Through the (J,K) = (1,1) and (2,2) inversion transitions of the 
\
molecule we have studied the dense gas in a sample of 15 regions with
signs of star formation, as indicated by the presence of outflow activity.
We have been able to detect and map the 
emission in 14 of these
regions. This high rate of detections is a clear indication of the strong
association between 
emission and outflow activity. This result
confirms the young nature of the observed objects, since they appear to be
still associated with (and most of them embedded in) the dense gas from
where they have been formed.
In all cases where we have mapped the 
emission associated with
molecular outflows, the ammonia maximum is located near (
) the
position of a candidate for driving the outflow. This result gives
support to their identification as the outflow exciting sources, following
the criterion proposed by Anglada et al. (1989) for such identification.
For the sources of our sample that are only associated with optical signs
of outflow, the 
emission is weak and does not coincide with any known
object in the observed field, suggesting that probably none of the nearby
objects that have been detected up to now is related to the outflow
excitation.
The sizes of the condensations we have mapped in 
range from 
to 
(a somewhat larger value of 2 pc is obtained for the
condensation associated with IRAS 20188+3928, but this value may be
overestimated by an order of magnitude, since we have adopted the upper
limit for the distance). We find evidence for some elongation in the
condensations we mapped (as was noted, in general, by Myers et al. 1991).
However, our angular resolution, in general, is not good enough to allow
us to further discuss on the morphology of the sources. In particular, we
are not able to establish whether or not these condensations play a
relevant role in the collimation of the outflows in these regions (as
suggested, e.g., by Torrelles et al. 1983). We note, however, the very
high degree of elongation of the condensation associated with IRAS
22376+7455 in L1251, reminiscent of the 
structure observed in L1448
(e.g., Anglada et al. 1989). In both cases (L1251 and L1448), several
objects are seen in projection towards the elongated 
structure. A
high angular resolution study of the L1251 region may be relevant in order
to investigate a possible fragmentation of the structure. A higher
angular resolution study may also be relevant for other sources of our
sample that appear unresolved in our present single-dish study.
The intrinsic line widths obtained (
; see Table 2 (click here)) are
significantly larger than the expected thermal line widths, since the
kinetic temperature estimates for these regions give very low values (see
Table 3 (click here)). Only for the region associated with IRAS 20188+3928 the line
widths reach 
(in some positions). This is also the region
where we have obtained the highest value for the kinetic temperature, as
estimated from our ammonia observations (we note that in many sources we
only could obtain upper limits or we were forced to use estimates from CO
data, which are less accurate). Despite the uncertainty in the distance,
the luminosity of IRAS 20188+3928 appears to be higher than for most of
the objects studied in our sample. Thus, this object appears to produce a
larger perturbation in its molecular high density environment than the
others objects we have observed.
In part due to our lack of angular resolution, we are not able to measure
in detail the velocity gradient in our regions. However, it is remarkable
that in the condensation associated with IRAS 22376+7455 in L1251, our
results show the presence of a strong velocity gradient with sudden
velocity shifts of up to 
between contiguous positions. A
study with high angular resolution of this region appears to be very
promising.
The 
column densities we have obtained are 
(assuming [
/
] 
), implying mean visual extinctions
. For L483 we have obtained the highest 
column density
(
, corresponding to a visual extinction of
), suggesting that this object is very deeply embedded.
The masses obtained for the observed regions lie typically in the range
from 1 to 100 
. The values derived coincide, in general, with the
virial masses within a factor of 3. This general trend, observed for the
regions of our sample, suggests that the condensations are near virial
equilibrium and that the assumed 
abundance is adequate. L483 is the
region for which the calculated mass exceeds the virial mass by the
largest factor. This could imply that for this source the cloud is still
in the process of gravitational collapse. We note here that recently
Myers et al. (1995) have detected asymmetric line profiles in this region
consistent with infall motion, according to the modeling of Anglada et
al. (1987). Although the uncertainties involved are still large, it
seems clear that this object is among the youngest sources in our sample,
in agreement with the results of Fuller et al. (1995), which classify
this as a very young object.
We have detected and mapped the 
emission in 12 out of 13 regions with
molecular outflow in our sample. Only in one region, L1048 (for which, in
fact, no published map of the CO outflow is available), we failed in
detecting ammonia emission (Table 1 (click here)). The 
emission is faint (
; see Table 2 (click here)) only in 3 regions (RNO 43, HH 83 and L100) of the
12 regions associated with molecular outflow we have mapped. On the other
hand, in the two regions without molecular outflow (HH 84 and HH
86/87/88), the ammonia emission is very faint. These results tentatively
suggest that the ammonia emission tends to be more intense for those
sources associated with molecular outflow than for the sources associated
with only ``optical'' signs of outflow (such as jets and Herbig-Haro
objects).
In order to substantiate this possible relationship between the type of
outflow and the intensity of the 
emission, we have complemented the
sample of regions observed in this paper with the results of other
Haystack 
observations reported in the literature. We have studied
the distribution of the intensity of the 
emission, as measured by the
main beam brightness temperature towards the outflow exciting source, in
this larger sample of regions. We note here that the 
brightness
temperature is a good measure of the intensity of the 
emission only
for sources that fill the beam of the telescope. For unresolved sources,
a more adequate comparison should be made in terms of the distance
corrected flux density of the ammonia emission (``ammonia luminosity'').
As we expect that for nearby sources the angular size of the ammonia
emission will be, in general, larger than the telescope beam, we have used
the main beam brightness temperature to make the comparison, restricting
our sample to nearby enough regions. Thus, we have used sources with
, and completed our sample with the data from Torrelles et
al. (1983) (9 sources), Anglada et al. (1989) (13 sources),
Benson
& Myers (1989) (5 sources, and 2 additional sources observed at Green Bank),
Verdes-Montenegro et al. (1989) (3 sources), and Persi et al.\
(1994) (1 source).
Our final sample is shown in Table 4 (click here). It contains a total of 47 sources,
with 21 sources associated only with molecular outflow, 19 sources
associated both with optical and molecular outflow and 7 sources with only
optical outflow. In Fig. 21 (click here) we present the distribution of the 
main
beam brightness temperature towards the position of the proposed outflow
exciting source (Table 4 (click here)) for the three groups of sources. The mean values
of the 
brightness temperature are 1.7 K (only molecular outflow), 1.5
K (optical and molecular outflow) and 0.5 K (only optical outflow).
Despite the relatively small number of sources with only optical outflow,
it is clear that these sources tend to present lower values for the 
\
brightness temperature, while for the sources with molecular outflow the
distribution is displaced to higher values of the 
brightness
temperature. We conclude, thus, that the ammonia emission is in
general more intense in molecular outflow sources than in sources with
optical outflow.
Figure 21: Distribution of the 
main beam brightness
temperature for sources with only molecular outflow (top), for sources
with both molecular and optical outflow (middle), and for sources with
only optical outflow (bottom)
Table 4: Regions associated with molecular or optical outflow
observed in 
We should note that recent sensitive studies have detected weak CO outflows in regions where previous studies failed in the detection (e.g., in HH 1-2 or HH 34; Chernin & Masson 1995). We have not attempted to take into account the effect of the intensity of the molecular outflow in our study, and we have only considered whether or not an outflow detection has been reported in a given region.
The fact that the sources of molecular outflow present more intense
ammonia emission can be interpreted as indicating that these sources are
deeply embedded in the high density gas, and surrounded by a larger amount
of molecular gas, while those sources with only optical outflow have
already dispersed the molecular core or escaped from it. This
interpretation can be corroborated by comparing the estimated column
density in the sources listed in Table 4 (click here). We found that the ammonia
column density towards the outflow exciting source decreases as the
outflow activity becomes prominent in the optical. The mean values of the
column density are 
(only molecular
outflow), 
(optical and molecular outflow)
and 
(only optical outflow). In Fig. 22 we
show the distribution of the 
column density for the three groups of
sources. A similar correlation is obtained if the comparison is made in
terms of the estimated mass of the associated core, but in this last case
it is unclear to what extent should the 
emission be considered as
associated with a given object.
Figure 22: Distribution of the 
column density for
sources with only molecular outflow (top), for sources with both molecular
and optical outflow (middle), and for sources with only optical outflow
(bottom)
These results suggest an evolutive sequence of the sources, traced by the intensity of the ammonia emission and the observational appearance of the outflow. Molecular and optical outflow would be phenomena that dominate, observationally, at different stages of the early stellar evolution. In the younger objects molecular outflows will be prominent, while optical outflows will progressively show up as the star evolves. However, this result does not exclude that both phenomena could coexist simultaneously as is required in the so-called ``unified models'', in which molecular outflows are driven by highly collimated jets (e.g., Raga et al. 1993); only the observational appearance of the outflow evolves in time as the star loses progressively the surrounding high density gas. In this scenario, the driving optical jet is becoming visible as a consequence of the ambient molecular material being progressively removed by the effect of the molecular outflow itself. Alternatively, the observed differences could represent intrinsic differences in the amount of molecular high density gas from one to another region.
