All reconstructions

of single neurons were based on neurobiotin injected cells. Confocal image stacks were acquired with the 25× objective. For two-dimensional neuron reconstructions, image stacks were loaded into Photoshop software and arborizations INCB024360 order were traced with the pencil tool. According to neuropil boundaries visible from background staining, the resulting image was finally projected onto a three-dimensional reconstruction of the central-complex neuropils. Three-dimensional reconstructions of neurons were achieved by using a supplemental tool for Amira 4.2 as described by Schmitt et al. (2004). The updated version of this tool was kindly provided by J.F. Evers (Cambridge, UK). For obtaining neuropil reconstructions from the dye-injected brains, unspecific background staining was used analogous to anti-synapsin staining. For recordings, animals were waxed onto a plastic holder. Legs and wings were removed and the head capsule was opened frontodorsally. Recordings were all performed on the left side of the brain. For accessing the recording site, the left antenna was removed, while the right antenna was left intact; behavioral studies in a flight simulator have shown that one antenna is sufficient for proper time-compensated sun compass orientation (P.A. Guerra and S.M.R., unpublished data). To increase stability, the oesophagus was transected and the gut

was removed check from the opened abdomen. The neural sheath was locally removed mechanically with forceps after brief enzymatic find protocol digestion and intense rinsing with ringer solution (150 mM NaCl, 3 mM KCl, 10 mM TES, 25 mM sucrose, 3 mM CaCl2; pH = 6.9; King et al., 2000). The animal was then mounted in the recording setup, with the vertical axis of the compound eye aligned horizontally. Thus the dorsal side of the eye faced the stimulation setup, while the

recording electrode could be inserted vertically from the frontal side. Intracellular recordings were performed with sharp electrodes (resistance 60–150 MΩ), drawn from borosilicate capillaries. Electrode tips were filled with 4% Neurobiotin dissolved in 1 M KCl and backed up with 1 M KCl. Intracellular signals were amplified (10×) with a SEC05-LX amplifier (NPI), digitized, and stored on a PC (details in Supplemental Experimental Procedures). After applying all stimuli, depolarizing current was applied (1–3 nA, 1–5 min) to iontophoretically inject Neurobiotin when stability of recording allowed. Two different types of visual stimuli were applied during the experiments. First, linearly polarized light was presented from the zenith (as seen by the animal). Second, unpolarized light spots were presented at an elevation of 25°–30° (above the animal’s horizon). Both stimuli were connected to a rotation stage, which could be rotated by 360° in either direction.