MORPHOLOGICAL VARIABILITY AND EVOLUTION OF THE BACULUM (OS PENIS) IN MUSTELIDAE (CARNIVORA)

673

Journal of Mammalogy, 84(2):673–690, 2003

MORPHOLOGICAL VARIABILITY AND EVOLUTION OF THEBACULUM (OS PENIS) IN MUSTELIDAE (CARNIVORA)

GENNADY F. BARYSHNIKOV, OLAF R. P. BININDA-EMONDS, AND ALEXEI V. ABRAMOV*

Zoological Institute, Russian Academy of Sciences, 199034 St. Petersburg, Russia (GFB, AVA)Institute of Evolutionary and Ecological Sciences, Leiden University,

Kaiserstraat 63, P.O. Box 9516, 2300 RA Leiden, The Netherlands (ORPB-E)Present address of ORPB-E: Lehrstuhl fur Tierzucht, Technical University of Munich,

Alte Akademie 12, 85354 Freising-Weihenstephan, Germany

We examined morphological variability and evolution of the baculum (os penis) across theMustelidae through allometric analyses and character mapping. Fifty-four species and 26genera (including 2 fossil forms) were examined with numerous caniform out-group spe-cies. Allometric analyses showed that bacular length is relatively constant across mustelidsand caniforms; only a tendency to a slightly shortened baculum in mephitines was observed.Character mapping revealed the ancestral mustelid baculum to be an elongated rod-shapedbone that lacks a urethral groove and possesses a simple, nondistinct distal tip. This formis largely retained in mephitines and, to a lesser degree, in lutrines. From the ancestralcondition, it is possible to derive forms with a more complicated head that has projectionsand openings (e.g., melines, Eira barbara, Galictis, Gulo gulo, Martes) or spoon-shapedand cup-shaped processes (e.g., Ictonyx, Mellivora capensis). Another evolutionary trajec-tory involves the distal tip of the baculum becoming hook-shaped and the urethral groovewell developed (e.g., Mustela, Vormela peregusna). Although the structure of the baculumdistinguishes closely related species, many features are derived independently in more dis-tantly related forms. Therefore, bacular structure provides restricted phylogenetic infor-mation and should be analyzed in concert with other data sources (e.g., morphology of thebasicranial region).

Key words: allometry, baculum, independent contrasts, morphology, Mustelidae, phylogeny

The baculum (os penis) is a heterotopicbone (Romer and Parsons 1986) that is de-rived from connective tissue and is situateddorsal to the urethra and medial to the cor-pora cavernosa. It is found across a diverserange of mammals—carnivores, bats, ro-dents, insectivores, flying lemurs, and someprimates (Weber 1928)—and often showsgreat morphological diversity among evenclosely related species. This especially istrue among rodents and carnivores.

The reason for the high degree of mor-phological variation is unclear; no single in-terpretation of bacular function seems sat-

* Correspondent: [email protected]

isfactory and well supported (Dixson 1995;Lariviere and Ferguson 2002; Long andFrank 1968). Two primary hypotheses exist(Patterson and Thaeler 1982). One expla-nation is that the variation is simply a pleio-tropic by-product of phylogenetic diver-gence (Burt 1936; Contreras et al. 1993;Hafner 1978; Hamilton 1949). Conversely,other hypotheses explain bacular differ-ences as proceeding from the functionalrole of the os penis in copulation (e.g., asa supporting or stimulating structure—Ewer1973; Long and Frank 1968) and may ex-tend to a potential role as a reproductiveisolating mechanism. For example, Mec-zynski (1974) suggested that the bacular

674 Vol. 84, No. 2JOURNAL OF MAMMALOGY

teeth in ground squirrels (Spermophilus ci-tellus and S. suslicus) apparently interdigi-tate with the rugae vaginalis of the femaleduring copulation, providing functionalspecificity of genital morphology. However,in Carnivora at least, such functional expla-nations do not seem to pertain (Lariviereand Ferguson 2002).

Baryshnikov and Abramov (1997, 1998)described bacular morphology in the ma-jority of genera of Mustelidae. Herein, wepresent a comprehensive comparative anal-ysis of bacular length and morphologyacross all mustelids with a view toward elu-cidating the evolution of the os penis withinthe family. Specifically, we examine the re-lationship between changes in the baculumas an internal heterotopic structure and themajor pathways of adaptive evolution ofmustelids (Sokolov 1968). We also clarifywhether only unique divergence of its struc-tural characters has occurred or whetherthere also have been convergent changes indifferent phylogenetic lineages.

MATERIALS AND METHODS

Data sources.—Data on mustelid bacularlength and morphology were obtained primarilyfrom museum specimens. Three hundred andtwo bacula, representing 46 extant species and22 genera, were examined. Data for 8 additionalmustelid species from 4 genera (including 2 fos-sil forms) were obtained from the literature.Data on the bacular morphology of other speciesof caniforms (‘‘dog-like carnivores’’: canids,mustelids, pinnipeds, procyonids, and ursids)also were collected personally or obtained fromthe literature. A list of all species studied, to-gether with museums visited or literature sourc-es appears in Appendix I. Unless stated other-wise, all observations are for bacula of adults.

Taxonomy of the family follows Wozencraft(1993) except for the following changes. Werecognize 2 species of Meles: the European bad-ger, Meles meles, and the Asian badger, Melesanakuma (Abramov 2001; Baryshnikov 2001;Baryshnikov and Potapova 1990). Also, we rec-ognize Mustela subpalmata as distinct fromMustela nivalis (Abramov and Baryshnikov2000; van Zyll de Jong 1992) and Mustela itatsi

as separate from Mustela sibirica (Abramov2000a; Masuda and Yoshida 1994).

Allometry.—We investigated the allometry ofbacular length in mustelids and in caniforms(Appendix II) using both simple cross-speciesanalysis and the method of independent con-trasts to account for possible phylogenetic ef-fects. Measurements of maximum bacular length(L) without considering the curvature of theshaft were made on pertinent specimens. Bodysize was estimated using condylobasal length(CBL) of the skull of adult males.

Although cross-species analysis provides anaccurate estimate of the regression coefficientfor any relationship (Pagel 1993), it is generallyinvalid because it assumes that species are in-dependent (Harvey and Pagel 1991; Pagel 1993;Purvis et al. 1994). This assumption especiallyis problematic when trying to ascertain function-al reasons for any observed trends. Hierarchicaldescent with modification during evolutionmeans that species that share a common ancestorare likely to be more similar than distantly re-lated species. For example, 2 sister speciesmight share relatively long bacula simply be-cause they both inherited it from their commonancestor. Independent contrasts account for thispotential ‘‘phylogenetic inertia’’ and provide an-swers independent of the phylogenetic relation-ships among species, thereby allowing true func-tional trends to be explored.

Independent contrasts were obtained using the‘‘crunch’’ procedure in the computer programComparative Analysis Using Independent Con-trasts (CAIC—Purvis and Rambaut 1995) basedon the complete and dated (‘‘best estimate’’branch lengths) phylogeny for all extant carni-vores of Bininda-Emonds et al. (1999). Basedon a phylogeny, branch length information, andsome specified predictor variable (CBL here),CAIC estimates nodal values for both the pre-dictor and dependent (L) variables from theirvalues in the node’s descendent taxa and thengenerates ‘‘contrasts’’ (i.e., differences) betweenindependent nodes or species.

We excluded the species Meles anakuma,Mustela itatsi, Mustela subpalmata, Plesiogulomarshalli, and Sardolutra ichnusae from theanalyses because they are not present on thephylogeny of Bininda-Emonds et al. (1999).Other species were excluded because data foreither L or CBL were unavailable: Galictis cuja,Melogale moschata, and Mustela felipei. Both L

May 2003 BARYSHNIKOV ET AL.—BACULAR MORPHOLOGY IN MUSTELIDS 675

and CBL were log-transformed (base e) to ac-cord better with the underlying random walkmodel of evolution used in CAIC (Felsenstein1985; Purvis and Rambaut 1995) and also to re-duce heterogeneity of variances for the regres-sion analysis, improve normality, and convertthe allometric relationship into a predicted linearone (Harvey 1982). Various diagnostic tests in-dicated that this transformation was appropriate,that the branch lengths were standardized ade-quately, and that the Brownian motion modelwas appropriate for the data (Purvis and Ram-baut 1995).

We used cross-species analysis largely to cal-culate ‘‘relative bacular length’’ (with respect toskull length, CBL) for the subsequent character-mapping analysis. Relative bacular length wascharacterized by using the residuals derivedfrom fitting raw species data to the regressionequation obtained using the species data formustelids. Species with residuals .1 SD aboveor below the regression line were held to possessrelatively long or short bacula (compared withthe mustelid average), respectively; the remain-ing species were defined as having bacula of av-erage length. Species with residuals .1.96 SDsabove or below the regression line were held tohave significantly longer or shorter bacula, re-spectively.

Character analysis.—Fifty-two mustelidsspecies were scored, as well as several out-groupspecies (Appendix II) for the following 10 char-acters, described here with their alternativestates (Fig. 1 and Appendix III). Character 1—abrupt upward bend in distal tip: 0, absent; 1,weak or slightly curved; 2, well marked or pro-nounced (Fig. 1a). Character 2—presence of adistal hook: 0, absent; 1, present (Fig. 1b). Char-acter 3—shape of median portion of stem incross section: 0, triangular, dorsal crest well pro-nounced; 1, rounded/triangular, dorsal crest notpronounced; 2, rounded, dorsal crest absent.Character 4—pronounced urethral groove: 0,long; 1, short, only present distally; 2, absent(Fig. 1c). Character 5—pronounced head: 0, ab-sent; 1, slightly pronounced; 2, well marked(Fig. 1d). Character 6—opening on head: 0, ab-sent; 1, present (Fig. 1e). Character 7—subdi-vision of distal tip in ventral plane: 0, absent; 1,present (Fig. 1f). Character 8—shape of distaltip: 0, symmetrical; 1, asymmetrical (Fig. 1g).Character 9—complexity of distal tip: 0, addi-tional projections absent; 1, additional projec-

tions present beyond those formed by subdivi-sion of distal tip (Fig. 1h). Character 10—rela-tive length of baculum (compared with mustelidaverage): 0, shorter; 1, average; 2, longer.

Evolutionary transformations of these char-acters (Appendix III) were examined by map-ping them on the carnivore phylogeny of Bin-inda-Emonds et al. (1999; Fig. 2). This phylog-eny is a supertree (Sanderson et al. 1998) de-rived from 274 individual estimates of carnivorephylogeny (including 31 for mustelids alone); assuch, it probably represents the best current es-timate of the evolutionary relationships of all ex-tant (mustelid) species. For each character, theancestral state for mustelids was reconstructedusing parsimony methods. Except for character10, the out-group consisted of a restricted set ofcaniform species representing all major caniformlineages; for character 10, the ancestral state wasinferred using a broader selection of species fromthe allometry analysis (Appendix II).

RESULTS

Allometry Using Independent Contrasts

Although there are clear differences insize between the different groups, cani-forms are remarkably homogeneous in therelative length of their bacula (Fig. 3a).When ln(L) is regressed on ln(CBL), mostspecies cluster relatively tightly around theregression line, with only a few being .1SD away (Appendix II). However, someclade-specific trends are apparent. Most ca-nid and ursid species fall below the line,whereas most pinniped and procyonid spe-cies are located above the line; however,these trends generally are not significant.Species in the various mustelid subfamiliesare scattered equally on either side of theline, although mephitines do possess rela-tively short bacula compared with othermustelids (Fig. 3b).

Analyzing the species data using inde-pendent contrasts confirmed that a signifi-cant linear relationship exists between thelength of the os penis and size across can-iforms (Fig. 4a). This also revealed signif-icant trends that would have been missed ifonly the raw species data were examined.Although the regression is highly signifi-


FIG. 1.—Characters examined, as shown by representative bacula of mustelids. Bacula are shownin lateral view from right side (lat), ventral view (vent), or as distal tip from left side (dist). a)Character 1, bend of tip—Melogale personata (top, lat); Martes pennanti (middle, lat); Mustelaerminea (bottom, lat). b) Character 2, presence of hook—Eira barbara (top, lat); Mustela eversmanii(bottom, lat). c) Character 4, presence of urethral groove—Lontra canadensis (top, vent); Pteronurabrasiliensis (middle, vent); Mydaus javanensis (bottom, vent). d) Character 5, size of head—Mephitismephitis (top, lat); Martes foina (middle, lat); Meles anakuma (bottom, vent). e) Character 6, presenceof opening in head—Arctonyx collaris (top, vent); Meles meles (bottom, vent). f) Character 7, divisionof tip—Lyncodon patagonicus (top, vent); Lutrogale perspicillata (bottom, vent). g) Character 8,shape of tip—Enhydra lutris (top, vent); Lutra lutra (bottom, vent). h) Character 9, complexity oftip—Ictonyx libyca (top, lat); Martes flavigula (middle, dist); Mustela strigidorsa (bottom, lat). (Notillustrated—character 3, cross-sectional shape and character 10, relative length.)


FIG. 2.—A recent estimate of mustelid phylogeny used for the independent contrasts analysis andcharacter mapping. The tree is a simplified version of the supertree in Bininda-Emonds et al. (1999)showing only those species examined in the current study. Taxa marked with asterisks do not appearin the original study but were placed with their putative sister species based on other sources; theywere included for the character mapping only. Branch lengths are not proportional to time.


FIG. 3.—Regression of ln bacular length onbody size (ln condylobasal length) using rawspecies points a) across caniforms and b) withinmustelids. Named species are those with signif-icantly shorter or longer bacula or those thatcontribute to the outliers in Fig. 4.

FIG. 4.—Independent standardized contrastsfor ln bacular length regressed on standardizedcontrasts of body size (ln condylobasal length)a) across caniforms and b) within mustelids. Sig-nificant outliers are labeled.cant (P , 0.0001), the coefficient of deter-

mination is rather low (r2 5 0.366) due to4 contrasts, all of which are significant out-liers (i.e., possess residuals .1.96 SDs fromthe regression line): Helarctos malayanusversus Melursus ursinus (ursids), Neophocacinerea versus Otaria byronia (otariids),Lutrogale perspicillata versus Pteronurabrasiliensis (lutrines), and Ictonyx libycaversus I. striatus (mustelines). The removalof these 4 points improves the fit of the re-gression equation dramatically (r2 increasesto 0.582). Similarly, a highly significant lin-ear relationship is present within mustelids(Fig. 4b). Again, the coefficient of deter-mination is low (0.356) due to the Lutro-gale–Pteronura and Ictonyx contrasts beingsignificant outliers. Removal of these out-liers raises r2 to 0.662.

Each outlier reveals significant devia-tions away from the general tendency

among caniforms (and within mustelids) forthe baculum to increase in length as sizeincreases. For instance, although this posi-tive relationship holds when H. malayanusand M. ursinus are contrasted, connectingthe species points yields a line with a muchmore positive slope than that for all cani-forms (Fig. 3a), thereby emphasizing therelatively short os penis of H. malayanuscompared with its sister species M. ursinus.In contrast, the highly negative residual forthe node connecting Lutrogale perspicillataand P. brasiliensis indicates a trend in theopposite direction. Thus, although P. bra-siliensis is larger than its close relative Lu-trogale perspicillata, its baculum is rela-tively (and absolutely) shorter (Fig. 3b).


Character Analysis

The 10 bacular characters are highly in-congruent with the carnivore phylogeny ofBininda-Emonds et al. (1999), showing anensemble consistency index (CI) of 0.147and individual CIs for each character of0.500 or less (and often ,0.200). This isreflected by numerous parallel derivations ofthe different states and reversals within eachcharacter. Below, we discuss each of the 10characters described in the ‘‘Materials andMethods’’ and illustrated in Fig. 1.

Character 1.—A baculum without anabrupt distal bend is primitive for caniformsand mustelids. Any upward bend in the dis-tal tip is peculiar to mustelids and is derivedindependently on several occasions. Aweak bend is found independently in Melesanakuma, Spilogale putorius, Poecilogalealbinucha, and the clade of Gulo gulo, P.marshalli, and Martes. Within the latterclade, the bend is more prominent (state 2)in Martes gwatkinsii and Martes flavigula.Independent derivations of this morphologyalso occur in Lontra, Mustela, Taxidea tax-us, and Vormela peregusna.

Character 2.—A hook on the distal partof the os penis is a derived feature presentonly within mustelids. It unites Mustelawith independent reversals to the primitivecondition occurring in Mustela erminea andMustela frenata. However, note that the lat-ter would be a synapomorphy for the 2 spe-cies under the phylogeny of Abramov(2000b).

Character 3.—Shape of the stem in crosssection varies among out-group taxa. Mostspecies possess a purely triangular shape.Only the 2 procyonid species examined(state 2) and several pinniped species (state1) display the 2 derived conditions. Assuch, although the ancestral mustelid formdiffers from the plesiomorphic caniformcondition (state 0), it is equivocal betweenthe 2 apomorphic states (1 and 2). This un-certainty, combined with the polytomy atthe base of the mustelid radiation, makestracing the evolution of this character dif-

ficult. The ancestral condition for most lin-eages is equivocal. Only Mustela clearly re-verse primitively to rederive the triangularmorphology, which universally is present inall the species examined (also independent-ly in Lontra and V. peregusna). Both lutri-nes and mephitines display the apomorphicconditions, as do ‘‘badgers’’ (i.e., Arctonyx,Meles, Mellivora, Melogale, Mydaus, Tax-idea). However, the intermediate morphol-ogy (state 1) does diagnose the clade ofArctonyx collaris, Meles, and Mydaus ja-vanensis. Most of the remaining species,notably the clade of G. gulo, P. marshalli,and Martes, possess a baculum with arounded stem.

Character 4.—The urethral grooveshows a complicated evolutionary history,with a number of gains and losses in mus-telids to which it is largely confined; onlythe canids Canis lupus and Vulpes vulpes(state 0) and the procyonid Procyon lotor(state 1) possess a urethral groove amongthe out-group taxa. Lack of a groove alsois ancestral for mustelids. The polytomy atthe base of the mustelids prevents an exactreconstruction of these characters, but theabsence of a groove is found primitively inmost of major lineages: mephitines, all thebadgers with the exception of Mellivora ca-pensis (state 0), P. albinucha, and fromLyncodon patagonicus to Eira barbara. Ex-cept for S. ichnusae, all lutrines possess agroove of some form, although there is noapparent pattern. G. gulo, P. marshalli, andmost species of Martes possess a short, dis-tal groove, which becomes increasinglypronounced in the sister species Martesgwatkinsii and Martes flavigula (state 0).All species of Mustela also have a long,strongly formed groove.

Character 5.—The primitive conditionamong caniforms is the absence of a dis-tinct head on the baculum. Only otarioids,procyonids, and mustelids derive a head ofsome description. Unfortunately, the distri-bution of this character makes the ancestralcondition for all mustelids and the majorlineages within mustelids equivocal be-


tween states 0 and 2. Mephitines and lutri-nes lack a distinct head ancestrally, as doesMustela. This condition remains in mostspecies within the 2 clades. The lutrines En-hydra lutris and the sister species Lutra lu-tra and S. ichnusae together with Mustelafelipei, Mustela nudipes, and Mustela stri-gidorsa independently derived the inter-mediate state 1. The remaining mustelidspecies generally possess a bacular head.This includes all badgers (except Mydausjavanensis), which possess a well-markedhead. The clade spanning from Galictis vit-tata to Martes primitively has a pronouncedbacular head, before it weakens to becomeonly slightly pronounced in G. gulo, P.marshalli, and most species of Martes.

Character 6.—The absence of an open-ing is primitive in mustelids. Derived oc-currences of an opening are found indepen-dently in Meles and within Martes (all spe-cies except Martes gwatkinsii and Martesflavigula). The polytomy within Martesprevents a more exact reconstruction of thischaracter within the genus.

Character 7.—The primitive conditionamong caniforms is of an undivided tip,which is found in all species except Phocavitulina, in both procyonid species exam-ined (Potos flavus and P. lotor), and amongmustelids. The ancestral condition for mus-telids is equivocal; however, most lineagesare characterized primitively by an undivid-ed tip (e.g., lutrines plus mephitines, mostbadgers, and within mustelines). Baculawith subdivided tips occur independently inmost lutrine species, Melogale personata,Mellivora capensis, throughout the cladespanning from L. patagonicus to Martes,Mustela nudipes, Mustela strigidorsa, Mus-tela felipei, and Mustela africana.

Character 8.—An asymmetrical baculartip is found only within mustelids and in-dependently is derived a number of timeswithin the family. It occurs among the ma-jority of lutrines, including the clades ofLutra lutra plus S. ichnusae, and Lutrogaleperspicillata plus P. brasiliensis. Other der-ivations of an asymmetrical tip include

Melogale personata, T. taxus, V. peregus-na, all species of Mustela, and the clade ofG. gulo, P. marshalli, and Martes, withMartes pennanti regaining the ancestralcondition.

Character 9.—The derived conditionwhere the bacular head has additional pro-jections occurs infrequently among cani-forms: some otariids, P. flavus, and mustel-ids. Within mustelids, V. peregnusa, G.gulo, G. vittata, Mustela felipei, and Martesgwatkinsii plus Martes flavigula all possessthe derived condition independently.

Character 10.—The distribution of thischaracter is shown in Appendix II. Of thespecies listed as having relatively short orlong bacula, this difference was significant(.1.96 SD from the regression line) in onlya few species. Those with significantlyshorter bacula were Ailuropoda melanoleu-ca, Ailurus fulgens, Conepatus mesoleucus,and H. malayanus. Only Odobenus rosma-rus and S. ichnusae had significantly elon-gated bacula.

Bacula of average length are ancestral forcaniforms and for all the caniform familiesexamined herein. For nonmustelid species,relatively shorter bacula are derived inde-pendently in A. melanoleuca, A. fulgens,and H. malayanus. Relatively longer baculaare limited to independent appearances inP. lotor and several pinnipeds, N. cinerea,O. rosmarus, and Phoca groenlandica.

Most mustelids maintain the ancestralstate of a baculum of average length. Rel-atively shorter or longer bacula are derivedindependently a number of times. Shorterbacula occur in Mydaus javanensis, P. al-binucha, P. brasiliensis, and the mephitinesC. mesoleucus and Mephitis mephitis. Arelatively short baculum may represent asynapomorphy for all mephitines, followedby a reversal back to the ancestral condi-tions in S. putorius (accelerated transfor-mation optimization). However, parallelderivations in the 2 mephitine species areequally parsimonious (delayed transforma-tion optimization—Swofford and Maddison1987). Relatively long bacula are found in-


dependently in I. striatus, P. marshalli, andthe lutrines E. lutris, Lontra canadensis,and S. ichnusae. Mustela felipei, for whichwe lacked data on bacular length, was in-ferred by parsimony to possess bacula ofaverage length.

Overall, we infer the ancestral mustelidbaculum to be a remarkably simple struc-ture. It is of average length (relative to thatfound in extant mustelids), with a straight,not triangular, stem that likely lacks a ure-thral groove and has a simple distal end thatis for the most part not distinct from therest of the bone (it may possess a distincthead, however).

DISCUSSION

Functional correlates of baculum length(size).—The general uniformity in length ofthe os penis observed in this study, whetheracross caniforms or across mustelids only,belies an important dichotomy present with-in Carnivora. To some extent, the unifor-mity is an artifact of examining only cani-forms, which all possess the putative an-cestral condition among carnivores of amuch-elongated baculum (Dixson 1995). Incontrast, some families within the other ma-jor carnivore clade, the Feliformia (‘‘cat-like carnivores’’: felids, herpestids, hyaen-ids, and viverrids), are characterized by anos penis that is sharply reduced or incom-pletely ossified (felids) or is lost completely(hyaenids and some viverrids). This lengthdichotomy is thought to be related to theduration of copulation, with a correlationbetween extended copulation times and anelongated os penis being observed, sug-gesting a structural or functional role forthis bone such as protecting the urethra(Dixson 1995; also Ewer 1973; Long andFrank 1968). Fractures of the baculum havebeen reported in mustelids (Laidler 1982;Ruprecht 1994), which points to a largeload being placed on the penis during theprolonged intromission pattern of copula-tion.

Caniforms are remarkably similar in rel-ative bacular length. Most species have a

baculum that is not significantly differentfrom the average length found in mustelids.Even measuring bacular lengths relative tothe caniform average did not alter this pat-tern much. When we performed this cal-culation, only V. zerda shifted categories(from having an average-sized to relativelyshort baculum). Therefore, it is probablynot worthwhile seeking a functional expla-nation for what we believe is largely ran-dom variation in baculum length in mustel-ids or across caniforms as a whole. Ourview is substantiated by the analyses ofLariviere and Ferguson (2002), who founda lack of support in carnivores for variousfunctional explanations of bacular lengthvariation. Willemsen (1992) associated therelatively elongate baculum in otters withtheir aquatic lifestyle, but this observationis at best only weakly substantiated by ourdata. Only E. lutris, Lontra canadensis, andS. ichnusae have elongate bacula, whereasP. brasiliensis possesses a relatively shortos penis; all remaining otters have ones ofaverage length. A similar relation betweenelongate bacula and aquatic copulation inpinnipeds (Scheffer and Kenyon 1963) alsois not supported by our data (also Dixson1995). Finally, although some significantshortenings or lengthenings of the baculado occur, they arise independently in di-verse lineages with no obvious functionalconnection between the species. Otherwise,there is no consistent phylogenetic patternfor changes in baculum length among can-iforms except possibly for mephitineswhere it is shorter in 2 of the 3 species ex-amined.

Instead, functional explanations are per-haps more profitably directed at variation inqualitative bacular morphology. However,even here, different explanations may applyto different taxa (Dixson 1995). Therefore,we concentrate on possible evolutionarytransformations of bacular morphology inmustelids.

Evolution of the mustelid baculum.—It isclear that a large os penis (compared withthat found in feliforms) is primitive for can-


FIG. 5.—Putative evolution of the head of thebaculum. Top—in Martes and Taxidea taxus: a)Martes zibellina, b) Martes martes, c) Martesamericana, d) Martes foina, e) Martes pennanti,f) Taxidea taxus. All views are of the distal tipin dorsal view. Bottom—in Mustela: a) Mustelaerminea, b) Mustela africana (from Izor and Pe-terson 1985), c) Mustela nudipes, d) Mustelastrigidorsa, e) Mustela felipei (from Izor and dela Torre 1978), f) Mustela lutreola, g) Mustelasibirica, h) Mustela putorius, i) Mustela vison,j) Mustela nivalis (abnormal specimen from theNetherlands). All views are of the distal tip fromthe left side.

iforms and mustelids. Furthermore, we inferthat the baculum of the mustelid commonancestor was a remarkably simple rod-likestructure resembling the morphology foundin extant species of mephitine. Unfortunate-ly, paleontological evidence to support thislatter assessment is lacking. Bacula for Ol-igocene and early Miocene ‘‘palaeomustel-ids,’’ which are diagnosed as such by theretention of a carnassial notch on the uppercarnassial tooth (Wolsan 1993), are virtu-ally unknown. A fragment of the baculumfrom Megalictis (5Aelurocyon) from theearly Miocene of North America (Matthew1907) does reveal that this archaic grouphad a relatively large os penis. The suc-ceeding group of ‘‘neomustelids,’’ whichare characterized by having lost the carnas-sial notch on the upper fourth premolar, al-ready possessed a baculum with a complexhead (e.g., P. marshalli from Pliocene ofNorth America; Harrison 1982).

From the ancestral mustelid condition, itis possible to derive what we consider to bethe major bacular forms in mustelids. Oneform is characterized by the presence of adistinct head with projections. These pro-jections are usually paired, although the dis-tal tip often is also divided into 3 (e.g., G.gulo, Melogale moschata), which also isobserved in juvenile E. lutris. To some ex-tent, we can place G. vittata and G. cuja inthe same group because the small-pairedknobs (‘‘horns’’) and a large central lobe oftheir bacula resemble the trifid tip in theabove-mentioned forms. E. barbara alsocan be considered to belong to this group.

The paired projections can fuse to forman opening (e.g., most badgers and Martes)or not (e.g., Melogale personata, P. mar-shalli, and S. ichnusae). In the former case,a putative transformation sequence for theos penis is observable (Fig. 5). The fork-shaped distal tip in Martes zibellina (alsopresent in juvenile specimens of badgersand lutrines) is closed to form a ring inMartes martes, Martes americana, andMartes foina (the projections are alwaysfused in the latter species, but not neces-

sarily in some specimens of the former 2species), with the foramen thus formed be-coming smaller in Martes pennanti. Contin-uation of this sequence would result in thecomplete closure of the foramen and pos-sibly a flat, nearly triangular shaped head asin T. taxus. We base the polarity for thissequence on ontogenetic data where it isnoteworthy that young individuals of bad-gers and lutrines also possess a bifurcatedhead. However, the reverse polarity makesmore sense according to the tree of Binin-da-Emonds et al. (1999) and the times ofdivergence that it presents.

A different evolutionary pathway is dem-onstrated by Mellivora capensis, Ictonyx, V.peregusna, and Mustela in which the headis either absent or flattened, spoon-shapedor cup-shaped. In those cases where thehead is not pronounced, the distal tip of theos penis is hook-shaped. V. peregusna andMustela are further characterized by astrongly developed urethral groove. Finally,a tendency toward complication of the bac-ulum by the formation of a distal hook is


observed within Mustela (Fig. 5). From asimple, weakly curved baculum (Mustelaerminea and Mustela frenata), the distal tipis flattened (Mustela africana) with smallprojections developing on it (Mustela feli-pei, Mustela nudipes, and Mustela strigi-dorsa), which are subsequently lost whilethe tip is bent upward (Mustela itatsi, Mus-tela lutreola, and Mustela sibirica) to final-ly form an actual hook (Mustela altaica,Mustela eversmanii, Mustela kathiah, Mus-tela nigripes, Mustela nivalis, Mustela pu-torius, Mustela subpalmata, Mustela vison).Evidence for this series, with a late appear-ance of hook in the evolution of the bacu-lum in Mustela, again derives from onto-genetic information; it is not easily recon-cilable with the phylogeny of Bininda-Emonds et al. (1999). In immature males ofthe last group, the hook is not yet developedand baculum is a rod-like, slightly curvedstructure resembling that of Mustela ermi-nea and Mustela frenata (e.g., Mustela vi-son; Tarasov 1984).

Therefore, 3 major apomorphies are ap-parent in the development of the mustelidos penis: a slight reduction in length, in-creased complication of the head, and de-velopment of the urethral groove. Althoughthe baculum as a whole is highly diagnosticof most genera, these pathways have prob-ably occurred independently in all adaptivelineages to varying extents. Some parallelsare particularly noteworthy. The division ofthe distal tip in Mustela felipei is reminis-cent of that in Melogale moschata or G.gulo. The characteristic wide, leaf-shapedlobe and opposing finger-shaped outgrowthof the bacular head of P. marshalli is alsofound in Mustela strigidorsa and Mustelanudipes. Special attention also should begiven to the similarity in structure betweenthe distantly related genera G. gulo andMelogale moschata, or P. marshalli and S.ichnusae, in which the cranial and dentalcharacters otherwise differ considerably.

Analogous examples of morphologicalparallelism also are observed with speciesfrom other caniform families, notably pro-

cyonids. For example, the bifurcated headin P. lotor resembles that of Lutra lutra,and the distal trifid tip of P. flavus is quitesimilar to that in G. gulo. Nasua nasua alsohas a trifid tip on its baculum. In contrast,Bassariscus astutus has a simple rod-shaped baculum with a slightly extendedand flattened distal tip (Burt 1960), whichpartly resembles the bacula of L. patagon-icus and some mephitines.

Systematic value of the baculum.—Thetaxonomic importance of bacular structureamong closely related forms has long beenappreciated. The species-specific nature ofbacular morphology causes it to be an ex-cellent character for species diagnosis andtaxonomy, assessing the affinity of fossilspecies, and elucidating relationshipsamong closely related genera (Baryshnikovand Abramov 1997, 1998; Lavrenchenkoand Baskevich 1996; Packard 1960; Patter-son and Thaeler 1982 and references there-in; Pocock 1918; Rabeder 1976; White1953). This especially has been the case inCarnivora (e.g., Abramov 2000b; Barysh-nikov 2001; Burt 1960; Chaine 1926; Di-dier 1947a, 1947b, 1948; Morejohn 1975;Pocock 1941; Tumlinson and McDaniel1984; van Zyll de Jong 1972). This utilitymay derive from the fact that many pro-posed functional hypotheses to explain var-iation in bacular morphology (e.g., a pleio-tropic side effect or a mechanism for pre-reproductive isolation) relate to species-lev-el phenomena.

Despite this, 2 factors limit the utility ofusing bacular characters to elucidate phy-logeny. First, there is the high amount ofconvergence that we noted, both withinmustelids and across caniforms as a whole.Second, even on a more restricted scale, no-table differences in the shape of os peniscan occur between closely related species.This potentially hinders the discovery ofshared derived features. For example,strong differences occur between S. ichnu-sae and other lutrines, between Lutra lutraand Lontra, or between the sister speciesMartes flavigula and Martes gwatkinsii and


remaining species in Martes. These differ-ences could have arisen either due to anearly phylogenetic separation or rapidchanges in the structure of the os penis dur-ing a short period of time. The latter sup-position appears to be more realistic in caseof the insular isolation of S. ichnusae,where its separation from the Eurasian ge-nus Lutra occurred relatively recently (Mid-dle Pleistocene—Willemsen 1992).

Still, we feel that the os penis can pro-vide useful phylogenetic information. Muchof the convergence we observed resultsfrom defining our characters too generallyso that they were applicable across all mus-telids. However, gross bacular structure isinformative at very inclusive levels withinCarnivora, where it distinguishes the cladesCanidae (large bacula consisting of 2 bonyplates grown together), Arctoidea (large,rod-shaped baculum), and Feliformia (re-duced bacula that are lost entirely in somegroups). Bacular morphology also shouldbe informative at less inclusive levels, par-ticularly around the generic or subgenericlevels, with more specific coding schemesindividually tailored for each genus andwhen combined with other data sourcessuch as morphology of the basicranial re-gion.

ACKNOWLEDGMENTS

We thank the following for access to the mu-seum collections under their care: V. Eisenmann,H. Endo, J. Englund, A. Forsten, P. Jenkins, A.Lister, R. Masuda, I. Pavlinov, J. Purdue, B.Stanley, R. S. Voss, J. Werner, V. Yudin, and E.Zholnerovskaya. We also thank A. Averianov,H. N. Bryant, E. H. Miller, M. Zaitsev, and 2anonymous reviewers for critically reading thismanuscript. G. F. Baryshnikov and A. V. Abra-mov were supported by a grant from the RussianState Scientific Program ‘‘Biodiversity.’’ O. R.P. Bininda-Emonds was supported by a NaturalSciences and Engineering Research Councilpostdoctoral fellowship (Canada) and the vander Leeuw Fonds (the Netherlands). G. Barysh-nikov obtained financial support for his visit tothe Natural History Museum in London from theRoyal Society (United Kingdom).

LITERATURE CITED

ABRAMOV, A. V. 2000a. The taxonomic status of theJapanese weasel, Mustela itatsi (Carnivora, Mustel-idae). Zoologicheskii Zhurnal 79:80–88 (in Russian,English summary).

ABRAMOV, A. V. 2000b. A taxonomic review of thegenus Mustela (Mammalia, Carnivora). Zoosyste-matica Rossica 8:357–364.

ABRAMOV, A. V. 2001. Notes on the taxonomy of theSiberian badgers (Mustelidae: Meles). Proceedingsof the Zoological Institute Russian Academy of Sci-ences 288:221–233 (in Russian, English summary).

ABRAMOV, A. V., AND G. F. BARYSHNIKOV. 2000. Geo-graphic variation and intraspecific taxonomy of wea-sel Mustela nivalis (Carnivora, Mustelidae). Zoo-systematica Rossica 8:365–402.

BARYSHNIKOV, G. F. 2001. Order Carnivora. Pp. 11–425 in Mammals of Russia and adjacent territories.Carnivores and pinnipeds (A. A. Aristov and G. I.Baranova, eds.). Russian Academy of Sciences, St.Petersburg, Russia (in Russian).

BARYSHNIKOV, G. F., AND A. V. ABRAMOV. 1997. Struc-ture of baculum (os penis) in Mustelidae (Mamma-lia, Carnivora), communication 1. ZoologicheskiiZhurnal 76:1399–1410 (in Russian, English sum-mary).

BARYSHNIKOV, G. F., AND A. V. ABRAMOV. 1998. Struc-ture of baculum (os penis) in Mustelidae (Mamma-lia, Carnivora), communication 2. ZoologicheskiiZhurnal 77:231–236 (in Russian, English summary).

BARYSHNIKOV, G. F., AND O. R. POTAPOVA. 1990. Var-iability of the dental system in badgers (Meles, Car-nivora) of the USSR fauna. Zoologicheskii Zhurnal69:84–97 (in Russian, English summary).

BININDA-EMONDS, O. R. P., J. L. GITTLEMAN, AND A.PURVIS. 1999. Building large trees by combiningphylogenetic information: a complete phylogeny ofthe extant Carnivora (Mammalia). Biological Re-views 74:143–175.

BURT, W. H. 1936. A study of the baculum in the gen-era Perognathus and Dipodomys. Journal of Mam-malogy 17:145–156.

BURT, W. H. 1960. Bacula of North American mam-mals. Miscellaneous Publications of the Museum ofZoology, University of Michigan 113:1–76.

CHAINE, J. 1926. L’os penien, etude descriptive et com-parative. Actes de la Societe Linneene de Bordeaux78:5–195.

CONTRERAS, L. C., J. C. TORRES-MURA, A. E. SPOTOR-NO, AND F. M. CATZEFLIS. 1993. Morphological var-iation of the glans penis of South American octo-dontid and abrocomid rodents. Journal of Mammal-ogy 74:926–935.

DIDIER, R. 1946. Etude systematique de l’os penien desmammiferes. Mammalia 10:78–91.

DIDIER, R. 1947a. Etude systematique de l’os peniendes mammiferes. Carnivores 2. Famille des Muste-lides. Mammalia 11:30–43.

DIDIER, R. 1947b. Etude systematique de l’os peniendes mammiferes. Carnivores 2. Famille des Muste-lides. Mammalia 11:139–152.

DIDIER, R. 1948. Etude systematique de l’os penien desmammiferes. Mammalia 12:67–93.

DIDIER, R. 1950. Etude systematique de l’os penien des


mammiferes. Famille des Procyonides. Mammalia14:78–94.

DIXSON, A. F. 1995. Baculum length and copulatorybehaviour in carnivores and pinnipeds (Grand OrderFerae). Journal of Zoology (London) 235:67–76.

EWER, R. F. 1973. The carnivores. Cornell UniversityPress, London, United Kingdom.

FELSENSTEIN, J. 1985. Phylogenies and the comparativemethod. American Naturalist 125:1–15.

GITTLEMAN, J. L., AND B. VAN VALKENBURGH. 1997.Sexual dimorphism in the canines and skulls of car-nivores: effects of size, phylogeny, and behaviouralecology. Journal of Zoology (London) 242:97–117.

HAFNER, J. C. 1978. Evolutionary relationships of kan-garoo mice, genus Microdipodops. Journal of Mam-malogy 59:354–365.

HALL, E. R., AND K. R. KELSON. 1959. The mammalsof North America. Ronald Press, Co., New York 2:547–1083 1 79.

HAMILTON, W. J., JR. 1949. The bacula of some NorthAmerican vespertilionid bats. Journal of Mammal-ogy 30:97–102.

HARRISON, J. A. 1982. The baculum of the Plesiogulomarshalli (Carnivora: Mustelidae). Journal of Pa-laeontology 56:1266–1273.

HARVEY, P. H. 1982. On rethinking allometry. Journalof Theoretical Biology 95:37–41.

HARVEY, P. H., AND M. D. PAGEL. 1991. The compar-ative method in evolutionary biology. Oxford Uni-versity Press, Oxford, United Kingdom.

HEPTNER, V. G., K. K. CHAPSKY, V. A. ARSENJEV, AND

V. E. SOKOLOV. 1976. Mammals of the Soviet Union.Pinnipeds and whales. Vysshaya Shkola, Moscow,Russia 2(3):1–720 (in Russian).

IZOR, R. J., AND L. DE LA TORRE. 1978. A new speciesof weasel (Mustela) from the highlands of Colom-bia, with comments on the evolution and distributionof South American weasels. Journal of Mammalogy59:92–102.

IZOR, R. J., AND N. E. PETERSON. 1985. Notes on SouthAmerican weasels. Journal of Mammalogy 66:788–790.

LAIDLER, L. 1982. Otters in Britain. David & Charles,London, United Kingdom.

LARIVIERE, S., AND S. H. FERGUSON. 2002. On the evo-lution of the mammalian baculum: vaginal friction,prolonged intromission or induced ovulation? Mam-mal Review 32:283–294.

LAVRENCHENKO, L. A., AND M. I. BASKEVICH. 1996.Variation of generative system structure in somespecies of the genus Mastomys Thomas, 1915 (Ro-dentia, Muridae). Mammalia 60:277–288.

LONG, C. A., AND T. FRANK. 1968. Morphometric var-iation and function in the baculum, with commentson correlation of parts. Journal of Mammalogy 49:32–43.

MALATESTA, A. 1977. The skeleton of Nesolutra ich-nusae sp. n., a quaternary otter discovered in Sar-dinia. Geologica Romana 16:173–209.

MASUDA, R., AND M. C. YOSHIDA. 1994. A molecularphylogeny of the family Mustelidae (Mammalia,Carnivora), based on comparison of mitochondrialcytochrome b nucleotide sequences. Zoological Sci-ence 11:605–612.

MATTHEW, W. D. 1907. A lower Miocene fauna from

South Dakota. Bulletin of the American Museum ofNatural History 23:169–219.

MEAD, R. A. 1967. Age determination in the spottedskunk. Journal of Mammalogy 48:606–616.

MECZYNSKI, S. 1974. Morphohistological structure ofthe female genital organs in sousliks. Acta Therio-logica 19:91–106.

MONDOLFI, E. 1987. Baculum of the lesser Andean co-ati, Nasuella olivacea (Gray), and of the larger gri-son, Galictis vittata (Schreber). Fieldiana: Zoology39:447–454.

MOREJOHN, G. V. 1975. A phylogeny of otariid sealsbased on morphology of the baculum. Rapports etProces-verbaux des Reunions Conseil Internationalpour l’Exploration de la Mer 169:49–56.

OGNEV, S. I. 1931. The animals of the Eastern Europeand Northern Asia. Gosizdat, Moscow–Leningrad,Russia 2:1–776 (in Russian).

PACKARD, R. L. 1960. Speciation and evolution of pyg-my mice, genus Baiomys. University of Kansas Pub-lications, Museum of Natural History 9:579–670.

PAGEL, M. 1993. Seeking the evolutionary regressioncoefficient: an analysis of what comparative meth-ods measure. Journal of Theoretical Biology 164:191–205.

PATTERSON, B. D., AND C. S. THAELER, JR. 1982. Themammalian baculum: hypotheses on the nature ofbacular variability. Journal of Mammalogy 63:1–15.

POCOCK, R. I. 1918. The baculum or os penis of somegenera of Mustelidae. Annals and Magazine of Nat-ural History 1:307–312.

POCOCK, R. I. 1941. Mammalia. The fauna of BritishIndia, including Ceylon and Burma. Taylor andFrancis, London, United Kingdom 2:1–503.

PURVIS, A., J. L. GITTLEMAN, AND H.-K. LUH. 1994.Truth or consequences: effects of phylogenetic ac-curacy on two comparative methods. Journal of The-oretical Biology 167:293–300.

PURVIS, A., AND A. RAMBAUT. 1995. Comparative anal-ysis by independent contrasts (CAIC): an AppleMacintosh application for analysing comparativedata. Computer Applications in the Biosciences 11:247–251.

RABEDER, G. 1976. Die Carnivoren (Mammalia) ausdem Altpleistozan von Deutsch-Altenburg 2. MitBeitragen zur Systematik einiger Musteliden undCaniden. Beitrage zur Palaontologie von Osterreich1:5–119.

ROMER, A. S., AND T. S. PARSONS. 1986. The vertebratebody. 6th ed. Saunders College Publishing, Phila-delphia, Pennsylvania.

RUPRECHT, A. L. 1994. Ausgeheilter Bruch des Pen-isknochens bei einem Mink, Mustela vison Schreber,1777. Saugetierkundliche Informationen 3:684–685.

SANDERSON, M. J., A. PURVIS, AND C. HENZE. 1998.Phylogenetic supertrees: assembling the trees of life.Trends in Ecology and Evolution 13:105–109.

SCHEFFER, V. B., AND K. W. KENYON. 1963. Baculumsize in pinnipeds. Zeitschrift fur Saugetierkunde 28:38–41.

SOKOLOV, I. I. 1968. Origin and systematic position ofthe Mustelidae family and the major trends of itsevolution. Bulleten Moskovskogo Obstcestva ispy-tatelei prirody, otdel biologii 73(6):5–16 (in Russian,English summary).


SWOFFORD, D. L., AND W. P. MADDISON. 1987. Recon-structing ancestral character states under Wagnerparsimony. Mathematical Biosciences 87:199–229.

TARASOV, S. A. 1984. Age peculiarities of baculum insome mammals. Zoologicheskii Zhurnal 58:120–125 (in Russian, English summary).

TUMLINSON, R., AND V. R. MCDANIEL. 1984. A descrip-tion of the baculum of the bobcat (Felis rufus), withcomments on its development and taxonomic impli-cations. Canadian Journal of Zoology 62:1172–1176.

VAN ZYLL DE JONG, C. G. 1972. A systematic reviewof the Nearctic and neotropical river otters (genusLutra, Mustelidae, Carnivora). Life Sciences Con-tributions, Royal Ontario Museum 80:1–104.

VAN ZYLL DE JONG, C. G. 1992. A morphometric anal-ysis of cranial variation in Holarctic weasels (Mus-tela nivalis). Zeitschrift fur Saugetierkunde 57:77–93.

VAZ-FERREIRA, R. 1968. South American sea lionOtaria flavescens (Shaw 1800). Pp. 39–65 in Hand-book of marine mammals (S. H. Ridgeway and R.J. Harrison, eds.). Academic Press, New York 1:1–235.

WEBER, M. 1928. Die Saugetiere. Einfuhrung in dieAnatomie und Systematik der recenten und fossilenMammalia. 2nd ed. G. Fischer, Jena, Germany 1:1–430, 2:431–879.

WHITE, J. A. 1953. The baculum in the chipmunks ofwestern North America. University of Kansas Pub-lications, Museum of Natural History 5:611–631.

WILLEMSEN, G. F. 1992. A revision of the Pliocene andQuaternary Lutrinae from Europe. Scripta Geologica101:1–115.

WOLSAN, M. 1993. Phylogeny and classification of ear-ly European Mustelida (Mammalia: Carnivora).Acta Theriologica 38:345–384.

WOZENCRAFT, W. C. 1993. Order Carnivora. Pp. 279–348 in Mammal species of the world: a taxonomicand geographic reference (D. E. Wilson and D. A.Reeder, eds.). Smithsonian Institution Press, Wash-ington, D.C.

Submitted 18 March 2002. Accepted 5 September2002.

Associate Editor was Robert D. Bradley.

APPENDIX I

Specimen list.—The following species (withsample sizes in parentheses; extinct species arepreceded by †) were examined from the follow-ing museum collections: American Museum ofNatural History in New York; Biological andSoil Research Institute of the Far East Division,Russian Academy of Sciences, Vladivostok,Russia; Field Museum of Natural History inChicago, Illinois; Illinois State Museum inSpringfield, Illinois; Illinois University Cham-paigne-Urbana in Urbana, Illinois; Museum Na-tional d’Histoire Naturelle, Paris, France; Natu-ral History Museum, London, United Kingdom;

Naturhistoriska Riksmuseet, Stockholm, Swe-den; National Science Museum, Tokyo, Japan;Zoological Institute, Russian Academy of Sci-ences, St. Petersburg, Russia; Zoological Mu-seum, Helsinki, Finland; Zoological Museum ofMoscow University, Moscow, Russia; and Zoo-logical Museum, Siberian Branch of RussianAcademy of Sciences, Novosibirsk, Russia.Eight additional species where information wasobtained from the literature are listed with thesources. Subfamily taxonomy again followsWozencraft (1993) for convenience.

Lutrinae.—Amblonyx cinereus (4), Enhydralutris (8), Lontra canadensis (2), Lontra longi-caudis (1), Lutra lutra (2), Lutrogale perspicil-lata (1), Pteronura brasiliensis (1), †Sardolutraichnusae (from Malatesta 1977).

Melinae.—Arctonyx collaris (2), Meles anak-uma (10), Meles meles (8), Melogale moschata(from Pocock 1941), Melogale personata (2),Mydaus javanensis (1).

Mellivorinae.—Mellivora capensis (from Po-cock 1918).

Mephitinae.—Conepatus mesoleucus (1), Me-phitis mephitis (2), Spilogale putorius (fromMead 1967).

Mustelinae.—Eira barbara (3), Galictis cuja(5), Galictis vittata (from Mondolfi 1987), Gulogulo (9), Ictonyx striatus (5), Ictonyx libyca (3),Lyncodon patagonicus (2), Martes americana(4), Martes flavigula (4), Martes foina (6), Mar-tes gwatkinsii (1), Martes martes (17), Martespennanti (2), Martes zibellina (45), Mustela af-ricana (from Izor and Petersen 1985), Mustelaaltaica (5), Mustela erminea (29), Mustela ev-ersmanii (8), Mustela felipei (from Izor and dela Torre 1978), Mustela frenata (3), Mustela itat-si (10), Mustela kathiah (2), Mustela lutreola(5), Mustela lutreolina (1), Mustela nigripes (2),Mustela nivalis (31), Mustela nudipes (1), Mus-tela putorius (11), Mustela sibirica (25), Mus-tela strigidorsa (2), Mustela subpalmata (1),Mustela vison (5), †Plesiogulo marshalli (fromHarrison 1982), Poecilogale albinucha (1), Vor-mela peregusna (5).

Taxidiinae.—Taxidea taxus (3).Caniformia excluding Mustelidae (out-

groups).—Ailuropoda melanoleuca (1), Ailurusfulgens (1), Callorhinus ursinus (3), Canis lupus(10), Eumetopias jubatus (4), Leptonychotesweddellii (1), Odobenus rosmarus (6), Phocagroenlandica (3), Phoca vitulina (4), Potos fla-vus (1), Procyon lotor (2), Ursus arctos (3), Vul-pes vulpes (8), Zalophus californianus (2).


APPENDIX IICondylobasal length of skull and baculum lengths for adult male caniforms. Relative length refers

to length of baculum relative to the mustelid average (see text): 0 5 shorter by 1 SD, 1 5 average,2 5 longer by 1 SD. Those species with an asterisk were significantly shorter or longer than themustelid average (i.e., residuals greater than 1.96 SD from the regression line). Data for condylobasaland bacular lengths not collected by the authors were obtained from the following sources, as indi-cated: 1, Burt (1960); 2, Didier (1946); 3, Didier (1950); 4, Gittleman and Van Valkenburg (1997);5, Hall and Kelson (1959); 6, Harrison (1982); 7, Heptner et al. (1976); 8, Izor and Petersen (1985);9, Malatesta (1977); 10, Mead (1967); 11, Mondolfi (1987); 12, Ognev (1931); 13, Pocock (1918);14, Scheffer and Kenyon (1963); 15, Vaz-Ferreira (1968). Nonmustelid taxa in boldface were usedas outgroups for the parsimony analysis and for the mapping of characters 1–9; mapping of character10 used all taxa listed.

Family,species

Condylobasallength (mm) Source

Baculumlength(mm)

Relativebacularlength Source

Canidae

Alopex lagopusCanis aureusCanis latransCanis lupusNyctereutes procyonoidesUrocyon cinereoargenteus

127.90156.20193.60237.10119.20125.30

1212

3

4

59.3062.6776.80

100.8372.5053.10

111111

1211, 221

Vulpes vulpesVulpes zerda

145.5084.70

57.6031.00

11

1, 22

Otariidae

Arctocephalus tropicalisCallorhinus ursinusEumetopias jubatusNeophoca cinereaOtaria byroniaZalophus californianus

240.60232.30389.70295.38330.41279.20

77

128.00142.00206.00260.00146.00148.50

111211

14141414151, 14

Odobenidae

Odobenus rosmarus 369.00 7 528.68 2* 1, 14

Phocidae

Cystophora cristataErignathus barbatusHalichoerus grypusHydrurga leptonyxLeptonychotes weddelliiLobodon carcinophagusMirounga angustirostrisMirounga leoninaMonachus schauinslandiPhoca fasciata

271.50220.50261.90374.04284.81281.37503.98479.98238.55191.70

77

7

210.00140.00163.00233.00218.00220.00274.00331.00183.00142.00

1111111111

14141414141414141414

Phoca groenlandicaPhoca hispidaPhoca vitulina

210.60185.70205.00

777

185.00118.00137.00

211

141414

Procyonidae

Potos flavusBassariscus astutusBassariscus sumichrastiNasua naricaNasua nasuaProcyon cancrivorusProcyon lotor

87.8082.00

115.00109.50125.80114.60

45

5454

71.3046.8047.0082.5385.0094.6598.68

1111112

1, 3111, 31, 31, 31, 3


APPENDIX II.—Continued.

Family,species


Baculumlength(mm)


Ursidae

Ailurus fulgensAiluropoda melanoleucaHelarctos malayanusMelursus ursinusUrsus americanusUrsus arctosUrsus maritimusUrsus thibetanus

105.40251.00238.40307.30272.20340.00375.00291.60

23.0024.0052.00

151.00145.23134.00168.00118.25

0*0*0*11111

331, 3333

Mustelidae

Amblonyx cinereus 86.00 39.00 1Enhydra lutrisLontra canadensisLontra longicaudisLutra lutraLutrogale perspicillataPteronura brasiliensis

137.50110.00114.80119.50122.10154.80

155.0095.0069.6065.0050.0039.20

221110*

Sardolutra ichnusae 102.00 9 152.00 2* 9Arctonyx collarisMeles melesMeles anakumaMelogale personataMydaus javanensisTaxidea taxus

134.00130.80125.50

84.0090.10

125.00

82.0070.0068.3052.0030.3092.00

111101

Mellivora capensis 138.00 65.50 1 13Conepatus mesoleucusMephitis mephitis

73.0075.70

13.0020.60

0*0

Spilogale putorius 50.70 4 20.50 1 10Eira barbaraGalictis vittataGulo gulo

110.0072.00

144.70

78.0055.8080.10

111

11

Plesiogulo marshalli 169.50 6 182.00 2 6Ictonyx libycaIctonyx striatusLyncodon patagonicusMartes americana

54.0065.0055.6085.20

33.9057.0027.5036.00

1211

Martes flavigulaMartes foinaMartes gwatkinsiiMartes martesMartes pennantiMartes zibellina

108.0083.0090.0081.00

117.2082.70

82.0055.5057.3037.0096.7037.40

111111

Mustela africana 51.20 8 31.00 1 8Mustela altaicaMustela ermineaMustela eversmaniiMustela frenataMustela itatsiMustela kathiah

43.0042.0071.0046.0053.0050.10

24.0025.0041.0028.0029.0030.50

111111


APPENDIX II.—Continued.

Family,species


Baculumlength(mm)


Mustela lutreolaMustela lutreolinaMustela nigripesMustela nivalis

63.0058.5068.0033.00

36.0032.7037.0014.00

1111

Mustela nudipesMustela putoriusMustela sibiricaMustela strigidorsaMustela subpalmataMustela visonPoecilogale albinuchaVormela peregusna

60.7067.0060.0057.4050.2067.0054.2056.40

46.0037.0032.0040.7025.9048.0018.8042.05

11111101

APPENDIX IIICharacter states for the 10 morphological characters of the os penis used for cladistic analysis.

(See ‘‘Materials and Methods’’ for descriptions of characters and states.) Missing value (character10 for Mustela felipei) is indicated by a question mark.

Character

1 2 3 4 5 6 7 8 9 10

Canis lupusVulpes vulpesCallorhinus ursinusEumetopias jubatusZalophus californianusOdobenus rosmarusLeptonychotes weddelliiPhoca groenlandicaPhoca vitulinaPotos flavusProcyon lotorAilurus fulgensAiluropoda melanoleucaUrsus arctosAmblonyx cinereusLontra canadensis

0000000000200002

0000000000000000

0001110012200020

0022222222122210

0012120002100000

0000000000000000

0000000011100010

0000000000000010

0010100001000000

1111121211200112

Lontra longicaudisLutra lutraLutrogale perspicillataEnhydra lutrisPteronura brasiliensisSardolutra ichnusaeConepatus mesoleucusMephitis mephitisSpilogale putoriusArctonyx collarisMeles melesMeles anakuma

200000001001

000000000000

012211122111

010012222222

010101000222

000000000011

011011000000

011011000000

000000000000

111202001111


APPENDIX III.—Continued.

Character

1 2 3 4 5 6 7 8 9 10

Mellivora capensisMelogale personataMydaus javanensisTaxidea taxusEira barbaraGalictis vittataGulo guloIctonyx libycaIctonyx striatusLyncodon patagonicusMartes americanaMartes flavigulaMartes foina

0002001000121

0000000000000

1212112222222

0222221202101

2202221220111

0000000000101

1100011000010

0101001000111

0000011000010

1101111121111

Martes gwatkinsiiMartes martesMartes pennantiMartes zibellinaMustela africanaMustela altaicaMustela ermineaMustela eversmanii

21112222

00001101

22220000

01110000

11210000

01110000

10011000

11011111

10000000

11111111

Mustela felipeiMustela frenata

22

10

00

00

10

00

10

11

10

?1

Mustela itatsiMustela kathiahMustela lutreolaMustela lutreolinaMustela nigripesMustela nivalisMustela nudipesMustela putoriusMustela sibiricaMustela strigidorsaMustela subpalmataMustela visonPlesiogulo marshalliPoecilogale albinuchaVormela peregusna

222222222222112

111111111111000

000000000000220

000000000000120

000000100100100

000000000000000

000000100100100

111111111111101

000000000000001

111111111111201

MORPHOLOGICAL VARIABILITY AND EVOLUTION OF THE BACULUM (OS PENIS) IN MUSTELIDAE (CARNIVORA)

Documents

Transcript of MORPHOLOGICAL VARIABILITY AND EVOLUTION OF THE BACULUM (OS PENIS) IN MUSTELIDAE (CARNIVORA)