Suburban Melbourne yeast capture. The rancid smell of failure.

Wild yeast hunting doesn’t always strike gold. Sometimes you end up capturing something that faithfully recreates the odours of vomit, or sweat, or cheese, or maybe all of them at once. Even if it’s disgusting, I still find it immensely fun to be hit with a surprise nostril assault and find the words to describe it.

In the interests of publishing negative results and sharing my yeast capture method, here’s the grotesque outcome of a recent attempt to capture brewing microbes from a Eucalypt blossom.

9 April 2020
The yellow gum (Eucalyptus leucoxylon) in my Melbourne front yard is currently going bonkers. Every day its teeming with lorikeets, bees, and flies, which I assume must be moving nectar-inhabiting microbes around, spiking the very open nectaries with potentially useful fermenters.


I clipped off several branchlets with fresh-looking flowers, and brought them inside, careful to not handle or touch the flowers I was going to use for capture. In my improvised biological safety cabinet (stove-top, gas hob on low, extraction fan on), I picked flowers off with flame-sterilized forceps and dumped them into four sterile jars of media, each jar had two flowers in about 80mL. All incubation took place in a warm spot above the heat lamp in the python’s tank. The jars probably cycle through swings from 10 – 25 degrees C.


Wild capture media (1 litre)
86g pils DME for a target 1.030
25mL EtOH for 2.5% abv
1 drop of tetrahop
Adjust pH down from 5.5 to 4.5 with 88% lactic using narrow range pH papers


18 April 2020
After just 9 days, there’s already a terrific pellicle growing on a couple of the jars. I removed the flowers from all of the jars with sterilized forceps, decanted some of the liquid, and topped up with fresh capture media.

My notes on appearance for the four jars below.
#1: Pellicle forming. Cloudy.
#2: Slightly cloudy, no pellicle.
#3: Robust pellicle. Cloudy.
#4: Very cloudy, no pellicle.


01 May 2020
By now, I should have cultured up the bugs that were going to grow, and filtered out the ones that can’t survive in the presence of low pH/alcohol/alpha acids. I swirled each sample jar and decanted a small amount into a fresh and sanitized 250mL flask to inoculate fresh 1.037 LDME wort within.

05 May 2020
I think #4 has possibly caught a Saccharomyces infection. Froth and bubbles present, but were not in the initial capture.

11 June 2020
Time to smell, taste, and measure. I decanted a 100mL sample from each flask, then replaced the volume with more fresh 1.037 DME wort.

#1: Thick mat of white mould on the surface. Discarded.

#2: Gravity is 1.024. 34% apparent attenuation.
Appearance: Solution is bright. Nothing in suspension. No pellicle.
Aroma: Smells of Eucalyptus oil (most likely still hanging around in solution after coming off the flower), heavily phenolic, smokey, leather, solventy like acetone or ethyl acetate. Then there’s a faint background sick and cheesy vibe. Just faint. The whole sample is very pungent. Aroma is overpowering, and not something I want to taste.

#3: Gravity is 1.024. 34% apparent attenuation.
Appearance: Cloudy yellow. Robust white pellicle.
Aroma: Smells sharply and impressively of sour sweaty feet, with a dash of body odour, and slightly like vomit. Could be isovaleric acid. I don’t like foot smell and I am not tasting it.

#4: Gravity is 1.015, so this one has achieved a reasonable 59% attenuation, which fits with the bubbling I noted on 5th May.
Appearance: Cloudy yellow. Rocky white pellicle…. and then I tried to decant the sample and boy was it thick and ropey. It has become sticky and glutinous like Chinese takeaway sweet-corn soup.
Aroma: Smells sort of intermediate to #2 and #3. It has the light Eucalyptus and smoke presence of #2, along with the vomit and cheese of #3. Again, I’m not putting this in my mouth.

10 July 2020
A last chance at redemption. Have any of these turned a corner and presented anything remotely appetising to warrant continuing the experiment?

#2 is dropping clear and bright. It smells phenolic, but there’s some fruitiness in there too! A white-wine like fruitiness. There’s also an enteric slightly vomity backdrop there. I dared to taste this one, just a small sip, then spit, and it does taste quite fruity too. Nothing fantastic, and the wort is still very sweet indicating no further attenuation.

#3 is hazy. Stills smells strongly of sour foot sweat. It’s going down the drain.

#4 is still grotesquely goopey and smells like it did one month ago. Kill it with fire.

My first attempt to capture an estate culture from my own small patch of suburban Melbourne has failed. Yet the impressive bouquet of bodily odours produced by these microbes was a fun display of nature in action.

Recipe: Make your own paleo dog food

The raw food dog diet is based on the simple principle that raw meaty bones, organs, meat and vegetable scraps are what dogs have evolved to eat, and therefore what they are most likely to thrive on. In contrast to the ubiquitous mass-market dog food full of cheap carbs and offal, this is not a hard argument to accept. It’s basically the paleo diet for dogs.

As well as prey, wild dogs would typically eat vegetables and leaves from their environment, as well as the semi-digested vege-matter from the stomachs of their herbivore prey. Although she would probably survive in the wild, Pixel doesn’t have access to the stomach contents of Moose-carcasses, so we instead feed her a porridge substitute called “Vet’s All Natural” (or VAN for short). It’s a product developed by an entrepreneurial vet with good credentials in animal nutrition. When rehydrated, left to begin a natural ferment, then served with raw mince, VAN provides the nutrients and roughage required to round out the raw diet. Alternating this with raw meaty bones, and occasional organs, eggs, fruit and veg, completes Pixel’s diet.

Using the ingredients reported on the packet and website, and a series of sieves, I  reverse engineered VAN so that we could make it ourselves from scratch.

The back of the packet quotes the following ingredients:

Rolled oats, cracked barley, flax seed meal, whole cracked oats, carrots, split peas, calcium carbonate, parsley, kelp, lecithin, barley grass, vitC, dried garlic.

Assuming they are listed by conventional order of largest fraction to smallest, you can see that VAN is mostly grains (oats and barley), with some minor fractions for nutrition and taste.

The dishes photographed below show sieved fractions from 50g of VAN. I mixed the bag very well by shaking, before drawing this sample. For fractions containing multiple ingredients of similar size, I picked out the pieces with forceps (for example, all the carrot pieces). I then weighed the fractions to convert the ingredient to an overall percentage by weight. Due to the limit of measurement on my equipment, there are errors around these estimates and the totals do not add to 100%. But hey, this is nutrition, not chemistry.


From top to bottom: 1: split peas, 2: rolled oats, 3: cracked barley and oats, 4: carrot, 5: cracked barley and oat fragments, 6: fine grain/fibre/greens/carrot, 7: powder and meal

Screenshot 2020-04-08 20.45.38

Weights for VAN fractions pictured above

So VAN is mostly cheap grain—around 50-55% oats and barley by weight, divided between rolled oats, cracked oats, and cracked barley. Dehydrated peas and carrots make up close to 10% of the mix. Judging by the substantial meal fraction and flax seed meal’s appearance as 3rd in ingredients list, it composes something like 20 – 25%. Then the remaining 10 – 20% is formed by the dried greens and supplements.

Using this guide, I concocted my own raw dog food supplement porridge recipe:

Screen Shot 2020-08-07 at 12.29.32 pm

Recipe for 1 kg dry dog porridge. To be prepared in batches by mixing with equal volumes of room temperature water, and allowed to Lacto ferment for some hours.

One could just feed raw carrot from time to time instead of dried, or sub grain fractions for other grains if necessary. The important thing with these bulk ingredients is to get fiber, carbs, and protein in there. Oats have a different protein and fat content to barley, so I wouldn’t sub them wholesale for each other, however the form they take could be flexible.

Supplements. VAN contains VitC, Lecithin, and Calcium carbonate as supplements in the mix. The calcium amount is easy to dose given the information on the package (2800mg/kg), however the others are difficult to extrapolate from my weighing. The safest policy is therefore to feed foods rich in these elements as well. e.g. fish/eggs/organs/soy for lecithin, and fruit and veges for vitamin C. VAN needs to be nutritionally complete when mixed with mince, yet my own mix does not need to be as rigorous if we are sensible in supplementing Pixel’s diet with other foods.

Puppies. This is not designed for puppies. We trusted products nutritionally designed for growing dogs when Pixel was a pup.

Some tips on sourcing ingredients

  • Homebrew shops are an excellent and cheap source for grains like torrefied oats, and cheap calcium carbonate
  • Bulk dry goods in middle eastern, medditerranean, African supermarkets are good for cracked barley
  • Cheap stockfeed kelp here
  • Many other of the ingredients are available from bulk food/health food stores


Edit: I have updated the original recipe with less garlic after the original turned out to be rather pungent.

Wild yeasts are everywhere. Some of them will even make beer for you.

I spend a lot of time thinking about flowers and beer. Thinking about flowers is part of my job, and beer—that’s my current obsession. Thoughts collide, and I recently found myself dwelling upon what they have in common: that most marvellous microbe, yeast.

Yeast is that critical fungus that converts sugar solutions into beer and wine, and while we’ve got a handful of domesticated strains harnessed for beverage production, diverse and untamed wild yeasts are everywhere. They are in the air, on plants and animals, on your skin, in your hair. Wild yeasts are particularly abundant in flowers, and that’s because flowers provide a source of freely available sugar by way of nectar.


Brewers/baker’s yeast (Source: Wikimedia Commons)

Given this obvious overlap, a natural and totally irresistible work-hobby collaboration sprang to mind. Could I capture a wild yeast from flowers of my study species and use it to make beer?

Now I am not the first to try this. Wild yeasts and other microbes have had a long history of use in creating beer. Belgian brewers have perhaps the most celebrated and storied traditions in this area—their Lambic beers are created by leaving fresh, unfermented beer (wort) to be inoculated by whatever yeast and bacteria the atmosphere may gift them. Wild fermentation is a growing global trend now, with numerous craft breweries here in Australia (e.g. La Sirene and Wildflower), and internationally (e.g. Allagash, Trinity) establishing strong reputations for artisanal ales fermented with the help of local microbial biodiversity.

While the diversity of wild yeasts might be wide, not all are useful for producing beer however. Many yeasts die in the presence of moderate alcohol, many cannot ferment all but the simplest of sugars, many produce unpalatable by-products during fermentation.

So where are we most likely to find the best, most useful wild yeasts for beer production? This is where floral biology meets brewing.

Floral biology meets beer brewing

Nectars are produced by flowers as rewards for the service of pollinating animals. Because some flowers specialize in being pollinated by particular kinds of animals, they evolve specific traits that cater to the biology of those animals. For example, moth-pollinated flowers are white so that they are visible in low light, bee-pollinated flowers evolve UV-reflective runway markers to guide accurate landing and foraging, carrion-fly pollinated flowers smell like rotting flesh. In the same fashion, nectar is shaped by evolution to cater to the specific creatures most likely to consume it.

One way nectar becomes tailored to its consumer is by its sugar concentration, which varies wildly. At the concentrated end, exceeding 50% sugar by weight, nectar is very viscous and sticky and difficult to suck up through long or thin mouth parts. These nectars cater to insects with short tongues like bees, flies, wasps and beetles. On the other end you have dilute nectars, with 10 – 25% sugar concentrations, and these are perfect for birds to lap up. By a happy coincidence, the sugar concentrations of bird-adapted nectars are in the same range as unfermented wort. Recognizing this was what led me to try hunting for yeasts in the flowers of my study species—the bird-pollinated shrub Prostanthera walteri.


Monkey Mint-bush (Prostanthera walteri)

Also known as the Monkey Mint-bush, this is a rare shrub growing amongst boulders on a few misty, granite peaks in remote East Gippsland, Victoria. I have been getting to know the plant for a couple of years now, using it in a study to understand how bird-pollination might differ from insect-pollination. And so on a January field trip to collect some data, I took the opportunity to collect some fresh flowers and take them back to my home lab (kitchen bench) for bioprospecting. At home, I made up a test wort: a low concentration malt-extract solution to mimic the conditions of beer, then I syringed out the nectar from several flowers and spiked the test jars with whatever might be living in the nectar.

Screenshot 2019-05-14 14.50.19.png

Kitchen-bench inoculation of test wort

And it worked. Most of the test jars began fermentation, and sniffing the results revealed various aromas of bright apple juice, white wine, earth and smoke. After months of re-culturing these initial samples I now have what I think are two different strains* of nectar yeast, one of which just produced its first beer.

So how does it taste?

Interesting, and not bad… and that’s all I’m willing to venture at this stage of the experiment! The yeast fermented very quickly, and chewed through 79% of the available sugars (which is more than some domesticated brewing strains). It has a somewhat Belgian Saison-like character, with strong pear and floral esters, some smoke and spice, and a very slight tartness.


I was quite blown away at how well this yeast performed, fermenting much like a domesticated yeast, yet with a much bigger, bolder, dare I say “wild” flavour. It is stunning to think that it has probably existed in flowers in remote eastern Victoria for some thousands to many thousands of years, and one can just go and pick it up and persuade it to make interesting beer. And as I get to know it better, perhaps that beer will become both interesting and delicious.


Thanks to Ruth Barry (Boatrocker Brewery) for inspiring conversation and advice on this.

*These are technically mixed cultures, but I believe they each have come to be dominated by single strain of yeast.