Chapter VI, Difficulties Of The Theory
This challenge has not been met. In the ensuing 140 years, no such thing has been found. Plants give away nectar and fruit, but they get something in return. Taking care of other members of one's own species (kin selection) doesn't count, so ants and bees (and mammalian milk) don't count.
Darwin pointed out that the Madagascar Star orchid has a spur 30 centimeters (about a foot) long, with a puddle of nectar at the bottom. Now, evolution says that nectar isn't free. Creatures that drink it pay for it, by carrying pollen away to another orchid. For that to happen, the creature must rub against the top of the spur. So, Darwin concluded that the spur had evolved its length as an arms race. Some creature had a way to reach deeply without shoving itself hard against the pollen-producing parts. Orchids with longer spurs would be more likely to spread their pollen, so Darwin's gradualistic scenario applied. The spur would evolve to be longer and longer. From the huge size, the creature must have evolved in return, reaching deeper and deeper. So, he predicted in 1862 that Madagascar has a species of hawkmoth with a tongue just slightly shorter than 30 cm.
The creature that pollinated that orchid was not learned until 1902, forty years later. It was indeed a moth, and it had a 25 cm tongue. And in 1988 it was proven that moth-pollinated short-spurred orchids did set less seed than long ones.
A thousand years ago, just about every remote island on the planet had a species of flightless bird. Evolution explains this by saying that flying creatures are particularly able to establish themselves on remote islands. Some birds, living in a safe place where there is no need to make sudden escapes, will take the opportunity to give up on flying. Hence, Evolution predicts that each flightless bird species arose on the island that it was found on. So, Evolution predicts that no two islands would have the same species of flightless bird. Now that all the world's islands have been visited, we know that this was a correct prediction.
The "same" protein in two related species is usually slightly different. A protein is made from a sequence of amino acids, and the two species have slightly different sequences. We can measure the sequences of many species, and cladistics has a mathematical procedure which tells us if these many sequences imply one common ancestral sequence. Evolution predicts that these species are all descended from a common ancestral species, and that the ancestral species used the ancestral sequence.
This has been done for pancreatic ribonuclease in ruminants. (Cows, sheep, goats, deer and giraffes are ruminants.) Measurements were made on various ruminants. An ancestral sequence was computed, and protein molecules with that sequence were manufactured. When sequences are chosen at random, we usually wind up with a useless goo. However, the manufactured molecules were biologically active substances. Furthermore, they did exactly what a pancreatic ribonuclease is supposed to do - namely, digest ribonucleic acids.
An animal's bones contain oxygen atoms from the water it drank while growing. And, fresh water and salt water can be told apart by their slightly different mixture of oxygen isotopes. (This is because fresh water comes from water that evaporated out of the ocean. Lighter atoms evaporate more easily than heavy ones do, so fresh water has fewer of the heavy atoms.)
Therefore, it should be possible to analyze an aquatic creature's bones, and tell whether it grew up in fresh water or in the ocean. This has been done, and it worked. We can distinguish the bones of river dolphins from the bones of killer whales.
Now for the prediction. We have fossils of various early whales. Since whales are mammals, evolution predicts that they evolved from land animals. And, the very earliest of those whales would have lived in fresh water, while they were evolving their aquatic skills. Therefore, the oxygen isotope ratios in their fossils should be like the isotope ratios in modern river dolphins.
It's been measured, and the prediction was correct. The two oldest species in the fossil record - Pakicetus and Ambulocetus - lived in fresh water. Rodhocetus, Basilosaurus and the others all lived in salt water.
The point is not that these prove evolution right. The point is that these were predictions that could have turned out to be wrong predictions. So, the people who made the predictions were doing science. The Theory of Evolution was also useful, in the sense that it suggested what evidence to look for, and where. http://chem.tufts.edu/AnswersInScience/evo_science.html