2004MNRAS.349..889A

We have observed three clusters at z ∼ 0.7, of richness comparable to the low-redshift sample of Butcher & Oemler (BO), and have determined their fraction of blue galaxies. When adopting the standard error definition, two clusters have a low blue fraction for their redshifts, whereas the fraction of the third one is compatible with the expected value. A detailed analysis of previous BO-like studies that adopted different definitions of the blue fraction shows that the modified definitions are affected by contaminating signals: colour segregation in clusters affects blue fractions derived in fixed metric apertures, differential evolution of early and late type spirals potentially affects blue fractions derived with a non-standard choice of the colour cut, and the younger age of the universe at high redshift affects blue fractions computed with a colour cut taken relatively to a fixed non-evolving colour. Adopting these definitions, we find largely varying blue fractions. This thorough analysis of the drawbacks of the different possible definitions of the blue fraction should allow future studies to perform measures in the same scale. Finally, if one adopts a more refined error analysis to deal with BO and our data, a constant blue fraction with redshift cannot be excluded, showing that the BO effect is still far from being detected beyond doubt.