AMER. ZOOL., 31:707-725 (1991)
Functional Constraints on the Evolution of
Larval Forms of Marine Invertebrates:
Experimental and Comparative Evidence1
R I C H A R D B. EMLET
Department of Biological Sciences, University of Southern California,
Los Angeles, California 90089-0371
SYNOPSIS. Marine invertebrate larvae are well known for their distinctive body shapes and
elaborate patterns of ciliation. In this study I take a physically based approach to investigate
the functional consequences of variations in body shape and patterns of ciliation. With
experimental models I demonstrate that shape as well as surface area contributes to drag of
larval forms. Based on flow fields around larvae tethered in still water and flowing water I
argue that drag, which acts as a partial tether, may influence how water is processed and
food is captured by cilia. With mechanical models of cilia I show that placement of cilia
on the surfaces can influence the effectiveness with which water is moved and the steepness
of the velocity gradient through the ciliary layer. These models indicate that placement of
cilia on ridges, at extreme anterior ends, and at extreme posterior ends of larval bodies
increases the volume of water moved per ciliary stroke relative to placement of cilia on a
flat surface. A comparative survey of 46 larval forms indicates that distributions of body
shape and patterns of ciliation reflect functional requirements of swimming and feeding by
larvae. The experimental and comparative approaches together suggest functional constraints
on the evolution of larval forms which may lead to convergence in patterns of ciliation and
conservation of larval forms within taxa.
INTRODUCTION
Larval forms of marine invertebrates are
perhaps best known for their importance in
studies of phylogenetic relationships among
the metazoa (e.g., Garstang, 1951; Jagersten, 1972; Nielsen and N^rrevang, 1985;
Nielsen, 1987), their bizarre beauty, and
their lack of resemblance to the adult forms
into which they grow. Beginning with the
discovery that microscopic larvae were
actually the early developmental stages of
larger, familiar benthic forms, many distinctive morphological features of larvae
were interpreted as adaptations for planktonic existence (e.g., Balfour, 1881; Garstang, 1928, 1951). There is still much to
be learned about the functional aspects of
microscopic organisms. The purpose of this
paper is (1) to present a novel approach for
examining functional consequences of variations in body shape and patterns of ciliation of microscopic organisms, (2) to evaluate whether invertebrate larvae have sets
of traits that reflect functional constraints,
and (3) to identify general guidelines for
expected morphological changes in body
shapes and patterns ciliation associated with
shifts in developmental mode. I will take a
physically based approach to the study of
larval body shapes. I will use engineering
principles to explore the swimming and
feeding performance of marine invertebrate
larvae and to determine how they interact
with the fluid environment around them.
With experimental approaches I will examine the following two questions: What are
the functional consequences of body shape?
How does placement of cilia on the body
influence water movement? I will follow with
a comparative survey: what is the distribution of body shapes and patterns of ciliation across taxa? I will examine only larvae that use cilia to swim and feed. However
the small size and slow movement of other
planktonic organisms, such as copepods and
protozoa, means the results may apply to
these creatures also.
Larval forms and trophic mode
Variation in larval forms is related to
1
From the Symposium on Experimental Approaches phylogeny, planktonic adaptations, and
to the Analysis of Form and Function presented at the
Centennial Meeting of the American Society of Zool- developmental life histories. The majority
ogists, 27-30 December 1989, at Boston, Massachu- of invertebrate larvae are free swimming
setts.
and small, approximately 1 mm or less in
707
708
RICHARD B. EMLET
length (though some forms can grow to be
longer than 1 cm). Because of their small
size and slow movement (1 to 10 mm/sec),
dispersal of free swimming larvae is largely
due to currents and waves during the hours
to months that larvae spend in the plankton.
Important activities of pelagic larvae include
swimming and, for many, feeding on particulate food such as phytoplankton (Chia
etai, 1984; Strathmann, 1987).
Closely related invertebrates can have
very different larval morphologies that
reflect whether or not larvae feed during
development (Miall, 1897; Strathmann,
1985). Strathmann (1978a, b) evaluated the
evolution and loss of larval feeding by various marine invertebrates and found that
feeding was repeatedly lost in different
groups but only rarely regained. This loss
of feeding ability by larvae was often associated with changes in morphology, including the loss of a functional digestive system,
and changes in body shapes and patterns of
ciliation that resulted in simplified morphologies. These changes have been interpreted (e.g., Strathmann, 1974) as loss of
structures that were no longer required
because of shifts in trophic mode. However,
general guidelines for expected changes that
lead to certain body shapes and patterns of
ciliation have yet to be established.
Cilia, terminology, and definitions
In the following experimental studies, I
will examine fluid movement around
mechanical models of larvae and models of
cilia. Some larval forms have surfaces that
are uniformly ciliated, but most have cilia
tightly packed into distinct rows called ciliated bands. Typical locomotor cilia go
through a power stroke, where the straight
or mildly curved cilium bends at its base,
and this motion carries the fiber though an
arc. During the power stroke, fluid moves
downstream in the direction of the swing.
The power stroke is followed by a recovery
stroke, where a bend in the cilium propagates from the base toward the tip. By moving the tip along a line near the surface of
the body, this movement returns the cilium
to its position at the beginning of the power
stroke. Relatively little fluid is moved
upstream during the recovery stroke. Adja-
cent cilia beat slightly out of phase so that
a metachronal wave travels along the ciliated band as individual cilia go through their
beat cycles. The metachronal wavelength is
the distance along the band between cilia
that have the same beat phase. For more
information on structure and function of
cilia see Sleigh (1984) and references therein.
Drag is a force exerted by a fluid on a
moving object, resisting its movement.
Shear is the slipping of adjacent layers of
water relative to one another. The Reynolds
number (Re) is a standard measure in fluid
mechanics of the ratio (and thus the relative
importance) of inertial and viscous forces
in a flow. Reynolds number is defined as
follows:
Re = UL/v
where C/is the velocity of the fluid or of the
object; L is characteristic length of the object;
and v is the kinematic viscosity of the fluid.
An alternative formulation for cilia is (Blake,
1973):
Re = Ira/v
where the changed terms are /(cilium length),
r (cilium radius), and a (angular velocity of
the cilium).
For a 1 mm long larva swimming at 1
mm/sec, Re = ca. 1. For a cilium 50 nm
long and 0.1 /xm in radius, with an angular
velocity of 100 radians/sec, Re = ca. 10~4.
The important point here is that below Re
= 1, viscous forces become dominant (Vogel,
1981). Because of the small size and relatively slow movement of larvae, the fluid
viscosity dominates the interactions of the
larvae and cilia with the water and with food
particles (Vogel, 1981; LaBarbera, 1984). At
the level of the cilium, viscous forces are
overwhelmingly predominant. This has
several important consequences. (1) As soon
as a force exerted by an organism on a fluid
ceases, all fluid motion stops and so does
the organism (i.e., there is almost no inertia,
see Purcell, 1977). (2) Objects influence
water movement around themselves for
large distances relative to their sizes
(approximately 20 diameters for Re = 1 and
for an increasing relative distance as Re
drops [Vogel, 1981]).
I also used Reynolds number as a scaling
INVERTEBRATE LARVAL FORMS
parameter to generate dynamic similarity
between larvae or cilia and the models of
larvae or cilia. In the present studies, when
the Reynolds number for a model matches
that of the real life situation being modeled,
dynamic similarity assures that theflowfield
around the model will look the same as that
around the real object, despite changes in
size, but depending on the accuracy of the
model. Also, with dynamic similarity
between a model and a real life situation,
the force coefficients will be the same, even
though the absolute magnitudes of forces
may differ (Vogel, 1981).
709
EXPERIMENTAL STUDIES
What are the functional consequences of
variations in body shape?
Larval shapes vary from very simple to
complex with many projections extending
from the body (Fig. 1). These shape differences will influence the drag experienced by
larvae as they swim or sink in the water
column. It is possible to measure the drag
experienced by larval forms by building
models and directly measuring drag. Earlier, I presented a very simplified approach
to predicting the forces exerted during
FIG. 1. A collage of some invertebrate larval forms showing variations in shapes and patterns of ciliation.
Larvae are not drawn to the same scale. Larval forms were redrawn or modified from references cited: A,
protobranch bivalve (after Drew, 1899); B, ophiuroid, nonfeeding (after Grave, 1903); C, gastropod veliger (after
Garstang, 1928); D, polychaete nectochaete (after Blake, 1975a); E, echinoid pluteus (after Strathmann, 1971); F,
articulate brachiopod (after Percival, 1960); G, phoronid actinotroch (after Silen, 1954); H, flatworm (after Jagersten,
1972); I, gastropod veliger (after Dawydofi; 1940a); J, nemertean pilidium (after Dawydoff, 19406); K, sipunculid
pelagosphera (after Jagersten, 1972); L, cnidarian planula (Emlet, personal observation); M, enteropneust hemichordate, nonfeeding (after Burdon-Jones, 1952); N, enteropneust tornaria (after Strathmann and Bonar, 1976);
O, entoproct (after Jagersten, 1972); P, inarticulate brachiopod (after Jagersten, 1972); Q, archeogastropod trochophore (after Kessel, 1964); R, holothuroid auricularia (after Strathmann, 1971); S, ophiuroid pluteus (after Strathmann, 1971).
710
S
RICHARD B. EMLET
40
'2.B30J
fel20
2
5
8
•>
•
g
io
o 0 o 0 \f
B
C D E F
Larval form
FIG. 2. The drag (•) and surface area (O) of models
of simple shapes and selected larval forms (A-H)
[redrawn from Emlet, 1983]. Drag and surface area of
the models are indicated by the diameters of spheres
with equivalent drag and surface area, respectively. At
these low Reynolds numbers {ca. 1), drag on a sphere
is proportional to its diameter. Drag on models was
measured in a flow tank containing circulating, undiluted corn syrup.
swimming by pluteus larvae of sea urchins
and brittle stars (Emlet, 1983). Pluteus larvae (Fig. IE, S) have anteriorly projecting
larval arms which increase in both number
and length during larval development. These
arms bear the rows of cilia that create the
Crassostrea gigas
•-Flow
o-NoFlow
>
swimming and feeding currents (Strathmann, 1971). Brass models of pluteus larvae, 50 times life size, were immersed in a
flow tank with circulating corn syrup, a viscous fluid that was used to match the Reynolds number (ca. 1) between the models and
the real larvae (Emlet, 1983). The models
were constructed such that both number and
length of arms could be manipulated. The
drag on these different models was measured and then compared with the diameter
of spheres with equal drag (Fig. 2; see also
Emlet, 1983). At this low Reynolds number
drag on a sphere is proportional to its diameter (Vogel, 1981).
Changes in body shape will influence the
drag experienced by an organism. Addition
and lengthening of arms increases drag (Fig.
2). There is a four-fold difference in drag
between the larval body without arms and
the larvae with four long arms (Fig. 2D vs.
H). Figure 2 also shows the diameters of
spheres with equivalent surface area to each
of the models. Though surface area increases
with increasing arm number and arm length,
it does not increase as rapidly as drag. Thus
at Re near 1, shape as well as surface area
g Calliostoma ligatum
4.0 x
«-Flow
C Mesochaetopterus taylori
7.0T •—Swimming
o-Self-tethered
O-NOFIOYT
6.0|
4.05.0
3.03.0-
"o
o
4.0
2.0 +
8 3.0
"5
1,20
2.0-
6
I
s
1.0
0
25
50
75
H
100 1.0 0
25
50
75
- 1.0 0
100 125
V
*•••*'
o.'c>
1 H
1
1
1
1
5 10 15 20 25 30 35
Distance
from Body
Distance from Velar Edge (/xm)
Surface (pan)
FIG. 3. The ratio of cilium velocity to particle velocity as a function of distance away from the larval surface
is a measure of the relative movement of cilia to water (from Emlet, 1990). This ratio is higher for the animals
tethered in place (no-flow) than for the same animals swimming freely (flow). The lines are linear regressions
fit to the different data seta: lines in A have different elevations (P < 0.001), lines in C have different slopes (P
< 0.001), and in B the means of ratios are different (P< 0.001) [see Emlet, 1990]. For the first two larvae (A,
B), animals were tethered in a miniature flow tank and flow fields were compared around the organisms in
stationary fluid and in fluid moving past the animals at their approximate swimming speeds. This latter condition
mimicked free swimming. For the polychaete larvae (Q, animals were studied freely swimming and when
tethered by their own mucous strands.
711
INVERTEBRATE LARVAL FORMS
B
FIG. 4. The model ciliated band. Figure 4A is an end view, showing the model in the tank; the point of
observation (P) is perpendicular to this view. Figure 4B shows a view of the ridged model (R) from the observation
point. The model consisted of a single row of 5 steel "cilia" (C), 25 mm in length, 0.37 mm in radius, and
rotated at 0.36 radians/sec. Cilia spaced Vi cilium length (12.5 mm) apart, were mounted on an axle (A) turned
by a gear motor (M) and drive belt (B). During a power stroke, the cilia rotated downward through slits (S) on
the upstream side of a Plexiglas surface (R) and moved in an arc in the working section of the tank before passing
through slits on the downstream side of this surface. Thus the only motion in the working section of the tank
was that of the moving cilia. Dye lines (D), positioned midway between two adjacent cilia, were placed one
cilium length upstream, one cilium length downstream and immediately below the axis of rotation of the cilia.
The model operated in a tank filled with undiluted corn syrup (Karo light, TM), with a dynamic viscosity of 5
to 10 kg/m sec. Reynolds number ranged from 1 x 10~3 to 5 x 10"4, comparable to that of real larval cilia. The
model was positioned in a glass aquarium (40 cm long, 20 cm wide, and 20 cm deep). The clearance between the
cilia tips and the tank bottom was a 5 cilium lengths (not drawn to scale).
influences the magnitude of the drag (Emlet,
1983). Increased drag will translate directly
into an increased cost of transport (energy
consumed per unit distance covered) and a
reduced sinking rate.
Drag on a swimming larval form may have
other effects. When swimming, an animal's
cilia sweep through the water and move
some water along with them during their
power strokes; as a consequence, the animal
moves forward. Because of viscous forces
the fluid is sheared, and a velocity gradient
is established out from the moving cilium
during the power stroke. Water near the cilium is moving at or close to the speed of
the cilium, water farther away moves less
rapidly (Sleigh, 1984). Emlet and Strathmann (1985) argued that drag acts as a
retarding or tethering force which resists
movement of the animal relative to the
water, and in doing so increases the relative
movement of cilia through the fluid. Negative buoyancy can have the same effect
(Strickler, 1982). A rough calculation of the
body length at which the tethering effects of
buoyancy and drag are equal is 0.6 mm for
animals shaped like prolate spheroids (length
to width ratio, 3:1), with an excess density
relative to sea water of 0.03 g/cm3, and a
swimming speed of 1 mm/sec (Emlet and
Strathmann, 1985). For animals with these
characteristics and smaller than 0.6 mm,
drag would yield a greater retarding effect
than negative buoyancy.
I (Emlet, 1987, 1990) have examined
experimentally the effects of tethering on the
712
RICHARD B. EMLET
FIG. 5. Schematic line drawing of four model surfaces. The oblique, heavy lines represent cilia and the
dashed lines show the arcs through which the cilia
traveled. Data were collected by double exposure photographs that showed displacement of dye lines during
a 180 degree power stroke. The power strokes always
started from and ended in horizontal positions (3 o'clock
and 9 o'clock). This convention insured cilia moved
the same amount for each model and that dye lines
were always placed in similar positions relative to cilia
at the beginning of the power stroke. The three parallel
vertical lines below each surface are dye lines at the
start of the power stroke, and the curved lines are the
same dye lines at the end of the power stroke. For the
anterior and posterior models, an additional Plexiglas
plate (S) was mounted in the plane of the air-syrup
interface.
flow fields animals produce. Results from 3
different ciliated larvae, Crassostrea gigas
(a bivalve), Calliostoma ligatum (a gastro-
pod), and Mesochaetopterus taylori (a
polychaete) show that animals held stationary experience greater shear of water past
cilia than do these same animals under conditions that mimic free swimming (Fig. 3).
These experimental results represent the
extreme case where animals were completely tethered. For animals swimming
freely, the drag on their bodies can be
expected to act as a partial tether and to
cause increased movement of cilia relative
to water. A number of organisms capture
food particles by direct interception, where
cilia or setous appendages contact a particle.
Enhanced shear of cilia to water may
enhance the probability of particle capture.
How does the placement of cilia on the
body influence water movement?
Under the physical conditions in which
cilia function, viscous interactions dominate. Stationary surfaces retard fluid movement and moving surfaces move fluid; in
both cases the surfaces influence fluid
movement at a distance. The placement of
cilia on larval bodies should influence how
effectively they move water. To examine
this prediction empirically, I constructed a
set of mechanical models of ciliated bands
with four different surface arrangements
observed on marine invertebrate larvae.
Larvae have ciliated bands either on (1) flat
surfaces or on edges of their bodies. When
on an edge, the ciliated bands can be (2)
elevated on a ridge, (3) located at the extreme
anterior end, or (4) located at the extreme
posterior end of the larval body. For some
larvae, a combination of these arrangements occurs: e.g., velar cilia of gastropod
and bivalve larvae are located at the extreme
anterior ends and raised on ridges called
velar lobes (Fig. 1C, I).
The model ciliated band consisted of five
steel "cilia," spaced lh cilium length apart,
and a Plexiglas "larval surface" (Fig. 4).
Model cilia were much farther apart than
adjacent cilia in an actual ciliated band.
Because all model cilia had the same phase
of beat during the power stroke, the spacing
represents the metachronal wavelength.
Metachronal wavelengths measured for
velar cilia in several species of gastropod
larvae varied from V6 to over Vi the length
of cilia (Emlet, unpublished data). The
Plexiglas surface and frame were adjustable
so that three different surface configurations
could be formed at the base of the cilia: a
flat surface, a ridged surface (with angles off
horizontal of 25 degrees), and a right-angled
surface (Fig. 5). By reversing the direction
of ciliary movement, the right-angled surface could be used for both the anterior and
posterior ciliated band. For the anterior and
posterior models, an additional Plexiglas
plate (Fig. 5S) was mounted in the plane of
the air-syrup interface. This additional plate
acted as a wall that minimized fluid movement in the region near the air-syrup interface. The addition of this plate increases the
robustness of the comparison of fluid move-
713
INVERTEBRATE LARVAL FORMS
Flat (•)
8 e
4.0
4.0
35
3.5
u u
w3 0
s*
Q..C1
Ridge (O)
| |
25
20
3 0
1 £2.5
20
1 5
15
Posterior ( T )
Anterior
to
2.0
25
3.0
35
Displacement Area of
Flat Model (cm2)
40
ooJH
a »
2.0
I Q 12
6
1.5
15
2.0
2.5
30
35
40
DOWNSTREAM Displacement
Area (cm^)
45
10
5
10
15
20
25
d — Distance from Base
of Cilia (mm)
FIG. 6 Displacement offluidby four different surface configurations of model ciliated bands. All reported areas
of displacement were adjusted by a magnification factor to give the actual area or distance of displacement. A.
The mean areas displaced during a power stroke of 180 degrees for ridged (O), anterior (A), and posterior (T)
models are each compared to that displaced by a similar power stroke for the flat surface model. The mean area
displaced is upstream + central + downstream areas/3. The diagonal line represents equality between compared
areas. Mean displacement of each model was compared to that of a flat model with Wilcoxon's signed-ranks
test (Sokal and Rohlf, 1981), assuming a priori that cilia on the flat model moved less fluid (1-sided test): T =
0 and P < 0.05 for each comparison). B. The area displaced upstream compared to area displaced downstream
for each model surface. Symbols are the same as in A but include (•) for a flat surface model. The diagonal line
represents equality between compared areas. Wilcoxon's signed-ranks tests between upstream and downstream
displacement areas for each of the models were as follows (1-sided tests): flat, T = 9, P = 0.25; ridge, T = 10,
P > 0.25; anterior, T = 0, P < 0.05; posterior, T = 0, P < 0.05. For A and B all comparisons are made within
trials. The areas enclosed by dye lines were measured by digitizing the area or by cutting out and weighing their
traces. The mean of 3 replicate measurements for each displacement was used. Measurement error was < 5%.
C. Location of the peak displacement for central dye line. The magnitude (w) is distance displaced downstream.
Distance from the axis of rotation (d) is measured vertically from the base of cilia (see insert). Values in Figure
6C were not compared statistically; only trends are discussed.
ment around the band on a flat surface to
that around anterior and posterior bands.
To visualize the movement of fluid, vertical lines of dyed corn syrup were injected
through small holes intermediate between
adjacent cilia (Figs. 4, 5). The displacement
of dye lines was measured as the area
between positions of a single dye line before
and after a single 180 degree power stroke
(Fig. 5). This area is approximately proportional to the volume of fluid displaced
along this plane in the downstream direction. Due to the placement of dye lines
intermediate between adjacent cilia, the displacement area measured is the minimum
displacement of any position along the ciliated band. My models had only a power
stroke for the model cilia. (Later models
have incorporated a model of a recovery
stroke that works during the power stroke,
and results indicate that effects of the recovery stroke are localized near the surface
[Emlet, unpublished data].)
Models were compared in three ways. (1)
The mean area of displacement of all three
dye lines was compared between models
(Fig. 6A). (2) The areas of displacement of
the upstream and downstream lines were
compared to each other for all the models
(Fig. 6B). (3) The points of maximum displacement of the central dye line of all surfaces were compared to one another (Fig.
6C). Replications of experiments consisted
of replicating trials in which double exposure photographs were taken for each model.
Comparing mean area of displacement,
cilia on ridged or right-angled surfaces
moved more fluid per stroke than cilia on
flat surfaces (Fig. 6A). Within each trial and
without exception, the displacement area of
the central dye line (not figured) and the
mean displacement area of the 3 dye lines
were lower for the flat surface than for the
other surfaces (Emlet, 1985). The ridged
model and the posterior model had the
greatest mean displacement (Fig. 6A).
714
RICHARD B. EMLET
Comparison of displacement areas of the
upstream and downstream dye lines demonstrates the interactions of cilia with different surface configurations within a power
stroke. The data for anterior and posterior
model bands show how effectively surfaces
retard fluid movement (Fig. 6B). For these
models the greatest displacement area was
for the upstream line of the anterior band
and for the downstream line of the posterior
band. A difference in areas of displacement
upstream and downstream implies that
movement of cilia relative to the fluid
changed over the power stroke. For the
anterior band, ciliary movement relative to
the fluid was greatest at the end of the power
stroke and for the posterior band ciliary
movement relative to the fluid was greatest
at the beginning of the power stroke. Data
for the flat and ridged surfaces fell on both
sides of the diagonal line (Fig. 6B). Because
these surfaces were symmetrical about a
plane passing through the axis of rotation
and the central dye line, the displacement
areas of upstream and downstream lines
should have been more similar for these
models than for lines of the right-angled surfaces. The variation of data for the ridged
and flat models to both sides of the diagonal
line is probably due to differences in placement of the dye lines prior to the power
stroke.
Because of its location, the central dye
line marked fluid through which the cilia
pass during the power stroke. It is fluid in
this region that may be cleared of food particles by some bands of cilia. For this reason
the location of the peak displacement is a
relevant comparison to make between the
model surfaces. The magnitude and location of the peak displacement of the central
dye line showed distinctive patterns for each
type of model surface (Fig. 6C). The maximum displacement was least for the flat
surface. In order of increasing distance from
the axis of rotation and decreasing magnitude, the four surfaces rank as follows: posterior, ridged, anterior, and flat (Fig. 6C).
These patterns reflect the influence of adjacent surfaces in the development of the
velocity profile and are consistent with the
results presented above for the influence of
surfaces on displacement of the outer dye
lines. The differences in distances of peak
displacement from axis of rotation (d)
between the anterior and posterior models
are probably less than the models indicate;
the placement of the central dye line biases
the posterior model toward reduced distances (d), but also suggests that the best
geometry to reduce d would be a ciliated
band on a sharp ridge, such as the velar edge
of mollusk larvae (Fig. 1C, I).
Summary of experimental studies on
body shape and cilia
The above studies suggest that both body
shape and patterns of ciliation can influence
fluid movement and therefore performance
of both feeding and swimming. Body shape
influences drag which can affect sinking rates
and cost of locomotion. Ciliated bands on
ridges and at the anterior and posterior ends
of larval forms are more effective in moving
water than if placed on a flat surface. Placement of cilia in these locations may also
enhance particle capture by increasing the
steepness of the velocity gradient out from
the surface of the body. A steeper shear gradient would bring particles closer to the body
in the region of the ciliated band and in this
way might increase the probability of particle capture. For cilia placed at anterior or
posterior locations, the shear of water by
cilia changes within the power stroke. For
an anteriorly located band, cilia move more
water during the first part of the stroke and
in the second part they may overtake particles in the water. Bands on ridges have an
overall effect of decreasing shear of water
relative to cilia and thus may reduce particle
capture but increase swimming effectiveness. Larval shape can also modify shear of
water relative to cilia; all other things being
equal, simple shapes have relatively lower
form drag and thus cilia should shear less
through surrounding water.
COMPARATIVE SURVEY
What are the distributions of body form
and patterns of ciliation among
invertebrate larvae?
A survey of invertebrate larvae was carried out to determine whether invertebrate
larvae have body shapes and patterns of ciliation that may enhance swimming or feed-
INVERTEBRATE LARVAL FORMS
ing performance. Fifteen marine invertebrate phyla were surveyed (Table 1,
Appendix) and forms were identified by a
number of criteria including level of phylogenetic representation (usually taxonomic
class), their characteristic body shapes and
patterns of ciliation. A total of 46 larval
forms are included in the survey, and most
of these bear distinct names assigned by
morphologists. These forms have not
evolved entirely independently, but are
either morphologically distinct or have been
separate from phylogenetically and morphologically related forms since the origin
of the taxon. For example, trochophore larvae (Fig. 1Q) occur in polychaetes, sipunculids, echiurans, and mollusks, and their
common occurrence is considered to be evidence that these groups share a common
ancestor. These phyla have been separate at
least since the early Paleozoic and, therefore, their larval forms have had a considerable amount of time to diverge morphologically. Within phyla and classes,
nonfeeding larval development has evolved
a number of times from feeding larval
development (Strathmann 1978a, b) and
often involves morphological changes.
These morphological changes are evidence
that larval forms can change and thus are
justification for counting (as separate) the
phylogenetically related forms from separate phyla. In the survey, feeding and nonfeeding forms that were morphologically
distinct were also counted. Morphologically
distinct nonfeeding forms within a phylum
were counted but each separate, nonfeeding
lineage was not counted.
Each larval form was scored for body
shape, pattern of ciliation, trophic mode,
and ciliary band location (if present). Body
shape was categorized as simple or complex
by comparing larval forms within phyla to
one another and to the mechanical criteria
of high drag and low drag shapes. Complex
shapes are those with projections from their
bodies or with high surface areas that contribute to high drag during locomotion.
Simple shapes are those that approximate
prolate and oblate spheroids or cylinders
and typically have low surface area and low
drag (see Appendix for further explanation).
At low Reynolds numbers the geometry with
the lowest dray relative to volume is a pro-
715
late spheroid with a length to diameter ratio
of 2:1 moving parallel to its axis of rotation
(Vogel, 1981). Categorization as simple or
complex is not meant to imply structural or
cellular properties. Patterns of ciliation were
categorized according to whether cilia were
uniformly distributed over the larval surface or collected into distinct ciliated bands.
Several larval forms with body surfaces generally covered with cilia but also with longer
cilia in discrete rows that function as ciliated
bands were scored as banded only {e.g.,
Muller's larvae of flatworms [Fig. 1H], larvae of articulate brachiopods [Fig. IF]).
Trophic mode was based on whether larval
forms feed on paniculate food (Feeding) or
not (Nonfeeding). When a particular larval
form had both feeding and nonfeeding representatives, each was counted. Larvae with
ciliated bands were scored for whether these
bands were located in positions to enhance
water movement. As indicated above, ciliated bands on ridges and at the anterior and
posterior ends of bodies enhance fluid
movement relative to ciliated bands on flat
surfaces. If a larval form had representatives
with and without bands on edges, each pattern was counted once.
Contingency table analyses of data in
Table 1 were conducted (Table 2) to evaluate whether associations existed between
the following set of variables: (A) larval ciliation pattern and body shape; (B) larval
trophic mode, body shape and ciliation; (C)
body shape and location of ciliated bands
(only for larvae with ciliated bands). To further analyze part of the results from the preceding comparison, I also evaluated (D) larval trophic mode and location of ciliated
bands for larvae with simple shapes.
Body shape and patterns of ciliation are
not randomly associated (Table 2A). Larvae
with complex body shapes always have cilia
arranged in bands. Cilia restricted to bands
are the most common pattern and are found
in approximately equal numbers of larvae
with complex and simple body shapes; uniformly ciliated larvae always have simple
body shapes (Table 2A). Feeding larvae
more commonly have complex shapes than
simple shapes. With a single exception,
feeding forms, whether complexly or simply
shaped, have bands of cilia (Table 2B). The
exceptional feeding larval form is uniformly
716
RICHARD B. EMLET
TABLE 1. Ciliated larval types and categories ofbody shape, patterns ofdliation, trophic mode, and band location.*
Larval type
Porifera
Amphiblastula
Parenchymula
Parenchymula
Cnidaria
Planula
Platyhelminthes
Muller's larva
Direct
Nemertea
Pilidium
Iwata's & Direct
Polychaeta
Trochophore
Nectochaete
Rostraria
Mitraria
L-metatrochophore
Echiura
Sipuncula
Trochophore
Pelagosphera
Pogonophora
Entoprocta
Mollusca
Veliger (Gastropod)
Rotiger (Bivalve)
Scaphopod
Protobranch
Chiton
Aplacophoran
L-veliger
L-rotiger
Bryozoa
Cyphonautes
Coronate
Other
Brachiopoda
Articulate
Inarticulate
Phoronida
Actinotrocha
L-larva
Hemichordata
Tornaria
L-larva
Pterobranch
Echinodermata
Bipinnaria (A)
Echinopluteus (E)
Ophiopluteus (O)
Band locations on
edges (enhance water
movement) (Y = Yes,
N = No)
Body shape
(S = Simple,
C = Complex)
Ciliation pattern
(U = Uniform,
B = Bands)
S
S
U
U
B
N
N
N
Y
s
U
F,N
—
c
s
B
U
?
N
Y
—
c
s
B
U
F
N
Y
—
s
s
c
c
s
s
B
B
B
B
B
B
F,N
F
F
F
N
F
Y
Y
Y
Y
Y
Y
s
c
s
s
B
B
B
B
N
F,N
N
F,N
N
Y
Y
Y
c
c
c
s
s
s
c
c
B
B
B
B
B
B
B
B
F
F
N
N
N
N
N
Y
Y
Y
N
N
Y/N
Y
Y
s
s
s
B
B
U
F
N
N
Y
Y
—
s
B
B
N
F
Y
Y
c
s
B
U
F
N
Y
—
c
s
s
B
B
U
F
N
N
Y
N
—
c
B
B
B
F
F
F
Y
Y
Y
s
c
c
c
Trophic mode
(F = Feeding,
N = Nonfeedmg)
?
—
717
INVERTEBRATE LARVAL FORMS
TABLE 1. Continued.
Larval type
Auricularia (H)
L-Asteroid
L-Echinoid
L-Ophiuroid
L-Holothurian
L-Holothurian
L-Crinoid
Body shape
(S = Simple,
C = Complex)
Ciliation pattern
(U = Uniform,
B = Bands)
c
s
s
s
s
s
s
B
U
U
B
B
U
B
Trophic mode
(F = Feeding,
N = Nonfeeding)
F
N
N
N
N
N
N
Band locations on
edges (enhance water
movement) (Y = Yes,
N = No)
Y
Y
Y, N
N
* See APPENDIX for a fuller explanation of characters chosen for this table. A larval type beginning with
"L-" indicates a nonfeeding form of the group.
ciliated planula larvae of several Cnidaria.
These larvae feed by trailing a strand of
mucus to which particles adhere (Trantor et
ai, 1982; see also Strathmann, 1987); the
mucous strand is subsequently drawn into
the planula's mouth. Complex forms always
have ciliated bands in locations to enhance
fluid movement. However, for simply
shaped larvae, ciliated bands may or may
not be arranged to enhance fluid movement
(Table 2C). When larvae with simple shapes
and bands of cilia were examined for an
association between band location and trophic mode, feeding larvae always had bands
in a location to enhance water movement,
but nonfeeding forms were split such that
some had bands located to enhance fluid
movement and others did not (Table 2D).
These comparisons reveal that there is a
tendency for feeding larvae to have complex
body shapes and a strong tendency for nonfeeding larvae to have simple body shapes.
There appears to be a tradeoff of body shape
(and cost of locomotion) that is associated
with larval trophic mode. The increased
energetic expenditure during locomotion due
to the high drag of complex body shapes
may have little cost for planktonic larvae,
dependent of currents for transport, that feed
while swimming. Because larvae are subject
to currents for horizontal transport, locomotory expenditures may be devoted to
depth regulation and moving water to catch
food particles. If complex body shapes
enhance particle feeding by influencing the
flow field, these costs may be largely offset.
The spatial arrangement of cilia {i.e., tight
bends in the ciliated bands) may also change
the local flow conditions near cilia (Strathmann, 1988; Emlet, 1990). In contrast, the
tendency for nonfeeding forms to have simple shapes that experience lower drag may
reflect requirements for mechanical effectiveness in forms that lack the ability to feed
on paniculate materials. Even though these
nonfeeding forms are also subject to currents for most of their transport, restrictions
on energy reserves may lead to selection for
low drag shapes because of their reduced
energetic costs of locomotion.
With the one exception already noted,
feeding larvae have bands of cilia. This
observation suggests that bands of cilia are
required for feeding on paniculate food.
Larvae with ciliated bands have several
methods of feeding (Strathmann, 1987); all
involve interactions of cilia with particles
moving through the ciliary band. Band cilia
may capture particles by direct interception
or may redirect a particle to a trajectory that
results in capture by other cilia (Strathmann, 1987). Feeding larvae always have
ciliated bands in a position to enhance water
movement. Because this pattern occurs in
larvae with both complex and simple shapes,
the increased flux of water through the ciliated region appears to be of prime importance. Increased flux may enhance feeding
by bringing more particles through the ciliated region where they can be caught. A
complex body shape may partially counteract the effects of placing cilia on edges
because it should increase the shear of water
relative to cilia, leading to a reduced flux
through the ciliary layer. For animals that
depend on direct interception increased
718
RICHARD B. EMLET
TABLE 2. Associations of body shape, patterns ofciliation, trophic mode, and band pattern; contingency tables
for data from Table 1*
A. 2 x 2 contingency table
Body shape
Complex
Ciliation
Patterns
16
(12.2)
0
Uniform
(3.8)
16
Total
G = 10.59, significant = P < 0.005
Bands
B. 2 x 3 contingency table
Simple
Total
19
(22.8)
11
(7.2)
30
35
11
46
Larval trophic mode
Feeding
Nonfeeding
Total
13
(6.3)
5
(7.9)
1
(4.8)
19
3
(9.7)
15
(12.1)
11
(7.3)
29
16
Ciliation
Complex body
Bands
Simple body
Uniform
Simple body
Total
20
12
48
G = 18.75, significant at P < 0.001
C. 2 x 2 contingency table
Band locations on edges
Yes
No
Total
Banded forms only
16
(12.6)
14
Simple
(17.4)
30
Total
G = 9.62, significant at P < 0.005.
Complex
Body
Shape
D. 2 x 2 contingency table
0
(3.4)
8
(4.6)
8
22
38
Band locations on edges
Yes
Banded forms only
Simple
Body shape
16
Feeding
5
(3.6)
Nonfeeding
10
(11.4)
Total
15
G = 3.47, not significant, 0.05 < P < 0.1.
No
0
(1.4)
6
(4.6)
6
Total
5
16
21
* Numbers in parentheses indicate the expected number if patterns ofciliation and body shape are randomly
distributed. For all tables, a G-test of independence with the Williams' correction for small sample sizes was
used (Sokal and Rohlf, 1981, p. 736).
shear of water relative to cilia may enhance
particle capture. In addition, the placement
of cilia on edges increases the steepness of
the shear gradient relative to the body in
the ciliated layer and this may persist despite
the high drag body shape. The interactions
of body shape and placement of ciliated
bands on larvae with different feeding
mechanisms require more empirical investigation for further interpretation.
Nonfeeding forms can either have bands
of cilia or be uniformly ciliated. Of those
which have ciliated bands, these bands may
or may not be in locations to enhance water
movement. A simple body shape with bands
of cilia on ridges is an effective morphology
for swimming. The existence of some simple shaped, nonfeeding larvae with cilia that
are not elevated to enhance water movement may imply that increased flux is not
INVERTEBRATE LARVAL FORMS
important energetically. Other developmental or structural factors may also prevent formation of elevated bands (e.g., dense
yolk granules in cells may restrict formation
of ridges). Because many nonfeeding forms
have evolved from feeding forms that always
have bands on ridges, some nonfeeding
forms may have retained the ancestral condition of elevated bands. Similarly the few
nonfeeding forms with complex shapes may
also have retained this form from an ancestor with feeding larvae. Why some nonfeeding larvae are uniformly ciliated whereas
others are banded cannot be determined
from the present analysis. Though several
models of fluid propulsion and mechanical
efficiency have been developed for uniformly ciliated surfaces (Blake, 1973; Keller
et al, 1975; Keller and Wu, 1977) comparable models for simple shapes with cilia in
bands are not yet available
CONCLUSIONS
719
swimming performance (and may enhance
feeding performance) in their microscopic
world. I have provided a rationale for
expecting functional convergence among
larval forms. For instance, some of the posteriorly located bands of cilia ("telotrochs")
commonly found among larvae probably
represent functional convergence. In addition, given the constraints of the physical
conditions and the genetic and cellular
structure of the organisms examined, I
believe I have provided a rationale that may
explain evolutionary conservatism of larval
forms. Within a taxon, functional requirements probably maintain feeding larval
forms as variations of complex body shapes
with ciliated bands on ridges and maintain
nonfeeding larvae as simple forms that can
become evenly ciliated. Furthermore, when
larval life histories change, the physical constraints indicate the directions and limits of
permissible change because functional
requirements are maintained. For example,
when larval feeding structures are lost, the
simplification of body shape is in the direction of reduced drag (and probably reduced
cost of locomotion).
Experimental analyses of body shape and
patterns of ciliation indicate that both of
these morphological features have consequences for movement of fluid and feeding
by marine invertebrate larvae. ComparaACKNOWLEDGMENTS
tive analyses of form indicate that body
I am grateful to the organizers of the symshapes and patterns of ciliation are not randomly distributed with respect to each other posium for the opportunity to present this
or with respect to trophic mode of larvae. information and for the impetus to write
Either an experimental or a comparative this analysis. I sincerely thank the many
approach by itself is not sufficient to eval- people who have contributed to the analyses
uate the effects that functioning in a viscous presented here or provided constructive
environment has had on the evolution of criticism of the manuscript, most notably
larval forms among marine invertebrates. R. Strathmann, M. Koehl, T. Daniel, J.
However, when combined, the results of the Hanken, T. Mace, J. Morin, R. Zimmer, the
experimental and comparative analyses U.C.L.A. Seminar in Invertebrate Zoology,
imply that the functions of feeding and and two anonymous reviewers. Any
swimming are probably important in lim- remaining errors are my own. I also thank
iting the possible body shapes and arrange- T. Mace for preparing the figures. This
ments of cilia among these larvae. The asso- research was supported by NSF grants OCEciations of body shape with trophic mode, 8400818 to R. Strathmann and OCEpattern of ciliation with trophic mode, and 8510834 to M. A. R. Koehl. Symposium
pattern of ciliation with body shape can be support was provided by a National Science
understood in the context of the hydro- Foundation grant (BSR-8904370) to J.
mechanical consequences that these mor- Hanken and M. H. Wake.
phological features have for the performance of swimming and feeding.
REFERENCES
The combination of mechanical and Atkins, D. 1955. The ciliary feeding mechanism of
comparative evidence indicates that larval
the cyphonautes larvae (Polyzoa Ectoprocta). J.
Mar. Biol. Assoc. U.K. 34:451^166.
forms possess characteristics that enhance
720
RICHARD B. EMLET
Baba, K. 1938. The later development of a solenogastre, Epimenia verrucosa (Nierstrasz). J. Dept.
Agric. Kyushu Imp. Univ. 6:21-40.
Bakke, T. 1974. Settling of the larvae of Siboglinum
fiordicum Webb (Pogonophora) in the laboratory.
Sarsia 56:57-70.
Balfour, F. M. 1881. ^4 treatise on comparative embryology, Vol. 2. Macmillan, London.
Bergquist, P. R., M. E. Sinclair, C. R. Green, and H.
Silyn-Roberts. 1979. Comparative morphology
and behaviour of larvae of Demospongiae. Colloques Int. C. Nl. R. S. 291:103-111.
Bhaud, M. and C. Cazaux. 1982. Les larves de polychetes des cdtes de France. Oceanis 8:57-160.
Birkeland, C. F., F.-S. Chia, and R. R. Strathmann.
1971. Development, substratum selection, delay
of metamorphosis and growth in the seastar Mediaster aequalis Stimpson. Biol. Bull. 141:99-108.
Blake, J. A. 1975a. The larval development of Polychaeta from the northern California coast. III.
Eighteen species of Errantia. Ophelia 14:23-84.
Blake, J. A. 19756. The larval development of Polychaeta from the northern California coast. I. Cirriformia spirabrancha (family Cirratulidae). Trans.
Amer. Microsc. Soc. 94:179-188.
Blake, J. A. 1975c. The larval development of Polychaeta from the northern California coast. II.
Nothria elegans (family Onuphidae). Ophelia 13:
43-61.
Blake, J. R. 1973. A finite model for ciliated microorganisms. J. Biomech. 6:133-140.
Burdon-Jones, C. 1952. Development and biology of
the larva of Saccoglossus horsti (Enteropneusta).
Phil. Trans. Roy. Soc. London B 236:553-589.
Bury, H. 1888. The early stages in the development
of Antedon rosacea. Phil. Trans. Roy. Soc. London
179:257-301.
Cantell, C.-E. 1966. Some developmental stages of
the peculiar nemertean larva Pilidium recurvatum
Fewkes from the Gullmarfjord (Sweden). Ark.
Zool., Ser. 2, 19:143-147.
Chia, F.-S., J Buckland-Nicks, and C. M. Young. 1984.
Locomotion of marine invertebrate larvae: A
review. Can. J. Zool. 62:1205-1222.
Chuang, S.-H. 1977. Larval development of Discinisca (inarticulate brachiopod). Amer. Zool. 17:
39-53.
Coe, W. R. 1899. Development of the pilidium of
certain nemerteans. Trans. Conn. Acad. Arts Sci.
10:235.
Conklin, E. G. 1908. Two peculiar actinian larvae
from Tortugas, Florida. Carnegie Inst. Washington 103:173-186.
Dawydoff, C. 1940a. Quelques veligers geantes de
Prosobranches provenant de la Mer de Chine. Bull.
Biol. Fr. Belg. 74:497-508.
Dawydoff, C. 19406. Les formes larvaires de Polyclades et de Nemertes du plancton Indochinois.
Bull. Biol. Fr. Belg. 74:443-496.
Dawydoff,C. 1959. Classe des Echiurens. In P. Grasse
(ed.), Traiti de zoologie, pp. 885-907. Masson et
Cie, Paris.
Dilly, P. N. 1973. The larva of Rhabdopleura compacta (Hemichordata). Mar. Biol. 18:69-86.
Drew, G. A. 1899. Some observations on the habits,
anatomy and embryology of members of the Protobranchia. Anat. Anz. 15:493-519.
Emlet, R. B. 1983. Locomotion, drag, and the rigid
skeleton of larval echinoderms. Biol. Bull. 164:
433^145.
Emlet, R. B. 1985. Functional morphology and ecology of larvae of clypeasteroid echinoderms and
other ciliated larvae. Unpubl. Ph.D. Diss., University of Washington, Seattle. 154 pp.
Emlet, R. B. 1987. Flow fields around ciliated larvae:
The effects of natural and artificial tethers. Amer.
Zool. 27:13A.
Emlet, R. B. 1990. Flow fields around ciliated larvae:
The effects of natural and artificial tethers. Mar.
Ecol. Prog. Ser. 63:211-225.
Emlet, R. B. and R. R. Strathmann. 1985. Gravity,
drag, and feeding currents of small zooplankton.
Science 228:1016-1017.
Fell, P. E. 1974. Porifera. In A. C. Giese and J. S.
Pearse (eds.), Reproduction of marine invertebrates, Vol. 1, pp. 51-132. Academic Press, New
York.
Fenaux, L. 1963. Note preliminaire sur le developpement larvaire de Amphiura chiaje (Forbes).
Vie Milieu 14:41-46.
Garstang, W. 1928. The origin and evolution of larval forms. Rep. Br. Assoc. Advance. Sci. Sect. D:
77-98.
Garstang, W. 1951. Larvalforms and other zoological
verses. Blackwell, Oxford. (Paperback reprint 1985,
University of Chicago Press, Chicago.)
Gemmill.J. F. 1912. The development of the starfish
Solaster endeca Forbes. Tran. Zool. Soc. London
20:1-58.
Grave, C. 1903. On the occurrence among echinoderms of larvae with cilia arranged in transverse
rings, with a suggestion as to their significance.
Biol. Bull. 5:169-186.
Hadfield, M. G. 1975. Hemichordata. In A. C. Giese
and J. S. Pearse (eds.), Reproduction of marine
invertebrates. Vol. 2, pp. 185-240. Academic Press,
New York.
Hatschek, B. 1880. Uber Entwicklungsgechichte von
Echiurus und die systematische Stellung der
Echiuridae. Arb. Zool. Inst. Univ. Wien 3:45-78.
Hendler, G. 1975. Adaptational significance of the
patterns of ophiuroid development. Amer. Zool.
15:691-715.
Iwata, F. 1958. On the development of the nemertean
Micrura akkeshiensis. Embryologia 4:103-131.
Jagersten, G. 1972. Evolution of the metazoan life
cycle. Academic Press, London.
Keller, S. R. and T. Y. Wu. 1977. A porous prolatespheroidal model for ciliated microorganisms. J.
Fluid Mech. 80:259-278.
Keller, S. R., T. Y. Wu, and C. Brennen. 1975. A
traction-layer model for ciliary propulsion. In T.
Y. Wu and C. Brokaw (eds.), Swimming and flying
in nature, Vol. 1, pp. 253-272. Plenum Press, New
York.
Kessel, M. M. 1964. Reproduction and larval development of Acmaea testudinalis (Mttller). Biol. Bull.
127:294-303.
INVERTEBRATE LARVAL FORMS
Komatsu, M. 1975. On the development of the seastar, Astropecten latespinosus Meissner. Biol. Bull.
148:49-59.
LaBarbera, M. 1984. Feeding currents and particle
capture mechanisms in suspension feeding animals. Axner. Zool. 24:71-84.
Lacalli, T. C. 1980. A guide to the marine flora and
fauna of the Bay of Fundy: Polychaete larvae from
Passamaquoddy Bay, New Brunswick. Can. Tech.
Rep. Fish. Aquat. Sci. 940:1-27.
Lacalli, T. C. 1982. The nervous system and ciliary
band of Muller's larva. Proc. R. Soc. B217:37-58.
Lacalli, T. C. 1986. Prototroch structure and innervation in the trochophore larva of Phyllodoce
(Polychaeta). Can. J. Zool. 64:176-184.
Lester, S. M. 1988. Ultrastructure of adult gonads
and development and structure of the larva of
Rhabdopleura normani (Hemichordata: Pterobranchia). Acta Zool. 69:95-109.
McEuen, F. S. and F.-S Chia. 1985. Larval development of a molpadid holothuroid, Molpadia
intermedia (Ludwig, 1894) (Echinodermata). Can.
J. Zool. 63:2553-2559.
Miall, L. C. 1987. Presidental address. Rep. Br. Assoc.
Advance. Sci. Sect. D:667-683.
Mladenov, P. V. 1979. Unusual lecithotrophic development of the Caribbean brittle star Ophiothrix
oerstedi. Mar. Biol. 55:55-62.
Mladenov, P. V. and F.-S Chia. 1983. Development,
settling behaviour, metamorphosis and pentacrinoid feeding and growth of the feather star Florometra serratissima. Mar. Biol. 73:309-323.
Mortensen, T. 1920. Studies in the development of
crinoids. Pap. Tortugas Lab. 16:1-94.
Mortensen, T. 1921. Studies on the development and
larvalforms ofechinoderms. G. E. C. Gad, Copenhagen, Denmark.
Newth, H. G. 1916. The early development of Cucumaria. Preliminary account. Proc. Zool. Soc. London. Pt. 2, pp. 631-641.
Nielsen, C. 1970. On metamorphosis and ancestrula
formation in cyclostomatous bryozoans. Ophelia
7:217-256.
Nielsen, C. 1971. Entoproct life-cycles and the entoproct/ectoproct relationship. Ophelia 9:209-341.
Nielsen, C. 1987. Structure and function of metazoan
ciliary bands and their phylogenetic significance.
Acta Zool. 68:205-262.
Nielsen, C. and A. N^rrevang. 1985. The Trochaea
theory: An example of life cycle phylogeny. In S.
Conway Morris, J. D. George, R. Gibson, and H.
M. Platt (eds.), The origin and relationships oflower
invertebrates, pp. 28-41. Oxford University,
Oxford.
Ohshima, H. 1921. On the development of Cucumaria echinata v. Marenzeller. Q. J. Microsc. Sci.
65:173-246.
Olson, R. R., J. L. Cameron, and C. M. Young. 1988.
Larval development of the pencil urchin Phyllacanthus imperialis. A lecithotrophic echinopluteus. In R. D. Burke, P. V. Mladenov, P. Lambert,
and R. Parsley (eds.), Echinoderm biology, Proc.
6th Int'l Echinoderm Conf, Victoria, p. 807. A.
A. Balkema, Rotterdam.
721
Parks, A. L., B. W. Bisgrove, G. A. Wray, and R. A.
Raff. 1989. Direct development in the sea urchin
Phyllacanthus parvispinus (Cidaroidea): Phylogenetic history and functional modification. Biol.
Bull. 177:96-109.
Parks, A. L., B. A. Parr, J.-E. Chin, D. S. Leaf, and R.
A. Raff. 1988. Molecular analysis of heterochronic changes in the evolution of direct developing sea urchins. J. Evol. Biol. 1:27-44.
Pearse, J. S. 1979. Polyplacophora. In A. C. Giese
and J. S. Pearse (eds.), Reproduction of marine
invertebrates, Vol. 5, pp. 27-85. Academic Press,
New York.
Percival, E. 1960. A contribution to the life-history of
the brachiopod Tegulorhynchia nigricans. Q. J.
Microsc. Sci. 101:439-457.
Purcell, E. M. 1977. Life at low Reynolds number.
Am. J. Physics 45:3-11.
Rice, M. 1975. Sipuncula. In A. C. Giese and J. S.
Pearse (eds.), Reproduction of marine invertebrates, Vol. 2, pp. 67-127. Academic Press, New
York.
Runnstrom, J. and S. Runnstrom. 1919. Uber die
Entwicklung von Cucumaria frondosa Gunnerus
und Psolus phantapus Strussenfeld. Bergens Mus.
Arboknr. 5:1-100.
Runnstrom, S. 1927. Uber die Entwicklung von Leptosynapta inhaerens (O. F. Muller). Bergens Mus.
Arboknr. 1:1-80.
Ruppert, E. E. 1978. A review of metamorphosis of
turbellarian larvae. In F.-S. Chia and M. E. Rice
(eds.), Settlement and metamorphosis of marine
invertebrate larvae, pp. 65-81. Elsevier, New York.
Sastry,A.N. 1979. Pelecypoda (excludingOstreidae).
In A. C. Giese and J. S. Pearse (eds.), Reproduction
of marine invertebrates. Vol. 5, pp. 113-292. Academic Press, New York.
Schroeder, P. C. and C O . Hermans. 1975. Annelida:
Polychaeta. In A. C. Giese and J. S. Pearse (eds.),
Reproduction of marine invertebrates, Vol. 3, pp.
1-213. Academic Press, New York.
Siebert, A. E. 1974. A description of the embryology,
larval development and feeding of the sea anemones Anthopleura elegantissima and Anthopleura
xanthogrammica. Can. J. Zool. 52:1383-1388.
Silen, L. 1954. Developmental biology of the Phoronidea of the Gullmar Fiord area. Acta Zool. 35:
215-257.
Sleigh, M. A. 1984. The integrated activity of cilia:
Function and coordination. J. Protozool. 31:1621.
Sokal, R. R. and F. J. Rohlf. 1981. Biometry. W. H.
Freeman, San Francisco.
Southward, E. C. 1988. Development of the gut and
segmentation of newly settled stages of Ridgeia
(Vestimentifera): Implications for relationship
between Vestimentifera and Pogonophora. J. Mar.
Biol Assoc. U.K. 68:465-487.
Strathmann, R. R. 1971. The feeding behavior of
planktotrophic echinoderm larvae: Mechanisms,
regulation, and rates of suspension feeding. J. Exp.
Mar. Biol. Ecol. 6:109-160.
Strathmann, R. R. 1974. Introduction to function
722
RICHARD B. EMLET
and adaptation in echinoderm larvae. Thalassia
development of Phoronida. Unpubl. Ph.D. Diss.,
Jugosl. 10:321-339.
University of Washington, Seattle.
Strathmann, R. R. 1978a. The evolution and loss of Zimmer, R. L. and R. M. Woollacott. 1977. Structure
and classification of gymnolaemate larvae. In R.
feeding larval stages of marine invertebrates. EvoM. Woollacott and R. L. Zimmer (eds.), Biology
lution 32:894-906.
of bryozoans, pp. 57-89. Academic Press, New
Strathmann, R. R. 19786. Progressive vacating of
York.
adaptive types during the Phanerozoic. Evolution
32:907-914.
Strathmann, R. R. 1985. Feeding and nonfeeding
APPENDIX
larval development and life-history evolution in
Explanation and sources for features chosen for Table
marine invertebrates. Annu. Rev. Ecol. Syst. 16:
1. A characterization of larval forms for each phylum
339-361.
Strathmann, R. R. 1987. Larval feeding. In A. C. is given below.
Giese and J. S. Pearse (eds.), Reproduction ofmarine Porifera
invertebrates, Vol. 9, General aspects: Seeking unity
Two larval forms, amphiblastulae and parenchyin diversity, pp. 465-550. Blackwell Scientific, Palo
mulae are distinguished in the literature by whether
Alto.
they are hollow or solid (Fell, 1974). In both larval
Strathmann, R. R. 1988. Functional requirements types, shapes are simple: spherical, ovoid, or elongate
and the evolution of developmental patterns. In cylindrical. Larvae are generally uniformly ciliated,
R. D. Burke, P. V. Mladenov, P. Lambert, and R. though often lacking complete cover by cilia, especially
Parsley (eds.), Echinoderm biology, Proc. 6th Int'l at the posterior end. In several parenchymulae, in addiEchinoderm Conf., Victoria, pp. 55-61. A. A. Bal- tion to the uniformly ciliated surface, there is a ring of
kema, Rotterdam.
longer cilia around a bare posterior end (e.g., Haliclona
Strathmann, R. R. and D. Bonar. 1976. Ciliary feed- spp., Callyspongia sp., Adocia sp., Aplysilla sp.
ing of tornaria larvae of Ptychodera flava (Hemi- [Bergquist et al, 1979, and Fell, 1974, Fig. 18]). Bergchordata: Enteropneusta). Mar. Biol. 34:317-324. quist et al. (1979, p. 104) mention that several of the
Strathmann, R. R. and L. R. McEdward. 1986. forms that possess the longer posterior cilia are vigCyphonautes' ciliary sieve breaks a rule of infer- orous swimmers.
ence. Biol. Bull. 171:694-700.
Strieker, S. A. and C. G. Reed. 1985. The ontogeny Cnidaria
of shell secretion in Terebratalia transversa Planulae larvae are uniformly ciliated prolate spher(Brachiopoda, Articulata). I. Development of the oids (Fig. 1L). Most planulae do not feed on paniculate
mantle. J. Morphol. 183:233-250.
food. Planulae that feed with mucous threads have
Strickler, J. R. 1982. Calanoid copepods, feeding cur- been described for two species, Anthopleura xanthorents and the role of gravity. Science 218:158-160. grammica (Siebert, 1974) and Caryophyllia smithi
Thompson, T. E. 1960. The development of Neo- (Trantor et al, 1982). Particles swept past the larval
menia carinata Tullberg (Mollusca Aplacophora). body by cilia adhere to a trailing mucous strand. PeriProc. R. Soc. B 153:263-278.
odically the strand is brought into the mouth and the
Trantor, P. R. G., D. N. Nicholson, D. Kinchington. larval body becomes distended with food. See also
1982. A description of spawning and post-gas- Strathmann (1987).
trula development of cool temperate coral, CarTwo forms of zoanthid larvae are exceptions to uniyophyllia smithi. J. Mar. Biol. Assoc. U.K. 62: form ciliation of planulae: the zoanthella and the zoan845-854.
thina. The zoanthella larva has a longitudinal band that
Vogel, S. 1981. Life in moving fluids, the physical extends anteriorly along ~% of the body from the trailbiology of flow. Willard Grant, Boston.
ing oral end. This band undulates to produce the
Webber, H. H. 1977. Gastropoda: Prosobranchia. In locomotory current. The zoanthina larva has a transA. C. Giese and J. S. Pearse (eds.), Reproduction verse ciliated ring located % of the way from the oral
of marine invertebrates, Vol. 4, pp. 1-98. Aca- end. When actively swimming, larvae are conical with
demic Press, New York.
the ciliated ring posteriormost and the trailing oral face
Williams, D. H. C. and D. T. Anderson. 1975. The flattened. Swimming is by rapid and synchronous beats
reproductive system, embryonic development, of the ring cilia, so that larvae move in spurts (Conklin,
larval development and metamorphosis of the sea 1908). Both these zoanthid larval forms are relatively
urchin Heliocidaris erythrogramma (Val.) (Echi- large (2-4 mm body length) and their cilia are reported
noidea: Echinodermata). Aust. J. Zool. 23:371- to be similar to ctenophore cilia. Because of their large
403.
size (to 8 mm) and different operation of cilia, zoanthid
Wilson, D. P. 1932. On the Mitraria-larva of Owenia larval forms were not included in the analysis.
fusiformis Delle Chiaje. Phil. Trans. Roy. Soc.
Platyhelminthes
London B 221:231-334.
Wilson, E. B. 1904. Experimental studies on germinal
Muller's larvae (Fig. 1H) and directly developing
localization. I. The germ-regions in the egg of Den- larval forms are known. Muller's larva is uniformly
talium. J. Exp. Zool. 1:1-72.
ciliated but has a band of cilia that runs along projecYatsu, N. 1902. On the development of Lingula ana- tions of the larval epidermis called larval arms (Ruptina. J. Coll. Sci. Imp. Univ. Tokyo 17:1-112.
pert, 1978). Ruppert argues that larvae show all the
Zimmer, R. L. 1964. Reproductive biology and necessary equipment to be able to feed, but direct evi-
INVERTEBRATE LARVAL FORMS
dence for feeding is lacking at present (Lacalli, 1982).
Directly developing forms lack larval arms, are simply
shaped, and are uniformly ciliated (references in Ruppert, 1978).
Nemertea
Pilidium larvae are the common, indirectly developing, form (Fig. 1J). These feeding larvae are helmet
shaped with lateral lobes directed away from the direction of swimming. The larval epidermis is sparsely
ciliated; the anterior, posterior, and lateral lobes are
all lined by a band of long cilia at the point of curvature
of the epidermis to the ventral (oral) region (Coe, 1899).
One pilidium larva, Pilidium recurvatum Fewkes, has
a posterior ciliated ring that is probably important for
locomotion (Cantell, 1966).
With a modified indirect development, Iwata's larva
(Iwata, 1958) is a prolate spheroid with a uniformly
ciliated epidermis. The adult develops inside this epidermis. Because Iwata's larva and the direct developing
forms are similar in shape, ciliation, and trophic mode,
I group them under one heading in Table 1. Development of Desor's larva is similar to Iwata's larva, but
occurs inside an egg capsule, so it was not considered.
Polychaeta
Thefirstthree larval forms below are defined according to descriptions of Schroeder and Hermans (1975,
p. 143).
Trochophore larvae. Polychaetes show the greatest
variation of this form, also found in other phyla. Larvae are prolate spheroids with a prototrochal band of
compound cilia at the equator. Commonly (e.g., in
serpulids, sabellarids, and polygordids) a metatrochal
band and food groove are located immediately posterior to the prototrochal band, and these bands are
often raised on a ridge. Other trochophore larvae (e.g.,
phyllodocids, nephthyds, and polynoids) lack the
metatroch and ciliated food groove, though some still
feed, as evidenced by algal cells in their guts (Strathmann, 1978a; Lacalli, 1986).
Metatroch larvae are later stages of development in
which the additional ciliation shows a segmental
arrangement. Often these larvae possess a telotrochal
band of cilia located at the extreme posterior end of
the larval body. Because trochophore and metatrochophore stages intergrade, I combined them in Table 1.
Nectochaete larvae have functional segmentation and
setae (Fig. ID). Larval bodies are prolate spheroids or
cylinders. Keys to polychaete larval fauna show that
these larvae have cilia located on equatorial ridges or
anterior or posterior ends of their bodies (e.g., Lacalli,
1980; Bhaud and Cazaux, 1982).
Mitraria larvae (Owenidae) are unique among polychaetes in that the prototrochal ciliated band is greatly
lengthened and wraps sinuously around a pronounced
ridge on the body (Wilson, 1932). The posterior region
of the larva is attenuated, making the larva appear
roughly as an oblate spheroid with a ruffled edge. Two
bunches of very long provisional setae arise on the
posterior (hyposphere) surface and point posteriorly. I
considered the mitraria to be complexly shaped because
of the elaborated ciliated band and because of possession of the very long setae.
723
Rostraria larvae (Amphinomidae) are vermiform in
shape and characterized by a pair of tentacle-like extensions of the prototrochal and metatrochal bands (Jagersten, 1972). Rostraria larvae also possess a telotrochal
band of cilia that is used for swimming.
Lecithotrophic polychaete larvae are included as an
additional category (L-metatrochophore). In survey of
nonfeeding larvae from 5 families and 6 genera, all
have simple prolate or cylindrical shapes and transverse ciliated bands on ridges (see Blake, 1975a, b, c).
These larvae are trochophores, metatrochophores, and
in some cases polytrochophores (Dorvillidae, especially
Ophryotrocha larvae, are polytrochous, possessing 5 to
6 transverse ciliated bands on an elongate cylindrical
body).
Echiura
Pelagic larvae are simply shaped, feeding trochophores with cilia raised on a ridge at the equator of the
body (e.g., Hatschek, 1880; Dawydoff, 1959). With additional growth larvae resemble polychaete metatrochophores, but metamerism is not permanent. One nonfeeding larva (Bonellia) is known but reported to be
nonplanktonic (Dawydoff, 1959) so is not considered
here.
Sipuncula
Two larval forms are known: trochophore and pelagosphera (Fig. 1K). The trochophore larvae are all nonfeeding and have a prototrochal ring of cilia at the
equator of the spheroidal form. This band occurs at
the widest spot on the larva but is not raised on a ridge.
Pelagosphera can be either feeding or nonfeeding and
has a very prominent metatrochal band of cilia on a
distinct ridge (collar). Some of these larvae retain the
prototroch and this is sometimes on a slight ridge as
well (Rice, 1975).
Pogonophora
Some species may have only freely swimming larvae,
but only stages from species with brooded development
have been described (Southward, 1988). Nonfeeding,
vermiform larvae of' Siboglinumfiordicum are brooded;
however, the later stage larvae have a short pelagic
period (Bakke, 1974). There are two transverse ciliated
rings, one anterior and one posterior; in addition a
ventral longitudinal stripe of shorter cilia. While planktonic, the larva has the anterior band on the widest
part of the body (Bakke, 1974).
Entoprocta
Larvae range from slightly to highly modified trochophore larvae (Fig. 1O; Nielsen, 1971). Prototrochal
cilia are on a pronounced ridge, often at the trailing
end of the body (Nielsen, 1971). Many of these larvae
have short pelagic periods and an alternative mode of
locomotion of gliding over surfaces with a ciliated "foot"
region. When feeding they swim slowly and trail a
mucous thread. Some larvae have complete guts and
may feed, but this is poorly known. Other larvae without guts do not feed.
Mollusca
Gastropods and some bivalves.—Free swimming larvae of these two classes are called veligers, though Gar-
724
RICHARD B. EMLET
stang (1928) recommended the bivalve veliger be distinguished by the name rotiger. These larvae possess
anteriorly directed velar lobes that are round to tentaculate (Fig. 1C, I). Coursing the border of the velar
lobes are compound prototrochal cilia, and (in most
cases) immediately posterior to these are a ciliated food
groove and the metatrochal band of cilia. The cilia of
the velum are used for swimming and feeding. These
larvae also have negatively buoyant shells which can
be ornate but their shapes are roughly prolate or oblate
spheroids. Because the velar lobes are often large flaps
of tissue and are extended in front of the shell, larval
shapes are considered complex. Nonfeeding larvae (Lveliger and L-rotiger of Table 1) of these two types
retain the general from, including shell and velar lobes,
and most of the ciliation of planktotrophic relatives
(see Webber, 1977; Sastry, 1979).
Protobranch bivalves.—These larvae are ovoid and
their exterior consists of five rows of large cells that
make up the larval test. Free swimming species (Yoldia
limatula and Nucula proximo) have 3 distinct transverse bands of longer cilia on 3 central rows of test
cells, and they have shorter cilia covering the anterior
and posterior rows of test cells (Fig. 1A; Drew, 1899).
Drew's drawings indicate no clearly defined ridges where
the bands are located. (One brooding species, Nucula
delphinodonta, has a larva with a test surface that is
uniformly ciliated, but this larva is never free swimming and therefore is not included in the analysis.)
Polyplacophorans. —All free swimming trochophores are prolate spheroids, with a well developed
band of compound of cilia at the widest part of the
body. Later stages retain the ciliated band, but there is
no clear ridge on which it sits (Pearse, 1979).
Aplacophorans.—Larvae are more or less elongate
ovoid or cylindrical. Larvae have anterior and posterior bands of cilia. Both bands for the larvae of Epimenia verrucosa are located for effective movement of
water, whereas the anterior band of Neomenia carinata
is not on a clearly defined ridge or at the anterior end
of the body (Baba, 1938; Thompson, 1960).
Scaphopoda.— The larvae that have been described
from this group originally form a prolate spheroid surrounded by up to 7 transverse rows of cilia (reminiscent
of a protobranch bivalve larva), though 3 equatorial
rows are most prominent. With the appearance of the
conical shell, the larval form elongates and the 3 prominent rows of (prototrochal) cilia are shifted anteriorly
and are still elevated on a ridge (Wilson, 1904; Garstang, 1928).
Bryozoa
The cyphonautes larva is a feeding form, simply
shaped, laterally compressed, and with a triangular
bivalve shell. Bordering the two valves along one side
of the triangle is a flattened ring of cilia called the
corona (Atkins, 1955). The larva moves with this ring
of cilia posteriormost. The corona surrounds openings
to incurrent and excurrent chambers between the two
valves. These two chambers are separated by a pair of
lateral ciliated ridges that create the feeding currents
and transport particles to the mouth (Strathmann and
McEdward, 1986). The coronate larvae of gymnolaemates are nonfeeding forms and can be barrel shaped
or spherical. In some coronate larvae, the corona is a
narrow band of cilia, located either equatorially or along
the trailing edge, adjacent to a flattened surface. The
cyphonautes and this form of coronate larva are considered to have effectively placed ciliated bands (e.g.,
Zimmer and Woollacott, 1977). In other coronate larvae the corona is expanded and covers much of the
lateral surface of the larvae, leaving only small leading
and trailing areas devoid of cilia (Zimmer and Woollacott, 1977). Larvae of stenolaemates (e.g., cyclostomes) are spherical and uniformly ciliated (Nielsen,
1970). The larvae of these last two types are grouped
as Other in Table 1.
Brachiopoda
Articulates. —Only nonfeeding larvae are known. The
larval body consists of three regions: apical, mantle,
and pedicle lobes (Fig. IF). The form is simple and
conular, tapering toward the posterior with deep annuli
between lobes. Only the anterior apical lobe is generally
ciliated, but it also possesses a distinct transverse ring
of longer cilia at its posterior end (Strieker and Reed,
1985).
Inarticulates. —Only feeding larvae are known. Larvae possess a bivalved shell that surrounds an extensible body with mouth and multiple pairs of tentacles
which become the first tentacles in the lophophore (Fig.
IP; Yatsu, 1902). Larvae swim and feed with the tentacles protruded forward and laterally from the shell
(Yatsu, 1902; Chuang, 1977). Though the bivalved shell
is a simple and relatively low drag shape, when tentacles are extended the larva has an elaborate shape
with cilia located on the tentacles.
Phoronida
Actinotroch larvae (Fig. 1G) are distinctive in this
phylum. The larva is roughly cylindrical with a hypertrophied anterior region that folds toward the ventral
surface and forms the preoral lobe. Just posterior to
this lobe, numerous paired tentacles develop during
the free swimmming larval stage. Cilia on these tentacles beat in metachrony and are longer than elsewhere
on the generally ciliated epidermis (Zimmer, 1964). At
the extreme posterior end of the larva there is also a
telotrochal band of compound cilia (Nielsen, 1987)
that is recognized as a swimming organ (Jagersten,
1972).
Also known is a nonfeeding larva (Phoronis ovalis)
that is slug-like in shape, atentaculate, and uniformly
ciliated (Silen, 1954).
Hemichordata
Enteropneusts have a feeding tornaria larva, roughly
spheroidal, but the anterior (upstream) half of the larval body is complexly shaped because it is cut longitudinally and transversely by sunken food grooves (Fig.
IN). A convoluted ciliated band is arranged meridionally on the ridge separating the outer (aboral) surface
from the food groove (oral) surface. This band of simple cilia creates the feeding current. In some species
this ciliated band is so convoluted that small tentacles
occur along the border of the aboral and oral surfaces.
A separate band of longer cilia forms a ring around the
posterior end of the larva. This telotrochal band is
INVERTEBRATE LARVAL FORMS
composed of compound cilia and functions in swimming (Strathmann and Bonar, 1976).
A nonfeeding enteropneust larva, Saccoglossus horsti,
is also known (Fig. 1M; Burdon-Jones, 1952). The larva
is elongate ovoid, uniformly ciliated with a posterior
transverse band. This band is not on a distinct ridge;
nor is it at the extreme posterior end of the larva.
Pterobranchs. —Nonfeeding larvae are elongate, ovoid
and uniformly ciliated, much like planulae (Dilly, 1973;
Hadfield, 1975; Lester, 1988).
Echinodermata
All feeding larvae of asteroids, echinoids, holothuroids, and ophiuroids possess ciliated bands that occur
on ridges and edges of their complexly shaped bodies
(Fig. lE,R,S;e.£., Mortensen, 1921; Strathmann, 1971).
The elaborate larval forms of these 4 classes are so
distinct that they bear special names used in Table 1.
Nonfeeding forms from these 4 classes and nonfeeding crinoid larvae also occur. The nonfeeding forms
are relatively simply shaped but the patterns of ciliation
vary from complete and uniform to ciliated bands on
flat or ridged surfaces. The nonfeeding echinoid larvae
vary from uniformly ciliated, prolate spheroids to forms
resembling feeding larvae but with a reduced number
of larval arms (Williams and Anderson, 1975; Parks
725
etai, 1988, 1989; Mortensen, 1921; Olson tf al, 1988).
In this analysis all were considered simply shaped and
uniformly ciliated.
Nonfeeding asteroid larvae are uniformly ciliated
and are shaped more or less as oblate or prolate spheroids but some also bear brachiolar arms {e.g., Gemmill,
1912; Birkeland et al., 1971; Komatsu, 1975).
Most nonfeeding ophiuroid larvae are cylindrical to
spheroid with a series of incomplete transverse ciliated
bands that occur on ridges (Fig. IB; Hendler, 1975).
Several other species have larvae which resemble feeding forms but have a reduced number of larval arms
and ciliated bands (Fenaux, 1963; Mladenov, 1979).
All were considered to be simply shaped with bands
of cilia.
Nonfeeding holothurian larvae can either have rings
of cilia or be uniformly ciliated (banded—e.g., Ohshima, 1921; Runnstrom, 1927; McEuen and Chia, 1985;
uniformly ciliated—e.g., Runnstrom and Runnstrom,
1919; Newth, 1916).
All described crinoid larvae are nonfeeding doliolaria with transverse ciliated rings that do not occur on
ridges (e.g., Bury, 1888; Mortensen, 1920; Mladenov
and Chia, 1983). The posteriormost band is at the end
of the larval body, as are the telotrochs of other larval
forms.