Method

How we know

Most dumpling maps tell you where a dish is from. This one also tells you how sure anyone can be. Every entry and every date carries an evidence grade, and legends are kept apart from evidence.

Where things stand

465dumplings in the database
122graded above G
131researched and on the map
62with checked sources listed
26with a recipe
343unverified leads, not published

This is a working prototype. Map positions, eras and recipes are drafts that still need review, and the illustrations are not yet commissioned: the stamps and photographs you see are stand-ins.

Evidence grades

A Archaeological or primary source: An archaeological find or a primary dated document has been identified. The Astana dumplings, the Hocheppan fresco, Czerniecki's 1682 cookbook.
D Dated text: A dated text, reported by scholars or a high-quality secondary source. The Yinshan Zhengyao of 1330; EU protected-name regulations.
S Secondary consensus: Secondary or journalistic consensus, with no primary text cited yet. Enough to publish, flagged as such.
L Legend: A legend is attached. Legends are told as stories and never set a date or an origin.
G Unverified: General knowledge only. Never published until someone finds a source.

A dumpling's headline badge is its strongest non-legend grade. Where a legend is attached, an L sits beside it, never instead of it.

What counts as a dumpling

A discrete, hand- or tool-formed portion of starchy dough, batter or mash, cooked mainly by moist heat (boiled, steamed, simmered), or by moist heat and then fried. It may be filled or plain.

Some foods sit on the edge: fried and baked pastries, leaf-wrapped grain parcels such as tamales and zongzi, bulgur shells such as kibbeh, steamed buns, swallows such as fufu, and cloth-boiled puddings. They are included, flagged as borderline, with the reason.

Eras

A dumpling's era is the era of its earliest dated, non-legend evidence. Undated dumplings stay off the era filter until someone finds a date.

01Ancientbefore 220 CE
02Silk Road & Tang220–907
03Islamic & Mongol world900–1400
04Renaissance & print1400–1700
05Empires & migration1700–1900
06Modern & global1900–now

Families

Dumplings were invented many times. The atlas groups them into families, and some of them are separate inventions rather than branches of one tree. Browse the families.

A · East Asian wheat-wrapper: Chinese wheat-dough foods (bing) splitting into mantou/baozi, jiaozi, hundun/wonton and shaomai, anchored by Tang dumplings and wontons from the Astana tombs at Turpan; spread across East and Southeast Asia and later with the Chinese and Japanese diasporas.
B · Steppe / Silk Road manti: The MANT- cluster (Uyghur manta, Central Asian manti, Anatolian mantı, Afghan mantu) carried by Turkic and Mongol peoples, with small boiled (joshpara/chuchvara), Mongolian (buuz, bansh), Himalayan (momo) and Caucasian (khinkali, contested) lines. Second pass adds the tutmaj cluster (Kashgari 1072 → Yinshan zhengyao 1330 → Ottoman tutmaç) and the khinkal cluster.
C · Slavic & Baltic boiled pocket: Boiled filled half-moons and rings of Eastern Europe (pierogi, varenyky, pirohy, kołduny, koldūnai, uszka), with kreplach and North American pierogi as offshoots; Ural pelmeni sit at the B/C junction.
D · Mediterranean filled pasta: Filled fresh pasta descending from 14th-century ravioli (Datini papers, Libro per cuoco, Forme of Cury) into Emilian rings, Piedmontese agnolotti, Alpine and Swabian forms, and Italian-diaspora inventions.
E · Central European unfilled: An independent European origin of unfilled lumps in three branches - bread (Knödel, Hocheppan fresco c. 1180), batter (Spätzle 1725, nokedli, halušky) and potato (18th–19th c. Klöße, gnocchi, cepelinai, kroppkakor).
F · British suet/flour: Suet and flour dumplings and cloth puddings of Britain, carried by settlers to the Caribbean, North America, Australia, New Zealand and the Pacific; Dutch kluitjies at the Cape are grouped here by the atlas.
G · Rice dough (several origins): Rice- and glutinous-rice-dough dumplings with several independent origins - East/Southeast Asian (tangyuan, mochi, songpyeon, klepon) and South Asian (modak, kozhukattai, pitha, yomari). Not one tree.
H · Leaf-wrapped grain parcels (separate inventions): Grain or starch pastes cooked in leaves or husks, invented separately in Mesoamerica (tamal), East/Southeast Asia (zongzi) and West/Central Africa (kenkey, kwanga, moi moi). Second pass adds independent Austronesian parcels (cinavu, abai) and a fermented-cassava sub-cluster; we also park Pacific leaf-wrapped tuber puddings here at low confidence.
I · Shell: Bulgur, rice or semolina shells around a filling - Levantine kibbeh, Iraqi/Kurdish kubba soups (with the boiled kutilk/kutleh sub-cluster), Persian-Jewish gondi, Brazilian quibe; coxinha is a borderline relative.
J · Fried/baked pastry relatives (borderline): Filled pastries that fail the moist-heat criterion - samsa/samosa/sambusa, empanada/pastel, börek, knish, pirozhki, curry puff, gujiya. Shown for context, badged borderline.
independent · Independent / unaffiliated: Our addition - dishes the atlas does not place in A–J that look like local, convergent inventions (e.g. Haudenosaunee boiled cornbread, Ladakhi pinched-dough soups, Breton farz, Arabian dough-in-stew dishes). No ancestry is implied; resemblance to a family is recorded as `convergent` only.

Contested origins

When sources disagree, the atlas shows every serious claim and says how strong each one is. Some examples:

Mantou, manti, mandu, momo: One word family across Asia. Whether Chinese wrote down a foreign word or lent its own is disputed; both readings are shown.
Pelmeni: Udmurt or Komi-Permyak? Scholars disagree on which Ural language gave Russian the word.
Khinkali: Shepherd food of the Georgian highlands. The theory that it arrived with the Mongols is contested.
Kutilk and kutleh: Kurdish and Assyrian names for the same boiled bulgur shell. Both are recorded; no single origin is assigned.
Cinavu and abai: Taiwan Indigenous leaf-wrapped millet. Sources say they are independent traditions, not a kind of zongzi.

Sources and images

The research behind the atlas is open: the source reportsand the structured data are in the project repository. Recipes are written for the atlas in our own words from several cross-checked sources, which each recipe lists. Photos are openly licensed placeholders from Wikimedia Commons, credited on each page, until the atlas's own illustrations replace them. The base map is Natural Earth in the Equal Earth projection.