I’ve talked at you all before about how easy it can be to do historical brewing research and recreation. We often attempt to take the principles of period processing methods and attempt to translate them into modern methodology, to give a sense of historical practice by varying the familiar.
We can also alter ingredient bills, to attempt to emulate the flavor profiles that may have existed at the time. This is all well and good, and it’s an important part of the process of experimental recreation.
Sometimes, though, the task is not so clear-cut, and attempting accurate recreation becomes a real challenge. How were the ingredients grown? What units of measurement were at play? Water quality? We can’t always answer all of these questions, but the attempt to do so can yield valuable information, and the process of extrapolating will teach us things whether or not we get a useful end-product.
So let’s talk about wood.
This bottle is part of an excavation of Trossingen grave 58, a find in Germany that dates to the 6th century CE. The picture above links to a gallery of the find.
This bottle is identified as a vessel with the remains of a hopped barley beer. This is sort of A Big Deal in the historic brewing world, because this would constitute the oldest existing physical evidence of the use of hops in a fermented beverage ever found. Not only that, but this is solid physical evidence of the use of hops a good 500 years before we had thought hops were really coming into use. This find has the power to really re-shape what we think of the history of brewing and hopped beverages. Neat stuff.
There is a publication which details the find (and its numerous artifacts) which you can obtain here; of course, the entire thing is available exclusively in German, so you may have to find a linguistically-inclined friend to help you out with it. Fortunately, I have some connections, and I managed to acquire the part of the journal detailing the bottle find. A bit of OCR, Google translate, dictionary consultation, and linguistically-inclined friend consultation, and I managed to figure out most of what the find was about.
Evidently, there was pollen residue in the bottle (~3500 grains), and researchers were able to identify the sources of the pollen grains:
Gut 17% davon stammen von Getreide, wobei der Gerste-Typ überwiegt. Getreideunkräuter machen zusammen fast 11% aus, Hopfen und die Weinrebe sind mit jeweils 0,4% vertreten. Mit gut 29% die größte und auch die artenreichste Gruppe sind Pflanzen…
If my translation is right, the contribution is 17% barley, 11% cereal weeds (possibly rye or oats?), 0.4% hops, 0.4% grapes, and 29% “bee pollen” (which is taken as a marker of honey). The bottle also contained evidence of fermentation (oxalate crystals), and so the author concludes that the beverage was probably a mixture of the above ingredients in the mentioned proportions, fermented together and hopped. The beer came first, and it was “enriched” with honey – or so the author concludes.
But I don’t like that analysis. For one thing, the author doesn’t seem to try to figure out the actual proportions of the plant matter represented by the pollen; the text seems to assume that all ingredients will convey the same amount of pollen, which may not be the case. They also don’t elaborate too much on their rationale for their experiments or on the type of hop present – which is too bad, because this is a pretty big find!
So let’s tear this down and show how you can extrapolate a recipe from scant information. What if you wanted to try recreating a beverage like this? No recipe, no method, just some pollen grains in a bottle – how can we do it?
Watch and learn.
When we do this kind of analysis, we often have to make lots and lots of assumptions and extrapolations. In archaeology, the variables are often well beyond our control – so experimental archaeology must try to control what it can or accept the limitations of uncontrolled variables. I’ve advocated a sort of “mapping” approach to redacting and analyzing ancient recipes, and that principle will aid us here as well; by listing out my assumptions and reasoning, I can go back and nitpick and refine and strengthen my arguments.
The goal here is to get to something that resembles a more accurate technique, and in the process to enumerate some other possible and plausible methods. Most of the time, these sorts of analyses are rarely definitive, and tend to leave us with more questions than when we started – but it helps us to focus our inquiries, so that our questioning can be more productive. This is the heart of science.
Let us assume:
1) That a total of 28% of the 3500 pollen grains are attributable directly to barley which has been malted (that would be 17% attributed mostly to barley and 11% attributed to “cereal” weeds – we know that barley is not generally insect-pollinated, so the “bee pollen” probably does not cross with this group);
2) That 29% of the pollen grains are attributable to raw honey (bee pollen shows up often in raw honey);
3) That 0.4% of the pollen grains are attributable to Hallertau hops (they’re alleged to be the first hops that were ever domesticated, and the Trossingen area was close-ish to Hallertau);
4) That 0.4% of the pollen grains are attributable to grapes (though as you will see shortly, I haven’t rolled grapes into my analysis yet because I can’t find information about them);
5) That the ingredients were fermented together in a single beverage (as opposed to the pollen contribution coming from, say, 3 different beverages which all touched the bottle at some point);
6) That a single kernel of barley (which contains three anthers) will produce ~4500 pollen grains, about half of which can be removed relatively freely – so ~2250 pollen grains will survive through malting and will make it into the final beverage;
7) That a single kernel of dry barley weighs one grain (0.06 grams – the origin of the term “grain” is the weight of one kernel of barley), and that malted barley is ~10% less dense than unmalted barley;
8) That raw honey contains, on average, 6000 pollen grains per gram (based on estimates of average pollen load of “normal” New Zealand honey);
9) That hops used were wild, and thus grew at a ratio of 1:1 male:female plants (hops are a dioecious plant, and wild-type examples of such plants grow in a ratio pretty close to 1:1 – this indicates that the pollen load of a male plant reported represents a single female flower);
10) That hops pollinate in a manner similar to their nearest botanical relative, Cannabis (note that hops are a cannaboid) – which produces an average of 36,500 pollen grains per male flower;
11) That the mechanism of wind pollination results in ~95% of the pollen accumulating on the windward (i.e. exterior) surfaces of the plant, and that this pollen load would be removed in hop processing (i.e. the pollen that didn’t make it into the interior of the female flower just falls off);
12) That there are 100 wet hop flowers (we use the female flower of the hop in brewing) per 50 grams of hops, or 0.5 grams wet per hop flower (which translates to roughly 0.1 grams per dried flower);
13) And that these estimates actually apply to 6th century German plants.
So, basically, I’m making shit up. “Educated guesses” if you’re feeling generous – but I’m basically winging it in the absence of any more useful information.
One thing that we can definitely see by my analysis so far: it is a great mistake to assume that all of the ingredients going into a beverage would have the same pollen representation per gram.
Let’s look at my numbers. Each barley grain produces 2250 pollen grains, each gram of honey has 6000 pollen grains, and each hop flower has 1825 pollen grains (5% of 36.5k). Let’s convert these to a standard measure: pollen grains per gram of plant matter.
Barley: 37.5k pg/g
Honey: 6k pg/g
Hops: 3650 pg/g
Now, how about the proportional representation of pollen grains in the find? 3500 pollen grains total, so:
Barley: 28% = 980 pg
Honey: 29% = 1015 pg
Hops: 0.4% = 14 pg
And then we just do the math to figure out the possible mass of plant matter that delivered that pollen load!
Barley: 0.026 g
Honey: 0.17 g
Hops: 0.0038 g wet (1/5 as much dried)
That gives us a ratio of barley:honey:wet hops (by weight) of 26:170:3.8, or to make things easier: 7:45:1
So let’s turn this into amounts that make more sense, shall we? Let’s also not forget that malted barley weighs 10% less than “green” barley:
63 g malted barley (about 2 oz)
450 g honey (about 1 pound)
10 g wet hops (2 g dried)
The first thing I notice straight away – this ain’t a barley beer. Not by any stretch. The mass of barley is so small that it really seems much more like a flavoring or additive than anything else. The vast majority of sugar here is coming from the honey – enough that I’d really call this a “mead.”
Of course, as you will remember, the word “beor” (which is a root of “beer”) is glossed with “hydromel,” which refers to a honey-based strong beverage. So really, it’s not outside the realm of possibility that one could call a honey-based drink a “beer” in the ancient world – it seems to have fulfilled that role.
In fact, the amount of barley is so small that I really think about a starter biscuit more than I do an actual source of grain sugar. Remember how I’ve been hypothesizing about Viking-era “breads” really being used as yeast starters? This may be the sort of thing I’m looking at here. And remember how I’ve talked about those same breads really being grain/herb mixtures? And how that grain/herb mixture, once fermented, could be used as the basis for fermenting a strong drink?
Pliny specifically discusses the various methods of making “leaven,” and one method is to incorporate grape must into barley flour and make a biscuit. Grape must incorporated into such a “bread” as I’ve talked about previously could explain the grape pollen in the original find. The use of herbs in the bread may give us a clue as to how the hops came into play; perhaps grape must and hops were mixed into barley flour, and the resultant “cake” was used as a yeast starter to then ferment a honey/water solution.
We can make a wide number of recipes simply by varying the amount of water that goes into such a thing. Generally, “hydromel” was a 1:4 honey:water ratio. A pound of honey occupies a space of about 10 fluid ounces, so we’d need about 40 fluid ounces of water to properly dilute that honey. Do that, add in your 65 grams of barley/dried hop mix (which has been previously fermented), and wait a bit. Yeast from the grapes eat those sugars, and you get a little more than a quart (about 1.5) of slightly hopped mead.
How hopped? Well, 2 dried grams of hops at that density of sugar yields ~12 IBU – roughly the same bittering content of Budweiser. For reference, an English Ordinary bitter is somewhere in the 25 – 35 IBU range. American pale ales are in the 50′s, and IPAs are up in the 70′s or more.
You could even add a bit more water – maybe go to half a gallon of final volume (1:5 ratio) with all that honey, which would give you a lighter-bodied beer with only 8 IBU. A little less sweet, a little less hoppy. The evidence still supports such an idea. Hell, it supports a lot of ideas.
Or you could go heavier (1:3 ratio) and make something really sweet with about 16 IBU. It’s all up to you and what you prefer!
Therefore, based on my analysis of the evidence, I conclude that the Trossingen bottle may have contained the remnants of a lightly hopped mead, which may have been fermented using the residue of a light grain fermentation.
Possible OG (Original Gravity) Range: 1.059 – 1.120
Possible bitterness (IBU) Range: 8 – 16
Possible volumes (quarts) Range: 1 – 2
The lesson here: archaeological evidence always requires interpretation. Using the same set of facts, we can come up with very different conclusions simply by varying the manner of our interpretation and the set of assumptions used to perform an analysis.
This is far from a definitive answer. I have thirteen listed assumptions, any variation on any of which can completely alter my outcome. I have no idea how much water was added, or how long it was fermented, or what proportion the grapes represent. We could re-analyze the model with an attempt to figure out what “cereal weeds” means and re-evaluate the contribution of plant matter from those (here’s a hint: rye produces ~10x the pollen that barley does – so there may be even less grain in this recipe than I’ve indicated).
But at least for now, I have something to work with – and that’s how science works.