5 Results

Based on the observations presented in this work, which we reiterate does not comprise a complete sample, we derive the following preliminary results.

5.1 Radii r_H(20.5)

The isophotal radii in this work are derived at the 20.5 mag arcsec^-2 H-band isophote, which represents a rather bright level, even for H-band measurements. For example in Papers I and II we were able to measure similar quantities one mag arcsec^-2 fainter, i.e. up to 21.5 mag arcsec^-2. This is not due to lower signal-to-noise of the present data compared with the past data, but rather to the different method adopted for deriving the light profiles. In the previous papers the radial surface brightness profiles were derived by azimuthally integrating the counts in concentric elliptical coronae of fixed center and ellipticity up to indefinite radii. Here, instead, the center and ellipticity are kept as free parameters and the fitting is halted when the mean surface brightness within a given elliptical corona equals the corresponding rms. fluctuation. The drawback is that the new fitting routine halts at higher surface brightnesses than before. The lowest surface brightness reached in each image is given in Fig. 4 as a function of the integration time.

Figure 4: The distribution of the lowest surface brightness reached in the outer light profiles, as a function of the exposure time

The comparison between the isophotal B band r_B(25.0)=(a/2)₂₅(B) radii (Gavazzi & Boselli 1996) and the infrared r_H(20.5) isophotal radii determined in this work is shown in Fig. 5. Apart from the curvature apparent at small radii, the data are consistent with r_H(20.5) = 0.7 r_B(25.0) between these two arbitrary isophotal levels.

This indicates that at the adopted isophotal level, the H observations cover a substantial fraction of the light, and are not restricted to the bulge. The relationship between optical and H-band isophotal radii is consistent with that expected from our limiting H surface brightness of 20.5 mag arcsec^-2 and with the B-H color of the outer portion of normal galaxies.

The curvature seen in Fig. 5 is only marginally an artefact of the seeing: the effect is reduced slightly by removing the objects observed in the worse seeing conditions. The deviation from linearity is not due to an overestimate of the H band radii, rather it reflects an underestimate of the B band diameter of small ($\leq~10$ arcsec) early-type galaxies which are intrinsically red (see Paper V for a more comprehensive discussion on this issue).

Figure 5: The relation between the apparent major isophotal radius rH(20.5) as determined in the infrared (this work) and in the optical rB(25.0). E+S0 are plotted as open circles, S+Irr as filled circles. The solid line represents the relation rH(20.5) = 0.7 rB(25.0)

5.2 Magnitudes (H_T, H_B25)

$H_{{\rm B25}}$ magnitudes listed in Col. 18 of Table 1 are obtained by extrapolating the circular aperture measurements to the optical r_B(25.0) radius (as in Gavazzi & Boselli 1996). $H_{\rm T}$ mag instead are obtained by extrapolating to infinity the magnitude integrated along elliptical isophotes using combinations of exponential and de Vaucouleurs laws. As expected, $H_{\rm T}$ are brighter than $H_{{\rm B25}}$ by $0.10\pm0.2$ mag on average.

5.3 Concentration index (C₃₁)

The concentration index C₃₁ is a measure of the shape of light profiles in galaxies, independent of a bulge-disk decomposition. Values larger than $C_{31}\,>\,2.8$ indicate the presence of substantial bulges.

We confirm the presence in our sample of a general correlation between C₃₁ and the H band ($H_{\rm T}$ or $H_{{\rm B25}}$) luminosity (computed from the redshift distance). We find that C₃₁ generally increases toward higher absolute magnitudes (Fig. 6). High C₃₁ are found only among high luminosity systems, but the converse is not true: there are several high luminosity systems (namely late-type giant spirals) with no or little bulge ( $C_{31}\sim 3$).

Figure 6: The dependence of the near-infrared concentration index C31 on H band luminosity