THE PHOTOGRAPHS on the opposite page tell the story of an amateur's investigation of mountain geology. They were taken by Carol De Decker, who lives in the Owens Valley of California and snapped the pictures practically in her own back yard. Carol is only 17, but she is already a veteran of nine years of research on the Sierra Nevada mountain range. Last year a paper she wrote about it won her a top award in the Westinghouse Science Talent Search, and she is now at Pomona College preparing for a career in science.

Besides making her mark in geology, Carol is also an amateur astronomer and has found time to master horsemanship, win letters in tennis and basketball, serve as vice president of her high school's student body and graduate at the head of her class.

From her pigtail days Carol wondered about the forces that built the Sierra. It seemed evident to her that the Sierra Nevadas, with their slender peaks and knifelike ridges, must be young mountains, since she had learned in books that old mountains were worn down in time to gently rolling contours. Yet here and there in the Sierras are stretches of smooth terrain, sometimes on the floors of valleys, sometimes on the summits of peaks. Could the Sierra be both young and old?

As soon as she was old enough to go hiking, Carol began to spend part of every summer tramping and camping the range. She questioned geologists and learned to read their journals. She made lengthy field notes and supplemented them with photographs. Finally her assorted data began to form an orderly pattern. She could answer part of the riddle out of her own experience.

"To the casual observer," she wrote, "the Sierra appears to be a jumbled mass of steep canyons and jagged pinnacles- a landscape devoid of all order. Everywhere, it seems, patches of ancient terrain mingle with the new features of today. Yet with enough care and study the old topography, now glacially carved assumes a clear and logical pattern, and thus gradually the mystery of the Sierra, like its ancient glaciers, melts away."

Carol's prize-winning paper presents a concise geological outline of the range. Essentially the Sierra Nevada is a single block of granite 400 miles long and 40 to 80 miles wide, now badly cracked by the processes of its formation and partly eroded by time. It was thrust up and tilted to the west some 110 million years ago. Thereafter the range was successively elevated and eroded until it reached its present height about a million years ago. Throughout its length the eastern edge descends in steep and irregular cliffs, while on the western side the broad back of the range slopes more gently, finally disappearing under a deposit of alluvial debris. Along the base of the range on the eastern side is a series of faults, some of them geologically recent, indicating that the range is still being elevated.

The seeming age differences in sections of the Sierra's terrain are explained by the fact that the range has undergone four major uplifts, each followed by a period of erosion. Thus it has four distinct types of surfaces, known to geologists as summit upland, subsummit plateau, high valley and canyon. A typical example of each is shown in Carol's photographs. She took these in the Kings-Kern Divide region, an area 150 miles north of Los Angeles and about 50 miles west of Death Valley.

Kings-Kern Divide marks the approximate southern limit of the Ice Age glaciers on the Pacific Coast. The Sierra Nevada owes much of its distinctive appearance to the work of the glaciers. At least three periods of glaciation left their marks on the northern portion of the range, and Carol has recorded evidence of two glacial excursions that reached as far south as Kings-Kern Divide. Here she found a great pile of small rocks mixed with earth and other debris, the deposit of an early glacier. This moraine is partially buried under a thick layer of lava, laid down by the subsequent eruption of an ancient volcano. Over the lava a still more recent glacier deposited another moraine-positive evidence of at least two glacial periods. The comparatively uneroded condition of the topmost moraine indicates that it was deposited recently, perhaps as late as 10,000 years ago.

"The main canyons which open into Owens Valley," Carol reported, "are all alike, in that at about the same elevation each changes from the characteristic form of glacial erosion to that caused by the action of streams. Georges Creek Canyon, which is situated near my home, is typical. In the canyon's upper portion you find rocks with rounded surfaces and there the valley is littered with assorted moraines. At its highest elevation the canyon is broad and open. But lower down, at about 9,000 feet, its whole character changes. The sides narrow into a sharp V. The moraines disappear, and even to the unpracticed eye it is clear that here the rock has been eaten away by the action of running water.

"There are many glacial cirques to be seen throughout the Sierra. These are semi-circular recesses, usually situated at the head of a canyon or basin. In some instances it is evident that two glaciers ate into opposite sides of the same mountain and came so close to uniting at their heads that only a narrow, steep ridge connects the peak with the divide of which it was once a major part.

"From University Peak, a few miles north of Mt. Whitney, one can get an excellent view of several hanging valleys. These are valleys whose floors are considerably higher than the main valley into which they drain. The tributary glaciers once occupying them did not dig as deeply into their beds as did the main glacier, hence the difference in levels when the ice melted.

"I have found few moraines on the western slope of the Sierra. Perhaps the glaciers carried the products of erosion to the western base of the range before depositing them."

THIS MONTH and next the space in this department that is usually devoted to amateur telescope-making is given to another aspect of optics: the ruling engine that makes diffraction gratings. At mention of the ruling engine the average amateur optical worker assumes an attitude of extreme respect for this tricky machine is an even higher example of controlled ultraprecision than telescope mirror-making. In precision optics the standard tolerance of error is one-500,000th of an inch; in diffraction gratings it is one-millionth of an inch.

It is not a coincidence that three men who have been close to precision optics or in it-John Strong, Wilbur Perry and Dave Broadhead-have had much to do with the advanced ruling-engine described elsewhere in this issue. Strong's researches in experimental physics were performed for several years in the optical shop at the California Institute of Technology only a few feet from where the 200-inch Hale telescope mirror was being made, indeed, it was Strong who aluminized this mirror. Perry and Broadhead were members of the team that Strong later assembled at the Johns Hopkins University to build his ruling engine. Perry had been the chief technician at the Johns Hopkins ruling-engine laboratory since 1930, and both he and Broadhead have long been telescope makers.

Perry was born in 1905 in Vermont. Before entering the Worcester Polytechnic Institute he spent a year with Russell Porter in Springfield, Vt., making the six-inch mirrors for 100 of the Porter garden telescopes which were marketed by the Jones and Lamson Machine Company. Perry worked two years as a toolmaker. He was an early member of the Telescope Makers of Springfield, and his portrait, done by Porter, may still be seen at Stellafane, their clubhouse.

In 1930 Perry became technician for the physicist R. W. Wood, long in charge of the Rowland engines. Since then Perry has ruled on these engines about half of the total number of gratings that have been made by the machines since the first one was built in 1882. A pleasant, calm, poised and always deliberate man, Perry is the exact opposite of the tense, hurried, mercurial "slambanger" type who would be a bull in any ruling engine china shop. His hands have four special qualifications for ruling-engine work: 1) they are dry-damp hands corrode gratings; 2) they are both "right" hands-an ambidextrous technician is two men; 3) they are sensitive, and 4) they are sure-fingered-in 22 years he has never dropped as much as a feather on the delicate ruling engine. Habitually "pessimistic" (as the realist's viewpoint appears to the optimist), he gets results where chronic optimists only "op." Throughout the ruling-engine world Perry is a legend. AS an avocation he still makes telescope mirrors.

Broadhead became ensnared in amateur optics in 1936. AS his first adventure he made three 4-inch optical flats and an objective lens, then a 6-inch Newtonian telescope, an 8-inch Dall-Kirkham and a 10-inch Maksutov. In 1941 he joined SCIENTIFIC AMERICAN'S wartime roof-prism group on learning that roof-prism-making called for "men who get results despite all obstacles." The first prism he made was so good that Frankford Arsenal asked for 50 more, 47 of which proved acceptable. In spare hours while doing other work he then made 2,850 roof prisms in the cellar shop of his home in Wellsville, N. Y. Government inspectors accepted more than 98 per cent of them.

After the war Strong selected Broadhead to make a pair of matched 36-inch plate-glass paraboloids for research in infrared spectroscopy and asked for them in 90 days. Broadhead extemporized equipment and, working alone, delivered them in 60 days. This showed Strong where to look for a resourceful builder for the vital parts of his ruling engine. Without ever having seen a ruling engine, Broadhead made these parts working alone in his shop, with a lathe, shaper, grinder and drill press. He cu the threads and pivots of the screws for the Strong engine, built the lapping ma chine shown in the drawing below, lapped the threads and lapped their axis into coincidence with that of the pivots, cut and lapped the end-thrust bearings and built and fitted the working nuts. At Baltimore Perry helped with the design and assembled the parts: the dividing wheels on the ends of the screws, the cross ways, the connector bar and the diamond mounting.

A high-precision screw can only be roughed out on a lathe, because even in the best possible lathe the gears and lead-screw are less precise than the thing we are going to make. After the first roughing out we cannot borrow precision from any machine; we must generate it by some primitive (i.e., prime) method. To make a "perfect" screw (in mechanics perfect means good enough) Henry A. Rowland of Johns Hopkins invented a method so simple in principle, though not in application, that Robinson Crusoe could have made one with the iron crowbars, scrap metal, lead and grindstone that he had-if Robinson had needed a screw and had been a mechanic.

The Rowland method was to make a nut almost the length of the screw out of softer metal, split it lengthwise and equip the halves with tightening clamps. One half of a typical lapping nut for this work, Broadhead's first, is shown in the drawing. It consists of a thick and rigid steel backing into which thin half ring sectors of bronze are screwed and soldered. The sectors are separated by about an eighth of an inch for cleanliness, for even distribution of the abrasive and to minimize flexure from differential expansion. The nut is then threaded as a whole, and is equipped with six push-pull clamps. This lapping nut is run up and down the screw for weeks in emery and oil. The emery grains embed themselves in the softer nut and precision-shape the screw by refined, slow abrasion. This is lapping.

Rowland explained the screw-perfecting principle in negative form: "A long solid nut, tightly fitting in one position, cannot be moved freely to another position unless the screw is very accurate. If grinding material is applied and the nut is constantly tightened, it will grind out all errors of run, drunkenness, crookedness, and irregularity of size.... The condition is that the nut must be long, rigid and capable of being tightened as the grinding proceeds." A short nut would lap only locally, failing to integrate the screw as a whole. A flexible lap would condone the errors. A nut that could not be tightened would soon cease to lap because of wear.

The entire literature on the art of screw-making consists of Rowland's 1,400-word article in the ninth edition of the Encyclopaedia Britannica, George F. Ballou's fascinating but technically unrewarding account of the five-year task of making a ruling-engine screw in an 1895 issue of American Machinist, Edward K. Hammond's article on making screws for scientific instruments in a 1917 issue of Machinery, and J. A. Anderson's short article in Richard Glazebrook's Dictionary of Applied Physics, Volume 4. None of this literature is detailed. Broadhead made his apprenticeship screw without seeing even these articles; sometimes it is more profitable not to read up on an art. Instead he worked and thought-a method of demonstrated value in mirror-making. His first screw came out uniform in diameter over substantially all its length within one-eighth of a wavelength of light (one-400,000th of an inch ). A common machine screw of equal diameter has a diameter tolerance 15,000 times as great.

For several years this department received progress reports in an informal correspondence with Broadhead as he wrestled with the screws. The resulting fat file of interlineated scraps of paper would equal a book in length if fit to print. From them the following narrative, with sidelights on precision screw-making, is abstracted. This account, to be continued in the next issue, should make future precision screw-makers Broadhead's debtors for hundreds of hours of time.

Broadhead's apprenticeship screw, begun early in 1947, was made from a bar of dead soft steel two inches in diameter and 16 inches long. Nine inches of its length was threaded with 20 threads per inch at a 45-degree thread angle. This angle gives better positional stability than the standard 60-degree thread angle, but the steeper angle makes the work much more difficult at the root. Broadhead had to work for an angle accuracy within two or three minutes of arc, and a root-radius accuracy within one-thousandth of an inch.

In roughing (a relative term, since the "rough" screw is rough only under the microscope) the threads were given 10 passes, each .002 inch deep, and then five passes, each .0005 inch deep, with a carbide tool in the lathe. "This is only an approximation," Broadhead says, "since you reduce the cuts progressively by feel, to avoid tearing." Finally came enough passes at .0001 inch each to reach the desired depth. All the cuts were made on one side of the thread and then the screw was reversed. In finishing, each pass advanced .0005 inch or less for a total of .002 or .003 inch, with guided 1,200-grit diamond honing at each pass.

This was as far as the lathe could carry the precision. The "figuring" by lapping was done with the counterweighted split lapping-nut with bronze sectors, shown in the drawing, anointed with oil and emery (at the start, 303-1/2 emery in white lead and olive oil, later 303-1/2 emery alone, then 1200 emery). The nut was constantly run up and down the screw by an electric motor reversed every five minutes by the automatic switches shown. But only the drive is automatic; Broadhead had to make frequent adjustments in tightness. The rub is in deciding which changes to make; the screw is made by the man rather than by the machine.

Broadhead had become pessimistic by autumn and wrote that "the work has progressed like a dumb amateur's first telescope mirror." He found that precision screw-making "parallels mirror-making in a general way but goes at snail's pace because it requires a day to do appreciable lapping in order to test a change in technique, and the possible changes are countless. The screw improves rapidly at first and you celebrate perhaps prematurely. Progress stops beyond the goal if the technique is good enough, otherwise you have failed. I am one of the Strong team and have the screwy end."

Interferometric tests showed the following deviations (or non-deviations) from the desired diameter at 19 stations along the screw-threads: zero, zero 1/500,000 inch, 1/1,000,000 inch, zero, 1/1,000,000 inch, zero, zero, zero, 1/2,000,000 inch, 1/2,000,000 inch, zero, zero, zero, 1/500,000 inch, 1/1,000000 inch, 1/500,000 inch, minus 1/1,000,000 inch, zero.

The experimental screw successful, the screws for the Strong engine came next. These are an inch and a quarter in diameter, threaded on 10 inches of their length, with 45-degree threads at 40 per inch instead of 20. Before beginning, Broadhead made a precision tool of his old $500 South Bend lathe, scraping its ways straight within .0001 inch with the aid of a cast-iron straightedge and Prussian blue, and making them parallel. "That's the catch: parallel," he wrote. He fitted the lathe with new gears, lapped a new lead-screw (no problem for him now), took periodic errors out of its thrust bearing, fitted the headstock and tailstock bearings with high precision to the bed and brought the carriage-travel errors down to .0001 inch in 10 inches. "It's an old lathe," he explained, "but instrument makers use such lathes for centuries, just scraping 'em over-which they'd have to do even with new ones, for this work."

In summer Strong arrived and "for two days," Broadhead wrote, "we worked like beavers. He brought two garbage cans with inner cells, one for electric heat, the other with dry ice, so I could artificially age the screws before the finish cuts were made." After roughing out the screws and thus setting up internal stresses in the metal, Broadhead spent a week relieving them of stress by dipping them in what he called "tincture of skunk cabbage" (overheated vegetable oil, Mazola) at 400 degrees F., then at 100 degrees, and next in "hobo cocktails" (dry ice and alcohol) at 10 degrees, minus 60 degrees, 10 again and again 100. He repeated the whole process for 50 cycles.

Early in 1949 I found work well started on the Strong engine at Baltimore, while the veteran Rowland engines, hovered over by the always careful Perry, plodded along, falling more and more behind the growing backlog of orders for gratings. A little later Strong, Perry and John F. McClellan of Strong's team drove to Broadhead's house in Wellsville to coordinate the work and try to consume a side of Broadhead's venison.

In the spring Broadhead philosophized: "My experience with ruling engines has tempered my amateur eagerness I agree with R. W. Wood that making gratings is not a project for amateurs in the sense that mirrors are. It also takes cooperative effort. I think I could build an engine in about five years, but prefer participation in this team. Working all alone, a rabid, egotistical amateur could easily kill himself if egged on, so don't commit manslaughter. Some guys might think that because they hit millionth-inch accuracy on a mirror they could as easily do so on ruling engines. They might be surprised. A mechanic working to .0001 inch is as good as a telescope maker working to .00001 inch."

Strong praised the upgraded Broadhead lathe: "On the screws Dave can take a cut as small as .00004 inch thick and hold it over the length of the thread, or turn a cylinder 10 inches long accurate to .00005 inch diameter."

I journeyed to Wellsville and found Broadhead peering downward through a 50-power toolmaker's microscope attached to the lathe. The tool was smoothly peeling off a shaving only one micron thick. Without the microscope it seemed to be cutting nothing. As he worked he discussed the philosophy of screws and telescope mirrors.

"A long lapping nut corresponds in theory to a mirror lap. A 'full-size' nut, 10 inches in this instance, would run out over the ends of the 10-inch screw and affect the screw as a full-diameter lap affects a mirror: the screw would become barrel-shaped. On a mirror an 83-per-cent-diameter (neutral) lap on top will hold the curve, neither deepening nor shallowing. On the screw I use a neutral lap of about 90 per cent length, mirrors and screws not being wholly alike."

That summer he finished the threads and began work on the mounting of the screw-several times more difficult than the screw itself. "Progress is slow and what is learned is mainly what not to do. I feel like a sapper crawling in a mine field." And in November: "Anyone who schedules a ruling-engine job by advance dates is my idea of an optimist." When he was asked "What's wrong?" the answer came: "How do I know? Trouble is, there's no Foucault test for screws. A mirror can be tested with great precision throughout the process of making, and the test takes only a few moments, but on a screw the use of the interferometer, nut tension scales, recorders, feel, appearance, all add up to little, and the first real test is the cross-rulings made with the assembled engine. With the screw it takes weeks to evaluate each change in technique. I am now working on getting the pivots concentric with the helix of threads, as mean a job as Anderson said, and on eliminating periodic end-thrust errors, an awful job, far more nerve-wracking than making the screw." ("Mean" and "awful" were not the actual adjectives Broadhead used.) The factors involved are pressure distribution, "stroke" length, oil viscosity, types and sizes of abrasive, speed, tooling method. Nobody seems ever to have made systematic records of these, or perhaps they hid them.

Rowland depended largely on feel as a test for constancy of screw diameter and progressive error; the two are mixed, or additive, not separately identifiable because of the shape of a screw-thread. Broadhead states that if the diameter, as given by the three-wire method, is good to .0001 inch, the error may be attributed to lead. "If the nut moves from end to end without friction," Rowland wrote, "the screw is uniform in diameter." To Rowland's method of testing, Anderson added a lever and spring balance, and Broadhead, with a background in electronics, devised electrical recording apparatus, with an Esterline-Angus tape recorder that makes a continuous ink record of the torque of the nut as it keeps changing while the screw is lapped. The change of torque, Broadhead said, is due mainly to temperature changes in the work, to variations in the lead and diameter of the screw, to the additions of fresh grit and to overhang With this recorder the trends can be watched and the lap adjusted often to prevent seizing. The recorder indicates when fresh abrasive is needed, when a bit of steel has got in, and so on. Even with the recorder it still amounts to making a whole engine to test its screws.

A few days later McClellan arrived at Wellsville with a truck, and the parts of the engine that Broadhead had made were taken to Baltimore.

In January, 1950, Broadhead was queried as he hung at home on tenterhooks, awaiting reports on the newly assembled engine. "Dave," I wrote, "knowing mechanism as you do, did you really expect it to go together and start right off working satisfactorily?" No answer came for several days, and then: "No. But it did. It DID! Strong, Perry and I got on the long-distance phone and celebrated. First grating came off today; only 56 hours to rule it. I just wander around in a daze. Can't think, plan or do anything, except drum on piano,

Since that was written, Broadhead has all but completed the important parts for the engine's duplicate, for gratings with 28,800 grooves per inch.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.