All swimming all dancing fossil page

The story of one man's plight to make fossils swim

Ugly mugWelcome to my homepage, which is really just a description of my Ph.D. project on fossil biomechanics.

For the last three and a half years I have been working at the University of Manchester trying to work out how a strange group of extinct creatures called eurypterids would have swam. The problem is not at all straightforward. Since eurypterids are known only as fossils I have had to reconstruct their biomechanics with no behavioural data and often unreliable physical data. Indeed, the problem covers aspects of palaeontology, biology and engineering. I am now just finishing off writing up my Ph.D. thesis and this page presents the edited highlights (!) of that.


The problem

One of the babies Eurypterids were chelicerates which thrived during the Ordovican to Devonian and went extinct at the end of the Palaeozoic. Eurypterids came in variety of shapes and sizes and some reached two metres in length. The species I am going to be talking about, a close relative of Eurypterus (right), was a more modest 20 cm or so as an adult.

Some eurypterids possess paddles which many workers have regarded as an adaptation for swimming (others have suggested that they were used in digging). This assertion is supported by the remarkable similarities between the eurypterid paddle and the swimming paddle of living portunid crabs. The similarities are not just morphological: the two paddles show further similarities in their hydrodynamic characteristics and the positions and sizes of the muscle attachment sites.

There have been two suggestions as to how eurypterids might have used their paddles to swim. The first, by Paul Selden, had the paddles used as oars. In this hypothesis the paddle blade is swept backwards relative to the body and oriented perpendicular to its own movement to maximise drag. The blade is turned ninety degrees during recovery to minimise drag, which here would be disadvantageous. Essentially, this is a rowing technique, similar to that used by water beetles and muskrats.

The second suggestion, by Roy Plotnick from the University of Chicago, had the paddles used in a more sophisticated figure-of-eight mechanism, with the thrust being generated mainly through lift. Plotnick's model is comparable to the hovering phase of insect flight.

Neither model was tested experimentally. Both look feasible on paper, but both are essentially speculations with no quantitative evidence to back them up.

My wonderful invention

I have attempted to approach this problem from a biomechanical perspective. However, whereas most biomechanical studies have attempted to explain observed behaviour in mechanical terms, the aim of this study is to predict the likely behaviour of an animal from only morphological information.

It has been necessary to develop a mathematical model for swimming in which the majority of the parameters can be obtained from fossil evidence. The models which currently exist were primarily developed for living animals and are therefore inappropriate. Very few of the variables which are used in my model cannot be obtained from fossil evidence; in some cases, however, such as body density, it is necessary to make comparisons with living animals, usually portunid crabs, of a similar size.

Flume tank The problem can be broken down thus. Swimming is achieved when the thrust forces produced by the paddle exceed the resistance forces of the body, essentialy the resultant of body weight and drag. We want to know how can the paddle produce sufficient thrust to overcome the forces resisting swimming in such a way that the power required to do so is minimised.

It is necessary to know the hydrodynamic forces the paddle and body would have generated and to obtain such data from fossils, physical models must be built. The models are tested in a water flume tank at a range of speeds to measure the drag and the lift on them (left). Once such data has been obtained, it is a simple task to calculate the body resistance from standard equations. Calculating the thrust is a little more complicated. There are two reasons for this. First, the stroke angle is variable. The stroke angle is the range in which the paddle can be swept relative to the body. Its range can be estimated from the detailed reconstruction of the joint mechanics of Baltoeurypterus provided by Paul Selden in 1981. Second, the animal can vary the angle of attack of the paddle blade to optimise the ratio between the drag and lift production. In fact there are nine ways in which the animal can do this.

Screen shot Thus for each animal there are many thousands of equations to solve from which one stroke angle and stroke technique combination emerges as the most efficient. A computer program has therefore been developed to perform the calculations (right) and it takes about half an hour to arrive at an answer.

The computer program, which is called "SimSwim" (I thought would cash-in on the popularity of SimCity and SimEarth, but it didn't quite work out), was tested with data from a portunid crab in order to see if its predictions are accurate. Since the paddles of portunid crabs are analogous to those of eurypterids and since their swimming behaviour is well documented, they would seem the obvious candidate for this test.

In fact, the computer model provides a reasonably accurate reflection of the swimming mechanics of the swimming crab. Furthermore, it suggests that crabs swim most efficiently with a lift-based technique with steep power and recovery angles, which is in general agreement with observations made of crabs swimming. The swimming speeds and acceleration potentials it predicts are also in accordance with observed behaviour. SimSwim therefore holds up rather well.

IconIf you're biomechanically minded, feel free to try out SimSwim (Mac-only). Download it from here.

So how did the fossil animal swim?

Swimming I used the computer model to work out how an animal called Baltoeurypterus tetragonopthalmus, which was about 22 cm long, would have swam. Baltoeurypterus makes a good candidate for study because it was the subject of both Selden's and Plotnick's earlier hypotheses. The results show that both Selden and Plotnick got it wrong! Well they didn't get it wrong, but their theories do not represent the most efficient technique for swimming in Baltoeurypterus. The animal would have swum using an entirely lift-based stroke, with the paddle being swept forwards and down on the fore-stroke and backwards and up on the backstroke at an angle of about 45 degrees to the horizontal (right). Interestingly, the stroke mechanism suggested is identical to that observed in sea lions.

The maximum swimming speed of Baltoeurypterus can be estimated if the maximum possible paddle speed is known. Obviously it is impossible to know the maximum forces that the eurypterid was capable of generating with its paddle, but comparisons with portunid crabs suggests an upper limit of around 1 m/sec. This translates into a maximum swimming velocity of between 3 and 4 m/sec which hardly makes Baltoeurypterus a fast swimmer, but it does compare well with turtles and sea otters. The acceleration potential of Baltoeurypterus can also be estimated, again by making comparisons with living animals. The large error bars here account for the fact that the aerobic efficiency of the muscles of Baltoeurypterus is not known. The acceleration potential of Baltoeurypterus is at best comparable with that of crabs and does not reach the impressive escape responses of fish.

From its swimming technique it appears that Baltoeurypterus was a generalist, probably swimming not to hunt but to seek out new sites at which to feed on substrate dwellers like worms and smaller arthropods. This fits in with previous workers' suggestions regarding the ecology of the animal.

Finally, I examined paddles from four growth stages of Baltoeurypterus and found that the paddles from smaller animals have much higher drag coefficients than the paddles from larger animals. In other words, smaller eurypterids could produce higher drag forces than larger eurypterids. This effect is partly caused by paddle growth, which increases at a rate slower than the increase in body length.

One effect of this change in paddle drag is that smaller animals swim relatively faster than adults. Their actual speeds are, of course, lower, but juveniles were definately more sprightly than the older generation. Some workers foresaw this scenario and attributed it to the avoidance of predation and canibalism at a vulnerable stage in the young eurypterid's life. Another effect of the trend in the drag coefficient is that in smaller animals lift-based swimming mechanisms are less efficient than drag-based mechanisms. The lift:drag ratio of the paddle increases dramatically during the life of the animal. In fact, at body lengths of below 1 cm it is most likely that Baltoeurypterus used a drag-based technique, changing to the adult technique at higher swimming speeds and greater body lengths like the switch from walking to running in humans.

Hairy paddle The stroke technique employed by juvenile eurypterids is similar to that suggested by Selden, in that it is essentially a rowing mechanism, but here the stroke angle is tilted so that the paddle moves backwards and down. This is to overcome the submerged weight of the animal: Baltoeurypterus is negatively buoyant, unlike water beetles, and so must generate some upward thrust to compensate for that. The idea that juvenile eurypterids swam by drag-based techniques is supported by the observation of small hairs on the lateral margins of the paddles of small eurypterids (left). Such hairs flag the use of drag-based techniques, increasing surface area on the power stroke and feathering on the recovery stroke.

The future...?

I don't know how likely it is that this work will continue. The people with research grants just don't seem interested in knowing how fossil animals swam (for some reason). However, I think that with a little development a computer model such as mine could provide palaeontologists with a fairly reliable tool to reconstruct the functional morphology of long-extinct animals and allow them to, as it were, swim again.

Links to related sites:

IconCheck out Jason Dunlop's research into the origin of spiders.

IconLyall's fossil surgery is open for business.

IconThe homepage of my supervisor, Dr. Paul Selden. Includes more fossil/wildlife related Macintosh software.

IconReconstruction of archosaurian forelimb musculature by Mason B. Meers at John Hopkins University, Baltimore.

Made with a Mac
Last updated: 25 March 1997