Remove detritus from MidRatStr reference

When allomorphs were first documented in dcd83b, `MidRatStr` still
existed as a three-superclass allomorph. That's no longer the case, so
remove language referring to hypothetical three-class allomorphs.

But leave open the possibility so this passage isn't brittle against
future allomorphs.

Author: treyharris

doc/Language/numerics.pod6

@@ -401,7 +401,7 @@ Note: there is no C<FatRatStr> type.
 
 =head2 Coercion of allomorphs
 
-Keep in mind that allomorphs are simply subclasses of the two (or three) types they represent. Just
+Keep in mind that allomorphs are simply subclasses of the types they represent. Just
 as a variable or parameter type-constrained to C<Foo> can accept any subclass of C<Foo>, so will
 a variable or parameter type-constrained to L<Int|/type/Int> will accept an L<IntStr|/type/IntStr> allomorph: