The sensory bristles and epidermal hairs of Drosophila have proven to be valuable model cell types for studying the role of the cytoskeleton in cellular morphogenesis. We have recently begun to use the arista laterals as a third model cell type. The laterals display a combination of bristle and hair characteristics and provide a system where we can compare the relative importance of specific genes and subcellular structures for the morphogenesis of different polarized cellular extensions. We have characterized the lateral phenotype of a collection of mutations selected because of their phenotypes in hairs and bristles. In many but not all ways the lateral phenotypes are similar to the hair and bristle phenotypes. We provide compelling genetic evidence for the importance of the actin cytoskeleton in lateral elongation, shaping and integrity. Our observations provide evidence that defects in actin bundling can destabilize laterals so that they split during growth. Temperature shift experiments suggest that a defect in lateral initiation can lead to subsequent splitting. These observations provide a link between multiple hair and lateral cells forming by both multiple initiation events and by the splitting of individual cellular extensions. We also found that mutations that lead to lateral splitting typically alter the stereotypic arrangement of actin filament bundles and microtubules in laterals.