On the morning of January 14, 1898, a man named Ernest Hawkinson loaded a horse-drawn sled with four hundred pounds of salt pork, flour, lamp oil, and a box of medical supplies, and drove it out onto the frozen surface of Lake Superior from the pier at Bayfield, Wisconsin. The temperature stood at twenty-three degrees below zero. The ice beneath the sled runners measured two feet thick in the shallow passages between the Apostle Islands, less in the deep channels, and nowhere was its thickness marked, because there was no one whose job it was to mark it. Hawkinson had made this crossing before.
He knew the route the way people know roads that have no map, by the ridges the ice forms when it freezes over open water, by the sound of the runners on ice of varying depth, by the particular color the ice takes on when the depth of the water beneath it shifts from safe to uncertain. He was headed for Madeline Island, the largest of the Apostles, where a settlement of several hundred people had been waiting since the last ferry run in November for the supplies that a winter road across the ice could bring them. The ferry had stopped when the ice began to form. The ice road began when the freezing grew thick enough to bear weight.
Between those two events, Madeline Island was cut off from the mainland in a way that no modern equivalent can quite describe. There was no telephone. There was no radio. There were no airplanes.
There was the island, and two miles of Lake Superior water between it and Bayfield. And in November and early December, when the ice was forming but not yet solid, there was nothing at all. The islanders had planned for this. They stored what they could before the freeze, but planning only extends as far as what you can anticipate.
Illness does not respect the ice calendar. Children are born without consulting the lake. Salt pork runs out before the calculations say it should. Hawkinson drove the sled across two miles of open ice on a January morning in the coldest winter of the decade, and delivered what the island needed.
His name appears in no official record of that crossing. It appears in a letter his wife wrote to her sister in Milwaukee. The letter describes the cold, the weight of the load, the sound of the ice beneath the horses, and the fact that he came home safely, which was not guaranteed, and which she knew was not guaranteed when he left. This is the story of the ice roads of the Apostle Islands, winter routes that connected a region the rest of Wisconsin could not reach by any other means for four months of every year.
It is the story of a transportation system that moved timber, fish, supplies, and people across one of the most dangerous bodies of fresh water on earth for nearly a century before the highways replaced it. It is the story of the communities that depended on it, the deaths that resulted from it, and the natural formations it left behind: the sea caves of sandstone along the Apostle Island shore that freeze into something otherworldly in deep winter, drawing a hundred thousand visitors a year who know nothing of the ice roads that once ran past them. To understand what the Apostle Islands were like in the winter of 1898, you have to understand what northern Wisconsin was in 1898, and what it was not. Wisconsin had been cutting its northern forests since the 1850s.
The timber that covered the upper third of the state, mostly white pine mixed with fir, hemlock, birch, and maple, was among the most commercially significant timber stocks in the United States. Mills at Eau Claire, Chippewa Falls, and Marinette had been running since before the Civil War, processing logs that came down the rivers each spring and turning them into lumber for houses, commercial buildings, and railroad ties for a country that was consuming wood faster than anything else. By the 1880s, the most accessible timber had been cut. The industry moved north into the Chequamegon region of Ashland and Bayfield Counties, toward the shores of Lake Superior, and onto the Apostle Islands themselves, where timber that had not been commercially available a decade earlier had come within reach thanks to rail lines that reached Ashland in 1877 and Bayfield in 1883.
The Apostle Islands are a group of twenty-two islands scattered across the western end of Lake Superior off the tip of the Bayfield Peninsula in northern Wisconsin. The largest, Madeline Island, is about fourteen miles long and three miles wide. The smallest is barely large enough to support a stand of trees. The islands sit in water that is twenty to forty feet deep in the shallow passages between islands, sixty to a hundred feet deep in the main lake channels, with the surrounding open lake reaching depths of more than seven hundred feet.
The islands are made of Precambrian sandstone and basalt, formed six hundred million years ago when this part of North America was a shallow tropical sea. Waves shaped the soft rock over ten thousand years to create formations along the shore that have no equal anywhere else in the Great Lakes: arched caves, rock columns, and undercut cliffs. In summer the islands could be reached by boat from Bayfield in minutes. In winter, when Lake Superior froze in its western arm between the peninsula and the islands, they could be reached across the ice.
In the transition seasons, November and early December when the ice formed, and March and April when it broke up, they could not be reached at all. The Ojibwe people had understood this geography for centuries before white settlers arrived. The name the Ojibwe gave to Madeline Island is Moningwunakauning Minis, the island of the yellow-shafted flicker. The island had been a center of Ojibwe life in the western Great Lakes for generations.
The Ojibwe habitually crossed the ice in winter, moving between the islands and the mainland along routes known through experience and inherited community knowledge. They traveled lighter than the commercial operators who later used the same routes, on foot or by dog sled, in ways that the winter surface of the lake made possible for those who understood it. They understood that the ice was not a uniform surface. It was a landscape with its own features: pressure ridges where ice sheets had collided and piled up, open leads where wind or currents had separated the surface, clear smooth ice where water had frozen undisturbed, rough snow-covered ice where drifts had accumulated and concealed what lay beneath.
Safe travel across the winter lake required reading those features accurately and adapting accordingly. The Ojibwe had developed the knowledge to do this across generations. The commercial operators who arrived in the 1880s did not have that knowledge, and they acquired it at a cost the Ojibwe had already paid. The logging operations that came to the Apostle Islands in the 1880s required a system of supply and transport that summer navigation could not fully provide.
Island timber had to be cut in winter, when the ground was frozen hard enough to be skidded by horse-drawn sleds without sinking into the swamps and bogs that covered much of the island interiors. Winter-cut timber was hauled to the shore and stacked on the ice to await the spring thaw, when it could be floated or towed to mainland mills. The men who did the cutting lived in camps on the islands all winter, supplied from the mainland across the same ice routes they used to move timber. A logging camp of forty men required a steady supply of food and equipment across the ice from Bayfield.
The ice road was not optional infrastructure. It was the factor that made the islands commercially productive. The road itself was built, to the extent that it was built, from its first use of the season. The teamster who drove the first load of winter across a fresh ice surface broke the trail.
Subsequent use packed the snow and compressed the ice surface into something like a road without quite becoming one. The route ran from Bayfield to Madeline Island, two miles across the passage between the tip of the Bayfield Peninsula and the western shore of Madeline, over ice that ranged from eight to twenty-four inches thick depending on the severity of the winter and the exact location of the crossing. Crossings to the outer islands, Outer Island, Stockton Island, and Oak Island, passed over deeper water and required more careful assessment of conditions before each crossing. Ice thickness was measured with an auger.
The person who wanted to know whether the ice would bear the load he intended to take across it drilled a hole by hand and measured the thickness with a ruler. Four inches of ice would hold a man. Eight inches would hold a horse and a light load. Twelve inches would hold a horse-drawn sled with a full commercial load.
Eighteen to twenty-four inches was what a cautious driver wanted to see before hauling heavy timber loads. The measurement told you what was beneath your feet at the point of measurement only. It did not tell you what was beneath your feet fifty yards ahead, where a spring, a current, or a change in lake depth might thin the ice to half the measured value or less. The people who used the ice roads understood this, and managed it through a combination of route knowledge, seasonal experience, and what can only be described as a deliberate tolerance of risk imposed by the commercial necessity of the crossing.
Supplies had to reach the islands. Timber had to move. A teamster who refused to drive the ice road when the ice was technically sufficient but not comfortably thick was a teamster who did not work. The economic structure of the timber industry did not allow for excessive caution about ice depth.
The ice was thick enough or it was not, and the determination of sufficient thickness was made by the person driving the load, based on the auger measurement, his experience, and his knowledge that the load had to reach the other side. The sea caves of sandstone along the Apostle Island shore were not, for the men who worked the ice roads, a tourist destination. They were landmarks. The caves formed where wave action over thousands of years had carved horizontal passages at the base of the sandstone cliffs.
Arched openings ten to forty feet high extended sometimes hundreds of feet into the cliff face in passages that dripped with water, echoed, and smelled of cold stone and lake water. In summer, the caves could be reached only by kayak or small boat paddled inside the openings from the lake side. In winter, when the lake froze at the cliff face and the water inside the cave mouths froze too, the caves became something entirely different. The wave spray that had entered the caves all autumn froze in place on the walls and ceilings.
Curtains of ice hung from the sandstone ledges. Columns of ice rose from the cave floors where drips fell and froze in accumulated layers. The formations followed the shapes of the stone beneath them, translucent and blue-white in the low winter light, lit from within by the ice itself as daylight seeped into the cave mouth. Teamsters and lumberjacks who worked the ice roads in the 1880s and 1890s used the cave formations as navigation points.
A certain arched opening was the marker at which you turned north for the crossing to Oak Island. An ice column in the mouth of a specific cave marked the point where the road forked toward Stockton Island from the road to Madeline. The caves were fixed features in a landscape that changed with every storm: pressure ridges that had not been there yesterday, open leads that appeared overnight where the ice had been solid the day before. But the caves did not change.
They were in the face of the cliff, and the cliff was in its permanent place, and the teamster who had worked these routes for ten years knew which cave marked which turn. At its peak in the 1890s and early 1900s, the ice road season was an intensive commercial operation moving enormous quantities of material across a surface that was never predictable and that existed for only about four months of the year. The timber trade was the dominant commercial engine, but it was not the only one. Commercial fishing operations established on the islands since the 1840s depended on the ice roads for winter supplies and for moving their catch to the mainland for rail shipment east.
The fishing families who lived on Madeline Island year-round, mostly Scandinavian immigrants, Norwegians and Swedes who had come to the Great Lakes fisheries because the work resembled what they had grown up with in the fjords of their home countries, used the ice road for everything the summer ferry normally provided. Medical care. Mail. Goods the island’s small stores could not supply.
The social connections that isolated communities need to hold together. The men, women, and children who crossed the Apostle Island ice roads between November and April crossed a surface the lake had created on its own schedule and dissolved on its own timing, with no regard for the needs of the people who depended on it. The lake froze when it wanted to and opened when it wanted to, and the island communities planned around those dates as best they could, which meant imperfectly, because the lake did not hold to a consistent schedule from year to year. A cold November meant early freezing and an early start to the ice road season.
The winter of 1891 to 1892 saw ice thick enough for loaded sleds by December 10. The winter of 1896 to 1897 did not see reliable crossing ice until January 8. The difference meant twenty-nine days of isolation for Madeline Island. Twenty-nine days when the ferry did not run because the open water had not yet become an ice road, and during which there was no solution for any medical emergency or supply need on the island.
The islanders knew this. They prepared for it. They stored what they could, kept their own supplies, and tended their own sick. Having a doctor on Madeline Island in the 1890s was not a luxury but a necessity, because the period between open water and ice road was a period when outside medical help was impossible to reach.
A man who broke his leg, or a woman with a fever that lingered, might live or die depending on whether there was someone on the island who knew what to do. The deaths the ice roads caused were of a particular kind. They were not slow deaths from disease or old age, which the island’s isolation made more likely. They were fast deaths by open water when the ice broke without warning and took with it everything that had been on top of it.
On March 3, 1901, a team of horses pulling a loaded timber sled went through the ice in the crossing between Sand Island and the Bayfield Peninsula, through ice that had been sufficient by the measurements taken that morning but was no longer sufficient by the time the sled reached the middle of the passage. The channel between Sand Island and the peninsula ran over water forty feet deep. The ice had measured ten inches that morning at the point of departure. Halfway across, where a current ran beneath the surface that the teamster could not see, thinning the ice to six inches over a stretch of a hundred yards, the surface gave way.
The first horse went through, the second followed, then the sled caught up with them, while the driver, a man from Cornucopia named Arvid Larson, jumped clear as the ice broke, landed on the surface eight feet from the hole, then lay flat and crawled on his hands and belly toward solid ice, while the horses and the load sank into forty feet of water in about forty seconds. Larson survived. The horses did not. Nor did the load of timber, which in 1901 was worth more than the horses.
Larson walked back to Sand Island over ice he was no longer confident would hold his weight. He reached the logging camp in a state the camp cook described in a letter to the Ashland Daily Press as blue from cold and fright equally, got whiskey, dry clothes, and a place by the stove, and was back driving the ice road within two weeks, because the timber had to move and he was the man who moved it. Larson’s account appeared in the local press. The accounts of men who went through the ice and did not come out appeared in the death certificates filed with the Bayfield County courthouse, and in the records of the small cemeteries on the islands and the mainland where they were buried when their bodies were recovered, which did not always happen in the same season they drowned.
The specific mechanism of drowning in ice road accidents on Lake Superior followed a sequence understood by the men who worked these routes, and that understanding did not prevent the accidents, because the accidents were not the result of ignorance. They were the result of ice thinner than it appeared at a location the measurements had not covered, at a time when the load had to be moved regardless of the risk. A man falling through ice into water deeper than he could stand faced immediate immersion in water that in February ranged between thirty-three and thirty-six degrees Fahrenheit. At those temperatures, cold shock occurs within the first thirty seconds of immersion.
It is an involuntary gasp reflex triggered by the sudden drop in skin temperature, causing rapid hyperventilation. If the head is under the surface during that gasp, water is inhaled immediately. The ability to swim is lost within three to five minutes as muscles cool and stop responding to voluntary commands. A man falling alone into February water in the Apostle Islands had a survival window measured in minutes.
The length of that window depended on whether anyone else was close enough to see him fall and reach him before it closed. The ice roads had no rescue teams. The people who traveled them were often alone. A teamster hauling a load from Sand Island to the Bayfield pier on a February morning was the only person on that stretch of ice as far as his eyes could see in either direction.
If he went into the water, the person who found out was the one waiting for him to arrive at the pier and noting that he had not arrived on time. The lake is forty degrees in February. The bottom of the channel between Sand Island and the Bayfield Peninsula is forty feet down, and it has been there for ten thousand years. The bottom does not record the names of the men it swallowed.
The Ojibwe communities of the Apostle Islands region understood the ice in a way the commercial operators who arrived in the 1880s did not, and traveled on it in ways that had no commercial application but kept more of their people alive. Ojibwe winter crossings relied on reading the surface rather than measuring it, on the color, texture, and sound of the ice under their feet rather than the single-point auger measurement that the timber industry used as its primary safety assessment. Dark ice was bad ice regardless of thickness, because dark ice is ice over deep water where the cold seeps in slowly. Ice that booms underfoot is ice under stress.
It formed quickly in a hard cold snap and has not yet bonded with the ice around it. Ice covered with a light dusting of snow that shows no trace of your footsteps is thin ice, because thick ice is dense enough to hold a footprint while thin ice is not. These distinctions were recorded nowhere. They were passed from those who knew them to those who needed to know them across generations, through specific conditions of crossing a particular body of water in a particular season.
The commercial operators had no access to this knowledge, and in most cases did not seek it. The economic pressure to move timber, supplies, and fish produced an approach to ice that made the auger measurement the primary standard and treated everything else as hesitation that cost money. The Ojibwe approach to reading ice had been developed over centuries of necessity. The commercial approach to reading ice was developed from the necessity of meeting shipping deadlines.
The two approaches produced different loss rates. The fishing families on the outer islands knew the ice better than the logging crews, because the fishing families lived on the water year-round for decades and learned what the Ojibwe learned through a different but equivalent accumulation of hard experience. A Norwegian fisherman who had crossed between the outer island and the mainland for twenty winters possessed detailed knowledge of that specific route: where the currents ran, where the springs were, where the ice thinned first in spring and froze last in autumn, facts acquired through surviving that crossing enough times to learn them from the ice itself. This knowledge was local, personal, and largely unrecorded.
It lived in the people who possessed it and went with them when they died. The timber produced by the Apostle Islands moved across the ice roads to the mills at Ashland and Bayfield through the last two decades of the nineteenth century and into the first years of the twentieth. By 1910, the commercially exploitable timber on the islands was nearly gone. The logging camps closed.
Vegetation grew over the timber roads leading into the island interiors. The piers where logs had been loaded onto barges eroded and fell into the lake. What remained on the islands after the cutting was a depleted landscape, a scene of stumps and brush and second growth that would take seventy years to return to anything like a forest. The ice roads continued after the logging ended, because Madeline Island kept a year-round resident population that still needed supplies and connection to the mainland.
Madeline Island’s population in 1910 was about three hundred people, fewer than its peak during the logging years, but stable enough to form a community with schools, churches, and a small commercial district in the town of La Pointe on the island’s western shore. The fishing industry on the outer islands continued, but on a smaller scale. The ice road season remained the central fact of winter life in the Apostle Islands, as it had been for thirty years. What changed in the 1920s was the automobile.
The car came to Bayfield County the same way it came everywhere in rural America. Slowly, then quickly, over about fifteen years during which horse owners and car owners coexisted on the same roads and the same ice surfaces. A car weighed more than a horse-drawn sled. It required a more solid ice surface to operate safely.
You could not jump clear of a car if the ice broke. A car that went through the ice in February sank faster than a horse and killed its occupants faster, because car doors were harder to open underwater than a sled was to roll clear of. The car also provided no buoyancy and nothing to cling to. The horse at least tried to get out.
The car tried nothing. The first cars crossed the Madeline Island ice road in the winter of 1921. The ice road between Bayfield and Madeline Island had been used by vehicles of all kinds for forty years by then. The route was well established and well known.
What cars did to the ice road was make it faster and more convenient in the seasons when it was usable, and more lethally catastrophic in the moments when it was not. Records of car accidents on the ice road in Bayfield County between 1921 and 1940 can be recovered from the county clerk’s office in Washburn and from the Bayfield County Press, which covered drownings and near-drownings with the matter-of-fact brevity that local newspapers of that era used for events that were part of the understood risks of living in that place. A car going through the ice was news in the sense that it was recorded and reported. It was not news in the sense of being surprising.
People drowned in the ice road crossing between Bayfield and Madeline Island in 1924, 1931, and 1938, and in other years for which the records are less complete. The crossings that ended without incident are not in the records, because they did not merit recording. They were the routine of winter life. They were like Ernest Hawkinson and his four hundred pounds of salt pork arriving safely, and going home to his wife, appearing in no document anyone ever found.
The ice caves that form along the sandstone shore of the Apostle Islands became a subject of tourist interest in the 1960s and 1970s, when their designation as a national lakeshore in 1970 placed the islands under federal protection and began their development as a recreational destination. The caves existed before the tourist infrastructure. They existed before the logging. They formed every winter when conditions were cold enough and the lake was calm enough for ice to accumulate on the sandstone face.
The ice formations that appear in photographs of the winter caves, the curtains and columns and icicles of lake ice that make the caves look like something imagined rather than real, are the product of the same cold that made the ice roads possible and the same lake that made them necessary. In deep winter, ice builds along the Apostle Island shore from the beach outward in layers. The spray of waves hitting the cliff face in late autumn freezes on the sandstone. Each subsequent cold night adds another layer.
Water seeping from the sandstone itself, groundwater moving through the rock and emerging at the cliff face as seepage and slow drips, freezes on contact, adding to the formations from within. What results, in a winter cold enough and long enough, is a transformation inside the caves into something with no equivalent in temperate climates. The ice follows the geometry of the cave. Arched ceilings produce arched formations of hanging ice.
Angled walls produce sheets of frozen spray. Floors where drips have fallen for years produce accumulated ice formations, recording each freeze in a layer the way a tree records each year in a ring. The ice is not permanent. It arrives in December or January, grows through February, and begins to collapse in March, first at the cave mouths where sunlight reaches the ice directly, then deeper inside as the warmth advances.
By April it is gone entirely. The next winter, the process begins again. The formations that appear in any given year are the result of that year’s particular sequence of temperatures, winds, and lake conditions. An early cold and calm water produce the most spectacular formations.
A late cold or a January warm spell produces lesser ones. No two winters produce quite the same caves. The people who enter the ice caves in February enter a formation that has never existed before and will never exist again in quite that form. The National Park Service began tracking winter visitor numbers to the ice caves in the 1980s.
In a year when the ice is available, and when the ice route from Meyers Beach on the mainland to the cave shore has been certified safe by the park service, visitors can reach one hundred thousand people. In a year when the ice does not form sufficiently, the caves are inaccessible and the numbers are near zero. The ice road that carries visitors to the caves is a physical descendant of the ice road that carried Arvid Larson’s timber load and Ernest Hawkinson’s salt pork and every commercial load that crossed the western arm of Lake Superior between the 1880s and the mid-twentieth century. It is assessed by the same essential question: is the ice thick enough to bear the load that will be put on it?
It is answered the same essential way, by drilling a hole and measuring. And the answer it gives carries the same essential uncertainty it always has, that the ice is this thick at this point, and what it will be at the next point depends on the water beneath it. And the water does not make itself available for inspection. The highway system that reached the Apostle Islands region in the 1930s and 1940s did not make the ice road obsolete overnight.
Highway 13 was paved around the Bayfield Peninsula during the 1930s under the New Deal public works programs, connecting Bayfield to Washburn, to the rail junction at Ashland, and to the wider Wisconsin road network. This made the mainland communities of the Apostle Islands region accessible year-round by land. It did not make Madeline Island accessible year-round. Madeline Island is an island, and the road stops at the Bayfield waterfront, and the crossing still requires covering two miles of lake that is still there.
What the highway system changed was the alternative to the ice road. Before Highway 13 was paved and reliably maintained in winter, the people of Bayfield County depended on the ice road for winter mobility, which made the risk of crossing a necessity that could not be avoided. After the highway became passable year-round, crossing the ice road to Madeline Island remained necessary for winter residents, but the consequences of late freeze or early thaw became less severe, because the mainland was accessible, and it provided the goods and services the island needed. The isolation shrank to two miles of water, rather than two miles of water plus an unreachable mainland.
The Madeline Island ferry began operation in 1940, replacing older, smaller operations that had served the open-water season. The ferry made summer crossing reliable and frequent. It did not change the winter crossing, which still depended on the ice road when the ice allowed it and on the interruption of the road when it did not. What the ferry created was a transition season that had not existed in quite the same form before.
The period between ferry service and ice road, when crossing was unavailable by either means, was managed after 1940 by running the ferry as long as the ice allowed and restarting it as early as possible, shortening the period of interrupted crossing from what it had been in the pre-ferry era. The ice road between Bayfield and Madeline Island still operates today. It is certified by Bayfield County after ice thickness measurements. Depth markers are placed when it opens, and it is used by a variety of vehicles, from trucks supplying island residents and businesses to tourists who cross specifically to say they drove a public highway across a lake.
The ice road is a state-certified highway during the winter season. It appears on official maps. It has a speed limit. The people who drive on it now are not in the same position as Ernest Hawkinson in January 1898.
They have, in most cases, another option. The ferry will run if the ice is not suitable. The mainland remains accessible if the ferry cannot run. The two miles of ice between Bayfield and La Pointe do not represent the difference between supply and starvation, or between connection and isolation, as they did for the fishing families and logging camps of the 1890s.
The risk is still there. Cars still cross the Madeline Island ice road. But the calculation behind the decision to drive on it has changed, because the cost of not driving on it has changed. What has not changed is the ice itself.
The ice forms over the same water, over the same lake bottom, in the same location relative to the sandstone cliffs where the caves form every winter. The thickness of the ice at any point is still a function of water temperature, current, depth, and the severity of that particular winter’s cold. And the only way to know it is to drill a hole and measure. And that measurement tells you what is here, not what is there.
And there is where you are going. The men who created the commercial ice road system in the Apostle Islands in the 1880s and 1890s did not build it in the sense of construction. They created it through repeated use. The first crossing of each season established the route.
Every subsequent crossing packed it down. The men who drove the first loads of each season onto new ice were people the commercial operation could not afford to do without and could not afford to lose. Both facts were true at once, and the industry lived with them without visible difficulty, because the alternative to this contradiction was not moving the timber, and not moving the timber was not an option the bondholders, mill owners, and railroad operators who financed these operations were prepared to consider. Wages for ice road teamsters in the 1890s ran from a dollar and a half to two dollars a day plus board.
The board was a bed in a logging camp or a room in a Bayfield boarding house with meals included. The wages compensated skill and risk at a price set by the industry, a price that men looking for work were willing to accept. Whether this was adequate compensation for driving a loaded sled across two feet of ice over forty feet of cold February water at twenty-three below zero is a question the records do not answer, because they were written by the people who set the wages. Their names exist in the payroll records of the logging operations, in the county death records, in family letters, and in no other document that has been widely available to the public.
The ice roads they drove form in the ice every winter from the same water they crossed, linking the same shores over the same lake bottom, because the lake has been here for ten thousand years, and the cold returns every November, and the water freezes in the same passages, and the same currents run beneath the same ice, and the same caves freeze on the same cliff face, and the same light penetrates the same ice with the same blue-white color that no photograph quite captures, and that you must stand inside the cave in February to actually see. The Apostle Islands ice road opens when the ice is ready. The Bayfield County Highway Department posts the certification on its website, with the thickness in inches, the weight limit in tons, and the route markers. A hundred thousand people come when the caves are accessible.
They walk across the ice, enter the caves, take photographs of formations that have never looked quite this way before and never will again, and walk back to Bayfield, and drive away up Highway 13, which connects to Highway 2, which connects to the interstate system, which leads to wherever they are going. The ice under their feet as they walk to the caves is two feet thick in a good year, and thinner where the lake is deeper, exactly as it was in 1898. The sandstone above the cave mouth is the same sandstone it has always been. The water below is thirty-three degrees in February, and it is very deep, and the current still runs.
The first man to cross the ice this season is the same as the first man who crossed the ice in 1883. He goes because something needs to reach the other side, and he is the one who takes it there. His name is in some database somewhere.
It is not written on the cave.
