Developmental toxicity risk assessment has typically relied on the estimation of reference doses or reference concentrations based on the use of no-observed-adverse-effect levels (NOAELs) divided by uncertainty factors. The benchmark dose (BMD) approach has been proposed as an alternative basis for reference value calculations. In this analysis of the developmental toxicity observed in rats exposed to boric acid in their diet, BMD analyses have been conducted using two existing studies. By considering various end points (rib XIII effects, variations of the first lumbar rib, and fetal weight changes) and various modeling approaches for those end points, the best approach for incorporating all of the information available from those studies could be determined. Particular emphasis has been placed on methods for combining data across studies and for combining potentially related effects (on rib XIII and on the first lumbar rib). The issues of study and end point selection are ones that will arise frequently in the process of estimating reference values. This example of boric acid suggests that the BMD approach provides a reasonable basis for appropriately comparing and combining study data, as opposed to ad hoc combinations of study results. Moreover, it is shown that the BMD approach can be used with combinations of end points considered to differ in severity. In this case, the preferred approach involved combining the data from the two studies, which were similarly designed and were conducted in the same laboratory, to calculate BMDs that were more accurate and more precise than those that could be derived from either study alone. It was determined that decreased fetal body weight provided the best basis for BMD calculations; BMDs calculated for fetal body weight changes were less than those for all other relevant end points. The appropriate BMD to use as the basis for boric acid reference dose calculation appears to be 59 mg/kg/day, which is very similar to the NOAEL observed in the second of the two studies (55 mg/kg/day). Although the first study failed to establish a NOAEL, the BMD approach could have been applied to that study, thereby avoiding the need for a repeat study. Similar BMD results were obtained in both studies.