@a->... and %h->{...} are described in perldsc Data Struct Cookbook.

@a->[$foo] is correctly written $a[$foo], while %h->{$bar} is correctly written as $h{$bar}. No references involved at all.

-- Randal L. Schwartz, Perl hacker
Be sure to read my standard disclaimer if this is a reply.

But your example structures aren't anonymous; if they were they wouldn't have names e.g., a and h! An anonymous structure is declared like:

my %hash = ();
$hash{foo} = [qw/bar/];
[download]

Here foo in %hash points to a reference to an anonymous array containing bar.

Another common usage is in parameter passing:

sub some_func {
  my $args = shift; 
}
some_func( { foo => 'bar' } );
[download]

Here an anonymous hash is created to hold the function arguments, which is then referred to by the scalar $arg inside the subroutine.

I couldn't find the section you were referring to in perldsc. Can you be more specific?

Update: clarified second example; the scalar $arg contains a reference to the anonymous hash.