The rifles were first tested in the early 1920s in rifle and light machine gun forms. While it failed to beat the Garand M1921 in early tests, it showed enough promise that 20 examples were ordered for trials in 1924.
The testing was not positive: the Thompson Autorifle was long, heavy, poorly-balanced, difficult to disassemble, required oiled cartridges to function at all, and extracted under extremely high pressure. One test report concluded with "Unless radical and unforeseen improvements are made in the guns submitted to the Board, further testing of these guns is not warranted."
It continued to be entered into trials until 1929, when one of the prototypes destroyed its bolt after firing 1,106 rounds during an endurance test. Thompson withdrew the Autorifle from the testing and never entered a new version.
The Thompson Autorifle, like the Blish pistol before it, was an attempt to incorporate a mechanism based on John Bell Blish's "principle of metallic adhesion" into a functional firearm.
Conceptually, the mechanism was supposed to be delayed by the adhesion under pressure of a steel front bolt section and a bronze rear section. According to Blish's principle these two dissimilar metals would solidly lock together until the chamber pressure dropped, whereupon the two would be able to slide against each other and the bolt would open due to normal blowback operating principles.
In practice, what happened was that the breech unlocked the instant there was any pressure in the barrel, with the main force resisting motion being the cam path of the bolt handle. This placed the forward section of the bolt group at a severe mechanical disadvantage, creating a friction-delayed blowback action. Because the cartridges had to cope with being pulled out of battery the instant they detonated, the Autorifle required an integral oil pad.
The enormous operating pressures placed on the Blish Lock resulted in an incredibly violent extraction, with the gun throwing cartridges over 30 feet, with such velocity that some were reportedly found with their necks embedded in solid wood. A series of locking lugs in the bolt were supposed to mitigate this, but achieved very little in that regard as their design assumed the Blish principle would be in effect.
Variants were built with both five-round fixed magazines fed with stripper clips, and 20-round detachable magazines. In the prototypes, the latter were Browning Automatic Rifle magazines with metal pads tack-welded to their sides for fit. Oiling pads were added to the sides of the magazine well to ensure cartridges were sufficiently lubricated before entering the action.