METHODS

Two roughly parallel razor cuts were made through the head of the drone to give a slice 300-800 pm thick as shown in FIGURE 2, in which the axes of the retinulae lie parallel to the surface: The photoreceptors were not cut, but the glial syncytium was. The slice was superfused on both faces with Ringer solution containing NaCl, 200 mM; KC1, 7.5 mM; MgCl,, 10 &, CaCl,, 1.6 mM; NaHCO,, 10 mM, and bubbled with 5% C0,/95% 0,. For the first 4-7 minutes, the superfusate contained 16 mM Caz+. A light beam coming from above could be used to stimulate all the photoreceptors, meeting them perpendicular to their axes instead of passing through the cornea as in vim The intracellular ion measurements were done with doublebarreled microelectrodes,” one barrel containing a liquid ion-selective sensor and the other, reference, barrel containing electrolyte as a conductor. The electrode was advanced into the retina until a recording site was identified, mainly by its electrical response to a light flash, as a photoreceptor, a glial cell, or extracellular space.‘* The ion signal, the difference in potential between the sensor barrel and the reference barrel, depends mainly on the activity of the primary ion, but may be affected by interference from other ion species, leakage of ions from the reference barrel, and junction and tip potentials.” The principle of calibration, that we have not always followed ourselves, is to attempt to choose a solution that mimics the properties of the cytosol. We will revert to this in the section on K+.