A Statistical Comparison of Line Strength Variations in Coma and Cluster Galaxies at z0.3

Discussion

We have made a statistical comparison of spectral line strengths in three rich galaxy clusters at with the nearby Coma cluster. Using the morphological information from Hubble Space Telescope images and a principal component analysis to analyse the relationships between spectral line strengths in the nearby and distant clusters, we have found two specific results.

First, more independent parameters are required to account for the line strength variations in the distant cluster data than in the Coma data. Couch & Sharples (1987) have demonstrated that the colour-magnitude diagrams of these clusters show little scatter in their early type galaxy populations and the same has been shown for Coma by Colless & Dunn (1996). In conjunction with the small photometric scatter of the early type galaxies in both the distant clusters and Coma, the change in spectroscopic properties between the distant clusters and Coma serves to confirm the ambiguity of broadband colours for investigating the star formation histories of galaxies. Although this is not a surprise, it should be considered as yet another reminder of the fact that broadband colours to a large degree do not reflect changes in the absorption feature strengths of the underlying spectrum, while the absorption features are tied more directly to the stellar populations in the galaxies. The different number of parameters in the data sets suggests that galaxies in clusters have been homogenized such that some fraction of the information from the past has been lost. Hence, it may be possible that the present day homogeneity of early type galaxy properties, e.g. fundamental plane relations, does not require a single formation scenario, but that a variety of formation scenarios for different galaxies could still yield the observed behaviour of nearby galaxies. Some evidence of this has been seen in the models of Shioya & Bekki (1998).

Second, the distributions of line strengths versus PC weights for the Coma and high redshift clusters are different. One of the differences is manifest in the fact that the first principal component for Coma is dominated by metal line strengths of both light and heavy elements, while in the high redshift clusters it is controlled by only the light elements plus the Balmer lines. One interpretation of this difference could be that the most important factor in understanding the high redshift clusters is the prevalence of recent star formation. The correlation between Balmer lines and metal lines could then stem from the general principle of weakening metal features by adding A star light to a spectrum. Although since only the light elements are involved, the reason may come from an intial post-starburst overabundance of Type II supernovae relative to Type Ia supernovae, which have a longer characteristic time scale. This could lead to large intial changes in the light elements while leaving Fe-peak elements temporarily stable. The behaviour of the first component for Coma could be demonstrating that Coma is relatively inactive in relation to the high redshift clusters and that the range in metallicity of the galaxies is the most relevant factor in understanding the properties of galaxies in that cluster.

The larger number of independent parameters and the importance of the Balmer lines in the high redshift spectra suggest that we are seeing a period of relative chaos in the evolution of galxies in rich clusters and, hence in the evolution of the clusters themselves. Perhaps the next avenue to pursue is to determine the relationship of the global properties of the clusters to the spectral properties of the galaxies. To do that will require more high quality spectral data and morpological information in galaxy clusters over a range of redshifts.

© Copyright Astronomical Society of Australia 1997