SIMBAD references

The Gini coefficient, a nonparametric measure of galaxy morphology, has recently taken up an important role in the automated identification of galaxy mergers. I present a critical assessment of its stability, based on a comparison of HST ACS imaging data from the GOODS and UDF surveys. Below a certain signal-to-noise level, the Gini coefficient depends strongly on the signal-to-noise ratio, and thus becomes useless for distinguishing different galaxy morphologies. Moreover, at all signal-to-noise levels the Gini coefficient shows a strong dependence on the choice of aperture within which it is measured. Consequently, quantitative selection criteria involving the Gini coefficient, such as a selection of merger candidates, cannot always be straightforwardly applied to different data sets. I discuss whether these effects could have affected previous studies that were based on the Gini coefficient, and establish signal-to-noise limits above which measured Gini values can be considered reliable.