I'm trying to use cURL to download the contents of webpages that require forms to get to.

In a browser it looks something like this

1.) Login using POST
2.) Pick which page to go to using another POST
3.) Pick another page... using POST
4.) etc.. until I get to the page I want, 
    then download all textfiles linked to on that page.

I am attempting to do this using a bash script and some loops with the values that change for each POST.

My problem is how do I do multiple POSTs with cURL? Must there be cookies involved?

FYI the website is http://metagenomics.nmpdr.org/ (MG-RAST).