diff --git a/anki/find.py b/anki/find.py
index 0202488b4..be2dab02f 100644
--- a/anki/find.py
+++ b/anki/find.py
@@ -159,8 +159,9 @@ and c.nid=n.id %s""" % (q, order)
         extra = "not" if neg else ""
         if not self.full:
             self.lims['args']["_text_%d"%c] = "%"+val+"%"
-            self.lims['preds'].append("flds %s like :_text_%d escape '\\'" % (
-                extra, c))
+            self.lims['preds'].append("""\
+(sfld %s like :_text_%d escape '\\' or
+flds %s like :_text_%d escape '\\')""" % (extra, c, extra, c))
         else:
             # in the future we may want to apply this at the end to speed up
             # the case where there are other limits
diff --git a/tests/test_find.py b/tests/test_find.py
index 98710c08f..161666499 100644
--- a/tests/test_find.py
+++ b/tests/test_find.py
@@ -111,12 +111,18 @@ def test_findCards():
     # full search
     f = deck.newNote()
     f['Front'] = u'hello<b>world</b>'
-    f['Back'] = u''
+    f['Back'] = u'abc'
     deck.addNote(f)
+    # as it's the sort field, it matches
+    assert len(deck.findCards("helloworld")) == 2
+    assert len(deck.findCards("helloworld", full=True)) == 2
+    # if we put it on the back, it won't
+    (f['Front'], f['Back']) = (f['Back'], f['Front'])
+    f.flush()
     assert len(deck.findCards("helloworld")) == 0
-    assert len(deck.findCards("helloworld", full=True)) == 1
+    assert len(deck.findCards("helloworld", full=True)) == 2
     assert len(deck.findCards("front:helloworld")) == 0
-    assert len(deck.findCards("front:helloworld", full=True)) == 1
+    assert len(deck.findCards("back:helloworld", full=True)) == 2
     # searching for an invalid special tag should not error
     assert len(deck.findCards("is:invalid")) == 0