The HII regions of NGC4736 are mainly distributed in a ring of
in diameter centered at the nucleus (Fig. 1 (click here)).
A series of short arms detach from the ring outwards. Inwards, we find
regions 55, 56 and 72 on a weak H
emitting arm, which twist itself
for more than 
, from Southwest to East, leaning against the
ring to the East. Weak H emission is also found in a featureless,
hook like structure, surrounding the nucleus to the East, which covers
an area of about 
. Images of the remaining
emission lines are presented in Fig. 2 (click here).
Table 2: Background brightness at each wavelength as measured in the
galaxy and standard star frames. The integration time in Col. 4
corresponds to the data of Col. 3. The integration time for Col. 2
is that quoted in Table 1 (click here)
Figure 3: The lines are 
.
Full line: 
; dotted lines: extreme values of
(500cm-3 and 40cm-3)
Figure 4: Numerical diameters distribution for D100. The diameter bins
are 
We have defined an isophotal diameter as D =
, with 
,
, where A and
are the areas inside the isophote of a defined brightness
for the HII region and in the half-maximum of a stellar profile
respectively. The isophotal
diameters have been measured on the deconvolved H image (Sect.
3 (click here)).
Figure 5: Cumulative diameters distribution of the HII regions of NGC4736.
Points shows the number of HII regions with diameters greater than the
value indicated in the horizontal axis. Full line: fit to the law
(see text) with D0=115 pc and a
correlation coefficient of 0.97
Figure 6: Differential luminosity function of NGC4736 HII regions.
Points: average number of HII regions per
0.2dex luminosity bin. Dashed line gives the fit to the full set of
points with the index of the power law 
and the correlation
coefficient is -0.85. Full lines: data set was fitted in two parts. In the
low luminosity part we obtained 
and a correlation
coefficient of -0.85; in the high luminosity part 
whith a
correlation correlation coefficient of -0.96
We obtained for each HII region a halo diameter (D100, Kennicut
1988), defined by the isophote at level 
,
roughly equivalent to the halo diameter used in distance scale
applications by Sandage & Tammann (1974). In Fig.
3 (click here) we show a plot of D100 vs. 
. The
average slope of the set of points is 3 (
, where C is a constant value; full line in the figure), similar to
that obtained by Kennicutt (1988) for the brightest
HII region of a sample of 95 nearby spiral and irregular galaxies.
As quoted by this author, for the simplest case of a constant density
ionization-bounded HII region, the Stromgren sphere should scale as
the cube root of the ionizing luminosity. But the average regression
in Fig. 3 (click here), also represents a locus of constant electron
density . We show in that figure the lines embracing 95% of the
HII regions which represent the extreme values of in this
sample. can be obtained from the following relations:
where N0912 is the number of
photons capable to ionize the Hydrogen. By comparing relation
(1 (click here)) with
where r1 is the Strömgren radius,
is the recombination coefficient to all excited levels, and
, the hydrogen atoms number density (Osterbrock
1989). We obtained for the average and the two extreme
values, 150
cm-3, 
cm-3 and 
cm-3 respectively. As a matter
of comparison, from Kennicutt's (1988, Fig. 6.) data we
calculated 
cm-3, 
cm-3 and 
cm-3
respectively. For the three regions in common with Kennicutt (regions 12,
16 and 87 in our sample), his mean is 
cm-3, while our
one is cm-3. The difference appear mainly from the measured
diameters. For these regions, Kennicutt's mean diameter is 
, while we obtained 
. The
deconvolution process might explain such a difference.
Figure 4 (click here) shows D100
histogram. The distribution of halo diameters is fairly symmetric and
peaks at 
.
Figure 7: Line ratio pairs relations.
a) 
vs. 
;
b) 
vs. 
;
c) 
vs. ;
d) vs. .
Full lines in a), b) and c) the MRS's predicting
equations. Black dots represents regions with 
Table 5: HII regions oxygen abundance 12+log (O/H) and temperature
index 
Figure 5 (click here) shows the cumulative distribution of D100.
It is fitted by the law (e.g. van
den Bergh 1981; Hodge 1983, 1987), where N (D)
represents the number of HII regions with a diameter larger than a
characteristic diameter D. A least square fit gives 
,
with a correlation coefficient of 0.97 and it is represented by the full
line in Fig. 5 (click here).
As a matter of comparison, for 
cm-3, D0 is
approximately five times larger than the radius of the Strömgren's sphere
of an O5 star, indicating that the largest HII regions are certainly
ionized by associations of massive stars. The total number of
HII regions furnished by the regression is 
. The fall
down of N(D) for 
with respect to the quoted
regression indicates that we detected only 2 to 5% of the weakest
HII regions.
Figure 8: HII regions internal
extinction C(H
) vs. 
.
Black dots represents HII regions with 
HII regions luminosity functions are generally well
represented by power laws d
dL, where dN(L)
is the number of HII regions emitting an H luminosity
between L and L+dL. In Fig. 6 (click here) we show the
differential luminosity function for the HII regions of NGC4736,
where the points are the average number of HII regions per
0.2dex luminosity bin.
The general properties coincide with that
obtained by Kennicutt et al. (1989). Basically, it deviates
strongly from a single power law, as happens with HII regions
in early-type spirals.
NGC4736, as M51, NGC3521 and NGC3627, presents a turn
over in the LF for HII regions
brighter than 
, a signature of
late-type spirals and irregulars. The number of this type of HII\
region is lower than that which would correspond to the extrapolation
of the power law representing the luminosity function of the faintest
HII regions. A
discussion about the implications of this property can be found
in Kennicutt et al. (1989).
Figure 9: Relation of the temperature parameter with H equivalent width
Figure 10: H -continuum luminosity vs. H intensity
compared to the evolutionary tracks of theoretical ionizing clusters
(full lines).
The models are:
;
;
;
;
.
The tick marks on each model represent the model runs
from right to left with t=0, 1, 3, 5, 7, 9 and 12 Myr
The Oxygen chemical abundance, 
, was estimated by
means of the ratio 
(Pagel et al.\
1979; McCall et al. 1985, hereafter MRS), calibrated
according to Zaritsky et al. (1994):
where 
. Since we observed only [OIII]
,
the total intensity was obtained from [OIII]
[OIII]. Table 5 (click here) gives the
set of derived Oxygen abundances, whose mean from 65
HII regions is 
with
.
Since the HII regions are practically located at the same radial
distance, we can not prove radial metallicity gradients.
In Fig. 7 (click here) we show MRS diagnostic diagrams, 
vs. , vs. , 
vs. , vs.
, where we plot our data set and MRS's predicting equations.
Except for vs. , the positions of
the HII regions brighter than 
(erg
) are in good agreement with MRS's predicting equations.
In Fig. 8 (click here) we plot the HII regions internal
extinction C(H ) vs. R23, which shows that higher
extinction corresponds with lower metallicities. A possible reason for
the observed correlation is that, on average, the higher metallicity, the
stronger are the stellar winds and consequently they are more efficient in blowing
out and breaking the dust grains within the HII region. We point
out that this correlation cannot be a consequence of the [OII]
extinction correction, since R23 is more influenced by
[OIII]/H as can be seen in Figs. 7 (click here)a and 7 (click here)d.
The effective temperature of the HII regions ionizing source 
can be estimated through the
parameter 
(Vı lchez & Pagel 1988),
which can be obtained from the observed line ratio
as
where t is the electronic temperature 
in units of 104 K.
is a good criterion of effective temperature of ionizing
stars and it is relatively insensitive to chemical
composition and ionization conditions found in observed nebulae,
particularly when measurements of the [SIII] lines in the far red
are available, as happens in our case.
Assuming t=1,
we can compute [SIII]
= 3.48
[SIII]
(Mendoza & Zeippen 1982).
Table 5 (click here) gives the values of for 45 HII regions.
Figure 9 (click here) gives the distribution of the HII regions in the
plane 
vs. 
. If the formation of stars in
the ionizing cluster occurs in a burst, the evolution of the HII regions
as a function of time in this diagram will be in the sense of
decreasing and increasing (decreasing 
)
(Copetti et al. 1986, CPD). The full line in Fig.
9 (click here) limits the highest values. It is the
locus of the youngest HII regions for a given , according to CPD.
As a consequence, the upper envelope of the data points
(dashed line) would represent the evolutionary track or fading of the
most luminous HII regions (black dots, 
).
The lower envelope (dotted line) shows a higher slope than the upper one,
indicating that the less massive HII regions suffer more drastic
change in temperature during their fading. In the next section we carry
out a quantitative analysis of the HII regions evolution.
Figure 11: Continua images at 
Å, 
Å,
Å, 
Å and 
Å. North is
up, East is to the left. A saw-tooth lookup table was used; levels in
the color bars are logarithm of flux density 
Figure 12: Result of the isophote fits to the Å image.
Left pannel: the isophotal map; brightness levels are 15.4,
16.0, 16.6, 17.0, 17.6, 18.0, 18.2, 18.5, 18.6, 18.8, 19.2, 19.8, and
21.0 magarcsec-2. Right pannel: fitting for the same
levels, which corresponds to ellipses with semi-major axis length of
10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 130, 170 and 220 arcsec
Figure 13: Ellipticities and position angles vs. distance along
the semi-major axis in Å image. Error bars are
obtained from the residual intensity scatter, combined
with the internal error in the harmonic fit, after removal of the
first and second fitted harmonics
Figure 14: Radial brightness profiles (data points) and its best fits
(full lines) for a), b),
c), d) and
e). Individual components of each fit are shown separatelly:
de Vaucouleurs' profiles in dotted lines; Freeman profiles in dashed
lines. We give the values of 
, 
(arcsec), 
,
(mag) and the RMS deviation to each fit. Error bars were
calculated directly from the RMS scatter in luminosity along the fitted
ellipse. f): Zoom of the central part of d), showing the fit
and data set behaviour
We discuss the HII region distribution in
the plane H intensity vs. H -continuum luminosity, which is
very suitable to study the ionizing clusters properties, because the
H intensity
measures the amount of ionizing photons furnished directly by the more
massive
stars in the Main Sequence, while the H -continuum luminosity
is mainly contributed by low-mass stars and evolved massive stars.
In this context, we derive age and mass of the HII region ionizing
clusters. The data set is compared to theoretical models of the
ionizing clusters, which are computed with the
"ET" code (Cid-Fernandes et al. 1992), which retrives the
temporal evolution of a cluster spectrum by inputting (i) An IMF formed in a
burst, (ii) Stellar evolutionary tracks and (iii) Model atmospheres,
besides the cluster total mass 
. We assume as input data for
"ET": (i) A Salpeter's IMF with a lower mass
limit 
and upper mass limit
, for clusters with total mass
; , for those
with 
; and ,
for those with 
, suitable for the range of
sources temperature derived in the previous section.
Particularly, for the two last cases, a larger 
would
require a fractional number of most massive star.
(ii) The set of stellar evolutionary tracks of Maeder &
Meynet (1988). (iii) The model atmospheres of Kurucz
(1979). For a theoretical cluster of a given mass, "ET" furnishes,
beside other data, the H -continuum luminosity and the H -line
intensity, which are computed by assuming a radiation bounded HII region
(case B of Backer & Mentzel 1936).
In Fig. 10 (click here)
we have plotted the data set in the plane
H -continuum luminosity vs. H intensity, where we have
superimposed the evolutionary tracks of cluster models, up to 12 Myr.
It becomes clear that the right-hand side
limit of the data set distribution corresponds to a Zero Age Sequence
for clusters
of different masses. The more massive ionizing clusters are those of
regions 5, 51
and 87 (
), with ages of 
5, 6 and 8
Myr, respectively.
The models indicate that the H intensity of
12 Myr old ionizing clusters, more massive than 
,
should be detectable
in our observations (sensitivity 
),
nevertheless we fail to. It might indicate that clusters as massive
as those produce enough Supernovae and massive star winds to blow
away the surrounding gas,
transforming the HII region after 9Myr in an optically thin one,
which would produce an
additional fadding of the emitting gas due to the leakage of ionizing
photons. Regions 3 and 54 are just two border cases of the fadding
process.
Figure 15: Result of subtraction of a bulge + 2 disks two dimensional
model image from the image