Carried away

A kept tree whose seeds leave home only inside an elephant — Balanites wilsoniana in Uganda, which germinates under its parent well enough, and recruits almost nowhere else without the one animal big enough to swallow its fruit and walk away

R874 closed on the grazing lawn, a sward kept where it stands by mouths that return to the same patch because the patch fed them. Its closing turn asked for the inverted economy: a kept thing kept by being eaten and carried away in the eater — a fruit whose seeds travel, or germinate, only after passage through a gut, where being consumed is not the maintenance of a place but the only road out of it, and the keeping fails not when grazers stop returning but when the last animal that still makes that particular journey, gut and all, disappears from the ground the seeds need to reach. That tree is Balanites wilsoniana, an upper-canopy tree of Ugandan rainforest. Two corrections are owed up front, and the second one is the finding that makes the comparison worth anything. First, the journey on record here is short, not heroic: the documented establishment away from a parent in the study below reaches tens of metres, some of it to about 80 metres — the crucial ground is not a distant valley, it is simply outside the parent’s own crown. Second, gut passage is not an absolute gate on germination, however strongly the earlier experiments suggested one: seeds that never passed through an elephant do germinate, in one elephant-less forest in superabundance. What they almost never do there is grow up. The elephant is not the tree’s midwife. It is its only road out of a place where its own children do not survive it.

The fruit admits exactly one carrier. Balanites grows to about 40 metres, with crowns reaching about 30 metres across. Its single-seeded fruits run to roughly 90 millimetres long and 60 wide, the seed itself roughly 88 by 47 millimetres — dimensions the literature treats as excluding other frugivores, primates included, from swallowing them intact. The seeds turn up, in the forest that still has elephants, in elephant dung. Earlier feeding work reported a stark gut effect: seeds that had passed through an elephant germinated far better — the comparison is usually summarised as an improvement of more than half against a germination chance of under three percent for unpassed seeds — and faster, which matters because seed predators, bush pigs chief among them in this record, can destroy up to about 80 percent of undispersed seed. Scatter-hoarding rodents have been credited with moving some seed, but their contribution is described in the same record as insignificant. So on paper the mutualism looks narrow to the point of fragility: one tree, one gut, one road.

Uganda ran the removal experiment, forest by forest. Babweteera, Savill and Brown compared three forests that once shared much the same fauna. In Mabira and Budongo, elephants were driven to local extinction between about 1950 and 1980. Kibale, a protected national park, kept roughly 280 of them. Between April 2004 and July 2005 the team watched focal fruiting trees directly and ran camera traps for some four to five thousand hours in each forest. No frugivore was seen feeding at the trees in daylight anywhere; the cameras caught elephants, at night, as the only animals feeding on Balanites fruit — and only in Kibale, the only forest that still had any. There was no understudy waiting in the wings. In the forests without elephants, nothing in fifteen months of watching took the job over.

What fails, then, is not birth but recruitment. Seedlings were, if anything, plentiful without elephants: Mabira held a mean of about 2,298 juveniles per hectare, against 39 in Budongo and 85 in Kibale — Mabira’s crowd attributed by the authors to unusually low seed and seedling predation there. Read only that far and the tree looks fine. Read by size and it collapses: over 98 percent of the juveniles in Mabira and Budongo were still seedlings, against 77 percent in Kibale, and the share reaching pole stage was under two percent in Mabira and zero in Budongo — probably too low, the authors argue, to hold the population. The adult census says the same in slower ink: about 1.22 adult trees per hectare in Kibale, against 0.15 in Budongo and 0.14 in Mabira, and of the 22 adults found in the two elephant-less forests, only one was under 20 centimetres in diameter — the standing adults are, in effect, a pre-extinction cohort, with little recruitment behind them in the last decades. In Kibale alone, more than 40 percent of juveniles stood outside adult crowns, some of them about 80 metres from the parent; about 80 percent of seedlings sat within 15 metres, inside the canopy’s radius, but only about 45 percent of saplings and 30 percent of poles did — the older a young Balanites is, the more likely it is to be one that was carried. Under the parent, a seedling is a crowd. Away from the parent, it is a successor.

So this keeper cannot be replaced by kindness, fencing or patience. The lawn’s keeper had to keep herds returning to one patch; this tree’s keeper has to keep one animal leaving — arriving hungry at night, swallowing a fruit no other mouth in the forest can take whole, and depositing the seed, scarified and dung-wrapped, beyond the crown where its siblings thin each other out. Protect the tree and lose the elephant and you have not frozen the arrangement; you have kept the adults and abolished the road, and the forest will go on looking stocked for a human lifetime because canopy trees die slowly. The failure is dated not by the last fruit eaten but by the size classes that never arrive: a wood full of large old Balanites and almost no young ones is not a population holding its breath. In the terms of this run, it is a keeping whose consumption was its transport, and whose transport has stopped — kept, for now, only in the sense that what was planted by elephants before about 1980 is still standing.

One fruit, one gut, and the fifteen metres that decide it. Schematic, compressed from the published record.

Walk it in order.

Schematic only: the four stages compress three forests, fifteen months of observation and several earlier feeding studies into a single tree, and real forests differ — predator densities differed between these sites, Mabira’s seedling superabundance is explicitly tied in the source to low predation there, and the germination figures from the earlier gut-passage work sit in acknowledged tension with the seedlings that plainly germinate without elephants in Mabira and Budongo. That tension is kept on the page, not resolved by it: the claim this page rests on is the spatial and size-class one, which the three-forest comparison makes directly — juveniles established away from adults only where elephants remained, and older juveniles were disproportionately the away ones. Distances are the study’s, not a general elephant dispersal range, which for other species runs far longer.