ABSTRACT
The Uspallata Pass is the most famous Andes crossing. Once a trail on the southern fringes of the Inca empire, the pass became important under Spain--first as the link between Chilean settlements on both sides of the Andes, and later as the imperial transandean routeway between Buenos Aires and Santiago. Although steep, and regularly closed by snow, the Uspallata Pass provided the quickest, most direct route over the mountains between Mendoza and Santiago, since the upper valleys of the Mendoza and Aconcagua Rivers provide a high, but relatively straight east-west crossing. After independence, the summit of the Uspallata Pass marked the boundary between Chile and Argentina, and with it the beginning of a long period of divided, rather that shared, interests. Traffic through the pass declined, a decline which continued even after the construction of the Transandine Railway (1887-1910), the first rail crossing of the Andes. This paper examines both the use, and the neglect, of the Uspallata Pass over five centuries, and considers finally how recent changes in Chile-Argentine economic and geopolitical relations are bringing new life to the Uspallata Pass, and encouraging feasibility studies for an ambitious new transport project for this ancient route way.

The Uspallata Pass is the best known of all the Andean passes. Indeed, it is the only well-known pass throughout the entire Andean system, an extraordinary distinction given the vast length of the range. Identified also as the Bermejo, Los Libertadores, La Cumbre, and EI Cristo Redentor Pass, the Uspallata (3,855m) lies below the magnificent south face of Mount Aconcagua (6,960m), the highest peak in the Americas. The pass located one of the early Indian trails through the mountains, it carried an important highway link in the Spanish colonial period, and became the route of the first railway ever built across the Andes.

Even so, in the long term, the Uspallata Pass has been characterized more by neglect than by development. Its fitful use over the centuries has served to emphasize the limited importance attached to transandean movement in general, and to Chile-Argentine exchange in particular. The existence of 'ways through' can so often remain an abstraction. The significance of all crossings is based on what lies at each side of the connection, and on the advantages (if any) of facilitating the contact between them. Mountain 'passes' have no intrinsic value in human geography; they can stimulate growth in their adjacent foothill zones, as well as growth in more distant centers whose trade routes depend on the pass, or they can remain unused. In the Andes, many potential 'passes' have never earned the name.

New transport technologies can provide new opportunities, but the Uspallata Pass has already been the focus of such ventures in the past. In the 21 st century, the future of the Uspallata will still depend primarily on the nature of the relationship between Chile and Argentina, but now it will also be influenced by the outcome of dramatic geopolitical and macroeconomic developments that are currently underway in many parts of the Southern Cone and the Pacific Rim. Nevertheless, both geographical [end p. 35] and historical perspectives are necessary for any objective assessment of the present and future prospects of this region, and nowhere can these prospects be more clearly assessed in the long term than by consideration of the past successes and failures of the Uspallata connection.

TRANSANDEAN INITIATIVE: SPAIN'S IMPERIAL CROSSING
Before the Spaniards came, the high but relatively direct east-west mountain crossing provided by the upper valleys of the Mendoza and Aconcagua rivers had already led to the Uspallata Pass becoming one of the Incas' transandean routeways. The two main north-south Inca highways--one close to the Pacific, one through the highlands and eastern foothills-were linked at intervals by transverse roads. The Uspallata appears to have been the most southerly of these crossings, a line of movement which helped to demarcate the frontier between Inca and non-Inca territory. The remains of the official Inca rest-houses (Quechua, tampu) along the main highways led to the frequent occurrence in the Jujuy, Salta, San Juan, and Mendoza regions for example, of Spanish place-names such as tambo, tambillo, tambería,and tamberías (De Aparicio, 1963,1: 278; VIII: 321-39). The names Tambillos, Puente del Inca, and Laguna del Inca occur within the Uspallata Pass itself (Figure 1).

The early selection of the Uspallata Pass as Spain's official routeway across the southern Andes[end p. 36] had been based on the location of Santiago, Chile. After the founding of the city by Pedro de Valdivia in 1541, an approach to Santiago from the east was opened byFrancisco de Villagni, who on his return to Chile from a provisioning expedition to Peru in 1550-51 had traveled along the eastern side of the Andes before following the old Inca (Uspallata) trail westward over the mountains. Villagni was impressed with the eastern foothill zone, and anxious to extend Chile's interests east of theAndes. Claims were established there in 1551, while by 1559, Villagni had again explored the eastern slopes, including the valleys of the Neuquen and Limay. Governor Mendoza of Chile now began the founding of Chilean settlements on the eastern flank, via the Uspallata Pass: Mendoza in 1561, San Juan in 1562, and a third oasis site which was established at San Luis in 1594.

This transandean frontier region was known as Cuyo (Araucanian, sandy land), and remained part of Chile for the next two hundred years, although its trade in livestock and foodstuffs to Santiago, like Santiago's administrative contacts with Cuyo, was confined to the summer months when the Uspallata Pass was free of snow. The eastern approach to the pass was wide and relatively easy; the trail swung north from the extensive irrigated orchards, vineyards, grain land and pastures around Mendoza to reach Uspallata (Quechua, the settlement among the ashes). At this small mining center, where silver had been discovered in 1638, fresh supplies of alfalfa were available before the trail reached the Mendoza River and followed it upstream to the great rest area and ample pastures of Punta de Vacas (2,395m). So far, the gradients had presented few problems as the broad route way wound through the front ranges, composed for the most part of well-shattered sandstones, quartzites and conglomerates, ablaze in red, russet and ochre. Beyond Punta de Vacas, however, the trail steepened as it entered the zone of granites, metamorphics, and strongly cemented breccias around Puente del Inca (2,720m),w ile beyond Las Cuevas (3,151m), as the gorge narrows and steepens again severely in the andesites, basalts and slates of the Aconcagua region, the trail became so difficult and slippery that it was dangerous even for mules, and always vulnerable to avalanche, rockfall and landslip.

The descent from the summit of the Uspallata Pass was both short and steep-down the west Andean escarpment, with the trail confined to the Aconcagua River gorge for much of the way. The Aconcagua valley is deeply incised into well-folded, resistant conglomerate and breccia beds, while in some sections the valley walls close in to form spectacular ravines where the river cuts through clusters of granitic and porphyritic intrusions (Figure 2).

During the seventeenth and early eighteenth centuries, cattle, mules, salt, foodstuffs and wine continued to be dispatched westward over the Uspallata Pass from the assembly point of Mendoza, bound for Santiago or the mines of Peru. But Cuyo settlements had also established their own trade links along the east Andean foothill zone, notably with Tucuman. When the new Viceroyalty of the Río de la Plata was created in 1776, these links, reinforced[end p. 37] by the isolation from Santiago, including the 4-6 months closure of the Uspallata Pass by snow, resulted in the transfer of the Mendoza, San Juan, and San Luís provinces to Buenos Aires' authority. For the first time, the Andes became a continuous political divide between the regional spheres of Santiago and Buenos Aires --between the new Captaincy-General of Chile and the new Viceroyalty on the River Plate. Patagonia was allotted to neither: the boundary of the new Viceroyalty was on the Río Negro, while much of the far south of South America, from the Pacific to the Atlantic, remained unorganized (and unconquered) territory.

Strengthening the Imperial Crossing of the Andes in the Final Phase of Empire
At this stage, the political divide running along the Andes between Chile and La Plata was still an internal division of empire. Indeed, Spain's decision to promote Buenos Aires as part of the late eighteenth-century imperial reform actually boosted trade across the pampas to Mendoza, and from Mendoza to Santiago via the Uspallata Pass. Long-distance trade through Córdoba added tobacco and yerba maté from Paraguay, while Buenos Aires itself added imported manufactured goods to the traditional foodstuffs, fodder, hides, salt, and livestock being herded or pack-muled over the Uspallata Pass.

Greater protection and better maintenance of the imperial trade route were authorized by Spain, Buenos Aires and Santiago: the construction of forts across the pampas to reduce the dangers of Indian attack, the building and restoration of mountain shelters (casuchas, refugios) at intervals through the pass to assist couriers and other travelers, and the introduction of a regular system of repair and improvement of the trail itself after rockfalls and landslips. On the Chilean side, the most effective improvement of the Uspallata routeway was organized by Ambrose O'Higgins, Governor and Captain-General of Chile from 1788 to 1796. He supervised the construction and provisioning of casuchas on both sides of the Andes, eight of them through the pass itself (Figure 1), while at the Chilean entrance to the Uspallata Pass, O'Higgins founded in 1791 the 'gateway' town of Santa Rosa de los Andes. By the 1790s, about two-thirds of Chile's imports were coming across the Andes, although Peru was subsequently able to reassert its dominance in the Chilean markets. Nevertheless, an estimated 25,000 or so mules were involved in a prosperous transandean carrying trade via Uspallata at the end of the eighteenth century (Encina, 1946, V: 381-2).

Since the Uspallata routeway was open only in the summer months, the desire to speed the mails and extend business by finding alternative passes led Spanish officials to authorize fresh surveys (De la Cruz, 1806; Robinson, 1970: 34-38). The existence of other transandean passes was already known, several of them lower than the Uspallata, less steep, and less severely blocked by snow. But the Uspallata was the only pass in which the imperial mountain refuges had been constructed. When, in addition, the different approach routes at each end of the pass were taken into account, to say nothing of the vested interests in Mendoza and Santiago that were opposed to any change in the existing arrangements, it soon became clear at this late stage of empire that the Uspallata Pass had no rival as the official pathway of Spain.

The Defeat of the Spanish Empire: The Army of the Andes
The continental strategy devised by Argentina's San Martín for the liberation of South America focused attention once more on the Uspallata Pass. San Martín's plan to defeat Spain in Peru by first liberating Chile, had been prepared in the province of Cuyo between 1814 and 1817. In the summer of 1817, the vanguard of the Army of the Andes --Los Libertadores-- had moved from Mendoza into the Uspallata Pass, while other units were dispatched over passes to the north (Los Patos) and south (Planchón). San Martin's own account of the Andes crossing by a total of some 5,000 men, horses, mules, cattle, weapons and supplies, underlines the hazards of this daring project (Miers, 1826, I: 157-60; II: 10-14. Rojas, 1933: 153, 168-70).

The transandean assault took Spain's Chilean troops completely by surprise. The defeat of the Royalist forces at Chacabuco, just north of Santiago, proved to be one of the decisive battles of the revolutionary period. But in one sense, this was to be a final flourish for the Uspallata Pass, a dramatic ending to any further serious interest among Chileans or Argentines in an Andean crossing for the best part of the next sixty years. [end p. 38]

THE DECLINE OF LINKAGES AND THE FRUSTRATIONS OF THE "RAILWAY AGE"
Reduction in traffic and increasing neglect of the Uspallata Pass followed Argentina's and Chile's independence from Spain in 1810-18, which in adopting the existing colonial division confirmed the summit of the pass as the boundary between the two states.

The loss of unified control and maintenance of the Uspallata routeway proved critical, and evidence of its decline was reported by several early-nineteenth-century travelers (Schmidtmeyer, 1824; Caldcleugh, 1825; Proctor, 1825; Head, 1826; Miers, 1826; Haigh, 1829). Crossing the Andes by this historic pass was a memorable experience for all of them, and all commented on the dilapidated condition of the mountain refuge huts, once the distinguishing feature of Spain's Camino de los Andes. Trade had been severely affected, Schmidtmeyer noted in discussing Mendoza's surviving remnants of Spain's Asia-Pacific trade: "Some of the silk and cotton goods, which come directly from China and Bengal to Chile, are also brought over the mountains to this place. But its transit and carrying trade has considerably decreased, .. and it is likely to cease almost entirely, as the safety of the road cannot be relied on." It was now safer to trade round Cape Horn than across the Andes (Schmidtmeyer, 183-184).

One of the earliest and most vivid accounts is that of Lieutenant Charles Brand, R. N., forced by circumstances to make both a winter crossing to Chile, and a return summer crossing to Mendoza, in 1827. Struggling on foot from one casucha to the next, sheltering in "these hovels, miserable and wretched as they truly are," Brand nevertheless found them a godsend in the blizzards, which he found more frightening than any sea storm or desert sandstorm he had ever experienced. The sheer quantity of snow trapped in the Uspallata Pass was astonishing (Brand, 1828). Charles Darwin also recorded the deterioration in the Uspallata routeway when, in 1835, he crossed the Andes twice between Santiago and Mendoza--eastward by the Portillo Pass and westward by the Uspallata. Commenting on the eight casuchas in the Uspallata Pass, Darwin noted that "under the Spanish Government they were kept during the winter well stored with food and charcoal, and each courier had a master-key. Now they only answer the purpose of caves, or rather dungeons." There was evidence too that Mendoza had lost business: "The prosperity of the place has much declined of late years." Mendoza now had a "forlorn aspect" (Darwin, 1845).

The isolated mountain shelters in the Uspallata, tiny against the vastness of their setting, were often described individually in travelers' journals. They punctuated the ordeal; they highlighted the drama of the pass and the memory of empire. Lieutenant Isaac Strain, USN, for example, crossed the Uspallata in 1849 in the company of one of the mail couriers, whose tenuous links between Argentina and Chile had been maintained even though the huts now offered no more than minimum protection. Strain learned that the real value of the casuchas was in late autumn and early spring when the couriers sought to extend their summer service; but the risks were daunting.

"The courier with whom I crossed the plains to Buenos Ayres, had upon one occasion been shut up in a Casucha by a snow storm for some eighteen days, and was finally obliged to sally out and pursue his journey by the immediate danger of starvation ... Some merchants [in Chile] had offered him twenty ounces of gold to take a letter to Mendoza very late in the autumn, and he had undertaken it; but no bribe, he assured me, would ever induce him to renew the attempt, after the terrible experience he had already gained."
(Strain, 1853: 168).

In fact, the neglect of the route as a well-maintained throughway was highlighted by the limits of Chilean and Argentine jurisdiction. Although the summit of the pass remained the official boundary between Chile and Argentina, the whole transandean zone had become an effective no-man's land between two widely separated points of national authority. Chile maintained a passport control post during the summer at Guardia Vieja (Guardia del Resguardo/Guardia de los Hornillos), with year-round passport and customs control farther west at Santa Rosa de los Andes (Figure 1). In Argentina official control began and ended at the small guardhouse in Uspallata:

"Passports and goods are examined here at Uspallata, Argentina, for here its territory ends," "Proctor reported, westbound in 1823." ... "five leagues further [from the casucha ojo de Agua] we came to the Guardia del Resguardo, where the territory of Chili commences"
(Proctor: 62-63, 83). [end p. 39]

The failure of Argentina and Chile to join forces to maintain basic services in the pass was emphasized by Lieutenant MacRae, USN, on his official reconnaissance across the Andes and the pampas in 1852-54. This investigation formed part of a series of wide-ranging surveys of South America authorized by the United States Congress in the 1850s (Fifer, 1991), and in all, MacRae made four transandean crossings while carrying out his work. The famous shelters in the Uspallata were totally abandoned to the elements; the outside stairs that had once given access above deep snow, and the shelves along the inside walls that had enabled travelers to sleep above the floor, had all collapsed (MacRae, 1855: 5).

Mendoza's potential for agricultural expansion was severely hampered by the town's increased isolation and lack of market, MacRae reported. The dispatch of cargo over the Uspallata Pass was minimal, but alfalfa was the principal crop, and trailing stock westward remained a vital source of revenue. In 1851-52, about 14,000 cattle, 2,000 horses, and 600 mules had been sent to the Chilean market, although as usual, many perished on the way. The poor condition of the trail, prolonged exposure, and wind buffeting all took their toll. At the section near Las Cuevas, MacRae had passed slopes strewn with skulls and bones--"the remains of a large drove of cattle which was caught in a heavy snow-storm on its way to Chile." (MacRae,18)

Most of the inhabitants of Mendoza, however, were looking east, not west. As MacRae reported, Mendoza was waiting for the railroad, a rail link across the pampas to Rosario and Buenos Aires. In fact, the railroad did not reach Mendoza for over another thirty years, when the Argentine government's Andino railway was extended to Mendoza and San Juan in 1885.

The Revival of Interest in the Uspallata Pass: Proposals for a Transcontinental Railway
The first proposal for a transcontinental railway across South America was made by William Wheelwright in the early 1850s and did not involve the Uspallata Pass. Wheelwright, originally from Newburyport, Massachusetts, had spent many years on the west coast of South America and pioneered numerous urban and transport improvements there, including the founding of the Pacific Steam Navigation Co. in 1840, and the construction of Chile's first railroad (from Caldera to Copiapó) in 1851. This he regarded as the initial stretch of a transcontinental line across Chile and Argentina, via Córdoba, to Rosario and Buenos Aires (Wheel wright, 1861). Against the background of the USA's Pacific Railroad Surveys being undertaken at the time across the trans-Mississippi West, Wheelwright had already begun to extend the line beyond Copiapó. Significantly, he rejected the traditional overland link between Santiago and Buenos Aires via the Uspallata Pass in favor of a much longer northwest-southeast diagonal routeway, since one of the stated purposes of the transcontinental railroad was to promote the economic development of the interior. The Andes crossing was to be made via the San Francisco Pass, higher than the Uspallata but generally less steep, and free of the heavy snow blockage and deep drifting that bedevilled the Uspallata. Year-round overland connection between the Atlantic and Pacific Oceans was fundamental to Wheelwright's plan. In practice, however, the scheme was premature and too ambitious; it broke with tradition and attracted official support in neither Chile nor Argentina.

Earlier, in 1842, Wheelwright had also proposed to the Chilean government the construction of a railway from Valparaiso to Santiago, not as the first section of a transcontinental railway, but as a means of improving the link between the capital city and the port, and thus expanding cargoes and mail service on the Pacific Steam Navigation line. This proposal was approved, and Wheelwright's surveyor, Allan Campbell, completed the initial work in 1853. The Valparaiso-Santiago railway was eventually completed in 1863 and among its other achievements, had the immediate effect of reviving the idea of a transcontinental railway, although now it was to link Santiago and Buenos Aires. This proposal had a much more attractive ring to it--it emphasized the importance and prestige of the two capital cities. Buenos Aires had just been confirmed as the capital of the new Argentine Confederation (1862), while Santiago, having been completely omitted from Wheelwright's transcontinental project, was now to be the focus of the Chilean traffic. Not least, the proposed line had historical justification, since it followed the 300-year-old imperial routeway over the Andes. [end p. 40]

Work Begins on the First Transandean Railway
The project was now advanced with great enthusiasm by two Anglo-Chilean brothers, John and Matthew Clark, whose company (Juan y Mateo Clark, Santiago and Valparaiso) had completed the transandean section of the telegraph link between Santiago and Buenos Aires in 1872 (via the Uspallata) and then, in 1872--4, had been awarded the concessions in both Chile and Argentina to construct the transcontinental railway by means of two lines across Argentina (Buenos Aires-Mendoza and Mendoza-Chilean frontier), and a line in Chile running up from the main network to meet the Argentine line at the mountain boundary. After consideration of some neighboring passes, the Uspallata alone gained general support, despite the Clarks' misgivings about gradients and snowfall.

The Uspallata Pass had proved to have an inexorable hold on the Andes Crossing. The route of the telegraph line had undoubtedly influenced that of the railway, while the telegraph route in turn had been influenced by the line of the old mail and messenger trail, which was stil1 in use. Working in the pass to comp lete the final section of the telegraph, however, John Clark had been left in no doubt about the severity of the snow conditions, and not only in winter. As the United States Minister to Chile reported at the time, after a personal investigation of the Uspallata and Portillo passes: "A violent [summer] snow storm continued for several days, making the summit impassable for a time, and preventing several thousand head of cattle from continuing their march from the Argentine Republic to Chile, thereby causing great loss of stock-vast numbers of these starved and frozen animals becoming food for the condors." (Root, 1871).

John Clark was a particularly strong supporter of Wheelwright's belief in the need for transcontinental rail linkage across southern South America. Indeed, the two men had been together at the celebrations in July 1872 to mark the completion of the transcontinental telegraph over the Andes. Clark now worked tirelessly to raise funds, but it was not until 1886 that the Buenos Ayres & Valparaíso Transandine Railway Co. was registered in London, with responsibility for the construction of a 177-kilometer, meter-gauge line from Mendoza to the Chilean border, via the Uspallata Pass. Work began in 1887, and in 1891 the first 91 kilometers were opened, the track making a wide southern loop alongside the Mendoza River to reach the front ranges, rather than following the northern loop of the old Uspallata trail (Figure 3). The line was extended to Punta de Vacas in 1892-93, but then work was suspended; the Clark brothers had run out of capital, and the Baring crisis saw the collapse of the London money market. Construction was resumed in 1899 and continued until the problems associated with the international Summit Tunnel again brought the work to a halt.

Meanwhile, the Clark brothers had also begun building in Chile in 1889, after engaging a Swiss mountain-railway engineer to design the most difficult sections. Costs soared. Three years later, as noted above, the Clarks were bankrupt, having invested considerable resources of their own to keep the work going when the main company fell into difficulties. The Chilean portion of the Transandine Railway was much steeper, and relatively more expensive, than the Argentine. The Chilean government had already extended a branch from its Longitudinal Railway eastward from Llai-Llai as far as Los Andes (the old Santa Rosa de los Andes), which at 833m., stands on the great alluvial gravel fan of the Aconcagua River, at the entrance to the Uspallata Pass. This was to be the starting-point of the Clarks' enterprise, with the Chilean Transandine closely following the line of the trail up the Aconcagua valley towards Juncal and the summit of the Pass (Figure 4). The railway was to comprise 70 kilometers of meter-gauge track, thereby making a total length of 247 kilometers narrow-gauge across the Andes between Los Andes and Mendoza. There was no through connection with the Chilean or Argentine rail networks at either terminal, since both Los Andes and Mendoza were on the broad gauge of 1.676 m. (5 feet 6 ins).

Shortage of funds was to cancel the original plan for more tunneling on the Transandine Railway. Eventually, six sections of rack-rail were required on the Chilean side, and another on the Argentine. Only one major tunnel was built-between Las Cuevas and Caracoles--the relatively short 3. 2-km. bore which crossed the international boundary at La Cumbre, 665 m. below the summit of the Uspallata Pass. Work on the tunnel began in 1906, and the line was finally opened in April 1910, in time for the Chilean President to use it on his journey to Buenos [end p. 42] Aires for the official centenary celebration of Argentine independence.

The Problems of the Transandine Railway
The entry of this historic corridor into the railway age did not, after all, mark a dramatic improvement in Chile-Argentine exchange. From the start, the Transandine Railway was unsatisfactory both as a transandean and as a transcontinental connection. Apart from the lack of uniform gauge and administration, the problem of seasonal closure was no nearer solution than it had been in the days of the Spanish empire. Insufficient tunneling, combined with insufficient snowsheds and snow ploughs, often resulted in annual closure of all or part of the Transandine Railway in Mayor June for up to six months at a time. Rockfalls, mudflows, and flooding added to the crippling maintenance costs. Although one commentator rejoiced in 1910 that the railway heralded "a reopening of ancient paths," the ancient problems of the Uspallata Pass were still there.

The route soon became little more than a minor tourist attraction, with passengers accepting mule, horse, or stagecoach transport through the mountains, wherever sections of the track were closed. It was part of the Andes adventure, and harked back to the years when 'end-of-track' excursions had been organized for foreign visitors to view work in progress on the line, before transferring to horse or stage to continue the sight-seeing journey to the summit of the pass. John Clark was an enthusiastic promoter of such excursions while he remained in charge of operations, as more than one visitor recorded in travel notes (e. g. Pereira, 1894). Between 1904 and 1910, trips to, or over, the pass included a viewing of the Christ of the Andes statue (EI Cristo Redentor), which had been erected jointly by Chile and Argentina at the summit of the Uspallata Pass. With the completion of the tunnel, however, this point of interest was completely lost to view from any point on the railway. Passengers passed beneath it, and a bell was rung on the train as it went through the tunnel to mark the crossing of the boundary between Chile and Argentina. "The Christ statue is buried in snow half the time, and out of sight the whole of the time," wrote one disappointed traveler.

Before long, motor transport and especially air transport, were competing successfully with the Transandine Railway. At best, Buenos Aires and Santiago had been placed within 36 hours of each other by rail, but the break of gauge, the early lack of coordination in train timetables, and the frequent delays or closure of the Transandine, made it a slow and unreliable form of transport. PANAGRA (PanAmerican-Grace Corp.) and NYRBA (New York, Rio and Buenos Aires) had both introduced airmail flights between Buenos Aires and Santiago, via Mendoza, in 1929, and passenger service by PANAGRA and Lufthansa was expanded in the 1930s. Ominously for the railway, the peak period for passenger movement on the Transandine had been the two opening years--351,962 1st- and 2nd-class through-passengers in 1910-11, and 349,930 in 1911-12. After that, quite apart from the effects of World War I, the novelty factor began to wear off. Unlike railroads in the United States, there was very little company promotion or advertisement to create and expand business, and fewer than 18,000 through-passengers were recorded in 1920. Passenger[end p. 43] volumes fell steadily in the 1920s, and by the early 1930s 1st - and 2nd-class annual passenger traffic had slumped to around 16,000 (Fifer, 1991: 96-97).

Freight revenues were even worse. Freight was always second to passenger traffic as the Transandine's source of income-mostly Argentine flour, together with small quantities of timber, minerals, miscellaneous foodstuffs and domestic items. An annual average of less than 38,000 tons was carried between 1910 and 1932, and less than half that figure by 1934, when heavy financial losses over many years were coupled with a catastrophic season (Johnson, 1943: 180-1; Annual Company Reports, passim). Severe snowstorms, avalanches, flooding and mud-slides in 1934 destroyed or badly damaged many bridges, tunnels, stations, and track, resulting in the closure of the Transandine as a through service for more than ten years. The Transandine's fundamental problems, however --physical, political, and economic-- were never solved. After further years of intermittent closure, the Transandine Railway finally closed in 1978-79, a century after the Clark brothers obtained the original concessions for its construction. The steepest parts of the track have been torn up or simply abandoned (Figure 5). In Chile, a 40-km. section is still in use, linking the Saladillo copper mine to Los Andes, while in Argentina, part of the railway is occasionally open for tourists traveling between Mendoza and Puente del Inca.

In addition to the problems of the transandean line itself however, the problems of the line as the vital link in a transcontinental railway had been growing since the 1880s. Indeed, much of the early conviction that the Buenos Aires-Uspallata-Valparaíso crossing would become a development axis across South America had been sapped by a growing sense of the futility of the enterprise.

Even before the delayed completion and dismal performance of the Transandine (as well as tariff barriers) had reduced the commercial potential of new trade with Chile, the effect of terminating Argentina's broad-gauge system at the foot of the Andes had proved decisive. The completion of the rail link in 1885 had turned Mendoza and San Juan firmly towards the east. Cuyo wines, fruit, vegetables, alfalfa, and even livestock now found their market around the River Plate, as the railway narrowed the physical frontier of separation between the Cordillera and the Littoral that had formerly been created by the dry pampa. The British Buenos Ayres & Pacific Railway Co. however became the dominant factor in the regional transport geography after its founding in 1882. By 1907 the company had established a controlling interest in the Argentine foothill zone by taking over the rapidly expanding broad-gauge network in the Cuyo provinces owned by a rival British company, and also by agreeing to manage the Argentine Transandine Railway (formerly known in the Clarks' day as the Argentine section of the Buenos Ayres and Valparaíso Transandine Railway), then nearing completion. In 1909, the Buenos Ayres & Pacific finally consolidated its control of the Cuyo traffic by acquiring the Andino railway west of Río Cuarto-- the pioneering line that had first brought the railway[end p. 44] to Mendoza and San Juan in 1885. Such unified control of what had been four separate railway companies could have had some beneficial effect in fostering Argentine-Chilean exchange, but in practice, despite its name, the Buenos Ayres & Pacific made no sustained attempt to capitalize on its through-route to Chile nor, as a matter of policy, to boost freight and passenger traffic on the international line. The Company's crucial objective had always been to eliminate rail competition to and within the central Argentine foothill zone, and to service the Buenos Aires domestic and export markets. The physical and political problems of working the mountain railway reinforced this view, and in 1923, the Buenos Ayres & Pacific rid itself of the 1907 agreement to manage the Argentine Transandine (Fifer, 1991: 95-99).

During the railway age therefore, the international boundary between Chile and Argentina had become more divisive, more consciously Andean. Paradoxically, the Transandine Railway over the Uspallata Pass had served to separate rather than to join. While Chile and Argentina concentrated on their own national rail networks, the Transandine became a drain on resources, and increasingly irrelevant to two states which basically had little or no use for it.

GEOPOLITICAL TRENDS IN THE LATE TWENTIETH CENTURY
New geopolitical and economic imperatives have revolutionized parts of southern Latin America since the 1970s. Once again, Chile has pioneered the movement, since Chile's emergence during the last twenty years as a new and spectacular economic growth zone has exploited its location both on the Pacific Rim and as the western edge of the Southern Cone.

While Chilean exports, and return investment, have been flowing around the Pacific, Chile in turn has been investing increasingly in Latin America, particularly in Argentina. Indeed, Chile is now the largest foreign investor in Argentina, with private Chilean investment rising from US$50 million in 1990 to more than US$500 million in 1992. These two neighboring states, whose relations since independence have been marked more by distrust than by collaboration, now have a bilateral trade exchange of over US$l billion.

By any reckoning this is a dramatic turn of events. The economic collapse of Argentina in the 1980s after forty years of decline, however, confirmed that economic restructuring could not be confined west of the Andes. The basic strategy for Argentine recovery was to be 'the Chilean Model' --massive privatization and freer trade. Argentina suddenly found itself in the unfamiliar position of seeking cooperation with Chile (Fifer, 1994).

One important aspect of the growth in Chile-Argentine trade has been the substantial increase in freight traffic over the Uspallata Pass, traffic which has emphasized the fact that throughout their 5000-km. Andean border, this old routeway is still the only major overland connection between Chile and Argentina. The focus of this paper on the Uspallata's 'Andes Crossing' is not symbolic: the Uspallata "is" the Andes Crossing. The bulk of the cargo movement between the two states remains seaborne, but 90-95 percent of the total transandean traffic between Chile and Argentina uses the Uspallata Pass (Figure 6). The number of freight trucks from Brazil is also increasing, since Brazil, like Argentina, now seeks direct access to the Pacific ports, and the Pacific Rim, via the old Andes crossing. Customs control at the Uspallata Pass (Los Libertadores) recorded a total, two-way traffic of over 83,000 freight vehicles in 1992--a 22 percent increase over the 1990 figure. The total two-way tonnage of goods crossing the pass in 1992 was 1,194,143 tonnes--a 29 percent increase on the 1990 figure (Interview, 1993).

The bilateral agreements concluded between Chile and Argentina in 1991 (Acuerdo, 1991; Tratado, 1991) made improved transport via the Uspallata (Cristo Redentor) Pass a centerpiece of the agreement--the creation of a new multimodal corridor involving new railway and tunnel construction. Farther south, a new transandean gas pipeline from Argentina's Neuquen field was approved. Chile currently imports about 85 percent of its petroleum requirements and, in addition, needs to expand the use of natural gas in the Santiago metropolitan area, for industry and pollution control. But the proposal for a multimodal corridor implies a determination to tackle the gradients, rockfalls, mud-slides, and snow accumulation which have always characterized the Uspallata Pass. Although the copious snowfall has fostered the growth of ski resorts at Portillo (Chile) and Los Penitentes [end p. 45] (Argentina), it continues to cause long delays to winter traffic, and occasionally to close the route altogether, despite the expensive efforts by teams of snowploughs to keep open one lane of the highway. The old Transandine Railway tunnel today carries the highway under the crest of the pass.

The basis of the proposed improvement to this old Andes crossing thus rests on the construction of a new tunnel, provisionally estimated at 21 kilometers, between Puente del Inca and Juncal (Figure 3). This long rail tunnel (which might also carry another gas pipeline) would form part of a new transandean railway, which envisages the replacement of the original narrow-gauge track between Mendoza and Los Andes with a broad-gauge, fully electrified line for year-round service. The new tunnel, like the new rail track, has been designed for double-stacked container trains, and for passenger trains with automobile-carrying service. In the context of the 1991 transport proposals, plans have already been prepared by the Houston-based PRECO group which, with good reason, regards the scheme as one of the world's major 20th-century tunneling and railroad projects. In fact, PRECO has linked the transandean project to staged outline proposals for a new growth corridor across the Southern Cone, involving seaport expansion in Chile, new handling facilities at Mendoza, and improved navigation on the Parana-Paraguay (i. e. the existing international regional project Hidrovía, which plans to develop the Paraguay-Paraná river system into a 3,442-km year-round navigable waterway). At this point, it is worth recalling that with the obvious exception of pipeline and air transport, the other components of this entire transcontinental project for integrated road, rail, and river transport were all originally proposed by United States engineers surveying the Southern Cone in the 1850s, although the Uspallata Pass was not recommended by any of them as the first choice for the Andes Crossing. Indeed, the faint shadow of late-18th century Spain is still discernible in the continued selection of the Uspallata as the major overland connection in southern South America between the Pacific and Atlantic Oceans.

Parts of this latest multiple project have attracted immediate attention. Feasibility studies for the new tunnel began soon after the completion of the 1991 Accord, which was designed to promote economic exchange (Figure 5). Efforts are underway to form a multinational investment consortium with initial private-sector financing. This is still an ambitious, if overdue, project, although it is one whose central task began to be undertaken in Europe well over a century ago. The first of the great Alpine tunnels, the Mt. Cenis (Frejus), 12 kms.long, was opened in 1871; the St. Gotthard (14.9 kms.) in 1882; the Arlberg (10.2 kms.) in 1884; the Simplon (19.3 krns.) in 1906; and the Lütschberg (13. 6 kms.) in 1912. In fact, the Simplon remained the world's longest rail tunnel for almost the whole of the twentieth century.

As in Europe, North America, and elsewhere, the long-term economic and political advantages of such connection are weighed against the heavy initial construction costs and ongoing maintenance. The implementation of the latest rail-and-tunnel project for this old Andes Crossing will depend in the final analysis on the extent of the need and the desire to facilitate international trade within the Southern Cone, and on the degree of urgency among the Atlantic states to gain overland access to the Pacific ports, and Chile to reach those on the Atlantic. Indeed, it is the possibility of faster access to the Pacific Rim that attracts most of the interest among the states of MERCOSUR---the Southern Cone Common Market comprising Brazil, Argentina, Uruguay and Paraguay. Argentina has already expressed the hope of acquiring facilities at as many as nine Chilean ports. Chile has so far declined the invitation to join MERCOSUR (where it holds observer status) until the bloc's state export subsidies and protectionist policies have been abolished, or significantly reduced. Instead, Chile continues to expand its trade in Latin America through bilateral agreements, including those with Argentina and Brazil. But Chile is also on track to be the first new state to join NAFTA, if the existing three members remain willing to extend the Agreement.

CONCLUSIONS
For nearly five centuries, the main thrust to use, and improve, the Uspallata Crossing has almost always originated in Chile. Though punctuated by long periods of neglect, the sixteenth-century founding of the Cuyo settlements, the eighteenth-century trail improvements by O'Higgins, and the nineteenth-century railway construction by the Clark [end p. 46] brothers, were all initiated in Santiago. In the early twentieth century, as the Panama Canal neared completion, Chile undertook extensive survey work in the Andes to promote new transcontinental railroads into Argentina. Several suitable passes were found and mapped, but the project was dropped. No support for Pacific-Panama linkage existed in Buenos Aires, or in the boardrooms of the British-owned railway companies in Argentina (Fifer, 1991: 99-102). This proved decisive.

In practice therefore (despite periodic, mostly verbal threats of attack or invasion), ever since independence Chile and Argentina have rationalized the Andes as a 'natural' boundary between them. Time and custom harden attitudes, earlier chances for greater flexibility are lost. For the Andes are not a 'natural' political boundary; indeed in many respects, the survival of the boundary between Chile and Argentina, particularly in the way it has operated, has been Latin America's most damaging boundary problem. Unlike frontiers in the tropical deserts, savannas and rain forest, those in the temperate South were effectively never changed by war, annexation, arbitration, or purchase. The polarization of Chile-Argentine interests --of Pacific-Atlantic interests-- discouraged mobility and the chances of productive cooperation at local level in the nineteenth and twentieth centuries, and it also severely restricted opportunities for economic development and complementarity across the Southern Cone, especially within the interior. As late as the start of the 1990s, for example, combined investigation of the mineral resources along the Chile- Argentine borderland was prohibited, while Argentina banned Chilean investment from a zone 40 kms. inside the Argentine boundary.

The survival of the Andean international boundary in southern South America has had more than a regional impact; it has been one of the greatest impediments to the successful geopolitical and economic development of the entire subcontinent. Recent changes reflect both internal and global movement, and it is too early to predict how far some of these changes will be consolidated. Given the [end p. 47] powerful influence of history rather than geography in the development of Hispanic America, no one should underestimate the scale of the required shiftin the traditional attitudes of both Chile and Argentina in significant cooperation is to be achieved in the long term (Fifer, 1994).

Never before, however, have Chile and Arrengina jointly acknowleged the importance of achieving new levels of growth within their own economies through the improvement of access and exchange across the Andes--in goods, services, investment, energy, and infrastructure. As part of this increasing exchange between South America's Pacific and Atlantic worlds, the Uspallata Pass is still, after five centuries, the crucial Andes Crossing.

REFERENCES

Acuerdo de Complementación Económica entre la República Argentina y la República de Chile, 2 August, 1991; Tratado entre la República Argentina y la República de Chile sobre promoción y protección recíproca de inversiones, 2 August, 1991. (Buenos Aires and Santiago de Chile).

Annual Reports, The Argentine Transandine Railway Company Ltd.; The Chilian Transandine Railway Company Ltd. (London).

Brand, C., RN. 1828. Journal of a Voyage to Peru: A Passage Across the Cordillera of the Andes, in the winter of 1827, performed on foot in the snow; and a Journey Across the Pampas. (London).

Caldcleugh, A. 1825. Travels in South America, during the years 1819-20-21; containing an account of the present state of Brazil, Buenos Ayres, and Chile, >2 vols. (London).

Darwin, Charles. 1839, 1845. Journal of Researches into the Natural History and Geology of the countries visited during the voyage of HMS 'Beagle' round the world. (London).

De Aparicio, F. and Difrieri, H. A. 1963. La Argentina: Suma de Geografía, 9 vols. (Buenos Aires).

De la Cruz, L. 1806. Extracts from the Journal of Don Luís de la Cruz, employed by the Government of Buenos Ayres in conjunction with that of Chili to open a communication between the two Viceroyalties through the intermediate Indian Tribes, 1806, with map, Archives, Royal Geographical Society, London.

Encina, F. A. 1946. Historia de Chile desde la prehistoria hasta 1891, 20 vols. (Santiago de Chile).

Fifer, J.Valerie. 1991. United States perceptions of Latin America, 1850-1930: a 'New West' south of Capricorn? (Manchester and New York: Manchester University Press).

____ . 1994. "Chile's pioneering location: Pacific Rim and Southern Cone," Geography, vol. 79: 129-146.

Haigh, S. 1829. Sketches of Buenos Ayres and Chile. (London). Head, F. B. 1826. Rough Notes taken during some Rapid Journeys across the Pampas and among the Andes. (London).

Johnson, E. R. 1943. The Railways of Argentina, US Department of Commerce. (Washington, DC).

MacRae, A., USN. 1855. The US Naval Astronomical Expedition to the Southern Hemisphere during the Years 1849-50-51-52, vol. II. (Washington, DC).

Miers, 1. 1826. Travels in Chile and La Plata, 2 vols. (London).

Pereira, G. 1894-95. Notes of a Tour Round South America, Archives, Royal Geographical Society, London.

Proctor, R. 1825. Narrative of a journey across the Cordillera of the Andes, and of a residence in Lima, and other parts of Peru, in the years 1823 and 1824. (London).

Robinson, David J. 1970. "Trade and Trading Links in Western Argentina during the Viceroyalty," Geographical Journal, vol. 136: 24-41.

Rojas, R. 1933. El Santo de la Espada: Vida del General San Martín. (Buenos Aires).

Root, J. P. 1871. Root to US Secretary of State, Santiago, 15 February, 1871. Diplomatic Despatches from US Ministers to Chile, General Records of the Department of State, Record Group 59, Washington, DC.

Schmidtmeyer, P. 1824. i>Travels into Chile, over the Andes, in the years 1820 and 1821. (London).

Strain, I. G., USN. 1853. Cordillera and Pampa, Mountain and Plain. Sketches of a Journey in Chili, and the Argentine Provinces, in 1849. (New York).

Wheelwright, W. 1861. "Proposed Railway Route across the Andes, from Caldera in Chile to Rosario on the Parana, via Córdova," Journal of the Royal Geographical Society, vol. XXXI, 154-162.

RESUMEN
EI Paso Uspallata es el cruce más famoso en los Andes. Una vez un pequeño sendero en la franja sur del imperio Inca, que se vuelve importante bajo el mando español primero como el eslabon entre la colonización chilena, y más tarde como la ruta imperial transandina entre Buenos Aires y Santiago. Después de la Independencia, la cima del Paso Uspallata delimitó el límite entre Chile y Argentina, y fue el principio de un largo período de intereses no-compartidos. EI tráfico a través del paso disminuyó y continuó decreciendo aún después de la construcción del primer ferrocarril transandino (1887-1910). Los recientes cambios económicos y geopolíticos en Chile y Argentina, han brindado una nueva vida al Paso, y han fomentado estudios sobre un nuevo y ambicioso proyecto para esta antigua ruta. [end p. 48]