Spherical Cows & Yeast Starters

A farmer wants to increase milk production, and asks his physicist neighbor for advice. The physicist does some calculations and says “I have a solution, but it only works for spherical cows in a vacuum”.

An old joke…

I was listening to Episode 160 of the Experimental Brewing Podcast, which was all about oxygen in brewing – the good and the bad. While good overall, near the end Denny made a statement so wild that I rewound and re-listened to the section to make sure I had heard it correctly. That claim was that stirring a starter does not introduce oxygen into the culture. This came as quite a shock, seeing as I and hundreds of thousands of microbiologists have been using stirring and shaking for over 90 years to oxygenate our cultures.

In the podcast Denny mentioned that there was an article on their webpage containing the evidence for this claim, so naturally I tracked down that article on the Experimental Brewing website. It made my inner microbiologist cry. There are a lot of myths in brewing. Most of them are harmless, but some do cause problems. And then there was this disaster of an article, which made numerous false and easily disproven claims – one, after another, after another. All based on, and I quote, the authors self proclaimed expertise of anyone who has studied physics and chemistry…hence the title of this article.

I’ve emailed Denny and Drew about this, and I hope they issue some sort of correction, but I think it’s worth going through their article on a point-by-point basis and comparing its claims to what the actual science says. It’s a good chance to learn a bit about yeast, the process of scientific discovery, and also to see how myths form and propagate in the brewing community.

Myth #1: Stirring starters does not introduce oxygen

From the original article: It is claimed that stir plates oxygenate yeast cultures.  However, anyone who has studied physics and chemistry knows that the shape of an Erlenmeyer flask does not lend itself to O₂ absorption.  The conical shape of an Erlenmeyer flask combined with its rapidly narrowing cone leads to a small specific surface area in which O₂ can be absorbed.  People claim that spinning the bar fast enough to create a vortex improves oxygenation.  To a point, that claim is true because it results in an increase in specific surface area...

Stirred cultures were invented by Albert Kluyver in 1933, who developed the method in order to grow oxygen-dependent fungi. In that original paper (Kluyver A, Perquin L (1933) Zur methodik der schimmel-stoffwechseluntersuchung. Biochem Z 266:68–81), published over 90 years ago, Albert measured the CO₂ and O₂ levels found in the flasks, and via those measurements determined that oxygen was indeed continually infused into the culture and consumed by the growing fungi. Since then, these measurements have been repeated in dozens of other studies, with investigations into every minute detail of how it works and how it can be improved. As two examples, this study from 1969 investigated how mixing speed and the type of cap used affects oxygenation. Their conclusions say it all: Flasks shaken at 230 to 385 rev/min gave sulfite oxidation rates of 1 to 8 mmoles of oxygen per liter per min over a useful working volume range (40 to 150 ml in 300-ml flasks). These rates are as high as those obtained in agitated fermentors under usual operating conditions.

This second example is more recent, and compares flask types and the use of baffles on oxygenation rate. As with the 1933 and 1969 studies, mixing the cultures provides high levels of oxygenation, and that the oxygen is rapidly consumed by the growing culture. The conclusions of this study isn’t really applicable to brewers as their main focus was on using flask types that are not readily available to the homebrew market and which require a rotary table* for optimal efficiency. But it is a modern study, using modern measurement tools, and confirms that rotary motion is sufficient to provide good oxygenation.

*In microbiology labs we tend to use rotary tables rather than stir plates as one table can oxygenate multiple flasks, while a stir plate is limited to one flask. In both cases you are aiming to produce a vortex to draw air into the flask and media – the only difference is whether that vortex is generated internally via a stirbar versus externally by moving the flask in a circular orbit. In general, a properly configured stirplate provides better oxygenation than a shaken cultures unless you use a flask with built-in baffles to increase splashing of the culture.

How does oxygenation occur?

The physics of how this oxygen exchange occurs in stirred and rotated cultures is also well understood. Experimental brewing’s erroneous claims were a product of two spherical-cow type oversimplifications. The first error was in their assumption of how oxygen is introduced. They assumed surface absorption. This is incorrect; when properly configured, the vortex in a stirred culture contacts (or gets near too) the stir bar, which is represented by the light-yellow region in the image to the left. The stir bar then forces some air, in the form of small bubbles, into solution (small circles in the image). Small bubbles are great for gas exchange, leading to efficient oxygenation and extraction of CO₂.

The second error was that they forgot that friction exist. There is friction between the air and liquid, meaning that the air in the flask also rotates, and is not simply a static layer of air. This means that there is a moving vortex of air within the flask. This is not a passive mass of air either: the air forced into solution by the stirbar needs to be replaced, meaning that there is a continued flow of air down the vortex to replace that air (blue arrow in image). Meanwhile, the air forced into the culture bubbles upwards, eventually hitting the inward curvature of the flask (black dotted lines in image). This inward curvature concentrates the raising bubbles along the wall of the flask. When this reaches the surface, this produces a sheath of upward moving air (green arrows) surrounding the inner cone of downward moving air (blue arrow). This creates a lot of turbulence, and so long as you cap is not too tight (or you’ve used a permeable cap), that turbulence drives a significant exchange of air with the outside environment (circular arrow at the top).

In other words, this is a spherical cow error. The author made some simplistic assumptions about fluid and air flow in flasks, didn’t bother to check if there was any experimental evidence on how things work in the real world, and a myth was born.

Myth #2: Stirring damages yeast cells

From the origonal article: [Stirring] comes at a cost to yeast cell wall health due to shear stress caused by the spinning bar (i.e., the spinning bar is a source of friction for the yeast cells in a starter).  Shear stress is something that has been well studied when it comes to the production of dry yeast.

This is something I’ve seen more and more in the brewing community, and the frustrating thing about this claim is that it arises purely from a misunderstanding of some really good work done by Fermentis. In 2019 Fermentis published some work on how to best rehydrate dried yeast. One of their findings is that stirring during rehydration damages the yeast and lowers viability. This is unsurprising – yeast cell strength is largely a product of turgor pressure; a fancy way of saying that yeast cells are pressurized. Because the yeast cell is surrounded by a cell wall, this outwards pressure tensions the cell membrane against the cell wall, and also tensions the cell wall. This provides a great deal strength to the cell (much like how stretching a cloth between your hands makes it more rigid than the cloth hanging loose). This is not an insignificant amount of pressure either, and can be up to 2 MPa of pressure – about 20 atmospheres, or 290 PSI of turgor pressure. This creates an incredibly rigid and tough cell that allows the yeast to resist enormous shear forces; far beyond anything you could generate in a stirred culture. In fact, yeast are more resistant to shear forces than are our own tissues – if you can stick your finger into your starter without the skin being torn off, than the shear forces are tolerable to yeast.

So why are dried yeast different? The answer has to do with that turgor pressure and what happens to the cell wall when you dehydrate yeast. Keep in mind that this pressure and the resulting strength is maintained by the plasma membrane of the yeast pressing against the cell wall. The cell wall itself is largely made of a beta-glucan hydrogel; essentially a meshwork of interlocked sugars that is “inflated” by water. This forms a substance similar in character to a hard rubber like that of a car tire – tough, but with some give. The water is key to that character, and when dehydrated, that meshwork collapses somewhat, distorting the cell wall and disrupting its structure. When dehydrated, the plasma membrane also pulls away from the cell wall, resulting in a loss of that stabilizing turgor pressure. To make matters worse, Pir proteins on the surface of the cell wall are critical for maintaining cell wall structure during rehydration – but their connections to the cell wall are weakened during dehydration, meaning that minor mechanical forces – including those of prilling and packaging dried yeast – may remove these from the cell. Meaning that as dried yeast “reinflate”, they have a cell wall that has been distorted and fractured by drying, and which may be further damaged due to handling by the yeast processor. There is no turgor pressure to stabilize this wall, and as such, as the yeast rehydrate their cell walls can be easily destroyed by even minor mechanical forces. But once the yeast has had time to inflate and repair their cell walls (a few hours at most), they are every bit as resistant to shear forces as are never-dried yeast.

Myth #3: Forcing yeast into suspension is bad…somehow

From the origonal article: Many amateur brewers use the argument that stir plates keep yeast cells in suspension.  This claim is true, but the cells that a stir plate keeps in suspension are non-viable and cells prone to early flocculation, neither of which are desirable. The counter to this argument is that brewing yeast cells do not need to be stirred to remain in suspension because they express the NewFlo phenotype.

This is nonsensical and doesn’t even have a consistent internal logic. Firstly, stirring cannot selectively keep non-viable cells in suspension. All cells will be put into suspension. Secondly, I could not find any literature suggesting that suspending dead cells in a culture in any way affects the growth of the viable cells (How could they? They’re dead and not absorbing nutrients). In fact, at least with bacteria, stirring in the absence of aeration improves both yield and viability – possibly due to active mixing allowing for quicker nutrient acquisition and waste removal than allowed by passive diffusion.

I’m actually preparing a series of experiments on this topic using my new toy – look for those in a few months. But the peek behind the curtain is that stirring greatly increases the viability and vitality of cultures (measured microscopically via trypan blue and iodine staining) compared to static cultures.

As for the flo stuff, they author has mis-understood how flocculation and flo genes work. They are referring to this article from the 1990s, which identified two flocculation phenotypes in yeast. The genetics and mechanics of these flo genes are now well understood…but would take a 5000-word article to explain, so we’ll leave that aside for now. In simple terms, yeast have a bunch of flocculation (FLO) genes. Yeast will increase the expression of these genes as fermentation progresses. When expressed, FLO genes bind to the sugars found in the cell walls of neighbouring yeast, thus cross-linking the yeast into larger clumps, which accelerates sedimentation. But those clumps are not required for flocculation – yeast cells have a density of 1.113 g/ml (for comparison, water has a density of 1.000 g/ml), meaning that gravity alone will cause yeast to sediment in any wort with a gravity below 1.113 (26.5 plato) – active biology is not required. The clumping of yeast by FLO genes speeds that process, but is not required. The reason fermenting yeast float has little to do with FLO genes – rather, yeast cells stick to the bubbles of CO₂ formed during fermentation, and the upward movement of the bubbles carries the yeast to the top of the fermenter.

Myth #4: Stirring starters won’t increase cell count

From the original article: The reality is that all propagation mediums have a maximum cell density that places an upper bound on the number of viable cells that can be produced given a specific volume.

This is half reality, half nonsense. There is a finite amount of growth you can get from a particular medium – assuming all other factors are taken care of, this limit is most often determined by the amount of yeast assimilable nitrogen (free nitrogen) in the medium. That assumes that all other macronutrient (e.g. sugar, oxygen) and micronutrient (e.g. calcium, zinc, iron) needs are met. This too will be covered in more depth in my upcoming experimental series – look for those in a few months.

But back to the science. Oxygen is a limiting factor for yeast yield, as oxygen is necessary for the production of sterols (12 oxygen molecules (24 atoms total) are needed to make 1 sterol) and unsaturated lipids (1 oxygen molecule is needed per lipid). A lack of sterols or unsaturated lipids caused by insufficient oxygen will stop yeast growth, regardless of the amount of protein and sugars available. Yeast also produce what is called quorum factors, which yeast secrete in order to provide a “measure” of the density of cells in the growth medium. Once these quorum factors reach a certain concentration, growth ceases and the cells begin to prepare for dormancy. Oxygen slows quorum sensing molecule production, allowing for higher cell densities to be achieved.

Other factors enter into play, but I’ve written about these previously, so I’m not going to re-hash it here. But long story made short, proper oxygenation is required to maximize the amount of yeast you can get from a particular starter. Not to give too much of my upcoming series away, but I see 1.5 to 2 times more yeast production with stirring than I do in static cultures, across a range of yeast strains and medium compositions.

Myth #5: Stirring causes stress that damages the health of yeast

From the original article: There is not a more troubling claim in amateur brewing than stir plates produce healthier cultures. Not only is this claim wrong, the practice of allowing a starter to enter quiescence and settle out, so foul smelling medium that is the result of stressed yeast cells can be decanted misses the point…Allowing a culture to enter quiescence extends lag time because the cells have to reverse the morphological changes they underwent before they can go about replenishing their ergosterol and UFA reserves.

Again, this is total nonsense that has no basis in reality. Firstly, you can pitch a starter at high kraussen, which would avoid these issues…if these were actually issues. But even if you let your starter settle, decant, and pitch that yeast, you’re still fine. Yeast are not idiots; they simply don’t run at top speed until the food is gone and then immediately starve. They use a mixture of the aforementioned quorum sensing factors and direct detection of nutrients to determine both the current nutrient availability in the environment and the level of competition for those resources. When the yeast “foresee” upcoming nutrient limitations they undergo a transition from log phase (when yeast are actively dividing) to stationary phase (when growth ceases and the yeast flocculate). During this transition yeast undergo a number of changes: they build up stores of trehalose (which aids in resistance to alcohol and osmotic stress), build up stores of glycogen (backup energy), and if oxygen is available, they also synthesize additional stores of unsaturated lipids and sterols. Yes, yeast plan for the future! And that plan is essentially “save everything I need to be ready to ferment when conditions get better”. In other words, contrary to Experimental Brewing’s claims, yeast that have recently sediment led out of a starter are in prime condition for fermentation. In fact, that trehalose makes the yeast better at dealing with osmotic stress – like that experienced when being pitched into fresh (e.g. high sugar content) wort. Glycogen stores can help the yeast survive a stressful fermentation such as a high gravity or low nutrient (wine, mead) ferment. Sterols and unsaturated lipids allow for a greater number of cell divisions, reducing the risk of stalled fermentation and some off-flavours.

There is a minor cost to this – the yeast will have a longer lag phase than yeast pitched at high kraussen, although yeast in this state typically have a lag phase of only 3 to 4 hours.

Now, if you leave your yeast in the flask for weeks or months things change as the yeast deplete those stores. But for relatively fresh yeast, yeast that have been allowed to sediment will have more of these valuable stores than yeast pitched at high kraussen, which in some cases may be a benefit to the brewer. Those stores come at the minor ‘cost’ of a slightly longer lag phase.

Stirring Spherical Cows

Congratulations for making it this far. Truly, Brandolini was correct: The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it.

I hope that the message that “simple assumptions lead to incorrect answers” was clear, as was the message that there is over a century of research into yeast biology out there that you can check before making silly and easily refuted claims.

Buried in there are also some suggestions for maximizing what you can get out of a starter. But to condense that into a easier-to-digest set of suggestions:

1. Stirring will produce more yeast than an unstirred culture, even without oxygen ingress, but you can maximize growth by ensuring that:
   • Your flask is not overfilled and the vortex reaches your stir bar (or at a minimum, stirring is vigorous enough that you see bubble detaching from the central vortex),
   • Your flask is loosely capped or has a foam topper, so that air exchange can occur.
   • You avoid spinning your stir plate at 10,000 RPM (I jest, but if it looks and sounds like an egg-beater, you’re probably going too fast).
2. Lots of factors can affect the amount of yeast a starter will produce, but so long as there is adequate sugar, oxygen, and micronutrients, it is the availability of amino acids (protein) that will determine your yield. Missing any of those (including oxygen) will reduce yield and potentially may reduce vitality as well.
3. Yeast are tough little buggers…so long as they haven’t been recently rehydrated.
4. Yeast are adaptive organisms that respond “intelligently” and with “foresight” to changes in their environment. Harnessed correctly, you can use those features to produce high-viability, high-vitality yeast that will contribute positively to your brews.