OBJECTIVE Animal studies have suggested that peripheral nerve transection results in substantial loss of ganglion cells and the selective survival of cells based on size. The implications are that subsequent repair of peripheral nerve injuries will be determined by the numerical density and character of the surviving cells. The purpose of this study was twofold: First, to determine the effect of mental nerve transection without repair on trigeminal ganglion cell density and morphology in adult rats, and second, to determine the variation of trigeminal ganglion cell density and morphology after immediate and delayed repair. METHODS In the first part of the study, 12 adult male Sprague-Dawley rats had their mental nerves exposed bilaterally (n = 24). Twelve mental nerves were then transected and prevented from regenerating, and the remaining 12 nerves were uninjured. Ninety and 180 days after transection or sham surgery, the trigeminal ganglia were serially cut into 5 microm longitudinal sections along the dorsoventral axis. The volume and volume density of the mandibular mental subdivision containing sensory cells was determined at each section level with point-counting methods. The numerical density and total number of cells was estimated on the same section, using an unbiased three-dimensional stereological probe, the dissector. Cell size and shape determinants were estimated using the dissector and computerized planimetry. In the second part of the study, six rats had the mental nerves transected bilaterally and immediately repaired by microscopic sutures. In six additional rats, the repair was delayed for 90 days. In both groups, the trigeminal ganglia were serially cut at 30, 60, and 90 days post-repair and stereologic estimates of numerical density and histomorphometry were examined using the dissector and computed planimetry. RESULTS In the trigeminal ganglia of the 12 sham-operated animals, the mean number of cells was 20.6 x 10(3) (+/-2.9 X 10(3)). After nerve section, the mean number of cells was 10.88 X 10(3) (+/-0.9 X 10(3)), representing a 47% reduction. The mean volume of the mandibular subdivision cells in the ganglia of the sham surgery animals was 0.3 mm3 (+/-0.05) and 0.22 mm3 (+/-0.04) in nerve-sectioned ganglia, a 38% difference. There were no ganglia cell size or shape differences between the two groups. The mean number of cells in the ganglia of immediately repaired nerves was 10.66 x 10(3)(+/-1.1 X 10(3)), and it was 12.45 x 10(3) (+/-0.9 x 10(3)) after delayed repair. The numerical density was significantly less than in the sham surgery ganglia but not different from that of the transection/no repair ganglia. The weighted mean reference volume of the mandibular subdivision after immediate and delayed repair was similar and was significantly greater than the transection/no repair group, but not different from the sham surgery group. The cell size was slightly larger in delayed-repair ganglia compared with immediate-repair ganglia, but the differences were not significant. There were no significant differences in any of the stereologic estimates when analyzed across treatment time. CONCLUSIONS The results of this study agree with previous reports that peripheral nerve transection produces a substantial loss of nerve cells within specified regions of sensory ganglions. However, the results conflict with evidence that cells survive transection based on size and shape. These findings also indicate that in the adult rat the substantial loss of nerve cells was unaltered by the reconnection of their severed axons. Neither immediate or delayed repair of the transected nerve altered the spectrum of surviving cells based on size or shape. The reestablishment of the mean reference volume of the mandibular subdivision after section and repair suggests that demands made on regenerating axons appear to result in the restoration of ganglionic volume normally lost after axotomy, probably the result of axo