I would agree with Arne that the top three panels of a cambered sail are very effective in sailing to windward in strong winds.  They have less camber in them than the lower panels, akin to what a trysail and storm jib would have.  As Arne further comments, wave action is the most significant factor in the ability to drive a boat to windward in open water, and hull shape plays a factor.  Arion needed to be cracked right off to about 60° in rough seas, due to the boat's great weight and beam.  Then the boat just went for it.  I had some memorable rides, and had no fear about beating off a lee shore in anything less than hurricane force winds.  God help anybody then.

If caught on a leeshore while anchored, which is possible with sudden windshifts and gale-force squalls, it is best to sail out of the anchorage before the sea has a chance to get up.  As the old sailor's saw puts it:  if in doubt, clear out!  If anchoring in open roadsteads, it pays to have a passage plan worked out to let you leave in the middle of the night.  I have never done it, but have stood an anchor watch all night on a few occasions, ready to go at a moment's notice.

I suspect Ingeborg's Folkboat lines would make to windward better than Arion, and might be my choice for a crossing of the North Atlantic against the westerlies.  Arion would triumph off the wind, however!  Despite the lack of volume, a Folkboat remains on my list of potential boats for my return to junk-rig cruising.  I'd prefer the cruising fibreglass version with higher coachroof (it may also have a little more hull freeboard, I am not sure.)  Either way, it would sail to windward very sweetly.  As would Mehitabel!

Mark wrote

A typical junk sail is 2:1 aspect, reefed perhaps 1:1. 

The top triangular panels do improve that.  In fact with a flat cut sail, they are the major contributor to windward drive. In that case the well reefed sail would not suffer as much.

But for a cambered sail not so. 

Mark, is this something you have actually experienced?

All my cambered panel sails have a bit camber in the top panels too. I can ensure you that we go quite well to windward with only three panels set, and attached airflow is shown by the flying telltales. Only with only two panels up, the performance to windward gets marginal. Still, my IF Ingeborg can be tacked with only two panels set, but I think she is exceptional.

My limited offshore experience tells me that it is mainly the seastate which in gale force winds hinders any progress upwind. 

Arne

(Below, Ingeborg 4-up in gusty but inshore conditions...)


(members album of Photos, Section 8)

This is all great fun, but 50 sq m...eeek!!!  I think a 40 sq m sail more than big enough for me.  Arion had 35 sq m, and that was quite big enough for my liking.  It's not really the sweat needed to hoist it (I could use an electric winch or winch handle), and reefing is as easy as ever, but gybing the sail in a fresh breeze would intimidate me, despite the comparative ease of doing so compared to a bermudan or gaff sail of that size.  I like to control the sheet in a gybe, so as to avoid it causing mischief as it sweeps across the deck.  I'd have to consider a schooner rig, but the limitation there, for Mehitabel, is that the mainmast would need to be in the same place as the current bermudan rig mast.  It cannot move further aft because the keel bolts start just aft of the current position, and there is insufficient room in the bilge for the mast step.

However, I agree with the general principle that the ideal sail area is at least the equivalent of main plus 150% genoa, to give superior light-air performance and reduce use of auxiliary power.  For me, realistically, assuming I could not configure (or afford) a schooner rig for Mehitabel, a smaller yacht of about 26 ft would make more sense.  I'd love to build a very simple plywood/epoxy boat, but don't have the facilities to do so, and rents are prohibitive.  I am canvassing the second-hand market for a cheap but solid hull and deck, so far without success.  As so often, everything is a compromise, and I may have to stick with Mehitabel.  The good thing about Mehitabel is that it is an easily driven hull, and the good thing about the skipper is that he is not interested in going fast.  If I can make steerage way and tack, I am content.  Arion was a slow boat, at least with the wind forward of the beam, and we sailed happily together for 23 years, the first 10 without an inboard engine.  Mehitabel is a rocket-ship by comparison.

The final challenge is that I have very limited funds.  I don't think my books are likely to make me a lot of extra cash, and I'm reluctant at 72 to go back to a regular job, even if I could find one suitable for an old seadog.  Besides, I already have enough writing projects in the pipeline to keep me busy.  It's lucky that junk rig remains the most economical approach to re-rigging a boat!

A full size JR for a Cavalier 32

By checking the Bermudan rig a little more closely, now with mainsail plus Genoa1, I found their combined area to be 49.6m².

Then I wondered, could one make a junk sail of that area which fitted on that mast through the hatch, and still with its CE at the Bermudan mast?
I put my secret, but not too accurate formula in use:

SA = B² x AR x 0.78.

Trying different batten lengths, I ended up with B=5.80m and AR=1.90 (..it helps to have experience with these numbers...)
This gave SA= 5.80² x 1.90 x 0.78 = 49.85m² .

By assuming that the CE sits on the middle of the boom, I drew up a 5.80m test boom with the middle of it where the other JR (40.1sqm) had it.
This resulted in a mast balance of about 28%. None of my master sails would handle this, but I had made a couple of prototype drawings with yard angle of 55°, which would accept about 25-30% mast balance.
As luck would have it, I had drawn such a master sail with AR = 1.90 (same as Ingeborg is using). However, the lower yard angle would push the CE a little aft of the middle of the boom, so I found its true CE the old-fashioned way; by balancing a printed out sail on a ruler.

This new master sailplan was imported into the Cavalier’s sailplan, and scaled up to B=5.80m. A little cross was added to the boom, representing the true CE. Then the new sail was moved into place on the sailplan.

It surely looked big, now with its accurate area found to be 49.6sqm! However, it is almost identical in size to the sail I had on my Johanna  -  and she was one ton lighter. The batten length is also the same, at 5.80m.
Thanks to the slightly corrected CE position, the sail was now shifted a few cm forward, resulting in a mast balance of 29%.

As you can see on the diagram, the sail fits well on the mast, and can actually have its mast balance adjusted between 25 and 30%.

Still, I do not recommend this sail, with this high mast balance, until a smaller version has been tested. This is only a study, driven by curiosity, and not a proposal for Graham...

It remains to be found if...

1.      ..the sail feathers reliably when the sheet is being eased.

2.      ..the boat sails well on both tacks.

3.      ..the battens do not S-bend at the mast

This post is already too long so I stop here...

Arne

(.Check Arne's sketches, section 7-44...)

Ok, good point for light winds.

But what about when you have to claw off a lee shore in strong winds?  
Is a well reefed large sail less efficient than a less reefed smaller sail?

One advantage, dare I say it, of a pointy sail is that it maintains its aspect ratio as you reef down