Philippa Kaur

OBITUARY
10 January 2022

Edward O. Wilson (1929–2021)

Naturalist, conservationist and synthesizer who founded sociobiology.

Bert Hölldobler

0

Bert Hölldobler

Bert Hölldobler holds the Robert A. Johnson Chair in Social Insect Research and is Regent’s Professor in the School of Life Sciences at Arizona State University, Tempe. He began working with Wilson in 1970.

View author publications

You can also search for this author in PubMed
 Google Scholar

Twitter

Facebook

Email

Download PDF

Harvard University Professor E.O. Wilson in his office at Harvard University in Cambridge, MA. USACredit: Rick Friedman/Corbis via Getty

Edward (Ed) Wilson began by exploring the systematics, geographical distribution, social organization and evolution of ants. He became one of the great scholarly synthesizers, winning two Pulitzer prizes. A superb naturalist who enjoyed challenging dogma, he fought for conservation, brought ideas of biodiversity into the mainstream and set ecology on a rigorous conceptual footing. He has died aged 92.Wilson’s book Sociobiology, published in 1975, was the first to address the evolution and organization of societies in organisms ranging from colonial bacteria to primates, including humans. The final chapter, on human social interaction, ignited controversy. Wilson argued that human behaviour, although adaptable to environmental conditions, is rooted in a genetic ‘blueprint’. Opponents claimed that nothing in human behaviour is grounded in genetics, except sleeping, eating and defecation. In a letter to The New York Review of Books, a group of academics including evolutionary biologists Stephen Jay Gould and Richard Lewontin associated Wilson’s view with racism and genocide. Wilson responded with elegance and humour; in my view, most scholars now agree that he won this argument.
Conservation: Glass half full
Wilson was born in 1929 in Birmingham, Alabama, and grew up, as he admitted in his 2006 autobiography, Naturalist, “mostly insulated from its social problems”. After studying biology at the University of Alabama in Tuscaloosa, he did graduate studies at Harvard University in Cambridge, Massachusetts. He felt its Museum of Comparative Zoology, with the world’s largest ant collection, was his “destiny”.In 1955, he obtained his PhD on the systematics of the ant genus Lasius, which includes the widespread black garden ant. Systematic biology and the study of biodiversity remained his mission, but he made significant contributions to other fields, such as animal behaviour and chemical ecology. His early work on chemical communication in animals, particularly social insects, inspired a generation of scientists to explore a new area in behavioural physiology.In 1954, Wilson set out for Melanesia, including New Guinea, to study ant taxonomy and biogeography. On the basis of his data, he elaborated the critique that he and his Harvard colleague William Brown had previously developed on the idea of subspecies. They argued that the distinctions between species should be more clearly defined, allowing for variability within species. Equally influential was their thinking on character displacement — when similar species in the same area diverge genetically to avoid competing for resources.Through his fieldwork in Melanesia and later in the Caribbean, Wilson drafted a principle of biogeography that he called the taxon cycle. Species evolve back and forth between being able to live in marginal habitats, and thus disperse widely, and restricting their distribution to species-rich habitats in island interiors. He tested this and other original hypotheses in the Florida Keys in the 1960s, in collaboration with his former student Daniel Simberloff. With ecologist Robert MacArthur, he proposed that species maintain their populations through trade-offs between number of offspring and quality of parental care (the concept of r/K selection). Their 1967 book The Theory of Island Biogeography had far-reaching effects on studies of evolution and conservation.
A revolution in evolution
From early in his career, Wilson wondered about ways to understand the evolution of social organization, from primates to social insects (such as honeybees and ants). “A congenital synthesizer,” he wrote in his autobiography, “I held on to the dream of a unifying theory.” He developed a theory of adaptive demography — that certain kinds of social structure might increase reproductive fitness — and the evolution of division of labour between castes, such as insect queens and worker groups. First brought together in The Insect Societies (1971), these concepts were elaborated in Caste and Ecology in the Social Insects, with mathematical biologist George Oster, in 1978.Sociobiology was a much more far-reaching synthesis on the evolution of social systems. The furore that ensued stimulated Wilson to write an even more provocative book, On Human Nature (1978). This garnered his first Pulitzer. His highly original book Biophilia (1984) was the first to use the term to mean human empathy for the natural world. He argued that pleasure in being surrounded by diverse living organisms is a biological adaptation. These books prepared the ground for Consilience (1998), which one reviewer called a biologist’s dream of the unity of knowledge. It proposed the kind of intellectual annexation that occurs when one field can be explained in terms of a more fundamental discipline, and received a mixed response.To his and my utmost surprise, in 1990, the huge monograph The Ants, on which we worked for years, won another Pulitzer. Wilson continued to publish on human evolution and humanity’s relationship with the planet into his 90s. Half-Earth (2016) is a passionate plea to leave half of our world to nature.Ed was not a team builder. He preferred to work alone, although in a few cases he found colleagues who complemented his abilities. He thrived on controversy. In the past two decades, he had rejected the theory of inclusive fitness — the idea that the reproductive success of an individual increases when it helps to raise the offspring of its close relatives — that he once propagated. This led to heated debates, and I opposed some of his views. When we reached a compromise and submitted the manuscript of our book The Superorganism (2009), Ed’s concluding remark was: “Bert, there is one thing we agree on 100%. That is: my co-author is wrong.” One could disagree with Ed over scientific issues and remain good friends.

Nature 601 (2022)
doi: https://doi.org/10.1038/d41586-022-00078-7

Competing Interests
The author declares no competing interests.

Related Articles

A revolution in evolution

Conservation: Glass half full

Ecology: Wilson in Africa

Evolution of eusociality

Subjects

Evolution

Ecology

Genetics

Latest on:

Evolution

SARS-CoV-2 infection in free-ranging white-tailed deer
Article 23 DEC 21

A Species-Level Timeline of Mammal Evolution Integrating Phylogenomic Data
Article 22 DEC 21

Large-scale migration into Britain during the Middle to Late Bronze Age
Article 22 DEC 21

Ecology

Emergence of methicillin resistance predates the clinical use of antibiotics
Article 05 JAN 22

SARS-CoV-2 infection in free-ranging white-tailed deer
Article 23 DEC 21

Sustainability at the crossroads
Editorial 21 DEC 21

Genetics

A blood test to predict complications of pregnancy
News & Views 05 JAN 22

Decoding gene regulation in the fly brain
Article 05 JAN 22

RNA profiles reveal signatures of future health and disease in pregnancy
Article 05 JAN 22

Jobs

Faculty Positions at Institute of Physics (IOP), Chinese Academy of Sciences: Beijing, China

Institute of Physics (IOP), Chinese Academy of Sciences (CAS)
Beijing, China

Professor of Metabolic Medicine, Nutrition and Diabetes Research

Lee Kong Chian School of Medicine
Singapore, Singapore

Post-doctoral researcher position in Automated Reasoning

University of Luxembourg
Luxembourg, Luxembourg

Doctoral candidates (PhD students) in Mathematics (Algebra and Number Theory)

University of Luxembourg
Luxembourg, Luxembourg More

1.Zehr, J. P. & Capone, D. G. Changing perspectives in marine nitrogen fixation. Science 9514, 729 (2020).
Google Scholar 
2.Karl, D. et al. Dinitrogen fixation in the world’s oceans. Biogeochemistry 57–58, 47–98 (2002).
Google Scholar 
3.Dugdale, R. & Wilkerson, F. in Primary Productivity and Biogeochemical Cycles in the Sea (eds Falkowski, P. G. et al.) 107–122 (Springer, 1992).4.Carpenter, E. J. & Capone, D. G. in Nitrogen in the Marine Environment 2nd edn (eds Capone, D. G., Bronk, D. A., Mulholland, M. R. & Carpenter, E. J.) Ch. 4 (Elsevier, 2008).5.Gruber, N. & Sarmiento, J. L. Global patterns of marine nitrogen fixation and denitrification. Global Biogeochem. Cycles 11, 23–266 (1997).
Google Scholar 
6.Buchanan, P. J., Chase, Z., Matear, R. J., Phipps, S. J. & Bindoff, N. L. Marine nitrogen fixers mediate a low latitude pathway for atmospheric CO2 drawdown. Nat. Commun. https://doi.org/10.1038/s41467-019-12549-z (2019).7.Monteiro, F. M., Follows, M. J. & Dutkiewicz, S. Distribution of diverse nitrogen fixers in the global ocean. Global Biogeochem. Cycles 24, 1–16 (2010).
Google Scholar 
8.Church, M. J., Björkman, K. M., Karl, D. M., Saito, M. A. & Zehr, J. P. Regional distributions of nitrogen-fixing bacteria in the Pacific Ocean. Limnol. Oceanogr. 53, 63–77 (2008).CAS 

Google Scholar 
9.Monteiro, F. M., Dutkiewicz, S. & Follows, M. J. Biogeographical controls on the marine nitrogen fixers. Global Biogeochem. Cycles 25, 1–8 (2011).
Google Scholar 
10.Dutkiewicz, S., Ward, B. A., Monteiro, F. & Follows, M. J. Interconnection of nitrogen fixers and iron in the Pacific Ocean: theory and numerical simulations. Global Biogeochem. Cycles 26, 1–16 (2012).
Google Scholar 
11.Walworth, N. G. et al. Nutrient-colimited Trichodesmium as a nitrogen source or sink in a future ocean. Appl. Environ. Microbiol. 84, 1–14 (2018).CAS 

Google Scholar 
12.McGillicuddy, D. J. Jr. Do Trichodesmium spp. populations in the North Atlantic export most of the nitrogen they fix? Global Biogeochem. Cycles 28, 103–114 (2014).CAS 

Google Scholar 
13.Carpenter, E. J. & Romans, K. Major role of the cyanobacterium Trichodesmium in nutrient cycling in the North Atlantic Ocean. Science 254, 1989–1992 (1991).
Google Scholar 
14.Bergman, B., Sandh, G., Lin, S., Larsson, J. & Carpenter, E. J. Trichodesmium – a widespread marine cyanobacterium with unusual nitrogen fixation properties. FEMS Microbiol. Rev. 37, 286–302 (2013).CAS 
PubMed 

Google Scholar 
15.Capone, D. G. Trichodesmium, a globally significant marine cyanobacterium. Science 276, 1221–1229 (1997).CAS 

Google Scholar 
16.Gallon, J. R. The oxygen sensitivity of nitrogenase: a problem for biochemists and micro-organisms. Trends Biochem. Sci. 6, 19–23 (1981).CAS 

Google Scholar 
17.Saito, M. A. et al. Iron conservation by reduction of metalloenzyme inventories in the marine diazotroph Crocosphaera watsonii. Proc. Natl Acad. Sci. USA 108, 2184–2189 (2011).CAS 
PubMed 
PubMed Central 

Google Scholar 
18.Dron, A. et al. Light-dark (12:12) cycle of carbon and nitrogen metabolism in Crocosphaera watsonii WH8501: relation to the cell cycle. Environ. Microbiol. 14, 967–981 (2012).CAS 
PubMed 

Google Scholar 
19.Mohr, W., Intermaggio, M. P. & LaRoche, J. Diel rhythm of nitrogen and carbon metabolism in the unicellular, diazotrophic cyanobacterium Crocosphaera watsonii WH8501. Environ. Microbiol. 12, 412–421 (2010).CAS 
PubMed 

Google Scholar 
20.Flores, E. & Herrero, A. Compartmentalized function through cell differentiation in filamentous cyanobacteria. Nat. Rev. Microbiol. 8, 39–50 (2010).CAS 
PubMed 

Google Scholar 
21.Burnat, M., Herrero, A. & Flores, E. Compartmentalized cyanophycin metabolism in the diazotrophic filaments of a heterocyst-forming cyanobacterium. Proc. Natl Acad. Sci. USA 111, 3823–3828 (2014).CAS 
PubMed 
PubMed Central 

Google Scholar 
22.Sherman, D. M., Tucker, D. & Sherman, L. A. Heterocyst development and localization of cyanophycin in N2-fixing cultures of Anabaena sp. PCC 7120 (Cyanobacteria). J. Phycol. 941, 932–941 (2000).
Google Scholar 
23.Lamont, H. C., Silvester, W. B. & Torrey, J. G. Nile red fluorescence demonstrates lipid in the envelope of vesicles from N2-fixing cultures of Frankia. Can. J. Microbiol. 34, 656–660 (1988).CAS 

Google Scholar 
24.Saino, T. Diel variation in nitrogen fixation by a marine blue-green alga, Trichodesmium thiebautii. Deep Sea Res. 25, 1259–1263 (1978).
Google Scholar 
25.Saino, T. & Hattori, A. Aerobic nitrogen fixation by the marine non-heterocystous cyanobacterium Trichodesmium (Oscillatoria) spp.: its protective mechanism against oxygen. Mar. Biol. 70, 251–254 (1982).
Google Scholar 
26.Berman-Frank, I. et al. Segregation of nitrogen fixation and oxygenic photosynthesis in the marine cyanobacterium Trichodesmium. Science 294, 1534–1537 (2001).CAS 
PubMed 

Google Scholar 
27.Ohki, K. & Taniuchi, Y. Detection of nitrogenase in individual cells of a natural population of Trichodesmium using immunocytochemical methods for fluorescent cells. J. Oceanogr. 65, 427–432 (2009).CAS 

Google Scholar 
28.Eichner, M. et al. N2 fixation in free-floating filaments of Trichodesmium is higher than in transiently suboxic colony microenvironments. New Phytol. 222, 852–863 (2019).CAS 
PubMed 

Google Scholar 
29.Ohki, K. Intercellular localization of nitrogenase in a non-heterocystous cyanobacterium (cyanophyte), Trichodesmium sp. NIBB1067. J. Oceanogr. 64, 211–216 (2008).CAS 

Google Scholar 
30.Ohki, K., Zehr, F. & Fujita, Y. Regulation of nitrogenase activity in relation to the light-dark regime in the filamentous non-heterocystous cyanobacterium Trichodesmium sp. NIBB 1067. J. Gen. Microbiol. 138, 2679–2685 (1992).CAS 

Google Scholar 
31.Finzi-Hart, J. A. et al. Fixation and fate of C and N in the cyanobacterium Trichodesmium using nanometer-scale secondary ion mass spectrometry. Proc. Natl Acad. Sci. USA 106, 9931 (2009).CAS 

Google Scholar 
32.Sandh, G., El-Shehawy, R., Díez, B. & Bergman, B. Temporal separation of cell division and diazotrophy in the marine diazotrophic cyanobacterium Trichodesmium erythraeum IMS101. FEMS Microbiol. Lett. 295, 281–288 (2009).CAS 
PubMed 

Google Scholar 
33.Küpper, H. et al. Traffic lights in Trichodesmium. Regulation of photosynthesis for nitrogen fixation studied by chlorophyll fluorescence kinetic microscopy. Plant Physiol. 135, 2120–2133 (2019).
Google Scholar 
34.Ohki, K. & Fujita, Y. Aerobic nitrogenase activity measured as acetylene reduction in the marine non-heterocystous cyanobacterium Trichodesmium spp. grown under artificial conditions. Mar. Biol. 98, 111–114 (1988).CAS 

Google Scholar 
35.Waterbury, J. B. & Willey, J. M. Isolation and growth of marine planktonic Cyanobacteria. Methods Enzymol. 167, 100–105 (1988).CAS 

Google Scholar 
36.Chen, Y. B., Zehr, J. P. & Mellon, M. Growth and nitrogen fixation of the diazotrophic filamentous nonheterocystous cyanobacterium Trichodesmium sp. IMS 101 in defined media: evidence for a circadian rhythm. J. Phycol. 32, 916–923 (1996).
Google Scholar 
37.Berman-Frank, I., Bidle, K. D., Haramaty, L. & Falkowski, P. G. The demise of the marine cyanobacterium, Trichodesmium spp., via an autocatalyzed cell death pathway. Limnol. Oceanogr. 49, 997–1005 (2004).
Google Scholar 
38.Bell, P. R. F. et al. Laboratory culture studies of Trichodesmium isolated from the Great Barrier Reef lagoon, Australia. Hydrobiologia 532, 9–21 (2005).
Google Scholar 
39.Tzubari, Y., Magnezi, L., Be’Er, A. & Berman-Frank, I. Iron and phosphorus deprivation induce sociality in the marine bloom-forming cyanobacterium Trichodesmium. ISME J. 12, 1682–1693 (2018).CAS 
PubMed 
PubMed Central 

Google Scholar 
40.Held, N. A., McIlvin, M. R., Moran, D. M., Laub, M. T. & Saito, M. A. Unique patterns and biogeochemical relevance of two-component sensing in marine bacteria. mSystems 4, 1–16 (2019).
Google Scholar 
41.Aryal, U. K. & Sherman, L. A. in Cyanobacteria Omics Manipulation (ed. Los, D. A.) Ch. 6 (Caister Academic Press, 2017).42.Held, N. A. et al. Co-occurrence of Fe and P stress in natural populations of the marine diazotroph Trichodesmium. Biogeosciences 17, 2537–2551 (2020).
Google Scholar 
43.Klugkist, J., Haaker, H., Wassink, H. & Veeger, C. The catalytic activity of nitrogenase in intact Azotobacter vinelandii cells. Eur. J. Biochem. 146, 509–515 (1985).CAS 
PubMed 

Google Scholar 
44.Zehr, J. P., Wyman, M., Miller, V., Capone, D. G. & Duguay, L. Modification of the Fe protein of nitrogenase in natural populations of Trichodesmium thiebautii. Appl. Environ. Microbiol. 59, 669–676 (1993).CAS 
PubMed 
PubMed Central 

Google Scholar 
45.Rodriguez, I. B. & Ho, T.-Y. Diel nitrogen fixation pattern of Trichodesmium: the interactive control of light and Ni. Sci. Rep. 4, 4445 (2014).PubMed 
PubMed Central 

Google Scholar 
46.Eichner, M., Kranz, S. A. & Rost, B. Combined effects of different CO2 levels and N sources on the diazotrophic cyanobacterium Trichodesmium. Physiol. Plant. 152, 316–330 (2014).CAS 
PubMed 
PubMed Central 

Google Scholar 
47.Hutchins, D. A. et al. Irreversibly increased nitrogen fixation in Trichodesmium experimentally adapted to elevated carbon dioxide. Nat. Commun. 6, 1–7 (2015).
Google Scholar 
48.Levitan, O. et al. Combined effects of CO2 and light on the N2-fixing cyanobacterium Trichodesmium IMS101: a mechanistic view. Plant Physiol. 154, 346–356 (2010).CAS 
PubMed 
PubMed Central 

Google Scholar 
49.Villareal, T. A. & Carpenter, E. J. Buoyancy regulation and the potential for vertical migration in the oceanic cyanobacterium Trichodesmium. Microb. Ecol. 45, 1–10 (2003).CAS 
PubMed 

Google Scholar 
50.Rabouille, S., Staal, M., Stal, L. J. & Soetaert, K. Modeling the dynamic regulation of nitrogen fixation in the cyanobacterium Trichodesmium sp. Appl. Environ. Microbiol. 72, 3217–3227 (2006).CAS 
PubMed 
PubMed Central 

Google Scholar 
51.Breitbarth, E., Wohlers, J., Kläs, J., LaRoche, J. & Peeken, I. Nitrogen fixation and growth rates of Trichodesmium IMS-101 as a function of light intensity. Mar. Ecol. Prog. Ser. 359, 25–36 (2008).CAS 

Google Scholar 
52.Chen, Y. B. et al. Circadian rhythm of nitrogenase gene expression in the diazotrophic filamentous nonheterocystous cyanobacterium Trichodesmium sp. strain IMS101. J. Bacteriol. 180, 3598–3605 (1998).CAS 
PubMed 
PubMed Central 

Google Scholar 
53.Rabouille, S., Staal, M., Stal, L. J. & Soetaert, K. Modeling the dynamic regulation of nitrogen fixation in the Cyanobacterium Trichodesmium sp. Appl. Environ. Microbiol. 72, 3217–3227 (2006).CAS 
PubMed 
PubMed Central 

Google Scholar 
54.Capone, D. G., O’Neill, J. M., Zehr, J. & Carpenter, E. J. Basis for diel variation in nitrogenase activity in the marine planktonic cyanobacterium Trichodesmium thiebautti. Appl. Environ. Microbiol. 56, 3532–3536 (1990).CAS 
PubMed 
PubMed Central 

Google Scholar 
55.Gründel, M., Scheunemann, R., Lockau, W. & Zilliges, Y. Impaired glycogen synthesis causes metabolic overflow reactions and affects stress responses in the cyanobacterium Synechocystis sp. PCC 6803. Microbiology 158, 3032–3043 (2012).PubMed 

Google Scholar 
56.Jackson, S. A., Eaton-Rye, J. J., Bryant, D. A., Posewitz, M. C. & Davies, F. K. Dynamics of photosynthesis in a glycogen-deficient glgC mutant of Synechococcus sp. strain PCC 7002. Appl. Environ. Microbiol. 81, 6210–6222 (2015).CAS 
PubMed 
PubMed Central 

Google Scholar 
57.Boatman, T. G., Davey, P. A., Lawson, T. & Geider, R. J. The physiological cost of diazotrophy for Trichodesmium erythraeum IMS101. PLoS ONE 13, 1–24 (2018).
Google Scholar 
58.Chappell, P. D., Moffett, J. W., Hynes, A. M. & Webb, E. A. Molecular evidence of iron limitation and availability in the global diazotroph Trichodesmium. ISME J. 6, 1728–1739 (2012).CAS 
PubMed 
PubMed Central 

Google Scholar 
59.Chappell, P. D. & Webb, E. A. A molecular assessment of the iron stress response in the two phylogenetic clades of Trichodesmium. Environ. Microbiol. 12, 13–27 (2010).CAS 
PubMed 

Google Scholar 
60.Walsby, A. E. The properties and buoyancy-providing role of gas vacuoles in Trichodesmium Ehrenberg. Br. Phycol. J. 13, 103–116 (1978).
Google Scholar 
61.Villareal, T. A. & Carpenter, E. J. Diel buoyancy regulation in the marine diazotrophic cyanobacterium Trichodesmium thiebautii. Limnol. Oceanogr. 35, 1832–1837 (1990).
Google Scholar 
62.Romans, K. M., Carpenter, E. J. & Bergman, B. Buoyancy regulation in the colonial diazotrophic cyanobacterium Trichodesmium tenue: ultrastructure and storage of carbohydrate, polyphosphate, and nitrogen. J. Phycol. 30, 935–942 (1994).
Google Scholar 
63.Wang, L. et al. Molecular structure of glycogen in Escherichia coli. Biomacromolecules 20, 2821–2829 (2019).CAS 
PubMed 

Google Scholar 
64.Berman-Frank, I., Cullen, J. T., Shaked, Y., Sherrell, R. M. & Falkowski, P. G. Iron availability, cellular iron quotas, and nitrogen fixation in Trichodesmium. Limnol. Oceanogr. 46, 1249–1260 (2001).CAS 

Google Scholar 
65.Kustka, A. B. et al. Iron requirements for dinitrogen- and ammonium-supported growth in cultures of Trichodesmium (IMS 101): comparison with nitrogen fixation rates and iron:carbon ratios of field populations. Limnol. Oceanogr. 49, 1224 (2004).CAS 

Google Scholar 
66.Paerl, H. W., Prufert-Bebout, I. L. E., Guo, C. & Carolina, N. Iron-stimulated N2 fixation and growth in natural and cultured populations of the planktonic marine cyanobacteria Trichodesmium spp. Appl. Environ. Microbiol. 60, 1044–1047 (1994).CAS 
PubMed 
PubMed Central 

Google Scholar 
67.Rubin, M., Berman-Frank, I. & Shaked, Y. Dust- and mineral-iron utilization by the marine dinitrogen-fixer Trichodesmium. Nat. Geosci. 4, 529–534 (2011).CAS 

Google Scholar 
68.Polyviou, D. et al. Desert dust as a source of iron to the globally important diazotroph Trichodesmium. Front. Microbiol. 8, 1–12 (2018).
Google Scholar 
69.Basu, S. & Shaked, Y. Mineral iron utilization by natural and cultured Trichodesmium and associated bacteria. Limnol. Oceanogr. 63, 2307–2320 (2018).CAS 

Google Scholar 
70.Held, N. A. et al. Mechanisms and heterogeneity of in situ mineral processing by the marine nitrogen fixer Trichodesmium revealed by single-colony metaproteomics. ISME Commun. 1, 35 (2021).
Google Scholar 
71.Basu, S., Gledhill, M., de Beer, D., Prabhu Matondkar, S. G. & Shaked, Y. Colonies of marine cyanobacteria Trichodesmium interact with associated bacteria to acquire iron from dust. Commun. Biol. 2, 1–8 (2019).CAS 

Google Scholar 
72.Tyrrell, T. et al. Large-scale latitudinal distribution of Trichodesmium spp. in the Atlantic Ocean. J. Plankton Res. 25, 405–416 (2003).CAS 

Google Scholar 
73.Robson, R. L. & Postgate, J. R. Oxygen and hydrogen in biological nitrogen fixation. Annu. Rev. Microbiol. 34, 183–207 (1980).74.Zehr, J. P. Nitrogen fixation by marine cyanobacteria. Trends Microbiol. 19, 162–173 (2011).CAS 
PubMed 

Google Scholar 
75.Bergman, B. & Carpenter, E. J. Nitrogenase confined to randomly distributed trichomes in the marine cyanobacterium Trichodesmium thiebautii. J. Phycol. 27, 158–165 (1991).CAS 

Google Scholar 
76.Inomura, K., Wilson, S. T. & Deutsch, C. Mechanistic model for the coexistence of nitrogen fixation and photosynthesis in marine Trichodesmium. mSystems 4, 1–13 (2019).
Google Scholar 
77.Janson, S., Matveyev, A. & Bergman, B. The presence and expression of hetR in the non-heterocystous cyanobacterium Symploca PCC 8002. FEMS Microbiol. Lett. 168, 173–179 (1998).CAS 
PubMed 

Google Scholar 
78.Zhang, J. Y., Chen, W. L. & Zhang, C. C. hetR and patS, two genes necessary for heterocyst pattern formation, are widespread in filamentous nonheterocyst-forming cyanobacteria. Microbiology 155, 1418–1426 (2009).CAS 
PubMed 

Google Scholar 
79.Moore, J. K., Doney, S. C., Glover, D. M. & Fung, I. Y. Iron cycling and nutrient-limitation patterns in surface waters of the world ocean. Deep Sea Res. 2 Top. Stud. Oceanogr. 49, 463–507 (2001).
Google Scholar 
80.Chisholm, S. W. in Primary Productivity and Biogeochemical Cycles in the Sea (eds Falkowski, P. G. et al.) 213–237 (Springer, 1992).https://doi.org/10.1007/978-1-4899-0762-2_1281.Young, K. D. The selective value of bacterial shape. Microbiol. Mol. Biol. Rev. 70, 660–703 (2006).PubMed 
PubMed Central 

Google Scholar 
82.Lu, X. & Zhu, H. Tube-gel digestion: a novel proteomic approach for high-throughput analysis of membrane proteins. Mol. Cell Proteom. 4, 1948–1958 (2005).CAS 

Google Scholar 
83.Saito, M. A. et al. Multiple nutrient stresses at intersecting Pacific Ocean biomes detected by protein biomarkers. Science 345, 1173–1177 (2014).CAS 
PubMed 

Google Scholar 
84.McIlvin, M. R. & Saito, M. A. Online nanoflow two-dimension comprehensive active modulation reversed phase-reversed phase liquid chromatography high-resolution mass spectrometry for metaproteomics of environmental and microbiome samples. J. Proteome Res. 20, 4589–4597 (2021).CAS 
PubMed 

Google Scholar 
85.Lee, M. D. et al. Transcriptional activities of the microbial consortium living with the marine nitrogen-fixing cyanobacterium Trichodesmium reveal potential roles in community-level nitrogen cycling. Appl. Environ. Microbiol. 84, AEM.02026-17 (2017).
Google Scholar 
86.Zhang, Y., Wen, Z., Washburn, M. P. & Florens, L. Refinements to label-free proteome quantitation: how to deal with peptides shared by multiple proteins. Anal. Chem. 82, 2272–2281 (2010).CAS 
PubMed 

Google Scholar 
87.Gallien, S., Bourmaud, A., Kim, S. Y. & Domon, B. Technical considerations for large-scale parallel reaction monitoring analysis. J. Proteom. 100, 147–159 (2014).CAS 

Google Scholar 
88.Pino, L. K. et al. The skyline ecosystem: informatics for quantitative mass spectrometry proteomics. Mass Spectrom. Rev. 176, 139–148 (2019).
Google Scholar 
89.Held, N. A. et al. Mechanisms and heterogeneity of in situ mineral processing by the marine nitrogen fixer Trichodesmium revealed by single-colony metaproteomics. ISME Commun. https://doi.org/10.1038/s43705-021-00034-y (2021).90.White, A. E., Spitz, Y. H. & Letelier, R. M. Modeling carbohydrate ballasting by Trichodesmium spp. Mar. Ecol. Prog. Ser. 323, 35–45 (2006).
Google Scholar 
91.Morrison, F. A. An Introduction to Fluid Mechanics (Cambridge Univ. Press, 2013).92.Hunter, J. D. Matplotlib: a 2D graphics environment. Comput. Sci. Eng. 9, 90–95 (2007).
Google Scholar 
93.Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat. Methods 17, 261–272 (2020).CAS 
PubMed 
PubMed Central 

Google Scholar 
94.Hagberg, A. A., Schult, D. A. & Swart, P. J. Exploring network structure, dynamics, and function using NetworkX. In 7th Python Scientific Conference (SciPy 2008) 11–15 (2008). More

Study areaAll simulations were run at a 100 × 100 m resolution for the entire European Alps, which cover ~200,000 km². Elevations reach 4,810 m above sea level at the highest peak (Mont Blanc, elevational data were obtained from ref. 44). Mean annual temperature ranges from approximately −13 up to 16 °C and annual precipitation sums reach up to ~3,600 mm (climatic conditions were obtained from WorldClim45).Species dataTrue presences/absences were derived from complete species lists of 14,040 localized plots covering subalpine and alpine non-forest vegetation of the Alps, compiled from published46 and unpublished data sources. For more information see the supplementary information in ref. 21.Data on demographic rates as well as dispersal parameters were taken from ref. 21, Supplementary Table 1. Detailed values are provided in Supplementary Table 1.Environmental variablesCurrent climate dataMaps of current climatic conditions were generated on the basis of mean, minimum and maximum monthly temperature obtained from WorldClim45 and monthly precipitation sums derived from ref. 47 at a resolutions of 0.5 arcmin and 5 km, respectively. Temperature and precipitation data were downscaled to 100 m as described in ref. 21 and using ordinary kriging with elevation as covariable. As the reference periods of these two datasets do not match (temperature values represent averages for 1950–2000, while precipitation covers 1970–2005) temperature values were adapted using the E-OBS climate grids available online (www.ecad.eu/download/ensembles/download.php). We used these spatially refined temperature and precipitation grids to derive maps of mean annual temperature and mean annual precipitation sum. We chose only two climatic variables to keep models simple and, therefore, interpretation of results more straightforward. In fact, the climatic drivers of population performance and distribution can be more complex48 and vary among species, life history stages and vital rates49. However, as correlations between different descriptors of temperature (such as coldest month or warmest month temperature, Pearson correlation of 0.84) as well as between precipitation variables are high in the European Alps, we chose mean annual temperature and mean annual precipitation sum as they give the most basic description of how climatic conditions change over geographical and elevational gradients.Future climate dataMonthly time series of mean temperature as well as precipitation sums predicted for the twenty-first century were extracted from the Cordex data portal (http://cordex.org). We chose two IPCC5 scenarios from the RCP families representing mild (RCP 2.6) and severe (RCP 8.5) climate change to consider the uncertainty in the future climate predictions. Both scenarios were generated by Météo-France/Centre National de Recherches Météorologiques using the CNRM-ALADIN53 model, fed by output from the global circulation model CNRM-CM5 (ref. 50). The RCP 2.6 scenario assumes that radiative forcing reaches nearly 3 W m−2 (equal to 490 ppm CO2 equivalent) mid-century and will decrease to 2.6 W m−2 by 2100. In the RCP 8.5 scenario, radiative forcing continues to rise throughout the twenty-first century and reaches >8.5 W m−2 (equal to 1,370 ppm CO2 equivalent) in 210024.These time series were statistically downscaled (delta method) by (1) calculating differences (deltas) between monthly temperature and precipitation values hindcasted for the current climatic conditions (mean 1970–2005) and forecasted future values at the original spatial resolution of 11′; (2) spatially interpolating these differences to a resolution of 100 × 100 m using cubic splines and (3) adding them to the downscaled current climate data separately for each climatic variable21,36. Subsequently, we calculated running means (averaged over 35 years) for every tenth year (2030, 2040 through to 2080) for each climatic variable and finally derived, on the basis of the monthly data, mean annual temperature and mean annual precipitation sums for these decadal time steps. The application of 35-yr running means ensures that we use the same time interval for parameterization and prediction. Furthermore, for long-lived species such as alpine plants, running means over long time intervals appear most appropriate to characterize climatic niches33.Soil dataIn addition to the climatic data we used a map of the percentage of calcareous substrate within a cell (5′ longitude × 3′ latitude dissolved to 100 m resolution; further referred to as soil) as described in the supplementary information of ref. 21.Environmental suitability modellingWe parameterized logistic regression models (LRMs) with a logit link function using species presence/absence data as response and the three environmental (two bioclimatic and one soil) variables as predictors. All predictor variables entered the model as second-order polynomials in agreement with the standard unimodal niche concept.From the models, we also derived a threshold value to use for translating predicted probability of occurrence (as a measure of site suitability) into predicted presence or absence of each species at a site (called occurrence threshold, OT, henceforth). The threshold was defined such that it optimized the true skills statistic (TSS), a measure of predictive accuracy derived from comparing observed and predicted presence–absence maps51.Genetic model and niche partitioningSpecies-specific suitability curves for the annual mean temperature gradient derived from the LRMs were partitioned into suitability curves of ecologically different haplotypes of a species as defined by allelic variation in up to three diploid loci (Extended Data Fig. 6). The number of alleles was varied (n = 1, 2, 3, 5 and 10 alleles) as was the proportion of the entire species’ (temperature) niche covered by each haplotype. Models with more than one locus were run with diallelic loci, as otherwise computational efforts would have increased excessively (for each haplotype the number of seeds, juveniles and adults has to be stored and all seeds have to be distributed separately). Each combination of haplotype number and allelic niche size used in a particular simulation is further referred to as setting. Species-specific suitability curves along the other two dimensions (precipitation and soil) remained unpartitioned to ease interpretation of results. The implications of relaxing this assumption by modelling niche partitioning along different environmental gradients are hard to predict. Outcomes would probably depend on the direction and strength of individual specialization along these gradients, whether they are positively or negatively correlated, as well as on how both temperature and precipitation patterns will be affected by climate change. As a consequence, the patterns we found could be re-enforced, for example when the replacement of cold-adapted haplotypes within the species elevational range is further delayed, for example, because those haplotypes adapted to warmer conditions can cope less well with higher precipitation at higher elevations. Vice versa, maladaptation to the warming temperatures might be attenuated, for example, if temperature increase is paralleled by precipitation decrease and emerging drought stress. If, in this case, haplotypes from lower elevations can better cope with both higher temperatures and less water availability than those of median elevations, they may replace the latter faster at these median elevations and hence shorten the phase of maladaptation.Allelic effects were assumed to define the temperature optimum additively. Hence, the heterozygotes’ optimum is always exactly between the optima of the two corresponding homozygotes, corresponding to a codominant genetic model. Finally, all haplotypes corresponding to one setting were assumed to have constant (temperature) niche size, defined as a proportion (ω = 50%, 75% and 100%, for one haplotype only 100%) of the entire species’ (temperature) niche. The temperature niche was computed as the difference between the upper and lower temperature values at which the LRM-derived suitability curve predicted a suitability equal to OT (with precipitation and soil set to the respective optima of the species, also derived from the LRMs). To derive the same geographic distribution under current climatic conditions for each setting, the union of the niches of all haplotypes of one set has to approximate the niche of the single-species model (Extended Data Fig. 6). Note, however, that, the aspired equality of niches is impossible, as the niches resulting from a logistic regression with quadratic terms are always elliptic in shape. Therefore, the optima of the two extreme homozygotes (that is, those carrying haplotypes adapted to the coldest or warmest margins of the entire species’ niche) are fixed at: niche limits ± 1/2 × ω × niche size and all other optima are distributed between them at equal distances (Extended Data Fig. 6 gives a schematic illustration). As a consequence, the performance of the extreme haplotypes, both coldest and warmest, is modelled to be somewhat higher towards the cold and warm margins of the temperature niche, respectively, compared with the performance modelled for the species without intraspecific niche partitioning (compare the overlap of the black with the red and blue curves in Extended Data Fig. 6a). However, as haplotype number did not affect modelled range loss (compare Table 1), this marginal mismatch does not apparently impact our results. Homozygotes were ordered from the cold-adapted A1A1 up to the warm-adapted AnAn.Finally, the suitability curves of the genotypes were assumed to have the same value at their optimum as the species-specific suitability curve at this point (Extended Data Fig. 6).Artificial landscapesArtificial landscapes were defined as a raster of 50 × 112 cells (of 100 × 100 m). These rasters were homogeneous with respect to precipitation and soil carbon conditions which were set to the values optimal for each species according to the LRMs. With respect to temperature, by contrast, we implemented a gradient across the raster running from the minimum –9.1 °C to the maximum +3.8 °C temperature for which the LRM predicts values >OT across all six species. Buffering by 1 °C at both limits was done to avoid truncating simulation results. Further 4 °C at the lower limit were added to consider simulated temperature increase (below). The final temperature range represented by the raster ran from −14.1 to +4.8 °C. Temperature increased linearly along this axis by an increment of 0.171 °C per cell, derived from a 20° slope and a temperature decrease of 0.5 °C per 100 m of elevational change. Along the 50-cell breadth of the landscape, temperature values were kept constant. On the basis of these grids, we implemented a moderate and a severe climate change scenario, characterized by temperature increases of 2 and 4 °C, respectively, over 80 yr. Therefore, temperature of each raster cell increased annually by 0.025 and 0.05 °C, respectively.Simulations of spatiotemporal range dynamicsCATS21 is a spatially and temporarily explicit model operating on a two-dimensional grid (of 100 m mesh size in this case). CATS combines simulations of local species’ demography with species’ distribution models by scaling demographic rates relative to occurrence probabilities (suitabilities) predicted for the respective site or grid cell by the latter. Dispersal among grid cells is modelled as a combination of wind, exozoochoric and endozoochoric (that is, animal dispersal via attachment to the fur or ingestion and defecation, respectively) dispersal of plant seeds. Time proceeds in annual steps. The annually changing occurrence probabilities at each site affect the demography of local populations and hence, eventually, the number of seeds that are produced in each grid cell in the respective year. As a consequence, local populations grow or decrease, become extinct or establish anew and hence the species’ distribution across the whole study area changes as a function of the changing climate.Demographic modellingClimate-dependence of local demography was modelled by linking demographic rates (seed persistence, germination, survival, flowering frequency, seed yield and clonal reproduction) and carrying capacity to occurrence probabilities predicted by LRMs by means of sigmoidal functions. Furthermore, all rates were fixed at OT at a value ensuring stable population sizes; for more information see refs. 21,33. Demographic rates were confined by zero and a species-specific maximum value (Supplementary Table 1), which was assumed to be the same for all genotypes of a species. As a corollary, the demographic rates of all genotypes of a species are the same under optimal environmental conditions but their actual values for a particular site in a particular year differ due to different temperature optima of genotypes. In addition, germination, survival and clonal reproduction were modelled as density-dependent processes to account for intraspecific competition between genotypes. In our application, for all density-dependent functions, the species abundance is defined as the sum of all adult individuals in a given cell, regardless of their genotypes. Density dependence is commonly achieved by multiplying rates with (C – N)/C, where N is the species abundance and C is the (site- and genotype-specific) carrying capacity. This correction for density dependence causes the functions to drop to zero when N approximates C. To avoid the subsequent collapse of population sizes, we defined density-dependent rates as (C – N)/C × rate() + N/C × rate(OT), which ensures stable population sizes at densely populated sites occupied by only one genotype. To account for uncertainty in parameters of demographic rates, we assigned each species two value sets representing the upper and lower end of a plausible range of values on the basis of information derived from databases and own measurements (Supplementary Table 1).The simulations allowed for cross-pollination between genotypes. We used the relative amount of flowers (genotype-specific flowering frequency as defined by the sigmoid curve for the given suitability in the given raster cell for the given year × number of adults of that genotype in the population of that cell) to derive an estimate of the haplotype frequencies in the total pollen produced by the population within a grid cell. For the multiallelic case we allowed for recombination between loci with a recombination rate of 0.1%. These frequencies were set equal to the probability that particular haplotypes are transmitted to each year’s seed yield by pollination. Spatial pollen dispersal was accounted for in the following way: in each cell, 5% of the pollen involved in producing the annual seed yield, was assumed to stem from outside the respective raster cell. The proportions of different haplotypes in this 5% were derived from the overall pollen frequencies in all cells within a 700 m radius around the target cell (following estimates in ref. 52). Subsequently, produced seeds of each genotype were divided into resulting genotypes regarding the adult’s haplotype composition and the haplotype frequencies in the cells’ entire pollen load.Dispersal modellingFor wind dispersal of plant species we parameterized the analytical WALD kernel53 on the basis of measured seed traits and wind speed data from a meteorological station in the Central Alps of Austria. Exozoochorous and endozoochorous plant kernels were parameterized on the basis of correlated random walk simulations for the most frequent mammal dispersal vector in the study area, the chamois (Rupicapra rupicapra L.). For more details, see ref. 33. To account for uncertainties in species-specific dispersal rates, the proportion of seeds dispersed by the more far-reaching zoochorous kernels was assumed either as high (1–5%) or low (0.1–0.5%), setting upper and lower boundaries of a credible range of the dispersal ability of species.Simulation set up and simulation initializationTo test for the effects of climate change on genetic diversity in 2080, we ran CATS over the period 2000 to 2080 for each of the six study species across the entire Alps under a full factorial combination of (1) three niche sizes (50%, 75% and 100%); (2) six numbers of haplotypes (equal to two, three, five and ten alleles for one locus and four and eight for the diallelic two- and three-locus models, respectively); (3) three climate scenarios (current climate, RCP 2.6 and RCP 8.5); and (4) two sets of demographic and dispersal parameters. As a ‘control’ we also ran simulations for all climate scenarios and the two demographic and dispersal parameter sets for a setting with one genotype filling the whole niche of the species. To account for stochastic elements in CATS four replications were run for each combination of ‘treatments’.For simulations in artificial landscapes we used, instead of RCP 2.6 and RCP 8.5, ‘artificial’ climatic scenarios with an assumed warming of 2 and 4 °C, respectively, and no change in precipitation.All simulation runs were started with homozygotic individuals only. As initial distribution, for each simulation run each cell predicted to be environmentally suitable to the species (that is, occurrence probability of species >OT)—and within the real distribution range of the species28 (not relevant for simulations in artificial landscapes, of course)—was assumed to be occupied by an equal number of adults of each (homozygotic) genotype, with the total sum equal to the carrying capacity of the site. To accommodate this arbitrary within-cell genetic mixture of homozygotes (and missing heterozygotes) to actual local conditions we started simulations of range dynamics with a burn-in phase of 200 yr, run under constant current climatic conditions. To have a smooth transition from the burn-in phase under current climate (corresponding to the climate of the years 1970–2005; see current climate data) to future climate projections (starting with 2030) and to derive annual climate series, climate data were linearly interpolated between these two time intervals.Statistical analysisWe evaluated the contribution of climate scenario, haplotype number and haplotype niche size to overall species’ range change as well as the change in the frequency of the warm-adapted haplotype by means of linear models. In these models, log(range size in 2080/range size in 2000) and log(percentage of warm-adapted haplotype in 2080/percentage of warm-adapted haplotype in 2000), averaged over the four replicates and the two demographic and dispersal parameter sets, were the response variables. For the analysis of the change of the warm-adapted haplotype simulation settings with 100% niche size were ignored, as in this setting all haplotypes have the same temperature optimum (that is, neutral genetic variation). Approximate normality of residuals was confirmed by visual inspection.As indicator of the ‘topographic opportunity’ remaining to the species in the real world we calculated the area colonizable at elevations higher than those already occupied at the end of the simulation period. We therefore drew a buffer of 1 km around each cell predicted to be occupied in 2080 and then summed the area within these buffers at a higher elevation than the focal, occupied cell. Overlapping buffer areas were only counted once. This calculation was done for simulations conducted with one full-niche genotype per species only.Sensitivity analysisWe interpret the simulated relative decrease of warm-adapted haplotypes mainly as an effect of (1) the unrestricted expansion of cold-adapted haplotypes at the leading edge and (2) the resistance of the locally predominating haplotype that becomes increasingly maladapted with progressive climate warming, to recruitment by better-adapted haplotypes from below that are either rare or not present at all initially. However, the results achieved, and our conclusions, may be sensitive to assumptions about particular parameter values. Parameters that control the longevity of adult plants, and indirectly the rate of recruitment of new individuals, as well as those controlling gene flow via pollen (instead of seeds) may be particularly influential in this respect. We additionally ran simulations on artificial landscapes under alternative values of these parameters. In particular, we set the maximum age of plants to 10 yr instead of 100 yr and raised the proportion of locally produced pollen assumed to be transported up to 700 m to 10%. Both of these values are thus probably set to rather extreme levels: a maximum age of 10 yr is much shorter than the 30–50 yr assumed to be the standard age of (non-clonal) alpine plants31; and a cross-pollination rate between cells of 10% is high given that among the most important alpine pollinators only bumblebees are assumed to transport pollen >100 m regularly54,55. We add that we ran these additional simulations only in combination with the demographic species parameters set to high values because a short life span combined with low-level demographic parameters did not allow for stable populations of some species, even under current climatic conditions.For individual species, adapting plant age and cross-pollination rate between cells (Extended Data Fig. 7), did change the magnitude of loss of the warm-adapted haplotype. Nevertheless, for all of them the warm-adapted haplotype still became rarer when climate warmed and this effect increased with the level of warming. We are confident that our conclusions are qualitatively insensitive to variation of these parameters within a realistic range.Finally, in the simulations where we assumed a multilocus structure of the temperature niche, the recombination rate may also affect simulation results because it determines the rate by which new haplotypes can emerge. We also tested sensitivity of our simulations to doubling the recombination rate to 0.2%. Again, we found that a higher recombination rate had little qualitative effect on the results. Quantitatively, it resulted in a slightly pronounced relative decrease of the warmth-adapted haplotype in most species (Extended Data Fig. 8).Reporting SummaryFurther information on research design is available in the Nature Research Reporting Summary linked to this article. More

1.Science Plan and Implementation Strategy IGBP Report No. 53/IHDP Report No. 19 (Global Land Project, 2005).2.Millennium Ecosystem Assessment—Ecosystems and Human Well-Being: Desertification Synthesis Encyclopedia of the Anthropocene vols 1–5 (MEA, 2017).3.Huang, J., Yu, H., Guan, X., Wang, G. & Guo, R. Accelerated dryland expansion under climate change. Nat. Clim. Change 6, 166–171 (2015).Article 

Google Scholar 
4.Pimm, S. L. The complexity and stability of ecosystems. Nature 307, 321–326 (1984).Article 

Google Scholar 
5.Tilman, D. & Downing, J. A. Biodiversity and stability in grasslands. Nature 367, 363–365 (1994).Article 

Google Scholar 
6.Belnap, J. Surface disturbances: their role in acceleration desertification. Environ. Monit. Assess. 37, 38–57 (1995).Article 

Google Scholar 
7.Zhao, Y., Jia, R. L. & Wang, J. Towards stopping land degradation in drylands: water-saving techniques for cultivating biocrusts in situ. Land Degrad. Dev. 30, 2336–2346 (2019).Article 

Google Scholar 
8.Rodriguez-Caballero, E. et al. Dryland photoautotrophic soil surface communities endangered by global change. Nat. Geosci. 11, 185–189 (2018).CAS 
Article 

Google Scholar 
9.Coe, K. K. & Sparks, J. P. Physiology-based prognostic modeling of the influence of changes in precipitation on a keystone dryland plant species. Oecologia 176, 933–942 (2014).Article 

Google Scholar 
10.Ferrenberg, S., Tucker, C. L. & Reed, S. C. Biological soil crusts: diminutive communities of potential global importance. Front. Ecol. Environ. 15, 160–167 (2017).Article 

Google Scholar 
11.Belnap, J. & Gillette, D. A. Soil surface disturbance: impacts on potential wind erodibility of sand desert soils in SE Utah, USA. Land Degrad. Dev. 8, 355–362 (1997).Article 

Google Scholar 
12.Rutherford, W. A. et al. Albedo feedbacks to future climate via climate change impacts on dryland biocrusts. Sci. Rep. 7, 44188 (2017).13.Duniway, M. C. et al. Wind erosion and dust from US drylands: a review of causes, consequences, and solutions in a changing world. Ecosphere 10, e02650 (2019).14.Ferrenberg, S., Faist, A. M., Howell, A. & Reed, S. C. Biocrusts enhance soil fertility and Bromus tectorum growth, and interact with warming to influence germination. Plant Soil 429, 77–90 (2018).CAS 
Article 

Google Scholar 
15.Eldridge, D. J. et al. The pervasive and multifaceted influence of biocrusts on water in the world’s drylands. Glob. Change Biol. 26, 6003–6014 (2020).16.Ferrenberg, S., Reed, S. C. & Belnap, J. Climate change and physical disturbance cause similar community shifts in biological soil crusts. Proc. Natl Acad. Sci. USA 112, 12116–12121 (2015).CAS 
Article 

Google Scholar 
17.Reed, S. C. et al. Changes to dryland rainfall result in rapid moss mortality and altered soil fertility. Nat. Clim. Change 2, 752–755 (2012).CAS 
Article 

Google Scholar 
18.Concostrina-Zubiri, L. et al. Biological soil crusts across disturbance-recovery scenarios: effect of grazing regime on community dynamics. Ecol. Appl. 24, 1863–1877 (2014).CAS 
Article 

Google Scholar 
19.Weber, B., Bowker, M., Zhang, Y. & Belnap, J. in Biological Soil Crusts: An Organizing Principle in Drylands (eds Weber, B., Büdel, B. & Belnap, J.) 479–498 (Springer, 2016).20.Reynolds, J. F. et al. Global desertification: building a science for dryland development. Science 316, 847–851 (2007).CAS 
Article 

Google Scholar 
21.Berdugo, M. et al. Global ecosystem thresholds driven by aridity. Science 367, 787–790 (2020).CAS 
Article 

Google Scholar 
22.Bestelmeyer, B. T. et al. Analysis of abrupt transitions in ecological systems. Ecosphere 2, e03360 (2011).Article 

Google Scholar 
23.Bestelmeyer, B. T. et al. Desertification, land use, and the transformation of global drylands. Front. Ecol. Environ. 13, 28–36 (2015).Article 

Google Scholar 
24.Beisner, B., Haydon, D. & Cuddington, K. Alternative stable states in ecology. Front. Ecol. Environ. 1, 376–382 (2003).25.Fukami, T. & Nakajima, M. Community assembly: alternative stable states or alternative transient states? Ecol. Lett. 14, 973–984 (2011).Article 

Google Scholar 
26.Belnap, J. & Büdel, B. in Biological Soil Crusts: An Organizing Principle in Drylands (eds Weber, B., Büdel, B. & Belnap, J.) 305–320 (Springer, 2016).27.Belnap, J. & Warren, S. D. Measuring restoration success: a lesson from Patton’s tank tracks. Ecol. Bull. 79, 33 (1998).28.Belnap, J. & Elderidge, D. in Biological Soil Crusts: Structure, Function and Management (eds Belnap, J. & Lange, O. L.) 363–383 (Springer, 2001).29.Turner, M. G. Disturbance and landscape dynamics in a changing world. Ecology 91, 2833–2849 (2010).Article 

Google Scholar 
30.Scheffer, M. & Carpenter, S. R. Catastrophic regime shifts in ecosystems: linking theory to observation. Trends Ecol. Evol. 18, 648–656 (2003).Article 

Google Scholar 
31.Sala, O. E. & Lauenroth, W. K. Small rainfall events: an ecological role in semiarid regions. Oecologia 53, 301–304 (1982).32.Cayan, D. R. et al. Future dryness in the Southwest US and the hydrology of the early 21st century drought. Proc. Natl Acad. Sci. USA 107, 21271–21276 (2010).CAS 
Article 

Google Scholar 
33.Christensen, N. S., Wood, A. W., Nathalie, V., Lettenmaier, D. P. & Palmer, R. N. The effects of climate change on the hydrology and water resources of the Colorado river basin. Clim. Change 62, 337 (2004).Article 

Google Scholar 
34.Herrick, J. et al. Field soil aggregate stability kit for soil quality and rangeland health evaluations. Catena 44, 27–35 (2001).Article 

Google Scholar 
35.Escolar, C., Martínez, I., Bowker, M. A. & Maestre, F. T. Warming reduces the growth and diversity of biological soil crusts in a semi-arid environment: implications for ecosystem structure and functioning. Phil. Trans. R. Soc. B 367, 3087–3099 (2012).Article 

Google Scholar 
36.Scheffer, M. et al. Creating a safe operating space for iconic ecosystems: manage local stressors to promote resilience to global change. Science 347, 1317–1319 (2015).CAS 
Article 

Google Scholar 
37.Collins, S. L., Micheli, F. & Hartt, L. A method to determine rates and patterns of variability in ecological communities. Oikos 91, 285–293 (2000).Article 

Google Scholar 
38.Rillig, M. C. et al. The role of multiple global change factors in driving soil functions and microbial biodiversity. Science 366, 886–890 (2019).CAS 
Article 

Google Scholar 
39.IPCC. Climate Change 2014: Impacts, Adaptations, and Vulnerability (eds Field, C. B. et al.) (Cambridge Univ. Press, 2014).40.Mirzabaev, A. et al. in IPCC Special Report on Global Warming of 1.5 °C (eds Masson-Delmotte, V. et al.) Ch. 3 (WMO, 2018).41.Torres-Cruz, T. J. et al. Species-specific nitrogenase activity in lichen-dominated biological soil crusts from the Colorado Plateau, USA. Plant Soil 429, 113–125 (2018).CAS 
Article 

Google Scholar 
42.Omernik, J. M. & Griffith, G. E. Ecoregions of the conterminous United States: evolution of a hierarchical spatial framework. Environ. Manag. 54, 1249–1266 (2014).Article 

Google Scholar 
43.Kuske, C. R., Yeager, C. M., Johnson, S., Ticknor, L. O. & Belnap, J. Response and resilience of soil biocrust bacterial communities to chronic physical disturbance in arid shrublands. ISME J. 6, 886–897 (2011).Article 

Google Scholar 
44.Tucker, C. L., Ferrenberg, S. & Reed, S. C. Climatic sensitivity of dryland soil CO2 fluxes differs dramatically with biological soil crust successional state. Ecosystems 22, 15–32 (2018). https://doi.org/10.1007/s10021-018-0250-445.Cable, J. M. & Huxman, T. E. Precipitation pulse size effects on Sonoran Desert soil microbial crusts. Oecologia 141, 317–324 (2004).Article 

Google Scholar 
46.Karl, T. R., Knight, R. W. & Plummer, N. Trends in high-frequency climate variability in the twentieth century. Nature 377, 217–220 (1995).CAS 
Article 

Google Scholar 
47.Kunkel, K. E., Easterling, D. R., Redmond, K. & Hubbard, K. Temporal variations of extreme precipitation events in the United States: 1895–2000. Geophys. Res. Lett. 30, 1895–2000 (2003).Article 

Google Scholar 
48.Kim, J. A projection of the effects of the climate change induced by increased CO2 on extreme hydrologic events in the Western US. Clim. Change 68, 153–168 (2005).CAS 
Article 

Google Scholar 
49.Smith, S. J. et al. Climate change impacts for the conterminous USA: an integrated assessment part 1. Scenarios and context. Clim. Change 69, 7–25 (2005). https://doi.org/10.1007/1-4020-3876-3_250.Schwinning, S., Belnap, J., Bowling, D. R. & Ehleringer, J. R. Sensitivity of the Colorado Plateau to change: climate, ecosystems, and society. Ecol. Soc. 13, 28 (2008).51.Daly, C. et al. Physiographically sensitive mapping of climatological temperature and precipitation across the conterminous United States. Int. J. Climatol. 28, 2031–2064 (2008).Article 

Google Scholar 
52.Jonasson, S. The point intercept method for non-destructive estimation of biomass. Phytocoenologia 11, 385–388 (1983).Article 

Google Scholar 
53.R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2019).54.Wickham, H. ggplot2: Elegant Graphics for Data Analysis (Springer-Verlag, 2016).55.Oksanen, A. J. et al. Vegan: Community Ecology Package. Rpackage version 2.5-7 https://CRAN.R-project.org/package=vegan (2020).56.Anderson, M. J. A new method for non-parametric multivariate analysis of variance. Austral Ecol. 26, 32–46 (2008).Article 

Google Scholar 
57.Venables, W. & Ripley, B. Modern Applied Statistics with S. (Springer, 2002).58.Lenth, R., Singmann, H., Love, J., Buerkner, P. & Herve, M. Package ‘emmeans’ https://github.com/rvlenth/emmeans (2018).59.Signorell, A. DescTools: Tools for Descriptive Statistics (2021).60.Hallett, L. M. et al. codyn: an R package of community dynamics metrics. Methods Ecol. Evol. 7, 1146–1151 (2016).61.Wood, S. N. Generalized Additive Models: An Introduction with R 2nd edn 1–476 (CRC/Taylor & Francis, 2017).62.Bürkner, P. C. brms: an R package for Bayesian multilevel models using Stan. J. Stat. Softw. 80, 1–28 (2017).63.Gelman, A. & Rubin, D. B. Inference from iterative simulation using multiple sequences. Stat. Sci. 7, 547–511 (1992).
Google Scholar 
64.Vehtari, A., Gelman, A. & Gabry, J. Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Stat. Comput. 27, 1413–1432 (2017).Article 

Google Scholar 
65.Modrák, M., Barrett, M., Weber, F. & Coronado, E. bayesplot: Plotting for Bayesian Models. R package version 1.8.0 https://mc-stan.org/bayesplot/ (2021).66.Makowski, D., Ben-Shachar, M. & Lüdecke, D. bayestestR: describing effects and their uncertainty, existence and significance within the Bayesian framework. J. Open Source Softw. 4, 1541 (2019).Article 

Google Scholar 
67.Phillips, M. L., Howell, A., Lauria, C. M., Belnap, J. & Reed, S. C. Data and software code from two long-term experiments (1996–2011 and 2005–2018) at three sites on the Colorado Plateau of North America (US Geological Survey, 2021); https://doi.org/10.5066/P9RUN1TP More

1.Thompson, R. S., Parish, J. C., Maidment, S. C. R. & Barrett, P. M. Phylogeny of the ankylosaurian dinosaurs (Ornithischia: Thyreophora). J. Syst. Palaeontol. 10(2), 301–312 (2012).Article 

Google Scholar 
2.Arbour, V. M. & Currie, P. J. Ankylosaurid dinosaur tail clubs evolved through stepwise acquisition of key features. J. Anat. 227, 514–523. https://doi.org/10.1111/joa.12363 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
3.Brown, C. M. An exceptionally preserved armored dinosaur reveals the morphology and allometry of osteoderms and their horny epidermal coverings. PeerJ 5, e4066. https://doi.org/10.7717/peerj.4066 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
4.Butler, R. J. & Barrett, P. M. Palaeoenvironmental controls on the distribution of Cretaceous herbivorous dinosaurs. Sci. Nat. 95, 1027–1032 (2008).CAS 
Article 

Google Scholar 
5.Brown, C. M. et al. Dietary palaeoecology of an Early Cretaceous armoured dinosaur (Ornithischia; Nodosauridae) based on floral analysis of stomach contents. R. Soc. Open Sci. 7, 200305. https://doi.org/10.1098/rsos.200305 (2020).ADS 
CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
6.Bourke, J. M., Porter, W. R. & Witmer, L. M. Convoluted nasal passages function as efficient heat exchangers in ankylosaurs (Dinosauria: Ornithischia: Thyreophora). PLoS ONE 13, e0207381. https://doi.org/10.1371/journal.pone.0207381 (2018).CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
7.Mallon, J. C. & Anderson, J. S. The functional and palaeoecological implications of tooth morphology and wear for the megaherbivorous dinosaurs from the dinosaur park formation (Upper Campanian) of Alberta, Canada. PLoS ONE 9(6), e98605. https://doi.org/10.1371/journal.pone.0098605 (2014).ADS 
CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
8.Ősi, A., Prondvai, E., Mallon, J. & Bodor, E. R. Diversity and convergences in the evolution of feeding adaptations in ankylosaurs (Dinosauria: Ornithischia). Hist. Biol. 29(4), 539–570. https://doi.org/10.1080/08912963.2016.1208194 (2017).Article 

Google Scholar 
9.Hayashi, S., Carpenter, K., Scheyer, T. M., Watanabe, M. & Suzuki, D. Function and evolution of ankylosaur dermal armor. Acta Palaeontol. Pol. 55(2), 213–228 (2010).Article 

Google Scholar 
10.Norman, D. B. Scelidosaurus harrisonii from the Early Jurassic of Dorset, England: Cranial anatomy. Zool. J. Linn. Soc. 188(1), 1–81. https://doi.org/10.1093/zoolinnean/zlz074 (2019).Article 

Google Scholar 
11.Galton, P. M. Endocranial casts of the plated dinosaur Stegosaurus (Upper Jurassic, Western USA): A complete undistorted cast and the original specimens of Othniel Charles Marsh. In The Armored Dinosaurs (ed. Carpenter, K.) 103–129 (Indiana University Press, 2001).
Google Scholar 
12.Galton, P. M. Skull bones and endocranial casts of stegosaurian dinosaur Kentrosaurus Hennig, 1915 from Upper Jurassic of Tanzania, East Africa. Geol. Palaeontol. 22, 123–143 (1988).
Google Scholar 
13.Kuzmin, I. et al. The braincase of Bissektipelta archibaldi—new insights into endocranial osteology, vasculature, and paleoneurobiology of ankylosaurian dinosaurs. Biol. Commun. 65(2), 85–156. https://doi.org/10.21638/spbu03.2020.201 (2020).Article 

Google Scholar 
14.Leahey, L. G., Molnar, R. E., Carpenter, K., Witmer, L. M. & Salisbury, S. W. Cranial osteology of the ankylosaurian dinosaur formerly known as Minmi sp. (Ornithischia: Thyreophora) from the Lower Cretaceous Allaru Mudstone of Richmond, Queensland, Australia. PeerJ 3, e1475. https://doi.org/10.7717/peerj.1475 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
15.Paulina-Carabajal, A., Lee, Y. N., Kobayashi, Y., Lee, H. J. & Currie, P. J. Neuroanatomy of the ankylosaurid dinosaurs Tarchia teresae and Talarurus plicatospineus from the Upper Cretaceous of Mongolia, with comments on endocranial variability among ankylosaurs. Palaeogeogr. Palaeoclimatol. Palaeoecol. 494, 135–146. https://doi.org/10.1016/j.palaeo.2017.11.030 (2018).Article 

Google Scholar 
16.Pereda-Suberbiola, X. & Galton, P. M. On the taxonomic status of the dinosaur Struthiosaurus austriacus Bunzel from the Late Cretaceous of Austria. C. R. Acad. Sci. Paris II 315, 1275–1280 (1992).
Google Scholar 
17.Pereda-Suberbiola, X. & Galton, P. M. Revision of the cranial features of the dinosaur Struthiosaurus austriacus Bunzel (Ornithischia: Ankylosauria) from the Late Cretaceous of Europe. Neues Jahrb. Geol. Palaontol. Abh. 191, 173–200 (1994).
Google Scholar 
18.Pereda-Suberbiola, X. & Galton, P. M. Reappraisal of the nodosaurid ankylosaur Struthiosaurus austriacus Bunzel from the Upper Cretaceous Gosau Beds of Austria. In The Armored Dinosaurs (ed. Carpenter, K.) 173–210 (Indiana University Press, 2001).
Google Scholar 
19.Pereda-Suberbiola, X. A revised census of European Late Cretaceous nodosaurids (Ornithischia: Ankylosauria): Last occurrence and possible extinction scenarios. Terra Nova 4, 641–648 (1992).ADS 
Article 

Google Scholar 
20.Pereda-Suberbiola, X. Ankylosaurian dinosaur remains from the Upper Cretaceous of Laño (Iberian Peninsula). Est. Mus. Cienc. Nat. de Álava 14(Número especial 1), 273–288 (1999).
Google Scholar 
21.Pereda-Suberbiola, X., Astibia, H. & Buffetaut, E. New remains of the armoured dinosaur Struthiosaurus from the Late Cretaceous of the Iberian peninsula (Lafio locality, Basque-Cantabric Basin). Bull. Soc. Géol. Fr. 166, 207–211 (1995).
Google Scholar 
22.Garcia, G. & Pereda-Suberbiola, X. A new species of Struthiosaurus (Dinosauria: Ankylosauria) from the upper cretaceous of Villeveyrac (Southern France). J. Vertbr. Paleontol 23(1), 156–165 (2003).Article 

Google Scholar 
23.Nopcsa, F. Dinosaurierreste aus Siebenbürgen V. Geologica Hungarica, Series Palaeontologica 4, 1–76 (1929).
Google Scholar 
24.Codrea, V. et al. More than just Nopcsa’s Transylvanian dinosaurs: A look outside the Haţeg Basin. Palaeogeogr. Palaeoclimatol. Palaeoecol. 293, 391–405. https://doi.org/10.1016/j.palaeo.2009.10.027 (2010).Article 

Google Scholar 
25.Ősi, A. & Prondvai, E. Sympatry of two ankylosaurs (Hungarosaurus and cf. Struthiosaurus) in the Santonian of Hungary. Cretac. Res. 44, 30–38. https://doi.org/10.1016/j.cretres.2013.03.006 (2013).Article 

Google Scholar 
26.Bunzel, E. Die Reptilfauna der Gosaformation in der Neuen Welt bei Wiener-Neustadt. Abhandlungen der Kaiserlich-Königlichen Geologischen Reichsanstalt. 5, 1–18 (1871).
Google Scholar 
27.Franzosa, J. & Rowe, T. Cranial endocast of the Cretaceous theropod dinosaur Acrocanthosaurus atokensis. J. Vertebr. Paleontol. 25, 859–864 (2005).Article 

Google Scholar 
28.Witmer, L. M. & Ridgely, R. C. New insights into the brain, braincase, and ear region of tyrannosaurs (Dinosauria, Theropoda), with implications for sensory organization and behavior. Anat. Rec. 292(9), 1266–1296. https://doi.org/10.1002/ar.20983 (2009).Article 

Google Scholar 
29.Schade, M., Rauhut, O. W. M. & Evers, S. W. Neuroanatomy of the spinosaurid Irritator challengeri (Dinosauria: Theropoda) indicates potential adaptations for piscivory. Sci. Rep. 10(9259), 1613–1616. https://doi.org/10.1038/s41598-020-66261 (2020).Article 

Google Scholar 
30.Witmer, L. M., Ridgely, R. C., Dufeau, D. L. & Semones, M. C. Using CT to peer into the past: 3D visualization of the brain and ear regions of birds, crocodiles, and nonavian dinosaurs. In Anatomical imaging: towards a new morphology (eds Endo, H. & Frey, R.) 67–87 (Springer, 2008).Chapter 

Google Scholar 
31.Evers, S. W. et al. Neurovascular anatomy of the protostegid turtle Rhinochelys pulchriceps and comparisons of membranous and endosseous labyrinth shape in an extant turtle. Zool. J. Linn. Soc. 187, 800–828. https://doi.org/10.1093/zoolinnean/zlz063 (2019).Article 

Google Scholar 
32.Ősi, A., Pereda Suberbiola, X. & Földes, T. Partial skull and endocranial cast of the ankylosaurian dinosaur Hungarosaurus from the Late Cretaceous of Hungary: Implications for locomotion. Palaeontol. Electro. 17(1), 18p (2014).
Google Scholar 
33.Sampson, S. D. & Witmer, L. M. Craniofacial anatomy of Majungasaurus crenatissimus (Theropoda: Abelisauridae) from the Late Cretaceous of Madagascar. J. Vertebr. Paleontol. 27, 32–104. https://doi.org/10.1671/0272-4634(2007)27[32:CAOMCT]2.0.CO;2 (2007).Article 

Google Scholar 
34.Walsh, S. A., Barrett, P. M., Milner, A. C., Manley, G. & Witmer, L. M. Inner ear anatomy is a proxy for deducing auditory capability and behaviour in reptiles and birds. Proc. R. Soc. B 276, 1355–1360 (2009).Article 

Google Scholar 
35.Porter, W. R. & Witmer, L. M. Vascular patterns in the heads of dinosaurs: Evidence for blood vessels, sites of thermal exchange, and their role in physiological thermoregulatory strategies. Anat. Rec. 303, 1075–1103. https://doi.org/10.1002/ar.24234 (2019).Article 

Google Scholar 
36.Benson, R. B. J., Starmer-Jones, E., Close, R. A. & Walsh, S. A. Comparative analysis of vestibular ecomorphology in birds. J. Anat. 231, 990–1018. https://doi.org/10.1111/joa.12726 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
37.Bronzati, M. et al. Deep evolutionary diversification of semicircular canals in archosaurs. Curr. Biol. 31(12), 2520-2529.e6. https://doi.org/10.1016/j.cub.2021.03.086 (2021).CAS 
Article 
PubMed 

Google Scholar 
38.Hanson, M., Hoffman, E. A., Norell, M. A. & Bhullar, B. S. The early origin of a birdlike inner ear and the evolution of dinosaurian movement and vocalization. Science 372(6542), 601–609. https://doi.org/10.1126/science.abb4305 (2021).ADS 
CAS 
Article 
PubMed 

Google Scholar 
39.Miyashita, T., Arbour, V. M., Witmer, L. M. & Currie, P. J. The internal cranial morphology of an armoured dinosaur Euoplocephalus corroborated by X-ray computed tomographic reconstruction. J. Anat. 219, 661–675. https://doi.org/10.1111/j.1469-7580.2011.01427.x (2011).Article 
PubMed 
PubMed Central 

Google Scholar 
40.Paulina-Carabajal, A., Lee, Y. N. & Jacobs, L. L. Endocranial morphology of the primitive nodosaurid dinosaur Pawpawsaurus campbelli from the Early Cretaceous of North America. PLoS ONE 11, e0150845. https://doi.org/10.1371/journal.pone.0150845 (2016).CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
41.Spoor, F. & Thewissen, J. G. M. Comparative and functional anatomy of balance in aquatic mammals. In Sensory Evolution on the Threshold, Adaptations in Secondarily Aquatic Vertebrates (eds Thewissen, J. G. M. & Nummela, S.) 65–81 (University of California Press, 2008).
Google Scholar 
42.Sobral, G. & Müller, J. Archosaurs and their kin: The ruling reptiles. In Evolution of the Vertebrate Ear (eds Clack, J. A. et al.) 285–326 (Springer, 2016).Chapter 

Google Scholar 
43.Arbour, V. M. Estimating impact forces of tail club strikes by ankylosaurid dinosaurs. PLoS ONE 4(8), e6738. https://doi.org/10.1371/journal.pone.0006738 (2009).ADS 
CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
44.Park, J. Y. et al. A new ankylosaurid skeleton from the Upper Cretaceous Baruungoyot Formation of Mongolia: Its implications for ankylosaurid postcranial evolution. Sci. Rep. 11, 4101. https://doi.org/10.1038/s41598-021-83568-4 (2021).ADS 
CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
45.Witmer, L. M., Chatterjee, S., Franzosa, J. & Rowe, T. Neuroanatomy of flying reptiles and implications for flight, posture and behaviour. Nature 425, 950–953. https://doi.org/10.1038/nature02048 (2003).ADS 
CAS 
Article 
PubMed 

Google Scholar 
46.Walsh, S. A. et al. Avian cerebellar floccular fossa size is not a proxy for flying ability in birds. PLoS ONE 8(6), e67176. https://doi.org/10.1371/journal.pone.0067176 (2013).ADS 
CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
47.Ferreira-Cardoso, S. et al. Floccular fossa size is not a reliable proxy of ecology and behaviour in vertebrates. Sci. Rep. 7(1), 2017. https://doi.org/10.1038/s41598-017-01981-0 (2005).CAS 
Article 

Google Scholar 
48.Ezcurra, M. D. et al. Enigmatic dinosaur precursors bridge the gap to the origin of Pterosauria. Nature 588(7838), 445–449. https://doi.org/10.1038/s41586-020-3011-4 (2020) (Epub 2020 Dec 9).ADS 
CAS 
Article 
PubMed 

Google Scholar 
49.Carpenter, K., Sanders, F., McWhinney, L. A. & Wood, L. Evidence for predator–prey relationships: examples for Allosaurus and Stegosaurus. In The Carnivorous Dinosaurs (ed. Carpenter, K.) 325–350 (Indiana University Press, 2005).
Google Scholar 
50.Mallison, H. CAD assessment of the posture and range of motion of Kentrosaurus aethiopicus HENNIG 1915. Swiss. J. Geosci. 103, 211–233. https://doi.org/10.1007/s00015-010-0024-2 (2010).Article 

Google Scholar 
51.Arbour, V. M., Lech-Hernes, N. L., Guldberg, T. E., Hurum, J. H. & Currie, P. J. An ankylosaurid dinosaur from Mongolia with in situ armour and keratinous scale impressions. Acta Palaeontol. Pol. 58(1), 55–64. https://doi.org/10.4202/app.2011.0081 (2013).Article 

Google Scholar 
52.Scheyer, T. M. & Sander, P. M. Histology of ankylosaur osteoderms: Implications for systematics and function. J. Vertebr. Paleontol. 24, 874–893 (2004).Article 

Google Scholar 
53.Carpenter, K. Ankylosauria. In The Complete Dinosaur 2nd edn (eds Brett-Surman, M. K. et al.) 505–526 (Indiana University Press, 2012).
Google Scholar 
54.Jerison, H. J. Evolution of the brain and intelligence 482 (Academic Press, 1973).
Google Scholar 
55.Marugan-Lobon, J., Chiappe, L. M. & Farke, A. A. The variability of inner ear orientation in saurischian dinosaurs: Testing the use of semicircular canals as a reference system for comparative anatomy. PeerJ 1, e124 (2013).Article 

Google Scholar 
56.Benoit, J. et al. A test of the lateral semicircular canal correlation to head posture, diet and other biological traits in “ungulate” mammals. Sci. Rep. 10, 19602. https://doi.org/10.1038/s41598-020-76757-0 (2020).ADS 
CAS 
Article 
PubMed 
PubMed Central 

Google Scholar 
57.Coombs, W. P. Jr. Forelimb muscles of the Ankylosauria (Reptilia, Ornithischia). J. Paleontol. 52, 642–657 (1978).
Google Scholar 
58.Maidment, S. C. R. & Barrett, P. M. Osteological correlates for quadrupedality in ornithischian dinosaurs. Acta Palaeontol. Pol. 59(1), 53–70 (2014).
Google Scholar  More

Symbiont diversity and distributionThe present study provides the first evidence of symbiosis in R. alisae, a species of nudibranchs. This is the most multiple symbiosis that have ever been recorded for marine invertebrates. While many organisms establish an exclusively one-on-one relationship with a single microbial species or microbes belonging to the same functional group5,12, there are also organisms that harbor multiple microbial species, in which symbiont–symbiont and host–symbiont interactions occur. Six phylotypes of chemoautotrophic bacteria were reported for mussel Idas sp. from a cold seep area11 and five extracellular symbionts for the gutless oligochaete worm Olavius algarvensis34. However, in these cases, symbioses involving bacteria and marine invertebrates are either endosymbiotic microbes co-occurring inside the host bacteriocytes5,11 or ectosymbiotic microbes associated with the external surfaces of the animals3,4,9,15,34, with the exception of scaly-foot snail from hydrothermal vents having partnerships simultaneously with epi- and endosymbiontic microbes35.Bacterial symbionts in R. alisae have appeared to be more diverse than was previously known for marine invertebrates. It is evident that the detected symbiont phylotypes differ greatly from all other known symbionts found in marine invertebrates. Labrenzia (Rodobacteriales) and Maritalea (Rhizobiales) have not been recorded as forming symbiotic associations with invertebrates or plants so far, although other members of the families Rodobacteriales and Rhizobiales are well known symbionts14. Strains of Bradyrhizobium, Burkholderia, Achromobacter, and Stenotrophomonas are reported as symbionts of plants, interacting with a vast majority of nodulating legume species and efficient in biological nitrogen fixation36. This may be important when considering the nature of these symbionts in the nudibranch. Symbioses between cyanobacteria and marine organisms are commonly found among marine plants, fungi, sponges, ascidians, corals, and protists37,38. Synechococcus, identified as dominant symbiont clones of R. alisae (Table S2), is a unicellular cyanobacterium common in the marine environment, providing a range of beneficial functions including photosynthesis, nitrogen fixation, UV protection, and production of defensive toxins8,9,37. Symbiotic interactions between actinobacteria and their host have been observed in insects, human, animals, and plants, where the bacteria provide the host with protection against pathogens and produce essential nutrients39. However, none of the members of the clade Actinobacteria recorded in R. alisae are known to live symbiotically.Arrangement of symbiotic associationDespite the high diversity of bacteria, they are well organized in the host. Dense clusters of rod-shaped bacteria, Labrenzia, Maritalea, Bradyrhizobium, Burcholderia, Achromobacter, and Stenotrophomonas, were found within host-derived vacuoles, referred to as bacteriocytes, inside epithelial cells of R. alisae (Fig. 3). Although such arrangement differs from that typical of bacteriocytes, which are usually considered as specialized cells of the hosts for harboring bacteria, it resembles that reported for scaly-food snail from hydrothermal vents, which harbor symbionts in the esophageal gland35. Bacteriocytes in the gastropod Lurifax vitreus found near hydrothermal vents also constitute a portion of the mantle epithelium; they have large vacuoles containing many live and dividing bacteria40. Each bacteriocyte was densely packed with certain symbionts, and the bacteriocytes were randomly distributed within the epithelium cells. A distinctly regular distribution pattern was observed in the gill epithelium of the mussel Bathymodiolus sp.: the thiotrophic symbionts occupy the apical region, and the methanotrophic symbionts are more abundant in the basal region of bacteriocytes4. In the mussel Idas sp., however, there is no spatial pattern of the six distinct bacterial phylotypes, and the symbionts are mixed within bacteriocytes11.Synechococcus dominated the cytoplasm of intestinal epithelium and, more rarely, epidermis cells, mainly as specialized cell type referred to as nitrogen-fixing heterocysts. They are visually similar to cyanobacteria from corals and sponges8,37.The phylogenetic diversity and the spatial organization of the symbiotic community in R. alisae were determined by the 16S rRNA analysis, which was consistent with the results of FISH and TEM. Unlike most symbioses of marine invertebrates when bacteria house specialized host cells5,11 or cover epidermis7,15, symbiotic association of R. alisae exhibited spatial partitioning between symbionts, which were unevenly distributed between the tissues (Table S2). It has been established that different members of the microbial community can complement each other in acquisition of various restrictive nutrients, confirming the importance of the functional diversity of symbionts41. Thus, Stenotrophomonas rhizophila and Bradyrhizobium build a beneficial association in the rhizosphere and can act synergistically on promoting growth and nutrient uptake of soybean36. Cyanobacteria can interact synergistically with beneficial members from the endophytic microbiome of rice seedlings42. The location of bacterium in the organism of R. alisae may, in fact, depend on the specific metabolic and ecological roles that the symbionts play, and also on the interaction with bacterium belonging to different physiological groups.Nature of symbiosisSymbiotic associations between microbes and invertebrates are acquired mainly in a nutrient-depleted environment where symbionts usually provide their hosts with essential nutrients and high-energy compounds1. In contrast to known symbioses between microbes and gutless invertebrates, which obtain nutrients exclusively from the bacteria, R. alisae, like most nudibranch species, is a sponge-eating predator. However, due to the lack of adipose tissue, sponges are distinguished by a low lipid content (0.4 to 1.5% of wet weight)43 and also by specific proteinaceous spongin fibers and chitin, a polysaccharide similar to cellulose that can be indigestible for some predators, which together indicate their low nutritional value. Furthermore, R. alisae feeds exclusively on the sponge O. pennata; therefore, in habitats with low prey availability, this nudibranch has to survive starvation while searching for sponge assemblages. We suppose that symbiotic bacteria of R. alisae contribute to the utilization of low-quality food, similarly to symbiotic bacteria from the genera Rhodobacter, Burkholderia, and Aeromonas associated with the detritivorous isopod Asellus aquaticus44.A fatty acid analysis, as a useful approach to clarifying the nature of symbiosis5,20,32, has confirmed the trophic interaction between symbionts and the nudibranch host (Table S2). Among the fatty acids of symbiotic bacteria in R. alisae, OBFA are a major acyl constituent of membranes in Stenotrophomonas45 and also in Actinobacteria, Arthrobacter, Iamia, Ilumatobacter, and Kocuria46. Cis-vaccenic acid is a major component of Maritalea30. Omega-cyclohexyl tridecanoic acid (cyclo19:0) is specific to Bradyrhizobium47, Burkholderia, and Achromobacter48. Linoleic acid is produced by cyanobacteria including marine species of Synecoccocus33; in nudibranch, it obviously serves as a precursor in the synthesis of arachidonic acid (20:4n-6), thus, providing additional evidence for the transfer of fatty acids from symbionts to the host. Mollusks are capable of converting linoleic acid to arachidonic acid, since they have enzymes required for its synthesis21. The presence of these bacteria-specific markers and the abundance of arachidonic acid confirm the metabolic role of symbionts in the nudibranch host.Among nutrients, biologically available nitrogen can be considered a restrictive nutrient for the sponge-eating R. alisae, which can be acquired with the help of nitrogen-fixing symbionts, also referred to as diazotrophs. R. alisae harbors Bradyrhizobium, Burkholderia, Achromobacter, and Stenotrophomonas that are efficient in biological nitrogen fixation previously found to be associated with nodulating legume species36. Symbiotic nitrogen fixers are known to be associated with a variety of marine invertebrates such as wood-boring bivalves, corals, sponges, sea urchins, tunicates, and polychaetes7,8,37. Moreover, the protection of the enzyme nitrogenase that catalyzes N2 fixation against oxygen is an important physiological requirement in bacteria such as symbiotic Bradyrhizobium, Burkholderia, Achromobacter, and Stenotrophomonas that are located in bacteriocytes and provide this protection. Synechococcus is known as a nitrogen-fixer37,49. It performs N2 fixation in heterocysts where nitrogenase is restricted under oxic conditions. Indeed, heterocysts of Synechococcus are abundant in the intestine cells of R. alisae (Fig. 5B–D).Nitrate assimilation is one of the major processes of nitrogen acquisition by many heterotrophic bacteria and cyanobacteria50,51. Symbionts of R. alisae can play an important role in the process of nitrate utilization through denitrification, dissimilatory nitrate reduction, and assimilatory nitrate reduction as a nitrogen source and synthesize it into organic nitrogen. The nitrate reducers, Labrenzia52, Stenotrophomonas53, Maritalea30, and Rhodobacteraceae29 are widely represented in R. alisae. Synechococcus also utilizes nitrate, nitrite, or ammonium for growth50. Thus, symbiotic bacteria may play a significant role in the N-budget of the nudibranch mollusk.The symbiotic bacteria of R. alisae, including Bradyrhizobium, Maritalea, Labrenzia, Burkholderia, Achromobacter, Stenotrophomonas, Arthrobacter, Iamia, Ilumatobacter, and Kocuria, are known as carboxydotrophic or carbon monoxide (CO) oxidizers54,55. Despite the toxicity of CO for multicellular organisms, numerous aerobic and anaerobic microorganisms can use CO as a source of energy and/or carbon for cell growth56. The marine worm Olavius algarvensis establishes symbiosis with chemosynthetic bacteria using CO, a substrate previously not known to play a role in symbiotic associations with marine invertebrates, as an energy source57. We do not rule out that the R. alisae symbionts also might exploit CO as carbon and energy source. Despite this, assumption may seem impossible taking in account the CO toxicity, but, since many invertebrates (mollusks, tube worm, etc.) use toxic sulfate, thiosulfate, and methane as an energy source1,15, this hypothesis is worth to be addressed.An important component of skeleton in marine sponges of the family Microcionidae, including O. pennata, is the structural polysaccharide chitin58. Some bacteria are capable of hydrolyzing chitin via the activity of chitinolytic enzymes and can utilize chitin as a source of carbon, nitrogen, and/or energy59. Chitinase activity was documented for strains of Labrenzia60, Burkholderia61, Arthrobacter62, Achromobacter63, Stenotrophomonas64, Alcaligenes65, and actinobacteria59 associated with R. alisae. Thus, these bacteria can work synergistically to digest chitin and spongin, contributing to feeding success of the host nudibranch which depends solely on low-quality, nitrogen- and carbon-deficient food available.Furthermore, direct evidence has confirmed that many bioactive compounds from invertebrates are produced by symbiotic microorganisms66,67. Many biologically active compounds including toxic and deterrent secretions have been identified in nudibranchs of the family Discodorididae68. Symbiotic bacteria may exhibit toxic activity to provide the host nudibranch with chemical defense against predators and environment. Bacteria, especially actinobacteria, living in a symbiotic relationship with R. alisae may help the host in defense, since nudibranch lack a shell, and secondary metabolites of bacteria can provide them with chemical defense against predators and environment, as has been reported for some marine invertebrates2,9,10.In complex associations, the integration and coexistence of symbionts depend on supplementary partnerships and mutual contribution to the host’s metabolism41. The most intensively studied cases are highly specialized associations, where both partners can only exist in close relationship with one another. The relatively high diversity of microbes in R. alisae complicates understanding the complex pattern of molecular and cellular interactions between the host and its symbionts. More

Description of study areasIGR plant (33° 42/N, 46° 13/E) is located along the edge of the mountains of Zagros forests and 25 km from Ilam city. Its main activity, to supply gas to the western provinces of Iran, started in 2007. It converts sour gas to sweet gas and also produces various products such as pastil sulfur, ethane, and liquefied gas. The refinery has two chimneys, which release waste gases into the atmosphere. Oak trees are the main tree species of the Zagros forests around the refinery; these are exposed to various air pollutants and different elements from this source. Based on random analysis of exhaust emissions, sulfur dioxide and sulfide hydrogen are the major pollutants emitted from the flare gases of this refinery plant34. The sampling points have an average altitude of about 1000–1250 m and a slope of less than 20%. The climate of the region is semiarid and influenced by Mediterranean winds. The predominant wind direction was west and southwest. The highest and lowest air temperatures were 41.4 °C and − 11.3 °C, respectively. The average annual rainfall was 71.94 mm (http://www.amarilam.ir).Samples collection and analysesAll methods were carried out in accordance with the relevant institutional, national, and international guidelines and legislation. Besides they were discussed and approved by the Research Ethics Committee of Tarbiat Modares University. The formal identification of the Quercus brantii Lindl. was performed by H. Dadkhah-Aghdash based on colorful Flora of Iran35. The permissions or licenses to collect Brant oak (Quercus brantii Lindl.) trees in Zagros forests were obtained. A voucher specimen of Brant oaks were collected and deposited at the Herbarium of department of Plant Biology of Tarbiat Modares University.We studied different distances (1000, 1500, 2000, 2500, and 10,000 m [control]) in an easterly direction from the gas refinery. The map of study area was drawn by software of ArcGIS version of 10.5, https://desktop.arcgis.com (Fig. 5). At each distance, three soil samples taken from the depth of 0–20 cm with a plastic gardening shovel, 30 healthy and mature leaves were collected from a certain height (nearly the middle of the canopy) and the outer canopy of three Brant oak trees in the late spring, summer, and autumn of 2019. These trees with average height and diameter at breast height of 5.5 m and 45 cm were selected randomly. The leaf and soil samples were put into polyethylene bags and transported to the laboratory for analysis36.Figure 5Locations of collection sites of soil samples and Brant oak leaves at five different distances (1000, 1500, 2000, 2500 and 10,000 m [control]) from the gas refinery (drawn by H. Dadkhah-Aghdash using software of ArcGIS Desktop. version of 10.5. ESRI, California, US. https://desktop.arcgis.com).Full size imageIn the lab, firstly the leaves were categorized into two types: unwashed leaves and leaves washed with ethylenediaminetetraacetic acid (EDTA) solution to remove some atmospheric dusts and particles deposition. The leaf and soil samples were dried for 10 days until they reached a constant weight at lab temperature. The leaves were grinded and homogenized, soils were sieved with ASTM mesh (DAMAVAND, Iran) with a diameter of 2 mm and homogenized.To determine the pH and electrical conductivity (EC) of soils, 2 g of the soil samples were shaken in 10 ml of double-distilled water with a ratio of 1:5; after 1 h, the pH and electrical conductivity (EC) of the solution were measured by a digital pH meter (Fan Azma Gostar Company, Iran) and EC meter (Sartorius, PT-20, USA). The analysis of the particle sizes of the soil was carried out using the hydrometer method and texture class was determined with a soil texture triangle37.According to different U.S.EPA protocols that were modified by following references, the soil and leaf samples were prepared and dissolved. The digestion of soil samples was conducted with a mixture of concentrated HF–HClO4–HNO338. Approximately 0.5 g of dry soil sample was digested with 10 mL of HCl on a hot plate at ~ 180 °C until the solution was reduced to 3 mL. Approximately 5 mL of HF (40%, w/w), 5 mL of HNO3 (63%, w/w), and 3 mL of HClO4 (70%, w/w) were then added and the solution was digested. This process was continued with adding 3 mL of HNO3, 3 mL of HF, and 1 mL of HClO4 until the silicate minerals had fully disappeared. This solution was transferred to a 25 mL volumetric tube, and 1% HNO3 was added to bring the sample up to a constant volume for the element’s determinations. After filtering the digested samples, the concentrations of sulfur (S), arsenic (As), chromium (Cr), copper (Cu), lead (Pb), zinc (Zn), manganese (Mn), and nickel (Ni) were measured via inductively coupled plasma mass spectrometry (ICP-MS,7500 CS, Agilent, US). The procedures of quality assurance and quality control (QA/QC) were performed.To quantify element contents from soil samples, external standards with calibration levels were used. The precision and the repeatability of the analysis were tested on the instrument by analyzing three replicate samples.According to Liang et al.39 leaf samples were acid digested and sieved powder samples were placed in the acid-washed tubes and 10 mL of 65% nitric acid was added to it. The solution was placed at room temperature overnight (12 h) after that, it was placed for 4 h at 100 °C and then 4 h at 140 °C until the solution color was clear. After cooling, the solution was diluted by deionized water to 50 mL and then passed through Whatman filter paper until 25 mL of the filtrate volume was provided. Each sample was digested three times and the average of measurements is reported. Total plant elements were measured by using the ICP-MS (7500 CS, Agilent, US). A control sample was also used beside each sample to determine the background pollution during digestion. To confirm the accuracy of the methodology and to ensure the extraction of trace elements from the leaf samples, the standard solution of each studied elements was used.Measuring of pollution levels of different elements in soils and leavesFor assessment of contamination levels (concentration) of different elements in soils and trees, common indices of pollution including geoaccumulation index (Igeo), pollution index (PI), pollution load index (PLI), enrichment factor of plants (EFplant), bioconcentration factor (BCF), air originated metals (AOM ), metal accumulation index (MAI) were used.Igeo was calculated using the following (Eq. 1):$${text{I}}_{{{text{geo}}}} = log_{2} left[ {{text{C}}_{{text{n}}} / 1.5{text{ B}}_{{text{n}}} } right]$$
(1)
where Cn is the measured concentration of the element n, Bn is the geoaccumulation background for this element and 1.5 is a constant coefficient used to eliminate potential variations in the baseline data40. The Igeo classifies samples into seven grades:  5 for extremely polluted30.The first PI is expressed as (Eq. 2):$${text{PI }} = {text{ C}}_{{text{i}}} /{text{S}}_{{text{i}}}$$
(2)
where Ci is the concentration of element i in the soil (mg kg−1) and Si is the soil quality standard or reference value for element i (mg kg−1). The PLI for different elements is calculated via the (Eq. 3):$${text{PLI}} = left( {{text{PI}}_{{1}} times {text{ PI}}_{{2}} times {text{ PI}}_{{3}} times cdots times {text{PI}}_{{text{n}}} } right)^{{{1}/{text{n}}}}$$
(3)
The PLI of soils is classified as follows: PLI  More

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript. More