Three of our sample galaxies are in common with the study by Franx et al. (1989; hereafter FIH). Our RC of NGC 1379 is in close agreement, while our VDP has a shallower gradient compared to that by FIH, but agrees with the CVD measure of $\sim$ 135 km s^-1. It is interesting to note that the minor-axis VDP of FIH is instead fully consistent both with ours and with the VDP of D'Onofrio et al. (1995). Both the VDP and the RC of NGC 1399 and NGC 1404 agree with those by FIH. Bicknell et al. (1989) and Stiavelli et al. (1993) presented kinematical profiles of NGC 1399 at PA = 84 $^{\circ}$ . Our VDP is consistent with their data, while the comparison with the RC by Bicknell et al. (1989) is made difficult by the larger uncertainties in their results (their RC is actually consistent with any RC not exceeding 30 km s^-1). Notice, however, that there is a difference of 28 $^{\circ}$ in PA. We are also consistent with the kinematical profiles by Van der Marel & Franx (1993) for NGC 1374, the only galaxy from their sample in common with ours.

The study with the largest number (8) of galaxies in common with us, is that by D'Onofrio et al. (1995; hereafter D95, see also Longo et al. 1994). Their kinematical profiles are folded around the photometric centers. In general, the agreement between the two sets of data is good, with only some of the discrepancies in the profiles worth mentioning (the CVDs will be addressed below). The first discrepancy is the up-turn in the VDP of NGC 1339, apparent in the data of D95 at 4-6 arcsec. This difference cannot be attributed to the folding process used by D95 and it is at odds with our data which give no indication of its presence. Another discrepancy is the hump in the RC of NGC 1427 over the inner 6 arcsec. D95 show it to reach a maximum rotational speed of $\sim$ 45 km s^-1, whereas our data suggests it does not reach any more than $\sim$ 25-30 km s^-1. The data for NGC 1404 are in agreement in the inner 10'', even if the central dip in the VDP is less pronounced in the D95 data. The comparison is difficult beyond that radius, due to the heavy smoothing of the D95 data, that eventually lead to a wrong negative gradient of the RC. Finally, the RC of NGC 1339 in D95 is steeper than ours in the inner 5'', while the opposite happens for NGC 1374, where D95 fail do detect the maximum at 3'' from the center.

The comparison of the CVDs needs some prior comments. Generally, our CVD measurements agree with the values given in the above-mentioned works, and with those listed by McElroy (1995), with the exception of NGC 1399, for which McElroy gives a value of 308 km s^-1. D95 give 420 $\pm$ 27 km s^-1, that is still under the 430 km s^-1 derived by Stiavelli et al. (1993) using core resolution techniques applied to observations performed under 0.7 $\hbox{$^{\prime\prime}$}$ seeing. Our CVDs are systematically lower than those by D95, that were obtained under better seeing conditions (private communication). On the other hand, since the smoothing effect of the seeing is stronger for the steeper VDPs, and since the gradient of the VDP is correlated to the CVD (Busarello et al. 1997), we expect the lower CVDs to be less affected by the smoothing. The two sets of data are related by: $\log(\sigma_0^{\tiny D95})=1.10(\pm.04)\times \log(\sigma_0^{\tiny \rm this\ work})-0.17(\pm.1)$ (see Fig. 1). We can then use this relation to estimate the "corrections'' to the CVDs for the other objects of our sample. These "corrected'' CVDs are also listed in Table 2.

$\begin{figure} \resizebox {8.8cm}{!}{\includegraphics{ds7478-fig1.eps}}\end{figure}$

Figure 1: Comparison of the CVDs in this work to those by D'Onofrio et al. (1995). The dashed line is the least square fits between the two samples ( $\log(\sigma_0^{\tiny \rm D95})=1.10(\pm.04)\times \log(\sigma_0^{\tiny \rm this\ work})-0.17(\pm.1)$ ), while the continuous line represents the equality between the two quantities. The difference between the two sets is accounted for by the difference in the seeing conditions

4 Comparison with previous data