WIXLIB

Shell variations in European Ocenebra (Mollusca, Muricidae)discontinuity questions the biological reality ofboth species. In addition, the legendary intraspecific variation of O. erinaceus (no less than 49superfluous names! [Houart 2001a]) and the taxonomic individualisation of O. brevirobusta fromO. erinaceus, which is not based on a modernanalysis of the shell variation, also justify thisinterrogation. These observations constitute themain source of our thought. Combining severalmethods of comparative morphology, the goal ofthis work will be to analyse the shell variation ofthese Ocenebra in order to know if the populations reflect a process of morphological differentiation. O. erinaceus and O. brevirobusta will beclosely analyzed, but Ocinebrellus inornatus willbe also included in order to compare the specieswith European Ocenebra. We used the followingmethods:– cladistic analysis to search for the relationshipsbetween the populations;– traditional biometry to describe transformations of the shell shape and the sculptural characters during the ontogeny;– Procrustes analysis to delineate variations ofsculptural characters;– Fourier analysis to study variations of outlines.ABBREVIATIONSText-conventions used to describe the spiral sculpture ofthe teleoconch (after Merle 1999, 2001)abisabapical secondary cord of the suturalramp;ABPabapical primary cord of the siphonalcanal;absabapical secondary cord of the siphonalcanal;adisadapical secondary cord of the suturalramp;ADPadapical primary cord of the siphonalcanal;adsadapical secondary cord of the siphonalcanal;Ddenticles of the outer lip;D1 to D6 abapical denticles;IDinfrasutural denticle;IPinfrasutural primary cord;MPmedian primary cords of the siphonalcanal;msmedian secondary cord of the siphonalcanal;Pprimary cords (first appearance sequence);P1shoulder cord;GEODIVERSITAS 2004 26 (2)P2 to P6ss1 to s6RepositoryMNHNprimary cords of the convex part of thewhorl;secondary cords (second appearancesequence);secondary cords of the convex part of thewhorl.Muséum national d’Histoire naturelle,Paris, domaine de collection Sciences de laTerre.Other abbreviations concerning the localities: seeAppendix 1.MATERIALNew material (V. B.) of Ocenebra erinaceus hasbeen sampled on the French coast (AtlanticOcean and Mediterranean Sea) in order to complete the available collections and to closely coverthe geographic range of the species (Fig. 1B). Thetidal zones have been closely prospected duringthe low tides for a better knowledge of its repartition. For each sample, the environmental characteristics (orientation of the sample stations,substrates, associated fauna and flora) have beenalso registered the soonest possible. The detailedlist of the studied populations including fossiland Recent O. erinaceus and O. brevirobusta ispresented in the Appendix 1. Among all populations, a random sampling of the 32, 10 or fiveselected specimens for the different analyses(cladistics, Fourier and Procrustes) has beenprocessed by the computer program MicrosoftExcel.METHODSIDENTIFICATION OF SCULPTURAL CHARACTERSAND STRUCTURAL HOMOLOGIESThe way to consider the muricid sculpture has adirect influence on the perception of their diversity and therefore on the results of morphologicalanalyses (Merle 1999). This observation is particularly true for cladistic studies, in which structural homologies should be defined in a previousstep of the analysis. For the axial sculpture, Miller267

Berrou V. et al.(1999) demonstrated that characters (constructional characters) are usually considered as similar, but may be built by different constructionalpathways and then should not be regarded ashomologous. For the spiral sculpture, the mostaccurate method to propose a hypothesis ofhomology consists in using a criterion of topological correspondence (Hylleberg & Nateewathana1992), in associating another criterion, the ontogenetic correspondence (homologous sequencesof appearance of the cords). This last criterion isfundamental because during the growth, the relative relief of cords may dramatically change andconsequently represents a pitfall when the topological position of these cords is only identifiedby the adult morphology (Merle 1999, 2001).The propositions of structural homologies arecoded by text-conventions used at familial level(Merle 1999, 2001, 2002; Houart 2000b,2001b, 2002a, b; Houart & Dharma 2001;Merle et al. 2001; Merle & Pacaud 2002a, b) andare given in the Abbreviations above.CLADISTIC ANALYSISThe objective of the cladistic analysis is focusedon the relationships between the populations ofOcenebra erinaceus, O. brevirobusta and Ocinebrellus inornatus. It will allow obtaining a phylogenetic hypothesis. The results will be confrontedwith the further analyses.The ocenebrine phylogeny is poorly studied(Vermeij & Vokes 1997) and only the works byKool (1993a, b), Merle (1999), Marko &Vermeij (1999), Oliverio & Mariottini (2001)and Oliverio et al. (2002) provide scattered data.Using anatomical characters, Kool (1993a, b)demonstrated that several genera, as NucellaRöding, 1798 (type species by subsequent designation [Stewart 1927]: Buccinum filosumGmelin, 1791, junior synonym of Nucella theobroma Röding, 1798 and senior objective synonym of Buccinum lapillus, Linnaeus, 1758; for acomplete discussion see Kool & Boss [1992]) andAcanthina Fischer von Waldheim, 1807 (typespecies: Buccinum monodon Pallas, 1774 by original designation), which were previously attributed to the subfamily Rapaninae Gray, 1853, are268in fact more closely related to Ocenebra(Ocenebrinae). This result is also reached byMerle (1999) using shell characters and byMarko & Vermeij (1999) using molecular data.At subfamilial level, Oliverio & Mariottini(2001), using molecular data (12S RNAsequence) suggested that Nucella may be regardedas a basal taxon, which is more closely related toPhyllonotus Swainson, 1833 (type species by subsequent designation [Swainson 1833]: Mureximperialis Swainson, 1833 non M. imperialisFisher von Waldheim, 1807 [new name: M. margaritensis Abott, 1958]; subfamily MuricinaeRafinesque, 1815) than Stramonita Schumacher,1817 (type species by original designation:Buccinum haemastoma Linnaeus, 1767; subfamilyRapaninae). Conversely, in a further study basedon the secondary structure ITS2, Oliverio et al.(2002) suggested, with reserve, that Nucella maybe regarded as sister group of the Rapaninae. Forthe present study, all the results are useful toselect two functional outgroups: N. lapillus andOcinebrina edwardsi (Payraudeau, 1826). Thesetaxa correspond to European ocenebrine speciessharing homologous structures and bearing cleardifferences with Ocenebra.One or several specimens of more than 30 populations of Ocenebra have been included in theanalysis. In the matrix, each specimen has beenregarded as a taxon. We have also carefullychecked that all specimens correspond to a samestep of growth (see Results: Growth of the outerlip). The matrix has been processed by the computer programs Winclada99 (Nixon 1999) andHennig86 (Farris 1988).BIOMETRIC ANALYSISThis approach consists in using linear measurements (raw values and indices), assessing to theproportions of the organism (shell), and describing the morphogenesis of a structure. For twostudied populations (O. erinaceus from Malagaand O. brevirobusta from Atlantic coast ofMorocco) containing numerous young specimens, the analysis of ontogenetic series willattempt to simulate the growth of a “consensualindividual”, for which the size will be assimilatedGEODIVERSITAS 2004 26 (2)

Shell variations in European Ocenebra (Mollusca, Muricidae)to the age. In many cases, the size and the age arelargely correlated and for the paleontologists, thesize represents a convenient estimation, but notwithout pitfalls (Dommergues et al. 1986;Allmon 1994). The retained measurements are:H (total height), HA (height of the aperture),WA (width of the aperture), HLW (height of thelast whorl), HS (height of the siphonal canal),WS (width of the shoulder), HP1 (height of theP1 cord spine), TP1 (thickness of P1), and BCS(distance between the base and the old siphonalcanal) (Fig. 2). The bivariated plots provideinformations on the type of growth. They arebased on the law of allometric growth having asequation Ly ß(Lx)α or log Ly α log Lx log ß;Lx and Ly corresponding to the sizes of organs,a and b corresponding to the constants.PROCRUSTES AND FOURIER ANALYSESThe muricid shells bear a complex morphologycharacterized by varices and cords varying intheir size, form and number. For this reason,the methods of traditional biometry are insufficient to fully delineate intra- and interpopulation differences among the studied Ocenebra.Therefore, the Procrustes and the Fourieranalyses are used in order to complete the biometrical study. Considering the technicalnecessities of these methods and time-consuming mathematical treatments, the followingprotocol is adopted:– the number of localities is prioritized to thenumber of specimens among each population inorder to cover the widest geographic range.Consequently, a limit of five to 10 specimens hasbeen retained for the Procrustes and the Fourieranalyses;– in order to minimize effects of ontogenetic variations, only adults specimens sharing a relativeequivalent size and a same degree of constructionof the outer lip (see Results: Growth of the outerlip) have been considered in the analyses;– the landmarks in ventral view (Procrustesmethods) and the outlines in ventral and dorsalviews (Fourier transform) have been digitalizedusing drawings made with a camera lucida, andthen have been scanned.GEODIVERSITAS 2004 26 (2)WSHTP1HLWHP1HSBCSF IG . 2. — Linear measures used for the biometric study ofOcenebra Gray, 1847. Abbreviations

Quaternary and Recent shells of Ocenebra erinaceus (Linnaeus, 1758) . Murex eurypteronReeve, 1845 [a junior synonym of M. aduncusSowerby, 1834]), but has been recently transferred (Houart . s1 to s6 secondary cords of the convex part of the who