Have you tried it?

Without even trying to understand what was required, a lot is already possible with DBD::CSV.

$ cat pm1072615.csv
Month,Zone,Replicate,SpeciesA,SpeciesB,SpeciesC
Sept,1,1,5,10,15
Sept,1,2,0,5,10
Sept,1,3,5,0,5
Sept,2,1,5,5,5
Sept,2,2,10,15,10
Sept,2,3,0,0,5
$ cat pm1072615.pl
#!/usr/bin/perl

use 5.14.2;
use warnings;

use DBI;

my $dbh = DBI->connect ("dbi:CSV:", undef, undef, {
    f_ext       => ".csv/r",
    });

my $sth = $dbh->prepare (qq;
    select month, zone, replicate, avg (speciesa)
    from   pm1072615
    group  by month,zone, replicate;
    );
$sth->execute;
while (my $row = $sth->fetch) {
    printf "%-5s %3d %3d %8.5f\n", @$row;
    }
$ perl pm1072615.pl
Sept    2   1  5.00000
Sept    2   2 10.00000
Sept    1   1  5.00000
Sept    1   3  5.00000
Use of uninitialized value in printf at pm1072615.pl line 19.
Sept    1   2  0.00000
Use of uninitialized value in printf at pm1072615.pl line 19.
Sept    2   3  0.00000
$
[download]


Enjoy, Have FUN! H.Merijn

If you have postgresql on hand, then it's as easy as :

create table data(
month text,
zone integer,
replicate integer,
speciesa integer,
speciesb integer,
speciesc integer);

copy data from stdin with delimiter ',';

sept,1,1,5,10,15
Sept,1,2,0,5,10
Sept,1,3,5,0,5
Sept,2,1,5,5,5
Sept,2,2,10,15,10
Sept,2,3,0,0,5
\.

select month, zone, replicate, avg(speciesa), stddev(speciesa) from da
+ta group by month,zone, replicate;


 month | zone | replicate |          avg           | stddev 
-------+------+-----------+------------------------+--------
 Sept  |    1 |         2 | 0.00000000000000000000 |       
 Sept  |    2 |         3 | 0.00000000000000000000 |       
 Sept  |    1 |         3 |     5.0000000000000000 |       
 Sept  |    2 |         2 |    10.0000000000000000 |       
 sept  |    1 |         1 |     5.0000000000000000 |       
 Sept  |    2 |         1 |     5.0000000000000000 |       
(6 lignes)
[download]