Morphospace studies are rich in pattern and process. Techniques for adequate description and mapping of morphologies have been increasingly refined and applied, the same being true of metrics for relevant parameters (like disparity). However, the testing of process hypotheses for specific patterns of morphospace occupation in time and space is less refined and demands more intensive scrutiny. The polarization of ecological and developmental explanations entails a need to properly tease apart their respective contributions. There are different ways to go about this problem. Here I describe one approach: the isolation of development as a target for testing via the construction of developmental morphospaces. <p> Comparison of differently constructed morphospaces (one reflecting development directly, the other indirectly) provides a way of consistently studying the impact of development in constraining or facilitating changes in diversity. Congruence in range and/or location of "developmental" and "non-developmental" morphospaces, when properly interpreted to account for possible stochastic effects, is powerful evidence for a controlling influence of ontogeny. Appropriate choices involve three kinds of contrasts (with qualifications): theoretical (generative) vs. empirical (descriptive), abnormal (teratological) vs. normal, and juvenile vs. adult morphospaces. Theoretical morphospaces encapsulate logically simple principles of form generation, mimicking the potential simplicity of epigenetic processes. Teratological morphospaces are based on unsuccessful experiments in evolution, but any regularities encountered must reflect internal constraints and opportunities. Juvenile morphospaces are likely to depict a range of pronounced allometries that can serve as raw material for evolution. I provide an extended example of juvenile vs. adult morphospace comparisons with a case study on the evolutionary history of the echinoid order Spatangoida. <p> The broad pathways of morphospace occupation for adults in the order are contrasted with the distribution of juveniles in a developmental morphospace constructed from geometric morphometrics of post-metamorphic specimens. To ensure comparability, the same homologous landmarks are used for adults and juveniles. Disparity between pooled adults and pooled juveniles is compared, and also interpreted in a phylogenetic framework. The distributions of resampled total juvenile and adult disparities are indistinguishable, but there is much dissociation of growth trajectories in morphospace, with many instances of taxonomically heterogeneous juvenile/adult clusters. There are cases of demonstrable association of derived juveniles and primitive adults, pointing to peramorphosis, and of primitive juveniles and derived adults, pointing to paedomorphosis. Instances of clustering of juveniles from different taxa also occur, thus partially supporting Von Baer's second law. Different composite variables, however, suggest no consistent support for VBSL, with cases of similar disparity in juveniles and adults (contra VBSL), of higher disparity in adults (pro VBSL), and of higher disparity in juveniles (contra VBSL). When developmental disparity is mapped on a phylogeny, lack of support for VBSL is again implied: juvenile/adult disparity contrasts based on clade ranks indicate that within subclades juveniles tend to be more variable than adults. There is thus no single vector of change in disparity through ontogeny. Still, all such developmental action potentially driving evolution occurs within a constrained morphospace which is similarly partitioned in ontogenetic and evolutionary time. These results are in disagreement with what can be inferred from vertebrate model systems; their generality should be further explored with analyses of developmental disparity in other invertebrate groups. <p> <p> To appear in Chapman, R. E., Rasskin-Gutman, D. and Wills, M. (eds.): Morphospace Concepts and Applications. Cambridge, UK: Cambridge University Press (publisher and date of publication to be confirmed).
