Reconstructing the evolution of morphogenesis and dispersal among velatid asteroids

MISSING IMAGE

Material Information

Title:
Reconstructing the evolution of morphogenesis and dispersal among velatid asteroids
Physical Description:
xi, 123 leaves : ill. ; 29 cm.
Language:
English
Creator:
Janies, Daniel Andrew, 1966-
Publication Date:

Subjects

Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 107-121).
Statement of Responsibility:
by Daniel Andrew Janies.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002070879
notis - AKQ9161
oclc - 34404534
System ID:
AA00003604:00001

Full Text









RECONSTRUCTING THE EVOLUTION OF MORPHOGENESIS AND
DISPERSAL AMONG VELATID ASTEROIDS












By


DANIEL ANDREW JANIES


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA

1995


UNIVERSITY OF FLORIDA LIBRARIES

























Dedicated to Andrew and Mary Ellen Janies













ACKNOWLEDGMENTS

I thank the public schools that have been my intellectual and physical

homes for over the past decade: the Universities of Michigan, Washington,

and Florida. Accordingly, I thank my parents Mary Ellen and Andrew Janies

for their support and for giving me the values to strive for an education. I

thank Rebecca Shammami and Cyril for love and companionship. I thank

my advisor, Dr. Larry McEdward, for sharing his good humor, knowledge,

experience, and mostly for believing in me. I also thank Dr. Greg Wray for

support and helping me to see a future in science. I thank my committee for

the hard work they devoted towards my degree. I thank the agencies that

were very generous in their financial support of my studies: Lerner-Gray

Fund for Marine Science of the American Museum of Natural History,

American Scandinavian Foundation, American Society of Zoologists,

Aylesworth Foundation for the Advancement of Marine Science, Canadian

Society of Zoologists, Friday Harbor Labs of the University of Washington,

Harbor Branch Oceanographic Institution, Office of Naval Research and the

Hopkins Marine Station of Stanford University, Sigma Xi, National Museum

of Natural History of the Smithsonian Institution, Department of Zoology of

the University of Florida, Graduate Student Council of the University of

Florida, Graduate School of the University of Florida, National Geographic

Society, National Science Foundation. I thank the faculty and crew of the

research vessels of the Kristineberg Marine Biological Station, Fiskebackskil,

Sweden, the Marine Research Institute of Reykjavik, Iceland, the University

of Tromso Marine Station, Norway, The Huntsman Marine Science Centre,








New Brunswick, Canada, and Stan Harmon of Sitka, Alaska for their skill,

bravery, and perseverance at sea in such an esoteric endeavor as collecting

starfish.














TABLE OF CONTENTS


A CKN O W LED G M EN TS............................................... .......................................... ii

LIST O F TA BLES ................................................................................................... viii

LIST O F FIG U R ES.................................................................................................... ix

A B ST R A C T ................................................................................................................. x

CHAPTERS

1 OVERVIEW OF DISSERTATION............................................. ..................1

The Evolution of Complex Life Cycles..........................................................1
W hat is a Larva? .................................................................................................. 1
What is Direct Development? ........................................................................4
Direct Development Among Echinoderms.....................................................7
Ramifications of Evolutionary Changes in Development...............................7
O overview By Chapters........................................................................................ 9
C h ap ter 2 .......................................................................................................... 9
C h ap ter 3 ........................................................................................................ 10
C h ap ter 4 ........................................................................................................ 10
C hap ter 5 ........................................................................................................ 11

2 PHYLOGENY AND THE EVOLUTION OF DEVELOPMENT .......................12

Most Asteroids Develop as Larvae.......................... ........................................13
Asteroid Systematics and Development............... .................................15
What is the Ancestral Condition for Asteroid Development? ...................22
Patterns of Asteroid Development and Coelomogenesis.............................24
Paxillosids With Indirect Development and Pelagic Feeding
L arv ae .......................................................................................................25
Paxillosids with Indirect Development and Pelagic Nonfeeding
L arv ae .......................................................................................................28
Asteroids of the Forcipulatida and Valvatida with Indirect
Development and Pelagic Feeding Larvae...........................................29
Asteroids in the orders Valvatida and Forcipulatida with
Indirect Development and Pelagic or Benthic Nonfeeding
L arv ae .......................................................................................................30








Features of Coelomogenesis in Taxa Closely Related to Pteraster
(Superorder Spinulosacea) .......................................................................32
Unity and Diversity in the Coelomic Development of Asteroids ...............34


3 DEVELOPMENT WITHIN THE GENUS PTERASTER................................36

Introd u action ......................................................................................................... 36
M materials and M ethods........................................................................................ 37
Collection and Rearing of Pteraster tesselatus.....................................37
Collection and Rearing of Pteraster militaris............................................38
New Brunswick, Canada .......................................................................38
Trom s N orw ay .............................................................................. .. .....39
N orth of Iceland ........................................................................................39
Collection of other Species of the Genus Pteraster.....................................41
Collection of Pteraster acicula ......................................... ............... 42
Collection of Pteraster sp. (cf. P. hastatus)....................................43
Collection of Pteraster stellifer..............................................................43
Collection of Pteraster sp........................................................................44
Collection of Pteraster sp. (cf. Marsipaster)......................... .............44
Collection of Pteraster temnochiton....................................................44
Collection of Pteraster pulvillus..........................................................44
Collection of Pteraster obscurus........................................................44
M icroscopy........................................................................................................... 45
Preparation of young of Pteraster tesselatus for Scanning
Electron Microscopy ......................................................................45
Preparation of young of Pteraster tesselatus for physical
sectioning................................................... .........................................45
3-d reconstruction of Pteraster tesselatus........................................46
Preparation of other species of the genus Pteraster for
Confocal Laser Scanning Microscopy ..............................................47
Preparation of other species of the genus Pteraster for
Scanning Electron Microscopy....................................................48
R esu lts ..................................................................................................................... ...48
Development of Pteraster tesselatus......................................................48
G astrulation ............................................................................................ 48
A xes and sym m etry ............................................................................... 59
Morphogenesis of the water-vascular system...................................59
Formation of the perivisceral coeloms and axial complex.................61
Development of the gut and anterior compartment.........................62
Development of external features of the mesogen and
juvenile..............................................................................................63
Development of Species of Pteraster that Brood.......................................64
Development of external and internal features of Pteraster
m ilitaris..............................................................................................65








Development of external and internal features of Pteraster
acicu la .............................................. ..................................................... 69
Development of external and internal features of Pteraster sp.
(cf. P hastatus) .......................................................... .......................... 73
Development of external and internal features of Pteraster
stellifer.......................................................... ..... .... ...................... 76
Development of external and internal features of Pteraster
sp ...................................... ............ ............................... .................. 76
Development of external and internal features of Pteraster sp.
(cf. M arsipaster) ....................... .......................... ....... ................... 77
External features of Pteraster temnochiton.....................................80
Pteraster pulvillus .................................................. ......................80
Pteraster obscurus .............................................. .....................................80
D iscu ssion ....................................................................................... .. ................... 80
Reproductive Periodicity of Pteraster miltaris..........................................80
Developmental Habitat of Pteraster militaris.............................................. 82
The Polarity of Developmental Transitions................................................82
The evolution of brood-protection.....................................................82
The evolution of direct development.................................. ..........83

4 THE DEVELOPMENT OF HYMENASTER PELLUCIDUS, A
PTERASTERID ASTEROID WITH BROODED BRACHIOLARIAN
L A R V A E ................................................................................................. ...................85

In tro d u ctio n ................................................................................................ ......... 85
Materials and Methods...................................................................................86
Collection of Hymenaster pellucidus........................................................ 86
Survey of Hymenaster species in the USNM.............................................87
Microscopic Methods Used to Study Hymenaster pellucidus..................88
Preparation for Scanning Electron Microscopy.....................................88
Preparation for Confocal Laser Scanning Microscopy .........................88
R esu lts.................................................................................................................. 89
Reproduction of Hymenaster pellucidus collected off Iceland................89
Development of External and Internal Features of Hymenaster
p ellu cid u s .................................................................................... .............90
D iscu ssion ................... ..................................................................................... ....95
Reproductive Periodicity of Hymenaster pellucidus...............................95
Distribution of Hymenaster pellucidus ................................................95
Polarity of Developmental Transitions ................................................97
Brachiolaria are Plesiomorphic among the Pterasteridae...................97
Direct Development Evolved Subsequent to Brooding among
the Family Pterasteridae...........................................................98

5 SUMMARY AND GENERAL CONCLUSIONS................................................99








Direct Development Occurs in a Small Subset of the Asteroids that
Brood-Protect Their Young..........................................................................99
Evolutionary Pathways Between Pelagic and Benthic Development
do not Depend on Changes in Morphogenesis.......................................100
In Most Asteroid Species with Benthic Development there are
Functional Reasons to Explain the Evolutionary Persistence of
Larval Structures........................................................................................102
Do Asteroid Lineages with Benthic Development Re-evolve Pelagic
D evelopm ent?............................................................................................ 103
Under Which Circumstances Does Direct Development Evolve?.............104

LIST O F REFEREN CES .........................................................................................107

BIOGRAPHICAL SKETCH.................................................................................122













LIST OF TABLES


Table page

1. Collection and Development of Species of the genus Pteraster....................40

2. Abbreviations Used in Figures 5 and 6....................................... .......................49

3. Collection and Development of Hymenaster pellucidus..............................96


viii













LIST OF FIGURES


Figure............................................................................................ ....................... a

1. Diagram of an Idealized Bilaterian Life Cycle............................................ 6

2. Cladogram of Selected Asteroid Taxa and Developmental Patterns.............17

3. Three Topologies for the Cladogram of Asteroids..........................................21

4. Comparison of Coelomic Origins and Fates During Development in
D different A steroids ........................................................................................... 27

5. Internal Morphogenesis of the Early (6-8d) Mesogen of Pteraster
tesselatus. .... ................................................................. ..................................... 51

6. Internal Morphogenesis of the Late (9-11d) Mesogen of Pteraster
tesselatus ............................................................................................................ 55

7. Internal and External Morphogenesis of Pteraster militaris.........................67

8. Internal and External Morphogenesis of Pteraster acicula............................72

9. Internal and External Morphogenesis of Pteraster sp. (cf. P. hastatus).........75

10. Morphogenesis of Pteraster stellifer, Pteraster sp. (cf. Marsipaster),
Pteraster temnochiton, and Pteraster pulvillus...........................................79

11. Internal and External Morphogenesis of Hymenaster pellucidus ..............94













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

RECONSTRUCTING THE EVOLUTION OF MORPHOGENESIS AND
DISPERSAL AMONG VELATID ASTEROIDS

By

Daniel Andrew Janies

December, 1995




Chairman: Dr. Larry R. McEdward
Major Department: Zoology


Most echinoderms develop via bilaterally symmetrical larvae. A

juvenile forms within the larval body as an imaginal rudiment. At

metamorphosis the radially symmetrical juvenile reshapes, or resorbs, or

casts off portions of the larval body.

Development via a larva is indirect development because specialized

non-adult morphologies and metamorphosis are used to reach the adult body

plan. Very few species of echinoderms lack larvae and undergo direct

development. In direct development the adult body plan develops from the

embryonic organization thus obviating metamorphosis.

I studied the evolution of direct development and dispersal within

clade of deep-sea starfish, the family Pterasteridae. Pterasterids have a unique

adult body plan including a supradorsal membrane under which embryos are








brooded. Several species of the genus Pteraster (family Pterasteridae) have
direct development. 3-d reconstructions of serial sections were used to

illustrate that the direct developing young of Pteraster have been remodeled

evolutionarily with respect to larvae. In Pteraster, evolutionary changes in
the points of origin and fates of the early coeloms occurred to produce a
juvenile directly from the embryo.
One species, Pteraster tesselatus, has pelagic direct development, an
extremely rare pattern. Pelagic direct development likely evolved from an

ancestral pattern of brooded direct development rather than pelagic larval
development. P. tesselatus uses juvenile tube feet in lieu of larval

settlement structures at the end of the pelagic period. The re-evolution of
pelagic development is evolutionarily and ecologically significant because it

provides for high dispersal potential.

Hymenaster pellucidus (family Pterasteridae) broods larvae. Larvae

such as those of H. pellucidus are plesiomorphic among the Pterasteridae
because they are similar to the larvae of the outgroup family Solasteridae.
The prevalence of brooded young and the ubiquity of a body plan specialized

for brooding among the Pterasteridae indicates that brooding is plesiomorphic

among the family. The discovery of a larva in H. pellucidus supports the
hypothesis that direct development evolved subsequent to the evolution of
brooding and that the unique brooding morphology of pterasterids is not a

result of direct development. The ubiquity of direct development among the
genus Pteraster supports the hypothesis that direct development is not an

adaptation for the re-evolution of pelagic dispersal.













CHAPTER 1
OVERVIEW OF DISSERTATION


The Evolution of Complex Life Cycles

The majority of marine invertebrates have complex life cycles divided

into two (or more) discrete life history phases, larval and adult. Many

arguments have been proposed for the evolution and persistence of complex

life cycles such as: 1) advantages afforded by larvae for increased dispersal, and

exploitation of habitat and nutritional niches that differ from those occupied
by the adults (Strathmann, 1993) and 2) decoupling of developmental phases

from adult phases thus allowing larvae and adults to respond independently

to selective forces without affecting the other phase of the life cycle (Moran,

1994). These arguments are related to the specific issues in this dissertation. I

have investigated an instance of the evolution from a complex life cycle to a

simple one. Specifically, I analyzed the evolutionary loss of larval

development (the transition to direct development) in a clade of asteroid

echinoderms to explore the following question: Do changes in ecology of

development require changes in morphogenesis?


What is a Larva?

"A larva should be defined by developmental criteria as an
intermediate stage in the life cycle that is produced by post-embryonic

morphogenesis and is eliminated by the metamorphic transition to the

juvenile. In addition, a larva should be defined by a morphological criterion,




2

in that it must possess transitory structural features that are not

developmentally necessary for morphogenesis of the juvenile stage."

(McEdward and Janies, 1993).

Previously, an ecological paradigm was favored by invertebrate

biologists concerned with classifying the diversity of developmental patterns

(reviewed in Young, 1990). Consensus for definitions of many basic terms

necessary for interpreting the diversity of development in marine

invertebrates has been elusive. The problem stems from multiple contexts,

ecological and morphological, inherent to the discussion of the evolution of
marine invertebrate larvae. For example, there has been ambiguity in the

definition of a larva and a developmental pattern.

One may define a larva, ecologically, as having a habitat distinct from

that of the adult (Chia, 1974). This criterion fails in many lineages that have

recently undergone evolutionary transitions from pelagic larvae to brood-

protection of larvae or vice versa (examples from asteroids are discussed in

Chapter 2). Pelagic or benthic development can occur in two closely related

species within a genus (e.g., the asteroid genus Patiriella, Byrne, 1991; or the

asteroid genus Pteraster discussed herein). However, pelagic -- benthic

transitions have little or no immediate effect on the morphology and
function of the young (treated in detail in Chapters 2). In other words,

brooded young often undergo the same morphogenesis as pelagic larvae.

Under certain ecological definitions of the term "larva", classification of

developmental stages as "larvae" was based on were development occurs

(e.g., separate from the parent; sensu Chia, 1974) rather than on what stages
are actively involved. Under this definition one could not call any brood-

protected larvae, "larvae" because these young are not separate from the

parent. Such definitions hamper evolutionary studies of life cycle diversity.







We have argued that a larva is best defined by morphogenesis (not ecology)

because larvae can have a variety of ecological specializations (feeding or

nonfeeding, pelagic or benthic) (McEdward and Janies, 1993).

Traditional classifications of developmental patterns (Thorson, 1950;

Chia, 1974; Mileikovsky, 1971, 1974, Oguro et al., 1976; Chia et al., 1993) have

been ambiguous because they defined patterns based on a single character

such as: 1) presence of a feeding larva or a nonfeeding larva or 2) having a
certain larval stage or not (such as the presence or absence or a brachiolaria in

asteroids or 3) having pelagic or benthic development (see also McEdward

and Janies, 1993, for a review of traditional classification schemes). Schemes

based on a single character fail to capture the full diversity of developmental

patterns. Some traditional patterns, such as "planktotrophy" (larvae that feed
on phytoplankton) are defined and named on the basis of one character

(nutrition), whereas others, such as brooding, are defined on the basis of a

different character (habitat) (McEdward and Janies, 1993). Furthermore,
"lecithotrophy" (larvae that are nonfeeding but rather live off yolk reserves)

is often equated with direct development without distinguishing between

developmental habitat or type of morphogenesis (Wray and Raff, 1991).

Multifactor classifications of developmental patterns have recently

been proposed to incorporate the ecological and morphological diversity of

larvae (McEdward and Janies, 1993; Levin and Bridges, 1995). These
incorporate factors such as nutrition, habitat, and type of morphogenesis. A

developmental pattern is a unique combination of life cycle character states.

The two character states for morphogenesis are complex larval morphology

(indirect development) and simple morphology (direct development). The
two character states for the mode of nutrition during development are

feeding and nonfeeding. Offspring that feed require exogenous, particulate







food to fuel growth and development. Nonfeeding development is
supported by endogenous nutritional materials provided in the egg. With
regard to habitat, development can occur either in the water column (pelagic
development) or on the sea floor benthicc development). We arranged these
features (nutrition, habitat, and type of morphogenesis) into a matrix of three
independent characters, each with two states, to yield a total of eight
developmental patterns (McEdward and Janies, 1993). The advantage of this
system for classification of pattern, over traditional classifications is that it is
explicitly multifactor and therefore unambiguous. Our classification scheme
can be readily expanded by adding additional factors (sexual vs. asexual
reproduction) or adding levels of detail within a factor (e.g., benthic
development with or without parental care) (Levin and Bridges, 1995).
However, the three main characters seem sufficient to define patterns that

encompass much of the diversity within the echinoderms.


What is Direct Development?

Larvae are extremely common among bilaterian phyla (i.e., metazoan
phyla that evolved after Cnidaria and Porifera). Larvae of bilaterian phyla
share many features that evolved at the base of the Cambrian (circa 590 ma)

(Wray, 1995). These features include: feeding with ciliary currents (some
taxa, such as crustaceans, have secondarily lost this) and complex
metamorphosis (JAgersten, 1972; Strathmann, 1978).
Larvae evolved at the base of the bilaterian lineage and characteristic
larval types evolved within each phylum as they diversified. Subsequently,
direct development evolved via the independent loss of larvae in lower taxa
(genera, species) and a few phyla (e.g., Gastrotricha, Kinorhyncha, and
Nematoda). Some argue that there may have been independent evolution of





5

two types of feeding larvae after major lineages split (i.e., deuterostomes and

protostomes) (Neilsen and Norrevang, 1985). Others argue that nonfeeding
larvae are primitive at the level of the bilateria (Haszprunar et al., 1995).

Nevertheless, for the purposes of this study, larval (indirect) development is
primitive with respect to nonlarval (direct) development.

In direct development, embryonic stages are followed by the

morphogenesis of the juvenile, without intervening larval stages (McEdward

and Janies, 1993) (Fig. 1). Since there is no larval morphology,

metamorphosis is obviated. For example, the asteroid Pteraster tesselatus is
radially symmetrical throughout development, lacks all larval structures, and

does not undergo a metamorphosis (Janies and McEdward, 1993). In P.

tesselatus the coeloms arise in a pattern that forms the adult organs and body

plan without being repositioned or removed by metamorphosis. Direct

development is an alternative path to the adult body plan that does not

include the formation and subsequent rearrangement of a larval body plan.

Direct development has been similarly defined by others describing
nonlarval morphogenesis. Bonar (1978) described direct development as

"ametamorphic development" among opistobranch mollusks. Schatt (1984)

referred to direct development in the echinoid Abatus cordatus as "ni larve,

ni metamorphose" (neither larva nor metamorphosis). Fell (1948) used the

terms "recession of metamorphosis and the loss of larval forms" to describe a

trend towards direct development among ophiuroids.

McEdward and Janies (1993) introduced the term "mesogen" (middle

stage of development) to recognize that even with direct development there

can be a prolonged period of morphogenesis between embryo and juvenile.

Use of the term "mesogen" also serves to avoid the aforementioned
ambiguity of defining a "larva" by ecological criteria. The term "mesogen"








defines only nonlarval morphogenesis and thus is equally applicable in cases

in which stages of direct development are free-living (e.g., Pteraster tesselatus,

McEdward, 1992, 1995) or brooded (e.g., several species of Pteraster described in

Chapter 3).


Grow
Mature


Juveni


Sexually
mature adult Gametogenesis

th &
nation
Gametes





le
D-. ir ect e Zy
Differentiation & Mesogenv
Morphogenesis ... 4 ff et

Differentiation & ..
r Morphogenesis Embry


Fertilization




gote

Cleavage &
Gastrulation

0o


Metamorphosis


Larva


Differentiation &
Morphogenesis


Figure 1. Diagram of an Idealized Bilaterian Life Cycle.

The small typeface identifies the major life cycle processes (e.g.,
Differentiation & Morphogenesis); The medium-sized typeface identifies the
major life cycle stages (e.g., Juvenile); The large typeface indicates the two
alternative types of development in the life cycle (e.g., Direct Development).
(After McEdward and Janies, 1993).








Furthermore, the term "mesogen" is applicable to any taxon with true direct

development. Mesogen should be adopted by others to: 1) describe stages of

direct developing echinoderms such as: Abatus (Schatt, 1985, 1987),
Neosmilaster (Bosch and Slattery in prep.; Janies, unpublished data,
Ophionereis schayeri (Fell, 1941 a), and Ophiomyxa (Fell, 1941 b), or 2)
replace the terms "paralarva" or subadultt" for the stages of direct developing
cephalopods (Young and Harman, 1988), and 3) to replace the term "froglet"
in direct developing anuran vertebrates (Diesel et al., 1995).


Direct Development Among Echinoderms

Known direct developers include all species of the asteroid genus
Pteraster (described herein), the echinoid genus Abatus from the shallow
regions of the Southern Indian Ocean and Antarctica (A. cordatus, Schatt,

1985, 1987; and perhaps A. shackletoni and A. nimrodi, Pearse and
McClintock, 1990).

Putative direct developers that have not been extensively sampled and

sectioned to rule out larval stages include the ophiuroids Amphiura
squamata (Russo, 1891), Ophiomyxa brevirima and Kirk's ophiuroid (cf.

Ophionereis schayeri) from New Zealand (Fell, 1941 a), the brooding asteroid,

Neosmilaster georgianus, from Antarctica (Bosch and Slattery in prep.; Janies,
unpublished data), and the brooding asteroid Patiriella vivipara, from
Tasmania (Dartnall, 1969; Chia, 1976; Byrne, 1991).


Ramifications of Evolutionary Changes in Development

The comparison of embryos and larvae of closely related species is a
rigorous means of addressing questions of how development evolved (Wray
and Bely, 1994). Many authors have addressed the alterations in








developmental mechanisms associated with the evolutionary changes in

development by combining cellular methods and species comparisons (e.g.,

cell lineage specific gene-expression in ascidians by Jeffrey and Swalla, 1991;

cell lineage tracing in sea urchins by Wray, 1990; and Wray and Raff, 1991).

Others have shown that gene flow between populations is limited as a

consequence of the shift from feeding to nonfeeding pelagic development in

sea urchins (these species both have indirect development, albeit with long

and short dispersal capability) (McMillan et. al, 1992).

Despite the advances in understanding mechanisms of evolutionary

change in development afforded by these approaches, few studies have

addressed the ecological ramifications of the evolutionary shifts to brood-

protection and true direct development (as defined above). In contrast, the

study of developmental diversity among asteroids presented herein provided

a rare opportunity to investigate not only the evolution of true direct

development but also the role of direct development in distinct ecological

transitions from dispersive to nondispersive to dispersive young.

In the study presented herein, development was compared among

several starfish species of the order Velatida to test whether the ecological

transitions between pelagic and brooded development and vice versa depend

on radical modifications in development. The major research questions

were:

Are transitions between pelagic and brooded development dependent

on changes in morphogenesis or are they simple ecological shifts in that occur

in spite of changes in morphogenesis?

Does the evolution of direct development occur as a transformation

series of developmental modifications, and if so, which modifications are

associated with the shift from pelagic larval development to brooded direct







development and which are associated with the evolution of pelagic direct

development?


Overview By Chapters


Chapter 2

This chapter includes a review of: asteroid morphogenesis, ecological

features of asteroid development, and the controversy surrounding ordinal

level phylogeny within the class Asteroidea. Cladograms based on molecular

sequences or adult morphologies provide an independent framework with
which one can study evolutionary changes in development among asteroids.

An important example of this methodology is provided. There is

disagreement among systematists on which is the most primitive group of

asteroids. Hence the primordial asteroid larval type remains unresolved.
Despite the discordance at certain levels of asteroid classification, a

recent consensus has emerged regarding the families included in the clade of

asteroids in which I have studied development (Superorder Spinulosacea).

This has allowed me to postulate important preadaptations and a

transformation series of developmental and ecological modifications that

lead to the evolution of direct development in the genus Pteraster.

Relationships within families remain unresolved. As a postdoctoral project, I
will attempt to resolve a species-level phylogeny among available species of

the Superorder Spinulosacea. This phylogeny will provide an independent

test of the polarity and tempo of evolutionary changes in development.








Chapter 3

This chapter is an in-depth description of the development of Pteraster
tesselatus and descriptions of several other species of the genus Pteraster. P.

tesselatus has pelagic direct development whereas all other species of the
genus Pteraster brood-protect direct developing young. Despite the ecological
differences, the morphogenesis of the genus Pteraster is relatively uniform.
It has been possible to rear young of Pteraster tesselatus in vitro and
study its morphogenesis intensively. However, rearing of live embryos and
mesogens has not possible for most other species of Pteraster. Most of the
specimens are known only from serendipitous collections of brooding adults,
often from the deep-sea. Thus, I have used Pteraster tesselatus as a model to

which the other species are compared.
The near ubiquity of brooded direct development among Pteraster and

the fact that closely related outgroup species (i.e., Hymenaster pellucidus)
brood, suggests that brooding is the ancestral state for the family Pterasteridae
and genus Pteraster. Differences in development among Pteraster species

provide support for the hypothesis that accelerated development of tube feet

may have allowed for the evolution of pelagic development in Pteraster

tesselatus. The evolution of pelagic development is ecologically significant

because it contradicts the conventional theory that brood protection is an
irreversible specialization.


Chapter 4

This chapter is a description of the development of Hymenaster
pellucidus. This species is in the family Pterasteridae, as is the genus
Pteraster. H. pellucidus is the most closely related outgroup to the genus







Pteraster for which development is known. H. pellucidus has a brooded

brachiolarian larva that is similar in some external and all internal features to

the brachiolarian larvae of other taxa in the superorder Spinulosacea. One
significant feature in the development of H. pellucidus distinguishes it from

other brachiolarian larvae. In H. pellucidus, a supradorsal membrane forms

at the end of larval development and during juvenile development (the

supradorsal membrane forms the brood chamber in adult pterasterids). H.

pellucidus brood-protects it young thus, it represents an ecological

intermediate in the shift from pelagic brachiolarian larvae among the

Spinulosacea to brooded direct developers among the Pterasteridae. Despite

the formation of the supradorsal membrane, there is little basis from which

one can argue that there are morphogenetic changes in the larval body plan of

H. pellucidus associated with the ecological transition to brooding.


Chapter 5

In each of Chapters 2 through 4, conclusions specific to the information

in that chapter are discussed. In chapter five, conclusions and new working

hypotheses based on the project as a whole are discussed. The special nature

of brooding in the family Pterasteridae is discussed with respect to the role it

played in the evolution of direct development. The re-evolution of pelagic

development may not be unique to Pteraster tesselatus. Many brooding

lineages do not lose larval features and hence may revert freely between

pelagic and brooded development.

Fossil data indicate that brooding lineages are more subject to

extinction than lineages with pelagic larvae in coastal regions that experience
sea level fluctuations. I discuss a hypothesis of a deep-sea refuge for lineages

with direct development.













CHAPTER 2
PHYLOGENY AND THE EVOLUTION OF DEVELOPMENT




Phylogenetic relationships among taxa can be used as an a priori

hypothesis from which one can study the evolution of development among

the taxa. In this study, asteroid classification and phylogeny are the
frameworks on which developmental transitions are mapped. With this

method it is possible to reconstruct the polarity of changes, the frequency of

independent events, and the sequence in which events may have occurred.

In taxa where the phylogeny is controversial, developmental characters are

sometimes used to help define clades (Wray, 1995).

The development of selected asteroid taxa is presented in this chapter

with respect to characters such as nutrition (feeding or nonfeeding), habitat

(pelagic or benthic), and type of morphogenesis (indirect = larval or direct =
nonlarval). This background is important to the specific hypotheses on the

relationship between developmental habitat and morphogenesis in

subsequent chapters on the evolution direct development among the genus

Pteraster. Conservative features of asteroid morphogenesis (coelom

formation and organogenesis of the juvenile) provide a basis from which to
explore the changes associated with evolution of direct (nonlarval)
development. Many asteroid clades have independently evolved nonfeeding

larvae and benthic development (Fig, 1) but have not lost larval

morphogenesis. In the following sections and in Chapters 3 and 4, I show








that the starfish most closely related to Pteraster superorderr Spinulosacea)
share unique derived features of coelomogenesis (synapomorphies) that
preadapted this clade for the evolution of direct development (see also Janies
and McEdward, 1993). Two recent independent phylogenetic studies have
argued that the Spinulosacea is a monophyletic clade using synapomorphies

such as: adult morphology (Blake, 1987) and molecular sequence data (Lafay
et. al., 1995).


Most Asteroids Develop as Larvae

The asteroid larval body plan is characterized by: 1) bilateral symmetry,
2) three pairs of larval coeloms arranged serially from anterior to posterior:

axocoels, hydrocoels, and somatocoels, 3) an anterior lobe lined by coelom

that bears the brachiolarian arms (if present), 4) the axis of radial symmetry of
juvenile rudiment forms in a plane orthogonal to the anterior posterior axis
of the larva (McEdward, 1995). In larval asteroids the coeloms originate as
one or two lateral evaginations from the anterior of the archenteron
(enterocoels) and establish the initial bilateral symmetry. Later, coeloms

divide, grow, change shape, and reconnect in complex patterns to form
various juvenile structures (reviewed in Erber, 1985, Chia and Walker, 1991,

Chia et al., 1993, and Janies and McEdward, 1993) (discussed in detail in

subsequent sections and chapters).
The bipinnaria is a the first larval stage that occurs in the life cycle of all
asteroids with feeding larval development (e.g., Asteriasforbesi, A. vulgaris,
Agassiz, 1877; Luidia sarsi, Wilson, 1978; Patiriella regulars, Byrne and
Barker, 1991). A bipinnarian larva is characterized by the bilateral
arrangement of the pre- and post-oral ciliated swimming and feeding bands

that are borne on arms (MacBride, 1914; Kum6 and Dan, 1968). Bipinnarian








arms are hollow extensions of the body wall that contain blastocoelic space.
Unlike the larval arms of ophiuroids and echinoids, bipinnarian arms are not
supported by calcareous skeletal rods.
The brachiolaria is the second larval stage in asteroids that have

feeding development (except those of the order Paxillosida). The brachiolaria
is the only larval stage in asteroids that have nonfeeding development. The
brachiolaria is characterized by (often 3) arm-like processes of the anterior lobe

that serve for reversible attachment to, and movement along, the substratum
at settlement. An adhesive disk is often present between the brachiolar arms
for more secure attachment to the substratum during metamorphosis. The
brachiolar arms are extensions of the ectoderm of the anterior lobe of the

larva that are lined by coelom. Brachiolar arms may have evolved in

conjunction with the tube feet of adults based of the structural similarities
(Fell, 1967).
The juvenile rudiment is a pentaradial, disc-shaped complex of
coeloms that form the putative water-vascular system, digestive, hemal,

perihemal, and major body coeloms of the adult. The juvenile rudiment

forms in the left-posterior of the larva, in a sagittal plane with respect to the
larval anterior-posterior axis. Hence the juvenile rudiment determines the

oral surface of the adult on the left side of the larva. Many purely larval
structures such as the right axocoel, anterior coelom, and anterior lobe (with

brachiolar and feeding structures where applicable) are resorbed at

metamorphosis.
A taxonomically based description of asteroid developmental diversity
is presented following the discussion of asteroid phylogeny. The considerable
morphological and ecological diversity of asteroid larvae seems to argue

against the idea of a typical larva or pattern of development. However,







almost all asteroids undergo larval (indirect) development followed by
metamorphosis whether or not the young feed during development. The
morphogenesis of coeloms and their post-metamorphic fates are relatively
conservative among almost all asteroids (Hyman, 1955) despite diversity in
larval form, habitat, and nutrition (Janies and McEdward, 1993). Direct
developing asteroids use an alternative pattern of coelomogenesis to produce
a juvenile that is identical in body plan to that of asteroids that develop via

larvae. On the basis of this understanding, it was possible to infer the
coelomic rearrangement associated with evolution of direct development in
Pteraster tesselatus as modifications of the pattern of coelom formation found
in many asteroid larvae (Janies and McEdward, 1993). These inferences were

accomplished by tracing the formation of coeloms from juvenile organs back
to points of origin in the archenteron of the embryo.


Asteroid Systematics and Development

The topology of orders in the cladogram of representative asteroid taxa
used in this study is based on the results of Lafay et al. (1995) (Fig. 2, 3). The

developmental information results from original study (presented in

Chapters 3 and 4) and a review of the literature on asteroid development.
Lafay et al. (1995) based the ingroup topology of asteroid orders on an
unrooted cladogram resulting from strict consensus parsimony analysis of

morphological characters. However, Lafay et al. (1995) rooted their

cladogram with parsimony analysis of sequence data (28s rRNA) from
asteroids and other echinoderms. Within the cladogram of representative
asteroid taxa (Fig. 2), the positions of Echinasteridae, Solasteridae, and
Pterasteridae are based on the family-level phylogeny of Blake (1987) and the
species-level nominal classification of Clark and Downey (1992). I will













o (16 '42
C 0
1 & ,2 |



S-o .4 .




(U
o 3, 0e
0\0 z 5 b



M n (vo






-4
0 Si v < 6
e o







S 0 > m 0
$-4 a
1 ta '5 go
0 cu ^
.2s a A



S3g a 5>
0g 8 aZ

;-4 t M -1








,a o 0 ( U o
E .^: go






0 4.


o 50 o.
CM < o .
X g5-sj Cht



















ID
0
CI

a.



^ e
\'~


IK


,a

CL
aX

Q.

Is
S *
ll
s'
o*t Q.


M~oF


%


,"~s~~


b

I



~.,








attempt to resolve the topology within the Pterasteridae in a future study.
Relationships within Pterasteridae are important as independent tests of the
hypotheses on the polarity of character transitions presented in this study.
The phylogeny of asteroids is controversial. Blake (1987) significantly

reorganized the families that comprised orders with respect to the non-
cladistic classification of Spencer and Wright (1966). Blake (1981) split Spencer
and Wright's large and likely paraphyletic order, Spinulosida, into three
orders: Velatida, (e.g., families Solasteridae, Pterasteridae, and some rare deep-
sea families), Spinulosida (composed of a single family, Echinasteridae), and
Valvatida (e.g, family Asterinidae). Blake (1987) placed the orders
Forcipulatida and Brisingida at the root of his cladogram. Blake (1988) argued

that the Paxillosida are highly derived because they are specialized for soft
sediments. The Paxillosida lack features of other asteroids (e.g., brachiolarian
larval stages, suckered tube feet, and in some species an anus) and have been

historically considered to be the most primitive extant asteroids (Spencer and
Wright, 1966). Blake (1987) placed the Valvatida as the sister group of the
Paxillosida (Fig. 3).

Gale (1987) analyzed morphological characters of extant and fossil taxa
and used synapomorphies to support a tree (Fig. 3). However, Gale did not
test for alternative topologies using parsimony or any other analyses. Unlike
Blake, Gale reasserted that the Paxillosida are primitive. In addition, Gale's
ordinal concepts differed from Blake's. Gale recognized only four extant
orders and lumped the Echinasteridae, Pterasteridae, and Solasteridae as stem
groups of his largest order, the Valvatida. In doing so, Gale largely
maintained Spencer and Wright's (1966) likely paraphyletic order
Spinulosida, the only significant difference was that Gale named it the








Valvatida (e.g., Asterinidae, Echinasteridae, Solasteridae, and Pterasteridae).
Gale places his Valvatida as the sister group of the Forcipulatida (Fig. 3).

The families in Gale's concept of the Valvatida do not share derived

features of development. For example, several species among the family

Asterinidae have simple pelagic or benthic nonfeeding larvae (MacBride,

1896; Byrne, 1991, 1995; Byrne and Barker, 1991). However, because there are

pelagic feeding larvae among the Asterinidae, it is clear that simple pelagic or

benthic nonfeeding larvae evolved within species at the crown group, rather

than the stem lineage of the family (Fig. 2). Gale includes the families

Solasteridae, Echinasteridae, and Pterasteridae in his concept of Valvatida.

These families lack feeding larvae and share derived features of development

that originated at the stem lineage of their clade. Thus any similarities in
derived features of development among the Asterinidae and the families

Solasteridae, Echinasteridae, and Pterasteridae must be regarded as

convergent.

Blake's concepts of the Velatida (e.g., families Solasteridae and

Pterasteridae) and Spinulosida (i.e., Echinasteridae but not Asterinidae) are

much more in accord with developmental similarities documented by

independent analyses (Masterman, 1902; Gemmill, 1912, 1916, 1920; Janies and

McEdward, 1993). Together as a clade, the orders Velatida and Spinulosida

form the superorder Spinulosacea (e.g., families Echinasteridae, Solasteridae,

and Pterasteridae) that are the focus of the study described here) (Fig. 2).

Lafay et al. (1995) combined and recorded the morphological data sets of
Gale and Blake. Lafay et al. (1995) produced a tree that is largely similar to
Gale's with respect to the placement of the Paxillosida as the sister taxon to

the rest of the asteroids and Forcipulatida as sister taxa to the remaining

orders. However, Lafay et al. (1995) use the ordinal concepts defined by Blake














OcO






0)4

cuc;
'4 0 (t'







(u
-40)
0 a







IMa
00
boo








0

N 0(U
;0)0
mg



oo.
4-~






(U "14


000

cn -4-4
o4bl
o
0
-d








0)0
4-
o



0)
Itl
ob




(U 03 bo (
o
H4
0) (Ut E
0 ) 8







ir: cnc(U(U










(Jalseu!qo3 pue eioJuaH) ep!solnu!dS


ep!lelaA



(eBuaIJsy) BP!leleAA


ep!IelndiooJo


(ualjodoltsy) ep!sol!xed


JalsessoJo


Jaiseu!o3-3 puBe ePIIJUa


ep!lBIndlidoj.


( epn7I pue uaeoadoilsy) ep!soll!xed




(jlaseuiqo3 pue epIiJUaH) ep!solnu!ds,

(JalsBaeJd 'JsasessoJo) epelepa A

( eBp!n7 pue ualoadoisV) ep!soll!XBd ,




(euiJalsV) Bep!eAleA ,


ep!Ielnd!ojo-I







(i.e., Valvatida, Velatida, and Spinulosida). By accepting Blake's ordinal
concepts, Lafay et al. (1995) maintain the taxa important to this study

superorderr Spinulosacea) as a clade (Fig. 2, 3). (Note that a synapomorphy

was erroneously coded by Lafay et al, 1995, regarding features of coelom
formation that are convergent between the Asterinidae and Spinulosacea;
however, it does not affect their choice of most parsimonious topology).

Thus the cladogram used in this study (Fig. 2) is based on ordinal concepts of
Blake (1987) and the methods of phylogenetic inference of Lafay et al. (1995).
The cladogram (Fig. 2) provides a framework in which taxa with

independent sets of developmental modifications are separate.


What is the Ancestral Condition for Asteroid Development?

The following review is an important example of how views on
asteroid phylogeny strongly influence perceptions on the evolution of
development. Indirect development with a complex, feeding, pelagic larva is

considered the ancestral condition for echinoderms (JAgersten, 1972;

Strathmann, 1978). However, considerable debate over the larval stage that

comprised asteroid life cycle has occurred (Bather 1921a, b, 1923; MacBride,
1921, 1923a, b; Mortensen, 1922, 1923).
Paxillosids are considered by many to be the most primitive of the

starfish orders. Thus indirect development via a pelagic feeding bipinnaria
with complex larval morphology, but without a brachiolaria stage, is
characterized as the ancestral developmental pattern of asteroids (Mortensen
1921; Fell, 1967; Oguro et al., 1988; Wada et al., 1994; Lafay et al., 1995). The

brachiolaria was thought to have evolved later in asteroid evolution, at about

the same time as the appearance of suckered tube feet (Fell, 1967).







In contrast, MacBride (1921, 1923b) argued that paxillosids are

specialized, not primitive, and the lack of a brachiolarian larva is an

adaptation for life on soft substrata. Blake's phylogenetic (1987) and
functional analyses (1988, 1989) renewed support for the view that paxillosids
are highly derived asteroids. Blake (1988, 1989) refuted the traditional lines of

argument for paxillosids as primitive. Moreover, Blake (1988, 1989) provided
evidence that paxillosids are adapted to sandy or muddy habitats, such as

pointed rather than suckered tube feet and oral rather than anal elimination
of shell and large-grain sediment waste. Some paxillosids were reported to

have suckered tube feet for a short period immediately after metamorphosis

(Astropecten latespinosus, Komatsu, 1975a; Astropecten scoparius, Oguro et
al., 1976). The presence of suckered tube feet on paxillosid juveniles has been

used by Blake (1987) to refute the putative evidence that paxillosids diverged
from the rest of the asteroid lineage before the evolution of suckered tube

feet. However, the claim that species of the families Luidiidae and

Astropectinidae have suckered tube feet as juveniles has been retracted

recently (Komatsu et al., 1994; M. Sugiyama, pers. comm.).

We had adopted the view that the original asteroid life cycle consisted
of pelagic feeding indirect development via a complex bipinnaria and a

complex brachiolaria (McEdward and Janies, 1993). This adoption was a

byproduct of accepting the topology of Blake's (1987) entire phylogeny for the

class Asteroidea along with the ordinal concepts that well suited our view of

evolution of development within a subset of the asteroids superorderr
Spinulosacea) (Janies and McEdward, 1993). The most robust phylogenetic
information to date (Lafay et al., 1995) argues that Paxillosids (at least the clade

that includes the genus Astropecten) are primitive asteroids. Another group

makes similar claims, but they have not yet published complete analyses







(Wada et al., 1994). Thus it is most prudent at this point to accept the view
that a feeding bipinnaria (without brachiolaria) was the ancestral larval stage

among asteroids. The phylogeny of Lafay et al. (1995) is significant because it
allows me to consolidate the adoption of two likely scenarios of the evolution
of development among asteroids: 1) the most primitive life cycle (i.e., larval
stage) among extant asteroids is represented by the feeding bipinnaria without
a brachiolaria in the Paxillosida and 2) a highly derived pattern of coelom
formation (i.e., posterior enterocoely, Janies and McEdward, 1993) is

synapomorphic among the Spinulosacea. Previously, if one accepted that the
paxillosids were primitive, then one had to accept the ordinal concepts of
Gale (1987) that lumped too many taxa together to allow resolution of the
evolution of derived developmental patterns. On the other hand, one had to

accept the controversial conclusion that paxillosids were not primitive if one
stressed the ordinal concepts of Blake (1987) that made possible the mapping
of independent cases of the evolution of derived development patterns. For
example, nonfeeding brachiolaria evolved among the Valvatida
independently of the nonfeeding brachiolaria and direct developing

mesogens of the Spinulosacea.


Patterns of Asteroid Development and Coelomogenesis

Subsequent chapters present a description of highly derived
developmental patterns among the asteroid family Pterateridae, including
direct (nonlarval) development. Coelom formation in the direct developing
asteroid Pteraster tesselatus has been reconstructed using serial sections
(Janies and McEdward, 1993). In this work the process of tracing juvenile
structures back to original enterocoels was facilitated by making comparisons
of coelomic fates among asteroids with various patterns of development







including: 1) pelagic, feeding bipinnarian and brachiolarian larvae 2) pelagic,
nonfeeding brachiolarian larvae, and 3) benthic, nonfeeding brachiolarian

larvae. A brief review of development and coelom formation among
asteroids is presented in this chapter to provide a context for understanding
the hypotheses for the evolution of direct development presented in Chapters

3, 4, and 5. Examples of studies of morphogenesis in representative species,
rather than exhaustive lists of species, are discussed.


Paxillosids With Indirect Development and Pelagic Feeding Larvae

Asteroids in the order Paxillosida with pelagic feeding development have

only a bipinnarian larval stage. This has been documented in the genera

Astropecten, Platyasterida, and Luidia (Chia et al., 1993). These larvae are

sometimes termed, "non-brachiolarian larvae" because they lack brachiolar

arms and the associated coelomic projections found on the anterior lobe of

other asteroid larvae (Oguro et al., 1976). In bipinnaria, the coeloms develop
from a pair of enterocoels that evaginate from the tip of the archenteron (Fig.

4 I A-B). The enterocoels develop into three pairs of coeloms (from anterior

to posterior: axocoels, hydrocoels, and somatocoels) (Fig. 4 I C). The

somatocoels pinch off from the enterocoels as separate coeloms early in larval

development, but the axocoels and hydrocoels are often confluent via small

coelomic tubes and hence are considered axohydrocoels by some authors

(Hyman, 1955). However, each coelomic sac or region can be considered to

have a separate role in organogenesis (Fig. 4). The left and right axocoels fuse
in the anterior region of the larva to form a single coelom (axocoel) in the

shape and orientation of an inverted U (Fig. 4 I D) (Gemmill, 1914). The
axocoel extends to the anterior into the anterior lobe (e.g., Astropecten

scoparius, Oguro et al., 1976, p. 561, 566; Luidia clathrata, Komatsu et al., 1991,























Figure 4. Comparison of Coelomic Origins and Fates During Development in
Different Asteroids.


Comparison of coelomic origins and fates during development (A through
D) in different asteroids (I, II, and III). All drawings are oriented as dorsal views
of longitudinal sections with the anterior end up. A key to the stippling
indicating coelomic fates is provided. I A-D. Ancestral pattern of
coelomogenesis via feeding bipinnarian and feeding brachiolarian larval stages.
II A-D. Coelomogenesis in the nonfeeding brachiolarian larvae of the superorder
Spinulosacea (e.g., the families Solasteridae and Echinasteridae). III A-D.
Coelomogenesis in Pteraster tesselatus. (After Janies and McEdward, 1993).













Left /
SOral





Left/


Stipple Cavity


Preoral coelom

Axocoel

Hydrocoel

Somatocoel

Somatocoel

Archenteron


Resorbed

Inner oral perihemal coelom
Axial perihemal coelom

Water-vascular system

Oral perivisceral coelom
Outer oral and radial perihemal coeloms

Aboral perivisceral coelom


Gut


C





0(I


Adult fate


El


Eli



El







p. 497). This portion of the axocoel is termed the anterior coelom (or preoral

coelom). These larvae lack brachiolar apparatus. At settlement from the

plankton following metamorphosis the juvenile tube feet are used to attach
to the substratum.


Paxillosids with Indirect Development and Pelagic Nonfeeding Larvae

Three species (Astropecten latespinosus, Komatsu, 1975a; A.

gisselbrechti, Komatsu and Nojima, 1985; and Ctenopleura fisher, Komatsu,

1982) of the order Paxillosida develop via pelagic nonfeeding larvae. These

are termed barrel-shaped larvae because they lack the external morphology of

the bipinnaria such as: feeding structures including complex ciliated bands,

bipinnarian arms, and larval mouth. Barrel-shaped larvae also lack

brachiolar structures, the adhesive arms and disk. Barrel-shaped larvae occur

only in the Paxillosida. This strongly suggests that barrel-shaped larvae are

derived from ancestors that develop via feeding bipinnaria stages only (i.e.,

they lacked brachiolaria) (Komatsu et al., 1988; McEdward and Janies, 1993).

Furthermore, barrel-shaped larvae are the only nonfeeding larval type among

asteroids that are known to form paired coelomic compartments (Erber, 1985;

Komatsu, 1975a, 1982; Komatsu and Nojima, 1985; Komatsu et al., 1988).

Typically bipinnaria produce paired coelomic compartments, which

subsequently fuse anteriorly, at the brachiolaria stage. Nonfeeding
brachiolarian larvae, which are never preceded by a bipinnaria stage, form

only a single, unpaired coelomic compartment (Erber, 1985).

Although few details of internal development are known, coelom

formation in barrel-shaped larvae appears similar to the pattern described for

feeding bipinnaria because it includes the formation of a U-shape and an

anterior coelom. McEdward (1995) has pointed out the ambiguity regarding







the orientation of the juvenile rudiment with respect to the larval body

among barrel-shaped larvae. For example, the figures of Ctenopleura fisher
(Komatsu, 1982, p. 202 Fig. 4 I and K) suggest a frontal (rather than sagittal)

orientation of the rudiment. It is also ambiguously reported that barrel-

shaped larvae attach to the sea floor with the tip of the larval "stalk" (anterior

lobe) (Komatsu, 1975a, p. 52-54 her fig. 23). This is perhaps simply a

settlement posture and not an attachment because the author does not

explain how settlement is accomplished if the stalk lacks brachiolar arms and

attachment disk.


Asteroids of the Forcipulatida and Valvatida with Indirect Development and
Pelagic Feeding Larvae

Many species of these orders have feeding bipinnarian and

brachiolarian larval stages (e.g., Forcipulatida: Asterias rubens, Gemmill 1914;

Valvatida: Porania pulvillus, Gemmill, 1915, Patiriella regulars, Byrne and

Barker, 1991). A brachiolarian larva is defined by the presence of specialized
attachment structures on the anterior lobe: the brachiolar arms and

attachment disk. Brachiolar arms are hollow, but contain extensions of the

larval anterior coelom, and are thereby distinguished from bipinnarian arms

(Gemmill, 1914). Brachiolar arms are used by larvae to test the substratum

and provide initial, temporary adhesion during settlement. The adhesive

disk secretes cement and provides stronger attachment at metamorphosis

(Barker, 1978). In asteroids with feeding development, the brachiolarian stage

only occurs after the bipinnaria. These are not independently evolved types

of larvae but simply sequential developmental stages.

The enterocoels develop into three pairs of coeloms (from anterior to

posterior: axocoels, hydrocoels, and somatocoels) (described above and in Fig.







4 I A-D). The only significant difference in coelomogenesis between
bipinnaria and brachiolaria is that the anterior coelom extends into the

lumen of the brachiolar arms (Barker, 1978).

Asteroids in the orders Valvatida and Forcipulatida with Indirect
Development and Pelagic or Benthic Nonfeeding Larvae

Several asteroids of the order Valvatida develop via pelagic
nonfeeding brachiolarian larvae (e.g., Asterina coronata japonica, Komatsu,

1975b; Asterina batheri, Kano and Komatsu, 1978; Patiriella gunni and P.

calcar, Byrne, 1991; Mediaster aequalis, Birkeland, et. al., 1971; and Iconaster

longimanus, Lane and Hu, 1994). Demersal nonfeeding larvae are inferred to
be present in the orders Forcipulatida based on studies of egg size (Tyler et al.,
1984) however there are no descriptions of the development of the larvae.

There are few studies of internal structure in pelagic nonfeeding brachiolaria
of valvatids (e.g., Patirella gunni and P. calcar, Byrne, 1991). Nonfeeding
brachiolaria show simplification of external development with respect to

feeding larvae. Nonfeeding brachiolaria lack feeding structures such as the

ciliated band, the larval mouth, and the convoluted extensions of the larval

body.
A variety of means of keeping young on the benthos have evolved in
the orders Valvatida and Forcipulatida. Several species attach eggs to the sea

floor (e.g., Valvatida: Asterina gibbosa MacBride, 1896; Patiriella exiuga,
Byrne, 1995; and Forcipulatida: Leptasterias ochotensis similispinis Kubo,

1951). In many species, female adults brood by posturing their arms in a cup
like fashion and trapping young amongst the spines and tube feet of the adult

(e.g., Valvatida: Asterina phylactica Emson and Crump, 1976; Strathmann et
al., 1984; Forcipulatida: Leptasterias hexactis, Osterud, 1918; Chia, 1968; L.







miller, Masterman, 1902; L. polaris, Hamel and Mercier, 1995). A few

species of Valvatida develop while brooded in the ovary (e.g., Asterina

pseudoexiuga pacifica Komatsu et al., 1990) or in the ovary and subsequently

in the adjacent coelom (e.g., Patirella vivipara, Dartnall, 1969; Chia, 1976;

Byrne, 1991). A few species of Forcipulatida develop while brooded in the

stomach of the adult (e.g., Leptasterias groenlandica, Liberkind, 1920; L.
tenera, Worley et al., 1977; and Smilasterias multipara, O'Loughlin et al,

1994). Despite the evolutionary shifts from feeding to nonfeeding and or

pelagic to benthic development, all of these species develop via a brachiolaria

(Patirella vivipara is a possible exception. Chia, 1976 and Byrne, 1991 report

that P. vivipara lacks a larval stage and has direct development. However

this has not been proven conclusively with sections of all stages of

development).

There are few morphological and functional differences between the

larvae of related species with pelagic nonfeeding development and benthic

nonfeeding development. In taxa that do not have internalized brood-

protection, it is apparent that brachiolar arms are maintained by selective

forces that favor protection and retention of larvae (treated in Chapter 5). As

stated earlier, pelagic nonfeeding development has evolved independently in

species or genera not at the stem of the orders Forcipulatida and Valvatida.

Simplified nonfeeding brachiolaria or nonfeeding brachiolaria with benthic

adaptations are not shared due to common ancestry between the
Forcipulatida and Valvatida, but rather represent convergent evolution

subsequent to the loss of feeding larvae (Fig. 2).

Coelomogenesis in pelagic and benthic nonfeeding brachiolaria shows
a few differences from the pattern described above for feeding larvae. For
example, in Leptasterias hexactis (Chia, 1968) and Asterina gibbosa (MacBride,







1896) a single, large enterocoel evaginates from the anterior tip of the
archenteron. This unpaired coelomic sac expands anteriorly and laterally,

then extends backwards with a pair (left and right) of short, posterior
projections, thus producing the characteristic U-shaped coelom. The lateral
projections eventually give rise to separate coelomic regions: a sagittally

oriented hydrocoel on the left side of the larval body and a somatocoel on
each side of the larva. The anterior region of the unpaired U-shaped
coelomic sac extends into the lumen of the brachiolar arms and is considered

homologous to the anterior coelom of feeding larvae (Erber, 1985;
Strathmann, 1988). Thus despite the simplifications, the larval body plan is
maintained and there is no justification to consider nonfeeding brachiolaria

as representative of direct development (some authors disagree, e.g., Chia et
al., 1993).


Features of Coelomogenesis in Taxa Closely Related to Pteraster (Superorder
Spinulosacea)

Detailed studies on internal features of development have been

performed for the following species: order Spinulosida, Henricia

sanguinolenta (formerly Cribella oculata) Masterman (1902); Fromia

ghardaqana, Mortensen (1938); and order Velatida: Solaster endeca, Gemmill
(1912, 1916); Crossaster papposus, Gemmill (1920). The archenteron constricts

into unpaired anterior, middle, and posterior portions (Fig. 4 II A, B). The
anterior portion and the posterior portion of the archenteron are
presumptive coelom, while the middle portion is presumptive gut
(Masterman, 1902; Gemmill, 1912, 1920; Hyman, 1955, p. 298). The axocoels,
hydrocoels, and right somatocoel develop from the unpaired anterior coelom,

as is typical in nonfeeding larval development (Fig. 4 II B-D). However, the







left somatocoel develops from the posterior enterocoel (Fig. 4 II B, C) in
proximity to its definitive larval location. The anterior, unpaired enterocoel
develops from the anterior portion of the archenteron (Fig. 4 II A-II B). Two

long lateral projections extend from the anterior enterocoel to the posterior of

the larva, producing a U-shaped coelom. A hydrocoel pinches off from the

left lateral region of the U-shaped coelom. A somatocoel (adult aboral

perivisceral coelom) forms from the posterior portion of the right side of the

U-shaped coelom. The left anterior region of the U-shaped coelom forms the

axocoel (adult perihemal coelom of the axial complex and the inner oral
perihemal ring coelom). An anterior coelom lines the brachiolar arms (Fig. 4

H B-D).

There is one striking difference in the pattern of coelomogenesis in

solasterids and echinasterids that is a radical departure from the ancestral

pattern of coelom development. There has been a change in the site of origin
of the left somatocoel. This heterotopic modification, which produces
multiple (i.e., anterior and posterior) enterocoels, is a synapomorphy of the

asteroids of the superorder Spinulosacea (Fig. 2).

The evolution of posterior enterocoely dissociated the origin of the left

somatocoel from the anterior region of the archenteron, in the lineage

leading to P. tesselatus. I believe that this modification of coelomogenesis

was a necessary precursor to the radical rearrangements of coelomogenesis
that occurred in P. tesselatus to produce direct development. This was an

important preadaptation for the novel positioning and acceleration of water-

vascular morphogenesis that occurred in the evolution of pterasterids.

Because P. tesselatus lacks the bilateral pattern of enterocoely and rather has a
radial and transverse pattern, it is improper to use the terms "left" and
"right". Although Gemmill (1912, 1916, 1920) and Masterman (1902) referred







to the posterior coelom of the spinulosaceans as the "left posterior coelom,"

because of homology with the left somatocoel of other asteroid larvae, we
refer to this cavity, which is also found in P. tesselatus, as the large posterior

coelom.


Unity and Diversity in the Coelomic Development of Asteroids

In spite of considerable diversity in external larval morphology, the

pattern of coelomic development is very similar among asteroids regardless
of shifts from feeding to nonfeeding or from pelagic to benthic development.

On the basis of the similarity in anterior coelomic development, Erber (1985)

argued that an anterior coelom develops in the anterior region of the larva in

all asteroids and therefore a homologous brachiolaria stage occurs throughout

the Asteroidea, including species in which brachiolar arms are absent (e.g.,

paxillosids and P. tesselatus). Erber's (1985) criterion for brachiolarian

development is sufficiently general to include all known asteroids, yet also

leads to at least two conclusions that I believe are false. First, Erber (1985)

concluded that paxillosids have a "brachiolarian" stage of development.

However, as stated earlier, paxillosids develop as feeding bipinnaria or

nonfeeding simplifications thereof -- they are not brachiolaria nor are they

derived from (i.e., homologous with) brachiolaria (Komatsu et al., 1988).

Second, Erber (1985) assumed that the mode of coelom formation in P.

tesselatus would be identical to that of its close relatives, the echinasterids and

solasterids. This assumption was based, in part, on the prevailing

interpretation of the evolution of the "larva" of P. tesselatus. Fell (1967)
considered the absence of brachiolar structures in P. tesselatus to be the result

of minor simplifications of the external structures of a typical pelagic

nonfeeding brachiolarian larva.







In all larval types, most of the anterior of the larva (including the

anterior coelomic structures) is resorbed at metamorphosis. This resorption

is accomplished by translocation of the larval structures to the oral side of the

juvenile rudiment, followed by histolysis and transfer to the juvenile
digestive tract (Chia and Burke, 1978). The postmetamorphic (juvenile and
adult) fates of the larval coeloms are highly conservative among asteroids as
illustrated by a comparison of Figures 4 I and 4 II. The left side of the axocoel

forms the perihemal coelom of the axial complex and the inner oral

perihemal ring coelom (Fig. 4) (Hyman, 1955; Dawydoff, 1948). The left

hydrocoel forms the entire coelomic lining of the water-vascular system. The

left somatocoel forms the oral perivisceral coelom by growing dorsal and

ventral extensions medial to the hydrocoel. The outer oral perihemal ring

coelom forms from small interadial evaginations of the oral perivisceral

coelom. The right axohydrocoel is mostly resorbed at metamorphosis yet

becomes the madreporic vesicle (an aboral component of the axial complex)

in some species (Hyman, 1955). The right somatocoel lies on the side of the

gut opposite the hydrocoel (i.e., on the aboral side of the juvenile) and forms

the aboral perivisceral coelom.

Given the conservative nature of asteroid coelomogenesis, analysis of

the evolution of direct development can be focused on the transition from

the larval pattern of coelomogenesis in the nonfeeding brachiolaria of taxa

closely related to P. tesselatus (e.g., Solaster, Crossaster, and Henricia) to the
highly derived pattern of P. tesselatus.













CHAPTER 3
DEVELOPMENT WITHIN THE GENUS PTERASTER

All species of Pteraster in which development has been studied have

direct development. In this chapter, I discuss the collection of specimens and

describe the internal and external features of development of several species.

I discuss topics specific to Pteraster in the concluding sections of this chapter

and topics of significance to the entire study in Chapter 5.


Introduction

Pteraster tesselatus is unique among species in the genus because it has

pelagic development. However, it is the only species of Pteraster for which a

complete sequence of developmental stages has been studied because it is the

only species that has been reared in laboratory culture. For this reason,

Pteraster tesselatus serves as the model against which all other species of

Pteraster have been compared.

The development of Pteraster tesselatus is morphologically different

from that of all other starfish (McEdward, 1992; Janies and McEdward, 1993).

P. tesselatus has pelagic development but does not pass through the typical

asteroid larval forms, the bipinnaria or the brachiolaria. Important features

that distinguish the development of P. tesselatus from other asteroids
include absence of specialized larval attachment structures (brachiolar arms

and adhesive disc); accelerated development of the water-vascular system and

the use of podia (tube feet) for attachment to the substratum at settlement;

radial rather than bilateral embryonic symmetry; parallel rather than







orthogonal embryonic and adult axes of symmetry; a transverse orientation of
the juvenile disc, and complex morphogenesis of a supradorsal membrane.

McEdward (1992) concluded that this set of unusual developmental

features in P. tesselatus characterized a novel type of pelagic larva in the

Asteroidea. Subsequent investigation of internal morphogenesis led to the

interpretation that P. tesselatus completely lacks a larval stage and undergoes

direct development (Janies and McEdward, 1993). By means of comparisons

with asteroids that undergo the indirect development via pelagic, feeding

bipinnarian and brachiolarian larvae, as well as with asteroids that develop

via pelagic and benthic nonfeeding brachiolarian larvae, we found the

coelomic and water-vascular development in P. tesselatus are unlike

anything previously described in the development of asteroids. Our

interpretation of the development of P. tesselatus is radically different from

the historically accepted views (Janies and McEdward, 1993). Chia (1966, p.

507-508) originally described this species as developing via a nonfeeding

bipinnaria. Others have interpreted the pelagic stage as a reduced

lecithotrophic brachiolaria (Fell, 1967; Erber, 1985; Oguro et al., 1988). Our

work has shown that these interpretations are incorrect. Pteraster tesselatus

has a highly derived pattern of direct development (Janies and McEdward,

1993).


Materials and Methods


Collection and Rearing of Pteraster tesselatus

Adults of Pteraster tesselatus Ives, 1888 were collected using SCUBA,
from subtidal populations (5 to 20 m) at several sites near the Bamfield

Marine Station (4849'N, 125"08'W) in Barkley Sound, Vancouver Island,







British Columbia, Canada and from depths of 15 to 30 m near the Friday
Harbor Laboratories (4832'N, 1230'W) in the San Juan archipelago,
Washington, U.S.A. Spawning was induced in adults of P. tesselatus by
intracoelomic injection of 2 to 5 ml [10-4 M] (molar) of the hormone 1-methyl
adenine (Sigma Chemical Co.). Eggs (= 1000-1400 pm in diameter) were
released within 1 to 3 h after injection. The eggs developed without artificial
insemination. This species may have natural facultative parthenogenesis
however this phenomenon has been extremely difficult to study due to the
very opaque, yolky cytoplasm of the eggs. There are no differences in the
coelomogenesis of the young whether eggs are inseminated or not. Young
were cultured in mesh-bottomed baskets suspended in a shallow tank of
flowing sea water at ambient temperatures (10-13C).


Collection and Rearing of Pteraster militaris

New Brunswick. Canada
Adults of Pteraster militaris were collected from Passamaquoddy Bay,

New Brunswick, Canada at depths of 15-30 m using SCUBA in September
1993 and October 1994. The brood chambers of large numbers of adults were
dissected. Many juveniles yet very few mesogens were found. The yield of
mesogens was not sufficient to provide information on the pattern of
coelomic and water-vascular system morphogenesis.

Spawning was induced in Pteraster militaris by means of 1 to 3 ml (10-4
M) injections of the hormone 1-methyl adenine. Approximately 100 females

were induced to spawn during the study periods. Cultures were established
from eggs that were released from the brood chamber. These eggs were
fertilized with a suspension of spermatozoa from dissected testes of the largest
male adults or spermatozoa that was sometimes spawned by males after







injection. The fertilized eggs were maintained in vitro in mesh-bottomed

baskets in flow through sea water table at 11-13 "C. Attempts were also made

to fertilize eggs that remained within the brood chamber of adult females by

immersing females in sperm suspensions or by placing spawning males and

females in the same beaker for several seconds or minutes. In all culture

systems, early embryonic cleavages often occurred, but embryos never

developed beyond the wrinkled blastula stage. This failure was unlikely to be

the consequence of poor culture conditions since several antibiotics were

tested and culture techniques known to be successful for most asteroids.

Tromso. Norway

A small number of adult P. militaris were collected by SCUBA off

Tromse, Norway at depths of 10-30 m during the last week of November 1993.

Spawning, fertilization, and culturing of released and brooded eggs was

conducted as described above. Embryonic development was similar to but at

a much slower rate than observed in New Brunswick because the sea water

temperature was much lower (4C) in Tromse. Development proceeded

through cleavage stages, wrinkled blastula stages, hatching of gastrulae from

the extraembryonic membranes, and elongation of the gastrula as described

for P. tesselatus. However upon examination of these stages with confocal

microscopy, no consistent pattern of morphogenesis among embryos could be

discerned. Thus development was abnormal in these specimens.

North of Iceland

Over 100 mesogens were dissected from a single specimen of Pteraster

militaris collected on July 16, 1993, by the BIOICE program off the north coast

of Iceland at a depth of 719 m (Table 1, see reference BIOICE 2581). A detailed

description of the internal structure of these mesogens is provided below.









- tI _I 7I 1_ 7I 1 -I I I I


& & & & -^ & & & g g






O ( \ 0 0 0 00
-1 !P qa





S00 C7 C C M M 0 4
0%r 0 m C Z 9

nw O 0 m M" iQ A

ON a% )I IZ










C'4 C1 ('4 (4 Ct Cl) C) ) w c(C4 0 V
Me a,

rN g 0o % o,



















'5 u 05 0 z
S a a
W M L m t im O M
ah Fl a ^ g- ^^^ S
iBQOO ^ .C^ -e ^ tI t ^ .S ,."
ha 'a< MIT? pl tQ V) t o ^
U~~ ~ ~ & -A~ ~~w
- ---- A
'2'M~
i 811118~ 11~1 !sz








Collection of Other Species of the Genus Pteraster

The brood chambers were searched for young in all adult individuals

of the genus Pteraster collected on various cruises of The Benthic

Invertebrates of Icelandic Waters (BIOICE) program between from May 1991

and August 1994. All specimens were fixed at the time of collection in 10%

formalin buffered in sodium borate. The BIOICE cruises sampled primarily

within the Icelandic Economic Zone, an area of sea banding the entire coast of

Iceland to 200 km offshore. BIOICE cruises have also sampled along the

Faroe-Iceland Ridge which extends beyond the 200 km limit from the

southeast coast of Iceland. From these searches, young (mesogens and/or

juveniles) were discovered in seven species, including: Pteraster militaris

(Muller and Troschel, 1842) (as described above), Pteraster sp. (cf. P. hastatus,

Mortensen, 1913), Pteraster acicula (Downey, 1973) Pteraster obscurus

(Doderlein, 1900), Pteraster pulvillus (Sars, 1861), Pteraster sp., and Pteraster
sp. (cf. Marsipaster, Sladen, 1882, 1889). (Note, the taxonomic references

provided are representative of the first authority to use the genus and species

designation as listed in this study. Older descriptions of the same species

under a different name may precede the reference provided. See Clark and

Downey, 1992, for a complete systematic history of each species).

The brood chambers were searched for young in many individuals of

14 species of the genus Pteraster from the non-type, wet specimen collection

of the National Museum of Natural History of the Smithsonian Institution,

Washington, DC USA (USNM) including: Pteraster abyssorum, Pteraster

acicula, Pteraster affinis lebruni, Pteraster coseinopeplus, Pteraster jordani,

Pteraster lebruni, Pteraster marsippus, Pteraster obesus myonotus, Pteraster

obscurus, Pteraster pulvillus, Pteraster stellifer, Pteraster temnochiton,








Pteraster tesselatus arcuatus, and Pteraster militaris. Among these species, a
few individuals were found to be brooding juveniles that were fully
developed and hence not of interest to this study. However, mesogens that
provided valuable information on the type of development were discovered
in two species: Pteraster stellifer (Sladen, 1882, 1889) and Pteraster
temnochiton (Fisher, 1910) (Table 1, see references USNM E10070 and 31741).
Most the species in the wet collection of the following genera of the order
Velatida were also searched for young, including: (expressed as family: genus
or genera) Pterasteridae: Diplopteraster, Euretaster, Hymenaster; Solasteridae:
Heterozonias; Korethrasteridae: Peribolaster; Myxasteridae: Pythonaster;
Caymanostellidae: Caymanostella. Although, young were not found among
any of these specimens, it is important to report that they have been searched
to avoid duplication of effort by other students of velatid asteroids.

Collection of Pteraster acicula
Mesogens and juveniles were dissected from the brood chambers of
adult specimens of Pteraster acicula. Adults collected off the southwest coast
of Iceland in September 1992 at depths of 426 550 meters contained juveniles
(Table 1, see references BIOICE 2245 and 2247). Adult collected along the
Faroe-Iceland Ridge off the southeast coast of Iceland in May 1993 at depths of
550-698 meters contained mesogens and juveniles (Table 1, see references
BIOICE 2332 and 2343). Thus I infer the spawning season of Pteraster acicula

to include April in the northern hemisphere. However the year-round
collections necessary to evaluate continuous or other reproductive schedules
have not been compiled for this species.
A number of specimens of P. acicula were examined. These were

collected by the USNM at depths of 3460 to 5011 meters, in the Venezuelan
Basin in November 1981. Many pieces of tissue or eggs that protrude through








the body wall encased in dermal papillae were collected from these specimens

and some BIOICE specimens. Clark and Downey (1992) reported, "One

specimen has few large eggs visible through the actinal membrane of this

species." Dissection and confocal microscopic examination was performed in

an attempt to confirm the origin of these "eggs". The tissues did appear to
originate from ripe gonads but no indication of embryonic development

could be confirmed with optical sections. A single adult specimen was found

to contain both protruding "eggs" and young mesogens. However these

mesogens were not encased in the sac of the dermal papillae. Either the

mesogens are never in the papillae or they break out shortly after

development commences. I conclude that the tissues are eggs from ripe
gonads protruding through the body wall. However, I cannot determine

whether this protrusion is due to artifacts of preservation, or pressure change

upon collection, or a novel process of spawning via dermal papillae rather

than gonopores.

Collection of Pteraster sp. (cf. P. hastatus)

Two specimens of Pteraster sp. (cf. P. hastatus) were collected off the

north coast of Iceland on July 8, 1992 at a depth of 208 meters (Table 1, see

reference BIOICE 2126). The brood chambers and gonads were dissected from

both specimens (one male and one female). The female specimen contained

about 20 juveniles in interadial and apical regions of the brood chamber. In

this female specimen, there were also eggs that had not developed protruding

through the body wall.

Collection of Pteraster stellifer

Adult specimens of Pteraster stellifer were collected by the staff of the

USNM off the south east coast of East Falkland Island at depths of 137-146

meters on January 1, 1932. Specimens were fixed in formalin at the time of







collection and stored in 70% ethanol for 62 years (Table 1, see reference

USNM E10079). The brood chambers of two adults specimens were dissected

and 2 early juveniles were found in each.

Collection of Pteraster sp.
A single adult female specimen of Pteraster sp. was collected off the

north coast of Iceland on July 8, 1992, (Table 1, see reference BIOICE 2128).

Two early juveniles were dissected from the brood chamber of the adult

specimen.

Collection of Pteraster sp. (cf. Marsipaster)

Three adult male and one adult female specimens of Pteraster sp. (cf.

Marsipaster) were collected off the north coast of Iceland on July 14, 1994, at a

depth of 435 meters (Table 1, see reference BIOICE 2604). A few mesogens and

late stage juveniles were dissected from the brood chamber of the female

specimen.

Collection of Pteraster temnochiton

Adult specimens of Pteraster temnochiton, were collected by the staff of

the USNM off Akun Island, Unimak Pass (Aleutian Islands) Alaska at a depth
of 102 meters (Table 1, see reference USNM 31741). Many juveniles were

dissected from the brood chamber of a single adult.

Collection of Pteraster pulvillus

A single female specimen of Pteraster pulvillus was collected in a fjord

off the northwest coast of Iceland on July 11, 1994 at a depth of 138 meters

(Table 1, see reference BIOICE 2530). A few late juveniles were dissected from
the brood chamber of this specimen.

Collection of Pteraster obscurus

A single female specimen of Pteraster obscurus was collected off the

north coast of Iceland on July 14, 1994, at a depth of 435 meters (Table 1, see








reference BIOICE 2604). Several very old juveniles were dissected from the

brood chamber of this specimen.


Microscopy


Preparation of young of Pteraster tesselatus for Scanning Electron Microscopy

Specimens were fixed for scanning electron microscopy (SEM) in cold

osmium tetroxide (2% for 1 h) in 0.45 4m filtered sea water, rinsed twice in
distilled water, dehydrated through a graded ethanol series (30%, 50%, 70%, 15
minutes each), and stored in 70% ethanol. Dehydration of specimens was

continued to absolute ethanol (90%, 100%, 15 minutes each), then dried

specimens with hexamethyldisilazane (HMDS, Sigma Chemical Co.) for

several hours in air at room temperature (Nation, 1983) in a dust-free

chamber. Specimens were mounted on aluminum stubs, sputter coated the

specimens with gold-palladium, and stored the specimens under desiccation.

Specimens were viewed with a Hitachi SEM and images of a small
cathode ray tube were captured using Polaroid type 50 film. These images

were scanned at 300 dpi. Contrast, size, and tonal qualities of the images were

adjusted in Adobe Photoshop 2.5.1. Final copies of the images in the figures
were printed on a dye sublimation printer.

Preparation of young of Pteraster tesselatus for physical sectioning

Specimens were fixed for serial histological sectioning in Bouin's fluid
for 24 h, dehydrated them through a graded ethanol series for 15 minutes in

each concentration (30%, 50%, 70%), and stored specimens in 70% ethanol.

Later, specimens were dehydrated to absolute ethanol (90%, 100% 15 minutes

each, 100% overnight), transferred them to absolute ethanol with eosin y for
30 min., transferred them to xylene for 30 minutes, infiltrated them with a

graded series of Paraplast-xylene mixtures (at 56 OC under vacuum), then







embedded individual specimens in blocks of Paraplast. Embedded specimens

were sectioned serially at 7 or 12 p.m, and stained the slides in Hematoxylin -

Eosin. Some specimens were partially sectioned. The tissue remaining on

the block was prepared for SEM by dissolving the Paraplast in xylene and

drying the tissue with HMDS, as described above.

3-d reconstruction of Pteraster tesselatus

Serial sections were examined and photographed swith a compound
light microscope. For 3-d reconstruction, a camera lucida drawing tube was

used to trace each section (= 80 per mesogen) onto paper. Colored pencils

were used to trace the outer edges of the body wall, coelomic compartments,

and developing gut as a set of contours. Sequential sections were aligned

visually for registration. Each tracing was marked with a set of fiduciary

points for alignment during digitization. The x, y, and z coordinates of points

along the contours were digitized and stored as files in ASCII file format. The

coordinate data were plotted as graphic images of each section and stored as

raw 8-bit color bitmaps. Each bitmap file was converted to MCID file format

with programs written for this purpose by Dr. L. R. McEdward. Entire series

of sections were imported (series of MCID files) into the program NIH Image
1.44 (Rashband, 1992). Regions of images bounded by mesodermal and

endodermal contours were filled with color to aid visualization. Images

filled with color were saved as a series of files in TIFF file format. The 3-d

reconstructions were accomplished with stacking and projection routines

(brightest point algorithm in the macro provided by Mike Castle and Tom
Ford-Holevinski, University of Michigan, Ann Arbor) in NIH Image. In

different reconstructions, the transparency bounds were adjusted to remove

specific layers, such as the ectoderm, thus allowing visualization of internal
structures.







Preparation of other species of the genus Pteraster for Confocal Laser
Scanning Microscopy

Mesogens were fixed at various times of collection as described above

and stored in 70% ethanol within the brood chamber of the adult for two

months to two years for BIOICE specimens and 62 and 103 years for the

USNM specimens.

Mesogens were stained for confocal microscopy in 0.03% acridine

orange in 70% ethanol for 2 hours, dehydrated through a graded ethanol

series to 100% ethanol, transferred to methanol, and cleared for one hour in a

solution composed of a 2:1 or 3:1 ratio of benzyl alcohol to benzyl benzoate.

Acridine orange was used as a general stain of epithelia because it

stains nucleic acids and fluoresces well under the laser light created by the 488

nm channel and BHS cube of the BIO-RAD MRC 600 microscope. To

visualize the coelomic walls with sharp contrast to the background, images

were captured using an aperture setting of 50%, neutral density filter = 0

(100% transmission), electronic gain = 4, black level = 4, normal scan speed,

and 10 over scans of the Kalman filter.

Data sets composed of multiple images termed "z-series" were collected

to visualize the entire embryo in three dimensions. Z-series used in this

study were created from serial planar (horizontal axes x, y) optical sections

along the vertical (z) axis with a 10 x microscope objective at 5 jtm steps

through a total depth of 100 to 300 pjm. Each planar image of the z-series was

captured with a frame grabber at a resolution of 768 x 512 pixels. With a 10x

objective this capture resolution corresponds to physical dimensions 1359 im

x 906 gim. Thus each pixel diameter corresponds to 1.77 gpm of linear distance

on the specimen.







Z-series were imported into the program NIH Image 1.57 (Rashband,

1995) for analysis. Adobe Photoshop 2.5.1 was used for manipulation of

contrast and tonal qualities of images of representative sections. Images were
printed on a dye sublimation printer.

Preparation of other species of the genus Pteraster for Scanning Electron
Microscopy

Dehydration of specimens was continued through absolute ethanol

(90%, 100%, 15 minutes each), then specimens were dried with
hexamethyldisilazane (HMDS, Sigma Chemical Co.) for several hours at

room temperature (Nation, 1983) in a dust-free chamber exposed to the air.

Specimens were mounted on aluminum stubs, sputter coated with gold-

palladium, and stored them under desiccation.


Results


Development of Pteraster tesselatus


Gastrulation

Between day 2 and 3 of development in P. tesselatus, the archenteron

formed as a wide depression at the posterior end of the embryo. The

archenteron nearly filled the interior of the gastrula, and caused the

ectodermal and mesendodermal cell layers to lie close together. The

blastopore dosed between days 3 and 4 of development resulting in a

completely enclosed archenteron sac. The archenteron did not join with the

ectoderm to form a mouth opening. P. tesselatus produced lecithotrophic

mesogens (Fig. 5 A).

The gastrula elongated along the animal-vegetal axis and acquired an

ovoid body form, just before hatching from the extraembryonic membranes








(vitelline envelope and remnants of the jelly coat) at 3 days. Within 1 to 2

days of hatching, an ectodermal depression produced a groove (termed the

circumferential groove) completely around the circumference of the gastrula

and divided the body into anterior and posterior regions (Fig. 5 A, C).



Table 2. Abbreviations Used in Figures 5 and 6.


Abbreviation


Description


lp. coelomic lining of primary podia (tube feet) of water-
vascular system
20p. coelomic lining secondary podia (tube feet) of water-
vascular system
a. ampulla of podia (tube feet) of water-vascular system
a.c. anterior compartment of the gut
a.pv.c. aboral perivisceral coelom
a.r. anterior region of the mesogen
ax.c. axial coelom
c.g. circumferential groove
co.r.c. circumoral ring canal of water-vascular system
g. gut
h.l. hydrocoel lobe of water-vascular system
hp.c. hydropore canal
i.ph.c. inner oral perihemal ring coelom
I.c. lateral canal of water-vascular system
l.p.e. large posterior enterocoel
m.b. marginal bulge of the mesogen
m.1. marginal lobe of the juvenile
o.ph.c. outer oral perihemal ring coelom
o.pv.c. oral perivisceral coelom
p. ectodermal covering of podia (tube feet)
p.c. preoral coelom
p.r. posterior region of mesogen
r.c. radial canal of water-vascular system
s.d.m supradorsal membrane
s.p.e. small posterior enterocoel
t.p. Terminal podium (tube foot) of water-vascular system












Figure 5. Internal morphogenesis of the early (6-8d) mesogen of Pteraster
tesselatus.

Magnification is the same in all panels and scale bar equals 0.2 mm. In all
lateral views the posterior of the mesogen is oriented to the left of the page. See
Table 2 for descriptions of abbreviations. A. SEM, lateral view of 6d mesogen,
showing circumferential groove dividing anterior and posterior body regions.
Podia are visible in the circumferential groove. B. Drawing of lateral view of
mesogen in panel A showing location and orientation of planes of section for the
following panels of this figure. C-T. Paired light micrographs and interpretive
diagrams of histological sections. See Table I for abbreviations. C, D.
Longitudinal section of 6d mesogen showing origin of the enterocoels from the
archenteron and relationship of circumferential groove to the hydrocoels. E, F.
Transverse section of 5d 8h mesogen showing the first five evaginations
(hydrocoels) in pentaradial symmetry around the circumference of the
archenteron and their early contact with the overlying ectoderm. G, H.
Transverse section of 6d mesogen showing the bifurcation of the distal part of
each hydrocoel lobe that forms the coelomic linings of the first pair of podia. I, J.
Oblique section of 7d mesogen showing the coelomic lining of the terminal
unpaired podium forming as an extension from the cleft of the original
bifurcation and the second pair of podia coelomic linings evaginating between
the terminal podium and the first pair. K, L. Transverse section through a 5d 8h
mesogen showing the large posterior enterocoel evaginating as a large crescent
shape from the extreme posterior region of the archenteron. M, N. Transverse
section of 6d mesogen showing the large posterior enterocoel encircling the gut
and enveloping the posterior side of the hydrocoel lobes. 0, P. Transverse
section of 6d mesogen showing four of the five initial small pouches growing
orally from the oral perivisceral coelom between the hydrocoel lobes to originate
the outer oral perihemal coelom. Q, R. Longitudinal section of 6d 16h mesogen
showing the enterocoels especially the small posterior enterocoel and its mixed
set of fates of the (i.e., axocoelic and somatocoelic). S, T. Longitudinal section of
6d 16h mesogen (same mesogen as section in Q R) showing the somatocoelic
derivative (the aboral perivisceral coelom) separating from the distal region of
the small posterior enterocoel and moving to the extreme posterior of the body.
The proximal region of the small posterior enterocoel develops into the inner oral
perihemal ring coelom and likely contributes to the coelomic and hemal axial
complex, such as the axial sinus, ampoule, axial gland. (After Janies and
McEdward, 1993).

































s.p.e. D)










1p.


l.p.e. -


h.l.
o.ph.c.












































o.pv.c.
a.pv.c.
ax.c. &
i.ph.c.











Figure 6. Internal Morphogenesis of the Late (9-11d) Mesogen of Pteraster
tesselatus.

Magnification is equal in all panels (except G, H) and all scale bars equal
0.2 mm. In all lateral views the posterior of the mesogen is oriented to the left of
the page. See Table 2 for descriptions of abbreviations. A. SEM, lateral view of
9d mesogen, showing five marginal bulges and well developed podia. B.
Drawing of lateral view of mesogen in panel A showing location and orientation
of planes of section for the following panels of this figure. C, D. SEM and
interpretive diagram of 9d mesogen sectioned in two planes and oriented
obliquely to the viewer. Transverse face revealing the oral perivisceral coelom
and elements of the hemal and water vascular system in the oral region of the
mesogen, especially the ampullae of the podia. Longitudinal face revealing an
ampulla and the ectodermal covering of a cluster of podia. E-V. Paired light
micrographs and interpretive diagrams of histological sections. E, F. Slightly
oblique section of 9d mesogen showing the proximal lateral evaginations of the
radial canals that fuse to form the circumoral ring canal of the water-vascular
system. Also the proximity of the hydropore canal to the small posterior
enterocoel is visible. G, H. Transverse section at high magnification of an 8d
ambulacrum revealing the proximal lateral evaginations of the radial canals that
form the circumoral ring canal and showing the early development of a portion
of the outer oral perihemal coelom. I, J, K, L. Two longitudinal sections from the
same 10d mesogen. These illustrate the ectodermal origin of the hydropore
canal, it's communication with the proximal region of the small posterior
enterocoel to form part of the axial complex, and the connection of the
hydropore canal to the circumoral ring canal. M, N. Transverse section though
the completed water-vascular system of an l1d mesogen. O, P. Transverse
section through the same 11d mesogen as in panels M, N. This section is
slightly aboral to M, N and hence reveals the outpockets of the oral perivisceral
coelom that form the outer oral perihemal coelom. Q, R. Oblique section
through 9d mesogen showing formation of the crescentic shaped inner oral
perihemal ring coelom from the proximal region of the small posterior
enterocoel. This view illustrates the proximity of the inner oral perihemal ring
coelom with the circumoral ring canal of the water-vascular system. S, T.
Oblique section of 7d 16h mesogen showing the formation of the axial complex
from the communication between the inner oral perihemal ring coelom and the
hydropore canal. U, V, Longitudinal section of lid mesogen showing an adult
internal organization. At this stage the anterior compartment is in the process of
transferring the contents of the anterior region of the mesogen to the gut in
order to fuel development. (After Janies and McEdward, 1993).



























B














.ph.c.



D




-t.p.

co.r.c.
lp.
2p.


lp.


o.pv.c.l


co.r.c.;
r.c.



t.p.















o.ph.c.


CO.r.C.






J


























































o.pv.c.:


lp.





58












a.


co.r.c.


i.pv.c.
20 p.
hp.c. 10 To
ax.c. e i.ph.c.
o.pv.c. a.

S T

1.c.
2p.


o.pv.c.
co.r. c. 0 .ph.c.

a.py.c. a.c. .r.


.V.. o.py.c.
aco.r.c.
r.c.

U V







Axes and symmetry
The oral surface of the juvenile corresponded to the anterior region of

the mesogen (= animal pole of the embryo) and the aboral surface of the

juvenile corresponded to the posterior end of the mesogen (= embryonic
vegetal pole and blastopore) (Fig. 5 A, B the posterior end is on the left and

anterior end is on the right side of these figures). Consequently, the juvenile

disc developed in a transverse orientation with respect to the anterior-

posterior and animal-vegetal axes. All the stages of development were

characterized by radial, rather than bilateral, symmetry.

Morphogenesis of the water-vascular system

At 4 days, five lateral coelomic pouches (enterocoels) evaginated

simultaneously from the equatorial region of the archenteron (Fig. 5 D, 4 E-

H). The five enterocoels were arranged symmetrically around the

circumference of the archenteron in a transverse plane (i.e., perpendicular to

the anterior-posterior axis of the mesogen; Fig. 5 B). These five enterocoels

were hydrocoelic in nature because they became the coelomic lining of the

water-vascular system (i.e., radial canals, podia, ampullae, and circumoral

ring canal). Initially, these hydrocoel lobes were broad, simple evaginations

from the archenteron. The hydrocoel lobes elongated radially towards the

ectodermal body wall. The hydrocoel lobes remained connected to the

archenteron throughout much of the development of the water-vascular

system. The hydrocoel lobes and their relationship to the archenteron were

evident in transverse section (Fig. 5 E-H). At 5 days, the distal ends of the

hydrocoel lobes begun to contact the overlying ectoderm (Fig. 5 E, F). At the

time and location of this contact, the ectoderm folded inward to produce the

circumferential groove. The groove formed just anterior to the hydrocoel, so







that the contact with the hydrocoel lobes occurred along the posterior wall of
the groove (Fig. 5 C-J).
The proximal and central portions of each hydrocoel lobe developed
into a radial canal. The distal part of each hydrocoel lobe widened and
bifurcated to form the coelomic lining of the first pair of podia (Fig. 5 G, H).
The coelomic lining of each terminal (unpaired) podium developed as an
extension from the cleft of the original bifurcation. At 7 days, the coelomic
lining of the second pair of podia evaginated between the terminal podium
and the first pair (Fig. 5 I, J). In the region of the hydrocoel lobes, the
ectoderm of the circumferential groove thickened, enveloped the developing
coelomic linings, and formed the epidermal covering of the podia. In
subsequent development, additional pairs of podia were added immediately
proximal to the terminal podium as the radial canals elongated (Fig. 6 M, N).
The five hydrocoel lobes remained connected to the gut but were

otherwise independent of each other during the early development of the
radial canals and podia. Later, these five independent evaginations were
connected by a coelomic tube (ring canal) that encircled the gut. At 8 to 9 days,

each radial canal separated from the archenteron by a constriction of the oral

side of the proximal end of the canal. Concurrently, the aboral portion of the
proximal end of each radial canal produced small lateral evaginations in the

same (transverse) plane of orientation as the hydrocoel. These evaginations
grew around the circumference of the archenteron, met in the interadii, and
fused to complete the circumoral ring canal (Fig. 6 E-N).
The ampullae of the podia developed by an expansion of the coelomic

lining on the aboral side of each podium. By 8 days, the ampullae of the first

two pairs of podia expanded posteriorly against the oral perivisceral coelom
(Fig. 6 C, D). At 8 to 9 days, the hydropore canal developed as a long tubular







invagination from the aboral ectoderm in the extreme posterior of the
mesogen. The hydropore extended orally and joined with the developing

circumoral ring canal in an interadial position (yet only slightly to the side of

a radius) (Fig. 6 E, F, I, J, K, L, S, T). The hydropore canal was an ectodermal

structure that was conspicuous in early development because it stained very

deeply. I infer that the hydropore canal formed the stone canal of the adult as
is typical of asteroids although its development was not traced beyond 12 days.
The completion of the major water-vascular components (circumoral

ring canal, radial canals, podia with expanded ampullae, and hydropore canal)

at 8 to 9 days corresponded to the time at which the mesogen had functional

podia that were extended from the body and used to adhere to the substratum.
In laboratory cultures, settlement occurs as early as 1 to 2 days after

completion of a functional water-vascular system (McEdward, 1992).

Formation of the perivisceral coeloms and axial complex

Two additional enterocoelic evaginations followed the evagination of

the hydrocoel lobes from the archenteron (Fig. 5 C, D, K, L, Q-T): 1) a large

posterior enterocoel (= 6th of 7 enterocoels) whose fate was the oral

perivisceral coelom which in turn produced the outer oral perihemal ring

coelom in the juvenile; and 2) a small posterior enterocoel (= 7th of 7

enterocoels) that formed the axial and inner oral perihemal coeloms and the

aboral perivisceral coelom.

The large posterior enterocoel developed from the extreme posterior

region of the archenteron at 4 to 5 days. Initially, this enterocoel had the

shape of a very large crescent lying in a plane transverse to the anterior-

posterior axis of the mesogen (Fig. 5 K, L). By 5 days, this large posterior

enterocoel assumed the adult location of the oral perivisceral coelom: it
encircled the gut and enveloped the posterior side of the hydrocoel (especially







the ampullae of the podia) (Fig. 5 M, N, 6 C, D) At 6 days, the oral

perivisceral coelom developed five small pouches that grow orally (=

anteriorly) between the hydrocoel lobes in interadial positions (Fig. 5 0, P).

These pouches grew laterally to the radii where the pouches met and fuse to
neighboring pouches to form a single outer oral perihemal ring coelom (Fig.

60, P).

At 4 to 6 days, the small posterior enterocoel evaginated from the

archenteron posterior to the hydrocoel and slightly anterior to the site of

evagination of the large posterior enterocoel (Fig. 5 C, D, 6 M, N). The small

posterior enterocoel subdivided into a complex set of coeloms. At 6 days, the

aboral perivisceral coelom separated from the distal region of the small

posterior enterocoel and moved to the extreme posterior of the body (Fig. 5

Q-T). The proximal region of the small posterior enterocoel formed a

crescentic coelom near the aboral, inner surface of the circumoral ring canal

of the water-vascular system (Fig. 6 Q, R). This crescentic coelom developed

into the inner oral perihemal ring coelom. The middle region of the small

posterior enterocoel was confluent with the hydropore canal (Fig. 5 Q-T, 6 E,

F, I-L, S, T). The region of contact between these coeloms likely formed the

substructures of the coelomic and hemal axial complex, such as the axial
coelom and madreporic vesicle. A comprehensive and definitive description

of the development of the axial complex will require substantial additional

study.

Development of the gut and anterior compartment

A great deal of coelomic morphogenesis occurred in P. tesselatus while

the seven enterocoels remained confluent via the central portion of the

archenteron. Masterman (1902) described a similar arrangement of connected

enterocoels and archenteron as a "mesenteron" in Henricia sanguinolenta







(formerly Cribella oculata). The term, mesenteron, is very useful in

visualizing the internal development of P. tesselatus. At 4 to 5 days, in P.
tesselatus, despite the confluence among all seven of the enterocoels, the gut

endodermm) was histologically differentiated from the coelomic (mesodermal)
regions of the mesenteron. The gut was delineated as a deeply staining,

thickened epithelial region located posterior to the five hydrocoel lobes and

between the two posterior enterocoels (Fig. 5 C, D, Q-T).
The region of the mesogen anterior to the circumferential groove

contained yolk and fibrous material in its periphery but also contained an

epithelial layer that enclosed a large, central internal compartment (Fig. 5 C,

D, Q-T, 6 I-L). The epithelial compartment was not the product of enterocoely

but was simply the anterior of the archenteron produced by gastrulation.

Hence, the anterior of the archenteron is not a coelom but rather an anterior

compartment of the gut (Janies and McEdward, 1993; McEdward, 1995). The

anterior compartment remained confluent with the gut region of the

archenteron throughout juvenile development. Identifiable structures did

not develop in the anterior of the mesogen and its blastocoelic contents were

depleted during development (Fig. 6 U, V). Hence the anterior region of P.

tesselatus is not a preoral lobe as found in bipinnarian and brachiolarian
larvae.

Development of external features of the mesogen and juvenile

External development of P. tesselatus was described in detail by

McEdward (1992). The summary below is important to this document

because it provides a basis to which one can compare the development of

other pterasterids. Morphogenesis of the supradorsal membrane in P.

tesselatus began at 5 to 8 days with the formation of five broad marginal

bulges around the circumference of the body, immediately posterior to the







circumferential groove (Fig. 5 A, 6 A). The marginal bulges became bilobed at
8 to 10 days (Fig. 6 A, 0) and eventually divided to produce a total of 10
distinct marginal lobes at 16 days. Five additional lobes developed at the
aboral pole of the body at 11 to 13 days. Fusion of all 15 lobes occurred at 17 to
19 days and resulted in a complete supradorsal membrane above the aboral

body wall (McEdward, 1992). In the adult, the supradorsal membrane encloses

a space, the nidamental chamber. These structures (discussed in Chapter 5),
unique to pterasterids, serve to protect (physically and chemically with
mucus, Nance and Braithwaite, 1979), and perhaps nourish young in
pterasterids that brood (McClary and Mladenov, 1990).
In P. tesselatus, functional podia, arranged in five clusters, emerge
from the circumferential groove at 9 days, long before the development of

juvenile arms (at 2.5 months) (McEdward, 1992). Early development of the
podia was important because the mesogen of P. tesselatus lacked the
brachiolar arms and adhesive disc of pelagic nonfeeding larvae (Fig. 5 A, 6 A).
Podia of P. tesselatus were used for attachment to the sea floor at the end of

the pelagic period (which is assumed to be 10 to 12 days based on the behavior
of laboratory cultures) (McEdward, 1992). Each podial duster initially
consisted of a pair of podia and a terminal podium. Three to four additional
pairs of podia were added by 28 days. The juvenile mouth did not form until
the second month and distinct arms were not present in most juveniles of P.

tesselatus until the third month (McEdward, 1992).


Development of Species of Pteraster that Brood

All species of the genus Pteraster that have been studied brood-protect

direct developing young (except Pteraster tesselatus which has pelagic
development). The young of nine species are described below. However, the







samples are from random and serendipitous collections from the deep-sea of

adults that contained broods and fixed at single points in time. Rather
limited information can be garnered from these samples compared to that

presented for P. tesselatus which could be reared in vitro. In some species of
Pteraster, only juvenile stages were found (Table 1). I was able to determine

whether or not juveniles of a particular species develop through larval stages.

In the metamorphosis of an asteroid larva there is often a remnant of the
brachiolar complex (larval stalk or preoral lobe in paxillosids) that is resorbed

on the oral surface (or aboral surface in paxillosids) of the juvenile (e.g.,

Hymenaster pellucidus, Fig. 11 E). Many young stages of juvenile among

several species of the genus Pteraster were found. No species of genus

Pteraster has brachiolarian larval stages and direct development via a

mesogen is likley universal for the genus.

Development of external and internal features of Pteraster militaris

With the exception of the early works that incorrectly interpret the

development of external structures in Pteraster militaris (Sars, 1861; Koren

and Danielssen, 1856; and Kaufman, 1968) this is the first description of

development of Pteraster militaris. The following description is based on the

study of mesogens collected from a single specimen of Pteraster militaris from

Iceland (Table 1, see reference BIOICE 2581) and juveniles from a number of

individuals collected in New Brunswick.

The brooded mesogen of P. militaris was exceptionally similar to the
pelagic mesogen of Pteraster tesselatus (compare figures 5 and 7) in external

and internal features. Like the mesogen of Pteraster tesselatus, the mesogen

of P. militaris differed from all asteroid larvae in the following features:

absence of specialized larval attachment structures (brachiolar arms and

adhesive disc); radial rather than bilateral symmetry; parallel rather than



















Figure 7. Internal and External Morphogenesis of Pteraster militaris.
A. SEM, oblique view of mesogen (anterior region of mesogen is facing
viewer) showing the five marginal bulges on the posterior region, the anterior
region, and five pair of podia. B. Longitudinal optical section of the posterior
region of the mesogen (posterior region of mesogen is oriented towards the top
of the page). The following structures are visible: small posterior enterocoel,
large posterior enterocoel, hydrocoel lobes, and gut. C. SEM, lateral view of the
mesogen (posterior region is oriented towards the left of the page) showing five
marginal bulges on the posterior region, anterior region, and five pairs of podia.
D, E. Transverse optical sections through hydrocoel (water-vascular system) of
the mesogen showing the gut, the coelomic linings of primary podia, secondary
podia, terminal podium, radial canals, and the circumoral ring canal. (Note that
panel E shows an abnormal specimen with only four ambulacura, five
ambulacura is normal for the species). F, G, H. Three longitudinal optical
sections. F. Longitudinal superficial optical section of the posterior region
showing two hydrocoel lobes, two podia, and a portion of the large posterior
enterocoel. G. Longitudinal deep optical section of the posterior region showing
two hydrocoel lobes, two podia, a small portion of the large posterior enterocoel,
the small posterior enterocoel, and the gut. H. Longitudinal deep optical section
of the anterior region showing hydrocoel lobes, podia, and the anterior
compartment.

















































































































1







orthogonal embryonic and adult axes of symmetry; a transverse orientation of
the juvenile disc, and complex morphogenesis of a supradorsal membrane.
From external view, the brooded mesogen of P. militaris consisted of

three major regions: the posterior region that formed the juvenile, an

anterior region that was nutritive, and a ring of ten podia that were

positioned in the groove separating the two regions (Fig. 7 A, C). The
posterior region of P. militaris was identical to P. tesselatus with five broad
marginal bulges around the circumference of the mesogen immediately

posterior to the circumferential groove. These bulges overlie each

ambulacrum. The ectoderm of these bulges produced elaborate folds (Fig. 7
E). Late mesogen stages of P. militaris were not available. However, by

combining data on the early mesogens that I have with the sketches (but not
the interpretations) of Kaufman (1968, p. 508), I infer that the pattern of

development of the supradorsal membrane in P. militaris was very similar
to that described by McEdward (1992, 1995) for P. tesselatus. In P. militaris
each of the original five lobes split to produce a total of ten. The ectoderm of

these ten lobes produced elaborate folds that fused with five additional lobes

from the upper hemisphere or the aboral region. Later the fused lobes were

supported by aboral spines to produce the supradorsal membrane of the
juvenile and adult.
Longitudinal sections of the coeloms of the posterior region are shown
in Figures 7 B, F, and G. In the posterior region the following structures are
visible and appeared to be produced from the archenteron in a pattern similar
to that described in P. tesselatus. The most conspicuous feature in the

posterior region is the large posterior enterocoel that can be seen connected to
the gut (Fig. 7 G). The large posterior enterocoel formed the oral perivisceral







coelom, a major body cavity that encircles the body of the juvenile between

the hydrocoel and aboral body wall (Fig. 7 B, F, G).
The small posterior enterocoel of P. militaris formed from the

posterior end of the archenteron (Fig. 7 B, G) as in P. tesselatus. The fates of

the small posterior enterocoel are likley similar in P. militaris and P.

tesselatus: forming the axial and inner oral perihemal coeloms and the

aboral perivisceral coelom.
The hydrocoel of P. militaris formed as five independent enterocoels

that radiated from the archenteron in a plane transverse to the anterior-

posterior axis of the mesogen (Fig. 7 B, D, E, G, H). This feature of
morphogenesis (identical to P. tesselatus) is one or the major non-larval

features that distinguishes the development of the genus Pteraster from that

of all other asteroids. The five hydrocoel lobes remained connected to the gut

but were otherwise independent of each other during the early development

of the radial canals and podia. Later, the five enterocoels became radial canals

from which rows of podia were produced (Fig. 7 D). After the formation of

two pair of podia, the radial canals were connected by a coelomic tube (ring

canal) that encircled the gut (Fig. 7 E).

The anterior region of the mesogen of P. militaris was lined with an

anterior compartment of the archenteron identical to that of P. tesselatus. In

both species the anterior region completely lacked brachiolar structures and

had no structural fate in the mesogen or juvenile. As argued earlier, the

anterior compartment is simply an extension of the gut and likely functions

to resorb the yolk in the anterior region.

Development of external and internal features of Pteraster acicula

Pteraster acicula has brooded direct development via a mesogen that

lacks any trace of vestigial larval features. This is apparent from external and







internal views of the large anterior region of the mesogen (Fig. 8 A, B, C).

Optical sections of the anterior region of P. acicula show a large anterior
compartment (Fig. 8 B, C) similar to that described for Pteraster tesselatus and
P. militaris. The anterior region and its compartment lacked larval

structures and are resorbed during development of the juvenile (Fig. 8 D).
The posterior region of P. acicula had five marginal bulges that split
into ten and were likely fated to form the supradorsal membrane as described
for Pteraster tesselatus and P. militaris (Fig. 8 A). However, the structure of

the aboral surface of adults of P. acicula was distinct from that of Pteraster

tesselatus and P. militaris. The supradorsal membrane of Pteraster acicula is
extremely thin and transparent (Clark and Downey, 1992). The aboral paxillae

that support the supradorsal membrane have numerous glassy spines at their
tips. These spines interlock to form much of the enclosure that retains

juveniles. The early aboral paxillae in P. acicula are visible in lateral and
aboral views of juveniles (Fig. 8 F).

The posterior region of the early mesogen of P. acicula lacked podia.
This was confirmed by observations of external and internal structure (Fig. 8

A, B). Mesogens of Pteraster tesselatus and Pteraster militaris had a distinct

hydrocoel with at least ten podia at comparable stages of development (i.e.,

five marginal lobes splitting to ten in the posterior region). This indicates

that the development of the water-vascular system in P. acicula developed at
a slow rate relative to the development of other structures. In P. acicula, the

water-vascular system is visible in transverse sections of late mesogens and
juveniles (Fig. 8 C, E) and views of the oral and lateral surfaces of juveniles
(Fig. 8 D, F). The circumoral ring canal of the water-vascular system is visible
in the section through a juvenile of P. acicula (Fig. 8 E). In P. acicula the
circumoral ring canal appeared to form via proximal lateral evaginations that





















Figure 8. Internal and External Morphogenesis of Pteraster acicula.

A. SEM, anterior /oral view of mesogen, showing the anterior region and
the five marginal bulges of the posterior region. B. Longitudinal optical section
of mesogen (posterior of the mesogen is oriented towards the left of the page)
showing the anterior compartment and the large posterior enterocoel. C.
Transverse (slightly oblique) optical section through the anterior region of the
mesogen showing the anterior compartment and a few podia. D. SEM anterior /
oral view of juvenile, showing podia and partially resorbed anterior region. E.
Transverse optical section through the hydrocoel of the juvenile showing the
coelomic linings of the primary podia, secondary podia, terminal podium, and
radial canals of two ambulacura. The circumoral ring canal, some podia, and the
gut are also visible. F. SEM lateral view and aboral view of juvenile showing
podia, supradorsal membrane, and osculum.




















































































































































I1







met and fused in the interadial regions as described above for Pteraster

tesselatus.

Development of external and internal features of Pteraster sp. (cf. P. hastatus)
The morphogenesis of the juvenile of Pteraster sp. (cf. P. hastatus) is

very similar to the juvenile of Pteraster tesselatus and Pteraster militaris.

External views of the juvenile of Pteraster sp. (cf. P. hastatus) show an

anterior region that lacks all vestige of the brachiolar apparatus (Fig. 9 A).

Internal views show that the anterior region is filled by an anterior

compartment, a simple extension of the gut, rather than preoral coeloms that

are typical of larvae (Fig. 9 B, C). Although mesogens were not found, it is

clear that Pteraster sp. (cf. P. hastatus) has brooded direct development. The

anterior region of young juveniles lack brachiolar apparatus in external (Fig.,

9 A) and internal views.
The fully developed water-vascular system of Pteraster sp. (cf. P.

hastatus) is visible in external views and optical sections. In Figure 9 A

ambulacra with two pair of podia and a terminal podium are evident. In

Figures 9 B and C the radial and circular canals of the water-vascular system

of a juvenile with a single pair of podia are visible. Aboral sections of

Pteraster sp. (cf. P. hastatus) show ampullae of podia and the portion of the

oral perivisceral coelom that envelops them. A deep aboral section reveals

the development of the small posterior enterocoel and stone canal (Fig. 9 F).

All of these sections show no significant differences between the

development of the juvenile of Pteraster sp. (cf. P. hastatus) and Pteraster

tesselatus or Pteraster militaris.

In Pteraster sp. (cf. P. hastatus) 10 marginal lobes that form the

supradorsal membrane are evident in external views and sections. The

initial concave shape of the 10 marginal lobes (Fig. 9 B, C, E, F) was similar to
















Figure 9. Internal and External Morphogenesis of Pteraster sp. (cf. P. hastatus).

A. SEM, anterior /oral view of juvenile showing primary podia,
secondary podia and terminal podium in each ambulacrum. Also visible are the
anterior region and ten marginal lobes of the posterior region that form the
supradorsal membrane. B. Transverse superficial optical section through the
hydrocoel showing the coelomic linings of the primary podia and terminal
unpaired podium in an ambulacrum and the gut. Also visible are several
marginal lobes of the posterior region. C. Transverse deep optical section
through the hydrocoel showing the coelomic linings of the primary podia,
terminal unpaired podium, and the radial canals in several ambulacura. Also
visible are the circumoral ring canal, the gut, and several marginal lobes of the
posterior region. D. SEM posterior /aboral view of the juvenile showing 15
marginal lobes of the posterior region that form the supradorsal membrane. E.
Transverse (slightly oblique) deep optical section through the posterior region of
the juvenile showing the ampullae of podia and portions of the oral perivisceral
coelom that surrounds them in three ambulacra. The coelomic linings of the
primary podia, radial canal, and circumoral ring canal are dearly visible in one
ambulacrum. Ten marginal lobes of the posterior region are also visible. F.
Transverse superficial optical section through the posterior region of the juvenile
showing portion of the axial system, the small posterior enterocoel, and the gut.
Portions of the oral perivisceral coelom and ten marginal lobes of the posterior
region are visible.
























A D


B E







16 day old juveniles of Pteraster tesselatus (McEdward, 1992). Late in the

development of the juvenile of Pteraster sp. (cf. P. hastatus) the 10 marginal
lobes and five aboral lobes that form the supradorsal membrane were similar
in shape but more widely spaced than those of Pteraster tesselatus (Fig. 9 D).

Development of external and internal features of Pteraster stellifer
The basic organization of the early juvenile of Pteraster stellifer (Fig. 10

A) was similar to that of the other species of Pteraster. The early juvenile of

Pteraster stellifer was radially symmetrical. The posterior region has bulges

that likely form the supradorsal membrane. The anterior region lacked any
external or internal sign of brachiolar structures (Fig. 10. A, B). However, the
details of the morphology of the early juvenile of P. stellifer are different
from other species of the genus Pteraster. Podia developed between the two
regions however there was not a groove dividing the regions. The podia

were fully exposed and appear to develop rapidly relative to other structures

(i.e., two pair are visible in each ambulacrum in this young juvenile) (Fig. 10
A).

Development of external and internal features of Pteraster sp.

The young juveniles of Pteraster sp. had an anterior region that lacks

any sign of brachiolar apparatus, a posterior region that formed the body of

the juvenile and the supradorsal membrane, and a circumferential groove
contained podia that were positioned between the two regions. A transverse

section of Pteraster sp. shows a pattern of folding of 10 marginal bulges (Fig.

10 C) that is very similar to the pattern described during the formation of the
supradorsal membrane in Pteraster tesselatus (McEdward, 1992). In the same

transverse section of Pteraster sp., a large pair of 1" podia, a smaller pair of 2

podia, and an unpaired terminal podium are visible in each ambulacurum of
the water-vascular system (Fig. 10 C.). In Pteraster sp. the marginal bulges







were arranged such that a pair surrounds each ambulacrum. This

morphology is identical to that described above regarding the completion of

the water-vascular system in an 11 day old juvenile of Pteraster tesselatus

(Fig. 6 M, N).

Development of external and internal features of Pteraster sp. (cf.
Marsipaster)
The early mesogen of Pteraster sp. (cf. Marsipaster) shows a

morphology similar to that described for Pteraster acicula. The external view

of the mesogen of Pteraster sp. (cf. Marsipaster) shows three main regions: a

posterior region that bears marginal bulges, a circumferential groove, and an

anterior region that lacks brachiolar apparatus (Fig. 10 D). A longitudinal

section of this mesogen shows that the interior of the anterior region lacks

any sign of brachiolar structures (Fig. 10 E). The early mesogen of Pteraster

sp. (cf. Marsipaster) had an irregular yet essentially radial pattern of marginal

lobes. This particular specimen was likely fixed while its marginal lobes were

in the process of splitting from five to ten lobes to later form the supradorsal

membrane (Fig. 10 D). The circumferential groove of the early mesogen lacks

podia in an external view (Fig. 10 D). A longitudinal section of this mesogen

shows an early enterocoel between anterior and posterior regions that is likely

a portion of the hydrocoel (Fig. 10 E). This indicates that the development of

the water-vascular system in this species is slow relative to other structures.

The bulk of the coelomic morphology of Pteraster sp. (cf. Marsipaster) was

difficult to decipher from the existing specimens. These mesogens were

among juveniles that were much further developed. One must consider the

possibility that the mesogens are developmentally retarded and thus do not

represent good examples of morphology.



















Figure 10. Various stages of Pteraster stellifer. Pteraster sp., Pteraster sp. (cf.
Marsipaster ), Pteraster temnochiton. and Pteraster pulvillus.

A. SEM, lateral view (the posterior region is oriented towards the left of
the page) of early juvenile of Pteraster stellifer showing two pair of podia in each
ambulacrum. The anterior region and several marginal lobes of the posterior
region are visible. B. Transverse optical section through the anterior lobe of
Pteraster stellifer showing the anterior compartment. C. Transverse optical
section through the hydrocoel of the juvenile of Pteraster sp. showing the
coelomic linings of the primary podia, secondary podia, terminal podium, radial
canals and circumoral ring canal of the water-vascular system. Ten marginal
lobes of the posterior region and the gut are also visible. D. SEM, lateral view
(posterior region is oriented towards the upper left corner of the image) of the
mesogen of Pteraster sp. (cf. Marsipaster) showing the posterior region and its
marginal lobes, and the anterior region. E. Longitudinal optical section through
the mesogen of Pteraster sp. (cf. Marsipaster) showing the anterior compartment
and an early hydrocoel lobe. F. SEM of anterior / oral view of Pteraster
temnochiton showing several pairs of podia in each ambulacrum and the partially
resorbed anterior region. Portions of the supradorsal membrane are also visible.
G. SEM anterior/oral view of Pteraster pulvillus showing two pair of podia and a
terminal unpaired podium in each ambulacrum and the almost completely
resorbed anterior region. Portions of the supradorsal membrane and aboral
paxillae (spines) are also visible.










., B




r C "
_ __
,% *d Alls
^ ^ !*
SL A^JA C







External features of Pteraster temnochiton

A surface view of the juvenile shows two pairs of podia and a terminal

podium in each ambulacrum. The remnant of the anterior region that

remains on the oral side of the juvenile of Pteraster temnochiton was

uniformly shaped and lacks any indication of brachiolar structures (Fig. 10 F).

Thus it is clear that brooded direct development occurs in this species.

Pteraster pulvillus

The oral side of juvenile specimens showed no remnants of the

anterior region of brachiolar apparatus of a larva (Fig. 10 G). However, these

late stage juveniles did not have a significant remnant of the anterior region.

Thus the available specimens did not provide conclusive proof of

development via a mesogen.

Pteraster obscurus

The juveniles I dissected were very old. They ranged in sizes from 6-10

mm in diameter, approximately the same size as those described and sketched

in D'yakonov (1950). These juveniles were fully developed and hence do not

provide any information that suggests or refutes development via a mesogen

or a larva.


Discussion


Reproductive Periodicity of Pteraster miltaris

McClary and Mladenov (1990) report from a multi-year study of the

populations of Pteraster militaris in Passamaquoddy Bay, New Brunswick,

Canada that P. militaris reproduces throughout the year. McClary and

Mladenov (1990) suggest that in P. militaris, new offspring are continually







added to the brood but develop rapidly through the embryonic stages and
become incorporated into an assemblage of juveniles of mixed ages.
Kaufman (1968) observed spawning and dissected early embryos from

the brood chamber of Pteraster militaris collected in Kandalaksha Bay of the

White Sea of Russia in October 1965 and 1966. He reported a sea water

temperature of 2.3"C. Although Kaufman's (1968) interpretation of the

features of the development of P. militaris are not correct (McEdward, 1992)

his sketches are very similar to the mesogens of P. militaris collected from
the Icelandic specimen that is described herein.

I combine the existing data on P. militaris reproduction as follows:

Observations that were made on the specimens of P. militaris obtained from
Iceland, in July clearly demonstrate synchrony among early stages within

broods. Thus a peak period of spawning (June, July) occurs in the early
summer at this location. Brooding of young of the same age is extended over

several months. I collected some extremely large (and thus old) juveniles

were discovered in broods in New Brunswick. In a typical female collected in

New Brunswick, the brood chamber often included a bimodal distribution of

juvenile sizes (e.g., many juveniles 2 mm and few 7 mm in outer diameter,

as measured from arm tip to opposite arm tip). Kaufman's (1968) report and

my own observations have verified that spawning can be induced in P.

militaris throughout the autumn (September, October). Thus a period of

spawning of a few eggs in the autumn may supplement the primary

spawning in early summer, the resulting broods are held together for several

months before mature young are released.







Developmental Habitat of Pteraster militaris

McClary and Mladenov (1990) have reported that eggs are trapped in
the brood chamber of P. militaris but some eggs escape via tears in the
chamber created by the release of earlier juveniles. McClary and Mladenov
(1988) suggest that P. militaris may employ a novel reproductive strategy --
mixed brooding benthicc) and broadcasting (pelagic) modes of development.

Kaufman (1968) reported that "larvae" (actually mesogens) that were
removed from the brood chamber continued to develop normally. I was not
successful in rearing eggs in vitro, thus making the hypothesis that eggs that
are broadcast develop as pelagic larvae suspect. However, because brooded
cultures were also unsuccessful, I believe that the hypothesis of brooding and
broadcasting of young in P. militaris deserves further study, especially during
the early summer (June, July) spawning period.


The Polarity of Developmental Transitions

The evidence presented here is partial support for the hypothesis that

direct development evolved in brooding species of the genus Pteraster and
pelagic direct development evolved subsequently. I present an argument
based on the available data with one caveat. A phylogenetic analysis is

required to corroborate the polarity of developmental transitions (this study
will be conducted as a postdoctoral project).

The evolution of brood-protection
The family Pterasteridae is characterized by a supradorsal membrane
(Perrier, 1875; Blake, 1987) that serves as a protective brood chamber in
almost all species. I confirmed or discovered brooding in nine species of the
genus Pteraster and one species of Hymenaster, another genus of the family







Pterasteridae (presented in Chapters 4 and 5). At least one species of the genus

Pteraster does not brood (Chia, 1966). Adults of P. tesselatus possess the
highly specialized brood chamber characteristic of the family, however P.
tesselatus is a free-spawner that has pelagic direct development. It is unlikely

that a specialized brood chamber would evolve in a lineage with pelagic

development. Hence, I hypothesize that brooding is plesiomorphic among

the family Pterasteridae and that pelagic development is a derived feature

that evolved within the genus Pteraster.

The evolution of direct development
I discovered that seven species of the genus Pteraster brood-protect

direct developing young. Brooded juveniles were discovered in two other

species P. obscurus and P. pulvillus. However, I did not find any stages that

were young enough to allow a confirmation of direct development.

Nevertheless direct development appears to be ubiquitous in the genus. The

ubiquity of direct development among the genus Pteraster supports the

hypothesis that direct development, by itself, is not an adaptation for the re-

evolution of pelagic dispersal as seen in Pteraster tesselatus. As described

above, Pteraster acicula and Pteraster sp. (cf. Marsipaster) have podial

development that is slow relative to other structures in each mesogen.
McEdward (1992, 1995) has argued that rapid development of the podia

(relative to other structures) is a developmental adaptation for the re-

evolution of pelagic life cycle in Pteraster tesselatus. At the end of the pelagic

period, Pteraster tesselatus uses podia for settlement to the sea floor in lieu of

any brachiolar structures. McEdward (1992) described a heterochronic pattern

in which development of podia occurs before arm elongation in Pteraster

tesselatus whereas podia develop after arm elongation in asteroids that settle
to the sea floor as brachiolarian larvae in related taxa such as Henricia. A test







of this hypothesis will be possible via analysis of character evolution. Once a

phylogeny of the Velatida based on characters independent of development is
resolved, developmental characters such as "mesogens that have early

podia", "mesogens that lack early podia", "pelagic development" or benthicc

development" can be mapped onto this phylogeny. For example, species that

brood and have slow podial development might group as a basal clade of

pterasterids and species that have rapid podial development might group as a

derived clade of pterasterids. This example would provide corroboration of

the hypothesis that the evolution of pelagic development from brooding is

correlated with heterochronic acceleration of podial development.












CHAPTER 4
THE DEVELOPMENT OF HYMENASTER PELLUCIDUS, A PTERASTERID
ASTEROID WITH BROODED BRACHIOLARIAN LARVAE


Introduction

Hymenaster pellucidus has indirect development via a brachiolarian

larva. These larvae are brooded in a chamber created by the supradorsal

membrane of the adult. The brooding of brachiolarian larvae is a unique and

previously unknown feature among the family Pterasteridae. In this chapter

I discuss the collection of specimens, describe the internal and external

features of development of larvae and juveniles, and review the previous

studies on reproduction in the genus Hymenaster. I discuss topics specific to

Hymenaster in the concluding sections of this chapter but discuss topics of

significance to the entire study in Chapter 5.

The genus Hymenaster of the family Pterasteridae is one of the few

deep-sea asteroids in which reproduction has been studied. Pain et al. (1982)

inferred continuous reproduction in H. membranaceus from their studies of

seasonal trends in gonad index (the ratio of gonad weight to body weight) and

the frequency of egg sizes within sections of gonad. Continuous

reproduction, the production and release of a small number of mature eggs at

various times throughout the year, is thought to be the dominant

reproductive mode among deep-sea megafauna (Gage and Tyler, 1991). Pain,

et al. (1982) did not find any brooded young in Hymenaster membranaceus

collected from the Rockall Trough of the North-East Atlantic between April

1978 and October 1981. As a result Pain et al. (1982) inferred nonfeeding







demersal (at or near the sea-floor) development in Hymenaster

membranaceus. based on this apparent lack of brooding.
In their systematic monograph of the asteroids of the Atlantic, Clark

and Downey (1992) report that many of the characters historically used by

taxonomists to separate species of Hymenaster are growth-dependent and

preserve poorly. Hence species-level distinction among specimens that differ

in size and preservation method are suspect. Clark and Downey (1992)

determined Hymenaster membranaceus to be synonymous with Hymenaster

pellucidus (W. Thomson, 1873). Thus the specimens collected by Pain et. al.

(1982) are Hymenaster pellucidus.

Sladen (1889), Thorson (1936), and Clark and Downey (1992) reported

brooded young in specimens of Hymenaster pellucidus and other species of

Hymenaster. However, none of these researchers described the morphology

and development of the young. This information was the only data

indicative of brooding among the genus Hymenaster preceding the study

presented in this chapter.


Materials and Methods


Collection of Hymenaster pellucidus

Adult individuals of Hymenaster pellucidus were collected and fixed

in 10% formalin buffered in sodium borate on various cruises of the Benthic

Invertebrates of Icelandic Waters (BIOICE) program in spring and summer

seasons only between May 1991 and August 1994. The sampling occurred

primarily within the Icelandic Economic Zone an area of sea banding the

entire coast of Iceland to 200 km offshore. BIOICE cruises have also collected

along the Faroe-Iceland Ridge that extends beyond the 200 km limit from the







south east of Iceland. The waters around Iceland lie on the border between
the North Atlantic Ocean and the Arctic Ocean. Although Iceland is a small
land mass (103,000 km2) lying just south of the Arctic Circle, hydrographic

conditions vary significantly across small geographic distances and there are

some seasonal differences. South and West Icelandic waters masses originate

in the North Atlantic. These water masses have annual low temperatures of

7C and annual high temperatures of 11C at depths less than 50 meters but

range between annual low temperatures of 5C and annual high

temperatures of 7C at depths of 200-1000 meters. North and East Icelandic

waters are Arctic in character (originate in the Greenland Sea and the

Norwegian Sea). These water masses have annual low temperatures of 0C

and annual high temperatures of 7C at depths to 50 m, but are < 00C year

round in layers deeper than 300 m (Malmberg and Kristmannsson, 1992).


Survey of Hymenaster species in the USNM

I dissected adult specimens of Hymenaster giboryi, Hymenaster

perissonotus, Hymenaster quadrispinosus from the National Museum of

Natural History of the Smithsonian Institution (USNM). I searched the

brood chambers for young and opened the body wall to make an assessment

of the maturity of gonads. Adults of Hymenaster giboryi collected in the

month of February off Cuba contained no young. Hymenaster perissonotus

collected off Washington, Oregon, and central California in the months of

March, June, and July contained mature gonads but no young. However one

adult female collected in May off Unimak Island (Aleutian Islands, Alaska)

contained one late-stage juvenile. Adult females of Hymenaster

quadrispinosus collected off Oregon in the month of May contained very few

individuals with mature gonads but many individuals with late-stage