Unless I've missed something, you've never provided an example of such. You've certainly never shown that they exist in significant numbers at DP.
Unless I've missed something, PG doesn't publish download numbers on anything other than the most popular books. However, TIA does publish download numbers which one can use as proxy: 2,583,382 Downloads of the Most Popular PG Book 8 Downloads of the Least Popular PG Book Bang-for-the-Effort Ratio of Over 300,000 to 1. You can query this yourself using the TIA "Advanced Search" option on "collection:gutenberg" fields to return = downloads + title HTML table Sort Results by: either downloads desc or downloads acs But one should be forewarned that it does not appear to me that patterns of downloads from TIA is identical to pattern of downloads directly from PG -- TIA users are more sophisticated users aka nerdy than PG direct users. Personally I would rather work on a book that is towards the 2,500,000 download end of the spectrum than on the 10 downloads end of the spectrum! Again, there are literally about 1,000 more books out there that can be saved than we have the time and effort to save. The question then becomes, which books do we save? If one is doing the entire job oneself then the answer is easy: That book which you are willing to work on. If one is picking a book and imposing the work on other volunteers then the question becomes who should have the right to make that decision and how? "First come first serve" I suggest is a horrible way to make this choice because it encourages the most greedy and inconsiderate submitters to get there first rather than to take a thoughtful approach to picking which books to save and then doing a really really good job of digitizing and OCR'ing them.