diff --git a/news/load-data-pdfparser.rst b/news/load-data-pdfparser.rst
new file mode 100644
index 0000000..3863d17
--- /dev/null
+++ b/news/load-data-pdfparser.rst
@@ -0,0 +1,23 @@
+**Added:**
+
+* <news item>
+
+**Changed:**
+
+* <news item>
+
+**Deprecated:**
+
+* <news item>
+
+**Removed:**
+
+* <news item>
+
+**Fixed:**
+
+* Read the PDF metadata in the data file header in ``PDFContribution.load_data``, which used the generic ``ProfileParser`` and lost the scattering type and qmax from NOMAD files with a free-text instrument comment.
+
+**Security:**
+
+* <news item>
diff --git a/src/diffpy/cmipdf/pdfcontribution.py b/src/diffpy/cmipdf/pdfcontribution.py
index 0e69bed..a275bc5 100644
--- a/src/diffpy/cmipdf/pdfcontribution.py
+++ b/src/diffpy/cmipdf/pdfcontribution.py
@@ -23,7 +23,8 @@
 
 from diffpy.cmipdf.debyepdfgenerator import DebyePDFGenerator
 from diffpy.cmipdf.pdfgenerator import PDFGenerator
-from diffpy.srfit.fitbase import FitContribution, Profile, ProfileParser
+from diffpy.cmipdf.pdfparser import PDFParser
+from diffpy.srfit.fitbase import FitContribution, Profile
 
 
 class PDFContribution(FitContribution):
@@ -109,14 +110,17 @@ def __init__(self, name):
     # Data methods
 
     def load_data(self, datafile):
-        """Load the data from a datafile.
+        """Load the data from a data file.
+
+        This uses the PDFParser to load the data and then passes it to
+        the built-in profile with load_parsed_data.
 
         Parameters
         ----------
-        data : str or Path
-            The path to the data file.
+        datafile : str or Path
+            The path to the file that contains the data.
         """
-        parser = ProfileParser()
+        parser = PDFParser()
         parser.parse_file(datafile)
 
         # Pass it to the profile
diff --git a/tests/test_pdfcontribution.py b/tests/test_pdfcontribution.py
index 466a0d1..36e30fd 100644
--- a/tests/test_pdfcontribution.py
+++ b/tests/test_pdfcontribution.py
@@ -71,6 +71,29 @@ def test_get_qmax_from_profile():
     assert actual_qmax == expected_qmax
 
 
+# load_data reads the PDF metadata in the data file header, so the
+# contribution takes its scattering type and qmax from the file.
+@pytest.mark.parametrize(
+    "input_filename, expected_settings",
+    [
+        # C1: The header holds PDFgetX3 "name = value" lines.
+        # Expected: The scattering type "X" and qmax 27 from the header.
+        ("si-q27r60-xray.gr", ("X", 27.0)),
+        # C2: The header holds a free-text NOMAD instrument comment.
+        # Expected: The scattering type "N" and qmax 31.414 from the comment.
+        ("nom-mno-neutron.gr", ("N", 31.414)),
+    ],
+)
+def test_load_data_reads_metadata(datafile, input_filename, expected_settings):
+    contribution = PDFContribution("pdf")
+    contribution.load_data(datafile(input_filename))
+    actual_settings = (
+        contribution.get_scattering_type(),
+        contribution.get_qmax(),
+    )
+    assert actual_settings == expected_settings
+
+
 def test_savetxt(datafile):
     # C1: A contribution with data on r = 0 to 10 is evaluated and saved.
     # Expected: The file has a header line and one line for each of the