1: \begin{abstract}

2: This main paper develops a general theory on rates of convergence of

3: penalized spline estimators for function estimation

4: when the likelihood functional is concave in candidate functions, where the likelihood is

5: interpreted in a broad sense that includes conditional likelihood,

6: quasi-likelihood, and pseudo-likelihood. The general theory is applicable to obtain results

7: in a variety of contexts. This supplementary document contains three

8: main topics: i. a literarture review of related asymptotic theory for smoothing splines

9: and polynomial splines; ii. additional examples to illustrate the

10: application of the general theory; iii. extension of the general

11: theory in the main paper to penalized tensor product splines and

12: penalized bivariate splines on triangulations.

13: \end{abstract}

14: