In 1905 Gottfried Merzbacher, a geologist, published an account of two years exploring the Tian-Shan. The Central Tian-Shan Mountains, 1902-1903 (John Murray, London) is an engaging read, partly because of Merzbacher’s insights about the landscape, and descriptions of the people he meets, but also because it is an account of his efforts to solve the puzzle of where the prominent mountain of Khan Tengri actually was. This 7000m peak was readily visible from lowlands outside the Tian-Shan, but previous mappers had merely been able to theorize about how one might get to its base.
Merzbacher’s book originally came with a map, but, as is so commonly the case, the map was not properly scanned for the digital copies of the book that one can download today from archive.org. You get a fragment like this:
It’s a shame. Hopefully someone who owns the book will post a good scan of the complete map one day.
What fascinates me about these exploration accounts is the question How would we follow his route today? To that end I’ve begun mapping Merzbacher’s travels as described in The Central Tian-Shan Mountains. So far I’ve mapped Chapter 1.
The key decision in this kind of mapping is what base map to use. Use a period base map (something from around 1900) and you may not be able to follow it today: place names change, roads move, and mountains get mapped more precisely. (See my previous post on the yet-moving location of Khan Tengri.) Use a modern map and you lose the sense of travel in an era of inaccurate mapping.
I tend more to the modern map. To me, the question Where did he go? means Where was his route relative to modern landmarks? But many modern base maps are unsuitable. Most online mapping services do not label rivers or mountain ranges. A modern road map (like Freytag+Berndt’s Central Asia) is too small-scale (1:1,750,000) to see what’s going on. So in this case I settled for a 1980s aeronautical chart (Tactical Pilotage Chart F-6C) which gives shaded relief at a reasonable scale (1:500,000) plus the added benefit of named rivers and mountain ranges. It ain’t that pretty though.
I’ve been reading Gottfried Merzbacher’s Central Tian Shan which was published in 1905. (Archive.org has a free copy here.) This German geologist spent 1902-3 in the Tian Shan exploring valleys, noting strike and dip, and, among other things, trying to figure out the actual location of the peak Khan Tengri. This 7000m peak was not marked correctly on the 40-verst Russian map he was using. (40-verst maps were a series at 40 versts to the inch, or 1:1,680,000.)
While reading Merzbacher’s account, I looked at my map of Charles Howard-Bury’s journey in the Tian Shan (posted previously) and discovered I too had Khan Tengri in the wrong place!
So I did a little homework to make sure I had it moved it to the right place, because not being sure where Khan Tengri is located should be the kind of problem they only had 100 years ago, right? Well, maybe not. Let’s compare Google maps, on the left, and Bing maps on the right.
Wow, that’s some pretty serious difference. Not only is Khan Tengri in a different place (relative to the glaciers, for example), but the Chinese-Kyrgyz border is also in a different place. In Bing’s world the border runs atop the highest peaks, while in Google’s the upper portion of the South Inyl’chek glacier is actually in China.
Bing, by the way, is using Russian peak names, like Pobedy Peak, while Google has its Kyrgyz name, Jengish Chokisu.
Turns out that OpenStreetMap has the same border Google does; but Yahoo and Here have the border that Bing does!
A quick web search reveals that China and Kyrgyzstan settled their border disputes in 2009. It’s likely that one of the sets of lines we see above dates from prior to that settlement, and one from after. But which is which?
Administrative boundaries downloaded from GADM and DIVA indicate that Bing/Yahoo/Here are actually more up to date.
But as far as the location of Khan Tengri is concerned, I’m afraid Google/OSM are correct.
In chapter 3 of Charles Howard-Bury’s Mountains of Heaven, written in 1913 but published in 1990, the author details a ten-day journey on Russian post roads. (I discussed a map I made of the central part of his journey in this previous post.)
He arrives in the town of Semipalatinsk by steamer (coming up the Irtish river from Omsk) on June 7th,1913, and then he and his servant John Periera prepare for this overland journey by buying a tarantass, a type of 4-wheeled cart. Although in the post road system one would typically receive a new horse and tarantass at each post station (about 20 to 30 versts apart — a verst being 1067 metres or slightly longer than a kilometre), Howard-Bury has strategized that by providing his own tarantass he’ll avoid repacking the heaviest boxes at each change of horses.
The list of luggage is revealing. A valise and bedding, camp bed, bath, table and chair, baking powder and jam in tins, photographic films, a few varieties of soups, some condensed milk and “one or two delicacies such as sardines and potted meats;” two rifles, a shotgun, and ammunition; tea, bought in Moscow; and “a plentiful supply of literature.”
They depart Semipalatinsk on June 9th in the late afternoon. The first three days of post-road travel takes them to the town of Sergiopol, covering on those days, by his reckoning, 47, 101 and 124 versts respectively (53, 108 and 132 km). Mountains of Heaven does not give us a detailed map, but it’s not hard to determine the route to Sergiopol, which is an old name for the town of Ayagoz, Kazakhstan. By following the most direct road between the two towns on Soviet maps made in the 1980s, I can get just these distances that Howard-Bury reports.
After Sergiopol, things get a bit trickier. Howard-Bury and Pereira stop sleeping at post stations and the author takes to giving distances from one midnight to the following midnight. At the end of day 6, they are one stage beyond the town of “Kapal,” which it’s easy to be confident is the still-extant town of Kapal, Kazakhstan. But by which route did they get from Sergiopol to Kapal?
Howard-Bury mentions all sorts of landmarks over these three days, which we can try to match up with the terrain. He gives the travel distances as 140, 153, and 105 versts, respectively (150, 163, 112 km). The most crucial landmark is a town named “Lepzinsk,” which they pass through towards the end of Day 5. If we can locate this Lepzinsk, we’ll be close to drawing the right post-road connection from Sergiopol to Kapal.
Here are the clues which Howard-Bury gives us.
Day 4 (140 versts, starting at Sergiopol):
For some distance we followed the valley of the Aksu, crossing and recrossing the river…
Towards evening we got into flatter country again, where the plains were covered with wormwood as far as the eye could see.
Day 5 (153 versts):
We continued travelling all night but at eight o’clock in the morning had only succeeded in covering 32 versts since midnight, owing to the sandy nature of the soil…
We had to cross several arms of the desert which stretched across the road from the west in broad rivers of sand several miles wide…
We now had to cross a range of rocky hills…
We are now not far from the shores of Lake Balkhash… there is good fishing to be got at this end of the lake where the Sergiopol River flows into it.
Low rocky ridges alternate with broad grey and yellow plains…
Further on the plain became at times white as snow from the alkali in it that had come to the surface…
Towards evening we got our first glimpse of lofty snowy mountains to the south…
The going was very sandy and heavy. We kept crawling slowly up and down over big sand dunes, when all of a sudden in the midst of the sand, we came across a fair sized river on the far side of which, across a wooden bridge, stood the town of Lepzinsk, set in a bower of green. What a delightful contrast!… There were pleasant grassy meadows all around, well irrigated and with many a wild flower growing on them. The contrast to the barren country through which we had been passing made Lepzinsk look doubly beautiful. …From now on to the mountains, the rainfall is greater and the climate milder, so that vegetation becomes everywhere more luxuriant.
Day 6 (105 versts):
It took us ten hours to do 47 versts, as we had to cross a range of rocky hills that formed a kind of buttress to the snowy Ala-tau mountains. The road… followed the bottom of a narrow and stony gorge…
On arrival at the top of the pass, which was only about 4,000 feet in height, there was a glorious view of the snowy mountains, only separated from us by a narrow grassy valley.
From now onward, the road kept at the height of over 4,000 feet and we drove across beautiful grassy meadows…
Soon after mid-day we arrived at Kapal… Except for a few trees at Sergiopol and Lepzinsk, this is the first place in the 650 versts that we have traversed since leaving Semipalatinsk, where they are at all plentiful.
So there are some wonderful clues here: the arms of a sandy desert coming in from the west, the proximity of Lake Balkhash, verdant Lepzinsk in the midst of sandy dunes and first seen on the far side of a “fair sized river,” a pass at 4,000 feet in a range of rocky hills “that formed a kind of buttress to the snowy Ala-tau mountains,” and a plateau over 4,000 feet that leads on to Kapal.
To make initial sense of all this, let’s take a look at this map. This is a portion of a Soviet topographic map at 1:1 million produced in 1973.
As you can see, Kapal lies near the large northeast-trending mountain range at the bottom of the map, the Dzhungarskiy Ala-tau, which Howard-Bury calls the Ala-tau mountains.
Starting from the end and working backwards, Let’s look at the area around Kapal for a traveller arriving from the north.
This topography seems to correspond well with what Howard-Bury describes on Day 6. There’s a line of rocky hills forming “a kind of buttress to the snowy Ala-atu mountains,” and there a pass through them marked, on this Soviet map, with an elevation of 1052 metres, or 3500 feet. From the pass to Kapal the route continues to gently ascend.
The town at the foot of that 1052 m pass was marked as Джансугуров/Dzhansugurov in Russian. Now it is Жансугиров/Zhansugirov, Kazakhstan. We have reduced our task to determining the route from Sergiopol to Zhansugirov.
Let’s back up for a moment and look at rivers in the area.
There are three main rivers that feed Lake Balkash in this area. From north to south today they are called the Ayaguz (Аягуз), Lepsi (Лепсы) and Aksu (Аксу). These are three of the famous Seven Rivers that give the region its name in Kazakh (Jetyysu) and Russian (Semirechinsk).
Howard-Bury mentions following the Aksu as he leaves Sergiopol, but confusingly the Aksu is the southernmost of the three rivers we see here. It’s possible that after staying up all night for so many days became confused about which river he saw on which day. But it’s also possible that someone told him the Ayaguz was called “Aksu,” because Aksu is a very common name for rivers in Central Asia (meaning, essentially, white water or clear water).
He also refers to a Sergiopol river, which, from his reference to it entering Lake Balkhash we can assume is the Ayaguz.
We might expect a town called Lepzinsk to be on the Lepsi river, since “Lepzinsk” is essentially a Russian construction meaning “of or pertaining to Lepsi.” Surprisingly, there are three candidates for this town, and, correspondingly, three main route choices.
This town of Lepsy is absent on historical maps, but it exists today and can be found, for example, on Google maps. We’ll call this one “Low Lepzinsk” because it is lowest on the Lepsi river. Its satellite image shows that one would indeed come across this river “all of a sudden in the midst of the sand.”
This route is initially appealing because it mimics the railway corridor shown on the 1973 map. It passes close to Lake Balkhash. But, there are a few problems. The town of Lepsy is on the north side of the river, and Howard-Bury says that he sees the river “on the far side of which, across a wooden bridge, stood the town of Lepzinsk” As well, prior to getting close to Lake Balkhash, arms of desert crossing this route would come from the east.
This Lepzinsk still exists, and we can see it shown as Lepsinsk/Лепсинск on the 1973 Soviet topo. It’s much higher on the Lepsi river, near its headwaters, so we’ll call it “High Lepzinsk.”
There are a number of problems with this route as well. Overall the eastern route is too long to match the distances Howard-Bury gives. It goes nowhere near Lake Balkhash. A satellite picture indicates that rather than travelling “up and down over big sand dunes” before coming upon the Lepsi river in this area, one is in a well-vegetated region.
This town of Lepzinsk, which we’ll call Mid Lepzinsk because it is mid-way along the Lepsi river, is shown on a 1912 map of Central Asia, by Alexander Keith Johnson:
Notice he actually has two Lepzinsks. There is Mid Lepzinsk, labelled ‘Lepsinski,” and also High Lepzinsk, here labelled ‘Lepsinskaia (Lepsa).” Also, importantly, he shows the post road from Sergiopol to Kopal running through Mid Lepzinsk.
Other period maps corroborate. Here’s Edward Stanford, 1901:
Again we have the two Lepsinsks, and the same post road passing through Mid Lepzinsk.
Here’s Vivien de Saint-Martin & Schrader, Atlas Universel de Geographie, published in 1935:
They too show both Lepsinks, although High Lepzinsk is bigger, and Mid Lepzinsk is ‘Lepsinskoïe.” At this point the railway has been built to the west and the older post road is not shown
It’s also probably worth looking at the map that comes with Mountains of Heaven.
Lepzinsk is there, but it is hard to say from this map whether Mid or High Lepzinsk is intended. Looking at other maps we can see that a bearing from Kapal to Mid Lepzinsk is more north than east, whereas one to High Lepzinsk is more east than north. Without longitude lines on this map we’re making a guess, but the Lepzinsk shown seems to be north-northeast of Kopal, and therefore would be Mid Lepzinsk.
What’s confusing is that the only river shown is the Aksu, which we know to be south of the Lepsi, and in fact the author’s route is shown as going to Lake Balkhash and then up the Aksu to Lepsinsk. This seems an unlikely route choice, as if one actually went up the Aksu he would have to turn north to get to a town on the Lepsi.
An argument against Mid Lepzinsk is that it no longer exists by that name. At this site today there is a town and a crossing of the Lepsi (and it is suitably situated with sandy areas north of the river, as the satellite image below shows) but it is called Kokterek (or “Blue Tree”), Kazakhstan. I can’t find any evidence it was ever named Lepzinsk.
In favour of Mid Lepzinsk and the Central Route option is that on this route it is easy to make distances agree with the daily distances Howard-Bury gives.
So I vote that Howard-Bury’s Lepzinsk was the town today known as Kokterek, Kazakhstan, and that he took what I have shown above as the central route from Sergiopol to Kapal.
I’ve just finished a map for a book that had no map: Mountains of Heaven: Travels in the Tian Shan, 1913. This version of the Charles Howard-Bury’s journals from his travels in what are today eastern Kazakhstan and the adjacent region of China, was published in 1990 and although it’s out of print, it’s generally available in the used book market.
Here’s the map.
What I set out to do here was to resolve all the geographic references in the text. That is, you’re reading along, and he mentions the Yulduz plains or the village of Karachekinskaia, and you don’t know where that is. The various maps, atlases and website you can check don’t have it. So the book needs a supporting map. Pretty simple.
But it evolved into something else. Initially I planned a 24″x24″ sheet of paper, a size typical of folded maps that used to be glued inside the back covers of books. It would contain three maps. There would be a 1:1 million map detailing the central portion of his journey, the area where he spends the most time and mentions the most local details. There would be a 1:3 million map showing how he got from Semipalatinsk to Jarkent, the ten-day journey along Russian post roads for whose specific route he gives intriguingly few clues. Last, there would be a map at 1:7 million showing his exit route from Central Asia, from Jarkent to Tashkent, and then on by rail to the Caspian Sea.
Then things happened which illustrate some general hazards about mapping for old books.
The first map became the main project, and eventually pushed both other maps off the sheet. It’s what you see above. There were simply too many details to squeeze it in at a smaller scale onto part of the sheet. I let go of the idea that all the mapping one would need could fit into 24″x24″. I began building a hillshade of the area in Blender, and figured out an hypsographic colouring scheme inspired by an old Soviet topographic map of the area (К-44-Б): low elevations are white, and then rise through oranges and brown to white summits.
At the same time, as I tried to figure out just where Howard-Bury had gone, I found myself buried in old maps, wikipedia articles, and all sorts of other documents such as you get when you search on obscure terms like “Kuitun River.” I was reading maps in Russian and putting German websites through Google Translate. With each Aha! answer to a question, I felt more convinced that these puzzles — for the map had come to represent a series of puzzles — would make good stories.
At this point the map became a poster, the bottom half of which contained explanations of stuff. Here’s the poster:
The sections in the bottom half cover the reason for the map, the plethora of alternate place names in Central Asia, a glossary of obscure terms Howard-Bury uses, an explanation of the Soviet topographic mapping system, some Turkic geographical equivalents, and a list of references. It is, in short, a brief tour of all the places I had to go to produce the map.
I’ll do another post here to present an example of one of the interesting puzzles: where was Lepzinsk?