An animal the weight of a motorcycle moves through dry scrub at the pace of a slow tide.
It stops at a fallen cactus pad and takes it whole, spines and all.
Behind it lies a flattened track of crushed stems where it came through.
Neither of those things is incidental.
The meal is a delivery contract and the crushed track is unpaid groundskeeping, and an island at the southern edge of the Galápagos spent most of a century without either service.
What a tortoise is actually for
Think of giant tortoises as machinery rather than scenery.
The giant prickly pear here drops fruit and pads that one of them swallows entire. The seeds survive the gut, travel for days at tortoise speed, and land somewhere well away from the parent plant wrapped in a parcel of fertilizer.
Nothing else on the island moves cactus seed that far.
The second job is cruder. A shell that heavy going through low brush over and over does not leave the brush standing, and cactus seedlings need open ground and light to get away.
So the animal plants the crop and clears the field, which is why the people managing the island describe it as an engineer rather than a resident. Take the engineer out and both jobs stop on the same day.
What the island did while nobody was walking on it
Goats came ashore and did the obvious damage first, eating the place down toward bare rock and taking the cactus with them. Then the goats were hunted out, and the island did something nobody had planned for.
With the goats gone and almost no tortoises left, woody scrub took over.
Nothing was eating it and nothing heavy was walking through it, and thornscrub is very good at holding ground once it has it.
That closed the door on the cactus. A prickly pear seedling in deep brush does not get the light it needs, and the old adults kept dying without replacements behind them. The island had been rescued from one problem straight into another.
Fifteen animals and a pen on another island
A survey in the 1960s turned up fourteen tortoises, twelve females and two males, so thinly spread across the lava that they were no longer finding each other. Every one was collected and moved to a breeding pen on Santa Cruz.
The fifteenth arrived by ship in 1977, a male taken off the island as a young adult decades earlier who had spent about thirty years in an American zoo.
The goats meanwhile had been there since 1905 and were not cleared out until 1978, three quarters of a century of grazing.
Releases started in 1975 and ran in batches for forty five years. Nearly 1,900 young tortoises went back, and when the program closed the original fifteen were shipped home too.
More than 2,300 animals live there now, all descended from that group.
The famous male gets the credit, though the genetics say he fathered about 40 percent of the output. A second male nobody has heard of produced most of the rest.
The number that shows the island did not come back
Demographically it worked. Roughly half the released tortoises were still alive three decades on, breeding on the lava is normal and unassisted, and no further releases are planned.
Then there is the cactus count. In the zone the tortoises actually occupy, researchers found about 1,310 adult plants, which works out near one stem per acre.
On comparable dry Galápagos islands the same cactus runs between 120 and 180 stems per acre.
That is not a slow recovery. It is two orders of magnitude short. The tortoises also hold under five square miles of an island covering twenty three, and soil carbon readings show the vegetation swinging toward woody plants over recent centuries.
Compare that with a Mississippi couple who spent twenty years rebuilding woodland before anybody released rare frogs into it.
The bird that needs the rock left bare
Which brings in the animal nobody counted. Española is the world’s nesting ground for the waved albatross, a bird so heavy in the air that it cannot lift off from anywhere cluttered.
It needs open lava to run along before it gets airborne, and it needs the same ground clear to come back down on.
As the scrub closed over the bare rock, colonies were abandoned and eggs left behind. The tortoises would clear that ground if they were on it, but their range barely touches the albatross end of the island.
The study behind all this wrote the conclusion plainly. Ecosystem level success takes far longer than population level success, and putting an engineer back often means rebuilding the habitat alongside it.
Bermuda understood that and gave a seabird an island rebuilt for it. Española saved the animal first and is still waiting on the island.
Read the whole thing?
Get the week's signal, not the noise
Our sharpest reporting on energy, climate and nature — free, once a week.