At one point my (purchased new) 91 Civic's water pump failed prematurely after
something less than 40k miles and a couple of years. I suspect the independent
shop that had done the last timing belt/water pump replacement used non-OEM
coolant. But I know I had replaced the coolant with green Prestone in the past,
so I might be the one to blame.

The car is now 1.75 years into a Havoline Dex changeover (this followed a very
thorough flush of the cooling system). I changed the thermostat a few months ago
and the Dexcool looked very clean, though that may mean nothing.

I plan to change the Dex in a few months, following the two year rule, despite
the advertisements on the Dex saying it will last longer. Why am I not heeding
the advertisements? Because of (1) all the suspicion here over non-OEM coolant;
(2) the lack of data on Dexcool for Hondas; and (3) my own experience with
non-OEM coolant.

When there's more positive data for Dex in Honda engines, I might reduce the
coolant changeout frequency.

Despite this little radiator coolant experiment, I do not plan on doing a
similar one with synthetic oil in my car's engine.

Caroline
(Pass the marshmallows.)