2 Sample definition, observations and data reduction

The present paper contains the observation of 170 galaxies, primarily selected among late-type objects belonging to the Virgo cluster. Out of the 99 observed Virgo (12$^{\rm h}$ $\le$ RA $\le$ 13$^{\rm h}$, 0$^{\rm o}$ $\le$ dec $\le$ 18$^{\rm o}$) galaxies, 84 belong to the Virgo Cluster Catalogue (VCC) of Binggeli et al. (1985) and 15, in the ouskirts of the cluster, were selected from the CGCG (Zwicky et al. 1961-68). These galaxies have velocities V<3000 km s^-1, and can thus be considered bona-fide cluster members. Observations of 71 filler objects are also given, so subdivided:
20 are CGCG galaxies in the A262 cluster (1$^{\rm h}$43$^{\rm m}$ $\le$ RA $\le$ 2$^{\rm h}$1$^{\rm m}$, 34$^{\rm o}$31$^{\prime}$ $\le$ dec $\le$ 38$^{\rm o}$33$^{\prime}$), 23 are CGCG objects in the Cancer cluster (8$^{\rm h}$11$^{\rm m}$ $\le$ RA $\le$ 8$^{\rm h}$25$^{\rm m}$, 20$^{\rm o}$30$^{\prime}$ $\le$ dec $\le$ 23$^{\rm o}$) and 28 are CGCG galaxies in the region 11$^{\rm h}$30$^{\rm m}$ $\le$ RA $\le$ 13$^{\rm h}$30$^{\rm m}$, 18$^{\rm o}$ $\le$ dec $\le$ 32$^{\rm o}$ containing the Coma supercluster, which includes the Coma and the Abell 1367 clusters and relatively isolated galaxies in the bridge between these two clusters.

By themselves these observations do not form a complete sample in any sense. However, combined with data published in Paper I (Gavazzi et al. 1996c), Paper II (Gavazzi et al. 1996b) (which were devoted to observations of disk galaxies), Paper III of this series and in Boselli et al. (1997: B97) (containing mainly measurements of Virgo galaxies taken with the Calar Alto 2.2 m telescope), the present survey contains a complete set of NIR observations as follows: out of the 646 galaxies, of both early and late-types in the CGCG ( $m\rm _{\rm p} \le$ 15.7) which are members to the Coma supercluster ( $\rm 18^{\rm o} \le \delta \le 32^{\rm o}$; $\rm 11.5^{\rm h} \le \alpha \le 13.5^{\rm h}$) according to Gavazzi et al. (2000a), i.e. 5000 < V < 8000 kms^-1, 625 (97%) have a NIR image available. Moreover the survey contains 221 out of 248 (89% complete) VCC galaxies brighter than $m_{\rm p}=14.0$. Thus the giant members of the Virgo cluster (excluding VCC galaxies which are found in the background of the Virgo cluster) are sampled in a quasi-complete manner. A less complete coverage is at $m_{\rm p}~\leq~16.0$: 277/587 objects were observed (47% complete). However, we have observed all but one the 88 late-type VCC galaxies selected as part of the central program of the Infrared Space Observatory (ISO) (see B97) brighter than $m_{\rm p}=16.0$. These are objects lying either within 2 degrees of projected radial distance from M87 or in the corona between 4 and 6 degrees. Thus the H band survey contains a complete ( $m_{\rm p}~\leq~16.0$) sample of late-type dwarf members of the Virgo cluster, restricted however to a region smaller than the VCC.

2.1 Observations

The observations were carried out in three photometric nights of February 26, 27 and 28, 1997 with the Calar Alto 2.2-m telescope. The Cassegrain focus of the telescope was equipped with the MAGIC $256 \times 256$ pixel NICMOS3 infrared array (Herbst et al. 1993). In order to observe galaxies with large apparent sizes, the optical configuration of the detector was chosen to give the largest possible field of view, i.e. $6.8 \times 6.8$ arcmin², with a pixel size of 1.61 arcsec. The observational technique and the data reduction procedures, here just briefly summarized, are similar to the one described in B97 and in Paper III.

The seeing ranged between 2 and 3 arcsec with an average of 2.4 arcsec, as shown in Fig. 1. These seeing conditions are mostly due to the large pixels in the selected optical configuration, and as such represent a necessary disadvantage, because they also provide the large field-of-view fundamental for our observations.

Figure 1: The seeing distribution

At H the sky brightness (typically 13.8 mag) varied over the time scale of an observation by typically 3% in photometric conditions, by up to 8% in the worst conditions encountered. Reaching a brightness limit 8 mag arcsec^-2 fainter than the sky requires a careful subtraction of the sky, necessitating mosaicing techniques.

As in B97 we used three types of mosaic maps, obtained by programming the telescope pointing along different patterns.

Galaxies with optical diameter larger than half of the size of the field of view of the array were observed using a mosaic in which 50% of the time is devoted to the target of interest and 50% to the surrounding sky ("A'' mosaic, Fig. 2a in B97). This pattern was obtained alternating 8 fields centred on the target with 8 observations of the sky chosen along a circular path around the galaxy (off-set by a field of view from the centre). The 8 on-target fields were dithered by 10 arcsec in order to help the elimination of bad pixels.

Galaxies with optical diameter smaller than half of the size of the field of view of the array were observed with a mosaic consisting of 9 pointings along a circular path and displaced from one-another by 2 arcmin such that the target galaxy is always in the field ("B'' mosaic; Fig. 2b in B97). To avoid saturation each pointing was split into 32 elementary integrations of 1 s which were added by the on-line MAGIC software. There were 7 galaxies with angular sizes larger than the dimension of the detector; these were mapped using mosaics prepared according to the shape and orientation of the galaxy in the sky in order to cover the entire surface of the target. In order to get a higher signal-to-noise two observation cycles were secured for the low surface brightness galaxies. Some galaxies were serendipitously observed in the sky frames of other targets. For these objects the number of available frames is generally $\leq$ 8 (see Table 1), thus their signal to noise is lower than the average value obtained for pointed galaxies.

  

Table 1: The program galaxies. This is a one page sample. The entire table containing 170 entries is only available in electronic format

The observations were calibrated and the fluxes transformed into the Hphotometric system using standard stars (Elias et al. 1982), observed hourly throughout the night. Calibration stars were observed with a third mosaic ("C'', Fig. 2c in B97). This is composed of 5 pointings, starting with the target star near the centre of the array followed by pointings in each of the 4 quadrants of the array. The observations of the standard stars were obtained with a defocused telescope to avoid saturation.

The typical uncertainty on the photometric calibration is $\le$ 0.05 mag.

2.2 Image analysis

The reduction of two-dimensional IR frames follows procedures identical to those reported in B97 and in Paper III. These procedures are based on the IRAF data reduction package developed by NOAO and on the SAOIMAGE and PROS packages developed at the Center for Astrophysics and on STSDAS.

To remove the detector response, two sets of flat-field exposures were obtained on the telescope dome with (lamp-on) and without (lamp-off) illumination with a quartz lamp. The response of the detector is then contained in the normalized frame FF = [(lamp-on) - (lamp-off)] / $\langle$ (lamp-on) - (lamp-off) $\rangle$ (per pixel).

Specific reduction strategies were used for the various mosaics, according to the stability of the sky during the observations. When the sky was stable to within a few percent during the observation of a galaxy (the large majority of the observations), the 8 SKY exposures ( ${\rm SKY}_i$) were combined using a median filter to obtain $\rm\langle SKY \rangle$ for type "A'' mosaics. For type "B'' mosaics $\rm\langle SKY \rangle$ was obtained by combining the 9 frames containing target+SKY with a median filter.

The mean counts $\langle c_{\rm T} \rangle_i$ and $\langle c_{\rm sky} \rangle$ were respectively determined for the $i^{\rm th}$ target observation and the median sky. Individual "normalized'' ${\rm SKY}_i$ frames were then produced such that ${\rm SKY}_i$ = $\langle {\rm SKY} \rangle \times \langle c_{\rm T} \rangle_i$ / $\langle c_{\rm sky} \rangle$. This removed the time variations of the sky level, but, due to the source emission, introduced an (additive) offset; this was subsequently removed (see below). Occasionally, when the average response of the detector to the sky changed by more than 3% during an observation, significant temporal variations in the spatial response of the detector to the sky became discernable. Under these circumstances, only the three sky frames closest in time to each target frame were used to determine the sky. After sky removal, each target frame (${\rm T}_i$) was processed to obtain a flat-field, sky subtracted, corrected frame: ${\rm T}_{i,{\rm corr}}$ = [ ${\rm T}_i - {\rm SKY}_i$]/FF.

Sky-subtracted and flat-fielded frames were then registered using field stars and combined together with a median filter. This provided a satisfactory removal of the bad pixels in the final combined image. Tests on the data showed that the photometry obtained from this use of a median filter was identical to that obtained with averaging techniques.

Star-subtracted frames were produced by a manual "editing'' of the contribution from pointlike sources which are clearly not associated with the target galaxies.

The residual sky background and its rms noise ($\rm\sigma$) (in the individual pixels) were determined in each star-subtracted frame in concentric object-free annuli around the objects of interest.

We checked the quality of the final images on large and small scales. On small scales the measured noise was always consistent with the expected statistical fluctuations in the photon count from the sky background accumulated over the total integration time. The typical pixel to pixel fluctuations are $\sim$ 22 mag arcsec^-2, i.e. 0.05% of the sky (see Fig. 2).

Figure 2: The distribution of the sky rms as a function of integration time