Central to biological processes is the regulation rendered by GTPases. Until recently, the GTP hydrolysis mechanism, exemplified by Ras-family (and G-α) GTPases, was thought to be universal. This mechanism utilizes a conserved catalytic Gln supplied "in cis" from the GTPase and an arginine finger "in trans" from a GAP (GTPase activating protein) to stabilize the transition state. However, intriguingly different mechanisms are operative in structurally similar GTPases. MnmE and dynamin like cation-dependent GTPases lack the catalytic Gln and instead employ a Glu/Asp/Ser situated elsewhere and in place of the arginine finger use a K(+) or Na(+) ion. In contrast, Rab33 possesses the Gln but does not utilize it for catalysis; instead, the GAP supplies both a catalytic Gln and an arginine finger in trans. Deciphering the underlying principles that unify seemingly unrelated mechanisms is central to understanding how diverse mechanisms evolve. Here, we recognize that steric hindrance between active site residues is a criterion governing the mechanism employed by a given GTPase. The Arf-ArfGAP structure is testimony to this concept of spatial (in)compatibility of active site residues. This understanding allows us to predict an as yet unreported hydrolysis mechanism and clarifies unexplained observations about catalysis by Rab11 and the need for HAS-GTPases to employ a different mechanism. This understanding would be valuable for experiments in which abolishing GTP hydrolysis or generating constitutively active forms of a GTPase is important.