Example of how we can add broad-brush constraints to the nonparametric bayesian model

e.g. statements of continuity, concavity, etc, without having to be more explicit about the functional form than can be justified by the biology.  

Illustrate how the addition of these constraints improves performance.  
