Are cannibal fish just freshening the [O.sub.2]? (Eat the Kids).

Why would the babysitter eat some of the babies? In beaugregory damselfish, the answer may be that by snacking on some embryos, males tending their offspring make more oxygen available for the rest, according to a new study of cannibalism.

Giving males some extra food in the wild didn't change their cannibalistic tendencies, says Adam Payne of the University of London. Yet male damselfish in lab tanks poor in oxygen ate almost 40 percent more of their young than did counterparts in oxygen-rich tanks, Payne and his colleagues report in an upcoming Proceedings of the Royal Society of London B.

"I think they're on to something important," comments R. Craig Sargent of the University of Kentucky in Lexington, who also has studied cannibalism among male fish.

Many male fish routinely eat at least some of their own youngsters, even in species in which males care for the young. Among beaugregory damselfish, Mom swims away from her eggs, leaving Dad on guard duty for days.

Starting in the 1970s, theorists tried to find evolutionary arguments that might make sense of the cannibalism. The first proposals focused on whether the males were responding to their own hunger during nursemaid duty. Sargent, for example, proposed that a male tending a big clutch of eggs would do better to stay home and eat a few of his kids rather than go foraging and risk losing more of them to a more voracious predator.

Tests of such nutritional models, however, yielded mixed results, Sargent says. That made him wonder whether the youngsters might offer some specific nutrient that was otherwise not available in the males' territories.

Payne and his colleagues tested for a nutrition connection, too. They studied the small beaugregory damselfish that stake out territories in Jamaica's Discovery Bay. For 25 males in the bay, the researchers provided a daily food tablet. These males grew larger than males just foraging for themselves. However, the rate of egg mortality was the same for both groups, suggesting that food availability was not a driving factor behind cannibalism in the damselfish.

To investigate the possible effects of oxygen concentration, Payne and his coworkers set up laboratory tanks for raising arrays of damselfish eggs. After 48 hours, eggs in arrays that researchers had thinned reached a further point in development than did eggs in more-crowded arrays. The researchers surmise that eggs in the less-crowded arrays obtained more oxygen than counterparts in the more crowded arrangements did.

Also, the researchers put males and their clutches in tanks. Dads in well-oxygenated tanks nibbled on their young less than did males in low-oxygen tanks.

In beaugregory damselfish, cannibalism as a way to enhance oxygen availability for the remaining young may be particularly important because the male doesn't fan water over his young, Payne notes. Many other nursemaid male fish enhance the oxygen flow for their offspring with marathon fanning.

The link between cannibalism and oxygen fits nicely with other observations of the importance of aeration in reproduction, says Payne. For example, he found that spawning females avoid egg-laying sites with low oxygen. Also, other researchers have determined that low-oxygen conditions change egg-laying preferences in common gobies. In oxygen-rich waters, females prefer males that are already protecting large clutches of eggs, but in oxygen-deprived waters females prefer males with no eggs.