The kinetics of repair of sublethal radiation damage was examined in the cervical spinal cord of developing, 1-week-old rats. Split-dose irradiation treatments were given with time intervals of 0.5-96 h. The data, supplemented with fractionation data from previous experiments, were analysed using direct analysis based on the incomplete repair (IR) model. The best fit to the monoexponential repair model resulted in an estimated half-time of repair (T1/2) of 1.5 (1.3-1.8) h. No indications of a slow or second component of repair could be detected in the 1-week-old cervical spinal cord. This is in contrast with literature reports of experiments with the adult rat cervical spinal cord, suggesting bi-exponential repair, with 65% of the damage repaired with a slow repair T1/2 of about 4 h. Two-step methods, although statistically inferior to direct analysis, are still in use for analysis of repair experiments. A number of two-step methods used for data analysis in previous reports concerning sublethal damage repair, are dose (un)repaired, proportion of dose unrepaired, and proportion of damage unrepaired. It is argued that of the methods discussed, only analysis of the data expressing the results as proportion unrepaired damage (delta Eu) and using split-dose experiments, does not result in introduction of an artificial second repair T1/2 in tissues with a high fractionation sensitivity. Two-step analysis of the present data using delta Eu suggested monoexponential repair with a T1/2 value of 1.5 (SE 0.2) h.