JOURNAL OF MORPHOLOGY 214:l-41 (1992) Skeletal Development in Xenopus laevis (Anura: Pipidae) LINDA TRUEB AND JAMES HANKEN Museum of Natural History, and Department of Systematics and Ecology, The University ofliansas, Lawrence, Kansds 66045-2454 (L.T.)and Department of Environmental, Population, and Organismic Biology, University of Colorado, Boulder, Colorado 80309-0334 (J.H.) ABSTRACT Postembryonic skeletal development of the pipid frog Xenopus laevis is described from cleared-and-stained whole-mount specimens and sec- tioned material representing Nieuwkoop and Faber developmental Stages 46-65, plus postmetamorphic individuals up to 6 months old. An assessment of variation of skeletogenesis within a single population of larvae and comparison with earlier studies revealed that the timing, but not the sequence, of skeletal development in X. laevis is more variable than previously reported and poorly correlated with the development of external morphology. Examination of chondrocranial development indicates that the rostral cartilages of X. laevis are homologous with the suprarostral cartilages of non-pipoid anurans, and sug- gests that the peculiar chondrocranium of this taxon is derived from a more generalized pattern typical of non-pipoid frogs. Derived features of skeletal development not previously reported for X. laevis include 1) bipartite forma- tion of the palatoquadrate; 2) precocious formation of the adult mandible; 3) origin of the angulosplenial from two centers of ossification; 4) complete erosion of the orbital cartilage during the later stages of metamorphosis; 5) development of the sphenethmoid as a membrane, rather than an endochon- dral bone; and 6) a pattern of timing of ossification that more closely coincides with that of the pelobatid frog Spea than that recorded for neobatrachian species. o 1992 Wiley-Liss, Inc. The literature describing skeletal develop- observations on laboratory-reared frogs. Os- ment in the pipid frogxenopus laevis is exten- sification sequence was tabularized for 32 sive and, in recent years, the species has cleared-and-single-stained specimens rang- become a "model system" for experimental ing in age from 20-365 days and arranged in studies of skull development (e.g., on laryn- four stages: premetamorphic Stages I (20-29 geal development: Sassoon and Kelley, '86; d) and I1 (31-39 d), metamorphic Stage I11 Sassoon et al., '86; Kelley et al., '89; on Meck- (40-49 d), and postmetamorphic Stage IV el's cartilage: Thomson, '86, '87, '89; on tooth (56-365 d). Bernasconi ('51) did not define development: Shaw, '79, '85, '86, '88). The his stages with morphological criteria; thus first accounts of the skull were provided more they are not easily compared to those of than a century ago by Parker (1876, 1881; = Nieuwkoop and Faber ('56). Moreover, no Dactylethra capensis, auctorum). The devel- data are provided on the development of the opment of the hyobranchial skeleton in X. cartilaginous precursors of the bony skele- laevis was described by Ridewood (1897) and ton, and Bernasconi's ('51) illustrations re- that of the larval chondrocranium by Kott- veal several errors and misinterpretations of haus ('33).Paterson ('39) reviewed the devel- the anatomy of the developing and adult skull. opment of the head of larval and adult X. A detailed account of the ontogenesis of laevis and corrected several errors of Kott- the vertebral column of Xenopus laevis was haus ('331, which unfortunately had been provided by Smit ('53), who examined sec- perpetuated by de Beer ('37) in his book on tioned specimens representing Stages 24-66 the development of the vertebrate skull. of Nieuwkoop and Faber ('56) of individuals The first comprehensive description of cra- collected from natural populations in South nial and postcranial ossification in Xenopus Africa and staged by J. Faber. As part of their laevis was by Bernasconi ('51),who based his study of the visceral arches, larynx, and vis- o 1992 WILEY-LISS,INC 2 L. TRUEB AND J. HANKEN ceral muscles of Xenopus laevis, Sedra and lessens the utility of the normal tables for Michael (’57) described changes in the skull staging wild populations.” during metamorphosis. Their specimens also Given these extensive and diverse research were collected from wild populations in South efforts, one would expect the osteology and Africa and comprise a total of 12 Nieuwkoop development of Xenopus laevis to be well and Faber (’56) stages: viz., Stage 55 (232 d; known. However, comparison of figures and “typical larval stage”), Stages 56-57 (k38-41 descriptions in Kotthaus (’33),Paterson (’39), d; “premetamorphic” stages before the expo- Weisz (’45a,b), Bernasconi (’51), Nieuwkoop sure of the forelimbs), Stages 58-65 (k44-54 and Faber (’56), Sedra and Michael (’571, d; “metamorphosing” stages), and Stage 66 Deuchar (’751, and Reumer (’85)reveals sig- (258 d; “newly metamorphosed”). The au- nificant discrepancies in the terminology of thors sectioned the heads of an unspecified both the larval chondrocranium and the adult number of specimens and based their descrip- skeleton. Moreover, none of these authors tions and illustrations on hand-generated described the striking differences in the devel- graphical reconstructions from these sec- oping and adult cranium of X. laevis as com- tions. Given the tedious nature of this method pared to non-pipidl anurans. and the absence of any statements about Sokol (’75) pointed out that the ethmoidal variation, it seems likely that Sedra and region of the skull of larval pipids is distinct Michael’s (’57) descriptions may be based on from that of all other anurans; he suggested only a single specimen of each stage. that on the basis of this and many other Cranial osteological data from Sedra and characters (e.g.,opercular structure, hyobran- Michael’s study were included among the chid and filter apparatus, prootic foramina criteria used to define the developmental and their nerves, palatoquadrate suspensori- stages in the widely used normal table of um), pipoid (i.e., Rhinophrynidae + Pipidae) development of Xenopus laeuis (Nieuwkoop larvae were derived with respect to all other and Faber, ’56). This work also contains gen- anurans exclusive of the Microhylidae. The eral, but abbreviated, narrative accounts of peculiarities of the ethmoidal region with the development of various parts of the skele- respect to the formation of the nasal capsule ton. These were written by a variety of con- in pipids were emphasized by RoEek (’89)and tributors and are based on examination of development of the ethmoidal area in Pipa sectioned material obtained from natural pop- pipa was described by RoEek and Veseljl(’89). ulations. The editors (Nieuwkoop and Faber, RoEek (’90) likened the ethmoidal endocra- ’56:7) carefully pointed out that intrastage nium of pipids to that of “labyrinthodont- variation was not studied, but that “It has grade” tetrapods, claiming that the two lar- . become evident . that a general varia- val ethmoidal configurations in anurans (i.e., tion of approximately half a stage to either the pipoid and non-pipoid types) could not side has to be taken into account.” have been derived from one another (contra Variation in osteological development in Sokol, ’75). Xenopus laevis was addressed briefly by An understanding of the structure of the Brown (’go), who examined small samples of mature skeleton (especially the cranium) re- wild-caught and laboratory-reared animals quires knowledge of its development. Fur- between Stages 49 and 60. Brown noted dis- ther, one can expect interspecific osteological crepancies in the timing and sequence of variation among adults to be correlated with ossification of certain elements when he com- variation in developmental patterns. In the pared his results with those of Bernasconi case of pipid frogs, the studies of Xenopus (’51). Brown (’80) reported that ossification laeuis clearly have historical precedence in is delayed in laboratory-reared X. laeuis as the literature as a baseline against which compared with wild-caught specimens and noted specific differences between his ob- ‘Throughout this paper, reference will be made to pipid and served timing and sequence of osteological pipoid frogs versus non-pipid and non-pipoid frogs. There are five Recent genera in the Family Pipidae-viz., Pipa in the New development in wild-caught X. laevis and World, Hymenochirus, Pseudhymenochirus, Xenopus, and Sih- those recorded by Nieuwkoop and Faber (’56). runu in the Old World (Cannatella and Trueb, ’88a,b).The super- family Pipoidea is composed of the Pipidae plus its sister taxon Brown (’80:28-29) correctly observed that Rhinophrynidae (containing only one Recent taxon, the New there is “considerably more variation in time World Rhinophrynus dorsalis).Currently, pipoid frogs are consid- ered to be a highly derived group of primitive frogs that are of ossification than the one half stage stated thought to be most closely related to pelohatoid murans (e.g., in the normal table . [and that this] . Pelobatidae and Pelodytidae) (Cannatella,’85). SKELETAL DEVELOPMENT IN XENOPUS LAEVIS 3 other pipid data will be compared. However, TABLE 1. Comparison of three common normal tables several problems must be resolved before of development for anurans' such comparisons can be made. Disparities Taylor and Nieuwkoop and among published anatomical descriptions Kollros ('46) Faber ('56) Gosner ('60) must be resolved and intraspecific variation I 46 26 in the timing and sequence of osteogenic I1 47 27 events should be evaluated. - 48 - Herein, skeletal development of Xenopus I11 49 28 - 50 - laevis from Stage 46 (premetamorphic, feed- IV 51 29 ing larva) through Stage 66 (metamorphosed V 52 30 froglet) and for 6 additional months is de- VI 53 31 scribed and compared with results of previ- VII - 32 VIII - 33 ous studies on X. laevis and other anuran M 54 34 taxa. The account is organized into four ma- X 55 35 jor sections-chondrocranial development, XI - 36 metamorphic development of the cranium, XI1 - 37 XI11 56 38 postmetamorphic development of the cra- XIV 57 39 nium, and postcranial development.