杏吧原创

IT鈥橲 one of the biggest questions in biology: is the outcome of evolution deterministic and predictable? In particular, was the evolution of human beings, or something similar, inevitable?

Jonathan Losos, an evolutionary biologist at Harvard University, approaches this through the contrasting views of the late Stephen Jay Gould and University of Cambridge palaeontologist Simon Conway Morris.

Gould famously argued that if we 鈥渞eplayed the tape of life鈥� we would get very different outcomes, because the pattern of evolution is unpredictable. In contrast, Conway Morris claims that convergent evolution 鈥� the idea that similar conditions produce similar adaptations 鈥� is 鈥渃ompletely ubiquitous鈥�.

Improbable Destinies focuses on the evidence underlying these opposing positions. However, chatty writing and an unclear structure mean that Losos does not explain the reasons behind Gould鈥檚 and Conway Morris鈥檚 ideas. Nor does he fully explore how their contrasting world views (Conway Morris is a devout Christian; Gould was a Marxist) influence their thinking.

Losos initially focuses on well-known examples of convergent evolution, such as the tendency of island animals 鈥� for instance hippos and mammoths 鈥� to become smaller than their continental counterparts. He also describes in some detail a series of experimental studies on lizards and fish that provide support for the centrality of convergent evolution, and thus for Conway Morris鈥檚 view.

But in the chapter on Richard Lenski鈥檚 ongoing study of bacterial evolution, Losos appears to switch sides. Lenski鈥檚 experiment began in 1988 and has, to date, involved nearly 70,000 generations and quadrillions of cells. Initially, the 12 identical lines of bacteria all grew faster and produced larger cells over the generations, so showed convergent evolution. But after around 31,000 generations, one line exhibited a unique adaptation 鈥� the ability to feed on citrate. Due to a series of random mutations, this line took a very different evolutionary path from the rest. Lenski鈥檚 attempts to encourage other lines to follow suit have failed. 鈥淪o much for predictability and parallel evolution!鈥� Losos writes.

Losos鈥檚 conclusion is that neither Gould nor Conway Morris is right. Faced with similar selection pressures, similar populations will indeed often produce convergent evolutionary outcomes. Even distantly related groups, such as marsupials and placental mammals, may do this 鈥� think of the marsupial and placental moles, separated by over 150 million years.

鈥淪tephen Jay Gould argued that if we 鈥榬eplayed the tape of life鈥� we would get very different outcomes鈥�

But the process isn鈥檛 ubiquitous. Sometimes, stuff happens and evolution goes a little crazy. In New Zealand, there were no terrestrial mammals (bats aside) until humans arrived, but in a striking example of non-convergent evolution, the islands鈥� birds did not evolve forms resembling mammals elsewhere that have a similar ecological niche and environment.

Alongside the widespread phenomenon of convergent evolution, life produces many unique forms. The human lineage is one such.

But before the reader can conclude that our uniqueness suggests we are the whole point of evolution, Losos plays his trump card: the duck-billed platypus.

This monotreme mammal has hair and a beak, and lays eggs. Like ours, its lineage is unique in the fossil record. Losos concludes that humans are no more the end-point of evolution than is the platypus, with its singular and slightly comical assemblage of characteristics. Not all evolution is convergent, he argues, and uniqueness does not imply destiny. That seems about right.

This article appeared in print under the headline 鈥淧laying dice with the animals鈥�