I have said that when testing optimization models, one is not testing the hypothesis that nature optimizes. But if it is not the case that the structure and behavior of organisms are nicely adapted to ensure their survival and reproduction, optimization models cannot be useful. What justification have we for assuming this?
The idea of adaptation is older than Darwinism. In the form of the argument from design, it was a buttress of religious belief. For Darwin the problem was not to prove that organisms were adapted but to explain how adaptation could arise without a creator. He was quite willing to accept that some characteristics are ''selectively neutral.'' For example, he says (26) of the sterile dark red flower at the center of the umbel of the wild carrot: ''That the modified central flower is of no functional importance to the plant is almost certain.'' Indeed, Darwin has been chided by Cain (8) for too readily accepting Owen's argument that the homology between bones of limbs of different vertebrates is nonadaptive. For Darwin the argument was welcome, because the resemblance could then be taken as evidence for genetic relationship (or, presumably, for a paucity of imagination on the part of the creator). But Cain points out that the homol-ogy would not have been preserved if it were not adaptive.
Biologists differ greatly in the extent to which they expect to find a detailed fit between structure and function. It may be symptomatic of the times that when, in conversation, I raised Darwin's example of the carrot, two different functional explanations were at once suggested. I suspect that these explanations were fanciful. But however much one may be in doubt about the function of the antlers of the irish elk or the tail of the peacock, one can hardly suppose them to be selectively neutral. In general, the structural and behavioral traits chosen for functional analysis are of a kind that rules out neutrality as a plausible explanation. Curio (23) makes the valid point that the ampullae of Lorenzini in elasmobranchs were studied for many years before their role in enabling a fish to locate prey buried in the mud was demonstrated (40), yet the one hypothesis that was never entertained was that the organ was func-tionless. The same could be said of Curio's own work (24) on the function of mobbing in birds;behavior so widespread, so constant, and so apparently dangerous calls for a functional explanation.
There are, however, exceptions to the rule that functional investigations are carried out with the aim of identifying particular selective forces, and not of demonstrating that traits are adaptive. The work initiated by Cain and Sheppard (9) on shell color and banding in Cepaea was in part aimed at refuting the claim that the variation was selectively neutral and explicable by genetic drift. To that extend the work was aimed at demonstrating adaptation as such; it is significant, however, that the work has been most successful when it has been possible to identify a particular selection pressure (e.g., predation by thrushes).
At present, of course, the major argument between neutral and selective theories concerns enzyme polymorphism. I cannot summarize the argument here, but a few points on methodology are relevant. The argument arose because of the formulation by Kimura (43) and King and Jukes (44) of the ''neutral'' hypothesis;one reason for proposing it was the difficulty of accounting for the extensive variation by selection. Hence the stimulus was quite different from that prompting most functional investigations; it was the existence of widespread variation in a trait of no obvious selective significance.
The neutral hypothesis is a good ''Popperian'' one;if it is false, it should be possible to show it. In contrast, the hypothesis of adaptation is virtually irrefutable. In practice, however, the statistical predictions of the neutral theory depend on so many unknowns (mutation rates, the past history of population number and structure, hitch-hiking from other loci) that it has proved hard to test (53). The difficulties have led some geneticists (e.g., 14) to propose that the only way in which the matter can be settled is by the classical methods of ecological genetics—i.e., by identifying the specific selection pressures associated with particular enzyme loci. The approach has had some success but is always open to the objection that the loci for which the neutral hypothesis has been falsified are a small and biased sample.
In general, then, the problems raised by the neutral mutation theory and by optimization theory are wholly different. The latter is concerned with traits that differ between species and that can hardly be selectively neutral but whose selective significance is not fully understood.
A more serious difficulty for optimization theory is the occurrence of maladaptive traits. Optimization is based on the assumption that the population is adapted to the contemporary environment, whereas evolution is a process of continuous change. Species lag behind a changing environment. This is particularly serious when studying species in an environment that has recently been drastically changed by man. For example, Lack (48) argued that the number of eggs laid by a bird maximizes the number of surviving young. Although there is much supporting evidence, there are some apparent exceptions. For example, the gannet Sula bassana lays a single egg. Studying gannets on the Bass Rock, Nelson (71) found that if a second egg is added, the pair can successfully raise two young. The explanation can hardly be a lack of genetic variability, because species nesting in the Humboldt current off Peru lay two or even three eggs and successfully raise the young.
Lack (48) suggests that the environment for gannets may recently have improved, as evidenced by the recent increase in the population on the Bass Rock. Support for this interpretation comes from the work of Jarvis (39) on the closely related S. capensis in South Africa. This species typically lays one egg, but 1 percent of nests contain two. Using methods similar to Nelson's, Jarvis found that a pair can raise two chicks to fledg-ings, but that the average weight of twins was lower than singles, and in each nest one twin was always considerably lighter than its fellow. There is good evidence that birds fledging below the average weight are more likely to die soon after. Difficulties of a similar kind arise for the glaucous gull (see 45).
The undoubted existence of maladaptive traits, arising because evolutionary change is not instantaneous, is the most serious obstacle to the testing of optimization theories. The difficulty must arise;if species were perfectly adapted, evolution would cease. There is no easy way out. Clearly a wholesale reliance on evolutionary lag to save hypotheses that would otherwise be falsified would be fatal to the whole research program. The best we can do is to invoke evolutionary lag sparingly, and only when there are independent grounds for believing that the environment has changed recently in a relevant way.
What then is the status of the concept of adaptation? In the strong form—that all organs are perfectly adapted—it is clearly false;the vermiform appendix is sufficient to refute it. For Darwin, adaptation was an obvious fact that required an explanation; this still seems a sensible point of view. Adaptation can also be seen as a necessary consequence of natural selection. The latter I regard as a refutable scientific theory (60); but it must be refuted, if at all, by genetic experiment and not by the observation of complex behavior.
Was this article helpful?