🐡 Japan keeps a list. It names the pufferfish species a licensed handler may work with, and for each one it specifies which parts may be sold: muscle, skin, testes. Decades of toxicology went into filling in those cells. Then the sea warmed up, and fish that don't fit any row started coming up in the nets.

A fish crossed the Tsugaru Strait

Around 2013, boats along Japan's eastern seaboard began landing hybrid pufferfish in quantity. Hiroshi Takahashi, now a professor at the National Fisheries University in Shimonoseki, has said that when the first reports reached him he could not take them as real. He had spent years studying hybridization in Takifugu, the genus that covers nearly every pufferfish Japan eats, and until then hybrids had reached him as rare oddities from a market's catch. When he ran the genetics, the scale was unlike anything on record.

Fish collected off Iwate, Fukushima and Ibaraki between 2012 and 2014 yielded 131 first-generation hybrids and 18 backcrosses. Health ministry figures put the hybrid share of puffers landed off Fukushima and Ibaraki in those years at up to roughly 40 percent.

The parents were shosaifugu (Takifugu snyderi), a Pacific species, and gomafugu (Takifugu stictonotus), which spawns in the Tsushima Warm Current on the Sea of Japan side. As those waters warmed, gomafugu pushed north through the Tsugaru Strait that separates Honshu from Hokkaido, out into the Pacific, and then down the Tohoku coast to where shosaifugu spawn. Sea surface temperatures around Japan have risen 1.36°C per century through 2025, according to the Japan Meteorological Agency, a little over twice the global rate.

On the docks this shows up as uncertainty. Asked by the Yomiuri Shimbun, a representative of the Kesennuma Fisheries Cooperative in Miyagi said "puffers hard to tell apart started coming ashore four or five years ago."

Tiger puffer skin, yes. Purple puffer skin, no.

Tetrodotoxin, the neurotoxin pufferfish carry and cooking cannot destroy, is not distributed the same way in every species. Japan's response, set out in a health ministry notice from December 1983 and still the spine of the system, is a table. Species down the side, three parts across the top: muscle, skin, testes. Every cell is a circle or a dash.

Torafugu, the tiger puffer (Takifugu rubripes) and the expensive one, gets circles in all three columns. Mafugu (Takifugu porphyreus) gets circles for muscle and testes and a dash for skin. Same genus, same fish counter, opposite answer. Gomafugu: skin no. Shosaifugu: skin no, plus a footnote recording weak toxicity in testes the table permits.

The table is geographic as well. Nashifugu muscle may be sold only from four named areas. Komonfugu and higanfugu are excluded outright if they came from three named bays in Iwate and Miyagi. Selling anything outside the table carries up to three years in prison or a fine of ¥3 million, roughly $19,000 at late-August 2026 rates.

The fussiness is the point, and it works. Japan logged 10 cases of pufferfish poisoning in 2025, with 15 patients and no deaths, according to ministry figures updated in April 2026. Nine of the ten happened in private homes. In 2003 there were 38 cases and three deaths. The commercial side of fugu has become close to uneventful; what remains is people cooking their own catch.

Nothing on the list quite matches

A hybrid cannot be on that table, because the cells were filled by testing large numbers of individual fish of each species over many years, and nobody has done that for hybrids.

So hybrids are thrown away. Takahashi puts the current situation plainly: "at present we have no choice but to discard them." That is not costless. If hybrids keep increasing, separating them out of mixed catches and binning the lot becomes a fisheries problem of its own.

The safety problem is subtler. How hard a hybrid is to spot by eye depends on which two species produced it. Work from Takahashi's group has shown that hybrids of torafugu and mafugu, and even backcrosses to torafugu, carry toxin in the skin the way mafugu does rather than the way torafugu does. The circle in the torafugu row does not travel with the offspring.

There is still no settled criterion for telling those hybrids apart by appearance, and the ministry's own page cautions that the features vary between individuals and between hybrid generations. Which is why the instruction is not to judge but to exclude: any puffer whose two parent species cannot be identified is to be taken out of the chain. On the evidence of the poisoning figures, that is holding.

How the rulebook caught up

In October 2019 the health ministry issued a national certification standard for fugu handlers. Until then, requirements had been written prefecture by prefecture, difficulty varied, and a licence earned in one prefecture was not necessarily accepted in another. The notice ended experience-based certification, requiring knowledge and skill to be confirmed by examination, and made licences mutually recognized in principle. Prefectures still write the ordinances that implement it.

Two items in that standard are about warming water, though the document never says so. Under species identification, candidates must distinguish the 22 permitted species and reject everything else. Under a heading of its own, "pufferfish hybrids," the stated objective is to understand that unidentified fish must be removed, and that handlers are to keep checking the nationwide occurrence status of hybrids. The ministry undertook to survey that occurrence regularly and pass the results to the prefectures.

The result is a public page. Japan's natural-toxin risk profile for pufferfish now lists which hybrids are turning up where. Torafugu crossed with mafugu: in 2022–2023 surveys, up to roughly 10 percent of catches between Fukushima and the mouth of Tokyo Bay, around 1 percent elsewhere. Shosaifugu crossed with gomafugu: up to roughly 20 percent in Miyagi.

The same page also corrects the 40 percent figure that gets quoted. The surge off Ibaraki and Fukushima collapsed, and ministry data show the combination had become nearly absent there by 2023. It did not stop, it moved north, which is roughly what you would expect if the driver is water temperature.

Whether hybrids ever get rows of their own is an open question. Toxicity testing is under way as a health ministry research project, and Takahashi has said that once the data exists, rules written for particular parent combinations become conceivable. His laboratory has since worked out how to read both parent species from a fish's DNA, and is now trying to make that usable on a working dock.

The look-alike from the south

A second route north is also open. Dokusabafugu (Lagocephalus lunaris) lives in Southeast Asian waters, and has been appearing far enough north that Miyazaki and Kumamoto have sent specimens to Shimonoseki for identification. On the ministry's table it is the row with dashes all the way across: every part is toxic, muscle included. It also closely resembles shirosabafugu (Lagocephalus spadiceus), which is permitted in all three columns. The dependable differences are how far the small spines extend along the back and the shape of the tail fin.

Takahashi says it is hard to study precisely because it turns up so rarely in mixed catches. He has watched for years without obtaining a fresh specimen, while poisonings attributed to the species have continued; discovery usually follows someone getting sick. In a 2024 interview for Japan's national climate adaptation platform, he said that if sea temperatures keep climbing at this pace, both hybrids and dokusabafugu could increase.

Rules are written from the past

Japan's fugu rules are unusually fine-grained: species, then part, then sometimes catch area. That precision is what let a country eat a poisonous fish for centuries with a shrinking casualty list. It is also what makes the rules brittle. Every cell assumes the fish in front of you belongs to a category somebody has already tested.

Other countries drew the line elsewhere. Under EU rules, fishery products from the pufferfish families Tetraodontidae, Molidae, Diodontidae and Canthigasteridae may not lawfully be placed on the market at all. No species list, no parts, no exceptions, and taxonomy-proof by design.

It has not been enough. At least thirteen pufferfish species have been recorded in the Mediterranean. Lagocephalus sceleratus, described by researchers as the most harmful invasive fish in Greek waters, arrived through the Suez Canal and was first reported off Türkiye in 2005. It has since been recorded from Greece, Italy, Spain and the North African coast. It carries tetrodotoxin and it has poisoned people in Greece, Israel and Lebanon. None of those cases passed through a market; recreational fishers caught the fish and ate it. A blanket market ban does not reach a home kitchen, which, as Japan's own numbers show, is where nearly all the poisoning happens there too.

The American arrangement is the mirror image, and it is built out of Japanese categories. Pufferfish has sat on the FDA's automatic detention list since 1980, and personal importation is prohibited. The one opening is an exchange of letters between the FDA and Japan's health ministry. It covers the muscle, skin and testes of tiger puffer, one species and three parts, certified by the city of Shimonoseki. The fish travels from a single recognized exporter there to a single recognized importer in New York, and only for "special occasions." Read that specification and you are reading one row of Japan's 1983 table, exported.

None of which means anything unsafe is on sale; Japan's instruction has moved the other way, toward stricter exclusion. Food safety rules are assembled from accumulated observation, this species, in this place, with this toxin, in these tissues, because there is no other way to write them. So every such rule hides an assumption: that the organisms will stay roughly where they were when the data was gathered.

The pufferfish are not staying. Does the food safety system where you live assume its ingredients hold still?

参照