For at least the past 20 m.y. the Bouvet triple junction appears to have existed predominantly in a ridge-fault-fault rather than a ridge-ridge-ridge configuration when it has an isosceles velocity vector triangle that allows either configuration to be stable. In this paper two conceptual models originally developed to explain the general orthogonality of ridges and transform faults in spreading center systems are examined to see if they can explain the observed predominance of the ridge-fault-fault geometry for this triple junction. The first hypothesis is that a ridge axis will form and evolve to remain oriented perpendicular to the direction of the maximum local tens fie stress at the ridge axis; at a triple junction this can be strongly modified by the local plate boundary geometry. The second hypothesis is that plate boundary geometries adjust to minimize the power dissipation associated with plate boundary motions. Both hypotheses successfully predict the observed predominance of the ridge-fault-fault geometry. We suspect that the minimum dissipation hypothesis is successful for spreading center problems, despite its questionable theoretical underpinning, because for these problems it is equivalent to plate geometries assuming a minimum resistive force configuration. Since a minimum resistive force configuration is a partial corollary of a local stress-controlled ridge axis geometry, the resulting minimum dissipation hypothesis plate geometry predictions will be similar to stress-controlled plate boundary geometries. Because of its simplicity, the minimum dissipation hypothesis may remain a useful tool for predicting plate boundary configurations provided it is realized that this is more an approximate rule of thumb than a hard physical theory. Observations of the occurrence and orientation of episodic overshooting of ridge segments past the stable triple junction location should provide a better test that both fine-and large-scale triple junction tectonics are stress controlled.