Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 319, 747-756 (1997)

Previous Section Next Section Title Page Table of Contents

7. Comparison between structural types and traditional ones

In order to assess the relation between our morphological types and the 'traditional' ones, we now present a detailed comparison between our types and published ones for galaxies in the Coma cluster. Because of the scarceness of Es and S0s classified into detailed classes (diE/boE, SA0/SB0), mainly in the comparison samples, but also in ours, we can only estimate the quality of the broad Hubble types (E,S0,S).

This comparison of Hubble type estimates is based on a magnitude complete sample of 190 galaxies in Coma brighter than [FORMULA] mag and within one degree from the cluster center. Types determined by structural morphology are presented in Andreon et al. (1996) and Andreon et al. (in preparation), traditional types are taken from the literature. In summary, our structural classification is based on (digitized) Schmidt plates for obvious spirals and on the 'quantitative analysis' of a digitized KPNO 4m prime focus plate, of CCD images taken under good or excellent seeing conditions at Pic du Midi and CFH for all the other types (as well as for non obvious spirals). Significant overlap exists among our observations, allowing us to assess the relative quality of the classifications (Sect. 6.3). For the classifications from the literature, particular attention was paid to the meaning given to the type notation by each morphologist: for example, the S0/a type is the intermediate type between S0 and Sa types for some authors and a sign of the inability to discriminate between the two types for others.

7.1. Saglia, Bender & Dressler (1993)

Saglia, Bender & Dressler (1993, hereafter SBD) determined the coarse Hubble type of galaxies in Coma following a morphological scheme similar to ours, in the sense that structural components of galaxies are measured, and not visually estimated. They take however a conservative approach, changing the traditional type of the galaxies only when their analysis shows that the type listed in Dressler (1980) is wrong.

Table 2 shows the results of the comparison of SBD's Hubble types with ours for common galaxies.


Table 2. SBD vs us

The fraction of discrepant types is 15 % (7 galaxies out of 47).

- One galaxy (GMP 1201) was classified as S0 by us (and by all the other morphologists, including Dressler) and S by SBD.

- Out of the two galaxies classified E by SBD and S0 by us, one (GMP 1878) shows a faint but extended disk, and is certainly not a boE, as our [FORMULA] resolution CFH images show, and the other one was assigned the same type by us from the analysis of two independent images. All the other galaxies with discrepant types are faint ([FORMULA]) and have low contrast spiral arms, and were thus classified as S0/a by us and S0 by SBD.

The fraction of discrepant types (15%) is very low, and fully understood but for two galaxies out of 47 (4 % of the sample).

7.2. Butcher & Oemler (1985)

Butcher & Oemler (1985, hereafter BO) classified galaxies in the Coma cluster core by inspection of a Schmidt plate (scale [FORMULA] arcsec mm-1) and/or of an unspecified "4m telescope" plate. Their classification scheme has a fine resolution for bright and large galaxies, and a coarse one for faint ([FORMULA]) galaxies and for galaxies difficult to classify (e.g. face-on S0s). At their intermediate resolution, which corresponds to our coarse Hubble types, most of the galaxies have been assigned a Hubble type by BO, whereas a small percentage could not be put in a single class by BO (some galaxies that are noted as EL, i.e. E or S0). BO put the S0/a galaxies in the S bin in the intermediate and coarse resolution schemes (see their Table 12). Since we use their intermediate resolution, we put our S0/a in the S bin for the purpose of comparison.

Table 3 shows the results of the comparison of the morphological types for common galaxies. Out of 123 galaxies in common, 23 (=19 %) have discrepant Hubble types.


Table 3. Butcher & Oemler vs us

- Out of the 8 early-type galaxies with discrepant types, four (GMP 999, 1035, 694, 552) are faint ([FORMULA] 16), two (GMP 1373 and 908) have an uncertain type in our work, one (GMP 565) has a disk but not bright enough to allow one to classify it as S0, in spite of BO's S0 classification, and finally our data for the last galaxy (GMP 1834), classified as E by BO, suggest the presence of a small bar (undetected by BO) which lead us to classify this galaxy as SAB0.

- Out of the 15 galaxies classified as S0 by BO and as S by us, 7 show a spiral pattern not resolved into spiral arms and without HII regions on our CCD images. Such faint features were probably not detected on BO's plates, or judged of null importance by them for the classification.

The galaxies with discrepant types are not distributed uniformly among the Hubble types. As a consequence, the morphological composition of the cluster differs in the two works, even though the sample is composed of the same galaxies. The spiral fraction (here defined as the ratio of the number of spirals to the total number of galaxies) in Coma's core rises from 17 %, when computed with BO's morphological types, to 26 % when computed with our types.

7.3. Dressler (1980)

Dressler (1980) classified the Coma galaxies in a slightly larger region than ours, by inspection of a plate taken at the Cassegrain focus of a 2.5m telescope (scale [FORMULA] arcsec mm-1). His sample is the largest one in the literature for Coma galaxies. It is part of the largest published survey of morphological types of galaxies in nearby clusters. The plates used by Dressler have one of the smallest scales, and therefore meet our first requirement, to be of good quality. Dressler's types have been widely used in the literature for studies of morphological segregation (e.g. Whitmore & Gilmore 1991, Whitmore et al. 1993, Sanromà & Salvador-Solé 1990), and most of our knowledge on morphological segregation of galaxies in clusters rests on data listed in this catalogue. Dressler's X/Y notation for types points out his care in not definitely classifying galaxies for which he does not have suitable data (Dressler 1980). Therefore Dressler's X/Y types are not meant to be a transition type between X and Y.

Table 4 shows the comparison of types for common galaxies. We put Dressler's E/S0 and D types in the same bin as the E. As for previous comparisons, we put Dressler's S0/a galaxies and ours in the S class, but consider them separately in the following discussion. The fraction of discrepant types is 23 % (35 galaxies out of 153).


Table 4. Dressler vs us

- 9 of Dressler's S0s were classified as E by us; most of them are roundish galaxies difficult to classify because face on.

- 12 of Dressler's E (+D+E/S0) were classified as S0 by us, of which 6 are faint galaxies, 5 were classified by Dressler as E/S0 (i.e. E or S0) and the last one (GMP 1931) is a barred galaxy.

- 11 of Dressler's S0 (+S0/a) were classified as S by us. The fact that at least 7 of them just have a spiral pattern not resolved into arms and HII regions in our images suggests that low contrast spiral arms are not detected or are considered of null importance in Dressler's morphological scheme.

- Two of Dressler's Sa galaxies (GMP 1925 and GMP 1844) and one S0/a (GMP 1154) were classified as S0 by us. Dressler's S0/a (i.e. S0 or Sa), classified S0 by us, is not a discrepant case since Dressler's classification is uncertain. Furthermore our images of this galaxy show a slight asymmetry that may explain why Dressler's classification is uncertain. The two other galaxies show uncommon characteristics of lenticulars, an isophotal twist in the outer envelope (GMP 1925) and an important ring-lens (GMP 1844). It is possible that Dressler used such characteristics to classify the galaxies as S, because they do not present spiral arms or dust on our images. In such a case, the differences in the morphological estimates are due to the different weights given to structural components in the definition of the galaxy types.

The fraction of discrepant types is 23 %, (at least) half of which are accounted for by galaxies difficult to classify, and the other half by different weights given to the presence of a spiral pattern or other morphological structures in classifying galaxies.

7.4. Rood & Baum (1967)

Rood & Baum (1967, hereafter RB) classified the galaxies in the Coma cluster by visual inspection of a prime focus plate from the 5m Palomar telescope (scale [FORMULA] arcsec mm-1).

Table 5 shows the comparison of RB's morphological types and ours for galaxies in common. We put the S0/a galaxies in the S bin as in previous comparisons, and consider them separately in the following discussion.


Table 5. Rood and Baum vs us

- Five of the 6 early-type galaxies with discrepant types are faint. The last one, classified E by us and S0 by RB, is the second ranked galaxy of the Coma cluster. RB probably classified it as S0 because it has an outer envelope, which identifies this galaxy as an S0 according to Hubble's definition of S0.

- All the other discrepant types concern RB's S0s classified as S by us. Since RB used similar plates to ours (with similar scales), they probably did not miss the structural components that we detected on these galaxies, but they presumably judged these components of null importance for the classification. We note, furthermore, that the missed spirals do not have obvious spiral arms or HII regions but just a spiral pattern. From the classical morphologist's point of view, these galaxies do not resemble the standard Sa, which explains their classification as lenticulars.

The fraction of galaxies with a discrepant type is 16 %, if we put the S0/a in the S class, or 12 % if we put the S0/a in the S0 class, as RB seem to do.

7.5. RC3

The RC3 catalogue (de Vaucouleurs et al. 1991) lists a large number of galaxies (more than 25000) distributed over the whole sky together with their morphological type. It is not fair to request that this huge catalogue be as good in selected regions as detailed studies of these regions. Furtermore, the galaxy types are not all based on the same observational material, but on a variety of plates, films, prints or copy of them, and have been assigned by different morphologists.

Because RC3 is the main catalogue of reference for types, and because morphological types are often claimed good if they agree with the ones listed in RC3 (e.g. Doi, Fukugita & Okamura 1993), we have to investigate how our types are related to the RC3 ones. A major caveat is in order: Coma galaxies are fainter than the majority of galaxies listed in RC3, and, as a consequence, we are making a comparison with the faint end of the catalogue only.

Among the 190 galaxies of our complete sample of galaxies in Coma, 129 are listed in RC3. Among those, only 71 have been assigned a definite morphological type in RC3, whereas 14 others have an uncertain morphological type. Limiting our sample to the magnitude of the faintest galaxy classified in RC3 ([FORMULA]), 79 brighter galaxies (out of 158) are not classified (or listed at all) in RC3. This catalogue is therefore incomplete to a large degree for such faint galaxies. To reach a reasonable degree of completeness (90 %), RC3 has to be limited to [FORMULA], and in this case the sample of classified galaxies only numbers 53 galaxies.

Tables 6 and 7 show the comparison between our types and the ones listed in RC3 for the whole sample ([FORMULA]) and for a brighter sample ([FORMULA]) for which RC3 is 90 % complete. First of all, 60 % of the S0s are missing in RC3 whereas only 30 % of Es and 43 % of Ss are missing. The missing galaxies are therefore not distributed uniformly among types. We stress again that these missing galaxies are brighter than the chosen magnitude limit.


Table 6. RC3 (all RC3 Coma galaxies) vs us


Table 7. RC3 (bright RC3 Coma galaxies) vs us

The fraction of galaxies with discrepant types is 30 % for the two RC3 subsamples, a factor two higher than the value found for studies dedicated to the Coma cluster. A similar disagreement is found between RC3 and Dressler (27 %), based on 66 common galaxies. The disagreement between RC3 and SBD is similar to the one between us and BO, Dressler or RB (i.e. [FORMULA] %).

RC3 agrees extraordinary well with BO and RB (less than 5 % of disagreement). This is inconceivable for two reasons. First, no pair of morphologists shows such a good agreement, and never does the RC3 type agree so well with de Vaucouleurs' estimate of the morphological type reported in Naim et al. (1995). Second, RC3 is a collection of types estimated by different morphologists. The only explanation for such a good agreement is that the morphological types of Coma galaxies in RC3 were taken from BO and RB.

In summary, RC3 is incomplete to a large degree in this region and for such faint galaxies, and, moreover, the missing galaxies are not distributed in a uniform way among types. Finally, one third of the galaxies with a morphological type in RC3 have discrepant types with respect to ours and Dressler's. Therefore, the use of this catalogue for morphological studies of nearby clusters is not recommended. Moreover, due to the high misclassification of the RC3 Coma sample of galaxies, this sample is not a good comparison sample to estimate the quality of other morphological classifications, as has unfortunately been done in the literature.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998