I used all of this advice and a couple of long wires to run part of the test. After I found a good ground while sitting in the driver's seat (mounting screw for the courtesy light, which burns plenty bright), I found the gauge works.
That said, the voltage coming down the wire to the sender fluctuates wildly. It peaks in the 6 V range but drops into millivolts every second or two. I hadn't found the specs on resistance before I had to run some errands (daily driver is getting badly needed new tires and alignment), so in order to not have things dangling reattached everything. Now, I'm getting a reading on the temperature gauge again. It pulled from pegged cold into the low range. This is plausible since it's in the high 30s here, and I was doing highway driving with fan that always rotates with the crank. The radiator and hoses did not feel hot.
2+2, I have no reason to believe that the current sender is of reliable stock. Indeed, when the car overheated this summer, I would have liked it to have gotten much higher on the meter before puking fluid, but it was the hottest day of the year and the thermostat seized. Still, it was at 3/4 when I turned the wipers on and took evasive action to get to the roadside.
At this point I think I have some old wires that are not reliable, a sender of dubious quality, and likely some problems with the voltage from the instrument cluster. I'll post on the last in a separate thread since the problems extend beyond the gauges, but the short version is that the cluster lighting intensity and frequency of the turn signal varies with engine speed, while headlight intensity does not.
Sorry for the long post and thanks for all of the responses.