Regarding the use of "half-peak" C.V. values... for the most part, I
think that this is a wonderful way to make data look just a bit better.
If the population that you are measuring is truely Gaussian (and the
instrument and DNA stain are working appropriately), then half peak, and
the actual peak CV (population Std. Deviation/Mean) should be similar. If
your data has a shoulder, or the peak is assymetrical, then the half-peak
will provide a "lower" C.V. I am convinced that this concept was
developed for flow cytometry by somebody's marketing department, not by
scientists working in the trenches.
Vince Shankey
Loyola Univ. Med. Cntr.