Ditto some of what Dan Houg wrote about the wasserboxer
idle.

My '84 has shown similar symptoms over the past decade.
This summer, however, a new situation came to my attention
that still has me puzzled.

When temperatures approached and passed the 100 mark,
suddenly the idle speed on my Vanagon began to increase by
itself.  I thought "lubricate the cable" or "something is
stuck" or some other possibility.  Once temperatures went
back down, the idle did too.

Now I wonder if the conditions Dan outlines are simply the
first step (fan kicks on) and if that doesn't work
*eventually* the idle will pick up to a point that the
coolant will circulate properly.

Just a hunch.  No empirical data.  Anybody else experience
this?

-Jim Bryant