BIPOLAR CELLS AND HORIZONTAL CELLS To record from a cell in the nervous system is one thing: it is another to record from a cell and know exactly what kind of cell it is. This microscopic picture shows a single bipolar cell in the retina of a goldfish, recorded in 1971 by Akimichi Kaneko, then at Harvard Medical School. The fact that it is a bipolar cell and not an amacrine or horizontal cell was proven by injecting a fluorescent dye, procyon yellow, through the microelectrode. The dye spread throughout the cell, revealing its shape. In this cross section, receptors are on top. Horizontal cells and bipolar cells occur, along with amacrine cells, in the middle layer of the retina. The bipolar cells occupy a strategic position in the retina, since all the signals originating in the receptors and arriving at the ganglion cells must pass through them. This means that they are a part of both the direct and indirect paths. In contrast, horizontal cells are a part of the indirect path only. As you can see from the diagram later in this section, horizontal cells are much less numerous than bipolar cells, which tend to dominate the middle layer. Before anyone had recorded from bipolar cells, the big unknown was whether they would prove to have center- surround receptive fields, as ganglion cells do, and come in two types, on center and off center. If the answer was yes, it would almost certainly mean that the organization discovered by Kuffler for ganglion cells was a passive reflection of bipolar-cell organization. The knowledge that the receptive fields of bipolar cells were indeed center-surround and were of two types came from intracellular recordings first made by John Dowling and Frank Werblin at Harvard Biological Laboratories and by Akimichi Kaneko at Harvard Medical School. The next question is how these receptive fields are built up. To answer it we have to begin by examining the connections of receptors, bipolar cells, and horizontal cells.