H. Kimball, Meteorologist
Mary Celeste, the Argosy of an Abandoned Ship, by Charles Edey
Foy, invites a critical examination of the weather leading up
to the mystery of that modern Flying Dutchman found forsaken
on the High Seas 70 years ago. There can be no further question
as to the trustworthiness of her owners, her underwriters, her
master, Captain Briggs or her crew, neither can any doubt remain
as to seamanship or soundness of vessel.
A waterspout at sea In: The Atmosphere
James Glaisher, 1873 From the work of Camille Flammarion
Figure 65, p. 351
Undoubtedly the ship left New York properly loaded. Captain
Briggs came to New York on October 20th, 1872 for the purpose
of supervising the loading, selecting his officers and crew
and seeing that all was well for the start. On November 3rd,
four days before he sailed, he wrote his mother: “Our
vessel is in beautiful trim and I hope we shall have a fine
passage”. Furthermore, that a man of Captain Briggs’
proved seamanship would willfully abandon a staunch vessel is
a suggestion so completely at variance with tradition as to
make it incredible as well as repellent to men of the sea.
Some catastrophic situation must be envisaged before one can
believe that the Captain would judge an overloaded lifeboat
in a rough sea as providing a greater measure of safety for
his family and crew than did his own vessel, which he had ample
reason to think to be sound. Yet many thousands, both landlubbers
and mariners, have believed and do now believe that the entire
ship’s company of ten put off from the Mary Celeste in
her small boat and were lost: all this in despite of the fact
that there is nothing in the testimony brought out in the Vice-Admiralty
Court salvage proceedings at Gibraltar to establish the belief
that a lifeboat was lowered.
Notwithstanding, those who have given the most painstaking consideration
to all available facts are inclined to consider tentative abandonment
as the least objectionable solution– the intention being
to bring the boat back to the ship by the aid of a towline if
and when the threat had disappeared. This presupposes a state
of panic; otherwise, how could the assumption, that the boat
was safer than the ship, be maintained for a period of time
sufficient to prepare and launch the boat? The state of panic
presupposes a terrifying surprise occasioned, as is presumed
by many, by an explosion in the hold. But if there was an explosion
it occurred after two hatch covers had been removed and properly
placed on deck. This methodical handling of hatch covers does
not suggest the work of men in a state of panic, neither does
the collection of the navigating instruments and the ship’s
papers, presumably to be carried off the ship; for all of them
were missing when the vessel was discovered. The vessel was
loaded with alcohol.
To support the assumption that the vessel was abandoned, it
is necessary to presuppose a lapse of time sufficient to unlash
and right the boat on deck, stow the instruments, ship’s
papers, a cask of water and some provisions, get the boat over
the side and onto a rough sea, make fast the towline and get
ten people, including the Captain’s wife and child, into
the boat. Certainly all this could not be done in less than
five minutes by a panicky crew.
It would seem likely that in the five long minutes of preparation,
at least someone would have regained his sense of reality and
have refused to leave the vessel; and, in that event, there
is a strong probability that a survivor would have been found
by the boarding party. Had the vessel been abandoned in a panic
the boat’s lashings would have been cut, not loosened
by fumbling fingers; but neither their frayed ends nor any attached
part of the towline were observed by Mate Deveau, commanding
the boarding party from the Dei Gratia, which sighted the desolate
Mary Celeste December 4, 1872.
The abandonment theory could better be supported if complete
separation were envisioned. This would eliminate the suggestion
of the use of a towline, and, at the same time justify the spread
of canvas as left for the purpose of securing the greatest possible
distance from the vessel. The explosion surmise calls for more
careful consideration of the probable loading than appears in
Nine barrels were found empty at Genoa on March 26th, 1873:
they had been in the hold since November 3rd, when Captain Briggs
reported that the vessel was in “beautiful trim”.
The barrels were not reported at Genoa as broken, merely empty,
so it so it may be assumed that their contents had seeped away
gradually. Each barrel probably contained 50 gallons. Such barrels
would be approximately 30 inches high, 20 inches in diameter
at the head and 24 inches at the bung. This figures to a total
of about 17,000 cubic feet, to which must be added 25% for broken
stow, bringing the total to 21,250 cubic feet. This is only
a few hundred cubic feet less than the capacity of the hold
of the vessel, which was approximately 100 feet long, 30 feet
wide and 12 feet deep.
Consequently, the vessel must have been loaded to within a few
inches of the deck. Such a loading would leave, in conformance
with standard practice, a freeboard amidships of about 3 feet.
Mate Deveau, of the Dei Gratia, testified that on boarding the
Mary Celeste he found “a great deal of water below decks
- the forward house, full of water up to the coaming–
everything wet in the cabin in which there had been a great
deal of water”. Second Mate John Wright testified that,
on boarding the vessel with Mate Deveau, “there was water
in the cabin, between decks and in the forehouse”. Augustus
Anderson, seaman, also in the boarding party, testified, “the
cabin was wet– everything was wet in it, the clothes and
all. There was three foot and a half of water in the hold, there
was a good deal of water between decks and also in the forwardhouse”.
This testimony relates to conditions found on December 4th.
On December 23rd the surveyor at Gibraltar reported having found
under the forward hatch “a new hawser which had never
been used and was perfectly dry– had any quantity of water
found its way through this hatch the hawser would have exhibited
signs of having been wetted. It exhibited none nor did any other
of the articles which I observed here”. Presumably, then,
some of the water found by the boarding party on December 4th
had come into the hold after November 24th through leaky floors.
Ordinarily, in the Mary Celeste’s time, floors in cargo
vessels were not water-tight. Loosely planked floors would permit
the escape of alcohol vapor into the living quarters if there
were a dangerous accumulation of so odorous a fume in the nine
days of voyage when the sleeping quarters were almost constantly
occupied. Had there been fumes in sufficient quantity to cause
apprehension the fact would have been noted in the ship’s
No doubt the barrels were stowed to within a few inches of the
deck– stowed in tiers on their sides, bungs up and end-to-end.
Such a stow suggests gradual leakage in several places and localized
accumulation of alcohol with subsequent vaporization and permeation
of the approximately 400 cubic feet of otherwise unoccupied
The discovery of the nine empty barrels was made when the vessel
was unloaded at Genoa, about twenty weeks after loading in New
York. On the assumption that each barrel originally held 50
gallons the loss would have been 450 gallons, at the rate of
3.2 gallons per day. If the leakage was gradual and if it began
at the time of sailing, the total leakage up to November 25th
would have been approximately 73.4 gallons. The fact that the
free-space consisted of a thin layer just under the deck supplemented
by many small interstices does not in itself weaken the explosion
theory; but the fortuitous occurrence of a detonating spark
caused by the rubbing together of metal barrel-hoops in a space
containing a sufficient explosive-mixture concentration, during
the relatively quiet period after four days of storminess throws
grave doubt on it.
In the 70 years since the disaster fancy has run rife in search
of an explanation. There have been stories built up from conceptions
of fraud, mutiny, drunkenness, panic, yellow fever and, to cite
perhaps the most imaginative, the picking off of the ship’s
company, one by one, by an octopus. But nowhere does there appear
more than a superficial consideration of weather as a possible
cause beyond speculation that the vessel did not go through
a severe storm. And yet, all that happened could have been done
by a waterspout.
The plan of the voyage would be laid down before the vessel
left Sandy Hook, and sine the ship was to be heavily laden the
course selected would be that recommended for sailing vessels,
i.e., east until reaching the 50th meridian; then about east-southeast
until nearing the island of Santa Maria; then east-northeast
to pass north of the island and reach the current normally flowing
Normally, such a course would avoid much of the storminess well
known to prevail farther north in November.
Plottings now available show the likelihood of gales in the
5° zone next north of the 40° parallel between longitude
40°W and the Azores to be nearly twice as great as in the
corresponding 5° zone to the south of the 40° parallel.
Also it would be good seamanship to pass to the northward of
the island of Santa Maria since from that location the set of
the current leads directly into Gibraltar. Moreover, in November
the set leads through a region less liable to calms than does
the more southerly approach. The choice of course, to the north
or to the south of Santa Maria would necessarily remain open
until the ship neared the island, since the Dollabaret Shoal,
21 miles to the northeast of the island, presents a difficult
Presumably, it was Captain Briggs who chose the northerly course.
A less experienced navigator would not care to risk the more
dangerous passage which the circumstances requires must be made
on a stormy night. This suggestion is in refutation of the charge
that a mutiny had deprived the Captain of his command.
On November 7th, when the voyage was begun, and for the two
following days, an area of low pressure was moving slowly eastward
across New England and the maritime provinces. Then, on the
10th, there followed an area of high barometer holding to westerly
the winds that had come from that direction since the beginning
of the voyage. On the 12th another low approached from the lower
Lake Region and the winds off-shore responded by shifting to
southerly, but, as the low passed rapidly across New England,
they soon resumed their original course. Then followed another
high sustaining the westerly winds. This succession of lows
and highs, none of which was of considerable development, gives
reasonable assurance of fair winds and moderate weather for
the first part of the passage.
The assumption of the officers of the Dei Gratia that storms
prevailed for most of the time between the departure of the
Mary Celeste on November 7th and her abandonment on the 25th,
because such was their experience in starting from New York
eight days later, disregards the normal behavior of Atlantic
storms. Storms of the Atlantic ocean that would affect the Dei
Gratia usually bear northeastward, crossing the middle part
of the European Shipping Route. They are of only moderate frequency
in November, and rarely are they diverted to easterly, as would
be necessary to reach the location of the Mary Celeste some
1200 miles ahead.
The theory of similar weather experienced by both vessels is
offered as the basis for the belief that the hold of the Mary
Celeste had not been ventilated from the time the vessel left
The weather experienced during the latter part of the voyage
is indicated in the experience of the Dei Gratia, then in the
same meteorological neighborhood; the heading of the derelict
when found, and official reports provided by the meteorological
Service of the Azores.
According to the testimony in the Vice Admiralty Court the Dei
Gratia was in winds from the northward on December 3rd and part
of the night of the 3–4th. These subsided in the forenoon
of the 4th and probably shifted to southerly in the middle of
the day. Under such circumstances, the greatest distance that
the Mary Celeste could have been to southward of a direct course
to Gibraltar would be less than 10 miles. During the four preceding
days she must have been drifting to southward under the then
prevailing northerly winds as reported by the Dei Gratia. That
surmise places the derelict well to the north of her course
on December 1st.
There seems to be no sound approach to an estimate of this departure
other than to make it fit as nearly as possible to a straight-line
drift (378 ½ miles) between the known locations, i.e.,
those of November 25th and December 4th; and hold the average
speed to be less than 4 knots, the near-maximum for an uncontrolled
vessel of this sort with shortened sail. If the vessel is considered
to have been propelled at an average speed of 2 knots for the
222 hours between noon of November 25th and noon of December
4th, the distance covered is found to be 444 miles, or a departure
of 65 ½ miles from the direct course. But from December
1st to the night of December 3–4th, the drift was to southward.
Consequently, between November 25th and December 1st the Mary
Celeste must have been pushed to the northward. Assuming the
drifts to be approximately equal, the vessel would have been,
on December 1st, near 39° 30' N, 22° west.
From this speculation, the Mary Celeste must have experienced
winds prevailingly south of west from November 25th as reported
by the Meteorological office at the Azores and another period
of light airs on December 4th as reported by the Dei Gratia.
The winds immediately preceding the lull on November 25th may
be inferred from the state of the sails that were found set,
i.e, the lower topsail hanging by the four corners, the jib
and fore-topmast staysail set on the starboard tack, the middle
staysail and topmast staysail lowered, and the mainsail, gaff-topsail,
topgallant-sail and royal furled. The foresail and upper foretopsail
had been blown away.
That both sails on the mainmast and the two uppermost sails
on the foremast were furled indicate that increasing storminess
was anticipated early enough to allow time for the completion
of this much of the job. Why the main-staysail and topmast-staysail
were not furled is open to speculation, but it is more likely
their condition indicates the end of an uncompleted job than
that the hoisting of more canvas had been started with these
Persistent north winds recorded in the Dei Gratia’s log
as prevailing for four days preceding December 4th suggest a
slow moving low proceeding eastward on a course well to the
northward of the Azores. This is not unusual. A file of daily
Atlantic charts for the first week in December shows such a
storm on December 3–6th, 1925 and another on December
1–5th, 1929. Besides, it is to be expected that a fresh
to strong northerly wind continuing for four days along the
Dei Gratia’s course would reach to the Mary Celeste never
more than 250 miles to the southward. A fresh 4-day northerly
wind of such spread is a logical development in the Azores region
when a slow-moving low approaches Southwest Europe.
The records of the meteorological stations in the Azores show
stormy conditions prevailing on November 24th and 25th, the
wind shifting from southwest to northwest between 3 P.M. and
9 P.M. November 25th. At the time of this shift (9 P.M.) The
station at Ponta Delgarda reported a wind 38 miles an hour.
During the forenoon of the 25th the winds were light but later
they became of gale-force. The record does not state how much
later, but presumably it was on the same day. This suggests
that storm moving to the eastward provided the gales after the
calm of the 25th, also the fresh to strong westerly winds experienced
by the Dei Gratia on December 1–4th.
Only a short quiet time-interval need be assumed between the
differing winds since it is a common experience for vessels
on a north Atlantic passage to report successive days of strong
“westerlies” shifting from southwest to northwest
Presumably, then, the mainsail, gaff-topsail, topgallant-sail
and royal were lowered and furled at about 8 P.M. of the 24th
when the weather became threatening and a large part of the
crew was available for the task.
Since the Dei Gratia made 150 miles between noons December 3rd
and 4th on a southeasterly course or 6.2 m.p.h. she must have
had made only 90 miles in the 24 hours next preceding the sighting
of the Mary Celeste. According to the testimony the wind in
this period had been strong but it became light by the time
the derelict was sighted. No mention was made of a shift so
it would seem that the change to east-southeast occurred after
the wind had subsided.
This indicates that the two vessels were on approximately the
same course on December 3rd and 4th , the Mary Celeste being
to the southward and leading but being overtaken during the
night, and that a short time before, say, in the forenoon of
December 4th, the wind shifted to east-southeast and the Mary
Celeste came to a west-northwest course when probably not more
than 10 miles from the place where she was discovered. The record
shows that the Dei Gratia made in the preceding 24 hours ending
December 1st, 200 miles; 2nd, 162 miles; 3rd, 150 miles, and
4th, 90 miles.
Likely, in the neighborhood of the Mary Celeste, the wind was
south of west most of the time from November 25th to December
1st, since from November 27th while the Dei Gratia was approaching
from the 20th parallel– that vessel had winds that were
mostly west-northwest to northwest.
According to Mate Deveau, when the Mary Celeste was discovered,
“The wind was to the northward, not much then, although
it had been blowing heavily in the morning with rain and squalls–
her head to the northward”, and from the testimony of
the second-mate: “The head of the other vessel was northwest-by-north
as far as I could judge”. The wind must have been from
northwesterly all night to account for the speed, 8 knots, made
by the Dei Gratia.
But, when sighted, the Mary Celeste was 4 to 6 miles on the
Dei Gratia’s port or windward bow when the Dei Gratia
was heading southeast by east. Under these circumstances it
would be impossible to judge very closely the course of the
Mary Celeste, but if she was making 1 to 2 knots, it seems likely
that she must have been headed about west-northwest, assuming
that the wind direction was the same in the location of both
vessels; and, that since shortly after 4 A.M., the Mary Celeste’s
wind was light east-southeast.
The gradual lulling of a northwest wind in the morning of November
25th and other circumstances to be commented on later, suggests
a waterspout as the cause of the disaster.
Waterspouts are more frequent in the general neighborhood of
the disaster than elsewhere in the eastern part of the Atlantic,
and many have been observed there in November; in fact, the
November frequency is only 2% less than that of October, the
month in which the largest number has been observed in that
area. The spouts usually occur in the lull between southerly
and northerly winds, for at that time the cool air aloft overflows
the warm surface drift producing a steep temperature lapse rate
and the convective instability essential to their formation.
In the development of a waterspout warm moist air rising to
an elevation of 1,000 feet or more develops a thick black cloud
from which a vortex-tube thrusts downward. In the meantime,
the ocean near the point of approach becomes agitated and later
the disturbed surface is circumscribed by a ring of spray. As
the whirling tube grows, it becomes a cylinder containing rising
air mixed with fresh and salt water, while the encircling spray
mounds up around the tip of the vortex giving it a bulbous appearance.
Well developed spouts remain intact usually for more than 15
minutes, during which time large quantities of water are accumulated
in the parent cloud, whence rain descending around the vortex
and mixing with the spray falls to produce the so-called cascade.
In many other respects waterspouts appear to be lawless. They
form in both settled and stormy weather; they may travel at
any angle, down wind, athwart wind, or against wind. Sometimes
they have only a slight forward movement, at other times their
speed exceeds 75 miles an hour. Sometimes they are alone, and
again in groups numbering two to ten or even twenty. They have
been reported in all hours of the day and night but more frequently
around dawn and mid-day. The average waterspout has a diameter
of 20 feet and travels 13 miles an hour.
Meteorological literature contains many accounts of vessels
that have been seriously damaged and others that have been destroyed
in encounters with these weather outlaws. Usually, in the daytime,
a steamer has but little difficulty in avoiding waterspouts
but for a sailing vessel, particularly at night and becalmed,
they are among the major perils of the sea.
One of the largest waterspouts of which there are dependable
measurements was observed from Cottage City, Massachusetts,
during the generally fair weather prevailing on August 19, 1896.
The following are some of the measurements:
Diameter of the spout at bottom............... 240 feet
Diameter of cascade............... 720 feet
Height of cascade............... 420 feet
Diameter of vortex at middle............... 144 feet
Diameter just at base of cloud............... 840 feet
Approximate height of tube............... 3600 feet
Another waterspout was observed from the Weather Bureau office
overlooking the Upper Bay, New York Harbor, September 5, 1924.
The tube of this one did not reach to the water surface, but
the invisible whirl of air in the intervening space was damaging
to a south-bound steam lighter that was unable to get clear
of it. The parent cloud covered almost the entire upper Bay
at an elevation of about 1,000 feet, while the funnel, at the
stage of its greatest development, reached to within 500 feet
of the water surface.
At the cloud base the diameter of the tube was between 300 and
400 feet, whence it tapered to a flattened extremity 100 feet
across. Directly below this funnel, and advancing northeastward
with it, there arose a conical whirl of mist and spray. From
a water area about 100 feet across, the cone extended to a flattened
top having a diameter of 50 to 60 feet.
The phenomenon was viewed from a distance too great to disclose
whether or not at any time harbor water was elevated in quantity
sufficient to produce a cascade. It was visible for four minutes,
during which time it traveled about 4,000 feet or to within
one half mile of the point of observation. There was no visible
rain in the immediate vicinity of the spout, but a sharp shower
occurred before it formed and another after it dissipated.
The mound of whirling air, spray and mist enveloped the lighter,
tearing away the tarpaulin that had covered the deck-load and
transported overboard seven bales each weighing 125 pounds.
The bales that were amidships and those on the port side were
not disturbed, nor was the lighter damaged materially.
Another spout, described to the writer as part of the early
sea-going experience of a thoroughly trustworthy witness was
encountered while he was mate of a barkentino loaded with lumber
and bound from Brunswick, Georgia for Las Palmas, Canary Islands.
The vessel had just crossed the Gulf Stream and the time was
between 11:00 p.m. and midnight. The southwest wind had become
a light air. But the sky was inky black and so threatening that
sail had been shortened in anticipation of a squall. The spout
came suddenly out of the northeast, passing over the forward
part of the vessel.
In a matter of seconds, the whirling wind struck, first on the
port side causing the vessel to lurch violently, then on the
starboard side, throwing her in the opposite direction. Quantities
of water were shipped. When the spout struck, the mate was on
one side of the ship and three sailors were on the other; but
when the spout had passed, all four found themselves crowded
in a space left in the cargo, up to their necks in water.
The lower sails, yards and rigging of the foremast were found
in a tangle wound around the fore-topmast, while all spars,
sails and rigging farther aloft had been carried away. Yet where
the Captain stood at the wheel the passing of the spout was
In establishing the time of the catastrophe of the Mary Celeste
as the middle of the forenoon the earlier investigators had
access to the ship’s log, which has sine disappeared,
and to the testimony of Mate Deveau as to entries found on the
The log write-up to noon of November 24th appeared to have been
complete while as to the memoranda on the slate, all that was
recalled was the following: “At 8, Eastern Point bore
SSW”. All writings on the slate were erased by Mate Deveau
to provide space for the record beginning when he took charged.
On the witness stand he did not recall whether or not there
were other notes on the slate. This uncertainty weakens his
statement as to his recollection of the note quoted.
The only other support for the widely accepted timing is the
testimony regarding the condition in which the galley and cabin
were found when the derelict was first boarded. Two items are
important, namely, (1) There was no evidence of preparation
of a meal, nor of one having been recently served, and (2) the
beds were not made.
In sailing ship’s routine, the cook is called at 5 a.m.
and shortly thereafter preparations for breakfast are begun.
Then there is but little respite until after the mid-day dinner.
Such a schedule leaves the middle afternoon and the night and
the very early morning as the only considerable periods in the
24 hours in which there are no signs of activity in the galley.
Furthermore, the surmise that the catastrophe occurred in the
period between breakfast and dinner runs afoul of the ritual
of the housewife of that period.
It is hardly believable that a woman of Mrs. Briggs’ upbringing
would leave her beds unmade until the middle of the forenoon.
Mate Deveau may not have recalled correctly the note he found
on the log slate, but whatever may have been the hour of the
disaster the following situation seems to fit the known facts.
The log book was found written up to and including noon of November
24th, on which day the winds were fresh to strong northerly.
During the following morning the wind subsided as shown by the
records of the Meteorological Service.
The main and lazarret hatches were found removed. This must
have been done on the morning of the 25th since no other suitable
period for ventilating the hold had occurred for several days
Sine the wind before November 25th had for four days been blowing
from some northerly direction, the air over the region must
have been abnormally cold. With the lulling of the wind, the
air-mass would begin to warm, the process proceeding upward
from the water’s surface. This would set up a steepening
temperature lapse rate, and produce the state of convective
instability precedent to formation of a waterspout.
On the assumption that the vessel was struck by a waterspout,
some such situation as the following becomes reasonable.
A fully developed spout is observed approaching from the southwest.
It bears down on the vessel so rapidly that nothing can be done
to mitigate the onslaught. The whole shop’s company gathers
in a terrified huddle. The southerly blast to the right of the
center of the vortex strikes the vessel, causing it to lurch
violently to port; as it is being righted the tube passes over;
its circling wind from the north careens the ship to starboard,
the deckload of water shipped when the Mary Celeste careened
swoops across the deck.
Passage of the waterspout’s vortex over the vessel would
require but a few seconds, the twinkling of an eye, and in that
moment the entire ship’s company, and the boat they hoped
would save them, are swept into the sea, which buries its secret
with the dead.
October 10, 1942