4 The SEDs

Figure 2 shows the SEDs of the sample galaxies obtained using the data given in Table 3 (only for those galaxies with at least 2 photometric data points). UV, optical and near-IR data are corrected for galactic and internal extinction as described in the previous section. FIR data at 60 and 100 $\mu$m are average values between IRAS and ISOPHOT data when both are available. When one of the two data is an upper limit, we take the detection. To be as consistent as possible with IRAS, ISOPHOT data have been corrected for the average ISOPHOT/IRAS ratio found by Tuffs et al. (2002) for Virgo galaxies detected with both instruments, $\rm ISOPHOT/IRAS=0.95$ and 0.82 at 60 and 100 $\mu$m respectively.

The morphological type given in Table 2 and the logarithm of the H band luminosity, defined as $\log L_H = 11.36 - 0.4H_{\rm T} +2\log D$ (in solar units), where $H_{\rm T}$ is the total H band magnitude and D is the distance to the source (in Mpc), are labeled in Fig. 2. For few objects we derive the H luminosity from K band measurements assuming an average H-K colour of 0.25 mag (independent of type; see Gavazzi et al. 2000). A minority of the objects in our sample have an H band magnitude obtained from aperture photometry, thus with no asymptotic extrapolation. For these we use the H magnitude determined as in Gavazzi & Boselli (1996) at the optical radius which is on average 0.1 mag fainter than $H_{\rm T}$ (Gavazzi et al. 2000).

The continuum line in the optical domain gives the integrated spectrum obtained by Gavazzi et al. (2002a). The two dashed lines at $\lambda< 10$ $\mu$m are the Bruzual & Charlot stellar population synthesis models (GISSEL 2001). The upper curves represent the models which best fit the extinction corrected data, as determined by Gavazzi et al. (2002a). The lower curves represent the same models attenuated by dust extinction using the inverse relations of Sect. 3.1. For galaxies with insufficient photometric points for fitting a model, we adopt the Bruzual & Charlot model that best-fits a template SED of similar morphological type (Fig. 9 in Gavazzi et al. 2002a). To be consistent with Gavazzi et al. (2002a), all models are normalized to the V band photometric data when available, or to the K band. Given the poor quality of the fit, models are not shown for the galaxies VCC 1217 and VCC 1313.

We have preferred not to give fits in the Mid-IR range for two reasons: 1) because the very small grains and the carriers of the Aromatic Infrared Bands responsable for the mid-IR dust emission are not in thermal equilibrium with the radiation, but are stochastically heated (mostly) by UV photons (Boselli et al. 2003). Thus modified black-body functions cannot be used to fit the mid-IR data. 2) mid-IR spectra obtained with the CVF camera onboard ISO in various galactic and extragalactic environments has shown a variety of strong emission lines with fluxes comparable with the continuum. It is thus difficult to estimate a typic mid-IR spectrum of galaxies.

The dashed line in the FIR domain (20-2000 $\mu$m) reprsents a two dust components model. Two modified blackbodies $F(\nu)\sim\nu^{\beta}B(\nu)(T_D)$, with $\beta=2$, one with a fixed warm temperature of $T_{\rm w}=47$ K (tracing the star forming regions), the other with a (variable) cold temperature $T_{\rm c}$ (tracing the cirrus emission), were determined consistently with Popescu et al. (2002). The two components are calibrated to match the 60 and 170 $\mu$m data respectively. For galaxies not observed by PHOT but detected by IRAS at 60 and 100 $\mu$m, we adopted a modified blackbody with $T_{\rm w}=47$ K for the warm component and we assume $T_{\rm c}=18$ K (the average value of Popescu et al. 2002), for the cold component. They are calibrated to match the 60 and 100 $\mu$m fluxes respectively.

The far-IR to mm domain, from 170 $\mu$m to $\sim$1 cm, is totally unexplored. Submillimetric observation should provide constraints on the cold dust temperature and on the total dust mass of the sample galaxies. From $\sim$1 mm to 1 cm, data are needed to estimate the relative contribution of the thermal and synchrotron radio emission.

The dashed line in the centimetric domain, given for all galaxies with more than two detections, represents the power-law regression to the radio continuum data. The best-fit parameters are given in Table 8.