Query restricted bottom-up evaluation of normal logic programs. David B. Kemp, Divesh Srivastava and Peter J. Stuckey. Several program transformations---magic sets, envelopes, NRSU transformations and context transformations, among others---have been proposed for efficiently computing the answers to a query while taking advantage of the query constants. These transformations use sideways information passing strategies (sips) to restrict bottom-up evaluation to facts potentially relevant to the query. It is of interest to extend these transformations to all logic programs with negation, and identify classes of programs and sips for which these transformations preserve well-founded models with respect to the query. In a previous paper we identified classes of programs and sips for which the magic sets transformation preserves well-founded models wrt the query. We continue this line of research to other transformations that use sips. We identify classes of programs and sips for which the context transformations and the envelopes transformations preserve well-founded models wrt the query. We also define a new program transformation based on magic sets that preserves well-founded models with respect to the query for any choice of sips. Finally, we compare and contrast the performance of envelopes with our new program transformation using the Aditi deductive database system.