Measurements of head capsule, mandible, metanotum, and body weight were done on larvae of Tenebrio molitor L. (Coleoptera: Tenebrionide) from the second to the last instar. Instar number varied from 14 to 18, but 15 or 16 instars were the most common. The value of dimensional measurements was evaluated as a tool for instar determination and to improve understanding of instar variation in T. molitor. Three methods of analysis were used, including discriminant, cluster, and frequency distribution analyses. Stepwise regression analysis was used to determine the optimal combination of dimensional variables. Head capsule width, left mandible length, and body weight at the beginning of stadia were the most significant variables impacting instar. Discriminant analyses of instar using these three variables resulted in 47.74% misclassifications, indicating a high degree of dimensional overlapping among instars. Dimensional overlap and variability increased in older instars. Cluster 15-group analysis using measurements of head capsule width and mandible length resulted in the lowest level of group variability; however, there was low degree of correspondence between cluster groups and instars. Frequency distribution analysis of head capsule width revealed 11 peaks, but only the early 8 peaks corresponded with observed instars. All three methods of morphometric analysis failed to identify all instars of T. molitor correctly. Frequency distribution analysis provided the best match, but only among instars 2 to 10. It appears that instar variation in the larval development of T. molitor occurs after the 10th instar. A model of T. molitor developmental plasticity is proposed.