An excerpt from a book ....... While wrestling with this incubator problem a wonderful discovery was made. This was that a regulator should be attached to the outside of the machine as well as the inside, in order to anticipate any change of temperature on the inside. This outside regulator works in conjunction with the one on the inside and the two keep an absolutely even temperature. You see, all the other inventors of incubators have been trying to regulate their machines from the inside and anyone can see that it is the outside temperature that causes variations. So I put my regulator on the outside. The inside temperature of a machine does not change as the outside temperature changes. If the machine is to be regulated from a thermostat on the outside, then the temperature must vary before the regulator can work. The inside of a machine is a poor place for a thermostat, for it cannot work with accuracy. Anyone can see that the inside temperature only changes with that outside. I also have a thermostat inside to take care of the animal heat that arises from the eggs as incubation proceeds. This is connected with the one on the outside and the two work in conjunction on the flame at the lamp. If the outside temperature lowers the least the flame is increased to take care of it. If the inside temperature tends to raise from the animal heat the flame is lowered accordingly. Only juet so much flame as is needed is used. The gates on the flame open and shut with the variations of temperature and can never clog up as those sleeve devices sometimes do. These gates tend to create a gas from the oil so that the wick is not consumed and complete combustion and a very hot flame are secured. So absolutely delicate is this regulating device that when once set the thermometer can be taken out and the hatch run through successfully. When the doors of the machine are opened to turn the eggs the flame immediately comes up to take care of this variation so that in a very few minutes after eggs are turned the temperature is back to normal. Any thoughts on this???? I thought this was interesting for those of us that build our own incubators. How would you even go about doing such a thing? How exactly would it work??