I just did an edit to your question, and realized that the link I put in there has the answers you were looking for. They are not definitive, because people use books and ebook readers differently form individual to individual, and the technology is new so we haven't collected all the data yet. However, from a blog entry on The Guardian (so take it for what it's worth)
I've only managed to find one report – on the Kindle (by The Cleantech Group) – but it backs up suggestions that so long as e-readers are used as book replacements rather than supplements, they soon start to pay back in carbon terms. The report states that a book uses up "approximately 7.46 kilograms of CO2 over its lifetime" and that the Kindle produces "roughly 168 kg" during its lifecycle, making it "a clear winner against the potential savings: 1,074 kg of CO2 if replacing three books a month for four years; and up to 26,098 kg of CO2 when used to the fullest capacity of the Kindle."
For a full report from The Cleantech Group, here is a link to a PDF document. The author of the blog goes on to say
Nevertheless, I'm part-way convinced. There are clear advantages to using e-readers in schools and academe. At home, I'm less sure – especially when you factor in side-issues such as the toxicity of the heavy metals used in ebook readers and their batteries. I also hesitate because the devices are so new we still know little about how they're used.
Here, I'm hoping an informal survey here might shed more light. So tell me: if you own an e-reader, how often do you use it? (Have you for instance topped off the 22.5 books The Cleantech Group require to break even with traditional books in carbon terms?)
So basically, if you put 23 books on the Kindle, you are ahead of the game according to this.
There are some other things to consider too. For instance
The current book ordering system encourages bookstores to order more books than they can reasonably sell. That increases the number of published copies and consequently the number of returns (and the amount of shipping costs).
Additionally
A MS candidate named Greg Kozak pitted textbooks against e-book devices [PDF] in 2003. He found that paper production, electricity of printing operations, and personal transportation were the main factors affecting the book footprint, while electricity was the main issue for e-readers; and that books were responsible for four times the greenhouse emissions as e-readers.
Self Publishing Review states
On Salon, there’s a highly technical summary of the amount of energy it takes to read on a computer vs. printing out an article to read it. The verdict: “The contrast is quite convincing, one-tenth of a pound of greenhouse gas emissions from reading the document on the computer, versus almost one-quarter of a pound of greenhouse gas emissions for printing it out!”
Or more specifically to the Kindle
Another article lays out the environmental impact of reading a newspaper on a reader and reading the New York Times on an ebook. Their conclusion: “Reading the physical version of the NY Times for a year uses 7,300 MJ of energy and emits 700 kg of co2. Reading it on a Kindle uses 100 MJ of energy and emits 10 kg of co2.”
Hope that helps.