-- 12 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 13 14 change in resistnnce by ~rosop~i~u nelanogasler to the parasitoid wasp Asobara tabida, J. Insect Behuo 9.477-491 Henter. H.J. (1995)TRe potential for soevoolution iE a h&-psrasitcid system II. Genetic variation within a population of wasps in the ability to parasitize an aphid host, Euo!lltion49.439-44s Henter, H..i. and Via, S. (1995) The potential for c~volution in a lost-~a~itoi~ system. 1. Genetic vatiation within an @id population in susceptibility to a parasitic wasp, Eodmion 49,427-438 Rohani, P., Godlray. H.C.J. and Hassell. M.P. (1994) Aggregation and the dynamics of host-parasitoid systems: a discrete generation model with wtt~~~-generation redistribution, Am nht. 144. 491-5O9 Holt, R.D. and Hassell, M.P. (1993) Environmental heterogeneity and the stability of host-parasitoid interactions, J. Anim Ed 62,89-100 Holt, RD. (1996) Demograplrie constrai’aints In evolution: tor.ards unifying the evolutionary theoties of senesccnee and niche conservation, I%/. Erol. 10, l-1 I Gomulkiewicz, R. and Holt, R.D. (1995) Wheam does natural stlection pmvent extinctim~? Euolulion 49,201-207 arefullycontrolled transplantation ex- periments of populations ofler some of most convincing evidence for natural selection as a major agent of evolutionary change’,“. The evolutionary rates observed during such experiments contribute to our understanding of how microevolution im- pinges on macroevolution. Of great inter- est in this context is the recent work by D.N. Reznickand colleagues’on life-history evolution in Trinidadian guppies (foecilia r&ulatu). Previous studies on this spe- cies have demonstrated parallellism zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA in life- history phenotypes in different localities with similar predator regimes, and these findings indicate that predator-induced natural selection has been a major agent in shaping populaticm differentiation in life-history traits”. The results from these observational studies have now been confirmed in field experiments whereby guppies were transplanted between en- vironments with different predator re- gimes. After a few years, large evolution- ary responses had occurred for several fife-historytraits’. ation On the island of Trinidad, off South America, guppies occur in river systems at localities that are characterized by either low-or high-predation pressure. In high-predation communities, guppies are preyed upon by cichlids, and this leads to high mortality rates and creates strong selective pressures om several life-history traits. Guppies from these populations mature early and at a small size, produce many small offspring and allocate more resources to reproduction (as opposed to somatic maintenance and growth) com- pared with guppies from low-predation IQ- calities. The zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA differences in life-history traits between the guppy populations have a genetic basisd. Low-predation localities are located upstream in the river systems where predators are excluded by water- falls. Reznick and his co-workers trans- planted guppies from hi@-predation local- ities to low-predation localities upstream. After only 4-l 1 years (6-18 generations) several life-history traits had changed sig- nificantly in the experimental populations compared to the ancestral populations: age and size at maturity increased, off- spring size illcreased, offspring number decreased and reproductive effort de- creased. The direction of all these changes was consistent with expectations from life- history theory, which predicts restrained reproductive effort to evolve when pre- dation is relaxecl and adult life-expectancy increasess. Moreover. the differences be- tween ancestral and experimental popu- lations persisted in common laboratory environments demonstrating that evolu- rages, rather than pheno- ere responsible for the pattern. Evolutionary responses to selection depend both on the strength of direct se- lection on the character of interest, se- lection on correlated characters and the level of additive genetic variance for the trait in question”. Inthe present study, evo- lutionary changes in the life-history traits were more pronounced among males than among females.This was caused by higher levels of additive genetic variance for the traits among males compared with fr- males. Moreover, a zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFED poskive genetic corre- lation between age and size at maturity among males contributed to the fast evo- lution of male traits -selection on one of the traits thus largely influenced evolution of the other. Males evolved faster than females because selection had more gen- etic variation to act upon, rather than be- cause of weaker selection on females. A possible explanation for this sexual differ- 18 19 20 21 Pimentel. D. (1986) Population dyn&~§ a~! tie ~mpo~~ce of evoIotion in successful biological cotatr~l, in Riologid P/WZI and Hedfh Profec~ton (Franz, J.M., ea.), pp. 3-18, Springer Hochberg, M.E. and Holt, R.D. (1995) Reh@e evolution and population dynamics of coupled host-parasitoid associations, Euol &co/ 9, 633-661 Dawkins, R. and Krebs. J.R. (1979) Arms races between and witbin species, PIVC. R Sm. Londm ser. B 205,489~511 Abrams. P.A. (1986) Adaptive responsesof predators to prey and prey to predators: the faiture of the arms race anatqy, Evolution 40.1229-1247 ence may be that some of the genetic vari- ation for the traits among males is linked to the Y c~~ro~~oso~~e, as has been dem- onstrated for other Poeciliid fishes?. Evolutionary biologists express evo- lutionary rates in units called ‘darwins”. The number of darwins is equal to (In+- InxJAt, where x; and X? are trait values at the beginning and end of the time period respectively, and Al is the length of the time period in millions of years. This unit thus gives biologists the opportunity to compare the speed of evo- lution for different characters under vari- ous selective regimes. The evolutionary rates observed in artificial selection ex- periments in plant and animal breeding genetics typically range from 12000 to 200000 darwins, with a geometric mean of 58000 darwins. In cuntrast, evolutionary rates observed in the fossil record are much lower, ranging from 0.7-3.7darwins. Interestingly, the evolutionary rates ob- served by Reznick et al. I ranged from 3700 to 45000 darwins, and thus are much closer to rates observed in the artificial breeding experiments than those in the fossil record, Another recent study by J.B. Lcsos and colleagues’ on morphological evolution following experimental transplantation ex- periments of Anolis lizards on Bahamian islands has also demonstrated rapid evo- lutionary rates, ranging from 0 to 2117 darwins. Taken together, these two micro- evolutionary studies clearly demonstrate that iapid bursts of selection can cause considerable evolutionary change, com- pared to averaged estimated rates inferred from paleontologicai data. How do the results presented by Reznick ef al.’ contribute to our under- standing of macroevolution? Gould and Eldredge8have suggested that macroevo- lution is decoupled from microevo~ut~o~~, and that microevolutionary processes (natural selection and genetic drift) cannot fully account for macroevolutionary pat- terns such as speciation, phylogenetic Copyright 0 1997. Elsevier Science Ltd. All rights reserved. 0169-5347/97/$17.08 Pll: SU169.5347(97)01145:2 TREE ooi I%, no. 10 October 19Yi’