A model of spinal cord trauma in guinea-pigs, using lateral compression to a set thickness, produces a delayed functional loss at one to two days, followed by a partial recovery over several weeks, as measured using hindlimb motor behavior, vestibulospinal reflex testing, and mapping the receptive field of the cutaneous trunci muscle reflex. The role of inflammatory events in these secondary changes, was investigated with intraperitoneal injections of the macrophage toxin, silica. In one experiment, 11 matched pairs of animals were injured. One of each pair was selected randomly and injected with a suspension of 1.2 g of silica dust in sterile saline, immediately after injury and surgical closure. In a second experiment, involving 10 pairs of guinea-pigs, a similar dose of silica was administered to one of each pair at either one or two days before the injury. The animals survived up to three months, then were fixed by perfusion with glutaraldehyde. Histopathology of the lesion was quantified by line sampling of myelinated axons, and by measurement of blood vessels, in plastic sections through the center of the lesion. Surgery, injury, analysis of behavior and histology were all performed without knowledge of the experimental status of the animal. The secondary onset of functional loss below the lesion appeared to be delayed by one to two days in silica-treated animals with respect to controls. The number of myelinated axons at the center of the lesion, examined at two weeks to three months after injury was higher in the animals injected with silica immediately after surgery, most significantly in the dorsal quadrant of the cord. Myelin sheath thickness and axon caliber distribution were not different. Hypervascularity of the lesion was significantly reduced in animals injected with silica within one day of injury. These findings support the hypothesis that inflammatory activity plays an important role in secondary tissue damage, and that it may be responsible for some proportion of long-term neurological deficits, but do not suggest a prominent role for early macrophage activity in the mechanisms of demyelination.