The Fresh Loaf

News & Information for Amateur Bakers and Artisan Bread Enthusiasts

San Francisco Sourdough Bread: Was the secret worth revealing?

dmsnyder's picture
dmsnyder

San Francisco Sourdough Bread: Was the secret worth revealing?

San Francisco Sour Dough French Bread

according to

Kline, Sugihara and McCready

(The Bakers Digest, April, 1970)

 

In their 1970 Bakers Digest articles, Kline, Sugihara and McCready first deliniated the microbiology of San Francisco Sourdough bread. Their discovery of the special relationship between the dominant yeast – S. exeguna – and a unique species of Lactobacillus provided understanding of the special flavor of San Francisco sourdough and the stability of the sourdough cultures over time.

In the first of their articles, they also described the process San Francisco bakeries used to maintain their starters and make their breads. They describe the process as if all of the bakeries used the same process without actually stating this was the case. Regardless, the process they describe is significantly different in several respects from those found in any of the currently popular baking books in my collection. Because of that, and because I'm curious about whether the process they described can be successfully replicated in my own kitchen and produce bread like that of the traditional San Francisco sourdough breads with which I'm familial, I made a couple of loaves following their procedures.

The formulas and procedures described by Kline, et al. in the articles cited are as follows:

Sponge

Bakers' %

Firm starter

100

High-gluten flour

100

Water

46-52

  1. Ferment for 7-8 hours at 80º F

  2. In the bakery, fed every 8 hours

  3. Can feed less often by fermenting for 6 hours and retarding at 50-55ºF or fermenting 3-4 hours and retarding for longer times with refreshments 3-4 times per week, rather than 3 times per day.

 

Dough

Bakers' %

Sponge

20

AP flour

100

Water

60

Salt

2

  1. Mix ingredients (No mix method or time is given.)

  2. Rest for 30 minutes at room temperature.

  3. Divide and pre-shape.

  4. Proof for 1 hour at 90ºF

  5. Shape.

  6. Proof for 6-8 hours at 85-90º F.

  7. Score loaves and Bake with lots of steam for the first half of the bake at 375-390ºF for 45-55 minutes.

To make 2 kg of dough, I proceeded as follows: 

Sponge

Bakers' %

Wt (g)

Firm starter

100

88

High-gluten flour

100

88

Water

50

44

Total

250

220

1. I built the firm starter for the sponge with two elaborations, starting with my firm stored starter.

2. The sponge was mixed and fermented at 80º F for 10 hours.

 

Dough

Bakers' %

Wt (g)

Sponge

20

220

AP flour

100

1099

Water

60

659

Salt

2

22

Total

182

2000

Note: Accounting for the flour and water in the sponge, the actual final dough hydration was 58.9%.

Procedures

  1. All the dough ingredients except the salt were mixed and allowed to rest, covered, at room temperature for one hour.

  2. The salt was sprinkled over the dough and mixed in a Bosch Universal Plus mixer at First (lowest) speed for 1-2 minutes, then at Second speed for 5 minutes.

  3. The dough was then divided into two 1 kg pieces, pre-shaped as rounds and proofed at 90º F for 1 hour. (I placed the pieces on a bakers' linen-covered 1/4 sheet pan in a Brød & Taylor Folding Proofer.)

  4. The pieces were than shaped tightly as boules, transferred to floured bannetons and placed in plastic bags.

  5. The loaves were then proofed in the Folding Proofer at 85º F for 6 hours.

  6. 45 minutes before baking, the oven was pre-heated to 450º F/convection with a baking stone on the middle shelf and a cast iron skillet filled with lava rocks on the bottom shelf.

  7. The loaves were dusted with semolina, transferred to a peel, scored and loaded on the baking stone. A perforated pie tin filled with ice cubes was placed on the lava rocks. The oven was turned town to 380º F/conventional bake.

  8. After 23 minutes, the skillet was removed from the oven. The oven setting was changed to 360º F/convection bake, and the loaves were baked for another 22 minutes.

  9. The oven was turned off, but the loaves were left on the baking stone, with the oven door ajar, for another 15 minutes.

  10. The loaves were then cooled on a rack before slicing.

This is a low-hydration dough. After mixing, it was very smooth and barely tacky in consistency. It was still very dry and firm before the final shaping. After the six hours final proofing, the dough was very soft and puffy. I was afraid it was over-proofed and would deflate when scored or, at least, have poor oven spring and bloom. However, It took the scoring well and had good oven spring and bloom.

The baking temperature was lower than I usually employ for lean dough hearth loaves, and the crust color was thus lighter than most of my bakes. The crust color was quite characteristic of the San Francisco Sourdough breads I remember from the 1960's and 1970's. The loaf profile also was typical – a rather flat loaf.

The aroma of the vented oven air was remarkably sour during the final 10-15 minutes of the bake. The bread, once baked, cooled and sliced, revealed a very even crumb. The crust was somewhat crunchy but more chewy. The flavor was minimally sour and had little complexity or sweetness to it. All in all, a handsome loaf with undistinguished eating quality.

I expect I could tweak more sourness out of this dough by retarding the loaves overnight, but I don't particularly feel inclined to experiment further when there are so many other breads that are so much better.

 Comments regarding the process described would be very welcome. I am curious regarding the disparity between the San Francisco Sourdough's I have had, supposedly produced by this method, and what I baked at home. Any thoughts?

David

 

Comments

GSnyde's picture
GSnyde

Now we don't have to.

I wonder if the bread would taste different if made with a sourdough starter that lives in San Francisco.  

The loaves are very pretty, and I hope they make good toast or croutons or something.

Glenn

Elagins's picture
Elagins

your starter produced/produces superb loaves: wonderfully sour with a tantalizing note of sweetness owing to the long ferment and my custom of retarding my sourdoughs for 12-24 hours.

Stan

Syd's picture
Syd

The hydration is (what I would have thought)  on the low side for this type of loaf.  The crumb is consistent with a low hydration, but I would have expected a higher profile. Interesting.  Perhaps that is due to only 7 minutes of mixing and no stretch and folds during the bulk ferment. 

Perhaps this formula with the addition of 5% rye or 10% WW would make for a more interesting flavour. 

Best,

Syd

dmsnyder's picture
dmsnyder

I think the gluten was well-developed by the mixing. The character of the dough at the end of proofing suggested to me that the long, warm proof resulted in some proteolysis, and that accounts for the loaf profile.

The addition of some whole grain rye or wheat would certainly enhance the flavor, but this bread, traditionally, was very white, and the best of it had pronounced sourness with a nice balence of flavors. This bake just didn't hit the mark.

David

Mebake's picture
Mebake

Great looking Loaves, David! Neat scoring, and beautiful color on the crust. The crumb is typical of a low hydration, as Syd said.

I havn't had a sourdough bread from a bakery in Dubai save from my kitchen, let alone San. Fransisco sourdough. However, your bakes, especially the beautifully crafted San Joaquin SD inspired me. I think you should stick to those breads you like most.

Yerffej's picture
Yerffej

David,

I have made similar, but not identical, bread using half of the dough as you did and cold retarding the other half for baking one day later.  The dough that cold fermented for 24 for hours longer was so incredibly superior to the dough baked on the first day that it  made you believe that they could not possibly be the same dough.  As for your original recipe and technique...who knows wheher or not it is accurate, how many actually did it that way, and what the quality of that particular bread might have been.

Jeff

BurntMyFingers's picture
BurntMyFingers

I used to come up from Los Angeles and would invariably grab a Larraburu baguette and a whole crab and munch contentedly on the waterfront. That light, chewy crust is something I've seen replicated along with a very strong sourness at some "artisinal" bakeries in surprising locations (such as supermarket bake shops); there is a good version made by Rock Hill Bakery in my upstate NY location. I've wondered if maybe some lemon juice, yogurt or vinegar is added to get such sourness.

Interesting that the loaf smelled sour when baking, but the final bread wasn't sour. Thanks for sharing.

Otis

txfarmer's picture
txfarmer

Very interesting procedure. Two things stood out for me:

1) Relatively frequent feeding of the starter. Maybe the temp at the bakery was high? Every 8 hours, it's almost like the schedule for Italian sweet starters, and those are fed to avoid sourness!

2) That's a VERY long proof, and barely a bulk rise!

As someone who just spent this past holiday season turning out sourdough panettones, this fermentation schedule is similar, which is interesting, since I'd think that these are two styles of breads with very different aims.

dmsnyder's picture
dmsnyder

I have to believe that there are some important details left out of the procedures reported by Kline. One cannot believe that all of the 1960's San Francisco bakeries used exactly the same formula and procedures, and there were unique qualities to each bakery's sourdough bread.

I think we know that the main variables affecting flavor (assuming a white, wheat bread) are time, temperature and hydration. The percent starter in the final dough mix also will impact fermentation time and bread flavor, especially sourness.

I think the safe to assume commonalities among the SF bakeries' formulas and procedures were:

1. Frequent feedings of a firm culture based on a pâte fermenté. In other words, a piece of the complete dough from one day's bake was saved and fed every 8 hours, until needed for the next day's dough. (Although Galal's description of the Larraburu Bros. process has the pâte fermenté refrigerated until the next day, when it had a single feeding before mixing in the final dough.)

2. Dough hydration of 60-65%.

3. Around 20% pre-ferment in the final dough.

4. A very short bulk fermentation and relatively long proofing at warm temperature.

5. Baking at low temperatures (for hearth loaves) for a long time.

6. Most bakeries did not retard their formed loaves, although I cannot recall ever seeing bread from Boudin without the crust bubbles characteristic of loaves that have been retarded.

This still leaves lots of room for important variations.

David

CarlSF's picture
CarlSF

I made this bread before using the same fermentation procedure that you stated, and the bread came out sour for me.  It didn't quite have that tang that is usually associated with SF Sourdough, but it was sour.  The only thing I did differently was that I proofed the loafs for 8 hours instead of 6 hours.  I'm not sure if that would make a difference or not.

Carl

dmsnyder's picture
dmsnyder

Very intersting!

Did you proof at the recommended temperature? Maybe a longer proof is what it needs. Maybe it is worth fiddling with.

David

CarlSF's picture
CarlSF

David,

Yes, I proofed the loaves (2nd fermentation) at the recommended temperature (85 F).

Carl

CarlSF's picture
CarlSF

David,

Since you took Artisan II from SFBI, did they ever demonstrate to the class on how to make bread that's really sour?  If I remember correctly, Artisan II was about creating, maintaining, and working with a starter.

Carl

dmsnyder's picture
dmsnyder

To get more sour flavor:

1. Feed starter "only" once a day.

2. Use a firm starter.

3. Either use a small inoculation with a long, cool fermentation or use a large inoculation (% pre-fermented dough)

4. Cold retardation of high-hydration dough in bulk or low-hydration dough as formed loaves.

However, the SFBI staff are not big fans of super-sour bread. They are after the complex, sweet flavors of well-fermented doughs with a "balanced" contribution of lactic and acetic acid. Probably it's the French influence on bread preferences. 

David

CarlSF's picture
CarlSF

David,

Did they give you any recipes for making several different sourdough breads?

Carl

dmsnyder's picture
dmsnyder

David

CarlSF's picture
CarlSF

When I took the class, the instructor said feeding the starter 3x a day is the ideal situation.  The instructor said that feeding the starter on time was ideal as well.  Were any of the sourdough breads that you made in class came close to being like SF Sourdough bread?

Carl

dmsnyder's picture
dmsnyder

Who was your instructor?

Frank Sally was my instructor. He favored a twice a day feeding schedule, it seemed.

There was no attempt to make a very sour bread. We made 3 versions of a mostly white flour pain au levain. The most sour was one that used a firm starter at about 40% bakers' % of the final dough, as I recall. I should go look it up. As I recall, of all the breads we made, the miche was the most sour, actually. I'm pretty sure the levain was fed twice a day for that bread. 

I really need to review the handouts and my notes. If there is any good, new information, I'll share it.

David

CarlSF's picture
CarlSF

David,

My instructor was Didier Rosada.  He said feeding 3x a day was optimal, but for home bakers this may not be the case.  Feeding your starter 2x a day is fine as well.  When I took the course, we explored the differences of having high %s and low %s  of levain in the final dough as well as using liquid and stiff levains, so we know what to expect in terms of taste.

Carl

ehanner's picture
ehanner

David,

I was reminded by your first photo that these breads from the 60-70's didn't have a dark bold bake. It was a long time ago for me to remember but the golden color stuck with me. The crumb you got with this procedure is nothing like I remember. As you point out, it seems unusually low in hydration for this type of bread.Perhaps all that is needed to change is an increase in hydration and a bulk ferment based on performance (seeing the gas pockets) and proofing to a greater degree.

The use of a sweet starter is the beginning of the flavor profile. Similar to Robertson's feeding regime, the lactic acids should be the prominent flavor component I would think. If I recall, Debra Wink makes the point that LAB's are more active at higher temperatures. Add to that idea the encouraging of LAB's in the short cycle feedings. The concept seems to make sense to me.  I don't see the bakers using a high temperature proof for production reasons. It had to be a flavor control consideration.

The one other thing that would have an impact on crumb quality and flavor is the flour. If you have an inroad to discuss this with Nicky, perhaps he could shed some light on what was being used back then and what is available today.

I'm really glad you have published this old document and your work to recreate those old breads David. The process is so unique it must be authentic to some degree. This was the  bread that made me take notice of good food. I was only 18 years old at the time and I have a clear memory of that day on the Wharf with my parents. That boule  was wonderful. My 89 yo Father also remembers how good it was to this day. I don't think we would have been so struck by the flavor if it had been  overly sour.

Eric

chris319's picture
chris319

In rereading this thread in which dmsnyder describes his results as:

The flavor was minimally sour and had little complexity or sweetness to it. All in all, a handsome loaf with undistinguished eating quality.

One thing jumped out at me

2. In the bakery, [sponge] fed every 8 hours

My question for dmsnyder if he sees this post: I've read the same articles posted by doc.dough. I'm curious where you got the notion that the bakery refreshed their sponge/starter periodically, every eight hours over a span of ??? hours for a total of ??? refreshments.

Quoting the article:

Table I

Sour Dough Starter Sponge
100 parts previous sponge
100 parts flour (Hi-gluten)
46-52 parts water

Make up and hold 7-8 hrs. at 80°F

The article goes on to describe the preparation of the dough.

What I get from this is that they made their sponge/starter, held it for 8 hours, then used it in their dough to make bread. I don't see where they refreshed the starter periodically on an eight-hour cycle for two, three or however many refreshments.

Quoting the article on Larraburu:

Each day a piece of straight dough or starter sponge known as the "Mother" is saved and refrigerated to be used as a starter sponge the following day.

This suggests their starter got a 24-hour "cold retard". There is nothing about a periodic eight-hour refreshment.

The [sponge] ingredients are mixed and fermented for 9-10 hr at 80°F.

The dough rests 1 hr and then is divided, molded, and deposited on canvas dusted with corn meal or rice flour.  The dough is proofed for 4 hr

baked at 420°F (216°C) for 40-50 min

If we consider the process from the standpoint of a bakery with workers putting in eight- or nine-hour shifts, the following timetable makes sense:

Sponge mix and ferment: 10 hours

Dough rests, divide and mold: 1 hour

Proof: 4 hours

Bake: 1 hour

Total: 16 hours = two 8-hour shifts.

The other bakeries worked on the following timetable:

(Sponge) Make up and hold 7-8 hrs

(Dough) Make up – approx. 1 hr floor time (includes 30 min. relax, scale, 20 min overhead proof) mold

then proof 8 hr

Bake 1 hour

Total: 17 - 18 hours = two 8.5 or 9-hour shifts

dmsnyder's picture
dmsnyder

The three times a day feeding routine was from the Kline, Sugihara and McCready (The Bakers Digest, April, 1970) article. The routine you are citing is from the Galai, et al. article ("Lactic and volatile (C2-C5) organic acids of San Francisco sourdough French bread." Cereal Chemistry 55(4): 461-468; Copyright 1978 The American Association of Cereal Chemists)

The latter article describes Larraburu's method. The Kline, et al. article generalizes without attributing the method they describe to a specific bakery.

David 

chris319's picture
chris319

I quoted both articles in my post. One of them was the Kline - Sugihara - McCready article from 1970 for Baker's Digest. Here is doc.dough's citation:

the (hard to find) 1970 paper "Nature of the San Francisco Sour Dough French Bread Process, I. Mechanics of the Process" by Leo Kline, T. F. Sugihara, and Linda Bele McCready in Baker's Digest 44(2), p48-50.

Below is doc.dough's excerpt. Where do you get that the starter was periodically refreshed every eight hours? All it says is that the starter was made up and held for 7 - 8 hours -- nothing about periodic refreshment every eight hours. What I get from it is that they made the starter, held it for 8 hours, then presumably used it for that day's bake. Connecting the dots, the next day they presumably used the previous day's sponge to make that day's starter and repeated the cycle.

In my efforts to replicate Larraburu (or any of the other S.F. sourdoughs of the time), I've found the flavor is the easiest part to get right, so I'm trying to understand why you had trouble with it. I remember well what Larraburu, Parisian, Colombo, etc.tasted like. When I moved to L.A. in 1979 we had an excellent sourdough made in Venice, California, which compared very favorably to the S.F. sourdoughs. It has since changed and the flavor is all gone. I had some Boudin back in December and it was awful. As you know, Boudin was never a major contender back in the day.

Table I

Sour Dough Starter Sponge
100 parts previous sponge
100 parts flour (Hi-gluten)
46-52 parts water

Make up and hold 7-8 hrs. at 80°F
Starting pH = 4.4 to 4.5
Final pH = 3.8 to 3.9

Table II

Sour Bread Dough Formulation
20 parts starter sponge (11% of final mix)
100 parts flour (regular patent)
60 parts water
2 parts salt
Make up – approx. 1 hr floor time – then proof 8 hr. 86°F
Starting pH = 5.2 to 5.3
Final pH = 3.9 to 4.0

Preparation and Handling
The bread dough, as shown in Table II, is made up simply with the fully developed starter sponge, flour, water and salt. None of the other usual ingredients of white pan bread such as yeast, sugar, shortening, non-fat dry milk, monoglycerides, dough conditioners, oxidants, mold inhibitors, etc. are needed or used. The starter sponge is used at a level of approximately 9 to 15 percent of the final bread dough which , after make-up is allowed to relax for at least 30 minutes. Then it is scaled (divided), rounded and given an overhead proof of about 20 min at 90°F, after which it is molded, placed on canvas dusted with rice flour and/or corn meal and allowed to proof about six to eight hours at 85 to 90°F. This long proof time may be reduced somewhat by increasing the proportion of starter sponge or by lengthening the floor time before molding, but is generally essential for development of the acidity and the coarse grain typical of this bread as well as for volume.

The pH of the bread dough on make-up is about 5.3 and drops to about 3.9 when the long proof is completed, or roughly to the same point reached by the starter sponge itself.

Baking

Baking is carried out in a hearth (generally carborundum) oven for a relatively long time (45 to 55 minutes) at a relatively low temperature (375 – 390°F). It is quite essential to slash or make cuts on the surface of the fully-proofed dough just before it is placed in the oven, otherwise the crust character will be wrinkled and generally unsatisfactory and the eating quality of the crust is probably the most essential part of this bread. Also an absolute requisite to achieving the desired crust character is the use of a very wet oven, particularly the first part of the baking cycle and continued until the crust attains a light tan color. Generally this is achieved by saturating the oven with low pressure steam.

dmsnyder's picture
dmsnyder

For the 3rd time: The every 8 hour feeding is described in the Kline, et al. article. What you are quoting from doc.dough appears in the Galal article on the Larraburu process.

Quoting from the Kline article, The Baker's Digest, April, 1970, pg.48, 3rd column,1st paragraph, 5th line and following:

"In commercial practice, the starter sponge is rebuilt about every eight hours or at least two to three times a day, seven days a week. Presumably it has been carried in this fashion for 100 years, although we can only guess how it got started originally."

David

ananda's picture
ananda

Hi David, Chris,

If it helps for clarification, I worked for several years in a bakery which operated 3 bread production shifts of 8 hours every 24.   Each shift made a batch of Pain de Campagne using wheat levain.   In practice, this meant a refreshment cycle of every 8 hours for that levain held in stock.   Reading these most recent posts in the thread, I'm reflecting that the authors David cites were referring to a very similar production system to the one I am describing.

All good wishes

Andy

dmsnyder's picture
dmsnyder

Thanks for chiming in, Andy. In several of the baking books aimed at professionals, this levain feeding regimen is documented. I suspect it's used especially when a sweet pain au levain is desired.

I'm not sure why Chris repeatedly questioned the description in the Kline paper. I suspect he misunderstood Doc.Dough's mentioning it and then citing only the Larraburu regimen in the other paper. I hope it's clear to all now.

Regards,

David

 

 

chris319's picture
chris319

Andy, if you're still following this, do you recall whether each shift baked with a starter which had been refreshed 8 hours previously, or whether each bake used starter which had been refreshed every eight hours for some number of refreshments?

I'm going by doc.dough's two posts from July 2011. David, are you looking at something different that I'm not looking at? If so, can you post a link if such exists? Doc.dough clearly separates the Larraburu process from the "other bakeries" process. You keep claiming, erroneously, that I'm quoting the Larraburu process which, if you'll look again at doc.dough's post is clearly identified as the "other bakeries" process.

Doc.Dough's picture

The following is transcribed from the (hard to find) 1970 paper "Nature of the San Francisco Sour Dough French Bread Process, I. Mechanics of the Process" by Leo Kline, T. F. Sugihara, and Linda Bele McCready in Baker's Digest 44(2), p48-50. (Note the use of the two word form of sourdough)

Table I

Sour Dough Starter Sponge
100 parts previous sponge
100 parts flour (Hi-gluten)
46-52 parts water

Make up and hold 7-8 hrs. at 80°F

See the names above my excerpt? They are Kline, Sugihara and McCready. See the words "Baker's Digest" and the year 1970?

Here is doc.dough't post from July 7, 2011:

This quotation: In commercial practice, the sponge is rebuilt on the average of about every 8 hours, or at least 2 to 3 times a day, seven days a week. The procedure has been carried on in this fashion for at least 100 years, and the origin of the initial sponge is veiled in mystery.

Appears in the K & S PATENT... https://www.google.com/patents/US3734743?dq=kline+sugihara+sourdough+french+bread&hl=en&sa=X&ei=ccQXU4LzJNjtoAT9uoCgDg&ved=0CDMQ6AEwAA

There is this, also from the patent:

A principal defect of the conventional sponge system lies in its rigidity of scheduling. When the sponge has been developed, it must be used within a short time, i.e., within l to 3 hours after reaching the peak of development,

Best of luck in your efforts to replicate Larraburu.

 

dmsnyder's picture
dmsnyder

Doc.Dough transcribed Table I from the Kline article in a July 18, 2011 post. I see that now. He discussed the Larraburu process in an earlier post, on July 7, 201, as you said. I saved both posts in the same document file. I apologize for the confusion.

However, I took your primary question to be where I got the 3 times a day feeding schedule. The answer is, from that same Kline article, as documented in my preceding reply.

I have tried the methods described  in both the Kline article and the Galal article. In my kitchen, neither produced remarkably good San Francisco-style Sourdough. I have moved on and have no plans to return.

David

chris319's picture
chris319

David -

I'm sorry to hear you abandoned your efforts to replicate Larraburu. It isn't easy, that's for sure. I face the opposite of your dilemma: I've got the flavor nailed but the other aspects, such as rise and crumb texture, leave much to be disired.

I was in San Francisco in December. Nowadays they can't even make S.F. sourdough in S.F.! Except for one, maybe two bakeries, it's all vastly inferior to Larraburu and the other sourdoughs we remember from that era. Boudin was around in those days but was at the bottom of the heap. The bread they bake nowadays tastes like vinegar. Part of the problem is that so few people know what the genuine article tasted like so they have no frame of reference.

I'm still plugging away at it.

Best.

dmsnyder's picture
dmsnyder

If you've nailed the flavor, that's an accomplishment! 

When we visited SF in the '50's and '60's, the restaurants we ate in tended to serve Parisian rather than Larraburu. I may have had Larraburu, but I have no distinct recollection. I have gotten close to the old Parisian SF SD flavor and the crust texture, but never the crumb texture. I think some of the difficulty duplicating old time SF SD is that it cannot be done in a home oven.

Good luck with your quest!

David

chris319's picture
chris319

All of those breads were fairly similar in flavor, except Boudin.

A Perkins oven is mentioned in connection with Larraburu. It is apparently a huge thing for use in a huge bakery, nothing a home baker could come close to.