1: \begin{abstract}

2: \noindent

3: This paper examines sparse grid quadrature on weighted tensor products (\textsc{wtp}) of

4: reproducing kernel Hilbert spaces on products of the unit sphere $\Sphere^2$.

5: We describe a \textsc{wtp} quadrature algorithm based on an algorithm of Hegland~\cite{Heg03}, 

6: and also formulate a version of Wasil\-kowski and Wo\'znia\-kowski's \textsc{wtp} algorithm~\cite{WasW99}, 

7: here called the \textsc{ww} algorithm.

8: We prove that our algorithm is optimal and therefore lower in cost than the \textsc{ww} algorithm,

9: and therefore both algorithms have the optimal asymptotic rate of convergence given by

10: Theorem~3 of Wasil\-kowski and Wo\'znia\-kowski~\cite{WasW99}.

11: Even so, the initial rate of convergence can be very slow, 

12: if the dimension weights decay slowly enough.

13: \end{abstract}

14: