I don’t say it will be impossible for them. An intelligent and determined person can accomplish remarkable things. I do say that very, very few of them are going to consider it worth much effort. The majority, however intelligent, are unlikely to be determined.

Of the few that will make the effort, none will have much confidence in their own skill, because they will never have had a chance to check it except on each other. They’ll be like a social club which has decided to learn Sanskrit as a project and" has only books to learn from. There’ll be some uncertainty even in matching an engineering text with the machinery it’s supposed to describe. Given the choice between using the original maintenance manual, printed in chicken tracks which really stand for sounds they’ve never heard, and using the notes made for their convenience by the maintenance workers who already know the machines — which are the kids going to do for homework?

Of course, the original books are still available as the years go on. They certainly aren’t getting worn out. Unfortunately, as the years go on the original books become less and less useful. They need modern texts, in one sense; but there are two strikes against the modern text.

First and obviously, they can’t read it. Second, it’s about as directly useful on machines designed and built a century or so ago as the manual on a power lathe would be to a flint-ax maker of thirty thousand B.C.

The machines designed and built so long ago have lasted well, but not perfectly. Routine maintenance must, more and more often, give way to major repair and even replacement; the original books don’t cover these problems even if they could be read. The notes of the maintenance engineers certainly don’t cover them.

So these people need helpers from the surface, either engineers who can do the necessary work without following a manual, or else harder-to-define experts who can take modern books and transfer their meaning to the local maintenance specialists. Maybe schoolteachers would be the best term.

In other words, they need Joey, and Bert, and Marie, and me. They need practically anyone they can recruit from the surface. Need us. Marie’s hypothesis was perfectly right. They’ve been getting people like us for decades past — the people whose writings enabled me to figure all this out — and their survival depends on keeping it up.

But that gave food for another thought.

It was easy enough to believe that a certain percentage of the people who had come to this place, either accidentally or as a result of surreptitious recruiting, had been persuaded to stay of their own free will. It was much harder to believe that all of them had been. What had happened to those who had not agreed?

I could see two possibilities. One was the fate which Marie seemed to expect if she tried to leave. The other was the explanation Bert had offered, that they had been allowed to return to the surface unharmed but that the Board had covered up their stories or reports.

But Bert was a proven and admitted liar. He might also be wrong.

There were references in the books I had read to visitors who had arrived, but of whom nothing more was mentioned. Of course if they hadn’t stayed it was unlikely that anything would be — either way. I didn’t like to believe that violence had been used — I preferred to believe that Bert was right. Still, Marie was far from stupid, and the morals of this isolated culture might well be those of a century or so back. In fact, in some ways they obviously were.

It was enough for me that there was even a possibility that Marie might be in danger.

For once, I was in complete agreement with Bert; she had to be persuaded to leave at once. Furthermore, she should be guarded until she was well away from here. Guarded by me. That meant two jobs, of which the first was likely to be the harder. Marie had listened to Bert’s arguments about her leaving for several weeks, with no result except a complete undermining of her trust in Bert. How could I possibly do any better?

I claim to be a reasonably good engineer, as I’ve said before, and I can run a competent investigation when the subject is an essentially technical one like tracking down where power is going. I’m not a plotter, though, in the real, old-style meaning of the term, and for a while I was completely stumped by this problem. I suppose what blocked me so long from a working idea was a natural reluctance to tell anything but the truth to Marie, backed up by an even greater dislike of causing her unhappiness.

I don’t know what finally broke through that block. Suddenly, though, it seemed as clear as day that if Marie were bound and determined to stay as long as she believed that Joey might be alive down here, she would presumably go if she were to be convinced that he had died down here.

I didn’t like the idea. I don’t like lying, especially to people who trust me and most especially to Marie. I went through the usual stage in childhood where lying seemed the easiest way out of all troubles, but some very good teachers and a pair of understanding parents, assisted by a close friend with a good right cross who outweighed me by fifteen pounds, had helped me outgrow it. In the present case, I had to tell myself repeatedly that it was for Marie’s own safety before I could decide it was proper to do.

How I convinced myself that it would also be worth the unhappiness it was certain to cause her is something I choose not to discuss. Once I was convinced, the plan was so simple that I wondered why Bert had never thought of it. After all, he seemed to lack my prejudice against falsehood.

I suggested it to him at the first opportunity, and he couldn’t see why he hadn’t thought of it either. He approved strongly, and complimented me as eloquently as developing writer’s cramp would permit. Then he set to work on arrangements.

The plan was simple enough. Joey’s sub was still here, of course. We would simply wreck it, tell Marie we had found the remains, and if necessary show them to her. A little care would make sure that the registry number and enough other identification features remained recognizable. With that much agreed, we set out for the dock where the boat lay. We’d have been able to get to work the moment we reached it, except for the fact that the half-hour swim without communication had enabled each of us to work out all the details. When we resumed conversation, the details didn’t jibe, and it took half an hour or so to reconcile them. With that, actual work and Bert’s search for people to help us with transportation, more than six hours passed before we were really ready to move the sub outside.

We didn’t attempt to run it out under its own power, though that would have been possible. It had been allowed to fill with the living-liquid at local pressure after Joey had been converted. We were able to work on its inner plumbing with no trouble. We thought of bringing it back to the ‘operating room” and connecting it with the transfer lock so that we could pump room and sub back down to surface pressure, but an easier plan had occurred to me.

Like all deep-work machines, Joey’s vessel had very large lift and ballast tanks. The former still worked, not having leaked enough flotation liquid to matter, judging by the sub’s present buoyancy. The latter, of course, were now full of the liquid which formed our regular environment. They were in two major units extending nearly the full length of the hull parallel to the keel, with each unit divided into four cells by bulkheads containing valves and transfer pumps.

We opened all these valves. Then we cracked the seals on the maintenance ports without opening them completely, so that fluid could bleed between the main hull interior and the ballast tanks. The ballast scavenger pumps would now, given time, empty the hull as well as the tanks.