diff --git a/pandas/core/dtypes/missing.py b/pandas/core/dtypes/missing.py
index af3a873bc2866..9913923cb7807 100644
--- a/pandas/core/dtypes/missing.py
+++ b/pandas/core/dtypes/missing.py
@@ -394,3 +394,10 @@ def na_value_for_dtype(dtype):
     elif is_bool_dtype(dtype):
         return False
     return np.nan
+
+
+def remove_na_arraylike(arr):
+    """
+    Return array-like containing only true/non-NaN values, possibly empty.
+    """
+    return arr[notnull(lib.values_from_object(arr))]
diff --git a/pandas/core/series.py b/pandas/core/series.py
index e1f668dd3afda..98b548f8ab3b5 100644
--- a/pandas/core/series.py
+++ b/pandas/core/series.py
@@ -36,7 +36,7 @@
     maybe_upcast, infer_dtype_from_scalar,
     maybe_convert_platform,
     maybe_cast_to_datetime, maybe_castable)
-from pandas.core.dtypes.missing import isnull, notnull
+from pandas.core.dtypes.missing import isnull, notnull, remove_na_arraylike
 
 from pandas.core.common import (is_bool_indexer,
                                 _default_index,
@@ -2749,7 +2749,7 @@ def dropna(self, axis=0, inplace=False, **kwargs):
         axis = self._get_axis_number(axis or 0)
 
         if self._can_hold_na:
-            result = remove_na(self)
+            result = remove_na_arraylike(self)
             if inplace:
                 self._update_inplace(result)
             else:
@@ -2888,13 +2888,6 @@ def _dir_additions(self):
 # Supplementary functions
 
 
-def remove_na(series):
-    """
-    Return series containing only true/non-NaN values, possibly empty.
-    """
-    return series[notnull(_values_from_object(series))]
-
-
 def _sanitize_index(data, index, copy=False):
     """ sanitize an index type to return an ndarray of the underlying, pass
     thru a non-Index
diff --git a/pandas/plotting/_core.py b/pandas/plotting/_core.py
index f8e83aea03594..9cceebb5c4cdb 100644
--- a/pandas/plotting/_core.py
+++ b/pandas/plotting/_core.py
@@ -11,7 +11,7 @@
 
 from pandas.util._decorators import cache_readonly
 from pandas.core.base import PandasObject
-from pandas.core.dtypes.missing import notnull
+from pandas.core.dtypes.missing import notnull, remove_na_arraylike
 from pandas.core.dtypes.common import (
     is_list_like,
     is_integer,
@@ -21,7 +21,7 @@
 from pandas.core.common import AbstractMethodError, isnull, _try_sort
 from pandas.core.generic import _shared_docs, _shared_doc_kwargs
 from pandas.core.index import Index, MultiIndex
-from pandas.core.series import Series, remove_na
+from pandas.core.series import Series
 from pandas.core.indexes.period import PeriodIndex
 from pandas.compat import range, lrange, map, zip, string_types
 import pandas.compat as compat
@@ -1376,7 +1376,7 @@ def _plot(cls, ax, y, style=None, bw_method=None, ind=None,
         from scipy.stats import gaussian_kde
         from scipy import __version__ as spv
 
-        y = remove_na(y)
+        y = remove_na_arraylike(y)
 
         if LooseVersion(spv) >= '0.11.0':
             gkde = gaussian_kde(y, bw_method=bw_method)
@@ -1495,13 +1495,13 @@ def _args_adjust(self):
     @classmethod
     def _plot(cls, ax, y, column_num=None, return_type='axes', **kwds):
         if y.ndim == 2:
-            y = [remove_na(v) for v in y]
+            y = [remove_na_arraylike(v) for v in y]
             # Boxplot fails with empty arrays, so need to add a NaN
             #   if any cols are empty
             # GH 8181
             y = [v if v.size > 0 else np.array([np.nan]) for v in y]
         else:
-            y = remove_na(y)
+            y = remove_na_arraylike(y)
         bp = ax.boxplot(y, **kwds)
 
         if return_type == 'dict':
@@ -1969,7 +1969,7 @@ def maybe_color_bp(bp):
 
     def plot_group(keys, values, ax):
         keys = [pprint_thing(x) for x in keys]
-        values = [remove_na(v) for v in values]
+        values = [remove_na_arraylike(v) for v in values]
         bp = ax.boxplot(values, **kwds)
         if fontsize is not None:
             ax.tick_params(axis='both', labelsize=fontsize)
diff --git a/pandas/tests/test_panel.py b/pandas/tests/test_panel.py
index e19e42e062932..445611c1696f5 100644
--- a/pandas/tests/test_panel.py
+++ b/pandas/tests/test_panel.py
@@ -10,11 +10,11 @@
 import pandas as pd
 
 from pandas.core.dtypes.common import is_float_dtype
+from pandas.core.dtypes.missing import remove_na_arraylike
 from pandas import (Series, DataFrame, Index, date_range, isnull, notnull,
                     pivot, MultiIndex)
 from pandas.core.nanops import nanall, nanany
 from pandas.core.panel import Panel
-from pandas.core.series import remove_na
 
 from pandas.io.formats.printing import pprint_thing
 from pandas import compat
@@ -155,7 +155,7 @@ def _check_stat_op(self, name, alternative, obj=None, has_skipna=True):
         if has_skipna:
 
             def skipna_wrapper(x):
-                nona = remove_na(x)
+                nona = remove_na_arraylike(x)
                 if len(nona) == 0:
                     return np.nan
                 return alternative(nona)
diff --git a/pandas/tests/test_panel4d.py b/pandas/tests/test_panel4d.py
index e1995316e7b7c..18643aff15e9b 100644
--- a/pandas/tests/test_panel4d.py
+++ b/pandas/tests/test_panel4d.py
@@ -7,10 +7,10 @@
 import numpy as np
 
 from pandas.core.dtypes.common import is_float_dtype
+from pandas.core.dtypes.missing import remove_na_arraylike
 from pandas import Series, Index, isnull, notnull
 from pandas.core.panel import Panel
 from pandas.core.panel4d import Panel4D
-from pandas.core.series import remove_na
 from pandas.tseries.offsets import BDay
 
 from pandas.util.testing import (assert_frame_equal, assert_series_equal,
@@ -118,7 +118,7 @@ def _check_stat_op(self, name, alternative, obj=None, has_skipna=True):
 
         if has_skipna:
             def skipna_wrapper(x):
-                nona = remove_na(x)
+                nona = remove_na_arraylike(x)
                 if len(nona) == 0:
                     return np.nan
                 return alternative(nona)