Marine Biology (1996) 125:735-742 9 Springer-Verlag 1996

M . A. T o d a r o 9 J. W . F l e e g e r 9 Y. P. H u
A. W . Hrincevich 9 D . W . F o l t z

Are meiofaunal species cosmopolitan? Morphological and molecular

analysis of Xenotrichula intermedia (Sastrotricha: Chaetonotida)

Received: 3 January 1995/Accepted: 8 February 1996

Abstract Many meiofaunal species are reported to be tions based on morphology suggest that individuals
cosmopolitan, but due to uncertainties of identification, characterized by different haplotypes are genetically
the affiliation of specimens from geographically distant isolated sibling species.
areas to the same species-taxon is problematic. In this
study, we examined morphological and molecular vari-
ation in samples of Xenotrichula intermedia Remane Introduction
(Gastrotricha: Chaetonotida) from the Mediterranean
Sea, the northwestern Atlantic and the northern Gulf of Meiofaunal species are defined as cosmopolitan if they
Mexico. Univariate analysis of 16 morphological traits
are reported from two or more oceans, including con-
was unable to detect differences among populations,
nected seas (Sterrer 1973). Within the meiobenthos,
except for the length of the pharynx, which was signifi- cosmopolitan species have been reported from a wide
cantly shorter in the Gulf of Mexico specimens. Ca- array of major taxa displaying a broad range of life
nonical discriminant analysis separated the Gulf of styles (Nematoda: Gerlach 1962; Ostracoda: Hulings
Mexico specimens from the other two populations,
1971; Polychaeta: Westheide 1971; Tardigrada: Re-
with pharynx length contributing about half of the
naud-Mornant and Pollock 1971; Harpacticoida: Wells
total discrimination. Molecular analysis based on re-
1986; Gastrotricha: Hummon et al. 1994; Gnathos-
striction-fragment length polymorphisms (RFLPs) in
tomulida: Sterrer 1973). Although frequently reported,
a 710-base pair polymerase chain-reaction (PCR) prod-
the high proportion of cosmopolitan species has been
uct representing roughly half of the mitochondrial
questioned, because (1) distributional records usually
cytochrome oxidase I (COl) gene detected four haplo-
come from different sources at different times, and (2)
types: one each from the Mediterranean and the Gulf of
the number of specimens examined is generally too low
Mexico populations and two coexisting within the At-
to allow meaningful comparisons within and among
lantic population. The estimated nucleotide-sequence
populations (see also Sterrer 1973). Indeed, because of
divergence calculated for each pairwise combination of
shared life-history traits such as a short life cycle, a rela-
haplotypes (based on the proportion of shared frag-
tively small number of offspring, the general absence of
ments) ranged from 5.3 to 11.5%. The high genetic
a pelagic larval stage, and a relatively limited swim-
divergence and the inability to clearly separate popula-
ming ability of adults, meiofaunal species (particularly
interstitial forms) would be expected to have restricted
geographical ranges.
Communicated by N. H. Marcus, Tallahassee A central point of the debate over the presumed
M. A. Todaro ~ . J. W. Fleeger (12~) cosmopolitan distribution of meiofaunal taxa concerns
Y. P. Hu 2 . A. W. Hrincevich 9 D. W. Foltz species identification. In particular, critics have ques-
Department of Zoology and Physiology, tioned the reliability of species identifications from geo-
Louisiana State University, Baton Rouge,
Louisiana 70803-1725, USA
graphically distant areas, especially when made by
different investigators using different methods (often
Present addresses: low-resolution microscopy) and probable personal "in-
Dipartimento di Biologia Animale, Universit/t di Modena stincts" to affiliate specimens with a given taxon. In
1-41100 Modena, Italy
2 Center for Theoretical and Applied Genetics, fact, careful morphological analysis has shown that
Cook College of Rutgers University, New Brunswick, some species with a presumed wide geographic range
New Jersey 08903, USA are actually composite assemblages of different species
736

(R. Huys, cited in Giere 1993; Evans 1994). On the (36 ~ 76 ~ and Pensacola Beach, Florida, USA (30 ~
other hand, recent surveys using highly reproducible 87 ~ in February, January and March 1995, respectively; they
techniques (e.g. high-resolution video microscopy) have were kept at ambient temperature (20 ~ and were brought to the
laboratory within one week. In the laboratory, specimens of Xeno-
confirmed that cosmopolitanism appears to be a wide- trichuIa intermedia Remane were immediately extracted from the
spread phenomenon among certain meiofaunal groups sediment by the narcotization-decantation technique (Pfannkuche
(Hummon 1994; Todaro et al. 1995). Because of these and Thiel 1988) using 7% MgC12 aqueous solution, and were pre-
contrasting results and the awareness of the possible pared for analysis. Morphological measurements were conducted on
existence of cryptic species, different approaches to 15 living, relaxed specimens from each population. Individuals were
transferred singly by glass micropipette to a slide, and were observed
species identification are needed to either confirm or using either differential interference-contrast optics with a Micro-
disprove the morphological species concept. Physical phot-FXA Nikon microscope, or phase-contrast using a Wild M 20
characteristics of the interstitial habitat have prevented microscope. Several gastrotrichs from each location were photo-
direct observations of behavioral/ecological differences graphed to serve as voucher specimens (Fig. i). Measurements of 16
combined metric and meristic traits (Table 1) were obtained with
among geographically distant populations. Moreover, an ocular micrometer, or from photographs. Comparisons among
small body size and the scarcity of specimens from the three populations were performed using both univariate analysis
single sites have, in most cases, prevented a more accu- (one-way ANOVA) and multivariate analysis (canonical dis-
rate species identification based on a genetic approach. criminant analysis and principal-components analysis) in the
The polymerase chain reaction (PCR: Saiki et al. 1988) Statistical Analysis System (SAS Version 6: SAS Institute Inc. 1990).
Canonical discriminant analysis requires that individuals be pre-
allows one to amplify large quantities of specific D N A classified into groups (here, "populations"), and is designed to reveal
sequences from very small initial amounts of template the linear combination of the original variables that maximizes the
(i.e., Harris et al. 1990). The PCR products can between-group variance. In contrast, principal-components analysis
subsequently be analyzed for nucleotide sequence does not require any prior assumption about group membership, so
it has the limitation that within-group variance is confounded with
differences and/or restriction-fragment length poly-
between-group variance.
morphisms (i.e., G/~rate et al. 1991; Karl and Avise For DNA analysis, 25 living specimens from each population were
1993; Castagnone-Sereno et al. 1994; Geller et al. 1994; extracted, rapidly rinsed in deionized water, and transferred singly
Litvaitis et al. 1994). by either a micropipette or an Irwin loop to sterile 500 gl microfuge
Here we focus on Gastrotricha (using both mor- tubes containing 4 itl PCR buffer (Promega, Madison, Wisconsin).
They were stored at - 70 ~ for future DNA amplification. Within
phological and molecular genetic analysis) to examine 3 wk, the gastrotrichs were thawed, homogenized and boiled at
the paradoxical existence of cosmopolitan meiofaunal 95 ~ for 10 min to release DNA. For each specimen, 1 gl of the
species. Species of this phylum are ubiquitous in unpol- homogenate was transferred to a sterile 500 gl microfuge tube and
luted sandy habitats, where gastrotrichs often rank mixed with 30 pl PCR reaction mix. Each 30 ~1 reaction consisted of
second or third in abundance after nematodes and 1 U Taq polymerase (Promega), 3 gl 10 x PCR buffer, 3 gl 25 m M
MgC12 (both solutions supplied with the polymerase), 0.6 gl of each
harpacticoid copepods (Coull 1988). A fairly high num- primer [1.25 mM, LCOI1490 and HCOI2198: Folmer et al. (1994);
ber of gastrotrichs are reported as multi-regional cos- the numbers refer to the primer binding-locations relative to the
mopolitans (i.e., Ruppert 1977; H u m m o n et al. 1994; Drosophila yakuba mtDNA sequence of Clary and Wolstenholme
Todaro et al. 1995). We investigated Xenotrichula inter- (1985)], 4.8 gl dNTPs (total 5 raM, New England Biolabs, Beverly,
media Remane, a small interstitial chaetonotid charac- Massachusetts), and 18~1 sterile (autoclaved) distilled water.
Reactions were amplified through 50 cycles in a Perkin-Elmer
teristic of (but not limited to) the intertidal zone of DNA thermal cycler using the following parameters: denaturation
high-energy beaches, that is frequently reported in at 95 ~ for 1 rain, annealing at 40 ~ for 1 min, extension at 72 ~
faunal surveys of coastal areas in northern Europe, for 1 min. Amplification was confirmed by electrophoresing 5 gl
across the Mediterranean Sea, the Atlantic and Gulf of each PCR product and appropriate negative controls in a 2%
agarose gel containing 1 x tris-borate-EDTA buffer with a 1 kbase
coasts of the USA, and also from India and Somalia DNA size-ladder (GibcoBRL, Gaithersburg, Maryland). Gels
(i.e., Ganapati and Rao 1967; Balsamo et al. 1992; Jouk were stained with ethidium bromide, and bands were visualized
et al. 1992; Todaro et al. 1995). We compared speci- on a UV transilluminator. When additional DNA was needed,
mens from the Mediterranean Sea, the northwestern 1:100 dilutions in sterile water of the initial PCR product were
Atlantic, and the northern Gulf of Mexico. At the used as a template in subsequent amplifications. PCR amplification
of specimens from Florida was less successful than for other loca-
morphological level we measured 16 traits, whereas at tions, and reamplification of the initial PCR product was also
the molecular level we focused on a portion of the difficult for this material. The reasons for this difference are un-
cytochrorne 0xidase I (COI) mitochondrial gene. Gen- known.
etic data on gastrotrichs are virtually non-existent (see Possible differences in the nucleotide sequence within and among
Balsamo and Manicardi 1995), so this gene was chosen populations were examined by digesting separate 8 to 24 gl aliquots
of individual PCR products with 11 restriction enzymes having 4 to
for analysis without prior knowledge of the amount of 6 base recognition-sequences (AseI, AvaII, CfoI, DdeI, HinfI, MseI,
intra- and inter-specific variation in these species. NlaIII, RsaI, ScrFI, TaqI, XbaI; protocol according to New Eng-
land Biolabs, Beverly, Massachusetts). Restriction fragments were
separated by electrophoresis in 15% acrylamide gels, and were
visualized on a UV transilluminator as for the uncut PCR product.
Materials and methods Fragment sizes were estimated by log-linear regression of mobilities
relative to the DNA ladder (DraI-digested lambda DNA, Gib-
Sandy sediments were collected from the intertidal of Petacciato coBRL). Preliminary comparisons between the banding pattern
Marina, Molise, Italy (41 ~ l l'N; 14~ 51'E), Norfolk, Virginia, USA generated from whole versus purified PCR product (purifying kit,
 737

Fig. 1 Xenotrichula intermedia

(specimen from Italian
population). A dorsal view;
B internal anatomy; C ventral
view (Scale bar 50 gin)

Table 1 Xenotrichula intermedia. Summary statistics for 16 morphometric variables mean • 1SD range in three populations (measurements
in txm; N = 15 for each collection)

Variable Italy Virginia Florida

mean • 1SD(range) mean _+ 1SD(range) mean _+ 1SD(range)

Total body length 203.5 i 8.2 (188.0-219.0) 202.3 • 4.4 (196.0-210.0) 203.0 • 12.1 (185.0-228.0)
Pharynx length 52.0 _+ 1.1 (48.0-59.6) 51.0 _+ 2.0 (47.0-54.0) 45.6 _+ 2.3 (45.0-52.0)
Furca length 34.7 • 1.1 (31.6-36.0) 34.3 • 1.1 (33.0-37.0) 34.6 • 1.8 (29.6-38.0)
Adhesive tube length 14.8 • 0.7 (13.7 19.0) 15,0 _+ 0.6 (13,9-19.0) 15.7 • 1.6 (13.6-18.0)
Pharynx anterior diameter a 9.1 • 1.0 (7.8 11.0) 10.0 • 0.5 (8.5-11.0) 9.1 + 0.9 (7.0-10.3)
Pharynx middle diameter 75 • 0.5 (6.6-8.5) 8.5 • 0.8 (6.0-9.2) 7.4 • 0.7 (6.0-9.0)
0
Pharynx posterior diameter 9.9 • 1.0 (7.8 11.1) 10.5 _+ 0.5 (10.0-11.2) 10.0 • 0.9 (8.2-11.2)
Head width 34.0 • 2.4 (31.3-41.1) 33.2 _+ 1.6 (30.8 36.0) 34.5 _ 3.7 (28.5-43.0)
Neck width 26.9 ___1.0 (25.4-28.4) 27.9 • 1.1 (26.0-29.4) 27.1 • 2.7 (21.6-31.4)
Trunk width 48.4 _+ 1.0 (47.0-50.0) 48.8 • 1.5 (45.0 51.0) 52.4 • 6.4 (41.1-70.6)
Base furca width 23.4 • 1.3 (21.0-25.5) 23.4 • 1.3 (22.0-27.0) 23.1 • 1.4 (21.5- 25.0)
Anus distance from
indentation between furcal branches 24.7 • 1.4 (22.2-28.0) 24.2 • 0.6 (23.2-25.0) 24.1 _+ 1.4 (20.8-26.0)
Mouth diameter 4.1 • 0.1 (3.94.3) 4.1 • 0.1 (4.04.3) 4.2 • 0.4 (3.94.8)
No. of dorsal columns of scales 19.9 • 1.0 (19.0-21.0) 20.1 _ 1.0 (19.0-21.0) 19.3 • 1.3 (17.0-19.0)
No. of scales in dorsal median column 52.3 • 1.4 (50.0-55.0) 51.5 • 1.4 (49.0-54.0) 52,5 _+ 1.1 (50.0-54.0)
No. of scales covering
inner margin of furca 5.0 _+ 0.0 (5.0-5.0) 5.0 _+ 0.0 (5.0-5.0) 5.0 • 0.0 (5.0-5.0)

a Differences among populations statistically significant at P < 0.05 (ANOVA)

Wizard T M PCR Preps System, from Promega, Madison, Wiscon- any two haplotypes x and y, the proportion of shared fragments, F~y,
sin) yielded no differences, so we used mostly whole PCR products in was estimated and then used to obtain the similarity coefficient G~r
our analysis. by successive approximation. The following formula was then used
Estimated sequence divergence at the nucleotide-site level (d), was to estimate dxr:
calculated for each pairwise combination of haplotypes as per Nei
(1987), using the shared-fragment approach (Nei and Li 1979). For d~y = - (2/r) loge(Gx,),
738

where r = the average number of nucleotides per recognition se-

quence (here, 4.57).

Results

Fully mature (hermaphroditic) specimens of Xenot-

richula intermedia ranged from 185 to 228 gm in total
body length. The smallest and largest values were both
found within the Florida population (Table 1). The
furcal appendages varied from 29.6 to 38.0 ~tm in
length, with extremes again among the Florida speci-
mens. The longest adhesive tubes (19.0 gm) were among
the Italian specimens, and the shortest (13.6 gm) among
the Florida specimens. Pharynx length varied from
45.0 gm (Florida worms) to 59.2 gm (Italian worms). Fig. 3 Xenotrichula intermedia. Representative PCR-RFLP (poly-
Regardless of the region of origin, the cuticular cover- merase chain-reaction-restriction-fragment length polymorphisms)
ing of the specimens consisted of typical pedunculated banding patterns after separate digestion with XbaI, ScrFI, RsaI and
scales, arranged on the dorsal side in 17 to 21 columns NlaIII and electrophoresis in 15% acrylamide gels containing 1 x
Tris-borate-EDTA buffer [Numbers on ordinate sizes of mobility
containing 50 to 54 scales each. Univariate statistical standard, in base pairs; U amplified uncut PCR product; M marker
analysis (ANOVA) performed on each of the above (DraI-digested lambda DNA); B Virginia B haplotype; A Virginia
characters, as well as on 10 additional morphometric A haplotype; I Italian haplotype] Fragments of < 40 base pairs
parameters (Table 1), could not detect significant differ- could not be visualized
ences among specimens of the three populations, except
for the length of the pharynx. The pharynx was signifi-
cantly (P < 0.05) shorter in the Florida specimens components analysis could not readily distinguish
(mean = 48.7 gm) than either the Italian (52.3 gm) or individuals from different locations (i.e., 10 principal
Virginian (51.0 gm) specimens. Canonical discriminant components were required to explain 89% of the vari-
analysis separated the Florida specimens from the ance).
others (Fig. 2), with pharynx length alone accounting The cytochrome oxidase I gene was successfully am-
for roughly half of the total discrimination (univariate plified from all 25 Italian specimens, 20 of 25 Virginian
r 2 =0.331 for pharynx length versus multivariate specimens, and 2 of 25 Florida specimens. Among the
r 2 = 0.623). In contrast to this result, principal- 11 endonucleases used in DNA digestions, one enzyme
(MseI) produced numerous small fragments that were
difficult to size accurately (these were excluded from the
4 analysis), and three enzymes (AvaII, CfoI and TaqI)
9 / 9 Italy produced no cuts within the PCR product. Seven en-
9 Virginia zymes (AseI, DdeI, HinfI, NlaIII, RsaI, ScrFI and XbaI)
3 9
l 9 Florida produced informative restriction-fragment length poly-
morphisms revealing a high degree of variation among
2 the specimens examined (Figs. 3 and 4 and Table 2).
II
(N Each of the fragment-length profiles in Table 2 summed
1 to a value that exceeded the size of the amplified
p,,
product (710 base pairs). Similar discrepancies have
0
0 9 m9 9
also been reported in other P C R - R F L P studies,
9 9
mml
9 9 including some in which the estimated restriction-
-1
9 9 mi fragment size exceeded the size predicted from direct
9 9
9 9 sequencing of the same product (Hrincevich and Foltz
unpublished observations). These differences are prob-
-2
ably due to the need for heavy loading of DNA in the
experimental lanes for ethidium bromide visualization
-3 I I I E i i i of small fragments. Incomplete digestion cannot ac-
-4 -3 -2 -1 0 1 2 3 4 count for the discrepancy, since within a profile differ-
Canonical 1 ent fragments were not related in size in a manner that
would be consistent with partial digestion.
Fig. 2 Xenotrichula intermedia. Canonical discriminant analysis of
morphological characters of specimens from three geographic areas.
In total, four different fragment profiles (correspond-
Florida specimens separated from others mainly by difference in ing to an equal number of different mtDNA haplo-
pharynx length (see first paragraph of "Results") types) were detected. The Italian and the Florida
 739

Table 3 Xenotrichula intermedia. Number of shared fragments (above

diagonal) and proportion of shared fragments (below diagonal)
among three mtDNA haplotypes. Total number of restriction frag-
ments for each haplotype is on diagonal line

Haplotype I Va Vb

I ~ 4 6
Va 0.296 ~ 1 4 3
Vb 0.500 0.240 ~ 11

Table 4 Xenotrichula intermedia. Nei's (1987) index of similarity

(above diagonal) and estimated sequence divergence (d, below diag-
onal) among three mtDNA haplotypes

Haplotype I Va Vb
Fig. 4 Xenotrichula intermedia, Representative PCR-RFLP banding
patterns after separate digestion with AseI, HinFI and DdeI. Further
details in legend to Fig. 3 I ~ 0.802 0.885
Va 0.097 ~ 0.770
Vb 0.053 0.115

Table 2 Xenotrichula intermedia. Restriction-fragment profiles for

three mtDNA haplotypes following separate digestion with in-
dicated enzyme. Fragment sizes are shown in basepairs and have Discussion
been rounded to nearest 5
Specimens of Xenotrichula intermedia from the three
Enzyme Haplotype
areas (Mediterranean Sea, northwestern Atlantic, and
I Va Vb northern Gulf of Mexico) looked remarkably alike.
Morphological traits varied little both within and
AseI 740 665, 95 740 among populations. With the exception of pharynx
DdeI 595, 200, 50 665, 200 740, 50 length, differences among populations were statistically
HinfI 730, 60 530, 375 765
NlaIII 715, 50, 40 715, 130 715, 130 non-significant (Table 1), and only pharynx length was
RsaI 810 710, 165 810 effective in separating one of the three populations
ScrFI 755, 70 755, 70 755, 25, 20 using multivariate analysis. Because data were
XbaI 800 695, 165 800 gathered using consistent methodology, our results re-
inforce the notion that morphologically similar taxa
may inhabit distant geographic areas. Metric and mer-
istic traits of our specimens fall well within the range of
populations each had a unique mtDNA haplotype measurement reported in the literature for X. inter-
(designated I and F, respectively), whereas two different media (Levi 1950; Gerlach 1953; Rao and Ganapati
mtDNA haplotypes (Va and Vb, with respective fre- 1968; Luporini et al. 1973). However, when we com-
quencies of 80 and 20%) were identified within the pared our specimens with X. intermedia from Ar-
Virginia population. Overall, the seven informative re- cachon, France (Ruppert 1979), we found statistically
striction endonucleases produced 13, 14, and 11 restric- significant differences in several major traits. Specifi-
tion fragments in Haplotypes I, Va, and Vb, respective- cally, the means of body length, pharynx length, furca
ly (see Table 2 and Figs. 3 and 4 for details of fragment length and adhesive tube length of the French speci-
profiles). The number of shared fragments between the mens were consistently smaller than the means of the
haplotypes ranged from 3 to 6 (Table 3), corresponding specimens from each of the other populations. This last
to estimated sequence divergences (d) of 5.3 to 11.5% finding may be a morphological indication that cryptic
(Table 4). Due to the small number of successful PCR species exist within the X. intermedia taxon (see later in
products for Florida specimens (see second paragraph "Discussion").
of "Results") and the fact that they were cut with The second goal of our investigation was to deter-
a reduced set of restriction endonucleases compared to mine the level of genetic variability in allopatric popu-
the other locations, the Florida specimens were ex- lations of cosmopolitan species. Our results show
cluded from the analyses shown in Tables 2 to 4. a very high degree of restriction-fragment length poly-
However, the estimated sequence divergences between morphism among populations of Xenotrichula inter-
the F haplotype and the other three (I, Va and Vb) were media. Four mtDNA haplotypes were represented from
comparable in magnitude (5.7 to 8.3%) to those shown three geographically distant areas. The estimated nu-
in Table 4. cleotide sequence divergence for the COI gene
740

(d = 5.3% to 11.5%), calculated for each pairwise com- results are in accordance with the prediction by Avise
bination of haplotypes, is quite high compared with et al. (1987) that gross mtDNA genetic discontinuities
previous studies of intraspecific mtDNA divergence, in the absence of spatial separation are to be expected
but consistent with estimates of mtDNA divergence where reproductively isolated sibling species are inad-
from potential or confirmed interspecific comparisons. vertently assayed as if belonging to a single species.
For the 16S rRNA gene, Bucklin et al. (1995) found that Several mechanisms of long-range meiofaunal dis-
the level of mtDNA sequence variation within species persal have been suggested (e.g. meiofauna rafting on
of Calanus ranged from 0.3 to 2.6%, while the inter- drifting materials or trapped in the ballast of sailing
specific variation ranged from 7.3 to 23%. The most vessels; see also Gerlach 1977), and even plate tectonics
comparable study of COI divergence in terms of has been suggested as a mechanism explaining amphi-
mtDNA region analyzed is that of Burton and Lee Atlantic distribution (Sterrer 1973). Short-range disper-
(1994; see also Burton 1994), who sequenced a 500- sal via the water column associated with benthic storms
base-pair PCR product in the copepod Tigriopus caIi- could, over time, also lead to a long-range dispersal
fornicus corresponding to mtDNA Positions 1756 to (Palmer 1988; Giere 1993). However, our results sug-
2255 in Drosophila yakuba (Clary and Wolstenholme gest that rates of dispersal among disjunct locations are
1985). They reported sequence differences of 1 to 15.4% very low and that extant regional populations exist as
within single samples of T. californicus, and differences discrete units among which gene-flow appears to be
of up to 18% in allopatric (but nominally conspecific) negligible.
samples. Other studies of COI sequence variation in The marked phenotypic resemblance but genetic dis-
marine invertebrates have used the COI-a and COI-f similarity indicate that levels of morphological and
primers of Palumbi and Benzie (1991), which amplify molecular divergence are remarkably decoupled in the
Positions 2131 to 2811 of the D. yakuba sequence. For Xenotrichula intermedia species complex. Similar find-
example, Knowlton et al. (1993) reported interspecific ings have been emphasized by evolutionary biologists
sequence-distance values of 6.4 to 20.4% in pairwise working with other invertebrates and vertebrates (e.g.
comparisons of snapping shrimp (Alpheus spp.) across Cherry et al. 1978; Cunningham et al. 1992). Recently,
the Isthmus of Panam/t, and comparable interspecific large differences in mitochondrial DNA among mor-
COI-based distances were reported in horseshoe crabs phologically similar species have been reported in
by Avise et al. (1994). penaeid shrimps (Palumbi and Benzie 1991) and in the
The high nucleotide-sequence divergence found "living fossil" horseshoe crabs (Arise et al. 1994). Two
among Xenotrichula intermedia haplotypes raises ques- hypotheses could explain large genetic differences in
tions regarding their affiliation with the same taxon, phenotypically similar species: (1) the rate of molecular
and suggests that the specimens with different haplo- evolution might be rapid, or (2) the rate of morphologi-
types may in fact be distinct species. Because our cal divergence might be slow. Because there are no data
specimens from different populations are generally to substantiate the first hypothesis, Palumbi and
morphologically indistinguishable, they may be con- Benzie (1991) contend that the phenomenon may be
sidered sibling species (Mayr 1948). Recently, Knowl- best explained by stabilizing selection acting on mor-
ton (1993) reviewed the occurrence of marine sibling phological or ecological characters while molecular
species and noted that, in comprehensive studies of differences accumulate at "typical" rates. Previous
single geographic regions, discovery of sibling species comparisons of morphological and molecular evolu-
often results in four-fold or greater increase in diversity. tion in amphibians have in fact revealed similar pat-
Overall, one can expect the number of all marine spe- terns (Wallace et al. 1971). Such a scenario is probably
cies to increase by an order of magnitude if sibling also true among meiofauna, in which slow morphologi-
species are considered. Thus, the discovery that popula- cal divergence may be related to the relatively stable
tions of X. intermedia are actually sibling species falls and uniform physiographic nature of their habitat. In-
well within Knowlton's prediction. Further study of terstitial animals, including gastrotrichs, are probably
specimens from other areas within the nominal range of subject to selective pressure and stabilizing selection
X. intermedia may well reveal additional species within that favor distinctive phenotypic and perhaps physiolog-
the complex, and more detailed examination of mater- ical characters (as is shown by the convergence of
ial from the Gulf of Mexico would also be worthwhile. morphological and life-history traits among unrelated
The occurrence of sibling species at the same site (in our meiobenthic groups) but may constrain morphological
case at Norfolk, Virginia) should not come as a sur- divergence when an adaptive peak is reached.
prise, since the phenomenon has previously been re- To our knowledge, this is the first study that has
ported in other taxa including meiobenthic forms (i.e., investigated the paradox of meiofauna cosmopolitan-
in the harpacticoid copepod genus Tisbe: Volkmann- ism using both morphological and molecular ap-
Rocco 1972; Fava and Volkmann 1975), and plank- proaches. While morphology has been examined by
tonic forms (i.e., several species of Calanus with previous investigations (see earlier in "Discussion"),
extensive overlapping geographical range in the north molecular genetics is a relatively new field in meioben-
Atlantic Ocean: see Bucklin et al. 1995). Moreover, our thic studies. We are aware of only one published paper
 741

(Litvaitis et al. 1994) dealing with this subject. Molecu- dasyida: Macrodasyidae) with description of seven new species.
lar genetic studies have never been carried out in the Proc biol Soc Wash 107:239255
Gastrotricha, and evolutionary aspects of the COI Fava G, Volkmann B (1975) Tisbe (Copepoda: Harpacticoida) spe-
cies from the Lagoon of Venice. I. Seasonal fluctuations and
mitochondrial gene (particularly within meiofaunal ecology. Mar Biol 30:151-165
groups) are poorly known (but see Palumbi and Benzie Eolmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA
1991). Although there is evidence that this is one of the primers for amplification of mitochondrial cytochrome
most conservative protein-coding genes in the c oxidase subunit I from diverse metazoan invertebrates. Molec
mar Biol Biotechnol 3:294--299
mitochondrial genome (Brown 1985), we acknowledge Ganapati PN, Rao GC (1967) On some marine interstitial Gastro-
that further studies are needed before making any con- trichs from the beach sands of Waltair coast. Proc Indian Acad
clusive statement as to what extent the COI gene can be Sci (Sect B) 66:214 225
used to infer phylogeny within "lower" metazoans such G~rate T, Robinson MP, Chac6n MR, Parkhouse RME (1991)
as gastrotrichs. Characterization of species and races of the genus Meloidogyne
by DNA restriction enzyme analysis. J Nematol 23:414-420
Gellcr JB, Carlton JT, Powers DA (1994) PCR-based detection of
Aeknowledgements We thank F. Dobbs and L. Drake for providing mtDNA haplotypes of native and invading mussels on the
us with sand from Norfolk, M. Hollay and R. Macchiavelli for northeastern Pacific coast: latitudinal pattern of invasion. Mar
statistical advice, R. Bouchard of the LSU Light Microscopy Facil- Biol 119:243-249
ity for his invaluable assistance, S. Herke for comments on the Gerlach SA (1953) Gastrotrichen aus dem Kuestengrundwasser des
manuscript, and S. Nunez for technical advice. Mittelmeeres. Zool Anz 150:203 211
Gerlach SA (1962) Freilebende Meeresnematoden von den Mal-
ediven. Kieler Meeresforsch 18:81-108
Gerlach S (1977) Means of meiofauna dispersal. Mikrofauna
Meeresbod 61:89-103
References Giere O (1993) Meiobenthology. The microscopic fauna in aquatic
sediments. Springer-Verlag, Berlin
Avise JC, Arnold J, Ball RM, Bermingham E, Lamb T, Neigel JE, Harris TS, Sandall LJ, Powers TO (1990) Identification of single
Reeb CA, Saunders NC (1987) Intraspecific phylogeography: the Meloidogyne juveniles by polymerase chain reaction amplifica-
mitochondrial DNA bridge between population genetics and tion of mitochondrial DNA. J Nematol 22:518-524
systematics. A Rev Ecol Syst 18:489-522 Hulings NC (1971) Summary and current status of the taxonomy
Arise JC, Nelson WS, Sigita H (1994) A speciational history of and ecology of benthic Ostracoda including interstitial forms.
"living fossils": molecular evolutionary patterns in horseshoe Smithson Contr Zool 76:91-96
crabs. Evolution 48:1986-2001 Hummon WD (1994) Trans- and cis-Atlantic distributions of
Balsamo M, Manicardi GC (1995) Nuclear DNA content in Gas- three marine heterotardigrades. Trans Am microsc Soc 113:
trotricha. Experientia 51:356-359 333 342
Balsamo M, Todaro MA, Tongiorgi P (1992) Marine gastrotrichs Hummon WD, Hummon MR, Mostafa HM (1994) Marine Gas-
from the Tuscan Archipelago (Tyrrhenian Sea). II. Chaetono- trotricha of Mediterranean Egypt. Am Zool 34: p. 10A (Ab-
tida, with description of three new species. Boll Zool 59: stract)
487-498 Jouk PEH, Hummon WD, Hummon MR, Roidou E (1992) Marine
Brown WM (1985) The mitochondrial genome of animals. In: McIn- Gastrotricha from the Belgian coast: species list and distribu-
tyre RJ (ed) Molecular evolutionary genetics. Plenum Press, tion. Bull Inst Sci nat Belg (Biol) 62:87-90
New York, pp 95-130 Karl SA, Avise JC (1993) PCR-based assay of Mendelian polymor-
Bucklin A, Frost BW, Kocher TD (1995) Molecular systematics of phism fi'om anonymous single-copy nuclear DNA: techniques
six Calanus and three Metridia species (Calanoida: Copepoda). and applications for population genetics. Molec Biol Evolut 10:
Mar Biol 121:655-664 342-361
Burton RS (1994) Inferring the genetic structure of marine popula- Knowlton N (1993) Sibling species in the sea. A Rev Ecol Syst 24:
tions: a case study comparing allozyme and DNA sequence 189 216
data. Rep Calif coop ocean Fish Invest (CalCOFI) 35:52-60 Knowlton N, Weigt LA, Sol6rzano LA, Mills DK, Bermingham
Burton RS, Lee BN (1994) Nuclear and mitochondrial gene genealo- E (1993) Divergence in proteins, mitochondrial DNA, and re-
gies and allozyme polymorphism across a major phylogeo- productive compatibility across the Isthmus of Panama.
graphic break in the copepod Tigriopus caIifornicus. Proc natn Science, NY 260:1629-1632
Acad Sci USA 91:5197 5201 Levi C (1950) Contribution a l'6tude des gastrotriches de la region de
Castagnone-Sereno P, Vanlerberghe-Masutti F, Leroy F (1994) Roscoff. Archs Zool exp g6n 87:31-42
Genetic polymorphism between and within Meloidogyne species Litvaitis MK, Nunn G, Thomas WK, Kocher TD (1994) A molecu-
detected with RAPD markers. Genome 37:904-909 lar approach for the identification of meiofaunal turbellarians
Cherry LM, Case SM, Wilson AC (1978) Frog perspective on the (Platyhelminthes, Turbellaria). Mar Biol 120:437-442
morphological difference between humans and chimpanzees. Luporini P, Magagnini G, Tongiorgi P (1973) Chaetonotoid gas-
Science, NY 200:209-211 trotrichs of the Tuscan Coast. Boll Zool 40:31 40
Clary DO, Wolstenholme DR (1985) The mitochondrial DNA mol- Mayr E (1948) The bearing of the new systematics on genetical
ecule of Drosophila yakuba: nucleotide sequence, gene organiza- problems. The nature of species. Adv Genet 2:205-237
tion, and genetic code. J molec Evolut 22:252 271 Nei M (1987) Molecular evolutionary genetics. Columbia University
CoulI BC (1988) Ecology of the marine meiofauna. In: Higgins RP, Press, New York
Thiel H (eds) Introduction to the study of meiofauna. Smith- Nei M, Li WH (1979) Mathematical model for studying genetic
sonian Institution Press, Washington, DC, pp 18-38 variation in terms of restriction endonucleases. Proc natn Acad
Cunningham CW, Blackstone NW, Buss LW (1992) Evolution of Sci USA 7',6:5269 5273 ,
king crabs from hermit crab ancestors. Nature, Lond 355: Palmer MA (1988) Dispersal of marine meiofauna: a review and
539-542 conceptual model explaining passive transport and active
Evans WA (1994) Morphological variability in warm-temperate and emergence with implication for recruitment. Mar Ecol Prog Ser
subtropical populations of Macrodasys (Gastrotricha: Macro- 48:81-91
742

Palumbi SR, Benzie J (1991) Large mitochondrial DNA differences SAS Institute Inc. (1990) SAS user's guide: statistics. Version 6 edn.
between morphologically similar penaeid shrimp. Molec mar SAS Institute Inc. Cary, North Carolina
Biol Biotechnol 1:27-34 Sterrer W (1973) Plate tectonics as a mechanism for dispersal and
Pfannkuche O, Thiel H (1988) Sample processing. In: Higgins RP, speciation in interstitial sand fauna. Neth J Sea Res 7:200-222
Thiel H (eds) Introduction to the study of meiofauna. Smith- Todaro MA, Fleeger JW, Hummon WD (1995) Marine gastrotrichs
sonian Institution Press, Washington, pp 134-145 from the sand beaches of the northern Gulf of Mexico: species
Rao GC, Ganapati PN (1968) The interstitial fauna inhabiting the list and distribution. Hydrobiologia 310:107-117
beach sands of Waltair coast. Proc natn Inst Sci India (Ser B) 34: Volkmann-Rocco B (1972) Species of Tisbe (Copepoda: Harpac-
82-125 ticoida) from Beaufort, North Carolina. Archo Oceanogr
Renaud-Mornant J, Pollock LW (1971) A review of the systematics Limnol 17:223-258
and ecology of marine Tardigrada. Smithson Contr Zool 76: Wallace DG, Maxson LR, Wilson AC (1971) Albumin evolution in
109-117 frogs: a test of the evolutionary clock hypothesis. Proc natn
Ruppert EE (1977) Zoogeography and speciation in marine Gastro- Acad Sci USA 68:3127-3129
tricha. Mikrofauna Meeresbod 61:231-251 Wells JBJ (1986) Biogeography of benthic harpacticoid copepods of
Ruppert EE (1979) Morphology and systematics of the Xeno- the marine littoral and continental shelf. Syllogeus (Nat Mus
trichulidae (Gastrotricha, Chaetonotida). Mikrofauna Meeres- Can) 58:126-135 [Proc 2nd int Conf Copepoda (Schriever G,
bod 76:1 56 Schminke HK, Shih C-t, eds). Ottawa, Can]
Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R, Horn GT, Westheide W (1971) Interstitial Polychaeta (excluding Archianel-
Mullis KB, Erlich HE (1988) Primer-directed enzymatic amplifi- lida). Smithson Contr Zool 76:57-70
cation of DNA with a thermostable DNA polymerase. Science,
NY 239:487-491