Lone Survivors by Chris Stringer Page B

easily have been inadvertently lost or even stolen in the two thefts that I suffered from my car.
    When I returned to Bristol, I took months to laboriously transcribe my data onto punch cards and start the computational wheels turning on the single, massive (but in modern terms ridiculously underpowered) computer that served the whole of Bristol University. Now, a single researcher could (if he or she knew where to look) achieve the equivalent amount of data collection, which took me four months and about 5,000 miles of travel, in a few days sitting at a computer console, summoning up online measurements and CT scans. Far more sophisticated comparative analyses of skull shape than I achieved in two further years could probably be accomplished in a few more days! Still, I have no doubt that by directly studying the fossils, I did obtain insights into their nature that would not have been apparent had I been sitting at a remote workstation. Plus I had the thrill and honor of holding and studying firsthand such iconic fossils as the skulls from the Neander Valley and the Cro-Magnon rock shelter.
    The approach I used to measure and compare my samples of fossil and recent human skulls is now called conventional morphometrics , although in 1971 this was very much the standard approach and had been in use since before the time of Charles Darwin. The human skull has various points where bones meet, where muscle markings cross a bone, or where there are specific locations such as the external earhole or the widest breadth of the nasal opening. These “landmarks” are used as measuring points so that an instrument can, for example, be laid across the nose to record its breadth at its widest point, or can measure the total length of the braincase from the top of the nasal bones at the front to the farthest point away in the middle of the occipital bone at the rear. Measurements and their variations can then be directly compared between specimens, either singly or through the calculation of an index or angle, using two or more of the measurements. For example, the cephalic index ( CI ) was a much used ratio of the breadth of a skull to its length. This index was a basic measure of how long or broad-headed a skull was, and in some of the racist science of the last two centuries it was taken as a crude measure of “primitiveness,” on the assumption that the most backward “races” had the longest heads.
    My Ph.D. work made use of angles and indexes but also extended to the then relatively novel area of multivariate analysis , where a large number of measurements could be assessed together, with specimens compared in a computed space of many dimensions, or via a single distance statistic —a bit like a ratio or index but one calculated from many measurements combined, rather than just two. However, I realized even then that my measurements were not capturing the whole complex shape of a skull, particularly some of its curved surfaces, which were poorly marked by suitable landmarks. And it was evident when comparing small and large skulls even within a single population that they might change their relative proportions as they changed in size (the study of this is known as allometry ), which was difficult to capture and visualize effectively with the techniques I had available in the 1970s.
    Today a new approach called geometric morphometrics allows far more effective visualization and manipulation of the shape of a complex three-dimensional object such as a skull. The whole shape is captured through scanning or digitizing, and virtual landmarks can be created by software at intervals across the surface of the object in question, such as a skull or jawbone. However, these secondary landmarks are still usually anchored to a network of primary points that correspond between the different objects to be compared, to provide a common frame of reference. A grid of points that reflect the overall shape (for example, of a