Of course any estimates that scientists make are "made up". But estimating the number of unseen species is a very old problem in statistics with a lot of rigoro

Of course any estimates that scientists make are "made up". But estimating the number of unseen species is a very old problem in statistics with a lot of rigorous research, including an important approach by Alan Turing. https://en.wikipedia.org/wiki/Unseen_species_problem https://en.wikipedia.org/wiki/Unseen_species_problem https://en.wikipedia.org/wiki/Good%E2%80%93Turing_frequency_estimation https://en.wikipedia.org/wiki/Good%E2%80%93Turing_frequency_... For intuition of how this is possible, imagine you sample k butterflies independently and all of them come from different species. By the birthday paradox, if there were less than k^2/2 different species, you would expect to see at least one pair from the same species, so you can estimate that there are more than k^2/2 different species.