WHEN immature mammalian red cells, or reticulocytes, are stained with brilliant Cresyl Blue they may be differentiated from adult red cells and enumerated in a sample as a percentage. If these cells were scattered according to chance, reticulocyte counts would be expected to follow the binomial distribution. In a recent paper, however, Jacobsen, Plum and Rasch1 have shown that, in their laboratory, reticulocytes appear to be more regularly arranged than this when present in a proportion greater than 10 per cent. They are therefore able to make counts on one sample with remarkably close agreement. Much of the important work on reticulocyte maturation carried out in their laboratory depends on the great accuracy with which this technique is used. It was therefore thought that an independent attempt to confirm this deviation from the binomial distribution would be of interest.