Ali and the Changtang to the Silk Road

<< (previous section) (next section) >>

Ali

It snowed all night, but the sun was out when we got up, and the landscape was white and beautiful. We finished the small climb over a pass we had started the night before, and began a descent. Halfway down, we found something we hadn't seen in 1900km: paved road. The sun had heated the black pavement, and the road stretched like a black ribbon slicing through the white out into the distance. This was a big change. Cycling in Tibet has always been a dirt road, mountain bike kind of thing, but this will change. There was some sense of loss of adventure, but I was tired, and honestly welcomed the easy ride to Ali. We cycled 85km that day, more than we had covered in a day for over 5 weeks. We summitted the last pass to Ali, over what had previously been a shortcut, but was now obviously officially part of route 219, and Ali came into view, the biggest town west of the Golmud highway.

I had been in Ali in 1999; what I found now was nearly unrecognizeable, a boomtown being built seemingly from scratch. The place had tall buildings, a new post office, a department store, and a massive fleet of taxis that made crossing the street difficult. We were referred to the Shiquanhe Hotel after being rejected from the Postal Hotel for being foreigners (there are some rules in China still governing where foreigners can stay the night: apparently the Postal Hotel in Ali is not officially acceptable, although it looked a fair bit nicer than the hotel we now checked into. The check-in process was a study in the old China: the door was open, but only a dim light was on over the deserted reception desk. We wandered the halls, knocking on doors, pushing them in, finding noone. The small empty desk in the lobby with the sign "Asst. Manager" laughed at us. I sat down on one of the plush leather couches in the lobby and waited. Martin grew impatient: he was desperate for a shower. A man came in, said he worked there, but wasn't reception, and couldn't do anything. He suggested we go out and eat: surely someone would be back by the time we finished. We wandered out and ate at the closest restaurant, a hotpot place where you toss skewers of vegetables and meat into a boiling bowl of broth heated from a flame under the table. When we came back, it was discovered that the reception had simply been sleeping in their room. We asked for a room with a shower, but it was prohibitively expensive, so we just took a dorm room, sharing it with a Japanese guy who had ridden a crummy Chinese mountain bike part way from Shigatse, hitching much of the way. He said he was sick of bad roads, and was eager to get onto something paved. He took a bus to Yecheng, down on the (paved) Silk Road the next day.

Ali is a garrison town, with three different army installations. Tibetans are a minority here, making up the class of day laborers you see wandering the city with a shovel balanced over their shoulders, looking or waiting for work. There are lots of Hui, and a Han majority, and a few Uighurs here from Xinjiang. What had been a two street town now had traffic lights (albeit moveable ones) and clothing stores lining the main street. Hair salons, fronts for prostitution, were open till the wee hours, with telltale red flourescent light bulbs illuminating women in pyjamas knitting or warming themselves by the stove in the middle of the room. And, of course, food of every kind. We spent three days eating our way around town, shopping for a few essentials, and taking a long hot shower to clean up. Tomorrow, hopefully, we will be off to the north, making our way towards the turnoff to Bandag Tso in the northwestern corner of the Changtang, and approximately three weeks of offroad travel over the Kunlun Mountains down to the Silk Road.

Leaving Ali: The Route to "The Turnoff"

We lingered in Ali one extra day when I found my rear tire flat and my Chinese patches wouldn't hold air: this was really an excuse to take another day off, although it also cost us a day on our visa. (The visa extension process in Ali is a study in bureaucratic nonsense: the one-month extension that starts at the end of your current visa anywhere else in China begins at the date of issue here. This meant that I was paying a full month's extension fee for an additional 15 days: Martin was only gaining 10 days. Unfortunately, we had little choice since there was no hope of arriving in a city on the Silk Road before either of our visas expired.) I used the day to patch my tube and eat more junk food: Martin spent some 7 hours in the wangba writing a long correspondence filling in folks on his last 2 months.

The next morning we packed up and headed out for breakfast. I crossed paths with an Australian hippy backpacker, who asked about Mount Kailas and how to get to Lhasa and so on. I mentioned Rajinder Swaroop, the Indian who was probably on that day making his 159th circumambulation of Kailas, thinking he would be right down her alley. Her response was "Oooh, I spent a lot of time in India man (this was obvious)...I just love sadhus." Perfect. She was accompanied by a fleshy-faced Chinese named Peter from Nanjing, who could only say "hello" in English. I wished them lots of luck, and made for breakfast, picking up Martin at the net cafe.

We left Ali sometime in the afternoon. The road was sandy, but vastly improved from years past since the Chinese have built up some sort of roadbed, and we made good time to the pass 40km north of town. A strong tailwind blew us down the other side, and we were able to find a campsite with a bit of standing water in a dried out grassy plain.

The following day we made Rutog, 130km from Ali, bumping along a rocky road through a canyon draining down to Bangong Tso. A road was being built over a small hill, taking out a bend in the road and effectively removing Rutog Xian, the Tibetan settlement, off the map: this was the equivalent of being passed by the railroad in the 19th century - the town will slowly wither and die, replaced by the new Chinese town of the same name 6km away. Chinese Rutog has also grown, with a paved street, a wangba, and a mixed population of Han, Hui, Uighur, and Tibetans. We wandered in, got a Chinese dinner, and then headed out of town to camp. The following morning we got breakfast and did some of the shopping we needed to do for Chang Tang supplies, two or three days before we actually left the main road. I tried to keep these to a minimum, counting on Domar, the final settlement before our turnoff, for the basics (instant noodles, more or less). Martin stocked up on milk powder ("I gotta have this, or I'm no good") and we both bought 2.5 liters of fuel here for the crossing at a ridiculously high price (about a dollar a liter), not knowing what was available in Domar.

The ride to Domar was straightforward, a day and a half. About 15km north of Rutog, the road hugs the southern shore of Bangong Tso, a large freshwater lake which is part of a lake complex that straddles the Indo-Chinese border: if the border was open here, one could get to Ladakh and Leh in three or four days, but continuing tension and unresolved disputed territorial claims from a brief border war in the early 1960s, and a heavy military presence on both sides of the line make this a pipedream for now. The lake was beautiful and the weather perfect: one could see the stony bottom well out from the shore. A fish restaurant sat by itself along the cliff lined shore, a derelict building and several even more derelict boats hauled up onto the shore. A Kyrgyz yurt was on a spit of land, next to a sign designating the area as part of the Changtang Nature Reserve. The proprietor came out as we took pictures of this bizarre collection of boat, yurt, and phone booth in the middle of nowhere. I asked him how long he had been set up here. Five years, he said.

"Where are you from originally?"

"Sichuan" It is a truism in China that nearly all the transmigrant Han in the west are originally from Sichuan and the Red Basin area, one of the most densely populated areas in the world. I asked about incentives to move here. "No, not really. It was just some kind of business opportunity. Well, I pay no taxes."

"Would you pay taxes in Sichuan?" He just laughed.

"How's business out here?" I knew that pretty much every cyclist who came along this road knew the place as "the fish restaurant": the guy walks out to the fish pen next to the shore, hauls out a fish in a net, clubs it in the head, chops it up and boils the hell out of it. Not good, but fish is a real novelty in Tibet.

"Its fine. Can't complain."

"Any plans on heading back to Sichuan?"

He smiled and shrugged.

We camped a few kilometers from the north end of the lake, and the following day cycled over a 4800m pass to Domar, arriving near sunset. Domar has also grown, although much less spectacularly: it is still a dusty collection of mud buildings alongside a river with good pasture. We found a Hui restaurant in town, which raised our spirits, but the food was lackluster, and we were urged to clear out when a bus heading for Yecheng stopped and disgorged a couple dozen passengers for the dinner stop. The bill was an odd number, which meant the guy was just making something up, something quite a bit more than it should have been. I had a few firm words with him and brought down the price a bit, speaking loudly and clearly in front of the other patrons, but it left a poor impression. My first dishonest Hui encounter...

There was a well-stocked shop across the street, but this guy ran that shop too, and his prices were outrageously high. We despaired of being able to buy enough food, and wandered off to camp 200m from town.

The next morning things looked up. We found a Chinese restaurant with fair prices (noodle soup for 8Y, about a dollar): the proprietors were just waking up, running around in their underwear, but they fired up the stove and in short measure had two soups on the table. The door was an imaginatively shaped concrete forest scene, complete with roots and two "stumps" to sit on in the sun. We wandered the town looking for the shop with the best deal on noodles, and anything else. This ended up being a Tibetan shop, with the shopkeeper hailing from Markam. Several others crowded in, and I was invited to have some thugpa and tsampa before heading off. It turned out they were all from Markam, apparently the Tibetan equivalent of Sichuan, a breeding ground for migrant businessmen. The guy was friendly and honest, selling instant noodles for 2Y each (much better than the 5Y the Hui had asked me for the previous night). I said we wanted a lot. How many? How about 130? No problem, he said, and soon enough we had stuffed our bags with noodles. I fished a rice sack out of the trash and used this to carry about half of the noodles. I also bought a kilo of tsampa to add to the half kilo I was already carrying, and a kilo and a half of sugar. With the addition of some 90 bars (a sort of high energy compressed food - we each got 10, cleaning out the town), I was ready for about 22 days in the wilderness fairly comfortably, albeit with a short menu. The last thing I needed was another lighter for my stove as backup, which was in a Chinese-run shop. I knew the lighter cost 1Y, but I asked anyway. The boy working there blurted out "yi kuai" ("1 yuan") before his father, the shopkeep, could get out "liang kuai" ("2 yuan"). I frowned and tossed 1Y at him as I walked out the door. It was time to get out into the wilderness and forget about this stuff.

We cycled a half day, and the following day managed to get over the 5400m Lalei La near dark, heading to Ngasashan (a small Tibetan settlement and truckstop) in the dark, the temperature dropping to -8C in the last twilight as I reached the truckstop. Five years before, there had been a rather lively restaurant tent catering to truckers, but this seemed to be packed up: perhaps the road upgrade had made this place an unneccessary stop, and the restaurateur had given up. A collection of Tibetan pilgrim trucks was in the front, and twenty to twentyfive Tibetans stared at me in the darkness as I waited for Martin to arrive. I poked around the place, but there was no restaurant for a last meal. Several nomad tents were set up for the winter: these were folks coming in from the western fringes of the Changtang to a more hospitable place (we were at 5200m, which was a stretch to call hospitable, but nevertheless...) for the winter months. Women jangled in their jewelry as I walked through the tangle of tents and yaks and sheep. A couple of guys ran a shop out of one room in the front building; I was not surprised to learn that they were from Markam. They invited us to stay the night in the shop, and we warmed ourselves by the yak dung fire while they melted large pieces of ice for water. Nomads and pilgrims wandered in and out of the place to have a look at us, occasionally buying something. Dinner was instant noodles, as was breakfast, something I wasn't very eager to get started with. I slept well, a last night around people, and in the morning we bundled up and headed out for the turnoff to the east and the beginning of the Changtang crossing.

Crossing the Changtang: Stage One - The Roaded Piece

The turnoff for the route we had planned on was about 15km from Ngasashan, just after a short 5300m pass. The road has been rerouted in the last year or so, so we didn't get to laugh at the sign marking the "6000m" pass that all the truckers and backpackers refer to (in reality the pass is about 5300m: if it were 6000m, a lot of our unacclimated friends would be experiencing very serious problems at that point...). We stopped, Martin said a few words into the camera, made a short speech to the effect of "I hope it goes well", and pedalled off on the track to the east.

The track had changed since Martin's encounter with it in 2002: there were bulldozer tread marks. This we found curious, and even moreso when they continued the next day. We passed an untended herd of sheep milling about near the lake in the valley, then rose up over another short pass marked with prayer flags and crossed the threshold of the Changtang. A few kilometers in, there were hot springs and a permanent structure down by the large lake spreading to the east from the main Ali-Yecheng road. A nomad woman was spreading yak dung on the ground to dry, but she was startled by our approach and quickly left the area without exchanging any words. The hot springs were nice: bubbling pools of warm water feeding a stream running down to the lake. Unfortunately none of them were truly hot, and the day was not much above freezing, if at all, so the best I did was dip my fingers in the water. We carried on across a flat plain crossed by a couple of rivers reduced to a tiny trickle by the cold, collected water, and headed up another hill before descending to another plain for camp. It crossed my mind that this could be a fool's errand as I put up my tent in the subfreezing temperatures, but I crawled into my sleeping bag and forgot about it, drifting off to sleep quickly.

The next day we continued on the road, over an almost imperceptible pass at 5300m, and then descended along a frozen river bed in a blustery snow shower. After lunch, during which I set up my tent to get us out of the elements (this became a daily practice, since the wind was incessant and the temperatures never above freezing), we made it to a small lake, where up on a ridge we saw a brown nomad's tent and a blue Dong Feng truck: I didn't see the people, and though we didn't know for sure (we did have some idea, however) this was to be the last visible sign of current human residents of the Changtang for the next 14 days. We had to break up ice for water, which I carried in my rice sack, something I was to be very happy to have in the following two weeks.

The next day the road continued, in quite passable condition, and we made good time to the western end of Bandag Tso. The road turned north, although there were tracks heading east: we followed the northern route for 3km before deciding this wasn't going where we wanted to go. We consulted the TPC map and GPS, and decided to head off-road to the southeast back to the southern shore, where we hoped to find the eastward-bearing tracks. The sun was gone, and the wind had picked up to a steady 25kph with gusts up to about 40kph. My face went numb as I crossed the odd phenomenon of a frozen salt marsh in the plummeting temperature. Martin was crossing more slowly and following a more eastward path: I resolved to make sure we stayed in more close contact in the future. A large salt storm blew over us, obscuring Martin from view and leaving an alkaline taste in my mouth and nose. We rejoined on the southern edge of the marsh, and after a quick reconnaissance climb found the tracks to the east. We picked these up and headed around the lake, finding a gushing spring along the way, and evidence of a truck camp. We filled our bottles and carried on another eight or nine kilometers to an area which was obviously popular with nomads in the summer, with yak dung and sheep tracks aplenty, and camped in a stone animal stockade on a soft bed of sheep dung - not ideal, certainly, but sheltered from the wind by a cliff face just to our west. The reason for the camps became apparent: a river gushed out from underground here, cutting a deep gorge of perhaps 60m out of the rock and flowing down to the lake. We traversed the southern rim of the canyon the next morning, following truck tracks along the way, and saved ourselves an unneccessary river crossing.

That night we received a few centimeters of snow, but the day dawned clear, and we took several photos of the white landscape, the bright blue lake, and the now visible massif to the north marking the southern skirts of the Kunlun Shan. We rounded the lake to the east and then the road struck out along a plain, crossing a dry sandy riverbed and a salt lake on its way to a pass. We saw several kiangs cantering north of us, and then a couple of groups of chiru. Hares bolted from holes as I approached, and ravens and birds of prey circled overhead. I hadn't expected much wildlife at all this late, but this part of the Changtang still had a fair number of animals eating the last of the summer grass. Along the road there were rusting tin cans crushed by truck tires, the occasional bottle of beer or whisky, and at one point, a woman's high heeled shoe (?!?). I picked up a rusting Uighur knife as a souvenir, evidence of either a prospector or poacher. I also picked up a chiru horn, perhaps 30cm long, as another souvenir, although if customs recognized it, there could be some explaining to do...

The afternoon was spent climbing a long slow pass, winding up a sandy riverbed that made pushing the bikes a neccessity. At the top, I collected ice from a small frozen pond, and then we descended to a plain at 5150m, camping near a spot where a group of several chiru had been standing before bolting at my approach.

The next day the track continued well-defined, and we crossed over a low pass to find a weather station tower (unmanned, of course) on a ridgetop overlooking Pur Tso, and the Toze Kangri massif (around 6300m) came into view, huge and covered by an extensive icecap. We lunched at the lake shore, getting water after breaking a hole in the ice of the not-quite-completely frozen body of water. I had never heard the sound of a large body of water freezing, being from California, and it was eerie: groans and creaks resounding across the lake surface and the surrounding valley. The track was well travelled, and we rounded the northern shore of the lake before striking eastward. At one streambed, there was an overturned and abandoned jeep, with broken cables attached to it: evidently there was some attempt to reclaim the vehicle, but in a strong current, an overturned truck is difficult to pull out without a crane or lifted winch, so there it sat, rusting as a reminder to foolhardy drivers.

The tracks became jumbled on the eastern shore, and we eventually decided to leave the tracks and make a straight line to the east northeast, heading around an unnamed 6000m massif to our north: this was to be a sort of pivot point - from here we bent our direction of travel to the northeast, towards Heishi Beihu and the entrance to the fault formation which was to be our way down and out of the Changtang to the deserts of Xinjiang. We weren't finished with the truck trail however: we crossed it again near dark, and followed it most of the next day, down to a small salt lake at the southern edge of what was named "Antelope Plain" by one of the 19th century British army officers who travelled the area, finding tens of thousands of chiru grazing the plain. What we found when we got there was a dessicated plain with chiru tracks but no wildlife: after the small lake we were to see no more wildlife, other than birds, for the remainder of our Changtang trek. After pushing our bikes through a crusty snow left from the storm several days before, we struck out away from the track, which continued along a more easterly path into the Changtang. By day 8, we were looking forward to our way out. We had made it past the starting point for the 2002 American expedition, written up in National Geographic and detailed in Tom Ridgeway's book "The Big Open". One thing puzzled the both of us: how was it that he could claim to have seen no tracks? While there wasn't much to call a road, or even really a trail, we couldn't ride or walk much more than 3 hours without coming across truck or 4WD tracks plunging into the wilderness. Something didn't add up: had the tracks been covered by snow, was the story better by using the word "trackless", or had poachers and prospectors driven ever deeper into the area in search of chiru skins or gold?

Crossing the Changtang: Stage 2 - The Walk

Crossing the antelope plain took a day and a bit. We were able to keep a pace of 20-25km per day (12-15 miles), which was more or less what we had planned on. At one point, Martin said "My biggest fear now is 20cm of snow", which set my mind racing. That could be serious, because it would slow us down substantially. Food was not a concern, but we were beginning to see that fuel could be: we hadn't really considered that we would have to melt ice for every bit of water we used. Nights typically ranged from -28 to -31 degrees Celsius, and the days never really got above freezing. The wind, incessant all day, but thankfully a tail wind, died down shortly after dark, but picked up an hour or two after sunrise. Our days were fairly short, usually only 5 or 6 hours moving, because neither of us was particularly eager to get up and go when the temperature was -28 at sunrise: we huddled in our tents lingering over breakfast (instant noodles, of course) for an hour or so before packing up.


After crossing the plain, a ridge rising perhaps 200m separated the plain from the depression around Heishi Beihu ("Blackrock Northlake" in Chinese). We could see clearly Qong Muztagh, a huge nearly 7000m massif capped in ice to our north about 40km, and Toze Kangri, still visible to the southwest, was receding quickly. The ridge looked simple enough - only one contour line on our map - but turned out to be rather more difficult. One of the things we discovered was that the map's 500 foot contour interval, which looks so impressive when sitting around daydreaming in a hotel room, misses key points like 300 foot climbs or cliffs, and so on. What looked flat on the map was actually a broken up volcanic badland, with black igneous rock piled up on top of ruthlessly soft sand. Beautiful, without a doubt, but no easy pull. A river ran up to a pass to our southeast, but Martin favored a direct line, and I demurred, though quickly cursing our decision after entering the badlands. We made no better than 1.5km per hour, hauling our cycles up over boulders, around dried out salt ponds, through small gorges. After 4 hours of this we found a way out, down to the river which fed Heishi Beihu from the southeast. The river had frozen up into the volcanic rock, creating a fairyland of black rock surrounding frozen ponds and rapids. As we left, I mentioned that although I was frustrated at first, I was glad we had seen such a spectacular landscape. Martin for his part picked up a couple of volcanic rocks to take home as souvenirs - I was more pragmatic and decided to save the weight.

After passing over another sandy ridge, we saw Heishi Beihu, another milestone on our trip. Martin (a fount of information about the Changtang and its explorers) mentioned that only 5 westerners had ever seen the place before: Schaller in 2001, and the 4 Americans in 2002. A long high ridge hemmed in the lake from the north, with the pass to the northeast delineating the border between the Chinese provinces of Tibet and Xinjiang. We decended down through a steep canyon to the lakeshore, which had the sound of waves lapping on the sand: Heishi Beihu is very saltine, and there wasn't a trace of ice anywhere on the lake. There were, however, very substantial truck tracks along the southern shore where we were, enabling easy progress (we could cycle without difficulty) to the northeast end of the lake for camp. We pondered this again: could the American expedition two years before have possibly missed these tracks? It seemed impossible they didn't see them. Once again, we wondered whether the wilderness was under serious pressure in the last few years.

The next morning we woke to a beautiful sunrise along the lakeshore, and I watched two ravens chase a small white bird in circles in the sky. I had been reading The Odyssey at night during the crossing, and omens were usually sent from Zeus in the form of an event with birds: my mind wondered what this portended. It didn't look good...

We crossed the pass into Xinjiang and made our way eastward over a new route, passing between two mountains and decending through a very narrow sandy gorge out onto a plain leading to the Kunlun fault. The soil was miserably sandy, and my front tire sunk deeply into the loose ground. The next two hours felt like I was plowing a 3km furrow by hand: Martin had a much wider tire which made things easier for him, but both of us were frustrated and exhausted when we decided to set up camp. The wind was howling from the north, more or less in our faces, and the temperature around sunset was -8C: the Kunlun was welcoming us. At least from here we could see the mountains on the other side of the fault, and guessed that in a day or a bit more, we should be headed down, off of the plateau that we had spent 3 months on. Both of us wondered at the prospect of dropping below 3000m, an elevation we had last seen about 3 months before, and talked from tent to tent about how nice it would be, and how burnt out we were on Tibet. Just before camping, we saw a large fuel drum abandoned in a wash...apparently this place, described by Schaller not so long ago as "one of the most inaccessible places on earth", was no longer out of range for a truck, let alone a 4WD vehicle.

Our final day on the plateau took us over a sandy ridge and down to a frozen salt lake. A dried out salt bed made cycling possible, and we rounded a mountain to see the final barrier, a 200m ridge above us to the north, giving onto the fault splitting the western and eastern Kunlun mountains. Here, too, there was a truck track, and the next morning on the final climb, we found a heavily trafficked track running from the fault to the southeast, compressed by dozens of truck trips and nearly good enough to call a road. We took a shortcut up a small pass, leaving the tracks which most likely crossed the official pass into the gorge, and sometime before lunch, we were looking across a valley at a 6200m peak and down into the fault. It was relieving to see our way out: we had plenty of food, but I had perhaps 6 days of fuel left, having used so much for melting ice over the previous 14 days. There was another short talk into the video camera, a handshake, and then we decended down into the fault, trying to get out of the wind and down to a warmer place.

Out of Tibet

We took one last look at the Tibetan Plateau, home for 3 months, and then I at least started eagerly down the fault formation heading for the Taklimakan and the Silk Road. There was a small sandy plain without drainage, the edge of which was a steep wall of sand down to the valley floor. We more or less skied down this slope and had lunch, setting up the tent in gusty winds which whipped up sand storms that blew quickly down to the northeast. A road was visible across the frozen river, so we crossed the stream, finding that at 4800m (our lowest point in more than 15 days) the ice was almost slushy in the sun. The stream meandered back and forth, feeding a lake about 8km away. A spring also fed the lake, and the liquid water made both of us smile: our fuel would be enough. We estimated 2 1/2 days to get down the fault formation to the village of Qarasay, where we both fantasized about a restaurant with something other than instant noodles on the menu. The spring came out of the hillside underneath the road that we now came upon. The road was a real road, with signs of maintenance along the route (embankments, cuts, and the like). Heavy machinery had been driven up at least this far, perhaps to the verge of the Changtang Plateau: although we couldn't confirm this, it seemed very likely, given what we had seen at the rim of the fault. This made me hope for a 2 day end to this section, and an easy ride down. What lay in front of us, while easier than a roadless section would have been, was far from easy. The people building the road had paid little attention to the Chinese road engineer's mantra of "five percent", preferring to make the route shorter and steep.

We passed by the dormant gold mine that the American expedition had described in 2002 (link here). It stunned the both of us that George Schaller, the notable Changtang biologist (link here), had had to approach the plateau from this entrance in a mule train in 2001: three years later, you could almost drive a family sedan along certain sections near the top. The mine left large piles of rock along the now frozen river: there were a few large metal graders and sifters, discarded shoes, and rubber hoses used as siphons for the mining process (which is water-intensive). The machinery itself was pulled out each winter, I guessed, because there was none in evidence. We pushed our bikes through the mess and then found the road again, hugging the skirts of the range on the north side of the fault.

Our first night left us feeling better: at 4500m, a positively balmy -18C (0F) for a low. Farther down, the sun even melted the river and liquid water was easily available along the route whenever the road dipped down near the waterway. On the second day, the road undulated over steep skirts and canyons, with sandy sections making the cycling difficult, sometimes impossible. We passed a homestead a couple of hours before sunset, but it appeared to be unoccupied: Martin suggested it could be a (for now) unmanned checkpoint or ranger station for inspection of trucks to nab chiru poachers. I thought this was overly optimistic: the Changtang appeared to be a nature preserve in name only - there was no money or will for enforcement, and as the mine showed, money talks. I told Martin that the US doesn't set a high standard, with drilling on Alaska's North Slope and pressure to expand these operations dramatically. If oil is found in significant quantities in the Changtang, the "preserve" will be finished: China's energy consumption is rising at least as fast as its economy is growing (about eight or nine percent annually), and its own oil reserves are far from sufficient. The political and financial incentives for exploiting mineral resources in Tibet and the Changtang far outweigh the conservation impulse that Schaller and others have worked hard to cultivate here.

The road descended to the riverbed, which was in a deep canyon by now, slicing into the arid landscape. We had to make a few river crossings (the first liquid river crossings in a long time for us), one of which I removed my boots for and ran across from one icy edge to the other: Martin asked me to repeat the feat for the camera...The canyon rose 80 meters over our heads, with sandy cliffs cut into by water erosion making for columns, caves, and some scree slopes looking like elephant's knees alongside us. We climbed back out of the canyon on a section of road that was so steep I had to do a double-take: what loaded truck could make it up this slope? It didn't seem possble, no matter how muscular the engine. As it was, we panted and pulled our way up and out, and camped near a side canyon issuing out of the range on the north side of the fault.

The next morning, we awoke to see a large herd of sheep grazing just over another side canyon. "We'll meet people in the next three kilometers" said Martin. We didn't have to move at all: as we were packing, we were approached by two men and a young woman, one of them on a horse. These were the first people we had seen in 15 days, and it was hard to say who was more surprised. They asked us to join them for naan, which tasted like your favorite food after two weeks of nothing but instant noodles. The men lit a small fire for keeping our hands warm, and we huddled behind a rock to get out of the chilly wind. The woman busied herself by digging through our bags, pulling out nearly everything, including our wallets. We had relatively little Chinese money (about 30 dollars worth), so they weren't overly impressed, and the US dollars I had as reserves were merely a novelty. They asked to have almost evey item in our possesion, but in the end I only parted with a cigarette lighter and Martin with a pen - we didn't have much extra, having pared down our gear about as much as possible. We thanked them as they packed a few more pieces of naan into my bag, and off we went, hoping for Qarasay.

We saw no other people or homesteads until the lowpoint of the fault, where the river we had been following made an abrupt 90 degree turn to the left and cut its way through the mountains to the desert 20 kilometers to the north. The fact that the road continued up another river valley made me think that that gorge must be fantastically narrow, because the road we continued on climbed up a gorge which was at times no wider than a truck, with no room to spare. It seemed incredible to me that someone would push excavators, dump trucks, and bulldozers through such a tight squeeze. Most of the time the road was simply a path up through the river, with little or no effort to make cuts or embankments. We camped our last night in the Kunlun at 3500m, 3km from the pass. At night, a couple dozen horses cantered past us to get water downstream, and when we woke, it was to a dusting of snow, although it hadn't been cold (only -5C).

As we packed, a shepherd came down the canyon driving perhaps a hundred sheep, and he watched quietly as I stowed my tent away. I gave him a light after he rolled a cigarette in a piece of newspaper, and asked him how far it was to Qarasay in the little Uighur I could remember from having travelled in the Central Asian republics 4 years before: "Qancha kilomet Qarasay ge?". He replied "Bir sad daban ge. Qarasay ikki sad." ("An hour to the pass, then two hours to Qarasay.")

I told Martin this, and in 20 minutes we were climbing the last hill over the pass in the sand and the snow. We reached the top, took a breather, and then pushed down the other side quickly, looking forward to something to eat. The road wound down through a rocky riverbed, and we passed another couple of shepherds along the way who just stared in silence at the two outlandishly dressed cyclists crashing over boulders and icy stream crossings in a hurry. There was one last bend in the canyon, and beyond, nothing, no horizon, just the sand blending into the haze in the distance: we had made it to the Taklimakan.