diff --git a/manual/modules/lang/pages/quickref.adoc b/manual/modules/lang/pages/quickref.adoc
index 60dd80ed..aae37548 100644
--- a/manual/modules/lang/pages/quickref.adoc
+++ b/manual/modules/lang/pages/quickref.adoc
@@ -14,7 +14,7 @@ print.
 
 xref:sugar.adoc#unicode[Unicode escapes]: ::
 
-`\x{3C4}` is exactly equivalent to character U+03C4.
+`+\x{3C4}+` is exactly equivalent to character U+03C4.
 
 xref:execution.adoc#predicates[Rule definitions]: ::
 
diff --git a/manual/modules/lang/pages/sugar.adoc b/manual/modules/lang/pages/sugar.adoc
index 01dc0fcc..0f649f09 100644
--- a/manual/modules/lang/pages/sugar.adoc
+++ b/manual/modules/lang/pages/sugar.adoc
@@ -403,8 +403,8 @@ Slash-expressions may only appear in rule heads, never inside rule bodies.
 Dialog can support just about any characters in source code, whether that's A, Ÿ,
 Σ, 五, or ✨. However, these characters are not always easy to type. Sometimes it's
 more convenient to refer to characters by Unicode codepoint; when this happens,
-you can instead refer to those characters as `\x{41}`, `\x{178}`, `\x{3A3}`,
-`\x{4e94}`, and `\x{2728}`. The two are exactly equivalent to the compiler, which
+you can instead refer to those characters as `+\x{41}+`, `+\x{178}+`, `+\x{3A3}+`,
+`+\x{4e94}+`, and `+\x{2728}+`. The two are exactly equivalent to the compiler, which
 allows for some truly awful source code:
 
 [source]
@@ -413,8 +413,8 @@ allows for some truly awful source code:
 ----
 
 But this is far less readable than `(program entry point)` and should be avoided
-at any cost. Still, `\x{200D}` is much clearer than a literal zero-width joiner,
-and `\x{301}` much clearer than a literal combining acute accent. Sometimes the
+at any cost. Still, `+\x{200D}+` is much clearer than a literal zero-width joiner,
+and `+\x{301}+` much clearer than a literal combining acute accent. Sometimes the
 difference between codepoints can be easier to read than the difference between
 the actual characters.