Okay, here's the scoop from Bentley.

 First of all, what I said earlier about the throttle switch enriching the 
 mixture was correct ONLY if you have a fuel injected bus built before 1977.
 The enriching function was "incorporated into the control unit" on later 
 models.

 Here's the rap on oxygen sensors on '79 and later California air-cooleds.

 "The oxygen sensor generates form 0-1 volt depending on sensor temperature
 and exhaust gas composition. The varying voltage is measured by the injection
 systems control unit which alter the injector-open-time to give low emissions.
 The sensor operates neither below 300 degrees C nor when turned off at full
 throttle by a microswitch on the throttle valve housing."

 So there you are. What Bentley doen't say is WHY it is switched off at full
 throttle. My guess is that the aforementioned enrichening would drive the O2
 circuit to correct for it if the sensor wasn't switched out of the circuit,
 but that's just a guess.

 Response to David Carment on EGR:

 Your remarks about when the EGR is on and off are backwards, I think. The EGR
 on my '74 is off at idle and off at full throttle and switched on in between.
 If the EGR valve is open at idle it behaves as a nasty vacuum leak and trashes
 the idle. Mine is electrically switched by a cam on the right hand carb 
 linkage. Pushing the switch and opening the valve at idle speed cause the   
 engine to die. Incidentally, that's what the smog inspector did to make
 sure the EGR was connected and working.

 Back to Nitrogen oxides - I think that the combustion temperature is affected
 by many things. For example, my '72 controlled NOX emissions by turning off 
 the vacuum advance when the engine was warm. I read that richening up the 
 mixture a little will cool the combustion temp somewhat. What I'm getting
 at is that all these things are related:  you may not want to 
 dis-connect your EGR because your injection control unit is setting the
 mixture assuming that some of the molecules in the mixture don't have an
 oxygen to contribue to combustion. The spark may be timed with the same
 assumption. 

 Dave Kautz
 dkautz@hpsidms1.sid.hp.com