Rare Japanese Last Names
Rare Japanese last names sit at the edge of a naming system built on land, work, and lineage. Every surname listed below is held by only a few hundred people in Japan, and several are recorded with about ten. This page lists uncommon Japanese surnames with their kanji, readings, and element meanings, shows where they cluster, and explains how to check whether a name is genuinely rare or only unusual outside Japan.
What Counts as a Rare Japanese Last Name
A Japanese surname is rare when very few households carry it. That is a different question from whether it looks unusual in English: Satō, Suzuki, and Takahashi are common in Japan and distinctive abroad, while a name such as Watanuki is uncommon in both places. Rarity also has two scales. A surname can be rare nationally yet ordinary in one prefecture, because many family names never travelled far from the place that produced them. And there is no official count of how many surnames Japan has: published estimates run from roughly 100,000 to more than 300,000, depending on whether variant spellings and readings are counted separately.
This page stays on that question: which surnames are uncommon, how they came to be uncommon, and how to verify one. The neighbouring pages cover the rest. The A–Z surname browser lists last names by first letter, the common last names page ranks the names held by the most households, the last names and meanings page filters surnames by the images inside their kanji, and the popularity ranking puts the well-known names in order. If you arrived looking for a name for a character or a business instead of a family, read the section on creative use before you borrow one.
Rare Japanese Last Names and Meanings
Most rare surnames are not exotic inventions. They are ordinary place-based, occupational, or lineage names that simply never spread beyond a region. Reading the kanji usually shows which of those three families a name belongs to, because the elements describe where a family lived, what it did, or which house it served. House names became permanent from the fourteenth century onward, when estates began passing to a single heir; in 1870 the government allowed commoners to use surnames and in 1875 required it. The restriction before that applied to official use, not to having a name, which is why so many families already had one to register.
Kanji Origins: Why Rare Surnames Read the Way They Do
The largest group is topographical, and it is the easiest to read. Two kanji describe a piece of landscape — 山 (mountain), 川 (river), 田 (rice field), 島 (island), 松 (pine) — and the pairing points to the kind of ground the family held. 松部 (Matsube) sets a pine beside 部, a section or group; 起島 (Okishima) adds the unexpected 起 to “island”; 上県 (Uegata) turns an administrative word into a surname.
A second group is occupational: 樵 (Kikori) simply means “woodcutter”, and 師田 (Morota) pairs a title (師, master) with 田, farmland. A third group carries lineage or service markers, above all 藤 (fuji), the element that spread from the Fujiwara line into thousands of surnames. Court families kept longer, more literary names such as 綾小路 (Ayanokōji) and 甘露寺 (Kanroji), which is part of why they still read as formal today.
A fourth group is written, not spoken: ateji and gikun surnames. 鴨脚 (Ichiyō) uses the kanji for “duck’s foot” and is read “ginkgo”, because the leaf was compared to a duck’s foot. 四月一日 (Watanuki) and 小鳥遊 (Takanashi) work the same way: the characters paint a picture instead of stating a sound, and no rule lets a reader guess them. Names of this kind are rare partly because they are easy to misread and easy to drop — but hard to read is not the same as rare. 五十嵐 (Igarashi) is read against its characters and is one of the most common surnames in the country, and reference works catalogue roughly 13,500 to 14,000 hard-to-read surnames in total.
Rare Japanese Last Names Attested in Surname Dictionaries
These are the names Japanese surname dictionaries record, and the ones readers most often meet in rare-surname lists: month-and-day names, rebus readings, and names built from kanji so unusual that the reading has to be learned, not guessed. Where a dictionary attests the reading but publishes no household count, none is claimed here.
| Romaji | Kanji | Reading | Elements and meaning | Where it is found and why it is rare |
|---|---|---|---|---|
| Watanuki | 四月一日 | わたぬき | The first day of the fourth month in the old calendar, when the padding was pulled out of winter kimono — 綿抜き (wata-nuki), which became the reading. | One of only two month-and-day surnames still in use, and read Watanuki regardless of the characters. The 四月朔日 spelling is reported as the common one in Hokkaidō and Toyama, so this spelling is itself scarce. |
| Hozumi | 八月一日 | ほずみ | The first day of the eighth month, when the earliest rice ears ripened and were picked and offered; the name is linked to 穂積 (Hozumi), “stacked grain ears”. | The second surviving month-and-day surname. Dictionaries record four readings (Hozumi, Hozumi, Hassaku, Yabumi) and a variant spelling 八月朔日; it is scattered nationwide, so it is rare by count, not by geography. |
| Yamanashi | 月見里 | やまなし | “A village where you can see the moon” — that is, a place with no mountain (山なし, yama-nashi). A rebus reading rather than a kanji-by-kanji one. | Documented in the southern Kantō region and as the name used by a branch of the Kai Yamanashi clan. Surname dictionaries attest the reading but publish no household count. |
| Takanashi | 小鳥遊 | たかなし | “Small birds at play” — safe because there is no hawk: 鷹なし (taka-nashi). The characters show the picture, not the sound. | Concentrated in Nachikatsuura, Wakayama. The reading is well attested, but the documented line is the Takanashi (高梨) clan, so the image is a memory aid rather than a proven etymology. |
| Minai | 薬袋 | みない | 薬 (medicine) + 袋 (bag): a medicine pouch, read Minai rather than Kusuribukuro. | Attested in surname dictionaries with the single reading Minai. No household count is published, and the place of the same name in Yamanashi does not by itself show where the families live. |
| Dōmeki | 百目鬼 | どうめき | 百 (hundred) + 目 (eye) + 鬼 (ogre) — the hundred-eyed ogre of Utsunomiya folklore, whose name the characters spell out. | Surname dictionaries record six possible readings (Todomeki, Dōmeki, Domeki, Hyakumeki, Momeki, Momomeki), which is the clearest sign of a name outsiders cannot guess. Places of the same name exist in Yamagata and Tochigi. |
| Mitarai | 御手洗 | みたらい | 御 + 手 + 洗: the place at a shrine where visitors wash their hands. The same characters are also the everyday word for a washroom, which is why the name reads oddly. | Documented clan lines include the Kai Mitarai (Fujiwara/Takeda), the Chikuzen line, and the Bungo Ōga line. Dictionaries list six readings, and the clan reading is Mitarai. |
| Omameuda | 大豆生田 | おおまめうだ | 大豆 (soybean) + 生 + 田 (field), read as a single local word; variants such as おおまみゅうだ also exist. | The National Diet Library reference service names Sōnobe in Tochigi as the densest area. Surname dictionaries list around thirty readings for this one spelling — the clearest possible sign of a very small, very local name. |
| Kamatsuka | 一尺八寸 | かまつか | One shaku eight sun — about 54.5 cm, the length of a sickle handle. The name is effectively a written-out way of saying “sickle-handle”, a variant of 鎌塚. | Attested in Shizuoka and in surname dictionaries, read Kamatsuka or Kamazuka. The reading かざはな that circulates online has no dictionary headword and should not be repeated. |
| Shinonome | 東雲 | しののめ | “Eastern cloud” is a rebus for shinonome, the dawn sky; it is said to come from 篠の目, light coming through a gap in a bamboo blind. | Attested in surname dictionaries with a single reading. No household count or concentration is published, and the alternative reading Tōun belongs to place names only. |
| Kenmochi | 剣持 | けんもち | 剣 (sword) + 持 (bear, hold): a sword-bearer. Also written 劔持 or 釼持, which are rare variants of the same characters. | Attested in surname dictionaries, which specialise in uncommon names, but with no published count. The literal gloss “sword-bearer” is a reading of the characters, not a documented family history. |
| Wanibuchi | 鰐淵 | わにぶち | 鰐 (crocodile) + 淵 (deep pool): a rare-kanji combination with a strong landscape image. | A single attested reading. No count is published. Surnames built on characters this rare are hard to write and hard to pass on, which is part of why they stay small. |
| Kutsukake | 沓掛 | くつかけ | 沓 (a form of ancient footwear) + 掛 (hang): a place where footwear was taken off or hung. The same characters name post towns on the old highways. | Surname dictionaries give several readings — Kutsukake, Kusakake, Kutsugake, Tōgake, Mizukake, Mitsukake — so the reading has to be asked for, not guessed. No count is published. |
| Abiko | 我孫子 | あびこ | 我 (my) + 孫子 (grandchild), literally “my grandchild” — an unexpected phrase to find on a family name. | A genuine surname with documented living bearers. The city in Chiba and the district in Osaka share the characters, but those are place names and say nothing about how many families use the surname. |
| Nekoyashiki | 猫屋敷 | ねこやしき | 猫 (cat) + 屋敷 (residence, estate): a name taken from a place, and one that contains no landscape or clan element at all. | Documented as taken from a place name in Karumai, Iwate, with a single reading in the dictionaries. A one-locality origin makes it geographically rare; no count is published. |
| Shōji | 東海林 | しょうじ | 東 (east) + 海 (sea) + 林 (grove): three kanji that do not spell the reading at all, which is the whole difficulty of the name. | Attested in surname dictionaries and strongly associated with the Shonai district of Yamagata. A textbook nandoku name: the reading must be learned, because no rule leads to it. |
The Smallest Names in Published Household Surveys
The names below come from published surname surveys and are counted in people, not described vaguely: several are recorded with about ten people nationwide, and two show how a name can be tiny nationally while still belonging to one region. Survey figures are estimates taken at a survey date, so read them as scale, not as a register.
| Romaji | Kanji | Reading | Elements and meaning | Where it is found and why it is rare |
|---|---|---|---|---|
| China | 地名 | ちな | 地 (chi, ground) + 名 (na, name): literally “place name”, a surname taken from the name of a locality. | Roughly ten people nationwide. Recorded with the readings China, Jimei, and Chimei; the characters match a former district in Shizuoka, and the surviving families are reported mainly in Aomori. |
| Ringo | 林檎 | りんご | The kanji for “apple”, chosen as a surname rather than as a landscape or occupation word. | About twenty people. Traditionally traced to Mutsu, and now concentrated in Ōno, Hirono, Iwate. |
| Kikori | 樵 | きこり | A single kanji meaning “woodcutter”, a surname that records an occupation directly. | About forty people. Described as coming from the woodcutter’s trade, with most families in Fukuoka. |
| Nakabeppu | 中別府 | なかべっぷ | 中 (naka, middle) + 別府 (beppu), a place element often linked to separate allotments or a separate estate. | About 240 people, the rarest surname inside the published 2025 top-12,000 ranking. Origins are given as Kagoshima and Miyazaki, with the largest group in Miyazaki. |
| Morota | 師田 | もろた | 師 (moro, teacher or master) + 田 (ta, rice field), a compound that pairs a title with farmland. | About 180 people. Traced to Gunma, but the largest single group now lives in Tokyo — a reminder that a locally rooted name can be rare nationally and still appear in the capital. |
| Ichiyō | 鴨脚 | いちょう | 鴨 (duck) + 脚 (foot): the ginkgo leaf was compared to a duck’s foot, so the characters are read Ichiyō, meaning “ginkgo”. | A rare rebus surname whose reading comes from the shape of the leaf rather than from the characters themselves. |
| Achichi | 兄父 | あちち | 兄 (elder brother) + 父 (father), an unexpected pairing of two kinship kanji. | Readings vary by family — Ajichi, Achichi, Keifu, or Ajichi — which is typical of very small surnames whose reading was never standardised. |
| Mafune | 馬舟 | まふね | 馬 (horse) + 舟 (boat), also read Umafune; a pairing of two unrelated vehicle words. | Only around ten people anywhere in the country, listed near 89,500th in the published survey. |
| Kaji | 鹿耳 | かじ | 鹿 (deer) + 耳 (ear), a nature compound that reads Kaji rather than the expected Shika-mimi. | A recorded population of about ten, ranked near 77,600th in the published survey. |
| Kinuhata | 絹畠 | きぬはた | 絹 (silk) + 畠 (a variant of 畑, cultivated field), pairing a textile word with farmland. | Held by roughly ten people; ranked near 71,400th in the published survey — the highest place in this group. |
| Ebinishi | 鰕西 | えびにし | 鰕 (shrimp, written with a kanji that is itself rare) + 西 (west). | Around ten people in the survey, near 92,500th. |
| Okishima | 起島 | おきしま | 起 (rise) + 島 (island), read Okishima; the first kanji is unusual in a topographical surname. | Among the smallest names reported: about ten people, ranked near 95,800th in the published survey. |
How This List Was Checked
Surnames were checked against Japanese surname dictionaries and reference sources, not reverse-translated from English keyword lists. Where a reading is disputed, the table gives the reading used by the surname dictionaries and the text says so. Rarity notes are based on published household estimates, which are surveys, not exact counts: they measure households, they are collected in different years, and they do not cover every name in use.
One limit is worth stating plainly. A name that does not appear in a top-ranked list is not automatically rare — it may simply be too small to rank. For that reason, this page calls a surname rare only when a dictionary entry, a regional distribution note, or a reading note supports it.
Reference sources used for this list: 名字由来net (Myoji Yurai Net surname database) · ReseMom — surnames with about ten people nationwide (2025) · ReseMom — 2025 national surname ranking, top 12,000 · ReseMom — the rarest surnames of 2016 · JMnedict (EDRDG Japanese name dictionary) · 名字辞典 (Myōji Jiten) surname-dictionary entries via weblio.jp · National Archives of Japan — Surname Day and the 1870/1875 surname laws · Ministry of Justice — who may request a family register (koseki)
Uncommon Regional Surnames
Rarity is usually local before it is national. A surname can be carried by hundreds of households in one prefecture and by almost none anywhere else, which is why "rare in Japan" and "rare where you live" are two separate questions. The groups below follow the broad regions where unusual surnames are concentrated.
Where Rare Surnames Cluster
Hokkaidō and Tōhoku
China (地名), Ringo (林檎), Shōji (東海林)
Northern examples: Ringo is concentrated in Ōno, Hirono, Iwate, and China is reported mainly in Aomori. Shōji belongs to the Shonai district of Yamagata, and Shonai is also the classic case of a surname that is ordinary locally and unfamiliar elsewhere.
Kyūshū
Kikori (樵), Nakabeppu (中別府)
Kyūshū supplies several of the smallest names in the surveys: Kikori is most numerous in Fukuoka, and Nakabeppu, the rarest entry in the 2025 top-12,000 ranking, is traced to Kagoshima and Miyazaki.
Kantō
Morota (師田), Omameuda (大豆生田), Abiko (我孫子)
Morota originates in Gunma but its largest group is now in Tokyo, and Omameuda is densest at Sōnobe in Tochigi. Regional distribution matters more than a national total: a name can be locally ordinary and nationally rare.
Okinawa and the Ryukyu Islands
Higa (比嘉), Kinjō (金城), Tamaki (玉城)
Okinawan surnames run on a separate track. Higa, Kinjō, Tamaki and their neighbours are everyday names in Okinawa while staying far down the national rankings, and mainland distribution tables under-represent them because the registers and readings developed differently.
Okinawa and the Ryukyu Surnames
Okinawan family names developed along a separate path from mainland names. Many are distinct in both kanji and reading, and mainland frequency tables under-represent them because those tables are built from national household surveys in which Okinawan names are a small share. A name that looks rare in a national ranking can be ordinary in Naha or on a neighbouring island.
Hokkaidō is the mirror case. Settlement from the mainland in the late nineteenth and early twentieth centuries brought common mainland surnames to the island, so genuinely local naming traditions are the minority there. Reading regional distribution instead of a national count is what separates a rare name from a merely displaced one.
Rare vs Unique vs Extinct Surnames
These three words get used as if they meant the same thing. They do not, and the difference matters when you are checking a name you have found.
| Term | What it means | How to tell |
|---|---|---|
| Rare | Few households carry it, but the name is documented and still in use. | A dictionary entry exists, and household estimates are low or regionally limited. |
| Unique | The name reads as distinctive, new, or unusual in style — it need not be numerically rare. | Judged by sound, kanji choice, or fashion; a unique name can still be common. |
| Extinct | No household is recorded using it. The line, not merely the spelling, has ended. | Absence from current registers plus a historical record. Absence alone is not proof. |
Frequency Caveats: What the Numbers Do Not Show
Published surname counts measure households, not people, and they are collected at different times from different registers. A name with a low household count may still be perfectly familiar in the town where it lives, and a name with a high count may belong to one extended family. Coverage is another limit: small surnames fall below the reporting threshold, so the bottom of a ranking is a fog, not a floor.
There is also no reliable public listing of extinct surnames, and the logic runs against you. Proving that a surname exists takes a single confirmed bearer. Proving that it does not exist would mean inspecting every family register in the country, which privacy law forbids, so the honest wording is “no confirmed living bearer could be found” rather than “this surname is extinct”. A name can disappear from a commercial ranking because it was always small, because families moved, or because the reading changed — none of which means the family ended.
Creative Use: What These Names Are Not
If you are writing a novel, running a game, or naming a business, these surnames are still real family names. Do not combine kanji into a new string and present it as a rare Japanese surname: invented combinations usually read as unpronounceable or accidentally comic to a native speaker. Do not claim a lineage or a samurai ancestry that the name does not support, and do not put a real rare family name on a product in a way that implies the family is involved.
Using a genuine, well-documented surname for a character is fine. In that case, get the reading right, keep the name spelled consistently, and remember that a reader who knows the name may recognise it. If a name is famous locally, that recognition is part of the effect you are choosing.
How to Research a Japanese Family Name
Researching a surname is mostly a process of eliminating the wrong readings. Work through these steps in order, and expect the kanji to come last, not first.
- 1Write the name in romaji, kana, and kanji if you have them. If you only have romaji, list every kanji combination that shares that reading — most readings have several.
- 2Check the reading against a Japanese surname dictionary. These entries usually give the element meaning, the plausible readings, and a note on where the name is found.
- 3Look at prefecture-level distribution, not the national ranking. A surname concentrated in one prefecture is a regional name, and that changes what the kanji are likely to mean.
- 4Confirm the family’s own record. The koseki (family register) is the authoritative source. It is held by the municipality of the registered domicile, and it can be requested by the person recorded on it, their spouse, and their lineal ascendants or descendants — siblings, cousins, and in-laws have no automatic right.
- 5Check the kamon (family crest) and any hinomaru or grave records the family holds. These often survive when documents do not, and they can distinguish two families that share a surname.
- 6For emigrant families, work backwards through immigration, census, and naturalisation records in the destination country. Spelling in these documents is a clerk’s best attempt at the sound, so expect several versions of the same name.
What You Cannot Find Online
Japanese family registers are not public. A koseki can be requested only by the person named on it, their spouse, or their lineal ascendants or descendants, at the municipality where it is registered, so no website can look up another family’s ancestry for you. The register is also organised by registered domicile and head of household, not by surname, which means a searchable national surname index cannot exist even in principle. Sites that offer a surname-to-address lookup are selling aggregated data of uncertain origin, and the result will not tell you who your relatives are.
What you can do online is verify the name itself: the reading, the kanji options, the element meanings, and the regional pattern. That is enough to rule out impossible answers and to ask a better question of the records that only the family can open.
Rare Japanese Last Names: Frequently Asked Questions
What makes a Japanese last name rare?
A surname is rare when few households carry it. The usual signals are a low household estimate, a distribution limited to one prefecture or town, or a dictionary entry that records the name as unusual. Rarity is measured in households, not people, so the counts are estimates, not exact totals.
What are some rare Japanese last names?
Documented examples include 四月一日 (Watanuki), 八月一日 (Hozumi), 小鳥遊 (Takanashi), 月見里 (Yamanashi), 百目鬼 (Dōmeki), 薬袋 (Minai), and 大豆生田 (Omameuda), whose readings are recorded in Japanese surname dictionaries. Others are small by count or by place: 地名 (China), 馬舟 (Mafune), and 鹿耳 (Kaji) are reported with about ten people nationwide, and 中別府 (Nakabeppu) with roughly 240.
Is my Japanese last name rare?
Check two things: the household estimate in a surname dictionary, and the prefecture-level distribution. A name that looks rare nationally may be ordinary in one prefecture, and a name absent from a national ranking may simply be too small to be listed. Only the second pattern, combined with a documented low count, supports calling a name rare.
Why are some Japanese surnames so hard to read?
Hard-to-read surnames are called nandoku myōji. Many use ateji or gikun, where kanji were attached to an existing reading, so the characters suggest an image rather than a pronunciation. Others use rare kanji or an unusual combination of common ones, and a Japanese speaker may have to ask, not guess.
Do Japanese surnames ever go extinct?
A surname can fall out of use when a family line ends or when the name is changed, but there is no reliable public register of extinct surnames. Absence from a current ranking only shows that the name is small, not that it has ended, so treat extinction as something to establish from records rather than from a list.
Can I use a rare Japanese surname for a character or a business?
You can use a real, documented surname, but do not invent kanji combinations and present them as genuine. Get the reading right, keep the spelling consistent, and avoid implying a family lineage or a samurai ancestry that the name does not have. For a business, check trademarks and domains as well.
From Rare Names to the Full Surname List
You now have a documented set of uncommon Japanese surnames, the elements behind them, the regions where they cluster, and a method for checking a name properly. Continue with the full A–Z surname browser, the most common last names, or the surname meaning pages.