I then asked Coggeshall to join in because we were beginning to map connections between cells and he was a very good anatomist. So it then dawned on us that maybe we could map connections between cells. And we found something very interesting. We started to record from cells, two at a time, instead of one at a time, to look at connections between them. And we found something very odd. We found if we recorded from certain cells, that there were inhibitory synaptic potentials in one cell that were synchronous with excitatory synaptic potentials in the other cell. Wow, that's odd. That suggested that a single cell was producing different synaptic actions on different follower cells. That had never been described before. There had been what Eccles has called Dale's principle. A neuron released only one transmitter which produces only one synaptic action at all of its terminals. This seemed to be a violation of Dale's principle. So we hunted around and we hunted around and boom we found a cell, L10, that produced excitation on some cells and inhibition on the others. We could clearly show, this was mediated by the same transmitter: acetylcholine excited one cell, inhibited the other. Acetylcholine receptor blockers blocked it.

Moreover, not only was this a fascinating finding, and Coggeshall helped us actually trace the connections, but it showed for the first time you could map connections between single cells in a neural structure. This had never been done before, but this really opened up the studying of neural circuitry, and that was really very encouraging. And then with Howard Wachtel and with Daniel Gardner, we went on to show a further variation of it, that some cells receive from the same synaptic neuron, L10, both excitation and inhibition. If you stimulate it slowly, you had excitatory synaptic potentials, if you now speeded it up, it converted to inhibition - really quite remarkable - mediated by acetylcholine. So just by combining receptors in different ways with different kinetics, you can get all kinds of synaptic properties.

So that convinced me we could begin to work out the neural circuitry of the behavior.