3 files changed, 7 insertions, 6 deletions
diff --git a/share/doc/papers/memfs/1.t b/share/doc/papers/memfs/1.t
index a065844..e079be9 100644
--- a/share/doc/papers/memfs/1.t
+++ b/share/doc/papers/memfs/1.t
@@ -33,7 +33,7 @@
 .\"
 .nr PS 11
 .nr VS 13
-.SH
+.NH
 Introduction
 .PP
 This paper describes the motivation for and implementation of
@@ -150,7 +150,7 @@ We await the results of others trying this approach.
 .[
 Ohta
 .]
-.SH
+.NH
 Implementation
 .PP
 The current implementation took less time to write than did this paper.
@@ -268,7 +268,7 @@ but have not been recently accessed (or have been deleted).
 .[
 leffler
 .]
-.SH
+.NH
 Performance
 .PP
 The performance of the current memory-based filesystem is determined by
@@ -313,7 +313,7 @@ Here, the low latency of the memory-based filesystem is
 noticeable compared to the disk-based filesystem,
 as a synchronous operation can be done with
 just two context switches instead of incurring the disk latency.
-.SH
+.NH
 Future Work
 .PP
 The most obvious shortcoming of the current implementation
diff --git a/share/doc/papers/memfs/A.t b/share/doc/papers/memfs/A.t
index c1938c8..09d8e93 100644
--- a/share/doc/papers/memfs/A.t
+++ b/share/doc/papers/memfs/A.t
@@ -34,7 +34,7 @@
 .bp
 .nr PS 10
 .nr VS 12
-.SH
+.NH
 Appendix A - Implementation Details
 .LP
 .nf
diff --git a/share/doc/papers/memfs/tmac.srefs b/share/doc/papers/memfs/tmac.srefs
index 6245118..96199b8 100644
--- a/share/doc/papers/memfs/tmac.srefs
+++ b/share/doc/papers/memfs/tmac.srefs
@@ -142,7 +142,8 @@ in \\f2\\*([B\\f1\c
 .\}
 .el .sp 27p
 .Li 2 30.5P
-\fBReferences\fP
+.NH
+References
 .br
 .if \\n(Ns<2 \{\
 .	nr Ns 1