Industries / Food, farming and forestry / Fishing and aquaculture / Shellfish farming
Hatcheries and seed
This is the work of getting young shellfish, called seed or spat, either by catching wild larvae as they settle or by spawning adults in a hatchery and rearing the larvae.
7
Milestones achieved, from 1865 to 2015.
3
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1865
Limed tiles catch wild spat
Growers in south-western France began setting out roof tiles coated with lime and fine sand for drifting oyster larvae to settle on. The spat (young shellfish that have just settled) were scraped off after 6 to 10 months and grown on, and coated tiles became the main way to gather oyster seed in France and the Netherlands, by FAO's account.
Credit
oyster growers of south-western France; FAO names no inventor
1879
Oyster eggs fertilised by hand
On 21 May 1879, at Crisfield, Maryland, William Keith Brooks opened ripe American oysters, mixed their eggs and sperm in a watch glass, and within two hours saw the eggs being fertilised. He showed that this oyster's eggs are fertilised in open water, not inside the mother as had been described for European oysters, and he raised the embryos through their early development.
LessonBrooks first tried to disprove his own idea, opening more than 1,000 oysters without finding an embryo inside a mother.
Credit
William Keith Brooks, Johns Hopkins University, for the Maryland Fish Commission
1950s
Larvae reared on cultured algae
At the federal shellfish laboratory in Milford, Connecticut, Victor Loosanoff and his staff learned to bring bivalves (two-shelled molluscs) into spawning condition nearly all year and to rear every life stage, feeding the larvae selected single-celled algae grown in the lab. NOAA calls this the "Milford Method" and says it is still used worldwide; Loosanoff and Harry Davis set it out in a 1963 review, "Rearing of Bivalve Mollusks".
Lessonfood was the bottleneck, so Loosanoff hired an algae specialist, Ravenna Ukeles, to find which algae suited each larval stage.
Credit
Victor Loosanoff, Harry Davis, Robert Guillard and Ravenna Ukeles, Milford Laboratory of the US Bureau of Commercial Fisheries (now NOAA Fisheries)
1970s
Eyed larvae shipped to farms
By the late 1970s, hatcheries in the US Pacific Northwest were selling growers billions of "eyed" larvae a year, larvae at the stage when they are ready to settle. Growers let them settle on shell in their own tanks, a practice called remote setting, which FAO calls the method of choice on North America's Pacific coast. It replaced seed imported from Japan, which had become too expensive. By 2015 the region's hatcheries produced 40 to 60 billion eyed larvae a year, by Alan Barton and colleagues' figures.
Credit
Pacific Northwest commercial hatcheries; no source read names who devised the method
1981
Sterile triploid oysters
J. G. Stanley, S. K. Allen and H. Hidu published the first report of triploid oysters (University of Florida extension calls it the first), meaning oysters with three sets of chromosomes, which leaves them sterile. They made them by treating fertilised eggs with the chemical cytochalasin B, which rarely gives 100% triploids, so in 1993 Ximing Guo and Standish Allen bred tetraploid oysters, with four sets, whose offspring with ordinary oysters are all triploid. Sterile oysters put energy into growth instead of spawning and stay plump in summer; the extension service cites triploids at about half of US West Coast Pacific oyster production and all French hatchery seed.
LessonRutgers patented the tetraploid method, and the patent expired in January 2015.
Credit
J. G. Stanley, S. K. Allen and H. Hidu (triploids); Ximing Guo and Standish K. Allen, Rutgers University (tetraploids)
2007
Corrosive seawater kills larvae
From July 2007, larvae at Whiskey Creek Shellfish Hatchery in Netarts Bay, Oregon, died in numbers unseen in its 30 years. Staff and scientists traced the deaths to upwelled seawater high in carbon dioxide, in which larvae struggle to build shell. Output in 2008 fell to about 2.5 billion eyed larvae, about a quarter of a normal season. Monitoring the water from 2010 and adding sodium carbonate to the intake water from 2011 won back 30 to 50% of the lost output, by the authors' account, which the hatchery co-wrote.
Lessonthe hatchery first blamed the bacterium Vibrio tubiashii, but larvae kept dying after it was removed, until someone measured the water chemistry.
Credit
Alan Barton (Whiskey Creek), Burke Hales and George Waldbusser (Oregon State University), Richard Feely (NOAA)
2015
Hatchery seed for mussels
New Zealand's Greenshell mussel farms had relied on wild spat, mostly tiny mussels that wash ashore on seaweed near Kaitaia on Ninety Mile Beach. In April 2015 the SPATnz hatchery near Nelson opened to breed selected mussels, and by 2024 its spat were among batches tested on ten farms by an industry research group. In SPATnz's own 2019 trial its mussels reached harvest size in 16.7 months, against 28.3 months for wild spat.
LessonCawthron reports "extreme variability" in the supply of beach-cast spat in recent years.
Credit
SPATnz (owned by Sanford), Cawthron Institute, and the Ministry for Primary Industries, which matched Sanford's NZ$13 million
Watch for
Oyster seed counted by camera
Louisiana Sea Grant's oyster lab and Louisiana State University's Math Consultation Clinic trained a deep-learning model on photos to count and size oyster seed 2 to 6 mm across. It runs on a Raspberry Pi with a 3D-printed phone stand, and the team reports 90 to 94% accuracy and a count in one minute against three by hand, with the code released as open source.
Achieved whena hatchery or farm outside the team uses it routinely and publishes its counts against hand counts.
Claimed by
Louisiana Sea Grant (Elizabeth Robinson, Sarah Bodenstein), February 2026
Sterile oysters without triploidy
NOAA's Northwest Fisheries Science Center and the University of Washington say they have found genes that mark primordial germ cells, the embryo cells that later become eggs or sperm, in Pacific oysters. They present this as a step toward sterile oysters made by stopping those cells from forming, an alternative to triploids, which NOAA says may be more vulnerable to warm water and disease.
Achieved whena hatchery produces sterile oysters this way and independent farm trials show they are sterile and grow at least as well as triploids.
Claimed by
NOAA Fisheries (Mackenzie Gavery) and the University of Washington, March 2026
Scallop seed from elsewhere
Yesso scallop farming depends almost entirely on spat collected from the wild (FAO). Kyodo News reports that summer heat has killed farmed scallops in Mutsu Bay, Aomori, Japan, every year since 2023, and that 2026 output is expected at about 10,000 tonnes against more than 100,000 in good years. In July 2026 Aomori Prefecture said it would study farming young scallops brought in from outside the prefecture and encourage other shellfish; local fishing cooperatives are testing oysters and clams.
Achieved whenseed from outside the bay is stocked at commercial scale and prefectural figures show it surviving a summer.
Claimed by
Aomori Prefecture, July 2026, reported by Kyodo News
Open
01
Warn of larval crashes early
A test, run when a batch of oyster larvae is three to five days old, that predicts whether the batch will crash (die off suddenly), so a hatchery can drain the tank and start again days sooner. The test reads which bacteria live with the larvae.
The briefCredit · unclaimed02
Forecast mussel seed wash-ups
A daily forecast of when seaweed carrying young Greenshell mussels will wash up on Ninety Mile Beach, New Zealand, so harvesters and farms are ready to collect and seed it.
The briefCredit · unclaimed
Open 01 · the brief
Warn of larval crashes early
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A test, run when a batch of oyster larvae is three to five days old, that predicts whether the batch will crash (die off suddenly), so a hatchery can drain the tank and start again days sooner. The test reads which bacteria live with the larvae.
Why it is still open
Hatcheries "regularly experience massive unexplained die-off of stock (termed "crashes") without clear causative factors", write researchers at the University of Maryland's Horn Point Laboratory oyster hatchery. In 15 batches from one season, the bacteria on 3 to 5 day old larvae were statistically related to whether the batch later crashed, but the authors say the bacteria they found are not "predictive", because they had too few batches to test them.
Built with what exists
Hatchery batch records; 16S rRNA gene sequencing (reading one gene that tells bacteria apart), done in-house on Oxford Nanopore's MinION sequencer or sent to a commercial lab; a standard classifier such as a random forest in Python's scikit-learn library.
First step
Ask one hatchery for last season's batch records with dates and outcomes, and start freezing a larval sample from every new batch on day three.
Done looks like
Samples from at least 100 batches at two or more hatcheries; on batches kept out of training, the day-four test flags at least 80% of crashes with no more than 20% false alarms; data and code published.
Open 02 · the brief
Forecast mussel seed wash-ups
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A daily forecast of when seaweed carrying young Greenshell mussels will wash up on Ninety Mile Beach, New Zealand, so harvesters and farms are ready to collect and seed it.
Why it is still open
The industry is "heavily reliant on one major source of wild spat at Ninety Mile Beach", where spat "arrives intermittently" on seaweed, and shortfalls have "caused major disruption", in the 2010 account of A. C. Alfaro, B. McArdle and A. G. Jeffs. Their study linked spat arrivals to strong offshore winds, low onshore swell and El Niño years but published no forecast, and searches found none since. Researchers at the University of Auckland and Cawthron Institute report that from 2015 to 2020 an average of 344 billion spat a year were harvested there and less than 1% survived on farms to harvest.
Built with what exists
Wind and wave hindcasts from ECMWF's ERA5, free from the Copernicus Climate Data Store; NOAA's Oceanic Niño Index; dated spat harvest records; Python.
First step
Ask Fisheries New Zealand or the harvesting companies for dated records of spat harvests from Ninety Mile Beach, and line them up with ERA5 winds for the same days.
Done looks like
A forecast issued every day for a full season and scored afterwards against recorded harvests, catching at least 70% of large wash-up days three days ahead, with hit and false-alarm rates published.