We have studied multichannel patch-clamp recordings in earthworm axon septal membranes that contain gap junctions. Though all channels have the same conductance and selectivity, the probabilities of the conductance levels in the majority of the recordings could not be fit by assuming independent and identical channels; in these cases, we found that at least two different open probabilities were required to explain the data. The data thus suggest that, within one junctional membrane complex, there exists a heterogenous channel population of similar but not identical channel types. The analysis also revealed cases where cooperativity between individual channels was the only explanation for the amplitude histograms of the observed multichannel activity. The conclusions drawn are based on a theoretical analysis of multichannel current-amplitude histograms. We derive two tests for independent and identical channels. We analyze the effects of mode shifting. These results are based on the ratio of peaks in the histograms; they are independent of the number of channels in the patch and the model of channel gating. In some cases failure to fulfill the criteria of these tests implied an interdependence or cooperativity between channels. Lastly, we have devised statistical tests for stability of the recording in the presence of variance due to finite sample size.