ndomi sulewayo had pursued his occupation on terraces of earth's grand canyon, on cliffsides of fomalhaut four's highest range and in badlands on the dimly lighted antares twelve. the physical hazards of his present position troubled him little. mclaughlin had agreed that the ledge where the paleontologist had been left was inaccessible to the larger carnivores, and had merely issued a final warning about poisonous "lizards." the primary danger, as nearly as sulewayo could see, was that something might happen to the helicopter. he certainly could not rejoin the others on foot. he was facing a sheer wall some sixty feet high. a score of yards behind him the terrace ended in another straight drop of several hundred feet. a quarter of a mile on either side, the flat surface ended; to the west, by narrowing until the two walls became one; at the other end, it was cut off as far as he was concerned by a joint penetrating apparently the full depth of the canyon.

there were several other cracks in the wall facing him. like those in the tributary canyon explored by krendall and mitsuitei, these were packed with volcanic detritus. this was hard to reconcile with the suggestion that erosion had been long at work. in such a case, the higher portions should have washed away long before the material found at the canyon bottom.

examination at close range suggested a possible explanation. the tuff at this point was fairly well cemented. it seemed reasonable to suppose that the joints had been present before the mountains had started to rise; that a volcanic mud flow had filled them with detritus; that the new material had then been cemented by dissolved material coming from above. this would make the top levels of the tuff more resistant than those lower down, where the cementing minerals had not reached, and account for what had been seen so far.

the hypothesis also implied a plentiful supply of fossils. volcanic mud flowing into a crack in the ground should carry plenty with it. sulewayo set to work with a hammer, and was presently soaking with perspiration.

he was tempted to remove some of his clothing; but this had been treated chemically to repel viridian insects and caution prevailed. mclaughlin had not mentioned any dangerous biters or stingers, and in all probability his blood would not be to the taste of any such creatures on this world—but if the mosquito or tick did not learn that fact until after it had tried, sulewayo would hardly profit by it. in any case the temptation to strip passed quickly. in only a few minutes, his attention was fully occupied by his work; for the expected fossils proved to be present in very satisfactory numbers.

most seemed rather fragmentary. apparently the original creatures had been tumbled about rather badly before the medium hardened. however, the remains were definitely bones, as he had expected and hoped. for some time sulewayo was occupied alternately digging out more fragments and trying to fit the more hopeful-looking specimens together, although he had no success at the latter job. then evidence of a more complete set of remains appeared, and he instantly slowed down to the incredibly meticulous procedure which marks a paleontologist anywhere in the universe.

at this time he had cut perhaps a foot into the tuff for the full three-foot width of the crack and from terrace level up to about his own height. in spite of its apparently firm texture, the rock was extremely soft; and the old question about erosion was reappearing. big pockets of extremely crumbly material had been responsible for most of his speed. now, however, with the usual perversity of the inanimate, a firmer substance was encountered, apparently encasing the bones he suspected of existing a little farther on. this combined with his increased care to bring almost to a halt the removal of rock from the cleft.

the bones were there. perhaps they had been betrayed by a discoloration of the rock too faint for him to have noticed consciously; perhaps something more subtle is involved in the makeup of a successful field worker in paleontology, but as flake after flake of the matrix fell away under his attack a shape gradually took form.

at first a single bone which might have been an unusually short digit or an unusually long carpal—or, of course, something totally unrelated to either—was outlined. then another, close enough to suggest that their lifetime relationship might have been maintained. and another—sulewayo failed to hear the approach of the helicopter until its rotor wash from a hundred feet above lifted the dust about his ankles.

knowing that lampert would be having trouble holding that close to the cliffside, the paleontologist reluctantly hooked his equipment to his belt and started up the ladder. five minutes later they were back in the camp, with krendall listening eagerly to sulewayo's description of his find.

"it's certainly a vertebrate, hans. that stuff can't possibly be shell or wood. it's almost certainly a land dweller—"

"likely enough in that sort of rock, anyway."

"—because i got enough uncovered to be nearly certain that it's a foot. certainly a limb that would not be needed by a swimmer."

"like an ichthyosaur?" queried lampert innocently. sulewayo grinned.

"quite possibly. more likely one of our ubiquitous amphibids, though. certainly something worth getting out, since the general idea is to get an evolutionary sequence of some sort."

"i suppose that means you'll want me to date the eruption which filled all these cracks with detritus, then."

"sure. but there's no hurry. tomorrow will do." lampert found he had no answer to this, and mitsuitei managed to edge into the discussion. he had spent the day with mclaughlin, as he had hoped; and mere failure to find paving stones had not damped his ardor.

"i suppose you and hans will both want to go up the cliff tomorrow," he remarked. "in that case, rob might as well stay with string and me. it will speed up the digging back at my hill."

"are you still scraping dirt off that thing?" asked sulewayo in mock surprise. "didn't one day indicate that it was a joint pattern like the rest?"

"not yet. we haven't gotten down to rock over any place where your cracks should be. the root tangle of the taller trees slows the digging. i admit the rock is limestone like the cliff, but there's still no evidence why those trees grow so regularly."

"that's just what we've been saying all along; but you keep looking for the remains of a city."

"i gathered, ndomi, from your recent conversation that you were digging for a land animal on the basis of three bones. either you are working on hunch, which destroys your right to criticize, or you are reasoning from knowledge not available to the rest of us. in the latter case, you should be at least open-minded enough to credit me with equivalent knowledge in my own field."

it was sulewayo's turn to have nothing to say; he had honestly supposed that the archaeologist had been taking the "city" hypothesis no more seriously than the rest. he apologized at once, and peace was restored. lampert sealed it by agreeing to mitsuitei's suggestion.

the rest of the evening was spent in detail planning by the two groups. at sunset, all turned in to sleep behind the protection of the electrified fence. even the guide regarded this as an adequate safeguard.

apparently his opinion was shared by at least one other. the felodon had spent most of the day under water, part of the time in the canyon fairly close to lampert and krendall and later down the stream by the site where the guide and archaeologist had been working. at neither place had it emerged, or shown the slightest sign of wanting to attack. mclaughlin's reference to the strange instinct of the creatures seemed justified. it certainly could not see the men, but just as certainly was aware of their presence.

what it was about the alien visitors which exercised such an influence on the minute brain of the carnivore, no one could have said—then. any watcher who had supposed, from its earlier actions, that it was moved by a desire for new and different taste sensations would have had to discard the notion now.

with the men safely settled down behind their fence, the beast suddenly turned back downstream. it had returned to the camp site at the end of the working day. in an hour it was in the jungle below the canyon; in another it had killed, and was feeding as it had the moment before the hum of the helicopter had first attracted its attention. this time it finished the meal in peace; and once finished, did not show immediate signs of its former obsession.

instead it sought a lair and relaxed, blending so perfectly into the undergrowth and remaining so silent that within a few minutes small animals were passing only feet away from the concealed killer.

robin lampert was only a fair statistician, but if he had been acquainted with the moves of that felodon during the last few days, even he would have been willing to take oath that more than chance was involved. he would probably have wanted to dissect the animal in search of whatever mechanism was controlling it.

but robin lampert knew nothing of the creature. neither did takehiko mitsuitei; and that was rather unfortunate, for the lair it had selected was on the same hill as the archaeologist's digging site, and a scant quarter mile away from the pit mitsuitei had left.

the rising of the green sun was not visible the next morning. the ever-present mist had thickened into a solid layer of cloud, and hissing rain cut the visibility to a few hundred yards. the helicopter felt its way down to the hill with radar, landed on the river, taxied on its floats to the bank and was moored. lampert, mclaughlin and mitsuitei emerged, the scientists laden with apparatus, and started up the hill toward the site. the guide carried only his weapons.

the equipment was not of the sort mitsuitei was accustomed to using. it actually belonged to lampert. normally it would not be used in an archaeological dig, any more than it would have been had they been fossil hunting; for neither activity takes kindly to any sort of automatic digging machinery. lampert had suggested its use, however, in order to get a rapid idea of the nature of the soil cover, bed rock and joint structure of the hill. if evidence warranted, it would be abandoned for the slower methods of digging. if not, a few hours would permit them to learn as much about the area as many days of work with slower equipment.

the hole mitsuitei had already dug was part way up the hill, in a space cleared of underbrush by a flamethrower. several other such clearings were in the neighborhood. as the archaeologist had said, he had made more than one attempt at digging which had been frustrated by roots.

somewhat to lampert's surprise, it was possible to tell even from ground level the orientation of the taller trees which had been so prominent from the air. even the smaller plants showed signs of some underground influence. between the tallest trees, tracing out the straight lines the men had seen from above, the underbrush formed an almost impenetrable wall. elsewhere foot travel was easy, though the surface was by no means barren. lampert understood how there might indeed have been difficulty in digging on one of the fertile lines, and admitted as much.

"that's the trouble," responded mitsuitei. "i'd like to get down right at such a point, to see what's underneath. it seems to me that paving might be responsible, if they'd used the right materials. lots of civilizations have used organic substances which decay to good fertilizer. then there might be the remains of a sewage system, which would account for richer soil—"

"after the time which must have passed since the place was buried?"

"it has happened. in such a case, of course, trace elements rather than nitrates or phosphates are responsible. that's what i suspect here."

"but wouldn't it be better to dig where you actually have—in the middle of a block, if that's what it is? then you'd be fairly certain to hit a building, which should be richer ground than a street."

"only if you actually strike artifacts. the building itself might be much less well preserved than a paved street. however, you are the one who's handling that mechanical mole. dig where you want, and see what you can learn about this hilltop. just get me at least a couple of cores from my 'streets' before you're done, please."

lampert nodded and proceeded to assemble his equipment. the "mole" was a cylinder about five centimeters in diameter and three times as long. a cutter-lined mouth occupied one end, while the other was attached to a snaky appendage which was wound on a fair sized drum. a set of control knobs and indicators were mounted near the center of the drum.

the geophysicist set the cylinder on the ground mouth downward, pushing it into the soft earth far enough to assure its remaining upright. then he turned to his controls and after a moment, with very little noise, the cylinder began to sink into the ground. in a few seconds it was out of sight, trailing its snaky neck after it.

the men watched it in silence. perhaps thirty seconds after it disappeared, there was a minor convulsion in the neck, a momentarily rising hum from the machinery, and a plug of dirt about two centimeters in diameter and five long was ejected from a port in the center of the drum. this was seized by lampert and examined briefly, then tossed aside. "the soil is pretty deep," he remarked.

"how far down did that come from?" asked mitsuitei.

"one meter. that's the sampling interval i've set in it, for now. if it meets anything much harder or easier to penetrate, it will warn me and i'll grab them more frequently." conversation lapsed while two more samples arrived and were inspected. then a light flickered on the panel, and lampert reset one of his knobs; and almost immediately a core of light gray limestone was produced.

"apparently the same stuff as the cliffs," said lampert after examining the specimen. "do you want to go any deeper, or drill a few more holes to get an idea of the contour?"

"how fast will that thing go through limestone?"

"a couple of centimeters per minute. it's too small to pack a real power unit."

"give it five minutes, just to make sure it isn't a building block."

"ten centimeters wouldn't give you a whole building block."

"a sample from that far inside one would tell me what i want to know. you rock-chippers don't seem to think that archaeology is a science yet. let me have that first core, too." mitsuitei looked confident to the point of being cocky, and lampert let the mole burrow on. the second core came in due time, and the little man set merrily to work with tiny chips from the two stone cylinders, a pinch of the lowest soil sample which had been acquired, a small comparison microscope and a kit full of tiny reagent bottles. lampert used the time the tests consumed in reversing the mole and resetting the equipment on a new spot. by the time the little mechanism had gnawed its way once more to rock, mitsuitei was forced to admit that the formation appeared to be natural.

he did not seem as disheartened by the discovery as might have been expected. he simply waited for more cores, his narrow face reflecting nothing but the utter absorption lampert knew he experienced whenever a problem arose in his line. in spite of his apparent tendency to jump to conclusions, takehiko mitsuitei was an experienced and respected member of his profession. lampert knew enough about his record to be perfectly willing to accept his instructions for the present.

a series of holes was drilled, from the original position toward one of the "streets" forty yards away from it. after each the archaeologist admitted with perfect cheerfulness that there was nothing inconsistent with the idea that the hill was a perfectly natural formation. he still insisted, however, that the regular lines of trees, reinforced as they were by the undergrowth pattern, required explanation.

lampert admitted this, but felt that he knew what the explanation would be. after all, volcanic residue is more than likely to contain the trace elements vegetation requires, even on viridis.

finally the time came to get verification—or the opposite. the flamethrower had to be used this time, and for several minutes clouds of steam swirled about the men as its blue-white tongue fought the sappy, rain-soaked undergrowth. then the mole and its controls were wheeled into place, and the little robot once more nosed its way out of sight.

"i don't suppose you want any samples above the regular rock level, do you?" asked lampert as the machine disappeared.

"i think it would be best if we took them as usual," was the reply. "for one thing, we should try to learn the depth at which the soil composition changes—we are at least agreed that it changes in some manner, after all."

"true enough." the geophysicist set his controls, and the process continued—a process familiar now to mclaughlin as well as the scientists, for the guide had caught numerous glimpses of what was going on while he prowled about the work area on self-imposed guard duty.

mitsuitei took the crumbly soil cores as they came, examined them quickly—they were arriving every few seconds—and filed them in numbered compartments in a specimen case he had opened. detailed stratigraphy would come later. for some time there was no gross evidence of change in the soil; not, in fact, until his first case had been filled.

"can you stop that thing for a moment, rob?" he asked at this point. "i don't want to lose track of these, and will have to hold up while i open a new case."

"all right. i thought you'd want to stop for thought soon anyway."

"why?"

"because the mole is nearly four meters down, well below the depth at which we hit bedrock before, and is still in soil."

"eh? but—but it's still ordinary soil; none of your volcanic ash."

"tuff had been eroded out of a lot of the joints in the cliffs. there's no reason to expect it to be at the same level as the surrounding rock."

"that's true." mitsuitei paused in thought for a moment. "if we keep on going straight down, we may just be working into a natural crack, as you say. might it not be better to drill several holes within a few square yards here, to determine whether it is a narrow joint such as you expect or an actual edge to the rock at this level?"

"maybe the edge of a roof, eh?" lampert chuckled, but spoke in a manner which could give no offense. "i can do better than that. don't need to pull up and start over; simply drill horizontally from where we are now. shouldn't take long to get dimensions, if that's all you want." he halted the robot momentarily, and from a compartment in the drum removed something like a small theodolite mounting. this he set up on a short tripod over the point where the neck of the mole emerged from the ground, and set a pointer at right angles to the line of tall trees. then he started the digging again.