A stab at synthesis...

One definite limit to circulation in when -- at a certain distance off-axis -- the permeability drops to zero. A recent re-evaluation of "sealing" age (the age where the observed heat flow approximately equals that predicted, suggesting hydrothermal heat transfer has been largely ceased) by Stein and Stein, (1994) concluded that crust in the Pacific, Atlantic, or Indian oceans seals at 65 +/- 10 Ma. They argue that overlying sediment plays a secondary role to the age of the crust in determining the lateral (off-axis) extent of hydrothermal circulation; their data demonstrate that even ~200 m of sediment is "neither necessary or sufficient to seal the crust" prior to an age of >50 Ma.

Juxtaposition of Stein and Stein's results with those of Rohr (1994) and Davis (1996) suggests that although seismic velocity increases dramatically at the topographic step where sediment thickens suddenly 0.6 Ma off-axis, hydrothermal circulation within the crust will continue all the way to (and through?!) the subduction zone. That hydrothermal circulation of reduced intensity is taking place at 1.2 Ma is evidenced by the high-resolution data of Davis et al (1996).

It is of couse tempting at this point to strut to the edge of the precipice and leap -- via breath-taking swan dive -- into the raging surf of speculation. If your curiosity inspires you, go ahead and multiply the distance from Endeavour ridge crest to subduction zone by the Layer 2A thickness and overlay your guess of where the growth temperatures and pressures of subsurface microbes might lie relative to the isothermic and isobaric surfaces, and -- presto! -- (with a bit of thought about porosity and permeability) you'll have an estimate of subsurface habitat. I'll join you in a bit; first I want to work out some of the details...

Outline