2nd Update, as I think now the charging system is operating normally.

I checked. It drops to something between 11.8 and 13 volts *momentarily*.

After the slight drop, it rises pretty quickly (3-10 seconds?) to something
between 14 and 14.5 volts.

When I turn off the loads, it stays up over 14 volts for a little while. (I
didn't wait for it to drop.)

If I then turn off the car, then turn it back on (no loads), the voltage is back
to about 12.7 volts (here in the summertime).

I can repeat the above events merely by turning on the headlights, though it's a
bit more dramatic with more electrical loads.

I'm buying this.

My best guess as to what was happening in February (when unloaded, the battery
terminal voltage was 14.4 volts) is that it is somehow related to the
particularly cooler temperatures then compared to now.

Darn tootin'. I'm sure as heck not going to replace an OEM, dealer installed
alternator after a lousy five years and 50k miles... :-) That puppy better make
it to 100k miles or eight years, AFAIC.

I hope!

Autozone Aside: I drew on the "expertise" of the fine folks at Autozone today.
They did what they said was an "alternator test." The guy hooked up to the
battery terminals a sophisticated-looking electrical tester. (IOW, it does more
than my $15 Radio Shack digital voltmeter.) He read off voltages while operating
the throttle control manually (under the hood). He got the same sub-13 volts I
got. He never changed the electrical loads. His conclusion: Buy a new
alternator. My conclusion, "All he did was what I pretty much did, so not so
fast." I thought he'd at least take some readings at the alternator terminal.
But one gets what one pays for, I suppose. This check was free.

Firestone (where I bought my new Insterstate battery) aside: After buying and
installing (in the Firestone parking lot) my new battery Monday, I checked the
voltage and saw no change from my previous readings of about 12.7 volts. I
popped into the Firestone showroom and asked the guy if he'd give me his
opinion. He said I was wrong about the 14+ volts being usual when the car was
running. I said okay and left, not wishing to start a row and knowing this was
my problem, anyway. At this point, I feel the guy was speaking with some
legitimacy.

Lastly: Thanks, Jim.

does the voltage drop under load?  you know, fans, lights, defrost etc.,
on?  if it drops to say 10V, then you have a charging problem.  if not,
and it kicks up to 13/14V when loaded, everything's working just fine.

the alternator should have a regulator circuit that determines whether
it needs to "charge hard" or not.  if the battery is charged and there's
minimal load, the alternator is not required to produce max output, and
indeed it shouldn't in order to not fry the battery.

alternators generally fail when the diodes go.  semiconductors have a
limited lifetime at high temperatures, and hot climates and/or full
electrical loads will keep them nice & toasty.  oem alternators can last
a good long time when treated conservatively.

1991 Civic LX 4-door sedan, 1.5 Liter, manual transmission, no air conditioning,
156k miles, here.

Voltage at the battery terminals when the car is
-- idling = about 12.7 volts
-- stopped, ignition off, also about 12.7 volts

In February, these numbers were 14.5 volts and 12. 4 volts.

Internet sources say a voltage when the car is idling of around 14 volts or so
indicates a properly operating charging system. So something is wrong, IMO.

I put in a new battery (Interstate) today, replacing the old one (4-years-old;
Diehard; wrong climate design as I moved from up North to the Southwest in the
past year?). The voltages above didn't change.

The car is on its second alternator (OEM). This 2nd alternator is 5 years and
50k miles old.

I installed a new alternator belt in June. I originally had the belt too loose,
as indicated by a squeal at cold startup. I tightened it and the squeal stopped.
Unfortunately I did not at this time check the battery terminal voltage.

I did shake out quite a lot of white powder (indicating some corrosion) from the
battery's positive terminal's cable connector while changing the batteries.

I'm going to do the checks at
http://www.honda.co.uk/owner/ConcertoManual/62sk301/16-66.pdf this weekend.

Meanwhile, has anyone seen a condition like this? If so, what was the fix?

I think this is not critical unless I do a lot of driving with the lights on
(which I do not, as I drive mostly during daylight hours). But I do feel my
battery isn't going to be maintained at optimal charge and so, as winter
approaches, I need to fix this.