Orders of magnitude separate the scales at which field ecologists work relative to the scales of grid boxes of global climate models. We suggest a procedure, strategic cyclical scaling (SCS), to help cross disciplinary and scale boundaries. Process based "bottom-up" relationships are used to predict behavior at large scales, which is tested against large scale data for a "top down" evaluation. To the extent that discrepancies are revealed (very likely), the bottom up formulae are modified and further top down tests performed. Cycling between large and small scales should thus produce more credible overall results - if the cycling process converges. Issues of convergence, the strategic importance of problems chosen for research and specific case studies (e.g., animal-climate range limits and thermohaline circulation collapse) are presented to illustrate the problems and prospects for the SCS approach to bridging gaps across disciplines and scales.