Here's a Google search using "Haigh diagram mean stress ratio":

http://www.google.com/search?q=Haigh+diagram+mean+stress+ratio&btnG=Search&hl=en&source=hp&gbv=2&gs_sm=e&gs_upl=11607l17934l0l18247l35l33l0l22l22l0l188l1390l4.7l11l0&oq=Haigh+diagram+mean+stress+ratio&aq=f&aqi=&aql=
 

There is a mean stress effect calculator available for a
number of steels. If the drive shaft were Low carbon steel
for example, some of these files, (and calculators)
would probably be fine:
  http://fde.uwaterloo.ca/Fde/Materials/Steel/Lowcarbon/lowcarbon.html
Try the one for SAE1015, it should be reasonable for everything from
about SAE1005 to SAE1020.   The "Calculator", when one hits the submit
button after filling in the elastic hot spot stresses, returns 5 different mean stress calculation types
#xcalc3  
StrainLife_Reps : Ignores mean stress effect
SWaT_Life_Reps  :  Smith Watson Topper mean stress correction
                  (good for tensile mean stresses)
StressLife_Reps :  Ignores mean stress effect
Morrow_Reps     :  Morrow mean stress correction (best for Compressive
                  mean stresses
Goodman_Reps    :  Uses Goodman
(Reps= Repetions)
The "Reps" are how often one's entered stress history can be repeated
before failure.

Also returned is a graph of the expected plasticity corrected
local hot-spot stress strain behavior.

Note that for a drive shaft one should also evaluate what stress is
caused by the full throttle reverse.  This usually combines with the
drive stall stress to form the biggest cycle.  Probably one should enter
several of the expected largest cycles in the calculator.  The largest
cycle does not always cause the most damage.

If you are near a weld on the drive shaft you will need to
throw in an extra  stress concentration factor, perhaps 3.
This can be simulated in the calculator by changing the magnification
factor from 1.0 to 3.0