Abstract The origin of the long virus rods associated with insect granuloses is investigated. The short thick virus rod is thought to emerge from its occluding crystal and then to elongate in the cell cytoplasm of the host, thus giving rise to the long rod. Uncomtrolled branching then takes place inside the cell; the branching is mostly at right angles and produces numerous bizarre shapes. The long rods have an inner core which appears to be a helix; negative staining with phosphotungstic acid tends to split open the long rods. Electron micrographs of the various phenomena are presented and some possible explanations of these phenomena are offered.