My first reaction was to split and grep, but since the numbers are so tightly defined, I think a regex should do,

my @keepers;
open my $fh, '<', '/path/to/file.dat' or die $!;
while (<$fh>) {
    while ( /(\d+\.\d+)/g ) {
        push @keepers, $_ and last
            if $1 >= 296.1 && $1 <= 314.0;
    }
}
close $fh or die $!;
[download]

The while (/(foo)/g) construction loops over all the matches, putting the matching text in $1.

Update: Missed the exclusion for 305.1, grep it is:

my @keepers;
open my $fh, '<', '/path/to/file.dat' or die $!;
while (<$fh>) {
    my @data = grep {
        /^\d+\.\d+$/ &&  $_ >= 296.1 && $_ <= 314.0
    } split ',';
    if (@data) {
        push @keepers, $_ 
            unless @data == 1 and $data[0] eq '305.1';
    }
}
close $fh or die $!;
[download]

Update2: CountZero++ is correct, code amended.