In 1809, three years after the review of the Pyrolophore, Lazare reports on another fire machine, that of Cagnard de Latour.10 This ingenious device embodies the principle of the modern gas turbine cycle that Sadi will describe in idealized form in his memoir.11 Cool air is brought in by an Archimedean screw (compressor) at the bottom of the hot water vat (heat exchanger, hot source). The air forms bubbles that rise through the hot water receiving heat as they rise. In their upward movement, the bubbles are caught in the inverted buckets of a paddle wheel (turbine) wholly immersed in hot water and make it turn. The expansion of the bubbles has a double cause: the drop in pressure as they rise and the transfer of heat from the hot water to the air in the bubble. The result is an excess of power on the buckets of the paddle wheel that more than compensate for the power needed to operate the Archimedean screw. This excess power is then used for activating another paddle wheel propelling the boat in which the engine is fitted.

The report claimed the excess power to be five times the original power, an overly optimistic appraisal, no doubt. The Cagnard idea provoked a sustained interest. Ten years later, in 1819, Nicolas Clement, a professor at the Conservatoire National des Arts et Métiers, together with Désormes, his friend, associate, and father-in-law, from whom he will even later borrow the name and be known as Clément-Désormes , present a paper12 before the Science Academy in Paris on the determination of the work obtainable during the expansion stroke in an ideal heat engine from various substances under different conditions of pressure and temperature. Their solution was based on an idealized Cagnard engine where the mechanical effect, or work, developed could be evaluated simply by measuring the water outflow from a column at the base of which a heated working substance was introduced and left rising in bubbles. In this case, however, the bubbles expanded practically in adiabatic mode as they rose by buoyancy through the water.
Reconstruction of the astronomical clock of Giovanni de Dondi of Padua. The weights activatng the clock have to be brought back regularly at thir initial position to perpetuate the cycle.