Philippa Kaur

Adeleye, M. A., Haberle, S. G., Gallagher, R., Andrew, S. C. & Herbert, A. Nat. Ecol. Evol., https://doi.org/10.1038/s41559-022-01943-4 (2023).Mokany, K. et al. Ecography 2022, e06426 (2022).Article 

Google Scholar 
Violle, C. et al. Oikos 116, 882–892 (2007).Article 

Google Scholar 
Birks, H. J. B. Front. Ecol. Evol. 8, 166 (2020).Article 

Google Scholar 
Reitalu, T. et al. J. Veg. Sci. 26, 911–922 (2015).Article 

Google Scholar 
Brussel, T. & Brewer, S. C. Front. Ecol. Evol. 8, 564609 (2021).Article 

Google Scholar 
van der Sande, M. T. et al. Ecol. Lett. 22, 925–935 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Veeken, A., Santos, M. J., McGowan, S., Davies, A. L. & Schrodt, F. Ecol. Lett. 25, 1937–1951 (2022).Article 
PubMed 
PubMed Central 

Google Scholar 
Suárez-Castro, A. F., Raymundo, M., Bimler, M. & Mayfield, M. M. Ecography 2022, e05844 (2022).Article 

Google Scholar 
Biggs, C. R. et al. Ecosphere 11, e03184 (2020).Article 

Google Scholar  More

To investigate whether vagrancy is associated with geomagnetic disturbance and solar activity, we developed a method for quantifying the relative vagrancy of spatiotemporal records for 152 North American landbird species (nfall = 150, nspring = 124). While vagrancy is often treated as a binary classification (i.e., an individual is either a vagrant or not) and then used as a discrete variable (i.e., a count of total vagrants in an area)16,18, here we calculated it as a continuous variable by combining two large-scale ornithological datasets—captures and encounters of individually marked birds from the USGS Bird Banding Lab (BBL)49 and weekly, species-specific abundance maps for the continental United States from the eBird Status and Trends (hereafter, eBird S&T; via the R package ‘ebirdst’, version 2.1.0)69. Banding records have the advantage over other potential databases of vagrancy records (such as eBird or rare bird lists) in that efforts are long-term, continent-wide, have limited false positives, and have only one record per individual. Additionally, eBird S&T has the advantage over static range maps in that they provide weekly predictions and incorporate relative abundance. With these two data sources, we constructed a species-specific vagrancy value (Fig. S1), that measures the spatiotemporal rarity for every banding record. Inclusion of all banding records rather than just rare records allowed for the analysis of the dispersion of whole species populations, mitigating the potential bias of effort in banding operations (i.e., more vagrant records with greater effort). We then used hierarchical Bayesian random-effects models to estimate the strength of the association between geomagnetic disturbance, solar activity, and avian vagrancy.Species data and inclusionWe considered all full—or partial-migrant landbird species with a breeding, non-breeding, or migratory range in the United States or Canada. To do this, we used species distribution maps accessed through Birds of the World70. Landbird species likely to be caught through banding efforts (excluding species like raptors, nightjars, and swifts) that regularly occur in  > 3 but  10 km. Each banding record included the date, latitude and longitude (and precision), species, and age (if known;71). Banding records were filtered to those captures that occurred during the species-specific migration period as defined by eBird S&T69. eBird S&T approximates stationary and migratory periods by determining when the distribution of whole species population is moving69. Our use of banding records within species-specific eBird S&T migratory periods was designed to maximize the proportion of migrant birds in the analysis, but likely excludes some early and late records of migrating individuals.Banding records of species that underwent taxonomic divisions or aggregations during the study period were eliminated if the date occurred during a period in which the species identity according to modern taxonomy is indeterminate (see Supplement 2). Taxonomic reclassifications were not considered when species divisions/aggregations would only affect records from outside North America, such as the split of a Southern American taxon, Chestnut-collared Swallow (Petrochelidon rufocollaris) from its North American counterpart, Cave Swallow (Petrochelidon fulva). In these cases, we assumed all banding records during the study period were of the North American species. For a full list of periods where species records were excluded, see Supplement 2. Species with  More

Samples collection and preparationFreshwater and sediment samples were collected from 5 irrigation drains (EL-Shikah, EL- Tramsa, EL-Mahrosa, EL-Aslia, and EL-Rawy) located in the geographical area of Qena city, the capital of Qena Governorate, 600 km south of Cairo, (Figs. 1 and 2). 3 sites inside each drain were randomly selected as sampling site; one of these sites represents the outlet of the drain into the Nile River. In addition, one site facing each drain in the main stream of the Nile River was selected to collect freshwater only, thus the total number of samples are 20 freshwater and 15 sediment samples.Figure 1Location map of the area under study (ArcGIS software 10.8.1; ArcGIS Online).Full size imageFigure 2Irrigation drain under study.Full size imagePolyethylene Marinelli beakers with a capacity of 1.4 L are used as collection and measuring containers. The beakers were washed with dilute hydrochloric acid and distilled water before use, filled to brim, and then pressed the tight lid to eliminate the internal air. Drops of HNO3 were added to the samples to prevent the adhesive of radionuclides with bottle walls8.Sediment samples were collected by Ekman grab sediment sampler. The collected samples were dried using electrical oven at a temperature of 105℃ for 24 h, then sieved through 200 mesh size. The dried samples were filled in hermetical sealed 500 ml polyethylene beakers. The prepared water and sediment samples were stored for 4 weeks to reach a secular equilibrium of radium and thorium with their progenies9.Measuring systemsGamma-ray spectrometer consisting of ″3 × 3″ NaI (Tl) detector enclosed in 5 cm thick cylindrical lead shield to reduce the background radiation and connected with 1024 multichannel analyzer was used. The spectrometer was calibrated for energy using 60Co and 137Cs standard point sources, and calibrated for efficiency using a multi-nuclides standard solution which covers a wide range of energy10. The spectrum was accumulated from each sample over 24 h and analyzed by Maestro software. The background was measured under the same condition of sample measurement.226Ra was determined using 214Bi and 214Pb gamma-lines at 609 keV and 352 keV, respectively, while 232Th from gamma-lines of 228Ac (911 keV) and 212Pb (238 keV). 40K was determined from its single gamma-line at 1460 keV. The activity concentration was calculated using the following formula (Eq. 1)11.$$A = frac{{C_{n} }}{{T times varepsilon { } times {text{P}} times {text{V }}left( {{text{or}}} right){text{M}}}}$$
(1)

where A is the activity concentration (Bq kg−1) or (Bq l−1), Cn is the net counts under a given peak area, T the sample counting time, (varepsilon) is the detection efficiency at measured energy, P is the emission probability and V is the sample volume in liter, M is the sample mass in kilogram. Minimum detectable activity (MDA) was estimated according to Currie definition using Eq. 212 and the MDA values were 0.031, 0.035 and 1.94 Bq L−1 for 226Ra, 232Th, and 40K, respectively.$${text{MDA}} = frac{2.71 + 465sqrt B }{{T times varepsilon times P times V}}$$
(2)

where B is the background counts under a given peak area,T,ɛ, P, and V are defined above.Doses for aquatic organismsThe external and internal absorbed dose rate for aquatic organisms (Phytoplankton, Mollusca, and Crustacean) in the studied irrigation drains was calculated based on the measured activity concentrations of 226Ra, 232Th, and 40K in environmental media (water and sediment) and using dose conversion coefficients of a given radionuclide for the reference organisms according to the method outlined by Brown et al. described below13,14.$$begin{aligned}& left( {Sediment,, conc. ,,wet} right)_{radionuclide} = (Sediment ,,conc. ,,dry)_{radionuclide} times left( {solids ,,fraction} right) \& qquad qquad + (water ,,conc.)_{radionuclide} times (1 – left( {solids ,,fraction} right). \ end{aligned}$$
(3)
$$begin{aligned}& left( {user2{External ,,dose ,,rate}} right)_{radionuclide,, organism} = DPUC_{radionuclide, ,organism}^{external} times left[ {Sediment ,conc. ,wet_{radionuclide} times left( {fsed_{organism} + fsedsur_{organism} /2} right)} right. \& quad quad left. { + left( {fwater_{organism} + fsedsur_{organism} /2} right) times water ,conc._{radionuclide } /1000} right] \ end{aligned}$$
(4)
$$left( {user2{Internal,dose,rate}} right)_{{radionuclide,,organism}} = ~left( {water,conc.} right)_{{radionuclide}} times CF_{{radionuclide}}^{{organism}} times DPUC_{{radionuclide,,organism}}^{{internal}}$$
(5)

where sediment conc. is the sediment activity concentration of a given radionuclide in Bq kg−1,water conc. is the water activity concentration of a given radionuclide in Bq m−3, CF is distribution coefficient factors for given radionuclide in freshwater sediment in m3 kg−1, DPUC is the dose rate per unit concentration coefficients (fresh weight) in μGy h−1 per Bq kg−1 weighted for radiation type (alpha = 10, low energy beta = 3, and high energy beta and gamma = 1), solids fraction of wet sediment (0.4), fsed organism is the time fraction spends by organism in sediment, fsedsur organism is the time fraction spends by organism at the sediment/water interface, fwater organism is the time fraction spends by organism in the water column. All parameters used in calculation are taken from Pröhl (2003)15 and Vives i Battle et al. (2004)16. The total dose is then calculated by summating the external and internal doses. More

Wingard, G. L., Bernhardt, C. E. & Wachnicka, A. H. The role of paleoecology in restoration and resource management—the past as a guide to future decision-making: review and example from the Greater Everglades ecosystem, U.S.A. Front. Ecol. Evol 5, 11 (2017).Article 

Google Scholar 
Gillson, L., Dirk, C. & Gell, P. Using long-term data to inform a decision pathway for restoration of ecosystem resilience. Anthropocene 36, 100315 (2021).Article 

Google Scholar 
Nieto-Lugilde, D. et al. Time to better integrate paleoecological research infrastructures with neoecology to improve understanding of biodiversity long-term dynamics and to inform future conservation. Environ. Res. Lett. 16, 095005 (2021).Article 

Google Scholar 
Leo, G. A. D. & Levin, S. A. The multifaceted aspects of ecosystem integrity. Conserv. Ecol. 1, 3 (1997).
Google Scholar 
Mason, N. & Mouillot, D. in Encyclopedia of Biodiversity (ed. Levin, S. A.) 597–608 (Elsevier, 2013).Carvalho, F. et al. A method for reconstructing temporal changes in vegetation functional trait composition using Holocene pollen assemblages. PLoS ONE 14, e0216698 (2019).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Brussel, T. & Brewer, S. C. Functional paleoecology and the pollen-plant functional trait linkage. Front. Ecol. Evol 8, 564609 (2021).Article 

Google Scholar 
Brussel, T., Minckley, T. A., Brewer, S. C. & Long, C. J. Community-level functional interactions with fire track long-term structural development and fire adaptation. J. Veg. Sci. 29, 450–458 (2018).Article 

Google Scholar 
Barboni, D. et al. Relationships between plant traits and climate in the Mediterranean region: a pollen data analysis. J. Veg. Sci. 15, 635–646 (2004).Article 

Google Scholar 
Reitalu, T. et al. Novel insights into post-glacial vegetation change: functional and phylogenetic diversity in pollen records. J. Veg. Sci. 26, 911–922 (2015).Article 

Google Scholar 
Blaus, A. et al. Modern pollen-plant diversity relationships inform palaeoecological reconstructions of functional and phylogenetic diversity in calcareous fens. Front. Ecol. Evol 8, 207 (2020).Article 

Google Scholar 
Morris, J. L. et al. Stable or seral? Fire-driven alternative states in aspen forests of western North America. Biol. Lett. 15, 20190011 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Ordonez, A. & Svenning, J.-C. Greater tree species richness in eastern North America compared to Europe is coupled to denser, more clustered functional trait space filling, not to trait space expansion. Glob. Ecol. Biogeogr. 27, 1288–1299 (2018).Article 

Google Scholar 
van der Sande, M. T. et al. A 7000-year history of changing plant trait composition in an Amazonian landscape; the role of humans and climate. Ecol. Lett. 22, 925–935 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Lacourse, T. & Adeleye, M. A. Climate and species traits drive changes in Holocene forest composition along an elevation gradient in Pacific Canada. Front. Ecol. Evol 10, 838545 (2022).Article 

Google Scholar 
Lacourse, T. Environmental change controls postglacial forest dynamics through interspecific differences in life-history traits. Ecology 90, 2149–2160 (2009).Article 
PubMed 

Google Scholar 
Veeken, A., Santos, M. J., McGowan, S., Davies, A. L. & Schrodt, F. Pollen-based reconstruction reveals the impact of the onset of agriculture on plant functional trait composition. Ecol. Lett. 25, 1937–1951 (2022).Article 
PubMed 
PubMed Central 

Google Scholar 
Ellis, E. C., Antill, E. C. & Kreft, H. All is not loss: plant biodiversity in the anthropocene. PLoS ONE 7, e30535 (2012).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
GILL, A. M. Fire and the Australian flora: a review. Aust. For. 38, 4–25 (1975).Article 

Google Scholar 
Crisp, M. D., Burrows, G. E., Cook, L. G., Thornhill, A. H. & Bowman, D. M. J. S. Flammable biomes dominated by eucalypts originated at the Cretaceous–Palaeogene boundary. Nat. Commun. 2, 193 (2011).Article 
PubMed 

Google Scholar 
Keith, D. A. Australian Vegetation (Cambridge Univ. Press, 2017).Woinarski, J. C. Z., Burbidge, A. A. & Harrison, P. L. Ongoing unraveling of a continental fauna: decline and extinction of Australian mammals since European settlement. Proc. Natl Acad. Sci. USA 112, 4531–4540 (2015).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Broadhurst, L. & Coates, D. Plant conservation in Australia: current directions and future challenges. Plant Divers. 39, 348–356 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Adeleye, M. A., Connor, S. E., Haberle, S. G., Herbert, A. & Brown, J. European colonization and the emergence of novel fire regimes in southeast Australia. Anthr. Rev. https://doi.org/10.1177/205301962110446 (2021).Gallagher, R. V. et al. High fire frequency and the impact of the 2019–2020 megafires on Australian plant diversity. Divers. Distrib. 27, 1166–1179 (2021).Article 

Google Scholar 
Gallagher, R. V. et al. An integrated approach to assessing abiotic and biotic threats to post-fire plant species recovery: lessons from the 2019–2020 Australian fire season. Glob. Ecol. Biogeogr. 31, 2056–2069.Mariani, M. et al. Disruption of cultural burning promotes shrub encroachment and unprecedented wildfires. Front. Ecol. Environ. 20, 292–300 (2022).Article 

Google Scholar 
Williams, A. N., Mooney, S. D., Sisson, S. A. & Marlon, J. Exploring the relationship between Aboriginal population indices and fire in Australia over the last 20,000 years. Palaeogeogr. Palaeoclimatol. Palaeoecol. 432, 49–57 (2015).Article 

Google Scholar 
Bird, M. I., O’Grady, D. & Ulm, S. Humans, water, and the colonization of Australia. Proc. Natl Acad. Sci. USA 113, 11477–11482 (2016).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Adeleye, M. A., Haberle, S. G., Connor, S. E., Stevenson, J. & Bowman, D. M. J. S. Indigenous fire-managed landscapes in Southeast Australia during the Holocene—new insights from the Furneaux Group Islands, Bass Strait. Fire 4, 17 (2021).Article 

Google Scholar 
Fletcher, M.-S., Romano, A., Connor, S., Mariani, M. & Maezumi, S. Y. Catastrophic bushfires, Indigenous fire knowledge and reframing science in Southeast Australia. Fire 4, 61 (2021).Article 

Google Scholar 
Fletcher, M.-S., Hall, T. & Alexandra, A. N. The loss of an indigenous constructed landscape following British invasion of Australia: an insight into the deep human imprint on the Australian landscape. Ambio 50, 138–149 (2021).Article 
CAS 
PubMed 

Google Scholar 
Adeleye, M. A. et al. Long-term drivers of vegetation turnover in Southern Hemisphere temperate ecosystems. Glob. Ecol. Biogeogr. 30, 557–571 (2021).Article 

Google Scholar 
Kershaw, A. P., D’Costa, D. M., McEwen Mason, J. R. C. & Wagstaff, B. E. Palynological evidence for Quaternary vegetation and environments of mainland southeastern Australia. Quat. Sci. Rev. 10, 391–404 (1991).Article 

Google Scholar 
Colhoun, E. A. & Shimeld, P. W. in Peopled Landscapes: Archaeological and Biogeographic Approaches to Landscapes (eds. Haberle, S. G. & David, B.) 297–328 (ANU Press, 2012).Madani, N. et al. Future global productivity will be affected by plant trait response to climate. Sci. Rep. 8, 2870 (2018).Article 
PubMed 
PubMed Central 

Google Scholar 
Squire, D. T. et al. Likelihood of unprecedented drought and fire weather during Australia’s 2019 megafires. npj Clim. Atmos. Sci. 4, 64 (2021).Article 

Google Scholar 
Ukkola, A. M., De Kauwe, M. G., Roderick, M. L., Abramowitz, G. & Pitman, A. J. Robust future changes in meteorological drought in CMIP6 projections despite uncertainty in precipitation. Geophys. Res. Lett. 47, e2020GL087820 (2020).Article 

Google Scholar 
Mori, A. S., Furukawa, T. & Sasaki, T. Response diversity determines the resilience of ecosystems to environmental change. Biol. Rev. 88, 349–364 (2013).Article 
PubMed 

Google Scholar 
Arias, P. A. et al. In Climate Change 2021: The Physical Science Basis (eds. Masson-Delmotte, V. et al.) (Cambridge Univ. Press, 2021).Laliberté, E., Legendre, P. & Shipley, B. FD: measuring (FD) from multiple traits, and other tools for functional ecology. R package version 1.0-12 (2014).Laliberté, E. & Legendre, P. A distance-based framework for measuring from multiple traits. Ecology 91, 299–305 (2010).Article 
PubMed 

Google Scholar 
Tilman, D. in Encyclopedia of Biodiversity (ed. Levin, S. A.) 109–120 (Elsevier, 2001).Fletcher, M.-S. & Moreno, P. I. Have the Southern Westerlies changed in a zonally symmetric manner over the last 14,000 years? A hemisphere-wide take on a controversial problem. Quat. Int. 253, 32–46 (2012).Article 

Google Scholar 
Markgraf, V., Bradbury, J. P. & Busby, J. R. Paleoclimates in Southwestern Tasmania during the last 13,000 years. PALAIOS 1, 368 (1986).Article 

Google Scholar 
Moros, M. et al. Hydrographic shifts south of Australia over the last deglaciation and possible interhemispheric linkages. Quat. Res. 102, 130–141 (2021).Article 

Google Scholar 
Perner, K. et al. Heat export from the tropics drives mid to late Holocene palaeoceanographic changes offshore southern Australia. Quat. Sci. Rev. 180, 96–110 (2018).Article 

Google Scholar 
Mariani, M. & Fletcher, M.-S. Long-term climate dynamics in the extra-tropics of the South Pacific revealed from sedimentary charcoal analysis. Quat. Sci. Rev. 173, 181–192 (2017).Article 

Google Scholar 
McWethy, D. B. et al. A conceptual framework for predicting temperate ecosystem sensitivity to human impacts on fire regimes. Glob. Ecol. Biogeogr. 22, 900–912 (2013).Article 

Google Scholar 
Baker, A. G., Catterall, C. & Benkendorff, K. Invading rain forest pioneers initiate positive fire suppression feedbacks that reinforce shifts from open to closed forest in eastern Australia. J. Veg. Sci. 32, e13102 (2021).Article 

Google Scholar 
Lambeck, K. & Chappell, J. Sea level change through the last glacial cycle. Science 292, 679–686 (2001).Article 
CAS 
PubMed 

Google Scholar 
Sloss, C. R., Murray-Wallace, C. V. & Jones, B. G. Holocene sea-level change on the southeast coast of Australia: a review. Holocene 17, 999–1014 (2007).Article 

Google Scholar 
Adeleye, M. A. et al. Holocene heathland development in temperate oceanic Southern Hemisphere: key drivers in a global context. J. Biogeogr. 48, 1048–1062 (2021).Article 

Google Scholar 
McWethy, D. B., Haberle, S. G., Hopf, F. & Bowman, D. M. J. S. Aboriginal impacts on fire and vegetation on a Tasmanian island. J. Biogeogr. 44, 1319–1330 (2017).Article 

Google Scholar 
Hope, G. Vegetation and fire response to late Holocene human occupation in island and mainland north west Tasmania. Quat. Int. 59, 47–60 (1999).Article 

Google Scholar 
Sim, R. The Archaeology of Isolation? Prehistoric Occupation in the Furneaux Group of Islands, Bass Strait, Tasmania. PhD thesis, Australian National Univ. (1998).Lourandos, H. Intensification: a late Pleistocene-Holocene archaeological sequence from Southwestern Victoria. Archaeol. Ocean. 18, 81–94 (1983).Article 

Google Scholar 
Bowman, D. M. J. S. The impact of Aboriginal landscape burning on the Australian biota. New Phytol. 140, 385–410 (1998).Article 
CAS 
PubMed 

Google Scholar 
Iversen, J. in Systematics of Today (ed. Hedberg, O.) 210–215 (Acta Universitatis Upsaliensis/Uppsala Universitets Årsskrift, 1958).Colhoun, E. A. Application of Iversen’s glacial–interglacial cycle to interpretation of the late last glacial and Holocene vegetation history of western Tasmania. Quat. Sci. Rev. 15, 557–580 (1996).Article 

Google Scholar 
Adeleye, M. A., Haberle, S. G., Ondei, S. & Bowman, D. M. J. S. Ecosystem transformation following the mid-nineteenth century cessation of Aboriginal fire management in Cape Pillar, Tasmania. Reg. Environ. Change 22, 99 (2022).Article 

Google Scholar 
Mccann, K. The diversity–stability debate. Nature 405, 228–233 (2000).Article 
CAS 
PubMed 

Google Scholar 
Hallett, L. M., Stein, C. & Suding, K. N. Functional diversity increases ecological stability in a grazed grassland. Oecologia 183, 831–840 (2017).Article 
PubMed 

Google Scholar 
Bello, Fde et al. Functional trait effects on ecosystem stability: assembling the jigsaw puzzle. Trends Ecol. Evol. 36, 822–836 (2021).Article 
PubMed 

Google Scholar 
Lucini, F. A., Morone, F., Tomassone, M. S. & Makse, H. A. Diversity increases the stability of ecosystems. PLoS ONE 15, e0228692 (2020).Article 

Google Scholar 
Tilman, D., Reich, P. B. & Knops, J. M. H. Biodiversity and ecosystem stability in a decade-long grassland experiment. Nature 441, 629–632 (2006).Article 
CAS 
PubMed 

Google Scholar 
Gallagher, R. V., Hughes, L. & Leishman, M. R. Species loss and gain in communities under future climate change: consequences for functional diversity. Ecography 36, 531–540 (2013).Article 

Google Scholar 
Song, Y., Wang, P., Li, G. & Zhou, D. Relationships between and ecosystem functioning: a review. Acta Ecol. Sin. 34, 85–91 (2014).Article 
CAS 

Google Scholar 
Li, W. et al. Plant can be independent of species diversity: observations based on the impact of 4-yrs of nitrogen and phosphorus additions in an alpine meadow. PLoS ONE 10, e0136040 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
Lewis, C. J., Huang, Y., Siems, S. T. & Manton, M. J. Wintertime orographic precipitation over western Tasmania. J. South. Hemisphere Earth Syst. Sci. 68, 22–40 (2018).Article 

Google Scholar 
Andrew, S. C. et al. Functional diversity of the Australian flora: strong links to species richness and climate. J. Veg. Sci. 32, e13018 (2021).Article 

Google Scholar 
Biswas, S. R. & Mallik, A. U. Species diversity and relationship varies with disturbance intensity. Ecosphere 2, art52 (2011).Article 

Google Scholar 
Gallagher, R. V. et al. A guide to using species trait data in conservation. One Earth 4, 927–936 (2021).Article 

Google Scholar 
Siefert, A. et al. A global meta-analysis of the relative extent of intraspecific trait variation in plant communities. Ecol. Lett. 18, 1406–1419 (2015).Article 
PubMed 

Google Scholar 
Harris, S. & Kitchener, A. From Forest to Fjaeldmark. Descriptions of Tasmania’s Vegetation (Department of Primary Industries, Water and Environment, Tasmania, 2005).Adeleye, M. A., Haberle, S. G., McWethy, D., Connor, S. E. & Stevenson, J. Environmental change during the last glacial on an ancient land bridge of southeast Australia. J. Biogeogr. 48, 2946–2960 (2021).Article 

Google Scholar 
Hopf, F. V. L., Colhoun, E. A. & Barton, C. E. Late-glacial and Holocene record of vegetation and climate from Cynthia Bay, Lake St Clair, Tasmania. J. Quat. Sci. 15, 725–732 (2000).Article 

Google Scholar 
Stahle, L. N., Whitlock, C. & Haberle, S. G. A 17,000-year-long record of vegetation and fire from Cradle Mountain National Park, Tasmania. Front. Ecol. Evol 4, 82 (2016).Article 

Google Scholar 
Michael-Shawn, F. et al. The influence of climatic change, fire and species invasion on a Tasmanian temperate rainforest system over the past 18,000 years. Quat. Sci. Rev. 260, 106824 (2021).Article 

Google Scholar 
Climate and Water Availability in South-Eastern Australia: A Synthesis of Findings From Phase 2 of the South-Eastern Australian Climate initiative (SEACI) (CSIRO, 2012); https://doi.org/10.4225/08/584af3986fe96Australian Climate Influences (Commonwealth of Australia, Bureau of Meteorology, 2010); http://www.bom.gov.au/watl/about-weather-and-climate/australian-climate-influences.shtmlRisbey, J. S., Pook, M. J., McIntosh, P. C., Wheeler, M. C. & Hendon, H. H. On the remote drivers of rainfall variability in Australia. Mon. Weather Rev. 137, 3233–3253 (2009).Article 

Google Scholar 
Mariani, M., Fletcher, M.-S., Holz, A. & Nyman, P. ENSO controls interannual fire activity in southeast Australia. Geophys. Res. Lett. 43, 10891–10900 (2016).Article 

Google Scholar 
Mariani, M. & Fletcher, M.-S. The Southern Annular Mode determines interannual and centennial-scale fire activity in temperate southwest Tasmania, Australia. Geophys. Res. Lett. 43, 1702–1709 (2016).Article 

Google Scholar 
Herbert, A. V. & Harrison, S. P. Evaluation of a modern-analogue methodology for reconstructing Australian palaeoclimate from pollen. Rev. Palaeobot. Palynol. 226, 65–77 (2016).Article 

Google Scholar 
Blaauw, M. et al. rbacon: Age-depth modelling using Bayesian statistics. R package version 4.2.0 (2022).Hogg, A. G. et al. SHCal20 Southern Hemisphere calibration, 0–55,000 years cal BP. Radiocarbon 62, 759–778 (2020).Article 
CAS 

Google Scholar 
Falster, D. et al. AusTraits, a curated plant trait database for the Australian flora. Sci. Data 8, 254 (2021).Article 
PubMed 
PubMed Central 

Google Scholar 
Pérez-Harguindeguy, N. et al. Corrigendum to: New handbook for standardised measurement of plant functional traits worldwide. Aust. J. Bot. 64, 715–716 (2016).Article 

Google Scholar 
Wright, I. J. et al. Global climatic drivers of leaf size. Science 357, 917–921 (2017).Article 
CAS 
PubMed 

Google Scholar 
Wright, I. J. et al. The worldwide leaf economics spectrum. Nature 428, 821–827 (2004).Article 
CAS 
PubMed 

Google Scholar 
Westoby, M., Falster, D. S., Moles, A. T., Vesk, P. A. & Wright, I. J. Plant ecological strategies: some leading dimensions of variation between species. Annu. Rev. Ecol. Syst. 33, 125–159 (2002).Article 

Google Scholar 
Moles, A. T. & Westoby, M. Seed size and plant strategy across the whole life cycle. Oikos 113, 91–105 (2006).Article 

Google Scholar 
Leishman, M. R. & Westoby, M. The role of seed size in seedling establishment in dry soil conditions—experimental evidence from semi-arid species. J. Ecol. 82, 249–258 (1994).Article 

Google Scholar 
Falster, D. S. & Westoby, M. Plant height and evolutionary games. Trends Ecol. Evol. 18, 337–343 (2003).Article 

Google Scholar 
Díaz, S. et al. The global spectrum of plant form and function. Nature 529, 167–171 (2016).Article 
PubMed 

Google Scholar 
Mason, N. W. H., Mouillot, D., Lee, W. G. & Wilson, J. B. Functional richness, functional evenness and functional divergence: the primary components of functional diversity. Oikos 111, 112–118 (2005).Article 

Google Scholar 
Pakeman, R. J. Functional trait metrics are sensitive to the completeness of the species’ trait data? Methods Ecol. Evol. 5, 9–15 (2014).Article 

Google Scholar 
Scheiner, S. M., Kosman, E., Presley, S. J. & Willig, M. R. Decomposing. Methods Ecol. Evol. 8, 809–820 (2017).Article 

Google Scholar 
Ripley, B. et al. MASS: Support functions and datasets for venables and Ripley’s MASS. R package version ??? (2022).Moy, C. M., Seltzer, G. O., Rodbell, D. T. & Anderson, D. M. Variability of El Niño/Southern Oscillation activity at millennial timescales during the Holocene epoch. Nature 420, 162–165 (2002).Article 
CAS 
PubMed 

Google Scholar  More

Zupan Hajna, N. Dinaric karst: Geography and geology in Encyclopedia of Caves (eds. White, W. B. & Culver, D. C.) 195–203 (Academic Press, 2012).Jug-Dujaković, M., Ninčević, T., Liber, Z., Grdiša, M. & Šatović, Z. Salvia officinalis survived in situ Pleistocene glaciation in ‘refugia within refugia’ as inferred from AFLP markers. Plant Syst. Evol. 306, 1–12 (2020).Article 

Google Scholar 
Bănărescu, P. M. Distribution pattern of the aquatic fauna of the Balkan Peninsula in Balkan Biodiversity. Pattern and Process in the European Hotspot (eds. Griffiths, H. I., Kryštufek, B. & Reed J. M.) 203–217 (Kluwer Academic Publishers, 2004).Sket, B. Diversity patterns in the Dinaric Karst in Encyclopedia of Caves (eds. White, W. B. & Culver, D. C.) 228–238 (Academic Press, 2012).Griffiths, H. I., Kryštufek, B., & Reed, J. M. Balkan biodiversity. Pattern and Process in the European Hotspot (eds. Griffiths, H. I., Kryštufek, B., & Reed, J. M.) 1–332 (Kluwer Academic Publishers, 2004).Culver, D. C., Pipan, T. & Schneider, K. Vicariance, dispersal and scale in the aquatic subterranean fauna of karst regions. Freshw. Biol. 54, 918–929 (2009).Article 

Google Scholar 
Gottstein Matočec, S. et al. An overview of the cave and interstitial biota of Croatia. Nat. Croat. 11, 1–112 (2002).
Google Scholar 
Hewitt, G. The genetic legacy of the Quaternary ice ages. Nature 405, 907–913 (2000).Article 
ADS 

Google Scholar 
Bilandžija, H., Morton, B., Podnar, M. & Ćetković, H. Evolutionary history of relict Congeria (Bivalvia: Dreissenidae): Unearthing the subterranean biodiversity of the Dinaric Karst. Front. Zool. 10, 1–18 (2013).Article 

Google Scholar 
Bedek, J., Taiti, S., Bilandžija, H., Ristori, E. & Baratti, M. Molecular and taxonomic analyses in troglobiotic Alpioniscus (Illyrionethes) species from the Dinaric Karst (Isopoda: Trichoniscidae). Zool. J. Linn. Soc. 187, 539–584 (2019).Article 

Google Scholar 
Vörös, J., Márton, O., Schmidt, B. R., Gál, J. T. & Jelić, D. Surveying Europe’s only cave-dwelling chordate species (Proteus anguinus) using environmental DNA. PLoS ONE 12, e0170945. https://doi.org/10.1371/journal.pone.0170945 (2017).Article 

Google Scholar 
Delić, T., Švara, V., Coleman, C. O., Trontelj, P. & Fišer, C. The giant cryptic amphipod species of the subterranean genus Niphargus (Crustacea, Amphipoda). Zool. Scr. 46, 740–752 (2017).Article 

Google Scholar 
Delić, T., Trontelj, P., Rendoš, M. & Fišer, C. The importance of naming cryptic species and the conservation of endemic subterranean amphipods. Sci. Rep. 7, 1–12 (2017).Article 

Google Scholar 
Delić, T., Stoch, F., Borko, Š., Flot, J. F. & Fišer, C. How did subterranean amphipods cross the Adriatic Sea? Phylogenetic evidence for dispersal–vicariance interplay mediated by marine regression–transgression cycles. J. Biogeogr. 47, 1875–1887 (2020).Article 

Google Scholar 
Podnar, M., Grbac, I., Tvrtković, N., Hörweg, C. & Haring, E. Hidden diversity, ancient divergences, and tentative Pleistocene microrefugia of European scorpions (Euscorpiidae: Euscorpiinae) in the eastern Adriatic region. J. Zool. Syst. Evol. Res. 59, 1824–1849 (2021).Article 

Google Scholar 
Beron, P. Zoogeography of Arachnida (ed. Beron, P.) Meth. Ecol. Evol. 1–987 (Springer Cham, 2018).Ćurčić, B. P. M. Cave-dwelling pseudoscorpions of the Dinaric karst (ed. Ćurčić, B. P. M.) 1–192 (Slovenska Akademija Znanosti in Umetnosti, 1988).Harms, D., Roberts, J. D. & Harvey, M. S. Climate variability impacts on diversification processes in a biodiversity hotspot: A phylogeography of ancient pseudoscorpions in south-western Australia. Zool. J. Linn. Soc. 186, 934–949 (2019).Article 

Google Scholar 
Muster, C., Schmarda, T. & Blick, T. Vicariance in a cryptic species pair of European pseudoscorpions (Arachnida, Pseudoscorpiones, Chthoniidae). Zool. Anz. 242, 299–311 (2004).Article 

Google Scholar 
Ozimec, R. List of Croatian pseudoscorpion fauna (Arachnida, Pseudoscorpiones). Nat. Croat. 13, 381–394 (2004).
Google Scholar 
World Pseudoscorpiones Catalog. Natural History Museum Bern. https://wac.nmbe.ch (2022).Ćurčić, B. P. M., Dimitrijević, R. N., Rađa, T., Makarov, S. E. & Ilić, B. S. Archaeoroncus, a new genus of pseudoscorpions from Croatia (Pseudoscorpiones, Neobisiidae), with descriptions of two new species. Acta Zool. Bulg. 64, 333–340 (2012).
Google Scholar 
Ćurčić, B. P. M. et al. On two new cave species of pseudoscorpions (Neobisiidae, Pseudoscorpiones) from Herzegovina and Dalmatia. Arch. Biol. Sci. 66, 377–384 (2014).Article 

Google Scholar 
Ćurčić, B. P. M. et al. Roncus sutikvae sp. n. (Pseudoscorpiones: Neobisiidae), a new epigean pseudoscorpion from central Dalmatia (Croatia). Arthropoda Sel. 30, 205–215 (2021).Article 

Google Scholar 
Ćurčić, B. P. M., Rađa, T., Dimitrijević, R., Ćurčić, N. B. & Ćurčić, S. Roncus ladestani sp. n. and Roncus pecmliniensis sp. n., two new Pseudoscorpions (Pseudoscorpiones, Neobisiidae) from Croatia and Bosnia and Herzegovina, respectively. Zool. Zhurnal. 100, 159–169 (2021).
Google Scholar 
Hebert, P. D. N., Cywinska, A., Ball, S. L. & DeWaard, J. R. Biological identifications through DNA barcodes. Proc. Royal Soc. B. 270, 313–321 (2003).Article 

Google Scholar 
Page, R. D. DNA barcoding and taxonomy: Dark taxa and dark texts. Philos. Trans. R. Soc. Lond., B. Biol. Sci. 371, 20150334. https://doi.org/10.1098/rstb.2015.0334 (2016).Article 

Google Scholar 
Ratnasingham, S. & Hebert, P. D. A DNA-based registry for all animal species: The Barcode Index Number (BIN) system. PLoS ONE 8, e66213. https://doi.org/10.1371/journal.pone.0066213 (2013).Article 
ADS 

Google Scholar 
Kekkonen, M. & Hebert, P. D. DNA barcode-based delineation of putative species: Efficient start for taxonomic workflows. Mol. Ecol. Res. 14, 706–715 (2014).Article 

Google Scholar 
Christophoryová, J., Šťáhlavský, F. & Fedor, P. An updated identification key to the pseudoscorpions (Arachnida: Pseudoscorpiones) of the Czech Republic and Slovakia. Zootaxa 2876, 35–48 (2011).Article 

Google Scholar 
Gardini, G. A revision of the species of the pseudoscorpion subgenus Chthonius (Ephippiochthonius) (Arachnida, Pseudoscorpiones, Chthoniidae) from Italy and neighbouring areas. Zootaxa 3655, 1–151 (2013).Article 

Google Scholar 
Gardini, G. The species of the Chthonius heterodactylus group (Arachnida, Pseudoscorpiones, Chthoniidae) from the eastern Alps and the Carpathians. Zootaxa 3887, 101–137 (2014).Article 

Google Scholar 
Gardini, G. The Italian species of the Chthonius ischnocheles group (Arachnida, Pseudoscorpiones, Chthoniidae), with reference to neighbouring countries. Zootaxa 4987, 1–131 (2021).Article 

Google Scholar 
Zaragoza, J. A. Revision of the Ephippiochthonius complex in the Iberian Peninsula, Balearic Islands and Macaronesia, with proposed changes to the status of the Chthonius subgenera (Pseudoscorpiones, Chthoniidae). Zootaxa 4246, 1–221 (2017).Article 

Google Scholar 
Gams, I. Kras v Sloveniji v prostoru in času. (ed. Gams, I.) 1–516 (Postojna: Inštitut za raziskovanje Krasa, 2004).European Union, Copernicus Land Monitoring Service. https://land.copernicus.eu (2016).Maddison, W. P., & Maddison, D. R. Mesquite: a modular system for evolutionary analysis. http://mesquiteproject.org (2019).Katoh, K., Rozewicki, J. & Yamada, K. D. MAFFT online service: Multiple sequence alignment, interactive sequence choice and visualization. Brief. Bioinform. 20, 1160–1166 (2019).Article 

Google Scholar 
Villesen, P. FaBox: An online toolbox for fasta sequences. Mol. Ecol. Notes. 7, 965–968 (2007).Article 

Google Scholar 
Felsenstein, J. Maximum likelihood and minimum-steps methods for estimating evolutionary trees from data on discrete characters. Syst. Biol. 22, 240–249 (1973).Article 

Google Scholar 
Minh, B. Q. et al. IQ-TREE 2: New models and efficient methods for phylogenetic inference in the genomic era. Mol. Biol. Evol. 37, 1530–1534 (2020).Article 

Google Scholar 
Hoang, D. T., Chernomor, O., Von Haeseler, A., Minh, B. Q. & Vinh, L. S. UFBoot2: Improving the ultrafast bootstrap approximation. Mol. Biol. Evol. 35, 518–522 (2018).Article 

Google Scholar 
Muster, C. et al. The dark side of pseudoscorpion diversity: The German Barcode of Life campaign reveals high levels of undocumented diversity in European false scorpions. Ecol. Evol. 11, 13815–13829 (2021).Article 

Google Scholar 
Ontano, A. Z. et al. Taxonomic sampling and rare genomic changes overcome long-branch attraction in the phylogenetic placement of pseudoscorpions. Mol. Biol. Evol. 38, 2446–2467 (2021).Article 

Google Scholar 
Rambaut A. FigTree v1.4.3 http://tree.bio.ed.ac.uk/software/figtree/ (2016).Letunic, I. & Bork, P. Interactive Tree Of Life (iTOL) v5: An online tool for phylogenetic tree display and annotation. Nucleic Acids Res. 49, W293–W296 (2021).Article 

Google Scholar 
Talavera, G. & Castresana, J. Improvement of phylogenies after removing divergent and ambiguously aligned blocks from protein sequence alignments. Syst. Biol. 56, 564–577 (2007).Article 

Google Scholar 
Guindon, S. et al. New algorithms and methods to estimate maximum-likelihood phylogenies: Assessing the performance of PhyML 3.0. Syst. Biol. 59, 307–321 (2010).Article 

Google Scholar 
Ronquist, F. et al. MrBayes 3.2: Efficient Bayesian phylogenetic inference and model choice across a large model space. Syst. Biol. 61, 539–542 (2012).Article 

Google Scholar 
Lanfear, R., Frandsen, P. B., Wright, A. M., Senfeld, T. & Calcott, B. PartitionFinder 2: New methods for selecting partitioned models of evolution for molecular and morphological phylogenetic analyses. Mol. Biol. Evol. 34, 772–773 (2017).
Google Scholar 
Rambaut, A., Drummond, A. J., Xie, D., Baele, G. & Suchard, M. A. Posterior summarization in Bayesian phylogenetics using Tracer 1.7. Syst. Biol. 67, 901–904 (2018).Article 

Google Scholar 
Miller, M. A., Pfeiffer, W., & Schwartz, T. Creating the CIPRES Science Gateway for inference of large phylogenetic trees in Proceedings of the Gateway Computing Environments Workshop (GCE). https://doi.org/10.1109/GCE.2010.5676129 (2010).Kimura, M. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16, 111–120 (1980).Article 
ADS 

Google Scholar 
Paradis, E. & Schliep, K. ape 5.0: An environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics 35, 526–528 (2019).Article 

Google Scholar 
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/ (2020).Brown, S. D. et al. Spider: an R package for the analysis of species identity and evolution, with particular reference to DNA barcoding. Mol. Ecol. Resour. 12, 562–565 (2012).Article 

Google Scholar 
Meier, R., Shiyang, K., Vaidya, G. & Ng, P. K. L. DNA barcoding and taxonomy in Diptera: A tale of high intraspecific variability and low identification success. Syst. Biol. 55, 715–728 (2006).Article 

Google Scholar 
Puillandre, N., Lambert, A., Brouillet, S. & Achaz, G. J. M. E. ABGD, Automatic Barcode Gap Discovery for primary species delimitation. Mol. Ecol. 21, 1864–1877 (2012).Article 

Google Scholar 
Puillandre, N., Brouillet, S. & Achaz, G. ASAP: Assemble species by automatic partitioning. Mol. Ecol. Resour. 21, 609–620 (2020).Article 

Google Scholar 
Zhang, J., Kapli, P., Pavlidis, P. & Stamatakis, A. A general species delimitation method with applications to phylogenetic placements. Bioinformatics 29, 2869–2876 (2013).Article 

Google Scholar 
Kumar, S., Stecher, G., Li, M., Knyaz, C. & Tamura, K. MEGA X: Molecular evolutionary genetics analysis across computing platforms. Mol. Biol. Evol. 35, 1547–1549 (2018).Article 

Google Scholar 
Karney, C. F. Algorithms for geodesics. J. Geod. 87, 43–55 (2013).Article 
ADS 

Google Scholar 
Leigh, J. W. & Bryant, D. POPART: Full-feature software for haplotype network construction. Meth. Ecol. Evol. 6, 1110–1116 (2015).Article 

Google Scholar 
Bregović, P., Fišer, C. & Zagmajster, M. Contribution of rare and common species to subterranean species richness patterns. Ecol. Evol. 9, 11606–11618 (2019).Article 

Google Scholar 
Hunter, J. D. Matplotlib: A 2D graphics environment. Comput. Sci. Eng. 9, 90–95 (2007).Article 

Google Scholar 
Young, M. R. & Hebert, P. D. Patterns of protein evolution in cytochrome c oxidase 1 (COI) from the class Arachnida. PLoS ONE 10, e0135053. https://doi.org/10.1371/journal.pone.0135053 (2015).Article 

Google Scholar 
Yin, Y. et al. DNA barcoding uncovers cryptic diversity in minute herbivorous mites (Acari, Eriophyoidea). Mol. Ecol. Resour. 22, 1986–1998 (2022).Article 

Google Scholar 
Doña, J. et al. DNA barcoding and minibarcoding as a powerful tool for feather mite studies. Mol. Ecol. Resour. 15, 1216–1225 (2015).Article 

Google Scholar 
Blagoev, G. A. et al. Untangling taxonomy: A DNA barcode reference library for Canadian spiders. Mol. Ecol. Resour. 16, 325–341 (2016).Article 

Google Scholar 
Aliabadian, M., Kaboli, M., Nijman, V. & Vences, M. Molecular identification of birds: Performance of distance-based DNA barcoding in three genes to delimit parapatric species. PLoS ONE 4, e4119. https://doi.org/10.1371/journal.pone.0004119 (2009).Article 
ADS 

Google Scholar 
Moritz, C. & Cicero, C. DNA barcoding: promise and pitfalls. PLoS Biol. 2, e354. https://doi.org/10.1371/journal.pbio.0020354 (2004).Article 

Google Scholar 
Dellicour, S. & Flot, J. F. The hitchhiker’s guide to single-locus species delimitation. Mol. Ecol. Resour. 18, 1234–1246 (2018).Article 

Google Scholar 
Polak, S., Delić, T., Kostanjšek, R. & Trontelj, P. Molecular phylogeny of the cave beetle genus Hadesia (Coleoptera: Leiodidae: Cholevinae: Leptodirini), with a description of a new species from Montenegro. Arthropod Syst. 74, 241–254 (2016).
Google Scholar 
Lukić, M., Delić, T., Pavlek, M., Deharveng, L. & Zagmajster, M. Distribution pattern and radiation of the European subterranean genus Verhoeffiella (Collembola, Entomobryidae). Zool. Scr. 49, 86–100 (2019).Article 

Google Scholar 
Casale, A., Jalžić, B., Lohaj, R. & Mlejnek, R. Two new highly specialised subterranean beetles from the Velebit massif (Croatia): Velebitaphaenops (new genus) giganteus Casale & Jalžić, new species (Coleoptera: Carabidae: Trechini) and Velebitodromus ozrenlukici Lohaj, Mlejnek & Jalžić, new species (Coleoptera: Cholevidae: Leptodirini). Nat. Croat. 21, 129–153 (2012).
Google Scholar 
Andersen, T. et al. Blind flight? A new troglobiotic Orthoclad (Diptera, Chironomidae) from the Lukina Jama-Trojama Cave in Croatia. PLoS ONE 11, e0152884. https://doi.org/10.1371/journal.pone.0152884 (2016).Article 

Google Scholar 
Velić, J. et al. A geological overview of glacial accumulation and erosional occurrences at the Velebit and the Biokovo Mts., Croatia. The Min. Geol. Petrol. Eng. Bull. 32, 77–96 (2017).
Google Scholar 
Bickford, D. et al. Cryptic species as a window on diversity and conservation. Trends Ecol. Evol. 22, 148–155 (2007).Article 

Google Scholar 
Trontelj, P. Adaptation and natural selection in caves in Encyclopedia of Caves (eds. White, W. B., Culver, D. B. & Pipan, T.) 40–46 (Academic Press, 2019).Beier, M. Die Höhlenpseudoscorpione der Balkanhalbinsel. Studien aus dem Gebiete der Allgemeinen Karstforschung, der Wissenschaftlichen Höhlenkunde, der Eiszeitforschung und den Nachbargebieten. 4, 1–83 (1939).
Google Scholar 
Antić, D., Dražina, T., Rađa, T., Tomić, V. T. & Makarov, S. E. Review of the family Anthogonidae (Diplopoda, Chordeumatida), with descriptions of three new species from the Balkan Peninsula. Zootaxa 3948, 151–181 (2015).Article 

Google Scholar 
Pretner, E. Koleopterološka fauna pećina i jama Hrvatske s historijskim pregledom istraživanja. Krš Jugoslavije. 8, 101–239 (1973).
Google Scholar 
Zaragoza, J. A. & Šťáhlavský, F. A new Roncus species (Pseudoscorpiones: Neobisiidae) from Montseny Natural Park (Catalonia, Spain), with remarks on karyology. Zootaxa 1693, 27–40 (2008).Article 

Google Scholar 
Myers, N., Mittermeier, R. A., Mittermeier, C. G., Da Fonseca, G. A. & Kent, J. Biodiversity hotspots for conservation priorities. Nature 403, 853–858 (2000).Article 
ADS 

Google Scholar 
Médail, F. & Diadema, K. Glacial refugia influence plant diversity patterns in the Mediterranean Basin. J. Biogeogr. 36, 1333–1345 (2009).Article 

Google Scholar 
Borko, Š., Trontelj, P., Seehausen, O., Moškrič, A. & Fišer, C. A subterranean adaptive radiation of amphipods in Europe. Nat. Commun. 12, 1–12 (2021).Article 

Google Scholar 
Fišer, C. et al. The European green deal misses Europe’s subterranean biodiversity hotspots. Nat. Ecol. Evol. 6, 1403–1404 (2022).Article 

Google Scholar 
Moritz, C. Defining ‘evolutionarily significant units’ for conservation. Trends Ecol. Evol. 9, 373–375 (1994).Article 

Google Scholar  More

Study species and spider collectionAustralomisidia ergandros is a subsocial spider inhabiting South-Eastern Australia. They live in communal kin-groups in nests usually built with leaves from Eucalyptus trees bound by silk threads. Group size usually ranges from 5 to 45 spiderlings. Groups are comprised of the offspring of a single female who provides maternal care until her death28. Offspring continue to live in groups for 5 to 7 months after the mother’s death29,30. One of the females inherit the natal nest while the remaining females disperse to found new nests. It is not entirely clear if A. ergandros inbreed with natal kin or if spiders show a mandatory pre-mating dispersal.We collected 29 A. ergandros nests from a population along Yass River Road in New South Wales, Australia (34° 55′ 20.50′′ S, 149° 6′ 15.53′′ E) in February 2016. At this time of year, the spiderlings are very young and the presence of immigrants, who might influence the extent of social foraging, is improbable9,31. For our experiments, we transferred the original nests to the laboratory at Macquarie University in Sydney.Group composition effectsOur experiments spanned a duration of 56 days. To investigate group composition (cooperators vs. defectors) effects, we first assessed the hunting types of individuals within ‘initial’ groups (phase 1) and subsequently composed and tested ‘sorted’ groups of cooperators or defectors only (phase 2). The formation of initial groups was dictated by special requirements. Basically, we randomly selected up to 30 individuals per original nest and split these individuals into two to three initial groups of ten (Nnests = 10, N groups = 25). Each selected individual received a unique color mark (©Plaka-Farbe) and was weighed to the nearest 0.01 mg on an electronic balance (Mettler Toledo New Classic MS). Each group was then transferred to a petri dish (100 mm in diameter) which served as the test arena for the hunting type assessment. An acclimatization period of four days ensured that the spiders weaved silk threads which amplify vibrations by prey32.Phase 1We assessed hunting types with a modified version of the ‘communal feeding experiment’ originally used by Dumke et al.16 to establish hunting specialization in A. ergandros. For each initial group, we completed 7 feeding trials over 24 days (1 trial every 4 days), during which we offered living Musca domestica flies and observed the foraging behaviour of all group members (Fig. 1). Each fly was weighed before being placed into the petri dish and either removed after two hours if not captured, or after two hours post capture. For each trial, we documented the attack latency, the attacker IDs and the IDs of the feeding individuals in 10-min intervals over two hours. From these data, we determined the feeding frequency of each individual (i.e. the number of trials it was feeding) and calculated the proportions to which it cooperated vs. defected. We thus obtained comparable quantifications of hunting types16. All individuals except those that died during the assessment (56 of 250 spiders) were weighed two days after the last trial to assess weight gain1 (= log (end weight1/start weight133).Phase 2Following phase 1, we regrouped individuals into ‘sorted’ groups of cooperators or defectors only, and this time gave three days acclimatization time since the re-grouping took one day. We formed experimental cooperator groups by selecting nine to ten individuals with the highest cooperating tendencies from the original colony. Next, we formed experimental defector groups analogously from that same pool (Fig. 1). Thus, we achieved paired relatedness between cooperator groups and defector groups, to control for nest origin and nest experience (matched pairs design). We further ensured comparability of cooperator groups and defector groups in the individuals’ physical state (details in Supplementary Methods). Owing to mortality in three nests and restricted possibilities to realize balanced conditions between groups in two nests, we could establish five cooperator-defector group pairs with nine individuals per group.To explore group composition effects on social foraging behaviour and individual fitness payoffs, we tested each sorted group over another seven feeding trials over 24 days. The trials were conducted in exactly the same manner as for the feeding type assessment (phase 1). From the recorded data (attack latency, IDs of attackers, IDs of feeding individuals), we calculated a set of variables that quantified social foraging behaviour (data points per trial and group). To examine individual fitness payoffs, we checked the petri dishes for dead individuals and noted their IDs prior to every trial. As an additional fitness payoff measure for those individuals still alive at the end of phase 2, we determined individual weight gain2 (= log (end weight2/start weight2)).The role of sex in cooperator vs. defector typesTo examine the role of sex in cooperation-defection scenarios, we collected another eight nests from Yass River Road in June 2016. Around this time, A. ergandros individuals reached the subadult stage, at which sex can be visually determined17. Three nests contained subadult males and females in sufficient numbers, so that we formed three groups, each with ten males and ten females from the same nest (in total: N males = 30, N females = 30). All group members were weighed and color marked before they were tested in another, extended feeding type assessment over ten trials.Based on the IDs of attackers and individuals that hunted in these trials, we generated social network graphs and visualized the foraging interactions within groups31,34,35. Individuals were represented by ‘nodes’; a directed line (‘edge’) was drawn from one node to another if the specific individual had cooperated by sharing prey with the other. The lines received weights reflecting the frequency of the respective interaction. We quantified individual prey sharing tendencies using the node-level metric out-strength: the weight sum of all outgoing edges from a particular node35. This metric comprehensively reflects an individual’s prey sharing tendency, as it incorporates the frequency and the spread of prey sharing behaviour. To visualize social networks and calculate the individuals’ out-strengths, we used the software UCINET 636.Statistical analysesAll model analyses were performed in R version 3.2.2, whereas all social network analyses were conducted in UCINET 636.Group composition effectsWe modelled the effect of group composition on social foraging behaviour separately for each response variable with binomial or gamma GEEs (generalized estimation equations). GEEs are adequate to analyse data from repeated measurements over time within same groups because they allow adjustment for the dependence of these measurements37. Defining the dependence structure of our data, we set sorted-group ID as a grouping variable and specified the temporal correlation AR-1. Group composition constituted the explanatory variable of interest, fly weight and group size were included as additional variables to control for prey mass and mortality. An exception was the model for the scrounging degree, in which group size was controlled by the variable itself. We assessed the significance of group composition effects by dropping each explanatory variable in turn and then comparing the full model to its nested models based on Wald test statistics. The least significant variable was removed, and model comparisons were repeated until all remaining variables were significant.Mortality was compared between cooperator groups and defector groups using a Chi-squared test. The difference between group compositions in individual weight gain2 was analysed in a GLS (generalized least squares) model that incorporated an exchangeable correlation structure with sorted-group ID as the grouping variable.Sex differencesWe conducted a node-based Monte Carlo randomization test to determine whether the observed difference in mean out-strength between sexes deviated significantly from the difference expected if producing associations occurred randomly and hence independent of sex. The observed data were shuffled in 10,000 node-label randomizations that preserved group membership. The sum of the differences between mean male out-strength (σm) and mean female out-strength (σf) within groups was used as the test statistic A (A = sumnolimits_{(i = 1)}^{3} {left( {sigma_{{(m_{i} )}} – sigma_{{left( {f_{i} } right)}} } right)}^{ – }), where i denotes group identity. To produce a probability value, we compared the observed test statistic to the distribution of random test statistics drawn from the 10,000 Monte Carlo simulations34. More

Bacterial strains, media and batch culturesWe used the wildtype strain Vibrio natriegens ATCC 14048 and Alteromonas macleodii sp. 4B03 (non-clumping variant) isolated from marine particles [8]. Strains were cultured in Marine Broth (MB, Difco 2216) and grown overnight at 25 °C. In total, 1 ml of cell culture was centrifuged (13,000 rpm for 2 min) in a 1.5 ml microfuge tube. After discarding the supernatant, the cells were washed with 1 ml of MBL minimal medium medium without carbon source. Cells were centrifuged again and the cell pellet was resuspended in 1 ml of MBL (marine minimal medium) [30, 34] adjusted to an 0.002 OD600. Cells from these cultures were used for experiments in MBL minimal medium containing 0.1% (weight/volume) Pentaacetyl-Chitopentaose (Megazyme, Ireland). The carbon source was added to the MBL minimum medium and filter sterilized using 0.22 μm Surfactant-Free Cellulose Acetate filters (Corning, USA). A total of 500 µl of the prepared cultures (250 µl + 250 µl for co-cultures) were added to 9.5 ml of MBL + 0.1 % chitopentaose (v/w) in serum flasks. This resulted in a starting OD of 0.0001. The flasks included a stirrer and were sealed with a rubber seal. Serum flasks were stored on a bench top magnetic stirrer (500 rpm) and connected to the microfluidics setup via Hamilton NDL NO HUB needles (ga21/135 mm/pst 2).MicrofluidicsMicrofluidics experiments were performed as described previously [35,36,37,38]. Cell growth was imaged within mother machine channels of 25 × 1.4 × 1.26 μm (length × width × height). Within these channels, cells could experience the batch culture medium that diffused through the main flow channels. The microfluidic device consisted of a PDMS flow cell (50 µm/23 µm). The PDMS flow cell was fabricated by mixing the SYLGARD 184 Silicone Elastomer Kit chemicals 10:1 (w/v), pouring the mix on a master waver and hardening it at 80 °C for 1 h. The solid PDMS flow cell was cut out of the master waver and holes were pierced at both ends of each flow channel prior to binding it to a cover glass (Ø 50 mm) by applying the “high” setting for 30 s on the PDC-32G Plasma Cleaner by Harrick Plasma. The flow cell was connected via 40 mm Adtech PTFE tubing (0.3 mm ID × 0.76 mm OD) to a Ismatech 10 K Pump with 40 mm of Ismatech tubing (ID 0.25 mm, OD 0.90 mm) which again was connected via 80 mm Adtech PTFE tubing (0.3 mm ID × 0.76 mm OD) via a 5 mm short Cole-Parmer Tygon microbore tubing (EW-06418-03) (ID 0.762 mm OD 2.286 mm) connector tubing to a Hamilton NDL NO HUB needle (ga21/135 mm/pst 2) that was inserted into the feeding culture. During the whole experiment the pump flow was set to 1.67 µl/min (0.1 ml/h).Time-lapse microscopyMicroscopy imaging was done using fully automated Olympus IX81 or IX83 inverted microscope systems (Olympus, Japan), equipped with a ×100 NA1.3 oil immersion, phase contrast objective, an ORCA-flash 4.0 v2 sCMOS camera (Hamamatsu, Japan), an automated stage controller (Marzhauser Wetzlar, Germany), shutter, and laser-based autofocus system (Olympus ZDC 1 and 2). Detailed information about the microscopy setup has been described by D’Souza et al. [39]. Channels on the same PDMS Chip were imaged in parallel, and phase-contrast images of each position were taken every 5 min. The microscopy units and PDMS chip were maintained at room temperature. All experiments were run at a flow rate of 0.1 ml h−1, which ensures nutrients enter the chamber through diffusion. Four biological replicates were performed. These replicates consist of four independent microfluidics channels (two for each of the strains). These channels were connected to one of two independent batch cultures.The microscopy dataset consists of 200 mother machine channels; 49 channels for the degrader on co-culture, 51 for the degrader on mono-culture, 40 for the cross-feeder on mono-culture and 60 for the cross-feeder on co-culture.Image analysisImage processing was performed using a modified version of the Vanellus image analysis software (Daan Kiviet, https://github.com/daankiviet/vanellus), together with Ilastik [40] and custom written Matlab scripts.Movies were registered to compensate for stage movement and cropped to the region of growth channels. Subsequently, segmentation was done on the phase contrast images using Ilastik’s supervised pixel classification workflow and cell tracking was done using the Vanellus build-in tracking algorithm.After visual curation of segmentation and tracking for each mother machine and at every frame growth parameters were calculated using custom written matlab scripts [36]. Lengths of individual cells were estimated by finding the cell center line by fitting a third-degree polynomial to the cell mask; then the cell length was calculated as the length of the center line between the automatically detected cell pole positions (see Kiviet et al. [33] for details).We quantified cell growth by calculating single-cell elongation rates r from measured cell length trajectories: L(t) = L(0)∙e^(r ∙ t). Cell lengths and growth rates varied drastically over the time course of the experiment; we thus developed a robust procedure that can reliably estimate elongation rates both for large fast-growing cells as well as for small non-growing cells. We first log-transformed cell lengths, which were subsequently smoothed over a moving time window with a length of 5 h (60 time points). We used a second order local regression using weighted linear least squares (rloess method of Matlab smooth function) in order to minimize noise while maintaining sensitivity to changes in elongation rates. Subsequently the instantaneous elongation rate was estimated as the slope of a linear regression over a moving time window of 30 min (7 time points). Time points for which the fit quality was bad (χ2  > 10−4) were removed from the analysis [32]. All parameters were optimized manually by visually inspecting the fitting procedure of many cell length trajectories randomly selected from across all replicates.As cells are continuously lost from the mother machine channels it is non-trivial to calculate the total amount of biomass produced in the chip. We thus need to estimate this quantity from the observed single-cell elongation rates. Specifically, we estimated the total amount of biomass produced per individual mother machine until a given time point as:$$B_T = e^{Delta tmathop {sum}limits_{i = 1}^T { < r_i > } }$$Where is the average growth rate of all cells in a given replicate at time point i, and where Δt is the time interval between two timepoints. By using the average growth rate, we ignore the variation in growth rates between cells. However, it is difficult to calculate population growth when growth rates vary both with time and between cells and the current method still allows us to capture the overall effect of interactions on cell growth.Datasets and statistical analysisAll microfluidics experiments were replicated four times. No cells were excluded from the analysis after visual curation. For V. natriegens 2227 cells were analyzed on mono-culture, and 1707 cells were analyzed on co-culture. For A. macleodii 2657 cells were analyzed on mono-culture, and 3901 cells were analyzed on co-culture. Each mother machine channel was treated as an independent sample. All statistical analysis was performed in Rstudio v1.2.5033. Percent increases were calculated using the relative differences of estimated between the corresponding values. For mixed effect models analysis the LmerTest package (Version 3.1-3) [41] with the following equation were used: y ~ Batch + (1 | Replicate) The Tukey Post hoc test was performed using the Multcomp package (Version 1.4-15) [42].Chitinase assayDegrader and Cross-feeder cells were cultured in Marine Broth (MB, Difco 2216) and grown overnight at 25 °C. In total, 1 ml of cell culture was centrifuged (13,000 rpm for 2 min) in a 1.5 ml microfuge tube. After discarding the supernatant the cells were washed with 1 ml of MBL minimal medium medium without carbon source. Cells were centrifuged again and the cell pellet was resuspended in 1 ml of MBL adjusted to an 0.002 OD600. A total of 10 µl of cell culture was added to 190 µl of MBL containing 0.1% Chitopentaose (w/v). Cultures were grown to exponential phase in a plate reader (Eon, BioTek) at 25 °C. Cell free supernatants were generated by sterile filtering cultures using a multi-well filter plate (AcroPrep) into a fresh 96 well plate. Chitinase activity of cell free supernatants was measured using a commercially available fluorometric chitinase assay kit (CS1030, Sigma-Aldrich) following the protocol. In short, 10 µl of sterile supernatant was added to 90 µl of the assay mix. The solution was incubated in the dark at 25 °C for 40 min before measuring fluorescence (Excitation 360 nm, Emission 450 nm) in a plate reader (Synergy MX, Biotek). Logarithmic chitinase activity per OD600 was analyzed for eight replicates.Chitinase activity in units per ml was calculated using a standard concentration. Using the following Formula: ({{{{{rm{Units}}}}}}/{{{{{rm{ml}}}}}} = frac{{left( {{{{{{rm{FLU}}}}}} – {{{{{rm{FLUblank}}}}}}} right) times 1.9 times 0.3 times {{{{{rm{DF}}}}}}}}{{{{{{{rm{FLUstandard}}}}}} times {{{{{rm{time}}}}}} times {{{{{rm{Venz}}}}}}}})Here, FLU indicates measured fluorescence, DF indicates the dilution factor, and V indicates the volume of the sample in ml [43].Acetate assayCell cultures were prepared and grown in serum flasks as described above. At different time intervals 1 ml of culture was removed and OD600 was measured. Cultures were filter sterilized using 0.22 μm Surfactant-Free Cellulose Acetate filters (Corning, USA) into a 1.5 ml microfuge tube. Cell free supernatants were stored at −4 °C until they were used for acetate measurements. Acetate concentrations were measured using a colorimetric assay kit (MAK086, Sigma-Aldrich) following the protocol. In short, 50 µl of cell free supernatant was added to 50 µl of assay mix. The solution was incubated in the dark at 25 °C for 30 min. Acetate concentrations were measured in a plate reader (Eon, Biotek) at 450 nm [44].Growth on spend mediaDegrader and Cross-feeder cells were cultured in Marine Broth (MB, Difco 2216) and grown overnight at 25 °C. In total, 1 ml of cell culture was centrifuged (13,000 rpm for 2 min) in a 1.5 ml microfuge tube. After discarding the supernatant, the cells were washed with 1 ml of MBL minimal medium medium without carbon source. Cells were centrifuged again and the cell pellet was resuspended in 1 ml of MBL adjusted to an 0.002 OD600. A total of 10 µl of cell culture was added to the 190 µl cell free supernatant described above. Cultures were grown in a plate reader (Eon, BioTek) at 25 °C. Cell free supernatants after this growth assay were generated by sterile filtering cultures using a multi-well filter plate (AcroPrep) into a fresh 96 well plate. These supernatants were used as described above to measure acetate levels after growth on spend media. More

Funicular bodies: structure, function, and developmentThe ultrastructural and functional complexity of funicular bodies in bryozoans25,46,47 (our data), as well as the reduction of the genome in their symbiotic bacteria (e.g., Bugula neritina48), point at a long-term co-evolution between these organisms, also suggesting that infected bryozoan colonies spend a significant part of their energy budget supporting numerous bacteria inside FBs. All known prokaryote symbionts are apparently non-pathogenic for bryozoans49,62. Instead, the overall evidence indicates a mutualistic relationship. Symbiotic bacteria are known to produce toxic substances (bryostatins in Bugula63; bryoanthrathiophene in Watersipora64) that protect bryozoan larvae from predators42,44,45. Another potential role of bacterial secondary metabolites is the chemical defense of early developmental stages during larval settlement and metamorphosis40,65. Such chemical defence may also prevent epibiotic overgrowth of the bryozoan colonies by other bacteria and algae (reviewed in66,67). Similar functions can be assumed for symbiotic prokaryotes of Dendrobeania fruticosa, although experimental evidence is still required.FBs with bacterial symbionts in colonies of D. fruticosa show signs of high specialization. The walls of the funicular body completely isolate its internal cavity from the surrounding zooidal cavity: the cells of the outer layer overlap each other, whereas cells of the inner layer have tight Z-shaped contacts. Such isolation probably creates a specific environment inside the symbiont-containing space that helps maintain a growing population of bacterial symbionts. A massive protein-synthesizing apparatus was observed in the cells of the inner layer. In addition, the ‘pocket’-like structure of the internal cavity of FB and the abundance of cytoplasmic processes, some of which protruded deep into this cavity, contribute to increasing both the general inner surface area of FB and the contact area between its inner cells and bacteria. From their side, bacteria provide this contact by numerous pili. Numerous pits and microvesicles associated with the apical membrane of the inner cells imply an active exchange of substances between the host and its symbionts (e.g., nutrition provided to growing and multiplying bacteria and absorption of the potential wastes they produce).The ultrastructure and functional morphology of FBs in Aquiloniella scabra, the only species in which the ultrastructure of these organs was described, are very similar to those in D. fruticosa. Its FBs also consist of two cell types, but the inner cell layer is formed by one or a few cells, whereas external cells form a multilayered envelope25. In both species, the main function of FBs was considered to be the chambers/organs for symbiont incubation and nourishment.FBs with symbiotic bacteria have been found in several bryozoan species from four different families20,25,47,49,51 (our data). Remarkably, all these families (Bugulidae, Beaniidae, Candidae, Epistomiidae) belong to the clade Buguloidea. Encapsulated aggregations of bacteria within the zooidal cavity were also described in two species of Watersipora from the phylogenetically distant family Watersiporidae (Smittinoidea)62,68. In that case, prokaryotes (“mollicutes”) were also enveloped by flattened bryozoan cells, although the current scarcity of data makes it difficult to compare these cell aggregations with funicular bodies.Among Buguloidea, FBs show the same basic structure, although species described by Lutaud49, Dyrynda and King51, and Mathew with co-authors46 based on light microscopy require re-examination by TEM. Moreover, FBs are in all instances associated with the funicular system of the colony. This similarity could indicate the single origin of the bacterial symbiosis and FBs within Buguloidea. Still, it remains unknown (although rather probable) whether that system transports nutrients from feeding polypides to FBs because no communication between the lacunae of the funicular cords and the FB internal cavity was detected. Another, more likely option, however, is the independent acquisition of bacteria (see also below) and a similar ‘reaction’ of host tissues to invaders, resulting in the formation of bacterial organs (FB) with a similar bauplan. The presence of incapsulated bacteria in non-related Watersipora supports the second interpretation. Answering this question will require both ultrastructural studies and molecular identification of bacteria.As for the initial source of somatic cells and development of FBs, two variants have been suggested. Describing the development of FBs in Bugulina turbinata, Lutaud49 stated that the epithelial cells of the cystid wall were transformed into the inner cell layer of the FB, whereas peritoneal cells formed its external lining. By contrast, while studying Aquiloniella scabra, Karagodina with co-authors25 suggested that bacteria are engulfed by one of the funicular cells, which becomes a ‘bacteriocyte’ that is later enveloped by neighboring funicular cells. In the latter case, FBs are considered to be modified expanded parts of the funicular system. This is consistent with experiments by Sharp and co-authors40, who detected groups of labeled bacteria in the funicular cords of Bugula neritina, and also multiple bacteria developing inside enlarged funicular cords (in fact, very large FBs) in the related Paralicornia sinuosa20.Our TEM study of FBs in D. fruticosa showed that they are not swollen parts of the funicular cords, as was stated by Vishnyakov with co-authors for B. neritina20. It is more likely that the funicular cords and processes of their cells contact the external cell layer of FBs. According to the third scenario that we present here, the inner cell layer of FBs in D. fruticosa, as well as in other studied bugulids, most likely originates from the coelomocytes which accumulate bacteria via phagocytosis. Such solitary cells were described inside the zooidal cavity in B. neritina20. Instead of being digested, engulfed bacteria could trigger the coelomocyte divisions resulting in the formation of the inner cell layer. In contrast, in A. scabra, the coelomocyte can remain single or undergo only a few divisions. We propose that the external cell layer of FB in D. fruticosa originates from the funicular cells because these cell types are ultrastructurally similar. Finally, it is also possible that the exact process and sources of FB formation differ in different species.Multiple and lobed FBs found in two zooids of D. fruticosa could indicate a potential mode of their multiplication. The case of P. sinuosa requires additional study, but currently we believe that its bacteria-bearing ‘funicular cords’ are very large, elongated FBs, as well (see20).Symbiont circulation in the bryozoan life cycleThe taxonomic diversity of bryozoan hosts and their symbiotic bacteria—supported by a variety of sites in the bryozoan zooids and larvae where symbionts have been found—unambiguously point to multiple independent origins of symbiotic associations between bacteria and cheilostome Bryozoa20,25.Rod-shaped bacteria are the most common symbionts in the superfamily Buguloidea. Although superficially similar, these bacteria strongly differ in their maximum size, suggesting the presence of different procaryote species. Thus, bacteria detected in the larvae of Bugulina simplex and in FBs of Aquiloniella scabra can reach 10 μm in length25,69, while healthy symbionts (see below) inside coelomocytes and presumably peritoneal cells of Bugula neritina were only 2.5 μm long20. The maximum length of bacteria in FBs of D. fruticosa never exceeded 5 μm. Coccoid bacteria in the tentacles of B. neritina were 0.5–0.7 μm in diameter20. Else, oval or irregularly-shaped mycoplasma-like α-Proteobacteria were detected in the genus Watersipora62,68,70. By contrast, the symbionts identified in B. neritina and B. simplex belong to γ-Proteobacteria69,71,72.Apart from FBs, prokaryote symbionts were described extracellularly in colonies of different cheilostome species: in vestibular glands of autozooids, inside polymorphic zooids (avicularia)38,39,50, in tentacles and funicular cords20. They have also been found intracellularly: inside coelomocytes, epithelial and peritoneal cells of the body wall, and pharyngeal cells20,50. In addition, Woollacott and Zimmer73 described bacteria in the ‘channels’ of the funicular cords associated with brood chambers. However, the TEM image they published shows bacteria inside large vacuoles of the funicular cells—seemingly not in the lacunae between these cells, recalling the aforementioned idea of a ‘bacteriocyte’. Finally, bacteria have also been found in the pallial sinus of bryozoan larvae (62,68,74; reviewed in72). All these diverse data have led to two opposite views on the acquisition and circulation of symbionts in the bryozoan life cycle.Discovery of bacteria in both colonies and larvae of B. neritina was regarded as possible evidence of their transmission from larvae to adults74. This assumption was experimentally proven using both labeled bacteria and their metabolites (bryostatins) in the larvae and preancestrulae developing from them40. Moreover, the presence of bacteria within brood chambers (ovicells) in this species was considered as proof for the next step—the transition of symbionts from the colony to the incubated larvae, i.e., the vertical transfer of symbionts. Symbiotic bacteria populating larvae (and making them unpalatable for predators) are incorporated into the preancestrular tissues during larval metamorphosis, and then found inside zooidal buds in early colonies (a symbiont association with the host cells was not specified) and funicular cords of rhizoids in adult colonies40.The next step of the bacterial development could be the formation of FBs as a locus of symbiont reproduction. Mature FBs, full of bacteria, were considered to be the starting point for the transfer of prokaryotes (by an unknown mechanism) from the zooid to the brood cavity (via funicular cords associated with both FB and ovicells), and then to the incubated larvae47. Light microscopic data demonstrated: (1) the association of FBs with tube-like funicular cords47, and (2) the presence of groups of ‘bacterial bodies’ (small aggregations of bacteria) inside the ooecial vesicle (membranous-epithelial ‘plug’ that closes the entrance to the ovicell; mentioned in B. neritina47, and shown in images of the related Bugulina flabellata75,76,77). In addition, ultrastructural data proved the presence of bacteria inside funicular cords, more precisely—inside their funicular cells (see above), “extending to the ooecial vesicle” (73, p. 362). Elsewhere, Sharp and co-authors (40, p. 697) used fluorescence microscopy to demonstrate the presence of bacteria “within the ovicells”, and suggested that they are transported there across the colony via funicular cords that also house bacteria. Combined, all these data imply that bacteria move from FBs to the ooecial vesicle, accumulate there, and then somehow enter the brood cavity, which contains a larva, either through or in-between the epithelial cells and the cuticle of the ooecial vesicle. Findings of bacteria inside larvae and adult colonies of two species from the non-related genus Watersipora62,68 further strengthened the hypothesis of vertical transfer, which has subsequently been widely accepted by many authors20,30,47,78. Despite extensive TEM studies, no bacteria have been found inside the funicular cords in B. neritina (Vishnyakov & Ostrovsky, unpublished data), which contradicts the data of Sharp and coauthors40 obtained by fluorescent microscopy. Accordingly, it was suggested that coelomocytes carry symbionts to the ooecial vesicle instead20.Nonetheless, the hypothesis of vertical transfer faces serious objections based on life history, molecular and morphological data. In Dendrobeania fruticosa in the White Sea, for example, larval production occurs predominantly in autumn (mainly in the distal parts of branches, Fig. 1A), and no bacteria are present in colonies during this period. In addition, molecular population studies revealed that B. neritina is a complex of sibling species, both symbiotic and aposymbiotic, some of which live in sympatry, with the horizontal transfer between colonies being the most parsimonious explanation for the distribution of bacteria between siblings79. A study of the genome of symbiotic bacteria showed that they may be able to live outside the host48, which is consistent with the hypothesis of horizontal transfer (which is not the same as environmental transmission, see below).TEM data showed no communication between the FB cavity and lacunae of the funicular cords in the studied species, in particular in B. neritina (Vishnyakov & Ostrovsky, unpublished data), Aquiloniella scabra25 and D. fruticosa (this study). Coelomocytes (and presumed peritoneal cells) with bacteria embedded in their cytoplasm were indeed recorded inside the ooecial vesicle in B. neritina20. Nevertheless, extensive TEM studies of B. neritina ovicells at various stages of placental development (Vishnyakov and Ostrovsky, unpublished data) have not revealed bacteria between placental cells adjoining a developing embryo. The fact that these cells are provided with both tight and adherens junctions (e.g., B. neritina and Bicellariella ciliata73,80), and additionally are covered by a cuticle (albeit thin) raises the question of whether coelomocytes with bacteria and/or bacteria alone can move through the very thick hypertrophied placental epithelium. Interestingly, Miller with co-authors48 detected a gene encoding chitinase in the genome of the symbiont of B. neritina that could potentially be used for cuticle piercing.Another opportunity for the vertical transfer of symbionts is their transport via the supraneural coelomopore during oviposition (see20,81). In this case, free bacteria in the cavity of the maternal zooid could stick to the ovulated oocyte before its transfer into the brood cavity via the coelomopore. However, free bacteria were never recorded in the zooidal cavity. For the Watersipora species, an assumed variant of symbiont transmission is through a strand of mucus extending from the maternal zooid to the released larva and tethering it for a few minutes68.Environmental transmission, when bacteria are acquired from the surrounding seawater, is an alternative option for symbiont acquisition. It may potentially occur either via infection of brooded larvae inside the ovicell by bacteria entering the brood cavity from the external environment or via infection of larvae during the free-swimming period by bacteria from the water column. Published images by Sharp and co-authors40 showed the presence of both symbiotic bacteria and bryostatins both inside the ooecial vesicle and in the peripheral part of the brood cavity, close to the entrance of the ovicell. Although these authors stated that such close “locations of both the bacterial symbionts and the bryostatins demonstrate that the B. neritina–‘E. sertula’ association has a delivery system for both the symbionts and the bryostatins to embryos within the ovicell” (40, p. 699), we argue, based on the above-mentioned data, that this statement remains a probable yet unproven speculation. Our numerous unpublished TEM images indicate the presence of a large number of bacteria filling the brood space between the embryos and the ovicell wall in the brood chambers with and without developing larvae. These bacteria, attracted by some chemical signal(s), can enter the ovicells from outside and infect larvae. Until the transfer of bacteria through the wall of the ooecial vesicle or during oviposition is documented, environmental transmission remains the more probable method. Notably, recent studies on sponge microbiotas showed that the environmental transmission is widespread in this group of suspension feeders82.The hypothesis of external acquisition of bacteria by the bryozoan hosts leaves different infection pathways open. Some of these could potentially develop into vertical transfer. Beyond the infection of larvae, prokaryotes could enter feeding autozooids via the mouth (and further through the intestinal epithelium into the zooidal cavity), through the coelomopore—a presumed entrance for alien sperm76,81,83, or by direct infection of the tentacles (probably by penetration through the outer epithelium).We should stress that FBs were absent in zooidal buds and the youngest zooids with functional polypides in the growing branch tip of one colony of D. fruticosa collected in June. This suggests that bacteria are not transmitted from the older colony parts (and, thus, are not inherited from the founding larvae), but obtained from the external medium since older (and more proximal) zooids had FBs. This idea is supported by the lack of signs of transfer of bacteria between zooids via communication pores and their pore-cell complexes in Dendrobeania fruticosa and Aquiloniella scabra in our TEMs. In contrast, fluorescence microscopy showed symbionts in the non-feeding zooidal bud of the newly-formed small colony of B. neritina40. Bacteria were also present in the preancestrula formed during larval metamorphosis. It remains unknown whether they can move from the preancestrula to the bud along with coelomocytes, with growing funicular cords or both before the formation of transverse walls that isolate newly budded zooids. So, interzooidal transfer to budding sites is possible, and the youngest zooids with functional polypides in the growing tips of D. fruticosa could already receive bacteria too, but FBs were not yet developed. Thus the question of interzooidal/intracolonial transport of bacteria remains open.Two infection pathways—via larvae and by direct penetration through the tissues of the functional polypide, potentially exist in the same species. This is the case in B. neritina, which has morphologically different symbionts in FBs and in the tentacles20. Bacteria in the epithelial wall of the tentacle sheath and ooecial vesicle (see above) could potentially get there via both pathways, or enter the zooid via the coelomopore (third way), subsequently becoming entrapped by coelomocytes or cells of the cystid wall.Finally, the presence of bryozoan sibling species, some of which have symbionts while others do not, and the presence of symbiotic and aposymbiotic colonies within the same species79, suggests that bacteria can be lost and acquired anew at both short- and long-term time scales, as occurs in hermatypic corals and their symbiotic zooxanthellae84,85,86. In this light, it would be important to know whether FBs can develop anew in D. fruticosa after overwintering in the same zooids, or whether they appear only in newly budding zooids. What is the source of bacteria in overwintered colonies? Is it an external infection, or some ‘survivor’-cells (descendants of the bacterial pool from the larva that overwintered inside epithelial cells and/or coelomocytes), or both? This will require further study.Symbiont population dynamics in Dendrobeania fruticosa and its potential driversWhatever the route used by bacteria to enter zooids, they are apparently immediately ‘trapped’ by somatic cells. Free bacteria have never been observed inside the zooidal cavity, another argument against their passage through the wall of the ooecial vesicle to the ovicell.We have shown that regardless of the as-yet-unknown mode of FB development, these temporary organs and the symbionts inside them undergo seasonal changes. Early and mature FBs with non-modified morphology and ‘healthy’ bacteria were found in young zooids only in the colonies collected in June. In one of these colonies, older zooids contained FBs at the initial stage of degradation. In the same month, one colony possessed FBs either at the early-advanced degradation stage (in young zooids) still containing numerous bacteria or at the late-advanced and even terminal stages in old, presumably overwintered zooids. At the initial stage of modification, the slightly developed ‘interlayer’ space between the inner and outer FB cell layers contained fibrils (presumed virus-like particles). Interestingly, the cells of the inner lining apparently engulfed some of the bacterial cells by phagocytosis, supporting our interpretation of the origin of these cells from coelomocytes in this species.One to two months later (August–September), all examined FBs were either at the middle-, or the late-advanced stages of degradation. The number of bacteria in the FB internal cavity distinctly decreased, the inner layer of cells became thinner, and in some regions remained only as a double membrane. The protein-synthesis apparatus was seen only occasionally, and engulfed bacteria were no longer visible inside the inner cells. A wide ILS, formed between the cells of the outer and inner layers, contained abundant putative virus-like particles (Figs. 2, 10). All these FBs were recorded in feeding and non-feeding zooids in the non-growing distal parts of colony branches.Polypide recycling and a seasonal drop of planktonic food alone cannot explain these changes. Firstly, modified FBs were found in zooids with both degenerated and functioning polypides. Secondly, the initial stages of FBs degradation were detected in June when phytoplankton is abundant in the White Sea (e.g.87). We therefore propose the following scenario for the sequence of changes in FBs and their possible causes. In June, newly-formed zooids build funicular bodies containing bacteria that were acquired either from outside or via internal transfer from older colony parts. During that month, FBs begin to degrade. This process continues throughout the rest of the summer. The mid-advanced stage of FB degradation, with few modified bacteria surviving and distributed on the periphery of the FB cavity, was recorded in August. This stage is reminiscent of the final stage in Lutaud’s49 descriptions of the gradual destruction of FBs accompanied by the disappearance of bacterial symbionts in Bugulina turbinata. In late September, FBs change and bacteria disappear, probably through viral lysis (the induction of prophages) (Fig. 2). Young and non-modified FBs were never encountered in August and September, indicating that development of FBs occurs only in young zooids at the growing tips of colony branches in June.Vishnyakov et al.20 recently described the degradation of symbiotic prokaryotes in Bugula neritina and Paralicornia sinuosa accompanied by a change in bacterial morphology similar to bacteriophage-mediated lysis. Degradation process was accompanied by the appearance of polyhedral VLPs in B. neritina and by the formation of structures similar to the so-called metamorphosis-associated contractile complexes (MACs) in P. sinuosa. These complexes are phage-related structures whose activity eventually results in the cell lysis, see88. Although the fate of FBs was not analyzed in their paper, these two VLP variants were observed both inside the bacteria and in the FB internal space by Vishnyakov and co-authors20.In D. fruticosa, presumed VLPs in the form of spherical complexes (as clusters of straight filaments) are present inside bacteria and, together with their fragments, inside the ILS. The filaments were also observed in the free state, frequently curved (apparently flexible) in the internal cavity of FBs. ILS was mostly filled with ‘fibrils’ (potentially representing modified/corrupted filaments) and ‘globules’, although filaments were incidentally recorded inside ILS too. We suggest that, in D. fruticosa, filaments may represent bacteriophage virions. This interpretation is supported by their appearance being associated with the degradation of bacterial cells, as in the case of VLP in B. neritina and P. sinuosa. Nonetheless, the morphology of the spherical complexes built from the filaments in D. fruticosa is unique: they do not resemble any known group of bacterial viruses. We found these putative VLPs inside ILS between the outer and inner FB cell layers. It remains unclear whether they travel there from the FB internal cavity or self-assemble inside ILS from individual filaments that were incidentally met there too. We add that spherical complexes, complete or partial, were recorded inside ILS of the funicular bodies in non-overwintered (collected on 31 September 2019) and presumably overwintered (14 June 2021) zooids.A filamentous morphology is known from only one bacteriophage from the order Tubulavirales, which includes two families Inoviridae and Plectoviridae89. Although inoviruses or plectoviruses have never been reported to assemble in any regular macrocomplexes, the ability of the filamentous Pf phages (inoviruses) of Pseudomonas aeruginosa to form nearly regular liquid crystalline assemblages was recently demonstrated90 (see also review91). Interestingly the formation of such crystals required interactions with bacterial or eukaryotic polymeric molecules such as polysaccharides, DNA and probably mucin90,91.The development of filamentous phages in bacterial cells usually does not kill the cells because these viruses assemble, along with extrusion from the infected bacterium, without disrupting its cell wall (reviewed in92). However, cell death mediated by filamentous prophage induction has been reported in P. aeruginosa due to the emergence of so-called superinfective phage variants93,94,95. Accordingly, the degradation of bacteria in D. fruticosa associated with bacterial viruses is possible.Since the assembly of known filamentous phages is associated with their extrusion from the cell, virions should not accumulate inside bacteria. Although filamentous assembly intermediates may be present (see92), they are not expected to accumulate in such large quantities and/or form superstructures in the bacterial cytoplasm like the ones we found in D. fruticosa (Fig. 9). Our observations revealed nothing resembling the extrusion of a filamentous phage from the surface of a bacterial cell. Therefore, if the described filaments are indeed VLPs, they may represent a new type of bacterial virus.Even though many details remain unknown, we assume that the filaments, ‘globules’, ’fibrils’, and spherical structures in D. fruticosa are of viral origin. Their development following the total disappearance of bacteria in FBs indicates their bacteriophage nature. If so, our observations support the idea that viruses control the number of symbionts in their bryozoan host20, as has been reported in some insects85,86,96. More