I agree with tobyink, except that recreating the arrays when entering the second routine is not really optimal:

sub secondRoutine {
    my @array1 = @{ shift(@_) };
    my @array2 = @{ shift(@_) };
    my @array3 = @{ shift(@_) };

    print ("Test first value in array1: $array1[0]");
    print ("Test third value in array2: $array2[2]");
    print ("Test fifth value in array3: $array3[4]");
}
[download]

I would usually prefer something like this:

<c>sub secondRoutine {
    my ($array1_ref, $array2_ref, $array3_ref) = @_;

    print ("Test first value in array1: $$array1_ref[0]");
    print ("Test third value in array2: $$array2_ref[2]");
    print ("Test fifth value in array3: $$array3_ref[4]");
}
[download]

This is matter open to discussion. Sometimes, I also create intermediary arrays from arrayrefs or hashref to make sure that I am really understanding my complicated data structure, but, here, it seems quite simple, no point of using memory to copy data structures.

Update: fixed the copy-and-paste error seen by Monk::Thomas.