Bagels for Sunday brunch have become altogether too easy, and while I keep telling people I no longer cook to show off, what I really mean is that I no longer cook only to show off. Of course there remains an element of demonstrating my mastery, to myself as much as to anyone else. What finer way, then, to do so casually than to conjure up a basket of proper croissants, here in the land of the oversweet, unbuttery cornetto?

I knew the principle already; essentially a sweet puff pastry, dough finely laminated with butter. And I knew, in theory, how to make it. Dan Lepard, whose recipes are generally really good, had a detailed one for croissants in Exceptional Breads. Best of all, croissants are, they say, equally good reheated from frozen, which would be easy enough in an oven already hot from bagels. So, on Friday evening, I made my dough, following the recipe precisely.

It was very stiff, stiffer by far than bagel dough.

I checked and rechecked the recipe and reaffirmed both that I had followed it to the letter and that 220ml of liquid for 500gm of flour was both correct, as printed, and very, very wrong. Naturally, I flew to the internets. Dan Lepard, who in the past has been very helpful, was MIA. Last Friday evening, his website read “14th October 2013 Working on a new website design, should be ready soon.” It still does.

Search engines were no help; nobody else appears to have even tried that particular recipe. Or if they did, they failed to publish their difficulties.

I searched further, for croissant recipes, and though there seemed to be lots of them, the vast majority were from the US and measured volumetrically, making any quick calculation of hydration a pain. At last, I found Jeremy Shapiro’s series of posts detailing his adventures in pursuit of the perfect croissant. Now, I’ve been slightly snarky about this other Jeremy before, and he took it in very good part, so I decided to ask his advice, even overcoming my reluctance to sign in to leave a comment. And, a true gent, he replied: 58% hydration, as against the recipe’s 44%.

Long story short, next morning I sweated and strained to work the extra milk into the dough, schlepped my way through the butter lamination, the rolling out, the cutting and the shaping, and baked the little buggers.

They tasted great — with that much butter, sawdust would taste great — but they had failed to rise and really were not presentable. With hindsight, they were destined to fail. As Jeremy S. told me (too late, alas): “Start over! Or the detrempe will be overworked!”

And that’s what I plan to do, half a batch at a time, until I have failed often enough to get them right.

The failures are still in the freezer; too good to throw away, too bad to serve to anyone else.

When I was growing up I was fortunate enough to have a mother who really cared about “healthy” food. She wasn’t quite what one might today call a health-nut, but the bookshelves featured Adelle Davis and John Yudkin, and her baking featured plenty of wholemeal flour. The bread, in particular, was a leaden brown brick, far better toasted than fresh, tasty and filling and quite a lot of work to chew. I remember too a batch of homemade blackberry jam, made with brown sugar, probably Muscovado rather than Demerara, which was always under suspicion of being white dyed a bit brown. That jam set so solid it needed brute force to prise it from the jar. It had to be sliced, rather than spread, but sliced onto that bread, made the chewing worthwhile.

The other thing about the bread was that it was easy as pie; just roughly mix flour, water and yeast with some salt and sugar, pour the resulting goop into greased tins and by the time the oven was hot it had risen a bit and was ready to be baked. This was the famous Grant loaf, perfected and made famous by Doris Grant, another pioneering health-food nut, who as it happened lived to be 98. The Grant loaf was pretty popular, and I’d be lying if I said I didn’t enjoy it. I even made it myself a few times, back then, but I also think it was those loaves that started me baking my own bread, for no other reason than to have something else under my peanut butter and sliced jam.

The Smithsonian article set me thinking. First, was the Grant loaf as simple as I remembered it? Yes; although one of the recipes I found suggested sieving out the bran. Huh? The same recipe said to make it half and half with white flour. Huh? squared. But then, all Hamelman’s recipes for wholemeal breads seemed to call for half and half. Elizabeth David had a few versions of Doris Grant’s own recipe, and so did Saint Delia, and all of them seemed to ask a lot of yeast to do its work in a little time.

I can do better than that, I thought, making use of what I’ve learned in the past few years. So I had a go, using a natural leaven not commercial yeast. And it worked. Beyond my craziest expectations.


350 gm ripe wholemeal leaven at 75% hydration
800 gm strong wholemeal flour
700 gm water
17 gm salt


Prepare your leaven by feeding your starter with 200 gm of wholemeal flour and 150 gm of water. As it happens I keep two starters going, one a 75% wholemeal starter, the other a 100% white starter, and I used the brown, but I don’t think it matters much what your starter is. Allow to ripen for 12 hours or so.

Weigh the leaven into a large bowl, add 700 gm of water and stir about a bit to break up the leaven. Add the salt and the flour. I added the flour in three batches, two of 300 gm and one of 200 gm, stirring well after each batch to make sure all the flour was fully wetted. At this stage the dough is very soft, pretty sticky, and has very little structure. Don’t worry. Turn the dough onto a work surface and do a couple of stretch and folds. The S&F technique is much harder to describe than to do, and every time I do it my hands are too messy to use my camera, but my bread-making friend Dan had some good pics on his site. In essence, and usually with the help of a scraper to start, you pull the dough out to one side and fold it back to the middle. Then the other side. Then backwards, and finally forwards. You end up with a parcel of dough, which even after the first four stretch and folds is way more structured than it was when you began. I do this twice, for what I call one set of stretch and folds.

Return the parcel of dough to the bowl, cover, and allow to ferment for 30 minutes. Turn it out, do another set of stretch and folds, and return to the bowl for another 30 minutes. Feel that structure building up? Return to the bowl, this time for an hour before another set of stretch and folds.

Stretch and fold after 30, 60 and 120 minutes

To recap, you S&F immediately after mixing, and then after 30 minutes, one hour and two hours. My notes say:

  • 9:15 Holding shape.
  • 9:35 Stiff, holding shape.
  • 9:50 Really elastic, 2nd fold hard to stretch.
  • 11:00 Well structured, still a bit sticky, into fridge.

Put the bowl in a plastic bag and put it in the fridge to continue its slow fermentation. I left it in the fridge for 3 hours, but 6 or even overnight would probably be fine. The dough had risen well, easily doubling in volume. Bring the dough out and allow it to warm up for about an hour. Gently scrape the dough out onto the counter and do a gentle stretch and fold. At this stage, handling needs top be very light as you don’t want to get rid of all the gas and bubbles that have formed. Divide the dough parcel in two and shape each as you prefer. I formed batard shaped loaves and put them into a well-floured couche to rise for another hour. That may have been a mistake. The loaves are still very soft, and hard to handle; a banetton or tin might have made life easier.

Preheat the oven to 230°C and prepare for steam too. Gently, again, turn the loaves onto a piece of baking parchment, slash to suit, and slide the parchment onto a tray or baking stone in the oven. Bake for 30 minutes, at the end of which remove the source of steam, turn the loaves and reduce the heat to 210°C and bake for another 20 minutes. Test that the bread is done (hollow knocking sound, or internal temperature around 95°C) and if it is turn the oven off and leave the loaves in there for another 5–10 minutes with the door slightly ajar. Cool the bread on a rack and, if you can, wait until thoroughly cool before slicing.

The result? As I said, beyond expectations. This bread is light, the interior soft and open, but not too open, the crust crunchy but not too thick. And the taste has the nuttiness of wholemeal and that interesting sweetness, without any added sugars. A greater contrast with the Grant loaves of my youth I cannot imagine, and yet it was those loaves that made me able to bake this one. Hard to bake? I suppose so, but not really, given a bit of experience. Worth repeating? For sure.

And there’s a set of the images here, if you’re interested.

A couple of weeks ago some of the alumni of Sourdough U and potential new students were in Tuscany, celebrating a milestone birthday, and asked if I could do another bread-making class. Of course. We toyed with bagels, but what’s the point, really, for people who can buy a good bagel any day of the week? And we toyed with pizza, given that the place had a wood-fired oven, but without an expert in wood-fired ovens, which I am not, that would be too tricky. In the end, we settled on Hamelman’s Sourdough Seed Bread, which has become a recent favourite, and it went very well. Not content with that, though, I decided to try my hand at ciabatta.

Ciabatta does actually share one key characteristic with bagels: abominations abound, little breads that have nothing in common with the original beside their shape. The secret to a good ciabatta is a very wet dough — Hamelman’s formulae are 73% and 72% — that permits the characteristic giant holes and blistered crust. Building strong gluten that will hold the gas for holes is a challenge in such a wet dough, and one that seems to require a machine, but I remember reading ages ago a piece by Susan of Wild Yeast about making ciabatta rolls by hand. The key is a technique called double hydration: you make a good strong dough with some of the water, and when that is done incorporate the rest of the water to increase the hydration. I’ve had to add extra water to a dough before, when I’ve made a mistake in measurements, and under those circumstances it can be a real pain. Doing it deliberately sets up different expectations, although working with an 75% dough is not for the faint-hearted. I followed Susan’s recipe and method pretty carefully, although I use slightly less salt.


610 gm ripe starter at 100% hydration
465 gm bread flour
75 gm whole wheat flour
26 gm olive oil
15 gm salt
355 gm water
Extra flour (lots of it) for dusting


Combine the starter with 280 gm of the water, the additional flour, the salt and the olive oil and bring this dough together as a rough mass. Turn it out onto a counter and knead until the dough is smooth and elastic. At this stage hydration is 62% and so the dough might be a bit sticky but should nevertheless be quite easy to knead without any extra flour. It took me about 10 minutes to get a well-developed dough.

Now comes the fun. Return the dough to the bowl and add the remaining 75 gm of the water. At this point, you need to get the water into the dough, rather than slipping around entertainingly on the surface. A combination of squidging through your fingers, folding the dough in on itself repeatedly, whirring your hand around pretending to be a dough-hook; whatever it takes. Hydration is now 70%, and I’d leave it there the first time you try this. As you become more proficient you can go up to 75% (which would mean 115 gm of water in the second hydration) or beyond. The tricky part is not so much hydrating the dough but handling it after the bulk fermentation.

Sticky ciabbata dough.

With all the water in the dough, lightly oil a clean bowl and pour the dough into it. Cover, and set aside for 30 minutes. Give the dough a couple of sets of stretch and folds, either in the bowl or, better yet, on a floured work surface, brushing off any surplus flour so as not to incorporate it into the folded dough. Return to the bowl and stretch and fold again at 60 minutes and 120 minutes (i.e. 3 sets of stretch and fold during the 2 hours of bulk fermentation). It is truly remarkable how much structure the dough gains after a couple of rounds of stretch and fold; Susan has great pictures, I was too preoccupied to take any. At this point, put the bowl in a plastic bag and pop the whole lot into the fridge for 7-12 hours, or overnight.

Remove the bowl from the fridge and allow it to warm up to room temperature for a couple of hours. Dust your worktop with flour (Susan recommends a mixture of half and half flour and semolina, but I used plain flour) and gently turn the dough out of the bowl, trying not to degas it at all. Gently stretch the dough out into a rectangle and divide it into 10-12 portions with a well-floured dough cutter. Equally gently, transfer these to a well-floured linen couche, keeping the floury side (the bottom) down and spacing them well so they do not stick to one another. The cut edges of the pieces are sticky, which makes the whole process of cutting and transferring fraught with difficulty, but with a little bit of luck and a lot of dexterity you can get them into the couche for a rest and further rise, under a cover, of about 90-120 minutes. The rolls will be very light.

Preheat the oven, preferably with a baking stone, to 250°C, or as hot as you can get, and prepare also to be able to steam the oven using whichever method you prefer. Gently flip the rolls onto a piece of baking parchment, slide the whole parchment onto the stone and bake with steam for 5 minutes. After five minutes, turn the oven down to 240°C and bake without steam for another 15-20 minutes until the rolls are the colour you like. I prefer them quite pale. Remove from the oven and cool on a wire rack.

I confess, I was very pleasantly surprised by how well these turned out, and so were the Sourdough U alumni. To consolidate my learnings (sic) I made another batch last Friday, pushing the final hydration up to 75%, and they were just as good, if not better. A lot of work, but really worthwhile.

p.s. Thanks to Christy Lichtenstein for pointing out an error in my original quantities for the total water. Now fixed.

Finding a better vocabulary for tastes and smells is definitely something I need to do, because faced with putting the sensation of baked kalonji into words, I fall back on US Justice Potter Stewart’s famous definition of pornography : “I know it when I see it”. And I like it. Kalonji, not hard-core pornography. I’ve used it before in bread but not lately and not kneaded into the dough. So, with a bunch of ripe 100% starter at hand, I thought I would do an experiment.


530 gm ripe 100% starter, made with tenero flour.

100 gm wholewheat flour (10%)

100 gm whole rye flour (10%)

15 gm salt (1.5%)

40 gm kalonji (4%) I would normally aim at 5% but had only 40 gm.

25 gm olive oil (2.5%) because I wanted a little of the softness

535 gm tenero forte 0 flour (taking total flour weight to 100 gm (100%)

400 gm water (taking total hydration to 66.5%)


Add the water, wholewheat flour, rye flour, salt and kalonji to the starter and mix to incorporate. Now add the oilive oil and the rest of the white flour, stir until mostly mixed, and tip out onto the counter. Work in the rest of the flour, kneading and folding as you go. The rye makes to dough a little sticky, but it will quickly develop some structure. Knead quickly for about a minute after it has all come together properly and then return to a bowl to rise. Cover with a plate and bulk ferment for approximately 2 hours. Turn the dough out and knead again, quite rapidly, for about 30 seconds. Return to the bowl and rest again for about an hour.

Turn the dough out, degas gently, divide in two or three and shape loaves.

At this point I refrigerated the shaped loaves overnight, or you could leave them out until proved, about 2 hours.

Preheat the oven to 230°C and prepare to use steam.

Slash the loaves and place them in the oven, with steam. Bake for about 22 minutes. Then remove the steam, turn the temperature down to 220°C and bake for a further 22 minutes. Test the loaves for doneness and if done remove and place on a wire grill to cool.

There are some more pictures here on Flickr.

The way rye always gives the starter a big boost is really interesting, although the still-cold weather meant that it was Saturday night before I could feed it up again. I decided to go for Jeffrey Hamelman’s 40% caraway rye, without caraway seeds because my client doesn’t much like bits in her bread. Which is why she doesn’t get bread as often as I bake, but that’s another story.

I’m not going to repeat the recipe and method here because other than omitting the caraway seeds I didn’t do anything to make it my own. Normally I don’t add the little bit of yeast that Hamelman’s sourdough loaves often call for, but this time I did, partly in order to be able to bake it on Sunday night. Rye doughs, even at only 40%, are often sticky and so I tend to work it by smooshing it along the counter with one hand and using a scraper in the other to bring it back together.

In addition to omitting the caraway seeds, I also began building the rye starter with 150 gm rather than 20 gm. The idea was to have enough left-over starter to try something else: crispbread. For this I turned to Dark Crisp Rye Bread in Dan Lepard’s The Handmade Loaf, another favourite book. Again, I’m not going to copy out the recipe. I hadn’t made quite enough extra starter, so I ended up reducing quantities to 75% overall. Rolling out the dough proved not nearly as awkward as I expected, probably because I followed the instructions to dredge it with rye flour. I even managed to make the whole thing almost rectangular. I suspect that if I really got into this I might try harder for neatness in future, but not this first time.

The loaf looked good coming out of the oven — I’ve developed a new scoring pattern that tends to elongate a round loaf, which I like — and the timing was perfect for the crispbread too. I’m not going to win any prizes for the regularity of my dimples, but they worked reasonably well, and after scoring the dough roughly into squares, into the oven it went “for 40-50 minutes, or until the upper surface is lightly tipped with golden patches against the grey rye crust”. And off I went, timer in hand, to do some other work, out of smelling range of the oven.

Big mistake.

Forty minutes later the oven greeted me with an acrid smell and even some smoke. The bread was charred, rather than merely crisp. It seemed so bad, I really was wondering whether there was a mistake in either the temperature or the time. Seems like I wasn’t the only one.

I’m adding in a comment about the Dark crisp rye bread as well (p. 167). I was very excited to try this as we love knackerbrod. Despite my initial misgivings about the temperature and the bake time (425 F for 40 – 50 minutes), knowing that the light rye flatbreads were a mere 400 F for all of 25 – 30 minutes, I went ahead and followed the instructions to the letter, as I usually do the first time around. What a mistake. 30 minutes into the process, the bread was burnt to a crisp. It was black in several places. Now as I mentioned I’ve had my oven re-calibrated twice and I do not have this problem with other books. So if you do buy the version out right now, go ahead and lower the temperature and bake time on that recipe, and don’t be afraid to tweak things ahead of time if they look a little off.

That’s actually strangely reassuring. I am not alone. In the end, the flavour of these crispbreads was rather good, making allowances for the bitterness of the burnt bits, and I will definitely be trying them again.

Whether that will be before the next entirely unpredictable Pasta Madre Day rolls around, I cannot say. And in its own strange way, the very unpredictability of Pasta Madre Day (there seems to have been one in December 2011, but no indication of whether it is annual, but roves around, like Easter, or happens only when people remember to do it) is at odds with baking with a natural leaven. You can’t just suddenly decide to make a sourdough loaf. You need to have a starter, and you need to feed it up before you begin to make your dough. Unless, of course, you keep a large amount fed all the time just in case, which strikes me as incredibly wasteful. So I’m confused. I didn’t need Pasta Madre Day to bless my baking. But I rather like the idea of my baking helping to contribute to a wider appreciation of the joy and satisfaction to be had from baking with natural leavens.

Maybe next time I (and Pasta Madre Day) will be better prepared.