Bull. nat. Hist. Mm. Lond. (Geol.) 57(2): 95-107 



}<Y C3^oo:i\ A J 



Issued 29 November 2001 



Gough's Cave 1 (Somerset, England): a study 

 of the pectoral girdle and upper limbs 



STEVEN E. CHURCHILL 



Department of Biological Anthropology and Anatomy, Duke University, Durham NC 27708, USA 



Synopsis. The pectoral girdles, arms and forearms of Cheddar Man (Gough"s Cave 1 ) are well preserved and, vi'ith the 

 exception of a missing left scapula, are completely represented. These remains, which derive from early Holocene deposits in 

 Cough's Cave, are described here. Comparative evaluation of the Cough's Cave 1 remains reveals that his upper limb skeleton was 

 somewhat gracile, but certainly within the range of variation in strength measures of his contemporaries. 



INTRODUCTION 



The upper appendicular skeleton of Cheddar Man is represented by 

 both claviculae, the right-side scapula, the humeri, ulnae and radii of 

 both sides, and elements of both hands. The preservation of the upper 

 limb elements is generally good. 



The elements of the pectoral girdles, arms and forearms are 

 described below (the manual remains are described elsewhere), 

 followed by a discussion of the overall morphology of the upper 

 limbs. 



MATERIAL 



The description of the Cough's Cave 1 upper limb remains includes 

 both traditional osteometric (Tables 1, 2, 4-7, 9, 10, 12 and 13) and 

 diaphyseal cross-sectional data (Tables 3, 8, and 1 1 ). Mean osteo- 

 metric and cross-sectional data are provided for other European 

 Mesolithic and later Upper Paleolithic specimens to provide a com- 

 parative framework for evaluating the morphology of Cough's Cave 

 1. These additional specimens range in geological age from the 

 terminal Pleistocene to the mid-Holocene, and include Arene Candide 

 2 to 5, 10, and 12 to 15, Bruniquel 24, Cap-Blanc 1, Chancelade 1, 

 Farincourt 1, Gramat 1, Hoedic 1. 2, and 4 to 10, Neuessing 2, 

 Oberkassel 1 and 2, Rochereil 1, Romito 3 and 4, Romanelli 1, St. 

 Germain-la-Riviere 4, Teviec 1, 7 to 9, 11 and 16, and Veyrier 1 

 (Stasi & Regalia, 1904;'Verwomera/., 1919;Grazzini, 1921;Pequart 

 (fePequart, 1934;Bonin, 1935;Boule&Vallois, 1937, 1946; Vallois, 

 1941-1946, 1972; Lacam et al, 1944; Pittard & Sauter, 1945; 

 Sauter, 1957; Graziosi, 1962; Genet- Varcin & Miquel, 1967; Billy, 

 1969; Gieseler, 1977; Paoli et al, 1980). All comparative data were 

 collected by the author on original specimens. 



Diaphyseal cross-sections of the claviculae, humeri and ulnae 

 were reconstructed from radiographs and external contour molds for 

 the midshaft (claviculae and humeri) or mid-proximal (ulnae) dia- 

 physes. Subperiosteal contour molds were taken perpendicular to the 

 diaphyseal axis, using dental putty molds (Cuttersil Putty Plus; 

 Heraeus Kulzer Inc.), at 50% (midshaft) or 65% (mid-proximal) of 

 biomechanical length (measured from the distal end). The molds 

 were photostatically reproduced on paper to provide the subperio- 

 steal (outside) contour of the cross-section. In the case of claviculae. 

 ventral, dorsal, superior and inferior cortical dimensions were meas- 

 ured from superoinferior and dorsoventral radiographs. For humeri 

 and ulnae, anterior, posterior, medial and lateral cortical thicknesses 

 were measured from mediolateral and anteroposterior radiographs. 



Subperiosteal dimensions from the original specimens were com- 

 pared with those from the radiographs to determine the degree of 

 parallax distortion and thus allow for algebraic correction of cortical 

 thickness measurements. The cortical dimensions were used along 

 with the subperiosteal contour to interpolate the endosteal contour. 

 The resultant cross-sections were manually digitized and geometric 

 properties were computed using a PC-DOS version (Eschman, 1 990) 

 of SLICE (Nagurka & Hayes, 1970). 



SLICE calculates the total subperiosteal (TA) and cortical (CA) 

 areas, 2nd moments of area about the superoinferior (clavicle) or 

 anteroposterior (humerus and ulna) (I ) and dorsoventral (clavicle) 

 or mediolateral (humerus and ulna) (I ) axes, and the maximum (I ) 



^ ' ^ y/ ' ^ max' 



and minimum (I^^^) 2nd moments of area. Geometric analysis of 

 cross-sections provides measures of the contribution of bone geom- 

 etry to the resistance of biomechanical loads: in the case of cortical 

 area, to axial compressive and tensile loads; for 2nd moments of 

 area, to bending loads. Medullary area (MA) can be determined from 

 total and cortical areas (MA = TA - CA). The polar moment of area 

 (J, or I ) is a measure of torsional rigidity and overall strength, and 

 can be determined as the sum of any two perpendicular 2nd moments 

 of area ( J = [I^ -I- 1 ] = [I^^^ + I^.J )• Since cross-sectional measures of 

 strength co-vary positively with body size, humeral and ulnar cross- 

 sectional values were standardized by powers of bone length to 

 produce measures of relative strength, or robusticity. Following the 

 rationale described in Churchill (1994), measures of cross-sectional 

 area (TA, CA and MA) were standardized as: 



AREA , , = (AREA , , , ,/AL-) * 10-' 



slandardized ^ unslandarai/ed 



where AL = humeral or ulnar articular length. 2nd (and polar) 



moments of area (I^, I , I_^^_^, I^^^, J) were similarly standardized as: 



2nd Moment \ [ "= (27d Moment , , , JAL*) * 10". 



slandiirdtzed ^ unMandardized 



In addition to the measures of bone rigidity outlined above, three 

 cross-sectional shape indices were computed to better illustrate the 

 morphology of the Cough's Cave 1 upper limb material. Percent 

 cortical area (%CA = 100*CA/TA) serves as a simple measure of the 

 degree of cortical occlusion of the medullary space. Ratios of 2nd 

 moments of area provide information about diaphyseal shape (at the 

 location of the cross section) with respect to anatomical axes (I,/!^) 

 or with respect to the axis of maximum bending rigidity {i^Jl^J- 



CLAVICULAR REMAINS 



Both claviculae are preserved and are virtually complete (Figs 1,2). 

 Both lack the sternal epiphysis but have well preserved metaphyseal 

 plates, making it clear that the sternal .secondary centres of ossifica- 

 tion were unfused. The right clavicle is damaged on the inferior and 



© The Natural History Museum. 2001 



