diff --git a/pandas/core/frame.py b/pandas/core/frame.py
index f5681142a0048..cb3bd6d7ef183 100644
--- a/pandas/core/frame.py
+++ b/pandas/core/frame.py
@@ -4818,6 +4818,8 @@ def apply(self, func, axis=0, broadcast=False, raw=False, reduce=None,
             while guessing, exceptions raised by func will be ignored). If
             reduce is True a Series will always be returned, and if False a
             DataFrame will always be returned.
+
+            .. deprecated:: 0.22.0
         args : tuple
             Positional arguments to pass to function in addition to the
             array/series
@@ -4850,6 +4852,10 @@ def apply(self, func, axis=0, broadcast=False, raw=False, reduce=None,
         axis = self._get_axis_number(axis)
         ignore_failures = kwds.pop('ignore_failures', False)
 
+        if reduce is not None:
+            warnings.warn("the reduce keyword is deprecated",
+                          FutureWarning, stacklevel=7)
+
         # dispatch to agg
         if axis == 0 and isinstance(func, (list, dict)):
             return self.aggregate(func, axis=axis, *args, **kwds)
@@ -5760,8 +5766,10 @@ def f(x):
                         # numeric_only and yet we have tried a
                         # column-by-column reduction, where we have mixed type.
                         # So let's just do what we can
-                        result = self.apply(f, reduce=False,
-                                            ignore_failures=True)
+                        result = self._apply_standard(f,
+                                                      axis=axis,
+                                                      reduce=False,
+                                                      ignore_failures=True)
                         if result.ndim == self.ndim:
                             result = result.iloc[0]
                         return result
diff --git a/pandas/tests/frame/test_apply.py b/pandas/tests/frame/test_apply.py
index 55f04bf189f0c..4867496123f99 100644
--- a/pandas/tests/frame/test_apply.py
+++ b/pandas/tests/frame/test_apply.py
@@ -82,22 +82,24 @@ def test_apply_empty(self):
         assert_frame_equal(xp, rs)
 
         # reduce with an empty DataFrame
-        x = []
-        result = self.empty.apply(x.append, axis=1, reduce=False)
-        assert_frame_equal(result, self.empty)
-        result = self.empty.apply(x.append, axis=1, reduce=True)
-        assert_series_equal(result, Series(
-            [], index=pd.Index([], dtype=object)))
-
-        empty_with_cols = DataFrame(columns=['a', 'b', 'c'])
-        result = empty_with_cols.apply(x.append, axis=1, reduce=False)
-        assert_frame_equal(result, empty_with_cols)
-        result = empty_with_cols.apply(x.append, axis=1, reduce=True)
-        assert_series_equal(result, Series(
-            [], index=pd.Index([], dtype=object)))
-
-        # Ensure that x.append hasn't been called
-        assert x == []
+        with tm.assert_produces_warning(FutureWarning,
+                                        check_stacklevel=False):
+            x = []
+            result = self.empty.apply(x.append, axis=1, reduce=False)
+            assert_frame_equal(result, self.empty)
+            result = self.empty.apply(x.append, axis=1, reduce=True)
+            assert_series_equal(result, Series(
+                [], index=pd.Index([], dtype=object)))
+
+            empty_with_cols = DataFrame(columns=['a', 'b', 'c'])
+            result = empty_with_cols.apply(x.append, axis=1, reduce=False)
+            assert_frame_equal(result, empty_with_cols)
+            result = empty_with_cols.apply(x.append, axis=1, reduce=True)
+            assert_series_equal(result, Series(
+                [], index=pd.Index([], dtype=object)))
+
+            # Ensure that x.append hasn't been called
+            assert x == []
 
     def test_apply_standard_nonunique(self):
         df = DataFrame(
@@ -371,14 +373,16 @@ def test_apply_dict(self):
         B_dicts = pd.Series([dict([(0, 0), (1, 2)]), dict([(0, 1), (1, 3)])])
         fn = lambda x: x.to_dict()
 
-        for df, dicts in [(A, A_dicts), (B, B_dicts)]:
-            reduce_true = df.apply(fn, reduce=True)
-            reduce_false = df.apply(fn, reduce=False)
-            reduce_none = df.apply(fn, reduce=None)
+        with tm.assert_produces_warning(FutureWarning,
+                                        check_stacklevel=False):
+            for df, dicts in [(A, A_dicts), (B, B_dicts)]:
+                reduce_true = df.apply(fn, reduce=True)
+                reduce_false = df.apply(fn, reduce=False)
+                reduce_none = df.apply(fn, reduce=None)
 
-            assert_series_equal(reduce_true, dicts)
-            assert_frame_equal(reduce_false, df)
-            assert_series_equal(reduce_none, dicts)
+                assert_series_equal(reduce_true, dicts)
+                assert_frame_equal(reduce_false, df)
+                assert_series_equal(reduce_none, dicts)
 
     def test_applymap(self):
         applied = self.frame.applymap(lambda x: x * 2)