I sort of agree with Steve here, and sort of don't. Focusing on the ICC regs (and ARA car classes) that bracket the presumptive build date (1922) of these NTCX TMI cars, I would offer the following:
1. Not all TMI cars were Class IV (ICC 104) cars; there were insulated Class III/ICC 103 cars as well. I have not discerned any way to tell them apart unless the photo is sharp enough to be able to read the ARA/ICC car class in the stencil to the right of center on the car side.
2. Regardless of car class, a great many insulated cars also had steam heater coils, suggesting that at least their primary purpose was hauling viscous commodities rather than highly volatile inflammables.
3. Also regardless of class, TMI cars were not that numerous, at least in the 1920s. In my 1930 ORER, TMI cars comprise <3% of the UTLX fleet, and about 5% of the Sinclair fleet.
4. In the 1920 ARA tank cars specs, the only example commodity mentioned in connection with the Class IV cars is casinghead gasoline. Note that Class 5 cars were also coming on-line at this time for specific (and dangerous) cargoes such as liquid chlorine and sulfur dioxide. Later on, ethyl chloride starts to appears as a commodity mentioned in the context of Class IV cars.
5. By 1923, the ICC had very detailed specifications for calculating the needed "outage" for volatile inflammables, including methanol, ethanol, acetone, and gasoline/naptha (the last broken down by density using the API scale). These could all be carried in Class III/ICC 103 cars as long as sufficient headspace was provided to allow for expansion. That outage volume depended on the loading temperature and the coefficient of expansion of the particular commodity.
6. That expansion volume might be provided by the dome alone but, if inadequate, the shell had to be filled less than level full. Each car design had an outage table that showed how much additional head space was gained for each inch below level-full.
7. For convenience, many owners started buying cars with domes greater than the 2.0% required for all Class III cars and, in my experience, domes right at 2% are more the exception than the rule. Cars built in 1920s routinely had dome volumes of 2.3 to 3.0%, and 3.5+% cars are easy to find.
Bottom line: the use of Class IV cars, while certainly permissible, wasn't necessary for routine hauling of refined gasoline, regardless of season (ANAICT). It might have been more common with particularly "light" blends or components because of their greater volatility, but that's a guess.
PS: The primary difference between the Class III and IV cars that jumps out at me is the required pressure test for the tank: 60 psi for Class III versus 75 for Class IV. The safety valves were to be set at the same 25 +/- 3 psi.
That's my take with my 1920s perspective -- other viewpoints and information are of course welcome.
Swall Meadows, CA