Thursday 16 April 2015


Erebus drifts south in the ice, a still from Franklin's Lost Ships.
© Lion Television and 90th Parallel Productions.
Image provided by Andrew Gregg, used with permission.

Last week, Lion Television, in collaboration with 90th Parallel Productions, aired a wonderful new documentary, Franklin’s Lost Ships, on CBC’s The Nature of Things. The show was directed by Ben Finney and produced by Andrew Gregg for CBC, PBS, and Channel 4.

The documentary chronicles the 2014 Victoria Strait Expedition, culminating in the discovery of HMS Erebus by Parks Canada archaeologists and their collaborators. The film uses detailed computer graphics, historical reenactments, and interviews to bring the Franklin mystery into sharp focus - in ways we have never seen before.  

One of the most unique aspects of the documentary is the use of computer animations to bring Franklin’s ships, HMS Erebus and Terror, to life. I was fortunate to be asked to consult on the team’s attempt to recreate the ships and I have been corresponding with the producers since December of last year. I know that I wasn't the only researcher they consulted on the ships, so it was definitely a crowdsourced project.

Erebus and Terror push their way through the ice - a still from Franklin's Lost Ships.
© Lion Television and 90th Parallel Productions.
Image provided by Andrew Gregg, used with permission.

Andrew Gregg, a producer for 90th Parallel, oversaw the model construction process, which was undertaken by Redlab Digital, a Toronto-based visual effects studio. Mike Brown of Lion Television was also heavily involved, and Karen Kershaw led the artists at Redlab. The team’s goal, as stated to me, was to produce the most historically accurate reconstruction of the ships possible. My admittedly small role was to provide feedback and comments on the various drafts of the model build.

I sent Redlab copies of my plans and they consulted historical data such as paintings, drawings, contemporary ship models, and the original draughts of the ships. As they constructed the model, the artists would provide rendered images of different angles of the vessel to Andrew and Mike, who would then pass them on to me for comment. The process involved many drafts and I marked-up many images as, polygon by polygon, the ships began to take on a recognizable shape.

Until late February the ships were untextured grey geometric models, but I remember smiling broadly when Andrew emailed the first textured (fully coloured and detailed) render. Here was a complete view of one of Franklin’s ships, with the correct colour scheme and all of her unique systems, and I was thrilled to see it.

Fully textured render created by Redlab, showing a view of the hull of the ship. Note the iron bow plating.
© Lion Television and 90th Parallel Productions.
Image provided by Andrew Gregg, used with permission.

Fully textured render created by Redlab, showing a view of the upper deck. Note the diagonal deck planking.
© Lion Television and 90th Parallel Productions.
Image provided by Andrew Gregg, used with permission.

Fully textured render created by Redlab, showing the ship
covered in ice and snow.
© Lion Television and 90th Parallel Productions.
Image provided by Andrew Gregg, used with permission.

Of course, readers of my blog know that I could never be fully satisfied with the accuracy of any model, and I worried that I would annoy Andrew and RedLab with all of my suggestions for changes. However, I learned quickly that they wanted to include every detail they possibly could, and they made changes to the models right up the last minute possible before their airdate.

Erebus and Terror sail again, a still from Franklin's Lost Ships.
© Lion Television and 90th Parallel Productions.
Image provided by Andrew Gregg, used with permission.

Are there errors in the models? Sure there are. But these are relatively minor given the project and they are primarily because of time constraints. There are issues with the ice channels, the masts and rigging, and the various esoteric fittings bolted to the ship. But I suspect there are only a handful of people on the planet who could point out all of the specific issues (and most of them are currently diving on the Erebus).

I’m convinced that these are the most accurate models of Franklin’s ships ever committed to film, and I want to thank Andrew and Mike for involving me in their great project. I’ve spent the better part of two years trying to build a wooden model of HMS Terror, but I never thought I’d have the opportunity to see her sail in arctic waters again. Watching Erebus and Terror plough their way through the ice allows us to better comprehend the astonishing journey of these remarkable exploration vessels.

If you read my blog, don't miss this wonderful documentary.

Erebus takes advantage of her screw propeller. A still from Franklin's Lost Ships.
© Lion Television and 90th Parallel Productions.
Image provided by Andrew Gregg, used with permission.

Acknowledgments: I would like to thank Andrew Gregg and Mike Brown for allowing me to post renders of the ships and stills from the documentary. It was a pleasure to work with them.

Tuesday 14 April 2015


I haven't posted an update regarding my model in several months. While I've kept busy with side projects, the real reason for my delay is that I had reached an impasse with Terror’s stern.

As I've discussed in previous posts, the sterns of Franklin’s ships were modified in 1845 to accommodate a new auxiliary screw propulsion system – to be used as a time saving device “providing the wind should prove contrary or a dead calm” [1]. There are two sources of data on these modifications: Oliver Lang’s original design plan [1], and its counterpart, a contemporary model of the design [2]. I had purchased full resolution copies of the plan many months ago, but unfortunately Lang did not include a cross section in his draught. That information could only be gleaned from the contemporary model held at the National Maritime Museum’s storage facility in Chatham.   

The contemporary model of Oliver Lang's 1845 design. 
National Maritime Museum, Greenwich, London (SLR2253 [L2251-001]).

Fortunately, I recently had an opportunity to visit the Chatham model ship facility. Assisted by the expert curators, I was able to study the stern model in detail. It is quite unique, being constructed using a series of carved blocks arranged to conform to the position of major structural and engineering elements of Lang’s design.  The information I gathered has allowed me to complete my construction of the stern;  below, I’ll reveal the new information I've learned from the contemporary model, while documenting my final assembly of Terror’s stern:

1) The propeller well used to raise and lower the screw was rectangular, almost square-sided, with the sternpost and rudderpost forming the fore and aft sides of the well, respectively. To accomplish this, thick timbers were bolted to the sides of the rudderpost and sternpost [3]. The rudderpost bolsters were much more complex than I originally assumed and were each constructed of at least two pieces, with the lower portions tapering gently to the width of the rudderpost, following the lines of the body plan (see here for my original conceptualization of the design).

The stern pieces prior to assembly. The bolster on the left is the old design I intended to use, which was incorrect. 

The overkill method I used to glue the bolsters to the stern and rudderposts. Thankfully this was just a dry-run
(note the older bolster design). 

The  new bolster timbers glued on the rudderpost. Note the groove for the "Lihou" rudder on the
rudderpost. I may need to sand the bolsters somewhat to match the run of the planking as they
may be slightly oversized - but no by much. 

Another angle showing the bolster timbers on the sternpost. The NMM model shows that the bolsters on the
rudderpost are longer than those on the sternpost. 

2) The rudderpost and sternpost were each tenoned into the keel extension, as was typical, but each was secured with a single bolt, which was not indicated on Lang’s plan.  

Marking the precise position of the tenon bolts. 

The bolts were simulated with 20 gauge copper wire, precisely the same as that used on the keel scarphs. 

3) The propeller well was framed on the port and starboard sides in three distinct sections. The upper section included stout rectangular framing fayed to the deck beams, which formed a ledge for a scuttle on the upper deck. Below this, the well was probably enclosed by watertight planking down to the height of the stern timbers. Because of the construction of the contemporary NMM model, such planking was not shown, but it is unlikely that solid timber pieces would have been used, as these aren’t shown in contemporary models.

The heavy framing used to form the top of the propeller well. The upper part of these timbers formed a lip
for a scuttle to the well. 
Planking on the upper section of the well. I've estimated a width of 12 inches. The actual width is unknown.
Note that this section of the model will be covered so I haven't simulated bolts or spikes here. 

A view of the topside of the well. The upper pieces of the sternpost and rudderpost bolsters will be
trimmed at a later stage of the build , but are useful for alignment at this stage. 

4) A new section, clearly visible in the well of the model, started at the position of the stern timbers. This suggests the stern timbers were bolted to the sides of the rudderpost and sternposts to provide major structural support to the new rudderpost and well. This makes good sense, and Lang’s 1845 stern plan clearly shows the stern timbers as a major element of the design. In fact, these new timbers are substantially more robust than Terror’s original stern timbers, suggesting they were an integral part of the strength of the new structure. Again, this type of structure is supported by contemporary models.

The bottom portion of the framing planks were trimmed to match the run of the stern timbers. Note
the rabbet on the rudderpost on the right. 

5) The lower section of the propeller well was composed of the second layer of hull planking where it ran aft, horizontally.  Eventually, the run of the higher planks would have veered away from the straight-sided wall of the well. At this point, straight horizontal planking would have been used to frame the sides of the well. The position where this occurs is marked by a block seam on the contemporary NMM model.

Unfortunately, Lang’s contemporary  model does not include any of the ironwork used to strengthen the stern, nor does it include the propeller rail/track mechanism. I've based these portions of the model on Lang’s plans and extensive research on other contemporary models and designs. This research is outlined in several blog posts (and here, here, and here).

Oliver Lag's stern design. Note the extensive ironwork and the propeller systems.
National Maritime Museum, Greenwich, London (ZAZ5683 [J1529]).

The iron staple knee glued in place. The knee provided essential support for the rudderpost. 

Mini-Crozier inspects the staple knee in dry dock. 

Lang used iron strapping to further reinforce the stern structure. Here they are made from chemically blackened copper. 

Each strap was glued in place and then the bolt holes were drilled out by hand. 

Bolts glued in place. These were simulated using blackened brass. 

Another view of the completed iron work.

Mini-Crozier frets over the modifications. 

The staple knee was protected by a fitted  chock bolted to the keel section. I carved this using a simple
chisel blade. 

The finished chock compared to the plans. 

Image showing how the chock fits over the knee. Unfortunately it had to be glued in place to permit 
the propeller rails/tracks to be installed. At least I know the knee is there. 

The chock glued in place.

The propeller was raised and lowered using rails or "tracks". These have been modified slightly
from my original versions based on new data. Copper bolts were simulated using wire. 

The rails glued in final position. Note  the  rabbet on the rudderpost 
for the second layer of hull planking. The rabbet will be modified  to 
accommodate the precise run of planking when it is installed. 

 View of the rails installed on the sternpost. 

View of the rails installed on the rudderpost. 

Another view. 

Wooden bolt plugs added to the chock. The bolts were "counterbored and plugged".

The staple knee was bolted to the rudderpost; these bolts were also counterbored and plugged.
I'm not entirely happy with the contrast here and may redo them at a later date.

The completed stern assembly.

Lowering the screw propeller in place (it raise and lowers - and the propeller spins). 

The propeller in position. Unfortunately the angle of the photo makes it look slightly crooked,
but it is not - is spins freely, with very small tolerances as shown on Lang's original plans. 

A view from the stern.

Another angle showing how the propeller was seated.  

Looking down the well from the position of the upper deck .

Mini-Crozier contemplates how the stern will fare in the ice. 

How successful was Lang’s stern at protecting the ship from the pack ice? Parks Canada divers are assessing that currently, and with luck they’ll find the answers soon. We know from historical sources that the Admiralty was concerned about the strength of the design, and that while Lang believed the “sternposts” (sternposts and rudderposts) were as strong as those on other ships, he would not certify that the strength of the filling chocks was sufficient to protect the Erebus and Terror [4].

No matter how vulnerable it made the ship, we can suspect that Lang’s radical redesign also altered the sailing qualities of Terror. Contemporary sailing reports indicate that Vesuvius class bomb vessels were rather lumbering and could not carry sail well, and Ross reported that Terror was constantly falling behind Erebus during his Antarctic voyage, delaying and endangering the expedition.

Recently, Regina Koellner, assisted by William Battersby, transcribed a letter [5] from Francis Crozier to his friend John Henderson, written shortly after the ships arrived at Whalefish Islands in Greenland. In the letter, Crozier provides a brief report of Terror’s sailing qualities: "Our steering is decidedly improved by the alterations on the counter we now sail much more evenly with Erebus which is advantageous to us in many ways." I suspect that the effective lengthening of the keel to accommodate the propeller allowed Terror to sail closer to the wind, finally permitting her to keep up with the more nimble Erebus. It seems the final conversion of Terror to screw propulsion made her a more capable vessel under sail, an irony certainly not lost on Crozier.

I sincerely  thank Regina Koellner for permitting me to post the transcription from her recent research, and William Battersby for alerting me to this exciting primary report on Terror’s sailing qualities.

[1] Object ID: ZAZ5683

[2] Object ID: SLR2253

[3] Filling frames, similar to fashion pieces, appear to have been fayed to these bolster timbers on the sternpost to accommodate the smooth run of planking over the upper part of the well. My belief is that space between these frames and behind the transom pieces was filled in with chocks, similar to the bow, to add more strength to the stern.

[4] Admiralty Digest, cited from Battersby, William, and Carney, Peter, 2011, Equipping HM Ships Erebus and Terror, 1845. International Journal for the History of Engineering & Technology 81(2):192-211.

[5] AGC/C/5/; MS62/007