Here's a different strategy, based on using split() to separate a string into pieces using non-capitalized words as separators, then discarding empty parts. This scheme needs some tweaking to honor sentence boundaries, but the general trick is a useful one in situations like this.

#!/usr/bin/perl -w
use strict;

my $source = join('', <DATA>);
my @capgroups = grep { $_ } split(/(?:^|\s+)(?:[^A-Z]\S*\s*)+/, $sourc
+e);
foreach ( @capgroups ) {
    print "$_\n";
}

__DATA__
Douglas built five Douglas World Cruisers to attempt his first flight 
+to Buenos Aires.
These were the predecesors of the modern AH-64D and AH-64D Apache.
[download]

And how does the script differentiate between "These" or "Douglas" at the beginning of a sentence? One is a proper noun, one is just at the start of the sentence, but there is no way to clearly tell.

</ajdelore>

#!/usr/bin/perl -w
use strict;

$_ = <DATA>;
my @matches = /([A-Z].*? )[a-z].*?/g;
my $biggest = 0;
my $biggestCluster;
for my $m ( @matches ) {
    print "$m\n";
    $_ = $m;
    my @cs = /([A-Z].*? )/g;
    if ( $#cs > $biggest ) {
        $biggestCluster = $m;
        $biggest = $#cs;
    }
}
print "Biggest Cluster is [$biggestCluster]\n";
__DATA__
Douglas built five Douglas World Cruisers to attempt his first flight 
+to Buenos Aires. There were the predecessor of the modern AH-64D and 
+AH-64D Apache.
[download]

#/usr/bin/perl

$sentence = "Douglas built five Douglas World Cruisers to attempt his 
+first flight to Buenos Aires. These were the predecesors of the moder
+n AH-64D and AH-64D Apache.";

# split on spaces and some common punctuations
@words = split( /\s+|[.,\/\\;]/, $sentence );

# roll through the words and grab the cap words
# and their position 
$cnt = 0;
foreach( @words ) {
        push( @lol, [ $_, $cnt ] ) if /^[A-Z]/;
        $cnt++;
}

# print out the first match
$prev = $lol[0]->[1]-1;
print $lol[0]->[0];
shift( @lol );

# roll through the rest. print out a new line
# if it's not the next word by count, space if it is
foreach( @lol ) {
        print $prev != $_->[1]-1 ? "\n" : " ";
        print $_->[0];
        $prev = $_->[1];
}
print "\n";
[download]

Like I said ... ugly.

-derby

Abigail

A slightly more straight-forward approach (IMO). It picks up 'These,' unless you uncomment the first line in the regexp (but then you run the risk of missing proper names at the begining of sentences; you still need to decide how you want to handle that).

#!/usr/bin/perl -l
use warnings;
use strict;

my $data = 'Douglas built five Douglas World Cruisers to attempt his f
+irst flight to Buenos Aires. These were the predecesors of the modern
+ AH-64D and AH-64D Apache.';

# What do you call a capitalized "word"?
my $cap_word = qr/[A-Z][\w-]*/;

my @clusters = $data =~ /
  #(?<!\.\s)           # Ignore words at begining of sentences?
  (
    $cap_word          # Capitalized word, followed by any number
    (?:\s+$cap_word)*  # of other cap words (separated by spaces)
  )
/gx;

# Update: Oh, you wanted the largest...
print "Largest cluster: ", (sort { length $b <=> length $a } @clusters
+)[0];
[download]

bbfu
Black flowers blossom
Fearless on my breath