Astron. Astrophys. 318, 729-740 (1997)

1. Introduction

Deep imaging surveys have revealed a large population of blue galaxies at faint magnitudes (e.g. Tyson 1988; Lilly et al. 1991; Metcalfe et al. 1991, 1995; Driver et al. 1994a), which becomes increasingly important at fainter limits. The exact nature of these objects remains uncertain; for instance it is not clear whether these blue galaxies themselves have more extreme properties at fainter magnitudes, or whether their relative numbers increase. Spectroscopic surveys suggest that they are found predominantly at moderate redshifts ( to 0.6) (e.g. Colless et al. 1990, 1993; Cowie et al. 1991; Lilly et al. 1991; Glazebrook et al. 1995a) and that they tend to have strong [OII] emission (Koo et al. 1995). Since they are not predicted by standard galaxy models, these blue galaxies have often been cited as the strongest evidence for galaxy evolution with redshift (Broadhurst et al. 1988; Broadhurst et al. 1992).

Any convincing explanation of their nature must account for the absence of any clear local counterpart to the faint blue galaxies. One possible model envisages that these objects have faded over time, implying luminosity evolution of either the entire galaxy population (e.g. Lilly et al. 1991) or of a subset of it (Broadhurst et al. 1988; Phillipps & Driver 1995; Driver et al. 1995a). Alternatively, the blue galaxies may have experienced density evolution. Merging (Rocca-Volmerange & Guiderdoni 1990; Broadhurst et al. 1992; Carlberg 1992) of the faint galaxies, either with each other or with more luminous objects, provides a mechanism for reducing their numbers by the present time (Koo 1990; Carlberg & Charlot 1992).

Giraud (1992) performed high resolution imaging of samples of blue galaxies, identifying three distinct morphological classes. More recently, Colless et al. (1994) have studied the light profiles of a sample of faint blue galaxies from the Colless et al. (1990, 1993) redshift survey using images taken in excellent seeing with the Canada-France-Hawaii Telescope. They found that faint galaxies exhibiting [OII] line emission often have nearby companions, suggesting interactions are important in activating star formation.

Recent HST results show a large fraction (around a half) of faint galaxies to have irregular or peculiar morphology (e.g. Casertano et al. 1995; Driver et al. 1995a, b). Of these maybe one third (i.e. 15% of the total population) appear to be interacting (e.g. Driver et al. 1995b; Glazebrook et al. 1995b). This clearly suggests a role for interactions or mergers in the evolution of the faint blue galaxy population. Furthermore, Burkey et al. (1994) found that 34% of HST galaxies at redshifts 0.4 to 0.7 had close companions compared to 7% locally, suggesting that 13% of the distant population may have disappeared by merging. However, Woods et al. (1995) using a similar technique find no excess pairs in the deep data.

Various different merging models have been advocated. The faint blue galaxies may gradually merge with one another to form more massive objects, or may be accreted into massive dark haloes (Rocca-Volmerange & Guiderdoni 1990). Cowie et al. (1991) suggested that the faint galaxies have 'parent' giant galaxies with which they have since merged, implying a physical (clustering) association between the two (see also Cowie et al. 1995). Kauffmann et al. (1994) have fitted faint number counts using detailed models of hierarchical galaxy formation (in a cold dark matter context) in which satellite galaxy haloes merge with more massive dark haloes of giant galaxies. On the other hand, Dalcanton (1993) has argued that the conservation of luminosity during mergers of the blue galaxies would lead to an excess integrated luminosity over that observed. The observed thinness of the discs of spiral galaxies may also constrain the importance of mergers in the evolution of these objects (e.g. Tóth & Ostriker 1992). The general level of clustering among faint blue galaxies appears to be low (Efstathiou et al. 1991; Pritchet & Infante 1992; see also Neuschaefer & Windhorst 1995) which may constrain direct merger models. On the other hand, Cole et al. (1994) find that dwarfs and giants at moderate redshifts occupy the same general structures and have very similar large-scale clustering properties.

A model of mergers of dwarfs with giants might therefore be tested by measuring the small-scale clustering of faint blue galaxies around a sample of candidate giant galaxies; this is the approach we adopt here (see Jones et al. 1994). A preferential clustering of the faint blue galaxies around giants would imply that they are dwarfs at similar redshifts to the giants and might favour merging models over fading in accounting for the lack of these low luminosity systems in the nearby Universe. Conversely, the absence of any excess around giants might be interpreted as evidence against the merging of dwarfs with more luminous parents (see also Bernstein et al. 1994). Using data from the Hitchhiker camera, we study the numbers of these objects around photometrically selected candidate giants and compare them with a random distribution in order to search for an excess consistent with the blue galaxies being dwarfs associated with parent giants.

In Sects. 2 and 3 we describe the observational data and the data reduction methods. In Sect. 4 the image detection process is detailed together with the techniques used for aperture photometry. The definition of samples of faint blue galaxies and candidate luminous galaxies are discussed in Sect. 5, and the association of the blue galaxies with the giants is determined. We model the distribution of random samples of faint images to account for the effects of the limited areas of data frames and to demonstrate that the light of bright images does not mask faint images to affect significantly the statistics of fainter objects. Sect. 6 considers the association between other photometrically selected samples. Sect. 7 presents a detailed investigation of the errors in the analysis. Finally, the implications of the results for the evolutionary history of the galaxy population are briefly considered.

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998
helpdesk.link@springer.de