Number 7	SMITHSONIAN   ANNALS  OF   FLIGHT






&

SMITHSONIAN   AIR   AND   SPACE   MUSEUM

SERIAL PUBLICATIONS OF THE SMITHSONIAN INSTITUTION
The emphasis upon publications as a means of diffusing knowledge was expressed
by the first Secretary of the Smithsonian Institution. In his formal plan for the Insti-
tution, Joseph Henry articulated a program that included the following statement:
"It is proposed to publish a series of reports, giving an account of the new discoveries
in science, and of the changes made from year to year in all branches of knowledge."
This keynote of basic research has been adhered to over the years in the issuance
of thousands of titles in serial publications under the Smithsonian imprint, com-
mencing with Smithsonian Contributions to Knowledge in 1848 and continuing with
the following active series:
Smithsonian Annals of Flight
Smithsonian Contributions to Anthropology
Smithsonian Contributions to Astrophysics
Smithsonian Contributions to Botany
Smithsonian Contributions to the Earth Sciences
Smithsonian Contributions to Paleobiology
Smithsonian Contributions to Zoology
Smithsonian Studies in History and Technology
In these series, the Institution publishes original articles and monographs dealing
with the research and collections of its several museums and offices and of professional
colleagues at other institutions of learning. These papers report newly acquired facts,
synoptic interpretations of data, or original theory in specialized fields. These pub-
lications are distributed by mailing lists to libraries, laboratories, and other interested
institutions and specialists throughout the world. Individual copies may be obtained
from the Smithsonian Institution Press as long as stocks are available.
S. DILLON RIPLEY
Secretary
Smithsonian   Institution

The Curtiss D-12 Aero Engine


Curtiss D-12-E engine, 435 hp, 1930.  (Smithsonian photo A-4593.)

SMITHSONIAN    ANNALS    OF    FLIGHT	•	NUMBER    7
The Curtiss D-12 Aero Engine
by
Hugo T. Byttebier

SMITHSONIAN INSTITUTION PRESS
City of Washington
1972

UNITED STATES GOVERNMENT PRINTING OFFICE
WASHINGTON :   1972
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 - Price 75 cents
Stock Number 4705-000i

Acknowledgments
This history of the Curtiss D-12 engine is the result of research extend-
ing over several years. It has been made possible through assistance
provided by many persons and organizations.
  From its inception the project received the careful attentions of
Mr. Robert B. Meyer, Jr., curator, aero propulsion division, National
Air and Space Museum, Smithsonian Institution. Other contributors
include the following:
  Dr. Arthur Nutt, designer of the D-12 engine; Mr. Lee M. Pearson
and Mr. Eric Collins of the Bureau of Naval Weapons; Mr. Royal Frey
of the Air Force Museum at Wright-Patterson Air Force Base; General
James Doolittle; Major G. E. A. Hallett; Mr. Roland Rohlfs; Mr. Erik
Hildes-Heim; Mr. Harold Morehouse; Mr. George Page; Mr. Theodore
Wright; Technical Sergeant Merle Olmsted; Mr. Steve Wittman; Mr.
Alan Phillips; Mr. Paul Matt; Mr. Thomas Foxworth; the Curtiss-
Wright and Pratt & Whitney companies; Lieutenant Colonel F. E.
Rudston Fell, D.S.O., O.B.E., Wing Commander Norman Macmillan,
and the Rolls-Royce and Fairey companies, of Great Britain; Mr. Anto-
nio M. Biedma of Argentina; and General Edmundo Vaca Medrano,
former aeronautical attache at the Bolivian Embassy in Buenos Aires.
  Special thanks are due Mr. William Lewis, who interviewed Charles
B. Kirkham, and Mr. Paul E. Garber, historian emeritus, National Air
and Space Museum, who reviewed the original manuscript.
  

  
Contents
Page
Acknowledgments   		v
Introduction		1
Development of the Curtiss C-12		30
Development of the Curtiss CD-12		43
The Curtiss D-12 Engine		51
The Influence of the D-12		86
vn



Introduction
T
he date 28 September 1923 is an important one in the annals of
aviation. On that day, at Cowes in England, the sixth interna-
tional race for the Schneider Cup was held, and on that day the Schnei-
der race changed drastically from a contest to evaluate speed, reliability,
and navigability of seaplanes to an all-out, high-speed competition
between world powers.
  This change was brought about by two sleek, diminutive Curtiss racers
that had been brought from America to challenge the European sea-
planes. The Curtiss planes ran away from their competitors so convinc-
ingly that even those who had foreseen the result were surprised at the
extent of the victory. The apparent ease of the American win was
forcefully significant, and the observant press was quick to praise the
excellent combination of flying skill and advanced technology. The
most important element of the victory soon was pinned down to the
Curtiss D-12 engines. They proved to be powerful, light, and reliable,
and they introduced a new dimension in power plants by having a
frontal area for which one commentator found the correct adjective—
"piccolissimo."
  The D-12 engine is an excellent example of a first-line design in the
history of aero engine development. It was the result not only of
brilliant inspiration but of patient application, perseverance, and hard
work on the part of its developers.
  To understand the origins of the D-12 it is necessary to go back to the
autumn of 1915. World War I was about to enter its second year, and
to everyone concerned with aeronautics it had become obvious that the
Germans had taken a lead in aero engine development. The quickly
rising importance of the air arm caused the Allies to become increasingly
aware of their shortage of adequate engines, especially of high-power
units that would be decisive in the battle for aerial supremacy.
  Several American engineering firms tried to develop aero engines that
would be acceptable to the Allied flying services, but only a few were
able to secure orders.   The inherent problems of aircraft engine devel-
  
opment were made more acute by wartime difficulties. The fighting
forces were continually exerting pressure on the technical limitations of
an already highly sophisticated piece of machinery, thus imposing an
unrelenting stress on designers and manufacturers.
  One of the greatest efforts by an American firm toward the achieve-
ment of a first-class, high-power aero engine was undertaken by the
Wright Aeronautical Corporation, successor to the original Wright
Company. Wright Aeronautical, whose directors had seen the great
future that lay in engine manufacture, already had gained control of
the Simplex Automobile Company; and it had acquired the services of
that firm's chief engineer, Henry M. Crane, who had built up quite a
reputation because of the excellence of his Crane-Simplex motor cars.
  In making the decision to embark upon the manufacture of high-
power aero engines, the executives of Wright Aeronautical were aware
that no quick results would be reached by starting on the design of yet
another type of engine. Rather, they believed that the quickest and
safest way to achieve big production potential was to build an engine
based upon the most promising design being evolved in Europe. Thus,
the firm sent Crane and another representative to France to evaluate
the different types of aero engines being developed there and to deter-
mine the design best suited for production in the United States.
  The story of the D-12 engine starts in the fall of 1915 with the arrival
of the Wright mission in France. At that time a novel type of aero
engine was passing through its tests with fair results. It had been built
according to the ideas of Marc Birkigt, a young Swiss engineer. Some
ten years earlier Birkigt had linked his talent to capital from Spain
to enable that country to compete in the field of automobile manu-
facture. In this he had succeeded brilliantly, with the trade name
Hispano-Suiza having become well known in the automotive world.
  At the start of World War I, Birkigt had begun building an aviation
engine along original ideas evolving from his company's successful auto-
mobile racing engines. When Wright's representatives arrived this
aero engine was passing through a period of intense development prior
to its being launched on large-scale production.
  The usual water-cooled engine at that time consisted mainly of a
very stiff and relatively heavy crankcase upon which steel cylinders were
bolted individually. Each cylinder was surrounded by its own cooling
mantle, while the valve mechanism in the head was, in most cases, left
totally exposed. Birkigt's Hispano-Suiza engine departed radically from
that type of construction. Its cylinders were formed from an aluminum,
single-block casting into which were screwed four forged steel barrels
  

FIGURE   1 .—Cross-sectional  arrangement,  Hispano-Suiza,   200  hp, of  World   War  I.
(Smithsonian photo A-53098.)
that were threaded on the outside for their entire length (Figures 1, 2).
These barrels were closed at the top, forming a compartment that served
as a combustion chamber. Because the barrels were screwed into the
aluminum blocks, the passages for the cooling water were cored from
the block; as a result, the heat imparted by the combustion process to
the cylinder barrel had to pass through a part of the aluminum casting
before it was affected by the cooling medium. This was one of the
disadvantages of the design.
  A very light crankcase was attached to the cylinder blocks; thus, the
Hispano could be described as having the crankcase hung on the cylin-
ders instead of having the cylinders on the crankcase. The valves, two
per cylinder and mounted vertically in the cylinder head, were actuated
3


FIGURE 2.—Hispano-Suiza, Type I, 150 hp.  (Smithsonian photo A-4593B.)
by a single overhead camshaft which operated the valve stems directly,
without the interposition of either pushrods or rockers. The entire
valve mechanism was enclosed in an oiltight cover that fitted closely
over the cylinder block. The overall design resulted in a compact,
clean, light, and apparently simple unit which inevitably attracted the
attention and admiration of the engineering world.
  The most outstanding feature of the Hispano-Suiza, however, was
not the novelty of its design but its operating efficiency, something that
hardly could be said of other designs, advanced or not, which many
inspired and enthusiastic inventors were trying to adapt for use in air-
planes. The Hispano's clean lines and reliability in operation derived
more from its designer's outstanding talents in precision engineering
than from any simplicity of design—a fact which several engine manu-
facturers, deluded into trying to improve upon the type, would soon
discover.
  By 1915 the Hispano-Suiza V-8 engine was developing 150 bhp at
1,450 rpm, and,  thanks  to  Birkigt's long  experience  with   precision
  

FIGURE 3.—Charles B. Kirkham
(1882-1969) .
machine tools, it was ready for large-scale manufacture.
  Henry Crane, through his appreciation of good engineering practice,
recognized the Hispano-Suiza's superior design and great potential.
Arrangements were made immediately between Wright Aeronautical
and the French and Spanish parties involved for the granting of a
license for manufacture in the United States. At the same time, an
order for 450 engines was secured from the French government.
  While these activities were in progress, developments also were taking
place at the Curtiss firm, at that time the largest and oldest manu-
facturer of aero engines in the United States. In 1915 the company had
secured the services of Charles B. Kirkham, a young engineer of great
talent and long an acquaintance of Glen Curtiss. Kirkham, a builder
of motorcycle engines since the the turn of the century, had built his
first aero engine in 1910. He started his assignment as chief motor
engineer at Curtiss by improving the existing OX and VX models, of
90 and 160 hp respectively (Figures 5, 6). The OX-5, destined for a
long and famous career, already had been the object of important pro-
duction orders from the Allied air services.
  

FIGURE 4.—John North Willys (1873-1935).    (Photo courtesy of Automotive
Old Timers, Inc.)
  With an eye on developments in Europe, Curtiss, together with
several other American engine builders, began, toward the end of 1915,
a design study for the construction of a big aero engine of 300 or more
horsepower. The quickest way to achieve this was thought at the Cur-
tiss Company to be through enlarging the 8-cylinder VX model into a
12-cylinder engine. The first 12-cylinder Curtiss, designated the V-4,
was, like the products of the other American manufacturers of big
engines, of great weight and bulk (Figure 7); in fact, no such design had
weighed under 1,0001b, and the Curtiss V-4, when finished, turned the
scales at nearly 1,100 lb, dry. Although the engine eventually developed
about 400 bhp, it was not produced because a successful rival called the
Liberty was lighter and had the same horsepower. In addition, Kirk-
ham became interested in an entirely new design. The sole example
of the V-4 ended as the power plant of a speedboat, Miss Miami.
  The sudden appearance, in 1916, of the Wright Corporation as an
engine manufacturer—introducing the 150-hp Hispano-Suiza in the
United States with  the avowed intention of mass-producing it—was
  

FIGURE 5.—Curtiss OX-5, 90 hp, World War I.   United States National Air and Space
Museum  (NASM)  specimen number 1920-8.  (Smithsonian photo A-1832.)
viewed by the Curtiss management as a challenge, especially by John
North Willys, Curtiss's energetic president and financial backer. No
one seemed to doubt the Wright Corporation's ability to achieve what it
had set out to do, and Willys feared that Curtiss's engine department,
hitherto leading in the United States, was about to face strong competi-
tion. He also knew the Hispano's reputation, and he saw that Curtiss's
unwieldy V-type engines would never be able to answer Wright's chal-
lenge. He instructed Kirkham to design an engine to compete with
the Hispano-Suiza; and Kirkham, whose engineer's imagination was as
much stirred by the high efficiency of monoblock construction as had


FIGURE 6.-Curtiss V-X, 160 hp, 1916.   NASM number 1949-21.
(Smithsonian photo A-38654.)
been Crane's, decided that the only way to counter the threat from the
Hispano was to build something better.
  From Europe news had come that the Hispano-Suiza, by having a
reduction gear fitted and its revolutions raised, was developing over 200
bhp. So Kirkham, very judiciously calculating the future needs of
fighter aircraft as around 300 hp, determined, probably with full ap-
proval of the Curtiss management, to produce the world's foremost
fighter engine. Thus, his first venture into high-speed engine manu-
facture was to be a 300-hp unit of small dimensions that would outper-
form the Hispano in every way.
  

FIGURE 7.-Curtiss V-4, 400 hp, 1917.   NASM 1950-97.    (Smithsonian photo A-4989.)
  The new engine was to have twelve cylinders. At that time this was
most unusual, especially for engines of fighter planes, as it was believed
that the engine should be as short as possible so as not to impair the
aerobatic properties of pursuit planes, then being driven by compact
rotary engines. But Kirkham was after speed, and a 12-cylinder V-type
engine would permit a much smaller frontal area than would a radial
configuration or even a V-8, like the Hispano.
  In his efforts to create an engine that would surpass the Hispano,
Kirkham used the Hispano's progressive features only as basic inspira-
tion for an engine so audaciously designed that even Birkigt's revolu-
tionary project would appear conservative in comparison. Kirkham's
first aluminum monoblock engine was known, in Curtiss nomenclature,
as the model AB (Figure 8). It had a detachable cast-aluminum cylinder
head into which six individual closed-end cylinder barrels, or sleeves,
were screwed and heat-shrunk at the top only. The single aluminum
casting included crankcase and water jackets, on top of which was
bolted the cylinder head casting containing the sleeves. The steel
sleeves were made to fit neatly inside their jackets without actual con-
tact. A packing ring, fitted under heavy pressure at the lower end,
sealed the cooling water in the jackets from the crankcase.
  
* I
U/l I I

FIGURE 8.-Curtiss AB, 300 hp, 1916.  (Smithsonian photo A-5235.)
  Thus, by a masterful stroke, Kirkham had avoided the disadvantages
of the "dry-sleeve" construction as found in the Hispano. In his engine,
the cooling water was in direct contact with the barrel for nearly its
entire length. This "wet-sleeve" construction was easy to obtain with
individual cylinders as on the Mercedes or Liberty designs, which have
their own cooling mantles neatly welded around them, but in the case
of a monoblock engine it constituted a technical tour de force. The
AB (and the later K-12) construction was to become the hallmark for
all subsequent high-power design. It is interesting that Hispano-Suiza
engineers eventually switched over to wet-sleeve construction, but not
until 1928.
  As it originally was proposed to have the AB run at very high crank-
shaft speeds, a reduction gear was embodied in the design. The hous-
ing of the gears even formed a part of the large crankcase and water-
jacket casting, making the design even more complicated.
  For improved breathing at high crankshaft speeds the AB had four
valves per cylinder, against the Hispano's two. The valves were actuated
by a single overhead camshaft through T-shaped cam followers, one
follower operating the intake valves and the other the exhaust valves.
The drive of the camshaft was at the center of each cylinder block, an
arrangement that diminished the torsion problems but made the engine
10

slightly longer. As in the Hispano, the valve mechanism was entirely
enclosed in an oiltight cover and continuously lubricated.
  The crankshaft, supported by four main bearings, was balanced only
statically, with counterweights. The reduction gear pinion on the front
end of the crankshaft had no outboard bearing because the design of
the propeller shaft and its gear did not allow sufficient propeller
clearance.
  As compared to the 8-cylinder Hispano, Kirkham's AB had a smaller
(4-in.) bore for a smaller frontal area, and a 5Vi-in. stroke for improved
volumetric efficiency.
  By the time the AB was ready for testing, April 1917 had arrived and
the United States had entered World War I. The engine developed
300 bhp at 2,250 rpm but it saw very little test running for several
reasons. Kirkham found that it was too heavy (at 725 lb) for the power
developed. Then the gearing arrangement gave trouble, a portent of
much that was to come. The valve-actuating mechanism was not satis-
factory either, but still worse, in Kirkham's view, was the news that
Hispano-Suiza was building an improved V-8 engine of 300 hp, also
to be manufactured by Wright—which in the meantime had become
Wright-Martin. So, Kirkham scrapped the AB and proceeded without
delay on a new design.
  Kirkham's new engine, first known as the D-1200 (Figure 9) had a half
inch greater bore and stroke (4/2 by 6 in.) than the AB, displaced 1,145
cubic inches against the new Hispano's 1,127, and was intended to
furnish 400 bhp and run at an unheard-of 2,500 rpm—all this with a
designed weight of only 625 lb.
  From today's viewpoint, one wonders at the daring and foresight of
Kirkham, who brought out a fighter engine of 400 bhp at a time when a
proposed United States Standard Engine (known as the Liberty), be-
lieved to cover all existing needs, was planned as a V-8 of between 250
and 275 hp. The fact that the big new Hispano-Suiza eventually
materialized as a direct-drive, slower turning (1,800 rpm) engine did not
impress Kirkham; or, if it did, he paid no heed, for so possessed had he
now become by his inspiration that he no longer was to be deterred by
technical difficulties or problems with metallurgy.
  The greatest improvement over the AB was in the flow of the exhaust
gases. In the D-1200 the valves were actuated by two camshafts, one for
the exhaust and one for the inlet valves. Consequently, all exhaust valves
could be located on the outside of each cylinder block, where they
opened into individual ports for unhampered escape of burnt gases.
The exhaust ports fitted into a big collector which ejected the gases
11


FIGURE 9.—Curtiss D-1200, 400 hp, 1917.  (Photo courtesy of U.S. Air Force Museum.
Smithsonian photo A-4618A.)
directly upward through a large central chimney at each side of the
engine. Carburetion was by a new Ball and Ball Duplex type DVA-3,
and the two magnetos were Ericsson-Berling model D-66 X-4. Each
magneto fired one bank of cylinders, but this procedure was not well
thought out—a magneto failure resulted in an immediate loss of half
the engine power. The A.C. spark plugs were all located on the inside
of the V formed by the two banks of cylinders. As in the AB model,
the crankshaft was supported by four main bearings, but there was an
added fifth outboard bearing in front of the reduction gear for extra
support.
  Inaugurating a new breakthrough in aero engine technology, a me-
chanically driven, partly built-in supercharger was to be installed at the
rear of the new engine. It was to be driven by a long shaft running
through the center of the V and to be geared to the reduction gear at the
12


FIGURE lO.-Curtiss K-6, 150 hp, 1919. (Smithsonian photo A-4991A.)
front of the engine. A friction clutch was inserted near the front end
of the driving shaft to eliminate all possibility of any torsional or
accelerating forces being transmitted to the supercharger impeller.
Four mechanically driven superchargers built by General Electric were
ordered in 1917. Two of these were requisitioned by McCook Field's
engineering division, at that time busy with development of the exhaust-
driven turbosupercharger for the Liberty engine. The other two units
were delivered to the Curtiss Company, but development work on the
superchargers lagged because the engine itself demanded every possible
effort from Kirkham and his team. Later, the supercharger was found
to be unsatisfactory, but the idea for its use was significant in that it
showed Kirkham's early preoccupation with a feature that would come
into use a full ten years later.
13


FIGURE 11.—Curtiss K-12, 375 hp, 1918.   (Photo courtesy of U.S. Air Force Museum.
Smithsonian photo A-4618.)
  The new engine soon became truly identified with Kirkham by its
official designation, the K-12; and usually it was referred to as the
Kirkham or Curtiss-Kirkham engine. The prototype K-12 was bench-
tested (in the last week of 1917) at a weight of 625 lb, as designed.
A six-cylinder unit, the K-6 (Figure 10), was developed along parallel
lines. It really was half a K-12 (without reduction gears and other
pretentions toward high power) that was rated at 150 hp at 1,700 rpm.
  Undoubtedly, Kirkham had become fully aware that his K-12 (Figures
11, 12), with its very small frontal area and high power, stood unchal-
14


FIGURE 12.-Curtiss K-12, 1918.  (Smithsonian photo A-4594A.)
lenged; and, afraid lest his brainchild be wasted on some cumbersome
or indifferent airplane, he undertook at once to design a fighter that
would be able to use the K-12 to its best account—a battle plane that
15

would outperform any existing airplane as completely as he believed
the K-12 would eclipse every existing engine. Here was a case of a
single designer being able to devote his talents not only to developing
a new, untried engine but also to designing a new airplane that would
utilize it.
  The new airplane had characteristics that were extremely progressive
and original, following the true Kirkham touch. Full advantage was
taken of the low frontal area of the K-12, with the fuselage closely
tailored around the engine. This was a new departure in design, as
hitherto the engines somehow had been fixed onto the airplanes. A
possible exception was the French SPAD, which was built to take full
advantage of the compactness of the Hispano-Suiza, and it may have
inspired Kirkham to some extent. Pure speed had not been uppermost
in the minds of aircraft designers up to that time, as so many other
qualities had to be considered. The trend toward the combining of a
small, powerful engine with a wholly streamlined aircraft for obtaining
higher speeds had started, however, in 1917.
  Kirkham's airplane is described in the first edition of the Aircraft
Yearbook (1919, page 121): "The craft is a triumph in streamlining.
Not a square inch of the machine, from tail skid to exhaust pipes, has
been omitted from the careful plan of shaping, which has cut resistance
to the minimum."
  With high speed now secured by high power, complete streamlining,
and a small engine, Kirkham proceeded to expand on the theme by
using three wings on the plane's fuselage. Kirkham believed that a
triplane would increase maneuverability and climbing capability. To
top it all, the fighter was to be a two-seater, with the pilot having two
synchronized guns at his command and with the gunner, in the rear,
having two movable guns, thus adding maximum firepower to highest
speed and optimum climbing ability. This promised to become by
far the most formidable airplane designed since the beginning of the
war, and more than others it deserved the name of "battleplane" that
was applied to it.
  Capitalizing on the big promise of the K-12 and the triplane, Willys
approached the United States Navy and pressed for a big contract as
soon as possible. In spite of entreaties from the developers of the
geared Liberty, then on test, that that engine was more suited for its
needs, the Navy did not have the fixed idea for "one standard engine in
one standard plane." On 30 March 1918 the Navy placed an order
with Curtiss for two triplanes and four K-12 engines.
On 5 July 1918 the K-12 went up for the first time, powering the
16

triplane on its maiden flight. Roland Rohlfs, Curtiss test pilot, was at
the controls, and his take-off was significantly short.
  Progress of both airplane and engine was watched with keen interest.
In the first flights the four-blade propeller braked the engine at 1,900
rpm, restraining power and speed, so a new two-blade propeller was
fitted. The side radiators cooled too much, so smaller ones were
installed; but then the engine overheated on the fast climbing tests.
The problem was solved by adapting a by-pass device that eliminated
use of part of the larger radiator during horizontal flight.
  It was clear from the first flights that the triplane stood up magnifi-
cently to its claim for speed and climb; and it already had shown itself
capable of outdistancing a DH-4 with ease. Rohlfs claimed that the
climb was in the neighborhood of 2,000 ft/min, and maximum speed
was assumed to be around 150 mph. Both figures represented perform-
ances greater than any other airplane had been known to achieve.
  The first official flight finally was made on 19 August, and trials were
held from 1 September to 5 September. The witnessing Navy officers,
Holden C. Richardson and C. N. Liqued, may hardly have believed
the figures when it was shown that the 18-T—as the triplane was now
designated—on its first official flight had reached a best speed, over a
measured course, of 162 mph in full military trim. The previous
officially recognized speed record had been set at 126 mph, in September
1913.
  The first triplane (Figures 13, 14) was given Bureau No. A-3325; the
second was assigned the number A-3326. After the first official trials
the Navy was willing to order 750 K-12 engines at once, but the Bureau
of Aircraft Production decreed that this order was not to interfere with
the priority of materials destined for the Liberty engine, the mass pro-
duction of which already was underway. But to the great chagrin and
disappointment of everyone concerned, including the naval officers
delegated at the Curtiss plant to supervise manufacture, no large orders
were placed.
  The reasons for the Navy's reluctance were valid enough. The K-12
had not yet passed a 50-hour acceptance test; also, interest in triplanes
was on the wane at that time. Allied authorities had stated that tri-
planes were not satisfactory, and certainly not for two-seat fighters,
because of poor visibility and lack of stability as a gun platform. It was
clear also that the war was nearing its end and the time for large con-
tracts was past.
  The Army had already approached Kirkham for a biplane version of
the 18-T to be built along similar lines; and the Navy also had stated
17


FIGURE   13.-Curtiss-Kirkham   18-T   Wasp   triplane,  about   1919,   with  Curtiss  K-12
engine, 1918.  (Smithsonian photo A-4597.)
a preference for a biplane—if possible, a seaplane—and after the Sep-
tember tests it had recommended conversion of the triplane into a sea-
plane. As a result, in the summer of 1919 the A-3325 was converted,
using floats of the N-9 type, and it reached 126 mph during tests. In
April 1920 this aircraft set a world's seaplane speed record at 138 mph.
Later, both triplanes were prepared as racing landplanes, with the aft
seat faired in.
  The Army first had evinced interest when it had asked Curtiss to
lend it the Navy's A-3325 for test flights. This plane was flown at
Wilbur Wright Field on 18 September but it showed the same troubles
with overheating, due to the smallness of the radiator, as it had on the
earlier Navy tests. In the meantime, the Army's engineering division
had written a report on the biplane design the Army had asked for.
This report, based on the data given by the Curtiss Company and signed
by Alexander Klemin, bears the date 13 August 1918 and the serial
number 267.
  Criticism was made of the single gun on the turret for the observer,
though there was another gun on a flexible mount in the floor. When
the time came to evaluate the K-12 engine, it was noted that 400 bhp
was claimed at 2,350 rpm on a 6:1 compression ratio with a weight of
18


FIGURE 14.—The 18-T triplane with floats.  (Smithsonian photo A-4597A.)
680 lb. The weight figures were stated as being "Much below present
best practice. There either is an error in the weights or the motor may
have to be carefully watched." The phrase "present best practice"
meant, of course, the Liberty engine. It also was noted that the weight
of the K-12, given as 625 lb in June 1918, had been raised to 660 lb and
finally, in August, to 680 lb. Curtiss then was given Order CS-152, for
two biplanes and two triplanes. One plane of each type was to be stati-
cally tested to ascertain the true limits of strength, while the others were
to be thoroughly tested in flight. The static tests of the first triplane,
delivered to McCook Field in February 1919, revealed serious weak-
nesses in several components. The first biplane was delivered in June
1919, after it had been displayed in New York at the Madison Square Gar-
den show in March. It received number SC-40454 and became the
P-86. Its crash, soon afterwards, ended the Army's interest in the type.
Meanwhile, the Wright-Martin Corporation had been experiencing
difficulties in its efforts to bring the Hispano-Suiza to mass production.
It was clear from the start that the design of that engine was not one
that could be copied easily, and considerable delays were suffered for
different reasons. To make things worse, new specifications for im-
proved ratings were coming from France, raising the power first to
19


FIGURE 15.—Hispano-Suiza Type H, 300 hp, 1919.  (Smithsonian photo A-4594B.)
180 hp, then to 200 hp, and later to 220 hp. It was not noted, however,
that as horsepower ratings went up, reliability went down.
  In the face of these conflicting interests, Wright decided, wisely, to
stay with the original design and direct all efforts towards bringing the
150-hp model to production as soon as possible. But not much headway
was made at first. The chief difficulties lay in the extreme stringency
of the specifications for the different materials and in the excessively
small tolerances to which every piece had to be finished, making the
work more difficult than watchmaking. The famous cast-aluminum
cylinder blocks were giving nightmares to all concerned, and most of
the troubles were solved only when the company had established its
own aluminum foundry, working at superior refinements. Production
finally started at the beginning of 1918, but by then the proud "Hisso"
fighter engine of 1916 had become suitable for trainer duties only.
  When the blueprints for the new 300-hp model finally arrived, prepa-
rations were made promptly for its speedy production, but the solving
20



FIGURE 16.— Crankshaft and propeller shaft, showing reduction gears, of Curtiss K-12,
1919. (Smithsonian photo A-4594D.)
of problems similar to those with the smaller types resulted in the new
model (Figure 15) not being ready for production until near the end of
1919. Still, the fact remains that the small Hispano finally was pro-
duced successfully in satisfactory numbers. Two British attempts to
improve on the Hispano form of construction—the Siddeley Puma and
the Sunbeam Arab—had not been successful, and subcontractors were
running into trouble everywhere. (The fitting of steel cylinder liners
in aluminum blocks still is considered a major problem by automobile
manufacturers today.)
  Against this background it is easier to understand why the K-12,
which was intended to outperform the Hispano in many ways, encoun-
tered numerous difficulties. The main troubles were related to the reduc-
tion gear, the large aluminum casting, and the crankshaft (Figure 16).
  The reduction gear troubles were particularly irritating because they
were a necessary evil. The problem basically stemmed from the oppos-
ing relationships between the engine, which, in becoming lighter and
smaller, could run faster, and the propeller, which gained in efficiency
by running slower. In the case of the propeller, the rotational speed
was further limited by enormous problems that arose when the velocity
of the tips reached the speed of sound. This put an upper limit to
propeller speed at about 2,000 rpm. A prudent safe maximum being
considered at the time was around 1,800 rpm. The only solution
lay—and still lies—in the provision of a form of gearing which reduces
21

the speed of the propeller shaft of the engine to a figure better suited
to propeller efficiency. The problem is as old as aviation itself. The
Wright brothers used, from the first, a crude but efficient form of gear-
ing with bicycle chains.
  The Hispano-Suiza was readily adaptable to high-speed running. The
original model of 150 hp had been enabled to turn out increasing power
ratings by the simple expedient of making the engine run faster. With
crankshaft speeds of up to 2,150 rpm coming into use, the introduction
of a reduction gearing became imperative. But at the same time it was
found that the gearing was a delicate and tricky affair, and seizures were
frequent. The big, 300-hp Hispano had been conceived in geared and
ungeared form, but the geared form was not followed up and the crank-
shaft speed of the direct-drive production model was kept down to
1,800 rpm.
  The K-12, which from the onset was designed for highest power and
smallest size, could not dispense with the gears, which had a reduction
ration of 5:3 (or 0.6), but here, too, continuous trouble-free running
could not be attained in spite of the addition of a fifth bearing in front
for the support of the pinion.
  The integral construction of crankcase and water jackets in one
aluminum casting, while making for great saving in weight and theo-
retical rigidity, really was a type of construction that bordered on the
impossible. The blocks were cast at the Buffalo foundry of the Alu-
minum Company of America, which worked diligently and conscien-
tiously to produce good castings, but such a complicated and large
casting was beyond the art of the time. The results: too many blocks
condemned because of warping, too many misruns, and too much
porosity, doubtless due to the high pouring temperatures used in the
attempt to avoid misruns.
  The crankshaft, supported on four bearings and with counterweights
for static balance only, was not reliable and breaking occurred easily.
The situation could not be improved, and an exasperated Kirkham
ended his attempts by refusing to strengthen the design and preferring
to blame the material as faulty.
  The first of six attempts to run an endurance test as required by the
armed services was started at the Buffalo plant of the Curtiss Corpora-
tion on 9 August 1918 using K-12 engines 7, 8, and 9. The normal
schedule to which any engine had to be submitted consisted of ten
separate five-hour runs. The first half hour of each run was at full
throttle and rated power, while the remaining four and a half hours
were at 90 percent of rated power.    Engines 7 and 8 weighed 665 lb,
22

dry; engine 9 had a 24-lb counterweight added in the center of the
crankshaft, raising its dry weight to 689 lb.
  Test 1, with engine 8, was discontinued after 1 hour 36 minutes, full-
throttle, when connecting-rod bearing 2 burned out because of break-
age of the oil manifold.
  Test 2, with engine 7, ended after 20 minutes of the third five-hour
period (i.e., after 10 hours 20 minutes) when six teeth of the reduction
gear on the crankshaft broke.
  Test 3, with engine 8 after its oil manifold had been repaired and
strengthened, was discontinued after only 4 minutes running when
cylinders 1, 3, and 5 ceased firing. The breakdown occurred when
parts of an auxiliary air valve of the front carburetor were sucked into
the cylinders, holding the valves open and bending the valve stems.
The auxiliary air valve was then redesigned.
  Test 4 was the third attempt with engine 8, after it had been put in
condition again. This time the K-12 ran well and hopes soared, but
after 21 minutes of the third five-hour period the articulating boss on
master connecting-rod 6 broke; as a result, the crankcase cracked
between cylinders 5 and 7 and the base of cylinder 5 ;was damaged.
This test marked the end of engine 8. The boss on master rods was
strengthened on the other engines.
  Test 5, with engine 7, was an all-out effort to make the 50 hours,
but it was marred by numerous interruptions. The first stoppage came
after 10 minutes running when the tachometer drive housing broke.
It also was discovered that one camshaft bearing had broken. Heavier
bearings were installed and the test was resumed, but after 5 minutes
more there was a second interruption—the right-hand magneto drive
shaft had frozen to the bushing in the tachometer drive gear housing.
When this trouble was cured the engine went well for the remainder
of the first five-hour period. After 20 minutes at full throttle in the
second period there was another interruption when the clamping bolt
of connecting-rod 2 broke away. The bolts were replaced with others
having thicker heads. A fourth interruption, after 3 hours 8 minutes,
was caused by the breaking of the lock screw for the tappet of the rear
intake valve of cylinder 11, letting the valve drop into the cylinder.
This lock screw had been bent in tightening. Lock washers were fitted
and the test was resumed. After half an hour at full throttle and 4 hours
37 minutes at 90 percent power, the fifth interruption occurred when
the oil line to the pressure gage at the crankcase cracked. After another
4 hours 9 minutes running, this test was at last discontinued when a
lock nut on the rear of the propeller shaft broke.   This time the engine
23


FIGURE 17.—Assembly of pistons and connecting rods, Curtiss K-12, 1919.
was ruined beyond repair because the propeller shaft shifted out. As
the reduction gears on the propeller shaft and crankshaft were of the
herringbone type, the propeller shaft pulled the crankshaft with it.
This broke the oil feed line to the bearing and interrupted the oil feed
to connecting-rod 2, causing its bearing to burn out together with the
crankpin and thus spoil the crankshaft. This was the end of engine 7.
Test 6, with engine 9, was interrupted after half an hour wide open
and 2 hours 57 minutes at 90 percent power when an exhaust valve
tappet lock screw loosened on cylinder 12 and the valve dropped into
the cylinder. The valve tappets then were more carefully fitted and all
lock screws were wired to prevent loosening. A new start was made
but the test was interrupted after 25 minutes of the second five-hour
period when the A.C. Titan spark plugs began to leak badly. All plugs
then were replaced by a complete set of Bethlehem Aviation plugs.
24


FIGURE 18.—Cylinder head casting of Curtiss K-12, 1919, with cylinder sleeves, cam-
shafts, and valve gear assembled.  (Smithsonian photo A-4602A.)
The test was discontinued after 51 minutes of the third five-hour period
when two rear camshaft bearings and three teeth of the reduction gear
on the crankshaft broke.
  The report ended on an optimistic note: "The design of the engine is
excellent and it is believed to be capable of passing the 50 hrs test
without difficulty when the troubles mentioned are overcome." The
results of the different tests showed an average power of 389 hp at
2,250 rpm and a maximum of 418 hp at 2,550 rpm. Compression ratio
was 5.5:1. In the official report the engine was mentioned as Kirkham-
Curtiss model D-1200, rated at 375 bhp at 2,250 rpm.
  Toward the end of 1918, or very early in 1919, two new engines,
numbers 14 and 15, were sent to the engineering division at McCook
Field to be submitted again to a 50-hour endurance test. Engine 14
broke its crankshaft during the preliminary calibration test, so only
engine 15 was given a test, which is described in McCook's Serial Report
811 of 25 April 1919. At the request of the Curtiss Company, the
engines were not run above 2,400 rpm, which was about the peak of
the power curve. The average rated power was 396 bhp at 2,250 rpm,
with a maximum of 406 hp at 2,420 rpm.
  The report concluded with the statement that the engine, even though
well designed and of high power for its weight, appeared to be some-
what complicated and of difficult maintenance. The tested engine
suffered from water leaks throughout the duration of the tests, which
terminated with a cracked cylinder jacket.   The report ended on a note
25


FIGURE 19.—Curtiss Eagle, 1919, powered by three Curtiss K-6 engines.
(Smithsonian photo A-4602E.)
less optimistic than the previous one: "The [K-12] engine is still in the
experimental stage and is not as yet suitable for actual service."
  It has sometimes been stated that the K-12 was turned down because
the government was concentrating on the Liberty engine. This is only
partly true. The United States Navy retained its interest throughout
the K-12 engine's career. Besides, the Curtiss Company never had been
associated with the design or production of the Liberty. Undoubtedly
it would have, had it not believed that it had something better. A
great part of the millions of dollars spent on the American aeronautical
effort was poured into the simultaneous attempts to mass-produce and
improve the Liberty engine, a high-power design based on proved con-
structions. Though these efforts made it possible for America to get
aero engines to the battlefront before November 1918, when the
Armistice occurred, the men actively connecetd with the K-12 never
could shake off the belief that the United States would have had a
superior product had a fraction of the millions been put into the
development of Kirkham's aluminum-block design.
  After the Armistice, the interest in experimental airplanes and engines
waned rapidly, and no new contracts were made for development or
26


FIGURE 20.—Curtiss Oriole for Amundsen polar expedition, with Curtiss K-6 engine,
1922.  (Smithsonian photo A-4602F.)
production of the K-12; however, existing contracts were not canceled.
  With the advent of peace, men were confident that weapons would be
abolished for all time and the aviation manufacturers were eager to get
going with plans for new commercial aircraft, of which great things were
expected. The Curtiss Company, for one, brought out a beautiful new
trimotor transport, the Eagle, that was powered by three 150-hp K-6
engines (Figure 19). Another use for the lower-stressed, direct-drive
K-6 was found in a new three-seat sports plane, the Oriole (Figure 20),
Which was also available with the war-surplus OX-5.
  Curtiss President Willys was even more anxious now to get the K-12
past its teething troubles and to turn it into a salable product. He
decided to call Finlay Robertson Porter to assist Kirkham in bringing
the costly experiments with the K-12 to an end as soon as possible. In
1918 Porter had been chief motor engineer at McCook Field, and he had
acquired some reputation in the automobile field. Porter and Wright's
chief engineer, Crane, were old competitors. They had marketed,
respectively, the FRP and the Crane-Simplex, the two most expensive
motor cars built in the United States during 1916. Porter also was an
old acquaintance of Willys, having worked with him in 1915 on a racing
car powered by a Knight engine.
27

  But Kirkham was not willing to accept any kind of assistance or
supervision. The antagonism thus created between him and the Curtiss
management culminated in his resignation from the Curtiss Corporation
early in 1919. He then founded his own firm, Kirkham Products, which
was devoted to consulting engineering in aeronautics and which would
be responsible for a number of highly original airplanes and engines.
His assistant, chief engineer, T. S. Kemble, left Curtiss in April. Kirk-
ham was retained for a time as consultant at Curtiss's Garden City
plant to advise on the triplane, the trials of which were continuing; but
his connection with the K-12 was severed forever, and Curtiss no longer
was to use his name in relation to the design or use of the engine. The
fate of the K-12 came completely under the responsibility of Porter,
who became chief motor engineer at Curtiss.
  About twenty K-12 engines had been built and several of these were
still around; so, when the first big postwar aero show opened in
Madison Square Garden on 1 March 1919 the K-12, rated at 375 hp,
was exhibited on the Curtis stand. The Army's 18-B, now with the
name Hornet, also was shown. For the Hornet, a maximum speed of
163 mph was claimed, and a speed of 80 mph was mentioned for normal
use, which needed much less than half the maximum engine power. It
perhaps was piously hoped that the K-12, if sold, would not be used
too much at full throttle. At that time there were no type certifications
or FAA inspections, and manufacturers were free to offer any product;
but there is no record of any person or any service in the United States
buying one of those "fastest planes in the world," as they then un-
doubtedly were, nor was the K-12 sold for installation in any other
civil or military airplane in the United States during that year.
  Through vigorous publicity and the efforts of a special sales delega-
tion some commercial success was achieved in South America. One K-12
was installed in a triplane flying boat, ex-Navy A-2277, which was used
for exploration, and one K-12-powered 18-T triplane was sold to the
Bolivian government, of which more later.
  The 18-T triplane had been given the name Wasp, and, by one of
fate's tricks, the names Wasp and Hornet were the very ones that would
be assigned to the two engines that were to cause, eight years later, the
displacement of the K-12's successors in military airplanes.
  It had been an idea of Chief Test Pilot Roland Rohlfs, who remained
enthusiastic about the triplane's climbing abilities, that an attempt
could easily be made to win the world's altitude record. Triplane
A-3325 (or possibly A-3326), still in abeyance at the Curtiss Garden
City   plant,  was  converted  for   that  purpose  with  enlarged   two-bay
28


FICURE 21.—Curtiss  18-T Wasp with Curtiss  K-12 engine, destined for Bolivian  air
force, October 1919.  (Smithsonian photo A-4602G.)
wings. A specially prepared K-12 with a higher compression ratio was
used. When the first attempts were made, two or three engines burned
out because the scavenging oil pump in the forward part of the engine
failed to work at a high angle of climb.
  In July 1919 an altitude of 30,500 ft was reached, and finally, on 18
September, with engine 13—which had produced 465 bhp at 2,600 rpm
on a compression ratio of 6:1 during a bench test in January—Rohlfs
attained a height of 34,910 ft and claimed a world's record. The first
part of the climb was made with a nursed engine until 31,000 ft was
reached, but then, due to the steep climbing angle maintained and the
lessened density of the air, the plane began to stall and dropped 600 ft.
Rohlfs then applied full throttle and lowered the angle of climb for
the remaining 4,000 ft. This record was never officially recognized as
a world's record by the Federation Aeronautique Internationale, ap-
parently because the barograph, after the necessary corrections, did not
show a sufficient improvement over the preceding record, made on a
French Nieuport with a special high-compression Hispano-Suiza of
300 hp.
29

  The climbing and high-flying demonstrations of the Wasp triplane
were decisive in convincing the Bolivian authorities of the worth of
that aircraft. They were anxious to have an air force like everybody
else, but had been confronted by the problem of having their capital city
high in the Andes mountains, with its airfield, El Alto, situated at
13,500 ft above sea level. Prior to the Bolivian government's purchase
of the Curtiss Wasp (Figure 21), and in spite of some earnest attempts,
no airplane had been able to take off from that airfield, but with the
arrival of the K-12-powered triplane the situation changed dramatically.
On 17 April 1920, in the able hands of Donald Hudson, who was on
loan from the Curtiss Company to the Bolivian government, the Wasp
took off from El Alto with such ease and grace that the assembled
officials were wildly enthusiastic. The Bolivian Air Force was created
on the spot, and Donald Hudson was made a lieutenant colonel and
chief pilot.
  The first flights in Bolivia were phenomenal. A short time after the
first take-off Hudson reached 31,000 ft, and on 12 June he crossed the
Andes, flying at 30,000 ft. Unfortunately, the elation was not to last
long. In the following year, on 19 May 1921, Hudson made an over-
land flight of 160 miles to Oruro in 75 minutes, but on the return flight
the inevitable happened—the K-12 gave up, and the plane crashed and
was destroyed beyond repair, but with no injury to the pilot. The
Bolivians, who by now had become aware of the experimental status
of their purchase, did not accept the loss graciously; they felt cheated,
and the contract with Hudson and Curtiss was canceled. Five years
later, however, relations with Curtiss became cordial again when Jimmy
Doolittle flew a Curtiss Hawk fighter to El Alto field and sold several
such craft to the Bolivian government.
Development of the Curtiss C-12
Meanwhile, the Curtiss Company, under the presidency of John North
Willys, was not sure how to proceed with the K-12. Apart from the
preparations for the height-record attempts, it appeared unlikely that
more work would be devoted to that engine.
  In the summer of 1919 Clement Keys, Curtiss's vice president, began
a tour of Europe as a member of the American aviation mission. On
that  trip he  visited  the  most  important aircraft and engine manu-
30



FIGURE 22.—Sectional drawings of Napier Lion, 450 hp, 1919.
(Smithsonian photo A-4619.)
31

facturers in England and France, and he quickly became aware of some
interesting situations in regard to aero engine development. In Eng-
land, for instance, development of water-cooled engines had virtually
stopped, except for the broad-arrow Napier Lion (Figure 22), and any
money made available was being spent on perfecting the high-power
air-cooled radials that had been designed in 1917 and 1918. In France
there was such a large stock of aero engines on hand that development
had stopped altogether, and the only activity was concentrated on main-
taining normal standards of reliability for the number of such engines
as were needed. The leading engine there was the 300-hp Hispano-
Suiza, the same model that was being readied for service by Wright
Aeronautical in the United States.
  Keys discovered that no engine in Europe could match the K-12 in
regard to power, low weight, and small size. Also, he saw that several
French airplane manufacturers were adapting the new 300-hp Hispanos
to modified prototype fighters for attacks on the world's speed record,
set in 1913. Indeed, that record was broken in October 1919 by a
Nieuport-powered, specially prepared single-seater with clipped wings
that flew at 167 mph. In contrast, the 162 mph attained by the Curtiss
triplane in full military trim showed the promise of the K-12 for speed
work, even if it was not then adaptable for commercial use. Keys argued
that the speed record should be brought to America, and he wrote to
Willys that he thought it advisable to continue the development of the
K-12, the fate of which was hanging in the balance.
  In April 1919 Porter had started on his assignment at the Garden City
plant; and in May he took over direction of the operations in Buffalo,
where the K-12 was being tested under direction of Arthur Nutt, the
firm's young motor test engineer of whom we will hear more. Porter
took a good look at the K-12, was able to form an opinion about the
principal weaknesses of the design, and proceeded to redesign the engine
accordingly.
  The changes introduced in Porter's new design thus were a reflection
of the experiences with the K-12. The large aluminum casting was
eliminated by separating the cylinder block from the crankcase and
bolting them together, which was more in accordance with Hispano
practice. This arrangement eliminated the tooling problems that had
appeared with the integral castings, and it eased maintenance to some
extent because repairs to any cylinder could now be handled by remov-
ing a cylinder block instead of removing the whole engine, which the
K-12 type of casting demanded.
  The crankshaft was changed to a seven-bearing type and the counter-
weights of the K-12 crankshaft were eliminated, but since the cylinder
32



FIGURE 23.-Views of Curtiss C-12, 400 hp, 1920.
(Smithsonian photos A-4881C, A-4879.)
33


FIGURE 24.—Crankshaft, reduction gear, propeller shaft, and propeller hub of Curtiss
C-12, 1921. (Smithsonian photo A-4619F.)
blocks were not redesigned the total bearing surface could not be aug-
mented. Consequently, the introducing of more bearings was largely
a step backwards.
  The spark plugs, which on the K-12 had been placed side by side
on the inside of the V, were relocated to both sides of the cylinder block
so that six plugs were on the outside and six on the inside of each
cylinder bank facing each other. Now, the failure of a magneto would
not cause the loss of half engine power as with the K-12.
  New carburetors of the Claudel-Hobson double-inverted type were
installed. The magnetos were Berkshire D-6s, each firing one A.C.
plug per cylinder.
  The large exhaust collector was eliminated, obviously because it
burned out too easily. All the new engines and the remaining K-12
units were fitted with short individual exhaust stacks, one for each
exhaust orifice, so that there were twelve stacks on each side of the
engine, a distinctive feature of all Curtiss water-cooled engines ever
since. One source of trouble, the reduction gearing, was retained on
the new design, but there was no way to eliminate it without suffering
a great reduction in power.
  The new engine (Figures 23-25) became known as the C-12. It was
rated at 400 hp for 2,250 rpm, and the weight, at least for the first units,
was given as 675 lb, about that of the K-12. The six-cylinder K-6 was
redesigned along the same lines and renamed Curtiss Six, or C-6. It
had a rating of 160 bhp at 1,750 rpm and was to have a long and suc-
cessful life, without further modifications.
34


FIGURE 25.—Pistons and connecting rods of Curtiss C-12, 1921.
(Smithsonian photo A-4624G.)
  The first C-12 engines were ready at the end of 1919, and the first
unit was tested in January 1920. With the commercial market still in
mind, Curtiss gave the new engine its first publicity showing at the New
York aero show in March 1920 as the power plant of a new version of
the Eagle transport, the Eagle II—the same airplane that had flown in
1919 as a trimotor with three K-6 engines. The airplane made its new
appearance with two C-12 engines installed.
  Roland Rohlfs piloted the new Eagle for one test flight, but during
take-off the shaft driving the water pump seized on the left engine,
which then lost all power within seconds. The plane barely cleared a
building, actually rolling its wheels on the roof. This was the only
flight of the Eagle II, as the C-12 engines were removed immediately.
But it was not the end of the Eagle. In 1920, when the commercial
value of the new projects was beginning to appear doubtful, the Eagle
was offered to the Army for use in a project for a day-bomber with a
single Liberty engine. Three such bombers were ordered, and the
existing Eagle was immediately converted with a single Liberty engine
and delivered to Mitchell Field.   No others were built.
  About the time the Eagle II made its unsuccessful test flight, both
Willys and Porter were becoming convinced that there was no future
in the development of high-powered aero engines, which apparently
were unwanted. The Liberty had stolen a march on all new designs,
and there were too many Liberties in storage.    Both men withdrew
35

from the Curtiss Company to devote their talents and energies toward
the promotion of the automobile industry. Porter again would market
America's most expensive motor car, the Porter. The products of Willys
are well known.
  The C-12 was to be saved from oblivion by the enthusiasm of Clement
Keys, Curtiss's new president. Keys decided to bury all attempts at
possible commercial use and to go ahead with a program for high-speed
flying, using all the qualities of the new C-12 engine toward that
purpose.
  The first important opportunity for the C-12 to show its mettle
presented itself with the first postwar race for the Gordon Bennett Cup.
This event, to be held in France in the summer of 1920, promised to
become one of international importance. Each competing country was
permitted three entries. The race itself was to be a pure speed event
over a course of 300 km (186.5 miles). Each contestant could choose the
time of day for his entry.
  For Curtiss, the winning of the Gordon Bennett Cup was particularly
desirable, as a Curtiss plane and engine had won the first such event in
1909 and derived much prestige from it. Officials at Curtiss decided
to make a special effort, and the victory appeared well within reach
because the decisive element, the engine, already was at hand.
  A special committee met in March 1920 at a luncheon in Dayton at
the invitation of Texas oil millionaire and aviation enthusiast S. E. J.
Cox, who already had entered an airplane under the auspices of the
Aero Club of Texas. A contract between Cox and the Curtiss Corpora-
tion that called for the building of a racer capable of at least 200 mph
was signed on 19 June.
  Apart from Curtiss, two other contestants from 'the United States
entered racers that represented completely different approaches to the
problem of obtaining high speed. An entry by the United States Army
Air Service represented the expression of the "power is everything"
philosophy. This was a Verville VCP pursuit biplane, originally built
to carry the 300-hp Hispano but for this occasion fitted with Packard
Engineering Vice-President Jesse Vincent's latest masterpiece, a 12-
cylinder blockbuster based on the Liberty but displacing 2,025 cubic
inches. At McCook Field the new engine had been thoroughly tuned,
the compression raised, and a special fuel concocted, resulting in a
power on the bench of well over 600 bhp. The reformed pursuit,
designated the VCP-R, was declared capable of 200 mph.
  The third American entry reflected the other extreme of design. In
this  airplane,   built   at  the   Dayton-Wright   Company   where  Orville
36


FIGURE 26.—Curtiss-Cox Texas Wildcat, 1920, with Curtiss C-12 engine, 400 hp.
(Smithsonian photo A-4619A.)
Wright was consultant, extreme fineness of penetration had been the
goal, making up for a lack of engine power. The Dayton-Wright racei
was a full cantilever monoplane with wings of variable camber. The en-
gine, built by Hall-Scott, was a six-cylinder, in-line Liberty of maximum
250 hp, making this racer the lowest powered airplane entered—but the
in-line engine permitted the least possible frontal area. Another special
feature was the retractable undercarriage, a novelty at this time. The
pilot was enclosed so completely in the fuselage that he had limited
visibility, only through a window at each side. This racer, like the
VCP-R, was advertised as being capable of 200 mph.
  Curtiss, of course, had the C-12, giving 427 hp at 2,250 rpm—without
even being pushed very much—and capable of good durability with full-
throttle running. Not content to play up only the engine, Curtiss also
gave the airframe maximum attention—perhaps too much. Under the
direction of William Gillmore, the Curtiss engineers, carried away by
the prevailing high spirits, designed and built two extremely small racers
around the C-12 engine. Mrs. Cox named them Texas Wildcat and
Cactus Kitten (Figures 26, 27). The first flight was made on 25 July, five
weeks after the signing of the Cox contract. As usual, the undauntable
Roland Rohlfs was in the cockpit. With large test wings, a speed of
183 mph was reached, but a special minimum-surface wing already was
in the works and with these special race wings the maximum speed was
calculated as 214 mph. These were certainly the fastest airplanes of
their time, and the Gordon Bennett Cup appeared as good as won. As
events turned out, one of the Curtiss racers easily could have won with
the large test wings or any sort of wings, but such was the furore aroused
37






FIGURE 27.—Drawings of general arrangement of Curtiss-Cox Cactus  Kitten,  1920,
powered by Curtiss C-12 engine.  (Smithsonian photo A-21253.)
by the press that it was thought to be of no use to enter a racer not
capable of at least 200 mph.
  The first tests of the Curtiss racers were held at Roosevelt Field,
which was not large enough to test the maximum-speed wings. As time
was runing out, the decision was made to ship the two planes to France,
in the belief that the big, smooth airfields near Paris would make testing
and adjusting easy. After the usual delays with shipping and customs,
the machines arrived at their assigned airfield, still in crates, a bare ten
days before the race. After another five days of work the Texas Wildcat
was ready for a first test and Roland Rohlfs took his place at the
controls.
  The C-12 was started and warmed up. Soon Rohlfs gave the signal
and applied full throttle, but at once he found himself in a distressing
situation. The geared-down, fixed-pitch propeller had been calculated
to pull hard at 200 mph, but at zero airspeed its pull was hardly notice-
able. Add to this the effect of wings which did not lift until an im-
pressive speed was reached, and the result can readily be imagined. The
engine roared its utmost but the start was frustratingly slow.
  Rohlfs used up the entire length of the airfield and an adjoining piece
of farmland before he finally was able to get the unwilling plane air-
borne. He continued at a slight climb until the airplane reached a
high speed and became fairly controllable.    Under continuining full
38

throttle the Wildcat accelerated still more until suddenly, above 190
mph, it went into an oscillating up-and-down motion of such violence
that Rohlfs had to land as soon as he could. Again the plane overshot
the field, and to all intents and purposes the Curtiss racers appeared
to be out of the race.
  The C-12 never had missed a beat during the long take-off and the
high-speed run, and it had demonstrated that it was fully dependable.
So the race manager, Mike Thurston, one of the design engineers who
had accompanied the planes, decided to have another try. In his hotel
room that night he laid out plans for a completely new set of biplane
wings and empennage, and arranged with the men of the Morane-
Saulnier plant, whose airfield was being used, to manufacture them.
The Texas Wildcat had the new wings and tail surfaces on the after-
noon of the day before the race. With so little time remaining it was
necessary to take off and fly the aircraft to Etampes Field, 30 miles
away, where the Gordon Bennet Cup race was to be held.
  The Curtiss racers, however, were deficient in one other detail—the
undercarriage was not fitted with shock-absorbers, and an untested new
kind of Dunlop wheel, with radial spokes, was used. At near take-off
speed a sudden obstruction in the field caused the Wildcat to bounce
into the air, where, luckily for Rohlfs, it stayed, as the new wings were
lifting beautifully. But both wheels had been so severely damaged that
they collapsed when the plane touched down on Etampes Field. The
airplane turned over and broke in two near the cockpit. Rohlfs, of
whom it truly might be said that he feared God but not the products
of the Curtiss Company, was saved miraculously by his parachute pack
and by the fact that a minimum supply of petrol had been carried for
the short flight. The other Curtiss racer had been left in its crate. So
ended the high hopes of the Curtiss people to win the Gordon Bennett
Cup in 1920. Through no fault of its own, the C-12 had lost an oppor-
tunity to show again what small size and high power were worth for
speed.
  The two other American entries showed up no better; they reached
the starting line, but the Verville had to land soon after take-off because
the big Packard overheated so badly that the pilot, Rudolph ("Shorty")
Schroeder, expected to see his plane go up in flames. The hurried land-
ing proved to be a blessing, as the fuselage later was found to be cracked
so severely that the tail certainly would have come off at the first turn.
The other American entry, Dayton-Wright, was found to be unable to
turn to the left and had to withdraw, never to be raced again.
So ends  the story of a  formidable American  effort that  gave  the
39

Frenchmen the fright of their lives. The race itself turned out to be a
surprising anticlimax. Of twelve aircraft entered, only six were present
on race day. The remaining two Americans were out at once, and the
result was a lone French win at 168.5 mph. A second French racer also
reached the finish, but only after two forced landings and in sorry con-
dition. Its average speed was 112 mph. A Liberty-engined DH-4
would have done better! The postrace reports pointed out how
extraordinarily delicate were these high-power engines as soon as any-
thing like a sustained effort was demanded of them. The engines in the
three French racers and in the lone British contender all were 300-hp
Hispano-Suizas.
  In the course of the year 1920, in a competition between the French
SPAD and Nieuport companies with racers that were all equipped
with the 300-hp Hispano, the absolute speed record was pushed upward
from 171.5 to 194.5 mph, at which mark the record stood on 31 December.
  The Pulitzer race, the next high-speed event after the Gordon Bennett
fiasco, was won by the same Verville VCP-R racer with the Packard
engine that had proved so unmanageable in France, but this time the
engine was throttled down to 1,700 rpm and the average speed achieved
was 158.5 mph. A 300-hp Wright-Hispano-powered MB-3 Army pur-
suit plane was second, only 8 mph slower than the winner. The
two K-12-powered triplanes proved very fast but both dropped out
with engine trouble—one because of the breaking of a so-called main
engine part, possibly the crankshaft, though no exact information has
been obtained. This then was the end of the K-12 as far as service
flying was concerned. A year later the K-12 made one surprisingly
effective racing appearance, of which more mention will be made.
  Regarding the C-12, during 1920 there was an increasing awareness
of the qualities peculiar to that type of engine—a high rate of rotation
and low weight coupled to a very small frontal area—and of the possible
uses of those qualities for attaining very high flying speeds.
  The whole theory was admirably expounded by Grover Loening in
an article which vividly describes the benefits to be gained from
engines with low frontal area.1 Loening was an early champion of
small fuselages, and he knew what he was taking about. In 1919
he had demonstrated that at 160 mph a 400-hp Liberty had less effec-
tive power than a 300-hp Hispano because the Liberty's greater frontal
area used up 128 hp more to propel a plane at that speed than did the
Hispano.
   1 Grover  Cleveland   Loening,   "Engine   Shape  as  Affecting   Airplane   Operation."
S.A.E. lournal, vol. 6, no. 6 (20 June, 1920) , pp. 417-421.
40

  In his article, Loening referred to the then-current sentiment in avia-
tion circles that small and fast airplanes should be powered by light,
powerful, radial engines cooled by air, while water-cooled engines,
mainly regarded as heavy and bulky but reliable power plants, were
best suited for slow carriers of high payloads This, wrote Loening, was
an error; and he referred to the formula R=KSV2 which gives the rela-
tion of power to speed, where AT is a constant, dependent upon the shape
of the airplane and for all contemporary fuselages of about the same
value, and where S represents the cross-sectional or frontal area.
  Loening explained that the importance of factor S should be plain
to everybody, but he applied himself to convince the doubters by a clear
example based on some reasoning of his own. As a standard he chose
a theoretical airplane powered by the C-12 engine (which shows plainly
enough whose inspiration had enligtened him), said airplane having
a theoretical maximum speed of 232 mph. If, he went on, the weight of
the engine, by some laborious work, were reduced to half the original
amount (a fact of nearly impossible achievement, he hastened to add)
the maximum speed of that airplane would be raised from 232 to 241
mph. If, on the other hand, the head resistance of the engine were
halved, and the weight remained the same, the maximum speed would
take a jump to 290 mph. This showed how much more important the
head resistance is than the weight, and he urged builders to abandon
the unwarranted race for lighter weight per horsepower and suggested
that they start afresh on a new line offering far greater possibilities by
making the shape of the engines more suitable to smaller airplanes.
  A statistical table in Loening's article included data for the frontal
areas of several contemporary engines, giving the value for the 650-hp
Rolls-Royce Condor as 7.32 square feet, for the air-cooled 400-hp
Bristol Jupiter radial as 12.65, the 450-hp Napier Lion as 5.68, the 400-hp
Liberty as 5.16, the 420-hp Curtiss C-12 as 4.66, and the 300-hp Hispano-
Suiza as 4.75. The C-12 clearly was well ahead where horsepower per
square foot of frontal area was concerned. This tabulation is inter-
esting because sooner or later all the engines mentioned would be in
competition with the successors of the C-12. It was, of course, equally
important to include the frontal area of the radiators, and this also had
been anticipated by Kirkham when the 18-B and 18-T fighters—and
later the Cox racers—had their radiators installed on the sides of the
fuselages.
  The C-12 could not be blamed for the unlucky outcome of the
Gordon Bennett attempt in 1920, even though it still was not fit for
service use.   The armed forces again were becoming increasingly inter-
41

ested in what was still potentially a formidable engine for a pursuit
plane. Toward the end of 1920, C-12 engine 4, probably the last built,
was sent to McCook Field for evaluation and testing. McCook Field was
the seat of the engineering division of the United States Army Air Force
and a rather impressive concentration of engineering talent had been
gathered there under Captain (later Major) George Hallett. The gen-
eral idea prevailed that the engineering division should design the
armed forces' own aircraft and engines, using the industry solely for
production with contracts to the lowest bidder, an idea which may have
resulted more from lack of funds than from self-confidence.
  At McCook the C-12 was received with curiosity, and with quite a
show of interest, too, as several engineers, notably William Warner and
Glenn Angle, were former Curtiss engineers who had worked under
Kirkham on the layout of the K-12 engine before they came to McCook
Field in 1918. Captain Hallett had been chosen to accompany Lieuten-
ant Cyril Porte as technical expert on the Wanamaker-Curtiss trans-
atlantic flight of 1914, which had been canceled because of the outbreak
of the war. In fact, everybody concerned had heard since 1918 of the
"super" engine that was being developed at Buffalo.
  The tests of the C-12 at McCook Field formed the basis of Serial
Report 1780 of 25 November 1921. The engine had been submitted to
preliminary tests for obtaining data on power, fuel consumption, and
general efficiency. No run exceeded one hour, and there was no attempt
to have the engine undergo a test of some duration because, when stripped
after the calibration test, it was found to have suffered more than nor-
mal wear—especially in the reduction gearing—with the propeller gear
teeth showing uneven wear. The reduction gearing still was the weak
part of the engine, as no unit had been able to withstand more than 25
hours running.
  Another weakness concerned oil consumption, which was much too
high and had caused fouling of the A.C. spark plugs on the outside of
the cylinder banks. These plugs were replaced with Mosler mica plugs
which gave satisfaction at first but overheated quickly and caused pre-
ignition; so, they, in turn, were replaced with A.C. plugs. A slight water
leakage also was observed. The cause of the excess oil consumption
was traced, in part, to the seven 1-in. bearings which were too narrow
for long life and proper control of the oil flow.
  The inverted Claudel-Hobson carburetors were severely criticized.
They gave low fuel consumption, but only at full-power operation. At
lower engine speeds the specific consumption rose very quickly and there
was danger of flooding and fire because of the external air vents.
42

  The report ended with the statement that the Army's engineering
division was not interested in a geared engine of this size that could not
be adapted to cannon mounting. Tests of such a cannon engine had
been going on since 1919 with an H-type Wright-Hispano, but never
with successful results.
  The C-12 tested, rated at 400 hp/2,250 rpm, was shown to give a maxi-
mum of 427 bhp at 2,250 rpm on a compression ratio of 5.5:1. The
weight of the tested engine was 698 lb, or 712 lb with the necessary
auxiliaries. No attempt was made to run the engine at speeds over
2,300 rpm.
Development of the Curtiss CD-12
  It was by the Curtiss Company rather than by the military that the
policy in regard to the future development of the C-12 was to be shaped.
After Porter left, the company for a while was again without an engi-
neer in charge of motor development, but then an engineer of supreme
ability took over—Arthur Nutt, who would influence decisively the
C-12's future career.
  Arthur Nutt had entered the Curtiss Company after his graduation
in 1918, and in July of that year he started in the experimental motor
department as test observer; and later, as test engineer, he became inti-
mately connected with the development of the K-12. In 1919, after
Kirkham left, he became motor engineer under Porter and finally, after
Porter's resignation, he was appointed chief motor engineer in charge
of engine development. His talents were linked with Curtiss engines
until the 1929 merger and then with Wright engines until October 1944
when, as Dr. Nutt, he resigned from his position as vice-president of
Wright Aeronautical to join the Packard Motor Car Company. He
remained with Packard until it discontinued its aircraft engine activities
in 1949.
  When Arthur Nutt assumed charge of the engine department at
Curtiss he was only 26 years old. Yet, since the future of the company's
only high-power product, the C-12, was at stake, he was at once called
upon to make some important decisions and to offer advice on other
questions.
  Clearly, the needs of Curtiss were to stop costly experimenting and to
start selling, or else abandon the C-12 design entirely. The first step,
which was to be reached in the shortest possible time, was to make the
43


FIGURE 28.-Arthur Nutt (1895-).
C-12 reliable enough to stand an official 50-hour test and then sell it to
the only available customers, the armed services. It was obvious that
with the enormous amount of cheap Liberty engines around there was
no question of entering a nearly nonexistent commercial market. Even
the Eagle had ended by being powered with a Liberty engine.
  Arthur Nutt put forth his ideas after consultation with William Gill-
more, chief aeronautical engineer. The Curtiss management agreed
that there was no time to develop the reduction gearing to suitable
reliability, so the best and only thing to do was to dispense with the
gears altogether. This entailed a lowering in rotational speed of the
crankshaft to retain propeller efficiency and, in fact, a complete derating,
which would further increase reliability.
  The Curtiss officials were confident that even in derated form the
C-12 would still remain a competitive product. Without reduction
gears it would weigh 650 lb, against the 300-hp Wright-Hispano's 625
lb and Liberty's 850 lb, which, in high-compression form, gave a very
theoretical maximum of 420 bhp. The decision to derate the C-12
could, of course, have been taken much earlier, and even applied to the
K-12, but the continuous obsession with highest power had prevented it
heretofore.
Both services were interested in the design but were then severely
44

short of funds, so an agreement was reached to divide the expense of
the development of the new model. The Navy agreed to buy one derated
C-12, whose power was provisionally taken from a point down the
original K-12 curve and set at 325 hp at 1,800 rpm crankshaft speed.
The Army agreed to do the testing at McCook Field, and to participate
in the development; thus, it would both gain experience and help pay
its share of the costs.
  On 19 October 1920 a supplemental agreement was signed with Curtiss
by the Navy, amending contract 46553 for one C-12 engine without
reduction gears at a total cost of $31,000—no small sum for an aero engine
at that time. Contract 46553 had been pending with Curtiss since 1918
for the furnishing of four K-12 engines, but because none of those
engines had passed its official acceptance test a report of September 1920
still mentioned that contract as zero percent completed. The rating of
325 hp was quite acceptable to the Navy, whose low-compression Liberty
engines had an output of about 330 hp at sea level and were much bigger
and heavier than the C-12.
  The first C-12 Direct-Drive, or CD-12, as the engine was renamed, was
shipped from the Curtiss plant to McCook Field for testing on 25 Feb-
ruary 1921. It differed from a C-12 only by having all of its spark plugs
on the inside of the V, doubtless to avoid the oil fouling of the plugs
on the outside of the cylinder banks, from which the C-12 had suffered.
The rest of the design was kept unchanged, even as to the Claudel car-
buretors.   (Figure 29.)
  There was some bickering at first between the Army and the Navy
over the duration of the test to which the new engine should be sub-
mitted. The Navy insisted on 60 hours but the Army contended that
50 hours was enough. A telegram dated 9 March states that if there was
no agreement on 50 hours the Army would ship the engine immediately
to the Navy and proceed with work on the geared C-12, from which it
appears that the engineering division at McCook Field was still hoping
to see the C-12 brought to satisfactory status. It was finally agreed that
the engine would be submitted to a preliminary ten-hour test for the
Navy, followed by a normal 50-hour test for the Army. These tests are
described in McCook Serial Report 1853, dated 10 January 1922.
  The ten-hour run for the Navy was made on a low-compression ratio
of 5.23:1 in two of the usual five-hour periods. The test of CD-12
engine 1 (Bureau Number 9673) started on 4 April and was completed
on 21 April 1921. The power developed on the dynamometer prior to
the start of the ten-hour test was 335 bhp at the normal speed of 1,800
rpm, 360 bhp at 2,000 rpm, and 368 bhp at 2,100 rpm.   After 8 hours
45

57 minutes running, the failure of a spring collar retaining nut caused
one of the intake valves to drop into number 3 cylinder. The repair was
made and three new cam followers were installed at the same time. The
ten-hour test for the Navy then ran to completion.
  After this test the engine was torn down for complete inspection prior
to the 50-hour Army endurance test, and it was found to be still in excel-
lent condition. This gave rise to an optimistic mood, and it was decided
to let the 50-hour endurance test be run under more severe conditions.
The compression would be raised by installing different pistons and the
normal speed would be 2,000 rpm instead of 1,800. The Army 50-hour
test started on 4 May and ended 24 June. It was the first 50-hour test
to be completed since the inception of the design four years earlier.
  At the end of each five-hour period the engine was stopped and checked
and minor adjustments were made when necessary. Because of the
higher power developed and the duration of the test, the engine did not
do as well as in the Navy test. During the warmup two of the new high-
compression pistons seized. It was found that the compression ratio
was 5.65:1, and, as this appeared too high, all the pistons were again
replaced with ones giving a compression of 5.37:1.
  In the course of the test there were numerous forced stops, and advan-
tage was taken during these times to redesign or strengthen all the parts
that failed, though all major parts held. The performance of the engine
remained good and it ran very smoothly throughout the test.
  During the calibration test, preliminary to the endurance test, the
power developed at the normal speed of 2,000 rpm was 383 bhp; the
maximum reached was 393 bhp at 2,090 rpm. The average power devel-
oped during the full-throttle runs of the 50-hour test was ,367 bhp at
1,976 rpm. The fuel and oil consumption was good. When the engine
was torn down for final inspection after the tests it was discovered that
two of the lower bearing caps had failed just before the end of the test,
but the bearings were still in very good condition. The overall wear
of the engine, especially of the valves, was such that the engine would
have been due, shortly afterwards, for a complete overhaul.
  The Navy had at once placed an order for two more CD-12 engines
(Contract 53933) destined for new service aircraft in replacement of the
Wright-H (the 300-hp Wright-Hispano), but plans soon were formed to
use the CD-12 for racing. This shows that even then, in derated form,
the CD-12 was thought to be an unbeatable proposition for high-speed
work. The ordering of two new racing aircraft from Curtiss by the Navy
in June 1921 may be considered to represent the opposite pole of the
attitude that had led to the construction of the Cox racers and a direct
46


FIGURE 29.-Views of Curtiss CD-12 aero engine of 1922, which had an electric starter.
(Smithsonian photos A-4781B, A-4781.)
47

result of the experience gained and the disappointments suffered at the
1920 Gordon Bennett race.
  Due respect was now given to a low landing speed. The Navy had
placed a ban on anything above 75 mph. The new racers, therefore,
were biplanes with an appreciable wing area and, it may be presumed,
with a strong landing gear. In fact, they relied nearly exclusively on
the power and possibility for aerodynamic penetration of their CD-12
engines—following Loening's teaching—for the attainment of adequate
speeds in competition. The history of the first Navy Curtiss racers or
CR airplanes is not to be related here; suffice it to say that after ordering
two planes in June the Navy lost interest but permitted one of the
planes to be entered privately by the Curtiss Company for the 1921
Pulitzer races, which were held in November.
  The CD-12 meanwhile had shown definite promise for improved
reliable output and, by being pushed a little, had been able to develop
405 bhp at 2,000 rpm after the compression ratio had been raised to
6.1:1 by using a fuel with a high percentage of benzol. Whereas the
power of the latest CD-12 had not risen above that of the original K-12,
it now was achieved at 2,000 rpm instead of at 2,500, showing how appre-
ciable an increase in mean effective pressure had been reached in two
years.
  The Pulitzer race of 1921 was won convincingly by the Navy Curtiss
racer (Figure 30) piloted by Bert Acosta at 176 mph. Second place was
taken by the other 1920 Gordon Bennett racer, the Cactus Kitten. This
racer still was powered with the geared C-12, giving some 435 hp, but
the airplane was changed beyond recognition by the application of a
set of impressive-looking triplane wings. These, suitably clipped, had
been taken from one of the early triplanes. With these wings it was
hoped to keep the landing speed down to 70 mph. The Cactus Kitten
appeared to be the faster of the two Curtiss entries. Its maximum speed
was evaluated at 196 mph, but the CR won, due presumably to its supe-
rior handling qualities. The Cactus Kitten nevertheless achieved the
fame of being the fastest triplane ever built.
  Third place was taken by a Thomas Morse MB-6 powered by the
inevitable competitor, the Wright-H, of well over 300 hp now. Since
the previous January Wright had dropped the "-Hispano" addition to
its name, showing how far the design had been Americanized; indeed,
the Americanized version was fast becoming better than the French
original! There was another Curtiss engine entered—an old 400-hp K-12
whose swan song this race would prove to be. Powering an Ansaldo
Balilla, it had won the American Legion Derby a few days earlier, and
48


FIGURE 30.—Navy Curtiss Racer CR-1, powered by Curtiss CD-12 engine, 400 hp.
This plane, piloted by Bert Acosta, won the Pulitzer trophy at Omaha, Nebraska,
on 3 November 1921.  (Smithsonian photo A-4624B.)
it now finished the Pulitzer race without trouble, securing fourth place.
  Cox had tried to enter his repaired Texas Wildcat biplane too, but
it was ruled out because of its high landing speed. The result of the
race showed that three of the first four placed were powered by Curtiss
engines, one of each model developed since 1918. The advantage of
low frontal area, coupled with high power, again was proven conclusively.
  Only four CD-12 engines had been built, and three of these were sold
to the Navy. Engine 2 was retained by Curtiss as a prototype to prepare
for series production, and plans and specifications were drawn up together
with a comparative study of the CD-12 versus the Liberty and the
Wright-H. The specifications offered the engine as the CD-12A in both
a low- and a high-compression version, rated at 375 and 400 bhp respec-
49

tively, both ratings at 2,000 rpm for a weight of 700 lb dry. The limit
of 2,000 rpm, imposed by the propeller, could not be raised at that time.
  Other than the United States Navy, there was no prospective pur-
chaser, but in England, as early as March 1921, Frank Barnwell, the
celebrated designer of the Bristol Fighter, suggested the adoption of a
Bristol Scout*F airframe—hitherto powered by radial air-cooled engine
types—as a flying testbed for one of the new Curtiss engines. His pro-
posal was not accepted, but in view of the outstanding developments in
Britain, to be described later, it provided a touch of the prophetic.
  Two CD-12 units were used by the Navy as power plants for the new
Curtiss Torpedodropper, or CT (A-5890), in replacement of the originally
installed Wright-H engines which were prone to excessive vibration.
The prototype CT was delivered to Anacostia in January 1922 and fitted
with CD-12 engines 3 and 4, but before June both engines, one in dam-
aged condition, had been shipped back to the Curtiss Company for
reconditioning. Nine CT planes had been ordered, but by 1 July the
eight remaining planes of that order were canceled as being "of doubtful
value for Naval Purposes."
  After the 50-hour test, CD-12 engine 1 was shipped back to Curtiss on
8 September 1921 and was the object of negotiations for converting it
from a test engine into an operationally useful unit at minimum cost.
The result was BuAero contract 55792 of 14 March 1922, ordering a
variety of changes—including correction of the weaknesses encountered
during the tests—to be made at a cost of $1,625. The principal changes
were a new intake manifold, a Ball and Ball carburetor, a new water
pump impeller and shaft, new camshaft drive shafts, and gears of new
design. During conversion it was decided to install high-compression
pistons so the engine could be used for racing.
  During the summer of 1922 the Navy decided to use two CD-12 engines
to stage a comeback of the 18-T triplanes. These, after the breakdown
of the Navy's K-12 engines at the 1920 Pulitzer races, had remained in
storage and were now being reconditioned again as single-float seaplanes
at the Naval Aircraft Factory in preparation for their use in competition.
Their maximum speed was calculated to be 131 mph. Reconditioned
engine 1 was fitted to A-3325, while one of the later engines was installed
on A-3326. The two triplanes then were entered for the 1922 Curtiss
Marine Trophy race on 8 October 1922. One of these, piloted by Lieuten-
ant Sanderson, was nearing victory when it ran out of fuel in the last lap.
After the race A-3325 was sent to Selfridge Field, where it was wrecked
shortly afterwards. The second triplane, which remained in service for
another year, will be referred to later.
A-3325's undamaged engine, CD-12 engine 1 (Bureau No. 9673), was
50

in the Navy Museum in Philadelphia until 1926, when it was shipped
to the Smithsonian Institution. This was the first engine to pass a
50-hour test at 375 bhp. Certainly it is worthy of preservation, as it
pointed the way for the history-shaping engine that will now be discussed.
The Curtiss D-12 Engine
In July 1921, after the CD-12 finally showed itself capable of passing a
service test, it then was possible to put the engine into production with
a fair possibility of sales, which in the case of the Navy had already
begun. However, the experience gained had shown Arthur Nutt and
the engineering staff at McCook Field, where most of the testing of the
C-12 and CD-12 had been done, that the design could be reworked into
a really good engine by correcting the few basic weaknesses still remaining.
  After the official test of the CD-12, the engineering division at McCook
Field made several recommendations that were detailed in official letters
to the Curtiss Corporation. The first such letter was dated 22 August
1921, and the Curtiss Corporation replied on 6 October; a second letter
from McCook Field was dated 14 October. The principal criticisms
and recommendations were as follows, (a) The crankcase and oil system
should be changed to conventional dry sump, (b) The crankshaft main
bearings, especially bearings 2, 4, and 6, were considered too narrow. This
condition could be remedied only by lengthening the bearings or by
using a four-bearing crankshaft like that on the K-12. (c) The Claudel
inverted carburetor was considered unsatisfactory. An approved make
of upright carburetor was recommended to be fitted inside the V of the
engine and, if necessary, after redesigning the intake manifold. (The
Navy had already returned to the K-12 carburetor for its first CD-12.)
  The engineers at McCook Field had gone to great lengths in order to
prove their points, and, as a result of a detailed study comparing the
C-12 and K-12 crankshaft construction, Report 1996 was issued. This
report contained a complete calculation of all the crankshaft stresses
and bearing loads of the K-12, C-12, and CD-12 engines and a theo-
retical direct-drive KD-12 engine with a four-bearing crankshaft. The
conclusions showed that Porter's seven-bearing crankshaft was inferior
to Kirkham's four-bearing one because the seven bearings were of insuf-
ficient surface, especially the central bearing, the one most heavily
loaded.
Kirkham later improved his crankshaft by installing more counter-
51

weights, and this improvement caused McCook Field's engineers to
recommend it for future development. The K-12-type crankshaft ap-
peared amply sufficient, especially for a direct-drive engine run at not
more than 2,000 rpm.
  The report by McCook Field's engineering division allowed the Curtiss
engineering staff two alternatives. One alternative was the simple one
of adopting the K-12-type, four-bearing crankshaft on a CD-12-type
engine, and this would lead to a satisfactory 375-hp engine in the shortest
possible time. The other alternative, appearing to be a longer and more
arduous undertaking, was to redesign the crankshaft as a seven-bearing
unit with adequate bearings for which McCook's engineering division
specified the minimum lengths that would be satisfactory. This second
alternative would necessitate a complete redesigning of the cylinder
blocks due to the respacing of the cylinder center distances.
  The difficult but more rewarding policy was chosen by Arthur Nutt. On
13 October 1921 he submitted to the engineering division at McCook
preliminary drawings of a new crankshaft having adequately propor-
tioned main bearings. The seven one-inch bearings were replaced with
six bearings of VA inches and a central bearing of 1% inches. Future
experience would show that the width of the new bearings was sufficient
to cope with the increased loads of the later, larger-bore developments.
  The cylinder blocks were completely redesigned, as were the connect-
ing rods, the cylinder attachments, and all other accessories that had
given trouble or that could be lightened without impairing reliability.
The lubrication oil was taken from a separate oil reservoir through an
oil cooler instead of from the crankcase sump. This construction, cur-
rently called dry-sump lubrication, would permit a smaller and stiffer
crankcase, a more effective lubrication by control of the oil temperature,
and a small gain in reduced frontal area.
  The new design required new patterns and castings, and the new
crankshaft, which was drop-forged, required the sinking of massive dies
by the Wyman-Gordon Company of Worcester, Massachusetts. Carbu-
retion also was completely modified. Frank Mock, of Stromberg, went
to Buffalo in January 1922 to lay out and design a new Stromberg car-
buretor for the new Curtiss engine. The patterns were made at Curtiss
under his supervision, while the castings, machining, and assembly were
effected at the Stromberg plant. The result was the Stromberg NA-Y5,
which was to become standard for all D-12 engines.
  The Stromberg carburetors were not ready to be installed on the first
engines in time for their tests. These engines were fitted with American
Zenith 54-mm carburetors which, following the rejection of the earlier
Claudels by McCook Field's engineering division, had been chosen for
52

installation on the CD-12A engine that was not put into production.
The magnetos of all early D-12 engines were Splitdorf SS-12s.
  This, then, was the engine that was to achieve fame as the Curtiss D-12
and was to have a profound influence on subsequent aero engine design.
At first it was regarded as an improved version of the CD-12 and was
rated accordingly, 375 hp at 2,000 rpm, but it soon would show itself
capable of much better outputs and much higher crankshaft speeds.
The limits imposed on the propeller by the tip speed prevented higher
speeds for a time, but a new development in propeller design by S. A.
Reed would finally put the D-12 in a class far above its contempories.
  The D-12 was not only a mechanical masterpiece; it also was a beauti-
ful engine, with a grace that derived from the simple and yet purposeful
form, and its lines set a fashion for all high-power, liquid-cooled power
plants down through World War II. Henceforth, it would be impossible
for any engine with individual liquid-cooled cylinders—no matter how
much it might be based on up-to-date concepts—to overcome the impres-
sion of being somehow "antique."
  The Navy ordered twelve engines straightaway (Bureau Contract
55824), even before the first D-12 had passed the regulation 50-hour
endurance test. Two D-12s of that contract were to be delivered to the
Army. This first order was followed by another in April and a third
soon afterwards. By the first of July 1922 there were 57 additional
engines on order under Contract 56230.
  The first D-12 engines received by the Navy were rated at 350 hp at
1,800 rpm. The contract weight, limited to 720 lb with reference to the
latest improved CD-12, was found to have been reduced by 52 lb. The
Navy did not regard its first D-12 engines in their true capacity as power
plants for high-speed or fighter airplanes but used them as a sort of
smaller and improved Liberty on a series of newly evolved Torpedo or
Observation aircraft which barely passed the 100-mph mark. These
aircraft were the Douglas XNO-1, Martin MO-1, and Martin M20-1;
also, the Curtiss CT ended its career with two D-12 engines installed.
  The Army did not relish being left with only two D-12 engines, and
the Navy's monopolizing of the D-12 production was not well received.
The Navy, of course, considered itself the principal beneficiary by right,
as, from the K-12 onward, Curtiss high-power engine development had
been under some form of Navy sponsorship while the Army had partici-
pated mostly by evaluation of design and by testing, not by actual pur-
chases. Under an agreement eventually reached by the two services the
Army Air Corps would receive 27 of the 57 engines ordered by the Navy
under Contract 56230. Afterwards, production would be able to supply
both services.
53

  The first D-12 delivered to the Navy completed its preliminary accept-
ance tests and passed the official 50-hour endurance test between 12 April
and 27 May 1922. It was submitted to an additional 250 hours of testing,
as was Navy practice at the time, and then to several more full-load tests,
bringing the total testing time to 275 hours on 19 July 1923. The engine
then was set up at Buffalo for an additional 200-hour duration test at
part load. After 88J/2 hours running, this test ended on 3 August when
the rear left piston broke and the corresponding connecting rod assembly
at once went through the crankcase and wrecked the engine. After dis-
mantling, all remaining pistons showed cracks near the bosses, but the
other parts were found in very good to excellent condition. A new
stronger piston was designed for all remaining engines.
  There was unforeseen delay in the acceptance of the remaining eleven
engines of Contract 55824 because the Navy claimed to have found an
excessive quantity of nonmetallic material in the forged crankshafts.
Neither Curtiss nor the forger agreed with the Navy about this, but they
had to replace the shafts.
  As we have seen, the Army received in 1922 two of the first twelve
engines ordered by the Navy. The Army was more alive to the poten-
tialities of the D-12, and it was eager to test its capacity as a high-speed
engine.
  Two Fokker fighters, still crowned with the laurels of a famous name,
then were on order by the U.S. Army Air Forces. These fighters, powered
by Wright-H 300-hp engines, had been allocated Army Air Forces desig-
nations PW-5 and PW-6 (PW means "pursuit watercooled"). In August
1922 a contract was placed in Holland for a third Fokker to be designated
PW-7 and powered by the D-12 engine. It was hoped that the Fokker/
D-12 combination would become the most advanced fighter airplane of
its time.
  There also were some thoughts regarding modification of the PW—1
fighter, with its already obsolete Packard engine, so that it would accom-
modate the D-12 with a nose radiator, but nothing came of that.
  Before the Army's first two D-12 engines were used as prototype fighter
engines they showed their true pace as racing engines when two high-
speed airplanes were ordered from Curtiss to compete as R-6s in the
1922 Pulitzer races, which were to be held in October at Detroit and
where the D-12 would make its first public appearance. For these two
new racers, the Curtiss engineers exercised utmost care to eliminate all
unnecessary drag and went to the extreme of incorporating the radiators
as part of the wing surfaces.   Otherwise, the new Army R-6 racers were
54



%'-*fr*m


FIGURE 31.—Curtiss R-6 Army racer powered by Curtiss D-12 engine, 450 hp, 1922.
(Smithsonian photo A-4624E.)
still fully braced biplanes with fixed undercarriages (Figure 31) and low
landing speeds.
  The race ended in a sweeping victory for the Army and a decisive one
for the D-12 which secured the first four places and was hailed as the
speed champion of its time. The two Army R-6 racers, two of the most
beautiful airplanes ever built, came in first and second, followed by the
two Curtiss Navy racers. The winners' average speed was 206 mph, 30
mph more than the 1921 winner had achieved. The exact horsepower
developed by the victorious D-12 engine was not divulged and contempo-
rary press reports differ widely in that respect; but as a small-diameter
55


FIGURE 32.—Wright T-2 engine, 700 hp, about 1922.  (Smithsonian photo A-4624F.)
wooden propeller capable of withstanding a speed of 2,200 rpm had been
fitted, the D-12 was able to deliver the 440/450 bhp at that speed when
fitted with high-compression pistons and running on a 50 percent benzol
mixture.
  This 1922 Pulitzer Trophy race was not only an interesting event
because of the D-12's performance; it also was noteworthy for the com-
petition that was offered by other engine manufacturers.
  The greatest effort for the race had been made by the Wright Aero-
nautical Corporation, still the principal purveyor of fighter engines.
That company had had two aces up its sleeve. The first was a refined
high-compression model of its 300-hp, eight-cylinder model H that pro-
duced close to 400 bhp. These engines were fitted to several racers, two
of which—Verville-Sperry R-3's with retractable landing gears and mono-
plane wings—finished fifth and seventh. Approximately 400 bhp would
prove the maximum of which the model H engine was capable, and that
engine's normal proneness to vibration was much aggravated when run-
ning with high-compression pistons. But the Wright engineers had
developed their second ace, a big new V-12 of 1,950 cubic inches
capacity. This engine—also of monoblock construction with dry liners,
following the original Hispano-Suiza concept—had open-end sleeves and
four valves per cylinder.    Known as the T-2,  it had already been
56


c*Xf
FIGURE 33.-Packard aeroengine l-A-2025, 630 hp, about 1920-1922.
(Smithsonian photo A-2325.)
accepted by the Navy for heavy-duty service rated at 525 hp.
  The Navy, always on the alert for healthy competition between manu-
facturers, had eagerly backed the construction of a Wright racing air-
plane to be powered by a high-compression version of the T-2 engine
(Figure 32) that developed over 700 bhp. This racer, first known as the
Navy Mystery because of the secrecy that surrounded its building, was a
dubious monoplane or half-hearted biplane. It was built along the lines
of a Nieuport racer, to which the French had applied the designation
Sesquiplan, meaning one-and-a-half plane. The T-2 engine proved
inadequate to demands for high power, and the Mystery was retired from
the 1922 Pulitzer race after making only one lap at high speed.
  Another firm still competing for honors was Packard. Its 2,025 cubic
inch Gordon Bennett engine of 1920, now reliable and good for about
57

630 hp (Figure 33), was installed in several racers, the oldest of which
was the Verville R-l that had won the first Pulitzer Trophy in 1920.
The Verville could place no better than sixth, while the other, and newer,
Packard-powered racers finished in the last places. Thus, it was again
brought home that brute force alone was not sufficient to win.
  The 1922 Pulitzer race was to be the last appearance of the big Packard
engine, as the company found that it needed something more refined if
it wanted to continue in competition against the new Curtiss and Wright
monoblock engines. The result was that Packard soon would make a
great effort to come up with its own advanced design.
  The apotheosis came a few days after the race when General Billy
Mitchell, using the Army's Pulitzer winner, brought the world's speed
record to the United States for the first time with a mean speed over
four runs of 223 mph. Lieutenant Russell Maughan, the Pulitzer winner,
had attained the unofficial speed of 220 mph before the race when testing
the speed of the new plane.
  Speed, that most glamorous of all records, had been pushed upwards
very slowly after 1920: first to 205 mph in September 1921 by the intro-
duction of the Nieuport Sesquiplan of high wing loading and landing
speed, and a year later to 212 mph by the same Nieuport, powered by
the same Hispano-Suiza and manned by the same pilot, Sadi-Lecointe,
who had become a French national hero. Sadi's second record was made
only a few days before Maughan reached the 220 mph mark.
  Toward the end of 1922 the Army asked for bids on a new ambulance
plane, and Curtiss immediately proposed the Eagle again. This ubiqui-
tous transport already had been considered for the nonstop crossing of
the American continent that was to be performed by a Fokker T-2,2 and
Curtiss now wanted to sell the Eagle as an ambulance plane with a D-12
engine. The Army concluded that the Eagle indeed appeared to be the
most suitable design from a medical standpoint but, with a touch of
regret, added that the planes would have to be fitted with Liberty engines
because so many of these engines were on hand. And so it happened.
The Liberty-powered Eagle was converted into an ambulance plane in
1923, but soon afterward its career ended in a crash.
  After the Pulitzer races of 1922 the Army's two D-12 engines were
removed from the racers and sent to Curtiss for reconditioning. One
of the engines then was shipped to Holland for installation in the Fokker
PW-7, which, with the Curtiss engine, showed a maximum speed of 151
mph, against 138 mph for the earlier two Fokkers with their 300-Wright
   2 See Louis S.  Casey, "The First Nonstop Coast-to-Coast Flight and  Historic T-2
Airplane," Smithsonian Annals of Flight, no. 1   (1964) .
58

engines. But the PW-7 was not to be put into production. The first
new fighter built expressly around a D-12 engine was, naturally, a
Curtiss product, and it received Army designation PW-8. The proto-
type was ready early in 1923, showing that the Curtiss designers had lost
no time. The first PW-8 machine was based on the winning Pulitzer
racers to the point of having wing radiators. In service trim and with a
high-compression D-12 engine, it achieved a maximum speed of 170 mph.
  Also in 1923, Boeing brought out the PW-9, which was to become a
worthy competitor of the Curtiss fighter. The Thomas Morse Company,
manufacturer of the standard Army Pursuit MB-3, received a D-12 on
loan for installation in its new TM-22 prototype; but that plane was
abandoned early because it proved to be too small and impossible to
handle.
  Wright tried to counter the success of the D-12-powered fighters with
the introduction of an all-metal fighter built around the 700-hp T-3
engine. This fighter made its first flight during 1923 but was not able
to out-perform the Curtiss-powered planes. The Wright competition for
high-performance, aluminum-monoblock engines would soon come to
an end, but Wright was to attack on another front, as its H-3 engine
was to gain a belated victory over the D-12 in the struggle for the world
speed record.
  When the French learned that the Wright-H engines were able to
deliver much more power than their own Hispano-Suizas, they took a
drastic step. In 1922 they imported one of the American Wright-H
engines, capable of 400 bhp at 2,000 rpm, and installed it in a new Nieu-
port (model 37) racer. When this plane proved unsuccesesful the Sesqui-
plan racer was fitted with the Wright engine, or one closely patterned
after it. Wing radiators were installed, and the wings were clipped to
the bare minimum with the handicap of a 112-mph landing speed. With
this modified Sesquiplan, Sadi-Lecointe was able to raise the speed record
in February 1923 to 233 mph and beat the American marks established
in 1922.
  The French record was to last only a few weeks. Early in 1923 McCook
Field received the first D-12 engines fitted with the new Stromberg
NA-Y5 carburetors. The Strombergs were found to be a big improve-
ment over the earlier Zeniths as they functioned perfectly up to 2,500
rpm, while with the Zenith the D-12 would not function over 2,300 rpm.
On 29 March 1923 Lieutenant Maughan established a new world's record
of 244.9 mph with a biplane R-6 racer powered by the Pulitzer-winning
D-12 with Stromberg carburetors and a new all-metal propeller. The new
propeller was the result of the experimental work of Dr. Sylvanus Albertus
59

Reed, who was to add his fame to that of Dr. Nutt in helping to boost the
D-12 to a new level of performance by permitting it to unleash its power
at high speed without having to use the troublesome reduction gear again.
In view of this development's importance, its story is included here.
  In 1915, while engaged in research on high-frequency acoustics, Dr.
Reed was employing an apparatus with a shaft revolving at 36,000 rpm.
It occurred to him that it would be interesting to conduct similar experi-
ments with high-speed propellers. He was fully cognizant of the prob-
lems that were believed to arise when any object moving through air
approached the speed of sound, a belief based on the experience from
tests conducted with artillery shells. It was known that at or near the
speed of sound there is a critical point on the curve of the relation to air
resistance, but Reed professed a truly scientific attitude of mind by not
believing anything before it had been proven—all the more as he had
been told that the subject of very high propeller tip speeds was a field
that had not been explored.
  Reed thought that the difficulties encountered when the speed of
sound was reached might well be dependent upon the shape of the
moving object. He then proceeded to conduct a long series of experi-
ments, first with propellers 20 inches long which he revolved at 14,000
rpm—ascertaining, perhaps a bit rashly, that the supposed physical limit
at the speed of sound did not exist and that the thrust of his propellers
did not undergo any appreciable variation when exceeding that limit.
  This was truly a revolutionary theory at that time, considering that as
late as 1919 the British Advisory Committee for Aeronautics had issued
an official report of its own experiments in which it was affirmed that
thrust practically disappeared at high tip speeds, a conclusion corrobo-
rated by American tests conducted at McCook Field. Nevertheless, Reed
was able to interest officials of the Curtiss Company in his research,
especially William Gilmore, chief airplane engineer in charge of airframe
design, who placed full facilities for further testing at Reed's disposal.
  The final result of Reed's experiments was the Z-l propeller of 1921,
which had very thin tips and razor-sharp edges. In fact, Reed had antici-
pated what the aeronautical engineers would discover 25 years later when
the speed of sound would be approached by full-size aerofoils. To be
able to obtain the necessary thin section, the Reed propeller was made
of aluminum in one forging. It gave excellent results, being capable of
very high efficiency at hitherto unheard-of speeds of rotation. It became
the perfect associate for the direct-drive D-12 engine, which seemed
eager to turn at speeds higher than the 2,000 rpm to which it had been
restricted by the older types of airscrew.
60

  With all activity directed towards the winning of the Pulitzer races,
the speed records, and the contract with Fokker, it was May 1923 before
a D-12 engine could be spared for a duration test at McCook Field. The
first engine to arrive was D-12 No. 22, A.S. 23-112, fitted with low-
compression pistons. Shortly afterwards it was joined by the Pulitzer
winner and holder of the speed record, an engine which was to be sub-
mitted to a high-compression (CR. 5.7:1) test. The racing engine was
wrecked during a preliminary fuel consumption run when a master con-
necting rod broke and ruined the crankcase.
  The low-compression (CR. 5.3:1) engine was submitted to a complete
50-hour test on the dynamometer (McCook Serial No. 2266). The test
was conducted in July 1923, so the engineering division at McCook Field
took over testing of the D-12 at the time the Navy finished its testing of
that engine. The Army test was tougher than the earlier Navy ones
because the engineering division at McCook Field aimed at once at high
power and allowed the engine to be run at speeds up to 2,310 rpm, at
which speed 428 bhp was developed. The engine stood up satisfactorily
to the tests, but at the end of the performance it was found that a cylinder
block had developed a light crack.
  The same weakness also became apparent when a new Curtiss PW-8
fighter was used for an attempt at a flight that was to become mem-
orable as the "dawn-to-dusk flight." This was an effort to cross the
American continent during daytime, a feat that could be performed
only by an engine and plane capable of maintaining racing power and
speed for a long period of time, a much tougher undertaking than the
short burst of speed necessary to establish a record or win a race.
  The first PW-8 plane was prepared for Lieutenant Maughan, the
winner of the 1922 Pulitzer race and the pilot who had established a
world speed record in March 1923. Maughan took off for the big try
on 10 July 1923. In contrast to the transcontinental flight of the Fokker
T-2 a few months earlier, the PW-8 flight was organized with inter-
mediary stops, but with the engine full out during the flying laps. This
first attempt ended when Maughan was forced down 15 miles from the
intended stop at St. Joseph, Missouri, because of a clogged gasoline line.
He had covered 1,330 miles in nine hours.
  A second attempt was made on 19 July. When over North Platte,
Nebraska, Maughan noted that a layer of oil was forming on the cockpit
floor. He flew on, and after the scheduled landing at Cheyenne,
Wyoming, a leak in the oil cooler was repaired. But after taking off
from Cheyenne the trouble appeared to get worse, and Maughan landed
at Rock Springs, his cockpit covered with an inch-deep mixture of water
61

and oil. He had already covered two-thirds of the total distance at
racing speed. A close inspection showed that two cylinder blocks had
cracked at the rear so that the engine was losing a gallon of water per
minute and was on the verge of seizing. Maughan then had to postpone
his gallant effort for a year.
  Meanwhile, there was another challenge overseas—the race for the
1923 Schneider Trophy, a tough, high-speed event for seaplanes preceded
by navigability trials. The United States Navy originally entered three
planes. Two were float-mounted CR planes of 1921 and 1922 Pulitzer
vintage. Each of these two aircraft was fitted with a finely tuned, high-
compression D-12, a Reed propeller, and wing radiators (see Figure 34).
With these modifications, the planes were designated CR-3s. The third
Navy entry was the NW-1, the 1922 Mystery which had been turned
into a biplane, on floats, with a Wright T-2 engine of some 700 hp.
  The announcement of these entries created a stir among the tradi-
tional European participants, especially in England where a special effort
had resulted in the winning of the event in 1922. For the 1923 race
England entered a light Supermarine flying boat based on the 1922 win-
ner and powered by the Napier Lion. France was relying on the 300-hp
Hispano-Suiza.
  The Navy's NW-1 crashed during the preliminary tests when the en-
gine disintegrated after 20 minutes of flight. The demise of this racer left
the two diminutive CR-3 seaplanes as the only American challengers.
The CR-3s looked trail against the assembled flying boats of Europe,
but the D-12 engine, running on a 50 percent benzol mixture, was able
to turn out 475 bhp at 2,300 rpm, with the Reed propeller operating at
high efficiency. The British favorite, however, with 560 bhp and a
geared-down propeller, surely was not lacking in power.
  The result of the race was surprising. The two D-12 engines ran
perfectly and the Curtiss planes made a clean sweep, attaining speeds
that left everyone stunned. The average speed of the CR-3s was 30 mph
more than that of the 1922 winner and 20 mph above that of the third-
placed Supermarine, which was clearly outclassed in spite of its greater
power. The victory had been achieved largely through the D-12's small
frontal area, high power, and uncommon reliability. This race had a
profound effect on British military aircraft and engine design, as we
shall see later.
  The year 1923 witnessed another D-12 victory for the Navy when two
Curtiss biplane racers model R2 C-l, powered by a new slightly enlarged
D-12 A engine capable of 500 bhp, took first and second places in that
year's Pulitzer race, the winner flying at a formidable 243 mph.   This
62


FIGURE 34.—Curtiss CR-3 seaplane racer, 1923, powered by Curtiss D-12 engine.
(Smithsonian photo A-4655F.)
event took place a few days after the Schneider Trophy victory. Third
and fourth places were taken by two new Navy Wright fighters (at 230
mph) powered by improved Wright T-3 engines capable of 780 bhp on
the test bench. All other finishers were powered by D-12 engines. This
was the last time the Wright-T engine was used for racing. Henceforth
it would be found, in derated form, only on flying boats and other heavy
aircraft.
  In a two-day contest between Navy Lieutenants Brow and Williams
after the Pulitzer race, the world speed record was spectacularly boosted
to 266.6 mph, thanks to the new R2 C-l and the D-12A. This record
was to stand unchallenged for more than a year.
  The last of Kirkham's triplanes, the A-3326, was entered for the
Liberty Engine Builders Trophy, which preceded the 1923 Pulitzer race
by a few days. While running well in the race, the triplane's overextended
CD-12 engine broke its crankshaft and was completely wrecked and
caused the A-3326 to crash-land. The incident perhaps would have been
forgotten had the Curtiss Company not found it necessary to issue a
statement. The company was anxious to make clear that the engine
had been an experimental unit, originally rated at 325 bhp and installed
in the triplane for the race. Against the wishes of the company, the
engine had been fitted with high-compression pistons and boosted to
give 450 bhp at 2,200 rpm, the highest rating on record for a CD-12.
  
The crankshaft of that engine had not been drop-forged. It was made
from a slab forging and did not have proper grain flow. There had
been no time available to make proper forging dies and the disastrous
result had been anticipated. After this race the last remaining triplane
was scrapped.   A page of colorful aviation history had been turned.
  A correspondent of the leading French aviation paper L'Aeronautique,
after having had a look at the Curtiss and Wright engines toward the
end of 1923, found it strange that in France only the Hispano-Suiza firm
was engaged in the manufacture of monoblock engines, "a form of con-
struction that appears to be so much in vogue in America." It had not
dawned upon him that what really had happened was that during the
previous six years the Curtiss and the Wright engineers had been trying
to "out-Hispano" each other, and the original Birkigt concept had been
left far behind in the process.
  It has been recorded that in 1920 and 1921 France had made little
progress toward very high speeds, in spite of the introduction of a com-
pletely new and special airplane, the Nieuport Sesquiplan. This lack
of progress should be blamed on the lack of research in aero engines
and the inability of the French Hispano to develop more than about
330 bhp. As previously stated, it had a greater frontal area than the
D-12. The Hispano-Suiza could be run at higher revolutions per min-
ute, but here the propeller tip speed problem loomed dangerously near.
Sadi-Lecointe had nearly lost his life in 1921 when a propeller burst at
high speed.
  The Americans, meanwhile, had jumped into the competition for
engine and propeller progress with considerable dash, and their efforts
were beginning to bear fruit. The critical period from 1919 to 1923 may
be termed decisive for France's loss of top place in aviation design. The
Curtiss Company enjoyed an immense advantage in that Kirkham had
anticipated, in 1917, the horsepower requirements of five years later.
Wright had done almost as well by refining the original 300-hp Hispano
into a much superior engine.
  The year 1923 had been one of glory for the D-12, and just before the
year ended the Army's engineering division at McCook Field started to
test that engine in earnest.
  A D-12 was installed in a DH-4 (Project XP-277) for service tests
that were conducted with success. And D-12 engine 125, AS-23-154,
was set up for a 50-hour endurance test, running with high compression
and at high speed—that is, an endurance test under racing conditions.
The results of the test are described in McCook serial report 2348. With
compression of 5.8:1, the average power developed during the 50 hours
64

was 440 bhp at 2,170 rpm. Maximum rating during the first full-throttle
half hour of each of the five ten-hour runs, as well as during the last
hour of the test, averaged 468 bhp at 2,240 rpm, with an average fuel
(50 percent benzol) consumption as low as 0.48 lb/hp/hr. Highest
power developed was 475 bhp at 2,320 rpm, which may be taken as
indicative of the power developed during the Schneider and Pulitzer
races of 1923. This test, which lasted from 3 December 1923 until 12
February 1924, revealed the remaining weaknesses of the design. The
splines of the vertical drive shaft were unsatisfactory, and three of these
had to be replaced during the test. Also, the cylinder block was found
to be too fragile below the top flange. After 35 hours the right block had
to be replaced by one from another D-12, and at the end of the test the
left block was beginning to leak.
  The high-compression D-12 ran with Stromberg NA-Y5A carburetors,
already standard, and Splitdorf magnetos. The originally fitted B.G.
plugs had to be replaced during the endurance test by A.C. plugs because
the B.G. units became too hot, causing pre-ignition.
  After the final inspection the engine's major parts that had not given
trouble during the power runs were found in good condition. The
lessons of the high-compression endurance test were immediately put
into practice. The vertical drive shafts were redesigned, and henceforth
the cylinder blocks were cast from Duralumin and gave no more trouble.
  The result of these improvements soon was making headlines. On
23 June 1924 Lieutenant Maughan began a dawn-to-dusk continental
crossing in a PW-8—the fifth one produced—that had been equipped
with one of the improved D-12 engines and the already obligatory
Reed propeller (see Figure 35).
  Maughan started from Mitchel Field, Long Island, New York, at 3
a.m. Eastern Standard Time. The D-12 ran beautifully, and the first
lap of 590 miles was covered in four hours when Maughan landed
at McCook Field, Dayton, Ohio. The second landing was made at the
Municipal Flying Field, St. Joseph, Missouri, at 10:50 a.m. Central
Standard Time, 30 minutes sooner than the year before. The third
stop (not official) was at North Platte, Nebraska, where Maughan arrived
at 1:30 p.m. Twenty minutes later he was off again for Cheyenne,
Wyoming, the next official stop, where he arrived at 2:17 p.m. Mountain
Time. Then came the worst part, the crossing of the Rocky Mountains;
but the PW-8 surmounted every obstacle in its path without trouble
and arrived at the next landing place, Saldura, Utah, at 5:20 p.m. Pacific
Time. The last lap also was covered without incident, and Crissy Field
at San Francisco was reached at 9:48 p.m. Pacific Time, still in daylight.
65


FIGURE 35.—Curtiss Pursuit PW-8, 1924. This plane, powered by a Curtiss D-12
engine and piloted by Lieutenant R. L. Maughan, flew the "Dawn to Dusk Flight"
of 2,700 miles from New York to San Francisco in 21 hours and 48 minutes on
23 June 1924. (Smithsonian photo A-4655G.)
  The entire flight had taken 21 hours 48 minutes, but actual flying time
for the 2,645 miles covered was 18 hours 20 minutes. The average flying
speed was over 156 mph. Compared to the 26 hours, nonstop, of the
Liberty-powered T-2 of May 1923, the advance was indeed impressive.
  Since Maughan had felt ill during some stretches of the flight in the
previous year, presumably because of exhaust fumes, the 1924 PW-8
had been fitted with long exhaust pipes leading behind the cockpit. This
is the only example known of a D-12-powered airplane being so modified.
  A couple of months after the epoch-making dawn-to-dusk record, one
of the new D-12 engines arrived at McCook Field for another endurance
test. It was engine 93, AS 23-144, and it embodied all the improvements
resulting from the experiences of all the previous Navy and Army tests:
Duralumin cylinder blocks, on which the length of the end water-jacket
upper bosses had been doubled; redesigned vertical drive shafts; new
type short-skirt pistons; and steel-backed connecting rod and main bear-
ings. The engine, running with low-compression pistons, was given what
may be termed an all-out test at high crankshaft speeds—higher than
that of any test of a Curtiss engine since the K-12.
  This latest D-12 stood up magnificently to all that was demanded
of it, and, according to McCook report 2428, it "operated remarkably
66


FIGURE 36.—Views of Curtiss D-12 engine, 375 hp, 1922.
(Smithsonian photos A-4655C, A-4655D.)
well without a single forced stop!" All the tests—endurance, calibration,
fuel consumption, and others—were effected between 30 September and
13 October 1924, a record low in test duration. The average maximum
power developed was 444.5 bhp at 2,480 rpm, and, during a calibration
run before the 50-hour test, 460 bhp at 2,490 rpm was reached, running
on 40 percent benzol mixture. After the endurance test, the engine was
stripped for a thorough inspection, as was customary. Apart from some
warped exhaust valves