Racing car evolution part III: 1919-1921

We have great pleasure in presenting herewith a continuation of the "Evolution of the Racing Car" series, this time contributed by Laurence Pomeroy, Technical Editor of "The Motor" and author of the "Milestones of Speed" articles now appearing in that journal. We have decided to present Mr. Pomeroy's article in three parts, the first instalment of which covers the period 1919 to 1921. The text and illustrations are published by kind permission of Temple Press, Ltd., and remain their copyright. – Ed.

The two articles written by Messrs. Clutton and Heal which have preceded the period with which I am about to deal have set a very high standard of ability and literary skill, but although similar in these qualities they show a very divergent approach to the subject. Mr. Clutton cites all the facts relevant to his conclusions; Mr. Heal cites all the facts he can assemble and lets the reader draw his own verdict from the evidence so presented. I propose to combine both methods and will, necessarily, be more long-winded in the process. This article will be divided into three parts, of which this is the first. Each one will be sub-divided into two sections, in which I will summarise the racing history of the period dealt with, together with the specifications of the cars running, and will then separately consider the engineering aspects of their design and performance. As with Messrs. Clutton and Heal, I shall restrict my interest to road racing cars of Grand Prix formula and (or) these of 1 1/2-litres capacity.

With this by way of preface let us address ourselves to our muttons.

Between 1916 and 1919 racing virtually ceased all over the world, and it was natural that when it was revived in the latter year the countries taking an interest should be those which had been the least involved in the recent war, viz., the U.S.A. and Italy.

In France, the Manufacturers' Association decided to sanction no competitions. A similar policy was followed by the R.A.C. But in the U.S.A. the Indianapolis Race was organised to take place on May 31st, with a capacity limit of 300 cub. in., which is equivalent to 4,917 c.c. For this the large entry of 43 vehicles was received – 31 of American design and 12 representing Europe. Of these 12, four were privately entered Peugeots, all of which had been manufactured in 1914 or previously. There were three 1914 cars, two ex-Grand Prix 4 1/2-litre models and one 2 1/2-litre, which had been built for a race which did not take place owing to the outbreak of war.

There were also entered two six-cylinder Sunbeams which had been designed and made in the factory at Wolverhampton during the course of the war under the guise of being experimental aviation work. The subterfuges by which these jobs were covered away when Air Ministry inspectors went round the works is a story in itself, but, alas! all this effort was in vain as the cars were withdrawn; some say because the engines were found to have a capacity of 4,924 c.c. (and were thus over size), others because Resta's report on the torsional oscillations in the crank at 2,000 r.p.m. was decisive.

Belief in the latter version is reinforced by measurements made later by Brooklands scrutineers, which gave the capacity as 4,914 c.c. Would the pistons and blocks have been changed after the cars came back to England? I think not.

The remaining four European cars may be justly considered the sensation of the race, for they introduced a new name to automobilism – Ballot. They were designed by Ernest Henri, who had been responsible for the highly successful Peugeots of 1912 to 1914, and were built in the amazingly short time of 101 days. They had eight-cylinder engines, made of two blocks of four, bore and stroke 74 x 104 mm.

Ernest Ballot was the head of a company which had been making stationary engines during the war and he decided to make an entry on Christmas Eve, 1918. The cars had to leave Paris not later than April 26th, so there remained but 120 days in which to design and test the vehicles. M. Henri was retained as designer and, naturally, the general layout of the cars bore a certain resemblance to the four-cylinder Peugeot he had made previously. The decision to use a straight-eight layout was undoubtedly the result of being brought into contact, during the war, with a Bugatti-designed straight-eight engine for aviation use.

The novelty of this construction and the renown of its designer naturally led to the highest expectations in regard to performance. In this the cars were not disappointing, and during practice the Ballots definitely proved themselves the fastest cars on the track. Unfortunately, they also proved to the team manager's satisfaction that the gear ratio was rather too high and, as they had no spare crown wheels and pinions, the wheels were changed to American types with straight-sided tyres. In the race this proved a disaster; trouble started within the first 100 miles and persisted throughout the race. On two cars the wheels broke and the others were so much delayed that fourth and eleventh were the highest places they could occupy. The winner, interestingly enough, was Henri's ex-1914 Grand Prix model Peugeot, which, driven by Wilcox, averaged 88.06 m.p.h. and 11 m.p.g., in comparison with the best Ballot speed of 84.44 m.p.h. A Durant was second at 87.1 m.p.h., the general design of this car being a cleaned-up and simplified version of the winning Peugeot, with cylinders enlarged to 98 x 162 mm. to give a capacity of 4.90 litres and with only a single camshaft.

The most original American car was a twelve-cylinder (60 x 114.5 mm.) Packard, driven by Ralph de Palma. This car led up to half distance, then dropped back to finish sixth. Another sufferer from wheel trouble was André Boillot on the small 2 1/2-litre Peugeot, which despite its small size was running third with only 10 laps to complete. Boillot, however, had a great compensation in the next and only other event of 1919, viz., the Targa Florio, run on November 23rd.

The 21 entries included Alfa, making their bow in the competition world; three Fiats of the 1914 Grand Prix type; two Nazzaros and an Aquila-Italiana of similar specification, and a miscellany of lesser-known types.

The battle pre- versus post-war designs was maintained by the entry of one Ballot, driven by René Thomas, but despite having double the cylinder capacity of the small Peugeot, the larger car never managed to get to the lead, and Boillot ran persistently at the head of affairs.

On the last lap Thomas went out with a broken differential and the remaining challenger to the leader was an Itala. The finish was dramatic in the extreme. Boillot throughout the race had been making up for lack of engine size by the most fantastic exhibition of driving without regard to personal safety. He had been six times off the road; his refuelling consisted of picking up a can when on the move and filling the tank without stopping and, finally, as he arrived at the finish he found the crowd had overflowed on to the road. He jammed on the brakes, spun the car round, and crashed 10 yards before the finishing line. The excitable Sicilian crowd endeavoured to right the car, ignorant of the fact that they would thus ensure its disqualification. Boillot and his mechanic forcibly defended themselves, drove back 30 yards, turned round, and thus won. Utterly exhausted, Boillot cried, "C'est pour la France," and collapsed over the wheel.

Strangely enough, after this event the 2 1/2-litre Peugeot seemed to disappear from the racing world, but several of the Ballots, fortunately, were brought to Brooklands and in their original form and tune they were able to lap the Track at over 112 m.p.h. (Chassagne did 112.17 m.p.h. on September 25th, 1920). It is reasonable to assume that their maximum speed was of the order of 115-118 m.p.h.

In 1920 the 3-litre capacity limit came in for the first time in the Indianapolis Race and the eight-cylinder Henri-type engine was widely copied. The Ballot was produced in a scaled-down version, with cylinders 65 x 112 nun., and in the U.S.A. Duesenberg followed the practice. This was a natural development, for during the war the Duesenberg Co. had been engaged on developing those same eight-cylinder-in-line Bugatti aviation engines which had provided Henri with his inspiration.

The cylinder proportions chosen by the American designer were slightly different., viz., 63.5 x 117 mm., and whereas the Ballots followed Henri's invariable practice of having four-valve heads, the Duesenberg had three valves per cylinder. Further, the valves were operated by one camshaft only, as compared to the Ballots' two. Easily the record for valves was broken by Peugeot, who decided to build some new cars of their own, the design of which was more noted for its novelty than its success. The four-cylinder engines were 80 x 149 mm. and had five valves per cylinder with two sparking plugs, also put into the top of the cylinder, cooling, presumably, being more by courtesy than H2O. They were, however, in advance of their time in embodying unit construction for engine and gearbox.

In the preliminary trials, R. de Palma, on one of the 3-litre Ballots, put up the best time, and during the race he led until only 35 miles remained. Unfortunately, his car caught fire and, although he restarted, it again caught fire and put him out of the race, so that the best Ballot performance was third position. The race was won by a Monroe which had an engine exceedingly like the 3-litre 78 x 156-mm. 1913 Peugeot engine and thus was another successful follower of a Henri formula. The dimensions had been slightly changed to 80 x 151.5 mm.

The three new Peugeots were never in the running; one withdrew at 300 miles with plug, magneto and carburetter trouble; another, two laps previously, with engine trouble, whilst André Boillot's car lasted only 16 laps before breaking its crankshaft. The winning speed was 88.5 m.p.h. – a slight advance on the previous year's time, despite a reduction in engine capacity of 33 per cent.

As a matter of interest, the race would quite possibly have been won by Chassagne on Ballot No. 3, for he was making up time rapidly at the end when he hit an outer retaining wall, bending his steering when having only 50 miles to go, an incident which put him down to seventh position.

Of the Duesenbergs, one finished third, despite a broken valve spring and having to change three tyres; another fourth and another sixth. Of the first half-dozen finishers, one only (admittedly the winner) had a four-cylinder engine – the others were all straight-eight types.

So far as international racing is concerned this concluded the lesson for 1920, for although the Targa Florio was enlivened by the entry of a touring Buick, it was won by a Nazzaro of no particular technical interest. The only other racing in 1920 of any significance was an event near Le Mans for cars with a capacity limit of 1,400 c.c. Three Bugattis of pre-1914 construction proved themselves easily superior to the rest of a rather poor field, and although one was disqualified and another had minor trouble, Friederich made no mistake about winning.

The 1919 Indianapolis Sunbeam did, however, obligingly appear at Brooklands and, with a lap speed of 108.98 m.p.h., disclosed itself as being about 4 m.p.h. slower than the 1920-type Ballot.

The 3-litre formula was continued for 1921, and this year saw the revival of racing of the Grand Prix order in Europe.

It was preceded by the 500 Miles Race at Indianapolis, where the entry list showed that the lesson of eight-cylinder superiority driven home in the previous year had been well digested. In 1919 the proportion of various numbers of cylinders amongst the entries, omitting fractions, was four-cylinder, 79%; eight-cylinder, 13%; six-cylinder, 6%; twelve-cylinder, 3%. In 1920 the proportions were four-cylinder, 75%; eight-cylinder, 25%; whilst in 1921 the eight-cylinders for the first time could show a clear majority, the proportions being eight-cylinder, 56%; four-cylinder, 28%; and six-cylinder, 16%.

In the race itself the Ballot, for the third year in succession, proved itself the fastest car on the track, and for the third year in succession failed to win. After covering 200 miles at 93.6 m.p.h., easily a record, de Palma broke a connecting rod and let a straight-eight Frontenac into the lead. Only the one Ballot was entered, but there was a good further representation from Europe in the shape of Peugeot, Sunbeam, and Talbot-Darracq. The two former were amongst the four-cylinder representatives, and although one driven by Wilcox lay third at 25 miles, it failed through a broken connecting rod.

The Sunbeams were highly interesting. They were the first completely post-war racing models for which Louis Coatalen was responsible, and there was no doubt that his team of designers modelled the engine fairly closely on the successful Ballot. The timing gears and camshaft layout were very similar; the bore and stroke (65 x 112 min.) identical. But four carburetters were fitted and owing to the use or plain in place of ball-bearing mains the cylinders were more widely spaced and the engine was longer. The Talbot-Darracq was the same design, but ran with a different radiator. At 350 miles two of these cars were lying third and fourth, but various incidents intervened, and eventually the best that could be managed was fifth place at 84 m.p.h., over 5 m.p.h. slower than the winner's average of 89.62 m.p.h. Seven Duesenbergs similar in design to those running the previous year were entered and finished second, sixth and eighth with four retirements. Again five out of the first six cars had eight-in-line engines, which type also contributed 90% of the finishers in this year.

In the French Grand Prix held later in the year these Duesenbergs reappeared to mark the first occasion for many years when an American team had competed in a European race.

The revival of this race, with the glamour attached by its traditions, attracted an excellent entry, other teams being provided by Sunbeam, Talbot and Talbot-Darracq (all of these were of identical design), also by Ballot and F.I.A.T.

The Fiats failed to start and the Coatalen cars only just got to the line after they had been formally withdrawn owing to lack of preparation. Their performance was impaired owing to this cause, and the race was fought out between the Duesenbergs and the Ballots. The former cars were handicapped for road racing by having only three speeds, but they had first-class brakes, hydraulically-operated, were in good tune, and had excellent powers of acceleration, the horsepower claimed being 75 at 2,000 r.p.m. and 92 at 3,000 r.p.m. By comparison the Ballot claimed 82 h.p. at 2,500 r.p.m. and a maximum of 107 h.p. at 3,300 r.p.m. as compared to the Duesenberg's maximum of 115 h.p. at 4,250 r.p.m. In consequence, although a Ballot led at one time, when it was put out by a leaky fuel tank, a Duesenberg finished first at an average of 71.8 m.p.h. De Palma on a Ballot finished second a quarter of an hour behind.

A second International Grand Prix was held later in the year on a circuit near Brescia. In this race the Fiats turned up to compete with the Ballots.

The Italian team had secured the services of Wagner, who bad finished third in the 1914 French Grand Prix at the wheel of a Mercédès, and had been lent to drive in the French Grand Prix in the Ballot team. This was a sound tactical move, because he was able to give the F.I.A.T. racing manager and engineer, Fornaca, inside information concerning the Ballot performance. Unfortunately for this pretty scheme, Ernest Ballot and Henri had improved the output on some later type engines, and these power units were secretly installed at the last moment after practice was over.

In the race two Ballots ran non-stop to finish first and second at 90 m.p.h., with Wagner a bad third at 86.1 m.p.h. Learning the cause of his defeat, the wretched driver went to carry the story to his chief. The latter, however, was in no kindly mood and said, "There is the door. Get the other side of it!" And that was the end of Wagner as a member of the F.I.A.T. team.

In Sicily the Targa Florio attracted better support, but here two of the 3-litre Fiats were rather surprisingly beaten by an old design from the same factory, consisting of a 1914 G.P. engine installed in a rear-brake-only chassis, which had been prepared for the 1915 Indianapolis event, but had been prevented from running therein. Thus, the Turin concern could scarcely congratulate themselves on the success of their new car, although its potential performance is indicated by a record lap by Bordino in the Brescia race at 96.31 m.p.h.

The design had been prepared by a young man called Bertarione and, although adhering to the straight-eight motif, it represented a considerable breakaway in constructional features, and differed in almost every material point from the racing cars which were then dominated by the Henri technique. Only two valves per cylinder were employed, the cylinder blocks were not made from iron castings, but from steel forgings welded together and enclosed by a sheet steel jacket, and finally both main and big-ends were of the roller-bearing type.

Henri's pre-1914 four-cylinder models had had ball-bearings for the crankshaft and plain big-ends, and so did his eight-cylinder types. The F.I.A.T. design embodied a startling novelty, for whereas in the past roller big-ends had been associated with built-up crankshafts, Bertarione daringly decided to split the big-end in the conventional fashion for a plain bearing and to use a divided roller cage.

This engine ran up to 4,600 r.p.m., and one might reasonably suppose that it developed 120 b.h.p. The cylinder dimensions were not quoted, but as the four-cylinder 1 1/2-litre supercharged Fiats of 1923 had the external appearance of being identical to the previous 3-litre models there was reason to suppose that the internal dimensions of these two engines were the same. If so, the bore and stroke were 65 x 112 mm.

In the Brescia race this car proved itself an average of about 10 secs. per lap faster than the Ballot, the best lap speeds being of the order of 96 and 93 m.p.h., respectively.

The last race of 1921 was an astonishing affair, of which the implications have never received their due. In October there was a 50 Mile Invitation Race at Indianapolis. The Duesenberg, with its laurels of the French Grand Prix still metaphorically draped around it, again met the Frontenac to which it had run second in the 500 Mile Race on the same track earlier in the year. Also competing was the 1914 4 1/2-litre Peugeot which surprised everyone by beating the 1921 cars fairly easily, averaging 97 m.p.h., whereas the best of the 3-litres was Murphy on the Duesenberg, who averaged 95 m.p.h. to finish 1 min. 33 secs. in arrear.

Completing our summary of motor racing history we must mention the Voiturette races of 1921, which were dominated completely by the four cylinder Talbot-Darracq cars; these had engines which were exactly half-size reproductions of the straight-eight models of the same concern. The only serious opposition to them came from some improved Bugatti models with cylinder dimensions 69 x 100 mm., 16 valves per cylinder and ball-bearing crankshaft, which were called the "Brescia type" in celebration of their win in the Light Car Class on this circuit, the Talbot-Darracqs not running in this event.

Now let us turn to a technical analysis of the results. First of all let us consider how the cars developed in respect of the maximum speed, for after all this is one of the most interesting points about a racing car.

If we take the 1914 G.P. Peugeot which won the first post-war event as a logical starting point we find we have a car capable of 110-112 m.p.h., with a probable engine output of 110-115 b.h.p. The 1919 Ballot cars were certainly faster than this and 116 m.p.h. was within their power. One might argue that this seems not remarkable in view of the larger engine capacity (4.9 litres in place of 4.5 litres), but, against this, Peugeot had a long tail and the Ballot a very square one made by a bolster tank. When the Ballot developed a "Peugeot-like" tail its speed rose to 119.75 m.p.h. Nevertheless, one is forced to admit that the difference in speed between these models is rather less than one would have supposed. When we come to the 3-litre cars some real advance begins to make itself felt. The old four-cylinder Henri-design Peugeot covered the kilometre at Brooklands at 105.81 m.p.h. and lapped at 101.64 m.p.h. The 1920 successor by the same designer was capable in its original form of lapping at 107.84 m.p.h., and hence must have had a maximum speed of over 112 m.p.h.

The winning speeds at Indianapolis are another indication of the comparative merit of various vehicles. The old four-cylinder 4 1/2-litre Peugeot put up its best performance in 1915 when it averaged 88.91 m.p.h. for the whole distance, and in practice lapped at just under 99 m.p.h.

In 1919 the 4.9-litre Ballot was able to lap at 105 m.p.h., and although in 1920 none of the 3-litre cars could approach this speed, in 1921 the Ballot was able to average nearly 93.4 m.p.h. quite comfortably, and but for its broken connecting rod would doubtless have maintained an average of over 90 m.p.h. to the finish.

As regards the Fiat the only timed performance is the 96 m.p.h. average for Brescia, but this speed on a road circuit indicates an abnormally high maximum. At that time the F.I.A.T. engineering staff said that these 3-litre models were too fast for any human being to control, and I think we may surmise that they could steam up to nearly 120 m.p.h. It is on record that in April, 1922, Bordino averaged 115 m.p.h. on one of these cars for 25 miles on the Beverly Hills Track in the U.S.A. It is thus fair to claim that the best specimens of the 1920 and 1921 3-litre models were between 6 and 14 m.p.h. faster than the best pre-war cars of equivalent capacity.

Were they faster than the Grand Prix cars of the 1914 formula? In view of the results of the 500 Mile Race at Indianapolis and the way in which Masetti won the 1922 Targa Florio at 39.2 m.p.h. on a 1914 Mercédès (i.e., 5 m.p.h. faster than the winning speed in 1919, when the 5-litre Ballot was competing), one hesitates to answer this question in the affirmative. There is, however, one other piece of evidence which is not commonly known. In 1921 Masetti, driving this same Mercédès over the Brescia circuit, won the Grand Prix des Gentlemen (the correct title, believe it or not) at 72m.p.h., which compares very unfavourably with the Ballot's speed of 90 m.p.h. Further, one must remember that the older cars had been the subject of a good deal of development and may not, necessarily, have been so fast in their original tune.

It would not, therefore, be unfair to sum up a difficult problem by saying there was little to choose between the 1914 4 1/2-litre cars and 1921 3-litre models.

From this we may argue that power per litre must have been increased by at least 33%. This brings us to the question of engine performance, a thing which can be analysed under several different headings.

There is first this matter of power per litre, in which one naturally expects smaller engines with large numbers of cylinders to show an advantage compared to units having cylinders of large capacity. From an efficiency standpoint two good guides are power developed per square inch of piston area and the mean effective pressure realised at peak r.p.m. From a mechanical angle the piston speed figure is of interest, whilst the skill of the designer in getting good breathing is indicated by the power developed per square inch of inlet valve area. Fortunately, I have sufficient data to provide an answer on all these items for the following cars:– 1914 4 1/2-litre G.P. Peugeot, 1921 Ballot, 1921 Duesenberg and 1921 Fiat. It is, therefore, reproduced in a table.

A study of these figures will indicate some interesting facts. It will be seen in the first place that the post-war designs represented an increase in m.e.p. of less than 15% above the best type of Grand Prix car in the pre-war period, but power output per litre went up by around 55%.

Rather strangely the piston speed figure for the old Peugeot was only exceeded by the all-roller-bearing Fiat, and was not really approached by various other designs, so that the power per square inch of, piston area went up by only 10%. When we come to consider the breathing of the engines we find that the big Peugeot had an enormous valve area in relation to its swept volume, and hence a comparatively low power figure expressed in relation to square inches. A really significant thing on this score is the much better figures of 9.6 h.p. and 9.2 h.p. realised by the single-inlet-valve-per-cylinder designs of Duesenberg and F.I.A.T., compared to the 7 h.p. per sq. in. obtained on the Ballot. This seems a definite indication that the four-valve head does not offer such good flow values as the two-valve type. It is also apparent that the Duesenberg and Fiat engines were really the best of the 3-litre types, and their performance figures are strikingly similar. This is rather odd because they were radically different in design at every point. The Duesenberg had three-bearing crankshaft, plain big-ends, tubular connecting rods, crankcase and cylinder block cast in one and a detachable cylinder head with one inlet and two exhaust valves operated from a single camshaft. The inclined angle between the valves was approximately 50 degrees, and the drive from crank to camshaft by bevel gears and vertical shaft in the front of the engine.

The Fiat, on the other hand, had a ten-bearing crankshaft, roller-bearing big-ends, welded cylinder construction and the valves inclined at an angle of 96 degrees. This naturally gave two widely separated camshafts, which were driven by a train of gears from the rear of the engine. These two engines between them foreshadowed future developments, but for the time being there was no doubt that the Ballot represented the classic school of racing design throughout the world.

This being so, let me give some more details of this car beyond those mentioned in the table of specifications. The two camshafts were driven from a train of gears at the front of the engine and the cams (markedly different in shape) operated two exhaust and two inlet valves per cylinder through the medium of inverted steel pistons which fitted over the triple valve springs. The valves were inclined at an angle of 60 degrees and the sparking plugs, although mounted in the top of the cylinder head, were appreciably offset to the exhaust side. The diameter of the combustion chamber was slightly greater than the cylinder diameter and was not symmetrical but had a bulge, so to speak, in the fore and aft plane. Light alloy pistons were employed with moderate compression ratio (about 5.3 to 1), the gudgeon pin being only 15 mm. in diameter or 23% the size of the cylinder. The big-ends were also of comparatively small size, being 40 mm. in diameter. This part of the engine was, indeed, its Achilles heel, and is definitely a bad piece of design, the bearing consisting of a floating, white metal-lined bronze bush between the rod and the crank pin, thus impairing the heat flow from the white metal. The big-end was also asymmetrical in that there was a greater length of bearing from one side of the centre line than there was from the other, looking at the rod from sideways on. It is known that such a design does not by any means effectively utilise the theoretical bearing area, and despite a projected big-end area of 15.5 sq. in., it is not surprising that breakage of connecting rods limited the speed of revolution on this engine.

The crankshaft in itself was an ingenious piece of work, made up in four sections so as to provide for the five ball-type main bearings. The three front sections of the crank were interchangeable for easy production and repair when necessary. The crankcase was of light alloy casting, to which the cast-iron cylinder blocks (two, containing four cylinders each) were mounted, whilst lubrication was at low pressure and of the dry-sump type. Two magnetos were mounted on platforms and driven from the back of the camshaft, whilst each block was fed with mixture by a single Claudel carburetter through a Y-branch manifold. Water circulation was by a centrifugal pump at the front of the engine, which fed water into side plates on the exhaust side of the blocks, the return to the radiator being through four vertical offtakes between each pair of cylinders. From this angle the engine represented really advanced practice.

The whole power unit was mounted in a channel section sub-frame together with the gearbox. This had four speeds giving ratios of 3, 4.2, 5.4 and 7 to 1, the respective maxima being, therefore, approximately 80 m.p.h. in third gear, 62 m.p.h. in second gear and 48 m.p.h. in bottom gear. The exceedingly wide gap between top and third gear is to be noted, and the comparatively low speed on this ratio is a little surprising. On top gear the car did 34 m.p.h. at 1,000 r.p.m. The frontal area of the car was approximately 12 sq. ft., the engine thus conferring a little under 9 1/2 h.p. per sq. ft. The weight of the car with crew and fuel may be taken as 22 cwt., which gives 82 lb. per 100 c.c. and 22.5 lb. per h.p. The wheelbase was 8 ft. 9 1/2 in., the track 4 ft. 4 1/2 in., and, as a final detail, these cars were amongst the last to carry a spare wheel so as to enable a change to be made away from the pits. Although their main competition work was done in 1921, these cars were designed in the winter of 1919-20, and in general features they are similar to the 4.9-litre cars which had been produced with such astonishing celerity 12 months previously. Together these two cars undoubtedly set the fashion of the time.

It would, indeed, be hard to name any designer who has exerted more influence over his contemporaries than Ernest Henri, as I have disclosed. As Mr. Heal described in his narrative of racing history, Henri's Peugeots were almost invincible from 1912 to 1916, and, in fact, were beaten only twice in major races in this period, viz., in the 1914 French Grand Prix and the 1915 Indianapolis Race.

Readers of my opening remarks will have seen how the first post-war race was won by Henri's pre-war design, as was the second, viz., the 1919 and Targa Florio events. Then we have the 1920 Indianapolis victory by a direct copy of the Peugeot, although only ill-luck prevented Henri's latest Ballot design from coming home first. By 1921 his successful design work three years previously on the straight-eight engine had demonstrated its merit to such an extent that this type was established in a dominant position in the raging world. Again and again Henri's post-war cars narrowly missed winning races, and it is an ironic thing that he only finished first in a solitary major event, viz., the Italian Grand Prix. Nevertheless, he had certainly set the pace for all other engineers during seven years of active motor racing in Europe and the U.S.A.

So far as chassis design is concerned, the 1919 to 1921 period does not display any considerable novelty. The successful use of front wheel brakes in 1914 naturally led to these components being standardised on the cars of the post-war era. In this field Perrot was responsible for most of the equipment, with Birkgit designing ingenious servo mechanism which would get a good push on the brake shoes with reasonably light pedal pressure, despite the shocking mass of friction-full compensating mechanism which existed between the pedal and the brake drum.

Duesenberg provided the first example of successful hydraulic brakes in racing. A small pump driven off the gearbox provided pressure which was directed to a cylinder within the brake shoes on movement by the driver of the pedal. The brake shoes themselves were made from a flexible strip of metal with multiple segments of lining attached thereto. Although they did not become popular they were undoubtedly efficient. Semi-elliptic springs with friction shock absorbers were standard wear at the four corners of the chassis, whilst four-speed gearboxes were used by everyone except Duesenberg.

The relative merits of Hotchkiss drive and torque tube remained fairly evenly balanced in designers' minds. So far as body designs are concerned, although the 1919 Ballot (and the 1921 Sunbeams which were derived from it) had square tails the other cars were notable for a degree of effort towards streamlined form that was really remarkable. The Duesenbergs particularly, possibly due to their track racing traditions, had really beautiful lines, whilst the 3-litre Ballot by following the lines of the 4 1/2-litre 1914 Peugeot was a shapely car in which wind resistance had undoubtedly been studied by the designer. Typical body width at the largest section can be taken as 32 in., i.e., about 5 in. less than that of the Grand Prix cars of 1914. Despite this fact, it was possible to have both the driver and mechanic well enclosed by staggering the seats and putting the latter back about 4 in., so that he could put his right arm behind the driver's back. Generally speaking, by the end of 1921 racing car design seemed to have settled into something of a groove. The typical specification included a straight-eight engine with cast-iron cylinder block and narrow angle valves worked through twin overhead camshafts driven from the front of the engine. The crankshafts ran on ball or roller bearings, but the big-ends were plain. The engine ran at about 4,000 r.p.m. or a little more. The gearbox was separately mounted and the body had a long tail with rather rounded flowing lines.

These cars were fast and reliable and there seemed at the time to be no reason why they should not, as a general type, be used for many years. It is true that for 1922 to 1925 a 2-litre formula was to be enforced, but there was nothing in this which, of itself, demanded any radical change in design.

It was reasonable to expect a Continuance of the U.S.A. versus French struggle, with England perhaps gaining the Grand Prix win which had eluded her entries in the pre-1914 years. The facts proved contrary to these expectations.

By the end of 1922 the Henri technique had "gone like the wind." By the end of 1923 atmospheric induction was finished. By 1924 power per litre was doubled, and by 1925 speeds of road racing cars had risen to over 130 m.p.h. During these years U.S.A. intervention in Europe ceased to be a menace and almost became a farce. Details of this highly exciting period in racing car development Will be the subject of the second part of this article.

Meanwhile I should like to thank the Acting Editor of Motor Sport for extending to me the invitation to review the 1919 to 1927 races and racing cars, and to the Directors and General Manager of Temple Press, Ltd., for according me permission to do so and to draw upon data included in The Motor "Milestones of Speed" and other unpublished notes.