_A Study for a Scale Model_
_The original plans of the pioneer transatlantic steamer_ Savannah
_no longer exist, and many popular representations of the famous
vessel have been based on a 70-year-old model in the United States
National Museum. This model, however, differs in several important
respects from contemporary illustrations._
_To correct these apparent inaccuracies in a new, authentic model,
a reconstruction of the original plans was undertaken, using as
sources the ship's logbook and customhouse description, a French
report on American steam vessels published in 1823, and Russian
newspaper accounts contemporary with the_ Savannah's _visit to St.
Petersburg on her historic voyage of 1819. The development of this
research and the resulting information in terms of her
measurements and general description are related here._
THE AUTHOR: _Howard I. Chapelle is curator of transportation in
the United States National Museum, Smithsonian Institution._
The United States National Museum has in its watercraft collection a
rigged scale model purported to be of the pioneer transatlantic
steamer _Savannah_. For many years this model was generally accepted
as being a reasonably accurate representation and was the basis for
countless illustrations. Curiously enough, the model (USNM 160364)
does not agree with the published catalog description[1] as to the
side paddle wheels. Neither does it agree with the material in the
Marestier report,[2] which is accepted as the only source for a
contemporary picture of the _Savannah_.
The recent naming of an atomic-powered ship in honor of the famous
steamer greatly increased popular interest in the pioneer ship and its
supposed model. Consequently, the National Museum undertook the
research necessary to correct or replace the existing model. This
research has been carried out by the staff of the Museum's
transportation division with the aid of Frank O. Braynard of the
American Merchant Marine Institute, Eugene S. Ferguson, curator of
mechanical and civil engineering at the Museum, and others.
The _Savannah_ crossed from Savannah, Georgia, to Liverpool, England,
in the period May 22 to June 20, 1819; and proceeded to the Baltic,
where she entered at St. Petersburg (now Leningrad), Stockholm, and a
few other ports. On her return she reached Savannah on November 30,
and on December 3 she sailed for Washington, D.C., arriving on
December 16. Her original logbook now on exhibition in the Museum,[3]
covers the period between March 28, 1819, when she first left New York
for Savannah, to December 1819 when she was at Washington.
[Illustration: Figure 1.--Old model of the _Savannah_, built under the
supervision of Captain Collins. This model has been removed from
exhibition in the United States National Museum because of
inaccuracies. (_USNM_ 160364; _Smithsonian photo_ 14355.)]
The old model (fig. 1) was built about 1890-1892 by Lawrence Jenson, a
master shipwright and model builder of Gloucester and Rockport,
Massachusetts, under the supervision of Capt. Joseph Collins of the
U.S. Fish Commission. Notes in the records of the Museum's
transportation division show that the research for this model was done
by Captain Collins through use of an unidentified lithograph, printed
after the transatlantic voyage, and what then could be learned about
American sailing ships contemporary with the _Savannah_. In these
notes the complaint is made that no contemporary representation of the
steamship had then been found.
The old, inaccurate model, built to the scale of one-half inch to the
foot, represents an auxiliary, side-wheel, ship-rigged steamer. The
model scale measurements are about 120 feet in over-all length, 29
feet in beam, and 13 feet 6 inches depth in hold. The tonnage is
stated on the exhibit card to have been about 350 tons, old
measurement. The model has crude wooden side paddles of the radial
type, a tall straight smokestack between fore and main masts, a small
deckhouse forward of the stack, a raised quarter-deck, and a round
stern.
[Illustration: Figure 2.--The United States National Museum's new
model of the _Savannah_. This model was built by Arthur Henning, Inc.,
of New York City, from the ship's plans as reconstructed by staff
members of the Museum's division of transportation. (_USNM_ 319026.)]
The first step in the research for creating a more faithful
representation of the _Savannah_ was to obtain the customhouse
description of the ship. It was readily established that she was built
as a sailing packet ship by the Fickett and Crockett shipyard[4] at
Corlaer's Hook, East River, New York, and that she was launched August
22, 1818. Her register shows that she was 98 feet 6 inches in length
between perpendiculars, 25 feet 10 inches in beam, 14 feet 2 inches
depth in hold, of 319-70/94 tons burthen, and with square stern, round
tuck, no quarter galleries, and a man's bust figurehead.
These dimensions of the _Savannah_ required the researchers to
investigate the method of taking register dimensions in 1818. It was
found that the customhouse rule then in effect measured length
between perpendiculars above the upper deck, from "foreside of the
main stem" to the "after side of the sternpost." The beam was measured
outside of plank at the widest point in the hull, above the main
wales. If a vessel were single-decked, the depth was measured
alongside the keelson at main hatch from ceiling to underside of deck
plank; if double-decked, one-half the measured beam was the register
depth.[5] However, inspection of the register of a number of ships of
1815-1840 showed that, in practice, double-decked ships commonly were
measured as single-decked ships; this obviously was the case in the
_Savannah_. Also, due to the lack of precise measuring devices, the
register dimensions were not always accurate, particularly those of
the length, which often were in error as much as one foot in a
hundred, as was found by investigation of various classes of vessels.
Because of inherent difficulties in measuring to the required points,
this condition lasted even after steel tapes were introduced late in
the 19th century.
The Museum's researchers next turned their attention to examination of
the Marestier work, a French report on early American steam vessels
that had become known to some American marine historians in the
1920's. The author was a French naval constructor who, on orders from
his government, had spent two years in the United States between 1819
and 1822 studying American steam vessels, schooners, and naval
vessels. The published report contained only material on steam vessels
and schooners. The portion dealing with naval vessels was not
published, and the manuscript has not been found to the present time
(1960). The publication, a rare book, was available in only a few
collectors' libraries or public institutions in the United States. In
1930 the writer translated the chapter on schooners,[6] and in 1957
Sidney Withington translated most of the remainder.[7] As a result of
these publications and earlier published references, the Marestier
material became widely known to persons interested in ships.
[Illustration: Figure 3.--Marestier's sketch of the _Savannah_ (from
plate 8 in Withington's translation of the Marestier report). Heights
of lower masts are excessive by all known American masting rules; and,
according to Marestier's drawing of the engine (see figure 4), the
deckhouse is too short.]
Withington's translation states that the _Savannah_ measured 30.48
meters (100 feet) in length and 7.92 meters (26 feet) in beam and that
she drew 3.66 meters (12 feet) in port and 4.27 meters (14 feet)
loaded. Marestier's sketch (see fig. 3) of the outboard of the
_Savannah_ shows a ship-rigged, flush-decked vessel with a small
deckhouse forward of the mainmast and nearly abreast of the side
paddle wheels. The stack is a little forward of the deckhouse and has
an elbow at its top. Netting quarter-deck rail is shown and a bust
figurehead is indicated. The position of the hawse pipe shown at the
bow indicates the wheel shaft to have been at or about deck level. For
structural reasons, and in compliance with the sketch, the wheel shaft
would have been just above the deck.
[Illustration: Figure 4.--Marestier's drawings of the _Savannah's_
engine (from plate 7 in Withington's translation of the Marestier
report). The graphic dimensions do not precisely correspond to the
scale of dimensions in Marestier's text, nor with other recorded
measurements.]
Marestier's drawings of the engine and paddle wheels[8] are reproduced
in figure 4. The nonoscillating engine is inclined toward the
paddle-wheel shaft. The connecting rod operates a crosshead to which
is pivoted a pitman, or oscillating rod, that operates the
paddle-wheel crankshaft. Alongside the steam cylinder is an air pump
cylinder, also connected to the crosshead. The steam inlet and outlet
pipes enter a valve chest on top of the steam cylinder, which is
described as being 1.035 meters (3.4 feet) in diameter, and of 1.5
meters (4.9 feet) in stroke.
The paddle wheels are shown as being of iron, with two fixed arms
opposite one another on the hub. The other arms (four above and four
below the fixed arms) are pivoted to the hub and held spread by chain
stays. These eight blades fold, in pairs, to each of the fixed arms.
The wheels are shown in elevation, with the upper pivoted arms folded
on top of the fixed arms, and in cross section; the latter shows the
shape of the buckets, hub, and outboard bearing of the shaft. The
wheels are described as being 4.9 meters (16 feet) in diameter, while
the buckets are 1.42 meters (4.65 feet) wide and 0.83 meters (2.72
feet) deep. The two outer corners of each bucket are snyed off at
nearly 45 deg. The wheels are shown folded in the sketch; according to
the description, they could be unshipped from the shaft and stowed on
deck when desired. The method of removing the wheels from the shaft is
not described, but from the drawings it seems probable that they were
detached from the shaft by removing a lock bolt outboard and sliding
the wheels off the square shaft. The hub seems adequate for this.
Marestier states that this removal could be accomplished in 15 to 20
minutes; the logbook shows that it took 20 to 30 minutes to perform
this operation at sea.
Marestier states that the ship had spencer masts and trysails on fore
and main, and a spencer mast on the mizzen for a spanker; he
illustrates these as having royal poles, but with no royal yards
crossed.[9] The smokestack is described as pivoted. The mainstay is
double, setting up at deck, near rail, and forward of the foremost
shrouds of the foremast to clear the stack and foremast.
The boilers were in the hold, but Marestier gives no dimensions.
However, he comments that, in American steamers, the space for steam
in the boilers varied from 6 to 12 times the capacity of the cylinder.
He gives the _Savannah's_ boiler pressure as 2 to 5 pounds per square
inch and the maximum revolution of the wheels as 16 revolutions per
minute. The boilers could burn coal or wood. Judging by Marestier's
sketch of the ship, the stack was at the firebox end; the boiler or
boilers were underneath the engine.
The log of the _Savannah_ gives little useful technical information
other than that the ship readily made 9 to 10 knots under sail in
fresh winds, showing she could sail well. Under steam alone the log
credits the ship with a speed of 6 knots; Marestier estimated her
speed at 5-1/4 knots in smooth water. The log shows that she usually
furled her sails when steaming, though on a few occasions she used
both steam and sail. In her crossing from Savannah to Liverpool she
appears to have been under steam for a little less than 90 hours in a
period of about 18 days (out of the total of 29 days and 11 hours
required to cross). There is no evidence of any intent to make the
whole passage under steam alone, for the vessel was intended to be an
auxiliary, with sails the chief propulsion.
Captain Collins states in his notes that the ship was built by Francis
Fickett as a Havre packet, that she stowed 75 tons of coal and 25
cords of wood, and cost $50,000. Apparently quoting Preble[10] to a
great extent, he also states that the engine developed 90 horsepower
and had a 40-inch diameter cylinder with a stroke of 5 feet.
Preble states that the ship was purchased for conversion to a steamer
after launching and gives statements by Stevens Rogers, sailing master
of the _Savannah_, to the effect that the ship was built as a Havre
packet and that the project ruined financially one of the investors,
William Scarborough. Rogers, who made these statements in 1856, also
said the ship was built by "Crocker and Fickett." Contemporary
newspapers, quoted by Preble, state that the ship had 32 berths in
staterooms for passengers.
Morrison[11] credits the building of the _Savannah_ to Francis Fickett
and says she was intended for the Havre packet run. He states that the
vessel cost $50,000; that her paddle wheels, each with eight buckets,
were 16 feet in diameter; and that she had canvas wheel boxes
supported by an iron frame. Morrison also relates the history of the
ship after her return from Russia--the removal and the sale of her
machinery to James P. Allaire, the operation of the ship as a sailing
packet between New York and Savannah under the ownership and command
of Captain Holdridge, and her stranding and loss during an
east-northeast gale on November 5, 1821, at Great South Beach, off
Bellport, on the south shore of Long Island. He also states that the
steam cylinder of her engine was exhibited at the Crystal Palace Fair
in New York during 1853, and that the ship proved uneconomical due to
the large amount of space occupied by the engine, boilers, and fuel,
leaving little space for cargo. Morrison apparently used some of the
statements made in 1836 and 1856 by Stevens Rogers, who was the
sailing master on the famous voyage.
Tyler[12] names the stockholders of the Savannah Steamship Company,
owner of the _Savannah_. The company was proposed by Capt. Moses
Rogers, and its shareholders were William Scarborough, John McKenna,
Samuel Howard, Charles Howard, Robert Isaacs, S. C. Dunning, A. B.
Fannin, John Haslett, A. S. Bullock, James Bullock, John Bogue, Andrew
Low, Col. J. P. Henry, J. Minis, John Sparkman, Robert Mitchell, R.
Habersham, J. Habersham, Gideon Pott, W. S. Gillet, and Samuel Yates.
Tyler establishes, by the company's charter, that the objective was to
institute a New York-Savannah packet service, for which the _Savannah_
was to be the first ship. He shows that, due to the economic
depression of 1819, the _Savannah_ sailed to Liverpool in ballast and
without passengers. Her fuel capacity is given as 1,500 bushels (75
tons) of coal and 25 cords of wood. [It should be noted that 1,500
bushels of bituminous coal does not quite equal 75 tons.] Tyler quotes
S. C. Gilfillan[13] as to criticisms of the engine and its design.
Partington[14] estimated coal consumption to be nearly 10 tons a day;
remarked on the uneconomical arrangement of the ship, with the engine
and boiler occupying the greater part of the space amidships, between
fore and main masts; and located the axle of the paddle wheel "above
the bends," that is, in the topsides above the wale. The description
he gives of the unshipping of the wheels is that the pivoted blades
were removed and the fixed blades, in horizontal position, were left
on the shaft. This agrees with a Russian description referred to
later. The logbook repeatedly speaks of "shipping" and "unshipping"
the paddle wheels, indicating that the wheels were entirely removed
from the shafts and stowed on deck.
Watkins[15] showed, by the account books of Stephen Vail, owner of the
Speedwell Ironworks near Morristown, New Jersey, that the engine was
built by Vail, but apparently to designs by Daniel Dod. The latter
built the _Savannah's_ boiler at Elizabeth, New Jersey, and made some
parts of the engine, which he furnished, incomplete in some instances,
to Vail. These account books, which were in the possession of John
Lidgerwood of New York City in 1890, show the steam cylinder to have
had an inside diameter of 40-3/8 inches and a 5-foot stroke. Reference
in the account books to an error in Dod's draught of a piston proves
that Dod designed the engine.
Watkins states that the engine was rated at 90 horsepower. He does not
give the diameter of the pump cylinder, but, judging by the scaling of
Marestier's drawing and by a rather indefinite entry in the Vail
account book, it appears to have been between 17 and 18 inches.
Quoting Captain Collins at some length, Watkins writes that the
mainmast was placed farther aft than was usual in a sailing ship, and
that the vessel had a round stern. Collins apparently based his
opinion upon an unidentified "contemporaneous lithograph" and upon
"all other illustrations of this famous vessel." Collins' conception
of the appearance of the _Savannah_ is shown in a drawing by C. B.
Hudson that is reproduced as the frontispiece in Watkins' publication.
A statement by Stevens Rogers that was published in the _New London
Gazette_ in 1836 appears to have been the original source for
statements regarding the _Savannah's_ fuel capacity, her sale, and her
loss in 1821 while owned and commanded by Capt. Nathaniel Holdridge,
"now master of the Liverpool packet ship _United States_." Watkins
also gives a picture of Stevens Rogers' tombstone, on which there is a
small carving purported to be of the _Savannah_. The tombstone was
made in 1868.
From a Russian newspaper contemporary with the _Savannah's_ visit to
St. Petersburg, Frank Braynard found a statement that the vessel had
two boilers, each 27 feet long and 6 feet in diameter.[16] It was also
shown she had at least one chain cable. Considerable information on
the cabin arrangement and the method of folding the wheels was also
obtained from this Russian source.
In spite of a very extensive bibliography on the _Savannah_, the basic
sources for reliable technical description are Marestier's report on
American steamers, the logbook of the ship, Watkins' extracts from the
Speedwell Iron Works account book, the customhouse records, and some
of the statements made by Stevens Rogers between 1836 and 1856. Plans
of the ship, or a builder's half-model, have not been found.
Marestier's sketch of the _Savannah_, which is not a scale drawing,
and his drawings of the engine and paddle wheels were the only
available illustrations upon which reconstruction could be based.
Through the efforts of Malcolm Bell, Jr., of Savannah, Georgia, and
Frank Braynard, a search was made by Russian authorities at Leningrad
for contemporary references to the ship. This work resulted in
information as to how the side wheels were folded, the dimensions of
the boilers, and some description of the cabins and fittings.
As to the ship itself, the customhouse registered dimensions are of
prime importance; they fix the over-all hull dimensions within
reasonable limits. A vessel of 1818 measuring 98 feet 6 inches between
perpendiculars would have been 100 to 104 feet long at rail. The type
of ship represented by the _Savannah_ is well established. All
references are in agreement that she was built as a packet ship--a
Havre or transatlantic packet in most accounts.
The packet ships listed by Albion[17] show that all the pioneer ships
of the transatlantic Black Ball Line--which began operation with the
sailing of the 424-ton _James Monroe_ on January 5, 1818--measured at
least 103 feet 6 inches between perpendiculars. Two of the pioneer
ships of the first Havre Line--which did not begin operation until
1822--were under 98 feet between perpendiculars. The second Havre Line
began operation in 1823; of its four pioneer packets, two were
purchased general traders measuring under 98 feet between
perpendiculars. The coastal packets built between 1817 and 1823 were
all under 100 feet between perpendiculars. It is apparent, then, that
the size of the early packets did not indicate, with any degree of
certainty, the trade in which they might be employed.
Belief that the _Savannah_ was built as a Havre packet is based upon
Stevens Rogers' statements, and her size obviously does not make this
impossible; nevertheless, it seems highly improbable that she was
built for the Havre service because no Havre line of packets had been
organized as early as 1818 out of New York or Savannah so far as can
be found. However, the matter is not of very great concern as it is
probably true that the models of coastal and transatlantic packet
ships were quite similar at the period of the _Savannah_. This
statement is supported by the plan of a coastal packet built seven
years after the _Savannah_.
The hull-type of these early packets can be established. While no
half-models or plans of packets built before 1832 could be found,
offset tables of a Philadelphia-New Orleans packet of 1824-1825 were
obtained through the courtesy of William Salisbury, an English marine
historian who had been studying the British mail packets. These offset
tables had been sent from Washington on March 25, 1831, by John
Lenthall, U.S. naval constructor, to William Morgan and Augustin
Creuze, London editors, for publication.[18] The offset tables were
for a packet ship 103 feet between the perpendiculars of the builder
(rather than between those of the customhouse) and 27 feet moulded
beam. An examination of the files on American packet vessels in the
collection of Carl C. Cutler, curator emeritus of the Mystic Marine
Museum, showed with certainty that the offsets were for the _Ohio_,
built at Philadelphia late in 1825. The drawings of this ship (fig. 5)
were made from the offset tables and from other measurements; minor
details are from portraits of packet ships, particularly of the first
_New York_ (1822-1834) of the Black Ball Line.
[Illustration: Figure 5.--Lines of the coastal packet ship _Ohio_,
built at Philadelphia in 1825 for the Philadelphia-New Orleans run.
The _Ohio_ represents the general type of early American packet
ships.]
The _Ohio_ was two-decked, with the upper deck flush. She had rather
straight sheer, 27-inch bulwarks, a moderately full but easy entrance,
a fine, long run, and little drag to the keel. The midsection was
formed with moderately short and rising floor, round and easy bilge,
and some tumble-home in the topside. The stem raked a good deal for a
ship-rigged vessel; the post raked slightly. There was a distance of 6
feet between upper and lower deck planks. The stern was of the square
transom, round tuck form, as mentioned in the _Savannah's_ register.
Lenthall reported the _Ohio_ to have been a good sailer and to have
had other desirable qualities. She was registered as being of 351.86
tons burthen, 105.5 feet between perpendiculars, and 27.4 feet in
extreme beam. She was, therefore, about 7 feet longer and about 2
feet 3 inches wider than the _Savannah_. The plan shows she was about
2 feet 4 inches deeper in hold than the _Savannah_, and, according to
Cutler, she had "an unexpected degree of sophistication for a coastal
packet of that period."[19] By modern standards, the _Ohio_ shows a
well-advanced design for the period.