1 Introduction

Several theories of bulge formation are currently discussed: primordial scenarios (monolithic collapse: Eggen et al. 1962 (ELS); clumpy collapse: Kauffmann et al. 1993, Baugh et al. 1996; inside-out formation: van den Bosch 1998; Kepner 1999), secular evolution scenarios (merger: Wyse et al. 1997 and references therein, dynamical evolution due to gravitational instabilities, such as bars: Combes et al. 1990), and combinations of both (Combes 2000). Therefore statistics of bulges are needed to test the proposed dynamical processes of bulge formation. Statistics of b/p bulges are of particular interest to demonstrate the importance of these structures for disk galaxies and to point out the relevant evolution scenarios of bulges in general. From the frequency of b/p bulges the likeliness of their formation process can be estimated.

External cylindrically symmetric torques (May et al. 1985) or mergers of two disk galaxies (Binney & Petrou 1985; Rowley 1988) as origins of b/p bulges require very special conditions (Bureau 1998). Therefore such evolutionary scenarios can explain only a very low frequency of b/p bulges. Accretion of satellite galaxies is a formation process of b/p bulges (Binney & Petrou 1985; Whitmore & Bell 1988), which could produce a higher frequency of b/p bulges. However, an oblique impact angle of the satellite is needed for the formation of a b/p bulge and a massive accretion event would disrupt the stellar disk (Barnes 1992; Hernquist 1993). The origin of b/p bulges by bars was first noticed by Combes & Sanders (1981). Combes et al. (1990) and Raha et al. (1991) substantially revived this idea. Comparisons between the frequency of barred galaxies and galaxies with a b/p bulge can test the probability of evolutionary scenarios of b/p bulges based on bars.

The estimated frequency of galaxies with a b/p bulge has steadily increased in former statistics from 1.2% (Jarvis 1986), over 13% (de Souza & dos Anjos 1987, hereafter SA87) and 20% (Shaw 1987), up to 45% (Dettmar & Barteldrees 1988). This rise results from differences in sample selection, sample size, detection method, and criteria to identify b/p bulges. However, the later three statistical studies have shown that b/p bulges are not really as peculiar as they were supposed to be in the past. Therefore very common processes are required to explain their high frequency. No complete statistics and list of galaxies with b/p bulges, based on observations, have been published in the last ten years. However, the knowledge about galactic evolution would greatly benefit from a reliable value for the frequency of b/p bulges by the determination of the most likely scenario for the formation of b/p bulges and subsequent conclusions on the evolution of bulges in general. A complete list of b/p bulges is a base for a detailed analysis of these structures.

In Sect. 2 we present our investigated sample of edge-on disk galaxies, and our classification of the bulge shape is explained in Sect. 3. The results of our statistics and of additional follow-up CCD observations in the optical and in the NIR are given in Sect. 4. Section 5 compares our findings to former statistics. In Sect. 6 we discuss our results and finally, in Sect. 7 we give the conclusions from our statistical analysis of the sample.