not quite:

The idea is to create one age and one name for each subdirectory. This grouping can be done by a hash. However you should use the "grouping-criteria" as hash-key. In your case it is $File::Find::dir. The rest of your logic can remain unchanged.

use strict;
use warnings;

use File::Find; 
@ARGV = ('.') unless @ARGV; 
my %age;
my %name;

sub oldest { 
  return if ((defined $age{$File::Find::dir}) && ($age{$File::Find::di
+r} < (stat($_))[9])); 
  $age{$File::Find::dir} = (stat(_))[9]; 
  $name{$File::Find::dir} = $File::Find::name; 
} 

find(\&oldest, @ARGV); 
 
 foreach my $n (keys(%age))
 {
    print $name{$n}."  " . scalar(localtime($age{$n})) . "\n"; 
}
[download]

btw.: I'm unsure whether this is a homework-question or not ... if it is be sure that you totally understand the solution. I have seen many situations where a student couldn't explain some code supposedly written by himself. You don't want that happen to you! ;-)