The Great Arc

by John Keay

  ‘Those only,’ says an 1850s manual on Indian surveying, ‘who have had any practical experience with such delicate and expensive instruments as Mercurial Mountain Barometers can be fully aware of the disappointments met with in a country like this.’ Indeed, barometers were so prone to breakage that almost any alternative method was reckoned to be worth a try.

  The manual in question recommended a simpler ‘thermometrical’ solution, which was precisely that to which Hodgson and Webb duly had recourse. All that was needed in this case was a pocket-size thermometer and a kettle. You boiled the kettle and when, as Hodgson put it, the water ‘reached full ebullition’, you took its temperature. Since the boiling point decreases as the altitude increases, reference to a simple conversion table would give your height. Accuracy depended on very careful reading of a well-calibrated thermometer, and could not be counted on to within less than a few hundred feet. It could also be a problem getting your kettle to boil. Snow and ice, often the only form of water available at high altitudes, took time to melt, let alone ‘reach ebullition’. Meanwhile the high winds common to exposed locations toyed with the fire, blowing it either sideways, out, or sometimes away. Nevertheless it is curious that this comparatively simple test seems to have been very little used until Hodgson and Webb adopted it as standard practice.

  In the summer of 1816 Hodgson worked north, triangulating where possible, perambulating where not, and planting potatoes as he went. (The potato had been rightly identified by the then Governor-General as having an important contribution to make to Himalayan diet and economy.) But, ‘having some doubts as to the precise latitude of my grand station on The Chur on which everything depends’, Hodgson returned there in October and again in 1818 after an adventurous foray to the most distant of the Ganges’ sources. Near the top of The Chur he had constructed a thirty-foot pyramid from which, as well as from Saharanpur in the plains and from Surkananda to the east of Mussoorie, he plotted the angles of some fourteen major peaks and many more minor ones. In fact The Chur for Hodgson, as later for Everest, became a second home. He wrote of it with feeling – although not in situ, since the ink invariably froze to the pen.

  At a place like the summit of this vast mountain no one who has not resided in such a stormy region can have an idea of the violence of the wind, and the suffering of an observer by night from the cold … On the 10th of October, water instantly froze when poured out at 9 in the morning, tho’ the sun shone out, the thermometer being then at 31°. Judge then what it is by night, accompanied by a wind which peels the skin from the face, and blows with a violence which seems to shake the very ground. I had a tent to protect the instrument [his theodolite] but at the time of observation the wind rushes in and shakes the instrument, and blows out the lights and creates confusion, and people holding the tent to prevent its being carried away are apt to touch or shake the stand [of the instrument] so that I found it impossible to keep the adjustments in order from night to night.

  Nor was this the worst of Hodgson’s problems. The latitudes of his Saharanpur-Chur base as established by zenith observations simply could not be reconciled with those computed by triangulation from a third point. Trying other points from which to complete the triangle made no difference. And the discrepancy remained when his colleague and friend James Herbert carried out a check. It was said that if experienced observers, taking all possible precautions, found themselves confronting an anomaly for which they could not account, they were probably ‘on the verge of some important discovery’. Hodgson quoted this maxim with approval and was perfectly candid about the discrepancy, hoping that others might someday be able to identify what it was that he had discovered.

  In fact, of course, they already had. But Webb, Hodgson and Herbert seem to have been totally unaware that mountain masses exercise an attraction over the plumb-line (so critical in zenith readings). Nor were they aware that, as per Lambton’s conjectures, the earth’s density might have a completely contrary effect. Moreover, they were in the worst possible place for this distortion. A geodetic paper on the subject would later confirm that the plumb-line deflection amounted to fifteen seconds of a degree at Saharanpur and at least thirty-six seconds at The Chur. The resultant discrepancy would be something like a third of a mile in the sixty-one miles of Hodgson’s base, a quite unacceptable ratio of error which, according to the Survey’s historian, ‘rendered all Hodgson’s care and labour of no avail whatsoever’.

  To resolve the problem, Hodgson did in the end revert to his original plan of a ground measurement. But, after long exposure on the hill tops, his health was no longer up to the task and it therefore fell to his assistant and successor. To James Herbert in 1819 belongs the honour of measuring the first Himalayan base-line.

  Herbert was clearly a resourceful individual. Although he had no experience of base-line measurement, had never even witnessed it, and lacked much of the necessary equipment, he entered into the business with enthusiasm. His only complaint was that he had no assistant. Lambton reckoned that at least three Survey officers were essential to oversee such an operation and Everest, for his Calcutta line, would muster no fewer than ten officers. Herbert was alone except for his Indian staff, none of whom knew the procedures any better than he, and none of whom he could trust to take a reading. How many times he had to scuttle from one end of each measurement to the other he does not say, but it must have been many.

  A site was found in the Dun which, though on a slope, was as near flat as anywhere in Garhwal and Kumaon, and the necessary vegetation clearance was soon underway. Meanwhile Herbert wrestled with the means of measurement. He had a chain but no coffers, no tripods and no elevating screws with which to level the tripods. Assembling and making these things would be the work of months. Reserving the chain for comparison purposes, therefore, he decided to use rods. A well-seasoned roof joist of Himalayan cedar was removed from a derelict house and sawn into four lengths, each of twenty-five feet. William Roy had tried rods on his Hounslow base, trussing them to reduce bowing and fitting the ends with ivory markers. Herbert used brass buttons for the critical contacts but, with his rods being only 13/4 inches square and trussing being out of the question, he was much concerned how to prevent them from bowing. In the end he devised a retractable stand on which the rods would rest, and ordered thirty-seven of these, one being required every six feet. As a visual check on bowing he also tensioned a thin brass wire which ran the hundred-foot length of his rods in a specially cut groove.

  A bench incorporating a brass standard was designed for the daily check on the expansion or contraction of each rod, and various markers, pickets and flagstaffs were also improvised. All in all, and considering that everything except the rods themselves had to be made out of unseasoned pine, Herbert was not displeased. In the final measurement of about four miles he reckoned that the error could not exceed two feet, ‘an uncertainty which will only affect the distances of the remotest peaks by about sixty or seventy feet’.

  Once the carpenters had finished, the whole operation took no more than a month. It then remained only to connect the base by triangulation to Hodgson’s problematic Saharanpur-Chur-Surkananda triangle. This done and the doubtful zenith readings of the triangle corrected, the observations previously conducted to all the snowy peaks could be adjusted and reduced to a table entitled ‘Latitudes, Longitudes and Elevations of principal Peaks and Stations in the Survey’. It was published in 1822, just as an apoplectic Everest was laying out his Sironj base-line on the Great Arc.

  Of the forty-six Himalayan peaks in the Garhwal-Kumaon region which Hodgson and Herbert had located and listed, only fourteen were less than 20,000 feet above sea-level. Five were over 23,000 feet, and three of these giants, which Hodgson had numbered as ‘A1’, ‘A2’ and ‘A3’, were far enough east to have also come within Webb’s Garhwal survey, with which they provided a useful link. Webb too had remarked the height of this threesome and had produced very similar values. At Hodgson’s 25,749 feet, ‘A2’ proved the
biggest of the three; and as Herbert discreetly put it in a note printed so small as to be barely legible, ‘A2 is thus, so far as our knowledge extends, the highest mountain in the world.’

  Reviewing the Garhwal and Kumaon surveys, Andrew Waugh, Everest’s successor as Surveyor-General, would find them ‘highly creditable to the scientific ability of the officers employed’, although he greatly regretted the ‘inartistical’ nature of their maps. Everest himself applauded Herbert’s efforts to determine a base-line and rated the Garhwal survey ‘quite sufficiently accurate for geographical purposes’. Where not duplicated by his own triangulations, it would be incorporated into the Great Trigonometrical Survey. He was less happy with Webb’s effort in Kumaon. Lacking a base-line, he thought it ‘little better than what is called a wheel and compass survey’.

  Webb’s observations had, though, been good enough to win a rather oblique retraction from the Quarterly Review. Tucked away in an 1820 notice of a recent paper in French ‘On the Height of India’s Mountains’, the Review’s contributor cited findings lately received from Webb. Evidently Webb, incensed by what he had read at Kedarnath in the earlier Review, had quickly set up his barometer. The readings, forwarded to London, convinced the reviewer that Kedarnath stood at 12,000 feet above sea-level and that a nearby pass into Tibet was over 16,000 feet. Both places were snow-free at the time. According to Alexander von Humboldt, the distinguished author of the paper under review, who had made extensive observations in both South America and Europe, such altitudes should be well above the level of perpetual snow and must therefore be exaggerated. Earlier the Review had also used this argument. Now it conceded that snow-lines were not necessarily consistent and that its conclusions had been ‘as erroneous as those of the Baron de Humboldt’. Moreover, if the heights of such observation points were correct, then so probably were those of the peaks observed from them.

  Webb’s sightings, back in 1808, of Dhaulagiri from south of the Nepal border were still suspect. The distance had been too great, the angle too low, refraction too uncertain, and his instrument too primitive. Even Webb now agreed. Crucially, the distance between the places from which he observed the mountain and calculated its position was also suspect, in that it had been deduced from zenith observations just like Hodgson’s at Saharanpur and The Chur. Until the Central Himalayas in Nepal could be approached as closely as the Kumaon and Garhwal Himalayas, or until such time as they could be observed from points in the plains whose locations were more precisely known, such long-range observations were reckoned a waste of time.

  But the heights given in Herbert’s 1822 listing for Garhwal looked good. They had been grounded on his base-line, however crude, and they had been carefully observed at comparatively close quarters from triangulated positions whose altitudes had sometimes been verified barometrically. ‘A2’ therefore went unchallenged as the world’s highest peak. Curiously, it also went unnamed. Herbert’s list included a Mount Moira (after Governor-General Lord Moira) and Mounts St Patrick and St George, all clustered around Gangotri. They had been so christened by Hodgson, either in an onset of patriotic fervour or in a gesture of Anglo-Christian bravado. Contemporaries disapproved and the names have since been expunged from the map.

  But for ‘A2’, neither Herbert nor Webb had proposed a name. Observing it from a distance of about fifty miles, they had been unable to ascertain whether there was already a local name for it. Moreover, they probably felt disinclined to suggest one. For were it to remain the world’s highest, this honour would almost certainly be contested by their superiors.

  Such anxieties, however, proved groundless. In the course of time it emerged that it did indeed have a local name – Nanda Devi – and that it was not the world’s highest. But Nanda Devi is, still, the highest mountain in India (as opposed to Nepal, Tibet and Pakistan-controlled Kashmir); and at 25,645 feet its official height today is only a hundred feet less than that given by Herbert.

  NINE

  Through the Haze of Hindustan

  The outermost limit of the Himalayas in Garhwal is defined by the Siwalik hills, a brown sierra of modest height which intervenes between Mussoorie’s alpine slopes and the baking Indian plains. Here, on the morning of 12 November 1833, a small army might have been seen breasting the ridge and, with the dew still on their boots, slithering down parched and gravel-strewn gullies, past Saharanpur whence Hodgson had begun his Garhwal survey, and on into the dusty immensity of the Gangetic plain. After an interlude of eight years, George Everest was taking the field again; and with its Superintendent back at the helm, the Great Trigonometrical Survey was launching itself into the dreaded haze of Hindustan. Departing Hathipaon with two assistants, three sub-assistants, four elephants, forty-two camels, thirty horses and ‘about 700 natives’ – in that order – Everest was about to address what he considered the most difficult terrain ever to be triangulated, let alone trigonometrically surveyed to the exacting standards of the Great Arc.

  During the previous season he had done little more than assess the challenge. From his well publicised base-line measurement in Calcutta, he had journeyed west to Sironj, where the Arc had been abandoned seven years previously, and had then reconnoitred north through Gwalior, Agra and Delhi along the line which it now must follow. On the way, he had met up with Olliver, Rossenrode and some of the newly recruited assistants. All were already marching and counter-marching in search of hills, mounds, barrows, buildings, anything with a view which might serve as a trig station. But the season was exceptionally dry, the atmosphere a ‘pea soup’, and the heat soon became like nothing Everest had ever experienced. One of his parties was shot at, another robbed. All reported numerous fever cases.

  Rossenrode, currently rated the most reliable of veterans, had managed well in the more undulating country south of Agra; but the dependable Olliver had let Everest down on the section thence to Mathura (Muttra); and a man called Boileau, lately transferred from topographical surveying, had been an unmitigated disaster wherever he was deployed. Everest had hoped to complete the selection of stations by May 1833. In fact it had barely begun.

  Continuing north, he had found that whereas up to Delhi he might get away with building towers on only one side of the Arc, from there onwards every station would require a tower. According to the co-ordinates worked out for places in the plains by Herbert and Webb using their peaks as ‘fixes’, the 78-degree meridian passed straight up the ‘two rivers’ region between the Jumna and the Ganges. Known in Hindi as the Doab, this was a congested flat even by the horizon-choked standards of the Hindustan plain. Up to Delhi, the Arc could advance with one ‘foot’ on rising ground to the west of the Jumna. Thereafter it would have to cross to the eastern bank and plunge both ‘feet’ into a cloudy murk whose floor was strewn with villages and encumbered with trees.

  Fuming over the season having been such an abject failure, Everest had then retired for the monsoon to his new property in the hills. It was his first visit to Hathipaon. There were leaks to be fixed, workshops to be built and accommodation to be organised. As the monsoon battered the ridge, Everest’s spirits revived. In a make-or-break gesture, he vowed to take the field personally at the head of his entire establishment as soon as the rains were over. Hence the cavalcade which crossed the Siwaliks in November 1833.

  They headed straight for Mathura, halfway between Delhi and Agra. There and elsewhere Everest had left instructions for the stockpiling of timber, bamboo poles, ropes, blocks and pulleys. These items were now divided up into numerous cart, camel and elephant loads and despatched to depots along the line of the Arc.

  Triangulating such impossible terrain involved different procedures to those followed by Lambton when droog-hopping across Mysore or by Everest himself when working up through central India. A simple linear progression in which an advance party chose the best hills while the Superintendent, following along behind with the Great Theodolite, took the angles, was out of the question. ‘I was about to make my first essay in a new career,�
�� wrote Everest, ‘wherein all my former experience would avail me but little.’

  Except where some existing eminence invited attention, stations in the plains could be located almost anywhere. The skill lay in finding the least objectionable position, somewhere sufficiently distant from, and at an appropriate angle to, other stations whose lights might eventually be sighted through the dense atmosphere after a minimum in the way of tree-felling, house-demolition and ground-levelling. But just how much clearance would be necessary depended on where the sight-line would actually fall. Would lopping a single branch solve the problem? Or would a whole village have to be moved? With sight-lines grazing the ground for up to thirty miles, it was hard to tell.

  It was therefore essential to predetermine the direction of each line mathematically. A chain of small triangles conducted between the two positions could provide this information, but the procedure was time-consuming and often abortive. Better was the system developed by Everest which he called ‘ray-tracing’. For this a party advanced along the supposed line with a perambulator and compass, traversing to left or right at exactly ninety degrees in order to work round any obstructions, and likewise in order to home in on the final position. From the angle followed, the traverses involved and the distances recorded, the true bearing of the sight object could be calculated. Telescopes and heliotropes could then be trained accordingly and the ‘ray’ duly cleared of obstructions. It was an extremely trying business. Yet the cost of a mistake, like erecting a tower where it would not serve its purpose, was unthinkable.