For most of the period this course covers, more goods moved between China and the west by sea than by land, and the reason is a freight rate.
The last few lessons kept pointing here. The plague accelerated fourfold when it reached shipping. The porcelain of the technique lesson went east to west entirely by water. The freight model of the third lesson excluded from the overland route precisely the goods that dominate the maritime archaeology. This lesson gives the sea its own treatment, and the effect is to shrink the caravan route to its real proportions without diminishing what was distinctive about it.
Wind that reverses
The Indian Ocean has a property no other ocean has: its winds reverse twice a year. From roughly June to September the southwest monsoon blows steadily from Africa and Arabia towards India; from November to March the northeast monsoon blows back. A sailing vessel can therefore make an open-sea crossing in one season and return in the other, without coasting.
The Periplus of the Erythraean Sea, a merchant's handbook written in Greek around the middle of the first century CE, describes the route in operational detail: which ports to use, what to carry to each, what to buy, when to sail and how long each leg takes. It attributes the open-water crossing to a navigator named Hippalus, and whether or not one man deserves the credit, the effect is clear in the traffic. Strabo reports that in his day 120 ships a year sailed from Myos Hormos on the Egyptian Red Sea coast for India, where under the Ptolemies only a handful had gone.
Note what the monsoon does to the shape of the business. The wind is a timetable, not a convenience: miss the season and you wait months, so cargoes, credit and warehousing all had to be organised around a fixed annual cycle. That is a different commercial rhythm from the caravan trade, and it favours large operators with capital to leave idle.
What one cargo was worth
The Roman trade with India is the best quantified exchange anywhere in this subject, thanks to two very different documents.
Pliny the Elder supplies the rhetoric. In the Natural History he complains that India drains fifty million sesterces a year from the empire, and elsewhere that India, the Seres and Arabia together take a hundred million, all of it to pay for luxuries that in his view do the Roman character no good. He is a moralist writing a diatribe, and the figures have often been dismissed as invention.
Then there is a contract. A papyrus in Vienna, known as the Muziris papyrus, records the terms of a loan for a cargo shipped from Muziris on the Malabar coast to Egypt on a vessel called the Hermapollon, in the second century CE. It values the cargo at 1,154 talents and 2,852 drachmas, and it specifies the Roman quarter-tax, the tetarte, on arrival.
Example. Take the drachma as equal to one sesterce in this Egyptian context, and legionary pay in the second century as 1,200 sesterces a year. Value the cargo and the tax on it, then test Pliny's hundred million against it.
The cargo is drachmas, about 6.93 million sesterces. That is soldier-years of pay in one ship's hold. The quarter-tax alone is 1.73 million sesterces, or 1,440 soldier-years, from a single arrival, which tells you why the Roman state was interested in the Red Sea. Now Pliny: his hundred million is cargoes of this size. Against Strabo's 120 ships a year, that is one ship in eight carrying a Hermapollon-sized cargo, which is entirely plausible and possibly conservative. Pliny's figure, long treated as a rhetorical exaggeration, turns out to be the right order of magnitude, and against a total imperial revenue usually estimated at something like 800 million sesterces it represents about an eighth, which is large, striking and not absurd.
Now you. What is the strongest objection to using one papyrus this way, and how much does it damage the conclusion?
Answer
That it is one document and probably not a typical one. A contract elaborate enough to be written, notarised and preserved is likely to concern an unusually large consignment, since small shipments were handled informally, so the Hermapollon may sit at the top of the distribution rather than in the middle of it. If the average ship carried a tenth of that value, Pliny's hundred million would require 144 cargoes, more than Strabo's entire annual fleet, and the figure would look inflated after all. The conclusion is therefore sensitive to an unknown, and the honest statement is a range: Pliny's number is achievable if a substantial minority of ships carried cargoes of this order, and not otherwise. What survives the objection is the more robust point, which does not depend on the average at all: single cargoes worth thousands of soldier-years existed, they were taxed at twenty five per cent, and the Roman state's Red Sea customs were a serious revenue stream. Even the sceptical reading leaves maritime trade far larger than anything the overland route can be shown to have carried.
The freight rate that explains everything
The mechanism behind all of this is one ratio. Working from Diocletian's Price Edict of 301 CE, A. H. M. Jones and Richard Duncan-Jones estimated the relative cost of moving goods by sea, by inland waterway and by road at roughly 1 to 5 to 40.
Take that seriously for a moment. Forty to one means that carrying a cargo 1,000 km by sea costs about what carrying it km by road costs. Distance almost stops mattering on water and dominates everything on land. Every consequence in this course follows from that single ratio: which goods can move, how far, in what quantity, and why the same commodity is a luxury at one end of a continent and a commonplace along a coast.
Sixty thousand bowls
The eastern half of the maritime route was invisible until divers found it. In 1998 fishermen located a wreck off Belitung Island in Indonesia: a sewn-plank vessel of Arabian or Indian construction, sunk in the 820s or 830s, carrying a cargo of Chinese goods. Among them were roughly 60,000 ceramic pieces, mostly bowls from the Changsha kilns in Hunan, packed in storage jars, along with fine gold and silver work. One bowl carries an inscription with a date corresponding to 826.
Example. Take a Changsha bowl at 0.4 kg and use the overland freight figure of 1.73 labour-days per kilogram derived earlier. What would it have cost to send this cargo by caravan?
One bowl weighs 0.4 kg, so its overland freight is labour-days, most of a labourer's day's wage to move one cheap bowl. The whole cargo is tonnes, which is 160 camel-loads, and the freight bill is labour-days, about 138 working years. A Changsha bowl was a mass-produced object worth a small fraction of a day's wage. Its overland transport would have cost several times its value, which is not a difficulty to be overcome by clever merchants; it is a prohibition. So this cargo could never have existed on the land route, in any century, and every generalisation about Silk Road trade built from caravan evidence is a generalisation about the small, dense, expensive minority of what actually moved between these regions.
Now you. The Belitung ship was built in the Persian Gulf or India and sank in Indonesian waters loaded with Chinese goods. What does the combination tell you about who was running this trade?
Answer
That it was run from the western end, by Persian and Arab shipping sailing all the way to China, rather than by Chinese vessels sailing west. The hull is the evidence: sewn planks, stitched with coir rather than nailed, is a western Indian Ocean tradition and not a Chinese one, and Chinese shipbuilding of the period used bulkheads and iron nails. That fits the textual record, which describes a substantial foreign merchant community at Guangzhou by the ninth century, large enough that its massacre during a rebellion in 878 is reported by Arab writers as a commercial catastrophe. It also fits the cargo, which is mass-market Chinese export ware made to suit foreign taste and packed for a long voyage, meaning a buyer at the far end had specified what he wanted. The picture is a genuinely intercontinental shipping business operated by Muslim merchants from the Gulf, and it is worth setting beside the ninth lesson's conclusion about Islamic commercial law, which is what made a partnership between Basra and Guangzhou enforceable.
Example. A merchant sailing from Egypt to India must leave on the southwest monsoon between roughly June and September and can only return on the northeast monsoon between November and March. Work out the round trip and say what it does to the business.
Leaving in July and arriving perhaps two months later puts him in India in September, and the earliest useful return wind is November, so a fast round trip is about a year and a slow one is two. Compare the caravan, which takes about 130 days each way and can in principle set off again whenever it is loaded. The consequences are commercial rather than nautical. Capital is tied up for a year at a time, so the trade needs financiers rather than traders with their own purse, which is exactly what the Muziris papyrus documents: a loan against a cargo, with terms. The waiting months must be spent somewhere, which produces permanent expatriate merchant communities at the far end rather than passing visitors, and those communities appear in the record at Indian ports, at Guangzhou and later everywhere in the Indian Ocean. And the annual rhythm makes information old: a merchant chooses his cargo on prices a year out of date, which raises risk and rewards anyone with a resident agent. The monsoon, in other words, produced the institutions of the trade as directly as the camel produced the caravanserai.
Now you. The Roman trade with India ran a persistent deficit settled in gold and silver coin, and Pliny complains about it. Was he right to?
Answer
He was right about the flow and confused about what it meant, which is the usual shape of a bullionist complaint. The flow was real: India wanted little that Rome made, Rome wanted pepper, gems, ivory and textiles, and the balance went in coin, which is why Roman gold and silver turns up in south Indian hoards and Indian goods do not turn up in Roman ones in comparable value. Where Pliny goes wrong is in treating the coin as loss. Romans who bought pepper got pepper, which they wanted more than the coin, and the state took twenty five per cent of the value of every cargo at the customs house, so the trade was a substantial net contributor to the treasury even as the private balance ran outwards. There is a serious version of his worry, which is that an empire exporting its coinage exports its money supply, and that is a real monetary problem rather than a moral one. But his actual argument is moral: he objects to luxury, to women's ornament and to the corruption of Roman character, and the economics is recruited to support the sermon. It is a good example of why a number in a rhetorical passage can be simultaneously accurate and used for an argument it does not support.
The scale later on
The maritime trade grew steadily and the archaeology grows with it. The Nanhai One wreck, a Southern Song merchantman raised whole off the Guangdong coast in 2007, has yielded well over a hundred thousand items. Song and Yuan China ran maritime trade superintendencies at Guangzhou, Quanzhou and other ports specifically to tax this traffic, and Quanzhou, the Zayton of European writers, was described by both Marco Polo and Ibn Battuta as one of the greatest ports on earth.
The most spectacular episode was also the least commercial. Between 1405 and 1433 the Ming sent seven fleets into the Indian Ocean under the admiral Zheng He, reaching Hormuz, Aden and the East African coast, with the first expedition reported at 317 ships and about 27,800 men. The Ming Shi gives the largest vessels as 44 by 18 zhang, which at a Ming zhang of about 3.11 m is 137 m long and 56 m in the beam.
Those dimensions deserve the treatment this course has given other chronicle numbers. A wooden hull of 137 m is far longer than anything Europe built before iron: HMS Victory, launched in 1765, is 69 m. And the proportions are wrong, since a length to beam ratio of describes a floating rectangle rather than a ship. Naval architects have generally concluded that the figures are exaggerated or refer to some other measure, while accepting that the treasure ships were genuinely very large by contemporary standards. The voyages themselves are not in doubt; a stele the fleet erected records them.
What happened next is the more important fact. The voyages were state enterprises for prestige, diplomacy and tribute rather than trade, they were enormously expensive, they were opposed by officials who thought the money belonged on the northern frontier, and after 1433 they simply stopped. Private Chinese maritime trade continued and often thrived, legally or otherwise, but the state withdrew from the sea at the moment Portuguese ships were beginning to work down the African coast.
Why the sea is under-taught
If the maritime route carried more, why is the overland one famous? Three reasons, and none of them is evidential.
The name is one. Richthofen coined a term for a land route, and the phrase "maritime Silk Road", now used freely and adopted into Chinese policy language, is a back-formation from it.
The sources are another. Overland travel produced travellers' tales, and shipping produced bills of lading and then sank. Until scuba diving made wrecks accessible from the 1960s, and marine archaeology matured after that, the physical record of the sea route was inaccessible while the desert record had been lying in museum drawers since 1907.
And the third is romance, which is not nothing: camels, deserts, passes and lost cities are simply better material than a monsoon timetable.
The corrective is not to swap one route for the other. It is to hold both, and to say what each did. The sea carried the volume: ceramics, spices, timber, cotton, bulk goods, and the great majority of the value. The land carried what could pay its freight, and it also carried the things that had no weight, monks, scripts, faiths, techniques and, once, a bacterium.
That leaves the ending, and the ending most people have been taught is wrong. The next lesson takes apart the story that the Ottomans closed the road in 1453, and replaces it with a slower and better evidenced account in which the overland trade outlived its own supposed death by four hundred years.