We analyze the three catalogs of nearby loose groups compiled by A. M. Garc
ia. She identified groups in a magnitude-limited redshift galaxy catalog, w
hich covers about similar to 2/3 of sky within cz = 5500 km s(-1) using two
methods, a percolation method and a hierarchical method. The free paramete
rs of the group-selection algorithms were tuned to obtain similar catalogs
of groups. The author also proposed a third catalog of groups, defined as a
combination of the two. Each catalog contains almost 500 groups. In agreem
ent with previous works on earlier catalogs, we find that groups can be des
cribed as collapsing systems. Their sampled size is in general considerably
larger than their expected virialized region. We compute the virial masses
and correct them by taking into account the young dynamical status of thes
e groups. We estimate corrected group masses, M, for two reference cosmolog
ical models, a flat one with a matter density parameter Omega(0) = 1 and an
open one with Omega(0) = 0.2. We calculate the mass function for each of t
he three catalogs. We find that the amplitude of the mass function is not v
ery sensitive to the choice of the group-identification algorithm. The numb
er density of groups with M > 9 x 10(12) h(-1) M., which is the adopted lim
it of sample completeness, ranges in the interval 1.3-1.9 x 10(-3) h(3) Mpc
(-3) for Omega(0) = 1, and it is about a factor of 15% lower for Omega(0) =
0.2. The mass functions of the hierarchical and combined catalogs have ess
entially the same shape, while the mass function of the percolation catalog
shows a flattening toward large masses. However, the difference decreases
if we do not consider the most massive groups, for which reliable results c
ome from galaxy cluster studies. After having estimated the mass contained
within the central, presumably virialized, regions of groups by adopting a
reduction in mass of similar to 30%-40%, we make a comparison with the resu
lts from the virial analysis of nearby rich clusters. All three group mass
functions turn out to be a smooth extrapolation of the cluster mass functio
n at M < 4 x 10(14) h(-1) M., which is the completeness Limit of the cluste
r sample. The resulting optical virial mass function of galaxy systems, whi
ch extends over 2 orders of magnitude, is fitted to a Schechter expression
with a slope of similar to 1.5 and a characteristic mass of M* similar to 3
x 10(14) h(-1) M.. We also verify that our group mass function agrees reas
onably well with the Press-Schechter predictions of models which at large m
asses describe the virial mass function of clusters.