How should NLM DTD records be formatted?


The following elements are required on all NLM DTD records.   For additional help see the National Library of Medicine documentation at:

See the links at the end of this page to sample records and a template.


A DTD describes the structure of an XML record as well as the
relationship of the elements to the application. It must appear at the
beginning of each set of records, or on each record if a file contains only
one record. Records without a schema or DTD will be rejected. 

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE ArticleSet PUBLIC "-//NLM//DTD PubMed 2.6//EN" "">

Journal Title:

The journal title should be transcribed as it appears on the journal.  Do not use the abbreviated form the title.

Capitalization guidelines follows the same policy as stated in <ArticleTitle>. Do not include a journal’s subtitle unless it’s needed to clarify the meaning of the main title.

<JournalTitle>Journal of food, agriculture and environment</JournalTitle>

ISSN (International Standard Serial Number):

Metadata samples will not be approved without an ISSN that is registered with the ISSN International Centre in Paris. ISSNs can be obtained in each country through the local ISSN Centre, which is often the country’s national library.

For more information see:

All ISSN’s appearing on the publication should be transcribed. This usually means ISSNs for both the electronic and print versions.

Volume Number/Issue Number:

Transcribe the volume and issue numbers as they appear on the publication. Roman rather than Arabic numbers may be used if they appear on the publication.


Article Title:

The first word, proper nouns and names should be capitalized. All other letters should be in lower case. Capitalize the first word of biological names. Do not use HTML coding for italics.

Article titles must be in English. If the title appears in a second language, it may be added as a translated title using the element <VernacularTitle>, even if it is in the language of the text.

a) Rheological characterisation of the fractions separated from pork lards through dry fractionation
b) Developing a savanna burning emissions abatement methodology for tussock grasslands in high rainfall regions of northern Australia
c) Susceptibility of developmental stages of Tetranychus urticae (Acari: Tetranychidae) to infection by Beauveria bassiana and Metarhizium anisopliae (Hypocreales: Clavicipitaceae)

d)Text is in Portuguese; titles and abstracts in Portuguese and English.

<ArticleTitle>Drippers flow disturbances due to application of CO2 for irrigation water</ArticleTitle>
<VernacularTitle>Disturbios de vaz~ao em gotejadores devido a applic~ao de CO2 via agua de irrigac~ao</VernacularTitle>

Links to full-text:

Only DOIs (Digital Object Identifiers) may be used with the NLM DTD.  Publishers who wish to use URLs to link records to online content must use either the MODS or JATS metadata formats.

<ELocationID EIdType="doi">10.7127/rbai.v8N400213</ELocationID> 


PubMed DTD language code should be "EN" for English, not En-US


For other language codes, see the link below.

Extent (e.g. pagination):

Transcribe the pagination as it appears in the article. HTML created articles do not have page numbers, so do not include them in records.


If the letter “P” appears with the page number, include it in the metadatda:

Publication date:

<PubDate pubstatus="epublish"> electronic publication

PubMed DTD:
<Pubdate pubstatus="ppublish"> print publication

Personal Authors/Contributors and their affiliations:

Last name, first name or initial, middle name should be in tagged elements as required by the format.

Records will be rejected if first name, middle name and last name are in one tagged element.

Diacritics can be included with names.

Include affiliations of all authors, if possible; affiliation is mandatory for the first named author.
More than one affiliation per author may be recorded within the affiliation tag. Separate the locations with a period.

<Affiliation>Theoretical High School "J.L. Calderon," Pestalozzi 14, 300115 Timisoara, Romania; (link sends e-mail)</Affiliation>

<LastName>Furtada da Silva</LastName>
<Affiliation>IF Goiana, Campus Rio Verde</Affiliation>


The abstract should be in plain Unicode text. It should not include HTML or other format encoding elements. Please refer to the FAQ "How are entity references coded (symbols that need to be converted to codes)?" for the exceptions.

An English abstract is required on each record in order for it to be indexed. If an abstract is available in a language in addition to English, that abstract must be in a separate tagged element. The schema used to code the metadata will determine how the multiple language abstracts should be coded.

Records with multiple languages in one <abstract> element will not be accepted.

Abstracts should follow same capitalization rules as titles. There should be a period (.) at the end of the abstract.

Example (in English only)

Example (In English and Portuguese)
<OtherAbstract language="pt">…</OtherAbstract>