Figure 1 summarizes the four models for a range of effective stellar
temperatures 
. The points are the average value of
each quantity over the first 100¸ out of the plane. The
vertical lines are the corresponding 
variations from this
mean. Regions which are excluded by the observational constraints
discussed in §1 and summarized in Table 1 have been shaded gray.
Notice that none of the models with only photoionization heating
(dashed lines) satisfy the observational constraints. Lower values of
predict electron temperatures which are too low while
higher values predict too much ionized helium. In all cases the
and 
ratios are either too low or near
the lower limits in Table 1.
When the turbulent heating term is included (solid lines) there are
many effective temperatures which satisfy all of the observational
constraints. The He observations constrain the radiation field to be
softer than 
for all models. This limit has a slight
dependence on the density (see Figure 1). The higher density models
produce lower electron temperatures, especially for the models where
turbulent heating is included. The and
ratios predicted by the model are consistent with the observations in
Table 1 only when the turbulent heating is included. These ratios
also have a moderate dependence on the density.
Moreover, the linewidths predicted by the this model
are consistent with observations in the DIG.
All four models predict that the 
ratio is < 0.005,
while Reynolds (1985b) measured an ratio of 0.06 in
the diffuse ISM. The origin of this 
emission is not known.
Reynolds suggested that it could come from the DIG, from hot 
gas along the line of sight, or from emission
originating in H 
regions and reflected off of dust grains. Our
models suggests that the observed emission does not originate
from the oxygen in the DIG. Reynolds (1990a) found an upper limit for
the 
ratio of 
. In all cases, our
models predict an ratio at least one order of magnitude
below this limit. Our models also predict that the
line ratio ranges from 0.15 at 
to 0.85 at
. This ratio is at least five times greater than
the values predicted by the models without turbulent heating.
A simple physical picture of the DIG evolves naturally from these
results. The 
and 
recombination lines are produced almost
entirely by the stellar radiation field because the turbulent heating
does not produce enough energy to ionize these species. The forbidden
lines, however, are excited primarily by collisions; thus the addition
of thermal heating via the dissipation of turbulence increases the
and 
intensities as well as the kinetic temperature.
Therefore, turbulent heating provides enough additional heating to
explain the observed conditions in the DIG while remaining negligible
in ``classic'' H regions. Models of classic H regions which
include the turbulent heating discussed in §2 confirm
this conclusion.
Another heating mechanism which could be important in the DIG is
photoelectric heating from dust grains ([Reynolds & Cox 1992, Draine 1978]). This
mechanism provides approximately the same heating rate as does the turbulent
heating up to 
in a gas where the hydrogen is
fully ionized. For 
the heating due to the grains falls off quickly due to cooling of the gas
via collisions with the cooler dust grains ([Draine 1978]). In our models,
the dust grains' heating rate is 
of the total heating rate of
the DIG.
From the measured parameters of the local ISM 
from the
Sun; Wood & Linsky 1997 and [Frisch & Slavin]
,
we estimate that the turbulent heating rate
is 
of the cooling rate in the local interstellar cloud (LIC).
The photoelectric heating from dust grains should also be important in the
LIC ([Frisch & Slavin]).
Thus the
dissipation of turbulence should be important in the local ISM and
may also be important in other phases of the ISM.
We would like to thank Ron Reynolds for discussions on the origin of
the emission and the photoelectric heating via grains,
Gary Ferland for helpful discussions about
CLOUDY and Jay Lockman for comments on the
manuscript. We also thank Chris McKee and Don Cox for their comments.