GEOL2301: Palaeoecosystems

The birth of animal complexity

The Cambrian “explosion” marked the sudden onset of complex animal-dominated ecosystems, and the origin of the modern animal groups, in a geologically brief window of time.

Was there something special about the Cambrian that precipitated this rapid radiation – or does this evolutionary event only stand out from others because there was so little there before it?

This is the focus of my own research. Today we're going to look in some depth at a Cambrian community.

Clueless about the Cambrian? Catch up with this optional refresher video, which will set the scene for this session.

The Cambrian Fauna

In a seminal palaeontological study, Jack Sepkoski showed that a ‘Cambrian fauna’ (marked I in figure) with a distinctive taxonomic composition arose in this period, before waning almost immediately during the Ordovician radiation.

Sepkoski compiled a database of the stratigraphic occurrence of marine families, allowing the first detailed analysis of the diversity of each invertebrate class (the Linnean rank below a phylum) through time. Elegant statistical analysis showed, remarkably, that the diversity of each class could be described as a blending of three ‘fundamental trends’ (factors).

Blending these three factors together accounts for the vast majority (>90%) of the pattern of diversity through time; the residuals largely correspond to times of mass extinction or rapid diversification.

Some classes’ trends (e.g. Trilobites, articulate brachiopods, gastropod molluscs) map almost exactly onto one of the three factors – these groups are considered to represent a ‘Cambrian’, ‘Palaeozoic’ and ‘Recent’ fauna. Other groups (e.g. ostracod arthropods, corals) represent a blending of two or three factors.

• Which factors blend to account for the diversity of corals (Anthozoa, rightmost column) through time?
• Factor I
• Factor II
• Factor III
• Many classes map to a single factor (find some examples). Why might corals represent a blending of two?
• Factor II maps to the success of rugose and tabulate corals; Factor III might represent the emergence of scleractinians.

Characterizing the faunas

Sepkoski suggested that members the three different faunas might be characterised by different evolutionary parameters. We’re going to explore this now with an ecological analogy. Could the Precambrian oceans, essentially devoid of macroscopic life, be analogous to a patch of ground recently cleared and about to undergo ecological succession? If so, the first animal species to evolve in the Cambrian might have something in common with the first individuals to arrive on a patch of new earth.

Let’s refresh our memory of succession. Imagine that a nettle seed, a hawthorn berry, and an acorn (oak seed) germinate on a patch of bare earth. In the left-hand plot, draw three curves, corresponding to the population size of each population (of nettles, hawthorn shrubs, and oak trees). In the right-hand plot, stack these curves vertically to depict the proportion of individuals that are (i), nettles; (ii), hawthorns; (iii) oaks.

• Double-click here to show my approximate answer.
• Your image should resemble Sepkoski's plot of the success of faunas I, II and III through time, albeit without the effect of mass extinctions. Compare with the first main figure in the Succession session.

Developing the analogy

Let’s pretend for a moment that families, each containing many species, come to fill empty ecospace over evolutionary timescales in the same way that populations, each containing many individuals, come to fill empty ecospace over ecological timescales.

Let’s explore this analogy further (without worrying, yet, whether it is a good analogy). Complete this table:

Ecological succession	Evolutionary analogue	Hint
Population of individuals	Family of species
Birth of new individual	🖉_______ of a new species	Birth is when an individual appears for the first time. What is the equivalent for a species?
Death of an individual	🖉_______ of a species	Death is when an individual's existence comes to an end. What is the equivalent for a species?
Reproductive rate	🖉_______ rate	Reproduction sees one individual become more than one individual. What is the equivalent for a species?
Death rate	🖉_______ rate	The death rate is how many individuals no longer exist after a unit of time. What is the equivalent for a species?
Life expectancy	🖉_______	Life expectancy is the average time from birth to death. What is the equivalent for a species?

Remind yourself of the trajectories of the Cambrian (I), Palæozoic (II) and Recent (III) faunas in Sepkoski's diversity plot.

Assuming that the analogy with ecological succession holds true, infer the properties of each fauna:

	Cambrian Fauna (I)	Palæozoic Fauna (II)	Recent Fauna (III)
Speciation rate	🖉high / medium / low	high / medium / low	high / medium / low
Species longevity	🖉high / medium / low	high / medium / low	high / medium / low
Competitivity	🖉high / medium / low	high / medium / low	high / medium / low

• If you get stuck, come back to this table later.

• In an ecological context, what sort of life strategy is associated with the lifespan and reproductive rate you reconstructed for the Cambrian and Recent faunas?

• Cambrian: rapid speciation / high species turnover is equivalent to the high reproductive rate / short lifespan we'd see in an r-selected species.

  Recent: vice versa for K-selected species.

• What other characteristics do you associate with these life strategies?
• r-selected species tend to be generalists / opportunists; K-selected species tend to be specialized competitors.
• Assuming that the analogy with succession holds, what characteristics would you predict might characterise (i), the Cambrian fauna; (ii), the Recent fauna?

• Might the Cambrian fauna be dominated by generalists / opportunists, and the Recent fauna specialized competitors? Does this fit with what you know of the increased specialization through time, for example of predators during the Mesozoic Marine Revolution?

As it happens, the predictions of our ‘succession analogy’ are not too wide of the mark. In particular, the Cambrian is often viewed as 'less competitive' than later periods, conjuring a harmonic idyll in which predation only exerted a secondary influence on ecosystem structure.

This raises an interesting question: do we need to invoke large-scale macroevolutionary processes to explain the replacement of biotas and the pattern of diversity through time, or do the first-order patterns in biodiversity simply represent the null expectation as life gradually conquers an uninhabited realm?

Preservation

The Burgess Shale is an exceptional fossil deposit that offers an unrivalled glimpse into Cambrian ecology. Event beds, packed with fossilized 'soft' anatomy, range from 2–200 mm thick, separated by fossil-poor background sedimentation; each corresponds to a single obrution event.

As always, it's worth worrying to what extent the fossil assemblage is representative of the living community, and to what extent the living community was representative of Cambrian communities more generally.

• What factors make the Burgess Shale promising for ecological study?
• • Soft tissue preservation: Everything's there [?]
  • Discrete burial horizons: Snapshots in time [?] (and space??)
• Is it legitimate to draw conclusions about Cambrian communities in general from the Burgess Shale?
• • Low Palaeolatitude (15° N) → potentially increased diversity, as more primary productivity possible
  • Connection to open ocean present, so not isolated basin; supported by pelagic trilobites; few endemic species
  • Shelly taxa match other Cambrian assemblages… so presumably non-shelly taxa in common too?
• How is the taxonomic composition of fossil assemblages likely to relate to that of the original (live) communities? What biases might be in play – and how might we evaluate their effects?

• Size? (Might large organisms manage to estape obrution events by swimming away or burrowing free?)

  Habitat? (Could burrowers be protected from transport, or better able to escape when transported?)

  Time averaging? (What was dying on the sea floor during an event?)

  Mineralized parts? (Might they provide ballast and reduce transport?)

  Decay?

  Collection bias? (What don't we spot, because we don't prepare it out, or we leave it in the field because it's not interesting enough to lug back to camp, or because the rock doesn't split along that plane?)

  Finding the same taxon in multiple beds is instructive: when that taxon is better preserved, do we see other taxa that don't appear elsewhere (and thus are presumably the most succeptible to decay)? As it happens, this sort of analysis suggests that decay does not systematically distort assemblage composition.

How to count the data

We're going to look at data from two studies: Conway Morris 1986, which uses ‘familiar’ techniques on ‘old’ collections, which incorporate talus material and were not collected systematically (anything dubbed inferior was chucked down the hillside); and Caron & Jackson 2008, which performs a bed-by-bed analysis and some more sophisticated stats for additional insight.

Conway Morris takes three different counts: all individuals; living individuals only (e.g. discounting empty shells); and biovolume. These results differ in certain respects (see figure).

Explain the following trends:

Systematic sampling allows Caron (foreground) to link samples to specific horizons.

• Super-abundant in all ways of counting (arthropods)
• Arthropods just really are dominant – however you measure them! And about 50-60% dominant too.
• Many individuals, little biovolume (hemichordates)
• Individuals are small.
• Few individuals, much biovolume (e.g. sponges)
• Individuals are disproportionately large. Interesting caveat with sponges – they are mostly hollow, so mostly water – does biovolume overestimate their true biomass?

• Many individuals, few live individuals (molluscs, brachiopods)
• Most individuals are empty shells. If the shells weren't biomineralized, they'd've rotted (and thus not be countable).
• Consistently in the middle (priapulid worms)
• Perhaps counting individuals overestimates because they are smallish, whereas biovolume overestimates because priapulids are mostly (metabolically inert) gut?

Habitat and feeding

Trends in habitat and feeding habit also vary depending on counting mechanism – but some observations are robust to this.

Use the slides below to interpret the ecology of the Burgess Shale. (Advance the slides to colour-code the habits.)

• Habitat bars are ordered from ‘more to less active’ from left to right; feeding habit categories are DC, DS: Deposit; SU: Suspension; PC: Predator / carnivore / scavenger.
• Not loading? Log in to a Microsoft website (Outlook?) using your Durham account, or open in a separate tab.

• How does the mode of counting influence the patterns in life habits (habitat)?
• Hemichordates are many but tiny, so “I.S.” inflated by counting individuals.
• Did predators play an important role in Cambrian ecosystems?

• Yes, a surprisingly large role, if you look at biovolume. No, if you look at numbers of individuals. Which measure is more appropriate here?

  Biovolume is important in quantifying the contribution of different trophic roles to ecosystems, because that’s how energy is passed up a food chain.

  Clearly deposit feeding AND suspension feeding are both important. Perhaps there are periodic currents so food sometime settles? Perhaps suspension feeders are less efficient in the Cambrian, so more food slips through the net?

Use the percentage data to (approximately) plot the Burgess Shale assemblage on Scott's (1978) ternary diagrams of Cenozoic environments.

• • • • ×

• Does the composition of individuals in the Burgess Shale correspond to the distribution of habitats and feeding habits in a Cenozoic marine ecosystem? Account for any differences.

• By biovolume: Fewer infaunal suspension feeders than any modern setting! Was burrowing limited in the Cambrian? (YES) Would infaunal feeders necessarily get swept up in mudflows? (NO)

  By individual count though: good match for middle shelf.

  Scott used abundance data (not biovolume) to generate these plots.

The trophic nucleus

The trophic nucleus (remind yourself of the definition…) gives a quick-and-dirty overview of the dominant consituents of an ecosystem.

• Use the trophic nucleus (below) to discuss the prevalance of predation in Cambrian ecosystems.
• Assuming that tbe Burgess Shale is representative (…) predators feature prominently in the trophic nucleus calculated by biomass; clearly a lot of energy was stored up in higher trophic guilds – even if piddly little hemichordates and mini-Marrellas swamp the census of individuals.

• Not loading? Log in to a Microsoft website (Outlook?) using your Durham account, or open in a separate tab.

Miniature Marrella (~10 mm long)

Digging deeper

We've now got an overall picture of the Burgess Shale community though bulk analysis of all its inhabitants. But do these trends over the ‘averaged’ data across the entire Phyllopod bed truly represent individual living communities?

The ‘core species’ (defined, ad hoc, as 5 most abundant) are pretty consistent from bed to bed. So is evenness and diversity (species richness), with the occasional outlier.

• What are the strengths and limitations of the 'core species' approach?

• Strengths: quick and easy to perform, gives a first-order overview.

  Limitations: arbitrary definition (less relevant if number of species is close to 5); emphasizes number of individuals, so will be biased to small/readily preserved species.

Coordination analysis

The table suggests that some beds are more similar than others. Coordination analysis allows such differences to be picked apart. (See Caron & Jackson 2008 for details of the analysis.)

• Why might certain species consistently occur together?
• • Similar environmental preferences (high / low O₂ / salinity?)
  • Similar ecological role (generalists / specialists; r/K-selection)?
  • …
• Reconcile your suggestions with the evenness and richness trends through the bedding assemblages.
• Here's one possibility – there are many others! Recently disturbed settings tend to have a low diversity, and are dominated by r-selected species. Stable settings tend to be dominated by competitors – K-selected species. r-selected species will thus tend to co-occur with other r-selected species, and in low-diversity settings.

Whittaker (rank-abundance) plots

Analyses of the whole Burgess Shale assemblage, and of individual beds, and of individual niches all fit a log-normal distribution best (with the odd exception – but run enough tests and you’ll always find an odd one out!).

• Provide an ecological interpretation of this log-normal distribution

• Log-normal suggests climax communities: taxa specializing to their niches.

  How does this fit with the presence of generalists in the core species? Or the changes between different taxonomic assemblages? Or the diversity / dominance data through the series of horizons?

  Provide a taphonomic, statistical or ecological explanation for the apparent disconnect

Conclusions for the Cambrian

Now bring together all the aspects we’ve addressed today. If you've been working in a group, see whether you agree on the following points:

• To what extent was the Burgess Shale a ‘typical’ Cambrian community?
• What were the ecological properties of the Burgess Shale life assemblage? In what environment did its denizens live?
• How is the Burgess Shale (and Cambrian ecosystems more generally?) similar to and different from modern ecosystems?

Linking to the Wenlock

You now have slab-by-slab data for the whole class, and have hopefully started to ponder how to approach the analysis.

How might you apply the ideas discussed today, in combination with the techniques introduced in the first two sessions? Here are some questions you might think about:

• How might you characterize the fauna of the Wenlock biota as a whole?
• What is its trophic nucleus?
• What feeding habits and substrate niches predominate?
• What does this tell you about the environment in which the fauna lived?
• Is each group’s sample simply a microcosm of the biota as a whole?
• Do any slabs stand out as having a distinct taxonomic or ecological composition?
• Do slabs generally have the same diversity and dominance? What accounts for any exceptions?
• What does this tell you about the environment in which the organisms lived?
• How might the collected data differ from the original life assemblage?
• What is the direction and extent of any biases you can identify?
• If these potential biases are extensive, how would they affect your interpretation?
• Are you able to quantify or correct for these biases? If not, what caveats should be raised when you discuss your results?

Take the opportuntity to chat with a demonstrator today to be confident in how you will approach any of these questions that you feel might be relevant – and any other questions that might occur to you. Make a start on some of the analyses as soon as you can.

I would suggest conducting your analyses this week. Next week you can discuss your interpretations, and get help on any analyses that proved intransigent. That’ll give you a week to finish writing up your results, before receiving feedback on your draft submission.