Jaager

Siggi,
 
I am going to face a similar construction - if I can re-gain the motivation.  You have the top inner support as a molded half circle. 
Once the whole is assembled, is there any reason  not to use a solid half circle instead?   I know that the original vessel would have been as you built it, but if it is hidden  no one would know but the builder.

I do not know.  I think they were for defecation - so with glass there would be protection from the wind - a welcome thing is some seasons - without glass there would be ventilation.  Maybe there was a solid cover on a hinge or hook to allow either option.   I do not see that the diet was sufficient in fiber, so perhaps the job done there required a bit of study?😉

Siggi,
 
I am going to face a similar construction - if I can re-gain the motivation.  You have the top inner support as a molded half circle. 
Once the whole is assembled, is there any reason  not to use a solid half circle instead?   I know that the original vessel would have been as you built it, but if it is hidden  no one would know but the builder.

I have a DustDeputy  in-line - so except for some really fine particles the filter is clear and bin is empty.  Besides, it started life with me having a short on cycle.   I figure that a sensor is faulty.  I used to be big on using IPL and ordering and replacing faulty and worn out parts.  Either living alone or getting beyond retirement age has greatly dampened my enthusiasm  or I guess it could be my black dog.   I am betting that replacement sensors and the trouble involved with taking things apart and reassembled would match the cost of the Rigid, now that I think on it.   It annoyed me that the instructions said not to use a cyclone trap.  There is no constricting aperture involved with the trap, so what possible effect could it have?   Trap or no trap,  remote control on/off or straight to 110V wall outlet - none of it affected the short cycle unreliable nature of the Festool - so I am more than a little displeased with it.

Some random thoughts on your project:
in 1799 three super frigates were built using the same plans as the starting point:  United States, President, Constitution.
The United States started life with a round house.  The first captain of each ship had some say about how their new command would be built.  It was a wild hare of an idea and the resulting poor sailing quality caused the addition to be removed.
A round house is a tall deck at the stern - in the 17th century it was were the poop deck would be but a round house is a lot longer.
 
The United States lasted until the US Navy's base here in Norfolk was captured by Confederate forces.
The ship would have likely undergone topside cosmetic changes about every 20 years as the then current fad changed.  I would not be surprised if United States did not gain an elliptical stern at some point.
Since it is a sister, the AOTS monograph for Constitution may help with masting and rigging and deck details if the era of your model is the same as the monograph's.
 
Plastic masts and yards at 1:96 might prove to be a bit bendy or fragile.  You might give a thought to replacing them with wood.  And also a thought to upgrading the kit supplied line,
 
@rwiederrich has some 1:96 clippers with exquisite masting and rigging - to give you something to shoot for.
USS  was not part of the official name for US Navy vessels until TR decreed it early in the 20th century.

Siggi,
 
I am going to face a similar construction - if I can re-gain the motivation.  You have the top inner support as a molded half circle. 
Once the whole is assembled, is there any reason  not to use a solid half circle instead?   I know that the original vessel would have been as you built it, but if it is hidden  no one would know but the builder.

I do not know.  I think they were for defecation - so with glass there would be protection from the wind - a welcome thing is some seasons - without glass there would be ventilation.  Maybe there was a solid cover on a hinge or hook to allow either option.   I do not see that the diet was sufficient in fiber, so perhaps the job done there required a bit of study?😉

I do not know.  I think they were for defecation - so with glass there would be protection from the wind - a welcome thing is some seasons - without glass there would be ventilation.  Maybe there was a solid cover on a hinge or hook to allow either option.   I do not see that the diet was sufficient in fiber, so perhaps the job done there required a bit of study?😉

I cannot keep the Festool running - the overheat switch or whatever keeps tripping - the Rigid that I bought said it was quiet.  It is quiet enough not to need ear muffs, but not quiet enough for Mozart to be heard.

Some random thoughts on your project:
in 1799 three super frigates were built using the same plans as the starting point:  United States, President, Constitution.
The United States started life with a round house.  The first captain of each ship had some say about how their new command would be built.  It was a wild hare of an idea and the resulting poor sailing quality caused the addition to be removed.
A round house is a tall deck at the stern - in the 17th century it was were the poop deck would be but a round house is a lot longer.
 
The United States lasted until the US Navy's base here in Norfolk was captured by Confederate forces.
The ship would have likely undergone topside cosmetic changes about every 20 years as the then current fad changed.  I would not be surprised if United States did not gain an elliptical stern at some point.
Since it is a sister, the AOTS monograph for Constitution may help with masting and rigging and deck details if the era of your model is the same as the monograph's.
 
Plastic masts and yards at 1:96 might prove to be a bit bendy or fragile.  You might give a thought to replacing them with wood.  And also a thought to upgrading the kit supplied line,
 
@rwiederrich has some 1:96 clippers with exquisite masting and rigging - to give you something to shoot for.
USS  was not part of the official name for US Navy vessels until TR decreed it early in the 20th century.

I used an old Craftsman 16 gal that could have substituted for a 747 jet engine for noise in a closed area.
I bought a Festool Midi when the Craftsman burned out.  It is quiet, not enough to allow music, but I could ditch the ear muffs.  
Two problems though - it says not to have a cyclone trap in-line - which I insist on - and trap or not, it turns itself off after 15 minutes at most.
Bandsaw ripping of framing stock takes longer than 15 minutes.  Doing the thickness sanding on the ripped stock takes much longer.  A hull of any size (with a reserve for mistakes) can need 60 pieces 2" x 2'  planks.  And that always takes me longer than I imagine it will take  before I start.
The Midi was ~$500 - I checked HomeDepot and bought a 14 gal Rigid for  < $100.  It is about as quiet as the Festool, likes the DustDeputy, and says on as long as I can endure a session.   A radio control On/Off - (budget models bad)  - is a help - it lives in a condo, and I have to unplug it when done because neighbor garage door remotes can turn it on. 
The Festool is $500 that I will never get back and is now a support for the 5 gal cyclone trap catch drum..
My ideal solution for a shop vac is to have it in a room where I ain't.  Alas, this is not an option in a condo.

Unless you own stock in a shop vacuum filter company, if you plan to hook your vacuum up to a major woodworking tool, a cyclone installed between the tool and vacuum is a good idea.  The cyclone separates the “big chunks” that would ordinarily plug up your vacuum’s filter and dumps them into a dedicated container.  These cyclones have no motors or moving parts.  
 
Roger

I take it that you are soliciting feedback by writing this:
 
 
"Mill"
If you are going to work metal,  an actual mill is needed
If it is just wood that will be worked - a rebranded EuroTool small drill press - sharp cutters - and a quality XY table will probably suffice.
 
"Table saw (I like the Byrnes machine) (I have a full size machine for rough cuts already)"
Nothing else approaches this.
 
"Thickness planer (I also like the Byrnes for this)"
Nothing else approaches this.
 
"Band saw (cheapo WEN)"
not really a wise choice.  All bandsaws are squirrely to fine tune.  Even benchtop models need to be quality.  Quality is important.
Your entry here says that no serious resaw of boards is an objective.
For effective resawing, a floor model 14" at least will save on frustration
Even for a benchtop paying a bit extra for something like a Rikon 10" for scroll cutting will save you on heartache.
 
"Reciprocating spindle sander (cheapo WEN)"
Not wishing to deal with sandpaper tubes that are expensive, and have limited suppliers,  and compressed drums to hold them at can be made out of round.  I opted for sleeveless - which meant that I supply the motor and mount it vertical.
 
"Disk and belt sander combo (cheapo WEN)"
I have an old Dremel combo -  the disk part is not accurate and I could not get much use from a 1" belt.
The 5" Byrnes disk sander is about as good as it gets.   The only downside is if you plan to sand styrene or other sheet plastic - the Byrnes can not be set to have a reduced speed.
For a belt - a HF 4"x36"  -  it is low cost and shows it - I did not even mount the disk part - the housing on the back is thin - my shopVac pulled it into the backside of the belt. A steel angle brace keeps it away from the belt.  A belt sander does not really need to be precision to do rough bulk wood removal.   Using the rounded end for getting at inside curves works, but keeping the environment from being a dense cloud of sawdust is beyond most any vac.
 
"Lathe (though I think this will be purchased somewhat down the road - BTW, those vintage Unimat machines look the BUSINESS.  They seem to scream quality!)".
I have a Unimat SL 1000 and yes - it is both precise and accurate.  It was a real loss when the company went out of business.
A lathe is mostly an important machine if you intend to work metal.   Most anything it can do to wood can be faked using much less expensive machines - except maybe using a duplicator attachment for multiple copies of cannon and such.
 
Something that I bit on because my budget allowed for it is a Foredom TX flex shaft -  It is a high quality and versatile machine.  Besides the obvious rotary function - there are attachments for a small belt sander, angle grinder - drill,  attachments to make it into a miniature router, and sturdy drill press.  So far for me it is a tool looking for a job,  but it is anything but junk.   The TX model is about power and torque, not blindingly fast rotary speed.

I take it that you are soliciting feedback by writing this:
 
 
"Mill"
If you are going to work metal,  an actual mill is needed
If it is just wood that will be worked - a rebranded EuroTool small drill press - sharp cutters - and a quality XY table will probably suffice.
 
"Table saw (I like the Byrnes machine) (I have a full size machine for rough cuts already)"
Nothing else approaches this.
 
"Thickness planer (I also like the Byrnes for this)"
Nothing else approaches this.
 
"Band saw (cheapo WEN)"
not really a wise choice.  All bandsaws are squirrely to fine tune.  Even benchtop models need to be quality.  Quality is important.
Your entry here says that no serious resaw of boards is an objective.
For effective resawing, a floor model 14" at least will save on frustration
Even for a benchtop paying a bit extra for something like a Rikon 10" for scroll cutting will save you on heartache.
 
"Reciprocating spindle sander (cheapo WEN)"
Not wishing to deal with sandpaper tubes that are expensive, and have limited suppliers,  and compressed drums to hold them at can be made out of round.  I opted for sleeveless - which meant that I supply the motor and mount it vertical.
 
"Disk and belt sander combo (cheapo WEN)"
I have an old Dremel combo -  the disk part is not accurate and I could not get much use from a 1" belt.
The 5" Byrnes disk sander is about as good as it gets.   The only downside is if you plan to sand styrene or other sheet plastic - the Byrnes can not be set to have a reduced speed.
For a belt - a HF 4"x36"  -  it is low cost and shows it - I did not even mount the disk part - the housing on the back is thin - my shopVac pulled it into the backside of the belt. A steel angle brace keeps it away from the belt.  A belt sander does not really need to be precision to do rough bulk wood removal.   Using the rounded end for getting at inside curves works, but keeping the environment from being a dense cloud of sawdust is beyond most any vac.
 
"Lathe (though I think this will be purchased somewhat down the road - BTW, those vintage Unimat machines look the BUSINESS.  They seem to scream quality!)".
I have a Unimat SL 1000 and yes - it is both precise and accurate.  It was a real loss when the company went out of business.
A lathe is mostly an important machine if you intend to work metal.   Most anything it can do to wood can be faked using much less expensive machines - except maybe using a duplicator attachment for multiple copies of cannon and such.
 
Something that I bit on because my budget allowed for it is a Foredom TX flex shaft -  It is a high quality and versatile machine.  Besides the obvious rotary function - there are attachments for a small belt sander, angle grinder - drill,  attachments to make it into a miniature router, and sturdy drill press.  So far for me it is a tool looking for a job,  but it is anything but junk.   The TX model is about power and torque, not blindingly fast rotary speed.

For carvel, I have seen a snake-like spilling pattern on a wide board.
For clinker   I am thinking that only the edge that is overlapped is spilled?

I think it is in Underhill, that after planking the hull, he 'flooded the bilges with varnish'.   I am fairly sure that at the time it was written, varnish was generally taken to be essentially "boiled" linseed oil in mineral spirits.  It was probably a 'one off, seems like a good idea' choice.  Not at all an experiment with controls.   At the time, most of those writing how-to books were few in number, in isolation as far as consensus about what were optimal practices for stability over time.  The closest to that I have seen to objective procedure  are the USN museum standards for models that they would purchase.    I am not sure that any finish will be 100% successful at preventing the hull wood from equilibration with the temperature, O2, and water vapor of its environment. 
 
Clear coating a closed interior may or may not have any practical benefit.  But one definite negative would be if you coat a place where a subsequent wood to wood bond would be.  PVA bonds by intercalating its synthetic hyphae onto a microscopically rough surface and into micro pores that all wood has. Filling the pores and smoothing the surface with a clear finish, or sanding finer than 220 tends to defeat the bonding process.  
 
Coming at this from a POF - with no windows cut into the hull - point of view - if you intend to shellac an interior space before it is closed - I recommend  that any site where a PVA bond is yet to be  have a masking material there before the shellac is applied.  A really tedious and no fun at all sort of prep work and worse, clean up after.  Then, if your OCD about this is intense enough: what to do about the sides and bottoms of the deck beams and the underside of the deck planking between the beams?  It would take the skill of a fiber optic surgeon to get at it after assembly.  If done prior, this adds significant time and work to assembly.
 
To be logical,  worry about  failing due to environmental forces demands  another factor.  Certainly if it is a properly precise wood to wood bond that was clamped with enough pressure to yield a close surface to surface distance.  The polymer chain length should not too long, for a strong bond to form.    This other factor is what I call a belt and suspenders bond.   That is, the bond be both chemical and physical.  The physical part is a dowel at every bond.  The dowel, if visible, should have an in scale diameter.   Pulling bamboo thru a draw plate - especially in the high #70's range is difficult and the yield is low.   It is grad school level - 15% finish  - rather than med school level - >95% - if you get in you will finish unless you work at failing.
 
The bottom line - for a PVA wood to wood bond - always have it raw wood to raw wood.  If you use an agent that allows smooth surface to smooth surface adhesion ( epoxy (?) or CA ("the horror, the horror") )  the worry goes up stream.  The weak point is the clear finish to wood bond.  And no finish ever flakes off?
 
addendum:   In the shower just now, I remembered another option:
For a clear coat for a part of the hull that will never been seen again - give a thought to brushing lacquer.  
It is visibly thick, drys in 2 hours, has been around long enough to show that it lasts.
Too shiny to use where it can be seen.  It is out of scale thick, also.
No spray version - small enough to spray is small enough to breathe - any mitigation will never be as good as never making it airborne to begin with.

An older way
Use a screw plate or sharp angle needle file to cut spiral grooves into the long arm
rotate it in using 2 part epoxy as the chemical  bonding agent.   - I think now there are clear versions of epoxy.
It should be thick and viscous enough to stick after the long arm is dripped.
A pin can be used to poke some epoxy into the hole.
the hole should be a tad smaller than the diameter of the long arm before it is cut for mechanical grab as well as chemical.
Do not leave any wet epoxy on an open surface. when set it is difficult to remove.
It does not soak in and effect subsequent clear finishes like CA does. 
You might wish to place a small piece of masking tape- the old crepe paper type as surgical drape.  Epoxy is sneaky about spreading.

An older way
Use a screw plate or sharp angle needle file to cut spiral grooves into the long arm
rotate it in using 2 part epoxy as the chemical  bonding agent.   - I think now there are clear versions of epoxy.
It should be thick and viscous enough to stick after the long arm is dripped.
A pin can be used to poke some epoxy into the hole.
the hole should be a tad smaller than the diameter of the long arm before it is cut for mechanical grab as well as chemical.
Do not leave any wet epoxy on an open surface. when set it is difficult to remove.
It does not soak in and effect subsequent clear finishes like CA does. 
You might wish to place a small piece of masking tape- the old crepe paper type as surgical drape.  Epoxy is sneaky about spreading.

An older way
Use a screw plate or sharp angle needle file to cut spiral grooves into the long arm
rotate it in using 2 part epoxy as the chemical  bonding agent.   - I think now there are clear versions of epoxy.
It should be thick and viscous enough to stick after the long arm is dripped.
A pin can be used to poke some epoxy into the hole.
the hole should be a tad smaller than the diameter of the long arm before it is cut for mechanical grab as well as chemical.
Do not leave any wet epoxy on an open surface. when set it is difficult to remove.
It does not soak in and effect subsequent clear finishes like CA does. 
You might wish to place a small piece of masking tape- the old crepe paper type as surgical drape.  Epoxy is sneaky about spreading.

I used an old Craftsman 16 gal that could have substituted for a 747 jet engine for noise in a closed area.
I bought a Festool Midi when the Craftsman burned out.  It is quiet, not enough to allow music, but I could ditch the ear muffs.  
Two problems though - it says not to have a cyclone trap in-line - which I insist on - and trap or not, it turns itself off after 15 minutes at most.
Bandsaw ripping of framing stock takes longer than 15 minutes.  Doing the thickness sanding on the ripped stock takes much longer.  A hull of any size (with a reserve for mistakes) can need 60 pieces 2" x 2'  planks.  And that always takes me longer than I imagine it will take  before I start.
The Midi was ~$500 - I checked HomeDepot and bought a 14 gal Rigid for  < $100.  It is about as quiet as the Festool, likes the DustDeputy, and says on as long as I can endure a session.   A radio control On/Off - (budget models bad)  - is a help - it lives in a condo, and I have to unplug it when done because neighbor garage door remotes can turn it on. 
The Festool is $500 that I will never get back and is now a support for the 5 gal cyclone trap catch drum..
My ideal solution for a shop vac is to have it in a room where I ain't.  Alas, this is not an option in a condo.

Using an aniline wood DYE does not affect a PA bond since the surface is unchanged.
A wood STAIN is a surface coat that blocks the teeth of a wood surface.  PVA has nothing to intercalate its polymer chains with = no effective bond.
Agents using surface bonding will bond with a stain layer. Then the bond strength is probably most likely dependent on the strength of the paint to wood bond.  We 'know' that paint "never" flakes off?

Using an aniline wood DYE does not affect a PA bond since the surface is unchanged.
A wood STAIN is a surface coat that blocks the teeth of a wood surface.  PVA has nothing to intercalate its polymer chains with = no effective bond.
Agents using surface bonding will bond with a stain layer. Then the bond strength is probably most likely dependent on the strength of the paint to wood bond.  We 'know' that paint "never" flakes off?

lets see if we can get this going.
I am writing in theory here - so perhaps someone who actually knows will help.
If you did not loft the molds yourself, it may be that you are placing too much trust in their being accurate.
Is not the sheer strake the plank that is just under and up against a rail?   The rail location is the more important component.
If you can fix the hull to a base - and have it midline (stem and sternpost) vertical  - jigs can be made that measure the height of the rail at every mold.
The molds can be trimmed or shimmed as required.  The sheer strake can be spilled to fit the space.  It will be up to you whether to do the rail first or place the rail on top of the sheer. 
 
If the space is not a perfect fit - many ships had a small decorative under the rail strip. It was not as far out as the rail.  It would sit on the sheer strake.   These are NMM multi decker plans and they appear as though the rails were super thick.  I think a more reasonable construction would be two parts.  The under piece would be like crown molding.  There is no reason that a boat could not have this.

An advantage of a machine carved hull is that it can be easily planked using a thin veneer of a scale appropriate species. 
The thin veneer can be spilled using a sharp knife, a steel straight edge, and a sanding block.
It looks to me as though the original Mamoli sold kits for Bounty, Endeavour, and Beagle that used the same lines for all 3, so a careful check of the hull shape using the S.I. plans would prudent.  It is also possible that the planking material provided is all various species of some open pore African Mahogany relative 
As far as a look,  I doubt that even an impecunious merchant owner would send a vessel with raw wood to sea.  Lamp black in a paint carrier would be less expensive than frequently replacing planks.

An advantage of a machine carved hull is that it can be easily planked using a thin veneer of a scale appropriate species. 
The thin veneer can be spilled using a sharp knife, a steel straight edge, and a sanding block.
It looks to me as though the original Mamoli sold kits for Bounty, Endeavour, and Beagle that used the same lines for all 3, so a careful check of the hull shape using the S.I. plans would prudent.  It is also possible that the planking material provided is all various species of some open pore African Mahogany relative 
As far as a look,  I doubt that even an impecunious merchant owner would send a vessel with raw wood to sea.  Lamp black in a paint carrier would be less expensive than frequently replacing planks.