s
ell
Abte
3, Smithsonian Institution, P.O. Box 37012, Washington, DC 20013-7012, USA
ians-
, and are remarkable for their
y, and aesthetically appealing
07). Leys et al. (2007, p. 117)
since
; Mens
nd Col
single erroneous or possibly pseudogenic 18S rDNA sequence from
GenBank (see also Voigt et al., 2008). This same sequence may also
have had an adverse effect on the rest of the phylogeny.
Here, we have addressed these issues by including 10 addi-
tional species that were previously unavailable to us (i.e., Dohr-
Euretidae. Whereas the position of Farrea did not change, we
?nd that Lyssacinosida is monophyletic but two of the three
subfamilies of Euplectellidae are not. Implications of these re-
positional homology introduced by the new sequences. The ?nal
concatenated alignment [TreeBase (www.treebase.org) matrix
accession number M4266] consisted of 3426 bp for 60 taxa, includ-
ing the 17 outgroup species (from Choano?agellata, Demospon-
giae, Homoscleromorpha, Calcarea, Placozoa, and Cnidaria) used
in our previous study (see Dohrmann et al., 2008). The total
amount of missing data decreased from 18.75% to 16.82% because
of the more complete sequencing of Farrea and one Hyalonema spe-
cies (see Table 1).
* Corresponding author. Fax: 49 89 21806602.
Molecular Phylogenetics and Evolution 52 (2009) 257?262
Contents lists availab
ne
.e lsE-mail address: woerheide@lmu.de (G. W?rheide).et al., 2008), and description of many more new species and even
genera is in progress, suggesting that biodiversity of glass sponges
is much higher than currently appreciated. A well-resolved phylog-
eny of this important yet understudied group of sponges is highly
desirable in order to understand how this diversity evolved.
The ?rst molecular phylogeny of Hexactinellida (based on three
ribosomal DNA markers) has only recently been published, and
revealed a surprisingly high level of congruence with morphol-
ogy-based systems (Dohrmann et al., 2008). However, monophyly
of the most species-rich order, the Lyssacinosida, remained ambig-
uous, being model dependent. Furthermore, limited taxon sam-
pling prevented resolution of subfamily-level relationships
among the Euplectellidae (the ?venus-?ower basket? family). Final-
ly, the accuracy of the placement of Farreidae (Sceptrulophora) was
also questionable because this taxon was only represented by a
2008) and highlight the non-trivial relationship of taxon sam-
pling and model choice in deep metazoan phylogenies inferred
from rDNA data.
2. Materials and methods
2.1. New specimens and sequences
Taxonomic and collection information of the new specimens,
and sequence accession numbers, are given in Table 1. Full-length
18S, and partial 28S and mitochondrial 16S rDNA sequences were
obtained as described, and added to the previously reported align-
ments (see Dohrmann et al., 2008). A few additional sites (9 bp
overall) had to be excluded due to further ambiguities regardingof benthic deep-water communities
unique tissue organization, physiolog
skeletal morphology (Leys et al., 20
reported 531 described species, but
increased to 550 (Lopes et al., 2007
Reiswig et al., 2008; Tabachnick a1055-7903/$ - see front matter  2009 Elsevier Inc. A
doi:10.1016/j.ympev.2009.01.010then this number has
henina et al., 2007a,b;
lins, 2008; Tabachnick
sults for evolution of morphological characters are discussed.
We also elaborate further on comparison of RNA (paired-sites)
substitution models (see Savill et al., 2001; Dohrmann et al.,Glass sponges (Porifera, Hexactinellida) are major components sentative of the most species-rich family of Sceptrulophora, theShort Communication
New insights into the phylogeny of glass
Monophyly of Lyssacinosida and Euplect
position of Euretidae
M. Dohrmann a,c, A.G. Collins b, G. W?rheide c,*
aGeorg-August-Universit?t G?ttingen, Geowissenschaftliches Zentrum G?ttingen (GZG),
bNMFS, National Systematics Laboratory, National Museum of Natural History, MRC-15
cDepartment f?r Geo- und Umweltwissenschaften & GeoBioCenterLMU, Ludwig-Maximil
a r t i c l e i n f o
Article history:
Received 7 October 2008
Revised 3 January 2009
Accepted 23 January 2009
Available online 30 January 2009
1. Introduction
Molecular Phyloge
journal homepage: wwwll rights reserved.Universit?t M?nchen, Richard-Wagner-Str. 10, 80333 M?nchen, Germany
 2009 Elsevier Inc. All rights reserved.
mann et al., 2008). Among them was a Farrea species that
yielded high-quality sequences of all three markers, and a repre-ponges (Porifera, Hexactinellida):
inae, and the phylogenetic
ilung Geobiologie, Goldschmidtstr. 3, 37077 G?ttingen, Germany
le at ScienceDirect
tics and Evolution
evier .com/locate /ympev
ch O
m A
but
eale
(se
enet2.2. Phylogenetic analyses
Models for the loop partitions and 16S were selected based on
the Akaike Information Criterion (AIC; Akaike, 1974) as described
(Dohrmann et al., 2008). Results were the same as with the previ-
ous alignments, except that the inclusion of a proportion of invari-
able sites (+I) was now favored for the 16S partition. For stem
regions, we employed the four paired-sites models (RNA6A, -6B,
-7A, and -7D; see Savill et al., 2001) investigated previously and as-
sessed their ?t to our data with Bayes factors (Kass and Raftery,
1995; see Dohrmann et al., 2008 for details). In addition, we also
ran an analysis under a 16-state model to assess whether discrim-
inating between all possible base pairings could improve model ?t
(in 6- and 7-state models, mismatch base pairings are ignored or
lumped into a single state, respectively; see Savill et al., 2001).
We chose the RNA16C model because it is equivalent to models
7D and 6B (Savill et al., 2001); more general variants (16A, 16D)
would have probably resulted in over-parameterization as this
was already the case with the most general 6- and 7-state models
6A and 7A (see Results and Dohrmann et al., 2008). Phylogenetic
analyses were performed with PHASE v2.0 (http://www.bio-
inf.manchester.ac.uk/resources/phase/) essentially as in Dohrmann
et al. (2008), except that chains were run for 5  107 post-burnin
generations and sampled every 103 generations. Because currently
Table 1
New hexactinellids sampled for this study. GW, collection of G.W.; HBOI, Harbor Bran
National Museum of Natural History, Smithsonian Institution; ZMA, Zo?logisch Museu
Family ID Collection region
Hyalonematidae Hyalonema sp.3a Bahamas
Hyalonema sp.4 Gulf of Mexico
Euplectellidae Docosaccus n. sp. California
Hertwigia sp. Florida
Saccocalyx sp.* Hawaii
Bolosoma n. sp. Hawaii
Rossellidae Rossella sp.*,b Antarctica
Bathydorus n. sp. California
Aphrocallistidae Aphrocallistes beatrix Florida
Euretidae n. gen. n. sp. Bahamas
Farreidae Farrea sp.c Florida
* Only sequenced for 16S and 30-half of 28S partition.
a Probably the same species as ??Hyalonema sp.2? used in Dohrmann et al. (2008),
b Originally identi?ed as Anoxycalyx joubini; re-investigation of this specimen rev
c Sequenced for all three partitions; the erroneous 18S sequence from Farrea occa
258 M. Dohrmann et al. /Molecular Phylogavailable Bayesian Markov Chain Monte Carlo (MCMC) implemen-
tations might not be able to accurately infer branch lengths under
certain conditions (e.g., Kolaczkowski and Thornton, 2007; Yang
and Rannala, 2005; Yang, 2007), we re-estimated branch lengths
of the consensus phylograms with the maximum likelihood (ML)
program optimizer of the PHASE package, using exactly the same
models as in the Bayesian analyses.
3. Results
Results of Bayes factor comparisons between 6- and 7-state
models were the same as in Dohrmann et al. (2008), i.e.,
6B > 6A >> 7D > 7A. We only present results obtained under 7D
and 6B (Fig. 1), because Markov chains employing 6A and 7A had
not converged for many parameters, making the results untrust-
worthy. Several considerations lead us to conclude that the phylog-
eny obtained under 7D likely provides the best estimate (Fig. 1A).
First, under 6B Aphrocallistes and Heterochone (Aphrocallistidae)
were not monophyletic; instead the two species with the longer
terminal branches (A. beatrix and Heterochone sp.) grouped to-
gether (Fig. 1B), potentially indicating long-branch attraction
(LBA; Felsenstein, 1978). In addition, the phylogeny obtained un-der 7D has generally more robust support values than those ob-
tained under 6B (Fig. 1). RNA16C performed worst of all ?ve
models (2lnBF = 3018 compared to 7A), and also did not recover
monophyly of the aphrocallistid genera (although the branching
order was different than under 6B; results not shown).
Overall, the phylogeny is congruent with previous results
(Dohrmann et al., 2008): both subclasses (Amphidiscophora,
Hexasterophora) and all families as well as most genera are
monophyletic, and all sceptrule-bearing taxa (Hexactinosida
part.) form a highly supported clade (Sceptrulophora; see Mehl,
1992) to the exclusion of Iphiteon panicea (Hexactinosida,
Dactylocalycidae). However, Lyssacinosida is recovered as mono-
phyletic under both 6B and 7D (Fig. 1) [and also 16C (not shown)],
due to a more stable position of I. panicea, which was nested with-
in Lyssacinosida in the previously reported phylogeny obtained
under 6B (see Dohrmann et al., 2008). Furthermore, the euplectel-
lid subfamily Euplectellinae is monophyletic, due to a change in
position of Walteria leuckarti (Corbitellinae) from within Euplec-
tellinae (Dohrmann et al., 2008) to sister of Hertwigia sp. + Sacco-
calyx sp. The grouping of the latter two species implies paraphyly
of the other two subfamilies, Corbitellinae (Hertwigia and Walte-
ria) and Bolosominae (Rhabdopectella, Bolosoma, and Saccocalyx).
The position of Farrea remains the same as in Dohrmann et al.
(2008), but the terminal branch leading to it is much shorter here.
ceanographic Institution; MBARI, Monterey Bay Aquarium Research Institute; USNM,
msterdam.
Voucher-No./source Acc.-Nos. (18S, 28S, 16S)
GW5454/Deep Scope 3 FM946129, FM946128, FM946109
USNM 1122180 FM946130, FM946131, FM946108
GW5429/MBARI FM946116, FM946115, FM946105
USNM 1122181 FM946121, FM946120, FM946104
USNM 1097540 ?, FM946119, FM946103
USNM 1097546 FM946118, FM946117, FM946102
ZMA POR16769 ?, FM946112, FM946107
GW5428/MBARI FM946114, FM946113, FM946106
USNM 1122182 FM946127, FM946126, FM946110
HBOI 16-XI-02-1-001/H.M. Reiswig FM946125, FM946124, FM946101
USNM 1122183 FM946123, FM946122, FM946111
sequenced for all three partitions.
d the presence of calycocomes, which are diagnostic/apomorphic for Rossella.
e Dohrmann et al., 2008; Voigt et al., 2008) was not included in this study.
ics and Evolution 52 (2009) 257?262The representative of the sceptrulophoran family Euretidae is
more closely related to Farrea + Aphrocallistidae than to Tre-
todictyidae. Monophyly of Hyalonema (Hyalonematidae), Bathydo-
rus (Rossellidae), and the Antarctic Rossella species (i.e., excluding
the North Atlantic R. nodastrella) is further corroborated, while
support for the inclusion of Clathrochone clathroclada (Lyssacinos-
ida incertae sedis) in the Leucopsacidae (Dohrmann et al., 2008)
substantially decreased. Under 16C, C. clathroclada was sister to
Leucopsacidae + Rossellidae, but support for the latter clade was
only 54% (not shown).
The branching order of the outgroup taxa (not shown) did not
change, but some support values differed (Table 2). Support for
monophyly of Porifera was 84% under 7D, compared to 72% ob-
tained with the previous taxon set. Under 6B, sponge monophyly
had 59% Bayesian support, not signi?cantly different from the
60% reported previously. Likewise, paraphyly of Demospongiae
sensu stricto with respect to Hexactinellida (see Dohrmann
et al., 2008) remained weakly supported under 6B (73%), but
had 92% support under 7D, compared to 79% obtained with
the previous taxon set. Under 16C, sponge monophyly and dem-
osponge paraphyly had 98% and 93% support, respectively (not
shown).
enetM. Dohrmann et al. /Molecular Phylog4. Discussion
4.1. Model comparison and support for outgroup nodes
With our extended dataset, comparison of likelihood-values
under the two paired-sites models RNA6B and RNA7D (see Savill
et al., 2001) again revealed the former to be higher. However, in
the phylogeny obtained under 6B, there is an apparent case of
Fig. 1. Ribosomal DNA phylogeny of Hexactinellida (new taxa in bold). Bayesian consen
(see Materials and methods for details). Outgroups not shown. Posterior probabilities onl
(B) Phylogeny obtained under RNA6B for stem regions. Models for loop regions and 16S w
gamma distribution (four categories) and a proportion of invariable sites (+I+G) indepe
Aphro, Aphrocallistidae; D, Dactylocalycidae; E, Euretidae; En, Euplectellinae; F, Far
Tretodictyidae. Scale bar, expected number of substitutions per site. The tree in (A), inc
number S2246.ics and Evolution 52 (2009) 257?262 259LBA regarding the genera Heterochone and Aphrocallistes, and
the phylogeny is less robust, in terms of clade support, than
the 7D phylogeny. Although this has to be further tested with
additional data, we therefore suspect that although being pre-
ferred by the Bayes factor, the 6B phylogeny is less accurate than
that obtained with 7D. This is presumably because mismatch
base pairings are treated as missing data in 6-state models, thus
potentially decreasing the amount of phylogenetic signal that
sus trees based on 50,000 post-burnin trees; branch lengths re-estimated with ML
y shown where <95%. (A) Phylogeny obtained under RNA7D applied to stem regions.
ere GTR (18S, 16S) and TrN93 (28S). Rate heterogeneity was modeled with a discrete
ndently for each partition. Further details can be found in Dohrmann et al. (2008).
reidae; Hyalo, Hyalonematidae; Leu, Leucopsacidae; Phero, Pheronematidae; T,
luding outgroups, has been deposited in TreeBase (www.treebase.org) under study
Fig 1. (continued)
Table 2
Comparison of Bayesian support for monophyly of Porifera and paraphyly of Demospongiae sensu stricto, respectively, obtained with the taxon set used in Dohrmann et al. (2008)
and that used in this study. ??Demosponge paraphyly? refers to a branching order where the clade (Spongilla lacustris(Mycale ?brexilis, Suberites ?cus)) is more closely related to
Hexactinellida than to the remaining included Demospongiae sensu stricto, (Dysidea, Aplysina ?stularis). With both taxon sets, Demospongiae sensu stricto and Hexactinellida form
a highly supported clade, whereas the Homoscleromorpha (traditionally assigned to Demospongiae) are the sister group of calcareous sponges (see Dohrmann et al., 2008 for
discussion).
Model for stem sites Posterior probability (%)
Sponge monophyly Demosponge paraphyly
Dohrmann et al. (2008) taxon set This study Dohrmann et al. (2008) taxon set This study
RNA6B 60 59 75 73
RNA7D 72a 84 79a 92
a Not reported in Dohrmann et al. (2008).
260 M. Dohrmann et al. /Molecular Phylogenetics and Evolution 52 (2009) 257?262
enetcan be extracted from the alignment. On the other hand, the
poor performance of the 16-state model indicates that account-
ing for all possible mismatches is super?uous in our case, and
a single mismatch state is suf?cient to describe our data well.
These interpretations are also in better agreement with the con-
clusions of Savill et al. (2001), who gave a general preference to
models 7A and 7D.
A ?nding that deserves further attention is the fact that in-
creased taxon sampling resulted in markedly greater support for
monophyly of Porifera and paraphyly of Demospongiae sensu stric-
to, but only in the 7D phylogeny (Table 2). Signi?cant support
(>95%) for sponge monophyly was only found with the 16C model
(not shown). However, this result has to be interpreted with cau-
tion because of the poor performance of this model. Likewise,
although receiving >90% Bayesian support under 7D and 16C with
the present data set, paraphyly of Demospongiae s.s. appears to be
an artefact of insuf?cient taxon sampling (Dohrmann et al.,
unpublished results). Thus, there appears to be a complex relation-
ship between model choice and taxon sampling in case of rDNA
data extracted from early-branching metazoans, suggesting that
the in?uence of both these factors and their interplay need to be
thoroughly addressed in studies aiming to resolve deep metazoan
relationships from this kind of data.
4.2. Monophyly of Lyssacinosida and hexasterophoran skeletal
evolution
In contrast to our previous study (Dohrmann et al., 2008), the
most species-rich hexactinellid order, Lyssacinosida, is now recov-
ered as monophyletic, independent of model choice, re-emphasiz-
ing the importance of taxon sampling for phylogenetic inference
(e.g., Zwickl and Hillis, 2002). Another factor that might have con-
tributed to this result is the replacement of the erroneous or
pseudogenic 18S sequence of Farrea occa (see Voigt et al., 2008)
with new sequences from Farrea sp. Exclusion of F. occa in our pre-
vious study (Dohrmann et al., 2008) resulted in substantially de-
creased support for the position of I. panicea as sister to
Euplectellidae. Thus, replacement of the F. occa sequence had an
overall bene?cial effect on tree reconstruction.
Lyssacinosida is a well-established taxon (Reiswig, 2002a), and
corroboration of its monophyly with molecular data further in-
creases the high congruence of molecular and morphological sys-
tems in Hexactinellida (compared to other sponge groups; see
Erpenbeck and W?rheide, 2007). It is noteworthy, however, that
no morphological autapomorphies are known for Lyssacinosida
(Mehl, 1992); the taxon is basically de?ned by a skeletal organiza-
tion of mainly unfused spicules (a plesiomorphy that also charac-
terizes Amphidiscophora), combined with hexasters as
microscleres (the de?ning autapomorphy of Hexasterophora)
(Mehl, 1992). In contrast, members of ??Hexactinosida? (here repre-
sented by Sceptrulophora and Iphiteon panicea) possess rigid skel-
etons composed of fused hexactine megascleres (dictyonal
frameworks) in addition to loose spiculation (see Leys et al.,
2007). The position of I. panicea as the sister taxon to Lyssacinosida
raises the possibility that dictyonal frameworks were inherited
from the last common ancestor of Hexasterophora and subse-
quently lost in the lineage leading to Lyssacinosida. Although the
occurrence of ??basidictyonal frameworks? as attachment struc-
tures in many lyssacinosidans (see Leys et al., 2007) might provide
some support for this idea, homology of these structures to true
dictyonal frameworks is questionable, and reasons for a reversal
to the ancestral type of skeletal organization are hard to imagine.
It is indispensable to determine the phylogenetic positions of addi-
M. Dohrmann et al. /Molecular Phylogtional non-sceptrulophoran dictyonal taxa such as Aulocalycoida,
Lychniscosida, and Dactylocalyx in order to reconstruct the evolu-
tion of hexasterophoran skeletal organization.4.3. Position of Farreidae and Euretidae
In contrast to all other sceptrulophoran taxa, whose sceptrules
are of the scopule-variant, Farreidae possess clavules (see Reiswig,
2002b). Thus, the nested position of Farrea suggests that the divi-
sion of Sceptrulophora into Scopularia and Clavularia (Schulze,
1886; Mehl, 1992) is arti?cial, because Scopularia is paraphyletic.
Our results imply that scopules belong to the groundplan of Sceptr-
ulophora and were replaced by clavules in Farreidae. The occur-
rence of ??intermediate? forms of sceptrules (aspidoscopules,
sarules) in Farreidae (see Reiswig, 2002b) lends some support to
this idea.
We also provide the ?rst estimate of the phylogenetic position of
Euretidae (Sceptrulophora) with molecular data. Its position (see
Fig. 1) does not con?ict with the Linnean system, simply because
no statements about interrelationships of the families of ??Hexacti-
nosida? are made in the current classi?cation (see Hooper and van
Soest, 2002, p. 1281 ff). Since Euretidae is themost species-rich fam-
ily of Sceptrulophora, and its taxonomic de?nition is weak, it might
be suspected that this taxon is paraphyletic. This hypothesiswas not
tested here and awaits sampling of additional species.
4.4. Phylogeny and evolution of Euplectellidae
In our previous study, we only included one species each of the
euplectellid subfamilies Bolosominae (Rhabdopectella tintinnus)
and Corbitellinae (Walteria leuckarti), and Euplectellinae was
recovered as paraphyletic (Dohrmann et al., 2008). Here, mono-
phyly of Euplectellinae is highly supported, whereas the other
two subfamilies appear to be non-monophyletic since Saccocalyx
sp. (Bolosominae) groups with Hertwigia sp. (Corbitellinae).
Although these two genera have extremely different body shapes,
their spicule compositions are very similar (see Tabachnick,
2002), so this grouping gains some support from morphology.
The topologies reconstructed here further imply that a) attachment
to substrate by means of spicule tufts (lophophytous mode), which
distinguishes Euplectellinae from the other subfamilies (Tabach-
nick, 2002), was derived from the basiphytous condition (attach-
ment via basidictyonal frameworks; see above) in Euplectellidae,
and b) tubular peduncles (stalks), the main character de?ning
Bolosominae (Tabachnick, 2002), evolved at least twice in this fam-
ily. The latter ?nding mirrors the situation in Rossellidae, where
the stalked genera Crateromorpha and Caulophacus are not closely
related (see Fig. 1 and Dohrmann et al., 2008). These results high-
light the susceptibility of poriferan body shape features to homo-
plasy. We suggest that a more stable system of Euplectellidae
(and also Rossellidae) requires thorough revision of morphology-
based taxonomy, combined with increased taxon sampling of
molecular markers.
4.5. Further results
In addition to the above ?ndings, this study further corrobo-
rates monophyly of Hyalonema, Bathydorus, and the Antarctic Ros-
sella species. Although monophyly of Aphrocallistes is also
recovered (in the 7D phylogeny), this remains poorly supported.
However, given that Aphrocallistidae contains only seven species,
complete sampling of this family?s diversity in order to robustly re-
solve its phylogeny appears to be an attainable goal.
Reasons for the decreased support for inclusion of Clathrochone
clathroclada in the Leucopsacidae (Dohrmann et al., 2008) are
somewhat unclear, but we suspect that the exclusion of additional
alignment positions (see Materials and methods) may be responsi-
ics and Evolution 52 (2009) 257?262 261ble. C. clathroclada is only represented by the 16S and one half of
the 28S partition, so con?dent placement of this species awaits col-
lection of additional sequence data.
Acknowledgments
We thank Amanda Kahn, Justin Marshall, Shirley Pomponi,
Henry Reiswig, and Rob van Soest for providing specimens, Henry
Reiswig and Konstantin Tabachnick for help with identi?cation,
and Burkhard Morgenstern for providing access to computational
resources of the Department of Bioinformatics, University of G?t-
tingen. Three anonymous reviewers are acknowledged for critical
comments that lead to improvement of the manuscript. This study
was funded by grants Wo896/5-1 & 2 of the Deutsche Forschungs-
gemeinschaft (DFG) to GW and bene?ted from a Mini-PEET award
of the Society of Systematic Biologists to MD.
References
Akaike, H., 1974. A new look at the statistical model identi?cations. IEEE Trans.
Automat. Contr. 19, 716?723.
Dohrmann, M., Janussen, D., Reitner, J., Collins, A.G., W?rheide, G., 2008. Phylogeny
and evolution of glass sponges (Porifera, Hexactinellida). Syst. Biol. 57, 388?405.
Erpenbeck, D., W?rheide, G., 2007. On the molecular phylogeny of sponges
(Porifera). Zootaxa 1668, 107?126.
Felsenstein, J., 1978. Cases in which parsimony or compatibility methods will be
positively misleading. Syst. Zool. 27, 401?410.
Hooper, J.N.A., van Soest, R.W.M., 2002. Systema Porifera. A Guide to the
Classi?cation of Sponges. Plenum, New York.
Kass, R.E., Raftery, A.E., 1995. Bayes factors. J. Am. Stat. Assoc. 90, 773?795.
Kolaczkowski, B., Thornton, J.W., 2007. Effects of branch length uncertainty on
Bayesian posterior probabilities for phylogenetic hypotheses. Mol. Biol. Evol. 24,
2108?2118.
Leys, S.P., Mackie, G.O., Reiswig, H.M., 2007. The biology of glass sponges. Adv. Mar.
Biol. 52, 1?145.
Lopes, D.A., Hajdu, E., Reiswig, H.M., 2007. Taxonomy of Euretidae (Porifera,
Hexactinellida, Hexactinosida) of Campos Basin, southwestern Atlantic, with a
description of a new species. Mar. Biol. Res. 3, 243?255.
Mehl, D., 1992. Die Entwicklung der Hexactinellida seit dem Mesozoikum.
Pal?obiologie, Phylogenie undEvolutions?kologie. Berl. Geowiss. Abh. E 2, 1?164.
Menshenina, L.L., Tabachnick, K.R., Janussen, D., 2007a. Revision of the subgenus
Neopsacas (Hexactinellida, Rossellidae, Crateromorpha) with the description of
new species and subspecies. Zootaxa 1436, 55?68.
Menshenina, L.L., Tabachnick, K.R., Lopes, D.A., Hajdu, E., 2007b. Revision of
Calycosoma Schulze, 1899 and ?nding of Lophocalyx Schulze, 1887 (six new
species) in the Atlantic Ocean (Hexactinellida, Rossellidae). In: Cust?dio,
M.R., L?bo-Hajdu, G., Hajdu, E., Muricy, G. (Eds.), Porifera Research:
Biodiversity, Innovation & Sustainability. Museu Nacional, Rio de Janeiro,
pp. 449?465.
Reiswig, H.M., 2002a. Order Lyssacinosida Zittel, 1877. In: Hooper, J.N.A., van Soest,
R.W.M. (Eds.), Systema Porifera. A Guide to the Classi?cation of Sponges.
Plenum, New York, p. 1387.
Reiswig, H.M., 2002b. Family Farreidae Gray, 1872. In: Hooper, J.N.A., van Soest,
R.W.M. (Eds.), Systema Porifera. A Guide to the Classi?cation of Sponges.
Plenum, New York, pp. 1332?1340.
Reiswig, H.M., Dohrmann, M., Pomponi, S.A., W?rheide, G., 2008. Two new
tretodictyids (Hexactinellida: Hexactinosida: Tretodictyidae) from the coasts
of North America. Zootaxa 1721, 53?64.
Savill, N.J., Hoyle, D.C., Higgs, P.G., 2001. RNA sequence evolution with secondary
structure constraints: comparison of substitution rate models using maximum-
likelihood methods. Genetics 157, 399?411.
Schulze, F.E., 1886. ?ber den Bau und das System der Hexactinelliden. Phys. Abh.
k?nigl. preu?. Akad. Wiss. Berlin 1, 1?97.
Tabachnick, K.R., 2002. Family Euplectellidae Gray, 1867. In: Hooper, J.N.A., van
Soest, R.W.M. (Eds.), Systema Porifera. A Guide to the Classi?cation of Sponges.
Plenum, New York, pp. 1388?1434.
Tabachnick, K.R., Collins, A.G., 2008. Glass sponges (Porifera, Hexactinellida) of the
northern Mid-Atlantic Ridge. Mar. Biol. Res. 4, 25?47.
Tabachnick, K.R., Janussen, D., Menshenina, L.L., 2008. New Australian
Hexactinellida (Porifera) with a revision of Euplectella aspergillum. Zootaxa
1866, 7?68.
Voigt, O., Erpenbeck, D., W?rheide, G., 2008. Molecular evolution of rDNA in early
diverging Metazoa: ?rst comparative analysis and phylogenetic application of
complete SSU rRNA secondary structures in Porifera. BMC Evol. Biol. 8, 69.
Yang, Z., 2007. Fair-balance paradox, star-tree paradox and Bayesian phylogenetics.
Mol. Biol. Evol. 24, 1639?1655.
Yang, Z., Rannala, B., 2005. Branch-length prior in?uences Bayesian posterior
probability of phylogeny. Syst. Biol. 54, 455?470.
Zwickl, D.J., Hillis, D.M., 2002. Increased taxon sampling greatly reduces
phylogenetic error. Syst. Biol. 51, 588?598.
262 M. Dohrmann et al. /Molecular Phylogenetics and Evolution 52 (2009) 257?262