Upstream of
the Aquarium

In the mid-1990s, researchers Ning Labbish Chao and George Prang estimated that close to 20 million live ornamental fish were exported annually from the middle Rio Negro region of Brazil, generating more than USD 2 million for the local economy and accounting for over 60 percent of the municipal revenue in the area they studied. A follow-up monograph put the 1998 export value at just over USD 2.2 million and estimated at least 10,000 Amazonian residents were involved in some part of the trade.

One species carried most of that weight. Port records from Barcelos between April 1998 and December 2008, analyzed later by Brazil's National Institute for Amazonian Research, logged 376.7 million ornamental fish shipped over that decade — an average of roughly 35.5 million a year. The cardinal tetra, Paracheirodon axelrodi, accounted for 86 percent of it. Everything else — rummy-nose tetras, green neons, hatchetfish, Apistogramma, pencilfish, several catfish species — made up the remaining 14 percent, spread across dozens of other species.

These are historical figures, tied to specific years and specific records, not a description of the trade as it stands today. What they establish is scale: for a stretch of years, one river system and one small fish supported an export economy large enough to matter to an entire municipality, not as a side activity but as one of its main sources of income.

"Low-impact, manually operated gear" undersells what this actually looks like. A 2020 study of 89 ornamental and artisanal fishers in Barcelos documented the specific tools in use: rapiché, a large dip net worked from a canoe along the edge of blackwater streams; cacuri, a traditional woven trap; puçá, a simple hand net; talha and zagaia, used less often. Rapiché was the most common by a clear margin, used by roughly 44 percent of urban fishers and 42 percent of rural ones surveyed.

The fishing happens in specific, unglamorous places — igarapé streams, floodplain lakes, riverbanks, and igapó forest that floods seasonally — using a pole roughly four meters long with netting fixed to the end, worked from a canoe in water too shallow or too obstructed by submerged branches for anything larger. None of this requires imported equipment or a supply chain of its own. It requires knowing exactly where cardinal tetra school, at what time of year, in water that changes character with the season.

That's a form of expertise this site has described before in a different setting: reading a system closely enough, over enough repetition, to know where and when something will be true, without a chemical vocabulary attached to it. Here it's applied to a river instead of a tank. The underlying skill is the same one.

The argument for this fishery has always rested on incentive: if a household's income depends on a healthy river producing healthy fish, that household has a direct reason to protect the river rather than clear it for farmland or exhaust it for short-term gain. That logic held up well enough in the 1990s and 2000s to be documented, published, and cited as a working example of conservation through livelihood, not through restriction alone.

It hasn't held steady. Trade through Barcelos is reported to have fallen by something like 40 to 50 percent over several recent years, driven by competition from higher-quality farmed fish, high mortality among wild-caught stock in transit, and inconsistent export quality, not by a shortage of fish in the river. One industry review, which should be read as an indicator rather than settled data, put the drop even more starkly: from 56.6 million fish exported in 2000 to 6.4 million in 2014.

A socioecological study of the same fishery found something worth sitting with directly: the fishery may be ecologically and socially viable in principle and still shrink anyway, because cash transfer programs, other employment, and shifting markets gave households real reasons to fish less, reasons that have nothing to do with whether the river could still support the trade. An incentive only carries weight while people can actually earn a living from it. When something else pays better, or pays more reliably, the incentive doesn't argue back.

None of this settles the mortality question, and it's worth being honest about that instead of picking whichever number sounds better. Estimates for this trade range from Project Piaba's own reported 3 to 5 percent during storage and transport, up to figures several times higher cited elsewhere for comparable handling conditions. The real rate for any specific shipment today depends on species, handling, packaging, and transit time, not on which side of the wild-versus-farmed argument a keeper prefers.

A separate study of a commercial supply chain feeding the UK market, tracking fish from wholesaler to retailer, found that water chemistry and pathogen exposure shifted measurably at every stage before a fish was ever sold. That's evidence for a general principle, not a Rio Negro-specific one: a fish's condition on day one in a new tank reflects a journey that started well before the keeper was involved.

So the honest version of this isn't "wild-caught is more ethical" or "farmed is safer." It's narrower and less comfortable than either. A fish arrives carrying a river, a fisher's skill, a household's economics, a transport route, and a market's current mood, none of which show up in the tank. Reducing all of that to one label on a price tag was always going to lose most of the story.