diff --git a/docs/downloading.rst b/docs/downloading.rst
index 2bf6949..255a2e0 100644
--- a/docs/downloading.rst
+++ b/docs/downloading.rst
@@ -47,6 +47,25 @@ download the *summary file* of the ORCID Public Data File
works on the compressed file; there is no need to expand it (it expands
to about 0.5 TB).
+PubMed data
+~~~~~~~~~~~
+
+You can populate a database with the data from the PubMed/MEDLINE database
+from the National Library of Medicine (NLM).
+The data are available from
+`here `__.
+You can also download the data from the FTP server which is documented in
+`this readme file `__.
+Publication data are overlapping with the Crossref dataset,
+but the PubMed data contain additional information
+such as MeSH terms and grants.
+They also complement the Crossref data as for example some affiliations
+are only available in the PubMed data.
+
+The shell script ``examples/common/fetch-pubmed.sh`` or the
+``$(PUBMED_DIR)`` dependency of the ``examples/common/Makefile``,
+which uses it, can be used to download all PubMed data.
+
USPTO data
~~~~~~~~~~
@@ -88,26 +107,6 @@ The shell script ``examples/common/fetch-uspto.sh`` or the
``$(USPTO_DIR)`` dependency of the ``examples/common/Makefile``,
which uses it, can be used to download all USPTO data.
-PubMed data
-~~~~~~~~~~~
-
-You can populate a database with the data from the PubMed/MEDLINE database
-from the National Library of Medicine (NLM).
-The data are available from
-`here `__.
-You can also download the data from the FTP server which is documented in
-`this readme file `__.
-Publication data are overlapping with the Crossref dataset,
-but the PubMed data contain additional information
-such as MeSH terms and grants.
-They also complement the Crossref data as for example some affiliations
-are only available in the PubMed data.
-
-The shell script ``examples/common/fetch-pubmed.sh`` or the
-``$(PUBMED_DIR)`` dependency of the ``examples/common/Makefile``,
-which uses it, can be used to download all PubMed data.
-
-
Other data sources
~~~~~~~~~~~~~~~~~~