3 Observations and data reduction

  The observations of the first sample of 19 sources have been discussed in Paper I and will not be described again here. The new 26 stars have been observed during several runs at TIRGO (11 fields in J, H and K) and at the Nordic Optical Telescope (NOT, 15 fields in K only) using the Arcetri near infrared camera (ARNICA). ARNICA is equipped with a NICMOS3 256$\times$256 HgCdTe detector and a complete description of the instrument and its performances at TIRGO can be found in Lisi et al. (1996) and Hunt et al. (1996). The TIRGO observations were obtained between 1993 and 1996. The observing setup was the same as in Paper I, with a field of view of $\sim 7^\prime\times 7^\prime$ for all objects but Z CMa, for which we covered only an area of $\sim 4^\prime\times 4^\prime$.The ARNICA pixel scale at the TIRGO is $\sim 0.96^{\prime\prime}$, which well matches the typical seeing conditions ($2^{\prime\prime}-3^{\prime\prime}$).

The NOT observations were carried out during a five night run from August 31$^{\rm st}$ to September 4$^{\rm th}$ 1996. At NOT the ARNICA plate scale was $0.52^{\prime\prime}/$pix in order to match the better seeing (in fact, during the whole run the images have been pixel-limited due to the excellent, sub-arcsecond, seeing conditions). The field of view of each image is thus $\sim2^\prime$. In order to cover a larger field around each star, we used a mosaicing technique that yields a constant signal to noise ratio on a field of $\sim8.5^\prime$ diameter centered on the target star; each mosaic consists of at least 42 partially overlapping frames. Flat fielding has been performed using differential flat frames constructed by subtracting one from the other two sky frames at different illumination, obtained by median averaging various sets of exposures at sunset and sunrise. The differential flat fielding was necessary due to the high, spatially non-uniform emissivity of the telescope (see also Hunt et al. 1996). After flat fielding, sky subtraction was performed on each frame using sky frame obtained by median averaging a set of exposures in the mosaic, as described in Hunt et al. (1994). After reduction the images were registered and combined to form the final large mosaics. All the data reduction has been performed using the IRAF and ARNICA (Hunt et al. 1994) software packages.

Photometric calibration was performed observing a set of near infrared photometric standard stars from the ARNICA (Hunt et al. 1998) list. The final calibration accuracy is $\sim3\%$ for all fields at the NOT and better than 8% at the TIRGO (except for Z CMa). The calibration accuracy for each field is reported in Table 1.

As discussed in Paper I, the automatic star finding algorithms have not proven to be completely reliable in finding all (and only) the point sources in every field, primarily because of the bright diffuse emission associated with some of the Herbig AeBe stars. Source lists in all fields have thus been individually edited and corrected by inspecting the images at different contrast levels.

Aperture photometry on the detected point sources was performed using the IRAF DAOPHOT package and a 4 pixel aperture for both TIRGO and NOT observations, corresponding to $\sim4^{\prime\prime}$ and $\sim2^{\prime\prime}$ respectively. The $3\sigma$ limiting magnitudes are reported in Table 1. By checking the cumulative source count plots, we estimate our data to be complete down to one magnitude brighter than the limiting magnitude of each field. Note that, as in Paper I, the TIRGO limiting magnitudes refer to the "edges'' of the mosaics. Since the fields have not been imaged with constant signal to noise, the central regions of the mosaics are 0.5-1 magnitude deeper.

The completeness absolute magnitude in K ($M_{K}\rm ^c$) has been computed from the observed completeness magnitude assuming the distance reported in Table 1 (Col. 11). In computing $M_{K}\rm ^c$, we have neglected the effect of extinction. We note however that in the K band the correction due to interstellar extinction is expected to be very low for all the sources, with the possible exception of the most distant ones.