Confocal Scanning Optical Microscopy of a 3-Million-Year-OldAustralopithecus afarensis Femur

T. G. BROMAGE1,2, H. M. GOLDMAN

3, S. C. MCFARLIN4, A. PEREZ OCHOA

5, A. BOYDE6

1Department of Biomaterials and Biomimetics, New York University College of Dentistry, New York,

New York2Department of Basic Science and Craniofacial Biology, New York University College of Dentistry, New York,

New York3Department of Neurobiology and Anatomy, Drexel University College of Medicine, Philadelphia, Pennsylvania4Department of Anthropology, Center for the Advanced Study of Hominid Paleobiology, The George Washington

University, Washington, District of Columbia5Centro Superior de Estudios Universitarios, Lasalle University, C/La Salle, Madrid, Spain6Department of Dental Biophysics, Centre for Oral Growth and Development, Queen Mary University of London,

London, United Kingdom

Summary: Portable confocal scanning optical micro-scopy (PCSOM) has been specifically developed forthe noncontact and nondestructive imaging of earlyhuman fossil hard tissues, which here we describe andapply to a 3-million-year-old femur from thecelebrated Ethiopian skeleton, ‘‘Lucy,’’ referred toAustralopithecus afarensis. We examine two bonetissue parameters that demonstrate the potential ofthis technology. First, subsurface reflection imagesfrom intact bone reveal bone cell spaces, the osteocytelacunae, whose density is demonstrated to scalenegatively with body size, reflecting aspects ofmetabolism and organismal life history. Second,images of a naturally fractured cross section near toLucy’s femoral mid-shaft, which match in sign thoseof transmitted circularly polarized light, revealrelative collagen fiber orientation patterns that are animportant indicator of femoral biomechanicalefficacy. Preliminary results indicate that Lucy wascharacterized by metabolic constraints typical for aprimate her body size and that in her femur she was

adapted to habitual bipedalism. Limitations imposedby the transport and invasive histology of unique orrare fossils motivated development of the PCSOM sothat specimens may be examined wherever andwhenever nondestructive imaging is required.SCANNING 31: 1–10, 2009. r 2009 Wiley Period-icals, Inc.

Key words: confocal microscopy, early hominid,Australopithecus afarensis, collagen fiber orienta-tion, osteocyte lacunae

Introduction

Most fossils are either translucent or, if they aresurface reflective, are not flat. In both cases, lightinteracts with the sample over a considerable verticalrange and is reflected (or the fluorescent light ema-nates) from a thick layer. The challenge we face forthe nondestructive examination of unique earlyhuman fossil ancestors (hominins) is how to obtainresearch-grade images of microantomical featuresof the skeleton in the field setting from such speci-mens as has been demonstrated for recent tissues(Boyde et al. 1983). We have found a solution indevelopment of portable confocal microscopy, ourobjective being to image microanatomical featuresin fossil bones and teeth for the purpose of describingaspects of the hard tissue biology and the organismallife and evolutionary histories of our extinctancestors.

The principle of the portable confocal scanningoptical microscope (PCSOM), as for any confocalPublished online in Wiley InterScience (www.interscience.wiley.com)

DOI 10.1002/sca.20139

Received 14 November 2008; Revised 29 December 2008;

Accepted with revision 5 January 2009

Address for reprints: T. G. Bromage, Departments of Biomaterials

and Biomimetics and Basic Science and Craniofacial Biology,

New York University College of Dentistry, 345 East 24th Street,

New York, New York 10010.

E-mail: tim.bromage@nyu.edu

Contract grant sponsors: L.S.B. Leakey Foundation; The Blan-

quer Foundation; The March Foundation; National Science

Foundation; Contract grant numbers: 9727689; 0202823.

SCANNING VOL. 31, 1–10 (2009)& Wiley Periodicals, Inc.

microscope, is to eliminate the scattered, reflected,or fluorescent light from out of focus planes,allowing only light originating from the plane offocus of the objective lens to contribute to imageformation. It does this at the several conjugate focalplanes (each plane representing the image of theother i.e., intermediate, eye point, and imagerecording device), and thus eliminates light comingfrom all out of focus planes. In practice, an illumi-nated spot in the plane of focus is scanned across thefield of view and an image is compiled. Confocalscanning optical microscopy thus differs from con-ventional light microscopy, where light from thefocus plane of the objective lens, as well as from allout of focus planes across the entire field of view, isobserved. The history and various technicalachievements in confocal microscopy are summar-ized by Boyde (1995).

Studies of early hominin life history and skeletalfunctional adaptation typically focus on the study ofbone at the organ level. There are, however, bonemicroanatomical features that reflect physiologicalprocesses in relation to developmental and me-chanical loading histories. For instance, osteoblastsare cells that produce the bone organic matrix thatis subsequently mineralized, which regularly becomeincorporated into the bone matrix during thegrowth of bone. Osteoblasts, once embedded in thebone matrix, are called osteocytes, which occupymicroanatomical spaces called lacunae. Osteocytelacunar density in mammalian bone is inverselyproportional to body mass (Mullender et al. 1996),which is an indicator of organismal life history; thesmaller the mammal the faster its metabolism andlife history, which is reflected in faster rates of os-teoblast proliferation and osteocyte lacunar density.

The preferred orientation of collagen fibers inbone is another microanatomical feature, which hasbeen demonstrated to shed light on the biomecha-nical history and competence of bone (Ascenzi et al.1967; Simkin and Robin 1974). Indeed, examina-tions of several material properties of bone havedetermined that collagen fiber orientation makes themost significant contribution to the overall strengthof bone tissue (Martin and Boardman 1993; Martinand Ishida 1989; Riggs et al. 1993).

The orientation of collagen (and its associatedmineral) is most appropriately investigated by cir-cularly polarized light (CPL) microscopy of bonehistological thin cross sections (e.g., 100 mm thick).Collagen fibers oriented transversely, or those per-pendicular to the long axis of the bone, resist axialcompressive forces and appear bright (i.e., white),those oriented in the longitudinal axis of the boneresist tensile forces and appear dark (i.e., black), andfibers having intermediate orientations appear indifferent levels of gray in a monochrome image. The

brightness in histological thin sections observed byCPL is identical in sign to PCSOM images of thesame fields of view, and though we as yet have noexplanation for this similarity, it is a characteristicof PCSOM that may be exploited for interpretingcollagen fiber orientations in bone.

The PCSOM provides Z-axis through focusimaging of topographically complex surfaces atrelatively high magnifications revealing a planeview of fossil hominid hard tissue microstructure.Osteocyte lacunae are visible because of specularlyreflected light from the inner walls of the space andcrystals precipitated from ground waters withinlacunae during the fossilization process. With theemploy of CPL-like images, the PCSOM alsoprovides information on subsurface collagenorientation contrast as well. Here, we examinebone microanatomical details from the femur of‘‘Lucy,’’ a celebrated 3-million-year (my) fossilhominid skeleton from Ethiopia, representative ofAustralopithecus afarensis.

Materials and Methods

The Microscope

We employ a PCSOM based on the Nipkow disktechnique (Nipkow 1884) described in detail byPetran and Hadravsky (1966) and first commercia-lized in the early 1980s. The Petran and Hadravskydesign uses a so-called ‘‘two-sided’’ disk; the speci-men is illuminated through an array of pinholes onone side of the disk while detected through a con-jugate array of pinholes on the other (via a numberof delicately aligned mirrors). Applications of thistechnology to bone and tooth microanatomy havebeen described (Boyde et al. 1983). Another Nipkowdisk design (the one used here) employs a ‘‘single-sided’’ disk in which the illumination and detectionpinhole is one in the same (Kino 1995): that is,illuminating light and its reflections from the objectpass through the same pinhole, which is imaged bythe eyepiece objective or camera. This latter designis robust and able to tolerate our relatively extremeportable applications.

The PCSOM is illustrated in Figure 1, whichemploys a one-sided Nipkow disk TechnicalInstrument Co. K2S-BIO confocal module (ZygoCorp., Sunnyvale, CA). Like other confocal scan-ning optical microscopes, the final image derivesfrom the plane of focus, thus it eliminates the fogowing to the halo of reflected, scattered, or fluor-escent light above and below the plane of focus,which otherwise confounds image content in con-ventional light microscopy.

2 SCANNING VOL. 31, 0 (2009)

An interesting feature of the single-sided diskdesign used here (Kino 1995) is the approach takento suppress internal, nonimage-related reflectionsthat are a significant problem in this type of system;light reflecting from internal components of themicroscope, having nothing to do with forming animage, degrades the image. This method is theclassical method of illuminating with polarized lightto stop light reflections from within the opticalsystem (e.g., from optical hardware within the bodyof the microscope), but not the useful light reflectingfrom the specimen and returning through theobjective lens. Linear polarizing light filters and asingle quarter-wave plate filter, described furtherbelow, provides the means for eliminating theunwanted reflected light. A consequence of the sin-gle-sided disk design is that it significantly reducesthe number of mirrors in the light path making thealignment of the optics less critical. The result is avery robustly constructed instrument able to toler-ate transport and relatively rough handling (e.g., aschecked-in baggage for air travel).

The microscope configurations include severalother features critical to our research. Considerationwas given to obtaining objective lenses with rela-tively long working distances (i.e., ca. 20 mm)because often we have little control over the geo-metry of broken fossil bone surfaces examinedunder remote field or museum conditions, and so wemust be prepared to image through long Z-heightpositions to avoid interference between the fossilsurface and the objective nosepiece. Objectives

chosen include 5� and 10� lenses (34 and 19 mmworking distances respectively; Thales-Optem Inc.,Fairport, NY) and Mitutoyo 20� and 50� lenses(20 and 13 mm working distances respectively; Mi-tutoyo Asia Pacific Pte Ltd, Singapore). Flexibilityin magnification is achieved by both the introduc-tion of a Thales-Optem 0.5� or 1.9� CCD adapteror by converting the fixed magnification optical as-sembly described above into a zoom system, whichinvolves the introduction of a Thales-Optem 70�Lzoom module (1–7� ) between the K2S-BIO modulecoupler and the manual coarse/fine focus module.For fully automated image acquisition, we motor-ized the Z focus (below).

Automation in X, Y, and Z axes has been var-iously implemented onto the PCSOM, whichincludes a KP53 motorized precision micro-steppingXY stage from the Semprex Corporation (Campbell,CA), and a Vexta 2-phase Z-axis stepping motor(Oriental Motor USA Corp., Torrance, CA).Integrated XYZmovement is performed by an Oasis4i PCI stepper motor controller board for XY stageand Z focus. A three-axis trackball/mouse controlof XYZ axes allows manual stage and focus move-ment to aid real-time viewing.

Portable image acquisition are transmittedthrough the FireWireIEEE 1394 digital interfacenow common on notebook and desktop computers,thus eliminating the need for a framegrabber. ThePCSOM uses a JVC KY-F1030U 6-pin IEEE 1394digital camera containing a 1

2in color progressive

scan interline CCD and 1360� 1024 output pixels,operating at 7.5 frames per second live.

The 300W Lambda LS Xenon Arc Lamp (SutterInstrument Company, Novato, CA) transmits a flatand intense beam of light via a liquid light guide. Itoperates at wavelengths suitable for both fluores-cence and white light illumination (320–700 nm out-put in an ozone-free bulb), is robustly constructedand pre-aligned, and is economically packaged andlightweight, housing its own power supply.

A Shuttle XPC SB52G2 computer with a Pen-tium4 Intel processor and Windows XP Professional(Shuttle Computer Group Inc., Los Angeles, CA)supports fully automated XYZ stage movement andimage acquisition. A reasonably lightweight and thinstandard 1024� 768 15 in monitor (Dell Inc., RoundRock, TX) was chosen for our real-time viewing.

The microscope returns image detail from a verythin optical plane at and immediately below theobject surface (1–50 mm, depending upon specimencharacteristics). To obtain two- or three-dimen-sional projections from a surface which is anythingbut perfectly flat, potential fields of view must becompiled from a through-series of captured imagesat all optical planes represented in the Z-axis.Computerized control over image acquisition using

Fig 1. Diagram of the portable confocal scanning opticalmicroscope. See text for details.

T. G. Bromage et al.: Portable CSOM of a 3 my femur 3

Syncroscopy Auto Montage software (SyncroscopyInc., Frederick, MD) permits an even and fullyrepresentative image of either a pseudo-planar fieldof view or a three-dimensional reconstruction ofsurface or subsurface details. For extensiveautomated XY image montaging, SyncroscopyMontage.Explorer software is employed, which canoperate in ‘‘3D mode’’ to acquire useful Z focalplanes over fields as large as 40,000� 40,000 pixels.

A simple and lightweight stand is manufacturedfrom aluminum and includes an upright cylinder,containing within a lead screw operable from above,which drives the Nipkow disk module platform upor down; the drive is sensitive enough to be used as acoarse focus adjustment. The cylinder inserts into asleeve at the base from which two hollow rectan-gular feet slide forward and rotate out at any angleappropriate for the balance of weight and requiredworkspace. The platform for holding the K2S-BIOattaches to a sleeve around the cylinder, which rideson a bearing that conveys the module in any rota-tional position within the workspace.

Microscope electronics automatically switch be-tween 110 and 220 V electrical supplies (only theNipow disk motor requires an optional 110/220 Vadaptor), fits into two suitcases (Pelican Products,Inc., Torrance, CA), and may be set up and testedwithin 1 h of arrival at museum locations.

Circularly Polarized Light

In a conventional CPL microscope, unpolarizedlight passing through the linear polarizer filterbecomes in-phase linearly polarized light (LPL).LPL becomes polarized into helical vibration planeswhen passing through a quarter wave retardingplate. A quarter wave plate has two transmissionaxes at 901 to one another: fast and slow. Althoughthese axes do not diminish the intensity of incomingpolarized light, they do resolve the beam into twoemerging orthogonal components, one along thefast axis and one along the slow axis when the in-cident polarized light lies between them at 451. LPLresolved by the quarter wave plate along the slowaxis lags that of the fast axis and emerges from thefilter one quarter wave length behind and out ofphase; the light is no longer linearly polarized. Thenet vector of each component now vibrates aroundan elliptical axis, originally turned toward the slowaxis to produce a right circular trajectory (clockwisetoward the observer). This is CPL.

Polarizing filters in the K2S-BIO are arranged sothat stray light within the optical system is rejectedfrom light paths to the eyepieces and recordingsystem (Fig. 2). Light scattered from within the

optical system is rejected by an analyzer crossed to apolarizer. A quarter wave plate diagonally orientedbetween the objective front lens and the specimensurface imprints a wave shift, which results in LPLin a direction of vibration of which is rotated 901from the original direction. This puts it parallel tothe analyzer and what is transmitted forms theclassic incident light bright field image at reducedglare. This filter arrangement invites an explanationfor how the K2S-BIO generates reflected CPL.

CPL will refract in a birefringent material suchthat its elements oriented parallel with the plane ofsection will transmit peak light intensities irrespec-tive of their 3601 rotational position. Incident rightCPL (i-RCPL) reflecting from the specimen doesnot maintain a right circular trajectory (as would betransmitted through a birefringent thin section ob-served by a typical transmitted light microscope),but reverses to take a reflected left CPL helical path(r-LCPL; clockwise away from the observer). Thislight is converted back into LPL as it passes onceagain through the quarter wave plate whose fast andslow axes are effectively reversed (extinction posi-tion) from the first, when one considers the reversalfrom right to left CPL. The LPL polarization sense

Fig 2. Diagram of the K2S-BIO filter arrangement. Thelamp source generates unpolarized light having vibrationplanes in all 3601. This light is first converted into linearlypolarized light (LPL) along the transmission axis of thepolarizer and then bent into incident right circularlypolarized light (i-RCPL) waves (clockwise to the observer)in the direction from fast (F) to slow (S) axes of the quarterwave plate filter. As i-RCPL reflects from a birefringentspecimen, the light reverses its helical path and is refractedback in all 3601 rotational positions as a reflected leftcircularly polarized light (r-LCPL). Refracted light isre-linearly polarized 901 to the polarizer at the quarter waveplate filter, which is then allowed to pass along thetransmission axis of the analyzer. The emerging LPL isobserved at the eyepiece objectives or the recording system.

4 SCANNING VOL. 31, 0 (2009)

is shifted 901 owing to birefringence of the collagenfibers, aligning now with the transmission axis of theanalyzer.

Despite this logic, we remain unsure of its vera-city. Irrespective, repeated comparisons of samefield of view images of bone by conventionaltransmitted CPL with those obtained by the K2S-BIO reveal the same signal. Thus, for all tense andpurposes, gray-level images of bone obtained by theK2S-BIO may be interpreted in respect of variationin collagen fiber orientation.

The Specimens

We examine here the right femur from the 3 myearly human fossil ancestor attributed to A. afarensis,catalogue number AL 288-1a-p and commonlyreferred to as ‘‘Lucy.’’ The long bone shaft fragmentshad been previously assembled, but were de-glued byNational Museum of Ethiopia staff, thus presenting adistal shaft whose intact surfaces and mid-shaft crosssection were available for imaging.

Microanatomical details acquired from Lucy werecompared with 67 modern human (Goldman et al.1999) and 13 chimpanzee (Pan troglodytes) (McFarlinet al. 2008) mid-shaft femurs derived from individualsof partially known life history. Those employed in thedetermination of osteocyte lacunar density, togetherwith a variety of other primates and nonprimatemammals, are listed in Table I. Comparisons of the

distributions of Lucy’s mid-shaft collagen fiber or-ientations were made relative to quantitative ob-servations represented as color-coded maps acquiredfrom the human and chimpanzee study sample.

We cleaned all modern bone and tooth specimenswith 1% Terg-a-Zyme (Alconox, NY) solution at50 1C, subject to dehydration by graded ethanolsubstitution and then 50:50 isopropanol:heptanereflux (intermittent exchange of fresh solution in aSoxhlet apparatus) for 7–14 days. All modern andfossil samples excepting the AL 288-1a-p homininspecimen were polymerized in plastic (poly-methylmethacrylate). Each cured block was sec-tioned to a uniform thickness of 100 mm, 74 mm inmost cases. AL 288-1a-p was examined by PCSOMand thus required no preparation.

Microscopy and Analysis

Osteocyte density

Imaging of the early hominin was obtained byPCSOM. A 270 mm field width by 200 mm fieldheight field of view was imaged through 50 mm offocus from the first surface. The focal plane wastranslated vertically from the outer shaft surfaceinto the bone cortex by ca 5 mm virtual increments;we used the formula 1.52� 3 mm absolute verticaltranslation, where 1.52 is taken as the refractiveindex (RI) of ethylene glycol coverslip medium and

TABLE I Selected primate and nonprimate mammal osteocyte density and body mass

Specimen(common name) Genus Age and sexy

Osteocytedensity

Body mass(kg)z

Specimensourcey

PrimateLesser Galago Galago moholi Adult 51,724 0.244 DUPCGreater Galago Cheirogales major Adult 31,526 0.4 DUPCDwarf Lemur Otolemur crassicaudatus Adult 44,353 1.15 DUPCVervet Chlorocebus aethiops Adult female 32,012 3.515 UCSCMacaque Macaca mulatta 4.16 yr female 22,222 3.0 MIBPChimpanzee Pan troglodytes M2 erupted (ca. 6 yr) 18,706 33.7z MNH‘‘Lucy’’ Australopithecus afarensis Female adult 23,333 27.5J NMEHuman Homo sapiens 49 yr female 19,166 62 VIFMNonprimate mammalRat (Wistar) Rattus norvegicus 3 mo female 58,148 0.3 HTRUPygmy Hippo Phanourios minutes Adult 23,641 200yy TRNCHippo Hippopotamus amphibius Adult 16,667 2,000zz HLD

yAge and sex are given when known. No adult specimen was geriatric.zExcept where indicated, body mass was measured directly from study individuals.yDUPC, Duke University Primate Center; HLD, Hessisches Landesmuseum Darmstadt; HTRU, Hard Tissue Research Unit,NYUCD; MIBP, Morgan Island Breeding Program, Yemassee, South Carolina; MNH, Museum fur Naturkunde der HumboldtUniversitat; NME, National Museum of Ethiopia; TRNC, Turkish Republic of Northern Cyprus, Department of Antiquities;UCSC, Sherwood L. Washburn collection, University of California at Santa Cruz; VIFM, Victorian Institute of ForensicMedicine, Melbourne, Australia.zSmith et al. (2002).JJungers (1982).yySimmons (1999).zzEltringham (1999).

T. G. Bromage et al.: Portable CSOM of a 3 my femur 5

a 170 mm thick coverslip. The through-series ofimages from each ca. 5 mm virtual focal plane wasacquired using Syncroscopy Automontage until50 mm depth had been achieved. A red-green ana-glyph image with 201 angular separation was cal-culated from this series with Automontage.

Imaging of the comparative samples was per-formed with the Edge R400 real time 3D micro-scope, configured to provide transmitted obliqueCPL contrast when useful. Histological sections of10074 mm were coverslipped with ethylene glycol,and XY fields of view conforming to 270 mm fieldwidth by 200 mm field height, respectively, andcontaining no dehydration cracks or vascular canalswere located. Subsequent imaging procedures fol-lowed the same PCSOM imaging routine.

Each through-focus series of images was presentedfor analysis on a PC configured with dual monitors.The Automontage through series was placed on onemonitor, whereas the analglyph image was openedseparately in Adobe Photoshop CS2 (San Jose, CA)on the other. This side-by-side arrangement allowedthe viewing of lacunae with red-green anaglyphglasses through the full 3D thickness of the histolo-gical section while also scrolling through and crosschecking source image details in Automontage.

All lacunae were counted for each XYZ volumeaccording to standard stereological procedure; thoseintersecting each of one X, Y, and Z surface werelabeled as excluded, while those intersecting theircontralateral X, Y, and Z surfaces were labeled asincluded. After all lacunae had been labeled, theywere manually counted.

For each histological thin section a single XYZvolume was selected wherever a complete field couldbe located without vascular canals or cracks. Theperiosteal cortex was imaged in all individuals, thoughfor purposes of another study on femoral bio-mechanics, five periosteal and five endosteal volumeswere measured around the entire cross section of thehuman sample, which were averaged into the valuegiven in Table I. All measurements were standardizedby simple extrapolation of the number of lacunae permeasured bone volume, typically from a 5,400,000 mmcubic measurement volume, to a 1mm3 unit value.

Collagen fiber orientation

Imaging of the early hominin was obtained byPCSOM. The distal AL 288-1a-p shaft was situatedupright so that the naturally fractured surface wasmore or less plane to the incident beam. There was,however, an approximately 1 cm relief that had to benegotiated by extended focus to be presented as asingle two-dimensional image. Extensive automatedXY image montaging was performed by SyncroscopyMontage.Explorer software control of the XY stage,

which was operated in ‘‘3D mode’’ to acquire all per-tinent Z focal planes using the Z-axis stepping motor.

Images of whole histological cross sections ob-tained on the comparative human and chimpanzeespecimens were acquired by transmitted CPL mi-croscopy using a Leica DMRX/E Universal micro-scope (Leica Inc., Wetzlar, Germany) configuredwith an automated Marzhauser stage controlled bySyncroscopy Montage Explorer.

Results

A count of 63 osteocytes was made in the samplevolume of AL 288-1a-p, which extrapolates to 23,333osteocytes per mm3 (Fig. 3), which is consistent withher body size of 27.5 kg. Osteocyte lacuna densitiescalculated from several comparative primate andnonprimate mammals are presented in Table I andgraphically represented in Figure 4. It is noted thatthe primate regression of osteocyte lacuna density vs.body mass falls below that of a generalized mam-malian relationship, which indicates that, for theirbody size, primates incorporate fewer osteoblastsinto the bone matrix than other mammals.

Quantitative analyses of CPL employ a simplemethod for mapping preferential collagen fiber or-ientation patterns in whole cross sections (Boydeand Riggs 1990). Eight-bit gray-scale images arevisualized and pixels distributed by color into eightgray-level bins. Although the comparative humansand chimpanzees were so evaluated, interpretationsof whether the AL 288-1a-p bone cortex containsprimarily transversely or longitudinally orientedcollagen is qualitative, because it is not now possibleto arrive at a quantitative solution to the resultsowing to the vagaries of fossil preservation (Brom-age et al. 2003). Hence, the comparison betweenfossil and modern specimens is relative.

Readily apparent from color-coded gray levelimages of collagen fiber orientation, where cool col-ors denote collagen lying longitudinal with the axis ofthe femur and hot colors represent collagen lyingtransverse to the femoral axis, the human medialfemoral cortex (that cortex along the inner leg) con-tains relatively more transverse collagen than othercortices (Fig. 5). This is interpreted to mean that themedial cortex experiences significant alignment ofcollagen to resist compression during locomotoryfunction. The chimpanzee anterior (front), lateral(outer leg), and posterior (rear) cortices are im-pressively dark, though the intermediate posterior-medial cortex does contain some transverse collagen(Fig. 6). The entire AL 288-1a-p cross section wasimaged, providing basic details of the microanatomy(Fig. 7). Preliminary examination of the fossilhominid femur at higher magnification illustrates

6 SCANNING VOL. 31, 0 (2009)

that it is more similar to that of a modern humanfemur, containing more transverse collagen, particu-larly in the medial cortex than elsewhere (Fig. 8).y

Discussion

There is much interest in obtaining details ofearly hominid skeletal microanatomy from natural

Fig 4. Log osteocyte density linearly regressed against logbody mass for primates � and selected nonprimate mammals~ (least squares model). Raw data are provided in Table I.Summary statistics for primates are: r2 5 0.73, r5 –0.86,po0.01, y5 10.473–0.156�x, and for selected mammals it isr2 5 0.99, r5 –0.99, p5 0.01, y5 10.804–0.144�x. When allprimates and mammals examined here are treated together,the relationship is described by r2 5 0.72, r5 –0.85, po0.01,y5 10.513–0.124�x.

Fig 5. Human mid-shaft femur; 27 yr female. Specimenwidth5 26 mm (medial at left, anterior at top, lateral at right,posterior at bottom). Collagen fiber orientation is repre-sented by a color look-up-table (LUT). See text for details.

Fig 3. Real color stereopair image of osteocyte lacunae immediately beneath the bone surface of A.L. 288-1ap. Color derivesfrom minerals, which have precipitated inside the lacunae. FW5 200 mm.

yTo confirm the presence of collagen in 3 my fossils, we employedFourier Transform Infrared (FTIR) Spectroscopy in the range4,000–500 cm�1. Birefringent faunal lamellar bone tissue patternswere identified in a partially demineralized thin section derivedfrom the early hominid site of Makapansgat, South Africa.Absorbance peaks followed very closely those represented byFTIR spectroscopy of bovine collagen standards, and the onlyrational hypothesis about the nature of this material is that it isrelatively intact collagen.

T. G. Bromage et al.: Portable CSOM of a 3 my femur 7

and fractured surfaces, but such surfaces are rarelygiving of all the desired detail. The resolving power,such as that of stereo-zoom microscopy, and therecovery of microanatomical detail from sig-nificantly below the surface by scanning electronmicroscopy, has been wanting. With the develop-ment and employ of portable circularly polarizedconfocal scanning optical microscopy, we are nowable to resolve microanatomical details relevant toreconstructions of early hominid development andfunctional life history.

It is now possible to use bone microanatomy toplace early hominids in the organismal life historygamut (Harvey and Clutton-Brock 1985). For in-stance the lower height of the primate regression ofosteocyte lacuna density vs. body mass (Fig. 4)agrees with growth law predictions indicating thatprimate production energy (i.e., that energy re-quired to develop body mass) is significantly lesscompared with other mammals of the same bodysize (Charnov and Berrigan 1991). These predictionsare part of a larger model, which explains thatgrowth energy determines also how much energy isavailable for reproduction later. This suggeststhat energy-limited rates of production, boneand body size, and reproduction are centrally linkedto the regulation of organismal life historyaccording to metabolic constraints (Brown andSibly 2006).

Collagen fiber orientations in humans are com-plex because of the variety of axial and compoundforces during the various stages of bipedal loco-motion, but in general the anterior and the inter-mediate anterior–lateral aspects of the cortex aresubject to tension and the posterior–medial andmedial aspects are subject to compression, whichmore or less agrees with the average human fiberdistribution (Goldman et al. 2003). The posteriorcortex is consistently dark, however, because of thenet tensile influence of muscles attaching by collagenfibers at a high near-longitudinal angle to the femurshaft. Of note, the load axis of a chimpanzee femuris entirely medial to the shaft, thus some resistanceto compressive strain may be expected on the medialcortex as would be resistance to tension on thelateral cortex, as reflected in their preferred fiberdistributions (McFarlin et al. 2008). That the AL288-1a-p bone cortex contains a modicum oftransversely oriented collagen in all anatomicalsectors indicates that Lucy was adapted to habitualbipedalism and unlike the locomotory strategies ofliving apes.

Conclusion

The portable confocal scanning optical micro-scope was specifically developed to offer superbanalytical light microscopy of early hominid skeletalmaterial. The microscope is, in the main, a heuristicdevice, enabling us to access and learn about themicrostructures of objects normally out of reachfrom such instrumentation. Because it is both asuperb circularly polarized light and reflectionmicroscope, we have added new information ondevelopmental and functional processes in the bonetissues of an early human ancestor.

Fig 7. The entire mid-shaft femoral cross section of A.L.288-1ap imaged by PCSOM. Specimen width5 20.5 mm(medial at left, anterior at top, lateral at right, posterior atbottom).

Fig 6. Chimpanzee mid-shaft femur; 12–13 yr female.Specimen width5 21 mm (medial at left, anterior at top,lateral at right, posterior at bottom). Collagen fiber orienta-tion is represented by a color look-up-table (LUT) illustratedin this figure. See text for details.

8 SCANNING VOL. 31, 0 (2009)

Acknowledgements

Additional support was provided to S.C.M. byThe George Washington University’s SelectiveAcademic Excellence Initiative to the Center for theAdvanced Study of Hominid Paleobiology. To eachof the numerous individuals and institutions thatprovided specimens listed in Table I we are grateful,particularly to Mamitu Yilma, Director, and scien-tific staff and technical assistants of the National

Museum of Ethiopia. We also thank Jan Hinsch forcritical comments on an earlier draft of this article.

References

Ascenzi A, Bonucci E, Bocciarelli DS: An electron micro-scope study on primary periosteal bone. J UltrastructRes 18, 605–618 (1967).

Boyde A: Confocal optical microscopy. In Image Analysis inHistology: Conventional and Confocal Microscopy,

Fig 8. Higher magnification medial (a; FW5 1150 mm), posterior (b; FW5 1,150 mm), lateral (c; FW5 1,192 mm) and anterior(d; FW5 1,002 mm) CPL image fields of A.L. 288-1ap. Collagen fiber orientation is represented by a color look-up-table (LUT)illustrated in Figure 6. See text for details.

T. G. Bromage et al.: Portable CSOM of a 3 my femur 9

Wootton R, Springall DR, Polak JM (Eds.), CambridgeUniversity Press, Cambridge, 151–196 (1995).

Boyde A, Petran M, Hadravsky M: Tandem scanning re-flected light microscopy of internal features in wholebone and tooth samples. J Microsc 132, 1–7 (1983).

Boyde A, Riggs CM: The quantitative study of the orientationof collagen in compact bone slices. Bone 11, 35–39 (1990).

Bromage TG, Goldman HM, McFarlin S, Warshaw J,Boyde A, et al.: Circularly polarized light standards forinvestigations of collagen fiber orientation in bone. AnatRec New Anat 274B, 157–168 (2003).

Brown JH, Sibly RM: Life-history evolution under a pro-duction constraint. Proc Natl Acad Sci 103, 17595–17599(2006).

Charnov EL, Berrigan D: Why do female primates have suchlong lifespans and so few babies or life in the slow lane?Evol Anthropol 1, 191–194 (1991).

Eltringham SK: The hippos. Academic Press-London (1999).Jungers WL: Lucy’s limbs: skeletal allometry and loco-motion in Australopithecus afarensis. Science 297,676–678 (1982).

Goldman HM, Bromage TG, Thomas CDL, Clement JG:Preferred collagen fiber orientation at the human mid-shaft femur. Anat Rec 272A, 434–445 (2003).

Goldman HM, Kindsvater J, Bromage TG: Correlative lightand backscattered electron microscopy of bone. Part I:specimen preparation methods. Scanning 21, 40–43 (1999).

Harvey PH, Clutton-Brock TH: Life history variation inprimates. Evol 39, 559–581 (1985).

Kino GS: Intermediate optics in Nipkow disk microscopes.InHandbook of Biological Confocal Microscopy, Pawley JB(Ed.), Plenum Press, New York, 155–165 (1995).

Martin RB, Boardman DL: The effects of collagen fiberorientation, porosity, density, and mineralization on

bovine cortical bone bending properties. J Biomech 26,1047–1054 (1993).

Martin RB, Ishida J: The relative effects of collagen fiberorientation, porosity, density, and mineralization onbone strength. J Biomech 22, 419–426 (1989).

McFarlin SC, Terranova CJ, Zihlman AL, Enlow DH,Bromage TG: Regional variability in secondaryremodeling within long bone cortices of catarrhine pri-mates: the influence of bone growth history. J Anat 213,308–324 (2008).

Mullender MG, Huiskes R, Versleyen H, Buma P: Osteocytedensity and histomorphometric parameters in cancellousbone of the proximal femur in five mammalian species.J Orthop Res 14, 972–979 (1996).

Nipkow P: Elektrisches Teleskop. Patentschrift 30105,Kaiserliches Patentamt, Germany (1884).

Petran M, Hadravsky M: Method and arrangement forimproving the resolving power and contrast. USA(1966). United States Patent No. 3,517,980, priority05.12.1966, patented 30.06.1970 US.

Riggs C, Vaughan L, Evans G, Lanyon L, Boyde A:Mechanical implications of collagen fibre orientation incortical bone of the equine radius. Anat Embryol 187,239–248 (1993).

Simkin A, Robin G: Fracture formation in differing collagenfiber pattern of compact bone. J Biomech 7, 183–188(1974).

Simmons AH: Faunal Extinction in an Island Society: PygmyHippopotamus Hunters of Cyprus. Kluwer Academic/Plenum, New York (1999).

Smith RJ, Cheverud JM: Scaling of sexual dimorphism inbody mass: a phylogenetic analysis of Rensch’s rule inprimates. Int J Primatol 23, 1095–1135 (2002).

10 SCANNING VOL. 31, 0 (2009)