I had departed WVI for a long cross-country trip to PHX in
a Cessna 172.  On the climb from WVI, the oil temp
was just below the red line, and I had to level out at
7500 and 9500 for cooling.  Even so, the OAT was
15 degrees C at 9500 (very hot).  I leaned the mixture
to 50 degrees rich of peak EGT and had the throttle
full forward at 2500 RPM.  It was the middle of a hot 
afternoon flight.

  About twenty minutes of cruise had passed when the engine
started to vary from 2100 to 2500 RPM. 

  So fuel selector left, mixture rich, carb heat on
(helped a little, 200 RPM), mag check normal (a little
backfire when off), primer locked, mixture all the way
out (engine stops), mixture full in (engine back to
2200), throttle all the way back (engine stops), throttle
forward (engine back to 2200).

  By now the engine has stopped surging and stabilized at 
2200 RPM with the carb heat on and 2000 with carb heat off.

    I checked for the closest runway.  New Coalinga 
was about 15 miles away, so I continued to it.  

  I knew I could glide to the airport, so I advised 
Oakland Center I was making a precautionary landing at
C80 due to engine trouble.  They asked if there was anything
else they could do, and I replied I would let them know.

  I had at least ten minutes to diagnose the problem.
It wasn't carb ice, but pulling carb heat helped, so it
could have been a clogged air filter OR richening the
mixture with carb heat helped.  I think the richening
the mixture was what was helping.  

  Two things make an engine run - fuel and spark.  Since
the mag check yeilded normal 100 RPM drops, spark 
wasn't a problem.

  Of course many things can make an engine stop.  Low oil
pressure, redline oil temperature, broken cylinder rod,
stuck exhaust or intake valve, blown gaskets, etc.
In flight, these are hard to diagnose, and as far as I know,
impossible to fix (OK so some endurance record holders
have an in-flight catwalk, but not on a 172).  Since I had
good oil pressure and the oil temperature was steadily
decreasing, there was no oil problem.  It could still
be a valve problem, but I had no way to diagnose or fix
that in the air.

  So I tried to fix a possible fuel problem.  Water or
other contaminants in the fuel were possible, so I
switched from right and left tanks (leaving each for a
few minutes) and had no change.  Since there was no
water in the tanks (I shake them frequently and have 
new seals for the fuel sending units and the gas caps)
when I drained them, gas wasn't a problem, at least
not a fuel problem in the lines before the fuel selector.

  In the hot weather, fuel may have boiled and vapor
locked after the fuel selector, but would that allow
some fuel to get through for 2200 RPM?  I don't know.

  Fuel is gas AND air.  Maybe something had clogged the
air filter.  Carb heat is also alternate air, so that
would solve that problem.  But carb heat only partially
restored power, so that didn't make sense.  I've heard
of labels falling off the carb intake box and clogging
the air intake, but this is something I can't fix in
flight, and if such a label blocked the air intake,
carb heat would cause a reduction in RPM, not an
increase.

  At about 4000 MSL, I called Center and asked for an
800 number to call on the ground.  They gave me a number
to call collect, and I told them to send out a search team
if I didn't call in 15 minutes.  I was directly over the
airport spiraling down, but I wasn't going to take any
chances.  They gave me squawk VFR and good luck.  

  At about 3000 MSL, the engine magically started running
normally again.  I continued down for a normal landing,
and was able to radio Center from the ground to tell them
I was safe.

  On the ground, I checked oil, plugs (externally), fuel,
oil cap, and everything seemed normal.  I drained
normal fuel samples from both sides.  I couldn't find
anything wrong.  After some hemming and hawing, I decided
to depart back to WVI.  Coalinga has two runways in a 
T, so I did several runups and fast taxis on the
runway, and then set up for short field takeoff into the
wind.  At any point I was in a position to land on a
runway if the engine failed.  I circled the airport
all the way to 10,500 feet, and then headed to King City
KIC.  

  Based on the distances and 0 winds aloft, the GPS showed
I could glide to either airport at any point of the trip.
I continued uneventfully over KIC, Clark (pvt), Chalone
Vineyard (pvt), Salinas (SNS), and then cut power at
a point that would put me 1000 feet over WVI.  I cleared
the throttle occasionally, and the glide was perfect.
I landed right on the numbers and taxied in.  I had
6 gallons of gas left (I kept the plane light for
better climb).

  I tied down the plane, and then pulled the cowling off.
The left exhust pipe had come loose from the friction
fitting.  The pipe clamp was installed wrong, and I
rotated it 90 degrees and put the pin through the
pipe and outlet.  Problem easily solved.

  The problem happened when the pipe was improperly
re-installed after the annual.  I had thought it was
a stuck exhaust valve, but to prevent future valve problems: 

  I won't allow the oil temperature to get close to redline.

  I will use 15W-40 multi-viscosity oil instead of 40 weight.

  I will use lower lead fuels (80/87) and get an autogas STC.
 
  I will not lean as aggressively.  100 to 150 rich of peak.

  I will get an STC oil cooler to keep the temperatures down.