diff --git a/doc/Language/numerics.pod6 b/doc/Language/numerics.pod6
index de25e9e10..1d0253fee 100644
--- a/doc/Language/numerics.pod6
+++ b/doc/Language/numerics.pod6
@@ -210,8 +210,20 @@ say ½ + ⅓ + ⅝ + ⅙; # OUTPUT: «1.625␤»
 If a I<mathematical operation> that produces a L<Rat|/type/Rat> answer would produce
 a L<Rat|/type/Rat> with denominator larger than 64 bits, that operation would instead
 return a L<Num|/type/Num> object. When I<constructing> a L<Rat|/type/Rat> (i.e. when it is not
-a result of some mathematical expression), however, a "fattish" L<Rat|/type/Rat> with
-larger denominator will be created.
+a result of some mathematical expression), however, a larger denominator can be used:
+
+    my $a = 1 / (2⁶⁴ - 1);
+    say $a;                   # OUTPUT: «0.000000000000000000054␤»
+    say $a.^name;             # OUTPUT: «Rat␤»
+
+    my $b = 1 / 2⁶⁴;
+    say $b;                   # OUTPUT: «5.421010862427522e-20␤»
+    say $b.^name;             # OUTPUT: «Num␤»
+
+    my $c = Rat.new(1, 2⁶⁴);
+    say $c;                   # OUTPUT: «0.000000000000000000054␤»
+    say $c.^name;             # OUTPUT: «Rat␤»
+    say $c.Num;               # OUTPUT: «5.421010862427522e-20␤»
 
 =head2 C<FatRat>