The clustering of warm and cool IRAS galaxies

We use a series of statistical techniques to compare the clustering of samples of IRAS galaxies selected on the basis of their far-infrared emission temperature, to see whether a temperature-dependent effect, such as might be produced by interaction-induced star formation, could be responsible for the increase in clustering strength with redshift in the QDOT redshift survey that has been reported by several authors. The temperature-luminosity relation for IRAS galaxies means that warm and cool samples drawn from a flux-limited sample like QDOT will sample quite different volumes of space. To overcome this problem, and to distinguish truly temperature-dependent results from those depending directly on the volume of space sampled, we consider a pair of samples of warmer and cooler galaxies with matched redshift distributions, as well as pairs of samples selected using a simple temperature cut. We find that the redshift-space autocorrelation function of warm QDOT galaxies is significantly stronger than that of cool galaxies on large scales, but that this difference disappears when we come to consider the warmer and cooler samples with matched redshift distributions. A counts- in-cells analysis reveals no significant difference between the clustering of the warm and cool QDOT samples, while the use of a new, symmetric estimator reveals that the cross-correlations of warm and cool IRAS galaxies with Abell clusters do not differ significantly. A higher signal-to-noise ratio test is provided by computing the projected cross- correlations of the matched samples with the parent two-dimensional catalogue from which QDOT is drawn, and this does yield a marginal detection of greater large-scale power for warmer galaxies. A direct comparison of the distributions of the warmer and cooler samples, using a new technique which tests the null hypothesis that they are drawn from the same population, reveals that the two classes of galaxy do cluster differently on small scales in redshift space, while their ξ (σ,π) plots suggest that the apparent concentration of more warm IRAS galaxies into richer environments reflects the fact they sample richer volumes of space within the QDOT survey, rather than illustrating a correlation between temperature and richness. We conclude that there may be a temperature-dependent component to the observed increase in the clustering strength of QDOT galaxies with redshift, but that it is less important than a sampling effect, which reflects the local cosmography, rather than the physical properties of the galaxies and their environment. We discuss the implications of this work for the use of IRAS galaxies as probes of large-scale structure and for models accounting for their far-infrared emission by interaction- induced star formation.