Morphology of a New Deep-Sea Acorn Worm (Class
Enteropneusta, Phylum Hemichordata): A Part-Time
Demersal Drifter with Externalized Ovaries
Nicholas D. Holland,1* Linda A. Kuhnz,2 and Karen J. Osborn3
1Marine Biology Research Division, Scripps Institution of Oceanography, University of California at San Diego,
La Jolla, California 92093-0202
2Monterey Bay Aquarium Research Institute, 7700 Sandholdt Road, Moss Landing, California 95039-9644
3Department of Invertebrate Zoology, Smithsonian Institution, National Museum of Natural History,
Washington, DC 20560
ABSTRACT Ten individuals of an enteropneust in the
family Torquaratoridae were videotaped between 2,900
and 3,500 m in the Eastern Pacific—one drifting a few
centimeters above the bottom, two exposed on the sub-
strate, and seven partly burrowed, reflecting a bentho-
pelagic life style. Here, we describe a captured specimen
(26 cm living length) as the holotype of Allapasus auran-
tiacus n. gen., n. sp. The small proboscis is dome-shaped,
and the collar is only slightly wider than deep; both of
these body regions are more muscular than in other tor-
quaratorids, which presumably facilitates burrowing.
The proboscis complex, in contrast to that of shallow-liv-
ing enteropneusts, lacks a pericardial sac and is located
relatively posteriorly in the proboscis stalk. The stomo-
chord is separated from the main course of the gut by
the intervention of a small, plate-like proboscis skeleton
lacking posterior horns. The most anterior region of the
trunk houses the pharynx, in which the pharyngeal
skeletal bars are not connected by synapticles. The post-
pharyngeal trunk comprises three intestinal regions:
prehepatic, hepatic (with conspicuous sacculations), and
posthepatic. On either side of the worm, a flap of body
wall (lateral wing) runs the entire length of the trunk.
The two lateral wings can wrap the body so their edges
meet in the dorsal midline, although they often gape
open along the pharyngeal region. The holotype is a
female (presumably the species is gonochoric) with
numerous ovaries located in the lateral wings along the
pharyngeal region. Each larger ovary contains a single
primary oocyte (up to 1,500 lm in diameter) and bulges
outwards in an epidermal pouch attached to the rest of
the body by a slender stalk. Such externalized ovaries
are unprecedented in any animal, and nothing is yet
known of their role in the reproductive biology of A. aur-
antiacus. J. Morphol. 000:000–000, 2012. ! 2012 Wiley
Periodicals, Inc.
KEY WORDS: Hemichordata;Enteropneusta;Torquaratoridae;
Allapasusaurantiacus
INTRODUCTION
Hyman (1959), in reviewing acorn worms (class
Enteropneusta, phylum Hemichordata), general-
ized that they typically live as benthic infauna at
depths from the intertidal zone to the continental
shelf—the only exceptions she noted were one shal-
low-living species that sometimes swims by undu-
lating the body and a single deep-sea species
dredged by the ‘‘Challenger’’ expedition. During the
closing decades of the twentieth century, these older
ideas needed to be modified because deep-sea photo-
graphs (summarized in Table 1 of Smith et al.,
2005) revealed more and more enteropneusts living
entirely exposed on the ocean floor. During that pe-
riod, it was difficult to image an animal at great
depth and then collect the same specimen, although
this was accomplished in 1979 when the manned
submersible ‘‘Alvin’’ collected a deep-sea acorn
worm later described as Saxipendium coronatum by
Woodwick and Sensenbaugh (1985). Fortunately,
during the last decade, the use of remotely operated
vehicles (ROVs) has facilitated in situ video record-
ing of deep-living animals followed by collection of
the same specimens for anatomical and molecular
phylogenetic studies.
By now, two enteropneust species brought to the
surface by ROVs have been named—Torquarator
bullocki (Holland et al., 2005) and Tergivelum
baldwinae (Holland et al., 2009)—and others have
been collected but not yet formally described. In a
recent molecular phylogenetic study of named and
unnamed acorn worms, Osborn et al. (2012) dem-
onstrated that most enteropneusts living in the
deep sea belong to a single clade, the family Tor-
quaratoridae. In this article, the torquaratorid
Contract grant sponsor: David and Lucile Packard Foundation.
*Correspondence to: Nicholas D. Holland, Marine Biology
Research Division, Scripps Institution of Oceanography, University
of California at San Diego, La Jolla, CA 92093-0202.
E-mail: nholland@ucsd.edu
Received 17 October 2011; Revised 22 December 2011;
Accepted 22 January 2012
Published online in
Wiley Online Library (wileyonlinelibrary.com)
DOI: 10.1002/jmor.20013
JOURNAL OF MORPHOLOGY 000:000–000 (2012)
! 2012 WILEY PERIODICALS, INC.
that Osborn et al. (2012) designated as Genus D,
species 1 is described from deep-sea recordings as
well as histological reconstructions and named Alla-
pasus aurantiacus n. gen, n. sp. We directly demon-
strate that these worms can drift demersally a short
distance above the bottom, although we more often
observed them extended on the surface of the ocean
floor or partly burrowed in the sediment.
MATERIALS AND METHODS
Between 2002 and 2009, the ROVs ‘‘Tiburon’’ and ‘‘Doc Rick-
etts’’ of the Monterey Bay Aquarium Research Institute (MBARI)
videotaped 10 specimens of an enteropneust in the Monterey
Submarine Canyon and at the base of Davidson Seamount
between 60 and 100 km offshore of Monterey, California, at
depths between 2,900 and 3,500 m (Table 1, specimens A–J).
Only two of the imaged animals were brought to the surface. One
was recovered in good condition (Fig. 1A,B; specimen A in Table
1) and is the holotype described here. The other (Fig. 1D; speci-
men H in Table 1) was badly damaged during collection and was
suitable only for molecular sequencing (Osborn et al., 2012).
The holotype was videotaped for 3 min as it drifted a few centi-
meters above the ocean floor (Fig. 1A) before being captured in a
funnel-mouthed suction sampler facing into the approaching cur-
rent (Fig. 1B). After the captured worm had been brought to the
surface, several large oocytes were removed and frozen in liquid
nitrogen for molecular analysis (Osborn et al., 2012), and the
body, which had broken into two pieces at the level of the anterior
trunk, was fixed in 10% formalin-seawater. The fixed specimen
was photographed under a dissecting scope to show the overall
morphology as well as details of the pharyngeal pores, pharyn-
geal skeletal bars, and parts of the gonadal region. For light mi-
croscopic reconstruction of the internal morphology, we embed-
ded selected regions of the body in paraplast and prepared them
as 15-lm serial sections stained in 0.1% aqueous azure A (Spicer,
1963). After paraplast embedding, the larger oocytes and hepatic
region of the intestine tended to shatter when sectioned. There-
fore, these tissues were embedded in Spurr’s resin, sectioned at 4
lmwith a glass knife, and stained in 0.1% aqueous azure A.
RESULTS
Systematics
Class Enteropneusta Gegenbaur, 1870; Family
Torquaratoridae Holland et al., 2005 (as rediagnosed
by Osborn et al. (2012), limiting it to enteropneusts
whose proboscis skeleton is either absent or reduced
TABLE 1. Collection data for Allapasus aurantiacus
Specimen Date Dive no. Depth (m) Latitude N Longitude W Behaviora
Ab,c 12 Jun ’02 T438 2,994 36834048@ 122829030@ Drifting
B 13 Jun ’02 T439 3,492 36819041@ 122853051@ Epibenthic
C 11 Mar ’05 T829 3,455 36814021@ 122853027@ Burrowed
D 16 Dec ’05 T930 3,266 35848034@ 12283405@ Epibenthic
E 21 Jan ’06 T938 3,266 35848038@ 12283405@ Burrowed
F 10 Jan ’07 T1069 2,893 36836048@ 12282609@ Burrowed
G 19 Dec ’07 T1162 2,891 36836048@ 12282608@ Burrowed
Hc 19 Nov ’09 D098 2,893 36836033@ 12282603@ Burrowed
I 19 Nov ’09 D098 2,891 36836032@ 12282603@ Burrowed
J 19 Nov ’09 D098 2,891 36836032@ 12282603@ Burrowed
aAll the ‘‘Burrowed’’ worms were partially emergent onto the sea floor.
bMorphological holotype (Voucher ID: SIO-BIC-H20).
cMolecular sequence (in Osborn et al., 2012).
Fig. 1. Allapasus aurantiacus alive in situ; single frames from video recordings. (A) Holotype (A in Table 1) drifting in current
near bottom. (B) Holotype entering collecting funnel. (C) Burrowed worm (C in Table 1) with its anterior fourth protruding. (D)
Partially burrowed worm (H in Table 1) being collected with push-core (bivalves at bottom right). Scale bar in A,C,D 5 5 cm; in B
5 10 cm.
2 N.D. HOLLAND ET AL.
Journal of Morphology
to a small medial plate and whose adult stomochord
is either absent or not in communication with the
main course of the gut).
Genus Allapasus n. gen. Type and only species:
Allapasus aurantiacus n. sp.
Etymology: name of genus derives from Latin
masculine noun Allapasus 5 gliding approach. Di-
agnosis: Small, plate-like proboscis skeleton inter-
vening between stomochord and main course of
gut, lateral wings running along entire length of
the trunk, and with sacculate hepatic intestine
shorter than posthepatic intestine.
Allapasus aurantiacus n. sp. (Figs. 1–5). Cap-
ture site and labeling of holotype: ROV Tiburon,
MBARI dive T438, 36834’48"W, 122829’30"N, 2994
m, 12 June 2002, Karen J. Osborn collector; forma-
lin-fixed female prepared as histological sections
conserved along with unsectioned body parts in
the Scripps Institution of Oceanography Benthic
Invertebrate Collection (SIO-BIC-H20); no para-
types. Diagnostic features same as for genus. In
life, proboscis and collar light orange, with trunk
ranging from beige to light orange depending on
individual. Etymology: species name derives from
Latin masculine adjective aurantiacus 5 orange,
in reference to the light orange color of the probos-
cis and collar.
Behavior, Living Appearance, and General
Anatomy
In the 3-min videotape of the living holotype of
A. aurantiacus (Fig. 1A), the worm was drifting
with the current a few centimeters above the bot-
tom. During the recording, there was no evidence
of active swimming: the loosely curled posture of
the worm underwent no detectable change, and no
peristaltic or undulatory movements were
observed. Moreover, there appeared to be no gut
contents in the intestinal region running through
the relatively translucent posterior part of the
worm. The living body length, as estimated by a
comparison with the dimensions of the collecting
funnel (Fig. 1B), was about 26 cm. Other individu-
als of A. aurantiacus were videotaped on the sea
floor—of these, two were entirely exposed, and
seven were partly burrowed beneath the surface
(Table 1; Fig. 1C,D).
From anterior to posterior, the main body
regions are the proboscis, collar, and trunk. A lat-
eral wing (a wide flange of body wall) runs along
either side of the entire length of the trunk. In
life, the right and left wings typically wrap around
the body and appose their free edges dorsally,
except in the pharyngeal region of the trunk,
where they often leave a gap dorsally (Fig. 1A,C).
In the drifting holotype (Fig. 1A), the gaping of the
lateral wings in the pharyngeal region exposed the
numerous externalized ovaries (described below)
as cream-colored spheres.
The body of the holotype (Fig. 2A,B), broken into
two parts at collection, lost its pigmentation soon
after fixation and shrank to a total length of about
18 cm. Along with the fixed worm, the collecting
jar contained a large clump of flocculent material
(Fig. 2A, bottom left), evidently a precipitate of
mucus that either had surrounded the living ani-
mal as a transparent sheath or had been secreted
in response to the fixative. Moreover, the collection
jar contained no feces, and the gut lumen was
empty after fixation. Thus, all the evidence indi-
cates that the gut of the drifting holotype was
truly empty.
The major regions of the holotype are visible in
Figure 2A,B, and the external details of the ante-
rior end of the body are shown in Figure 2C–E. All
measurements refer to the fixed specimen. The
proboscis is a pointed dome (0.6 cm anterior to
posterior, 0.6 cm dorsal to ventral, and 1 cm wide)
and is deeply indented by a narrow groove running
along the ventral midline. The collar measures 0.5
cm anterior to posterior, 0.6 cm dorsal to ventral,
and 1 cm wide. The trunk measures about 17 cm
anterior to posterior, 0.5 cm dorsal to ventral, and
1 cm wide (except along the region of the hepatic
intestine, which is about 0.3 cm from dorsal to
ventral and 1.4 cm wide). The right and left lateral
wings, running along either side of the trunk,
arise dorsolaterally in the pharyngeal region and
ventrolaterally in the intestinal region; the apposi-
tion of these flaps obscures the dorsal side of the
trunk in the fixed holotype. All along the ventral
side, the trunk is indented by a groove in which
runs the ventral nerve cord (Fig. 2A, arrowhead).
The dorsal nerve cord (Fig. 2B, arrowhead) is visi-
ble running along the dorsal midline of the trunk.
Histological Structure
Proboscis and collar. A cross section near the
anterior tip of the proboscis shows a core of
smooth muscle fibers overlain by the epidermis
(Fig. 2F).The muscles are organized as a moder-
ately compact network with a tendency for radially
oriented fibers to predominate. The epidermis of
the proboscis, like that of the body generally, com-
prises supporting cells, gland cells (many contain-
ing acid mucopolysaccharides), and elements of the
intraepithelial nervous system, most conspicuous
as a basal fibrous layer (arrowed in inset in Fig.
2F). Within the muscle fiber mass of the proboscis,
the protocoel opens up (Fig. 2G,H, arrows). Probos-
cis pores are absent, a lack that seems to be a fea-
ture of torquaratorids in general (Holland et al.,
2005, 2009).
Figure 2I,J shows where the proboscis stalk
begins to merge with the more dorsal collar tis-
sues. The dorsal nerve cord (single arrow) is asso-
ciated with a pair of proboscis vessels (arrow-
heads) and is underlain by the anterior tip of the
MORPHOLOGY OF A NEW DEEP-SEA ACORN WORM 3
Journal of Morphology
Fig. 2. Allapasus aurantiacus holotype. (A) Formalin-fixed worm broken into anterior fourth (left-side view) and posterior
three-fourths (ventral view at left twists to left-side view at right); pr, proboscis; co, collar; ph, pharynx; pri, prehepatic intestine;
hi, hepatic intestine; pi, posthepatic intestine; arrowhead indicates ventral nerve cord and arrows indicate folds in ventral body
wall (possibly artifacts); extraneous mucus collected with living worm is at lower left. (B) Same, broken into anterior fourth (right-
side view) and posterior three-fourths (dorsal view at left twists to right-side view at right); arrowhead indicates dorsal nerve cord;
levels of cross sections in Figure 5 are indicated. (C) Enlarged right-side view of anterior end; levels of cross sections in Figures 2,
4, and 5 indicated. (D) Ventral view of anterior end; arrow indicates ventral groove in proboscis. (E) Dorsal view of anterior end.
(F) Cross section of proboscis (inset: enlargement of epidermis with basal nerve fiber layer indicated by arrow). (G) Cross section of
proboscis through protocoel (arrow). (H) Cross section showing anterior rim of collar (top) and proboscis (containing muscle masses
with posterior extension of protocoel (arrows). (I,J) Rectangle in former enlarged in latter, showing dorsal nerve cord (single
arrow), proboscis vessels (arrowheads), and anterior extremity of heart (twin arrow). (K,L) Rectangle in former enlarged in latter,
showing dorsal nerve cord (single arrow), proboscis vessel (arrowhead), heart (twin arrow), and anterior extremity of stomochord
(tandem arrow). Scale bar in A,B 5 1 cm; in C–I, K 5 2 mm; in J,L 5 500 lm; for inset in F 5 100 lm.
4 N.D. HOLLAND ET AL.
Journal of Morphology
heart (twin arrow). The anatomical relations of
these features are shown in Figure 3, which dia-
grams the proboscis complex and related haemal
vessels. In a slightly more posterior section (Fig.
2K) the anterior extremity of the stomochord is
visible ventral to the heart. The enlargement (Fig.
2L) shows that the paired proboscis vessels have
joined at this cross sectional level to form a single
proboscis vessel. As diagrammed in Figure 3, the
latter arises from dorsal side of the heart. More
posteriorly, the dorsal side of the heart also gives
rise to the dorsal longitudinal vessel, which runs
in a posterior direction (Fig. 3). The proboscis com-
plex of A. aurantiacus, in comparison to that of
spengelids, harrimaniids, and ptychoderids, lacks
a pericardial sac and is located relatively posteri-
orly in the proboscis stalk. This situation is inter-
mediate between the absence of a proboscis com-
plex in most torquaratorids (Holland et al., 2009)
and its prominence in spengelids, harrimaniids,
and ptychoderids (Hyman, 1959).
Figure 4A,B, a short distance posterior to the
level of the anterior neuropore, shows the collar
nerve cord, heart, and stomochord. The last is
packed with vacuolated cells and has no obvious
lumen anywhere along its length. In contrast to its
location in nontorquaratorid hemichordates, the
glomerulus is more closely associated with the sto-
mochord than with the heart. Just posterior to the
heart (Fig. 4C,D), the proboscis complex comprises
only the stomochord and a narrowed posterior
extension of the glomerulus. At this level, the col-
lar nerve cord is underlain by a pair of muscle-
filled perihaemal coeloms, which are separated by
a mesentery containing the dorsal longitudinal
vessel.
Approximately midway between the anterior
and posterior limits of the collar (Figs. 3 and
4E,F), a peribuccal vessel arises in the muscle
mass of the body wall on either side of the worm
and runs posteriorly. Figure 4G–J shows the
reduced proboscis skeleton, as a small, nearly ver-
tical plate that lacks the posterior horns character-
izing the proboscis skeleton of shallow-living enter-
opneusts. The plate-like proboscis skeleton of A.
aurantiacus completely separates the stomochord
from the main course of the digestive tract.
Underlying the conspicuous collar tissues, the
posterior region of the proboscis stalk (Fig. 4G–J)
hangs down into the roof of the buccal cavity as
sacculations, probably equivalent to the racemose
(5 cauliflower) organ of some other enteropneusts.
More posteriorly in the collar (Fig. 5A), no ele-
ments of the proboscis stalk remain. The body wall
of the collar, like that of the proboscis, comprises a
moderately compact meshwork of muscle fibers
among which no peribuccal coeloms were detected.
At cross sectional levels shown in Figure 5B–F,
left and right parabranchial ridges protrude dorso-
laterally into the pharyngeal lumen, partially
dividing it into a dorsal branchial region and a
ventral digestive region; presumably, in feeding
animals, the gut contents are limited to the ven-
tral region of the pharynx. No coelomopores are
present in the collar of A. aurantiacus, as may be
typical of torquaratorids generally (Holland et al.,
2005, 2009; our unpublished observations). The
most anterior tissues of the trunk appear dorsal to
the collar-trunk septum near the level of the poste-
rior neuropore (Fig. 5B,C).
Pharyngeal region of trunk (and ovary).
Figure 5D shows trunk tissues and collar tissues,
respectively, dorsal and ventral to the collar-trunk
septum. Left and right lateral wings arise as out-
foldings of the dorsolateral body wall, and the peri-
buccal vessels have joined medially beneath the gut
to form the ventral longitudinal vessel. At a level a
little more posteriorly, cross sections show the ven-
tral nerve cord (Fig. 5F,G), which begins at the
anterior end of the trunk. The body wall of the
pharyngeal region of the trunk, like that of the pro-
boscis and collar, includes a moderately compact
meshwork of muscle fibers. The branchial region of
the pharynx (Fig. 5E) opens to the exterior via
about a hundred pairs of slot-like pharyngeal pores
Fig. 3. Diagram of Allapassus aurantiacus proboscis complex
(actually located in anterior half of collar) and related haemal
vessels. At left are indicated levels for the anterior neuropore
(ANP), posterior neuropore (PNP), and Figures 2I and
4A,C,E,G. At right, abbreviations (top to bottom) are: PrV, pro-
boscis vessel; Ht, heart; StC, Stomochord; PSk, proboscis skele-
ton; PBV, peribuccal vessel; DLV, dorsal longitudinal vessel; and
VLV, ventral longitudinal vessel.
MORPHOLOGY OF A NEW DEEP-SEA ACORN WORM 5
Journal of Morphology
Fig. 4. Allapasus aurantiacus holotype. (A,B) Rectangle in former enlarged in latter, showing stomochord (tandem arrow)
underlain by glomerulus (single arrow) and overlain by heart (twin arrow) and collar nerve cord (arrowhead). (C,D) Rectangle in
former enlarged in latter, showing stomochord (tandem arrow) underlain by posterior extension of glomerulus (asterisk); collar
nerve cord (arrowhead) is underlain by muscle-filled perihaemal coeloms (twin arrow) separated by mesentery containing dorsal
longitudinal vessel (single arrow). (E,F) Rectangle in former enlarged in latter, showing peribuccal vessels (arrowheads); stomo-
chord (tandem arrow) underlain by posterior extension of glomerulus (asterisk) and overlain by space (single arrow) continuous
with buccal cavity. (G) Composite based on successive sections H-J showing posterior end of stomochord separated from buccal cav-
ity by proboscis skeleton comprising a thin plate (single arrow). (H–J) Enlargements of rectangle in G showing three contiguous
sections; posterior extremity of glomerulus indicated by asterisks; in H, sacculations of racemose (5 cauliflower) organ indicated by
single arrows. Scale bar in A,C,E,G 5 2 mm; in B,D,F,H-J 5 500 lm.
6 N.D. HOLLAND ET AL.
Journal of Morphology
(Fig. 5H) and is strengthened by primary and sec-
ondary pharyngeal skeletal bars not connected by
synapticles (Fig. 5I,J). Water exits the pharyngeal
pores into a dorsal space bounded on either side by
the lateral wings and open to the surrounding sea
water by a mid-dorsal gap. The epidermis facing
this dorsal space in this region of the trunk is asso-
ciated with numerous ovaries (Fig. 5F, arrow).
Each immature ovary, which lies just beneath
the epidermis, comprises a germinal epithelium
surrounding a small mass of nongerminal cells
(Fig. 5L, arrowhead). Somewhat more developed
ovaries still lie just beneath the epidermis, but the
germinal epithelium now encloses single, small
primary oocyte (Fig. 5L, single arrow). Each oocyte
contains a large germinal vesicle that includes a
prominent nucleolus. At a later stage, ovaries con-
taining medium-sized or large oocytes (Fig. 5K,L)
bulge outward, beyond the surface of the animal.
Such ovaries are surrounded by two very thin
membranes, the inner comprising the germinal ep-
ithelium and the outer continuous with the epider-
mis. Each externalized ovary remains attached to
the outside of the worm by a narrow stalk (Fig.
5K, arrow). We know of no comparable instance of
externalized ovaries in any other animal group.
The largest oocytes, of which there were several
dozen (Fig. 5M), had a diameter of around 1,500
lm, equaling the size of those in Tergivelum bald-
winae (Holland et al., 2009). In the present study
of A. aurantiacus, the only specimen available for
histological study was the female holotype: pre-
sumably, the sexes are separate in this species,
and males will ultimately be discovered.
Intestinal region of trunk. The body wall of
the intestinal region of the trunk contains only a
sparse meshwork of muscle fibers. The right and
left lateral wings, which arise ventrolaterally
along either side of this trunk region can fold dor-
sally and meet in the dorsal midline of the living
worm (shrinkage of the fixed tissues causes the
lateral wings to pull slightly apart dorsally, as in
Fig. 5N). The dorsal and ventral nerve cords (and
their underlying dorsal and ventral longitudinal
vessels) continue all along this region of the body
(Fig. 5O,P). In many enteropneusts, the gut region
just posterior to the pharynx has been called the
esophagus; however, in A. aurantiacus there is no
justification anatomically or histologically for mak-
ing such a distinction, and we will simply call this
region the prehepatic intestine.
In the pre- and posthepatic intestine (Fig.
5N,T,U), the lining epithelium is relatively flat
ventrally, but is corrugated into plicae dorsally
and laterally. From the dorsal midline, each plica
runs ventro-anteriorly at an angle of about 458.
The hepatic intestine intervenes between the pre-
and posthepatic regions and is morphologically dis-
tinct from them, at both the gross anatomical and
cytological levels of organization. In the hepatic
region (Fig. 5Q–S), the lining epithelium is rela-
tively flat on the ventral side, but the dorsal side
bulges with large-scale outpocketings (hepatic sac-
culations) oriented perpendicular to the long axis
of the animal. The epithelium lining the saccula-
tions of the hepatic intestine includes many cells
filled with inclusions that range from light to dark
brown in unstained sections.
DISCUSSION
Enteropneusts with a Bentho-Pelagic Life
Style
Travel in the water column. Postlarval enter-
opneusts living in shallow water almost never float
or swim above the bottom; the only known excep-
tions are individuals of the genus Glandiceps that
can swim by undulating the body (Ikeda, 1908;
Yoshimatu and Nishikawa, 1999). In contrast,
deep-living enteropneusts, as adults, are much
more prone to spend time in the water column.
The first report of a deep-living enteropneust leav-
ing the sea floor was by Gaillard (1991) who
reported that the animal, probably in Genus A of
Osborn et al. (2012), ‘‘swam away’’ in some unspe-
cified manner when disturbed. Subsequently, a
more passive drifting behavior has been directly
observed in three more deep-sea acorn worms
(Osborn et al., 2012). In addition, and is strong cir-
cumstantial evidence that a fourth species, Tergi-
velum baldwinae can similarly drift in the water
column, because time-lapse photography showed
one of these worms appearing abruptly, foraging
for a day and a half, and suddenly disappearing,
presumably into the water column (Smith et al.,
2005). It would not be surprising if further work
shows that most, if not all, torquaratorids are
part-time demersal drifters.
Nothing is yet known about how deep-sea enter-
opneusts control their ascent into the water col-
umn, although rising from the bottom is presum-
ably facilitated by emptying the gut and possibly
also by the secretion of a thick coating or balloon
of mucus (Osborn et al., 2012). It is also not known
how a demersally drifting worm descends to a new
benthic feeding site. Moreover, there is no informa-
tion to indicate the horizontal or vertical distances
typically traversed in a given drifting episode or
the overall proportion of time spent above the bot-
tom.
Life on the bottom: burrowing versus
surface crawling. Although many shallow-living
enteropneusts inhabit burrows in soft substrates,
this generality has been overstated by some (like
Barrington, 1965), who concluded that ‘‘Enterop-
neusts, then are essentially subterranean and bur-
rowing animals . . .’’ On the contrary, some shal-
low-water species can live epibenthically either
temporarily or permanently. Benham (1899) found
specimens of Balanoglossus otagoensis crawling on
MORPHOLOGY OF A NEW DEEP-SEA ACORN WORM 7
Journal of Morphology
Fig. 5.
8 N.D. HOLLAND ET AL.
Journal of Morphology
the surface of kelp holdfasts; Ikeda (1908) claimed
that Glandiceps hacksii is ‘‘a creeper, but not a
burrower as other enteropneusts are’’; Stiasny
(1910) reported, with some skepticism, that fisher-
men told him that Balanoglossus clavigerus crawls
out of its burrow to sun itself when cold; Knight-
Jones (1953) found that Glossobalanus minutus
leaves its burrow and crawls on the surface when
oxygen is limiting; Cameron et al. (2010) proposed
that Saccoglossus rhabdorhynchus might crawl on
the surface of rocks; and Hadfield (personal com-
munication) informed us that a still-undescribed
ptychoderid lives on the surface of hardened lava
at a depth of 25 m at Ahihi-kinau, Maui, Hawai’i.
In spite of these exceptions, however, it is fair to
generalize that enteropneusts living at shallow
depths are essentially burrowers.
In contrast to their shallow-living relatives,
most torquaratorids, when associated with the bot-
tom, are seen in their entirety, fully exposed on
the ocean floor (Holland et al., 2005; Smith et al.,
2005; Cannon et al., 2009; Holland et al., 2009;
Anderson et al., 2011; Osborn et al., 2012). More-
over, the worms are often observed at the head of
a fecal trail laid out on the ocean floor. Depending
on the species, trail patterns may be spiral, looped
in approximate raster patterns, or loosely mean-
dering. Allapasus aurantiacus contrasts with other
torquaratorids, first because no individual has yet
been observed leaving a fecal trail on the surface
of the sea floor and second because worms of this
species are frequently encountered partly bur-
rowed in the substratum (Fig. 1C,D); the burrow-
ing is presumably facilitated by its proboscis and
collar muscles, which are well developed in com-
parison to those of other torquaratorids. At pres-
ent, it is not known if the burrow of A. aurantia-
cus has a conspicuous wall, as in Saccoglossus
inhacensis (van der Horst, 1934), or is indistinct,
as in Balanoglossus simodensis (Miyamoto and
Saito, 2007).
Feeding in shallow- versus deep-living
enteropneusts. It was once assumed that most
enteropneusts swallow the substratum indiscrim-
inately and pass it through the gut earthworm-
style (van der Horst, 1939). Subsequently, how-
ever, Barrington (1940) and Carey and Mayer
(1990) claimed that shallow-living enteropneusts
deposit feed by protruding the proboscis from the
burrow entrance and using ciliary-mucoid tracts to
pick up the top few millimeters of sand plus or-
ganic matter for transport to the mouth. It is
widely believed that this kind of feeding is not
selective: thus the size spectrum of sand particles
in the gut lumen mirrors that in the top few milli-
meters of substratum (Knight-Jones, 1953; Colin
et al., 1986; Dobbs and Guckert, 1988; Miller,
1992). Such surface deposit feeding stuffs the gut
lumen with sand and results in the production of
sandy fecal casts. Although Karrh and Miller
(1996) claimed that Saccoglossus kowalevskii is an
obligate deposit feeder incapable of suspension
feeding, many other enteropneusts in shallow
depths have been found to augment their deposit
feeding by sucking suspended particles directly in
at the mouth and capturing them within the phar-
ynx (Thomas, 1972; Cameron, 2002; Gonzalez and
Cameron, 2009). Finally, according to Ikeda (1908),
Glandiceps hacksii deviates strikingly from other
shallow-living enteropneusts in having sand-free
gut contents comprising only ‘‘microorganisms.’’
The absence of ingested sand could indicate that
Fig. 5. Allapasus aurantiacus holotype. (A) Cross section of collar showing peribuccal vessels (arrowheads) and collar nerve
cord (single arrow). (B,C) Rectangle in former enlarged in latter; level of posterior neuropore, showing peribuccal vessels (arrow-
heads), parabranchial ridges (tandem arrow) and dorsal nerve cord (twin arrow) underlain by paired perihaemal spaces; single
arrows indicate collar-trunk septum. (D) Collar-trunk septum (single arrows) delimiting anterior extremity of trunk (towards top)
from posterior extremity of collar (at bottom); lateral wings indicated by twin arrows; parabranchial folds (tandem arrow) separate
pharynx into dorsal branchial part (single asterisk) and ventral digestive part (twin asterisks); ventral longitudinal vessel (arrow-
head) underlies pharynx. (E) Enlargement of rectangle in D; dorsal nerve cord (arrowhead) overlies dorsal longitudinal vessel
(single arrow); tandem arrows indicate pharyngeal skeletal bars. (F) Pharynx divided by parabranchial ridges (tandem arrows)
into respiratory and digestive parts; dorsal to pharynx, lateral wings almost completely enclose space including some oocytes
(single arrow). (G) Enlargement of rectangle in F, showing ventral nerve cord (arrowhead) and ventral longitudinal vessel (twin
arrow). (H) Dorsal surface of pharynx showing dorsal nerve cord (arrowhead) flanked on either side by pharyngeal pores (single
arrows). (I) Inner surface view of one side of respiratory pharynx (dorsal toward left) showing primary and secondary skeletal bars
(arrow and arrowhead, respectively). (J) Cross section of primary and secondary skeletal bars (arrow and arrowhead, respectively).
(K) Inner surface view of lateral wings in pharyngeal region; medium sized oocytes bulge outward, but remain connected to under-
lying tissue by a stalk (single arrow). (L) Section through inner side of lateral wing in pharyngeal region showing ovaries contain-
ing either nongerminal cells (arrowhead) or small oocytes (single arrow); twin arrow indicates medium-sized oocyte in extruded
ovary connected to underlying tissue. (M) Section of oocyte of maximum size freed from outer epithelial layer, but surrounded by
thin jelly layer (single arrow) and germinal epithelium (arrowhead). (N) Cross section of prehepatic intestine with lateral wings
indicated by single arrows. (O) Enlargement of upper rectangle in N, showing dorsal nerve cord (arrowhead) and dorsal longitudi-
nal vessel (single arrow). (P) Enlargement of lower rectangle in N, showing ventral nerve cord (arrowhead) and ventral longitudi-
nal vessel (single arrow). (Q) Cross section of hepatic intestine with sacculations (tandem arrow), dorsal nerve cord (arrowhead)
and ventral nerve cord (single arrow). (R) Dorsal surface view of hepatic intestine showing sacculations flanking the dorsal nerve
cord (single arrow); line between arrowheads indicates orientation of section in S. (S) Parasagittal section of hepatic sacculations;
gut lumen is toward bottom; sacculations overlain by thin epidermis (single arrow). (T,U) Cross sections of anterior and posterior
regions, respectively, of posthepatic intestine with dorsal nerve cord (arrowhead) and ventral nerve cord (single arrow). Scale bar
in A,B,D,F,N,Q,T,U 5 2 mm; in I,M 5 500 lm; in K 5 400 lm; in R 5 4 mm; in C,E,H,J,S 5 300 lm; in L 5 200 lm; in G,O,P 5
100 lm.
MORPHOLOGY OF A NEW DEEP-SEA ACORN WORM 9
Journal of Morphology
this species is an exclusive suspension feeder and/
or a highly selectively deposit feeder that can pick
up organic matter while excluding mineralized
sediment particles.
Most torquaratorid enteropneusts are epibenthic
deposit feeders that differ from most of their shal-
low-water relatives in having gut contents virtu-
ally devoid of sand grains (Holland et al., 2005;
Smith et al., 2005; Holland et al., 2009; Holland,
unpublished observations). Evidently, these deep-
living worms can select particles rich in organic
material from the surface of the deep sea without
an appreciable admixture of mineralized sedi-
ments. It is likely that much of the selection takes
place in ciliary-mucoid tracts on the ventral side of
the collar lips, which are exceptionally wide in
many torquaratorids. As already mentioned, A.
aurantiacus has a tendency to burrow not seen in
other torquaratorids; this raises the possibility
that its gut contents may include a relatively high
proportion of mineralized sediments. At present,
however, nothing is known about the diet A. aur-
antiacus, because the floating holotype had an
empty gut and the one partially burrowed individ-
ual was captured too badly damaged to yield any
information on gut contents.
Reproduction compared for shallow- and
deep-living enteropneusts. Hadfield (1975), in
his review of reproductive biology in shallow-living
enteropneusts, generalized that the sexes are sepa-
rate, with females having hundreds of small ova-
ries and males having hundreds of small testes
distributed in the anterior trunk region. Depend-
ing on the species, nongerminal ‘‘yolk’’ cells may or
may not be present in the gonads. In females of
species with indirect development, each ovary con-
tains numerous primary oocytes that are relatively
small when spawned (on the order of 100 lm in di-
ameter). In contrast, females of species with direct
development, commonest in the family Harrimanii-
dae, have only a few, large oocytes per ovary (the
largest oocytes of Harrimania kupfferi measure
1,300 lm 3 1,000 lm, the maximum sized oocytes
known for any shallow-water enteropneust). Typi-
cally, when shallow-living enteropneusts spawn,
the gametes pass through small pores in the epi-
dermis and enter the sea water where fertilization
takes place. No specialized copulatory interactions
between males and females have been reported.
Aspects of the reproductive biology for deep-liv-
ing enteropneusts have recently been considered
by Osborn et al. (2012). Many have separate sexes,
but some are hermaphrodites, which is one solu-
tion for facilitating fertilization of sparsely distrib-
uted animals in the deep-sea. Because torquarator-
ids are so sparsely distributed in the deep sea, it
would also not be surprising to find that males
and females associate at spawning time (for exam-
ple, either by copulating or by surrounding them-
selves with a common mucous cocoon) to help
insure fertilization. To date, there is no direct evi-
dence for such interactions.
In torquaratorid species for which the gonadal
morphology has been studied, the ovaries contain
one or a few oocytes with large maximum diame-
ters: 500 lm in Torquarator bullocki (Holland
et al., 2005) and 1,500 lm in both Tergivelum
baldwinae (Holland et al., 2009) and A. aurantia-
cus (present study). The large oocytes of torquara-
torids could indicate that these worms are conven-
tional direct developers. Alternatively, however,
they might be unconventional indirect developers:
that is, the large oocytes of torquaratorids may be
the source of giant (up to 28 mm) Planctosphaera
larvae. Hadfield and Young (1983) have previously
suggested that such larvae might someday be
traced back to the adult stage of enteropneusts liv-
ing in the deep sea. As pointed out by Osborn
et al. (2012), this possibility could be conclusively
tested by preserving specimens of Planctosphaera
appropriately for molecular analysis.
Certainly the most remarkable feature of the
reproductive biology of A. aurantiacus is the exter-
nalization of the ovaries. Each ovary containing a
medium- to large-sized oocyte hangs off of the out-
side of a ripe female in a stalked, hollow bag. The
wall of the bag evidently comprises the remains of
the germinal epithelium closely invested with a
thin-stretched epidermis. Such an arrangement is
without precedent, not only in enteropneusts, but
in animals generally. At present, it is difficult to
imagine a role for the externalized ovaries at mat-
ing time. Perhaps the sperm in this species are
invasive, as they are in salps (Boldrin et al., 2009),
and can burrow through nongerminal tissues to
fertilize female gametes within the mother’s body.
Should this be so, the extreme peripheral location
of the oocytes would shorten the invasion route.
Finally, although, the externalized ovaries of the
holotype contained only oocytes, capture of addi-
tional specimens of A. aurantiacus might reveal
that the fertilized female gametes are retained in
the externalized ovaries and undergo embryonic
development there. Viviparity in enteropneusts
has a precedent; one shallow living harrimaniid
was found brooding a larva within the body of the
mother (Gilchrist, 1925).
ACKNOWLEDGMENTS
The authors thank the pilots and crew of the
MBARI R/V Western Flyer, mother ship for the
ROVs Tiburon and Doc Ricketts, for their skill and
dedication to deep-sea biology. They are also
indebted to Greg Rouse for assistance with the
photography, to Bruce Robison, Bob Vrijenhoek,
and Shannon Johnson for valuable cruise and/or
collections, and to Linda Holland for critical com-
ments on the manuscript.
10 N.D. HOLLAND ET AL.
Journal of Morphology
LITERATURE CITED
Anderson TJ, Przeslawski R, Tran M. 2011. Distribution, abun-
dance and trail characteristics of acorn worms at Australian
continental margins. Deep-Sea Res II 58:970–978.
Barrington EJW. 1940. Observations on feeding and digestion
in Glossobalanus minutus. Q J Microsc Sci 82:227–260.
Barrington EJW. 1965. The Biology of Hemichordata and Proto-
chordata. San Francisco: Freeman.
Benham WB. 1899. Balanoglossus otagoensis, n. sp. Q J Microsc
Sci 42:497–504.
Boldrin F, Martinucci G, Holland LZ, Miller RL, Burighel P.
2009. Internal fertilization in the salp Thalia democratica.
Can J Zool 87:928–940.
Cannon JT, Rychel AL, Eccleston H, Halanych KM, Swalla BJ.
2009. Molecular phylogeny of hemichordates, with updated
status of deep-sea enteropneusts. Mol Phylogenet Evol 52:
17–24.
Cameron CB. 2002. Particle retention and flow in the pharynx
of the enteropneust worm Harrimania planktophilus: The fil-
ter feeding pharynx may have evolved prior to the chordates.
Biol Bull 202:192–200.
Cameron CB, Deland C, Bullock TH. 2010. A revision of the ge-
nus Saccoglossus (Hemichordata: Enteropneusta: Harrimanii-
dae) with taxonomic descriptions of five new species from the
Eastern Pacific. Zootaxa 2483:1–22.
Carey DA, Mayer LM. 1990. Nutrient uptake by a deposit-feed-
ing enteropneust: Nitrogenous sources. Mar Ecol Prog Ser
63:79–84.
Colin PL, Suchanek TH, McMurty G. 1986. Water pumping and
particulate resuspension by calianassids (Crustacea: Thalassi-
nidea) at Enewetak and Bikini Atolls, Marshall Islands. Bull
Mar Sci 38:19–24.
Dobbs FC, Guckert JB. 1988. Microbial food resources of the
macrofaunal-deposit feeder Ptychodera bahamensis (Hemi-
chordata: Enteropneusta). Mar Ecol Prog Ser 45:127–136.
Gaillard C. 1991. Recent organism traces and ichnofacies on the
deep-sea floor off New Caledonia. Palaios 6:302–315.
Gegenbaur C. 1870. Grundzu¨ge der vergleigchenden Anatomie,
Zweite Auflage. Leipzig, Engelmann.
Gilchrist,JDF. 1925. Xenopleura vivipara, g. et sp. n. (Enterop-
neusta). Q J Microsc Sci 69:555–570.
Gonzalez P, Cameron CB. 2009. The gill slits and pre-oral cili-
ary organ of Protoglosssus (Hemichordata: Enteropneusta)
are filter-feeding structures. Biol J Linn Soc 98:898–906.
Hadfield MG. 1975. Hemichordata. In: Giese AC, Pearse JS,
editors. Reproduction in Marine Invertebrates, Vol. II. Ento-
procts and Lesser Coelomates. New York: Academic Press. pp
185–240.
Hadfield MG, Young RE. 1983. Planctosphaera (Hemichordata:
Enteropneusta) in the Pacific Ocean. Mar Biol 73:151–153.
Holland ND, Clague DA, Gordon DP, Gebruk A, Pawson DL,
Vecchione M. 2005. ‘‘Lophenteropneust’’ hypothesis refuted by
collection and photos of new deep-sea hemichordates. Nature
434:374–376.
Holland ND, Jones WJ, Ellena J, Ruhl HA, Smith KL. 2009. A
new deep-sea species of epibenthic acorn worm (Hemichor-
data, Enteropneusta). Zoosystema 31:333–346.
van der Horst CJ. 1934. The burrow of an enteropneust. Nature
134:852.
van der Horst CJ. 1939. Hemichordata. In: Bronn HG, editor.
Klassen und Ordnungen des Tierreichs. Band 4, Abteilung 4,
Buch 2, Teil 2. Leipzig: Akademische Verlagsgesellschaft. pp
1–739.
Hyman LH. 1959. The Invertebrates, Vol. V. Smaller Coelomate
Groups. New York: McGraw-Hill.
Ikeda I. 1908. On the swimming habit of a Japanese enterop-
neust, Glandiceps hacksii Marion. Annot Zool Jpn 9:255–257.
Karrh RR, Miller DC. 1996. Effect of flow and sediment trans-
port on the surface-deposit feeder Saccoglossus kowalevskii.
Mar Ecol Prog Ser 130:125–134.
Knight-Jones EW. 1953. Feeding in Saccoglossus (Enterop-
neusta). Proc Zool Soc Lond 123:637–654.
Miller DC. 1992. Response of Saccoglossus kowalevskii (phylum
Hemichordata, class Enteropneusta) to changes in diet. Mar
Ecol Prog Ser 87:41–54.
Miyamoto N, Saito Y. 2007. Morphology and development of a
new species of Balanoglossus (Hemichordata: Enteropneusta:
Ptychoderidae) from Shimoda, Japan. Zool Sci 24:1278–1285.
Osborn KJ, Kuhnz LA, Priede IG, Urata M, Gebruk AV, Hol-
land ND. 2012. Diversification of acorn worms (Hemichor-
data, Enteropneusta) revealed in the deep sea. Proc R Soc B.
Published on line 16 November 2011, doi: 10.1098/
rspb.2011.1916.
Smith KL, Holland ND, Ruhl HA. 2005. Enteropneust produc-
tion of spiral fecal trails on the deep-sea floor observed with
time-lapse photography. Deep Sea Res I 52:1228–1240.
Spicer SS. 1963. Histochemical differentiation of mammalian
mucopolysaccharides. Ann NY Acad Sci 106:379–388.
Stiasny G. 1910. Zur Kenntniss der Lebensweise von Balano-
glossus clavigerus. Zool Anz 35:561–565.
Thomas IM. 1972. Action of the gut in Saccoglossus otagoensis
(Hemichordata: Enteropneusta). New Zeal J Mar Freshwat
Res 6:560–569.
Woodwick KH, Sensenbaugh T. 1985. Saxipendium coronatum,
new genus, new species (Hemichordata: Enteropneusta): The
unusual spaghetti worms of the Galapagos Rift hydrothermal
vents. Proc Biol Soc Wash 98:351–365.
Yoshimatu S, Nishikawa T. 1999. Swimming swarms of a usu-
ally benthic enteropneust, Glandiceps sp. in the Seto Inland
Sea, Japan, found in 1998. Zool Sci 16:39.
MORPHOLOGY OF A NEW DEEP-SEA ACORN WORM 11
Journal of Morphology