4 Results

4.1 Model-free photometric parameters and radial profiles

Integrating the light of the galaxy from the center to the outskirts the growth curve approches an asymptotic value. This total intensity, $I_{\rm T}$, is related to the total apparent magnitude, $m_{\rm T}$, through:

$$m_{\rm T} = -2.5\,\log(I_{\rm T}) + c.$$ (1)

c is a constant derived from the calibration. $m_{\rm T}$ translates to $B_{\rm T}$, $V_{\rm T}$, or $R_{\rm T}$ depending on the colour band.

The model-free effective radius can be read at half of the total growth curve intensity. As the integration has been done with elliptical apertures, this radius refers to an equivalent radius, $r = \sqrt{ab}$, where a and b are the major and minor axis of the galaxy, respectively. Together with the total apparent magnitude this yields the effective surface brightness by:

$$\langle\mu\rangle_{\rm eff}\,[{\rm mag}/\ifmmode\hbox{\rlap{$... ...$ }]=M+5\,\log(r_{\rm eff}[\hbox{$^{\prime\prime}$ }]) +1.995.$$ (2)

The global photometric parameters of the observed galaxies are listed in Table 2. All magnitudes, surface brightnesses, and colours listed are corrected for galactic extinction based on Schlegel et al. (1998). The columns of Table 2 represent:

Column (2): name of the galaxy;

Column (3): total apparent magnitude in the B band;

Column (4): galactic absorption in B using the extinction maps of Schlegel et al. (1998);

Columns (5) and (6): total apparent magnitude in the V and R band, respectively;

Columns (7), (8) and (9): effective radius in B, V and R, respectively, in arcseconds;

Columns (10), (11) and (12): effective surface brightnesses in B, Vand R, respectively, in mag/ $\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\hfil \penalty50\h... ...$ } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi\hbox{$^{\prime\prime}$ }$;

Column (13): total B-V colour index;

Column (14): total B-R colour index.
Doubtful values due to low-quality observations (frames hampered with reflected light or with a bright background due to moon light) are flagged with a colon. For the error assessment (see Sect. 4.4) these cases have not been taken into account.

Surface brightness profiles were obtained by differentiating the growth curves with respect to equivalent radius. The resulting profiles, with a resolution, or bin size of 2 $\hbox{$^{\prime\prime}$ }$, are shown in Fig. 2. The profiles are traced down to the level where the uncertainties owing to the fluctuations in the sky level on the profile become dominant, which is at $\sim$ $28\,{\rm mag}/\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\hfi... ...$ } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi\hbox{$^{\prime\prime}$ }$ in B, $\sim$ $27.5\,{\rm mag}/\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\h... ...$ } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi\hbox{$^{\prime\prime}$ }$ in V, and $\sim$ $27\,{\rm mag}/\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\hfi... ...$ } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi\hbox{$^{\prime\prime}$ }$ in R.

4.2 The exponential model: Fits and parameters

Radial intensity profiles can be fitted by different models. The exponential model (De Vaucouleurs 1959; Binggeli & Cameron 1993) works very well for dwarf elliptical galaxies, because of their steady decrease of brightness from the center to the outer parts. Owing to their shape and, if present, bright star forming regions, the brightness of irregular galaxies is not so uniform, but parts of almost all intensity profiles of the irregulars considered here exhibit an exponential behaviour. The exponential intensity profile can be written as

$$I(r) = I_0\,{\rm e}^{-\alpha r},$$ (3)

which in the surface brightness (magnitude) representation becomes a straight line:

$$\mu(r) = \mu_0+1.086 \alpha r.$$ (4)

The central extrapolated surface brightness $\mu _0$ and the exponential scale length $1/\alpha$ are the two free parameters of the exponential fit.

Using a standard least squares fitting procedure, we determined a best-fitting exponential within that part of the galaxy profile, usually in a medium radius range, that is looking reasonably straight (exponential). This fitting range was defined for each galaxy individually (but is the same for each colour band). However, some noisy profiles (e.g. PGC 17716, A0554+07, DDO 64) were hard to fit and the definition of the fitting region was not obvious. In these cases we used rather large radius ranges for the fits to compensate for the fluctuations. The best-fitting exponential parameters are listed in Table 3. The best-fitting exponential profiles are plotted as dashed lines along with the observed profiles in Fig. 2, where also the fitting range is marked by ticks along the upper axis.

The deviation of the observed profile from a pure exponential law is expressed by the difference $\Delta m$ between the total magnitude of an exponential intensity law given by

$$m_{\exp} = \mu^{\exp}_0 + 5\,\log(\alpha) - 1.995,$$ (5)

and the actual measured total magnitude. The results are shown in Table 3. The difference $\Delta m$ is a measure of the goodness of fit of the exponential intensity profile. As one can see from the table, and of course Fig. 2, most of our galaxies are "good'' exponentials.

The columns of Table 3 are as follows (again all values being corrected for galactic extinction based on Schlegel et al. 1998):

Column (2): name of the galaxy;

Columns (3), (4) and (5): extrapolated central surface brightness according to equation (4) in $B\,[{\rm mag}/\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\hfi... ... } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi\hbox{$^{\prime\prime}$ }]$, $V\,[{\rm mag}/\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\hfi... ... } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi\hbox{$^{\prime\prime}$ }]$, and $R\,[{\rm mag}/\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\hfi... ... } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi\hbox{$^{\prime\prime}$ }]$, respectively;

Columns (6), (7) and (8): exponential scale length in arcsecs $[\hbox{$^{\prime\prime}$ }]$, in B, V, and R, respectively;

Columns (9), (10) and (11): difference between the total magnitude as derived from the exponential model and the true total magnitude in B, Vand R, respectively.

4.3 Colours and colour profiles

As expected from a sample of late-type galaxies, the median integrated colour, with (B-V) = 0.52, is rather blue. However, the scattering is considerable (see Table 2). Four galaxies (UGC 685, UGC 2905, PGC 17716 and A0554+07) appear extremly blue ( $0.07 \leq B-V \leq 0.18$). While UGC685 might indeed be a blue dwarf compact (BCD, see Fig. 1), the blue colour of the others is probably not real: all three galaxies lie close to the zone of avoidance and have therefore high corrections for galactic extinction ( $1.35 \leq A_B \leq 3.17$, see Table 2). Any slight uncertainty in the correction with respect to colour will produce a large error in the "true'' colour index. On the other hand, there are two very red galaxies: UGC 1281 with (B-V) = 1.10, and UGC 2689 with (B-V) = 1.02. UGC 1281 is seen perfectly edge-on, and so its red colour is clearly caused by internal dust absorption, while UGC 2689 is a probable background S0 (see notes on individual galaxies) and therefore quite plausibly red.

B-R colour profiles, along with the differences of the B and R exponential fits, are shown in Fig. 3. Most colour profiles, where the noise is small enough to allow any significant gradient to be seen, follow the trend of an increasing B-R with increasing galactocentric radius, i.e. a blueing inwards. This is the stellar population gradient known and expected for dwarf irregulars: regions of active or recent star formation, and therefore of blue colour, are concentrated to the inner part of the galaxies. There are two notable, and plausible, exceptions showing the inverse trend, i.e. a reddening inwards (see Fig. 3): UGC 2689, the suspected background S0, and UGC 3303, a large spiral in the zone of avoidance.

Figure 4: Comparison of our data with data from the RC3: our published data vs. RC3 (left panel); data from our "uncleaned'' frames vs. RC3 (right panel).

4.4 Photometric accuracy

The photometric zero point determination represents the greatest source of error on the global parameters in this work. Other uncertainties, stemming from the flat-fielding and background subtraction etc., were mentioned in Sect. 3. Instead of trying to assess the overall error from all these single error sources, which seems impossible anyway, it is easier and more reliable to compare our results with data from the literature. Ten galaxies of our sample can be compared to RC3 data (Third reference catalogue of bright galaxies, deVaucouleurs et al. 1991). Figure 4 shows the result of this comparison for total magnitudes (where this time no correction for galactic extinction was applied). A first plain comparison of our magnitudes with RC3 magnitudes is disappointing: deVaucouleurs et al.'s magnitudes are systematically brighter by up to one magnitude (as evidenced in the left panel of Fig. 4). However, if we take the magnitudes derived from our uncleaned images and compare them with RC3, the agreement is quite satisfactory, with a standard deviation in magnitude of only $\Delta m = 0.143$ mag (shown in the right panel of Fig. 4). This likely means that the RC3 magnitudes, at least for these galaxies, are based on images that were not cleaned from disturbing foreground stars. In fact, they are based on aperture photometry where the contamination by stars is unavoidable and inherently large for faint and diffuse dwarf galaxies.

If we distribute the contribution to the above-mentioned $\Delta m$evenly between our photometry and deVaucouleurs et al.'s, had they properly accounted for foreground contamination, we arrive at a global (one sigma) uncertainty of $\Delta m = 0.1$ (for us and them), applying also to surface brightness. This is indeed the typical photometric accuracy achieved in our previous papers.

Previous photometry for a number of our dwarfs is provided also by some other studies based on CCD imaging, notably by Makarova et al. (1998). A comparison with our $B_{\rm T}$ and B-V values often shows large differences. However, since no details on how the photometry was done is given in that and other papers, we cannot judge the reliability of the comparison. More details are given in the notes on individual galaxies below.

As far as the typical uncertainty of the model parameters is concerned, given the amount of subjective judgement in choosing the fitting range etc., this is impossible to assess without comparison with independent data from the literature.