A traditional sourdough bread is made without commercial baker's yeast. Instead, the dough rises through the activity of a community of wild microorganisms - principally yeasts of the genus Saccharomyces and Candida, together with lactic acid bacteria of the genus Lactobacillus - that have colonised the flour and water mixture maintained by the baker as a starter culture. The starter is kept alive indefinitely by regular feeding with fresh flour and water, and some bakeries maintain cultures that can be traced back many decades or, in a few cases, more than a century. The microbial community inside the starter is a genuine ecosystem whose dynamics have become the subject of a surprisingly active scientific literature.

The first microbiological studies of sourdough starters were conducted in the early twentieth century, but systematic understanding of their microbial composition has come much more recently with the availability of DNA sequencing techniques that can identify organisms without needing to culture them first. The Italian microbiologist Lucia Bertolini has coordinated an international survey of more than five hundred traditional starters from bakeries in Europe and North America, and her findings have been widely cited. Most mature starters contain between two and five species of yeast and three to fifteen species of lactic acid bacteria, with the specific combination varying between starters but the general composition showing considerable consistency.

The microbial dynamics are driven by several factors. The flour itself contributes a starting community of microorganisms, including many species that cannot survive under the specific conditions of a maintained starter. Over successive feedings, the community shifts toward species that can thrive under the combination of conditions the starter provides: a pH that falls rapidly after feeding as the lactic acid bacteria produce organic acids, a rich supply of sugars released from the flour by enzymes, and a temperature in the range typical of baking environments. Within a few weeks, a new starter typically settles into a stable microbial composition that changes only slowly thereafter.

What makes each starter different is harder to pin down than what makes all starters similar. Bertolini's survey showed that even starters maintained with essentially identical practices - same flour, same feeding schedule, same storage temperature - can develop different species compositions depending on the particular organisms that happened to colonise them early. The role of the flour is particularly interesting: different grains, different milling practices, and different geographic sources of flour produce different starting communities, and some of these differences persist into the mature starter. Ancient grain varieties, such as the heritage wheats used by some artisan bakeries, often produce starters with distinctive microbial communities. The romantic claim that a starter captures the specific microbial flora of its home location has some basis in fact, though it is more complicated than popular accounts usually suggest.

The flavour of sourdough bread depends on the microbial community in several ways. Lactic acid bacteria produce the organic acids - mainly lactic and acetic acid - that give sourdough its characteristic tangy taste, and the balance between these acids produces noticeably different flavours. Yeasts contribute alcohols and esters that add complexity and fruitiness. Enzymatic activity releases amino acids from the flour that, during baking, participate in the Maillard reactions responsible for the browning and crust development. The combination of these effects produces a bread significantly more complex than one made with commercial yeast alone, and the specific flavour profile of any given sourdough reflects the specific microbial community in the starter from which it was made.

Health properties have been studied, though with the caution that applies to all claims about fermented foods. Long fermentation with lactic acid bacteria partly digests gluten and reduces the concentrations of phytic acid and other compounds that can interfere with mineral absorption. The glycaemic index of sourdough bread is typically lower than that of equivalent non-sourdough bread, which may be beneficial for metabolic health. A fraction of people with non-celiac gluten sensitivity report better tolerance of sourdough than of commercial bread, though studies have produced mixed results and the effect is not universal. Bertolini's laboratory has been careful to emphasise that sourdough should not be treated as a solution for medically diagnosed gluten intolerance, but she has argued that the combination of flavour improvement and modest health benefits justifies the attention the tradition has attracted.

The renewed popularity of sourdough bread has had practical consequences. Artisan bakeries have grown in many countries, and both commercial and home bakers maintain starters as a matter of course. The online exchange of starter cultures, either by post or through local clubs, has spread specific starters across wide geographic areas. Bertolini's survey included several starters that, according to their owners, had been sent across oceans and maintained in their new homes for years; she found that the microbial communities of such transported starters typically shifted toward compositions more similar to local starters, reflecting the arrival of locally adapted microorganisms from new flour and feedings. The geographic identity of a starter, on her account, is a genuine thing but not immutable; a starter transported to a new climate and fed with new flour will, over time, become a hybrid of its original and its new environment.

Looking forward, the scientific study of sourdough starters offers a convenient model system for research on microbial community dynamics more generally. A starter is small, well-defined, easy to sample, and runs through many generations per year. Bertolini has argued that similar approaches should be extended to other traditional fermented foods, including cheeses, wines, and vegetables, each of which contains its own microbial community whose dynamics have been insufficiently studied. The larger prize, she has suggested, is a general theory of how traditional food cultures maintain themselves despite continuous microbial turnover; and such a theory would have implications reaching well beyond the bakeries and dairies where its inputs come from.