The galaxy density field extracted from a complete redshift survey of
IRAS galaxies brighter than 1.936 Jy is compared with the mass-density
field reconstructed by the POTENT procedure from the observed peculia
r velocities of 493 objects. Both density fields have been filtered wi
th a Gaussian of smoothing length 1200 km s-1. Noise considerations li
mit the present analysis to a volume approximately (5300 km s-1)3 cont
aining approximately 12 independent density samples. We find a strong
correlation between the galaxy and mass-density fields; both feature t
he Great Attractor, part of the Perseus-Pisces supercluster, and the l
arge void between them. Monte Carlo noise simulations show that the da
ta are consistent with the hypotheses that the smoothed fluctuations o
f galaxy and mass densities at each point are proportional to each oth
er with a '' biasing '' factor of IRAS galaxies b(I), and that the pec
uliar velocity field is related to the mass-density field as expected
according to gravitational instability theory. Under these hypotheses,
the two density fields can be related by specifying two parameters, b
(I), and the cosmological density parameter OMEGA. The Monte Carlo sim
ulations are then used to estimate the random errors, to correct for s
ystematic errors in POTENT, and to constrain the parameters via a like
lihood analysis. Our strongest result is OMEGA0.6/b(I) = 1.28(+0.75/-0
.59) at 95% confidence. Small nonlinear effects allow weaker, separate
constraints on OMEGA and on b(I). Thus, if OMEGA = 1, then b(I) = 0.7
(+0.6/-0.2), and if b(I) > 0.5, then OMEGA > 0.46, both at the 95% con
fidence level. Inhomogeneous Malmquist bias could cause an overestimat
e of OMEGA; our 95% confidence limit for b(I) > 0.5 could be reduced b
y correction for this bias at most to OMEGA > 0.3. The constraints on
OMEGA are limited to the simple biasing relation assumed, but the effe
ct of undersampling cluster cores by IRAS is negligible, and the resul
ts are independent of the cosmological constant LAMBDA.