clarification dictionary format
diff --git a/zstd_compression_format.md b/zstd_compression_format.md
index d7f2193..9d27f6c 100644
--- a/zstd_compression_format.md
+++ b/zstd_compression_format.md
@@ -1156,6 +1156,7 @@
               FSE table for match lengths, and FSE table for literals lengths.
               It's finally followed by 3 offset values, populating recent offsets,
               stored in order, 4-bytes little-endian each, for a total of 12 bytes.
+              Each recent offset must have a value < dictionary size.
 
 __`Content`__ : The rest of the dictionary is its content.
               The content act as a "past" in front of data to compress or decompress.