diff --git a/SAMtags.tex b/SAMtags.tex
index e19ec290b..37287f1b0 100644
--- a/SAMtags.tex
+++ b/SAMtags.tex
@@ -92,6 +92,7 @@ \section{Standard tags}
   {\tt MI} & Z & Molecular identifier; a string that uniquely identifies the molecule from which the record was derived \\
   {\tt ML} & B,C & Base modification probabilities \\
   {\tt MM} & Z & Base modifications / methylation  \\
+  {\tt MN} & i & Length of sequence at the time {\tt MM} and {\tt ML} were produced \\
   {\tt MQ} & i & Mapping quality of the mate/next segment \\
   {\tt NH} & i & Number of reported alignments that contain the query in the current record \\
   {\tt NM} & i & Edit distance to the reference \\
@@ -621,6 +622,16 @@ \subsection{Base modifications}
 {\tt ML} values for ambiguity codes give the probability that the modification is one of the possible codes compatible with that ambiguity code.
 For example {\tt MM:Z:C+C,10; ML:B:C,229} indicates a C call with a probability of 90\% of having some form of unspecified modification.
 
+\item[MN:i:\tagvalue{length}]
+\hfill\\
+Tools may edit the {\sf SEQ} sequence data, such as modifying the alignment with hard-clipping.
+If the sequence is shrunk in this manner then the base offsets in {\tt MM} and {\tt ML} become invalid unless they are also updated accordingly.
+
+There may be hard-clipping tools which update {\tt MM} and tools which do not, so the {\tt MN} tag offers a simple sanity check.
+It holds the length of the sequence at the time {\tt MM} was last written.
+Tools that wish to validate {\tt MM} should compare the length of the {\sf SEQ} field with the contents of the {\tt MN} tag.
+The tag is optional, but recommended, and if it is absent then there is an implicit assumption that the {\tt MM} data is valid unless evidence implies otherwise (such as having coordinates beyond the end of the sequence).
+
 \end{description}
 
 \section{Draft tags}