D with Line Through It: đ in Croatian, dj in ASCII, Not ₫

Which d with a line through it is the one you saw?

Most of the time it is đ, U+0111, the d with stroke of Croatian, Bosnian, Serbian and Vietnamese. A search for d with line through it can also mean one of several other characters, each with its own job. The notes below are the ones Unicode prints in its code charts for Latin Extended-A (opens in a new tab), Latin Extended-B (opens in a new tab), Phonetic Extensions (opens in a new tab), Latin Extended-D (opens in a new tab) and Currency Symbols (opens in a new tab). Every name except the dong sign's starts with LATIN SMALL LETTER, left out here.

Character	Code point	Unicode name	The line	What it is for
đ	U+0111	D WITH STROKE	across the stem	Croatian, Serbian, Vietnamese, Sami, Moro
ð	U+00F0	ETH	across a curved stem	Icelandic and Faroese, in AccentPad's facts
₫	U+20AB	DONG SIGN	across the stem, plus one under the letter	the currency of Vietnam
ꟈ	U+A7C8	D WITH SHORT STROKE OVERLAY	a short stroke	tau gallicum, a dental affricate in Gaulish
ƌ	U+018C	D WITH TOPBAR	a bar on top	former-Soviet minority scripts, old Zhuang spelling
ᵭ	U+1D6D	D WITH MIDDLE TILDE	a wavy line	a phonetic letter; the chart gives no note

Two notes in the đ entry matter. The chart points from đ to ð, so Unicode itself flags the pair as easy to confuse; the eth letter page covers that side. It also says Americanist phonetic writing may draw the stroke through the bowl, the round part, instead of the stem. The character is still U+0111.

Unicode's confusables list (opens in a new tab), defined in UTS #39 (opens in a new tab), reduces đ to a d with a short stroke overlay. It reduces ₫ to the same thing plus a macron below, so a spoofing check reads the dong sign as an underlined đ. In lowercase it keeps đ and ð apart: ð reduces to the partial derivative sign ∂ with a stroke.

Why does a Croatian d turn into đ in mlađi and rođak?

Because many Croatian words with đ are built on a word with d. The Institute of Croatian Language and Linguistics lists the cases in Hrvatski pravopis (opens in a new tab), its spelling manual: đ stands where the base word has d, in comparatives, in verb forms and in derived words.

Base word with d	Form with đ	What kind of form
mlad, young	mlađi, younger	comparative
lud, crazy	luđi, crazier	comparative
naći, to find	nađem, nađu	present tense
grad, town	predgrađe, suburb	derived noun
rod, kin	rođak, relative or cousin	derived noun
grozd, cluster of grapes	grožđe, grapes	derived noun

The glosses come from Wiktionary's entries for mlad (opens in a new tab), lud (opens in a new tab), naći (opens in a new tab), grad (opens in a new tab), predgrađe (opens in a new tab), rod (opens in a new tab), rođak (opens in a new tab), grozd (opens in a new tab) and grožđe (opens in a new tab).

Drop the stroke and the word changes. Wiktionary's entry for mlad lists mladi, with a plain d, as the definite form, the young one, and mlađi as the comparative. Its entry for lud pairs ludi and luđi the same way.

Other words have đ in the root, with no d form behind them. Pravopis lists anđeo (opens in a new tab), angel, lađa (opens in a new tab), a riverboat, and đak (opens in a new tab), a pupil, among them.

Does đ come before or after dž in the Croatian alphabet?

After. The letters chapter (opens in a new tab) of Hrvatski pravopis fixes the order as d, dž, đ, e, and names the letter đe. Dž counts as one letter, though it takes two signs: a word set in capitals writes it DŽ, as in DŽIBUTI. The CLDR Croatian collation (opens in a new tab) carries that order into software.

We measured what it does to a sorted list on 2026-10-10, in Node.js v26.0.0 with ICU 78.3 and CLDR 48 data. The words are the same in both columns; only the rules differ.

Sorted under English rules	Sorted under Croatian rules
đak	dan
Đakovo	Dubrovnik
dan	dupin
đon	džem
Dubrovnik	džep
dupin	đak
džem	Đakovo
džep	đon

English rules treat đ as a d with a mark, so đak lands among the da- words. Croatian rules wait until every d and dž word is done. Bosnian and Serbian Latin rules, run the same way, gave the Croatian column. An index sorted with English rules puts đak where a Croatian reader would not look for it.

Should đ become dj or d when a name goes into plain ASCII?

It depends on the language, and on who will read the result. Wiktionary's usage note for Serbo-Croatian says đ is sometimes written dj. The Serbian Academy of Sciences and Arts shows the habit on its own site. On 2026-10-10 the title of its page on Đura Daničić (opens in a new tab) spelled Đura, while the web address spelled djura. Vietnamese drops the stroke instead: the Đà Nẵng city portal lives at danang.gov.vn (opens in a new tab).

Passports follow a third rule. The transliteration table in ICAO Doc 9303 (opens in a new tab), the standard for machine-readable travel documents, sends U+0110, Đ, to a plain D. Its Cyrillic table sends U+0402, the capital of Serbian ђ, to D as well.

Code that strips accents follows none of them. The usual strip is NFD (opens in a new tab) followed by deleting the combining marks, and Unicode gives đ no decomposition to delete. We ran it on 2026-10-10 in Node.js v26.0.0. Željko became Zeljko and Ćiro became Ciro, while Đorđe came out exactly as it went in. Thủ Đức came out as Thu Đuc, only half stripped.

Name, from AccentPad's facts	With the dj convention	ICAO machine-readable line	After NFD and deleting marks
Đorđe	Djordje	DORDE	Đorđe
Srđan	Srdjan	SRDAN	Srđan
Anđela	Andjela	ANDELA	Anđela
Đurađ	Djuradj	DURAD	Đurađ

The second and third columns are those rules applied by us, letter by letter. Where a system accepts Unicode, keep đ. Where it does not, a Croatian or Serbian name reads best with dj, and a form that has to match a passport should ask for the machine-readable spelling. Either way, ASCII-only code needs an explicit entry that maps đ, because the accent strip never will.

How does Serbian Cyrillic ђ become the Latin đ?

Letter for letter. The UN system for romanizing Serbian, documented by the UNGEGN Working Group on Romanization Systems (opens in a new tab), maps Ђ ђ to Đ đ and calls the table unambiguous and reversible. The report says it is used in Serbia, Montenegro, and Bosnia and Herzegovina. The letters around đ in the table work the same way.

Serbian Cyrillic	Serbian Latin
Ђ ђ	Đ đ
Џ џ	Dž dž
Ћ ћ	Ć ć
Љ љ	Lj lj
Њ њ	Nj nj

The digraphs dž, lj and nj each stand for a single Cyrillic letter, and Croatian spelling also counts each of them as one letter. So a Serbian text can switch scripts without losing anything. Turning it to Cyrillic gives ђ for every đ, and back again.

Macedonian is the trap. The UN system for Macedonian (opens in a new tab) also writes đ, but for ѓ, a different letter, and it writes g in its place before e and i. A đ in a romanized Macedonian place name is therefore not the Serbian ђ.

What does đ sound like outside Croatian and Vietnamese?

Different things, in the handful of other alphabets that use it. The rows come from the language sections of Wiktionary's entry for đ (opens in a new tab). The Unicode chart adds Moro to the list of users, without a sound.

Language	Sound of đ	Note
Serbo-Croatian	/d͡ʑ/	as in đak, /d͡ʑâːk/
Vietnamese	[ʔɗ]	the letter is called đê or đờ
Northern Sami, Skolt Sami	/ð/	the th of this
Sicilian	/ɖ/	less common today, often written dd
Sylt North Frisian	[l], or [r] for some speakers	it stood for an earlier [ð]
Slovene	/d͡ʒ/	mostly in loanwords from Serbo-Croatian
Brahui	/ɖ/	the voiced retroflex stop, as Unicode describes ɖ

The Sami row is the sound of the eth letter, written with another character. The Sylt row explains its own shape: Wiktionary says the đ was chosen because the sound used to be [ð]. For the Vietnamese side, with D and Đ as two letters, see the Vietnamese alphabet page.