Sarsia marii n . sp . ( Hydrozoa , Anthomedusae ) and the use of 16 S rDNA sequences for unpuzzling systematic relationships in Hydrozoa *

Molecular data and parsimony analysis have become powerful tools for the study of phylogenetic relationships among extant animal taxa. In particular, DNA sequence data have added important and surprising information on the phylogenetic relationships at almost all taxonomic levels in a variety of animal groups (for references see Avise, 1994; DeSalle and Schierwater, 1998). We may not forget, however, that inferring phylogenetic relationships from molecular data is neither trivial nor a final solution to systematics, since the molecular option can be prone to a large number of pitfalls and misunderstandings (for refs. see Miyamoto and Cracraft, 1991; Schierwater et al., 1994; Swofford et al., 1996). Nevertheless molecular data are especially useful or even indispensable in groups where other characters, like morphological data, are limited or hard to interpret. One classical problem to animal phylogeny is the evolution of cnidarians and the groups therein (Hyman, 1940; Brusca and Brusca, 1990; Bridge et al., 1992, 1995; Schuchert, 1993; Schierwater, 1994; Odorico and Miller, 1997). Particularly the systematic relationships within the Hydrozoa represent one of the most difficult and long-standing problems to evolutionary biologists. In lack of fossil records and a sufficient number of unambiguous morphological characters, neither the phylogenetic relationships between nor within the orders have been understood and are still subject of controversial debates (see for SCI. MAR., 64 (Supl. 1): 117-122 SCIENTIA MARINA 2000


INTRODUCTION
Molecular data and parsimony analysis have become powerful tools for the study of phylogenetic relationships among extant animal taxa.In particular, DNA sequence data have added important and surprising information on the phylogenetic relationships at almost all taxonomic levels in a variety of animal groups (for references see Avise, 1994;DeSalle and Schierwater, 1998).We may not forget, however, that inferring phylogenetic relationships from molecular data is neither trivial nor a final solution to systematics, since the molecular option can be prone to a large number of pitfalls and misunderstandings (for refs. see Miyamoto and Cracraft, 1991;Schierwater et al., 1994;Swofford et al., 1996).Nevertheless molecular data are especially useful or even indispensable in groups where other characters, like morphological data, are limited or hard to interpret.
One classical problem to animal phylogeny is the evolution of cnidarians and the groups therein (Hyman, 1940;Brusca and Brusca, 1990;Bridge et al., 1992Bridge et al., , 1995;;Schuchert, 1993;Schierwater, 1994;Odorico and Miller, 1997).Particularly the systematic relationships within the Hydrozoa represent one of the most difficult and long-standing problems to evolutionary biologists.In lack of fossil records and a sufficient number of unambiguous morphological characters, neither the phylogenetic relationships between nor within the orders have been understood and are still subject of controversial debates (see for example, Petersen, 1979Petersen, , 1990;;Cunningham and Buss, 1993;Schuchert, 1996).Historically, hydroid systematics has been hindered by problems of deciding the importance of polyp versus medusoid characters (Rees, 1957;Naumov, 1960;Boero, 1980;Bouillon, 1981;Boero and Bouillon, 1987).Since in most hydroids the medusa stage has been reduced to sessile gonophores, reconstructions of phylogenetic relationships have mainly been based upon morphological characters of the hydroid stage and the assumption of a progressive reduction of the medusoid stage (Werner, 1984;Boero and Sarà, 1987).
Here, the arrangement of tentacles and the inventory of nematocysts have been two of the most important characters.The morphological simplicity and plasticity of hydroids, however, makes morphological homoplasy likely to be common in this group.We suggest that molecular characters must be added to morphological and life cycle characters in order to resolve hydrozoan systematics.By applying molecular data to previously undescribed species we might in many cases be able to resolve their relative taxonomic position even in the absence of morphological or developmental data.This approach promises valuable results if molecular data from closely related taxa are available for comparative analyses.We present an example from a new athecate hydroid species which we assign to the genus Sarsia based on both morphological and molecular data.

Animal material
Based on morphological characters we used the following closely related species of the infra-order Capitata (Werner, 1984) for embedding the new species into a taxonomic framework: Stauridiosarsia producta Wright (1858), Sarsia tubulosa Sars (1835), Sarsia reesi Vannucci (1956)  Hydroids of Eleutheria dichotoma, Cladonema radiatum and Sarsia marii n. sp. were collected by the authors from the green alga Ulva lactuca, obtained 10-50 cm under the water surface at the rocky shore of the Laboratoire Arago, Banyuls sur Mer, France.Single polyps were isolated from the algae and grown in small finger bowls in the lab using our standard culturing conditions for hydrozoans (Schierwater et al., 1992;Schierwater and Hadrys, 1998).Polyps were maintained in artificial seawater (30‰ salinity) at 19-21°C and fed ad libitum two times weekly on 3-4 day old brine shrimp larvae (Artemia salina).Under these conditions polyps of all three species, including Sarsia marii, grew into colonies of polyps connected by a horizontal net of stolons (the latter grew attached to the glass surface of the finger bowl).While medusae of Eleutheria dichotoma and Cladonema radiatum are found regularly year round, Sarsia marii medusae were found only once in spring 1998.Since the latter stayed alive just for a few days, morphological traits could be observed on subadult medusae only.
The original polyp was found on a thallus of the green alga Ulva lactuca collected on 5 April 1997 at the Laboratoire Arago, Banyuls sur Mer, France, 42°29'N, 3°08'E.

DNA sequence alignment and phylogenetic analyses
16S rDNA sequences were aligned with the aid of CLUSTAL (Higgins and Sharp, 1989) and ambiguous regions (often surrounding variable loops) were deleted.Since computer based alignment programs did not reveal consistent alignments, sequences were controlled by eye and compared to a secondary structure model of the Eleutheria dichotoma 16S rRNA molecule (Ender, 1997).A detailed description of the 16S rDNA data collection and analyses will be given elsewhere (together with a larger hydrozoan tree; Ender and Schierwater, in prep.).For tree inference by maximum parsimony or neighbor-joining data were analyzed using PAUP 3.1 (Swofford, 1993) andPHYLIP 3.75c (Felsenstein, 1995), respectively.

Sarsia marii n. sp.
The morphology of Sarsia marii polyps and medusae (Fig. 1) suggests relationships to the Cladonematidae and Corynidae (cf.Werner, 1984;Petersen, 1990).The polyp of S. marii looks rather similar to a Cladonema radiatum polyp and exhibits in addition to the oral whorl of capitate tentacles a second whorl of 4 filiform sensory tentacles.The number of capitate tentacles was always four, with each 10-11 endodermal cells (Table 1).The medusa looks like a typical Sarsia medusa and lacks any Cladonema typical tentacle branching patterns, suggesting that Sarsia marii is related to the genus Sarsia.Here the new species appears to be very similar to S. reesi (synonym: Dipurena reesi Vannucci, 1956;Brinkmann-Voss and Petersen, 1960), both in the polyp and medusa generation (Table 1).The typical long manubrium of S. reesi medusae, however, was not found in the immature medusae of S. marii.The most striking difference between S. marii and S. reesi medusae is found in the catching tentacles, which are capitate in S. marii and filiform in S. reesi.TABLE 1 -Distinguishing morphological traits for Cladonema radiatum, Sarsia reesi (Brinkmann-Voss and Petersen, 1960;Schuchert, 1996) and Sarsia marii n. sp., Filifera).Shown is the single most parsimonious tree (PAUP, Branch-and-Bound search, 500 bootstrap replicates, reweighting of characters RC=0.734,CI=0.847,RI=0.867).Note, the shown clade is a monophyletic clade within a more comprehensive tree that includes 27 hydrozoan species (Ender and Schierwater, in prep.).The number on nodes corresponds to the bootstrap value and branch lengths are drawn proportionally to the number of changes on each branch.The tree supports a close relationship between known Corynidae and the new species Sarsia marii, which appears within the Sarsia group.Note, that the Corynidae form a paraphyletic clade which varies according to the outgroup used.Please also note that the Staurocladia polyp shown is the drawing of a Staurocladia wellingtoni polyp (Schuchert, 1996) since polyps of Staurocladia bilateralis and S. oahuensis have not been found yet.

Characters
While it cannot be excluded from organismal observations alone that the observed differences between S. reesi and S. marii medusae are the result of phenotypic plasticity (we have not seen any mature S. marii medusae yet), 16S rDNA sequence data unambiguously identify them as different species.

Taxonomic placement
The 16S rDNA phylogram shown in Figure 2 supports the placement of S. marii within the genus Sarsia.The given branching patterns are the result of a branch and bound search using character reweighting according to the rescaled consistency index, and using bootstrap analyses with 500 resampling replicates.In the single most parsimonious tree (Fig. 2) S. marii clusters as a sister species to S. tubulosa.The genus Sarsia forms a paraphyletic group, depending on the outgroup used.The latter reflects the diversity -and consequently also the classification difficulties -of genera within the Corynidae, which are well known for their considerable morphological variability (including the number, placement and type of tentacles of hydroids; cf.Boero and Bouillon, 1987).For instance, the genera Sarsia and Dipurena were distinguished by a single character (gonads in one ring or gonads in two or more rings; Kramp (1961), and the genus Stauridiosarsia from Sarsia solely by the presence or absence, respectively, of aboral filiform tentacles (Bouillon, 1985).
The genus Sarsia appears near the base of a clade that gave rise to the invention of walking (Fig. 2).The derived genera Cladonema, Staurocladia, and Eleutheria all exhibit branched tentacles with specialized tentacles for attaching the medusae temporarily to a substrate.The tree supports a placement of Stauridiosarsia close to the family Corynidae.It is noteworthy that the above conclusions are supported by high bootstrap values and also neighbor-joining analysis (not shown), which revealed the same branching patterns as shown in Figure 2.

DISCUSSION
Hydroid taxonomists have traditionally faced difficulties many other taxonomists have not, including controversies over the relative importance of polypoid versus medusoid characters and the lack of one of the two bauplans in many groups.Furthermore, the independent gains and losses of morphological features in any one generation may produce high levels of morphological homoplasy (convergence), especially since the number of morphological characters in Hydrozoa is severely limited (cf.Boero, 1980;Boero and Bouillon, 1987;Petersen, 1979Petersen, , 1990)).The main advantages of adding molecular data to phylogenetic analyses of Hydrozoa include that molecular characters are independent of any developmental or life cycle stage and can be generated at cum grano salis unlimited numbers.Our example, in which we describe a new species, Sarsia marii, demonstrates the usefulness of 16S rDNA sequence data for verifying the taxonomic placement for a new species and simultaneously resolving its phylogenetic relationships to closely related species.
Both the morphology and the 16S rDNA sequence data provide unambiguous evidence that Sarsia marii n. sp.must be regarded as a member of the Corynidae, closely related to Sarsia tubulosa.Based on morphological data alone this taxonomic placement would be less clear, especially since we are yet lacking knowledge of the morphology of sexually mature medusae.However, it was the morphology which allowed us to identify closely related species for comparison in a molecular tree.The tree itself is widely congruent with our expectations from morphological data, and seems to be well supported by the high bootstrap values, and the independence of the tree topology from the tree building algorithm used.Furthermore, the clade discussed here also remains stable in a larger scale phylogeny, which includes a total of 27 hydrozoan species (Ender and Schierwater, in prep.).Nonetheless we would like to note that any molecular phylogeny is sensitive for example to the selection of taxa, to the choice of the outgroup, to the algorithm used for tree reconstruction, and in case of DNA sequence data, particularly also to the alignment of putative homologous nucleotide positions (for refs.see DeSalle and Schierwater, 1998).While future progress is to be expected -and needed -with respect to finding the most robust means for analyzing DNA sequence data, it seems clear that 16S rDNA data provide informative characters for unpuzzling phylogenetic relationships in Hydrozoa at different taxonomic levels.
FIG. 1. -Drawings of live specimens of Sarsia marii n. sp. raised in the laboratory.A polyp with a well developed medusa bud above the aboral tentacle whorl and a four-day-old medusa are shown.Explanations are given in the text.