"Yes, Doctor. But we believe, based on extensive analyses by computer and on our experience with catapulting, that today the problem can be solved. Two of our larger firms, the LuNoHo Company and the Bank of Hong Kong in Luna, are ready to head a syndicate to do it as a private venture. They would need help here on Earth and might share voting stock--though they would prefer to sell bonds and retain control. Primarily what they need is a concession from some government, a permanent easement on which to build the catapult. Probably India."

(Above was set speech. LuNoHoCo was bankrupt if anybody examined books, and Hong Kong Bank was strained; was acting as central bank for country undergoing upheaval. Purpose was to get in last word, "India." Prof had coached me that this word must come last.)

Dr. Chan answered, "Never mind financial aspects. Anything which is physically possible can always be made financially possible; money is a bugaboo of small minds. Why do you select India?"

"Well, sir, India now consumes, I believe, over ninety per cent of our grain shipments--"

"Ninety-three point one percent."

"Yes, sir. India is deeply interested in our grain so it seemed likely that she would cooperate. She could grant us land, make labor and materials available, and so forth. But I mentioned India because she holds a wide choice of possible sites, very high mountains not too far from Terra's equator. The latter is not essential, just helpful. But the site must be a high mountain. It's that air pressure you spoke of, or air density. The catapult head should be at as high altitude as feasible but the ejection end, where the load travels over eleven kilometers per second, must be in air so thin that it approaches vacuum. Which calls for a very high mountain. Take the peak Nanda Devi, around four hundred kilometers from here. It has a railhead sixty kilometers from it and a road almost to its base. It is eight thousand meters high. I don't know that Nanda Devi is ideal. It is simply a possible site with good logistics; the ideal site would have to be selected by Terran engineers."

"A higher mountain would be better?"

"Oh, yes, sir!" I assured him. "A higher mountain would be preferred over one nearer the equator. The catapult can be designed to make up for loss in free ride from Earth's rotation. The difficult thing is to avoid so far as possible this pesky thick atmosphere. Excuse me, Doctor; I did not mean to criticize your planet."

"There are higher mountains. Colonel, tell me about this proposed catapult."

I started to. "The length of an escape-speed catapult is determined by the acceleration. We think--or the computer calculates--that an acceleration of twenty gravities is about optimum. For Earth's escape speed this requires a catapult three hundred twenty-three kilometers in length. Therefore---"

"Stop, please! Colonel, are you seriously proposing to bore a hole over three hundred kilometers deep?"

"Oh, no! Construction has to be above ground to permit shock waves to expand. The stator would stretch nearly horizontally, rising perhaps four kilometers in three hundred and in a straight line--almost straight, as Coriolis acceleration and other minor variables make it a gentle curve. The Lunar catapult is straight so far as the eye can see and so nearly horizontal that the barges just miss some peaks beyond it."

"Oh. I thought that you were overestimating the capacity of present-day engineering. We drill deeply today. Not that deeply. Go on."

"Doctor, it may be that common misconception which caused you to check me is why such a catapult has not been constructed before this. I've seen those earlier studies. Most assumed that a catapult would be vertical, or that it would have to tilt up at the end to toss the spacecraft into the sky--and neither is feasible nor necessary. I suppose the asswnption arose from the fact that your spaceships do boost straight up, or nearly."

I went on: "But they do that to get above atmosphere, not to get into orbit. Escape speed is not a vector quantity; it is scalar. A load bursting from a catapult at escape speed will not return to Earth no matter what its direction. Uh... two corrections: it must not be headed toward the Earth itself but at some part of the sky hemisphere, and it must have enough added velocity to punch through whatever atmosphere it still traverses. If it is headed in the right direction it will wind up at Luna."

"Ah, yes. Then this catapult could be used but once each lunar month?"

"No, sir. On the basis on which you were thinking it would be once every day, picking the time to fit where Luna will be in her orbit. But in fact--or so the computer says; I'm not an astronautics expert--in fact this catapult could be used almost any time, simply by varying ejection speed, and the orbits could still wind up at Luna."

"I don't visualize that."

"Neither do I, Doctor, but-- Excuse me but isn't there an exceptionally fine computer at Peiping University?"

"And if there is?" (Did I detect an increase in bland inscrutability? A Cyborg-computer-- Pickled brains? Or live ones, aware? Horrible, either way.)

"Why not ask a topnotch computer for all possible ejection times for such a catapult as I have described? Some orbits go far outside Luna's orbit before returning to where they can be captured by Luna, taking a fantastically long time. Others hook around Terra and then go quite directly. Some are as simple as the ones we use from Luna. There are periods each day when short orbits may be selected. But a load is in the catapult less than one minute; the limitation is how fast the beds can be made ready. It is even possible to have more than one load going up the catapult at a time if the power is sufficient and computer control is versatile. The only thing that worries me is-- These high mountains they are covered with snow?"

"Usually," he answered. "Ice and snow and bare rock."

"Well, sir, being born in Luna I don t know anything about snow. The stator would not only have to be rigid under the heavy gravity of this planet but would have to withstand dynamic thrusts at twenty gravities. I don t suppose it could be anchored to ice or snow. Or could it be?"

"I'm not an engineer, Colonel, but it seems unlikely. Snow and ice would have to be removed. And kept clear. Weather would be a problem, too."

"Weather I know nothing about, Doctor, and all I know about ice is that it has a heat of crystallization of three hundred thirty-five million joules per tonne. I have no idea how many tonnes would have to be melted to clear the site, or how much energy would be required to keep it clear, but it seems to me that it might take as large a reactor to keep it free of ice as to power the catapult."

"We can build reactors, we can melt ice. Or engineers can be sent north for re-education until they do understand ice." Dr. Chan smiled and I shivered. "However, the engineering of ice and snow was solved in Antarctica years ago; don't worry about it. A clear, solid-rock site about three hundred fifty kilometers long at a high altitude-- Anything else I should know?"

"Not much, sir. Melted ice could be collected near the catapult head and thus be the most massy part of what will be shipped to Luna--quite a saving. Also the steel canisters would be re-used to ship grain to Earth, thus stopping another drain that Luna can't take. No reason why a canister should not make the trip hundreds of times. At Luna it would be much the way barges are now landed off Bombay, solid-charge retrorockets programmed by ground control--except that it would be much cheaper, two and a half kilometer-seconds change of motion versus eleven-plus, a squared factor of about twenty--but actually even more favorable, as retros are parasitic weight and the payload improves accordingly. There is even a way to improve that."