diff --git a/src/calibre/ebooks/mobi/debug/index.py b/src/calibre/ebooks/mobi/debug/index.py
index 861ba16bee..3b03431628 100644
--- a/src/calibre/ebooks/mobi/debug/index.py
+++ b/src/calibre/ebooks/mobi/debug/index.py
@@ -25,35 +25,36 @@ Elem = namedtuple('Chunk',
 
 GuideRef = namedtuple('GuideRef', 'type title pos_fid')
 
-INDEX_HEADER_FIELDS = INDEX_HEADER_FIELDS + ('last_index', 'tagx_block_size', 'tagx_block')
+INDEX_HEADER_FIELDS = INDEX_HEADER_FIELDS + ('indices', 'tagx_block_size', 'tagx_block')
 FIELD_NAMES = {'len':'Header length', 'type':'Unknown', 'gen':'Index Type (0 - normal, 2 - inflection)',
-               'start':'IDXT Offset', 'count':'Number of Index entries or records', 'code': 'character encoding', 'lng':'Unknown',
-               'total':'Total number of Index Entries in all records', 'ordt': 'ORDT Offset', 'ligt':'LIGT Offset', 'nligt':'Number of LIGT',
-               'ncncx':'Number of CNCX records', 'last_index':'Geometry of index records'}
+               'start':'IDXT Offset', 'count':'Number of entries in this record', 'code': 'character encoding', 'lng':'Unknown',
+               'total':'Total number of actual Index Entries in all records', 'ordt': 'ORDT Offset', 'ligt':'LIGT Offset', 'nligt':'Number of LIGT',
+               'ncncx':'Number of CNCX records', 'indices':'Geometry of index records'}
 
 def read_variable_len_data(data, header):
     offset = header['tagx']
     indices = []
+    idxt_offset = header['start']
+    idxt_size = 4 + header['count'] * 2
     if offset > 0:
         tagx_block_size = header['tagx_block_size'] = struct.unpack_from(b'>I', data, offset + 4)[0]
         header['tagx_block'] = data[offset:offset+tagx_block_size]
-        offset += tagx_block_size
+        offset = idxt_offset + 4
         for i in xrange(header['count']):
-            strlen = bytearray(data[offset:offset+1])[0]
-            text = data[offset+1:offset+1+strlen].decode('ascii')
-            offset += 1 + strlen
-            num = struct.unpack_from(b'>H', data, offset)[0]
+            p = struct.unpack_from(b'>H', data, offset)[0]
             offset += 2
+            strlen = bytearray(data[p])[0]
+            text = data[p+1:p+1+strlen]
+            p += 1 + strlen
+            num = struct.unpack_from(b'>H', data, p)[0]
             indices.append((text, num))
     else:
         header['tagx_block'] = b''
         header['tagx_block_size'] = 0
-    idxt_offset = header['start']
-    idxt_size = 4 + header['count'] * 2
     trailing_bytes = data[idxt_offset+idxt_size:]
     if trailing_bytes.rstrip(b'\0'):
         raise ValueError('Traling bytes after last IDXT entry: %r' % trailing_bytes.rstrip(b'\0'))
-    header['last_index'] = indices
+    header['indices'] = indices
 
 def read_index(sections, idx, codec):
     table, cncx = OrderedDict(), CNCX([], codec)