(A) Timing of Eyeless reduction in INP lineages. Transient inactivation of ts.Gal80 (29°C heat pulse; gray bar) results in transient EyelessRNAi during the time in which E-PG neurons are normally generated. (B,C) The manipulation in A does not alter CX neuropil morphology as seen by nc82 (neuropil) staining (EB, shown; other neuropils, data not shown). (D) Schematic of experimental apparatus for sun navigation experiments. The wing stroke amplitudes of a tethered, flying fly were monitored with an IR camera; the stroke difference determined the angular velocity of a 2.4° sun stimulus. Modified from (Giraldo et al., 2018). (E) Example flight (top panel) and summary data (bottom panel) from ts.EyelessRNAi control with no heat pulse. Top panel: left plot shows headings over 5 min flight; 0° is sun position in front of fly. Right histogram is distribution of headings in this example flight; sideways green triangle is the mean. Bottom panel: summary data, with each 5 min flight represented by radial lines. The angle of each line is the mean flight heading. The length of each line is vector strength of flight, varying from 0 (center of circle; no stimulus stabilization) to 1 (edge of circle; perfect stabilization). Each fly flew for two 5 min flights separated by a 5 min rest period. Cyan line shows mean heading, across flights with vector strength >0.2, as well as 95% confidence interval, calculated via resampling across flies. 44 flights, 22 flies total. (F) Example and summary data from ts.mcherryRNAi control. Same plotting convention as (E). 48 flights, 24 flies. (G) Example and summary data from ts.EyelessRNAi flies with heat pulse. Same plotting convention as (E,F). 50 flights, 25 flies. (H) Cumulative probability distribution of vector strengths from both control groups (black and gray) and from experimental group (red). There was no significant difference between means (ts.EyelessRNAi heat pulse, 0.34, red; ts.EyelessRNAi no heat pulse, 0.31,black;ts.mcherryRNAi heat pulse 0.30,gray; p>0.1, permutation test). (I) Cumulative probability distribution of mean absolute headings. The heading distribution for the experimental group, ts.EyelessRNAi heat pulse, was skewed significantly to frontal headings (mean 63.7°, 37 flights in 23 flies with vector strength >0.2) compared to control distributions (p<0.01, permutation test; ts.EyelessRNAi no heat pulse, mean 107.9°, 35 flights in 19 flies; ts.mcherryRNAi, mean 106.9°, 31 flights in 21 flies). Scale bars, 10µm.