9 The mass-to-light ratio

To put the mass-to-light ratio of NGC 3992 into perspective one should compare it with values of other galaxies. Palunas & Williams (2000) give maximum disc M/L I-band ratios for a sample of 74 spiral galaxies. The luminosities of their galaxies have been corrected for absorption as if observed face-on. Unfortunately this correction does not take into account the demonstrated dependence of such a correction on total luminosity of a galaxy (Giovanelli et al. 1997; Tully et al. 1998). In practice Palunas & Williams make a correction more or less as if all their galaxies belong to the most luminous category. To circumvent this problem, 32 galaxies of the sample were selected having a $^{10}\log (2v_0) > 2.6$ or a maximum rotation approximately larger than 200 km s^-1. For that subsample the average (M/L)_I^i-0 (so corrected to face-on) is $3.1 \pm 0.7$with a total range of 2.2 to 4.3. There is another problem, however. The sample is concentrated in the great attractor region and galaxies or groups of galaxies may have large peculiar velocities rendering the adopted Hubble flow distances uncertain. For example, galaxies in the vicinity of the Centaurus cluster have a median (M/L)_I^i-0 of $1.8 \pm 0.6$ while galaxies near the Hydra cluster have a median (M/L)_I^i-0 of $3.4 \pm 1.0$.

As mentioned in the introduction, the $K^{\prime }$-band rotation curve decomposition of NGC 3992 by Verheijen (1997) indicates that this galaxy has an abnormally large $(M/L)_{K^{\prime}}$ value compared to the other HSB galaxies of the UMa cluster. Is this also the case for the present more precise rotation curve and the I-band photometry? In Paper II (Bottema 2002) a conversion is made to absorption corrected face-on (M/L)_I^i-0 maximum disc values for ten UMa cluster HSB galaxies not including NGC 3992. The average value of these ten galaxies, $\langle (M/L)_I^{i-0} \rangle = 1.65$ with a range of 0.7 to 2.2. The same absorption procedure for NGC 3992 requires that the observed (M/L)_I ratio has to be divided by a factor 1.27. Then the maximum disc fit (Fig. 13b) has a (M/L)_I^i-0 of $3.68 \pm 0.1$.

Compared to the other HSB galaxies of the UMa cluster the mass-to-light ratio is very large, also in the I-band. However, when compared with $(M/L)_I^{i-0} = 3.1 \pm 0.7$ for the luminous galaxies of the Palunas & Williams sample the value of NGC 3992 is still large, but not exceptional. When the notion holds that barred galaxies are closer to maximum disc, then one would expect (on average) that maximum disc M/L ratios of barred galaxies are determined to be smaller than those of non-barred galaxies. Consequently NGC 3992 contradicts this notion. Remains to be explained why the mass-to-light ratios of the different clusters can be so different.

A comparison with M/L ratios generated by population synthesis models is in principle not possible. Those ratios depend on the employed IMF, and in particular on the low mass end of the IMF. Changing the functionality or the low mass cutoff also changes the M/L ratio (Jablonka & Arimoto 1992; Bottema 1997; Bell & de Jong 2001). A detailed discussion of the relation of NGC 3992 to its companions and its position on the Tully-Fisher relation is deferred to Paper II.