The growth of ten exoskeletal structures has been studied throughout the post-embryonic development in Stagmatoptera biocellata. The selection of an appropriate statistical procedure to analyse the data was discussed. In spite of theoretical prediction the least-squares regression technique turned out to be as efficient as the Bartlett (1949) procedure for studying growth. Longitudinal treatment of longitudinal data proved to be the best method for this kind of analysis. All the dimensions investigated showed some statistically significant degree of allometric growth and could be ascribed to certain class of allometry. The adaptative value of some allometric growth patterns was discussed in terms of physiological efficiency. A hypothesis put forth by Balderrama & Maldonado (1973) about the anatomic relationships between the parts involved in the mantis strike, received clear support from this analysis. A quantitative assessment of Dyar's (1890) and Przibram's (1912) rules and their modifications led us to conclude that they are untenable at least for the mantid Stagmatoptera biocellata. However, some periods of constant growth ratio could be identified during the post-embryonic development and the concept of maximum growth period was put forward as a complementary element for typifying structures.