The low frequency of b/p bulges (1.2%) in the sample of Jarvis (1986) seems to result from the limitation to objects of very small diameters (Shaw 1987). Such a restriction, coupled with the used visual detection method on the plates of the ESO/SERC J sky survey, results in a non-detection of many b/p bulges and in the erroneously small frequency of b/p bulges. However, his division in box (B) and peanut-shaped (P) bulges is in agreement with our classification of b/p bulges. Considering only galaxies included in both samples, 3/4 of the galaxies of type P correspond to our bulge type 1 and type B to bulge type 2, respectively. All galaxies with b/p bulges listed by Jarvis (1986) are also classified here as b/p bulges. However, his statistics are based on a too small number of galaxies being representative. Additionally, his sample is biased with galaxies of known b/p bulge characteristics from the literature, therefore the conclusion with regard to a dependence of b/p bulges on the Hubble type is not meaningful.
Shaw (1987) uses a similar definition for his selection criterion of b/p bulges (depression along the minor axis or a general flattening of the isophotes) as we do, but without classification. The low frequency of 20%, he states, can be explained by the fact that galaxies with b/p bulges of type 3 are almost completely neglected. Counting only types 1 and 2 in our investigated sample results also in a frequency of 20% (Table 2). However, some prominent b/p bulges are missed by excluding peculiar disk galaxies (e.g. NGC 3628), galaxies which are far from edge-on (e.g. NGC 7582: Quillen et al. 1997), or by simply ignoring others (e.g. NGC 2424: Pohlen et al. 2000). On the other side there are two galaxies (NGC 1596 and NGC 4958) in his list of b/p bulges which have obviously an elliptical bulge. The sample also shows a maximum in the distribution of galaxies with b/p bulges concerning the morphological type of these galaxies at Sb/Sbc galaxies (Table 5). However, the frequencies in the individual bins are much smaller reflecting the total low frequency of b/p bulges. Additionally, the lack of Sd galaxies and the small frequency of Sc/Scd galaxies with b/p bulges are remarkable. That is in contrast to the rather high percentage of late type galaxies with b/p bulges which we have found in our study. The low frequency of Sc/Scd galaxies and the lack of Sd galaxies with b/p bulges are likely a result of the faint nature of bulges and therefore also of their b/p structures. Our method of detection using the DSS and new data analysis systems could explain the differences to Shaw (1987) in the bins of the late types, although both investigations are based on the same photographic material.
SA87 define three classes of galaxies marked by the degree of separation
between disk and bulge, and by the form of the bulge.
Additional to the mixture of these two criteria, it must be mentioned that
their class III is defined as "cigar shaped'' bulge. Therefore this class
has nothing in common with b/p bulges.
This reveals
their prototype for class III (NGC 1380) and the high fraction of
elliptical bulges (68%) in the class III sample (Table 4).
Therefore the title of their list "box-shaped galaxies''
is misleading. However, it should be clear by their definition of the classes
that they do not present a list and classification of b/p bulges.
This was and is
still misunderstood in the literature (e.g. Bureau & Freeman 1999).
Therefore a strong correlation between their types
and our new introduced bulge types of the
galaxies in both samples cannot be expected
and is indeed not present (Table 4).
However, their types I and I-II ("clear rectangular or peanut shape'')
should be classified also in our system as
b/p bulge, but four galaxies (
25%) out of these types
have clear elliptical bulges
(e.g. NGC 6504, Fig. 1, bottom).
They also missed, likely due to the bad resolution of the images,
some bright galaxies with
prominent b/p bulges
(e.g. NGC 3079: Shaw et al. 1993; Veilleux et al. 1999;
NGC 7582: Quillen et al. 1997). Therefore their list is far from being
complete and
the frequency of b/p galaxies derived by SA87 is not comparable to
the frequency of b/p bulges derived in our investigation.
The statistics with regard to the Hubble type are
not comparable due to the different definition of b/p bulges.
In this way it is not surprising that
the maximum in the distribution of b/p galaxies is
reached for S0 galaxies and clearly
lower values for intermediate types (Sa/Sb) in SA87 (Table 5)
are not detectable in our distribution of b/p bulges.
| bulge type | |||||
| 1 | 2 | 3 | 4 | ||
| I | 6 | 2 | 4 | 4 | |
| [-1.5ex]galaxy class | II | 4 | 9 | 8 | 12 |
| [-1.5ex]from SA87 | III | 0 | 2 | 4 | 13 |
Our derived frequency of b/p bulges is in excellent agreement with
the result of Dettmar & Barteldrees (1988)
and Dettmar (1989) who find a frequency of 
%.
Their detection method by use of CCD images seems to
prevent a misclassification. All five galaxies
in their sample of b/p bulges (published in Shaw et al. 1990) and
our sample have bulge types 1 or 2.
Dettmar's (1989, 1996) statistics concerning the Hubble type
reveal also a maximum in the distribution of galaxies with b/p bulges
at Sb/Sbc galaxies.
However, the
differences between the individual bins are much larger
than in our statistics (Table 5).
This can be explained by statistical errors due to the small sample size of
73 galaxies.
Dettmar (1989, 1996) has also not found any Sd galaxy
with b/p bulge in contrast to the frequency of 40% Sd
galaxies with b/p bulges in our study.
However, the lack of Sd galaxies is likely a result of the small number of
investigated Sd galaxies in his sample.
Additionally, the already mentioned
faint nature of bulges of late type galaxies and
the uncertainties in the classification of bulges increasing
to later types (see above) lead to the fact that
the boxiness of late type bulges could be missed.
Even in small samples, where the
possibility of comparison between the different forms of Sd bulges is not
given, the detection of b/p bulges could be influenced.
No detection of a type 1 b/p bulge in galaxies later than Sc in our
sample could have similar reasons, if bulge type 1 in late type
galaxies exists at all.
Galaxies in previously studied samples, which do not fulfill the selection
criterion of our investigated sample due to
morphological misclassifications of the Hubble type or
too small diameters or axis ratios,
are listed in Table 7. However, most of these galaxies have
and therefore their classification with the DSS
and ESO Lauberts-Valentijn Archive is uncertain.
| S0/S0a | Sa/Sab | Sb/Sbc | Sc/Scd | Sd | |
| SA87 | 33% | 13% | 15% | 3% | -1 |
| Shaw2 | 25% | 33% | 36% | 8% | 0% |
| Dettmar3 | 46% | 29% | 61% | 42% | 0% |
| this study | 41% | 45% | 48% | 46% | 40% |
Copyright The European Southern Observatory (ESO)