Reassessment of the La Ferrassie 3 Neandertal ossicular chain Rolf Quam Ignacio Martinez b.d, Juan Luis Arsuaga b.e 'Depoment of Anthropology, Binghomton University (SUNY), Binghomton, NY 13902-6000, USA Cenwo UCM-ISUII de Invertigoci6n robre 10 Evoluci6n y Comportomiento Humonor, Avda Monforte de iemor, 5, 28029 Modrid, Spoin 'Division of Anthropology, Americon Museum of Noturol History, Cmtml Pork West ot 79th sweet, New York, NY 10024-5192, USA *oniverridod de Alcold, Deportomento de Geologio (Areo de Poleontologio), EdlJicio de Ciencior, Compus Univerritorio, 28871 Alcold de Henorer, Spoin eUniverridod Complutense de Modrid, Deportomento de Poleontologio, Focultod de Ciencios Geol6g3cor, Ciudod Univerritorio r/n, 28040 Modrid, Spoin Keywords: Malleus 1"C"I C..... 3Ldprb Middle ear Evolution Homo neondertholensis A B S T R A C T The ossicular chain in La Ferrassie 3 was briefly described in the monograph on the La Ferrassie Nean- dertal children, but to date has not been the subject of detailed study. We provide new data on these important fossils and re-examine some previous suggestions of derived Neandertal features in the middle ear ossicles based on more limited evidence. The malleus shows a curved lateral margin of the manubrium and a relatively large head. The incus shows a tall articular facet, a depressed area on the medial surface of the body, a straight anterior border of the long process and a more closed angle be- tween the processes. The stapes shows an asymmetrical configuration of the crura, with an anteriorly skewed head. and eenerallv small dimensions. includine a smaller and relativelv wider staoedial foot- ~olar i tv of manv of these features remains to be clarified. the asvmmetrical staDes and anteriorlv skewed footplates in Homo sapiens cannot be explained by changes in body mass. Indeed, H sapiens seems to depart from the general mammalian pattern in combining an increase in stapes footplate size with a decrease in body mass. Although the malleus/incus lever ratio in La Ferrassie 3 is similar to that in H sapiens, Neandertals appear to be characterized by a slightly different spatial relationship and artic- ulation of the ossicular chain within the tympanic cavity. While only limited inferences can be drawn regarding hearing ability based on the ossicles, the few physiologically relevant dimensions in the La Ferrassie 3 ear bones are similar to H sapiens. Introduction The La Ferrassie 3 (LR) Neandertal is one of the very few human fossil specimens to preserve a complete ossicular chain (malleus. incus and stapes). The bones were briefly described and a few metric dimensions were provided by Heim in the monograph on the La Ferrassie Neandertal children (Heim. 1982)? Heim suggested that the Neandertal ossicles were distinct from those of fossil and * Corresponding author E-moil oddrerr: rquam@binghamton.edu (R Quam). Heim (1982) also mentions the presence of an incur within the ympanic caviy inLa Ferrarrie4 bir. This specimenwar raid to clorely resemble the incur in the LF3 individual. However examination of the original rpecimen did not reveal the presence of an incur, nor ir a photo of the rpecimen included in the monograph on the La Ferrarrie children. Thus, the existence of this rpecimen ir currently known only from the reference to it by Heim. recent Homo saviens in a number of subtle asvects. In varticular, the malleus in LR was slightly longer, with a larger head and a more oven anele between the manubrium and the headineck. In addi- tion, the manubrium was relatively straight with a well-developed lateral (short) process, the latter implying a greater protrusion of the tympanic membrane. The incus shows a longer long process. which is relatively straight and thin. In contrast, the short process is shorter and does not show a notch in its lower border. hi articular surface is described as large and the bodv as sliehtlv flatter than in H. sapiens. The stapes is slightly smaller and shows a pronounced asymmetry between the shorter anterior crus and the longer and more curved posterior crus. The head of the stapes is skewed anteriorly and the crura are thicker in LB. In sum, the Neandertal ear ossicles are said to differ from H. sapiens in the larger di- mensions of the malleus and incus, the smaller size and asymmetry of the stapes, the more open angle of the malleus and a more closed angle of the incus (Heim. 1982). Nevertheless. Heim was appro- oriatelv cautious in interoretine whether these slight differences held any evolutionary significance or were merely the result of individual variation. Several more recent studies have provided additional data for the LF3 ossicles (Masali et al.. 1991; Spoor, 2002; Crevecoeur. 2007; Quam and Rak 2008), based largely on measurements obtained from the scaled photo published by Heim. Since the original spec- imens were previously unavailable for study, this represented the only approach possible to obtain further metric data on these specimens. However, given the quality of the photograph, the ~ ~ imprecise orientation and very small size of the ossicles, this data should be considered tentative. On a recent trio, one of us IRMOl . - was able to study the original fossil specimens. Several developments since the original publication make a reassessment of the LF3 ear ossicles warranted. Studies of the inner ear in Neandertals have revealed differences from H sapiens in several asoects iSooor et al.. 2003). and desoite their diminutive ~. size, the middle ear ossicles are particularly well-suited to drawing phylogenetic inferences. At birth, these tiny bones are fully formed and have already reached adult dimensions (Scheuer and Black 2000). The embryological origin of each of the ear bones has been thoroughly studied, and their development is under tight genetic control (Mallo. 1998.2001; Frenz et al.. 2001). Comparative genomic studies have revealed accelerated rates of evolution in several eenes related to hearine in humans comoared with chim- panzees (Clark et al.. 2003), including some (e.g.. Eyal) that may be directly related to the formation of the ossicles (Xu et al.. 1999). A standardized measurement protocol, relying in part on previous studies (Masali. 1964). has been developed for the auditow ossicles - , Maureille. 2002; Spoor, 2002; Lisonek and Trinkaus. 2006; Quam et al.. 2006; Crevecoeur. 2007). In addition. the ear ossicles play an important physiological role in audition (Wever and Lawrence, 1954; Kirikae. 1960). Anatomical differences in the ear ossicles across ~r imates have been shown to be correlated with asoects of their hearing sensitivitv (Coleman and Ross. 2004; Coleman and Colbert. 2010), and fossil hominin ear ossicles have been included in models that reconstruct the auditorv caoacities in hominin . . species (Martinez et al.. 2004, in press; Quam et al.. 2012). The preservation of a complete ossicular chain in LF3 is excep- tional. The only other published fossil hominin specimens to pre- serve all three ear ossicles are the Middle Paleolithic H. sapiens specimen from Darra-i-Kur (Angel. 1972) and the recently redis- covered Le Moustier 2 Neandertal infant (Maureille. 2002). In addition to LF3, published data on Neandertal ear ossicles is mainly limited to the staves from Subalvuk 2 and the incudi from Amud 7 and Le Moustier 1 (Arensburg et al.. 1996; Ponce de Le6n and Zollikofer, 1999; Quam and Rak. 2008). The study of the ear ossi- cles in LF3, then, provides an opportunity to confirm the limited observations made on these isolated specimens and promises to reveal new insights into the evolution of the auditory apparatus in Neandertals. Materials and methods Tke La F e m s i e 3 ear ossicles The three ear ossicles were recovered from the right tympanic cavity. Since the original publication, the LF3 ear ossicles have suffered some damage, which limits the information that can be obtained. The malleus appears to be complete, but is broken in two pieces, corresponding to the manubrium and the headlneck (Fig. 1). These have been glued together in the past, although they are F l y r e I It.' 11 9 :rill'#. It.' ' :c' ~ p p . ' ~ . . 'c' :~p l ' ' ' 8 ,% 8rnl I.,..,.'. ,,..'I' . . ' I\ ,l..:.'., p. h', l..,. 1:.l,..I,,.',\ I b.' ' :c' It',,.',\ : b.' n. 1h.r . . e . p :.I. >pp. \#:n 1.1) t ' . A ' % p#'I#.h' . l '\ I 'I:? . .. Ih' other three views show slightly different orientations of the bone. Note the curvature of the lateral surface of the manubrium. Scale bar = 5 mm. currently not in their proper anatomical position. In particular, the headlneck is glued to the superior border of the lateral process. rather than ioining with the manubrium lateralhi. This has the ef- - feet of artificially elongating the specimen and the typical angula- tion between the head and the manubrium is not accuratelv represented. Close comparison of the present-day specimen with the photograph published by Heim (1982) indicates that the specimen was already in its present state at the time of its publi- cation. This means that the metric dimensions published previ- ously by Heim and others need to be revised. ~ximination under a standard light microscooe reveals that both the tio of the ma- - nubrium and the lateral process are complete and undamaged. However, the top of the head shows some erosion of the surface bone near the margin of the articular facet. The incorrect ori- entation of the two main pieces and the thick coating of adhesive prevent observation of a few anatomical features. Although the true maximum length cannot currentlv be deter- . . mined, the maximum length of the bone as it is presently (incor- rectlvl reconstructed was measured in several sliehtlv different . , orientations of the bone (Fig. l), including the view that appears to correspond to that by Heim. h he highest valueobtained in anv of these views was 8.74 mm 18.54-8.74 mm range). This - . value is larger than that reported by Heim (8.3 mm), but shorter than that measured by Spoor (2002) (9.0 mm) and Crevecoeur (2007) (8.9 mm) on the published photograph. Although all of these estimates are unreliable since they were taken on the incorrectly reconstructed malleus, the true maximum length is unlikely to be as long as 9.0 mm, as suggested by Spoor (2002). Given the preservation conditions, every attempt was made to measure as many standard dimensions as possible, paying close attention to preservation of the bone and identification of reliable anatomical landmarks. Despite these preservation issues, mea- surements of the manubrium and headlneck in the present study should be considered reliable. However, the angulation between the manubrium and headlneck as well as the corpus length and the total length of the bone cannot be accurately assessed. The published photograph of the stapes reveals that this bone was complete and intact at the time of publication (Heim. 1982). Currently, the bone is nearly complete, but broken in four pieces and covered with adhesive (Fig. 2). Given their extremely fragile state, no attempt was made to separate the pieces to reconstruct the bone more accurately. The largest piece corresponds to the intact head and includes approximately the superior one-third of the anterior crus. A second piece comprises the superior two-thirds piece corresponds to the inferior one-third of the posterior crus and a portion of the footplate. The footplate itself appears to be missing a vortion of the bone in the middle section. The four vieces were glued together in their approximate relative positions, but do not articulate directly with one another, making the correct anatomy difficult to discern. The only reliable measurement that can be taken on this bone in its current state of vreservation is the head height. Heim 119821 provides values for the total height of the stapes and the foot- plate length, measured when the bone was complete. In addition. Masali et al. 119911 provide the lengths of the anterior and vosterior . . - crura, as well as the area of the stapes footplate. We have relied on two sources to evaluate the accuracy of the stapes height and the lengths of the anterior and posterior crura. We have measured the total stapes height and the lengths of the crura in the published scaled photograph of the original specimen. finding a close correspondence with the values reported by Heim (1982) and Masali et al. (1991). In addition. we have estimated the total stapes height and the anterior crus length in the scaled photograph taken for the present study, even though the bone is damaged Again, a fairly close correspondence with~the published values is found, suggesting that these measurements are accurate. -- - The posterior crus length could not be evaluated in the present state of preservation of the stapes, but given good correspondence with the total height and the anterior crus length, we have accepted this value to be accurate as well. The footplate length and area cannot currently be measured in the LR stapes since the footplate is damaged. However, the foot- plate dimensions can be reliably estimated from the oval window visible on the medial wall of the tympanic cavity. The footplate Figure2. The LF3 rtaper. Despite some damage, the asymmetrical configuration of the crura and anteriorly rkewed orientation of the head can be appreciated. Scale bar = 2 mm. length provided by Heim (Heim. 1982) (2.7 mm) is considerably greater than the maximum length of the oval window measured in the present study (2.26 mm). Since the stapes footplate inserts into theoval window, its dimensions should be slightly smaller than that of the oval window. Measurement of these two varameters in a recent human sample (see below) suggests that the linear di- mensions of the stapes footplate are about 95% those of the oval window. This would predict a footplate length of 2.15 mm for LF3. considerably smaller than the 2.7 mm estimate of Heim (1982). Similarly, the value for the area of the footplate provided by Masali et al. (1991) (3.3 mm2) is considerably larger than that estimated from the size of the oval window. The measured area of the oval window in LR in the present study is 2.70 mm2. Com- parison of the size of the oval window with the stapes footplate in a stapes footplate area of 2.43 mm2 for LF3, considerably smaller than the 3.3 mm2 estimate of Masali et al. (1991). Taken together, the oval window dimensions clearly indicate a smaller stapes footplate than the values published by both Heim and Masali. Importantly, this revised estimate of footplate area in LR is very similar to the values in two other individuals from the same site. In contrast to the malleus and stapes, the incus is complete and well preserved (Fig. 3). Microscopic examination reveals the tip of the long process is intact, and this bone presents no complications with assessine its moroholoev and metric dimensions. Table 1 . provides the final list of measurements that should be considered ieliable on all of the LF3 ossicles. The finalvalues are based on those measured in the vresent studv as well as a careful consideration of previously published data, including that of Heim (1982). Comparative sample The early formation of the ear ossicles means that the adult dimensions are already reached early in ontogeny, and the di- mensions of LR can thus be compared with both juvenile and adult specimens. The comparative samples (Table 2) comprise the vast maioritv of Neandertal and fossil H. saviens ear ossicles known to . . date. In most cases the original specimens were studied, and this was complemented with a careful selection of metric data drawn from the literature to ensure that all measurements conform to the protocol outlined below. Many of these specimens were used in a orevious studv where details of the data collection are orovided (Quam and Rak 2008). In addition, the original ear ossicles of the Middle Paleolithic H. inpiens individual from Darra-i-Kur and the stapes of the Subalyuk 2 Neandertal were recently studied, and the Figure 3. The LF3 incur in (from left to right) lateral, anterior and medial views. Note thecloredangle between the processes, adeprerrion on the medialrurfaceofthe body and the presence of a notch in the inferior margin of the short process. Scale bar = 5 mm. data for these individuals are included in the present study. A number of fossil soecimens do not oreserve ear ossicles. but the oval window can be measured to provide an estimate of the size of the stapes footplate. We have also studied a sample of recent hu- man ear ossicles removed during cadaveric dissection in gross - - anatomy instruction at the New YorkChiropractic College in Seneca Falls. New York (USA). This samole is comorised of individuals of ~, known sex and race. ranging in age from 51 to 100 vears old. The individuals included in the present study were selected based on a criterion of maximizing the number of individuals that oreserve all three bones. Morphological fenhires The expression of several morphological features in the ear os- sicles was assessed in LF3 and the comoarative samoles. Several features described by Heim (1982) in the LF3 ossicles have been cited by subsequent authors as potential derived Neandertal fea- tures (Maureille. 2002; Crevecoeur, 2007; Quam and Rak. 2008). Among these are a relatively straight malleus manubrium and incus long process, a flatter incus body, asymmetrical stapedial crura and an anteriorly skewed stapedial head. Importantly, all of these fea- tures can be assessed in-the LF3 ossiclis in their current state of oreservation. Although the functional significance of manv of these - - features (if any) is not clear, the tight genetic control over the development of the ear ossicles means that variation in the expression of these features can generallv be taken to reflect ge- . . netic differentiation between populations, the soundest basis for drawing phylogenetic inferences (Lieberman. 1995). Measurement definitions and protocol Numerous previous studies on the auditory ossicles have relied Eouchet and Giraud. 1968; Elumer et al.. 1982; Mutaw, 1986; Siori et al.. 1995). Desoite some variation in methodolow. the reoorted , . U<. mean sizes of the ossicles are generallv similar across studies . (Arensburg et al.. 1981). suggesting that measurement error. when controlled for. is not a significant source ofvariation in the reoorted dimensions of the auditory ossicles. The two most influential - et al. (1981). and several subsequent studies of fossil hominin ear ossicles have followed one or the other of these orotocols iSooor. ~. . 2002; Lisonek and Trinkaus. 2006; Quam. 2006; Crevecoeur. 2007; Quam and Rak. 2008). The original LF3 specimens were studied by one of us (RMQ). The specimens were examined under a standard light microscope to evaluate the present state of preservation and to aid in assessing some morphological features. For measurement purposes, scaled photographs were taken of the ossicles using a digital camera and ~ - measured on a computer using the PhotoshopTM software program. The malleus was oositioned so that the manubrium was olaced flat on the surface of the table with the posterior aspect of the bone iexoosine the articular facet) facine the camera lens. The incus was ~. positioned so that the long and short processes were flat on the surface of the table with thelateral aspe;t of the bone (exposing the lower-most margin of the articular facet) facine the camera lens. The stapes was placed flat on the surface of the table with the su- perior aspect of the bone facing the camera lens. In this position. the superiormost margin of the obturator foramen is exposed. For calibration purposes, a 10 mm scale was placed next to the bone on thesurfaceof the table. For those specimens that preserved a stapes footplate, the bone was inverted so the footplate was facing the camera lens. The long axis of the footplate was oriented perpen- dicular to that of the camera lens, resulting in the maximum exposure of the footplate, and a 10 mm scale was placed at the same distance from the camera as the stapes footplate. With a few exceptions, we have generally relied on the mea- surements defined by Masali (1964). A detailed account of the measurement definitions for the malleus and incus iincludine some inconsistencies in measurement definitions in previous studies) has been published previously (Quam and Rak. 2008) and is fol- lowed here. The current state of preservation of the L R stapes limits the metric data that can be collected on this specimen. The measurement definitions for the stapes variables are provided here and in Fie. 4. The total staoes heieht is the maximum heieht from the lower margin of the footplate to the tip of the head, taken perpendicular to the long axis of the footplite. The length of the anterior crus is the maximum distance from the antero-suoerior corner of the footplate to the tip of the head. The length of the posterior crus is the maximum distance from the postero-superior corner of the footplate to the tip of the head. The head height is the . minimum distance between the superior margin of the obturator foramen and the too of the head. Footolate area is the directlv measured area of the stapes footplate. The oval window area was measured following similar pro- cedures as for the auditory ossicles. The oval window was oriented so that the maximum area was exposed and the long axis of the oval window was oeroendicular to that of the camera lens. A scale . . was included in the digital photo for calibration purposes and was placed at the same distance from the camera lens as the oval window itself. The photos were then transferred to the computer and measured using the PhotoshopTM software program. A similar Estimating stapes footplate area The footplate area in the LR stapes was estimated from the dimensions of the preserved oval window on the medial wall of the tympanic cavity. This approach has been used previously in other studies [Martinez et al.. 2004; Coleman and Colbert. 2010). and a close correspondence (3 = 0.961) has been found between the size of these two structures across primates (Coleman and Ross. 2004) since the stapes footplate inserts into the oval window. The stapes footplate is slightly smaller than the oval window since it is held in place by the annular ligament, which attaches to the Height 4 Total Height Stapes Footplate Figure 4. Definitions of the rtaper measurements. See tent for additional information. circumference of the footplate. To account for this size discrepancy, The lateral margin of the manubrium is curved, and shows a later- we have measured the staoes footolate area and oval window area allv inflected tio. In contrast. the medial marein is relativelvstraieht in a sample of 29 neonatal individuals of known sex and race, all and the manubrium increases in thickness from the tip to the neck. aged between six months intrauterine and 14 months postnatal. The lateral process is well-developed and protrudeslaterally. The deriving from the Johns Hopkins Fetal Collection housed at the medial surface of the malleus head shows some erosion of the bone Cleveland Museum of Natural History (USA). Importantly, both di- near the margin of the articular facet. This could be attributed to an mensions were measured in each individual. The leneth and width erosive eoisode of otitis media. but could eauallv be taohonomic in of the stapes footplate was found to be approximatelv 95% of the value of the length and width of the ~ v ~ l w i n d o w . ~ imi l a r l~ , the measured area of the staoes footolate was found to corresoond to approximately 90% of the value of the area of the oval window. A Atapuerca (Martinez et al.. 2004). Statistical analysis Statistical analysis of the ossicular variables includes univariate and multivariate approaches. In addition to descriptive statistics. . nature. The head shows a prominent protrusion antero-superiorly. opposite the articular facet. Just below this is a depressed area filled with sediment, on the anterior aspect of the head and neck. The articular facet itself is separated from the malleus head by a groove. The groove on the lateral neck described in Qafzeh 11 does not appear to be present in LF3. It is not possible to discern whether a prominent crest was present on the posterior neck since the specimen is brokenjust at the level where this crest would be. Nor is it possible to tell whether an anterior (gracile) process was pre- sent in the current state of preservation. Heim (1982) described the L R malleus manubrium as relatively straight with a well-developed lateral process. The description of . . . correlations between ossicular variables within the recent human the manubrium as straight seems to be an inference drawn largely samole were exolored and significant correlations io < 0.051 are from the oublished ohoto of the incorrectlv reconstructed soeci- discussed when relevant to interpreting the fossil specimens. Cor- relations above r > 0.7 were considered high. Subsequently, prin- cioal comoonents analvsis IPCAl was undertaken for the incus and . . stapes. he incus is the only 'bone that is complete and well- preserved, and all relevant dimensions could be used. Although the stapes is damaged, the few preserved measurements do largely capture the overall dimensions of this bone. In contrast, the pre- served malleus measurements are largely restricted to the manu- brium, limiting the utilitv of a multivariate analvsis. PCA reduces . the number of variables to a smaller number of components whose constituent variables can then be analvzed. Bv default. PCA eener- ates the same number of components as variables in the analysis. but only those components that yield eigenvalues > 1.0 can be considered to exolain more variation than the individual variables in isolation (Kachigan. 1991). Factor loadings of >0.7 for the indi- vidual variables. indicatine their correlation with the orincioal component, were generally considered high. The La Ferrassie 3 ear ossicles Malleus (Fig. l ) The description of the malleus morphology refers to the bone when oriented in anatomical position in the tympanic cavity, with the manubrium tio olaced inferiorlv, the manubrium itself laterallv. . . the head medially and superiorly and the articular facet posteriorly. Table 3 Malleus dimensions in the La Ferrarrie 3 Neandertal and Pleirtocene and recent human men. Examination of the original specimen makes it clear that the manubrium shows a curved lateral margin with a laterally inflected tip. This same morphology is found in 80% of fossil H. sapiens specimens where the manubrium tip is preserved (Quam and Rak. 2008). A curved manubrium is also the common condition in recent H. sapiens, being present in 66.2% of our recent human sample (Quam and Rak. 2008). The lateral process in LF3 is well-developed. but is clearly encompassed within the normal range of variation seen in both fossil and recent H. sapiens. Among the fossilH. sapiens specimens, only the Lagar Velho 1 individual apparently lacks a lateral process, and this is an uncommon variant in recent humans as well. The lateral process in the human malleus is embedded in the superior fibers of the tympanic membrane. However, it is not clear that degree of development of the lateral process in LF3 implies a greater protrusion of the tympanic mem- brane in Neandertals, as suggested by Heim (1982). Heim asserted that the L R malleus was slightly longer with a larger head and more open angle between the manubrium and headlneck than in H sapiens. The total length and angle of the axes cannot be assessed in the LF3 malleus. The manubrium length in LR (4.89 mm) is slightly shorter than the Middle Pleistocene AT- 3746 specimen but longer than all but one of the fossil H. sapiens individuals (Darra-i-Kur) (Table 3). The value in L R is, however, very close to the mean value in our recent H. sapiens sample. In contrast, the manubrium thickness (1.26 mm) is quite high, falling above the recent H sapiens range of variation and only being Specimen/rample Manubrium Manubrium Manubrium arc Head Manubrium length (mm) thickness (mm) depth (mm) width (mm) roburticiy index La Ferrarrie 3 4.89 1.26 0.44 2.92 25.8 AT1746 5.16 Biache-Saint-Vaart 1 >4lO 2.70 Qafzeh 11 2.41 Qafzeh 12 4.02 1.11 0.33 2.62 27.4 Qafzeh 15 4.48 1 0 2 0.28 2.44 22.3 Darra-i-Kur 5.02 1.34 0.40 2.50 26.7 Dolni Vertonice 14 4.60 0.87 0.14 2.50 18.9 Nazlet Khater 2 4.67 0.90 0.27 2.55 19.3 Lagar Velho 1 (left) 4.38 081 0.19 2.33 18.5 Fossil H. sopiens (mean & rd.) 4.53 10 .33 1.01 & 0.20 0.27 & 0.09 2.48 & 0.10 22.2 & 4.0 Forril H. ropiens range (n) 4.02-5.02 (6) 0.81-134 (6) 0.14-0.40 (6) 2.33-2.62 (7) 18.5-27.4 (6) Recent H. ropienr (mean 1 rd.) 4.94 1 0 3 1 100 & 0.09 0.33 & 0.15 2.43 & 0.17 20.3 & 1.9 Recent H ropienr range (n) 4.22-5.59 (43) 0.81-119 (43) 0.05-0.64 (43) 2.03-2.79 (43) 16.6-25.6 (43) surpassed by the value in Darra-i-Kur. The resulting robusticity index (25.8) is similarly high, again only being surpassed by two of the fossil H. sapiens individuals (Qafzeh 12 and Darra-i-Kur) and falling nearly 2 9 standard deviations (%d.) above the mean in our recent H. saviens samole. The arc deoth of the manubrium is a measure of the degree of curvature of the lateral margin. The value in LF3 (0.44 mm) is higher than all of the fossil H. sapiens individuals but is within 1 s.d. above the recent H. sapiens mean. indicating a curved lateral margin of the manubrium in LF3. The head width in L R (2.92 mm) is larger than the value provided by Heim (1982), falling above the head width in the Middle Pleisto- cene Eiache 1 specimen and outside the range of variation in both the fossil and recent H. saviens samoles, confirming that a large - - head size characterizes LF3. Among the few preserved malleus dimensions, the thickness and curvature (arc deothl of the manubrium were not correlated . . with its length within the recent H. sapiens sample. However, moderate positive ( r = 0.45) and negative ( r = -0.39) correlations were found between the manubrium length and head width and robusticity index, respectively. The robustiky index shows a much stronger correlation 1r = 0.771 with the manubrium thickness. indicating that the values for this index are driven largely by var- iation in thickness. Incus (Fig. 3) The description of the incus morphology refers to the bone when oriented in anatomical position in the tympanic cavity, with the tip of the long process placed inferiorly and angled posteriorly. the short process placed superiorly and protruding posteriorly, the articular facet facing anteriorly and the lowermost margin of the articular facet facing laterally to articulate with the malleus head. The L R incus is complete and well-oreserved. The contour of the superior margin of the short process is concave and the tip is thick suoero-inferiorlv. On the medial asoect of the short orocess. a clear notch is present along the lower margin. There is a depressed area in the central portion of the medialaspect of the body, just pos- terior to the articular facet. The articular facet itself is seoarated from the surrounding bone by a groove, particularly visible in lat- eral view. The articular facet shows a skewed orientation, with the Table 4 I ~ C U I dimensions in the La Ferrarrie 3 Neandertal and Pleirtocene and recent humans. Specimen/rample Short ~ o n g Functional Long process process process length arc depth length length (mm) (mm) (mm) (mm) La Ferrarrie 3 4.74 6.89 4.08 0.40 AT-3747 4.33 Biache-Saint-Vaart 1 7.50 Amud 7 5.07 6.98 4.08 0.28 Le Mourtier 1 6.80 Qafzeh 11 5.07 6.94 4.05 0.30 Qafzeh 12 4.70 6.73 3.84 0.43 Qafzeh 15 4.64 6.34 3.57 0.35 Qafzeh 21 4.55 6.47 3.93 0.44 Darra-i-Kur 4.78 6.75 4.46 0.63 Dolnivertonice 14 5.50 7.10 4.11 0.56 Dolni Vertonice 15 4.10 6.40 3.89 0.35 Lagapar Velho 1 4.76 FollilH.sopiem 4 7 6 & 0.40 6.68 & 0.29 3.98 & 0.27 0.44 & 0 1 2 (mean & rd.) Fossil H. sopiem 4.10-5.50 (8) 6.34-710 (7) 3.57-4.46 (7) 0.30-0.63 (7) range (") Recent H. ropienr 5.07 & 0.37 6.83 & 0.32 4.00 & 0.21 0.56 & 0.14 (mean & rd.) Recent H. soptens 4.02-5.86 (41) 6.17-7.59 (43) 3.61-4.46 (42) 0.25-0.80 (43) inferior margin being lower on the lateral aspect, where it articu- lates with the malleus head, than on the medial. The anterior margin of the long process is relatively straight, particularly in lateral view, but the tip does deflect medially for articulation with the staoes head. Heim (1982) suggested that the long process of the incus in L R is relativelvstraieht and thin when comoared with H. saoiens. while . . the short orocess does not show a notch in its lower border and the body is slightly flatter than in H. sapiens. The long process in L R does show averv straight anterior border (Fie. 31 and resembles the condition seen in both the Amud 7 and Le Moustier 1 Neandertals. Nevertheless, some variation among the fossil and recentH. sapiens samoles is also oresent. Both Oafzeh 11 and Dolni Vestonice 15 - show a straight long process and this condition is found in just under 25% of recent H. sapiens (Quam and Rak. 2008). The lower margin of the short process shows a clear notch in medial view (contra Heim). This is a well-known anatomical variant in the incudi of fossil and recent H. sapiens (Arensburg and Nathan. 1971. 1972), and it is also found in both the Amud 7 and Le Moustier 1 Neandertals (Quam and Rak 2008). The form of the short process tip is variable in fossil and recent H. sapiens, ranging from a very thick, bulbous appearance to a slender, tapering point. This range of variation is seen among the four incudi from Qafzeh (Quam and Rak 2008), but the tip of the short process in all three known Neandertal incudi is thick supero-inferiorly with a rounded pos- terior margin. The body of the L R incus also shows a deeply excavated depression on its medial surface (Fig. 3), resembling Amud 7 in this regard. The majority of fossil and recent H. sapiens specimens also show a depression on the medial surface of the incus body. In both Qafzeh 12 and 15 and 16% of recent H. sapiens. this depression is as deep as in LF3 (Quam and Rak. 2008). In general, the morphological details of the incus in Neandertals seem less variable than in H. sapiens. The length of the shir t process in L R (4.74 mm) is slightly shorter than in Amud 7 (5.07 mm) (Table 4). Nevertheless. LF3 falls within 1 s.d. below the recent H. sapiens mean and is nearly iden- tical to the mean value in fossil H. saoiens. The lone orocess leneth U . in LF3 16.89 mm) is clearlv shorter than that in Eiache 1. but falls between the values of t h e ~ m u d 7 and Le Moustier 1 Neandertals. However, the values in the three Neandertal individuals are similar Articular Inter-process Inter-process Angle lncudal facet length arc depth (mm) between index height (mm) (mm) the axes 3.34 5.51 1.43 48.2 68.8 Figure 6. The Subalyuk 2 rtaper. Despite some obvious rignr of pathology, the head (arrow) is clearly anteriorly rkewed, like in LF3 Scale bar = 2 mm. that calculated for the Middle Pleistocene specimens AT-374613747 (119.2). Although both of these specimens fall toward the upper end of the fossilH. sapiens range of variation, they are close to the recent H. sapiens mean (123.4). Thus, there appears to be little difference in the lever ratio between Neandertals and H. sapiens. Stapes (Fig. 2) The descrivtion of the staves morvhologv refers to the bone skewed head (contra Arensburg et al., 1996). In addition, asym- metrical crura are also said to characterize the stapes in the recently rediscovered Le Moustier 2 Neandertal infant (Maureille. 2002). While the stapedial head normally occupies a central po- sition (65.4%) in recent H. sapiens, resting atop both crura equally, an anterior orientation (32.7%) is a common variant (Quam. 2006). Thus, while some variation in the orientation of the stapedial head characterizes recent H sapiens, the Neandertals show a consistent anatomical pattern. The height of the LF3 stapes (3.00 mm) is slightly taller than that estimated for the Subalyuk 2 Neandertal but is shorter than the Darra-i-Kur specimen and falls more than 2 s.d. below the recent H. sapiens mean value (Table 7). Thus. Heim's (1982) assertion that the Neandertals have a slightly smaller stapes would appear to be correct. The head height in LR is similar to that in Subalvuk 2 but again falls toward the lower end of the - recent H. sapiens range of variation, and the relative head height in all of the fossil soecimens falls below the recent H. saoiens mean. Both Neandertal individuals have an anterior crus that is much shorter than the posterior crus, while in recent H. sapiens the lengths of the crura are much more similar. The resulting crural index in both Neandertal individuals falls below the recent H. sapiens range of variation, consistent with the highly asym- metrical configuration of the crura in Neandertals. - The stapes footplate in LR is damaged and cannot be measured directlv. However the area of the footolate can be estimated from the preserved oval window on the medial wall of the tympanic cavity. The size of the oval window in LF3 (2.70 mm2) (Table 8 ) is consistent with an estimated stapes footplate area of 2.43 mm2. This value is identical or vew close to that measured in two other -. when oriented in anatomical position in the tympanic cavity, with specimens from La Ferrassie and is within 1 s.d. of the Neandertal the head olaced laterallv to articulate with the tio of the incus lone mean. It is sliehtlv smaller than the three Middle Pleistocene Ata- crus, the crura located anteriorlv and ~osteriorlv and the footplate puerca (SH) individuals but well below the lower limit of the range located medially with its long axis oriented antero-posteriorly. The ofvariation in fossil and recent H. sapiens.The oval window index in current state of vreservation limits the morvholoeical observations LR 154.0). comvaring the length and width of the oval window, is that can be made on this specimen. Heim (1982) described the LF3 stapes as showing a pronounced asymmetry between the shorter anterior crus and the longer and more curved posterior crus. In addition, the head of the stapes was described as skewed anteriorly and the crura as thicker in LB. The footplate was described as 'reniforme', or kidney-shaped, like in H sapiens. The combination of a pronounced asymmetry between the anterior and posterior crura and an anteriorly skewed head are evident in the published photo of the complete stapes (Heim. 1982). In addition, both of these features can be appreciated in the LR stapes, even given its current state of preservation. The Subalyuk 2 Neandertal infant also preserves a partial stapes, and this specimen is described as showing asymmetrical crura [Arensburg e t al.. 1996: Pav et al.. 1996). Although the head on this . . . - - nearly identical to the Neandertal mean (54.7) but is higher than all but two of the fossil H. sapiens individuals and is more than 2 s.d. above the recent H. sapiens mean based on the stapes footplate. Thus, although the stapes footplate is 'kidney-shaped', like in H. sapiens. Neandertal specimens show a slightly wider oval win- dow, and hence stapes footplate. The total stapes height is strongly correlated with the lengths of the anterior ( r = 0.83) and posterior ( r = 0.84) crura. While a moderate correlation is also found with head height ( r = 0.59), no correlation with the relative head height was found. Thus, relative head height is driven primarily by variation in the height of the stapes head ( r = 0.91). The length and width of the footplate are only weakly correlated with one another ( r = 0.39) but show strong correlations 10.73 < r < 0.83. resvectivelv) with the measured ~pecimen-shows some s&ns of pathology, examination of the footplate area. The footplate index is only weakly correlated with original fossil (Fig. 6) reveals that it clearly shows an anteriorly the measured area (r=0.37), andvariation in this indexseems to be Table 7 Staper dimensions in the La Ferrarrie 3 Neandertal and Pleirtocene and recent humans. Specimen/rample Total Head Anterior crur Posterior crur Crural Relative head height (mm) height (mm) length (mm) length (mm) indexd heightb La Ferrarrie 3 3.00 0.92 2.70 3.40 79.4 30.7 Subalyuk 2 2.70 est. 0.93 2.45 est. 2.90 ert 84.5 ert 34.4 ert Darra-i-Kur 3.13 0.83 2.82 3.17 89.0 26.5 Recent H. ropienr 3.44 & 0.20 1.21 & 0.16 3.40 & 021 3.35 & 0.20 1016 & 5.6 35.2 & 3.9 (mean & rd.) Recent H. ropienr 2.94-3.87 (40) 0.87-152 (40) 2.94-3.93 (40) 2.77-3.72 (40) 884-1088 (40) 26.3-43.5 (40) rrlnce I n > a crural index = (anterior crur lengthlporterior crur length) x 100. Relative head height = (Head heightltotal height) x 100. Table 8 Comvariron of oval window dimensions and footvlate area in La Ferrarrie 3 with Pleirtocene and recent humans Oval Oval Oval Measured Estimated Footplate Footplate Footplate Measured window window window oval footplate length width index footplate length width index window aread (mm) (mm) area (mm) (mm) area (mm" [m2) (mm2) La Ferrarrie 3 2.26 1.22 54.0 2.70 2.43 Neandertal~ La Ferrarsie 4b La Ferrarsie 5 Arcy IUT Cure Keh2r2 l ~ Neandertal mean f s.d. Neandertal range (m) Fossil H. sopiem Qafzeh 3 Qafzeh 11 Qafzeh 12 Qafzeh 13 Qafzeh 21 Darra-i-Kur Skhul 1 Border Cave 3 Cm Magnon 2 Parpallo 1 Fossil H. s o p i m mean f s.d. Fossil H. s o p i m 2.60-3.27 (9) 1.20-1.61 (9) 36.7-55.8 (9) 3.06-3.88 (9) 2.75-3.49 (9) range @l Recent H ropienr 294* 0.14 1.39&0.10 47.3 &3.3 3.36&0.33 mean* rd . Recent H ropienr 2.47-3.27 (46) 1.20-163 (43) 40.3-571 (43) 2.95-4.29 (41) range (n) Estimated footplate area = (measured oval window area) x 0.9 largely a product ofvariation in the width ( r = 0.78), rather than the footplate. Similarities with H sapiens include the curved lateral length, of the footplate. margin and absolute length of the malleus manubrium, the pro- Principal components analysis was carried out on the four portions of the incudal long and short processes and the malleus/ preserved linear variables in LF3, and only the first component incus lever ratio. Although the polarity of many of these features is yielded an eigenvalue > 1.0 (Table 9). This component explains difficult to discern, a few generalizations do seem possible. 76.8% of the total variance (Fig. 7). and all of the variables showed The Stw 151 stapes attributed to Aushalopithecus africanus high negative loadings. This component is clearlv size-related, with shows a centrallv placed staoedial head and largelv svmmetrical . . . . . . . higher scores being ;ssociated with smaller stapis dimensions. The configuration of the anterior and posterior crura (Moggi-Cecchi and two Neandertal soecimens fall towards the uooer end ILR) or Collard. 2002). Thus. the Neandertal condition seems to be derived. . . , , outside (Subalyuk 2) the modern human range of variation along the first component (Fig. 7), while the single fossil H. sapiens in- dividual. Darra-i-Kur, also falls within but toward the uooer end of the recent H. sapiens range of variation. Discussion Phylogenetic polarity of ear ossicle morphologv The study of the original LR ear ossicles has revealed a number of differences from H. saoiens in the ossicular chain. It is not oos- sible to confirm based on LF3 whether Neandertals had an abso- lutely longer malleus and more open angle between the manubrium and corpus, as suggested by Heim (1982). Never- theless. Neandertals can be distinguished from H. sapiens in showing a larger mallear head, taller incudal articular facet. deoressed area on the medial surface of the incudal bodv, straight - incudal long process, more closed angle between the incudal pro- ~ ~ cesses, smaller stapes with asymmetrical crura, an anteriorly skewed stapedial head and a smaller and relatively wider stapedial even though this same condition can be found as an uncommon variant among recent H. sapiens. At the same time, the mean footplate area in Neandertals (2.58 mm2) is only slightly larger than that reported for A. africanus (2.33 mm2) (Moggi-Cecchi and Collard. 2002) despite considerable differences in body size, sug- gesting that this may represent a primitive retention in Neandertals. The human-like lever ratio in LF3 appears to be a derived con- dition that is shared with H. sapiens. Most primate species have Table 9 Results of the orincioal comoonentr analvrir of rtaoer variables. Factor 1 Factor 2 Eigenvalue 3.07 0.48 %Total variance 76.8% 11.9% Variable factor loadiner Head height 0 8 0 1 0.509 Total height 0 9 6 1 0 . 2 0 5 Posterior crur length 0 . 8 5 8 0.152 Anterior crur length 0 . 8 7 7 0 . 3 8 9 Figure 7. Scatterplot distribution of the firrt and second principal components bared on the rtaper variables. me firrt component largely reflects rile differences, with t..tc'. i l l 8 % ' . '.'I:. 1.. n '.I t .:nllc'. . >pc'. It.' ,% \C'>: 1%'. 11 .pc' :w:. n l l ,%>..I tc' #ppc'. .':.I . 1 ' .c' t ~ ' . ~ ' ~ ' : t8.n: r i3.n. : 11 :. tc'r . . :?p ..c'.' .C ' I~ ' I:. t ~ ' . .:rill . I~~'..II:~C':.. :. higher lever ratios (Coleman and Colbert. 2010), including chim- panzees (Martinez et al.. 2004). The presence of a human-like lever ratio in LF3 and the Middle Pleistocene Atapuerca (SH) individual sueeests this is a shared derived condition inherited from the last common ancestor of Neandertals and H. sapiens. Although many of the features mentioned above are variably present in fossil and recent H. sapiens, the Neandertals seem to show a nearly constant expression of these features. Thus, they consistently express a portion of the modern human range of var- iation and show less anatomical and metric variation than H. sapiens. This is true for the incus and stapes, but is less clear for the malleus since the LF3 specimen is the only Neandertal malleus published to date. Given the tight genetic control over the mor- . . phology and development of the ear ossicles (Mallo. 2001), this lower level of variation in Neandertals is consistent with a lower genetic diversity in Neandertals compared with H. sapiens, as suggested by ancient DNA analyses (Eriggs et al.. 2009). Body size and ear ossicle dimensions Given the extremelv small dimensions of the ossicles and the differences in body mass between Neandertals and H. sapiens, the oossible relationshio of the ossicular dimensions to asoects of bodv size should be considered. Although bodv mass estimates in . fossil hominins involve a number of complicating factors, the emergence of H. saoiens seems to have coincided with a reduction in body mass compared with their Pleistocene precursors (Ruff, 2002). The results of the present study suggest that the malleus head (and, by implication the articular facet) and incus articular facet are larger in Neandertals than H. sapiens. Although correla- tion of the incudomalleolar joint dimensions with body mass has not been investigated, ioint size in the mammalian postcranial skeleton refleits differences in body size (Godfrey et al.. 1995: Lieberman et al.. 2001). Thus. the lareer incudomalleolar jointin Neandertals might betaken to refleccdifferences in body mass. In general, the ossicle dimensions do not appear to be related to the size of the skull in modern humans (Arensburg et al.. 1981). but the lever ratio shows a weak negative correlation with skull size across primates (Coleman and Ross. 2004). Cranial dimensions do differ between Neandertals and H. sapiens, with the former show- ing, on average, longer, lower crania and somewhat larger brain sizes (Arsuaga e t al.. 1997). Nevertheless. the lever ratio in LF3 is close to the modern human mean, and the dimensions of the sta- pes, including the size of the footplate, are smaller in Neandertals. Thus, the relatively minor differences in absolute cranial size do not seem to be resoonsible for the differences in ossicular dimensions reported here. More broadly, the mass of the ossicles is negatively allometric with skull mass iNummela. 1995). and the staoes footplate area is correlated with body mass across mammals (Rosowski. 1992). The comparison of stapes footplate size between Neandertals and modern humans is revealing in this regard. Neandertals are char- acterized by a smaller stapes footplate area and a heavier body mass. In contrast. H. sapiens shows an increase in the stapes foot- plate area and a decrease in bodv mass, and seems derived in both these aspects. Interestingly, the change in stapes footplate size in H. saoiens is in the oooosite direction of that oredicted based on . . body mass, and represents a departure from the general mamma- lian pattern. The mechanism behind this increase in stapes foot- plate size in H. sapiens is currently not clear, and further study of this question is clearly warranted. Articulation of the ossicular chain Several aspects of the Neandertal ear ossicles suggest subtle differences fromH. sapiens in the articulation of the ossicular chain. Although the larger incudomalleolar joint in Neandertals may reflect differences in bodv size. ioint surface area is also correlated with the transmission of forces across the loint IGodfrev et al.. 1995). Thus. the larger incudomalleolar jointsurfaces in ~ e a n d e r - tals are consistent with the caoacitv for transmitting lareer forces . . through the ossicular chain. More importantly, the more closed angle between the pro- cesses of the incus and the anteriorlv skewed stapedial head clearly suggest a slightly different spatial relationship and artic- ulation within the tvmoanic cavitv in Neandertals comoared with G . H. sapiens. Although the fossil record is limited, these are constant feat"res in the few Neandertal fossils known. This arrangement sueeests either a more posterior placement of the oval window (for the insertion of the stapes footplate) on the medial wall of the tympanic cavity or a more anterior position and101 change in the angle of the tvmpanic membrane Ifor the insertion of the - . . malleus manubrium) at the medial end of the external auditory canal. The inner ear, like the ear ossicles themselves, is under a vew tight genetic control, being fullv formed and adult size at . - - - . birth and is a highly conservative region evolutionarily ueffery and Sooor. 2004). In addition. the tieht relationshio between . . the osseous structures of the stapes footplate and oval window would seem to be more resistant to changes in their spatial arrangement within the tympanic cavity than the fibrous and elastic structure of the tympanic membrane, for insertion of the malleus manubrium. Perhaps relevant in this regard, an embryological link between the malleus and the temporo-mandibular joint (TM]) can be seen in the discomalleolar ligament (Rodriguez-Vizquez et al.. 1993). This ligament passes anteriorly from the base of the manubrium through the glasserian fissure to insert onto the medial aspect of the cartilaginous articular capsule of the TM] (Rodriguez-Vizquez et al.. 1998). Tension applied to the discomalleolar ligament has been shown to produce movement of the malleus, suggesting a functional interdependence of the malleus and the TM]. Nean- dertals have been shown to display a suite of derived features in both the TM] and the mandibular condyle (Quam and Smith. 1998; Jabbour e t al., 2002; Rak et al.. 2002; Martinez et al.. 2008), raising the possibility of a pleiotropic effect on the ossicular chain, as has been suggested previously in early hominins (Rak. 1994). Nevertheless, further study of the Neandertal tympanic cavity and middle ear ossicles are clearly necessary to confirm these hypotheses. Auditory implications Although the relationship between individual auditory struc- tures and hearing performance is complex (Quam. 2006; Coleman and Colbert. 2010; Quam et al.. 2012), the functional lengths of the malleus and incus and the stapes footplate area are important physiological variables in modeling audition (Rosowski. 1996; Aibara et al.. 2001; Martinez et al.. 2004.. in press). These three variables are all strongly correlated with low frequency sensitivity (i.e. < l kHz) across primates. In addition, the incus functional length has been shown to be correlated with the presence of a notch (indicating decreased auditory sensitivity) in the audio- gram around 4 kHz across primates (Coleman and Colbert. 2010). Thus, taxa with shorter incudal functional lengths show a more pronounced notch (and hence greater decrease in auditow sensi- . tivity) in the audiogram. Compared with chimpanzees, humans are characterized bv a considerable increase in the incudal functional length (Masali and Chiarelli. 1965; Masali et al.. 1991; Quam. 2006). andthe absence of a notch is one of the most salient distinctions in the hearine oattern of modern humans. one that differentiates us -. fromnearly all other primates tested to date (Coleman. 2009; Quam e t al.. 2012). The of a human-like incus functional length and lever arm ratio is consistent with a hearing pattern in Nean- -. dertals that is similar to that of living humans in showing a broad- ened region of heightened sensitivity in the midrange frequencies. This suggestion is further supported by the analysis of the auditory capacities in the Middle Pleistocene Atapuerca (SH) hominins, since a close phylogenetic relationship is widely recognized between these two groups of hominins (Arsuaga et al.. 1997; Hublin. 2009). The Atapuerca (SH) hominins have been reconstructed to show a hearing pattern that resembles that of modern humans very closely (Martinez et al.. 2004.. in press; Quam et al.. 2012), sug- gesting this is a shared derived feature inherited from the last common ancestor of Neandertals and H. saviens. Given the ana- tomical distinctions in the Neandertal ossicular chain reported here, analysis of the auditory capacities in Neandertals should be a future research priority Conclusion The present study of the LF3 ossicular chain has clarified a number of aspects regarding the anatomy and metric dimensions of the Neandertal middle ear ossicles. Several previous assertions regarding the LE4 ear ossicles have been revised (e.g.. presence of a straight malleus manubrium, differences in the lengths of the incus processes, size of the stapes footplate). At the same time, we have also confirmed several observations made by Heim (1982) in his original study (e.g.. presence of a large malleus head and incus articular facet, straight anterior margin of the incus long process. asvmmetrical stapes crura and an anteriorlv skewed stapes head). In addition, the present results have confirmed several suggestions made oreviouslv reeardine the Neandertal ear ossicles based on more limited evidence (Quam and Rak. 2008). Far from showing no difference from H. sapiens (Arensburg et al.. 1981. 1996). the ear ossicles have revealed subtle yet real distinctions in the ossicular chain between these two Pleistocene taxa, suggesting that the auditory ossicles are an important source of phylogenetic and paleobiological information. Acknowledgments We wish to thank the following institutions for access to fossil hominin and recent human specimens included in the present studv: Musee de I'Homme (Paris. France), lnstitut de Paleontoloeie - Humaine (Paris. France). New York Chiropractic College (Seneca Falls. USA). Cleveland Museum of Natural History (Cleveland. USA). Tel Aviv Universitv (Tel Aviv, Israel), lnstituto Portugu@s de . . Arqueologia (Lisbon. Portugal). ~ a t i i n a l Museum o f -~a tu r a l Histow-Smithsonian Institution iwashineton. DC. USA). Universitv . . ,. of Witwatersrand Uohannesburg. South Africa). Ditsong National Museum of Natural History (Pretoria. South Africa) and the Museo de Prehistoria v Etnoloeia ivalencia. Soainl. We also thank T. . , Greiner for access to the recent human sample. This research was funded by the Leakey Foundation, t h e ~ m e r i c m Museum of Natural Histow and the Ministerio de Ciencia e InnovaciCln of the Govern- ment i f Spain. Project No. CGL2009-12703-C03-03/02 References Aibara, R , Welsh, J., Puria, S., Goode, R, 2001. Human middle-ear round oanrfer function and cochlear inout imoedance. Hearine Rer. 152.100-109. Aneel, IL, 1972. A Middle ~a i eo l i t h i temvoral bondfrom Darra-i-Kur, Afehanirtan. . . TranS. Am. Phi1 Soc. 62, 54-56. Arenrburg, B., Harrell, M , Nathan, H., 1981. The human middle ear orricler: mor~hometrv. and taxonomic imolicationr. 1. Hum. Evol 10. 199-205. *. Arenrbure, B., Nathan, H., 1971. Obrenrationr on a notch in the short frmerior or posterior) process of the incur. Acta Anat. 78, 84-90. \ IC~.,# r . l) u?lt.?n I1 l., > \ p r p . 1% 1 . 1 \ ...1.1. 1% I r c# l l c n. ycnnc 1 8 n " " l 1 l ,, c ' . , t . I l t . , p l 8 I 3 I 3 , \rcn.,# r . h 1'10 I l l . . C t. l I.,.,< I t . 8 l k > . . I l l . ? r rtaper. 1nt J. Orteoarcheol. 6, 185-188. Arruaga, J L , Martinez, I., Gracia, A., Lorenzo, C., 1997. The Sima de lor Hueror crania (Sierra de Atapuerca, Spain). A comparative study. J. Hum. Evol. 33, 219-282. Blumer, W., Freedman, L, Lofgren, M , 1982. Middle ear orricler of Australian ab- origines. Archaeol. Oceania 17, 127-131. Bouchet, A., Giraud, M , 1968. Contribution a I'etude morphologique et radiologique der orreletr de I'ouie. CR. Arroc. Anat 53, 588-600. Briggr, A., Good, J., Green, R., Kraure, J., Maricic, T., Stenzel, U, Lalueza-Fau, C., Rudan, P, Brdkovii, D., Kuian, 2 , Guiii, I., Schmin, R , Doronichev, V., Golovanova, L, de la Rarilla, M , Fortea, J., Rorar, A., PBBbo, S., 2009. Targeted retrieval and analysis of five Neandertal mmNA genomer. Science 325. 318-321. Clark, A., Glanowrki, S., Nielren, R , Thomar, P, Keiariwal, A., Todd, M , I,".">,l n. I1 .I\. I1 I1 1 1 I l p t . l l > " .. /h". \ t . I \n#n.ky I \ I?n. . \I l \I > 3 Inr~rr8n. n n n l o l n l . \ . I I , n i r 0. t.h".," t . , " " " . I . . < , t . l . . . " C , C . . > , C " C l > 1960-1963. Coleman, M.N., 2009. What do primates hear? A men-analysis of all known non- human primate behavioral audiograms Int. J. Primatol. 30, 55-91. Coleman, M.N., Colbert MW., 2010. Correlationr between auditory rtructurer and hearing renritiviy in non-human primates. J. Morphol. 271, 511-532. Coleman, M.N., Rorr, C., 2 0 M Primate auditory diverriy and itr influence on hearing performance. Anat Rec. Part. A 281A, 117,-1137. Crevecoeur, I., 2007. New discovery of an Upper Paleolithic auditory orricle: the right malleur of Nazlet Khater 2. J. Hum. Evol 52, 341-345. Frenz, D., McPhee, J., van de Water, T., 2001. Souctural and functional development of the ear. In: Jahn, A., Santor-Sacchi, J. (Edr.), Physiology of the Ear. Singular, San Diego, pp. 191-214 (second ed.). Godfrey, L, Sutherland, M , Paine, R, William, F., Boy, D., Vuillaume- Randriamanantena. M , 1995. Limb joint surface areas and their ratios in Mal- agasy lemurs and other mammalr. Am. J. Phyr Anthropol. 97, 11-36. Heim, J L , 1982. Ler Enfantr Neandertalienr de La Ferrarrie. Marron, Paris. Heron. I.. 1923 Measurements and observations uoon the human auditonr orricler. . . Am. J. Phyr Anthropol. 6,11-26. Hublin, J., 2009. The origin of Neandertalr. Proc. Natl. Acad Sci. 106, 16022-16027. Jabbour R , Richardr, G., Anderron, J., 2002. Mandibular condyle oaitr in Nean- derthal~ and other Homo: a comoarative. correlative. and ontoeenetic stud". Am. 1. Phvr Anthrovol. 119,344-355 Kirikae, 1, 1960. The Structure and Function of the Middle Ear Univerriy of Tokyo Prerr, Tokyo. Lieberman, D., 1995. Testing hypotheses about recent human evolution from rkullr: integrating morphology function, development and phylogeny. Curr Anthro- p01 36, 159-197. Lieberman, D.. Devlin. M , Pearron, 0 , 2 0 0 1 Articular area rerponrer to mechanical loading: effects of exercise, age, and skeletal location. Am. J. Phyr Anthropol. , ,G ,,X-,,, . . -, -. . . Lironek, P, Trinkaur, E., 2006. The auditory orricler. In: Trinkaur, E., Svoboda, J. (Edr.), Early Modern Human Evolution in Central Europe. The People of Dolni Vestonice and Pavlov. Oxford Univerriy Prerr, Oxford, pp. 153-155. Mallo, M , 1998. Embryological and genetic aspects of middle ear development. Int. J. Dev. Biol. 42, 11-22, Mallo, M , 2001. Formationof the middle ear: recent progress on the developmental and molecular mechanirmr. Dev. Biol. 231. 410-419. Martinez, I., Quam, R, Arruaga, J., 2008. Evolutionary trends in the temporal bone in the Neandertal lineage: a comparative study between the Sima de lor Hueror (Sierra de Atapuerca) and Krapina rampler. In: Monge, J. (Ed.), New lnrightr on the Kra~ina Neandertalr: 100 Years after Gorianovii-krambereer. Croatian Natural History Museum, Zagreb, pp. 75-80. \ l? \l \C .# ? L ? l l ? r ? > l' L# ?n. h lLrcn, . , c ? 8 , \ 111.1.1 1 1h.l": 11, 1% . 1.11 1 . > c > "Cl1 l > . 18 l., r, >p> # ,#C. 8 " \ l . l l l l 8 . " C t . . ,, 0. t . 8 c, , l \ " C , . 0 , " l U l l l Acad Scl. 101, 9976-9981 . - Marali,, M , 19% Dati rulla variabiliti morfometrica e ponderale degli orricini delludito nelluomo. Arch. ltal Anat. Embriol. 69, 435446. Marali, M , Chiarelli, B., 1965. Analiri morfometrica comparata degli orricini del- l'udito dei Primati 11. L'incudine nelle rcimmie del Vecchio Mondo e nell'uomo. Riv. Antropol. 52, 147-157. Marali, M , Maffei, M , Borgognini Tarli, S.M., 1991. Application of a morphometric model for the reconstruction of some functional characteristics of the external and middle ear in Circeo 1. In: Piperno, M , Scichilone, G. (Edr.), The Circeo 1 Neandertal Skull: Studies and Documentation. lnrtituto Polierafico e Zecca Mutaw R.. 1986. Human Auditory Orricle Variation and Function. Department of Anthropology Univerriy of Colorado, Boulder. Nummela, S., 1995. Scaling of the mammalian middle ear. Hearing Rer. 85, 18-30. Pao. I.. Tillier. A.. Arenrbure. B.. Chech. M. 1996. The Subalvuk Neanderthal remains .. . . . ". (Hungary): a re-examination. ~ n n l r . ~ i r t - "a t . Mur. Piat Hung. 88, 233-270. Ponce de Leiin, M.,Zollikofer, C., 1999. New evidence from Le Mourtier 1: computer ?....I% 11% n.lr# ( 8 n i n 1 n. rpt. n..lr\ rrt.. .k# l1 \m( U. >-. . . . r , I.# ln. U > I 1 % n p nlh n \ l t . l l n \ I . I # . ? p > # l # . 8n Quam, R , Martinez, I., Loremo, C., Bonmati,A., Rora, M.,Jarabo, P.,Arruaga,J., 2012. Studying Audition in Forril Homininr: a New Approach to the Evolution of Language? In: Jackron, M. (Ed.), Psychology of Language. Nova Science Pub- lishers, Inc., Hauppauge, NY pp. 47-95. Ouam. R. Rak. Y.. 2008. Auditor" orricler from southwest Arian Mourterian rites. J. Hum. Evol. 54,414-433. Quam, R, Smith, F.H., 1998. Rearrerrment of t h e n b u n C2 mandible. In: Akazawa,T. (Ed.), Neandertal~ and Modern Humans in Western Aria. Plenum Prerr, New York 00. 405-421. Rak, Y., 1994. The middle ear ofAurtrolopithecur robustus. Doer it bear evidence of a specialized masticatory system? In: Corruccini, R , Ciochon, R. (Edr.), lnte- grative Paths to the Part: Paleoanthropological Advancer in Honor of F. Clark Howell. Prentice Hall. Enelewood Cliffs. oo. 223-227. Rak, Y.. Ginzburg, A., Geffen,"~., 2002. ~o i r ' f i omo neondertholenrls play a role in modern humanancertry7The mandibular evidence. A m J Phyr Anthropol. 119, 399-204. RodriguezVizquez, J., M6rida-Velarco, J., Jim6nez-Collado, J., 1993. Relationrhipr between the temooromandibular ioint and the middle ear in human fetuses. I I I , ~ U . > l U 11.. c \11.1 , l \1'1111\111. 1 \1;1111\111. l I#".C'"C/ ( l l > l l l \ 0 . 8 l n . # I c n l # n. n rt.. I.. n.? l lc h r I # . ? n . ~ n l I \mr 1 . J I I <> l Rorowrki, J., 1992. Hearing in transitional mammalr: predictions from the middle ear anatomy and hearing capabilities of extant mammalr. In: Webrter D. (Ed.), The Evolutionary Biology of Hearing. SpringerVerlag, New York, pp. 615-632. Rorowrki, J., 1996. Models of external and middle-ear function. In: Hawkinr, H. (Ed.), Auditory Computation. Springer, New York, pp. 15-61. Ruff C., 2002. Variation in human body size and shape. A. Rev. Anthropol. 31, 211-232. Scheuer. L. Black. S.. 2000. Develoomental luvenile Orteolom Academic Prerr. \ I " I I.. \I I, \I \I n t.,.,, I \ l 1 l l l p t . " C . I t.h".," 11 11, r, .I 1 % . I, n. \"l," c U, r p I.. 1.?'p, I", I \",l., p I I >.> > c 0 r I > > It. 1 181 r \ . ..I 1 % . In /#It.? I l r # n k ? # . I I I. I' r l n l l i r h Artist ar a Child. The iravettian Human Sheleton from ;he ~ b r i g o Do Lagar Velho and 1 s Archeological Context. lnrtimto Portuguer de Arqueologia, Lisbon, pp. 293-296. S0001 F.. Hublin. 1.. Braun. M. Zonneveld. F.. 2003. The bonv labvrinth of Nean- . . . . * * dertiae. J H&. EVOI 44,141-165. Tillier AM., 1999. Ler Enfants Mourterienr de Qafzeh: Interpretation Phylogene- tique e t Paleoauxologique. CNRS Editions, Paris. Wever E., Lawrence, M , 1954. Phyriological Acourticr. Princeton Univerriy Prerr, Princeton. Xu, P-X., Adamr,J., Peterr, H.,Brown, M , Heaney, S., Maar, R, 1999. Eyal-deficient mice lack ears and kidneys and show abnormal apoptorir of organ primordia. Nat Gen. 23, 113-117. 1-s2.0-S0047248413000109-main_Página_01.jpg 1-s2.0-S0047248413000109-main_Página_02.jpg 1-s2.0-S0047248413000109-main_Página_03.jpg 1-s2.0-S0047248413000109-main_Página_04.jpg 1-s2.0-S0047248413000109-main_Página_05.jpg 1-s2.0-S0047248413000109-main_Página_06.jpg 1-s2.0-S0047248413000109-main_Página_07.jpg 1-s2.0-S0047248413000109-main_Página_08.jpg 1-s2.0-S0047248413000109-main_Página_09.jpg 1-s2.0-S0047248413000109-main_Página_10.jpg 1-s2.0-S0047248413000109-main_Página_11.jpg 1-s2.0-S0047248413000109-main_Página_12.jpg 1-s2.0-S0047248413000109-main_Página_13.jpg