Project Background

In the late 1990s, my mother mentioned that she had acquired an old metal dispatch chest from her sister after she died that contained a bunch of diaries their father had kept of his travels across the world between 1906 and 1910. Joseph C. Pearson had previously acquired the equivalent of a Masters degree in physics from Harvard University. After teaching physics and math at Bowdoin College in Maine for a few years, he had been hired by the Department of Terrestrial Magnetism of the Carnegie Institution in Washington, DC (DTM/CIW; today incorporated into the Carnegie Science Earth and Planets Laboratory), as a geomagnetic data observer aboard the US West Coast brigantine Galilee. This vessel was originally a Pacific South Seas packet ship, built by Captain Matthew Turner in 1891 at his shipyard in Benicia, California. As far as I can tell, this ship still holds the San Francisco to Tahiti speed record under sail!

Joseph Cleveland Pearson—my grandfather (c. 1907, photo courtesy of Carnegie Science Library)

Captain Matthew Turner aboard Galilee (c. 1905, courtesy of Carnegie Science Library)

The Galilee herself was chartered by DTM after the laboratory conducted a search for a suitable vessel to be a testbed of sorts for developing and employing new scientific techniques for measuring the earth’s magnetic field at sea. This kind of work had been done only a few times before in maritime history but with inconsistent results. This lack of success was mainly due to the vagaries of the ocean environment as well as sea-going vessels’ extensive magnetic signatures from iron fasteners and fittings, which affected the magnetic data. For Galilee's charter period, the type and locations of all ferrous fasteners and materials were documented. During her three charter cruises throughout the Pacific Ocean between 1905 and 1908, the non-essential iron components were, as much as possible, progressively removed without compromising the ship's safe operation.

Broadside view of the Galilee configured for her geomagnetic survey work in the Pacific Ocean (courtesy of the Carnegie Science Library, c. 1907)

Concurrently during Galilee’s charter period, DTM was building a nonmagnetic sailing ship for this purpose, the Carnegie, which could be the topic of another story . . .

Original Inspiration

This reconstruction project was instigated by the receipt of my grandfather’s diary in 1998, following the death of my eldest aunt. It partially documented the return leg of the second DTM charter voyage in Galilee. The diary is comprised of eight lined pages bound by a blue ribbon along their left edges, with a thread stitching them together. The diary was hand-written in black ink on both sides (but included pencil annotations). Its narrative covered the dates 8 September to 21 October, 1906, which described the ship’s return to San Diego, California, beginning several days after departing Yokohama, Japan. Unlike my grandfather’s later diaries that he kept while employed by DTM, this fragmentary document was probably the only remnant of a lost diary of this particular cruise. Or, more likely, a diary he cobbled together after leaving port, having lost the original diary. The ship had been sunk along a breakwater after dragging her anchors during a typhoon that hit Yokohama harbor on 24 August, 1906. The details of this storm were described by the expedition’s commander, J. F. Pratt, in a letter to his wife.

Two pages of my grandfather's Galilee diary—or what's left of it! (Personal photo)

The Galilee alongside a breakwater in Yokohama harbor after the typhoon in August, 1906. (Courtesy Carnegie Science Library)

Having a maritime background myself (see my profile), I was taken with the prospect of perhaps building a model of the Galilee that would become part of our family’s legacy. The main problem with this idea was acquiring reliable plans and information about the ship itself ...

The Search for Plans and Information

(Continued from my previous post)

My first task was to locate some information about the ship. According to my project journal, my initial source was the San Francisco Maritime National Historical Park (SFMNHP) museum, from which I obtained a set of basic hull lines of Galilee that were originally published in The Rudder magazine in 1899 (see following images). The plans were signed by C.W. Davis, who was the author of many of our classical shipmodeling reference books. The librarian there suggested that I contact the San Diego Maritime Museum. They provided me a copy of their periodical, Mains’l Haul, which contained an overall history of the ship, as well as the same lines plans. These are the relevant pages from The Rudder magazine.

Pages from an 1899 edition of The Rudder magazine. (Public domain)

These plans seemed like a good start for understanding the ship. However, several years later, when I actually started developing my own, larger scale plans, things began to fall apart. The waterlines in the profile view were not parallel to the keel—they diverged from the stern to the bow! So, this situation clearly wouldn’t produce any consistency among the three views. I suspected this was a draftsman’s error when the plates were prepared for the magazine. I doubt that an authority such as Mr. Davis would have been responsible for such a gaff!

During the same time-frame, the SFMNHP librarian had connected me with the librarian at the DTM library (now Carnegie Science Library) who was Shaun Hardy at the time. He provided me with a trove of printed content from the DTM archives to help me on my way.

Early in 2001, I had been participating in a forum on Yahoo Groups (remember them?), where I met a ship modeler, John Kowalla, who periodically conducted model shipbuilding workshops at the SFMNHP. He suggested that I should check out the Smithsonian Institution’s collection of ship’s plans, which include the plans produced by the Historic American Merchant Marine Survey (HAMMS) project from the late 1930s.

I ended up purchasing from the Smithsonian a large scanned hard copy of the Galilee produced by G.C. Berger of the Pacific Marine Research Society, San Francisco. Someone told me that the Society was a predecessor to the Nautical Research Group (NRG). These plans became my go-to source for reconstructing Galilee’s hull. As it turned out, these plans were riddled with image and dimensional errors due to a history of multiple scans at the Smithsonian. The staff there evidently didn’t keep track of which sheet was the original, resulting in their archived source copy for subsequent scans being a scanned version itself (i.e., a scan of a scan of a scan ...). Consequently, the image had lost its original resolution, with blurry lines, and was distorted in a number of ways that required significant rework over the following years.

Copy of the G.C. Berger plans as obtained from the Smithsonian Institution. The lines and fidelity of the views were terrible!

During the summer of 2001, I had the privilege of visiting the DTM (now Carnegie Science) Library in Washington, DC. Mr. Hardy spent an entire day sifting through their archives to provide me with dozens of photos of the ship actually taken onboard during her charter period, as well as numerous other documents and sources. This was a windfall of details normally not available to a researcher of ships more than 100 years old!

The library held several original plans, apparently drawn by non-nautical DTM engineers, based on the fact that they omitted key information from their drawings, such as whether the plans were exterior dimensions or moulded dimensions, as well as scale bars and other similar content we have come to expect in ship's drawings. These are some examples of the kinds of plan I was able to obtain (the DTM plans were created around 1905):

Profile and halfbreadth plans of the ship. No scale bar so there is no way to determine the actual dimensions. Also, no indication if this was a molded or outboard plan.

Illustrative sail plan. Dimensional analysis later determine this was not to scale. (DTM/CIW)

Body plan view of Galilee showing the typical fastener placement in a mid-body frame. Not to scale/for information only. (DTM/CIW)

A reference plan of the Galilee showing deck beams and cabins immediately before her conversion to her charter configuration.

Many details in this drawing were not consistent with photographic evidence obtained later. (DTM/CIW)

Without belaboring the topic, the DTM Library was an extraordinary help in providing the detailed information I needed to begin work on this project. In subsequent years, as even more content became available online, I was able to obtain additional plans of the ship after her DTM charter period, when she was sold to the Union Fish Company following the death of Captain Turner in 1909. There are several plans available through the SFMNHP library of Galilee as a bald-headed three-masted fishing schooner. These drawings were of limited use to me because of the changes in topside structures and rig.

One of several plans of the Galilee showing her as a three-masted fishing schooner after 1910. (Courtesy SFMNHP).

Interestingly, it has been documented that a diesel engine was added to the ship in this employment, but no evidence of a propeller can be seen.

The SFMNHP library also provided me a copy of a technical document that probably contains the best compilation of the history of the ship found in one place. This historical narrative was published by an academic named Richard A. Stradford, who wrote a paper for a presentation to the Corps of Engineers. He described most of the the ship's documented history and included justification for preservation via the National Registry, in a bid to conserve her remains. The paper was titled "Brigantine, Schooner, Houseboat: Journeys of the Galilee." I have attached the publication for readers' interest.

Important technical information regarding the ship was provided via another branch of the National Park Service (NPS). Gina Bardi connected me with an individual who was the project manager for a proposed Historic American Engineering Record (HAER) project. The intention of this project was to preserve what information that could be obtained regarding the Galilee in spite of the ship's advanced state of decay and dismemberment. The ship project was designated HAER No. CA-2271. The engineer in direct charge of the survey and data collection was Michael R. Harrison. His draft report was submitted in 2011. However, subsequently, a decision was made by the NPS to table the project, evidently because there was not enough left of the ship to create a report that adhered to the HAER program documentation requirements. I received a draft copy of the report that included laser scans of the ship's stern remnant plus several dozen photos taken at Fort Mason in San Francisco and in Benicia, California.

Well, I could go on for quite a while documenting many other sources of information, but these are the major ones. Not too many years ago, Gina Bardi, a SFMNHP librarian, informed me that they held a hi-resolution digital copy of the G.C. Berger plans that I had obtained from the Smithsonian. Its resolution was so fine and the file so large that they had to place the document on a special file transfer server to get it to me. Its lines are far cleaner than my hard copy and it has formed the base image for all subsequent work in the past three years or so.

My research not only turned up a variety of plan drawings of the ship, most of which I won't show here because they were obviously derivative plans and/or drawn up after the fact, but I was able to uncover some written histories of the ship as well. The DTM Library showed me original copies of the published DTM field reports as well as correspondence by the key players in the DTM's administrative staff regarding the utilization of the ship during the charter period. In more recent years, I have been able to recover detailed photos of the magnetic instruments, as well as other content pertaining to the ship's employment. One historical narrative was published by an academic named Richard A. Stradford, who wrote a paper for a presentation to the Corps of Engineers describing the ship's history and consideration for preservation via the National Registry, in a bid to preserve her remains. The paper was titled "Brigantine, Schooner, Houseboat: Journeys of the Galilee." I have attached the article for your interest.

So, I think this gives you a sense of the amount of information that was available when I finally figured out where to look for it! As appropriate, other sources of information will be presented in this thread that contributed to the creation of a digital model and plans.

To be continued . . .

Proceedings.11Stradford.pdf

Early Reconstruction Challenges

I’ll try to condense the first few years of my reconstruction work as much as possible for the benefit of others who may want to model this ship, because they were filled with frustrations and false starts—never having done this kind of work before. My research journal that I began back in 2001 is full of sketches and notes pertaining to details of the ship—cabin construction, windows and doors, the magnetic instruments, general notes on masting and sails, and so-on. These notes were derived from photos that I had obtained from the DTM Library.

As noted in the previous post, I was able to obtain a hard-copy of large-scale plans of Galilee as a brigantine (even though the undated plans—if compiled as part of the HAMMS program—had supposedly been developed by G.C. Berger decades after the ship had been converted to a three-masted fishing schooner). These plans were gifted to me by a fellow modeler, John Kowalla, sometime late in 2001. Because I intended to work up a set of digital plans, I had to figure out how to scan the sheet into a digital file. I found a Kinko’s store (remember them?) that could handle scanning a large-format document and obtained a reasonably good resolution digital copy of Berger’s plans.

The only attribution relating to the large-scale plans of the Galilee. G.C. Berger was identified as the draftsman. No date information. No provenance at all!

I had been using the 2D graphics program CorelDraw for technical and educational illustrations since 1993 (and still do), so my intent was to eventually create a 2D vector graphics version of Galilee’s hull using CorelDraw X4 at the time. However, my research journal noted numerous defects with the scanned plans:

• First of all, there were no indications of what constituted the Fore and Aft Perpendiculars, which define the registered length of the ship. Nada. There was just a forward station line somewhere aft of the bow and an aft station line 10-feet forward of the aft face of the stern post , neither of which corresponded to anything that related to traditional Perpendiculars. (This suggested that the aft face of the sternpost could be considered the Aft Perpendicular.)

• While the original overall scale of the plans was likely intended to be 1:48, that wasn’t what I had to work with based on distances between station lines (nominally 10 feet). The horizontal scale was actually slightly larger than that, while the vertical scale was spot on, based on the distance between waterlines. Again, this suggested issues with the scanning process.

• The distance between the nominal 10-foot stations was variable, up to several inches difference by scale.

• The station and buttock lines were not quite perpendicular to the waterlines. In other words, the plans were skewed—also likely a scanning artifact.

• I also noted in the body plan section that the distances varied between buttock lines—since these lines were perpendicular to the scanning direction, probably yet another scanning defect.

My decision at this point was to simply try and work around these issues without altering the plans.

All of these printed plan-related problems were noted in my research journal in late 2001 and early 2002.

While a lot of my preliminary efforts evaluating the plans remained undocumented in my research journal, I recall that over several years in the 2000s, there was a lot of fiddling around in CorelDraw, off and on, trying to get the station-, water-, and buttock-line intersections to correlate in the three views. That didn’t happen. Trying to accomplish any meaningful work in a 2D digital graphics environment just led to chasing the edits in a circle. Changing a curve in the body plan knocked the corresponding intersections out of place in the profile or halfbreadth plans, or often both. Around and around. So I gave up for a while, turning my attention to other projects, including transcribing my grandfather's 12 land magnetic survey journals between 1908 and 1910, and house renovations.

The latter years of the 2000s and into the early twenty-teens involved periodic forays into analyzing and resolving discrepancies among the the three main ship plan sources: the 1899 CG Davis plans, the 1905 DTM project engineering reference plans, and the "lack-of-provenance" G.C. Berger plans obtained from the Smithsonian. My journal documented key dimensional discrepancies among the plan sources for things like distances between top of keel and rail, length of hull between key landmarks (none showed standard FP/AP locations), sweep of sheer of deck and main railing, deck camber/cabin roof cambers, and so-on. Very frustrating for a novice ship modeler and researcher! And all this was documented only in 2D diagrams. At the end of November, 2012, I made the decision that the G.C. Berger plan would by my standard reference to this project. Any further modifications to the plan would be justified by dimensional analysis based on photogrammetric data. I also settled on a LBP of the registered 132.5 feet (the documented/registered length), which assumed the AP was the aft face of the sternpost and the FP being near the underside of the main deck where it intersected the stem. This seemed to work until I discovered the correct definition of LBPP applicable to the Galilee several years later. That will be a separate discussion farther on in this narrative.

Over a decade ago, in spite of home renovations and other commitments, I returned to working in the 2D digital environment as often as possible, without much overall real progress, until I joined the "other site," MSW, in December, 2014. My initial posts in that thread addressed the history of this project up to that point, describing the issues I had encountered with my hull plan resources themselves, as well as my frustrations working in the 2D digital medium. The biggest visual problem I had to deal with was trying to resolve the shape of ship’s transom, which affected the shape of the aft third of the hull. This issue was eventually resolved by switching to a 3D graphics program.

To be continued . . .

Transition to 3D Graphics Modeling Tools (and That Transom)

[Continued from previous post.]

While creating posts in this thread late in 2014 on the "other" site, I documented my wrestling with reconstructing the Galilee’s transom, which I felt was key to moving forward with converting all the information I had acquired over the previous 13 years into a usable 2D digital model. This lack of progress continued until March, 2015, when I discovered the free version of the for-the-purpose 3D naval architectural software DELFTShip-Free which was and is produced by a small Dutch engineering company. The free version is probably the best all-around digital program for modelers who seek to construct plans of any-sized vessel or reconstruct ship’s hulls from existing plans. As is the case with most advanced digital software, DELFTShip has a steep learning curve, and some of its ancillary features are quirky, to say the least!

The main and most useful advantages of this software are that, when one makes an adjustment in one dimension to revise an intersection line, you can see the consequences instantly in the associated intersections. Recall that, in the context of ship’s plans, an “intersection” is any line generated by the intersection of the 3D hull surface with:

• a vertical transverse plane erected at a station, or

• a horizontal longitudinal plane laid down at a waterline, or

• a vertical, longitudinal plane erected at a buttock line.

Since the most pressing issue in my project at this point in time was resolving the transom shape and its effects on the aft end of the hull, I began learning the program in that context.

This is the shape of the molded transom according to the G.C. Berger plans:

G.C. Berger body plan of Galilee. Note the shape of the transom, then compare it to the photos below. (Smithsonian Archives)

These are views of the shape of Galilee's transom documented by contemporary photos:

Different views of Galilee's transom (Courtesy of Carnegie Science Library, c. 1905–6)

As one can see, there are significant differences between the Berger plans' shape of the ship's transom and the real thing at the time of concern. The outboard edges in the photo are distinctly curvy compared to the line plans.

My project journal documents a prolonged period of time wrestling with the convex shape of the transom, which also factored into how it would intersect with the hull sides to create the curvy shape documented in the photos. I considered three cases:

• The least probable was that the transom was simply eyeballed by the shipbuilders to enclose the stern—or that it lay on a very large-diameter, conical surface. This was deemed to be too unlikely and too difficult to produce the uniform, eye-pleasing shape seen in the photos. In addition, the buttock lines formed by a conical surface would uniformly converge in the profile view. There was no clear evidence of that being the case visually or in the reference plans.

• It was possible that the transom lay on a tilted surface formed by a cylinder with an elliptical cross section. This option was evaluated with some success. However, it was considered more unlikely due to the extra effort to reliably acquire this shape compared to a cylinder with a circular cross section.

• Third—and there is ample evidence from the existing plans and photos for this—the curved surface formed a uniform horizontal curve (buttock lines on the transom face were parallel) and the face angle of the transom with the horizontal was consistent across the entire transom (also parallel buttock lines). This could result from the transom lying on a tilted cylindrical surface with a uniform, circular cross section.

After several weeks of tweaking this configuration, I tentatively decided that the transom probably lay on a skewed, circular cylindrical surface, which, in practice, probably was easier to develop on a horizontal lofting floor. The circular profile thus formed by the angled transom surface had a horizontal radius of about 52 feet.

Continuing with adjustments of the molded model surface of the aft quarter of the hull adjoining the transom, I achieved a relatively smooth hull surface under the counter adjacent to the transom. This work involved making many fine adjustments to the hull's surface, checking the buttock lines for farness, then tweaking the control net iteratively until the intersection of the transom edge with the hull was smooth, the hull surface was smoothly shaped, and the transom face contained straight, parallel buttock lines. Added to this was verifying that the DELFTShip program waterlines across the transom were smooth, circular arcs as well.

By April, 2015, after much trial and error, I was able to come up with a reasonable approximation of the aft half of the molded hull to work with, and from which I could begin tweaking the shape of the transom. Sadly, I don’t have any of these early screen shots to illustrate these efforts. The following image is a current reconstruction of how the transom approximately looked after adjusting the hull and transom surfaces to bring them into alignment, while staying as true as possible with the hull stations shown on the plans. This image also includes the yet-to be accomplished adjustments to the transom face itself to shape the edges to agree with the contemporary photos.

Current reconstruction of the transom as it approximately appeared about mid-2015. (DELFTship-Free)

The previous image illustrates all the features that I worked toward during 2015. Because this is how the model currently appears in DELFTShip, it also includes numerous minor tweaks added during the last decade. Those do not materially change what the goal was back when this digital model was begun. Note that the green station lines across the face of the transom are smooth ellipses, the blue waterlines are sectors of horizontal circular arcs, the tan buttock lines are straight, and the outer edges of the transom smoothly connect the transom to the molded hull, which are the results I was working toward 11 years ago.

As an aside, at this stage of the project, I experimented with using the Offset Table import feature that DELFTShip (supposedly) supports. I took the values off the CorelDraw plans after updating the transom construction and formatted the text file table according to the program manual. The import operation worked quite well, surprisingly. Sadly, that was the last time I was able to make it work, even in later updates of the program! This is what the aft end of the hull looked like (sans keel) with the Shaded surface option:

Solid molded hull (shape inside the planking) showing the use of "intersections" to check the fairness of the surface. The keel and rails were added later.

The dark edges are artifacts of the DELFTship program (Both Sides view). The waterline was just estimated in this version of the model.

In June, 2015, after adding the keel and adjusting the main rail edges, these are screenshots of the hull at that point:

More advanced versions of the ship's molded hull after numerous adjustments to the shape to bring it fair and smooth. The keel was included at this point. (DELFTShip)

At least one contemporary painting showed the ship with green antifouling paint.

To be continued . . .

Erecting the Hull—The Backstory

In the two previous versions of this thread on the “other” website, I had initially presented the incremental research progress and resulting developmental work as they chronologically occurred over the past 25 years since 2001. Since I was a novice nautical researcher for much of that time, the results of this trial-and-error effort predictably included having to revise/retract in the later posts former observations or conclusions.

To spare readers having to experience those issues, in this thread I'll just present the development of the model in a way that more-or-less reflects the construction stages of the actual ship.

As mentioned in the previous post, after struggling for more than a decade in the CorelDraw, 2D graphics environment, I initially began using DELFTShip Free in 2015 to create in a 3D environment the molded (inner) and outer hull surfaces to establish the shape of the hull. Then I added hull railings, deck surfaces, and keel. When I was satisfied with the fidelity of the basic modeled hull compared to the reference plans and DTM photos, then, from 2015 until 2023, I constructed in that program nearly all the major external items visible topside (cabins, deck furniture, fittings, instrument observation bridge, bulwark details, etc.).

Early during that latter phase, since the ship had open bulwarks, I realized that the exposed top timbers of the hull frames, along the middle portion of the hull, needed to correctly relate to the ship’s internal framing. So I decided to create, as much as possible from known references, the hull frames, deck beams, hold stanchions, beam shelves, and most other key structures not directly visible from outside the hull, just to settle in my mind that all these details would actually work together in my reconstruction. That didn’t go as smoothly as I had intended, which will be discussed in later posts.

As the number of modeled details accumulated, I faced greatly increased edit times using the DELFTShip software (the reasons for which I can share with anyone who is interested). So, in January, 2023, I switched to using the high-end graphics program Blender, which was much faster for the things I was doing with the ship model—let’s just leave it at that.

In order to not reinvent in Blender the parts that I had already built in DELFTShip (a.k.a., D/S), I have spent a good part of the past three years importing, as STLs, all the numerous detailed D/S parts into Blender at full size (1:1). The most annoying aspect of this work was removing from the imported objects the vast majority of triangular faces, which is the way STL objects are constructed and exported, in order to minimize the polygon and vertex counts in the final Blender model.

Along the way, I also built from scratch in Blender the remaining structural and hardware components that I had avoided in D/S when those had become too hard to deal with in that program. These new items included virtually all of the spars and masts, and some of the more complicated mechanical details (e.g., anchors and ground tackle gear, deadeyes, etc.).

One of the visual benefits (and pleasures) working in Blender is employing realistic 3D surface textures and related visual enhancements to the digitally-rendered model surfaces. These properties were added to the basic 3D-modeled objects during the 2023–2025 period. These embellishments included planked weather deck textures, reflective glass, specular metallic surfaces, and the subtle hull planking seams. These features would have been impossible to do in D/S. I relied on members skilled in digital-modeling in the ship-modeling forums, who are truly talented in Blender, to help me through these phases. I'm still learning.

Examples of metallic reflections and material textures on otherwise flat surfaces. (Blender)

Hull planking and image textures (Blender)

Environmental details in specular glass reflections (Blender)

So this discussion brings this topic narrative to the point of rebuilding the model—literally from the keel up. I will include some research and technical discussions of problems and the decisions I faced—or still have to deal with to this day.

I apologize for the wordiness of this post, but I felt it was necessary to set the stage for the following posts!

The Keel and Related Structures

According to the book, The Elements of Wood Ship Construction by W.H. Curtis (1911—from now on, referred to as The Elements), construction of a wooden sailing ship proceeded in a sequence of logical steps determined by the shipyard. While this reference was written more than 20 years after Galilee was launched, the general methods it describes were pretty much well established by the end of the 1800s. Commercial and private wooden-hulled ship construction around the start of the last century began with laying the keel on an inclined shipway. Next, the stem and sternpost, with the associated deadwood were erected. Then workers attached the square frames followed by cant and half-frames. In Galilee’s case, all her frames were square, and she employed squared half frames at the extremities of the hull, based on my direct observation of her bow and stern remnants in the San Francisco Bay area in 2001.

The Keel

One photo online depicts the beginnings of a ship’s keel (inverted) showing the characteristic round-up of a rocker-type bow, which Galilee had. Some sources state that this photo actually depicted the beginning of Galilee’s keel construction, probably in 1890, at the Turner Shipyard in Benicia, California.

 

A photo of a keel being fabricated in place at the Turner shipyard in Benicia, California. This photo can be found at various places on the Web.

Some sites listing the photo attribute the keel to Galilee, while others claim its provenance is unknown. (c. 1890, courtesy of Museum of History, Benicia, California)

I obtained the cross-sectional dimensions of the keel from a G.C. Berger body plan of the ship. The width of the lower keel itself was 14 inches. For the model's elevation reference, I chose the level of the inner rabbet (the lower frame/keel intersection) at the mid-body point as the zero-height reference. The keel height below the baseline was about twice that width, or 28 inches. The plans do not show a false keel or shoe, so I assumed the plan’s keel dimensions included this item.

Excerpt of the G.C. Berger plans showing the typical hull structures at the mid-body location of the hull. (Source: SFMNHP; No date.)

Keelsons

The plans also show a keelson resting on top of the floor framing with cross-sectional dimensions of about 31 inches high by 14 inches wide. Since the individual keelson timbers were typically and approximately square in cross-section, it is likely there were two keelson timbers in this hull, the main keelson and a rider keelson on top. I assumed this structure ran the entire length of the keel from the bow to stern deadwood.

One question I have had for a long time was whether the spaces between frame floors enclosed by the keel and keelson were left open or filled in. A drawing from The Elements (Fig. 85) shows fillers between the frame floors. The book calls them “chocks.” That particular diagram actually illustrates two separate chocks to fill this space—a lower horizontal one that lies on the top of the keel and a vertical one that spans the gap between adjacent frame floors (horizontally) and between the lower chock and the keelson above (vertically). However, Fig. 84 from the same book shows no chocks or other structures between frame floors on top of the main keel. It didn’t make sense to me to leave these big gaps between frames, so, for simplicity sake, I modeled these spaces as solid and continuous with the lower keel. (Comments?)

Extract of Figure 83 from "The Elements" (1911) reference showing the keel and associated structures most similar to those in Galilee (p. 82).

Note the two types of chocks at the midline between frames.

The G.C. Berger plans also show substantial sister keelsons to both sides of keelson. I assume that these timbers ran horizontally the entire length of keel adjacent to the main keelson, while resting on the hull frame floors. Where the frames angled upward at the extremities of the hull, the sister keelsons were tapered to about 6 inches in depth according to the reference. The typical sister keelson construction was shown in Fig.86 of the referenced book.

Bow and Stern Deadwood

I don’t have any good reference for the actual arrangement of the wood components comprising the deadwood in Galilee, so I made my best estimate for their overall structure based on photos of the bow and stern remnants still in existence in the San Francisco Bay area. The following images show the current state of the Blender model's backbone structures:

Orthogonal view of my reconstruction of Galilee's keel, rabbets, and related structures. The square-ish piercings show the locations of the full square frames. (Blender)

Perspective view of Galilee's backbone structures. Keelsons, rabbets, and bearding lines are clearly seen in this view. (Blender)

As a matter of interest, the Department of Terrestrial Magnetism (DTM) field reports noted that the ship had a 4–5 inch hog in the keel at the time of the ship’s charter. I chose not to model this feature because that would have greatly complicated modeling the rest of the vessel.

At the bow, the keel rounds up into the stem, forming a classic “rocker” bow. This design promoted better speed, since the moving water can flow smoothly along the keel. The absence of the slightly-extra buoyancy provided by the more familiar “clipper" bow was offset by a slightly broader bow. The stem itself was tapered in the plan view—narrower toward the leading edge than at the rabbet line.

In Galilee, the stem was topped with a beak structure in lieu of a figurehead or scroll. The simple beak, shown in both the plans and in photos, would have been typical for a Pacific packet ship at this time in history.

At the stern, the keel supported the lower bearing of the rudder as well as the extensive deadwood there.

The sternpost and the aft deadwood both supported the aft half-frames and the rudder box—to be shown in later posts.

Full and Half Frames

In the normal sequence of erecting the hull, the full frames were usually installed next. In the case of Galilee, the full frames comprised approximately the middle half of the hull. The type of wood for the frames was identified as "live oak" in a scaled DTM drawing of the dimensioned body plan.

It was my decision to develop the hull's framing and internal structural details only to the extent necessary to ensure that any of these items that might interact with the framing (e.g., deck furniture, masts, etc.), as well as the components of the frames that would be visible topside, would work to accurately support the exterior views of the model. This was particularly true of the open bulwark stanchions that were the exposed parts of the hull frame top timbers.

My research log documents that, over a period of many months, I agonized over the dimensions of the frames, where the DTM drawings described these as something like 10 inches by 12 inches in one place, but in at least one of their drawings, the depth of the mid-ship frames were 14 inches deep at the centerline. Also, the DTM deadflat drawings stated: "frames spaced 28"," which roughly agreed with spacing of the exposed bulwark stanchions in the Berger plans, if measured between corresponding frame faces. Sadly, the Berger plans didn't provide a consistent-enough spacing of frames (varying by up to more than an inch in places), nor the sided dimensions of the frames. Only after repeated measurements on the drawings, plus extracting laser scan dimensions taken by the NPS engineer from the bow and stern remnants in 2011, was I able to say with some confidence that the typical room-and-space of at least the mid- and after-body frames were 16 and 28 inches, which seemed to agree with the DTM plan notations. Interestingly, the room-and-space of the bow frames forward of the forecastle break changed to 14 and 28 inches, as documented by both my direct measurements in 2001 and the NPS scans.

Another issue I had with developing the hull frames was identifying which of the frame top-timbers formed the exposed bulwark stanchions. While my research notes indicated I had determined as far back as 2015 that the aft member of the doubled hull frames in way of the open bulwarks formed the visible 8-by-8-inch stanchions, I nevertheless went through the whole process of modeling the hull frames in DELFTShip using the forward member of those frames. I have no idea why I did that! That error initially shifted all the frames 8 inches out of position in the digital model, which resulted in a significant rebuild work many months later.

The initial DELFTShip model of the hull framing structures and deck beams, where I had mistakenly positioned the double hull frames,

so that the forward top timbers formed the bulwark stanchions. This mistake caused me several weeks of rework months later when I recognized the error!

The process of developing the frames began by referencing the moulded (inner) surface of the hull (see next image), which, as previous posts noted, was done in DELFTShip. Then the midship frame positions were lined off on this hull surface using the Intersection tool provided by the software to define the forward, middle, and aft surfaces of each frame as projected onto the inner hull. This was accomplished by measuring the longitudinal coordinate of each frame face in CorelDraw, and entering that data in a spreadsheet for each frame. Then I entered that data as Station Intersections for each frame in DELFTShip. Those intersections then appeared as lines projected onto the moulded hull surface depicting the surfaces of the adjacent frames. Next, these intersection lines were viewed in the forward or aft view, as appropriate, in the Export Plans feature, where I exported the view as an image file and imported it into CorelDraw in order to develop the actual transverse structure of the frame as depicted in the Berger body plan. (I can discuss this process in more detail in a PM for anyone that is interested.) Recall that all of Galilee's frames were square, so, for this vessel's case, this process was used for all 57 frames. It was tedious, and finicky, but worked, for the most part. For ships with cant frames, a different method would need to be used.

This image shows how I used the moulded inner surface of the hull to help define the shapes of the hull frames to which they conform.

The blotchiness represents differences in position measured in a few hundredths of an inch. (Blender)

Using a cross-sectional DTM drawing as a basis for establishing frame shape and dimensions (see next image), I basically retained the moulded dimensions of the frames at each corresponding height above the baseline throughout the length of the hull. This resulted in the cross-sectional dimensions of the frames remaining more or less consistent at a given height above the keel. While this may not reflect actual practice, it seemed to produce reasonable results.

The marked-up DTM drawing I used to determine hull-frame moulded dimensions. Drafted dimensions were original to the drawings.

(c. 1905; Courtesy of Carnegie Science Library)

The full frames were constructed to the centerline of the hull, inserted into the keel structure, and digitally mirrored. The half-frames had to be mated to the deadwood and other structures inside the hull. So each frame presented a unique case. Not only that, but as the frames approached the extremities of the ship, the top timbers, in particular, became skewed to conform to the run of the main rail (see the images following). This effect presented a problem, because, as the skewing became more extreme, the effective cross-sectional width perpendicular to the frame face became narrower. This tendency reduced the support area underneath the main rail toward the ends of the hull. Consequently, I had to adjust the moulded depth of the top timbers of these frames in order to retain the same effective width of the frame under the rail. Just another thing to keep track of.

Images showing the cross-sectional shapes of the hull frames at the level of the main rail. Mid-ship (top) versus near the bow (bottom). {Blender)

Another issue I had to deal with was the extreme curves that the aftermost hull frames had to conform to. Galilee had a shapely stern with significant overhangs at the quarters, so the half frames had to create this surface. The last two hull frames were partially or completely supported by the rudder box.

A hull-frames-only view of the aft end of the hull. (Blender)

The following image shows all the main hull frames in place from both end viewpoints.

Views of the hull showing the arrangement of the square hull frames. Hawse timbers and transom framing are—um—even more complicated!

At this point in actual construction, there are numerous support timbers, ribbands, and harpins (not shown) to keep everything in place

until planking can begin. (Blender)

According to The Elements reference, after the full frames were erected, then the keelsons and sister keelsons were installed. The half-frames at both ends were then "dubbed in" to seat properly against the keel and deadwood. Not sure how that would have worked for Galilee, but the next image shows how these items relate to the frames:

Keelsons and sister keelsons laid in place after the framing is erected and braced on the ways. (Blender)

The last major hull framing I will address before moving on to deck beams and related structures involves the hawse timbers and transom, which will be the topics of the next posts. As noted in the cross-sectional DTM drawing shown earlier in this post, the interior of the hull was lined with ceiling planks, riders, bilge logs, etc., which all strengthen the hull. Since these details don't affect the exterior appearance of the vessel, I chose to omit them as unnecessary, and which would simply have added to the complexity and digital size of the model.

To be continued . . .

The Hawse Timbers

Finishing off the framing at both ends of the ship would normally be the next priority after erecting the whole- and half-frames.

The hawse timbers filled in the bow supporting structure between the forward-most half frames and the stem. I suspect that, due to the short distances involved, coupled with the stresses imposed on the bow structure from the ground tackle, it was necessary to construct a basically solid foundation for this area of the hull.

For Galilee, most (if not all) of her original hawse timbers survive in the bow remnant on display at the Museum of History–Benicia (California). When I visited the museum back in 2001, the bow was under a shed and cross-braced to prevent its collapse due to its advanced deteriorated condition. Most of the internal hull structures appeared to be original except for an oversized hawse pipe casting on the port side, which was probably installed after her conversion to a fishing schooner in the 1910s. (This permitted open-ocean anchoring for the type of fishing the ship was employed in.) Also, the bowsprit bedding and fittings were probably not original due to her revised schooner rig.

Port view of Galilee's bow remnant at the Museum of History Benicia (MoHB). (2001, personal photo)

Stbd view of bow showing the original hawse pipe and its location relative to the hawse timbers. (2011, Michael Harrison, NPS)

The three hawse timbers on each side of the bow had about 9–10 inches sided measures based on photogrammetric ratios. I estimate that their molded depths were similar to the foremost hull frame's but may have been reduced toward the stem. They were set at an angle to the adjacent hull frame, and more-or-less paralleled the stem. They mostly filled the space between these structures in order to support the shape of the bow in that area.

When digitally modeling these timbers back in 2021, a variety of significant issues cropped up. With the positions of the hull frames as originally modeled, and assuming the forward top timbers of the hull frames represented the open bulwark stanchions, nothing fit in the hawse timber area of the model with reference to these MoHB photos. After weeks of agonizing over alternative interpretations of these structures, and even reassessing the length of the ship in the Berger plans, I finally realized the following:

• The hull was about 1-foot too long based on the working plans. (In a side rabbit trail, I had discovered early on in the 2000s that the construction rules published in the Record of American and Foreign Shipping, 1891—the predecessor to the American Bureau of Shipping, or ABS—made reference to longitudinal positions inside the hull that were not visible in the plans—nor from outside the actual vessel.)

• The exposed hull frame top timbers in the open bulwarks were actually the aft top timbers, not the forward ones.

As a consequence, all 57 of the hull frames had to be shifted longitudinally and their shapes revised in order to accommodate these revisions. (Their locations in the previous post show their final configuration.)

Another difficulty in working up the details of the hawse timbers was correctly locating the future hawse pipes. The Berger plans do not show the hawse pipes. So this information had to be obtained from photos of the hull. No one DTM nor SFMNHP photo of the ship shows a truly orthogonal view of the bow. So the process to obtain this information was one of iteration among many sources. Eventually, with the aid of a number of more current photos of the bow remnant taken by staff from the MoHB, I was able to make a reasonable estimate of the hawse pipe locations.

The following images show the installation of the hawse timbers in the most recent version of the hull framing.

Two views of Galilee's hawse timbers. In reality, the hawse holes would probably have been cut in farther along in construction,

once the deck beams, waterways, and related timbers were erected. (Blender)

Next up: Stern/transom framing . . .

Great to see your research and 3D efforts here on DDM Terry, you have certainly put a lot of time and effort into this vessel.

cheers

Pat

The Stern/Transom Framing

As I have previously discussed, modeling the shape of the ship's transom at the time of interest (1908) was subject to some interpretation. The original transom during the DTM charter period (1905–1908) was quite curvy at its lateral points. None of the extant plans of the hull in any way match the photos from this period. To complicate matters, photos of the stern after the ship was retired in the 1930s also do not match the DTM photos. In posts discussing these issues with MSW members several years ago, a number of us concluded that the hull and transom may have been modified. This could have especially been the case after the ship was altered to employ a centerline propellor with its diesel engine, which has been reported in the historical literature of the ship.

Photos of the Galilee's stern in 1906 and much later sometime in the 1930s. The sides of the transom are distinctly different.

(Photos courtesy of Carnegie Science Library [left/top] and SFMNHP Archives [right/bottom])

A modified stern/transom may have also explained some of the unusual structures photographed inside the stern. As noted earlier, NPS engineer Michael Harrison was given the privilege in 2011 of thoroughly documenting the stern remnant of the Galilee at its Fort Mason, San Francisco site. In the process, he was able to obtain a number of photos from within the stern, which greatly aided the digital reconstruction of this end of the ship. Following are a number of photos pertinent to this effort:

View of the poop deck of Galilee's stern remnant, under an added canopy, looking aft and to port.

Lazarette access to the right. Two poop-deck bitt timbers center and left. (Photo 2011, Michael Harrison, NPS)

View from the forward end of the lazarette of the stern remnant looking aft toward the two portholes in the transom.

The transom framing and overhead aft poop deck beams can be seen. The rudder box is to the right. (Photo 2011; Michael Harrison, NPS)

View looking to starboard and outboard in the lazarette, showing the waterway and interior hull sheathing.

Poop deck and beams up and to the right. (Photo 2011; Michael Harrison, NPS)

View in the hold below the lazarette decking, looking to starboard and aft. The lower rudder box is in the center. Lower portion of the aft starboard bitt in back.

Main deck and deck beams are above. An unusual built-up transverse bulkhead to the right lies aft of the rudder box and is attached to it.

An original hold main deck beam stanchion is to the left. (Photo 2011, Michael Harrison, NPS)

From these photos, I was able to obtain a good idea of how the stern internal structures were arranged. Even so, I had to guess at the other essential structures, not otherwise shown, required to build up the framing of the stern by referring to contemporary books and illustrations addressing similar hull structures.

The key structure around which the entire stern was built was the rudder box. The rudder box itself was apparently a . . . um . . . rectangular box built up of heavy side timbers, creating a solid structure and sandwiched between two heavy flat timbers, front and back, bolted into shape. From what I could surmise, the box was constructed continuous with, and mounted on top of, the stern post. The transom and hull planking were rabbeted into the completed structure.

The combined keel and rudder post with the rudder box showing the associated rabbets and rudder stock fairings. (Blender)

During construction, the rudder box was hollowed out in a quasi-conical shape, elongated in the longitudinal direction so that, when the rudder had to be unshipped, the rudder stock could tilt forward in order to clear the lower bearing as it was extracted:

Looking down into the rudder box to observe the tapered space within the box to accommodate shipping the rudder stock. (Blender)

Once the rudder box was installed, then, in Galilee's case, the last two hull frames were attached to it and the underlying sternpost and deadwood. Frame 57 was totally supported by the rudder box alone.

A view showing how the last two doubled hull frames were attached to the rudder box. (Blender)

The transom framing was attached to and supported by Frame 57. First, let's return to that built-up bulkhead mentioned earlier. It appears to have spanned the hold space aft of the rudder box and even provided support for it. The photos indicate that there was a limber hole/slot at the centerline. I am suspicious that this structure was not originally part of the hull but may have been added on, if and when the stern was modified.

According to recent photos, a transverse, built-up bulkhead spanned the last frame and was attached to the aft side of the rudder box. (Blender)

Next, fashion pieces on each side were attached to the aft face of frame 57 to help shape the lateral edges of the stern and to support the upper transverse framing of the transom. These timbers were clearly visible in the Fort Mason photos. They look in bad shape or were poorly fabricated, so I wonder if they, too, were not original.

Short fashion pieces were attached to the last hull frame. They formed the outboard underlayment for the hull planking at the lateral extremities of the transom.

The image of the reconstructed fashion pieces in the model shows the entire fashion piece structure in combination with frame 57. (Blender)

While not directly related to the transom, I needed to include the poop deck beams because the aft-most poop-deck beam supported the upper shelf of the transom. The lower shelf was supported at its ends by the lower fashion pieces. The notched shelves supported the upright transom frames.

View showing the notched upper and lower transom frame shelves. The upper shelf was attached to the aft-most poop deck beam. (Blender)

Next, the seven upright transom frames were installed. I determined the total number of these by counting the vertical rows of fasteners in present-day photos of the outer transom planking. None of the interior photos taken by Harrison showed all the transom frames inside the lazarette. (It is possible that there might have been another pair of outboard frames, but the outside fastener pattern didn't support this idea.)

Reconstruction of the seven transom frames which supported the transom planking. (Blender)

And, lastly, I surmised that there needed to be blocking or chocks between the lower ends of the transom frames in order to support the shorter, lower transom planks.

Completed reconstruction of the stern/transom framing. (Blender)

Next up will be establishing the structures for supporting the main deck beams and the beams themselves.

And I believed wooded ship modeling is buying a kit and researching the best wood glue.

Respect.

An interesting topic – one to keep an eye on.

Main Deck Beams and Related Structures

The next typical hull structure to be installed were the deck beams. In a single-deck vessel like Galilee, the main deck was the only continuous, longitudinal deck. Construction of multideck vessels normally began with the lower decks because the upper decks were dependent on the lower decks for support.

According to both the Berger and DTM renditions of the body planking plans, there was a beam shelf that evidently ran the entire length of the hull. The drawings suggest a timber 6 inch thick and 14 inches high, lying on a 4-inch ceiling layer. Since I have omitted the ceiling layer, the digitally modeled shelf is 14 inches high by 10 inches thick for the majority of the shelf run. Toward the stern, as the hull frames sharply angled in toward the centerline, the shelf had to twist somewhat, which is not documented in any of my reference plans or photos, so I did the best I could to maintain the overall support surface. The shelf may have terminated at the built-up transverse bulkhead (if the bulkhead was original) or terminated in the lower transom shelf.

The representative main deck beam shelf, which incorporates the hull ceiling thickness along its length.

The beam shelf was normally fastened on top of the ceiling. (Blender)

This image illustrates how the shelf had to twist near the stern. (Blender)

Once the shelf was in place, the deck beams were shaped according to the lofting measures and installed per plan. For the digital model, I used the beam locations as indicated in the G.C. Berger plans, since they properly corresponded to original deck structures and openings. In the "other" forum, there was a lot of discussion regarding the camber of deck beams. Many members suggested that the camber varied along the length of the deck. From my research, I concluded that the deck camber was constant along both the length of a deck and among decks. If nothing else, it simplified the construction and modeling of the decks. This permitted the deck sheer to be consistent across the width of the hull as well, which helped to determine the locations of the outboard ends of the deck beams as a smooth curve. The outboard deck line, then, was simply the intersection of the modeled deck beam surface with the inboard surfaces of the hull frames.

Main deck beams and framing for hatches and partners. (Blender)

According to a description of deck framing provided by Crothers in his book The American-Built Clipper Ships, deck beams in way of hatches

and partners were typically 10 percent deeper, so in my model I let them into the beam shelves (arrows). Whether the beams or shelves were notched is not known.

As with the hanging knees, I omitted the bosom and other horizontal knees between beams. (Blender)

Many reference sources mention that after the ends of deck beams were fastened, they were sprung to their full camber by jacks, then stanchions were installed to keep them in place. Finally, the hanging knees were installed to hold their ends in place. That method, apparently was not the case with Galilee's hull. Photos taken during the 1960s when her hull was still relatively intact in the Sausalito mud showed that, not only did she not have knees at every deck beam, but the beams themselves show no evidence of having been sprung into a camber. Evidently, Captain Turner constructed the hull with hanging knees every other deck beam. So I modeled the lower surface of the main deck beams as horizontal, since there was no need to anchor the ends of the sprung beams. The following photos show unidentified men evaluating the potential for salvaging the Galilee's hull for posterity. In the end, only her stern and bow were salvaged. The hull was too riddled with shipworms to recover it.

View of Galilee's hold showing deck beams, beam shelf, ceiling, and hanging knees. The very bottom of the hull was basically consumed by shipworms.

(SFMNHP, 1960s)

Another view of the ship's hold. A modern steel beam jack supports a main deck beam. The pad under which it is installed was an original stanchion pad.

(SFMNHP, 1960s)

Wooden main deck stanchions supporting the beams, looking forward. (Blender. Clipped program views do not support renders)

Next, installation of the poop deck framing and beams.

Poop Deck Construction

According to some of my references along the way, once the hull frames and main deck beams were in place, the yard would typically apply the exterior hull planking. However, since the hull planking needed to be "lined in" with reference to the main rail and any other structures that dictated the run of the main rail, I believe it was needful that those structures probably should have been erected/installed before attempting the outside planking of the hull.

Consequently, I will address the construction of the poop deck (and then, later, the forecastle deck), as well as their structural members, before moving on to the hull planking.

However, before any of this can be considered, several other structural elements need to be added . . .

Immediately after the deck beams were in place, the yard had to provide a walking surface for subsequent steps in construction, so it seems logical that the the main deck should be put in place next. However, the outboard edges of the main deck were defined by the waterways, which were major longitudinal timbers essential to stiffening the hull. Galilee had unusually prominent waterways emplaced directly on top of the deck beams and adjacent to the hull frames. Based on photogrammetric dimensions, the waterway cross-sectional dimensions were about 11 inches high above the deck beam and 14 inches wide, inboard of the hull frames. Wherever the waterway was exposed to human traffic, the upper inboard corner was rounded for safety. Under the elevated decks, the photo farther along in this post shows that the waterways were squared off where not accessible to passengers.

This is a screenshot of the model showing the run of the waterway. In order to seal the spaces between the frame stanchions in the open bulwark, waterway fillers (or "chocks") were installed to prevent water from flowing into the frame bays inside the hull.

View showing the heavy waterway installed on top of the main deck beams and inboard of the hull frames. Filler pieces seal the space between frames. (Blender)

Toward the ends of the hull, where the bulkheads were closed, there were no waterway fillers between the frames.

In Galilee, there was blocking added to the waterway as reinforcement for the hawse pipe. (Blender)

Once the waterways were in place, it seemed reasonable that the main deck was laid down, if for no other reason than the deck provided a walking surface for erecting the structures pertaining to the poop and forecastle decks, as well as all the other inboard structures. From several sources in my documentation, the main deck was 4 inches thick. The average width of the deck planks was about 4¼ inches, which I figured out after many weeks of agonizing analysis that resulted in modifications of other major parts of the deck furniture . . . Thankfully, the Harrison photos confirmed the thickness of the deck planks. So, here is a view of the modeled main deck planking, including the joggling along the outboard edges. The planking is worked right up to the waterways, which is why they had to be installed first.

This view shows the main deck when Galilee was built, including the two holes for the cargo hatches. For the DTM charter period, these were decked over. There was another access aft to the hold through the main deck via the lazarette, but I have no information on that. (Blender)

At this point, I can begin the construction of the poop deck. According to latter-day photos taken of the lazarette interior by Harrison back in 2011, the poop deck beams were supported at their ends by heavy sheathing timbers fastened to the inside surfaces of the hull frames. These timbers served as poop deck beam clamps as well as a form of ceiling within the lazarette. In addition, the sheathing was stopped about 6 inches short under the main rail to provide an air gap for ventilation of the frame spaces. These details can be seen in this photo of the Galilee's stern remnant on display at Fort Mason in San Francisco.

A closeup of the ship's lazarette space looking outboard and to starboard, showing the heavy sheathing planks that supported the poop deck beams. The two-by lumber is recent to support the awning over the stern.

Note the air gap between the sheathing and the main rail in the upper left corner of the photo. The waterway and main deck planks are below.

(Photo: NPS/Harrison, 2011)

Here is the model showing the side sheathing and deck beam clamp for the poop deck. This sheathing actually was carried forward to the aft-most extent of the open bulwark.

This view shows the extent of the poop deck beam shelf that turned into a closed bulwark farther forward. (Blender)

Next, it is likely that the full-width poop deck beams were laid in, probably secured by notching the sheathing.

The framing of the poop deck is complete. The forward bulkheads enclose the lazarette and the upper sills were continuous with the main rail. (Blender)

Farther forward, the poop deck beams to the side of the (future) main cabin had supports and inboard clamps. I discovered how this was done by examining photos of the refit of the C.A. Thayer, a West-Coast contemporary of Galilee. It makes perfect sense. The vertical stanchions were supported by the underlying main deck beams.

Photo showing the construction of the side poop deck framing in the lumber schooner C.A. Thayer, during her most recent refit. (SFMNHP)

Another major feature of the poop deck structure was the four heavy-duty mooring bitts, two on each side of the poop deck. These, evidently, were secured to the hull frames at their lower ends and their upper ends were captured with carlings and blocking/chocks between the poop deck beams. These structures also supported the ends of the deck planks in way of the bitts.

Here is a photo of the bitts as they appear in the stern remnant.

The two starboard mooring bitts on Galilee's poop deck—much worse for wear.

The bitt norman pins were long gone, probably taken for their scrap metal, before the stern was salvaged. (Photo 2011, Harrison)

And here is how the wooden parts of the bitts are modeled. No metal parts at this point in the project.

Arrangement of the four mooring bitts and their supporting framing at the stern of the ship.

A clipped view of the model showing how the aft mooring bitts were secured to the hull frames and main deck beams. (Blender)

Just as I was going to add the poop decking itself, it occurred to me that the main rail significantly constrains the edges of the poop (and forecastle) decks. So, I'm going to include the main rail reconstruction without going into a lot of detail at this point—simply so I can show the poop deck weather surface. Not only that, but, as I mentioned in the beginning of this post, lining out the exterior hull for planking required that the main rail be constructed as well to provide a reference for spiling. So this is probably a good time to include the rail, which forms another significant longitudinal structural member.

Overall view of the entire poop deck, the mooring bitts, and the main rail. (Blender)

And that pretty much describes the construction of the poop decking and related structural members. Later, when the main cabin is constructed, there will be some additional moldings and details related to the poop deck, not to mention all the deck furniture, railings, and other hardware located here.

Next up: the forecastle deck . . .

Constructing the Forecastle Deck

In Galilee, the forecastle deck was the only other deck above the main deck besides the poop deck, being a relatively small brigantine. As was typical for this type of vessel, the forecastle deck provided an elevated working surface for essential shiphandling evolutions, such as anchor handling, being towed, and managing the foresails. The space below the forecastle deck was reserved for the windlass, securing the anchor chains, and probably providing storage for seaman gear, such as the capstan bars and miscellaneous equipment and materials.

The elevated forecastle deck was supported by less substantial deck beams compared to the main deck beams. Because of the congestion of the numerous items of deck furniture required to penetrate the deck here, the deck beams had to be closer together, and thus their dimensions could be lighter. The beam cross sectional dimensions were about 7 inches by 6 inches and they were placed a little more than two feet apart, compared to the main deck beams, which were about four feet apart on centers. Interestingly, these beams appear to have been milled with a camber on both the lower and upper surfaces, based on current-day photos of the bow remnant I took of the bow remnant in Benicia, California, back in 2001. Also based on the same photos, the deck beams were let into the hull sheathing lining the bulwarks below the forecastle deck and above the waterways. The planks that served as the beam clamps were somewhat thicker than the rest of the sheathing in this area.

The forecastle bulwarks were lined with thick planks to weatherproof the frame bays in this part of the ship.

The planks that served as the beam clamps were nearly twice as thick as the bulwark sheathing. (Blender)

Due to the strain placed on the deck structure by the samson post and the capstan/windlass, additional bracing was provided by carlines between the relevant deck beams. Also, there were a pair of substantial vertical timbers connecting the forecastle deck with the main deck, to which the windlass was secured. As with the poop deck, the forecastle deck was delimited by both the main railing and a substantial transverse timber that connected the port and starboard main rails. This timber was attractively milled on its aft face to match the beaded contours of the inboard main rails. It also probably included a shelf to which the aft ends of the deck planks were secured.

Forecastle deck beams and related framing. The carlines will brace the samson post and windlass gear. The latter required substantial bracing

knees, which are secured to the main deck beams underneath. The forecastle deck aft sill beam was also braced with knees.

As with the poop deck, the main railing was included in this view in order to help define the extent of the forecastle deck. (Blender)

As part of the bracing for the samson post, there was laid down during the decking phase a thicker centerline plank braced against the inner surface of the stem in order to provide longitudinal support for the samson post, which would probably have been constructed during this phase, since it penetrates the main deck.

The centerline "king plank" bracing for the samson post. (Blender)

With these preliminary structures in place, the forecastle decking could be installed so as to provide a working surface for installing the the upper components of the hull, and in preparation for lining off and planking.

The forecastle decking installed. There were probably chocks between the deck beams along the main rails in order to

secure the ends of the deck planks.

I probably should include them for completeness.🤔(Blender)

Next up, more discussion of the main rail.

The Main Rail

For most wooden ships (at least in the later 1800s), the main railing that caps the lower bulwarks performs several functions. It usually is one of the main longitudinal structural elements that ties together the tops of the hull frames. As such, it was usually a fairly substantial piece of wood. The specs called for the longest pieces available using a minimum of scarfs. Aesthetically, the main rail also was one of the main visual elements that defined the sweep of the sheer line, drawing the eye along the ship's hull in a smooth, sweet curve. All the outboard visual elements of the ship were visually tied to sheer line of the rail. In Galilee, the main rail was 5⅜ inches thick by direct measure.

I measured the main rail attached to Galilee's bow remnant at Fort Mason in San Francisco back in 2001. (Personal photo)

However, it took some work for me to determine the width of the ship's main rail, because, in the various plans I had to refer to, the rail width varied between about 8 inches and 12 inches and even appeared to taper in the drawings. After studying the photos to get a feel for the shape of the rail, I finally settled on a constant overall width of 11½ inches.

Another visual aspect of a ship's rail was the practice of adding one or more rounded ridges to the edges of the rail, which were usually called beading. Galilee's main rail had two beads on the outward side and a single bead on the upper half of the inboard edge. The upper outer bead was 3 inches high and the lower was then 2⅜ inches. I had initially thought that the rail timbers must have been milled with these two beads along the entire length (before scarfing). But a photo I took of the bow shows that, apparently, the two beads were milled separately. The upper bead seemed to have been let into a rabbet into the main part of the railing above the lower bead. At least that is what it seems from the photo.

Photo showing the cross section of the ship's main rail. Either the upper bead was milled separately from the main part of the rail,

or there is a very advantageous crack here separating the upper bead from the rest of the rail!

The ship's main rail sweeps around the entire hull uninterrupted except at the bow, where both ends terminate in the bowsprit. That doesn't mean there aren't some modifications along the way! First, in way of the chainplates that anchor the mast shrouds and stays, the main rail is horizontally expanded to anchor the chain plates within the width of the rail. The metal straps that anchor the lower deadeyes pass through the rail at these points. The rail beading wrapped around this expansion in the rail.

A preview of the rigging chainplates showing how the main rail was modified to secure the chainplates.

The other modification occurred at the stern. Because the ship's mainsail was fore-and-aft rigged, the main boom had a sheet horse mechanism, which was designed to snub the shock of a jibe or swinging boom being stopped by the sheet blocks. The horse had to be anchored firmly to the deck or the whole sheet rig could be carried away. So the main rail was substantially broadened at the stern to securely mount the sheet horse. In fact, the main rail was fastened to two deck beams at this location.

Another preview showing how the main boom sheet horse was mounted on the main railing at the stern.

The horse design was typical to the US West Coast at this time.

So, this post explains the shape and function of the main rail in Galilee. In my opinion, it and the cabin roof moldings were the two most attractive features of the ship!

Next: Planking the hull . . .

Your Blender skills are impressive!

Planking the Hull

I'm not going to go into how the an actual hull was lined off for planking because I really don't know that much about the actual, real-world process. The only references I have for the real-world hull planking are from the contemporary photos taken mainly by DTM photographers, one of whom was my grandfather when he was assigned to the expedition crews. Thankfully, those reference photos provided a world of information of how the ship was planked! The real trick was converting that information into a 3D plan.

My primary resources were provided by the following photos, as well as reference to other photos in my collection for details:

Drydock photo of the ship, probably in early 1906. The irregular edge to the boot top paint job suggests that the ship wasn't coppered, at least at this point in time.

Photos of other Turner brigantines in drydock show that their hulls were coppered, and the stern remnant shows scraps of copper sheathing still attached.

(Photo: Courtesy of Carnegie Science Library, c. 1906)

A photo of the ship alongside a pier after the observation bridge was lengthened. (Photo: Courtesy of Carnegie Science Library, c. 1907)

Another reference photo of the Galilee's stern well after retirement, Sausalito Bay, c. mid-1930s. (Source: SFMNHP Library)

One of the more difficult questions to answer, surprisingly, was the physical location of the actual painted waterline of the ship while in service with DTM, from which I could estimate the base widths of the hull planks. I ended up having to use multiple dimensional references from other parts of the actual ship that I was fairly confident were accurately reproduced in the model to provide a guide to where the waterline was located in the photos. Having determined that, then it was fairly straight forward to determine the average width of the hull planks at the middle of the ship, which probably should have been fairly consistent. Based on what little I know about merchant ship planking methods, hull planks tend to taper toward one or both ends of the hull, so the midpoint is where they would be widest. For Galilee, her hull planks below the bulwark were about 7-7/8 inches wide and 3 inches thick (according to DTM body plan drawings). Using a body plan of the hull in CorelDraw, I was able (eventually) to establish that the waterline in the photo was 123 inches above the model baseline.

Another issue that had to be resolved were the number of rows of hull planks. My research journal revealed a long iterative process early last year trying to figure out how many planks were required to span the distance from the bottom of the bulwark to the garboard strake. That number turned out to be 38–39. I won't go into the details about how that was determined. Interestingly, the top edge of the top 3-inch thick hull plank is beveled at nearly a 45 degree down-angle to promote water drainage off the top of the plank. The bulwark planks above the top hull plank were only 2 inches thick.

And before I could actually start modeling the planks, I had to deal with the garboard plank, which was a heavy timber about 6-3/4 inches thick and about 14 inches wide at the midbody point. In addition, it tapered toward the bow and stern to the hull planking thickness of 3 inches. Such fun!

The bulwark height was more or less constant the full length of the ship, and the planking there was finished smooth, so I was never able to determine the plank widths in the bulwark itself. So I treated the bulwark as a single surface in the model, with no visibly calked edges.

In order to take into account the taper of the planks at each end, normally, this was done through the process of spiling at each successive station, which is apparently a very laborious process for anyone who has done it. (I haven't.) Blender has a function that can equally subdivide an Edge (curved or straight between two endpoints) into equal-length segments of any number desired. So, theoretically, if you place a station line curve along the surface of hull in the program, then tell it to subdivide that curve into a given number of planks, say 38, it will place the vertices equidistant apart along that curve to give you 38 segments, which is effectively what spiling does to a station line drawn on the hull from rail to keel. The trick was to take into account the garboard strake in the middle 2/3 of the hull. Some interpolation was necessary at those points. Also, where the bow and stern stations did not represent the full depth of the hull, some creativity was needed to designate the length of the line to be so subdivided. This isn't intended to be a tutorial, so I won't go into the details of how this was accomplished.

The only other issue with laying out the planks was that, apparently, there was need of a single stealer in the plank runs on both sides of the stern. I'm told that is not unusual for very curvy hulls, as is the case with Galilee. Thankfully, there was only one.

Once the plank runs were determined on the inner molded hull surface as modeled, I had to figure out where the plank seams were for the individual strakes, because single planks didn't run from bow to stern. That required a map, using a nominal plank length and the planking rules that established the minimum distance between plank ends on any given hull frame. The following diagram was the result. Who knows what the actual planking pattern was?

My planking map for the digital model. The horizontal scale was the frame number, starting at the stern. The horizontal scale is the frame numbers.

The ends of adjacent planks could be no closer than three frames. This detail was probably overkill!

Once this pattern was set, some tweaking was done as needed, particularly around the garboard strake, as well as up under the stern out to the rudder port. That is where the stealer plank occurred. Again, there was some interpolation required out by the stern post and the planked hull surface is still not as smooth as I like.

I won't go into the process for actually creating the 3D planks. It's quite laborious and not really intuitive. I thank Nate (3Dshipwright), with MSW, for guiding me through the process. And this is the result for the main hull. I especially like the subtle plank edges that simulate the caulking, and were visible in the photos of the ship.

Bow view of the freshly planked hull before painting.

View showing the one stealer required. The weird markings along the rudder post are due to overlap of the digital meshes there. Needs some tweaking.

Stern view showing how the light plays off the planks. The transom planking was 3 inches thick, so there had to be some creative molding work done between the transom and the 2-inch bulwark planking!

Next up: the upper railings and the rudder.

Upper Railings, Etc.

It makes sense in the course of building a ship to provide as much safety for the workers as is feasible by constructing those components of the vessel that would aid in keeping workers from falling off the hull early in the the process. At least those are my thoughts, having actually served on ships in drydock for as many times as I have done. So, the next logical phase in priority would seem to be to add the upper railing structures to the hull, at least while beginning the other major construction tasks.

The Fancy Rail

Forward, Galilee was equipped with what was called a "fancy rail" which ran along the outboard sections of the main rail in way of the forecastle deck. It not only added some height to the main rail, but it was flared in height toward the bow to help somewhat in keeping the forecastle a bit drier at sea. Its function also included securing the catheads for handling and stowing the anchors. And the railing also provided a means of tying together the upper hull structures above the main deck in the vicinity of the bow. The fancy rail also included a large knee at the very forward ends of the railings to further tie the hull structure together in the vicinity of the bowsprit.

This view shows the "fancy rail" installed on top of the main rail at the bow. The forward intersection of the fancy rail was reinforced by a massive horizontal knee structure.

This knee will eventually provide a solid base for forestay hardware and headsail sheet belaying. (Blender)

The two catheads were about 11 inches square at their outer ends. Their inboard ends were solidly secured to the underlying forecastle deck beams. Eventually, they would be fitted with sheaves and cleats for the anchor securing gear as well as release mechanisms for letting the anchors go when anchoring.

Monkey and Taffrails

At the other end of the ship, the poop deck was protected by a pair of railing structures. On top of the main rail, was one called the "monkey rail." I tried to find the origin of this term but was unsuccessful. Evidently, the first use of the term in a written work was in "Two Years Before the Mast," by Richard Henry Dana, in the 19th century. The term generally referred to a smaller rail secondary to the main railing on which it was mounted.

For Galilee, the monkey rail was made of solid timber about 6-1/4 inches high between the main rail and its double-beaded cap rail, and 4-1/4 inches wide. It wrapped its way around the stern and extended nearly as far forward as the aft extent of the open bulwarks. The monkey rail cap was double-beaded outboard, single-beaded inboard, and its cross-sectional dimensions were 3 inches high by slightly under 10 inches wide. There are fairleads through the rail in the vicinity of the mooring bitts, port and starboard. Furthermore, the monkey rail provided a base for the taffrail, to be discussed next.

Arrangement of the monkey rail on top of the main rail at the stern of the ship.

The forward ends of the rail were equipped with shoulder plates to keep rigging lines from fouling the rail

The monkey rail also formed the base for the taffrail balusters and related structures. (Blender)

The Taffrail

Above the monkey rail was the taffrail. Since the poop deck was continually manned at sea, it makes sense that the additional railing height would be warranted here for the safety of the crew. The taffrail was supported by 23 turned balusters, staked into the monkey rail, and the two end-boards. Again, the end-boards were rounded to avoid fouling the running gear lines.

.

A view of the taffrail installed on top of the monkey rail.

Photos indicate that the outboard balusters were slightly tilted inward to follow the slight tumblehome at the stern. (Blender)

The Rudder

According to my resource, about this time in the construction process, the rudder was often installed, depending on the launch schedule for the project. Galilee's rudder is not very well documented in the DTM photos. The actual profile outline agrees well with the G.C. Berger plans, but there are a lot of questions regarding how it was secured to the ship. There is no visual or drawing evidence of the standard gudgeon-and-pintle-type hardware normally used for mounting the rudder. However, there is ample evidence that the rudder was secured to the ship at its bottom only by a bearing assembly mounted on the lower rudder post and, evidently, at its head by the steering mechanism. The rudder profile also shows "coning," which is an enlargement of the rudder stock above the blade, which tapers down in size to the width of the rudder blade itself. Evidently, this detail reduced the tendency for rudder stocks to snap due to the asymmetrical forces exerted on the rudder at sea.

The rudder installed. Note the coning (enlargement) of the rudder stock as it enters the hull. (Blender)

Next up: Painting the hull prior to launching; continuing the outfitting