YEARBOOK OF PHYSICAL ANTHROPOLOGY 35:243-298 (1992) Craniofacial Evidence for the Origin of Modern Humans in China GEOFFREY G. POPE Department of Anthropology, William Paterson College, Wayne, New Jersey 07470 KEY WORDS Replacement Model, Assimilation Model, Regional Continuity Model, Cladistic method, Chronophenetic assessment ABSTRACT Previously known and recently discovered hominid fossils in China exhibit a mosaic of regionally restricted and extra-regional morphological traits which evolve in the general direction of modern humans. Specifically East Asian features are concentrated in the midfacial region, while other craniofacial regions (upper face, lower face, occipital, and some basicranial features) show a gradual and general divergence in the direction of all extant human groups. Some extra-regional traits suddenly appear in the Middle and Late Pleistocene, suggesting a period of increased and intensified long range gene flow. Although difficult to define, Homo erectus exhibits a few autapomorphic and synapomorphic features (with Archaics and Neanderthals) which do not exclude it from human ancestry. The strict application of cladistic method and theory has supported the interpretation that Homo erectus was not ancestral to modern humans. An alternative chronophenetic assessment of the fossil Asians does not support a Replacement Model, but instead supports a variant of the Regional Continuity Model, a n Assimilation Model, a s the most parsimonious means of understanding the origins of modern humans in China and other parts of the Old World. o 1992 Wiley-Liss, Inc Recent advances in Chinese paleoanthropological research have major implications for the ongoing debate over modern human origins. Understandably, Western anthropologists have had only limited access to the important new discoveries and recent theoretical developments of the past few years. This review not only examines and analyzes the new chronological and morphological data deriving from recent Chinese discoveries, but also places these finds in the context of the continuing debate over the origin of modern humans. In the last two decades two important developments in Chinese paleoanthropology have significantly influenced attempts to understand the emergence of anatomically modern humans in China. The first of these was the discovery of Middle Pleistocene [0.73-0.125 million years ago (mya)] fossil hominids with craniofacial morphologies that diverge significantly from Homo erectus as it had been previously known from Zhoukoudian. The 1958 find of the relatively complete Maba cranium from southern China was easily incorporated in Weidenreich‘s (1943, 1945) gradualistic, monophyletic scheme of hominid evolution that recognized a “Neanderthal stage” between Homo erectus and modern humans. Also, several hominid specimens found in the 1970s at Xujiayao in the northern Chinese province of Shanxi that are intermediate in robusticity and morphology between Homo erectus and modern humans posed 0 1992 Wiley-Liss, Inc 244 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 relatively few problems for Weidenreich’s phylogenetic interpretation. The same was true of the Dali cranium found in 1978. However, the recent discovery and dating of the Hexian hominids and the Jinniu Shan specimen have occasioned a serious reconsideration of Weidenreich’s orthogenetic scheme that directly united extant Asians and Chinese Homo erectus. Although Jinniu Shan was originally announced as a Homo erectus (Lii, 1985), it is now accepted as an early form of Homo sapiens (Lii, 1989, 1992). In the discussion of the new Replacement Model (Wu M, 1989; Wu R, 1989, Wu X, 1989) and its implications for China, the new discoveries and their dates have become critical elements in arguments concerned with modern human origins. The second important development, the introduction of new dating techniques such as thermoluminescence and uranium series, established the possibility that the Hexian fossils, the upper levels a t Zhoukoudian, and the Jinniu Shan speciKMPS were of approximately equai antiquity. Because the Hexian specimen has features which are more “progressive” than the Zhoukoudian hominids, simple gradual unilineal change has become an inadequate explanation. Patterns of geographic variation and links to fossil Indonesian populations were also discussed (Huang et al., 1982) and continue to be debated (Dong, 1989).Also, according to the currently accepted dates, the robust Dali cranium is later than the larger-brained and thin-walled Jinniu Shan skull. In an important article addressing these new developments in Chinese paleoanthropology, Wu Xinzhi (1990) divided the craniofacial traits originally inventoried by Weidenreich (1937,1943) a t Zhoukoudian into “Common” (Asian) traits shared among fossil and modern humans in China, “Progressively Changing” (translation mine), and characters that underwent a “. . . similar evolutionary trend as those in other part[s] of the world (Wu, 1990:321).Although Wu regarded simple variation as the chief explanation of the coexistence of Homo erectus-like traits and modern traits in the same individual, in considering why Hexian had thicker cranial bones, but less postorbital constriction, he suggested: “For instance, there might be different small populations, in which the development of features might have different evolutionary speed (1990:321). In the same English summary he concluded that: “The mosaic coexistence of the possible Homo erectus autapomorphic features with features usually belonging to Homo sapiens in the same specimen indicates that in China, Homo sapiens had been evolved from Homo erectus” (1990: 321). The reasoning, though not the conclusion, represented a significant departure from Weidenreich’s (1943, 1945) original emphasis on the essentially static characters which link “Sinanthropus” with modern Asians. This approach results, to a considerable extent, both from the need to incorporate new Premodern Homo sapiens (see below) in phylogenetic reconstructions and the growing realization that while the Premodern forms are transitional in their morphology, modernization proceeded a t different rates in different anatomical traits and complexes. This is a particularly interesting moment to review and analyze the late Homo erectus and early Homo sapiens finds from China in the context of the controversy over the temporal and geographic origin of modern humans, commonly referred to as the Replacement versus Regional Continuity debate. This entails tracing the probable sequence of the appearance, persistence, or subsequent modification of individual morphological traits in the Chinese fossil record. For example, the differences in the “evolutionary speed of some character modifications which Wu discerned may in fact represent the introduction of traits originating outside of the Far East (Pope, 1991). The Replacement Model posits a chronologically circumscribed framework of between 0.2-0.05 mya for the emergence of anatomically modern humans in Africa and the subsequent replacement in all parts of Eurasia of Homo erectus andfor its descendants (Andrews, 1984a,b; Wood, 1984; Bilsborough and Wood, 1986; Cann, 1988; C a m et al., 1987; Stringer, 1981, 1988, 1989a,b; 1991; Stringer and Andrews, 1988; Delson, 1986, 1988; Rightmire, 1989, 1990; Wilson and Cann, 1992). As applied to China, the model logically predicts few, if any, indications of Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 245 admixture of Chinese fossil populations with extra-regional populations. Furthermore, any morphological traits shared between Chinese Homo erectus and extant Asians must be the result of plesiomorphic retentions, parallelisms, or rapidly developed convergences (Stringer and Andrews, 1988).One variant of the Replacement Model, the Hybridization and Replacement Model (Brauer, 1984a,b), admits the likelihood of limited admixture with immigrant populations (see also Hublin, 1985, 1986) but still postulates the complete replacement of all non-modern populations. It was originally developed to accommodate the European data bearing on the “fate” of Neanderthals and identifies the demise of the Neanderthals as the result of displacement into marginal environments andlor competitive exclusion at the hands of modern humans. Yet another recent variant of the Replacement Model postulates a n extreme form of genetic and cultural isolation that allowed for the evolution and eventual extinction of all Far Eastern Hame erectus (Clarke, 1990). This is extremely similar to Movius’ (1944) view of the cultural retardation and isolation of the Far East. Clarke argues that Asian Homo erectus never progressed beyond the “Oldowan” stage of technological complexity. This would mean that in China and other regions of the Far East the invading populations were technologically superior (contra Pope, 1984, 1985, 1989b). On the other hand, proponents of the Regional Continuity Model have argued that African morphologies are not only absent in the fossil record of peripheral areas such as China, but also that a number of morphological traits are shared between Asian Homo erectus and Asian or Asian-derived anatomically modern humans after the hypothetical period of replacement (Weidenreich, 1935, 1937, 1939,1943,1945, 1946; Coon, 1962; Wolpoff et al., 1984,1988; Wolpoff and Nkini, 1985; Wolpoff, 1980, 1985, 1986, 1989; Pope, 1988, 1989a; Wu X, 1988b, 1989, 1990, 1991; Smith et al., 1989; Thorne and Wolpoff, 1981, 1992; Jones, 1989). In Europe, Frayer (1991) has documented the persistence of regionally specific characters shared by non-modern populations and anatomically modern populations which have undergone a diachronic shift in the frequency of traits. A variant of the Regional Continuity Model, the Assimilation Model of Smith et al. (1989), emphasizes multidirectional gene flow between various regions. In this model, extra-regional genes become incorporated in long established local populations through a process of immigration or demic diffusion (Wolpoff et al., 1984; Wolpoff, 1989). As with other versions of the Regional Continuity Model, gene flow, whether of a continuous or intermittent nature, is viewed as a significant contributor to the make-up of local morphological patterns. Most of these competing models are based primarily on morphological data drawn from the paleontological record. However, in a recent defense of the “Eve Hypothesis,” Wilson and Cann (1992) criticized all morphological arguments for their reliance on anatomical traits whose functional and biological independence has not been demonstrated. This is a n astute and undeniably valid criticism of morphologically based objections to the Replacement Model. Useful and independent morphological characters are those whose presence or absence (or degree of development) do not correlate with each other. Ideally, they should be functionally unrelated, especially when considering craniofacial traits whose functional interrelationships have been historically difficult to demonstrate (See Pope, 1991, and below for a discussion). It is of the utmost importance to establish that traits are independent of each other in order to prevent the generation of long lists of separate traits which are really manifestations of the same selection pressures or developmental complexes (Brace and Hunt, 1990). Such lists give the false impression t h a t a number of different lines of morphological evidence support a particular conclusion. Specific morphological trends that have been successfully linked with a number of traits include the reduction of masticatory stress (Rak, 1983), increased basicranial flexion (Weidenreich, 1943), reduced nuchal musculature (Weidenreich, 1943; Knusel, 19911, and thinning of cranial bones in general (Kennedy, 1991; see below). Attempts to isolate biologically independent traits 246 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 explicitly independent are becoming more common (Tobias, 1991; Kennedy, 1991; Andrews, 1984b), but this important goal remains largely unrealized in attempts to elucidate hominid phylogeny. Cladistic approaches to the question of modern human origins using long lists of traits have been particularly unconvincing and have reached diametrically opposite conclusions about the role of Homo erectus (especially Asian members of the taxon) in the ancestry of modern humans. Some studies have suggested that Homo erectus is defined solely by autapomorphic traits that preclude an ancestral relationship with modern humans (Stringer, 1984, 1991; Andrews, 1984a,b; Wood, 1984, 1985). Other studies have concluded that the total morphological pattern of Homo erectus is composed entirely of plesiomorphiccharacters which may preclude recognition of the taxon as a distinct biological or taxonomic entity (Brauer, 1990; Kennedy, 1991). This study finds that many, but not all. of the chararters which have been used to define Homo erectus are plesiomorphic when judged by strict cladistic standards (cf. Delson et al., 1977). Additionally, the insistence of autapomorphic traits to define the taxon and the failure of the cladistic method to incorporate chronometric data limit its usefulness in elucidating infrageneric phylogenetic relationships necessary to determining the likely origins of anatomically modern Homo sapiens. By employing the concept of the total morphological pattern (Le Gros Clark, 1955; Tobias, 1991) within a chronometric framework, this analysis concludes, along with Wu X (1988a, 19901, that the craniofacial traits that the Chinese fossil hominids exhibit are a mosaic of both pangeographic traits diverging in the direction of modern humans and specific regional midfacial traits which link them with extant Asians. Although previous cladistic approaches and results are examined in detail, the principal methodological technique emphasized here is chronophenetic in its orientation. It is similar to the stratophenetic approach of Gingerich (1979), in that it explicitly examines and incorporates the dating and morphology of the hominids in an attempt to test the predictions of the various phylogenetic models that have been advanced to describe modern human origins. This review and analysis assumes that phylogenetic interpretations are most reliable when they integrate paleontological and chronological data. CHRONOLOGY OF THE LATER CHINESE HOMINIDS Pleistocene chronology, long a controversial component of Asian paleoanthropological studies, has matured dramatically in the past decade and remains critical in the evaluation of fossil hominid discoveries. In addition to isotopic and paleomagnetic data, mammalian biostratigraphic and paleoclimatic studies have been extensively employed and refined in the past few years as a means of dating critical hominid specimens. Because of limited technical resources, the dating of late Middle Pleistocene and early Late Pleistocene (ca. 0.125-0.010 mya) hominids may not be as reliable as in Europe (Cook et al., 1982; Poulianos, 1971; Stringer, 1981), Africa (Klein, 19891, and the Middle East (Mercier et al., 1991; Stringer and Grun, 1991; Grun et al., 1990; Schwarcz et al., 1988, 1989; Arensberg et al., 1985). However, China currently affords the most securely dated evidence for addressing the question of the timing of modern human origins in Asia. In China, the same dating techniques have been used as in the West, but only recently. These independent techniques have produced consistent results (see Wang, 1989, for an extensive summary of Chinese dates; Guo et al., 1991; Li and Wang, 1982,1991).Few convincing data have yet emerged for the synchronic and sympatric existence of different species of Homo, although the possibility of the synchronic and sympatric existence of Homo erectus and Premodern Homo sapiens has been raised (Zhang, 1989; Chen and Zhang 1991; Semah et al., 1990). Mammalian biostratigraphy (Han and Xu, 1985,1989; Qi, 1989,1990) and loess chronology (An et al., 1991; Ding, 1991; Liu, 1991a,b) continue as important means of establishing dates for Chinese Pleistocene finds. These have been further extended in attempts to produce oxygen isotope stage correlations for many hominid Popel CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 247 bearing sites (Qi, 1989,1990). Uranium series, thermoluminescence, fission track, amino acid racemization, and, more recently, electron spin resonance (ESR) techniques have been the primary means of generating radiometric dates for Chinese Middle and Late Pleistocene hominid localities (Wang, 1989; Yuan and Chen, 1991). The methodological reliability of some of the dates has been the subject of some criticism (Grun, personal communication). Aigner (1986, 1988) has also suggested on the basis of paleoclimatic data that the important site of Zhoukoudian Locality 1 actually spans only a single interglacial. The majority of Chinese and also Western workers who possess a firsthand familiarity with the Chinese data have not supported these criticisms. The divisions of the Chinese Pleistocene continue to be based to a large extent on biostratigraphic and paleoclimatic data now tied to regional tectonic processes influencing a succession of paleoclimatic oscillations (Liu, 1991a; Liu and Ding, 1984a.h) The old aqwmptian thst the biostratigraphic ailLiessiuii and paieociimatic fluctuations of China were synchronous with European glacial stages (Lee, 1939; Sun, 1991) has now been replaced by a regionally specific tectonic-paleoclimatic framework (Xu and You, 1982,1984).This new model links biostratigraphic and lithological sequences to the continuing uplift of the Qinghai-Xizang (Tibetan) Plateau and its influence on paleomonosoonal fluctuations in East Asia (Zhou and Wang, 1991; Wang and Li, 1991; Derbyshire, 1991; Kuhle, 1991; Rutter, 1991; Itihara et al., 1991; Agrawal, 1991). This interpretation has begun to take on almost paradigmatic status in Chinese Pleistocene chronology (An et al., 1991; Liu, 1991a,b; Li, 1991a; Wang and Yang, 1991). The tectonic uplift of western China possibly may be responsible for the world wide global climatic oscillation of the Quaternary. As part of this maturing multifaceted model, Pleistocene China is not only divided into the previously recognized northern and southern faunal zones, but also into a three-tiered progression from east to west characterized by increasing altitude and climatic seasonality (Qi, 1990). In fact, during periods of low Pleistocene sea levels, a fourth step, the now inundated continental shelf, undoubtedly played a significant role in the migration and dispersal of hominids and other vertebrates. The altitudinal distribution of the Chinese hominid and Lower Paleolithic localities (Fig. 1) reveals that no hominids have been discovered from above 1,000 meters in altitude. Some hominid fossils actually derive from within the “second step,” but these localities are situated within erosional features (river valleys, canyons) which are actually less than 1,000 meters in elevation. Although the lowland nature of fossil faunas associated with hominid finds has been discussed before iZheng and Han, 1991), the implications for the paleoecology of fossil hominids have rarely been explored (Pope, 19881. As in other regions of the world, certain faunal and lithostratigraphic sequences have become regional chronometric and biostratigraphic “yardsticks.” In northern China (north of the Qingling Mountains), Zhoukoudian Locality 1 has been the traditional biostratigraphic sequence with which all other northern Middle Pleistocene finds are compared. More recently, the extensive fluviolacustrine Nihewan Formation has become the type sequence for the Pliocene, Pleistocene, and Holocene of northern China, but no hominid specimens have yet been recovered there (Liu and Xia, 1983; Wei et al., 1985; Wei and Xie, 1990; Luo et al., 1988; Cheng et al., 1978; Jia and Wei, 1980, 1987; J i a and Huang, 1990; Wei, 1978, 1988, 1990; Wei and Xie, 1990; Wei et al., 1985; Pope et al., 1990; Zhang et al., 1991; Pope and Keates, 1992). The loess-paleosol accumulation a t Luo Chuan and other areas in Shanxi from the Central Loess Plateau is another important lithostratigraphic reference sequence for the Quaternary of northern China (Fig. 2) (An and Lu, 1984, 1987; Yuan et al., 1991; Liu, 1991a,b). The sequence is not only considered to be well dated, but also widespread enough to permit reliable correlations with the oxygen isotope sequence of deep-sea cores. The BrunhesiMatuyama Boundary and its dating is thought by the majority of Chinese workers to fall in the seventh to eighth paleosol coincident with oxygen isotope Stage 20 (Liu T et al., 1985), but 248 YEARBOOK OF PHYSICAL ANTHROPOLOGY Late Homo erectus m Premodern (early Homo sapiens) 0 Anatomically modern Homo sapiens (Modified from Qi 1990; Liu and Ding 1984) A [Vol. 35, 1992 ezzI Mean altitude = 4,000 hz=3 Mean altitude = 1,000-2,000 rn 0Mean altitude = ~ 1 , 0 0 0rn Fig. 1. The geographic distribution Chinese fossil hominids. From east to west, China is divided into three altitudinal steps resulting from the continuing uplift of the Qinghai-Xizang Plateau throughout the Pleistocene. Chinese fossil hominids are restricted to lowland sites at altitudes lower than 1,000 meters. Their apparent occurrence on the second step actually represents specimens recovered from erosional features equal in altitude to sites of the first step. some disagreement remains as to where this actually occurs (see Qi, 1989). The correlations and the chronological positions of the horninid localities (in northern China) shown in Figure 2 are based to a great extent on these sequences. The result is still schematic, but a great deal of tectonic, geomorphological, biostrati- Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 249 graphic, and radiometric data has been assembled in support of this approach to Chinese Pleistocene chronology (Liu, 1991a,b). In southern China this chronological framework has been difficult to use, not only because of the very limited distribution of loess paleosol sequences, but also because most of the hominid localities are cave deposits (Qi, 1990; Zhou and Wang, 1991; Tang, 1991; J i , 1991; Huang, 1991a,b; Huang and Fang, 1991). Instead, the sequence of the advance and retreat of ostensibly “cold adapted” mammalian forms has served as the primary biostratigraphic means of estimating the ages of hominid localities. Although the extensive and highly fossiliferous fluviolacustrine Yuanmou Basin series has sometimes been correlated with cave faunas, the series has only rarely been used to estimate the age hominid finds from southern China (Qian et al., 1991). The extensive coastal plain in north and south China facilitated faunal migration that resulted in climatically “mixed” faunas over widely varying 1d- l . t d e b , further complicating biostratigraphic correlations (Qi, 1989, 1990; Han and Xu, 1989; Huang, 1991b). The establishment of chronologically restricted index taxa has remained a problem throughout China, and is especially acute in southern China (Huang, 1991a,b). CHINESE FOSSIL HOMINIDS The descriptions and discussion of specimens presented here are based on the author’s study of originals, casts, photographs, line drawings, and written descriptions assembled over the course of a number of years. With the exception of Jinniu Shan, Yunxian, and Yiyuan, all of the recently discovered specimens discussed in this paper have been figured in Wu et al. (1989). Although some finds have been partially described and figured in previous compendia (Aigner, 1981; Wu and Olsen, 1985), most of the new data have appeared in Chinese (Wu et al., 1989). Some English translations and reviews have recently become available (Wu and Lin, 1985; Wu and Xu, 1984; J i a and Huang, 1990; Etler, 1990; Wu, 1991; Liu, 1991a,b; Li and Etler, 1992), but most of these largely descriptive works do not present detailed phylogenetic analyses. In the discussion which follows the descriptions and dating of the specimens, the emphasis is on the identification of biologically independent traits which can be tentatively identified as regionally persistent, gradually modified, or rapidly appearing non-indigenous traits. The assumption that rapidly appearing traits are extra-populational in origin andlor punctuational in situ mutations is of course open to argument. However, most, and quite possibly all, of the traits identified in this paper as rapidly appearing are known earlier outside of China (Pope, 1991‘1 T t is strongly suggested that this approach is currently the most parsimonious means of constructing testable hypotheses arising from the Regional Continuity versus Replacement debate. Taxonomic and morphological considerations The specimens considered here range from middle to terminal Middle Pleistocene members of Homo erectus to what this paper terms Premodern Homo sapiens (or simply Premoderns) and to anatomically modern Homo sapiens. For reasons presented in Pope (19911,the term Archaic Homo sapiens or Archaics is used only for Middle Pleistocene specimens from Africa and Europe (e.g., Bodo, Jebel Irhoud, Kabwe, Petralona, Arago, Steinheim) which have been generally recognized a s exhibiting distinctive, if informally recognized common morphologies (Cook e t al., 1982; Klein, 1989; Howell, 1978, 1986; Howells, 1981). Neanderthals are not considered to be Archaics. The term Premoderns is used in both a geographic and morphological sense. This category includes only early Homo sapiens from China t h a t are not demonstrably modern in their morphology. With the possible exception of Maba, no broad morphological equivalency between EuroAfrican and East Asian samples of early Homo sapiens can be recognized. Although Premoderns are thought to be restricted to China, the Narmada specimen (Kennedy et al., 1991) may also be ultimately assigned to this group on morpho- YEARBOOK OF PHYSICAL ANTHROPOLOGY 250 (Modifiedfrom Qi, 1990, 1989; Yuan and Chen, 1991; Liu et al., 1984) most probable chronological position - well dated estimated range paieosol UiEd loess iiii Fig. 2 [Vol. 35, 1992 Popel CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 251 logical grounds. It is important to continue to employ separate terminologies for these widely separated geographic groups because there are substantial morphological differences between the Archaic Euro-African and Neanderthals on the one hand and the Premodern Chinese hominids on the other. The fossil samples A number of Chinese localities have yielded remains which have been assigned to Homo sp. (Gao, 1975; Pope, 1977, 1988; Wu and Olsen, 1985; Wu et al., 1989; Zhang, 1988) or early Homo erectus on both morphological and chronological grounds. However, these specimens are too early and fragmentary to be directly relevent to a discussion of the origin of anatomically modern Homo sapiens. These include Ji a n Shi, Hubei; Gongwangling and Chenjiawo, Shaanxi; and Yuanmou, Yunnan (see Wu and Dong, 1985a; Dong, 1989, for a complete list). While the main focus here is on late Homo erectus from China, Hvmo erectus specimens trom J a v a (Sangiran 17, Sangiran 4,and Ngandong) are also included for comparative purposes. Late Chinese Homo erectus specimens directly relevant to the question of modern human ancestry include Zhoukoudian I11 and V and Hexian, Anhui. Homo erectus specimens of more doubtful age are known from Yiyuan, Shandong and Yunxian, Hubei and are considered in some detail. The newly discovered Yunxian material is particularly important and allows for morphological comparisons to the well preserved Premodern Asians from Dali, Jinniu Shan (Fig. 3), Maba, and also Hexian (Fig. 4). The following specimens also document the widespread distribution of Homo erectus in China, but are too fragmentary for morphological comparisons: Yunxian Dragon Bone Cave, Hubei (not the Yunxian material discussed below, Wu and Dong, 1980); Xixian, Henan (Wu and Wu, 1982); Yunxi, Hubei (Qun, 1983); Nan Zhaoxian, Henan (Qiu et al., 1982); and Luonan Xian, Shaanxi (Xue, 1987). Specimens of Premodern Asians complete enough for inclusion here are identified by their site name (synonym, local name, or previous orthography) and Province. These include: Chaoxian (Chaohu), Anhui; Changyang, Hubei; Yunxian, Hubei; Maba (Mapa), Guizhou; Dingcun (Ting Ts’un), Shanxi; Dali, Shaanxi; Jinniu Shan (Yingkou), Liaoning; and Xujiayao, Shanxi. The inclusion of the fragmentary, but chronologically important Xujiayao hominids in the Premodern group ultimately may prove to be premature. In this study only one demonstrably anatomically modern Homo sapiens is considered in any detail. This is the well preserved Liujiang cranium from Guangxi. I t is important not only because of its excellent preservation, but also because it has been assigned a n extraordinarily early date, which if correct would predate all other known modern humans. Anatomical foci Weidenreich’s (1943) original descriptions, measurements, and comparisons of the Zhoukoudian cranial specimens continue to influence contemporary analyses of Chinese fossil hominids. From Weidenreich’s (1943) list of 121 features “characterizing ‘Sinanthropus’,” subsets of traits ranging from a few dozen (Howell, 1978; Stringer, 1984) to fewer than ten (Andrews, 1984a,b; Brauer, 1990) have been used in phylogenetic arguments. The choice here is to emphasize a few anatomical regions which seem to best display variations between the Premoderns and Chinese Homo erectus: i.e., the face (in frontal and lateral aspects), the glenoid region (in basal aspect), and the occipital region (in posterior and lateral aspect). These particular regions are cataloged below. Fig. 2. Chronometric positions of Middle and Late Pleistocene Homo erectus and Premodern Homo sapiens. Recent attempts a t Chinese chronology have been based on radiometric calibration of the loess paleosol sequence and its correlation with oxygen isotope stages. The exact placement of the Brunhesi Matuyama boundary is open to question, but usually considered to occur in S7 or L8. The dating of the hominid bearing section of Zhoukoudian Locality 1 is well dated by a variety of isotopic techniques. 252 YEARBOOK OF PHYSICAL ANTHROPOLOGY Fig.3 IVol. 35, 1992 Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 253 1. Face. The regions of the face have been divided in various ways. Here the face is defined as any part of the cranial skeleton that can be observed in frontal view (cf. Rak, 19831, from vertex to gnathion; the upper facial skeleton extends from nasion to vertex. The midfacial region extends from nasion to prosthion (or alveolare), and is bounded horizontally by the distance between the most lateral extent of the cheekbones (zygion-zygion);the lower face consists of the mandible. In frontal view, the lateral protrusion of the alveolar torus of the maxilla is sometimes also visible posterior to P3 (lateral alveolar prognathism, Pope, 1991). The zygomaticomaxillary region of the middle face, including the IZM (Inferior Zygomaticomaxillary Margin, Pope, 1991) that extends from the lateral maxillary wall to zygion (Rak, 1983) is discussed as well. Weidenreich (1943) also noted a prominently developed malar notch (“incisura malaris”) and its consistent association with the malar tubercle of “lateral prominence.” The dentition of the midface (especially maxillary inrisors) have also figured prominently in past rnorphologicai analyses of Asian hominids. These are also included here, where appropriate. 2. Glenoid region. The glenoid region, on the inferior surface of the temporal bone, is the area surrounding and including the temporomandibular joint (Fig. 5). It was described in detail by Weidenreich (1943) and more recently by Kennedy (1991) and Hill et al. (1992). Once again it has come to play a major role in phylogenetic arguments. The major features of the glenoid that have been widely cited are discussed below and do not include all of the features Weidenreich (1943) thought to be phylogenetically significant. The plesiomorphic state of the hominoid glenoid region (Fig. 5A) takes the form of a low, poorly developed mastoid (1)whose anterior surface is separated from the posterior surface of a roughly tubular (in sagittal cross-section) tympanic bone (3) by a broad mastoid groove (2) or tympanomastoid fissure (Tobias, 1991). The entire region is highly pneumaticized. The tympanic plate which forms the anterior wall of the tympanic bone is separated from the glenoid fossa [(4) mandibular fossa or temporomandibular fossal by the postglenoid process (51, which may take a variety of forms. In most hominoids, the fossa is broad and shallow. The fossa is bounded medially by the entoglenoid process (6),which contributes to the medial wall of the fossa, but is separated from the tympanic bone by a deep gap or medial recess (7). Anteriorly, the fossa is bounded by a rounded, gently sloping anterior wall that grades into an extensive preglenoid planum (8). The derived condition of modern humans (Fig. 5B) is variable, but exhibits a broadly contrasting morphology. In comparison with hominoids, the entire region appears generally “compressed” in the anterioposterior direction. The results is that the tympanomastoid groove is absent or reduced to a very narrow fissure <1 mm in breadth). The posterior wall of the tympanic is pressed against the anterior wall of the mastoid. In anatomically modern humans the bone is much less pneumaticized, and thinner. The entoglenoid process is greatly reduced and participates more fully in the medial wall of the fossa. A medial recess or gap between the entoglenoid process and the petrous may be variously developed. The glenoid fossa as a whole is much deeper and oval shaped and is limited posteriorly by a more vertical tympanic plate and a much steeper anterior wall that grades into a much shorter preglenoid planum or articular eminence. In cross-section, the tympanic bone is not as round and can have an elliptical, triangular, or rectangular shape. The anterioposterior dimensions of the preglenoid planum are greatly reduced. In contrast to hominoids and Homo erectus, a distinct styloid process (9), partially surrounded by a flange of bone [vaginal process (VP in Fig. 5B)I is present in modern humans. In Homo erectus (see below), there is neither a styloid process nor vaginal process, but there is a process supratubalis (PS in Fig. 5 C ) . Fig. 3. Line drawings of facial views of Premodern Homo sapiens from Dali (A) and Jinniu Shan (B) and Homo erectus, Yunxian I1 in norma lateralis (C). 254 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 a I Hexian --* - Yunxian b Fig. 4. Yunxian I1 compared with Dali and Hexian. The sagittal contours reveal that the face of Yunxian I1 is noticeably taller than Dali (A). The great anterior-posterior length of Yunxian I1 (B) may be only minimally partially an artifact of postdepositional deformation. 3. Occipital. The thickness and curvature of the occipital bone have been important parameters in morphological discussions and phylogenetic analyses of Homo erectus and other early hominids. In lateral and posterior aspects, the evaluation Popel CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 255 Fig, 5. The left glenoid region of a chimpanzee (A); anatomically modern Homo supiens (B), and the Weidenreich (1943)Homo erectus ( C ) viewed in norma basalis. 1, Mastoid process; 2, tympanomastoid fissure; 3,tympanic bone; 4,glenoid fossa; 5,postglenoid process; 6, entoglenoid process; 7, medial recess; 8, preglenoid planum; 9,styloid process. Also identified are the vaginal process (VP) and the process supratubalis (PS). of the prominence of the occipital torus have been considered important. In posterior aspect, the lateral extent of the torus and its development near asterion has led to the idea that more fusiform lateral portions of the torus are more like 256 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Val. 35, 1992 modern humans. In other words, a robust torus has generally been associated only with Homo erectus, while an evenly rounded or bun morphology (Neanderthals) has been considered characteristic of Homo sapiens. In internal aspect, the areas of the cerebellar and cerebral fossae were identified originally in “Sinanthropus” as being distinctly different from modern humans in their relative proportions. Moderns have a cerebellar fossa to cerebral fossa ratio of about 2:1, or a 100%larger area devoted to the cerebellar fossa. In “Sinanthropus,’) the ratio is reversed, with the cerebellar fossa only 50% as large as the cerebral fossa (Weidenreich, 1943). This trend is related to the forward displacement of the cerebrum and also possibly to the relative growth of the cerebellum. Related to this trend and the more rounded sagittal contour of the occipital is the close approximation (<11 mm) of endinion (center of the cruciate eminence) and inion (midpoint of the occipital torus in the sagittal plane) in modern humans. In general, a lesser endinion-inion distance seems directly associated with rounding of the occipital as a whole. In “Sinanthropus” endinion and inion are well separated (Weidenreich, 1943; Stringer, 1984). Although the degree of separation is generally a good indicator of anatomical modernity, it is important to point out that lesser values for this character are also exhibited by Neanderthals and some phylogenetically more distant forms such as Olduvai Hominid 5 (Kennedy, 1991, and discussion) which also have more rounded brain cases in comparison with Homo erectus. In sum, besides cranial capacity, features that have been most widely employed in morphological analyses of early Homo sapiens and Homo erectus are the structure of the face, development of ectocranial superstructures, cranial bone thickness, the glenoid region, general vault curvature, and, less frequently, dental characters of the maxillary and mandibular dentition (I1and I’ shoveling and M3 agenesis). Dental and gnathic characters form a relatively small part of the analysis which follows, especially since no mandibular fragments are known from early Chinese Homo sapiens (but see Weidenreich, 1937). Of course, not all the specimens preserve parts which allow the assessment of all these characters in each individual (Table I). Zhoukoudian Locality 1 Any review and analysis of the fossil evidence for the origin of modern humans in China must begin with the 40 m of stratified hominid and artifact bearing sediments at Zhoukoudian Locality 1. The continued manifold importance of the sequence lies not only in the quality of the multidisciplinary research (see Weidenreich, 1935; Zhou, 1990), which led to the initial formulation of the Regional Continuity Model, but also in the geochronological framework which was established there. In China, Zhoukoudian has served for over half a century as the chronometric and paleontological “yardstick for all subsequent discoveries. The careful descriptions and thorough comparisons of the Zhoukoudian hominids, as well the excellent casts produced by Black and Weidenreich, originally defined the morphological parameters of the current debate about modern human origins in China. However, recently the Zhoukoudian Locality 1 sequence has been the focus of renewed examinations which have questioned both its chronology (Aigner, 1986, 1988) and cultural significance (Binford and Ho, 1985; Binford and Stone, 1986). Binford and colleagues’ “reexamination” of the evidence from Zhoukoudian strongly contends that the inferences of cultural behaviors at Zhoukoudian (the use of fire, hunting, repeated occupation, and sapient behavior in general) have been based on faulty interpretations of faunal remains and site features. In fact, Binford and colleagues’ analyses were based on highly dubious statistical manipulations of a small fraction of the total faunal evidence, studied over a short period of time. There is no reason to believe that this stratigraphic sequence does not result, for the most part, from repeated hominid occupations. Aigner has argued on paleoclimatic grounds that the 40 m hominid bearing sequence spans only about 0.1 my and not the ca. >0.2 my indicated by Chinese Popel CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 257 studies (Wu and Dong, 1985a,b). The concurrence of most of the 35 or more dates (Li and Wang, 1982) strongly suggests that the actual time represented is in excess 0.2 my. Aigner’s own palynological and faunal data does, however, convincingly suggest that severe climatic fluctuation has been exaggerated in previous attempts to correlate the sequence with the long spans of European glacial oscillations (See Qi, 1989, 1990, for a summary). The Locality 1 sequence is now a reasonably well dated hominid occupation site that spans most of the late Middle Pleistocene (but not the terminal Middle Pleistocene). The hominid bearing section all falls within the later Brunhes paleomagnetic Chron (Liu et al., 1977). Although a t least 45 hominids have been recovered from layers 3-11, three-quarters of the individuals are from layers 8-10, and the majority of the remaining specimens are from layers 3-4 (Aigner, 1981; Wu et al., 1985). Usually treated as a single stratigraphic unit, layers 1-3 (breccia about 5 meters thick) have hem dated hy ura.ninrn series QE mammalian teeth te bet.h-eeil 0.23 2 0.030 and 0.23-0.26 2 0.08010.050 mya (see Wang, 1989, for a complete list of Zhoukoudian dates by layer). Thermoluminescence dating of burned cherts from layer 4 (4-8 m in thickness) has produced values of 0.292 2 0.26 and 0.312 0.026 mya (Pei, in Wang, 1989). Fission-track dating of sphene separated from ash in layer 4 has produced an age of 0.29 2 0.050 mya (Guo et al., 1991). This is consistent with previous fission track dates (Liu S et al., 1985). This value is also consistent with ESR (Electron Spin Resonance) dates recently obtained by Huang et al. (1991). Guo et al. (1991) also conclude that both the layer 10-11 thermoluminescence dates derived from burnt cherts of 0.610, >0.417, and <0.592 mya and a uranium series date on bovid horn of 0.340 0.110.06 mya (Wang, 1989) are valid. One of the most important dates in terms of reliability was obtained on burnt crystal from layer 10 which yielded an age of 0.462 2 0.045 mya (Liu S et al., 1985). Amino acid dates have produced less reliable values which are as much as 200% older than the results of radiometric techniques (Zhou, 1989). The level by level faunal changes remain unclear despite the adoption of a rigorous plotting system (Aigner, 1986). Faunal elements considered to be diagnostic of the terminal hominid occupation interval a t Zhoukoudian Locality 1 include; Macaca robusta, Equus sanmenensis, Megaloceros pachyosteus, Canis lupus, and Crocuta ultima. At Zhoukoudian Locality 1,very few faunal differences separate the upper and the lower hominid bearing levels. The most important contrasts which do differentiate faunas of the earlier Middle Pleistocene (layers 8-10) from the later (but not terminal) Middle Pleistocene (layers 1-4) are the absence of Megantereon inexpectatus and Hyaena brevirostris sinensis. The chronological issues raised by Aigner (1981) center around the question of whether the general lack of faunal turnover is due to a relatively short span of time documented by the sequence or prolonged climatic equability (Pope, 1983, 1988). The stability of most other Middle Pleistocene localities in northern China strongly supports the interpretation that the sequence of hominids here spans approximately 0.2 my. The question of diachronic trends at Zhoukoudian Probably because of his belief in a “ ‘Pithecanthropus’-‘Sinanthropus’ stage” and also because he was unconcerned with stratigraphic relationships in general, Weidenreich never discussed the evidence for diachronic change a t Zhoukoudian (see especially von Koenigswald, 1951, 1949). Although he argued for the overall reduction of size and massiveness in the course of hominid evolution, the changes he perceived were between very broadly conceived stages. However, with the 1966 recovery of fragments from layer 3 that fit with the Zhoukoudian V fragments from the 1934 and 1936 excavations, an early and late comparison between Zhoukoudian V and Zhoukoudian XI became feasible. Unfortunately, as Table 1shows, the comparisons were between a juvenile and a female(?). In their description of Zhoukoudian V, Qiu et al. (1973) identified a number of “progressive” morphological changes evident in the sequence of Zhoukoudian hominids. They include: 1) thinner supraorhital torus; 2) thinner cranial boncs; 3) a * * YEARBOOK OF PHYSICAL ANTHROPOLOGY 258 [Vol. 35, 1992 Table I . Chinese Fossil Hominids Cranial specimens (one X J Y dental specimen) SexiAge ZhouS.O. S.O. S.O. S.O. S.O. koudian Cranial EndinionTorus Torus Torus Torus Torus layer capacity inion thick. thick. (L. thik. (R- thk. (L.(R(level) (ML) (alt. value) (R- medial) medial) middle) middle) lateral) TI. T. ZKD . __._ Wjuv.? (Calvarium)* ?lad. ZKD I1 (Calvarium) ZKD X (Calvarium) Wad. ZKD XI Flad. (Calvarium) ZKD XI1 Wad. (calvarium) ZKD V (Calvarium) ?F/Mlad. Yunxian I (Parti. ? cranium? Yunxian I1 (Partl. Wad. cranium ) Wad. Hexian (Calvarium) Yiyuan (Cranial ?lad. frags.) Dali (Partl. Wad cranium) Jinniu Shan (Partl. Wad. cranium) Maba (Partl. Flad. cranium) Chaoxian (Maxilla F? and occ. frag. Changyang (Partl. 7 maxilla) XYJ 1 (L. partl. Wjuv.? max. frag.)** ? XYJ 2 (L.U..M. 2) XYJ 3 (L ?. parietal ? frag.) XYJ 4,5 (L. parietal ?!ad frags.) XJY 6 (Parietal F?lad. frag.) XJY 7 (Occipital ?lad. frag.) XJY 8, 9 (Parietal ?!ad. frags.) XJY 10 (R. parietal frag.) Misub.-ad. XJY 10 A (R. Wad.? mand. frag.) XJY 11,12 (L.? parietal frags.) ? XJY 13 (R. parietal ? frag.) XJY 14 (R. and L. ? frag. a t bg.) XJY 15 (Occipital Wad. frag.) ? XJY 16 (L.L.molar? XJY (Temporal M?lad. frag.-no #) Chaoxian (occipital F?!ad. frap.) Liujiang (partial cranium) M? Y 11 (46) 8-9 (40) 8-9 (40) 915 1030 1225 27.5 38.0 14.0 8-9 (40) 1015 34.0 13.0 8-9 (40) 3 (23) 1030 1140 35.0 34.0 (36) 16.0 12.1 >1350?? 1025 - - - - - - 22.0 18.0 19.0 17.0 16.0 13.0 - - 1120 11.0 14.3 - 10.4 <1300? - - 22.0 - - - - 10 (8) - - - 22.0 - +ZKD = ZHOUKOUDIAN LOCALITY 1 **XJY = XUJIAYAO __ 12.0 - 12.8 19.6 13. 16.2 12.8 11.5 12.7 12.7 16.5 - 13.2 13.8 14.6 10.1? 15.6 12.5 - 10.8 - 12 14 (7) CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS Pope1 259 Table 1. Chinese Fossil Hominids (continued) S.O. TONS Thk. Llateral) 12.0 11.2 Parietal (thk. bregma) 9.6 9.0 7.5 PariThk. etal (Emi- frontal nence) gl . 23.0 11.0 20.0 11.0 23.0 12.5 CereTem- bellar poral fossa (Thk. thickcenter) ness Frontal (Thk. center) Parietal (Thk. center) 10.0 10.0 7.0 9.3 10.0 5.2 9.3 10.0 5.2 6.8 Ccerbli Crb ratio % Ft-Ft Temporal length Temporal height - - Temporal index - (5.0) 50 50 50 81.5 84.0 98.0 70.5 39 55.3 - - 7.0 16.0 18.7 11.0 6.0 6.0 2.8 50 84.0 62 35.5 57.3 (14.5) 11.1 9.7 9.1 14.5 12.4 22.0 7.0 11.0 7.0 10.0 7.0 10.0 2.5 4.5 50 50 91.0 97.0 74.0 71.5 36.5 35 49.3 49 - - - - - - - - - - - - - 13.8 - 11.2 - 9.0 7.0 7.0 3.9 66 106.0 - 5.5 6.5 - 5.0 - 4.0 3.0 - 114.0 - - - - 7.0 9.0 - - - - - - 96.0 - - - 70 42 60 260 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 higher and rounder parietal border of the temporal squama; 4) a decreased distance between endinion and inion (internal occipital protuberance); and 5) reduced occipital torus. Their assertion is supported by only some of the metric data. In Zhoukoudian V, portions of the supraorbital torus areas small or smaller than the lowest values for the rest of the Zhoukoudian sample (Table 1).However, in all cases the difference is within the margin of error. Similarly, in Zhoukoudian XI, the thickness of the vault bones (measured at the center of the frontal squama, the parietal eminence, the center of the temporal squama, and at the cerebellar fossa) is within the Zhoukoudian range. In comparison with the other Zhoukoudian specimens, the height of the temporal bone in Zhoukoudian V is less than in any other, but it is more arched in shape along the parietal border. The lengtwbreadth index of Zhoukoudian V is nearly identical to Zhoukoudian XII. On the occipital bone, the endir,ion-ir,ion value for Zhoukoudian V is equal to t h e (XI) female from layers 8-9, and well within the Zhoukoudian range. The value given by Weidenreich (1943)was in fact a minimum. My own measurement of the cast was of 36 mm, the second highest value in the Zhoukoudian sample (see below). The central thickness of occipital torus was also a minimal estimate (Weidenreich, 19431, but still falls within the range of the Zhoukoudian sample. The prominence of the occipital torus is slightly reduced and only a very shallow supratoral sulcus is present. Therefore, three of the five observations of Qiu et al. are not supported by the metric data. In a more recent examination of temporal variation at Zhoukoudian, Zhang (1991) took issue with Weidenreich’s (1935) claim that dental variation a t Zhoukoudian was due only t o sexual dimorphism and not to temporal change. Using length and breadth measurements from a sample of 336 teeth from the upper and lower layers at Zhoukoudian, Zhang (1991) showed that the mandibular teeth underwent clear reductions in size. The lower canine and P, reduced slightly, while P, reduced the most. M, remained virtually unchanged. The upper sample M, and all the upper sample M, specimens fell below the range of the lower group. He could demonstrate little change in the maxillary teeth, possibly because the maxillary sample from the upper layers was inadequate. Zhang’s study clearly shows dental reduction a t Zhoukoudian. The possibility of increasing cranial capacity in Homo erectus in general has been the subject of considerable controversy (Leigh, 1992; Lestrel, 1976; Wolpoff, 1984; Rightmire, 1986a,b, 1990). A stratigraphic comparison of cranial capacities reveals part of the reason for the confusion. As Table 1shows, the most complete crania are represented by six calvaria which include one female and one juvenile. The difference in cranial capacity between the earliest and smallest specimen. Zhoukoudian I11 (a male juvenile), and the latest specimen, Zhoukoudian V (an adult male), is 225 ml. This represents an increase of approximately 20% in cranial capacity in the 40 m of the vertical sequence. If only the average of 1,075 ml for the four adults (including the one female) is considered, then over a vertical span of less than 40 m, cranial capacity increases by less than 15% (160 mi) of the layer 8-9 average. If the female (XI) is removed, and only adult males are included, the average cranial capacity is 1,095 ml for the layer 8-9 hominids. Compared to Zhoukoudian V this is a difference of 45 ml, or about 4%. However, the greatest difference in cranial capacity, 25% or 310 ml, is between the earliest specimens, Zhoukoudian I11 (a juvenile) and Zhoukoudian X (an adult male) from layer 8-9, rather than between the earliest and latest specimens. It is therefore possible to manipulate the Zhoukoudian cranial capacity data to obtain a variety of widely differing results. The fact that the range of the layer 8-9 hominids (210 ml) is nearly as large as the range of the entire sequence of Zhoukoudian hominids (225 ml) may indicate that the sample has the characteristics of a single contemporaneous population and should be treated as such when included in larger statistical arguments about trends in cranial capacity over the course of the Pleistocene. However, there is a direct correlation between cranial capacity and a decreased Popel CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 261 endinion-inion distance. In fact, Weidenreich’s estimate for Zhoukoudian V is a minimum. Using his estimate of 34 mm yields a correlation coefficient of r = .79 between cranial capacity and endinion-inion distance. My own estimate of 36 mm (based on a cast) yields a coefficient of r = .82. However, as with cranial capacity, there is no directional diachronic trend for this dimension in the Zhoukoudian sequence. The relationship between cranial capacity and endinion distance is important since they have been frequently used as independent traits which define Homo erectus. In summary, there is no obvious monotonic correlation between stratigraphic provenience (and presumable date) and the reduction of supraorbital thickness or thinner cranial bones. The thickness of the supraorbital torus of Zhoukoudian V, in both its middle and medial portions, is smaller than all the other Zhoukoudian hominids except for Zhoukoudian X, which it exceeds by 0.1 mm. The temporal of Zhoukoudian V is mticeablj: more arched along its parietal margin, but its absolute height is less than any other Zhoukoudian hominid and most comparable in its temporal index to Zhoukoudian XII. The glenoid region The glenoid region has also figured heavily in phylogenetic analyses of Asian Homo erectus (Stringer, 1984; Andrews, 1984a; Kennedy, 1991). Features of this region have been taken alternatively as plesiomorphic or autapomorphic characters of Homo erectus. Similarly, Weidenreich’s (1943) description pointed to both primitive and unique features (see Fig. 5C). One of the major features which he emphasized was a deep mandibular fossa bordered anteriorly by a very vertical anterior wall and posteriorly by a posterioinferiorly inclined and very thick tympanic plate, sometimes fused to an indistinct postglenoid process. Medially, the fossa is not bounded by the concave and weakly developed entoglenoid process usually seen in modern humans. Instead, a well developed entoglenoid process protrudes into the fossa (Fig. 5C). The entoglenoid process is separated from the tympanic bone by a wide medial recess. The width of this gap, when it is present, is much less in modern humans. The retention of a wide tympanomastoid fissure separating the coronally oriented tympanic and the mastoid process is also obvious. As noted previously, this is reduced to a very thin fissure in modern humans. The petrous pyramid is not aligned with the transversely oriented tympanic, but instead intersects the medial end of the tympanic at an obtuse angle. The petrous portion is thick. Styloid and vaginal processes are absent and instead a pronounced and blunt tubercle (“process supratubalis”) is present. Weidenreich viewed most of these features of the glenoid region as either primitive retentions shared with “anthropoids” (by which he clearly meant hominoids), or as features which were intermediate in morphology between hominoids and modern humans. However, two features which he emphasized as unique to “Sinanthropus” were the extreme depth of the glenoid fossa and its medial placement near the sagittal plane relative to both modern humans and hominoids. In these and other characters of the region he regarded Neanderthals as more similar to modern humans. He also believed that most of the differences between modern humans and “Sinanthropus” resulted from thinner vault bones and larger brains in modern Homo sapiens. These and the other features identified a t Zhoukoudian continue to function as major parameters of phylogenetic comparisons. Hexian The Hexian specimens not only document geologically recent Homo erectus from China, but also what is perhaps the last known occurrence of this taxon anywhere. The 1980-81 finds from Longtandong (Dragon Pool Cave) in coastal Anhui Province are very important discoveries. One of the specimens, a calvarium, represents the first reasonably complete Chinese specimen of Homo erectus from outside the northern temperate zone. Despite its location in subtropical southern China, its morphology closely approximates the Zhoukoudian crania (Dong, 1989). However, 262 YEARBOOK OF PHYSICAL ANTHROPOLOGY IVol. 35, 1992 Huang et al. (1982) have argued that it is more similar to the Ngandong hominids. Regardless of its geographic affinities, it is widely regarded as the latest surviving and most “progressive” representative of Homo erectus in China (Huang et al., 1982; Wu and Dong, 1982; Wu M, 1983, 1989). A few of the features may be progressive, but as the comparison below demonstrates, most of the supposedly progressive features of Hexian fall within the Zhoukoudian sample. Some of the features actually exceed the Zhoukoudian sample in robusticity and size. Of more importance is the possibility, which the discovery raises for the first time, that early Homo sapiens and Homo erectus were contemporary. This possibility is unique to China. The initial 1980 hominid discoveries represent a t least three individuals and consist of 1)a calvarium lacking most of the cranial base, 2) a partial left mandibular fragment preserving M i and Ma, and 3 five isolated teeth (right P4, left M1, left MI, left M1, and left M,) (Wu and Dong, 1982). In 1981, Wu (1983) recovered a strongly weathered frontal fragment preserving the right supraorbital region and a small part of the squama, a right parietal fragment, and additional isolated teeth (right 11, left M1, right M2, and two left MJ. These specimens were recovered from a different part of the deposits, but from the same level (Wu, 1983). The dating of the locality is based on a uranium series date (Chen et al., 1987), an amino acid racemization date, and two TL (thermoluminescence) dates derived from above and below the 0.5-1 m thick layer that yielded the hominid specimens. TL dates on chert bracket the cranium between 0.184 t 0.015 and 0.195 t 0.016 mya (Li and Mei, 1983). A uranium series date on mammalian teeth and distal podia1 bones has a value of 0.15-0.27 mya and a racemization age based on animal teeth is given simply as 0.2-0.3 mya (see Wang, 1989). The associated mammalian fauna are a mixture of southern and northern faunal elements which include Ailuropoda melanoleuca, Ursus arctos, and U . thibetanus (Huang and Huang, 1985). The biostratigraphic age, based to a large extent on the presence of Macaca robusta, Hyaena breuirostris sinensis, Megaloceros pachyosteus, and Megantereon sp., has been taken as indicating an antiquity approximately equal to or slightly later than the upper-most hominid occupation levels at Zhoukoudian Locality 1 (ca. 0.2 mya). The I1 is noticeably shovel shaped and displays a well developed centrally placed basal tubercle and mesial and distal marginal ridges that extend to the occlusal surface. Wu (1983) has pointed out that this morphology is very similar to homologues from Zhoukoudian and Krapina. In its mesiodistal and buccolingual dimensions of 11.7 mm and 8.1rnm respectively, it falls well beyond the dimensions of all other known Homo erectus homologues, but within the range of Krapina. With the exceptions of M,, the occlusal dimensions of the teeth are broader and longer than Zhoukoudian. The M, is smaller than any of the Zhoukoudian specimens (see Wu and Dong, 1985b, for tabulated dimensions). These findings are very consistent with the dental trends established for Zhoukoudian (Zhang, 1991) and Pleistocene China in general (Brace et al., 1984). The mandibular corpus is massive and exceeds the Zhoukoudian ranges in thickness, height, and robusticity. Three mandibular foramina are present below P, and M,. Dong (1989) emphasized that the robusticity and maximum width of the mandible were consistent in development and dimensions with those of the calvarium, implying that both specimens probably represented a single individual. Previous workers have agreed that the morphology of the Hexian cranium can easily be accommodated within Homo erectus (Wu and Dong, 1982; Dong, 1989). However, much debate has centered around whether it is more similar to Zhoukoudian Skull V (Wu, 1983; Dong, 1989) or more similar to the Ngandong specimens (Huang et al., 1982). The cranial capacity of 1,025 ml is smaller than the average for both Zhoukoudian (1,059 ml) and the Ngandong hominids (1,096 ml). The cranial vault is thick and falls within the Zhoukoudian range at the parietal eminence and temporal squama and the center of the frontal squama. There is a Popel CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 263 well developed supraorbital torus, a n angulated occipital, a bregmatic eminence, and a posteriorly inclined frontal squama. The thickness of the supraorbital torus substantially exceeds that of all the other Zhoukoudian hominids except in comparison with the medial and middle thickness of Zhoukoudian 11. Slightly more convincing “progressive” features of Hexian are found in the frontal region. These include less postorbital constriction; the greatest inferiosuperior thickness of the supraorbital torus restricted to the middle portion of the brow ridge (in the 1981 fragment); larger and slightly more laterally expanded frontal sinuses; a broad and shallow supraglabellar fossa extending to the middle portion of the orbits; and a broader frontal with ft-ft (frontotemporale) falling between the averages for Zhoukoudian and Ngandong, but closest to the Zhoukoudian sample (Wu and Dong, 1982; Wu X, 1989). However, if the value of 93 mm for ft-ft is used, Zhoukoudian V exceeds this value. However, Huang et al. (1982) give a value of 101 mm which would put Hexian beyond the Zhoukoudian range. The frontal squama exhibits a less developed sagittal keel or torus which extends from the posterior half of the squama past bregma and along the sagittal suture where it disappears unaccompanied by parasagittal depressions. In norma occipitalis, the greatest cranial breadth occurs a t the level of the mastoid crest. There is considerable, though asymmetrical, development of the angular tori. The occipital is thick, and the internal and very weakly developed external occipital protuberances are separated by a distance of approximately 22 mm (Wu and Dong, 1982; Dong, 1989). This is 5.5 mm lower than the lowest adult Zhoukoudian value of 27.5 mm and close to the 22 mm values of Ngandong 6 and 11.There is a slight supratoral sulcus (Dong, 1989).The occipital torus is restricted to the medial portion of the occipital and disappears completely before reaching the lamdoidal suture. Wu (1990, 1991) accurately described it as “fusiform.” The occipital angle (ca. looo) falls within the Zhoukoudian range (98-105’) and the occipital squama is shorter than the nuchal squama. The Hexian cranium is broad and exhibits a reduction of those ectocranial superstructures normally not associated with more robust crania, such a s Zhoukoudian. In basal aspect, the Hexian cranium exhibits a deep glenoid fossa with a massively developed inferiorly convex articular eminence. The glenoid fossa is larger and more expanded in both its anterioposterior width and transverse length than the Zhoukoudian specimens. Though not as deep a s the glenoid fossa in Ngandong 6 (Solo V), the overall shape is similar. The right glenoid fossa exhibits only a slight medial recess and the medial end of the fossa is not patent, again similar to the conditions seen in Ngandong 6. The tympanic bones are situated very medially as in Zhoukoudian and oriented perpendicularly to the midsagittal plane. In crosssection, they are triangularly shaped and thicker than in the Ngandong 6, Trinil 11, and Sangiran 4 specimens. The petrosal crest of the tympanic is well separated from the mastoids by a broad tympanomastoid fissure. On the internal aspect of the basicraniurn, it is possible to note that the cerebellar fossa is about 75% the size of the cerebral fossa (Wu X, 1989). With the exception of the size of teeth, no strong diachronic trends have been established at Zhoukoudian. For this reason, the Hexian calvarium should not be regarded as being the end of any strong morphological trajectory within Homo erectus. I t is, however, very close to Zhoukoudian V. One undoubted value of the Hexian material lies in its extension of the range of variation for Homo erectus. The material also demonstrates, in conjunction with the Zhoukoudian material, that there is no simple continuous diachronic decrease in robusticity. The contemporaneous 1981 parietal and frontal fragments document the presence of a more gracile individual that falls a t the lower range of the Zhoukoudian sample in cranial thickness. Taken together with the Hexian dental sample, Chinese Homo erectus displays a large range of variation in dental robusticity throughout its entire record. The massive dentition and small cranial capacity also fail to suggest any strong divergences in the direction of modern humans. 264 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 Yiyuan The Yiyuan specimens are from an excavated karstic fissure at Qizi Hill in Yiyuan County, Shandong. With the exception of an abundance of Melgaloceros pachyosteus specimens, the associated fauna are not chronologically diagnostic. A tentative terminal Middle Pleistocene age has been assigned to the hominids (Lu et al., 1989). The hominid specimens consist of a left and a right farietal, frontal fragments, occipital fragments, and seven teeth (C, two P4, two P , and two M2)from at least two adult individuals. The occlusal dimensions of the teeth all fall within the known Zhoukoudian range. All the specimens are very fragmentary. In thickness the cranial bones are comparable to Zhoukoudian specimens. The Yiyuan specimen is 9 mm thick at bregma. This value falls within the middle range of values for Zhoukoudian (7-9.7 mm) and is greater than those for either Maba or Jinniu Shan. However, some specimens from Xiijiayao have Comparable thickness (Xujiayao 4 and 5 at 9 mm) or even greater thickness (Xujiayao 3 a t 12.4 mm). At asterion, the cranial bone thickness in Yiyuan (13 mm) is only 0.5 mm less than the minimum value found in the Zhoukoudian sample, and well within the range of error. The frontal fragments from two different individuals preserve a well developed (especially laterally), well defined but postdepositionally deformed, continuous supraorbital torus and a sulcus (Lu et al., 1989). In the more robust toral fragment, the outer table is thicker than the inner table. No sinuses are observable in either specimen. It is difficult to tell from the fragmentary specimens if a sagittal keel was present. All the available information on the specimens warrants their provisional inclusion in Homo erectus. Given the geographic location of the Yiyuan finds, in the southern part of Northeast China, future radiometric dating of this locality could be critical in arguments concerning the synchronic presence of Homo erectus and Premodern Homo sapiens in China (see especially Chen and Zhang, 1991). Along with the Hexian find, the fragmentary Yiyuan finds may represent one of the latest occurrences of Homo erectus in China. Yunxian Yunxian, Hebei, has yielded two very complete, though distorted crania which contrast sharply with previously known crania from the later Middle Pleistocene of China. It is important to emphasize that this description is based on the study of published scientific descriptions, scaled photographs, and slides. However, the Yunxian crania are important enough that even the preliminary data warrant as detailed a treatment as possible. Reports of these two specimens are still based on a few articles (Li J, 1991; Li et al., 1991; Li and Etler, 1992).Except for the Lantian specimens (Gongwangling and Chenjiawo) and a few isolated teeth, they are the only Homo erectus in China deriving from fluvial deposits. The first specimen, Yunxian I (EV 900) was originally discovered in the pit fill taken by a farmer from a nearby bank of earth (Li T, 1991; Li et al., 1991). Subsequently, more than 60 artifacts and a second, more complete specimen, Yunxian I1 (EV90021, were recovered in situ. The two crania were reported as being separated by a vertical distance of less than 0.4 m and a horizontal distance of 3.3 m. The artifacts include bifacially worked artifacts such as “choppers” and trihedral picks, as well as smaller flake artifacts (Li and Etler, 1992). The dating is extremely problematic, with the presence of Hyaena brevirostris licenti suggesting middle Middle Pleistocene age. However, the presence of very hypsodont bovine of Bos sp. and Bubalus sp. suggests a later or terminal Middle Pleistocene age. On the bases the large size of associated specimens of Tapirus, Hystrix hodgsoni, and rhinoceros it is only possible to place the fauna in the Middle Pleistocene. A uranium series date of 0.40 mya has been reported by Li T (Public lecture August 9, 1991; Beijing). Yunxian I preserves nearly the entire middle and upper face, but the specimen Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 265 is severely distorted and deformed. The supraorbital torus is prominent and most developed laterally. The original size of the supraorbital planum (no true sulcus is present) is impossible to determine as most of the superior neurocranium is flattened. The nasal aperture is extremely wide at the base, spanning a t least one third of the width of the maxilla, and a prominent nasal spine is present. The specimen also exhibits pronounced subnasal prognathism. A deep incisura malaris is present. The second specimen (Yunxian 11)is much more complete and preserves most of the cranium and facial regions, with only the mandible lacking. I t is clear that this specimen is morphologically extremely similar to Yunxian I. The neurocranium is noticeably depressed in the post-bregmatic area. The right side of the frontal may retain its original contour. However, it is unclear whether the right side of the frontal has been inferiorly depressed, or the left side everted (Fig. 3C). The rest of t h e qperirnen appears t9 be disterted to some degree. The entire specimen possibly has undergone superior-inferior compression, possibly causing posterior elongation. Li T (Public lecture, August 9, 1991, Beijing) originally suggested a cranial capacity in excess of 1,100 ml, but later increased the estimate to greater than 1,350 ml. However, no basis for this estimate was given. Most recently any estimation of cranial capacity was considered to be impossible (Li and Etler, 1992) but Yunxian I1 is clearly the largest Homo erectus cranium ever recovered in China. In norma lateralis, the distance from glabella to opisthocranion in Yunxian I1 is also longer than in any other previously recovered fossil hominid in China though the reported length of 217 mm (Li and Etler, 1992) must be confirmed by reconstruction. Figure 4 shows traced scaled sagittal contours of Yunxian 11, Hexian, and Dali, with A showing the positioning of the Yunxian I1 cranium at the occipital contour. Figure 4B shows a comparison with the Yunxian I1 specimen aligned with the anterior contour of the uppermost Hexian face and the much more complete Dali face. Regardless of the degree of distortion which one attributes to the Yunxian cranium, it is clear that the cranium is both relatively and absolutely long and that the face is massive by Chinese standards. The Yunxian I1 face is set perpendicularly and the degree of distortion and crushing resulting from postdepositional deformation appears minimal on the left side of the face. The inferior medial corner of the orbit has been pushed forward, creating a false impression in norma lateralis of a very horizontally oriented anterior nasal contour. In fact this contour is actually created by the anteriorly everted orbital portion of the maxilla. A broad and shallow postorbital sulcus clearly demarcates the supraorbital region from a posteriorly positioned squamosal portion of the frontal. The anterior portion of the supraorbital torus is broken and not preserved. In sagittal contour it is apparent that the frontal bone has only very slight bregmatic bossing, less than that observable in Zhoukoudian homologues. The frontal of Yunxian I1 is more similar to Dali and the poorly preserved Jinniu Shan frontal than to the Ngandong specimens. In general, ectocranial superstructures are only moderate in comparison with other Homo erectus. In lateral view, the orbits are rather short. Judging from the left side of Yunxian 11, the original shape of the orbit is rectangular with a n angular superiomedial corner and a rounded inferiolateral corner. The impression of roundness in the right orbit of Yunxian I1 is probably a n artifact of deformation. The midline contour of the nasal aperture is straight and vertical. The sagittal contour of the subnasal portion of the maxilla reveals strong alveolar prognathism. From the inferior border of the nasal aperture to interdentale the contour of the subnasal region appears “hook shaped.” This probably results from damage to the anterior alveolar portion. Nothing like this is known in other Chinese fossil hominids. The zygomatic process of the maxilla is higher and much thicker than in Gongwangling, Sangiran 17 and Sangiran 4. An occipital torus is present, but i t is fusiform and associated with a shallow supratoral fossa and a distinct rounded or buniform structure. In comparison with other Chinese finds, this feature of Yunxian I1 seems most similar to the Jinniu 266 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 Shan specimen, whose marked “occipital bun” was an artifact of the original reconstruction that failed to articulate the posterior part of the neurocranium with fractured margins of the parietals (see below under Jinniu Shan). This has been corrected and the new Jinniu Shan reconstruction reveals a smaller, but still protruding buniform structure. The morphology of this region is similar to Swanscombe and Vertesszollos. In norma verticalis the minimal frontal breadth is wider than that found in Zhoukoudian Homo erectus homologues. There is no lateral supraorbital trigon. A distinct but shallow continuous postorbital sulcus is visible. Based on estimates of the dimensions and proportions, the size of the face in both specimens is unprecedented among known Far Eastern Homo erectus specimens. The right inferior zygomaticomaxillary margin of Yunxian I1 seems to show a distinct and well developed incisura rnalaris sitaated medial to, or in the saint: plane, as the lateral wall of the orbit. This is very similar to the condition found in Zhoukoudian specimens (Pope, 1991) and is more developed than in Sangiran 17 and Dali. There is no posterior alveolar prognathism. This contrasts to the condition seen in Archaic Homo sapiens. Despite their damaged condition, the inferior lateral corners of the orbit are rounded. The supraorbital torus is thickest in the medial two-thirds of the orbit. In comparison with other Chinese and Javanese specimens of Homo erectus and Premodern Homo sapiens, the torus is foreshortened. Compared with Yunxian I and other Premodern Chinese, the margins of the orbits are sharper or more salient. In norma basalis, the Yunxian specimens appear to be extremely long from opisthion to opisthocranion. This is partially the result of post-depositional crushing. The dental arcade is “u” shaped with the posterior dentition arranged in parallel rows. The incisors are reduced relative to the other teeth. All are well worn and have distinct basal tubercles. Though worn, the lateral incisors also display distinct marginal ridges suggesting an originally well developed, shovelshaped morphology (contra Li and Etler, 1992). The canine is broad. The first maxillary molar is preserved only on the right side and is slightly narrower than M2 and M3. The left maxillary third molar of Yunxian I is larger than M2, and the size relationship of the posterior dentition is M3 > M2 > M1. However, Yunxian I1 has a small, almost vestigial “peg shaped” left M3. This would be the earliest appearance of this trait, which has its highest frequency in extant East Asians. The glenoid fossa of Yunxian I1 is deep and the tympanic appears oval, but perhaps originally round in cross section and is oriented perpendicularly to the midsagittal plane with its lateral margin extending nearly to the edge of the squamous portion of the temporal. The petrous portion is oriented in a similar fashion. There is a broad tympanomastoid fissure. The anterior third of the foramen magnum lies anterior to the tympanic-tympanic axis. The shape of the foramen magnum is noticeably rounded rather than oval. In summary, the Yunxian crania, though badly distorted, are well enough preserved to discern a combination of both previously known Asian features and novelties. The obvious Asian features are an incisura malaris, flat vertical zygomatics, shovel-shaped incisors, a peg-like M3, and possibly a very long cranium. Features similar to specimens from western Eurasia and Africa include massive faces, broadly arching nasal bones, and highly placed massive cheek bones. In their total morphological pattern, they are unique in East Asia. The significance of this pattern will become apparent in light of the Premodern specimens described below. Chaoxian The Chaoxian hominid finds were recovered in 1982 and 1983 from the excavation of a brecciated fissure-fill in Chaoxian, Anhui. Two distinct fauna-bearing levels are present a t the locality. The hominid material comes from the upper of the two fossiliferous layers and is apparently temporally distinct from the lower pocket, which has yielded fauna of an Early Pleistocene age. A uranium series date of 0.167-0.2 mya has been reported from a sample of nine dental and postcranial Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 267 mammalian specimens from the upper hominid bearing level (Chen et al., 1987). The stratigraphic divisions and mammalian fauna listed by Huang and Huang (1985), Xu et al. (1984,1986a,b), and Han and Xu (19891,and summarized by Etler (1990), differ slightly from the more recent data presented by Xu and Zhang (1990), which delete Bison sp. and Caprinae from the upper hominid-bearing layers. Fifteen species of mammals have been reported from the hominid bearing section of the deposit (Liu et al., 1982). These have been correlated with layers 1-4 a t Zhoukoudian (Wu X, 1989). All reports agree that the hyaenid present is Hyaena brevirostris sinensis. As mentioned previously, the presence of this species and the absence of Crocuta ultima is considered an indicator of a time period earlier than Zhoukoudian layer 4. The absence of Megantereon (which last appears in layer 5) is usually taken as an indicator of the middle Middle Pleistocene. The hominid material consists of an occipital without the portion below the inferior nuchal line, which is damaged slightly near lambda. Also preserved are a maxilla containing right P3 to M1 and three isolated maxillary teeth (P3,M1, and M2).As the M’s are antimeres, the dental and gnathic specimens probably belong to the same individual. On the bases of the relatively thin occipital bones, estimated biasterionic breadth, and the lack of well developed ectocranial markings on the occipital, Wu X (1989) suggested that the specimen was that of a 26-year-old female, but such an exact determination is speculative. The occipital torus is confined to the medial portion of the occipital and diminishes laterally. Although discernible, a supratoral sulcus is not well developed. On each side, the torus curves downward, and is considerably thinner than all the Zhoukoudian specimens. The distance between endinion and inion is the same as the Hexian specimen (22 mm). The maxilla has less pronounced subnasal prognathism than Homo erectus, including Yunxian. The inferior nasal aperture is bounded by a palpable nasal sill that appears to differ from Homo erectus homologues, which exhibit a smooth transition from the subnasal clivus into the floor of the nasal aperture. The size of the Chaoxian cheek teeth falls within the upper-mid range of Chinese Homo erectus, but between the values for Changyang. The incisors have been described as “broken off a t their roots” (Etler, 1990). However, Zhang (1989), who worked with the original material, maintains that the lateral incisors are worn, particularly on the right. Taken as a whole, the fragmentary nature of the Chaoxian specimens and their estimated dating suggest that, a t least for the present, these specimens should provisionally be regarded as early Homo sapiens (Premodern Asians) that display both morphological (occlusal crenelations) and metric similarities to other Homo erectus dental remains, as all the dimensions of the teeth fall within the middle and upper range of the Zhoukoudian sample (Zhang, 1986). However, neither the presence of a Carabelli’s cusp nor the reconstructed occlusal dental morphologies offer definitive evidence of a direct phylogenetic relationship with Chinese Homo erectus. Neither the uranium date nor the associated fauna preclude an age which is intermediate between the upper Zhoukoudian Locality 1hominids and anatomically modern humans. According to Xu et al. (1986), the most convincing reasons for placing this specimen within Homo sapiens are the thinness of the occipital and the vertical orientation and anterior placement of the incisive foramen. Xu et al. (1986a,b) also acknowledge that the more gracile characters of the occipital may be attributable to a female and thus do not necessarily indicate the specimens’ placement in early Premodern Homo sapiens. Changyang In 1956, a partial maxilla and isolated teeth were recovered from a cave near Changyang, Hubei. Based on uranium dates on mammalian teeth, the specimens have been accorded ages of 0.194 2 0.024/0.02 and 0.196 0.020/0.017 mya (Yuan et al., 1986). Most of the associated cave fauna are not temporally diagnostic. However, the absence of Crocuta and the presence of Hyaena brevirostris sinensis sup- * 268 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 port a late Middle Pleistocene age. The fauna in many cases has only been identified to the familial or generic level and a more detailed re-study of the sample might allow a more precise age estimate. The partial left maxilla preserves the canine, p,and M2 as well as the broken roots of the incisors. An isolated left P, also was recovered from the cave deposits. Wu and Wu (1985) and Wu X (1989) suggest that the anterior position of the incisive foramen, a large maxillary sinus extending anteriorly forward of P3, and a very orthognathic subnasal planum indicate affinities with Homo sapiens. From a measurement of the orientation of the root sockets of the incisors, Wu X (1989) concluded that the anterior dentition distinguished them from other early Chinese Homo sapiens in exhibiting a much more anteriorly sloping dentition. The extension of a large canine jugum superior to the floor of the nasal aperture may also indicate non-modern morphology. Xujiayao Xujiayao is the largest single sample of Premodern Asian Homo sapiens recovered in the Far East. The numerous Xujiayao hominids have remained one of the least discussed, but one of the most important samples (Table 1). These specimens are both temporally and morphologically intermediate between anatomical moderns and Premoderns. Excavations in the 1970s yielded a number of hominids from Locality 74093 on the banks of the Liyi stream at Xujiayao, Shanxi. At least 16 hominid specimens representing at least 11 individuals of various individual ages have been recovered (Jia et al., 1979; Jia and Huang, 1990; Wu, 1980, 1986; Wu X, 1989). All the hominids derive from a 4 m-thick sequence of sediments 8-12 m below the surface from a large excavation (surface area of approximately 5,000 m2). However, the horizontal relationship of the finds is unclear (Aigner, 1981), and the dating of these specimens has been the subject of continuing controversy. One uranium date from four samples of teeth and distal phalanges of equids from “below 12 meters” in the section yielded a value of >0.1 mya (Chen et al., 1982). Another date based on rhinoceros teeth from “below 8 meters” in the section yielded an age of 0.104-0.125 mya (Chen et al., 1982, 1984). Two 14C dates on molluscan specimens and rhinoceros teeth yielded contradictory ages of >0.04and 0.1692 2 0.002 mya (Chen et al., 1982). The mammalian fauna includes supposedly cold adapted forms (Equus przewalskyi, Megaloceros ordosianus, Cervus elaphus, Coelodonta antiquitatis, and Bos primigenius) and the paleoclimatic regime has been characterized as cold and possibly earlier than L1 of the loess sequence (oxygen isotope Stage 5) and coincident with the penultimate glaciation (Qi, 1989). The numbers of individuals of the more than 22 species of mammals (some of them apparently burned) are heavily skewed toward Equus przewalskyi. Together with thousands of artifacts Xujiayao has been regarded as evidence for a kill/butchery or occupation site (Jia et al., 1979). Given the numbers and diversity of the hominid sample from Xujiayao, it will be important to establish a more precise date for the Xujiayao hominids. The Xujiayao hominid specimens have been described in detail by various authors as showing a mosaic of characters similar to both Zhoukoudian Homo erectus and modern hominids (Jia et al., 1979; Wu, 1980, 1986; Wu and Wu, 1985). The hominid specimens include a nearly complete left temporal bone, 11parietal fragments, two largely complete occipital fragments, the ascending ramus of a mandible, a partial left maxillary fragment preserving an unerupted canine, M’, and M2, and two isolated molars and one incisor (M, and M2, and Ill. The cranial vault specimens are as thick or thicker than Zhoukoudian specimens. The occipital angle of the one measurable specimen yields a value of 116”,well above the Zhoukoudian range, and 1”below the modern range given by Wu (1980). This specimen also shows weak development of an occipital torus and associated inferior sulcus. In curvature and size the vault bones are intermediate between Zhoukoudian and Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 269 modern humans. Both occipitals are believed by Wu X (1989) to have possessed Inca bones. One left temporal bone has been tentatively referred to a young male. It is comparable to Amud I in its thickness, position of the zygomatic process, and the dimensions of its less rectangularly shaped squama (Wu, 1986). The glenoid fossa is as deep as those seen in Zhoukoudian specimens and preserves a distinct postglenoid process similar to Dali, a well developed articular eminence, but no articular tubercle (Wu X, 1989). The fossa is much longer transversely than any of the Zhoukoudian specimens, but similar in depth to the smaller Zhoukoudian individuals. The oval shape of the fossa is decidedly more similar to modern humans and Dali (Wu X, 1989) than it is to Zhoukoudian. Also, as in modern humans, there may have been a styloid process which is not seen in Homo erectus. There is a wide tympanomastoid fissure and the mastoid is very thick, falling at the lower end of the Zhollkoudian range. The maxillary fragment is that of a child and exhibits less anterior prognathism than the Zhoukoudian specimens and no lateral alveolar prognathism. The occlusal enamel crenelations visible on the teeth are similar to Zhoukoudian specimens. The incisor is shovel shaped, but with a reduced basal tubercle and shorter marginal ridges which do not extend to the occlusal surface. All the teeth are large in occlusal dimensions and fall just within or outside the upper range of Zhoukoudian specimens. Not only do the Xijiayao finds appear to date to a time period equivalent to the appearance of anatomically modern humans in Africa and the Middle East (Grun and Stringer, 1991; Grun et al., 1990, 1991), but they also provide the largest sample of hominids which bridge the temporal and morphological gap between Homo erectus and Homo sapiens. Furthermore, the specimens present a combination of characters of less developed ectocranial superstructures (a reduced occipital torus, reduced occipital angulation, reduced parietal curvature, and expanded cerebellar fossae) in conjunction with greater general robusticity in some specimens (thick cranial bones and large dentition) and probably more marked sexual dimorphism than is generally seen in modern populations. This combination occurs elsewhere in roughly contemporaneous specimens. Despite claims to the contrary (Stringer, 1989a,b; Delson, 1988), the absolute anatomical modernity of the earliest African and Middle Eastern “moderns” has not gone without challenge (Smith et al., 1989; Caspari and Wolpoff, 1990; Wolpoff and Caspari, 1990). The Xijiayao specimens seem to represent a similar grade of morphological organization that not only exhibits similarities to locally derived Homo erectus, but also to extant populations occupying the same region (Wu X, 1989). Though the date for the Xujiayao specimens must remain tentative for the present, their discovery underscores the possibility that “progressive” morphology is just as old in China as in other parts of the Old World. Dingcun Dingcun refers to a series of localities of different ages discovered and excavated in 1953 and 1954 on the eastern bank of the Fen River in Shanxi. The uranium date of 0.16-0.21 mya for the one hominid locality (Locality 54:lOO) is neither supported nor contradicted by the associated fauna. The fauna includes Myospalax fontanieri, Megaloceros sp., Bubalus sp., and Bos primigenius. There is also some doubt as to the original provenience of the hominid specimens (Pei et al., 1958; Aigner, 1981). The Din cun hominids are few in number and consist of a right M,, and a right shoveled IB and I2 from a juvenile estimated a t 12 to 13 years of age. These teeth are much smaller than comparable ones from Zhoukoudian and Xujiayao and provide some of the earliest and most complete evidence of dental reduction in Homo sapiens. A subsequent find of a right parietal may include an Inca bone (Wu X, 1989). The primary significane of Dingcun lies in the archeological documen- 270 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 tation of bifacially worked “picks”bearing a striking resemblance to artifacts from Chongok-Ni, Korea (Pope, 1988). Dali The morphology of the Dali specimen contrasts with Homo erectus, Neanderthals, and European Archaics and has assumed a pivotal role in arguments about the origins of modern humans. Stringer (1988, 1989a,b, 1990) has repeatedly argued that its facial morphology is similar to Archaics and therefore offers no support for uniquely East Asian characteristics. In fact, the facial morphology shows a number of features which make it an excellent transitional form intermediate between Asian Homo erectus and extant Asians. The Dali cranium was recovered from a fluvial sequence in northern Shaanxi in 1978 (Wu; 19811 nespite a iiraniiim d a t e nf 0 209 0.23 mya (Chen et 3!., 19841, considerable doubt remains as to the age of this important specimen. The geological and chronological contexts of the Dali hominid are unclear and the presence of river gravels, fish fossils, and the abraded condition of the Dali artifacts (Keates, personal communication) strongly suggest that the uranium dates can not be confidently associated with the hominid. Also, at least some of the mammalian elements and artifacts have been subject to fluvial transport. The few mammalian taxa ostensibly associated with the cranium present a confusing biostratigraphic picture. Middle Pleistocene elements such as Elephas sp., Megalocerospachyosteus, and Equus sanmenensis were recovered in association with Equus przewalskyi, a supposedly Late Pleistocene taxon, and the identification of the equids is very uncertain (Wu and You, 1979). Different excavation teams have also divided the Dali stratigraphy in different ways (Wu and You, 1980). However, the cranium, artifacts and dated mammalian fossils were reported in a published cross-section (Zhang and Zhou, 1984) as being in close proximity. The correlation of paleosols found in the sequence with the oxygen isotope stages identified in the central Chinese loess sequence suggests that the hominid bearing level may correlate with L2 of the loess sequence and oxygen isotope Stages 6-7 (see Qi, 1989). Qi (1989) also suggests that the Dali hominid may have occupied the area during a transition from a wet and warm period to a drier and colder period. None of the chronological data preclude assigning an age of ca. 0.3 mya to the hominid (Pope, 1988). Dali (Fig. 4A) has been only partially described in English (Wu and Wu, 1985; Pope, 1991) and Chinese (Wu X, 1989). Wu X (1989) considers the cranium to be that of a male, about 30 years old, with an estimated cranial capacity of approximately 1,120 ml. The sex determination is strongly supported by the great thickness of the cranial vault bones and the massively developed supraorbital tori. The Dali face can be accurately characterized as broad and short. Although otherwise well preserved, controversy has surrounded the preservation of the cranium’s midfacial area. Stringer (1990, 1991) contends that postdepositional vertical crushing and displacement is largely responsible for the impression of a short face. In fact the degree to which the maxilla has been superiorly displaced is minimal and there is no doubt that the original antemortem morphology of the maxilla was short and broad (Pope, 1991). The alveolar margin which is perfectly preserved in some areas supports this observation. It is also clear that the original height of the maxilla is comparable in its vertical dimensions to Sangiran 17 and dissimilar to Neanderthals and Archaic Homo sapiens (Pope, 1991; Wu X 1988a; contra Stringer, 1990). The maxilla also exhibits a superior lateral inflation of the bone surrounding the nasal aperture similar to that of Sangiran 17. The nasal aperture is very broad and bell-shaped. Laterally, there is an infraorbital depression which is separated from the orbital margin by a marginal torus. This depression further accentuates the appearance of paranasal inflation. The anterior surface of the depression intersects the lateral maxillary wall almost at a right angle and does not exhibit the oblique anterior zymomatic roots found in Neanderthals and Archaics. In this feature it is comparable to Jinniu Shan (see below). The zygomatic bones are + Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 271 oriented horizontally and a definite but shallow incisura malaris is present on the right side. Wu X (1988a) describes this feature as similar to the Xujiayao maxilla. It is also similar to Yunxian 11. No lateral alveolar prognathism can be observed on the preserved posterior right side of the maxilla. In midfacial dimensions and areas, the Dali specimen is most similar to Jinniu Shan, Steinheim, Arago, and Jebel Irhoud (Pope, 1991). However, the IZM contour is very different, because of the absence of pronounced lateral prognathism. Jebel Irhoud shares with Dali a n incisura malaris, a n infraorbital depression and a distinct, more medially situated malar tubercle. In other features, such as its tall middle face and high frontal, the Jebel Irhoud cranium differs noticeably from Dali (see below). Virtually no morphological evidence supports the interpretation t h a t Dali is substantially the same as the Archaics or Neanderthals. This is a n important point since Stringer (1989a, 1990, 1991) has repeatedly suggested that the short face of Dali is not a unique Asian characteristic but ar? artifact sf postdepcsitiofia! crushing. The orbits of Dali are clearly rectangular in overall shape, but have inferior lateral margins which are round. The interorbital distance is broad and shares the distinct internasal crest seen in other Chinese specimens. The supraorbital tori are massively developed, distinctively peaked and greatly thickened near the midline of the orbits. In frontal view, there is a supraglabellar depression joining the separate tori. This may be similar to the same area as it is preserved in Steinheim and Arago (Wu X, 1988a), but this is hard to ascertain since these latter two specimens are distorted. The supraorbital tori are much more massively developed in Dali t h a n in Jebel Irhoud, where they are much thinner and broadly arched rather than peaked. The frontal process of the Dali maxilla is much more anteriorly oriented and broader in comparison with Archaics, in which this feature is less developed. The frontal squama of Dali is short and has a distinct and very broad sagittal keel which widens posteriorly and then disappears anterior to bregma. Postorbital constriction is less than in the Zhoukoudian specimens, but greater than in Maba (Wu X, 1989, 1990, 1991). In lateral view, the Dali cranium is long and displays a distinctly flattened occipital contour which can be accurately described as a buniform protrusion. The nuchal plane is demarcated by a torus-like thickening. The occipital angle is obtuse and similar to other Chinese fossil hominids such a s Jinniu Shan and Yunxian. A definite Inca bone is also present in Dali. In basal view, a distinct mastoid fissure separates the mastoid from a triangular tympanic, but i t is much narrower than in Homo erectus. The Dali glenoid region is shallower than in the Zhoukoudian specimens. It is, however, still deep by modern standards. The tympanic bone is situated relatively medially as in Homo erectus. It is separated from the mastoid process by a distinct though narrow tympanomastoid fissure. This fissure is narrower than those seen in Homo erectus. The entoglenoid process is reduced and a blunt process supratubalis is present. A styloid process appears to be absent, but this area is not well preserved. In features of the glenoid region Dali is intermediate in its morphology between Homo erectus and anatomically modern Homo sapiens. It is for this reason, and its late Middle Pleistocene date, that paleoanthropologists continue to await a detailed description. Jinniu Shan The total morphological pattern of the Jinniu Shan cranium could not have been predicted on the bases of previously known finds in Asia. It is possible, though as yet unsubstantiated, that the specimen possesses the earliest known large cranial capacity and thinnest cranial bones of any of the early Homo sapiens. These anatomically modern characteristics were unexpected in China where only thickboned crania had been recovered previously. The Jinniu Shan hominid locality is one of several brecciated fissure fills located in a n isolated karst tower in Yingkou County, Liaoning Province. It is the most northerly of all the Premodern Chinese hominids. Only two lithic artifacts are 272 YEARBOOK OF PHYSICAL ANTHROPOLOGY Wol. 35, 1992 known from the site and these were recovered in the uppermost of the seven layers originally recognized a t the Locality Site A (Lu, 1985). Bone artifacts also have been recognized in the deposits (Lu, 19901, but their artifactual origin is unconvincing (personal observation). Recently, the stratigraphy of Locality Site A has been revised and the hominid fossils are now reported as deriving from layer 7, not layer 6 as originally reported (Lu, personal communication). The total deposits are about 14.5 m in thickness and dip at an angle of approximately 60-70". Huang and You's (1987) study and my own observations a t the site agree that the deposits represent a natural fissure fill. The faunal remains show few traces of modification by carnivores or other animals and many of the mammalian species are represented by very complete skeletons. There is no indication that Jinniu Shan was a hominid occupation site. The relative isolation of the hominid from the other mammals also supports this assessment (Pope, personal observation, see also Figtlre 3 in Anonymous, 1990). Additionally. the dense accumulations of well stratified microfauna suggest the remains of an autocthonous cave fauna. Taken as a whole, the evidence indicates that Locality A is a natural faunal trap. However, the excavator of the site, Lu Zun 'E (personal communication) has interpreted the hominid bearing level as indicative of a cave that perforated the side of the older breccia and served as a hominid occupation site. Because the dates are based on the faunal remains (Huang and You, 1987), an age of greater than 0.2 mya may be seriously in error and the actual age of the hominid may be much younger than that of the mammals. Twenty-six species of mammals were ostensibly recovered from above and a t the same depth as the hominid layer (ArcheologicalUnited Team, 1976).Homotherium ultima has been reported from layers 3 and 4 and Crocuta ultima from layers 5 and 7. These genera do not overlap a t Zhoukoudian Locality 1 or other Zhoukoudian localities. Other species include Felis chinensis,F. teilhardi, Ursus arctos, Microtus brandtioides, Trogontherium sp., and Megaloceros pachyosteus. All of these taxa suggest a correlation with layers 3-4 a t Zhoukoudian. These taxa are of undoubted later Middle Pleistocene aspect and are also consistent with the uranium dates of 0.24 mya (layer 4), 0.27 mya (between layers 5 and 6) and 0.28 mya for layer 7 (formerly layer 6) (Lu, 1985). Chen and Yuan (1988) have reported uranium ages for the sequence, based on mammalian teeth and bones, ranging from 0.164-0.314 mya. Lu (1990) reported a uranium-thorium age of 0.28 mya for layer 7 containing the hominid remains. There is every possibility that the dated mammalian fauna which is the basis of the radiometric dates has nothing to do with the age of the Jinniu Shan hominid. Despite dating problems, the Jinniu Shan hominid is one of the most important and interesting recent finds in China. The hominid was originally assigned to Homo erectus, but is now assigned to Homo sapiens. The specimen is represented by a cranium, vertebrae, arm and hand elements, foot elements from both sides, and a left innominate (Lu, 1990, 1992). All have been attributed to an adult male on the bases of a robust innominate, a narrow sciatic notch, the large triangularshaped obturator foramen, and relatively deep cotyloid fissure (Lu, 1985, 1990, 1992). In fact, all the post cranial elements are very robust. The Jinniu Shan cranial capacity originally was estimated a t ca. 1,390 ml (Lu, 1992) but this estimate was later revised to 1,330 ml (Lu, 1990). The new reconstruction suggests an even smaller capacity of less than 1,300 ml, but still within the error range of the original estimate (Lu, personal communication). A CT scan revealed faulty reconstruction of the occipital area that had given the impression of a occipital bun. However, even in the new reconstructioin, an occipital buniform structure is clearly present. The Jinniu Shan cranium has been most frequently compared to Dali (Wu X, 1988, 1989, 19901, but differs from Dali in its extremely thin cranial bones. Of the currently known Archaic Homo sapiens and Premodern Asians, only Steinheim may have cranial bones as thin as Jinniu Shan (Wolpoff, personal communication). In norma frontalis, the supraorbital tori are arched, but a continuous and rela- Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 273 tively shallow postorbital sulcus is present. The tori are thick, especially in the median portions, but not as thick or as arched as in Dali. The supraglabellar area is shallower than Dali. Large frontal sinuses are developed in the medial one-third to one-half of the frontal, but are asymmetrical in their size, shape, and lateral development, with the right sinus exhibiting a much more rounded shape and greater lateral extension (Lii, personal communication). The impression of postorbital constriction is perhaps slightly more pronounced than in Dali and is accentuated by an anterioposteriorly wide lateral trigon of the supraorbital torus that tapers to a more acute prominence a t the point where the anterior portion of the superior temporal line terminates. A weakly developed sagittal keel is present on the frontal; a slight prominence on one of the parietal fragments may also indicate the presence of a keel. Bregma is not preserved in Jinniu Shan, so it is impossible to tell if this area had a cruciform bregmatic eminence of the kind described by Weidenreich (1943 and Santa Luca (1980) for the Zhoukoudiari arid Ngandong specimens, respectively. The frontal process of the malar bone is oriented anteriorly, but not as anteriorly as in Dali. Judging from the right side of the specimen, there is a distinct notch a t the frontomalar suture, very similar to that seen in Dali and Sangiran 17. The original shape of the orbits is difficult to determine, but judging from the left orbit they are slightly more rounded than in Dali. The lower lateral corner of the orbit is rounded and, like Dali, situated below the inferior medial corner of the orbit. The inferiolateral rim of the orbit forms a torus which overhangs the anterior malar surface. This feature is also seen in Dali where it is more medially situated, forming the superior margin of an infraorbital depression. The face has a large nasal aperture and the interorbital area is wide. The preserved nasal bones indicate a higher nasal angle than Dali. A well defined sill divides the subnasal clivus from the floor of the nasal cavity and the remnants of a distinct nasal spine are present, suggesting that the nasal spine was very well developed in this individual. The superior paranasal region of Jinniu Shan has a distinct anteriorly facing swelling or inflation that is nearly identical to the condition seen in Dali and Sangiran 17. This morphology is also reported in high frequencies in Amerinds and is distinct from the condition usually found in American Caucasians (Gill and Gilbert, 1990). The subnasal clivus exhibits even less anterior alveolar prognathism than Dali or Sangiran 17. This is the result of postmortem anterioposterior compression. Like these two specimens, the Jinniu Shan maxilla is also short in comparison with Archaics. The IZM contour displays a shallow incisura malaris in association with the medially placed malar tubercle and, though noticeably horizontally oriented in comparison with Archaics, the zygomatic root intersects the lateral maxillary wall at a more obtuse angle (Pope, 1991) and is situated more superiorly than in Dali or Sangiran 17. Its greater height is comparable to Yunxian 11. In norma occipitalis, the greatest breadth of the cranium is on the parietals at the level of the parietotemporal suture. There is no discernible occipital torus. However, as mentioned previously, there is a rounded buniform swelling which is not associated with occipital flattening. In norma lateralis, the sagittal contour of the occipital squama is rounded and similar to the contour seen in Dali, but in Dali there is a distinct occipital flattening of the nuchal squama not seen in the more rounded contour of Jinniu Shan. The angle between the occipital and nuchal squama approximates that of Dali, but this is partly a function of the inferior contour of the buniform structure. Most of the parasagittal portions of the parietals are not preserved, and the original contour of this area cannot be reconstructed with any accuracy. The sagittal portion of the frontal is also poorly preserved. However, the fragments that are preserved may indicate a broadly inclined frontal squama situated behind a broad and shallow postorbital sulcus. The sagittal outline of the face is similar to that seen in Dali, with moderate anterior alveolar prognathism. In basal view, the large foramen magnum tapers posteriorly to form a tear-drop shape. The anterior margin of the foramen magnum is not preserved. The glenoid 274 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 fossa is preserved on one side and is broader and shallower in comparison with most specimens of Homo erectus. There is a definite articular eminence. The tympanic plate is separated from the mastoid process by a distinct tympanomastoid groove, except at its most medial portion where spongy bone has been deposited. The tympanic and petrous portions are not oriented parallel to each other, but the degree to which their relative orientations represent postdepositional deformation cannot be determined. The palate is deep and the specimen possesses a “ U shaped dental arcade. The teeth exhibit crenelated enamel surfaces and the incisors are shovel shaped with clear marginal ridges apparent even in their worn condition. The maxillary M3 is noticeably smaller than M2.The incisive foramen is vertically oriented and situated very anteriorly (Lii, 1990). In these features Jinniu Shan is most similar to other Asians. Perhaps the most perplexing aspect of the Jinniu Shan hominid is its complex mixture of characters which recall both fossil hominids and modern humans. Among the Premodern Asians, the thin and large cranial bones are most like Maba, the horizontally oriented IZM is more sloping than in Dali, but the broadly arching supraorbital tori are unlike those seen in Dali and Neanderthal-like morphology. In sum the total morphological pattern seen in the Jinniu Shan specimen fits neither the Replacement nor Regional Continuity Model in their simplest forms. Maba In comparison with all other Far Eastern finds, the Maba specimen recalls “classic” Neanderthals more than any of the other Premoderns. The Maba callote was found in a cave in Guandong, southern China, associated with faunal elements that are chronologically undiagnostic. Uranium dates from rhinoceros teeth range between approximately 0.129 -+ 0.011/0.01 and c0.169 2 0.018/0.016 mya for the strata that yielded the hominid and most of the other mammalian fossils (Yuan et al., 1986). The Stegodon-Ailuropoda fauna contains Crocuta ultima, but like the other faunal constituents, it is impossible to place this find with any certainty in either the terminal Middle Pleistocene or the early Late Pleistocene. Song and Zhang (1988) originally suggested that some of the associated fauna is indicative of a colder climate, and that the Maba hominid may date to a time equivalent to the penultimate glacial, but none of the associated fauna specifically indicate the presence of cold-adapted forms. The climatically diagnostic forms that can be identified indicate a tropical to subtropical climate. Following its discovery in 1958, Maba was often regarded as a Chinese Neanderthal (Wu and Peng, 1959; Wu X, 1988b1,primarily because of presumed age and its prominent supraorbital tori. The tori are similar in thickness to those of Neanderthals, with the thickest portion in the medial one-third of each torus. The orbits are distinctly rounded and lack a supraorbital notch, again similar to Neanderthals. There is a shallow and slightly developed supraorbital sulcus that resembles Dali more than Neanderthals (Wu X, 1989). However, it is deeper than in Dali, especially in its posterior portion a t the base of the frontal squama. In superior view, the tori are separated by a distinct glabellar notch. Also like Dali, the Maba specimen possesses large frontal sinuses separated by a thick septum (Wu and Wu, 1985). The supraorbital torus and sulcus complex is less inflated in its general morphology than other specimens discussed in this paper. In superior view, there also appears to be a more pronounced postorbital constriction than Hexian (Wu, 1990). Wu has pointed to the similar degrees of postorbital constriction in Zhoukoudian Locality 1 hominids and Dali as an example of mosaic and non-synchronous morphological change. Like other Chinese specimens, the Maba cranium also exhibits an anteriorly facing frontal sphenoidal process of the malar bone and a broad anteriorly oriented frontal process of the maxilla. The preserved right orbit is clearly round in shape and unlike any other known Homo erectus or Premodern. The inferior lateral Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 275 corner is not well preserved, but it appears less angular than other Chinese specimens. As with other Premoderns, the interorbital distance is great and a distinct ridge is formed at the sagittal sutures of the nasal bones. The nasofrontal suture is horizontally oriented. The frontal squama has a distinct sagittal keel, but it is not as developed as Dali and is comparable in development to that seen in Jinniu Shan. As a whole, the sagittal contour of the frontal displays a prominence which is not dissimilar to that seen in Zhoukoudian and other Chinese hominids. However, the sagittal profile of the nasal area does not resemble other Chinese fossil hominids in its marked curvature. Although biogeographically hard to defend and controversial, it is hard to avoid the observation that this specimen resembles Neanderthals from the West. Furthermore, the contrasts between Maba, Dali, Jinniu Shan, and Xujiayao constitute a roughly contemporaneous range of variation that is without precedent in either Africa cr Europe. It was exactly this realization, coupied with increasing evidence for a more tightly restricted terminal Pleistocene framework, that has occasioned the current re-thinking of simple Replacement Models. The situation is even more complex in light of the extremely early ages given to Liujiang, the oldest anatomically modern Homo sapiens from China. Liujiang The earliest well-preserved Chinese hominid exhibiting unequivocally modern morphology is the Liujiang cranium recovered from a cave in Guangxi Province. None of the few associated mammalian specimens, Rhinoceros sinensis, Megatapirus sp., and Sus sp. are chronologically undiagnostic, but Liujiang has been dated to perhaps as old as 0.101-0.227 and >0.067 mya by the uranium method (Yuan et al., 1986). These dates are also consistent with a 14Cdate of >0.040 mya (Yuan et al., 1986). However, it is difficult to accept the older dates in view of what we know about human evolution. On the basis of the antiquity claimed for other anatomically modern specimens from western Eurasia, a date of approximately 0.1 mya cannot be dismissed out of hand. Study of the Liujiang specimen reveals that it is completely modern in every aspect, although Wu (1959) considered it “more primitive” than the Upper Cave specimens. It has been regarded as a middle-aged male (Wu, 1959). Liujiang has the pronounced submaxillary prognathism of earlier specimens and the relatively horizontally oriented cheekbones that arise from a moderately high (in comparison with the Premoderns) zygomatic process of the maxilla. The orbits are rectangular in shape. The superciliary arches are thickest in their medial portions and there is an anteriorly projecting glabellar prominence. The interorbital distance is great and there is a distinct sagittal ridge where the nasal sutures meet. By Premodern Asian standards, the nasal aperture is small and there is a slight sill delimiting the subnasal clivus from the floor of the nasal aperture. The nasal spine is prominent, as is a distinct incisura malaris with medially placed tubercles. In lateral view, the squamous portion of the temporal bone is rounded and not rectangular as in Homo erectus. The occipital is noticeably inferiorly rounded in lateral view. The frontal squama is somewhat posteriorly inclined in comparison with many modern specimens and slight frontal bossing is observable. In basal view, the glenoid fossa is broad and a distinct articular eminence is observable. There is no mastoid fissure. The very anteriorly placed incisive foramen is not oriented vertically. The Liujiang cranium has many traits which can be traced back to Premoderns and Chinese Homo erectus. These include rectangular orbits, frontal bossing, a flat midface with a distinct incisura malaris, a relatively short maxilla, and no lateral alveolar prognathism. In the context of the Regional Continuity Model, Liujiang presents no morphological anomalies. However, the extremely early date is remarkable, leading most workers to doubt the radiometric values. For now this may be the most logical reaction, but there is currently no valid stratigraphic or methodological reason for disassociating the dates and the cranium. Taken a t face 276 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 value, the implications of the Liujiang specimen are that modern humans in China are as old as, or older than, anatomically modern humans from any other geographic region. DISCUSSION In China, the Regional Continuity and Replacement Models can be tested by distinguishing between relatively static or slowly changing traits and rapidly appearing traits which may indicate immigration to the area during the course of the Pleistocene. This focus on craniofacial trait complexes indicates that the morphology of some anatomical regions of the cranium are better indicators of regional clades than others (Howells, 1973,1989; Pope, 1991; Brace and Hunt, 1990). Howells (1989:83) notes that a “preoccupation with the shape of the braincase” has tended to obscure the importance of the face. In East Asians it is the midfacial region which displays the largest number of regional traits. Other workers have reached the same conclusion for racial differentiation of modern humans (Brues, 1990; Gill and Gilbert, 1990; Howells, 1989). In China, the upper and lower faces (as defined in this study) are not particularly useful in addressing the question of regional continuity. However, they do document a number of general trends that diverge in the direction of modern humans also seen in other geographical areas. The same is true of the glenoid and occipital regions. Both cladistic and phenetic studies rely on morphological traits whose independence from each other has only rarely been established. Additionally, it is important to establish the “morphocline polarity” (in cladistic terms) of traits or the “diachronic trends” (in chronophenetic terms) of morphological features and complexes. Both polarities and trends are much more convincing when an attempt has been made to ensure that lists of traits are not manifestations of the same underlying cause. In our ongoing attempts to understand the phylogenetic relationships between Homo erectus, early Homo sapiens, and modern humans, only rarely has the question of independence been considered. Therefore, an explicit discussion of the independence of the traits which define Homo erectus is extremely important. The Independence and Polarity of Traits A number of cladistic studies have suggested that Asian Homo erectus is defined only by a few autapomorphic traits which preclude it from the ancestry of modern humans (Stringer, 1984; Andrews, 1984a,b; Wood, 1984; Groves, 1989). Still other studies (Brauer, 1990; Habgood, 1989; Kennedy, 1991) have suggested that features such as those used by Andrews (198413) are in fact plesiomorphic traits which can be found in other hominoid and hominid taxa from Africa and Europe. Employing what might be termed a “semi-cladistic” approach, which emphasized characters derived in relation to early Homo, Rightmire (1984, 1990) concluded that Homo erectus is a definable taxon that may have given rise to later hominids. Brauer’s (1990) analysis emphasized the viewpoint that a taxonomic entity composed of plesiomorphic traits cannot be used to exclude it from the ancestry of anatomically modern Homo sapiens. At the same time he suggests that all the traits analyzed by Andrews (1984b) may simply be related to cranial thickness. The present study finds this is demonstrably not the case. Kennedy (1991) contends that Homo erectus, as presently defined by cladistic analyses, may not be a valid species, but that it is likely that the cladistic method is unworkable a t the generic level. Both Brauer’s and Kennedy’s analyses not only raise the distinct possibility that Homo erectus should be considered as a plesion, but also cast doubt on the utility of relying on a strictly cladistic analysis. The concept of a plesion (Patterson and Rosen, 1977; Schwartz et al., 1978) embodies a taxonomic entity that is defined solely as a collection of symplesiomorphic traits shared only with lower taxonomic sister groups. Such a situation removes the need for formal taxonomic units. In this case, the lower taxonomic groups would be those represented by Neanderthal and Archaics. This concept may also be particularly applicable to anatomically modern Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 277 humans since the only widely accepted autapomorphic “discrete trait” of extant Homo sapiens sapiens is a true chin. Even this trait, as exhibited by the samples from Klasies River Mouth, may not always be ascertainable (Wolpoff and Caspari, 1990; Caspari and Wolpoff, 1990). Drawing primarily on a list of autapomorphic traits distilled from Stringer (1984) and Wood (1984),Andrews’ (1984b)compiled a list of ostensibly autapomorphic traits of Homo erectus. Brauer (1990) found the same list to be composed entirely of plesiomorphic traits. These include: 1)a frontal keel; 2) a parietal keel; 3) thick cranial vault bones; 4)endinion well separated from inion; 5) a mastoid fissure (tympanomastoid fissure); 6) a medial recess between the entoglenoid and tympanic plate; 7) an angular torus (mounding of bone developed on the mastoid angle of the parietal). To this list of traits those of Kennedy (1991) can also be added: 8) a thick tympanic plate; 9) a bregmatic eminence; 10) a coronal keel: 11) an oeeipital torus; 12j increased craniai capacity. Rightmire (1990) lists characters which, though “not straight forward,” are diagnostic of Homo erectus and derived with respect to early Homo. These include traits 1, 6, 7, 11, and 12 listed above. He also includes: 13) “thickened continuous brow and associated flat supratoral shelf behind”; 14) wide and sharply angled occipital squama with distinctive transverse occipital torus (this morphology is listed as a single trait); 15) a “blunt process supratubarius [process supratubalis]” situated on a tympanic bone with a strong petrosal crest (Weidenreich, 1943). Rightmire points out that not all of the traits are expressed in all Homo erectus crania. All of the bilateral features (5-8 and 15) are highly variable in any one individual. The review of this list presented below shows that only some of these 15 features should be treated as definitely independent variables. Such a critical examination is absolutely essential to accomplishing the task of identifying the origin of modern humans in China or any other part of the Old world. The independence of traits 1 (frontal keel) and 2 (parietal keel) has not been firmly established. However, it is clear that frontal keels are not restricted to Homo erectus or East Asians. A frontal keel is present in KNM-ER 3733, Neanderthals, and Archaics, but salient parietal keels are apparently absent in all extra-Asian forms (Kennedy, 1991; Brauer, 1990). This may suggest that the two kinds of keels are truly independent traits. However, Brauer is correct in his contention that damage precludes the accurate assessment of parietal keeling in KNM-ER 3733. None of the other known African Homo erectus show definite parietal keels. Parietal keels are present in tandem with frontal keels in Far Eastern fossil hominids (Zhoukoudian, Sangiran, and Ngandong). Keeling is less developed and restricted to the frontal of Hexian, Dali, Maba, and Jinniu Shan. The presence and development of the frontal and parietal keels may possibly be independent because a frontal keel is found in the absence of a parietal keel. However, a parietal keel or eminence is never found in the absence of a frontal keel or some other form of bregmatic eminence or coronal keel (traits 9 and 10). This may possibly be interpreted in the Asian fossil record as the gradual disappearance of a single feature influencing both the frontal and parietal bones. Treated in a cladistic manner, the presence of both features in tandem is an autapomorphy of the Far Eastern clade only if KNM-ER 3733 lacks both keels. In specimens throughout the Old World there are also significant morphological differences in the formation of the frontal keel. The development of this feature is most prominent in extant Asians and East Asian derived populations (Rhine, 1990). It is premature to treat these features as definitely independent, but this may be the case. Thick cranial vault bones (trait 3) are one of the most commonly listed features of Homo erectus. This characteristic has also usually been used in cladistic studies as an autapomorphic feature which eliminates Homo erectus from the ancestry of modern humans. The phylogenetic significance and etiology of cranial bone thickness and its relationship to other traits has been discussed for decades, and recently reviewed a t length by Kennedy (1991), who, like other workers, reaffirms 278 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 that thickening of the outer table in the formation of the occipital torus and basicranium is a synapomorphic character of hominids. Kennedy contends that the thick tympanic plate (trait 8 of Homo erectus) is also a plesiomorphic character of the Hominoidea, since it is supposedly composed of spongy bone. However, this explanation is unconvincing, given the thinness of the Sangiran 4 and Sangiran 17 tympanic plates (the former, 2.0 mm; the latter 2.8; measured a t the same position specified by Kennedy’s data-the middle part of the lateral margin). Kennedy (1991) lists a mean 2.7 mm for modern humans. The Sangiran 4 and 17 cranial vaults exhibit massively thick vault bones but distinctively thin tympanic plates well within the range of moderns who do not show the “general symplesiomorphic thickness” of expansion of the “medial, spongy layer of bone” (Kennedy, 1991:389). The estimate of this dimension on the damaged KNM-ER 3733 measures 2.7 mm, the same as the modern average given by Kennedy. There is no doubt that all the Zhoukoudian and most Ngandong specimens exhibit thickened tympanic plates in association with thick vault bones, but this feature is a derived condition and not a plesiomorphic retention. It may be possible that in Chinese Homo erectus thick tympanic plates are linked with general cranial thickness, but this cannot be shown to be the case for all Homo erectus, especially in Java. Kennedy’s conclusion that, in general, thick cranial bones are not an autapomorphic character of Homo erectus because thick cranial bones are also present in Archaics and Neanderthals is valid. However, thick cranial bones are definitely derived with respect to early Homo and thus are a useful means of defining late members of Homo erectus and early members of Homo supiens. Kennedy also finds no correlation between bone thickness and ectocranial superstructures (traits 1, 2, 9, 10, 11).The thin walled KNM-ER 3733 and Trinil 2 specimens clearly demonstrate that supraorbital torus and sulcal development (trait 13) is not directly related to overall cranial thickness. Both the thin vaulted Jinniu Shan and the thick vaulted Dali specimens show substantial supraorbital development of the tori. The degree of postorbital constriction also seems to be unrelated to cranial thickness. Maba also shows much greater postorbital constriction than the much thicker Hexian specimen, which has an index value that falls within the Zhoukoudian range (Wu X, 1989).The prominence of the occipital torus (trait 11)is also apparently not directly related to cranial thickness as it is more pronounced in KNM-ER 3733 than in the much thicker Hexian cranium. Most other features also seem not to be influenced by cranial thickness. For example, only in Chinese Homo erectus is the association between thick cranial bones and small frontal sinuses convincingly correlated. The much thinner crania of Maba and Jinniu Shan are associated with large frontal sinuses. The Zhoukoudian (with the exception of Zhoukoudian 111) and Hexian crania exhibit small frontal sinuses. The frontal sinuses of all the Indonesian fossil hominids are large and in this group there is no correlation between thick cranial bones and small sinuses. The linkage between great cranial thickness and small frontal sinuses seems to vary between geographic groups and is strongest in Chinese Homo erectus. In Archaics, the trait is highly variable and there is no relationship between cranial vault thickness and frontal sinus development. Neanderthals consistently exhibit large sinuses. The large sinuses of Maba are extremely interesting given other features of Maba which recall Neanderthal morphology. The relationship between cranial thickness and separation of endinion and inion (trait 4) is demonstrably complex. Despite Kennedy’s contention (1991) that the separation of these two points is plesiomorphic for hominoids as a whole, her own data (1991) make it clear that within the Hominidae a high degree of separation is synapomorphic for most Homo erectus and Archaics in Asia and Africa. Marked separation (‘‘well separated,” Stringer, 1984) of 220 mm also characterizes most Asian adult Homo erectus. Furthermore, an inspection of the endinion-inion distances taken in conjunction with cranial bone thickness indicates no apparent correlation within the Hominidae of thinner cranial bones with shorter endinioninion height. A detailed comparison of this dimension with general occipital round- Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 279 ing would also be useful. An inspection of Kennedy’s data indicates, t h a t at least within different taxa of Homo, there is a correlation of endinion-inion distance with general rounding of the occipital. Among Chinese Homo erectus, Hexian (22 mm) Ngandong XI (21 mm) fall near the Neanderthals. Dali at 11 mm falls very close to the modern human average of 10.4 mm given by Kennedy. On the basis of current evidence, a well separated endinion-inion distance should be considered as a derived trait of Homo erectus that is shared with some Archaics and is not independent of occipital rounding. Trait 5 , the degree of development of the mastoid fissure (tympanomastoid fissure), is highly variable, but definitely not a n autapomorphic feature of Homo erectus, a s this trait is visible in many specimens of fossil Homo. It is also clearly present in both gracile and robust australopithecines and is indisputably symplesiomorphic for the hominoids as a whole. Premoderns diverge in the direction of the condition seen in moderns in their tendency to close the fissure. In this feature they are morphologically intermediate between Homo erectus and modern humans. Trait 6, a n entoglenoid or medial recess, may, like traits 5 and 8, be related to the overall morphology of the glenoid region, but not to overall cranial thickness. Both Weidenreich (1943) and Kennedy (1991) speculate that these traits may be related to anterioposterior compression or basicranial flexion, respectively. The feature is enormously variable ranging from a small fissure-like structure (Sangiran 4; Ngandong 4; left side) to a broad, but medially closed recess (Sangiran 17; right side) and to a medially patent fossa (KNM-ER 3733; left side, Sangiran 2; Sangiran 1; left side; Zhoukoudian I11 and V). It appears to show progressive closure through time toward the condition seen in the majority of modern humans. As Kennedy (1991) suggests, the patent condition is most probably the plesiomorphic condition for hominoids. Trait 7, a n angular torus, is most prominent in Zhoukoudian and in some Javanese individuals, but is variable within Homo erectus as a whole (Brauer, 1990). Within single individuals (cf. Zhoukoudian V, Sangiran 2, Sangiran 4, and most Ngandong hominids) the trait is bilaterally variable (personal observation). Its degree of development also varies with age (Santa Luca, 1980). The presence of this feature in Archaics is disputed (Kennedy, 1991). It is definitely present in Bodo. However, its morphology is also clearly different from that seen in the Zhoukoudian hominids, where i t is a more inferiorly restricted mound, confined to the posterior part of the cranium. In OH 9, it extends much more anteriorly along the contour of the superior temporal line. The feature is much less developed in Javanese hominids and not bilaterally represented in Sangiran 2 and Sangiran 4. Kennedy is clearly correct that i t is definitely not a n autapomorphic trait of Homo erectus. Trait 8, a thick tympanic plate, has been discussed above in relation to cranial thickness. It is highly unlikely that this is simply a plesiomorphic retention or that i t can be related to cranial flexion. In early Homo and Homo erectus, as judged from KNM-ER 1813 and KNM-ER 3733 respectively, the wall of the tympanic plate is thin, and in the former specimen bony exostoses are present. Trait 9, a bregmatic eminence, is most developed in Zhoukoudian and Javanese and is present in all of the Premodern Chinese where the region is sufficiently preserved. There has been considerable discussion about the exact form which i t takes and whether or not i t is related to frontal, coronal, and parietal keeling. Santa Luca (1980) reports that i t does not occur in Maba, but my examination of the original agrees with Wu (1990) that both a frontal keel and bregmatic eminence is present. It is apparently a synapomorphic feature of Asian Homo erectus (contra Kennedy, 1991), Premodern Asians, and some Archaics. Trait 10, a distinct coronal keel, is present in the Javanese forms, but not in the Zhoukoudian sample or many other specimens of Homo. Kennedy’s contention that this and other ectocranial features are not related to cranial hyperostoses seems correct. This structure’s development presents a problem similar to traits 1 and 2 280 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 in that it is difficult, if not impossible, to separate definitively a coronal ridge from a frontal keel and bregmatic eminence. All these features may be degrees of expression of the same trait. A distinct occipital torus, trait 11, is most pronounced in Asian Homo erectus. The plesiomorphic condition for Homo is a weakly developed torus as found in early African Homo. Within Asian Homo erectus, Hexian shows the weakest torus and the least angulated occipital bone, though it is similarly expressed in the Yunxian hominids. Its degree of expression seems definitively linked to the overall rounding of the back of the skull. All of the Ngandong hominids possess a distinctive form of this feature where it is more pronounced than in any other currently known Homo erectus. It is an autapomorphic character of Homo erectus and most strongly developed in the Asian clade. As Rightmire (1990) has noted, trait 12, cranial capacity, does show some increase over time and is definitely larger in Homo erectus than in early Homo However, the rate of increase can be made to vary markedly, depending on the dating assigned to the Ngandong individuals and the dating and inclusion or exclusion of KNM-ER 1813, Gongwangling, and Sangiran 17 (see Rightmire, 1990). Within Homo erectus, there is some increase in cranial capacity over time, though it is arguably not statistically significant (Rightmire, 1990). The future inclusion of Yunxian estimates will undoubtedly bolster cases for a demonstrable increase in cranial capacity over time, but their distorted condition will no doubt be a continuing point of debate. At Zhoukoudian there is a significant difference between Skull X and Skull I11 which exceeds the range of the other crania from layers 8 and 9, but there is no discernible trajectory of diachronic change. The independence of ectocranial superstructures, occipital and frontal angulation, and cranial capacity remains difficult to demonstrate. Traits 13, a thickened continuous brow and associated flat supratoral shelf, and 14, wide and sharply angled occipital squama with distinctive transverse occipital torus, probably should not be regarded as independent in cladistic studies. Like trait 4 (endinion well separated from inion), traits 13 and 14 are quite possibly related to the overall rounding, angulation, and size of the frontal and occipital bones. Both traits are almost certainly dependent on the overall rounding of the neurocranium which, at least in Homo erectus, is in turn dependent on brain size. However, smaller-brained hominids such as KNM-ER 1813 and OH 5 both exhibit very round neurocrania which suggests that rounder crania may result from more than one biological cause, and not just increased brain size. A variety of mechanical explanations of specific facial morphologies related to masticatory stress and muscular development have been advanced over the years (Russell, 1985; Rak, 1983; and Pope, 1991, for a review). However, recently it has been suggested that trait 13, a well developed and continuous supraorbital torus and supraorbital sulcus, is not involved in the resistance to masticatory stress, but instead reflects the positioning of the brain (Hylander et al., 1991). By this reasoning, the reduction of trait 13 would therefore be related to brain enlargement and repositioning, which in Homo is characterized by progressively larger frontal lobes. The Yunxian material may eventually throw more light on this matter. The comparatively reduced brows of Jinniu Shan and Maba may support the general contentions of Hylander et al. (1991). Trait 15, the process supratubalis, also seems to be a derived trait of most Homo erectus, if the trait is qualified as both “blunt” (Rightmire, 1990) and well developed. Like the other traits, this feature shows a wide range of variability. It is not always blunt in Homo erectus. It is a sharp crest or swelling in KNM-ER 3733 and WT 15000 and a bulbous, thickened swelling in the Zhoukoudian specimens. The Javanese hominids exhibit this feature, but its development in the Ngandong crania and Sangiran 4 and 17 hominids is sharper and more similar to the African specimens than to Chinese hominids. The feature is also present in OH 5 but again it is not sharp. In its expression within Chinese Homo erectus, the feature does not Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 281 seem to be independent of cranial thickness since thicker crania exhibit more developed and blunter processes. This discussion shows just how difficult it is to identify unambiguously independent autapomorphic features for Homo erectus. In fact, many of the claimed autapomorphies used to eliminate Homo erectus are synapomorphies shared with Archaics, Premoderns, and Neanderthals. Synapomorphies of Homo erectus, early Homo sapiens, and Neanderthals strongly suggest a phylogenetic continuity in these taxa (see Weidenreich, 1943, Kennedy, 1991). Strict cladistic studies have been unsuccessful in their attempts to demonstrate a large number of autapomorphic features which preclude Homo erectus from modern human ancestry, although some features do seem to be autapomorphic in terms of their extreme degree of development. The studies of Brauer (1990) and Kennedy (1991) ostensibly have demonstrated that no autapomorphic traits define the taxon. These studies illustrate the importance of considering temporal relationships, even when they are poorly known, in phylogenetic studies. Probably one of the most useful conclusions to emerge from all of these investigations, particularly Kennedy’s, is the likelihood that the number of biologically independent features has been greatly overestimated. This seems especially to be true for the glenoid and occipital regions. Conclusions presented here on the independence of characters (see Table 2), are of course to some degree tentative as they are based on a still limited fossil record. However, they can be considered working hypotheses that allow a new approach to the delineation of regional differences in complexes of interrelated characters. Whether or not these relationships are functional or idiosyncratic is beyond the scope of this paper. However, there is a growing body of evidence which suggests that very few craniofacial characters can be directly linked to specific stresses: “The relationship between human craniofacial morphology and the biomechanical efficiency of bite force generation in widely varying muscular and skeletal types is unknown” (Hanam and Wood, 1989:429;see also Pope, 1991; Hylander et al., 1991). With these provisos in mind the following observations can be made for the Chinese hominids. 1. There is no direct relationship between supraorbital prominence and supratoral sulcus development to cranial thickness as evidenced by the Maba and the Jinniu Shan crania. This has also been demonstrated for Neanderthals by Smith and Ranyard (1980). The Hexian and Dali crania both exhibit thick cranial bones in conjunction with a prominent supraorbital torus and tori, but they have relatively shallow supratoral sulci. In Africa, a similar independence is suggested by the prominent torus of the thin walled KNM-ER 3733 specimen. There may be a direct correlation between cranial thickness and small frontal sinuses in Chinese Homo erectus. With the exception of Zhoukoudian 111, all of the Zhoukoudian hominids and Hexian have small sinuses. This condition is derived with respect to early Homo, Javanese hominids, and Jinniu Shan. The degree of sinus development remains undescribed in the Dali and Yunxian specimens. 2. There is a definite and direct relationship between ectocranial structures of the occipital and the occipital angle. Weidenreich (19431, Tobias (19911, and Kniisel (1991) related these morphologies to the development of nuchal musculature. As the nuchal musculature reduces, the occipital angle becomes more obtuse and rounder, and the occipital torus and attendant sulci diminish in prominence and extent. Additionally, the endinion and inion distance decreases. Neither of these features should be treated as independent variables. Although an occipital torus can be found in many specimens outside of Asia, the most extreme forms are found in Asian Homo erectus. Within the Hominidae, a well separated endinion and inion is a synapomorphic character of a few Archaic Homo (Kabwe, Petralona, Bilzingslaben), Neanderthals (Gibraltar), and Homo erectus. This character is not independent of occipital rounding. 3. There is no relationship between the cross-sectional morphology (triangular or Unrelated to facial flatness No demonstrated relationship with other features Subnasal alveolar prognathism Lateral alveolar prognathism Orbital shape Independent of midfacial flatness, but not IZM height Related to facial flatness Independent of midfacial flatness Independent of midfacial flatness Related to midfacial flatness? Related to facial flatness Interdependent? Independent of vault thickness? Dependent on squama inclination? Independent of vault thickness In ZKD, dependent on vault thickness? Related to frontal lobe expansion Independent of vault thickness Gradually diminishes Sym. wi some early Homo Absence is aut. for Asians Present in WT 15000, S17, YX, I)L, and JNS Absent in all fossil Asians ZKD reconstruction is wrong, present in JI Flat in S4 and GWL Inferolateral rounding is variable in Asians Low in S17, S4, GWL, and DL: higher in ZKD and YX JNS inclined DL more horz.; Rapidly appearing more oblique orientation in some horz. in early Eur. AMH PMA Persistent and variable in Known first in ZKD and YX, Asians later in Jbl. Irhoud Gradual elevation in Asians Broad in Asians and some Arch. “Tented in Asians Syn. for fossil Asians and some Arch. and extant H. sapiens Variable in early Homo. Syn. Very rect. in fossil Asians, with some AMH, rect. in persistent in Asians fossil Asians Sym. for Homo Persistant in Asians Hoz. = Aut. for Asians Syn. wi Arch. H. erectus and PMA Low = Aut. for Asians Sym. for Homo? Sym. for Homo? Prominent, but thin in ER 3733, I hin but vert. thick in JNS, thick and tall in DL HX, NG, and DL slight Vertical in ZKD receding in NG Specimen notes Less pronounced in later More pronounced in DL and MB Homo erectus Slight enlg. from ZKD to HX ZKD I11 has large sinus, JNS, and YY. Rapid late enlg. in MB, and NG large, Variable in PMA Arch. Both present in ZKD Frontal HX, DL, and JNS frontal keels and ant. parietal only in PA only and Arch. Rapid-late thinning Syn. thickening wl Arch. Progressive broadening Syn. wi all AMH Sym. in ZKD? Syn. in NG, Arch., and Nean.? Variable Chronophenetic change Variable Cladistic polarity” Table 2. Morphological Traits of Chinese Fossil Hominids Phenetic independence - Incisura malaris IZM inclination Midfacial Wide interorbital breadth Cross-sectional shape of nasal root IZM height Frontal and parietal “keels” Frontal sinus Postorbital constriction Postorbital sulcus Traits Upper face Frontal squama inclination Supraorhital torus Sym. with Asians and some Nean. Syn. for Asians and Jbl. Irhoud Related to vault expansion Syn. large and late with and encephalization Archs. Related to vault expansion Syn. with Africian H . erectus and encephalization Related to facial flatness Related to small faces‘) Persistent in all fossil Asians Less apparent in Maba Persistent in all fossil Asians High frequency in all Asians ? erectus Sym. for hominids Variable Most developed in Asian H . erectus. esueciallv S4 The biological independence of each trait has been assessed as well as cladistic polaritry and chronophenetic observations of evolutionary change in Chinese Pleistocene hominids. “In comparison with early Homo. AMH, Anatomically modern humans; Arch, Archaics; Aut.. autdpomorphic; DL, Dali; GWL, Gongwangling; HX, Hexian; IZM, inferior zyg)matico-maxillary height; JNS, Jinniu Shan; MB, Maba: Nean, Neanderthal; NG, Ngandong; PMA, Premodern Asians; S, Sangiran; Sym., Symplesiomorphic; Syn., Synapomorphic; wi,with regard to; YY, Yiyuan; YX, Yunxian; ZKD, Zhoukoudian. Process supratubarius Gradual but not marked over GWL = 780 ml, NG = 1.200 ml time Angulation Gradually diminishes with Noticeably diminished in HEX Chinese H . erectus and later and PA Homo Occipital torus Related to vault expansion Syn. with African H . erectus Gradually diminishes within Bun-like morphology in JNS and encephalization H . erectus Endinion-inion vertically Ralated to rounding and Aut. in extreme form, syn. Gradually diminishes within Present in Petralona and Kabwe well separated encephalization with some Archaics H . erectus Angular torus Related m. temporalis Syn. with some Archs., highly Gradually diminishes within ‘Most developed in fossil Asians, development variable H. erectus less developed in OH9 Basalitemporal region fossa Depth of glenoid Partially developmental in Extreme depth aut. in H Divergence in direction of i3roader and shallower in S17, AMHerectus AMH YX. DL. and JNS General “inflation” of Not associated with vault Syn. with African H . erectus Gradual closure or variable Uniquely thin in S4 and thinner glenoid region thickness in KNM-ER 3733 Anterioposteriorly thick Independent of vault Syn. with Arch. Persistent ]Moderate thickness in DL and tympanic thickness in Java, JNS? possibly not in China Medial recess Related to basicranial Syn. with African H . erectus Variable Variable in AMH flexion? but highly variable Mastoid fissure Related to basicranial Syn. in regard to increased Narrows gradually? Absent in most AMH flexion? closure with African H. Frontosphenoidal process orientation Occipital and parietal Cranial capacity Shovel shaped incisors 284 YXARBOOK O F PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 rounded) of the tympanic and the presence or absence of a tympanomastoid fissure. A tympanomastoid fissure is definitely present in early African Homo. It is variable in Archaics and Neanderthals and present in all Homo erectus individuals and Premodern Asian individuals. Although it is found in Africa and Asia, it is much more developed and obvious in the early Asians (Sangiran 4 and 17). However, in later Asian Homo erectus, it is sometimes closed laterally. In Jinniu Shan and Dali one end of the fissure is filled by cancellous bone. Though the mastoid fissure may be considered as a separate character from tympanic morphology, it may not be independent of functional constraints in the evolutionary changes in the glenoid region as a whole. 4. The Asian fossil record demonstrates that lateral alveolar prognathism is independent from anterior alveolar prognathism. In the Archaics from Africa and Europe this does not seem to he the rasp as all of them exhibit anterior alceolar prognathism in conjunction with lateral alveolar prognathism. Future research may add to this necessarily tentative list of observations. An important objective should be the identification and delineation of morphological traits which are unambiguously independent. This is requisite to the generation of useful inventories of traits which can be used in both cladistic and phenetic studies designed to examine the competing hypotheses of Regional Continuity and Replacement. Chronophenetic observations The midfacial region (zygion-zygiodnasion-prosthion) A number of authors (Weidenreich, 1943,1939; Wu X, 1989,1990; Wolpoff, 1989; Wolpoff et al., 1984; Pope, 1991) have presented lists of phenetic traits which may be generally regarded as related to midfacial flatness. However, except as noted above, the independence of these traits has rarely been discussed. These traits include: 1)the anterior orientation of the frontosphenoidal process of the malar; 2) a wider interorbital distance (associated with an anteriorly facing frontal process of the maxilla); 3) more acute angulation of the junction of the lateral and anterior portions of the zygoma; 4) the perpendicular-everted orientation of the anterior facies of the cheek bone to the lateral wall of the maxilla; 5) a more horizontal orientation of the zygomatic process of the maxilla and; 6) the form of the IZM contour; 7) an incisura malaris associated with a more medially situated malar tiihercle; and 8) the intermediate to low height of the anterior root of the zygvrriatic process on the lateral maxillary wall. Features associated with midfacial prognathism include an anterior, obliquely inclined zygomatic root which arises directly from the alveolar margin is a condition which is of high frequency in Neanderthals and modern European and African populations. Rectangularly shaped orbits with inferior laterally round corners, a broad interorbital distance and a low nasal saddle are found in high frequency in fossil Asians. The absence of lateral alveolar prognathism is a trait which characterizes all of the known Asian fossil hominids (Pope, 1991). Wu Xs (1989,1990,1991) list of “common” Asian traits implied stasis in certain facial features. However, there is gradual modification in many of these features. The earliest known maxillae from the Far East (Sangiran 4, Sangiran 17, and Gongwangling) exhibit low anterior zygomatic roots and, judging from these fragmentary specimens, horizontally oriented zygoma (Pope, 1991). These contrast with the condition seen in the later Yunxian 11, Zhoukoudian, and Jinniu Shan hominids where the root of the process is situated in a more superior position. The plesiomorphic Asian condition differs from early African Homo in which the process is higher and no incisura malaris is present. The degree of development of the incisura malaris varies from deep in Zhoukoudian and Yunxian to moderately shallow in Dali, Jinniu Shan, and Liujiang. The Jinniu Shan hominid exhibits a modified combination of chcck features with a straighter, slightly more oblique Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 285 cheek bone, comparatively low zygomatic maxillary process, and a slight incisura malaris. In these features it is most similar to Dali. The nasal region (nasal root and anteriomedial wall of the orbit) has received considerable attention in racial identification (Brues, 1990;Gill and Gilbert, 1990). Brues has divided nasal root morphology into “Quonset Hut,” “Tented,” and “Steepled categories for Afroamericans, Asian-Amerinds, and Caucasians respectively. Gill and Gilbert (1990) observe the Afroamericans and Amerinds share a broader frontal process of the maxilla that contrasts with the condition seen in Caucasians. Wu X (1989, 1990) has further added that a flat or broadly curving nasofrontal suture is common to East Asians. The shape categories of Brues (1990) listed above supposedly differentiate Asian and African fossil hominids. In contrast, the nasal morphology used by Gill and Gilbert (1990) does not consistently and reliably separate fossil Asians and Africans, since Jebel Irhoud (see below) conforms to the Asian pattern in this and other features which are related to facial flatness. An anteriorly oriented frontosphenoidal process of the zygomatic relates the Asians and perhaps WT 15000, which exhibits the anteriorly oriented condition and the flexed “African” condition. The superior paranasal area is noticeably inflated in Dali, Jinniu Shan, and Sangiran 17. This condition is most prominently developed in Asians and differs from the condition seen in Neanderthals, where the superior paranasal margins extend more anteriorly and parallel to the sagittal plane. Subnasal prognathism is emphatically not a unique East Asian clade characteristic. However, Brooks et al. (1990) contend that different sagittal contours of the submaxillary region can be used to differentiate Afroamericans, Amerinds, Caucasians, and South Asians. Afroamericans supposedly present a much flatter or concave contour in comparison with East and Southeast Asians who display a rounder and more inflated inferior contour. KNM-ER 3733 and KNM-ER 1470 are consistent with this observation. KNM-ER 1813, Sangiran 17, and all the Chinese fossil hominids exhibit the Amerind pattern. Sangiran 4 may exhibit the African pattern. It is the lack of lateral alveolar prognathism, independent of subnasal alveolar prognathism, that is characteristic of the Asian clade. The maxillary dentitions of Asians also show a high frequency of certain traits. The best known of these is shovel shaping. Shovel shaped medial and lateral maxillary incisors occur in every Chinese fossil where this character can be assessed. However, there is no doubt that shoveling is plesiomorphic for Homo as it occurs in early African Homo erectus (WT 15000) and in Neanderthals (e.g., Krapina), though the specific morphology of shoveling varies greatly (Mizoguchi, 1985). In the Chinese fossil record, a basal tubercle is characteristic of these teeth, but the prominence of the tubercle diminishes during the course of the Pleistocene. Among modern populations, shovel shaping has by far highest frequency in East Asian and East Asian-derived populations (Turner, 1989, 1990; Cadien, 1972). The lower face (the mandible) and dentition With the exception of M, agenesis, these are few definite mandibular features specifically indicative of an Asian clade, though this can certainly not be said for dental features (Turner, 1990, 1989). Zhang’s study (1991) taken with the dental evidence from Yunxian suggests that the reduction of M, and M, began in Homo erectus times. Angel and Kelly (1990) have proposed the degree of posterior ramus inversion and gonial flare as a means of discriminating Africans from Amerinds and Caucasians. It is hoped future fossil finds in China will allow us to employ this and other traits (Rhine, 1990) as a means of testing the hypotheses of human origins in China. In regions other than the midface, the same temporal trends seen in other parts of the world are also seen in Chinese hominids. Weidenreich (1943) reported that one specimen from Zhoukoudian may perhaps show an incipient mental trigon. However, one of the most perplexing problems in Chinese paleoanthropology is that no mandibles of Premodern Asians have yet been recovered. 286 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 The upper face (nasion-vertex) In China, the upper face has evolved in a manner consistent with trends elsewhere in the Old World. The frontal squama become gradually more vertical, higher, and transversely expanded. There is a continuance throughout the Pleistocene in China of a bregmatic eminence which arguably distinguishes the Chinese and Javanese hominids (Weidenreich, 1943; Santa Luca, 1980). There is no evidence that the thickness of the frontal squama decreases gradually over time. Similarly, there is no evidence for a gradual enlargement of frontal sinuses. In fact, judging from the Jinniu Shan and Maba crania, thin cranial bones and possibly large sinuses appear suddenly. However, the thicknesses of the Xujiayao crania are variable and fall just above the modern average. The Xujiayao specimens are temporally restricted to a period dating to about 0.1 mya and thus provide a picture of contemporaneous variation at t h e beginning of the Late Pleiatocene. As noted previously, there is minimal enlargement of the frontal sinuses in Chinese Homo erectus, though they are well developed in the Ngandong hominids. Future discoveries may show that there is a more gradual transition to larger sinuses, but on the basis of the current fossil record, it is quite possible that thin frontal bones (and vault bones in general) may have been rapidly introduced in China. The same may be said of reduced supraorbital torus thickness. No obvious diachronic trend is evident in the Zhoukoudian sequence. The Hexian specimen also exhibits a massive torus, though the supraorbital sulcus is shallower and less defined. The Dali specimen also shows no reduction in torus development. The first indications of brow ridge reduction occur in Maba and Junniu Shan. These are the earliest forms that can be shown to be intermediate in thickness between modern East Asians and Homo erectus. This conclusion may change with the recovery of new material, but for now there is no evidence for the gradual reduction in brow size. The occipital region The same general statement about the lack of specifically regional features in the frontal region also applies to the occipital region except for the high frequency of Inca bones in Chinese fossils. The same seems to be true for the Dali, Xujiayao, and Dingcun samples (Wu, 1991). There is a gradual decrease in the angulation of the occipital bone, diminution of the occipital torus, and a decrease of endinioninion separation, associated with both the rounding of the occipital and relative enlargement of the cerebellar fossa. If the preliminary Yunxian date of ca. 0.4 mya is correct, and if the occipital contour is not largely a product of postdepositional processes, moderately buniform occipitals (not associated with lamdoidal flattening) may have a considerable time depth in China. This feature is also present in Jinniu Shan, although it is much less pronounced than the original reconstruction indicated. Occipital flattening of the kind seen in Neanderthals does not appear to be part of the Chinese morphological pattern. A buniform rounding of the occipital is also present in the anatomically modern Liujiang specimen. In China, the pattern of progressive rounding is very similar to that seen in other parts of the Pleistocene Old World. However, the Ngandong hominids exhibit a distinctly “hooked inion” which has been considered a Caucasoid characteristic among modern racial groups (Rhine, 1990). The glenoid region It is possible to make a case for gradual evolutionary change in the glenoid region, involving the closure of the mastoid fissure, the reduction of the entoglenoid recess, and the progressive shallowing of the glenoid fossa. However, it is important to emphasize that these traits may not be independent and are bilaterally variable in both fossil and modern crania. There is no doubt that the Chinese fossil specimens are derived in these characters from early African Homo. How- Popel CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 287 ever, the immature WT 15000 exhibits a much shallower and wider glenoid fossa in comparison with most other Homo erectus, which, judging from modern humans, could be expected to already show the deeper adult condition. The Sangiran 17 specimen is also extremely interesting in that it exhibits a massive medial recess and a broad, shallow glenoid fossa which is not only similar to WT 15000, but also similar to the condition seen in KNM-ER 1813 and KNM-ER 3733. Since Sangiran 17 is among the earliest Javanese hominids, such morphological afthities are biogeographically and temporally consistent, as the specimen is geographically and chronologically closer to early African Homo habilis and Homo erectus. However, the effect of developmental pressures, especially mastocatory stresses (or lack thereof) have not been investigated in sufficient detail. If the depth and development of the glenoid region can be shown to be a s sensitive to environmental stress in fossil hominids as it is in modern humans (Hinton, 1979,1981; Aiello and Dean, 1990) this avenue of investigation may make It possible to address the question uf biobehavioral differences in fossil hominids. The Jebel Irhoud-Cro-Magnon and Junniu Shan-Maba “Problem” One of the fundamental problems at the center of the Replacement versus Regional Continuity debate is the necessity of explaining how and why traits that were once regionally restricted became widespread near the end of the Pleistocene. A single origin interpretation, such as the Replacement Model, presents a n explanation for this. Why some of the earliest anatomically modern humans and some of the latest Archaics exhibit very similar features to other geographically widely dispersed contemporaneous hominids also demands a n explanation. African Archaics and early anatomically modern humans from Europe highlight the importance of addressing this question. Jebel Irhoud and Cro-Magnon are prime examples of non-Asian specimens which display traits that had been previously confined to the Asian clade (Pope, 1991). These traits have been unconvincingly attributed solely to general facial flatness (Stringer, 1989a,b). Many of the Asian zygomatic features are present in Jebel Irhoud (and some later North Africans) and the cheek bones of Cro-Magnon are oriented horizontally and are associated with a vertically short maxilla and remarkably rectangular orbits. However, the Jebel Irhoud 1 cranium displays marked lateral alveolar prognathism, as do all the other African and European Archaics. Stringer (1989a,b) has proposed that these similarities suggest that socalled Asian characteristics were never confined to Asia, or that parallel evolution accounts for these geographically separated features. Simmons and Smith (1991) argue that fossil North Africans actually form a part of a continuum of interconnected populations distributed across Africa and Eurasia. Thus, there is no need to invoke geographically disparate parallelisms in highly contrasting environments. Jinniu Shan and Maba also display features that are not found in other fossil Chinese, notably large frontal sinuses and relatively thin cranial bones, and in the case of Maba, very round orbits. Both of these terminal Middle Pleistocene specimens exhibit the first Asian instances of thin cranial bones and round orbits. While the Jinniu Shan cranium exhibits some features similar to Dali (in the supraorbital region and in cranial shape), Maba in its entirety exhibits features that are totally consistent with Neanderthal morphology. It could well have been in “Maba” times that individuals with rounder orbits entered the Chinese clade. If Maba had been recovered in Europe, it would have been classified as a Neanderthal. Although this seems biogeographically illogical, on chronophenetic grounds i t is a n accurate observation. Framed in terms of the Replacement versus Regional Continuity debate, Jinniu Shan and Maba may represent immigrating populations from outside of China. Alternatively, Jinniu Shan and Maba may be the result of admixture with Archaic Homo sapiens or Neanderthals. Similarly, Jebel Irhoud and Cro-Magnon populations could be interpreted as representing Premodern Asian and anatomically modern human populations from the Far East. There is no convincing reason for 288 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 35, 1992 assuming that Europe and Africa were the original sources of these traits, especially as these morphological features are known far earlier in the Far East. Movement out of the Far East is just as likely. Proponents of the Replacement Model must postulate exactly this kind of long range movement of populations. However, these fossils (Jebel Irhoud and Cro-Magnon)do not support the movement of genes and/or populations in only one direction (“Out of Africa”). As Smith et al. (1989) note, in this type of model, new genes would appear in old gene pools and old genes would appear in new gene pools. The most logical way of employing morphological evidence in arguments about the origin and direction of dispersal of anatomical traits is to establish the temporal and geographical location of their first appearance. This is a fundamental paradigm of biostratigraphy which cannot be invalidated on the basis of contradictory biochemical interpretations. In Europe, Africa, and China, the first appearance o f novel traits is preceded by the ear!ier establishment of these traits ir, other regions. Normal populational variation must also be taken in account when assessing traits, but one feature which both the Replacement Model and the Regional Continuity Model share in common is the necessity for long-range gene flow. In the former model, the replacing genes must arrive in a distant geographic area or else there could be no replacement and in the latter, long distance gene flow must be maintained in order to preclude cladogenic speciation. Furthermore, assimilation is the most plausible and parsimonious explanation of the persistence of some regional traits in specimens which also display suddenly appearing novel traits. Specifically, traits such as rounder orbits (Maba), less horizontally oriented cheek bones (Jinniu Shan), thin cranial bones associated with larger frontal sinuses (Jinniu Shan), and perhaps a markedly increased cranial capacity were western Eurasian “imports” which appeared in “Jiniu Shan-Maba Times” in both northern and southern China. On the other hand, if the Jinniu Shan specimen isjust under 0.3 mya in age, larger neurocrania may have appeared first in China. The tentative age and large estimated cranial capacity of Yunxian I1 would also support this conclusion. A similar kind of extra-regional origin of midfacial morphology (midfacial flatness and a distinct incisura malaris) is found in the fact of Jebel Irhoud. These particular features are known earlier in the Far East. The Asiatic-appearing faces (short maxillae, inferiorly situated and horizontally oriented cheekbones, IZM contour, incisurae malaris, more medially situated malar tubercles, and markedly rectangular orbits) of fossils like Cro-Magnon, Predmosti, and Maldec may possibly represent even later introductions from Asia (Wu X, 1988a, 1990; Pope, 1991). Although this is a highly contentious suggestion, it is no different from the picture of Asian migrations developed in historical times. As mentioned previously, it is exactly this kind of model which Smith et al. (1989) have posited to accommodate the European and African fossil record. A Late Pleistocene Assimilation Model is just as parsimonious as a Replacement Model and also has the advantage of being highly consistent with the Far Eastern archeological record which shows no evidence for the advent of novel technologies until the late Late Pleistocene (Pope, 1988a,b). Demonstrably early morphologies in a given region continue to persist (though in many cases with modification) and do not disappear after the introduction of new morphologies. With the exception of the large faces of the Yunxian crania (Fig. 31, the small faces of Premodern Asians resemble extant East Asians. The genes responsible for the large Yunxian faces did not persist in subsequent Asian populations. On the other hand, the large faces of Archaics and Neanderthals did persist and resemble the faces of modern Europeans much more than the small Cro-Magnon faces do (Pope, 1991). Similarly, the large Jebel Irhoud face resembles extant North Africans in both size and the presence of lateral alveolar prognathism. Interestingly, in comparison with Asia and Europe it is in Africa that introduced extra-regional features seem to persist the longest. Examples include the persistence of facial flatness and an incisura malaris in anatomically modern Pope1 CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 289 North Africans such as in Wadi Halfa. Although the Assimilation Model does not have the advantage of the compact simplicity of a single origin model, neither does it labor under the constraints of geologically rapid divergences which must be invoked to explain the polytypism of modern humans. SUMMARY The chronophenetic discrimination of long persisting and gradually evolving regional midfacial traits from rapidly and late appearing traits makes i t possible to support a n Assimilation Model, rather than a Replacement Model, for the origin of modern humans in Asia. The increasingly firm chronology for Chinese Pleistocene hominids strongly suggests accelerated interregional gene flow beginning in the terminal Middle and/or Late Pleistocene in both western and eastern directions. The hominid fossil record in China is rapidly becoming a s precisely dated as in Europe and the Middle East. Although the chronology of the Chinese hominids is not uncritically accepted by most Western workers, i t is based on the very same techniques in use in the West and has produced isotopically, climatically, and biostratigraphically consistent results. The new Chronometric data show that there is a n enormous amount of roughly contemporaneous morphological diversity among Pleistocene hominids in China. However, unlike the West, there is very little to suggest the overlap of anatomically modern humans with Premodern forms. However, new discoveries and future refinement of the chronological framework eventually may show this to be the case. One possible exception is the very old date obtained for Liujiang, which would, if correct, document the synchronic existence of modern and Premodern forms. There has, however, been some suggestion t h a t Premoderns are synchronic with Homo erectus (Chen and Zhang, 1991). This suggestion is based on the possible chronological overlap between the upper deposits at Zhoukoudian Locality 1,and new discoveries from Hexian, Chaoxian, Dali, Jinniu Shan, and Xujiayao. On the basis of current evidence no such overlap can yet be confirmed. A chronophenetic approach (Table 2) reveals that both Chinese Homo erectus and Premodern Homo sapiens display a mosaic of craniofacial characters. Some of these characters reflect a continuous long-term presence in China while others seem to represent rapid and late introductions. Most regional clade features are found in the midface. Slowly evolving persistent features include the superior movement of the zygomatic process of the maxilla, more horizontal orientation of the IZM, a n incisurae malaris, facial flatness (anteriorly oriented frontosphenoidal process, anteriorly oriented frontal process of the maxilla, and more vertical malar facies), and a n absence of lateral alveolar prognathism. Other traits, not restricted to the midfacial area, include the persistence of thick cranial bones in association with small frontal sinuses. With the exception of Zhoukoudian 111, Maba, and Jinniu Shan, Chinese fossil hominids show a consistent correlation between thick cranial bones and small frontal sinuses. Such a correlation is not found in either the Javanese hominids or the Archaics (where thick crania may or may not have large frontal sinuses). One of the important findings of this study is that there is no consistent divergence of these features in the direction of modern humans and that cranial thickness is independent of a number of anatomical features in Homo erectus. Thin cranial bones, large frontal sinuses, large cranial capacities, but not the reduction of supraorbital tori, seem to make rapid and geologically late appearances in China. However, the Xujiayao evidence is equivocal and may be used to argue for either the in situ development of at least a few modern features, or for the rapid introduction of some features during a short chronological interval. If both the large preliminary cranial capacity estimate for Yunxian and its early estimated age (0.4 mya) are correct, the specimen would document considerable cranial expansion at a n unprecedented early age in China. However, both the chronological and cranial capacity estimates are preliminary. 290 YEARBOOK OF PHYSICAL ANTHROPOLOGY IVOl. 35, 1992 Other portions of the cranium such as the occipital, parietal, frontal and glenoid regions show a gradual modification in the direction of anatomically modern Homo sapiens which is not regionally restricted. The occipital becomes rounder and the ectocranial superstructures diminish (including the angular torus of the parietal), the frontal gradually becomes broader and more vertical with less postorbital constriction, and frontal and sagittal keels gradually reduce over time. All of these trends have also been found in regional fossil records of Africa and Europe. The glenoid region remains deep in most Homo erectus compared with modern humans. In this regard, the dating of the Yunxian crania is of crucial importance as the glenoid fossa is apparently somewhat shallower than in other Homo erectus homologues. Poorly understood features such as the medial recess, the mastoid fissure, thickness of the tympanic plate, and the process supratubalis need to be investigated in detail with regard t o their occurrence in modern regional populations. Cladistically, most of these features appear to be derived in Homo erectus with respect to early Homo and A ustralopithecus. However, the Premodern Asians show an intermediate morphology between Homo erectus and modern humans which indicates that they are suitable transitional forms which were ancestral t o modern humans in China. This contradicts a major prediction of the Replacement Model, which precludes transitional forms outside of Africa. Cladistic claims that Homo erectus cannot be defined morphologically by autapomorphic characters, or that the autapomorphic characters which do define the taxon preclude its place in human ancestry, are not sustained by this chronophenetic study. The contradictory results of cladistic studies demonstrate the difficulty of using this approach to resolve phylogenetic relationships within the genus Homo. Many of the supposed autapomorphic traits of Homo erectus now have been clearly shown to be plesiomorphies or manifestations of the use of traits whose biological independence has not been established. However, the extreme endinioninion distance, a prominent occipital torus, a well developed continuous supraorbital torus, a deep glenoid fossa, and a blunt process supratubalis appear to be autapomorphic features which do define the taxon. There is strong evidence that the majority of morphological traits exhibited by Chinese Homo erectus and Homo erectus are in general plesiomorphic for the Hominidae, and in some cases the Hominoidea as a whole. However, many of the traits identified as non-autapomorphic are shared with later or contemporaneous groups and argue strongly for a phylogenetic connection between Homo erectus, Archaics, and Neanderthals. The failure to produce an extensive, valid list of independent morphological traits is a failure of the cladistic method itself (Kennedy, 1991) and not an indication that the taxon Homo erectus has no reality. The absence of autapomorphic traits does not preclude ancestor-descendent relationships. If taxa must be defined soley on the basis of autapomorphic traits, then the strict cladistic method becomes an inappropriate means of determining phylogenetic relationships. Previous cladistic studies, with the exception of Kennedy (19911, have failed to address adequately the question of the independence of morphological traits. Specifically regional clusters of traits associated with Chinese Homo erectus and its descendants are exhibited in the midfacial region (Table 2). These traits appear first in East Asia and later spread to other regions. The possibility that these traits are not all just a reflection of general facial flatness is indicated by the fact that they become “decoupled” in modern humans with traits appearing in various combinations in various parts of the world (see Pope, 1991). With the exception of the midfacial traits, the majority of Homo erectus characters indicate that we are dealing with a species whose total morphological pattern is distinctive enough that it can be recognized in widely dispersed geographic localities which span more than one million years. Whether or not the taxon Homo erectus can be defined cladistically or phenetically, the fact remains that members of the taxon can be consensually recognized in most instances. As Wolpoff (1989:83) notes, characterizations of “regional plesiomorphy cannot possibly apply in attempting to discount what any observer can see”. Difficulties in recognition are Popel CRANIOFACIAL EVIDENCE FOR MODERN HUMAN ORIGINS 291 greatest with some of the earliest and latest members of fossil homo (i.e. KNM-ER 1813 and Ngandong). It is exactly this situation which should be expected from temporally continuous lineages and regional geographic clades. Finally, non-anatomical evidence, in the form of mammalian biogeography and archeology, also fails to support either replacement or complete isolation a s a model for the emergence of modern East Asians. Many mammalian taxa are shared throughout the Pleistocene between northern China and Europe (Pope, 1982). The archeological record, although distinctly different from Europe and Africa, shows a temporal continuity which mirrors the anatomical evidence (Pope, 1982, 1983, 1988, 198913; Huang e t al., 1987; Li, 1989; Pope and Keates, 1982; Zhang, 1990), and argues for continuity. The archeological evidence from China remains as a n especially large impediment to the acceptance of complete replacement of indigenous populations by invaders possessing a superior technology (contra Klein; 1992). There is n o evidence for the introduction of major technological innovations. In sum, the Chinese archeological evidence is in complete agreement with the fossil evidence. ACKNOWLEDGMENTS I express my sincere gratitude to my Chinese colleagues, Wei Qi, Huang Wanpo, Wu Rukang, Wu Xinzhi, Zhang Zhenbiao, Zhang Yinyun, Yuan Dongya, and Zhou Guoxing for access to specimens and valuable discussions. Very special thanks are due to Liu Dengsheng. S.G. Keates, J. Ostlund, and D. Bakken provided valuable suggestions and exhaustive editorial assistance. I also benefited from the comments and suggestions of the Yearbook of Physical Anthropology Reviewers and the Editor. The drawings are by Kai Nakbunlung. This research was supported by the Research Board of the University of Illinois, which provided pivotal financial support. LITERATURE CITED Agrawal DP (1991) Interdisciplinary paleoclimatic research in Asia. In Special Proceedings Review Reports XI11 Inter. Cong. Beijing: INQUA, pp. 76-77. Aiello LC and Dean MC (1990) Introduction to human evolutionary anatomy. London: Academic Press. Aigner JS (1981)Archaeological Remains in Pleistocene China. Miinchen: C.H. Beck. Aigner JS (1986)The age of Zhoukoudian Locality 1: The newly proposed 0" correspondences. In VV Novotnv and A Miserova (4s.):Fossil Man. Kew Facts, New Ideas. Papers in Honor of J a n Jelinek's Life Anniversary. Anthropos (Brno)23: 157-173. Aigner JS (1988) Dating the earliest Chinese Pleistocene localities: The newly proposed 0" correspondences. In P Whyte, JS Aigner, NG Jablonski, G Taylor, D Walker, and P Wang (eds.): The Palaeoenvironment of East Asia From the Mid-Tertiary, Vol. 11. Hong Kong: Centre of Asian Studies, pp. 1032-1061. An 2 and Lu Y (1984) A climatostratigraphic subdivision of late Pleistocene strata named by Malan Formation in North China. Kexue Tongbao 29t1240-1242. An 2, Liu T, Kan X, Sun J , Wang P, Gao W, Zhu Y, and Wei M (1987) Loess-Paleosol sequence and chronology a t Lantian Man localities. Aspects of Loess Research. Beijing: China Ocean Press. An Z, Wu X, Wang P, Wang S, Sun X, and Lu Y (1991) An evolution model for paleomonsoon of China during the last 130,000 years. In T Liu (ed.): Quaternary Geology and Environment in China. Beijing: Science Press, pp. 237-244. Andrews PA (1984a) An alternative interpreta- tion of the characters used to define Homo erectus. In P Andrews and J L Franzen (eds.): The Early Evolution of Man, With Special Emphasis on Southeast Asia and Africa. Cour. Forsch. Senckenberg 69t167-175. Andrews PA (1984b) On the characters that define Homo erectus. In P Andrews and J L Franzen (eds.): The Early Evolution of Man With Special Emphasis on Southeast Asia and Africa. Cour. Forsch. Senckenberg 69:176-179. Angel J L and Kelley J O (1990) Inversion of the posterior edgt: uf Ute jaw ramus: A new race trait. In GW Gill and S Rhine (eds.): Skeletal Attribution of Race: Methods for Forensic Anthropology, Maxwell Mus. Anthropol., Anthropol. Pap. No. 4, pp. 33-39. Anonymous (1990) Jinniu Shan. In Wei F (ed.): The Valuable Cultural Relics and Historic Sites of Liaoning Province. The Liaoning Artistic Publ. House, pp. 6-11. Archaeological United Team (1976) New material of Pleistocene mammalian fossils in Jinniushan from Yingkou, Liaoning Province and their significance. Vert. PalAs. 14t120-130. Arensburg B, Bar-Yosef 0, Chech M, Goldberg P, Laville H, Meignen L, Rak Y, Tchernov E, Tillier A-M, and Vandermeersch B (1985) Une sepulture neandertalieene dans la grotte de Kebara (Israel). C.R. Acad. Sci. Paris 300 Serie I1 6,227230. Bilsborough A and Wood BA (1986) The nature, origin, and fate of Homo erectus. In B Wood, L Martin, and P Andrews (eds.): Major Trends in Primate and Human Evolution. Cambridge: Cambridge University Press, pp. 295-316. Binford LR and Ho CK (1985) Taphonomy at a 292 YEARBOOK OF PHYSI(3AL ANTHROPOLOGY distance: Zhoukoudian, “the cave home of Beijing man”? Curr Anthropol26r413-442. Binford LR and Stone NM (1986) Zhoukoudian: A closer look. Curr Anthropol27:453-475. Brace CL and Hunt KD (1990) A non-racial craniofacial perspective on human variation: Afustralia) to Z(uni). Am J Phys Anthropol 83:341360. Brace CL, Shao Z, and Zhang Z (1984) Prehistoric and modern tooth size in China. In FH Smith and F Spencer (eds.):The origin of modern humans: A world survey of the fossil evidence. New York: Alan R. Liss, pp. 485-516. Brauer G (1984a) The “Afro-European supiens hypothesis” and hominid evolution in East Asia during t,he later middle and upper Pleistocene, In P Andrews and J L Franzen (eds.):The Early Evolution of Man, With Special Emphasis on Southeast Asia and Africa. Cour. Forsch. Senckenberg 69r145-165. Brauer G (1984b) A craniological approach to the origin of anatomically modern Homo supiens in Africa and implications for the appearance of modern Europeans. In FH Smith and F Spencer (eds.): The Origin of Modern Humans: A World Survey of the Fossil Evidence. New York: Alan R. Liss, pp. 327-410. Brauer G (19901The occurrence of some controversial Homo erectus cranial features in the Zhoukoudian and East African hominids Acta Anthropol Sin 9 352-358 Brooks S, Brooks RH, and France D (1990) Alveolar prognathism contour, a n aspect of racial identification. In GW Gill and S Rhine (eds.): Skeletal Attribution of Race: Methods for Forensic Anthropology. Albuquerque Maxwell Mus. Anthropol., Anthropol. Pap. No. 4, pp. 41-46. Brues AM (1990)The once and future diagnosis of race. In GW Gill and S Rhine (eds.): Skeletal Attribution of Race: Methods for Forensic Anthropology. Albuquerque Maxwell Mus. Anthropol., Anthropol. Pap. No. 4, pp. 1-7. Cadien J D (1972) Dental variation in man. In SL Washburn and P Dolhinow (eds.): Perspectives on Human Evolution. Val. 2. New York: Halt Rinehart and Winston, Inc., pp 199-222 Cann RL (1988) DNA and human origins. Ann Rev. Anthropol. 17:127-143. Cann RL, Stoneking M, and Wilson AC (1987) Mitochondrial DNA and human evolution. Nature 329:31-36. Caspari R and Wolpoff MH (1990) The morphological affinities of the Klasies River Mouth skeletal remains. Am J Phys Anthropol 81:203. Chen T and Yuan S (1988) Uranium-series dating of bones and teeth from Chinese Palaeolithic sites. Archaeometry 30:59-76. Chen T and Zhang Y (1991) Palaeolithic chronology and possible coexistence of Homo erectus and Homo supiens in China. World Arch. 23t147-154. Chen T, Yuan S, Gao S, Wang L, and Zhao G (1982) Uranium series dating of Xujiayao (Hsuchia-yao) site. Acta Anthropol. Sin. 1:91-95. Chen T, Yuan S, and Gao S (1984) Using the uranium method to investigate important Palaeolithic dates in northern China. Acta Anthropol. Sin. 3:259-269. Chen T, Yuan S, Gao S, and Hu Y (1987) Uranium series dating of fossil bones from Hexian and Chaoxian fossil human sites. Acta Anthropol. Sin. 6r249-254. [Val. 35, 1992 Cheng G, Lin J , Li S, and Liang Q (1978) A preliminary paleomagnetic survey of the Nihewan bed. Scient. Geol. Sin. 7r247-252. Clarke FLJ (1990) The Ndutu cranium and the origin of Homo supiens. J. Hum. Evol. 19:699-736. Cook J , Stringer CB, Currant SP, Schwarcz HP, and Wintle AG (1982) A review ofthe chronology of the European Middle Pleistocene hominid record. Yrbk. Phys. Anthropol. 25:19-65. Coon CS (1962) The Origin of Races. New York: Alfred A Knopf. Delson E (1986) Human phylogeny revised again. Nature 322:496-497. Delson E (1988)One source not many. Nature 332: 206. Dc!son E, Eldrcdg:: N, 3nd Tattersdl I (19??! IEc construction of hominid phylogeny, a testable framework based on cladistic analysis. J . Hum. Evol. 6:264-278. Derbyshire E (1991) The quaternary evolution of the Qinghai-Xizang (Tibet) Plateau. In Special Proceedings Review Reports XI11 Inter. Cong., Beijing: INQUA, pp. 69-71. Ding Z (1991) Pedostratigraphy of Chinese loess and quaternary climatic fluctuation. In Liu D (ed.): Quaternary Geology and Paleoenvironment. Beijing: Science Press, pp. 168-172. Dong X (1989) Homo erectus in China. In Wu R, Wu X, and Zhang S (eds.): Early Humankind in China. Beijing: Science Press, pp. 9-23. Etler DA (1990) A case study of the “erectus”-“sapiens” transition in Asia: Hominid remains from Hexian and Chaoxian Counties, Anhui Province, China. Kroeber Anthropol. Sac. Pap. 71-72:l-19. Frayer D (1991) The persistence of Neanderthal features in post-Neanderthal Europeans. In G Brauer and FH Smith (eds.): Continuity or Complete Replacement: Controversies in the Evolution of Homo supiens. New York: Balkema. Gao J (1975) Australopithecine teeth associated with Giguntopithecus. Vert. Palas. 13231-88. Gill GW and Gilbert BM (1990) Race identification from the midfacial skeleton: American blacks and whites. In GW Gill and S Rhine (eds.): Skeletal Attribution of Race: Methods for Forensic Anthropology. Maxwell Mus. Anthropol., Anthropol. Pap. No. 4, pp. 47-53. Gingerich PD (1979) The stratophenetic approach to phylogeny in vertebrate paleontology. In J Cracraft and N Eldridge (eds.): Phylogenetic Analysis and Paleontology. Proceedings of a Symposium entitled “Phylogenetic Models” convened at North American Paleontological Convention 11, Lawrence, Kansas, Aug. 8,1977. New York: Columbia University of Press. Groves CP (1989) A regional approach to the problem of the origin of modern humans in Australasia. In P Mellars and CB Stringer (eds.): The Human Revolution: Behavioural and Biological Perspectives on the Origins of Modern Humans. Edinburgh Edinburgh University Press, pp. 274-285. Grun R and Stringer CB (1991) Electron spin resonance dating and the evolution of modern humans. Archaeometry 33:2-47. Grun R, Beaumont PB, and Stringer CB (1990) ESR dating evidence for early modern humans at Border Cave in South Africa. Nature 344:537539. Grun R, Stringer CB, and Schwarcz HP (1991) Popel CRANIOFACIAL EVIDENCE F(7 R MODERN HUMAN ORIGINS ESR dating of the teeth from Garrod's Tabun cave collection. J. Hum. Evol. 20.231-248. Guo S, Liu S, Sun S, Zhang F, Zhou S, Hau X, Hu R, Meng W, Zhang P, and Liu J (1991) Fission track dating of 4th layer of the Peking Man site. Acta Anthropol. Sin. 10.73-77. Habgood PJ (1989) The origin of anatomicallymodern humans in Australasia. In P Mellars and CB Stringer (eds.): The Human Revolution: Behavioural and Biological Perspectives on the Origins of Modern Humans. Edinburgh: Edinburgh University Press, pp. 245-273. Han D and Xu C (1985) Pleistocene mammalian faunas of China. In Wu R and JW Olsen (eds.): Palaeoanthropology and Palaeolithic Archaeology in the People's Republic of China. Orlando, Academic Press, pp. 267-286. Han U and Xu C (1989) Quaternary mammalian faunas and environment of fossil humans in South China. In Wu R, Wu X and Zhang S (eds.): Early Humankind in China. Beijing: Science Press, pp. 338-391. Hanam AG and Wood WW (1989) Relationship between the size and spatial arrangement and morphology of the human masseter and medial pterygoid muscles and the craniofacial skeleton and jaw. Am. J. Phys. Anthropol. 80r429-465. Hill A, Ward S, Deino A, Garniss C, and Drake R (1992) Earliest Homo. Nature 355:719-722. Hinton RJ (1979) Form and function in the temporomandibular joint. In DS Carlson (ed.): Craniofacial Biology, Vol. 1, Monograph No. 10, Craniofacial Growth Series. Ann Arbor: Center for Human Growth and Development. Hinton RJ (1981) Changes in articular eminence morphology with dental function. Am. J. Phys. Anthropol. 54t439-455. Howell FC (1978) Hominidae. In VJ Maglio and HBS Cooke (eds.): Evolution of African Mammals. Cambridge: Harvard University Press, pp. 154-248. Howell FC (1986) Variabilite chez Homo erectus, et probleme de la presence de cette espece en europe. L'Anthropologie 90.447-481. Howells WW (1973) Cranial variation in man. A multivariate analysis of patterns of difference among recent huiuari populations. Papers of the Peabody Museum, Harvard University, Vol. 67. Howells WW (1981) Homo erectus in human descent, ideas and problems. In BA Sigmon and JS Cybulski (eds.): Homo erectus. Toronto: University of Toronto Press, pp. 63-85. Howells WW (1989) Skull shapes and the map. Craniometric analyses in the dispersion of modern humans. Peabody Mus. Arch. and Ethnol., Harvard University. Cambridge: Harvard University Press. Huang R, J i n S, Liang R, Lu Z, Yuan Z, Cai B, and Fang Z (1991) Study of ESR dating for burying age of the first skull of Beijing Man and chronological scale of the cave deposit in Zhoukoudian site LOC.1. Acta Anthropol. Sin. lOr107-115. Huang W (1991a) Quaternary cave accumulations and mammalian fauna in China. In Liu D (ed.): Quaternary Geology and Environment in China. Beijing: Science Press, pp. 73-77. Huang W (1991b) The late Pleistocene mammalian faunas of China. In Contributions to the XI11 INQUA, Institute of Vertebrate Paleontology and Paleoanthropology, Academia Sinica, Beijing Sci- 293 entific and Technological Publishing House, Beijing, pp. 44-54. Huang W and Fang Q (1991) Wushan hominid site. Beijing: China Ocean Press. Huang W and Huang C (1985) Mammal fossils and sporo-pollen compositions at Hexian Man locality and their significance. In (Selection) Symposium of National Conference on Quaternary Glacier and Periglacial. Beijing: Kexue Chubanshe, pp. 180-183. Huang W and You Y (1987) On the problems of the karst cave and the deposits of the site of Jinniushan Man. Carsologica Sin. 6t61-67. Huang W, Fang D, and Ye Y (1982) Preliminary study of the fossil hominid and fauna from Hexian, Anhui. Vert. PalAs. 20r248-256. Huang W; T,iu Y, Li C , Yuan X, Zhang Z, Zat~gX, and Xie G (1987) Tentative opinions on the age of Baise industry. In Treatises in Commemoration of the 30th Anniversery of the Discovery of Maba Human Cranium. Guangdong Provincial Museum and the Museum of the Qujiang County. Beijing: Cultural Relics Publishing House, pp. 95-101. Hublin JJ (1985) Human fossils from the north African middle Pleistocene and the origin of Homo supiens. In E Delson (ed.): Ancestors: The Hard Evidence. New York: Alan R. Liss, pp, 283288. Hublin JJ (1986) Some comments on the diagnostic features of Homo erectus. In VV Novotny and A Mizerova (eds.): Fossil Man. New Facts, New Ideas. Papers in Honor of J a n Jelinek's Life Anniversary. Anthropos (Brno) 23r175-188. Hyland WL, Picq PG, and Johnson KR (1991) Masticatory stress hypotheses and the supraorbital region of primates. Am. J . Phys. Anthropol. 86: 1-36. Itihara M, Arkhipov SA, Wang P, and Kumai H (1991) Major subdivisions of the Quaternary of Asia, and their litho- and biostratigraphic significance. In Special Proceedings Review Reports XI11 Inter. Cong., Beijing: INQUA, p. 75. J i H (1991) The middle Pleistocene mammalian faunas of China. In Contributions to the XI11 INQUA, Institute of Vertebrate Paleontology and Paleoanthropology, Acadeiliia Sinica. Beijing Scientific and Technological Publishing House, Beijing: 38-43. J i a L and Huang W (1990) The Story of Peking Man. Hong Kong and Oxford: Foreign Language Press and Oxford University Press. J i a L and Wei Q (1980) Some animal fossils from the Holocene of north China. Vert. PalAs. 18: 327-333. J i a L and Wei Q (1987) Stone artifacts from Lower Pleistocene at Donggutuo site near Nihewan (Nihowan), Hebei Province, China. L'Anthropologie 91r727-732. Jia L, Wei Q, and Li C (1979) Report on the excavation of Hsuchiayao Man site in 1976. Vert. PalAs. 17:277-293. Jones R (1989) East of Wallace's line: Issues and problems in the colonization of the Australian Continent. In P Mellars and CB Stringer (eds.): The Human Revolution: Behavioural and Biological Perspectives on the Origins of Modern Humans. Edinburgh: Edinburgh University Press, pp. 743-782. Kennedy GE (1991) On the autopomorphic traits of Homo erectus. J Hum. Evol. 20t375-412. 294 YEARBOOK OF PHYSIf7 A L ANTHROPOLOGY Kennedy KAR, Sonakia A, Chiment J, and Verma KK (1991) Is the Narmada hominid an Indian Homo erectus. Am. J of Phys. Anthropol. 86:475496. Klein RG (1989) The Human Career: Human Biological and Cultural Origins. Chicago: University of Chicago Press. Klein RG (1992) The archeology of modern human origins. Evol. Anthropol. I:5-14. Knusel CJ (1991) Biomechanics of the hominie cranium with special reference to Homo erectus and the archaic Homo sapiens. PhD dissertation, Simon Fraser University. Kuhle M (1991)The Pleistocene glaciation of Tibet and the onset of ice ages-an autocycle hypothesis Tn Special Proceedings Review Reportc: XTTT Inter. Cona., Beiiina: INQUA4,I). 72. Lee JS (1959) Th; 6eolog; of China. New York: Nordeman Publishing. LeGros Clark WE (1955) The Fossil Evidence for Human Evolution. Chicago: University of Chicago Press. Leigh SR (1992) Cranial capacity evolution in Homo erectus and early Homo sapiens. Am. J . Phys. Anthropol. 87:l-13. Lestrel PE (1976) Hominid brain versus time: Revised regression estimates. J . Hum. Evol. 5:207212. Li H and Mei Y (1983) The dating of the upper section of Hexian man deposits. Scientia Sinica 28r703. Li H and Wang J (1982) Magnetostratigraphic study of several typical geologic sections in north China. In Quaternary Geology and Environment of China, Quaternary Research Association of China. Beijing: China Ocean Press. Li H and Wang J (1991) The latest advance in Quaternary magneto-statigraphy of China. In Liu T (ed.): Quaternary Geology and Paleoenvironment of China. Beijing: Science Press, pp. 158-167. Li J (1991) The uplift of the Qinghai-Xizang Plateau and its effect on environment. In Liu T (ed.): Quaternary Geology and Environment of China. Beijing: Science Press, pp. 265-272. Li T (1991: Unearthing of Chaoxian man's fossil skull. China Cultural Report 5.1. Li T and Etler D (1992) New Middle Pleistocene crania from Yunxian in China. Nature Vol:357 404-407. Li T, Wen Z, Li W, Ma X, Hu K, and Liu W (1991) Survey and preliminary excavations of the Chaoxian Quyuan river fossil locality. Jiang Han Archeology 2:l-15. Li Y (1989) The early Palaeolithic of China. In Wu R and Wu X (eds.): Early Humankind in China. Beijing: Science Press, pp. 159-194. Liu C, Zhu X, and Ye S (1977) A palaeomagnetic study on the cave deposits of Zhoukoudian (Choukoutien), the locality of Sinanthropus. Scien. Geol. Sin. 1:25-30. Liu J, Zhen S, and Jin F (1982) Some new species of mammals in the middle Pleistocene from Chaoxian, Anhui, with a discussion of the transitional district of Huai river. Mem. Beijing Nat. Hist. Mus. I9:1-26. Liu S, Zhang F, Hu R, Liu J , Gou T, and Zhou S (1985) Dating Peking man by fission track. In Wu R, Ren M, Zhu X, Yang Z, Hu C, Kong Z, Xie Y, and Zhou S (eds.): Multidisciplinary Study of the IVol. 35, 1992 Peking Man at Zhoukoudian. Beijing: Science Press, pp. 241-245. Liu T (1991a) Loess, Environment and Global Change. Beijing: Science Press. Liu T (1991b) Quaternary Geology and Environment. Beijing: Science Press. Liu T and Ding M (19844 The characteristics and evolution of the palaeoenvironment of China since the late Tertiary. In RO Whyte (ed.): The Evolution of the East Asian Environment. Vol. I. Hong Kong: Centre of Asian Studies of the University of Hong Kong, pp. 11-40. Liu T and Ding M (1984b)Mammoths in China. In PS Martin and RG Klein (eds.): Quaternary Extinctions: A prehistoric revolution. Tucson: The University of Arizona Press, pp. 517-527. Liu T, Lu Y, Zheng H, Wu Z, and Yuan B (1985) Loess and the Environment. Beijing: China Ocean Press. Liu X and Xia Z (1983) A suggestion on the division and correlation of the Nihewan formation. Marine Geology and Quaternary Geology 3:7585. Lu Z (1985) The exacavation and significance of the Jinniushan site. Beijing: Peking University Archaeological Department Publication. Lu Z (1989) On the time of Jinniushan (Gold Ox Mount) man and the position of its evolution. Liaohai Cult. Relics J . 1:44-55. Lii Z (1990) La decouverte de l'homme fossile de Jing-Niu-Shan premiere fitude. L'Anthropologie 94.899-902. Lu Z (1992) Mosaic evolution of Jinniushan Archaic Homo sapiens. In D Croess (ed.): Proceedings of the Circumpacific Prehistory Conference, Washington State University, Seattle, in press. Lu Z, Huang Y, Li P, and Meng Z (1989) Yiyuan fossil man. Acta Anthropol. Sin. 8:303-313. Luo B, Wang Y, Lin Z, Chen M, Lan C, and Fu M (1988) Paleontological research in Nihewan Beds. In Chen M (ed.): Study of the Nihewan Beds. Beijing: Ocean Press, pp. 40-62. Mercier N, Valladas H, Joron J-L, Reyas J-L, Leveque F, and Vandermeersch B (1991) Thermoluminescence dating of the late Neanderthal remains from Saint-Cesaire. Nature 351:737739. Movius HL (1944) Early man and Pleistocene stratigraphy in southern and eastern Asia. Papers of the Peabody Museum, Harvard University 9:l-125. Mizoguchi Y (1985) Shovelling: A statistical analysis of its morphology. The Univers. Mus. Univers. of Tokyo Bull. 26. Tokyo: University of Tokyo Press. Patterson C and Rosen DE (1977) Review of Ichthyodectiform and the Mesozoic teleost fishes and the theory and practise of classifing fossils. Bull. Am. Mus. Nat. Hist. 158:83-172. Pei W, Wu J, Chia L, Chou M, Liu H, and Wang C (1958) Report on the excavation of palaeolithic sites at Ting-ts'un Hsiang-fen-hsien, Shansi Province. Institute of Vertebrate Palaeontology and Palaeoanthropology Mem. 2:l-111. Peking: I.V.P.P. Pope GG (1977) Hominids from the Lower Pleistocene of South China. Kroeber Anthropol. SOC. Pap. 50:63-73. Pope GG (1982) Hominid evolution in East and Southeast Asia. PhD thesis, University of California, Berkeley. Pope1 CRANIOFACIAL EVIDENCE F(3R MODERN HUMAN ORIGINS 295 Pope GG (1983) Evidence of the age of the Asian Rightmire G P (1986a) Body size and encephalization in Homo erectus. Anthropos (Brno) 23r139hominidae. Proc. Natl. Acad. Sci. U.S.A. 80: 150. 4988-4992. Rightmire GP (1986b) Stasis in Homo erectus dePope GG (1984) The antiquity and paleoenvironfended. Paleobiology 12:324-325. ment of the Asian hominidae. In RO Whyte (ed.): The Evolution of the East Asian Environment. Rightmire GP (1989) Middle stone age humans Vol. 11. Hong Kong: Centre of Asian Studies of the from eastern and southern Africa. In P Mellars University of Hong Kong, pp. 822-847. and CB Stringer (eds.): The Human Revolution: Pope GG (1985) Taxonomy, dating, and paleoenviBehavioural and Biological Perspectives on the ronment: The paleoecology of the early Far EastOrigins of Modern Humans. Edinburgh: Edinern hominids. Mod. Quat. Res. in Southeast Asia burgh University Press, pp. 109-122. 9:65-80. Rightmire G P (1990)The evolution of Homo erecPope GG (1988) Recent advances in Far Eastern tus. Cambridge: Cambridge University Press. paleoanthropology. Ann. Rev. Anthropol. 17:43Russell MD (19851The supraorbital torus. A most 77. remarkable peculiarity. Curr. Anthropol. 26: Pope GG (1989a) Asian Homo erectus and the 337-350. emergence of Homo supiens: An alternative to Rctter N (13311 Paieociimatic interpretations of Iepiacement modeis. In A Sahni and R Gauer loess-soil sequence. In Special Proceedings Re(eds.): Perspectives in Human Evolution. SRK view Reports XI11 Inter. Cong., Beijing: INQUA, Chopra Festschrift Volume. Delhi: Delhi Renaispp. 73-74. sance Publishing House, pp. 21-32. Santa Luca AP (1980)The Ngandong fossil homiPope GG (198913) Bamboo and human evolution. nids. Yale Univers. Publ. Anthropol. 78:l-175. Nat. Hist. Oct.:48-57. Schwarcz HP, Grun R, Latham AG, Mania D, and Pope GG (1991) Evolution of the zygomaticomaxBrunnacker K (1988) The Bilzingsleben archaeillary region in the genus Homo and its relevance ological site: New dating evidence. Archaeometry to the origin of modern humans. J. Hum. Evol. 30:5-17. 21:is9 -2i3. Schwarcz HP, Buhay WM, Griin R, Valladas H, PoDe GG and Keates SG (1992) The evolution of Tchernov E, Bar-Yosef 0, and Vandermeersch B h i m a n cognition and cultural capacity: A view (1989) ESR Dating of the Neanderthal site, Kefrom the Far East. In preparation. bara Cave, Israel. J Arch. Sci. 16:653-659. Pope GG, An 2, Keates S, and Bakken D (1990) Schwartz JH, Tattersall I, and Eldredge N (1978) New discoveries in the Nihewan Basin, Northern Phylogeny and the classification of Primates reChina. The East Asian TertiaryiQuaternary visited. Yrbk. Phys. Anthropol. 21r95-132. Newsletter. 11:68-73. Poulianos AN (1971) Petralona: A middle Pleis- S h a h F, Semah AM, and Djubiantono T (1990) They Discovered Java. Jakarta: Pusat Penelitian tocene cave in Greece. Archeology 24:6-11. Arkeologi and Museum National DHistoire NaQi G (1989) Quaternary mammalian faunas and turelle. environment of fossil humans in North China. In Wu R, Wu X, and Zhang S (eds.): Early Human- Simmons T and Smith FH (1991) Human population relationships in the late Pleistocene. Curr. kind in China. Beijing: Science Press, pp. 237Anthropol. 32:623-627. 337. Qi G (1990) The Pleistocene human environment Smith FH and Ranyard GC (1980)Evolution of the supraorbital region in Upper Pleistocene fossil of North China. Acta Anthropol. Sin. 9:340-349. hominids from South-Central Europe. Am. J . Qian F, Zhou G, et al. (1991) Quaternary geology Phys. Anthropol. 53~589-609. and palaeoanthropology of Yuanmou, Yunnan, Smith FH, Falsetti AB, and Donnelly SM (1989) China. Beijing: Science Press. Modern human origins. Yrbk. Phys. Anthropo!. Qiu Z (1989) The middle Palaeolithic of China In 32135-68. Wu R, Wu X, and Zhang S ieds.1: Eariy HurnanSong S and Zhang Z (1988) The Maba fauna. In: kind in China. Beijing: Science Press, pp. 195Treatises in Commemoration of the 30th Anni219. versary of the Discovery of Maba Human CraQui Z, Gu Y, Zhang Y, and Zhang S (1973)Newly nium. Guangdong Provincial Museum and the discovered Sinanthropus remains and stone artiMuseum of the Qujiang County, Beijing: Cultural facts at Choukoutien. Vert, PalAs. 11:109-131. Relics Publishing House, pp. 23-35. Qui Z, Xu C, Zhang J, Wang W, and Zhao C (1982) Stringer CB (1981)The dating of European middle Hominid and mammalian fossil discoveries from Pleistocene hominids and the existence of Homo Zhaoxian. Acta Anthropol. Sin. 1tl-9. erectus in Europe. L'Anthropologie 19:3-14. Qun L (1983) Fossil ape-man teeth discovered at White Dragon Cave, Hubei. Acta Anthropol. Sin. Stringer CB (1984)The definition of Homo erectus and the existence of the species in Africa and Eu2:203. rope. Cour. Forsch. Senckenberg 69:131-145. Rak Y (1983) The Australopithecine Face. New York, Academic Press. Stringer CB (1988)The dates of Eden. Nature 331: 565-566. Rhine S (1990) Non-metric skull racing. In GW Gill and S Rhine (eds.): Skeletal Attribution of Stringer CB (1989a) The origin of early modern Race: Methods for Forensic Anthropology. Maxhumans: A comparison of the European and nonwell Mus. Anthropol., Anthropol. Pap. No. 4, pp. European evidence. In P Mellars and CB Stringer (eds.): The Human Revolution: Behavioural and 9-20. Biological Perspective on the Origins of Modern Rightmire GP (1984)Comparisons of Homo erectus Humans. Edinburgh: Edinburgh University from Africa and Southeast Asia. In P Andrews Press, pp. 232-244. and J L Franzen (eds.): The Early Evolution of Man, With Special Emphasis on Southeast Asia Stringer CB (1989b) Documenting the origin of and Africa. Cour. Forsch. Senckenberg 69r83-98. modern humans. In E Trinkaus (ed.): The Emer- 296 YEARBOOK OF PHYSIf?AL ANTHROPOLOGY gence of Modern Humans. Cambridge: Cambridge University Press, pp. 97-141. Stringer CB (1990) The Asian Connection. New Scientist 178:33-37. Stringer CB (1991) Asia and recent human evolution. In: Special Proceedings Review Reports XI11 International Congress. Beijing: INQUA, p. 342. Stringer CB and Andrews P (1988) Genetic and fossil evidence for the origin of modern humans. Science 239:1263-1268. Stringer CB and Grun R (1991) Time for the last Neanderthal. Nature 351:701-702. Sun D (1991) The hunting for the Quaternary glaciers in China. In Liu D (ed.): Quaternary Geology and Environment in China. Beijing: Science Press, pp 1-1 5 Tang Y (1991) The early Pleistocene mammalian faunas of China. In Contributions to the XI11 INQUA, Institute of Vertebrate Paleontology and Paleoanthropology,Academia Sinica, Beijing Scientific and TechnologicalPublishing House, Beijing: 21-31. Thorne AG and Wolfpoff MH (1981) Regional continuity in Australasian Pleistocene hominid evolution. Am. J. Phys. Anthropol. 55:337-349. Thorne AG and Wolpoff MH (1992) The Multiregional Evolution of Humans. Scientific American 266:28-33. Tobias PV (1991) The skulls, endocasts and teeth of Homo habilis. Vol. 4, Parts I-IV, V-IX. Cambridge: Cambridge University Press. Turner CG I1 (1989) Teeth and prehistory of Asia. Sci. Am. 260:88-96. Turner CG I1 (1990) Major features of Sundadonty and Sinodonty, including suggestions about East Asian microevolution, population history, and late Pleistocene relationships with Australian aboriginals. Am. J. Phys. Anthropol. 82:295317. von Koenigswald GHR (1949) The Pleistocene. In General Geology of Indonesia and Adjacent Archipelagos Vol. 1A. The Hague: Government Printing Office. von Koenigswald GHR (1951) Introduction. In Weidenreich F (ed.):Morphology of Solo Man. Anthropol. Pap. Am. Mus. Nat. Hist. 43:211-221. Wang L (1989) New progress in chronology in Chinese paleoanthropology. In Wu R, Wu Z, and Zhang S (eds.): Early Humankind in China. Beijing: Science Press, pp. 392-431. Wang N and Yang J (1991) Major morphotectonic patterns and regions of China. In Liu T (ed.): Quaternary Geology and Environment in China. Beijing: Science Press, pp. 253-264. Wang S (1988) Palaeomagnetic stratigraphy of Nihewan beds. In Chen M (ed.): Study of the Nihewan Beds. Beijing: Ocean Press, pp. 117-123. Wang S and Li J (1991) Late Cenozoic lake sediments in China. In Liu T (ed.): Quaternary Geology and Environment in China. Beijing: Science Press, pp. 41-57. Wang Y, Xue X, Yue L, Zhao J , and Liu S (1979) Discovery of Dali fossil man and its preliminary study. Kexue Tongbao 24:303-306. Wei Q (1978) New discoveries from the Nihewan formation and other aspects regarding its study. In Treatises on Paleoanthropology. Beijing: Science Publishing Society, pp. 136-150. Wei Q (1988) The stratigraphic, geochronological and biostratigraphir hackground of the earliest IVol. 35, 1992 known palaeolithic sites in China. L'Anthropologie 92:931-938. Wei Q and Xie F (1990) Treatises on Nihewan. Beijing: Cultural Relics Publ. House. Wei Q, Meng H, and Cheng S (1985) New Palaeolithic site from the Nihewan (Nihowan) beds. Acta Anthropol. Sin. 4:223-232. Weidenreich F (19351 The Sinanthropus population of Choukoutien (Locality 1) with a preliminary report on new finds. Bull. Geol. SOC.China 14:427-461. Weidenreich F (1937) The dentition of Sinanthropus pekinensis: A comparative odontography of the hominids. Pal. Sin. New Series D 1:1-180. Weidenreich F (1939) On the earliest representatives of modern mankind recovered on the soil of East Asia. Peking. B d l . Nat. Hist. 9oc. 13r191174. Weidenreich F 11943) The skull of Sinanthropus pekinensis: A comparative study on a primitive hominid skull. Pal. Sin., New Series D 1O:l-484. Weidenreich F (1945) Giant early man from Java and South China. Anthropol. Pap. Am. Mus. Nat. 4O:l-134. Weidenreich F (1951) Morphology of Solo Man. Anthropol. Pap. Am. Mus. Nat. Hist. 43:205-290. Wilson AC and Cann RL (1992) The recent African genesis of humans. Sci. Am. 266(4):66-73. Wolpoff MH (1980) Paleoanthropology New York Knoff. Wolpoff MH (1985) Human evolution a t the peripheries: The pattern a t the eastern edge. In PV Tobias (ed.): Hominid Evolution: Past, Present, and Future. New Y o r k Alan R. Liss, pp. 335365. Wolpoff MH (1986) Describing anatomically modern Homo sapiens. A distinction without a definable difference. In VV Novotony and A Mizerova (eds.): Fossil Man New Facts-New Ideas. Papers in Honour of Jan Jelinek's Life Anniversary. Wolpoff MH (1989) Multiregional evolution: The fossil alternative to Eden. In P Mellars and CB Stringer (eds.): The Human Revolution: Behavioural and Biological Perspectives on the Origins of Modern Humans. Edinburgh Edinburgh University Press, pp. 62-108. Wolpoff MH and Caspari R (1990) Metric analysis of the skeletal material from Klasies River Mouth, Republic of South Africa. Am. J. Phys. Anthropol. 81:319. Wolpoff MH and Nkini A (1985) Early and middle Pleistocene hominids from Asia and Africa. In E Delson (ed.): Ancestors: The Hard Evidence. New York Alan R. Liss, pp. 202-205. Wolpoff MH, Wu X, and Thorne AG (1984) Modern Homo sapiens origins: A general theory of hominid evolution involving the evidence from East Asia. In FH Smith and F Spencer (eds.): The Origin of Modern Humans. New York Alan R. Liss, pp. 411-483. Wolpoff MH, Spuhler JN, Smith FH, Radovic J , Pope G, Frayer DW, Eckhardt R, and Clark G (1988) Modern human origins. Science 241 :772773. Wood B (1984) The origin of Homo erectus. Cour. Forsch. Senckenberg 69:99-113. Wood B (1985) Early Homo in Kenya and its systematic relationships. In E Delson (ed.): Ancestors: The Hard Evidence. New York: Alan R. Liss, pp. 206-214. Popel CRANIOFACIAL EVIDENCE FOIR MODERN HUMAN ORIGINS Wu M (1980) Human fossils discovered at Xujiayao site in 1977. Vert. PalAs. 18t229-238. Wu M (1983) Homo erectus from Hexian, Anhui found in 1981. Acta Anthropol. Sin. 2t109-115. Wu M (1986) Investigation of Xujiayao man fossils. Acta Anthropol. Sin. 5:220-226. Wu M (1989) Late Chinese Homo sapiens. In Wu R, Wu M, and Zhang S (eds.): Early Humankind in China. Beijing: Science Press, pp. 42-61. Wu R (1959) Human fossils found in Liukiang, Kwangsi, China. Palaeovertebrata et Palaeoanthropologia 1:97-104. Wu R (1988) The reconstruction of the fossil human skull from Jinniushan, Yingkou, Liaoning Province and its main features. Acta Anthropol. Sin. 7:97-101. Wu R (1989) A new theory about the origin of modern hnmans. A& Arlihropoi. Sin. 6:182-185. Wu R and Dong X (1980) Fossil ape-man teeth from Yunxian, Hubei. Acta Anthropol. Sin. 18: 142-149. Wu R and Dong X (1982) Preliminary study of Homo erectus remains from Hexian, Anhui. Acta Anthropol. Sin. 1.2-13. Wu R and Dong X (1985a) Homo erectus in China. In Wu R and-JW Olsen (eds.): Palaeoanthropology and Palaeolithic Archaeology in the People’s Republic of China. Orlando: Academic Press, pp. 79-89. Wu R and Dong X (1985b) Retrospect and prospect on the study of Peking Man. In Wu R, Ren M, Zhu X, Yang 2, Hu C, Kong Z, Xie Y, and Zhou S (eds.): Multidisciplinary study of the Peking Man at Zhoukoudian. Beijing, Science Press, pp. 8694. Wu R and Lin S (1985) Chinese paleoanthropology: Retrospect and Prospect. In Wu R and J W Olsen (eds.): Palaeoanthropology and Palaeolithic Archaeology in the People’s Republic of China. Orlando: Academic Press, pp. 1-27. Wu R and Olsen JW (1985) Palaeoanthropology and Palaeolithic Archaeology in the People’s Republic of China. Orlando: Academic Press. Wu R and Peng R (1959) Fossil human skull of early Paleo-anthropic stage found at Mapa, Shaokuan, Kwangtung Province. Palaeovert. Palaeoanthropol. 4:159-164. Wu R and Wu X i1982! Hominid fossil teeth from Xichuan, Henan. Vert. Palas. 2O:l-9. Wu R and Xu X (1984) Hominid fossils from China and their relation to those of neighboring regions. In RO Whyte (ed.): The Evolution of the East Asian Environment Vol. 11. Hong Kong: Centre of Asian Studies of the University of Hong Kong, pp. 787-795. Wu R, Ren M, Zhu X, Yang Z, Hu C, Kong Z, Xie Y, Zhao S, et al. (1985) Multidisciplinary Study of the Peking Man Site at Zhoukoudian. Beijing: Science Press. Wu R, Wu X, and Zhang S (1989) Early Humankind in China. Beiiing: Science Press. Wu X (1981) The wefi-preserved cranium of a n early Homo sapiens from Dali, Shaanxi, China. Scient. Sin. 2:200-206. Wu X (1988a) Comparative study of early Homo sapiens from China and Europe. Acta Anthropol. Sin. 7r287-293. Wu X (1988b) The place of Maba man in human evolution. In: Treatises in Commemoration of the 30th Anniversary of the Discovery of Maba Human Cranium. Guangdong Provincial Museum 297 and the Museum of the Qujiang County, Beijing: Cultural Relics Publishing House, pp. 3-7. Wu X (1989) Early Homo sapiens in China. In Wu R, Wu X, and Zhang S (eds.): Early Humankind in China. Beijing: Science Press, pp. 24-41. Wu X (1990) The evolution of humankind in China. Acta Anthropol. Sin. 9t312-321. Wu X (1991) Fossil humankind of China. In Liu T (ed.): Quaternary Geology and Paleoenvironment in China. Beijing: Science Press, pp. 129-136. Wu X and Wu M (1985) Early Homo sapiens in China. In Wu R and JW Olsen (eds.): Palaeoanthropology and Paleolithic Archaeology in the People’s Republic of China. Orlando: Academic Press, pp. 91-106. Wu X and You Y (1979) A preliminary observation . 17.294-303. of Dali mansite V P ~Ps!As. Wu X and You Y (1980) Dali man and its culture. Archaeological and Cultural Relics 1:2-6. Xu C and Zhang Y (1990) Human fossils discovered at Chaoxian, Anhui, China. Hum. Evol. 4: 95-96. Xu C, Zhang Y, Zheng C, and Fang D (1984) The discovery of a human occipital fossil and mammalian fauna at Chaoxian, Anhui. Acta Anthropol. Sin. 3:202-209. Xu C, Zhang Y, and Fang D (1986a) Human fossil newly discovered a t Chaoxian, Anhui. Acta Anthropol. Sin. 5:305-310. Xu C, Zhang Y, and Fang D (1986b) New finds from the human fossil localities at Chaoxian, Anhui. Prehistory (Shiqian Yanjiu) 3-4.41-45. Xu Q and You Y (1982) Four post-Nihewanian Pleistocene mammalian faunas of North China: Correlation with deep-sea sediments. Acta Anthropol. Sin. 1:180-187. Xu Q and You Y (1984) Hexian fauna: Correlation with deep-sea sediments. Acta Anthropol. Sin. 3: 102-113. Xue X (1987) Human teeth from Luonan, Shaanxi and their geological dating. Acta. Anthropol. Sin. 6t284-288. Yuan B, Cui J , and Zhao X (1991) Typical Quaternary Sections in North China. Excursion Guide Book International Union for Quaternary Research, XI11 International Congress. Beijing Yuan S and Chcn T (1991) Chinese Quaternary Radioscopic Chronology. In Liu T (ed.): Quaternary Geology and Environment in China. Beijing: Science Press, pp. 179-184. Yuan S, Chen T, and Gao S (1986) The dating of southern Chinese Palaeolithic sites by uranium method. Acta Anthropol. Sin. 5t179-190. Zhang S (1990) Regional industrial gradual advance and cultural exchange of Paleolithic in North China. Acta Anthropol. Sin. 9.322-333. Zhang S and Zhou C (1984) A preliminary study of the second excavation of Dali man locality. Acta Anthropol. Sin. 3:19-29. Zhang Y (1986) The dental remains of early Homo sapiens found in China. Acta Anthropol. Sin. 5: 103-113. Zhang Y (1989) Tooth wear from early Homo sapzens from Chaohu and the hypothesis of use of anterior teeth as tools. Acta Anthropol. Sin. 8: 314-319. Zhang Y (1991) An examination of temporal variation in the hominid dental sample from Zhoukoudian Locality 1. Acta Anthropol. Sin. 10.8595. Zhang Z (1988) An analysis of the physical char- 298 YEARBOOK OF PHYSICAL ANTHROPOLOGY acteristics of Modern Chinese. Acta Anthropol. Sin. 8~314-323. Zhang 2, Wang Q, and Zhao Y (1991)Quaternary sediments in Nihewan Basin, Hebei Province. In Liu T (ed.): Quaternary Geology and Environment in China. Beijing: Science Press, pp. 343350. Zheng S and Han D (1991) Quaternary mammals of China. In Liu T (ed.): Quaternary Geology and Environment of China. Beijing: Science Press, pp. 101-114. Zhou M 11990) Discovery of the “Peking Man” and [Vol. 35, 1992 international scientific cooperation. Acta Anthropol. Sin. 9~300-302. Zhou M and Wang Y (1991) Pleistocene environmental changes, mammalian faunas and fossil man of China. In Contributions to the XI11 INQUA, Institute of Vertebrate Paleontology and Paleoanthropology,Academia Sinica, Beijing Scientific and Technological Publishing House, Beijing: 1-14. Zhou Y (1989) Amino acid dating of Peking man and Dingcun man. Acta Anthropol. Sin. 7:177181.