So: hydrate Iris, introduce Terran genomes, occupy. F itself would then be a gorgeous marble in Iris’s sky, a gas giant full of volatiles they would surely need. A giant ball of feedstocks right next door, and its huge beauty helping them occupy Iris by way of its reflected sunlight, which would illuminate all of Iris as Iris slowly turned, and not just one hemisphere, as it had been on Aurora. Really, it looked very promising.

But how long would it take to terraform Iris?

Nothing but guesses could answer this question, and the guesses depended on many assumptions one necessarily had to enter as numbers into one’s models. The median times calculated by the models was privately judged by the ship to be about 3,200 years, with outlier estimates ranging from 50 to 100,000 years. Obviously the models and parameters chosen made quite a difference. In fact the problem was poorly constrained. Still it seemed fair to assume that the median estimate had some kind of theoretical validity.

Many people in the ship didn’t want to wait three thousand years, or however long it might take to terraform Iris. Others didn’t think they could last that long. Others didn’t think it would take that long. “The models must be wrong,” some said. “Surely once life got started on a planet, it would change things fast. Bacteria reproduce very quickly in an empty ecological niche.”

“But on Earth it took a billion years.”

“But there was only archaea on Earth. With the full suite of bacteria it would go fast.”

“Not where there isn’t an atmosphere. Bacteria on rock, exposed to the vacuum, doesn’t grow very fast. It mostly dies, in fact.”

“So we need self-replicating robot machinery to make soil, to make air, to add water.”

“But the selfreps need feedstocks. Collecting the necessary materials can only be accomplished by a first generation of robots, and that won’t be fast.”

“We can print printers and thrive! It can be done. We can do it. Our robots can do it.”

“It’s going to take too long. In the meantime we’ll die out. Evolve at differential rates, and diverge right inside our own bodies. Zoo devolution. Codevolution. Sicken and die and go extinct. Sicken and die and never once leave this ship.”

“So that maybe,” Freya kept saying, “we need to go back home.”

The day came when they tried to make a choice.

Strange, perhaps, to wake up one morning, get dressed, eat breakfast, all the while knowing that one was going off to a meeting that would change the world. Decisions are hard. Everyone has the halting problem. Freya sat next to Badim at their kitchen table, restlessly pushing around cut strawberries with a fork.

“What do you think will happen?” she asked.

Badim smiled at her. He looked unusually cheerful, and was eating heartily, chomping on pieces of buttered toast and washing them down with milk.

“It’s interesting, eh?” he said between bites. “Up until today, history was preordained. We were aimed at Tau Ceti, nothing else could happen. We had to do the necessary.” He waved his bread in the air. “Now that story is over. We are thrust out of the end of that story. Forced to make up a new one, all on our own.”

They walked together to the tram station, then got on a crowded tram car and headed east to Costa Rica. In the biomes along the way the tram stopped and more full cars were linked to it, first in Olympia, then Amazonia. Mostly the people in the tram cars were subdued. People looked pensive. 102,563 conversations had been recorded in the previous month that were about this topic, and there were conflict markers in grammar and semantics in 88 percent of these conversations, which were inevitably held mostly between people well-known to each other.

Now they were done with that. 170.170: the general assembly called for in Costa Rica brought 620 people to the Government House plaza. Most of the rest of the ship’s population watched the assembly on screens throughout the ship, but another meeting, called “in opposition to the tyranny of the majority,” drew 273 people to the plaza in Kiev, in the Steppes.

San Jose’s Government House plaza occupied much of the middle of town. It was surrounded by four-and five-story buildings, all faced with white stone cut to a fussy pattern of interlocking rectangles. The overall impression was of some stage set imitating a European capital, but that could be said of many real European capitals, so it was possibly based on a real square somewhere back on Earth. Ship saw resemblances to Vienna, Moscow, Brasilia.

Now about a third of the population of the ship stood in the plaza, listening to speakers rehash various aspects of the matter at hand. People clumped in cohorts based mostly on home biomes. After the speeches began, the flow between groups was minimal. Some people sat on the smooth flagstones of the plaza; others had brought folding chairs and stools; still others stood. There were some open-walled dining tents to provide food and drink, and what circulation there was in the crowd mostly led to and from these tents.

One sequence of speakers described the plan to shift their efforts to F’s second moon, all of them now calling it Iris. They would establish a base on its surface and move down into that base as they built it to full size. They would add water to the surface by way of a comet bombardment, which would also create the beginnings of an atmosphere. The self-replicating robots and factories would build shelters, break volatiles into gases, create an atmosphere and soil, and shape the growing hydrosphere as it fell from the sky. They would introduce their bacteria to this virgin surface, and it would quickly expand to fill this truly empty ecological niche. After the archaea and bacteria and fungi were established on the ground, they would help to bulk up the atmosphere and to create the soil, and soon further plants and animals from the starship could be introduced, in waves of succession similar to what had occurred in Terran evolution, and the planet terraformed thereby in rapid order, at a speed literally a million times faster than it had happened naturally on Earth: meaning three thousand years instead of three billion. With the chance of doing it in three hundred years also very real, if things went faster than expected.

The various components of this plan were described in some detail by Heloise, and in this effort she was joined by Song. They had joined forces, Song having agreed to the Iris plan with the idea that as an extension of it, her plan for returning to Aurora could be pursued. For now, she agreed with Heloise that terraforming Iris was the best plan, whether interim or permanent.

People stood or sat in silence, listening.

Then Aram was invited to the podium. He stood for a moment staring down at the people, then spoke.

“The problem is this: the spaces we have available to live in are too small to survive in for three thousand years. The main problem is the differential rates of evolution between the various orders of life confined to the space. Bacteria generally mutate at a rate far faster than larger species, and the effect of that evolution on the larger species is eventually devastating. This is one cause of the dwarfism and higher rates of extinction seen in island biogeography studies. And we are an island if there ever was one. And this Iris is not an Earth twin, nor an Earth analog. It is a Mars analog.

“There are also chemicals we need that can’t be found on a rock planet that has never harbored life. In short, the supraorganism that we all constitute together can’t survive over that long a time in the confinement we would be subjected to.”