Thursday, October 29, 2015

Gueuzerie Tilquin Visits

This post (which is a bit of a long one) about Gueuzerie Tilquin is based on what I've learned from a couple visits there as well as general visits in the lambic region of Belgium. Gueuzerie Tilquin is Belgium's newest lambic blender. The Gueuzerie is run by Pierre Tilquin, who trained at both Cantillon and 3 Fonteinen before opening up his own blendery and releasing his first products in 2011. Pierre is not as outspoken as other lambic producers, but he's serious about quality products and I've found him to be very welcoming and willing to share information with those passionate about lambic who make the effort to get there (and you should definitely make that effort and visit if/when you are in Belgium).
In front of Gueuzerie Tilquin
Location - The Gueuzerie is located in the municipality of Rebecq in Wallonia (the French speaking part of Belgium) in the Walloon Brabant province. Tilquin is very near to the borders of both the Hainaut province (home of saison) and Flemish Brabant, the province where most of the lambic producers are found. The Pajottenland (the region traditionally thought of as the home of lambic) lies between Brussels and Gueuzerie Tilquin, and even though Tilquin is across the Belgian language border from the Pajottenland, they are seperated by less than one kilometer. Note that the exact extent of the Pajottenland depends on the source you chose to define it. And whether or not you can find certain other main lambic producers in the Pajottenland region (excluding Cantillon which is in the Brussels district) depends on whether you combine the Zennevallei (Senne valley) region in with the Pajottenland (Beersel is in the Senne valley region). See these couple wikipedia pages (in Dutch) if you're into the exact locations of these regions of Belgium (Pajottenland and Zennevallei). The point of all that was to say that while Tilquin is technically not within the borders of the Pajottenland, given their proximity to each other, the variable extent of these boundaries, and given that the boundary definition is not necessarily based on significant differentiating natural features, I don't think that distinction is important here.
Boon Meerts (top) and Girardin lambic (lower 2 rows)
Lambics at Tilquin - As a lambic blendery, Tilquin does not brew his own lambics and instead receives lambic that is brewed at other lambic breweries and is cooled overnight in their coolships. This means that he receives inoculated but unfermented wort the morning after the brew day, and he then ferments the wort in his barrels, ages it and blends it to create his products. Tilquin regularly uses 5 worts from 4 suppliers: Lindemans, Girardin, Boon (both a normal lambic wort and a lower strength wort called 'Meerts', more on this below and in a post to come), and Cantillon. He is currently the only blender using Cantillon wort and the Boon Meerts. And based on some logistical challenges he may also soon be the only blender using another one of those worts. He is also giving other suppliers a trial run (he received Timmerman's lambic in the 2014-2015 season), but first he needs to determine if their wort will work out for the final products he wants. So any full production with other suppliers in his blends awaits the results of the development of those lambics.

Cantillon Lambic at Tilquin (L = lambic, P = 2012-2013 season, 24th brew)
Pierre schedules his deliveries of wort from different suppliers to come in the same week. He feels it makes his job easier if he can do the same sort of work for one week, rather than switching tasks all the time. So the deliveries from his 4 suppliers will come on consecutive (or nearly consecutive) days. This means that he has, in his stock of aging lambics, lambics brewed at basically the same time by different suppliers for multiple different seasons. This is an amazing opportunity for research as well as tasting.

So far much of the lambic microbe development research has been focused on following a batch or two as they age in one brewery. This has given a great profile of the evolution of lambic, but this work is limited in its coverage of lambics from different producers. Cantillon has been the center of some of the work (see Spitaels et al., 2014), and based on their very urban location and high tourist traffic, they may be different in terms of microbe presence and balance than other locations, most of which are more rural and see fewer visitors (or maybe Cantillon is not different in regard to microbes, which would be interesting as well). This is not to pass any quality judgement as I quite like Cantillon and am happy whenever I get a chance to stop in, but just that Cantillon's fermentation may or may not be representative of all lambic breweries. And so far the lambic of other breweries has been under-represented in lambic research.

The Lindemans symbol on a barrel at Tilquin.
So a unique strength of the lambics at Gueuzerie Tilquin is their breadth of coverage from lambics of different producers being gathered all in one place. The weather conditions of the brews would be generally similar among the different lambics because they were brewed at the same time in the same general region. And they've all been stored in the same warehouse since the morning after the brew. So it would allow researchers to look into how more localized influences (a brewery's more immediate surroundings, different brewhouse-resident microbes, slightly different brewing procedures/recipes, possibly very localized weather conditions) might influence the characteristics of different lambics. Pierre expressed interest in having researchers come and study some of his lambics, so hopefully they take him up on that. Given the variability within the small number of lambic producers in terms of things like recipe, process, fermentation behavior, and house flavor, looking into the microbe and metabolite differences would be a great study! And this could provide good initial data to do a more thorough study following batches through time at other producers such as what has been done at Cantillon.

Tilquin is one of the few producers using a 'Meerts', or March lambic. This is a lower strength lambic that was traditionally brewed along with 'normal' lambic. A post on march lambic is in the works (-and it will be linked here when it is up-) so I don't want to spill all the beans about what this historically was. In the case of Tilquin the March lambic comes from Boon and it is used for the draft gueuze (which helps to differentiate it from the bottled version) as well as the faro. The March lambic finishes around the mid 3% abv range, so that would make the OG roughly 6-7° Plato or so, assuming a low FG.

Lambic with blowoff tubes, from 2014.
Managing Fermentation and Aging
Pierre likes to see a healthy start to fermentation rather that a prolonged (weeks to 1+ months) delay in active visible fermentation. He was previously seeing this sort of delay at times when he received deliveries in colder stretches. So he is now warming his cellar when he receives deliveries at times when it is especially cold to help encourage a good fermentation. In the warmer months Pierre is cooling the cellar to keep temps around or below 21° C (70° F).

Barrels are topped up once, after the active fermentation and before the first summer. This usually occurs in May or June. At first Pierre was doing this topping up by using only barrels of the same origin and brew to top other barrels of that origin and brew as much as possible, but this is quite logistically challenging. So now he is topping up barrels without restricting it to only the same origin and brew by using the contents of a what I think of as a 'sacrificial barrel' to top up the other barrels. When there are barrels left that are half full (i.e. if a delivery of wort is not a full unit number of barrel fills), different origins or strengths of wort are blended in to fill them. When I last visited he had a barrel of 1 year old lambic which was a blend of three different origins from the same week that he gave me a taste of. While one source was the main component, and this was identifiable in the lambic, there were some differences brought in by the others which were cool to see.

Rows of lambic from different suppliers
Different lambic origins and blending -
In the blendery, each wort has its own character based mostly on the origin. In that way, for example, Girardin lambics of different ages are identifiably 'Girardin' and the same can be said for the other producers. Of course there will be some batch to batch differences and differences with ages, but the 'origin signature' remains a feature. This is found not only in the flavor of the lambics but also how they behave during fermentation. For example lambic from some producers starts sooner and/or ferments more vigorously (and may blow off more) while other producers may have a longer lag time and/or more sluggish fermentation.

The 'origin signature' of different lambics gives Tilquin and other blenders a pretty unique tool when it comes to producing a house character in their gueuze - they can establish a house character by incorporating different proportions of different origins. Pierre does exactly this, choosing proportions and ages for each constituent origin but keeping the general balance of different origins relatively stable in his blends from batch to batch, which in turn helps keep his house character relatively stable. Desired ages for use of different lambic sources are based on their fermentation behavior and flavor. Pierre chooses certain origin lambics to use when young, others to use when older, and some that he feels are good to use at anytime. The main blending tank at the Gueuzerie is 3800 L, so blends are generally composed of 10 x 400 L barrels.

Lambics and empty barrels, warehouse room 2.
Pierre's approach to blending is fairly straightforward, and he feels that blending is not some great mystery or some exceptionally complicated process. He doesn't go through and try all the possible combinations that he is considering. Instead, he goes down the line barrel by barrel to the lambics of appropriate age (based on origin and subsequently desired age for use) and measures the gravity to make sure that they are ready. If the gravity and taste are good then they are included in the blend until an appropriate number of lambics of a given origin and age are selected. He is certainly making consistently great lambic, so whether you prefer this approach or the approach of trials of many possible combinations, the Tilquin approach definitely works well.

The generally accepted driver of gueuze carbonation is that the younger lambic provides the sugars for bottle refermentation. In the case of Tilquin, it is actually the opposite. The lambic origin that he is preferentially using young may be at 0° P (1.000) as quickly as 6 months from brew day, while the 3 year old lambic that he is using may still be as high as 2-3° P (1.008-1.012) at 3 years old. So the carbonation is derived from the extra gravity in these older lambics, which based on their origin (and microbes and/or brew process) characteristically don't finish as low/as quickly. At one point Pierre was using a bit of sugar in addition to unfermented gravity points to achieve carbonation, but that is no longer the case.

Warehouse  room 1, with a heater for warmth in winter
While blenders get the unique opportunity to use different origin worts and can use this to their advantage, they also have the unique challenge of not having as good of an idea how their lambic blends are going to finish out in terms of final gravities. A producer always working with their own lambic, with their house microbe population, recipe, and brewing process, can be much more sure about final gravity in bottling and is therefore better able to dial in carbonation. For blenders, combining different microbe populations and FGs, there is the potential for the final blend FG to go lower or not as low as the blender suspects. Even when choosing lambics of the lowest possible gravity from a given producer and age to produce his blends, the carbonation is sometimes higher than desired. This is what led to the (Gueuze Tilquin)2which was a gueuze blend that overcarbonated considerably in bottles so it was emptied into barrels and re-bottled. This turned out quite well, and if a similar problem arises in the future I think Pierre is prepared to make another batch of Gueuze2 (though it certainly isn't his intention to have such a situation arise). And, as he currently has an overcarbonated batch from 2013-2014, a new Gueuze2 might be in our future. Pierre is careful to hold onto his blends to monitor carbonation and he packages a few bottles from each blend with crown caps for testing of the carbonation levels in the blend.

Barrel Care-
Barrels are thoroughly cleaned with multiple steps of spraying with increasingly hot water (top temperature = 80° C/176° F). This washes out anything in the barrel (including helping to strip stronger wine characteristics in new barrels). After this the barrels are left open for a few days to dry a bit and then sulfur is burned in them and they are bunged up and stored stacked on their heads until their next use. Before the next use, the barrels are swelled by putting a bit of water on the outside of one head and some water inside the barrel (and therefore inside the other head). The dry storage time is ideally minimal and gueuze blending occurs in the wintertime shortly before batches of wort are due to arrive. In the case of prolonged dry storage the barrels remained sealed and no additional sulfur is used.

The new second warehouse space. Stainless tanks for fruit, empty barrels on their heads and full barrels pyramid stacked.

New Stuff-
The cellar capacity at the Gueuzerie is expanding and now the barrels are split into two warehouse spaces side by side in the same building (in 2013-2014 the barrels were all in the original cellar space in one of the warehouse rooms). Pierre has recently acquired the whole building, which gives him more space to work with. This expansion will allow more production volume as well as more efficient processes. He is first expecting to expand production of the gueuze and draft gueuze, and then possibly fruited products. Of course, with something like lambic which takes multiple years in the making, expansion won't happen overnight...

On the facilities and equipment side, Pierre is looking into getting a new bottling line. It currently takes two days to bottle the 3800 L blending tank. He will be moving the warm conditioning room from the back of the warehouse next to the bottling line to one of the two barrel rooms. He's also planning on moving the office from its current location in the back of the warehouse to the front of the warehouse, and possibly opening up a tasting room toward the front of the warehouse. This will free up more storage room at the back, which in turn will free up even more room for barrels in the two barrel cellar rooms.

In addition to the new spaces and changes being brought with that, there are also some new projects/offerings in the works. I'll leave disclosing details of these to Pierre when he feels it is time. Keep an eye out on the Tilquin Facebook page if you want to know when those projects in the works come to fruition. Something that I think it is safe to say - future batches of Quetsche (both Alsace plums and Prunes de Namur) as well as future batches of Mure are in stainless tanks now.

This is all great news for lovers of Tilquin lambic!

Last updated 30-Oct-2015

Thursday, October 15, 2015

Grisette Recipe

July 2017 edit: I made this post in 2015 and I've learned a lot about grisette since then. I think this is a fine recipe for a good beer that is similar to grisette, but I think you could make a more 'grisette-y' grisette by following another recipe. For an updated and more grisette-like recipe, please see: What is Grisette part II.

This is a follow up to my What is Grisette post with a recipe for a beer that I felt was a grisette. In June 2015 I gave a talk about saisons at a local brewery and I wanted to talk a bit about grisette (prompted by some of the same factors that influenced the first grisette post). And I felt that if I was going to make claims about what I thought grisette was, and that it was different than what people may have previously tasted, then I had better show up with what I thought grisette was. In order to be able to serve it, I brewed this as a 'pilot batch' for a local brewery. But my continued involvement with this brewery will be focused elsewhere and I don't intend this grisette to go into any sort of one-off, seasonal or normal production in Victoria. So I don't think it will be going anywhere commercially for now, but that's not a reflection of what I think about the recipe. I intend to keep brewing something like this on a regular basis at home.

I brewed this before I did much of the research that went into the grisette post. I said in that post that at the time I was calling this beer a grisette, and I stand by that now. But there are probably some parts about this recipe that might not make for the most traditional grisette recipe. First and foremost is the hopping. This is fairly typical of my approach to hopping saison-oriented beers but it is probably inconsistent with what would have been done historically. My approach to hopping these beers is more influenced by the 'heavy on the late hops' sort of approach found in many North American hoppy beers. But I've been really happy with how this sort of hopping using noble-type hop varieties translates to saison-oriented beers. So to me a 'bittering addition' in most of my beers is probably either a first wort hopping or something more like an old school IPA flavor addition of roughly 45-30 minutes left in the boil. - Sidenote, so far I've felt that translating this sort of hopping to fruitier varieties (like southern hemisphere hops) doesn't quite produce the perceived bitterness that I would like, but that's a work in progress...

The second reason that this might not be a classic grisette recipe is the use of flaked oats. I have no reason to say that this was (or was not) employed in historic grisettes. But with such a low OG beer and highly attenuative yeast, I think it is a reasonable addition to build some body into the final beer while staying lighter with the grist. At a comparable time, wit beers would have been using oats (and to some grisette is described as a type of 'biere blanche', though that likely refers to the use of wheat and pale color rather than any other similarities to witbier) so it is not unheard of for Belgian brewers to be using oats at the time. And maybe oats made it into some grisettes, but I currently have no information to substantiate that so for now I'll treat the inclusion of oats as atypical.

So I stand by calling this a grisette, at least based on the 1900's characteristics that formed the basis of my analysis in the "What is Grisette" post. Unfortunately I don't have any of this left, so I can't sit down and write up some notes. But I did drink a reasonable amount of it. At first (before packaging) I thought it was perhaps too hoppy, but in the end I really liked it. It was crisp, light (quite light actually, perhaps a bit too much so) and had a pleasant hoppyness and bitterness to it. At first I thought it was too bitter, but that passed and I was happy with the level of bitterness. The head retention and appearance were nice - lightly hazy but not obtrusively so. Definitely a beer I would happily drink a lot of. I'll have to re-brew this and come up with detailed notes to know exactly how I'd like to change it, but probably the only thing I'd do at this point is shoot for a touch more malt presence, perhaps by slightly increasing the vienna portion.

The Grisette


Brew day: 4-Apr-2015
Batch Size: ~ 12.5 gals (47.5 L) in primary
OG: 1.033
IBU: ~33 (Tinseth)
Transferring wort into my primary fermenter.
FG: 1.001
ABV: 4.2%
Bottled: 12-May-2015

71.3% Doehnel #30 Pils malt
13.6% Wheat malt (1/2 CMC, half Weyermann)
6.8% Flaked oats
5.4% Vienna malt (Best)
2.9% Acidulated malt (Weyermann)
-I should note that this acidulated is for a minor pH adjustment and I think the recipe would probably work quite well (and possibly better) without it.

40g Slovenain Aurora pellets, 8.0% aa at 45 min left in boil (~17 IBU)
34.5g Hallertauer Tradition pellets, 6.7% aa at 50 min left in boil (~10 IBU)
70g Czech Saaz whole hops, 3.1% aa at flame out, 25 minute whirlpool (~4 IBU)
31g Tettnang pellets, 4.4% aa at flame out, 25 minute whirlpool (~2 IBU)

A blend of Wyeast 3724 (Belgian Saison) and 3711 (French Saison), roughly 75%/25% respectively.

1 tab whirlfloc
6g Wyeast yeast nutrients
Water: +~95 ppm Ca2+, +~147 ppm SO42-, +~55 ppm Cl- to water that is basically equivalent to distilled to start

Mash: One step at about 151-152 F (66.1-66.7 C) for an hour. This didn't really result in a higher FG than mashing for fermentability, and sort of backs up my experience with 3711. I've always equated this yeast to brett in some ways as I felt that it kept working on more dextrinous stuff, both in the carboy and in the bottle (even when bottling at/below 1.000). And while everybody mentions crazy attenuation with this yeast I hadn't heard much about continued slow attenuation of 'non-fermentable' material after reaching 'FG' until hearing a couple pro brewers talking about it recently. So, in the same way as I'd treat brett, I probably won't go back to mashing higher to try to decrease attenuation with 3711 in the future, and instead will come up with other ways to increase the malt character and body.

My pressurized racking cane.
This beer was fermented in a de-speared sanke keg. For this batch I just shoved a #11 bung into the top, but that forms a slightly imperfect seal (airlock still bubbles normally and I pressure racked out of it fine). So a bit more O2 may have gotten into the beer than ideal, but the beer was consumed pretty quickly and I didn't notice any issues from O2. It probably helped that I was using noble-type hops. I've since gathered the necessary supplies to adapt the de-speared keg and eliminate this potential for O2 pickup (a tri-clamp gasket that is flat on one side, a short tri-clamp stem and a clamp, see this Milk the Funk wiki page for some ideas).

Because this beer fermented at the brewery instead of at my house in conditions that I know how to control well, the temperature regulation wasn't everything I would have hoped for. But overall it worked out pretty well. The yeast was pitched around 68 F (20 C) and the beer was allowed to free rise/was raised to about 78 F (25.6 C) by day 4. The temp was held in the high 70s (~25 C) for about 10 days or so, and then it dropped to ambient for about 3 weeks. It was then cold crashed for a couple days and bottled/kegged. I used this cane to bottle/keg it (at right), but in the future I'd probably use another racking cane (as described by this Milk The Funk thread, thanks to my friend Shawn for construction help) for that purpose due to the easier ability to adjust cane depth in the fermenter (especially since I could't see where the yeast cake was).

As my research on grisette in ongoing, I'll periodically publish other posts on grisette (which might influence how I'd change brewing one in the future). I'll compile those here:

Hopping Historic Grisettes
Categories of Grisette and Grisette Strength
What is Grisette part II (including an updated recipe)

Wednesday, October 7, 2015

What is Grisette?

(edit Oct 2016: I expand upon and update this post in What is Grisette part II)

This post, which was some time in the making, was prompted by beer naming. This is a very difficult task for many brewers and I don't want to overplay issues in beer labeling or call any specific people out for poor labeling, but I do think that brewers should be thinking about appropriate labels and why that label is applied to a given beer. Labeling terms and beer styles have mutually agreed upon meanings and this is why they are valuable. But when those words are used to describe things outside of the scope of their mutually agreed upon meanings (or when brewers and/or consumers don't know the meanings of the words) then they loose their value. I'll save most of the philosophy and semantics discussion for a potential later post, but in brief, this need for integrity in the terms applied to beers was really brought home to me this past March when, on the same day, two different breweries close to me released a beer that they were calling a Grisette. Grisette seems to be one of the new buzzwords sweeping Belgian-inspired craft beer. It's like the new cool lost and obscure style. While I think grisette is a cool style and I want to see lots of good examples of it brewed, seeing beers popping up with the name grisette can be concerning because not many people have a good idea of what it should mean (both brewers and consumers included). And this is for a good reason - there is very little information out there about it! So consumers might learn grisette to be something that actually shouldn't be represented by the name, but that's all they've seen with the name so that's what they know. These local "grisettes" were, to my knowledge, the first two "grisettes" released in my area. And I didn't think either of these were very Grisette-y. So that got me thinking:

What is Grisette? What does it mean to me? Is that an accurate meaning? What else could it mean?

Grisette Grenier - note the industrial-themed background

After going back and looking through the available information I knew of on Grisette (there's not much), I started a couple discussions with other brewers and beer drinkers who think about these styles (one on the Milk the Funk facebook page and another on the Saison, BdG, Farmhouse appreciation society facebook page, I think you'll have to be a member of those pages to see the discussion). I got some good insight from other brewers and compiling that with the published information I knew about and some reasoned conjecture, here's my take on Grisettes with some historical context:

While Saisons are (historically) beers rooted in the farm, Grisettes are beers inherently rooted in industry and the industrial revolution. They were brewed for miners in the Hainaut province (and possibly adjacent mining areas) in south of Belgium, which until the World Wars was one of the centers of the Industrial Revolution in Europe (see this wikipedia page and the cities mentioned within). The Hainaut province is the same region that saisons come from, and there were likely significant similarities between the two beers. Both mining and farming are rural activities done on (or in) the land with the need for an urban area nearby to sell/process goods and for trading. And in addition both the styles are beers to refresh the workers and, as such, would have been lower strength and would have fulfilled the same basic requirements.

As the world industrialized as a whole, so did breweries. And this would have influenced beers (which influences grisettes as well as saisons). This means larger breweries with better/more specialized equipment and more desire for consistency. And more specific roles and trained people to work those roles rather than brewing being a part time task in the off season. For beer, and especially grisette as it would have been more central in industrialization than saison, this would mean beers trending toward cleaner fermentation and the regular use of certain ingredients rather than more variable recipes. It also might mean that with brewing more as an occupation and less as a role secondary to another occupation such as farming, the amount of time given to each beer for fermentation and conditioning would start to come down, resulting in 'cleaner' beer (by this I mean less influence of slower growing wild yeasts and bacteria and not a reflection on Saccharomyces yeast character, which is often developed and complex for strains common in saison) whether or not 'cleaner' pitches of yeast were used for fermentation. And beers being sold sooner would also lead to brewers wanting to produce beers that were ready to be released sooner.

Malting practices were becoming better at the same time, meaning more consistent and also likely paler malts were available. Consumer taste preferences were starting to shift as beers traveled further and imported beers came in. All of this combines to yield a very dynamic landscape for beer. And as writers of the time still for the most part didn't care about smaller styles of beer being brewed in these rural (though sometimes not so rural) areas of Belgium, there isn't a lot of sound documentation on what was going on.

Some breweries would make both saisons and grisettes in addition to many other beers (see these label collections from Brasserie Bavery and Brasserie Fauconnier for example). It is likely that with the dynamic beer landscape of the time, these styles would have drifted both on their own and in relation to each other. What we know about grisettes from historical records is often contradictory. This could be because authors may be discussing grisettes from different time periods and/or different producers. Beer styles definitely changed over time and grisettes were around long enough to see some of that change. This caveat should be kept in mind for the rest of this discussion, as much of the grisette info I'll reference comes from the 1900s. So perhaps grisette would have been different in the mid 1800s. Differences in historical information could also be because the information at the time was not always accurate/representative and objective or something might be lost in translating outdated/seldom used terms and units of measure. Whatever the case, here's what we've got-

From Farmhouse Ales by Markowski, pp.131-133:

  • No longer produced but used to be fairly common
  • Modern beers in Belgium called grisette are not like historic grisette
  • low abv (~3-5%)
  • light bodied, dry and refreshing
  • saison-like, but with hops giving all of the refreshing quality rather than acidity as well as hops as was possible in old saisons
  • golden in color, clean blond ales
but also
  • OG of 16.3-17.5
  • Amber in color
From Brewing with Wheat by Stan Hieronymus (p.xi, 23):
  • Modern beers called grisette in Belgium are not like historic grisette
  • Disappeared around 1960 (coincidentally around the same time as the major downturn of industry and the subsequent economic downturn of the region)
  • Included malted wheat, around 10% of the grist
  • The wheat was kilned at a very low temperature
  • Grisette was brewed with relatively modern (for the time) brewing techniques
  • Ranged in color from white to brown
  • three strength classes
    • Jeune/young: brewed all year, summer version 'stronger' at 10.2-11 P (1.043-1.046)
    • de garde/de saison: brewed in the winter, 12 P and up (1.048+)
    • Supérior/double: stronger, used first 2 runnings from mash (rest went to a small beer)
  • Fermentation in metal and then wooden barrels
The Brewing with Wheat source is from a text that falls probably on the earlier half of the history of Grisette (the 1870's, source of info mentioned here thanks to Browne and Bitter) and that could explain why there are three versions mentioned there but not in Farmhouse Ales. The Farmhouse Ales info seems to fit best with the Jeune/young Grisette mentioned in Brewing with Wheat. The conflicting Farmhouse Ales description could be discussing the Supérior/double, or there could be something else going on. [edit Sept 2016: at the time that this was written I was guessing that the 1870s were the early half of the history given the info I had available, but was unsure of this as reflected in the original wording. The earliest source I have found so far is from 1812 and mentions a brewery that does at least some regional shipping of grisette, so that puts the likely origin of the style, at the latest, around 1800. Given this info, it is likely that the 1870s were either around the middle of the time range of grisettes or on the later half. Either way, the original point that modern English-language sources which offer inconsistent info could do so due to referencing historic sources from different points in grisette history is valid - grisette in the 1870s would probably have been different from grisette in 1812 and around 1950.]

Old beer labels can also give some insight on grisette. A possible concern about basing this analysis partly on old labels is that I won't really be able to disambiguate marketing speak from truthful information about the beers. Another concern is that I am drawing only from the breweries that are large enough or sufficiently well known to have records of their labels. But given that grisette is by its nature tied to industrialization I am ok with that concern (it may be more of a problem for historic saison studies). Jacques Trifin's website is an awesome resource of old Belgian beer labels and all of the labels I'm showing here come from this site.

Taking the labels with the above precautions in mind, the old labels for grisette that give a strength indication list 2nd Catégorie or 3rd Catégorie. This category system likely has tax class roots, but it also gives us an idea of the strength range of the beer (see this link from the Netherlands, article 4, also this wikipedia page). According to the modern Dutch categories, a Cat. 2 beer has an OG of 7-11 P (1.028-1.046) and a Cat. 3 beer has an OG of 1-7 P (1.004-1.028).  The specific gravity ranges of such a system have likely changed over time and country to country, but presuming that the categories haven't completely changed, category 2 and category 3 beers would be lower strength than what we would now consider normal strength beer (Cat. 1). Of all the numbered strength categories I saw, there were 3, 2, and 1 corresponding to beers I would expect to be increasing in strength in that order, so that gives us some confidence that the 1900's Belgium system is somewhat similar to the modern Dutch one. If someone knows what specific strength ranges these old Belgian categories meant that would give us a more definite answer, but I suspect it won't be extremely different from the ranges here. There is one label for a beer termed "Grisette double" (see here, note that this brewery closed in 1923, early compared to the other labels here) but mostly when I'm seeing a strength modifier it is 2nd category (see below left), something like table beer/house beer (here as well as Francois and sons above left and Grenier top of the page right) or 3rd category versions (here and here, the same shown below center and right).

Cat. 3 Grisette
Cat. 3 Grisette
Cat. 2 Grisette

Synthesizing this, the bulk of the information seems to be pointing toward a low strength, pale, hoppy wheat beer that would not have been aged especially long, though some versions may have been a bit stronger and aged a bit longer. But it seems that most of the grisettes that we have records of now were lower strength beers. Moving ahead with this, from something in the range of category 2-3, the 3-5% abv sounds about right, with grisettes perhaps more likely falling on the lower end of that range later in their history (category 3 would be on the lower end of that range, or quite easily below it, given the modern Dutch OGs associated with these categories). Unfortunately the labels don't offer any specific recipe info, but we can be reasonably sure that (at least at one point) wheat was a fundamental part of the beers given the info from Brewing with Wheat. Many of the breweries that the labels came from were brewing a number of different beers (saison, pils, bock, brune, abbey beers...) so it seems that in the 1900's at least they were confident working with different yeasts in the same brewery. Along with the general identity of grisette being tied to industrialization, this lends support the grisettes being more of a pure culture fermentation than historic saisons. Because the grisette labels have no indication of being an aged beer, while some other labels of the time do include a note for aged beers (and since it would add cost to a brewery to keep a beer around longer you'd think they would want to advertise it), we can also assume that grisettes were generally served younger. This would mean that the hop character would be more forward than an aged beer of a similar recipe and there would not be much time for bacteria (and in addition wild yeast) development, lending support to the Farmhouse Ales information that grisettes were crisp from hop character and without acidity.

Saison Cavenaile - "matured in out cellars"
This gives us a good ballpark for the sort of beer that a grisette was. I think it is reasonable to think of grisette as a lost branch in the saison family fitting somewhere between historic farmhouse saison and modern saison. Many modern saisons are also more clean (using only Saccharomyces). Modern saison gravities are higher than historic saison, and given the gravity category on saison labels from around the time of grisettes, saison gravities were also higher than those of grisettes. Based on descriptions of "Vieille" or "mûrie dans nos caves" (matured in our cellars) on saison labels, saisons of the time were still beers with more age. With imperfect yeast management this would still allow for the expression of wild yeast and bacteria character, though at the time saisons were also moving toward being pure culture products. In looking at the modern day petite saisons, some of these may have considerable overlap with grisettes. I think a microbially clean (or at least mostly clean) but expressively fermented, shorter-aged, low strength beer with wheat and a more prominent hop profile could definitely be a description of either a grisette or a modern petite saison. To me petite saison is more open to the possibility of lactic acid and/or brettanomyces character, but that may be one of multiple sides of petite saison. So although I don't think the two are interchangeable, I think there are beers out there that could be appropriately described by either term.

A category 1 saison
It would be interesting to get a bit more information on Grisette de Saison as I didn't find anything with that name or mentioning longer aging for grisette in the label collection. That could be something more in line with saisons given the higher gravity and longer aging that Markowski mentions. I'd also like to find out some more info on earlier grisettes. So still some questions remaining. I'll update this post as I find more answers (or if there is enough new stuff then I'll do a follow up post).

Ok, so there's some thoughts on grisette. Here's a recipe for a beer that I was calling a Grisette, at least at the time I brewed it.

Apr 2016 Update: I will be presenting this and continued research on grisette at the 2016 AHA National Homebrewers Conference. Hope to see you there if you are headed to the conference! If not, I am hoping to post a follow with my ongoing research sometimes shortly after the conference.

June 2016 Update: I've started a series of posts expanding on various aspects of grisette based on the research from my NHC presentation. Here are those posts:

Hopping grisette
Categories of Grisette and Grisette Strength
What is Grisette part II