You know what they say... Life is a journey. But have you ever been pulled down a path that you otherwise would have walked on by? That's what happened to me when I started playing with sourdough. I didn't even like sourdough, or so I thought until about seven and a half years ago. I was watching the food network one day, and Daniel Leader appeared as a guest on Cooking Live. He was demonstrating how to make a sourdough starter from nothing but flour and water. How fascinating! I had no idea you could do that. Bread science was nowhere in the curriculum when I went to university almost 30 years ago, let alone sourdough. But for someone with a microbiology degree and a passion for baking, sourdough is the perfect marriage of two loves. I had to try it for myself. And so began the journey. The path so far has taken some surprising twists and turns, and led to meeting many interesting people along the way. Including Peter Reinhart, who has been nothing but gracious and supportive throughout, and who is ultimately responsible for getting me to sit down and record this story in my own words. Now, back to the beginning...
In early 2002, I was a member of the King Arthur Baking Circle, a message board for baking enthusiasts. With this brand new interest in sourdough, I found myself reading all of the threads under that category. In March another member, Pat Doucette began posting of the difficulties she was having in getting a starter going. She had tried a few different recipes with no success. Now she was following the formula in The Bread Baker's Apprentice and... still no luck. Newly armed with all of the advice that she was getting from others on the forum, she started over. And once again, she got results that were nothing like the book describes. But a pattern was becoming clear. On the second day, her seed cultures would fill with bubbles and expand to over three times the starting volume when minimal growth was expected. And then do nothing on the third and fourth days when they were supposed to be expanding more and more. They came on strong and then died at the same point each time. A pattern holds a clue, so I offered to do the procedure myself and see if I could reproduce what she was seeing. I followed the directions to the letter and, lo and behold, my results duplicated Pat's perfectly. I may be the only person on the planet who would be excited about this, but it gave me something to study and troubleshoot. There was something unexpected going on at the microbial level. Living things are funny that way, and microorganisms don't always follow directions.
One by one, other people on the message board began to speak up and post that they had experienced the same thing. In fact, it seemed that many more saw that scenario than the one described in the book. This phenomenon had nothing to do with local strains of lactobacilli and yeast as some had surmised, because Pat was making starter in Massachusetts and I was in Missouri. Others chiming in represented various regions of the country from coast to coast. This pattern is apparently quite common. We ran the gamut of theories on why yeast were coming on like gangbusters, only to quit and become non-responsive. We tested each theory by trying different flours at various points, increasing the feeding frequency, changing the hydration and water source, cooling it down, and anything else that anyone thought might help. But in the end, nothing fixed the problem, and the results weren't making much sense to me.
At that point, I had to do what microbiologists do when things don't add up---go back to the microscope and take a look. That meant packing up my starters, taking them to work, and having to answer all the curious questions about what I was doing and why. But the microscope answered a few of my questions, and that day proved to be the turning point. No wonder things didn't make any sense! We were operating on the assumption that we were growing yeast. What I found was that there were no yeast or lactobacilli to be seen anywhere in all the activity of day two. Not a single one. But it was like a three-ring circus in there---different kinds of bacteria, some round, some rod-shaped, some motile, some not. Some were spinning, some were twirling, some flipping or zigzagging, and some were just darting back and forth across the field. What were these bacteria, and which one was responsible for all the gas?
I knew, from having made so many starters by now, that this pattern does turn into sourdough if given more time. So, I looked at cultures each day in the process, comparing them to my established starter which was yeasty and stable. Everything quiets down in there and yeast emerges a few to several days later. They don't appear to be coming from the air as many people believe, because it happens even in a covered container. But if they are already in the flour as the more reliable sources say, then why couldn't I find any? Obviously, there was more to this than just a symbiotic relationship between lactobacilli and yeast gradually increasing in number, good guys out-competing bad---the usual explanation. It was evident that there are many more bacterial and fungal species present in flour than just sourdough lactobacilli and yeast. But where were the good guys? Why weren't they growing? It was time to close the cookbooks and open the textbooks.
I turned to a large, newly updated food microbiology tome, and was disappointed to find only two brief paragraphs on sourdough, and not much more on yeasted breads. So it became a challenge to find the information, mostly borrowed from chapters on wine, beer, dairy, and other food fermentations that share something in common with sourdough. I was able to narrow down the gas producer to a Leuconostoc species. The tip-off was reading that almost 90% of spoiled doughs are caused by Leuconostoc mesenteroides or Leuconostoc dextranicum. When I started searching for more on the genus, I found Leuconostoc mesenteroides is considered the primary agent in the fermentation of an Indian steamed bread called idli. (Spoilage is a subjective thing from culture to culture.) After soaking grains for a day and then grinding them with water into a paste, there is a 15-24 hour fermentation during which the idli batter increases in volume by about one and one half to three times---the same as our wild day two growth.
Leuconostocs are also occasional spoilage bacteria in wine making, "but they undergo little or no growth during the alcoholic fermentation and tend to die off because of competition from yeasts. Nevertheless, these bacteria are capable of abundant growth in the juice and, if yeast growth is delayed, they could grow and spoil the juice or cause stuck alcoholic fermentation." Many microorganisms produce characteristic aroma compounds, and so smell is also an important clue. I had previously described an unamended, all-white seed culture as smelling like sour milk with a hint of rotten cheese. Then I learned that some leuconostocs are added to dairy fermentations (such as cultured buttermilk, and cheeses like Gouda, Edam, blue cheese and havarti) for their carbon dioxide and aroma compounds. Together, these pieces all fit what we were seeing, and according to the chapter on fermented vegetables, leuconostocs are quite common in nature and found routinely on all kinds of produce and plant material. So, we can expect them to be present on grains and in flour.
Knowing how bland a flour-water mixture starts out, and seeing how the microscopic picture becomes more subdued as the sourness increases, it was apparent that the shift in populations and activity are tied to changes in acidity. pH is a fundamental factor in microbial growth. Some like it neutral while others need more acidity or alkalinity, but each species has its own pH range. The reason that the starters had become quiet on day three was because the pH had fallen and the gas-producing bacteria were no longer growing. Even though I still wasn't sure what these bacteria were, it was clear that whenever the gas-producing one or ones grew, the starter would subsequently become still and take longer to finish---sometimes by several days. I reasoned that the best solution might simply be to keep them from growing. And since they stop growing as the pH drops, why not add an acidic ingredient to the mixture to lower the pH and inhibit them from the outset?
It was May now and Evan Shack had entered the picture. Unaware that this was already a hot topic, he began posting to the message board seeking help after having just tried to make starter and getting the same result that we had. Evan was interested in learning the science behind it, and he and Pat were both eager to get to the bottom of the problem, so they volunteered to do some testing. Soon after we joined forces, Gary Wray contacted me and we invited him to join our little task team. With so many different recipes to choose from, it was clear that there are several approaches to making starter. But we needed to pick a direction to focus our problem-solving efforts. And because so many people on the message board were loyal fans of The Bread Baker's Apprentice, the group decided the goal would be to use that formula, altering it as little as possible, and make it proceed as described in the book. The fix should be simple, with ingredients readily available at home or in the average grocery store. Our choices for the acids were ascorbic (vitamin C), citric (sour salt), tartaric (cream of tartar), acetic (vinegar), lactic (yogurt), and mixed acids (fruit juices).
For our first trial we chose ascorbic acid, because it is readily available in the vitamin supplement section, known to be beneficial, and widely accepted in bread-making. Pat and I used vitamin C tablets that we had on hand. We crushed them and mixed the powder with the flour and water on day one. And much to our amazement... it worked! No gassy bacteria, and we were both growing yeast on or before day four, where it had been taking about seven days. But I discovered a little problem with supplement pills, which is that some are buffered without being labeled as such. I was not getting the pH to drop in mine even though I kept adding more and more vitamin C. When I took a closer look at the bottle, I found two ingredients listed which together, formed a buffer system that was keeping me from reaching the pH I was aiming for. Pat's vitamin C was not buffered and her starter took off in only three days.
Buffer problems aside, neither one of us enjoyed the task of crushing pills. And whirring them in a blender with the water only worked so-so. We also had no idea what the best dose would be. Gary and I both had ascorbic acid powder, so we did another experiment testing different doses ranging from 1/8 to over 1 teaspoon mixed with the 4.25 ounces of flour on day one. It was a fun experiment to do. With the jars lined up next to one another, they looked like perfect stair steps as the starters began to rise. It was easy to see which doses were most effective by how fast and how high the cultures rose. For me, the most active jars were the ones with 1/4 and 1/2 teaspoon of ascorbic acid powder. For Gary, the best results came from 1/2 and 3/4 teaspoon, and so we settled on 1/2 teaspoon as the recommended dose. While the ascorbic acid worked quite well, and may be the ingredient of choice for purists or professionals, the average person must go a little out of their way to find or mail-order it. So we decided to press on.
All of the acids that we tried, inhibited the gassy bacteria effectively, but sour salt (sometimes found with canning supplies) was so strong that it was hard to measure the tiny amounts accurately. Cream of tartar (found in the spice section) was too weak, and required an impractical amount to effectively lower the pH. We dismissed lactic acid because we didn't want to deal with dairy or go to the trouble of draining yogurt for the whey. And vinegar was so highly inhibitory to yeast in the doses required to lower the pH, that it was no solution at all. That left fruit juices. I tested the pH of various juices and made a list for the group to try---apple cider, orange, lemon, grapefruit and pineapple juices seemed like the most suitable candidates based on wide availability. But whenever trying a new juice or acid, I had the group run a negative control alongside---a duplicate to the test in every way, except using plain water. This would show whether changes in the result were due to the ingredients under evaluation, or to chance or variation in experimental conditions. Time after time though, the control jars followed the familiar pattern, while the test jars proceeded by the book.
While the trials were under way I went back to basics, monitoring the changes in acidity and examining seed cultures under the microscope every day. I recorded pH readings, growth measurements and observations at the beginning and end of each 24-hour feeding cycle. After a number of runs, I gathered my notes to compare and look for patterns. (My pH paper was only sensitive to the nearest 0.5 increment, so readings are approximate.) I found that when I acidified the day one mix to 4.5, it stayed at 4.5 until I fed it again on day two. If I didn't add more acid at that time, the freshly fed starter would read 5 and the gassy bacteria grew on day two and followed the oh-so-familiar pattern. If I acidified the day one mix to 4, it stayed at 4 until I fed it on day two, after which it read 4.5. The gassy bacteria did not grow and the culture started producing its own acid as other lactic acid bacteria were increasing in activity. During the second 24 hours, the pH dropped to 3.5 and the starter tasted really sour. Yeast usually appeared the day after. When I acidified the day one mix to 3.5, I actually got some yeast growth on day two. I'm not sure that this is the best way to go, though. I've only done it once with citric acid and yeast were not as vigorous the next day as I had hoped to see them. More testing could be done. But the key points here are that the gassy bacteria grew at or above pH 5, not at or below 4.5, and the cultures I was growing all failed to produce acid of their own in the first 24 hours. That is important because a day one flour-water paste measures about 6---quite inviting to leuconostocs. And even more importantly, in all my trials I have never seen yeast before a starter gets sour, but it usually follows very soon after.
I was hoping orange juice would perform well, since it is a good source of Vitamin C and a staple in many homes. But, it turned out not to be acidic enough to meet the group's objective, which was to use it only on the first day. However, Orange juice and apple cider do work well if they are used in place of the water for two or three days. Pat was the first to try pineapple juice, which has a lower pH than most other juices, and just happens to come in handy 6-oz cans (exactly the right measure for day one). She liked it so well that she stopped testing anything else and started recommending it to others. Almost everyone who tried it was thrilled with the results, and so pineapple juice became the solution that stuck. While the group's mission was accomplished, the story doesn't end here. But the rest will have to wait until next time, so please stay tuned...
1. Doyle, Michael P., Larry R. Beuchat, and Thomas J. Montville. 2001. Food Microbiology Fundamentals and Frontiers, 2nd ed. American Society for Microbiology Press, Wahington, DC.
This article was first published in Bread Lines, a publication of The Bread Bakers Guild of America.
Vol. 16, Issue 1, March 2008
The Pineapple Juice Solution, Part 2 | The Fresh Loaf