A Brief History of Mount Everest’s Altitude and The Configuration of the Summit of Mount Everest
GEORGE EVEREST WAS APPOINTED SUPERINTENDENT of the Great Survey of India in 1823. He became Surveyor General in 1830. He left India in 1843 and Sir Andrew Waugh took his place at that time. From November 1849 to January 1850, Mount Everest, then known as Peak XV, was observed by Mr. J.O. Nicholson from six stations on the northern edge of the Indian network, using a 24-inch theodolite. All of these stations were at altitudes of only about 200 feet above sea level and all were located between 108 and 118 miles from Peak XV. The unweighted mean of altitudes computed from these observations yielded such an exciting result that Radhanath Sikhdar (the Chief Computer) is reported to have rushed breathless into the office of Sir Andrew Waugh with the exclamation, “Sir, I have discovered the highest mountain in the world!” Peak XV was 29,002 feet high. The six altitudes that yielded this result were:
Unweighted Mean 29,002 "
In the midst of these observations, computations and discussions, sincere efforts were made to name this huge peak. Because both Nepal and Tibet were exceedingly difficult to enter, no local names could be discovered (they are now known, respectively, as Sagarmatha and Chomolangma), so in 1854 it was decided to name it Mount Everest in honor of the eminent, recently retired Surveyor General of India. Sir Andrew Waugh, Everest’s successor, proposed the name and it was enthusiastically accepted.
The altitude of 29,002 feet did not last long before professional criticism arose and further efforts were made to do the job better—at least differently.
From 1880 to 1902, a new series of observations were made from the Darjeeling area—from survey stations that were much higher, but still a long way off: 85 to 108 miles from Everest:
These more recent figures were not only from points higher and closer to the peak, but also the computations involved somewhat more sophisticated corrections for atmospheric refraction and a new problem, about which more and more was under discusssion: deflection in the vertical.
This is simply and beautifully described by the authors of Mount Everest: The Reconnaissance, 1921 on pages 10 to 12. Another source of error arises from the varying effects of gravitational attraction. “The attraction of the great mass of the Himalaya and Tibet,” says Burrard, “pulls all liquids towards itself, as the moon attracts the ocean, and the surface of water assumes an irregular form at the foot of the Himalaya. If the ocean were to overflow northern India, its surface would be deformed by Himalayan attraction. The liquid in levels is similarly affected and theodolites cannot consequently be adjusted; their plates when levelled are still tilted upward towards the mountains, and angles of observation are too small by the amount the horizon is inclined to the tangential plane. At Darjeeling the surface of water in response is inclined about 35? to this plane, at Kurseong about 51?, at Siliguri about 23?, at Dehra Dun and Mussooree about 37?. For this reason, all angles of elevation measured from the plains, as Mount Everest was measured, are too small and consequently all our values of Himalayan heights are too small. Errors of this nature range from 40 to 100 feet.”
Another correction which these early surveyors took into careful consideration was the fact that Everest is a snow summit and its altitude must therefore be in a state of flux: higher after the heavy snowfalls of the monsoon months (June, July and August) and lower from December through May after the great winter winds had blown away most (if not all) the summer accumulation.
This new altitude was 29,141 feet, but it never seemed to have the universal appeal that its curious predecessor 29,002 had enjoyed. However, circumstances did change significantly from 1952 to 1954. For the first time, in the late 1940s, the Survey of India was allowed to enter Nepal, in order to provide a triangulation network to control the irrigation program for the huge new Kosi dam. This new survey was under the direction of Mr. B.L. Gulatee, director of the Geodetic and Research Branch of the Survey of India, whose remarkable Technical Paper Number 8 (1954) is now a classic in survey history.
A system of crossed quadrilaterals carried the Indian network northward for 60 miles from Ladnia and Harpur to two new stations in Nepal, about 60 miles almost due south of Everest. From these new points a very complex array of new stations was developed, from which crossed quadrilaterals went eastward for nearly 100 miles to Darjeeling and northward for 20 to 30 miles toward Everest.
This very sophisticated program ended in a complex nine-sided figure with two center-points (See Gulatee, Chart III between pages 9 and 10). Vertical angles were observed from eight points in this figure. Another system went westward for 50 miles, then back southward for 60 miles to tie everything solidly back into the northern arc of India’s survey system.
All this work, which was accomplished from 1952 to 1954, yielded Mount Everest’s presently accepted position and altitude:
North Latitude: 27°59' 15.85? or 96.830.017
East Longitude: 86°55'39.51? or 492.904.957
Altitude: 8848 meters or 29,028.81 feet
These angles were observed with a Swiss Wild T-2 or a British Tavistock and repeated numerous times to achieve the best possible closures—as neither of these is considered to be a first-order instrument. At one of these stations (Lower Rauje), 30 miles south-southeast of Everest, the observers encountered the greatest deflection of the vertical ever recorded anywhere: 70 seconds of arc.
These are still considered as the official data of the world’s highest peak.
Chart V of Gulatee’s Historic Report
Height of station
Spheroidal height difference
Goidal rise between station & Everest
Goidal height of Everest
The new height of 29,028.81 involved a series of major corrections to the raw field figures which had developed as these extraordinary forebears of ours learned more and more about the complex science of geodesy in the presence of huge mountains. Everest also has a snow summit which rises during the monsoon (June to September) and falls from early October till mid winter while the jet-stream descends to well below 25,000 feet and speedily removes all of the new monsoon snow. (See the second memorandum, The Configuration of the Summit of Mount Everest.)
In addition to these basic problems, another one now begins to loom in the inevitably complex adjustment of the huge array of triangles involved in Gulatee’s remarkable Everest system—and also in carrying reliable altitudes, which, of course, start at the Bay of Bengal, nearly 500 miles to the southeast of Everest’s summit. This was the first major mountain where the geoid was seriously taken into account.
Anyone undertaking a new survey of any sort in this area should read every word of Gulatee’s remarkable report and then move onward with any new program of observations made to improve and modernize results.
Nobody made a determined effort to question Everest’s height for over twenty years—until the Chinese climbed to its summit in 1975, and, for the first time, erected a survey target there—one that survived those incredible winter winds for several years. This Chinese survey was a massive effort which involved a triangulation and leveling system that went all across China and Tibet, an overall distance of no less than 1500 miles. This is described in G. Danshang’s report of 1979 and J.Y. Chen’s of 1980. Surveyors worked up onto Everest to altitudes as high as 23,000 feet, leveling was carried to a new benchmark only 16 miles north of the summit and vertical angles observed to the summit target from nine different stations in Tibet. Also, an exhaustive study was made of the local geoid.
The new Chinese altitude of Everest was, extraordinarily, only one foot higher than Gulatee’s! It was also different, as they made clear that this altitude was to the rock beneath the summit snowdrifts: 0.93 meter below the snow surface, to which they are said to have carefully probed. This snow-depth will soon appear as an important part of the Everest-altitude story.
For over a hundred years, K2, some 800 miles from Everest, has had an accepted altitude of 28,250 feet (8611 meters). It is the second highest mountain in the world, and was surveyed by a British team led by Col. Thomas C. Montgomerie in 1854.
In March 1987, Dr. George Wallerstein, a teacher of Astronomy at the University of Washington in Seattle, announced that, using a new tool involving satellites, he had determined that K2 was 11 meters higher than Everest— 29,065 feet. Despite the fact that his new height depended on only one pass of one satellite, the Italians, who had made the first ascent of K2 in 1954, were thrilled. Professor Ardito Desio, the leader of that first-ascent team, quickly assembled a group of professional surveyors to resurvey both Everest and K2 (speedily and in a single month) in August 1987—they properly used the new satellites to establish their initial positions in Tibet and Pakistan, and normal triangulation procedures to secure the altitudes of the two peaks.
To their dismay, their figures showed K2 to be 8616 meters high and Everest to be 8822—not only confirming Everest’s dominance but widening the difference between the two peaks from 778 to 840 feet!
This was very disappointing to the Italians, but, at the same time, it whetted Professor Desio’s interest in the true altitudes of both K2 and Everest. He realized that Wallerstein’s work had been very sloppy, to say the least—but he also simultaneously realized that the speedy GPS checks of his 1987 team were indeed just checks and not the sort of slow, meticulous surveys that had characterized the work of Gulatee.
My personal interest in Everest started on October 9, 1926 when I listened, spellbound, to a lecture by Captain John Noel on the great 1924 British attempt to climb Everest, on which George Mallory and Andrew Irvine lost their lives. Almost exactly 60 years later, on December 20, 1984, our National Geographic team flew over Everest in a Learjet at an altitude of 12,000 meters and took the aerial photographs which yielded our new and detailed map of the Everest area—published in November 1988 by the National Geographic and Boston’s Museum of Science. No new field survey work on the ground was used to make this map.
The only point of immutably-fixed control on this National Geographic map was the top of Everest—from Gulatee. All of the other 90 control points that hold our photo-system together were from the maps of Morshead and Wheeler (British) 1921, Spender 1935 (British), Schneider 1957 (Austrian) and Wang Wenying 1960-1975 (Chinese). All of this mapping, over a long period of years, has assumed that Everest is where the Surveys of India and China have agreed that it is. This has been the only more-or-less indisputable fact—though the detail on the maps which have surrounded this “fixed” point have changed more and more, as photogrammetry has become more and more sophisticated and the airplanes that carried more sophisticated cameras have been able to fly higher and higher.
During the last four or five years, a number of Global Positioning Systems (GPS) experts have begun to focus their interest on the Himalaya—not to secure new altitudes for these summits, but to study their dramatic plate-tectonics movement. One of the leaders in this group has been Roger Bilham of the University of Colorado, who, working with the Survey Department of Nepal, launched a long-range program in 1991 to study the activity of the Nepalese plates. I have been proud to be able to have him as one of my best friends—as, at the very beginning, I helped him to establish friendly relations with the surveyors of both China and Nepal.
Many of these new ultra-precise GPS control points now appearing in Nepal and China are greatly strengthening the rather flimsy local control systems of these remote regions. They are also becoming very helpful to those interested in the possible changes that this new high-quality control will bring to the XYZ of many of the huge peaks in these two countries. But, the GPS plate-tectonics people are first to point out that they show little or no interest in the positions and altitudes of the great peaks of the Himalaya—as the summits involved are almost invariably deeply covered with snow and worthless in the determination of the minuscule movements that are being recorded so accurately by GPS along the boundaries between the plates. You don’t measure heights of snowdrifts in millimeters!
The only reason that the top of Everest might be an exception to this rule is the fact that the Chinese team which targeted it on May 27, 1975, reported that the snow under their survey-work was only 0.93 meters deep. This seemed reasonable, as, only a few yards west and north of the final drift, there has always been a bare windswept rocky area from which climbers have been able to collect tiny “summit-rock” specimens!
Having had a large amount of experience in the use of lasers in our mapping in both Alaska and the Grand Canyon, I decided that it would be very interesting to experiment with the use of the laser, T-3 theodolite and GPS to redetermine the position and altitude of Everest.
In the spring of 1991, Michael Jackson, a brilliant young student of Roger Bilham’s, occupied several GPS stations for Bilham in the Khumbu area just south of Everest at Lukla, Namche, Pheriche and Kala Pattar. Another, just north of Everest, was also set up at Rongbuk. This work began to develop a succession of GPS points that Bilham wished to establish more or less all along the 87th meridian from India northward into Tibet. This work was done (along with many, many other stations) with the enthusiastic interest and support of the Survey Department of Nepal—of which Buddhi N. Shrestha was then the Director.
In January 1992, a three-man team from the Nepal Survey Department, working for Boston’s Museum of Science, set up a new concrete observing stand at Namche Bazar to facilitate high-precision theodolite and laser work—and also made a number of extremely precise theodolite observations to Everest from this area … possibly the first ever to use a Wild T-3 instrument as close as this to Everest.
In May 1992, my wife and I, helped by Amir Shakya of the Nepal Survey Departments traveled to Namche a second time—with a view to doing additional T-3 work, as well as to attempt to measure a laser slope-distance to the top of Everest. This work was closely coordinated with two climbing teams of an Alpine Ascents International (USA) expedition which was planning to ascend Everest via the regular Khumbu-South Col route.
We were also keenly interested in attempting to determine the exact snow depth on the summit of Everest—for, if the snow, indeed, was only a meter thick, it might be possible to make a minor excavation there, drill a hole in bedrock, and establish a meaningful GPS station for Bilham and his associates.
Accordingly, two teams of our climbers reached the summit of Everest on May 12 and 15, 1992. The first was led by Vern Tejas and Skip Horner, the second by Todd Burleson and Peter Athans—both in superb weather. These men not only set up our two-prism assembly there, but they also drove and drilled the 1-inch aluminum pipe to which the prisms were attached 90 inches into the summit snow and never hit bedrock. However, they did hit a hard ice-layer about four feet down and had to screw our pipe through it. It could not be hammered through this thin but resistant obstacle. The two prisms were added to this assembly three days later by our second team. Tragically, they did not bring any more pipe to permit deeper penetration of the summit snow. We had listened too carefully to the Chinese report of 1975 and 90 inches of pipe (three 30-inch sections) had seemed an ample supply.
On May 17, 18 and 19, Amir Shakya and I had total success in measuring our 18-mile slope distances to the prisms on Everest’s summit—and in making T-3 theodolite sights at the same time. As nobody was on the summit that day, we could not record summit air temperature and barometric pressure. We accurately knew Namche temperature and pressure, but we must estimate the summit data. We deliberately did this work from the Namche area in order to keep our slope-angle in the 10° range (10°09'43?)—so that slight angular errors could not have a major effect on Everest’s altitude. This was twice as steep as the British angle observed in 1953 from Gulatee’s nearest survey station at Upper Rauje: 5°05'01?. That was 29 miles away instead of our 18.43 miles.
We retreated from Namche on May 19, delighted with the results of our efforts, but well aware of the fact that these new data needed much additional information before any new altitude of Everest could be calculated: significant data on the local geoid, deflection of the vertical and atmospheric refraction. And, of course, if we wanted to compare these with the present Everest data, a reasonable way must be found to convert our WGS/84 GPS altitude at Namche (3523 meters) to the old “Everest Ellipsoid” system—or vice versa.
Little did we realize that 95-year-old Ardito Desio had big plans in mind—nor did he know what we were doing!
During 1991 and 1992 Ardito Desio and a team of Italian experts, working with the collaboration of the Italian National Research Council and with the financial support of the Swiss firm of Baume et Mercier of Geneva and Leica, planned a major GPS and laser attack on Everest from both China and Nepal toward the end of 1992.
The Chinese set up survey teams at Rongbuk (16 miles north of Everest) as well as at three other excellently positioned points north and northwest of Everest along the margins of the Rongbuk and East Rongbuk Glaciers. Simultaneously, three Italian survey stations were occupied in the Khumbu area of Nepal, closely adjacent to the excellent Italian Research Pyramid headquarters at Lobuche.
On September 28, 29 and 30, this carefully-planned scientific scenario was successfully activated. GPS, laser and theodolite measurements were used to link the top of Everest rapidly to this array of valley stations in both Nepal and China. Even weather balloons were used to determine the temperatures aloft along the laser slope-lines. Wild-Leica equipment of the most modern types were used in this surveying blitzkreig—superbly orchestrated by Professor Giorgio Poretti.
During the last year I have devoted much of my time to the coordination of the work of the Italians and Chinese with our activities and those of Dr. Bilham. Dr. Klaus-Peter Schwarz of Calgary, a Canadian expert on geoids, has recently joined us, as have experts of the U.S. Defense Mapping Agency. We are now sharing all of our data with each other—rather than becoming involved in an international competition—as it is clear that no one of us can secure all of the complex array of data needed to determine a meaningful Everest geoid— without which the precise position and elevation of Everest (or any part of it) cannot possibly be secured.
However, the recently published Italian report on Everest’s altitude adds not only the possibility of an error in calculations but also the fact that the Italian team seems to have chosen to assume that bedrock lies only 2.55 meters below the highest snowdrift—a figure which I seriously doubt.
The altitudes of most snow-capped peaks (like Mont Blanc or K2) have always been measured to what was considered to be the average level of the snow. But Everest has been different—and that difference was initiated in 1975 when the Chinese climbing team (not surveyors) reported that the snow on Everest’s peak was only 0.93 meters deep. If one assumes that this was true (and is still true), then the slight excavation to bedrock makes good sense—but our 1992 measurement of at least 92 inches of snow belies this assumption. The pictures of the summit of Everest by Kurt Diemberger (1978) and Dawson Stelfox (1993) lead one to believe that there are many more feet of snow there than our preliminary measurement of 92 inches … and it now seems very clear that the Italians simply encountered our ice layer rather than bedrock when they probed the summit snows in 1992. It’s a pity that (at least up to today) the equipment for measuring snow depth electronically is so heavy.
With this data in hand, we have decided that GPS measurements to the top of Everest should be made to a permanent survey station set into the nearest solid bedrock adjacent to Everest’s snow summit—and Peter Athans (of Burleson’s Everest team) drilled and marked such a station on May 13, 1994—slightly on the Chinese side of the summit. At that time, the Nepalese Government had started to prohibit all scientific research on or near to Everest. The constantly-slightly-changing difference in elevation between this GPS point and the snow summit must then be determined by GPS or by a short (and frigid!) leveling procedure. After all, the important scientific data about Everest’s summit have to be recorded on bedrock anyway, as the changes in these data, which have now become of such interest, must be measured in millimeters which cannot be measured to the top of a snowdrift!
This is all rather tragic from the Italian point of view, as their 1992 research team could easily have done its work on the nearby bedrock, had they not been misled by that layer of ice.
The Chinese National Bureau of Surveying and Mapping is enthusiastically interested in carrying out a bedrock GPS observation program atop Everest. We all hope that the government of Nepal will join us in a truly international program by lifting its extraordinary prohibition of scientific field-work on Everest. An intense program to study Everest’s geoid, coefficient of refraction and deflection of the vertical is already well underway in China and the full involvement of Nepal and India is essential to a thorough understanding of the unusual characteristics of this lofty part of the world.
Part 2: The Configuration of the Summit of Mount Everest
Gulatee’s observations, as described above, were deliberately made in the middle of the winter (December-March) “when the amount of snow on top is likely to be least, as the northwest wind is in its full stride. There is heavy precipitation of snow during the monsoon months (June to September) and, although there is no observational evidence regarding the change of snowfall at the summit, it is likely to be well over 10 feet.” (Gulatee, page 14)
I now believe that this monsoon-snow accumulation is approximately two meters—beginning to build up in June, reaching its maximum around mid September to early-October—and almost all is blown off by the “winter winds” by January 1. From that date until late May or early June, the highest snowdrift of Everest’s summit stays at about the same altitude.
When Amir Shakya of the Nepal Survey Department and I measured our extremely precise T-3 vertical angles from Namche Bazar to Everest’s summit (May 17, 18 and 19, 1992) we were well aware of two things: first, that Everest’s highest point has for many years been a massive snowdrift; second, that there has to be a seasonal change in the depth of the snow there because of the monsoon. On the other hand, I felt strongly that the “winter winds” do not seem to lessen significantly until early June (when the monsoon normally starts) and that, therefore, the average minimum altitude of the summit drifts would be more likely to occur then than in midwinter.
Our 1992 sights to Everest’s top were unusual because they were not made to the crest of the highest drift, but rather to a laser prism which was set there for us on May 15 by the climbing team led by Todd Burleson. We were aiming at a very precise tiny point of light instead of setting our hair-wire tangent to a snow-surface.
Another unique aspect of this 1992 vertical angle was that we were keenly aware at that time of the desire of the plate-tectonics experts to study the changes in the top of Everest rather than simply to record the position of that drift. We therefore gave Burleson’s expedition 90 inches of strong aluminum pipe, with an ice-screw at its end, to try to determine the depth of this ever-changing drift of snow.
When the Chinese survey target was set on the top of Everest on May 27, 1975, the party reported that the snow was slightly less than one meter deep. I reasoned that if this was, indeed, still true, it would be possible for a scientific team to dig that amount of snow away and insert a tiny stainless steel pin in the bedrock, which could then be relocated very easily to determine the precise movement of the summit for geophysical research studies.
Our party (Todd Burleson, Pete Athans, Skip Horner and Vern Tejas of Alpine Ascents International) reached the summit on May 12 and 15, 1992 and drove the pipe easily down for a little more than a meter, then ran into a solid surface through which it could no longer be pounded. They then screwed it through 4 to 6 inches of ice or very hard snow—after which it was easily driven to a depth of 90 inches (3 30-inch pipe sections) without any more resistance at all. Our two laser prisms were then added to an adapter on the top of this pipe: one prism was 8 inches above the snow surface, the other 14 inches above the snow. This assembly was carefully aimed at our survey station at Namche Bazar and it was successfully observed several times about a week later, both from Namche and from another survey station drilled in bedrock at the Thyangboche Monastery.
Three-and-a-half months later, on September 30, 1992, Giorgio Poretti's powerful and experienced survey team brought both laser and GPS equipment to Everest’s summit and made a series of observations from six different valley-survey-stations in both China and Nepal. At that time our prism assembly was not to be seen anywhere. The Italians left a very large and sophisticated target about two meters in height, with three prisms aimed at Tibet and three at the Lobuche area of Nepal.
Its main vertical aluminum pole was not driven deeply into the snow and was kept in position both by three guylines and three rigid aluminum poles which propped it from the sides. This party also made an attempt to determine the depth of the snow with a rigid avalanche-probe and ran into a “very hard surface” which they felt might be rock at a depth of about 2.55 meters. Their probe did not have the ability to screw through this hard surface—and I feel that this is very likely to have been our hard layer of icy crust—and not bedrock.
To this array of data, the Irish Everest Expedition which reached the summit on May 27, 1993 adds some extremely interesting new facts. This party (Dawson Stelfox and Frank Nugent) approached the summit from the northeast and made a unique series of photographs of both the First and Second Steps and their approaches. Stelfox (then alone) also took one on the summit looking northeast, with our two-prism assembly in the foreground, with the snow- surface exactly where it had been exactly a year earlier! The prisms were still aimed at Namche Bazar—and behind them was a neat bundle of the Italian target equipment jammed in the snow a few yards northeast of our prisms. This assembly had apparently been blown down during the 1992-3 winter winds and had been carefully tied together and placed upright in the snow by someone in the spring of 1993. One of the three aluminum braces of the Italian target was separated from this package and was jammed vertically in the snow right next to our prisms. This dramatically indicates that, for one year at least, the snow depth on the summit increased significantly during the monsoon, by an amount at least enough to conceal our prisms—then the winter’s gales had scoured the top so that it had exactly resumed its “average minimum” appearance of the year before.
A second photograph made by the Irish party is of at least equal interest. It was taken from a point perhaps 100 yards from the top, looking southwest. It shows the summit drift in profile, with the gently-sloping north slope at the right and the top of the very steep Kangshung face on the left. The rocks in the right foreground, trending toward the summit drifts, seem to be the crest of the rock ridge in the northeast of which the summit snowdrifts accumulate. An oxygen bottle and prayer flags are visible in the snow, perhaps a dozen feet below the present surface on the vertical northeast face of the summit snows!
This picture clearly indicates that there may be a lot of snow directly below Everest’s summit, as it blows up the gentle, rocky northwest side of the summit, then piles up on the almost-vertical northeast side. These objects that are buried in the snow were clearly left on the top a good many years ago. More snow drifted over them and slowly the weight of this snow has caused this mass of layered snow to settle, year after year, down the Kangshung (northeast) side of the top. These data are dramatically confirmed by Kurt Diemberger’s unique picture of Everest’s summit from the bottom of the Hillary Step, 300 meters south of the top, which shows a very large accumulation of snow below Everest’s peak at the crest of its Kangshung Face. This photo was taken on October 15, 1978.
I am therefore predicting that the top of Everest seems to be staying in pretty much the same place, but a little snow keeps being added to the summit each year or two—and, as this snow accumulates there, its weight is tending to make the summit slowly settle downward, even though the altitude of the top remains always pretty much the same.
If one could remove several feet of this huge summit drift one would doubtless find fascinating collections of things left there by triumphant summiters over quite a long period of years! I’ve asked Todd Burleson (who is going back to Everest in May 1995) to make a determined effort to actually get down to that oxygen bottle and prayer-flags and see if the date of the expedition that left them there can be determined. This would instantly reveal the rate at which this summit drift is accumulating. This 1995 team will also bring along a light tape and measure exactly how far down these “artifacts” are from the summit.
An interesting bit of information about cornices on the upper part of Everest is emerging from these studies of the snow on its summit. I have long been interested in the fact that none of the climbers who have reached the crest of the northeast ridge over the last seventy years have commented on cornices there. They have all agreed that the dropoff at the top of the Kangshung Face is dramatic, impressive. But they have all been able to reach the ridge’s crest, somewhat northeast of the First Step and look almost straight down—never a comment on a cornice. The 1993 Irish expedition took some remarkable pictures of the crest of this ridge, looking back, downward (northeastward) along the crest of the ridge—only a few significant cornices.
At the same time, if one looks up to the summit of Everest from its South Peak, the opposite is the case. Although the configuration of the cornices varies a lot from year to year, there are always dramatic cornices between the South Peak and the summit. On reflection, this is exactly what one should expect— because the great and well-known winter winds blow from the west-southwest. When our Learjet made its mapping flights over Everest in December 1984, it encountered 165 miles of wind from the west-southwest at 39,000 feet. These terrific gales sweep the summit of Everest more-or-less at right-angles to the ridge between the South Peak and the top and more or less parallel to Everest’s northeast ridge. The change in direction of these ridges is clearly apparent in our large-scale map of Everest’s summit. As one studies the snow-formation at the summit, the wind is clearly piling up a lot of snow on the Kangshung side, but it also appears to be swirling around the top in a counter-clockwise manner, which tends not only to build up big masses of snow there, but simultaneously to trim off the cornices that may begin to develop there. The snow that builds up on the northeast side of Everest’s summit seems to droop dramatically over this side, but does not seem to develop the extraordinary overhanging cornices that develop between the South Peak and the top, where the prevailing winds blow at almost exactly right-angles to the trend of the ridge.