6 Discussion and conclusions

We have shown that a large fraction of late-type galaxies in the Virgo cluster have their current star formation rate significantly quenched with respect to isolated objects. These systems coincide with the Virgo gas deficient galaxies. Since the "gas'' deficiency parameter is dominated by the HI phase (H₂ contributes only to 15% of the HI), it is concluded that, to the first order, the star formation properties of galaxies in the Virgo cluster are determined by the pattern of HI deficiency. As earlier recognized by Kennicutt (1998), this is a somewhat surprising result, because the typical scales of HI and of star formation are very different in disk galaxies. HI reservoirs extend some 2 $\times$ the scale where the star formation takes place (Cayatte et al. 1994). We will re-examine this issue in more details in our forthcoming paper dedicated to the morphology of the star formation regions in galaxies, where a comparison between the scale-length of $\rm H_\alpha$, $\rm HI$ and $\rm H_{2}$ will be carried on specifically. Limiting ourself to the global quantities, they indicate that infall of HI gas occurs in the disks on time scales similar to the star formation time. If the gas replenishment fails, because the HI reservoir is reduced by some ablation mechanism (e.g. ram pressure), the star formation adjusts itself to significantly lower rates.

Figure 14: Newly observed galaxies with substantial $\rm H\alpha +[NII]$ structure. The NET (ON-OFF) frames are given with grey-scale, with superposed contours of the OFF frames. J2000 celestial coordinates are given.

Figure 14: Continued.

Figure 14: Continued.

Galaxies with $b_{{\rm resid}}< -0.7$ and $Def_{{\rm gas}}>0.4$ ("quenched'') are plotted in Fig. 13 with empty symbols, together with their "healthy'' counterparts (filled symbols). Beside a marginal clustering of deficient objects around M 87 (cluster A) and M 49 (cluster B) the two populations appear mixed in position. There is for example a considerable fraction of "healthy'' objects projected onto the center of cluster A. However Virgo is known to be a complex dynamical entity, composed by the main cluster (A) a secondary cluster (B), several Mpc behind A, and a number of clouds at approximately the distance of A, but with significantly discrepant velocities, suggesting infall (see Gavazzi et al. 1999a).

By considering galaxies with projected angular separation <3.7 deg from M 87 we isolate 68 bona fide members of cluster A. We divide them into 48 "quenched'' and 20 "healthy''. For a considerable fraction (22/48 and 13/20 respectively) their distance is available from Gavazzi et al. (1999a) based on the H band Tully-Fisher relation (Tully & Fisher 1977) (distances of few galaxies whose H magnitudes were not yet available to Gavazzi et al. 1999a were recomputed by us). To our surprise we find that, while the average distance modulus of the deficient objects ( $\mu_{\rm o}=$ 30.85) is in perfect agreement with the distance modulus of cluster A as a whole ( $\mu_{\rm o}=$ 30.82) (Gavazzi et al. 1999a), the distance modulus of the non-deficient galaxies projected onto A is $\mu_{\rm o}=$ 31.77 on average, thus almost one mag more distant. It is thus concluded that "healthy'' spirals projected onto the Virgo center belong in fact to a background cloud with a distance comparable with that of cluster B. This cloud has not yet entered the dense environment of cluster A.