BULLETIN OF MARINE SCIENCE, 35(!J: 38-S3, 1984 CORAL REEF PAPER 
A GENERIC REVISION OF THE STYLASTERIDAE 
(COELENTERATA: HYDROZOA) 
PART 2: PHYLOGENETIC ANALYSIS 
Stephen D. Cairns 
ABSTRACT 
A phylogcnelic analysis was performed on the 23 genera ofstylasterid corals. Hydractinia, 
a genus of athecate hydroid, was chosen as the out-group based primarily on morphological 
homology and secondarily on ontogeny, fossil record and advocacy. The evolutionary po- 
larities of the 19 characters used in the analysis were established by out-group comparison 
and transformation series of mullislate characters were ordered by apparent structural com- 
plexity and the process of reciprocal illumination. Several equally parsimonious cladograms 
are discussed and the justifications for choosing one in preference to the others are given. 
The interrelationships of the genera arc discussed; l^pidopora is considered to be the most 
plesiomorphic genus, Pseudocryptheiia the most apomorphic. The final cladogram is com- 
pared to the evolutionary tree proposed by Mosclcy (1881), Within the context of the final 
cladogram, the relative value of the characters and degree of homoplasy are discussed. The 
stylasterids are considered as a family oFathecate hydroids and the subfamilial designations 
are recommended to be abolished. 
Stylasterid corals arc fragile, usually small, uniplanar to slightly arborescent 
colonial hydrozoans of the phylum Coelenterata. Their calcium carbonate skel- 
etons are often brightly pigmented orange, red, blue or violet. The approximately 
185 known species (Cairns, 1983b) occur in all ocean basins from continental 
Antarctica to the Arctic Circle at depths between 0-2,800 m. They are most 
diverse and abundant at depths of 200-500 m. They arc known from the Paleocene 
to the Recent. Opinion is divided as to whether they should be considered a 
separate order in the Hydrozoa or simply a family of calcified hydroids in the 
Hydroida. This analysis is based on the redescription of the 23 stylasterid genera 
as revised by Cairns (1983b). 
Ideally, a phylogenetic analysis should be based on out-group comparison, 
supplemented by evidence derived from ontogeny (Stevens, 1980). Unfortunately, 
the ontogeny of stylasterids is virtually unknown and the out-group chosen for 
this analysis is a genus of uncalcified athecate hydroids. All characters used in the 
classification of stylasterids at all taxonomic levels are based on the calcium 
carbonate skeleton, which makes comparison to an uncalcified out-group difficult. 
Nonetheless, certain characters can be polarized from the out-group, and those 
that could not were ordered into transformation series by their apparent structural 
complexity and by the process of reciprocal illumination, a method of testing 
hypotheses of character state series against one another (discussed later). The 43 
taxa analyzed represent 23 presumably monophyletic genera (Cairns, 1983b). 
I Some generalized references on phylogenetic analysis, particularly on how to 
' determine polarity and order multistate characters are: Eldredge and Cracraft 
{ (1980), Watrous and Wheeler (1981) and Michcvich (1983). 
' This is the second application of phylogenetic systematic methods to a coelen- 
tcrate group. The first was by Schmidt (1972; 1974), concerning the ordinal clas- 
sification of the class Anthozoa. 
38 
CAIRNS: PHYLOOENETIC ANALYSIS OF THE STYLASTERIDAE 39 
METHODS 
choke of Out-Group. ?Hydro\ds of the genus Hydractinia were chosen as Ihe out-group for this 
analysis as they are hypothesized to be the sister group of the stylaslerids. This decision was based 
primarily on morphological homology, supported by ontogeny and advocacy, and was not contradicted 
by the fossil record. 
HoMO[,OGV. A decalcified stylasterid coral is indistinguishable from an athecate hydroid, a fact that 
no one has disputed since Moseley (1876) showed that stylastcrids were not scieractinian corals but, 
in fact, belonged to the Hydrozoa. Within the subclass Athccata sensu Petersen, 1979, stylasterids are 
most closely allied to the order Filifera, because they both have filiform, noncapilale, gastrozooid 
tentacles. Within the Filifera, stylastcrids are most similar to the Pandeida Petersen, 1979, one of 
three suborders in the Filifera. The Pandeida and stylasterids are characterized by having spindle- 
shaped gastrozooids with tentacles arranged in one whorl around a conical hypostome. Within the 
Pandeida, stylasterids are most similar to the Hydractinoidea Bouillon, 1978, one ofthree supcrfamilies 
in the Pandeida. The most important character in common at this level is the high degree of polyp 
polymorphism of the two tasa. Of the three or four families in the Hydractinoidea, stylasterids are 
most similar to the Hydractiniidae Agassiz, 1862. Hydractiniids have developed the potential for 
calcification, as evidenced by Janaria, Uydrocorelta and Poiykydra. which are the only hydroid genera 
to do so. For this reason, Stechow (1921) suggested that one of these genera may have been the 
evolutionary link between hydroids and stylasterids. Another character in common between the hy- 
dractiniids and stylasterids is their simple, noncapilate dactylozooids. Within the hydractiniids it is 
tempting to think, as did Stechow (1921), thai one of the three calcified genera is most closely related 
to the stylastcrids; however, detailed examination indicates otherwise. The coenosteal texture of the 
calcified hydroids is quite different from the reticulate-granular or linear-imbricate coenosteum of 
stylasterids. Furthermore, vesicles of unknown function are found in great abundance in Janaria and 
Hydroi'OteHa. Stechow (1921; 1962) identified these vesicles as gonophores but histological exami- 
nation reveals that they are not gonophores, gastrozooids or dactylozooids; there is no counterpart of 
this structure in any other hydroid or in the stylasterids. The stylasterids are, in fact, tnorc similar to 
species of Hydractinia. particularly because they both have spines and they both lack the medusoid 
stage. Stechow ( 1962: 418) suggested that the surface spines of Hydractinia were the predecessors, and 
thus homologs, ofthe stylasterid gastrostyle, achieved by deposition of calcium carbonate around the 
hydractiniid spine. Certain hydractiniid spines are very similar to stylasterid gaslrostylcs (Fig. 1) and 
thus fulfill one ofthe most important criteria for homology: similarity of positional hierarchy (Rieger 
and Tyler, 1979). Furthermore, these two structures contradict most of Rieger and Tyler's (1979) 
criteria for analogy, i.e., (1) they are not under the infiucncc of a common selective pressure, (2) they 
are composed of different materials (chitin vs. calcium carbonate), (3) they are not the only possible 
means to accomplish a particular function (the double-chambered gastropore chamber without a 
gastrostyle retains the gastrozooid as well as a gastropore without a style), 4) they both develop from 
ectoderm (Fritchmati, 1974) and 5) they are not under selective pressure to evolve mimicry. Therefore, 
I agree with Stechow (1962) that the Hydracttnta spine is homologous to the stylasterid gastrostyle. 
To summari/c, the dilTerenccs between stylasterids and Hydractinia are minor, mostly involving a 
constellation of changes associated with the deposition of a calcium carbonate skeleton, i.e., gonophores 
encapsulated as ampullae, gastro- and dactylozooids encased in calcified tubes and the transformation 
ofthe protective spines into a supportive gastrostyle. 
ONTOGENY. Very little is known about the ontogeny of stylasterid corals. In one ofthe few studies 
on stylasterid development, Fritchman (1974) noted a similarity ofthe gland cells ofthe planulae of 
the stylasterid Allopora petrograpta and the hydractiniid Hydractinia echinaia and stated that the 
method of skeleton formation of stylasterids and Hydractinia was so similar that it was undoubtedly 
an homologous structure, even though one is chitinous and the other is calcareous. 
FOSSIL RECORD. Very few hydroids are known from the fossil record but Hydractinia is one ofthe 
exceptions, known from the Eocene to Recent and questionably as far back as the Cretaceous (Hill 
and Wells, 1956). The earliest known stylasterids are from the Paleocene. This is certainly not proof 
of an evolutionary connection, but the hypothesis is not contradicted by the fossil evidence. 
Aovot ACY. Stylasterid corals customarily have fjeen placed in a separate order (Boschma, 1956), 
the Stylaslerina; however, as early as 1914. Broch considered them as a family of hydroids, closely 
allied to either Clathrozoon or Hydractinia. Stechow (1921; 1922; 1923; 1925; 1962) agreed with 
Broch that the stylasterids represented a family of hydroids closely related to the Hydractiniidae, 
especially the calcified hydractiniids. The stylastcrids were considered as one of four families in 
Bouillon's (1978) superfamily Hydractinoidea, one ofthe other families being the Hydractiniidae. 
Finally. Petersen (1979: 1 I 2). in his reorganization of the higher taxa of the Athccata. placed the 
Slylasteridae and Hydractiniidae as sister groups. I concur with these authors in considering the 
stylasterids to be a family of calcified hydroids within the superfamily Hydractinoidea. 
40 BULLETIN OF MARINE SCIENCE, VOL. 35. NO. I, 1984 
Figure I.    Scanning electron micrographs of the calcareous gastrostyle of Errina ?spera (left, 250 x) 
and the chitmous spine of Hydraainia echinaia (right, 125 >? ). 
Coding of Character States and Computer- Generated Cladograms. - N incleen characters were used in 
the phylogenetic analysis of the stylasterid genera. Most of these characters have more than two 
character states and one has as many as 10 character slates. Because the multistate characters arc not 
always interpreted as being linear in their evolution and because these data must be coded for the 
computer, often more than one data column was required lo code each character state (Appendices 
1 and 2). Ultimately, 44 columns were used to code the 19 characters. 
The characters used in the analysis were, for the most part, conservative at the generic level; however, 
sometimes species or groups of species within genera dilfcred in one or more character states. For 
instance, most species oiSienohelia have randomly arranged ampullae (Appendix 1 ; character 3: state 
A), but S. profunda has its ampullae concentrated around its gastropores (character 3: state B). To 
allow for an accurate coding of this genus, it was divided into two components: Stenohelia 1 and 
CAIRNS: PHYUMENETIC ANALYSIS OF THE STYLASTERIDAE 41 
Stenohelial. the former coded as having random?y arranged ampullae, the latteras having concenlraled 
ampullae. In theory, these two component laxa should reunite in the final cladogram as a monophyletic 
unit, as they did in this case. It should be stated that autapomorphies for genera that were subdivided 
were still considered as autapomorphies, not synapomorphies of the subdivided genera. It was necessary 
to use this technique for 8 of the 23 genera, some of which were divided into as many as six component 
taxa. A total of 20 additional taxa were added in this manner (Appendix 2). Not all of the component 
taxa regrouped into monophyletic units in the final cladogram, indicating that, based on these data, 
these genera are evidently not monophyletic. The implications will be discussed later. With a total of 
23 genera, 20 additional subdivided "genera," and the out-group, a total of 44 taxa were considered, 
producing a 44 x 44 data matrix (Appendix 2). 
In iwo cases, both concerning dactyloporc spine shape, all of the species of a genus had two character 
stales for the same character. For instance, trrinopsis always has both conical (coded: OlOOOO) and 
abcauline (coded: ??0010) dactylopore spines (character 19). Il was therefore coded as 010010. 
The cladograms discussed in Ihc remainder of the paper were produced by the Wagner 78 algorithm, 
which is discussed by Farris (1970) and Wiley (1981: 178-192). The program was installed on the 
Smithsonian's Honeywell computer by James S, Farris in 1979. The advantages of a Wagner analysis- 
tree stability, allowance for reversals, usage of all data and adherence to parsimony?are discussed by 
Michcvich (1978) and Farris ( in press). 
The first cladogram generated (not illustrated) was based on only those II of the 19 characters 
(Appendix 1: characters 1-10, 19) that could be polarized from out-group comparisons. As a simple 
example, the random arrangement of ampullae (character 3: stale A) is considered plesiomorphous 
because Hydractinia has randomly arranged gonophores; ampullae concentrated around gastropores 
is thus considered to be a derived stale (character 3: stale B). As a more complex example, the random 
coordination of gastro- and dactylopores (character 9: state A) is considered plesiomorphous because 
this is the condition found in Hydraclirtia. However, there are another eight character stales to which 
out-group comparison cannot be applied. In these cases, the character slates were cither coded in a 
very noncommittal manner, in which they were all independently derived from the plesiomorphous 
stale, or estimates were made as to their transformation series based on increasing morphological 
complexity. In this particular case, six of the nine states were provisionally linked to the ancestral 
stale but the OVz-opora-lype and cyclosystem arrangements (states I and J) were hypothesized to have 
derived from the Errinopora-type condition (stale H). This was based on the observation that some 
species of Errinopora have pseudocyclosystems very similar to those of the Stylaslerinae and some 
species have linearly arranged adjacent dactylopore spines very similar to those of Gyropora- Thus, 
the gastro-dactylopore coordination o? Errinopora was interpreted as a transition between those genera 
with randomly arranged daclylopores and those in which the dactylopores are coordinated into a 
cyclosystem. As another example, the dactylopore arrangement o? DisHchopora I (slate F) was hy- 
pothesized to be a less derived predecessor of the more highly coordinated pore row of Di.vtichopora 
2. Therefore, the character diagram illustrated in Appendix 1 (Fig. 4, drawing 9) was used for ihis 
character. These hypotheses of character slate order were considered provisional and subject to change 
if contradicted by a more parsimonious cladogram resulting from two or more other more reliable 
characters. This process of testing one hypothesis against other hypotheses of character state trans- 
formation series has been called reciprocal illumination (Hcnnig, 1966; Wiley, 1981) and will be 
discussed again later. 
The preliminary cladogram, based on these 11 polarized characters, was highly resolved in the upper 
levels but poorly resolved in the lower levels of the Wagner tree, with 20 of the 43 taxa originating 
directly or indirectly from one basal polycholomy. Therefore, the remaining eight characters were 
polarized and ordered based on the same principles described above, only this time the out-group was 
considered to be the 20 taxa in the basal polycholomy. A second, much more highly resolved cladogram 
resulted (not illustrated), which was not very different from the finally proposed cladogram. 
At this point m the analysis, the character state changes for each character were reanalyzed in relation 
to the branching pattern ofthe second cladogram in the process of reciprocal illumination. For instance, 
for character 17 (shape of gaslropore chamber) both the Fliohoihrus-lype (state C) and the cylindrical 
gaslropore (state B) were previously hypothesized to have originated from the ancestral condition 
(state A); however, cladogram 2 implied that it would be more parsimonious to derive the PHoholhrus- 
type from the cylindrical. Seven minor changes of this type were made in the character coding. A 
scries of computer runs was then made, each run differing in the order of taxa in the data matrix 
("shuffling the deck"). After eight runs a consistently most parsimonious tree was used for cladogram 
3 (Fig. 2). 
The changes made between cladogram 3 and the final cladogram 4. resulted from: I) a r??valuation 
of character 19: dactylopore shape, 2) two equally parsimonious alternatives for minor branches of 
the cladogram, 3) the addition of characteristics of coenosleal texture and 4) the addition of auta- 
pomorphous characters. 
Dactylopore shape was coded in a very generalized manner for cladogram 3 (Appendix I: character 
4Z BULLETIN OF MARINE SCIENCE, VOL- 35. NO, 1, ?984 
v 
o 
n 
M C? 
?n ^ 
q H 
;3 
r- C (U U 
c F u a 
E K 3 
Z o 
?. 1^ 
c 
CU 
J3 u 
CTl -C 
J5 
?T1 ? 
i o 
E? 
o   ? 
T3   C 
?a ?U 
- ?" 
1^^ 
u  C 
fr, 
^ ? 
3    O 
D. ^ 
B fti 0 o 
u x; 
-d .. 1? I? 
?S li 
?n iJ 
C    C 
ID   C 
?? F 
"'  r) 
NS ^ 
b, 3 
o   ?a T) 
?a T3 r. 
?   a; u 
05 a 
OJ <r 
rj   cz 
H ? ?o 
3  o 
n 
IJi S o 
CAIKNS: PHYLOGENETIC ANALYSIS OF THE STYLASTERSDAE 43 
! 9); 18 Steps, including 7 convergences and 5 reversals, were required to fit the character states to the 
taxa. If, however, the conical dactylopore is assumed to be a generalized structure that gave rise to 
the Hush, elliptical, abcaulinc, and cone of platelets dactylopore spines (Appendix 1: character 19'). 
then only 16 steps are required, including 7 convergences and only 2 reversals. Making this change 
produces a stem for the Lepidotheca-1nferioiahiaia branch andcollapses SieUapora into the polychoto- 
my with Sporadoporu and Disiichopora. 
A minor change, resulting in an equally parsimonious tree, was made by uniting Lepidopora 5 and 
Lepidopora 4 with the synapomorphy of character 9 (slate C: gastropores restricted to anterior face). 
This change created another convergence for character 13 (state C: pointed branch tips), but this was 
considered justified based on the high variability of the latter character and the stability of the former. 
Another equally parsimonious change united Oyropora and Errinopora by the synapomorphy of 
mullitipped gaslrostylc spines (character 1?: state C). This created a convergence for the presence of 
daclylostyles (character 2; state B), occurring once for Entimpora and again for the next segment of 
the cladogram. This change was fell to be justified because the daclylostyles of Errinopora arc much 
more robust than any of the other slylasterid dactylostyles and suggests a reinterpretation as a different 
kind of style. 
Coenosteal texture was recently introduced (Cairns, 1983a) as an easily distinguishable (with scanning 
electron microscopy) character that is usually conservative at the generic level. Initially. [ thought that 
it might serve as an important character in the phylogenetic analysis. Unfortunately, the character 
states of coenosteal texture are very unstable, occurring in parallel and reversing with great frequency, 
defying attempts to polarize or order the character states. Therefore, coenosteal texture (Appendix I: 
character 20) was not used to produce the computer-generated eladograms, but was added to the final 
cladogram in an unpolarizcd fashion to increase resolution. The five minor changes it made in the 
final cladogram were: 1) lo unite Lepidopora 2 and lepidopora 3, 2) to help produce a monophyletic 
group of Errinopsis 1 and 2, 3) to unite Errina 5 and Errina 6, 4) to unite StyiasWr 3 and Sty lasier 
4 and 5) to produce a monophyletic group of Stenohelia 1 and 2. 
Finally, the addition ofaulapomorphous characters helped to unite Errinopsis 1 and 2. Ordinarily 
the addition ofaulapomorphous characters does not change the branching of a cladogram, but because 
eight of the genera were subdivided, there was a potential for their reunion using aulapomorphous 
characters. 
RESULTS AND DISCTJSSION 
Discussion of the Cladogram.?The results of the phylogenetic analysis are sum- 
marized in cladogram 4 (Fig. 3). The genus with the least number of derived 
characters is Lepidopora, specifically the Lepidopora 1, 2 and 4 components. 
Although three other genera are linked to a common polychotomy, all of the 
Lepidopora components are at least two character state changes less derived, even 
Lepidopora 6. Lepidopora, unfortunately, does not resolve as a monophyletic unit, 
which is an indication that, pending further study, it should be divided into more 
than one genus or that more characters should be used in the analysis. Moseley 
(1881), without explanation, designated Sporadopora as the "ancestral" stylasterid 
genus. Broch {1914) suggested that Piiobothrus was the most "primitive" genus 
based on its lack of coordination of gastro- and dactylopores, simple dactylozooids 
and lack of gastrostyles; however, in 1942 (Broch, 1942: 7, 33) he vacillated 
between Sporadopora and Piiobothrus as the most primitive. The phylogenetic 
analysis places Piiobothrus near the root of the cladogram but, because the absence 
of gastrostyles is considered as a derived state, Lepidopora results in having the 
least number of derived characters. 
Three genera, Piiobothrus, Adetopora and Phalangopora. arc grouped by their 
lack of gastrostyles, and their stem is placed in the polychotomy with Lepidopora. 
Such lack of resolution in the lower levels of a cladogram is apparently not 
uncommon (pers. comm., V. Funk, 1982). It is interesting to note that Adelopora, 
having perhaps the most sophisticated adaptation of all the stylasterids?the hinged 
operculum ?is otherwise quite undcrived. 
Lepidopora 6 is distinguished from the first polychotomy by a relatively minor 
and variable character: sharply pointed gastrostyle spines. 
44 BULLETIN OF MARINE SCIENCE, VOL. 35. NO. 1. I9K4 
D||aL{id^j3QpfiaS(j 
oiiaMtd?J:^) 
"i DJodouoo 
I Djodouoo 
DDBMiUD|A4s 
C    jet?D|*iS 
Z    JB?30JA4S 
I    JB;SD|A[S 
OJOdojXc} 
ojodoujjj^ 
g    DUjji^ 
3   DUiJ^a 
CAIRNS: PHYLOGENETIC ANALYSIS OF THE STYLASTERiDAE 45 
The next branch of the cladogram, Lepidotheca and Inferiolahiata, is charac- 
terized by the transformation of the conical dactylopore to the abcauline dactylo- 
pore spine. The three components of Lepidotheca are adjacent (paraphyletic) but 
are not monophyletic on the cladogram. 
Paraerrina is distinguished by having both abcauline and flush dactylopores, 
both presumably derived from the conical dactylopore. The abcauline spines of 
Paraerrina are much smaller than those of Lepidotheca and found only on branch 
tips, therefore suggesting a reinterpretation as a different kind of dactyiopore spine 
instead of a convergence with those o? Lepidotheca. 
Ridged gaslrostyles are found in the remainder of the genera that have gas- 
trostyles. Stellapora has moderately ridged gastrostyles and very tall, clustered, 
composite, abcauline dactylopore spines, again presumably derived from the con- 
ical dactylopore; however, because these dactyiopore spines are so different from 
those of Lepidotheca or Paraerrina. they might also be reinterpreted as a different 
kind, not necessarily a parallelism with the other abcauline dactyiopore spines. 
Sporadopora and Distichopora seem to form a natural unit, united by the syn- 
apomorphies of very long, deeply ridged gastrostyles; internal ampullae (a case 
of convergence); and long dactyiopore tubes (a reversal). Distichopora. which 
resolves as a monophyletic unit, is distinguished from Sporadopora by its more 
highly coordinated gastro- and dactylopores and its elliptical dactylopores. Dis- 
tichopora 1 {=D. providentiae) forms an intermediate between Sporadopora and 
Distichopora 2, evidenced by its intermediate level of gastro- and dactyiopore 
coordination; D. (Haplomerismos) appears to be a highly derived offshoot from 
Distichopora 2. 
The remaining genera arc characterized by thick, adcauline or adcauline-like 
dactyiopore spines. The cladogram branch containing Errinopsis and Cheilopori- 
dion has individualized adcauline dactyiopore spines, and shares the synapo- 
morphy of branches that arc rectangular in cross section and fencstrate in ar- 
rangement. The two Errinopsis resolve as a monophyletic group distinct from 
Cheiloporidion. The latter genus is distinguished by an unusual modification of 
the conical dactyiopore. 
The six component laxa of Errina resolve as a monophyletic unit, united by 
the presence of both adcauline and flush dactylopores. The further resolution 
within Errina is based on characters subsequently interpreted as being highly 
variable. 
The remaining genera all have exclusively adnate dactylozooids, no conical 
dactyiopore, and a higher degree of gastro-dactylopore coordination, ranging from 
lines of adjacent dactylopores to cyclosystems. Errinopora and Gyropora are united 
by the synapomorphy of multihcaded gastrostyle spines. Gyropora is slightly more 
derived, having internal ampullae (a convergence) and common walls between 
adjacent dactyiopore spines. The dactyiopore spine walls of Errinopora are ad- 
jacent but discrete; Errinopora also has very well developed daclylostyles. 
The remaining genera, traditionally called the subfamily Stylasterinae, all have 
true cyclosystems. The four taxa immediately following Gyropora?\\\c non- 
monophylctic assemblage of Stylaster?havc dactylostyles and constricted gas- 
tropore chambers, each with a ring palisade. Stylaster 1 and 2, previously known 
as Ailopora, are differentiated from each other only on the basis of having sharp 
or blunt gastrostyle spines, a highly variable character. Stylaster 3 and 4 are 
differentiated from "Allopora" by having pointed branch tips, cyclosystems re- 
stricted in distribution, and a mixture of imbricate and granular cocnosteal texture. 
Stylaster 4 is distinguished by having coenostcal papillae and cyclosystems ar- 
ranged exclusively on the branch edges. 
46 BULLETIN Of" MARINE SCIENCE. VOL. 35. NO. I. 1984 
The clade consisting of Calyptopora and Stenohelia is distinguished from Sty- 
laster by the synapomorphy of unifacial cyclosystem orientation. Calyptopora is 
distinguished by its enlarged pseudosepta, which approximate lids, and the two 
Stenohelia are united by the synapomorphy of imbricate coenosteal texture. 
The remaining genera have their ampullae concentrated around their gastro- 
pores. The next branch, Stylantheca, is distinguished by a series of reversals, all 
related presumably to its reversion to the ancestral state of an encrusting habit. 
It also has the aulapomorphy of more than one gastrozooid per cyclosystem. 
The group consisting of the four remaining genera has the largest number of 
derived characters and is strongly differentiated by the loss of gastro- and dac- 
tyloslyles; the gain of large, round nematopores containing large nematocysts; the 
transition to a double-chambered gaslropore chamber; and a reversion to imbri- 
cate coenosteal texture. This represents a change equivalent in magnitude to that 
which occurred with the advent of cyclosystems. The two Conopora resolve as a 
paraphyletic (not monophyletic) group. Conopora 2 is distinguished from Co- 
nopora 1 by having cyclosystems arranged only on the edges of slender, pointed 
branches and by having internal ampullae. 
The last three genera all have unifacially oriented cyclosystems. Astya has the 
aulapomorphy of nematopores concentrated exclusively on the edges of the pseu- 
dosepta, and a prong jutting into the gastropore tube. Crypthelia and Pseudo- 
crypthelia share the synapomorphy of having both randomly arranged and con- 
centrated nematopores and a lid over each cyclosystem. Pscudocrypthelia is 
considered the most highly derived genus with its unusual gastrostyle and textured 
upper gastropore chamber. 
Moseley {1881; 98-101) is the only person to have proposed a phylogeny of 
the stylasterid genera; at that time there were only 12. He stated that the descent 
of the genera "from a parent form seems to be traceable with especial clearness." 
His tree included several hypothetical ancestors and the evolution of genera from 
other Recent genera. His approach was intuitive. Only that part of his tree dealing 
with the most advanced genera, those with cyclosystems, corresponds to my 
cladogram; the remainder is at variance with my results. 
The cladogram of Figure 3 does not allow for the monophyletic separation of 
the four traditionally recognized subfamilies of stylastcrids and I therefore suggest 
the abolishment of the subfamily level in the stylastcrids. 
Discussion of the Characters. ?One hundred fourteen character state changes were 
required to distribute the 19 characters within the tree in the most parsimonious 
manner. Of these I 14 changes, 14 are reversals and 49 are parallelisms or con- 
vergences. This relatively high rate of homoplasy, including 55% of the character 
slate changes, is apparently not uncommon (pcrs. comm., V. Funk, 1982), and 
suggests two related explanations; (1) stylastcrids were quite convergent in their 
evolution, developing similar structures many limes and even reversing the trend 
of evolution on occasion, or (2) the characters chosen to produce the cladogram 
are not conservative at the generic level; more characters should be analyzed and/ 
or the polarity of the original characters reevaluated. Both of these explanations 
are probably responsible, to varying degrees, for the high rate of homoplasy. 
However, once the 19 characters were chosen and polarized, none was dropped 
from the data matrix, regardless of its apparent homoplasy. This was done to 
avoid prejudicing the results by using only "good characters" subjectively chosen 
to support an a priori hypothesis. In theory, a well-corroborated cladogram would 
not be influenced by several highly homoplastic characters but would, in fact, 
serve to illustrate where these homoplasics occurred. With regard to the final 
CAIRNS: PHVLOGENETrC ANALYSIS OF THE STYLASTERIDAE 47 
1   3. 
B 
-      i \ 12,18   , 15      V      /^ 
T B^   ^D t X 
Ai A 
A 
? '  ^ D E    / I 10    C        ^E 
r A       Xj A 
>^' 
t 
A 
DBB 
A A A A 
"' D      E 
A A 
Figure 4. Character stale transformation series for the 19 characters used to construct cladograms 3 
and 4. Numbers and letters correspond lo the characters and character stales, respectively, as listed 
in Appendix I. Drawing 11' is an equally parsimonious interpretation derived from cladogram 4. 
Drawing 19 was used for cladogram 3; 19' for cladogram 4. 
cladogram, two characters were particularly homoplastic: prominence of ampullae 
(characler 7; 7 convergences, ] reversal; CI = 0.22) and condition of the branch 
tips (character 13: 6 convergences, 1 reversal; CI = 0.37). [Consistency indices of 
characters, CI, are defined by Farris (1969).] Not surprisingly, these are the two 
characters most often used to divide genera into smaller units to facilitate coding, 
which is an indication that they are probably not conservative at the generic level. 
On the other hand, character 9 (coordination of gastro- and dactylopores) has a 
high consistency index of 0.9 and thus yielded much information for its construc- 
tion and interpretation. Other highly consistent characters were: 2, dactylostyle 
type (CI = 0.75); 4, presence of gastrozooid tentacles (CI = 1.0); 8, position of 
dactylopore spines (CI = 0.5); 12, branch anastomosis (CI = 0.67); 15, presence 
of gastrostyle ridges (CI = 1.0) and 17, gastropore chamber shape (CI = 0.83). 
Characters 8, 9 and 19 had some degree of overlap. 
48 BULLETIN OF MARINE SCIENCE, VOL. 35, NO, I. ilM 
It is interesting to note that an alternative way of coding character 11 (Appendix 
1, Fig. 4: 11'), implying the evolution of the fixed cyclosystem lid from the prong 
of Astya, produces the same cladogram in an equally parsimonious manner. Mo- 
seley (1881: 101) vacillated on the interpretation of this interrelationship but 
eventually drew his tree to reflect this alternative. 
Other Observations.?The fossil record of stylasterids is not well known despite 
the fact that 28 of the 231 nominal species are known exclusively as fossils; most 
of these are from the Paleocene of Denmark (Nielsen, 1919) and the Eocene of 
Eua, Tonga (Wells, 1977). Also, most of the fossils arc not well preserved and are 
of dubious generic identity (Cairns, 1983b). One fossil genus, Congregopora, 
containing only one known species from the Paleocene, is not included in this 
analysis because of the lack of diagnosable characters. Speculations concerning 
the evolutionary position of this genus and Axopora will be made at a later time. 
When the poorly known geological ranges are superimposed on the generic clado- 
gram, only a very generalized picture emerges. One of the most derived genera, 
Cryplhetia, was present in the Eocene, and the least derived genus, Lepidopora. 
was only questionably present in the Paleocene (Cairns, 1983b). The implication 
is that many, if not all, of the genera evolved in a rapid radiation in the late 
Paleocene or early Eocene, shortly after diverging from the hydractiniid hydroids. 
ACKNOWLEDGMENTS 
It is wilh pleasure that I acknowledge Ihc interest and guidance received from V, Funk (Smithsonian 
Institution) regarding the practical aspects ofcladism and for her advice at every step of the process 
resulting in the cladogram. I also thank D. Brooks (University of British Columbia) for introducing 
me to the theory of phylogenetic analysis. I am grateful to R, Vari (Smithsonian Institution) and J. 
Russe (National Manne Fisheries Service) for their advice on procedural matters regarding the coding 
ofcharactcrs and running the computer program. H, Zibrowius (Station Marine d'Endoume. Marseille) 
and N. Platnick (American Museum of Natural History) read the manuscript and oflFered suggestions. 
This work was supported by NSF grant DEB-8102776, 
LITERATURE CITED 
Boschma, H.   1956.   Milieponna and Stylasterina, Pages F90-F106, figs, 75-85 in R. C. Moore, cd. 
Treatise on invertebrate paleontology. Part F, Coclenterata. University ofKansas Press, Lawrence, 
Kansas. 
Bouillon, ,1.   1978,  Sur un nouveau genre et une nouvelle esp?ce de Ptilocodiidae Hydrichthelloides 
reticulata et la super-famille des Hydractinoidea (Hydroida-Athecata). Steenstrupia 5(6): 53-67, 
4 figs. 
Broch, H.   1914.   Stylasteridae. Dan. Ingolf-Exped, 5(5): 1-28, 5 pis., 6 figs. 
 .   1942.   Investigations on Stylasteridae (Hydrocorals). Skr. Norske Vidensk.-Akad., I. Mat.- 
Naturv. KJasse 3: 1-113, 6 pis., 38 figs. 
Caims, S. D.   1983a.  Antarctic and Subantarctic Stylasterina. Antarctic Res. Ser, 38: 61-164, 50 pis. 
 .   1983b.   A generic revision of the Stylasterina (Coelenterata: Hydrozoa). Part 1. Description 
of the genera. Bull, Mar. Sei. 33: 427-508, 28 figs. 
Eldredge, N. and J. Cracraft.   1980.   Phylogenetic patterns and the evolutionary process. Columbia 
University Press, New York. 349 pp. 
Farris, J. S.   1969.   A successive approximations approach to character weighting. Sysl. Zool. 18(4): 
374-385. 
 .   1970.   Methods for computing Wagner trees. Syst. Zool. 19(1): 83-92. 
 .   In Press.  The logical basis of phylogenetic analysis. Pages 7-36 in N. I. Platnick and V. A. 
Funk, eds. Advances in cladistics 11. Columbia University Press, New York. 
Fritchman, H. K..   1974.   The planula of the stylasterinc hydrocoral Altopora pelrograpia Fisher: Its 
structure, metamorphosis and development of the primary cyclosystem. Proc, Second Int. Coral 
Reef Symp. 2: 245-258, 27 figs. 
Hennig, W,   1966,   Phylogenetic systematics. University of Illinois Press, Urbana. 263 pp. 
Hill, D. and J. W. Wells.   1956.   Hydroida and Spongiomorphida. Pages F81-F89, figs. 65-74 in R. 
CAIRNS; PHYLOGENETIC ANALYSIS OF THE STYLASTERIDAE 49 
C. Moore, cd. Treatise t)n invertebrate paleontology. Part F. Coclenterata. University of Kansas 
Press, Lawrence, Kansas. 
Michevich, M. F.   1978.   Taxonomic congruence. Sysl. Zool. 27(2): 143-158. 
 .   1983.   Transformation series analysis. Syst, Zool. 31(4): 461^78. 
Moscley, H. N.   1876.   Preliminary note on the structure of the Stylasteridae, a group of stony corals 
which, like the Milleporidae, are Hydroids, and not Anthozoans. Proc. R. Soc. Lond. 25: 93- 
101. 
 .   1881.   Report on certain Hydroid, Alcyonarian and Madreporarian corals procured during 
the voyage ofH. M. S. Challenger in the years 1873-1876. Part 1. On the HydrocoraUinae. Rep. 
Scient. Res. Voyage Challenger, Zool. 2: 1-101, 209-230, pis. 1-14. 
Nielsen, K. B.   1919.   En Hydrocoralfauna fra Faxe. Danm. geol. Unders. (4)1(10): 1-66, 2 p!s., 9 
ligs. 
Petersen, K. W.    1979.   Development of coloniahly in Hydrozoa. Pages 105-139, figs. 1-12 in G. 
Larwood and B. R. Rosen, eds. Biology and systematics of colonial organisms. Academic Press, 
London. 
Ricgcr, R. and S. Tyler.   1979.   The homology theorem in uttrastructural research. Am. Zool. 19: 
655-~6fi4, 4 figs. 
Schmidt. H.    1972.   Die Nesselkapseln der Anlhozoa und ihre Bedeutung f?r die phylogenetische 
Systematik. Helgolander Wiss, Meersunters. 23: 422-458. 
 .   1974.   On the evolution of the Anthozoa. Proc. Second Int. Coral Reef Symp. 1: 533-560, 
16 figs. 
Stechow, E.    1921.   Neue Gruppen skelettbildender Hydrozoen und Verwandtschaftsbeziehungen 
rezenter und fossiler Formen, Verh. dt. Zool. Ges. 26: 29?31. 
 .    1922.   Zur Systematik der Hydrozoen, Stromatoporen, Siphonophoren, Anthozoen und 
Ctenophoreti. Arch, Naturgesch. 88(A)3: 141-155. 
 .   1923.   ?ber Hydroiden der Deutschen Tiefsce-Expedition, nebst Bemerkungen ?ber einige 
andre Formen. Zool. An/. 56(5-6): 97-119. 
 .   1925,  DieHydroidenderDeulschen Tiefsee-Expediti on ( Valdi vi a), Wiss. Ergebn, dt. Tie fsee- 
Exped, "Valdivia" 17(3): 383-546, 54 figs. 
1962  (posthumous). ?ber skclettb?dende Hydrozoen. Zool. Anz. 169 (9-10): 416-428, 7 
figs, 
Stevens, P. F.   1980,   Evolutionary polarity of character stales. Ann. Rev. Ecol. Syst. I I: 333-358. 
Watrous, L. E. and Q. D. Wheeler.   1981.  The out-group comparison method of character analysis. 
Syst. Zool. 30(1): 1-11. 
Wells, J. W.    1977,   Eocene corals from Eua, Tonga. Prof. Pap. U.S. Geol. Surv. 640-G: 1-13, 17, 
18, pis. 1-3. 
Wiley, E. O.    1981.    Phylogcnetics: The theory and practice of phylogenetic systematics. Wiley- 
Interscienee, New York. 439 pp. 
DATE ACCEPTED:    May 17, 1983. 
ADDRESS:    Deparlmem of ?nyerlebrale Zoology. Smithsonian Instilulion, Washington, D.C. 20560. 
APPENDIX 1 :    CODING OF CHARACTER STATES 
Character!: Shape of Colony 
DATA COLUMN 
1 
A Encrusting 
B Branching 
Character 2: Dactylosivlcs 
0 
1 
2 
A Absent 
B Type 1 (one row of slender elements per dactylopore) 
C Loss of Type 1 
D Type 2 (several rows of thick elements 
per dactylopore) 
Character 3:  Location of Ampullae 
0 
1 
2 
-1 
3 
A Randomly arranged on branch 
B Concentrated around gastropore 
Character 4: Gastrozooid Tentacles 
0 
1 
4 
A Present 
B Absent 
0 
1 
50 BULLETIN OF MARINE SCIENCE, VOL. 35, NO. 1, 1984 
APPENDIX 1:    CONTINUED 
Character 5: Daciylozooid Tentacles 
A Simple 
B Simple and adnate 
C Exclusively adnate 
Character 6; Nematopores 
A None 
B Papillae 
C Round pores, randomly arranged 
D Round pores, randomly arranged and concentrated 
around gastropore 
E Round pores, concentrated around gastropores 
Character 7:  Prominence of Ampullae 
A No skeletal evidence 
B Superficial 
C Internal 
Character 8:  Position of Dactylopore Spines 
A Lacking or widely spaced 
B Clustered 
C Adjacent, arranged in rows; separate walls 
D Adjacent, arranged in rows; common walls 
E Adjacent, arranged in cyclosystems 
Character 9: Coordination of Castro- and Dactylopores 
A Random 
B Gastropores at branch axils 
C Gastropores restricted to anterior face 
D Gastropores on both faces 
E Gastroporcs restricted to branch edges 
F Rudimentary pore rows 
G Pore rows 
H Dactylopores arranged in discontinuous lines adjacent 
to gastropores; pseudocyclosystems present 
I   Dactylopores arranged in lines; daclylopores have 
common walls; pseudocyclosystems present 
J  Cyclosystcm arrangement 
Character 10: Spination of Gastrostyles 
A Blunt 
B Loss of blunt 
C Muliihcaded 
D Sharp 
E Loss of sharp 
F Rudimentary {Pseudocrypihelia) 
Character 11 : Covering of Gastropore 
A None 
B Enlarged pseudosepta 
C Fixed Ud 
D Prong 
E  Hinged opcrculum 
Character 12: Branch Anastomosis 
A Encrusting, no branches 
B Branches free or slightly anastomolic 
C Branches regularly fenestrate 
Character 13;  Branch Tips 
A Encrusting, no branches 
B Blunt 
C Pointed, slender 
D Lobate 
5 
- 
0 
I 
2 
6 7 
- 
0 0 
1 0 
-1 0 
-1 
i 
1 
-1 
8 
0 
1 
2 
9 10 
- 
0 0 
1 0 
2 0 
2 1 
2 -1 
11 12 13 14 15 16 17 
0 0 0 0 0 0 0 
1 0 0 0 0 0 0 
0 1 0 0 0 0 0 
0 0 I 0 0 0 0 
0 0 0 1 0 0 0 
0 0 0 0 1 0 0 
0 0 0 0 2 0 0 
0 0 0 0 0 1 0 
0 0 0 0 0 1 1 
0 0 0 0 0 1 -1 
18 19 20 
0 0 0 
1 0 0 
0 ] 0 
0 0 1 
0 0 2 
0 0 3 
21 22 23 24 
- 
0 0 0 0 
1 0 0 0 
0 1 0 ? 
0 0 1 0 
0 0 0 1 
25 
0 
1 
2 
26 27 
0 
1 
1 
0 
0 
1 1 
1 
1 
-1 
CAIRNS; PHYLOGENETIC ANALYSIS OF "1"1!E STYLASTERIDAE 51 
APPENDIX 1:    CONTINUED 
Character 14: Orientation of Cyclosystcms 
A No cyclosystems 
B Random 
C Primarily on branch edges bul some on faces 
D Exclusively on branch edges 
E  L'nifacial 
Character 15; Ridges of Gastrostyle 
A No ridges on style 
B Loss of non ridged style 
C Moderately ridged 
D Deeply ridged 
E Loss of ridged style 
Character 16:  Branch Cross Section 
A Encrusting, no branches 
B Round to slightly elliptical 
C Rectangular 
D Lamellar 
Character 17: Shape of Gastroporc Chamber 
A No chamber 
B Cylindrical 
C Unique U'liobothrus) 
D Constricted 
E Constricted, with ring palisade 
F Double chamber 
Character I 8:  Length of Daclyloporc Tubes 
A None 
B Long, extending, down branch axis 
C Short, terminating within 2 mm 
Character 19: Shape of Dactylopore 
A None 
B Flush 
C Conical 
D Cone of platelets 
E Elliptical 
F Abcauline 
G Adcaulinc 
H Adcautine-typc, linearly arranged 
I   Adcauline-type, arranged in cyclosystems 
Character 19': Shape of Dactylopore (alternate) 
A None 
B Conical 
C Elliptical 
D Cone of platelets 
E Abcauline 
F  Rush 
G Adcaulinc 
H Adcauline-like, linearly arranged 
1   Adcauline-like, arranged in cyclosystems 
Character 20: Coen osteal Texture 
Linear-imbricate 
Reticulate-granular 
Both linear-imbricate and reticulate-granular 
Unique, each case being a dilferent texture 
28    29 
0 0 
1 0 
2 0 
2 1 
2 -1 
30 31 
0 0 
-1 
1 
1 
0 
0 
1 
1 -1 
32 33 
0 
I 
1 
0 
0 
1 
1 
34 
-1 
35 36 37 
- 
0 0 0 0 
1 
1 
0 
1 
0 
0 
0 
0 
1 0 1 0 
1 0 0 1 
1 0 0 2 
38 
0 
1 
2 
39 40 41 42 43 44 
0 0 0 0 0 0 
1 0 0 0 0 0 
0 1 0 0 0 0 
0 0 1 0 0 0 
0 0 0 1 0 0 
0 0 0 0 1 0 
0 0 ? 0 ? 1 
0 0 0 0 0 2 
0 0 ? 0 0 3 
39 40 41 42 43 44 
0 0 0 0 0 0 
1 0 0 0 0 0 
1 1 0 0 0 0 
1 0 1 0 0 0 
1 0 0 1 0 0 
1 0 0 0 1 0 
0 0 0 0 0 1 
0 0 0 0 0 2 
0 0 0 0 0 3 
45 
I 
Q 
I/O 
U 
52 BULLETIN OF MARINE SCIENCE. VOL. 35, NO, I, I9H4 
APPENDIX 2: DATA MATRIX FOR CLADOGRAMS 3 AND 4 
(Figures 2 and 3) 
;:3:=?tjOt-'!^i-?i-i?-iD 
CJOOOOOOOOC 
OOOOOOOO?'O 
OOOOOOOCDOO 
|_(  ^   M   I-?  o   (J   o 
o o a o o o o 
^    _H    _l     ^    ^    .-1    o 
o o o o o o o 
o o o C? o o o 
o o o o o ?I a 
o o o o ?' o -^ 
cN rj rj r4 H fM --< 
o a o o o o o 
o  o  o  o  "^ C3  o 
o o C3 o o o o 
o o c 
o o : 
o o 
a o 
o o 
,--iC3?'?  -JT^.-<.?if^c^fonnrO 
?ICJ   OOOOOOC?OOOOOOO 
?-^lM^^^c-^lt^l?^J?C^l^J?M^^l^^lolrJc^ltMtv|t 
OCOOOOOOOUOCJ?J^-'-H'-I' 
- ci l-i 
o o 
o o 
o o 
o o 
(-1 u 
o o 
o o 
o o 
o o 
o o 
o o 
o o 
ro r^ m  m rO 
? o o o o 
o o o a o 
o o o   o C3 
o o o o o 
o o o o o 
fvj oj rM es iN 
rsl ?-J (>1  CM f^l 
o o o o o 
o o o a o 
OOOOO0OOC3OOOOO0OOOO' SOOOOOQOOOOOaOOOOC?OOO 
oaaoooooaoooo?oo- lOOOOOOOOOOOOCOOOOOO' 
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOCh-^^?M?^?N?^oJ^-HCNC^nli-^J 
o o o o 
o o o o 
o o o o 
o o o u 
o o o o 
o o o o 
o o o o 
o o o o 
o o o o 
o o o o 
o o o Ci 
C! o o o 
o o o o 
o o o o 
o o o o 
C3   o   CJ o 
o o o ? 
o o o o 
o o o ? 
o o ? o 
o o o o 
o o o o 
o    o    -H    r-H 
a o o o 
a o o o 
o o o o 
o o o o 
o o o o 
^ ^ o o 
o  o o  '-t 
o  C3 o  o 
o o o o 
o o 
o o 
o  o 
o o 
o 1-^ 
o o 
.-( o 
Ci o 
o a 
o o 
a o 
o o 
o o 
o o o 
c: o C3 
o o o 
o o o 
r-l   o   C3 
o  Q  o 
o o o 
o o o 
o o o 
o o o 
o o o 
o C3 o 
o o o 
o ?I o 
o o o 
o o a 
o C3 o o o o o 
o o o o o o ? 
o  o  o o  o  o ? 
o o ci> o ? o o 
o   i-H   ?I   ?   ?'   -^ ?' 
o o o o o o o 
a c o o o o c:3 
? o C3 o o o o 
o o o o o o o 
o  o  o  o  ?' es (N 
o o o o o Ci o 
o o o o o o o 
o o  o  o  o  o C! 
o o o o o o o 
o o o o o o o 
o o -? o t^ ?^ ^ 
INI   ?-J   CM 
o o o 
o o a 
o o o 
o o o 
o o o 
C!   o   Cl 
o o o 
a o o 
o o o 
o o a 
o o i-i 
o o 0 
o o o 
o o o 
o o o 
o o o 
Q  o 
o o 
ct o 
o o 
o o 
o o 
o o 
o o 
o o 
a o 
o o 
o o 
o o 
o o o 
o o o 
o o o 
o o o 
o o o 
o o a 
o o o 
o o  o 
o o o 
o o o 
o o o 
o o o 
o o o 
o o o 
o o o 
o o o o o 
o o o o o 
o o o C] o 
o o o o a 
o o o o ?I 
o -* --I o o 
a o o o o 
Ci o  ?  --1 --I 
c)?3000Dao'-' oo 
OOOOOOOC?O   ?i'-i 
oo?'oooooo oo 
OOOOOOOOO    ?O 
OO0OOOOC?CZ5OO 
I    I    I I I I 
o ? C] o o 
o o  D ?  o 
o o o o o 
o o a o o 
o o tp o o 
o o 
I 
t m ? 
OOOOOO ac?o oo 
ooooooooo oo 
CJOOOOOOOO oo 
OOOOOOOO? o o 
OOCCOOOOOOQ 
7 1" 1   I   I   I   I   I   I   I   ' 
0000000000000-1000000000000000000000000000^-' 
oooooooooooooooo ? ooooooooooooooooo^^^^ ? ? ^-'-' 
OOOOOOOO?'OOOOOrvj?I  ?  oir-i?vi_?rO--!"--''^?"--^rjrj 
oooooooooooooooooooooooaooooooooooooooo 
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO^?I 
ooo?uuooooaoo^^?"ooooooooooooooo-^o--^i-^?1-^^^--''-''?' 
I     I     I     1     I 
I 
a r^ rn 
nJ [rt itV 
r?   ? ri n <n rr 
1. 
1^ r^ ^ r, i-i n r; r: u 
.p ?U 4-i 
-VI n r> -n -rt Ti 
?H ?H TH OJ fO 
n n n n M 1> n 
o J ^ J hJ J d. < P^ f-i .-i i-i UJ 
r. 
<)i ?H 
H !-l 
n n n. 
1?1 
)-i M 1* 
I-I n 
1^ W u 
_, m ? 
m in fj 
?H ?ri ?U 
-fl (]} QJ     di fli !% 
+J !1 lU Hi 
n r: r> fi 
n w u. 
n-l >-^ a 
u ?=- !^ ?>-, ^ :?) 11 >- r 
l-l 1- > .1  4J 
^ UJ M W o trt t/: UJ IH UJ UJ UJ 
CAIRNS: PHYUX?ENETIC ANALYSIS OF THE STYLASTERIDAE 53 
APPENDIX 2:    CONTINUED 
The character stales of character 20 (coenosteal texture) were not ordered and therefore were not 
used to produce the computer-generated cladograms. See Appendix 1 for a key to the characters and 
how they were coded. Eight genera were subdivided as follows: 
Lepidopora 1: Species with randomly arranged dactylopores, blunt branch tips, and linear-granular 
coenosteal texture: L. diffusa. I., granulosa. 
Lepidopora 2: One speeies with randomly arranged daclyloporcs, blunt branch tips, and reticulate- 
granular coenosteal texture: L. decipiens. 
Lepidopora 3: Species with randomly arranged daclylopores, slender branch tips, and reticulate-gran- 
ular coenosteal texture: L. carinara. L. sarmentosa. 
Lepidopora 4: One species with dactylopores restricted to lateral edges of branch, blunt branch tips, 
and linear-imbncate coenosteal texture; /,. eb?rnea (Calvet, 1903) (^?. hicksoni Boschma, 1963). 
Lepidopora 5: One species with dactylopores restricted to lateral edges of branches, slender branch 
tips, and a unique coenosteal texture: L. glabra. 
Lepidopora 6: One species with randomly arranged dactylopores, slender branch tips, and a unique 
coenosteal texture: L. acrolophos. 
Lepidotheca 1: Species with blunt branch tips, sharp gastrostyle spines, and without dactylostyles: L. 
cervicornis. L. hachijoensis. L. jap?nica. 
Lepidolheca 2: Species with slender branch tips, blunt gastroslylc spines, and without daciylostyles: 
L. ramosa, L. fascicularis, L. h?rrida. 
Lepidolheca 3: One species with slender branch tips, blunt gastrostyle spines, and dactylostyies: L. 
lenuistylus. 
Disiichopora 1: One species with rudimentary pore rows: D. providenttae. 
Distichopora 2: All other species o? Disiichopora. all having well-developed pore rows. 
Errinopsis I : One species with a cylindrical gaslropore chamber: E. reticulum. 
Errinopsis 2: One species with a constricted gastroporc chamber: E. feneslrata. 
Errina 1: Species with reticulate-granular coenosteal texture, superficial ampullae, and blunt branch 
tips: E. ant?rctica. E. cruenta. E. ?spera, E. capensis. 
Errina 2: One species with reticutatc-granular coenosteal texture, internal ampullae, and blunt branch 
tips: E. kergueiensis. 
Errina 3: Species with reticulate-granular coenosteal texture, superficial ampullae, and slender branch 
tips: E. gractlis, E. cheilopora, E. novaezeatandiae. E. rubra, E. dabneyi. E. atl?ntica, E. cochieata. 
Errina 4: One species with linear-imbricate coenosteal texture, superficial ampullae, and slender branch 
tips: E. macrogastra. 
Errina 5: Species with both reticulate-granular and imbricate coenosteal texture (the latter only on 
the dactylopore spines), superficial ampullae, and slender branch tips: E. fissuraia. E. boschmai. 
Errina 6: One species with both reticulate-granular and imbricate coenosteal texture (the latter only 
on the dactylopore spines), internal ampullae, and slender branch tips: E. laterorifa. 
Stylaster 1: One species in Stylaster (Group A) sensu Cairns, 1983b, with blunt gastrostyle spines: S. 
nomegicus. 
Siylaster 2: The remaining species in Siylasler (Group A): about 21 species. 
Stylaster 3: .Siylaster (Group B) sensu Cairns, 1983b: 16 species. 
Siylaster 4: Stytaster (Group C) sensu Caims, 1983b: 27 species. 
Srenohelia I: Species with randomly distributed ampullae: all species except for 5". profunda. 
Stenohelia 2: One species having ampullae clustered around gastropores: S. profunda. 
Conopora 1: Conopora (Group B) sensu Caims, 1983b: two species. 
Conopora 2: Conopora (Group A) sensu Caims, 1983b: three species.