Teranesia

A wave of vertigo swept over him, but he kept his grip on the branch beside him. The centre of the kampung had been reclaimed by the forest, but the trees couldn’t quite obscure the roofs of the huts: the matt-grey photovoltaic surface was still visible through a thin layer of creepers. The buildings had all become badly skewed, but none of them appeared to have collapsed completely. The six huts had been arranged in a regular hexagon, and in their current state he couldn’t tell them apart; with the path from the beach erased there were no cues to enable him to orient the view.

He looked away, remembering his purpose. There wasn’t much foliage around him, but he examined it dutifully. Then he spoke into his notepad.

‘There’s nothing else here. I’m coming down.’

Three more trees yielded five more pupae, but still no sign of the larval stage. It was mid-afternoon; Grant decided there was no point looking further. Prabir was dripping with sweat, and itching from all the contact with bark and sap. When they reached the beach, he handed his samples to Grant and swam out to the reef and back. After the heat of the forest, the water was glorious beyond belief.

He collected his clothes from the beach and waded back to the boat. As he climbed up on to the deck, Grant met him with the latest news from Brazil. ‘They’ve copied whole, purified pigeon chromosomes, using just SPP,’ she said. ‘And the error rate was the same as mine, for the cultured cells.’

It took Prabir a moment to interpret this result. ‘So there is no second protein after all?’

‘Apparently not,’ Grant concurred. ‘SPP alone in a test tube does just as good a job as SPP in an intact cell, if and only if the sequence being copied is the same. Which shows that these changes aren’t errors at all. Or at least, they’re not just random copying mistakes. They must depend in some way on the sequence itself.’

Prabir pondered this. ‘The pigeon genome has probably been copied in the presence of SPP dozens of times. So whatever transformation SPP causes must be convergent: the genome must change less and less with each iteration, until by now it’s virtually stable under the process.’

Grant nodded. ‘Whereas there’s no reason at all why the test sequences they first tried copying would have been stable. Randomly chosen input sequences would have undergone apparently random changes.’

Prabir had a minor epiphany. ‘And all the different fruit pigeons on Banda that ended up looking identical – the process must also be convergent for sufficiently similar genomes. Not only is there a stable endpoint for a given starting point, but similar starting points – closely related species – get dragged towards the same endpoint.’ He beamed with delight. ‘It all makes sense!’

Grant was pleased, but slightly less rapturous. ‘Except we still don’t know what SPP actually does, or how it’s doing it.’

‘But the Brazilians have all the information they need to crack this now, don’t they? They just have to look more closely at their model.’

‘Maybe. For a molecule as large as a protein you can never solve the equations for its shape and binding properties exactly, and it can be hard to choose a set of approximations that only cause trivial discrepancies. They’ve already tried simulating the pigeon chromosome being copied by SPP, and the simulation produced exactly the same error rate as for any other sequence.’

Prabir winced. ‘So their model has just proved that it’s missing the most important subtlety of the real protein.’

Grant didn’t see it quite so bleakly. ‘Missing it now, but they might yet be able to capture it with a little fine-tuning. At least they know what they’re aiming for, what they need to get right.’

Prabir said, ‘OK. So what do we do next?’ Grant had been posting all their results on the net, stating precisely where they’d been collecting their samples; the expedition biologists would already know that there was no need for anyone else to come here. So long as Grant didn’t cut corners.

She said, ‘I’ll have a proper look at these pupae, see what that tells us. I don’t know whether it’s worth going back to hunt for the larvae; I mean, the life cycle is of interest in itself, but larvae don’t make germ cells.’

Prabir filled a bucket with sea water and set about washing his sap-stained clothes, while Grant went for a swim. The travel shop in Toronto had sold him a detergent with enzymes that worked in the presence of salt; as long as you didn’t leave it too late you could remove almost anything with the stuff.

When he walked back into the cabin to get fresh water for rinsing, he glanced at the wire cage holding the adult butterflies they’d captured.

There was a pupa, similar to the ones he’d collected in the forest, hanging from the top of the cage. Except it couldn’t be a pupa. The adults had only been there for a day; at most they might have laid eggs. Grant had been in the cabin twenty minutes before. This had happened since then.

Prabir counted the adults. One of them was missing.