Technical Comments

Response to Comment on “A Complete Skull from Dmanisi, Georgia, and the Evolutionary Biology of Early Homo

See allHide authors and affiliations

Science  25 Apr 2014:
Vol. 344, Issue 6182, pp. 360
DOI: 10.1126/science.1250081


Schwartz et al. hold that variation among the Dmanisi skulls reflects taxic diversity. The morphological observations to support their hypothesis, however, are partly incorrect, and not calibrated against intraspecific variation in living taxa. After proper adjustment, Schwartz et al.’s data are fully compatible with the hypothesis of a single paleodeme of early Homo at Dmanisi.

According to Schwartz et al. (1), the fifth fossil Homo skull from Dmanisi (2) represents the species Homo georgicus (3), given its “highly idiosyncratic” mandible (D2600) and “unusual” cranium (D4500). They further implicate taxon-level differences between mandibles D211 and D2735 (1) and between crania D2280 and D2282 (4). According to these authors, Dmanisi would now comprise at least four different hominid taxa and thus hold the world record in hominid paleospecies diversity documented at a single site that extends over a mere 40 m2, and probably over a mere couple of centuries.

Morphological variation among the fossil hominids from Dmanisi offers the unique opportunity to formally test the null hypothesis (H0) that this ensemble represents variation within a paleodeme of a single species. Schwartz et al. suggest that our analyses start with H0 as an a priori truth, “primarily because [the fossils] come from the same site and a relatively short time period” (1). They do not consider, however, the logical difference between necessary and sufficient conditions. The unity of space and time of the Dmanisi fossils is a necessary—but not sufficient—condition to test H0. In other words, this condition must be fulfilled to test H0, but it does not preclude falsification of H0.

The sufficient condition for H0 is that variation among the Dmanisi hominids be similar to variation in demes of closely related extant taxa (2) (Fig. 1). We used quantitative data and statistical procedures to assess whether H0 should be accepted or rejected. Geometric morphometric (GM) analysis of three-dimensional cranial shape variation combined with resampling statistics shows that H0 cannot be rejected (2). Schwartz et al. fail to recognize that our application of GM represents a decent example of studying paleobiology “with, but not as, mathematics” (5). For example, they misrepresent the “Pinocchio effect” [which characterizes the influence of well-defined but extreme landmark positions during generalized Procrustes analysis (6)], and they confound the visual representation of patterns of shape variation [figures 4 and S7 in (2)] with the statistical analysis of shape distances [figures S8 and S9 in (2)]. Also, they did not recognize that resampling analyses were made per extant Pan deme, not pooled over Pan species (2). Furthermore, none of the GM analyses has been performed with the intention to address questions of phylogeny, as surmised by Schwartz et al. The only purpose was to test H0.

Fig. 1 Craniomandibular variation in adult Pan troglodytes troglodytes.

Frontal and lateral views of five female (A and C) and five male (B and D) skulls. To eliminate visual differences unrelated to morphology, the digital surface representations of all specimens were uniformly colored, oriented in the Frankfurt Plane, and rendered in orthographic projection (scale bar is 10 cm). Note substantial differences between individuals in gross morphology (e.g., facial to neurocranial orientation, size, and shape), as well as in morphological detail (e. g., shape of cranial vault, orbits, superstructures, maxillae, and mandibles). Compare with variation in the Dmanisi paleodeme [figure 2 in (2)].

Schwartz et al. ask for “morphological detail [that] might prove enlightening in deciphering the systematic identities of the Dmanisi hominins.” However, they ignore those analyses, which provide all the requested morphological detail [supplementary text S2 to S4, figures S3 to S5, and tables S2 to S4 in (2)]. Discrete character state variation in the Dmanisi sample and in a sample of African and Asian early Homo specimens [table S4 in (2)] in fact indicates that these fossil specimens cannot be attributed to distinct species. Schwartz et al. are incorrect to suggest that we “deny the utility of morphology in systematics.” Like other proponents of the multispecies hypothesis (7), they effectively deny the morphological evidence from Dmanisi that speaks against their hypothesis. As already noted by Darwin [(8), p. 45], “It should be remembered that systematists are far from pleased at finding variability in important characters.”

Similar arguments apply to Schwartz et al.’s description and interpretation of dentognathic trait variation in the Dmanisi sample. For example,

(i) Schwartz et al.’s assertion that premolar root number has “taxonomic valence” ignores basic comparative evidence from extant populations. Modern human sub-Saharan populations exhibit the full range of root variation, from single and Tomes’ roots to mesiobuccal + distal and buccal + two lingual roots (9). Variation in root morphology is also observed in the third mandibular premolars [P3s in (2)] of Pan troglodytes verus (10).

(ii) Schwartz et al.’s observations on P3 [same as P1 in (1)] root and crown morphology in Dmanisi are incorrect. The P3s of D2600 have two roots (not three) but three clear canals on the left P3. The P3s of D2735 also have two roots [2T configuration; see table S3B in (2)]. Both D211 and D2735 have P3 metaconids. D2735 has a large left but small right P3 metaconid. It is not possible to determine whether a fully formed entoconid was present on the heavily worn P3s of D2600.

(iii) Quantifying buccolingual compression of the mandibular canines (11) shows that D2735 has buccolingual (BL)/mesiodistal (MD) ratios of 0.88 and 0.90 (left and right canines, respectively), and D211 has ratios of 0.96 for both canines. D2600 has ratios of 0.86 (left) and 1.12 (right). Evidently, the two sides of the latter mandible do not represent two different taxa.

(iv) Schwartz et al. ignore a recent study quantifying the effects of wear-related dentognathic remodeling on symphyseal shape and orientation and on molar outline shape (12), such that they mistake in vivo modifications for taxonomic differences. Also, the symphyseal alveolar bone of D2600 (12) is far from being “disease-free” and “unaltered” (1).

Focusing on dentognathic traits might be the only practicable way to analyze large samples of fragmentary fossil remains. When applied to Dmanisi, however, this approach artificially reduces and fragments the total available craniodental evidence. Deliberately omitting integrated morphological information tends to distort natural patterns of covariation and artificially inflate perceived taxon numbers.

Clearly, addressing questions of taxonomic diversity in early Homo is of central importance, because it has major implications on inferred mechanisms of phylogeny and dispersal. Unfortunately, though, the taxonomic importance of morphological features has largely been determined by tradition, not by comparative evidence about intra- versus intertaxon variation in populations of living species (Fig. 1).

We used the name H. erectus ergaster georgicus to denote the Dmanisi sample as a paleodeme of the chronosubspecies H. erectus ergaster. The use of taxonomic quadrinomials is not regulated by the International Code of Zoological Nomenclature (article 1.3) but is not considered invalid. It should be noted that the Code of Botanical Nomenclature regulates the use of quadrinomials to denote infrasubspecific groups such as demes (local populations). Demic differences are also present in animal species, such that we maintain the long-held argument that nomenclature has to serve evolutionary and (paleo-) population biology, not vice versa (13).

Overall, Schwartz et al. (1) repeat old (14) but unsubstantiated (15) claims about taxonomic diversity in hominids, and specifically in the Dmanisi sample (4), and fail to refute the null hypothesis that the Dmanisi ensemble represents natural variation in a paleodeme of H. erectus ergaster (2).

References and Notes

  1. Buccolingual compression is measured as the ratio between BL and MD tooth dimensions at the lower (cervical) third of the tooth crown.

Stay Connected to Science


Navigate This Article