Normal driving of a Model T Ford causes the coolant to expand, sometimes filling the upper radiator tank. The excess volumn then discharges from the system via the overflow tube, if it does not leak out the radiator cap first. Most people with normal Model T Ford thermo syphon systems leave the coolant at a reduced level to prevent this overflow.
A fairly hot, but not boiling, engine will often gurgle and burp when it is shut down and the residual heat in the block and head causes the now static coolant to boil a bit. This often leads to putting water out the overflow tube onto the pavement. This along with other leaks causes the coolant level in the radiator to drop even lower during normal operations.
Use of the vehicle brakes can also cause the coolant to surge forward in the block and head. This momentarily drives up the level in the upper radiator tank, sometimes resulting in coolant overflow and a lower normal operating level in the system.
The usual temperature indicator used with Model Ts is the well known MotoMeter on the radiator cap. This reads the coolant temperature via a probe at the bottom end of the unit. This probe does not extend very far below the radiator cap. In a normal car the probe only sees water vapor and the heat transfer is not efficient. In addition, the upper metal of the radiator tank is not immersed in water and loses heat to the air flow so it runs at a lower temperature. This results in a MotoMeter indication that is lower than the actual coolant temperature. The MotoMeter tends to indicate low until the car approches the boiling point, when the coolant raises up to the level of the sensor, then the indication rapidly climbs up to the top of the MotoMeter scale.
I have run static tests of temperature vs water level below the Motometer. I drained the water to 1.85 inches below the overflow tube, 1.94 below the lower tip of the Motometer. I did a number of measurements with a calibrated temperature probe and found a 20 to 25 degree difference between the motometer location and the water below. This water level was just below the neck of the radiator. It is to be expected that the difference would be considerably more in moving air, where the radiator neck is cooled by the air flow when the car is being driven.
The onset of the periods of thermal syphon will also cause variations in the coolant level in the upper radiator tank, resulting in variations in the MotoMeter indications. In my 1926 Tudor, I had an experience where the water pump was not running. The temperature gauge and the motometer would cycle up to around 200 degrees and then cycle back down as the thermosyphon action set in - this car has a water pump with a fair amount of clearance around the impeller so that free flow of water is possible.
My experience is with my two Model T Fords, but it is probably applicable to Model A Fords as well. I have 180 degree thermostats, direct reading temperature gauges and Motometers up at the radiator cap. I also use a non-pressurized coolant recovery system so that the Motometer is always in water at the top of the radiator. The gauge and the Motometer have been cross calibrated so that I know how accurate they are. The top of the radiator almost always reads higher than the direct gauge at the head outlet. This appears to be due to the hottest water going through the thermostat bleed hole and, being lighter, collecting at the top of the radiator due to the very low flow with a closed thermostat. This minor flow is probably very smooth and laminar so that it does not mix with the other water in the outlet hose on its way to the radiator. I have checked this several times and have used a calibrated temperature probe in the water by the two locations to confirm my observations. The temperature gauge location is also subject to air flow from the engine fan, so some local cooling may result from that.
In a test of this difference phenomena, I warmed the car by driving about 15 miles. Air temperature was at about 70 F. The Motometer usually read about 180 with the car in constant motion; the normal temperature gauge in the coolant outlet typically read about 160. The thermostat did not open during this test - the regular temp sensor would have gone up more if it did. Both the Motometer and the temperature sensor went down during stops at traffic lights. The coolant level had been topped off cold prior to this test. The coolant recovery tank showed a max level of about 1.375 inches. This tank is 5 inches square. One has to assume that the bulk coolant temperature was a bit below the 180 thermostat value, try 170? The water expansion into the coolant recovery tank was just about what one would expect for a three gallon cooling system.
In another test, I removed the thermostat and replaced the water, topping it off to be sure the system was full. I ran car and found that it was difficult to get the temperature gauge over about 150 F partially due to a low outside air temperature. During this time the Motometer was generally in the 165 F range, but would vary a fair amount as water was expelled (going up to about 170 F) and returned to the system (dropping to 155 F). The lack of the restriction of the thermostat did not cause the water pump to eject extra water via the overflow tube. The expansion appeared to be normal for the temperature of the engine.
There are other tests that I would like to do, but this information should be useful to other Model T and A Ford owners.
At Shootout II, I had just replaced the water pump in my 1925 Touring. This car also had just accumulated some rust from an experiment with a standard high head that, unknown to me, was loaded with rust. The water pump was very tight and was able to produce a fair head of water. The car did not have a thermostat installed at that time. I experienced water loss and boiling on just about any sort of a hill. The water pump was pushing water into the top radiator tank and out the overflow tube. Backflushing the radiator and installing a thermostat pretty well cured the problem. The thermostat provides a controlled flow resistance that seems to aid in cooling.