diff --git a/NEWS.rst b/NEWS.rst
index 2c86998..383b01f 100644
--- a/NEWS.rst
+++ b/NEWS.rst
@@ -12,6 +12,7 @@ Release date: *not yet released*
 * Add :py:meth:`ConnectionPool.close` and support using the pool as a
   context manager
   (`issue 229 <https://github.com/python-happybase/happybase/issues/229>`_)
+* Add :py:meth:`Table.append`
 
 
 HappyBase 1.3.0
diff --git a/happybase/table.py b/happybase/table.py
index f1ea350..cfb6fd5 100644
--- a/happybase/table.py
+++ b/happybase/table.py
@@ -8,7 +8,7 @@
 
 from six import iteritems
 
-from Hbase_thrift import TScan
+from Hbase_thrift import TAppend, TScan
 
 from .util import thrift_type_to_dict, bytes_increment, OrderedDict
 from .batch import Batch
@@ -494,6 +494,36 @@ def delete(self, row, columns=None, timestamp=None, wal=True):
         with self.batch(timestamp=timestamp, wal=wal) as batch:
             batch.delete(row, columns)
 
+    def append(self, row, data, include_timestamp=False):
+        """Atomically append values to one or more columns in a row.
+
+        The `data` argument is a dictionary that maps columns to the values
+        to append. If a column did not exist, it is created with the given
+        value.
+
+        :param str row: the row key
+        :param dict data: the values to append
+        :param bool include_timestamp: whether timestamps are returned
+
+        :return: mapping of columns to their values after appending
+        :rtype: dict
+        """
+        # The server returns the cells without their names, sorted by
+        # family and qualifier, so send the columns in that same order.
+        columns = sorted(
+            data, key=lambda c: c.split(b':' if isinstance(c, bytes) else ':'))
+        cells = self.connection.client.append(TAppend(
+            table=self.name,
+            row=row,
+            columns=columns,
+            values=[data[column] for column in columns],
+        ))
+        return {
+            column: (cell.value, cell.timestamp) if include_timestamp
+            else cell.value
+            for column, cell in zip(columns, cells)
+        }
+
     def batch(self, timestamp=None, batch_size=None, transaction=False,
               wal=True):
         """Create a new batch operation for this table.
diff --git a/tests/test_api.py b/tests/test_api.py
index 9e1f797..484aab5 100644
--- a/tests/test_api.py
+++ b/tests/test_api.py
@@ -175,6 +175,20 @@ def test_atomic_counters():
     assert 10 == table.counter_dec(row, column, -7)
 
 
+def test_append():
+    row = b'row-with-append'
+
+    assert {b'cf1:c1': b'a'} == table.append(row, {b'cf1:c1': b'a'})
+    assert {b'cf1:c1': b'ab', b'cf2:c2': b'x'} == table.append(
+        row, {b'cf2:c2': b'x', b'cf1:c1': b'b'})
+    assert {b'cf1:c1': b'ab', b'cf2:c2': b'x'} == table.row(row)
+
+    value, timestamp = table.append(
+        row, {b'cf1:c1': b'c'}, include_timestamp=True)[b'cf1:c1']
+    assert b'abc' == value
+    assert isinstance(timestamp, int)
+
+
 def test_batch():
     with pytest.raises(TypeError):
         table.batch(timestamp='invalid')