There is a special satisfaction reserved for the cynic in the study of extinct animals. It must tickle his fancy no end to think of nature, having achieved the diplodocus, finding that the largest creature in the world was no good anyhow. (I believe that someone has calculated that if another animal bit the diplodocus' tail, it was a couple of minutes before the head knew anything about it. But perhaps it wasn't the diplodocus at all and anyhow I am getting off the point.) Nature, having discovered her mistake, had to go all the way back to the diplodocus' great-grandfather and start altering his other descendants before she got a nice rhinoceros, or a kangaroo, or whatever it was she wanted.

All this is by way of an analogy. It is fairly easy to trace the descent of your M.G. Magnette from M. Levassor's original car ; but in the course of those forty years, there have been all sorts of original people who have strayed off the beaten track, designing the most ingenious machines, which in the , end have proved to be merely diplocloci. But in the case of cars, the cynic must beware. Designs which appear to be as extinct as the diplodocus have a habit of reviving. After all, Georges and Maurice Sizaire were exponents of knee-action thirty years ago.

Opposed Pistons

Vial never find engines with opposed pistons nowadays, but years ago an enormous amount of energy was devoted to designs based on this principle. Many engineers undoubtedly slummed it for fear of suspicion that they were rembracing the fallacy of " power lost against the cylinder head." But others were bolder—anyone who has felt the vibrations of an early engine can guess their motive. Koch and Riancey in France, Gustav Mees in Germany, were all building horizontal engines with opposed pistons in the 'nineties. But it was those famous engineers Messieurs Gobron and Brillie who brought the principle to the world's notice by constructing racing engines embodying it. I well remember at school being enormously impressed with a car of this marque belonging to the parents of another boy. The latter informed me that the engine was far better than those of conventional design, but that its commercial success had been killed by inimical racing rules. I expect he was right ; but at first all went well.

at first all went well. The Gobron-Brillie Engine

The original Gobron-Brillie engine, designed towards the end of last century, had two vertical cylinders in which fact there was nothing remarkable, but in each cylinder there, were two pistons. The lower pair were both coupled to the same crank, while the inverted pistons working in the top halves of the cylinders were connected to a cross-head. At either end, this bar had a connecting rod of great length, which ran down by the side of its cylinder, and was coupled at its lower end to a crank at 180 degrees to the central one driven by the two lower pistons. The whole affair was balanced with immense care, but whether the

results in this respect were better than with a four-cylinder engine of the same date, I do not know. The automatic inlet and the exhaust valve were arranged in an ordinary valve pocket' situated about half-way up the cylinder. At any rate the design, one must admit, was one of great ingenuity and deserved that more should be heard of it. Quite a lot was!

Racing Experiences

The marque made its racing debut in the Paris-Toulouse-Paris race which, with a nett racing distance of 837 miles, was the great event of 1900. A single car started, in the heavy class, and was driven by one of its designees, M. Brillie. It had two-cylinder engine rated at 18 h.p., but, unfortunately, I have not been able to find its dimensions. It was, however, the heaviest car in the race, weighing 1,400 kilos. while the winning 24 h.p. Mors weighed but 1,200. The GobronBrillie certainly worked—for it finished the race ; but it does not seem to have been very fast, for Brillie only averaged 19.3 m.p.h. against the winner's 40.2

Perhaps MM. Gobron and Willie realised that the weight of their car militated against its performance. At any rate, the next year the three cars which they ran in the Paris-Berlin race were so much lighter that they qualified for the 650-kilo. " voitures legeres " class. The engine was still a twocylinder,. but this year was rated at only 8 h.p., the dimensions being, bore 110 -mm, stroke of each piston 100 mm, giving a capacity of 3,801 c.c. A slightly larger Gobron-Brillie engine, rated at 10 h.p. was used in the Manceenne cars, built by a firm which at this time concentrated on lorries fitted with this engine. One of these started in the light car class, but the other, which incidentally was driven by M. Brillie, weighed 725 kilos., and had to run with the big cars.

The, three Gobron-Brillies proper, driven by Roland, Dernier (who belied his name) and Rigolly all reached Berlin. The fastest one, driven by Roland, averaged 27.7 m.p.h. compared with the speed of 35.5 m.p.h. recorded by Giraud's Panhard et Levassor, which was the class winner. This Panhard, rated at 12 h.p. had a four-cylinder engine with a bore and stroke Of 90 by 130 mm. and a capacity of 3,296 c.c.. which was, therefore, smaller than that of the GobronBrillies. The two, however, were fairly comparable, in view of Panhard's reputation at the time, the opposed piston principle may be said to have acquitted itself adequately.

Later Races

The next year was the Circuit du Nord race, and three Gobron-Brillie light cars were in it. They were now called 16 h.p. but the dimensions seem to have been the same as in 1901. All the cars in this race had to use alcohol fuel, of which believe Gobron-Brillie had made a special

study ; but as at least some of the Panbards used it voluntarily in the ParisVienna Race later in the year, perhaps there is not much in that. At any rate Rigolly finished third in the light car class, averaging only 32.5 m.p.h. and being beaten only by a Darracq and a Panhard, which averaged 41.2 m.p.h. and 37.8 m.p.h:. respectively. By now, too, the Darracq had an engine of 120 by 130 mm. (5,618 c.c.) while the Panhard's was 100 by 130 mm. (4,084 c.c.). On this occasion, therefore, the Gobron-Brillie engine may be considered to have proved itself at least comparable to the best motors built on conventional lines.

Paris-Vienna later in the year again saw the Gobron-Brillie light cars and also a team of machines using the same engine, but built at the Nagaut works in Belgium. It was actually one of the latter that did best, but although it finished creditably, it was not in the first flight. The triumph of the opposed piston principle came in the Circuit des Ardennes at the end of the year. In that race, Rigolly, driving his Gobron-Brillie, finished first in the light car class at an average speed of 46.3 m.p.h., beating all that Panhard and Darracq could do agaipSt him, and all but four of the cars ie the heavy class. In the racing world, liabron-Brillie had arrived.

Eternal Youth

With this encouragement, the firm decided to go forward to greater things. The light cars were entered for ParisMadrid in 1903, but the interest had shifted from them, for Gobron-I3rillie had now built a team of racers to compete in the heavy, 1,000 kilos. class. The principle was extended to a four-cylinder, eight-piston engine. The bore was increased to 140 mm. and the piston strokes to 110 mm. each, which gives a capacity of 11,680 c.c. These new cars did not prove the leaders in the ParisMadrid, but all, three of them reached Bordeaux in creditable time, and that was no small achievement in the most famous race of all time. Again, greater success came in the Ardennes race, for although neither of the big cars which started succeeded ih finishing, Tavenaux on the 4-litre light car came in second in his class, being beaten only by a Darracq with a 7-1-litre engine. 1903 was practically the last year when any attention was paid to the light car class but in 1904, the same big GobronBrillies which had -run in the ParisMadrid started in the French eliminating trials for the Gordon Bennett team. None of them 'finished, but Rigolly continued to show that his 1903 racer possessed the secret of eternal youth. He appeared on it in the Circuit des Ardennes race later in 1904 and finished fourth, only 1.1 m.p.h. slower than the winner in a race of 367 miles. He started again in the Gordon Bennett eliminating race of 1905 and finished eighth, 3.8 m.p.h. slower than the winner. In the first Grand Prix in 1906, he was eleventh at the end of the first day, but fell out on the second, and started once again in the second Grand Prix in 1907 with his fouryear-old racer. Throughout this period,

he was able to compete on equal terms with the latest productions from other factories.

In 1908, the Grand Prix rules limited the bore of engines, and made no provision for these built on the GobronBrillie principle. This spelt the end of the firm's racing career, and perhaps, as my school friend claimed, of the design. But while it lasted, it was an interesting and not unsuccessful experiment.

The Arrol-Johnston

In the meantime, the development of the opposed piston principle was being developed in this country also. Here, its great protagonist was Arrol-Johnston, and it was brought very much to the public notice by the victory of one of these machines, designed by Mr. J. S. Napier, in the first Tourist Trophy race of 1905. In the case of the Arrol-Johnston, the engine was a 2-cylinder, but the cylinders were arranged horizontally, with the crankshaft below them. Each cylinder contained two pistons, but unlike the Gobron-Brilli6, neither was connected direct to the crankshaft. Instead, each had a connecting rod attached to the end of a rocking lever, the other end of which had another connecting rod all which drove the crankshaft. The cranks connected to the two pistons of the same cylinder were at 180 degrees to each other and at 90 degrees to the cranks of the other cylinder. The bore was 120 mm., combined stroke of the two pistons, 165 mm., giving a capacity of 3,733 c.c. The engine apparently reached a speed of 1,100 r.p.m., and was stated to develop 18 h.p. at 800 r.p.m. I do not know whether the peak of the power curve was there, or whether this statement was in deference to the R.A.C., which organised the race and which, by its formula, rated the engine at 18.05 h.p. The formula, incidentally, is based, as everyone knows, on the bore of the cylinder but, as perhaps everyone does not know, takes no account of the number of pistons in the cylinder. In this respect, incidentally, taxation does not follow the R.A.C. formula as the tax calculated on the bore is multiplied by the number of pistons, not by the number of cylinders. Perhaps, otherwise, opposed pistons would be the rage to-day. The Arrol-Johnston engine had automatic inlet valves and governed on the exhaust valves, which was useful with a

fuel COnsumption limit for the race. The engine, incidentally, was no novelty, having been used in the Arrol-Johnston dog-cart since 1900.

In The Tourist Trophy Of 1905

The race \vas run over four rounds of the Manx course on 14th September, 1905, the total distance being, to be exact, 208 miles 4 furlongs. Fuel was originally to be allowed at the rate of 25 m.p.g., but at the last moment the limit was altered to 22.54 m.p.g., a 'change which annoyed some of the entrants of the more economical cars. However, it made no difference to the winner of the race, as the Arrol-Johnston averaged over 25 m.p.g. Two of these cars actually. started, one driven by its designer, J. S. Napier, the other by E. J. C. Roberts. At the end of the first lap, Napier led from a field of 42 starters and he was never displaced. His average speed for the whole race was 33.9 m.p.h., which compared with a stated maximum speed of 45 m.p.h. Second place went to Percy Northey's Rolls-Royce, third to Norman Littlejohn Vinot et Deguingand and fourth to the second Arrol-Johnston.

From the point of view of the engine, it is interesting to compare the ArrolJohnston with the Rolls-Royce and the Vinot. The Rolls had a 4-cylinder engine of 100 by 1.27 mm. (3,990 c.c.), which was stated to develop 20 h.p. at 1,000 r.p.m., while the car's maximum speed was given at 45i m.p.h. The ArrolJohnston's win, even by less than two Minutes was thus highly creditable for the opposed piston principle, in view of the larger size of the engine of the Rolls which, incidentally, would not have finished if the fuel consumption limit had been left at 25 m.p.g. Nor for that matter would the Vinot, although it had a rather smaller 4-cylinder engine of 90 by 130 Mm. bore and stroke (3,308 c.c.), and was only stated to develop 14 h.p. at 900 r.p.m. The opposed piston principle was

applied to Arrol Johnstons which appeared in Subsequent Tourist Trophy races, but the same success did not attend them as in the first race of the series.

The Principle Re-emerges

The idea apparently passed into oblivion. The intensive study of aero engines during the war, however, in

eluded the problem of filling the cylinders with gas at high altitudes, and focused attention on the supercharger with which Marc Brikigt of Hispano-Suiza fame and others had been toying before the outbreak of hostilities, as a result of the capacity limit used for the Coupe de l'Auto and Grand Prix races. After the war Mercedes used their aero engine experience to apply superchargers to the racing car in a thoroughly secretive manner, but Fiat definitely placed the blower on the racing map by a spectacular failure in the. 1923 Grand Prix. After a year or two, everyone was using superchargers, superchargers made them think of two-strokes, and two-strokes made them think of—opposed pistons. For a moment it looked as if Gobron-Brilli6 and Arrol-Johnston had not been exploring a blind alley twenty years before, after all. Fiat .apparently experimented with An opposed-piston engine in 1927, but in the end abandoned it in favour of a 12cylinder with two rows of parallel cylinders and the crankshafts geared together. In France, however, experimenters were also at Work. In 19.27, the Cozette Carburetter Company built a vertical 4-cylinder, opposed-piston, supercharged 2-stroke engine, and a M. CauSan a similar engine, except that it was horizontal, which he operated successfully in a boat. Both the vertical principle of Gobron-Brillie and the horizontal design of .Arrol-Johnston was thus represented. The two more recent engines, however, differed from their prototypes in that they had two crankshafts. The Cozette engine which was intended to be applied to racing cars, had a bore of 56 ram. and a combined stroke of 110 mm., bringing the capacity at 1,084 c.c. just within the 1,100 c.c. class. The two crankshafts were united at the rear of the engine by spur-gearing with an intermediate pinion. This engine is believed to have run at very high speed, and to have developed 100 h.p. At that moment, however, the quest for volumetric efficiency in racing engines was brought to an end for the time being by the passive resistance of manufacturers who refused to build cars for races run under a capacity limit. So perhaps it was a blind alley after all. On the other hand, one of these days a fresh turning may show that it does lead somewhere.