Statistical and graphical methods for testing the hypothesis that quanta are made up of subunits.

It has been proposed that at the neuromuscular junction quanta are made up of subunits, because regularly spaced peaks appear in some histograms of miniature end-plate potential (MEPP) amplitudes. Many sets of MEPP sizes fit lognormal probability distribution functions. I compared sets of MEPP amplitudes from the literature to lognormal distributions, using cumulative plots and a robust test for significance. Some data sets fit a single lognormal distribution, but most required two lognormal subpopulations. This bimodal model is a parsimonious alternative to the subunits hypothesis. I also used graphs in which cumulative MEPP sizes were plotted on the ordinate against probability on the abscissa. Model data sets, generated on the subunit assumptions, produced concave plots. End-plate potential sizes at low quantal outputs--in which we know there are subunits--show the same concave shape. Most of the tested data from the subunit literature produced convex curves.