Phylogenetic Classification of the Halichondrids (Porifera, Demospongiae)
Total Page:16
File Type:pdf, Size:1020Kb
FAU Institutional Repository http://purl.fcla.edu/fau/fauir This paper was submitted by the faculty of FAU’s Harbor Branch Oceanographic Institute. Notice © 1990 Institute for Systematics and Population Biology (Zoological Museum), University of Amsterdam. This manuscript is an author version with the final publication available and may be cited as: Van Soest, R. W. M., Diaz, M. C., & Pomponi, S. A. (1990). Phylogenetic classification of the halichondrids (Porifera, Demospongiae). Beaufortia, 40(2), 15-62. BEAlJFOR TIA INSTITUTE OF TAXONOMIC ZOOLOGY (ZOOLOGICAL MUSEUM) UNIVERSITY OF AMSTERDAM Vol. 40, no. 2 July27, 1990 PHYLOGENETIC CLASSIFICATION OF THE HALICHONDRIDS (PORIFERA, DEMOSPONGIAE) R. W. M. VAN SOEST*), MARIA CRISTINA DIAZ") & SHIRLEY A. POMPON!***) *) Institute of Taxonomic Zoology, University of Amsterdam, P. 0. Box 4 766-1009 AT Amsterdam, the Netherlands **) Marine Sciences, U. C. Santa Cruz, 273 Applied Sciences Building, Santa Cruz, California 95064, U.S.A. ***)Harbor Branch Oceanographic Institution Inc., 5600 N Old Dixie Highway, Fort Pierce, Florida 34946, U.S.A. ABSTRACT Using a multicharacter approach and numerical cladistic computer programs a phylogenetic analysis is made of a newly defined order Halichondrida (which includes all Halichondrida and parts of the Axinellida sensu L6vi, 1973), with emphasis on the newly defined family Halichondriidae (which includes the families Halichondriidae and Hymcniacidonidac sensu LCvi, 1973, and some genera formerly assigned to the Axinellidae and Desmoxyidae). The newly defined order consists of the "old" families Axinellidae/Bubaridae and Desmoxyidac, a "new" family Dic tyonellidae (containing Dict_yonel/a, Acanthella, Liosina and Dactylella), and the new concept of the family Halichondriidae. On the basis of numerous specimens, including many type and other original material from all parts of the world oceans, the latter group has been analyzed in depth. Next to classical morphological characters also biochemical data (secondary metabolites) were used. This resulted in the recognition of several discrete generic groups, viz. Axilu>ssa + Colloralypta, A1yrmekioderma + Didiscus, Spongosorites, TojJscntia + Epipolasis + Petromica, Ciocalypta + Amorphinopsis, and Hymcniacidon + Hafichondria. A revision comprising definitions of and remarks on all genera, including all junior synonyms, of the newly defined Halichondriidae concludes the paper. INTRODUCTION at the subclass level and if used at lower levels are suspect as likely adaptive characters easily The Halichondrids s.s., until very recently rec developed independently. ognized as a separate order (cj. Levi, 1973; If reproductive strategies are abandoned as Bergquist, 1978; Hartman, 1982) are a group characters at the subclass-level, it is necessary heavily affected by the major changes of to reevaluate the contents and the position of classification (based on reproductive strategies) the Hali'thondrids. Characters of the sponges introduced by Levi (1953), which subsequently involved are discussed and evaluated, and a gained general acceptance. It has been argued new classification based on (numerical) cladistic elsewhere (e.g. Van Soest, 1987, 1990b; Diaz, techniques is presented. For a general outline of eta!., 1990a), that reproductive strategies can these methods see Wiley ( 1981); for sponge not be employed as synapomorphous characters examples see Van Soest (1987, 1990a and b). 15 The newly recognized valid families and genera (Axinellida) - Acanthella (?) - Dictyonella (?) - and their junior synonyms are listed, defined, Axinyssa (?) - Myrmekioderma (?) - Topsentia and illustrated, based on many type and other (Halichondrida) - Ciocalypta (Haliehondrida) - original specimens from all over the world Halichondria (Halichondrida). In fact, the oceans. definitions of the "orders" Axinellida and This study is the fourth part of a revision of Halichondrida apply only really well to the type the Halichondrids undertaken by Pomponi, genera Axinella and Halichondria, and it is only Dfaz & Van Soest. Preliminary papers on this through other characters (e.g. sinuous spicules revision have been published by Diaz et al. in Phakellia and Acanthella, and Axinella cannabina (1990a) and Pomponi et al. (1990), and a review (Esper, 1794); telescoped endings of spicules in of Central Atlantic representatives of the newly Axinella, Dictyonella, Axinyssa and Spongosorites; defined family Halichondriidae was made by size ranges of oxea in Myrmekioderma, Topsentia, Diaz et al. (in press). Ciocalypta and Halichondria), that typical genera are linked with intern1ediates. From these observations, it may be inferred that to provide AXINELLIDA-HALICHONDRIDA RELA definitions of two discrete groups (orders or TIONSHIPS suborders) encompassing all the associated To perform a cladistic analysis of the families genera is virtually impossible. It will not be and genera of the Halichondrids, it is necessary attempted here; instead it will be shown below to find the nearest outgroups which enables that the group of genera together comprising character polarization (Watrous & Wheeler, the presently recognized Axinellida s. s. and 1980). Van Soest (in the press) and Diaz et al. Halichondrida s. s. consists of four definable (in the press) showed that Halichondrids and genus-groups which can conveniently be given some Axinellida share many characteristics. family status. Also, pre-Levi (1953) classifiers (Topsent, Bur The phylogenetic position of Raspailiids ton, De Laubenfels) associated Halichondrids (including Euryponids) and Hemiasterellids with Axinellids, particularly the family has to be discussed briefly, since these have Axinellidae. In fact, both were considered been considered close to Axinellidae in the past, members of a single order Halichondrida and indeed showed up to be close in a recent Vosmaer, 1887 by Topsent and De Laubenfels, phylogenetic exercise (Van Soest, 1987), based until the Axinellida were given separate ordinal on possesswn of an axially condensed status in 1955. Halichondrida s.s., i.e. families choanosomal skeleton. Ongoing cladistic Halichondriidae and H ymcniacidonidac sensu analyses (Van Soest, 1990b), however, reveal Levi (1973), can be associated with Axinellida the probable homoplasic nature of the axially Levi (1955) (in part only: familes Axinellidae, condensed skeleton. Typical members of Bubaridae, and perhaps some Desmoxyidae, Axinellidae, Raspailiidae, Hemiasterellidae but excepting the Raspailiidae, Euryponidae, and Sigmaxinellidae share the axial condensa Rhabderemiidae, Trachycladidae, Sigmaxinel tion of the choanosomal skeleton; however, lidae ( ~ Hamacanthidae) and Hemiastercl each of them also shares indubitable lidae), by a convincing series of intermediate synapomorphous characters with widely dif genera and families. The major features of the ferent groups such as Hadromerids Axine1lida, i.e. axial condensation of the (Hemiasterellid asters) and Poecilosclerids choanosomal skeleton and indiscriminate (Raspailiid acanthostyles and Sigmaxinellid occurrence of styles and oxea, and the major sigmata). There is no solution for this distribu features of Halichondrida, I.e. ectosomal tion of characters other than to assume that one tangential skeleton and confused choanosomal or more homoplasic developments have occur skeleton, intergrade nicely in the series of red in the evolution of these groups. According genera Axinel/a (Axinellida) Plzake/lia to recent cladistic analyses (Van Soest, 1990b), 16 it is more parsimonious to assume that axial E 0"' 0 0 0 ..: ~ condensation has developed in four different ~ "' Q) 'C " a. ~ "' 0 "'Q) lines (Axinellidae, Hcmiasterelliidac, Raspailii Q) -f- J:"' J:"' c. J:"' ~ dae and Sigmaxincllidae), than to assume that asters, acanthostyles and sigmata have each developed independently in two separate lines (asters: Hadromerids and Hemiasterellids; acanthostyles: Raspailiids and Micro cionids/Myxillids; sigmata: Sigmaxincllids and Poccilosclerids). Speculations about presence of any of those characters in unknown ancestors of the groups, and subsequent independent losses in several major lines, seem to be all equally ( un-)parsin1onious and unnecessary. These results arc in support of Hooper (1990), who correctly pointed out the sitnilarities of Fig. 1. Revised classification of the Dcmospongiae based on the reevaluation of major group synapomorphies Raspailiids and Microcionids. h1oreover, axial (based on Van Socst, l990b). Character A: (sub-)radiatc condensation of a sort is found in several architecture, B: anisotropic reticulation, C: taurin syn Suberitid Hadrmnerids, viz. species of Rhizax thesis via cysteic acid, D: sigmatose microscleres. inella and the species originally described as Phakellia lobata Wilson, 1902. Comparisons of axially condensed species of all these groups show in1portant architectural differences vide strong (underlying) synapomorphies between various axial skeletons (see Van Soest, (Pomponi, el al., in press). 1990b), testifying for the probable homoplasic nature of this type of supporting skeleton. The Definition of the Halichondrida s .I.: revised classification of the Hadromerid DemosjJongiae with a plwnoreticulate skeletal architec Axinellid-Halichondrid-Poccilosclerid ture built of interchangeable styles and oxea and Demospongcs is presented in Fig. 1 (after Van intermediate spicules, of widely diverging sizes, and not Socst, 1990b). Junctional!;• localized. A problem remaining is the formal definition Remarks: