2 Sample selection

2.1 Former samples

The first search for b/p bulges was conducted by Jarvis (1986). By visual inspection of all fields of the ESO/SERC J sky survey he found therein 30 b/p bulges and listed additionally 11 galaxies with b/p bulges from the literature. Together with a control sample he estimated that about 1.2% of all disk galaxies possess b/p bulges.

Shaw (1987) inspected visually the ESO/SERC J and Palomar Observatory/National Geographical Survey (POSS) B- or R-band scanned survey plates. He derived a frequency of b/p bulges of 20($\pm4$)% as a lower limit detecting 23 b/p bulges in a sample of 117 disk galaxies. The sample was mainly selected from the RC2 (Second Reference Catalogue of Bright Galaxies, de Vaucouleurs et al. 1976) and included disk galaxies with a diameter log $D_{25} \ge 1.55$ ( $D_{25}\!\sim\!3.5\hbox{$^\prime$ }$) at the 25 (B) mag arcsec^-2 isophotal level and a "sufficiently'' edge-on aspect.

Until today the largest investigated sample of 555 galaxies was presented by SA87. Their list was extracted from several catalogues and contained all disk galaxies with restricted total B magnitude ( $B_{\rm T} < 13.2$ mag) and axis ratio (b/a < 0.5). The galaxies were inspected on film copies of the ESO Quick Blue survey or the POSS prints by means of a microscope. A "substantial number'' of candidates were scanned using the PDS microdensitometer at ESO and checked with an image processing system (IHAP). In this way SA87 found 74 b/p galaxies as a lower limit. This is 13% of their total sample.

A much larger frequency of b/p bulges was proposed after the investigation of 73 galaxies obtained with CCD surface photometry (Dettmar & Barteldrees 1988; Dettmar 1989). That sample gave a frequency of $45(\pm8$)% and Dettmar (1996) concluded from this sample a lower limit of the total frequency of b/p bulges of 35%. However, only 15 objects with b/p bulge from this sample were listed by name in a paper of Shaw et al. (1990) together with other known b/p bulges from the literature.

2.2 New sample selection

The former statistics in mind a new one should contain a large number of galaxies, be complete, and based on a detection method for the b/p bulges with reasonable accuracy. In the new statistics presented here the completeness with regard to the size of a galaxy is reached by the selection of all disk galaxies ( -3.5 < T < 9.5) out of an electronic version of the RC3 with diameters larger than 2 $\hbox {$^\prime $ }$ ( $\log D_{25} \ge 1.3$). For disk galaxies without known T parameter it is determined separately (see notes to Table 6). Extracting the sample from one catalogue is preferable to the method of SA87 since the parameters differ among the catalogues and an exact selection criterion is impossible. Similar to Shaw (1987) the selected sample is diameter limited, but compared to his sample the limit is nearly 2 times smaller. Additionally, the use of the newer catalogue (RC3 instead of RC2) leads to a much larger number of investigated galaxies. Restricting the sample in diameter and not in $B_{\rm T}$ magnitude is preferred because the RC3 attempts to be complete for galaxies with an apparent diameter larger than 1 $\hbox {$^\prime $ }$ at the D₂₅ isophotal level and in addition, the $B_{\rm T}$ magnitude is not even listed for all galaxies larger than 2 $\hbox {$^\prime $ }$. Therefore a magnitude limit would not result in a complete sample. Furthermore, a magnitude limit would prefer the selection of early type galaxies.

Since b/p bulges can only be observed at inclinations down to $i\sim$75$^{\circ}$ (Shaw et al. 1990), in a first step all face-on galaxies are excluded using the axis ratio according to SA87. S0 galaxies ( $-3.5 < T \leq -0.5$) with $\log R_{25} \ge 0.30$ and other disk galaxies with $\log R_{25} \ge 0.35$ are included ( $R_{25}\!=\!\frac{a}{b}$). These different limits are used since the transformation of the axis ratio into the inclination angle depends on the morphological type (Bottinelli et al. 1983; Guthrie 1992). With different formulae (the simplest is cos $i\!=\!\frac{b}{a}$; see references in Guthrie 1992) this restriction results for disk galaxies in a limit for the inclination angle i between 60$^{\circ}$and 70$^{\circ}$ degree. This limit ensures most likely a detection of all observable b/p bulges. The final sample meeting these selection criteria contains 1343 galaxies.

The preferable method to classify a b/p bulge is the use of CCD images, but for the whole sample the amount of observing time would be unreasonably large. Therefore all galaxies are inspected using the Digitized Sky Survey (DSS). This survey is complete over the whole sky and an investigation of the images of the galaxies with data analysis software is possible. The DSS is based on photographic surveys of the northern POSS E plates (R-band, $m_{\rm lim}\!=$ 20.0 mag), the southern SERC J plates (equivalent to B-band, $m_{\rm lim}\!=$ 23.0 mag), and the southern Galactic plane SERC V plates (V-band, $m_{\rm lim}\!=$ 14.0 mag) (McLean 1999) and has a scale of $1.7\hbox{$^{\prime\prime}$ }$ pixel^-1. Galaxies which are saturated in their central regions are checked (if possible) with the ESO Lauberts-Valentijn Archive (Lauberts & Valentijn 1989) kindly made available by ESO. While the images within the ESO Lauberts-Valentijn Archive are not saturated and have a better scale ( $1.35\hbox{$^{\prime\prime}$ }$ pixel^-1), they have a lower signal-to-noise ratio compared to the images of the DSS.