Oops, they're doin’ it again…The Permian–Triassic Extinction

New s& Comment shift in attacks toward the rear, w hereas the alternative hypothesis predicts the opposite. Cordero suggests that experim ents could be perform ed on butterfly species w ithout false heads (but that are related to species w ith false heads) in w hich artificial heads are experim entally added, or in species w ith the false heads experim entally rem oved. We feel that such m anipulations w ould be technically challenging, and the experim ents difficult to design if TRENDS in Ecology & Evolution Vol.16 No.8 August 2001 manipulations are not to affect prey behaviour, and the target part of the body of a butterfly in a given attack is to be definitively identified. We suggest that an effective way to explore how adding a false head affects predator attack targets would be to use a predatory bird trained to peck at computergenerated images on a touch-sensitive screen. This technique is already well established in the study of prey visual detection by predators (e.g. Ref. 2) and might well lead to exciting and rapid advances in a 427 system previously (but perhaps prematurely) considered to be well understood. 1 Cordero, C. (2001) A different look at the false head in butterflies. Ecol. Entom ol. 26, 106–108 2 Dukas, R. and Kam il, A.C. (2001) Lim ited attention: the constraint underlying search im age. Behav. Ecol. 12, 192–199 Graeme D. Ruxton [email protected] Stuart Humphries S.Hum [email protected] M aking sense of mammals The higher level relationships of eutherian mammals have long been a phylogenetic puzzle, the main reason being a temporal inconsistency: although most of the major lineages (‘orders’) of placental mammals originated well before the Cretaceous–Tertiary (K–T) boundary, they only diversified in a massive adaptive radiation after this event. Therefore, much of the early evolutionary history is difficult for morphological studies to address and often only the adaptive radiation is reflected as unresolved relationships at the base of the tree. Molecular studies hold more promise at seeing past the K–T boundary, but this promise has been largely unfulfilled owing to limited data. Comparatively few, distantly related eutherian species have been sampled for only a few molecular markers, a scenario for which accurate phylogenetic estimation is known to be difficult. Now, three research groups1–3 using two different approaches have attempted to address these limitations to present the first comprehensive mammalian trees. The results show a surprising degree of concordance. M adsen et al.1 and M urphy et al.2 independently constructed tw o of the largest m olecular data sets yet assem bled for any phylogenetic study (8655 bp for 26 species and 9779 bp for 64 species, respectively). M oreover, because the sequence data do not overlap betw een the studies, the estim ates provide independent tests of each other. Both studies indicate the sam e four m ajor clades of placental m am m als: Afrotheria, Xenarthra, Laurasiatheria and another large as yet unnam ed group. Tw o other unexpected results occur. First is http://tree.trends.com the grouping of rodents plus rabbits and pikas as Glires, a result previously supported prim arily by m orphological studies. Second is the grouping of elephants, sirenians, aardvark, elephant shrew s and Old World insectivores (together, the Afrotheria), a cluster w ith no m orphological support. Liu et al.3 take another approach by combining 430 morphological and molecular estimates of eutherian phylogeny to form a comprehensive family-level supertree. In spite of large differences between the molecular and morphological supertrees, the combined data reveal the same four major eutherian clades as Refs 1 and 2, with only two exceptions (slightly altered placements for Old World insectivores, and Primates, Dermoptera, plus Scandentia). The phylogenetic picture for m am m als is still far from com plete. Differences betw een all three studies still exist, especially w ithin the four m ajor clades. How ever, the root of the placental tree is m ore agreed upon – and better resolved – than ever before. With continued sequencing efforts and data accum ulation, agreem ent should increase up towards the tips of the m am m al tree. M olecular evidence has done m uch to, and w ill continue to, foster this grow th. How ever, as pointed out by Liu et al.3, the im portance of additional m orphological data cannot be ignored. Together, only the use of as m uch inform ation as possible, from all data sources, w ill allow us to finally put all m am m als in their place. 1 M adsen, O. et al. (2001) Parallel adaptive radiations in tw o m ajor clades of placental m am m als. Nature, 409, 610–614 2 M urphy, W.J. et al. (2001) M olecular phylogenetics and the origins of placental m am m als. Nature, 409, 614–618 3 Liu, F-G.R. et al. (2001) M olecular and m orphological supertrees for eutherian (placental) m am m als. Science, 291, 1786–1789 Olaf Bininda-Emonds [email protected] …The Oops, they’re doin’ it again… Permian–Triassic Extinction The Perm ian–Triassic (P–T) extinction (approxim ately 251 m illion years ago) was the largest in the history of the Earth, w ith ~90% of m arine species, nearly 70% of terrestrial vertebrates, and m any plants disappearing. How ever, in spite of being larger than its flashier cousin, the Cretaceous–Tertiary (K–T) extinction – w hen dinosaurs bit the big one – the P–T event has not received as m uch press. During the 1980s, a debate raged about the cause of the K–T extinction. Opinion polarized betw een tw o cam ps: those w ho attributed all extinctions to a single environm ental catastrophe caused by an asteroid im pact, and those w ho believed that Earth-based phenom ena such as sea-level, clim ate change and volcanism w ere to blam e. This polarization obscured the probably com plex link betw een Earth-based and 0169–5347/01/$ – see front m atter © 2001 Elsevier Science Ltd. All rights reserved. 428 New s& Comment extraterrestrial events and, eventually, the issue dropped from the scientific radar screen. We are m oving on to bigger extinctions. Tw o new papers have recently set up a sim ilar dichotomy at the P–T boundary. Hotinski et al.1 explored the hypothesis that w orld-w ide anoxia in the deep ocean contributed to P–T extinctions. The idea is not a new one. Pyrite-rich black shales com m on in the late Perm ian and early Triassic hinted at changes in deep-ocean chem istry. Excursions in the isotope ratios of carbon, sulfur and strontium suggested m ajor rearrangem ents in ocean circulation. This new w ork com bined m odels of m arine circulation w ith the additional feedback betw een nutrient availability and m arine prim ary productivity 1. The m odification was essential because the accum ulation of black shales requires both anoxia (w hich im plies reduced upw elling) and high m arine productivity (w hich requires an adequate nutrient supply). The authors found that warm ing at the Poles during the latest Perm ian could slow upw elling sufficiently to generate w ide spread m arine anoxia. How ever, the m echanism required a higher m arine nutrient inventory than is typical of the m odern ocean 1. In counterpoint, Becker et al.2 reported the discovery of interstellar helium trapped in fullerenes found at P–T boundary sites in China and Japan. The paper, w hich was w idely covered in the popular press, determ ined that the isotopic com position of trapped helium was sim ilar to that found in other extraterrestrial bodies, and was different from helium native to Earth. This observation, coupled w ith the particular cocktail of noble gasses inside the fullerenes, suggested that they form ed in a star or interstellar cloud and w ere transported to Earth by the im pact of large (approxim ately 9 km ) com et or asteroid 2. Therefore, they concluded that ‘such an event could have caused the severe endPerm ian m ass extinction’ 2. The stage seem s to be set for another polarizing debate. 1 Hotinski, R.M . et al . (2001) Ocean stagnation and end-Perm ian anoxia. Geology 29, 7–10 2 Becker, L. et al. (2001) Im pact event at the Perm ian–Triassic boundary: evidence from extraterrestrial noble gases in fullerenes. Science 291, 1530–1533 Nan Crystal Arens [email protected] http://tree.trends.com TRENDS in Ecology & Evolution Vol.16 No.8 August 2001 Bew are of trojans bearing fish Neither academic biologists nor politicians have been particularly well poised to confront a difficult question increasingly posed by the public: ‘Are genetically modified organisms harmful or benign?’. The answer ‘it depends’ satisfies neither the desire of academics for general theory allowing replicable predictions, nor the need of politicians for clear-cut, popular stances addressing public concerns – nor does it result in effective regulatory action. Because the necessary case-by-case examination of each genetically modified organism (GMO) is difficult, and the proper methodology has been unclear, a data vacuum has been left that extremists on both sides have been free to fill with fear mongering or hand waving, according to their biases. Muir and Howard’s1 new study of the potential effects of the escape of GM fish into wild stocks provides the rare detailed perspective that has been so needed. In response to the commercial creation of numerous species of growth-enhanced transgenic fish for potential use in aquaculture, the authors developed a stagebased population genetic model. They explored the hypothetical introduction of 60 transgenic fish into a panmictic wild-type population of 60 000 individuals using recurrence equations that included juvenile and adult viability, age at sexual maturity, female fecundity, male fertility and mating advantage. They then derived actual estimates of these parameters to test their model, by creating a model GMO – a Japanese medaka Oryzias latipes containing a human/salmon growth hormone construct – on which they conducted a series of comparative fitness experiments. They thus could explore changes in transgene frequency under multiple realistic scenarios, as well as identify parameters most influential in the spread of the transgene. Contrary to the common notion that viability has the greatest influence in transgene spread (and thus that transgenics, which often have reduced viability, present no threat), the authors found that age at sexual maturity was by far the most influential fitness component, giving early-maturing transgenic fish the advantage. Disturbingly, there appears to be a large area of realistic parameter space in which transgenic males might be large enough to dominate matings, but reduced enough in their juvenile survival that their introduction into a wild population could drive it extinct. This interaction allows transgenic fish to become ‘Trojan horses’ of sorts: they are more attractive to mates but eventually doom the population that they enter. Most sobering, however, might be the empirical demonstration by this study of a phenomenon currently under renewed scrutiny by quantitative geneticists: epistasis. The effect of the genetic insert varied enormously, even in a constant environment and among strains derived from a single founding fish. Growth in second-generation transgenic medakas ranged from −4% to 76% relative to wild-type controls, reflecting the large impact of the exact genetic background on transgene expression, and underscoring the need for extensive testing within each new GM strain. Thus, although the authors provide no easy answers, they take the crucial step of mapping the road to real regulatory solutions by providing a clear methodological framework. These experiments demonstrate that the rather arbitrary measures used in the past to assess GM fish (swimming speed, predator avoidance, gill size, etc.) could be replaced profitably with interacting fitnesscomponent models and fitness experiments. Although much remains to be explored (e.g. more realistic population structures incorporating genetic drift), this model provides a framework in which GM fish could conceivably be either approved (e.g. if performance relative to wild-type counterparts is much worse in all six fitness parameters) or unmasked as potentially Trojans. 1 Muir, W.M. and Howard, R.D. (2001) Fitness components and ecological risk of transgenic release: a model using Japanese medaka (Oryzias latipes). Am. Nat. 158, 1–16 Kathryn M. Rodríguez-Clark [email protected] Jon Paul Rodríguez [email protected] 0169–5347/01/$ – see front m atter © 2001 Elsevier Science Ltd. All rights reserved.