V grateful for Roger Todd's help in sorting this out.

The Perinelle-Dumay was designed by Louis Paul André de Perrinelle-Dumay, the captain of a frigate in the Fench Navy who transferred to the tanks and commanded a Groupement of Saint-Chamonds. The design perhaps reveals his maritime background (and the fact that there was some naval influence on Fench design as well as on the British). It's clearly from the Sueter/Crompton school. It was proposed in Feb 1918, so it is a WWI vehicle. However, he continued to tinker with it for some time after the War, apparently until 1921. What seems to have caused the confusion is that he wrote a book, Les Chars de Combat en 1933, and also made further changes to the design. It looks as if the book was a contemporary review of tank development, tactical theory, and so on, something quite a few writers were doing at the time. The sketch of the machine is from the 1933 book, and it seems no drawings of earlier versions are known. He died two years later.

By "floating tank," as it's described in Steelmasters, they mean just that - a watertight vehicle that could swim, presumably propelled by its tracks. The telescopic, hydraulic legs were nothing to do with the swimming or anything operational; they were to raise the vehicle off the ground for cleaning and maintenance of the tracks. I thought they might be to stabilize it while the guns were in action, but it seems not.

The 1933 changes included replacing the track under the prow (like the Saint-Chamond experiment) with an avant-bec much like the Schneider's. Sounds like a retrograde step, but P-M thought it would improve trench-crossing. A further mod was to replace petrol engines with gas-fired motors powered by coal or charcoal. The captain evidently continued to develop his idea long after the point where it clearly wasn't going to happen.

Details: Length 19.7 metres; h 3.7m; w 3.0m. Draught in water: 2.5m. Weight: 84 tonnes.

Armament: Two 65mm naval guns fore, one 47mm aft, 5 machine guns. Max armour: 6-8mm.

The article points out that M. P-D did not address the question of turning the vehicle. There is certainly no obvious way it could be done. Transport would have been by rail, in the manner of the FCM 2C. The vehicle was described as having roughly the dimensions of a railway restaurant car.