Sake rice is not table rice, and two of our six bottles were table rice

Six bottles went into bags in June and got scored blind by the room. Working back through them afterwards to find out which rice each one was brewed from turned up something I had not expected. Two were made from varieties bred for people to eat, and one of those two finished second.

Almost every English introduction to sake opens with the line that brewing rice is a separate crop from the rice on your plate. That holds up for the varieties and falls apart on the volumes, and the distance between those two facts is where the useful part of this sits.

What sits inside a grain bred for brewing

Japan grows roughly 270 varieties of japonica rice, and a subset of them are designated for brewing. The National Research Institute of Brewing, the state brewing laboratory, gives three features that separate those from the rest. Grains are large, protein content is low, and the middle of the kernel holds a shinpaku, which the institute defines as the white, opaque section at the centre formed by a matrix of starch granules pocked with voids.

Large carries a number. Anything weighing 26 g or more per thousand grains qualifies, and the institute's own side-by-side puts Yamadanishiki at 26.5 g against the table variety Nihonbare at 22.3 g. Hyogo Prefecture, which breeds Yamadanishiki and controls its seed, measured a thousand grains of the 2014 crop at 28.1 g and puts Koshihikari and other eating rice at about 22 g.

Those voids are the whole point of the number. Starch in a shinpaku is loosely packed rather than dense, so koji fungus can drive hyphae into the grain instead of sitting on the surface, and how far in they go is something the brewmaster decides. The institute names two styles. Sohaze covers the entire grain and dissolves the rice hard, which suits full-bodied sake and futsushu. Tsukihaze grows in a spotted pattern, with hyphae well developed in places and absent in others, and ginjo has to be made that way. When Hyogo surveyed 300 breweries on why they buy its Yamadanishiki, the answer given most often, at 45 percent, was that koji grows on it easily and comes out good. Aroma placed second at 12 percent.

Protein is the other half of the definition. Outer layers and germ carry the protein, fat and minerals that feed koji and yeast, and an overabundance speeds fermentation into imbalance and leaves what the institute calls zatsumi, an unrefined taste. Milling is the answer to that. Rice for the table loses its germ and bran, about 8 percent by weight, and arrives at a seimaibuai of 92. Rice for brewing gives up 30 to 70 percent of the grain. Before any of that happens, sake rice already costs on average more than 20 percent more than table rice.

The four names you will see

Yamadanishiki came from a cross registered on 27 February 1936 at the Hyogo Prefectural Agricultural Experiment Station, Yamadabo as the mother and Tankan Wataribune as the father. Around 70 percent of its grains show a shinpaku, and short side chains on its amylopectin make the starch accumulate slowly and dissolve well in the mash. The Hyogo Sake Rice Association counts it in use at 543 breweries and shipped into 43 prefectures.

Gohyakumangoku was bred at the Niigata experiment station in Nagaoka from Kikusui and Shin No. 200, certified in 1957 and named for the prefecture's rice production passing five million koku, about 750,000 tons. Grains run small around a large shinpaku. It is harder than Yamadanishiki and slower to dissolve, while being easier to polish and easier to grow koji on. Early ripening is what gets it off the field before winter reaches Hokuriku, and it carried the light dry Niigata style into the boom around 1980. It held the largest acreage of any brewing rice until Yamadanishiki passed it in 2001. Between them the two are roughly 70 percent of demand by type.

Miyamanishiki is the young one, created in 1978 at Nagano's agricultural experiment station by irradiating Takanenishiki. Takanenishiki grew well in the cold and yielded well, but small grains and a low rate of shinpaku expression made it poor brewing rice. The mutation kept the cold hardiness and added the core. Sake from it comes out light and sharp, since the grain is hard and dissolves reluctantly. Its name points at a shinpaku that looks like snow on a mountain.

Omachi is the old one, found in 1859 by a farmer in Omachi, in what is now Naka-ku in Okayama City, with the pure strain grown today isolated from it in 1922. Soft, absorbent and generous with its shinpaku, it dissolves easily and makes rich, full sake. Okayama grows about 90 percent of the crop, partly because the plant is late and tall and goes over in bad weather. It is also the ancestor of the two above it. Yamadanishiki's father Tankan Wataribune is a child of Omachi, and Gohyakumangoku's mother Kikusui is another.

Most of the rice that becomes sake is not sake rice

Here is the figure that complicates the opening line. Japanese breweries used 180,000 tons of polished rice in 2008, and sake rice accounted for 44,000 tons of it. The other three quarters was rice grown to be eaten. The institute puts it plainly, that certain varieties mainly grown as table rice are also used, and a 20 percent price gap on the input is not hard to connect to volumes that size.

Nothing about that makes a brewery using table rice a lesser brewery. It makes a named variety on a back label a claim about one specific bottle rather than a rule covering the category.

What our six were brewed from

Pacific International Liquor, which brings in three of the June bottles, publishes the rice on each product page, and SakéOne publishes a trade tech sheet for a fourth.

• Okunomatsu Tokubetsu Junmai, second at 4.0, is listed as Akitakomachi, milled to 60 percent.
• Murai Family Tanrei Junmai, fourth at 3.6, is Masshigura at 65 percent.
• Kurosawa Junmai Kimoto, fifth at 3.5, is Miyamanishiki and Hitogokochi at 65 percent.
• Hatsumago, last at 2.0, is Miyamanishiki at 60 percent.
• Hananomai Katana and Momokawa Pearl publish no variety at all. Hananomai says Shizuoka grown and leaves it there.

Akitakomachi is an eating rice. Akita released it in 1984, bred out of Koshihikari for flavour and a cold-hardy line for the weather, and named it after the poet Ono no Komachi. Fifteen years of work starting in 1988 then gave the prefecture a brewing variety of its own, released in 2003 as Akita Sake Komachi, because its breweries had been buying ginjo rice from elsewhere. Two rices, nearly the same name, different jobs. Second place in our only flight went to a bottle made from the one you cook for dinner.

Masshigura runs the same story one prefecture north. Aomori introduced it in 2006 as an affordable everyday rice, blast resistant and dependable in a cold wet climate, sold partly on how well it holds up after it cools. Momokawa mills it to 65 percent and makes a dry junmai that our room placed fourth of six.

Both bottles at the bottom of that flight were Miyamanishiki, and there is an obvious story to tell about that. Those same two bottles were also the only two kimoto in the flight, and both were poured cold against instructions from the brewery and the importer to warm them, which is the argument written up in serving temperature. One evening and six glasses cannot pull rice apart from starter method or from serving temperature. Naming the variety is worth doing because it is a fact about the bottle. Reading a score backwards into the variety is a different move and a worse one.

What all this changes on a D.C. shelf is small and worth having. A back label naming Yamadanishiki or Omachi is telling you the brewer paid up for the grain and wants credit for it, and a back label naming nothing is the ordinary case rather than a warning sign. Our score sheet carries the rice for Hatsumago and Murai Family and leaves it blank for Kurosawa and Okunomatsu, which is a gap the importer's own product pages close.