A seemingly obvious question is, If you want to find the best tree, why not write out all the possibilities & choose the shortest? The answer is that it's computationally impossible. My PhD thesis computation in 1983 with RFLP maps from N = 6 species of frogs (>100 trees) was worked out as a hand calculation on paper; without a shortcut, there would have been > 10,000. The evolutionary geneticist Joe Felsenstein (1942 - ) showed that the number of possible trees follows a hyper-exponential function, as # trees for t taxa = (2t-3)! / [ (2^t-2)(t-2)! ]. The Branch & Bound algorithm begins to require too much computer time to be practical at about t = 20, and the number of possible trees very rapidly exceeds what is possible to represent on a single 64-bit computer, or a network of such computers, or all computers available, or all computers anywhere. With N = 52 taxa, a modest phylogenetic study by current standards, the number of possible solutions (2.75 x 10⁸⁰) exceeds Eddinger's Number, the estimated number of molecules in the Universe (~10⁸⁰) !