The initial results for the Floating Buffer R-Tree show some promise in that they suggest that the technique represents an improvement over the Buffered R-Tree for answering bulk queries of the form described in this report. However, if the technique cannot be shown to outperform the one-by-one query method on the standard R-Tree, then it will be of very limited usefulness. It is expected that as further testing is done the I/O counting routines will be improved upon and more accurate results obtained. Furthermore, the tests that have been performed to this point are still quite limited using small query sets and small databases and unrealistically small blocksizes and memory limits. This is due to some extent to some instability in the current code with larger datasets and some limitations in the testing environment that still need to be worked out. Once these issues are resolved it is hoped that the advantages of the Floating Buffer Tree over the standard one-by-one query method can be demonstrated.