In an interview that recently appeared in the Guardian, neurobiologist Colin Blakemore has overstepped the mark in his discussion of the evolution of the human brain. There are a number of problems with Blakemore’s thesis that have been covered more than adequately by Jerry Coyne and John Hawks. I wish to focus on the claim that there was an “abrupt” increase in brain size in hominins around 200,000 years ago. Blakemore presents his argument as follows:
The question is: why is [our brain] so big compared to the brains of our predecessors, such as Homo erectus? Until 200,000 years ago, there had been a gradual increase in brain size among hominins, starting three million years ago. Then, abruptly, there was a remarkable increase of about 30% or so.
John Hawks is not convinced that there is any abrupt change in cranial capacity. Referring to the above graph showing endocranial volume against time he writes:
As you can see, there’s no sudden jump 200,000 years ago, or at any other time. The data, such as they are, are consistent with a single pattern of increase over time, as pointed out by Sang-Hee Lee and Milford Wolpoff (2003).
Heck, it’s the lack of a sudden jump that has gotten all the attention. Because if “modern” humans suddenly showed up in Africa 200,000 years ago, and all of a sudden had vastly larger brains than any other hominins, wouldn’t that be a simple and tidy story? Don’t you think we’d all be talking about the sudden origin of modern humans as reflected by their larger brains?
It just didn’t happen.
I decided to take the data from the Lee and Wolpoff paper and compare the periods prior and subsequent to 200,000 years ago. As Hawks eluded to, the data can be explained by a linear model. However, this is not very helpful since we can easily fit a line or curve to just about any data. More to the point, a single fitted line doesn’t tell us much about any changes in the data. The red line in the graph below corresponds to the best fit line for the entire dataset (r = 0.81). The green and orange lines are the best fit lines for the two time periods we are considering. We can see that slopes of all three lines differ appreciably from one another. An analysis of covariance test confirms that there is a significant difference in cranial capacity between the two time periods, after we control for time. The model is statistically significant: F(1, 84) = 107, p < 0.001.
Another way to consider our data is to look at the residuals. The residuals are simply the difference between our true values and the best fit line of our model. A good way to think about residuals is to imagine rotating our data above anticlockwise until the best fit line is horizontal. Since a horizontal line has a slope of zero, it also has a zero correlation with the x-variable, in our example time. In so doing, we can consider the differences in the residuals, having controlled for time. When we compare the residuals using the best fit line the means for the two time periods (separated by a grey dashed line) are significantly different. The model is also statistically significant: t(84) = -3.9994, p < 0.001. The mean difference in cranial capacity between the two periods is 122 cc; a difference of 31%. This corresponds well with Blakemore’s figure. However, it is important to note that this is the mean difference between the two periods and does not necessarily indicate an abrupt change at 200,000 years ago.
While the numbers seem to agree with the hypothesis of a marked increase in cranial size for the later period, I think the weight Blakemore gives it is rather foolish. The fossil record is patchy and likely unrepresentative of the true cranial variation of past hominins. As Jerry Coyne rightfully points out, a geologically sudden change in the fossil record may simply reflection how erratic it is. We already saw how cranial size can change markedly in 30,000 years – little more than a blip on the time scale that we are considering here. The gradual decrease in cranial capacity since the early Upper Palaeolithic would seem geologically sudden when considered on the above timescale. The size of the fossil record is small enough that the discovery of five or six new specimens could mean having to revise our figures once again.
Another problem is that calculating cranial capacity is not an exact science. While advances have been made in calculating cranial capacity, in many cases it should still be considered a best guess (de Miguel and Henneberg, 2001). This is particularly the case for palaeoanthropological material which tends to come out of the ground fragmented and deformed. With all its drawbacks, the fossil record is often all we have to answer some of our most pressing questions. At the same time, we need to always be conscious of what the record can and cannot tell us, and avoid the temptation to tell “fanciful tales”.
De Miguel C and Henneberg M (2001) Variation in hominid brain size: how much is due to method? Homo 52: 3–58.