Thanks for your replies everyone.

We weren't sure if Google would look at JS removing the page navigation as cloaking or not, so that's still a bit of a concern. We were reading Rand's post from 2008 on the subject http://www.seomoz.org/blog/white-hat-cloaking-it-exists-its-permitted-its-useful and Matt Cutts' replies on the subject.  We know it was a few years ago, but he still seemed to be saying to be over-cautious with that kind of thing.

Should we be worried about cloaking if we use JS to "hide" the page nav?

My company is looking at replacing our ecommerce site's paginated browsing with a Javascript infinite scroll function for when customers view internal search results--and possibly when they browse product categories also. Because our internal linking structure isn't very robust, I'm concerned that removing the pagination will make it harder to get the individual product pages to rank in the SERPs.

We have over 5,000 products, and most of them are internally linked to from the browsing results pages in the category structure: e.g. Blue Widgets, Widgets Under $250, etc.

I'm not too worried about removing pagination from the internal search results pages, but I'm concerned that doing the same for these category pages will result in de-linking the thousands of product pages that show up later in the browsing results and therefore won't be crawlable as internal links by the Googlebot.

Does anyone have any ideas on what to do here? I'm already arguing against the infinite scroll, but we're a fairly design-driven company and any ammunition or alternatives would really help.

For example, would serving a different page to the Googlebot in this case be a dangerous form of cloaking? (If the only difference is the presence of the pagination links.) Or is there any way to make rel=next and rel=prev tags work with infinite scrolling?