diff --git a/rdflib/graph.py b/rdflib/graph.py
index 9ba3dd396..d88fe0ac1 100644
--- a/rdflib/graph.py
+++ b/rdflib/graph.py
@@ -2349,7 +2349,7 @@ def parse(
         file: BinaryIO | TextIO | None = None,
         data: str | bytes | None = None,
         **args: Any,
-    ) -> Graph:
+    ) -> "ConjunctiveGraph":
         """
         Parse source adding the resulting triples to its own context (sub graph
         of this graph).
@@ -2406,8 +2406,7 @@ def parse(
 
         context = self.default_context
         context.parse(source, publicID=publicID, format=format, **args)
-        # TODO: FIXME: This should not return context, but self.
-        return context
+        return self
 
     def __reduce__(self) -> tuple[type[Graph], tuple[Store, _ContextIdentifierType]]:
         return ConjunctiveGraph, (self.store, self.identifier)
@@ -2644,7 +2643,7 @@ def parse(
         file: BinaryIO | TextIO | None = None,
         data: str | bytes | None = None,
         **args: Any,
-    ) -> Graph:
+    ) -> "Dataset":
         """
         Parse an RDF source adding the resulting triples to the Graph.
 
@@ -2681,8 +2680,11 @@ def parse(
         c = ConjunctiveGraph.parse(
             self, source, publicID, format, location, file, data, **args
         )
-        self.graph(c)
-        return c
+
+        for context in c.contexts():
+            self.graph(context)
+
+        return self
 
     def add_graph(self, g: _ContextIdentifierType | _ContextType | str | None) -> Graph:
         """alias of graph for consistency"""
diff --git a/rdflib/plugins/parsers/nquads.py b/rdflib/plugins/parsers/nquads.py
index 2ed2ab4a1..11815be42 100644
--- a/rdflib/plugins/parsers/nquads.py
+++ b/rdflib/plugins/parsers/nquads.py
@@ -7,7 +7,7 @@
 >>> g = ConjunctiveGraph()
 >>> data = open("test/data/nquads.rdflib/example.nquads", "rb")
 >>> g.parse(data, format="nquads") # doctest:+ELLIPSIS
-<Graph identifier=... (<class 'rdflib.graph.Graph'>)>
+<Graph identifier=... (<class 'rdflib.graph.ConjunctiveGraph'>)>
 >>> assert len(g.store) == 449
 >>> # There should be 16 separate contexts
 >>> assert len([x for x in g.store.contexts()]) == 16
diff --git a/test/test_conjunctivegraph/test_conjunctive_graph.py b/test/test_conjunctivegraph/test_conjunctive_graph.py
index 0fdb0af1d..78c6f6821 100644
--- a/test/test_conjunctivegraph/test_conjunctive_graph.py
+++ b/test/test_conjunctivegraph/test_conjunctive_graph.py
@@ -2,6 +2,7 @@
 Tests for ConjunctiveGraph that do not depend on the underlying store
 """
 
+from io import StringIO
 import pytest
 
 from rdflib import ConjunctiveGraph, Graph
@@ -83,3 +84,21 @@ def check(kws):
 @pytest.mark.parametrize("checker, kws", get_graph_ids_tests())
 def test_graph_ids(checker, kws):
     checker(kws)
+
+
+def test_parse_return_type():
+    g = ConjunctiveGraph()
+    g.parse(data=DATA, format="turtle")
+    assert type(g) is ConjunctiveGraph
+
+    g = ConjunctiveGraph()
+    g = g.parse(data=DATA, format="turtle")
+    assert type(g) is ConjunctiveGraph
+
+    g = ConjunctiveGraph()
+    g.parse(source=StringIO(DATA), format="turtle")
+    assert type(g) is ConjunctiveGraph
+
+    g = ConjunctiveGraph()
+    g = g.parse(source=StringIO(DATA), format="turtle")
+    assert type(g) is ConjunctiveGraph
diff --git a/test/test_dataset/test_dataset.py b/test/test_dataset/test_dataset.py
index 9f9bc9c26..36b1c1cf8 100644
--- a/test/test_dataset/test_dataset.py
+++ b/test/test_dataset/test_dataset.py
@@ -1,7 +1,9 @@
 import os
 import shutil
 import tempfile
-import warnings
+from io import StringIO
+from test.data import CONTEXT1, LIKES, PIZZA, TAREK
+from test.utils.namespace import EGSCHEME
 
 import pytest
 
@@ -24,6 +26,9 @@
 
 HOST = "http://localhost:3030"
 DB = "/db/"
+DATA = """
+<http://example.org/record/1> a <http://xmlns.com/foaf/0.1/Document> .
+"""
 
 pluginstores = []
 
@@ -270,6 +275,25 @@ def test_graph_without_identifier() -> None:
     ) == ("genid", genid_prefix)
 
     assert f"{g1.identifier}".startswith(genid_prefix)
+    assert f"{subgraph.identifier}".startswith(genid_prefix)
+
+
+def test_parse_return_type():
+    g = Dataset()
+    g.parse(data=DATA, format="turtle")
+    assert type(g) is Dataset
+
+    g = Dataset()
+    g = g.parse(data=DATA, format="turtle")
+    assert type(g) is Dataset
+
+    g = Dataset()
+    g.parse(source=StringIO(DATA), format="turtle")
+    assert type(g) is Dataset
+
+    g = Dataset()
+    g = g.parse(source=StringIO(DATA), format="turtle")
+    assert type(g) is Dataset
 
     # now add a preexisting graph with no identifier
     # i.e. not one created within this Dataset object