foreach (@array) {
    $hash{$_}++;
}
foreach (keys %hash) {
    delete $hash{$_} if ($hash{$_} > 1);
}
@array2 = keys %hash;
[download]

HTH

More compactly as,

$hash{$_}++ foreach (@array);
@array2 = grep( ($hash{$_} == 1), keys %hash );
[download]

I came up with this, at 46 chars:

@array=do{grep$_{$_}<2,grep{!$_{$_}++}@array};
[download]

Of course, if we were truely golfing, that comes out at:

sub find_duplicates {

#23456789_123456789_123456789_
grep$_{$_}<2,grep{!$_{$_}++}@_

}
[download]

30 characters :)

From RE: Return a Unique Array by sergio via Perl Cookbook:

@array = grep { ! $seen{$_}++ } @array ;
[download]

--Jim

Update:

Oops! Sorry, this doesn't eliminate the duplicated values.

#!/usr/bin/perl

use strict;

my @input = qw(a b c d e d e f);

# %seen is a hash used to count the number of times
# an element appears in the array.  @result contains
# the results of removing the duplicates.
my (%seen, @result);

foreach (@input)
{
        $seen{$_}++;
}

foreach (@input)
{
        # Snag the elements which appear only once.
        # Use the original elements of the input array
        # as hash keys rather than keys(%seen) to
        # preserve ordering
        if ($seen{$_} == 1)
        {
                push @result, $_;
        }
}
[download]

-Matt