Page 1 of 9
OLYMPIC & TITANIC: REFINING A DESIGN
By Mark Chirnside
This article was first published in the British Titanic Society’s
Atlantic Daily Bulletin December 2019: Pages 18-22.
Author’s Note: Back in 2005, I published information about some
previously unknown refinements to Titanic based on experience
Harland & Wolff gained from observing Olympic during a
particularly severe North Atlantic storm in January 1912. The
article was published on the Titanic Research & Modelling
Association (TRMA) website. It discussed some modifications to
some of her rivetted joints fore and aft: Olympic’s great length
meant that the stresses at these points – from about a quarter of
her length ahead of the stern and a quarter of her length abaft the
bow – required some additional reinforcement, beyond what
previous experience had suggested was necessary, to prevent
rivets in these areas becoming gradually slack in severe weather
conditions.
It goes to show how much we are still learning about the ‘Olympic’
class ships all these years later, but the demise of the TRMA
website offered an opportunity to publish this new article. It
contains the original article’s information about the changes to
Titanic, supplemented by additional material, including new
diagrams of both Titanic and Britannic, and contextual
information about other large liners of the period.*
B
y the time Titanic was completed she incorporated hundreds, if
not thousands, of changes compared to her older sister: ranging
from the more obvious changes, such as her first class
accommodation on B-deck, to an improved propeller configuration
which Harland & Wolff estimated would increase her speed by between
* When The ‘Olympic’ Class Ships: Olympic, Titanic & Britannic was published
(History Press; revised and expanded edition, 2011) I included this information on
page 226. For an analysis of these changes and their potential impact on Titanic,
see Parks Stephenson’s article ‘What Caused Titanic to Sink?’ in the Titanic
Historical Society’s Titanic Commutator 2014: Volume 39 Number 206. Pages 92100. See, also: Rudi Newman’s ‘A “Riveting” Article – an Historical Rejoinder to
Metallurgical Studies of the Titanic Disaster’ in the British Titanic Society’s Atlantic
Daily Bulletin 2012: Pages 18-30.
– Mark Chirnside’s Reception Room: www.markchirnside.co.uk 2004-present –
Page 2 of 9
one-eighth and one-quarter of a knot. These refinements represented
practical experience supplementing theoretical knowledge as the
shipbuilder learned from Olympic’s operation.
Olympic passed the Daunt’s Rock Light Vessel at 2.11 p.m. on 11 January
1912, starting her first westbound crossing of 1912. Her log abstract
recorded ‘rough sea’ on the first day out, followed by ‘snow’ on the
second day, then ‘fresh NW west winds to S’ly to whole WNW Gale’ and
‘Whole NNW Gale, Heavy Squalls, High Sea’. Olympic’s progress slowed
as the sea conditions deteriorated: the first day’s run of 485 miles fell
to a run of 389 miles on the third day and then only 301 miles on the
fourth day out. (On a normal westbound day, she might have been
expected to cover well over 500 miles.) The weather improved towards
the end of the voyage but, by the time she arrived at the Ambrose
Channel Light Vessel, her average speed was a mere 18.62 knots. This
was a strong indication of how bad the weather had been. Based on
average speed performance data covering almost 90 per cent of the
voyages she made, it was the third-slowest westbound crossing of her
entire career.
The New York Times reported she had ‘experienced very rough weather
on the voyage, and in spite of her enormous size she shipped one huge
sea over her bows on Sunday afternoon that tore off the cover of No.1
hatch on the foc’sle deck and lifted it bodily over the guard rails at the
break of the deck and deposited it safely on the well deck below’. It said
J. Bruce Ismay ‘had one of his ports smashed by the big sea’:
He has come over to discuss business in New York and also to see
how the Olympic behaved in bad weather. He found that, in spite of
her size, the ship rolled and pitched a good deal, but that her
promenade decks were dry and that the majority of passengers were
able to go into the saloon for their meals.
Harland & Wolff also gained valuable information. It was, sometimes,
only through observing their designs in service that shipbuilders were
able to confirm something was working well or identify improvements.
(Early in January 1911, for instance, Cunard were discussing plans for
their new Aquitania with their shipbuilder and Lloyd’s registry. They
drew on experience with Lusitania and Mauretania, which had been in
service since September and November 1907, respectively. The plans
showed ‘treble [original emphasis] riveted landings in bottom plating.
Lusitania and Mauretania have double [original emphasis] riveted
– Mark Chirnside’s Reception Room: www.markchirnside.co.uk 2004-present –
Page 3 of 9
landings. Olympic [and] Titanic…have treble riveted landings. As the
Lusitania and Mauretania have shown no signs of weakness in the
riveting of bottom shell plating it was agreed to accept double riveting
for shell landings on bottom, with the exception of the strake in way of
the outer line of docking blocks on each side…’) Harland & Wolff were a
proactive shipbuilder, quick to apply their practical experience to
supplement previous theoretical knowledge.
As a result of observing Olympic in the storm, which turned out to be
one of the worst of her career, Harland & Wolff decided to make a
number of refinements, including changes to Number 1 Hatch. On 13
February 1912, Francis Carruthers, Ship Surveyor to the Board of Trade
at Belfast, also reported that the shipbuilder was making changes to
Titanic as a result of her older sister’s experience. They were fitting a
one-inch-thick steel ‘strap’ on the port and starboard sides of the ship
‘in way of no. 6 boiler room and extending three frame spaces forward
of the watertight bulkhead at the forward end of the boiler room.’ The
strap extended from frame 63 to frame 81 at the landing of strakes J and
K, at the ‘upper turn of the bilge.’ At this area, the hull frames were
spaced thirty-six inches apart (the furthest distance between frames
throughout the entire ship).
Further aft ‘in way of the turbine room and extending two frame spaces
into the reciprocating [engine] room’, another one-inch-thick ‘strap’
was fitted from frame 50 to frame 73 at the landing of strakes K and L.*
In addition, ‘one extra row of holes has been drilled in the plate above
the landing, making it a quadruple riveted landing.’
Carruthers noted: ‘I am informed that this strengthening is in
consequence of observations made on board the Olympic during a
recent heavy passage across the Atlantic’.
* On both the port and starboard sides of the hull at this landing the condenser
injection openings, which were large and rectangular, probably contributed to
additional ‘working’ in this area on the Olympic. In spite of the double plating
already provided, Scott Andrews points out: ‘I’m sure the presence of the large
rectangular openings in what is basically the bottom two corners of the box
girder formed by the hull caused these seams to “work” a bit more than those of
the neighbouring strakes’.
– Mark Chirnside’s Reception Room: www.markchirnside.co.uk 2004-present –
Page 4 of 9
Left: An extract from a
midsection plan of Olympic and
Titanic, showing a cutaway view
of the ship at boiler room 2. The
letters in red mark the locations
of the strakes of hull plating, J, K
and L. The thick red lines mark
the approximate location of the
rivetted landings between (top)
strakes L and K and (bottom)
strakes K and J. (The Shipbuilder,
1911/Ioannis
Georgiou
collection)
Below: The ship’s hull shape changed considerably over its length,
narrowing towards the extremities of the bow and stern as the width
of the ship progressively decreased. This plan shows the area located
at frame 72 forward. (Bob Read © 2019)
Following a request for further information from London on 22
February 1912, Carruthers confirmed that the joints in question were
double hand rivetted joints. (This was a contrast to other similar joints,
which were double or treble hydraulically rivetted.) Near the end of the
month, the Board of Trade decided to take the opportunity to examine
Olympic when she was drydocked for the replacement of a port
propeller blade. Accordingly, Carruthers made a detailed inspection
and reported on 6 March 1912. The Board were concerned that there
might be other signs of stress, beyond the specific riveted joints which
they knew were being modified. Carruthers’ report allayed those fears:
Below the waterline starboard side forward in way of no. 6 boiler
room in the shell landing of J & K strakes from frame 63 to 74,
about 160 rivets were slack and were drilled out and & renewed.
– Mark Chirnside’s Reception Room: www.markchirnside.co.uk 2004-present –
Page 5 of 9
About four feet below this landing in the tank bar, from frame 71
to 75 about 50 slack rivets were drilled out & replaced.
Port side fwd. [forward] in the tank bar from about frame 71 to
78, about 90 rivets were showing a little slack & were caulked.
On both sides aft in way of turbine [engine] room, in the shell
landing of K & L strakes from about frames 52 to 69, 100 rivets
found slack were drilled out & replaced. I carefully inspected
the vessel inside in the neighbourhood of these slack rivets
but found no further signs of stress. [my emphasis]
The Board of Trade’s estimate of stress ‘on rivets due to shearing’
showed that increasing the number of rows of rivets in a joint could
make a substantial difference. A double rivetted landing had a stress of
9.8 tons; a treble rivetted landing a stress of 6.5 tons; and a quadruple
rivetted landing a stress of 4.9 tons.
Olympic herself saw similar, or identical, changes to Titanic. On 22 May
1912, the Board of Trade noted: ‘Some of the landings near the bilge
were only double riveted & gave trouble. Inside straps have now been
fitted’.
Nor did Harland & Wolff forget Britannic. She had been laid down on
30 November 1911 and her double bottom was not yet completed, so
they had ample opportunity to make sure her design was refined to take
into account their experience operating Olympic. On 5 February 1912,
Edward Wilding had already noted that ‘This landing [between strakes
K and L], & any other coming above the tank side level, to be treble
rivetted between frames 39 & 85 forward & aft (about)’.
The available evidence is that Harland & Wolff’s changes did their job
in correcting what could have become a recurring maintenance
nuisance during the ship’s annual overhauls.* There is no record of any
* It is ironic that these changes were misrepresented later on by conspiracy theorists.
When I published my article originally, my argument was that these minor changes
demonstrated the fundamental strength of Olympic. She came through one of the worst
North Atlantic storms in her career and did not show any significant weakness.
Carruthers’ careful inspection of the hull revealed ‘no further signs of stress’ beyond
the small number of slack rivets identified. The entire purpose of his inspection was to
look for such signs of stress, but he did not find any. However, writer Brad Matsen
claimed it was evidence ‘that Olympic’s hull was cracking’, which is completely the
opposite of what the documentation shows. (See Matsen, Brad. Titanic’s Last Secrets.
Twelve; 2008. Page 295. A detailed analysis of these and similar false claims can be found
– Mark Chirnside’s Reception Room: www.markchirnside.co.uk 2004-present –
Page 6 of 9
further issues on Olympic. Several years later, Edward Wilding noted
that: ‘We have had less repairs to the Olympic than to any large ship we
have ever built, due to external causes, of course’.*
COMPARISON OF TITANIC & BRITANNIC
Above, top: Longitudinal view of Titanic with the locations of the
‘straps’ fore and aft marked in red. (Bob Read © 2019)
Above, bottom: Longitudinal view of Britannic, in her completed
configuration, with the locations of the treble rivetted joints specified
by Edward Wilding marked in red. (Bob Read © 2019)
in my article, ‘Titanic: Allegations & Evidence’, published in the Titanic International
Society’s journal Voyage 94 December 2015: Pages 55-60 and online:
http://www.markchirnside.co.uk/TitanicAllegations&Evidence.htm.)
* In total, 310 rivets were replaced and another ninety were caulked on Olympic in
March 1912. Similar, more serious repairs needed to be undertaken to a number of
her contemporary peers. Shortly after Majestic entered service in 1922 she needed
thousands of rivets in the bottom of the hull caulking, and similar issues were reported
in the case of Empress of Britain and Aquitania.
– Mark Chirnside’s Reception Room: www.markchirnside.co.uk 2004-present –
Page 7 of 9
Above and below: These diagrams show close up views of the location
of the ‘straps’ added fore and aft on Titanic. The thicker vertical hull
frames are ‘web’ frames, which added enormously to the strength of
the hull structure. Unlike some other ships of the period, whose hull
frames were numbered consecutively, on Olympic and her sisters the
numbering system started from the middle of the ship. Therefore
numbers were duplicated and we see frame number 60 (forward) and
frame number 60 (aft). (Bob Read © 2019)
Liners such as Cunard’s Aquitania and HAPAG’s Bismarck, which
followed Olympic, had additional rivetting in landings at about a quarter
of the length of the hull from both the bow and stern, because
shipbuilders believed that this was where high shearing stresses might
be expected.
Bismarck, which was launched in 1914 and completed after World War
I, became White Star’s Majestic (1922). She was the largest liner in the
world when she entered service and the longest, as well, with an overall
length of 956 feet. During a drydock inspection near the end of March
1925, three years after she had entered service, a surveyor noted:
– Mark Chirnside’s Reception Room: www.markchirnside.co.uk 2004-present –
Page 8 of 9
‘Several hundreds of “weeping” rivets in the flat of the bottom and at
quarter length from forward have been caulked’. This was, of course,
underneath the ship rather than at the side of the hull, however the
Board’s Principle Ship Surveyor asked ‘I should like to know whether
the landing edges of the shell plating at ¼ length from forward and aft
in way of the neutral axis [36 feet above the ship’s keel, or around the
usual waterline] are double or treble riveted’.
On 20 April 1925, the Senior Ship Surveyor replied: ‘The landing edges
of the shell plating are riveted thus: XW and WV, treble for full length;
VU and UT treble for 6/10 lengths amidships; TS, SR and RQ treble at
zones between 100 feet and 320 feet from each end of vessel. Rest of
landings double’. The letters for the relevant strakes of hull plating and
corresponding landings are not comparable to the lettering scheme
used on Olympic or her sisters, however it is interesting to note that the
shipbuilder had specified an increase from double to treble rivetting for
these landings. These landings represented about a quarter of the
ship’s length (between perpendiculars) fore and aft.*
Cunard’s Aquitania had similar design features, whereby the number of
rivets in these areas was increased. The Shipbuilder noted:
The seams of the side plating in general are treble-rivetted for
three-fourths of the vessel’s length amidships and double rivetted
beyond this length; but for those portions of the side plating at
about the one-fourth length of the ship from each end, where the
maximum shearing stresses may be expected, the seams have
been quadruple rivetted.
*
*
*
The progressive changes we see from ship to ship aid our understanding
of shipbuilding before World War I, as builders such as Harland & Wolff
supplemented theoretical knowledge with practical experience down to
even the smallest details. Even today, there is still much to learn.
* On Majestic, strakes X and W were the side plating of D and E-deck; W and V, Edeck; V and U, F-deck; U and T, F and G-deck; T and S, G and H-deck; S and R, Hdeck; R and Q, H and J-deck. These strakes were not near the turn of the bilge,
but close to the normal waterline.
– Mark Chirnside’s Reception Room: www.markchirnside.co.uk 2004-present –
Page 9 of 9
ACKNOWLEDGEMENTS
I repeat my thanks to Bruce Beveridge, Scott Andrews and Steve Hall for
their guidance and assistance with the original article. I am grateful to
Ioannis Georgiou for sharing generously from his collection and Bob
Read for his extensive work drawing and supplying the longitudinal
views of Titanic and Britannic, as well as the diagrams illustrating the
changes.
– Mark Chirnside’s Reception Room: www.markchirnside.co.uk 2004-present –