDM) and therefore have, on average, greater vel
ocities, this leads to significant differences from Press & Schechter (PS)
predictions. Such predictions agree with simulations if both HDM and CDM ar
e used to define clusters. If this criterion is adopted, however, we see th
at (1) MDM corresponds to delta(c) values slightly but systematically great
er than CDM; and (2) such delta(c) exhibit a scale dependence: on scales si
milar to 10(14) M., we find delta(c) similar to 1.7 or 1.8 for CDM or MDM,
respectively, while at greater scales the required delta(c) decreases, and
a substantial cluster excess is found at the large-mass end (M > 10(15) M.)
. Clusters defined through CDM in MDM models, on the other hand, are less n
umerous than PS estimates by a factor of similar to 0.3 at the low-mass end
; the factor becomes similar to 0.6-0.8, depending on the mix, on intermedi
ate-mass scales (similar to 4-5 h(-1) 10(14) M.), and almost vanishes on th
e high-mass end. Therefore, (1) MDM models expected to overproduce clusters
over intermediate scales are viable; (2) the greater reduction factor at s
mall scales agrees with the observational data dependence on the cluster ma
ss M (which, however, may be partially due to sample incompleteness); (3) t
he higher spectral normalization is felt at large scales, where MDM models
produce more objects (hence, large clusters) than CDM. MDM1 even exceeds th
e findings of Donahue et al., while MDM2 is consistent with them. Simulatio
ns are performed using a parallel algorithm worked out from the Couchman AP
3M serial code, but allowing for different particle masses and used with va
riable time steps. This allowed us to simulate a cubic box with sides of 36
0 h(-1) Mpc, reaching a Plummer resolution of 40.6 h(-1) kpc, using (3 x)18
0(3) particles.