An Entertaining & Informative Vintage Automobile Internet Magazine

Charles F. Kettering with GM's Rear-Engined Experimental 1930s Economy Car Martia I and II and AD-800

GM’s Rear-Engined Experimental 1930s Economy Cars

By Michael Lamm:  From 1934 through 1938, General Motors built three small, rear-engined, experimental economy cars. This was in response to similar projects at Ford and Chrysler. All three manufacturers knew what the other two were doing, and while no American automaker felt serious about producing such cars, no one wanted to be left behind.

GM’s effort began in 1929 when Charles F. Kettering code-named the project “Research Light Car Body.” More than a body, though, it soon became an entire car, including an all-new and highly unusual engine.

  • In the lead photo Charles F. Kettering, head of GM Research Labs, inspects Martia I during a shakedown run near Gaylord, Mich., in 1935. A radical 2-stroke X-4 engine stood at rear of car, and it’s doubtful that GM Art & Colour had anything to do with the body design. This was strictly an experimental engineering project.

Kettering set up a formal Automotive Design Department in 1933 and put engineer O.E. (Olle) Schjolin in charge of it. Schjolin was assisted by Carl A. Lindblom and a small, dedicated staff. Together, they had the first car, which Schjolin named Martia I, running and ready for testing in February 1934. His team would soon build a second version, Martia II, and then a final one, named AD-800. All three used fully unitized bodies – fairly unusual for that day – and Dubonnet suspension units at all four corners. But what really set these vehicles apart were their engines – as strange as they were innovative.


  •         Both Martias had twin radiators up front. Marita I also had a third radiator above the engine.


  •                                 Both Martias and the AD-800 seated four and delivered up to 45 mpg.

Not only were they X-4s, placed in the rear of each vehicle, but they were two-stroke and used a Roots-type supercharger. Each arm of the X-4 contained two parallel cylinders and pistons, a U-shaped combustion chamber and two separate connecting rods that attached to a single throw of the crankshaft. Technically, then, these were eight-cylinder engines, but the twin pistons were offset by 15 degrees. One piston took care of intake and the other performed exhaust functions. Schjolin and Lindblom found that offsetting the pistons gave a smoother idle and got rid of the typical two-stroke’s poppety-popping on deceleration. Also, the 15-degree phasing in the side-by-side bores allowed the exhaust to open far ahead of the intake, yet both ports closed at the same time. Thus at top dead center, compression was held for a longer time.


  • Two-stroke engines used a Roots-type supercharger to extract the exhaust, not to boost intake volume. Twin pistons in each arm of the X made this an 8-cylinder engine. Left piston was for intake, right for exhaust, and they were phased 15 degrees. The rotating and reciprocating parts can be viewed (above and below) at two different crankshaft positions. 

The supercharger acted not as a blower but as a scavenger pump so that, rather than forcing the fuel mixture into each working cylinder, it removed spent gases after combustion had taken place. And because two-strokes don’t burn all the fuel mixture totally, part of the exhaust stream was rerouted around the supercharger and back into the intake. That, according to Schjolin, made these two-strokes 15-20% more efficient than comparable-displacement four-strokes. In these three experimental cars, which weighed 1844-2100 pounds, Schjolin’s team recorded up to 45 mpg in typical driving. Oil consumption was roughly 3000 miles per quart, and oil entered each engine through an external reservoir rather than being mixed with the gasoline.


The two Martia engines displaced 130 cubic inches, the AD-800 160. One of the great drawbacks of these powerplants was their exhaust odor at light loads. George Hallett, one of the engineers I talked to about this project back in 1971, said the exhaust smelled “…like a combination of tear gas and skunk.” Another engineer, Marion Fast, told me, “…these were the stinkiest engines ever built – absolutely impossible from an emissions standpoint, even in those days.” But by force-scavenging both the 130 and the 160 with the same-sized pump (supercharger), the team found that the larger engine had less blow-through and a less obnoxious exhaust smell.


  • Some of the exhaust stream was rerouted around the blower and back into the intake. Unlike the previous cutaway, this schematic shows exhaust gases being drawn from the crankcase.

All three engines were water-cooled. The first car had triple radiators, two in front and one above the engine. The Martia II and AD-800 used just the forward radiators. And the AD-800’s engine stood at 90 degrees to the rear axle, while the Martia I and II placed their crankshaft in line with the axle. All three cars used three-speed transaxles specifically engineered for these cars.

The light-car research experiment must have been quite expensive, even by General Motors standards. The project ended when Schjolin and Lindblom left GM to go to Volvo in 1938. Lindblom told me years later that the postwar Volvo PV-444 looked like a 1941 Ford because he and Schjolin wanted to cash in on Ford’s popularity.

Meanwhile, during WW-II, another team of engineers installed variants of the two-stroke X-4 in otherwise stock Chevrolets and Oldsmobiles. These engines performed well, with very little maintenance, as did a similar GM X-4 installed in a small Cessna airplane. Copyright © 2015 Michael Lamm.

  • Experimental 2-stroke engines powered not only the three prewar compact cars but also an Oldsmobile and a Chevrolet during WW-II and later a Cessna airplane. All engines worked well, but the main objection was their exhaust odor, which one engineer described as a combination of tear gas and skunk.


18 responses to “GM’s Rear-Engined Experimental 1930s Economy Cars

  1. Hi David
    These engines were clearly inspired by the extraordinary Trojan from 1922. Have a look at this page for more information:
    www dot practicalmachinist dot com/vb/antique-machinery-history/ot-unusual-engine-1-trojan-115778/
    All the best, Mark Walker

  2. Here is a short (3:23) Vimeo video: https colon //vimeo dot com/74395593 that shows a lot of information.
    It complements the Trojan info provided by Mark Walker (@25 sec) with the hand cranking of a cutaway engine clearly showing the biased stroke of the two pistons on one crank throw. The latter segment of the video is, “Now what is the Trojan” a contemporary silent ( 2+ mins – w/subtitles) showing the Trojan’s unique capabilities.

  3. The last photo is upside down, since sediment bowls don’t point up and collect moisture and debris from the fuel.

    • Scott, Fuel pumps have been made with the sediment bowl on the top.

      This photo appears to show a motor highly detailed for a display and the tubing that is part of its stand can be seen below it.

  4. I’m familiar with the rear engine Ford and the Chrysler Star Cars, but when did Chrysler try a rear engine layout?

  5. Reo was another of the manufacturers that toyed with a streamlined, rear-engine car during the early 1930s. Reo chief engineer H.T. Thomas completed the little prototype, which he called the Doodlebug, in the summer of 1933. Its 25-horsepower Hercules engine was not as radical as the Martias, but it was cooled by radiators mounted above each rear wheel. Air flowed into a set of louvers above the rear fenders and exited through vents on the engine cover. The lightweight car gave peppy performance of up to 60 miles per hour. Unfortunately, Reo was unable to finance production during the Depression and the company stopped building cars two years later to concentrate solely on trucks.

    • One other interesting tidbit about the Marti cars: They were intended to sell in the economy class at under $500 per car. In order to keep tooling and production costs to a minimum, only two door stampings were needed for the four-door bodies. The left front door interchanged with the right rear, and the right front door interchanged with the left rear.

    • That’s always my thought after reading a story like this; ‘I wonder if anyone squirreled one away somewhere? The engine alone is a fascinating bit of engineering.

  6. I’m fortunate to be the current caretaker of another X engine prototype from the early thirties, the Hoffman. It could be the only remaining example of this small but interesting genre. The Hoffman has a rear/mid mounted X-8 of approximately 260 cubic inch displacement. Michael also wrote about this car in the early seventies while it was still in the Brooks Stevens collection. It was a mystery back then and, even though we have discovered more of the story, there is still much to be learned. Thank you for posting this article.

  7. THE “TWINGLE” engine has been used for many years on MOTORCYCLES, Garelli starting in 1912, (German) Triumph, Steyr-Puch (SEARS), & others, its advantage being a combination of 2-stroke & 4 stroke technology.
    Most of this was well established by the time of any 20’s 30’s American experiments.
    The first car LOOKS LIKE DR. Ferdinand PORSCHE “written all over” its body & mechanicals (Like the wheels)
    Edwin – 30 –

  8. This design reminded me instantly of the pre-war DKW racing motorcycles. From 1931 Zoller had been constantly improving his ideas of 2-stroke “twinkle” engines using an offset second rod and piston as the transfer piston – in effect a supercharger. As with this Martia design one effect was to allow the exhaust port timing to be designed for better scavenging as well as reducing loss of fresh gas to exhaust on the upstroke. Power improved dramatically, eventually achieving 100 bhp per litre (from a 250cc engine) in 1937. But so too did the noise increase, and it was said during the Isle of Man TT week in June you could hear the DKW on “full noise” from the English coast, 60 miles away!

    But the Martia intrigues on a number of grounds. First the Roots “sucker”, instead of a blower. Not sure what the efficiency would be where the pressure at the inlet varies from +ve to -ve with exhaust cycle . Perhaps the 4 cylinder banks even it out? Second, recycling spent exhaust would do nothing to improve volumetric efficiency, so that claims of 45 mpg in a vehicle weighing not far short of a ton are difficult to credit.

    Nevertheless, an intriguing and courageous design adventure!

  9. It seems to me that the supercharger would be required for cylinder filling as crankcase compression common to two strokes would not be possible( I think) with a common crankcase. I cannot envision a workable crankcase for intake compression with an X4 configuration. I am probably wrong on that . I do own a rare MZ 2 stroke flat twin and that works well.Previous comments are correct about the long history of split single cylinders or twingles on motorcycles. I own several right now, I believe that the Iso version of the Isetta( the original) also used a split single.

  10. On another note as I look at the diagrams I see what looks like a conventional supercharged 2 stroke with the blower feeding the fuel and air mixture into either the intake ports or the crankcase. I see no evidence of the supercharger ‘drawing exhaust gases from the crankcase’ or removing spent gases after combustion as stated in the article. You would never pull exhaust thru the crankcase as the caption of the photo says. What about bearings etc.? That picture actually seems to show forced induction INTO the crankcase as is common on superhcarged two strokes, Maybe the reference to the supercharger sucking rather than blowing is referring to the fact that it is sucking thru the carb not blowing thru from upstream

  11. What?! Why would you have the scavenger pump on the exhaust side?! That would be quite dumb!!! Hopefully more correct info(It says nothing about a blower in the exhaust!!!) for people that want to read about this motor : www dot hemmings dot com/blog/2007/06/17/sia-flashback-gms-x-cars/

Leave a Reply to Steve Denner Cancel reply

Your email address will not be published. Required fields are marked *