Compiled by the UNGEGN Working Group on Romanization Systems
Version 4.0, March 2016


The United Nations recommended system was approved in 1987 (resolution V/19), based on the ELOT 743 conversion system of the Greek Standardization Organization. The table was published as an annex to the resolution1.

The system is used in Greece and Cyprus, also increasingly in international cartographic products. Many systems of romanization adapted to transcribing names from older forms of Greek continue to be used in other countries. The 1987 resolution acknowledged that "a suitable transitional period will be required before the system can be fully implemented".

The romanization table is unambiguous and simple to use. There are two versions of romanization which complement each other: transcription and transliteration. The latter differs from the former only by adding a sub-macron to certain ambiguous romanization equivalents.


The table below contains the transcription version of the romanization system. See note 1 for the transliteration equivalents.

1 Α α a
αυ avA, afB
2 Β β v
3 Γ γ g
γγ ng
γξ nx
γχ nch
4 Δ δ d
5 Ε ε e
ευ evA, efB
6 Ζ ζ z
7 Η η i
ηυ ivA, ifB
8 Θ θ th
9 Ι ι i
10 Κ κ k
11 Λ λ l
12 Μ μ m
μπ bC, mpD
13 Ν ν n
14 Ξ ξ x
15 Ο ο o
ου ou
16 Π π p
17 Ρ ρ r
18 Σ σ ς s
19 Τ τ t
20 Υ υ y
21 Φ φ f
22 Χ χ ch
23 Ψ ψ ps
24 Ω ω o


  1. In order to ensure reversibility a sub-macron may be added to the following characters and character combinations: 1 αυ av̠/af̠, 3 γγ ṉg, γξ ṉx, γχ ṉch, 5 ευ ev̠/ef̠, 7 η , ηυ i̠v̠/i̠f̠, 24 ω .
  2. Some examples of the ELOT 743 table contain also combinations in parentheses αι, άι, αϊ, γκ, ει, έι, εϊ, ντ, οι, όι, οϊ, υι which are converted according to the conversion rules of each character they consist of, resp. ai, ái, aï, gk, ei, éi, eï, nt, oi, ói, oï, yi.
  3. The combinations αυ, ευ, ου are converted according to the conversion rules of each character when the vowel before υ has an accent or when υ has the diaeresis sign.
  4. When the character combinations αυ, ευ, ηυ are stressed in Greek, in the Roman script the accent is set on the vowel (áv, áf, év, éf, ív, íf).

Other systems of romanization

The BGN/PCGN 1962 System that continued to be used until 1996, provided for the romanization, as a single block, of the following characters or character combinations differently from the UN system (the Greek character is followed in parentheses by the romanization according to the UN system):

αυ (av/af) av
γ (g) g, yA
γχ (nch) nkh
γκ (gk) gB, ngC
δ (d) dh, dD
ει (ei) iE
ευ (ev/ef) ev
ηυ (iv/if) iv
μπ (b/mp) bB, mbC
ντ (nt) dB, ndC, ntF
υ (y) i
υι (yi) i
χ (ch) kh

The transliteration ISO 843:1997 which is also based on ELOT 743 presents the table of type 1 (transliteration of Greek characters into Latin characters) whereby each Greek character has only one Latin equivalent. The following consonants are romanized slightly differently from the UN-approved system:

Η η Ī ī or I¯ i¯
Ω ω Ō ō or O¯ o¯

Combinations of two or more characters are transliterated according to the provisions laid down for each independent character. The only exception to this rule for transliteration is the conversion of the Greek double vowels ΑΥ, αυ, ΕΥ, ευ, ΟΥ, Ου which are mapped into Roman script as AU, au, EU, eu, OU, ou respectively.

The transliteration version of the system approved by the UN conference in 1987 is referred to in the standard as reversible transcription (type 2) and is given as informative Annex B.


  1. Fifth United Nations Conference on the Standardization of Geographical Names. Montreal, 18-31 August 1987. Vol. I. Report of the Conference, pp. 42–43.