Abstract
In recent years, a wide array of tools originally developed for molecular dating analyses has been adapted for use within a morphological clock perspective. This is of paramount relevance for taxonomic groups that cannot be sampled in forms suitable for DNA extraction. The Scolebythidae, a small family of stinging wasps, represents an interesting group to study the application of the morphological clock due to their limited extant diversity contrasted with their abundant Cretaceous fossil record. Here, we investigate the impact of model specifications on the Bayesian inference of dated phylogenetic trees for the family. A total of 78 models were evaluated through Bayes Factors, with a thorough comparison of topology and node ages across the resulting trees. We find that the partitioning of characters had slight effects on ages retrieved, even though partitioning by homoplasy improved model fit and yielded narrower ranges for age estimates. The uniform tree prior consistently provided older age estimates compared to the fossilised birth‐death prior. Models employing an Independent Gamma Rates (IGR) clock and lognormally distributed rate resulted in higher marginal likelihood values. While calibration priors had limited impact, the root age prior substantially affected both the estimates and overall model performance. We discuss the evolutionary history and biogeography of Scolebythidae, highlighting the relevance of dispersal events in shaping their modern disjunct distribution. Advances in the application of the morphological clock may be essential for bridging the extant diversity with the ever‐growing fossil record.