<*« LJ
i/mm
JOURNAL OF SHELLFISH RESEARCH
VOLUME 22, NUMBER 1
JUNE 2003
^ . .1 Laboraic,
AUG 1 2003
Wootis loie, r/..\ U25.J3
The Journal of Shellfish Research
(formerly Proceedings of the National Shellfisheries Association)
is the official publication of the National Shellfisheries Association
Standish K. Allen. Jr. (2004)
Aquaculture Genetics and Breeding
Technology Center
Virginia Institute of Marine Science
College of William and Mary
P.O. Box 1346
Gloucester Point. Virginia 23062
Shirley Baker (2004)
University of Florida
Department of Fisheries and Aquatic Sciences
7922 NW 71- Street
Gainesville, Florida 32653-3071
Bruce Barber (2005)
School of Marine Science
University of Maine
5735 Hitchner Hall
Orono. Maine 04469
Brian Beal (2004)
University of Maine
9 0"Brien Avenue
Machias, Maine 04654
Neil Bourne (2003)
Fisheries and Oceans
Pacific Biological Station
Nanaimo, British Columbia
Canada V9T 6N7
Andrew R. Brand (2003)
University of Liverpool
Port Erin Marine Laboratory
Port Erin, Isle of Man IM9 6JA
United Kingdom
Eugene BuiTcson (2003)
Virginia Institute of Marine Science
P.O. Box 1346
Rt. 1208 Create Road
College of William and Mary
Gloucester Point, Virginia 23062
Editor
Sandra E. Shumway
Department of Marine Sciences
University of Connecticut
Groton. CT 06340
EDITORIAL BOARD
Peter Cook (2004)
Austral Marine Services
Lot 34 Rocky Crossing Road
Warrenup
Albany, W.A. 6330. Australia
Simon Cragg (2004)
Institute of Marine Sciences
University of Portsmouth
Ferry Road
Portsmouth P04 9LY
United Kmgdom
Leroy Creswell (2003)
University of Florida/Sea Grant
8400 Picos Road. Suite 101
Fort Pierce, Florida 34945-3045
Lou D'Abranio (2004)
Mississippi State University
Department of Wildlife and Fisheries
Box 9690
Mississippi State, Mississippi 39762
Christopher V. Davis (2004)
Pemaquid Oyster Company. Inc.
P.O. Box 302
1957 Friendship Road
Waldoboro. Maine 04572
Ralph Elston (2003)
Aqua Technics/Pacific Shellfish Institute
455 West Bell Street
Sequim, Washington 98382
Susan E. Ford (2004)
Rutgers University
Haskin Shellfish Research Laboratory
6959 Miller Avenue
Port Norris, New Jersey 08349
Raymond Grizzle (2003)
Jackson Estuarine Laboratory
Durham, New Hampshire 03824
Karolyn Mueller Hansen (2004)
1524 Barley Circle
Knoxville, Tennessee 37922
Journal of Shellfish Research
Volume 22, Number 1
ISSN: 0730-8000
June 2003
Mark Luckenbach (2003)
Virginia Institute of Marine Science
Eastern Shore Lab
P.O. Box 350
Wachapreague, Virginia 23480
Bruce MacDonald (2004)
Department of Biology
University of New Brunswick
Saint John, New Brunswick
Canada E2L 4L5
Roger Mann (2004)
Virginia Institute of Marine Science
Gloucester Point, Virginia 23062
Islay D. Marsden (2004)
Department of Zoology
Canterbury University
Christchurch, New Zealand
Jay Parsons (2005)
Memorial University
Marine Institute
Box 4920
St. John's, Newfoundland
Canada AlC 5R3
Tom Soniat (2004)
Biology Department
Nicholls State University
Thibodaux, Louisiana 70310
J. Evan Ward (2004)
Department of Marine Sciences
University of Connecticut
1080 Shennecossett Road
Groton. Connecticut 06340-6097
Gary Wikfors (2004)
NOAA/NMFS
Rogers Avenue
Milford. Connecticut 06460
www.shellfish.org/pubs/jsr.htm
./,)/(/■//((/ „f Slwlljlsh Rcst'tinh. Vol. 22. No. 1. 1-20. 2003.
A REVIEW OF PUBLISHED WORK ON CRASSOSTREA ARIAKENSIS
MINGFANG ZHOU AND STANDISH K. ALLEN, JR.*
Aciiuwulture Genelics ami Breeding Technology Center. Virginia Institute of Marine Science. P.O. Box
1346. Gloucester Point. Virginia
INTRODUCTION
NOMENCLATURE
Field research on the Asian (Suminoe) oyster. C. ariakensis.
began in 1998 at the Virginia Institute (if Marine Science (VIMS)
in response to a resolution from the Virginia Legislature to initiate
investigations on alternative species. All field trials have used
sterile triploids. Initial research indicated promising performances
in C. ariakensis in a variety of salinities for growth and disease
resistance (Calvo et al. 2001). Research on this species cxmtinues
at VIMS today, but in the meantime, the Virginia Seafood Council
has run two commercial trials of C. ariakensis on their own v\ ith
similar promising results. They have proposed a third for 2003
with about a million triploid C. ariakensis. The direction taken by
industry clearly indicates a desire to proceed with larger and larger
scale-ups of aquaculture using triploids. This notion was addressed
in a symposium staged m 2001 (Hallerman et al. 2002) where the
general consensus found that "it is difficult to consider the risks of
aquaculture of triploid (infertile) C. ariakensis as separate from the
risks of diploid (fertile) C. ariakensis. That is. there was consensus
that triploid aquaculture would inevitably lead to some introduc-
tion of reproductive individuals in the Bay. with unknown out-
comes for population growth." Part of the difficulty in assessing
the risk of such a scenario comes from the inherent difficulty of
predicting the consequences of an introduction generally. Another
difficulty of assessing risk, especially for C. ariakensis. is the lack
of information on this species.
The aim of this review was to provide an unabridged overview
of the published works on this species. We may have missed some
references that were obscure or indirectly referred to C. ariakensis.
Many of the works on C. ariakensis were in other languages,
principally Chinese. For Chinese articles, they were translated and
are presented in somewhat more detail than those in English. Some
were obtained while traveling to specific laboratories in China and
would otherwise be difficult to obtain. We were as complete as
possible give the timely need for this review.
We present the information uncritically. That is. we present the
contents of the articles without analysis. Partly this is the result of
space constraints. More importantly, it is unclear that data reported
always apply to C. ariakensis. Morphologic confusion is common
with Crassostrea species. For example, a considerable number of
reports of C. ariakensis occur in west India and Pakistan, geo-
graphically isolated from the main populations in Japan. China,
and Korea. It seems unlikely that this is the same species, but to
judge so a priori would be to leave out this information. We e.xpect
scientists to consider the data critically and test it if appropriate.
The information we collected is organized into general catego-
ries so that one work may be cited repeatedly if it crosses catego-
ries. The content in each category in no way implies the impor-
tance of this information, merely what has been done. Conversely,
categories missing information reflect the absence of data.
*Corresponding author. E-mail: ska@vims.edu
Harry ( 1981 ) described the history of the genus name Crassos-
trea Sacco. 1 897 as follows: Over half a century ago Lamy ( 1929-
1930) surveyed the living oysters and put all species in the genus
Ostrea Linnaeus. 1758. including Crassostrea ariakensis. But
since 1930. other authors, chiefly those interested in the commer-
cial production of oysters (e.g.. Thompson 1954). have separated
Cras.wstrea from Ostrea on the basis that the proniyal passage on
the right side of the excurrent mantle chamber is closed in Ostrea
and open in Crassostrea. Other differences on morphology and
anatomy between these two genera can be found in Ahmed (1971
and 1975). Glude( 1971). and Stenzel (1971). In this review, please
note that Ostrea is cited from many old references.
Nomenclature is confusing for C. ariakensis (Carriker &
Gaffney 1996) because the traditional oyster classification meth-
ods rely mainly on conchological characters, i.e.. external and
internal morphology of the shell, which express high phenotypic
plasticity among environments (Hirase 1930). In addition, oyster
eggs are fertilized in mass spawns that increase the possibility of
hybridization and promote high variation (Guan & Li 1986).
Therefore, species with the same name might be genetically dis-
tinct whereas the ones with different scientific names might be
genetically the same. Species variously called C. rivularis, dis-
coidea. palmipes. or paiiliiceiae in previous literature (Carriker &
Gaffney 1996) might be the same as the species we call C. ariak-
ensis today. In general, it is accepted that rivularis is synonymous
with ariakensis. although it is still possible that rivularis/ ariak-
ensis was misclassified in certain publications. This review in-
cludes all the available publications with the above mentioned
species names.
The authorship of ariakensis has been credited to Fujita (1913).
However, we are confused by the description of Wakiya (1930) on
the origin of the name ariakensis. He wrote his reference as "O.
ariakensis (Wakiya M. S.) Fujita. ... 1913." Harry ( 1981) assumed
that "Fujita proposed the name in 1913. based on a manuscript of
Wakiya." Coan et al. (1995) seemed to agree by giving the refer-
ence in a way of "Fujita. 1913... e.x Wakiya MS." Who proposed
the name ariakensis first, Fujita or Wakiya? We were not able to
locate Fujita (1913), so we cannot answer that question for sure.
According to our publication collection, the species name aria-
kensis was not referred to as frequently as rivularis before mid
1990s, but it has been widely referred to in recent publications.
The history of species name rivularis can be traced back to
1 861 . when Gould described a new species called Ostrea rivularis,
which in Latin means "oysters in small brooks." His original de-
scription was written in Latin. Translated to English, the shells he
observed were "discoid, oblong, slender; inferior valve thick,
purple, with remotely radiate ribs and fortified small mbes: supe-
rior valve simple, with ramosing less purple veins; cavity mini-
mally deep, ovate; white ash-colored broad margin, weak hinge."
He emphasized "the rays of the little tubes below, and the veins
Zhou and Allen
above, are uniiMially clear, distinctive ciiaracters." The dimension
of the observed shells was "Diam. 60; Lat. 10 millim." It "inhabits
the China Seas, as indicated by shells adhering to it."
There is serious ambiguity in the source of Gould's specimen.
The title of his article indicates that his description was based on
the collection of "the North Pacific Exploring Expedition,"
whereas according to Hirase (1930), it was based on a single
specimen from China in Dunker's collection. Hirase did not ex-
plain whom Dunker is except for a reference listed as Dunker
(1882). Several other authors mentioned China as the source of
Gould's specimen (Ahmed 1971, Galtsoff 1964), but no additional
references were offered for further confirmation. Hirase (1930)
also questioned the completeness of Gould's description and its
value for identification because it seems based on a single speci-
men, which seems to be comparatively young according to its size
(60 mm). Gould's description of rividaris and those of others (see
Morphology section) are incompatible. Thus, it is quite possible
that nvitUms of Gould ( 1861 ) is different from the species we call
rividaris or ariakensis today.
O. (C ) rividaris Gould has been widely applied to oysters with
similar conchological characters in many Pacific coastal countries,
such as Japan. China, Pakistan, and India. Its taxonomic status in
each country is still muddled. A review is summarized below.
In Japan, Ariake-gaki, Suminoe-gaki, and Kaki ("gaki" in Japa-
nese means oyster) were some common names for O. rivtilaris
(Amemiya 1928). This species was once classified as O. gii^ax by
Fujimori (1929) but this was refuted by Taki ( 1933) and Imai and
Hatanaka (1949). Wakiya (1930) surmised that O. rividaris of his
in 1915 (Wakiya 1915) and that of Amemiya (1928) was the same
as O. ariakensis. whereas the O. rivularis described by Lischke
(1871) seemed to be the young of O. ariakensis.
In Pakistan, Awati and Rai ( 1931 ) indicated two names lor the
same species, O. discoidea and O. rividaris. Reeve (1871) de-
scribed O. discoidea based on specimens from Fuji Island and New
Zealand, but Ahmed (1971) stated that the figure and the shell
characters published by Reeve were different from that of O. dis-
coidea. According to Ahmed, Reeve's O. discoidea is rounded and
flat to the extent that it looks like the windowpane oyster, Placuna
placenta Linne, 1758, which is abundant in lagoons of Philippines
and South East Asia (Abbott & Dance 1986). Based on his own
experience, Ahmed believed that O. discoidea is not distinguish-
able from C. rividaris.
In China, the common name for O. (C.I rivularis is Jinjiang-
muli ("jinjiang" in Chinese means "close to river" and "muli"
means oyster). One of the long-standing debates on oyster classi-
fication involves two morphologically very similar variants that
occur in the Peari (Zhujiang) River estuary. One is called "white
meat" oyster and the other is "red meat." Very experienced oyster
farmers can separate these two variants by external appearance and
the color of the soft body. Fei ( 1928) believed that both are O.
gigas. However, Zhang and Lou (1956a) identified "white meat"
as O. rivularis and "red meat" as a variant. The "white meat"
oyster is considered better than the "red meat" because of meat
quality and productivity in aquaculture, thus has higher commer-
cial value. The "red meat" oyster is apparently more resistant to
harsh conditions according to observations of it in culture (Guan &
Li 1986). Further investigations by other researchers revealed
other differences. A comparative study on the physiologic and
biochemical indexes (Guan & Li 1986), such as oxygen consump-
tion rate, fatty acid composition, and amino acid composition,
demonstrated sufficient differences in physiology to suspect that
genetic differences are likely. Anatomically, Li (1989) found a
difference in the connection of the body with the gills. In "white
meat" both the left and right epibranchial chamber connect with
the promyal chamber, whereas in "red meat" only the right epi-
branchial chamber connects with the promyal chamber. He be-
lieved the two belong to two different species. A study on genetic
variation using starch gel electrophoresis (Li et al. 1988) demon-
strated that they should belong to different species because their
genetic identity was low (I = 0.548). The estimated divergence
time of the two is 3 x 10" years. The comparison of genetic
similarities and genetic distances suggests that "white meat" is C.
rivularis and "red meat" is probably C. iredalei. Guan and Zheng
( 1990) studied the esterase isoenzyme of the two groups by poly-
propylene amide gel electrophoresis and agreed that they are dif-
ferent species. Above all, it was generally agreed that "white meat"
is C. rivularis. but whether "red meat" is C. iredalei is still un-
confirmed.
MORPHOLOGY
Conchological Characters
References on conchological characters of naturally occurring
C. ariakensis come from three countries: China, Japan, and India.
References from the United States (Pacific Northwest) are also
included because the seed were introduced from Japan. Reports
containing conchological data are listed individually following a
general review to compare and contrast characters of what are
called O. (C.) rivularis, now C. ariakensis. The major conchologi-
cal characters presented in these reports are size; thickness and
shape of the valves; outer structure of the valves; comparison
between the left and right valve; color of outer and inner surface;
size and color of ligament; color, size and position of the muscle
scar; and hinge structure (Table I ).
Review
In China, it is commonly observed that valves of Ostrea (Cras-
saslrea) rivularis are large and thick with varying shapes, basically
round but sometimes elongated into oval, oblong, and even trian-
gular shapes. The right valve is thinner, flatter, and smaller than
the left. Both valves are covered with concentric lamellae (fluted
shell margins on the external shell), with fewer layers of but
stronger, lamellae on the left valve. Density and shape of lamellae
varies by age class, which are thicker and more layered in older
oysters (Zhang & Lou 1956a, Zhang et al. 1960). Color of lamellae
or the outer surface of valves ranged from gray, yellowish brown,
brown, to purple or dark purple. Dark purple coloration is apparent
in C. ariakensis grown in high-salinity areas of Chesapeake Bay
(Zhou & Allen, unpubl.). The inner surface of valves is white or
grayish white, purple on the edge. The ligament area is short and
wide, and the ligament is usually purple black. The muscle scar is
very large, mostly oval or kidney shaped, located in the mid-
dorsal area, purple or light yellow in color.
The coloration of valves and muscle scars of C. ariakensis
described in reports from Japan is different from those from China.
In Japan, the outer surface of the valves was described as cream-
buff or white, streaked with radial chocolate bands, violet bands, or
almost uniformly violet (Hirase 1930, Torigoe, 1981, Wakiya
1929). The inner surface of the valves was strongly lustrous or
partly opalescent (Hirase, 1930, Torigoe 1981). The muscle scar
was usually white or sometimes stained with olive-ocher spots or
CRASSOSTREA ARIAKENSIS REVIEW
TABLE 1.
Characteristics of oysters by citation.
Gould (1861). China. O. rmilahx
Valve shape size: Discoid, ohlong. slender.
Left, righl xalve: Inferior valve thick, purple, with remotely radiating ribs and fortified small lubes; superior valve simple, with ramosing less purple veins;
cavity minimally deep, ovate.
Shell color outer; Purple; white ash-colored broad margin.
Shell color inner; —
Ligament: —
Muscle scar: —
Hinge; Weak.
Zhang and Lou (I'J.Wl. China, O. (C.) nviilaris. includes figurelsl
Valve shape size: Large and thick with various shapes, round, oval, triangle, and oblong; concentric scarce lamellae on outer surface.
Left, right valve: —
Shell color outer; Yellowish brown.
Shell color inner; —
Ligament: —
Muscle scar; —
Hinge: —
Zhang and Lou I I956al. China. O. iC.I rivularis. includes l'igure(s)
Zhang et al. {I960). South China. O. rivuluris. includes figurets)
Similar descriptions from the above two references are combined as Ibllows.
Valve shape size: Valves large and thick with various shapes, round, oval, triangle, and oblong.
Left, right valve; Right valve flatter and smaller than the left one, with yellowish brown or dark purple concentric lamellae on its surface. In 1 to 2-y-old
individuals, lamellae thin, flat, and brittle, sometimes dissociated; on valves older than 2 ys old, flat but sometimes with tiny wavy
shape at the edge; on valves several years old. thickly layered, strong as stone. Left valve is larger and thicker than right valve,
stronger but fewer layers of lamellae. A few samples had inconspicuous radiating ribs or plication.
Shell color outer: Gray, purple, or brown.
Shell color inner; White, grayish purple on the edge.
Ligament: Ligament purple black. Ligament groove shon and wide, like an o,\ horn. The length from the ligament to anterior is one sixth to one
fourth of shell height.
Muscle scar; Muscle scar very large, light yellow, irregular shape, mostly oval or kidney shaped, located in the middle of the dorsal area.
Hinge; —
Cai et al. (1979), China, O. rivularis. includes figure(s)
Valve shape size; Shells large and thick with various shapes, such as round, oval, triangle and oblong.
Left, nght valve; Right shell (latter and smaller than the left shell, with yellowish brown or dark purple lamellae on its surface. The lamellae are thin and
flat, with not much layers and no radiating ribs, but usually with protuberance. The left shell is larger and thicker with irregular shape
and similar lamellae as the right shell.
Shell color outer; Yellowish brown or dark purple.
Shell color inner; White or grayish white.
Ligament: Ligament purple black, ligament groove short and wide.
Muscle scar; Muscle scar large, oval or kidney shaped, located in the middle of the dorsal area.
Hinge: No denticulate on the hinge.
Li and Qi 1 I994i. China. C rivularis. includes figurelsl
Valve shape size; Large variation in shell shape, usually oval or oblong.
Left, right valve; Concentric lamellae tend to coalesce, no radiant ribs.
Shell color outer; Light purple.
Shell color inner: White.
Ligament; Wide ligament groove.
Muscle scar: Light purple.
Hinge; —
Amemiya ( 1928). Japan. O. rivularis. includes figurets)
Valve shape size; It is either circular or oval in form, pronounced elongation as found in O. gigas is absent.
Left, right valve; —
Shell color outer: —
Shell color inner: —
Ligament: —
Muscle scar: —
Hinge: —
Cahn (1950), Japan. O. rivularis. includes riguretsi
Valve shape size: Round, Hat. smooth surfaced, plates thin, almost smooth, shell thick.
Left, right valve; —
Shell color outer; Pale pink, radiating burnt lake strikes on shells.
Shell color inner: —
Ligament: —
Muscle scar; —
Hinge; —
Hirase (I9.WI. Japan. O. (C.) rivularis. includes rigure(s)
Valve shape size: Orbicular, oval, elongated oval, though appearing somewhat subtriangular because of its rather long umbo. There are many intermediate
forms, but on the whole the specimens are oval. The shell is fairly strong and thick, though not to the extent of C. gigas.
continued on next page
Zhou and Allen
TABLE 1.
continued
Shell color outer:
Shell color inner:
Ligament:
Muscle scar:
Left, right vahe: The right valve is somewhat smaller. The conca\il> ot the Icit \alve is larger. The amerior depression of the left valve is very obscure. The
lamellae of the right valve are somewhat thin and almost smooth, and distinct placations are not apparent, but sometimes the lamellae
are covered with somewhat irregularly tubular projections. It is noteworthy that smooth lamellae are more common in the young than in
the adult. The color is cream-buff with many radial chocolate bands, but in adults these bands are fused into larger ones; their
arrangement differs in each individual. In the left valve, the lamellae are generally indistinct, and may be close together or separate.
The common color is pale rhodonite pink with radiating "burnt lake" striae.
The inner shell surface is generally white with strong luster, sometimes with a yellowish central part.
The ligament is "burnt lake" or black.
The muscular impressions, elongated oblong with concave anterior side, are equal in size for the two valves and rather large in porportion
to the inner shell area. The color of the impression is while, or rarely marked with olive-ocher spots; its surface is almost Mat.
Hmge:
Imai (1978), Japan. C. rivularis, includes figure(s)
Valve shape size: Round or elliptical
Left, right valve: The lower shell is shallow and the umbo cavity below the hinge plate is very small.
Shell color outer: The part near the hinge plate m the upper shell is violet-brown in color.
Shell color inner: —
Ligament: —
Muscle scar: —
Hinge: —
Kira (1962). Japan, C rivularis
Valve shape size: Has a large and rather flat shell, oi v\hich the surface bears very coarse and widely spaced concentric lamellae.
Left, right valve: —
Shell color outer: —
Shell color inner: —
Ligament: —
Muscle scar: —
Hinge: —
Torigoc (1981), Japan, C ariakensis, includes Figure(s)
Large sized (height 200 mm x length 1 12 mm, Hirase 1930). Outline orbicular to long spatulate form, mostly tongue form, subequivalves.
Attachment area is small to medium, commonly behind the umbonal area.
Both valves flat, but left valve weakly concave. Both valves have very faint dichotomous radial ribs, left valve more conspicuous than right
valve. Growth squamae flat and stretched parallel to the grow lines. No commissural plication, or very weak even if present.
Commissural shelf small to medium. Umbonal cavity shallow. No chomata. The dorso-ventral section has chalky deposits between soHd
shell layers and no hollow chambers. Both valves are thinner than those of C. gigas. so chalky layers are very thin. The parts of chalky
deposits are often intruded by worms.
"White in ground" (sic) color with pale purple streaks radiating from umbo.
Chalky white or partly opalescent.
Valve shape size:
Left, right valve:
Shell color outer;
Shell color inner:
Ligament:
Muscle scar:
Reniform. dorse -an ten or border concaved and close to ventro-posterior shell margin from the center ot the valve. Lustrous while or
sometimes with purple patches, particularly on nght valve.
Hinge; —
Wakiya (1929). Japan, Osirea ariakensis
Valve shape size;
Shell usually circular or oval in shape. However, its shape varies considerably according to the hardness of the bottom on which it lives.
When found imbedded in soft mud it has an extremely elongated shell so that it is very difficult to distinguish it from that of O.
Inperousi found on a mud bottom of lower salinity, only differing from O. kiperoiisi in having the hinge of lower valve not very long
and subequal to that of the upper one. O. rivularis Gould has. according to the original description, its lower valve provided with
radiating, tube-shaped ribs set distantly. Therefore the species in which the ribs are absent from the lower valve or only very weakly
developed, if present at all. cannot be the species of Gould.
Lamellae imbricated rather compactly, lower valve concave, not provided with ribs; upper valve flat, length of hinge nearly equal to that ot
lower valve. Occasionally, weakly developed ribs are observed on the lower valve of the young of the species, but never on full-grown
ones.
Whitish and streaked with violet, or almost uniformly violet.
Lead white; muscular impression faint, usually not specially colored but sometimes stained purple.
The hinge of the lower valve not so long as. as long as or a little longer than the breadth; no umbonal cavity below margin of hinge.
USA. C. ariakensis. includes fi2ure(s)
Left, right valve:
Shell color outer:
Shell color inner:
Ligament;
Muscle scar:
Hinge:
Coan et al. (1995
Valve shape size; Subtrigonal. flared ventrally, heavier and more rounded than C. gigas.
Left, right valve; Left valve moderately concave, with white to pale pink lamellae; right valve moderately flattend, with many thin commarginal lamellae,
sometimes with dark brown to purple radial color bands. Both valves with densely layered, thin lamellae.
Shell color outer; —
Shell color inner: —
Ligament; —
Muscle scar: White to purple to olive.
Hinge; —
Galtsotf (1964). USA, C. rivularis, includes figurels)
Valve shape size: Orbicular strong and large.
Left, right valve: Left, lower valve slightly concave, upper valve shorter and flat. The left valve has generally indistinct lamellae of pale pink color with
radiating striae. The lamellae of the right valve are thin and most smooth, sometimes covered with tubular projections.
continued on next page
Crassostrea ariakensis Review
TABLK 1.
continued
The color ol the right \alve is LTcaiii hiilt wilh nuiny radial chocolate bands, their arrangements greatly variable.
Situated near the eenler or a little dorsally. is while, occasionally with olive-ochre spots.
Shell color outer;
Shell color inner:
Ligament:
Muscle scar:
Hinge: —
Langdon and Robinson ( 1*^%), USA. C. ariakensis. includes figure{s)
Valve shape size: This species differs from the Pacific oyster morphologically in that the shell is typically more rounded and the edges of shell layers are llal
and no! rippled like those of Pacific oysters (Torigo. 1981 i
Left, right valve: —
Shell color outer: —
Shell color inner: —
Ligameni; —
Muscle scar: —
Hinge: —
Awali and Rai (1931). India. O. discnidea or O. rivularis
Valve shape size:
Left, right valve:
Shell color outer:
Shell color inner:
Ligament:
Muscle scar:
Hinee:
Shell flat and of large size, rounded, foliaceous with conspicuous lines of growth.
Lower valve lightly concave, upper valve of the same size and shape as the lower, slightly convex.
Clear and nacreous.
Ligament area small.
Oblong with a cloudy white or smoky white color.
No denticulations.
Rao (1987). India. C. rivularis. includes figure(s)
Valve shape size:
Left, right valve:
Shell color outer:
Shallow shell cavity
Imai (1978) has slated that the hinge part of the shell of C nvtilaris is violet brown in color. The coloration may be caused by ecological
conditions such as luxuriant growth of seaweeds in the vicinity or other factors and should not be considered of taxonomtc importance.
Shell color inner: —
Ligament: —
Muscle scar: Oblong white.
Hinge: —
Palel and Jetani (1991), India. C. rivularis
Valve shape size:
Left, right valve:
Shell color outer:
Shell color inner:
Ligament:
Muscle scar:
Hinee:
Shell oval, narrow at anterior end and broader with posterior end.
Left valve has deep radial ndges from the hinge and tightly inter locked with upper right valve.
Pink to brownish with tints.
Having narrow hinge-ligament
White.
Having narrow hinge-tigament.
purple patches (Hirase 1930. Torigoe 1981, Wakiya 1929). Rao
(1987) thought the difference in coloration might be caused by
ecological conditions and therefore not considered a character of
taxonomic importance. Reports from the United States are consistent
with reports from Japan for coloration, which indicates that at least
some part of coloration might be caused by genetic factors. O. (C.)
rivularis from India are similarly described. Coloration of the inner
surface of the vahes and the muscle scar are close to Japanese reports.
Reports from Japan were often comparative between C. ariak-
ensis and other species, such as O. (C.) gigas (Amemiya 1928. Hirase
1930. Torigoe. 1981) and O. lopenmsi (Wakiya 1929). O. (C.) gigas
were believed to have stronger, thicker, and more elongated shells
than O. (C.) rivularis. whereas O. rivularis is very difficult to distin-
guish from O. lapennisi foLind on muddy bottom in lower salinity. O.
rivularis differs from O. lapcmusi by having the hinge of the lower
valve not very long and subequal to that of the upper one. Japanese
reports agree that O. IC.) ariakensis has flat valves, with the left one
weakly concave (Cahn 1930. Kira 1962. Torigoe 1981). Wakiya
( 1929) thought the various shapes of O. ariakensis were influenced by
the hardness of the bottom because the ones with extremely elongated
shells were found imbedded in soft mud. This is also a character of
other Crassostrea spp. (Galstoff 1964).
The most confusing character through this review has been
what Gould (1861). who first named O. rivularis. described as
remotely radiating ribs and fortified small tubes on the outer sur-
face of left valve and veins on right valve. He emphasized that
these are usually clear, distinctive characters of this species. His
observation was based on a sample from China. However, no
reports from China agreed with his description of such characters.
Cai et al. ( 1979) and Li and Qi (1994) observed no radiating ribs
in this species. Based on a large-scale investigation of oyster spe-
cies all along the Chinese coast. Zhang and Lou ( 1956a) described
inconspicuous radiating ribs or plication in a few samples of O.
(C.) rivularis. Only one report from India described deep radial
ridges from the hinge on the left valve (Patel & Jetani 1991).
although the origin of the background specimen was unknown.
From Japan, similar characteristics were described as indistinctive
or occurring at very low frequency. Hirase (1930) and Galstoff
(1964) mention that the lamellae are sometimes covered with tu-
bular projections. Hirase (1930) and Cahn (1950) mentioned "ra-
diating burnt lake strikes," which might or might not be the same
feature we are discussing here. Torigoe's ( 1981 ) report said "both
valves have very faint dichotomous radial ribs, left valve more
conspicuous than right valve." Wakiya ( 1 929) is more helpful in
clarifying this confusion. He stated this species was "not provided
with ribs... occasionally, weakly developed ribs are obser\'ed on
the lower valve of young of the species (Ostrea ariakensis). but
never on full-grown ones." Either Gould's original descriptions
6 Zhou and Allen
were inappropriate for adult C. ariakensis. or he described a dif- GEOGRAPHIC DISTRIBUTION
ferent species ( Wakiya 1929). The latter possibility is quite high if
Gould did get his specimen from China because there are around ^''""^ ^" overview of the literature. C. ariakensi.s seems to have
20 oyster species there (Zhang & Lou 1956b. Cai & Li 1990, Li & ^ ^''^^ geographical range. According to Kuroda and Habe 1 1952).
Qi 1994. Guo et al. 1999). and classification based completely on ^^ '■""/<"■" encompassed latitudes 12-34'N. which covers the
morphologic characters is questionable. ^'"'^^ *ro'" southern Japan to southern India. Ranson (1967) listed
sources of C. ahakensis specimens in museums around the world.
ANATOMIC CHARACTERS
coming from Southern Japan to coasts bordering the South China
Sea. including Hong Kong. Vietnam, and Sabah (formerly North
Borneo), Malaysia. Several authors (Wakiya 1929, Cahn 1950,
Review Kira 1962, Coan et al. 1995) mentioned its distribution in Korea.
Anon (1996) mentioned that C. rivutaris was also found from the
Anatomic characters were not studied as broadly and com- Philippines and Taiwan to Thailand. Above all, this species seems
pletely as conchological ones. Reports mainly come froin Japan to occur all along the west coast of the Pacific Ocean, from south-
and China. Researchers had different emphases in their anatomic em Japan to Pakistan (Angell 1986). Sparks ( 1965) even reported
studies. The only character described by more than one researcher that C. rivtilaris was indigenous to Kenya. However, for most
is the mantle. Hirase (1930). Zhang et al. (1960), and Galtsoff areas outside of Japan and China, no references are available to
( 1964) were in agreement that the inner row of the mantle tentacles confinn these observations genetically as C. aiiakensis.
is aligned while the outer row is iiregular. Details of anatomic Quite a few literature reports are available listing specific lo-
characters are given in Table 2. cations in a country where this species occurs naturally. Below we
TABLE 2.
Anatomical characteristics of oysters by citation.
Hirase (1930). Japan, O. (C.) hvulahs
Mantle — In a specimen whose length and altitude are 96 mm and 45 mm. respectively, the mantle is united by the anterior 21 mm. or 0.22 of the
body length. There is no siphon. The mantle margin is dark nigrosine violet or pinkish vinaceous, and the tentacles are arranged in two rows,
the outer consisting of tentacles of irregular size and the inner of slender single tentacles. Fine tendons radiate from the posterior sides of the
adductor muscle as usiLal.
Adductor muscle — The adductor muscle measures 20 mm m altitude and 22 mm in breadth and is suborhicular. with somewhat concave anterior
face and convex posterior face. The distance between the anterior end of the adductor muscle and the anterior end of the body is 52 mm. A
small portion of the posterior part of the adductor muscle is white as usual.
Heiin — The pericardium, continguous to the anterior face of the adductor muscle, is oval and measures 19 mm in altitude and ti m in breadth. The
heart runs obliquely from the antero-dorsal corner of the pericardium to the postero-ventral corner. The ventricle and the auricles are both tlesh
color. The ventricle measures 8 mm m altitude and 6 mm in breadth, while one of the auricles measures 8 mm in altitude and 3 mm in breadth.
Ctenidium — The posterior end of the ctendium curls up along the posterior face of the adductor muscle.
Alimentary system — The palps are as usually found in Crassostrea. The rectum begins at the dorsal region of the pericardium and ends just above
the posterior end of the adductor muscle. About 3 mm of the terminal portion is free, differing from other oysters of this subgenus and shorter
than in Neopycnodonte cochlear, whose free portion is 5 mm. The anal end has a ring. The distance between the mouth and the anus is 55 mm,
its ratio to body length being 0.57.
Imai (1978). Japan. C. rivularis
C. ariakensis differs from C. gigas in that a part of the rectum and anus are away for the soft parts.
Torigoc (1981), Japan, Crassostrea ariakensis
Soft parts are similar to C. gigas but the coloration of soft parts is the palest of Japanese Crassostrea species.
Zhang et al. (1960), South China, 0. rivularis
Mantle — The inner row of the mantle tentacles is aligned while the outer row is irregular.
Heart — Heart chamber is flesh pink.
Li (1989), China. C rivularis
Promyal chamber — The left and right cpibranchial chambers connect with the promyal chamber all together. In the cross section of this type, the
ascending lamellae of the left and right outer demibranch attach to the mantel, whereas the other part of gills are free in the mantel cavity. The
whole epibranchial chamber is connected with the promyal chamber. On the lateral view from the right side of the oyster, the joint of the two
gills attaches to the visceral mass at and below the adductor muscle, while above the adductor muscle, the gills are dissociated so that the two
rows of water tubes on the left as well as the two rows on the right of oyster body can be seen. The "white meat" Jinjiang oyster from
Shenzhen Bay belongs to this group.
Nelson (1938) stated that oysters with a promyal chamber are adapted to low salinity and highly turbid waters, while oysters without it do
better in high salinity, less turbid waters. Thomson (1954) had similar reports. The occurrence of the promyal construction in commonly
cultured oyster species in China and their distribution are consistent with Nelson's statement. Oysters with the chamber inhabit mostly estuary
and intertidal zones, where salinity and transparency are both low and the environmental factors tluctuate. The ones without the chamber inhabit
mosdy shallow seas with higher salinity and relatively stable environments. It is likely that the promyal chamber is an adaptation stemming
from oysters moving into increasingly estuarine habitats.
Galsoff (19641. USA. C rivularis
Mantle — Margin of the mantle is dark \ uilcl; the tentacles are arranged in two rows; those of the outer row are of irregular size; the inner
tentacles in a single row are slender.
Crassostrea ariakens/s Review
summarize this intormation by country, Irom iiortii to south alony
the Pacific west coast.
Japiiii
Kira (1962) reported distribution of C. riviilans roughly from
central Honshu to Kyushu (Fig. 1). Honshu is the largest island of
Japan located in the center of the archipeligo. Kyushu is southern
most. Cahn (1950) reported the restricted range of its distribution
as western Kyushu, mainly in Ariake-kai C'kai" in Japanese means
sea) and Yatsuchiro-wan ("wan" means bay). It is most abundant
in the inner parts of Ariake-kai. the southern coast of Fukuoka and
Saga prefecture. Hedgecock et al. (1999) found a similar distribu-
Honshu Islantd
Pacific Ocean
'Kyushu Island
East China Sea
^
f
Figuri' 1. Locations reported v\ith C. ariakensis popiilutions in .lapan. 1. Ariake-kai; 2. \atsuchiro-\\an; .^. Fukuoka prefecture: 4. Saga
prefecture; 5. Shiranuhi Bay: 6. Kochi prefecture: 7. \ amaguchi prefecture; and 8. Okayama prefecture.
Zhou and Allen
tion in the Ariake Bay. Ariake-kai or commonly called Ariake
Bay, seems to be the most recognized natural habitat and the
namesake of C. ciriakcnsis. as it was mentioned most frequently
(Wakiya 1929. Hirase 1930. Cahn 1950, Galtsoff 1964. Imai 1978.
Hedgecock et al. 1999). In addition. Wakiya (1929) mentioned
Shiranuhi Bay on the northeastern coast of Kyushu, and Cahn
( 1950) listed the Pacific coast of Kochi. the coast of Yamaguchi
and Okayama prefecture.
China
China has an extensive coastline of about 18.000 km extending
from the cold temperate north to the tropical south. Based on an
extensive investigation on oyster species along the Chinese coast
in 1956, O. (C.) liviilaris was identified in each coastal province
(Zhang & Lou 1959: Fig. 2). As Zhang et al. (1960) later stated,
the distribution of this species covers the whole coastal region of
China, with a latitudinal range of 15-40°N and a longitudinal
range of 107-1 24'E. Table 3 lists the names of locations where
O. (C.) vivulaiis has been reported. The locations underlined were
considered by Zhang and Lou (1956b) as major production
areas, which might not be true today. Among those. Xiaoqing
River estuary in Yangjiaogou, Shandong province was specifi-
cally mentioned because a very large population of O. rividaris
was found there. In certain localities, the population was so large
that people call them "oyster hills" because individual oysters
grew attaching to each other (Zhang & Lou 1956b, Zhang et al.
1960). It would be interesting to try to determine whether natural
populations are still available in some locations, having possibly
been shielded from exploitation because of their rarity (Table 3).
India
Although Ahmed (1971) mentioned that C. riridaris was dis-
tributed on both east and west coasts of the Indo-Pakistan subcon-
tinent, other reports maintained that this species was found only on
the west coast of India (Fig. 3). It was first reported along the coast
of Bombay (Awati & Rai 1931). Durve (1986) gave a much wider
range between Ratnagiri and Okha along the coast of Gujarat and
Maharashtra area. Gujarat (Saurashtra) has a long coastline of
1500 km (Patel & Jetani 1991). Specific locations in this range
were described by Mahadevan (1987) as Aramra, Poshetra, Port
Okha, Porbandar, Sikka. Gagwa Creek. Singach Creek. Beet Kada.
Khanara Creek. Laku Point. Gomati Creek (Dwarka), Harsad.
Navibander (Madha Creek). Balapur. and Azad Island. In addition.
Rao (1987) mentioned creeks of Kutch and Aramda Creek in Gu-
jarat and Mahim, Ratnagiri and Jaytapur in Maharshtra. Durve
(1986) also mentioned some trawling areas around Bahrain in the
Arabian Gulf.
Pakistan
This species was found abundant on the coast of West Pakistan
(Ahmed 1971; Fig. 3). The following locations have been men-
tioned in the literature: the coast of Sind (Ahmed 1971 ). Korangi
Creek (18 miles south of Karachi) and Sonari (40 miles west of
Karachi: Asif 1978b). Sandspit backwaters (Qasim et al. 1985.
Barkati & Khan 1987. Aftab 1988), and Port Qasim (Gharo-Phitti
saltwater creek system near Karachi; Ahmed et al. 1987. Barkati &
Khan 1987).
ECOLOGY
Habitat
Below we summarize reports on the nature of the habitat de-
scribed for C. ariakeii\is and the vertical and horizontal ranges of
its distribution.
In Japan, O. riviilaris was only reported from muddy beds
(Ameiniya 1928, Wakiya 1929, Hirase 1930). It generally adheres
to other objects by the umbonal part of the left valve, but many
specimens appear to have lived separately (Hirase 1930), Its ver-
tical range is just above the low tide mark and closely restricted to
the vicinity of the low tide line (Amemiya 1928, Wakiya 1929). Its
horizontal range was determined by water temperature and salinity
(Imai 1978). The salinity range of its natural habitat under ordinary
conditions is 9-30 ppt (Amemiya 1928, Cahn 1950), the lower
range of which is lower than many Crassostrea species. As
Amemiya ( 1928) explained, these conditions are apt to change for
one reason or another. For instance, during ebb tide the exposure
of the beds to the air and sun inevitably inake the surrounding
water more saline due to evaporation. But because this species
lives close to the low tide mark, exposure to high salinities is short.
C. ariakensis can apparently tolerate low salinities as well. O.
rividaris was found in places where the salinity falls occasionally
much below 10 ppt, sometimes even in entirely fresh water
(Amemiya 1928).
In China, this species occurs widely among the river estuaries
along the coast. It is found from the low tide line to 7-10 m below
mean low water (Zhang & Lou 1956b. Zhang et al. 1960, Cai
1966, Cai et al. 1979. Xu et al. 1992). Sometimes it could be found
around the high water mark (Zhang et al. 1960). According to Lu
(1994). the temperature range of C. rividaris is 2-35°C. Normal
salinity range was reported as around 10-30 ppt (Zhang & Xie
1960. Lu 1994) or 9-28 ppt (Zhang & Lou 1956b). Optimum
salinity was reported as 10-25 ppt (Zhang el al. 1960) or 10-28 ppt
(Nie 1991 ). It was observed that C. rividaris could tolerate salinity
as low as 1-2 ppt for a short tenn (Zhang et al. 1960, Zhang & Xie
1960). as Nie (1991) reported its salinity range 1-32 ppt. Pure
fresh water could cause mortality (Zhang et al. I960). An inter-
esting exception to the normal distribution of C. ariakensis was
reported by Chen (1991) for Northern Jiangsu. The silty coast of
Jiangsu province was not originally suitable for O. rividaris. Ac-
tually, few oysters were found in this province. Things changed
when Spanina anglica was introduced. It was planted discontinu-
ously along the coast of Jiangsu province, and by 1991, it occupied
377 km of coastline and 1 80 km^ coastal area of the province. This
plantation changed the local ecology. Chen reported that this plant
kept clay around its growing area and gradually formed small
ridges and backwaters in that area, which he believed was a critical
condition for these oysters. O. rividaris was found at the seaward
boundary of the S. anglica planting area, which was between high
and middle tide mark with one-third to one-half time exposure.
The density of its distribution was as high as 107 per m"^ and the
average shell height of adult O. rivularis was 19.5 cm.
In India, C. rividaris was found on both hard grounds and in
muddy creeks (Mahadevan 1987. Patel & Jetani 1991). Patel and
Jetani (1991) reported its preference of muddy rocks, rocks cov-
ered by 3—4 inches of mud, although we have to think that settle-
ment preceded the mud deposits. This oyster has been found in
groups of four to five large and small individuals attached to
isolated rocks and coral stones that came up in trawl-nets (Durve
Cf<ASSOSTIit:A AR/AKENSIS REVIEW
Figure 2. Locations reported «ith C. ariakensis populations in China. No distinction is made between aquaculture sites and natural populations.
Underlined sites are considered major production areas. I. Xindao (dao: island): 2. Andon); (Dadonggoul ; 3. Zhuanghe: 4. Gaipin)>: 5. Fengnan;
6. Ninghe; 7. Beitang; 8. Tanggukou; 9. Yangjiaogou: l(). Ycxian; II. ^antai; 12. Rongchen ; 1.1. Dingzigang : 14. Shijiusuo: 15. Sheyang; 16.
Jianggang Bay; 17. Rudong; 18. Huijiao: 19. Daishan: 2(1. Zhenhai: 21. Dinghai; 22. Meilin: 23. Sannien : 24. Wenling; 25. I.ei|ing Bay: 26.
Wenzhou Bay: 27. Xiapu: 28. Ningde: 29. Luoyuan Bay: 3(1. Huian: 31. Tongan: 32. Xiamen : 33. I.onghai: 34. Haiclieng; 35. ^unxiao: 36.
.Shantou: 37. Haimen: 38. Lanbiao. Huilai County : 39. ,)iazi: 4(1. .lieshi: 41. (iaoluo: 42. .Shanwei : 43. Qingcao: 44. Baoan: 45. \iangzhou: 46.
Tangjiahuan : 47. Nanshui: 48. Hengshan : 49. Zhanjiang Bay: 50. Qinzhou\van(Longnien); 51. Baoping Bay : 52. Boao: 53. Qinglangang; 54.
Qiongshan: 55. Lofu Shan: and 56. Deep Bay.
1986) or solitary (unattached) in the littoral zone (Awati & Rai
1931). The vertical range of C rivularis was described as the
littoral zone (Awati & Rai 1931 ). sublittoral low waterline area or
submerged offshore area (Durve I9K6). intertidal (Mahadevan
1987. Rao 1987) or tidal region (Patel & Jetani 1991 ) and also at
9-15 m depth (Durve 1986).
In Pakistan, the preferred habitats of C. rivularis are the back-
waters and creeks along the coast (Moazzam & Rizvi 1983). It
seems that this species thrives in muddy environments (Ahmed
1971, Asif 1978b, Ahmed et al. 1987) and adheres to hard sub-
strate such as stones (Ahmed et al. 1987). It occurs near the low
water mark (Ahmed 1971, 1975, Ahmed et al. 1987, Barkati &
Khan 1987) and the preferred tidal height for spat settlemeni is 0.5
ft mark (Ahmed et al. 1987).
Predators, Harmful Organisms, and Diseases
According to Zhang and Lou (1956b). in China, "led tide" is
generally most hai-mful to oysters. It caused 509r mortality of
10
Zhou and Allen
TABLE 3.
Locations where C. (O.) riviilaris was reported in China.
Province
Locations Where C. (O.) riviilaris was reported
Liaoning Gaiping. Andong (Dadonggou). Xindao. Zhuanghe
( Zhang & Lou 1959)
Hebei Fengnan. Tanggukou. Beitaiig (Zhang & Lou 1959)
Tianjin City Ninghe (Zhao et aL 1991)
Shandong Rongchen (Zhang & Lou 1956b)
Yangjiaogou. Dingzigang (Zhang & Lou I956h. 1959)
Shijiusuo (Zhang & Lou 1959)
Yantai. Yexian (Zhao et al. 1991 )
Jiangsu Sheyang, Rudong (Zhang & Lou 1959)
Northern coast (north of Jianggang Bay; Chen 1991 )
Zhejiang Sanmen (Zhang & Lou 1956b)
Zhenhai, Daishan, Huijiano. Dinghai, Meilin, Wenling
(Zhang & Lou 1959)
Wenzhou Bay (Huang et al. 1981)
Leqing Bay (Zhou et al. 19821
Fujian Xiamen (Zhang & Lou 1956b. 1959)
Tongan. Haieheng (Zhang & Lou 1959)
Luoyuan Bay (.Xu el al. 1992)
Yunxiao. Longhai. Huian. Ningde, Xiapu (Cai 1966)
Guangdong Shanwei. Lanhiao (Zhang & Lou 1956b)
Baoan. Tangjiahuan. Hengshan (Zhang & Lou 1956b.
1959)
Shantou. Jiazi. Jieshi. Haimen, Nanshui (Zhang & Lou
1959)
Qingcao. Gaoluo, Xiangzhou (Zhang et al. 1960)
Zhanjiang Bay (Cai et al. 1992)
Peal River estuary (Guan & Li 1986)
Guangxi Longnien (Zhang & Lou 1959)
Hainan Baoping Ba> (Zhang & Lou 1956b, 1959)
Qiongshan. Qinglangang. Boao. (Zhang & Lou 1959)
Hong Kong Lofu Shan (Ke & Wang 2001 )
Deep Bay (Mok 1974)
cultured oysters in Baoan. Guangdong Province in 19.5.^. Red tide
could be caused by Noctiluca sp. diatom or the more harmful
Dityhun sp, The carnivorous oyster drills Thais gradata (known as
"huluo," which means tiger snail in China) and Naticidae sp.
(known as "yuluo," which means jade snail) are also very harmful
to oysters. Tiger snail can drill through the shell of a spat in 3 min
and in 8 h for a 3-y-old oyster (Wu et al. 1997). Beside these,
carnivorous crabs, such as Scylla. Portunidae. Lithodidae. sea ur-
chin Ecliiiioidea. and sea star Aseroidea. are also harmful to spat.
Below we list the available reports on these subject areas by
publication year.
Harmful organisms to C. riviilaris cultured in Zhanjiang Bay,
Guangdong Province, China (Cai et al. 1992)
The effects of the predator T. gradata and Balanus spp. were
reported in an important estuary for aquaculture. T. gradata was
found harmful to l-y-old oysters. Its density on oyster cultch could
be as high as seven individuals/m". Mortality caused by T. gradata
could be as high as 31%, 14% on average. T. gradata preferred
living in groups, usually hiding in the shaded area of concrete
posts. Its reproductive season was from the beginning of April to
the middle of June peaking from the beginning of April to the
beginning of May. Each female carried .50-100 oospores, with
about 100 eggs in each oospore. Hatchability was very high, al-
most 100%. Incubation period was about 15-30 days. Barnacle
Balanus spp. competed for setting space and food. In the worst
situation, the oyster seed could be smothered with a total covering
of Balanus spp. Balanus spp. set increased from the upper estua-
rine area toward the lower saltier regions. Highest density occurred
in the low intertidal area. Balanus spp. larvae preferred the sunny
side of a setting place.
Mass mortality putatively caused by Proroceiilnim sp. bloom in
Zhanjiang, South China (Zhang et al. 1995)
From late April to late May 1994, an episode of high mortality
occurred at an O. rivularis farm close to the port of Zhanjiang.
Fujian Province. South China. Mortality reached 98% o\er about
25 hectares. Water sampling and histopathological monitoring was
conducted. During the outbreak, the water temperature increased
from 18 to 30°C, pH fluctuated between 6.5 and 7.0. and salinity
ranged 25.6-29.1 ppt. The water was blue-brown in color and all
water samples revealed variable concentrations of phytoplankton.
of which 96% were composed of Prorocentruin sp. with concen-
trations of 201-667 cells/mL over the period of observation. The
temporal association of the mass mortality and a Prorocentrum
bloom suggested that the bloom was probably the cause of the
mortality. This assumption is supported by the histopathological
findings that suggest toxicosis. In particular, the observed lesions
were acute and corresponded with the outbreak.
Affected oysters were gray in color and had a softer than nor-
mal texture. The most outstanding microscopic lesion was intense
accumulation of hemocytes in and around hemolymph channels,
especially in the Leydig tissue. Close examination of the larger
vessels revealed that hemocytes were actively infiltrating the ves-
sel walls, as well as involved in transmigration into the Leydig
tissue and the formation of intravascular thrombi. A diffuse, and
less intense, hemocytosis was present in the interstitium between
the digestive tubules, while a mild hemocytosis was detected in the
gills. Oedematous changes were prominent around the digestive
tubules and in the Leydig tissues where they were accompanied by
tissue necrosis/lysis. The digestive tubules were empty and their
epithelia were dysplastic, varying from low columnar to cuboidal
and in some instances there was necrosis of the tubular epithelium.
Brown cells were pailicularly prominent in the intertubular tissues.
The pathology was consistent with a systemic toxicosis resulting
from absorption of toxins from the digestive gland.
Bouamia-\\V.e parasite found in C riviilaris reared in France
(Cochennec et al. 1998)
C. rivularis was imported from the Haskin Shellfish Research
Laboratory in New Jersey in 1994. Seven months after introduc-
tion, some mortality occurred in quarantine. Histologic examina-
tion revealed the presence of an intracellular protozoan parasite in
the connective tissues of nine dead specimens. Ultrastructure
analysis suggested that the protozoan might belong to the genus
Bonamia. Bonamia was likely transmitted to the experimental oys-
ters from neighboring waters, which are endemic for bonamiosis,
possibly when inlet water treatment lapsed.
An intracellular procaryotic micoorganism associated with lesions in
C. ariakeiisis in Pearl River estuary. South China (Wu & Pan 2000)
A series of mortalities of cultured oysters have occurred in
Pearl River estuary since 1992. usually from February to May. The
mortality peaks at 80-90%' during April and May. The diseased
CRASSdSTREA ARIAKENSIS REVIEW
11
Figure 3. Luculions reported with ('. ariakensis populations in India and Paliistan. India: 1. Ratnagiri (I6N, 73E); 2. Balapur (not locatedl: 3.
Porljander iPorbundar), Navibander (2IN, 69E|; 4. Dwarka (Gomati Creeli) (22N, 68El: 5. Oiiha. Aramda Creek, Posheira, Port Okha, Sikka
(22N, 69E). Pakistan: 1. Korangi Creek (24N, 67E): 2. Karaclii (24N. 64E); and 3. Port Qasini (27N, 68E).
oysters are generally aged 2-7 y. A rickettsia-like iiitracelliilar
microorganism is present in the tissue of diseased oysters.
PHYSIOLOGY
Natural Reproduction
Hermaphroditism and Sex Reversal
Crcissostrea are oviparous and protrandric hermaphrodites (c./..
Coe 1934). The occurrence of true hermaphrodites (both sexes
simultaneously) is rare. Hasan (1960) stated that hermaphrodites
do not exist in O. discoidea ( = C. rividaris). In a study of her-
maphroditism and sex reversal in C. rividaris from the coast of
Karachi, Pakistan, true hermaphrodites were absent (Asif, 1979).
Hermaphrodites observed were actually transitional stages of the
sexes and used to study sex reversal. According to Asif, gonad
generally appeared in C. rivukiris at the age of 2-3 mo at a length
of 0.4-0.6 cm and 62* were male. Protandric hermaphrodites
were found in summer and autumn, which indicates the time of sex
reversal. The percentage of males declines gradually with increas-
ing size as is true for other Cnissostrea spp. Cai et al. ( 1992) also
claimed that sex ratio of C. riviiUiris had an obvious regular change
during the reproductive season (usually summer and autumn) and
the ratio of females to males increased as the oysters got older.
Hasan (1960) also mentioned that individuals with undistinguish-
able sex are fairly common throughout the spawning season. In
Asif s study, the percentage of females increased over males be-
yond the size class 5.0-5.9 cm.
Spawning
Importance of temperature in gonad maturity and spawning of
oysters is well known. Temperature influences the development of
gonad (Orton 1936, Spark 1925. Nelson 1928). Temperature also
directly influences the abundance of food, which is necessary for
the development of gonad (Loosanoff & Engle 1942. Loosanoff &
Tomnier 1948). Periodic examinations of the gonad of O. dis-
coidea showed that normal growth of the reproductive products
was coincident with gradual rise of water temperature and food
abundance in the summer months (Hasan 1960).
The combined effect of temperature and salinity on the start of
12
Zhou and Allen
spawning was discussed by Hornell (1910. cited from Hasan.
1960) and confirmed by Hasan (1960) through an experiment on
O. discoidea in Pakistan. The rise in water temperature helps the
development of gonad, while decrease in salinity stimulates the
gonad for spawning. Cai et al. (1992) also mentioned that oyster
reproduction is closely related to environmental conditions. High
temperature and low salinity could cause mass spawning of C.
rividaiis in Zhanjiang Bay. Guangdong province. Hu et al. (1994)
presented a more detailed and slightly different discussion in his
study of C. lividaris spat collection in Jioulong River estuary.
Fujian province. He agreed that spawning is related to the change
of water temperature and salinity. Water temperature could change
with wind direction or strength. Salinity could be changed by
precipitation, water current, and tides. However, he seemed to
believe that simply a change of water temperature and salinity
could be the trigger for spawning, whether an increase or decrease.
According to his observation, whenever the tide changed from
neap to spring, spring to neap, or during spring tide, oysters would
spawn, as long as their gonad was well developed. If the wind
direction happened to change from northeast to southwest, or cold
air happened to pass by. spawning would increase. He explained
that a temperature change of only about 1-2°C would stimulate C.
rivularis to spawn.
Hasan (1960) studied two natural O. discoidea beds at Wau-
gudar Creek. Pakistan. Spawning starts by the first week of July
when temperature was about 28-29°C and salinity about 24 ppt.
Number of spawning individuals remains almost constant during
August and September, much reduced in November and almost nil
in December
Several authors talked about reproduction of C. rividaiis from
China. According to Zhang and Lou ( 1956a), the optimum salinity
for reproduction of C. rivularis is 10-25 ppt in China. Hu et al.
(1994) reported that in Jiulong River estuary. Fujian province,
gonad maturity reaches its peak from the middle of April until
mid-May. Oysters spawn twice each year: spring spawn is from
May to June and fall spawn, from the end of October to the
beginning of December. During spring spawn, water temperature
fluctuated between 20 and 30°C, salinity 5-25 ppt. Guan and Li
(1986) mentioned that in Zhujiang River estuary. Guangdong
province, the reproductive season is from June to September.
Spawning is mainly during June and July. There might be a second
spawning if appropriate environmental conditions are available.
Guan and Li did not report the environmental conditions associ-
ated with spawning. Cai et al. ( 1992) reported that the reproductive
season is generally from the beginning of April to the middle or
end of June in Zhanjiang Bay, Guangdong. Environmental condi-
tions in the study area (Shimen) are listed as follows: Annual water
temperature ranged from 14 to 31.8°C. Daily water temperature
changed 2 to 4°C. Water temperature was highest in June and
lowest in January. Salinity ranged from 7.52 to 22.18 ppt in sum-
mer (but could drop to 0.00 ppt when flooded). 18 to 30 ppt in
winter. pH ranged from 7.1 to 7.9 in summer and 7.9 to 8.1 in
winter. Zhang et al. (I960) mentioned that reproduction occurred
year round in South China Sea area. The reproductive peak is from
late May to eariy September. Zhang et al. did not report environ-
mental conditions during this time period.
According to Tanaka ( 1954). the spawning season of O. rivu-
laris ranges from late May (20-22°C) to early September (28-
26.5°C) in Ariake Bay, Japan. There are three major spawning
periods during this season: early June (22-23°C). late June to eariy
July (24-26°C). and the beginning to middle of August (30-
28.5"C). The eggs of U. rividtiris measure 49-53 ixm in diameter.
The relation between salinity and developmental condition is
shown in Table 4. The temperature varied from 24 to 27°C
(Amemiya 1928). The above results are neariy identical to those of
Hu/iniori (1920. cited from Amemiya. 1928).
Spalfall
The preferred tidal height of settlement for C. rivularis spat was
reported to be at the 0.5 ft mark in Pakistan (Ahmed et al. 1987).
A broader range was reported from China by Nie ( 1991 ): from the
low tide line to a depth of 10 m. with the maximum setting at
± 0.4 m low water mark. Hu et al. (1994) reported the optmial
water depth for spat collection is from the low tide mark to a depth
of 1 m in Jiulong River estuary. China. Larvae settle 12-18 days
after spawning. In southern China, spatfall occurs from June to
August, the period of highest temperature and lowest salinity (Nie
1991, Cai & Li 1990).
Three reports on spatfall seasons from Pakistan are summarized
below. One study was conducted at Paradise Point situated on the
west coast of Karachi (Moazzam & Rizvi 1983). This is basically
a rockv shore having frequent stretches of boulders and sand. The
subtidal area along this shore is generally more deeply inclined
than the rest of the coast. This is also a power plant site. C.
rivularis occurs in the cooling system of the power plant, which
has been made artificially "protected" and simulates conditions of
a backwater environment. The enxironnient conditions were re-
ported as follows. Temperature dropped to its minimum of 20-
22 C in December-January and reached its maximum of 28-30°C
m June-July. Salinity remained fairly constant (35-36 ppt) except
during the short spell of rains in July-August when salinity
dropped to 28 ppt. The contents of suspended matter fluctuated
between 0.003 mg/L in November and 0.1 16 mg/L in June. Trans-
parency was less than 1 m in June-July. Maximum settlement of
C. rivularis occurred in June and September-October. A consid-
erable number were also observed in July-August.
The second report came from two natural oyster beds (Hasan
1960). One is situated between Korangi and Kadero creeks, south
of the village Vagudar and about 16 miles southeast of Karachi.
The other one is about 6 miles south of Dhabeji. The temperature
and salinity profile were reported from Vagudar creeks. Tempera-
ture profile looks very similar to the one from the above report,
except that it dropped even lower to 16-17°C in January. Salinity
was reported only from April to September, with a maximum of
3(S-37 ppt in April-May and then dropped continuously to 21-22
ppt in September. The pattern of larval settlement of O. discoidea
in this report is different from the one mentioned above. Settlement
at Vagudar Creek occurred from July to December with mid-
TABLE 4.
Relationship between salinity and developmental condition,
accordini> to .\meniiya 1928.
Salinity ppt
Sp. gr. at C
Condition
ca. 7
ca. 1.0056
Minimum salinity
S-14
1.0064-1.0112
Much too low salinity
L'^-IX
1.0120-1.0144
Too low salinitv
1 9-25
1.0153-1.0200
Optimum salinity
26-30
1.0209-1.0241
Too high salinity
31-33
1.0249-1.0256
Much too high salinity
ca. 34
L-a. L0273
Maximum salinity
Crassostrea ariakensis Review
13
September being the peak permd. Moa//aiii and Ri/vi related
setting failure to the presence of high contents of suspended matter
in seawater during the southwest monsoon period (June-
September). This high content of suspended matter is believed to
interfere with larval settlement of many in\ertebrates in this area
(Ahmed et al. 1978).
The third report came form the Gharo-phitti saltwater creek
system (Ahmed et al. 1987). Spat fall occurred from April to
October with peak settlement from April to July. The maximum
settlement occuned during the period June 24 to July 23. No
environmental conditions were given in this report.
Growth
Growth Rate
C. uriakeiisis is well known for fast growth. In Pakistan. C.
rivularis spat reached the si/e of 0.5 mm in about one week and
2.0cm in about I mo (Ahmed et al. 1987). Hasan ( I960) found that
a size of 3.0 cm was reached 2 mo after settlement. In about one
and half years, they become ready for market. Temperature and
salinity data of Hasan's study is shown in the spatfall section. In
China. C. rivularis can growth to 10-16 cm in 2 to 3 y (Zhang &
Lou 1956b). In Japan, it attains full size (20 cm) in 2 or 3 y
(Amemiya 1928). The results of Fujiinori's study (1929) on the
growth rate of O. rivularis was presented in two parts: spat / young
oysters and the sexual adult. Fujimori found that the growth rate of
the spat varies considerably according to their time of attachment.
The size of adult O. rivularis in Kyushu was 5.5 cm shell height at
I y. 9.7 cm at 2 y. 12.4 cm at 3 y. 15.2 cm at 4 y. 17.9 cm at 5 yr.
and 19.7 cm at 6 y. In Japan, growth was most rapid in August and
September (Cahn 1950). Environmental conditions were unavail-
able for the above reports, if not mentioned.
Shell Dimension
C. ariakeiisis reaches a large size. As Cahn ( 1950) mentioned,
the maximum size attained by this species according to the litera-
ture is 257 mm with an estimated age of 20 y. The maximum
length he recorded in Japan was 240 mm. A maximum shell height
of about 200 mm was reported several times from Japan and the
United States (Amemiya 1928. Hirase 1930. Coan et al. 1995).
According to the growth rate of adult O. rivularis determined by
Fujimori (1929). the estimated age of such size is more than 6 y
old. Generally, adult specimens reach 6-7 inches (or 150-170 mm)
in height, as reported from four countries (Hirase 1936. Galtsoff
1964. Ahmed 1971. Rao 1987).
Allometric Growth
A study of the allometric (relative growth) relationship between
shells and tissues of C. rivularis was presented by Barkati and
Khan (1987) from Pakistan. Shell length was defined as the maxi-
mum distance between the tip of the anterior margin and the pos-
terior margin. Shell width was defined as the maximum distance
between the lateral maigins. The following points were reported.
Shell width increased faster than shell length (/■ = 0.85). Shell
length increased faster than dry tissue weight (/■ = 0.52). An
exponential relationship exists between shell length and shell
weight with faster growth in length compared with shell weight
(/■ = 0.84). Dry tissue weight increased faster than shell weight (c
= 0.74). Condition index (the proportion of dry tissue weight to
total dry weight of shell and dry tissue) increased with increasing
shell length (r = 0.41 ). No linear variable was useful to accurately
predict other variables due to low coefficient of correlation (/).
probably due to irregular growth in various shell dimensions
(length and width).
For example. Asif ( 1978b) reported variation in shell growth in
two populations of C. rivularis caused by setting density in Pak-
istan. One population in Korangi Creek was exploited and densi-
ties were low. Another population in Sonari was crowded. In the
Korangi Creek, the oysters are attached to rocks or stones hori-
zontally, whereas those of Sonari grow upward with the umbo
downwards. Generally, the wild stock of C. rivularis of the Kor-
angi Creek are round and shallow whereas the Sonari population is
elongated and deeply cupped. In the majority of the Korangi Creek
population, height plus width varies closely with length of the shell
while in the Sonari population, shell height plus width varies twice
as much as the length.
Feeding
Food .Selectivity
According to Cai et al. ( 1992). C. rivularis (collected in Zhan-
jiang Bay, Guangdong Province. China) is a selective feeder. It
prefened small articles to long-chain groups or large articles. The
majority of its food is composed of phytoplankton such as Cosci-
uihUscus sp., Nitzscliia sp. and Cyclotella sp.
Feeding Habits
Zhang et al. ( 1959) did an extensive study on the feeding habits
of O. rivularis in relation to time, tides, season (change of tem-
perature and salinity) and suspended particles. The experiment was
conducted in the Pearl River estuary and some nearby bays. Most
of the sampled oysters were 3 to 4 y old at the time of examination.
These oysters were collected from the wild as spat and cultivated
in oyster farms. The percent of O. rivularis that are feeding at any
given time (incidence of feeding) was not related to periods of
light and darkness, nor to the periods of tides, or the density of
suspended particles. Salinity and temperature did have certain in-
fluences, as summarized below.
According to examinations at five different times of the year,
the highest average incidence of feeding for O. rivularis was a
little more than 80%. It was also found that feeding time of O.
rivularis adds up to 16-19 h everyday with irregular intervals.
Feeding habits of O. rivularis were not related to change of sea
level or direction or speed of water flow caused by tidal change.
In Pearl River estuary, feeding incidence of O. rivularis was
highest from October to April (50-100%). when temperature
ranges between 10 and 25°C and salinity between 15 and 30 ppt.
During summer, the natural reproductive season of O. rivularis.
when temperature is much higher (22-30°C) and salinity is much
lower (3-26 ppt). feeding incidence is lower (0-70%). Feeding
incidence seems to be more closely related to salinity according to
monthly records. Although O. rivularis is known to tolerate low
salinity, feeding rate was significantly retarded if salinity was
lower than 5 ppt. Above 10 ppt, feeding was active.
Increase in suspended particles in the seawater (higher turbid-
ity) failed to influence feeding incidence of O. rivularis. In this
case, the authors maintained that these suspended particles served
as a food source for the oysters.
Oxygen Consumption
Guan and Li (1988) did an extensive study on oxygen con-
sumption of C. rivularis. A Warburg manometer was used to mea-
sure the oxygen consumption of dissected gill tissue of C rivularis
taken from the Shenzhen Bay Oyster Fann. Oxygen consumption
14
Zhou and Allen
varied with the change of seawater temperature. A negative cor-
relation was found between oxygen consumption and the oyster
age. Tlie older and heavier the oyster, the less oxygen was con-
sumed by its gill tissue. Oxygen consumption differed significantly
in different reproductive periods.
BIOCHEMISTRY
Biochemical composition
Qasim et al. (1985) determined the following biochemical pa-
rameters for C. hvidaris from Pakistan. Water contributes 787r of
soft body wet weight. Of soft body dry weight, 35.7% was crude
protein, 22.5% glycogen, 23% lipid, and 11.2% total inorganic
substances. These are the averages from sampling over a period of
time (sample interval was not stated in the article). Higher value
for lipids (31%) was reported from India (Patel 1979. cited from
Qasim et al. 1985). This difference is probably the result of geo-
graphical variation, seasonal variation, or both.
Qasim et al. ( 1985) mentioned that the ratio between glycogen
and protein changes with reproductive state of an oyster (no spe-
cific information available). Another report on biochemical in-
dexes of C. rivularis from the Pearl River estuary. China (Guan &
Li 1986) showed seasonal change of lipid content and its close
relationship with reproductive physiology of the oysters. As the
authors di.scussed, reproductive season in the Pearl River estuary is
from June to September, of which June and July are primary
spawning periods. There could be a second spawning in September
if environmental conditions were appropriate. In their study, lipid
content was highest in May (2.88% of wet weight), then dropped
dramatically from June until it reached the lowest point 1 .06% in
October, the end of the reproductive season.
For protein, amino acid profile determines the nutritive quality
of tissues. Such a profile of C. rivularis tissue protein has been
reported from the Pearl River estuary. China (Guan & Li 1986) and
Pakistan (Aftab 1988). There are only slight differences between
the two reports. From China, specimens were tested in May, and
the amino acid profiles are presented in Table 5 (Guan & Li 1986).
Glutamine and asparagines are most abundant. From Pakistan, 14
amino acids were analyzed. Methionine and arginine were not
detected. Glycine and aspartic acids were most abundant. Seasonal
variation in bound amino acid content is shown in Table 6 (from
Aftab 1988)
The shells of O. rivularis have been used as traditional Chinese
medicine. Zhao et al. (1991) examined the content of calcium
carbonate, trace elements and amino acids in shells of O. rivularis
collected from Tianjin. Shandong, Zhejiang, and Fujian provinces.
Calcium carbonate in raw shells was 92.0-95.5% and in calcined
shells, 96.4-96.9%. Calcined shells have had organic materials
removed. The raw shells contain large amounts of Ca, small
amounts of Mg, Na, Sr, Fe, Al, Si, and traces of Ti, Mn, Ba, Cu.
etc. Shell decoctions (an extract obtained by boiling the shells)
contain small amounts of Ca, Na, Mg. K. and trace element of Sr,
P, Pb, Zn, Ni, V, Ba. Li. Mn, Ti, Cu, Cr, Mo. As, Hg, etc. The
oyster shells contain 1 7 amino acids. Total amino acid content
amounted to 0.16 to 0.24% in raw shells.
Li et al. (1994) studied the medicinal value of "oyster complete
nutritional tablet," a dietary supplement made from extracts of
both shells and soft body of O, gigas and O. rivularis from South
China Sea. The tablet contains a high content of eighteen amino
acids, especially the eight essential to the human body. Putative
benefits are attributed to the liver, kidney, spleen and intestine to
a certain extent.
TABLE 5.
The amino acid compositions and their contents In C. rivularis
sampled in May, 1984 (Guan & Li 1986).
.\mino Acid
Contents In
Dried Samples ( % )
Alanine
Arginine
Asparagine
Cystine
Glutamine
Glycine
Histidine
Isoleucine
Leucine
Lysine
Methionine
Phenylalanine
Serine
Threonine
Tyrosine
Valine
2.04
L95
3.30
0.28
4.06
2.15
0.76
1.18
1.87
2.23
0.57
1.05
1.39
1.48
1.34
1.32
Heavy Metals and Toxins
Lu ( 1994) did a preliminary study on the feasibility of using O.
rivularis as a monitoring agent for heavy metals, like Cu, Zn, Cd,
Pb, along the Guangdong coast, China. He found that profiles of
Cu, Zn and Cd content in the oyster correlated with the distribution
of industrial discharge along Guangdong province. Also see Ke
and Wang 2001. Further investigations on the suitability of O.
rivularis as a biomonitor of specific metals or other chemicals are
presented below.
Zn
According to Lu et al. ( 1998a), Zn accumulated continuously in
the tissues of the oyster through 12 days of exposure. Accumula-
tion was linear with time. Loss of Zn from C. rivularis was not
observed over 35 days of depuration. Zn accumulated less readily
with increasing salinity. The author concluded that in general C.
rivularis is a reliable indicator of Zn in marine systems.
TABLE 6.
Seasonal variation in the protein and amino acid composition of
tissue protein hydrolysate of C. rivularis (Aftab 1988).
Component
February
May
August
November
Average
Protein % d.v\.
40.36
41.25
52.50
55.00
47.33
Alanine
6.04
6.91
9.00
9.67
7.90
Aspartic acid
6.78
9.83
12.56
6.58
8.94
Glutamic acid
6.. 30
7.98
11.26
5.08
7.65
GIvcine
13.27
11.88
6.55
9.10
12.10
Histidine
2.07
1.87
2.72
1.91
2.14
Isoleucine
2.38
1.80
3.12
2.08
2.32
Leucine
3.98
2.89
5.34
3.63
3.96
Lysine
1 .59
1.74
1.44
1.28
1.51
Phenylalanine
1.73
1..3()
1.94
1 .55
1.63
Proline
0.46
2.05
0.79
2.40
1.43
Serme
4.07
6.01
7.82
3.51
5.35
Threonine
3.85
5.74
6.92
3.24
4.93
Tyrosine
0.85
0.60
0.91
0.79
0.79
Valine
3.64
3.75
5.55
2.98
3.98
CflASSOSTRKA AK/AKENSIS ReVIHW
15
Cd
Lu et al. (1998b) studied Cd absdiplion in C. riviilaris. The
content of Cd in body tissues of C hviilaris accumulates in linear
proportion to Cd concentration in the water and to exposure time.
Accumulated Cd attenuates slowly with a biologic half-life of 77
days. With increased salinity, rate of accumulation decreases while
rate of Cd loss slows down. C. livuhiris seems to be a reliable
bio-monitor of Cd pollution.
Cu
Cu absorption in C. rivuUiri.s was examined by Lu et al.
(1998c). It continuously accumulated in the tissues of the oyster
through the e.\posure to a concentration of 100 (J-g/L over 12 days.
Accumulation was linear with time and decline of Cu concentra-
tion was slow, with a half-life about 1.^1 days. Rate of Cu accu-
mulation was significantly slower with increased salinity, but rate
of decline in Cu concentration was not signiticantly related to
salinity.
Total Petroleum Hydrocarbons (TPHs)
Lin et al. ( 1991 ) looked at concentration of TPHs in the Pearl
River estuary. China. TPHs in C. rivularis tissues decreased with
time during the period leading to sexual maturity. The rate of
decrease was about 0.24 |jig/g, dry weight. The biologic half-life
was 43 days. Aromatic hydrocarbon compounds with smaller mo-
lecular weight were released sooner from oyster tissues than those
with greater molecular weight. The concentrations of TPHs in
oyster tissues were not significantly related to those in waters and
sediments, and not clearly dependent on the contents of lipids in
oyster tissues during the study period (September 1986 until Feb-
ruary 1987).
GENETICS
Karyotype
So far, research on the cupped oyster species of the genus
Crassostrea shows a common diploid chromosome number of 2;;
= 20, and their karyotypes include only metacentric and submeta-
centric chromosomes. The proportion of these chromosome types
can be different interspecifically (Leitao et al. 1999).
Chromosome number of In = 20 was confirmed in C. aricik-
eiisis (leyama 1975) and in C. rivularis from West Pakistan
(Ahmed 1973) and China (Yu et al. 1993). Yu et al. reported the
karyotype of C rivularis sampled in Southern China had 10 meta-
centric pairs. A more recent karyological study (Leitiio et al. 1999)
on an American population of C. ariakensis originally introduced
from Japan shows that it consists of eight metacentric and two
submetacentric (nos. 4 and 8) chromosome pairs. A variable num-
ber of one to three Ag-NORs (nucleolus organizer regions) was
observed terminally on the metacentric pairs 9 and 10. About 68'7f
of the silver stained metaphases showed Ag-NORs only on pair 10.
Polyploidy
Rong et al. ( 1994) reported their attempts to produce tetraploid
C. rivularis. Newly fertilized eggs of C. rivularis from south Chma
were treated with physical and chemical methods in the first three
minutes before the cleavage of zygotes or at the onset of first
cleavage. Induction rates of tetraploids were 28% for heat shock,
30% for cold shock, 28% for chlorpromazinum treatment and
35,8% for "traditional Chinese medicine" treatment as indicated by
chromosome spreads from larvae. Production of viable spat was
not reported.
Hyhridizaliun
Gaftney and Allen ( 1993) reviewed previous hybridization re-
ports among Crassostrea species and pointed out that most of
reports of successful hybridization suffer from one or more of the
following: I ) ambiguities in classification; 2) possible contamina-
tion during spawning; 3) absence of experimental controls for
assessing the quality of gametes as well as larval viabilities; and 4)
the absence of genetic confirmation of hybrid status. They con-
clude that there was virtually no unequivocal evidence for the
formation of viable interspecific hybrids among Crassostrea spe-
cies.
Early studies on cross-fertilization between C. gigas and C
rivularis gained little success (Miyazaki 1939, Imai & Sakai
1961), but was reported successful by Zhou et al. (1982) and
Downing (I988a,b, 1991). Asif (1978a) reported successful pro-
duction of trochophore larvae 4-5 h for the cross of C. rivularis
with C. glomerata and Saccostrea cuccullata. For the reasons
mentioned above, these should be viewed with caution.
Hybridization of C gigas and C. rivularis was re-examined by
using specimens originally introduced from Japan to the United
States (Allen & Gaffney 1993). Such crosses are of interest be-
cause of the disease resistant properties of these species (Calvo et
al. 1999, 2001). In addition, the hardiness and apparent disease
resistance of C. gigas and the high temperature, low salinity tol-
erance of C. rivularis could lead to promising variants for aqua-
culture, especially if the diploid is sterile. Three replicates of a 2 x
2 factorial mating of C. gigas and C. rivularis were produced to
examine the viability of this cross. Fertilization rate, yield of 48-
h-old larvae, and survival of fertilized eggs was lower in the hy-
brids than in pure crosses. All crosses showed similar larval
growth rates, except C. rivularis (female) x C. gigas, which grew
more slowly. Isozyme electrophoresis and flow cytometry con-
firmed hybridization. Triploid hybrids were produced using tetra-
ploid C gigas and diploid C. ariakeusis (Que & Allen 2002).
Hybridization between C. ariakensis and C. virginica failed
(Alien et al. 1993). Cytogenetic and electrophoretic analysis re-
vealed the formation of hybrid zygotes and larvae between C.
virginica and C. rivularis. but larval survival was limited to a
maximum of 10 days. Larvae stopped growing at about day 4,
reaching a maximum length of about 80 um. Studies on larval
feeding using fluorescent beads indicated that growth limitation
apparently was not caused by an inability to feed. Induced triploidy
did not rescue hybrid failure.
Population Genetics
A number of studies have used molecular markers of various
sorts to distinguish among Crassostrea species, including C. ari-
akensis. Among the earliest was work by Buroker et al. (1979)
who estimated levels of genetic variation for six Crassostrea and
three Saccostrea species based on electrophoretic variation in pro-
teins in about 30 loci, C rivularis among them. Liu and Dai (1998)
used RAPD techniques to differentiate C. talienwhanensis and C.
plicatula froin C rivularis. Li et al. (1988) used electrophoretic
markers to separate four Crassostrea species, and concluded that
the "white oyster" was C. rivularis and the "red oyster," C. ired-
iilai.
C. rivularis was also among those used by Little wood (1994) to
establish the first phylogenetic estimates for this species based on
nuclear DNA. Since then, a number of other studies employing
16
Zhou and Allen
molecular markers have been applied to C. ariakeiisis. mostly to
discriminate among species (O'Foighil et al. 1995. GatTney &
O'Biern 1996. Hedgecock et al. 1999. Francis et al. 20(X)). Hedge-
cock et al.'s study confirmed the occurrence of C. ariakensis in the
northern regions of the Ariake Sea and re-emphasi/cd the need for
genetic confirmation for species identification.
AQUACULTURE
RefeiTences to aquaculture of C. ariakensis come mainly from
Japan and China, and are discussed accordingly.
Aquaculliire in Japan
Of the five edible oysters species in Japan, only O. gii;as and O.
rivuiaris were cultured commercially (Cahn 1950). O. i-i\'nlaris
was second to O. gigas in commercial importance (Amemiya
1928)
According to Amemiya (1928), cultivation of O. rivuiaris be-
gan in Ariake Bay in the late 1890s and seed were later trans-
planted to Kozima Bay in Okayama Prefecture around 1928. An
even earlier report of cultivation in Ariake Bay in the 186{)s was
given by Wakiya ( 1929). Both Wakiya and Langdon and Robinson
(1996) mentioned that the culture of Suminoe oyster were con-
ducted in the Suminoe river. Saga Prefecture from the beginning of
the Meiji period in the mid- 19th century. Discrepancy between
Cahn and Wakiya on the start of C. rivuiaris aquaculture might rest
on their definition of cultivation. Cahn ( 1950) described two types
of culture sy.stems at the mouth of the Suminoe-gawa ("gawa" in
Japanese means river or stream). Ariake Bay. a primitive one and
a more developed one. Cahn did not say when the primitive culture
started, but he implied that the more sophisticated culture started
after 1885. The primitive culture consisted simply of gathering
natural oysters and storing the larger individuals for a short time on
the bottom of the Sumino-gawa. later to be shipped to Nagasaki at
the proper season for sale.
Aquaculture of O. rivuiaris began fortuitously. For some rea-
son during the winter of 1884 these oysters were not shipped for
sale to Nagasaki. The ne.\t year they were considerably larger by
size and weight. From this observation, a new type of culture
evolved in the local area. Young oysters about 2.5 cm in length
were gathered from every possible growing place from July until
March and were placed on oyster beds at the mouth of the river. To
prevent loss, they were heaped close together in masses. They
were washed and cleaned twice or three times each month during
low tide. In April individual oysters were stuck in the mud verti-
cally, hinge down and ventral margins uppermost. As the mud was
very firm, the oysters fared and grew well. As they grew, they were
thinned and replanted to give them more growing space. Growth
was most rapid in August and September.
Aquaculture in China
C. rivuiaris is the most economically important marine shell-
fish species cultured in South China (Zhang et al. 1995), primarily
in Fujian, Guangdong and Guangxi Province. The history of its
culture in Guangdong is over .'^00 y old (Cai et al. 1979). The Pearl
River (Zhujiang) estuary. Guangdong was considered the most
famous cultivation site of this species (Zhang & Xie 1960). Some
other places mentioned in the literature are Yangjiaogou, Shan-
dong Province (Zhang et al. 1960), Leqing Bay, Zhejiang Province
(Zhou et al. 1982) and in Deep Bay, Hong Kong (Mok 1974). In
1996. China produced 2.3 million tonnes of oysters from aquacul-
ture, among which C rivuiaris accounts for 20-30% (Guo et al.
1999). In Guangdong province, C. rivuiaris production was about
40'>f of total sea culture production (Qiu & Li 1983).
The primitive method of oyster culture was to improve growth
and reproduction with procedures like fishing restrictions and pro-
tection from diseases and predators (Zhang & Xie I960). The
advanced method involves collecting natural spat and artificial
grow-out. Modern oyster culture includes larval culture and breed-
ing. Larval culture and breeding of C. rivuiaris larvae has been
successfully accomplished on a research scale in South China (Li-
ang et al. 1983. Cai et al. 1989) but has not been used in large-scale
commercial culture. Hatchery production of seed is seen as a step
to increase the reliability of seed production.
Spat collection and artificial grow-out is still the most popular.
This is composed of four steps: spat collection, grow-out, fatten-
ing, and harvest. For spat collection, cultch material to collect spat
was traditionally oyster shell and gravel (Nie 1991). Since the
1960s, cement plates (17-24 cm x 14-19 cm) or cement bars
(40-80 cm long x 4-6 cm") reinforced with embedded bamboo
stakes were used. Stakes are used increasingly since they are easier
to handle, provide more surface area, and are not so readily cov-
ered by silt. Season and location of spat fall is summarized in
Physiology. Oyster larvae in the water are monitored to ensure the
best time of planting the clutch. Spat collectors are placed in rows
in rectangular blocks, usually 30 to 37.5 x lo' stakes or 100 to 135
X 10' plates per hectare. Further details follow below for specific
culture techniques.
The age of harvest is generally 3.5 to 4 y (Qiu & Li 1983). but
\aries from 2 to 5 y depending on culture location where the
environment, the specific culture technique, and even the expected
market size could be different. For example, Guo et al. (1999)
reported 2 to 3 y in Guangxi where oysters maintain rapid growth
throughout the first 3 y and are usually harvested at a size of 10-15
cm. The culture technique used there is concrete bars or shell
strings hanging on rafts and long lines. In Pearl River estuary,
Guangdong, oysters were usually harvested at 3 y of age by bam-
boo stake culture (Zhang & Xie I960). Cai and Li (1990) reported
the period to be 3 to 5 y in Southern China.
Cai and Li (1990) summarized oyster culture techniques in
China. The ancient bottom culture techniques, including bamboo
stake, stone and concrete culture, are still the major methods, but
farmers are becoming increasingly aware of the advantage of off-
bottom culture, like the rack and raft culture. The various tech-
niques are described below (reproduced from Cai and Li's work,
1990).
Rock (Stone) Culture
Rock culture is usually applied in areas that have hard sub-
strate. Marble flagstones approximately 90 cm x 25 cm wide and
10-cm thick are preferred for this method. Stones may be arranged
one-by-one vertically, resembling tombstones or two stones may
be aiTanged in an "A" shape. Three stones may be ananged to form
a tripod. Average spacing between stone groups is 70 cm. Another
choice of rock is irregularly shaped natural boulders of 4 to 5 kg.
The traditional anangement of the boulders, called "stars in the
sky," involves uniform distribution over the substrate. Two modi-
fications were used along the coast of Guangdong and Hainan
Provinces. One is called "plum blossom" with five or six boulders
grouped together. Another is called "small house" with three flag-
stones aiTanged to form a shed or an upside-down "U." Both kinds
of rocks are thoroughly washed and then covered in limewash 10
davs before use.
Crassostrea ariakensis Review
17
111 Guangdong and [■uiian Proxinces. the rocks are set out in
early May to June or in November. Maxiniuni spatfall is expected
in May. Spat collected in June is usually subject to heavy mortality
due to high temperatures and strong sunlight during attachment.
Spat collected late in the season usually grew poorly because of to
low water temperatures. Oysters are grown to market size at the
site of spat collection.
Approximately 60,000 stones are required for one hectare, and
C. rivularis may be harvested in 3 to 5 y. Production is moderate,
ranging from 750 to 3000 kg per hectare. The oysters grown on
rocks are more subject to predation by starfish and other organisms
than are oysters grown on stakes, so considerable time must be
invested in predator control.
Concrete Culture
Prefabricated posts or tiles are a derivative of the traditional
rock culture technique for the culture of C. rivularis and has been
used since 1930 in Guangdong Province. Spatfall occurs most of
the year, but optimum periods are April and May. To prevent the
tiles or posts from sinking into the mud, they are removed and
reananged around May, September, and December. Concrete cul-
ture requires a 4-y cycle. Spat collection and growth occupies the
first year from June to April. The second and the third years
involve a cultivation period yearly from May to August. Market
size is attained in 2.5 to 3 y and involves a progressive increase in
the spacing of the concrete tiles or posts. The cultivation cycle is
completed by a fattening period extending from September to
January. For fattening, oysters are transferred from the spat col-
lection/grow-out area to prime growing grounds, usually in the low
intertidal zone. For this culture method, in Guangdong, harvest
generally occurs in February to April of the fourth year, when
growth rates begin to decline sharply. Expected production from
the concrete method is 7.5 to 15 tons of meat per hectare.
Rack Culture
Since 1965, rack culture has been used to cultivate C. rivularis
in Guangdong Province. The racks may be constructed of bamboo,
wood, stone or concrete. Because wood and bamboo are rapidly
destroyed by shipworms and stone is heavy and awkward to
handle, concrete is preferred. The forni of the rack varies greatly,
but consists basically of members driven into the substrate to form
a horizontal frame, which supports the oyster cultch 2.5 to 3 m
above the substrate.
Several types of material are used for spat collection. The most
popular one is punched oyster shells, separated by 3 cm bamboo or
plastic spacers, and strung on 2 m lengths of galvanized wire ox
polypropylene line. Concrete tiles, approximately 10 cm" with a
central hole, may be substituted for the oyster shell. Concrete poles
between 70 and 130 cm in length may also be used. The cultch is
suspended from the rack, with spacing proportional to the density
of spat settlement and the character of the growing area. The
number of racks accommodated varies widely between the grow-
ing sites. Production is estimated at 10 to 20 tons per hectare.
Raft Culture
According to Qiu and Li ( 1983). raft culture started in Japan in
1950. Since 1979. the Fisheries Research Institute of the South
China Sea has conducted experimental raft culture of C. rivularis
in Guangdong Province. The fattening period lasts from September
to May. and three crops may be harvested, because 2 mos are
sufficient under optimal seasonal conditions. The ratio of meat
production to shell is some 60'/^ higher in raft-fattened oysters than
in oysters harvested directly from bottom culture.
C. rivularis can be marketed in less than 3 y using rafts, and
that the condition factor will be increased by more that 22% and
the meat quality will be superior to oysters cultivated by the tra-
ditiimal bottom methods (Qiu & Li 1983). Though initial costs are
higher, the increased production and working advantages of float-
ing raft culture are apparent, and it is expected that raft culture will
account for a steadily increasing share of oyster production in
China (Qiu & Li 1983). Nie ( 1991 ) also mentioned that raft culture
gives faster growth and a higher yield. A raft of 84 n\' will produce
in 2 y what 667 nr of bottom culture will in 4 y. Rafts seem to
w ithstand typhoons better than originally thought.
DISCUSSION
C. ariakcusis shares many life history traits with other Cras-
soslrea species. It is clearly an estuarine species v\ith salinity
tolerances similar to C. virginica. Its occurrence in river systems
and apparent responsiveness to salinity changes for spawning cues
suggests that its reproductive strategy is somewhat different than
C. virginica. There are indications that larval behavior differs from
that of C. virginica (M. Luckenbach, VIMS, pers. comm.), perhaps
an adaptation to fluvial existence. Many other questions about its
ecology are unanswered or incomplete and a number of research
priorities have been identified (Rickards & Ticco 2002). One of the
principal problems with extrapolating life history from the avail-
able literature is the uncertainty over species designation. Some
reports are clearly referring to C. ariakensis. e.g.. those from
southeast China where aquaculture activity is concentrated and
there is a long history of working with this species. Other reports
are not so clearly C. ariakensis, especially ones deriving from
western India and Pakistan. Also because of likely morphologic
confusion, the geographic range for C. ariakensis is incompletely
described. For example, it seetns likely that its range should in-
clude the coast of Vietnam, yet there seem to be no direct accounts
of this. There are accounts of its occurrence as far as Borneo, the
Philippines, and Thailand, but these are unconfirmed. Froin a prac-
tical standpoint, C. ariakensis from China are probably an appro-
priate starting stock for an introduction, should that proceed, be-
cause of similarities in latitude. From that respect, this area seems
a most appropriate focus for obtaining more information on the
species. Korea and Japan are possible sources as well. We did not
encounter reports of C. ariakensis from Korea except as casual
remarks. Stocks in Japan seem to be limited in abundance.
It is unclear whether C. ariakensis is a "reef-forming" oyster,
depending on how you define "reef" Clearly, Crassostrea species,
and oysters in general, benefit from aggregation and adults or their
shells provide substrate for recruitment in subsequent generations.
Some accounts of C. ariakensis describe "oyster hills" that would
clearly qualify as reefs (Zhang & Lou 1956b. Zhang et al. I960).
Apparently, it is common knowledge among fishermen in China
that C. ariakensis forms reefs. Other accounts have C. ariakensis
occurring as small aggregates and singles. In our travels to China,
we encountered several sites that had "natural" populations of C.
ariakensis (Allen et al. 2002). There seem to be natural popula-
tions in proximity to Xiamen although we did not observe this first
hand. They were available in the local market and reportedly from
natural populations that were harvested. There are natural sets of
r. ariakensis near Hong Kong on the shores of Deep Bay. but this
18
Zhou and Allen
could be from culture activity in the area. Seed is imported froin
the Pearl River estuary, so there are likely sources of ""natural"
populations in the Pearl River delta system. We observed, first
hand, collection (harvesting) of C. ariakensis adults from sections
of the Shiman River near Guan Du in close proximity to Zhanjiang
Ocean University. According to the diver on hand, they occur in
various assemblages, mostly stuck onto available substrate such as
large rocks. They also occur in the Dafeng River in Guangxi
province near Beihai. There are probably many other natural popu-
lations along the coast of China. By way of caveat, it is difficult to
attest to the "naturalness" of resident C. ariakensis populations.
That is, those that we observed or heard about first hand were
populations that occurred relatively deep (3-10 m) in river sys-
tems. Whether at some time in the past populations of C. ariak-
ensis were distributed in higher reaches of the water column (i.e..
before they were exploited over the millennia) is difficult to es-
tablish. It is also difficult to distinguish whether spat fall is from
natural populations or from aquaculture operations.
There are clearly big questions concerning basic physiology in
the kind of detail that exists for other congeners. C. ariakensis
seems to exhibit growth rates that are extraordinary in head to head
trials with C. virginica. Yet, these trials have been carried out in
disease endemic areas where C. virginica could be sick or dying.
Growth rates of C. virginica in. for example, the Gulf of Mexico,
approach those seen in trials of C. ariakensis in the Chesapeake
Bay or reported growth rates from the literature. Similar knowl-
edge gaps exist for larval biology, reproductive physiology, pre-
dation. competition, etc.
In our opinion. C. ariakensis is an underused resource around
the world. It clearly has aquaculture applications in estuarine areas
that are marginal or unsuitable to C. gigas. the most popular cul-
ture species. It seems hearty, fast growing, and highly marketable.
Of course, utilization of this species would require introduction, as
in the Chesapeake Bay. From that perspective, it would be useful
to have more basic research on C. ariakensis with which to guide
decisions about movement of this potentially valuable oyster spe-
cies.
ACKNOWLEDGMENTS
The authors thank our Chinese colleagues for their warm as-
sistance in compiling many of the papers cited here, particularly.
Dr. KE Cai-Huan, Professor LI Fu-xue, Dr. CAl Lizhe. Dr. WU
Xinzhong, Dr. Catherine Lam. Dr. QILI Dequan. Dr YU Xiang-
yong. Professor CAl Yao-Guo (retired). Director LAO Zan. and
Dr. LIU Zhigang, among others. We also thank S. Shumway for
early editorial assistance. This work was supported by the Camp-
bell Foundation and an award to S. Allen, Jr. from the Virginia
Center for Innovative Technology. Contribution number 2541
from the Virginia Institute of Marine Science, College of William
and Marv.
LITER,4TURE CITED
Abbott, R. T. & S. P. Dance. 1986. Compendium of Seashells. Melbourne.
FL: American Malacologists, Inc.
Aftab, N. 1988. Amino acid composition of tissue protein from five species
of oysters. Pakistan J. Sci. Ind. Res. 31:200-202.
Ahmed. M. 1971. Oyster species of West Pakistan. Pakistan J. Zool. 3:
229-236.
Ahmed, M. 1973. Cytogenetics of oysters. Cytologia. 38:337-346.
Ahmed. M. 1975. Speciation in living oysters. Adv. Mar. Biol. 13:357-397.
Ahmed, M.. S. Barkati & Sanaullah. 1987. Spalfall of oysters in the Gharo-
Phitti salt-water creek system near Karachi (Pakistan). Pakistan J. Zool.
19:245-252.
Ahmed, M.. S. H. Naiz Rizvi & M. Moazzam. 1978. Studies on the settle-
ment and control of marine organisms in cooling systems of coastal
installations. Final Tech. Rep. Institute of Marine Biology. University
of Karachi. 234 pp.
Allen, S. K., Jr. & P. M. Gatfney. 1993. Genetic confirmation of hybrid-
ization between Crassostrea gigas (Thunberg) and Crassostrea rivii-
laris (Gould). .Aquaculture 1 13:291-300.
Allen. S. K.. Jr., P. M. Gaffney, J. Scarpa & D. Bushek. 1993. Inviable
hybrids of Crassostrea virginica (Gmelin) with C. rivularis (Gould)
and C. gigas (Thunberg). Aquaculture 113:269-289.
Allen. S. K., Jr.. K. Sellner & M. Zhou. 2002. Recommendations to the
Chesapeake Bay scientific community: Contacts and opportunitie^ for
C. ariakensis research in China. A white paper available from Chesa-
peake Research Consortium, sellner@serc.si.edu, 8 pp.
Amemiya. I. 1928. Ecological studies of Japanese oyster, with special
reference to the salinity of their habitats. Tokyo, [Impenal] University.
/ Coll. Agr. 9:333-382.
Angell, C. L. 1986. The biology and culture of tropical oysters. Manila.
Philippines: ICLARM (International Center for Living Aquatic Re-
sources Management) Studies and Reviews 13. 42 pp.
Anon. 1996. The geographical distribution of Crassostrea virginica. C
gigas and C. rivularis. VA. Mar. Resour. Bull. 28:11.
Asif M. 1978a. Some observations on interspecific cross in three species
of oysters from the coast of Karachi. Pakistan J. Zool 10:217-221.
Asif. M. 1978b. Variations in allometeric growth in the shells of Crassos-
trea rivularis (Gould). Saccostrea glomerata (Gould) and S. cuccullata
(Born) from the coast of Karachi. Pakistan J. Sci. Ind. Res. 23:46-49.
.Asif. M. 1979. Hermaphroditism and sex reversal in the four common
oviparous species of oysters from the coast of Karachi. Hydrobiologia
66:49-55.
Avvati. P. R. & H. S, Rai. 1931. The Indian zoological memoirs on Indian
animal types III. In: K. N. Bahl. editor. Ostrea cucullata (the Bombay
Oyster). Lucknow, Methodist Publishing House, p. 7.
Barkati. S. & R. M. Khan. 1987. Relative growth in three species of
oysters. Pakistan J. Sci. Ind. Res 30:624-627.
Buroker, N. E., W. K. Hershberger & K. K. Chew. 1979. Population
genetics of the family Ostridae. II. Interspecific studies of the genera
Crassostrea and Saccostrea. Mar. Biol. 54:171-184.
Cahn. A. R. 1950. Oyster culture in Japan. United States Fish and Wildlife
Service, fishery leaflet 383. 80 pp.
Cai. v.. C. Deng & Z. Liu. 1992. Studies on the ecology of Crassostrea
rivularis (Gould) in Zhanjiang Bay. Tropical Oceanol. 11:37—44.
Cai. Y. 1966. A preliminary survey on bivalvia from Fujian coast. / Zool./
Dong Wu Xue Bao iChinese) 2:76-80.
Cai. Y. & X. Li. 1990. Oyster culture in the People's Republic of China.
World Aquaculture 21:67-72.
Cai. Y.. Y. Zhang & R. Wei. 1979. Ostreacea. In Introduction of malacol-
ogy. Shanghai Sci. and Tech. Press. Shanghai, pp. 216-217.
Cai. Y.. Z. Liu & S. He. 1989. A study on the artificial rearing of spats of
Ostrea rivularis Gould. Mar. Sci./Hai Yang Ke Xue (Chinese) 1:53-56.
Calvo. G. W., M. W. Luckenbach & E. M. Burreson. 1999. Evaluating the
performance of non-native oyster species in Virginia. / Shellfish Res.
18:303.
Calvo, G. W., M. W. Luckenbach, S. K. Allen, Jr. & E. M. Burreson. 2001.
A comparative field study of Crassostrea ariakensis (Fujita 1913) and
Crassostrea virginica (Gmelin 1791) in relation to salinity in Virginia.
J. Shellfish Res. 20:221-229.
Carriker. M. R. & P. M. Gaffney. 1996. A catalogue of selected species of
living oysters {Ostreacea) of the worid. In: V. S. Kennedy, R. I. E.
Newell & A. F. Eble, editors. The eastern oyster: Crassostrea virginica.
College Park. MD: Maryland Sea Grant College Publication, pp. 1-18.
CHASSOSTHEA ARIAKENSIS REVIEW
19
Chen. H. 1941. The growth of oysters in Spuitina uni^Hca sea beach ni
Northern Jiangsu. Mar. Sci. / Hai )V»is; Ke Xiic (Chiiuwcl 5:68-69.
Coan. E. V.. P. V. Scott & F. R. Bernard. 199?. Bivalve seashells of
western North America, marine bivalve mollusks from Arctic Alaska to
Baja California. Santa Barbara: Santa Barbara Museum of Natural
History, pp. 213. 215 and 217 .
Cochennec. N.. T. Renault. P. Boudry. B. Chollet & A. Gerard. 1998.
Bonamia-Wke parasite found in the Suminoe oyster Crassoslrea rivii-
hiris reared in France. Dis. Aquatic Organisms 34:193-197.
Coe, W. R. 1934. Alternation of se.xuality in oysters. /4m. Wa/ 68:236-251.
Downing. S. L. 1988a. Comparing adult performance of diploid and trip-
loid monospecific and interspecific Crassoslrea hybrids. J. Shellfish
Res. 7:549.
Downing. S. L. l98Sh. Triploid and diploid hybrids between the oysters
Crassoslrea gigas and C. riviilaris: production, detection and potential.
J. Shellfish Res. 7:156.
Downing, S. L. 1991. Hybrids among the Pacific. American and Suminoe
oysters: induction and aquaculture potential. / Shellfish Res. 10:514.
Dunker, C. 1882. Index Molluscorum Maris Japonici. London. 301 pp.
Durve. V. S. 1986. On the ancestry and distribution pathways of three
species of Indian oysters. Ind. J. Mar. Sci. 15:56-58.
Fei, H. 1928. Oyster industry. New construclion 5
Francis, E., K. S. Reece & S. K. Allen, Jr. 2000. Species designation
among sympatric oysters Crassostrea ariakensis. C. giga.\. and C. .sika-
mea. J. Shellfish. Res. 19:662-663.
Fujimori. S. 1929. Studies on the marine aquiciilture of Ariake Bay.
Fukuoka Fish Exp Sta. 75 pp.
Fujita. I9I3. Nihon suisan dobutsugaku (Japanese aquatic fisheries ani-
mals). Vol. 2. Tokyo: Shokabu. 292 pp.
Gaffney, P. M. & S. K. Allen. Jr. 1993. Hybndization among Crassoslrea
species: a review. AquaculUire 1 16:1-13.
Gaffney. P. M. & F. X. O'Biern. 1996. Nuclear DNA markers for Cras-
soslrea species identification. / Shellfish. Res. 15:510-511.
Galstoff. P. S. 1964. The American Oyster. Crassostrea virginicu Gmelin.
United States Depanment of the Interior. Fish and Wildlife Service.
Fishery Bulletin 64. Washington. DC: United States Government Print-
ing Office. 480 pp.
Glude, J. B. 1971. Identification of oysters of the South Pacific Islands.
National Marine Fisheries Service, Seattle, WA: Mimeo.
Gould, A. A. 1861. Description of shells collected by the north Pacific
exploring expedition. Proc. Boston Soc. Natural Histoiy 8:14-40.
Guan, Y. & Y. Li. 1986. Comparative studies on some physiological and
biochemical indexes of the cultured oysters in the Zhujiang River es-
tuary. Proc. Conchol. 3:1 lO-l 16.
Guan. Y. & Y. Li. 1988. Preliminary studies on the oxygen consumption
of gill tissue separated the oyster, Crassostrea rivularis. Oceanol. Lim-
nol. Sin./Hai Yang Yu Hu Zhao (Chinese) 19:210-214.
Guan. Y. & Y. Zheng. 1990. A comparison study on the esterase isoen-
zyme of cultured oysters in China. Soiiili China Sea Fisheries Res.
2:32-35.
Guo. X.. S. E. Ford & F. Zhang. 1999. Molluscan aquaculture in China. /
Shellfish Res. 18:19-31.
Hallemian. E.. M. Leffler. S. Mills & S. K. Allen. Jr. 2002. Aquaculture of
triploid Crassostrea ariakensis in Chesapeake Bay: A Symposium Re-
port. Maryland Sea Grant Extension Publication UM-SG-TS-2002-OI
and Virginia Sea Grant Publication VSG-02-03. 20 pp.
Harry, H. W. 1981. Nominal species of living oysters proposed during the
last fifty years. The Veliger 24:39-t5.
Hasan. S. A. 1960. Oyster fishing resources of Pakistan. In: Proc. 4"'
Congress. PIOSA. Karachi. Section B. pp. 167-171.
Hedgecock. D.. G. Li. M. A. Banks & Z. Kain. 1999. Occurrence of the
Kumamoto oyster Crassostrea sikainea in the Ariaki Sea. Japan. Mar.
Biol. 133:65-68.
Hirase. S. 1930. Transactions I. On the classification of Japanese oysters.
Jap. J. Zool 3:1-65.
Hirase. S. 1936. A collection of Japanese shells with illustrations in natural
colors. Tokyo: M. Sanshodo. 217 pp.
Honicll. J. 1910. The practice of oyster culture at Aroachon and its lessons
for India. Madras Fisheries Bull. 5
Hu. J.. Y. Zhong & Y. Lin. 1994. Spat collection technique and spat
collection forecast of Ostrea rivularis Gould. / Xiamen Fisheries Col-
lege/Xiamen Shui Chan Xue Yuan Xue Bao I Chinese) 16:28-33.
Huang. Z.. C. Li. L. Zhang. F. Li & C. Zheng. 1981. On the marine fouling
and boring organisms off Zhejiang southern coast I. Notes on the
fouling organisms of Wenzhou harbor. Acta Oceanologica Sinica/Hai
Yang Xue Bao (Chinese) 3:634-638.
Huzimori. S. 1920. Experiments on Oyster-culture. Hukuoka-Suisan-
Sikenzyo-Gyomukotei-Hokoku.
leyama. H. 1975. Chromosome numbers of three species in three families
of Pteriomorpha. Venus. Jpn. J. Malacol. 34:26-32.
Imai. T. 1978. The revolution of oyster culture. In: A. A. Balkema and
Protterdam, editors. Aquaculture in shallow seas: progress in shallow
sea culture. Translation of Senkai Kanzen Yoshoki. Tokyo: Koseisha
Koseiiku Publishers. 1971. 614 pp.
Imai. T. & M. Hatanaka. 1949. On the artificial propagation of Japanese
common oyster. Ostrea gigas Thunberg by non-colored naked flagel-
lates. Bull. Inst. Agr. Res. Tohoku Univ. 1:1-7.
Imai. T. & S. Sakai. 1961. Study of breeding of Japanese Oyster Crassos-
lrea gigas. Tohoku J. Agr. Res. 12:125-171.
Ke. C. & W. Wang. 2001. Bioaccumulation of Cd. Se. and Zn in an
estuarine oyster (Crassoslrea rivularis) and a coastal oyster (Saccos-
trea glomerala). Aquat. Toxicol. 56:33-51.
Kira. T. 1962. Shells of the Western Pacific in color. Osaka. Japan:
Hoikusha Publishing Co.. Ltd. 224 pp.
Kuroda. T. & T. Habe. 1952. Check List and Bibliography of the Recent
Marine Mollusca of Japan. Tokyo: Leo W. Stach. 210 pp.
Lamy. E. 1929-1930. Revision des Ostrea vivants du Museum National
d'Histoire Naturelle de Paris. Journ. De conchyl. 73 (ser. 4, v. 27). No.
1 (30 April 1929): 1^6: 3 Figs; No. 2 (20 July 1929): 71-108: No. 3
(30 October 1929): 133-168; No. 4 (28 February 1930): 233-275;
Pit. 1.
Langdon. C. J. & A. M. Robinson. 1996, Aquaculture potential of the
Suminoe Oyster (Crassostrea ariakensis Fugita 1913). Aquaculture
144:321-338.
Leitao. A.. P. Boudry. J.-P. Labat & C. Thiriot-Quievreux. 1999, Com-
parative karyological study of cupped oyster species. Malacologia 41:
175-186.
Li. G., Y. Hu & N. Qing. 1988. Population gene pools of big-size cultivated
oysters (Crassostrea) along the Guangdong and Fujian coast of China.
In: Proceedings on marine biology of the South China Sea. Beijing:
China Ocean Press, pp. 51-70.
Li. K., L. Zou. J. Yang. P. Li. L. Yu. Z. Zheng, C. Chen & S. Wang. 1994.
Experimental studies on nourishing Yin effect of oyster complete nu-
tritional tablet. Chinese J. Mar. Drugs/Zhong Guo Hai Yang Yao \Vu
(Chinese) 13:12-16.
Li. X. 1989. A comparative morphology of mantle cavity in some Chinese
oysters. Oceanol. Linmol. Sin./Hai Yang Yu Hu Zhao (Chinese) 20:
502-507.
Li. X. & Z. Qi. 1994. Studies on the comparative anatomy, systematic
classification and evolution of Chinese oysters. Studia Marina Sinica
35:143-177.
Liang. G.. S. Chen & G. Xu. 1983. Report on artificial incubation of Ostrea
rivularis Gould. Mar. Sci./Hui Yang Ke Xue (Chinese) 5:41^3.
Lin. Q.. X. Jia & X. Lu. 1991. Petroleum hydrocarbons in oysters during
the fattening period. In: Proceeding of the 4"' Chinese oceanological
and limnological science conference (Qingdao. Shandong Province).
Beijing: Science Press, pp. 135-141.
Lischke. C. E. 1871. Japanische Meeres-conchylien. II. pp. 160-162.
Littlewood, D. T. J. 1994. Molecular phylogenetics of cupped oysters
based on partial 28S rRNA gene sequences. Mol. Phylogenetics Evol.
3:221-229.
Liu. B-Q. & J-X. Dai. 1998. Studies on genetic diversity in oysters. Cras-
sostrea (in Chinese). / Fisheries China 3:193-198.
20
Zhou and Allen
Loosanoff. V. L. & F. D. Tomnier, 1948. The effect of suspended silt and
other substances on the rate of feeding of oysters. Science 107:69.
Loosanoff, V. L. & J. B. Engle. 1 942. Effects of different concentrations of
plankton forms upon shell movements, rate of water pumping and
feeding and fattening of oysters. Anal. Rec. 84:86.
Lu. C. 1994. Oyster Ostrea rivularis as an indicator of heavy metals
pollution. J. Oceanogr. Taiwan Strait/Tai Wan Hai Xia (Chinese) 13:
14-20.
Lu. C, W. Xie & G. Zhou. 1998a. Crassosnea rivularis as a biomonitor of
zinc pollution of seawater. China Environ. Sci. 18:527-530.
Lu. C. W. Xie & G. Zhou. 1998h. Studies on Crassnslrea rivularis as a
biological indicator of cadmium pollution. J. Fish. Sci. Chimi/Zhon^
Guo Slmi Clian Ke Xiie (Chinese) 5:79-83.
Lu. C, W. Xie & G. Zhou. 1998c. Studies on Crassoslrea rivularis as a
biomonitor for copper pollution in seawater. Mar Environ. Sci./Hai
i'ang Hiian Jing Ke Xue (Chinese) 17:17-23.
Mahadevan, S. 1987. Oyster resources of India. In: K.N. Nayar and S.
Mahadevan, editors. Oyster culture-status and prospects, central marine
fisheries research institute bulletin 38. Cochin. India, pp. 14-16.
Miyazaki, I. 1939. Some notes on the cross-fertilization of Japanese oys-
ters. Bull. Jpn. Soc. Scientific Fisheries 7:257-261.
Moazzam. M. & S. H. N. Rizvi. 1983. Settlement of oyster lanae in Pakistani
waters and its possible implication for settnig up oyster culture. In: Pro-
ceedings of the Symposium on Coastal Aquaculture. Part 2: Molluscan
Culture. Marine Biological Association of India. Cochin. India.
Mok. T. K. 1974. Observations on the growth of the oyster. "Crassoslrea
gigas" Thunberg. in Deep Bay. Hong Kong. Honii Kong Fish. Bull.
4:45-53.
Nelson. T. C. 1928. Relation of spawning of the oyster to temperature.
Ecology 9:145.
Nelson. T. C. 1938. The feeding mechanisms of the oyster I. On the
pallium and branchial chambers of O. virginicu. O. edulis and O. un-
gulala. J. Morphol. 63:1-61.
Nie. Z. Q. 1991. The culture of marine bivalve moUusks in China. In: W.
Menzel, editor. Estuarine and marine bivalve mollusk culture. Boston:
CRC Press, pp. 261-276.
O'Foighil. D.. P. M. Gaffney & T. J, Hilbish. 1995. Differences m mito-
chondrial 165 ribosomal gene sequences allow discrimination among
American [Crassoslrea virginica (Gmelin)] and Asian [C. gi.gas (Thun-
berg) C. ariakensis Wakiya] oyster species. / &/). Mol. Biol. Ecol.
192:211-220.
Orton. J. H. 1936. Observations and experiments on sex change in the
european oyster (Ostrea edulis). Part 5. A simultaneous study of
spawning in 1927 in two distinct geographical localities. Mem. Miis.
Roy. Hist. Nat. Belg. Ser 2, 3. 977.
Patel. B. V. 1979. Bionomical. biochemical and some pollution studies on
oysters. Crassoslrea and Ostrea species. Ph.D. thesis submitted to
Suarashtra University.
Patel, S. K. & K. L. Jetani. 1991. Survey of edible oysters from the
Saurashtra coast. / Curr. Biosci. 8:79-82.
Qasim, R., N. Aftab & S. Barkati. 1985. Biochemical composition and
caloric values of four species of oysters from Karachi Coast. / Phann.
Univ. Kar. 3:51-55.
Qiu. L. & Y. Li. 1983. A report on the raft culture of oyster (Ostrea
rivularis Gould). Trans. Chinese Soc. Malacol. 1:149-156.
Que. H. & S. K. Allen. Jr. 2002. Hybridization of tetraploid and diploid
Crassoslrea gigas (Thunberg) with diploid C ariakensis (Fujila). J.
Shellfish. Res. 27:137-143.
Ranson, G. 1967. Les especies d'huitres vivants actuellement dans le
monde, definies par leurs coquilles larvaires ou prodisso-conchs. Etud
des collections de quelque,s-un de grand musees d'Histoire Naturell.
Rev. Trav. Inst. Peche Maril. 31:127-199.
Rao, K. S. 1987. Taxonomy of Indian oysters. In: K. N. Nayar and S.
Mahadevan, editors. Oyster Culture-Status and Prospects. Central Ma-
rine Fisheries Research Institute Bulletin 38. Cochin. India, pp. 1-6.
Reeve, L. A. 1871. Conchiologica Iconica. London.
Rickards. W. L. & P. C. Ticco. 2002. The Suminoe ovster, Crassoslrea
ariakensis, in Chesapeake Bay: Current status and near-term research
activities. Charlottesville. VA: Virginia Sea Grant. 6 pp.
Rong. S.. S. Shi. Q. Mo. S. Liu, Z. Liang & X. Zhao. 1994. Tetraploid
induced by physical and chemical methods in Jinjiang oyster (Cras-
soslrea rivularis). Acta Oceanologica Sinica/ Hai Yang Xue Bao (Chi-
nese) 13:275-283.
Spark. R. 1925. Studies on the biology of the oyster (Ostrea edulis) in the
Limfjord. with special reference to the influence of temperature on the
sex change. Rep. Dan. Biol. Stat.. No. 30
Sparks. A. K. 1965. A report on the status and potential of the marine
shellfisheries of Kenya.
Stenzel. H. B. 1971. Oysters. In: K. K. Moore, editor. Treatise on inver-
tebrate paleontology. Part N, vol. 3. Mollusca 6. Boulder. CO: Geol.
America Inc. and the University of Kansas.
Taki. I. W. 1933. On the report about the specific difference between
Ostrea gigas and Ostrea rivularis in Ariake Bay by Mr. Fujimori.
Venus 13:365.
Tanaka. Y. 1954. Spawning season of important bivalves in Ariake Bay- II.
Ostrea rivularis Gould and O. gigas Thunberg. Bidl. Jpn. Soc. Scien-
tific Fisheries 19:1161-1164.
Thompson. J.M. 1954. The genera of oysters and the Australian species.
Aiist. J. Mar. Freshwater Res. 5:132-168.
Thomson, J. M. 1954. The genera of oysters and the Australian species.
Aiisl. J. Mar. Freshwater Res 5:132-168.
Tongoe. K. 1981. Oysters in Japan. J. Sci Hiroshima Univ. Ser B. Div I
(Zool) 29:291-118.
Wakiya. Y. 1915. Oysters of Korea. Report of the Fish. Invest.. Govern-
ment General of Korea.
Wakiya. Y. 1929. Japanese food oysters. Jpn. J. Zool. 2:359-367.
Wakiya. Y. 1930. Japanese food oyster. Proceedings of the Fouth Pacific
Science Congress. Batavia. Java. 3:341-348 (Same as text part of Wak-
iya. Y. 19.30)
Wu. X. Z. & J. P. Pan. 2000. An intracellular prokaryotic microorganism
associated with lesions in the oyster. Crassoslrea ariakensis Gould. J.
Fi.^h. Dis. 23:409-414.
Wu. X. Z.. J. Pan & J. Jiang. 1997. Advances in studies on marine mol-
luscan diseases III. Pestology and neoplasia of marine moUusks. Mar.
Sci. BulL/Hai Yang Tong Bao (Chinese) 16(4):82-87.
Xu. Y.. Z. Gao. Y. Ji. X. Chen & C. Qin. 1992. Holocene oyster reef in
Shangyuan-Kengyuan Luoyuan Bay. Fujian and sea level variation. J.
Oceanogr. Taiwan Strail/Tai Wan Hai Xia (Chinese) 1 1:368-371.
Yu. J.. Z. Xiong. W. Tong. S. Shi. X. Zhao & S. Rong. 1993. The Karyo-
types of Jinjiang oyster Crassoslrea rivularis. J. Zhanjiang Fisheries
College/Zhanjiang Shui Chan Xue Yuan Xue Bao (Chinese) 13:27-29.
Zhang, X. & Y. Xie. I960. Culture of major bivalves in China. Bull. Biol./
Sheng Wu Xue Bao (Chinese) 5:197-201.
Zhang. X. & Z. Lou. 1956a. A study on Chinese oysters. Acta Zool. Sin./
Dong Wu Xue Bao (Chinese) 8:65-94.
Zhang. X. & Z. Lou. 1956b. Oyster. Bull. Biol./Sheng Wu Xue Bao (Chi-
nese) 2:27-32.
Zhang. X. & Z. Lou. 1959. Oyster. Beijing: Science Press. 156 pp.
Zhang, X., Z. Qi & Y. Xie. 1959. Moeurs de s'alimenter chez VOstrea
rivularis Gould. Oceunol. Limnol. Sin. /Hai Yang Yu Hu Zhao (Chinese)
2:163-179.
Zhang, X.. Z. Qi. J. Li. X. Ma. Z. Wang. X. Huang & Q. Zhuang. 1960.
Bivalves of South China Sea (Nanhai). Beijing: Science Press. 274 pp.
Zhang, Y.. B. L. Munday & J. Handlinger. 1995. Mass mortality of flat
oysters (Ostrea rivularis) associated with a bloom of Prorocenlrum sp.
in the port of Zhanjiang. South China. Bull. Eur Ass. Fish Pathol
15:61-63.
Zhao. Z.. P. Jiang & A. Li. 1991. Determination of calcium carbonate, trace
elements and amino acids content in oyster shells. Chinese J. Mar
Drugs/ Zhang Guo Hai Yang Yao Wu (Chinese) 10:11-14.
Zhou. M.. Y. Kao & Y. Wu. 1982. Preliminary studies on hybridization of
Crassoslrea gigas with Ostrea rivularis and Ostrea plicatida. J. Fish-
eries China 6:235-241.
Joiinial uj Shellfish Research. Vol. 22, No. 1, 21-3U, 20U3.
CONSUMER RATINGS OF NON-NATIVE (CRASSOSTREA G/GAS AND CRASSOSTREA
ARIAKENSIS) VS. NATIVE {CRASSOSTREA VIRGINICA) OYSTERS
JONATHAN H. (JRABOWSKI.'* SEAN P. POWERS/t CHARLES H. PETERSON,'
MONICA J. POWERS,' AND DAVID P. GREEN"
' University of North Carolina at Chapel Hill. Institute of Marine Sciences, Morehead City,
North Carolina 28557 and 'North Carolina State University; Center for Marine Science and
Technology, Morehead City, North Carolina 28557
ABSTRACT Given suggeslion^ that a non-native oyster be used to replace the depleted native oyster, consumer preference evalu-
ations were conducted to determine how two non-native oysters, Crassostrea gigas and C. ariakensis, when grown in North Carolina
estuaries, were rated by consumers. Tests compared the taste, appearance, and/or aroma of both raw and cooked non-native oysters to
similarly prepared native oysters, C. virginica. In the first series of tests, consumers exhibited a slight preference for raw C. virginica
over raw C. gigas. When cooked, both species were rated equal. In the second series of tests, a larger group of participants ranked the
taste, appearance, and aroma of C. virginica. C. gigas, and C. ariakensis. Participants that tasted raw oysters collectn ely preferred C.
virginica over both non-native species. This preference remained strong regardless of the frequency with which participants consumed
oysters. Preferences for appearance and aroma varied; however, ratings never indicated a preference for either non-native species over
C. virginica. Participants as a whole preferred the taste of cooked C. virginica better than C. gigas. whereas a taste preference did not
exist between cooked C. virginica and C. ariakensisis. Given that participants collectively preferred the taste of both raw and cooked
C virginica to C. gigas. the suitability of C. gigas for substitution in either the raw or steamed oyster market is questionable. For oysters
of similar length (80 to 1 10 mm), dry tissue weight of C. ariakensis was twice that of C. virginica. This higher per-oyster yield suggests
that C ariakensis might be more suitable for a steamed or packaged oyster market where oysters are sold by meat weight rather than
by number. However, these markets often command much lower prices, perhaps rendering unfeasible the aquaculture of this introduced
oyster. Before large-scale introduction of non-native oyster species occurs, consumer preferences should be incorporated into economic
evaluations that include additional economic (oyster prices, market demand and supply functions) and biological information (growth
and survivorship). Profitability expectations generated by the model then need to be weighed against the potential ecological risks and
ecosystem benefits of aquaculture or introduction to the wild for each non-native oyster species.
KEY WORDS: Crassostrea ariakensis. Crassostrea gigas. Crassostrea virginica. economic feasibility, native versus non-native
oysters, raw versus cooked oysters, frequent versus inexperienced consumers, taste test
INTRODUCTION
Landings of the eastern oyster, Crassostrea virginica (Gmelin
1791 ), have declined by over 90'7c during the past century in the
major estuaries of the eastern United Slates (MacKenzie 1983,
Hargis & Haven 1988. Frankenberg 199.S). Habitat degradation
from destructive harvesting techniques (Rothschild et al. 1994,
Lenihan 1999) and mortality induced by bottom-water hypoxia/
anoxia, sedimentation, and parasitic diseases (Seliger et al. 1985,
Ford & Tripp 1996, Lenihan & Peterson 1998, Lenihan et al. 1999)
collectively have contributed to this decline. In North Carolina,
efforts to sustain the oyster fishery over the past several decades
through shell plantings have contributed to but not restored land-
ings, which are less than K/r of historic maxima achieved in the
late 1800s (Frankenberg 1995). Introduction of non-native species
such as C. gigas (Thunberg 1793) or C. ariakensis (Fujita 1913) is
a possible alternative or supplement to continued efforts to restore
native populations, and could resuscitate the oyster industry in the
eastern United States.
The Pacific oyster, C. gigas. accounts for over 9,0'^i of the
world's aquaculture production of oysters (Ayers 1991), and
thrives in shallow, sub-tidal estuaries at higher salinities (Calvo et
al. 1999). Native to Japan and the Korean peninsula (Mann et al.
*Corresponding author. University of Maine at Orono, Darling Marine
Center. 193 Clarks Cove Road. Walpole. ME 04S73. E-mail: jgrabow@
maine.edu
tCurrent address: University of Southern Alabama. Dauphin Island ,Sea
Lab. Dauphm Island. AL 36528
1991). it has been successfully introduced to France, Oregon,
Washington, western Canada. Australia and New Zealand (Shatkin
et al. 1997). C. gigas often establishes populations successfully
when introduced and is successfully cultured in part because it is
highly resistant to the protozoan diseases MSX. Haplosporidiuin
nelsoiii. and dermo. Perkinsus inariiuis (Calvo et al. 1999). MSX
and dermo continue to impede recovery of native oyster popula-
tions along the eastern coast of the US (Ayers 1991. Mann et al.
1991 ). C. gigas also typically reaches harvest size more quickly
than native oysters, leading many culturists to prefer growing C.
gigas over native species (Pollard & Hutchings 1990. Ayers 1991,
Parameswar 1991 ).
In contrast to C. gigas. the Suminoe oyster, C. ariakensis,
currently does not contribute substantially to oyster fisheries of the
world. Despite some taxonomic confusion with C. riviilaris. the
native distribution of C. ariakenis is thought to range from Paki-
stan to Japan, and extends into quite low salinities within the
estuaries that it inhabits (Breese & Malouf 1977, Langdon & Rob-
inson 1996). Like C. gigas. C. ariakensis grows more quickly than
most other oyster species (Byrne 1996. Calvo et al. 2001 ). partly
explaining why many fishermen in North Carolina and Virginia
are advocating its introduction. This species can be grown to mar-
ket size in 12-18 mo in colder waters along the west coast of the
U.S. (Langdon & Robinson 1996). Calvo et al, (2001) also dem-
onstrated that C. ariakensis is resistant to MSX and dermo. Long-
term failure of management to restore native oyster populations
coupled with higher growth rates and disease-resistance of C. gi-
gas and C. ariakensis have created the impetus within industry to
promote triploid aquaculture of and even intentional introduction
21
22
Grabowski et al.
of diploid non-native species along the Atlantic coast of North
America.
Previous intentional and accidental introductions of commer-
cial fishery species have resulted in many well-documented nega-
tive impacts (Naylor et al. 2001 ). For example, the predatory oys-
ter drill, and both MSX and dermo. have been introduced unin-
tentionally through oyster introductions (Carlton 1999, Burreson et
al. 2000). Because of the risks associated with introducing a new
fisheries species, including possible introduction of non-native dis-
eases, competitors, and predators, importation of harmful mi-
crobes, and induction of competition with native species (Ruiz et
al, 2000, Naylor et al. 2001), assessing and contrasting the poten-
tial risks and benefits associated with any proposed introduction
should precede taking action. Here we present results of controlled
trials assessing how oyster consumers rate the palatability of the
two non-native species under consideration for introduction as
compared with C. virginica.
MATERIALS AND METHODS
Two series of tests were conducted to determine consumer
responses to non-native oysters grown in eastern North Carolina
and to compare those responses to native oysters. In both series of
tests, preferences among native, Crassostrea virginica (eastern
oyster) and non-native species, Crassostrea gigas (Pacific oyster),
and Crassostrea ariakensis (Suminoe oyster), were tested sepa-
rately for raw and cooked oysters. Regulations set forth by the
Shellfish Control Authorities in North Carolina mandated that we
inform participants that they were consuming raw or steamed oys-
ters, the location where oysters were grown (non-natives) or har-
vested (natives), and the species of oysters that were being offered.
Participants in the tests were drawn from the local coastal com-
munity surrounding Morehead City, NC and represented a diverse
range of ages (20-81 y old), professions, and knowledge of local
fisheries. Of the 31 individuals that participated in the first taste
test, a few also were among the 96 participants in the second. Each
participant completed and signed a waiver form regarding risk of
raw seafood consumption, completed a deinographic survey, and
provided information on oyster consumption. Finally, participants
were offered water and crackers to assist them to cleanse their
pallets between tasting oysters.
In the first series of tests (conducted on 21 August 2000), we
compared consumer responses to taste and appearance of C vir-
ginica to C. gigas. Triploid C. gigas (approx. 30 mm in length)
obtained from the Virginia Institute of Marine Sciences (VIMS)
had been cultured since February 2000 in plastic mesh vexar cages
held on racks above the sea bottom in Chadwick Bay, Onslow
County, North Carolina. C gigas achieved a length of approx. 80
mm by .August 2000 and were removed from the field and stored
in upwellers at the Institute of Marine Sciences in Morehead City,
North Carolina. Wild C. virginica oysters were harvested in Au-
gust 2000 from both the Newport River and Bogue Sound (Cart-
eret County, North Carolina). Participants were asked to rate un-
labeled raw or cooked oysters in paired contrasts. Separate trials
were performed for raw and cooked oysters. Some participants
were involved in both trials. To begin a trial, two oysters (either
raw or cooked) on the half-shell were presented to each participant,
who then rated each oyster's appearance and (separately) taste on
a scale of I (least desirable) to 5 (Fig. I). Each participant also
specified whether either oyster tasted unappetizing, and, if any
difference was perceived, which one tasted saltier, was more wa-
tery, and was more preferable overall (including an explanation for
any preference). A second pair of oysters was presented to each
participant, who then answered the same set of questions. One of
the pairs of oysters presented a contrast of the two species, whereas
Circle the most appropriate response
1 . Have you eaten raw oyster before? Yes
2. Approximately how many times a year do you eat raw oysters? 1 2
1st Test Oyster #
(A) vs. #
JBL
Yes
No
A
B
Yes
No
A
B
Yes
No
A
B
No
5 >6
1 . Rate tlie appearance of eacli oyster on a scale from 1 to 5 with 5 being the best and 1 being the worst.
A=12 3 45 B=12 3 45
2. Rate the taste of each oyster on a scale from 1 to 5 with 5 being the best and ) being the worst.
A= 12345 B= 12345
3. Did one or both of the oysters taste unappetizing?
Ifso whichone(s): A B Both
4. Did one oyster taste saltier than the other?
Ifso which one:
5. Did one oyster taste more watery than the other?
Ifso winch one:
6. If you preferred one oyster over the other briefly explain why.
Figure I, Survey form used in first taste test.
Consumer Ratings oi- Oysthrs 23
OYSTER TASTE PANEL
Panelist Code # Sample: Raw Steamed Date:
Procedure: Three samples of oysters (either raw or steamed) will be placed in front of you. We
would like for you to taste each of the oysters and evaluate them for their quality attributes by
answenng the questions listed below. Please rate each sample accordini; to their four diait code
by placin^ a mark across the unmarked line that best reflects your opinion, e.g.. like greatly (far
right), neither like or dislike (middle) and dislike greatly (far left).
Note that you are not required to chew or swallow the oyster samples. You may spit the sample
out at any time you need to into the cup provided. You are expected to drink (wash mouth)
between samples with water. If you feel a need to be less fatigued in terms of flavors, aromas and
textures blending together between samples, then you should eat crackers and drink some water.
Ql : When you eat oysters either at home or in a restaurant, what quality attributes are most
important to you?
Q2: How does the appearance of the samples appeal to you? What appearance characteristics do
you like? Dislike?
Dislike Greatly Like Greatly
Q3: How does the aroma of the samples appeal to you? What aroma characteristics do you like?
Dislike?
Dislike Greatly Like Greatly
Q4: How does the texture of the samples appeal to you? What texture characteristics do you
like? Dislike?
Dislike Greatly Like Greatly
Q5: How does the flavor of the samples appeal to you? What flavor characteristics do you like?
Dislike?
Dislike Greatly Like Greatly
Q6: What other attributes do you perceive in the samples?
Please dispose of any left over samples in the appropriate trash container. Be sure to turn your
sensory survey sheet to the project assistant when you leave the room. We appreciate your
time in this study! Results will be available from the project coordinator. THANK YOU!
Figure 2. Survey form used for second laste.
the other presented two C. virginica with one from each site to ters (one of each species, either all raw or cooked) on the half-
determine if grow-out location affected the test results. shell, and asked to rate the appearance, taste, texture and aroma of
The second series of taste tests (conducted on 6 and 7 February each oyster. To quantify a participant's ratings of each oyster, we
2002) evaluated consumer responses to appearance, aroma and measured the distance of the mark along the line, creating a scale
taste of C. virginica, C. gigas, and C. ariakensis. Triploid C. gigas from cm (least desirable) to 10 cm (most desirable). We asked
and C. ariakensis (approx. 30 mm in length) had been obtained participants to indicate profession, age group and the frequency
from VIMS and planted at Chadwick's Bay (3 April 2001 ) and in with which they eat oysters (either raw or cooked, depending on
the Newport River (23 March 2001). Oysters were cultured using whether they were tasting raw or cooked oysters) to determine if
the cage and rack method and achieved harvestable size by January these factors influence their ratings.
2002. C. virginica was also harvested in January 2002 from Chad- We also quantified the wet and dry weights of ^0 replicate
wick's Bay and the Newport River in close proximity to culture oysters (80-1 10 mm shell length) for each of the three species to
operations. In this second set of taste tests, we requested more determine whether percent dry tissue or total dry tissue differed
subtle distinctions by asking participants to rate each oyster tasted among the three species. We determined that the shell length of
by placing a mark on a continuous line that ranged from least to oyster specimens did not vary among the three species with a
most desirable (Fig. 2). Each participant was presented three oys- one-factor analysis of variance (ANOVA: F, ,47. 1.06. P = 0.35).
TABLE 1.
Results of Wilcoxon signed rank tests comparing consumer ratings for taste and appearance of Crassostrea virginica with C. gigas in the
first series of taste tests.
All Part
cipants"
Infrequent Oyster Consumers
Taste .\ppearance
Frequent 0\
Taste
ster Consumers
Oyster Feature
Taste
.\ppearance
.\ppearance
Raw oysters
No. of differences"
2
1
1
1
1
No. of ranlcs < 0"
4
11
3
8
1
3
No. of ranks > 0"
10
4
5
1
5
3
Z value
-1.38
-1.2.^;
-0.56
-2.19
-1.57
-0.63
P value
(1.17
0.21
0.58
0.03
0.12
0.53
Cooked oysters
No. of differences"
2
2
1
1
2
No. of ranks < 0"
6
7
1
2
5
5
No. of ranks > 0"
7
6
3
3
4
3
Z value
-0.21
-1.12
-0.37
-0.27
-0.06
-1.26
P value
0.S.3
0.26
0.72
0.79
0.9.S
0.21
^ Raw data were analyzed collectively and then reanalyzed by subgroup to determine whether those participants who rarely eat oysters have
preferences from those that frequently eat oysters.
" No. of differences indicates the number of participants that rated species equally, no. of ranks <0 indicate participants who rated C. gigns
than C. virginica. and the no. of ranks >0 indicates participants who rated C. virginicn as better than C. gigas.
different
as better
3 1
Participant Category
b. Cooked Oysters
5
Appearance
Participant Category
Figure 3. Results from taste test 1. Taste and appearance ratings of (a) ra« and (b) cooked oysters (Crassostrea virginica vs. C. gigas) are
presented for the following participant categories: I ) all participants, 2 1 infrequent consumers of raw oysters, and 3 1 frequent consumers of raw
oysters. The test in which ('. virginica was ranked significantly lower than C. gigas is marked with an asterisk. Error bars indicate +1 SE.
Consumer Ratings of Oysters
25
Soft tissue was removed from each oyster, placed in a pre-weighed
aluminum pan. and weighed using a Mettler balance (0.001 g).
Tissue was then dried at 60°C in a drying oven for 48 h. and
weighed again to obtain a dry tissue weight (dry weight with pan
minus pan weight). The proportion of each oyster's soft tissue that
is biomass was calculated by dividing the dry weight (tissue
weight minus water weight) by the initial wet weight.
Slatisliial A luilyses
Results from the first taste test were analyzed using the Wil-
co.xon signed rank test. C. virginicci from Bogue Sound were first
compared with C. viri^inica from the Newport River. Because
rankings of native oysters from Bogue Sound and the Newport
River did not differ from each other (in taste: P = 0.97; in ap-
pearance: P = 0..^.^). we concluded that grow-out site did not
affect the taste of native oysters in our study and we analyzed
rankings for C. gigas versus C. virginica from both sites collec-
tively. Separate C. virginica versus C. gigas tests were conducted
for appearance and taste of raw and cooked oysters. Additional
tests were conducted to determine if results varied between groups
that ( 1 ) rarely and (2) frequently (three or more limes per year) eat
oysters to determine if the frequency with which participants eat
oysters affected preferences for native versus non-native oysters.
Results from the second taste test were also analyzed using the
Wilcoxon signed rank test to determine whether participants pre-
ferred the taste, appearance, or aroma of raw and cooked C. vir-
ginica better than C. gigas or C. ariakensis. Each measure of C.
virginica quality was first compared with C. gigas and then to C.
ariakensis for raw and cooked oysters. Two additional series of
Wilcoxon signed rank tests were conducted on the results of the
second series of taste tests (raw and cooked] to determine if rank-
ings of people that eat oysters less frequently differ from those that
often consume oysters. Finally, percent and mean dry tissue
weights of all three species were coinpared using separate one-
factor ANOVA tests. Cochran's test for homogeneity of variance
was perfomied for both response variables (Underwood 1981).
Student-Newman-Keuls (SNK) post hoc tests were conducted on
significant ANOVA results {P < 0.05) to determine which of the
three species differed from each other. The SNK test was selected
because we conducted a balanced experiment with a priori pre-
dictions and a fixed factor (Day and Quinn 1989).
RESULTS
First Series of Tests (C. virginica versus C. gigas)
Collectively, survey participants ranked the taste of raw C.
virginica slightly higher and its appearance slightly lower than C.
gigas, but neither difference was significant (Table 1; Fig. 3). Of
the 16 participants offered raw oysters, 10 preferred C. virginica.
three preferred C. gigas, and three had no preference. Only two of
the 16 considered C. gigas unappetizing and only one replied that
C. virginica was unappetizing. Of the nine raw oyster tasters who
rarely eat raw oysters, the appearance of C. gigas was ranked
significantly higher than C. virginica. but the taste ratings were
similar. Among the seven raw oyster tasters who frequently con-
sume raw oysters, the taste of C. virginica was rated slightly higher
than C. gigas: five of the seven preferred C. virginica. but low
sample size more than likely rendered this difference non-
significant (Table 1). Ratings of the appearance of the two species
did not differ among this subgroup of tasters.
Collectively, tasters of cooked oysters did not distinguish be-
tween species in taste or appearance (Table 1; Fig. 3). Of the 15
TABLE 2.
Results of Wilcoxon signed rank tests comparing consumer ratings for taste, appearance, and aroma of Crassostrea virginica h ith C. gigas
and C. ariakensis during the second series of raw oyster taste tests.
C". virginica vs. C. gigas
C
virginica vs. C
ariakensis
Oyster Feature
Taste
.Appearance
.Aroma
Taste
Appearance
.\ronia
All participants'*
No. of differences"
2
3
15
5
2
16
No. of ranks < O"
22
35
31
32
40
35
No. of ranks > 0''
64
53
44
51
49
39
2 value
-4.75
-2.4
-0.88
-2.96
-0.14
-0.50
P value
<0.0001
0.02
0.38
0.003
0.89
0.62
Infrequent consumers of raw oysters
No. of differences''
2
3
1
5
No. of ranks < 0"
5
13
15
9
14
15
No. of ranks > O''
24
17
14
20
17
12
Z value
-3.43
-O.ftI
-0.28
-2.32
-0.27
-0.99
P value
0.0006
0.54
0.78
0.02
0.79
0.32
Frequent consumers of raw
oysters
No. of differences''
2
1
12
5
1
11
No. of ranks < O"
16
T)
16
T")
26
20
No. of ranks > O''
40
35
29
31
31
26
Z value
-3.65
-2.48
-1.29
-2.07
-0.35
-1.22
P value
0.()()()3
(1.1)1
0.2(1
0.04
0.72
0.22
■■ Raw data were analyzed collectively and then reanuly/ed by subgroup to delermuic v\hcther participants who rarely eat raw oysters liave dilferenl
preferences from those who frequently eat them.
No. of differences indicates the number of participants who rated species equally, no. of ranks <0 indicate participants who rated the non-native species
as better than C. virginica. and the no. of ranks >0 indicates participants who rated C. virginica as better than the non-native species.
26
Grabowski et al.
■ C. virginica
D C. gigas
■ C ariakensis
All Participants
Rarely Eat Oysters Frequently Eat Oysters
b. Appearance
10 1
■ C. virginica
a C. gigas
■ C. ariakensis
Al! Participants
Rarely Eat Oysters Frequently Eat Oysters
■ C. virginica
a C. gigas
■ C. ariakensis
All Participants Rarely Eat Oysters Frequently Eat Oysters
Figure 4. Results from taste test 2: raw oysters, (a) Taste, (b) appearance, and (c) aroma ratings of raw Crassostrea virginica. C. gigas, and C.
ariakensis for the following participant groups: I) all participants. 2| infrequent consumers of raw oysters, and 3) frequent consumers of raw
oysters. Tests in which C. virginica was ranked higher than non-nati\c oysters are marked with * for C. gigas and # for ('. ariakensis. Error bars
indicate +1 SE.
Consumer Ratings of Oysters
27
TAIU.E 3.
Results of Wilcoxon signed rank tests coniparinj; consumer ratings lor taste, appearance, and uronia of Crassuslrea yirf;iiiica with C. gigas
and C ariakeiisis during the second series of cooked oyster taste tests.
lire
C. virginica vs. C. gigas
C. virginica vs. C. ariakensi.
Oyster Feat
Taste
Appearance
Aroma
Taste
Appearance
Aroma
All participants'
No. of differences''
3
3
15
3
3
14
No. of ranks < 0"
33
49
37
38
40
36
No. of ranks > 0"
54
39
38
47
47
38
Z value
-2.5.'i
-0.89
-0.23
-0.68
-1.22
-0.003
P value
(1.01
0.38
0.82
0.49
0.22
0.99
Infrequent consumers of
cooked oysters
No. of differences"
2
4
2
3
No. of ranks < 0"
13
19
15
13
16
17
No. of ranks > 0'"
19
12
12
18
14
10
Z value
-1.98
-1.53
-0.32
-0.36
-0.51
-1.59
P value
0.05
0.13
0.75
0.72
0.61
0.1 1
Frequent consumers of cooked
oysters
No. of differences''
3
1
11
3
1
11
No. of ranks < 0"
19
29
21
24
24
18
No. of ranks > 0"
35
27
26
29
32
28
Z value
-1.83
-0.15
-0.57
-0.90
-1.82
-1.26
P value
0.07
0.88
0.57
0.37
0.07
0.21
" Cooked oyster data were analyzed collectively and then reanalyzed by subgroup to determine whether participants v\ ho rarely eat cooked oysters have
different preferences from those that frequently eat them.
'' No. of differences indicates the number of participants who rated .species equally, no. of ranks <0 indicate participants who rated the non-native species
as better than C. virginica. and the no. of ranks >0 indicates participants who rated C. virginica as better than the non-native species.
participants tasting cooked oysters, seven preferred C. virginica.
six preferred C. gigas. and two had no preference. Only one of the
1 .5 considered cooked C. gigas to be unappetizing, whereas two
replied that C. virginica was unappetizing. Splitting participants
out into inexperienced and frequent eaters of cooked oysters failed
to detect any pattern of species preference in taste or appearance of
the cooked oysters (Table 1; Fig. 3).
Second Series of Tests (C. virginica versus C. gigas or C. ariakensis^
In the second taste test, raw oyster tasters collectively ranked
the taste of C. virginica significantly higher than both C. gigas and
C. ariakensis (Table 2: Fig. 4). Appearance of C. virginica was
rated significantly above C. gigas but not above C. ariakensis
(Table 2: Fig. 4). Neither of the paired species contrasts distin-
gLushed native from non-native oysters by aroma. Infrequent oys-
ter eaters ranked the taste of raw C. virginica significantly above
both C. gigas and C. ariakensis. but rankings by appearance and
aroma did not vary among the three species (Table 2; Fig. 4).
Frequent oyster eaters ranked the taste of raw C. virginica signifi-
cantly above both ni)n-native species and the appearance of C.
virginica over C. gigas but not different from C. ariakensis. Aroma
rankings did not differ in either contrast of pairs of oysters (Table
2; Fig. 4).
Tasters of cooked oysters collectively rated the taste of cooked
C. virginica significantly more than C. gigas (Table 3; Fig. 5) but
did not distinguish between cooked C. virginica and C. ariakensis.
Ratings of appearance and aroma did not differ between cooked
native and non-native oysters in any contrast. The subgroup
formed by infrequent consumers of cooked oysters also ranked the
taste of cooked C. virginica significantly better than C. gigas but
failed to distinguish between cooked C. virginica and C. ariakensis
(Table 3; Fig. 5). These relatively inexperienced oyster eaters did
not rate the appearance or aroma of native oysters differently from
non-native species. Finally, frequent oyster eaters ranked the taste
of C. virginica marginally above C. gigas but not significantly
higher than C. ariakensis. For these experienced oyster eaters,
aroma and appearance rankings did not differ significantly be-
tween cooked native and non-native oysters, though the appear-
ance of C. virginica was ranked marginally higher than C. aria-
kensis (Table 3; Fig. 5).
Dry Weight
Percent dry weight of soft tissues (dry weight/wet weight) did
not significantly differ among the three species (Table 4). Prior to
this analysis, percent dry weight data were transformed using a
square root transformation to remove heterogeneity among vari-
ance groups. Total dry tissue weight (g) of C. ariakensis was
significantly greater than that of C. gigas or C. virginica. and the
dry tissue weight of C. gigas was greater than that of C. virginica
(SNK post hoc comparisons; Fig. 6). Because average shell length
did not differ among species, this analysis reflects biomass for
oysters of a fixed range of harvestable lengths (80-1 10 mm).
DISCUSSION
As managers consider use of non-native species to enhance or
restore fisheries, they should weigh carefully the risks and poten-
tial benefits. Decisions on species introductions are driven by a
variety of social and political pressures, often with insulTicient
attention to potential ecological risks or economic benefits (An-
drews 1980). In North Carolina, it is unclear, for example, how
current market prices would adjust to an increase in oyster supply
(Lipton & Kirkley 1994). Oyster and clam markets in the state
have already endured low demand and reduced prices that threaten
the economic viability of both culture operations and wild harvest
28
Grabowski et al.
■ C. virginica
D C. gigas
■ C. ariakensis
All Participants
Rarely Eat Oysters Frequently Eat Oysters
b. Appearance
All Participants
Rarely Eat Oysters Frequently Eat Oysters
c. Aroma
10
o
o
■ C. virginica
' D C. gigas
\ ■ C. ariakensis
All Participants
Rarely Eat Oysters
Frequently Eat Oysters
Figure 5. Results from taste test 2: cooked oysters, (a) Taste, (b) appearance, and (c) aroma ratings of cooked Crassostrea virginica, C. gigas, and
C. ariakensis for the following participant groups: I) all participants, 2) infrequent consumers of cooked oysters, and }) frequent consumers of
cooked oysters. Tests in which C. virginica was ranked higher than non-nati\e oysters were marked with * for C gigas and # for C. ariakensis.
Error bars indicate +1 SE.
Consumer Ratings of Oysters
TABLE 4.
Ki'siills III ANOVA comparison of pcrctnl (lr\ tisMii' Hciyhl ol' Mif(
tissiKs and tiilal dr> tissue weight for Crasso^lrcii \irf;iitica. C. giaas.
and C. ariakeiisis.
Source of Variance
df
MS
F Value P Value
Percent dry tissue weight
Oyster species
Residual
Total dry tissue weight
Oyster species
Residual
2 0.001 1.3X
147 0.001
0.26
2 2.32 28.82y <0.()Oni
147 0.08
fisheries. Although the transport of C. ,?/,i;(i.v from the west coast
for sale in the eastern United States has increased since the col-
lapse of native stocks on the east coast, it is unclear whether
consumers in the eastern United States prefer a particular oyster
species (Lipton et al. 19921 and how such preferences may vary
with targeted market (e.g.. raw on the half-shell versus steamed,
etc.). In this study, we set out to identify (1) whether the taste of
non-nati\e oysters is acceptable to oyster consutners in North
Carolina and (2) whether consumer ratings differ and preferences
exist among raw and cooked C. virginicci, C. gigas. and C. ciria-
kcnsis. Our purpose was to begin the process of evaluating the
market potential of the two non-native species of oyster in North
Carolina and. by extension, the other east coast states where con-
sumers are accustomed to eating native oysters.
Although consumer ratings of taste and appearance provided no
consistent pattern of preference in the first taste test (i.e.. for taste
C. virginicci > C. gigiis. and for appearance C. gigas > C. vir-
ginica). the majority preferred raw C. virginicii more than C. gi-
gas. These findings suggest that consumer preference for raw oys-
ters may be dictated more by taste than appearance. Cooking re-
moved any indication of a difference between species in taste or
appearance, indicating that non-native C. gigcis may be suitable for
local cooked oyster markets. When asked, few participants con-
sidered either C. gigas or C. virginica unappetizing regardless of
C. ariakensis
C. virginica
C gigas
Species
Figure 6. Mean dry tissue weight (g) of Crassostrea virginica. C. gigas,
and C. ariakensis. Error bars indicate +\ SE. Results of SNK post-hoc
mean comparisons are Indicated with letters above the error bars, and
species with different letters above them are signitkantl) different at
P < (1.(15.
preparation (raw or cooked), implying that non-native C. gigas
might be acceptable. Fisheries managers may wish to assess next
whether consumer demand exists for an acceptable but less pref-
erable oyster and if lower preference implies a reduction in market
price before allowing introduction of C. gigas to the east coast.
The larger numbers of participants in the second series of tests
pro\ided greater ability to resolve differences among oysters and
included contrasts with the second non-native species. C. ariak-
ensis. Participants in the raw oyster tests collectively indicated a
strong taste preference for C. virginica over either non-native spe-
cies. This preference held regardless of whether consumers rarely
or frequently eat oysters. Because frequent consumers eat a dis-
proportionately large amount of the raw oysters consumed in
North Carolina, these results raise concern about the suitability of
either non-native species for local raw oyster markets. Though
appearance and aroma preferences were not as definitive, consum-
ers collectively preferred the appearance of raw C. virginica to C.
gigas. which raises further doubt about the marketability of raw C.
gigas on the east coast.
Tasters of cooked oysters in the second test exhibited weaker
preferences among oysters. Yet participants collectively, as well as
the subset who rarely consume oysters, preferred the taste of
cooked C. virginica more than C. gigas. and frequent consumers of
cooked oysters expressed a slight preference for the taste of
cooked C. virginica more than C. gigas. Consumers as a whole, as
well as the subset who frequently eat cooked oysters, did not
exhibit a taste preference for cooked C. virginica or C. ariakensis.
suggesting that C. ariakensis may be more suitable for steamed
and packaged oyster markets. Because the weight of C. ariakensis
oysters was double that of C. virginica of a given length and C.
ariakensis grows to market size much more quickly than the native
oyster (Calvo et al. 2001). the Suminoe oyster might be more
successful in markets that sell by meat weight. However, the high
costs of triploid aquaculture need to be considered in assessing the
economic viability of this industry. On the other hand, our results
show that the most widely marketed and consumed oyster in the
world. C. gigas. is not rated as high by North Carolina consumers
as the eastern native oyster. C. virginica. The alternative non-
native oyster. C. ariakensis. is rated at least as high and in some
contrasts higher than C. gigas. Thus, if the Suminoe oyster could
be produced at sufficiently low cost, then it should compete fa-
vorably with C. gigas for market share.
Because of serious environmental risks associated with intro-
ducing a non-native species as a self-replicating wild population or
even for culture as triploids. we argue that an analysis of economic
viability is necessary for responsible decision making by fisheries
managers. Such an analysis would include our new information on
consumer perceptions, ratings, and rankings of alternative species
of oysters under consideration for use. A complete economic
analysis to follow our study of consumer ratings and preferences
would involve a model to convert these consumer ratings into
prices. Additional costs of each type of culture and impacts on
market supply and demand must also be assessed. Collapsing oys-
ter fisheries along the Atlantic coast and declining water quality
collectively have eroded consumer demand for oysters, such that
current oyster markets are probably less elastic. Therefore, an in-
crease in supply from successful introduction of non-native oysters
in North Carolina could result in a corresponding decrease in oys-
ter prices (Lipton & Kirkley 1994). especially within smaller raw
oyster markets. Biological information on growth and mortality
rates of non-native oyster species must be acquired and compared
with nati\e oysters. Given that non-native oysters were generally
30
Grabowski et al.
less preferable than the native eastern oyster in our study and that
producing cultured oysters from triploid seed is expensive, suc-
cessful culture of triploid oysters would require a substantial bio-
logical benefit in the form of shorter time to market and/or higher
survival. Inclusion of this information into a comprehensive eco-
nomic analysis of potential benefits and costs of introduction
would enable managers to assess whether the environmental risks
are worth taking. Finally, restoration of any oyster will have posi-
tive effects in restoring water quality and compensating for estua-
rine eutrophication (Jackson et al. 2001. Newell et al. 2002), such
that this ecosystem benefit should be included in a complete eco-
nomic evaluation of any potential oyster introduction. If the intro-
duced oyster were to form reefs, then further ecosystem benefits of
habitat enhancement (Lenihan et al. 2001) should also be incor-
porated.
ACKNOWLEDGMENTS
The authors thank Rachael Wagaman, Christina Tallent. David
Gaskill, Hal Sumnierson. and Chris Stewart for culturing the oys-
ters, assistance conducting the two food surveys and quantifying
oyster tissue weights. Stan Allen, Jr., of the Virginia Institute of
Marine Sciences provided disease-free triploid seed and much
guidance. This research was supported by the North Carolina Gen-
eral Assembly through the Rural Development Foundation and the
Fishery Development Foundation and the North Carolina Depart-
ment of Natural Resources.
LITERATURE CITED
Andrews. J. D. 1980. A review of introductions of exotic oysters and biologi-
cal planning for new importations. Mar. Fisheries Rev. 42:1-11.
Ayers. P. 1991. Introduced Pacific oysters in Australia. In: J. Sutherland
and R. Osman. editors. The ecology of Crussosrreu gi.i^as in Australia.
New Zealand. France, and Washington State. College Park. MD: Mary-
land Sea Grant, pp. .V7.
Breese, W. P. & R. E. Malouf. 1977. Hatchery rearing techniques for the
oyster Crassostrea rivularis Gould. Aquaciitiiire 12:123-126.
Burre.son. E. M.. N. A. Stokes & C. S. Friedman. 2000. Increased virulence
in an introduced pathogen: Haplosporidiiim nelsoni (MSX) in the east-
ern oyster Crassoslreu virt;inicu. J. Aqiianc Aiiim. Health 12:1-8.
Byrne. R. J. editor. 1996. Strategic plan for molluscan shellfish research;
including a rational plan for testing application of non-native oyster
species. Richmond, VA: Report of the Virginia Institute of Marine
Sciences. House Document No. 16. Report to the Governor and Gen-
eral Assembly of Virginia. Commonwealth of Virginia.
Calvo, G. W.. M. W. Luckenbach, S. K. Allen, Jr. & E. M. Burreson. 1999.
Comparative field study of Crasso.strea git^as (Thunberg. 1793) and
Crassostrea virginica (Gmelin. 1791) in relation to salinity in Virginia.
/ Shellfish Res. 18:465-473.
Calvo. G. W.. M. W. Luckenbach. S. K. Allen. Jr. & E. M. Burreson. 2001.
A comparative field study of Crassostrea ariakensis (Fujita 1913) and
Crassostrea virginica (Gmelin 1791) in relation to salinity in Virginia.
J. Shellfish Res. 20:221-229.
Carlton, J. T. 1999. Molluscan invasions in marine and estuarine commu-
nities. Malacologia 41:439^54.
Day, R. W. & G. P. Quinn. 1989. Comparisons of treatments after an
analysis of variance in ecology. Ecol. Monogr. 59:433—463.
Ford, S. E. & M. R. Tripp. 1996. Diseases and defense mechanisms. In: V.
S. Kennedy. R. I. E. Newell & F. Ebele. editors. The eastern oyster,
Crassostrea virginica. College Park. MD: Maryland Sea Grant, pp.
581-660.
Frankenberg, D. 1995. Report of North Carolina Blue Ribbon Advisory
Council on Oysters. Raleigh. NC: North Carolina Department of En-
vironment, Health, and Natural Resources.
Hargis. W. J.. Jr. & D. S. Haven. 1988. The imperiled oyster industry of
Virginia: A critical analysis with recommendations for restoration. Spe-
cial report Number 290 in applied marine science and ocean engineer-
ing. Gloucester Point. VA: Virginia Institute of Marine Sciences.
Jackson. J. B. C. M. X. Kirby. W. H. Berger. K. A. Bjomdal. L. W.
Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke, J. Erlandson, J. A.
Estes, T. P. Hughes, S. Kidwell, C. B. Lange, H. S. Lenihan, J. M.
Pandolfi, C. H. Peterson, R. S. Steneck, M. J. Tegner & R. R. Warner.
2001 . Historical overfishing and the recent collapse of coastal ecosys-
tems. Science 293:629-638.
Langdon, C. J. & A. M. Robinson. 1996. Aquaculture potential of the
Suminoe oyster {.Crassostrea ariakensis Fugita 1913). Aqiiaciilliirc
144:321-338.
Lenihan, H. S. 1999. Physical-biological coupling on oyster reefs: How
habitat structure infiuences individual performance. Ecol. Monogr. 69:
254-275.
Lenihan. H. S.. F. Micheli. S. W. Shelton & C. H. Peterson. 1999. The
inlluence of multiple environmental stressors on susceptibility to para-
sites: An experimental determination with oysters. Limnol. Oceanogr.
44:910-924.
Lenihan. H. S. & C. H. Peterson. 1998. How habitat degradation through
fishery disturbance enhances impacts of hypoxia on oyster reefs. Ecol.
.Appl. 8:128-140.
Lenihan, H. S., C. H. Peterson. J. E. Byers. J. H. Grabowski. G. W. Thayer
& D. R. Colby. 2001. Cascading of habitat degradation: Oyster reefs
invaded by refugee tlshes escaping stress. Ecol. Appl. 1 1:764-782.
Lipton. D. W. & J. Kirkley. 1994. A profile of the oyster industry: North-
eastern United States. Virginia Sea Grant Marine Resource Advi.sory
No. 54. Gloucester Point, VA: Virginia Institute of Marine Sciences.
Lipton. D. W., E. F. Lavan & 1. E. Strand. 1992. Economics of molluscan
introductions and transfers: The Chesapeake Bay dilemma. J. Shellfish
Res. 11:511-519.
MacKenzie, C. L.. Jr. 1983. To increase oyster production in the north-
eastern United States. Mar Fisheries Rev. 45:1-22.
Mann, R.. E. M. Burreson & P. K. Baker 1991. The decline of the Virginia
oyster fishery in Chesapeake Bay: Considerations for introduction of a
non-endemic species. Crassostrea gigas (Thunberg. 1973). / Shellfish
Re.'i. 10:379-388.
Naylor. R. L., S. L. Williams & D. R. Strong. 2(J0I. Aquaculture— a
gateway for exotic species. Science 294: 1 655- 1 656.
Newell, R. 1. E., J. C. Comwell & M. S. Owens. 2002. Influence of
simulated bivalve biodeposition and microphyrobenthos on sediment
nitrogen dynamics: A laboratory study. Limnol. Oceanogr. 47:1367-
1379.
Parameswar, D. 1991. Introduced Pacific oysters in New Zealand. In: J.
Sutherland & R. Osman. editors. The ecology of Crassostrea gigas in
Australia. New Zealand, France, and Washington State. College Park.
MD: Maryland Sea Grant, pp. 9-11.
Pollard, D. A. & P. A. Hatchings. 1990. A review of exotic marine organ-
isms introduced to the Australia region. II. Invertebrates and algae.
Asian Fisheries Sci. 3:223-250.
Rothschild. B. J., J. S. Ault, P. Goulletquer & M. Herat. 1994, Decline of
the Chesapeake Bay oyster population: A century of habitat destruction
and overfishing. Mar Ecol. Prog. Ser. 1 1 1:29-39.
Rui/. G. .M.. P. W. Fofonoff, J. T. Carlton, M. J. Wonham & A. H. Hines.
2000. Invasion of coastal marine communities in North America: Ap-
parent patterns, processes, and biases. Ami. Rev. Ecol. Sysrematics 31:
481-531.
Seliger, H. H.. J. A. Boggs & W. H. Biggley. 1985. Catastrophic anoxia in
the Chesapeake Bay in 1984. Science 228:70-73.
Shatkin, G., S. E. Shumway & R. Hawes. 1997. Considerations regarding
the possible introduction of the Pacific Oyster [Crassostrea gigas) to
the Gulf of Maine: A review of global experience. J. Shellfish Res.
l6:46.^-+77.
Underwood, A.J. 1981. Techniques of analysis of variance in experimental
marine biology and ecology. Oceanogr. Mar Biol. Anna. Rev. 19:513-
605.
Jo:inml of Slicll/isl, Rcscairh. Vol. 22. No. 1. .M-.^S. 20(13.
TAXONOMIC STATUS OF FOUR CRASSOSTREA OYSTERS FROM CHINA AS INFERRED
FROM MITOCHONDRIAL DNA SEQUENCES
ZINIU YU,'"* XIAOYU KONG,' LIUSUO ZHANG,' XIMING GUO.- AND JIANHAI XIANG'
^College of Fisheries. Ocean University of Qingihio. Qingclao 266003. Peoples Republic of China:
-Haskin Shellfish Research Laboratory. Institute of Marine ami Coastal Sciences. Riitiiers University.
Port Norrls. New Jersex 0S.U9: and ''Institute of Oceanology. Chinese Academy of Sciences, Qingdao
266071. Peoples Republic of China
ABSTRACT It has been presumed ihat there are tour eoaiiiion Cra\snstrea oyster species along the eoast ol China; the Pacitic oyster
(Crassostrea gigas), Zhe oyster (C plicatula). Suminoe oyster (C ariakeiisis). and Dalianwan oyster (C. talienwbanensis). Classifi-
cation and species identification of these Crassostrea oysters have been difficult because of morphologic plasticity. In this article,
phylogenetic analysis was performed to clarify taxonomic status of these species using mitochondrial DNA sequence data. Nucleotide
sequences of a 443-bp fragment of ribosomal RNA gene and a 579-bp segment of cytochrome c oxidase I gene were obtained through
sequencing and used for analysis. Genetic distances among the four species, using C. virgiiiica as outgroup, were computed based on
the sequence data, and phylogenetic trees for the five species were generated. The divergence between C. gigas and C. talienwhanensis
was very low. as was that between C. pticaiula and C ariakeiisis. Phylogenetic analysis showed that haplotypes of C. gigas and C.
lalieimhaiieiisis clustered in one clade and those of C. plicaluta and C ariakeiisis in another one. Our data suggest that C. gigas and
C latieimlumensis may be the same species. However, the lack of divergence between C. plicaltila and C. ariakeiisis samples may
indicate that the C. plicaliila specimen we sampled could actually be a morph of C. ariakeiisis living in high salinity habitats. More
work is needed for confirmation.
KEY WORDS: Crassostrea oysters, taxonomy, phylogenetic analysis, 16S rDNA, COI gene, nucleotide sequences
INTRODUCTION
Ainotig the over 20 species of oysters recorded in China, four
Crasso.strea species are most cotnmon and of commercial impor-
tance; the Pacific oyster (Crassostrea gigas), Zhe oyster (C. pli-
catula). Suminoe oyster (C. ariakeiisis). and Dalianwan oyster (C
talienwhanensis; Zhang et al. 1956, Qi I9S9). The Pacific oyster,
which occurs naturally along the coast of China, is a well-
recogni/ed species. However, most of the Pacific oysters cultured
in China were originally introduced from Japan or Korea (Wang et
al. 1993). The Zhe oyster is commonly found along the entire coast
of China. It is relatively smaller in body size than the Pacific and
Suminoe oysters and thin-shelled (Qi 1989, Guo et al. 1999).
Suminoe oysters are also distributed along most of the coast of
China with two major populations, one in the estuaries of Yellow
river and the other in Guangxi and Guangdong in southern China.
It can tolerate a wide range of salinity but prefers low-salinity
estuaries and riverbeds (Torigoe 1981. Li & Qi 1994). The Dalian-
wan oyster occurs mainly in areas along the coast of Liaoning and
Shandong provinces in the North (Zhang et al. 1956, Qi I9S9).
Because of the morphologic plasticity, there have been dis-
agreements about the taxonomic status of the four Crassostrea
types and difficulties in their identification. Some believed that the
Pacific and Dalianwan oysters are different species (Zhang et al.
1956, Qi 1989), whereas others argued that the Dalianwan oyster.
described by Zhang et al. ( 1956), is the Pacific oyster, or a variety
of Pacific oyster (Torigoe 1981, Li & Qi 1994). In addition, some-
times the discrimination of Pacific and Suminoe oysters was am-
biguous with shell morphology, although it is distinguishable w ith
some body anatomic features (Li & Qi 1994). The most common
oysters found in the rocky intertidal zone and extensively cultured
in the south are generally believed to be the Zhe oyster, although
♦Corresponding author. Tel: 856-785-0074; Fax: 856-7S5-I544; E-mail:
carlzyu @ hsrl.rutgers.edu
Li and Qi (1994) assumed it was the Pacific oyster. Liu et al.
{ 1 998 ) compared RAPD data from several Crassostrea species and
concluded that the Dalianwan oyster, Zhe. and Pacific oysters were
sister species with each other.
Because of this confusion, further study, especially with DNA
markers, is needed. DNA polymorphisms are useful tools for eco-
logical, genetic, and evolutionary studies of both terrestrial and
marine organisms, and DNA sequences can be used to detect dif-
ferences among species, populations, or individuals. Proper iden-
tification of oyster stocks will assist management, including con-
servation and the sustainable use of these resources. Past efforts to
investigate and identify differences among populations and species
of oysters along the coast of China have provided useful but in-
conclusive information {Liu et al. 1998, Yatig et al. 2000).
Because of its fast sequence evolution and inaternal. nonrecom-
bining nature of inheritance in animals, mitochondrial genes have
proved a powerful tool in phylogenetic studies and species iden-
tification (Banks et al. 1993, Littlewood 1994, Jozefowicz et al.
1998, Lapegue et al. 2002). The I6S rRNA and COI gene frag-
ments are popular choices for phylogenetic analysis (O'Foighil et
al. 1995, O'Foighil et al. 1998. Canapa et al. 2000). In this study,
mitochondrial 1 6S rRNA and COI gene fragments from these four
putative species were amplified and sequenced for phylogenetic
analysis.
MATERIALS AND METHODS
Sampling and Polymerase Chain Reaction (PCR) Amplifications
Crassostrea gigas samples (eight specimens) were obtained
from a hatchery broodstock in Shandong province; C ariakeiisis
samples (seven individuals) were collected from estuaries of the
Yellow River, in Yantai, Shandong province, which is a typical
habitat of this species in north China. C talienwhanensis was
sampled from Dalian (five individuals). Liaoning province and
Rongcheng (five individuals). Shandong province. C. plicatula
31
32
YU ET AL.
samples were collected from Qingdao (five specimens). Shandong
province and Wenzhou (five specimens). Zhejiang province. Sam-
pling sites are showed in Figure 1 . C. virf^iiiica was collected from
Delaware Bay in the United States. Morphologic identification was
made according to that described in Zhang et al. (1956). Qi (1989),
Torigoe (1981). and Li and Qi (1994).
Total DNA was e,xtracted from mantle tissue using an extrac-
tion kit (Pure Gene, Centra, USA). Fragments of the 16S rDNA
and COI gene were amplified using two pairs of universal primers:
1 6sar-L/ 1 6sbr-H: 5 ' -GCCTGTTTATCA AAA ACAT-3 75 ' -
CCGGTCTGAACTCAGATCACGT-3'(Palumbi 1991 );
COIL 1 490/CO1H2 1 98: 5 '-GGTCAACAAATCATAAAGATAT-
TGG-37 5'-TAAACTTCAGGGTGACCAAAAAATCA-3'
(Folmer et al. 1994).
Amplification of the products was performed using a PTC- 100
thermal cycler (MJ Research. USA). The 100-p.L amplification
reaction contained 2.0 niM MgCK; 200 (j.M of each dNTP: 0.2 (xM
each primer; 2.5 p.L of template DNA; and 2.5 units of Taq poly-
merase (Sangon. Canada) with supplied buffer. For all amplifica-
tions, hot-start PCR was initiated by addition of polymerase and
primers after an initial 2-min denaturization at 80°C. The PCR
cycling profile was as follows: 35 cycles at 94'C/45 sec. 48°C
(COI) or 50°C ( I6S)/1 min and at 72°C/1 mm. with a final exten-
sion at 72"C for 7 min.
Sequencing
PCR products were purified using UNIQ-5 Column PCR Prod-
uct Purification Kit (Sangon. Canada), ligated into pMD18-T Vec-
tor by following instniction of Takara DNA Ligation Kit ver.2
(Takara. Japan) and used to transform competent JM109 Escheri-
chia coli cells using standard protocols. Recombinant colonies
were identified by blue-white screening. Inserts of the correct size
jr^
VJliaoning J'
'NortK
KorSa
BEIJING^JJ
(^^^ 'On Inn y
"X
/hebei
/SHANDONG 1 j/
/fiurwo
1
o^^-i«'Ss4^ a fi y u 03 s cmj
^
r^JIANGSU
, anhV?^*''"'"
^^/>
'^'i/^-y ZHEJIANG
UIAN
K /^^ •
IC hii IJ
GXI^ ^/
Figure \. .V map nt sampling area »ith sampling sites underlined.
were detected via restriction enzyme digestion by EcoRI and
HiiicHU. Vector DNA containing the desired insert was further
purified using Pharmacia EasyPrep Kit. Sequencing was per-
fonned for both strands of every sample on an ABI PRISM 377XL
DNA Sequencer using ABI PRISM BigDye"^"^ Terminator Cycle
Sequencing Ready Reaction Kit w ith AmpliTaq DNA Polymerase.
FS (Perkin-Elmer. USA).
Dala Analysis
The 16S and COI sample sequences, along with those already
obtained for C. gigcis and C. ariakensis (0"Foighil et al. 1995.
1998; courtesy of Dr. D. OToighil) were aligned with CLUSTAL
W (Thompson et al. 1994). For clarity and convenience in com-
paring with other published sequences, the sequences were
trimmed to the same length as published sequences after align-
ment. Parsimony analysis was made with Phylip (Ver.3.56C.
Felsenstein 1989) using the program DNAPARS with C. virginica
as the out-group. Bootstrap analysis with 1000 replication was
performed by the SEQBOOT and CONSENSE programs. Consen-
sus phylogenetic trees were drawn with DRAWGRAM program in
the Phylip package. Pair-wise sequence divergence between hap-
lotypes and species were estimated by the DNADIST program of
Phylip according to Kimura's two-parameter model (Kimura
1980).
RESULTS
A PCR fragment of 488 bp from the mitochondrial IdS ribo-
sonial gene and a fragment of 649 bp from the mitochondrial COI
gene were obtained and sequenced for 37 individuals of five spe-
cies (including two C. virginica specimens). Figure 2 shows the
alignment of 1 68 sequences of the seven haplotypes detected
among all specimens in this study, along with those of C. gigas and
C. ariakensis from O'Foighil's study. Eight specimens of C. gigas
and 10 of C. plicatitla exhibited only one genotype, whereas seven
C. ariakensis and 10 C. lalienwhanensis individuals had two hap-
lotypes each. The two haplotypes of C. taliemvhanensis came from
different sampling locations. Including the outgroup. 80 nticleotide
positions were variable in the 16S data set. Six insertion/deletion
sites were detected between C. virginica and all other haplotypes.
Similarly, the alignment of the seventeen COI haplotypes de-
tected in our study and those two of C. gigas and C. ariakensis
from O'Foighil's study are shown in Figure 3. The 17 haplotypes
in our study included one for C. gigas (gigas 1. 8 individuals),
seven for C. plicatula (plical. 2. 3. 6 and 7. one individual for
each; plica4. three individuals; plica5. two individuals), three for
C. ariakensis (ariakenl. 4 individuals; ariaken2. two individuals
and ariaken3. one individual), five for C. ralienwhanensis
(talienwl. 2 and 3. one individual each; talienw4. four individuals;
talienw5. three individuals), and one for C. virginica (virgl. two
individuals). Including the outgroup. 170 positions are variable.
No insertions/deletions were detected for this protein-coding gene
fragment.
Pair-wise genetic distances of 16S sequences among all nine
haplotypes and those of COI sequences among all 19 haplotypes
were computed, then the mean genetic distances were obtained
(Table 1 ). In the 16S sequence, the genetic divergence between C.
gigas and C. talienwhanensis was low. 0.81%, and so was that
between C. ariakensis and C. plicatula. 0.13%. The sequence di-
vergences between C. gigas or C. ralientvhanensis and C. aria-
kensis or C. plicatula were higher, ranging from approx. 1.74 to
Taxonomic Status of Cr.assostrea Oysters
33
gigasl
talienwl
gigasO
talienw2
plical
ariakenl
ariaken2
ariakenO
virgl
gigasl
talienwl
gigasO
talienw2
plical
ariakenl
ariaken2
ariakenO
virgl
gigasl
talienwl
gigasO
talienw2
plical
ariakenl
ariaken2
ariakenO
virgl
gigasl
talienwl
gigasO
talienw2
plical
ariakenl
ariaken2
ariakenO
virgl
gigasl
talienwl
gigasO
talienw2
plical
ariakenl
ariaken2
ariakenO
virgl
gigasl
talienwl
gigasO
talienw2
plical
ariakenl
ariaken2
ariakenO
virgl
80
GCAATACCTG CCCAGTGCGA AATATTACTG TAAACGGCCG CCCTAGCGTG AGGGTGCTAA GGTAGCGAAA TTCCTTGCCT
C . ATAAGTC . . C T .
160
TTTGATTGTG GGCCTGCATG AATGGTTTAA CGAGGGTTTG ACTGTCTCTA AATTTTTTAT TGAAATTGTA CTGAAGGTGA
.A . .
.A . .
.A . .
.A . .
.A . .
.T G.
240
AGATACCTTC ATTTAAAAGT TAGACAAAAA GACCCCGTGC AACTTTGAAA A--TTAACTT TATTCAGGAG TAAAAGATTT
. .A. .
. .A. .
. .A. .
. .A. .
. AAG.
.GC.A.G. .G A.
320
TTAGGTGGGG CGCCTAGAAA GCAAG-TCTA ACCTTT-CTG AATAACT--A ACTCTTTCCG GATTTGACCC GATTATATTC
. -C.
. -C.
. AA . T C . GT .
. C.TT.--.
. .T. .ATA.
GT
.T. . .AA.TA
400
GATCATAGGA GAAGTTACGC CGGGGATAAC AGGCTAATCC TTTAGTAGAG TTCGTATTGG CTAAAGGGAT TGGCACCTCG
443
ATGTTGAATC AGGGATAATA GCTTCAAGGC GTAGAGGCTT TGA (8)
(5)
(7)
(5)
(10)
(5)
(2)
(5)
(2)
Figure 2. Mignnu'rit of seven oyster haplotypes of a 443-bp fragment of the mitochondrial I6S rDNA obtained in this study (C virgiiiica as
oulgroup) «ith published sequences for (. gigas and ('. ariakensis (O'Foighil et al. 1995, 1998). gigasO and ariakenO designate the sequences
of C. gigas and C. ariakensis from O'Foighil's study, respectively. Haplotype names are abbreviated as: gigas for C gigas. talienw for ('.
talienwhaiiensis. plica for ('. plicaliilu. ariaken for C. ariakensis. and virg for C. rirginica. Additional haplotypes per species are numbered
consecutively. Dots indicate nucleotide identity to the first sequence presented, gigasl. Dashes indicate inferred nucleotide indels relative to C.
rirginica. The number of individuals observed for each haplotype is indicated in parentheses at the end of sequence.
2.45%. The same pattern appeared in the COI data set: the coire-
sponding numbers were 1.08% between C. gigas and C. talien-
whanensis. 0.59% between C. ariakensis and C. plicaluUi. and
approx. 10.72 to 11.43% for the same comparisons mentioned
above. It is worth noting that the COI sequence was more variable
than the 16S sequence.
Consensus phylogenelic trees based on a parsimony analysis of
the 16S and COI fragments sequenced are presented in Figures 4
and 5. respectnely. Two groups (clades) in the 16S tree were
clearly distinguishable: C. ariakensis and C. pliiatida vs. C. gigas
and C. talienwhancnsis. whereas three groups (clades) were ap-
parent in the COI tree: (1) C ariakensis and C. plicatula: (2) C.
gigas and C. laliemvhanensis: (3) C. ariakensis from O'Eoighii's
study.
DISCUSSION
Oysters are among the most extensively studied and morpho-
logically variable marine invertebrates. However, our knowledge
of oyster phylogeny and systematics is still limited. There had been
over one hundred recorded species of oysters until 1970s, but two
thirds of them could be synonymous w ith each other according to
34 YU ET AL.
80
gigasl GCTGTTCTTG CGGGAACTAG GTTTAGGTCT CTTATTCGTT GGAGACTTTA TAACCCTGGA GCTAAGTTTT TAGACCCCGT
talienwl
gigasO
talienw2
talj.enw3
talienw4
talienw5
plical
ariakenl
plica2
plica3
plica4
ariaken2
plica5
ariakenS
plica6
plical
ariaKenO
virgl
gigasl
talienwl
gigasO
talienw2
talienw3
talienw4 C. .G.
G
r
G
r
-G
A
G
.A. .
r
r
G
. .C.
.G
.A.
G
.A. .
r
r
G
. .C.
.G
A
G
.A. .
r
r
G
. .c.
T
p^
.G
A
G
.A. .
r
r
G
. .c.
T
ft
.G
A
G
.A. .
r
r
G
. .c.
T
A
.G
C. . . .
A
G
.A. .
r
T
r
G
. .c.
.G
A
G
.A. .
r
r
G
. .c.
.G
.A.
G
.A. .
r
T
c
G
c
T
ft
.G
A
G
.A. .
r
T
r
G
. .c.
.G
A
G
.A. .
r
T
r
G
. .c.
. .c
TA.
.T.
GCA
. . .C. .A
.A.
TT
_G. .
r
A
G
c
T
. .A. . .T.A.
GACTTATAAT
.G. .C. .
GTTGTAA
T
CTAGGCATGC GTTGGTTATG
.A. .T. .A. .
ATTTTTTTCT
. .CT G
TTGTTATACC
A
TGTAATAATT
G. .T. .
160
GGGGGGTTTG
C
talienwS C. .G.
plical A G. . A. .G.
ariakenl A G. . A. .G.
plica2 A G.. A..G.
plica3 A G.. A..G.
plica4 A G. . A. . . .
ariaken2 A G. . A..G.
plicaS A G. . A. .G.
.A,
A.
. .G. .
A.
.G. . .
.A.
A.
. .G. .
A.
.G. . .
A. .C. . .
. .C
.C.
.A.
A.
.A.
A
C.
,T.
G.
. .C
TGTG. . .
. .c
. .T. .
,G. .
.G.
.T.
, .C. .
.A.
.A. .
.C. .
. .G. .
A.
.T. .
240
GTAACTGGCT
TATCCCTTTG
ATGCTTCTAG
TAGCAGACAT
GCAATTTCCT
CGATTAAATG
CATTTAGATT
TTGAGTTTTG
A ' '. . .
_ A
. . .T.
A
.A. .
. .T.
r
. . GC . .
, . C
c
. . .T.
A
.A. .
. .T.
A
r
. . GC . .
. .C .
c
.T.
A
.A. .
. .T.
B
r
. .GC. .
. C
c
. . .T.
.A
.A. .
. .T.
A.
. .GC. .
. .c.
.c. . .
. . .T.
A
.A. .
T
A
r
. .GC. .
. .c.
.c. . .
. . .T.
A
.A. .
T
A
r
. . GC. .
. .c.
.c. . .
. . .T.
A
.A. .
T
A
r
. . GC. .
. .c.
.c. . .
. . .T.
A
.A. .
T
A
r
. . GC . .
. .c.
.c. . .
. . .T.
.A
.A. .
A.
.c.
. .GC. .
. .c.
.c. . .
. . .T.
A
.A. .
. .T.
A
r
. . GC . .
. .c.
c
.A. .T. .
.A.
.A
.A. .
. .T.
.G.
. G. . .
C.
.G.
T. .
. . .T.
.GC
.T
. .A,
GA. .
.G.
.G.
.0.
.T.
C.
.A. . .
ariaken3 A G. . A. .G.
plica6
plica7
ariakenO
virgl
gigasl
talienwl
gigasO
taiienwZ
taJ ienw3
talienwj
talienwS
plical
ariakenl
plica2
plica3
plica4
ariaken2
plicaS
ariaken3
plica6
plica7
ariakenO
viryl
320
gigasl CCAGGGTCTC TTT.ATCTTAT GCTTATGTCT AACATTGTAG AAAACGGAGT TGGGGCAGGG TGAACAATTT ACCCTCCTTT
talienwl
gigasO
talienw2
talienw3
talienw4 C G
talienwS C
plical A.. A T G.
ariakenl A. .A T G.
plica2 A. .A T G.
plica3 A.. A T G.
plica4 A. .A T G.
ariaken2 A. .A T G.
plicaS A. .A T G.
ari3ken3 A.. A T G.
plicae A. .A T G.
plica7 A T G.
ariakenO C..A TC GT..G.. C A
virgl AT .GCTG..A.. AT . G A . . T . . . . CT . . G . GA T....A C GC .
Figure 3. Alignment of 17 oyster haplot.vpe.s of a 579-bp fragment of the mtCOI gene obtained in this study (C virginica as outgroup) with
published sequences for C. gigas and C. ariakensis (O'Foighil et al. 1995. 1998). gigasO and ariakenO designate the sequences of C. gigas and
C. ariakensis from O'Foighil's study, respectively. Haplotype names are abbreviated as: gigas for C. gigas. talienvv for C. talienwhaneiisis, plica
for C. plicalula, ariaken for C. ariakensis and virg for C. virginica. Additional haplotypes per species are numbered consecutively. Dots indicate
nucleotide identity to the first sequence presented, gigasl. The number of individuals observed for each haplotype is indicated in parentheses at
the end of sequence.
GG
C
GG
_ _ . . . r. _
GG
GG
. . .c
GG
C
GG ....
r
GG
GG
GG
GG
Taxonomic Status of Crassustrea Oysters 35
400
gigasl ATCAACTTAC TCTTATCATG GAGTTTGTAT AGACCTTGCA ATTCTAAGCC TTCACCTTGC TGGTATTAGC TCTATTTTCA
talienwl
gigasO
talienw2
talienw3
talienw4 C T C.
talienwS C T C.
plical G..G C. G T TT .A A.. ..
ariakenl G..G C. G T....TT .A A
plica2 G..G C. G T....TT .A A
plica3 G..G C. G T....TT .A A
plica4 G..G C. G T....TT .A A
ariaken2 G..G C. G T....TT .A A
pllcaS G..G C. G T....TT .A A
ariakenB G..G C. G G T....TT .A A
plicae G..G C. G T....TT .A A
plica7 G..G C. G T....TT .A A
ariakenO G..C..C TT.A A.
virgl G TT C C.. G..TT....C . . . T . . . . GT .A...T.A.. A....
480
gigasl GGTCAATTAA TTTCATAGTA ACGATTAGAA ATATGCGATC TGTTGGGGGC CATTTACTAG CACTATTCCC TTGATCTATT
talienwl
gigasO
talienw2
talienw3 G
talienw4 T T.. C
talienwS T.
plical T A.
ariakenl T A A T.G. .G..G..T.. C
plica2 T A A T.G. .G..G..T.. C
plicaB T A A T.G. .G..G..T.. C
pllca4 T A A T.G. .G..G..T.. C
ariaken2 T A A T T.G. .G..G..T.. C
plicaS T A A T.G. .G..G..T.. C
ariakenS T A A T.G. .G..G..T.. C
plicae T A A T.G. .G..G..T.. C
plica7 T A A T.G. .G..G..T.. C
ariakenO T C..T..G GT.G. .G T.. A..G C
virgl ....T T C C T ..CA..T T G..A...
560
gigasl AAGGTTACTT CATTCTTGCT TTTGACTACT CTCCCAGTGT TAGCTGGAGG TCTTACTATA CTTTTGACTG ATCGTCATTT
talienwl
gigasO
talienw2
talienwS
talienw4 G
talienwS G
plical TC.A A T G.. C G C
ariakenl TC.A A T G. . C G C
plica2 TC.A A T G.. C G C
plicaS TC.A A T G. . C G C
plica4 TC.A A T G. . C G C
ariaken2 TC.A A T G.. C G
plicaS TC.A A T G. . C G
ariakenB TC.A A T G.. C G
plicae TC.A A T G.. C G
plica7 TC.A A T G.. C G
ariakenO ..A..C..A T..A A..A..C ..T..G..AC C..G C..G
virgl ..A..G..A C... GC.T..C..G ..A..T..TC C. G G . . CC . T A
579
gigasl TAATACCTCT TTTTTTGAC (8)
talienwl ( 1 )
gigasO (20)
talienw2 ( 1 )
talienw3 (1)
talienw4 C . . . ( 4 )
talienwS C. . . (3)
plical . . .C. .G T (1)
ariakenl ...C..G T (4)
plica2 . . .0. .G T (1)
plicaS . . .C. .G T (1)
plica4 ...C..G T (3)
ariaken2 ...C..G T (2)
plicaS ...C..G T (2)
ariaken3 ...C..G T (1)
plicae . . .C. .G T (1)
plica7 . . .C. .G T (1)
ariakenO ...C..G T (5)
virgl A. .G (2)
Figure 3. (Continued)
Harry ( 1985 ). The inability to clearly classify closely-related oys- proven to be a powerful tool for oyster identification and discrinii-
ters has created problems for classification and species identifica- nation between closely related species or between nati\e and non-
tion worldwide. native species. Banks et al. (1993) discriminated closely related
Although morphologic identification of oysters often turned out oyster species, C. gigas and C. sikamea. via mitochondrial I6S
to be unreliable or ambiguous. mtDNA sequence analysis has rRNA gene sequencing and PCR/RFLP analysis. O'Foighil et al.
36
YU ET AL.
TABLE I.
Pair-wise sequence divergence (mean genetic distances! according to Kiniura's two-parameter model iKimura 198(1) among the five species
based on 443-nucieotide 16S rDNA and 579-nucleotide COI sequences.
16S
COI
Species
1
2
3
4
5
ft
1
2
3
4
5
6
C. gigas
C. lalienwhanensis
0.0()81
0.0108
C. plicanda
0.0233
0.0174
0.1113
0.1072
C. ariakensis
0.024?
0.01 S5
0.0013
0.1143
0.1 100
0.0059
C. ariakensisO
0.0450
0.04S7
0.0444
0.0462
11.1619
U.1639
0.1652
0.1691
C. virginica
0.1636
0.1 60S
0.1654
0.1673
0.1937
0.2569
0.2573
0.2510
0.2513
0.2849
C. ariakensisO indicates S. ariakensis sequence from OToighil's studies (1995. 1998).
Pair-wise comparisons yielding low genetic distances estimates are showed in boldface.
(1995) succeeded in distinguishing C. viri>iiuco from two closely
related oysters. C. gigas and C. ariakensis. and C. gigas from C.
ariakensis by employing sequencing and PCR/RFLP analysis of
pan of a fragment (443 bp) of the 16S rRNA gene. Sequence data
revealed that C. gigas and C. ariakensis showed higher levels of
similarity to each other (95%) than to C. virginica (84-86%).
Comparison of a 579-nucleotide fragment of the COI between the
Portuguese oyster. C. angiilala. and several Japanese oysters were
made by OToighil et al. (1998). showing that Portuguese oyster
haplotypes clustered firmly within a clade of Asian congeners and
were closely related to C. gigas (but not identical). This result
supports an Asian origin for the Portuguese oyster.
Reportedly, there are over 20 recorded species of oysters oc-
curring along the coast in China (Zhang et al. 1956, Qi 1989). and
for some of them classification and identification have been prob-
lematic or uncertain. Based upon extensive anatoinic studies of
almost all oyster species in China. Li and Qi ( 1994) concluded that
there were 15 species of oysters, and claimed that identification of
a few oyster species was clarified. Most of the species are rare and
found in South China Sea. However. Even for the four common
species (the Zhe oyster. Pacific oyster. Suminoe oyster, and
Dalianwan oyster), it is often not empirically easy even for marine
zoologists sometimes, to distinguish them clearly. This has caused
inconveniences and difficulties in broodstock management and
aquaculture practices. If the Dalianwan oyster is a discrete species.
ariakenO
— ariakenl
— ariaken2
— plical
talienw2
talienwl
gigasi
— gigasO
separate stock conservation and management should be applied.
Accordingly, clarification of the Zhe oyster's status would also
help oyster aquaculture practices. These are widespread concerns
for the oyster fishery along the coast of China
The molecular data provide some clarification on the species
status and phylogenetic relationships of these four species. For
Dalianwan and Pacific oysters, the 16S data show close similarity
between the samples of these two species, and the haplotypes of
Dalianwan and Pacific oyster formed a clear clade in the phylo-
genetic tree. This relationship is strongly supported by the COI
data set. in which all five haplotypes of the Dalianwan oyster and
the only haplotype of the Pacific oyster clustered closely. This is
also supported by the evident similarity in moi-phology between
these two species. The Dalianwan oyster samples were collected
from t> pical distribution areas, identified carefully according to the
plica5
— virgl
Figure 4. A consensus phylogenetic tree based on parsimony analysis
of 443-nucleotide mt I6S rDNA fragment according to Kimura's
model with C. virginica as an outgroup.
talienw4
talienwS
■ virgl
Figure 5. .\ consensus phylogenetic tree based on parsimony analysis
of 579-nucleotide mt COI gene fragment according to Kimura's model
with ('. virginica as an outgroup.
Taxonomic Status of Chassostrea Ovstbrs
37
descriptions of Zhang et al. ( 1956) and Qi ( 1989). Although there
are some morphologic differences compared with the Pacific oys-
ter, Dalianwan oysters share some morphologic characteristics
with Pacific oysters as described by Zhang et al. ( 1956) and Qi
(1989). A similar situation exists in scallops Pecten imiximus and
P. jacoheiis, where they share highly similar morphologic features
but have a surprisingly close genetic distance based on 16S se-
quences (Canapa et al. 2000). Our molecular data suggest that
Dalianwan and Pacific oysters belong to the same species, which
supports Li and Qi"s (1994) conclusion based on anatomy studies.
Results for the Zhe and Suminoe oysters are rather surprising.
The divergence between the two is much less than expected. The
genetic distances between them are as low as 0.1 39i- (for 16S) and
0.59% (for COl), even lower than that between the Dalianwan and
Pacific oysters (0.81 and 1.08'^H. They share a high degree of
similarity in these two gene fragments. In contrast, they showed
higher divergence from the Pacific and Dalianwan oysters in both
the 16S and the COl sequence data, though more strongly in the
latter. Also, haplotypes of the Zhe and Suminoe oysters clustered
in a single clade in both trees. This result is different from that
generally concluded from morphologic data. Morphologically, the
Zhe and Suminoe oysters are easy to distinguish in most cases.
Therefore, caution should be taken for the concern of status of
these two species. A possible explanation could be as follow, the
"Zhe oysters" we sampled could actually be a morph of Suminoe
oysters living in high salinity habitats. Because ecologically the
Suminoe oyster has a wide distribution and can tolerate a wide
range of salinities, morphologies could vary in different habitats.
Samples collected from the habitats other than an estuary may look
different from the Suminoe oysters from a typical habitat. It is
possible that Suminoe oysters from high salinity area and on rocky
shores are mistakenly classified as Zhe oysters because of mor-
phologic plasticity. It has been shown that the Zhe-like small oys-
ters found in the rocky intertidal zones of northern coast, once
removed to more productive waters, could grow to a bigger size,
which resemble the Suminoe oysters from an estuary habitat (R.
Wang, personal comm.). To confirm either of these possibilities, a
more extensive sampling and sequence analysis throughout their
natural range are needed.
An interesting finding from this study is that O'Foighil's COl
sequence of the Suminoe oyster showed a significant divergence
not only from that of the Dalianwan-Pacific oysters, but also of the
Suminoe-Zhe oysters. The divergence may be due to the fact that
mt protein-coding genes like COl are usually more variable than
iDNA (Hixson & Brown 1986) and the fact that OToichil's Sumi-
noe oyster samples, which came from a hatchery stock originated
from Japan, may represent a different population that is genetically
isolated from the Chinese population (our samples). However,
analysis of more specimens from Japan or other parts of their
natural range is needed for confirmation.
Li and Qi (1994) suggested that the Zhe-like oysters most com-
monly found in the rocky intertidal zone were Pacific oysters
instead of Zhe oysters as most people assumed. If so, the iiitDNA
sequences of these (Zhe) oysters should have higher similarity to
(or low divergence with) those of the Pacific oysters or Dalianwan
oysters we presented here and that of O'Foighifs. Actually this is
not the case. Our sequence data show that these smaller oysters
from rocky shores could be Suminoe oysters, rather than Pacific
oysters.
Additionally, in this study the COl sequences showed more
variations, as expected, than the 16S sequences. For instance, in
the 16S data, we detected only one haplotype for Zhe oyster, two
tor each of the Dalianwan and Suminoe oysters; but in COl data,
the numbers of haplotype are seven, five and three for these three
species, respectively. Also, the divergence between C. gigas and
C. plicatida or C. ariakensis is three times as high as that between
C. gigas and C. taUt'imhanensis in the 16S data, whereas the
divergence is eleven times higher in the COl data. The COl se-
quence is more sensitive in discriminating closely related species,
supporting the observation by Boudry et al. ( 1998) where no vari-
ability was detected with nine endonucleases among 253 individu-
als of C. gigas and C. angiilata with 16S rDNA, but reasonable
polymorphism was detected with four enzymes with COL Other
works have also proved that CO! sequence is a good choice for
similar purposes (Meyran et al. 1997. O'Foighil et al. 1998).
In summary, the mtDNA sequence data strongly suggest that C.
laliemvlianensis is not a discrete species and should be considered
as synonymous with C. gigas. Our data also indicate that the "Zhe
oyster" is different from the Pacific and Dalianwan oysters, but is
genetically very close to Suminoe oyster, at least for the ones we
sampled.
ACKNOWLEDGMENTS
This work was financially supported by National Science Foun-
dation of China (39600113) and Research Foundation (2001) of
Institute of Oceanology, Chinese Academy of Sciences, Qingdao
266071, P. R. China. Yu and Guo are partly supported by grants
from US Sea Grant and New Jersey Commission on Science and
Technology.
LITERATURE CITED
Banks. M. A., D. Hedgecock & C. Waters. 1993. Discrimination between
closely related Pacific oyster spp. iCrassoslreu) via mitochondrial
DNA sequences coding for large suhunit rRNA. Mol. Mar. Binl. Bio-
techiwl. 2:129-136.
Boudry, P.. S. Heurtebise, B. Collet, F. Comette & A. Gerard. 1998.
Differentiation between population of the Portuguese oyster, Crasso-
strea angiilata (Lamark) and the Pacific oyster, Crassostrea gigas
(Thunberg), revealed by mtDN.^ RFLP analysis. J. Exp. Mar. Biol.
Ecol. 226:279-291.
Canapa. .A., M. Barucca, A. Marinelli & E. Olmo. 2000. Molecular data
from the 16S rDNA gene for the phylogeny of Peclinldae (Mollusca:
Bivalvia). ./. Mol. Evol. 50:93-97.
Folmer. O.. M. Black, W. Hoeh. R. Lutz & R. Vnjenhoek. 1994. DNA
primers for amplification of mitochondrial cytochrome c oxidase sub-
unit I from diverse metazoan invertebrate. Mol. Mar. Biol. Biotechnol.
3:294-299.
Felsenstein, J. 1989. PHYLIP-Phylogeny Inference Package (Version 3.2).
Cladi.mcs. ."S: 164-166.
Guo. X.. S. Ford & F. Zhang. 1999. Molluscan aquaculture in China. J.
Shellfish Res 18:19-31.
Harry, H. W. 1985. Synopsis of the supraspecific classification of living
oysters. Vc/i^ijcr 23:121-158.
Hixson, J. E. & W. M. Brown. 1986. A comparison of the small ribosomal
RNA genes from the mitochondrial DNA of the great apes and humans:
sequence, structure, evolution and phylogenetic implications. M. Biol.
Evol. 3:1-18.
Jozefowicz. C. J. & D. O'Foiahll. 1998. Phvlogenelic analysis of southern
38
YU ET AL.
hemisphere flat oysters based on partial mitochondrial 16S rDNA gene
sequences. Mol Phylogenet. Evot. 10:426 — 135.
Kimura. M. 1980. A simple method for estimating evolutionary rates of
base substitutions through comparative studies of nucleotide se-
quences. / Mol. Evol. 16;1 11-120.
Lapegue, S., I. Boutet, A. Leitao. S. Heurtebise, P. Garcia, C. Thiriot-
Quievreux & P. Boudry. 2002. Trans-Atlantic distribution of a man-
grove oyster species revealed by 16S mtDNA and karyological analy-
sis. Biol. Bull. 202:232-242.
Li, X. & Z. Qi. 1994. Studies on the comparative anatomy, systematic
classification and evolution of Chinese oysters. Studia. Marina Sinica.
35:143-173. (In Chine.se).
Liu, B. & J. Dai. 1998. Studies on genetic diversity of oysters of genus
Crassostrea. J. Fish. China 22:193-198. (In Chinese).
Littlewood, D. T. J. 1994. Molecular phylogenetics of cupped oysters
based on partial 28S rRNA gene sequences. Mol. Phylogenetic. Evol.
3:221-229.
Meyran, J. C, M. Monnerot & P. Taberlet. 1997. Ta,\onomic status and
phylogenetic relationships of some species of the genus Gammarus
(Crustacea. Amphipodal deduced from mitochondrial DNA sequences.
Mol. Phylogenet. Evol. 8:1-10.
O'Foighil. D., P. M. Gaffney & T. J. Hilhish. 1995. Differences in mito-
chondrial 16S ribosomal gene sequences allow discrimination among
American (Crassostrea virginica) and Asian (C gigas. C. ariakensis)
oyster species. J. E.\p. Mar. Biol. Ecol. 192:211-220.
O'Foighil D., P. M. Gaffney & W. A. Wilbur. 1998. Mitochondrial cyto-
chrome oxidase I gene sequence support an Asian origin for the Por-
tuguese oyster Crassostrea angiilara. Mar. Biol. 131:497-503.
Palumbi. S. R., A. Martin & S. Romano. 1991. The simple fool's guide to
PCR. Honolulu, Hawaii: University of Hawaii Press.
Qi, Z. 1989. Mollusk of Yellow Sea and Bohai Sea. Beijing: Agricultural
Publishing House, pp. 176-180. (In Chinese).
Thompson, J. D., D. G. Higgins & T. J. Gibson. 1994. CLUSTAL W:
improving the sensitivity of progressive multiple sequence alignment
through .sequence weighting, position-specific gap penalties and weight
matrix choice. Nucleic Acids Res. 22:4672^680.
Torigoe. K. 1981. Oysters in Japan. J. Sci. of Hiroshima Univ.. Ser. B. Div.
1 (Zool). 29:291^19.
Wang, R., Z. Wang & J. Zhang. 1993. Marine molluscan culture. Qingdao,
China: Qingdao Ocean University Press. (In Chinese).
Yang. R. & Z. Yu. 2000. AUozyme variation within oysters Crassostrea
plicatiila and C. gigas from Shandong coastal waters. / Fish. China
24:130-133. (In Chinese).
Zhang, X. & Z. Lou. 1956. Studies on oysters in China. Acta. Zool. 8:65-
94. (In Chinese).
Jniiniul ,>f Shclirish Rcsi-anh. Vol. 22, No. I. 34-19. 2()().V
INCREASED BIOMASS YIELD FROM DELAWARE BAY OYSTERS (CRASSOSTREA
VIRGINICA) BY ALTERNATION OF PLANTING SEASON
JOHN N. KRAEUTER,' SUSAN FORD,' AND WALTER CANZONIER"
^Haskin Shellfish Research Lahnniiory. Iiistiliite of Marine and Cixistal Sciences. Rutgers University,
6959 Miller Avenue. Port Norris. New Jersey US349: and 'Aquarius Associates. Manasijuan. New Jersey
ABSTRACT The practice of moving oysters from low-salinity to high-salinity areas for improving growth and meat quality has been
practiced for well over a century. In the Delaware Bay. the practice was abruptly changed when MSX [Haplosporidium nelsoiii) caused
large-scale oyster mortality in the higher salinity portions of the bay. Similar disruptions occurred in Chesapeake Bay and other areas.
In lime the Delaware Bay. the oyster industry learned how to operate around the disease, but in early 1990s. Dermo (Pert^innis mariims)
began to cause serious mortality on transplanted oysters. Despite the historic and continuing movement of oysters within and between
estuaries, there is little published scientific literature indicating optimum conditions for transplantation. We investigated the effects of
transplantation from a low-salinity seed bed to a typical higher salinity leased ground. The transplants were designed to evaluate an
early, the traditional spring, and two fall transplant dates on the subsequent disease levels, growth, and survival of the oysters in three
size classes: market, submarket. and small. Environmental and oyster disease data suggest we conducted the experiment under nearly
worse-case conditions, high Dermo. and low food (chlorophyll). There were no significant differences associated with the timing of
transplant. We did not record significant growth on any size oyster and disease caused mortality exceeded 50% for early transplants.
Smaller oysters experienced greater mortality than market size individuals. Despite these conditions, meat dry weight nearly doubled
within 1 to 2 mo after transplant in all but the March transplant. Under these di.sease and environmental conditions the only economic
gain would be from the doubling of the meat weight and associated better meat quality. No gain can be expected from submarket
oysters growing into the market size classes.
KEY WORDS: oyster. Cnisso.strea. Delaware Bay. season, disease, growth
INTRODUCTION
In the Delaware Bay oysters have been transplanted from upper
bay low-salinity seed producing areas to lower bay higher-salinity
growing beds for more than 150 years (Ford 1997; Fig. 1 ). Similar
transplantation strategies have been used by oyster growers in
Chesapeake Bay (Andrews & McHugh 1957) and New England
(Ingersoll 1881, Goode 1887). Further, to increase production and/
or to supplement local seed as resources became depleted, oysters
were imported from distant sources. Despite these historic and
continuing large scale movement of oysters within and between
systems, there is little scientific literature indicating the optimum
conditions for transplantation.
Hopkins and Menzel ( 1952) developed a framework for study-
ing the transplantation of oysters based on the biomass yield of the
product, and Andrews and McHugh (1957) used biomass yield
estimates from trays of oysters to evaluate the effectiveness of
transplantation strategies. Reliance on biomass as a means of as-
sessment in both of these studies was based on the assumption that
the majority of oysters were destined to be shucked, and thus meat
yield was the most important aspect of production. This may not be
the case for those oysters that are grown to be sold for the half-
shell trade. In this latter case, assuming adequate meat quality,
numbers at market size are more important than total bicmiass.
Haskin et al. ( 1983) and Hargis and Haven (1988) both indicate
that the oyster planting industry in the Delaware Bay and the
Virginia portion of Chesapeake Bay, respectively, operated under
the assumption that transplanting was profitable if one bushel of
seed oysters yielded one bushel of market oysters. In the late
1950s, the parasite MSX, Haplosporidium nelsoni. caused epi-
zootic mortalities in both estuaries and forced major changes in
oyster industry practices. In the Virginia portion of Chesapeake
Bay, growers abandoned higher salinity grounds and concentrated
efforts in areas that historically produced higher than the 1:1 yield
(Hargis & Haven 1988), Despite H. nclsoni-c-Msed losses, the
Delaware Bay oyster industry continued to transplant oysters
based on the system developed in the KSOOs. Oysters were left on
the planted grounds, where high salinity favored the H, nelsoni
parasite, but for no more than 1 y (Ford 1997), and yields contin-
ued to be about 1:1 (Haskin & Ford 1983). After the 1950s H.
nelsoni epizootic, the importation of seed from out of state into the
New Jersey portion of Delaware Bay was banned.
In 1990, an outbreak of Dermo disease caused by Perkinsiis
niarlniis prompted a further change in strategy by the Delaware
Bay oyster industry. After 1990. P. inarinns infected most of the
oysters in the seed bed areas (Ford 1997), and oysters planted in
the spring of 1991 suffered high mortality in the late summer. The
oyster industry and the State of New Jersey responded by devel-
oping a program to market oysters directly from the seed beds.
This strategy produced oysters that had poorer meat quality and a
lower value than those from higher salinity waters.
At the same time, it was realized that although Powell et al.
( 1997) modeled the effect of transplant time, disease, and preda-
tion on market oyster populations, there were no real data on which
to base transplantation decisions in the presence of this new para-
site. The model predicted that fall (November) transplants left for
1 y yielded the best survival of market oysters compared with
transplants in January, March, or May that were harvested in No-
vember. In all cases the number of market oysters declined from
July to November. The model did not include an August transplant
with immediate harvest that fall, a strategy that would minimize
disease-caused mortalities while still taking advantage of typically
good fall "fattening" conditions. The industry requested data on
the following: 1) the best time of the year to transplant oysters;
2) the survival of transplanted oysters at various times after trans-
plant; 3) the numbers of market oysters expected from the net
result of growth and mortality; and 4) the gains that could be made
in iTieat quality and the length of time after transplant this gain
might take.
The industry, through a nonprofit foundation, collaborated with
39
40
Kraeuter et al.
NEW JERSEY
DELAWARE
CAPE HENLOPEN
Figure 1. Delaware Bay showing locations of the seed beds and Shell
Rock bed, leased grounds, and the ground used for transplant
studies.
state New Jersey Department of En\ironmental Protection
(NJDEP) and Haskin Shellfish Research Laboratory (HSRLl per-
sonnel to conduct an initial test of alternative planting dates. This
study (Canzonier 1998) moved oysters from the Shell Rock seed
bed to higher salinity grounds (527 D) in December. February.
May, and August. The effort clearly established that transplanting
in months different from the historical spring period was economi-
cally feasible, but cautioned that a single year's result could not
provide sufficient background for assessing year-to-year variation.
In addition, all months but the traditional spring transplant period,
represented by the May transplant, gave nearly identical results.
The May transplant had significantly less market oysters produced
than the other months (Canzonier 1998).
The information at the onset of the current study suggested that
transplantation strategy would depend on several factors: oyster
population size frequency distribution, source stock disease level,
seed bed used as a source, environment of the planted ground,
disease pressure, and harvest timing. In addition to biological vari-
ables, market factors, and industry seasonal work cycles affect the
economic impact of alternative planting seasons. The present study
builds upon earlier efforts and evaluates the effects of varying the
timing of transplanting oysters from one seed bed to a lower bay
planting ground.
MATERIALS AND METHODS
Experimental Design
Oysters from Shell Rock Bed were transplanted to ground 354
D (Fig. 1 ) in March, May, September, and October of 1999. Shell
Rock was selected because it represented a central seed bed source,
had a significant number nearly market size oysters, and pro\ ided
the oysters for the Can/onier ( 1998) study.
The transplant ground was subdivided into experimental plots.
each marked with navigation coordinates. A preliminary sampling
indicated that only a small number of large residual oysters (mean
99 mm) were present (mean 2.4 oysters bu~' from 8 one-bushel
samples). Approximately 1800 US Standard bushels (36.4 L;
herein after referred to as bushels or abbreviated as bu.) of oysters
were planted on each 24.4 x 91.4 m plot each transplant time
(3.200 bu.acre"' or 90.000 oysters hectare"').
At each transplant time, triplicate bushels of oysters were re-
moved from the deck load of the boat and analyzed in a manner
similar to the techniques used for the subsequent monthly samples
(see below). In addition, oysters were processed for disease diag-
nosis.
After planting, at least three dredge samples were collected
each month from each planting. All material was placed in the
bushels so that triplicate composite bushel samples of material
were examined from each planting each inonth. These were ex-
amined in the same manner as the source oysters, but with special
attention to growth, meat condition. P. inarintis level, and mortal-
ity (apportioned by oyster size). In the latter months, additional
oysters were set aside after the samples had been collected to be
sure enough material was available in all size classes to process P.
marinus and condition index samples. H. nelsoni levels were not
detemiined on the monthly samples, but were evaluated on the
fmal samples from each plot in No\'ember, as well as on the initial
transplants.
Sample Processing
All live oysters >20 mm, old, new boxes, and gapers in the
entire sample were counted. All oysters >20 mm were measured
and divided into market (>76 mm) and submarket (35-73 mm) and
small (<55 mm) classes. All parameters were normalized to a
standard bushel for comparison with other samples. Mortality was
estimated by calculating the percentage of new boxes and gapers in
each sample. This was considered recent mortality. Recent mor-
talities were accumulated to provide an estimated cumulative mor-
tality at the end of the study (Ford & Haskin, 1982).
Twenty oysters (six or seven from each of the 3 bu.) of each
size class were set aside for evaluation of condition index and an
additional group of similar size was examined for P. inariiuis
infection. Condition index was derived from the ratio of meat dried
at 50°C, and greatest shell dimension (height). P. marinus was
diagnosed after incubation of the rectum and a piece of mantle in
Ray's fluid thioglycollate medium. Infection intensity was scored
from to 5 (Ray 1954) and a weighted prevalence calculated as the
mean intensity, including zeros, of all oysters in a sample. Oysters
in the initial planting and final sampling were diagnosed for H.
nelsoni by tissue section histology. Infection intensities were rated
from to 4 (Ford 1983) and a weighted prevalence calculated as
for P. marinus.
Individual Oyster Growth and Mortality Study
To evaluate production requires size-class-specific growth and
mortality data. This was approximated from the bushel samples,
but a second method was utilized to provide a more precise evalu-
ation of individual oysters. A group of experimental oysters rep-
resentative of the source bed was deployed at the time of trans-
plant. This group consisted of five replicates of 20 oysters from
each of three size classes (63.5 to 69.9 mm, 70 to 75.9 mm, and
>76 mm) for a total of 300 oysters. Fishing leader tethers were
glued to the top valve of each oyster with Marine Tex. The tethers
Increased Biomass Yield ok Oysters
41
were then attached with cable ties along the side of a square
reint'orcing rod frame square (~l m on each side) that was held
approximately 5 cm above the bottom by a centrally located ce-
ment anchor. The entire array was attached to a surface lloat. Each
individually identified oyster was measured (height) and the array
deployed so that the oysters would lie on the bottom. Each month
each oyster was measured and mortality or loss noted. In this
instance, mortality was calculated directly because the history of
each oyster was known.
Environmental Data
The following environiuenlal data were collected on bottom
water on at least an every other week basis: temperature, salinity,
dissolved oxygen, pH, total suspended solids. Chlorophyll a. and
suspended organic material. In addition, temperature was moni-
tored continuously with an electronic recorder. Salinity was ob-
tained with a refractometer. All grab sample temperature and dis-
solved oxygen data were measured with a YSI oxygen meter, and
pH data were obtained with an electronic pH meter. Suspended
solids, chlorophyll and particulate nitrogen samples were obtained
from at least 500 ml of water filtered through Whatman GF/C glass
fiber filters, which were stored on ice until they were returned to
the laboratory. Chlorophyll samples were immediately placed in
buffered acetone and refrigerated. Particulate samples were dried
at 50°C. All en\ ironmental data were analyzed according to Strick-
land and Parsons ( 1968).
Data Analysis
Size frequency data were normalized by adjusting the base live
and recent dead (gapers and new boxes) frequency distributions
from all individuals collected in the three bushel samples (in 5-mm
increments) to 100 individuals. These frequencies were then ad-
justed to the number of live or dead bu."' by multiplying the
frequency of occurrence in all sizes by the average number of live
or dead bu."' Data were summarized and significant tests were run
using one-way analysis of variance, I tests, or other descriptive
techniques. Percentages were transformed using an arc-sine trans-
formadon before performing analysis.
RESULTS
Envirnnnicntal linla
Temperature on the transplant ground was 3.5°C in March, at
the beginning of the study, and peaked in August at 27.5"C. Sa-
linity was generally between 21 and 23 ppt.. with a low of 19 ppt
in April and a high of 26 ppt in October and December. pH
remained relatively stable, ranging from 7.8 to 8.6 with the excep-
tion of a low value of 6.9 on September I. Dissolved oxygen
ranged from a high of 13.5 mg L"' in March to a low of 5.6 mg L^'
on July 14. In general dissolved oxygen levels remained near or
above saturation at temperatures below 2()"C and near or slightly
below saturation above those temperatures. Total suspended solids
were typically between 30 and 55 mg L"'. with highest and lowest
values of 86 and 1 8 mg L~ ' on August 1 8 and May 5, respectively.
Chlorophyll a showed a typical spring (late March to early April)
bloom followed by generally lower vales in summer (Fig. 2).
There was an increase in Chlorophyll a in fall (October to early
November). Highest Chlorophyll a levels were found March 25,
April I, May 18 and November 5 with values of 54, 46, 38 and 39
mg m~' respectively.
60
50
^40
30
D-
O
U
20
10
Mar 9 Apr 1 May 5 May 18 Jun 18 July 14 Aug 18 Sept 23 Oct 8 Oct 29 Dec 15
Mar 25 Apr 18 May 1 1 Jun 7 Jun. 30 Aug 4 Sept I Oct 5 Oct 22 Nov 5
1999
1996/1997
Figure 2. Buttuni water chlorophyll a In samples taken from bottom water over ground 554 1) in Delaware liay in 1999 compared with similar
data taken over ground 527 D In Delaware Bay in 1997. Data are in mg per m'. 1997 data from Canzonier (1998).
42
Kraeuter et al.
Oyster Data
Because the samples taken at the time of transplant represented
the source bed and culHng machinery on the boat, not the ground
to which the oysters were transplanted and monitored, time (7",,)
for subsequent analyses was the first sample after transplant. The
samples taken from the deck at the time of transplant were utilized
to estimate the size, condition and numbers of oysters transplanted.
Numbers of Live and Dead Oysters
The numbers of oysters being transplanted, based on the initial
samples for each transplant period, suggests that all groups, with
the exception of the October transplant, received uppro.ximately
the same number of individuals per unit volume of material
moved. The October samples had fewer oysters than those groups
transplanted in March and September, but was equivalent to the
May transplant (Table I ). It seems likely that more live oysters
were moved in the May transplant than in October, but the high
variance in May precludes making a definite statement.
The total numbers of live oysters significantly decreased from
Tf, to the fmal samples {T,) in November. The numbers in the
March and May transplants fell approximately 50* from 200 in
initial post-planting samples to <I00 bu.~' in the final November
sample (Table 1 ). The mean oysters bu.~' in October and Novem-
ber, traditional harvest months, were greatest for the September
transplants, but the difference was statistically significant only in
October. The decrease in oysters from planting to November was
least in the September transplants, but the time between 7",, and T,
was only one month. No calculation can be made for the October
planting because Tq = Tf (Table 1).
Live oyster numbers were also analyzed by size (Table 2). Data
from dredged samples show that numbers of marketable oysters
declined about 50% for March transplants, but that subsequent
transplants experienced little or no loss. Submarket and small oys-
ter numbers also declined, and. with the exception of the Septem-
ber transplant, these declines were usually greater than for market
size oysters and often more than 50Vc. Despite large losses of
oysters, there were no statistically significant differences in No-
vember in the number of market size, or submarket size oysters in
any transplant period. Numbers of small oysters in the March and
May transplants had declined appreciably by November and there
were about half as many small oysters per bushel as in the other
two size classes, even though small oysters were most abundant at
the time of transplant. Numbers of small oysters remained high in
the final sample of the September transplant, but not in the October
group.
Recent mortality, for all size classes, was greatest in the fall
(Fig. 3). These losses occurred across all size classes, but, with the
exception of the October transplant, losses were greatest in the
smallest size classes. Estimated cumulative mortality from trans-
planting to the final sample of all size oysters was 54, 55, 15, and
9% for March, May, Septeinber and October transplants, respec-
tively. Total losses of small oysters were greater than those of
market or submarket oysters for the March and May transplants
(Table }). There were no differences between the market and
submarket oyster losses in any transplant group.
Disease Levels
H. nc'l.sdiii was detected, in initial and final sainples. only at
very low levels. There was no association with size or transplant
time. The highest infection level (prevalence) was 30%, but most
infections averaged <15%. The highest weighted pievalence (0.4)
was found in the fall samples.
In contrast. P. marinus levels were high in all plantings and all
size classes (Fig. 4). Infections were nearly as heavy and abundant
on the source bed as they were in oysters already transplanted to
the higher salinity experimental site. Percent infection (prevalence)
for the March and May transplants exceeded 80% by July and was
usually 90 to 100% until it dropped below 80% in November. For
the later transplants, P. marinus levels usually increased to 90 to
100% within 1 mo after transplant. Weighted prevalence was rela-
tively high in the March transplants, but underwent a typical drop
in April/May (Bushek et al. 1994). The same drop occurred on the
source bed as the May transplants had a weighted prevalence simi-
lar that of the March transplant at the same time. Intensities in both
groups then increased over the summer until September, when
levels in all size categories decreased concurrent with an increase
in mortality (compare Fig. 3 to 4). Levels increased again in the
October sample and then dropped by nearly 50% in November. At
TABLE L
Mean numbers of live oysters >20 mm bu."' by month with 95% conndence limits (h = 3 for each monthly sample).
March
May
September
October
Mean
95% Conf
. Limits
Mean
95% Conf.
Limits
Mean
95% Conf. Limits
Mean
95% Conf. Limits
M
323
363
283
A
212
124
156
105
119
80
108
76
230
189
209
155
144
120
131
87
195
59
103
56
93
40
84
65
M
296
403
188
J
J
A
I8,S
121
76
90
92
79
239
169
154
143
104
113
137
73
38
79
45
S
307
355 259
169*
146
203 136
205 87
243
254 232
N
78
101 56
Bold numbers indicate a significant difference from the prior month. The area in gray indicates samples removed from the deck of the transplant vessel.
These were not used in subsequent calculations.
* Significantly more oysters than in other transplants during the sample period.
Increased Biomass Yield of Oysters
43
I ABIE 2.
Mean number of live market (>7f) ninil, subniarket (75-55 mm), and small (55-20 mm) oysters bu. ' of dredfjcd material from transplants
in March, May, September, and October 1999.
March 1999
May
1999
September 1999
October 1999
Market
Submark
Small
Market
Submark
Small
Market Submark
Small
Market Submark
Small
M
58
115
150
(^
63
48
61
24
40
30
32
76
32
42
37
41
30
38
29
75
45
54
40
38
21
38
15
M
78
104
114
.1
J
34
38
23
25
38
34
64
34
25
37
33
33
87
49
28
39
21
16
s
56
86
167
54 70
o
29
27
47
35
94
84
120
N
25 26
28
Oys(ers were transplanted from Shell Rock to Ground 554D on the Delaware Bay leased grounds. Areas of gray indicate samples from deck loads of
transplanted oysters. All other samples were dredged from transplant plots. Submark = submarket.
this linic. heavy (iiortality was observed in the March transplants to those transplanted earher. but, unhke the former, infections
only (Fig. 3) and the drop was probably the beginning of the retnained at very high levels in these oysteis into November. The
overwinter loss of infections (Bushek el al. 1994). Oysters trans- persistence of high infection levels was as.sociated with low mor-
planted in September and October had weighted prevalence siniilar tality in both fall groups.
Mar >75mm
Mar 55 to 74mm
Mar <55mm
I Li
.....III!
Mil
Apt
Mav June Iul> Aug
May >75mm
Scpl
Oci
Nov
*7 1
30-L--
1
1
1"
1
1
1
■
■ - ■■
"
Apt Miy June July Aug Sept Do Nov
Sept >75nun
Hi
Mm Apr May June July Aug Sep( Oci Nov
Oct >75mm
II
Apr May June July Aug Sep! Ocl Nov
Sept 55 to 74 mm
li
Apt M»y June July Aug Sept Oct Nov
Oct 55 to 74 mm
h
Mai Api Mdy June July Aug Scpi Oct Nov
May <55mm
Mar Apr May June July Aug Sepi Oct Nov
Sept <55 mm
■ I
Ms Apr Miy tunc luly Aug Sepi Oa Nav
Oct <::55 mm
Apr Miy June July Aug Sept CJti Nov
Mat Apt Miy June July Aug Sepi Oci Nov
Mat Apt Miy June July Aug Sept Oct Nov
Figure 3. Interval percent mortality by month of market (>75 mm), submarket (55 to 74 mm), and small {<55 mm) oysters transplanted from
Shell Rock to Delaware Bay ground 554 D in 1999. Transplant months were March {top graphs), May (middle top graphs), September (middle
bottom graphs), and October (bottom graphs).
44
Kraeuter et al.
TABLE 3.
Estimated cumulative percent mortality, from plantin}< to November 19')9. by size category of dredged oyster samples collected in Delaware
Bay by transplant month.
March 1999
May 1999
September 1999
October 1999
Market
Submark
Small
Market
Submark
Small
Market Submark
Small
Market
Submark
Small
46
48
65
45
4,S
XS
19 16
14
11
11
5
Market (>76 iiini). Mibniarket (75-55 mm), and small (55-20 mml.
Growth and Condition
With the exception of the March transplants, there were no
differences in the sizes of oysters in the subniarket and small
categories through time. Mean dry meat weight of market oysters
for the March and May transplants increased significantly in June,
after .^ and I mo, respectively (Table 4). That of markel-si/e Sep-
tember and October transplants rose in November after 2 and I mo.
respectively. There were no significant differences in meat weight
among any of the transplanted groups by November. While not
statistically significant, there was a consistent increase in meat
weight in all transplants of market-si/e oysters between October
and November. In general, meat weight increases of submarket
and small oysters mirrored those of the market-size individuals.
Reflecting the increase in meat weight without increased shell
size in market oysters, the condition index increased during the
study period. With the exception of the March transplants, oysters
required one month after transplant to the lower bay to improve
condition, and they typically retained this condition throughout the
summer and into the fall. While not statistically .significant, there
Mar > 75 mm
Mar 55-75 mm
March April May June July Aug Sept Oct Nov
irxll
March April May June July Aug SepI Oct Nov
Mar < 55 mm
4-
3-
2-
0-
Mafch April May June July Aug SepI Oct Nov
May > 75 mm
5 :
llili
mil
T r
March April May June July Aug Sept Ocl Nov
May 55-75 mm
III T .
illi
1
Oct
nil
i
JUU
1
March April
May June July Aug Sept
Nov
May < 55 mm
March April May June July Aug SepI Ocl Nov
Sept > 75 mm
March April May June July Aug Sept Ocl Nov
4—]
Sept 55-75 mm
T i
[
1 1 I 1 1 !
March April May June July Aug
1 1 1
^ept Oct Nov
Sept < 55 mm
i
i ™ i™i ™i
— I — i — I — I —
March April May June July Aug Sept Oct Nov
Oct > 75 mm
5„
1 1 1 1 1 1 r
March April May June July Aug Sept Oct
i
Oct 55-75 mm
1 1 1 1 1 1 r
March April May June July Aug Sepl Ocl
I
Oct < 55 mm
t
1 1 1 1 1 1 r
March April May June July Aug SepI Oct
i
Figure 4. Monthly weighted prevalence of Dermo (/'. nuirinu\) infections in market (>75 mini, subniarket (55 to 74 mini, and small (<55 mml
oysters transplanted from Shell Rock to Delaware Bay ground 554 D in 1999. Transplant groups were March (top graphs). May (middle top
graphs), September (middle bottom graphs) and October (bottom graphs). For each transplant group, the llrst sample represents that on Shell Rock
bed when the oysters were moved. All subsetpient samples represent infection levels on ground 554 D. Error bars represent 95 Cr confidence interval.
Increased Biomass Yield of Oysters
45
TABI.K 4.
Mean dry meal Heijjht (jjl of markel-size (ijsterv b> month with 'tS'c conlldence Mmit.s.
March
May
September
October
Mean
95 "/f Conf.
Limits
Mean
95% Conf.
Limits
Mean
95% Conf. Limits
Mean
95% Conf. Limits
M
1,1
1.2
0.9
A
1..^
1..5
1.2
M
1,?
1..S
1,2
l.s
1,7
1.3
.1
2.4
2.8
2.0
2..^
2.6
1.9
J
2.5
2.8
T 1
2.3
2.7
2.0
A
-> 1
2.4
2.0
1 T
2.4
2.0
S
2.4
2.7
2.1
2.5
2.8
T 1
1.6
1.8 1.4
O
2 2
2.6
1.8
2.3
2.7
2.0
1.9
2.3 1.6
I.I
1.3 1,0
N
2.9
.V4
T 1^
3.1
3.6
2.6
2..S
2.9 2.1
2.7
3.2 2.2
Bold niinihers indicate a significant difference from the previous month.
was a general trend for market oysters to improve in condition
from October to November.
Condition index for submarket and small oysters generally fol-
lowed the same trends as for the market oysters with no significant
change from June to Nmember. In general, there was a significant
increase in condition w ithin 1 mo after transplant for all submarket
and stnal! oysters with the exception of the March transplants and
small oysters transplanted in October.
By November, the meat condition index of all si/e classes in
the March and September transplants was statistically the same.
Among the October transplants, condition of submarket and mar-
ket oysters was the statistically similar, and greater than that of
small oysters, while the condition of market oysters in the May
transplants was greater than that of either submarket or small oys-
ters.
(iriiHlli ami Mdilalily of hulividually Marked Oysters
For calculations of mortality, the data from the tethered oysters
were corrected for oysters lost during the experiment by reducing
the numbers of oysters present from the initial counts. A few
oysters were lost because of detachment of the adhesive, but one
entire rack was lost.
Mortality of tethered oysters mirrored that of oysters trans-
planted at similar times, with a few notable exceptions (Fig. 3). It
IS evident from the cumulative mortality data (Table 5) that the
tethered oysters (particularly those put out in May and September)
liad substantially more mortality than that estimated from exami-
nation of boxes and gapers in dredged samples. At times, shells on
one section of an airay were observed to have become blackened.
This suggests that some silting had taken place around these oys-
ters and may have elevated the mortality above that experienced by
the planted oysters, but we have no independent measure to evalu-
ate if some planted oysters were silted in and not adequately
TABLE 5.
Cumulative percent mortality of tethered oysters, and oysters in
dredged samples as a function of transplant time.
Month of Transplant
Method
March
May
September
October
Tethered
Dredged
76
,54
93
55
59
15
38
9
sampled with the dredge. There were no significant differences in
recent or cumulative mortality based on si/e of the tethered oys-
ters.
Because all tethered oysters were large and the growth incre-
ment was small relative to the potential error, the monthly growth
increment of tethered oysters was difficult to measure. This diffi-
culty is evident in the fluctuations in increment growth for the
various size classes (Fig. .5) and the negative growth measured for
some months. Growth, as indicated by new shell being accreted to
the oysters, was observed on some oysters in all but the coldest
months.
Because individual oysters were followed, cumulative growth
is the difference between the initial measurement and the measure-
ment of surviving oysters at any time period (Fig. 5). Because not
all oysters survived through all time periods, cumulative growth
reflects both survival and growth of individuals.
By November there were no differences in growth of surviving
tethered oysters classed as market-sized in March and May, but
individuals in both groups had grown more than those tethered in
September and October. There was no statistically significant
growth for either of these latter two periods. Growth of submarket
size oysters was also at the limits of detection. The 70- to 75-mm
size class showed >0 growth only for the May and September
groups when the mean were 4.8 and 1.8 mm. respectively. With
the exception of the March tethered individual (only one oyster
survived to October) oyster classed as small did not show tnea-
surable growth.
DISCUSSION
Hopkins and Men/el ( 1952) indicated that the major difficulty
in deriving estimates of production was not related to measurement
of growth, but to measurement of losses due to mortality. In our
case, where only large oysters were being evaluated and growth
was poor; it was also difficult to assess growth.
The dominant themes of Delaware Bay oyster transplantation
in 1999 were related to high Dermo (P. niciriiiiis) levels and the
associated high mortality and low chlorophyll and the associated
poor growth. There is a general hypothesis that mortality of trans-
planted, market-sized oysters, due to disease or other factors, can
be made up for by oysters growing from smaller sizes to the
market classes during the year Powell et al. (1997). This can hap-
pen in some years (Canzonier 1998), but in periods such as 1999
with high P. iiiarinus levels and relatively low food, growth may
46
Kraeuter et al.
Mar 70-75 mm
Apnl Nby June July Aug Sepi Oct Nov Dec
May 70*75 mm
-\ 1 ! ! ! '■ \ \ 1
Apnl May June July Aug Scpi Ocl Nov Dec
Sept 70-75 mm
1^1 May lunc July Aug SqX On Nov Dec
Oct >75 mm
i
1
■
1
■
1 1 1 1
Alril
Mv
luoe
July
Am
Sq«
On
Nov
Dec
Oct 70-75 mm
7J " —
1
S
i; 1 1
III
■ ! ' 1 1
Mar 63-70 mm
1
a
E
g ,5
■ ■ -
m
■■
^Hll
■
1
1
1 1 1 1 1 1 i
Apnl
my
June July Aug Sept
May 63-70 mm
Oci
Nov
!
Dee
1
g jj^
■ ■
3
.II.B
[ ; \
t
Apnl
Miv
June July Aug Scpl
Sept 63-70 mm
OCL
Nov
Dec
!
1
= 2S
s
■
t ----- - ■ -"'
Ap«l
Miy
1 1 !
June July Aug Sept
Oct 63-70 mm
Oci
Nov
Dee
i
i
^
! ! 1
1
i
^
1 } 1
!
!
- ^-4—1 ! I I H ! I I
Apnl May June luly Aug Scpi Oci Nov Doc Apnl May lunc luly /wg Scpl Oci Nov Dec Apnl May lunc July Aug Scpl Oci Nov Dk
Figure 5. Cumulative growtli of >75 nini, 7(1-75 mm, and 63- to 7(l-mni tetiiered oysters transplanted from Shell Rock to Delaware Bay ground
554 D in 1999. Transplant months Here March (top graphs), Ma_\ (middle top graphs), Septemher (middle hottom graphs), and October (bottom
graphs). Negative growth is due to measurement error. All oysters were followed as individuals and growth is the summation of all oysters alive
in that size class at the time of measurement.
be reduced to the point that this hypothesis is not valid. Neither the
tethered oysters nor the transplanted oysters in the dredged
samples, of any size class, in the present study showed statistically
significant growth.
The data did not show statistically sigiuficant differences in
numbers, based on month of transplant, of market, submarket. or
total oysters per bushel in final sampling in November. This sug-
gests that in periods of high P. imiriinis. high i-i-)ortality. and low
food the timing of transplantation is not a major consideration
from the point of view of the numerical yield of market oysters. In
addition to the nearly 50% losses of submarket and market oysters,
losses of si-nall oysters exceeding 65% suggest that transplantation
of small oysters with the expectation that they will grow into the
market-size category is not an efficient use of the resource under
high P. mariiuis conditions.
In view of lack of significant differences in the numbers of
i-i-)arketable oysters associated with transplant month, possible dif-
ferences in meat quantity need to be considered. In all cases (ex-
cept the March transplants when water temperatures were low)
total meat weight improved within one month following transplan-
tation (Table 4). Beyond this initial improvement in there was no
change durinc the summer months, but in all cases there was a
trend (not statistically significant) toward further improvement in
between the October and the November samples. Clearly the im-
provement in meat weight in the May to June period could be due
to the increase in gonadal tissue, but the weight did not decrease in
the summer or fall, after the spawning period, indicating that some
of this weight gain was more than gonadal production. The im-
provement in meat quality occurred in 1999 despite the high dis-
ease levels, high mortality and lack of shell growth.
Comparison with Previous Studies
Powell et al. (1997) modeled the effect of transplanting Dela-
ware Bay seed bed oysters in November. January. March. April,
and May on the number of market size oysters available the fol-
lowing July to November. The model predicted that a November
transplant with a November harvest provided the best yields, and
that growth of submarket sized oysters compensated for the losses
of market sized individuals. Mortality of submarket oysters was
less than for larger ones because the added scope-for-growth offers
these individuals some disease protection. Simulated P. nmrinus
levels peaked slightly above four weighted prevalence a level
nearly reached in the present study. The model simulated that
Increased Biomass Yield of Oysters
47
TABLE 6.
Comparison of niimbcrs of marktl and siihniarkil ovslurs hu.
plantt'd on leastd fjrounds in IMMft to IW7 and IWy.
Year of
Transplant
Market
95 '7r
Confidence
Limit
Submarket
Confidence
Limit
1999
1996/97
62
?6
tl3
232
576
±5?
tl05
Data from 1996 to 1997 are from Canzonier (199S). Data arc from samples
removed from the deck of the transplant vessels.
submarket size oysters were less susceptible to mortality from P.
mciriiuis than the market-sized oysters, which allowed them to
grow to market size and replace larger, individuals with lethal
infections. This simulation was not verified in the present studies.
One reason is that, in contrast with the model simulation, the
smaller oysters did not grow. Thus, they did not increase in bio-
mass fast enough to "outgrow" the parasite and maintain parasite
burdens below lethal levels. It is important to emphasize that the
food present in 1999. as indicated by Chlorophyll a. was lower
than that used in the model of Powell et al. ( 1997). It seems likely
that the low food concentrations in 1999 reduced the potential for
compensatory growth of submarket oysters to replace market oys-
ters that died during the study period. The lack of growth may also
have been a consequence of high disease levels (Men/el & Hop-
kins 19.'i.'i. Paynter 1996). Further, many of the assumptions of the
Powell et al. ( 1997) simulations were based on age/size relation-
ships observed in the Gulf of Mexico, which do not apply to
Delaware Bay. In Delaware Bay. for instance, submarket-sized
oysters (35-75 mm) obtained from seed beds are at least 3 years
old and many of the small oysters (<55 mm) are at least 2 y old.
All sampling of oysters in the Bay indicate that by age 2, oysters
have P. marimis infection levels that are equal to that of older
oysters. Thus, it is not surprising that cumulative mortality for our
submarket and small oy.sters was equal to. or greater than, that of
market-sized oysters. A second major difference between our
study and the model simulations is that significant numbers of
submarket oysters did not grow into market individuals in 1999.
Canzonier et al. (1998) reported on a similar transplant. He
moved oysters from the same seed bed (Shell Rock) in December
1996, and February, May and late August 1997. and sampled them
until November 1997. Growth of oysters into the market size cat-
egory was clearly evident in the 1996 to 1997 period (Canzonier
1998). The number of oysters bu. ' transplanted differed signifi-
cantly between this study and the present one (Table 6). There
were no differences (P = 0.43) in the numbers of market oysters
bu. ' from the deck loads of the two studies, but there were neariy
twice as many submarket oysters in the earlier trial (Table 6). In
1996 to 1997. the percentage of market oysters bu. ' ranged from
8 to 10% whereas in 1999 market oysters were between 18 to 26%
of the total. Canzonier (1998) found the number of market oysters
from dredge samples remained relatively constant throughout the
test period in spite of the substantial mortality. Thus despite twice
as many submarket size oysters and growing conditions that were
better than in 1999. there were no changes in the number of market
size oysters in any month of transplant in 1996 to 1997. Growth of
submarket oysters made up for the loss of older oysters.
As opposed to the 1999 results, in which a 21% decrease in the
numbers of market oysters was observed in all transplants. Can-
zonier ( 1998) reported an insignificant 4% decrease in the number
of market-size oysters at the end of the experiment in November.
P. mariniis levels were generally lower in 1996/97 when compared
to both the model and the 1999 data (Table 7). Cumulative mor-
tality was less for December and February transplants but appar-
ently higher for May and August transplants in 1996/97 when
compared with roughly similar transplant months in 1999 (Table
8). Chlorophyll ii in 1997 showed a slight peak in the spring, a
second peak in June and continued high levels (relative to 1999)
throughout the summer, but a general decline from late August to
November (Fig. 2). In this latter condition. Chlorophyll a in the
earlier period was similar to those in the Powell et al. (1997)
model. The presence in 1996 to 1997 of high summer food con-
centrations, lower P. mariiuis. and consequently lower moitality
than in 1999 suggests that the 1999 conditions may be nearly a
worst-case representation. The only exception would be the pres-
ence of the fall bloom in 1999 that would have allowed the oysters
to enter the winter in better condition. This may or may not be
important because there was no difference between the dry meat
weights in 1996 to 1997 when there was no fall bloom and 1999.
Canzonier (1998) reported that market oysters moved from
Shell Rock in December. February. May, and August averaged the
same dry meat weight (1.2 to 1.3 g) as those at the time of trans-
plant in the present study. His final product in November had a
meat weight of 2.8 g. the same weight as oysters in 1999.
How the increase in meat quality in transplanted oysters, vs.
those marketed directly Iriim the seed beds, would affect profit-
TABLE 7.
Initial and selected months.
December
February
May
.August
Market
Submark
Market
Submark
Market
Submark
Market
Submark
D
l.S
1.3
F
0.8
0.7
A
0.2
0.1
0.1
0.1
M
0.1
0.1
A
2.1
1.4
1.2
1.3
0.7''
1.1
1.0
0.6
S
1.2
1.3
1.9
1.3
1.3
2.3
1.7
1.8
N
0.4
0.6
0.2
1.2
0.5
1.2
0.5
0.9
Weighted prevalence oi P. marimts (Dermo) in oysters transplanted from Shell Rock to 527D m 1996 to 1997. Market >75 mm. Submark = Submarket
(55-75 mm). (From Canzonier 1998).
48
Kraeuter et al.
TABLE 8.
Cumulative percent mortality from planting to November of oysters
from Can/.onier (IWSt and present study.
Study
Month of
Transplant
Present study
March
54
May
55
September
15
October
4
Canzonier (I99SI
Decemtier
43
February
45
May
30
August
15
ability is dependent on the relationship among the following pa-
rameters: 1 ) the number of market oysters bu.~' and/or the amount
of meat bu."' that could have been harvested directly from the seed
beds; 2) the number of market oysters and/or the amount of meat
bur' that could have been harvested from the transplanted oysters;
3) the cost of re-harvesting the transplanted oysters; 4) the added
value that is derived from post-shucking processing (washing with
fresh water and blowing with air to help remove shell materials) a
higher salinity oyster; and 5) the value of the bushel of oysters to
the market. The latter \ alue is dependent on the season of harvest,
competing product and whether the oysters are shucked or sold in
the shell.
If oysters are used as shell stock, there would be little gain in
value to the harvester from an increase in meat yield, because in
current conditions, there is little chance (hat additional price would
be paid (S. Fleetwood. Bivahe Packing, pers. comm.). The best
that could be expected would be a longer term value increase
because of better market acceptance. Before the disease infesta-
tions. Delaware Bay oysters received a premium price because of
their high meat yields. Thus for shell stock oysters, in years of high
or moderately high P. marinus disease-caused mortality, there
would be little to gain from transplantation.
For oysters that are to be shucked, results of both the 1996 to
1997 and 1999 studies indicate a significant increase in meat yield
after transplantation. It is important to note that the meat yield
increase, during months with warm water, can be obtained in one
or at most two months. In 1999 the average meat yield increase by
November was about 1 13'-^. and in 1996 to 1997 the meat yield
increased by about 1339}- (Table 9).
Given that there was no difference in the number of oysters
available for market in November (Table 9) associated with trans-
plantation time, it would appear that there was no value added
from transplantation in any month or tor the average of all months.
It should be emphasized that under current conditions, market
oysters are culled on board. This means that nearly equal numbers
of oysters bu."' would be delivered to the packing house from both
the seed beds and the planted grounds. Under these conditions the
meat from oysters harvested from the planted grounds in both trial
periods would weigh approximately 1249^ more that of oysters
from the seed beds. In both cases the use of oysters for shucking
stock would result in increased yields. The higher salinity on the
planted grounds and the added meat weight, will provide addi-
tional gains during the washing and blowing of the meats during
processing.
CONCLUSIONS
When combined with the Canzonier (1998) study the data
cover two of a myriad of possible cases. In 1996 to 1997 there
were slightly elevated summer chlorophyll levels, moderate
growth and moderate P. marinus. whereas in 1999 there were low
or typical Delaware Bay sunmier chlorophyll levels, no growth and
high P. iiniriiuis. The month of transplant did not have a significant
effect on the numbers of market oysters available at the end of the
year. When P. marinus levels were elevated and food supply was
low. transplanted small oysters were lost at a higher rate than
market or submarket oysters. The data from both studies suggest
that food levels on the planted grounds in the warmer part of the
year are generally sufficient to support increases in meat yield 1 to
2 mo after transplant, but may not be sufficiently high to support
shell growth in all years. Under high to moderate P. marinus
conditions, exclusive of tiiarkel timing, meat weight or shucked
meat volume gain were the most important factors for economic
comparison of market oysters between the seed beds and the
planted grounds.
TABLE 9.
Estimated dry meat yield (g) of market oysters (>76 mm) bu. ' of dredged material at time of transplant (Shell Rock) and in November
1997 and 1999.
Shell Rock
Transplants
Transplant Month
Oyster/bu.
Dry Meat Wt
Dry Meat/bu.
Oyster/bu.
Dry Meat Wt
Dry Meat/bu.
1999
March 99
63
1.1
69
32
2.9
92
May 99
34
1.5
51
34
3.1
105
September 99
29
1.6
44
27
2.5
68
October 99
25
1.1
28
25
2.7
68
Average
38
1.3
49
30
2.8
84
1996/1997
December 96
110
1.1
121
108
2.8
302
February 97
92
1.2
110
110
2.5
275
May 97
95
1.5
143
93
2.7
251
August 97
133
1.3
174
106
3.0
318
Average
108
1.2
130
104
2.8
291
Oyster numbers for Shell Rock have been adjusted by using data from the first month of post transplant sampling to accommodate for differences culled
deck load samples and dredge samples. Oysters transplanted from Shell Rock by month of transplant.
Increasbu BioMASS Yield of Oysters
49
ACKNOWLEDGMENTS
The study was funded through funds supphed by the State of
New Jersey for evaluation of the Delaware Bay oyster resources,
and allocated through the Oyster Industry Science Committee of
the Delaware Bay Shellfish Council. The present study could not
have been completed without the on-the-water efforts of Royce
Reed and Russell Babb of NJDEP— Shellfisheries. Staff of the
Haskin Shellfish Research Laboratory (Bob Barber. Beth Brewster
and Meagan Cummings) were instrumental in carrying out much
of the sampling and sample processing efforts. The NJ Agriculture
Experiment Station also pro\ided support.
LITERATURE CITED
Andrews. J. D. & J. L. McHugli. \95T. The sLir\i\al and giowlh of South
Carolina seed oysters in Virginia waters. Pidc. Nur. Shclljlsh As.s<ic.
47:.V17.
Bushek, D., S. E. Ford & S. K. .Mien. 1994. Evaluation of melhods using
Ray's fluid thioglycollate medium for diagnosis of Perkinsiis mariniis
infection in the eastern oyster. Cnissostiva virginicn. Ann. Rev. Fi.sh
D/sraiM 4:201-217.
Canzonier. W. J. 1998. Increased oyster production hy alteration of planing
season. Commercial scale project in Delaware Bay — 1996 to 1998.
Ford. S. 1997. History and present status of molluscan shellfisheries from
Bamegat Bay to Delaware Bay. In: C. L. MacKenzie. Jr.. V. G. Burrell.
Jr., A. Rosenfield & W. L. Hobart. editors. The history, present con-
dition, and future of the molluscan fisheries of North and Central
America and Europe. Volume 1. Atlantic and Gulf Coasts. US Dept.
Comm. NOAA Tech. Rept. NMFS 127. pp. 119-140.
Goode. G. B. 1887. The fisheries and fishery mdustries of the United
States. Washington. DC: in ."i sections.
Hargis. W. J.. Jr. & D. S. Haven. 1988. The nnperilled oyster industry of
Virginia. A critical analysis with recommendations for restoration. Spe-
cial Report 290 in Applied Marine Science and Ocean Engineering,
Virginia Institute of Marine Science. Gloucester Point. VA. 130 pp.
Haskin. H. H.. R. A. Lutz & C. E. Epifanio. 1983. Ch. 13. Benthos (shell-
fish). In: J. H. Sharp (ed.). The Delaware Estuary: Research as hack-
ground for estuarine management and development. A report to the
Delaware River and Ba> .Authority. Unnersity of Delaware. Lewes.
Delaware. 326 pp.
Haskin. H. H. & S. Ford. 1983. Quantitative effects of MSX disease iha-
plosporidium nelsoni) on production of the New Jersey oyster beds in
Delaware Bay. USA. Int. Counc. E,\plor. Sea. CM 1983/Gen:7/Mini
Symp.. Goteborg. Sweden.
Hopkins. S. & R. W. Menzel. 1952. Methods for the study of oyster plant-
ings. Convention Addresses NaL Shellfish. Assoe. 1952:108-112.
IngersoU. E. 1881. The oyster industry. In: The history and present con-
dition of the fishery industries: Tenth Census of the United States.
Department of the Interior. Washington. DC 251 pp.
Menzel. R. W. & S. H. Hopkins. 1955. Growth of oysters parasitized by the
fungus Dermocystidium marinum and by the trematode Bucephalus
cueiihis. J. Parasitol. 41:333-342.
Paynter. K. T. 1996. The effects of Perkinsus mariniis infection on physi-
ological processes in the eastern oyster. Cnissosrren virginicn. J. Shell-
fish Res. 15:119-125.
Powell. E. N.. J. M. Klinck. E. E. Hoffman & S. Ford. 1997. Varying the
timing of oyster transplant: implications for management from simu-
lation studies. Fish. Oceanogr. 6:4. 213-237.
Ray. S. M. 1954. Biological studies of Dermocystidium mariiuim. Rice
Institute Pamphlet. Special Issue. (The Rice Institute. Houston. Texas).
Strickland. J. D. H. & T. R. Parsons. 1968. A practical handbook of sea-
water analysis. Fish. Res. Bd. Canada. Bull. 167. 311 pp.
.loiinuil oj Slu'ltfisk Rcscanh. Veil. 22. No. I, 51-59. 200.^.
U.S. CONSUMERS: EXAMINING THE DECISION TO CONSUME OYSTERS AND THE
DECISION OE HOW FREQUENTLY TO CONSUME OYSTERS
LISA HOUSE,'* TERRILL R. HANSON," AND S. SURESHWARAN'
^ Fo(xl and Resource Economics Di'purtnwnt. University of Florida. P.O. Box 1J024U, Gainesville.
Florida 32611: 'Department of Agricultural Economics. Mississippi State University, PO Box 5187.
Mississippi State, Mississippi 39762: and Higher Education Programs, Cooperative State Research,
Education and Extension Sen'ice, USDA, Mail Stop 2251, 1400 Independence Ave, SW, Washington, DC
20250-2250
ABSTRACT Oyster consumption has been decreasing in the United States. Investigating consumer attitudes and preferences can help
identify factors involved in this decrease. This study used data obtained through a nationwide survey in a douhle-hurdle regression
model to determine factors that influence both the decision to consume oysters and frequency of consumption. Results uidicate there
is a significant difference in the reasons people choose to eat oysters or not and the reasons oyster consumers choose how frequently
to eat oysters. Concern for product safety significantly influenced the decision of how frequently to consume but not whether to
consume oysters. Consumers also indicated a potential willingness to pay for measures that would increase product safety.
KEY WORDS: consumer preference, double-hurdle model, food satety . marketing, oyster industry
INTRODUCTION
METHODS
Overall per capita fresh shellfish consumption in the United
States has increased from 2.5 pounds in 1980 to a high of 4.7
pounds in 2000 (Fig. 1). Per capita consumption of oysters, how-
ever, has decreased from an average of 0.35 pounds per year
(average of 1980-1989) to 0.25 pounds in 1990 to 0.20 pounds in
1999 and 2001 (USDOC 2001; Fig. 2).
Food safety is a factor often blamed for decreases in consump-
tion of oysters. In a 1993 news release, a multi-state outbreak of
viral gastroenteritis related to consumption of oysters occurred iti
Louisiana. Maryland, Mississippi, and North Carolina (Centers for
Disease Control and Prevention 1993). In 1998. bacteria-tainted
oysters from Texas were identified as the cause of sickness for 368
people, and in the preceding summer, 209 laboratory-confirmed
cases of illnesses were linked to consumption of raw oysters har-
vested in the Pacific Northwest (ABC News 1998). The Center for
Science in the Public Interest has asked FDA "to take immediate
action to protect consumers from raw oysters contaminated with
deadly bacteria" (Center for Science in the Public Interest 2000).
They cite 36 deaths in the previous 2 years and 1 19 deaths since
1989 associated with Vibrio viiliiifuus — contaminated raw oysters
and other shellfish. In 1990. Billups (2001) showed only 9% of
respondents considered oysters "not at all safe" compared with
31% rate in a similar survey conducted 5 years later.
Although food safety is suspected to be a major factor in the
decision to consume oysters, other factors may be involved. Re-
gional and national oyster consumption can be affected by many
determinants that may vary across geographical region, ethnicity,
income levels, and perceptions of nutritiim (Wessells et al. 1994.
Gempesaw et al. 1995. Wessells & Anderson 1995. Manalo &
Gempesaw 1997, Wessells & Holland 1998, Holland & Wessells
1998). The goal of this study was to investigate the decision to
consume oysters and the decision of frequency of oyster consump-
tion.
*Corresponding author. E-mail: lahouse@'utl.edu
The data for this study was obtained through a mail survey.
After conducting a number of focus groups of seafood consumers
and nonconsumers (in three locations in the United States), and
conducting survey pretests, a questionnaire designed to elicit in-
formation on seafood consumption, specifically consumption of
oysters, shrimp, tuna, and catfish, was mailed to a sample of 9(J00
households in the United States, with 1000 mailed to each of the
nine major census regions (shown in Fig. 3: Hanson et al. 2002).
The stratified sample was chosen as the region is expected to be a
significant determinant of both the choice to consume and the
choice of how often to consume oysters. The surveys were mailed
in late 2000 and early 2001, with households receiving a second
copy of the survey if they did not return the first. This approach
resulted in a return of 1 790 surveys or a response rate of 20. 1 %
(after accounting for "return-to-sender" surveys). Because of the
length and complexity of the survey, a large number of respon-
dents did not answer all of the questions in the survey, therefore,
a total of 874 observations are included in this study.
Table 1 shows descriptive statistics for the responses used in
this study. Compared with U.S. Census data (United States Census
Bureau 2000), the results showed a larger percent of Caucasians
responded to the survey (89% in the survey compared with 75% in
the 2000 US Census). The survey results also contained a sample
slightly older than the US population, with 69% of survey respon-
dents over the age of 45. compared with 53% of the US adult (over
25) population. The tnean response for income in the survey was
in the S50,000-$59,999 category, compared with a US mean of
$42,148. Religious composition of the survey respondents corre-
sponds to that presented in the World Almanac and Book of Facts
(1999), i.e., 85% of the US population practices Christianity, in-
cluding 23% Catholic, and approximately 2% and 1% of the US
population practices Judaism and Islam, respectively. Our survey
results indicated 83% Christianity with 25% Catholic, and 3%'
practicing Judaism.
In a series of six questions, respondents were asked to indicate
how often they consumed oysters for breakfast, lunch, and dinner,
both at home and away from home. This differs from most previ-
ous studies (including Cheng & Capps 1988. Yen & Huang 1996)
52
House et al.
o>
o
^—
CN
CO
■V
lO
CD
h-
oo
CJ)
o
^—
oo
O)
CD
cn
cn
O)
O)
CT)
cn
o>
cn
o
o
a>
o>
O)
O)
cn
OJ
cn
<3)
cn
cn
cn
o
CM
o
CM
Figure 1. I'liited States per capita fresh and liozen shellllsh consiimplicin (Source: IISDA, ERS. 1999).
that analyze at-home consumption only. Overall. 56.9% of the
respondents indicated that they never ate oysters. The means and
ranges of the responses are shown in Table 2. As expected, con-
sumption of oysters, as well as other seafood products, differed by
region of the respondent's residence (Fig, 3),
Additionally, respondents were asked to identify and rank the
top three reasons they consumed and did not consume oysters.
Results from the question on reasons nonconsumers do not con-
sume oysters and why consumers do not consume more oysters
provide an interesting insight into the data (Fig. 4). Visual inspec-
tion of the results from this question may provide support for a
double-hurdle regression model because it appears nonconsumers
have different reasons for not consuming compared with consum-
ers decision on frequency of consumption.
A number of factors were hypothesized to be relevant to the
consumption and frequency of consumption decisions. The same
set of variables was used as regressors in both equations as theory
provides no guidance for differences and to allow for a specifica-
tion test. The dependent variable was constructed from responses
to a set of six questions regarding frequency of consumption of
oysters for breakfast, lunch, and dinner at-home and away-from-
home. If a respondent indicated they never consumed oysters for
each of the six questions, the value of the dependent variable was
set to zero. For the sample, 56,9% of the responses were zero. For
the remainder of the sample, the responses were summed to de-
termine the frequency of consumption in one month. For example,
if a respondent answered they consumed oysters once per month
for dinner at home and once per month for dinner away from
home, but never for lunches and breakfasts, their frequency of
consumption for the month was two. Those who did eat oysters
consumed oysters on an average of 2.2 times per month. Quantity
of oyster consumption was not obtained in this survey because
respondents were not asked how much was consumed (or by how
many in the household) because of time and space limitations of
the survey. Additionally, because the survey was asking for all
consumption, including away from home and recreational catch, it
was determined from the focus groups and test surveys that re-
spondents were having difficulty answering in terms of quantity
(i.e,, pounds or ounces — other quantities, such as number of oys-
ters, were not considered because of the fact other species were
considered and did not have comparable measures).
Independent variables included demographic variables (age,
gender, ethnicity, religion, household income), variables relating
to the respondents geographic location and variables relating to
slated preference. For geographic location, a dummy variable was
included representing the census region the respondent belonged
to, as well as one variable that represented how close the respon-
dent currently lives to a coast. It was hypothesized that persons
li\'ing closer to the coast would have a higher probability of con-
suming shellfish. Other expected explanatory variables included
perceptions of safety and top reasons for eating and not eating
oysters as indicated by the respondent. Descriptive statistics for all
variables are shown in Tables 1 (demographic) and 3 (other).
Model
Cheng and Capps (1988) and Yen and Huang (1996) recog-
nized the restrictions of using a tobit model in demand analysis for
finfish and shellfish. The tobit model assumes the factors that
affect level of consumption are the same as those that determine
the probability of consumption. Cheng and Capps (1988) used a
Heckman two-step procedure and Yen and Huang (1996) used a
generalized double hurdle model to analyze household demand for
finfish. As a result of information obtained in focus groups and the
preliminary visual appearance of the data, we have chosen to use
Cragg's ( 1971 ) double-hurdle model, similar to the model u.sed by
Yen and Huang (1996).
The double-hurdle model has separate participation and con-
sumption equations that are related in the following manner:
=
if V,* > and </, >
otherwise
(1)
(2)
U.S. 0\sTtR C0N.SUMPT10N - To Eat or Not to Eat
53
0.30 1
0.20
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Figure 2. I'nited Slates per capita consumption of oysters (Source: L'SDOC7N0.4.4/NMFS, 'Fisheries of the L.S., 2001,' September 2002).
where v,* represents the consumption decision and i/, is a latent
variable describing participation as shown below:
= .V,'P + £,
(3)
the same explanatory variables appear in all three equations, the
following value will be distributed as a x" random variable with
degrees of freedom equal to the number of explanatory variables
under the null hypothesis that the Tobit specification is correct:
a + Tii
(4)
where .v, and -, are vectors of explanatory variables and \i and a are
vectors of parameters. Estimation o( the double-hurdle model is
straightforward. Maximum likelihood estimation of a probit equa-
tion is used to evaluate the censoring rule (r,'a). whereas maxi-
mum likelihood estimates that account for a truncated normal dis-
tribution are used for the subsample of uncensored obser\ alions. A
specification test that evaluates the restrictions imposed by the
tobit specification (assumption that the decisions are based on the
same parameters) is obtained through a comparison of the log-
likelihood function \ alues of the tobit. probit. and truncated nor-
mal regression models (Greene 1993). Specifically, assuming that
'^ — -VTdhit ./pr.ihil ./Truncalcd'- ^-'j
where the /,s represent the respective log-likelihood function val-
ues.
RESULTS
Using the double-hurdle model with frequency of oyster con-
sumption as the dependent \ ariable. the model was estimated with
the variables described in Table 4. The coefficients from the probit
and truncated tobit equations, as well as the marginal effects (cal-
culated at the means) are reported in Table 5. The probit model
correctly predicted a consumer's likelihood to consume or not
1.00
0.80
0.60
0.40
0.20
0.00
T3
n I
III
I I I IIJUULIJ
-a c
^1
(/I
CS
C
CO
SI
2
3
2
4-»
z
^
4-1
C
U^
c
3
■t^
^mi
u
cd
u
«a
U
UJ
^
UL)
o 2
1) (J
c
2
c
3
O
2
a.
Figure 3. Percent consumption of oysters by region.
54
House et al.
TABLE 1.
Summary of demographics.
Oyster
Nonconsumers ( % )
Oyster
Consumers ( % )
Overall
Sample (%)
Age of Respondent
Greater than 6?
Between 50 and 65
Between 35 and 50
Under 35
Gender
Percent female
Household Income
Less than $29,999
Between $30,000 and $59,999
Between $60,000 and $99,999
100.000 or greater
Region of Residence
New England
Mid-Atlantic
Southeast Atlantic
East North Central
East South Central
West North Central
West South Central
Mountain
Pacific
Lives within 50 miles of Coast
Religion
Catholic
Christian
Other
Ethnicity
Caucasian
Noncaucasian
Education
High school or less
Some College
College degree(s)
17.3
34.0
39.4
9.3
52.7
16.3
37.2
29.2
17.3
13.1
10.7
9.3
14.7
8.2
13.9
7,0
13.5
9.7
29.4
26.4
56.1
17.5
90.5
9.5
17.1
32.2
50.7
19.6
39.3
33.7
7.4
67.4
11.4
34.0
28.1
26.5
9.8
8.8
14.6
8.0
12.2
9.3
13.8
13.0
10.6
30.0
23.6
59.4
17.0
87.3
12.7
14.9
30.0
18.3
37.0
36.3
8.5
59.0
14.2
35.8
28.7
21.3
11.7
9.8
11.6
n.8
10.0
11.9
10.0
13.3
10.1
29.6
25.2
57.6
17.3
89.1
10.9
16.1
31.2
52.6
consume oysters 87% of the time (incorrectly predicted consump-
tion 49c of the time and no consumption 9% of the time). The
results of the test shown in equation (5) indicate the double-hurdle
model is a better specification than the traditional tobit (\ =
264.9, df = 431. The results indicated that different variables
affected the decision to consume versus the decision of frequency
of consumption, as expected. A set of variables was included to
determine if the location of purchase of seafood affected either
decision. Results indicated that if a person bought seafood (any
seafood, not just oysters) at grocery stores (GRSOURCE) or spe-
cialty stores (OTHERCS; such as fish markets or gourmet stores),
they were more likely to be oyster consumers. However, these
variables did not significantly influence frequency of consumption.
The variables indicating if a person consumed seafood purchased
from restaurants (RESTSC) or obtained through recreational catch
(RECCATCH) were not significant in determining if a person
would consume oysters, but significantly decreased the frequency
of consumption. A potential explanation for these results is that if
a person purchases seafood (again, any seafood) from grocery
stores or specialty stores, they are a different type of seafood
consumer than someone who purchases from a restaurant or eats
recreational catch. Perhaps they are more "dedicated" seafood con-
sumers than those who eat at restaurants, hence more likely to eat
oysters, as well as consume different types of seafood than those
who eat recreational catch (unlikely to be oysters). Following this
line, a person who does eat oysters, but is a restaurant or recre-
ational catch consumer is likely to consume oysters less frequently.
Our results indicate the average oyster consumer consumes oysters
2.21 times per month. Respondents who purchased seafood from
restaurants were likely to consume oysters 1.16 times per month
and those who indicated recreational catch as a source of seafood
were likely to consume 1.84 times per month.
Respondents were asked to identify the top three reasons they
consumed oysters. These reasons give insight to the type of person
that both consumes oysters and what influences a person to con-
sume more or less frequently. If the person indicated they enjoyed
the flavor (FLAVOR) of oysters, as expected, they were both more
likely to consume oysters (66.5% more likely) and consume oys-
ters more frequently (0.46 more times per month). Tradition
(TRAD) plays a part in determining how frequently a consumer
eats oysters, but did not influence whether the person was a con-
sumer. In other words, those who indicated they eat oysters out of
tradition, or habit, were likely to eat oysters 0.62 times more often
per month. Importance of availability was shown in the probit. but
U.S. OvsThR Consumption - To Eat or Not to Eat
55
TABLE 2.
Statistics on frfqutiK> of ojsttr consumpliun (H = 1(167).
Mean
Mode
(Times
Consiinii
dAIonthl
(% Frequency)
Range
Breakfast at home
(1.0.^
Never (93.0%)
Never to less than weekly
Breakfast away from home
(1.01
Never (97.1%)
Never to less than 1 /month
Lunch at home
(1.14
Never (84.0%)
Never to 1/week
Lunch away from home
11.2(1
Never (74.8%)
Never to 1/week
Dinner at home
(1.21
Never (73.8%)
Never to 1/week
Dinner away from home
(1.34
Never (63.0%)
Never to 1/week
Respondents used a scale of to 6 to indicate frequency where = Never; 1 = Infrequently (<I/month); 2 = l/moiilh. 3
1/week .
Daily.
tiot truncated tobit equation. Consumers who believed availability
was an important reason for consumption were 22.4% more likely
to consume oysters. This may be reinforced by the results from the
regional variables. Additionally, those who indicated variety in
diet (VDIET) was an important factor were 30.3% more likely to
consume oysters. Although insignificant, it is interesting to note
the sign on the coefficient for VDIET in the results from the
truncated tobit equation was negative. Intuitively this is attractive,
as someone interested in adding variety might eat oysters, but not
that frequently. Factors that were indicated as a reason for con-
sutning oysters, hut were not significant, included health reasons
(HEALTH), price (PRICE), convenience (CONVl. preparation
knowledge (KNOWHOW). and aphrodisiac properties (APHROD).
Respondents were also asked to identify the top three reasons
they did not consume oysters, or did not consume oysters more
frequently. Three of these reasons significantly influenced the de-
cision to consume oysters: price (NOPRICE). allergic reaction
(ALLERGY), and taste (TASTE). Consumers who indicated they
did not like the taste of oysters or were allergic to oysters were
significantly less likely, 16.3% and 38.7%. respectively, to con-
sume oysters. Those who indicated price was a reason for not
consuming oysters were 17.9% more likely to be oyster consum-
ers, but were likely to consume 0.39 times less frequently than the
average oyster consumer. Oyster consumers who lacked prepara-
tion knowledge (LPKLDGE) were likely to consume 0.62 times
less frequently per month than average.
Perhaps the most interesting result is that "concerns about prod-
uct safety" (PRODSAFE) did not influence a person's decision
whether to eat oysters. Additionally, a variable that indicated the
respondent believed oysters were the least safe of all seafood prod-
ucts (UNSAFE) was not significant in the decision to consume.
Concern about product safety did. however, decrease frequency of
o
^^ ^ CC
(J- < J
c <u
S
c
o
O M
o
•♦-•
a "^
rt ^
o
s
(30
c
a>
Prepar
Know
3
u
D
H
o
o
H
S
C
o
a.
0-
CO
X
tj
INon- Consumers D Consumers
Figure 4. Reasons given for not consuming oysters or not consuming more oysters.
56
House et al.
TABLE 3.
Statistics on factors included in the double-hurdle model.
Mean.
Mean.
Overall
Nonconsumers
Consumers
Mean
Frequency of oyster consumption (dependent variable)
U/monlh (4M7 observations)
2.21/month (377 observations)
U.y5/month
Indicated oysters were the least safe of all
shellfish and fintlsh
products
34.6%
44.5%
39.0%
Indicated the following was a source of seafood for con^
iumption:
Grocery store
86.1%
89.4%
87.5%
Restaurant
86.3%
90.7%
88.2%
Recreational catch or fish farms
15.7%
27.1%
20.6%
Fish market or gourmet store
17.5%
37.1%
26.0%
Indicated the following was one of the top
three
reasons
for
consuming oysters
Enjoy flavor
4.4%
65.6%
31.8%
Variety in diet
2.2%
31.6%
15.3%
Availability
1.5%
21.9%
10.6%
Tradition/habit
2.2%
16.6%
8.6%
Health/nutrition
1.0%
16.4%
7.9%
Know how to prepare
0.5%
8.2%
3.9%
Convenience
0.5%
7.2%
3.5%
Price
1.0%
5.9%
3.2%
Aphrodisiac properties
0.3%
4.8%
2.3%
Other
0.3%
4.0%
2.0%
Indicated the followmg was one of the lop
three
reasons
for not
consuming oysters
Taste
49.7%
8.8%
31.5%
Texture
43.8%
10.1%
29.1%
Smell
26.7%
5.5%
17.2%
product safety concerns
20.9%
25.3%
22.9%
Price
12.7%
37.9%
23.9%
Fresh not available
5.1%
20.4%
11.9%
Lack of preparation know ledge
9.8%
12.0%
10.8%
Custom
4.2%
4,4%
4.3%
Health/nutrition
2.5%
6.3%
4.2%
Too time consuming to prepare
3.0%
5.9%
4.3%
Other
8.2%
3.2%
5.8%
consumption for oyster consumers, from the average of 2.21 to
1.63. a 0.58 per month decrease.
Demographics did have an effect on both the choice to con-
sume and the frequency decision. Persons living in the Southeast
Atlantic (SEATL) and West South Central (WSC) regions of the
country were more likely (17.891- and 33.29f respectively) to con-
sume oysters than persons living in New England. Other regions
did not significantly differ from the New England region. Persons
in the East South Central (ESC). West South Central (WSC). and
Pacific (PACIFIC) regions were likely to consume irtore fre-
quently (0.90. 1.08. and 0.80 times per month, respectively) than
those in the New England region. In the United States. 67% of
oyster landings come from the Gulf of Mexico and 23% from the
Pacific region (USDOC 2002). Given the three regions that con-
suined oysters significantly more frequently are closest to oyster
production, these results make intuitive sense.
All income categories above the base category of $30,000 or
less consumed oysters significantly more frequently. However,
income was not a factor in the decision to consume. Birlhdate (BD)
was a factor in both decisions, with younger ages significantly less
likely to consume oysters, or if they were oyster consumers, sig-
nificantly likely to consume less frequently. Education levels, re-
ligion, gender, and ethnicity did not significantly infiuence either
the participation or consumption decisions in this study. However,
the sample did not include a representative portion of the nonCau-
casian population in the United States. Future studies might benefit
from specifically targeting these populations for information on
seafood consumption.
DISCUSSION
The two main goals of this study were to determine whether the
factors that infiuenced the decision to consume oysters differed
from the factors that influenced the decision of how often to con-
sume oyster and to see what factors were significant that could be
used to develop marketing strategies for the oyster industry. Re-
sults showed that the two decisions were based on significantly
different factors, as suspected. Though food safety is often credited
as a reason why people do not consume oysters, this was not. in
fact, the case. Concerns about food safety did influence how often
oyster consumers ate oysters, but did not significantly influence
whether a person was an oyster consumer. In fact, the belief that
oysters are the least safe of all fish and seafood products did not
influence this decision either. Somewhat surprisingly, nearly 45%
of oyster consumers identified oysters as the least safe of all sea-
food products, while only 35% of nonconsumers identified oysters.
U.S. Oyster Consumption - To Eat or Not to Eat
57
\ariate
Source cil purchase
Reasims lor eating oysters
Reasons tor not eating oysters, or
not consuming oysters more
frequentls
TABI.F, 4.
Description ol independent \ariables.
\ ariable Name
Safety perception
Region of residence (U.S. Census
GRSOURCE
RESTSC
RECCATCH
OTHERSC
FLAVOR
HEALTH
TRAD
PRICE
AVAIL
CONV
VDIET
KNOWHOVV
APHROD
NOPRICE
NOFPAVAI
NOCUSTOM
LPKLDGE
TOOTIME
TEXTURE
SMELL
TASTE
TRAUMA
PRODSAFE
ALLERGY
UNSAFE
Description
I if seafood is purchased at a grocery store
1 if seafood is purchased at a restaurant
I if seafood is from recreational catch
I if seafood is purchased at specialty fish markets or gourmet stores
The following variables are 1 if this reason was listed as one of the top three reasons for
consuming oysters:
Enjoy flavor
Health/nutrition
Tradition
Price
Availability
Convenience
Variety in diet
Know ledge of how to prepare
Aphrodisiac properties
The following variables are I if this reason was listed as one of the top three reasons for
NOT consuming oysters, or not consuming MORE oysters:
Price
Lack of availability of fresh products
Custom
Lack of preparation knowledge
Too time consuming to prepare
Dislike texture
Dislike smell
Dislike taste
Traumatic experience
Product safety concerns
Allergic reaction
I if respondent believes oysters are the least safe of all seafood products
Religion
Race/Ethnicity
Income
Education
Proximity to Coast
Age
Gender
NEWENG New England (omitted category)
MIDATL Mid-Atlantic
SEATL Southeasit Atlantic
ENC East North Central
ESC East South Central
WNC West Nonh Central
WSC West South Central
MOUNTAIN Mountain
PACIFIC Pacific
CHRISTIA Christian (omitted category)
CATHOLIC Catholic
OTHERREL Other religions
CAUC 1 if Caucasian, otherwise
EMCI <$30.000 (omitted category)
INC2 $30.000-$.59.999
INC3 $60.000-S99.999
INC4 SIOO.OOO or above
EDUCATI High School degree or less
EDUCAT2 Some College
EDUC.AT3 At least one degree from College
PROXCST I if currently lives within 50 miles of a coast
BD Birth date
GENDER 1 if female
However, 25% of oyster consLimers indicated they ate oysters less
frequently due to product safety concerns.
Results indicated that people did not consume oysters, and did
not consume oysters as frequently, if they indicated price was an
inhibiting factor. Future studies are needed to address the issue of
willingness to pay for safer oyster products. Consumers who in-
dicated price was a reason they did not consume oysters more
frequently were likely to consume oysters 0.39 times per month
58
House et al.
TABLE 5.
Empirical results from double-hurdle model.
Variable
Probit
Truncated
Name
Coefficient
F(z)/X
Coefficient
E(Y*)/X
Source of seafood for consumption
GRSOURCE
0.391**" (0.197)"
0.155
1.949(2.054)
0.263
RESTSC
0.005(0.196)
0.002
-7.783* (2.142)
-1.050
RECCATCH
0.249(0.164)
0.099
-2.711*** (1.509)
-0.366
OTHERSC
0.699* (0.155)
0.277
2.039(1.362)
0.275
Top three reasons
for consuming
oysters
FLAVOR
1.682* (0.181)
0.665
3.434*** (2.016)
0.463
HEALTH
0.155(0.324)
0.061
-2.771 (2.968)
-0.374
TRAD
-0.223(0.241)
-0.088
4.579* (1.746)
0.618
PRICE
-0.201 (0.340)
-0.080
3.124(2.134)
0.422
AVAIL
0.566** (0.261)
0.224
-0.821 (1.445)
-0.111
CONV
0.411(0.467)
0.163
1.210(2.034)
0.163
VDIET
0.766* (0.223)
0.303
-1.576(1.361)
-0.213
KNOWHOW
0.164(0.417)
0.065
1.819(2.147)
0.245
APHROD
0.569(0.517)
0.225
-2.500 (3.286)
-0.337
Top three reasons
for not consuming oysters, or not consuming more oysters
NOPRICE
0.454* (0.155)
0.179
-2.852** (1.473)
-0.385
NOFPAVAI
0.172(0.209)
0.068
0.402(1.749)
0.054
NOCUSTOM
-0.217(0.296)
-0.086
-3.530(3.464)
-0.476
LPKLDGE
0.065(0.184)
0.026
-4.618** (2.170)
-0.623
TOOTIME
-0.314(0.307)
-0.124
0.008 (2.556)
0.001
TEXTURE
-0.030(0.175)
-0.012
3.312(2.523)
0.447
SMELL
-0.215(0.192)
-0.085
-3.531 (3.511)
-0.477
TASTE
-0.412** (0.169)
-0.163
-5.850** (3.054)
-0.790
TRAUMA
-0.727(0.519)
-0.288
14.509(9.523)
1.958
PRODSAFE
-0.145(0.152)
-0.057
-4.311* (1.708)
-0.582
ALLERGY
-0.977** (0.589)
-0.387
-4.596(7.728)
-0.620
BeMe\'ed oysters to be least safe of all seafood products
UNSAFE
-0.048(0.1.%)
-0.190
1.889(1.354)
0.255
Demographics
MIDATL
0.152 (0.279)
0.060
3.535 (3.207)
0.477
SEATL
0.450** (0.270)
0.178
2.263(2.918)
0.305
ENC
-0.118(0.290)
-0.047
4.071 (3.470)
0.549
ESC
0.480 (0.299)
0.190
6.632** (3.211)
0.895
WNC
0.040 (0.297)
0.016
3.991 (3.438)
0.539
WSC
0.840* (0.308)
0.332
8.017* (3.151)
1.082
MOUNTAIN
0.246(0.290)
0.097
3.851 (3.367)
0.520
PACIFIC
0.139(0.274)
0.055
5.927** (3.044)
0.800
CATHOLIC
0.039(0.150)
0.015
-1.259(1.538)
-0.170
OTHERREL
-0.008(0.168)
-0.003
1.183(1.688)
0.160
CAUC
-0.266(0.190)
-0.105
-0.944(1.796)
-0.127
INC2
0.151 (0.193)
0.060
7.973* (2.601)
1.076
INC3
0.099(0.210)
0.039
6.859* (2.634)
0.926
INC4
0.224 (0.229)
0.089
6.105** (2.701)
0.824
EDUCAT2
0.081 (0.190)
0.032
2.545(2.070)
0.343
EDUCAT3
-0.077(0.191)
-0.031
-0.831 (2.014)
-0.112
PROXCST
-0.220(0.185)
-0.087
2.112(1.705)
0.285
BD
-0.008* (0.0002)
-0.003
-0.007* (0.003)
-0.001
GENDER
0.106(0.130)
0.042
1.461 (1.453)
0.197
Log-likelihood function
-281.04
-635.67
Percent of correct
predictions in
prohit model
87.1%
■" One. two, and three asterisks indicate significance at the 0.01,
'' Standard errors of the coefficients are reported in parentheses.
0.05. and 0.10 levels, respectively.
less frequently than the average oyster consumer. However, con-
sumers who indicated concern for product safety was a reason for
not consuming were likely to consume oysters 0.38 titties per
month less frequently. The tradeoff between an increased price due
to increases in costs of implementing safety programs and in-
creases in consumption if consumers believe oysters to be safer is
an area for future investigation.
Overall, this study does identify characteristics that the oyster
industry can use to segment consumers for marketing purposes. As
expected, people living in regions nearest to oyster production are
U.S. O'l'STBR Consumption - To E.m or Not to Eat
59
more likely to consume oysters and more likely to consume more
oysters. Avuilubility ot fresh products also significantly increased
the likelihood of the respondent to consume oysters. Consumers
who purchase seafood products at grocery stores or specialty stores
may be a segment that could be targeted, as they are more likely
to consume oysters.
ACKNOWLEDGMENTS
This research was supported by the Florida Agricultural Ex-
periment Station and the following grants and approved for pub-
lication as Journal Series No. R-09388. This work is a result of
research sponsored in part by the National Oceanic and Atmo-
spheric Administration, U.S. Department of Commerce under
Grant #GMO-99-24. the Mississippi-Alabama Sea Grant Consor-
tium. Mississippi State University, and University of Florida. The
U.S. Government and the Mississippi-Alabama Sea Grant Consor-
tium are authorized to produce and distribute reprints notwith-
standing any copyright notation that may appear hereon. The views
expressed herein are those of the author(s) and do not necessarily
reflect the views of NOAA or any of its subagencies. This material
is based upon work supported by the Cooperative State Research.
Education and Extension Service, U.S. Department of Agriculture,
under Agreement No. 99-.388 1 4-8202. Any opinions, findings,
conclusions, or recommendations expressed in this publication are
those of the author(s) and do not necessarily reflect the view of the
U.S. Department of Agriculture.
LITERATURE CITED
ABC News. 1998. Oysters Cause Illnesses. Retrieved July. 23, 2U01 from
http://abcnews.go.com/sections/living/DailyNews/oysters980723.html.
Billups. A.L. 2001. Seafood Safety. University of Florida, Research and
Graduate Programs. Explore Magazine Vol. 2{ 1 ). March 3 1 .
Center for Science in the Public Interest (CSPI). 2000. FDA Inaction on
Raw Oysters Means More Deaths on the Half Shell. Retrieved May 10.
2001 from http://www.cspinet.org/new/oysters.html.
Center for Disease Control and Prevention (CDC). 1993. Multistate Out-
break of Viral Gastroenteritis Related to Consumption of Oysters —
Louisiana, Maryland, Mississippi, and North Carolina. 1993. Morbidity
and Morlality Weekly Report Series 42(49).
Cheng, H. & O. Capps, Jr. 1988. Demand analysis for fresh and frozen
finfish and shellfish in the United States. Am. J. Agri. Ecoii. 70:533-
542.
Cragg, J. 1971. Some statistical models for limited dependenl \ariahlcs
with application to the demand for durable goods. Ecimimietrua 39:
829-844.
Gempesaw, C. M. 11. J. R. Bacon, C. R. Wessells & A. Manalo. 1995.
Consumer perceptions of aquaculture products. Am. J. Agr. Econ. 77:
1306-1312.
Greene, W. 1995. Limdep version 7.0 user's manual. Econometric Soft-
ware, Inc.
Hanson, T.. L. House. S. Sureshwaran. B. Posadas & A. Liu. 2002. Opin-
ions of U.S. Consumers Toward Oysters: Results of a 2000-2001 Sur-
vey. Mississippi State University. Department of Agricultural Econom-
ics, AEC Research Report 2002-005.
Holland, D. & C. R. Wessells. 1998. Predicting consumer preferences for
fresh salmon: the influence of safety inspection and production method
attributes. Agr. and Res. Econ. Review 27:1-14.
Manalo. A. B. & C. M. Gempesaw, li. 1997. Preferences for oyster attrib-
utes by consumers in the U.S. Northeast. J. Food Dislrih. Res. 28:55-
63.
United States Census Bureau. U.S. Census 2000. Retrieved Februarv 4.
2003 from http://www.census.gov/main/www/cen2000.htinl.
United States Department of Agriculture. Economic Research Service.
1999. Food Consumption. Prices, and Expenditures. Edited by J. Put-
nam and J. Allshouse.
United States Department of Commerce. National Oceanic Atmospheric
Administration. National Marine Fishery Service. 2001. "Fisheries of
the U.S., 2000." Current Fishery Statistics No. 2000. Silver Spring.
MD.
United States Department of Commerce. National Oceanic Atmospheric
Administration, National Marine Fishery Service 2002. "Fisheries of
the U.S.. 2001." CuiTent Fishery Statistics No. 2001. Silver Spring.
MD.
Wessells. C. R. & D. Holland. 1998. Predicting consumer choices for
farmed and wild salmon. Aqtia. Eton, and Manag. 2:49-59.
Wessells, C. R. & J. G. Anderson. 1995. Consumer willingness to pay for
seafood safety assurances. J. Consiim. Affairs 29:85-107.
Wessells. C. R., S. F. Morse, A. Manalo & C. M. Gempesaw. li. 1994.
Consumer Preference for Northeastern Aquaculture Products: Repon
on the Results from a Survey of Northeastern and Mid-Atlantic Con-
sumers. Department of Resource Economics. University of Rhode Is-
land. Rhode Island Experiment Station Pub. No. 3100.
The Worid Almanac and Book of Facts. 1999. Mahwah. NJ: Wodd Al-
manac Books.
Yen, T. S. & L. C. Huang. 1996. Household demand for finfish: A gen-
eralized double-hurdle model. / of Ag. and Res. Econ. 21:220-234.
Juiinuil oj Shellfish Kcseanh. Vol. 22. No. 1. fil-67, 2U03.
REHABILITATION OF THE NORTHERN QUAHOG (HARD CLAM) (MERCENARIA
MERCENARIA) HABITATS BY SHELLING— 11 YEARS IN BARNP:GAT BAY, NEW JERSEY
JOHN N. KRAEUTER,' MICHAEL J. KENNISH," JOSEPH DOBARRO/
STEPHEN R. FEGLEY/ G. E. FLIMLIN JR.'
^Haskiii Shellfish Research Laboratory. Institute of Marine and Coastal Sciences. Rutgers University:
6959 Miller Avenue, Port Norris. New Jer.'iey 08349. 'Institute of Marine and Coastal Sciences. Rutgers
Utiiversity. 71 Dudley Road. New Brunswick. New Jersey 08901. ^Marine Field Station. Institute of
Marine Coastal Science. 132 Great Way Blvd. Tuckerton, New Jersey. '^Department of Oceanography
Castine, Maine 04421. ^ Maine Maritime Academy. Rutgers Cooperative Extension. 1623 Whitesvllle
Road. Thomas River. New Jersey 08753
.ABSTRACT The use of shell or other coarse material to enhance sur\ ival of newly set hard clams (Mcnciiana incrccnariu) has been
suggested as a management strategy to increase clam stocks. Barnegat Bay, New Jersey and surrounding areas supported a large clam
fishery throughout the 1950s and 1960s, but this resource has declined in recent years. We established replicate 20 x 70 m plots of high
shell densitv, low shell density, and no shell (control) in a Latin Square design in 1990 and have obtained periodic samples since that
time. The shell, obtained from ocean quahog processing plants, had been broken into a variety of sizes. High-density shell received
900 bu per plot, and low -density shell received .^00 bu per plot. Plots with high shell density had significantly more clams after 10 years
than those with low-density shell or controls. High shell density significantly increased hard clam recruitment, but this exceeded 1 m""
in only one year, from the years 1990 to 2000. In plots with low shell or in controls, recruitment never exceeded 0.4 nr-. and in half
or more of the years no recruitment was found. Some individual plots with shell did not enhance recruitment, indicating that factors
not investigated must be important as well. In spite of the low recruitment density, there appears to be an increase in survivorship when
the shell content is greater than 8000 gm"".
KEY WORDS: Merccnaria meicfiniha. shelling, hard clam recruitment, quahog
INTRODUCTION
Methods of increasing natural abundance of hard clams {.Mer-
cenaria mercenaria) are important to state resource managers and
the shellfish industry. There are several approaches a manager can
use to improve shellfish stock abundance: ( 1 ) increasing the num-
bers of spawners (spawner sanctuary); (2) reducing harvests or
providing alternate areas in some cycle so the stocks last longer;
(3) adding hatchery produced clam seed to a selected area; and (4)
protecting naturally set clams (shelling or other substrate modifi-
cation and use of chemicals to eliminate predators).
The theoretical concept underlying a "spawner sanctuary" is
that increasing the number or density of clams in an area will
increase the number of eggs, larvae, and, set clams. The potential
for an increased number or greater concentration of clams to pro-
duce more larvae when conditions are favorable is suspect because
it depends on the existence of a spawner-recruit relationship (more
spawners = more recruits) over a wide range of clam densities. In
addition, there are large numbers of clams in most bays even at low
densities, and thus the numbers of clams that inust be transplanted
to have even a small probability of significantly increasing the
number of active spawners in the region is extremely large. Fi-
nally, of those sanctuaries that have been created in New Jersey
and New York, preliminary evidence indicates that little detectable
enhancement of natural hard clam stocks may be expected (Kass-
ner & Malouf 1982, Barber et al. 1988).
Reducing harvests allows clams to be harvested over a longer
period of time while waiting for the next surviving .set. While this
appears to be attractive, hard clams are different than most species
harvested from the wild. Smaller sizes of hard clams (liltlencck)
command a premium price. Econoinic considerations suggest that
most of the clams should be harvested in the smaller sizes and that
larger clams should only be taken as a last resort. Growth rates in
most areas are such that clams remain in these premium si/c
classes only a few years. This suggests that the best economic
returns would be from intense harvest on these sizes. The only way
to manage the fishery for maximizing economic benefit would be
through an extensive monitoring program to delineate areas with
maximum concentrations of appropriate sizes (McHugh 1991).
The third option, the use of hatchery seed to enhance hard clam
production is well established in aquaculture (Manzi & Castagna
1989). In general, predation rates on high-density plantings of seed
without protection devices are too high to recoinmend this option
(Kraeuter & Castagna 1989). Preliminary experiments using low
density seeding of hard clams suggest this may yield higher sur-
vival rates than would be expected from dense plantings [Macfar-
lane (Orleans, MA), and Relyea (F. M. Flowers and Sons, pers.
comm.)]. These observations are supported by the work of Paulsen
and Murray (1987). They conducted a number of short-term (less
than one year) experiments using three seed sizes, at high and low
density, planted both on and below the sediment surface. They
reported that survival (58 days) of clams planted below the sedi-
ment surface at high densities was no greater than if seed were
broadcast. Low-density plantings of hard clams below the surface
significantly increased long-term survivorship when compared
with similar high-density plantings. Peterson et al. (1995) have
provided additional evidence indicating that low-density plantings
of large (>20 mm) seed may be an economically viable means of
increasing hard clam stocks in isolated basins.
The fourth option, modifying the substrate to increase post
settlement survival of juvenile hard clams, has been shown to
work. MacKenzie ( 1977, 1979) demonstrated that treating areas of
bay bottom with various pesticides significantly increased juvenile
hard clam survivorship by eliminating arthropod predators such as.
shrimp and crabs. Siinilar techniques provided additional protec-
tion to seed clams planted in mesh and gravel protected aquacul-
ture plots (Kraeuter & Castagna 1985). The use of this technique
61
62
Kraeuter et al.
is considered to be unacceptable because it requires introducing
toxic chemicals into tlie environment, and these may produce long-
term detrimental effects. Parenthetically, it is plausible that the
massive use of pesticides during the 1950s and 1960s, to control
insects in the coastal marshes of New Jersey, was the proximal
cause of the high abundance of hard clams in some of these shal-
low, poorly flushed systems.
An alternative of the fourth option, that has also been shown to
increase survival of juvenile hard clams in different habitats, is
"shelling" the bottom (Parker 1975. Kraeuter & Castagna 1977,
Kraeuter & Castagna 1989. Kassner et al. 1991). This practice
involves broadcasting pieces of broken shell or stone aggregate
over the bottom to increase the percent composition of larger par-
ticles (stone or shell) in the sediments. This technique was devel-
oped from the many studies revealing that hard clams are more
abundant in areas with a higher percentage of shell in the bottom
(Pratt 1953. Wells 1957. Saila et al. 1967. Walker & Tenore 1984.
Craig & Bright 1986. Papa 1994). The larger particles have two
mechanisms by which they can affect hard clam abundance. Wells
( 1957) suggested that shell might create areas of low current speed
in which small clams either collect {sensu Carriker 1961) or at
least are not swept away. He also proposed that the hard substrates
provide a byssal attachment point for newly set clams. A large
body of evidence indicates that coarser material can interfere with
the ability of many hard clam predators to detect or manipulate
small clams (Arnold 1984, Kraeuter 2001). Any or all of the.se
mechanisms can have a positive effect on natural set resulting in
greater numbers of clams surviving to market size.
The shelling option can be used, but it cannot he used with
confidence. Kassner et al. (1991) found no significant enhance-
ment of a clam area with low abundance in Great South Bay. New
York, one year after placing 12.5L shell m"" on a mud bottom.
Several important variables associated v\ ith construction of shelled
plots are unknown. For example, the amount of shell added must
fall within a bounded region; too little shell may not effectively
deter predators while too much shell may serve as a haven for the
same predators. There is uncertainty regarding the amount of shell
needed to afford protection. For example, most studies and surveys
of natural populations indicate a positive effect of larger particles,
but Day (1987) has observed in the laboratory that mud crab pre-
dation was greater in gravel and gravel and sand mixtures than in
sand alone. She suggests that the gravel substrates offer hiding
places for these small predators and thus increase the predation
rate. Further, little information exists on: density of shelling, shell
size, plot size, substrate type (grain size, percentage of organic
matter, redox discontinuity level, water content, etc.) and their
interactions (density of shell x shell size, density of shell x plot
size, density of shell x substrate type, etc.) relative to clam sur-
vival. This information is essential to allow some predictive capa-
bility concerning whether the increased numbers of clams avail-
able for harvest will justify the cost of the original shelling. In
addition to the effects of shelling on the clams, infonnation con-
cerning the shell size, shelling density, substrate type, and their
interactions is required to evaluate the increased effort that might
be required to harvest the potentially increased numbers of clams
(shell fragments could interfere with the harvest).
METHODS
This study was designed to determine whether shelling the
bottom, at a spatial scale large enough to be meaningful to habitat
management, produces significant increases in hard clam abun-
dance. A subset of the design examined two densities of shell
cover: low density and high density. The major uncontrolled vari-
ables were the sporadic nature of hard clam spat set and predator
populations.
The experimental design was a Latin Square matrix of 20 x 70
m plots (slightly more than 0.15 ha). A rectangular shape was
chosen because a boat was used to place the shell into the plots.
Plots were arrayed in a 3 x 3 Latin Square design with 30 m buffer
zones between each of the separate plots. The entire matrix was
surveyed using sextants: comers were marked with stakes and
buoys. Three treatments were arrayed within the plots: (I) 10
high-density shelling — 900 bushels per plot (15 L/m"); (2) 20 low-
density shelling — 300 bushels per plot (5 L/ni"); and (3) control —
no shell added. Most of the shell consisted of broken pieces (2-8
cm") of A rctica islandicci. although some Spisiila soliilissima shell
and the shell debris of other offshore species could be seen. This
shell is available in large quantity from several local clam-
processing plants.
Shell Spreading
The experiment was located approximately 200 m east by
northeast from Gulf Point in Barnegat Bay. New Jersey. The co-
ordinates of the matrix are: NE corner — 39^44. 23 'N by
74°9.05'W, NW comer— 39°44.23'N by 74°9.I2'W, SE comer
— 39°44.04'N by 74°9.05'W, and SW corner— 39°44'N by
74"9.I2'W. The site, characterized by sandy sediments with rela-
tively low silt-clay content and few naturally occuning shells,
experienced only moderate tidal currents. It had a fairly uniform
bottom composition and water depth (4 m). was protected from the
longest fetches that occur in the bay. and had long been a hard
clam habitat. The latter was determined through discussions with
several local watermen who aided in the site selection process and
designated this area as the best location for our project and least
disruptive to their activities.
Shell was spread onto the experimental plots during the week
of April 23 to April 27. 1990 using the Ocean County Bridge
Department's LCM. the Beujamin H. Mahie. The shelling required
2 days to complete.
The shell was stored in the middle of the ship and transferred
to a hopper with a small catloader. The viilume of the scoop of the
catloader was calibrated previously so that the shell volume going
overboard could be estimated. The shell moved from the hopper
via a conveyor belt to a highway salt spreader located in the bow
approximately 4 m above the water. This procedure produced an
evenly dispersed spread of shell on the bay bottom. SCUBA ob-
servation subsequent to spreading confirmed the even nature of the
shell on the bottom.
During the second day. it became apparent that the volume of
shells delivered was short. To accommodate this, we reduced the
size of the last high-density plot to 20 x 50 m to maintain the same
density of shell. To ensure that all plots received as nearly identical
disturbance as possible, the LCM was powered over each control
plot as if it was being shelled.
Sampling
Samples were retrieved from each plot using a diver-operated
suction sampler. Each plot was located with sextant coordinates (or
later GPS); the center was marked, and a diver was deployed
Rehabilitation of Mercenaria mercenaria
63
approximately 9 ni from the center mark. Diiriny the first year of
sampling (May 1991). a ring made from a bottomless galvanized
bucket was used to mark the area to be sampled. Samples were
collected approximately 1 m apart by removing all material from
the ring to a depth of 10 cm with a suction sampler. All materials
were collected in a .^ mm mesh bag. brought to the surface, and
preserved. During the first year of sampling. 9 samples were col-
lected from each plot each sample covering 0.043 m". Samples
were returned to the laboratory and numbers of clams removed and
the volume of the material was recorded. All hard clams were
measured in length, height and width.
Subsequent samplmg followed the same protocol except thai
the ring was modified and size of the area sampled was increased
to 0.25 m". The number of samples was reduced to five or six
during 1996 and increased to 10 during partial sampling in 1998 (.3
plots) and in 2001. The procedure in the laboratory remained the
same except that that the weight of the dried shell material was
measured rather than its volume. A factor of approximately 850 g
dry weight is equivalent to IL of this material. We chose <15 mm
as the size limit for seed clams (0 y class). For the 1996, 1998. and
2001 samples, we sectioned one of the valves of each clam that
was older than seed to determine approximate age. We counted the
annual growth rings in the valves to determine if the clams were
those that might have set since the 1990 shelling. We could not
accurately age animals older than 10 years; therefore, we consid-
ered these individuals to have been the residual population, even
though by 200! they may have recruited after the experiment
started.
RESULTS
Samples were retrieved from plots on May 2 1 to May 28. 1991 ;
September 30 to October 2, 1992; November 24, 1993; June 23 to
June 25, 1996: August 10 to August 12. 1998; and November 14
to November 15, 2001 . During the first year of sampling, only one
hard clam was found in the 72 samples that were sorted in the
laboratory. Because of insufficient numbers of hard clams in the
samples, these data were not analyzed further.
During the second year, we increased the sample size to 0.25
m~ and reduced the numbers of replicate samples per plot (Table
1). In general, setting was sparse. Data from the second year in-
dicated that on one heavy shelling treatment there was enhanced
setting. Shell weight data indicated that the other heavily shelled
plots were not sampled, and control plots were over represented.
None of the control plots had seed clams, and there were seed
clams on two of the three low-density treatment plots. Because of
the sampling difficulties, no Latin Square analysis was attempted
on the 1992 data. A linear regression of the effect of shell mass on
total clams and seed clams collected showed that shell density had
a significant positive effect on the presence of both total clams and
seed clams (Table 2).
Considerable effort was directed toward surveying the plots for
the third year, and weights of shell indicate we were successful in
sampling the stations in all but one case. Even with this effort, the
low-density plot (2-3). based on shell weight data (Table 1) ap-
pears to have had high-density shell. We conducted an ANOVA
with that plot characterized both "as-sampled" and "corrected". In
TABLE I.
Numbers of replicate samples removed from Barnegat Bay, NY shell plots by year.
Sample (Jrid
1-1
2-2
3-3
1-2
2-3
3-1
1-3
2-1
3-2
Shell Density
H
H
H
L
L
L
C
C
C
Year
1991
# replicates
Mean Shell DW
9
9
9
9
9
9
9
9
9
Total Clams
1
Recruits
1992
# replicates
6
6
6
5
5
8
5
5
4
Mean Shell DW
4994
1676
59.9
1059
0.2
1315
409
2.5
14.6
Total Clams
13
1
3
3
1
1
Recruits
11
1
2
1993
# replicates
5
5
5
5
5
.s
5
5
5
Mean Shell DW
5892
5305
3904
1463
4256
1538
89
27
23
Total Clams
10
6
2
9
3
5
1
1
Recruits
8
6
2
7
2
3
1996
# replicates
5
5
5
5
5
5
5
5
5
Mean Shell DW
6279
3966
4264
312
2986
1893
10
56
154
Total Clams
5
4
7
2
11
2
1
3
Recruits
5
4
6
1
5
2
2
1998
# replicates
Mean Shell DW
Total Clams
Recruits
13
4031
11
10
10
288
5
1
10
1004
4
4
2001
# replicates
10
10
10
10
10
10
10
10
10
Mean Shell DW
3285
2358
5095
639
1104
1039
24
->1
280
Total Clams
15
12
20
5
1
4
2
1
4
Recruits
12
11
IS
3
.^
">
1
H = High density shell. L = low density shell, and C = control. In 1991 replicates were 0.(143 m" all subsequent samples were 0.25 ni-
ls in grams. Clams and recruits are the totals for all samples.
. Shell drv wei.sht
64
Kraeuter et al.
TABLE 2.
Intercept, regression coefficient and correlation coefficient for the
effects of shell density (g) on total clams and those that have
recruited since 1990 (clams 0.25 m"").
Year
Intercept
Regression
IT
1992
Total Clams
0.005 NS
4.028 E-04***
0.47
Recruited Clams
-0.082 NS
3.338 E-04***
0.48
1993
Total Clams
0.297 NS
2.152 E-04**
0.18
Recruited Clams
0.1 IONS
2.080 E-04***
0.26
1996
Total Clams
0.363 NS
1.827 E-04**
0.19
Recruited Clams
0.118NS
1.876 E-04***
0.31
1998
Total Clams
0.314 NS
1.476 E-04*
0.14
Recruited Clams
0.073 NS
1.624 E-04**
0.26
2001
Total Clams
0.129 NS
3.713 E-04***
0.45
Recruited Clams
-0.003 NS
3.703 E-04***
0.53
NS = not sisnificant. *0.05. **0.01.
*0.001.
both cases the results with respect to treatments were similar, and
we have presented the as-sampled data (Table 3). High and low
shell density plots had similar numbers of total clams and seed;
both had significantly more total clams and seed than the controls.
We arrayed the data according to shell density and used linear
regression. Both total numbers of clams and seed clams (Table 2)
were significantly correlated with shell density.
Latin Square analysis of the 1996 data on shell weight indicated
that column 2 had significantly less shell than the other two. These
differences negated further use of the Latin Square. We evaluated
the total clams and recruited clams with ANOVA based on the
three treatments (high density shell, low density shell, and con-
trol): a linear regression for all satiiples (total clams and recruited
clams vs. shell weight) was then computed. There were no sig-
nificant differences in total clams with treatment (Table 3); how-
ever, the regression line showed a significant positive effect of
shell density (Table 2). In contrast, the ANOVA analyzing the
effect of shell on clams that had recruited since 1990 was signifi-
cant. A Tukey (HSD) test found that clam density in high-density
shell and low density shell were not significantly different, and low
TABLE 3.
Tukey (HSD) results (number 0.25 m"" for total number (Total) of
hard clams {Mercenaria mercenaria) and those that had recruited to
the population (Recruit) since the beginning of the
experiment ( 1990).
1993
High Low Control
1996
High Low Control
Total
Recruit
1948
Total
1.29
1.14
High
0.85
Recruit 0.69
1.13
0.73
Low
0.50
0.30
0.07
0.00
Control
0.40
0.10
Total
Recruit
2001
Total
Recruit
1 .07
1.00
High
1.53
1 .00
0.53
Low
0.33
1,40 0.20
0.27
0.67
Control
0.23
0.10
High = those areas covered with high density of shell, low = those areas
covered with low density shell, control = those areas that did not receive
shell. Underlines indicate those treatments that were not significantly dif-
ferent a (a = 0.05).
density shell and control areas had similar clam density (Table 3).
High density shelling increased clam recruitment over that ob-
served in the control areas.
In 1998, only 3 plots were sampled, and ANOVA results were
similar to 1996. There was no difference in total clams between
treatments, but the clams that had recruited since 1990 were more
abundant in high shell plots. There were no significant differences
between low shell and control (Table 3). Again, linear regression
indicated a positive effect of shell density on total and recruited
clams (Table 2).
In 2001. as with previous sampling, Latin Square analysis of
the shell distribution revealed significant differences between all
columns and some rows. The total numbers of clams and clam
recruitment were evaluated relative to shell weight and treatment
type with general ANOVA and linear regression techniques. After
I \+ years, most plots remained intact, but the increasing differ-
ences between rows and columns suggest that the shell is gradually
being dispersed. In contrast to 1996, when total clams were not
significantly different by treatment, both the total and recruiting
clams since 1990 exhibited significant differences by treatment. In
both the total clams and recruited clams, the Tukey (HSD) test
found that high shell density plots had significantly more clams
than either the low-density shell or the control. The latter two
treatments were not significantly different from each other. The
similarity between total and recruiting clams after I l-l- years may
have been greater than indicated by the base data. We were unable
to distinguish ages of clams >10 y. Thus, some of the clams in this
class may have recruited to the area since the shell was placed on
the bottom. In 2001. 20.7% of the sampled clams were in the age
10 or older category. As a comparison in 1996, 31.4% of the clams
were from classes that had recruited before the shell was placed on
the bottom).
Recruitment
We considered clams <15 mm in shell length to be seed clams.
Relatively few of these clatns were found (Table 4), and never in
the control areas. In some years, seed can be as large as 20 mm.
We found only one clam of this size in a control plot (Table 4).
We have attempted to evaluate annual recruitment (long-term
survival) of clams at this site by back calculating from the age data
to determine when particular clams had set (Fig. I). We have
averaged the data from the 1996, 1998, and 2001 samples, but,
because so few animals were obtained by sampling, have not at-
tempted to place error bars around these estimates. With the ex-
ception of 1993, there is a relatively good correspondence between
the back calculated data and that from animals recovered. The
TABLE 4.
Mean number of seed clams m'" bv treatment.
.Seed <15.1 mm
Seed <20.1 mm
Year
High
Low
Control
1992
1993
1996
1998
2001
2.00
2.67
(1
0.50
0.53
0.53
High
Low
Control
2.25
0.75
2.67
0.80
0.27
0.53
0.27
Seed = <15.1 nimor<20.1 mm Shell Length. Numher of 0.25 nr samples
is given in Table 1.
Rehabilitation of Mercenaria mercenaria
65
1.40
-High Shell
-Low Shell
-Control
Figure 1. RiTiuitnieiit of hard clams {Mercenaria mercenaria) into hi)>h-densit> shell. l<iH-dc'nsit\ shill and cdnlrol plots in Barnegeat Bay, New
Jersey. Data represent the average estimated recruitment, based on live animals collected in 1996, 1998, and 2001.
scarcity of animals precluded meaningful statistical analysis of
these data. For both estimates, areas with high shell density had
more recruiting clams than areas of low shell, and these in turn
receive more recruits than control areas. Based on these estimates,
clam recruitment to this area has generally been very low for the
past decade. Annual average recruitment, based on aged shells,
exceeded 1 ni"" only on the high shell plots sampled in 1992.
These same plots approached 1 m'" again in 1994. Recruitment in
the low shell density and control plots was below 0.5 m^" in all
years and has been since 1997. Average annual recruitment in the
high shell plots shows a general trend toward less recruitment from
1998 to present when it reached 0. Data from clams <20. 1 mm also
suggest there has been little or no recruitment since 1996.
Growth
Size-at-age was computed for clams from the 1996, 1998. and
2001 samples. These data were compiled and averaged to yield an
estimate of growth (Fig. 2). Although we have few clams of age 1
and 2, the data indicate that growth is rapid until age 3, and then
abruptly slows. Growth is sporadic after age 5. The largest clam
found at this site was 82.8 mm shell length. In addition, when the
clam meat was being removed to prepare the shell for sectioning it
appeared very dark brown, black, or gray in color, e.xcept in small
clams. This condition has existed in Barnegat Bay and Little Egg
Harbor clams for a number of years.
DISCUSSION
Age (Years!
Figure 2. Growth of hard clams based on average size-al-age of live
clams collected in all experimental plots in 1996, 199S, and 2001, and
all clams <20.1 nun collected in all plots from 1992, 199.V 1996, 1998,
and 2001. ,\ll plots were in Barnegat Bay, New Jersey. Data are mean
length (mm) and the 95'7f confidence limits. Bars lacking conlldence
limits are based on one individual, .\nimals older than 9 could not be
aged, thus all data for ages 10 and 11 come from animals collecled in
1996. .Animals >10 y are based on the average of all data from all
animals aged in all years.
We have demonstrated that shelling increased ihe number of
hard clams on the bottom at an experimental site in lower Barnegat
Bay. These data are consistent with observations about the effects
of shell on the bottom and wild hard clam populations. At this site,
the shell has persisted for 1 1 years and appears to continue to
support hard clam recruitment. After 1 1 years, linear regression of
both total and recruited clams shov\ed the positi\'e effect of shell
density, but the effect of shell on clam recruitment was not sig-
nificant until shell exceeded 8 kg m^" (Fig. 3). The larger numbers
of recruits between 1 992 and 1 994. as well as Ihe lack of difference
between clam abundance in high and low density shell during this
period, suggests that the shell continued to enhance recruitment.
Beginning with the 1996 samples, there was no statistical differ-
ence in clam abundance between the high and low shell density
plots. There was also no significant difference between the low
shell density and the control sites, and by 2001 the high-density
plots were significantly different from the low-density shell and
the control. This appeared to be coupled with a general loss in
overall recruitment at the sites. The low density shelling may have
started to lose its effectiveness, but we cannot determine whether
this reflects a drop in actual recruitment or some loss of effective-
66
Kraeuter et al.
Oto2.0 2IIO4.0 41108.0 8 I to 120 i:i(ol60 l6 1to26-CI
Shell Density (Kg/sq meterl
Figure 3. Numbers of surviving clams m"" based on survival in higli-
densitj shell, low density shell, and control plots placed in Barnegat
Bay New .Jersey in 1990. Data represent back calculated (from regres-
sion equations) mean and 95 "/r confidence limits of the number of live
clams and clams <10 y of age. Numbers of the latter clams are based
on shell sections and ages of animals. These represent the animals
recruited since 1990.
ness of the shell caused by its protracted residence time on the
bottom.
Our study indicates that in areas experiencing low recruitment,
several years of data may be required to thoroughly evaluate the
effectiveness of shelling on the survivorship of hard clam seed.
Similar experiments, perhaps of significantly smaller scale, should
be conducted on different types of bottom to ascertain how much
shell is required.
Economics is one of the many important factors to consider
before any large-scale shelling program commences. It clearly
costs more to add more shell to the bottom, but we do not have
sufficient data to determine full costs per unit of shell spread. The
cost of the shell is a direct multiple of the amount to be spread (3
X more shell will cost 3 x more), but the cost of spreading the
higher density shell will be somewhat less per unit on the bottom
than will the lower density shelling. Shell costs are not insignifi-
cant, and transportation adds to these costs. In New Jersey there
are large quantities of shell produced by the surf clam and ocean
quahog processing plants and these can be purchased for about
$0.50 bu~'. The logistics of handling the shell on a regular basis
have precluded it being available free for repletion. Private con-
tractors remove the shell and store it for roads and other purposes.
Oyster shell repletion, utilizing large boats (3,000 -l- bu load) cost
about $1,000 day"', for the boat. Smaller boats (1.000 bu. load)
cost about $600 day"'. Extrapolating from these basic data, it
would cost between $2,300 and $3,100 acre"' to spread shell at the
highest density used in this experiment, but boat availability, trans-
port of shell to the sites, and other logistical costs may make these
data unreliable. We know that this particular shelling lasted at least
1 1 years without substantial loss of shell. Figure 2 also makes it
clear that high-density shell increased the clam population from a
mean of 0.7 m""-7.6 m"". nearly a factor of 10 increase, during the
first few years. This population generally persisted throughout the
course of the experiment. It is impossible to know whether the
sporadic nature of the recruitment was due to changes in recruit-
ment, shell effectiveness, or a combination of the two.
It is also unclear how long a plot can continue to enhance clam
set. It is certain that high shell density continued to support more
clam, even after 1 1 years, but there has been a noticeable decline
in the number of clam .seed (those <15 mm) through time. This is
true in both the shelled and unshelled areas. As noted above,
whether this is due to loss of effectiveness of the shell or lack of
recruiting individuals cannot be determined, but there was a gen-
eral tendency for low-density shell to be somewhat effective at the
beginning. By 2001, low-density shell had clearly reduced capac-
ity to sustain clam recruitment, but high density shelling continued
to retain recruited animals. The different rates of loss of effective-
ness make it tempting to conclude this is a function of the shell
density; however, under conditions of low recruitment, other fac-
tors may be operative and the interpretation remains uncertain. It
will require placing shell out for a number of consecutive years on
different bottom types to allow evaluation of the length of time
shell remains effective. This requires differentiation of recruitment
processes on freshly planted shell and shell placed out for a num-
ber of years.
Disturbance of the shell either by natural physical forces, such
as burial by sediments, or human activities, such as clam harvest-
ers working within an area, could alter the effectiveness of the
shell. We have no data regarding the effects of increased clam
harvesting on the enhancement capability of each shell density.
The density of marketable hard clams was low in this area; there-
fore, we do not believe disruption of the shell or sediment by
harvesting was high during the study. Pieces of shell were covered
with fouling organisms so at least some of the material remained
near the sediment surface for the duration of the study.
A 2002 survey of hard clam populations by the New Jersey
Department of Environmental Protection in Little Egg Harbor Bay
stopped just south of our experimental area, but it reported a nearly
two thirds reduction in hard clam standing stocks since the last
survey in the middle 1980s (Joseph pers. Comm.). Commercial
clam harvesters working throughout the area also indicated that
they believe that clam populations have declined significantly in
recent years.
Low levels of recruitment made it dilTicult to detect statistically
significant effects, even with 0.25 m" samples. It was only through
time and repeated sampling that we were able to evaluate the
effectiveness of the shell in this low clam density, low recruitment
area. It is also clear that in the 1 1 years of this experiment that the
control areas had just sufficient recruitment to maintain the popu-
lation al the 1990 levels. This study only covered one type of
substrate and the results could be very different under different
substrate, depth, and current regimes.
While the relationship between shell in the bottom and in-
creased hard clam density occurs wherever studies of natural popu-
lations have been conducted (Gulf of Mexico to New England), the
types of predators and their effects are substantially different. Dur-
ing 1996. we enumerated other organisms in the samples. There
was an increase in species, mainly epifauna. on the shelled areas
relative to the controls. This clearly indicates that other species are
enhanced as well. The nature of the sampling (suction sampler and
a 3-mm mesh collection bag) precluded examination of the effects
on infauna. Many of the epifauna we found are known to prey on
hard clam seed (Kraeuter 2001 ). The "reef effect" from mounds of
shell may cause an increase in epifaunal predators. It is important
to spread the shell evenly and not allow mounds to form that would
attract and retain these organisms. The best combination is for
shell to become an integral part of the bottom with only a small
portion protruding above the sediment surface. In other areas, par-
ticularly where oyster setting is high, the effect of shelling on the
establishment of oyster populations needs to be carefully evalu-
ated. Extrapolation of shell density recommendations to different
Rehabilitation of Mercenaria mercenaria
67
environments should be examined carefully before large-scale at-
tempts are made.
The slow growth rate of clams after 3 to 5 years and the small
size oi clams >I0 years old. the small size of the largest clam
collected (82.8 mm shell length), and the dark color of the meat on
most clams suggests that conditions at this site are not optimal for
hard clam production at present.
CONCLUSIONS
Shelling the bottom of Barnegat Bay. New Jersey increased the
abimdance of hard clam seed by nearly a factor of 10. The shell
remained on the plots for at least 1 1 years and continued to en-
hance the set throughout that period. Settlement was 0.5 clams m"^
on the control plots and exceeded 1 m"" only once in the high shell
areas. Clams <I3 mm in shell lenizth were never found in control
plots. This method presents a potentially viable protocol for in-
creasing survivorship of small clams from natural set. but more
thorough evaluation is needed before it can be used on a variety of
bottom types.
ACKNOWLEDGMENTS
This study would not have been possible without a large num-
ber of volunteers, and the Ocean County Board of Chosen Free-
holders who allowed the use of their LCM, and its crew from the
Bridge Department. The initial grant to provide for shelling and
sampling in 1992 came from the New Jersey Department of En-
vironmental Protection. Intermediate sampling was based on vol-
unteer effort and limited fund from the New Jersey Agriculture
Experiment Station and the New Jersey Commission on Science
and Technology. The final sampling was provided by funds from
the Fisheries Information Development Center.
LITERATURE CITED
Arnold. W, S. I9S4. The effects of prey iv/e. prediitor size, and sediment
compositicm on the rate of predation of the blue crah {CcilUnecles
sa/Hdiis Rathhun) on the hard clam {Mercenaria mercenaria Linnel. J.
Experimental Mar. Biol. Ecol. 80:207-220,
Barber. B. J.. S. R. Fegley & B. J. McCay. 1988. The Lmie Egg Harbor
hard clam spawner sanctuary: a reproductive evaluation (report). Tren-
ton. New Jersey: The New Jersey Fisheries Development CommLssion.
Carriker. M. R. 1961. Interrelation of functional morphology, behavior,
and autecology in early stages of the bivalve Mercenaria mercenaria.
J. Eli.slm Mitchell Sci. Sac. 77:168-241.
Craig. M. A. & T. J. Bright. 1986. Abundance, age distributions and
growth of Texas hard clam. Mercenaria mercenaria texana in Texas
Bays. Contrih. Mar. Sci. 29:59-72.
Day, E. A. 1987. Substrate type and predatory risk: effects ot mud crah
interaction with juvenile hard clams. State University of New York,
Marine Environmental Science (MSc thesis). Marine Science Research
Center. 1 12 pp.
Kassner. J. & R. Malouf 1982. An evaluation of "spawner transplants" as
a management tool in Long Island's hard clam fishery. ./. Sliellfrsh Res.
2:165-172.
Kassner. J., R. Cerrato & T. Carrano. 1991. Toward understanding and
improving the abundance of quahogs (Mercenaria mercenaria) in east-
ern Great South Bay, New York. In: M. A. Rice, M. Grady & M. L.
Schwartz, editors. Proceedings First Rhode Island Shellfish Confer-
ence. Rhode Island Sea Grant, Kingston, RI. pp 69-78.
Kraeuter, J. N. 2001. Predators and predation. In: J. N. Kraeuler & M.
Castagna, editors. Biology of the hard clam. Developments in Aqua-
culture and Fisheries Science, pp. 441-589.
Kraeuter. J. N, & M. Castagna. 1977. An analysis of gravel, pens, crab
traps and current battles as protection for juvenile hard clams. Merce-
naria mercenaria. Proc. World Marie. .Soc. 8:581-585.
Kraeuter. J. N. & M. Castagna. 1985. The effect of clam size, net size and
poisoned bait treatments on survival of hard clam. Mercenaria merce-
naria. seed in field plots. J. World Marie. Soc. 16:337-385.
Kraeuter, J. N. & M. Castagna. 1989. Factors affecting the growth and
.survival of clam seed planted in the natural environment. In: J. J. Manzi
& M. Castagna. editors. Clam mariculture in North America. Devel-
opments in .'\quaculture and Fisheries Science, vol. 19. New York:
Elsevier Science, pp. 149-165.
MacKenzie. C. L. 1977. Predation on hard clam Mercenaria mercenaria
populations. Trans. Amer. Fish. Soc. 106:530-536.
MacKenzie. C. L. 1979. Management for increasing clam abundance. Mar.
Fish. Rev. 1979:10-22.
Manzi. J. J. & M. Castagna. (editors.) 1989. Clam mariculture in North
America. Developments in Aquaculture and Fisheries Science. Vol. 19.
New York: Elsevier Science 461 pp.
McHugh. J. L. 1991. The hard clam fishery past and present. In: J. R.
Shuhel. T. M. Bell & H. H. Carter, editors. The Great South Bay.
Albany: State University of New York Press. 107 pp.
Papa. S. T. 1994. Distribution and abundance of the hard clam in relation
to environmental characteristics in Great South Bay New York (MSc
Thesis). Stony Brook: MSRC SUNY. 147 pp.
Parker. K. M. 1975. A study of natural recruitment of Mercenaria merce-
naria. Report to North Carolina Division of Marine Fisheries. Wrights-
ville Beach. North Carolina.
Pratt, D. M. 1953. Abundance and growth of Venus mercenaria and Cal-
locardia morrhuana in relation to the character of bottom sediments. /
Mar. Res. 12:60-74.
Paulsen. R. & P. Murray. 1987. Test of hard clam seed survival as affected
by subsurface planting. Final report the New York State Urban Devel-
opment Corporation. AquacuUure Innovation Program. 31 pp.
Peterson. C. H., H. C. Summerson & J. Huber. 1995. Replenishment of
hard clam stocks using hatchery seed: combined importance of bottom
type, seed size, planting season, and density. / Shellfish. Res. 14:293-
300.
Saila, S. B.. J. M. Flowers & M. T. Cannario. 1967. Factors affecting the
relative abundance of Mercenaria mercenaria in the Providence River.
Rhode Island. Proc. Natl. Shellfish. Assoc. 57:83-89.
Walker. R. L. and K R. Tenore. 1984. The distribution and production of
the hard clam. Mercenaria mercenaria in Wassaw Sound. Georgia.
Estuaries 7:19-27.
Wells. H. W. 1957. Abundance of the hard clam Mercenaria mercenaria in
relation to environmental factors. Ecology 38:123-128.
Jouniiil I,] Slicllfhh Research. Vol. 22, No. I, 69-7.^, 2()().V
SPATIAL VARIATION IN THK BODY MASS OF THE STOUT RAZOR CLAM,
TAGELUS PLEBEIUS: DOES THE DENSITY OF BURROWING CRABS,
CHASMAGNATHUS GRANULATA, MATTER?
JORGE L. GUTIERREZ* AND OSCAR O. IRIBARNE
Deparhimento cle Biologia. FCEyN, Uuivcisidad Nacinncd de Mar del Plata. CC 573.
B76UUWAG Mar del Plata. Arqeiitina
ABSTR.ACT A series of functional-group hypotheses proposed for marine soft-sediment systems predict that either deposit-feeders
or hijihly mobile bioturbators exclude low-mobile supension feeders because of their sediment reworking activity. However, a
low-mobile suspen.sion-feeder — the stout razor clam Tagelus plebeius — coexists with highly mobile deposit-feeding burrowing crabs.
CImsmagnalhus gramduta, in several Southwestern Atlantic estuaries. In this study, we compared the body mass (as relationship
between shell length and dry weight of flesh) of the stout razor clam between replicated patches showing contrasting densities of
burrowing crabs. Spatial variation was observed in the slope of the relationship between shell length and dry weight of tlesh of T.
picbeiiii in the three samplmgs dates (July 1999, January 2(J00. and April 2000). However, the pattern of spatial variation in the slope
of this relationship was not consistent with the pattern of spatial variation in crab density. In addition, the pattern of spatial variation
in the slope of the relationship between shell length and dry weight of tlesh of the stout razor clams was not consistent between the
three sampling dates. These results suggest either that ( 1 ) body mass of the stout razor clam is affected by habitat features other than
crab density, or (2) effects of burrowing crabs on body mass of the stout razor clam are masked by spatial variation in other habitat
features that affect body mass of stout razor clams or the extent to which crabs are able to affect clams.
KEY WORDS: bioturbation. body mass, Cluisnuignailnis Kiuiudata. spatial variation, Tagelus plebeius
INTRODUCTION
The stdLit ra/or clam Tagelus plebeius Soiaiider (Veneroida:
Solecurtidae) is an euryhaline species that occurs in estuarine en-
vironments from North Carolina (34°N. United States) to the San
Matias Gulf (41' S, Argentina; see Holland & Dean 1977a, 1977b,
Viegas 1981. Gutierrez & Iribame 1998. 1999. Gutietrez & Valero
2001 1, This is a suspension-feeding species that construct perma-
nent burrows (up to 50 cm deep) lacking lateral mobility (Holland
& Dean 1977a. 1977b. Gutieirez & Valero 2001), In several
Southwestern Atlantic estuaries, this species coexists with the bur-
rowing grapsid crab Cluisimignathus gramduta Dana (Gutierrez &
Iribame 1998. Gutierrez & Valero 2001), C, granulata is one of
the dominant macroinvertebrates in tidal flats and salt marshes of
Southwestern Atlantic estuaries from Rio de Janeiro {23°S. Brazil)
to the San Mati'as Gulf (41' S. Argentina; Bo.schi 1964. Spivak et
al. 1994. Iribame et al, 1997), This is a gregarious species that
excavate and maintain semipermanent open burrows in the inter-
tidai, from soft bare sediment flats to areas vegetated by the
cordgrass Spuriina densiflora (Spivak et al. 1994. Iribarne et al.
1997). At sediment Hat areas, individuals of C. gruiuilata behave
as deposit-feeders, showing large (up to 1.4 1 volume) and mobile
burrows (up to 5 cm day"'; Iribame et al. 1997),
Coexistence between these two species, however, must not be
expected according to any of the functional-group hypotheses that
were proposed to predict species assembly in soft-substrate envi-
ronments. For instance, the trophic-group amensalism hypothesis
(Rhoads & Young 1970) predicts that deposit-feedeis, such as
Chasmagnalhus granulata. exclude suspension feeders, such as
Tagelus plebeius. by increasing the amount of sediment resus-
pended in the water column, which clogs the filtering appendages
of suspension-feeders. The adult-larval interaction hypothesis
(Woodin 1976) predicts that sediment reworking by deposit-
feeders kill the larvae of recently settled suspension-feeders be-
*Corresponding author. E-mail: jlgutie@mdp.edu. ar
cause of direct damage or burial to unsuitable depths. The mobil-
ity-mode hypothesis (Brenchley 1981. 1982) proposes that mobile
benthic species, such as C. gramdata. exclude more sedentary
forms, such as T. plebeius. by continually burrowing trough the
sediment. The coexistence between C. granulata and T. plebeius,
however, illustrates that sedentary suspension-feeders are not al-
ways excluded from areas inhabited by mobile burrowing deposit-
feeders. In fact, the latter is not a novelty: much evidence support-
ing the occurrence of the mechanisms predicted by the functional-
group hypotheses often refer to negative but non-lethal effects (see
Posey 1989 for a review). Therefore, regardless the lack of exclu-
sion between both species, we are still in conditions to expect for
negative, but nonlethal effects of C. granulata on stout razor
clams.
The patchy distribution of bun'owing crabs in the tidal flats of
several Southwestern Atlantic estuaries (see Botto & Iribarne
1999. 2000) provides a good opportunity to explore this possibility
at a realistic scale. In this study, we compare the body mass (the
relationship between dry weight of tlesh and shell length) of the
stout razor clam in patches with high and low density of burrowing
crabs. We recognize that this comparative approach does not allow
to address cause-effect relationships between the presence of crabs
and the body mass of stout razor clams, but comparing the body
mass of stout razor clatns among replicated areas with high and
low density of burrowing crabs allow to discern between the fol-
lowing logical possibilities:
( 1 ) The body mass of the stout razor clam vary between habi-
tats depending on crab density, which may indicate (a) that
burrowing crabs affect body mass of the stout razor clam
or. (b) that the habitat features that affect crab density also
affect body mass of stout razor clams.
(2) The body mass of the stout razor clam vary between habi-
tats but irrespective of crab density, which may indicate (a)
that body mass of the stout razor clam is affected by habitat
features other than crab density, or (b) that effects of bur-
rowing crabs on body mass of the stout razor clam are
69
70
Gutierrez and Iribarne
T.\BLE 1.
Mean (SD) density (burrows m~-) of burroHing crabs Chasinagnalliiis granulata in the locations under study and results of one way ANOVA
(df = 114) evaluating differences in crab density between locations.
Location
ANOVA
Sampling Date
1
2
3
4
5
6
MS
F
July 1999
January 2000
April 2000
l..\5 (0.81)
3.15 (1.35)
1.70(0.86)
1,.^0 (0.86)
3.35 (1.50)
1.60(0.99)
1.40 (0.75)
3.60 (1.43)
1.80(0.83)
0.15 (0.37)
0.45 (0.51)
0.60 (0.60)
0.25 (0.44)
0.50(0.61)
0.45 (0.51)
0.20 (0.52)
0.60 (0.50)
0.55 (0.51)
2.64
6.85
1.97
24.36*
53.87*
14.68*
' P < 0.01. Tiikey tests: (1 = 2 = 3) * (4 = 5 = 6) in all sampling dates
overwhelmed by spatial variation on other habitat features
that affect body mass of stout razor clams or the ability of
crabs to affect clams.
(3) The body inass of the stout razor clam did not vary between
habitats, which may indicate (a) that burrowing crabs does
not affect body mass of stout razor clams, or (b) that effects
of burrowing crabs are being compensated by spatial varia-
tion in other habitat features that affect body mass of the
stout razor clam.
MATERIALS AND METHODS
This study was conducted at the Mar Chiquita coastal lagoon
(37°S. Argentina), which is a 46-km" body of brackish water af-
fected by semidiurnal low amplitude (<l m) tides and character-
ized by mudflats and large surrounding marshes dominated by the
halophyte Spartina densiflora (Spivak et al. 1994, Iribarne et al.
1997). Samplings for crab density and collections of stout razor
clams were conducted in July 1999 and January and April 2000. in
an area approximately located 2.5 km upstream from the lagoon
inlet, which comprises about 700 m of shoreline. At this area, six
locations were selected; three of them characterized by high bur-
row densities of Chasina\(iiatluis f;raiu(lata (locations I. 2. and 3).
and the others by very low bun'ow densities (locations 4. .5. and 6).
Crab density at each location was estimated by random sampling
using a 1 X 1 m sampling unit (;; = 20). Single factor analysis of
variance followed by Tukey test (Zar 1984) was used to test for
differences between locations in the density of burrowing crabs.
Locations grouped under the same level of crab density did not
differed significantly in the density of crab burrows in all sampling
dates (see Results and Table I). Sixty clams per location were
collected at each sampling date by excavating the sediment using
hand shovels. The length of the clams was measured along the
anterior-posterior axis to the nearest 0.01 mm and their flesh was
removed from the gaping shell after a short immersion in boiling
water. The flesh was dried separately at 70°C for 48 h before their
dry weight was determined. Correlation analysis (Zar 1984) was
used to evaluate the existence of a significant relationship between
shell length and dry weight of flesh in clams at each location and
sampling date. Once significant relationships between the shell
length and the dry weight of flesh of the clams were observed at all
locations and sampling dates (see Results and Table 2), parallelism
tests followed by Tukey tests (Zar 1984) were used to compare the
slope of this relationship between locations at each sampling date.
Gi\ en that clams smaller than .50 mm occurred in low numbers and
not in all locations, we excluded these data from the analysis of
correlation and parallelism to cover the same range of sizes in all
locations. After removing these data, we also randomly discarded
some data from clams larger than 50 mm to attain an equal sample
size between locations (July 1999; /; = 57; January 2000. n = 56.
.A.pril 2000. n = 52).
RESULTS
Single-factor analysis of variance indicated that the density of
burrowing crabs significantly differed between locations in the
three sampling dates (Table I ). Tukey tests revealed that the six
locations can be subdivided in two clearly defined groups: loca-
tions with relatively high crab density (locations 1. 2. and 3) and
locations with low crab density (locations 4. 5. and 6): being this
pattern consistent in the three sampling dates irrespective of tem-
poral variations in the density of crab burrows (Table 1). Corre-
lation analysis indicated a significant linear relationship between
the dry weight of tlesh of stout razor clains larger than 50 mm and
their shell length in all locations and sampling dates (Table 2).
Parallelism tests indicated that the slope of the relationship be-
tween shell length and dry weight of flesh of T. pleheiiis differed
TABLE 2.
Site-specific regression equations and determination coefficients (between brackets) observed for the relationship between dry weight of flesh
and shell length of the stout razor clam Tageiiis pleheiiis in the three sampling dates.
Location
Julv 1999
January 2000
April 2000
y = 0.019X-0.467 (0.386)
y = 0.030X-1.135 (0.562)
y = 0.027x-^.897 (0.361)
y = 0.026X-0.927 (0.446)
y = 0.013X-0.286 (0.285)
v = 0.032X-1.169 (0.548)
y = 0.026X-0.739 (0.383)
y = 0.034.X-1.191 (0.282)
y = 0.014X-0.126 (0.182)
y = 0.014X-0.082 (0.086)
y = 0.033X- 1.088 (0.331)
y = 0.018X-0.458 (0.265)
y = 0.()21.x-0.509 (0.383)
y = 0.019X-0.427 (0.244)
y = 0.028.x- 1.024 (0.540)
y = 0.016x-0.3(.17 (0.333)
y = 0.033x^1. 144 (0.317)
V = 0.O26X-O.870 (0.416)
P < 0.05 in all cases.
Spatial Variation in Bod\- Mass of Tagelus flebeius
71
significantly between locations in the three sampling dates (Table pattern of spatial variation in crab density in any of the sampling
3, Fig. 1 ). Tukey multiple comparison of slopes indicated that the dates (Table 3).
patterns of spatial variation in the slope of the relationship between
shell length and dry weight of tlesh of T. pkbeius was not con-
sistent between sampling dates. In addition, spatial variation in the Recalling the logical possibilities established in the introduc-
slope of the dry weight-shell length relationship did not match the tion. our results suggest either ( 1 ) that body mass of the stout razor
DISCUSSION
3
h-
O
LU
>-
Q
15^ LOCATIOIMS
♦ 1 a2 m^ o4 a5 o6
0.5
July 1999
1
1-5 1
0.5
1.5
0.5
45
»
0.8
0.6
0.4
0.2
45 50 55 60 65 70 45
January 2000
«"o
1.2
1
0.8
0.6
0.4
0.2
45 50 55 60 65 70 45
April 2000
1.1
0.9
0.7
0.5
0.3
50 55 60 65
70
45
50 55
60
65 70
50 55
60 65
50
55
60
SHELL LENGTH (mm)
70
5C
65 70
Figure 1. Relationship between dry weight (if flesh and shell length of the stout razor elani Tardus pkbeius at each location and sampling date.
Left: Dry weight data points plotted against shell length. Right: Cur\es corresponding to the linear fit of data points in the figures on the left.
Locations are denoted by numbers beside each curve (locations 1, 2, and 3: High crab density: locations: 4, 5, and 6: low crab density). Curves
showing the same letter beside their respective location numbers have slopes that are not significant different [P > 0.(15) after Parallelism tests
followed by Tukev tests.
72
Gutierrez and Iribarne
TABLE .V
Results of tests for parallelism and Tukey tests used to e\aluate
differences between locations in the slope of tlie relationship
between dry weight of flesh and shell length of Tageliis pleheius
Sampling
Parallelism Test
Date
MS
F
Tukey Test
July 1999
0.015
24.12(5*
(1,2,4.6) (3.6) (5)
January 2000
0.021
20.418*
(1.3.4) (3.4.6) (1.2) (2.5)
April 2000
0.014
14.194*
(2.3.4,6) (1.2.4) (5l
Numbers between brackets indicate locations that did not significantly
differed in the slope of the dry weight-shell length relationship after Tukey
tests.
*P<0.01.
clam is affected by habitat features other than crab density, or (2)
that effects of burrowing crabs on body mass of the stout razor
clam are masked by spatial variation in other habitat features that
affect body mass of stout razor clams or the extent to which crabs
are able to affect clams. This is reasonable to occur because the
locations encompassed in this study differ in many features -as
sediment characteristics and orientation-iirespective of the pres-
ence of crabs (personal observation). Sediment characteristics rnay
directly affect clam body mass (e.g., by determining the costs of
burrowing: see Swan 1952. Newell & Hidu 1982) as well as the
nature and extent of habitat tnodifications derived from crab bur-
rowing that may be detritiiental for stout razor clams (e.g., sedi-
ment resuspension; see Turner & Miller 1991). Differences in the
orientation of the locations in relation to winds determine, for
example, the degree to which clams are exposed to events of
environmental disturbance by waves and cun'ents (see Turner &
Miller 1991, Bock & Miller 1995) as well as the degree to which
sediment reworking by crabs might be overwhelmed or not by
physical reworking (see Grant 1983).
The overall conclusion of this study is that crabs alone do not
promote a spatial pattern in body inass of the stout razor clam at
the scale of crab patches. It is uncertain, however, whether effects
of crabs on the body mass of stout razor clams are occurring at
locations with high density of crabs but overwhelmed by other
sources of spatial variation that affect clams. Several lines of evi-
dence suggest that crabs might have iinportant local effects on the
body mass of stout razor clams. For instance, organisms that are
known to exclude low-mobile suspension-feeders, such as calli-
anassid shrinips (see Posey 1989) excavate sediments at rates of
2.7-3.5 kg (dry) nr- d"' (Vaugelas 1984. Swinbanks & Luter-
nauer 1987. Witbaard & Duineveld 1989). whereas burrowing
crabs excavate sediments even at higher rates [5.9 kg (dry) m""
d"': Iribarne et al. 1997]. Consequently, the detrimental effects of
sediment reworking by crabs on the stout razor clam predicted by
the functional-group hypotheses are still possible.
However, considering the rates at which burrowing crabs and
callianassid shrimps remove sediments, the question at this point is
why burrowing crabs does not exclude stout razor clams as calli-
anassid shrimps do with a variety of suspension-feeders. The an-
swer is. perhaps, in the different modes by which callianassid
shrimps and burrowing crabs rework sediments. Callianassid
shrimps burrow and sift the sediments continuously for food, de-
stabilizing them and increasing water turbidity (Aller & Dodge
1974. Murphy 1985). Ht)wever, C. granulata reworks sediments
mostly during low tide eventually depositing mounds of fine, co-
hesive sediment above the surface, which are not likely to be easily
resuspended by tidal cuirents (e.g., Iribarne et al. 1997, Botto &
Iribarne 2000), This implies that some mechanisms predicted to
exclude suspension-feeders from areas dominated by deposit-
feeders, such as sediment resuspension (see Rhoads & Young
1970) might not take place in the case of buiTOwing crabs. Further,
the latter suggests that sediment reworking is not a good predictor
of the actual effect of burrowing deposit-feeders on suspension
feeders.
ACKNOWLEDGMENTS
This project was supported by grants from Universidad Nacio-
nal de Mar del Plata. CONICET. FONDECyT. and Fundacion
Antorchas. J.L.G. is supported by scholarships from CONICET
and this article is part of his Doctoral thesis.
LITERATURE CITED
Aller. R. C. & R. E. Dodge. 1974. Animal-sediment relations in a tropical
lagoon. Discovery Bay. Jamaica. / Mar. Res. 32:209-232.
Bock, M. J. & D. C. Miller. 1995. Storm effects on particulate food re-
sources on an intertidal sandflat. J. E.xp. Mar. Biol. Ecol. 187:81-101.
Boschi. E. E. 1964. Los Crustaceos Decapodos Brachyura del litoral bo-
naerense (R. Argentina). Bol. Inst. Biol. Mar. Mar del Plata 6:1-99.
Botto. F. & O. O. Iribarne. 1999. The effect of the burrowing crab Cluis-
magnathus granulata on the benthic community of a Southwestern
Atlantic lagoon. J. Exp. Mar. Biol. Ecol. 241:263-284.
Botto. F. & O. O. Iribarne. 2000. Contrasting effects of two burrowing
crabs {Chasmagnatlms granulata and ilea uniguayensis) on sediment
composition and transport in estuarine environments. Est. Coast. Shelf
Sci. 51:141-151.
Brenchley, G. A. 1981. Disturbance and community structure: an experi-
mental study of bioturbation in marine soft-bottom communities. /
Mar. Res. 39:767-790.
Brenchley. G. A. 1982. Mechanisms of spatial competition in marine soft-
bottom communities. J. E.vp. Mar. Biol. Ecol. 60:17-33.
Grant, J. 1983. The relative magnitude of biological and physical sediment
reworking in an intertidal community. ./. Mar. Res. 41:673-689.
Gutierrez. J. L. & O. O. Iribarne. 1998. The occurrence of juveniles of the
grapsid crab Chasmagnatlms granulata in siphon holes of the stout
razor clam Tagelus pleheius. J. Shellfish. Res. 17:925-929.
Gutierrez. J. L. & O. O. Iribarne. 1999. Role of Holocene beds of the stout
razor clam Tagelus pleheius in structuring present benthic communi-
ties. Mar. Ecol. Prog. Ser. 185:213-228.
Gutierrez. J. L. & J. L. Valero. 2001. La almeja navaja y su partlcipacion
en mecanismos ecologicos de comunidades intermareales mediante la
produccion de valvas. In: O. O. Iribarne. editor. Reserva de Biosfera
Mar Chiquita: Caracteristicas fi'sicas. biologicas y ecologicas. Mar del
Plata, Argentina: Editorial Martin, pp. 121-128
Holland. A. F. & J. Dean. 1977a. The biology of the stout razor clam
Tagelus pleheius. 1. Animal-sediment relationships, feeding mecha-
nism and community biology. Chesapeake Sci. 18:58-66.
Holland. A. F. & J. Dean. 1977b. The biology of the stout razor clam
Tagelus pleheius. 2. Some aspects of the population dynamics. Chesa-
peake Sci. 18:188-196.
Iribarne. O. O.. A. Bortolus & F. Botto. 1997. Between-habitat differences
in bunow characteristics and trophic modes in the southwestern At-
lantic burrowing crab Chasinagnathus granulata. Mar. Ecol. Prog. Ser.
155:137-145.
Murphy, R. C. 1985. Factors affecting the distribution of the introduced
Spatial Variation in Body Mass of Tagelus plehfjus
73
bivalve, Mercenaria mercenaria, in a California lagdon The impor-
tance of bioturbation. J. Mar. Res. 43:673-692.
Newell. C. R. & H. Hidu. 1982. The effect of sediment type on growth rate
and shell allometry in the soft-shelled clam Mya aremirki L. J. E.xp.
Mar. Biol. Ecol. 65:285-295.
Posey. M. H. 1989. Functional approaches to soft substrate coninmiiities:
how useful are they? Rev. Aqiiat. Sci. 4:2-4-41.
Rhoads. D. C. & D. K. Young. 1970. The influence of deposit-feeding
organisms on sediment stability and eonimunils trophic structure. J.
Mar. Res. 28:150-178.
Spivak, E., K. Anger, T. Luppi. C. Bas & D. Ismael. 1994. Distribution and
habitat preferences of two grapsid crab species in Mar Chiquita lagoon
lPro\ince of Buenos Aires. Argentina). Helaolaiuler Meeresimlcrs 48:
59-78.
Swan. E. F. 1952. The growth of the clam M\a arcinina as affected by the
substratum. Ecology 33:530-534.
Swinbanks. D. D. & J. L. Lutemauer. 1987. Burrow distribution of
thalassinidean shrimp on a Fraser delta tidal Hat, British Columbia. J.
Paleoinol. 61:315-332.
Turner, E. J. & D. C. Miller. 1991. Behavior and growth of Mercenaria
mercenaria during simulated storms events. Mar. Biol. 1 1 1:55-64.
Vaugelas, J. V. 1984. Preliminary observations on two types of calli-
anassids (Crustacea: Thalassinidea) burrows. Gulf of Aqaba. Red Sea.
Proc. 1st Int. Symp. Coral Reef Environ. Red Sea 1:520-539
V'legas. O. 1981. Dinamica populacional e produ^ao de Tagelus plebeiiis
no canal do Calunga. Maceio-Alagoas. M.Sc. thesis. Departamento de
Biologia Vegetal, Universidade de Brasilia, 86 pp.
Witbaard. R. & G. C. A. Duineveld. 1989. Some aspect of the biology and
ecology of the burrowing shrimp Callianassa siibterranea (Montagu)
(Thalassinidea) from the southern North Sea. Sarsia 74:209-219.
Woodin. S. A. 1976. Adult larval interactions in dense infaunal assem-
blages: patterns of abundance. J. Mar. Res. 34:25^1.
Zar. J. H. 1984. Biostatistical analysis. Englewood Cliffs. NJ: Prentice-
Hall. 718 pp.
Jmiriuil III Shellfish Research. Vol. 22, No. I. 75-X.^. 2()().^,
MARICULTURE SITING— TIDAL CURRENTS AND GROWTH OF MYA ARENARIA
WILLIAM R. CONGLETON, JR..' BRYAN R. PEARCE," MATTHEW R. PARKER,' AND
ROBERT C. CAUSEY'
DcjHiriiui'nt of Animal ami Veterinary Science. Univer.sin of Maine. Orono. Maine 04469 'Depanment
of Civil ami Enviroiuiiental Enginecrini>. Univer.'iity of Maine. Orono. Maine 04469
ABSTRACT Mariculture of the soft-shell clani Myii uienaria L. involves seeding juvenile shellfish on nitertidal niudtlats for
grow-out. Laborator>- studies have shown that constant current velocity affects shellfish growth. Few studies have determined the effect
of tidal currents on shellfish growth in siiii. Spot estimates of tidal currents can be generated with portable current meters and by
measuring the erosion of Plaster of Paris hemispheres called clod cards placed in the current. Current velocities for Geographical
Information System (CIS) coverages for entire estuaries can be estimated using numerical flow models. Although these different types
of measurement have different relative advantages of cost, ease of describing large areas, and accuracy, each can be potentially used
in evaluating sites for shellfish grow-out. Current velocities averaged over the flood tide were estimated by a numeiical flow model
and by clod cards for 16 locations at the same elevation in a bay in Eastern Maine and were compared with the annual shell increment
of clams collected at the same locations. Statistical models included main effects and interactions between initial shell size, year of
sample, and high-low current category estimated by clod cards or a numerical model. Models explained 57-58% of the variability in
growth increment with initial shell size and year affecting growth more than current. Faster tidal currents resulted in 22-24% greater
shell growth. Sites categorized as low flow had means for tidal currents {±SD) of 4.35 ± 0.37 cm/s and 2.99 ± 0,43 cm/s using the
numerical model and clod cards, respectively. Least squares means (±SE) for the annual increment in shell length increment was 9,56
+ 0.247 mm for the low flow sites identified using the numerical model and 9.5 1 ± 0.274 mm for the low flow sites idenfified using
clod cards. Sites categorized as high flow had current means (±SD) of 5.86 ± .62 cm/s using clod cards and 5.84 ± 0.46 cm/s using
the numerical model and least squares means (±SE) for growth increment of 1 1.90 ± 0.32 and 1 1.70 ± 0.33 mm, respectively. The
stimulatory effect of tidal currents on clam growth could be used in mariculture siting. Placing clod cards at specific intertidal locations
at the same elevation could be used to estimate relative current velocities. Current velocities estimated using numerical models and
displayed as CIS grids of entire regions will not have the same resolution as spot estimates from current meters or clod cards. However,
grids can be used for siting if the grid cells are comparable in si/e to area to be seeded.
KEY WORDS: numerical model. Geographical Information System (GIS), current, growth, Mya iireiuirui
INTRODUCTION
Seed planting and transplanting has been an integral part of the
hard clam and oyster industries (Malouf 1989). With hundreds of
miles of mudflats in Northeastern Maine and a 457f decline in state
landings over the past 13 y (DMR, 1997). mudflats with low
densities oi Mya arenaria L, are being seeded with juvenile clams.
Site-specific characteristics must be evaluated in selecting sites for
shellfish seeding (Beal et al. 2001. Peterson et al, 1995, Newell
1996). but determining environmental parameters capable of sus-
taining populations of bivalve seed is difficult in most cases (Mal-
ouf 1989).
Among a variety of biologic and environmental that influence
growth of bivalves in situ, sufficient current speed is recognized as
an important factor. Water velocity, horizonlal adveclion, and ver-
tical mixing in the water column influence the availability of phy-
toplankton to mussels (Frechette et al. 1989), Currents are needed
to avoid depletion of oxygen and food particles to suspension
feeders, especially at high-density levels (Jorgensen 1990). Newell
(1990) suggested a minimum current speed (about ,3 cni/s) below
which bottom culture of mussels may not be cost effective. An
actual reduction in food intake of bivalves was found when current
rates are not kept high enough (Bayne et al. 1976). Faster flow
results in a greater flux of organic particles (Peterson & Skilleter
1994). Shell growth rates for hard clams over a 15 wk period
increased by 10.7% in fast relative to slow current sites in coastal
lagoon in New Jersey (Grizzle & Morin 1989). Soft-shell clams
were found to orient perpendicular to the principal component of
current direction potentially to optimization energy acquisition
during an entire tidal cycle (Vincent et al, 1988),
The effect of water flow on growth varies with species of
bivalve. For infauna, northern quahogs displayed a consistent in-
crease in shell growth with higher flow speed in the range of
stream velocities between to 4 cm/sec (Grizzle et al. 1994).
Growth response in the soft-shell clams was similar to that ob-
served in hard clams with a proportional increase in shell length
for 4 y old, 40 mm clams with flow and no evidence of growth
inhibition between free-streatn velocities of 0.1 to 5.8 cin/sec (Em-
erson 1990),
For epifauna species, it has been speculated that growth is
maximized at water flows that match inhalant pumping speed.
Mussels grown in multiple flume trials at flow velocities of 0. I. 2.
4. and 8 cm/sec had a statistically nonsignificant increase in shell
growth at a flow of 2 cm/sec. which matched the approximate
inhalant pumping speed (Grizzle et al, 1994). Eastern oysters in-
creased growth at a flow of 1 cm/sec relative to tlows of and
>1 cm (Grizzle et al. 1992). The constant flow in flume studies,
however, is different from tidal currents, which vary in magnitude
and direction. Flume experiments with ascending and descending
flows have found clearing or grazing rates of scallops differed by
307f (Pilditch & Grant 1999).
Currents may affect shellfish growth, but estimating current
velocities can be difficult. A device coinmonly used to determine
flow rates is a current meter. However, collecting time series ve-
locity profiles with current meters over large areas is time con-
suming with conventional instrumentation, particularly in inter-
tidal waters. When current rates and flow patterns are needed for
large regions being considered as potential shellfish grow-out sites,
the use of current meters becomes impractical.
Two- and three-dimensional numerical computer models can
be used to describe the direction and magnitude of currents for
individual cells in grids covering coastal areas. The output data
from numerical models can then be used to create thematic maps
for Geographical Information System (GIS) coverages (Congleton
75
76
CONGLETON ET AL.
et al. 1999). Numerical models are supported by data for bottom
elevations for each cell in the grid and tidal amplitude at the ocean
boundaries of the models. They simulate time series estimates of
velocity vectors for grid cells covering the model domain. Veloc-
ities may be estimated for discrete layers in individual grid cells or
may be vertically averaged, as in this study. Model output can be
analyzed in the GIS to identify sites with optimum conditions for
shellfish growth. The major drawback, however, is the difficulty of
initializing and running a numerical model.
An alternative method for estimating currents is by measuring
a process, which is affected by the current magnitude. A physical
analog measurement of current velocity is the dissolution of cal-
cium sulfate (Plaster of Paris or gypsum) blocks or hemispheres,
called clod cards, placed in moving water (Muus 1968. Doty 1971,
Peterson & Skilleter 1994). Thompson and Glenn (1994) devel-
oped an equation for calculating mean water speed from field
deployed clod cards using clod cards from the same batch for
laboratory calibration in quiescent water of the same salinity tem-
perature as in the field. They concluded that proper execution of
field and calibration tests result in a simple and practical method
for measuring water motion over a wide range of temperatures,
salinities, and current speeds. Clod cards are inexpensive and
simple to construct, but the difficulty of deploying large numbers
limits their usefulness for estimating cunent magnitudes o\ er large
areas.
The objective of this study is to evaluate the relationship be-
tween ( 1 ) field measurements of tidal currents made with clod
cards; (2) average current estimates generated by a numerical flow
model; and (3) growth of soft-shell clams on a mudfiat in Eastern
Maine. The appropriateness of incorporating current estimates
from a numerical model into a GIS for the selection of sites for
grow-out of juvenile shellfish will then be considered.
METHODS
The study was conducted in Mason Bay in Eastern Maine on
the western side of Englishman Bay, which bounds the Gulf of
Maine. The bay (Fig. I A) is 2.39 km long by 1.03 km wide,
oriented in an east-west direction, and is located 9.7 km north of
Jonesport, Maine (44°61.80'N, 67°56.23"W). At low tide (mean
low water = -1.875 m nisi), mudflats are exposed along the entire
length of the bay with two channel inlets from Englishman Bay
joining on the west side of Spar Island and running the length of
the bay (Fig. IB). Water temperatures vary from 5°C in April to
I6'=C in September (Beal et al. 2001).
Soft-shell clams were collected at 15 sites at the same water
line spaced 40 m apart to the south of Spar Island and west of
Flake Point Bar (Fig. IB) at an elevation of -2.0 m msl in Spring.
1 996. These sample sites were close to one of the inlets of the bay
with a maximum separation of 485 m from the most easterly site
to the most westerly site. A sixteenth site between the tip of Flake
Point Bar and Spar Island was sampled during spring of 2000 to
increase the range of water velocities sampled. One of the low flow
sites in the center of the earlier sampling array was also sampled
the second year. Sites were relocated in the second year using their
global positioning system (GPS) coordinates.
Location of the 16 sites was determined by caiTier-phase GPS
measurements made with a Trimble GeoExplorer™ GPS receiver,
and post-processed. Carrier-phase GPS is commonly used for sur-
veying with sub-decimeter accuracy for measurements in the ho-
rizon plane. Measurements in the vertical plane are less accurate.
The range in the elevation measurements for the 10-min carrier-
phase GPS readings at the 16 sites was -1.5 to -2.7 m msl with
95% confidence range of ±0.55 m for individual measurements
(Congleton et al. 1999). Because inaccuracy in GPS measurements
alone could have resulted in a difference in elevation between
sites, locations were selected with simultaneous flooding and dry-
ing times.
Sample site coordinates were then imported into the Maplnfo'^'
GIS creating a layer of sampling site locations (Fig. IB ). Sediment
cores from four of the sites were analyzed for composition by the
Analytical Laboratory of the Maine Soil Testing Service using the
hydrometer method for particle size and 1050'"C combustion ana-
lyzer for total carbon. Fifty clams were dug with a clam rake at the
1 5 sites South of Spar Island at the end of the first growing season.
A single low flow site (sixth site counting from the most easterly)
and the high flow site SE of Spar Island were sampled in the
second growing season.
External annual rings were used to determine the increase in
shell length during the preceding summer. Brousseau ( 1979) found
winter rings to be a reliable method of determining age in soft-
shell clams from Gloucester. Massachusetts. However. Mac-
Donald and Thomas ( 1980) found external growth rings to be less
reliable for age determination than thin shell sections, and Lewis
and CeiTato (1997) found shell increment might be temporarily
decoupled from soft-tissue growth by high temperature or starva-
tion. However, external growth rings have been used for long-term
estimation of growth (Kube et al. 1996) and growth and age
(Jacques et al. 1984. Evans & Tallmark 1977) of /;; situ Mya
arenaria.
Because of limitations of using growth rings for measuring age.
length between the last shell check marks were used to measure the
size at the beginning of last growing season. Initial size was then
used as a covariate in the statistical analysis instead of age. Annual
growth increment was then calculated by subtracting the final shell
length from the initial size. The problem of lengthy shell abrasion
limiting the usefulness of external rings in aging was minimized by
taking measurements of growth only in the last growing season.
NUMERICAL MODEL OF TIDAL CURRENTS
Estimated currents for Mason Bay were obtained from the Ma-
son Bay Model (MBM). which is an adaptation of Princeton Ocean
Model (Mellor 1992. Blumberg & Mellor 1987) modified to de-
scribe intertidal areas (Congleton et al. 1999). Input bathymetry
data for the model were processed in the Maplnfo GIS including
sublidal depths from NOAA nautical chart no. 13325, the shoreline
boundary traced frotn an aerial photograph and 27 high accuracy
canier phase GPS measurements made at the waterline near low
water on a single Spring tide. To increase the accuracy of the
description of the bottom in the study area, fourteen of the GPS
measureinents were in the region, which is enlarged in Figure lb.
These data were used to generate a 100 by 76 grid covering the bay
composed of square cells with 36.125 m sides. The 7600 cells in
the grid gave increased resolution of depths between points with
known elevations without unnecessarily increasing computing
time for a run describing a tidal cycle. Grid cells (36 m sides) were
smaller than the distance between the clam sampling locations
(40 m) resulting in a different estimate of current velocity at each
sample site.
The model generated estimates of vertically averaged cuirent
velocities for each grid cell flooded by the tide at one-second
Tidal Currents and Clam Growth
77
O Sample site
+ High flow.
- Low flow i^^°del
H High flowj _
L Low flow*
Water displacement
per minute
-2 3 Depth m msl
Shoreline mean
high water
Figure 1. (A) Location of Mason Bay in Eastern Main near Jonesport. Maine with Englisliman Ba> and the Atlantic Ocean to the east connected
by channels north and south of Dunn Island. Lines and labels show locations and extent of 7.5 niin L'S(;S quadrangles. (Bl Aerial photos of
Mason Bay. Right image is the rectangular area in SE image of the entire Bay. The array of sample locations (-2 m msl) are spaced 40 m apart
except for the site nearest Spar Island. Vectors show water displacement/minute at maximum flood tide.
78
CONGLETON ET AL.
intervals for an average 2 m amplitude tide (Congleton et al. 1999).
A vertical average of the current velocity for each time step was
used because tidal amplitude and shallow water depths would in-
hibit stratification. Bottom friction was proportional to the square
of the veilically averaged bulk flow. Vectors showing the current
magnitude and direction estimated by the model for each grid cell
were imported into the GIS. Layers of cuirent vectors at different
times in a tide cycle described flow throughout the bay.
For the statistical analysis of clam growth, the time series of
velocities were averaged over the flood phase. The layer showing
the sample site locations was placed over a layer of average cuirent
velocities to estimate velocities at each site. Because the sample
locations were not centered on the grid used by the numerical
model, the mean velocity of adjacent grid cells with the same
approximate elevation were averaged.
FIELD MEASUREMENT OF CURRENTS— CLOD CARDS
Plaster of Paris hemispheres (clod cards) were used for mea-
suring relative water motion at each of the fifteen sampling sites.
In previous studies, rectangular clod cards were used (Doty 1971,
Thompson & Glenn 1994). Clod cards used in this study (Fig. 2)
were molded in hemispheric plastic capsules (32.36 cm'), creating
a uniform surface area exposed to the current regardless of card
orientation.
Commercial Plaster of Paris or gypsum was mixed two parts
powder to one part water. The slurry was poured into the capsules
and leveled off with a straightedge and left at room temperature for
a week to insure thorough drying. After attachment to a 9 x 6.5 cm
sheet of plastic with silicone epoxy. initial dry weights for each of
the clod cards were measured and recorded.
For field deployment, the backing sheet of each clod card was
attached to a brick with rubber bands. One clod card was placed at
each of the 16 clam sample sites (Fig. IB), and a total submersion
time was estimated for the period that included air exposure at low
tide. Because all clod cards were deployed on a spring tide in April
(-0.7 m mllw), they were recovered after 4 days while the loca-
tions were still accessible at low tide. After recovery, cards were
lightly rinsed to remove mud and were left to dry at room tem-
perature for one week and weighed. The percentage loss and the
change in weight were calculated.
The calibration of clod cards in quiescent water or under free
convection conditions is necessary for the overall calculation of
integrated field water speed. Four clod cards from the same lot as
those used in the field trial were suspended 5 cm below the surface
of a 22-1 cylindrical container containing seawater (30-32 ppt
salinity). The container was placed in a larger recirculating tank
maintained at 7''C. which corresponded to the average water tem-
perature during the field trial.
Every 24 h, the water inside the container was replaced with
fresh salt water and the dissolved Plaster of Paris on the bottom of
the container discarded. After the calibration period of four days,
each card was dried at room temperature for a week and then
weighed. The average of the initial weights and average of the final
weights for the four calibration cards were used in the water ve-
locity calculations.
The scalar arithmetic mean velocity of the water in the field (V)
was estimated for the 16 sites following the methods of Thompson
and Glenn (1994):
V = 4.31 (W,„„„„/A,„,„,,)"-^(S,„,'-^/S,,„„„„,,„)
(1)
where W,,,,,, ,, is the initial clod card weight of the field deployed
card: Ai,,,,,.,, is the initial exposed surface area: Si,^,,,, and S^^,|,„„on
are calculated as
|i-(W,-„,„/w„„„.„)"-'i/e
(2)
where W,,,,^^, and W„„„_,| are the final and initial weights of the
field and calibration tests, and H is time submerged in the field and
calibration tests.
Theta for the field trial was total time between deployment
and recovery even though the clod cards in the field experienced
air exposure during low tides. During periods of air exposure, field
clod cards remained wet and continued to dissolve. On average,
the cards were subjected to aerial exposure for approximately 1 h
during each low tide.
STATISTICAL ANALYSIS
Tidal velocities for each site were categorized as high or low
using estimates from the clod cards and numerical model. Because
the mean (±SD) velocity estimated using the clod cards (4.96 ±
0.88 cm/s) was higher than the mean estimated by the numerical
model (3.85 ± 1 .34 cm/s), high flow sites were identified as having
flow s greater than 5 cm/s using clod cards and 4 cm/s the numeri-
cal model. Mean flow at the seven high flow sites identified using
clod cards a\eraged 5.87 ± 0.63 cm/s and at the five high flow sites
identified using the numerical model averaged 5.85 ± 0.47 cm/s.
Mean flow at the nine low flow sites identified by clod cards
averaged 4,35 ± 0.37 cm/s and at the 1 1 low flow sites identified
by the numerical model averaged 3.02 ± 0.33 cm/s. High and low
flow means categorized using either clod cards or the numerical
model were statistically different {P < 0.001) using pooled vari-
ance f-tests.
Variability in shell growth increment during the preceding
growth season was analyzed by analysis of variance (ANOVA)
Plastic backing sheet
Figure 2. (Jypsum clod cards constructed from a plastic mold cemented to a 9 x 6.5 cm backing stieet of plastic.
Tidal Currents and Clam Growth
79
using the GLM procedure in SYSTAT v. 10. Shell length at the
beginning of the growing season (distance between the last most
anterior and posterior margins of the last growth check), years of
sampling (1996 and 2000), and water velocity category of either
high or low as indicated by either clod card weight loss or the
numerical model were the independent factors. The initial model
included main effects and all interactions.
y, = B,, + BiX, + B.X, + B^X, + B4.V1.Y, + fijA-iX, + B,,^,^,
+ fiyXiA'.A', + e
Where y, is the annual growth increment; B„ is the intercept; B,X,
are the coefficient and categorical variable for year; fi-,X, are co-
efficient and value for initial size and 5,^", are the coefficient and
categorical current estiinate (H,L) either from clod cards or the
model: B,X,X„ S^X^X, and fi^XiX^X, are coefficients for two and
three way interactions; and e is the random error term.
Terms that were statistically insignificant iP > 0.05) were de-
leted from the model using the backward elimination procedure
(Draper & Smith 1466).
To ensure independence of residual errors in predicting growth
increment of spatially proximate observations, the Durbin- Watson
Test Statistic was used to test for the existence of autocorrelated
errors. Because 50 clams were collected at each location, residual
eiTor terms remaining after fitting the GLM model might not be
independent if there is a site effect independent of local cun-ents.
First-order autocorrelation (lag = 1 ) results in the error term con-
sisting of a fraction of the previous error term plus a new random
disturbance tenn (Neter et al. 1996). Error terms are uncorrelated
only at the time the autocorrelation term (p) is statistically equal to
zero.
RESULTS
Composition of the sediments at the 16 sites ranged from 47-
55% sand. 29-tl<7f silt. 12-16% clay and 1.15 to 1.27% carbon.
Shell lengths at the start of the two years ranged from 6.9 mm to
55.7 mm. with average (±SD) of 23.1 ± 10.8 mm. After excluding
the juveniles or individuals without a growth check mark, sample
size was 724 with an average (±SD) growth increment of 9.4 ± 4,0
mm with clams sampled once.
Trends in estimated water velocities from the numerical model
and clod cards were similar. Velocities estimated with clod cards
were highest at the site nearest Spar Island and at the sites near
Flake Piiint Bar. Velocities decreased at the sites near the center
and increased at the western end of the cove (Fig. IB). Estimates
from the numerical model displayed a similar trend generally de-
creasing moving westward from Flake Bar. but without the in-
crease at the most western locations.
The correlation coefficient between the 16 estimates of current
velocity from the numerical model and Eq. I was 0.74 (P < 0.05).
Velocity averages over the flood tide at the sixteen sites ranged
from 2.2 cm/sec to 7.14 cm/sec as estimated by the numerical
model and ranged from 3.8 cm/sec to 7.52 cm/sec as estimated by
Eq. I. The estimation of current velocities by a numerical com-
puter model and Eq. 1 were similar although the estimates from the
numerical model were lower. The velocities estimated by the clod
cards were during a spring tide, which would be expected to be
higher than the velocities predicted by the numerical model during
an average tide. Clod card measurements, however, were near the
bottom where velocities are decreased by bottom shear.
The maximum water speed on the flood tide was also estimated
for the sixteen sites. Maximum velocities predicted by the numeri-
cal model ranged from 4.0 cm/sec for some of the western and
central sites to 21.4 cm/sec at the site closest to Flake Point Bar.
All linear models used growth increment as the dependent vari-
able, year ( 1996 vs. 2000) and flow (high vs. low as categorized by
either clod cards or the model) as categorical variables and in-
cluded initial size as a continuous variable. The Durban-Watson
Statistic indicated that the GLM models had statistically signifi-
cant first order autocorrelations. An inspection of autocorrelation
plots of correlation versus lag indicated significant but diminishing
positive autocorrelations up to lag 10 (Fig. 3). Autoconelation
significance (P < 0.05) was deterniined from the 95% confidence
interval for the sampling distribution of the autocorrelation of lag
k or i-f.. which is normal with (x^^ = and a^^- = l/n"" with a
sample size of n (Lin et al. 1995).
A difference transformation replaced values for the dependent
variable (growth increment) with the difference between it and the
preceding value. Differencing is a popular and effective method of
removing trend from spatial (location effect) and time series (tem-
poral effect) data. Autoconelation plots following the transforma-
tion had no trend because as lag increased there was a random
distribution of positive and negative autocorrelations (Fig. 3). To
ensure validity of significance tests using the transformed data, a
linear regression with a hierarchical layout with clams (or trial)
nested or stacked within site was used. The trial or clam within site
effects was insignificant for hierarchical models tested (P = 0.87).
Consequently, independence of error terms could be assumed and
significance tests based on the diffei-ence-transformed data would
be valid.
The ANOVA tables for the difference transformed growth in-
crement as the dependent variable and high-low current category
estimated by clod cards or the numerical model are in Tables 1 and
2. Both models explained 57-589^ of the variability in growth
increment. Estimates from both models indicated clams grew
slower the first year of sampling ( 1996) and that larger clams grew
less with -0.26 mm and -0,28 mm decrease in the growth incre-
ment for each mm increase in initial size depending on whether
1.0
0.5-
Jljrthm..Tnrrnii
0.0
-0.5
+0.5
0,0
-0,5 -
-1.0
f
10
20
30
Lag
40
50
60
Figure 3. Autocorrelations between residual linear model errors with
lags from 1 to 50 for predicting shell increment (top) and difference
transformed measurements of shell increment ( bottom ). Lines above
and below zero baselines are 95% amlldence Intervals for autocorre-
lation = 0.0.
so
CONGLETON ET AL.
TABLE 1.
ANOVA of growth increment with a difference transformation resulting from fitting a complete model reduced until only statistically
significant effects remain. Current categories were average current <5 cm/s or average current 55 cm/s as estimated from clod cards using
Eq. 1. R- of 58%.
Source
Sum-of-Squares
df
Mean-Square
F-Ratio
Year
Initial size
Clod card current
Current * size
Current * year
Error
39:.378
513S.024
143.122
33.430
77.765
435.^,1 OQ
392.378
513S.024
143.122
33.430
77.765
6,032
65.049
851.793
23.727
5..542
12.892
0.000
0.000
0.000
0.019
0.000
cuiTents were described with the numerical model or clod cards
(Fig. 4). Larger average currents also stimulated growth although
the effect on growth increment was less than that of year or initial
size (Table .3). The adjusted least squares mean (±SE) for the
growth increment at the sites identified by clod cards as low flow
was 9.6 ± 0.25 and at the high flow sites was 1 1.9 ± 0.32 (Table
.3). The least squares means (±SE) for growth increment at sites
identified by the numerical model as low flow was 9.51 ± 0.274
cm/s and at the high flow sites was 1 1.70 ± 0.33 cm/s.
There was a significant interaction between year and current
(Tables 1 and 2l. Increased growth for high flow was expected
during the second year because the highest flow site was only
sampled in the second year. There were also significant two-way
(clod card analysis) and three-way (numerical model analysis) in-
teractions involving the effect of initial size indicating an incon-
sistent stimulatory effect of current on growth for animals of dif-
ferent size. However, interaction terms involving initial size made
the smallest contribution to the model Sum of Squares or R".
DISCUSSION
A previous study (Congleton et al. 1999) also reported general
agreement between water velocities estimated by the numerical
model and measured by a portable current meter. The conelation
between flows estimated by the numerical model and Eq. 1 in this
study were lower than reported in Congleton et al. 1999. The 16
sites in this study, however, were a subset of the 25 sites in the
previous study and had a smaller range of current velocities.
Numerous factors affect the accuracy of using clod dissolution
in measuring currents. Mean current velocities estimated with the
clod cards were higher than the velocities estimated using the
model (Table 3). As previously noted, cards were deployed during
a Spring tide when cunents were stronger than an average tide that
is simulated by the model. High estimates of currents using clod
cards compared with other techniques ha\e been previously re-
ported with dissolution rates in field experiments 16-18% high
(Porter et al. 2000) compared with measured flows. Although flow
estimates using cards in this study were higher than estimates
using the model, there should also be some negative bias in the
clod card estimated flows because 9 in Eq. 1 included the time
when the cards were air exposed at low tide while the H used for
calibration was total emersion time. Clod card accuracy could be
increased by calibration in known steady flows rather than using a
diffusion index factor as in this study (Porter et al. 2000).
Flows were anticipated to be greatest at the most easterly and
most westerly sample locations because the flood tide entered the
cove on either side of Spar Island. This anticipated pattern was
seen in the flow rates estimated by the clod cards, but not the
numerical model. The failure of the numerical model to predict
increased currents west of Spar Island may be caused by the av-
eraging of flow rates of the surrounding grid cells, because sample
sites were not centered on the grid. Also, velocity estimates were
an average for a cell with an area of 1305 m". A model with greater
spatial resolution would show flow patterns in greater detail.
With a significant correlation between the current velocities
estimated by the clod cards and numerical model, the similarity in
the statistical analysis for the two sets of current measurements
was not unforeseen. As expected, initial size had a significant
effect on the grow th increment of M. arenaria. resulting in slower
growth in larger individuals (Fig. 4).
In an earlier study (Beal et al. 2001 ) placed clams at the same
intertidal locations in Mason Bay and measured increment in shell
length between time of removal from the hatchery and seeding on
the flats in April and removal from the flats at monthly intervals
until December. Mean shell length increased from 14.1 mm to 21.9
mm resulting in a 7.S mm increase between June and August to
December. Growth increment for the entire srowins season was
TABLE 2.
ANOV.\ of growth increment with a difference transformation resulting from fitting a complete model reduced until only statistically
significant effects remain. Current categories were average current <4 cm/s or average current >5 cm/s as estimated from the numerical
model. R' »{S19c.
Source
Sum-of-Squares
df
Mean-Square
F-Ratio
Year
Initial size
Model current
Current * year
Current * year * size
Error
304.81(1
3524.739
155.673
142.192
62.972
4458.564
I
1
1
1
1
722
,W4.810
3524.739
155.673
142.192
62.97
6. 1 75
49.360
570.781
25.209
23.026
10.197
0.000
0.000
0.000
0.000
0,001
Tidal Currents and Clam Growth
81
Annual Shell Increment
E
E.
*•>
c
0)
E
S.
u
c
<u
CO
10
20
30
40
50
Initial Size (mm)
Figurt 4. Imrement in shell length for the year 2(1(11) for clams In low
and high How sites as categorized using clod cards. (Card L. Card Hi
and the numerical model (Model L. Model H).
slightly less than 12 mm. Although juvenile clams without an
initial growth check were excluded from the sample in this study,
the growth increments predicted for 10 mm clams in Figure 4 is
similar to the value reported by Beal et al. (2001). Brousseau
(1979) predicted an asymptotic size of 108.12 mm and individuals
in age class 5 reaching a harvestable size on Georgetown Island.
Maine. Growth increments from both studies in Mason Bay would
also result in a market size of 51 mm being reached in approxi-
mately 5 y. Results from this study also indicate market size would
be reached earlier by clams at sites with average flows >5 cm/s
than flows <.'i cm/s.
Walne ( 1972) concluded that water current is a significant fac-
tor affecting filtration rates of bivalves, leading to higher growth
TABLE 3.
Adjusted least squares means for annual shell growth increment in
low and high flows as estimated b> clod cards and a numerical How
model. ANOV.A and signincance tests are in Tables 1 and 2.
Mean Flow
G
-owth Increment (mm)
Least
Flow Estimate
(cm/s)
Sq
uare Mean
SE
Clod card
Low flow
4.357 + 0.370
9.565
.247
High flow
5.860 ±0.618
11.899
.323
Numerical model
Low tlow
2.994 ± 0.428
9.505
.274
High flow
5.838 ± 0.457
1 1 .699
.327
rates. The relationship, however, varies with species of bivalve. As
velocities increase, an increased supply of particles corresponds to
increased consumption rates in mussels (Frechette et al. 1989).
Higher currents would afso cause sediment resuspension. Both
frequency of sediment resuspension and sediment food value were
found to be adequate to provide a nutritional benefit to scallops on
George's Bank (Grant et al. 1997). However, filtration and growth
rates were observed to be inhibited at higher flow levels. Mussels
reduce filtration rates on average by 4.8% at velocities >25 cm/sec
(Wildish & Miyares 1990). At a specified algal concentration,
Cahalan et al. (1989) found that growth rates of bay scallops
peaked at an intermediate fiow velocity of 6.5 cm/sec. Sea scallop
feeding is inhibited at currents >10 cm/sec (Wildish & Saulnier
1992. Wildish et al. 1987), and growth may even cease at 12
cm/sec (Kirby-Smith 1972).
Species differences in the stimulatory effect of water currents
on growth were explained by an "inhalant pumping speed" hy-
pothesis that predicts maximum growth at ambient flow the same
as the inhalant pumping speed of the species. Siphonate taxa gen-
erally ha\e greater inhalant pumping speeds. Hard clams (Grizzle
et al. 1992) and mussels (Grizzle et al. 1994). however, increased
growth rates over a wider range of currents.
Although year and initial size had more effect on clam grov\ th
in Mason Bay than did water velocity (Tables 1. 2), clams at high
flow sites did have a larger growth increment than the low flow
sites (Table 3). The results from this study show increasing shell
increments of Mya arenaria of 23-24% at higher average current
velocities. It is possible that the site closest to Flake Point Bar with
a inaximum estimated free stream flow 2 1 .4 cm/sec could have had
feeding inhibition at maximum flood tide. However, preliminary
data (Turner 1991 ) found no decrease in average pumping velocity
of Mercenaria mercenaria in flows between 20 to 30 cm. Addi-
tional studies need to be completed to identify the current velocity
at which physiologic inhibition of feeding occurs in clams and
other siphonate bivalves and also to determine the effect of a wider
range of tidal flows on feeding and growth.
The R" values for the linear models accounted for 57-58% of
the variability in the annual growth increment with differences in
initial size responsible for most of this variability in growth. The
range of water velocities across the study sites was not large. Some
of the unexplained variability inay have been partially caused by
error in counting external growth lines particularly for older indi-
\iduals as was reported for Geiikensia demissa (Brousseau 1981).
Error in predicting current velocities would also decrease R~
for the statistical models. Clod cards were wet and dissolving, but
air-exposed during part of the tidal cycle resulting in overestinia-
tion of 9 in Eq. 2 and a possible underestimation of current speed
in Eq. 1. Field deployed clod cards could be eroded by waves and
cuirents. Shallow water waves result in a local "to and fro" water
motion on the bottom increasing gypsum erosion resulting in over-
estimation of tidal currents using clod cards.
Different calibration techniques for clod cards could increase
accuracy of their use. Calibration of gypsum dissolution in flumes
with known flows was superior to still water calibration (Porter et
al. 2000) as used in this study. Porter et al. 2000 also found that the
gypsum dissolution method should not be used to compare flows
in different flow environments or to measure flows in an environ-
ment different from the calibration environment. These consider-
ations limit the usefulness of clod cards in tidal environments
because the flow environment changes during a tidal cycle. How-
ever, gypsum dissolution experiments should be interpreted as
82
CONGLETON ET AL.
measuring mass transfer relationships rather than flow speed. Bio-
logic response variables such as shell growth in this study may be
directly influenced by mass transfer of nutrients and indirectly
affected by flow.
Another limitation to the predictive capability measured in this
study is the bivalves in the present were not maintained in a con-
trolled environment. Numerous factors could cause stress and af-
fect growth. In a mariculture operation, trampling, predation. and
reburial after digging could be eliminated. Under these conditions,
the impact of water movement on variation in growth may be
greater.
Differences in clam density could also affect growth. Clam
density was not controlled in the present study. Beal et al. 2001
varied seed clam densities between 330 m"" and 1320 m"" at the
same location in Mason Bay without significantly affecting the
growth increment in shell length (Beal et al. 2001). Low clam
densities at all study sites were apparent during field sampling
from the digging effort required to collect the clams. Density was
also found not to have a significant effect on final shell length of
Mercenaria mc'rcenaria grown in bags (Fernandez et al. 1999).
Application to Mariculture Siting
The relationship between bivalve growth and the clod card
erosion should be useful in evaluating mariculture sites. Although
the contribution of cunent magnitude to the R" of the linear model
of growth was small relative to year and initial size, the increase in
growth predicted for clams of uniform size that are seeded at the
same time (or year) would be increased by 22-249f in high fiows
sites relative to low flow sites.
Relative water flow can be estimated by measuring percentage
weight loss of cards deployed at different sites. The use of Eq. 1
for calculating an estimated velocity requires laboratory measure-
ment of clod card loss in quiescent water, but determining the
percent weight loss of cards should be sufficient for estimating
relative flow rates at locations with the same air exposure and
water temperature.
The number of cells required in a grid with sufficient resolution
to estimate local tidal currents is a possible limitation on using a
numerical model. Grid scale is an important aspect of tide mod-
eling in the Gulf of Maine (Sucsy et al. 1993). For use in mari-
culture siting, grid cells should be of the same size or smaller than
the location where the clams are to be seeded. Ramming and
Kowalik ( 1980) considered using a grid with iiTegular steps with
the smallest grid distance in the region of primary interest with
larger grid cells away from the region of high resolution. The
solution for the irregular grid, however, is much more complicated
compared with an equidistant grid with spurious effects decreasing
the accuracy expected from grid refinement. Despite these limita-
tions. Kowalik and Murty ( 1993) gave a number of examples of
models using a combination of coarse and fine grids in their con-
sideration of the problem of using nested and multiple grids to
describe tidal flats.
A frequently used approach is to use the solution from a model
using a coarse grid as input for the boundary conditions for a fine
mesh grid for the area where higher resolution is required. The
development of multiple models at different scales would be fa-
cilitated by using an object-oriented approach. The object-oriented
feature of inheritance allows a general description of model com-
ponents in a base class to be inherited by a child or derived class
with the specific components to be added for a specific implemen-
tation. An object-oriented, two-dimensional landscape model with
biologic components has been pre\ iously developed (Congleton et
al. 1997).
For time series descriptions of current magnitude and direction
over large areas, obtaining estimates from a numerical model
would be the most practical. The incorporation of current estimates
from a numerical model in a GIS, as described by Congleton et al.
(1999), would make the information readily retrievable for use in
aquaculture siting and other applications.
ACKNOWLEDGMENTS
This project was supported by the Maine Agricultural Experi-
ment Station (MAES Pub. No. 2630). Assistance of Brian Beal in
digging clams and identifying growth checks is greatly appreciated.
LITERATURE CITED
Bayne. B. L.. R. J. Thompson & J. Widdows. 1976. Physiology I. In: B. L.
Bayiie, editor. Marine mussels: their ecology and physiology. Cam-
bridge: Cambridge University Press, pp. 121-206.
Beal. B. P.. M. R. Parker & K. W. Vencile. 2001. Seasonal effects of
intraspecific density and predator exclusion along a shore-level gradi-
ent on survival and growtli of the solt-shell clam, Mya arenaria L., in
Maine. USA. J. Exp. Mar. Biol. Ecol. 264:133-169.
Blumberg, A. F. & G. L. Mellor. 1987. Three-Dimensional Coastal Ocean
Models. In: N. Heaps, editor. A description of a three-dimensional
coastal ocean circulation model. Americal Geophysical Union. 208 pp.
Brousseau, D. J. 1979. Analysis of growth rate in Myii arenaria using the
Von Bertalanffy equation. Mar. Biol. .sl(no. 31:221-227.
Brousseau, D. J. 1981 . Use of internal growth bands tor age determination
in a population of Geuken.sia demi.'i.ui (ribbed mussel). Biol. Bull. 161:
323-324.
Cahalan, A., S. E. Siddall & M. W. Lukcnback. 1989. Effects of flow
velocity, food concentration and particle flux on growth rates of juve-
nile bay scallops .Argopecren irradiuns. J. Exp. Mar. Biol. Ecol. 129:
45-60.
Congleton, W. R., Jr.. B. Peaice. M. Parker & B. F. Beal. 1999. Mariculture
Siting: a GIS description of intertidal aras. Ecol. Model. 1 16:63-75.
Congleton. W. R.. Jr., B. R. Pearce & B. F. Beal. 1997. A C++ Imple-
mentation of mdividual/landscape model. Ecol. Model. 103:1-17.
Draper. N. R. & H. Smith. 1966. Applied Regression Analysis. New York:
John Wiley.
DMR. 1997. Fisheries landings statistics for Maine. Maine Department of
Marine Resources. Augusta: State House Station #21. 1 p.
Doty. M. S. 1971. Measurement of water movement in reference to benthic
algal growth. Botanica. Marina. 14:32-35.
Emerson, C. W. 1990. Influence of sediment disturbance and water flow on
the growth of the soft-shell clam. Mya arenaria L. Can. J. Fish. Aqual.
ScL 47:1655-1663.
Evans. S. & B. Tallmark. 1977. Growth and biomass of bivalve mollusks
on a shallow, sandy bottom in Gullmar Fjord (Sweden). Zoon-.ii-iH.
Fernandez. E. M., J. Lin & J. Scarpa. 1999. Culture of .Mercenaria Mere-
cenaria (Linneaeus): Effects of Density, predator exclusion device and
bag inversion. J. Shellfish Res. 18:77-83.
Frechette, M., C. A. Butman & W. R. Geyer 1989. The importance of
boundary-layer flows in supplying phytoplankton lo the benthic sus-
pension feeder. Mylihis edulis L. Limnol. & Oceanogr. 341:19-36.
Grant, J.. P. Cranford & C. Emerson. 1997. Sediment resuspension rates,
organic matter quality and food utilization by sea scallops {Placopecten
magellanicus) on Georges Bank. / Mari. Res. 55:965-994.
TiuAL Currents and Clam Growth
83
Grizzle. R. E.. R. Langan & W. H. Howell. 1992. Growth of suspension-
fceding bivalve molluscs to changes in water tlow : differences between
siphonate and nonsiphonate taxa. J. Exp. Mar. Biol. Ecol. 162:213-228.
Grizzle. R. E.. R. Langan & W. H. Howell. 1994. Growth responses of
Cni.s.so.srmi virgiiiica. Mercenaria merceiniria. tniil Mylilus eilulis to
changes in water flow: A test for the "inhalant pumping speed hypoth-
esis." 7. Shellfish Res. 13:315.
Grizzle. R. E. & P. J. Morin. 1989. Effect of tidal currents, .seston. and
bottom sediments on growth of Mercenaria nurcciiaria: Results of a
field experiniem. Mar. Biol. 102:85-93.
Jacques, A.. J-CF. Brethes & G. Desrosiers. 19X4. Growth o( Mya arcnaria
in relation with sedimentological characteristics and immersion period
on the Rimouski tidal flat. Sci. Tech. Eaii. 17:95-100.
Jorgensen. C. B. 1990. Bivahe Filter Feeding: Hydrodynamics. Bioener-
getics. Physiology and Ecology. Denmark: Olsen & Olsen. 140 pp.
Lewis. D. E. & R. M. Cerrato. 1997. Growth uncoupling and the relation-
ship between shell growth and metabolism in the soft shell-clam Mya
arenaria. Mar. Ecol. Prog. Ser. 158:177-189.
Lin. F.. X. H. Yu. S. Gregor & R. Irons. 1995. Time Series Forecasting
with Neural Network. Complexity International 2: www.csu.edu.au/ci/
vol02/ci2.hlml.
Kowalik. Z. & T. S. Murty. 1993. Numrerical Modeling of Ocean Dynam-
ics. World Scientific. Singapore.
Kirby-Smith. W. W. 1972. Growth of the bay scallop: the influence of
experimental water currents. / E.\p. Mar. Biol. Ecol. 8:7-18.
MacDonald. B. A. & M. L. H. Thomas. 1980. Age determination of the
soft-shell clam Mya arenaria using internal growth lines. Mar. Biol.
58:105-109.
Kube. J.. C. Peters & M. Powilleit. 1996. Spatial variation in growth of
Macoma balthica and Mya arenaria (Mollusca. Bivalvia) in relation to
environmental gradients in the Pomeranian Bay (southern Baltic Sea).
Arch. Fish. Mar. Res./Arch. Fisch. Meereforsch. 44:1-2; 81-93.
Malouf. R. E. 1989. Clam culture as a resource management tool. In: J. J.
Manzi & M. Castagna. editors. Clam Mariculture in North America.
Amsterdam: Elsevier, pp. 427-447.
Mellor, G. L. 1992. User's Guide for a three-dimensional, primitive equa-
tion, nuinerical ocean model. Progress in Atmosphere and Ocean. Sci-
ence, Princeton University. 35 pp.
Muus. B. J. 1968. Field measuring "exposure" by means of plaster balls —
a preliminary account. Sarsia 34:61-68.
Neter. J.. M. H. Kulner. C. J. Nachtsheim & W. Wassennan. 1996. Applied
Linear Statistical Models. 4th ed. Chicago: Irwin Inc. 1408 pp.
Newell. C. R. 1990. The effects of mussel {Mxtiliis cdiilis. Linneaeus.
1758) position in seeded bottom patches on growth at subtidal lease
sites in Maine. / Shellfish Res. 9:1 13-1 18.
Newell. C.R. 1996. Currents, food critical to shellfish farm siting. Fish
Farm. News 2:6-7.
Peterson. C. H. & G. .A. Skilleter. 1994. Control of foraging behavior of
individuals within an ecosystem context: the clam Macoma halihica.
flow environmenl. and siphon-cropping fishes. Oecologia 100:256-
267.
Peterson. C. H.. H. C. Siimmcrson & J. Huber. 1995. Replenishment of
hard clam stocks using hatchery seed: combined importance of bottom
type, seed size, planting season and density. J. Shellfish Res. 14:293-
300.
Pilditch, C. A. & J. Grant. 1999. Effect of variation in flow velocity and
phytoplankton concentration on sea scallop (Placopeclen magellani-
cus) grazing rates. J Exp. Mar. Biol. & Ecol. 240:1 1 1-136.
Porter, E. T., L. P. Sanford & S. E. Suttles. 2000. Gypsum dissolution is not
a universal integrator of "water motion." Limnol. & Oceanogr. 45: 145-
158.
Ramming. H. & Z. Kowalik. 1980. Numerical Modeling of Marine Hy-
drodynamics Applications to Dynamic Physical Processes. Amster-
dam: Elsevier.
Sucsy. P. v., B. R. Pearce & V. G. Panchang. Comparison of two- and
three-dimensional simulation of the effect of a tidal barrier on the Gulf
of Maine tides. / Phys. Ocean. 23:1231-1248.
Thompson. T. L. & E. P. Glenn. 1994. Plaster .standards to measure water
motion. Limnol. & Oceanogr. 39:1768-1779.
Turner. E. J. 1991. A new method for directly measuring bivalve pumping.
J. Shellfish Res. 10:275.
Vincent. B.. G. Desrosiers & Y. Gratton. 1988. Orientation of the infaunal
bivalve Mya arenaria L. in relation to local current direction on a tidal
flat. J. Exp. Mar. Biol. Ecol. 124:205-214.
Walne. P. R. 1972. The influence of current speed, body size and water
temperature on the filtration rate of five species of bivalves. J. Mar.
Biol. Assoc. 52:345-374.
Wildish. D. J. & M. P. Miyares. 1990. Filtration rate of blue mussels as a
function of flow velocity: preliminary experiments. J. Exp. Mar. Biol.
Ecol. 142:213-219.
Wildish. D. J. & A. M. Saulnier. 1992. The effect of velocity and flow-
direction on the growth of juvenile and adult giant scallops. J. of Exp.
J. E.\p. Mar. Biol. Ecol. 133:133-143.
Wildish. D. J., Kristmanson. D. D.. Hoar. R. L., DeCoste. A. M., McCor-
mick, S. D. & A. W. White. 1987. Giant scallop feeding and growth
responses to flow. J. Exp. Mar. Biol. & Ecol. 113:207-220.
Joiinuil of Shellfish Resfunh. Vol. 22. No. 1. S5-y(). 2003.
MATURITY AND GRO\VTH OF THE PACIFIC GEODUCK CLAM, PANOPEA ABRUPTA, IN
SOUTHERN BRITISH COLUMBIA, CANADA
A. CAMPBELL AND M. D. MING
Shellfish Section. Stock Assessment Divisio?) Science Branch. Fisheries and Oceans Canada. Pacific
Biologiccd Station. Nanaiiuo. British Columbia. Canada WT 6N7
.ABSTRACT Measurements were made to determme size and age at maturity and growth of the Pacific geoduck clam. Punopea
abnipm. from two areas in southern British Columbia. Canada. Growth rates were slower for P. ahrupm from Gabriola Island than
those from Yellow^ Bank. Histological examination of gonads indicated that at sizes <90 mm SL considerably more males matured than
females, but at sizes a90 mm SL the sex ratio was similar for males and females. Size at 50% maturity was similar for P. ahnipta
from both areas (58.-3 and 60.5 mm SL. respectively), but age at 50% maturity was slower for geoduck from Gabriola Island (3 y) than
those from Yellow Bank (2 y). Although one hermaphrodite was recorded, P. ahrupm was considered basically gonochoristic
(dioecious).
KEY WORDS: Pacific geoduck. Panopca ahrupia. maturity, sex ratio, hermaphrodite, reproduction
INTRODUCTION
The Pacific geoduck clam, Panopea abrupta (Conrad, 1849)
(Pelecypoda: Hiatellidae). is distiibuted along coastal areas from
southern California to Alaska and west to southern Japan (Bernard
1983. Coan et al. 2fW0). Geoduck are found buried up to 1 m deep
within soft substrates (e.g.. mud and sand) from the low intertidal
to at least 100 ni (Jamison et al. 1984. Goodwin & Pease 1989).
There are commercial fisheries for geoduck in Alaska, British
Columbia, and Washington State (Campbell et al. 1998, Bradbury
& Tagart 2000, Hand & Bureau 2000). Geoduck are long-lived,
reaching ages up to 168 y (Bureau et al. 2002). Adult geoduck
have separate sexes and broadcast spawn annually, usually during
summer (Andersen 1971. Goodwin 1976. Sloan & Robinson
1984). Planktonic larvae settle on substrates within 47 days, and
juveniles burrow into the substrate (Goodwin et al. 1979, Goodwin
& Pease 1989). Geoduck juveniles and adults feed by filtering food
particles (e.g., phytoplankton) from seawater (Goodwin & Pease
1989). Geoduck growth is variable but most rapid in the first 10 y:
thereafter, although growth in shell length is greatly reduced, shell
thickness and meat weight continue to increase at a slow rate
(Bureau et al. 2002).
Andersen (1971 ) found SO'^r maturity occurred at about 75 mm
SL in geoduck sampled in the Hood Canal. Washington State, but
little is known about the rate of sexual tnaturity for P. ahrupta.
especially in British Columbia. (Sloan & Robinson 1984). The
purpose of this paper is to present information on the sexual ma-
turity and growth rates of P. abrupta from two areas in southern
British Columbia.
MATERIALS AND METHODS
Samples from as wide a range as possible of P. ahnipia were
obtained from Yellow Bank, near Tofino on the west coast of
Vancouver Island, (Lat. 49°14.18'. Long. 125"55.48') during 28
May, 1991 and Gabriola Island, near Nanaimo in Georgia Strait,
(Lat. 49°07.6'. Long. 123°45.05') during 22 to 23 May, 1991, at
depths between 5-15 m for both areas. The clams were transported
to the laboratory in coolers (2°C) and kept in running sea water
(ambient temperature) until processed within 48 h of capture.
For each geoduck, shell length was measured as the straight-
line distance between the anterior and posterior margins of the
shell to the nearest mm with vernier calipers. The age of each
geoduck was estimated using the acetate peel method of Shaul and
Goodwin ( 1982). Each right valve was sectioned through the hinge
plate, the cut surface polished, etched with a \% hydrochloric acid
solution for 1.5 min. washed with distilled water, dried, and an
acetate peel made by applying an acetate sheet on the hinge surface
with acetone. Growth rings imprinted on the acetate peel were
counted on a digitizing table after x40 magnification using a Neo-
Promar projector. Although most individuals had their SL and age
ineasured. there were some that had only the SL or only the age
measured; these latter individuals were included in the analysis
where appropriate. Reproductive condition of each geoduck was
determined by removing a sample from the central portion of the
gonad and preserving the tissue in Davidson's Solution (Shaw &
Battle 1957). Histological slides were prepared with sections of the
gonad stained with heniatoxylin-eosin. Histological sections of the
gonads were classified into six stages according to Andersen
( 1971 ). Stage was immature (no differentiation in gonadal tissue:
loose vesicular connective tissue in gonad). The other stages were
for mature geoduck (connective tissue well developed, primary
cells evident on follicle walls or eggs or sperm development evi-
dent) and classified as: ( 1 ) early active: (2) late active: (3) ripe: (4)
partially spent: and (5) spent.
Average von Bertalanfy growth curves were fitted to all data
points of size at age using the equation;
L, = L fl
')
where t is age in years. L, is shell length (mm) at age t, L,, is
theoretical maximum size, k is a constant, determining rate of
increase or decrease in length increments, t„ is the hypothetical age
at which the organism would have been at zero length. The pa-
rameters L^ . k, and t^, were estimated using a non-linear Gauss-
Newton least squares method (SYSTAT 2000).
The proportion of mature geoduck (P) at shell length or age (X)
was estimated using the equation;
Px = X/(X -t- e'*-''^')
where A and B are parameters estimated using a non-linear Gauss-
Newton least squares method (SYSTAT 2000). Data for both sexes
were combined for each of the growth and maturity curve analyses
since sex could not be distinguished in the immature sizes.
85
86
I-
O
LU
200
150-
100-
LU
^ 50H
Campbell and Ming
200 n
o
o
O O On
OO OCPO (5>
n
oO Q?' O O
o
20 40 60
AGE (YEARS)
80
100
)J!i^
>^
T
T
20 40 60 80
AGE (YEARS)
100
Figure I. Growth curves for P. abnipla collected from (Al Gabriola Island, and (Bl Yellow Bank. Curves calculated from the von Bertalanfy
growth parameters (Table ll.
RESULTS
Growth
The oldest P. ahruphi collected was 77 y (146 mm SL| from
Gahriola Island, and 1 17 y ( 154 mm SL) from Yellow Bank. The
smallest and largest geoduck, respectively, was 10 mm SL (age
unknown, probably 1 y) and 163 mm SL (42 y) from Gabriola
Island, and 43 mm SL (2 y) and 180 mm SL (58 y) from Yellow
Bank. Growth was fastest in the first 10 y followed by slow growth
thereafter for geoduck from both areas (Fig. I). There was con-
siderable variability of size within each age group. Growth rates of
P. ahrupui from Gabriola Island were slower than those from
Yellow Bank (Fig. 1. Table I).
Gonadal Condition
Immature gonads comprised 10.85% and 12.10% of the total
geoduck gonads sampled from Gabriola Island {n = 129) and
Yellow Bank (n = 124, includes three individuals without SL
measurements), respectively (Fig. 2). The largest immature geo-
duck was 80 mm SL (5 y) and 72 mm SL (4 y) from Gabriola
Island and Yellow Bank, respectively. There were Insufficient data
to determine spawning periods because seasonal monthly samples
were not collected. However, most mature gonads were in the
ripe or partially spent condition for geoduck collected from both
areas (Fig. 2). There were no gonads that were spent (gonadal
TABLE I.
Von Bertalanfy growth parameters for P. abnipla from Ciabriola
Island and Yellow Bank during May 1991. Values in brackets are
approximate 95% confidence intervals.
Area
Lx
Gabriola Island
Yellow Bank
129.6 (±4,1)
147.7 (±5. Si
0.146 (±0.020)
0.189 (±0.055)
-1.02 (±0.951
-1.42 (±1.17)
120
108
condition 5). This suggested that geoduck spawning had begun at
both areas during mid to late May 1991.
Sex Ratio
For geoduck <90 mm SL. in both areas combined. 41.1 8% were
immature, and 54.41% were males (Table 2). The sex ratio for
mature geoduck <90 mm SL was predoniinantls (92.5%) male
70 n
12 3 4
GONADAL CONDITION
Figure 2. Frequency of gonadal condition stages found in gonads of all
/'. ahrupta collected from Gabriola (black bars) and Yellow Bank
(hatched bars). Gonads classified as = immature, and mature stages
that are I = early active: 2 = late active: 3 = ripe: and 4 = partially
.spent.
Geoduck Maturity
87
TABLE 2.
Pericnl of total gonads differentiated into mature males and females
and immature /'. ahnipla from (iahriola Island and \ eiloM Bank
during .Ma> IVMI. One 91 mm SI, hermaphrodite was found. N =
total nuniher. Includes onlv individuals with SL measurements.
Percent of Total
Area
Male
Female
Immature Hermaphrodite
N
<y() mm SL
Gahricila Island
56.76
5.40
.37.84
.37
\elk)W Bank
.51.61
.■i.2.^
45.16
31
Total
.54.41
4.41
41.18
68
>90 mm SL
Gabriola Island
57.61
42..^9
92
Yellow Bank
45.56
53..^-^
1.11
90
Total
5 1 .65
47.80
0.55
182
uith few (7. 3%) females for both areas combined. In contrast.
geoduck s90 mm SL had generally a more equal se,\ ratio, al-
though males were slightly more abundant than females in the
Gabriola Island sample, whereas there were slightly more females
than males in the Yellow Bank sample (Table 2),
Figure 3. Ph()liiriiicni;;ra|)lis iil /' ahnipla gonadal tissue cross-
sections of (.\l Male (x4(Mt magnilkation) showing spermatozoa-filled
follicle surrounded by connective tissue, (B) Female (x4()0) showing
oocyte-filled follicle surrounded bv connective tissue.
Hermaphroditism
.Although most of the histological material of mature P. ahnipta
gonads allowed differentiation between females (follicles with oo-
cytes) and males (follicles with .spermatozoa) (Fig. 3) there was
one individual that was a hermaphrodite, with a gonad showing
both male and female characteristics (Fig. 4). This gonad had some
follicles containing only either female or male gametocytes per
follicle, and other follicles, which contained spermatozoa and oo-
cytes in the same follicle. The geoduck was 91 mm SL (age was
not determined).
Malurity
Mean size at 50% maturity was similar for geoduck from
Gabriola Island, 58.3 mm SL (55.2-59.4 mm SL, lower and upper
95% confidence intervals, CI), and Yellow Bank, 60.5 mm SL
(51.1-64.0 mm SL. 95% CI) (Fig. 5, Table 3). Mean age at 50%
maturity was about 1 y slower for geoduck from Gabriola Island,
3.09 y (2.68-3.25 y, 95% CI), than at Yellow Bank, 2.04 y ( 1 .72-
2.16 y. 95% CI) for Yellow Bank geoduck (Fig. 6. Table 3). The
smallest mature male was 45 mm SL (2 y) and 60 mm SL (2 y),
the smallest mature female was 59 mm SL (4 y) and 88 mm SL
(2 y), and the largest immature geoduck was 80 mm SL (5 y) and
72 mm SL (4 y), respectively, in the samples from Gabriola Island
and Yellow Bank.
B
*■ ■
. *•
s-
Figure 4. Photomicrographs of hermaphrodite I', ahnipta gonadal tis-
sue cross-sections of (.\) (x250 magnification), and (B) (xl60) showing
single follicles containing oocytes and spermatozoa.
88
Campbell and Ming
1.0
UJ
Q: 0.8H
Z)
0.6-
01 0.4
O
CL
o
a: 0.2H
Q_
0.0-
I I I I
T
50 100 150
SHELL LENGTH (MM)
200
1.0
LU
Qi 0.8-
I-
<
^ 0.6-
z
o
fe 0.4-
O
Q.
o
01 0.2-
CL
0.0-
O OC
OO QCO
50 100 150
SHELL LENGTH (MM)
200
Figure 5. Size at maturity curves for P. abntpta collected from (A) Gabriola Island, and (B) Yellow Bank. Symbols indicate number of individuals
per shell length: "O" = I; "X" = 2; "+" = 3. See text for equation for the predictive curve and Table 3 for parameter values.
DISCUSSION
Our findings indicated that growtli rales were faster for geo-
duck from Yellow Bank than those from Gabriola. Results were
similar to those of Burger et al. (1998) and Bureau et al. (2002)
who found that geoduck from Georgia Strait were generally
smaller than those from the west coast of Vancouver Island. Rea-
sons for the differences in P. abntpta growth rates between areas
could be attributed to a variety of environmental and biological
factors associated with different habitats (e.g.. substrate type, tem-
perature, exposure to water surge activity, pollution, food avail-
ability, and geoduck density or genetic characteristics) (Breen &
Shields 1983. Harbo et al. 1983, Goodwin & Shaul 1984, Goodwin
& Pease 1991, Noakes & Campbell 1992. Hoffman et al. 2()0().
Bureau et al. 2(X)2).
Our examination of gonadal condition suggested thai the
spawning period for geoduck from both study areas was just be-
ginning in mid to late May 1991. Results agree with other gonadal
studies of geoduck, which found the main spawning period was
TABLE 3.
Parameter estimates for equation indicating relationships betv\een
proportion that are mature with shell length (SI. in mm) or age
(years! of P. abriipta from (iabriola Island and Yellow Bank during
May 1991. See text for equation formula. \ alues in brackets are
approximate 95% confidence intervals.
Parameter Estimates
Variable
.\rea
X
Gabriola Island
SL
Yellow Bank
SL
Gabriola Island
Age
Yellow Bank
Age
8.512 (±2.741) 0.076 (±0.044) 79
7.224 (±2.. ^14) 0.052 (±0.0.^3) 80
2.956 ( ± 1 .55 1 ) 0.59 1 ( ±0.435 ) 1 5
2..397 (±1.540) 0.828 (±0.644) 14
during June and July (.Andersen 1971. Goodwin 1976. Sloan &
Robinson 1984).
The male:female sex ratio of mature P. abntpta found in this
study (52:48) was similar to that reported by Goodwin (1976)
(.53:47) and Sloan and Robinson ( 1984) (37:43). The high percent-
age of males in the small sizes (young ages) in this study was
1.0-
LU
cn O.BH
3
0.6-
q: 0.4-
O
Q.
o
q: 0.2
Q.
0.0-
6 6 5 12 2 1 3
XM^^:^!^' ^ ig^ O » O
, ' " "^/Al 4 5 6 3 11
-1 1 1 \ 1 1 1 r
5 10
AGE (YEARS)
15
Figure 6. Age at maturity curves for P. abriipta collected from
Gabriola Island ("O" solid curve), and bellow Bank ("X" and dashed
curve). Number by each symbol indicates numlier of individuals per
age group. See text for equation for the predictive curve and Table 3
for parameter values.
Geoduck Maturity
89
similar to Andersen's ( 1971 ) findings of94.4'7i- males among geo-
duck with <100 mm SL.
Our findings indicated the first recording of a P. cibnipui her-
maphrodite. Most bivahe species are dioecious (sexes are sepa-
rate) although hermaphroditism does occur in some species of this
group (Coe 1943, Coan et al. 2000). Factors causing hermaphro-
ditism in P. ubnipia are unknown. Whether the "simultaneou.s"
hermaphroditism (Coe 1943. Eversole 1989) in this geoduck was
fully functional in producing viable eggs and sperm is unknown.
However, sexuality of different sizes (or ages) In F. nhntpta has
not been studied extensively. We estimated thai only -1.200 indi-
vidual gonads have been histologically examined to date from
mature P. dhniprn sampled in Washington State and British Co-
lumbia (Andersen 1971. Goodwin 1976. Sloan & Robinson 1984.
this study). Andersen (1971) and Goodwin (1976) suggested that
P. ahnipia might be gonochoristic where sex is determined by
development with males maturing at a smaller size (earlier age)
than females. Although we suspect that hermaphroditism is rare in
P. ahriipta. the probability that some level of protandry. sex re-
versal, or "simultaneous"" hermaphroditism in P. nhnipla (espe-
cially for sizes <I00 mm SL) ma\ occur and should be in\esti-
gated further.
Sexual maturity was variable between P. ahniplu individuals
and sexes. Males started to mature at an earlier age than female
geoduck in Yellow Bank than Gabriola Island. Although size at
SC/f maturity was similar for P. ahnipia from both areas (58.3 and
60.5 mm SL. respectively) age at 50% maturity was slower for
geoduck from Gabriola Island (3 y) than Yellow Bank (2 y).
Andersen ( 1971 ) found sexual maturity of geoduck to be variable,
the smallest sexually mature geoduck to be 45 mm SL, and 50%
size at maturity to be 75 mm SL (which Andersen estimated to be
an age of 3 y). Our study is the first to show that although size at
maturity may be similar for geoduck from two different areas,
differences in growth rates may influence the age at which geo-
duck matures sexually. These findings are siinilar to some studies
of other bivalve species, which suggest that onset of maturity may
depend more on size than age (e.g.. Nakaoka 1994). However, size
and age at sexual maturity can also vary between populations in
the same bivalve species (Ponurovsky & Yakovlev 1992. Sato
1994). Variation in environmental (e.g.. temperature, current pat-
terns, substrate type, and depth) and biological (e.g., genetics, food
supply, growth and mortality rates, predation. and parasitism) fac-
tors may affect maturity rates w ithin bivalve populations at differ-
ent locations (Thompson et al. 1980, Ponurovsky & Yakovlev
1992, Nakaoka 1994, Sato 1994, Taskinen & Saarinen 1999).
ACKNOWLEDGMENTS
The authors thank M. Boudreau, G, Hickie, D. Larson, M.
Lanoie. and N. Sorenson for the geoduck collections. S. Bower. W.
Carolsfeld. B. Clapp. S. Dawe. L. Lee. and T. White for technical
assistance, and J. Blackbiiurne, N. Bourne, S. Bower, and G.
Gillespie for helpful comments on early drafts of this manuscript.
LITERATURE CITED
Andersen, A. M. 1971. Spawning, growth, and spatial distrihution of the
geoduck clam Punope generosa (Gould), in Hood Canal, Washington.
PhD thesis. Seattle; University of Washington. I2S pp.
Bernard, F. R. 1983. Catalogue of the living Bivalvia of the eastern Pacific
Ocean: Bering Strait to Cape Horn. Can. Spec. Piihl. Fi.'ilt. Aquat. Sci.
61:102.
Bradbury. A. & J. V. Tagart. 2000. Modeling geoduck. Panopea abrupta
(Conrad. 1849) population dynamics. II. Natural mortality and equi-
librium yield. J. Shellfish Re.s. 19:63-70.
Breen, P. A. & T. L. Shields. 1983. Age and size structure in five popu-
lations of geoduc clams (Panope generosa) in British Columbia. Can.
Tech. Rep. Fish. Acpial. Sci. 1 169: 62 pp.
Bureau. D., W. Hajas. N. W. Surry. C. M. Hand. G. Dovey & A. Camphell.
2002. .Age, size structure and growth parameters of geoducks [Panupca
ahnipia. Conrad 1849) from 34 locations in British Columbia sampled
between 1993 and 2000. Can. Tech. Rep. Fish. Aquat. Sci. 2413:84 pp.
Burger, L.. E. Rome. A. Campbell. R. Harbo. P. Thuringer, J. Wasilewski
& D. Stewart. 1998. Analysis of landed weight information for geoduck
clams (Panopea abrupta) in British Columbia. 1981 to 1995. Can.
Tech. Rep. Fish. Aquat. Sci. 2214:363-372.
Campbell. A., R. M. Harbo & C. M. Hand. 1998. Harvesting and distri-
bution of Pacific geoduck clams, Panopea abrupta, in British Colum-
bia. Can. Spec. Publ. Fish. Aquat. Sci. 125:349-358.
Coan. E. v.. P. H. Scott & F. R Bernard. 2000. Bivalve seashells of western
North America. Santa Barbara: Santa Barbara Museum of Natural His-
tory Publication. 566 pp.
Coe, W. R. 1943. Sexual differentiation in mollusks. 1. Pelecypods. Quar-
terly Rev. Biol. 18:154-164.
Eversole, A. G. 1989. Gametogenesis and spawning in North American
clam populations: implications for culture. In: J. J. Manzi & M. Casta-
gna, editors. Clam mariculture in North America. New York: Elsevier
Science Publishing Company Inc. pp. 75-109.
Goodwin, C. L. 1976. Observations of spawning and growth ol suhlidal
geoducks {Panope generosa. Gould). Proc. Natl. Shellfish Assoc. 65:
49-58.
Goodwin, C. L. & B. Pease. 1989. Species profiles: life histories and
environmental requirements of coastal fish and invertebrates (Pacific
Northwest) — Pacific geoduck clam. U. S. Wildl. Serv. Biol. Rep. 82
( 1 1. 120). United States Army Corps of Engineers. TR EL-82-4.I5 pp.
Goodwin, C L. & B. C Pease. 1991 . Geoduck, Panopea abrupta (Conrad,
1849). size, density, and quality as related to various environmental
parameters in Paget Sound, Washington. / Shellfish Res. 10:65-77.
Goodwin, C. L. & W. Shaul. 1984. Age, recruitment and growth of the
geoduck clam (Panope generosa. Gould) in Puget Sound Washington.
Wash. Dept. Fish. Prog. Rep. 215:30.
Goodwin, L., W. Shaul & C Budd. 1 979. The larval development of the
geoduck clam [Panope generosa. Gould). Proc. Nat. Shellfish. .Assoc.
69:7,V76.
Hand. C M. & D. Bureau. 2()()(). Quota options for the geoduck (Panopea
abrupta) fishery in British Columbia for 2001 and 2002. Ottawa: Can
Stock Assessment Secretarial Res. Doc. 53 pp.
Harbo, R. M.. B. E. Adkins. P. A. Breen & K. L. Hobbs. 1983. Age and
size in market samples of geoduc clams (Panope generosa). Can. M.S.
Rep. Fish. .Aquat. .Sci. 1714:77.
Hoffman, A., A. Bradbury & D. L. Goodwin. 2000. Modeling geoduck.
Panopea abrupta (Conrad, 1849) population dynamics. I. Growth. /
Shellfish Res. 19:57-62.
Jamison. D.. R. Heggen & J. Lukes. 1984. Underwater video in a regional
benthos survey. Honolulu: Proceedings Pacific Congress on Marine
Technology. Mar. Tech. Soc. April 24 to 27, 1984.
Nakaoka. M. 1994. Si/e-dependent reproductive traits of Yoldia notahilis
(Bivalvia: Protobranchia). Mar. Ecol. Prog. Ser. 114:129-137.
Noakes, D. J. & A. Campbell. 1992. Use of geoduck clams to indicate
changes in the marine environment of Ladysmith Harbour. British Co-
lumbia. Environinetrics 3:81-97.
Ponurovsky. S. K. & Y'u. M. Yakovlev. 1992. The reproducuve biology of
90
Campbell and Ming
the Japanese littleneck, Ta/ws phitippmarium (A. Adams and Reeve,
1850) (Bivalvia: Veneridae). J. Shellfish Res. 11:265-277.
Sato, S. 1994. Analysis of the relationship between growth and sexual
maturation in Phacosoma japoniciim (Bivalvia; Veneridae). Muv. Biol.
118:663-672.
Shaul, W. & C. L. Goodwin. 1982. Geoduck iPanope geiierosa: Bivalvia)
age as determined by internal growth lines in the shell. Can. J. Fish.
.\qiuit. Sci. 39:632-636.
Shaw. B. L. & H. I. Battle. 1957. The gross and microscopic anatomy of
the digestive tract of the oyster Crassostrea virginica (Gmelin). Can. J.
Zool. 35:325-347.
Sloan, N. A. & S. M. C. Robinson. 1984. Age and gonad development in
the geoduck clam Pcmope ahrupia (Conrad) from southern British Co-
lumbia. Canada. J. Shellfish Res. 4:131-137.
SYSTAT'S' for Windows* software. 2000. Statistics Software SPSS Inc.,
Chicago, IL, USA
Taskinen, J. & M. Saarinen. 1999. Increased parasite abundance associated
with reproductive maturity of the clam AnodoiUci pisciiuilis. J. Para-
sitol. 85:588-591.
Thompson. I., D. S. Jones & J. W. Ropes. 1980. Advanced age for sexual
inaturity in the ocean quahog .Anica islandica (Mollusca: Bivalvia).
Mar. Biol. 57:35-39.
Joiinuil ,>/ Shellfish Ri'search. Vol. 22, No. I. 91-94. 2(K).\
THE EFFECTIVENESS OF N-HALAMINE DISINFECTANT COMPOUNDS ON PERKINSUS
MARINUS, A PARASITE OF THE EASTERN OYSTER CRASSOSTREA VIRGINICA
M. A. DELANEY,'* Y. J. BRADY,- S. D. WORLEY,' AND K. L. HUELS-
^ Aquatic Animal Health Research Laboratory. USDA-ARS. P.O. Bo.\ 952. Auburn. Alabama 36831:
'Department of Fisheries and Allied Aquaeultures. Auburn University. Auburn. Alabanui 36H49:
Department of Chemistry. Auburn University, Auburn, Alabama 36849
ABSTRACT The pathogenic protozoan Perkinsus marinus (Mackin. Owen and ColHer) is the cause of extensive mortalities in
Eastern oyster. Cwssosirea virgiiiicci. populations along the Gulf and East Coasts of the United .States. A series of experiments was
undertaken to determine the effect of N-hakutiine disinfectants on this protozoan parasite. The organic N-halamine disinfectants.
1.3-dichloro-2.2.5.5-tetramethyl-4-imidazolidinone (DC) and l-chloro-2,2.5,5-tetramethyl-4-imidazolidinone (MC). apparently dam-
age the permeability of the parasites outer membrane and alter the osmoregulatory functions of the cell. Damaged parasites were unable
to reproduce at concentrations as low as 14.9 mg/L DC at 8 h exposure, or for the chemical MC at 24.9 mg/L at 12 h exposure. The
chemical compounds appear to lyse the larger meronts first, followed by lysis of the daughter spores. These studies strongly suggest
that the chemical compounds DC and MC can be u.sed to disinfect seawater allowing the production of specific pathogen-free stock
in oy.ster hatcheries, and having the potential to prevent the spread of these parasites froin contaminated oysters to uninfected oysters.
KEY WORDS: ovster. Pcikinsii.\ marimis. disease, disinfection. N-halamine
INTRODUCTION
The Eastern oyster. Crassostrea vir^inica (Gmelin 1791 ) natu-
rally occurs in North America frotn the Gulf of St. Lawrence in
Canada to the Gulf of Mexico. It is common in estuaries in coastal
areas of reduced salinity, and is an important commercial species.
Once considered the most abundant source of oysters in the world,
eutrophication, overharvesting and the parasites Haplosporidiiim
nelsoiii and P. nuu'iiuis have caused the Chesapeake Bay oyster
population to be reduced to a critically low level (Andrews 1988.
Haskin & Andrews 1988, Hargis & Haven 1988). The parasites
inhibit growth, reduce fecundity, and lower the oyster's condition
and glycogen content (Menzel & Hopkins 1953, Newell 198.3,
Barber et al. 1988. Crosby & Roberts 1990). Oyster populations
that have incurred high infection prevalence and intensities typi-
cally have low mortalities during their first year, but suffer higher
mortalities in the following years (Paynter & Buneson 1991 ). The
parasite does not have the same drastic effects on the oyster popu-
lation in the Gulf of Mexico as it does in the Chesapeake Bay. An
oyster requires three or tnore years to reach marketable size in the
cooler waters of the Atlantic; however, only two years are required
in the warmer waters of the Gulf of Mexico. In the Gulf of Mexico,
this parasite infects over 80% of Eastern oysters with annual mor-
talities typically 50% of the adult oyster population. Transmission
of the parasite occurs through the water by release of infective
stages from the feces of living oysters, the tissues of dead oysters
(Ray 1932. Mackin & Hopkins 1962). and by the gastropod ecto-
parasitic snail, Boonea impressa (White et al. 1987).
Perkin.sus marinus has several life stages in the host oyster
(Mackin & Boswell 1956. Perkins 1969). These include immature
thalli. mature unicellular thalli (trophozoites), and presporangia.
When released into seawater. presporangia develop a resistant cell
wall, and then enlarge to become hypnospores. Under aerobic
conditions, hypnospores differentiate into sporangia and produce
*Corresponding author: E-mail: mdelaney@vetmed.auburn.edu
This study was funded by Mississippi-Alabama Sea Grant Consortiimi.
motile zoospores (aplanospores in Mackin & Boswell 1956) within
the hypnospores cell wall. One sporangium of P. marinus is ca-
pable of releasing approximately 354.700 zoospores (Chu &
Greene 1989). Zoospores are released from hypnospores and un-
dergo free-living stages in seawater.
Eradication of these pathogens in the wild is not possible be-
cause of the widespread nature of the diseases and the lack of
knowledge regarding other species that might carry the disease
(Elston 1990). Resistance to H. nelstmi. but not to P. marimis
(Barber & Mann 1991 ) has been achieved through selective breed-
ing of C. virgiiiicci (Ford & Haskin 1987. Foid et al. 1990. Bur-
reson 1991).
Developing and maintaining hatcheries to produce larval oys-
ters for grow out for co)nmercial production or to repopulate de-
pleted areas is one approach to alleviate the lack of natural repro-
duction. This method, however, requires the incoming seawater to
be specific pathogen free. The traditional methods of using ozone
and ultrafiltration are expensive for continuous production. Chlo-
rine is an inexpensive alternative for water disinfection; however.
Its chemistry changes when combined with seawater.
Observations of oyster larvae exposed to chlorine-treated sea-
water indicate a lethal concentration for 50% of the test organisms
(LC 50) for C. virgiiiica larvae of 0.005 mg/L free chlorine (CI+),
regardless of whether static or intermittent addition of chlorine was
used (Roberts et al. 1975, Bellanca & Bailey 1977. Roberts &
Gleeson 1978). Concentrations as low as 0.05 mg/L of bromate,
broiTtoform and chloroform caused some C. virgiiiica 48 h larval
mortality (Stewart et al. 1979). Galtsoff (1946) noted a 46% de-
crease in pumping action at a dose of 0.2 mg/L chlorine. He and
other workers concluded, however, that chlorine was an effective
means for disinfecting shells of contaminated oysters and that the
oxidant would not interfere with depuration if chlorine levels were
kept at a minimum. Later studies agreed with this finding but
cautioned that oysters reduce pumping when chlorine concentra-
tions exceed 0.01 mg/L. At chlorine concentrations above 1.0 ing/L.
pumping cannot be maintained; thus, the use of chlorine as an
effective means of depuration is limited by the tolerance of the
species. The ability of adult shellfish to respond to low concen-
trations of total residual oxidant and to cease pumping may be
91
Delaney et al.
beneficial because it allows the animal to survive chlorine-
produced oxidant (CPO) concentrations as high as 10 mg/L for 30
days (Galtsoff 1964). The corresponding decrease or cessation,
however, of shell growth and feeding is disadvantageous. The
most severe restrictions to chlorine use arise from the formation of
chemical compounds from adding this to seawater. Halogenated
organic compounds are formed that display complex chemistry.
The products of chlorination of seawater are complex and not fully
understood (Carpenter & Macalady 1973. Davis & Middaugh
1977, Wong & Davidson 1977, Carpenter et al. 1980). In seawater
and brackish water, chlorine replaces some of the bromine in hy-
pobromous acid releasing the bromine cation that is considered the
disinfecting compound. Full strength seawater has a bromide ion
concentration of 65 mg/L, and chlorine reacts with it to produce
hypobromous acid and hypobromite ion. Bromamines and
chloramines may be formed in the presence of ammonium ion. For
normal seawater of pH 8, the initial products of chlorination are a
mixture of hypobromous acid and hypobromite ion that are un-
stable with respect to decomposition and disproportionation
(Macalady et al. 1977).
The N-halamine compounds used in this study were 1,3-
dicliloro-2,2,3.5-tetramethyl-4-imidazolidinone ( DC; dichloro)
and 1 -chloro-2,2,5,5-tetramethyl-4-imidazolidinone (MC;
monochloro). Both compounds were synthesized at Auburn Uni-
versity in the laboratory of S. D. Worley. Department of Chemis-
try. The compound MC can be produced in the laboratory and as
a result of the hydrolysis of the compound DC. The compounds
will be marketed by Vanson/HaloSource Corporation, Seattle
WA'. These compounds are more stable in water and dry storage
than free chkirine and other commercial products, such as the
hydantoins and isocyanurates (Tsao et al. 1991). The N-halamine
compounds do not produce trihalomethanes or react with bromide
in seawater and should be more stable and more effective than free
chlorine. The compound DC is the faster acting compound and the
amide N-Cl moiety is more labile than the amine N-Cl group,
providing a small amount of free chlorine. The hydrolysis decom-
position product MC. having only the more stable amine N-Cl
moiety, acts more slowly as a disinfectant.
In this series of experiments, the parasites were exposed to the
chemical compounds in sterile artificial seawater (SASW) to de-
termine the effectivity of the compounds. A related compound,
3-chloro-4,4-dimethyl-2-oxa/olidinone, has been shown to kill
Giaidia lanihlia more effectively than free chlorine (Kong el al.
1988), and it was speculated that DC or MC would penetrate oyster
tissues and the thick parasite walls at a reduced level of chlorine.
A previous study using Anadara trapezia (blood cockles) and
Haliotis laevii-ata (greenlip abalone) showed that free prezoospo-
rangia of Peikinsus sp. (remo\ed from oyster tissues) died within
30 min in chlorine solutions of 40 mg/L (Goggin et al. 1990);
however, within tissues the parasites presumably are more pro-
tected and survived at least 2 h. Their study was concerned pri-
marily with disinfecting meats of abalone. The objective of this
study is to determine if the parasite P. iiniriiuis could be eliminated
in the water column. The possibility of controlling P. inariiuis in
an oyster hatchery by treating incoming water, or as an interim
control preventing the spread of the parasite between oysters,
could mean economic gains associated with increased health and
growth characteristics.
Use of trade or manufacture's name does tuh imply endorsement.
METHODS
A series of three experiments were conducted to evaluate the
effectiveness of these compounds on P. marinus.
Perkinsiis mannus cultures were obtained from the American
Type Culture Collection (ATCC), and cultured according to La
Peyre and Faisal (1995). In experiment one, an aliquot was re-
moved from culture, vortexed briefly to break up cell clumps, and
then centrifuged at 5(.)0i; for 5 min. These cells were rinsed twice
with 15 ppt sterile artificial seawater (SASW), then resuspended in
SASW at a concentration of approximately 5 x lO'' cells niL"'.
The chemicals DC and MC, which were synthesized according to
the method of Tsao et al. (1991), were prepared in three concen-
trations: 0.3. 14.9 and 29.8 mg/L and 0.5, 24.9 and 49.8 mg/L,
respectively. These concentrations are based on molar equivalents
of chlorine. Four replications of each chemical at each concentra-
tion were prepared in sterile. 50 niL. polypropylene centrifuge
tubes. Approximately 5000 parasites were added to tubes contain-
ing 50 mL of each chemical concentration. The same amount of
SASW with and without parasites served as the positive and nega-
tive controls. Contact time consisted of eight time intervals: 0.5. 1,
2, 4, 8, 12. 18. 24. and 48 h. At the appropriate time, the samples
were mixed and I niL removed from each tube. Sodium thiosulfate
(0.02 N) was added to neutralize the chlorine (i.e., to quench
disinfecting action! and the cells were observed microscopically at
xlOO with and without staining with Lugols Iodine.
A second experiment was initiated to determine the percent
mortality at \ arious concentrations and time intervals using a vital
dye, trypan blue, which distinguishes between living and dead
cells. This viability test evaluates the breakdown of membrane
integrity determined by the uptake of the dye to which the cell is
normally impermeable. Cell and chemical preparation was the
same as previously described. Contact time consisted of three time
intervals: 1. 2. and 8 h. At the appropriate time, the samples were
mixed and 1 niL removed from each tube. The cells were washed
with Hanks Balanced Salts Solution (HBSS) (Sigma, St. Louis,
MO) and resuspended in 0.3 niL HBSS to which 0.5 niL trypan
blue was added. The cell suspension was mixed and allowed to
stand at room temperature for 5-15 min. Living and dead cells
were counted and enumerated using a hemacytometer at xlOO.
Dead cells stained a dark blue, but living cells were able to exclude
the dye. Cells with an intermediate blue color stain were consid-
ered dead.
A third experiment was performed to detemiine the viability of
the cells after exposure to the two chemicals, targeting the cells
that lightly stained indicating damage to the membrane. It was
important to know whether these damaged cells would be able to
recover and initiate a new infection.
Cells were removed from culture, centrifuged to pellet the para-
sites then resuspended in SASW. Four concentrations of DC (7.4.
14.9. 29.8. 44.6 mg/L) and 4 concentrations of MC (12.9. 24.9.
49.8. 76.6 mg/L) v\ere prepared in sterile, polypropylene centri-
fuge tubes, and then 2 niL were transferred to individual wells of
tissue culture plates. Three replications of each chemical con-
centration were prepared. Approximately 20 |j.L of the P. marinus
(4.5 X lO'* parasites mL~') cell suspension were added to each
disinfectant chemical. The same amount of SASW with and with-
out parasites was added to the positive and negative controls.
Contact time consisted of four time intervals: 1. 2. 8. and 12 h. At
the appropriate time, the chlorine in the samples was neutralized
with 20 |jiL of 0.02 N sodium thiosulfate and the cells resuspended
Effectiveness of N-Halamine Compounds
93
TABLE 1.
Experiment 2: the etTect of DC and MC concentration and exposure
time on mortality of P. mariniis.
TABLE 2.
Experiment 3: the effect of DC and MC concentration and exposure
time on mortality and replication of P. mariniis.
'Jc Staining
'?c Staining
mg/L
I hour
2 hours
8 hours
DC 0..^
0.3
1.3
DC 14.9
11.4
77.4
DC 29.8
80.0
80.8
MC0.5
0.2
MC 24.9
.VI
16.3
MC 49.S
19.2
22.9
in 3 mL of culture media. A portion of the cell suspension was
removed and evaluated with typan blue staining as previously
described. The remainder of the samples were incubated in the
dark at 25°C and evaluated at 24 and 48 h.
RESULTS
In the first experiment, no visible effects on P. marimis were
observed for DC or MC treatments at any tested concentration up
to 4 h. At 8 h exposure to either DC or MC. all parasite cells
appeared to decrease in size, and at 18 h all cells were completely
lysed at all concentrations. The negative controls appeared free of
debris and bacterial contamination during the test. The positive
controls appeared unchanged and did not exhibit any decrease in
size, nor did they lyse.
The second experiment attempted to refine the earlier one by
determining viability at various contact times. The viability of the
cells exposed to DC has been reduced by 80% at I h at a concen-
tration of 29.8 mg/L (Table 1). At a concentration of 49.8 mg/L
MC at 1 h, a reduction of only 19.2% was observed. At the end of
8 h. 99.8% mortality was observed at 29.8 mg/L DC. as compared
with 25% with 49.8% MC.
In the study addressing the viability and the ability of the para-
site to recover from exposure to the DC and MC compounds
showed a trend towards more rapid deactivation of the parasites by
DC as compared with MC. at similar concentrations (Table 2).
Cells in the positive control treatment exhibited normal growth and
development.
DISCUSSION
Results of this study demonstrated that the compounds MC and
DC eliminated the pathogen P. mariniis in 15 ppt seawater under
laboratory conditions. It is important to kill all parasites because a
single sporangium of P. mariniis is capable of releasing approxi-
mately 354.70(J zoospores (Chu & Greene 1989).
Mortalities of 100% of P. mariniis can be achieved using the
faster acting chemical DC at concentrations of 14.9 mg/ for 8 or
12 h. 29.8 mg/L for 8-12 h or 44.6 mg/L for a minimum of I hour.
mg/L
I hour
2 hours
8 hours
12 hours
12.9
DC 7.4
4.1
11.0
6.9
13,4
88.2
DC 14.9
12.7
34.8
83.0"
33.0"
99.8
DC 29.8
10.4
29.8
36.0"
98.6"
0.2
DC 44.6
98.1"
99.6-'
100-'
100"
16.0
MC 12.4
3.4
6.2
7.1
25.0
MC 24.9
MC 49.8
10.9
17.2
14.3
14.6
20.7
82.0"
70.0"
87.2"
1 was
MC 76.6
98.6"
90.3"
" Indicates cultures in which all parasites died without producing viable
offspring when observed 48 hours after chemical treatment.
The slower acting chemical MC can achie\e 100% mortality at
concentrations of 24.9 mg/L for 12 h, 49.8 mg/L for 8 or 12 h. or
76.6 mg/L for 8-12 h. Additional testing would be desirable to
determine lower concentration effectivity against this pathogen.
Both DC and MC are effective against the oyster parasite P.
mariniis in vitro at concentrations less than the estimated LDg,, of
the oyster larvae exposed to these same chemicals (Delaney et al.
2002). Histologic and physiologic information would be required
on the long term effects of chemical exposure to oyster larvae;
however, either compound has the potential to be used in oyster
hatcheries to prevent infections of P. marimis from occurring, or to
prevent the spread of the disease through the water column if the
contact time is sufficient. Electron microscopy would provide ad-
ditional insight on the mechanism of damage to the parasite's cell
walls at different stages in the life cycle of the parasite.
N-halamines DC and MC at concentrations of total chlorine
within the lar\ al and adult oysters range of tolerance, are effective
for the control of a protozoan pathogen. P. marimis. of Eastern
oysters. These compounds have the potential to be used in oyster
hatcheries and in recirculating based systems to produce specific
pathogen free oysters. The use of these compounds as a substitute
for free chlorine or chloramines would mitigate deleterious physi-
ologic effects currently observed on oyster recruitment and sur-
vival in estuaries receiving chlorinated discharges.
ACKNOWLEDGMENTS
The authors thank Dr. David D. Rouse, Dr. Sharon R. Roberts,
and Dr. George W. Folkerts for their technical assistance and
Dr. Thomas McCaskey, for his attention to details which improved
this manuscript. Additional thanks to Dr. Jeffrey Williams of the
Vanson/HaloSource Company for providing the chemicals used in
this study and technical assistance, and Dr. John Supan for pro-
viding larval oysters.
LITERATURE CITED
Andrews. J. D. 1988. Epizootiology of the disease caused by the oyster
pathogen Perkinsus marimis and its effect on the oyster industry. Am.
Fisli. Soc. Special Pulylication 18:47-63.
Barber. B. J., S. E. Ford & H. H. Haskin. 1988. Effects of the parasite MSX
{Haplosporidium nelsoni) on oyster (Crassostrea virginica) energy me-
tabolism. 1. Condition index and relative fecunditv. J. Slwllfisli Res.
7:25-31.
Barber. B. & R. Mann. 1991. Triploid Crassostrea virginica (Gmelin,
1 791 ) grow faster than diploids but are equally susceptible to Perl<insus
mariniis. J. Sliellfisli Res. 1 0:445-450.
Bellanca. M. A. & D. S. Bailey. 1977. Effects of chlorinated effluents on
aquatic ecosystems in the lower .lames River. J. Water Poll. Control
Fed. 49:639-645.
Burreson, E. M. 1991. Effects oi' Perkinsus marimis infection in the East-
94
Delaney et al.
em oyster. Crassostrea virginica: Susceptibility of native and MSX-
resistant stocks. / Shellfish Res. 10:417-423.
Carpenter, J. H. & D. L. Macalady. 1975. Chemistry of halogens in sea-
water. In: R. L. Jolley. ed. Water Chlorination: Environmental Impact
and Health Effects, vol 1. Ann Arbor. Michigan: Ann Arbor Science
Publishing Inc. pp. 161-180.
Carpenter, J. H., C. A. Smith & R. G. Zika. 1980. Reaction products from
the chlorination of seawater. In: R. L. Jolley. ed. Water Chlorination:
Environmental Impact and Health Effects, vol. 3. Ann Arbor, Michi-
gan: Ann Arbor Science Publishing Inc. pp. 379-385.
Chu, F. E. & K. H. Greene. 1989. Effect of temperature and salinity on in
vitro culture of the oyster pathogen Perkinsiis maiiinis (Apicomplexa:
Perkinsea). J. Invert. Path. 53:260-268.
Crosby, M. P. & C. F. Roberts. 1990. Seasonal infection intensity cycle of
the parasite Perkinsus marimis (and an absence of Haplosporullum
spp.) in oysters from a South Carolina salt marsh. Dis. Aqua Org.
9:149-155.
Davis. W. P. & D. P. Middaugh. 1977. A revised review of the impact of
chlorination processes on marine ecosystems: Update 1977. In: R. L.
Jolley. ed. Water Chlorination: Environmental Impact and Health Ef-
fects, vol. I. Ann Arbor. Michigan: Ann Arbor Science Publishing Inc.
pp. 283-310.
Delaney, M. A.. Y. J. Brady. S. D. Worley & K. L. Huels. 2002 submitted
for publication. The acute toxicity of two N-halamine disinfectants to
larval oysters Crassostrea virginica (Gmelin).
Elston, R. A. 1990. MoUusk diseases: Guide for the shellfish farmer. Wash-
ington Sea Grant Program. Seattle, Washington: University of Wash-
ington. USA. pp. 1-77.
Ford. S. E. & H. H. Haskin. 1987. Infection and mortality patterns in the
strains of oysters Crassostrea virginica selected for resistance to the
parasite Haplosporiclium nelsoni (MSX). J. Parasitology. 73:368-376.
Ford, S. E.. A. J. Figueras & H. H. Haskin. 1990. Influence of selective
breeding, geographic origin and disease on gametogenesis and sex
ratios of oysters, Crassostrea virginica, exposed to the parasite Hap-
losporidium nelsoni (MSX). Aquaculture 88:285-301.
Galtsoff, P. S. 1946. The reaction of oysters to chlorination. Research
Report 1 1. United States Department of the Interior. Fish and Wildlife
Service.
Galtsoff. P. S. 1964. The American Oyster. Crassostrea virginica. Gmelin.
Fisheries Bulletin of the U.S. Fish and Wildlife Service. Washington.
DC: Government Printing Office, pp. 1^83.
Goggin, C. L.. K. B. Sewell & R. J. G. Lester. 1990. Tolerances of Per-
kinsus spp. (Protozoa, Apicomplexa) to temperature, chlorine and sa-
linity. J. Shellfish Res. 9:145-148.
Hargis, W. J., Jr. & D. S. Haven. 1988. Rehabilitation of the troubled oyster
industry of the lower Chesapeake Bay. J. Shellfish Res.7:n\-279.
Haskin. H. H. & J. D. Andrews. 1988. Uncertainties and speculations about
the life cycle of the eastern oyster pathogen Haplospondiiiin nelsoni
(MSX). Am. Fish. Soc. Special Publication 18:5-22.
Kong. L. I.. J. J. Swango. B. L. Blagburn. C. M. Hendrix. D. E. Williams
& S. D. Worley. 1988. Inactivation of Gianlia lamhlia and Gianlia
canis cysts by combined and free chlorine, .^pp. Environ. Microbiol.
.54:2580.
La Peyre. J. F. & M. Faisal. 1995. Improved method for the initiation of
continuous cultures of the oyster pathogen Perkinsiis nuiriniis (Api-
complexa). Trans. .Am. Fish. Soc. 124:144—146.
Macalady, D. L., J. H. Carpenter & C. A. M. Moore. 1977. Sunlight-
induced bromate formation in chlorinated seawater. Science 195: 1335-
1337.
Mackin. J. G. & J. L. Boswell. 1956. The life cycle and relationships of
Dermocystidium mariimin. Proc. Nat. Shellfish As.soc. 46:112-115.
Mackin. J. G. & S. H. Hopkins. 1962. Studies on oyster mortality in
relation to natural environments and to oilfields in Louisiana. Publi-
cation of the Institute for Marine Science en the University of Texas
7:1-131.
Menzel, R. W. & S. H. Hopkins. 1955. The growth of oysters parasitized
by the fungus Dermocystidium marinum and by the trematode Buceph-
alus cuculus. J. Parasitology 41:333-342.
Newell, R. I. E. 1985. Physiological effects of the MS.X parasite Haplospo-
ridiiim nelsoni (Haskin, Stauber, and Mackin) on the American oyster.
Crassostrea virginica. J. Shellfish Res. 5:91-96.
Paynter, K. T. & E. M. Burreson. 1991. Effects of Perkinsus mariiuis
infection in the Eastern oyster. Crassostrea virginica II: Disease de-
velopment and impact on growth rate at different salinities. J. Shellfish
««. 10:425-431.
Perkins, F. O. 1969. Electron microscope .studies of sporulation in the
oyster pathogen, Minchinia costalis (Sporozoa; Haplosporidia). J.
Parasitol. 55:897-920.
Ray, S. M. 1952. A culture technique for the diagnosis of infections with
Dermocystidium marinum. Mackin. Owen and Collier, in oysters. Sci-
ence 166:360-361.
Roberts, M. H., Jr.. R. J. Diaz. M. E. Bender & R. J. Huggett. 1975. Acute
toxicity of chlorine to selected estuarine species. J. Fish. Res. Canada
32:2525-2528.
Roberts, M. H and R. A. Gleeson. 1978. Acute toxicity of bromochlori-
nated seawater to selected estuarine species with a comparison lo chlo-
rinated seawater toxicity. Mar. Environ. Res. 1:19-30.
Stewart, M. E., W. J. Blogoslawski, R. Y. Hsu & G. R. Hetz. 1979.
By-products of oxidative biocides: Toxicity to oyster larvae. Mar. Poll.
Bull. 10:166-169.
Tsao. T.. D. E. Williams. C. G. Worley & S. D. Worley. 1991. Novel
N-halamaine disinfectant compounds. Biotechnol. Prog. 7:60-66.
White, M. E.. E. N. Powell, S. M. Ray & E. A, Wilson. 1987. Host-to-host
transmission of Perkinsus marinus in oyster {Crassostrea virginica)
populations by the ectoparasitic snail Boonea impressa (Pyramedel-
lidae). J. Shellfish Res. 6:1-5.
Wong. G. T. F. & J. A. Davidson. 1977. The fate of chlorine in sea water.
Water Research 11:971-978.
Jounnil of Shellfish Research. Vol. 22, No. 1, 95-W. 2003.
HATCHERY REARING OF THE BLACK SCALLOP, CHLAMYS VARIA (L.)
A. LOURO. J. P. DE LA ROCHE. M. J. CAMPOS, AND G. ROMAN*
lusiituto Espauol de Occauograjia. Centra Uceaiiognifico de A Conirui, FO Box JJU,
15080 A Conma. Spain
ABSTRACT Thi.s work describes methods used for conditioning, spawning, and growing larvae of Clihiiny. vuria in hatcheries and
the results obtained. Conditioning in winter results in fast ripening. Oocytes are easily obtained by injecting serotonin. Different
antibiotics were tested and the results compared. Different systems for setting were compared. C.vuria prefers flat surfaces rather than
monofilament as settlement substrate.
KEY WORDS: Chlaniys vuiiii. hatchery, conditioning, spawning, larval culture, settlement, antibiotics
INTRODUCTION
Worldvv ide production of pectinids has increased spectacularly
in recent years, rising from 200.000 t in 1970 to 1.7 million t in
1996. The rise is largely the result of an increase in production of
these shellfish by aquaculture. which accounts for W/e of the total
production (Bourne 2000).
In Spain, as in the rest of Atlantic Europe, Pectcn nui.xiiniis is
the most commercially valuable of the pectinid species exploited;
however, experiments have recently been peiformed to assess the
possibility of cultivating smaller pectinids. such as Aequipecteii
opercuUihs (Roman et al. 1999) and Chlaniys varia (Acosta &
Alvarez 1990. Acosta et al. 1990. Roman 1991),
Chlamys varia is found in the eastern Atlantic, ranging from
southern Norway to Senegal and also in the Mediterranean (Ansell
et al. 1991, Brand 1991). It displays rhythmic consecutive her-
maphroditism; most younger/smaller specimens are males that un-
dergo a gradual sex change so that most older animals are females
(Lubet 1956. Lucas 1965. Reddiah 1962. Burnell 1983),
This species is relatively scarce in Spain. It is therefore rarely
sold commercially, and there is very little information available
about its biology and ecology. However, the potential for culturing
the species in Galicia is presently being considered. Methods of
obtaining gametes have been determined (Roman & Fernandez
1990). and spat have been cultivated in suspension from rafts
(Acosta et al. 1990); Parada et al. (1993) provide information on
the reproduction of C varia cultivated in suspension. In Galicia
the use of collectors to capture spat in natural environments has
proven unsuccessful (Roman et al. 1987. Ramonell et al. 1990) and
therefore spat production must be conducted in hatcheries. Hatch-
ery cultivation of this species has been described by Burnell
(1983), Le Pennec and Dis-Menguss (1985, 1987), Acosta and
Alvarez (1990), and Roman (1991).
The aims of the present study were to investigate ( I ) the larval
behavior of Chlaniys varia under the standard conditions estab-
lished at the Centro Oceanografico de A Coruna (COAC), for the
culture of P. nia.\imiis larvae, as summarized below; (2) the effect
of different antibiotics on larval growth and survival: and (3) the
behavior of the larvae at settlement, with the aim of optimizing the
culture methods to increase the yield of spat.
At the COAC. culture of P. ntaxinuis larvae has been conducted
intermittentlv since 1976. with some modifications to the tech-
niques described by Roman and Perez (1979) and Roman (1986.
19911,
The use of antibiotics in larval cultures is controversial. In
general in Europe pectinid larvae cannot be consistently cultivated
without chloramphenicol (Gonzalez & Roman 1983. Samain et al,
1992, Torkildsen et al. 2000). the use of which is presently pro-
hibited by the EU. Other antibiotics must therefore be used com-
mercially.
During settlement of pectinid larvae, mesh bottomed cylinders,
or collectors made of different materials are often used. Pearce and
Bourget ( 1996) have reviewed the use of different materials for the
settlement of competent spat of various pectinid species, although
no reference is made to hatchery rearing of C. varia. Only Rod-
house and Burnell (1979) mention settlement preferences of C.
varia on undersurfaces or on shaded areas in sections, of PVC
slats, in laboratory experiments.
MATERIALS AND METHODS
Conditioning
*Corresponding author. Tel: +34 981 205362; Fa.\: -i-34 981 229077;
E-mail: guillermo.roman@co.ieo.es
Adult C varia of between 30 and 50 mm in height were trans-
ported from the sea to the COAC and conditioned from the end of
December 2000 until March 2001. The trial started when scallops
were totally spent. Scallops were placed in tanks (180 x 50 x 30
cm), through which sea water flowed at a rate of 6 L min"' at
ambient temperature ( 12-14°C), An average number of 21.8 x 10''
cells day"' of Skelelonema costatiini. 13.7 x 10'' cells day"' of
Tahitian I.uichiysis ajf. galbana and 14.0 x 10'' cells day"' of
Pavlova lutheri were added to the circulating sea water using a
dosing pump for density of 183 cells/|jiL. Males and females were
kept separately after their sex was established by microscopic ex-
amination of gonad samples taken by needle puncture.
Stimulation
When females were observed to have well-developed gonads
(swollen appearance and color range between white, cream or
yellow), they were injected intramuscularly with 0.2 mL of 0.2
mM serotonin (Roman & Fernandez 1990). Once spawning began.
8 to 10 males were injected, and the sperm suspension from vari-
ous specimens was mixed. The oocytes were sieved (100-(JLm
mesh) to remove large particles and feces. The number of oocytes
shed by each female was counted using a l-mL Gallemkamp
counting cell ( Sedge wik-rafter S50). Sperm suspension was added
to the containers in which the oocytes were held, so that there were
approximately five sperm per oocyte (Gruffydd & Beaumont
1970).
95
96
LOURO ET AL.
Inctihalion
Incubations were performed in 130-L conical-bottomed fiber-
glass tanks containing 0.45-|jim filtered sea water at 16-18°C with
slight aeration, for 3 days. Food was added on the second day (25
cells fj-L"' of a 1:1 mixture of Tuhitian /. aff. galhiuui and P.
lutheri) and on the third day the tanks were emptied and the larvae
collected in 60-|jim mesh sieves. Larvae of normal appearance
were counted and the hatch yield was calculated. Three ranges of
incubation density (<6; 6-10; >10 eggs/niL) was tested.
Lanal Culture
Larvae were cultured in 150-L tanks containing 0.45-p.m fil-
tered sea water at ambient temperature (16-18°C) at an initial
density between 0.5 and 8 larvae niL^'; 8 mg L~' of chloramphen-
icol was added, and a mixed diet of 50 cells |xL"' of Tahitian /. aff.
gathana. and P. Iiillieri (1:1) was provided. The water was changed
three times a week and the mesh size of the sieve used to retain the
larvae was increased depending on the si/e of the larvae; each time
the water was changed a sample of larvae retained was measured.
Larvae reached a final density of less than 1 larva niL"' at the time
of settlement. When competent pediveliger larvae appeared, the
culture was 140-|j.m mesh sieved. If the number of pediveligers
with eye spots was greater than 509c. they were placed in settle-
ment systems.
Effect of Different Antibiotics
Larvae were cultivated at three different treatments: chloram-
phenicol (8 mg L"' ). penicillin plus streptomycin (.^0 mg L"' -i- 50
mg L" ' ). and erythromycin { 8 mg L ' ) and no antibiotic as control
from hatching until settlement. The number of settled larvae was
counted for each treatment. All treatments were carried out in
duplicate.
Larval Sultleiiunt Systems
Three trials were perfonned with C. varia using two settlement
systems, i.e.. the traditional and the modified system. These two
settlement systems were compared in the first experiment. The
traditional system, consisting of a PVC cylinder that was 43 cm in
diameter and 40 cm in height with a 140-|ji,m mesh base through
which water was circulated in an upwelling system, was placed in
a 150-L tank. The method developed at the COAC (modified
system) using artificial seaweed as a settlement substrate was pre-
pared in another tank of the same size. A total of 172.500 pedi-
veliger larvae were added to each tank. Water was changed by
displacement. Food was added daily according to larval culture
and 5. costatum was included in the diet.
The effect of the substrate and the density of pediveligers on
settlement was investigated in a second trial. The traditional sys-
tem was used, but with a settlement substrate also provided. Nine
140-p.m mesh-bottomed cylinders 25 cm in diameter and 19 cm in
height (1983 cm~ internal surface area) were placed in 200-L ca-
pacity tanks (180 x 50 x 30 cm). Three larval densities (10.000.
20.000. and 30,000 larvae/mL) and two settlement substrates (ny-
lon monotllament, artificial seaweed, no substrate control) were
used.
In the third trial, different settlement substrates were tested. For
this, collectors comprising of artificial seaweed, nylon monofila-
ment filling and scallop shells were placed in a 400-L tank along
with 312.125 pediveliger larvae. The numbers of spat on each
substrate and on the tank walls were determined after approxi-
mately 45 days.
RESULTS
Conditioning
After 6 or 7 wk on the conditioning system, scallops were
observed to have swollen gonads, from which viable gainetes were
obtained after stimulation of spawning.
Stimulation
Scallops were artificially stimulated by serotonin injection, in
January. February, and March, and gametes were obtained on each
occasion. A total of 58.3'7f of the females and 80.0Vr of the males
responded to serotonin stimulation. The time needed to obtain
sperm and oocytes ranged between 7 and 43 min and 9 and 52 min,
respectively. An average number of 0.6 x 10^ (range: 0.05 x 10'-
2.4 - 10'\ n = 16) oocytes were obtained from each female; the
mean diameter of the oocytes was 68.8 ixm ± 1.9 (SD).
Incubation
Incubation yields for three eggs density ranks were 25.5% (0-5
eggs/mL, 11 = II): 34.1% (5-10 eggs/niL, n = 5); and 31.8%
(>I0% eggs/mL, /; = 6). Statistical differences were not found
between them (analysis of variance, P > 0.05). Mean size of larvae
D obtained was 1 10.28 |xm ± 2.61.
Standard Culture
Larval development (until 50% of the larvae developed eye
spots) lasted an average of 19.3 days ± 2.0 (/? = 16): 8 days after
the .spawning (larvae size = 134 (xm ± 1 a purple spot, which is
characteristic of this species, appeared on the dorsal posterior re-
gion of the larvae. Although larvae with eye spots may appear after
13 days, the proportion did not reach 20% until day 17 (larvae size
= 194. 1 p.m ± 13. 1 ). At the end of the culture period, the average
yield of pediveliger larvae was 31.2 ± 17% (larvae size = 21 1.8
240
5 10 15
Days after spawning
20
Figure 1. Larval growth of Clilainys varia (mean ± SD of 16 lartal
cultures).
Hatchery Culture of Black Scallop
97
TABLE 1.
Elfett of different antibiutics on lar>al yields.
Percent
Pediveliger
Percent
Settlement
Cloramphenicol
Penicillin + streptomycin
Erythromycin
Control
73.2 ±5.1
71.7 ±9.3
83.6 ± 4.5
78.0 ±8.3
10.1 ±4.9
10.8 ± 1.4
10.0 ±3.3
1.7 ± 1.0
|xm ± 9.9). The rate of growth from hatching initil the final day of
culture was 5.3 (j.m day"' (Fig. 1 ).
Effect of Different Antibiotics
The percentage .survival of the larvae, at the lime of recording
50% Vi'xlh eye spots, exceeded 709(- in all treatments, including the
control in which no antibiotics were used (Table I). However.
during settlement, \.19c larvae settled compared >\(Y/c for the
antibiotic treatments.
Settlement
First Trial
Similar spats settlement was recorded in the tanks in which
artificial seaweed and mesh bottomed PVC cylinders were used
(30.1% and 30.6%, respectively) and 31.907 and 52,7(10 spat were
obtained, respectively. More spat settled on the sides of the cyl-
inder than on the mesh bottoin. In the tank containing artificial
seaweed, most spat settled on the walls of the tank. Although the
spat on each substrate were not counted, there was a marked pref-
erence for vertical walls in both cases.
Second Trial
Effect of substrate and density of pediveligers on settle-
ment. The number of spat settled in each cylinder was deter-
mined, the numbers that settled on the walls and the substrates
provided were counted separately. The results are shown in Table
2. Most of settlement took place on the walls.
Third Trial
Settlement in 400 L capacity tank with various sub-
strates. The results are showed at Table 3. A total of 19.7% spat
settled were recorded. The higher settlement was on tank walls
(12.7%) with preference on bottom (Table 3).
DISCUSSION
Cultivation of C. variii larvae was performed using the tech-
niques developed over se\ eral years at the COAC for cultivating P.
inaximus (Roman, unpublished data). However, C. varia behaves
differently from P. inii.xiiiiiis. The most important differences were
associated with settlement and effect of antibiotics. At the COAC,
P. niaxiiniis larvae have not been successfully cultivated without
antibiotics (Gonzalez & Roman 1983. Ruiz 1996), and to date,
artificial seaweed has been found to be the best settlement sub-
strate for this species (Roman, personal communication). In con-
trast, C. varia can be cultivated to pediveliger successfully without
antibiotics and artificial seaweed was not a particularly good
settlement substrate for this species, the larvae preferring to settle
on the tank walls.
Part of the standard cultixation method of C. varia involves
discarding batches in which the oocytes are not spherical or in
which there is a low hatching rate (<10%'). Not all times of the year
are suitable for obtaining good quality larvae and hatcheries do not
have unlimited space, therefore when larvae are available the best
possible production rates must be obtained. Early removal of
batches of poor quality larvae allows culture of other batches ob-
tained from different spawns. With this method, time and money
are saved and better average yields are obtained, as cultures that
would probably die are eliminated.
Conditioning of C. vcirin during the winter months allows vi-
able gametes to be obtained from January onwards, thereby bring-
ing forward the natural spawning times, which usually take place
in spring and early summer (Parada et al. 1993). Unlike other
pectinid species that have been cultivated at the COAC (P. maxi-
iniis. P. jacobaeiis. and Aeqiiipeclen opcrciilaris) C. varia matures
quickly during the conditioning period (4-5 wk) and gametes are
obtained using serotonin, allowing the timing of the larval cultures
to be planned. Furthermore, there is no risk of self-fertilization and
polyspermy is easily avoided.
The result of the response of C. varia to stimulation by sero-
tonin was similar to those described by Roman and Fernandez
(1990) although complete emptying of the gonads was not always
observed in this study.
The average number of oocytes per female obtained in the
present study (0.6 x 10". ma.ximum 2.4 x 10") was less than those
previously reported: 1.54 x 10" (Roman & Fernandez 1990). 4.5 x
10" (Le Pennec & Diss-Menaus 1985) and 5 x lO" (Burnell 1983).
TABLE 2.
Effect of larval density and settlement substrates on yield of spat of C. varia (Trial 2),
Settlement Substrate Provided
Percent of Settlement
Number of Pediveligers
Kl.OOd
2(1,(1011
30,(100
Control (mesh bottomed cylinder only)
Mesh bottomed cylinder + monofilament
Mesh bottomed cylinder + artificial seaweed
Cylinder (Vr)
Cylinder (%)
Monofilament (%)
Total (%)
Cylinder ( a )
Artificial seaweed
Total (%)
35.1
32.2
1.5
33.7
9.1
9.9
19.0
52.3
4L4
9.2
50.6
37.7
20.8
58.5
48.3
19.1
4.1
23.2
13.8
8.4
22.2
98
LOURO ET AL.
TABLE 3.
Effect of settlement substrates on yield of spat of C. varia (Trial 3)
Collector Substrate
No. of Spat
Percent Settlement
Tank wall;,
39500
Standard net filling
10455
Scallop shell
9722
Artificial seaweed
1S85
Total
12.7
3.3
3.1
0.6
19.7
However, these authors used larger adult stock than in the present
study (30-50 mm; Roman and Fernandez used specimens of be-
tween 50-75 mm and Bumell, specimens >50 mm).
The mean diameter of the oocytes was similar (average range,
68-72 |jim) to those found by Bumell (1983; 65-70 |jim) but larger
than those found by Le Peiinec and Diss-Mengus (1985; 50-60
|xm).
The density of eggs incubated did not appear to affect the yield
of larvae. This is consistent with the results of Roman and Fernan-
dez (1990). who found no significant effect of density (using be-
tween 1 and 50 eggs mL"' ) on the yields. O'Connor and Heasman
( 1995) obtained yields of up to MV/r with cultures of C. asperrlma
using a density of 100 eggs mL^' and 48% with a density of 1 egg
mL"'. Le Pennec and Diss-Mengus (1987) obtained hatching
yields of 77.7% after a period of incubation of 2 days (density 2.3
eggs mL"'), after which D larvae of 90 jxm were collected (using
sieves of mesh size 43 |jLm). Roman and Fernandez (1990) also
incubated the eggs for 48 li and obtained a yields of 17.9%.
O'Connor and Heasman ( 1 995 ) reported that 54% of C. aspenima
eggs hatched, and veliger larvae were obtained, following 2 days
incubation. With the culture technique developed at the COAC,
larvae were incubated for 3 days, then 60 |jLm mesh sieves were
used to remove small or abnormal larvae. Although the yield of D
larvae (29.2%) was lower than that reported by the authors men-
tioned above, better results were subsequently obtained because
dead or abnormal larvae, which usually appear at the end of the
incubation period, have already been removed.
The duration of the larval period of C. varia has been reported
as 22 days at 18°C (Bumell 1983). 19 days at 16-18°C (present
.study and Acosta & Alvarez 1990), and 15 days at 17°C (Le
Pennec & Diss-Mengus 1985).
The characteristic purple spot that occurs in this species, has
been reported to appear at different ages and in different sizes of
larvae: on day 4, in larvae of 120 jxm (Le Pennec & Diss-Menguss,
1985); on days 10-12, in larvae of 130-140 [j.m (Bumell, 1983);
and on day 8, in larvae of 134 |j.iii, (present study).
Larvae with eye spots appeared from day 13 onwards. In the
present study, 20% of the larvae had eye spots on day 17 (average
size of larvae, 194.1 p.m). Acosta and Alvarez (1990) detected the
pigmentation on day 14 (161.7 fxm). whereas Burnell (1983) de-
tected it in 2()-day-old larvae (200 iJiml.
Similar growth rates have been reported: 5.3 (xm day"' (present
.study), 4.8 |j.m day"' (Acosta & Alvarez 1990), and 5.3 |j.m day"'
(Burnell, 1983). all of which are much lower than that reported by
Le Pennec and Diss-Mengus (1987; 10 (jliii day"').
The larval culture yield obtained (31.2%' pediveliger larvae,
average size 211.8 (jim) was lower than those obtained by Le
Pennec and Diss-Mengus (1985, 1987; of between 65.5% and
70%. of larvae of 210 ixm). Using the same conditions, Burnell
(1985) did not obtain more than 4% survival of larvae of size
215 |j.m.
Despite the fact that few studies have been made of this species,
there is considerable variation in the results obtained by different
authors. This may be because of genetic differences or more prob-
ably, to different culture conditions, such as the quality of the
gametes or the diet. De la Roche (pers. com.) cultivated C.voria
larvae obtained from adults originating from Malaga and from
Galicia simultaneously and did not observe any differences in the
diameter of the oocytes, the age and size at which the pigmented
mark appeared, size at the time of appearance of the eye spot or
growth rate. Of the studies compared, the best results (in terms of
growth rale and yields), were obtained by Le Pennec and Diss-
Mengus (1985, 1987), possibly because of the diet provided, which
included diatoms, and to better conditioning conditions.
It appears that antibiotics are necessary for successful cultiva-
tion of pectinid larvae but not all give good results. Chloramphen-
icol appears to give the most consistent results. Uriarte et al.
(2001) reported higher growth and survival rates in Argopecten
purpunitiis using chloramphenicol at doses of 2 and 8 mg L" than
without the antibiotic. Mendes et al. (2001 ) obtained survival rates
of 20-25% in cultures of Nodipecten nodosus using clorampheni-
col. in contrast with almost total mortality on using florphenicol.
Ruiz (1996) reported high mortality in Pecten maximiis larvae
cultured with erythronnycin and high rates of survival with tetra-
cycline and triniethoprim plus sulphamethoxazole. Gonzalez and
Romi'in (1983) reported no yield of Pecten maximus larvae cul-
tured without antibiotics, in contrast to cultures in which chloram-
phenical was used at a concentration of 2,5 mg L"'. Samain et al.
(1992) found much higher survival and growth rates einploying
antibiotics and Torkildsen et al. (2000) obtained larval yields of
30% when chloramphenicol was added to the cultures.
The percentage of settlement was variable in the different cul-
tures [30% (trial I), between 19 and 58% (trial 2). and 20% (trial
3); approximately 10% in the cultures conducted with different
antibiotics). This variability may have been due to intrinsic factors,
but there were also variations within the same culture batches,
depending on the quality of the substrates provided (extrinsic fac-
tors). It is clear that C. varia prefers to settle on the tank walls than
on nylon monofilament. O'Connor and Heasman (1994) found that
Clilamys asperrima also preferred the tank bottom and walls to the
collectors provided for settlement. C. varia showed a preference
for the more sheltered, poorly lit areas of the collectors (Rodhouse
& Bumell 1979). However, in experiment 3 of the present study,
we found a very low settlement rate on the scallop shells, despite
the fact that they were hung with the concave part of the shells
facing downwards, an arrangement which should have provided
the most sheltered conditions in the tank.
Although improvements in conditioning (quality of gametes),
larval diet and the substrate and settlement conditions must be
made, hatchery culture of C. varia larvae is possible, and com-
mercially viable numbers of spat can be obtained, which would
allow development of an industry dedicated to the production of
this species.
ACKNOWLEDGMENTS
This work was financed by FEDER, project IFD 1997-0201-
C()3-()l. The autliors thank .luan Feniandez-Feijoo and Carmen
Vazquez.
Hatchkry Culture of Black Scallop
99
LITERATURE CITED
AL'Osta. C. P. & M. J. Alvarez. 1990. Cullivo de /.amburina. I. Experiencias
en cultivo larvario. Adas III Congreso Nac. Acuicult. 445^M9.
Acosta, C. P., G. Roman & M. J. Alvarez. 1990. Cultivo de zamburina
(Chlamys varia). IL Preengorde en batea. Adiis III C(yiii;re.\o Ncic.
Acuicul. 533-538.
An.sell, A.. J. Mason & J. C. Dao. 1991. Three European scallops: Pcctcn
maxiiiiKs. Aequipeclen nperculuiis and Chlamys varia. In: S. Shum-
way. editor. Scallops: biology, ecology and aquaculture. New York:
Elsevier, pp. 715-751.
Brand, A. 1991. Scallop ecology: distributions and behaMoiir, In: S, Shuni-
way. editor. ScaUop.s: biology, ecology and aquaculture. New Y'ork:
Elsevier, pp. 517-579.
Bourne, N. F. 2000. The potential for scallop culture- the ne\t niillcnniuni.
Aquacidlure Int. 8:113-122.
Burnell. G. M. 1983. Growth and reproduction ot the scallop Chlamys
varia (L.) on the west of Ireland. Ph.D. Thesis. National University of
Ireland, Galway. 295 pp.
Gonzalez, G. & G. Roman. 1983. Larval culture of the scallop {Peclen
ma.ximtis). 4"' Pectiniil Workshop. Aberdeen, Scotland (mimeo).
Gruffydd, LL. D. & A. R. Beaumont. 1970. Determination of the optimum
concentration of eggs and spermatozoa for the production of normal
larvae in Pecten maximus (Mollusc, Lamellibranchia). Hflf^oUinder
wiss, Meeresunters. 20:486—497.
Le Pennec, M. & B. Diss-Mengus. 1985. Rearing of Chlamys varia in
commercial hatchery. 5''' Interiuilional Puciiiuil Wurk.shop. A Coruna,
Spain (mimeo).
Le Pennec, M. & B. Diss-Mengus. 1987. Aquaculture de Chlamys varia
(Ly. donnees sur la biologic de la larve et de la posllarve. Vie Marine
8:37-42.
Lubet. P. 1956. Recherches sur le cycle sexuel et laemission des gametes
chez les mytilides et les pectinides (Mollusques Bivalves). Rev. Trav.
Inst. Peches Marit. 23:387-548.
Lucas, A. 1965. Recherche sur la sexualite des mollusques bivalves. Thesis
d'Etat. Universite de Bretagne Occidentale, Brest. 136 pp.
Mendes de Bern, M., G. S. Rupp. A. Pereira & C. R. Poll. 2001. Effect of
antibiotics on larval survival and microbial levels during larval rearing
of the tropical scallop Nodipecten nodosus. 13"' International Pectinid
Workshop. Coquimbo. Chile. Book of Abstracts.
O'Connor, W. A., M. P. Heasman, A. W. Frazer & J. J. Taylor. 1994.
Hatchery rearing the doughboy scallop, Chlamys asperriiiia (Laniark).
Memoirs Queensland Museum 36:357-360.
O'Connor, W. A. & M. P. Heasman, 1995. Spawning induction and fer-
tilisation in the doughboy scallop, Chlamvs asperrimu. Aipiiuiiliure
136:117-129.
Parada, J. M., M. J. Cancelo, A. Fernandez & A. V. Guerra. 1993. Com-
portamiento reproductivo de la zamburifia {ChUunys varia L.) cultivada
en batea en Galicia (NO de Espana). .Actas IV Congreso Nac. Acuicult.
317-322.
Pearce. M. & E. Bourget. 1996. Settlement of larvae of the giant scallop.
Placopecten inagellanicus (Gmeiin) on various artificial and natural
substrata under hatchery-type conditions. Aquaculture 141:201-221.
Ramonell, R., G. Roman, C. P. Acosta & M. Malvar. 1990. Captacion de
semilla de pectfnidos en colectores: resultados de la campafia de
prospeccidn en Bueu (Ria de Pontevedra, Galicia) en 1988. In: A.
Landi'n & A. Cervino, editors. Actus III Congreso Nac. Acuicult pp.
439-444.
Reddiah, K, 1962. The sexuality and spawning of manx pectinids. J. Mar.
Biol. Ass. UK. 42:683-703.
Rodhouse. P. G. & G. M. Burnell, 1979. In situ studies on the scallop
Chlamys varia. In: J. C. Gamble & J. D. George, editors. Progress in
underwater science. 4 (NS). England, Pentech Press Ltd,, pp, 87-97,
Roman. G. & A. Perez. 1979. Cultivo de larvas de vieira. Pecten maximus
(L.) en laboratorio, Boletin del Inst. Espa. Oceano. Num 223.
Roman. G. 1986. Larvae rearing of bivalve molluscs. In; B. Loix, editor.
Production in marine hatcheries. Rovinj-Zadar (Yugoslavia): 10-28
Fef. 1986 MEDRAP, pp. 10-28.
Roman, G., F. Fernandez-Cortes. C. P. Acosta & E. Rodriguez-Moscoso.
1987. Primeras experiencias con colectores de pectfnidos en las rias de
Arosa y Aldan. In: A. Landi'n & A. Cervino. editors. Actas II Congreso
Nac. Acuic. Cuademos marisqueros: Santiago de Compostela, pp. 375-
380.
Roman, G. & I. Fernandez. 1990. Metodos de obtencion de gametos de
pectinidos para su cultivo larvario. In: A. Landi'n & A. Cervirio. editors,
Actas III Congreso Nac. Acuicult.: 433-438.
Roman, G. 1991. Fisheries and aquaculture. Spain. In: S. Shuniway, editor.
Scallops: biology, ecology and aquaculture New York: Elsevier, pp.
753-762.
Roman, G.. M. J. Campos, C. P. Acosta & J. Cano. 1999. Growth of the
queen scallop (Aequipecten opercularis) in suspended culture: influ-
ence of density and depth. Aquaculture 178:43-62.
Ruiz. C. M. 1996. Ecologia microbiana en cultivos larvarios de Pecten
ma.ximus (L.) Tesis doctoral. Univ. Santiago de Compostela.
Samain. J. F., C. Seguineau. J.-C. Cochard. F. Delaunay. J. L. Nicholas. Y,
Marty, R. Galois, M. Mathieu & J. Moal. 1992. What about growth
variability for Pecten ma.ximus production? Oceanis 18:49-66.
Torkildsen, L., O. B. Samuelsen, B. T. Lunestad & 0. Bergh. 2000. Mini-
mum concentrations of chloramphenicol, tlorfenicol. trimethoprim/
sulfadiazine and tlumequine in seawater of bacteria associated with
scallop [Pecten ma.ximus) larvae. Aquaculture 185:1-12.
Uriarte. I.. A. Farias & J. C, Castilla. 2001. Effect of antibiotic treatment
during larval development of the Chilean scallop Argopecten purpuru-
tus. Aquacult. Eng. 25:139-147.
Journal of Shellfish Research. Vol. 22. No. I. I()1-I(W, 2()(1.V
EFFECT OF DEPLOYMENT DATE AND ENVIRONMENTAL CONDITIONS ON GROWTH
RATE AND RETRIEVAL OF HATCHERY-REARED SEA SCALLOPS, PLACOPECTEN
MAGELLANICUS (GMELIN, 1791), AT A SEA-BASED NURSERY
LORELEI A. GRECIAN,' G. JAY PARSONS,'* PATRICK DABINETT,^ AND
CYR COUTURIER'
'Fisheries and Mciriiw Inslittite. Memorial University of Newfoundland. P.O. Bo.x 492U. St. John's,
Newfoundland, Canada AIC 5R3 and 'Department of Biology. Memorial University of Newfoundland,
St. John's. Newfoundland. Canada AIC 5S7
ABSTRACT The effect of date of deployment on jti'owth ;ind subsequent retrieval of hatchery-reared scallop spat from a land-based
hatchery to a sea-based nursery was studied to provide information for management of juvenile-size scallops, ranging from 1.4-7.0 mm
ui shell height. The objective of this study was to determine the optimal time period for spat deployment to a sea-based nursery to yield
commercially acceptable growth rates and retrieval (scallops reinaining after mortality and loss through nets). Spat of the same size
class and stocking density were deployed over five consecutive 16-23 day intervals beginning in August 19^7. Environmental factors
were monitored weekly. Scallops were sampled after each deployment period for determination of shell height and retrieval. Scallops
were then re-deployed and sampled before (November! and after (June! the winter season. Results demonstrated that there were
significant differences in scallop growth and retrieval among the t~ive consecutive deployments. Only scallops that had been deployed
in August were greater than 7 mm by November and could be sorted and transferred to larger mesh equipment for ongrowing prior
to winter. The findings of this study demonstrated that early deployment (August! to sea-based nursery yielded high growth rates and
retrieval. Deployment later than eariy September required over-wintering in nursery culture before transfer to ongrowing. Significant
correlations were found between both growth rates and retrieval and some of the environmental parameters (e.g., temperature,
chlorophyll-a, particulate organic matter!. Acclimation to the new farm conditions inay be necessary for nursery-sized .scallops to adjust
physiologically without a major lag in growth following transfer from the hatchery to the sea.
KEY WORDS: growth, nursery culture, Phuupeelen mai;ellanieus. scallop spat, sea star
INTRODUCTION
The aim of a nursery stage in bivalve aquaculture is to foster
the development of young postmetamorphic settled animals to an
optimal size for ongrowing and handling. For scallops, the nursery
stage starts with the transitional period between a planktonic larval
phase in a well-maintained hatchery setting and a benthic postlar-
val phase where the settled spat are deployed to a sea-based nurs-
ery or to a semicontrolled land-based growout environment. Sea-
based nursery culture can be improved by determining the varia-
tion of environmental factors at the nursery site and by
manipulating the liming of the deployment of spat to nursery cul-
ture to coincide with optimal conditions.
Determining the timing of deployment at the sea-based nui'sery
is necessary to optimize growth rates of hatchery-reared Pati-
nopecleii yessoensis (Bourne & Hodgson 1991 1. Spat deployed
during optimal food density and temperatures have higher growth
rates and survival.
The window of opportunity of deployment to the sea-based
nursery can be assessed by determining growth rates and retrieval
as functions of measurable natural factors, such as water quality,
food availability, and the presence of potential predators over time.
When adequate nursery conditions are provided, growth rates and
survival are maximal, and the time scallops spend in the nursery
stage exposed to other risk factors decreases.
Growth rates of scallops vary seasonally as a result of tluctua-
tions in food supply and temperature (Kirby-Smith & Barber 1974,
Vahl 1980, Grecian et al. 2000). Growth rates of cultured P. ma-
gellanicus are highest in the summer and lowest in the winter
•■^Corresponding author. Tel: 709-778-0331; Fax: 709-778-053.'^; E-mail:
Jay.Parsons@mi.mun.ca
(Dadswell & Parsons 1991. 1992, Cote et al. 1993. Kleinman et al.
1996. Parsons et al. 2002) and show no increase during the autumn
bloom compared with summer (Emerson et al. 1994). Sea scallops
in some areas of Atlantic Canada are able to naturally produce two
cohorts annually of which the summer (June to July) cohort grows
faster than the autumn (September to October) cohort over the
entire culture period (Dadswell & Parsons 1992). Dadswell and
Parsons (1992) proposed that the higher growth rates of the first
cohort were caused by the initial exposure of spat to the summer
food conditions in the water column and a longer, more favorable
period of warmer water. Thus, in bi\al\e hatcheries and nurseries,
the early production of scallop spat is important for deployment to
nursery culture early in the summer, as is the practice for oysters.
This may result in the growth of scallop spat to a size of 7 mm or
greater by the autumn, at which time spat would be large enough
to transfer to intermediate culture gear as well as for sale to com-
mercial growers. This growing period is much shorter than waiting
until the following summer, which is the current protocol in the sea
scallop industry (Dadswell & Parsons 1991, Couturier et al. 1995).
Salinity, temperature, and predation impact survival of scal-
lops. Salinity concentrations below 13 psu and 18 psu cause mass
mortality in scallops in short-term and long-term exposures, re-
spectively (Bergman et al. 1996, Frenette & Parsons 2001). As
well, sea star predation on scallops can be significant in wild or
bottom seeded scallops (Dickie & Medcof 1963, Scheibling et al.
1991, Barbeau & Scheibling 1994a). Sea star predation on scallops
is limited in suspended nursery culture gear, unless the nursery
gear is deployed prior to the settlement and growth of sea stars
(Dadswell & Parsons 1992, Parsons 1994). Survival of post larval
scallops, Pecten ma.ximus. transferred from hatchery to nursery
was dependent on the immersion time during transfer, temperature
differential and spat acclimation to the thermal regimen of the
101
102
Grecian et al.
sea-based nursery (Christophersen 2000. Christophersen &
Magnesen 2001).
Timing of deployment of nursery-sized spat at the sea-based
nursery is critical for optimizing growth rates and survival. The
objective of this study was to determine the window of opportunity
for deployment of hatchery-reared sea scallops at a sea-based nurs-
ery that enhances growth rates and retrieval and provides avail-
ability of spat for intermediate grow-out. Based on previous re-
search on sea scallops, the hypotheses for this study are: ( 1 ) growth
will be highest in scallops deployed earliest in the summer (Au-
gust) when temperature and food availability are highest and (2)
retrieval of scallops will decline with the onset of sea star settle-
ment.
MATERIALS AND METHODS
Study Site
Scallops were deployed on a scallop farm. Shell Fresh Farms
Ltd.. based in Poole's Cove. Newfoundland. Canada. The inain
study site was located in North Bay, head of Fortune Bay. NL at
the Ladder Garden lease (47°42'N, 55°26'W).
Experimental Design and Sampling Protocol
This experiment was designed to determnie the optimal period
for the deployment of nursery-size, post larval scallops at a sea-
based nursery. Scallops were deployed over consecutive treatment
intervals from the time they were first available from the hatchery
and were large enough to be handled Ol .4 mm shell height) until
no new cohorts of spat were available in the autumn. The spat were
reared at 15°C from several spawnings undertaken at the Belleo-
ram Sea Scallop Hatchery. Belleoram, Newfoundland (47^32 'N.
55°25'W). Spat were sorted by screening and those between 1.4
and 2.0 mm in shell height were used in the study. A sample of
spat was obtained for initial shell height measurements {n = 30)
for each deployment.
Scallops were counted and deployed on five occasions at 500
spat/collector in 1.2-mm-mesh collector bags on August 4, August
22. September 7. September 26. and October 19. 1997. Two col-
lector bags, each filled with 1 m of NetronT*' (34 g). were held in
individual plastic bread trays (69 cm x .'i7 cm x 15 cm) at a 5 m
depth (Grecian et al. 2000). The number of replicate bags varied
from two to four depending on scallop spat availability. The initial
"short-term" interval duration between successive deployment and
retrieval dates ranged from 16 to 23 days and depended on site
accessibility. Each short-term deployment interval ended when the
next set of collector bags was deployed and the final short-term
deployment interval ended on November 8, 1997.
Scallop retrieval (defined as number remaining after mortality
and any potential loss through the mesh of the nets) was assessed
by counting scallops remaining at the end of each interval and
scallops were measured for shell height (/; = 30). All scallop
treatments were then redeployed and again counted and measured
for shell height before and after the winter season on November 8.
1997 and June 24, 1998, respectively. During the experiment, all
scallop treatments were handled in a similar manner.
Water samples were pumped from a 5-m depth for phytoplank-
ton identification, density and determination of total particulate
matter (TPM), particulate inorganic matter (PIM). particulate or-
ganic matter (POM), and chlorophyll-iv concentration. Tempera-
ture and salinity were measured through the water column to a
depth of 10-m using a YSI Model No. 30 S-C-T meter. Sea star
settlement was also determined (see below). Each parameter was
sampled approximately weekly during the short-term intervals
(August to November).
Immediately after water samples were collected, the phy-
toplankton samples were fixed with Lugol's Iodine and 1% form-
aldehyde. These samples then sat undisturbed for at least two
weeks to allow the seston particles to settle. The top 909^ of water
was siphoned off and its volume was measured. The remaining
volume, which contained all settled algal particles, was also mea-
sured. This concentrated volume was mixed thoroughly and 10 mL
were transferred to a 10-mL Utermohl settling chamber for over-
night settlement. The sample was analyzed visually for total num-
ber of cells and species composition using a Zeiss Axiovert 35
microscope under phase contrast at 400x magnification.
The total plankton assemblage was categorized into 8 major
groups (McKenzie. 1997). Seven of these were on the basis of size
while the final group comprised "unidentified species." The size
categories included microzooplankton including tintinnids and
ciliates (>20 iJim in diameter), autotrophic and heterotrophic di-
notlagellates (12 to 60 |jLm). prymnesiophytes comprising small
(2 to 12 (xm in diameter) spherical nanofiagellates. auto-nano-
flagellates comprising spherical flagellates from 2 to 20 |j.m in
diameter, cryptophytes comprising small (8 to 18 |j.in in length)
tear-drop shaped biflagellates. centric diatoms (12 to 30 (jim in
diameter, connected in long chains), and pelagic pennate diatoms
(30 (jLin in length, single cells). Phytoplankton were identified
according to Rott (1981).
For TPM and chlorophyll-a samples, 15 L of seawater were
pimiped from a depth of 5 m and pre-screened at 300 p.m into
separate 20-L buckets and taken to the hatchery. Water samples (4
L) for TPM were filtered onto Whatman GF/C 45-mm diameter
glass microfiber filters, which had been previously combusted in a
muffle furnace at 500°C for 4 h to remove organic matter and were
then weighed. The filters were then stored frozen at -20'C and
ultimately oven-dried at 80°C for 24 h, weighed for TPM, trans-
ferred to a muffle furnace for 4 h at 500°C, and reweighed to
determine PIM. From these weights, ash-free dry weight or POM
was calculated according to the formula TPM = POM + PIM.
An additional 4 L of seawater was filtered onto Whatman GF/C
filters for chlorophyll-n and pheopigment determination. Filters
were frozen (-20°C) for later processing according to the fluoro-
metric methods of Strickland and Parsons ( 1968) and Parrish et al.
(1995).
Sea star settlement was monitored weekly from July 15 to
November 8, 1997. by deploying strings of eight empty pearl nets
(34-cm X 34-cm square base pyramidal-shaped nets. 6-mm mesh)
weekly at the farm with retrieval after approximately two weeks.
Individual pearl nets were washed and all material greater than 250
(xm was collected on a mesh screen and preserved in 40% metha-
nol. Samples were analyzed using a dissecting microscope for
determination of numbers of sea stars present.
Data Analysis
Data were analyzed using the SPSS statistical package (Version
8.0). All percent data were arcsine-square-root transformed before
statistical analysis (Sokal & Rohlf. 1995). Differences in growth
rates and retrieval were analyzed using an analysis of variance
(ANOVA) and the post hoc Tukey's b test was used to test for
differences among treatments. Equality of means was analyzed
Effect of Deployment Time on Sea Scallops
103
using an Independent sample Mest. Pearson correlation analyses
were also performed on growth and retrieval data with the envi-
ronmental parameters. The le\el of sisznificance was set at a =
0.05.
RESULTS
Grow til Rales
Initial shell height among the replicates was not significantly
different for all dates {P > 0.01) except September 7 (One-way
ANOVA;F = 9.735. df= 2, 87, P< 0.001). This was because the
scallops in one of the replicates were from a slow-growing batch
of larvae and they were not randomly assigned among the repli-
cates for that date, hence this replicate was not used for further
analysis or in figures. The initial mean size ranged from 1.41 to
1.62 mm shell height (Fig. I).
The mean shell heights of .spat at the end of each short-temi
deployment interval were significantly different from the initial
mean shell heights (Hests. P < 0.05. Fig. I). As well, mean shell
heights at the end of each short-term interval were significantly
different among the different deployment dates and decreased
from 3.54 mm to 1.51 mm shell height (one-way ANOVA; F =
556.621, df = 4. 445. P < 0.001 ).
Growth rates declined over the short-term intervals (Fig. 2).
Significant differences were found among growth rates for the
different intervals (one-way ANOVA: F = 95.162; df = 4. 1 1. P
< 0.001). Highest growth rates occurred during the first deploy-
ment interval at I 18 |jLm d"' (SE ± 1 .3). whereas the lowest growth
rates occurred during the last interval at 3.3 p.m d"' (SE ± 0.7).
The mean growth rate of all spat deployed between August 4 to
November 8. 1997 was 43.2 pim d"' (SE ± 0.8).
Growth rates of scallops from the earliest deployment were
higher in the autumn and over winter than those from the subse-
quent deployments (Fig. 3). For scallops deployed on August 4 and
22. growth rates were high until November 8. For the same scal-
lops, growth rates from November to June 24. 1998. declined to a
level similar to that of scallops deployed from September 7. 1997.
Scallops deployed on September 26 and October 19. had lower
overall giowth rates to November [1 1.4 |xm d"' (SE± 1.1) and 3.3
p.m d"' (SE ± 0.7). respectively] and to June |2I.5 [xm d~' (SE ±
1.4) and 7.2 fjim d~' (SE ± 1.3). respectively].
04-Aug
22-Aug
07-Sep
26-Sep
Initial deployment date
Figure 1. Mean shell height of scallops deployed over five consecutive
2-Heek intervals in 1997 and on November 8, 1997, and June 24. 1998,
at Shell Fresh Farms Ltd., Poole's Cove, NL. The initial date of an
interval was the final date of the previous short-term interval, t'oni-
mon letter denotes no significant difference among mean shell heights
for each sample period iTukey's b test). Vertical bars are ±SE.
I5U
120 -
90 ■
60 ■
30 ■
■
'd -■■■■'-1.
^^* Short-term Growth i
""■•■"' Short-term Retneval 1
-3
5.
1
"■•-.T i
2
M
li
b
r *
1
n
a
100
90
80
70 '
60 S
50 1
40 i
30
20
10
04-Aug
:-Aus; 07-Scp 2b-Sep 19-Ocl
Initial deployment date
Figure 2. .Mean growth rates and retrieval of scallops over consecutive
deployment intervals at Shell Fresh Farms Ltd., Poole's Cove, NL. The
initial date of an interval is the final date of the previous interval.
Common letter denotes no significant difference in growth rates or
retrieval among intervals (Tukey's b test). \ ertical bars are ±SE.
Retrieval
Retrieval of spat at the end of each deployment interval (num-
ber remaining after mortality and loss) declined over time (Fig. 2)
and was significantly different among the different short-term de-
ployment intervals (one-way ANOVA; f = 47.129, df = 4. I I, P
< 0.001 ). Highest retrieval was obtained from spat deployed during
the first interval (97%), whereas lowest retrieval was for spat
deployed on September 26 (53%). Retrieval of spat from their
initial deployment to November 8 was not significantly different
than their retrieval after the short-term intervals (Paired t-test; t =
0.013. df = \4.P = 0.990; Fig. 3).
En vironmeiital Characteristics
Water temperature declined over the deployment periods (Au-
gust-November. Fig. 4A). Mean temperatures for the five con-
secutive deployment intervals were 14.7, 13.6. 11.3. 11.2. and
7.9°C. respectively. Spat were not acclimated from 15 C in the
hatchery to ambient seawater temperatures before sea-based de-
ployment. Salinity increased over the study period (Fig. 4A). Mean
salinity was 28.3 psu whereas the range was from 26.5 to 31.5 psu.
Chlorophyll-fl concentrations (one-way ANOVA; F = 0.544.
df = 14. 24. P = 0.881). pheopigment concentrations (one-way
-^ .Autumn
'Spring
" November Retneval
04-Aug 22-Aug 07.Scp 26.Sep 19-Ocl
Initial deployment date
Figure 3. Mean growth rates and retrieval of scallops deployed at a
sea-based nursery at Shell Fresh Farms Ltd.. Poole's Cove. NL, on five
dates in 1997 and sampled on November 8. 1997, and June 24, 1998.
Common letter denotes no significant difference in growth rates or
retrieval among intervals (Tukey's b test). Vertical bars are ±SE.
104
Grecian et al.
u
3
«
u
u
a.
20
16
12
4 ■
■»- *-
Jul
22- 5- 19- 2- 16- 30- 14- 28-
Jul Aug Aug Sep Sep Sep Oct Oct
35
30
""i
B.
a
71)
,>i
IS
c
lU
"3
5
C/2
Nov
8- 22- 5- 19- 2- 16- 30- 14- 28- 11-
Jul Jul Aug Aug Sep Sep Sep Oct Oct Nov
=5.
C
O
o
c
o
U
20
16
12
8
4
8-
Jui
• Chlorophyll
" Phaeopignients
A.
A-A A-
Jul
5-
Aug
19-
Auti
2- 16-
Sep Sep
Date
30-
Sep
14-
Oct
Oct
11-
Nov
Kisure 4. Water quality at Ladder Garden site. Shell Fresli Farms Ltd., Poole's Cove, NL, from July IS to November 8, 1997. A) Temperature
and salinity (±SE: ;i = 3), Bl seston, C) chlorophyll and pheopigments at 5 m. (TPM, total particulate matter; POM, particulate organic matter).
ANOVA; F = 0.500. df = 14. 24. P = 0.910), and POM (one-
way ANOVA; F = 0.71.5. df = 14. 21. P = 0.737) were not
significantly different o\er the duration of the study.
TPM remained constant al Ladder Garden (Fig. 4B) with
weekly mean TPM being 5.6 mg L '. POM was also constant at
Ladder Garden with a mean of 1.9 mg L '. Chlorophyll-o and
pheopigments averaged 2.4 and 10.1 mg L"', respectively (Fig. 4C).
There was a significant difference in total phytoplankton den-
sity among the weekly samples (one-way ANOVA; F = 7.084. df
= 13. 28. P < 0.001; Fig. 5). The total phytoplankton density
peaked around the middle of August, followed by a decline. The
decline was also evident when the mean total phytoplankton den-
sity was calculated for each interval (Fig. 6). The autotrophic
nanoflagellates, pelagic pennate diatoms, and dinotlagellates were
the numerically dominant groups present (Fig. 7A and B). The
species that contributed to the peak abundance were Naviciila sp..
Cliluinydoinoiuis sp.. Ochronxnias sp.. Microinonas sp. (Fig. 8A
and B). Percent abundance of phytoplankton size groups indicated
that species <5 p,m had the greatest contribution to phytoplankton
biovolume (Fig. 9).
Sea star settlement at the Ladder Garden site peaked between
September 19 and October 23 (Fig. 10). There were significant
differences in sea star settlement over the different sampling dates
(ANOVA; F = 99.674. df = 13. 336. P < 0.001). Maxiinum
i
S-Jiil 2:-Jul 5-Aug 19-Aiig 2-Sep 16-Sep 30-Sep 14-Oct 28-Ocl 1 1-Nov
Date
Figure 5. Total phytoplankton density at Ladder Garden site of Shell
Fresh Farm, Poole's Cove, NL, from ,luly 15 to November 8, 1997.
Effect of Deployment Time on Sea Scaelops
105
Deploynicnt inlenai
Fij"ure 6. Mtun density of total phvtoplankton over five intervals of
scallo|) deplovMunt on a sea-based nursery at Shell Fresh Farm,
Poole's Cove, Nl.. Intervals hejjan on Ausiust 4 and ended on Novem-
ber S. IW7. \ertieal bars are ±SE.
setllenieiit was 311) sea stars per collector per day ami mean sea
star settlenienl was 79 sea stars per collector per day.
Most environmental factors were highly correlated with growth
rates and retrieval (Tables 1 and 2). TPM and dinoflagellates were
not correlated with growth rates and TPM and PIM were not
correlated with retrievals.
DISCUSSION
Effects of Deploymiiil Dale on Growth Rates anil Retrieval
The date of transfer or deployment of scallop spat from hatch-
ery to nttrsery was a useful predictor of growth and retrieval. The
higher growth rates and retrievals in the earlier deployments were
related to several parameters in this study, where ambient tem-
5-
19-
T.
16-
30-
14-
28-
11-
Aug
Aug
Sep
Sep
Sep
Oct
Oct
Nov
A
2500 ■
2000 ■
1500 ■
1000 ■
500 ■
ri ■
/ .♦■■♦■
--■•■-
" Pelagic pennale diatoms
~ Dinoflagellates
~ Unidentified phytoplankton
~ Autotrophic nanoflagellates
U
— e—
Q
V
^fe
*f»==«^«-,.>-*.T^,
8- 22- 5- ly- 2- 16- 30- 14- 2S- 11-
Jul Jul Aug Aug Sep Sep Sep Oct Ocl Nov
J
80
60
4U 1
20
c
D
• Microzooplankton
* Prymnesiophytes
- - o - - Centric diatoms
/\
.'^ A
/^^^:^:i\
8-
Jul
22- 5- 19- 2- 16- 30- 14- 2S- 11-
.lul Aug Aug Sep Sep Sep Oct Oct Nov
Date
Figure 7. Mean density of "(.V)" four dominant and "(B)" three less
dominant groups of major plankton at Shell Fresh Farms Ltd., Poole's
Cove, NL, from July 15 to November 8, 1997.
■ Rluzosolenia sp.
- Coccolithophore sp.
Prorocentruin sp.
Choanoflagellate sp.
Sirohilidnim innninuin
Dinophysis non'egica
-J
C
Q
8- 22- 5- 19- 2- 16- 30- 14- 28- 11-
Jul .lul Aug Aug Sep Sep Sep Oct Oct Nov
Date
Figure 8. Mean den.sity of "(A)" dominant and "(B)" less dominant
plankton species that showed a declining trend over intervals of scallop
deployment at a sea-based nursery at Shell Fresh Farms Ltd., Poole's
Cove, NL, from .luly 15 to November 8. 1997.
perature and food availability and quality (species composition,
organic content and lipid characteristics inferred from literature
reports) were higher initially, then declined after early August.
Predator (sea star) abundance peaked near the second deployment
date before declining. Spat growth and retrieval from the initial
deployment demonstrated that there is an optiinum time or window
of opportunity, which could be used to maximize nursery growth.
After this period, scallops face increasing adversity in terms of
declining temperature and food quantity and quality, and increas-
ing predation and temperature shock (the difference between
hatchery and ambient temperatures). In a similar study in Southern
Norway, Pecten inaximus spat transferred from hatchery to sea-
based nursery from March to August showed increased growth and
survival during the summer when water temperatures were >10°C
and when temperature differences between the hatchery and nurs-
ery were minimal (Christophersen & Magnesen, 2001).
Temperature and food availability declined from August to
November while sea star settlement began in mid-September.
Variations in temperature and food availability were similar to
those found in other areas of Atlantic Canada, including Concep-
tion Bay, NL, and Bedford Basin, NS (Mayzaud et al. 1989, Na-
varro & Thompson 1995). In earlier studies of sea scallop aqua-
culture, temperature and food availability were the main predictors
of growth (Parsons & Dadswell 1992, Cote et al. 1993. Emerson et
al. 1994. Kleinman et ul. 1996). Likewise, sea stars are a signifi-
cant predator of scallops (Barbeau & Scheibling 1994a. b).
Changes in these parameters may best explain the variation in
growth and retrieval of the scallops over the different deployment
intervals.
A negative correlation of salinity with growth and retrieval of
scallops in the deployinent study was probably a coincidence as
106
Grecian et al.
04-Aug 22-Aug 07-Sep 26-Sep
Initial deployment date
19-Oct
H <5 Hill
■ <10nm
D <20 urn I
D<50nm
O<100nm
■ <2()0 urn
D >200 urn
B Unidentified
Figure 9. Particle size frequency distribution of planlvton at Ladder Garden, Sliell Fresh Farms Ltd., Poole's Cove, NL, over five consecutive
deployment intervals of scallops at a sea-based nursery.
the salinity tolerance range for wild juvenile sea scallops is >25
psu (Frenette & Parsons 2001), which is lower than the salinity
during the present study. The increase in salinity over the study
period reflects the decreased runoff and the increased upwelling
that occurs in the autumn in this area.
Decreases in metabolic processes due to declining temperature
may explain why reduced growth rates were observed in scallops
deployed on different dates in this study as has been found for
Pecteii fiimatiis (Cropp & Hortle 1992). Respiration rates in sea
scallops decrease with declining temperature (Shumway et al.
1988), but clearance rates are coirelated with ambient temperature
in sea scallops (MacDonald & Thompson 1986) as well as in the
eastern oyster, Crassostrea virginica and the bay scallop. Ar-
gopecleii irradians (Rheault & Rice 1996). In the present context,
reduced clearance rates would be expected to decrease food intake
and result in reduced growth.
Declining retrieval over time was correlated with deployment
temperature. This however, does not indicate that scallops died as
a direct result of decreasing temperature. Scallops are able to live
within a temperature range of-2' C to 22"C (Dickie 1958). Hence
their survival should not have been influenced by decreasing tem-
peratures per se. Christophersen and Magnesen (2001 ) found that
when Pecten maximus spat were deployed at water temperatures
>10°C, spat had up to a 4-fold increase in survival compared with
scallops deployed at temperatures <10°C. The sea scallops were
likely influenced more by the temperature difference from the
hatchery to the sea-based nursery environment than their physi-
ologic condition or predation by sea stars.
Effects of Food Variation on Growth Rates and Retrieval
Scallops deployed when Prnrocciilninu DiiuipltYsis and Nav-
icitla spp. densities were elevated exhibited higher growth rates
than scallops deployed when densities of these phyloplankton spe-
cies had declined. All these mircoalgae have been found in gut
analyses of adult scallops (Shumway et al. 1987). We found all
three species in high abundance and the first two species are con-
sidered to add greatly to the energy uptake of scallops (Shumway
et al. 1987). Cryptophyte densities also peaked during August
when growth rates were highest. Cryptophytes are rich in the fatty
acids. 22:6w3 and 20:5w3 (Volkman et al. 1989. Viso & Marty
1993) and are important for a good diet and membrane fluidity in
bivalves (Enright et al. 1986, Napolitano et al. 1992). Crypto-
phytes are a preferred alga in mixed diets and are related to growth
350
8-Jul 22-Jul 5-Aug 19-Aug 2-Sep 16-Sep 30-Sep 14-Oct 2S-0ct 11-Nov
Date
Figure 10, Mean sea star settlement al Ladder Garden lease of Shell Fresh Farms Ltd., Poole's Cove, NL, from .July 15 to November 8. 1997
(;i = 8). Vertical bars are ±.SH
Effect of Deployment Time on Sea Scallops
107
TABLE 1.
Pearson's ciirrelalion coefficients of shorl-lcrni };ro"th rates and retrieval of nursery-size scallops «ilh mean water quality parameters at a
sea-based nursery at Shell Fresh Farms ltd.. I'oolc's Cove, NL. from August 4 to November 8, 1997.
%
Sea Star
Temperature
Salinity
Chlorophyll-a
Phaeopigments
TPM
PIM
POM
POM
Settlement
Growth rate
/■ value
0.840
-0.826
0.901
0.940
-0.043
-0.573
0.700
0.773
-0.796
Signifieancc (two-tailed)
<0.001
<0.0()1
0.001
<0.001
0.4.19
0.013
0.002
0.001
<0.001
Retrieval
/• value
0.828
-0.698
0.849
0.870
0.2.1.1
-0.358
0.714
0.644
-0.890
Significance (two-tailed)
<0.001
0.002
<().()01
<0.001
0.201
0.095
0.001
0.005
<0.0()1
15 for all parameters.
in sea scallops (Shumway et al. 1985. Pairish et al. 199.5). It is
expected that scalkips exposed to a higher quality diet allowitig
adaptation to declining conditions would peifomi better than scal-
lops exposed to a lower quality diet (see Shunnvay et al. 1997).
MacDonald and Thotnpson ( 1985) found that shell growth was
higher under favorable conditions of food and tetiiperature, and
that this was site specific. Location of sea-based nursery sites
should consider food quantity and quality. However, because there
have been so few growth studies of juvenile bivalves with respect
to natural phytoplankton composition, the actual quantity and qual-
ity of food required is not l<nown (Newell et al. 1989. Parrish et al.
1995, Grant 1996). Phytoplankton is a major component of the diet
of adults (Shumway et al. 1987, Cranford & Grant 1990); however,
further research is necessary to determine the actual quantity and
quality that allow juvenile scallops to perform optimally (Shum-
way et al. 1997).
Predation Effects on Retrieval
There was an increased negative correlation between retrieval
of scallop spat and sea star settleinent during the short-term inter-
vals. Increasing sea star settlement coupled with declining sea
scallop retrieval was expected (Dadswell & Parsons 1991. 1992.
Barbeau & Scheibling 1994a. Parsons 1994). Successful predation
may be due to the similar size of the settling sea stars and scallop
spat as well as debilitation caused by significant temperature
changes between hatchery and nursery environments (Dickie 1958.
Barbeau & Scheibling 1994a). In the present study temperature
difference between hatchery and nursery progressively increased
with deployment date from to 7.1°C. Although sea star predation
may be reduced with decreasing temperature (Barbeau & Scheib-
ling 1994a). the temperature shock may have rendered spat more
susceptible to sea star predation.
Dickie ( 1958) observed a lack of mobility of scallops for about
a month when they were exposed to drops of 4-7°C in ambient
temperature, which he speculated may be detrimental if predators
are unaffected. Temperature debilitation may have coincided with
highest mortality of scallops in the present deployment study.
which was during the period of peak sea star settlement on the
culture gear.
Importance of Acclimation on Growth Rales and Retrieval
The effect of increasing differences between hatchery and at
sea nursery conditions on the performance of scallops raises con-
cerns over handling protocols. Although acclimation to different
conditions was not specifically examined in this study, a few gen-
eral observations can be made regarding its importance. Sea-based
nursery conditions were within the natural environmental ranges
for scallops; however, scallops performed increasingly poorer with
each consecutive deployment interval. Other studies have found
that sudden changes in the environmental or rearing conditions can
decrease survival and growth (Thompson 1984. Cranford & Grant
1990. Cote et al. 1993. Christophei'sen 2000. Christophersen &
Magnesen 2001 ). Acclimation of Fccleii inaximiis to lower ambi-
ent water temperature did confer a small increase in survival in
TABLE 2.
Pearson's correlation coefficients of short-term growth rates of nursery-size scallops \\ith mean phytoplankton densities at a sea-based
nursery at Shell Fresh Farms Ltd., Poole's Cove, NL, from August 4 to November 8, 1997.
.Autotrophic Centric Rhizosolenia
Total Microzooplankton Choanoflagellates Dinotlagellates Pryninesiophytes Crjptophytes nanoflaj^ellales Diatoms t'nidentifled sp.
r value
0.994
(LS.S8
0.772
-O.O.'iS
0.895
0.789
0.991
-0.635
o.yyi
0.891
Significance
(two-tailed)
<().(XII
0.0.1 1
0.001
0.837
<0.001
<0.00l
<0.(K)l
0.011
<o.ooi
<0.(XJI
Navictila Chlamydomtmas Ochromoiias Micromoilas Prnrocenlruin Choanodajiellatc Strnmbilium Pelagic Pennatf
sp. sp. sp. sp. Coccollthophore sp. sp. minimum Diatoms
/■ value 0.726
Significance
(two-tailed) 0.(K)2
0.987
<0.001
0.687
0.005
o.y 1 1
<0,00l
0.980
<0.00l
0.895
<0.00l
0.944
<0.001
0.772
0.(X)l
0.974
<0.(X)1
" = 15 lor all parameters.
108
Grecian et al.
juvenile scallops (Christophersen & Magnesen 2001 ). Mylilus cJu-
lis requires 14 days to acclimate oxygen consumption, filtration
rates and assimilation efficiency (Widdows & Bayne 1971). Hall
( 1999) observed that in sea scallops 15-21 days were required for
membrane fluidity to adjust to a temperature decrease from 13 to
5°C. The temperature and diet differentials between hatchery and
nursery may have been too great or too abrupt for scallops to
maintain optimal peiformance without the opportunity to accli-
mate, pai'ticuiarly later in the deployment season.
Implications for Halcheiy, \'iirsery and Grow out
The findings of this study provide growers with a protocol for
working with animals in a dynamic environment, under optimal
and suboptimal conditions. Hatchoy tnanagers may be able to use
our results to improve decisions on when to deploy sea scallops
and nursery managers may now have the ability to optimize
growth and retrieval of sea .scallops reared in a sea-based nursery
system and to better plan for transfer to grow out when scallop spat
reach the desired target size,
CONCLUSIONS
Growth rates and retrieval of nursery-sized scallops were in-
fluenced by time of deployment at a sea-based nursery during a
period that spanned early summer to late autumn. Highest growth
rates and retrieval of nursery-sized scallops were observed during
August and early September when the nursery site water column
was characterized by high food densities, high temperature and
low sea star settlement. However, scallops deployed in late Sep-
tember and October had low retrieval as well as low growth rates
until the following spring or later.
The ability of nursery-sized scallops to grow and survive may
be related to the differences between hatchery and sea-based nurs-
ery environments in terms of food quality and temperature differ-
entials. There is a need to detennine the nutritional requirements of
nursery-sized scallops and to practice acclimation protocols.
ACKNOWLEDGMENTS
This research was supported by the Canadian Centre for Fish-
eries Innovation and the Canada/Newfoundland Economic Re-
newal Agreement - Aquaculture Componoit. Special thanks to
staff and management at Belleoram Sea Scallop Hatchery and
Shell Fresh Farms Ltd., where research was conducted. The au-
thors thank Dr, Cynthia McKenzie from the Ocean Sciences Cen-
tre, Memorial University for assistance in plankton identification
and enumeration, Elizabeth Hatfield, Ocean Sciences Centre for
assistance in chlorophyll analysis and Guilherme Rupp and Dr.
Michael Dadswell for reviewing the manuscript.
LITERATURE CITED
Barbeau, M. A. & R. E. Scheibling. 1994a. Temperature effects on pre-
dation of juvenile sea scallops [Placopecten magellaniciis (Gmelin)] by
sea stars (Asterias vulgaris Verrill) and crabs (Cancer irroralus Say).
J. Exp. Mar Biol. Ecol. 182:27-47.
Barbeau, M. A. & R. E. Scheibling. 1994b. Behavioral mechanisms of prey
size selection by sea stars {Aslcrias vulgaris Verrill) and crabs [Cancer
irroratus Say) preying on juvenile sea scallops (Placopecten magel-
lanicus (Gmelin)). J. Exp. Mar Biol. Ecol. 180:103-136.
Bergman, C, J. Parsons & C, Couturier. 1996. Tolerance of giant .sea
scallop to low salinity exposure. Bull. .Aquacul. Assoc. Canada 96:62-
64.
Bourne, N. & C. Hodgson. 1991. Development of a viable nursery system
for scallop culture. In: S. E. Shumway & P. Sandifer, editors. An
international compendium of scallop biology and culture, vol. 1. Baton
Rouge, LA: The World Aquaculture Society, pp. 273-280.
Christophersen, G. 2000. Effects of air immersion on sur\'ival and growth
of hatchery reared great scallop spat. Aquaculture Int. 8:159-168.
Christophersen. G. & T. Magnesen. 2001. Effects of deployment lime and
acclimation on survival and growth of hatchery-reared scallop (Pecten
inaximus) spat transferred to the sea. J. Shellfish Res. 20:1043-1050.
Cote, J., J. H. Himnielman, M. Claereboudt & J. C. Bonardelli. 1993.
Influence of density and depth on the growth of juvenile sea scallops
(Placopecten magellanicus) in suspended culture. Can. J. Fish. Aquat.
Sci. 50:1857-1869.
Couturier, C, P. DabineU & M. Lanteigne. 1995. Scallop culture in At-
lantic Canada. In: A. Boghen, editor. Cold-water aquaculture in Atlan-
tic Canada. Sackville, New Brunswick: The Tribune Press, pp. 297-
340.
Crantord. P. J. & J. Grant. 1990. Particle clearance and absorption of
phytoplankton and detritus by the sea scallop Placopecten magellani-
cus (Gmelin). J. Exp. Mar Biol. Ecol. 137:105-121.
Cropp, D, A. & M. E. Hortle. 1992. Mid-water cage culture ot the com-
mercial scallop Pecten fumatus Reeve 1852, in Tasmania. Aquaculture
102:55-64.
Dadswell, M. J. & G. J. Parsons. 1991. Potential for aquaculture of sea
scallop, Placopecten magellanicus (Gmelin, 1791), in the Canadian
Maritimes using naturally produced spat. In: S. E. Shumway & P. A.
Sandifer, editors. An international compendium of scallop biology and
culture, vol 1. Baton Rouge, LA: The World Aquaculture Society, pp.
300-307.
Dadswell, M. J. & G. J. Parsons. 1992. Exploiting life history character-
istics of the sea scallop, Placopecten magellanicus (Gmelin, 1791).
from different geographical locations in the Canadian Maritimes to
enhance suspended culture growout. J. Shellfish Res. 1 1:299-305.
Dickie, L. M. 1958. Effects of high temperature on survival of the giant
scallop. J. Fish. Res. Bd. Canada 15:1189-1211.
Dickie, L. M. & J. C, Medcof. 1963. Causes of mass mortalities of scallops
(Placopecten magellanicus) in the South-western Gulf of St. Lawrence.
J. Fish. Res. Bd. Canada 20:451-482.
Emerson, C. W., J. Grant, A. Mallet & C. Carver. 1994. Growth and
survival of sea scallops Placopecten magellanicus - effect of culture
depth. Mar Ecol. Prog. Ser 108:119-132.
Enright, C. T., G. F. Newkirk, J. S. Craigie & J. D. Castell. 1986. Evalu-
ation of phytoplankton as diets for juvenile Oslrea edulis L. J. Exp.
Mar. Biol. Ecol. 96:1-13.
Frenette, B. & G. J. Parsons. 2001. Salinity-temperature tolerance of ju-
venile giant scallops, Placopecten magellanicus. Aquacul. Assoc.
Canada Spec. Publ. No. 4:76-78.
Grant, J. 1996. The relationship of bioenergetics and the environment to the
field growth of cultured bivalves. J. Exp. Mar Biol. Ecol. 200:239-
256.
Grecian, L, A,. G. J. Parsons, P. Dabinett & C. Couturier. 2000. Influence
of season, initial size, depth, gear type and stocking density on the
growth rates and percent retrieval of sea scallop, Placopecten magel-
lanicus, on a farm-based nursery. Aquaculture Int. 8:183-206.
Hall, J. M. 1999. Temporal changes in the faUy acid composition and
tluidity of gill and hemocyte membranes during thermal acclimation of
the sea scallop Placopecten magellanicus. M.Sc. Thesis. Memorial
University of Newfoundland, St. John's, Newfoundland, Canada.
Kirby-Smith, W. W. & R, T, Barber. 1974. Suspension-feeding aquaculture
systems: effects of phytoplankton concentration and temperature on
growth of the bay scallop. Aquaculture 3:135-145.
Kleinman, S., B. G. Hatcher, R. E. Scheibling, L. H. Taylor & A. W.
Hennigar. 1996. Shell and tissue growth of juvenile sea scallops (Pla-
Effect of Deploymfnt Timk on Sea Scallops
109
copecten mageUaniciis) in suspended and holloni culture in Lunenhurg
Bay. Nova Scolia. Aijuacullnie 142:73-97.
MacDonald. B. A. & R. J. Thompson. I9S5. Inlluenee ol temperature and
tood availability on the ecological energetics of the giant scallop /*/</-
copecten magethmiciis. I. Growth rates of shell and somatic tissue.
Mar. Ecol. Prog. Ser. 25:279-294.
MacDonald, B. A. & R. J. Thompson. 1986. Influence of temperature and
food availability in the ecological energetics of the giant scallop Pla-
copecten magellunicus. 111. Physiological ecology, gametogenic cycle,
and scope for growth. Mar. Biol. 93:37-48.
May/aud. P., J. P. Chanut & R. G. Ackman. 1989. Seasonal variation of the
biochemical composition of marine particulate matter with special ref-
erence to fatty acids and sterols. Mar. Ecol. Prog. Ser. 56:189-204.
McKen/ie. C. 1997. Pictorial guide to the marine phytoplankton of New-
foundland. COPE Technical Report, Ocean Sciences Centre, Memorial
University of Newfoundland, St. John's, NL (Unpublished).
Napolitano, G. E., B. A. MacDonald, R. J. Thomp.son & R. G. Ackman,
1992. Lipid composition of eggs and adductor muscle in giant scallops
[Placopecten magellanicns) from different habitats. Mur. Biol. 1 13:71-
76.
Navarro. J. M. & R. J. Thompson. 1995. Seasonal fluctuations in the size
spectra, biochemical composition and nutritive value of seston avail-
able to a suspension-feeding bivalve in a subarctic einiionment. Mar.
Ecol. Prog. Ser. 125:95-106.
Newell, C. R., S. E. Shumway. T. L. Cucci & R. Selvin. 1989. The effects
of natural seston particle size and type on feeding rates, feeding selec-
tivity, and food resource availability for the mussel Mytilus ediilis
Linneaus, 1758 at bottom culture sites in Maine. J. Shellfish Res. S:
187-196.
Parrish. C. C, C. H. McKenzie. B. A. MacDonald & E. A. Hatfield. 1995.
Seasonal studies of seston lipids in relation to microplankton species
composition and scallop growth in South Broad Cove. Newfoundland.
Mar Ecol. Prog. Ser. 129:151-164.
Parsons, G. J. 1994. Reproduction and recruitment of the giant scallop
Placopecten magellanicns and its relationship to environmental vari-
ables. Ph.D. Thesis, University of Guelph. Guelph. Ontario. Canada,
196 p.
Parsons, G. J. & M. J. Dadswell. 1992. Effect of stocking density on
growth, production, and survival of the giant scallop, Placopecten ma-
gellanicns. held in intermediate suspension culture in Passamaquoddv
Bay. New Brunswick. Aqnacultnre 103:291-309.
Parsons. G. J.. S. E. Shumway. S. Kuenstner & A. Gryska. 2002. Polycul-
lure of sea scallops (Placopecten magellanicns) suspended from
salmon cages. Aquacniture Int. 10:65-77.
Rheault, R. B. & M. A. Rice. 1996. Food-limited growth and condition
index in the eastern oyster. Crassostrea virginica (Gmelin 1791 ). and
the bay scallop. .Argopecteii irrailians irrajians (Lamarck ISI9|. /
Shellfish Res. 15:271-283.
Rott. E. 1981. Some results from phylopkniklon couniiiig intercalibrations.
Schweiz. Z. Hydrol. 43:34-62.
Sclieibling. R. E.. B. G. Hatcher, L. Taylor & M. A. Barbeau. 1991.
Seeding trial of the giant scallop (Placopecten magellanicns) in Nova
Scotia. In: J. Barret, J.-C. Dao & P. Lubet. editors. Fisheries, Biology
and Aquaculture of Pectinids. 8th International Pectinid Workshop.
22-29 May 1991. Cherbourg. France. Actes de Colloques. No. 17.
Institut Erancais de Recherche pour L'E.\ploitation de la Mer. DCOM/
SE. Plouzane. France, pp. 12,3-130.
Shumway, S. E.. J. Barter & J. Stahlnecker. 1988. Seasonal changes in
oxygen consumption of the giant scallop, Placopecten magellanicns
(Gmelin). J. Shellfish Res. 7:77-82.
Shumway. S. E„ T. L. Cucci, M. P. Lesser, N. Bourne & B. Bunting. 1997.
Particle clearance and selection in three species of juvenile scallops.
.Aqnacniturc Int. 5:89-99.
Shumway, S. E.. T. L. Cucci, R. C. Newell & C. M. Yentsch. 1985. Particle
selection, ingestion, and absorption in filter-feeding bivalves. ./. Exp.
Mar Biol. Ecol. 91:77-92.
Shumway, S. E., R. Selvin & D. F. Schick. 1987. Food resources related to
habitat in the scallop Placopecten magellanicns (Gmelin. 1791): A
qualitative study. J. Shellfish Res. 6:89-95.
Sokal. R. R. & F. J. RohlL 1995. Biometry: the principles and practice of
statistics in biological research. New York: W. H. Freeman and Co.,
887 p.
Strickland. J. D. H. & T. R. Parsons, 1968. A Practical Handbook of
Seawater Analysis. Bnlletin Fish. Res. Board Canada No. 167,
Thompson, R. J. 1984. Production, reproductive effort, reproductive value.
and reproductive cost in a population of the blue mussel Mytilns ednlis
Irom a subarctic environment. Mar. Ecol. Prog. Ser. 16:249-257.
Vahl, 0. 1980. Seasonal variations in seston and in the growth rate of the
Iceland scallop. Chlamys islandica (O. F. Miiller) from Balstjord, 70 N.
J. E.xp. Mar. Biol. Ecol. 48:195-204.
Viso. A. & J. Marty. 1993. Fatty acids from 2S marine microalgae. Ph\-
tochemistiy 34:1521-1533.
Volkman, J. K.. S. W. Jeffrey, P. D. Nichols, G. 1. Rogers & C. D. Gariand.
1989. Fatty acid and lipids composition of 10 species of microalgae
used in mariculture. / E.xp. Mar Biol. Ecol. 128:219-240.
Widdows, J. & B. L. Bayne. 1971. Temperature acclimation of Mytilns
ednlis with reference to its energy budget. J. Mar Biol. Ass. U.K.
51:827-843.
.loiiiiNil <-/ Slwllfish Rcscanh. Vol. 22, No. I. 111-117, 2003.
DEVELOPMENT, EVALUATION, AND APPLICATION OF A MITOCHONDRIAL DNA
GENETIC TAG FOR THE BAY SCALLOP, ARGOPECTEN IRRADIANS
SP:IFU SEYOUM,' THERESA M. BERT,'* ami WILBUR.- WILLIAM S. ARNOLD,' AND
CHARLES CRAWFORD'
'F/.s7( ((/((/ Wilillife Conservation Commission. Florida Marine Research Institute. 100 Eighth Avenue SE.
St. Petersburg. Florida 33701-5095: and 'Department of Biologic Sciences. University of North
Carolina-Wilmington. 601 S. College Road. Wilmington. North Carolina 28403
ABSTRACT As a component of an aquaculuirc-based, bay .SLuiiop stock-restoration program in west-central Florida nearshore
waters, we developed a genetic tag for the bay scallop, Argopecien irradians. Using the polymerase chain reaction technique, we
amplified segments of highly purified scallop mitochondrial DNA using 10-base-pair random primers and generated fragments that we
investigated for use as genetic tags. We excised and cloned the amplicons obtained from six individuals to assess them for nucleotide
variability. We chose one, highly polymorphic umplicon of 1049 base pairs and designed a set of sequence-specific polymerase chain
reaction primers for it. We used these primers to sequence portions of the fragment, from both the 5' and 3' ends, respectively,
effectively dividing the fragment into two distinct segments separated by 66 nucleotide base pairs. The two segments contained
sufficient polymorphism such that 729^ (segment 1) and HO'f (segment 2) of the individuals were unique in a sample of 97 wild
individuals; 979f of these individuals were unique when both segments were considered. Nucleotide sequences appeared to be faithfully
transmitted from one parent to its presumed offspring; indications of heterozygosity and heteroplasmy were not observed for any
individual throughout the study. Our analysis of this DNA fragment suggested that it is an mtDNA component, but we were unable
characterize the gene region that it encompasses. To test our genetic tag, we used these two segments to preliminarily assess the
contribution of the stock restoration program to the bay scallop population at a single area targeted for stock restoration. Our analysis
suggests that the stock restoration effort either did not contribute or contributed at a low level to the local population, but our
postrestoration sample sizes may have been too small to detect very small contributions. Our work demonstrates the utility of using
random primers to develop mtDNA genetic tags for species for which little is known about the nucleotide sequence or gene order of
the mtDNA molecule and the potential for application of that tag as a preliminary evaluation tool in a stock restoration or stock
enhancement program.
KEY WORDS: Argopecten iiradicm.s. bay scallop, DNA sequencing, Florida, genetic tag, mitochondrial DNA, stock enhancement,
stock restoration, random primers
INTRODUCTION
Throughout the world, many commercially and lecreationally
valuable species of shellfish and finfish are declining as a result of
overfishing, pollution, habitat degradation, and disease (e.g., Har-
gis 1999, Marelli et al. 1999. Hutchings 2000). Management of
these fisheries through quotas and closures has not always been
effective in preventing further decline or allowing natural recovery
to take place. For this reason, aquaculture-based stock restoration
and enhancement are now accepted methods for restoring depleted
fishery stocks (Tettelbach and Wenczel 1993. Peterson et al. 1995.
Southworth and Mann 1998. Svasand et al. 2000. Ai-nold 2001 ).
The bay scallop, Argopecten irradians (Lamarck. 1819), a spe-
cies valuable to the people of Florida as both a commercial and
recreational resource, was once plentiful in west-Florida nearshore
waters and high-salinity embay ments. By the early 1960s, popu-
lation numbers and abundances had severely declined in many
regions, due in part to dwindling seagrass beds and pollution
during the 1950s (Haddad 1988, Blake et al. 1993, Arnold et al.
1998). Later, concerted efforts by state and federal governments
and environmental groups led to water-quality improvement and
restoration of habitats suitable for scallop propagation (Blake
1998), In the 1990s, the commercial fishery was closed, and area-
specific restrictions or prohibitions on harvesting were imple-
mented for the recreational fishery (Arnold el al. 1998). Despite
these efforts and those of a small-scale bay-scallop stock-
enhancement project ongoing throughout the 1990s (Blake 1998),
*Corresponding author. E-mail: theresa.bert@fwc. state. fi. us
bay scallop population numbers and abundance continued to de-
cline in west-central Florida waters. Therefore, in 1997, a multi-
year, aquaculture-based. stock restoration program was initiated in
west-central Florida nearshore waters to re-establish extirpated bay
scallop populations and enhance depleted populations.
One way to assess the success of a stock restoration endeavor
is to estimate the contribution of the original aquaculture bi'ood-
stock to the local or regional population. For bay scallops, the most
reliable method is a genetic tag, generated from nuclear or mito-
chondrial DNA (mtDNA). A genetic tag can be used to estimate
the success of a stock restoration or enhancement program because
an assessment of the contribution of hatchery-reared or hatchery-
derived individuals to the recipient population can be made (Bert
et al. 2002). The genetic tag must be sufficiently powerful to
discriminate aquaculture-derived individuals from wild individu-
als, or at least the tag should be composed of genotypes of suffi-
cient rarity to allow detection of the stock restoration contribution
through changes in the percentages of these genotypes in the popu-
lation. In either case, the contribution of the restoration effort must
be sufficient to enable detection by sampling.
There are several methods of genetic tagging (Palsboll et al.
1997, Palsb0ll 1999, Bert et al, 2002). but one of the easiest is to
find and use a highly variable portion of mitochondrial DNA
(mtDNA). Nontranscribed regions of the mtDNA molecule serve
as excellent genetic tags (Simon et al. 1994) because they typically
mutate more rapidly than most other DNA segments (Meyer
1993), In addition, if the mode of inheritance is uniparental, track-
ing it in first-generation offspring is straightforward.
In invertebrates, the mtDNA molecule can vary greatly in both
1 II
112
Seyoum et al.
gene order and nucleotide sequence, even among closely related
taxonomic groups (e.g.. Boore & Brown 1994. Boore et al. 1995.
Wilding et al. 19991. Thus, the universal mtDNA primers de\ el-
oped tor vertebrates (Palumbi 1996) are not always successful in
amplifying invertebrate mtDNA segments. Here, we report on the
development of a compound mtDNA genetic tag for the bay scal-
lop using an unidentified bay scallop mtDNA fragment initially
amplified by 10-base-pair (bp) random primers obtained commer-
cially. We evaluate its genetic diversity and applicability through
a preliminary assessment of the contribution of our slock restora-
tion program to the bay scallop population at one location (Ho-
mosassa Bay. Florida; Fig. 1 ). Last, we discuss the general utility
of single-gene genetic tags.
MATERIALS AND METHODS
Dcrelopineiit uf the Mitochiindrial DNA Genetic Tag
To search for an mtDNA fragment that could serve as a genetic
tag, we first obtained highly purified mitochondrial DNA from the
gonad tissues of sexually mature bay scallops from Homosassa,
Florida (;i = 6); mature bay scallops contain both male and female
reproductive tissues. We used a modified homogenization buffer
containing 100 (jlM sucrose and the standard extraction method of
cesium-chloride density gradient ultracentrifugation (Lansman et
al. 1981). The mtDNA band was collected in a 1-mL syringe h\
side puncture with a hypodermic needle and purified by dialysis.
The purified mtDNA yielded a single fragment of approximately
16.000-20.000 bp when run through a 29K ethidium-bromide-
stained. low-EEO. agarose gel (Fisher Biotechnologies. Pittsburgh.
PA). According to the methods described by Williams et al.
(1990), we amplified portions of the highly purified mitochondrial
DNA of these individuals using the twenty 10-bp random primers
supplied in RAPD Kit A (Qiagen Operon Technologies. Inc.,
Alameda. CA). Five microliters of each polymerase chain reaction
(PCR) product was run in a low-EEO agarose gel to view the
amplifications. Multiple bands were obtained for most primers
except for OPA-3 (AGTCAGCCAC) and OPA-18 (AGGTGAC-
Atlantic
Ocean
Gulf of Mexico
Figure L Collection locations for bay scallops {Argopecten irradians)
in Florida to estimate the frequencies of original-bruodstock haplo-
types in the wild prior to the restoration effort (sample sizes are given
in Table IB). HO was the location of the stock restoration evaluation
presented in this report. .Abbreviations: ST = Stcinhatchee; CK =
Cedar Key; HE = Hernando; HO = Homosassa Bay; AN = .Anclote
Estuary; TB = Tampa Bay; SS = Sarasota Bay.
CGT). OPA-3 gave a single band of approximately 1.000 bp and
OPA-18 gave two intense bands of approximately 1.000 and 1.6(X) bp.
The remaining 45 (xL of the OPA-3 and OPA-18 PCR reactions
were run in a gel of 2% low-melting-point agarose (NuSeive,
FMC. Rockland, ME). According to the standard method of Sam-
brook et al. (I9S9). the fragments were excised and cloned in a
plasmid vector (Bluescript PBC KS. Stratagene. La Jolla. CAl that
was initially cleaved with EcoRV and tailed with 2-mM dTTP
(Marchuk et al. 1990). Afterplating the transformed cells, the three
fragments were amplified from two colonies of each of the six
individuals using the T3 and T7 primers (Stratagene. La Jolla.
CA), which annealed to the Bluescript vector on either side of the
insert. The PCR products were electrophoresed in a 1.5%, low-
EEO, agarose gel, excised, and purified using a Strata Prep DNA
Gel Extraction Kit (Stratagene, La Jolla, CA). The purified DNA
was resuspended in 50 (xL of sterile distilled water.
Cycle sequencing was performed from both the 5' and 3' ends
of each fragment using 0.5 (xL of the purified DNA, two |j.l of
Big Dye'^' Terminator Cycle-Sequencing Ready-Reactions with
AmpliTaq FS DNA polymerase (PE Biosy stems, Foster City, CA)
and 0.5 |xL of 3.2-pM solutions of the T3 and T7 primers, in a total
volume of 5 p.L. The reaction product was then ethanol precipi-
tated and resuspended in 20 |j.L of Template Suppression Reagent
(PE Biosystems, Foster City. CA). according to the manufacturer's
instructions.
The resuspended product was analyzed by using an ,^BI
PrismT^' 310 Genetic Analyzer (PE Biosystems. Foster City, CA).
The sequences obtained were aligned and edited using the
AutoAssembler^"^ DNA Sequence Assembly Software (PE Bio-
systems. Foster City. CA); further electropherogram editing was
perfomied using Chromas v. 1.6 (Technelysium Pty. Ltd. Heles-
ville. Queensland. Australia). The two OPA-18 fragments cloned
and sequenced for the six individuals were composed of multiple
sequences, most of which matched very poorly when aligned and
were therefore considered nonhomologous. Some sequences
aligned very well and were thus presumed to be homologous, but
they were invariable in this fragment. However, sequences of the
OPA-3 fragment for the six individuals were homologous and
differed among all six individuals at one or more nucleotides. This
L049-bp fragment was named SCAOPA-3 and was identified as a
possible genetic tag. A representative sequence of the SCAOPA-3
fragment has been deposited in GenBank under accession number
AF261938. Highly specific primers for the fragment were de-
signed: SCA-I, composed of 5'-AGTCAGCCACCCACTAAA-
TTAGATCTCA-3' and SCA-2, 5'-AGTCAGCCACTGGTT-
TATAGTGGAATAGTT-3'. The first 10 bp of these primers con-
stituted the sequence of the 10-bp primer that was initially used to
amplify the SCAOPA-3 fragment.
Using each custom-made primer, we sequenced the
SCAOPA-3 fragment from the two ends toward the center, thereby
effectisely dividing it into two segments. The sequences for the
first portion (termed segment 1) consisted of 471 bp beginning at
position 33 and ending at position 503; the second segment
(termed segment 2) consisted of 450 bp beginning at position 569
and ending at position 1018. These segments did not overlap and
66 bp between these segments were not included.
For the remaining genetic analyses, bay scallops were obtained
alive and from each individual, a section of adductor muscle was
excised, labeled, and stored at -80°C. For each individual, we .
purified total DNA from the adductor muscle using the modified
Mitochondrial DNA Genetic Tag for Bay Scallop
113
PureGene DNA Extraction protocol for small tissue samples, ac-
cording to the manufacturer's instructions (Centra Systems. Min-
neapolis. MN).
To identify the origin of the SCAOPA-3 fragment (i.e.. mito-
chondrial or nuclear DNA). we used our broodstock bay scallops.
The bay scallop stock-restoration project involved two generations
of broodstock (an "original-broodstock" |parental| generation and
a '"restoration-broodstock" [F,] generation). The offspring of the
restoration-broodstock generation constituted the aquaculture-
derived "brood" (F^) generation that should supplement the wild
population. We determined the nucleotide sequences of 26 resto-
ration broodstock raised from eight original broodstock collected
from the wild Homosassa Bay population in 1997 and 23 restora-
tion broodstock raised from five original broodstock collected
Ironi the wild Homosassa Bay population in 1998. We examined
these sequences for among-individual heteroplasmy and for
within-individual heterozygosity. We also compared the sequence
of SCAOPA-3 to published sequences and to those a\ailable in the
computer database GeneBank.
Testing the Generic Tag
To assess the natural level of polymorphism of the SCAOPA-3
fragment and the potential of each segment to serve as an inde-
pendent component of a compound genetic tag, we sequenced
from one direction each of the two segments for 97 individuals
collected in 1997 and 1998, prior to the time of potential input
froin the stock restoration program. Twenty-three of these were
from Homosassa. of which 13 were the original-broodstock scal-
lops used in the Homosassa Bay stocking effort; the remainder of
these were collected from Tampa Bay (/; = 50) and the Anclote
Estuary (h = 24) (Fig. 1). Scallops from these nearby sites were
used because the Homosassa bay scallop population had collapsed
and thus individuals from that location had to be used with dis-
cretion To estimate the frequencies of the original-broodstock
haplotypes in the wild population, we collected and analyzed 54
individuals from Homosassa and 271 individuals from six other
west-Florida nearshore locations (Fig. I ). These individuals were
collected prior to 1999. the first year that aquaculture-derived in-
dividuals could ha\e contributed to the population.
To test the utility of our genetic tag, we examined the
SCAOPA-3 sequences from bay scallop recruits collected from
Homosassa Bay during appropriate years, determined as follows.
In west Florida, bay scallops, which are hermaphrodites, com-
mence spawning in October and generally cease by December
(Barber and Blake 1983; Arnold et al. 1998). Therefore, from
September through early October of each year, the original-
broodstock scallops were collected from wild populations at loca-
tions targeted for restoration, brought into the laboratory, and
spawned under controlled conditions. Their offspring (the restora-
tion broodstock) were reared in containment through the winter
and the following spring until they attained approximately 20-30
mm shell height. These scallops were then planted in cages in the
vicinities of collection of the original broodstocks. There, they
were to complete their growth through the summer and. hopefully,
contribute to the spawning stock when they sexually matured in
the fall. Their recruits (the brood generation), along with wild
recruits also inhabiting the restoration locations, would be of suf-
ficient size to be collected and tested for parentage in summer of
the year after they were spawned by the restoration broodstocks
and 2 y after collection and breeding of the original broodstocks.
Bay scallops can live for 2 y (Orensanz et al. 1991). but it is not
known whether they contribute to the spawning stock in the second
year of their lives. To insure that we accounted for this possibility,
we collected bay scallop recruits from restoration sites and assayed
them for the genetic tag for two years after the planting of the
restoration broodstocks. if those broodstocks survived for 2 y.
Thus, a single cycle of bay scallop stock restoration, including the
genetic monitoring, was a 3- or 4-y process.
We searched for haplotypes that could be from the offspring of
the restoration broodstocks that were planted in Hoinosassa Bay in
1998 and 1999; these were derived from original broodstocks col-
lected in 1997 and 1998. We removed the 1998 restoration brood-
stock after the 1998 spawning sea.son because most of those indi-
viduals died. However, we left the 1999 restoration broodstock in
their cages through both the 1999 and 2000 spawning seasons.
Therefore, we assayed bay scallop recruits collected from Homo-
sassa Bay in 1999 for genotypes that matched the 1997 original-
broodstock genotypes and assayed bay scallop recruits collected
from the bay in both 2000 and 2001 for genotypes that matched the
1998 original-broodstock genotypes.
To obtain these post-restoration "assessment" collections of
bay-scallop recruits, we randomly allocated 20 sampling stations
within an area of Homosassa Bay defined by the 0.7 m and 2.0 m
depth contours and by somewhat arbitral^ latitudinal borders that
were selected based upon our knowledge of the area. Using
SCUBA, at each station we .searched within I m on each side of a
300-m transect line and collected ail scallops within that zone (600
m~ per transect, 12,000 m" total). We also collected scallops using
vessel-deployed rollerframe trawling gear. Those samples were
obtained from deeper water sites (approximately 1.5-m to 3.5-m
depth). The Global Positioning System locations (available upon
request) of these collections were recorded. All assessment collec-
tions were potentially composed of an admixture of wild recruits
and hatchery-derived recruits, the latter of which could have as
parents either two restoration-broodstock indi\ iduals or one resto-
ration-broodstock and one wild individual. (We recognize that, if
mtDNA is maternally inherited in scallops, any recruit generated
by the union of an egg from a wild individual and a sperm from a
restoration-broodstock individual would not be identified as a pos-
sible aquaculture-derived bay scallop.)
Bert and Tringali (manuscript in preparation) describe the
samples needed to perform a complete assessment of a stock res-
toration or enhancement effort. Following their suggestions, we
analyzed the following individuals for their genetic-tag nucleotide
sequences. After they completed spawning, we assayed the eight
original-broodstock individuals used in fall 1997 and the five
original-broodstock individuals used in fall 1998 for both seg-
ments 1 and 2 of our genetic tag. Because bay scallops are her-
maphroditic, any or all of the original-broodstock individuals may
have passed their mtDNA on to the restoration broodstocks. We
did not assay the restoration broodstocks because many individuals
died before we could collect them. We assayed the following
numbers of bay scallop recruits: 199 collected in 1999. 253 col-
lected in 2000, and 242 collected in 2001. To detect individuals
with aquaculture-derived mtDNA haplotypes in these assessment
collections, we first compared the SCAOPA-3 segment-2 haplo-
type of each recruit to that of each original-broodstock scallop
from the appropriate year. We then sequenced for segment I any
recruit that had a segment 2 haplotype that matched that of an
appropriate-year, original-broodstock scallop. We used the sag-
114
Seyoum et al.
merit 2 component of our genetic tag first because it was slightly
more variable than segment 1 (see below).
Data Analyses
To examine the level of genetic diversity of our SCOPA-3
fragment, provide baseline data for estimating the contribution of
our stock restoration effort to the Homosassa bay scallop popula-
tion, and estimate the sensitivity of our genetic lag, we first cal-
culated a number of standard measures of genetic diversity for
each segment using the Arlequin statisfical package (Schneider et
al. 2000) on the 97 bay scallops collected from the three west-
Florida locations. We then estimated the frequencies of original-
broodstock haplotypes in the seven wild bay-scallop collections
and used the AMOVA program in Arlequin to obtain a baseline
estimate of the distribution of the original-broodstock haplotypes
in those collections, which included the Homosassa Bay collec-
tion. We also used Arlequin to quantify the genetic diversity of the
segment 2 haplotypes in each of the wild bay-scallop collections
and in the collective wild population. In addition, we searched for
original-broodstock haplotypes in the collections of wild bay scal-
lops. We tested our ability to detect original-broodstock haplotypes
in the wild population by calculating the minimum detectable fre-
quency iMDF) of the original-broodstock haplotypes, using the
basic binomial sampling equation
MDF= 1 -cxp
ln(a)
(1)
where a = 0.05 and /; = number of individuals, and defined as
the frequency below which the probability of detecting at least one
individual bearing an original-broodstock haplotype would be
<0.05. We assumed that our sampling and the distribution of the
haplotypes in the wild population were random.
To estimate the contribution of our stock restoration effort to
the Homosassa Bay scallop population, we first examined the ap-
propriate assessment collections for the presence of original-
broodstock haplotypes. Then we used Equation I to calculate the
probability of detecting those haplotypes in those assessment col-
lections. (Detailed mathematical and statistical approaches for the
overall assessment of the restoration effort will be described in a
later manuscript [Wilbur et al. in preparation]).
We further explored the limitations of our genetic tag by con-
ducting probability assessments on simulated data based on hap-
lotype frequencies observed in individuals from the 1999 and the
2000 + 2001 assessment collections. For the simulations, we ran-
domly eliminated 5%, 10%, 15%, 20%, or 25% of the individuals
in the collections and substituted at the designated frequency a
single, randomly chosen haplotype from the 1997 or 1998 original
broodstock, as appropriate for the assessment collection(s) under-
going the simulation analysis (Table 2). We then calculated hap-
lotype diversity, nucleotide diversity, and the percentage of differ-
ent haplotypes in the population for these simulated collections
and compared these statistics to those calculated for the corre-
sponding actual assessment collection without the hypothetical
stock restoration contribution. If the stock-restoration program was
successful, we would expect to see significant shifts in the fre-
quencies of haplotypes possessed by the original broodstock in
populations following restoration efforts. To determine the mini-
mum post-restoration frequency differences that would be needed
to detect contributions from restoration broodstock. we used the
V-test (DeSalle et al. 1987) to compare the haplotype frequency
distributions of our simulated as.sessment collections with the ap-
propriate actual assessment collection. For the 5% increment
within which we detected significance, we simulated assessment
collections for each 1 % increment of stock restoration contribution
and tested each of those simulated collections for significant dif-
ferences in haplotype distribution compared with our actual as-
.sessment collections.
RESULTS AND DISCUSSION
Evaluation of the SCAOPA-3 mtDNA Fragment
Our characterization the origin of the SCAOPA-3 fragment
suggested that it is of mitochondrial DNA origin. Each of the
SCAOPA-3 sequences from our 49 restoration-broodslock scal-
lops strictly matched only one of the haplotypes in the appropriate
pool of original-broodstock haplotypes. The DNA sequencing pro-
tocol that we used allowed for detection of heterozygous individu-
als if they were present; that is, heterozygous sequences charac-
teristically appear as two peaks of approximately equal intensity at
a given nucleotide site. However, none of these bay scallops were
heterozygous, and we found no heterozygous individuals in any of
our subsequent analyses. Therefore, we conclude that SCAOPA-3
is transmitted from parent to offspring as a haploid molecule and
we presume that it is mitochondrial DNA. At present, we cannot
say if SCAOPA-3 is inherited maternally: paternal mtDNA inher-
itance occurs in other bivalves (e.g., Mytilius: Liu et al. 1996.
Zouros et al. 1994). Despite comparing its nucleotide and pre-
sumptive amino acid sequences to those reported for other organ-
isms, including other mollusks (Hoffmann et al. 1992, Boore and
Brown, 1994) and to unpublished sequences, (e.g., GeneBank ac-
cession numbers AB055625, AB065375), we were unable to
characterize with certainty the gene region it encompasses. How-
ever, this will not affect the study, provided that SCAOPA-3 is
faithfully transmitted as a haploid molecule from parent to off-
spring.
Sensitivity and Application of the SCAOPA-3 Fragment
The eight 1997 original-broodstock individuals had only seven
different segment 2 haplotypes. However, the two individuals that
were identical for segment 2 differed for segment 1 . Thus, each of
our broodstock individuals had a SCAOPA-3 haplotype that was
unique in the aquaculture hatchery.
All differences among individuals in segment 1 and segment 2
of our mtDNA fragments were in the form of single bp substitu-
tions. In Table I A, we present estimates of genetic diversities for
the two segments as determined by sequencing the individuals
used to characterize the fragment. Separately, these segments dis-
tinguished high percentages of individuals: collectively, they dis-
tinguished nearly all of the individuals.
Results of the AMOVA analysis suggest that the bay scallops
comprising the west-Florida pre-restoration collections were ge-
netically homogeneous with respect to the SCAOPA-3 mtDNA
fragment. In Table IB, we summarize the segment 2 genetic di-
versities for these collections and for the west-Florida population.
Both the percentage of individuals with different haplotypes and
the proportion of haplotypes that were unique were very high in
the individual samples and high in the combined data. Eighty-five
individuals (26%) were defined by four haplotypes in the propor-
tion 42:19:18:6: thus, the most common haplotype present in the
population occurred in only 13% of the individuals. Conespond-
Mitochondrial DNA Genetic Tag for Bay Scallop
115
TABLE 1.
Kstiniates of bay scallop {Argopecten irradiaiis) genetic divcrsitifs for the SCAOPA-3 mitochondrial DNA genetic tag in (A) ')! wild
individuals from Tampa Bay, Homosassa. and Anclote, Florida, (segments I and 2 are defined in Materials and Methods! and (Bl western
Florida collections made prior to the stock restoration effort (segment 2 only).
Segment
No. bp
HN
HQ
RS
h
P
K
A.
1
451
72
^)l
0.19
(1.97 ±(1.(11
(1.86 + 0.48
3.9 ± 2.0
2
450
80
S6
(1.2(1
(1.49 ± (1.(1(1
L.Vi ±0.07
5.6 ±2.7
Total
901
97
98
U.20
1.0(1 ±(1.(10
L10±()..S6
9.5 ± 4.5
Location
Year
N
HN
HQ
h
P
K
B.
ST
1997, 1998
61
75
89
0.97 ±0.01
1.22 ±0.66
5.4 ± 2.6
CK
1997
20
75
87
0.94 ± 0.04
1.07 ±0.61
4.7 ±2.4
HO
1997. 1998
54
81
89
0.99 ±0.01
1 .38 ± 0.74
6.2 ±3.0
HE
1997, 1998
32
81
85
0.98 ±0.01
1.12 ±0.63
4.8 ±2.4
AN
1997. 1998
67
84
93
0.99 ±0.01
1.30 ±0.70
5.4 + 2.7
TB
1997
65
80
90
0.98 ±0.01
1.30 ±0.70
5.6 ± 2.7
SS
1998
26
85
86
0.98 ± 0.02
1.80 + 0.60
4.8 ±2.4
Total
325
62
85
0.98 ± 0.00
1 .26 ± 0.68
5.3 ± 2.7
Abbreviations: No. bp = number of base pairs; HN = percentage of individuals with different haplotypes: HQ = percentage of haplotypes unique to
single individuals; RS = number of polymorphic sites per nucleotide site; li = haplotype diversity; p = nucleotide diversity in %; K = number of
pairwise nucleotide differences; N = number of individuals; ST = Steinhatchee; CK = Cedar Key; HO = Homosassa; HE = Hernando; AN =
Anclote; TB = Tampa Bay; SS = Sarasota Bay.
h. />. and K are mean values ± standard deviations.
ingly. all standard measures of genetic diversity were conipura-
tively high (e.g., haplotype diversity ranged 0.94-0.99).
Twenty-four wild-individual haplotypes matched five 1997
original-broodstock haplotypes for segment 2. However, none of
the individuals that matched original-broodstock segment-2 hap-
lotypes also matched the same broodstock individual for seg-
ment 1. No wild individuals collected in 1998 matched any of the
original-broodstock segment-2 haplotypes. If our assumptions as-
sociated with Equation I were vahd, we could expect to obtain a
match between a wild-individual haplotype and an original-
broodstock haplotype if the broodstock haplotype was present in
our wild-population sample at a frequency of approximately \% or
greater {MDF,,^ = 0.00917). Thus, the estimated prerestoration
frequency of each of the 1997 and 1998 broodstock haplotypes in
the wild population probably was less than 1%.
Ten of the assessment scallops collected in 1999 matched three
of the 1997, segment-2, original-broodstock haplotypes. Eight of
those were identical to the single original-broodstock scallop with
the haplotype that was the second most common in the wild popu-
lation. However, the haplotypes of all of those individuals differed
from that original-broodstock individual's segment 1 haplotype.
No segment-2 haplotypes from assessment bay scallops collected
in 2000, and only one segment-2 haplotype from an assessment
bay scallop collected in 2001, matched any 1998, original-
broodstock, segment-2 haplotype. That individual did not match
for segment 1 the original-broodstock individual that it matched
for segment 2. Thus, our collective sample size of 694 individuals
gave no indication that the bay scallop restoration project contrib-
uted to the local Homosassa bay scallop population during 1999-2001.
The MDF^,^ for detection of an original 1997 or 1998 brood-
stock haplotype in the appropriate assessment collection(s) was,
respectively 0.015 (1999 assessment collection) or 0.0060 (2000 +
2001 assessment collections). Original-broodstock haplotypes that
were present in the putative admixed Homosassa population at
frequencies near or below the MDFg^s were at statistical risk of not
being detected. However, these frequencies were so low that stock
restoration contributions at or below these levels may essentially
be inconsequential.
Although haplotype diversity and nucleotide diversity in our
hypothetical assessment of stock restoration contribution were pro-
portionally reduced with increasing stock restoration contribution,
they were not as sensitive to the input of stock restoration contri-
bution as was the percentage of different haplotypes (Table 2).
Nevertheless, our simulations indicate that a stock restoration con-
tribution of at least 15% in the 1999 assessment collection and
\0% in the 2000-2001 combined assessment collection would be
needed to generate a significant difference between those assess-
ment collections with versus without stock restoration contribu-
tions.
Genetic Tags and Molluscan Stock Restoration
The general strategy in a stock restoration program is to collect
animals from the targeted restoration site, produce large quantities
of aquaculture-reared or. in the case of our bay scallop program,
aquaculture-derived (one generation removed 1 individuals, and use
them to supplement or replenish the population at the same site.
Determining the success of such an effort depends on the ability to
detect the contribution (in numbers or percentages) of hatchery-
reared or hatchery-derived offspring in the post-restoration re-
cruits. In supplemented populations, the frequency of aquaculture-
generated individuals can range from undetectable to a complete
swamping of the admixed population. A single-gene genetic tag
such as ours can indicate whether restoration effort has resulted in
essentially undetectable input, substantial input, or a complete
swamping of the local population. However, the capacity of this
tag to estimate the contribution of the stock restoration effort be-
tween the extremes of essentially no input and very high input is
116
Seyoum et al.
TABLE 2.
Hypothetical analysis of stock restoration contribution in the
assessment collections from Honiosassa with levels of contribution
varying from 0% (original assessment collection) to 25% (see
Materials and Methods for method of simulating stock restoration
contributions). (A) 1999 assessment collection (A' = 199
individuals). (B) 2I)(II) + 20(11 combined assessment collections
(A' = 495 individuals).
SRC(%)
Nl
N2
HNl
HN2
A.
199
0.72
0.72
5
189
10
0.73
0.70
10
179
20
0.73
0.66
15
169
30
0.72
0.62
20
159
40
0.75
0.60
25
149
50
0.77
0.5S
B.
495
0.69
0.69
5
470
35
0.69
0.66
10
445
69
0.69
0.62
15
421
104
0.70
0.73
20
396
139
0.72
0.57
25
370
174
0.73
0.55
Abbreviations: SRC = hypothetical stock restoration contribution; Nl =
number of individuals taken from the specified year assessment collection;
N2 = hypothetical number of individuals contributed from the stock res-
toration program (within a single percentage, all of which were taken from
a single, randomly chosen broodstock individual); HNl = percentage of
individuals with different haplotypes without stock restoration contribution
(calculated based on Nl only); HN2 = percentage of individuals with
different haplotypes with stock restoration contribution (calculated on Nl
+ N2).
related to the degree of statistical uniqueness, as measured by
statistical probability, of the tag in each application. To precisely
define an intermediate-level contribution from a stock restoration
effort, the assessment collection must consist of a very high num-
ber of individuals; the genetic tag must be complex (e.g., com-
posed of our compound mtDNA genetic tag plus several micro-
satellite loci), or. if it is a single-gene tag. extremely variable; or
the method for determining the contribution must differ from ours.
Because we found no original-broodstock haplotypes in either
the wild population or the assessment collections, we can combine
all of these collections to estimate the uniqueness of our original-
broodstock haplotypes and calculate the MDF above which we
might expect to encounter one of these haplotypes. We can esti-
mate with 95'7r probability that v\e would have detected at least
one original-broodstock haplotype in this combined sample ( 1,019
individuals) if the frequency of any of these haplotypes was 0.003
or greater. Clearly, frequencies below this MDF would represent
inconsequential contributions from a stock restoration effort. Thus,
our single-gene genetic tag should be useful for assessing the
success of our entire bay scallop restoration effort.
In many cases, a single-locus, preliminary genetic tag such as
ours could be useful in assessing the contribution of stock resto-
ration efforts. Multi-locus genetic tags can be laborious, time-
consuming, and expensive to develop, test, and apply. Fuilher-
more. in our case, the potential for reproductive mixing between
restoration broodstock and wild scallops limits the ability for
nuclear DNA-based assignment of individuals to either the brood
generation or to the wild population. Our genetic tag can be used
to preliminarily evaluate the success of a bay scallop stock en-
hancement or restoration effort and thereby to evaluate whether it
is worth the expense and effort to develop a more definitive ge-
netic tag. Then, if it appears that the stock restoration effort may
have contributed a potentially significant fraction of the recruits to
an area, a high-resolution, multi-gene tag can be developed. How-
ever, under certain conditions, the type of genetic tag presented
here may be sufficient for an entire study.
The advantages of using a single-gene genetic tag composed of
more than one hypervariable segment and in which the segments
can be used sequentially are increased resolution and reduced ef-
fort. In our genetic tag. both segment 1 and segment 2 had ample
and nearly equivalent variation. By sequencing first for segment 2.
the expense and time required were reduced significantly because
only the individuals that had segment 2 haplotypes identical to
those of the original-broodstock haplotypes also needed to be se-
quenced for Segment 1 ,
The utility of a single-gene genetic tag such as that presented
here is enhanced if the broodstock used possesses essentially
unique haplotypes or genotypes. However, there are limitations to
this type of approach. A large number of wild individuals or a high
percentage of the wild population must be assayed to establish the
frequencies of the genetic-tag haplotypes in the pre-restoration
population, and individuals with "unique" haplotypes should be
used as broodstock. Threatened or depleted populations can be
further endangered if they are flooded with aquaculture-derived
individuals that collectively possess only a few naturally rare
genotypes or haplotypes, if those individuals interbreed exten-
sively and successfully with the remnant wild population. Never-
theless, for some applications, the procedure that we described
here provides researchers with a method for finding an mtDNA
genetic tag in organisms for which little is known about their
mtDNA. This type of genetic tag can be used to screen individuals
and derive parentage or group associations for stock restoration
efforts, conservation biology, or other suitable applications.
ACKNOWLEDGMENTS
We thank M. Tringali for assistance in the designing of the
primers and notable suggestions in many aspects of the analysis.
We also appreciate the assistance of D, Marelli. M. Parker, M.
Harrison, and S. Peters with the field collections and C. Lund, T.
Thompson, and D. Warner for various types of assistance. We
additionally thank M. Tringali, A. McMillen-Jackson, and two
reviewers for valuable comments on our manuscript. This study
was funded by a grant from the National Oceanic and Atmospheric
Administration (NOAA), grant NA76FK0426 and project FWC
2234 and by the state of Florida. The views expressed herein are
those of the authors and do not necessarily reflect the views of
NOAA or any of its sub-agencies.
LITERATURE CITED
Arnold. W. S. 2001. Bivalve enhancement and restoration strategies in
Florida, USA. Hydmbiologia 465:7-19.
Arnold, W. S., D. C. Marelli. C. P. Bray & M. M. Harrison. 1998. Re-
cruitment of bay scallops Argopccteii irradimis in Ploridan Gulf of
Mexico waters: scales of coherence. Mar. Ecol. Prof;. Scr 170:143-157.
Barber, B. J. & N. J. Blake. 1983. Growth and reproduction of the bay
Mitochondrial DNA Genetic Tag for Bay Scallop
117
scallop, Argopecten iirmlians (Lamarck) al it^ southern distributional
limit. J. Exp. Mar. Biol. Ecol. 66:247-25fi.
Bert, T. M., M. D. Tringali & S. Seyoum. 2002. De\elopnient and appli-
cation of genetic tags for ecological aquaculture. In: B. Costa-Pierce,
editor. Ecological aquaculture. Oxford. UK: Blackwell Science, Ltd..
pp. 47-76.
Blake, N. J. 1998. The potential for reestablishing bay scallops to the
estuaries of the west coast of Florida. Trans. 63"' No. Am. Wildlife
Natural Re.murces Conf. 63:184-189.
Blake, N. J.. Y. Lu & M. Mover. 1993. Evaluation of Tampa Bay waters
for the survival and growth of southern bay scallop larvae and juve-
niles. Final Report. Tampa Bay National Estuary Program. Tech, Pub.
#04-93. p. 18.
Boore, J. L. & W. M. Brown. 1994. Complete DNA sequence of the
mitochondrial genome of the black chiton, Kallniriiia Uinicala. Gener-
ics 38:423-+43.
Boore. J. L.. T. M. Collins. D. Stanton. L. L. Daehlcr & W. M. Brown.
1995. Deducing the pattern of arthropod phylogeny from mitochondrial
DNA rearrangements. Naliire 376:163-165.
DeSalle, R.. A. Templton. I. Mori, S. Pletscher & J. S. Johnston. 1987.
Temporal and spatial heterogeneity of intDNA polymorphisms in natu-
ral populations of Drosophila merealonim. Genetics 1 16:215-223.
Haddad. K. 1988. Habitat trends and fisheries in Tampa and Sarasota Bays.
NCAA Estuary of the Month Series #11, Tampa and Sarasota Bays. pp.
113-128.
Hargis, W. J., Jr. 1999. The evolution of the Chesapeake oyster reef system
during the Holocene epoch. In: M. W. Luckenbach. R. Mann & J. A.
Wesson, editors. Oyster Reef Habitat Restoration: A Synopsis and
Synthesis of Approaches. Proceedings from the Symposium, Williams-
burg. Virginia, pp. 5-23.
Hoffmann, R. J., J. L. Boore & W. M. Brown. 1992. A novel inilochondrial
genome organization for the blue mussel. Mxlilus ediilis. Genetics 131:
397^12.
Hutchings, J. A. 2000. Collapse and recovery of marine fishes. Nature
406:882-885.
Lansman, R. A., R. O. Shade. J. F. Shapira & J. C. Avise, 1981. The use
of restriction enzymes to measure mitochondrial DNA sequence relat-
edness in natural populations. / Mol. Evol. 17:214-226.
Liu, H.-P., J. B. Mitton & S.-K. Wu. 1996. Paternal mitochondrial DNA
differentiation far exceeds maternal mitochondrial DNA and allozyme
differentiation in the freshwater mussel. .Anadimta grandis grandis.
Emiuium 50:952-957.
Marchuk. D.. M. Drumm, A. Saulino & E. S. Collins. 1990. Construction
of T-vectors, a rapid and general systein for direct cloning of unmodi-
fied PCR products. Nucl. Acid Res. 19: 1 154,
Marelli, D. C, W. S. Arnold & C. Bray. 1999. Levels of recruitment and
adult abundance in a collapsed population of bay scallops {Argopecten
irradiuns) in Florida. / Shellfish Res. 18:393-399.
Meyer, A. 1993. Evolution of mitochondrial DNA in fishes. In: P. W.
Hochachka & T. P. Mommsen. editors. Biochemistry and Molecular
Biology of Fishes, Vol. 2. Amsterdam: Elsevier, pp. 1-38.
Orensanz, J. M., A. M. Parma & O. O. Iribarne. 1991. Population dynamics
and management of natural stocks. In: S. E. Shumway, editor. Scallops:
Biology, Ecology, and Aquaculture. New- York: Elsevier, pp. 625-713.
Palsboll, P. J. 1999. Genetic tagging: contemporary molecular ecology.
Biol. J. Linnean Soc. 68:3-22.
Palsboll. P. J., J. Allen. M. Berube. P. J. Clapham, T. P. Feddeersen. P. S.
Hammond, R. R. Hudson, H. Jorgensen, S. Katona, A. H. Larsen, F.
Larsen. J. Lien, D. K. Maltila, J. Sigurjonsson, R. Sears R, R. Smith, R,
Sponer. P. Stevick & N. Oien 1997. Genetic tagging of humpback
whales. Nature },U:161 -1 b^ .
Palumhi, S. R. 1996. Nucleic acids II: The polymerase chain reaction. In:
D. M. Hillis. C. Moritz & B. K. Mable editors. Molecular Systematics,
2nd ed. Sunderland, MA: Sinauer Associates, Inc., pp. 205-247.
Peterson, C. H., H. C. Summerson & J, Huber. 1995. Replenishment of
hard clam stock using hatchery seed: combined importance of bottom
type, seed size, planting season, and density. / Shellfish Res. 14:293-
300.
Sambrook. J., E. F. Frit.sch & T. Maniatis. 1989. Molecular Cloning: a
Laboratory Manual. 2nd ed. Cold Spring Harbor, NY: Cold Spring
Harbor Laboratory Press.
Schneider, S., D. Roessli & L. Excoffier. 2000. Arlequin: A .software for
population genetics data analysis. Ver. 2.000. Geneva: Genetics and
Biometry Lab, Dept. of Anthropology, University of Geneva.
Simon, C, F. F. Frati. A. Beckenbach. B. Crespi. H. Liu & P. Flook. 1994.
Evolution, weighting, and phylogenetic utility of mitochondrial gene
sequences and a compilation of conserved polymerase chain reaction
primers. Ann. Entoino. Soc. Am. 87:651-701.
Southworth. M. & R. Mann. 1998. Oyster reef broodstock enhancement in
the Great Wicomico River, Virginia. / Shellfish Res. 17:1101-1114.
Svasand, T., T. S. Kristiansen, T. Pedersen, A. G. V. Salvanes, R. En-
gelsen, G. Naevdal & M. Nodtvedt. 2000. The enhancement of cod
stocks. Fish & Fr-^heries 1:173-205.
Tettelbach, S. T, & P, Wenczel. 1993. Reseeding efforts and the status of
bay scallop Argopecten irradians (Lamarck 1819) populations in New
York following the occurrence of "brown tide" algal blooms. / Shell-
fish Res. 12:423-431.
Wilding, C. S., P. J. Mill & J. Grahame. 1999. Partial sequence of the
mitochondrial genome of Littorina sa.xatilis: relevance to gastropod
phylogenetics. J. Mol. Evol 48:348-359.
Williams, J. G. K., A. R. Kubelik, J. J. Livak, J. A. Rafalski & S. V.
Tingey. 1990. DNA polymorphisms amplified by arbitrary primers are
useful as genetic markers. Nucl. Acid Res. 18:6531-6535.
Zouros, E., A. O. Ball, K. R. Freeman & C. Saavedra. 1994. An unusual
type of mitochondrial DNA inheritance in the blue mussel Mytilis.
Proc. Natl. Acad. Sci. USA 91:7463-7467.
Journal of Shellfish l<,:s<;urh. Vol. 22, Nu. I. I 19-123, 2003.
GAMETOGENESIS IN A SYMPATRIC POPULATION OF BLUE MUSSELS, MYTILUS EDULIS
AND MYTILUS TROSSULUS, FROM COBSCOOK BAY (USA)
A. P. MALOY,* B. J. BARBER, AND P. D. RAWSON
School of Marine Sciences, University of Maine, Orono, Maine 04469
ABSTRACT To lest the hypothesis that a temporal variation in species-specific spawning times is the mechanism Hmiting hybrid-
ization and maintaining genetic integrity in a Mylilits ediilis (L.) and M. irossiihi.s (Gould) hybrid /one in eastern Maine, mussels from
a low intertidal site in Cobscook Bay were histologically examined at monthly to semi-monthly intervals throughout the year 2000.
Analysis of gamete volume fraction and oocyte area measurements detected no difference in the timing of gametogenesis and spawning
between M. edulis and M. trossuliis. Differences in mature oocyte area measurements, however, indicated that M. ediilis spawned larger
eggs than M. trossuhis. At this location, low frequency of hybridization and maintenance of genetic identity for these two species is
unlikely the result of temporally distinct spawning times.
KEY WORDS: Myiihis. gametogenesis. hybridization, mussels
INTRODUCTION
The Mylilus species complex is composed of three closely re-
lated blue mussel species. M. edulis. M. trossuius, and A/, gullo-
provincialis. In the northern hemisphere, M. edulis occurs princi-
pally in the eastern and western Atlantic; M. trossuius is found in
the Baltic Sea. the northwestern Atlantic Ocean, and the northern
Pacific Ocean; and M. galloprovincialis occurs in the Mediterra-
nean Sea. the Atlantic coa.st of southern Europe, northern Africa,
and the Pacific coast of North America (Gosling 1984, 1992.
Koehn 1991, McDonald et al. 1991, Suchanek et al. 1997). An
early survey of Mytilus spp. on the east coast of North America
indicated the presence of only a single species, M. edulis (Koehn
et al. 19761, but in a later study, Koehn et al. (1984) identified two
genetically distinct taxa inhabiting Atlantic Canada. These two
genetically distinct groups (M. edulis and M. trossuius) form a
zone of sympatry from northern Newfoundland south to the east-
ern coast of Maine (Varvio et al. 1988. McDonald et al. 1991.
Bates and Innes 1995. Comesana et al. 1999, Rawson et al. 2001 ).
Hybridization is commonly reported wherever members of the
Mylilus complex are sympatric (Gosling 1992). In the Baltic Sea.
M. edulis and M. trossuius hybridize so readily that they are con-
sidered semi-species (Viiinola & Hvilsom 1991 ). M. edulis and M.
galloprovincialis hybridize extensively in a zone of sympatry that
extends from the coast of Spain through the British Isles. The
frequency of hybrid genotypes varies significantly among loca-
tions but can reach values as high as 80% in some populations
(Hilbish et al. 1994. Cotnesafia & Sanjuan 1997. Sanjuan et al.
1997). In contrast, the frequency of hybrid genotypes formed by
interspecific matings between M. edulis and M. trossuius in the
northwest Atlantic is much lower, ranging from 12 to 26% (Koehn
et al. 1984. Varvio et al. 1988, Bates & Innes 1995. Mallet &
Carver 1995. Saavedra et al. 1996. Comesana et al. 1999. Rawson
et al. 2001 ). Although variation among sampling locations and the
use of different methodologies (e.g., morphologic analysis, allo-
zyme electrophoresis, mitochondrial, and nuclear DNA-based
markers) may be partly responsible for the variation in the fre-
quency of hybrids observed, these studies suggest that hybridiza-
tion is less prevalent among blue mussels on the Atlantic coast of
North America than in the Baltic or European hybrid zones.
Mate choice, habitat specialization and differential environ-
mental tolerance, spawning asynchrony, and gamete incompatibil-
ity are processes that can initiate and maintain reproductive isola-
tion between closely related species in sympatric populations
(Palumbi 1994). In free-spawning marine invertebrates, mate
choice, per se. is unlikely to play an important role in limiting
hybridization. Increasing evidence, however, suggests that gamete
interactions can affect reproductive isolation. For example, rapid,
divergent evolution in sperm proteins (bindin and lysin) limits
interspecific hybridization in sea urchins and abalone (Swanson &
Vacquier 1998. Palumbi 1999). respectively. The existence of
similar mechanisms in bivalves has not been confirmed.
Additionally, any habitat-specific selection that creates patchy
species distributions may also limit hybridization because fertil-
ization is more likely among close neighbors. Gardner (1996) has
suggested that blue mussel hybrid zones occur in regions of envi-
ronmental discontinuity so that the general patterns of species
distribution are determined by differential adaptation. Several
studies have observed that the distribution of blue mussel species
is conelated with changes in environmental parameters, both in the
contact zone between M. edulis and M. galloprovincialis in west-
ern Europe (Hilbish et al. 1994, Gardner 1996, Gilg & Hilbish
2000. Hilbish et al. 2002) and between M. trossuius and M. gal-
loprovincialis on the Pacific coast of North America (Sarver &
Foltz 1993). In the northwest Atlantic, research has focused on
differences in salinity and wave exposure in structuring the species
composition of blue mussel populations. There has been little evi-
dence to directly link any of these factors with either the distribu-
tion, or the relatively low frequency, of hybrids within the region
where M. edulis and M. trossuius are sympatric.
Reproductive isolation and maintenance of genetic identity
may also be dependent on temporal variation in spawning events.
In sympatric populations of M. galloprovincialis and M. edulis in
southwestern Europe, low hybridization is observed when spawn-
ing periods are out of phase, whereas sites with a greater degree of
synchrony have a higher degree of hybridization (Gardner 1992,
Seed 1992). The objective of the present study was to determine
whether the relatively low rate of hybridization occurring between
M. edulis and M. trossuius in eastern Maine could be attributed to
temporal variation in spawning.
*Corresponding author. Department of Biochemistry. Microbiology, and
Molecular Biology. University of Maine, Orono. ME 04469. Fax: 207-
581-2801; E-mail: aaron.maloy@umit.maine.edu
MATERIALS AND METHODS
Adult mussels (35 to 50 mm in shell length) were collected by
hand from a sympatric. low intertidal population in East Bay (lati-
119
120
Maloy et al.
tilde 44°56'30"N; longitude 67°07'50"W: Cobscook Bay. Maine)
throughout 2000 (Table 1 ). Samples of 120 mussels were obtained
monthly from January through April. October through December,
and semi-monthly between 4 May and 14 September. Mussels
were transported on ice to the University of Maine, and a piece of
mantle tissue approximately 0.5 cm" was removed and preserved
in 95'/f ethanol for DNA extraction. The remainder of the mussel
was preserved in Dietrich's fixative (Gray 1954) for subsequent
histologic preparation. All preservation was completed within 24 h
of collection.
DNA was extracted from gonadal tissue following the protocol
of Rawson et al. (2001). Three polymerase chain reaction-based
nuclear markers, polyphenolic adhesive protein (Glu-5'). internal
transcribed spacer. Mytihis anonymous locus-I. and one mitochon-
drial marker (mtl6s-F: Rawson et al. 2001), were used to identify
mussels with M. edulis and M. trossidus genotypes from each
sampling period. Initially, the Glu-5' marker was run on al!
samples and used to identify 30 (n = 40 on 17 and 30 August)
individuals homozygous for both M. edulis and M. trossidus Glu-
5' alleles. These 60-80 mussels were subsequently genotyped at
the remaining three markers. Individuals not scored as inultilocus
homozygotes for M. edulis or M. trossulus alleles at all markers
(i.e., hybrids) were eliminated from further anal
bined results of all four markers were used to pick
(;; = 30 on 17 and 30 August) of each species for assaying re-
productive condition.
Preserved individuals were transversely sectioned (2- to 3-mm
thick) anterior of the byssal gland, dehydrated in an ascending
alcohol series, cleared with Xylenes, and embedded in Paraplast
(Howard & Smith 1983). Cross sections (5 ixm) of each block were
cut on a rotary microtome, placed on glass slides, stained with
Shandon instant hematoxylin and eosin Y, and permanently
mounted. Slides were examined using a compound microscope
(Nikon LABPHOT-2) equipped with a video camera (Dage CCD
72). Images were digitized with a fraine grabber (Flash Point 128,
vsis. The com-
20 individuals
Integral Technologies Inc.) and measurements made using image
analysis software (Image Pro Plus; Media Cybernetics).
Reproductive state was measured by two separate methods.
First, the gamete volume fraction (GVF) of all indixiduals was
calculated as the area of reproductive tissue present in one micro-
scopic t~ield divided by the entire area (Bayne et al. 1978). Thus,
estimates of GVF indicate the proportion of mantle that is com-
prised of reproductive tissue. The mean of five random fields
(300x) was calculated for each individual and used in subsequent
statistical analysis. In addition to the GVF. mean oocyte area was
estimated for each female from 50 measurements ( 1 200x ) of the
cross-sectional area of oocytes with a clearly visible nucleolus
(Garrido & Barber 2001).
GVF data were analyzed using a three-way ANOVA for
sample date, species, and gender. Oocyte data were evaluated with
a two-way ANOVA across sample date and species. Both data sets
were evaluated at a = 0.05 using simultaneous BonfeiToni pair-
wise comparisons of sample level means. Statistical analyses were
performed using Minitab 13.0. which automatically adjusts the
Bonferroni a lev el to compensate for the total number of possible
pairwise comparisons. Because all possible combinations of pair-
wise comparisons were not of interest, the a level was manually
readjusted to account for the appropriate number of comparisons
used in the analysis.
RESULTS
Gametogenesis in M. edulis (mean length 44.8 mm ± 3.7) and
M. trossulus (mean length 44.3 mm ± 3.5) was highly synchronous
at the East Bay site throughout 2000. Species-specific mean ga-
mete volume fractions (estimated for both male and female mus-
sels) were relatively low in February and increased steadily in both
species from February to June. The peak mean GVF of 0.89 in M.
edulis was identical to the 0.89 estimated for M. trossulus mussels
sampled on 4 June. GVF remained high in both species throughout
TABLE L
Mytilus edulis, Mytilus trossulus: relative number of males, females, and undifferentiated mussels sampled In East Ba>, 20(10.
Mytilus edulis
Mytilus trossulus
Males
Females
LndifTerentiated
Males
Females
Undifferentiated
Totals
19 Jan
7
9
4
5
5
1
31
20 Feb
9 .
6
5
8
11
1
40
21 Mar
11
7
2
11
8
1
40
17 Apr
7
11
2
9
10
1
40
4 May
8
10
2
8
12
40
1 8 May
12
8
-
11
9
40
4 Jun
8
12
8
11
39
18 Jun
11
9
13
7
40
30 Jun
8
11
9
11
39
17 Jul
12
8
12
8
40
1 Aug
10
10
9
11
40
17 Aug
19
10
1
17
11
58
30 Aug
15
14
13
16
58
14 Sep
7
10
3
13
4
3
40
15 Oct
9
11
7
5
8
40
17 Nov
10
9
1
7
7
4
38
9 Dec
6
12
1
6
8
6
39
Totals
169
167
21
16(1
154
25
702
Undifferentiated individuals were not used in statistical analysis.
Gametogenhsis in Sympatric Blue Mussels
121
June and July and then declined precipitously between 1 7 July and
1 August samples among mussels of both species. Following this
initial dramatic decline, a less pronounced decrease in GVF was
observed up to the 15 October sampling date, after which GVF
estimates were constant and nearly equal to those observed m
February (Fig. 1 ).
Analysis of gender-specific patterns of GVF \ariation indicated
that while gamete development in the females of both species was
comparable to that of males, it lagged behind that of the males. For
example, mean GVF estimates for females were consistently lower
than those observed in males from February to April but by June
these differences had disappeared. In addition, spawning in fe-
males resulted in a greater loss in GVF relative to males. Overall,
males had an average yearly GVF approximately lO'/r higher than
(enialcs for both Mytilus ediilis and M. trossuliis, Bonferroni pair-
wise comparisons (a = 0.05) indicated a significant difference in
GVF between males and females on 30 August (Fig. 2 A and B).
Consistent with the graphic analysis, a three-way ANOVA re-
vealed that significant differences in GVF occurred between date
and gender but not between species. Significant interactions oc-
curred between date and species and between date and gender
resulting from the seasonality of gamete development. Gametoge-
nic cycles (as defined by GVF) were the same for both species and
there were no significant interactions between species and gender
or date*species*gender (Table 2). With respect to the shaip de-
crease in GVF. Bonferroni analysis indicated that significant de-
creases in GVF at both the species and gender levels corresponded
with the initial spawning period between 17 July and 1 August.
Though differences occurred between sexes because of the high
postspawn variation, spawning times were still highly synchro-
nous.
Similar results were obtained using mean oocyte areas to assess
gametogenic cycles. Mean oocyte areas increased sharply for both
species from 21 March through 4 June. After 4 June, oocyte areas
gradually increased until maxima were observed on 17 July [Myti-
lus cdulis 678.6 ^JLm" and M. trossuliis 530.1 |j,m"). A sharp de-
crease in mean oocyte areas occurred between 17 July and 1 Au-
gust. After I August, there were increases in oocyte area until 30
August for M. ecluHs and 14 September for M. trossuliis. followed
by a less pronounced and protracted period of decline until 9
December (Fig. 3).
The two-way ANOVA for oocyte areas indicated a significant
l.U -
T - ■■ 1
c
0.9 -
- isc-
o
/- >
—
1) s -
•/
\
u
7
\
II 7 •
T ^ ■''
\. ^ 1
U.
.
\
— *— V/ fduIlK
Ol
0.6 ■
'r
L \
- -c- -M imsMilu''
b
'
3
U.S-
'
1 ,
.
/
\
O
y
^^^
\
>
114 ■
■/
»
\
Ol
,
\
u
,'
s
/
^
"
\
s
y.
]
'
'S
o
(1 : -
4 J
.:
...
-
(1.0 -
1 1 1 1 1 1 1 1 1 1 1 1 1 1—^' — r—
— p^
^-1
u! 5 <
6 _1 r^
s s
Sample Date
E
O
10
0.4
0.8
0.7
0.6
(1..^ ■
(14 -
Male
Female
ZIt jij 01)
3 3 3
< < <
# B
Sample Date
o
>
O
1
y ■
^ ;^-JHM
/:--^■--^--J\
O.S ■
/- \ : \
0.7-
/
A
/^ A
—•—Male
/
;\
- - - • Female
5 ■
/
■'
\
04 -
r /
.- J
\
n 1 -
/ .,' ;
,
V T
\ _/ '
\
II : -
I.'"': '
1
^' 'N
^
/
II 1 -
»--'
■•
';.
"
■ '
110 -
^ ^ , 1 ,
, — , —
1 , _:
"7 s
Sample Date
Figure 2. (A) Mytilus edulis. Mean (±1 SDl jjaniete volume fraction for
male vs. female mussels from East Bay, Maine. I'SA. (B) ,Mytilus tros-
suliis. Mean (±1 SD) gamete volume fraction for male vs. female mus-
sels from East Ba>, Maine.
interaction between date and species (Table 3). The difference
between species was caused by variation in mean oocyte size
rather than a variation in the timing of gametogenic events; aver-
age yearly oocyte area was 338.2 \i.m~ forM. edulis and 308.2 p,m"
for M. trossuliis. Significant declines in species-specific oocyte
area were observed between 17 July and 1 August, corresponding
with a period of spawning indicated by GVF analysis. Additional
TABLE 2.
Gamete volume fraction of Mytilus eilulis and Mytilus trossulus:
results of a three-«a> .\NO\ .\ testing the effects of date, species,
and gender on the gametogenic cycle.
Figure 1. Mytilus edulis, Mytilus trossulus. Mean (±1 SD) gamete vol-
ume fraction for mussels from East Bav, Maine, USA.
Source
df
Mean Square
F Value
Date
16
4.4869
142.53***
Species
1
0.0003
0.01
Gender
1
1.8650
59.24***
Date X species
16
0.1061
3.37**
Date X gender
16
0.0827
2.63**
Species x gender
1
0.0407
1.29
Date x species x
gender
16
0,029.1
0.93
Error
."^S?
** P< 0.01.
*p<o.m\.
122
Maloy et al.
800
*N
700
E
a
600
a
SOO
<
n
400
>,
(J
o
o
300
200
— ■ — M edlilis
T*" M :rossiiltis
C -= OU OU 5JJ C- T1
= ^ 3 3 3 O ^
g -g I £. ^ ^ g
V "r 5 < 5 S 3
2 c - r- ■ ' T ^ - _ _ r^ o • - r- •-..
— r. ^, _ -f =c — ^, _ „ — — _
Sample Date
Figure 3. Mytiliis ediilis. Mytilus troxsiiliis. Mean (±1 SD) oocyte area
for mussels from East Bav, Maine.
significant decreases between dates were slightly out of phase,
with the mean oocyte area of M. edulis decreasing from 14 Sep-
tember to 15 October and that of A^. trossidits from 15 October to
17 November. A significant difference in oocyte area (t = 7.24,
P < 0.001) was observed just prior to spawning on 17 July indi-
cating that M. edulis spawned larger eggs than did M. trossiihis.
Mean shell length of females sampled on this date was not sig-
nificantly different (t = 1.29, P = 0.220).
DISCUSSION AND CONCLUSIONS
The reproductive cycle of mussels in this population was highly
seasonal which is typical of many benthic marine invertebrates in
northern temperate zones. In this study, gonadal development was
minimal through the winter as indicated by low GVF and small
oocyte diameters. Increased gametogenic activity in spring corre-
sponded to increasing water temperature and presumably food
a\'ailability. A significant decrease in GVF and oocyte diameters,
indicative of a major spawning event, took place in late July and
involved a large proportion of the population. Interestingly, GVF
for females increased slightly in samples collected after this initial
spawning event. Such an increase could be caused by redevelop-
ment of the gonad in preparation for a second spawning. However,
we observed little histologic evidence of redevelopment in indi-
vidual mussels that had already spawned. The predominant histo-
logic feature at this time was empty follicles containing a few
refractory oocytes. Thus, the few individuals that did not spawn or
had only partially spawned after the peak-spawning event in late
July were responsible for the observed increase in GVF.
TABLE 3.
Mean oocyte area of Mytilus edulis and Mytilus trossulus: results of
a two-way ANOVA testing the effects of date and species on the
gametogenic cycle.
Source
df
Mean Square
F Value
Date
16
1.6398
53.67***
Species
1
0.1236
4.04*
Date X species
16
0.0665
2.18**
Error
285
0.0306
* P < 0.05. **P < 0.01, ***P < 0.001,
More importantly, the reproductive cycles of Myliliis edulis and
M. trossulus sampled from this population were highly synchro-
nous. For the year 2000 at the East Bay site, the results of this
study indicate that interspecific fertilization between M. edulis and
M. trossulus is possible based on spawning times. Similar findings
have been reported elsewhere. Freeman et al. (1992) and Mallet
and Carver ( 1995) observed synchronous reproductive patterns in
populations of M. edulis and M. trossulus from Lunenburg Bay,
Nova Scotia. Additionally, Toro et al. (2002) found that the ini-
tiation of spawning was coincident between these species and their
hybrids in Trinity Bay, Newfoundland; although M. trossulus dis-
played a more protracted period of spawning at this location the
variation alone was not sufficient to explain the limited numbers of
hybrids observed. Thus, four studies covering a wide geographic
region from Maine to Newfoundland have observed similar results
all suggesting that hybridization is not limited solely by species-
specific differences in spawning times.
It is possible that genetic identity is maintained between M.
edulis and M. trossulus by a factor other than different spawning
periods. Gamete recognition proteins have been shown to drasti-
cally reduce the hybridization potential between closely related
taxa of marine invertebrates. Interestingly, molecular phylogenies
suggest that M. trossulus is the most divergent of the blue mussel
taxa (Rawson & Hilbish 1995). It has been recently shown that M.
edulis and M. trossulus have also diverged significantly with re-
spect to amino acid sequence at a spenn lysin locus (C. Riginos.
pers comm). Divergence in gamete recognition proteins such as
sperm lysin could act to limit hybridization between M. trossulus
and other blue mussel taxa. Though no evidence of functional
differentiation has been documented as yet, preliminary data indi-
cates that cross-fertilization of M, edulis and M. trossulus is lim-
ited except at very high sperm concentrations (Rawson unpub-
lished). Thus, future effoils should focus on more detailed obser-
vations of the spawning behavior of these two species as well as
the potential for functional variation in gamete recognition pro-
teins.
The present study found that M. trossulus had smaller mean
oocyte size at maturity and presumably spawned smaller eggs than
M. edulis. Given that M. trossulus has a higher reproductive output
(Toro et al. 2002), it follows that similarly sized M. trossulus
produced more (but smaller) eggs than M. edulis. which might
provide a selective advantage for the more fecund M. trossulus.
Similarly, M. galloproviiicialis has a higher fecundity per unit
length than M. edulis at Croyde in S.W. England, but genotypic
ratios between these two species have not changed over time be-
cause of large numbers of small M, edulis (Gardner & Skibinski
1990). Smaller oocytes may also represent a response to environ-
mental stress. Cobscook Bay in eastern Maine is near the southern
distributional limit of M. trossulus (Rawson et al. 2001) and as
such, may be a less than optimal environment for this species.
However, M. tros.sulus from Newfoundland also produces smaller
eggs, has a smaller size at first maturity than M. edulis (Toro et al,
2(J02), as well as a population structure containing a higher fre-
quency of small M. trossulus individuals (Comesana et al. 1999).
Given that a difference in oocyte size has been observed in both
Maine and Newfoundland it is more likely that this difference is
the result of a difference in life history strategy rather than a
response to environmental stress. Additional data are needed on
extrinsic factors such as population structure, size at first maturity,
reproductive output, and size dependent mortality to draw coiiclu-
Gametogenesis in Sympatric Blue Mussels
123
sions concerning the intrinsic factors sinaping (he lite history evo-
lution of M. cdiilis and M. trossiiliis.
ACKNOWLEDGMENTS
Funding for this project was provided through a Maine Aqua-
cuhure Innmation Center crant to B. J. Barber and P. D. Rawson.
Maine Sea Grant, and Experiinent Station Hatch Funds to
P.D. Rawson. We are also grateful to D. Beane for histologic
preparations, and S. R. Fegley and P. A. Haye for helpful
comments on earlier versions of this manuscript. This is Maine
Agricultural and Forest Experiment Station external publication
#2627.
LITERATURE CITED
Bates. J. .A. & D.J. Iniies. 1995. Genetic variation amcmg popiikitidiis of
Myliliis Slip, in Eastern Newfoundland. Mar. Biol. 124:417—424.
Bayne. B. L.. D. L. Holland. M. N. Moore. D. M. Lowe & J. Widdows.
1978. Further studies on the effects of stress in the adult on the eggs of
Mylilus ediilis. J. Mar. Biol. Ass. U.K. 58:825-841.
Comesatia. A. S. & A. Sanjuan. 1997. Microgeographic allozyme differ-
entiation in the hybrid zone of Mylilus galloprnvincialis Lmk. and M.
ediilis L. on the continental European coast. Helgol. Meeresunters 5 1 :
107-124.
Comesana. A. S.. J. E. Toro. D. J. Innes & R. J. Thompson. 1999. A
molecular approach to the ecology of a mussel {Mylilus edulis-Mylitus
irossulus) hybrid zone on the east coast of Newfoundland. Mar. Biol.
13.1:213-221.
Freeman, K. R.. K. L. Perry & T. G. DiBacco. 1992. Morphology, condi-
tion and Reproduction of two co-occurring species of Mytihis at a Nova
Scotia mussel farm. Bull, Aquacul, Assoc. Canada 92-3:810.
Gardner, J. P. A. 1992. Mytilus galloprovincialis (Lmk) (Bivalvia. Mol-
lusca): The taxonomic status of the Mediterranean mussel. Ophelia
35:219-243.
Gardner, J. P. A. 1996. The Mytilus edulis species complex in southwest
England: Effects of hybridization and introgression upon interiocus
associations and morphometric variation. Mar. Biol. 125:385-399.
Garrido, C. L. & B. J. Barber. 2001. Effects of temperature and food ration
on growth and oogenesis of the green sea urchin, Strongylocentrotus
droebachieiisis. Mar Biol 138:447^58.
Gardner. J. P. A. & D. O. F. Skibinski. 1990. Genotype-dependent fecun-
dity and temporal variation in spawning in hybrid mussels {Mytilus)
populations. Mar Biol. 105:153-162.
Gilg, M. R. & T. J. Hilbish. 2000. The relationship between allele fre-
quency and tidal height in a mussel hybrid zone: a test of the differ-
ential settlement hypothesis. Mar. Biol. 137:378.
Gosling. E. M. 1984. The systematic status of Mytilus galloprovincialis in
western Europe: A review. Malacologia 25:551-568.
Gosling, E. M. 1992. Systematics and geographical distribution oi Mytilus.
In: E. M. Gosling (ed.). The mussel Mytilus: Ecology, physiology,
genetics and culture. Amsterdam: Elsevier, pp. 1-20.
Gray. P. 1954. The microtomist's fomiulary and guide. New York: The
Blakiston Co Inc.. pp. 108.
Hilbish. T. J.. B. L. Bayne & A. Day. 1994. Genetics of physiological
differentiation within the marine mussel genus Mylilus. Evol. 48:267-
286.
Hilbish, T. J.. E. W. Carson. J. R. Plante. L. A. Weaver & M. R. Gilg. 2002.
Distribution of Mylilus edulis. M. galloprovincialis. and their hybrids
in open-coast populations of mussels in southwestern England. Mar.
Biol. 140:137-142.
Howard. D. W. & C. S. Smith 1983. Histological techniques for marine
bivalve mollusks. NOAA technical memorandum NMFS-F/NEC-25:9.
Koehn. R. K. 1991. The genetics and taxonomy of the species in the genus
Mytilus. Aquiu-ulmre 94:125-145.
Koehn. R. K.. R. Milkman & J. B. Mitton. 1976. Population genetics of
marine pelecypods. IV. Selection, migration, and genetic differentia-
tion in the blue mussel Mytilus edulis. Evol. 30:335-342.
Koehn. R. K., J. G. Hall. D. J. Innes & A. J. Zera. 1984. Genetic differ-
entiation of Mytilus edulis in eastern North America. Mar Biol. 79:
117-126.
Mallet, A. L. & C. E. Carver 1995. Comparative growth and survival of
Mytilus trossulus and Mytilus edulis in Atlantic Canada. Can. J. Fish.
Aquat. Sci. 52:1873-1880.
McDonald, J. H., R. Seed & R. K. Koehn. 1991. Allozyme and morpho-
metric characters of three species of Mytilus in the northern and south-
ern hemispheres. Mar. Biol. 111:323-333.
Palumhi. S. R. 1994. Genetic divergence, reproductive isolation, and ma-
rine speciation. Annu. Rev. Ecol. Sysl. 25:547-572.
Palumhi. S. R. 1999. All males are not created equal: Fertihty differences
depend on gamete recognition polymorphisms in sea urchins. Proc.
Natl. Acad. Sci. USA 96:12632-12637.
Rawson, P. D. & T. J. Hilbish 1995. Evolutionary relationships among the
male and female mitochondrial DNA lineages in the Mytilus edulis
species complex. Mol. Biol. Evol. 12:893-901.
Rawson. P. D.. S. Hayhurst & B. VanScoyoc. 2001. Species composition
of blue mussel populations in the northeastern Gulf of Maine. / Shell-
ti.^h. Res. 20:31-38.
Saavedra, C. D. T. Stewart. R. B. Stanwood & E. Zouros. 1996. Species-
specific segregation of gender-associated mitochondrial DNA types in
an area where two mussel species {Mytilus edulis and M. trossulus)
hybridize. Genetics 143:1359-1367.
Sanjuan. A., C. Zapata & G. Alvarez. 1997. Genetic differentiation in
Mytilus galloprovincialis Lmk. throughout the world. Ophelia 47:13-
31.
Sarver. S. K. & D. W. Foltz. 1993. Genetic population structure of a
species' complex of blue mussels {Mytilus spp.) Mar. Biol. 117:105-
117.
Seed. R. 1992. Systematics evolution and distribution of mussels belonging
to the genus Mytilus: An overview. Am. Malac. Bull. 9:123-137.
Suchanek. T. H., J. B. Geller, B. R. Kresier & J. B. Mitton. 1997. Zoo-
geographical distribution of the sibling species Mytilus galloprovincia-
lis and M. trossulus (Bivalvia: Mytilidae) and their hybrids in the North
Pacific. Biol. Bull. 193:187-194.
Swanson. W. J. & V. D. Vacqier. 1998. Concerted evolution in an egg
receptor for a rapidly evolving abalone sperm protein. Science. 281:
710-712.
Toro. J. E.. R. J. Thompson & D. J. Innes. 2002. Reproductive isolation and
reproducti\e output in two sympatric mussel species {Mytilus edulis.
M. Irossulus) and their hybrids from Newfoundland. Mar Biol. 141:
897-909.
Vainola, R. & M. M. Hvilsom. 1 99 1 . Genetic divergence and a hybrid zone
between Baltic and North Sea Mytilus populations (Mytilidae: Mol-
lusca). Biol. J. Linn. Soc. 43:127-148.
Varvio. S. L.. R. K. Koehn & R. Vainola. 1988. Evolutionary genetics of
the Mytilus edulis complex in the North Atlantic Region. Mar. Biol. 98:
51-60.
Journal of Slwllfish Reseanh. Vol. 22. No. I. 125-134, 2003.
MODELING OF FILTER-FEEDING BEHAVIOR IN THE BROWN MUSSEL, PERNA PERNA (L.
EXPOSED TO NATURAL VARIATIONS OF SESTON AVAILABILITY IN SANTA
CATARINA, BRAZIL
F. M. SUPLICY,'* J. F. SCHMITTr N. A. MOLTSCHANIWSKYJ,' AND J. F. FERREIRA'
'School of Aquacuhiire. Tasmanian Aciuaciiluivc and Fisheries Institute. University of Tasmania,
Lockcd-Bag 1-370. Launceston. Tasmania. 7250. Australia: 'Lahoratorio de Ciiltiro dc Mohiscos
Marinhos iLCMM). Departamento de Aquicultura. Universidade Federal de Santa Catarina. P.O. Box
1 0-1. IS. Floriandpolis. Santa Catarina. CEP S,S062-6()I. Brazil
ABSTR.ACT The aim of this -.tiidy is lo quanlity and model the filter-feeding beha\ ior of the mussel Pcnui penui feeding on natural
seston. Models were generated that described each step of the feeding process and produced a predictive model of rates of food uptake
by P. perna in culture areas from Southern Brazil. Feeding experiments using the hiodeposition approach were conducted with mussels
ranging in shell height from 3.94 to 9.22 cm of three sites, including turbid and clear water environments. Organic content of the seston
tOCS. fraction) decreased as total particulate matter (TPM, mg L"') increased. The maximum filtration rate (FR. mg L~') measured
for an individual mussel was 156.7 mg h^' and was recorded when TPM was 33.9 mg L"' and OCS was 0.18. Rejection rate of particles
had a strong positive relationship with TPM, and an inverse relationship with OCS. Maximum rejection rate recorded was 124.1 mg
h ' and was measured under the same seston conditions as maximum filtration rate. Net organic selection efficiency by mussels (NOSE,
fraction) was related to the amount of particulate organic matter (POM. mg L"') and particulate inorganic matter (PIM, mg L"')
available in the water. NOSE was positive below PIM values of 2 mg L"', but had negative values when POM was above 3 mg L~'
and PIM between 2 and 15 mg L"', and positive values when POM was below 3 mg L~' and PIM above 15 mg L"'. Maximum NOSE
was 1.71. when PIM was 1.02 mg L"' and POM was 0.67 mg L"'. Organic content of ingested matter (OCI. fraction) had a positive
relationship with NOSE and TPM. Maximum OCI was 1.24 and was measured when TPM was 33.9 mg L"', OCS was 0.18, FR was
151.30 mg h~', and NOSE was 1.30. The net absorption efficiency of ingested organics (NAEIO) increased with increasing OCI in a
hyperbolic relationship. The net organic absorption rate (NOAR, mg h"') increased with both FR and OCI, The coupling of the
equations that described filter-feeding processes for P. pema in the STELLA software environment produced a robust model with
relatively low complexity and specificity. The model can predict the P. perna feeding behavior in turbid or clear water and can be used
with different species if the correct coefficients are used. The coupling of this feeding model with future models of energy budget,
population dynamics, seston hydrodynamics, and primary production will be valuable for the evaluation of shellfish carrying capacity,
KEY WORDS: mussel physiology, model, Perna perna. STELLA
INTRODUCTION
Assessing carrying capacity, the environmental capacity for
shellfish culture is generally approached using ecophysiological
modeling (e.g., Brylinsky & -Sephton 1991, Newell & Campbell
1998, Schcilten & Smaal 1998). The inclusion of processes relative
to rates of selectivity, rejection, and absorption by molluscan filter
feeders is of primary importance for both ecosystem and local
scales models (Smaal et al. 1998). Sessile suspension-feeders ob-
tain energy by selectively feeding on seston, which includes a
variable mi.xture of algae, detritus, and silt. Not only does the
seston have a small fraction with nutritional value f Smaal & Haas
1997), but also the composition changes on time scales of minute
to tnonths (Grant 1993). The available organic content of the
seston ranges from 5 to 809^ (Bayne & Hawkins 1990). Such
nutritional variability in the seston forces sessile organisms like
mussels to maximize their energy intake and ultimately their net
energy balance, by varying rates of feeding and digestion in re-
sponse to seston concentration and organic content (Bayne el al.
199.3).
The literature describing bivalve rates of filter feeding and
digestion is extensive (see reviews by Bayne & Newell 1983.
Griffiths & Griffiths 1987, Bayne 1993). However, recent findings
suggest that previous studies have limited application because they
used artificial diets, and it is unclear to what extent using artificial
diets provides a realistic representation of "(/; .■iitu" feeding behav-
*Corresponding author. Fax: -1-61-3-6324-3804; E-mail: fsuplicy@utas.edu.au
ior (Bayne & Hawkins 1990). Normal feeding processes and be-
havior are better measured in experiments where the animals are
allowed to feed on natural seston (Hawkins et al. 1996a. Wong &
Cheung 2001. Gardner 2002),
Most research on the ecophysiological processes in shellfish
has focused on temperate species (e.g., Mytihis ediilis). and there
has been limited work on tropical species and their environments
(Hawkins et al. 1998a, Wong & Cheung 2001 ), Although bivalves
use the same general selective mechanisms for food acquisition
(Hawkins et al. 1998b), there are both intra- and inter-specific
differences in feeding rates (Navarro et al. 1991). Describing the
physiologic responses characteristic of each species is needed,
rather than extrapolating data from other species (Gardner &
Thompson 2001, James el al. 2001). There are likely to be a
number of significant differences in tropical environments. Our
understanding of the feeding physiology of Perna perna (Lin-
naeus. 1758) (Berry & Schleyer 1983, Bayne et al, 1984, van
Erkon Schurink & Griffiths 1992) is limited to laboratory experi-
ments using microalgae monocultures or a mix of microalgae spe-
cies and silt. Furthermore, these studies were carried out in South
Africa where cold south Atlantic currents are predominant; in con-
trast, the Brazilian coast has warm waters brought by central At-
lantic currents. Such differences in temperature and productivity,
and consequently in food availability and its organic content, will
be reflected in ecophysiological differences of these filter feeders.
The aim of this study is to generate a model to predict food
uptake by P. perna in culture areas of Southern Brazil, based on
measurements of the filter-feeding process using natural seston.
123
126
SUPLICY ET AL.
The model reproduce the sequential passage of food through the
feeding steps of filtration, selection, rejection, ingestion, and ab-
sorption, and the calculation of each step is based on relationships
either with quantity and quality of seston or with some of the
preceding steps on the food processing sequence. Mussel aquacul-
ture is a fast growing industry in Brazil and problems regarding the
environmental capacity of this industry may occur in the near
future. This research will have the capability to deliver information
that can be incorporated into models of energy budget and growth
as a function of stocking density, for use in planning and managing
strategies of growing areas.
METHODS
Feeding experiments were conducted at three sites within mus-
sel farms in Southern Brazil; Bnto Cove (48°37'W, 27'46'S),
Porto Belo (48°33"W, 27°8'S). and Arma?ao de Itapocoroi
(48°38'W, 26°58'S). Rope-cultured P. perna were collected from
mussel farms at each site immediately before the experiments. All
experiments were done on one to three occasions at each site and
were exposed to natural differences in concentration and organic
content of seston at each site and time (Table 1 ). Each site was
arbitrarily classified as turbid or clear, based on total particulate
matter (TPM). The clear site had TPM <3 mg L"' (Porto Belo).
while the turbid sites had TPM between 10-40 mg L"' (Brito Cove
and Armagao do Itapocoroi).
The experiments were conducted on a raft containing a tray
with 10 individual 330-mL plastic chambers. Eight individual
mussels, cleared of epibiotic growth, were placed in separate
chambers, with two chambers left empty to act as blanks. Seawater
was pumped into the chambers with flow rates in each compart-
ment between 150 and 200 niL min"'; these were adjusted at the
beginning of the experiment. A battle between the mussel and the
inflow water provided a homogeneous distribution of water flow
inside the feeding chambers (Fig. 1). The mussels were initially
left undisturbed for 1 h to acclimate, after which time all biode-
posits on the bottom of the chambers were removed. Once the
experiment started the mussels were allowed to feed for four hours,
during which time all feces and pseudofeces for each mussel were
separately collected using a pipette immediately after being re-
leased. For each individual mussel the feces and pseudofeces col-
lected in each hour were stored in separate test tubes on ice. A 2-L
sample of inflow seawater was collected every 20 min for the
determination of seston concentration and organic content. Water
temperature and salinity were monitored every hour during the
experiment.
After 5 hours of feeding the experiment was terminated and the
mussels and samples were transported back to the laboratory on
ice. The biodeposit samples were homogenized by repeat pipetting
and filtered onto pre-ashed and weighed Whatman glass microfi-
bre (serie C) 1.2 p-m (GF/C) filters (25 mm or 47 mm diameter).
The samples were rinsed with 15 mL distilled water to remove
salts and dried at 60°C for 48 h before re-weighing and calculation
of the total sample dry weight. Each sample was then ashed at
450°C for 4 h prior to final weighing, allowing calculation of both
of the ash (inorganic) and ash-free (organic) mass of each filtered
sample. To account for settled material in the chamber, the mean
organic and inorganic weight of sediment material collected from
the blank chambers was subtracted from the mean organic and
inorganic weight of the collected feces and pseudofeces. To de-
termine seston concentration and organic content, three 300^00
niL samples from the 2 L of inflow seawater collected were fil-
tered onto pre-ashed and weighed Whatman GFC filters (25 mm
diameter) and dried, ashed, and weighed in the same way as the
biodeposit samples. The mean of the three values was calculated.
The seston concentration and organic content for each hour was cal-
culated as an average of the three 2-L samples taken during that hour.
To determine the lag time between when the mussels consumed
food and when feces and pseudofeces production occurs, mussels
starved for one day in the laboratory were fed green microalgae.
Green feces were observed within an hour of feeding therefore we
assumed the gut transit time to be 1 h. Green pseudofeces were
seen within minutes of the microalgae being added. Therefore, in
the analysis of the field data the quantity and content of the feces
was correlated with seston concentration and organic content in the
preceding hour. No time lag was assumed in correlation with
pseudofeces production. Feeding and absorption parameters were
defined and calculated (Table 2) using procedures outlined in
Hawkins et al. (1996a, 1998b), and using the mean of the hourly
feeding rate obtained for each mussel throughout the experiment.
For the regression analysis, seston concentration and organic con-
tent were the means of the hourly values obtained during each
experimental run.
From each mus.sel used in the experiments, total length was
measured and soft tissue removed, dried at 60°C for 48 h, and
weighed. To standardize findings and allow comparison of results
with other studies, feeding responses were expressed per 1 g dry
weight using Y„ = (W^AV^)*" * Y^„ where Y„ is the coiTected
TABLE L
Summary of envirunmental parameters and mussel size range for each day the experiments were run. Data of environmental characteristics
are the mean ± SD. TPM: total dry particulate mass; POM: total particulate organic matter; OCS = organic content of TPM; ND = no data.
Enviror
imental Characteristics n = 12
Mussels
Experiment
TPM
POM
OCS
Temperature
Turbidity
Shell Length
Dry Weight
Days
Location
(mg L"')
(mg L"')
(fraction)
(°C)
(NTU)
(cm)
<S>
14/().V0I
Brito's Cove
29.6 ± 11.9
4.7 + 3.7
0.15 + 0.05
25.7 ± 0.5
ND
5.05-8.90
0.398-3.522
14/04/01
Brito's Cove
12.4 ±3.0
1.2 ±0.3
0.10 ±0.02
25.5 ± 0.5
7.7 ± 1,7
5.70-8.16
0.485-2.034
05/06/01
Brito's Cove
9.8 ±3.1
1.0 + 0.1
0. 1 1 ± 0.03
22.2 ±0.3
4.5 ± 1.6
5.72-8.27
0.628-2.517
07/02/01
Porto Belo
1.7 + 0.3
0.7 + 0.3
0.41 ±0.17
29.0 ±0.4
0.5 + 0.2
5.74-8.28
1.177-3.257
31/O.VOl
Piirto Belu
1.6 ±0.4
0.3 ±0.1
0.20 ± 0.08
26.5 ± 0.4
1 .0 ± 0. 1
5.05-9.22
0.618-3.103
07/07/01
Porto Bell)
1.2 + 0.3
04 + 0.1
0.36 + 0.09
18.3 ±0.0
tl.3 + 0.1
4.11-8.22
0.343-2.757
26/0,';/01
A. Itapocoroi
4.6 ±0.7
2.3 ± 0.4
0.10 ±0.08
21.3 ±0.2
2.8 ± 0.8
6.00-8.49
0.857-3.087
Modeling Feeding Behavior in Perna fekna
127
Secondary tap
i
Water sample
outflow
Inflow
^
Main tap
Pump
lndi\ idual chamber
Figure 1. Schematic diagram ol' the feeding tray used in the biodeposition experiments.
parameter. W^ is the standard weight ( 1 g). W^, is the weight or
length of the experimental animal, Y^. is the uncorrected parameter,
and b is the average size exponent (Hawkins et al. 2001 ). However,
given the absence of spawning synchronicity (Marques et al.
1991), there is high variability in mussel dry weight within the
same population in every time of the year. Therefore, we used the
shell length equivalent of 1 g dry weight (6.26 cm) and the power
exponent that scales the feeding rates with SL (b = 1.85). The
power exponent has previously been used for Myltlus gcdlopnnin-
ciiiUs (Perez Camacho & Gonzales 1984, Navarro et al. 1996) and
for P. perna (Berry & Schleyer 1983).
All statistical analysis was done using SPSS for Windows,
Version 10 (SPSS Inc.. Chicago, ID and Sigma Plot. Multiple
regression models were fitted using the step-wise technique, en-
tering the most significant independent variable at the first step and
then adding or deleting independent variables until no further vari-
ables could be added to improve the overall fit. The coupling of the
equations to produce an integrated feeding model and the posterior
TABLE 2.
Dennitions and descriptions of the calculation of separate components of feeding behavior.
Parameter
.Acronym
Units
Calculation
Purticulated inorgunic matter
Particuluted organic matter
Organic content of seston
Clearance rale
Total filtration rate
Organic filtration rate
Inorganic filtration rate
Organic content ot tillered matter
Rejection rate
Inorganic rejection rate
Organic rejection rate
Net organic selection efficiency
Ingestion rate
Organic ingestion rate
Inorganic ingestion rate
Net organic ingestion rate
Organic content of ingested matter
Net absorption efficiency from
ingested organics
Net organic absorption rate
PIM
mg L-'
POM
mg L-'
OCS
fraction
CR
1 h-'
FR
mg h~'
OFR
mg h"'
IFR
mg h-'
OCF
fraction
RR
mg h"'
IRR
mg h-'
ORR
mil ir'
NOSE
fraction
IR
mg h"'
OIR
mg h"'
IIR
mg h*'
NOIR
mg h"'
OCI
traction
naeio
traction
NOAR
mg h
Asli tree dry weight of TPM
TPM-PIM
POM/TPM
(mg inorganic matter egesteU both as true feces and pseudoteces h~' -h (mg inorganic
matter available T' seawater)
(mg inorganic matter egested both as true feces and pseudoteces h~') -^ (I-OCF)
CR X mg total particulate organic matter r' seawater
CR X mg total particulate inorganic matter 1"' seawater
OFR ^ FR
mg total pseudoteces egested h~'
RR-ash free mg total pseudoteces egested h" '
RR-IRR
I (-(organic fraction within pseudoteces) -^ (OCS)l
FR-RR
OFR-ORR
IFR-IRR
(FR X (OC.S)|-|RR + (organic fraction within pseudofeces)]
NOIR ^ (FR-RR)
NOAR ^ NOIR
N01R-[(mg total true feces egested h"') x (organic fraction within true feces)]
128
SUPLICY ET AL.
sensitivity analysis was done using STELLA research software
(High Performance Systems, Inc.. Hanover. USA).
RESULTS
Organic content of seston (OCS) decreased as TPM increased
(Fig. 2, Table 3). Clearance rate of mussels decreased froin 10 to
5 L h"' as TPM increased from <3 to 30 mg L"' and OCS in-
creased from <0. 15 to 0.40. The parabolic relationship (Fig. 3A).
suggests that P. perna pumps more water under low TPM (<10 nig
L"') and OCS «0.20) conditions.
Filtration rate (FR. mg h"'). rejection rate (RR. mg h"'). in-
gestion rate (IR. mg h"' ). and net organic absorption rate (NOAR.
mg h"') were all related to TPM and OCS (Table 3. Fig. 3B, C. D.
and E). The nia.ximum filtration rate measured was 156.7 mg h''
when TPM was 33.9 mg L" ' and OCS was 0. 1 8. Rejection rate had
a strong positive relationship with TPM and inverse relationships
with OCS. The maximum rejection rate recorded was 124.1 mg
h"'. which represented 83% of filtered matter, and was measured
under the same seston conditions as the maximum filtration rate.
Pseudofeces production was observed when TPM levels were as
low as 2 mg L"', suggesting a very low threshold for pseudofeces
production in this species.
Net organic selection efficiency (NOSE, fraction) was con-
trolled by the proportion of particulated organic and inorganic
matter in the water (POM. mg L"' and PIM. mg L"' respectively).
Higher NOSE values were observed on the lower and higher ex-
tremes of PIM. Negative NOSE values, a minimum of -0.56, was
recorded at intermediate values of PIM and POM. and positive
values were recorded when POM was below 3 mg L"' and PIM
above 15 mg L"'. Maximum NOSE was 1.71 when PIM was 1.02
mg L"' and POM was 0.67 mg L"' (Fig. 3F. Table 3). Organic
content of ingested matter (OCI. fraction) had a positive relation-
ship with NOSE and it was not strongly affected by TPM. Maxi-
mum OCI was 1.24 when TPM was 33.9 mg L"', OCS was 0.18,
FR was 151.3 mg h*', and NOSE was 1.30 (Fig. 4A, Table 3). The
net organic ingestion rate (NOIR, fraction) was below 10 mg h"'
when mussels were feeding on TPM levels below 5 ing L"'. but
this increased to 25 mg h"' when TPM was above 30 mg L"' and
ingestion rate was ca. 50 mg h"' (Fig. 4B. Table 3).
O
.J
en
c
o
«
■n
=
^
3
=U
Of
c
2
f^l
■n
v:
n
c
CO
n
E
<
£
r"
Oi
c
c
£
"^r
^
c
■?
L.
at
a>
.C
C
s2
r
—
=
=
=
5=
5=:
5=
5c'
X.
3
■
■ — ■
__^-
__;
_■
_■
__■
_;
■ — ■
■ — '
^
—
■ — '
■ — ■
■ — ■
■ — ■
■ — ■
■^
\/
M
—
V
V
V
V
V
V
V
V
n,
n,
V
a.
O.
D-
n,
n.
U-
c
Dh
n
o
yr,
C2
c:
oc
00
c
«M
ri
o
oo
■y~i
ct
^
^t
E
_
■^■
■^■
-^■
•^
-1-
T
^
-t
II
II
L.
II
11
II
II
II
II
II
II
-3
■^
T3
~o
T3
"3
■a
■a
-f
— .'
-3
i/^,
tr.
^.
r^.
in
or)
r-
r*",
Tf
oc-
r-
oo
CO
oc
'^
"O
f*-!
^P
^o
-t
in
r-i
r^i
m
r*~i
r-l
r-l
ON
r^
Yi
r-
•rt
m
r\
r-
—
^
ri
t/"^.
II
II
II
II
II
11
II
II
11
II
II
t/j
U.
U-
U-
U-
O.
tJ-
U.
U.
tl.
u.
U,
r-
r-
o
r\
'^■
r^
O
m,
r-i
o
ri
r-,
(N
r-
r-
iri
r-
-i-
r-
ON
ON
Ti-
O
o
o
o
O
c
o
o
o
o
o
II
II
II
II
II
11
II
II
II
11
II
^1
^1
^J
J.,
fi
ri
Al
ri
ri
r-l
"-
■-
'-
■-
*-
*-
"-
'-
*-
c
o
_
r-l
r^.
-t
</",
sC
r--
OC
o-
o
—
=
■ '
■ ■
' "
■ ■
■ ■
~~'
' '
' '
'
O"
u
Figure 2. Relationship between the average organic content (OCS,
fraction) and average total particulate mass (TP^L mg L"') of seston
s\ithin the experhnentai feeding conditions. Data are the mean of three
replicate deternilnatiuns per condition. Reler to Table 3 for equation.
e
2
9
'oo
tu
oo
O
Z
r\
U
( )
c/1
U
O
ri
O
3C
ri
o
O
IT:
8
O
q:
u
o
3\
—
f.,-^
+1
-t
0-
+1
^_
+1
oc
+1
X
r-
+1
+i
2C
a-
§
+1
c>
+1
+
+1
3^
1
+
+
+
o
1
(-1
S
^
2
rr"
a.
s
a.
a.
+
n
s
a.
1
H
o
d
+
S
a.
5
-T
r^,
'^
f-
,:;^
Q.
p
=;■
— ■
^
■
isC
-t
o
t-
+1
Tl
+1
+1
Tl
Tl
5C
04
^
£
—
ri
O'
r^
o
O
H
— .
C:
+
-H
o
O
o
+1
+
t/5
8
+
-1-
t/1
O
-1-
+
fll
q
n
-n
( )
( )
s
f/^
n
O
X
y
O
O
X
n
n
X
o
O
X
X
X
z
X
X
r^
^o
o
. ,
^
,-,
00
' '
rr
+1
+1
+1
+1
+1
c
+1
+1
r~-
r^
r\
r^l
p
d
d
-H
8
-i-
.
+
o:
vC
a.
u.
ol
2
0-
f-
1
0-
H
X
2
D.
t-
CL
f-
X
o
+1
oT
X
X
3C
X
00
X
X
o
2
X
s
d
+1
Ol
d
-1-
a.
d
+1
^1
-r
o
+1
+1
?
?
ol
O
S
H
U",
ri
[^
+1
ri
+1
s
rj
ri
U
n
t
+
+1
T
in.
-t
ri
+1
o
+1
o
J,
n
'^
ri
+1
00
o
o
+1
d
c
Ti
d
+1
00
ri
OO
r-
'rf
'■~^
^
—
II
g
11
II
oo
II
II
II
1
II
II
c/1
O
II
II
II
a:
<
u
n<
rv
lY.
£
Di
O
U
O
o
<
o
U
U-
a
<
Z
o
^
z
z
•?■
u
8
8
C/5
8
00
U
n
u
o
O
0-
c
O
Z
c
^
-o
■a
c
a.
-a
^
r:
„^
a.
^
r5
S
E
S
S
a.
f-
U
O
u
o
t—
a.
r-
C/1
H
oi
OC
<
CJ
K
Di
^
Oi
O
U
O
u
<
o
U
u,
a:
<
^
O
Z
z
z
Modeling Feeding Behavior in Perna perna
129
of eiivironmeiits. Model predictions and observed data of FR. RR.
IR, NOSE. OCI. and AR of mussels in a range of TPM between 2
and 40 mg L '. are shown in Fig. 7 A. B. C, D. E. and F. respec-
tively, showing that predicted values satisfactorily reproduce the
main trends of feeding behavior observed in P. perna.
As bivalve feeding behavior is mainly controlled by concen-
tration and organic content of seston (Hawkins et al. 1998b). it is
likely that this model is sensitive to these forcing functions (TPM
and OCS). To verify the model sensitivity to changes in the coef-
ficients of the equation that predicts OCS as a function of TPM. we
ran the model three times, varying the coefficients values. Each
coefficient (EQ. ( I ). Table -■^. Fig. 2| was varied by ±10"^* from its
standard value, and the sensitivity was measured by the following
equation:
S = [x/x|/IP/P]
where (S) is a measure of sensitivity, x refers to model outputs at
the end of the integration period in the standard model, and r'»x is
the change in the value of x brought about by varying the model
Figure 3. Perna perna. The relationship between total particulate mat-
ter (TPM, mg !,"') and organic content of seston (OCS, fraction I and
(Al clearance rates (C'R I h '). (I?) filtration rate (FR, mg h"'), (C)
rejection rate (RR. mg h 'l, (I)) Ingestion rate (IR. mg h"'l, (E) net
organic absorption rate (N().\R, mg h 'l. Net organic selection effi-
ciency (NOSE, fraction) is plotted against particulated organic and
inorganic matter (PIM and POM, mg L"') (F). Refer to Table 3 for
equations and statistics.
Both the net absorption efficiency of ingested organics
(NAEIO, fraction) and the net organic absorption rate (NOAR. mg
h'') had a hyperbolic relationship with the organic content of
ingested matter (Fig. 4C and 5. Table ?}. NOAR was essentially
controlled by quantity (filtration rate) and quality (OCI) of food
passing through the digestive system (Fig. 4C, Table 3). The ab-
sorption rate across the experiments varied from 21.84 mg h"
(TPM .3.3.18 mg L '. OCS 0.18) to -0.69 mg h"' (TPM 10.09 mg
L"'. OCS 0.10).
The differential equations, logical functions, and starting values
of the state variables used to couple the equations describing the
filter-feeding processes for P. perna in STELLA are listed on
Table 4. We produced a robust model with relatively low com-
plexity and specificity. Figure 6A depicts the conceptual diagram
of the P. perna feeding process as a function of TPM and OCS.
The sub-model inserted inside the "ingested matter" variable (Fig.
6B ) reproduces the absorption of organic matter and the passage of
inorganic matter as inert material through the gut. As the model
was based on natural seston in both turbid and clear environments
and feeding rates measured in these environments, we believe that
it has incorporated feeding adaptations by P. perna for both kinds
B
Figure 4. Perna perna. The relationship between (.A I net organic se-
lection elTiciency (NOSE, fraction), total particulate matter (TPM. mg
L"') and organic content of ingested (OCI. fraction): (B) ingestion rate
(IR, mg h"'), TPM and net organic ingestion rate (NOIR. mg h'); (C)
net organic absorption rate (NO.\R, mg h"'), filtration rate (mg h"')
and OCI. Refer to Table 3 for equations and statistics.
130
SUPLICY ET AL.
O
• •
-0,2
0.0
02
1.0
12
1.4
04 06 08
OCI (fraction)
Figure 5. Penia perna. The relationship between the organic content
of ingested (OCI, fraction) and the net absorption efficiency from
ingested organics (NAEIO, fraction). Refer to Table 3 for equaliiins
and statistic.
coefficient. Similarly, the denominator measures the variation in
the coefficient of interest divided by its standard value. This equa-
tion compares the percentage change in the model outputs with a
given percentage change in one of the model parameters. The
value of (S) was averaged for positive and negative variations and
the results of the model outputs (absorbed matter, pseudofeces, and
feces produced) for the coefficients relating TPM and OCS are
shown in Table 5. The output most sensitive to variation in the
relationship between seston TPM and OCS was pseudofeces pro-
duction, as a result of increased or decreased rejection rate.
DISCUSSION
This study showed that P. perna, like other mussels, controlled
its feeding mechanisms to achieve an optimum organic absorption
rate independent of fluctuations in seston concentration and qual-
ity. It is important to note that the range of TPM recorded was
within normal values during the year for other bivalve aquaculture
locations in Southern Brazil (Suplicy, unpub. data). Therefore, the
TPM range experienced in the experiments and included in the
model are directly applicable to Brazilian shellfish famis condi-
tions. Although seasonal changes in feeding physiology were not
examined in this study, time series data of TPM, POM, and OCS
from 1998 to 2002 do not suggest strong seasonal changes in food
availability in the sub-tropical waters of Santa Catarina, (Suplicy et
al. unpublished data). Similarly, the condition inde.x of P. perna
does not follow a seasonal trend, as seen in Mylilus echtlis (Navanxi
& Iglesias 1995), because spawning occurs throughout the year
with small peaks in summer, autumn and spring (Marques et al.
1991 ). Therefore, we believe that the findings reported here can be
used to predict feeding physiology throughout the year.
Food availability (TPM and OCS) was the main forcing func-
tion of the models produced, therefore characterizing the available
seston is of primary importance to generate a model to predict food
uptake by P. perna. Data for Southern Brazil showed that the
organic content of available food decreased as TPM increased, a
common pattern in many estuaries and sheltered bays both in
teiTiperate and tropical waters (Hawkins et al. 1996a, 1998b). This
reduction of the organic proportion is a function of the dilution of
organic particles when resuspended silt increases particulate inor-
ganic matter on the water column (Frechette & Grant 1991. Wid-
dows et al. 1979)
The methods used in this study to estimate clearance rates of
filter feeders were less accurate than the methodology proposed by
Hawkins et al. (1998a, 1999) for measurements using natural
seston. The most appropriate method to accurately measure clear-
ance rates by bivalves is controversial (Cranford 2(M1. Riisgard
2001. Widdows 2001 ). As new methods are being developed, new
models about how these animals control their food uptake are
being produced. It is agreed that mussels do not always filter at
their maximal rate in their natural environment (Riisgard 2001.
Widdows 2001 ). This may be due to a regulation of feeding pro-
cesses in response to changes in quantity and quality of suspended
particles, salinity, temperature, and the presence of pollutants in
the water (Widdows 2001 ). In this study only ll^c of the variation
clearance rates of mussels using TPM and OCS as independent
variables was explained, and the significant proportion of the re-
maining variance in clearance rate in POM was not. In their ex-
periments, however, Hawkins et al. (1999) increased the amount of
the variability in clearance rate explained from 13-,').^'/f when they
included Chi and TPM as independent variables instead of only
POM. Although all precautions proposed by Iglesias et al. (1998)
in the use of the biodeposition method for suspension-feeding
TABLE 4.
Equations used in the formulation of feeding physiology model
in STELLA.
TPM = GRAPH (time-series)
OCS = 1/(2.55 -hO.47 * TPM)
PIM = 0.22 -1-0.81 * TPM
POM = TPM-PIM
PR = 68.77-0.12 TPM-370.10 OCS -i- 0.07 TPM" -i- 565.80 OCS"
Fillered matter (t) = Filtered matter (t - dt) + (FR - RR - IR) * dt
INIT Filtered matter = 219.81
RR = 52.43 + 0.97 TPM-362.47 OCS -i- 0.02 TPM" + 589.79 OCS"
Pseudofeces (5) = pseudofeces (t - dt) -f (rejection) * dt
IR = filtration-rejection
Ingested (t) = ingested (t - dt) + (ingestion' - NIIR - NOIR) * dt
INIT ingested - 36.46
NOIR = 1.37 - (.1.23 TPM -i- 0.1 1 IR -i- 0.01 TPM- + 0.004 IR"
NIIR = mgested-NOIR
Inorganic (t) = inorganic (t - dt) ■(- (NIIR - IM on gut) * dt
Organic (t) = organic (t - dt) + (NOIR - OM on gut) * dt
OM on gut = organic
Im on gut = inorganic
INIT organic = 13.17
INIT inorganic = 23.29
Ingested matter = food on gut + organic + ingested -i- inorganic
Food on gut (1) = food on gut (t - dt) -i- (OM on gut -i- IM on gut -
absorption' - egestion') * dt
INIT food on gut =
NOSE = 0.30 - 0.21 PIM -i- 1.03 POM + 0.01 PIM" - 0.20 POM-
OCI = 0.13 - 0.001 TPM + 0.27 NOSE -t- 0.0002 TPM" -i- 0.19
NOSE-
NOAR = -2.62 -I- 0.012 RF -i- 15,73 OCI -i- 0.0001 FR- - 9.22 OCI"
Ahsorption' = NOAR
Absorption = absoiption'
Absorbed matter (t) = absorbed matter (t - dt) + (absorption) * dt
INIT absorbed matter =
Egestion' = IM on gut + (NOIR-NOAR)
Egestion = egestion'
Feces (t) = feces il - do -i- (cgestioni * dt
Modeling Feeding Behavior in Perna perna
131
absofbed matter
-^
pseudofaeces
B
organic
organic matter
Figure 6. (A) Diagram of the feeding processes of a general niter-fceding bivalve, used on the modeling of P. perna feeding physiology. (B)
Diagram of the sub-model of a mussel gut showing the absorption of organic matter and feces production. Refer to Tables 2 and 3 for variables
and acronyms and Table 4 for logical and differential equations.
measurements were taken in this study, it seems that the new
methodology proposed by Hawkins et ai. (1998b, 1999) is more
appropriate for studies using natural seston. It seems that qualita-
tive features of seston may be just as important as availability of
food in mediating feeding responses (Hawkins et al. 1998b). The
general trend for decreasing clearance rates as seston concentra-
tions increase, however, is seen in other studies (Hawkins et al.
1999, Hawkins et al. 1998b, Wong & Cheung 2001). There are
many methods to quantify concentration and organic content of
seston in feeding experiments. Most use mass measurements of
total particulate matter available in the seston (TPM, mg L"'),
pailiculate organic matter available in the seston (POM, mg L"'),
and the ratio between these two variables, which is the organic
content of seston (OCS. fraction). Recent findings suggest that
clearance rate is primarily dependent on seston availability mea-
sured in terms of total volume, rather than mass. This helps to
explain the confusing variation in clearance rate reported by many
studies and stresses a need to consider volumetric constrains in
bivalve feeding studies (Hawkins et al. 2001 ). More detail about
the seston organic fraction can be obtained if the carbon;nitrogen
ratio is measured, which can vary from <4 to >26 (Bayne &
Hawkins 1990). The measurement of the biologically available
132
SUPLICY ET AL.
160
140
-. 120
E
— 80
cr
^ 60
1
08 -
I"'
o
m
^ 04
UJ
to
O 02
00
E,
a: 40
TPM (mg I"')
10 20 30 40 50
TPM (mgr')
Figure 7. Predictions of the P. pcnia filter-feediii}; model produced on
STELLA. (A» nitration rate IFR, nig h '), (B) rejection rate (RR, nig
h"'). (CI ingestion rate (IR. mg h '), ID) selection efficiency (NOSE,
fraction), (E) organic content of ingested matter (OCL fraction), and
(F) net organic absorption rate (NOAR, mg h~'), in the range of total
particulate matter (TPNL mg L ') observed in this study.
organic carbon and nitrogen in the water and in associated biode-
posits can provide, not only more accurate measurements of the
clearance rate, but also important information about the absorption
of these elements by filter feeders.
The biodeposiljon approach demands that the gut residence
time is correctly calculated to generate accurate physiologic feed-
ing rates. As starved animals were used to estimate gut passage
time this may have over-estimated the normal passage time. How-
ever, our estimates are comparable to those from other biodepo-
sition studies using Penui canaliculus, in which the gut passage
time for non-starved mussels was 80 min. and no delay time was
assumed for Perna viridis (Hawkins et al. 1998al.
Perna perna appeared to selectively enrich the organic content
of ingested matter by rejecting particles of higher inorganic con-
TABLE 5.
Sensitivity analysis of absorbed matter, pseudofeces and feces
production for the coefficients a and h in the equation OCS = l/(a +
b * TPM).
Absorbed
Matter
Pseudofeces
Feces
0.2(12
0.2-^2
0.(1.^7
0.7-14
0.118
0. 1 36
tent before ingestion. This selection efficiency was a function both
of filtration rate and the proportion between inorganic and organic
particulated matter available in the water. The increase in selection
efficiency at higher filtration rates is important, because this helps
to maintain nutrient acquisition independent of fluctuations in
seston organic content (Hawkins et al. 1998a). Extreme values of
net organic selection efficiency measured in this study (NOSE >l
or <0) must be considered with caution as they are probably mea-
surement errors associated inadvertently with collecting settled
sediment when collecting biodeposits. This would effectively alter
the organic ratio of pseudofeces. Extreme values were observed in
15% of measurements. Nevertheless, NOSE values recorded in
this study (>0.7) suggest that P. perna is efficient in selecting
organic particles available in the seston. Hawkins et al. (1996a)
recorded NOSE values of up to 0.5 in M. edulis. and Hawkins et
al. ( 1998b) report maximum NOSE of 0.7 for P. viridis.
Maximum net organic ingestion rate (NOIR) recorded for P.
perna was 24.05 mg h"' and occurred when TPM was 33.93 mg
L"' and OCS was 0.18. This is similar to values obtained for P.
canaliculus in New Zealand, that showed maximum organic in-
gestion rate of 27.3 ± 6.3 mg h"' (Hawkins et al. 1999), and for P.
viridis in Malaysia with a recorded rate of 24.8 ± 3.6 mg h"'
(Hawkins et al. 1998a). These rates are considerably higher than
the maximum organic ingestion rate of 6.5 mg h"' reported for M.
edulis (Hawkins et al. 1997). The growth rates of P. perna in
southern Brazil are among the fastest reported for mussels in the
Perna genus, reaching commercial size (80 mm) in 8-10 mo (Su-
plicy. unpub. data). This rapid growth is probably related to higher
weight-specific rates of energy acquisition and higher water tem-
peratures in the sub-tropical waters of southern Brazil.
Data from this study suggested that P. perna takes advantage of
the abundant organically rich seston available in Brazilian waters
throughout the year by maintaining high ingestion rates. There is
evidence that when ingestion rate is high absorption efficiency is
high and gut residence time is short (Bayne et al. 1988). Fuilher-
more, the proportion of gut volume occupied by ingesta may vary,
thereby facilitating an increase in absorption efficiency with little
change in the gut passage time (Bayne et al. 1987). Widdows et al.
( 1979) report that absorption efficiency declines as ingestion rate
increases and food progresses from the digestive gland to the in-
testine. However, this pattern may be counterbalanced by elevated
organic content of ingested matter due to selection processes (this
study, Hawkins et al. 1999) that positively increase the absorption
efficiency and ultimately the absorption rate. Similarly to the con-
siderations raised for NOSE values, negative absorption rate val-
ues are not biologically meaningful and must be considered with
caution as these could be caused by collection of inorganic sedi-
mented material together with mussel feces. Negative absorption
rates were measured in 7% of measurements.
The integration of all equations from Table 4 with STELLA
software resulted in a reductionistic and deterministic non-linear
model that reproduces the feeding processes of P. perna in both
clear and turbid environments. The general conceptualization of
the diagram was based on the description of the bivalve filter-
feeding process provided at the TROPHEE workshop (Bayne
1998. Hawkins et al. 1998b), and final equations were based on
intensive measurements that enabled calibration of the outputs.
This feeding model may not be a perfect reproduction of the bi-
valve feeding process, but the objective is to provide a useful tool
to understand and predict feeding processes of this species. The
model includes a complete sequence of steps in the feeding process
Modeling Feeding Behavior in Pekna perna
133
that may cause an accumulation of predictive error (Grant &
Baciier 1998). Its value lies in the ability to provide an understand-
ing of the interaction between a mussel farm and the environment,
for example, the amount and organic content of biodeposits re-
leased into the water column and sediment beneath the farm.
Sensitivitv analysis indicated that model predictions of ab-
sorbed matter and feces production were less affected by changes
in the relationship between TPM and OCS than model prediction
of pseudofeces production. This analysis suggests that predicted
absorption would stay reasonably invariable if the model is applied
to environments with different seston concentration and organic
content. Therefore, mussels maintain a reasonably constant or-
ganic ingestion rate in varying seston conditions by compensating
for low organic content of the seston through adjusting selection
efficiency and rejection of inorganic matter as pseudofeces.
This feeding model can be used as an important tool for the
understanding of how P. pcniu interact with the culture environ-
ment. Current studies are under way to integrate this feeding model
w ith energy budget and population dynamics of P. perna. Further
coupling of the P. perna biologic models with physical models of
seston hvdrodynaniics and models of primary production are also
planned, and this approach will allow the development of cairying
capacity analysis for suspended mussel culture in sub-tropical en-
vironments like the southern Brazilian coast.
ACKNOWLEDGMENTS
The research was supported by CNPq, a Brazilian govcrnnienl
agency for scientific and technologic development. The authors
thank two anonymous reviewers for their valuable criticism and
comments of the original manuscript.
LITERATURE CITED
Bayne. B. L. & A. J. S. Hawkins. 1990. Filter-feeding in bivalve molluscs:
control of energy balance. In: J. Mellinger. editor. Animal nutrition and
transport processes. I. Nutrition in wild and domestic aniniaK. Karger.
Basel, pp. 70-8.^.
Bayne. B. L. & R. C. Newell. I9S3. Physiological energetics of marine
molluscs. In: K. M. Wilbur. A. S. Saleuddin. editors. The Mollusca.
Vol. 4. New York: Academic Press, pp. 407-5 Li.
Bayne. B. L. 1993. Feeding physiology of the bivalves: lime-dependence
and compensation for changes in food availability. In: R. Dame, editor.
Bivalve filter feeders in estuarine and coastal ecosystems processes.
NATO ASI Series. G 33 Berlin: Springer-Verlag. pp. 1-24.
Bayne. B. L. 199X, The physiology of suspension feeding by bivalve mol-
luscs: an introduction to the Plymouth 'TROPHEE" workshop. J. Exp.
Mar. Biol. Ecol. 219:1-19.
Bayne. B. L.. A. J. S. Hawkins & E. Navarro. 1987. Feeding and digestion
by the mussel Myliliis edulis L. (Bivalvia: Mollusca) in mixtures of silt
and algal cells at low concentrations. / Exp. Mar. Biol. Ecol. 1 1 1 : 1-22.
Bayne. B. L., A. J. S. Hawkins & E. Navarro. 1988. Feeding and digestion
in suspension-feeding bivalve molluscs: the relevance of physiological
compensations. Am. Zool. 28:147-159.
Bayne. B. L.. J. I. P. Iglesias. A. J. S. Hawkins. E. Navarro. M. Heral & J.
M. Deslous-Paoli. 1993. Feeding behaviour of the mussel. Mylilus
(■(/»/«: respon,ses to variations in quantity and organic content of the
seston. J. Mar. Biol. Ass. U.K. 73:813-829.
Bayne. B. L., D. W. Klumpp & K. R. Clarke. 1984. Aspects of feeding,
including estimates of gut residence time, in three mytilid species (Bi-
valvia, Mollusca) at two contrasdng sites in the Cape Peninsula. .South
Africa. Oecologia 64:26-33.
Berry, P. F. & M. H. Schleyer. 1983. The brown mussel Perna pcnia on
the Natal coast. South Africa: utilization of available food and energy
budget. Mar. Ecol. Prog. Ser. 13:201-210.
Brylinski. M. & T. W. Sephton. 1991. Development of a computer simu-
lation model of a cultured blue mussel {.ifyiilns edulis) population.
Can. Tech. Rep. Fish, and .\qualic Sci. 1805 (vii + 88 pp.).
Cranford. P. J. 2001. Evaluating the "reliability" of filtration rate measure-
ments in bivalves. Mar. Ecol. Prog. Ser 215:303-305.
Frechette. M. & J. Grant. 1991. An in situ estimation of the effect of the
wind-driven resuspension on the growth of the mussel Mytihis edulis L.
/ E.xp. Mar. Biol. Ecol. 148:201-213.
Gardner. J. P. A. & R. J. Thompson. 2001. Naturally low seston concen-
tration and the energy balance of the Greenshell mussel (Perna canali-
cuius) at Island Bay. Cook Strait. New Zealand. N.Z. J. Mar. Frcsliw.
Res. 35:457-^68.
Gardner. J. P. A. 2002. Effects of seston variability on the clearance rate
and absorption efficiency of the mussel Aulacomya maoriami. Mylilus
galloprovincialis and Perna canaliculus from New Zealand. / Exp.
Mar. Biol. Ecol. 268:83-101.
Grant. J. & C. Bacher. 1998. Comparative models of mussel bioenergetics
and their validation at field culture sites. / Exp. Mar. Biol. Ecol.
219:2I-t4.
Grant. J. 1993. Working group report: modelling. In: R. Dame, editor.
Bivalve filter feeders in estuarine and coastal ecosystems processes.
NATO ASI Series, G 33 Beriin: Springer-Verlag. pp. 549-555.
Griffiths. C. L. & R. J. Griffiths. 1987. Bivalvia. In: T.J. Pandian. F.J.
Vemberg. editor. Animal energetics. Vol. 2: Bivalvia through Reptilia.
New York: Academic Press, pp. 1-88.
Hawkins. A. J. S., B. L. Bayne, S. Bougrier, M. Herat. J. I. P. Iglesias. E.
Navarro. R. F. M. Smith & M. B. Urrutia. 1998a. Some general rela-
tionships in comparing the feeding physiology of suspension-feeding
bivalve molluscs. J. E.xp. Mar. Biol. Ecol. 219:87-103.
Hawkins. A. J. S.. J. G. Fang, P. L. Pascoe. J. H. Zhang. X. L. Zhang & M.
Y. Zhu. 2001a. Modelling short-term responsive adjustments in particle
clearance rate among bivalve suspension-feeders: separate unimodal
effects of seston volume and composition in the scallop Chlamys fur-
reri. J. Exp. Mar. Biol. Ecol. 262:61-73.
Hawkins, A. J. S.. J. G. Fang, P. L. Pa.scoe, J. H. Zhang. X. L. Zhang & M.
Y. Zhu. 2001b. Modelling shon-term responsive adjustments in par-
ticle clearance rate among bivalve suspension-feeders: separate unimo-
dal effects of seston volume and composition in the scallop Chlamys
farreri. J. E.xp Mar. Biol. Ecol. 262:61-73.
Hawkins. A. J. S.. M. R. James, R. W. Hickman. S. Hatton & M. Weath-
erhead. 1999. Modelling of suspension-feeding and growth in the
green-lipped mussel Perna canaliculus exposed to natural and experi-
mental variations of seston availability in the Marlborough Sounds.
New Zealand. Mar Ecol. Prog. Ser. 191:217-232.
Hawkins, A. J. S., P. N. Salkeld, B. L. Bayne & E. Gnaigeri. 1996a. Novel
observations underlying the fast growth of suspension-feeding shellfish
in the turbid en\lronments: Mylilus edulis. Mar. Ecol. Prog. Ser. 131:
179-190.
Hawkins, A. J. S.. R. F. M. Smith. B. L. Bayne & M. Heral. 1996b. Novel
observations underlying the fast growth of suspension-feeding shellfish
in the turbid environments: Mylilus edulis. Mar. Ecol. Prog. Ser. 131:
179-190.
Hawkins, A. J. S., R. F. M. Smith, S. Bougrier. B. L. Bayne & M. Heral.
1997. Manipulation of dietary conditions for maximal growth in mus-
sels. Mylilus edulis L., from the Marennes-Oleron Bay, France. Aquat.
Living Resour. 10:13-22.
Hawkins. A. J. S.. R. F. M. Smith. S. H. Tan & Z. B. Yasin. 1998b.
Suspension-feeding behaviour in tropical bivalve molluscs: Perna viri-
dis. Crassotrea helcheri. Crassostrea iradelei. Saccostrea cucculala
and Pinctada margariferu. Mar. Ecol. Prog. Ser. 166:173-185.
Iglesias. J. I. P.. M. B. Urrutia, E. Navarro & I. Ibarrola. 1998. Measuring
feeding and absorption in suspension-feeding bivalves: an appraisal of
the biodeposition method. J. E.xp. Mar. Biol. Ecol. 219:71-86.
134
SUPLICY ET AL.
James, M. R.. M. A. Weatherhead & A. H. Ross. 2001. Size-specific
clearance, excretion and respiration rates, and phytoplankton selectivity
for the mussel Penia caiuiliciihis at low le\els of natural food. N.Z. J.
Mar. Fresim: Res 35:73-86.
Marques. H. L. de A.. R. T. L. Pereira & B. C. Correa. 1991. Study on
reproductive and settlement cycles of Perna pema (Bivalvia: Myril-
idae) at natural beds in Ubatuba .shore — Sao Paulo Stale. Brazil. Bo-
letim Institulo Pesca Sao Paulo. 18:73-81.
Navarro, E. & J. 1. P. Iglesias. 1995. Energetics of reproduction related to
environmental variability in bivalve molluscs. Haliotis 24:43-55.
Navarro, E., J. 1. P. Iglesias, A. Perez Camacho & U. Labarta. 1996. The
effect of diets of phytoplankton and suspended bottom material on
feeding and absorption of raft mussels {Mylihis gallopioviiiciali.s
Lmk.). J. Exp. Mar. Biol. Ecol. 198:175-189.
Navarro, E., J. I. P. Iglesias, A. Perez Camacho. U. Labarta & R. Beiras.
1991. The physiological energetics of mussels iMyriliis galloprovin-
cialis Lmk) from different cultivation rafts in the Ria de Arosa (Galicia.
NW Spain). Aquaculnire 94:197-212.
Newell. C. R. & D. E. Campbell. 1998, MUSMOD©, a production model
for bottom culture of the blue mussel, Mytilus edulis. J. Exp. Mar. Biol.
Ecol. 219:171-203.
Perez Camacho, A. & R. Gonzales. 1984. La filtracion del mejillon {Myri-
lus edulis L.) en laboratorio. In: Actas do Primeiro Seminario de Cien-
cias do Mar. As Rias Galegas, Cuadernos da Area de Ciencias Mannas,
Seminario de Estudos Galeaos, 1:427—437.
Riisgard, H. U. 2001. On measurements of filtration rates in bivalves — the
stony road to reliable data: review and interpretation. Mar. Ecol. Prog.
Ser. 211:275-29L
Scholten. H. & A. C. Smaal. 1998. Responses of Mytilii.^ edulis L. to
varying food concentrations: testing EMMY, an ecophysiological
model. J. Exp. Mar. Biol. Ecol. 219:217-239.
Smaal. A. C. & H. A. Haas. 1997. Seston dynamics and food availability
on mussel and cockle beds. Esruarine. Coastal Shelf Science 45:247-
259.
Smaal. A. C, T. C. Prins, N. Dankers & B. Ball. 1998. Minimum require-
ments for modelling bivalve carrying capacity. Aijuai. Ecol. 31:423-
428.
van Erkon Schurink, C. & C. L. Griffiths. 1992. Physiological energetics of
four South African mussel species in relation to body size, ration and
temperature. Coinp. Biochem. Physiology 10IA:779-789.
Widdows, J. 2001. Bivalve clearance rates: inaccurate measurements or
inaccurate reviews and misrepresentation? Mar. Ecol. Prog. Ser. 221;
303-305.
Widdows, J., P. Fieth & C. M. Worral. 1979. Relationships between seston,
available food and feeding activity in the common mussel Mytilus
edulis. Mar Biol. 50:195-207.
Wong, W. H. & S. G. Cheung. 2001. Feeding rhythms of the green-lipped
mussel, Perna viridis (Linnaeus. 1758) (Bivalvia: Mytilidae) during
spring and neap tidal cycles. / E.\p. Mar. Biol. Ecol. 257:13-36.
Joiinuil „f Shflllhh Research. Vol. 22. No. 1, 135-140. 200.^.
PHENOTYPES OF THE CALIFORNIA MUSSEL, MYTILUS CAUFORNIANVS, CONRAD (1837)
JORGE CACERES-MARTINEZ,'* MIGUEL A. DEL RIO-PORTILLA.'
SERGIO CURIEL-RWIIREZ GUTIERREZ,' AND IGNACIO MENDEZ GOMEZ HUMARAN"
^ Departamento de Aciiicultura del Centra de Investigacion Cientifica y de Ediicacion, Superior de
Ensenada. A. P. 2732 C. P. 22860 Ensenada. Bajci California. Mexico: 'Instiruto Nacional de la Pesca,
Pitdi^oras 1320 6° Piso, Col. Sia. Cruz Aroyac. C.P. 03310. Mexico D.F.
ABSTRACT The morphological variability of Mytilii.s ediilis complex species has been the subject of a variety of studies. However,
the morphological variability of Mytiliis califomiumis has not been studied. We found that there are some M. californiamis without
some of the shell characteristics mentioned by Conrad ( 1837) in the original description of this species. The most remarkable difference
was the absence of radial ribs on the exterior of the shell: thus, we tested the presence of at least two phenotypes in M. ealiforniainis.
Six hundred ninety five M. ealiforniainis of different sizes were collected from the locations La Mina del Fraile. La Bufadora, and La
Salina in Baja California. For comparison, 58 M. i^atloprovincialis were collected from an aquaculture facility at Rincon de Ballenas
in Bahi'a de Todos Santos. Baja California. Fourteen morphometric measures and the weight of the shell were measured and a principal
coinponent analysis (PCA) and a logistic regression (LR) were carried out to tlnd differences between mussels studied and for obtaining
a prediction to assign the phenotypes. The presence of ribs, small ligament margin, a narrow posterior byssal retractor muscle scar, and
shell weight were the discriminating characters between two groups in M. californiamis. These findings confirm the presence of at least
two phenotypes in this species, in all mussel sizes and the studied locations. The LR correctly assigned 99.28% of the shells to each
phenotype. and it considered only eight out of the fifteen morphometric measures. The PCA showed a clear morphologic difference
between both phenotypes of A/, ealifornianiis and A/, gulloprovineialis. The original description of this species by Conrad in 1837 was
done taking into account only the phenotype with ribs.
KEY WORDS: Mytilus ealifornianiis. Mytiliis eihilis complex species, morphological variability, phenotypes
INTRODUCTION
MATERIALS AND METHODS
The marine mussels of the genus Mytilus are widely distributed
in boreal and temperate waters of the Northern and Southern
Hemispheres (Soot-Ryen 1955), Prior to protein separation and
molecular genetics, about nine species of the genus Mytilus were
recognized (Gosling 1992). Today, about five species are consid-
ered belonging to this genus: Mytilus californiamis. Mytilus cor-
».«■;(.? (Gould 1861), Mytilus edulis (Linne 1758). Mytilus gallo-
provincialis and Mytilus. trossulus (Gould 1850) (Seed 1992), The
three later species are considered to be the M. edulis complex
species because they are very close in their external shell mor-
phology. These species have caused a variety of studies for their
differentiation, taking into account shell morphology, allozyme,
and molecular genetics (Beaumont et al, 1989. Figueras &
Figueras 1983, McDonald & Koehn 1988. Koehn 1991, McDonald
etal. l99l,Gelleretal. 1994. Inoue et al, 1995. Rawson & Hilbish
1995. Ohresser et al, 1997), Mytilus californianus has never been
questioned as a separate species from the Mytilus edulis-comp\ex
because of its characteristic radiating ribs, strong growth lines, and
heavy shell in larger specimens: these characters allow easy dif-
ferentiation from the other species in adult stage (Soot-Ryen 1955.
Koehn 1991). During a field study of Mytilus californianus in an
exposed rocky shore of the West Coast of Baja California, Mexico,
we found some specimens with typical external characteristics of
the shell described by Conrad in 1837. Other individuals, however,
showed a smooth shell without coarse ribs, similar to the M. edulis
complex form, but with heavy shells. A question arises from this
observation, are there two or more phenotypes of M. califor-
nianus'l This study focused on answering this question.
*Corresponding author. Mailing address: Department of Aquaculture.
CICESE. PO Box 434844, San Diego, CA 92143
In March 1997, 129 M. californianus (size range from 16,8-
1 13,5 mm, mean size 59,1 mm) were collected from an exposed
rocky shore along the intertidul zone during low tide in La Mina
del Fraile. B. C, Mexico, In August 2()(X), 278 mussels were col-
lected from La Salina (size range from 27.6-98.1 mm, mean size
56.9 mm) and 288 from La Bufadora (size range from 44.7-88.1
mm. mean size 54.1 mm). B. C. Mexico, both areas exposed rocky
shores, and the mussels were collected during low tide along the
intertidal zone. Additionally. 58 M. galloprovincialis were ob-
tained from culture long-lines placed at Bahi'a de Todos Santos,
B,C. (size range from 47.2-85.3 mm. mean size 61.4 mm) and they
were used to compare the morphological characteristics with M.
californianus (Fig. 1 ).
The shells of ail mussels were cleaned with a brush and water
stream and dried in an oven at 40"C overnight. The following
morphometric dimensions were measured for differentiation
among mussel groups and species (Fig. 2): number of ribs on the
external shell (rib), maximum shell length (si), height (sh) and
width (sw), the position of maximum shell width (a) along the
dorso- ventral axis, the maximum dimensions of the anterior (aams)
and posterior (pam) adductor muscle scars, the maximum length
(Ibr) and width (wbrs) of the posterior byssal retractor muscle scar,
the location of the center of the posterior adductor muscle scar
along both the anterior-posterior (pam-pm) and dorso ventral
(pam-vm) axes, the size of the hinge plate (hp) and number of
hinge teeth, the distance between the palial line and the \entral
shell margin (pl-vm) midway along the shell, and ligamentary
margin (Im), All measurements were taken with an electronic digi-
tal caliper to the nearest 0. 1 mm and were in accordance with those
taken by Beaumont et al, ( 1989), The dry shell weight (w) was also
measured for all mussels and it was included in the analyses, A
principal component analysis (PCA) was carried out to discrimi-
nate between phenotypes, followed by a logistic regression (LR)
135
136
Caceres-Martinez et al.
_32"3r
Pacific Ocean
V La Salina
\ Baja California
_32«
^^ <^\ Ensenada
Baja V
California'
Mussel culture facilityN.
• a]
La Bufadora \
_30°3r
^- — f
1 1 ,„ /La mina del Fraile
1 4
Figure 1. Map showing the three exposed rod^y shore localities where
Mytihis californiaiius was collected: La Mina del Fraile. La Bufadora.
and La Salina. The blue mussel yfyliliis f;allopr(niiiciali\ was collected
from a culture facility at Bahia de Todos Santos, Baja California.
Mexico.
(Sokal & Rohlf 1995) to fit the mussel phenotypes. A two way
ANOVA was used to determine possible differences between mus-
sel size among locations and phenotypes. A comparison through
the PCA between both phenotypes of M. culifonmimis with M.
gaUoprovincialis was carried out. These analyses were done using
the JMP statistical package by SAS Institute Inc.
RESULTS
Fifteen piincipal axes were extracted from the morphological
and shell weight data of M. califomiamts (Table I). The first
component explained 16% of total variance and was considered as
a size axis. A low correlation of size with number of ribs sug-
gests that the number of ribs does not change w ith mussel si/e. The
second component accounted for 1% of the variation indicating
morphological differences. Mussels with a high number of ribs
were correlated with this second component separating two groups
(Fig. 3). Also, in the second component, mussels with higher shell
weight (w). but with small ligament margin dm) and a naiTow
posterior byssal retractor muscle scar (wbrs) were coiTelated. The
rest of the components had eigen values smaller than the unit
accounting for about 139<^ of the total observed variance and thus,
no further explanation is necessary (Table 1 ). These data provide
statistical support to validate the presence of two phenotypes in M.
californiwms: A (with ribs) and B (without ribs), and they were
visually differentiated in mussels of different sizes (Fig. 4). After
separating both groups in all locations. 689f of the total mussels
belong to phenotype A and the rest to phenotype B.
Both phenotypes of M. caUfornianus were present in the three
locations. The two way ANOVA showed size differences among
mussel from different locations. (F -.f,^^ = 6.58. P = 0.001 ). but
the phenotype mean size was similar (F , ^^1) ~ 0.02. P = 0.892)
without interaction (F ,f,gg = 2.11. P = 0.122).
Once the PCA differentiated two phenotypes. the LR (Sokal &
Rohlf 1995) was used to determine whether it was possible to
assign any M. ctiUfonuanus to a particular phenotype. taking into
account morphological \ariables. excluding the number of ribs.
The LR considered only eight morphological measures from the
original fifteen to assign any mussel to a particular phenotype.
(X-
1 84.73. P < 0.0001 ; Lack of fit: x"
684.5. P
= 0.51). The coefficients of the eight morphometrical variables
were positive for: shell length (si = 0.104) and height (sh =
0.247). posterior adductor muscle scar (pain = 0.483). the dis-
tance between the palial line and the ventral shell margin (pl-vm
= 0.708). and weight (wO. 145); while the shell width (sw =
-0.281). the position of maximum shell width (a = -0.333). and
the ligamentary margin (Im = -0.348) were negative. After ap-
plying the LR we found that 99.28% were correctly assigned to
each phenotype. Thus, the visual. PCA and LR confirm the pres-
ence of two phenotypes in the Californian mussel.
Results of the PCA between morphometric data and v\eight of
Figure 2. Morphometric dimensions measured for Mytilus califoniiamis and Myliliix galldprovincialis: number of ribs on the external shell (ribi,
maximum shell length (si), height (sh) and width (sw), the position of maximum shell width (al along the dorso-ventral axis, the maximum
dimensions of the anterior (aams) and posterior (pam) adductor muscle scars, the maximum length (Ibrl and width (wbrs) of the posterior byssal
retractor muscle scar, the location of the center of the posterior adductor muscle scar along both the anterior-posterior (pam-pm) and dorso
ventral (pam-vm) axes, the size of the hinge plate (hp), and number of hinge teeth, the distance between the palial line and the ventral shell
margin (pl-vm) midway along the shell and ligamentary margin (Im).
Phenotypes of the California Mussel
137
TABLE 1.
Eigenvalues, explained variance (^rl. cumulative explained variance I "^i- 1 and eigenvectors (rounded to luo decimal places) from the
principal component analysis of Myliltis califoniUiiiiis niorphometric data from the Facillc coast of Baja California.
Principal Components
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Eigenvalue
11.45
1.07
0.59
0.45
0.29
0.25
0.22
O.IS
0,14
0,11
0.08
0.08
0.05
0.03
0.02
Variance!*)
76.31
7.12
3.93
3.02
1.94
1 .69
1.46
1.17
0.96
0,71
0.51
0.50
0.35
0.21
0.12
Cum. var.(%)
76.31
83.43
87.36
90.37
92.31
94.00
95.46
96.64
97.59
98.3
98.81
99.32
99.67
99.88
100.00
Eigenvectors
rib
0.02
0.95
0.17
-0.02
0.02
-0.02
-0.06
0.06
0.20
0,07
0,04
0,09
0.04
-0.03
0.01
si
0.29
-0.04
0.02
-0.15
-0.06
-0.10
-0.10
-0.03
0.20
0.03
-0. 1 3
-0. 1 2
0.19
0.16
-0.86
sh
0.28
0.05
-0.11
-0.01
0.04
0.01
0.03
0.32
0.07
-0.23
-0.16
-0.12
-0.82
0.15
-0.03
sw
0.29
-0.06
0.04
-0.09
0,02
-0.12
-0(16
-0.12
0,04
-0,01
0,24
0, 1 7
-0.17
-0.86
-0.11
a
0.27
-0.06
-0.27
0.09
0.14
-0.09
0.15
0.62
0.22
-0.30
0.07
0.26
0.42
-0.02
0.12
aams
0.26
0.01
0.05
-0. 1 1
-0.04
0.92
-0.14
0.03
-0.13
-0.09
0.06
0.01
0.10
-0.04
-0.01
pam
0.27
0.05
0.14
-0.33
0.29
-0.08
0.48
0.17
-0.39
0.34
0.24
-0.35
0.06
0.05
0.04
Ibr
0.28
-0.04
0.00
0.03
-0.08
-0.05
-0.20
-0.13
0.45
-0.04
0.00
-0.70
0.16
-0.07
0.35
wbrs
0.21
-0.13
0.76
0.59
0.04
-0.01
0.07
0.10
0.00
0.01
-0.05
0.06
0.00
0.04
-0.01
pam-pni
0.28
0.04
0.14
-0.31
0.00
-0.13
0.03
-0.17
-0.24
-0.20
-0.75
0. 1 3
0. 1 5
-0.08
0.22
pam-vm
0.28
0.00
0.15
-0.22
-0.04
-0.21
-0.17
-0.33
-0.12
-0.45
0.5 1
0.20
-0.01
0.38
0.11
hp
0.25
0.05
-0.34
0.35
0.61
0.09
0.16
-0.48
0.11
0.09
-0.07
0. 1 3
-0.02
0.10
-0.01
Im
0.27
-0. 1 8
0.04
-0.23
-0.16
0.00
-0. 1 3
0.04
0.33
0.64
-0.01
0,41
-0.10
0.21
0.23
pl-\iii
0.26
0.10
-0.27
0.32
-0.07
-0.17
-0,57
0.16
-0.54
0.23
0.00
-0.08
0.06
0.02
0.02
u
0.25
0.12
-0.25
0.28
-0.69
0.02
0.51
-0.20
-0.08
0.01
0.03
0.02
0.02
0.01
0.00
M. ccdifornianus and M.gaUoprovincialis are shown in Table 2.
The fist component explained 71% of the total variance and was
also considered a size axis. The number of ribs had a low corre-
lation with this axis. The second component explained 10% of the
total variance. The number of ribs (rib) and the shell width (sw)
were positively correlated with this component whereas the width
of the adductor muscle scar (wbrs) was negatively correlated.
Components 3 to 15 accounted for 19.4% of the total variance, but
their eigen values were smaller than one and they are not explained
further. The graphic presentation of the component scores shows a
Figure 3. Principal component scores plots between PC2 vs. PCI for
M. calif ornianiis. Phenotype A. open circles; Phenotype H, bold
squares.
clear difference between phenotypes of M. californianus and
among these phenotypes and M. gaUoprovincialis (Fig. 5).
DISCUSSION
Figure 4 shows different shell characteristics among M. cali-
fornianus specimens, and the PCA and LR support this visual
perception confirming that there are two phenotypes in M. cali-
fornianus. one with ribs and the other with a smooth shell, and
Figs. 4 and 5 show a morphological differentiation between both
phenotypes of M. californianus and M. galloprovincialis.
The original description by Conrad (1837) for Mytilus califor-
nianus was done from specimens collected by Thomas Nuttal in
upper California. Conrad describes "shell ovate elongated, in-
flated; anterior margin straight; posterior side emarginated; ribs
not very numerous, slightly prominent broad, rounded; lines of
growth very prominent"". This description agrees with phenotype A
studied here, where the rib number goes from 4 to 14 and they are
very prominent. In phenotype B. however, the ribs are not distin-
guishable and the growth lines are very prominent. Intraspecific
differences in shell sculpture on specimens from different habitats
have been noted in several gastropod species from the genus Lil-
torina (Struhsaker 1968. Johannesson et al. 199.3. Rush 1997).
These differences have been related to the degree of wave expo-
sure — extreme ribbed and with nodes forms live on dry raised
benches, not generally subject to horizontal water swash; while
extreme smooth forms predominate on low. moist benches subject
to strong wave swash. It is probable that a similar relation occurs
among M. californianus phenotypes and wave action or their po-
sition along the intertidal zone. We are carrying out a field study
to explore this. The presence of ribs has been conelated with shell
strength; the ribbed mussel Geukensia emissa has a stronger shell
than M. edulis. this strength was correlated with shell mass, shell
curvature and valve thickness (Majewski 1995). This could also be
138
Caceres-Martinez et al.
Figure 4. Mytiliis califoniianiis of different sizes showing the two phenotypes found in this study: ( Al with rihs and (B) no ribs. For comparison,
Mytilus gallopruvincialis of similar sizes also were included in figure (C). Note that different phenotypes appeared since young specimens.
TABLE 2.
Eigenvalues, explained variance ( 7r ), cumulative explained variance I '''< ) and eigenvectors from the principal component analysis between
Mytiliis califoniianiis and Mytilus galloprorincialis morphometric data from the Pacific coast of Baja California.
Principal Components
10
II
12
13
14
Eigenvalue
VarianceCvJ- )
Cum. var.(%l
rib
si
sh
sw
a
aanis
pam
Ihr
wbrs
pam-piTi
pam-vm
hp
Im
pl-vm
10.587
VCSSI
70.581
0.015
0.301
0.278
0.229
0.281
0.266
0.275
0.292
0.184
0.286
0.276
0.259
0.281
0.258
0.25.^
1.502
10.013
80.594
0.621
-0.043
-0.205
0.^94
0.016
0.000
0.05 1
0.037
-0.476
0.007
-0.201
0.030
-0.169
0.253
022.^
0.751
5.009
85.603
0.737
-0.010
0. 1 82
-0.339
-0.163
0.036
0.074
-0.083
0.428
0.091
0.190
-0.035
-0.094
-0.144
-0.101
0.459
3.063
88.665
0,001
-0.139
0.182
-0.291
0.220
-0.102
-0.334
-0.023
0.145
-0.326
-0.166
0.553
-0.223
0.325
0.274
0.3 I I
2.075
90.741
-0.044
-0.017
-0.268
0.279
-0.222
-0.114
-0.228
0.144
0.582
-0,061
-0.074
-0.373
-0.007
0.187
0437
0,277
1.844
92,584
-0,011
0,014
-0.142
0.269
0.073
-0.722
0.298
0.022
0.237
0.107
0.034
0.294
-0.128
0.074
-0,339
0.263
1 .755
94.139
-0.030
-0.118
-0.265
0.277
-0.169
0.592
0.048
0.030
0.287
-0,122
-0.197
0.336
-0. 1 .%
0.044
-0.432
0.216 0.173
1.442 1.155
95.781 96.937
Eigenvectors
-0.077
-0.125
-0.074
0.040
0.028
0.046
0.454
-0.179
0.094
-0.002
-0.211
0.222
-0.138
-0.585
0.516
0.020
-O.IOI
0.I7I
0.000
0.588
0.087
0.317
-0.216
0.185
-0.170
-0.344
-0.453
-0.075
0.215
-0.147
0.143
0.956
97.893
0.215
0.169
0.016
0.126
0.294
-0.100
-0.264
0.459
0.05 1
-0.318
-0.258
0.031
0.341
-0477
-0.129
0.100
0.667
98.560
0.052
0.007
-0.034
-0.211
-0.383
-0.094
0.344
-0.098
-0.020
-0.257
-0.309
0.093
0.652
0.268
0.027
0.079
0.527
99.086
0.062
-0.004
-0.077
0.077
0.166
-0.010
-0.388
-0.480
0.102
0.580
-0.317
0. 1 35
0.329
-0.053
-0.018
0.074
0.496
99.582
0.086
-0.161
-0.156
0.273
0.198
0.019
-0.080
-0.467
0.02!
-0.432
0.573
0.055
0.283
-0.062
0.037
0.045
0.299
99.882
-0.032
-0.209
0.764
0.467
-0.322
-0.065
-0.092
-Olio
-0.004
-0.073
-0.109
-0.019
-0.016
-0.074
-0.040
0.018
0.118
100.000
0.013
-0.866
-0.047
-0.074
0. 1 1 3
-0.006
0.036
0.353
-0.016
0.222
0.098
-0.012
0.210
0.016
0.000
Phenotypes of the California Mussel
139
CM
O
Q.
2
-2
4
■ ■
^ TtTT ▼
t'V
▼ ' T
-10
-5
5
PC1
10
15
Figure 5. Principal component scores plots between PC2 vs. PCI for
M. californiaiiiis Phenotype A, open circles; Phenotype B, bold
squares; and M. galloprmincialis bold triangles.
the case for M. caUfoniiiiniis where the presence of ribs might
indicate a stronger sheH.
In accordance with Seed (1968), variations in the M. cJiilis
shell form can be attributed to differences in age, habitat, growth
rate, and density. Old mussels have heavier shells, down-turned
divergent innboes, and varying degrees of incurvature of the ven-
tral shell margin than the young ones do. In this study, small and
large individuals showed similar morphometric characteristics;
therefore, the age or size of these mussels (which grow in the same
habitat) seems to have little influence on the variability of the
studied morphological characters. In relation to the habitat. Seed
(1968) comments that in areas free of predators (like the upper
shore) old individuals are common, whereas in areas where the
mussel turnover is rapid there is a predominance of young mussels.
Also, the presence of predators can affect shell morphology. M.
edulis has been found to ha\e a smaller shell lencth. heii;ht and
width with larger posterior adductor muscle, thicker shell, and
more meat per shell volume when a starfish was present (Reimer
& Tedengren 1996). In the Baja California region. M. califor-
niunus is the dominant species where there is high wave action,
whereas M. gallopravincialis is the dominant species in protected
bays with thinner shell and more meat than M. californianus
(Harger 1970. Harger 1972). It has been observed that shore level
has an influence on the morphology and physiology of M. gallo-
provincialis in the Adriatic see (Dalla Via et al 1987). Low shore
level mussels have higher and narrower shells and a higher dry
weight ratio whereas high shore mussels have a higher o.xygen
consumption rate. When cultivated Mytihis edulis was transplanted
between two different locations there were some morphological
differences that were considered to be due to genetic variation
(Stirling & Okumus. 1994). The same characters found in parents
of distinct ecotypes also occurred in progeny raised in the labora-
tory thereby indicating that the phenotypic differences have a ge-
netic basis (Struhsaker 1968). The presence of two phenotypes and
similar morphometric characteristics of the shell in small and large
M. californianus in all three locations indicates not only some
similarity among environments but it also strongly suggests that
the presence of ribs is genetically produced. To our knowledge.
there is no record on hybridization between M. californianus and
M. gallopravincialis. which could result in a heavy shell without
ribs. Our morphological results showed a clear difference between
both phenotypes of M. californianus and M. galloprovincialis,
which may suggest that phenotype B of M. californianus, is not the
result of hybridization with M. galloprovincialis. Further studies
using genetic markers will help to discard whether there has been
any degree of introgression between these two species due to hy-
bridization, which has been found in other Mvtilus species (Geller
et al 1994).
ACKNOWLEDGMENTS
The authors thank Antonio Figueras Jr., Antonio Figueras
Montfort. Andy Beaumont; for encouraging us to finish this study,
and Biologist R. Vasquez Yeomans from CICESE for his help with
the sample analysis. This work was supported by projects numbers
623106 and 6231 13 of CICESE.
LITERATURE CITED
Beaanuiiil. A. R.. Seed R. & P. Garci'a-Martfnez. 1989. Electrophorelic
and morphometric criteria for the identification of (he mussels Mytilits
edulis and M. gcdloprovincialis. In: J. S. Ryland & P. A. Tyler, editors.
Reproduction. Genetics, and Distribution of Marine Organisms. Den-
mark: Olsen and Olsen. pp. 251-258
Conrad T. A. 1837. Description of new manne shells, from upper Califor-
nia. Collected by Thomas Nuttall Esq. J. Acad. Nal. Sci. Phila. 7:227-
242
Dalla Via G. J.. U. Tappemer. & G. Bilterlich. 1987. Shore-level related
morphological and physiological variations in the mussel Mytiliis gtd-
loprinincialis (Lamarck. 1819) (Molusca Bivalvia) in the north Adri-
atic Sea. Mond. Zool. Ilal. 21:293-305
Figueras A. & A. J. Figueras. 1983. Variabilidad ecomorfica del mejilkin
silvestre y cultivado en Espafia (gen. Mytilii.s) y relacion con su
posicion sistematica. Investigacion Pesquera Al:yi-li
Geller J. B.. J. T. Carlton & D. A. Powers. 1994. PCR-based detection of
m( DNA haplolypes of native and invading mussels on (he nor(heas(ern
Pacific coast. Lati(udinal pauern of invasion. Mar. Biol. 117:243-249
Gosling. E. M. 1992. Systematics and geographic distribution of Mytiliis.
In: E. Gosling, editor. The mussel Mytihis: ecology, physiology, ge-
netics, and culture. A(ns(erdam: Elsevier Science Publishers B.V. pp.
1-20
Harger, J. R. 1970. Comparisons aniong grow(h charac(eris(ics of (wo spe-
cies of sea mussel. Mytiliis edulis and Mytiliis califoniianiis. The Ve-
liger 13:44-56
Harger. J. R. 1972. Compe(i(ive coexistence: maintenance of interacting
associations of the sea mussel. Mytiliis edulis and Mytihis califor-
nianus. The Veliger 14:387-410
Inoue, K.. S. Odo. T. Noda. S. Nakao. S. Takeyama. E. Yamaha. F. Ya-
masaki & S. Harayama. 1995. A possible hybrid zone in the Mytihis
edulis complex in Japan revealed by PCR inarkers. Mar. Biol. 128:91-
95
Johannesson, K., B. O. Johannesson & E. Rolan-Alvarez. 1993. Morpho-
logical differentiation and genetic cohesiveness over a microenviron-
mental gradient in the marine snail Littorina sa.\atihs. Evolution 47:
1770-1787
Koehn. R. K. 1991. The genetics and taxono(ny of species in the genus
Mytihis. Aquaculture 94:125-145
Majewski, M. 1995. Comparahve mechanical shell strength of the blue
mussel a Mylilus edulis and the ribbed mussel Ceukensia demissa. In:
J. P. Grassle. A. Kelsey. E. Dates & P. V. Snelgrove. editors. 23rd
140
Caceres-Martinez et al.
Benthic Ecology Meeting. New Brunswick, New Jersey. (Abstract
only).
McDonald, J. H. & R. K. Koehn. 1988, The mussels Mylilus gallopiovin-
cialii and M. trossulus on the Pacific coast of North America. Mar.
Biol. 99:111-118
McDonald. J. H., R. Seed & R. K. Koehn. 1991. Allozyme and morpho-
metric characters of three species of Mylilus in the Northern and South-
em hemispheres. Mm: Biol. 1 1 1:323-335
Ohresser, M., P. Borsa & C. Delsert. 1997. Intron-length polymorphistn at
the atin gene locus mac-1: a genetic marker for population studies in
the marine mussels Mytihis galloproviniiulis Lmk and M. ('(/»//,v L.
Mol. Mar. Biotech. 6:123-130
Rawson, P. D. & T. J. Hilbish. 1995. Distribution of male and female
mtDNA lineages in population of blue mussels, Mytilus trossulus and
M. galloprovincialis. along the pacific coast of North America. Mar.
Biol. 124:245-250
Reimer, O. & M. Tedengren. 1996. Phenotypical improvement of morpho-
logical defenses in the mussel Mytilus ediilis induced by exposure to
the predator Asterias rubens. Oikos 75:383-390
Rugh, S. N. 1997. Differences in shell morphology between the sibling
species Liltorina scutulata and Litlorinu plena (Gastropoda; Piosobran-
chia). The Veliger 40:350-357
Seed. R, 1968. Factors influencing shell shape in the mussel Mylilus edulis.
J. mar. bid Ass. U.K. 48:561-584
Seed, R. 1992, Systematics evolution and distribution of mussels belonging
to the genus Mylilus: An overview. Am. Malacol. Bull. 9:123-137
Soot-Ryen, T. A. 1955. A report on the family Mytilidae (pelecypoda).
.Mlait Hancock Pacific 20:1-175
Sokal, R. R. & F. J. Rohlf. 1995. Biometry. 3rd ed. New York: W. Freeman
and Company. 887 pp.
Stiriing, H. P. & 1. Okumus. 1994. Growth, mortality, and shell morphol-
ogy of cultivated mussel (Mytilus edulis) stocks cross-planted between
two Scottish sea lochs. Mar. Biol. 119:115-124
Struhsaker, J. W. 1968, Selection mechanisms associated with intraspecific
shell variation in Liltorina picia (Prosohranchia: Mesogastropoda).
Evolution 22:459-480
Jiiiiival oj Shellfish Research. Vol. 22. No. 1. 141-140. 2()(U.
ADJUSTMENTS OF LIMNOPERNA FORTUNEI (BIVALVIA: MYTILIDAE) AFTER TEN YEARS
OF INVASION IN THE AMERICAS
G. DARRIGRAN,' C. DAMBORENEA.' P. PENCHASZADEH," AND C. TARABORELLi'
^Division Zoologia Inveitehnulos. FCN v Miiseo, UNLP. Paseo del Bosque s/ir ( IWO) La Plata.
CONICET Argentina: -Dep. C. Bioldgicas. FCEyN. USA. Cnulad Univcrsitaria. Pah II. Niii'ic:.. Piso 4°.
Buenos Aires. MACN-CONICET. Argentina
ABSTRACT Limnoperna fortunei (Dunker, 1857) or golden mussel invaded South America through the Ri'o de la Plata estuary in
1991. Ten years later, the golden mussel lives in the main rivers of the Plata Basin. The gonadal cycle and the population settlement
in a temperate climate are discussed in this article. This basic knowledge is needed to assist industries that may suffer the effects of
macrofouling and also increment the ability to predict potential invasions of other countries. The study of population density and
reproductive cycle was performed in Ri'o de la Plata estuary. Argentina. The temporal variation of population density from data of
settlement and age structure collected between 1991 and 2001 is presented. The reproductive cycle between August 1998 and March
2000 was analyzed. Through the analysis of oocyte percentages four gonad spawning events were observed. The spawning events
appear regulated by temperature changes. After the initial increase in population density following the invasion, there was a decrease.
The population appeared stabilized at one third of the initial peak.
KEY WORDS: in\asion. Limnoperna foriimei. freshwater, bivalve, reproductive cycle. Neotropical Region
INTRODICTION
Limnoperna fortunei (Dunker 1857). or golden mussel, is a
freshwater invasive bivalve, from the southeast of Asia. It invaded
South America in 1991. through the Rio de la Plata estuary. This
represents the first record of L. fortunei for the American conti-
nent. Ten years later, the golden mussel lives in the main rivers of
one of the most itnportant Basins of the Neotropical Region
(Bonetto 1994). the Plata Basin (the Rio de la Plata, and the Uru-
guay. Parana, and Paraguay rivers). Since 1999. this species in-
vaded the Guaiba Basin in the south of Brazil (Mansur et al. 1999).
The golden mussel spreads 240 km/year, upstream along the Plata
Basin. (Darrigran & Ezcurra de Drago 2000).
The golden mussel attaches to every available hard substrate.
This lifestyle (epifaunal) is atypical in local freshwater bivalves.
The attachment capability and the great adaptability and reproduc-
tive capacity of these mussels make this species very effective
invaders (Darrigran 2000). The mussels impact on the natural en-
vironment (displacement of native species — Darrigran et al.
1998b, Darrigran et al. 2000 — or change of native fish diet —
Penchaszadeh et al. 2000) as well as on human activities (macro
fouling in fresh water (Darrigran 2000. Darrigran & Ezcurra de
Drago 2000).
Detailed infomiation about the life cycle of this harnitiil inva-
sive species provides a basis for the development and application
40
30
20
10
liC
Au Oc IDe Fe Ap Ju Au Oc I3e Fe
1998 1999 2000
Figure 1. Monthly variation of mean air tenipiTature (line) and water
temperature (bars) during sampling period in Bagliardi Htach. Rio de
la Plata. HVithout data.
of control strategies. The impact caused by this species in human
activities (plugging of water intake for industrial cooling, power
generation, and potable water) resembles what happened in the
north hemisphere with the zebra mussel Dreissena polynwrplia
(Pallas. 1771 ). The study of reproductive cycle, age structure and
temporal density variation, is essential to generate sustainable
techniques for golden tnussel prevention and control.
Details of the reproductive cycle, and the population settlement
in temperate climate are discussed in this article. This type of
knowledge is not only essential to assist biologists and ecologists
in the industries which tnay suffer frotn this new economic-
environtnental problem in the Neotropical Region, but it is also
necessary for predicting potential invasions of other countries in
the north hemisphere such as USA (Ricciardi 1998) and southern
Europe.
TABLE 1.
Date and number of specimens histologically processed per sample.
Date
Size
range (mm)
Males
Females
2.V08/98
27
0.6-2.5
17
10
25/09/98
30
0.6-2.6
23
17
30/10/98
29
0.4-2.5
18
11
27/1 1/98
17
0.5-2.6
14
13
23/02/99
14
0.5-2.9
13
11
19/04/99
20
0.8-2.2
7
13
15/05/99
24
0.7-2.2
14
10
30/06/99
29
0.7-1.9
13
16
26/07/99
25
0.7-2.1
10
15
27/08/99
28
0.5-1.8
19
9
21/10/99
32
0.6-2.1
22
10
27/11/99
34
0.5-1.7
23
11
16/12/99
31
0.5-1.7
14
17
26/01/00
27
0.6-2.1
16
11
22/02/00
35
0.7-2.1
25
10
12/03/00
29
0.6-2.2
18
11
Total
431
266
195
141
142
Darrigran et al.
DENSITY (ind/m-
200,000-
150,000-
100,000
50,000H
a
r^^ rh
r+i
I I I I I I I I I I I
Oct Oct March Oct March Oct March Oct Oct Oct Oct Oct
1991 1992 1993 1993 1994 1994 1995 1996 1997 1998 1999 2000
Oct
2001
Figure 2. Temporal variation of mean density (bars) and standard deviation (lines) oX Limnuperna fortumi in Ba^liardi Beach, Rio de la Plata.
• 4-5 ind/m". *\Vithout data.
MATERIAL AND METHODS
To study the golden mussel population density and reproduc-
tive cycle, samples were collected along the rocky banks of Ba-
gliardi Beach (34°55'S: 57°49'W). Rio de la Plata estuary. Ar-
locality has a temperate regimen ranging from approximately 1 I °C
to 3I°C (Fig. I). The physicochemical features of the Rio de la
Plata may be found in Darrigran (1999). The density data were
obtained partly from Darrigran. et al. (1998b) and through sam-
pling carried out ad-hoc (October 1998 and October 2001) in
gentina. South America. — where the mussel was found for the first Balgliardi Beach. Samples of mussels were collected for density
time in 1991 (Pastorino et al. 1993). The water temperature in this analysis from the fringes with macrobenthos from a rectangular
20
15
10
5
October 1992
n = 481
III...
March 1995
n= 1059
■.■iiMMlii
>N
15
u
c
OJ
in
J
cr
Ol
5
.1
October 1 993
n = 677
I..
I.ll
October 1998
n = 289
I
20
15
10
5
October 1 994
n = 631
llllll...
October 2001
n = 289
Illllllllll.llil..
length (mm)
Figure 3. Size frequency (%) ot Limnoperna fortunei in Bagliardi Beacli. Rio de la Plata.
L. FORTUNEi Ten Years After the Invasion
143
area, variable in size, according to Darrigran et al. ( 1998b). For the
age structure analysis, the niaxinium shell length was measured
and the length frequency distribution was made at 1 mm class
intervals (see Fig. 3 later).
The dates of sampling for reproductive cycle analysis, per-
formed at low tides, may be observed in Table 1. The maximum
shell length of the 431 collected individuals was taken. The ma-
terial was fixed in Bouin solution and the histologic processing
was performed according to Darrigran et al. (1999).
Approximately 25 oocytes with conspicuous nucleolus, both
free in the follicular lumen and attached to the follicle wall, for
each gonad were measured. The percentage of males with sper-
matozoids and the percentage of follicular occupation on the
mantle were calculated for each sample. The latter was calculated
using magnification (x200) in three different sections of the
mantle, (upper, middle, and lower) through the visual estimation of
field. The lysis periods were detemiined by microscopical analysis.
RESULTS
The temporal variation of population density found on the
rockv litoral zone of Baszliardi Beach between 1991 and 2001 is
given in Figure 2. From 1991 to 1995, the density increase was
remarkable (from four to five individuals/nr to over 100.000
ind/m~). The population density then decreases and stabilizes at
approximately 40,000 ind/nr. In Figure 3 it is shown that since
1 994 the population has had an age structure where most size class
intervals are represented.
The female and male tollicles grow in the mantle and in the
visceral mass. During this study 0.25% hermaphrodite specimen,
with female, male, and mixed follicles were recorded.
The gonad growth is characterized by growing follicles. In this
stage the follicles are small and there exists an abundant connec-
tive tissue between them. A more developed stage shows young
oocytes on the wall, many stalked oocytes (Fig. 4A) and abundant
spermatogoniums in the males (Fig. 4D). In a later stage the fol-
licles are bigger and the follicular lumen contains abundant oo-
cytes half-grown and also almost fully grown oocytes (60-80 p.).
When fully mature, the female and male follicles reach the maxi-
mum size. Male follicles are packed with spermatozoa (Fig. 4E)
and females" follicles with fully-grown (80-100 p.) oocytes.
When the gonads are spent and partially spent, the follicles
contain large spaces. Partially spent gonads retain genital products.
Figure 4. Female and male tollicles in different development stages. (A) Female follicle partiallj gro«n with voiing oocytes on the wall and many
stalked oocytes, scale bar = 1(10 fi. (B) .Spawned female follicles with abundant yellow bodies (arrows!, scale bar = 5(1 p. (C) Female follicles partly
spawned, scale bar = 100 p. (Dl Developing male follicles, scale bar = 50 fi. (E) Fully developed male follicles, scale bar = 100 \i. (Fl Male follicles
partly spawned, scale bar = 50 jj.
144
Darrigran et al.
IL
23/08/98
)(= 57 06
DS = 57 06
n = 224: N =
10
Ij.
23/02/99
X = 75 22
DS = 21 14
= 262, N=11
il
26/07/99
X = 45 19
DS = 45 19
n = 267, N
lll^
16/12/99
X = 43 02
DS = 17 93
n = 353, N=17
llx.
25/09/98
x = 56 91
DS = 20 27
n = 400. N= 17
III] L_ Liilijili.
19/04/99
X = 60.57
n = 225, N=13
27/08/99
X = 38 99
DS = 14 95
n= 177, N=9
I
L.
26/01/00
X = 47 65
OS = 16 £
n = 292; N=ll
31/10/98
X= 51.67
DS = 5167
n = 300. N = 1 1
u
15/05/99
X=46 52
DS = 46 52
n= 170. N=10
21/10/99
X = 51 92
DS= 17 91
n = 248; N
22/02/00
X = 49 94
n = 225. N=10
j1 iIL
LJ.
30/06/99
X = 44 87
DS = 21 72
n = 352. N=16
Jul
110 1» 130 10 K Ti
oocyte diameter (um;
12/03/00
X = 55 03
DS = 22 83
n = 320. N=11
K X) 40 so K 70 so so ICO 110 120 130
Figure 5. Frequency I in percentage) of oocyte sizes (ji) in different samplings, x, mean oocyte size; DS, standard deviation; n, number of oocytes;
N, number of females.
In males spennatozolds and spermatocites are observed (Fig. 4F|.
Partly developed oocytes, oogonies, and young oocytes are re-
tained on the female follicle walls (Fig. 4C). Oocitary lysis phe-
nomena (Fig. 4B). with yellow bodies are evident for a short time
after spawning is completed.
The body length at which the follicle, either female or male,
development is completed, varies seasonally. The smallest shell
length at which follicles differentiate is 5.5 mm. for both males and
females (Fig. 5). During this study (August 1998 to March 2000).
oocyte growth was always recorded. Froin May 1999 until August
1999, the oocytes smaller than 20 p. were 309f of the total oocytes
examined.
The change in frequency of oocytes <20 p, and >60 p, indicates
two reproductive peaks each year. The first peak occurs at the end
A
B
oocyte (%) S.
JJrlrl J rlbll.rhftrllJ
De I Fe
1999
* *
¥
males (%)
De I Fe
2000
n = 266
Figure 6. Temporal variation. (.\) Percentage of oocytes bigger than 60 fi (full bars) and smaller 20 yt (empty bars). The arrows indicate moments
of gamete liberation. (B) Percentage of males with sperm. ■ Without data, n, number of male individuals.
L. FOHTUNEi Ten Years After the Invasion
145
90
80
70
60
50
40
30
occupation of the mantle (%)
n = 431
1
M
Au Oc De Fe Ap Ju Au Oc De
1998 1999 2000
Finiirt' 7. Temporal variation of mantlt occupation. Female follicles
(full barsi and males (empt> bars), n, total number of considered males
and females. 'Without data.
of winter or beginning of spring (August to September of 1998.
October to November of 1999) and the second peak is recorded
during the summer (February of 1999. March of 2000). During
these periods in the female follicles the oocytes bigger than 60 |ji.
dominate, while smaller oocytes are scarce (<20%). During the
period of study gonad recuperation were observed (October 1998
and May to June of 1999). Through the analysis of oocyte per-
centages present in the gonad, four spawning events were observed
(Fig. 6A):
( i ) From September to October 1998.
(2) February 1999 to May 1999. It is the most important for its
duration and magnitude.
(3) in July to August 1999, the least important.
(4) between October and December 1999,
Figure 6B shows the percentage of males with sperm through-
out the period considered. The pattern agrees in general with that
observed for females.
The spawning pattern mentioned is similar to the follicular
occupation of the mantle (Fig. 7). The percentage of occupation
decreases during the spawning periods and stays low during the
recuperation period (June. July, and August 1999).
Lysis phenomena were observed in several samples (Fig. 8).
They are more important during May to August 1999, and coincide
with recuperating follicles or in partial evacuation.
DISCUSSION
The bivalve sexual processes are generally related to ambient
temperature (Lubet 1983). The results presented here for a popu-
lation of L. fiirliiiu'i. as well as those observed in the first study
(Darrigran et al. 1999), those performed for a Hong Kong popu-
lation (Morton, 1982), and the analysis of larvae density in the Ri'o
de la Plata (Cataldo & Boltovskoy 2000) show the strong relation-
ship between ambient water temperature and the reproductive
cycle. The spawning events are regulated by changes in tempera-
ture, and increases and decreases of temperature rule the gameto-
genesis in this species.
During the initial study (Darrigran et al. 1999). we found that
oocytes were always present in the mussels even during the resting
period. Periods of scarce proliferation were recorded from Decem-
ber 1993 to May 1994. This study was performed a short time after
the first record of L. forlimei in the Americas (Pastorino et al.
1993). The analysis of reproductive biology at that time differen-
tiated numerous spawning events (five were recorded), many of
them of low magnitude. Between September 1992 and January
1993 (the first period), two spawnings of reduced intensity were
recorded and between February 1993 and November 1994 (the
second period) three spawnings were recorded (two of these of
higher magnitude). During the first period, the oocytes bigger than
60 |xni and those smaller than 20 |jLm are always present and their
proportion is similar (about 309H. the spawnings are low in mag-
nitude but the proportion of oocytes bigger than 60 \vm is always
larger than 20%. During the second period, the spawnings are
more intense and result in a diminution of the bigger oocytes
proportion (by 10%). In contrast to the first period, the oocytes
bigger than 60 (xm reach more than 60% (Darrigran et al. 1999).
The population analyzed here shows a predictable reproductive
pattern. Only two major spawnings are observed throughout the
year, one when summer temperatures are recorded and the other
with spring temperatures. A small winter spawning is also ob-
served. This pattern, after 10 years of settlement in America, is
similar to that described by Morton (1982) for the population of
Hong Kong where the spawnings take place between May to June
and November to December. The pattern shown during the first
study [only after a year of settlement in the location considered
(Darrigran et al. 1999)] could be due to the recent invasion.
Morton (1982) describes short spawnings for a month in spring
and a month in autumn. In this study in South America, mainly in
autumn, the evacuation continues from April 1999 to May 1999.
The presence of larvae in the Ri'o de la Plata, between August and
April (Cataldo & Boltovskoy 2000). also indicates that the spawn-
ing periods are longer than those described by Morton (1982).
Similar to what was found in the first study of the golden
mussel reproductive cycle (Darrigran et al. 1998a) 0.25% of the
population was hermaphrodite.
According to the variation of population density, this species, at
the beginning of the invasion in temperate climate, presents a
noticeable increase of density. Then, it decreases its density to a
third part and stabilizes. At the same time, it presents an age
structure with most class intervals represented. These facts would
indicate a stable settlement of the population to the en\ ironment.
%
Ll
* *
Au Oc De
Ap Ju Au Oc De
I Fe
Figure 8. Percentage of females with follicles where lysis phenomenon occur. *Withoul data. n. number of considered females.
146
Darrigran et al.
The initial increase recorded in a temperate climate could also
be observed in a subtropical climate. Despite the preliminary stud-
ies of this species invasion in the south of Brazil, subtropical
climate (Mansur et al. 1999). the golden mussel presents an in-
crease in its population density similar to that observed in this
study. Two years after its first record (Mansur et al. 2001 a. Mansur
et al. 2001b). the maximum density is 62.100 ind/m".
The golden mussel, like other invasive species, is opportunistic.
This fact makes it difficult to relate the reproductive pattern with
environmental variables and to determine the different facts that
might be modified in the reproductive cycle. L. forliinei. for its
great adaptability and reproductive capacity, increases its distribu-
tion permanently by occupying environments of particular fea-
tures.
ACKNOWLEDGMENTS
The authors thank Renata Claudi for her comments on a draft
version of the manuscript. This work was partly financed by grants
BID 1201 OC/AR PICT 01-03453 from the Agenda Nacional de
Promocion Cientffica y Tecnologica, Argentina; Facultad Ciencias
Naturales y Museo, Universidad Nacional de La Plata (UNLP) and
Fundacion Antorchas.
LITERATURE CITED
Bonetto. A. A. 1994. Austral rivers of South America. In: R. Margalef.
editor. Limnology Now: a paradigm of planetary problems. Amslerdan:
Elsevier Science, pp. 425-472.
Cataldo, D. H. & D. Boltovskoy. 2000. Yearly reproductive activity of
Umnoperna fortuiu'i (Bivalvia) as inferred from the occurrence of its
larvae in the plankton of the lower Parana river and the Rio de la PUiia
estuary (Argentina). Aqiialic Ecology 34:307-317.
Darrigran, G. 1999. Longitudinal distribution of molluscan communities in
the Rio de la Plata estuary as indicators of environmental conditions.
Malacological Review- (Suppl.) 8:1-12.
Darrigran, G. 2000. Inva,sive Freshwater Bivalves of the Neotropical Re-
gion. Dreissena 11:7-13.
Darrigran, G. & I. Ezcurra de Drago. 2000, Invasion of Limnopenm for-
timei (Dunker, IS.'i7) (Bivalvia: Mytilidae) in America. Nautilus 2:69-
74.
Danigran, G., M. C. Damborenea & P. Penchaszadeh. 1998a. A case of
hermaphroditism in the freshwater invading bivalve Limnoperna for-
timei (Dunker. 1857) (Mytilidae) from Rio de la Plata. .Argentina.
Iberus 16:99-104.
Darrigran, G., S. M. Martin, B. Gullo & L. Armendariz. 1998b. Macroin-
vertebrates associated to Limnoperna fortunei (Dunker. 1857) (Bi-
valvia, Mytilidae). Ri'o de La Plata, Argentina. Hxdrobiologia 367:
223-230.
Darrigran, G., P. Penchaszadeh & M. C. Damborenea. 1999. The life cycle
oi Umnoperna forlunei (Dunker, 1857) (Bivalvia:Mytilidae) from a
neotropical temperate locality. J. Shellfish Res. 18:361-365.
Darrigran, G.. P. Penchaszadeh & M. C. Damborenea. 2000. An invasion
tale: Limnoperna fortunei (Dunker, 1857) (Mytilidae) in the neotropics.
International Aquatic Nuisance Species and Zebra-Mussels Confer-
ence. 10:219-224.
Lubet. P. 1983. Experimental studies on the action of temperature on (he
reproductive activity of the mussel (Myliliis eilnlis L. Mollusca, Lamel-
lihranchia). J. Mollusc. Studies (Suppl.) 12A: I(J0-I05.
Mansur, M. C. D., L. M. Zanirichinitti & C. Pinheiro dos Santos. 1999.
Limnoperna fortunei (Dunker, 1857), Molusco Bivalve invasor, na ba-
cia do Guafba, Rio Grande do Sul, Brasil. Biociencius 7:147-150.
Mansur, M. C. C. Santos, G. Darrigran, G. Heydrich, C. Quevedo & L.
Iranco. 2001a. Preferencias e densidades do mexilhao dourado Lim-
noperna fortunei (Dunker, 1857), em diferentes subtratos da bacia do
Guai'ba, Rio Grande do Sul. Brasil. RESUMOS V Congresso de Eco-
logia do Brasil: 246.
Mansur, M. C. D., C. Pinheiro dos Santos, G. Darrigran. 1. Heydrich. C.
Barbosa Quevedo & L. Bernades Iranco. 2001b. Densidade e cresci-
mento populacional do mexilhao dourado Limnoperna fortunei
(Dunker, 1857), na bacia do Guaiba e novos registros na Laguna dos
Patos, Rio Grande do Sul. Brasil. RESUMOS XVII Encontro Brasileiro
de Malacologia: 61.
Morton, B. 1982. The reproductive cycle in Limnoperna fortunei (Dunker,
1857) (Bivalvia: Mytilidae) fouling Hong Kong's raw water supply
system. Oceanologia el Limnologia Sinica 13:312-324.
Pastorino. G.. G. Darrigran, S. M. Martin & L. Lunaschi. 1993. Limno-
perna fortunei (Dunker, 1857) (Mytilidae), nuevo bivalvo invasor en
aguas del Ri'o de la Plata. Neotropica 39:34.
Penchaszadeh, P., G. Darrigran. C. Angulo, A. Averbuj, N. Brignoccoli, M.
Brogger, A. Dogliotti & N. Pirez. 2000. Predation on the invasive
freshwater mussel Limnoperna fortunei (Dunker, 1857) (Mytilidae) by
the fish Leporinus obtusidns Valenciennes, 1846 (Anostomidae) in the
Rio de la Plata, Argentina. J. Shellfish Res. 19:229-231.
Ricciardi. A. 1998. Global range expansion of the Asian mus.sel Limno-
perna fortunei (Mytilidae): Another fouling threat to freshwater sys-
tems. Bu'fouling 13:97-106.
J<:nniiil ofSlwllfish Research. Vol. 22. Nci. 1. 147-lh4. 2(KI,V
QUANTITATIVE EVALUATION OF THE DIET AND FEEDING BEHAVIOR OF THE
CARNIVOROUS GASTROPOD, CONCHOLEPAS CONCHOLEPAS (BRUGUIERE, 1789)
(MURICIDAE) IN SUBTIDAL HABITATS IN THE SOUTHEASTERN PACIFIC
UPWELLING SYSTEM
WOLFGANG B. STOTZ,* SERGIO A. GONZALEZ, LUIS CAILLAUX, AND JAIME ABURTO
Universidad Catolica del Norte, Facultud de Ciencias del Mar, DepariaiueiUo de Biolo^iia Marina,
Casilla 117. Coquiinhd. Chile
ABSTRACT Landings of Concholepas coinholepas. a carnivorous gastropod and valuable fishery resource, appear disproporlion-
ately high compared with herbivores or suspension feeding mussels. The species has been previously described as feeding on a great
variety of prey, the most important being barnacles, mussels, and tunicates. To quantitatively evaluate published information on the
diet of C. comluilepas. an analysis of the stomach contents of 92.'i individuals was performed, representing a wide size-range, broad
geographical distribution (29''30'S to 32°08'S), and different community types (variety of potential prey choices). The diet was based
principally on suspension feeders, such as barnacles iBaUiiui.s Uicvis and juveniles of .Aii.\tii)nu'.i;iihcilanus fi.siikicii.s) 175%) and the
ascidian Pyiira chilensis (16%). An additional sampling, in which abundance of prey in the habitat and microhabitats occupied by the
gastropod was determined, showed that the gastropod positively selects these prey species, the ascidian being the most preferred. The
rest of the diet was made up of Calyptraea Irochifoniii.s and mytilid bivalves. According to the literature, intertidal individuals of this
species only feed at night. To confirm this behavior for subtidal populations, two 24-h samplings (analyzing digestive tract contents)
were performed at a single location. No distinct circadian cycle of feeding for subtidal populations was found, most animals feeding
most of the time. This, together with the characteristics of diet, made mainly by suspension feeders, which transfer energy from primary
productivity in the water column which varies along the coast, to benthic carnivores, help to explain the high productivity of the
gastropod and its variability along the coast of Chile.
A'£l' WORDS: feeding beha\ ior, circadian rhythm, selectivity, carnivorous gastropod, Chile, subtidal, upwelling system
INTRODUCTION
The muricid gastropod Concholepas concholepas (Bruguiere
1789) ("Chilean abalone") is distributed from 12°S to 55°S along
the Peruvian and Chilean coasts and is an important predator oc-
cupying rocky shores (Castilla 1981. Castilla &. Paine 1987). It is
a valuable product in artesanal fisheries (Castilla & Jerez 1986)
along its entire distribution. In Chile, the highest landings ranged
between 6,369t and 25,()()Ot between 1978 and 1988 whereas the
fishery was unregulated; the maximum value was recorded in 1980
(SERNAP 1999). In region IV (between 29' 30'S and 32°08'S, 320
km coast) in the period between 1985 and 2000 landings fluctuated
between 258 and 2,2 19t for this carnivorous gastropod species. In
the same period and along the same stretch of coast the herbivo-
rous gastropods FissurcUa spp. (eight different species that are
fished) and Tegula aira (Lesson), which share the habitat with C
concholepas. together registered landings between 695 and 1525t.
The aim of this work was to investigate what kind of food sustains
the comparatively important production of this high trophic level
carnivore, the ecological position to which C concholepas is usu-
ally assigned, Stotz (1997) has shown that within management
areas the abundance of C. concholepas is related to the amount of
food, the species overexploiting its food source when not fished,
and then migration to other areas. Thus, the knowledge of diet and
feeding behavior is also of importance in developing a manage-
ment strategy of the species within management areas.
According to published literature, C. concholepas has been
observed feeding on a variety of prey, the most often mentioned
being barnacles, mussels, and tunicates (Viviani 1975, Castilla &
Cancino 1979, Castilla & Guisado 1979, Castilla et al. 1979.
DuBois et al. 1980, Castilla 1981, Guisado & Castilla I98.\ Sotn-
*Corresponding author. E-mail: wstotz@socompa.cecun.ucii,cl
mer 1991. Sommer & .Stotz 1991 ). Bui quantitative feeding infor-
mation is scarce; the number of published observations for indi-
viduals feeding in their natural subtidal habitats was less than 96,
observed at two localities (Castilla et al. 1979, Guisado & Castilla
1983, DuBois et al. 1980, Sommer 1991). These did not represent
the entire spectrum of subtidal communities in which the gastro-
pod lives. There are also qualitative observations (Viviani 1975,
Castilla et al. 1979, Castilla 1981 ) that increase the data regarding
the prey diversity of C. concholepas but do not allow evaluation of
the relative dietary importance of the different prey species of this
gastropod.
The published quantitative information on food types con-
sumed by C concholepas was obtained by feeding behavior ob-
servations (Castilla et al. 1979). DuBois et al. (1980) stated "an
individual is feeding when one observes an unusual extension of
the foot over a potential prey species or when the individual shows
movements to remove a prey." This includes lifting individuals to
check for empty shells, direct observations of ingestion of prey,
empty spaces on the substrate in front of the mouth or of the "shell
teeth", which the species has on the anterior border of the shell,
proboscis introduced into the prey, or prey held by the propodiuni
and directed to the mouth (Castilla ct al. 1979). This method gath-
ers information on the specific prey being consumed at the mo-
ment of observation. Thus, those prey species that are more diffi-
cult to consume and for which the process of ingestion lasts longer
will have a higher probability of being observed. Also, in order not
to disturb animals and thus record observations of natural feeding
behavior, observations have been limited to individuals found on
open surfaces. Feeding by individuals found in crevices or on the
undersides of boulders, including most juveniles and medium-
sized individuals of C. concholepas (Castilla & Cancino 1979,
Guisado & Ca.stilla 1983, Sommer 1991, Stotz & Lancellotti 1993)
cannot be easily observed. Thus, observations of C. concholepas
on open surfaces will focus only its feeding on prey abundant on
147
148
Stotz et al.
such places and food composition described using this method
may not necessarily reflect the relative importance of the different
prey species in the diet of C. concholepas.
In contrast, the analysis of digestive tract contents provides a
quantitative measure of food consumption over a certain time in-
terval, representing the range of prey species and their relative
importance in the diet of the predator. Only in case digestion rates
for different prey species differ greatly, some bias may occur. This
is the first work in which feeding of C. concholepas has been
studied through the analysis of the contents of the digestive tract.
According to published information. C. concholepas feeds only
at night (Castilla & Guisado 1979. Castilla & Cancino 1979.
Castilla et al. 1979. Guisado & Castilla 1983). However, this has
been conckided mainly from laboratory experiments mostly using
individuals collected in the intertidal zone. Only DuBois et al.
(1980) have made observations in the subtidal. recording the feed-
ing activity of 96 individuals of this species.
Intertidal gastropods search out and consume food mainly at
night to avoid desiccation (Underwood 1979, Branch 1981, Hawk-
ins & Hartnoll 198.^. Lowell 1984). Subtidal populations of C.
concholepas, not exposed to this stress, may feed mainlv at night
for other reasons: ( I ) to avoid visual predators active during da> -
time (Castilla & Cancino 1979) and/or (2) to capture prey that
respond to visual stimuli and may be able to escape predation by
C. concholepas during the day.
Visual predators, which are known to include C. concholepas
in their diet, such as the sea-otter Lmrafelina (Molina) (Castilla &
Bahamondes 1979), the sea lion Otariaflavescens (Shaw) (Aguayo
& Maturana 1973) and the fishes Pimelometopon macidatus
(Perez) and Sicyases sanguineus Miiller & Troschel (Viviani
1975). do not figure prominently in the monality of this gastropod
species. L. felina has been suggested to be highly specialized on
fish and Crustacea as prey (Sielfeld 1990); O. flavescens does not
appear to prey on gastropods firmly attached to substrates, as is the
case for C. concholepas (George-Nascimento et al. 1983); and the
fish species prey mainly on juveniles of C. concholepas which,
according to our observations, are hidden in crevices in the sub-
tidal. Prey selection is an unlikely factor promoting night time
feeding, as the main prey of C. concholepas are sessile species,
such as the barnacles Auslromegabalanus psittacus (Molina).
Balanus laevis Bruguiere, and Jehlius cirratus (Darwin); the tuni-
cate Pyura chilensis (Molina); the mitilid Peruniytilus puipuratus
(Lamarck); and the hemisessile gastropod Calyptraea trochifonins
(Boml (Castilla & Guisado 1979. Castilla et al. 1979. DuBois et al.
1980, Guisado & Castilla 1983, Castilla & Durdn 1985, Moreno et
al. 1986, Sommer 1991. Sommer & Stotz 1991). Therefore, there
appears to be no strong argument that subtidal populations of C.
concholepas feed exclusively at night. Nevertheless, this needs to
be investigated, which is one aim of this work.
This work reports food composition and feeding behavior (cir-
cadian feeding rhythm and food selection) for C. concholepas
based on the analysis of the food content in the digestive tract. A
greater variety of habitats than in previous studies were sampled,
including open surfaces, crevices, the undersides of boulders, hold-
fasts of the subtidal kelp Lessonia trabeculata (Villouta & San-
telices), and under the canopy of this algae along an extensive
stretch of coast from 29°30'S to 32°08'S (ca. 320 km). On one site
the sampling and analysis of the digestive tract contents of a large
number of individuals collected over a 24-h cycle was conducted.
For some of the individuals sampled along the coast and in dif-
ferent communities, the abundance of potential prey in the envi-
ronment is quantified to establish to what degree the food in the
gut represents the availability of prey. This allows us to study
whether there is some kind of preference for some prey species.
MATERIALS AND METHODS
Study Sites
Individuals of C concholepas were collected at several sites
along the ca. 320 km of coast of the Coquimbo Region, between
Pichidangui (32°08'S) and Punta Choros (29°30'S) (Fig. 1 ). The
sites were chosen considering accessibility and being representa-
tive of different coast and community types. A qualitative de-
scription of subtidal communities of each sampling site is provided
in Table I . Quantitative data of communities in which the gastro-
pod was sampled are provided in Tables 3 and 4. For the 24-h
sampling the site at Punta Lagunillas (30°05'S; 7|-26"W). located
ca. 15 km south of Coquimbo, was chosen. It is a rocky point
forming the northern border of Bahia Guanaqueros (Fig. I ). .-M-
though it is an exposed coast, it has an irregular configuration that
creates sheltered ponds that allow for safe diving through the surf
and at night. The substrate is formed by different sized boulders
that are covered by a dense kelp forest formed by small and bushy
(many blades, short stipes) individuals of Lessonia trabeculata. It
corresponds to community type I (Table I). Quantitative data for
the community at this site are given in Table 3. Larger individuals
of C. concholepas are found mostly within the kelp forest, w hereas
smaller individuals are mainly hidden in crevices or on the under-
sides of boulders.
PACIFIC
29
" S
OCEAN A
ELTEMBLADOR
Punla Choros
\
TOTORAULLO ^^S^
NORTE
\
P LINT A
LAGUNILLAS
J
30^
s
PUERTO X.'S
ALDEA ^
Coquimb
o
—
DEVACA --^^^BahlaTongoy
SAN y]
LORENZO
1
3,o
s
PUERTO 1
OSCURO >\
HUENTELAUQEN \
ISLA ^ I
HUEVO ^^^ /
LASTINICUNAS \ T
Vilos
32
s
TOTORAULLO ^V
SUR ""^f
7 Bah
aPichidan^
ui
Figure I. Location of the study sites along the coast in the region of
Coquimbo (region IV).
Diet and Feeding Behavior of C. concholepas
149
TABLK 1.
Siibtidal fommunitifs »herf ('. coiuhiilepus was collected a general description ol' each community is given.
Type
Communities
Localities
I Kelp hed nf Lcssonia traheciiUita El Tenihlador, Punla Lagimillas. Puma Lengua de Vaca, San Loren/o. Caleta Las Conchas. Totoralillo
Sur (isle and bay)
II Barren gniund Toralillo Norte (rock). Puerto Oscuro
III Barnacles and seaweeds Toioralillo Norte (isle)
IV Colonies of Pxiini chilen.\ii Puerto Aldca
General description of the subtidal community types
Type tcimmunity
General Description
Kelp bed of Lessonia traheculata
Barren ground
Barnacles and seaweeds
IV Colonies of Pxiira chileusis
Community characterized by the kelp Lessonia traheculata. Under the canopy, dense patches of
barnacles (e.g.. Balaims laevis) and to a lesser extent the ascidian Pyiiru chilensis are found. In
crevices and on the underside of boulders are observed aggregations of the gastropod Calyptraea
trocliifdiinis. sponges and small patches of barnacles.
Community characterized by an high cover of calcareaus crustose algae and high densities of the black
urchin Tetrapygiis niger. In crevices and on the underside of boulders are observed aggregations of
Pyiira chilensis, of C. trochyfonnis and patches of barnacles.
Community dominated by extensive patches of barnacles, specially by Austmmegabalaniis pssittacus,
which can be covered by a dense mat of the red algae Gelidiiim chllense. Also aggregations of the
ascidian Pyura chilensis may be present in crevices.
Community formed mainly by aggregations of the ascidian Pyiira chilensis. which covers most of the
surface. The ascidians could be partly covered by the algae Glgartina chamissoi. On the underside of
boulders aggregations of Calvpiniea trochiformis can be observed.
Sampling of C. concholepas Along the Coast to Describe Diet
Individtials were collected by Hookah di\'ing from the intertidal
down to a maximum depth of 25 m. At each site two divers
collected all C cimcholepa.t that they were able to find within
approximately 1 h of diving, which allows the inspection of an area
of about 200-500 m". Individuals of all sizes were collected and
the searches included the undersides of boulders. Table 2 summa-
rizes the number and size range of individuals collected at each site
of the samplings undertaken between January 1994 and December
1995.
Experiments for the Identification of Prey and Food Retention Time
in the Gut
The identification of each prey item was aided by a simple
experiment in which known prey were offered to individual C.
concholepas. Three groups of 10 adult individuals (70-110 mm
peristomal length) were collected at Punta Lagunillas and main-
tained in tanks with running seawater. Each group was offered one
of the most important prey items described in the literature (Som-
mer & Stotz 1991): the barnacles Aiistromegahulanus psittacits
and Bdhiiuis laevis. the gastropod Calyptraea trochiformis. and the
TABLE 2.
C. concholepas: Number of individuals collected in the field, number of entire digestive tracts analyzed in the laboratory, size range, number
of individuals with food in their tracts, and number of individuals with recognizable prey in their digestive tracts are given.
Sample
Sample
Individuals
Recognizable
Size
Size
Size
\>itb Food
Prey
Field
Labor.
Range
Localities
(N")
(N")
(mm)
No.
%
No.
Vf
Playa EI Teniblador
76
74
24-122
54
73.0
47
87
Totoralillo Norte (rock)
21
21
37-93
13
61.9
10
76.9
Totoralillo None (isle)
9
8
15-122
4
50.0
4
1 00.0
Punta Lagunillas (August)
166
166
2I-I2I
122
73.5
110
90.2
Punta Lagunillas (January)
282
260
7-125
235
90.4
225
95.7
Puerto Aldea
13
13
102-129
13
1 00.0
11
84.6
Punta Lengua de Vaca
52
45
51-116
33
73.3
22
66.7
San Lorenzo
188
158
16-126
105
66.5
93
88.6
Puerto Oscuro
7
7
59-100
5
71.4
5
lOO.O
Isla Huevos
54
54
24-131
52
96.3
52
1 00.0
Totoralillo Sur (isle)
95
72
69^7
65
90.3
62
90.4
Totoralillo Sur (bay)
51
47
26-125
40
85.1
39
97.5
Total
ini4
925
7-131
741
80.1
680
91.8
150
Stotz et al.
ascidian Pyiira chilensis. Individuals were maintained continu-
ously with food, sampling after the initial 48 h, and then daily, two
individuals. Sample animals were dissected and their stomach and
gut contents exaiuined. The physical characteristics of each prey
item after ingestion by C. conclwlepas were recorded and then
used as a reference in the analysis of stomach and gut contents
from individuals sampled in nature.
To measure the time the food is held in the digestive tract, a
field experiment was performed at Punta Lagunillas on October
25-26, 1995. Therefore, all the individuals collected during a
30-min period at 1800 h and again at 0600 h of the next day were
maintained in a mesh bag in the water in the study site, without
food. Every 2 h, six individuals of this mesh bag were sampled and
sacrificed, fixinc the visceral mass in 10% saline formalin. In the
laboratory, the proportion of individuals with food in the stomach
or gut in each sample was determined.
Samplings to Compare Diet with the Food A railable in
the Environment
At seven sites (El Temblador. Punta Lagunillas. Punta Lengua
de Vaca. Huentelauquen, Isla Huevos. Tinicunas. Totoralillo sur)
(Fig. 1), between January 1996 and March 1997. samplings were
repeated, but this time recording also abundance of prey in the
environment. For each C. conclwlepas individual collected, the
density and percent cover of species present on the spot, was
recorded. A 0.25-m" quadrant with 100 regularly distributed points
STOMACH
Pyura
chilensis
INDETERMINATE
89,6 CIRRIPEDIA
Q 3 Calyptrsea
''■^ ' trvctiiformis
INTESTINE
83 9 CIRRIPEDIA
INDETERMINATE 73
8,3 ^"^
chilensis
TOTAL
Pyura
chilensis 15,88
INDETERMINATE
Calyptraea trochifonvis o,65
74 67 CIRRIPEDIA
0,05 MYTILIOAE
m
CIRRIPEDIA
MYTILIDAE
^v8i Pyura chilensis
INDETERMINATE
Calyptraea trochiformis
Figure 2. Dietary composition of Conclwlepas conclwlepas.
Diet and Feeding Behavior of C. concholepas
151
was used. The quadrant was loL-ated with its center on the spot
were the C. conclioU-pas indi\ idual was captured.
For four of these seven sites (Isla Huevo. Punta Lengua de
Vaca. Punta Lagunillas. and El Temblador) (Fig. 1). a general
quantitative description of communities present on the site was
done. A 50-ni long and 2-m wide transect was placed parallel to
the coastline. For less frequent species their abundance in the
entire transect area (100 m") was counted, whereas for smaller,
more frequent species five 0.25-ni" quadrants, distributed regularly
along the transect, were used. To quantify the laminarian algae
Lcssduia inilicculata. the transect was divided into 2.^1 areas of 2 x
2 m. estimating percent cover within each of these areas. Within
these same areas the percent cover of each substrate type was
estimated in those cases in which the bottom was a mixture of sand
and rocks. This estimate was used to correct abundance and per-
cent cover estimates of species, in order that they refer only to
rocky bottom.
24-b Sampling al Punta iMgiiiiillas
The 24-h sampling was accomplished twice: on October 24 and
2.S. 1994 and August .S and 6. 1996. Dives took place at 1700.
STOMACH
>
u
c
(D
<1>
00
i
fe
^
EES
i
m
60 1
x^-'
:|;
40 i
5 - CN CM
i ^ ^ 1
1 Z 2 c
o
(A
Surl
Sur2
Total
o
o
o
O
illo
illo
i- - - m
n to
uj 2 2 -^
as
5
2 2
a
o
£
^—
INTESTINE
>
u
c
0)
3
a
u.
100 .
80 1
60
40
m
iiizza^
m
TOTAL
100
m
CIRRIPEDIA
MYTILIDAE
Pyura chilensis
Calyptraea trochlformis
INDETERMINATE
Figure 3. General dietary composition of Cnnchnlepas concholepas from each sampling site.
152
Stotz et al.
2100, 0100, 0500, 0900, and 1300 h. On each dive, two divers
sampled the subtidal at depths between 4 and 10 m, collecting each
C. concholepas they were able to tlnd within a half-hour dive.
Searches were concentrated beneath the canopies of L. trabeciilata
and included the undersides of boulders. At night searches were
conducted using underwater flashlights. Diving was conducted us-
ing a compressor on the beach that provided air to the divers
through a hose (Hooka diving). In the 1996 sampling, the indi-
viduals collected by each diver were considered as replicate
samples.
Processing of Samples
All samples of C. concholepas were processed immediately
after collection. Peristomal length of individuals were measured
with calipers and grouped into se\ en si/e classes from <30 mm to
>130 mm (see Figs. 4 and 5). Each specimen was taken out of the
shell and the visceral mass dissected and fixed in 10% saline
formalin. Visceral masses of all individuals from each size class
were stored together in a single container and transported to the
laboratory.
In the laboratory, the digestive tract of each individual was
dissected; the contents emptied separately for stomach and gut in
two Petri dishes, diluted with tap water, and spread on the bottom
of the dish. The relative abundance of each prey item was recorded
for each individual using a dissecting microscope. Therefore the
dish was put over a point matrix, recording the food item over each
point, and calculating its proportion to all the points covered by the
sample. Also the presence of prey species, which were present, but
not registered over any point, were annotated.
For C. concholepas from the 24-h sampling a measure of full-
ness was recorded. Fullness and digestion level was determined
Using the following scale:
Fullness:
Full: contents occupy ca. 100% of the volume of the stomach or gut.
Medium: contents occupy around 50% of the \ olume of the stom-
ach or gut.
Presence: contents occupy around 10% of the volume of the stom-
ach or gut.
Empty: no contents registered.
Digestion level:
Some digestion: entire structures are observed, such as pieces of
cirri, 2ills, muscles, etc.
STOMACH
>
o
z
LU
a
UJ
100
80 -
60
40
20
^^^
^
m
INTESTINE
>
o
z
tu
r>
o
lU
a:
u.
100
80
60
40
20
i^*c/j
m
CIRRIPEDIA
MYTILIDAE
^vvi Pyura chilensis
INDETERMINATE
Calyptraea trochiforwis
Figure 4. Dietary composition of Concholepas concholepas in different size classes (length of peristomal opening).
Diet and Feeding Behavior of C. concholepas
153
STOMACH
INTESTINE
>
o
z
lU
o
ai
u.
>-
o
z
lU
a
Ul
a:
u.
100)
sa
so
4a
20^
100
80i
6o^
4a
2a
I
+
^S3^^^SSS
80i
I
2(y
10Q
80,
601
4a
2a
EL
TEMBLADOR
^?
LAGUNILLAS
O
z
LU
a
UJ
(t
LL
o
z
UJ
o
UJ
100|
801
■^1
61
4a
2a
10Q
!
so]
40*
2a
^^
100|
8a
60
4a
20
''
Wi
100
80
60
40
20
LAS
CONCHAS
^^
TOTORALILLO
SUR
T- CO
SIZE CLASSES (Cm)
CIRRIPEDIA
Pyura chilensis
INDETERMINATE
Figure 5. Dietary composition of Concholepas concholepas in different size classes (lengtli of peristomal opening) Iroin lour sampling localities.
mate significance levels, using the following relations:
Species A Other spp. Total
Medium: structures could still be identified, but already with some one degree of freedom (Sokal & Rolf 1969, Pearre 1982) to esti-
digestion.
Total digestion: soft parts are completely digested, only pieces of
shells or hard skeletons can be identified.
Prey Selection Analysis
To determine the degree of selection of prey by C. cdiuhdlepas
an index proposed by Pearre (1982) was used. This allows the
estimation of the selection index C. but also using a x' tc'st with
In the diet
A,
«./
\
+ e.
= c
In the
environment
.4,,
fi„
.4,
+ s„
= D
■\,
+
.4,,
= ,4
«,,
+
«,,
= H
■\
+ ■'>„
+ Bj + B„
= N
154
Stotz et al.
10 12
18:30
Starvation period (hours)
Figure 6. Percentage of Individuals v\ith contents In the stomach and
Intestine during the starvation periods beginning In the morning (A)
and In the afternoon (B).
Where:
Aj = Proportion of species A in the stomach
/4,, = Proportion of species A in the environment
Bj = Proportion of the rest of species in the stomach
fi„= Proportion of the rest of species in the environment
The index "C" is obtained from the followina relation:
Where:
N
X'
N
(A,rB,,-A„- BJ--
A- B- CD)
The index C varies between -1 and +1. A significant positive
value indicates that the prey species was preferred and rejected
with a significant negatise value. Values around zero means that
the prey species is consumed in the same proportion it appears in
the environment.
For estimation of the index only those species found
in the diet of C. concluilepus where considered. For the cal-
culations, the density of invertebrates present in the quadrant
was transformed into percent cover to have all the values on
the same scale. For this, the area occupied by an average indi-
vidual was estimated, calculating its proportion within the
2.?00 cm" of the sampled area. This proportion was multiplied by
the number of sampled individuals, thus obtaining their percent
cover.
Once this proportions where estimated, a correction for poten-
Degree of
Fullness
n Empty
^ Presence
B Medium
■ Full
Digestion
Level
D Empty
^ Total
B Medium
■ Some
2 4 6 8 1012 1416 2 4 6 8 10 12 14 16
STARVATION PERIOD
Figure 7. Prey digestion level (first column: A, C) and degree of fullness (second column: B, D) of stomach and intestine during the starvation
periods beginning In the morning (first line: A, B) and in the afternoon (second line: C, D).
Diet and Feeding Behavior of C. concholepas
155
MORNING
SAMPLE
Cirripedia
(principally Balanus laevis)
Indeterminate
Totally digested
AFTERNOON
SAMPLE
Mollusca
Pyura chilensis
Calyptraea
trochiformis
Cirripedia
(principally Balanus laevis)
Mollusca
Indeterminate
Totally digested
Pyura chilensis
Figure 8. Prey composition of Concholepas concholepas in the starva-
tion experiment at Punta Lagunillas.
tial prey species was done. Therefore, the percent cover values for
algae and empty space was eliminated, calculating a new propor-
tion considering that potential prey species cover 100% of the
substrate.
For these analyses, only the content of the stomach was used
because this represents the most recently ingested food, most prob-
ably from the sampled spot. Also, empty or destroyed stomachs
were not considered.
RESULTS
Diet
Of the 1.014 individuals of C. concholepas collected at nine
sites (Table 2) visceral masses of 925 individuals were examined.
Of these, only 741 individuals (SO.I^r), covering a size range from
7-131 mm peristomal length, had food in their digestive tracts
(Table 2).
Only 8.2% of the digestive tracts had contents that could not be
identified because the process of digestion was already too ad-
vanced (Table 2). About 98% of the individuals examined fed on
one prey type. Only 18 individuals (2%) had more than one prey
item in the digestive tract.
The most important prey items were barnacles, representing
89.6% of the stomach contents, and 83.9% of intestinal contents
(Fig. 21. The second most important prey item, the ascidian P.
chilensis. represented 5.47r and 8.3% of the stomach and gut con-
tents, respectively. The remainder of the prey was Calyplraea
trochiformis, mitilids. and unidentified materials. Differences be-
tween stomach and intestine were produced by more advanced
digestion in the latter. That favored recognition of the ascidian in
the intestine because its remains were recognized mainly by color,
which was not affected by digestion. C. trochiformis was not found
in the intestine. But these different digestion rates of the various
prey did not change the general dominance of barnacles in the diet.
The dietary importance of barnacles was most pronounced at
Caleta Las Conchas, where they represented the only prey. In
contrast, at Puerto Aldea, where C. concholepas was introduced by
fishermen, barnacles were entirely replaced by P. chilensis (Fig.
3). With only two exceptions (Puerto Aldea and Lengua de Vaca).
in all sites the barnacles were the predominant prey (Fig. 3). even
though the basic community structure varied (Table I ).
Prey composition did not differ among the different size groups
within the pooled sample, where barnacles were always the dom-
inant prey item (Fig. 4). The same analysis made at selected sam-
pling sites, also showed in general, with only two exceptions (El
Temblador 9-1 I cm; Totoralillo Sur 5-7 cm) (Fig. 5) that the
barnacle was the predominant prey. Although in all cases the
smallest and the biggest indi\iduals only fed on barnacles, interme-
diate-sized individuals showed a slightly more varied diet (Fig. 5)
Identification of Prey and Food Retention Time
The feeding experiments with known prey items allowed gen-
eral descriptions of the prey after ingestion by the gastropod. Skel-
etal plates, cirri, and eggs were observed in the stomach and gut
when C. concholepas fed on barnacles. When the ascidian Pxuru
chilensis was the prey, an orange or red mass sometimes contain-
ing syphons was observed. In the case of Calyptraea trochiformis.
while-colored muscular tissue and egg capsules could be recog-
nized. Comparison of these characteristics with those observed in
the digestive contents of individuals collected in the field allowed
the identification of most prey items.
Regarding food retention, the percentage of individuals with
content in the digestive tract is highest (83.3%) in the morning
(0630 h) and in the evening (1830 h) when just sampled. As the
starvation period increases, the proportion of individuals with con-
tent in the digestive tract fluctuates, decreasing after 12 h of star-
vation (Fig. 6). The decrease is more evident and regular for the
stomach, not so much for the intestine. The stomach appears com-
pletely empty after 16 h of starvation. Accordingly, the percentage
of full stomachs or those with the content showing some digestion
decreases as the starvation period increases (Fig. 7). Nevertheless,
the tendency is not that clear, close to the end of the experiment
appearing again individuals with full stomach or intestine, and
showing just some digestion (Fig. 7). This suggests that some
contamination of the experiment may have occurred. The problem
probably stems on the fact that the shells of the individuals put
together in the mesh bag were not cleaned. Thus the barnacles,
which normally are attached to the shell, might have been con-
sumed by some of the experimental indi\ iduals. Considering this
possible contamination, the experiment suggests that the retention
time in the stomach is around 6 h. whereas in the intestine the food
seems to be retained up to 16 h. The prey species the experimental
individuals had ingested were the same as described above for the
individuals sampled along the coast (Fig. 8).
Prey Selection by C. concholepas
The most important prey species are not the most abundant
species in the habitat (Table 3). Barnacles appear in small patches.
156
Stotz et al.
TABLE 3.
Abundance of macroalgae and invertebrates (percent co>er and density, mean and standard deviation) in the rocky subtidal in which
Concholepas concholepas was collected at lour sites.
Temblador
Lagunillas
Lengua de Vaca
Isla Huevo
Percent cover (%)
Algae
Rhodophyta
Mesophytlwn sp.
Corallina officinalis
Gelidium chilense
Calcareus algal crusts
Phaeophyta
Glossophora kiinthii
Lessonia irtibccnUilii
Porifera annellida
Phnii^inalopoma sp.
Roiiunuiiellii piisudata
Spionidae
Crustacea
BaUmus laevis
Httluiut\ flosciitii\
Austromei>ubalaiuis psitlacus
Bryozoa
Bugula sp.
Briozoa indeterminated
Hemichordata
Pyura chilensis
Free space
Density (ind.m"')
Mollusca
Nassarius gayii
Crassilabnim crassilnhrwn
Tegula sp.
Mitrellii iinifasiiura
Crepiilula sp.
Tegula Iridentala
Calyptnwa Iroclnformis
Density (ind. 1 00m"-)
Cnidaria
Anemonia alicemartiinie
Phymactis clematis
Phymanlhea pluvia
Mollusca
Concholepas concholepas
Fissurella cosrara
Fissurella ciimingii
Crustacea
Paroxanthiis barbiger
Taliepus denumis
Homalaspis plana
Rhynchncinetes typiis
Echinoderniata
Aelionidiwn chilensis (Holoduiroidea)
Meyenaster gelalinosus
Stichaster strialus
Heliaster helianihus
Tetrapygiis niger
19.8 ±25.07
4.6 ± 10.29
5.0+ 11.18
2.4 ±2.51
68.0 ±28.72
29.8+ 17.04
0.2 ±0.45
5.6 ± 12.52
0.2 ± 1.79
0.8 ± 1.79
9.8 ± 10.43
21.2 ± l,V81
15.2 ±25.52
1.6±2.19
0.8 ± 1.79
20
361
16
2
45.6 ± 27.57
0.4 ± 0.55
4.0 ± 6.42
10.0± 11.16
70.0 ± 15.55
2.2 ±4.92
3.6 ± 5.68
1.6 ±2.30
10.6 ± 10.67
0.4 ± 0.8
1.2 ± 1.64
20.8 ± 23.86
0.6 ± 1 .34
1.6 ±3.58
1.0 ± 1.00
0.4 ± 0.89
95
2
55
1
3
17
1
5
57.0 ± 19.46
4.6 ±4.67
6.2 ± 8.90
1.2± 1.79
60.8 ±21.78
0.2 ± 0.45
6.6 ± 7.47
3.0 ±6.7 1
6.0 ± 7.04
15.2 ± 15.32
20.8 ± 29.04
4.8 ± 10.73
0.8 ± 1.79
1.6 ±2.19
0.8 ± 1.79
1
25
23
6
1
1
5
1
584
49.8 ± 23.22
0.4 + 0.55
10.0 ± 20.20
!3.8± 13.18
49.2 ±28. 12
1.0 ± 1.73
17.4± 13.92
0.4 ± 0.89
7.2 ± 16.10
124.8 ±265.85
164.0 ±257.74
125.6 ±265.38
26.4 ± 36.40
26
2
1
mostly associated to the area immediately around the holdfast of
Lessonia Irabeciilata. where fronds do not wipe the rock. Pyura
chilensis is mostly restricted to crevices. Percent cover of both
prey species together lluctuates between 10 and 20% cover. But C.
concholepas within the habitat selects microhahitats in which his
prey species, mainly barnacles, are more abundant. In those mi-
crohahitats percent cover of barnacles may increase up to almost
80% (Table 4). The polychaeta Phragmatopoma sp.. which con-
Diet and Feeding Behavior of C. concholepas
157
TABLE 4.
Proportion (%) of potential prtj in the different microhuhitats In Hhlcli Concholepas loiichohpas was captured on seven study sites.
lA-nj^ua de Totoralillo Las
El Temblador \ aca Huentelauquen Isla Huevo Sur Tinicunas Laguniiias
Main prey species
Pxura chilensis
14.34
6.16
1.01
Cirripedia
24.75
8.52
21.14
68.04
74.63
Phragmatopoma sp.
Other potential prey
Porifera annellida
45.66
8.76
43.74
6.88
8.13
27.68
2.52
4.88
14.63
Polvchaeta indeterniined
10.16
56.91
Romunclu'lla pusUilaui
3.68
6.47
Mollusca
Calyplniea InKJiifonnis
0.29
1.75
Fix.surella spp
0.81
0.19
0.49
Timiciii elegans
0.31
Brachiodomes granulaia
0.22
2.26
Crassilahnim crassiUihnim
0.22
0.10
0.56
Tegiila spp
Naisariiis gayii
0.51
0.41
0.41
5.37
Bryozo
Brvozoa
L59
8.42
9.76
Cnidaria
Hydriv.oa indeterminate
Echinodermata
Tetnipygus niger
Hemichordata
4.41
3.25
1.96
9.82
38.34
1.75
6.88
5.58
0.67
1.38
5.90
0.46
0.34
3.66
0.67
0.92
0.40
0.55
34.09
6.64
structs tubes of sediment attached to the rock surface, in some
areas gets very important, covering together with the barnacles
most part of the space in some sites (Table 4).
The digestive tracts of C. concholepas from the sampled sites
contained mainly barnacles and P. chilensis. Although barnacles
are the most abundant prey species in the environment. P. chilensis
was only rarely found, mostly in very low abundance. Only in one
site the ascidian was important in the environment (El Temblador,
Table 4). Barnacles appear in four of the seven sites as being
positively selected (Table 5. Fig. 9). In the remaining three sites
barnacles are consumed proportionally to their abundance in the
environment. P. chilensis was present only in four of the seven
sites (Table 4), being always positively selected (Table 5 1. On one
site (Las Tinicunas) P. chilensis did not appear registered in the
environment (its proportion less than 1%), but was in the digestive
tract of the gastropod. When the data from all the sites are grouped
and analyzed together, it is shown that only P. chilensis is posi-
tively selected, the rest of preys being consumed proportionally to
their abundance in the environment (Fig. 9H).
Circadian Feeding Rhythm
A total of 275 individuals were collected, representing a size
range between 29 to 120 mm of peristomal length in the first 24-h
sampling period. For the second period 88 and 84 individuals were
sampled by each diver, representing a size range between 2(1 to 1 19
mm of peristomal length (Table 6). Numbers collected during
individual sampling hours varied from 13 individuals at 2100 h to
71 individuals at 1300 h in the first sampling period and seven
individuals at 2100 h to 28 individuals at 1700 h for the second
sampling period (Table 6). As some of the samples were destroyed
during the transport to the laboratory, the analysis is based on 254
individuals for the first sampling period, and on 66 and 81 indi-
viduals respectively for the two replicate samples of the second
sampling period.
TABLE 5.
Selection index C and x" for main prey species of Concholepas concholepas on seven sites.
Lengua
de
El Temblador
Vaca
Huentelauquen
Isla Huevo
Totoralillo Sur
Las Ti
nicunas
Lagu
C
nillas
C
X'
C
X^
C
X-
C
X^
C
x'
C
x'
X^
Pxura chilensis
0.18*
6.64
0.47*
43.89
0.26*
13.39
0.34*
20.47
0.28*
16.22
Cirripedia
0.19*
6.95
0.05
0.45
0.80*
128.54
a 10
2.15
-0.(39
1.53
-0.43*
32.35
0.23*
10.37
Phragmatopoma sp.
-0.39*
30.42
-0.53*
.55.98
-0.63*
79.55
-0.32*
20.33
-0.16*
5.00
-0.09
1.76
Other species
0.03
0.21
0.01
0.0 1
-0.34
23.30
-0.13
3.23
0.14*
4.19
0.25*
11.11
-0.40*
3 1 .59
Values with * show significant positive or negative selection.
158
Stotz et al.
TABLE 6.
Date and time of 24-h samplings, number of individuals collected in
the field, number of entire digestive tracts analyzed in the
laboratory, and inditiduals with food in their digestive tracts
(number and percentage).
Sample
Individuals
Indi
viduals
Size
Field
Analyzed
in the
with Food
Date
Time
(No.l
Lab (No.)
(No.)
(%l
24 OCT 1994
17:00
44
42
31
73.8
21:00
13
13
11
84.6
01:00
46
40
31
77.5
05:00
44
44
37
84.1
09:00
57
52
41
78.8
13:00
71
63
45
71.4
Total
275
254
196
77.2
5 AUG 1995
17:00
28
19
15
78.9
(Replicate 1)
2 1 :00
7
5
5
100
01:00
S
7
6
85.7
05:00
15
7
6
85.7
09:00
15
15
14
93.3
13:00
15
13
11
84.6
Total
8S
66
57
86.4
5 AUG 1995
17:00
21
21
21
100
(Replicale 2)
21:00
7
7
7
100
01:00
12
12
9
75.0
05:00
16
13
12
92.3
09:00
15
15
13
86.7
13:00
13
13
10
76.9
Total
84
81
72
88.9
Considering all the individuals analyzed for the entire 24-h
sampling, the individuals with food in their digestive tract (stom-
ach and/or gut), represent 77.2% for the first sampling period and
86.4% and 88.9%. respectively, for the two replicate samples for
the second sampling period (Table 6. Fig. 10). During the different
sampling hours the proportion of individuals with food in their
digestive tract for all sampling hours represented at least 71.4%.
Although no clear pattern appears, in all sampling periods, the
highest values were always registered at the late afternoon and
early morning, thus suggesting that feeding intensity increases
during the afternoon and in the second half of the night, or at dawn
and dusk. Nevertheless, no statistical difference was detected be-
tween day (individuals sampled at 0900. 1300. and 1700 h) and
night (individuals sampled at 2100. 0100. 0500 h). as well as
between the different replicate samples (sampling in October 94.
and each diver in August 96) (3 x 3 G test, x" = 1 1.714: df = 7:
P > 0. 1 ). Neither statistical difference was detected between dif-
ferent hours (Contingency Table 6*3*2; x" = 32.7304; df = 27;
P > 0.05).
At the different sampling hours different degrees of fullness
were observed (Fig. 10). Although no clear pattern can be identi-
fied, the stomach shows a slight tendency of greater fullness in the
afternoon or late afternoon hours, decreasing during night, with the
same tendency repeating during the early morning hours. For the
intestine it is observed, that as the stomach empties, the intestine
increases in fullness (Fig. 10). Thus, again the data suggest that
intake of new prey tends to increases at dawn and dusk.
In all sampled individuals during both 24-h samplings, bar-
nacles appear as the main prey species, with proportions ranging
from 41.9% to 75.2%. with a mean value of 57.9% (Fig. 11). The
second most important prey was Pyiini chilensis, which comprised
17.7% to 40.3% of the digestive tract contents. The remaining
individuals had other preys of minor importance, such as Ciilrp-
tniea trochifonnis (Fig. 11). The food composition also did not
vary greatly with sampling time, and barnacles were always the
dominant prey item.
DISCUSSION
Concholepas concholepas fed almost exclusively on barnacles
and the ascidian Pyiini cliileiisis. The similarity in diet composi-
tion among individuals from different localities and among differ-
ent size classes, suggests that this is a general characteristic for
subtidal populations of this species.
These data support corresponding literature data (Castilla et al.
1979. DuBois et al. 1980. Sommer 1991 ), but show quantitatively,
that barnacles were usually the most consumed prey in the differ-
ent community or microhabitat types where C. concholepas was
found. The smaller individuals of C. concholepas live on the un-
dersides of boulders or in crevices (Stotz 1997. Guisado & Castilla
1983. Sommer 1991). where potential prey is probably different
from that present on the rock surfaces where larger individuals
live. Nevertheless, all size groups had consumed very similar food
types. This suggests a strong feeding preference for barnacles,
which nevertheless seems not always supported by the analysis
with the selection index. With the pooled data. P. chilensis appears
as the most preferred prey species. However, the preference is
better shown by the fact that the gastropod is always found in
microhabitats in which the barnacles predominate. And within
such microhabitat the index is not any more able to show a pref-
erence. Considering all the prey species described, the preference
extends in general to suspension feeders. A similar behavior has
been described for Acanthina lugubris angelica, the diet of which
was restricted exclusively to sessile suspension feeders (Vermeij
et al. 1994). The diet based on suspension feeders seems to be a
general pattern for benthic predators, such as diverse gastropods
and seastars (Table 7).
The most common barnacles in subtidal communities are Baki-
nus laevis and Aiistromegabalantis psituiciis. Individuals of the
latter species are mostly small individuals with 0.5-1 cm basal
diameter, while the species is able to growth to sizes of ca. 5-cm
basal diameter. But in the I'egion. barnacles of such big size are
seldom observed.
Feeding based on barnacles that are small, sessile, and form a
uniform cover on the substrate makes C. concholepas conceptually
resemble a grazer. The feeding of C. concholepas is similar to the
"grazing" of hydroid colonies by nudibranchs. or even to grazing
gastropods, for example, the keyhole limpets Fisurella spp.
(Moreno & Jaramillo 1983, Moreno et al. 1984. Godoy & Moreno
1989). This observation applies to many gastropods and starfishes
(Table 7). It is a well-described characteristic for intertidal whelks
(Dayton 1971. Paine 1966. Menge & Sutherland 1987). habitat in
which the sessile suspension feeders are the main space occupiers,
but less known for species living in the subtidal. where a wider
variety of potential prey species may be expected. In fact. C.
concholepas makes use of a wider variety of prey in such habitats,
including mobile predators as crabs and even fishes (personal ob-
servations), but quantitatively only the suspension feeders are im-
portant.
The feeding behavior of C. concholepas, not showing a clear
circadian rhythm, differs from what has been published previously
Diet and Feeding Behavior of C. concholepas
159
.2
Q
1U0
N=18
A
>.
o
50
**
*
r-1 n
S "-
<u
a
P. Cirr
orrzr
50
100
10O
50
TJ
Phrag. Other
N=7
Cirr Phrag. Other
50
10O
100
50
50
100
100
50
N=7
JZL
n
n
p. Cirr Phrag. Other
TJ
N=3
n
p.
Cirr
Phrag. Other
u
50
00
—
100
50
50
100
100
50
50
100
100
50
100"
Figure 9. Frequency of prey in the diet and In the niierohnhilut in which Concholepas concholepas was captured in xarious localities: (.\)
Teniblador, (B) I.engua dc Vaca, (C) Huentelauquen, (D) Isia Huevo, (E) Las Tinicunas, (F) Lagunillas, (G) Tutoralilio Sur, (H) Pooled Sample.
N=22
B
p. Cirr Phrag. Other
N=14
*
XIL
D
p. Cirr Phrag. Other
TJ
N=26,,
F
50
*
n
n
P Cirr Phrag. Other
50
100
100
50
P. Cirr Phrag. Other
T— rn \ZT-
N=
=97
H
*
.. n
for this species by Castilla and Guisado (1979), Castilla and Can-
cino (1979), Castilla et al. (1979). Guisado and Castilla (198.^).
and DuBois et al. (1980). Differences in the methodological ap-
proach may explain this. Previous studies have been based in the
intertidal zone, or in the laboratory, but using individuals collected
from the intertidal. Environmental characteristics of the intertidal
zone, principally dessication stress, often cause circadian rhythms,
with activity periods at night and resting periods during the day
(Underwood 1979, Branch 1981. Hawkins & Hartnoll 1983. Low-
ell 1984). Pino et al. (1993) compared the activity periods of the
intertidal gastropod FissureUa crassii Lamarck and the subtidal
species F. kitimurpnuta Sowerby and observed that the intertidal
species had a distinct day-night activity cycle whereas the subtidal
species did not. The novelty for C. concholepas is that in this case,
the difference is between different populations (intertidal and sub-
tidal) of the same species. However. DuBois et al. (1980) has also
reported a day-night activity cycle for a subtidal population of C.
concholepas.
DuBois et al. (1980). as all the published work done before on
the feeding of C. concholepas. based his conclusion on the direct
160
Stotz et al.
B
Individuals with Food(%)
Sample Size (N°)
100
Stomach
Intestine
Tide
]Day>Night
17 21 01 05 09 13
17 21 01 05 09 13
17 21 01 05 09 13 Time
FULL
i
MEDIUM
SOME
PREY
EMPTY
Figure 10. Circadian variations: Percentage of individuals witii contents in tlieir digestive tracts, corresponding sample sizes, and percentage of
individuals with different degrees of fullness of the stomach or intestine for each of the three replicate samplings (A) October 1994; (B) August
1996. replicate I: (C) August 1996, replicate 2.
observation of capture and ingestion, using criteria defined by
Castilla ( 1979). If the prey is small and the predator is positioned
directly over it, no sign of feeding will be seen. This may often be
the case when C. concholepas feeds on barnacles, its main prey
species. Study results may also be influenced by different condi-
tions of observation (day and natural light, night and artificial
light). For example, it is possible that at night the field observa-
tions are made mainly on more active individuals located on the
surface of rocks, whereas during the day individuals found in
crevices and between the algae might be included, and for these
individuals it would be more difficult to establish if they were
active or resting. Moreover, depending on the light conditions,
animals could react differently to the presence of the diver. Finally,
DuBois et al. (1980) also mention that some of the animals
included in their observations from Caleta Hornos were intro-
duced to the study site prior to the experiment. The behavior of
these individuals might differ from that of resident (subtidal)
animals.
In the approach used by DuBois et al. (1980). if capture and
ingestion of prey occurs rapidly and is of short duration, it is less
likely that observations will be recorded. The study of digestive
tract contents also includes the process of digestion, thus covering
a much longer time period, being less likely that a individual which
has been feeding is missed. But on the other hand, the long reten-
tion time shown by C. concholepas, may obscure the e.xistence of
a circadian feeding rhythm. Nevertheless, if no ingestion of food
took place over the day (or over the night), at the end of the day
(or night) most of the stomachs should be empty, as seen in the
experiment in which the individuals where starved. And this is not
A B
Total
Calyptraea
trochiformis 0.3%
Gastropoda 1.2%
ndetarminate
8.9%
Pyura chilensis
31.8%
Figure 11. Prey composition of Concholepas concholepas sampled
over 24 h at Punta Lagunillas. Composition of each replicate sampling
(A) October 1994: (B) August 1996, replicate 1: (C) .August 1996,
Replicate 2; and total diet are shown.
Diet and Feeding Behavior of C. concholepas
161
TABLE 7.
Summary of prey species for several gastropods and starflsh.
Predator
Main Prtv
Site
Author
Gastropods
Thais c'likiifiuuila
Thais tlaviiit'ta
Thais hi serai is
Acanthina hrcxideuuita
Thais emarginaia
Thais canaliculata
Thais lamellosa
Niicella lapilUis
Nmtlla lapiUns
Nmclla emargimna
Chicoseus capucinus
Siramonila haemastoma
Strtinionila liaemasloma
Concholepas concholepas
Starfishes
Leplasterias polaris
Astehas vulgaris
Asrerias rubens
Asterias vulgaris
Asterias forhesi
Aslcrias vulgaris
Crossasrer pappusus
Lepraslerias polaris
Coscinas calainaria
Cosmasieria luriila
Pisasier ochraceus
Asterias vulgaris
Stichaster australis
Leptasterias hexaclics
Pisasier ochraceus
Balanus gluiulula
Telractita squamosa
Balanus amphitrite
Siphonaria japonica
Barnacles
Bivalves
Barnacles
Semihalanus balanouUs. Balanus
creniUns.
Mytilus edulis and cither bisalves
Mytilus edulis. Seniihaltunis
halanoides
Bivalves
Barnacles
Bahuuts iunplurrite
Modiolus sp
Crassosrrea virginica
Brachiodomes pharatniis
Barnacles
Barnacles, tunicates
Mytilus edulis
Mytilus edulis
Mytilus edulis
Mytilus edulis
Balanus crenatus
Balanus balanoide
Mytilus edulis
Chlamys islandica
Mytilus edulis
Chlamys islandica
Ascidea sp.
Didemnum albidum
Mytilus edulis
Mya spp
Hiatella artica
Balanus sp
Halocvnthia pxrifinmis
Ascidea sp
Chlamvs asperrinius
Ascidaceas
Podoclavella cvlindrica
Botrylloides leachii
Stolonica australis
Aulacomya ater
Balanus spp.
Tunicata:
Styella melincae
Colonial tunicate
Mussels
Mussels
Mussels
Balanus cariosiis
Balanus glandula
Mytilus edulis
Cluhamahis dalii
Washington. USA
Cape d' Aguilar. Hong
Kong
Costa Rica
Washington. USA
New England. USA
Maine, Anglesey
Canada
Singapur
Gulf of Mexico
Israel
Chile
Canada
Canada
German Bight. North Sea
Outer Brewster Island
(Massachusetts)
Gulf of St. Lawrence
Gulf of St. Lawrence
St. Lawrence Estuary
Rapid Bay (Australia)
Puerto Toro (Chile)
Temperate NE Pacific
Temperate NW Atlantic
Temperate SE Pacific
San Juan Island,
Washington
Paine 1%6
Blackmore 2000
Paine 1966
Dayton 1971
Menge & Sutherland 1976.
1987
Hughes 1992. Dietl. 2000
Gosselin & Chia 1996
Koh-Siang Tan 2000
Brown & Stickle 2002
Rilov, Gasith & Benayahu
2002
This study
Gaymer et al. 2001
Gaymer et al. 2001
Saier 2001
Menge 1979
Himmelman 1991
Himmelman 1991
Himmelman & Lavergne
1985
Keough & Butler 1979
Vasquez& Castilla 1984
Menge 1992b
Menge 1974
162
Stotz et al.
TABLE 7.
continued
Predator
Main Prev
Site
Author
Heliasrer helUmllius
Pxcnopinliii hflitinthiiitles
Asterias vulgaris
Leptasterias polaris
Meyenaster gehitinosus
Meyenaster gelatinosus
Semimytitus algosus
Perumytilus purpuratus
Brachiodomes sp
Cbamidae sp
Jehlius cirratus
Chlhamalus scabrosus
Mylilus edulis
Bivalves
Balanus spp
Mytitus edulis
Macoma spp
Mya tiuncara
Mytilus edulis
Mya tiuncara
Mya arenaria
Macoma spp
Brachiodontes graiudara
Semele solida
Balanus sp
Aulacomya ater
Megahahmus sp
Pyura sp
Ancon Bav. Peru
Tokeshi el ul. 1989
Torch Bay. Alaska
Golf of St. Lawrence
Golf of St. Lawrence
El Frances. Chile
Golfo de Penas, Chile
Duggins 1983
Himmelman & Dutil 1991
Himmelman & Dutil 1991
Vasquez 1993
Dayton et al. 1977
the case. Although no statistical differences was detect between
individuals with food at different hours, a slight indication of the
existence of greater ingestion is suggested to happen at dawn and
dusk, at least in two of the three replicates.
Thus, the high percentage of individuals with stomach contents
throughout the day and night, show ing no distinct pattern of varia-
tion which could be associated with the circadian rhythm, suggests
that most animals are feeding at all day and night hours. Thus. C.
concholepas invests most of its time to feeding, as has been de-
scribed by Bayne and Scullard ( 1978) for the snail Thais (Nucella)
lapilhis. They estimated that this species spends between 45 and
63% of its time feeding.
The conclusion that C. concholepas feeds almost over the entire
24-h cycle is important for the validation of our study of the food
composition of this species because sampling time does not appear
to be an impoilant factor. Although our results show some minor
variation in the prey composition with time, this can probably be
attributed more to normal variability of the diet, rather than to
circadian rhythms of feeding.
The high production described for C. concholepas (Stotz &
Perez 1992) can be explained by its feeding on the lowest con-
sumer level, which shortens the energy pathway frotn the primary
producer level (Whittaker 1975). By feeding on barnacles and
ascidians, this benthic gastropod effectively shortens the food
chain. Through the consumption of suspension feeders C. con-
cholepas accesses the much larger energy pool of primary produc-
tion in the water column. For some coastal environments it has
been calculated that 509^ of the net primary production of the
water column is used by benthic animals (Grahame 1987). which
is the process C. concholepas is taking advantage of. By this
feeding habit, C. concholepas is taking advantage of the high
productivity provided by upwelling processes along the coastal
zone of the southeastern Pacific coast of South America (Raymont
1980. Bakum & Nelson 1991, Thomas et al. 1994).
Upwelling processes, being localized in certain coastal areas,
generate a spatial variability of primary production along the Chil-
ean coast (Fonseca & Fari'as 1987, Acufla et al. 1989). The possible
relation of this variability and the different production levels of C.
concholepas along the coast, as shown by variable landings in
different regions along the Chilean coast (Stotz 1997) is a hypoth-
esis of much interest for this valuable fishery resource. Stotz
(1997) showed that average landings for the period 1985-1995
along the entire coast of Chile, expressed as t per km of rocky
coast, shows two patterns: (1 ) a general trend of decreasing land-
ings from the south to the north, and (2) spots with higher landings
than observed in surrounding areas (see Fig. 10 in Stotz, 1997).
The first trend may be related to a similar trend for primary pro-
ductivity described by Thomas et al. ( 1994), who integrated infor-
mation for 8 y ( 1979-1986). These authors describe high primary
productivity year around for the area close to the coast (0 to 100
km from the coast) in front of region X (43°S)(see Fig. I for
location of regions). In front of region VIII (37°S) there are periods
of high primary productivity only during autumn and winter. In
front of region IV (29°S) the period of high primary productivity
is restricted to a short period in winter. Further north primary
productivity is year around low. The second pattern suggests
a close relation to upwelling centers located in the regions VIII
and IV. At a smaller geographic scale, for region IV, Stotz
(1997) also shows a similar pattern, with the highest landings
registered in the areas around the local upwelling center located
in front of Punta Lengua de Vaca (Fig. I). Variability of landings
may be produced by variations in productivity of the gastropod,
which, as shown by Stotz and Perez ( 1992) and Perez and Stotz
(1992) differs between sites along the 320 km of coast of the
Coquimbo region (region IV). Greater production of C. conchole-
pas associated to upwelling would be evidence for the hypothetical
alternative interaction webs in sites with differences in primary
production in the water column, as postulated by Menge (1992a).
Diet and Feeding Behavior of C. concholepas
163
In places with higher primary production, filter feeders get
more important, and consequently small carnivores, the category
to which C. concholepas would correspond, also increase. The
understanding of this variability and its causes are essential for
the management of this important fishery resource. The estima-
tion of catch quotas for different regions should consider this
variability. Knowledge of the quantitative relation between
primary production, production of suspension feeders and con-
sequent production of this gastropod, would improve predictive
capabilities, thus greatly aiding proper management of this
resource.
ACKNOWLEDGMENTS
We are grateful to the Servicio Nacional de Pesca for facilities
given special permission for the sampling as well as to the differ-
ent fishemien's organizations that allowed diving within their
management areas at Caleta Totoralillo Sur. Caleta Las Conchas,
Caleta San Pedro in Los Vilos. Caleta Huentelauquen, Caleta Puer-
to O.scuro. and Caleta Puerto Aldea. Thanks are given also to
Raymond Bienert and Louis DiSalvo, who improved the English
of the manuscript. This study was funded by Project FONDECYT
N° 1941146/1994.
LITERATURE CITED
Acuna. E.. J. Moraga & E. Uribe. 1989. La zona dc Coqulniho: iin sistema
nerftico de alta productivldad. Revista Paciticu Sur (Niiinero especial):
14.'i-153.
Aguayo. A. & R. Maturana. 1973. Presencia del lobii niarino comiin
{Otiiria flmesceiis) en el literal chileno. Biologia Pesquera. Chile 6:
45-75.
Bakun. A. & C. S. Nelson. 1991. The seasonal cycle of wind-stress curl in
sobtropical eastern boundary current regions. / Pliys. Oceanogr. 21:1815-
18,34.
Bayne. B, L. & C. Scullar. 1978. Rates of feeding by Thais (Niici'lhil
lupiUus (L.). / Exper. Mcirme Biol. Ecol. 32:1 13-132.
Blackmore. G. 2000. Field evidence of metal transfer from invertebrate
prey to an intertidal predator, Thais clavigera (Gastropoda: Muricidae).
Esniurine Coastal Shelf Sci. 51:127-139.
Branch. G. M. 1981. The biology of Umpets: physical factors, energy flow
and ecological interactions. Oceanogr. Marine Biol. Annu. Rev. 19:
235-280.
Brown, K. M. & W. B. Stickle. 2002. Physical constraints on the foraging
ecology of a predatory snail. Marine Freshwater Behav. Physiol. 35:
157-166.
Castilla. J. C. 1981 . Perspectivas de investigacion en estructura y dinamica
de comunidades intermareales rocosas de Chile central. II. Depreda-
dores de alto nivel trdfico. Medio Ambienle 5:190-215.
Castilla, J. C. & J. Cancino. 1979. Principales depredadores de Conchole-
pas concholepas (Mollusca: Gastropoda: Muricidae) y observaciones
preliniinares sobre mecani.smos conductuales de escape y defensa. Bio-
logia Pesquera. Chile 12:115-123.
Castilla. J. C. & 1. Bahamondes. 1979. Observaciones conductuales y eco-
logicas en Intra f el ina (Camivora: Mustelidae) en las zonas Central y
Centro-Norte de Chile. Arch. Biol. Med. Exper. 12:119-132.
Castilla, J. C. & L. R. Duran. 1985. Human exclusion from the rocky
intertidal zone of central Chile: the effects on Concholepas conchole-
pas (Gastropoda). Oikos 45:391-399.
Castilla, J. C. & Ch. Guisado. 1979. Conducta de alimentacion nocturna de
Coticholepas concholepas {MoWwica: Gastropoda: Miricidae). Biologia
Pesquera. Chile 12:125-130.
Castilla. J. C. & Jerez. G. 1986. Artisanal fishery and development of a
date base for managing the loco Concholepas concholepas resource in
Chile. Canadian Special Publication of Fishery and Aquatic Science
92:133-139.
Castilla, J. C. & Paine, R.. T. 1987. Predation and community organization
on Eastern Pacific, temperate zone, rocky intertidal coast. Revista Cliil-
ena de Historia Natural 60:131-151.
Castilla. J. C. Ch. Guisado & J. Cancino. 1979. Aspectos ecologicos y
conductuales relacionados con la alimentacion de Concholepas con-
cholepas (Mollusca: Gastropoda: Muricidae). Biologia Pesquera. Chile
12:99-114.
Dayton, P. K. 1971. Competition, disturbance, and community organiza-
tion: the provision and subsequent utilization of space in a rocky in-
tertidal community. Ecol. Monogr. 41:351-389.
Dayton, P. K., J. Rosenthal, L. Mahen & T. Antezana. 1977. Populatitm
structure and foraging biology of the predaceous Chilean asteroid Mey-
enaster geluimoMis and the escape hiology of its prey. Mar. Biol.
39:361-370.
Dietl. G. 2000. Successful and unsuccessful predation of the gastropod
Niicella lapillus (Muricidae) on the mussel Mytilus edidis from Maine.
VV//;??;- 43:319-329.
Dubois. R., J. C. Castilla & R. Cacciolatto. 1980. Sublittoral observations
of behaviour in the Chilean 'Loco' Concholepas concholepas (Mol-
lusca: Gastropoda: Muricidae). V'('//.?£'/- 23:83-92.
Duggins. D. 1983. Starfish predation and the creation of mosaic patterns in
a kelp-dominated community. Ecology 64:1610-1619.
Fonseca, T. R. & M. Farias. 1987. Estudio del proceso de surgencia en la
costa chilena utilizando percepcion remota. Invest. Pesquera 34:33^6.
Gaymer, C. P., J. H. Himmelman & L. E. Johnson. 2001. Distribution and
feeding ecology of the seastars Leptasterias polaris and Aslerias vul-
garis in the northern Gulf of St. Lawrence. Canada. J. Marine Biol.
Assoc. t/A' 81:827-843.
George-Nacimento. M.. R. Bustamante & C. Gyarzun. 1985. Feeding ecol-
ogy of the South American sea lion: Otaria flavescens: food contents
and food selectivity. Marine Ecol. Prog. Series 21:135-143.
Godoy. C. & C. Moreno. 1989. Indirect effects of human exclusion from
the rocky intertidal in southern Chile: a case a cross-linkage between
herbivores. Oikos 54:101-106.
Gossehn. L. A. & F.-S. Chia. 1996. Feeding habits of newly hatched
juveniles of an intertidal predatory gastropod, Nncella emarginata (De-
shayes). / Exp. Marine Biol. Ecol. 176:1-13.
Grahame. J. 1987. Plankton and fisheries. London: Edward Arnold Pub-
lishers, 140 pp.
Guisado, Ch.. & Castilla, J. C. 1983. Aspects of the ecology and growth of
an intertidal juvenile population of Concholepas concholepas (Mol-
lusca: gastropoda: Muricidae) at Las Cruces. Chile. Mar. Biol. 78:99-
103.
Hawkins. S. J. & R. G. Hartnoll. 1983. Grazing of the intertidal algae by
marine invertebrates. Oceanogr. Marine Biol. Aniui. Rev. 21:195-282.
Himmelman, J. 1991. Diving observations of subtidal communities in the
northern Gulf of St. Lawrence. In: J.-C. Therriault, editor. The Gulf of
St. Lawrence: small ocean or big estuary'.' Canadian Special Publication
of Fisheries and Aquatic Sciences. Fisheries and Oceans. Canada pp.
319-332.
Himmelman. J. & Y. Lavergne. 1985. Organization of rocky subtidal com-
munities in the St. Lawrence estuary. Naturaliste Canadiene 1 12:143-
1.54.
Himmelman. J. & C. Dutil. 1991. Distribution, population structure and
feeding of subtidal seastars in the northern Gulf of St. Lawrence. Ma-
rine Ecol. Prog. Series 76:61-72.
Hughes, R.. M. Burrows & S. Rogers. 1992. Ontogenic changes in foraging
behaviour of the dog whelk Nncella lapillus (L). J. Exper, Marine Biol.
Ecol. 155:199-212.
Keough, M. & A. Butler. 1979. The role of asteroid predator in the orga-
nization of a sessile community on pier piling. Mar. Biol. 51:167-177.
164
Stotz et al.
Koh-Sian Tan. 2000. Feeding habits of Chicoreiis copticimis (Neogas-
tropoda Muricidae) in a Singapure Mangrove. Menfit, Italy: 3° Work-
shop Inlernacionale di Malacologia.
Lowell. R, B. 1984. Desiccation of intertidal limpets: effects of shell size,
fit to substratum, and shape. / Exper. Marine Biol. Ecol. 77:197-207.
Menge. B. 1974. Competition for food between two intertidal starfish
species and its effect on body size and feeding. Ecology 53:635-644.
Menge, B. 1979. Coexistence between the seastars Asterias vulgaris and A.
forbesi in a heterogeneous en\ironnient: A non equilibrium explana-
tion. Oecologia 41:245-272.
Menge, B, 1992a. Community regulation: under what conditions are bot-
tom-up factors important on rocky shores? Ecology 73:755-765.
Menge. B. 1992b. Effect of feeding on the environment: Asteroidea. In: M.
A. Jangoux & J. Lawrence, editors. Echinoderm nutrition. Roterdam,
Netherlands. A Balkema, pp. 521-551.
Menge. B. & J. Sutherland. 1976. Species diversity gradients: synthesis of
the roles of predation, competition, and temporal heterogeneity. Am.
Naturalist 110:351-369.
Menge, B. & J. Sutherland. 1987. Community regulation: variation in
disturbance, competition, and predation in relation to environmental
stress and recruitment. Am. Naturalist 130:730-757.
Moreno, C. A. & E. Jaramillo. 1983. The role of grazers in the zonation of
intertidal macroalga of the Chilean coast. Oikos 41:73-76.
Moreno, C. A., J. P. Sutherland & H. F. Jara. 1984. Man apredator in the
intertidal zone of southern Chile. Oikos 42:155-160,
Moreno, C. A.. K. M. Lunecke & M. L Lepez. 1986. The response of an
intertidal Concholepas concholepas (Gastropoda) population to protec-
tion from man in southern Chile and the effects on benthic sessile
assemblages. Oikos 46:359-364.
Paine, R. T. 1966. Food web complexity and species diversity. Am. Natu-
ralist 100:65-75.
Pearre, S. 1982. Estiinating prey preference by predators: Uses of various
indices, and a proposal of another based on x2. Can. J. Fish Aquatic
Sci. 39:914-923.
Perez, E. P. & W. B. Stotz. 1992. Comparaciones nitiltiples de parametros
gravimetricos entre poblaciones submareales de Concholepas con-
cholepas (Bruguiere, 1789) en el Norte de Chile. Revista de Biologi'a
Marina. Valparaiso 27:175-186.
Pino. C, Oliva, D. & Castilla. J. C. 1993. Ritmos de actividad en Fissurcllu
crassa Lamarck 1822 y F. latimarginata Sowerby 1835: Efectos del
ciclo de marea, fotoperiodicidad y estacionalidad. Xlll Jornadas de
Ciencias del Mar. Vina del Mar, Chile, 26-28 de Mayo.
Powers, S. & J. Kittinger. 2002. Hydrodynamic mediation of predator-prey
interactions: differential patterns of prey susceptibility and predator
success explained by variation in water flow. J. Exp. Marine Biol. Ecol.
273:171-187.
Raymont. J. E. G. 1980. Plankton and productivity in the oceans. Vol. 1.
Phytoplankton. Oxford. UK: Pergamon Press Ltd.. 489 pp.
Rilov, G., A. Gasith & Y. Benayahu. 2002. Effect of an exotic prey on the
feeding pattern of a predatory snail. Marine Environ. Res. 54:85-98.
Saier, B. 2001. Direct and indirect effects of seastars Asterias rubens on
mussel beds {.Mylilus edulis) in the Wadden sea. J. Sea Res. 46:29^2.
SERNAP. (Servicio Nacional de Pesca). 1999. Anuario estadi'stico de
pesca. Chile: Ministerio de Economia. Fomento y Reconstruccion.
Sielfield, W. 1990. Dieta del chungungo iLutra felina (Molina. 1782))
(Mustelidae. Camivora) en Chile austral. Investigaciones Cientificas y
Tecnologicas. Serie Ciencias del Mar 1:23-29.
Sokal. R. S. & F. J. Rohlf. 1969. Biometry. San Francisco: W. H. Freeman
and Company. 776 pp.
Sommer. H.-J. 1991. Zur Okologie ausgewahlter Benthosarten im chilenis-
chen Felslitoral: Wiederbesiedlungsversuche mit Concholepas con-
cholepas. Diplomarbeit aus dent Zoologischen Instilul der Christhian-
Albrechts-Univeristat zu Kiel. Kiel. BRD. 81 p.
Sommer. H.-J. & W. Stotz. 1991. ,,Puede el loco Concholepas coiuholepas
(Bruguiere. 1789) vivir en fondos blanqueados? XI Jornadas de Cien-
cias del Mar. Viiia del Mar (Chile). 27-29 de Mayo.
Stotz, W. 1997. Las areas de manejo en la ley de pesca y acuicultura:
primeras experiencias y evaluacion de la utilidad de esta herramienta
para el recurso loco. Estiidios Oceanologicos 16:67-86.
Stotz, W. & Perez, E. 1992. Crecimiento y productividad del loco Conc-
holepas concholepas (Bruguiere. 1789) como estimador de capacidad
de carga en areas de manejo. Invest. Pesqueras 37:13-22.
Stotz, W. & D. Lancellotti. 1993. Asentamiento del loco Coiuholepas
concholepas (Bruguiere. 1789) en la zona intermareal: Excepcion o
regla. XIII Jornadas de Ciencias del Mar. Vit'ia del Mar (Chile). 26 al
28 de Mayo
Thomas. A.C.. F. Huang. P.T. Strubb & James, C. 1994. Comparison of the
seasonal and interannual variability of phytoplankton pigment concen-
trations in the Peru and California Current systems. J. Geophysical Res.
99:7355-7370.
Tokeshi. M.. C. Estrella & C. Paredes. 1989. Feeding ecology of a size
structured predator population, the South American sun star Heliaster
heliamhus. Marine Biol. 100:495-505.
Underwood. A. J. 1979. The ecology of intertidal gastropods. Adv. Marine
Biol. 16:111-210.
Vasquez. J. 1993. Abundance, distributional patterns and diets of main
herbivorous and carnivorous species associated to Lessonia traheculata
kelp beds in northern Chile. Facultad de Ciencias del Mar. Universidad
Catolica del Norte. Cocpnmho Chile Serie Ocasional 2:213-229.
Vasquez. J. & J. Castilla. 1984. Some aspects of the biology and trophic
range of Cosmasleria lurida (Asteroidea. Asteriinae) in belts of Mac-
rocystis pyrifera at Puerto Toro. Cliile Medio Ambiente 1:47-5 1 .
Vermeij. G. J.. H. A. Lescinsky, E. Zipser & H. E. Vermeij. 1994. Diet and
mode of feeding of the muricid gastropod Acanthinucella liigubris
angelica in the northern Gulf of California. Veliger 37:214-215.
Viviani. C. A. 1975. Las comunidades marinas litorales en el Norte Grande
de Chile. Iquique. Chile: Publicacidn ocasional Laboratorio de
Ecologi'a Marina. 196 pp.
Whittaker. R. H. 1975. Communities and Ecosystems (second edition).
New York: Macmillan Publishing Co.. Inc.. 387 pp.
.loiiriHil ,.f Shellfish Research. Vol. 22. No. I. I6.';-I64, 200.^.
FEEDING AND GROWTH IN THE KEYHOLE LIMPET, FISSURELLA FICTA (GMELIN, 17911
D. A. LOPEZ.* M. L. GONZALEZ. AND M. C. PEREZ
Ltihoiaiono de Cultivos Marinas. Dcpariumeiiio dc Aciiiculliira. Univcisidad de Los Liigos, Casilla 933,
Osorno. Chile
ABSTRACT The feeding habits and growth relationships ol the keyhole limpet ("lapa") hissiirelhi pieki were analyzed in the field
and under laboratory conditions. This species is of significant commercial value and considerable ecological importance in southern
Chile. F. picta is not strictly a herbivore, although it prefers algae; the quantity of vegetable items con.sumed compared with animal
items did not vary seasonally. The items most commonly found in the stomach of F. picta were the algae Ulva sp, Polysophonia sp
and Gelidhim .tp. The abundance pattern of the principal items did not vary seasonally. However, there was greater diversity in the
summer. The relative abundance of items in the diet was closely associated with their relative abundance in the environment. Under
laboratory conditions, adults showed a higher consumption rate for the alga Gracilaiia chileii.\is (artificial diet) than for Ulva sp
(natural diet). The preferred alga is not usually found in the natural habitat of F. picta and has a lower caloric value than that of Ulva
sp. C. chilensis proved to be the best source of energy available for growth in juveniles. Keyhole limpets feeding on the chlorophyte
alga Ulva sp show a negative energy balance. Specimens maintained in suspended systems and fed with the artificial diet (G. chilensis)
reached the average commercial size of 5.^ mm in -3 y; the average survival rate was 90"/?. The results suggest that keyhole limpets
prefer food with a high energetic scope for growth, although in field conditions they consume food with a lower energetic content but
high in abundance. Factors such as morphology or palatability of food are more important than caloric value or presence in the natural
habitats of keyhole limpets. This information is important for the culture of the keyhole limpet.
KEY WORDS: feeding, scope for growth, keyhole limpet. Fi.'.siirella picia
INTRODUCTION
Keyhole limpets ("lapas") of the genus Fissurella are grazing
molluscs that consume a wide variety of macroalgae in the inter-
tidal zone (Branch 1981, Hawkins & Hartnoll 1983). Previous
studies indicate that they also ingest other types of food, such as
crustaceans, small molluscs, coralline algae, ostracods, and
sponges, although they remain preferentially herbivores (Ward
1966. Bretos 1978, Santelices & Coirea 1985, Osorio et al. 1988).
Among Chilean species of lapas, Fissurella crassa is classified
as a generalist herbivore, which prefers to consume foliacious
algae, such as Ulva sp.. Emeromorpha sp and Porphyni sp (Bretos
1978, Santelices et al. 1986). Data available on F. maxima, based
on studies of its stomach contents, indicate that this species is
euriphycophagous (Osorio et al. 1988). Experimental field studies
on F. picta suggest that this species is a nocturnal herbivore, which
migrates during the night to the middle intertidal zone (Jara &
Moreno 1984. Moreno et al. 1984), to feed on the algae Iridaea
horycma and Ulva rigida.
F. picta. has an important commercial value, and over-
harvesting has resulted in the depletion of natural stocks in south-
ern Chile (Bretos 1978. Bretos 1988). In addition, human exploi-
tation of other species has, indirectly, had a negative effect on
keyhole limpet recruitment (Lopez et al. 1999). This species also
has ecological importance given that it can modify the spatial and
temporal distribution patterns of intertidal macroalgae (Mi)reno et
al. 1 984). Knowledge of the diet and dietary preferences of F. picta
is necessary to evaluate its growth rate in artificial cultures and to
interpret the ecological role of the population under field condi-
tions.
Published literature suggests that the interaction between quan-
tity and quality of food with factors such as pH. temperature and
salinity, influences growth in mobile marine invertebrates (Newell
1979. Frantzis & Gremare 1992). The effect of type of food in-
gested on growth can be determined by measuring the increase in
*Conesponding author. Fax: -1-56-6-420-5271; E-mail: dIopezCfl'ulagos.cl
weight or size of the animals, or in terms of energy through scope
for growth, established by evaluating the components of the energy
balance (Paine 1971, Bayne & Newell 1983. Gonzalez et al 1990,
Gonzalez et al. 1993, Thompson & MacDonald 1991, Navarro &
Torrijos 1994, NavaiTO & Torrijos 199.S). The aims of this study
are to determine the feeding habits of the keyhole limpet, F. picta
(Gmelin) in the field and under laboratory conditions and to es-
tablish the relationship between feeding and growth.
MATERIALS AND METHODS
Stomach Conteiil in the Wild
The feeding habits of keyhole limpets were observed in the
intertidal and subtidal zone of Metri Bay (4r36'S, 72°42'W), in
southern Chile. The stomach contents of 40 F. picta specimens
(between 32.9 and 64.8 mm total length) were analyzed per season.
Specimens collected at high tide were immediately injected with
formalin dissolved in seawater to stop digestion. The stomach
contents were analyzed over a 100-point grid (81 mm~). Thus, it
was possible to determine ( I ) the relative frequency of vegetable
and animal items; (2) the relative frequency of empty and full
stomachs; and (3) the quantity and frequency of each item in the
diet. A reference collection of all fronds of alga species present in
different habitats and at different periods of the year was estab-
lished to facilitate the identification of alga species consumed by
lapas. Analysis was carried out under a dissecting scope. The
relative abundance of sessile species present in the study area was
verified during each season, based on coverage, using a 100-point
grid 0.0625 m~ along ten linear transects of 15-18 m in the inter-
tidal zone (Bumham et al. 1980).
The statistical comparison between vegetable and aniinal con-
tent in keyhole limpets was carried out by the x" test. The differ-
ences in dietary preference and energy consumed and lost in ani-
mals feeding on Ulva sp. and G. chilensis were analyzed with a
r-tesl. Using correlation analysis, the relative abundance of algae in
the diet was associated with the food supply of algae available in
the environment.
165
166
Lopez et al.
"Scope for growth" with Natural and Artificial Diets
Juveniles of F. picta (length between 25.0-32.3 mm) were
collected from the rocky intertidal zone in Metri Bay. The animals
were separated into two groups and acclimated in the laboratory at
10°C ± 1°C for 20 days. During the experimental phase, each
group was fed ad libitum with Ulva sp (Chlorophyla) or C. chil-
ensis (Rodophyta).
All the parameters of energy balance were standardized as
joules per day per gram of shell-free dry weight (J • d~' • gdw~').
(using 1 cal = 4.18 J) (Lucas & Beninger 1985). Animal dry
weight was obtained using the regression equation for length ver-
sus dry weight, calculated for 150 keyhole limpets with lengths
between 20.0-36.0 mm.
The experimental procedures for the two groups were as fol-
lows:
To evaluate the effect of natural and artificial diets on ingestion
rate, 40 F. picta specimens of 42.2 ± 9.5 mm total length, collected
in the middle and lower rocky intertidal zone of Metri Bay, were
transferred to aquaria for an acclimation period of 13 days at 15°C
± 1°C. The specimens were permanently submerged and the water
was changed every 5-7 days. The ingestion rate of two types of
macroalgae was compared: Ulva sp, which is the most frequent
item found in the habitat of F. picta (natural diet) and G. chileiisis.
a rhodophycean species of alga, not present in the keyhole limpet's
natural habitat (artificial diet). G. chilensis is the principal species
used in artificial culture with an average annual production of
821 19.5 ton y~' (Semap 1998). The two alga species have distinct
forms: Ulva sp is foliaceus and G. chilensis is ramified. Each alga
species was offered ad libititm to two groups of twenty animals
of similar sizes kept in 1-L individual aquaria. The ingestion rate
was measured gravimetrically, at 7-day intervals. An aquarium
containing only alga samples was used as a control. The inges-
tion rate was obtained by comparing differences in alga weight
at the beginning and end of the experiment, expressed in grams
of dry weight of algae consumed per individual per day (gdw •
ind"' • d"'). Measurement of alga consumption was adjusted ac-
cording to percentage weight variation of algae in the controls. No
animal items were used as food because F. picta feed principally
on algae and an important fraction of animal items in its diet are
epiphytic organisms. The caloric contents of the Ulva sp and G.
chilensis used in the experiments was measured with a Parr bomb
calorimeter. Energy consumed (C) was determined using the ca-
loric value of the algae.
The energy loss due to metabolism (R) was measured in 39
animals as the standard oxygen consumption in a 145-mL hermetic
flask using a WTW-530 oxygenometer (0.01 mg 0,/l accuracy).
For conversion into energy, the Thompson and Bayne ( 1974) oxi-
caloric value of 1 niL O, = 19.95 J was used.
The excretion rate of ammonia (U) was determined in 40 in-
dividual keyhole limpets measuring the concentration of ammonia
accumulated over a period of 15 min in 200 ml aquaria, using the
Solorzano method (Solorzano 1969). Conversion into energy units
was carried out using the Elliot and Davison ( 1975) constant of 1
mg NH4* = 24.85 J. The energy loss through feces (F) was
measured in 15 keyhole limpets that were placed individually in
1-L aquaria containing filtered seawater (mesh size: 1 |j,m) that
was changed daily and with a constant supply of air. The feces
were collected every 1 2 h according to methods described by
Navarro and Thompson (1996). rinsed with isotonic solution of
ammonium formate, kept in containers, and dried in a Memmert
500 furnace at 75°C until a constant weight was reached. The
caloric value of the feces was determined in a Parr adiabatic bomb
calorimeter. Energy loss through mucus (M) was evaluated by
filtering water through 120-|jLm mesh.
The energy values of scope for growth were calculated accord-
ing to the following equation, using above average calculated val-
ues:
P = C-(F-^R + U-fM)
where P = scope for growth; C = energy from food consumed:
F = fecal energy loss; R = metabolic energy loss; U = energy
loss due to excretion and M = mucus.
Determination of Absorption Efficiency
Absorption efficiency was calculated using the Conover equa-
tion (Conover 1966):
AE =
(F-E)
(I -E)x F'
100
where AE = absorption efficiency (9<-); F = ash-free dry weight
food/total dry weight food and E = ash-free dry weight feces/total
dry weight feces.
To determine the algal and fecal organic matter content, algae
and feces were carefully rinsed with distilled water and then dried
in a Memmert 500 furnace at 75°C. until constant weight was
reached. The samples were then incinerated in a muffle furnace at
450°C for 4 h. The organic matter was obtained by establishing the
difference between the constant weight and the weight of the ash
of each sample after incineration.
The results of all the above determinations were then compared
(differences between animals fed with a diet of G. chilensis or
Ulva sp), using one-way ANOVA after logarithmic transformation
(Sokal & Rohlf 1979).
Dietary Preference — Natural and Artificial Diets
The same quantity of Ulva sp and C. chilensis (volume and
weight) was supplied simultaneously to a group of 20 individuals
of 46.7 ± 9.5 mm total length. The amount of algae consumed by
each specimen was determined daily, based on the biomass varia-
tions, with an electronic balance (±0,01g accuracy). A control was
also set up.
Growth of Keyhole Limpets in Suspended Systems Feeding on an
Artificial Diet
The direct effects of the artificial diet on keyhole limpets'
growth and mortality were determined in ailificial cultures.
This study was carried out over 12 months in Metri Bay. At this
location, average water temperature varies between 9.6°C (winter)
and I8.2"C (summer); salinity fluctuated between 28%f and 32%t
during the study period.
Two hundred and forty specimens of F. picia collected from the
intertidal zone were placed in trays ("lintemas") that were sus-
pended from a raft. Specimens were fed ad libitum with the red
alga G. chilensis. Four size categories were used. Initial average
size and the standard deviations of keyhole limpets placed in ex-
perimental growth systems (n = 20 per group) were; group 1 : 25.9
+ 1.3 mm; group 2: 31.9 ± 1.9 mm; group 3: 37.8 ± 0.7 mm and
group 4; 45.0 ± 0.9 mm. The experiments were replicated three
times. Total weight and maximun length were measured monthly.
Feeding and Growth in Fissurella picta
167
n Without gasinc content
■ With gasdic content
100
80
60
40
20
lL
Ax
s
w
Sp
Figure
without
(Sp).
SEASON
1. Relative seasonal frequencj of Fissiirella picta with and
gastric content. Summer (Si; Autumn (A); Winter (W); Spring
RESULTS
D = Vegetable items
80 n
J 1 = Animals iteins
Aimjr^
? «°
« 40
01
3
T
« 20 .
i
><
o
80
60
40
20
Sloinaih Contents in the Wild
WINTHR
The relative frequency of F. picta specimens with empty stom-
achs was less in autumn and winter than in summer and spring
(Fig. 1). The percentage of vegetable items was always signifi-
cantly higher than the animal items {P < 0.05), with no variation
between different periods of the year (Fig. 2).
The most frequent items present in F. picta stomachs were the
algae Ulva sp, Pohsiplioitia sp, and Gelidiuin sp, especially during
autumn. The main animal items were cirripedes and juvenile bi-
valves (Fig. 3). There was a positive correlation between the rela-
tive abundance of food items present in the stomachs throughout
the year and the relative abundance of these items in the environ-
ment (r = 0.891; n = 65: P < 0.05).
Scope for Growth
The diets used in scope for growth measurement had different
energy values. The energy content of Ulva sp ( 1 3,990.5 J • gdw^' )
was higher than that of G. chilensis ( 1 1.101 2 J • gdw"' ). The type
of food intluenced the energy balance and the scope for growth.
The scope for growth was highest when F. picta consumed G.
chilensis (Table 1 ). The negative energy balance in specimens fed
with Ulva sp was due to energy loss (Table 1 . 1 The amount of
energy consumed by F. /nrfcv juveniles did not vary significantly in
animals fed with G. chilensis and those fed with Ulva sp (t =
100
5- 80
5" 60
c
s «
e
ll 20
D Vegetable items
■ Animals items
S
w
Sp
SEASON
Figure 2. Relative seasonal frequency of vegetahle and animal items in
gastric content o( Fissiirella picta. Summer (S): Autumn (A): Winter
(W); Spring (Sp).
><
u
ou -
60-
SPRING
40
20
T
n 1.
80
£ 60
>t
u
S 40
20 4
U
SlMJBi
a
Ex
Ch
Jb
Items
Figure 3. Seasonal frequency (average ± standard deviation) of food
items in the stomachs of Fissurella picta. Ulva sp (U); Chondrus sp
(Ch); Gelidiuin sp ((i); I'olysiphonia sp (P); Fnteroinorpha sp (E); Cir-
ripeds (C); juvenile hivalves (Jb); Sodilittorina araucana (I.).
0.098; df = 28; P< 0.005) (Table 1.). The quality of food affected
the metabolic losses in F. picta (Fig. 4A). Oxygen consumption
was significantly higher in animals fed with Ulva_sp than in those
fed with G. chilensis (t = 5.48: df = 37; F < 0.001 ). The energy
loss due to excretion was significantly higher in animals fed with
G. chilensis, 23.0 J • d"' • gdw~', than in those fed with Ulva sp,
5.4 J • d"' • gdw-' (t = 8.10; df = 13; P < 0.001). The fecal
energy loss was also affected by the quality of food (Fig. 4B).
Specimens fed Ulva sp had significantly higher fecal energy los.ses
than those fed G. chilensis (ts = 6.56; df = 13; P < 0.001 ), Since
no mucus was found in the aquaria, and given that this value would
only represent IVc of the energy ingested in herbivorous molluscs
(Paine 1971 ). energy loss through mucus (M) was not considered.
168
Lopez et al.
TABLE 1.
Energy ingested, energy loss and scope for growth in Fissiirella piciii
juveniles fed with L'lva sp (natural dietl or Gracilaria chitensis
(artificial diet) in joule/day/gram dry weight of soft parts.
Food
Parameter
Ulva sp
Gracilaria
chileiisis
Range of energy ingested
(J ■ d"' gdw"')
Total energy loss (J • d~' ■ gdw"
Average scope for growth
(J -d"' -gdw-')
605.3-1.504.3
740-1.920.3
803 ±238.6 4(.)y ± 140.2
-10.4 390.6
Absorption Efficiency
Absorption efficiency was highest in specimens fed Ulva sp.
83.4%, and lowest in those fed G. cliilensis. 74,6% (x" = 0.49;
df = 1; 0.05).
Dietary Preference — Natural and Artificial Diets
In specimens of F. piclci. consumption rates of C. chilensis
(artificial diet) were higher than those of Ulva sp (natural diet)
(t = 76.12; df = 27; P < 0.001 ) and they also presented a greater
preference for C. chilensis than for Ulva sp (t = 19.89; df = 28;
P<0.00\).
Growth of Keyhole Limpets in Suspended Systems
The alga G. cliilensis proved to be suitable food for growth
and survival in keyhole limpets. The annual average survival rate
was 90'7(- under these experimental conditions. The growth rates of
the animals varied according to size. Using these data it was cal-
culated that F. picta reached 26.0 mm in about 14 mo. Thus, the
average commercial size of .53 mm would be achieved in approxi-
mately 3 y (Table 2).
DISCUSSION
The results obtained indicate that F. picta is preferentially a
herbivore, as has been described for other species of this genus.
(Osorio et al. 1988. Santelices et al. 19861. However, it also con-
sumes animal items. Similarly, the high consumption of foliaceus
species such as Ulva sp (Jara & Moreno 1984) was also confirmed.
This can be associated with the food supply available in the envi-
TABLE 2.
Growth in four groups in = 20) of Fissurella picta in suspended
cultures, feeding on Gracilaria chilensis (artit'icial diet).
Initial Length
Final Length
Time
Group
( mm 1
(mml
(Month)
1
25.9 ±1.31
38.6 ± 4,4
12
48.2 ±0.1
2!
2
31.98+ 1.97
46.3 ± 5.7
12
3
37.86 ± 0.76
50.0 ± 5.3
12
4
45.03 ± 0.95
54.6 ± 1.2
X
55.3 ± 2.3
14
<B —
600
500 <.
400
O O) 300
■S '
f 3 200
s
100
_^
B 600 n
0)
» 500
c
O — 400
t3 s
= E, 300 -
|'5 200-
01 100 -
u
0)
u. -
G U
Figure 4, P'nergy loss through metabolism (,\) and feces (B) in Fis-
surella picta feeding Gracilaria chilensis ((J) or Viva sp (Li I,
ronnient, as has been verified in other species of Fissurella (San-
telices et al. 1986). Ulva sp and Polysiphonia sp, the most frequent
items in the keyhole limpets" stomachs, are opportunist algae spe-
cies in the field. They densely colonize the intertidal zone of Metri
Bay (Buschmann 1991).
The higher consumption rates, trophic preference, and scope
for growth obtained with G. chilensis. which is not usually found
in the natural habitat of F. picta. compared with those for Ulva sp.
indicate that food items might not be selected due to their energy
characteristics. The trophic preference is not related to the caloric
value of food, given that Ulva sp has a higher caloric value than C.
chilensis, and the energy budget was not associated with the food
availability in the field. Although the laboratory results cannot be
reliably extrapolated to the natural habitat, it can be assumed that
the preference for macroalgae consumption may be associated
with their digestibility, morphology, or palatability (Lowe &
Lawrence 1976. Tugwell & Branch 1992). Although the chemical
defenses of algae are lower than in terrestrial plants, the secondary
compounds related to the plant-herbivore relationship, cannot be
discarded (Hay & Fenical 1992). Further research is required to
test these hypotheses.
The scope for growth in juvenile limpets varied according to
the algal food offered. Specimens fed with G. chilensis (artificial
diet) presented a positive energy balance. Considering the fact that
the specimens studied were juveniles that had not yet reached
sexual maturity, the balance of the energy budget can be consid-
ered as energy available for growth. In species such as the gastro-
pod Concholepas concholepas (Bruguiere 1789) and the echino-
derm Lo.xechinns alhiis (Molina 1782), it has been shown that the
type of food offered greatly influences both the "sign" of energy
balance and the amount of energy available for growth (Gonzalez
et al. 1990, Gonzalez et al. 1993). These results coincide with
those obtained in F. picta.
Feeding and Growth in Fissurella picta
169
Keyhole limpets maintained in suspended cultures and fed ex-
clusively on G. chilensis. had high sur\i\al rates. This study in-
dicates that the type of food offered can have a considerable in-
fluence on the growth rate of juvenile F. picta. Our data mdicate
that, under artificial conditions, it can be possible to maximize
reproduction and growth by selecting the food items offered. This
finding could ha\e significant consequences for cultivation of this
important resource. Other factors, however, such as culliire den-
sity, must be investigated to obtain higher growth rates.
ACKNOWLEDGMENTS
The authors thank FONDECYT for the financial support
through Grant 040-93. University of Los Lagos for pro\iding
the facilities. Dr. J. Jimenez and anonymous referees for the
critical review. J. M. LIribe. J. Castro, and C. Pino for the collabo-
ration in field and laboratory measurements. S. Mancilla for pro-
viding secretarial assistance, and S. Angus for translating the
manuscript.
LITERATURE CITED
Bayne. B. L. & C. Newell. 1983. The mollusca. physiological energetic of
marine mollusca. In: K. M. Wilbur, editor. The Mollusca. Vol. 4 Physi-
ology Part 1: New York: Academic Press, pp. 407-499.
Branch. G. H. 1981. The biology of limpet: physiological factors, energy
flow, and ecological interactions. Oceanogi: Mar. Biol. .Ann. Rev. 19:
235-280.
Bretos, M. 1978. Growth in the ke\hole limpet F/.s.vi(r<'//(i cni.s.sn Lamarck
(Mollusca: Archaeogastropoda) in Northern Chile. \'eHi>er 21:268-
273.
Bretos. M. 1988. Pesqueria de lapas en Chile. Medio ,\mlnen!e 9:7-12.
Burnham. K. P.. D. R. Anderson & J. L. Laake. 1980. Estimation of density
from line transect sampling of biological population. Wildlife Mono-
graphs 72:10-202.
Buschmann. A. H. 1991. Algal communities of a wave protected intertidal
rocky shore in Southern Chile. In: U. Seeliger, editor. Algal commu-
nities in Latin America Florida. New York: Academic Press pp, 91-
103.
Conover, R. J. 1966. Assimilation of organic matter by zooplankton. Lim-
nol. Oceanogr. 1 1 :338-354.
Elliot. J. M. & W. Davison. 1975. Energy equivalents of oxygen consump-
tion in animal energetics. Oecologia 19:195-201.
Frantzis. A. & A. Gremare. 1992. Ingestion, absorption and growth rates of
Paracentrotus lividus (Echinodermata: Echinoidea) fed with different
macrophytes. Mar. Ecol. Prog. Series 95:169-183.
Gonzalez. M. L., M. C. Perez. D. A. Lopez & M. S. Buitano. 1990. Efecto
de la temperatura en la disponibilidad de energia para crecimiento de
Coneholepas eoncholepas (Bruguiere). Rev. Biol. Mar. 25:71-81.
Gonzalez. M. L.. D. A. Lopez. M. C. Perez & J. M. Castro. 2002. Effect
of temperature on the scope for growth in juvenile scallops .Argopecien
piirpuraliis (Lamark. 1819). Aqiiaculture Intemalional.
Hay. M. E. & W. Fenical. 1992. Chemical mediation of .seaweed-herbivore
interactions in Plant-animal interactions in the Marine Benthos. In: D.
M. John. S. J. Hawkins & J. H. Price, editors. Systematics Association
Special Volume No 46. Oxford: Claredon Press, pp. 319-338.
Hawkins. S. J. & R. G. Hartnoll. 1983. Grazing of intertidal algae by
marine invertebrates. Oceanogr. Mar. Biol. Ann. Rev. 21:195-282.
Jara. F. H. & C. A. Moreno. 1984. Herbivore and structure in the mid-
littoral rocky community: a case in Southern Chile. Ecology 65:28-38.
Lopez. D. A.. M. L. Gonzalez, J. M. Uribe. R. 1. Martinez & P. A. Vergara.
1999. Effect of cirripeds on the recruitment of the keyhole limpet
Fissurella picta (Gmelin). Ciencias Marinas 25:75-90.
Lowe. E. F. & J. M. Lawrence. 1976. Absorption efficiency of Lyiechinu.s
variegatus (Lamarck) (Echinodermata: Echinoidea) for selected marine
plants. ./. E.xp. Mar Biol. Ecol. 21:223-234.
Lucas. A. & A. G. Beninger. 1985. The use of physiological condition
indices in marine bivalve aquaculture. Aquaculture 44: 1 87-200.
Moreno. C. A.. J. P. Sutherland & H. F. Jara. 1984. Man as predator in the
intertidal zone of Southern Chile. Oikos 42:155-160.
Navarro. J.M. & R.A. Torrijos. 1994. Seasonal variation in oxygen uptake
and ammonia excretion in the predatory gastropod Coneholepas con-
eholepas (Bruguiere. 1789). Compar. Biochem. Physiol. 108:39-46.
Navarro. J. M. & R. A. Torrijos. 1995. Fisiologia energetica de Conehole-
pas coneholepas (Bruguiere. 1789) (Gastropoda: Muricidae) en la
bahfa de Yaldad. sur de Chile. Rev. Chilena de Historia Natural 68:
61-77.
Newell. R. C. 1979. Biology ol intertidal animals. Mar. Ecol. Suney Ltd.
Fawerham. Kent. 757 pp.
Osorio. C. M. E. Ramirez & J. Salgado. 1988. Gastric contents of Fis-
surella nucxima (Mollusca: Archaeogastropoda) at Los Vilos. Chile.
Veliger 30:347-350.
Paine. R. T. 1971. Energy flow in natural population of the herbivorous
gastropod Tegula funebralis. Limnol. Oceanogr. 16:86-98.
Santelices. B. & J. Correa. 1985. Differential survival of macroalgae to
digestion by intertidal herbivore molluscs. J. E.xp. Biol. Ecol. 88:183-
191.
Santelices. B.. J. Vasquez & I. Meneses. 1986. Patrones de distribucion y
dietas en un gremio de moluscos y herbivoros en habitats intermareales
expuestos de Chile Central. Publicaciones Periodicas. P. Universidad
Catolica de Chile. Monogr. Biol. 4:147-171.
Servicio Nacional de Pesca. (Semap) 1998. Anuario Estadi'stico de Pesca..
Valparaiso. Chile. 240 pp.
Sokal. R. & J. Rohlf 1979. Biometria. San Francisco: Freeman & Com-
pany.
Solorzano. L. 1969. Determination of ammonia in waters by the phenol-
hypochlonte methods. Limnol. Oceanogr. 14:799-801.
Thompson. R. J. & B. L. Bayne. 1974. Some relationships between growth
metabolism and food in the mussel Myiihis edidis. Mar. Biol. 27:317-
326.
Thompson. R. J. & B. A. MacDonald. 1991. Physiological integrations and
energy partitioning. In: S. E. Shumway, editor. Scallops: biology, ecol-
ogy and aquaculture. Developments in aquaculture and fisheries sci-
ence. 21. Amsterdam: Elsevier, pp. 347-378.
Tugwell. S. & G. M. Branch. 1992. Effect of herbivore gut surfactants on
kelp polyphenol defense. Ecology. 73:205-215.
Ward. J. 1966. Feeding, digestion, and histology of the digestive tract in the
keyhole limpet Fissurella harbadensis Gmelin. Bull. Mar. Sci. 16:668-
683.
Joiimal oj Shellfish Research, Vol. 22. No. 1. 171-175. 2003.
A COMPARISON OF THE DIGESTIVE CAPACITY OF BLACKLIP {HAUOTIS RUBRA) AND
GREENLIP {HAUOTIS LAEVIGATA) ABALONE
MEEGAN E. VANDEPEER'* AND ROBERT J. VAN BARNEVELD"
'Soiith Australian Research and Development Institute. PO Box 120. Henley Beach. South Australia
5022 and 'Barneveld Nutrition Pty. Ltd.. 19-27 Coonan Rd. South Maclean.
Queensland. Australia 42S0
ABSTHACT In this study, the digestive capacity of blacklip ahalone. Haliolis nihni Leach, was compared whh that ol the greenlip
ahalone. Halioris Uicvigaw Donovan. This was performed by assessing each abalone species ability to digest the protein and energy
from 12 ingredients; semolina, defatted soytlour. fishmeal. casein, pregelatini/ed maiz.e starch, mung beans, whey powder, skim milk
powder, whole lupins [LiipifiKS anxKstifoliKs and Liipiiuis hileiis). dehulled lupins [L. cmfiiisiifolii(s). and bull kelp iDiinillea potci-
tonim). Significant differences were found between the two abalone species in their capacity to digest the protein and energy from some
of the ingredient.s assessed. Based on the differences observed, it was hypothesized that blacklip abalone are more efficient at digesting
protein and cellulose than greenlip abalone and greenlip abalone might have a greater capacity to digest soluble nonstarch polysac-
charides.
KEY WORDS: abalone. greenlip. blacklip. digestibility, protein, energy. Haliuiis rubra. Huliotis hievigiite
INTRODUCTION
Greenlip abalone [Huliotis laevigata) and blacklip abalone
{Huliotis rubra) are the predominant species commercially farmed
in Australia. Moratoriums on the collection of macroalgae for use
in commercial abalone production necessitate the use of manufac-
tured diets in these systems. To date, a significant amount of
research has been completed to characterize the nutritional quality
of ingredients and the nutritional requirements of greenlip abalone.
It is uncertain, however, whether this information is relevant to
blacklip abalone. If similarities exist between the digestive capac-
ity of greenlip and blacklip abalone, then a large proportion of the
research completed on the nutritional quality of ingredients for
greenlips need not be replicated for blacklips.
Studies investigating the feeding preference of blacklip and
greenlip abalone have shown that when given a choice, both spe-
cies prefer to eat red algae (Hone & Fleming, unpublished data;
Shepherd & Steinberg 1992, Fleming 1995). In the wild, however,
abalone are forced to eat what algae is available. For example.
along the coasts of Victoria blacklip abalone feed extensively on
the fronds of the large kelp Phyllopsara comosa whereas on Tas-
manian coasts they often feed on drifting blades of the giant kelp
Macrocxstis pyrifcra as well as on red algae (Shepherd 1975).
The structural and storage polysaccharides present in red and
brown algae are quite different. The storage polysaccharides in
brown algae are mannitol, a sugar alcohol, and laminaran, a glu-
can, whereas the storage polysaccharide for red algae is a starch
known as tloridean starch. The cell wall of brown algae are two
layered with an inner matrix of cellulose and microfibrils and outer
layer of alginic acid and sulphated fucans (Stewart 1974). The cell
walls of red algae consist of an inner rigid component made up of
microfibrils and an outer tnore amorphous component consisting
of mucilage or slime. The characteristic amorphous inucilages that
make up most of the rest of the cell wall (up to 709^) are usually
sulfated galactan polymers (Schweiger 1978). The two largest
groups are the agars and the carrageenans.
Because they differ in their structural and storage carbohy-
*Corresponding author.
Phone: -t-61 8 8 200 2466; Fax: -h61 8 8200 2481; E-mail: vandepeer.
meegan@saugov.sa.gov.au
drates, it is reasonable to suggest that different en/.ymes would be
required to digest red and brown algae. If, as the result of living in
different habitats, blacklip abalone consume different or a broader
range of algae than greenlip abalone, then it would be expected
that they might have a different digestive enzyme profile. If this
were so, then they may also differ in their capacity to digest the
nutrients from the ingredients that are used in manufactured diets,
particularly different carbohydrate sources.
Results from comparative studies conducted on other abalone
have shown there are differences between species in their nutri-
tional requirements or physiology. Mercer et al. ( 1993) examined
the nutritional value of eight algal diets for H. tuherctdala and H.
discus hannai by comparing feeding rates, growth rates, and bio-
chemical composition of the animals. The algae A. esculenta. L
saccharina. and U. lactuca were found to have different dietary
values for the two abalone species with quite different feeding
rates and feed conversion efficiency values being reported for
each. Significantly different responses in growth rates were also
recorded when fed particular diets. The lowest growth rates re-
corded for H. tuberculata occurred when it was fed with L. sac-
charina or C. crispus whereas the lowest growth rates recorded for
H. di.tcus hannai occurred when it was fed with U. lactuca. The
differences in dietary values of the algae to the two abalone species
were attributed to differences in their specific nutritive require-
ments and/or digestive physiology (Mercer et al. 1993).
Given that differences have been observed between other aba-
lone species in their ability to use the same algal diets (Mercer et
al. 1993), then it is possible that greenlip and blacklip abalone
differ in their digestive capacities and/or nutrient requirements.
This has important implications as feed costs represent a large
proportion of farm running costs in Australia and our current
manufactured diets are formulated based on results from research
done on greenlip abalone. The objective of this experiment was to
compare the protein and energy digestibility of a range of ingre-
dients for blacklip and greenlip abalone and thus establish whether
they differ in their digestive capacity.
MATERIALS AND METHODS
Diet Formulation and Manufacture
Twelve diets were fomiulated (Table 1 ) to evaluate the protein
and energy digestibility from semolina, defatted soyflour. Tasma-
171
172
Vandepeer and Van Barneveld
TABLE 1.
Composition of experimental diets (g/lig, air dry basis).
Diet
Ingredient
1
2
3
4
5
6
7
8
9
10
11
12
Semolina
400.0
_
_
-
-
-
-
_
_
-
_
-
Defatted soyHour
-
625.0
-
-
-
-
-
-
-
-
-
-
Tasmanian t'ishmeal
-
-
420.8
-
-
-
-
-
-
-
-
-
Casein
-
-
347.6
-
-
-
-
-
-
-
-
Pregelled starch
189.4
214.4
418.6
200.0
489.4
158.7
289.4
150.0
150.0
374.8
100.0
100.0
Mung beans*
-
-
-
-
-
630.7
-
-
-
-
-
-
Bull kelpt
-
-
-
-
-
-
500.0
-
-
-
-
-
Whey
-
-
-
-
-
-
-
600.0
-
-
-
-
Skim milk powder
-
-
-
-
-
-
-
-
600.0
-
-
-
Lupin Ij
-
-
-
-
-
-
-
-
-
389.6
-
-
Lupin 2§
-
-
-
-
-
-
-
-
-
-
421.1
-
Lupin 31
-
-
-
-
-
-
-
-
-
-
-
500.0
Jack Mackerel oil"
-
-
-
-
-
-
-
-
-
-
-
20.0
Mineral premix**
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
Vitamin premix**
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
Vitamin C
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
Vitamin E
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Sodium alginate
-
-
-
-
-
-
-
-
-
5.0
-
-
Kaolin
400.0
150,0
1 50.0
441.8
500.0
200.0
200.0
2.W.4
239.4
200.0
448.4
369.4
Chromic oxide
?.o
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
* Whole Vigna radiala.
t Dun'illea potatorum.
± Whole L liiteus.
§ Dehulled L angusiifoliKs.
']1 Whole L angusiifolius.
" Trachunis deciivis (Triahunna Fish Oils, Triabunna. Tasmania).
** Vitamin and mineral premi.xes as described by Uki et al. (1985).
nian fishmeal, casein, whey powder, skim milk powder, whole
mung beans {Vigna nidiata). pregeiatinized waxy maize starch,
bull kelp (Durviltea potatorum), and lupins (whole L. luteiis.
whole L aiigii.stifoIiK.s and dehulled L aiigKstifoliiis) by greenlip
and blacklip abalone. The crude protein and gross energy of each
of these ingredients is given in Table 2. Because of the wide range
in crude protein levels of the ingredients being evaluated, it was
TABLE 2.
Protein (g/kg, air-drj basis) and energ) (MJ/kg, air-dry basis)
content of the 12 ingredients used in the experimental diets.
Ingredient
Crude Protein
(.V X 6.25)
Gross Energy
(MJ/kg)
Semolina
104.0
Defatted sovtlour
480.0
Fi.shmeal
713.0
Casein
863.0
Pregelled starch
3.1
Mung beans
253.7
Bull kelp
69.0
Whey
135.0
Skim milk powder
361.0
Whole L Inteiis
385.0
Dehulled L cingiistifotius
380.0
Wht)le L. angiistift'
'lilts
320.0
15.51
17.45
18.71
22.00
15.65
16.54
10.77
15.20
17.26
18.03
18.28
17.74
not practically possible to formulate the diets to be isonitrogenous.
It is desirable for the diets to be isonitrogenous as it means that
unbiased comparisons can be made among the different ingredi-
ents in regard to the digestibility of their protein.
Before incorporation into diets, the mung beans and lupins
were cnjshed into a fine powder (<500 (xm) using a hammermill.
Each diet contained an equivalent amount of vitamin C (ascorbic
acid) and E (DL-alpha tocopherol) and vitamin and mineral pre-
mixes as described by Uki et al. (1985). Sodium alginate was
included in some diets to aid in binding. Kaolin and pregeiatinized
waxy maize starch were used in the diets as fillers. Chromic oxide
was included at 0.57r for use in subsequent digestibility calcula-
tions.
All diets were initially hand mixed and then mixed in a spiral
action dough mixer Clmpastrice". Hill Equipment and Refrigera-
tion. Adelaide, South Australia). The mixture was then fed through
a commercial pasta machine (La Prestigiosa medium. IPA. Vi-
cenza. Italy) where it was made into 300-mm long strips using a
die with slots 18 mm x 1.5 mm. The strips were dried on mesh
trays overnight in a forced draft oven at 55°C. They were then
broken into three pieces before feeding.
Diet Allocation
Each diet was randomly allocated to three digestibility tanks to
prmide three replicates per diet. Because there was only 18 tanks
in total, this meant that there were four separate collection periods.
Digestive Capacity of Abalone
173
Ahalone and Feeding
Juvenile greenllp and blacklip abaUme (shell length 40-60 mm)
were used in the experiments. The abalone had been obtained Irom
commercial hatcheries and raised on manufactured abalone feed.
The abalone were preconditioned for 1 week on the test diet as-
signed to their tank. During both the preconditioning and experi-
mental periods, the animals were fed lo excess every day at ap-
proximately 1700 h.
Tanks and Collcclion Syslem
Conical-shaped digestibility tanks were used. Abalone were
housed in 20-L buckets {approximately 80-100 per bucket) that
fitted inside the tanks. All the buckets were fitted with plastic mesh
bottoms ( 1 .3-cm x 1 .3-cm mesh ) allowing containment of the aba-
lone while permitting feces to drop into the collection tube at the
base of the tank. Three 25-cm lengths of PVC pipe (8 cm in
diameter) were placed in the buckets as shelters for the abalone.
Attached to the bottom of each digestibility tank was a screw-on
collection tube (11-cm long. 15-mm diameter). Tanks were on a
flow-through water system at a rate of about 2 L/min. The seawater
was filtered to 30 |jim by primary sand filters, then to 10 (xm by
secondary composite sand filters before entering the tanks. Aera-
tion was supplied at 0.5 L/min to each tank at all times by an air
stone. Water temperature and lighting were controlled during the
experiment with temperature maintained at I8.0°C ± 1.0 and a
light regime of 12-h light: 12-h dark. Salinity was 35-36SJ(
throughout the experiment.
Fecal Collection
Feces were collected by settlement every day until 5-6 g of
feces (dry weight) was collected for each replicate sample. This
took approximately 2 weeks. On each day of fecal collection the
buckets containing the abalone were removed and the digestibility
tanks were drained of water and all fittings were cleaned of feces
and uneaten feed. After cleaning, the tanks were refilled and the
buckets replaced. Collection tubes were fitted by 0900 h. A small
foam container was placed underneath each tube and filled with ice
to keep the collecting feces cold and reduce degradation by mi-
crobes. The feces were collected from the tubes at 1 630 h by gently
pouring the contents onto a 1-nim diameter mesh. The mesh was
then placed into a petri dish and frozen at -30°C. The following
day the frozen fecal sample was scraped off the mesh, pooled into
a composite sample, and stored in the freezer until required for
analysis. Before analysis, the samples were freeze-dried and
ground with a mortar and pestle.
Chemical Analyses
Gross energy was determined by a PaiT 1281 bomb calorimeter
(Parr Instrument Company, Moline, ID. Crude protein was deter-
mined by the combustion method using a LECO* CN-2000 Car-
bon and Nitrogen Analyser (RACI 1999).
Chromic oxide was determined using atomic absorption spec-
troscopy based on a modification of the methods described by
Hillebrand et al. (1953). The modified methodology involved pre-
liminary ignition of the sample at 500''C to remove organic ma-
terial and the dissolution of the sample in hydrochloric acid instead
of sulphuric acid (M. Frith, personal communication. University of
Tasmania. Launceston. Australia).
Digesliliility Delerminiilion
The apparent digestibilities of nutrients in the diets were cal-
culated using the following formula (Hardy 1997):
Apparent digestibility = 1
Cr^,.., X Nutriein,,
Cr,
X Niilrii'iil,,
where C, is chromium content and Niitriciil is nutrient or energy
content of the diet.
Statistical Analysis
The data were analyzed by analysis of variance using a gener-
alized linear model (SAS Institute Inc. 1988). Before analysis,
residuals were plotted to establish that the data were in fact nor-
mally distributed, which was the case. Within species treatment
means for nutrient digestibility of the twelve ingredients were
compared by least significant difference.
RESULTS
Significant differences were found between blacklip and green-
lip abalone in their apparent fecal digestibility of protein and en-
ergy of some of the ingredients evaluated (Table 3). Significant
differences in protein and energy digestibility were also found
among ingredients within each species (Table 3).
With respect to gross energy digestibility, blacklip abalone di-
gested the energy from whole L. aiii;iislifoliii.s. fishmeal. and skim
milk powder significantly better than greenlip abalone. and green-
lip abalone digested the energy from whey, bull kelp, and dehuUed
L. angustifoliiis significantly better than blacklip abalone (Table
3). No significant differences were found between the two species
in their ability to digest energy from semolina, defatted soyfiour.
casein, pregelatinized maize starch, mung beans, and L. luieiis
(Table 3).
Greater differences were found between the two species in their
capacity to digest protein from the ingredients with statistically
similar protein digestibility values only being obtained for mung
beans, whey and L. luteus (Table 3). Blacklip abalone digested
significantly more protein from defatted soyfiour, fishmeal. casein,
bull kelp, and skim milk than greenlip abalone. whereas greenlip
abalone digested significantly more protein than blacklip abalone
from semolina and dehulled and whole L. aiigustifolins (Table 3).
Comparisons among ingredients within species showed that
there were significant differences in their apparent protein and
energy digestibility for both species of abalone (Table 3). Whey
was the most digestible ingredient, having significantly higher
protein and energy digestibility than all other ingredients exaluated
for both blacklip and greenlip abalone I.P < 0.05). Bull kelp con-
tained the least-digestible protein for both species of abalone iP <
0.001), while semolina contained the least-digestible energy for
both species of abalone (P < 0.001 ).
DISCUSSION
The results from the current experiment demonstrate that black-
lip and greenlip abalone differ in their digestive capacity. Signifi-
cant differences were found in their ability to digest the protein and
energy from se\'eral ingredients.
With regard to protein digestibility it is interesting to note that
blacklip abalone can digest significantly more protein from, in
general, nonplant-derived proteins (excluding soyfiour and bull
174
Vandepeer and Van Barneveld
TABLE 3.
Comparison of the apparent faecal protein (PD) and energy (GED) digestibility coefficients obtained for 12 different ingredients fed to
blacklip and greenlip abalone.
PD
PD
GED
GED
Blacklip
Greenlip
Blacklip
Greenlip
Ingredient
Abalone
Abalone
Fu4
P
SEM
.Abalone
Abalone
f..4
P
SEM
Semolina
0.62''
0.84"
441
***
0.762
0.30''
0.34'
5.49
NS
1.265
Defatted soytlour
0.83'
0.82'
18.38
**
0.730
0.83"
0.78'
0.73
NS
1.507
Fishmeal
0.56'
0.46'
27.72
**
1.382
0.63^
0.52'
48.09
*
1.144
Casein
0.828
0.77"
27.42
**
0.624
0.79'=
0.78"
4.02
NS
0.579
Pregelled starch
-
-
-
-
-
0.92"
0.93"
1.80
NS
0.647
Mung beans
0.89''
0.9 1*"
5.13
NS
0.630
0.658
0.67'
2.40
NS
0.986
Bull kelp
0.46'
0.23'
105
»**
1.600
0.75'
o.sr
29.45
*
0.805
Whey
0.96"
0.95-"
1.46
NS
0.373
0.99''
LOO-*
43.20
*
0.106
Skim milk powder
0.94"
0.85'
510
***
0.286
0.95"
0.89"
1338
***
0.101
Lupin It
o.9r
0.91"
0.03
NS
0.804
0.79'
0.83'
2.83
NS
1.780
Lupin 2t
0.85=
0.92"
723
***
0.211
0.70'
0.82'
66.19
**
1.169
Lupin 3§
0.84"='
0.91"
371
***
0.284
0.63^
0.50'
202
:!=**
0.682
Within a species, superscripts have been used to identify Mgniticant differences among ingredients for their nutrient digestibility (within column
comparisons). Between species comparisons of nutrient digestion of each ingredient are made across rows and indicated by *.
NS, not significant
* P < 0.05
** P< 0.01
***/>< 0.001
"'8 Within a column, ingredient digestibility coefficients with different superscripts differ significantly (P < 0.05).
t Whole L luteiis.
± Dehulled L ungustifolius.
§ Whole L. anguslifolius.
kelp) than greenlip abalone. In contrast, greenlip abalone can di-
gest significantly more protein from plant-derived sources (lupins
and semolina) than blacklip abalone. This finding is in agreement
with that of Wee et al. (1994). who reported that blacklip abalone
digested significantly more protein than greenlip abalone from a
manufactured diet containing 50'7c fishmeal. It appears blacklip
abalone may not be able to digest the soluble nonstarch polysac-
charides found in terrestrial plants as efficiently as greenlip aba-
lone and that soluble nonstarch polysaccharides may actually in-
terfere with and reduce blacklip abalone's ability to digest nutri-
ents (both protein and energy I from plant feedstuff's which contain
them. As a consequence, use of exogenous enzymes that cleave
soluble nonstarch polysaccharides may improve the digestive ca-
pacity of blacklip abalone.
Dehulling had no effect on the digestibility of protein from L.
aiigHstifoHus when fed to blacklip abalone. Although a significant
increase was found in the digestibility of its energy for blacklip
abalone after dehulling it was much less than was found for green-
lip abalone (0.63 to 0.70 for blacklips compared with 0.50 to 0.83
for greenlips). After removal of the hull the energy from L. an-
guslifoliiis changed from being significantly less to significantly
more digestible for greenlip compared with blacklip abalone. The
hull of the lupin is composed primarily of cellulose. It appears that
blacklip abalone have a greater capacity to digest cellulose than
greenlip abalone given that the removal of the hull had a much
smaller effect on the capacity of blacklip abalone to digest energy
from this lupin compared with greenlip abalone.
Milk-based products (casein, skim milk powder, and whey) are
very digestible sources of protein and energy for both blacklip and
greenlip abalone. In particular, the sugar component of milk (lac-
tose) is very digestible for abalone given the extremely high gross
energy digestibility coefficients obtained for whey (the residue
from milk after removal of the casein and most of the fat). Lactose
is a disaccharide composed of galactose and glucose. Thus, it is a
much simpler carbohydrate than those found in many terrestrial
plant-based feedstuffs, such as lupins, which are composed of
complex structural and storage polysaccharides, p-galactosidase
(lactase) activity, needed for the hydrolysis of lactose, has been
found in abalone (Oshima 1931, Bennett et al. 1971). Obviously
P-galactosidase activity in wild abalone would not be for the di-
gestion of lactose, but probably for the breakdown of galactose,
one of the major components of carrageenan which is found in the
cell walls of red algae.
Pregelatinized waxy maize starch was also found to be a highly
digestible source of energy for both species of abalone. Again, this
is not surprising because the starch found in red algae, termed
floridean starch, is essentially the same as waxy starches found in
terrestrial plants in that it consists almost entirely of amylopeetin.
In addition Elyakova et al. (1981) found evidence for amylase-a-
1.4-glucanase activity against amylopeetin in extracts from the
hepatopancreas of W. asinina and H. vaiia. The fact that the starch
has been gelatinized, whereby the application of moist heat brings
about swelling and rupturing of the starch granules facilitating
amylolysis, would also increase energy digestibility.
The low protein digestibility of bull kelp by both species could
be caused by the presence of tannins, naturally occurring polyphe-
nols present in plants to protect them against herbivory. Their main
characteristic is that they bind and precipitate proteins. In vivo
studies have shown that protein digestibility is greatly reduced
when tanniniferous feeds are part of animal diets (Reed 1995).
Polyphenols are predominant in brown algae (Ragan & Glombitza
1986, Steinberg 1989). It should be pointed out that bull kelp has
Digestive Capacity of Abalone
175
a ver\' low crude protein content (69 g/kg) and that e\en though it
was included in the diet at a le\el of 500 g/kg the crude protein
content of the diet was onl\ 3.45 g/kg. Thus the endogenous N
contribution would ha\e had a much larger effect on the apparent
protein digestibility of kelp than for other ingredients, resulting in
these values being reduced as a result of an experimental artifact.
Neither species were able to digest the energy from semolina
very well, particularly blacklip abalone. In another study semolina
was found to affect the digestibility of other ingredients within a
diet (Vandepeer. unpublished data). The poor digestibility of
semolina and its effects on the digestibility of other ingredients is
a concern given that it is currently one of the major ingredients
used in manufactured diets in Australia. Further research is re-
quired lo establish the reasons why energy from semolina is so
poorly digested, however, it is possible that the starch component
significantly influences these results.
The results from this experiment demonstrate that greenlip and
blacklip abalone have different digesti\e capacities and thus a
different basis should be used for the formulation of manufactured
diets. Further comparisons of the nutritional requirements of
greenlip and blacklip abalone may also be justified.
ACKNOWLEDGMENTS
The authors would like to thank Dr. Ann Fleming for reviewing
and commenting on the manuscript. This research was funded by
a grant from the Fisheries Research and Development Corporation.
LITERATURE CITED
Bennett. R. Jr.. N. Thanassi & H. I. Nakada. 1971. Hepatopancreas gly-
cosidases of the abalone (Hcilioris rufescens). Comp. Biochem. Physiol.
406:807-811.
Coote, T. A. 1998. The protein, energy and lysine requirements of greenlip
abalone [Huliinis laevigata). Ph.D. Dissertation. University of Tasma-
nia. .Australia. 1 18 pp.
Ousel. G.. H. Kluge. K. Glaser, O. Simon, G. Harmann. J. v. Lengerken &
H. Jeroch. 1997. An investigation into the variability of extract viscos-
ity of wheat — relationship with the content of non-starch-
polysaccharide fractions and metaholisable energy for broiler chickens.
Arch. Anim. Niitr. 50:121-135.
Elyakova, L. A.. N. M. Shevchenko & S. M. Avaeva. 1981. A comparative
study of carbohydrase activities in marine invertebrates. Comp. Bio-
chem. Physiol. 69B:905-90S.
Fleming. A. E. 1995. Growth, intake, feed conversion efficiency and
chemosensory preference of the Australian abalone. Haliolis rubra.
.■Kqttaciilnire. 132:297-311.
Hardy, R. E. 1997. Understanding and using apparent digestibility coeffi-
cients in fish nutrition. Aquacultitre Mag. May/June:84-85.
Hillebrand, W. F., G. E. R Lundell. H. A. Bright & J. 1. Hoffman. 1953.
Applied inorganic analysis. New York: Wiley. 1034 pp.
Ikeda. K. & T. Kusano. 1983. In vitro inhibition of digestive enzymes by
indigestible polysaccharides. Cereal Chem. 60:260-263.
Mercer, J. P., K. S. Mai & J. Donlon. 1993. Comparative studies on the
nutrition of two species of abalone. Haliolis niberculala Linnaeus and
Haliolis discus hannai Ino I. Effects of algal diets on growth and
biochemical composition. Imerlebrale Reprod. Dew 23:75-88.
Cshima. K. 1931. Digestive enzymes appeared in abalone viscera. J. Ag-
ricult. Chem. 7:328-331.
Ragan. M. A. & K. W. Glombitza. 1986. Phlorotannins. brown algal
polyphenols. Prog. Phycol. Re.s. 4:129-241.
Reed. J. D. 1995. Nutritional toxicology of tannins and related polyphenols
in forage legumes. / Anim. Sci. 73:1516-1528.
Royal Australian Chemical Institute. 1999. Cereal Chemistry Division-
Official Methods. Dumas (combustion) total nitrogen determination.
Method No: 02-03.
SAS Institute Inc. 1988. SAS/STAT® Users guide. Release 6.03 Edition.
Cary. NC: SAS Institute Inc.. 1028 pp.
Schweiger. R. G. 1978. Carbohydrate Sulfates. A symposium sponsored by
the ACS Division of Carbohydrate Chemistry at the 174"' Meeting of
the American Chemical Society. Chicago. Illinois. August 30-31,
1977. ACS Symposium Series 77. pp. 214-243.
Shepherd. S. A. 1975. Distribution, habitat and feeding habits of abalone.
Aiisi. Fisheries 34:12-15.
Shepherd. S. A. & P. D. Steinberg. 1992. Food preferences of three Aus-
tralian abalone species with a review of the algal food of abalone. In:
S. A. Shepherd. M. J. Tegner & S. A. Guzman del Proo. editors. Aba-
lone of the world. Biology, fisheries and culture. Oxford: Blackwell
Scientific Publications, pp. 169-181.
Steinberg. P. D. 1989. Biogeographical variation in brown algal polyphe-
nolics and other secondary metabolites, comparison between temperate
Australasia and North America. Oecologia. 78:373-382.
Stewart. W. D. P. 1974. Algal physiology and biochemistry. In: W. D. P.
Stewart, editors. Oxford: Blackwell Scientific Publications, Osney
Mead. 989 pp.
Wee. K. L.. G. B. Maguire & S. Hindrum. 1994. Methodology for digest-
ibility studies with abalone 2. Comparison of markers in an artificial
diet for blacklip abalone (Haliolis rubra) and greenlip abalone (//.
laevigala). In: L. M. Chou. A. D. Munro. T. J. Lam, L. K. K. Cheong,
J. K. Ding, K. K. Hooi, H. W. Khoo. V. P. E. Phang, K. F Shim &
C. H. Tan. editors. Proceedings of the Third Asian Fisheries Forum.
Singapore. 26-30'" October. 1992. pp. 152-155.
JoiiiiKil ,>f Shellfish Research. Vol. 22. No. I. I77-IS4. 2()()_V
revip:vv of techniques to prevent introduction of zebra mussels
(dreissena polymorpha) during native mussel (unionoidea)
conservation activities
W. GREGORY COPE,'* TERESA J. NEWTON," AND CATHERINE M. GATENBY'
^ North Carolina State University, Department of Environmental and Molecular Toxicology, Box 7633.
Raleigh. North Carolina 27695; ^United States Geological Siiney. Upper Midwest Environmental
Sciences Center, 2630 Fanta Reed Road, La Crosse, Wisconsin 54603: Academy of Natural Sciences,
Patrick Center for Environmental Research. I'-MM) Ben Franklin Parkway. Philadelphia. Pennsylvania 19103
ABSTRACT Because ot the declines in diversity and abundance of native freshwater mussels (superl'amily Unionoidea). and the
potential decimation of populations of native mussels resulting from the rapid spread of the exotic zebra mussel Dreissena polymorpha.
management options to eliminate or reduce the threat of the zebra mussel are needed. Relocating native mussels to refugia (artificial
and natural) has been proposed to mitigate the threat of zebra inussels to native species. Relocation of native inussels to refugia such
as t~ish hatchery facilities or natural habitats within their historic range, which are unlikely to be infested by zebra mussels, necessitates
that protocols be developed to prevent the inadvertent introduction of zebra mussels. Several recent studies have developed such
protocols, and have assessed their effectiveness on the health and survival of native mussels during subsequent relocation to various
refugia. The purpose of this project is to synthesize and evaluate the current protocols and to develop a set of procedures that resource
managers and researchers should consider before conducting conservation activities in zebra mussel infested waters. We found that the
existmg protocols have many common points of concern, such as facility modification and suitability, zebra mussel risk assessment
and management procedures, and health and disease management procedures. These conservation protocols may have broad appli-
cability to other situations and locations. A summary and evaluation of the mformation in these main areas, along with recommended
guidelines, are presented in this article.
A'£)' WORDS: relocation, Unionidae, Dieissenu polymorphu. conservation, refugia
INTRODUCTION
Native freshwater mussels of the families Martiariliferiihw and
Unionidae (supeifamily Unionoidea) are one of the most rapidly
declining fauna! groups in North America. About 67% of the
nearly 300 native species found in North America are considered
vulnerable to extinction or already extinct (Bogan 1993, Williams
et al, 1993). The decline of native mussel populations in Noilh
America has occurred steadily since the mid 1 800s and has been
attributed to overharvest, construction of dams and impoundments,
sedimentation, navigation, pollution, and habitat degradation
(Fuller 1974, Bogan 1993, Naimo 199?, Brim Box & Mossa 1999,
Vaughn & Taylor 1999). An additional recent threat to the native
fauna has come from the introduction of the zebra tnussel Dreis-
sena piilyiiiorpha. This species colonizes native mussels and im-
pedes their movement, reduces the ability to feed and eliminate
wastes, and coinpetes for food and space ( Mackie 1 99 1 , .Schloesser
et al. 1996. Strayer 1999).
Because of the declines in diversity and abundance of native
mussels and the rapid and severe impacts of zebra inussels on
native mussels (Gillis & Mackie 1994. Nalepa et al. 1996). a
national strategy for the conservation of native freshwater mussels
was developed to provide a framework for preventing further
population declines and species extinction (National Native Mus-
sel Conservation Committee 1998). This document identified a
number of conservation needs and outlined goals, strategies, and
tasks to address these needs. Listed among these was the recom-
mendation to develop management options for eliminating or re-
ducing the threat of zebra mussels to native mussels. These options
included relocating native mussels to artificial and natural refugia.
Although tiiany mussel relocations have had poor success (e.g..
*Corresponding author. E-mail: greg_cope@ncsu.edu
Cope & Waller 1995), recent studies conducted with improved
techniques, experimental design, and monitoring programs, have
been successful (Dunn et al, 2000, Cope et al, 2003). Thus, with
the increased likelihood of successful relocation efforts, and the
continued range expansion and adverse effects of zebra mussels on
native tnussel populations, any relocation done to conserve native
mussels necessitates that protocols be developed to prevent the
inadvertent introduction of zebra mussels.
Several recent studies have developed protocols to ensure that
zebra mussels would not be inadveHently introduced during native
mussel conservation activities and have assessed the health and
survival of native mussels during subsequent relocation (Patterson
et al. 1997, Patterson et al, 1999. Gatenby et al, 2000, Nichols et
al. 2000. Hallac & Marsden 2001. Newton et al, 2001), The pur-
pose of this project was to synthesize and evaluate the current
protocols and to develop a set of procedures that resource manag-
ers and researchers should consider before conducting native tnus-
sel conservation activities in zebra mussel infested waters.
RESULTS AND DISCUSSION
Almost all of the recent native mussel salvage and relocation
projects have used some type of quarantine to prevent the inciden-
tal introduction of zebra mussels. The exceptions are those studies
intended to remove zebra mussels from fouled native mussels and
replace them back to their original location (e.g., Schloesser 1996,
Hallac & Marsden 2000), By necessity, most of the quarantine
protocols have been location and facility specific. For example,
Gatenby et al. (2000) reviewed procedures for relocating native
mussels from the Ohio River. Likewise, Newton et al, (2001)
developed a specific set of procedures for relocating native mus-
sels from the Mississippi River to artificial ponds and to fish
hatchery facilities. However, these and other protocols developed
for specific studies have many common points of concern, such as
177
178
Cope et al.
TABLE 1.
Summary of collection and quarantine-related conditions and procedures, and recommended guidelines for preventing introduction of zebra
mussels during native mussel conservation activities.
Condition or Procedure
Reference
Gatenbv et al. (2000)
Newton et al. (20011
Recommended Guidelines
Collection setting
Time of collection
July. September. October \W5 May 1W5
Species of native mussels
No. of native mussels
Native mussels analyzed for
disease and pathogens
before relocation
Air temperature (X)
Water temperature (°C)
Mechanism for removing
zebra mussels from native
mussels
Method for holding scrubbed
native mussels at collection
site
Emersion time (min) during
collection and processing
Transportation to quarantine
facility
Quarantine facility
Type
Mussel density (no./ni")
Water source
Water temperature ("O
Dissolved oxygen (mg/L)
pH
Potas.sium (mg/L)
Alkalinity (mg CaCOj/L)
Hardness (mg CaCO,/L)
Total ammonia nitrogen
(mg/L)
Unionized ammonia (|ji.g/L)
Total residual chlorine (p.g/L)
Nutrition/feeding
Amhiema plicata, Quadnila
ptistulosa, ElUptio
crassidens. Pleurobema
cordatum. Obliquaria
reftexu, Ponmnhis ulanis
27(»
No
20-28
Hand scrubbed vMth plastic-
bristled brushes
Mesh bacs in river*
20
Between moist burlap in
coolers with ice (no direct
contact of mussels and ice)
Above-ground tanks, l-t-500 L
Well water
LS0-2.'S()
2-28
6-14
7.2-8..'i
1.6
90
90
£1.0
2-66
■1 X 10'' cells/mL three times
per week in quarantine;
relocation ponds were
fertilized with a
nitrogen;phosphorous (N:P)
ratio of 10:1 (1.0 mg/L N.
0. 1 mg/L P) with NH^NO,
and NaHPOj salts
Early spring, before zebra mussel
spawning begins (water temperatures
<15°C) or mid to late fall when
natives have greater energy reserves
and juvenile zebra mussels are
visible (>2-5 mm shell length)
Amblema plicata, Fusconaia
flava, Leptodea fragilis,
Obliquaria reflexa. Quadrula
qiiailnihi
768
Yes
6-18
11-14
Hand scrubbed with plastic-
bristled brushes under x2
magnification
Hatchery truck with aerated
well water
Between moist burlap in
coolers with ice (no direct
contact of mussels and ice)
Pond (0.04 ha), mussels held in
8-2720 L mesh baas
If possible
Early spring or late fall temperatures;
minimize differences between air and
water temperature
Early spring or late fall temperatures;
minimize differences between air and
water temperature
Hand scrub with plastic-bristled brushes
under magnification
Hold in zebra mussel-free water after
scrubbing
Keep to minimum, but <20
Between moist burlap in coolers with
ice in plastic bags for transport
durations <12 h; no direct contact of
mussels and ice bags
-19-159
Keep
to
minimum, hut <1.50
water
Well
water
1.1-27
<28
6-20
>6
7.8-10.6
6.5-9.0
2.6
<4
110-160
>15
180-200
>50
0.03-0.2
<1.0
2-20
.3 g/m" of 10:10:10 N:P:K
fertilizer added to quarantine
pond 2 weeks prior to adding
unionids; relocanon ponds
were not fertilized
<25
<17
X 10' cells/mL or 4.0 mg dry wt./L
twice daily or 2.0-5.0 x lO"* cells/mL
or 1.9 mg dry wt./L on a continuous
basis (Gatenby 2000, 2002); suitable
algal species include Neochloris
oleoabimdans. Bracteacoccus
grandis. and Pliaeodactylum
tricormHum
continued on next page
Preventing Zebra Mussel Introduction
179
TABLE 1.
continued
Reference
Condition or Procedure
Gatenbv et al. (2(H)0)
Newton et al. (2001)
Recommended Guidelines
Da\s in quarantine Minimum of 30. but up In 120;
re-inspected under 4x
niagnirieation
Disinfection of equipment and Chlorine solution of 25 mg/L
supplies
Monitoring
Temperature, dissolved
oxygen, and pH
All other water quality
variables
Disease and inortalit\
Dessication for up to 4 d
Twice daily
Daily to weekly
Not specified
35; re-inspected under 2x
magnification
Not specified
Daily
Daily to weekly
Not specified
Minimum of 30; re-inspect under
magnification
Chlorine solution of 25-250 mg/L,
depending on type of material;
dessication in warm dry air for 3-5 d
At least daily
Daily to weekly
At least weeklv
' All native mussels were rinsed with a high pressure hose before being placed into the quarantine facility.
facility modification and suitability, zebra mussel risk assessment
and management procedures, and native mussel health and disease
management procedures, that may have broad applicability to
other situations and locations.
Facility-Specific Concerns and Procedures
The availability of aquatic facilities for long-term captive care
of freshwater mussels is limited. Thus, most of the salvage and
quarantine facilities have involved the short-term use of state and
US Government owned fish hatchery ponds and raceways or simi-
lar research aquaculture facilities (Dunn & Layzer 1997. Pinder et
al. 1999. Gatenbv 2000. Newton et al. 2001). The main facility
concerns have focused on the type of rearing or holding system
(e.g.. ponds, raceways, or above-ground tanks capable of housing
hundreds to thousands of mussels), the facility's proximity to the
source of relocated mussels (to reduce transportation time and
handling stress), on-site water quality for maintenance of mussel
health, and production of an algal-based food supply. The objec-
tives of any given conservation project will likely dictate the type
of facility or holding system used and any modifications that may
be required. Nonetheless, whether used for short-term quarantine
or for long-term captive care, all facilities should be able to pro-
vide space for isolation and quarantine, water quality characteris-
tics to meet requirements for shell growth and metabolic processes,
and food quantity and quality to support growth and reproduction
(Table 1).
Specific isolation and containment modifications are probably
necessary at most facilities to control and contain source water
inflow and potentially contaminated outflow. For example, the
outflow of water from quarantine units may need to be passed
through filtration or disinfectant treatments to remove or kill po-
tential zebra mussels before the water is discharged through nor-
mal routes. Containment procedures commonly used at facilities
conducting zebra mussel research have included filtration of out-
flow water through small mesh bags ( 100 (xm or smaller), chlorine
treatment tanks (230 mg/L for I h). and sand filtration units (J. J.
Rach, U.S. Geological Survey. Upper Midwest Environmental Sci-
ences Center. La Crosse, WI, pers. com.). Additional facility pre-
cautions may include the capping of all exterior drains to prevent
the release of potentially contaminated water from the affected
areas and the development of a flood risk assessment, if the facility
is within a designated floodplain.
The type of facility selected, however, may influence the rela-
tive success of the conservation project. Success could depend on
its use only as a short-term quarantine facility for subsequent re-
location to a natural or artificial system, or its use for long-term
captive care. For example. Newton et al. (2001) relocated five
species of native mussels (1,392 mussels total) from the Upper
Mississippi River to a fish hatchery pond after 35 d of quarantine
in an artificial pond (81% of mussels survived during quarantine).
Mussel survival in the hatchery pond averaged SO^c after 1 y. but
only 35% 3 y after relocation. Of the mussels in a handling-control
treatment that were placed back into the Mississippi River after
quarantine, survival was 80% after 1 y and 75% after 3.3 y. The
authors attributed the differences in survival between the hatchery
pond and riverine relocated mussels to inadequate nutritional re-
sources in the pond. This study illustrates the potential utility of
natural or managed refugia over artificial refugia for long-term
conservation (Nichols et al. 2000. Cope et al. 2003). Gatenby
(2000) observed similar decreases in survival of six large river
species relocated to pond refugia after a 30-d quarantine in above-
ground tanks. Mean survival of native mussels during quarantine
was 97%. Mean survival after 1 y in the ponds ranged between 82
and 93%. depending on species. Despite an abundance of a suit-
able algal food supply and adequate water quality conditions in the
ponds, however, the survival of relocated mussels decreased to
44%- after 2 y and to 5% after 3 y. Gatenby (2000) attributed the
mortality to high water temperatures in July and August during
years 2 and 3 of that study. Large river species of mussels relo-
cated (with no quarantine period) to fish hatchery raceways with
flowing water and sediment also showed high survival (95%) after
1 y (Dunn & Layzer 1997). but their long-term (3-5 y) success in
this type of system is unknown.
The relocation of native mussels after quarantine to natural
refugia or raceway systems supplied by natural river water will
likely have greater success for long-term preservation of the mus-
sels than retention in artificial pond refugia for two key reasons:
water temperature and food quality. These two cotnponents are
critical to the livelihood of any aquatic organism. Rapid fluctua-
tions in temperature, unnaturally high temperatures, and inad-
equate food supplies are known to cause stress in aquatic organ-
180
Cope et al.
isms, and can lead to mortality (Bayiie et al. 1973). Thus, tem-
peratute. food quality, and food quantity will also be key
components to the success of native mussel captive care programs.
Zebra Mussel Risk Assessment and Management Procedures
Because the threat of zebra mussels to native mussels has been
the primary causal factor for initiating most mussel conservation
activities, special precautions have been necessarily incorporated
into the collection and handling protocols where native mussels are
relocated. These precautions taken during collection, transport,
processing, and quarantine of native mussels are of utmost impor-
tance. Only the careful collection and handling of native mussels
from zebra mussel-infested waters will ensure that hatchery fish,
native mussels, and other aquatic species in the ecosystem are
protected from the incidental introduction of zebra mussels.
In situations where there is unceitainty in the co-existence of
zebra mussel populations in the watershed, the most prudent and
conservative approach is to treat all native mussels as if they
originated from zebra mussel-infested waters. A review of zebra
mussel range distribution and population dynamics in the particu-
lar river basin is also warranted. Particular items of interest in-
clude, the nearest known reproducing population of zebra mussels
to the native mussel collectiiin site, the relative density and poten-
tial spawning periods of zebra mussels at that site, and the likeli-
hood of an undetected presence at the native mussel collection site
(e.g.. lack of an active monitoring program).
The optimum time for collection of native mussels for a given
conservation project is largely unknown. Conservation projects,
however, should strive to select periods that reduce the stress
associated with handling as much as possible. Potential criteria
include choosing a period that coincides with the absence of zebra
inussel larvae in the water column, minimizes the temperature
differential between air and water, and does not inteiTupt the re-
productive cycle for most of the species being relocated. Zebra
mussel contamination can be minimized by collecting native mus-
sels during early spring or late fall periods when zebra mussel
larvae are likely not present in the water column (e.g., water tem-
peratures <15' C. Mackie 1991 ) or when the settled juveniles are of
a sufficient size to be easily seen (e.g.. 2-^ mm in shell length),
respectively. Freshwater mussels are categorized as either long-
term (bradytictic) or shon-term (tachytictic) brooders. Long-term
brooders, like many species of lampsilines and anodontines. be-
come gravid in late summer, retain the developing glochidia in the
gill marsupia throughout winter, and spawn in early spring (Mc-
Mahon & Bogan 2001 ). In contrast, short-term brooders, like many
species of amblemines. become gravid in early spring and spawn
in late summer (McMahon & Bogan 2001).
Newton et al. (2001) collected native mussels in early spring
when water temperatures ranged between 1 1 and 14°C. a period
before zebra mussel spawning, which generally occurs when water
temperatures reach 15 to 17°C (between May and June), in north-
em temperate regions of the United States and Canada (Mackie
1991 ). The collection of native mussels in early spring also has an
added potential benefit of reduced energetic stresses associated
with handling because of the cooler water temperatures (Jokela
1996, Newton et al. 2001). For example, glycogen concentrations
in Amhleina plicata were highest between May and July and
dropped precipitously thereafter — a pattern that closely paralleled
reproduction in this short-term brooder (Monroe & Newton 2001 ).
Similarly, Jokela et al. (1993) observed that glycogen concentra-
tions decreased substantially between July and October in An-
(uldiita pisciinilis. a long-term brooder. Ftirthennore. Jokela ( 1996)
suggested that transplanting females before fertilization or during
the early development of the brood had no detectable effect on
reproductive output.
Data on energetic reserves in marine bivalves contradict the
recently reported data in freshwater bivalves. In the marine envi-
ronment, it has been suggested that mussels collected in fall may
be able to better withstand handling stress because of their higher
energy reserves and because their metabolism is slowed by the
cooler water temperatures (Bayne et al. 1973). For example, by
mid to late fall, the marine species Mytihis edulis and M. trossulus
had accumulated abundant carbohydrate energy reserves (Hawkins
& Bayne 1985. Kreeger 1993, Kreeger et al. 1995). The differ-
ences between marine and freshwater species may be caused by
differing reproductive strategies. Results from a recent study with
native freshwater mussels, however, suggest that some species of
native mussels may build up their energy reserves in fall (Gatenby
2002). Obviously, this is an area where additional research is
needed.
When native mussels are collected from multiple sites in a
watershed with a known or suspected gradient in zebra mussel
density, working from the least infested site to the most infested
site will reduce potential zebra mussel contamination of boats and
other equipment. Optimally, boats used to collect or deploy native
mussels in zebra mussel infested areas should be cleaned (before
and after) by a high-pressure hot-water wash and diver wet suits,
supplies, and equipment (e.g.. ropes, buckets, etc.) used in the
study should be disinfected with a mild solution of chlorine bleach
(25 mg/L) or air dried (3-5 d) before use (Gatenby et al. 2000).
If the quarantine or relocation facility is also an operational fish
hatchery or aquaculture center, precautionary measures to protect
endemic wild species and cultured fish species should be consid-
ered. Before entrance into the facility, a subsample of native mus-
sels should be obtained from the collection site and submitted to a
United States Fish and Wildlife Service. National Fish Health
Center (Newton et al. 2001 ) or similar laboratory, to assess poten-
tial disease and pathogen presence (see section later on native
mussel health and disease management procedures).
After screening for diseases and pathogens, collection of native
mussels should proceed with procedures to minimize contamina-
tion from adult and larval zebra mussels. These include scrubbing
individual native mussels with plastic bristled brushes, visual in-
spection of all exterior surfaces of the shell with magnifying
lenses, and holding cleaned natives in zebra mussel-free water
(Table 1 ). Care should be taken during scrubbing and inspection to
avoid overlooking small zebra mussels that may be attached in
crevices, in areas of shell erosion (native mussels with severely
eroded or damaged valves should be discarded), or along the hinge
line (Gatenby et al. 2000, Newton et al. 2001). Only personnel
experienced in mussel biology should conduct the inspections to
ensure accuracy and efficiency of these procedures.
During collection and processing of native mussels, emersion
(exposure to air) and thermal stress should be kept to a minimum.
Recent studies have shown that handling mussels over a range of
emersion air temperatures ( 15-35°C) and emersion durations ( 15-
60 min) did not acutely impair survival, behavior, or biochemical
composition (Bartsch et al. 2000, Greseth et al. 2003). A minimal
emersion time (<20 min). however, is generally recommended
from recent efforts (Table 1 ). Moreover, water temperature and
Preventing Zebra Mussel Introduction
181
dissolved oxygen concentrations in the holding \esscls during col-
lection should be measured frequently (at least once per hour) and
maintained at or near (±2 'O the ambient stream conditions at the
time of collection with non-chlorinated ice and external aeration, if
possible (Gatenby et al. 2000).
Depending on the proximity of the native mussel collection site
to the quarantine facility (a transport time generally <12 h). mus-
sels should be transported in coolers covered with moist burlap and
kept cool (within ±2°C of the water collection temperature, if
possible) w ith ice in plastic bags without direct contact of ice bags
and mussels (Gatenby et al. 2000. Newton et al. 2001. Cope et al.
200.^). This method is advantageous over the use of water-filled,
aerated tanks (Chen et al. 2001) because of the reduced need for
costly and cumbersome trucks and equipment and of miniinizing
potential problems associated with maintaining stable dissolved
oxygen concentrations in water during transport.
At the quarantine facility, native mussels have generally been
held for a minimum of 30-35 d (Gatenby et al. 2000, Newton et al.
200 1 ) to allow any small or previously undetected zebra mussels to
become visually apparent on re-inspection. The 30-35 d quaran-
tine period is based on reported zebra mussel growth rates of
0.06-0.15 mm/d (Mackie 1991. Martel 1995. Chase & Bailey
1999), which would allow a newly settled zebra mussel to reach a
visible shell length of about 2-5 mm during quarantine. During
this time, basic water quality measurements (e.g., temperature,
dissolved oxygen, and pH) should be taken at least daily. Other
water chemistry variables such as alkalinity, hardness, potassium,
total ainnionia nitrogen (TAN), and unionized ammonia should be
measured at least weekly to ensure that water quality conditions
for minimum life requirements are met (Table 1 ). In addition,
mussels in quarantine should be monitored at least weekly for
disease (see section below on native mussel health and disease
management procedures) and mortality.
Isolation of native mussels from other aquatic species, their
contact water, nets, or other equipment at the quarantine facility is
necessary to protect organismal health and the physical facility.
These concerns can largely be addressed by applying standard best
practices for maintaining fish health. Disinfection of equipment
and supplies for native mussel quarantine should be guided by
National Fish Health Policy and Procedures, Part 713, sections
FWI and FW 3 (USFWS 1995): chlorine (200-250 mg/L for 1 h),
.sodium or potassium salts (saturated solutions) or other chemical
treatments (e.g., benzalkonium chloride at 100 mg/L for 3 h) and
desiccation (3-5 d) have been successfully used or recommended
(Reid et al. 1993, Waller et al. 1996. Gatenby et al. 2()()()).
After the minimum quarantine period (30-35 d). individual
mussels are thoroughly re-inspected by hand with magnifying
lenses to evaluate the presence of zebra mussels. If zebra mussels
are not found, the mussels are deemed zebra mussel-free and can
be relocated elsewhere (e.g.. to natural or artificial systems or to
other facilities for long-term captive care). Because no zebra mus-
sels were found after quarantine in the study of Newton et al.
(2001). the mussels were subsequently relocated to fish hatchery
ponds. In contrast. Gatenby et al. (2000) found zebra mussels on
initial re-inspection and consequently held native mussels in quar-
antine for additional 30 d intervals each time zebra mussels were
found, up to a total of 120 d. Because of declines in mussel health
and condition over time during quarantine (Patterson et al. 1997.
Newton et al. 2001). Gatenby et al. (2000) recommended re-
inspection of mussels at 7 d intervals after the initial 30 d period
when zebra mussels are found, and to hold them onlv for 30
additional days after the last zebra mussel is found, to shorten the
overall quarantine time. However, the added stress of handling
native mussels more frequently must be weighed against the prob-
ability of earlier detection of zebra mussels.
Additionally, native mussels could be treated with chemical
disinfectants. Certainly, the benefit of this type of treatment must
be weighed against the risk of added stress and reduced fitness in
the native mussels, but a study by Waller and Fisher (1998) found
that limited application of specific chemicals (e.g., 20,000 mg
NaCl/L for 6 h) may be feasible for certain tolerant native species.
They cautioned, however, that chemical disinfectants cannot guar-
antee the elimination of all zebra mussels from native mussel
shells and stated that pre-treatnient or multiple treatment (e.g.,
once per week) of native mussels and their holding tanks may be
most valuable for reducing the time held in quarantine. Many fish
hatchery and aquaculture facilities may already be using various
chemical treatments (Waller et al. 1996. Edwards et al. 2000.
Edwards et al. 2002) or hazard analysis protocols such as the
Aquatic Nuisance Species-Hazard Analysis Critical Control Point
(ANS-HACCP) approach (Gunderson & Kinnunen 2001) to pre-
vent the spread of zebra mussels and other aquatic nuisance spe-
cies during their activities, which may be adapted to the collection,
transport, and quarantine of native mussels.
,\ative Mussel Health and Disease Management Procedures
Although liltle is known about the diseases of native freshwater
mussels, recent studies have shown the potential for pathogen
transmission among native mussels and fish (Starliper et al. 1998,
Starliper & Morrison 2000). The primary concern for fish hatchery
or aquaculture facilities that contain native mussels is the potential
for transmission of disea.se and pathogens between host mussels
and hatchery fish. Transmissions from hatchery fish to mussels and
from mussel to mussel are also important vectors to control for
maintaining mussel health. Therefore, a pathogen and disease
monitoring plan for native mussels, similar to that commonly used
for hatchery-reared fish, should be considered. Hatchery personnel
are routinely trained in fish health protocols and record keeping:
these procedures could easily be adapted for monitoring mussel
health. The United States Government standards and protocols
currently exist for a disease control and classification system for
coldwater fish (salmonid) pathogens — similar guidelines for
warmwater fish or native mussels do not exist (USFWS 1995).
Revisions to the United States Fish and Wildlife Service, Fish
Health Policies and Procedures are currently underway to include
warmwater fish and other aquatic organisms (Richard Nelson,
United States Fish and Wildlife Service, La Crosse Fish Health
Center, Onalaska, Wl, pers. com.). Until those changes are imple-
mented, however, native mussels may only be screened in the near
term for reportable coldwater pathogens and diseases. On a posi-
tive note, a recent study evaluating the effect of depuration on the
transmission of the bacterial fish pathogen Aeromonas salmoni-
cicla (the causative agent offish furunculosis) between the unionid
Anihic'ina plicata and two strains of Arctic char Scilveliniis alpinus
found that the minimum 3()-d quarantine of native mussels recom-
mended for preventing the spread of zebra mussels was sufficient
for depuration of the fish pathogen and eliminating transmission of
the disease (Starliper 2001 ). Therefore, when adequate safeguards
and standard best practices for fish health are used in combination
with a 30-d quarantine, disease and pathogen transmission risks
should be minimal. Native mussels held in quarantine should be
182
Cope et al.
screened before being placed in tlie quarantine facility and moni-
tored monthly throughout the duration of their captive care to
document disease and pathogen incidence and history. More re-
search and policy development is needed in this area to ensure
protection of fish and native mussels.
Maintaining the physiologic condition of native mussels during
quarantine is difficult because diet and nutritional requirements are
poorly understood. Although the specific time course for changes
in biochemical indices of mussels caused by quarantine is un-
known, recent studies have shown that substantial decreases in
glycogen concentrations occur in as little as 7-35 d after quaran-
tine. For example. Patterson et al. (1997) found that glycogen
concentrations in mantle tissue in Amhieinii plicata and Quadriila
pustidosii dropped significantly after 7 d in quarantine and by day
30. concentrations had declined to only 15-31% of that measured
in wild-caught specimens. Likewise, glycogen concentrations in
foot tissue of A. plicata decreased 44% from 279 ± 191 mg/g dry
weight at day to 178 ± 105 nig/g dry weight after 35 days in
quarantine (Newton et al. 2001 1.
Based on the poor physiologic condition of native mussels after
quarantine shown by previous studies, it is critical to provide the
best source of nutrition during quarantine. Previous studies have
relied on an algal-based diet, either produced //; situ by stimulating
algal growth with fertilizers in ponds or cultured indoors on site
and added directly to mussel holding tanks (Gatenby et al. 1997,
Patterson et al. 1997. 1999, Gatenby 2000, Gatenby et al. 2000,
Newton et al. 2001 ). A number of algae have been tested as food
for juvenile and adult mussels (Gatenby et al. 1997, Gatenby 2000,
Beck 2001). Recent biochemical analysis of three algae (Neochlo-
ris pleoahmulans. Bnuteacticciis gnmdis. and Phacodactyliiiu lii-
ainuttuiii) indicate that these could be nutritionally suitable for
maintaining freshwater mussels in captivity (Gatenby et al. 2002).
If mussels are to be quarantined or relocated to ponds, the follow-
ing should be kept in mind: ( 1 ) standard commercial pond fertil-
izers should not be used to stimulate growth of algae; (2) the
potassium levels in commercial fertilizers are toxic to freshwater
mussels (Imlayl973); (3) the nitrogeniphosphorous ratio (N:P) of
the standard 10:10:10 nitrogen:phosphorous:potassiuni (N:P:K)
fertilizer will not promote suitable algae for mussels that typically
require an N:P ratio of 10:1 (McCombie 1953); and (4) an unsuit-
able, or indigestible filamentous blue-green algal bloom will result
when 10: 10: 10 N:P:K is used. Therefore, we recommend using the
fertilizers indicated in Table I, following Gatenby et al. (2000).
Although feeding requirements for native mussels will likely de-
pend on the species involved, temperature conditions, and meta-
bolic activity, Gatenby et al. (2000) recommended that native mus-
sels be fed 1 x 10'' cells/niL or 4.0 mg dry weight/L twice daily
(Table 1 ). This was a conservatively high recommendation based
on initial feeding studies and assimilation efficiencies. This con-
centration resulted in the greatest assimilation of organic carbon,
but a significant amount of this ration went unused by the animals
(Gatenby 2000). More recent data indicate that a diet ration of
2.0-5.0 X lO'* cells/niL or 1.9 mg dry weight/L per feeding cham-
ber should maintain mussel condition during summer growth pe-
riods (Gatenby 2002). Particle concentrations should be monitored
and not allowed to drop below 60% of this recommended ration.
Feeding frequency will depend on the species and total biomass
being held in captivity (Gatenby 2002). Thus, monitoring the par-
ticle concentration on a daily basis is necessary. Initially, particle
concentration may need to be monitored two to three times daily
until the manager is familiar with the particle depletion rate or
clearance rate of the native mussels held in captivity.
CONCLUSIONS AND RECOMMENDATIONS
Native freshwater mussels should only be relocated from ex-
isting areas as a la.st resort (Cosgrove & Hastie 2001 ). Other op-
tions to relocation and salvage, such as periodic cleaning of zebra
mussels from native mussels and replacement (Hallac & Marsden
2000, Hallac & Marsden 2001 ), and the use of natural or managed
refugia (Nichols et al. 2000), should be considered as first alter-
natives where practical. For example, Hallac & Marsden (2000,
2001 ) suggested that periodic cleaning and replacement might be
a viable option for conservation of native mussels, especially in
areas where food is not limiting and where collection and cleaning
are logistically feasible. If, however, freshwater mussel relocations
are required to conserve localized populations from zebra mussels
or other catastrophic events, the concerns and procedures de-
scribed in this article should provide general guidance for devel-
oping plans to prevent the incidental introduction of zebra mussels
during these activities and for maintaining the health of the native
refugees while under captive care.
In addition, procedures for ensuring long-term viability of na-
tive mussel populations need to be considered throughout the plan-
ning and Implementation process. For example, similarities in wa-
ter quality, substratum characteristics, food, and necessary fish
hosts among the systems are critical elements in a native mussel
relocation strategy. Additional ecological and evolutionary con-
cerns, such as retention of genetic diversity of the mussel popula-
tions, need to be carefully considered before relocating native
mussels to natural refugia, especially if the mussels are to be
relocated between river basins or between sub-basins of the same
river system (Villella et al. 1998, Storfer 1999).
Because of costs and limited availability of facilities for quar-
antine and captive care of native mussels, the United States Fish
and Wildlife Service and its resource conservation and manage-
ment partners may wish to designate several facilities within re-
gions of the United States that can accept, hold, and screen mussels
for disease and pathogens. These facilities may include state or
national fish hatcheries, research or aquaculture centers, and fish
health centers.
To our knowledge, this synthesis represents the "state-of-the-
science"" for minimizing the incidental introduction of zebra mus-
sels during native mussel conservation activities and for ensuring
their short-term and long-term health and viability. Readers of this
article should be cautioned that the information presented is only
recommended guidelines and that future improvements to proce-
dures will be made through research and policy development.
ACKNOWLEDGMENTS
This project was funded by the United Stales Fish and Wildlife
Service, through a contract with the Freshwater Mollusk Conser-
vation Society. Linda Drees and Tina Proctor provided valuable
insight on the relevance of the project to resource managers. Steve
Ahlstedt, Arthur Bogan, Heidi Dunn, Jerry Fairis. Doug Jensen,
Patricia Morrison, Pam Thiel, and Kurt Weike provided informa-
tion critical to preparation of the document. The authors thank
Robert Anderson, Heidi Dunn, Richard Neves, Jeixine Nichols.
Tom Watters. and Kurt WeIke for reviewing a draft of the docu-
ment.
Preventing Zebra Mussel Introduction
183
LITERATURE CITED
Barlsch. M. R. D. L. Waller. W. G. Cope & S. Giitreuter. 2000. Emersion
and thermal tolerances of three species ofunionid nuisscK: survival and
behavioral effects. / Shellfish Res. 19:233-:4().
Bayne. B. L., R. J. Thompson & J. Widdows. 1973. Some effects of
temperature and food on the rate of oxygen consumption by Mylilus
ediilis. In: W. Wieser. editor. Effects of temperature in ectothermic
organisms. Berlin: Springer-Verlag. pp. lSl-193.
Beck, K. M. 2001. Development of an algal diet for rearing juvenile fresh-
water mussels (MSc thesis). Blacksburg: Virginia Polytechnic Institute
and State University. 58 pp.
Bogan, A. E. 1993. Freshwater bivalve extinctions (Mollusca: Unionoidal:
a search for causes. Am. Zool. 33:599-609.
Brim Box. J. & J. Mossa. 1999. Sediment, land use, and freshwater mus-
sels: prospects and problems. J. N. Ant. Benrhol. Soc. 18:99-1 17.
Chase. M. E. & R. C. Bailey. 1999. The ecology of the zebra mussel
{Dieissena polymorpha) in the lower Great Lakes of North America: I.
Population dynamics and growth. / Great Lukes Res. 25:107-121.
Chen, L. Y.. A. G. Heath & R. Neves. 2001. An evaluation of air and water
transport of freshwater mussels (Bivalvia: Unionidae). .4m, Malacoi
Bull 16:147-154.
Cope. W. G. & D. L. Waller. 1995. Evaluation of freshwater mussel
relocation as a conservation and management strategy. Reg. Rivers Res.
Manage. 11:147-155.
Cope. W. G., M. C. Hove. D. L. Waller, D. J. Hornbach. M. R. Bartsch. L.
A. Cunningham. H. L. Dunn & A. R. Kapuscinski. 2003. Evaluation of
relocation of unionid mussels to in situ refugia. ./. Molluscan Stud.
69:27-34.
Cosgrove. P. J. & L. C. Hastie. 2001. Conservation of threatened fresh-
water pearl mussel populations: river management, mussel transloca-
tion and conflict resolution. Biol. Cons. 99:183-190.
Dunn. C. S. & J. B. Layzer. 1997. Evaluation of various holding facilities
for maintaining freshwater mussels in captivity. In: K. S. Cummings,
A, C. Buchanan, C. A. Mayer & T. J. Naimo, editors. Conservation and
management of freshwater mussels II: initiatives for the future. Pro-
ceedings of a UMRCC Symposium. 16-18 October 1995. St. Louis.
Missouri. Upper Mississippi River Conservation Committee, Rock Is-
land, Illinois, pp. 205-213.
Dunn. H, L.. B. E. Sietman & D. E. Kelner. 2000. Evaluation of recent
Unionid (Bivalvia) relocations and suggestions for future relocations
and reintroductions. In: R. A. Tankersley. D. I. Warmolts. G. T. Wat-
ters. B. J. Armitage. P. D. Johnson & R. S. Butler, editors. Freshwater
Mollusk Symposia Proceedings. Columbus. OH: Ohio Biological Sur-
vey, pp. 169-183.
Edwards. W. J.. L. Babcock-Jackson & D. A. Culver. 2000. Prevention of
the spread of zebra mussels during fish hatchery and aquaculture ac-
tivities. N. Am. J. Aquaculture. 62:229-236.
Edwards, W. J., L. Babcock-Jackson & D. A. Culver. 2002. Field testing
of protocols to prevent the spread of zebra mussels Dreissena poly-
morpha during fish hatchery and aquaculture activities. N. Am. J. Aqua-
culture. 64:220-223.
Fuller. S. L. H. 1974. Clams and mussels (Mollusca: Bivalvia). In: C. W.
Hart. Jr. & S. L. H. Fuller, editors. Pollution Ecology of Freshwater
Invertebrates. New York: Academic Press, pp. 215-273.
Gatenby, C. M. 2002. Seasonal changes in the biochemical and physiologi-
cal condition of native freshwater mussels (final report). Cincinnati:
Mussel Mitigation Trust Fund. 26 pp.
Gatenby, C. M. 2000. A study of holding conditions, feed ration, and algal
foods for the captive care of freshwater mussels (PhD dissertation).
Blacksburg: Virginia Polytechnic Institute and State University.
237 pp.
Gatenhy. C. M., B. C. Parker & R. J. Neves. 1997. Growth and survival of
juvenile rainbow mussel, Villosa iris (Lea 1829) (Bivalvia: Unionidae)
reared on algal diets and sediment. Am. Malacoi. Bull. 14:57-66.
Gatenby, C, M., D. M. Orcutt, D. A. Kreeger. B. C. Parker. V. A. Jones &
R. J. Neves. 2002. Biochemical composition of three algal species
proposed as food for capli\'e for freshwater mussels. J. Appl. Phycol.
15:1-11.
Gatenby. C. M., P. A. Morrison. R. J. Neves & B. C. Parker. 2000. A
protocol for the salvage and quarantine of unionid mussels from zebra
mussel-infested waters. In: R. A. Tankersley. D. 1. Warmolts. G. T.
Watters, B, J. Armitage, P. D. Johnson & R. S. Butler, editors. Fresh-
water Mollusk Symposia Proceedings, Ohio Biological Survey. Colum-
bus. OH: pp. 9-18.
Gillis. P. L. & G. L. Mackie. 1994. Impact of the zebra mussel. Dreissena
polymorpha, on populations of Unionidae (Bivalvia) in Lake St. Clair.
Can. J. Zool. 72:1260-1271.
Greseth. S. L.. W. G. Cope. R. G., Rada, D. L. Waller & M. R. Bartsch.
2003. Biochemical composition of three species of unionid mussels
after emersion. / Molluscan Stud. 69:101-106.
Gunderson. J. L. & R. E. Kinnunen. 2001. Aquatic nuisance species-
hazard analysis and critical control point training curriculum. Minne-
sota Sea Grant Publication No. MNSG-Fl 1 and Michigan Sea Grant
Publication No. MSG-00-400. 78 pp.
Hallac. D. E. & J. E. Marsden. 2000. Differences in tolerance to and
recovery from zebra mussel [Dreissena polymorpha) fouling by Ellip-
tio complanala and Lampsilis radiata. Can. J. Zool. 78:161-166.
Hallac, D. E. & J. E. Marsden. 2001 . Comparison of conservation strategies
for unionids threatened by zebra mussels (Dreissena polymorpha): pe-
riodic cleaning vs quarantine and translocation. / N. Am. Benthol. Soc.
20:200-210.
Hawkins. A. J. S. & B. L. Bayne. 1985. Seasonal variation in the relative
utilization of carbon and nitrogen by the mussel Mytilus edulis budgets,
conversion efficiencies and maintenance requirements. Mar. Ecol.
Prog. Ser. 25:181-188.
Imlay, M. J. 1973. Effects of potassium on survival and distribution of
freshwater mussels. Malacologia 12:97-1 13.
Jokela. J. 1996. Within-season reproductive and somatic energy allocation
in a freshwater clam. .Anodonia piscinalis. Oecologia 105:167-174.
Jokela. J., L. Uotila & J. Taskinen. 1993. Effect of castrating trematode
parasite Rhipidocotyle fennica on energy allocation of freshwater clam
Anodoma piscinalis. Fund. Ecol. 7:332-338.
Kreeger, D. A. 1993. Seasonal patterns in utilization of dietary protein by
the mussel Mytilus trossulus. Mar. Ecol. Prog. Ser. 95:215-232.
Kreeger, D. A.. A. J. S. Hawkins. B. L. Bayne & D. M. Lowe. 1995.
Seasonal variation in the relative utilization of dietary protein for en-
ergy and biosynthesis by the mussel Mytilus edulis. Mar. Ecol. Prog.
Ser. 126:177-184.
Mackie. G. L. 1991. Biology of the exotic zebra mussel. Dreissena poly-
morpha, in relation to native bivalves and its potential impact in Lake
St. Clair. Hydrobiologia 219:251-268.
Marlel. A. 1995. Deinography and growth of the exotic zebra mussel
[Dreissena polymtnpha) in the Rideaii River. Ccui. J. Zool. 73:2244—
2250.
McCombie. A. M. 1953. Factors influencing the growth of phytoplankton.
J. Fish. Res. Board Can. 10:253-282.
McMahon. R. F. & A. E. Bogan. 2001. Mollusca: Bivalvia. In: J. H. Thorpe
& A. P. Covich. editors. Ecology and classification of North American
freshwater invertebrates. New York: Academic Press, pp. 331—129.
Monroe. E. M. & T. J. Newton. 2001. Seasonal variation in physiological
condition of Amblema plicata in the Upper Mississippi River. J. Shell-
fish Res. 20:1167-1171.
Naimo. T. J. 1995. A review of the effects of heavy metals on freshwater
mussels. Ecotoxicol, 4:341-362.
Nalepa. T. F.. D. J. Hartson. G. W, Gostenik. D. L. Fanslow & G. A. Lang.
1996. Changes in the freshwater mussel community of Lake St. Clair:
from unionidae to Dreissena polymorpha in eight years. J. Great Lakes
R«. 22:354-369.
National Nafive Mussel Conservation Committee. 1998. National strategy
for the conservation of native freshwater mussels. J. Shellfish Res.
:7:I4I9-1428,
184
Cope et al.
Newton, T. J.. E. M. Monroe. R. Kenyon, S. Gutreuler, K. I. Welke & P.
A. Thiel. 2001. Evaluation of relocation of unionid mussels into arti-
ficial ponds. J. N. Am. Benlhol. Soc. 20:468-485.
Nichols. S. J., M. G. Black & J. D. Allen. 2000. U.se of on-site refugia to
protect unionid populations from zebra mussel-induced mortality. In:
R. A. Tankersley. D. 1. Warmolts, G. T. Watters. B. J. Arniitage. P. D.
Johnson & R. S. Butler, editors. Freshwater Mollusk Symposia Pro-
ceedings. Columbus. OH: Ohio Biological Survey, pp. 67-75.
Patterson. M. A., B. C. Parker & R. J. Neves. 1997. Effects of quarantine
times on glycogen levels of native freshwater mussels (Bivalvia:
Unionidae) previously infested with zebra mussels. Am. Malacol. Bull.
14:75-79.
Patterson. M. A.. B. C. Parker & R. J. Neves. 1999. Glycogen concentra-
tion in the mantle tissue of freshwater mussels (Bivalvia: Unionidae)
during starvation and controlled feeding. Am. Mulucol. Bull. 15:47-50.
Pinder. M. J.. M. A. McGregor. C. D. Slickley & J. J. Ferraro. 1999. The
use of state fish hatchery for the cultivation of freshwater mussels. In:
P. D. Johnson, R. S. Butler & G. Benz, editors. Musseling in on Biodi-
versity. Proceedings of a Freshwater Mollusk Conservation Society
Symposium, 17-19 March 1999. Chattanooga, TN. 36 pp.
Reid. D. F.. J. Bidwell. J. Carlton. L. Johnson. E. Marsden & S. J. Nichols.
1993. Zebra-mussel specific contaminant protocols. National Oceanic
and Atmospheric Administration. Contribution No. 890. 28 pp.
Schloesser. D. W. 1996. Mitigation of unionid mortality caused by zebra
mussel infestation: cleaning of unionids. N. .Am. J. Fish. Manage. 16:
942-946.
Schloesser, D. W.. T. F. Nalepa & G. L. Mackie. 1996. Zebra mussel
infestation of unionid bivalves (Unionidae) in North America. Am.
Zool. 36:300-310.
Slarliper. C. E. 2001. The effect of depuration on transmission of Aero-
monas salmonicida between the freshwater bivalve .Amhiciiui plicala
and Arctic char. J. Aqual. Animal Health. 13:56-62.
Starliper, C. E. & P. Morrison. 2000. Bacterial pathogen contagion studies
among freshwater bivalves and salmonid fishes. J. Shellfish Res. 19:
251-258.
Starliper, C. E.. R. Villella. P. Monison & J. Mathias. 1998. Studies on the
bacterial flora of native fieshwater bivalves from the Ohio River.
Biomed. Let. 58:85-95.
Storfer. A. 1999. Gene flow and endangered species translocations: a topic
revisited. Biol. Cons. 87:173-180.
Strayer. D. L. 1999. Effects of alien species on freshwater mollusks in
North America. J. N. Am. Benthol. Soc. 18:74-98.
USFWS. 1995. United States Fish and Wildlife Service. Service Manual
(700 Series), Part 713, Fish Health. Washington. DC.
Vaughn. C. C. & C. M. Taylor. 1999. Impoundments and the decline of
freshwater mussels: a case study of an extinction gradient. Cons. Biol.
13:912-920.
Villella. R. F.. T. L. King & C. E. Starliper. 1998. Ecological and evolu-
tionary concents in freshwater bivalve relocation programs. J. Shellfish
Res. 17:1407-1413.
Waller. D. L. & S. W. Fisher. 1998. Evaluation of several chemical dis-
infectants for removing zebra mussels from unionid mussels. Proi;.
Fish-Ctdt. 60:307-310.
Waller, D. L.. S. W. Fisher & H. Dabrowska. 1996. Prevention of zebra
mussel infestation and dispersal during aquaculture operations. Prog.
Fish Cidt. 58:77-84.
Williams. J. D.. M. L. Wan-en, Jr.. K. S. Cummings, J. L. Hanis & R. J.
Neves. 1993. Conservation status of freshwater mussels of the United
States and Canada. Fisheries 18:6-22.
Joiinial of Slwllfish Rcsi-anli. Vol. 2:, No. I. 1X5-192. 2()().V
A COMPARISON OF THE PARASITE AND SYMBIONT FAUNA OF COHABITING NATIVE
{PROTOTHACA STAMINEA) AND INTRODUCED (VENERUPIS PHILIPPINARVM AND
NUTTALIA OBSCURATA) CLAMS IN BRITISH COLUMBIA
W. L. MARSHALL, S. M. BOWER,* AND G. R. MEYER
Fisheries and Oceans Canada. Biological Sciences Brancli Pacific Biological Slalion Nanainm.
British Columbia. Canada. V9T 6N7
ABSTRACT Native littleneck clams iPronnlnii a sitiiiiincti). Manila clams {Vcncnipls pliilippiminim. inadvertently introduced in the
iy3().s), and varnish clams (NtittaUiu obiciiiaia, inadvertently introduced in the 1980s and lyyOs) were collected from the same
microsite at two different locations and examined for parasites and symbionts using histology and light microscopy. Varnish clams are
currently being assessed for their long-term fisheries potential but there is little knowledge of their parasite and symbiont fauna. This
study initiates the documentation of parasites and symbionts of varnish clams and adds to the continuing documentation of organisms
found within native littleneck clams and Manila clams. Host exposure to potential parasites and symbionts that were prevalent in at
least one of the clam species was assumed lo be similar for all clams due to their close proximity. This close association in the natural
environment allowed for the comparison of host specificity and response of the clams to multiple invasive species. All three of the clam
species had a different assemblage of parasites and this pattern was mostly consistent for both sites. Host preferences of each type of
parasite or symbiont v\'ere also consistent between sites and they were often restricted to a single host species. The most common
parasites of varnish clams were Nemaropsi.s-Wki! spores, pea crabs (Pinnixa fciha) and parasitic copepods (Mylilicolu sp.) and less
frequently a turbellarian inhabiting the kidney tubule. An undocumented eimeriorin-like kidney coccidian was found in 4% of Manila
clams and two previously undescribed inclusions bodies were found in native littleneck clams at low frequencies.
KEY WORDS: hixalve. Pniiotliaca suimiiwu. Vciicnipis philippiiuiniiu. Nuttullia (ihscitrala. parasites, symbionts
INTRODUCTION
In .Itnic of 2002 three species of clams (one native and two
introduced) were chosen for a survey of parasites and symbionts.
The native littleneck cluni \Prounhaca staminea: (Conrad 18.^7);
= Paphia suiininca. = Venus staininea] was the most important
fresh-market clam until the advent of the Manila clam [Venenipis
philippinanim: (Adams & Reeve 1850); = Riiditupes philippi-
nariim. = Tapes japonica. = Tapes philippinanim. = Tapes
semideciissata. = Venenipis japonica. = Venenipis semideciis-
satu\. another member of the family Veneridae with similar mor-
phology to the native littleneck clam but with a longer market
shelf-life. The Manila clam, also known as the Japanese littleneck
clam, was first observed in British Columbia near Ladysmith Har-
bour in 1936 (Quayle 1964). Introduction presumably occuned
during transplantation of Pacific oyster (Crassostrea gigas) .seed
from Japan, when young Manila clams of several millimeters in
shell length may have been trapped in the oyster shells (Quayle
1964). The dispersal of Manila clams was rapid, and by 1941 they
formed a significant proportion of the commercial catch and were
the doniniant lamellibranch of many beaches (Quayle 1964). They
are now established along both coasts of Vancouver Island, al-
though less abundant in the northern parts, and along similar lati-
tudes on the mainland coast (Bourne 1982).
Varnish clams {Niittallia obscurata (Reeve 1857); = Sole-
lellina ohsciirala, = Psammobia olivacea. = Satelettina japimica].
also known as purple mahogany or Savory clams, belong to the
family Psammobiidae. Originally native to Korea and the Japanese
Islands of Kyushu, Honshu, and Shikoku (Coan et al. 2000), they
have been recently introduced to the Georgia Strait, probably via
ballast water (Gillespie et al. 1999). They have since spread north
into Johnstone Strait, along the west coast of Vancouver Island
north to Checleset Bay. along the mainland coast, south into Puget
*Corresponding author. E-mail: BowerS@dfo-mpo.gc.ca
Sound and along the Oregon Coast to Port Townsend (Dinnel &
Yates 2000. Gillespie et al. 2001). There have been some trial
fisheries but the long-term potential of the fishery is currently
under investigation (Gillespie et al. 1999, 2001).
The purpose of this study was to compare the parasites and
symbionts found in each of the clam species at two different sites.
Clams from each site were gathered at close proximity to each
other and were assumed to have had similar exposure to the spec-
trum of parasites enzootic to that site. This sampling regimen helps
minimize suspicions that observed differences could be the result
of temporal or spatial variations, thereby increasing the interpre-
tative value of negative results. This survey is the first to examine
varnish clams for parasites and symbionts using histological meth-
ods and also contributes to the continuing documentation of para-
sites and .symbionts found in Manila and native littleneck clams.
MATERIALS ANU METHODS
On 10 June 2002. Manila clams, native littleneck clams, and
varnish clams (n = 25) were collected from each of two locations
within the Strait of Georgia on the coast of British Columbia for a
total of 150 clams. The first 75 clams were collected from Crofton
at a beach below a sewage outfall located between the ferry ter-
minal and pulp mill, the others were gathered 2 h later from Boul-
der Point. Ladysmith. At each location clams between 40 and 57
mm in length were dug from a single site (2.0-2.5 m" in area,
approximately 15 cm deep) within the mid-intertidal zone, away
from evidence of eutrophication and fresh water runoff, where
none of the target species were more than 1.5 times more abundant
than another. All clams appeared healthy and were held in tanks
(one tank per site) with flowing ambient seawater for 3—4 days.
Each clam was then shucked, the shell length and wet weight of
soft tissue recorded, superficially examined and pool fixed (5 per
jar) in Davidson's solution. Pea crabs were collected, preserved in
Davidson's solution and held for identification. After at least 24 h
in the fixative two cross sections, one through the region of the
185
186
Marshall et al.
stomach and digestive gland and the other through the kidney and
heart were made. The labial palps, siphon and posterior adductor
muscle were also sampled and processed with the cross-sections
using routine histological techniques. Sections (3-(jLni thick) were
cut and stained with Harris's modified hematoxylin and 0.5% al-
coholic eosin. Additional sections from selected speciinens were
stained with Brown and Hopps Gram stain and also tested for the
presence of DNA using the Feulgen stain reaction. All sections
were examined under a compound microscope (100 to lOOOx).
RESULTS
Average shell lengths of each clam species varied little between
sites but clams collected from the Crofton site had lower wet
weight to shell length ratios (Table 1 ). Native littleneck clams
ranged in length between 41.6 to 50.1 mm from Boulder Point and
41 .6 to 49.4 mm from Crofton; their wet weights were between 5.6
to 11.9 g from Boulder Point and 5.8 to 10.4 g from Crofton.
Manila clams ranged between 41 .4 to 56.4 mm from Boulder Point
and 40.2 to 55. 1 mm from Crofton: wet weights were between 6. 1
to 14 g from Boulder Point and 4.7 to 1 2.9 g from Crofton. Manila
clams showed the least difference in wet weight to shell length
ratio (Table 1 ). Varnish clams ranged between 41 .4 to 53.4 mm in
length from Boulder Point and between 40.0 to 5 1 .7 from Crofton,
wet weights ranged between 4.6 to 1 1 .6 g from Boulder Point and
3.9 to 7.8 g. from Crofton. The average wet weights to shell length
ratio was much less in varnish clams collected from the Crofton
site (Table 1 ).
Pea crabs (family Pinnotheridae) were collected from both Ma-
nila and varnish clams during the shucking process. Only one
immature Piiinixa fabci was found in the Manila clam sample,
however 16-24% of varnish clams contained one pea crab (Table
2). These were also identified as P. faha and were either immature
or male; the largest measured 13 mm across the carapace. The
presence of pea crabs had no obvious pathological effects and did
not affect the wet weight to shell length ratio. For example, the wet
weight to shell length ratio of the six varnish clams from Boulder
Point containing a pea crab was 0.18 g/mm whereas this ratio for
the 19 varnish clams from the same location without pea crabs was
0.17 g/mm. All other organisms were found during histological
examinations.
Colonies of intracellular prokaryotes (Rickettsiae or Chlamy-
diae) were observed within the epithelial cells of the gills and
digestive gland tubules in both Manila and native littleneck clams
(Fig. 1). Gill infections in Manila clams were less frequent (8-
20%) and were considered to be of light intensity (<80 colonies)
compared with native littleneck clams where there was a higher
prevalence (>88%) and many examples of moderate and high
(>200 colonies) intensities (Table 2). Infections within the diges-
tive gland were also more prevalent in native littleneck clams than
in Manila clams (Table 2). The digestive gland was the most
frequent site of infection in Manila clams whereas the gill infec-
tions greatly outnumbered digestive gland infections in native
littleneck clams. Most digestive gland infections were light (<10)
to moderate (10 to 24) in both species except for two cases of
heavy infection in native littleneck clams from the Crofton site
where as many as 55 colonies were counted. The identity of the
intracellular prokaryotes is unknown and may be representative of
more than one species. The colonies within the digestive gland
tubules appeared to be denser than those found within the gill
tissue where it was often possible to see the individuals within the
colony. Between hosts, the colony moiphologies were consistent
and appear to be the same agents as those described by Bower et
al. ( 1992). No associated host response was observed, however the
infected cells (especially gill epithelium) were often swollen be-
yond their normal size (Fig. 1 ). In many cases, host cells of gill
infections were ruptured and the prokaryotes were leaking out into
the water channel.
Colonies of large intracellular rod shaped bacteria (Fig. 1 ) were
obser\ ed at low intensities within gill epithelial cells of 4-52% of
native littleneck clams. The maximum size of these bacteria was
6.3 |xm long by 1 .4 (a,m wide but there were also smaller variants.
Staining characteristics ranged from strongly to very weakly ba-
sophilic and were predominantly gram positive, however, there
were also Gram-negative representatives throughout the entire size
range. Colonies were often 28 |xm in diameter but did not appear
to incite any hemocytic response or otherwise show any indication
of pathology. There was a weak correlation between intensity of
Rickettsia or Chalmydia-like infections and the number of colonies
of rod shaped bacteria observed, clams containing colonies of rod
shaped bacteria were usually infected with moderate to high num-
bers of Rickettsia or Chahnyilia-Wke colonies.
Another inclusion body, also unique to native littleneck clams,
was found in low intensities with 12% prevalence at both sites
(Table 2). These bodies were large, with an average diameter of 65
|j.m. and bound by hemocytes that appeared to have flattened
against the infected cell forming a thick eosinophilic membrane
(Fig. 2). The material within was basophilic, Feulgen positive and
Gram negative, it was of a very fine matrix and denser near the
edges of the colony. The infection was found in nearly every tissue
(heart, kidney, gonad, gill, and palps) and appeared to be the result
of an infected, extremely hypertrophied hemocyte.
Apicomplexan spores resembling Nematopsis sp. were ob-
served at least once in all three species, however, mainly in Manila
and varnish clams collected from the Crofton site (Table 2). The
prevalence in native littleneck clams was very low (4% and 12%)
and there were never more than two spores within an infected
clam. One spore was in the gill epithelium and the others were
found within the gill connective tissue, those found within the
TABLE 1.
.Average shell length and «et weight to shell length ratios of nati>e littleneck clams {Prnlolhaca slamiiiea). Manila dams iVenenipis
philippinarum). and varnish clams {Nitttallia obscurala) examined from two locations in British Columbia, Canada in = 25 for each species at
each location).
Native Littleneck Clams
Boulder Pt./Crofton
Manila Clams
Boulder Pl./Crofton
\ arnish Clams
Boulder Pt./Crofton
Average Shell length (nimi
Wet weight to shell length ratio (g/mm)
4,^.2 / 44.9
0.20/0.17
A5J /4i^.5
().[4/0.1S
47.7/44.7
0.17/0.12
Parasites of Three Bivalves in British Columbia
187
TABI.K 2.
Pre\ak'nce* and intensityt of parasites and synihionts in native littleneck clams {I'mtathaca stainiiuat. Manila clams {Venerupis
philippinarum), and varnish clams [Sutlallia iihsciirata) from two localities in British Columhia. Canada.
Parasite/Svmbiont
Native Littleneck Clam
Boulder Pt. / Crofton
Manila Clam
Boulder Pt. / Crofton
Varnish Clam
Boulder Pt. / Crofton
Rickellsia or Chlamydia in gill
Rickettsia or ChUiiii>dia in
digestive gland
Large intracellular rod shaped
bacteria
Fine-matrix inclusion bodies
Apicomplexan spores
Nemiirnpsis-like
Trichikliiki spp.
Order Rhynchodida
Einieriorin-like coccidian
(Apiconiplexal
Copepods, (Myiilicola-hke)
Other copepods
Tremalode metacercariae
Turbellarians
Pinnotheridae
88%; 19L, 3M (1-130)/ 100';;:
6L. 7M. I2H (12-600)
48%: 9L. 3M { 1-241 / 52%: 7L.
4M. 2H (1-35)
4%: L (l)/52%; L (1-16)
12%: L (1-9)/ 12%: L (3-5)
4%: L(l)/ 12%: L (1-2)
0% / 0%
0% / 0%
0% / 0%
4%: L(l)/8%: L(I)
4%: L(l)/0%
0% / 0%
0% / 0%
0% / 0%
8%: L (1-2)/ 20%:
L(l-
30)
0% / 0%
36%: 7L. 2M ( 1
-24
/ 32%: 5L.
0% / 0%
2M(l-20)
0% / 0%
0% / 0%
0% / 0%
0% / 0%
0%:/80%: 17L
2M
IH
4%; L(l)/ 100%: 16L. 6M. 3H
(3-200)
(3-230)
20%: L (1-3)/ 56%:
L(l
-11)
0% / 0%
20%: L (1-5)/ 44%:
L(l
-15J
0%/0%
4%:L(4)/4%:
L(3)
0% / 0%
4%:L(l)/0%
64%;L(1^)/60%:L(1^)
0% / 0%
4%; L(l)/0%
8%: L(l)/0%
0%/0%
0%/4%;L(l)
8%;L(l)/0%
4%;L(l)/0%
24%:L(1)/16%:L(1)
* Recorded as the percentage of each clam species infected with a given organism at each location.
t Recorded as the number of clams with heavy (H), moderate (M). or light (L) infections (as defined in text), followed by the range of colonies or
individuals of each parasiie/symbiont observed in parenthesis.
conneL'ti\e tissue were acctimpanied by a mild hemocytic re-
sponse. Manila clams were only infected at the Crofton site; the
majority of these infections were light (<60 spores per histological
section) with only a few cases of moderate or high (>150) inten-
sity. Gill connective tissue was the primary focus of infection but
was accompanied by a light infection (one to five spores) of the
palps in 289^ of the clams. There was also one instance where a
single spore was found in the gonadal tissue. The spores appeared
opaque, with no visible internal structures or nuclei, and were
usually accompanied by a focal hemocytic response, identical to
those described by Bovver et al. (1992). Spores found in varnish
clams also occurred predominately within the gill connective tis-
sue but there were also a few spores in the palps (three clams) and
kidney (two clams). A little over half of the spores were of the
same morphology typical to the Manila clams (Fig. .3) and usually
showed a hemocytic response. The remaining spores (Fig. 4) often
contained a nucleus and were clustered within clumps of
hemocytes. making them difficult to discern and accurately count.
These variations appeared to be part of the host immune response
since there was no evidence to suggest that the spores were alive
and capable of progenesis. There also appeared to be a size dif-
ference between the two spore morphologies with one having an
average length of 9.26 ± 1.33 p.m (n = 30) and the other an
average length of 7.91 ± 1.31 (jtm (/; = 30). However the si/e
differences were not statistically significant.
The remaining protistan parasites detected were an eimeriorin-
like coccidia (Apicomplexa) and two ciliates, a Sphenophyra-\\k&
ciliate of the order Rhynchodida and a Trichodina spp., and all
were found exclusively in the Manila clams. The coccidian was
observed within the kidney tissue in one Manila clam from each
location (Table 2) but only the macrogamont stage was observed
(Fig. 5). The macrogamonts were spherical, with a granular cyto-
plasm and a large central nucleus. These macrogamonts have not
been previously observed in Manila clams but ones with similar
morphology have been observed in native littleneck clams (Desser
and Bower 1997a). Although the large size of the macrogamonts
(32-33 |xm in diameter) was sufficient to stretch the kidney tu-
bules there appeared to be very little impact on the host due to the
low intensity and lack of other life stages. The Rhynchodyda-like
ciliate was attached by a stalk between the cilia of the gill epithe-
lium in IfWc and 44% of Manila clams (Table 2). They had a large
prominent nucleus and appeared to be the same as those described
by Bower et al. (1992). There was no evidence of a hemocytic
response and the intensity of infection appeared too light to have
a significant pathological effect. Trichodina spp. (similar to those
described by Bower et al. 1992) were found attached to or closely
associated with the foot, inner surface of the siphon and in one case
the mantle. The prevalence of these organisms was 209f and 56%
and the intensity was light (Table 2). There was no evidence of
tissue disruption or hemocytic response indicative of a pathologi-
cal impact.
Copepods resembling Mytdkola spp. (commonly called red
worms; Fig. 6) were observed at least once in all three clam species
(Table 2), although predominately in varnish clams (60% and 64%
infected) and rarely in the other species (4%^ to 8%). They were
usually found w ithin the lumen of the stomach or intestine but one
was found in the digestive gland duct of a native littleneck clam
(Fig. 7). Intensity was recorded as the number of cross sections and
therefore the same organism may be represented more than once.
In cases where there was more than one cross section in one part
of gut. the lumen was somewhat distended (Fig. 6). otherwise there
was no indication of serious pathology. These copepods have been
observed previously in Manila clams and native littleneck clams as
well as other bivalves (Bower et al. 1994).
188
Marshall et al.
« , ^i**3!F»e*7
» ♦
^' .
0.'
1 -^4:^iil^ R^ fly
JU
%
b
•
* .;
• •
/##
t
▼
^
»
4
•
■
^ $
3
N
'!_■«•
• ►4
Figures 1-5. Inclusion hiidies and protozoa «bser>ed in histological sections of clams from British Columbia, Canada (hematoxylin and eosin
stain, scale bars arc 2(t Mm).
Figure 1. Two strongly basophilic rod-shaped bacteria colonies (B) next to a Rickettsia or Chalmydki-WVx inclusion (R) in the gills of a native
littleneck clam [Pioldllima slamiiiea). Note the size difference in individuals in each type of colony. Both types of inclusions cause considerable
distortion of the host cell.
Figure 2. Large inclusion bound by hcmocytes within the gonad of a native littleneck clam {P. skiminea). The thick membrane surrounding the
inclusion appears to be the result of layers of flattened hcmocytes.
Figure 3. Four Nematopsis-Vke spores (arrowhead) surrounded by hcmocytes within the water channel of a varnish clam {Nuttallia obscuraui).
Figure 4. Three \iinalopsh-\\V.e spores (arrowheads) in the water channel of a varnish clam (,V. obscurata) gill. Note the smaller spore size and
greater number of responding hcmocytes compared with Figure 3.
Figure 5. Three macrogamonts of an eimeriorin-like coccidia in the kidney tubule of a Manila clam (Venerupis pbilippinarmn). The tubule is
greatly distended as a result of the large size of the macrogamonts.
All other metazoaiis observed wei'e copepods. turbellarians. or
trematode metacercariae and all occurred at low frequencies
(Table 2). Two different copepods were found, one in the gill of a
native littleneck clam and the other in the gonad of a varnish clam
(Table 2). The gill copepod (Fig. 8) was large, nearly 750 (im long
in the tissue section, and was observed within the water channel of
the gill. It did nol appear to be attached and despite its size there
was no significant tissue disruption. Two metacercariae were
found in Manila clams, one in the digestive gland (Fig. 9) and
another unencysled one within the pericaidial space. The metacer-
caria within the digestive gland was sunounded by a thick layer of
hcmocytes that caused some local tissue disruption. One turbellar-
ian was found within the intestine of a Manila clam (Fig. 10) and
two turbellarians were found in the kidney tubules of varnish
clams (Fig. 1 1 ). The turbellarians found in the varnish clams both
appeared to be of the same species and were quite large, one was
over 200 |xm in diameter, and therefore caused considerable swell-
ing of the tubule, otherwise no pathological effects were observed.
DISCUSSION
Comparisons of parasite and symbiont prevalences between
Manila, native littleneck. and varnish clams provide strong evi-
dence that there are host preferences. Each parasite/symbiont had
the same order of host preference at both locations except in the
case of Nematopsis-Wke spores. Nematopsis-Wke spores were
rarely observed at the Boulder Point site but were common in
clams from Crofton. Because Nematopsis spores do not reproduce
once they are inside the molluscan host (Sprague and Orr 1955) the
clams from Crofton had a significantly higher rate of invasion.
This may be related to the fact that known species o{ Nematopsis
require a decapod host to complete their life cycle (Lauckner
1983); possibly the Crofton site was more suitable for the alternate
host(s). Another possibility may be related to differences in expo-
sure, the Crofton beach was in a bay and had more protection from
waves and current due to the nearby ferry dock and marina. The
infectious agents may have been washed away from the Boulder
Point site before they reached the filtration field of their potential
bivalve host. These data are limited by time of year and are rep-
resentative of a small geographic area. Whether these patterns of
host specificity are constant throughout seasonal fluctuations and
at different locations is unknown. The assumption that clams of
similar sizes dug from the same micro-site have similar exposures
to potential parasites and symbionts does not work as well for
parasites that are accumulated at low intensities over a long period
of time. Because size is not an accurate measurement of age. clams
of similar sizes cannot be assumed to have the same exposure
times, also clams found in the same micro-site one year may have
been more widely separated in previous years.
Parasites of Three Bivalves in British Columbia
189
\
n % w^^ j- ,}■!
i •,
^v.
y*
* • V -^ f r "•/':
»•. v^
Figures 6-11. Metazoa in clams from British Cuiumbia Canada (hematoxylin and eosin stain).
Figures 6-8. Copepods found during histological examination (scale bars are 10(1 nm).
Figure 6. Mylilicola spp. in intestine of >arnish clam {\iillallia ohsciiiala). Multiple sections maj represent the same organism folding back on
itself. Damage to intestine wall (D) appears to be a sectioning artifact.
Figure 7. Mytilicola spp. in a duct of the digestive gland of a native littleneck clam [Prnlnlhaca slaminea). Note damage to intestinal wall in upper
left of photo between appendages of the copepod.
Figure 8. Section shown is through the appendages and abdomen of a copepod within the water channel of a native littleneck clam iP. slaminea)
gill-
Figure 9-11. Metacercaria and turbellarians (scale bars are 50 pm).
Figure 9. Metacercaria (arrov\ I within the digestive gland of a Manila clam iVenerupispliilippinanim) surrounded b> a focal hemocvtic response.
Figure 10. Turbellarian in the intestine of a Manila clam (\. philippinanim).
Figure 1 1. Turbellarian within a kidney tubule of a varnish clam (,V. ohscurala). The kidney tubule is grossly distended to accommodate the large
size of the turbellarian.
Nemalopsis-Vike spores are able to gain entry into many species
of bivalves (Sprague & Orr 1955, Bower et al. 1994) but do not
always remain viable (Bower et al. 1992). None of the Nemalop-
j/.T-like spores observed in these clams appeared to be alive and
were probably within the wrong host. Viable interactions between
bivalve host and Neiiiatopsis spp. are likely to be highly specific
(.Sprague & Orr 1955). There also appears to be some inhibition of
infection because native littleneck clams were not infected to the
same degree as varnish or Manila clams. Native littleneck clams
have been known to contain Nematopsis-Wke spores (Bower et al.
1994) but these may represent a different species than those en-
countered in this study. The few spores observed in native little-
190
Marshall et al.
neck clams here were slightly smaller and may have represented a
different species that was less abundant. It is uncertain whether the
spores of two different sizes found in the varnish clams were the
same species. However, both spore types were found in the same
tissues and were proportional in abundance so could represent
different stages of host response.
There were many instances where a parasite or symbiont was
unique to only one host, for example, Trichodina spp., Rhyn-
chodida-like and eimeriorin-like protizoa were only found in Ma-
nila clams. Trichodina spp. and Rhynchodida-like ciliates have
been observed on other bivalve species (Bower et al. 19941 and
have a worldwide distribution. Both of these ciliates can be found
in association with Manila clams throughout their range (Bower et
al. 1992); the particular species found on Manila clams may be
enzootic and introduced to British Columbia along with their host.
Both are belie\ed to be benign, large numbers of Rhynchodida-like
ciliates have been reported with no obvious host response or mor-
talities (Bower et al. 1994).
The presence of eimeriorin coccidia in Manila clams and not in
native littleneck clams was unexpected. An eimeriid coccidian
parasite from the kidney of the native littleneck clam has been
described in Washington State, USA (Morado et al. 1984). A
similar, presumably the same, parasite was described and named
(Maii>olisieIla liabatai) by Desser and Bower (1997a) in a low
percentage of native littleneck clams from Southern Vancouver
Island. The macrogamonts observed in the Manila clam appeared
similar to those described in native littleneck clams, however M.
kabatai shares many ultrastructural similarities to coccidian mac-
rogamonts found in California abalone (Hidiolis spp.; Friedman et
al. 1995). Because macrogamonts were the only stage obser\ed it
is impossible to determine whether this is a different species or if
M. kabatai is also able to invade Manila clams. More than one host
species is not unknown in eimeriorin coccidia (Leger 1897. Leger
& Duboscq 1915); however, a survey of 994 Manila clams (Bower
et al. 1992) came across no evidence of this parasite. A possible
explanation may be related to geographic distribution of the para-
site. The Manila clam survey performed by Bower et al. (1992)
only sampled 80 clams south of Nanaimo and those were sampled
in early spring. All records of the kidney coccidia lie further south
than the boundaries of the Manila clam survey, it is possible that
M. kabatai may only be infecting Manila clams from more south-
em populations. Although heavy infections of kidney coccidia in
native littleneck clams can da)iiage the architectural integrity of the
kidney due to lethal hypertrophy of parasitized cells containing
maturing macrogamonts (Morado et al. 1984), the intensity of
infection observed in this study probably had minimal effect on the
host. No link between clam beha\ ior and coccidian infection has
been established in British Columbia, unlike those reported in
Washington by Morado et al. ( 1984). Possibly this parasite has a
greater impact at lower latitudes.
Native littleneck clams weie the only clams infected with fine
matrix inclusion bodies and colonies of large rod shaped bacteria.
Both of these infections are previously undocumented and may be
unique to native littleneck clams. Native littleneck clams have not
been surveyed as intensively as inti'oduced and farmed species of
shellfish so these infectious agents may have escaped detection
until now. Those native littlenecks that have been surveyed were
collected at different locations (Bower et al. 1992), so range or
annual fluctuations may be an explanation. The fine matrix inclu-
sions have potential to be harmful to the host due to their extreme
size if they multiplied or accumulated in vast numbers.
The rod shaped bacteria were at first reminiscent of Rickettsia
or Clialinydici-hke prokaryotes but these individuals were larger
than others described from those groups (see review in Elston and
Peacock 1984). Most of the colonies were much more basophilic
and were usually Gram positive, unlike the paler Gram negative
colonies of what were more typical of Rickettsia-like prokaryotes.
The variations in Gram staining may be related to stages in devel-
opment; there was a tendency for the larger individuals to be Gram
positive but this was not always the case. The conelation between
the intensities of infection of colonies of typical Rickettsia-like
prokaryotes and rod shaped bacteria may be a function of clam
filtering activity or maybe some individuals are more susceptible
to gill infections than others. Unfortunately it was impossible to
compare clam size to infection intensity because the clams had
been pool-fixed. Although this paper separates these bacteria from
the more typical Rickettsia-like colonies it is not unusual to find
variations in the sizes of individual prokaryotes in bivalve inclu-
sions (Elston & Peacock 1984). However, the differences are not
usually as great as those observed here. The taxonomy of intra-
cellular prokaryotes from bivalves is very poorly understood and is
based on morphological observations as opposed to biochemical,
infective or taxonomic relationships with similarly named organ-
isms in higher animals.
Parasitic or commensal crustaceans are common within most
bivalve species; however, those encountered in this survey were
predominantly in varnish clams. Manila clams can be host to more
than one species of pea crab (Bower et al. 1992) but all accounts
to date have found only one species (P. faba) in varnish clams
(Gillespie et al. 2001 ). Immature P. faba are found in many species
of clams in British Columbia but mature pairs are most often found
in the horse clam, Tiesiis capa.x (Hart 1982). Pea crabs are usually
harmless to their host however one study of Manila clams in Japan
found that the presence of pea crabs was related to a decrease in
the ratio of wet weight to shell length compared with unexposed
clams (Sugiura et al. 1960). This relationship has not been ob-
served in any bivalves examined as such in British Columbia. The
prevalence of pea crabs found in the varnish clams is consistent
with a more extensive count by Gillespie et al. (1999) but the
reason varnish clams have so many is unknown.
None of the clams in the present survey were examined fresh;
thus, the specific identity of the Mylilicola-Wke copepod was not
determined. However the most common Mytilicola spp. encoun-
tered in British Columbia is Mytilicola orientalis. which was in-
troduced via Pacific oyster seed (Bernard 1969). It is improbable
that these copepods are enzootic to varnish clams and introduced
at the same time since varnish clams are presumed to have arrived
here in a larval form within ballast water. Rates of infestation of
Mytilicola intestinalis between individuals of the same bivalve
species is passively determined by the host's field of filtration
(Gee & Davey 1986) and are often found in greater abundance in
larger sized hosts (Goater and Weber 1996). This does not explain
their predominance in varnish clams since they are less dependent
on filter feeding and were not significantly larger. Either more
larvae are entering varnish clams or the survival rate is lower in
Manila and native littleneck clams. Varnish clams are deposit and
pedal feeders in addition to filtering (Gillespie et al. 1999), this
action may stir up the sediment more, re-suspending larvae and
increasing the incidence of infection. Some experiments using M.
intestinalis in Europe have been linked to poor growth, tissue
damage and gut metaplasia in oysters and mussels (Koringa 1952,
Parasites of Three Bivalves in British Columbia
191
Odlaug 1946. Sparks 1962) however no pathology has been re-
ported in British Columbia as a result of M. oricntalis (Chew et al.
1965, Bernard 1969).
Both gill and digestive gland Riekettsia or Chalm\dia-\\kc in-
fections showed the same order of host preference with a complete
absence from varnish clams. .Mthough there was no correlation
between numbers of gill colonies compared with number of di-
gestive gland colonies in infected individuals this trend in host
specificity may indicate a close relationship between these two
types of infections. Possibly they are the same species and only
appear different because they are found in different host cells. The
similarity in appearance between species supports this theory and
suggests that one agent may be responsible for these infections.
However, detailed ultrastructural observations, serological or ge-
netic analysis is necessary to make these distinctions. A greater
dependence on filter feeding does not completely explain why
nati\e littleneck and Manila clams have these colonies while var-
nish clams do not as the prokaryotes are not picked up indiscrimi-
nately by passive filtration. Gulka and Chang ( 1984) tried infecting
other bivalves with a rickettsia isolated from a scallop (Pla-
copeclen magelUinicus) but were unsuccessful. This suggests that
these organisms are fairly host specific and those found here were
not able to infect varnish clams. It is possible that these intracel-
lular prokaryotes are a natural parasite/symbionts of native little-
neck clams and are able to successfully colonize Manila clams at
a lower rate due to certain similarities between the hosts. The
prevalence found in Manila clams from this study is similar to that
found by Bower et al. (1992), in comparison the prevalence and
intensity found in native littleneck clams was very high. Infections
of this degree have been observed in farmed scallops without any
indication of pathology, in this case the intensity decreased after
the scallops were moved from contained aquaculture ponds to the
open environment (S. Bower & G. Meyer, personal communica-
tion). This was another case in which location had a pronounced
effect on frequency and intensity of infection, possibly related to
the differences in wave and current exposure between the two
locations. In general these types of prokaryotic infections are not
linked to a pathological response but it has been suggested that
heavy infections may reduce the metabolic efficiency and reduce
the nutritional status of the host (Otto et al. 1979. Elston 1986).
There are a few cases linking intensity of Rickettsia or Chalmydia-
like infections to mortality (Gulka & Chang 1983. Le Gall et al.
1988. Leibovitz 1989) but no detrimental effects have been re-
ported in British Columbia.
The low prevalence or absence of some organisms is also worth
noting. Native littleneck clams collected by Bower et al. (1992) in
1986 and 1990 and by Desser and Bower (1997b) in 1995 were
infected with the elongate sporozoites of a Coccidia-like Apicom-
plexan (37% to 100% prevalence), these organisms were also
found in Manila clams near the Northern end of their distribution.
Some of these samples were taken at the same time of year as the
samples in this study, so seasonal fluctuations are probably not the
cause. These parasites may have been in low abundance in 2002 or
possibly the unknown alternate host does not occur in the Georgia
Strait. There were also fewer turbellarians observed than expected,
this is may be due to an annual fluctuation since they are usually
common in both Manila and native littleneck clams.
ACKNOWLEDGMENT
A heartfelt thank you to J. Blackbourn for technical assistance
and help with staining procedures.
LITER.\TURE CITED
Bernard, F. R. 1969. The parasitic copepod Myulicola oricnkilis in British
Columbia Bivalves. J. Fish. Res. Bd. Can. 26:190-191.
Bower. S. M., S. E. Mc Gladdery & I. M. Price. 1994. Synopsis of infec-
tious diseases and parasites of commercially exploited shellfish. .-A;;".
Re\: Fish Dis. 4:1-199.
Bower. S. M.. J. Blackbourn & G. R. Meyer. 1992. Parasitic and symhiont
fauna of Japanese littlenecks. Tapes plulippinarwn (Adams and Reeve.
1850). in British Columbia. J. Shellfish Res. 11:13-19.
Bourne. N. 19S2. Distnhution. reproduction and growth of Manila clam.
Tapes plulippinanini ( Adams and Reeve ). in British Columbia. ,/. Shell-
fi.sh Re.^. 2:47-54.
Coan, E. V., P. Valentich Scott & F. R. Bernard. 2000. Bivalve seashells
of Western North America. Marine bivalve molluscs from Arctic
Alaska to Baja California. Santa Barbara. CA: Santa Barbara Museum
of Natural History, 764 pp.
Chew. K. K.. .\. K. Sparks & S. C. Katkansky. 1965. Preliminary results
on the seasonal distribution of Myiilicoki orientulis and the effect of
this parasite on the condition of the Pacific oyster Crassoslrea !iij>as. J.
Fish Res. Bd. Can. 22:1099-1101.
Desser. S. S. & S. M. Bower. 1997a. Margolisiella kahatai gen. et sp. n.
(Apicomplexa: Eimeriidae). a parasite of native littleneck clams. Pm-
lotluieea siaminea. from British Columbia, Canada, with a taxonomic
revision of the coccidian parasites of bivalves (Mollusca: Bivalvial.
Folia Parasilol. 44:241-247.
Desser. S. S. & S. M. Bower. 1997b. The distribution, prevalence, and
morphological features of the cystic stage of an apicomplexan parasite
of native littleneck clams (Protolhaca siaminea) in British Columbia. ./.
Parasitol. 83:642-646.
Dinnel. P. A. & E. Yates. 2000. Biological and ecological as.sessments of
Niiltallia obsciinua in north Puget Sound. ./. Shellfish Res. 19:630.
Elston. R. 1986. Occurrence of branchial rickettsiales-like infections in two
bivalve molluscs. Tapes japaniea and Patinopecten yessoensis. with
comments on their significance. J. Fish Dis 9:69-71.
Elston. R. A. & M. G. Peacock. 1984. A Rickettsiales-like infection in the
Pacific razor Clam. Siliqua patula. J. Invert. Pathol. 44:84-96.
Friedman. C. S.. G. R. Gardner. R. P. Hedrick. M. Stephenson. R. J.
Cawthom & S. J. Upton. 1995. Pseudoklossia haliotis sp. n. (Apicom-
plexa) from the kidney of the California abalone. Haliotis spp. (Mol-
lusca). J. Invert. Pathol. 66:33-38.
Gee, J. M. & J. T. Davey. 1986. Experimental studies on (he infestation of
Mxtilus edulis (L.) by Mylilicola intestinalis Steuer (Copepoda. Cyclo-
poida). / Con.seil 42:265-271.
Gillespie, G. E.. M. Parker & W. Merilees. 1999. Distribution, abundance,
biology and fisheries potential of the exotic varnish clam (Nuttallia
obsenraia) in British Columbia. Can. Stock Assess. Secret. Res. Doc.
99/l93:39p.
Gillespie, G. E., B. Rusch. S. J. Gormican. R. Marshall & D. Munroe.
2001. Further investigations of the fisheries potential of the exotic
varnish clam (.Nuttallia obscurata) in British Columbia. Can. Stock
Assess. Secret. Res. Doc. 143:59p.
Goater. C. P. & A. E. Weber. 1996. Factors affecting the distribution and
abundance of Mytilicola orientalis (Copepoda) in the mussel. Mytilus
tro.tsulus. in Barkley Sound. B.C. / Shellfish Res. 15:681-684.
Gulka. G. & P. W. Chang. 1983. Prokaryote infection associated with a
mass mortality of the sea scallop. Placopecten magellanicns. J. Fish
Dis. 6:355-364.
Gulka. G. & P. W. Chang. I9S4. Pathogenicity and infectivity of a rick-
192
Marshall et al.
ettsia-like organism in tlie sea scallop. Placopecten magellaniciis. J.
Fish Dis. 8:309-318.
Hart. J. F. L. 1982. Crabs and their relatives of British Columbia. British
Columbia Provincial Museum. Victoria. 267 pp.
Koringa. P. 1952. Epidemiological observations on the mussel parasite
Mytilicola intestinulis Steur. carried out in the Netherlands. 1951. ,4/;;).
Biol. Copenhagen 8:182-185.
Lauckner, G. 1983. Diseases of mollusca: Bivalvia. In: O. Kinne. editor
Diseases of marine animals, volume II: Introduction. Bivalvia to
Scaphopoda. Hamburg: Biologische Anstalt Helgoland, pp. 542-548.
Le Gall. G.. D. Chagot, E. Mialhe & H. Grizel. 1988. Brachial Rickettsi-
ales-like infection associated with a mass mortality of sea scallop
Pecren nia\inni.s. Dis. Aqiiat. Org. 4:229-232.
Leger. L. 1897. Sur la presence des coccidies chez les mollusques lamel-
libranches. C R. Soc. Biol. 49:987-988.
Leger. L. & O. Duboscq. 1915. Pseudoklossia glomenila n. g. n. sp..
coccidie de lamellibranche. Arch. Zool. Exp, Gen. 55:7-16.
Leibovitz. L. 1989. Chlamydiosis: a newly reported serious disease of
larval and postmetamorphic bay scallops, Argopecren irraiiians (La-
marck). /. Fish Dis. 12:125-136.
Morado. J. F.. A. K. Sparks & S. K. Reed. 1984. A coccidian infection of
the kidney of the native littleiieck clam. Prototlimca staminea. J. In-
vert. Pathol. 43:207-217.
Odlaug. T. O. 1946. The effect of the copepod Mytilicola orienlalis upon
the Olympia oyster. Oslrea liirida. Trans. .Am. Microscop. Soc. 65:3 1 1-
317.
Otto. S. v.. J. C. Harshharger & S. C. Chang. 1979. Status of selected
unicellular eucaryote pathogens, and prevalence and histopathology of
mclusions containmg obligate prokaryote parasites, in commercial bi-
valve molluscs from Maryland estuaries. Haliotis 8:285-295.
Quayle. D. B. 1964. Distribution of introduced marine Mollusca in Bntish
Columbia waters. J. Fish. Res. Bd. Canada 21:1155-1181.
Sparks. A. K. 1962. Metaplasia of the gut of the oyster Crassosirea gigas
(Thunberg) caused by infection with the copepod Mytilicola orientalis
Mori. J. Insect Pathol. 4:57-62.
Sprague, V. & P. E. Orr, Jr. 1955. Nematopsis ostreuni and N. prytherchi
(Eugregarinina: Porosporidae) with special reference to the host para-
site relations. / Parasitol. 41:89-104.
Sugiura, Y.. A. Sugita & M. Kihara. 1960. The ecology of pinnotherid
clams as pest in culture of Tapes japonica-l. Pinnotheres sinensis living
in Tapes japonicu and the influence of the crab on the weight of the
host's flesh. Bull. Jpn. Soc. Sci. Fish. 26:89-94.
Jounuil i>f Shcllt'ish Research. Veil. 22. No. I. 19.^203. 2003.
POPULATION DYNAMICS OF THE ASIATIC CLAM, CORBJCULA FLVMINEA (MULLER) IN
THE LOWER CONNECTICUT RIVER: ESTABLISHING A FOOTHOLD IN NEW ENGLAND
D. P:. MORGAN, M. KESER, J. T. SWENARTON. AND J. F. FOERTCH
Millslone Enviniiiiiicnkil Lab. DominiDii Nuclear Connecticut. Inc.. Wciterford. Connecticut 06.^85
ABSTRACT The founding population of Corhicula flwnmea ni the Lower Conneeticut River, discovered in 1990. was studied for
ten years ( 1991-2000). Seasonal abundance of si.x size classes was monitored near three electric power plants. Corhicula abundance
varied seasonally as well as annually, but peaked in 1992. Winter survival of clams was positively correlated with the average winter
water temperature and negatively correlated with frequency of daily mean water temperatures s 1 °C and with frequency of daily mean
April spring freshet flows ^1700 m'/s. Higher winter survival at Middletown Station sites during most years, when compared with
survival near Connecticut Yankee, was attributed to the influence of the Middletown Station thermal discharge. Thermal discharge did
not support a permanent population at Connecticut Yankee because of temperature extremes during power plant operation in summer.
Clam growth under ambient river temperatures began in May when water temperatures exceeded I0°C and ceased in December when
temperatures fell below this threshold. Cooling water discharges altered this seasonal growth pattern; growth began in November, as
temperatures fell below 35"C. and ceased in the summer, when discharge temperatures exceeded this upper thermal threshold.
Reproduction occurred in the river when water temperatures were between I7"C and 28'C. typically from June to October. Peak
spawning occurred in August. Discharge temperatures shifted clam reproduction back to spring (March to May). The key to Cor-
Ivcula's success in establishing a population in the Connecticut River is its ability to colonize refugia from winter temperature and
spring freshet flow extremes that often cause high clam mortality.
KEY' WORDS: Asiatic clams, Corhicula flumiueu. thermal discharges, electric power plants, winter survival, thermal tolerance,
reproduction, growth, invasive species
INTRODUCTION
The Asiatic clam {Corhiculu Jiuininea) is a freshwater bivalve,
native to southeast Asia, that is now common in Europe, Africa,
the Pacific Islands, and North and South America. Early evidence
of Corhicula in Noith America was empty shells collected in 1924
at a British Columbia site (Counts 1981 ) and at a Columbia River
site in Washington. United States in 1938 (Burch 1944). Today.
Corhicula is reported in 37 US states including, most recently.
New York and Connecticut (McMahon 1983; Foehrenbach &
Raeihle 1984; Morgan et al. 1992). The rapid spread and persis-
tence of Corhicula throughout North America is related to its rapid
growth rate, early onset of maturity, high fecundity, and its ability
to tolerate a wide range of environmental conditions (Mattice &
Dye 1976, Aldridge & McMahon 1978, Graney et al. 1980, Mc-
Mahon & Williams 1986a, McMahon & Williams 1986b, McMa-
hon 2002).
While Corhicula is considered an economically important food
species in its native range (Chen 1990), it is recognized as a
nuisance in North America (Ingram 1959, Sinclair 1964, Prokop-
ovich 1969, McMahon 1977, McMahon 1983, Isom 1986). Its
ability to clog water systems makes Corhicula a serious and costly
problem for the electric generating industry (Goss & Cain 1975,
Mattice 1979, Page et al. 1986). Thus, the discovery of Asiatic
clams in water systems at Connecticut Yankee Nuclear Power
Station (CY) on the Connecticut River in May 1990 (Morgan et al.
1992) received considerable attention. The range extension of Cor-
hicula to the Connecticut River, the northemtnost population in the
eastern United States, was not expected because river temperatures
frequently drop below 2''C, the minimum temperature tolerated by
this clam (Mattice & Dye 1976). This study was initiated in 1991
as a condition of a Connecticut Department of Environmental Pro-
tection (CTDEP) permit to allow CY to continuously chlorinate its
service water system to prevent Corhicula biofouling. Monitoring
was later expanded upriver to the Middletow n and South Meadow
power plant sites. This study examines the abundance, growth, and
reproductive phenology of Corhicula under ambient Connecticut
River conditions and under thermal discharge conditions at the
Connecticut Yankee and Middletown power plant sites.
SITE DESCRIPTION
The Connecticut River originates in northern New Hampshire
near the Canadian border and flows south for 660 km, dropping
800 m in elevation by the time it reaches the mouth at Long Island
Sound (LIS) (Merriman & Thorpe 1976 and Fig. 1). Annual av-
erage water fiow. measured at Thompsonville CT (102 km from
LIS), during the period 1991 to 2000 ranged from a low of 410
mVs in 1995 to a high of 735 mVs in 1996 (USGS 2002). Daily
maximal rates usually occur in April, often exceeding 1700 m /s.
The focus of this study is the lower Connecticut River extend-
ing downstream from Hartford, Connecticut to a point 30 km
above the mouth of the river (Fig. 1 ). The survey area extends over
a 51 kin section of river and encompasses three electrical power
plant sites: South Meadow Station (SM), a 68.5 megawatt, solid
waste-to-energy plant; Middletown Station (MS), an 856 megawatt
oil fired power plant; and Connecticut Yankee (CY), a 582 mega-
watt nuclear power plant (Fig. 1 ). River width varies between -400
m and 600 m over the study area. Depths at sampling sites were
1-6 m below mean low water. Semidiurnal tides affect river fiow,
bringing on average 425 mVs of additional flow to the lower
Connecticut River in the vicinity of CY (Merriman & Thorpe
1976), causing periodic fluctuations in river height of ~l m (NSI
1995, Rozsa 2001). The tidal influences are large in relation to
natural flow during periods of low river discharge, and absent or
nearly absent during freshet conditions (Boyd 1976, Rozsa 2001).
The study areas at CY and farther north at MS and SM are above
any seawater incursion. Daily average ambient water temperatures
were similar for all three power plants and ranged between -1.7
and 30.6°C during the 10-year study period (Fig. 2). The river
frequently freezes over during the winter in our study area, but the
duration of ice cover varies from year to year. Discharge water
temperatures at CY during plant operation were 8 to 12"C above
ambient river temperatures at a maximum flow rate of 25 m /s.
193
194
Morgan et al.
HARTFORD
South Meadow
■Station (SM)
VERMONT \NEW HAMPSHIRE
/ i
1 MASSACHUSETTS
cc
o
>-
I
z
OSPRtNGFIELD
HARTFORD pi \
Q
Z
<
CONNECTICUT
":°'^'r-"^->=^
NEW YORK y"^"^
Middletown
Station (MS)
Connecticut
Yankee (CY)
Figure I. Location of Asiatic clam study area and sampling sites on
the Connecticut River, showing the three electric power station sites
(SM. MS, and CY).
The CY cooling water discharge flows through a man-made canal
1 km long before mixing with ambient Connecticut River waters.
Connecticut Yankee ceased operation on July 22. 1996. At MS. the
average sustained discharge temperatures from 1992-1994 ranged
between 7 and 10"C above ambient river conditions with an av-
erage discharge of 3.6 niVs. At MS and SM. the cooling water is
discharged directly to the river.
MATERIALS AND METHODS
This study was conducted between August 1991 and November
2000. Data at CY were collected during the entire study at four
sampling sites located in the river near the power plant and one site
in the discharge canal (CY discharge). The four CY river sites
were similar in Corbuiiki abundance and the data from each were
combined for data analysis (CY). Sampling was extended to three
sites at MS in May 1992 and continued through November 1994;
two sites were grouped for data analysis as river sites (MS), and
the third, adjacent to the cooling water discharge (MS discharge),
was analyzed separately. At SM. a single river site downstream of
the cooling water discharge (minimal thermal influence) was
sampled between August 1993 and November 1994.
In the first year of the study ( 1991 ), field sampling was con-
ducted in August and November. For the remainder of the study
period (1992-2(M)). field sampling was conducted three times
each year, in May, August, and November. To collect Corbicula.
five 0. 1 ni" bottom sediment samples were obtained at each sam-
pling site using a weighted Peterson grab (Wildlife Supply Com-
pany. Buffalo. NY). Sample processing techniques were similar to
those of Gardner et al. (1976). Grab samples were sieved in the
field by passing the sample through a series of three screens (6.3.
2.0. and 1 .0 mm mesh size). Clams and sediment retained on the
I -mm screen were subsampled in the field by placing a well-mixed
1-L sample in an elulriator (Magdych 1981) for 3 min al a water
flow of 20-30 L/min. The overflow from the elutriator was col-
lected on a I -mm mesh sieve and sorted in the laboratory under a
dissecting microscope (lOx). Sediment and clams retained on the
6.3 and 2.0 mm screens were taken to the laboratory and washed
through a series of five US Standard Testing Sieves (19.0. 12.5.
6.3. 3.4. and 2.0-mm mesh sizes). Size classes were determined
based on the mesh size on which clams were retained. Clams
Figure 2. Intalve (-
91 92 93 94 95 95 97 98 99 00 01
and discharge (----) water temperatures at CV from January 1. 1991 to January 1, 2(M>0. Horizontal reference lines
represent upper and lower lethal temperature limits for Ciirhiculii Jluminea.
CORBICVLA IN THE LOWER CONNECTICUT RiVKR
195
c/)
E
O
0)
u
c
CD
■D
C
3
<
Month
Year
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
5
8
11
399 ±131
807 ±387
55 ±40
2,568 ±1,538
5,209 ±2,630
4.0 ±8.5
35 ±24
206 ±91
68 ±22
225 ±118
124+56
1,828+622
1,522 ±565
56 ±22
80 ±38
8.0 ±7.5
57 ±28
350±136
38 ±33
291 ±130
649 ±272
94 ±35
2.148 ±617
1.758 ±635
78+58
366±138
412±189
ANNUAL
-
2,610 ±1,136
89 ±40
98 ±45
1,158 ±328
45 ±16
138 ±59
326 ±116
1,334 ±362
286 ±85
Figure 3. Average abundance (# clams/m-) of CorbUiila fliiminea by size class (graph) and total (table. ±95% CI) at CV river sites.
retained on the 1.0-nim sieve averaged 2.0 mm in shell length: on
the 2.0-mm sieve. 4. 1 mm; on the 3.4-mm sieve, 6.7 mm; on the
6.3-mm sieve, 14.1 mm; on the 12.5-mm sieve, 19.3 mm; and on
the liJ-mm sieve, 31.1 mm.
Individual clam growth was monitored monthly in 1993 and
1994 using shell length measurements to the nearest 0.1 mm. In the
river near the CY plant intakes, marked clams maintained in lan-
tern nets were used for monitoring growth. In the CY discharge
canal. 12 clams collected randomly from lantern nets were mea-
sured monthly to assess growth.
Clam fecundity was determined monthly using techniques of
Aldridge and McMahon ( 1978). Several hundred adult clams (>8.0
mm in shell length) were collected from the river in May/June of
1991 through 1994 and held in lantern nets placed at two locations,
one in the river near CY plant intakes, the other in the CY dis-
charge canal. Clams were collected monthly from river nets until
winter, when no live clams remained in lantern nets. In the CY
discharge canal, all clams were dead by June (when water tem-
peratures at this site exceeded 37°C). In this study, data for fecun-
dity in the discharge canal were collected from November 1 992 to
July 1993, and June and July fecundity data were at ambient river
temperature due to a power plant shut down. Twelve clams were ANNUAL 11,482 ±4,416 616,.±227 555±253
subsampled monthly from each net. In the laboratory, each clam Figure 4. Average abundance (# clams/m") oi Corbiciila fluminea by
was held under static conditions at 20 "C for 24 h in a lOO-ml size class (graph) and total (table. ±95% CI) at MS river sites.
Month
Year
1992
1993
758 +688
10,413 ±7,233
23,275 ±5,494
43 ±36
878+451
928 ±339
1994
i'38"±68
786 ±577
533 ±295
196
Morgan et al.
beaker filled with filtered Connecticut River water. The number of
juveniles released during this period, determined with a lOx dis-
secting microscope, was recorded as an index of spawning activity.
Additional fecundity assessments were made by dissecting these
clams and noting the presence of brood. Maturity of gametes was
assessed by removing egg and sperm cells from the gonadal tissues
and examining the cells under a compound microscope (400x),
Statistical analyses were performed using SAS version 8 soft-
ware (SAS Inc., Cary. NC). Abundance data in figures are pre-
sented using arithmetic means and non-transformed data. Statisti-
cal comparisons of abundance data were always carried out after
log transformation. The relationships between winter clam survival
(detlned as the ratio of May clam abundance to November clam
abundance from the previous year, expressed as a percentage) and
temperature or river tlow indices were assessed using the rank-
order Spearman correlation. Growth and reproduction data were
not transformed prior to statistical testing.
RESULTS
Abundance
Corhiciila abundance exhibited high intra- and inter-annual
variability. Year to year abundance fluctuations were considerable
at all ambient temperature river sites (Figs. 3, 4. 5; note different
vertical scales). At CY. mean annual clam abundance in 1992,
1995, and 1999 (range 1.158-2,610 clams/nr) was significantly
higher (P < 0.05) than in all other years (range 45-326; Fig. 3). At
MS, mean annual abundance in 1992 (11.482 clams/m") was sig-
nificantly higher (f < 0.05) than in 1993 or 1994 (616 and 555
clams/nr, respectively. Fig. 4). At SM, mean annual abundance
was low, with 82 clams/nr in 1993 and 67 clams/nr in 1994
(Fig. 5).
Of ambient temperature river sites, seasonal abundance at CY
Month
Year
1993
1994
5
8
11
112 ±30
52 ±21
114 ±103
88+75
ANNUAL
82 ±27
67 ±43
Figure 5. .Average abundance (# clanis/ni") of Corhiciila ftuminea by
size class (graph) and total (table. ±95'7f CD at S.M,
over a 10-year period was significantly higher (P < 0.05) in No-
vember than in May or August. November abundance at CY
ranged from 80 clams/nr in 1996 to 5,209 clams/m" in 1992. By
contrast, over the 3 years surveyed at MS ( 1992-1994) and 2 years
surveyed at SM ( 1993 and 1994), abundance was not significantly
different (P > 0.05) between August and November samples. No-
vember abundance at MS in 1992 (23,275 clams/m~) was the
highest observed during the study. Lowest November abundance
occurred at SM in 1993 (52 clams/nr). At all sites, clam abun-
dance in May was significantly (P < 0.05) lower than that in either
August or November.
Of thermally infiuenced sites, seasonal clam abundance in the
CY discharge canal had significant differences (P < 0.05) among
the three sampling periods (Fig. 6). May abundance ranged from
0-92 clams/m". August abundance ranged from 0-12.174 clams/
m". November abundance ranged from 24 to 880 clams/m". At the
MS discharge. August and November abundance estimates were
not significantly different (P > 0.05). ranging from a low of 322
clams/nr in November 1993 to a high of 7.100 clams/m" in No-
vember 1992 (Fig. 7). As with river sites. May abundance at both
CY and MS discharge sites was significantly lower (P < 0.05) than
that in .August and November.
Annual abundance was variable at the CY discharge site. A
pooled f-test of total abundance during operational (1991-1996)
vs. post-operational years (1997-2000) indicated that clam abun-
dance increased significantly (P = 0.007) during post-operational
years. This increase was the result of higher abundance of larger
size class clams (7-14 mm and 19-31 mm) following power plant
shutdown. At the MS discharge site, total clain abundance was
significantly higher (P < 0.05) in 1992 (3.322 clams/nr) than in
1993 and 1994 (496 and 549 clams/nr. respectively: Fig. 7). Clam
abundance was not significantly different (P > 0.05) between the
river and discharge sites at MS, except for the largest clams (31
mm size-class), which were most abundant at the MS discharge
site. In fact, the largest clam measured during the entire study (37.6
mm) was collected at MS in August 1992.
Winler Snnival
Declines in clam abundance from November of one year to
May of the next were used to determine winter survival; values at
CY ranged from 0% in 1994 and 1996 to 55% in 1995 (Fig. 3). The
effects of winter water temperatures and peak river fiows on clam
winter survival were examined using Spearman-ranked correlation
(Table I ). The severity of winter water temperatures, as indicated
by the number of days with average water temperature <2°C, was
not significantly correlated (r^, = -0.65, P = 0.081) with clam
winter survival. The number of days, however, sl°C was nega-
ti\ely correlated (r., = -0.73. P = 0.040) with winter survival, and
average December through April water temperature was positively
correlated (r.,= +0.87. P = 0.004). Highest average monthly flow
in the Connecticut River typically occurs in April (Fig. 8). Ac-
cordingly, the number of days each year exceeding 1.700 mVs in
April was used as an index of spring freshet severity. This index
was negatively correlated with winter clam survival (r.. = -0.91.
P = 0.002). Data from 1993 were omitted from this analysis
because a single storm in March caused total mortality of clams at
our sampling sites.
Growth
Corbicida growth rates under ambient river conditions exhib-
ited seasonal cycles, and growth of marked clams was size-
E
m
O
c;
u
c
ro
■D
c
<
CORBlCfLA IN THE LOWER CONNECTICUT RlVER
12,096
197
95 11 5 8
96 11 5 8
97 11 5 8 r-
98 11 5 8
00
Month
Year
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
5
8
11
34 ±58
178±181
2 ±5.0
12, 174 ±30.269
96 ±121
32 ±22
42 ±26
60 ±54
880 ±1254
24 ±29
38 ±50
90 ±187
8.0 ±5.1
48 ±75
44 ±70
92 ±62
6 ±15
24 ±6.3
20 ±22
62 ±83
212 ±227
76 ±37
243 ±172
210 ±73
4±10
30 ±35
268 ±181
ANNUAL
-
4,091 ±8,454
25 ±14
313 ±397
51 ±53
33+28
41 ±27
98+78
173 ±63
101 ±83
Figure 6. .Average abundance (# liams/m'l of Cnrhiciiki Jhiiniiiea b> size class (graph) and total (table, ±95'7f CI) at C\ discharge.
dependent (Figs. 9 and 10). In 1993, clams with an initial shell
length of -14.5 mm had a higher growth rate (0.54 mm/wk) from
June to October than those starling at -17.5 mm (0.41 mm/wk).
and -21.7 mrt) (0.35 mm/wki. A similar size-dependent relation-
ship was also observed in the 1994 study; clams with an initial
length of -12 mm grew fastest from June lo October (0.51 mm/
wk), followed by -20 mm (0.32 mm/wk) and -30 mm (0.14 mm/
wk) clams. Growth rates were significantly different [P < 0.05)
among the three size classes through August. In September
through December, however, mean monthly growth rates for all
size classes were generally low and not significantly different from
each other.
Clam growth rates in the CY discharge canal from November
1992 to February 1993 were £0.18 mm/wk. when water tempera-
tures were 13-19°C, I0-12°C above ambient river temperatures
(Table 2). As these clams were not marked, negative growth rates
could occur as a result of mortality of large individuals. Growth
rates were as high as 0.27 mm/wk from March to May when water
temperatures ranged from I3-27°C. Maximum growth rates at this
site occurred during June (0.38 mm/wk) and July (0.33 mm/wk),
when canal temperatures were similar to those at ambient river
conditions because of a power plant outage. All clams died after
the power plant restarted and discharge water temperatures ex-
ceeded 37"C (July).
ReprodiictiDii
Microscopic examination of gametic tissues of clams held un-
der ambient river and CY discharge conditions show that eggs and
sperm were continually present as long as clams were alive (Fig.
1 1 ). For clams held at ambient river temperatures, the presence of
embryos and veligers in the demibranchs (brooding) and the active
release of juveniles occurred primarily over a 4-month period
(June to September). By October, only one clam out of 48 exam-
ined was still spawning. The maximum number of juveniles re-
leased per clam per day typically occurred in August across all 4
years in which reproduction was monitored (2.862 juveniles/clam/
day; Fig. 12). This pattern of juvenile release allowed maximum
recruitment to occur just after the period of maximum river water
temperature (July, with a 4-year average of 27.5°C). The number
of juveniles released per adult in August was positively correlated
with the size of the clam (r,= 0.77; P < 0.01; Fig. 13).
The reproductiv e cycle of Corbicuta in the CY discharge canal
was seasonally shifted (Fig. II). Brooding and releasing of juve-
niles first occurred in November 1992 when discharge tempera-
tures averaged I8.3°C, and ceased from December through Feb-
ruary when temperatures averaged <I4°C. Spawning began again
in March and increased through April when discharge tempera-
tures averaged I7'C. The sharp decrease in May was the result of
198
Morgan et al.
Month
5
8
II
1992
206"±i"22
2.666 ±836
7.100 ±1.896
ANNUAL 3,322 ±1,721
Year
1993
"326±14l'
840 ±450
322 ±69
496 ±186
1994
———
340 ±116
860 ±93 1
549 ±283
Figure 7. Average abundance (# clams/m") of Corbicula fluminea by
size class (graph) and total (table, ±95% CI) at MS discharge.
a power plant outage beginning on May 1 3, which dropped cooling
water temperatures from 30°C to 18°C in a single day (Fig. 14).
Spawning activity recovered and peaked in June and July as the
plant outage continued, similar to the pattern observed at ambient
river temperatures (17-27°C). On July 21. 1993 the power plant
restarted and temperatures increased to >35°C in 4 days. By Au-
gust 18. 1993 all clams held in the CY discharge were dead.
DISCUSSION
Corbicula fluminea was first documented in the Connecticut
River in May 1990 (Morgan et al. 1992). the first report of this
nonindigenous clam in New England waters. Before this discov-
ery. Coiiiicula was not expected to colonize the Connecticut River
because water temperatures routinely fall below 2^C for prolonged
periods. It is commonly accepted among researchers that the lower
lethal temperature limit for Corbicula is ~2°C (Homing & Keup
1964. Bickel 1966. Mattice & Dye 1976. Rodgers et al. 1979.
Cherry et al. 1980).
Corbicula abundance varied seasonally as well as annually, but
3500
3000
2500-1
to
5 2000'
I 1500-
1 1000'
<
500-
95 96
Year
Figure 8. Connecticut River daily flow rates (mVs) at the Thompson-
ville, CT gaging station in April from 1991 to 2000.
clearly peaked in 1992. Survival of clams from one year to the next
is positively coirelated with the average December to April water
temperatures and negatively correlated with the number of days
the river water temperature was below I °C and the number of days
that river flows exceeded 1700 mVs in April. For example, no
clams were observed in May at our Connecticut Yankee sampling
sites following the two coldest winters (1993-1994 and 1995-
1996). when river water temperatures dropped below 2^C for 12-
15 weeks and the highest winter survival occurred in 1995 when
daily average river flow in April never exceeded 1700 m /s.
Low survival at Connecticut Yankee and Middletown Station
during the winter of 1992-1993. when water temperature did not
drop below 2°C, was attributed to winter storm Joshua (March 13.
1993). This storm produced low water levels ( 1-2' below normal)
and left shoal areas, specifically our sampling areas, exposed to air
temperatures as low as -8°C. freezing sediment and clams
(NUSCO 1994).
Higher winter survival at Middletown Station sites, when com-
pared with those around Connecticut Yankee, was attributed to the
influence of the Middletown Station thermal dischaige. River wa-
ter temperatures seldom dropped below 2°C in the Middletown
Station discharge mixing zone (NUSCO 1994). Other over-
wintering populations likely exist in the river in refugia provided
by other industrial thermal discharges or in areas of the river
receiving regular influxes of groundwater that maintains a tem-
perature of 9.0 ± 2°C (R. Lewis. State of Connecticut Geologist;
pers. comm.). Graney et al. ( 1980) and Kreiser and Mitton ( 1995)
suggest that warm water refugia such as these were assisting the
Asiatic clam in expanding its geographical range northward.
Clam densities in the Connecticut Yankee discharge canal were
TABLE 1.
Spearman correlations coefficient ( rj for percentage winter survival of Cnrhicula fluminea at CY \ersus indices of winter temperatures and
Connecticut Rixerflow.
Variable
r^
Prob >lrl
n"
Mean
Std Error
Min.
Max.
Percentage Survival''
_
_
8
12.7%
6.23%
0%
54.9%
Ave. Winter Temp.*^
-1-0.87
0.004
8
2.93
0.37
1.32
4.86
No. Days S1°C
-0.73
0.040
8
54.9
8.75
17
93
No. Days s2°C
-0.65
0.081
8
70.6
8.12
28
103
Flow a 1 700 ni'/s''
-0.9 1
0.002
8
6.4
1.54
13
' 1993 data were omitted because of the mortality caused by the March storm Josliua (see text).
'% Survival = (May abundance/prior November abundance) x 100,
' Average Winter Temperature = the annual December to April mean daily Connecticut River temperature at CY.
' Number of days in April when the Connecticut River flow equaled or exceeded 1700 m'/s.
CORBICULA IN THE LOWtR CONNECTICUT RlVER
199
Jun
1.1 -
1.0-
^ 0.9-
-0.8-
^ 0.7-
Ld 0.6-
< 0.5-
^ 0.4-
|0.3.
g 0.2.
0.1 -
1 <-
-^^
V
^^-.
"^^
0.0-
Jul
Aug
Sep
Nov
Dec
Figure 9. Corhiciila Jhiminca growth rates (mni/«k) in 1993 for marked clams within initial size classes based on shell length. Vertical bars
represent two standard deviations around the mean growth rates for three individuals in each size class.
most \ariable. Large numbers of small (2 mm) clams that appar-
ently survived passage through the power plant cooling water sys-
tem characterized transient populations in the canal. A permanent
population, however, was not established during power plant op-
eration because summer water temperatures often exceeded 37°C,
the upper lethal temperature limit for Corbicula in our study. Mc-
Mahon and Williams (!986b) reported similar findings for Cor-
bicula living in the themial discharge of the Handley Power Sta-
tion in Texas. Following Connecticut Yankee closing in 1996, size
range of clams collected in the discharge canal has increased with
shell lengths now ranging from 2-19 mm. These results indicate
not only that clams are successfully over-wintering in the canal
under ambient river temperatures, but also surviving for >1 year.
The canal essentially has become a cove where circulation is de-
pendent on semidiurnal tidal exchange, and not \ ulnerable to high
spring freshet water Hows.
Clam abundance in the Middletown Station discharge area also
fluctuated, but was consistently higher than abundance at CY dis-
charge during the same period. Similar to the CY discharge, the
population near the MS discharge was dominated by clams 2 mm
in size. In contrast, however, to the CY discharge, clams of all size
clas.ses, including those in the 31 mm class, were regularly col-
lected at the MS discharge. The presence of larger size clams
suggests that this area provided a more stable refugium. The 37-
mm clam collected at this site in 1992. along with growth rates
observed during our study, suggests that Corbicula has been
I
n 4-
1—
5
o
0.3-
(r
o
0.2-
0.1 •
0.0-
Aug
Sep
Figure 10. Corbicula fluminca growth rates (mm/wkl in 1994 for marked clams within initial size classes based on shell length. Vertical bars
represent two standard deviations around the mean growth rates for two to five individuals in each size class.
200
Morgan et al.
TABLE 2.
Corbicula fluininea growth in the C\ discharge canal from November 1992 to July 1993.
Date
Growth
Weeli
Average Length
(mm)
SE
Minimum Length
(mm)
Maximum Length
(mm)
Growth Rate
(mm/wk)
1 1/10/92
20.20
12
0.69
15.3
1 2/22/92
6
20.04
12
0.38
18.1
01/26/93
11
20.97
12
0.59
17.7
02/23/93
IS
20.89
12
0.48
18.6
03/23/93
19
21.96
12
0.43
19.2
04/29/93
24
23.04
12
0.42
20.6
05/18/93
27
22.57
12
0.31
20.4
06/24/93
32
24.57
11
0.40
21.5
07/22/93
36
25.89
12
0.39
24.2
24,0
21.5
26.5
23.8
24.4
24.9
24.1
26.1
27.6
-0.026
0.185
-0.019
0.267
0.205
-0.172
0.378
0.330
.Ian i Feb Mar Apr May -lun
.Aug Sep
A
fnhicn
River Conditions
WINTER
MORTALITY
Uts
Sp.rn,
Brooding
Releasiim Ju\cniles i i i
1 1 i
Cooling
Water Discharge Conditions
1 1 i
Fees
SUMMER
MORTALITY
Spcnn
Brood inj^
Rclcasmc.luvcnilc.s
Figure 11. Summarization of the 1991 to 1994 annual reproductive
cycle of Corbicula Jhiiniiiea under ambient Connecticut River condi-
tion,s and the thermallv elevated conditions of the CV cooling water
discharge.
present in the river since 1988. Winter water temperatures were
moderated by the Middletown Station thermal discharge, and sum-
mer thermal stress was reduced because of rapid dilution of dis-
charge waters with ambient river water. In addition, the MS ther-
mal discharge flow was only ~\59r that of CY.
Ciiibicida growth in the Connecticut River under ambient wa-
ter temperatures is consistent with reports by other researchers in
North American (Morton 1977. Britton et al. 1979. Eng 1979.
Mattice 1979, McMahon 1983. Welch & Joy 1984, Joy 1983.
Matlice & Wright 1986. McMahon & Williams 1986b. Doherty et
al. 1990, French & Schloesser 1991). and was primarily influenced
by water temperature. Growth began in May when water tempera-
tures rose above IO°C and continued until December when water
temperatures dropped below this threshold. Other researchers re-
ported 9-15°C to be the lower temperature threshold for growth of
Corbicula fluininea in their studies (Hall 1984. Mattice & Wright
1986, McMahon & Williams 1986a, French & Schloesser 1991).
o
o
Q
40
30
5
o
;;^ 20
LU
>
3
O 10-
.
>^.
■
^
/
p
\
N.
-
1
^
.
/
1
/
I
\
\
\
/
1 \
\
,
/
1 \
\
■
9
/ \
\
\
■
1
1
I \
■
1
1 I
1 ^
\
*
/
■
\ «
-
A 1
\
.
y /
\ \
.
/
\ ^
■
■' 1
V
> -
=^ , 1 y . .
, ^> ^ 9 ^
30
25
2
<
20 3
15 73
4 5 B 7
MONTHS OF THE YEAR
10 11 12
10
Figure 12. Corhiciila fluininea fecundity
from 1991-1994.
(- -) and water temperature (- - : - -) for clams held in ambient temperature Connecticut River water
COHHICULA IN THE LoWhR CONNECTICUT RlVER
201
o
o
>-
<
<
_J
O
•\
>
3
O
d
75
50
25-
n =47
r=0.77
p<0.0001
.^<'-€>
100-
.■■0^-^^'
75;
a .^<^>>" ,„-'-
50-
o
o ,
Q' ^^^ , '' o
.'.^ ,-' o
yh-
^^^'' °
:
^^
'' rx.^
-■d' o
■
'--°'c
■^^-'
ol
T 1 1—
ro^
1 1
°o
1 1 1 1 1 1 1 1 1 1 —
12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42
LENGTH OF CLAM (mm)
Figure 13. Linear regression with 95 '7r CI on mean predicted \aliies for the number of Juveniles released per day in relation to shell length (mm)
of the spawning Corbktila jlumiiwa during the peak spawning month of August in the mainstem Connecticut River.
MONTHS OF THE YEAR
Figure 14. Corhiciila Jhiminea fecundity (-0-) and water temperature (- - - -) for clams held in the discharge canal at CV from November
1992 to August 1993.
202
Morgan et al.
Highest growth rates occurred in July and August, when river
water temperatures pealced (25-30°C). and growth rates were sig-
nificantly higher for the smaller clam sizes.
The upper temperature tolerance of Corhicithi determined in
this study is within ranges reported by other researchers in labo-
ratory and field experiments (Mattice & Dye 1976. Dreier 1977.
Mattice 1979. Cairns & Cherry 1983. McMahon & Williams
1986a). Corhuula growth in the CY thermal discharge canal was
initiated in November 1992 when water temperatures dropped to
<35°C. Growth continued until August 1993. when water tempera-
tures were >37°C and clams died.
Seasonal water temperatures also control reproductive cycles of
the Connecticut River Corbiciila population. The presence of eggs
and sperm was continuous in the Connecticut River population of
this species as long as water temperatures supported its survival.
Brooding and releasing of juveniles occurred when water tempera-
tures were between 17-28^C. typically from June to October.
Spawning temperatures of 14— 27°C were reported by other re-
searchers in North America (Eng 1979. Mattice 1979. Hall 1984.
Cherry et al. 1986. Foe & Knight 1986; McMahon & Williams
1986a: Doherty et al. 1987; Rajagopal et al. 2000).
A single annual spawning peak for the Corhicithi population in
the Connecticut River occurred in August. Others reported two
Corbiciila spawning peaks, one in spring and one in fall (Heinsohn
1958. Aldrige & McMahon 1978. Eng 1979. McMahon 1983. Foe
& Knight 1986. McMahon & Williams 1986a). Several others
have reported a single spawning peak (Bickel 1966. Homback
1992, Mouthon 2001). The presence of a single reproductive peak
in the Connecticut River population may be related to longer pe-
riods of cold-water conditions, more severe spring Hooding, and
the quantity and quality of available food.
The altered thermal regimen within the CY discharge canal
shifted the period of reproduction from the ambient river period of
June through September to November and March through May
when water temperatures in the canal ranged between 16-30'C.
Spawning during July and August 1993 occurred because the
power plant was off-line and the discharge water temperatures
were not elevated. These results demonstrate that thermal dis-
charges can alter the reproduction cycle of Corbiciila. Aldridge
and McMahon (1978) and Dreier and Tranquilli (1981) reported
that Corbicula fliiminea spawning activities stopped at tempera-
tures of 30-34°C. most likely due to thermal stress. Graney et al.
( 1980) speculated that elevated temperatures in thermal discharges
may e.xtend the spawning season into the winter.
In conclusion, this study showed that the Connecticut River
has supported a fluctuating Corbiciila population for at least 10
years. Cold water temperatures (<2°C) for several weeks, and
high water flow in the spring caused high mortality of clams in the
river during the winter and early spring. Growth and reproduc-
tion for Corbiciila in the Connecticut River peaked in July and
August when river temperatures ranged between 24-30°C and
only one spawning peak occurred each year. The key to Corbicii-
la's unexpected success in establishing a population in the
Connecticut River is its ability to colonize refugia from cold win-
ter water temperatures and spring freshet flows that cause high
clam mortality. Following the closing of (he CY power plant.
Corbiciila continued to populate the CY river sites establish-
ing a more mature population in the discharge canal. Based on
our observations of Corbiciila in the Connecticut River, we ex-
pect that this species will continue to successfully colonize other
rivers and lakes in New England, where similar winter refugia
exist.
LITERATURE CITED
Aldridge. D. W. & R. F. McMahon. 1978. Growth, fecundity, and hioen-
ergetics in a natural population of freshwater clam. Corhiciilii inanilen-
sis Philippi. from North Central Texas. J. Molt. Studies. 44:49-70.
Bickel. D. 1966. Ecology of Corbicula manileiuis in the Ohio River at
Louisville Kentucky. Sterkiana 23:19-24.
Boyd. W. A. 1976. Hydrology. In: D. Merriman & L. M. Thorpe, editors.
The Connecticut River ecological study, the impact of a nuclear power
plant. Bethesda. MD: Amer. Fish. Soc. Monogr. No. 1. pp. 25-34.
Britton. J. C. D. R. Coldiron, L. P. Evans. Jr.. C. Golightly. K. D. O'Kane
& J. R. TenEyck. 1979. Reevaluation of the growth pattern in Cor-
bicula fluminea (Miiller). In: J. C. Britton. editor. Proceedings. First
International Corbicula Symposium. October 13-15, 1977. Fort Worth.
Texas: Texas Christian University, pp. 177-192.
Burch. J. Q. 1944. Checklist of West North American marine mollusca
from San Diego. California to Alaska. Minutes of the Coiuhological
Club of Southern Califoniia 38:11-18.
Cairns. J.. Jr. & D. S. Cherry. 1983. A site-specitlc field and laboratory
evaluation of fish and Asiatic clam population responses to coal-fired
power plant discharges. Water Sci. Tech. 15:10-37.
Chen, L. 1990. Aquaculture in Taiwan. Cambridge MA: Fishing News
Books.
Cherry. D. S.. J. H. Rodgers. Jr.. R. L. Graney & J. Cairns, Jr. 1980.
Dynamics and control of the Asiatic clam in the New River. Virginia.
Bull. Va. Water Resour. Res. Cent. 123:1-72.
Cherry. D. S.. R. L. Roy. R. A. Lechleitner. P. A. Dunhardt. G. T. Peters
& J. Cairns. Jr. 1986. Corbicula fouling and control measures at the
CELCO Plant. Virginia. In: J. C. Britton & R. S. Prezant. editors.
Second International Corbicula Symposium, June 21-24. 1983. Hatties-
burg, Mississippi. Amer Malacol. Bull. Special Edition No. 2. pp. 69-81.
Counts. C. L.. lii. 1981. Corbicula jhnninea (Bivalvia: Corhiculidea) in
British Columbia. Nautilus 95:12-13.
Doherty. F. G.. D, S. Cherry & J. Cairns, Jr. 1987. Spawning Periodicity of
the Asiatic Clam Corbicula fluminea in the New River. Virginia. .Am.
Midi. Nat. 117:71-82.
Doherty. F. G., D. S. Cherry & J. Cairns. Jr. 1990. Multiseasonal tissue
growth trends in Corbicula fluminea (Bivalvia: Corbiculidae) from the
New River. Virginia. Nautilus. 104:10-15.
Dreier. H, 1977. Study of Corbicula in Lake Sangchris. In: The Annual
Report for Fiscal Year 1976, Lake Sangchris Project. Sect. 7. Urbana:
Illinois Natural History Survey, pp. 1-52.
Dreier, H. & J. A. Tranquilli, 1981. Reproduction, growth, distribution and
abundance of Corbicula in an Illinois cooling lake. In: R. W.
Lawrimore & J. A. Tranquilli, editors. The Lake Sangchris Study: case
history of an Illinois cooling lake. Bulletin Illinois Natural History
Survey, No. 32. pp. 378-393.
Eng, L. L. 1979. Population dynamics of the Asiatic clam, Corbicula
fluminea (Miiller). in the concrete-lined Delta-Mendota Canal of Cen-
tral California In: J, C. Britton, editor. Proceedings. First International
Corbicula Symposium. October 13-15. 1977. Fort Worth, Texas:
Texas Christian University, pp. 39-68.
Foe, C. & A. Knight. 1986. A thermal energy budget for juvenile Corbicula
fluminea. Amer. Malacol. Bull. 2:143-150.
Foehrenbach, J, & D. Raeihle. 1984. A further range extension of the
Asiatic clam. NY Fish and Game J. 31:224-225.
French, J. R. P. II & D. W. Schloesser. 1991, Growth and over-winter
survival of the Asiatic clam, Corbicula fluminea. in the St. Clair River.
Michigan. Hydrobiologia 219:165-170.
Gardner, J. A., Jr.. W. R. Woodall. Jr.. A, A. Staats, Jr. & J. F. Napoli.
CORBICVLA IN THE LOWER CONNECTICUT RiVER
203
1976. The invasion of the Asiatic clam (Corbicula munilensis Philippi)
in the Altanuiha River. Georgia. Naiilihis 90:117-125.
Goss. L. B. & C. Cain. Jr. 1975. Power plant condenser and service water
system fouling by Corbicula. the .Asiatic clam. Presented at the Bio-
fouling Workshop held at John Hopkins University, Baltimore. Mary-
land, June 16-19, 1975.
Graney, R. L., D. S. Cherry, J. H. Rodgers, Jr. & J. Cairns. Jr. 1980. The
influence of thermal discharges and substrate composition on the popu-
lation structure and distribution of the Asiatic clam, Corbicula flu-
minea. in the New River. Virginia. Nautilus 94:130-135.
Hall, J. J. 1984. Production of immature Corbicula fluminca. in Lake
Norman. North Carolina. Nautilus 98:153-157.
Heinsohn. G. E. 1958. Life history and ecology of the freshwater clam,
Corbicula fluniinea. MSc Thesis, Santa Barbara. CA: L'niversity of
California.
Homback, D. 1992. Life history traits of a riverine population of the Asian
clam Corbicula fluminca. .Am. Midi. Nat. 127:248-2.S7.
Homing. W. B. & L. Keup. 1964. Decline of the Asiatic clam in the Ohio
River. Nautilus 78:29-30.
Ingram, W. M. 1959, Asiatic clams as potential pests in California water
supplies. / Amer. Wateniorks Assoc. 51:363-370.
Isom, B. G. 1986. Historical review of the Asiatic clam {Corbicula) inva-
sion and biofouling of waters and industries in the Americas, June
21-24, 1983, Hattiesburg, Mississippi. Amer. Malacol. Bull. Special
Edition No. 2. pp. 1-5.
Joy. J. E. 1985. A 40-week study on growth of the Asian clam. Corbicula
fluminca (Miiller). in the Kanawha Ri\er. West Virginia. Nautilus 99:
110-116,
Kreiser, B, R, & J. B. Mitton. 1995. The evolution of cold tolerance in
Corbicula flinninea. (Bivalvia: Corbiculidae). Nautilus 109:111-112.
Magdych. W, P, 1981. An efficient, inexpensive elutriator design for sepa-
rating benthos from sediment samples. Hycirobiologia 85:157-159.
Mattice, J. S. 1979. Interactions of Corbicula sp. with power plants. In: J.C.
Britton, editor. Proceedings. First International Corbicula Symposium.
October 13-15, 1977. Fort Worth, Te.xas: Texas Christian University,
pp. 119-138,
Mattice. J, S. & L. L. Dye. 1976. Thermal tolerance of the adult Asiatic
clam. Pages 130-135 In: G. W. Esch & R. W. McFarlane. editors.
Thermal Ecology II. Conf. 750425. Nat. Tech. Info. Serv.. Springfield,
VA.
Mattice. J. S. & L. L. Wright. 1986. Aspects of growth of Corbicula
fluminca. June 21-24. 1983. Hattiesburg, Mississippi. Amer. Malacol.
Bull. Special Edition No. 2. pp. 167-178.
McMahon. R. F. 1977. Shell size-frequency distributions of Corbicula
manilensis from a clam-fouled steam condenser. Nautilus 91:54-59.
McMahon, R. F. 1983. Ecology of an invasive pest bivalve. Corbicula.
Pages 505-561 in W,D, Russell-Hunter (ed.). The Mollusca. Vol. 6.
Ecology. Academic Press, Orlando, FL.
McMahon, R. F. 2002. Evolutionary and physiological adaptations of
aquatic invasive animals: /■ selection versus resistance. Can. J. Fish.
.Aquat. Sci. 59:1235-1244.
McMahon, R. F. & C. J. Williams. 1986a. A reassessment of growth rate,
life span, life cycles and population dynamics in a natural population
and field caged individuals of Corbicula fluminca iMiiller) {Bi\alvia:
Corbiculacea), June 21-24, 1983, Hattiesburg. Mississippi. Amer. Ma-
lacol. Bull. Special Edition No, 2, pp. 151-166.
McMahon, R. F. & C. J. Williams. 1986b. Growth, life cycle, upper ther-
mal limit and downstream colonization rates in a natural population of
the freshwater bivalve mollusc, Corbicula fluminca (Miiller) receiving
thermal effiuent, June 21-24. 1983. Hattiesburg. Mississippi. Amer.
Malacol. Bull. Special Edition No. 2. pp. 231-239.
Merriman. D. & L. M. Thorpe. 1976. The Connecticut River ecological
study, the impact of a nuclear power plant. Amer. Fish. Soc. Mono-
graph No. 1. Bethesda. MD.
Morgan. D. E.. M. Keser. J. F. Foertch & J. M. Vozarik. 1992. The Asiatic
clam. Corbicula fluminca. extends its North American distribution to
the Connecticut River [abstract only]. In: W. J. Blogoslawski, editor.
In: Aquaculture and environmental stewardship: Milford Shellfish Bi-
ology Seminar 1991. Environmental management, pp. 521-529.
Morton, B. 1977. The population dynamics of Corbicula fluminca (Bi-
valvia: Corbiculacea) in Plover Cove Reservoir. Hong Kong. J. Zool.
181:21^2.
Mouthon, J. 2001. Life cycle and population dynamics of the Asian clam
Corbicula fluminea (Bivalvia: Corbiculidae) in the Saone River at
Lyon (France). Hydrobiologia 452:109-1 19.
NSI. (Nautical Software, Inc). 1995. Tides & Currents for Windows, Ver-
sion 2.0. nauticalswts'aol.com. Beaverton, Oregon.
NUSCO (Northeast Utilities Service Company). 1994. 1993 annual report
on Corbicula studies at Connecticut Yankee. As submitted to the Con-
necticut Department of Environmental Protection on January 27. 1994.
NUSCO Letter No. D07282. 19 pp.
Page. T. L.. D. A. Neitzel. M. A, Simmons & P. F. Hayes. 1986. Biofouling
of power plant service systems by Corbicula. June 21-24. 1983. Hat-
tiesburg. Mississippi. Amer. Malacol. Bull. Special Edition No. 2. pp.
41-45.
Prokopovich, N. P. 1969. Deposition of clastic sediments by clams. Sedi-
men. Petrol. 39:891-901.
Rajagopal, S.. G. van der Velde & A. bij de Vaate. 2000. Reproductive
biology of the clams Corbicida flumimdis and Corbicula fluminca in
the river Rhine. Arch. Hydrobiol. 149:403^20.
Rodgers, J. H., Jr.. D. S. Cherry. K. L. Dickson & J, Cairns, Jr. 1979.
Invasion population dynamics and elemental accumulation of Cor-
bicula fluminea in the New River at Glen Lyn, Virginia. In: J. C,
Britton. editor. Proceedings. First International Corbicula Symposium.
October 13-15. 1977. Fort Worth. Texas: Texas Christian University,
pp. 99-110.
Rozsa. R. 2001. Hydrology of the lower Connecticut River. In: G. D.
Dreyer & M. Caplis. editors. Living resources and habitats of the low er
Connecticut River. December 2001. New London. CT: The Connecti-
cut College Arboretum Bull. No. 37. pp. 18-22.
Sinclair. R. M. 1964. Clam pests in Tennessee water supplies. J. ,Am. Water
Works Assoc. 56:592-599.
USGS. (United States Geological Survey). 2002. Water resources of the
United States, surface-water data for Connecticut, August 15.
2002. <http://waterdata.usgs.gov/ct/nwis/sw.
Welch. K. J. & J. E. Joy. 1984. Growth rates of the Asiatic Clam. Cor-
bicida fluminea (Miiller). in the Kanawha River. West Virginia. Fresh-
mit. Imertebr. Biol. 3:139-142.
Joiinuil oj Shclllhh Rcwiinli. Vol. 22, No. I. 205-2()S. 201)3.
QPX, A PATHOGEN OF QUAHOGS (HARD CLAMS). EMPLOYS MUCOID SECRETIONS TO
RESIST HOST ANTIMICROBIAL AGENTS
ROBERT S. ANDERSON,'* BRENDA S. KRAUS,' SHARON MCGLADDERY," AND
ROXANNA SMOLOWITZ'
^Chesapeake Bi(>loi>ical Laboratory. University of Maryland. Center for Environmental Science. P.O.
Box .^(S. Solomons. Maryland 206HS: ^Department of Fisheries and Oceans. Canada. Gulf Fisheries
Center. P.O. Box 5030. Moncton. N.B. EIC 9B6: "'Marine Bioloi;ic Laboratory. 7 MBL Street. Woods
Hole. Massachusetts 0254.1
.ABSTRACT The thraustochytrid protist quahog parasite unknown (QPX) has caused mass mortalities of hard clams (Mercenaria
nicneiuiria) in Atlantic Canada and Massachusetts. It typically secretes copious mucus in vivo and in vitro. M. mercenaria plasma
contains naturally-occurring agents that modulate growth of QPX cultures. This activity was shown by exposing washed, mucus-free
QPX (wQPX) to filter-sterilized M. mercenaria plasma. Low plasma protein concentrations (<10 |xg/ml) in the medium tended to
stimulate QPX growth; higher concentrations (10-50 (jLg/ml) produced dose-dependent inhibition. If wQPX were incubated for various
times before exposure to an inhibitory concentration of M. mercenaria plasma, a time-dependent protection from the plasma was
observed; total protection was seen after -24 h preincubation. This effect was probably a result of the re-establishment of the mucoid
coats around the wQPX during preincubation. These data suggest th;it ihe mucoid secretion of QPX may represent an important
virulence factor.
KEY WORDS: quahog parasite unknown (QPX). Mercenaria mercenaria. virulence factors, clam diseases
INTRODUCTION
Whyte et al. (1994) described a protistan parasite that caused
high mortalities in a hard clam (Mercenaria mercenaria) hatchery
on Prince Edward Island, Canada; the causative agent was named
quahog parasite unknown (QPX). This organism was similar or
identical to the clam pathogen first observed by Drinnan and Hen-
derson ( 1963) in New Brunswick, Canada. Subsequently, QPX has
been cited as the cause of mass mortalities of M. mercenaria in
Massachusetts (Smolowitz et al. 1998) and has been reported in
several Virginia coastal embayments (Ragone Calvo et al. 1998).
Molecular phylogeny studies based on sequencing of I8S riboso-
mal RNA suggest that QPX is a member of the phylum Labyrin-
thulomycota (Maas et al. 1999. Ragan et al. 2000). in the thraus-
tochytrid phylogenetic group (Stokes et al. 2002).
A medium developed by Kleinschuster et al. (1998) has per-
mitted in vitro cultivation of QP,\. In culture, thalli were shown to
grow and mature into sporangia containing numerous vegetative
endospores. The endospores were released on rupture of the spo-
rangia and in turn matured to form thalli. and the stages of the
vegetative life cycle were repeated. Whyte et al. ( 1994) and Klein-
schuster et al. ( 1998) reported conversion of endospores to motile
zoospores in sterile seawater. Later studies (Brothers et al. 2000).
however, were unable to replicate these findings. The vegetative
life stages of QPX have been observed in the tissues of infected M.
mercenaria. In many instances, the QPX cells were seen in histo-
logic sections to be enclosed by a translucent space; this was
initially attributed to lysis of host tissue by enzymes secreted by
the parasite (Whyte et al. 1994). Subsequently. Smolowitz et al.
( 1998) determined that in live animals, the space is occupied by a
muco-fibrillar substance produced by the parasites; and that this
substance is removed by histologic processing. It was suggested in
that study that phagocytosis of the parasite in the clams' tissues is
inhibited by the mucofibrillar secretions of the parasite.
The disease caused by the Canadian strain (CA QPX) as de-
scribed by Whyte et al. ( 1994) is similar to that described for the
Massachusetts strain (MA QPX) by Smolowitz et al. (1998). MA
QPX. however, primarily infected the mantle and gill and some-
times produced nodules; CA QPX infections were more commonly
seen in the connective tissue of the foot and were rarely associated
with nodules. Areas of infection by CA QPX and MA QPX trig-
gered inflammatory responses involving extensive infiltration of
adjacent host tissues by hemocytes. with some evidence of phago-
cytosis and/or encapsulation of the parasites. Inflammatory foci
caused by MA QPX sometimes contained phagocytic multinucle-
ated giant cells similar to those produced /;) vitro by Anderson
(1987). Apparently QPX infection elicits a vigorous cellular re-
sponse, but this activity is insufficient to control the disease. Hu-
moral QPX modulatory agents in M. mercenaria plasma are de-
scribed for the first time in this article, and Ihe role of QPX mucoid
secretions in protection from them.
MATERIALS AND METHODS
*Corresponduig author. Tel.: -^ 1-4 10-326-7247; Fax; +1-410-326-7210;
E-mail; andersonts'cbl. umces.edu
QPX
These studies were carried out using MA QPX obtained from
Dr. R. Smolowitz, Marine Biologic Laboratory, Woods Hole, MA.
They were propagated in the medium of Kleinschuster et al.
(1998). The initial seeding density was 10"'/ml and the cultures
were maintained at 23°C and were har\ested at 7 d ( 168 h) while
still in exponential growth phase. The QPX cells were enveloped
by a heavy mass of mucoid secretion, which was routinely washed
off the cells by dilution with a saline solution. lO (25 ppt. Instant
Ocean®, Aquarium Systems Inc.; Mentor, OH), followed by re-
peated centrifugations (300 x g, 10 min, 21-0, x3). Washed QPX
(wQPX) were >90'7f viable by the trypan blue exclusion assay
(Hanks & Wallace 1958) and almost immediately resumed mucus
secretion. The numbers of QPX cells in particular cultures and cell
numbers required for subsequent experiments were quantified
spectrophotometrically using a standard curve of the numbers of
205
206
Anderson et al.
wQPX (as determined in a Ineniacytometer) as a function of tlieir
absorbance at 560 nm.
C9G
Another thraustochytrid. C9G. closely related to QPX (Ander-
son et al., in press) was isolated from gill tissues of Canadian M.
meirenaria and provided by Mr. G. S. MacCallum and Dr. S.
McGladdery, Gulf Fisheries Center. Moncton. Canada. Like QPX.
C9G was maintained in the medium of Kleinschuster et al. ( 1998)
at 25°C and subcultured at 7 d.
M. mercenaria Plasma
M. mercemiria. collected from the Ware River. VA by a com-
mercial supplier; were maintained with recirculating water (25 ppt.
10. 1 1°C). Hemolymph samples were withdrawn by syringe from
an adductor muscle hemolymph sinus and held on ice in polypro-
pylene tubes. The hemocytes were centrifuged out of suspension
(300 X g. 10 min. 4°C). The pooled supernatant (plasma) was
sterilized by filtration (0.2 |j.m pore size), and assayed for protein
content (BCA kit. Pierce Co.. Rockville. IL). Individual plasma
samples from three to four hard clams were pooled and were
frozen (-20°C) in aliquots. The frozen samples were used soon
because the QPX-modulatory activity declined after -2 mo in stor-
age. In one series of experiments, plasma was heat-treated by
exposure to 65°C for 10 min. the plasma was cooled to room
temperature (~25°C) before use.
Immediate Exposure of Thraustochylrids to Plasma
QPX cells from 7d cultures were washed, as described above,
and resuspended (2.5 x lUVml) in 25 ppt lO. Plasma protein con-
centration was standardized (usually to 0.2 mg/ml ) by dilution with
10 and serial dilutions prepared. Replicate culture flasks for each
protein concentration tested were prepared with experimental (1.9
ml Kleinschuster" s minimal essential medium (KMEM), 0.1 ml
QPX suspension, and 0.5 nil plasma dilution), control (1.9 ml
KMEM. 0. 1 ml QPX suspension, and 0.5 ml lO). and the necessary
blanks. After 7 d incubation at 24°C. the contents of each flask
were removed, and the QPX washed thoroughly and quantified, as
described previously. In related experiments. QPX or C9G were
incubated for 2 h in lO containing plasma, washed, and resus-
pended in KMEM. Percent inhibition was determined using the
following formula:
% inhibition = 1
experimental value
control value
X 100
Delayed Exposure to Plasma
In the delayed exposure experiments. wQPX were permitted to
incubate in KMEM for various time intervals <24 h before expo-
sure to 40 jjLg/ml M. mercenaria plasma proteins. The QPX cells
resumed typical secretory activities during these pre -exposure pe-
riods, as seen by microscopic examination. This plasma protein
concentration was selected because it had been shown in previous
immediate exposure experiments to inhibit -95% of the growth of
QPX cultures.
Viability Assays
QPX viability tests were carried out using viability/cytotoxicity
kit #1 (Molecular Probes, Eugene, OR). The test is based on the
differential permeability of live and dead cells to a pair of fluo-
rescent stains. Cell populations exposed simultaneously to both
dyes become differentially stained: live cells are stained green and
dead cells appear red. This assay was used to check wQPX viabil-
ity after exposure to 10 or M. mercenaria plasma.
RESULTS
Effects of M. mercenaria Plasma on Washed QPX
At the lower plasma concentrations tested, inhibition was low
and variable, with some pools actually stimulating growth (Fig. 1 ).
However, at plasma protein concentrations s 10-50 jxg/ml, a dose-
dependent inhibition was consistently recorded (-100% inhibition
was seen at >50 (j.g/ml). The inhibitory EC^,, was calculated to be
-19 p-g/nil. When this procedure was carried out with heat-treated
(65"C. 10 min) plasma, the stimulatory effects of the lower con-
centrations were not evident (Fig. 2). The inhibitory EC^,, for
heated plasma was -32 (xg/ml; therefore, this heat treatment only
partially inactivated (-40%) the growth inhibitory factor(s).
The inhibitory effects of M, mercenaria plasma were exerted in
a short period. When wQPX were exposed to 40 p.g/ml plasma for
2 h, washed free of plasma and cultured for 7 d in plasma-free
medium, the resultant QPX cell numbers were 80.7 ± 13.3% (n =
3) reduced as compared with untreated controls. A similar degree
of inhibition (94.3 ± 5.\%. n = 4) was seen when 40 |j,g/ml
plasma was left in the medium for the entire duration of the assay.
No significant difference was found between these means by way
of a 2-tailed, unpaired Mest. The inhibition produced by 2-h ex-
posure of wQPX to 40 p,g/ml plasma protein did not result from
QPX-cidal activity. Plasma-treated and untreated wQPX were
similar (treated: 94.0 + 1.7%r, n = 3; and untreated: 94.0 ± 3.0%-,
// = 3 viable). A degree of specificity for M. mercenaria plasma
is also indicted because exposure of wQPX to 40 |xg/ml produced
>90%i inhibition, whereas under the same conditions. C9G was
minimally inhibited (Fig. 3).
Reactions of M. mercenaria plasma with mucus-enveloped QPX
The typical response obtained by exposing wQPX immediately
to M. mercenaria plasma (Fig. 1) was not seen after comparable
100
75
50
25
-25
-50
-75
-100
c
o
10 20 30 40
— I 1
50 60
Protein Cone. (Mg/ml)
Figure 1. QPX-modulatory activity of M. mercenaria plasma ex-
pressed as percent inhibition of cultures after 7 d incubation. Final
plasma protein concentration in tlie medium is indicated. Linear re-
gression ly = lll..^[log xl - n.M: r- = 0.7497) of log-transformed
concentrations was used to calculate tlie inhibitory EC;,, = 18.99 (ig/ml.
QPX Mucoid Secretions
207
C
o
!c
c
100 n
75
50
25
-25
-50 H
-75
■100
10 20 30 40 50 60
Protein Cone, (pg/ml)
Figure 2. QPX-modulatory activity of heat-treated (65'C. 10 niin) M.
inerccnaria plasma expressed as in Figure I. Linear regression (\ =
48.22|log \| - 22.71; r" = 0.67151 of log-transfornied concentrations
was used to calculate the inhibitory V.C=,„ = 32.20 Mg/ml.
exposure of vvQPX that was incubated for 24 h before the addition
of plasma (Fig. 4). The lowest dose tested (3.75 |jLg/ml) apparently
produced some inhibition, whereas all other doses (:£60 jjig/ml)
seemed to stimulate the QPX cultures. The apparent inhibition
produced by the lowest concentration tested was not significantly
different from zero (P > 0.05. one sample Mest. 2-tailed). The
higher concentrations tested were all stimulatory. (P > 0.05. one
sample ;-test. 2-tailed). wQPX cells were either immediately ex-
posed to a highly inhibitory plasma concentration (40 |jig/ml) or
allowed to incubate in plasma-free medium for 2-24 h before
exposure; in these delayed exposure experiments, a time-
dependent linear decrease in growth inhibition was ob.served (Fig.
5 1. Unlike QPX. C9G cells in culture secreted no mucoid material
visible in preparations examined under the microscope. Preincu-
bation of washed CQG cells for 24 h before exposure to 40 p.g/ml
plasma had no significant protective effect as compared with cells
immediately exposed.
100
c
o
c
75-
50-
25
C9G
QPX
Figure 3. The effects of 40 pg/nil M. merciiiaria plasma proteins on
growth of 7 d cultures of QPX and C9G, a closely related thraus-
tochytrid also isolated from hard clams. The protists were exposed to
the plasma for 2 h. washed, and cultivated (7 d) in KMEM. Mean
percent inhibition and standard deviations are indicated.
o
luu-
•
•
n
•
u
t
•
•
100
•
•
•
•
•
•
•
•
•
•
200
r-
•
r
10 20 30 40 50 60
Protein Cone, (pg/ml)
Figure 4. (Irowth of wQPX preincubated lor 24 h in plasma-free me-
dium before exposure to 40 (ig/ml M. menenaria plasma. The QPX
cells continuously secreted mucoid material during the preincubation
period.
DISCUSSION
When wQPX cells were introduced into media containing vari-
ous concentrations of M. meixenahu plasma, their subsequent
growth was altered according to plasma concentration. This may
be seen in Figure 1 where 7 d QPX culture growth was often
stimulated in the presence of low plasma levels but consistently
suppressed at >I0 jjig/ml. These effects could be explained by the
presence of two QPX-modulatory agents in the plasma. Stimula-
tion at low protein levels might be caused by a factor with high
QPX-affinity and low to moderate activity. The effect of this
stimulator would be lost at higher protein levels if a low QPX-
affinity. higher activity inhibitor were present. The presence of two
growth modulators was also suggested by the differences in ther-
mal sensitivity (Fig. 2). Heat treatment of 65°C for 10 min seemed
to eliminate all stimulatory activity: however, the inhibitory effects
persisted with somewhat reduced activity. The growth modulating
activity of M. menenaria plasma takes place rapidly after inter-
action with wQPX. If wQPX was exposed to an inhibitory con-
centration of plasma (40 |jLg/ml) for 2 h. and then washed free of
plasma proteins before growing the culture in plasma-free me-
dium, culture growth was inhibited to about the same extent be-
cause it would have been if the cells had been continuously ex-
100
-25
-50
8 12 16 20 24 28
Time (hrs)
Figure 5. Effects of length of wQPX preincubation before exposure to
40 ng/ml M. menenaria plasma. Mean percent inhibition and standard
deviations are indicated.
208
Anderson et al.
posed to 40 ^Lg/ml plasma. These experiments could not establish
whether the inhibitory effects produced by M. mercenaria plasma
on the cell density of 7 d QPX cultures were caused by growth
inhibition or by cidal activity. Direct killing was ruled out by the
fact that 40 |xg/ml exposed, (potentially highly inhibited) wQPX
and unexposed wQPX were -95% viable.
Figure 3 presents evidence that the QPX-inhibilory plasma fac-
tor shows target specificity; C9G growth was hardly affected by 40
p.g/ml. Sequence analysis of C9G placed it in the thraustochytrid
phylogenetic group as a sister taxon to Thnnistocliytiium pachy-
denniim. and these sequences were grouped with QPX with a
parsimony jackknife support value of 100 (Anderson el al. in
press). Clearly. QPX sensitivity to low (<40 |xg/ml) plasma con-
centrations exceeds that of C9G; however, C9G growth was in-
hibited (-60%) by exposure to >180 |Jig/ml plasma (Anderson et
al. in press). Because the pathogenicity of C9G for M. Dierceinirici
has yet to be established, it is not known whether inhibition dif-
ferences caused by clam plasma between QPX and C9G reflect
differences in pathogenicity.
Incubation of wQPX in plasma-free medium allowed the cells
to resume mucus secretion. The cells underwent minimal division
for the first 48 h in culture, then proceeded lo grow w ith a doubling
time of -3 d (QPX growth curve not shown). The wQPX cells
were suspended in a loose gelatinous mass by 24 h. This mucoid
secretion often infiltrated the entire culture medium by 7 d in
culture. When the cells were permitted to develop their mucoid
covering for 24 h before the addition of plasma (Fig. 4). concen-
tration of -7-60 p-g/ml failed to inhibit QPX growth in 7 d cul-
tures. Unexpectedly, the lowest concentration tested (3.75 jjig/ml)
seemed to have inhibitory activity, but the mean of these experi-
mental values were not significantly different from zero. These
data suggested that the mucus material might protect QPX from A/.
iiicrceiuiria humoral defense mechanisms such as antimicrobial
factors. This hypothesis was supported by the results of the de-
layed exposure experiments, where protection from growth inhi-
bition was dependent on the time of incubation before exposure to
40 |j.g/ml plasma protein (Fig. 5). Because QPX cells in clam
tissues are typically enveloped by mucus, a role of this secretion as
a virulence factor seems likely. This is supported by a recent report
that clams injected with wQPX did not develop infections or dis-
ease (Smolowitz et al. 2001).
ACKNOWLKDGMENTS
This study was supported by Maryland Sea Grant, NOAA,
grant number NA06RG010I. This is Contribution No. 3642 of the
University of Maryland Center for En\ ironmental Science, Chesa-
peake Biological Laboratory.
LITERATURE CITED
Anderson. R. S.. B. S. Kruus, S. E. McGladdery, K. S. Reece & N. A.
Stokes. 2003. A thraustochytrid protist isolated from Meirciuiria mer-
cenaria: molecular characterization and host defense responses. Fish.
Shellfish Immunol, (in press).
Anderson, R. S. 1987. Polykaryon formalion by Meiccmirin iiicrceiiariu
hemocytes. Biol. Bull. l72:236-:45.
Brothers. C, E. Marks & R. Smolowitz. 2000. Conditions affecting the
growth and zoosporulation of the prollstan parasite QP.X in culture.
Biol. Bull. 199:200-201.
Drinnun. R. E. & E. B. Henderson. 1963. 1962 mortalities and a possible
disease organism in Neguac quahaugs. Annual Rept. No. Bl I. Biologi-
cal Station. St. Andrews, New Brunswick.
Hanks, J. H. & J. H. Wallace. 195S. Determination of cell viability. Proc.
Soc. E.\p. Biol. Med. 98:188-192.
Kleinschuster, S. J.. R. Smolowitz & J. Parent. 1998. /;; vino life-cycle and
propogation of quahaug parasite unknown. J. Shellfish Res. 17:75-78.
Maas, P. A. Y., S. J. Kleinschuster, M. J. Dykstra, R. Smolowitz & J.
Parent. 1999. Molecular characterization of QPX (quahog parasite un-
known), a pathogen of Mercenaria mercenaria. J. Shellfish Res. 18:
561-567.
Ragan. M. A., G. S. MacCallum, C. A. Muiphy, J. J. Cannone. R. R. Gutell
& S. E. McGladdery. 2000. Protistan parasite QPX of hard-shell clam
Mercenaria mercenaria is a member of Lahyrinthulomycota. Dis.
.Aqual. Org. 42:185-190.
Ragone Calvo. L. M.. J. G. Walker & E. M. Burreson. 1998. Prexalence
and distribution of QPX. Quahog Parasite Unknown, in hard clams
Mercenaria mercenaria in Virginia, USA. Dis, Aquat. Org. 33:209-
219.
Smolowitz. R.. D. Leavitt. B. Lancaster. E. Marks. R. Hanselmann & C.
Brothers. 2001. Laboratory based transmission studies of quahog para-
site unknown (QPX) in Mercenaria mercenaria. J. Shellfish Res. 20:
555.
Smolowitz, R., D. Leavitt & F. Perkins. 1998. Observations of a protistan
disease similar to QPX in Mercenaria mercenaria (hard clams) from
the coast of Massachusetts. J. Inverlehr. Pathol. 71:9-25.
Stokes. N. A.. L. M. Ragone Calvo. K. S. Reece & E. M. Bun-eson. 2002.
Molecular diagnostics, field validation, and phylogenetic analysis of
Quahog Parasite Unknown (QPX), a pathogen of the hard clam, Mer-
cenaria mercenaria. Dis. Aquat. Org. 52:233-247.
Whyte, S. K.. R. J. Cawthom & S. E. McGladdery. 1994. QPX (quahaug
parasite X), a pathogen of northern quahaug Mercenaria mercenaria
from the Gulf of St. Lawrence. Canada. Dis. Aquat. Org. 19:129-136.
J.Hinuil ofSlwllJhh Re.scanh. Vol. 22. No. 1. 209-212, 200.^.
A PORTABLE AND PRACTICAL METHOD TO MONITOR BIVALVE FEEDING ACTIVITY IN
THE FIELD USING TIME-LAPSE VIDEO TECHNOLOGY
BRl'CE A. MACDONALD* AND LISA M. NODVVELL
Dcpcinniciit of Biology. Centre for Coiisral Stiullcs and Aquaciihnrc. University of New Briins\vu± Saint
John. P. O. Box 5050 Saint Jolin. New Brnnswicl<. Canada. E2L 4L5
ABSTRACT We developed a simple iiielhod to measure leeding activity of Mylilii.s filiilis using a canicorder placed inside an
underwater housing, a plastic frame for holding mussels and time lapse videography. Exhalant siphon area, indicative of feeding
activity, was monitored in laboratory mussels exposed to filtered seawater and various concentrations of microalgae, including Pavlova
lulheri or TetraseUnis suecica. Exhalant siphon area increased as algal concentration increased from zero to -25-30 x 10' cells ml"',
hut declined again at higher concentrations. Advantages of this method include portability and relatively low cost, high resolution of
data over shon and long temporal scales, potentially large sample sizes, and minimum logistics required for deployment in a variety
of different environments. Once relationships between exhalant siphon area and other indicators of feeding such as filtration rate have
been established, this method could greatly miprove our understanding of bivalve feeding in situ and how they respond in dynamic
natural conditions.
KEY WORDS: Mvrilus etliilis. bivalve feeding, time-lapse recording, exhalant siphon area, particle concentration
INTRODUCTION
There have been numerous studies on measuring feeding ac-
tivity in a variety of suspension-feeding bivalves over the last
several decades. There has recently been much discussion and
debate on whether or not bivalves have the capability of physi-
ological regulation or are pumping at full capacity all the time
(Jorgensen 1996, Bayne 1998, Hawkins et al. 2001 ). This includes
numerous comments on the proper interpretation of the published
literature and diverse opinions on the reliability of some of the
methods used (Cranford 2001. Riisgard 2001. Widdows 2001),
One such method considered to have good potential for assess-
ing feeding activity remotely with little interference by the ob-
server and minimal disturbance to the bivalve is the estimation of
valve gape and siphon area in mussels (Newell et al. 2001 ). Posi-
tive relationships have been reported between pumping rates of
mussels, valve gape and the exhalant siphon area (Jorgensen I960.
Riisgard & Randlov 1981. Famme et al. 1986. j0rgensen et al.
1988. Jorgensen 1990) and between exhalant siphon area and mus-
sel filtration rates (Newell et al. 2001).
Filtration rates of mussels have been shown to be linked to
particle concentration with low levels observed for filtered water
but increasing with natural levels of seston before decreasing again
at higher seston loads (Foster-Smith 197.5. Winter 197.^. Bayne
1993). Riisgard and Randl0v ( 1981 ) found comparable reductions
in filtration rates and valve gape of blue mussels at densities of
Plmeodactylum trieonmutwn lower than 1.500 cells ml" and
higher than .W.OOO cells ml"'. Newell et al. (2001) found a similar
apparent threshold for the filtration response to particle concen-
tration to occur at 2.000-6,000 particles ml"' in a Hume environ-
ment. Dolmer (2000a, 2000b) observed that high algal concentra-
tions may lead to decreases in valve gape as well as estimates of
filtration in the field.
There is ample evidence to suggest that exhalant siphon area is
a useful indicator of feeding activity in mussels and it is responsive
to variations in the concentration of suspended particles. The pur-
pose of this study was to develop a ponable and reliable method to
*Corresponding author. E-mail: bmacdon@unbsj.ca; Fax: +1-5U6-648-581 1.
remotely estimate exhalant siphon area for numerous undisturbed
mussels simultaneously. It would be particularly advantageous if
the method could be deployed to the field where mussel response
could be continuously evaluated while natural seston and flow
conditions are monitored. The combination of time lapse capabili-
ties and high resolution image of a digital camcorder, a portable
underwater housing, a plastic frame for holding mussels, and
readily available image analysis software provides an effective
tool for studying mussel feeding activity. Exhalant siphon area was
monitored in this study in mussels exposed to various concentra-
tions of cultured microalgae in the laboratory en\ ironment.
MATERIALS AND METHODS
Mussels [Mytilus edidis Linnaeus 1758) were collected from an
inlet in the Pasamoquoddy Bay. New Brunswick and transported to
University of New Brunswick in Saint John. New Brunswick,
Canada. Mussels were acclimated to laboratory conditions for a
minimum of 2 d and a maximum of 7 d. Experiments were per-
formed in a 530 I (244 cm long, 66 cm wide, and 33 cm deep) tank
with well mixed recirculating seawater. flowing approximately
5-10 cm s"'. Experiments were performed in full room light and
temperature and salinity were maintained at I2°C and 35-36'^f.
respectively. Water was prc-filtered in the tank with three inline
filters of 20. 5. and I p.m. Mussels were exposed to filtered sea-
water and cultured microalgae ranging in initial concentration
from 5.000-85.000 cells ml"' while siphon area was monitored
over periods of hours using time-lapse videography. Mussels were
exposed to experimental conditions for 30-60 min prior to mea-
surements to ensure feeding activity had resumed. With a few
exceptions experiments for each series of mussels typically ran tor
2_t h to ensure a good time series of measurements and a detect-
able change in particle concentration. Algal concentration was
measured using an electronic particle counter (Coulter Multisizer
II) with a 100 p-m tube orifice diameter. Algal diets provided in
experiments were one of Pavlova lutlieri (Provasoli-Guillard
CCMP1325) or TetraseUnis suecica (Provasoli-Guillard
CCMP904) or a mussel spat formula of Nanocliloropsis ocidata.
Chaetoceros-B, and Phaeodaelyltim iricorniaum (Innovative
Aquaculture Products Ltd.).
209
210
MacDonald and Nodwell
At least one day prior to the experiments Velcro was attached
to the mussel shell using cyanoacrylate cement and. after drying,
mussels were attached to individual plastic posts also covered in
velero. The posts containing the mussels were secured to a plastic
plate and attached to a frame connected near the lens of a video
recording device (Fig. 1 A). The number of mussels observed (usu-
ally 9-12 adults) in the video frame depended on the size of the
mussels and the efficiency of arranging mussels to adequately
view the external siphon. A Sony Mini DV (model DCR-TRV900)
three ccd camcorder was enclosed in an Amphibico 900 underwa-
ter housing and set to an interval recording mode of 2 s every
30 s over the entire period of each experiment to capture siphon
activity.
Multiple images from the mini DV tapes were collected using
the photo feature of the camcorder and stored on memory cards
before being transferred to a personal computer (Fig. IB). Varia-
tion in siphon area was estimated for individual mussels using the
program Image J (NIH public domain Java image processing pro-
gram — URL: http://rsb.info.nih.gov/ij). Siphon area was calibrated
using a 1 cm mark on the mussel posts. The inherent variation in
measuring exhalant siphon area was 2.4-3.8%. To standardize
individual responses for different sizes of mussels to different algal
concentrations, exhalant siphon area data were converted to per-
cent of maximum values observed for each mussel.
RESULTS
There was a consistent decline in algae over time in all the
experiments, indicating removal of microalgae by the mussels in
the course of the experiments (Fig. 2). Exhalant siphons were
opened, confirming feeding activity by the mussels. The fitted
lines for the uptake rates of algae typically had r" values exceeding
0.90-0.95 in all examples.
The percent maximum exhalant siphon area in individual mus-
sels exposed to filtered seawater (no algae) was consistently lower
than the siphon areas reported for the same mussels exposed to
microalgae (Fig. 3A). A similar trend of greater exhalant siphon
area was akso observed for groups of mussels exposed to different
concentrations of microalgae compared to those held in filtered
seawater (Fig. 3B). Note that mus.sel exhalant siphon area was still
approximately 20-309f of the maximum when exposed to filtered
seawater.
The percent maximum exhalant siphon area in mussels in-
creased with increasing particle concentrations to a maximum of
near 90-95% at concentrations approaching 25-30.000 cells mP'
(Fig. 4) Further exposure to concentrations above 30.000 cells
ml"' resulted in a decline in percent maximum exhalant siphon
area.
DISCUSSION
By modifying an underwater housing and combining it with a
high resolution camcorder capable of time-lapse videography we
have developed a simple and relatively inexpensive method to
remotely study bivalve feeding behavior. There have been other
devices developed to remotely monitor bivalve activity but, for
various reasons, they have not been readily adopted by scientists
working on bivalves. This includes The Musselmonitor* devel-
oped as a biological early warning system containing sensors to
record shell opening and closing while mussels are exposed to
various pollutants (Baldwin & Kramer 1994). Manuel and Lob-
siger ( 1999) de\eloped the MarineCanary'^' as a biomonitoring
tool using an underwater camera and a time-lapse system to assess
the marine environment through changes in bivalves" valve gape
and mantle activity.
Using this new method we have established a positive relation-
ship between exhalant siphon area and the concentration of cul-
tured microalgae, also observed by Newell et al. (2001) in their
study. Feeding activity is this study was confirmed by the con-
tinuous decline in the concentration of microalgae in the experi-
y = -976.76x + 6671
R' = 0.9487
Figure I. {\). .\n adjustable plastic frame attached to the front of an
underwater video housing containing a high resolution camcorder
with time-lapse capabilities. Mussels are secured with \ elcro to move-
able posts inserted into a plate positioned in front of the video lens. (B)
A tvpicai black and while photo made from a video frame captured
from the mini DV tape. Exhalant siphons are clearly visible for several
mussels simultaneously.
Elapsed Time (h)
Figure 2. An example of variation in declining algal concentration,
attributable to mussel feeding, during a typical medium — low concen-
tration experiment.
Time-Lapse Video Technique to Estimate Mussel Feeding
w
<
e
o
a
R
0)
Mussel #2 Mussel #3 Mussel #5 Mussel #8
100
■ No Algae
=1 10-20000 cells ml
B 20-30000 cells ml
n 30-45000 cells ml"'
*
i
Figure 3. (A) Variation in individual mean percent maximum exhal-
ant siphon area of representative mussels held in filtered sea« ater and
exposed to microalgae in different algal concentrations (5-45.000 cells
nil"'). (B) Mean response for groups of mussels exposed to filtered
seawater and three different experimental concentrations of microal-
gae (5—15.000 cells ml"'). Values are means ± 1 SE.
mental tanks (Fig. 2). The shape of the line when fitted to semi-log
transformed data (i.e.. rate of clearance) was comparable to the
reduction observed by Riisgard (1991) when Myliliis edutis was
grazing on Rhodomonas baltica. The positive relationship between
particle concentration and exhalant siphon area was apparent until
concentrations reached 25.000-30.000 cells ml"' and exhalant si-
phon area appeared to decrease with further increases in concen-
tration (Fig. 3B). Clausen and Riisgard (1996) also observed that
mussels partly closed their valves and reduced the opening of the
exhalant siphon at high algal concentrations but they found this
reduction to occur at around 13-24.000 cells ml"'. Note that we
did observe some moderately high values for siphon area for mus-
sels at the highest algal concentrations. This may have been an
artefact of the experimental design where a group of starved mus-
sels were exposed initially to \ery high concentrations of microal-
gae.
There are several advantages to the time-lapse videography
method for the observation of feeding activity in bivalves. This
includes its size. cost, portability and readily available components
including public domain software. A variety of underwater hous-
ings are available today for most commercial camcorders capable
of using time-lapse technology. Because of the small size of the
housing, they can be placed unattended in a wide variety of habi-
tats for extended periods of time — up to 10-15 hours with the new
3 <
It
(/)
fli TO
LU
100
90
SO
70
60
50
40
30
20
10
z
i
i
s
10000 30000 .50000 70000
Algal concentration (cells ml'^)
Figure 4. Variation in percent maximum exhalant siphon area of mus-
sels exposed to different concentrations of microalgae. The closed dia-
mond represents an ex 
TVNews
OpenLibrary