Over the summer, I was lucky enough to work with a team led by Niel Davies, from the University of Cambridge, on describing the largest known arthropod fossil to date – a Carboniferous millipede. The paper has just been published:
Davies, N.S., Garwood, R.J., McMahon, W.J., Schneider, J.W. & Shillito, A.P. 2021. The largest arthropod in Earth history: insights from newly discovered Arthropleura remains (Serpukhovian Stainmore Formation, Northumberland, England). Journal of the Geological Society. doi: 10.1144/jgs2021-115 (More information: 1, 2, 3, 4)
We estimate that this creature was about 2.6 metres in length when alive, and weighed in the region of 50 kilos. The fossil suggests gigantism in this group began before high oxygen concentrations built up in the Late Carboniferous, and we also took the opportunity to chart the changes in palaeogeography of the iconic trace fossils associated with these animals. The Natural History Museum have posted an excellent overview of the findings, and you can find further press coverage by following the links at the end of the citation above.
Below you can see an image of the fossil itself, then a reconstruction of the animal in life. Clicking on either will take you to a copy of the paper.
I'm very pleased to report on a new paper, led by a colleague at the University of Cardiff, Dr Lesley Cherns, that has been in the works for several years. A team of use from across the UK used a combination of X-ray and neutron tomography to scan a fossil ammonite with soft tissues preserved. This is amongst the first fossils to show the preservation of soft parts in an ammonite, and the configuration of the muscles we found in our scan tells us something about the evolution of this iconic group. For example, we suggest that the animals had a mobile, retractable body, and swam using jet propulsion: sa mechanism that must have evolved relatively early in the evolution of ammonoids and their closest relatives – the octopodes, squid and cuttlefish. The paper details are as follows:
Cherns, L., Spencer, A.R., Rahman, I., Garwood, R.J., Reedman, C., Burca, G., Turner, M., Hollingworth, N.T. & Hilton, J.M. 2021. Correlative tomography of exceptionally preserved Jurassic ammonite implies hyponome-propelled swimming. Geology. doi: 10.1130/G49551.1 (More information: 1, 2, 3)
You can learn more about the fossil by following the links at the end of the above citation, which will take you to press coverage of the discovery.
A funded PhD opportunity is available for a project supervised by myself and my colleague Rob Sansom, titled The timing and modes of macroevolutionary change: molecules, morphology, and simulations. If you're interested in real-world data, simulations and evolution, please don't hesitate to get in touch if you have any questions! Deadline for applications is the 21st of January. here for more details of this PhD opportunity.
A new paper, led by my colleague Callum McLean at Manchester Metropolitan University, has recently been published. The study investigates shape complexity in the pedipalps (mouthparts) of whip spiders within and across species using elliptical Fourier analysis. Pedipalps are used for hunting, but also used in contest for mates. The paper has several implications, but a key finding is that there may be a trade-off between investment in pedipalp length (for use in selection/contest) v.s. spine length (for use in prey capture). You can read more about it in the (open access) paper, which is linked below:
McLean, C., Garwood, R.J. and Brassey, C. 2021. Assessing the patterns and drivers of shape complexity in the amblypygid pedipalp using elliptical Fourier analysis. Ecology and Evolution. doi: 10.1002/ece3.7882
A new paper, led by Euan Furness, from Imperial, has just been published:
Furness, E.N., Garwood, R.J., Mannion, P. D. & Sutton, M.D. 2021. Productivity, niche availability, species richness and extinction risk: Untangling relationships using individual-based simulations. Ecology and Evolution doi: 10.1002/ece3.7730
This contribution uses the Palaeoware package REvoSim to investigate the impact of productivity on species richness. It shows that species richness scales with productivity – matching some of the predictions of an explanation for this pattern called the more individuals hypothesis. For example, rare species are more extinction prone, as MIH predicts. But, we show that species richness only scales with productivity when species can partition niche space – the relationship doesn't hold when niche partitioning is prevented through saturation. Thus even when neutral explanations do help us understand patterns, we suggest niche theory is also important to consider.
One of the preprints mentioned below has just been published in Proceedings B. Whilst fossils are key to understanding past life, their impact on phylogenetic reconstruction is hard to assess: in real organisms we lack a true evolutionary tree. In this study, led by my colleague Nicolás Mongiardino Koch, we use the Palaeoware package TREvoSim to generate trees and associated character data, benchmark these against empirical datasets, and then evaluate the impact of including fossils in phylogenies. We do so using an analytical pipeline summarised in the graphic below:
The outcome? We show that fossils – irrespective of the levels of missing data or inference method – improve the accuracy of phylogenetic inference, and induce the collapse of highly uncertain relationships that tend to be incorrectly resolved. In our study, tip-dated analyses outperform undated inference methods: we suggest that the age of fossils contains important information that aids phylogenetic analyses. You can learn more in the open access paper linked below:
Mongiardino Koch, N., Garwood, R.J. & Parry, L.A. 2021. Fossils improve phylogenetic analyses of morphological characters. Proceedings of the Royal Society B: 288(1950):20210044. doi: 10.1098/rspb.2021.0044
A few new pieces of work that I've been lucky enough to contribute to have appeared since I last updated this website. The paper linked below, led by Imperial PhD student Euan Furness, uses the Palaeoware package REvoSim to investigate the impact of environmental disturbance on species richness. It shows that a key consideration is the scale of a disturbance, and the spatial heterogeneity of the environment. In a hetrogenous environment, small scale disturbance leads to a decrease in species richniess, but in a homoegeneous one, it increases species richness. In contrast, large scale disurbances always reduce the number of species. You can read more in the open access paper below:
Furness E.N., Garwood R.J., Mannion P.D. & Sutton M.D. 2021. Evolutionary simulations clarify and reconcile biodiversity-disturbance models. Proceedings of the Royal Society B: 288(1949):20210240. doi: 10.1098/rspb.2021.0240
A paper on smell in the sea catfish – for which I assisted with data collection – has also been published in recent weeks:
Cox, M.A., Garwood, R.J., Behnsen, J., Hunt, J.N., Dalby, L.J., Martin, G.S., Maclaine, J.S., Wang, Z. & Cox, J.P. 2021. Olfactory flow in the sea catfish, Ariopsis felis (L.): Origin, regulation, and resampling. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology p.110933. doi: 10.1016/j.cbpa.2021.110933
It's been a long since the last update of this site, in large part because of the COVID19 pandemic. This coincided with a relaunch of the degrees in the Department of Earth and Environmental Sciences. As a result I spent my time since June writing 1.5 course units for the first time, and then delivering them remotely. You can find the results at the two websites linked below:
The first of these sites (EART22101) has seven websites combining videos, quizzes, and other learning materials on topics surrounding evolution, and ancient life. The second (EART27201) has eight sites that focus on invertebrate macrofossils, and include 3D fossil models and the occasional virtual microscope. Materials are generally CC-BY and sources are always provided: they are available here in the hope that they are useful! There are other elements of course delivery provided through Blackboard (more quizzes, achievements to unlock) and via Zoom, which aren't included. Nevertheless, I hope they may be useful!
Early in lockdown, though, there was time for some science – and one of the results is a preprint which has just appeared on bioRxiv. You can find it by clicking on the link below:
In this paper, led by my colleague Nicolás Mongiardino Koch, we asses the impact of adding fossils to phylognies using a simulation based approach, based on the Palaeoware package TREvoSim. We demonstrate that adding extinct species always improves the accuracy of a phylogenetic analysis, even in the presence of missing data. We also show that tip-dated phloygenies – i.e. those which simultaneously infer tree topology and divergence times – outperform all other methods of inference. You can find out more by reading the freely available preprint!
A new paper, led by my colleague Jonathan Cox at the University of Bath, and a follow up to a paper released last year, has just appeared. This one uses CT scanning and CFD to study smell in the pike Esox lucius. As with the last such paper, although I am first author, this reflects authorship conventions in the field. I didn't lead the work, but did have the pleasure of conducting CT data collection. Details of this publication are as follows:
Garwood, R.J., Behnsen, J., Ramsey, A.T., Haysom, H.K., Dalby, L.J., Quilter, S.K., Maclaine, J.S., Wang, Z. & Cox, J.P.* 2020. The functional nasal anatomy of the pike, Esox lucius L. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology 244: 110688. doi:10.1016/j.cbpa.2020.110688 (* = corresponding author)
I am pleased to report that a new paper has just appeared, accompanied by two new Palaeoware evolutionary simulation software packages. The paper is:
Keating, J.N., Sansom, R.S.*, Sutton, M.D., Knight, C.G. & Garwood, R.J.* In press. Morphological phylogenetics evaluated using novel evolutionary simulations. Systematic Biology doi:10.1093/sysbio/syaa012 (* = corresponding author)
In it, my colleagues and I use two new evolutionary models to test methods of inferring evolutionary trees. We use the models to simultaneously generate trees and discrete character data, one in a stochastic manner, and one using an individual-based approach which includes selection. We use different approaches to infer trees from the character data, and then comparing how the approaches perform by measuring the distance between the true and inferred trees. There have been a number of recent works that have taken a simulation approach, generating stoachastic data onto a set tree topology, which have generally suggested that Bayesian inference outperforms parsimony. Our study supports this idea, but also suggests this is because of the differences in search strategy and consensus methods used by each, rather than resulting from the optimality criterion they use. Our approach also allows us to demonstrate that stochastic- and selection-generated data behave differently, and that the amount of homoplasy in our datasets are not a good indicator of how methods of inference will perform. Rather morphological coherence (how strongly linked characters differences are to the evolutionary, or branch length distances between taxon paris) is a key consideration, and this is a bigger problem for selection-generated data. Morphology is thought to evolve through selection, and so we suggest that asking whether parsimony or Bayesian is better may not be the best appraoch for the field in future: both methods struggle with selection data. Rather if we can future research on modes of morphological evolution, and then use those to build models for probabilistic inference, we will do a better job of reconstructing trees in the future.
You can find releases of the two software packages/models we created for this paper by clicking the banner below. TREvoSim is our selection-based model. It uses a similar fitness-algorith to REvoSim, but employs a different species concept, does not include sexual reproduction, eschews space, and has variable genome sizes. MBL2017 is a birth-death model, which simulates lineage splitting and extinction, as well as character evolution, in a stochastic manner. Both packages are available as binaries for Windows and Mac OS, and can be built on Linux systems using four lines of bash. We are releasing the two in the hope that they are versatile and useful tools for the community, and would very much welcome feedback, bug reports, and feature requests: these are best achieved using the issues system in GitHiub, or by emailing us at firstname.lastname@example.org.