MEMOIR ON MAPS
OF CHINESE TURKISTAN AND KANSU

From the Surveys Made During Sir Aurel Stein's Explorations 1900-1, 1906-8, 1913-5

Sir Aurel Stein
Trigonometrical Survey Office
Dehra Dun, India, 1923

Title page of Memoir on Maps

Second title page for this volume

Dedication page of Memoir on Maps

Stein published a special map showing the watch towers and Han 'limes' north of Tun-huang

He also published excellent general maps in Desert Cathay and Ruins of Khotan

The list of illustrations in the Memoir on Maps

The cover page to Volume IV of Innermostasia - the set of maps published with that report

The Table of Contents - xi
Preface - vii-ix
List of Abbreviated Titles - x
List of Plates - xii
List of Maps and Charts - xv
Supplementary Corrections - xv
Introductory - 1-2
Chapter I - History of the Surveys - 3 - 40
Section i - General character of the topographical work - 3
Section ii - First expedition 1900-1 - 5
Section iii -Surveys of the Second expedition 1906-08 - 10
Section iv - Surveys of the Third expedition 1913-15 - 24

Chapter II - The Regions Surveyed - 41 - 54
Section i - The Tarim Basin and its mountain ramparts - 41
Section ii - The Taklamakan Desert - 42
Section iii - The oases of the Tarim Basin - 43
Section iv - The terminal depression of Lop and the Turfan Basin - 46
Section v - The Su-lo-Ho Basin -49
Section vi - From the Central Nan-shan to the Etsin-gol Basin - 50
Section vii - The Pei-shan and the Easternmost T'ien-shan - 53

Chapter III - The Maps - 55-62
Section i - Compilation of Maps - 55
Section ii - Representation of physical details - 57
Section iii - Symbols and Local Names - 59
Chapter IV - Notes on Individual Map Sheets - 63 - 105
Appendix A - A short summary of, and discussion into, the merits of the Triangulation -107-114
Appendix B - Notes on height observations -151-154
Index of local names - 155-198
General Index - 199-207
Addenda et Corrigenda - 208

Preface
Stein never neglects to heap praise on and give thanks to all those who made his expeditions possible and rendered assistance during them. In this preface he focuses naturally on those responsible for the survey, leaving out those related to the archeological and cultural aspects of the projects. He claims deepest debt to Colonel Sir Sidney Burrard, R. E. K.C.S. I., F. R. S. who was Superintendent of the Trigonometrical Survey. He describes the assistance and his debt to many other officials of the Survey as well. And he includes his three principal Indian surveyors; Rai Sahib Ram Singh, Rai Bahadur Lal Singh, and Khan Sahib Afraz-gul; respectively Gurkha, Sikh and Pathan. For some reason he omits mention of the fourth surveyor.

Introductory
Stein states that this volume is to record the topographical surveys during the three expeditions. He notes that his primary object was the archeological exploration and geographical study. But that with the dedicated assistance of the expert Indian surveyors he was able to accomplish everything in an area from the 75th to 102nd degrees of longitude and from the 35th to 44th degrees of latitude. The surveying involved both plane table and triangulation work. The careful detailed work of surveying enabled also time for study of the local physical features as they moved. Where possible he expanded the survey area by sending the surveyors on separate tracks.
He notes that the maps omit large sections of the Tarim Basin and surrounding mountains, unavoidably as they are virtually inaccessible. This memoir describes all three expeditions, but was prepared specifically to accompany a new set of maps, better in both content and cartographic expression than those previously published.

Chapter I - History of the Surveys
Section i - General Character of the Topographical work
Although this memoir is occasioned by the publication of the new set of 40 maps, which are described in detail in subsequent chapters, this chapter on survey methods applies to all three of the expeditions. Stein begins by discussing his general survey methods for 'reconnaissance survey' work. By that I believe he means rapid general surveying without the precision required for official records. For instance, he did not have instruments for determining longitude directly, and the astronomical observations for latitude were hampered by atmospheric conditions. The trigonometric base depended on relatively few known points for resections and of those many depended on questionable early surveyors. For the 190-01 survey they used a scale of 8 miles to the inch on plane table, but for the second and third surveys the scale was 4 miles to the inch a very adequate scale to record many topographic and vegetation details. In the mountains they worked hard (that is climbed treacherous ridges) to find the best locations for the plane table. They used the ancient cyclometer to measure distances while in the desert and flat lands. Stein describes in detail the difficulties in astronomic observations to determine longitude. The weather was either do rainy and cloud covered, or the air was full of dust or the temperature in the mountains was too cold. Moreover the rapidity of movement, frequently 25 miles a day, left little time for astronomical observations. For example, he writes, during the second expedition of 28 months in China there were 488 moves of the camp and he moved 8300 miles. During the third expedition the rate of movement was similar. Stein accumulated nearly 7,000 miles and R. B. Lal Singh nearly the same. There was no way to ascertain longitude as chronometers were not available and not accurate anyway. But he does note his disappointment at not having a wireless time signal during the third expedition. Heights were measured by theodolite and by Watkin mountain aneroids provided by the Survey of India. (see appendix). During the second and third expeditions these were checked periodically by use of two mercurial mountain barometers. He notes the use of hypsometric observations. (Deemed worthless by experts, see appendix). Heights of observed peaks were calculated by use of clinometers.
In order to expand the area being surveyed Stein detached the experienced surveyors on separate routes when possible and safe. The plane table surveying enabled recording details of soil, vegetation, water, and local topography directly on the map sheets. Stein notes that he paid particular attention to recording place names as closely to phonetic values of the local Turkish names as possible and that Chinese names were transliterated on a different basis.

Section II - First Expedition, 1900-01
Stein notes right off that he had an excellent assistant in Rai Sahib Ram Singh who had previously participated in survey expeditions into China. They entered Chinese Turkistan at the Taghdumbash Pamir via Hunza over the Kilik pass by the end of June 1900. He notes his use of photogrammetric work with a Bridges-Lee photo-theodolite. The results of this - panoramic views - are included in his reports. And they were published separately as Mountain Panoramas from the Pamirs and Kwen Lun. Starting at the Wakhjur pass at the origins of the Oxus river the survey extended over the Sarikol valley to where the Zarafshan River bends eastward near Tashkurghan. Triangulation was begun based on points previously surveyed by the Pamir Boundary Commission. The survey continued north past the mighty Muz-tagh-ata massive and Shiwakte mountain. Triangulation continued around Lake Little Kara-kul. He writes that Ram Singh's work has set the elevation of Shiwakte (Kongur) at 25,146 feet - higher than Muz-tagh-ata, 24,341 feet. He followed the Gez and Yaman-yar rivers through narrow gorges as far as Kashgar. (See Ruins of Khotan.) Ram Singh's calculation of the longitude of Kashgar was 76 degrees 1' 0" very close to subsequent determination of 75 degrees 59' 5.64".
Stein departed Kashgar at beginning of September for a short visit to some ruins northeast of the city. Then he traveled southeast to Khotan but not on the main road. Rather he chose a desert track via the pilgrimage site at Ordan-padshah in order to study the desert. They rejoined the standard caravan route at Kizil. This led through Karghalik and Yarkand. The survey was mostly limited to the line of march plus excursions to several close-by ruins. They reached Khotan on 13 October. While awaiting winter for desert work, Stein moved south into the K'un-lun up the Yurung-kash River finding opportunities for survey stations on high peaks and ridges. They reached the gorges in the northern most range of the K'un-lun and made views at Tope (13,949 feet). At Karanghu-tagh they found a colony of exiles. Crossing the Yurung-kash they managed despite reluctance by the local labor to move for 2 more days up the river until blocked by impassable gorges. He at least determined that the Yurung-kash forces its way through the K'un-lun via gorges south and west of Muz-tagh (K5) a peak previously fixed by the Indian survey from Ladak (south side). To return to Khotan Stein learned of another trail across mountains to the north-west. Doing this he mapped the glacier-fed headwaters of other streams passing through the K'un-lun south of Karanghu-tagh. With yak transport he crossed the passes between the Kash, Nissa and Chash valleys with excellent opportunities to establish survey stations on the ridges. He reached the watershed at Yagan-Dawan between the Yurung-kash and Kara-kash Rivers. Details are in Ruins of Khotan. (A Map from that book.) Despite setbacks he finally managed on 7 November to tie his triangulation into the India survey at Ulughat-Dawan (9,890 feet), the last pass above the Kara-kash river, with sightings on 4 peaks. On 9 November he repeated the process from a ridge above the Kunat Pass (10,820 feet). He counted himself fortunate because observations of any distance were soon obscured by the dust storms carried up from the desert.
Back at Khotan Stein prepared for the winter explorations while dispatching Ram Singh on 23 November back south to carry the triangulations from Muz-tagh and Karanghu-tagh eastward through the K'un-lun foothills. Ram Singh established base stations close to Karanghu-tagh and on a peak (14,900) above Ulugh-Dawan. He then moved eastward over the ridges and foothills, while triangulating peaks to 21,000 feet between the rivers leading to Khotan and Keriya and the Yulung and Nurfa rivers. When winter cold grew too deep in the high mountains he descended and continued to survey along the lower foothills eastward to Chira. He then moved north to Keriya and followed the river of that name north to join Stein at Dandan-uiliq on 23 December. Meanwhile Stein studied the archeological remains around Khotan until December 7 when he moved northeast into the desert conducting plane table traverses over 120 miles to Dandan-uiliq. Ram Singh conducted his survey over 500 miles, the last 130 of which were in desert devoid of landmarks for triangulations. Remarkably upon comparing plane tables at their meeting they found the error of closure as only half a mile of longitude and less than a mile of latitude.
After finishing excavations at Dandan-uiliq Stein and Ram Singh returned rode to Keriya and from there to Niya arriving on 21 January 1901. Five days' ride north of Niya Stein found in the desert the large area of remains abandoned in the 3rd century. While Stein did the excavating Ram Singh prepared detailed plot plans of the area and individual ruins. From Niya they moved further east through unexplored desert to another ruined fort and village on the Endere River. They started back west on 26 February to Keriya. They then marched for 7 days down the Keriya river back into the desert to find Karadong at which they found another ruined fort. Completing that excavation they moved west again surveying as they went to Domoko, Gulakhma and Chira. Continuing west they surveyed more ground north of Hanguya and found a very special site at Rawak. With spring dust and heat closing in they rapidly returned to Kashgar via Yarkand and Kizil. From Kashgar Ram Singh returned to India in charge of the equipment and Stein entered Russia. He took the railroad across Russian Turkistan carrying the archeological materials to London. The survey maps were prepared by the Trigonometric Branch Office of the Survey of India and published in May 1903. But Stein comments that the technical ability of the map office prevented them from doing real justice to the excellent plane table drawings. A better product was published then for the Royal Geographic Society Journal and in his personal narrative. (see Ancient Khotan - map included here.)

Section III - Surveys of Second Expedition, 1906-08
This section is an excellent summary that provides added clarity to the lengthy but sometimes discursive accounts in the official record and personal memoir. Stein notes the continued support by officials of the Government of India from the Viceroy, Lord Curzon, on down and that of the Survey of India as well. He again received Ram Singh as his associate surveyor. They started on April 28, 1906 from Fort Chakdarra in the Swat Valley. Taking a new route as usual, Stein rode through tribal valleys of Chitral and Mastij across the Hindukush to the Chinese border at the Wakhjur Pass on 27 May. From there they again passed through the Taghdumbash Pamir and Sarikol valley to Tashkurghan. There Rai Ram Singh began his triangulation with a base near Chushman. He was then to survey down the Tashkurghan (Zarafshan) River after it bends east over very difficult unexplored ground until blocked by the high state of the river. Then he turned north up the Pas-robat stream and across the Merki Pass at 15,000 feet to survey down the Merki and Kara-tash rivers that flow east from Muz-tagh-ata. But he was blocked again by spring flood water so crossed the Ghijak Pass to return to the caravan route between Sarikol and Kashgar. He again found the Kara-tash River after it leaves the mountains west of Yangi-hissar at which point he joined Stein in June. Meanwhile Stein took the main caravan route for 180 miles in 6 days across the Chichiklik plateau and Chihil-gumbaz and Izhiz-yar. At Kashgar he had to organized a much more extensive team and caravan for two years' of exploration. En route to Khotan Stein again detached Ram Singh at Kizil to survey a route into foothills of Muz-tagh-ata to a point meeting the caravan route from Chihil-gumbaz to Yarkand. Together again at Yarkand in early July they crossed the irrigated area between the Yarkand and Tiznaf Rivers north-west to the desert edge and then to Karghalik with new surveying along the Tiznaf River. From Karghalik they rode back south into the foothills to Kok-yar. There Stein remained at work (see his memoir - Desert Cathay) while Ram Singh into the mountains toward Khotan. Stein applauds Ram Singh's success despite hardships and dangers. This effort involved crossing ravines and gorges eastward between the upper Tiznaf and Yarkand Rivers. He was unable to cross the Karlik-Dawan due to heavy snow. Instead he crossed the spurs eastward surveying new areas near the Kara-kash River. He then crossed the range south over the Kilian pass (17,910 feet) which is the main pass on the route from Ladak to Yarkand. He then ascended the Kara-kash valley to its junction with the Karakorum route. He hired Kirghiz guides to cross the Hindu-tash-Dawan at 17,750 feet. From there he reached the Pusha valley previously known only from native reports. He then ascended the Ak-tash-Dawan (15,250 feet) to view the narrow gorges of the Kara-kash valley Ram Singh then used the difficult route north east over high ridges and glacier fed streams to reach lower part of the Kara-kash River. He finally made it to the triangulation station from 1900 at Ulughat-Dawan. From there he reached Khotan where Stein had arrived from Kok-yar on 5 August. Stein had surveyed the lower foothills. With it still being too hot in the desert Stein set out back into the mountains for another attempt to reach the sources of the Yurung-kash south of Karanghu-tagh. They reversed their route of 1900 up to Nissa. They did not repeat the plane table surveying but did expand the photo-theodolite work. South of Nissa they started the plane table again with less obstruction from the local Taghliks.
They met with two glaciers, sources of the Nissa river. On August 19th Stein climbed the larger Otrughul Glacier to 16,000 feet (much to the worried objections of the Taghliks - see Desert Cathay.) He triangulated the height of the source peak at 23,071 feet. He determined also that the Nissa valley did not extend as far south as he had believed in 1900. From Nissa Stein was still determined to reach the sources of the Yurung-kash. He had all along carried a sketch that an explorer, Johnson, had made during a crossing from Ladak northward in 1865. (Think he was spurred on by desire to surpass this gentleman.) Again refer to Desert Cathay in which Stein records the minor battle he had with the local Taghliks who refused not only to help carry equipment further south but also to admit they knew any such route.) Undaunted, Stein pushed his way using yaks into the Busat valley to more big glaciers which he climbed to determine that the way south was indeed impassable. Another attempt up another side valley was thwarted by the recalcitrant Taghliks. Here in a footnote Stein writes much more about Johnson and about his own later efforts to achieve success by coming from the southern side. Stein is pleased to remark that his own explorations revealed significant discrepancies in Johnson's map and report.
Stein next moved north to Pisha, while Ram Singh surveyed a new route to Khotan further east (see map). On 15 September Stein started for excavations north and north-east of Khotan. Ram Singh went to the foot of the main K'un-lun to survey south of Keriya and continue his work from 1900 to the east. Having reached Imamlar, Ram Singh picked up his triangulation points from 1900 and those of Captain Deasy. He expanded the survey along the mountains as far east as Charchan. Meanwhile Stein finished his archeological work and then surveyed a new route to Niya which he reached on 14 October. Ram Singh arrived and was able to triangulate on peaks in the K'un-lun not usually seen from that far into the desert. They both proceeded north to the old Niya ruin explored in 1901. Ram Singh prepared detailed plans and area surveys while Stein supervised archeological work. After completing this work they again separated. Ram Singh returned south to the mountains while Stein went through the desert to the east to Endere again. After more digging there he moved on to Charchan. He continued plane table surveying from Charchan along the river of same name to Lashkar-satma and then to Vash-shahri and Charkhlik. Stein moved north-eastward while Ram Singh returned south to continue triangulation eastward. But the severe cold exaggerated his rheumatism forcing him to cease triangulations and revert to plane table surveying. On 6 December Stein left Charkhlik for Miran where he found the remains of a fort and stupa. Ram Singh was suffering so severely he had to ride one of the camels, thus reducing the total supply load that could be carried to Lou-lan on the other 20 camels. They left much baggage and artifacts at Abdal on the Tarim river and took 50 hired laborers north. After walking for 7 days they reached Lou-lan right where Dr. Hedin's map showed it would be.
From 18 to 28 December they excavated and surveyed the Lou-lan ruins. Lack of time and Ram Singh's illness prevented extensive survey, which Stein had to leave for the third expedition. Ram Singh took the main party back to Abdal while Stein explored a different route further west to find the delta of the Tarim river. He reached the Ilek branch of the river on 3 January 1907. There he found a small ruined fort named Merdek-tim. Then he surveyed the confluence of the Tarim and Charchan rivers were they finally disappear into marsh and desert. From Charkhlik Stein returned to Miran to continue archeological work until 11 February. From Abdal they both rode east for the 330 miles along the shore of the dry salt basin of the Lop sea and then along the Su-lo- Ho to Tun-huang. It was along the river that Stein found the western end of the Han Dynasty limes, a series of watch towers and low wall.
Stein notes that the wall and its topographical location were fascinating from both a military historical and a geographical view point. He spent two months doing a detailed survey extending eastward for 160 miles to An-hsi. The details are described in Desert Cathay and Serindia. Back at Tun-huang Stein visited the small oasis to the south west at Nan-hu and then found the ancient "Yang barrier' north of the village along the western length of the Su-lo Ho. From the plane table sheets and more detailed sketches a special map at half-inch scale was prepared for Serindia. (See map). Stein mentions here his extended stay at Tun-huang to collect the manuscripts and paintings from the famous 'caves'. He then moved to An-hsi to store all the 'loot' at the Yamen. He left that town on 24 June to explore and map the Nan-shan. He reached the ruin of Ch'iao-tzu and then mapped into the high mountains between the T'a-shih rivers and the Tsaidam plateau. They crossed the Su-lo Ho in the mountains near Ch'ang-ma and then reached the Ming wall at Chia-yu-kuan. At Su-chou he had great difficulty recruiting labors willing to brave the dangers and mystery of the high Nan-shan. With strong edicts from Chinese officials he managed to obtain enough support to depart on 28 July across the Richthofen Range to an elevation of 13,500 feet. The party continued into the wilderness without seeing another human for a month. They spent August crossing the three main ranges with peaks above 18,000 feet riding for over 400 miles. He found the sources of all the rivers flowing down to Su-chou and Kan-chou. They found ample sites on high ridges for the plane table in passes over 15,000 feet elevation. He reached the source of the Su-lo-Ho in a marshy basin over 13,000 feet elevation and two lakes - Khara-nor and Koko-nor amid the 20,000 foot high peaks.
From there Stein rode across a high pass through the Richthofen Range to the uppermost sources of the Kan-chou River and then down across a succession of ridges where he could use the plane table to Kan-chou city. On September 3rd he started back west - actually north-west along the Kan-chou River at first and then to Su-chou and on to An-hsi. Along the way he made detours north of the Su-chou River to find the location of the Han Dynasty wall.
Back at An-hsi Ram Singh had to start back to India as his rheumatism would not stand another cold winter's work in the desert. Ram Singh managed to accomplish plane table work through the foothills from An-hsi back to Khotan. Stein had written to the Survey months before about this and was delighted that Sir Sidney Burrand sent out the extremely experienced senior surveyor, Rai Lal Singh who met the party at An-hsi. (When one reads of the varied locations from Arabia to China in which Rai Lal Singh worked during his long career one gets but a hint at the extent of the Indian government survey work).
Stein started from An-hsi on the 900 mile ride to Kara-shahr via Hami across the stone desert of the Pei-shan hills. Stein surveyed the route, which was the main caravan route of the Chinese from the start of their occupation of the Tarim basin. (It today is still the route shown on maps for the main highway between Su-chou and Turfan.) While Stein looked into various archeological sites at Ara-tam and Lapchuk near Hami Lal Singh surveyed the slopes of the Karlik-tagh. with its peaks up to 14,000 feet. They next spent 3 weeks in the Turfan depression during which Stein focused mostly on archeology and Lal Singh surveyed surveyed the wider general area up to the southern slopes of the T'ien-shan. On December 1st Stein moved north on to Kara-shahr while Lal Singh went south to survey more of the Kuruk-tagh to Singer and then back west along a southern route to Korla. He returned to accomplish much more surveying of the Kuruk-tagh in 1914.
At Kara-shahr Stein found more ruins to excavate including Buddhist temples at Shorchuk. By January 1st 1908 the both were united at Korla. From there they explored the low desert and marsh between the Inchike and Charchak Rivers in futile search of 'ghost towns' steadfastly proclaimed to have been seen by several local shepherds. Stein concluded this was another example of local folk lore. But the surveys were useful. They then proceeded west to Kucha, Stein moving along a northern track by the foothills of the T'ien-shan while Lal Singh surveyed a southern track up the Inchike River basin via Shahyar.
After completing more archeological work around Kucha Stein got the idea to cross the Taklamakan directly south to the area where the Keriya River disappeared into the desert. In this account he down plays the dangers and difficulties, perhaps because the whole adventure was rather a digression from the main tasks. He mentions that Sven Hedin had accomplished a 'pioneer journey of this sort in 1896. But he does not mention that Hedin had traveled north from Keriya along the river to its end and then had only to continue directly north to hit the Tarim river at some point. But Stein was going south from Shahyar with a limited supply of ice in expectation of finding a narrow point in the midst of the high sand dunes where an error of only a few miles would miss the river entirely. (Throughout his expeditions the reader can catch a subtle feeling that Stein was always competing with Hedin). Anyway, after 15 days they found the river's end and sure enough it had shifted far west of its location when Hedin saw it and made a sketch map. Stein has to comment that well anyway the journey was full of interesting discoveries and topographical observations.
The next archeological site was Karadong reached from the river by a new desert route north of Domoko. Again, Stein excavated and collected remains while Lal Singh surveyed along the line of oases from Keriya to Khotan. More interesting sites were excavated north of Khotan (very interesting places described in the official report) but not specifically part of topographic surveying so omitted here. Then in April Stein and company moved right back north again across the desert, this time going along the Khotan River which is the united flow of both Kara-kash and Yurung-kash rivers. He makes this excursion appear to be business, so does not mention here his desire to meet again his old friend from 1900 who had been promoted to Tao-tai at Aksu. But along the river they did find the very interesting Tibetan fort - Mazar-tagh - on a high bluff adjacent to the river bed. From the Khotan River they crossed the Tarim River and rode along the left bank of the Aksu river to the Chinese administrative headquarters at Aksu, reaching it in early May.
At that point Stein and Lal Singh again went their separate ways. Stein traveled north into Uch-Turfan valley and across part of the T'ien-shan to Kelpin oasis. Then he returned southward across dry hills to the end of the Kashgar River and then stopping at Tumshuk on the Aksu -Kashgar caravan route. He accomplished plane table surveying of the hills toward Marl-bashi before the heat drove him back to Khotan via the left bank of the Yarkand river to that town and then along the main route east by 9 June. At Khotan Stein spent weeks sorting and packing the large volume of artifacts for their trip to India. Meanwhile Lal Singh surveyed the entire route from Aksu to Khotan in the T'ien-shan at first and then through the mountains south to Kashgar and on to Guma, Karghalik and Yarkand. Unfortunately the level indicator of the theodolite prevented trigonometric triangulations, but the plane table work provided new details of local conditions. Lal Singh then mapped the previously missing areas in the K'un-lun between Kilian and the middle Kara-Kash River above Pujiya. He also tied his survey into that accomplished by Ram Singh in 1906 south - up the Kara-kash to Kilian-kurgan.
(His detailed description here of routes taken is much easier to follow than that in his official report in which he continues to comment on many other topics at the same time. They were circuitous in the extreme.)
By the end of July Stein was able to send the huge convoy of camels bearing the recovered artifacts south toward the Karakorum Pass. He had decided to attack the Yurung-kash from the south. To do this he and Lal Singh set out eastward from Khotan to go around the outer K'un-lun. Turning south they reached Polur and crossed the K'un-lun to Seghiz-kol lake. Moving again westward with the guidance of a local hunter they found the deep valley of the Zailik River draining into the Yurung-kash. From the Zailik they could climb adjacent ridges and make triangulations on the prominent peaks. They could see glaciers up to 21,000 feet at distances of 60 miles from the survey stations. At Zailik they found some gold miners who agreed for a price to help with the baggage. They pushed on down the Yurung-kash and over spurs and passes. Their survey station at Mandar-kol-Dawan at 18,612 feet enabled them to establish triangulation ties. After 7 more difficult marches from Zailik they reached the main gorge of the Yurung-kash in a glacier basin at 16,000 feet. They then turned back eastward to reach their supply depot at Ulugh-kol on 3 September. They then crossed the main range of the K'un-lun via the Baba-Hatim Pass at 17,584 feet. For three more days then moved back southwest to the plateau that holds the source of the Keriya River at 17,000 feet. He showed that the glacier ridge from which the Keriya flows is the same as that from which the Yurung-kash flows in the other direction. From there they moved again back westward to survey unexplored ground. They found enclosed lakes at elevations of 15,000 to 16,000 feet. By that time the weather was turning cold. Fodder was running out. They could not spare more time for detailed survey work as they continued west. The high plateau was very complex with steep ridges and broad valleys coming of the very high crests of the K'un-lun. After 6 days of moving from the Polur-Lanak-la route they reached the east end of a large salt lake. Continuing north-west for 3 more days the found on September 17th traces of the forgotten medieval route that Haji Habibullah had attempted to develop for direct communication between Keriya and Ladak. This was the route Johnson had used traveling north to Khotan in 1865. There were remaining cairns and other evidences marking the route. They were able to sight peaks of the India survey. On 18 September they reached the eastern valley of the Kara-kash. Stein and Lal Singh continued to search for Johnson's route up side valleys blocked by glacier. On 22 September they climbed a glacier looking for the watershed. At 19,000 feet they reached a crest from which they could see far to the north enabling them to tie the survey to the points established in 1900 and 1906 from the Khotan side. But instead of a view of tributaries of the Yurung-kash they found one of the Kara-kash. The triangulated a peak at 23,071 feet that blocked view of the Yurung-kash while forming part of the divide between the two rivers.
It was while continuing with determination photo-theodolite and survey work in the ice on top the glacier than Stein lost feeling in his feet, frostbite ensued. From that point his priority was to be carried by litter over the Karakorum to Leh which he reached on October 12. His personal narrative describes the entire events graphically.
Meanwhile Lal Singh was left to supervise the huge caravan of relics that had been waiting for them in the Kara-kash valley to cross the Karakorum. Stein applauds the honors bestowed in Lal Singh with distinguished title of Rai Bahadur and an award.
The survey sheets were converted into 94 maps by the Trigonometrical Survey Office at Dehra Dun. Copies of these were distributed and then bound into a volume of Serindia, when it was finally published. Stein rates these maps as a significant improvement over the map that resulted from the first expedition. However, his personal absence for 3 years in London working on the archeological results prevented his direct editing of the maps. Maps were also published in Ruins of Desert Cathay. One is of the entire area at a scale of 1: 3,000,000 and there are two maps at a scale of 1:1,000,000 of the mountain region of the K'un-lun south of Karghalik-Khotan and of the western and central Nan-shan. (We have the series from both Serindia and Ruins of Desert Cathay included here.)

Section IV - Surveys of the Third Expedition 1913-15
Stein notes his continued recognition that the effort during the second expedition nevertheless left many gaps in both mapping and archeological discovery. He continued to press authorities for permission and funding to make a third expedition. The support was forthcoming for a slightly longer tour. He again chose a different route through the unexplored mountains of north-west India (Pakistan).
(It is interesting to realize that as late as 1913 these territories remained under control of independent tribal chiefs and up to that time no European had been permitted entrance. See Kipling's Kim - Stein was a friend of Kipling's parents).
Stein's request for the assistance of R. B. Lal Singh was granted. He also was assigned two other surveyors, Mian Afrazgul Khan of the Khyber Rifles, and Muhammad Yakub. Of course there were many other servants, cooks and handymen who are mentioned in the memoirs but not here, since they were not surveyors. They started on 31 July 1913 from Srinagar to the northwest through across the Chilas and Indus rivers into the Dard chieftaincy of Darel and Tangir. With protection provided by local chiefs they were able to survey and triangulate points along the way. In his memoir Stein describes the trying conditions that existed as he negotiated rocky gorges and crossed many very high passes. The results are to be published separately. They crossed the Indus-Gilgit watershed and the Darkot Pass to the Chitral river. They then crossed the Chhilinji pass into Hunza and then crossed the Ming-taka pass into Sarikol on 7 September.
Proceeding back to Tashkurghan they resurveyed Sarikol. Continuing, they crossed the Chichiklik pass toward Kashgar. After reaching the Tangitgar gorge they split up to expand the survey. Lal Singh turned east to move more directly to Yarkand and Khotan in order to complete the previous survey work in the K'un-lun near Kapa. Afrazgul led the main caravan directly over the Ighiz-yar and Yangi-hissar toward Kashgar while completing plane table survey as he went. Meanwhile Stein and Muhammad Yakub surveyed a new route across the Merki pass and down the Kara-tash river valley. Seasonal flooding from the massive Muz-tagh-ata usually prevented passage through the narrow canyons. Stein considered himself fortunate that late snows had prevented the usual melting and runoff. He managed to cross the Buramsal pass at 14,940 feet in deep snow. The passage through the narrow gorges was, nevertheless. difficult and dangerous as they repeatedly crossed the raging river. They reached Kashgar on 21 September.
Stein again thanks Sir George Macartney for his hospitality and generous assistance with all aspects of the critical preparations for the expedition. Stein departed Kashgar on 9 October. He had his sights set on reaching Lou-lan as quickly as possible to take advantage of the winter weather. His first idea resulted in a failure. Thinking to make a shortcut to Khotan he started due east from Kashgar to the low hill at Maral-bashi and from there directly to Mazar-tagh on the Khotan River. His previous survey indicated the route was practicable as the ridge at Mazar-tagh probably extended clear north-west to Maral-bashi. He moved on east from Maral-bashi on 25 October. But after a few days of tramping up and down over exceedingly high sand dunes and with water running short, he had to turn back. He mentions in passing that Sven Hedin had tried to march east into the desert there but the result was destruction of his caravan. After 4 more days of extremely difficult movement and with the animals nearing exhaustion, Stein realized the danger and reluctantly turned back. Still, he insists on noting some new discoveries to justify the effort, namely relics of Paleolithic settlements. The caravan returned to the Yarkand river and turned north to survey along the left bank as far as Aksu area and from there east to the Khotan river which provided a safe route south. Along that river previously surveyed they again stopped at Mazar-tagh. Further new survey was accomplished along the Kara-kash river. Stein stopped at Khotan to organize the further expedition. He left again on 28 November for the 700 miles to Lop-nor. The necessity for great speed Stein writes, apparently feeling the need to excuse his following the normal and previously surveyed caravan route. At Domoko he managed some more excavation and again visited Niya and Endere for more archeological work. Along the way he recorded the changes in irrigation and agriculture since 1906. While moving on to Charchan they found that the clear winter air enabled them to triangulate on K'un-lun peaks far to the south. They departed from Charchan on 31 December 1914 for 7 more day's ride along the Charchan river. Along the way Stein learned about an aborted coup in Charkhlik during which the Amban had been killed and then the rebels also killed by Tungan troop. (See Innermostasia). The situation enabled him to move on into the desert of Lop without interference. They remained in Charkhlik 6 days during which R. B. Lal Singh rejoined the main group.
Stein describes Lal Singh's efforts. Between September and 31 December he had surveyed the mountains beyond Tash-kurghan to Kapa and then extended Ram Singh's control from 1906 into the K'un-lun for 5 degrees of longitude, establishing triangulation points on high peaks. Then extreme cold and snow forced him to lower elevations where he continued with plane table surveying toward Tun-huang while continuing to record astronomical observations of latitude and barometric observations of elevation. He continued as far east as Nan-hu oasis and the turned north toward Miran.
Stein led the whole expedition from Charkhlik on 15 January 1914 once more to Miran where they spent two weeks for further archeological digging. On 23 January Lal Singh started along the Tarim to Tikenlik where Abdurrahim (the experienced desert hunter) again joined him for survey of the Tarim river bed (the former Konche-darya). Muhammad Yakub performed a leveling survey of the Su-lo-Ho basin to show its drainage into the marshes. On 1 February Stein started out for Lou-lan with 35 men - his own permanent team and local laborers. They had 30 camels to carry the heavy sacks of ice and food for a month. Along the route then found more ancient relics amid numerous dry channels of the ancient river that fed Lou-lan. Stein claimed to prove that the river continued further east before turning south as the Tarim does now. And that the area was a river delta of many streams rather than a lake. He arrived at Lou-lan on 10 February. While he supervised the diggers he sent Afrazgul Khan to explore and survey the areas north and northeast of the main Lou-lan settlement. He found many more ruins along the former Chinese route north-east toward Tun-huang. Among these he found a fully fortified garrison post (Stein calls it 'castrum' as always using Roman terms). With the archeological work completed and Lal Singh returned from survey of the Kuruk-tagh hills Stein moved to the spring at Altmish-bulak to refresh the camels and collect ice and fuel. They departed on 25 February. Lal Singh was sent northeast to survey around the northern edge of the Lop salt sea while Stein searched for the ancient Chinese track directly across the sea from Lou-lan toward the caravan route between Tun-huang and Charkhlik. The dangerous and trying trek is described in Innermostasia. (The resulting route is shown on the new maps prepared for this book but not on the set of 94 maps created from the Second Expedition and published in Serindia). They reached Kum-kuduk on the Tun-huang caravan route on 6 March. Lal Singh arrived at the same time after completing a much longer but perhaps easier route around the salt sea. The following year Afrazgul Khan set out back from Turfan to complete the survey around the western side of the Lop sea. On 7 March they continued survey of the eastern side of the Lop Sea to the Kuruk-tagh. Muhammad Yakub joined at that point, having completed a full leveling line of 60 miles eastward from Kum-kuduk to Besh-tograk to show there was a drop of 250 feet in that distance, proving Stein's theory that in ancient times the Su-lo Ho reached the Tarim river basin. While Lal Singh and Muhammad Yakub expanded survey work along the Su-lo-Ho, Stein returned to the line of the Han Dynasty wall to survey it further eastward than in 1907. By the end of March they gathered for a rest stop at Tun-huang. Then Stein visited the caves at Ch'ien-fo-tung again while Lal Singh moved south into the western end of the Nan-shan mountains. Muhammad Yakub went north to survey more of the Su-lo-Ho. Then they all met again at An-hsi in mid April. Stein's search for and survey of the Han wall eastward continued past An-hsi. They crossed the Su-lo-Ho and found the wall and towers there were on the northern bank near the Wahg-shan-tzu hills that form the northern side of the Kansu corridor. The Han wall continued east to the point where the Su-lo-Ho flowing north from the Nan-shan turns sharply west. But the wall continued north-eastward along a smaller stream to Ying-p'an close to the Pei-shan hills.
Abandoning further search for the wall Stein and crew went to Su-chow town. From there on 10 May they moved north along the bed of the united Su-chou and Kan-chou rivers (then named the Etsin-gol) to Mao-mei oasis. In this area they again found trace of the Han wall. Far to the north, along the dry Etsin-gol, they found the massive ruin of Khara-khoto (the Etzina of Marco Polo). Stein hired local Mongol labor to survey the fortress city while Lal Singh continued to survey north to the end of the Etsin-gol on the Mongolian border. With the heat of summer excavation in the desert ceased. Muhammad Yakub too the camels northeast to graze while mapping another new area. Stein and Lal Singh returned south past Mao-mei and turned east to survey more of the desert and foothills north of the Kan-chou river. From Kao-t'ai they took separate routes to Kan-chou city.
On 6 July they departed Kan-chou for more exploration into the high Nan-shan. Stein wanted to complete the survey begun in 1907 by covering the area west of that accomplished in the first survey and east of the western Nan-shan south of Tun-huang. They reached a high plateau above 11,000 feet when Stein's horse fell backward on his severely damaging the muscles in his left thigh. Rendered fully out of commission he had to wait in camp while Lal Singh completed survey as far as where the two branches of the Kan-chou unite, then followed the larger branch further up surveying the area of To-lai-shan and Richthofen Ranges. The Chinese again refused go further into the mountains for fear of bandits (as they did in the second expedition also).
In mid-August Stein was carried back to Kan-chou where Lal Singh joined him. They split again as Stein rode north-west along the Kan-chou river while Lal Singh surveyed again deep into the Richthofen Range before turning north to met Stein at Hsiang-p'u on 26 August. From there they returned to Mao-mei where Muhammad Yakub brought the camels from the Etsin-gol. On September 2 they began the 500 mile trek across the stony desert of the Pei-shan hills. Local 'guides' again became lost but Stein's dead reckoning and compass work saw them through. They pushed through the eastern extension of the T'ien-shan to the Bai village. During October they rode west along the northern slopes of the T'ien-shan (in Dzungaria) to Barkul and Guchen. After Guchen they surveyed Jimasa and then turned south into the Turfan depression via the previously unsurveyed pass in the Bogdo-ula range at 12,000 elevation. Lal Singh took the camels further east across the Ku-ch'uan pass above Jam-bulak. Early in November they were all again together at Kara-khoja (ancient capital) in the Turfan depression. The heavy loaded camels also arrived directly from An-hsi led by Naik Shams Din. And Muhammad Yakub arrived from his extended survey from Hami to Lapchuk and then to Shona-nor and Pichan.
Stein remained around Turfan for fall and winter archeological work. With Shams Din and Afraz-gul he surveyed and excavated many ruined locations until mid February 1915. Muhammad Yakub created a large-scale plane table survey at 1 one inch scale to record the numerous ancient remains on seven sheets. Both Muhammad Yakub and Afraz-gul also completed large scale plans of each ancient site including Kara-khoja and Yar-khoto.
On 12 November Lal Singh departed for a difficult triangulation survey of the Kuruk-tagh to Singer and on to Altmish-bulak he oasis near Lou-lan. For this extensive journey into the desert Lal Singh was aided by Abdurrahim with is camels. Due to the usual dust haze they had to wait at Altmish-bulak and then again at Astin-bulak for several weeks to obtain sights for triangulation across the desert to peaks in the K'un-lun, which they accomplished on 23 December. Then with Abdurrahim as guide Lal Singh continued further north-east along the Kuruk-tagh and then back west into the Turfan depression. On 4 February he moved again back into the Kuruk-tagh to complete a new survey from Singer west to Korla. The difficulty of this project in spring sand storms kept Lal Singh at work until reaching Korla in April.
Meanwhile, on 6 February Stein sent Afraz-gul Khan from Kara-khoja back east to the Lop desert to complete more surveys around Lou-lan and along the eastern edge of the Lop sea (noted above). Stein sent off a large caravan of relics to Kashgar and continued surveying at Yar-khoto. He then also went south into the Kurug-tagh leaving Muhammad Yakub to finish the one-inch survey of the Turfan including survey of the shore of Lake Baghrash. Stein went to Singer where he picked up one of the Abdurrahim's brothers as guide. From there he moved westward to Shindi mapping more previously unexplored areas around Turfan. He then continued south-eastwards back across gravel plateaus to Yardang-bulak south of the Kuruk-tagh. In mid March he explored burial grounds and mapped the remaining course of the dry Kuruk-darya.
Afraz-gul Khan joined him near Yardang-bulak after completing another arduous survey into the desert. For that trek he was guided by another brother of Abdurrahim from Altmish-bulak due south to Deghar. He then surveyed the ancient remains north east of Lou-lan. After replenishing his ice supply he moved south-east to the point where the Chinese route from Lou-lan crossed the Lop salt basin and then surveyed the 'shore line' to the south west reaching the point where the dying Tarim river disappears into marshes. Then back north he want to survey the dry basin of the ancient Kuruk-darya, completing the survey of that unexplored area west of Lou-lan. From there he rejoined Stein at Yardang-bulak.
Together Stein and Afraz-gul Khan rode west along the dry bed of the Kuruk-darya to Ying-p'an, where it was crossed by the Turfan-Lop route where they explored another ruined fort that had guarded the main caravan route from Lou-lan. The also surveyed the area in which the dry Kuruk-darya met the Konche-darya. This was the area that during the 3rd century the still existing Konch-darya to shift from the former bed of the afterwards dry Kuruk-darya that had flowed further east to provide water to Lou-lan. From Ying-p'an Afraz-gul Khan went to Tikenlik to survey the Lop - Kara-shahr route along the Tarim and across the Inchike-darya to Kara-kum on the Konche-darya.
(In all this repeated survey work between Turfan and Lou-lan Stein was determined to settle historical issues about Lou-lan and its abandonment. Stein always combined survey work desired by his patrons in the Indian Government with the historical issues he wanted to pursue.)
Afraz-gul Khan then surveyed the basin of the Konche-darya back west to Korla. Stein reached the same place by riding along the foothills of the Kuruk-tagh. Along this 100 mile stretch of now dry desert he found a line of watch towers similar to and dating the same as those on the Tun-lung section of the Han Dynasty wall. His point again was to show that the Han Chinese route west lay through Lou-lan and then along the northern side of the Taklamakan in the foothills of the T'ien shan.
By the end of April Stein met his other surveyors at Korla bringing their separate survey lines to a common point. From Korla they again split into three separate survey parties to complete three lines west to Kashgar. Lal Singh kept to the north in the foothills of the T'ien-shan as far north as the snow would allow. Muhammad Yakub went by a southern route with most of the camels across the Konche and Inchike Rivers to the Tarim and then on toward Yarkand. Stein focused on a route containing more archeological remains. That is the remaining 900 miles of the main ancient caravan route through the oases there. While he stopped to accomplish archeological work Afraz-gul Khan would survey other sites off the main route. They reached Kucha on 14 April where they stopped for 3 weeks of local exploration. Stein found that the difference between current cultivated area and the much larger ancient area revealed by location of many ruins showed that formerly the volume of water from the Kucha and Muz-art rivers was much greater enabling a wider area for irrigation.
Lal Singh pressed north up the several routes to the crest of the T'ien-shan from Yangi-hissar to the Kara-Dawan and from Kucha up toward the Tengri-han mountain peak but could not reach the summit due to the continued depth of snow even in April at high elevations. Eventually he rejoined Stein at Ak-su. Stein reached that major town on 17 May along with Afraz-gul Khan. They remained two days. From there Lal Singh again ventured into the T'ien-shan to survey new areas through the hills to Kelpin and then Kara-jol to Kalta-yailak and Kashgar. Meanwhile Stein rode rapidly (300 miles in 11 days) directly to Kashgar via Maral-bashi, reaching the city on 31 May. Lal Singh arrived the following week and Muhammad Yakub after another week. Stein was rushing to prepare 80 heavy camel loads of his recovered artifacts for their transport across the Karakorum to India. The surveyors kept busy preparing 157 new copies of the plane table sheets made during the expedition. Stein departed from Kashgar on 6 July 1915 with Lal Singh for a final survey of the mountains between Muz-tagh-ata and Kashgar. Finally Lal Singh, while also in charge of the large caravan, surveyed the route from Kok-yar up to the Karakorum pass border with India. Stein departed into Russia to transit Russian Central Asia to Persia and then through southeast Persia back to the border of India (Pakistan today).

Chapter II - The Regions Surveyed
In this chapter Stein writes a valuable description of the topography and climate prevailing during his expeditions with comments also on previous conditions. This description is a brief summary to accompany the new set of maps which will include both the previous surveys and the new one of the Third Expedition, which fill in many gaps.
Section I - The Tarim Basin and its Mountain Ramparts
Stein begins at the west edge, where naturally all three expeditions began. The range separates the Pamirs from the Tarim basin and appears on map sheets 2 and 3 - an area from the head waters of the Kashgar River to the eastern end of the Hindukush. From the Nuz-tagh-ata and Ulugh-art the rivers flowing east pass through deep and narrow gorges of the Gez, Kara-tash and Tash-kurghan rivers to reach the basin proper. This basin slopes gradually eastward as far as the Lop salt sea. This entire drainage flows into the Tarim River which then dies out in marshes of Lop-nor. The direct distance from Kashgar to the Lop sea is about 850 miles. It greatest width is between Kucha in the north and the K'un-lun south of Niya, about 330 miles.
On the southern side the basin is bordered by the abrupt heights of the K'un-lun. The mountain peaks are under perpetual snow with glaciers in the high valleys behind. The peaks in the western part reach 23,000 feet elevation. (maps 6, 9, 10, 14, 15) The Yarkand river breaks through on the western side and joins the Tarim. The Khotan, Keriya and Charchan also break from the K'un-lun but only the Khotan does not disappear in the desert. The high valleys of the K'un-lun lack vegetation for cultivation. East of Keriya the mountain chain is narrower and is backed to the south by the Tibetan plateau. The northern edge of this mountain chain consists of a barren clay and stony glacis some 40 miles wide.
On the northern side the basin is confined by the high T'ien-shan mountain range which joins on the west with the Pamirs around Alai and the head waters of the Kashgar River. The range extends eastward to near Korla where the Konche- darya coming from Lake Baghrash enters the basin. (maps 24, 25) The highest peak of the T'ien-shan, Tengri-khan, is north of Ak-su. (map 11) West of Ak-su the chain has a lower series of ranges between it and the plain. (maps 4, 5, 7) East of Ak-su the mountains gradually become lower. (maps 16, 17, 20, 21). In comparison with the K'un-lun the T'ien-shan has greater moisture and hence vegetation and it does not have the dry glacis but immediate areas for cultivation. East of Korla are the desert uplands of the Kuruk-tagh border the Tarim basin. (maps 25, 28, 29, 32) This region becomes more arid to the east.

Section II - The Taklamakan Desert
Stein divides the basin into four major regions. This desert forms the majority of the area in the Tarim basin comprised of bare dunes of drift sand. The desert is bordered on the west, north and east by narrow cultivated areas created around the Tiznaf, Yarkand and Tarim rivers. Along the southern side the border is along the northern edges of a string of oases from Karghalik to Niya created by rivers from the K'un-lun. (maps 6, 9 , 14, 19). Further east there are small areas of jungle near small rivers, (maps 19, 23) as far east as Charchan on the Charchan river that flows east to Lop-nor. In earlier times the southern rivers flowed further north into the desert enabling cultivation around ancient sites now ruins.

Section III - The Oases of the Tarim Basin
Between the mountains north, south and west and the desert there is the narrow belt of oases. Those along the western and northern edge are some what different from those along the southern border. The western and northern cultivated areas between Yarkand and Kashgar to Korla contain more oases and practically continuous cultivatable land in between. The result is that this strip contains the chief east-west trade and military communications route throughout historic times. South of the series of oases the Yarkand - Tarim river flows west to east and is joined at various points by rivers flowing south out of the T'ien-shan. There is no such river flowing parallel to the mountain - desert line in the south. Interesting is the fact that the area under cultivation along the north is not close to and on both sides of the river. This is because the elevation difference between the western end and eastern terminus of the Tarim is so slight that the river is continually changing its main channels and flooding, making permanent cultivation extremely difficult.
Besides the series of oases south of the T'ien-shan there is also the detached oasis area around Kara-shahr, located in another small basin northeast of the others, north of the western end of the Kuruk-tagh hills, where the Konche-darya flows out of the central T'ien-shan to Lake Baghrash-kol, which acts as a reservoir, and then flows past Korla on to join the Tarim. The favorable cultivation potential at Kara-shahr made it an important center in ancient times.
The southern strip from Karghalik to Lop is much different. There the steep mountain range changes into a dry, stone, gravel area that falls gradually to the desert. There is only one major oasis, Khotan, between two major rivers, the Yurung-kash and Kara-kash. The oases are separated by desert. East of Niya it is about 350 miles to the next small oasis, Vash-shahri and then Charchan. Except Khotan the other oases are north of the dry gravel line where their river meets the desert. These rivers are largely absorbed into the gravel plain between their exit from the mountains and the desert line. At that point they emerge as springs. The result is a curious double flood. In the spring time the ice melts lower down and the springs flood in the annual "black water". Then later in the summer after the snow and glaciers have melted high in the mountains comes the second flood the 'white water' that flows above ground in the river beds. In many places between these oases there is enough water from springs to create a jungle and a few spring-fed wells. All the southern oases except Khotan and Karghalik are 'terminal' that is at the point where their river disappears into the desert sand. Stein found the ruins of ancient villages north of these in the desert where in ancient times their river still flowed. These oases are made possible not only by their river's water but also by the fertile soil that the river brings.

Section IV - The Terminal Depression of Lop and the Turfan Basin
At its eastern end the Taklamakan ends in the Lop salt sea (Lop-nor). It was formed from ancient times by the terminus of the rivers flowing east. Currently the principal one is the Tarim river that ends in multiple channels and then marshes and lagoons bordering on the dry salt basin. Along the south-eastern edge flows the small Charchan river. The Tarim basin also includes the desert areas south-east of the Charchan and north-east of the Tarim but west of the next ranges of hills, the Kuruk-tagh north-east and the Altin-tagh south-east. Other than the very small oases at Miran, Abdal, and Charkhlik the area is uninhabited.
The Lop sea area has been of significant geographic and historical interest for years. And the nature of the terminal area of the Tarim has been controversial. Stein devoted much attention to this issue during his second and third expeditions. He determined that the Lop sea extends about 170 miles from south-west to north-east and has maximum width of 80 miles. The south-west corner of the evaporated salt flat still receives some water from the Tarim river. Around the time of Christ, when the Chinese established Lou-lan, the Konche-darya (now the dry Kuruk-darya) still flowed east into the northern side of the Lop salt flat. ( A modern map of China shows that again there is a river system flowing parallel with the eastern flow of the Tarim but continuing eastward right to Lou-lan and the northern side of the Lop-nor basin.) The north-east wind has created erosion by scouring the clay terrain, especially north and north-west of Lop up to the Kuruk-tagh leaving mesas and ridges with steep ravines in between. Changes in the river flow and direction in the 3rd century resulted in abandonment of Lou-lan. (maps 26, 29, 30)
The far 'shore' of the Lop sea is along the Kuruk-tagh. A narrower 'bay' extends further east. The southern edge of the salt basin is near the foothills of the Altin-tagh. Between them is a narrow desert corridor and the western end of the Su-lo-Ho. Through this passage occasional caravans are still able to move during the winter.
The deep depression of the Turfan oasis lies north of Lou-lan and Charkhlik and east of the main range of the T'ien-shan - east of the break in that mountain range leading north to Urumchi. It is close on the south side against the extension of the T'ien-shan eastward and north of the Kuruk-tagh hills and plateaus. (map 28). It is isolated from the Taklamakan and the main part of the Tarim basin. At its deepest it reaches 1000 feet below sea level. The Turfan basin as a whole is a small replica of the Tarim, with cultivated areas, desert and salt flat. Much of the irrigation depends on a unique system of underground tunnels dug for considerable distance that enable water to reach distant areas without evaporation en route.

Section V - The Su-lo-Ho Basin
The corridor though which the Su-lo-Ho flows from east to west is only 30-40 miles wide between the very high Nan-shan mountains to the south and the low but desolate, dry Pei-shan to the north. It extends some 220 miles west to east from its delta in marsh near Lop-nor eastward to where it flows north, breaking through the Nan-shan in a narrow gorge and then turns sharply west, then flowing through the Kara-nor lake. (maps 35, 38, 40) The Su-lo Ho is a major river flowing out of glaciers far to the south. The Yu-men-hsien oasis is located at the point where the river bends west. There is only one tributary along this stretch, the Tang-ho or Tun-huang river flowing north past Tun-huang oasis. Tun-huang is the only significant village in this entire corridor, which resulted in its being a significant base for Chinese operations into the Tarim. The corridor was the only access route west for the Chinese. For this reason the emperors in the 2nd century B.C. constructed the defensive wall and towers to protect the route to Lou-lan from the Huns. South and south-west of Tun-huang there are huge sand dunes against the Nan-shan created by the same north-east wind that erodes the corridor as is does the Tarim basin. Away from the river banks the corridor is dry gravel and clay. East of the Su-lo Ho is another interior basin.

Section VI - From the Central Nan-shan to the Etsin-gol Basin
The road south-east from Yu-men-hsien passes through the famous Chia-yu-kuan gate in the medieval Ming Great Wall. The first major town is Su-chou and southeast of it is Kan-chou. (map 46). This is the area in which the Kan-chou river flows north after coming out of the eastern Nan-shan. The southern area includes the three great ranges of the Nan-shan - Richthofen, Alexander III and with peaks over 18,000 feet elevation. The northern area includes the united Su-chou and Pei-ta-Ho which form the Etsin gol flowing north to the Mongolian border. The Central Nan-shan is wetter than the western and of course central Asia. This can be seen by the extensive forests on the ridges and grazing areas in some valleys. There is abundant snow and rain to provide for cultivation along the Kan-chou river from the city northward. This region was the base for Chinese expansion and commerce into Central Asia. Besides the two main cities there are numerous smaller villages between them. But the northern area is much different, more like Central Asia. The hills are barren and the Etsin gol flows intermittently through desert to two small lakes. There are narrow cultivated areas at the Mao-mei and Chin-t'a oases. The Etsin gol has always provided an avenue for commerce and invasion from Mongolia. (map 44)

Section VII - The Pei-shan and the easternmost T'ien-shan
The north-eastern area included in Stein's surveys contains the eastern end of the T'ien-shan and the barren Pei-shan hills. Together with his assistants they could only survey through this desolate region along several separate lines but had to skip regions in between. The Pei-shan forms the barren, dry hill range that borders the Su-lo Ho corridor on the north from the Etsin-gol west to join the Kuruk-tagh. On the north-west it meets the foothills of the T'ien-shan. The well known ancient caravan route from An-hsi to Hami crosses the Pei-shan. (And today this is still the main highway from Kan-chou to Urumchi.) The other route surveyed went from Mao-mei to near Karlik-tagh. (maps 37, 38, 40, 42), The Pei-shan barely reach 8,000 feet elevation. The eastern extension of the T'ien-shan extends for 300 miles from north of Turfan eastward with its main range at 13,000 feet and then gradually becoming lower as it turns into plateau and desert. The eastern section is called the Karlik-tagh from which short streams and springs provide water for local cultivation near the foothills. Hami is a significant oasis (map 34) providing provisions and rest along the main route from An-hsi to the T'ien-shan oases. Further west are the oases of Lapchuk - Kara-dobe. There is a 150 mile long gap of possible cultivation between Kara-dobe and the Turfan basin at Chik-tam. Stein also surveyed sections of the north eastern side of the T'ien-shan and noted the striking difference in climate and therefor agriculture. This is part of Dzungaria the home steppe of the nomads. The T'ien-shan peaks are high enough to retain perpetual snow which creates the water which enables forests to grow on the northern slope. The chief center here is Barkul near the lake of same name. On the southern side cultivation is only possible close up to the foothills supported by irrigation from the snow melt which appears from springs. The oases toward the east, Tash-bulak and Khotun-tam are very small but Hami and Kumul (map 34) are larger and important for their agricultural resources which from ancient times have supported the main caravan route from An-hsi, especially after the middle route via Lou-lan was abandoned.
Further west along the foothills are Lapchuk and Kara-dobe where more cultivation is possible due to sub-soil drainage from the snow melt. But, again, west of the foothills opposite Barkul there is again no drainage nor cultivation for 150 miles to Chik-tam in the Turfan basin.
On the northern slopes of the T'ien-shan a completely different climate exists. The wide Dzungaria plateau fill the expanse clear north to the Altai mountains. The ample grazing available on this steppe has made it the home of many nomadic groups from Huns to Mongols. At the eastern end, around Bai, there is the typical bare gravel slope. But further west along the Karlik-tagh there is extensive grazing. The snow melt from the higher peaks provides ample water for Barkul. Again, west of Barkul the mountain range is lower, hence is without perpetual snow. The result is much less water flowing north. Further west the mountains again reach higher elevations dividing the Turfan basin from the fertile area around Guchen. Cultivation there and at the ancient capital, Pei-t'ing, does not depend on irrigation. The vastly different geographical areas in the Turfan basin and around Guchen, separated by a narrow mountain range with low passes, is quite striking.

Chapter III - The Maps

Section I - Compilation of Maps
The set of 47 maps was prepared, beginning in 1916, at Dehra Dun. However, Stein notes, his absence in London to work on the archeological horde prevented his close supervision of this stage of map preparation. His return to India in 1917 enabled him to resume such supervision. This project was huge, as shown by the devotion of 15 draftsmen plus three senior supervisors to the effort during 1917-1919. Stein credits the personal attention of R. B. Lal Singh with assurance of accuracy. The project was not completed until 1922.
The maps were based on the triangulations and astronomical determined latitudes complied during the expeditions. In an appendix Major Mason has provided tables of latitude, longitude and elevation for the survey stations. In Chapter 4 there is a complete list of latitudes and other information about the methods used. The final compilations have also used information on latitudes and longitudes obtained from the travels of other explorers. But the topographic detail rests entirely on the plane table surveys from the three expeditions. Where terrain was surveyed more than once, the later results were preferred. Throughout Stein made use of his daily diary and detailed text descriptions of the vegetation, water, soil type and similar subjects. Elevations were obtained by use of mercurial, barometric, aneroid and hypsometric instruments. During the second and third expeditions Stein also used the clinometer - theodolite - to estimate elevations of points observed from known stations. Stein provides the latitude data for each station with his text discussion of each individual map sheet. For each map sheet Stein includes references to the relevant chapters in each of his reports and personal narratives.

Section II - Representation of Physical Details
Elevations have been shown by use of contour lines where sufficient data makes that possible. Depiction of the snow line and differentiation of snow from glaciers has been very difficult and of course on the far side of ranges seen from only one side is very approximate. Permanent water forms are shown in blue. Permanent river beds are shown by blue stipple, otherwise temporary rivers are shown by blue lines along the banks. The lettering of names of water features are also in blue. Depiction of salt crust and formation required use of three new symbols - all in black. Drift sand without vegetation is shown in brown. But sandy areas supporting some vegetation are shown in light yellow. Living vegetation is shown in green, dead in black. Ancient ruins are shown in red.

Section III - Symbols and Local Names
Standard symbols used by the Survey of India are used for man-made objects (other than ancient ruins). In Kansu practically all villages are walled. The fort symbol is restricted to larger towns. No symbolic distinction is made between different classes of roads. But those main routes marked by the Chinese with 2-mile towers are shown in double red lines. The actual survey routes are shown with series of black, blue or red crosses for the 1st, 2nd and 3rd expeditions respectively, and camps are numbered. For the first and 2nd expeditions only the camps occupied by Stein in person are shown. But for the third expedition the camps used by Lal Sing, Muhammad Yakub and Afraz-gul Khan are shown with different type faces. Various type faces are used for different categories of town, villages, and other locations. Rendition of the proper local name for villages and other places was attempted in Turki for Turkestan and Chinese for Kansu and was always difficult.

Chapter IV - Notes on Individual Map Sheets
In the notes for each map sheet Stein refers to his descriptions in Serindia, Innermostasia, Ruins of Desert Cathay, Ruins of Khotan, The Geographic Journal and Ancient Khotan, There is a list of locations at which astronomical observations for latitude were taken in each map sheet.

Notes on Sheet # I Turug-art-dawan
This map is compiled from Lal Singh's survey from Kashgar north to the Turug-art dawan (pass) on the Russian border.

Notes on Sheet #2 Kashgar, Muz-tagh-ata
This map depicts a large area along the mountain range that contains the Chinese-Russian border southwest of Kashgar. It connects the K'un-lun on the south with the T'ien-shan on the north. Muz-tagh-ata is the massive (in both elevation - 24,388 feet - and circumference) peak that dominates the area. The oases at Kashgar, Opal and Tah-malik are shown as well. Separate parts of the map were surveyed during each of the three expeditions. In addition information was obtained from the Pamir Boundary Commission and Captain Deasy. Kashgar's location is derived from several surveys as 39 degrees 28' 45" north latitude and 75 degrees 58' east longitude. The location of the Kosh-bel pass in the north is derived from Russian surveys. The map contains three regions, high plateau and valleys on the western side of the south-north main range, the much dryer valleys on the eastern side and the area north of the Gez defile which has greater moisture. The various Stein followed over parts of this map during his three expeditions are described in the several official reports and personal memoirs.

Notes on Sheet #3 Sarikol
This map is the southern part of the Taghdum-bash Pamir, south of Muz-tagh-ata, and the main part of Sarikol, the area of China directly adjacent to the Wakhan corridor to the west and to Russia to the northwest and Pakistan to the south. The area was thoroughly surveyed (triangulated) by the Pamir Boundary Commission in 1905 and Captain Deasy in 1896-8. Stein and his assistants made detailed plane-table surveys south of Tash-kurghan in 1900 and 1913. North of that town Rai Ram Singh completed triangulation in 1906. The map shows the several routes taken by Stein's team in different years. The route to Yarkand and Kashgar was surveyed in 1913.

Notes on Sheet #4 Yai-dobe
This map was surveyed along two routes followed by R. B. Lal Singh. One in 1907 was from Uch-Turfan to the Taushkan river along the T'ien-shan to Terek-dawan. The other was in 1915 and covered unexplored area from Kelpin via Kirghiz grazing at Yai-dobe and Chong-kara-jol to Kalta-yailak. Since no fixed points by astronomical observation were possible for this map it was fixed by adjustment from the adjacent sheets.

Notes on Sheet #5 Yangi-hissar, Yarkand
This map includes survey routes from all three expeditions. In 1900-01 the route was mostly within and between Kashgar and Yarkand. The routes in 1906-08 included ground all along the Yarkand river as well as portions of the hill area to the north-west and south-west corners of the map. The routes in 1913-15 were mostly from Kashgar to Maral-bashi and along the right bank of the Yarkand river below Yarkand. The map is based on very well established stations at Yarkand, Yangi-hissar, Kashgar, Maral-bashi and Karghalik. Ram Singh's plane table work has been matched with surveys made by other explorers. The topography is from Stein's personal observations. The area includes much variety from well-cultivated tracts to small oases and areas of drift sand.

Notes on Sheet # 6 Karghalik, Kilian
This map shows the mountains south of the Karghalik - Khotan road as far south as the upper reaches of the Yarkand river. This is the south-west corner fo the Tarim basin where the K'un lun mountains join the Hindu Kush and Pamirs. The surveys were accomplished in 1906 and 1908 plus survey along the road itself in 1900 and a survey completed in 1915 on the route south along the caravan route from Kok-yar to the Karakorum pass along the Tiznaf river and headwaters of the Yarkand river. Astronomical observations for latitude were accomplished at Karghalik and Kok-yar. A longitude for Karghalik of 77 degrees 26' 30" East. was interpolated from data from Captain Deasy and Colonel Trotter. The map shows an approximate elevation of the snow line at 17,500 feet. The oasis at Karghalik has extensive cultivated areas, but the other oases on this map sheet are small and located where rivers from the K'un lun leave the mountains: Kokyar, Yul-arik, Ushak-bashi, and Kilian. The high valleys contain sufficient vegetation to support annual grazing.

Notes on Sheet # 7 Ak-su, Uch-Turfan, Kelpin
This map is of the northern side of the Tarim basin and foothills of the T'ien-shan. It covers the area in which the feeder rivers of the Ak-su flow south from the mountains to the banks of the Yarkand-darya. The routes between Ak-su and Uch-Turfan south to Kelpin and Tumshuk wee surveyed in spring of 1908. Those further south toward Maral-bashi were made in 1915. Lal Singh's theodolite was damaged. Therefore the latitudes and longitudes shown are based on work by Mr. C. Clementi. The longitude for Ak-su is established as 79 degrees 55' 25" East and latitude of 41 degrees 7' 57" North. The large cultivated areas around Ak-su and Uch-Turfan are possible due to extensive irrigation from the Taushkan and Kum-arik rivers which unite close to Ak-su (Yangi-shahr). South of the area is a wide desert plain of bard clay and gravel covered by sand

Notes on Sheet # 8 Maral-bashi
This map is of the area adjacent to Maral-bashi and parts of the Yarkand river with the Taklamakan desert to the south-east. Stein's route in 1908 was mostly along the road from Tumshuk towards Yarkand. Additional areas were added from surveys in 1914 and 1915. Mr. Clementi gives the latitude for Maral-bashi as 38 degrees 46' 44" North. The longitude is 78 degrees 15' 15" East based on interpolation from Kashgar and Ak-su. The Marl-bashi oasis irrigation from the Kashgar river is insufficient but it supplemented by extensive dikes and reservoirs which create the largest irrigation system in the Tarim. Stein's attempt to short cut his route to Khotan was blocked by the huge sand dunes south of the Yarkand river.

Notes on Sheet # 9 Khotan, Sanju
This map shows the large mountain area from the edge of the Taklamakan between Guma and Khotan south into the main K'un-lun range and the upper Kara-kash river valley. The routes surveyed there were almost all from the first and second expeditions. Triangulations and astronomical observations of latitude in the southern part were accomplished in 1900. The longitude for Khotan adopted is 79 degrees 55' 51" East from a calculated mean of that obtained by several surveyors. In turn the positions adopted for Khotan and Karghalik were used as bases to fit the plane table surveys between those cities. The elevation of the snow line ( 17,000) adopted is estimated from observations during fall of 1900 and late summer of 1906. Although the elevation between the Taklamakan desert and the high ranges of the K'un-lun are extreme the entire area is extremely dry. The deep fertile soil around the oases has been created by the combination of deposits from dust storms and from the rivers. High in the mountains the valleys are so narrow that only limited areas contain sufficient vegetation for grazing. Movement on a north-south axis through the mountains is difficult due to the impassable nature of the river gorges.

Notes on Sheet # 10 Karakorum, Khitai-Dawan
Most of the area of this map was surveyed during Stein's and Lal Singh's epic trek east to west through the high plateau during the second expedition. The remainder, along the narrow route south to the Karakorum pass was surveyed again by Lal Singh during his movement of the camel caravan over that pass at the conclusion of the third expedition. The map shows the ranges of the K'un-lung around the massive peak Muz-tagh and the sources of the Yurung-kash river that eluded Stein for years. It shows the interior basin of the western Tibetan highland. The triangulation was based on three prominent peaks. The snow line is estimated at 18,000 feet relying on results of the expedition of Sir F. DeFilippi. The areas observed were naturally limited to those visible from the routes traveled. They comprise three distinctive regions; completely sterile basins of north-west Tibet, the main K'un-lun range with its deep valleys draining northward into the Yurung-kash river, and the high open plateaus between the Karakorum and Kara-kash river.

Notes on Sheet # 11 Muz-art
This map contains the area around the route leading from the south to the Muz-art pass in the main T'ien-shan mountains and a small part fo the outer hills above Kara-bagh. The plane table survey was accomplished by R. B. Lal Singh adjusted by the position of Tengri-khan Peak. But Lal Singh was unable to reach the summit of the pass due to the heavy snow remaining in May.

Notes on Sheet # 12 Kara-Yulghun, Bai
One route shown on this map was that between the end of the Khotan river to Ak-su and then north-east in 1908. The others were surveyed in 1915 by Lal Singh, Mian Afraz-gul, Muhammad Yakub and Stein. The locations are adjusted for those adopted for Ak-su and Kucha and for the junction of the Ak-su and Yarkand rivers. Dr. Hedin's observation of the latitude of that confluence was 40 degrees 28' 47" East. The longitude is the mean between values interpolated between Kashgar and Korla and between Khotan and Ak-su. Dr. Hedin also obtained latitudes for points along the Tarim river. The Bai basin between the T'ien-shan main range and another eroded hill range to the south in in the northern map section. The Muz-art river and its tributaries supply irrigation. The area between Ak-su and Jam, receives water from the T'ien-shan. Between the outer foothills and the Tarim river jungle belt the area is desert, but the locations for some wells make movement between Kucha and Ak-su possible along an ancient route. South of the Tarim river the Taklamakan desert takes over.

Notes on Sheet # 13 Mazar-tagh, Kara-dong
The surveys for this map were at the terminal areas of the Khotan and Keriya rivers in the Taklamakan. The first was explored during the second and third expeditions and the second during the first and second expeditions. There were numerous station at which astronomical observations for latitude were possible. But for longitude the routes had to be adjusted for the accepted location of Khotan and the Tarim-Ak-su river confluence and for those at Kucha and Koehkar-oghil. Extended archeological surveys with surveyed plans were made at Kara-dong and Mazar-tagh.

Notes on Sheet # 14 Sampula, Chira, Keriya
This map lies directly west of map sheet # 19. The map covers the mostly occupied ground between the Khotan and Keriya rivers along the cultivated stretch plus the desert areas to the north that include the ruins of ancient sites previously irrigated, and the gravel and stone plain stretching south to the northern ridges of the K'un-lun mountains. There were numerous surveys throughout the region during all three expeditions. Across the center lies the ancient caravan route connecting the string of oases. The norther surveys were made in conjunction with the archeological excavations of the ancient ruins. In addition to the Stein survey work that of Captain Deasy was used for control. There were numerous observations of latitude. Longitudes were interpolated from those established for Khotan and Niya. On the Keriya river the longitude for Kochkar-ughil was established from those at Keriya and Kucha and between Khotan and Domoko-bazar and Dandan-uiliq. The routes along the Yurung-kash and Kara-kash rivers were adjusted to the longitudes for Khitan and the Tarim-Ak-su rivers. The snow-line has been estimated at 16,500 feet.
The map shows the typical three topographic zones; in the north is the sand-dune desert with the two rivers cutting through in narrow but shifting beds; in the center is the zone of subsoil drainage from the rivers at higher elevations creating springs as they emerge from the gravel and clay to create a jungle belt and irrigated agriculture. At the ends of this belt are the larger oases of Yurung-kash, Sampula, Lop and Keriya. To the south is the barren plain of gravel and clay up to the mountain range. In a few of the wider valleys there is sufficient moisture and fertile soil to support summer grazing.

Notes on Sheet # 15 Yurung-kash and Keriya River Sources
This map is of the area east of that in map # 10. It was mostly surveyed by Stein and Lal Singh during their trek from east to west across the high plateau behind the norther ranges of the K'un-lun during August and September 1915. The Muz-tagh and northern K'un-lun peaks were seen in 1900-10. Many of the triangulated peaks were fixed by the Kashmir G. T. Survey in 1862 and by Captain Deasy. The snow -line was estimated at 17,500 to 18,500 feet elevation. During the second expedition Stein and Lal Singh circumvented the impassable K'un-lun ranges south of Khotan by traveling east and around the worst areas and then south and back west across the inner drainage-less plateau (ending on map sheet 10). During this trek they located the sources of the Keriya and Yurung-kash rivers. The area includes both wide basin plateaus and very deep cut narrow river gorges. There are many glaciers.

Notes on Sheet # 16 Kere-bazar, Bai
The survey for this small map was limited by atmospheric conditions that blocked observation of distant mountains during Lal Singh's survey in April 1915. The survey was fitted to the location of Kucha and latitude of Kara-kul and to the route lines from Kucha to Muz-art pass and between Kucha and Korla.

Notes on Sheet # 17 Kucha
The map was compiled from surveys in 1908 and 1915, especially the latter during which extensive archeological work was conducted there. Mian Afraz-gul worked on plane table details while Lal Singh surveyed the outer slopes of the T'ien-shan to the north. The routes south to Shahyar and further were mapped during the daring expedition directly across the Taklamakan in January and February of 1908. The astronomical latitude (41 degrees 42' 58" North ) results for Kucha agree closely with that of Mr. Clementi. Longitude (82 degrees 53' 30" East) was determined by taking the mean values and interpolating between Kashgar and Korla and between Korla and Tengri-khan.
The map shows part of the north-western section of the Bai basin. There are three other zones; north is the foot hill of the outermost spurs of the T'ien-shan and the wide alluvial fan from the Muz-art and Kucha rivers which create the oasis at Kucha. But this oasis was much larger in medieval times. To the south is the wide basin of the Tarim river with its jungle and various meandering beds that block northward movement of the Taklamakan sand dunes. This desert forms the third zone. In the norther part of this the dunes are oriented east to west parallel to the Tarim river but further south they are N. N. E. to S. S. W. corresponding to the lie of the Keriya river.

Notes on Sheet # 18 Keriya River End
Most of this map was compiled from Stein's trek in February 1908 north to south across the Taklamakan and up the Keriya river. Both an old bed and the current bed of the river were located. Latitudes were observed astronomically and longitudes were adjusted from those at Koehkar-oghil and Kucha. The Niya site was surveyed during archeological excavation work there.

Notes on Sheet # 19 Niya
This map adjoins map # lying to its west. The northern section of this map includes surveys conducted during all three expeditions. The important archeological ruin served to generate repeated detailed plane table surveying. The mountains far to the south were surveyed in the fall of 1906 as Rai Ram Singh was completing triangulations along the K'un-lun. In addition the surveys by Captain Deasy were incorporated. Niya-bazar itself was fixed in 1906 at Latitude 37 degrees 3' 34" East and Longitude 82 degrees 45' 32" North. The caravan route from Niya east to Endere and Charchan was checked by the position of Kalasti. Since no reliable observations of the actual snow-line were made the height at 17,500 feet is a guess.
The archeological discoveries at Niya and Endere generate questions about the historical geography of the region. it is divided into the same three topographic zones from north to south; desert, jungle belt and gravel plain up to the mountain spurs. In the northern desert belt are the terminals of the Niya, Yartungaz and Endere rivers. The first and third of these rivers once supported irrigated oases now ruins in the desert. There is only one small oasis in the jungle belt, at Niya-bazar. The mountains have a few valleys in which small hamlets manage to exist.

Notes on Sheet # 20 Kara-dawan, Kara-shahr River
The map includes part of the wide valley of the Kara-shahr river near its eastern end and a part of the outer range of the T'ien-shan that divides it from the Tarim basin. The range was surveyed by Lal Singh along its southern edge in 1915 (see map #21). The south-eastern corner was surveyed by Stein in 1907. The plane-table survey was adjusted to the positions for Kara-shahr, Korla and Bugur (maps 21 and 24). Lal Singh noted pine forest on the southern slopes of the mountain range at elevations above 8000 feet.

Notes on Sheet # 21 Bugur, Korla
This area is similar to that in map 49 of Serindia. The map contains part of the north-eastern corner of the Tarim basin. The survey from Korla to the Inchike-darya was accomplished in 1908 and the others were north and south of this in 1915. Unfortunately the map was compiled before it was realized that Lal Singh's triangulations based on observations of the K'un-lun across the Taklamakan were wrong. This requires that the location shown for Korla be moved 15' 30" East , but with the same latitude. The longitude for Korla in Serindia map 40 is correct. Again, Mr. Clementi's survey and that of Sven Hedin were incorporated. The longitudes along the Inchike-darya were adjusted for those at Peres and Shahyar. The Korla oasis is significant due to its extensive irrigation water from the Konche-darya and the reservoir of the Baghrash lake (sheet # 25). The large Bugur oasis appears to benefit from the Kizil river bringing water from the snow of the T'ien-shan.

Notes on Sheet # 22 Charchan
The map is of the south-east corner of the Tarim basin. The south-eastern corner of the map shows the area about the Charchan oasis and the Charchan river east. The survey routes here were the same west of Charchan during both expeditions but in 1906 and 1913 along opposite banks of the river to the east. On December 28, 1913 the exceptionally clear weather enabled Stein to triangulate the position of Kalasti (camp 116) from observations of four peaks in the K'un-lun previously established (Serindia map 46).

Notes on Sheet # 23 Kapa, Achchan
The map contains a part of the northern main K'un-lun range surveyed in 1906 and again in 1915 from the lower hills past the gold pits at Molcha and Kap to the Charchan river and also the ground crossed in both years along the desert route between the Endere river and Charchan. The K'un-lun range is based on the triangulation accomplished in 1906 by Rai Ram Singh and continued eastwards from Ushlung by R. B. Lal Singh in 1913. The desert route was adjusted from the positions of Niya and Kalasti. The time of the survey, late fall prevented accurate estimate of the elevation of the snow-line, but a guess is at 17,000 to 17,500 feet. The desert route is along the southern edge of the sand belt that contains some scrubby jungle and tamarisk-cones fed by underground drainage. Further south and higher in elevations the piedmont gravel and various temporary flood beds but other wise waterless.

Notes on Sheet # 24 Kara-shahr
The map shows along its southern edge the survey route along the main road from east to west through the Kara-shahr basin accomplished in 1907 and again in 1915. The map was adjusted for positions determined by Turfan (map 28) and Korla (map 21). The longitudes are in error due to the error at Korla (15' 30" too far west). (notes sheet 21). The location of Kara-shahr is also too far west. The latitudes come from those determined by Mr. Clementi and various Russian explorers. Except for the area east and north-east of Kumush the area is within the drainage from Baghrash lake.

Notes on Sheet # 25 Konche-darya
The map contains the surveys made north and south of the western end of the Kuruk-tagh hills and along part of the course of the Tarim and Konche-darya rivers lying south of it. The first mentioned was accomplished during the second and third expeditions and the second only during 1914-15. The traverses were adjusted to the locations determined for Korla and Altmish-bulak before the error in Lal Singh's triangulation was discovered. Thus all locations shown are too far west of their actual positions. The tables in Appendix A provide the correct positions based on Clementi's chronometrically determined value of Korla. The latitudes are given in a list with these notes. The map has three areas, the south-western corner of the Kara-shahr basin; The barren eroded hills in the north-east and to the south the wide belt of river scrub and jungle on the beds of the Inchike-darya, Tarim and Konche-darya rivers. To the west this river belt is bordered by the sands of the Taklamakan. On the east it is near the Lop desert. The Lop area divides the ancient route of the Konche-darya towards Lou-lan now the bed of the Kuruk-darya and the current (1915) course of the Konche-darya south-eastward.
(Stein placed great significance in the shift of the Konche-darya from its ancient flow further east to provide water for Lou-lan and its shift in turning south. But modern maps of the Tarim Basin now show the river deltas flowing east as before into the northern edge of the Lop sea.)
The oases at Kara-kum and Tikenlik have been created by the Chinese administration to improve communication routes along the northern and southern sides of the Tarim. (Again, modern maps now show a highway along this route.)

Notes on Sheet # 26 Vash-shahri
The map was compiled from surveys during the second and third expeditions. The mountains to the south are based on Lal Singh's triangulations of 1913 and Ram Singh's survey of 1906. The plane table survey along the Charchan-Charkhlik route was done by Stein in 1906 and again in 1913. Locations were adjusted to the position of Charkhlik (sheet 30). The southern part of the map shows the outer ridges of the K'un-lun and dry valleys. The modern village at Vash-shahri is located on the site of an ancient ruin.

Notes on Sheet # 27 Khadalik
The north-west corner of this map shows the spurs of the main K'un-lun mountains north of the Charchan river gorge to a portion of the gravel glacis near the gold pits at Khadalik. Triangulation is based on Lal Singh's survey in 1913. The estimate for the elevation of the snow line is probably too low.

Notes on Sheet # 28 Turfan
The map contains the survey route through the central and south-western part as surveyed in 1907. Much of this was again followed in 1914-15. At that time a large part of the entire Turfan depression was surveyed at the detailed scale of 1 mile to the inch. Thus much of the detail shown here is from the third expedition. The locations were adjusted for the position of Yangi-shahr near the map center with adjusted longitude of 89 degrees 6' 30" East. The latitude 42 degrees 55' 39" East. The longitude values for the routes south to Singer and Altmish-bulak are effected by the error previously described for Korla. The details of the Turfan basin were taken from Muhammad Yakub's detailed work of December 1914. The survey base line was measured at Kara-khoja. Of special interest is the great depth below sea level of part of this basin. The elevations were measured by mercurial barometer. Kara-khoja camp 242 was at -110 feet; Kara-khoja camp 242 at -140 feet; Sak-karez Deghar camp 275 at -630 feet; Camp 276 at - 980 feet.

Notes on Sheet # 29 Singer, Lou-lan
The map contains the area around ancient Lou-lan and the Kuruk-tagh mountains to the north in which Singer is in a valley. The surveys were mostly accomplished during the third expedition. But the routes in the desert from the Lou-lan site ruins to the Tarim river were followed in 1906-08 as was the route from Turfan to Singer and Korla. The detailed exploration and archeological work at Lou-lan shown in the south-eastern quarter of the map was done during both second and third expeditions. The detailed plane table results were adjusted for the position of Altmish-bulak and Korla, thus again containing the error from Lal Singh's triangulation on the K'un-lun.
The map is of an almost waterless area including the decaying Kuruk-tagh ridge west to east in the north. Singer, the only inhabited place in the region, is located close to this. The mountains are divided by wide plateaus and dry, salt encrusted basins. To the west the mountain chain reaches its highest extend. To the east of Singer there are a few saline springs. From the foot of the mountain chain a gravel plain descends southward in a waterless desert of drifting sand and clay mesas being eroded by the relentless north-east wind. This area is bordered on the west and south by the Tarim river and its marshy delta. On the east it is bordered by the salt of the Lop sea. In ancient times the Kuruk-darya flowed east to bring some water to Lou-lan. This enabled the Chinese to establish their frontier garrison there to support caravans from Tun-huang..

Notes on Sheet # 30 Lop-nor
This map lies south of # 29. It includes the terminal delta and marshes of the Tarim in the region called Lop-nor. Further south the map extends across desert ground to the northern foothills of the K'un-lun mountains. The routes in the area were used during the second and third expeditions. The peaks of the K'un-lun were mis-identfied by Lal Singh in his 1913 triangulation resulting in the locations of some points listed with longitude values too far east, in the case of Toghrak-chap as much as 1 minute and increasing to 5' 10" for Peak 1/75, the easternmost of the triangulation stations. The longitude determined for Charkhlik (88 degrees 2' 10") is the mean between Dr. Hedin's value and that of Lal Singh The longitude for Miran and Abdal were derived from plane table traverses to the Miran river. The routes leading north to Lou-lan and Tikenlik are adjusted to the other locations. The southern shore of the Lop-nor is adjusted to the location of Kum-kuduk (sheet # 32). .

Notes on Sheet # 31 Pichan, Chik-tam
The map includes areas on both slopes of the T'ien-shan and part of the western part of the Turfan basin with the desert east and south of it. The high road from Hami to Turfan was traversed in 1907. The other routes were surveyed in 1914-15. The latitude determined for Turfan is 342 degrees 30'. The traverse lines between Turfan , Hami and Barkul used latitude observations shown in the text. These values are in close agreement with those of Dr. Vaillant from 1908. The route at Donglik at its northern end intersected an old desert track from the Hami river. A northern route that M. Muhammad Yakub surveyed from the same basin to Chick-tam is no longer practical. In addition to these desert plateaus the map also shows the small portion of the Turfan basin around Pichan and Chik-tam. These oases, like the others, receive water from the Karezes which catch the subsoil drainage from the eastern T'ien-shan. (Karezes are man-made tunnels similar those formerly also in Persia prior to the Mongol conquest.) To the east of Chick-tam the crest line of the mountains falls considerably and receives even less moisture. To the west there are conifer forests on the northern slope of the range. The southern slope is far more barren. The Chinese main route along the foot is made possible only by rare springs and wells with some bits of vegetation in small basins.

Notes on Sheet # 32 Ancient Lop Lake Bed
The map shows the Lop salt sea. The route along the southern 'shore' was used in 1907. The other routes date from the third expedition. Those in the northern half were made by Lal Singh in winter of 1915. Those further south were made by Stein and Mian Afraz-gul. In the south-east corner is Kum-kuduk on the caravan route between Tun-huang and Charkhlik. its position was established by adjusting traverses from An-hsi (map 38) and Miran (map 30). The longitude was determined to be 91 degrees 55' 30" East. This is close to the number shown on map 67 of Serindia. The correction for latitude was made from observations in 1914. The traverses to Kum-kuduk from Altmish-bulak and the ruins north-east of the Lou-lan ruin were adjusted on the positions adopted for these two points as well as on Kum-kuduk. (This was the dangerous route Stein explored during the winter by riding directly south-east across the Lop salt sea.) The traverse Lal Singh made from Yetim-bulak north through unexplored parts of the Kuruk-tagh was adjusted to the adopted positions of Altmish-bulak and Deghar. Latitude observations were made during this survey.
The area in this map has two distinct regions. In the north there are the low desert ranges and plateaus of the Kuruk-tagh. In the south lies the great salt basin of the Lop sea. Both regions are empty except for seasonal movement of wild camels. There is no drinkable water outside the wells at Kum-kuduk in this entire region and no living vegetation except along the southern shore. This part of the Kuruk-tagh has never seen humans except for occasional hunters in the western parts. But across the salt sea and along the barren shore of gravel and eroded clay there was the Chinese caravan route from the 2nd century B.C. This route was finally closed some 4 centuries later by the increasing desiccation of the area. The caravan route along the southern 'shore' of the Lop sea was long out of use as well but now can be used during winter. .

Notes on Sheet # 33 Lowaza, Bash-Kurghan
This map shows the area between Tun-huang and Lop region. The two survey routes were used during both the second and third expeditions. Numerous latitude observations were taken. The survey routes were adjusted for the locations of Miran and An-hsi. But Miran is located about 2 minutes west of that shown in sheet # 30. The route shown in the northwest part of the map is part of the old caravan route along the southern shore of the Lop sea. All the springs along this route are saline and can be used only during the winter when ice forms at the top. Otherwise the caravan route follows the northern slopes of the Altun-tagh and eastern end of the K'un-lun. (See a modern map that shows the same route today.) The northern part of the Altun-tagh, which was surveyed, does not reach the snow line further south and it extremely arid. Between the Altun-tagh and the Lop sea is an even dryer area of sand dunes.

Notes on Sheet # 34 Barkul, Hami
This map shows an area north of the T'ien-shan as well as south. The surveys were accomplished in fall of 1907 and during the third expedition. One route is along the main caravan route between Hami or Kumul oasis. It was adjusted for the locations at Pichan and An-hsi. Hami surveyed latitude is 93 degrees 26' but that is east of Clementi's chronometric value of 93 degrees 18' 16". The location shown for Barkul was adjusted for Ku-ch'eng-tzu and Turfan. The accepted longitude is 92 degrees, 51' 20" East. The snow line shown at 12,000 feet is a guess. Hami has been a main station on the Chinese route west since the first and remains so today. The map shows two distinct regions divided by the eastern extension of the T'ien-shan. On the north are the plateaus and open valleys of Dzungaria where there is sufficient rainfall for grazing and agriculture in some areas. This is the historical homeland of the nomads. There is conifer forest on the northern slopes. The southern side of the range is dry. There is a gravel slope south of and below the mountains. The small oases along the foot of the gravel plain make use of sub-soil drainage. All the streams end in the basin at Shona-nor. South of the line of oases the ground continues with dry gravel and stony wastes.

Notes on Sheet # 36 Khanambal or Anambar
This map shows the outer ranges of the Altin-tagh and the descending table land to the Su-lo-Ho. It was surveyed by following the southern caravan route through the foothills between Tun-huang and the Lop area which Rai Ram Singh used in 1907 and R. B. Lal Singh used again in 1913. The locations along the route have been adjusted for the positions of Miran and Nan-hu. The observations for latitude of Khanambal (39 degrees 15' 36" North) and Su-mi-t'ou (39 degrees 49' 32" East) shown in map 75 of the Serindia series are confirmed within a minute.

Notes on Sheet # 37 Karlik-tagh
This map shows the eastern end of the T'ien-shan and the region east to the Pei-shan desert. The route west from An-hsi to Hami and some other areas were surveyed in 1907. The southern slopes of the Karlik-tagh were surveyed during the third expedition. The locations determined for Barkul, Hami, An-hsi and Su-chou on other maps were used plus latitude observations for places on this map. Several chronometrically determined longitudes made by other travelers were incorporated. The snow line at 12,000 feet is a guess because the observations were made in October when fresh snow had already fallen. The map includes a section of the ancient Chinese main route from An-hsi to Hami leading to the narrow area between the Taklamakan and the T'ien-shan. Along the eastern section of the T'ien-shan it forms a great divide between the more moist plain of Dzungaria to the north of the high perpetually snow-covered peaks and the dry Tarim basin to the south. Toward its eastern end the T'ien-shan becomes lower and lower and finally merges into the plains of Mongolia. Along this section the land even to the north becomes dryer. The southern slope of the Karlik-tagh is barren throughout. Only were there is some sub-soil drainage can irrigation of the fertile soil be attempted. The southern gravel glacis descends into a trough depression which separates it from the uplands of the Pei-shan. This map shows a northern ridge that enters the map in its south-east corner. This range is crossed by the Shuang-ch'uan-tzu pass. Further west is the caravan road between K'u-shui and Yen-tun.

Notes on Sheet # 38 Tun-huang, An-hsi
The northern part of this map contains the desert ranges and plateaus of the Pei-shan. The survey of this area was confined to the ground adjacent to the Chinese track between An-hsi and Hami which Stein followed in 1907. A few hills to the north east and south west were observed from the same route during the third expedition. The southern part of the map includes the wide valley of the lower Su-lo Ho and the foot-hills of the western end of the Nan-shan mountains. Along the south side of the Su-lo Ho Stein surveyed the Han wall and towers in both 1907 and 1914. The traverses were adjusted for the position of An-hsi at latitude 40 degrees 31' 38" East and longitude 95 degrees 57' North. This longitude differs from that shown on map 81 of the Serindia series. But the longitude for Tun-huang at 94 degrees 47 minutes East is affected by the incorrect adjustment from Lal Singh's triangulation. The maximum error along this section is 5' 10" too far east. Mr. Clementi's chronometric longitude value for An-hsi is 95 degrees 47' 20.6" East. Stein provides results from several other surveyors. He also refers to his descriptions in Desert Cathay and Serindia.
The map contains three distinctive regions. On the north the surveyed route towards Hami was first opened by the Chinese in 73 A. D. and has continued to be the main route west. It crosses a succession of decayed hill ranges of the Central Pei-shan. The low southern-most range of the Pei-shan to the west merges with an outer range of the Kuruk-tagh. Between this and the foothills of the Nan-shan to the south lies the depression of the Su-lo Ho river valley. The area west of Tun-huang is similar to the delta and basin of the Su-lo Ho seen on map # 35. The water supply from the Tang-Ho (the only tributary of the Su-lo Ho) provides the extensive irrigation for the Tun-huang oasis. The unique resources of this oasis provided the Chinese with the essential local logistic (food) support for caravans using the Lop desert route. This made the defensive Han wall important. East of Tun-huang the flat bottom of the Su-lo ho valley narrows and where the Hami road begins it is occupied by the An-hsi oasis (named Kua-chou in ancient Chinese). To the south on the map are seen the barren foothills of the Nan-shan. Its northern outer chain is immediately south of Tun-huang and covered with drift sand, hence the ancient name Sha-chou 'city of sands'. Behind the outer line of hills there is a wide plateau filled with gravel which slopes up toward the well-defined second low range in which there is another oasis at Tung-pa-t'u.

Notes on Sheet # 39 Nan-hu
This map shows the very small oasis at Nan-hu south-west of Tun-huang. The map shows the western end of the Nan-shan which my join the eastern end of the Altin-tagh. The area around Nan-hu was explored during the second expedition in 1907 as was the high plateau above Shih-pao-ch'eng. Lal Singh surveyed part of the area in 1914. Nan-hu lies at the south end of the Chinese "Yang barrier' (north-south) which defended the region south of the end of the Han wall (east-west). The deep valley of the Tang-Ho divides the area into two sections. The western side is a great gravel plane but contains a basin with fertile loess soil and irrigation from the subsoil drainage which enables some cultivation. To the west and south of Nan-hu the gravel glacis and low ridges are overrun by high sand dunes. Further south are two ranges shown on map 36. In the eastern part there are three chains of which the southern-most and highest forms the outer rim of the mountain wall behind which is the drainage-less plateau of Makhai and Tsaidam. To the east this chain joins another that continues east to the middle course of the Su-lo Ho and Central Nan-shan mountains. The southern most range sinks down with gentle slopes of gravel to the outer chain of dune covered foothills south of Tun-huang.

Notes on Sheet # 40 Yu-men-hsien
The map is directly adjacent to the east of map 38 and its area is divided in a similar way into two zones. The northern, much larger, sone was surveyed in 1914 along a single route. It contains desert ranges and plateau like valleys of the Pei-shan. To the south there is the eastern portion of the lower Su-lo Ho and the adjoining depression of Hua-hai-tzu. The former is flanked on the south by the outer ridges of the Nan-shan and the latter by the end of the hills continuing north-west to Su-chou. The southern region was surveyed from different routes during both the second and third expeditions. The plane-table work was adjusted for the routes in the southern zone on locations accepted for An-hsi and Su-chou on maps 38 and 43, and for the route in the northern zone on those of Su-chou and Barkul on sheet 34. Latitudes were observed for 6 points. The area in the south-western corner of the sheet was explored by Stein in summer of 1907 to visit the ruined sites near Ch-iao-tzu and Wan -fo-hsin as described in Desert Cathay. The historic Chinese caravan route from Su-chou past Yu-men-hsien to An-hsi and then Tun-huang appears. The Han wall was built to protect this caravan route. It crossed the Su-lo ho from south to north at the Wan g-shan-tzu ridge and from there lay along the right bank of the river as far as Shih-erh-tun near the river's sharp bend to the south. This section of the Han wall and its continuation further east to Hua-hai-tzu was surveyed in 1914.
The previously unexplored area in the Pei-shan mapped from Mao-mei (on map # 42) to the eastern end of the T'ien-shan is similar to the western area previously surveyed which is crossed by the An-hsi to Hami main road (mentioned in notes for map 38). This map shows the barren ranges oriented from east to west north of Ming-shui to Lo-t'o-ching. On their route north of camps 212 and 209 they found the highest pass elevations but without significantly higher ridges on either side. The drainage appeared to be to the east rather than to the west as found by other explorers. The chain of hills shown on the map extending north from the Su-lo Ho bend and the Hua-hai-tzu depression evidently represent the firth and southern most Pei-shan range.
The zone in the southern part of the map shows several geographically interesting features. In the west there is the head of the lower Su-lo Ho valley below the river's exit from the mountains. Further down the bed is narrowed by the defile between the Wang-shan tzu-ridge, which is the eastern extremity of the outer Nan-shan ridge, on the south and a flat spur of the southern most party of the Pei-shan range to the north. In the basis like head of the lower Su-lo-Ho valley the water shed slopes down gradually dividing it from the plateau between the two outer hill chains of the Nan-shan, which contains the small oases of T'a-shih and Ch'iao-tzu. The latter oasis receives its irrigation from springs. But above it is a ruin which indicates that irrigation was more extensive in former times.
The oasis of Yu-men-hsien derives its name from the ancient "Jade Gate" of the Han wall. This gate was originally located far to the west of Tun-huang (map 35 D-4). The cultivation here includes a level area of scrubby and in parts boggy ground dividing the Su-lo Ho valley from the Hua-hai-tzu depression to the east. This bit of topography divides the Su-lo Ho into its main branch flowing west and occasional small streams flowing east.
The Hua-hai-tzu basin is a drainageless area bordered on the north by the outer Pei-shan hills and in the south by the hill chain that goes south to Chia-yu-kuan west of Su-chou. There are several other small streams that flow into the Hua-hai-tzu basin.

Notes on Sheet # 41 Ch'ang-ma
This map shows the mountain area surveyed from the foot hills of the Western Nan-shan to the high snowy range dividing the upper Su-lo Ho valley and the head waters of the T'a-shih river from the plateaus drained by the sources of the Tang-Ho. The area was surveyed in 1907. The plane table work was adjusted to the positions accepted for An-hsi and Su-chou. These mountains contain the division between the Western and Central Nan-shan. But Stein notes that he cannot definitively point to the location of this dividing line. The western ranges up to the snow covered southern range are arid. But the Central Nan-shan has a moister climate. From the T'u-ta-fan pass eastward there is much more vegetation indicating more rain and snow. But similar signs were not seen when crossing the Su-lo Ho hear Ch'ang-ma. A that oasis, which is at 7,000 feet elevation cultivation depends on irrigation from sub soil drainage. Further east the climate is less arid and there is surface drainage supporting cultivation. This climate difference was seen also in the snow line observed in July and August of 1907. The high range south of Ch'ang-ma had snow at about 17,250 feet while further east in the ranges seen from the T'u-ta-fan pass the snow line descended much lower to 16,000 feet elevation. The Central Nan-shan is divided into parallel ridges seen on this map. But the valleys between them are not well marked. There is a gap in the survey between the mountains shown here and those on map 43.
Surveys by other explorers may indicate that the high snow covered range in the south, named the Suess Range, does continue into the high range south of Ch'ang-ma with peaks above 19,000 and 20,000 feet. The next range north is called the Emperor Alexander III range and it may be connected with the one which Stein's route from the T'a-shih river was seen with a west- east bearing. The massif of Erh-lung-shan south of Ch'ang-ma may be a link in that chain. The To-lai-shan range may be traced in the succession of high peaks this map shows as running towards the "Chu-chia-shan' just north west of Ch'an g-ma. One may consider the extension of the northern Richthofen Range to be in the points near the Su-lo Ho near Ch'ang-ma.

Notes on Sheet # 42 Chin-t'a
This map was surveyed in the spring and summer of 1914. It contains the south eastern end of the Pei-shan and the eastern part of the Hua-hai-tzu basin and part of the wide trough in which the Kan-chou and Su-chou rivers join to form the Etsin-gol. The map was adjusted for the locations of An-hsi and Su-chou. The survey through the Pei-shan was adjusted for the terminal points of Barkul and Mao-mei. For Mao-mei the latitude observation was 40 degrees 17' 49" East. Its longitude was the mean between two at Su-chou. The southern area contains a section of the Han wall traced from Hua-hai-tzu to the point it crosses the Etsin-gol below the northern end of the Mao-mei cultivation. The map shows the south eastern part of the Pei-shan, which is similar to that in the adjoining map (# 40), that is completely barren. The map contains the fifth, southern, range which was crossed by Stein's route south of the coal pits at Mou-wo. In the eastern part of the Hua-hai-tzu basin there are belts of drift sand and wind-eroded clay terraces such as appear throughout the areas further west. There is a stony plateau separating the basin from the valley containing the end of the Pei-ta-Ho above its confluence with the Kan-chou river.

Notes on Sheet # 43 Su-chou
The map was compiled mostly from extensive surveys made during the second expedition but supplemented on the north and eastern sides with more survey work from the third expedition. The surveys into the mountains from Su-chou all closed back at their starting point. The location of this significant city then was the basis point. Its latitude at 39 degrees 45' was established from astronomical observation made during both jounreys at the temple of Chiu-ch'uan outside the eastern city gate. The longitude of 98 degrees 33' is the mean between values established by Mr. Clementi (98 degrees 26' 56.3") and values from the Survey of India. This is slightly different from the longitude shown for Su-chou in map 88 of the Serindia series. The routes to Su-chou from An-hsi and Kan-chou were adjusted for those locations (maps 38 and 46). The southern part of this map contains the extensive mountain area of the four main ranges of the Central Nan-shan. The plane table work was enabled by taking positions on high passes and triangulating on numerous peaks in all directions. The plateaus between ranges is open, giving clear lines of sight. The area of the high spurs of the Richthofen Range east of Ma-yang-ho valley were enhanced by survey accomplished by Lal Singh in 1915. The approximate snow line was observed during August 1907 at 15,500 feet.
The map clearly depicts the three main regions of terrain between the Central Nan-shan and the Etsin-gol basin. The principal physical features of these regions have been described in Chapter II. The mountain region on the south shows the wide uplands at the sources of the Su-lo Ho and the Su-chou and Kan-chou rivers as well as the narrow, deep gorges by which these rivers cut their way out of the mountains. The increased moisture of the climate toward the east can be seen from the extensive forests at elevations above 8,000 to 10,000 feet in the valleys toward the Pacific water shed, marked by the Ta-t'ung river.
In this direction the change in the character of the Richthofen Range from steep mountain rampart to a system of broad spurs with easier slopes toward their top is noticeable. This change in the lateral expansion of the Richthofen Range determines the width of the second region, the plateau like belt stretching along its northern foot. Owing to the line of oases in this belt the corridor passage to the west was created. Su-chou is at the north western end of the belt in a broad alluvial fan of the Pei-ta-Ho and smaller rivers which flow out of the Richthofen Range. The rest of the belt along here is mostly scrubby steppe with enough water for grazing.
The northern edge of the second region is formed by the low, barren hills that are the western end of the Ala-shan. This range of hills with its slopes of bare gravel or sand is similar to the Etsin-gol basin described for maps 44 and 45. The cultivated area ends where the Kan-chou and Su-chou rivers break through the desert range to unite near the Mao-mei oasis in map 43.

Notes on Sheet # 44 and 45 - Etsingol, Etsin-gol delta
These two maps join together to show the course of the Etsin-gol flowing north from below Mao-mei to the terminal lake basins near the Mongolian border. Etsin-gol is the Mongolian name. The south-west end was fitted to the location for Kan-chou - map 46. and the rest of the surveys were compiled on the observed latitudes along the route north. All surveys were done in May and June of 1914 when dusty atmosphere prevented astronomical observations and distant views. The terminal course of the Etsin-gol and its drainage less basin that carries all the water from the Central Nan-shan are of great historical and well as geographical interest. The route along the river has always been important as a great natural highway from Mongolia into western Kan-su and thus into China and was used by Chingis Khan who captured Khara-khoto. The ruin of Khara-khoto adjoins the river. It was the Etzina city noted by Marco Polo.
The river area in the survey can be divided into three sections. From below Mao-mei to the outlying rocky spur of Bayin-bogdo the river is confined to a single wide bed in a trough flanked on the west by the steep gravel glacis of the Pei-shan and on the east by a slopping glacis from the Ala-shan. At the southern end of the Bayin-bogdo spur the river spreads out into a wider delta. Among the number of traceable branches few now receive water. For most of the year water can be obtained only from wells. In the narrow belts of riverine jungle flanking the beds there are wild poplars. The two terminal lakes are separated by a ridge and are at different levels, The Sogo-nor is 200 feet higher than the Gashun-nor. The map shows the shrinkage of the Sogo-nor in recent times.

Notes on Sheet # 46 Kan-chou
This map is of the furthermost eastern area visited and surveyed by Stein during both second and third expeditions. The work was completed mostly during June and August of 1914. The map locations were adjusted for the positions of Su-chou and Kan-chou. The latitude for Kan-chou was observed at 38 degrees, 55' 41" and then as 38 degrees 55' 36" North. A new longitude value was determined at 100 degrees, 38' 20" East. The map 94 in the Serindia series shows longitude as 100 degrees 49' 30" East. Stein lists different values obtained by other explorers.
The map shows all he basic topographic features in the three regions shown on map 43 to the immediate north. In the south there is the eastern part of the Central Nan-shan drained by the Kan-chou river. The broad valley of the O-po-Ho which is the main eastern tributary of the Kan-chou is included. The snow line appears to be some what higher on the eastern ends of the ranges. The narrow, difficult gorges in which the Kan-chou cuts its way through the Richthofen Range are impassable except in winter. The plateau stretching along the northern foot of the Richthofen Range gradually widens to the south-eastwards as the spurs descending from the latter recede. The large cultivated area around Kan-chou is possible due to the large flow of the river.
To the east of the longitude of Kan-chou cultivation along the foot of the Richthofen Range is independent of irrigation due to increased precipitation. The hill range to the north separates the inhabited plateauthe Kan from southern Mongolia rises to a height of over 10,000 feet. The route surveyed from Mao-mei to the Kan-chou river near Kao-t'ai provided evidence of the great aridity of the climate prevailing in the belt of low hills and wide desert valleys north of the middle course of the Kan-chou river. The route through the mountains followed in 1907 from the sources of the Kan-chou river to the city is described in Desert Cathay.

Notes on Sheet # 47 Kungurche
The map includes a narrow belt of hills visited only by M Muhammad Yakub while taking the camels to their summer grazing time. It lies along the Chinese- Mongolian border east of the Etsin-gol.

Appendix A
Discussion by Major Mason about the accuracy of the surveys and tables of latitude, longitude and elevation data.