In the Reference Catalogue of Brigth Galaxies (RC1, de Vaucouleurs & de Vaucouleurs 1964) the asymptotic magnitudes were derived from two parabolic growth curves, one for the early type galaxies and one for late types. Then in the RC2, the shape of the growth curves was determined for each revised morphological type using photographic surface photometry (de Vaucouleurs 1977). For each type, 4 to 12 galaxies were available. The set of growth curves used in the RC3 results from the evolution of the curves used in the RC2: the mean residuals from growth curves for each morphological type were used to compute differential corrections.
In the present work, surface photometry is available for 2774 galaxies. Hence we decided to reconsider the determination of the shape of the growth curves.
A major difference with RC3 is that we are determining the shape of the growth curves independently of the morphological type, hence, we call the parameter coding this shape the photometric type. We have chosen to write the photometric type in a scale comparable to the revised morphological type by setting the conditions:
for the dwarf galaxies with exponential profile.
for the elliptical galaxies with de Vaucouleurs
profile.
This scale extends toward 
, reflecting the existence of more
concentrated objects, as discussed in Caon et al. (1990).
We performed the fits with the three following sets of growth curves. Each of them spanning the full range of growth curves shapes, between exponential laws and structures even more concentrated than de Vaucouleurs law:
In the present case, we converted the Sérsic exponent into the photometric type with:
This relation returns the de Vaucouleurs law for and the
exponential law for .
The comparison between the fitted 
and the morphological type
(see Sect. 5) did not justify any sophistication of this relation.
The RC3 growth curves (RC3, Vol. 1, p. 28) were derived by averaging, for each morphological type, the growth curves of template individual galaxies with available surface photometry.
For consistency with our other sets of growth curves, we extrapolated the
RC3 set toward values of by Sérsic growth curves.
We finally used a linear combination between the de Vaucouleurs g0 and exponential g1 laws:
As in the Sérsic case, no further refinement was needed.
The residuals from the growth curve fitting are displayed in Fig. 2 (click here)
for the three cases above. For computing these mean residuals we only
selected the galaxies for which the fit was of the best quality
(
, see next section).
Figure 2: Mean residuals from growth curves. Full line: INTERP,
Dashed line: Sérsic, Dotted: RC3.
Each curve represents the mean residuals
(ordinate in mag; 
= Observed - Computed) vs. the normalized
aperture (abscissa) for the galaxies with
where the
photometric type T0, labelled on the right side, runs from -9 to 10.
The number of galaxies used to built each curve for the INTERP fit
is labelled on the left side (for the other fits, the numbers are comparable).
Each curve is shifted by a zero point 
We note that the RC3 growth curves for late types were not modified since RC2, and are probably out of date, in particular considering the significant contribution to the determination of the growth curves constituted by the inclusion of the photometry of 1355 late type galaxies from Mathewson et al. (1992). (Unfortunately, this latter photometry is restricted to the I band. Thus, color gradients had to be assumed or fitted on other data). This may partly explain the systematical wawy pattern apparent in Fig. 2 (click here).
Figure 3 (click here) shows the rms dispersion about the growth curves as a function of the photometric type for the three sets. Both Sérsic and INTERP growth curves are comparable, while RC3 is significantly worse.
Figure 3: Rms residuals from the growth curves fit vs. .
Dashed line: Sérsic,
Dotted: RC3, Continuous: INTERP
Finally, we adopted INTERP rather than Sérsic growth curves, because, in the course on this work, analysing a restricted sample of late type galaxies, we noted a systematic pattern in the residuals of the latter, absent from the former.
