Lots of people have asked me what exactly MSc Human Evolution means, and I generally reply that it’s a very anatomical course looking at how human anatomy has evolved to allow us to go from tree-dwelling chimp look-alikes to upright bipedal walkers. That’s not a very detailed answer though, so I’m going to answer it a bit more here by looking at a very important technique which allows us to track changes in our skeleton as they occurred during evolution.
That technique is geometric morphometrics.
No, I hadn’t heard of it either.
That is, until I looked at what I’d be learning during the next 12 months of my life. Maybe now is a good time to find out what it is before I have to start using it.
Why not join in?
(Everything that follows is based on reference ).
It turns out geometric morphometrics (GMM) refers to measuring the shape of body parts and how they relate to the overall form of the animal (the ‘morphometric’ part).
This modern computerised method of measurement uses a laser scanner to characterise an object such as a bone by its distinguishing features including its outline and bumps for muscle attachment, and assign each feature a set of co-ordinates like those used on an Ordnance Survey map or a road atlas (the ‘geometric’ part), only in 3D. These distinguishing features are known as landmarks.
The old method involved taking manual measurements of angles and straight-line distances between landmarks using rulers, protractors and other things that are clumsy and inaccurate compared to a computer-controlled laser. Although these manual measurements were the best available at the time, it’s just not practical to take enough measurements of enough landmarks by hand for really rigorous analysis.
Fortunately, the map/grid co-ordinates of landmarks generated by laser scanning are incredibly accurate and far more numerous. In addition to this, since the co-ordinates describe the landmarks in 3D, the computer automatically knows the distances and angles between each which means no information is lost compared to the manual method.
Specimen size and shape
On individual specimens, GMM can precisely map points of muscle attachments as well as the length of the bone and how straight it is. All of these can help tell us what purpose the bone served in the whole animal, as well as how the animal moved. For example, the long, curved phalanges of chimpanzee hands are much more suited to grasping and hanging from branches than our short, straight phalanges. The same can be said of our hip bones, which have changed shape dramatically to allow us to walk upright comfortably. (More on all that in later blog posts).
It doesn’t stop at measuring one thing, though. If you have enough specimens of the same bone you can scan them all and compare them to look for differences which are down to things like gender, species or age, but this throws up a few problems.
Does the size of your bone matter? What about the position you’re using?
Bones come in different sizes because people and animals come in different sizes. How does a large bone from a female avoid being confused with the same bone from an average-sized male?
If the laser scanner doesn’t move, don’t you have to put each bone in exactly the same place every time? This is because everything is scanned relative to centre of the laser’s view.
In short, no.
Fortunately, since the whole thing is hooked up to a computer, some good maths (Procrustes analysis) can be used to make up for all these problems. Admittedly, before Procrustes has his way you’ll end up with something that looks like this:
Initially, each specimen is digitally rotated and slid about to simulate being placed in the exact same place in the scanner so all the scans have the same centre point (they are optically superimposed on one another). Then they are scaled to same size so that shape alone can be analysed.
(The size scaling once they have been given the same centre point involves measuring the distance between a landmark on one specimen and its equivalent on another, and then multiplying that distance by itself (say 2cm, so 2 x 2cm = 4cm). This is done for the same pair of landmarks on every specimen. Then add them all together. Then take the square root of that number. Now do the same for every single landmark on every specimen and compare it to every other specimen.
Aren’t you glad there’s a computer to do that?)
Now you’ve got something that looks like this:
Corresponding landmarks of each specimen are then compared to look for statistical groups which may represent family, population or species relationships depending on what you’re looking at.
The average position of landmarks in any of the groups revealed can then be displayed on a simple ‘shape space’ graph to allow easy visualisation of the general difference between groups:
Although these differences may appear small, the large number of statistically rigorous tests mean that differences are very likely to be real.
If the object being looked at is a bone, the information can reveal how muscles attached to it and hence create models of how the animal it came from is likely to move. Do this for lots of different species and you can begin to see in what order skeletal changes took place when we came down from trees and started to walk upright.
The same thing can be done for tooth size and shape to tell us about what our ancestors ate, and hence what sort of environment they lived in. Interesting, because it feeds into the debate of whether climate forced us to move from forests to savanna as the planet became drier and trees were replaced by grassland.
But that’s not all!
As well as accurately classifying differences between groups that are known to exist, GMM can also be used to look for groups that are thought to exist. This is particularly useful in the study of evolution, where all we can study is the fossilised shape of an organism. If different groups are shown to exist within a bunch of specimens then you may well have two different species all hiding under the same name. (Or even the opposite, as has recently happened with three small dinosaur species. Time to re-write science!)
It can also be used to track movement of recent and modern populations of animals across their habitat and help to see if there are distinct sub-populations to help in conservation efforts. I described a similar thing using DNA here and here.
Goodbye to GMM
Well, only goodbye from you, the reader. GMM looks set to be central to the study of bones and fossils for a long time! It can look for differences in size between males and females and tell you who ruled the roost, it can tell you what order you evolved in, it can find two species where you thought there was one and it can do all of it a whole lot more quickly and reliably than you could yourself.
Pretty good, eh?
 Dennis E. Slice (2007). Annual Review of Anthropology, Vol. 36: 261 -281. (Not freely available)
- Homo floresiensis Contextualized: A Geometric Morphometric Comparative Analysis of Fossil and Pathological Human Samples. (Freely available)