Size Variation in Middle Pleistocene Humans

See allHide authors and affiliations

Science  22 Aug 1997:
Vol. 277, Issue 5329, pp. 1086-1088
DOI: 10.1126/science.277.5329.1086


It has been suggested that European Middle Pleistocene humans, Neandertals, and prehistoric modern humans had a greater sexual dimorphism than modern humans. Analysis of body size variation and cranial capacity variation in the large sample from the Sima de los Huesos site in Spain showed instead that the sexual dimorphism is comparable in Middle Pleistocene and modern populations.

Sexual dimorphism is potentially a major source of size variation in a population (1,2). Most samples in the human paleontological record consist of specimens that span large chronologic and geographic ranges; consequently, interpopulational variation, directional trends, or diachronic fluctuations can contribute more to the sample variation (even if the sample is large) than sexual dimorphism.

In the Sima de los Huesos site in Sierra de Atapuerca, Spain, there is a large sample of human fossils that comes from a single Middle Pleistocene biological population, which provides an opportunity to investigate intrapopulational variation (3, 4). All skeletal elements are represented in the Sima de los Huesos human collection in large numbers, and the minimum number of individuals has been estimated at 32 on the basis of the dental sample (5). The fossils have been directly dated by U-series and electron spin resonance to more than 200,000 years ago, and the probable age is >300,000 years ago (6). These dates are compatible with the faunal content of the site (7). The Sima de los Huesos hominids are attributed to Homo heidelbergensis and correspond to a population ancestral to Neandertals, exhibiting a mosaic of primitive traits, combined with some (in general, incipient) Neandertal-derived traits (3, 4).

Although many methods have been designed to evaluate the degree of sexual dimorphism (2), their calculations are based on individuals of known sex or skeletally diagnosed sex. A problem in paleoanthropology is that sexual dimorphism is determined on the same feature (size) used for sex diagnosis and presupposes sexual dimorphism in order to estimate itself. Some researchers have concentrated instead on a statistical approach in which the likelihood of obtaining by chance a fossil sample with a given variation is calculated (8). It is in essence a hypothesis test in which the null hypothesis is the variation of a living species analog. If none or few samples randomly generated from the extant species show a variation greater than that of the fossil sample, the null hypothesis (that is, the hypothesis that the variation of the fossil species is the same or less than that of the extant species) is rejected. Following this statistical approach we used the bootstrap method (9) to compare the intrapopulation variation between the Sima de los Huesos and modern humans. We simulated a large number of random samples selected with replacement from extant taxa using the same sample size as that of the fossil assemblages. Then we compared maximum ratios (MR = maximum value ÷ minimum value) and coefficients of variations (CV = standard deviation × 100 ÷ mean) of the simulated and the fossil samples to assess the likelihood of obtaining by chance a sample as variable as our fossil sample. Using only adult specimens we analyzed the cranial capacity and a number of postcranial variables that show sexual dimorphism in modern humans and are related, to overall robusticity and muscularity, and hence, weight. In all bootstrap simulations, we generated for each variable 1000 random samples, with each sample equal in size to the Sima de los Huesos fossil sample size (Table 1).

Table 1

Results of the bootstrap analysis. SH, Sima de los Huesos sample; GM, geometric mean; SD, standard deviation; ISD, index of sexual dimorphism (male mean/female mean). The values given in the last two columns are percentages of the 1000 random samples with coefficients of variation (CV) and maximum ratios (MR) above the Sima de los Huesos CVs and MRs. Variables include the following: glenoid fossa GM = (glenoid fossa height × glenoid fossa breadth)1/2; humeral proximal epiphysis GM = (proximal epiphysis breadth × head vertical diameter × head transverse diameter)1/3; ulnar proximal epiphysis GM = (olecranon breadth × coronoid breadth × olecranon height × trochlear anterior-posterior diameter × coronoid height)1/5; lumbosacral surface GM = (transverse diameter × anteroposterior diameter)1/2; subtrochanteric GM = (subtrochanteric anteroposterior diameter × subtrochanteric mediolateral diameter)1/2; patellar GM = (maximum thickness × maximum height × maximum breadth)1/3; talar trochlear GM = (trochlear length × trochlear breadth)1/2; calcaneus GM = (maximum length × body length × sustentaculum breadth × body height)1/4. We followed Martin and Saller (13) for all variable definitions except for the following: vertical diameter of the humeral head (4); ulnar olecranon height, ulnar coronoid height, and ulnar trochlear anterior-posterior diameter (14); vertical acetabular diameter (15); and calcaneus body height (16).

View this table:

It has been thought (10) that crania of European Middle Pleistocene humans show more sexual dimorphism than those of modern humans, although not significantly more than those of Neandertals. In the Sima de los Huesos sample, cranial capacity can be calculated in skull 5 (1125 cm3), cranium 4 (1390 cm3), and cranium 6 (1220 cm3) (11). The Sima de los Huesos crania almost span the whole European and African Middle Pleistocene range of cranial capacities. In spite of this, the maximum ratio of the Sima de los Huesos sample is in the central part of the distribution of modern human random samples of three individuals (Table1 and Fig. 1).

Figure 1

Frequency histograms of cranial capacity maximum ratio calculated from 1000 random samples of Euro-Americans and Afro-Americans of the Hamann-Todd collection, and the percentage of those samples with a MR larger than that of the Sima de los Huesos sample. The cranial capacity for this collection was measured by, or under the direction of, T. Wingate Todd between 1912 and 1938 by the seed or water method (18). The vertical dashed line marks the maximum ratio (1.24) of the Sima de los Huesos sample (n = 3). The Afro-American sample is more homogeneous than the Euro-American sample (more diverse origin of the individuals) in the Hamann-Todd collection, and this may account for the difference between the probabilities in this figure (25% versus 44.7%).

In the Sima de los Huesos collection, the postcranial skeleton also does not show an unusual size variation compared with the distribution of samples of the same size randomly generated from large samples of modern humans. Only the humeral midshaft perimeter shows an extreme maximum ratio in the Sima de los Huesos population, although it has a less exaggerated coefficient of variation (Table 1). Sexing the fossils a priori, Trinkaus (12) found that Neandertals exhibit sexual dimorphism in the postcranial skeleton to an extent similar to that of modern human samples. Consequently, the notion that sexual dimorphism has decreased in modern populations and that the sexual dimorphism of Middle Pleistocene hominids was greater than in modern humans is not supported by either cranial or postcranial evidence from Sima de los Huesos.


Stay Connected to Science

Navigate This Article