An elephant is very like a tree
The reason why the situations of string theory and evolution are analogous is that both of them are more or less inevitable given the known data, known approximate laws of Nature, and known and derivable logical constraints.
No minor, small observations would be enough to rule out the whole framework of evolution or the whole framework of string theory because we have much stronger reasons to be nearly certain that the general pictures are correct and our job is to complete the details. Surely, we don't abandon evolution after the first surprising observation of the genes of XY where XY does not stand for Xi Yin. In the same way, we don't abandon string theory once we find out that the simplest compactification - on torus - is not realistic.
Evolution is the only way how to logically reconcile the rich spectrum of complicated life forms, their shared molecular and other features, their diversity, currently observable mechanisms that decide about the survival, and the history of the Earth and the Universe including the obvious life-less beginning.
String theory is the only consistent way how to logically reconcile the known rich spectrum of particles and forces, including gravity, in a quantum setup, how to properly reflect their known interactions, and the known behavior of particle physics and general relativity at different energy scales. It is the only way to go beyond quantum field theory without sacrificing its proven features, and because we almost certainly know that we must go beyond ordinary local field theory, string theory seems to be the only game in town, despite 4 unsuccessful decades of attempts to show that there could exist alternatives.
In the case of evolution, the evidence is more of an observational character and the basic logic is comprehensible to many people; in the case of string theory, the evidence is much more strongly based on mathematical reasoning that is comprehensible to a limited set of experts.
Both string theory and evolution have critics who claim that string theory and evolution are not falsifiable. In both cases, the theories are viewed as "not falsifiable" because at least to some extent, the inevitability of the frameworks has already been proved which makes it hard to falsify it.
One slight difference: evolution predicted millions of things that have subsequently been observed. A few examples: The literally thousands of missing links, unknown to Darwin, but subsequently discovered. The great age of the Earth, not known or understood in Darwin's time. The fact that the mechanisms of heredity would support just such a system of evolution by natural selection that Darwin postulated. The fact that every organism on Earth shares the same basic genetic code and much the same biochemical mechanisms.
It would have been easy for evolution to have been falsified - no missing links might have been discovered, the Earth might have been as young as physicists of Darwin's time believed, different organisms might have proven to have different genetic machinery.
String theory also makes somes apparently unambiguous predictions: extra dimensions and supersymmetry for example. Neither is yet observed.