diff --git a/tts/text/sequencing/english_symbols_v1.py b/tts/text/sequencing/english_symbols_v1.py
index df4753a..d064316 100644
--- a/tts/text/sequencing/english_symbols_v1.py
+++ b/tts/text/sequencing/english_symbols_v1.py
@@ -1,14 +1,23 @@
-"""Adapted from https://github.com/keithito/tacotron """
+"""
+Originally adapted from https://github.com/keithito/tacotron
 
-'''
-Defines the set of symbols used in text input to the model.
+This defines the set of symbols used in text input to the model.
+
+This is the set of character encodings that FakeYou and UberDuck models have
+been trained on. We freeze it here as "V1", because these may change in future
+notebooks.
+
+Original documentation: 
+    The default is a set of ASCII characters that works well for English or 
+    text that has been run through Unidecode. For other data, you can modify 
+    _characters. See TRAINING_DATA.md for details.
+"""
 
-The default is a set of ASCII characters that works well for English or text that has been run through Unidecode. 
-For other data, you can modify _characters. See TRAINING_DATA.md for details.
-'''
 from tts.text.phonetization import cmudict
 from typing import List
 
+# NB: DO NOT CHANGE THE ORDER OF SYMBOLS! 
+# Models are trained on and depend upon the order.
 _pad        = '_'
 _punctuation = '!\'(),.:;? ' #NB: Includes space
 _special = '-'
diff --git a/tts/text/sequencing/test_english_symbols_v1.py b/tts/text/sequencing/test_english_symbols_v1.py
index 00b4438..6c9e71c 100644
--- a/tts/text/sequencing/test_english_symbols_v1.py
+++ b/tts/text/sequencing/test_english_symbols_v1.py
@@ -11,13 +11,19 @@ class TestEnglishSymbols:
         assert output == []
 
     def test_arpabet_word(self):
+        # NB: Models are trained on these integer encodings. 
+        # If these tests break, the order of symbols may have changed. 
         output = arpabet_to_sequence('T EH1 S T')
         assert output == [133, 94, 131, 133]
 
     def test_grapheme_word(self):
+        # NB: Models are trained on these integer encodings. 
+        # If these tests break, the order of symbols may have changed. 
         output = symbols_to_sequence('test')
         assert output == [57, 42, 56, 57]
 
     def test_grapheme_sentence(self):
+        # NB: Models are trained on these integer encodings. 
+        # If these tests break, the order of symbols may have changed. 
         output = symbols_to_sequence('this is a test.')
         assert output == [57, 45, 46, 56, 11, 46, 56, 11, 38, 11, 57, 42, 56, 57, 7]
\ No newline at end of file