Scaling of the first ethmoturbinal in nocturnal strepsirrhines: Olfactory and respiratory surfaces

Scaling of the First Ethmoturbinal inNocturnal Strepsirrhines: Olfactory and

Respiratory SurfacesTIMOTHY D. SMITH,1,2* KUNWAR P. BHATNAGAR,3 JAMES B. ROSSIE,4

BETH A. DOCHERTY,5 ANNE M. BURROWS,2,5 GREGORY M. COOPER,6,7

M.P. MOONEY,2,8 AND M.I. SIEGEL2

1School of Physical Therapy, Slippery Rock University, Slippery Rock, Pennsylvania2Department of Anthropology, University of Pittsburgh, Pittsburgh, Pennsylvania3Department of Anatomical Sciences and Neurobiology, University of Louisville

School of Medicine, Louisville, Kentucky4Department of Anthropology, State University of New York, Stony Brook, New York5Department of Physical Therapy, Duquesne University, Pittsburgh, Pennsylvania

6Department of Orthopedic Surgery, University of Pittsburgh, Pittsburgh, Pennsylvania7Department of Plastic Surgery, University of Pittsburgh, Pittsburgh, Pennsylvania8Department of Oral Biology, University of Pittsburgh, Pittsburgh, Pennsylvania

ABSTRACTTurbinals (scroll bones, turbinates) are projections from the lateral wall of

the nasal fossa. These bones vary from simple folds to branching scrolls. Con-ventionally, maxilloturbinals comprise the respiratory turbinals, whereas naso-turbinals and ethmoturbinals comprise olfactory turbinals, denoting the pri-mary type of mucosa that lines these conchae. However, the first ethmoturbinal(ETI) appears exceptional in the variability of it mucosal covering. Recently, itwas suggested that the distribution of respiratory versus olfactory mucosaevaries based on body size or age in strepsirrhine primates (lemurs and lorises).The present study was undertaken to determine how the rostrocaudal distribu-tion of olfactory epithelium (OE) versus non-OE scales relative to palatal lengthin strepsirrhines. Serially sectioned heads of 20 strepsirrhines (10 neonates, 10adults) were examined for presence of OE on ETI, rostral to its attachment tothe nasal fossa wall (lateral root). Based on known distances between sectionsof ETI, the rostrocaudal length of OEwasmeasured and compared to the lengthlined solely by non-OE (primarily respiratory epithelium). In 13 specimens, thetotal surface area of OE versus non-OE was calculated. Results show that thelength of non-OE scales nearly isometrically with cranial length, while OE ismore negatively allometric. In surface area, a lesser percentage of non-OEexists in smaller species than larger species and between neonates and adults.Such results are consistent with recent suggestions that the olfactory structuresdo not scale closely with body size, whereas respiratory structures (e.g., maxillo-turbinals)may scale close to isometry. In primates and perhaps othermammals,variation in ETI morphology may reflect dual adaptations for olfaction andendothermy. Anat Rec, 290:215–237, 2007. � 2007Wiley-Liss, Inc.

Key words: concha; ethmoturbinate; olfactory epithelium; pri-mate; respiratory epithelium

Among vertebrates, mammals possess the most com-plex nasal fossae in terms of architecture. Internally, asequence of thin osseous sheets called turbinals (turbi-nates, conchae) subdivides the airways into major com-partments. At their greatest complexity, these turbinalsdictate a maze of minor air passageways due to repeatedfolding, scrolling, and/or branching of the turbinal root,

*Correspondence to: Timothy D. Smith, School of PhysicalTherapy, Slippery Rock University, Slippery Rock, PA 16057.Fax: 724-738-2113. E-mail: [email protected]

Received 25 August 2006; Accepted 1 December 2006

DOI 10.1002/ar.20428Published online 15 February 2007 in Wiley InterScience (www.interscience.wiley.com).

� 2007 WILEY-LISS, INC.

THE ANATOMICAL RECORD 290:215–237 (2007)

or primary lamina. Correspondingly, inspired air isexposed to a vast expanse of nasal mucosa (Negus, 1958;Moore, 1981).Excluding transitional epithelia, several types of epi-

thelia line the nasal fossa: respiratory, olfactory, andstratified epithelia. These surface membranes, alongwith underlying lamina propria, provide air condition-ing, olfactory chemoreception, and fluid transport func-tions (Menco and Morrison, 2003; Harkema et al., 2006).These tissues are the primary functional components ofthe turbinals, and therefore the turbinals themselvesare seen as evolutionary innovations for augmentingolfaction and air warming/humidification (Negus, 1958;Moore, 1981; Hillenius, 1992, 1994). The predominantlyrespiratory function of the maxilloturbinal (often synon-ymous with the term ‘‘respiratory turbinate’’) has beenconsistently described (Negus, 1958; Hillenius, 1992). Incontrast, the more caudally positioned ethmoturbinals(olfactory turbinates) are described to be mostly linedwith olfactory mucosa (Le Gros Clark, 1959; Moore,1981; Novacek, 1993).The mucosal coverings of turbinals provide the basis

for functional interpretations of the bones themselves.Based on its glandular vascular mucosal coverings, themaxilloturbinal is of special interest for its role in theregulation of endothermy (Hillenius, 1992, 1994; VanValkenburgh et al., 2004). The remaining turbinals bearboth respiratory and olfactory mucosa, and a number ofauthors have urged further detailed study of the extentof each mucosa (Ankel-Simons, 2000; Smith et al., 2004).Such knowledge is critical to interpretation of osteologi-cal or computed tomographic (CT) data on individualturbinals (Van Valkenburgh et al., 2004; Rowe et al.,2005).Ethmoturbinals and nasoturbinals are the only endo-

turbinals to bear olfactory mucosa. Most of the ethmo-turbinal complex is sequestered in the dorsocaudal por-tion of the nasal fossa, abutting the recessus cupularisposterior (Smith and Rossie, 2006). In most eutherianmammals, the majority of this turbinal complex is sepa-rated from the respiratory airway (nasopharyngeal mea-tus) by a horizontal plate of bone, the lamina transver-salis posterior. This portion of the ethmoturbinal com-plex, housed within the enclosed olfactory recess, ispredominantly lined with olfactory mucosa in all mam-mals that possess ethmoturbinals (Negus, 1958). Amongthe ethmoturbinals, the first ethmoturbinal (ETI)—ter-minology relating to the first ethmoturbinal varies, andsome authors have combined the nasoturbinal into theethmoturbinal complex, a practice not followed here [seeSmith and Rossie (2006) for detail]—represents a nota-ble exception, as it has been shown to bear a varyingamount of nonolfactory epithelium in diverse mamma-lian taxa (Read, 1908; Negus, 1958; Loo, 1974; Bhatna-gar and Kallen, 1975; Kumar et al., 1993, 2000).Although the dual nature of the mucosa on the ethmo-turbinal complex has long been known, it is not consis-tently acknowledged. Both Read (1908) and Dieulafe(1906) wrote detailed accounts of the mucosal lining ofthe nasal fossa. Read (1908) stated that ‘‘about one-halfof the ethmoturbinal folds are olfactory’’ in the dog andcat (p. 41). In a work describing broader range of mam-mals (including dogs and cats), Dieulafe (1906) stated:‘‘We may say in general that the olfactory mucosa coversthe entire extent of the ethmoturbinals and the grooves

which separate them’’ (p. 294). Such statements areseemingly contradictory, and the degree to which theyreflect misinterpretation versus phylogenetic variabilityis not clear.Smith et al. (2004) reconstructed a mucosal map of

the ethmoturbinals in the mouse lemur (Microcebusmurinus), a strepsirrhine primate of distinguished olfac-tory abilities (Perret, 1992). This species possesses a re-stricted dorsocaudal distribution of olfactory mucosa onETI, while the remainder of the turbinal is primarilylined with respiratory mucosa. Moreover, the extent ofnonolfactory surface apparently increases with age afterbirth. Smith et al. (2004) suggested that as the strepsir-rhine nasal fossae grow in depth and length, dispropor-tionately more respiratory mucosa is added to this turbi-nal, while the olfactory epithelium is augmented to alesser degree. They suggested the differential scaling ofthese mucosae was best explained with an increaseddemand for air conditioning (filtration, moistening,warming) of inspired air as airways and nasal cavitiesenlarge. Since ETI projects rostral to the level of the cri-briform plate, it may be more directly exposed to air cur-rents than the more posterior ethmoturbinals.Although the maxilloturbinal may have a central role

in mammalian respiratory adaptations, the contributionof the ethmoturbinals to air conditioning is virtuallyunexplored. The present study examines the scaling ofmucosa on this turbinal rostral to its basal lamina. Inorder to investigate previous assertions of differentialscaling of mucosa on ETI (Smith et al., 2004), two fami-lies of strepsirrhine primates that exhibit a broad rangeof body sizes, cheirogaleids (dwarf lemurs) and galagids(bushbabies), are studied. Of interest is to what extentthese nocturnal keen-scented primates rely on the ETIto augment olfaction or air conditioning.

MATERIALS AND METHODSSample

The species under study included nocturnal strepsir-rhine primates from the family Cheirogaleidae (Cheiro-galeus medius, Microcebus murinus, Mirza coquereli)and Galagidae (Galagoides demidoff, Galago moholi,Otolemur crassicaudatus, O. garnettii). The adult prima-tes studied have a broad range in body mass from 109 g(M. murinus; Fig. 1a) to 1,495 g (O. crassicaudatus)(Rowe, 1996). Serially sectioned whole or half snouts of20 primates, including 10 neonates and 10 adults, wereexamined in the present study (Table 1). Most of thismaterial was previously prepared as described by Smithet al. (2002, 2004, 2005), and six additional heads wereprepared for this study. In addition, CT scans of threeadditional adult strepsirrhines (one Microcebus murinus,one Otolemur garnettii, and one Loris tardigradus) werereconstructed using Scion Image software (NIH) in orderto describe the spatial relationships of the nasal fossa tothe craniofacial skeleton and the turbinals within thenasal fossa.The salient features of the snout in nocturnal strepsir-

rhines are provided in Figures 1b–d and 2. The first eth-moturbinal is spatially distinguished from the other eth-moturbinals, which are nearly entirely sequestered inan olfactory recess by the transverse lamina that sepa-rates the dorsocaudal end of the nasal fossa (Recessuscupularis) from the nasopharyngeal meatus (Fig. 1b).

216 SMITH ET AL.

ETI projects extensively farther rostrally than other eth-moturbinals and broadly overlaps the maxilloturbinal inthe species under study.

Dissection and Light Microscopic Methods

All heads were stored in 10% buffered formalin afternatural deaths in captivity (specimens from Duke Uni-versity Lemur Center) or after sacrifice following use inan unrelated study (adult bushbabies from Duke Univer-sity Medical Center). The left ethmoturbinal was dis-sected from the contralateral side of one of the adultbushbabies for study using scanning electron microscopy(SEM). This specimen was stored in 10% buffered forma-lin until preparation for SEM. Prior to tissue processing,cranial measurements were taken, including dorsal pala-tal length (the midline distance from the rostral marginof the premaxilla at the base of the nasal aperture tothe caudalmost midline point of the hard palate) andcranial length (prosthion-inion). In addition, the lengthof the nasal fossa was obtained after sectioning, definedas the distance between the most rostral section showing

an enclosed nasal fossa and the end of the recessuscupularis posterior. Due to damage to specimens prior toprocessing (e.g., during necropsy) or after sectioning, itwas not possible to obtain all three measurements in allspecimens.Briefly, heads were prepared for light microscopic ex-

amination as follows. After decalcification (until negativetest with ammonium oxalate test), each was paraffin-em-bedded. Serial sections were made at 10–12 mm and ev-ery 5th (neonates) or 10th (adults) section was mountedon glass slides; intervening sections were saved ormounted for mucin staining. Slides were stained alter-nately with Gomori trichrome or hematoxylin-eosin.Selected intervening sections of each species werestained with Alcian blue-periodic acid-Schiff procedures(Humason, 1979; Roslinski et al., 2000).Slides were examined using light microscopy with a

Leica DMLB photomicroscope at different magnifica-tions. The nasal mucosa was examined in each specimento delimit rostral and caudal limits of olfactory neuroepi-thelium throughout the walls of the nasal fossa. Olfac-tory epithelium was identified based on the presence of

Fig. 1. The nocturnal strepsirrhines studied varied widely in size,including a representative of the smallest genus, (a) the mouse lemur(Microcebus murinus) and the cat-sized genus Otolemur (b). b: Areconstruction of the facial skeleton in an adult O. gametti in anteriorand oblique view. The outline of the facial skeleton is ghosted over thesoft tissues. The nasopharyngeal passageways are lightened and out-lined. Note the olfactory recess (R), positioned between the orbits (Or).c: Schematic view of the nasal fossa and nasopharyngeal passageway(shaded) in a lemur (Eulemur), showing the spatial relationship of thefirst ethmoturbinal (ETI) to the nasoturbinal (NT) and maxilloturbinal(MT), each name based on the bone from which they project. The eth-

moturbinal complex (I–IV) is highlighted within the nasal fossa. Notethat the more posterior ethmoturbinal projections (II–IV) are restrictedwithin the olfactory recess (R), which is separated from the nasopha-ryngeal meatus by a horizontal bony shelf, the lamina transversalisposterior (LTP). d: These same structures are shown in outline over aCT-based reconstruction of Microcebus murinus. Scion Image wasused to extract the turbinals, which were superimposed over thisreconstruction of the entire skull to trace their structure (dashed lines).The LTP is indicated by a solid bar. Note that ETI broadly overlaps theNT and MT. Ph, pharyngeal passageway. Illustration in a is copyrightedby Timothy D. Smith. c is modified after Smith and Rossie (2006).

217SCALING OF FIRST ETHMOTURBINAL

Bowman’s glands and olfactory axons in the laminapropria and an epithelium characterized by the presenceof three cell types: supporting cells (with an apical zoneclear of nuclei), olfactory receptor neurons (arranged instacks of round nuclei), and basal cells (Menco and Mor-rison, 2003). Turbinals were defined anatomically usingthe same system as Le Gros Clark (1959) and Ankel-Simons (2000), that is, the turbinals arising from theethmoid bone in the most medial row were ethmoturbi-nals [called endoturbinals by Novacek (1993) andothers]. Note that the nasoturbinal was not includedamong ethmoturbinals, as in Martin (1990). For furtherdiscussion of the anatomical nomenclature used herein,see Smith and Rossie (2006).

Computer 3D Reconstruction and

Histomorphometrics

For each genus that was represented by an adult andneonate, one specimen was selected for computer-basedthree-dimensional reconstructions of olfactory epithelialdistribution based on ideal preservation and integrity ofnasal cavity structures during sectioning. The ETI of anadult C. medius was partially reconstructed since it pos-sessed clearly identifiable mucosae for most of the turbi-nal (the caudal part of the nasal fossa of this specimenwas damaged during preparation). Reconstructions weremade using Scion Image software (release 4.02, NIH).For acquiring sections for reconstruction, the serial sec-tions of each nasal chamber were digitally photographedusing a Leica DMLB photomicroscope with a DKC-5000Catseye Digital Still Camera System (Sony Electronics,Montvale, NJ). Images were transferred to Adobe Photo-shop 7.0 and saved as TIFF files. Subsequently, individ-ual files for every 20th section (in adults) or 10th section(in infants) was marked according to fiducial landmarks(specific structures or contours) that delimited OE distri-bution along ETI. At the rostral and caudal ends, inter-

vening sections were also marked in adults so that accu-racy was comparable to that used for infants. Themarked digital images were then aligned for reconstruc-tion as described in detail by Smith et al. (2004).For measurement of olfactory surface on ETI, the

unattached rostral projection of ETI was studied as aunit. Figure 2 illustrates this part of the ethmoturbinalcomplex in a 3D reconstruction of the rostral nasal fossain an adult slender loris (Loris tardigradus), a memberof the same superfamily as galagids (Lorisoidea). Thecaudal endpoint was established as the first section inwhich ETI connected to the lateral nasal wall. This con-nection is shown in the most caudal cross-sectional levelshown in Figure 2. These spatial relationships can alsobe viewed online in serial sections of infant mouselemurs and bushbabies (http://www.interscience.wiley.com/jpages/1932-8486/suppmat). For the length or rostro-caudal extent of olfactory vs. nonolfactory surfaces, allspecimens were examined for the most rostral section inwhich olfactory neuroepithelium was present on ETI. Therostrocaudal distance from this point to the attachmentof ETI to the lateral wall was measured by counting theintervening sections and multiplying this number by sec-tional thickness. This distance was subtracted from theoverall length of the unattached projection of ETI inorder to obtain the nonolfactory extent of ETI.Thirteen specimens were suitable for calculation of

surface areas of the mucosa on ETI. The total surfacearea of ETI and the olfactory surface area was calcu-lated using Scion Image software. To obtain this mea-surement, the perimeter of olfactory epithelium wasmeasured in each section after calibrating to a digitalimage of a stage micrometer photographed at the samemagnification used for the images of sections. Every pe-rimeter measured was then recorded in millimeters, andmultiplied by the distance in millimeters to the next sec-tion, yielding a segment of olfactory surface areabetween sections. All segmental measurements were

TABLE 1. Species, sex, age, body mass and percentage of olfactory epithelium on ETI

SpeciesSample/

sex Age

Bodymass(g)a

ETI: % Olfactory surface

Length Area

Galago moholi 1M, 1F Adult 180.5 62.2–77.6 21.8Galago moholi 1 (sex?) Neonate 13.4 84.6 31.9Galagoides demidoff 1F Neonate 9 91.7 DOtolemur garnettii 2M, 1F Adult 730 51.2–55.9 7.0–9.3Otolemur garnettii 1M, 1F Neonate 49 68.5, 72.7 17.8, 21.8Otolemur crassicaudatus 2F Adult 1,115 51.5, 58.2 9.7Otolemur crassicaudatus 1M, 1F Neonate 44.5 79.2, 89.1 22.2Cheirogaleus medius 1F Adult 156 91.8 DCheirogaleus medius 1F Neonate 12 97.2 15.6Microcebus murinus 1M, 1F Adult 68 81.8–87.2 24.0Microcebus murinus 1M Neonate 5.8 95.5 39.4Mirza coquereli 1M, 1F Neonate 15.5 87.1–95.2 41.2

aFrom Kappeler and Pereira (2003); ETI ¼ first ethmoturbinal; M ¼ male; F ¼ female; D ¼poor epithelial preservation or tissue damage prevented accurate measurement.

Fig. 2. Four CT-based reconstructions of a portion of the rostralhalf of the nasal fossa in the slender loris (Loris tardigradus), afterremoval of the septum. Note that the four views are repeated in Fig-ures 9–12. The spatial relationship of the first ethmoturbinal (ETI; or-ange) and maxilloturbinal (MT; blue) is shown. Note that ETI broadly

overlaps MT; in Lorisoids (e.g., Loris, Galago, Otolemur), it reachesto the same level ventrally as MT. The lateral view is also used toindicate five coronal cross-sectional levels, with a level near the ros-tral tip of ETI on the left, and a level indicating the primary lamina ofETI on the right.

218 SMITH ET AL.

Figure 2. (Legend on page 218.)

219SCALING OF FIRST ETHMOTURBINAL

then summed and recorded as total olfactory surfacearea. The same process was repeated on the entire pe-rimeter of ETI in order to obtain total surface area. Ol-factory surface area was subtracted from total surfacearea to obtain nonolfactory surface area.

Scanning Electron Microscopy

One adult greater bushbaby (O. crassicaudatus) wasprepared for study by SEM as follows. The left ethmo-turbinal, which was dissected from the contralateralside of one of the adult bushbabies, was removed from10% buffered formalin and washed in distilled water for5 min. The tissue was dehydrated in a series of ethanolwashes, immersed in two HMDS baths for 15 min, andair-dried at ambient temperature. After air-drying, thesample was mounted on a stainless steel stub and sput-ter-coated with gold in a Blazers MED model 010 TurboMini Deposition System. Secondary electron (SE) imageswere taken with a CamScan Series 4 scanning electronmicroscope operated at an accelerating voltage of 5–20 kV and a working distance of 15 mm. Images wereacquired digitally using Princeton Gamma-Tech’s (PGT)IMIX system. The PGT system was calibrated using anMRS-3 SEM magnification calibration grid (Ted Pella).

Statistical Methods

Statistical analyses were performed after examiningthe data with Kolmogorov-Smirnov tests of normalityusing SPSS version 11.0 software (SPSS). Since tests ofnormality indicated that linear data were nonnormal, alllength data were log10-transformed except for analysisof surface areas. For the portion of the sample in whichall cranial measurements could be made, palatal lengthand nasal fossa length were significantly correlated withprosthion-inion length (r ¼ 0.951, P < 0.001; r ¼ 0.956,P < 0.001, respectively). Because of this and since thepalate was intact in all specimens, palatal length wasused for correlations and regression analyses. Area

measurements were converted to square root values inorder to correlate with palatal length. Since square rootareas were normally distributed, these were analyzedwith nontransformed palatal length.All variables were analyzed using Pearson correlation

tests (Table 2) and regression analyses (Table 3). Signifi-cance was set at P < 0.05. For regression analyses, theequation with the highest r2 value was selected. If twoequations were nearly equally suited to describe therelationship of ETI data to palatal length, the simplestequation was selected.

RESULTSDescriptive Findings

Adults. The rostral projection of ETI broadly over-laps the dorsal surface of the maxilloturbinal (Fig. 3aand b). In coronal cross-section, ETI begins as a nearlyflat or ventrally concave horizontal plate and becomesmore concave in a rostrocaudal sequence (Figs. 3c and4), culminating in an acutely folded structure withmedial and lateral plates hanging ventrally from a nar-row, rounded apex (Fig. 3a). This fold creates inner sur-faces of ETI, where the two plates oppose one another oroverlie the maxilloturbinal, and outer surfaces, wherethe two plates oppose the lateral nasal wall or the nasalseptum (Fig. 3a and b). The medial plate hangs fartherventrally than the lateral plate and more extensivelycovers the maxilloturbinal in the galagids. All along therostrocaudal progression, accessory plates or scrolls arisetransiently from the medial and lateral margins of ETI(Figs. 3c and 4). Within the caudal one-fifth of ETI, butrostral to the attachment to the lateral wall, a horizon-tal plate connects the lateral and medial plates, creatinga mucosal recess. This turbinal recess, which is blind-ended rostrally, expands in diameter in caudal sections.This recess becomes confluent with the nasal cavity nearthe connection of ETI to the lateral wall, caudal to the

TABLE 2. Pearson correlation coefficients ofETI variables with palatal length

ETI variable

log10palatalLength

palatalLength

log10 olfactoryextent

r ¼ 0.957**

log10 non-olfactory extent

r ¼ 0.902**

log10 totalETI length

r ¼ 0.969**

Square root ofolfactory surface area

r ¼ 0.864*

Square root of non-olfactory surface area

r ¼ 0.820*

Square root oftotal ETI area

r ¼ 0.874*

ETI ¼ rostral projection of the first ethmoturbinal; olfactoryextent ¼ the rostrocaudal distance of ETI covered with ol-factory epithelium; non-olfactory extent ¼ the rostrocaudaldistance of ETI with no olfactory epithelium.*¼ P < 0.05; **¼ P < 0.001.

TABLE 3. Regression equations

log10 CL vs. log10 BMY ¼ 1.14 þ (0.226 3 log10BM); R2 ¼ 0.84

log10 PL vs. log10 BMY ¼ 0.613 þ (0.281 3 log10BM); R2 ¼ 0.93

log10 ETI total length vs. log10 BMY ¼ �0.256 þ (0.484 3 log10BM); R2 ¼ 0.94

log10 PL vs. log10 CLY ¼ �0.94 þ (1.33 3 log10CL); R

2 ¼ 0.92log10 ETI total length vs. log10 CL

Y ¼ �2.68 þ (2.13 3 log10CL); R2 ¼ 0.78

log10 olfactory extent of ETI vs. log10 PLY ¼ �0.94 þ (1.3 3 log10PL); R

2 ¼ 0.92log10 non-olfactory extent of ETI vs. log10 PL

Y ¼ �3.5 þ (2.96 3 log10PL); R2 ¼ 0.81

log10 total ETI length vs. log10 PLY ¼ �1.25 þ (1.67 3 log10PL); R

2 ¼ 0.94Square root of olfactory surface area vs. PL

Y ¼ �0.87 þ (0.29 3 PL) þ (�2.75 3PL); R2 ¼ 0.94Square root of non-olfactory surface area vs. PL

Y ¼ �2.56 þ (0.38 3PL) þ (5.05 3 PL); R2 ¼ 0.97Square root of total ETI surface area vs. PL

Y ¼ �3.42 þ (0.67 3PL) þ (2.38 3 PL); R2 ¼ 0.98

BM, body mass; CL, cranial length; ETI, rostral projectionof ethmoturbinal I; PL, palatal length.

220 SMITH ET AL.

Fig. 3. Ethmoturbinal I (ETI) of Otolemur crassicaudatus, adultfemale. After approximately 2 mm, ETI has a folded appearance,with a medial and lateral plate (MP and LP; a and b). Caudally, ETIhas a turbinal recess (TR). c is a rostrocaudal sequence showing thetransition from a flat plate to a folded lamella with a central turbinal

recess. AP, accessory plate; IS, inner surface; MT, maxilloturbinal;NS, nasal septum; NT, nasoturbinal; OS, outer surface; VNO,vomeronasal organ. Dashed line denotes olfactory surface. Numbersin c refer to serial section numbers; sections at 12 mm. Scale bars ¼1 mm.

221SCALING OF FIRST ETHMOTURBINAL

attachment in galagids (Fig. 3c) and rostral to theattachment in cheirogaleids (Fig. 4). The lateral plate ofETI attaches to a horizontal process [‘‘lateral root:’’ seeSmith and Rossie (2006)] projecting from the lateralwall of the nasal fossa. More caudally, ETI attaches tothe nasoturbinal (dorsal root).In all species except M. murinus, olfactory neuroepi-

thelium is restricted to dorsolateral and dorsomedialcontours of the outer surfaces (Figs. 3c, 4, 5a, and 6a).The olfactory mucosa of ETI is typical in that olfactoryepithelium is greater in thickness than adjacent respi-ratory epithelium and has numerous alcian blueþ

Bowmans’s glands in the lamina propria. (Fig. 5a andd). The remainder of the turbinal is nearly completelycovered in respiratory mucosa, which comprises pseudo-stratified ciliated columnar epithelium and a vascular

lamina propria (Figs. 5a–c and f, 6, and 7). In contrastto the maxilloturbinal, subepithelial glands are rare inthe lamina propria of ETI (Fig. 5c). However, ETI pos-sesses more abundant goblet cells than the maxilloturbi-nal (Figs. 5c, e, and f and 6b) and intraepithelial glands(Fig. 6c). Aside from the respiratory and olfactory epi-thelium (Fig. 6b–g), only minute patches of other epithe-lial types (e.g., stratified) are found on the apices of mu-cosal folds. Observations of a single adult O. crassicau-datus by SEM confirm the extensive distribution ofciliated pseudostratified columnar epithelium. Figure 8shows that along the rostral half of the unattached pro-jection of ETI, respiratory cilia are clearly visible acrossthe dorsal surface.In all adult primates, venous sinuses are found in the

lamina propria and are especially abundant in more

Fig. 4. Ethmoturbinal I (ETI) of Microcebus murinus, adult male. Arostrocaudal sequence showing the transition from a flat plate to afolded lamella with a central turbinal recess. ETI has a similarly foldedappearance in cheirogaleids, with medial and lateral plates (MP andLP). Note that the lateral plate hangs farther ventrally than the medial

plate, the opposite of that observed for galagids. Caudally, note thatthe turbinal recess (TR) is lined in part with olfactory mucosa (dashedline). AP, accessory plate. Dashed line denotes olfactory surface.Scale bar ¼ 1 mm.

Fig. 5. Mucosal characteristics of ETI in an adult female Otolemurcrassicaudatus (a–c; Gomori trichrome) and an adult female Otolemurgarnettii (d–f; alcian blue-periodic acid-Schiff). The lateral and medialplates (LP and MP) of ETI had a vascular mucosa, possessing venoussinuses (VS; a–d and f) as well as intraepithelial unicellular (goblet) ormulticellar mucous glands (short arrows, e and f). The venous sinuseswere most pronounced in the specimen shown in a–c. Olfactory neu-roepithelium (ON) was limited to dorsal, dorsolateral, and dorsomedial

surfaces of ETI in galagids (a and d). Alcian blueþ Bowman’s glands(BG) were seen deep to the ON (d). Most of ETI was lined with glan-dular pseudostratified ciliated columnar epithelium (f; open arrowsindicate alcian blue þ rows of cilia. Subepithelial glands were rare inETI, but were seen in the maxilloturbinal. e shows the ventralmost partof the MP; f shows the lateral surface of ETI (lateral is up in this imageonly). MT, maxilloturbinal. Scale bars ¼ 1 mm (a); 600 mm (b); 150 mm(c); 100 mm (d); 300 mm (e); 150 mm (f).

222 SMITH ET AL.

Figure 5. (Legend on page 222.)

223SCALING OF FIRST ETHMOTURBINAL

ventral parts of ETI. One adult bushbaby (O. crassicau-datus) possesses dilated venous sinuses throughout allof the mucosal tissues of the ventral nasal fossa andclearly shows copious sinus distribution along medialand lateral plates (Fig. 5a–c). In cheirogaleids, venoussinuses are most abundant ventrally as well and thusare distributed in line with the nasopharyngeal meatusalong with the maxilloturbinal (Fig. 7). Figure 7b–dshows a cross-section just rostral to the root of ETI inM. murinus. ETI has distinctly different dorsal (olfac-tory; Fig. 7c) and ventral (respiratory; Fig. 7d) surfaces.The sinuses form an arc around the inferior surfaces ofETI that abut the maxilloturbinal. In the vicinity ofETI, venous sinuses are also plentiful in the maxillotur-

binal and, to a lesser extent, the ventral part of thenasal septum (Figs. 5a and 7d).Three-dimensional reconstructions of representative

adults from each species (Figs. 9–12) illustrate that ol-factory neuroepithelium is most broadly distributed inthe caudal one-half of ETI and is extremely restrictedrostrally. In cross-section, nonolfactory mucosa (almostexclusively respiratory type) lines more surface areathan olfactory mucosa at all rostrocaudal levels. Ros-trally, olfactory mucosa first appears as a small dorsal ordorsomedial patch. This strip reaches proportionatelyfarther to the rostral end of ETI in smaller species (Figs.10 and 12) than larger species (Figs. 9 and 11). In thisrespect, Otolemur spp. represent an extreme, since olfac-

Fig. 6. Highly magnified views of the mucosa along ETI of Otole-mur garnettii (adult female). a shows a dorsal view of ETI with the ol-factory surface indicated in green. Two different levels are indicated:level 1 indicates a section 25% caudal to the rostral tip; level 2 is atthe midpoint of ETI. The caudal attachment of ETI to the lateral wall isindicated by the root (R). At level 1 (b–d), only respiratory mucosa isfound, including both the inner (b) and the outer (c and d) surfaces.The surface epithelium was ciliated (open arrows) with goblet cells (G)as well as multicellular intraepithelial glands (Gl). The lamina propria

had few or no glands but had large venous sinuses (vs). At level 2 (e–g), olfactory epithelium was found dorsally, dorsolaterally (e), and dor-somedially (f), easily identifiable based on multiple rows of receptorneuron nuclei (RN) with a single apical row of supporting cell nuclei(SC). Olfactory epithelium was supported by a lamina propria withlarge Bowman’s glands (BG). Note the junction with respiratory epithe-lium shown in e (cilia, open arrows). Ventrally at this level, respiratorymucosa was found (g). Scale bar in b ¼ 20 mm for b–g.

224 SMITH ET AL.

tory mucosa barely projects beyond the midpoint of ETI(Figs. 3c, 6a, and 9). In rostrocaudal sequence, the olfac-tory mucosa extends first to the medial plate and thento the lateral plate, thus forming a dorsal arch along theouter surface of ETI (Figs. 9–12). Olfactory mucosaextends farther ventral on the outer surface of themedial plate, but on both plates, nonolfactory mucosacovers one-half or more of surface area.Based on three-dimensional reconstructions, a negative

relationship between palatal length and olfactory mucosatypify both galagids and cheirogaleids. The larger galagidhas much less extensive olfactory surface area (propor-tionately), restricted to little more than half the rostrocau-dal length of ETI and less than half of the vertical heightof the medial and lateral plates (Fig. 9). The full extent ofdifferences between the cheirogaleid genera is not yetclear due to damage to the caudalmost part of ETI in theadult C. medius, but the same trend is evident. The largerC. medius has nearly the same rostrocaudal extent of ol-

factory area as M. murinus, but olfactory area was farless extensive vertically (Figs. 11 and 12). This appears tobe due to proportionately greater dorsoventral dimensionsof the lateral plate in C. medius (Fig. 11). In addition tohaving the greatest rostrocaudal olfactory surface amongall adult primates in this study, M. murinus has thegreatest extent of olfactory mucosa within the turbinalrecess. Most of this space is a thin respiratory mucosa,but caudally the roof of the turbinal recess is lined witholfactory neuroepithelium (Figs. 4 and 12). A proportion-ally smaller amount of olfactory mucosa is seen in the tur-binal recess of C. medius, and none was found in the tur-binal recesses of galagids.

Infants. The rostral projection of ETI in infant pri-mates is less complex than adults in all species. In par-ticular, the medial plate is of smaller proportions verti-cally, particularly in cheirogaleids, where it does not pro-ject ventral to the turbinal recess (Fig. 13).

Fig. 7. Position of ETI and its mucosal characteristics in adult M.murinus. a: A rostrocaudal sequence of cross-sections through thenasal fossa, with serial section numbers indicated below. From left toright, these sections show ETI near the rostral tip (440), at the mid-point of ETI (660), near the attachment of ETI to the lateral wall (880),and the beginning of the nasopharyngeal meatus (NPM) and olfactoryrecess (R; 1250). Note that ETI is suspended ventrally so that its lowersurfaces are in line with the NPM and upper surfaces are in line withthe olfactory recess (all sections are aligned). b: A cross-section

through the nasal fossa just rostral to section 880 in a, showing therelationship of the ventral portion of ETI to the maxilloturbinal. c is amagnified view of the dorsal part of b, in which olfactory epithelium(arrows) was the predominant lining of ETI and adjacent structures. dis a magnified view of the ventral part of b, in which a highly vascularrespiratory mucosa lined ETI. Note that numerous venous sinuses (VS)occur in the mucosa of ETI in surfaces that are adjacent to the vascu-lar mucosa of MT. Scale bars ¼ 1 mm (a); 400 mm (b); 200 mm (c andd).

225SCALING OF FIRST ETHMOTURBINAL

Corresponding to the proportionally compressed ETIdorsoventral height, olfactory neuroepithelium moreextensively covers the outer surface of ETI in all infants(Fig. 13). However, the same trend shown in adults isseen in infants. Primates with larger infants possess agreater rostrocaudal extent of olfactory area. This dis-parity is greatest in galagids, where Otolemur spp. hasa dorsocaudally restricted olfactory area, although farless so than in adults (compare to Fig. 9).

Quantitative and Statistical Findings

Transformed (log10) cranial length, palatal length,and ETI length were highly correlated with log10 bodymass (r ¼ 0.919, 0.966, and 0.969, respectively). The

latter values were averaged from a range of bodymass data provided in Kappeler and Pereira (2003) andthen log10-transformed. ETI length was more positivelyallometric than either cranial length or palatal length(Fig. 14a, Table 1). All ETI measurements are highlycorrelated with palatal length. The lowest r-values are0.844 for the nonolfactory extent of ETI and 0.820 fornonolfactory area (Table 2). For a subset of the samplein which all measurements could be acquired, regressionline slopes show that both palatal length (as a proxy forsnout length) and total ETI length scale with positive al-lometry to cranial length (Table 3). ETI length exhibitsa higher slope, indicating that longer snouts have pro-portionally longer rostral extensions of ETI (Fig. 14b,Table 3).

Fig. 8. a: Scanning electron microscopic images of the dorsal sur-face of ETI in an adult Otolemur crassicaudatus. b shows a detail froma. Note three boxes indicating a series of locations, from lateral tomedial, that are magnified in c–e. c–e: Series of lateral (c), intermedi-

ate (d), and medial (e) locations along the dorsal surface of ETI. Clus-ters of cilia (Ci) are indicated. Scale bars ¼ 200 mm (a); 100 mm (b); 5mm (c); 3 mm (d); 5 mm (e).

226 SMITH ET AL.

Fig. 9. Computer three-dimensional reconstruction: four views of ETI in an adult greater bushbaby (Oto-lemur garnettii). The orientation of each turbinal is indicated by the rotating primate skull, positioned overeach view. Note the olfactory surface, indicated in blue. Nearly the entire nonolfactory surface (orange) wasrespiratory mucosa. MP, medial plate of ETI; TR, turbinal recess; R, root (primary lamella) of ETI.

Fig. 10. Computer three-dimensional reconstruction: four views of ETI in an adult lesser bushbaby(Galago moholi). For orientation relative to total cranium, see Figure 9. Note the olfactory surface, indi-cated in blue. MP, medial plate of ETI; TR, turbinal recess; R, root of ETI.

Fig. 11. Computer three-dimensional reconstruction: four views of ETI in an adult dwarf lemur (Cheiro-galeus medius). For orientation relative to total cranium, see Figure 9. The caudal most portions of ETIwere damaged in this specimen and are not shown in the reconstruction. Note the olfactory surface, indi-cated in blue. LP, lateral plate of ETI.

Fig. 12. Computer three-dimensional reconstruction: four views of ETI in an adult mouse lemur (Micro-cebus murinus). For orientation relative to total cranium, see Figure 9. Note the olfactory surface, indi-cated in blue. LP, lateral plate of ETI; MP, medial plate of ETI; R, root of ETI.

Fig. 13. Computer three-dimensional reconstruction: four views of ETI in neonatal strepsirrhines. Notethe olfactory surface, indicated in blue. Nearly the entire nonolfactory surface (orange) was respiratorymucosa. MP, medial plate of ETI; R, root of ETI.

In rostrocaudal distance, the total length of the rostralprojection of ETI, the extent covered by olfactory mu-cosa, and the length covered only by nonolfactory mu-cosa all increase with palatal length. The rostral nonol-factory extent of ETI scales closest to isometry with pal-atal length; the length of ETI and the extent coveredonly by OE are more negatively allometric by compari-son (Fig. 14c, Table 3). Corresponding to three-dimen-sional reconstructions, data on the square root of surfacearea indicate a lesser percentage of nonolfactory epithe-lium exists in smaller versus larger species and betweenneonates and adults (Fig. 14d). Relationships betweenvariables are best described by second-order regressionequations. Both total surface area and nonolfactory sur-face area are more positively allometric compared to ol-factory surface area (Fig. 14d; see regression line slopesin Table 3). Moreover, the olfactory surface area is bestdescribed by a second-order curve with a decaying slope,opposite to the regression lines for total surface areaand nonolfactory area (Fig. 14d, Table 3).A comparison of adults to neonates supports visual

impressions. In each species, a proportionally greaterpercentage of ETI surface area is covered with nonolfac-tory mucosa in adults compared to neonates (Fig. 15). In

galagids, this disparity between neonates and adultswas more than 10% for linear measurements (Table 1).For area data, the disparity was greater than 10% insome cases. Cheirogaleus medius had the least differ-ence in linear measurements between adults and neo-nates. Measurement of area data was not possible inthis species, but it appeared that in C. medius the respi-ratory mucosa may expand greatly in a different, morevertical dimension (compare Figs. 11 and 12).

DISCUSSION

Certain nocturnal strepsirrhines are regarded as appro-priate ecological models for stem primates (Fig. 1a) (e.g.,Cartmill, 1974; Gebo, 1988). Nocturnality may limit theutility of visual foraging and communication, but strepsir-rhines are presumed to compensate with a keen sense ofsmell. Hence, strepsirrhines have often been consideredthe macrosmats among primates (Negus, 1958).Although at one time olfactory reduction was sug-

gested as a defining characteristic of the order Primates(Elliot Smith, 1927), morphological and behavioral evi-dence indicates that this is overly simplistic (Schaal andPorter, 1991; Laska et al., 2000, 2006; Smith and Bhat-

Fig. 14. Scatter plots with best-fit regression lines for linear meas-urements relative to body mass (a), palatal length and ETI length rela-tive to cranial length (b), ETI linear variables relative to palatal length

(c), and square root of ETI area variables relative to palatal length (d).All linear measurements were log10-transformed except for the bottomright graph (see text). Open symbols, neonates; filled symbols, adults.

232 SMITH ET AL.

nagar, 2004; Smith and Rossie, 2006). The ethmoturbi-nal complex of primates is simpler than in some othermammals (Negus, 1958; Le Gros Clark, 1959), but thedisparity is greater for haplorhines (monkeys, apes,humans, tarsiers) than for strepsirrhines. Morphologi-cally, the latter appear to share many primitive nasalcharacteristics with basal primates and nonprimates(Smith and Rossie, 2006). Though the breadth of thesnout in strepsirrhines is somewhat reduced by theirfrontated convergent (moderately) orbits, these animalsstill possess spacious nasal cavities. More significantly,they possess a sequestered space in the dorsocaudal as-pect of each nasal fossa, termed the olfactory recess. Asin other eutherian mammals, these bilateral culs-de-sacand ethmoturbinals that reside within them are nearlycompletely lined with olfactory mucosa in mouse lemursand bushbabies (Smith et al., 2004). The ample olfactoryarea covering the ethmoturbinal complex in nocturnalstrepsirrhines accords well with the experimental andbehavioral evidence for their olfactory capabilities(Perret, 1992; Watson et al., 1999; Joly et al., 2004). Yetthe first ethmoturbinal, which has an extensive lamellarcomplex that projects rostral to the olfactory recess, isalso covered extensively by nonolfactory mucosa (Loo,1974; Smith et al., 2004).

Smith et al. (2004) described age-related differences inthe extent of nonolfactory mucosa that covered ETI andother regions of the nasal fossa. The present study con-firms their findings. Using a larger sample with a broadage range, these data indicate that primates of largerbody size, either according to species or age, bear agreater proportion of nonolfactory epithelium on ETIthan smaller species. Detailed histomorphology of ETI inthese strepsirrhines appears to reflect dual adaptationsfor olfaction and endothermy.

What Are Turbinals For?

Conventionally, the turbinals are referred to in ana-tomical nomenclature according to their osseous attach-ments in the adult. Other anatomical terms denote posi-tional associations (endoturbinal or ectoturbinal), andthese are discussed elsewhere (Moore, 1981; Smith andRossie, 2006). Another nominal distinction concerns tur-binal physiology. In this nomenclature, the turbinals areascribed respiratory or olfactory functions (Moore, 1981;Hillenius, 1992, 1994). Moore (1981) states that ethmo-turbinals are ‘‘generally coextensive with olfactory mu-cosa’’ (p. 243). Although the ethmoturbinals do bear ol-factory epithelium in all mammals described to date,

Fig. 15. Age changes in the percentage of ETI length lined with ol-factory epithelium. Adult G. moholi, O. garnettii, and M. murinus andinfant O. garnettii and O. crassicaudatus represent the average per-centage from more than one specimen. Note that the same trend

exists in all species: adults have a proportionately lesser rostrocaudalextent of ETI lined with olfactory epithelium. The least disparity wasseen in C. medius.

233SCALING OF FIRST ETHMOTURBINAL

they are also lined with respiratory epithelium. Descrip-tions of diverse taxa (ungulates, carnivores, bats, insecti-vores, primates, and marsupials) indicate at least somerespiratory mucosa covers the portion of ethmoturbinalsthat projects rostral to the cribriform plate (as opposedto the ethmoturbinal sequestered in the olfactory recess)(Read, 1908; Negus, 1958; Loo, 1974; Bhatnagar andKallen, 1975; Kumar et al., 1993, 2000; Rowe et al.,2005). The degree of interspecific variation is unclear atthis time due to the paucity of quantitative studies.Although discretely specialized, the functional muco-

sae of the nasal fossa do not operate in total autonomy.Numerous authors have discussed a protective functionof respiratory epithelium that probably augments olfac-tory function. It is thought that mucociliary filtration ofair prevents infection or accumulation of particles on theolfactory epithelium, which lacks motile cilia (Lucas,1932; Bang, 1961; Shepherd, 2004). Conversely, Shep-herd (2004) suggested that respiratory epithelium mayalso decrease the number of odorant molecules thatreach olfactory epithelium.In addition to regulating air quality, turbinals regu-

late airflow. Beginning with the external nares, the tur-binals are part of a larger suite of nasal fossa contoursthat dictate the flow of air throughout the nasal fossa(Wilson and Sullivan, 1999; Settles, 2005). In most mam-mals, the nasoturbinal and maxilloturbinal dominatethe rostral half of the nasal fossa, subdividing the airpassageways into ventral, middle, and dorsal meatuses(Evans and Christensen, 1979). ETI barely projects tothe midpoint of the nasal fossa in most mammals(Negus, 1958; Clancy et al., 1994; Clifford and Witmer,2004; but see ETI of the opossum in Rowe et al., 2005);however, in cheirogaleids and galagids, ETI projects intothe rostral 25% of the nasal fossa (Smith et al., 2004).Le Gros Clark (1959) also noted that ETI is excessivelybroad in lorisoids (lorises and bushbabies) and over-lapped the maxilloturbinal. He refrained from speculat-ing about the functional significance of this arrange-ment. The findings of the present study suggest thatETI of galagids and cheirogaleids invades the middlemeatus and subdivides it into two passages in a mannerunlike other mammals described to date. Based on theflattened shape of the rostral end of ETI in both groups,air currents could presumably be directed dorsal andventral to ETI, perhaps depending on velocity of inspira-tion. An investigation of airflow patterns throughout thenasal fossa of these primates may be invaluable. At thistime, the distribution of mucosa provides a preliminaryidea of the adaptations.

Variations in Turbinal Shape and Mucosa

In terms of complexity, Negus (1958) describes themaxilloturbinal to range from a single scroll (e.g.,humans) to branching (e.g., carnivores) morphology.Negus (1958) and Hillenius (1994) regard heightenedcomplexity of the maxilloturbinal as a means for increas-ing mucosal surface area for air conditioning. In lemursand bushbabies, the maxilloturbinal acquires an inter-mediate complexity as a double scroll (Fig. 3a).In keeping with his functional assessment, Moore

(1981) considered ethmoturbinal branching and folding tobe a macrosmatic trait. Although this is clearly accuratewithin the olfactory recess itself (Smith et al., 2004), ETI

must be more carefully assessed. The data herein showthat all accessory projections of ETI (in the form of simpleaccessory plates) emanate from the nonolfactory region ofETI. Indeed, the most complex surfaces of ETI, includingthe turbinal recess, descending processes, and accessoryplates, were lined predominantly with ciliated epithelium(Figs. 5 and 6). Based on the large number of venoussinuses, a possible collaborative air-warming functionwith the maxilloturbinal exists (Figs. 5–7). The majorityof the venous sinuses were distributed on the ventral sur-face of ETI in the cheirogaleids and throughout themedial and lateral plates of galagids. Based on the sub-stantial projection of ETI into the middle meatus, we sug-gest the entire space ventral to ETI functions in an analo-gous matter to the ventral meatus of most mammals. Inessence, the relatively simple maxilloturbinals of theseprimates are augmented by adjacent surfaces of ETI. Inthis context, it is noteworthy that in those strepsirrhinesthat lack a double-scrolled maxilloturbinal (e.g., lorises, inwhich the dorsal scroll is missing), the ventral projectionsof ETI are correspondingly enlarged (Kollmann andPapin, 1925), suggesting a structural trade-off betweenfunctionally equivalent elements.ETI may specialize in filtration of inspired air in addi-

tion to air warming in nocturnal strepsirrhine primates.The rostral projection of ETI has clear distinctions com-pared to adjacent mucosal surfaces. Both light micro-scopic and SEM observations indicate that the nonolfac-tory surface is nearly uniformly ciliated. This is in con-trast to the maxilloturbinal, which was only partlyciliated (data not shown). In addition, more numerousgoblets cells or intraepithelial glands are found alongthe respiratory surface of ETI compared to the adjacentmaxilloturbinal or nasal septal mucosa. Thus, ETI maybe important for producing acidic (ABþ) mucins, whichare secreted on other nasal surfaces (e.g., nasal septalmucosa) in some other mammals (Roslinski et al., 2000).Interspecific variation in mucosal covering is documented

for other turbinals. For instance, the guinea pig and rabbitpossess nonciliated maxilloturbinals according to Negus(1958; however, regarding rabbit ciliary streams, see Lucasand Douglas, 1934). In wild mice (Peromyscus spp.), Adams(1972) figured a nonciliated cuboidal epithelial covering ofthe maxilloturbinal. Thus, in different mammals, the max-illoturbinal and ETI are covered with ciliated epithelium toa different extent. These attributes suggest that greatercare is required in assessing functional attributes of iso-lated units of the turbinal complex, especially if a broadertaxonomic understanding is sought. The maxilloturbinalmay represent a fundamental respiratory adaptation incarnivores (Hillenius, 1992; Van Valkenburgh et al., 2004),but its function may be augmented to a greater degree byother nasal surfaces in some other mammals. In the noc-turnal strepsirrhines, respiratory surfaces of ETI may com-pensate for relatively simple maxilloturbinal morphology: asingle scroll in some taxa (Kollmann and Papin, 1925).Since the primate fossil record is poor with respect to turbi-nal morphology (Smith and Rossie, 2006), it is presentlyunclear whether a proportionately enlarged ETI is a primi-tive or derived morphology for nocturnal strepsirrhines.

Factors Influencing Turbinal Size

Since ethmoturbinals are not solely lined with olfac-tory mucosa in nocturnal strepsirrhines, their mor-

234 SMITH ET AL.

phology may derive from selective forces operating onboth chemoreception and respiratory functions. Selec-tive factors influencing the size of olfactory structuresinclude activity pattern (nocturnality vs. diurnality)and diet (Healy and Guilford, 1990; Barton et al.,1995; Alport, 2004; Barton, 2006). Although olfactoryepithelium and olfactory bulbs do increase with cra-nial or body size (this study; Stephan et al., 1981; Bar-ton et al., 1995), both are negatively allometric (Smithand Bhatnagar, 2004; Smith et al., 2004). Differencesin olfactory abilities reflect differences in number ofreceptor neurons, in addition to characteristics of thereceptor neuron dendrites (Schoenfeld and Knott, 2004).Since receptor neurons do not scale predictably withbody size, isometry is not expected. Olfactory neuro-epithelial surface area, therefore, may be expected tovary in a manner that is largely free of body size con-straints.The maxilloturbinal provides a different frame of ref-

erence for understanding ETI scaling, since it has beena pivotal nasal structure in understanding respiratoryadaptations for endothermy (Hillenius, 1992; Hilleniusand Ruben, 2004; Van Valkenburgh et al., 2004). By reg-ulating the amount of air that contacts mucosal surfa-ces, the maxilloturbinal warms and humidifies inspiredair (Hillenius, 1992, 1994). Its function in filtration ofinspired particles and pathogens that are trapped in sur-face mucous may be equally critical (Lucas, 1932; Negus,1958; Bang, 1961; Lale et al., 1998).The maxilloturbinal varies in both size (volume)

and complexity. Recent studies suggest that body size,rather than environmental factors (e.g., air humidity,temperature), are the most significant determinantsof maxilloturbinal size and form (Hillenius andRuben, 2004; Van Valkenburgh et al., 2004). Van Val-kenburgh et al. (2004) found that maxilloturbinalswere nearly isometric with body size in felids and can-ids. These authors suggested that maxilloturbinal sizemay relate to tidal volume (volume of air inhaled/exhaled with each normal breath), which in turnrelates to the size of the nasal cavities. This argumentimplies that increased volume of inspired air requiresincreased surface area for humidification and warm-ing. Thus, larger respiratory passageways that trans-mit more inspired air generally possess larger maxil-loturbinals to provide a mucosal interface. The sameargument could apply to filtration function; larger air-ways may necessitate more filtration to impede par-ticles and pathogens from entering the distal respira-tory tract.Within limits, metabolic rates are mass-specific (Suarez,

1995; White and Seymour, 2005) and respiratory meas-ures such as respiratory minute volume (tidal volume Xrespiratory frequency) scale positively in proportion tobody mass (Bide et al., 2000). The maxilloturbinal isknown to scale isometrically with tracheal size (Owerko-wicz and Crompton, 2001), and it follows that the mu-cosa supported by maxilloturbinals, not just the bonesthemselves, would scale in proportion to body size. ETImay be subject to the same scaling factors, but theseshould primarily affect the portion of the ETI that iscovered by nonolfactory epithelium.The present study shows that the nonolfactory surface

of ETI is more positively allometric than olfactory mu-cosa. It must be noted that nonolfactory epithelium

included some patches of nonciliated epithelium, butthese were minor constituents of this measurement.Thus, the high correlation of nonolfactory epithelial areais primarily due to respiratory mucosa. The growth pat-tern of ETI also supports this notion. In all species stud-ied, as ETI lengthened, disproportionately more nonol-factory length was added.A strong correlation to body mass exists for ETI: its

length is more positively allometric with body massthan either snout length or cranial length. Palatallength, a proxy for snout length, is positively allometricwith cranial length; ETI length is more positively allo-metric with cranial length. The regression lines of thecomponents of ETI to palatal length indicate that it isprimarily the nonolfactory portion that drives this allo-metric relationship.The differential scaling of mucosae has implications

for the overall growth of ETI. In surface area, nonolfac-tory mucosa is more positively allometric compared to ol-factory mucosa. Due to the different rates of expansionof its mucosal types, the extent of olfactory area on ETImay act as a constraint to ETI growth. In other words,because olfactory mucosa scales with negative allometry,ETI of strepsirrhines would not be expected to scale sim-ilarly (or grow as much) as the maxilloturbinal. Thisidea should be tested further with other mammals.The ethmoturbinal complex, like the primate snout

itself, is multifunctional (Martin, 1990). The charac-teristics of the primate nasal fossa are sometimes dis-cussed in terms of discrete function in primates. Forinstance, a greater or lesser number of turbinals isused to categorize olfactory abilities (Le Gros Clark,1959; Wako et al., 1999; Ding and Dahl, 2003). In amore general sense, decreasing snout length and com-plexity are identified as a trend in primate evolutionin which respiratory functions increase in emphasiswhile olfactory functions decrease (Siebert and Swin-dler, 2002). As a consequence of the differential scalingof mucosa, quantification of ethmoturbinals or snoutconfiguration may not be a valid criterion for inferringolfactory surface area in CT or osteological studies ofprimates and their ancestors.

ACKNOWLEDGMENTS

The authors are grateful to B.P. Menco for commentsregarding the SEM images, C.J. Bonar for providing theadult loris tissues for CT scanning, H. Ilg and K. Shimpfor preparing some of the serially sectioned specimens,G. Harding for the SEM preparation, A. Walker and T.Ryan for sharing HRXCT scans of Microcebus, T. Bar-bano and T. Rabold for providing CT scans of an adultbushbaby head for Figure 1b, and M.A. Smith for helpin generating movie files of primate nasal fossa. This isa Duke Lemur Center publication no. 1021.

LITERATURE CITED

Adams DR. 1972. Olfactory and non-olfactory epithelia in the nasalcavity of the mouse, Peromyscus. Am J Anat 133:37–50.

Alport LJ. 2004. Comparative analysis of the role of olfaction andthe neocortex in primate intrasexual competition. Anat Rec281A:1182–1189.

235SCALING OF FIRST ETHMOTURBINAL

Ankel-Simons F. 2000. Primate anatomy: an introduction. NewYork: Academic Press.

Bang FB. 1961. Mucociliary function as protective mechanism inupper respiratory tract. Microbiol Mol Biol Rev 25:228–236.

Barton RA, Purvis A, Harvey PH. 1995. Evolutionary radiation ofvisual and olfactory brain systems in primates, bats and insecti-vores. Philos Trans R Soc Lond B Biol Sci 348:381–392.

Barton RA. 2006. Olfactory evolution and behavioral ecology in pri-mates. Am J Primatol 68:545–558.

Bhatnagar KP, Kallen FC. 1975. Quantitative observations on thenasal epithelia and olfactory innervation in bats. Acta Anat91:272–282.

Bide RW, Armour SJ, Yee E. 2000. Allometric respiration/body massdata for animals to be used for estimates of inhalation toxicity toyoung adult humans. J Appl Toxicol 20:273–290.

Cartmill M. 1974. Rethinking primate origins. Science 184:436–443.

Clancy AN, Schoenfeld TA, Forbes WB, Macrides F. 1994. The spa-tial organization of the peripheral olfactory system of the ham-ster: II, receptor surfaces and odorant passageways within thenasal cavity. Brain Res Bull 34:211–241.

Clifford AB, Witmer LM. 2004. Case studies in novel narial anat-omy: III, structure and function of the nasal cavity of saiga(Artiodactyla: Bovidae: Saiga tatarica. J Zool 264:217–230.

Dieulafe L. 1906. Morphology and embryology of the nasal fossae ofvertebrates. Ann Otol Rhinol Laryngol 15:1–584.

Ding X, Dahl AR. 2003. Olfactory mucosa: Composition, enzymaticlocalization and metabolism. In: Doty RL, editor. Handbook ofolfaction and gestation. New York: Marcel Dekker. p 51–73.

Elliot Smith G. 1927. The evolution of man. Oxford: Oxford Univer-sity Press.

Evans HE, Christensen GC. 1979. Anatomy of the dog. Philadel-phia, PA: W.B. Saunders.

Gebo D. 1988. Foot morphology and locomotor adaptation in Eoceneprimates. Folia Primatol 50:3–41.

Harkema JR, Carey SA, Pestka JJ. 2006. The nose revisited: a briefreview of the comparative structure, function, and toxicologic pa-thology of the nasal epithelium. Toxicol Pathol 34:252–269.

Healy S, Guilford T. 1990. Olfactory-bulb size and nocturnality inbirds. Evolution 44:339–346.

Hillenius WJ. 1992. The evolution of nasal turbinates and mamma-lian endothermy. Paleobiology 18:17–29.

Hillenius WJ. 1994. Turbinates in therapsids: evidence for endo-thermy in mammal-like reptiles. Evolution 48:207–229.

Hillenius WJ, Ruben JA. 2004. The evolution of endothermy in ter-restrial vertebrates: who? when? why? Physiol Biochem Zool77:1019–1042.

Humason G. 1979. Animal tissue techniques. San Francisco, CA:W.H. Freeman.

Joly M, Michel B, Deputte B, Verdier JM. 2004. Odor discriminationassessment with an automated olfactometric method in a prosi-mian primate, Microcebus murinus. Physiol Behav 82:325–329.

Kappeler PM, Pereira ME. 2003. Primate life histories and socioe-cology. Chicago: University of Chicago Press.

Kollmann M, Papin L. 1925. Etudes sur lemuriens: anatomie com-paree des fosses nasales et de leurs annexes. Archs Morph GeExp 22:1–60.

Kumar P, Kumar S, Singh Y. 1993. Histological studies on the nasalethmoturbinates of goats. Small Rum Res 11:85–92.

Kumar P, Timoney JF, Southgate HHP, Sheoran AS. 2000. Lightand scanning electron microscopic studies of the nasal turbinatesof the horse. Anat Histol Embryol J Veterinary Med Ser C29:103–109.

Lale AM, Mason JDT, Jones NS. 1998. Mucociliary transport andits assessment: a review. Clin Otolaryngol 23:388–396.

Laska M, Seibt A, Weber A. 2000. ‘‘Microsmatic’’ primates revisited:olfactory sensitivity in the squirrel monkey. Chem Senses 25:47–53.

Laska M, Rivas Bautista RM, Hernandez Salazar LT. 2006. Olfac-tory sensitivity for aliphatic alcohols and aldehydes in spidermonkeys (Ateles geoffroyi). Am J Phys Anthropol 129:112–20.

Le Gros Clark WE. 1959. The antecedents of man. Edinburgh:Edinburgh University Press.

Loo SK. 1974. Comparative study of the histology of the nasal fossain four primates. Folia Primatol 21:290–303.

Lucas AM. 1932. The nasal cavity and direction of fluid byciliary movement in Macacus rhesus (Desm.). Am J Anat 50:141–177.

Lucas AM, Douglas LC. 1934. Principles underlying ciliary activityin the respiratory tract: II, a comparison of nasal clearance inman, monkey, and other mammals. Arch Otolaryngol 20:518–541.

Martin RD. 1990. Primate origins and evolution: a phylogeneticreconstruction. Princeton, NJ: Princeton University Press.

Menco BP, Morrison EE. 2003. Morphology of the mammalian olfac-tory epithelium: form, fine structure, function, and pathology. In:Doty RL, editor. Handbook of olfaction and gustation, 2nd ed.New York: Marcel Dekker. p 17–49.

Moore WJ. 1981. The mammalian skull. London: Cambridge Uni-versity Press.

Negus V. 1958. The comparative anatomy and physiology of thenose and paranasal sinuses. Edinburgh: Livingston.

Novacek MJ. 1993. Patterns of diversity in the mammalian skull.In: Hanken J, Hall BK, editors. The skull, vol. 2. Chicago: Uni-versity of Chicago Press. p 438–545.

Owerkowicz T, Crompton AW. 2001. Allometric scaling of respira-tory turbinates and trachea in mammals and birds. J VertebrPaleontol 21(Suppl):86A.

Perret M. 1992. Environmental and social determinants of sexualfunction in the male lesser mouse lemur (Microcebus murinus).Folia Primatol 59:1–25.

Read EA. 1908. A contribution to the knowledge of the olfactory ap-paratus in dog, cat and man. Am J Anat 8:17–47.

Roslinski DL, Bhatnagar KP, Burrows AM, Smith TD. 2000. Com-parative morphology and histochemistry of glands associated withthe vomeronasal organ in humans, mouse lemurs, and voles.Anat Rec 260:92–101.

Rowe N. 1996. The pictorial guide to the living primates. Charles-town, RI: Pogonias Press.

Rowe TB, Eiting TP, Macrini TE, Ketcham RA. 2005. Organizationof the olfactory and respiratory skeleton in the nose of the grayshort-tailed opossum Monodelphis domestica. J Mamm Evol12:303–336.

Schaal B, Porter RH. 1991: ‘‘Microsmatic humans’’ revisited: thegeneration and perception of chemical signals. Adv Study Behav20:135–199.

Schoenfeld TA, Knott TK. 2004. Evidence for the disproportionatemapping of olfactory airspace onto the main olfactory bulb of thehamster. J Comp Neurol 476:186–201.

Settles GS. 2005. Sniffers: fluid-dynamic sampling for olfactorytrace detection in nature and homeland security. J Fluids Engin127:189–218.

Shepherd GM. 2004. The human sense of smell: are we better thanwe think? PLoS Biol 2:E146.

Siebert JR, Swindler DR. 2002. Evolutionary changes in the mid-face and mandible: establishing the primate form. In: MooneyMP, Siegel MI (editors). Understanding craniofacial anomalies.New York: Wiley-Liss. p 343–378.

Smith TD, Bhatnagar KP, Shimp KL, Kinzinger JH, Bonar CJ, Bur-rows AM, Mooney MP, Siegel MI. 2002. Histological definition ofthe vomeronasal organ in humans and chimpanzees with a com-parison to other primates. Anat Rec 267:166–176.

Smith TD, Bhatnagar KP. 2004. ‘‘Microsmatic’’ primates: reconsider-ing how and when size matters. Anat Rec 279:24–31.

Smith TD, Bhatnagar KP, Tuladhar P, Burrows AM. 2004. Distribu-tion of olfactory epithelium in the primate nasal cavity: are‘‘microsmia’’ and ‘‘macrosmia’’ valid morphological concepts? AnatRec 281:1173–1181.

Smith TD, Bhatnagar KP, Burrows AM, Shimp KL, Dennis JC,Smith MA, Maico-Tan L, Morrison EE. 2005. The vomeronasalorgan of greater bushbabies (Otolemur spp.): species, sex, and agedifferences. J Neurocytol 34:135–147.

Smith TD, Rossie JB. 2006. Primate olfaction: anatomy and evolu-tion. In: Brewer W, Castle D, Pantelis C, editors. Olfaction andthe brain: window to the mind. Cambridge: Cambridge Univer-sity. p 135–166.

236 SMITH ET AL.

Stephan H, Frahm H, Baron G. 1981. New and revised data on vol-umes of brain structures in insectivores and primates. Folia Pri-matol 35:1–29.

Suarez RK. 1996. Upper limits to mass-specific metabolic rates.Ann Rev Physiol 58:583–605.

Van Valkenburgh B, Theodor J, Friscia A, Rowe T. 2004. Respiratoryturbinates of canids and felids: a quantitative comparison. J Zool264:281–293.

Wako K, Hiratsuka H, Katsuta O, Tsuchitani M. 1999. Anatomicalstructure and surface epithelial distribution in the nasal cavity of

the common cotton-eared marmoset (Callithrix jacchus). ExpAnim 48:31–36.

Watson SL, Ward JP, Davis KB, Stavisky RC. 1999. Scent-marking and cortisol response in the small-eared bushbaby (Oto-lemur garnettii). Physiol Behav 66:695–699.

White CR, Seymour RS. 2005. Allometric scaling of mammalian me-tabolism. J Exp Biol 208:1611–1619.

Wilson RM, Sullivan DA. 1999. Respiratory airflow pattern at therat’s snout and an hypothesis regarding its role in olfaction. Phys-iol Behav 66:41–44.

237SCALING OF FIRST ETHMOTURBINAL