In the autumn of 1884, a carpenter stood on a wooden trestle above Loop Creek in the Selkirk Mountains of British Columbia, driving spikes into a timber nailer at a height that left no practical option if his balance failed. The creek below was not far, but it was far in the way that mattered on a high construction platform. The platform was not level. Its planks were freshly sawn softwood, their surfaces still wet and rough underfoot.
Below the platform, the timber bridge that he and his crew had spent the previous days building and raising extended across a succession of columns, caps, and sway braces down to mudsills set on the creek bank and bed. The bridge was not yet complete. It was one of the structures still needed before the railway could cross here, and the railway had to cross here. It was coming from the east.
The men stringing telegraph cable were somewhere behind him on the Rogers Pass route, and the schedule William Cornelius Van Horne had set for completing the Canadian Pacific Railway’s main line through the Selkirks allowed no delay. The Selkirk Range is not the Rockies. The two are separated by the Rocky Mountain Trench, a long north-south valley that forms the structural divide between the mountain systems. A traveler crossing British Columbia from east to west meets the Rockies first, descends through Kicking Horse Pass to the trench at Golden, then climbs straight into the Selkirks up the Illecillewaet Valley toward the summit at Rogers Pass.
The two crossings are different in character. The Rockies at Kicking Horse are steep and open, their peaks rising above timberline, the valley floor wide enough in places to give a sense of space beneath the rock walls. The Selkirks at Rogers Pass are steep and walled in. The slopes are covered with dense forest to their upper reaches, Engelmann spruce and subalpine fir.
The peaks rise above ten thousand feet. The valleys between them are narrow and shaded. The snowpack that accumulates in the Selkirks through the winter months is among the deepest in North America. The Illecillewaet Valley above Glacier receives snowfall measured in metres, not inches, between November and April.
And that snowpack does not melt on any dependable schedule. Spring in the Selkirks is not a season of gradual warming but a season of competing processes, snow building in the high elevations while meltwater runs in the valley bottoms. The stability of the snow cover on the steep slopes above the valley deteriorates through the winter, then gives way in sudden events that carry no warning and follow no timetable that anyone standing below them could read in advance. The route through Rogers Pass had been under consideration since CPR survey parties began studying possible crossings of the Selkirks in the late 1870s.
The Canadian Pacific Railway had been incorporated in 1881 to fulfill the political commitment the Dominion government had made to British Columbia as a condition of the province’s entry into Confederation in 1871. British Columbia had joined Canada on the promise of a transcontinental railway to the east within ten years. By 1881, ten years had passed without a railway, and British Columbia had long since run out of patience. The CPR charter required a main line from central Canada to the Pacific coast.
The route through the mountains required crossing both the Rockies and the Selkirks. Kicking Horse Pass through the Rockies had been identified by the Palliser expedition in 1858. No pass through the Selkirks had been identified by any government survey. Major Albert Bowman Rogers, the CPR’s locating engineer assigned to cover the interior, located the pass in 1881 during a journey through the Illecillewaet drainage, a journey recorded in the survey records and in Rogers’ own account as arduous exploration through difficult terrain.
The pass that now bears his name sits at 4,534 feet above sea level, at the divide between the Illecillewaet and Beaver drainages. It is not a low pass, and it is not an easy one. It was the only pass in the Selkirks through which the CPR’s transcontinental line could be brought. The decision to route the railway through Rogers Pass rather than the Howse Pass to the north, which had been explored in earlier surveys and offered a lower crossing, rested on the engineering of the transcontinental line as a whole.
The CPR main line was designed to serve not only the coastal terminus at Vancouver but also the agricultural settlements of the prairies and the commercial development of the interior. A route through Howse Pass would have required a much longer line and a different relationship to the prairie settlement corridor. The Rogers Pass route was shorter but more difficult. Van Horne, the general manager brought from American railroading to drive the CPR’s construction at a pace matching the company’s financial position, understood both the difficulty and the necessity.
The line would go through Rogers Pass, and whatever that required would be done. Work on extending the line through Rogers Pass began in earnest in 1884. The contracting arrangements for the British Columbia section of the CPR main line were organized under Andrew Onderdonk, an American contractor from a New York engineering family who had won government construction contracts for the western sections beginning in 1880. Onderdonk was responsible for the contract for the hardest mountain section of the line.
He was responsible for laying the railway from the coast up the Fraser Valley to the interior. By 1884 his operations had extended through the valley to the mountain approach to the Selkirks, and work had begun on a bridge over the Illecillewaet stream at Rogers Pass. The problem of the timber bridge at Rogers Pass was not simply the distance between supports. Crossing the Illecillewaet Valley above Glacier was not a simple matter.
The valley was narrow, and the slopes on either side of the pass were so steep that the locating engineers worked under grade constraints that determined where and how the railway could be laid. The CPR main line across Rogers Pass was built with a maximum westward grade of 2. 2 percent, a rise of 2. 2 feet for every 100 feet of horizontal distance.
This was not the grade the engineers would have chosen. It was the result of terrain that offered no flatter option for crossing the pass at the height required to clear the valley walls without a tunnel that the 1884 construction schedule and budget would not allow. A 2. 2 percent grade was one that the locomotives of 1884 could climb with a suitable consist.
It also restricted the weight and length of trains operating over the pass. As long as the line operated at that grade, a constraint the CPR’s operating department recognized from the first season of service and recorded in the ledgers of every train movement over the mountain pass for the next thirty years. The loop at Rogers Pass was an engineering response to a specific problem in the alignment of the Illecillewaet Valley. The valley floor rose sharply over a short horizontal distance on the approach to the summit.
The railway had to climb through that section while holding the adopted 2. 2 percent grade that the engineers had set as the maximum for practical operation. A direct line up the valley floor would have exceeded that limit. The solution was to increase the horizontal distance over which the climb was distributed by laying the railway in a full circle, a curve that carried the line over itself on a bridge structure, gaining elevation on the outer portion, then turning over the inner portion and continuing toward the summit on the required course.
This loop added distance, which reduced the effective grade over the climb. The structure carrying the outer portion of the loop over the inner portion and over the course of Loop Creek was the timber bridge the carpenter was building in the autumn of 1884. It was not a large structure by the standards of the mountain crossings the CPR would later build. It was the structure required for the grade at the required location.
Without it, the track around the loop was not practicable. Without the loop, the grade over Rogers Pass exceeded the capability of the locomotives of 1884. The timber bridge over Loop Creek was a wooden structure of the type North American railway contractors had been building in the mountain west since the 1860s. The construction technique drew on the prevailing carpentry of the era: timber frames, sill plates, mudsills, posts, caps, sway braces, and deck beams, the essential vocabulary of a railway-building industry that had spent a generation crossing the mountains of Nevada, Utah, and the Sierra Nevada, and now the Selkirks.
The timber was available in the surrounding forest, where Engelmann spruce and subalpine fir grew to large diameters at the valley elevations where the construction camps stood. Some of the timber for the bridge came from the trees the construction crews felled to clear the right-of-way. The bridge’s structural design followed the standard arrangement of posts and bracing used in railway bridge construction since the continental crossings of the 1860s, adapted to the height and span requirements of the Loop Creek site. The foundation conditions at Loop Creek were not those of temporary embankments for crossing a major river.
The creek ran clear and cold over a gravel and sand bed in the valley bottom. The bedrock of the Selkirks lies close to the surface at most valley locations. The bridge’s mudsills rested on a timber grillage laid over the bearing material the construction crews had excavated. The creek passed beneath the structure in a trestle opening within the bridge frame.
The foundations were a problem that could be handled. The height above the creek bed and the grade the track had to maintain around the loop were the dimensions that governed everything else. The construction crews raised the bridge’s bents by hand, using long poles and ropes. Building a timber bent for a bridge of Loop Bridge’s height was heavy work.
The posts were of large-dimension timber, squared to twelve inches or greater, cut from the surrounding spruce and fir to the required length and roughly dressed to specified dimensions. The sill plates joining the bases of the posts and the cap plates joining their tops were of similar cross-section. The sway braces between the posts were of lighter material, but in total each bent was heavy enough that raising it from the ground to the vertical required coordinated force applied at the right points in the right sequence. Block and tackle provided mechanical advantage at the critical moments of rotation, while the raising poles guided the bent through its path from horizontal to vertical.
The men at the base held the heel at the intended pivot point while the raising gang lifted, and the gang at the top steered the bent’s head into the vertical. When the bent stood vertical, it was braced into place and the gangs moved on to the next one. The timber bridge rose from the valley floor in the order the bents were raised, each new bent adding height that had to be accounted for on the platforms above the creek. The men doing this work were Chinese.
The majority of the construction workforce on the British Columbia section of the CPR main line was Chinese. Onderdonk’s payroll records for the mountain section recorded a workforce of roughly six thousand Chinese labourers at peak hiring on the Rogers Pass approach alone. The men arrived at the construction site through a labour organization that was structurally parallel to the padrone system. Twenty years later, Pennsylvania railroad contractors would use Chinese labour contractors on the Rockville Bridge, though working through different populations and a different geography.
Men working within the commercial networks of Victoria and San Francisco, who had established connections with the labour migration networks coming out of Guangdong province in southern China, recruited the workers, arranged transportation, negotiated contracts with the construction company, managed the work camps, and supplied the camp stores from which the workers bought their food and living necessities at the construction site in the Selkirks. The wage paid by the construction company was paid to the contractor. What reached the worker was what remained after the contractor ran his organization, housed his men, and supplied his camp at the price the camp store allowed. Chinese workers on the Rogers Pass section were paid one dollar a day for ordinary labour.
White workers performing similar physical labour on Onderdonk’s contracts were paid between $1. 50 and $2. 50 a day, depending on classification and season. This difference was not a secret; it was the stated wage structure.
Chinese workers were classified in a separate labour category created and maintained by the contracting party. The wage assigned to that category was what Chinese workers were paid in the British Columbia labour market of 1884. The men who built the timber bridge structures above Loop Creek, set the timber, laid the deck planks, and drove the spikes on the railway above a creek in the Selkirks were paid one dollar a day for their work. The gap between that figure and the physical conditions and bodily risk of the work is not a gap that requires documentation.
The record records the wage, and the wage is what is recorded. The men who came to Rogers Pass from Guangdong province to British Columbia came for the same structural reasons that would bring men from southern Italy and Sicily to the Susquehanna Valley a generation later. The agricultural economy of the Guangdong delta had been under pressure since the 1840s. Population growth on a fixed land base, crop failures in bad monsoon years, the upheaval of the Taiping Rebellion through the 1850s and 1860s, and the structural poverty of a rural economy integrated into the world market in ways that drained value without returning it all pushed emigration from the Pearl River delta across the Pacific for thirty years before CPR construction began in British Columbia.
The men came looking for wages. The wages paid in British Columbia in the 1880s for railway construction, mining, and logging work, for men willing to do the work the labour market classified as unskilled, were the wages strong men who spoke no English and had no Canadian capital could earn. They were wages barely above subsistence. Enough to live on, and in good years to send some money to the village in Guangdong where the family remained.
Not enough to save meaningfully while the body was consumed by the work. The conditions at the Rogers Pass construction site in the autumn of 1884 were the conditions of the Selkirks in October and early November. The high mountain snowpack that would accumulate in the Illecillewaet Valley over the following months was beginning to build on the peaks above the construction pass. Temperatures in the Selkirk valley dropped below freezing regularly at night in October, and approached freezing in the afternoons on days when cloud built against the hills and the sun did not reach the valley bottom.
The Chinese workers in the construction camps on the Rogers Pass route were housed in accommodation provided by the Onderdonk contracting firm: timber structures adequate to keep the workers on site and fit for work, not designed for the coming winter. The CPR schedule required that as much of the Rogers Pass section be completed as possible before the winter of 1884-1885 closed the work down. What was built before the snow fell would still be standing after it melted. What was not built would have to be built in the following season.
The following season was the season in which the transcontinental line was supposed to be complete. Provisioning the workers through that winter was a matter for the contracting system. The company store run by the Chinese contractor near the construction camps sold foodstuffs at prices the store could charge at a site with no competing supply. Rice, dry goods, and the staple foodstuffs the workers depended on were available at the store.
Medical care for injured or sick workers on site was available at the level a construction camp budget provided, which was not a high level in 1884 on a mountain section of the CPR. Scurvy was documented among Chinese workers on the British Columbia sections of the CPR, a disease of nutritional deficiency that occurs when eating from limited sources. The camp store at a mountain construction site of the CPR in 1884 was not stocked to fight scurvy; it was stocked with what the contractor could profit from supplying. The avalanche problem at Rogers Pass was not a problem the CPR’s engineering department had fully accounted for in the original construction design.
Survey records documented snowfall and terrain steepness. They did not include systematic documentation of avalanche paths because the survey parties had passed through the region in summer and early autumn and recorded what they saw in those seasons. Avalanche paths on the slopes above the Illecillewaet Valley at Rogers Pass do not show clearly in summer. But the terrain that shapes them is evident: steepness, aspect, elevation, and the absence of old-growth trees on paths where avalanches periodically clear the forest.
The survey parties of 1881 and 1882 were working to find a route through the mountains, not to document the behaviour of the snowpack above the route they found in winter. The avalanche behaviour showed itself in the first operating season. Snow slid from the steep slopes above the railway, crossed it, buried it, and struck train cars and work crews on the line. In the severe winters of the late 1880s and early 1890s, avalanches on the Rogers Pass line killed a large number of CPR employees, a number documented in the company’s accident records at the level of incidents and fatalities, and in some cases at the level of victims’ names.
The winter of 1899 saw a series of avalanches that killed a number of company workers at Rogers Pass, prompting a formal engineering review of the protective measures available on the line. The company’s response to the avalanche problem was the construction of snow sheds: timber structures, roof-like constructions built over the track in known avalanche paths, designed to deflect the sliding snow mass over the roof and down the slope on the lower side, rather than through the track and whatever was on it. The snow sheds at Rogers Pass were a sustained construction and maintenance project that ran from the first operating seasons through the end of the nineteenth century and into the first decade of the twentieth. At their peak, more than thirty separate snow sheds protected different parts of the Rogers Pass line.
They were timber structures built and rebuilt as avalanches damaged or destroyed them, and as the CPR’s understanding of the most active paths developed through accumulated operating experience. The men who maintained the snow sheds on the Rogers Pass line through the winter were maintenance workers in the CPR’s track department, working on the most avalanche-prone section of main line in North America. Their names appear in CPR employee records, at the level those records were kept for section workers. The names of the avalanche victims appear in the accident records, and in some cases in local newspaper reports.
The names of the Chinese construction workers who built the timber bridges that the snow sheds were later built to protect appear in the construction accounts, at the level those accounts were kept for tracking the costs of the Onderdonk contract. Lopp Bridge carried the CPR’s mainline traffic from its completion in 1885 until the end of the original Rogers Pass line’s service life. The trains it carried in the CPR’s early years were the trains of the railway’s first decade. These included the passenger service between Montreal and Vancouver that the CPR had been built to provide, and the freight service generated by prairie settlement and the early development of the interior economy as the years passed.
The locomotives of 1885 that first crossed Lopp Bridge were the CPR’s initial equipment, machines whose axle loads were designed to be carried by the bridge’s structure at the operating speeds the mountain grades allowed. As locomotive weights increased through the 1890s and into the early twentieth century, and as trains grew longer and heavier, the demands placed on the bridge were no longer the demands its structure had been designed to carry. The CPR’s bridge engineering department inspected the timber structures on the Rogers Pass line as locomotive development outpaced the original design loads. Structures were reinforced and rebuilt wherever engineering analysis required.
The Lopp Bridge structure was maintained throughout the surface line’s operating life, the railway managing what could be managed. The operating problem on the Rogers Pass line was not any single structure but the grade, curvature, and avalanche risk that characterized the entire original route over the pass. The 2. 2 percent grade that the terrain had forced on the engineers of 1884 required helper locomotives on heavy westbound freight movements and careful handling in the eastbound descent.
The cost of helper service mounted with every train movement over the pass in every year of operation, a figure the operating accountants recorded consistently and the general manager’s office reviewed periodically. The avalanche risk required continuous investment in snow sheds and maintenance labour to keep those sheds safe through the winter. The combination of grade difficulty and avalanche risk made the surface Rogers Pass line an operating liability that the CPR management recognized from the first season of service, and the engineering department had been studying a permanent solution for twenty years. The solution was a tunnel.
The Connaught Tunnel, a single-track bore driven through the heart of the Selkirk Range below the summit of Rogers Pass, eliminating the highest and most avalanche-prone slopes by passing beneath them, opened in December 1916. The tunnel is 5. 02 miles long. At its completion, it was the longest railway tunnel in North America.
The tunnel reduced the maximum grade on the Rogers Pass line from 2. 2 percent to 1. 0 percent westbound, a reduction that transformed the operating economics of the mountain crossing. It also eliminated most of the avalanche risk that had dogged the original surface line.
After the tunnel was completed, the trains that had crossed the original route over Lopp Bridge and the other timber structures built by Chinese workers in 1884 and 1885 were diverted to the tunnel. The original surface line over the pass went out of mainline service. The abandonment of the original Rogers Pass route took it out of the CPR’s operating jurisdiction. The rail was lifted.
The bridge structures along the abandoned right-of-way, including the structure at Lopp Creek, were no longer subject to railway maintenance. The forest that had been cleared to build the line in 1884 began to reclaim the disturbed ground. The timber structures, no longer carrying traffic and no longer maintained, began the process of returning to the material they were made from. The Selkirks are not a dry environment.
Wood decays in the Selkirks at the rate wood decays in a wet mountain climate under the freeze-thaw cycle of the subalpine zone, which is faster than in a dry climate. Lopp Bridge was built in the autumn of 1884 from timber cut from the surrounding forest, carried traffic for thirty-one years. Then it was left to the weather and the fungi that favour it. The driving of the last spike of the Canadian Pacific Railway took place at Craigellachie in Eagle Pass on November 7, 1885.
The photograph taken at Craigellachie that morning shows Donald Smith, the CPR’s principal investor, driving the last spike into the final joint of the transcontinental main line. The photograph also shows railway officials, investors, and invited guests gathered around the spike site, along with the locomotive of the train carrying the official party to the ceremony. This photograph has been reproduced in every account of CPR construction since it was taken, standing for the moment of the transcontinental line’s completion. The Chinese workers who built the hardest parts of the line through British Columbia, who dug the tunnels in the Fraser Valley with hand drills and nitroglycerin, who cleared the route through the Selkirks in the seasons before the construction crews arrived, who built the bents over Lopp Creek, raised them, and fixed the rails to the deck of the line, are absent from the Craigellachie photograph.
They were not present at Craigellachie because they were not invited. The completion ceremony was organized for railway executives and their guests. The construction workers who built the line were not among the invitees. The absence of the Chinese workers was, moreover, an absence of numbers: by the time of the ceremony, many of them had already left the construction area.
The work was done, the contracts were closed, and the labour force the contracting system had assembled for the construction season was no longer needed. The men returned to the labour market they had come from, to Victoria and New Westminster and San Francisco, looking for whatever work was available to Chinese labourers in the Pacific coast economy of 1885. The Chinese workers who built the British Columbia sections of the CPR had worked under a contract that provided no benefits, no compensation for injuries beyond what the contractor determined, and no recognition of the physical effort of the work beyond the fixed daily wage. When the Canadian government passed the Chinese Immigration Act of 1885, the same year the transcontinental railway was completed, imposing a head tax of $50 on every Chinese person entering Canada, the men who built the railway were among the categories that tax was designed to restrict.
Fifty dollars was the equivalent of fifty days’ pay at the rate paid on the Rogers Pass section. The relationship between the completion of the work and the restriction of the workforce that completed it was not a coincidence requiring analysis to understand. The men who built the line understood it, as did the Chinese community organizations in Victoria and Vancouver that documented the head tax and organized campaigns against it. This has not escaped the studies of CPR labour history over the past century.
Ownership of the original Rogers Pass right-of-way passed to the Canadian Parks Service when Glacier National Park was created and its boundaries extended to include the historic pass. Today the original line is maintained as a walking trail through the park. The grade of the original railway is still evident in the terrain, and the locations of the original structures are documented in the park’s interpretive materials. Lopp Creek is there.
The site of Lopp Bridge can be located by the grade, the course of the creek, and the interpretive documentation of the original alignment. The bridge itself is not there. What remains is the site of the bridge, the grade that required it, and the creek still running below the level where the bridge timbers once lay. The CPR operates the Rogers Pass crossing today through the Connaught Tunnel, and since 1988 through the Mount Macdonald Tunnel as well.
The second tunnel was built to provide capacity for the double-stack container trains required by modern intermodal traffic. The current Rogers Pass route carries some of the heaviest and largest freight traffic on the North American rail network. The mountain that resisted the railway for thirty years now carries trains through a combined 14. 7 miles of tunnel.
The grade that Chinese workers worked around on the bridge curves at Lopp Creek, which helped the original line overcome it at 2. 2 percent, is now handled underground at 1. 0 percent through the tunnel that replaced the surface line. The men in the construction camps of 1884 built that line.
If you follow the abandoned railway right-of-way from the Illecillewaet camp in Glacier National Park, you walk on the grade of the CPR’s original main line. The trail follows the original alignment west toward Rogers Pass. You are walking where Chinese construction workers graded the roadbed in 1884. At the loop, a specific part of the route, the trail turns around itself at the place where the original line crossed over itself to gain elevation above the creek without exceeding the maximum grade the engineers had set.
Below this loop, Lopp Creek runs over the gravel of the Selkirk valley, as it ran when the timber bridge stood over it. The timber is gone, but the trail remains, the creek remains, and the mountains remain. The snow accumulates on the slopes above the old trail every winter, as it accumulated in the winters when the line was operating and the trains crossed on the surface built by the Chinese carpenters in the autumn before the CPR’s first winter in the Selkirks. There is no monument at Lopp Creek identifying the Chinese workers who built the timber bridge.
The interpretive plaque at the trailhead refers to the original CPR alignment, records the period of construction, and explains why the surface route was replaced by the Connaught Tunnel. The engineering achievement is recorded on the plaque. The Chinese workers who erected the bents, raised them, fixed the rails, and laid the deck of the railway in the autumn of 1884 are recorded in the CPR records at the level those records were kept for tracking the costs of the Onderdonk contract. They are also recorded in the head tax records of 1885, those who remained in Canada and paid the $50, at the level those records were kept for tracking payments.
And they are recorded in the records of the Chinese Benevolent Association and community organizations in Victoria and Vancouver at the level those organizations kept their membership records. But they are not named on the interpretive plaque. A CPR freight train crossing Rogers Pass today through the Mount Macdonald Tunnel carries containerized freight between the port of Vancouver and the interior markets of North America. The locomotive at its head weighs more than the entire train carried by the first CPR train across Lopp Bridge in 1885.
The mountain above the tunnel is unchanged, and the creek below the old trail still runs. The bridge is gone. The men who built it are recorded in the records of a contracting system that required their labour costs to be tracked. Their names are not in those records at the level required to identify them.
The last spike is in the photograph at Craigellachie, and they are not in the photograph. The line they built through the Selkirks is now a walking trail in a national park. The trains that replaced it run 600 feet beneath the valley floor through a mountain, crossed by the men in the construction camps of 1884 the only way the season, the budget, and the schedule allowed. They built it in the autumn before the first winter.
The winter came. The bridge stood. The trains crossed. The winters came again.
Thirty-one winters. Then the tunnel opened, the line closed, and the timber returned to the forest. The creek runs on as before.
