Fig. 5. Scanning electron micrographs of adult eye surfaces. (A) GMRp35. Mild eye roughening is manifested by subtle variations in facet size and shape and orientation of interommatidial bristles. (B) GMRDIAP1/+. Eye morphology was normal despite suppression of cell death (Hay et al., 1995). (C) GMRDIAP1/GMRp35. Eye morphology was rescued compared with GMRp35. If eye roughness was caused by survival of supernumerary cells, eye roughness would be expected to be enhanced or unaffected by co-expression of DIAP1 and p35. If p35 had effects independent of cell survival, eye roughness would be expected to be unaffected by co-expression of DIAP1 and p35. If Dronc activity perturbed eye development independently of cell survival, eye roughness would be expected to be reduced by co-expression of DIAP1 and p35 because DIAP1 inhibits Dronc (Hawkins et al., 2000; Meier et al., 2000b). (D) GMRDronc-DN/GMRp35. Eye morphology was rescued when compared with GMRp35, as predicted if Dronc activation were responsible for the eye roughening that occurred when p35 blocked cell death. (E) GMRDronc/GMRp35. Mild eye roughening was similar to GMRp35. (F) GMRDronc/+. Eye morphology was normal. (G) GMRDronc-DN/+. Eye morphology was normal. (H) GMRDronc/GMRDIAP1. Eye morphology was normal. (I) GMRDronc-DN/GMRDIAP1. Eye morphology was normal.