The larvae of fish, contrary to popular belief, are not simply swept along on dominant currents through the Hawaiian Islands.

Research under way in both the main and the Northwestern Hawaiian Islands is increasingly showing that you can't make generalizations about where larvae go. It seems to depend on the species.

Scientists aboard the NOAA ship Hi'ialakai are continuing genetic research begun last year that could be of critical importance to managing fisheries and refuges. It is beginning to show that the larval stage of certain marine creatures readily move from island to island, while others stay where their parents lived.

"Until 10 years or so ago, the theory was that they drift at the mercy of the currents," said Brian Bowen, a conservation geneticist and marine biologist with the Hawaii Institute for Marine Biology.

Today, it is becoming clear that two factors play a big role in larval movement. One is larval life. Some corals and clownfish larvae must settle on their new home sites in two weeks or they die. Others can live much longer — wrasses a month, surgeonfish two months, lobsters for eight months. Clearly, these animals can travel much longer distances.

But another factor is the larvae themselves. They are not casual drifters, said the institute's Iliana Baums.

"Larvae have behavior. They can choose depth. There may be a lot of behavior in larvae that enable them to come back to where they were born," she said.

"They have both navigation and swim ability," Bowen said. "If they want to stay near the reef, they can."

Damselfish, for instance, stay very close. They form a "halo" around an island during their larval stage, and then return to the same reefs when they're ready.

University of Hawai'i botany researcher Chris Bird is finding that among Hawai'i's three edible limpets, or 'opihi, larval behavior is quite different. The biggest, the kneecap 'opihi, or ko'ele, rarely travels beyond the island where it was spawned. The yellowfoot, or 'alinalina, and blackfoot, or makaiauli, more readily send larvae across channels.

For management purposes, that means for example that O'ahu, whose ko'ele population is severely overfished, can't count on other islands to replenish its stocks of that limpet species.

There are similar differences between fishes.

The research of the institute's Matt Craig, who is aboard the Hi'ialakai, indicates that big-scale soldierfish, or 'u'u are wide rangers, and if you catch one in Hawai'i, it's going to be genetically very similar to ones in other parts of the Pacific.

"That turns out to be a world-class disperser," Bowen said.

By contrast, the yellow tang, prized by the aquarium trade, has what geneticists call "significant structure."

That means that they don't get around much. The larvae from one part of the Hawaiian chain don't seem to make it to distant islands very often.

Researcher Jeff Eble, who conducted studies in the Northwestern Hawaiian Islands aboard Hi'ialakai last year, found that the Northwestern yellow tangs are different from the ones in the main islands.

From a management perspective, that means that if yellow tangs are fished out in the southern end of the chain, there won't be quick re-establishment of the species from the abundant populations up in the Northwestern islands.

These pieces of information are critically important in management of Hawai'i's marine resources, said Randy Kosaki, marine biologist with the Northwestern Hawaiian Island Coral Reef Ecosystem Reserve and chief scientist on the Hi'ialakai voyage.

Genetic research is able to determine how closely related different island's stocks are. That can tell them whether the larvae of one species regularly sweep up and down the archipelago like the big-scale soldierfish, or whether they stay close to home and have little outside contact, like the yellow tang and the kneecap 'opihi.

The research voyage Hi'ialakai is undertaking now will attempt to make similar determinations for a range of other marine species in the Northwestern Hawaiian Islands.

Reach Jan TenBruggencate at jant@honoluluadvertiser.com.