First Application of Morphometrics in a Study of Variations in Uncinial
Shape Present within the Terebellidae (Polychaeta)
André Rinaldo
Senna Garraffoni*1,2
and Maurício de Garcia Camargo1
1Centro de Estudos do Mar, Universidade Federal do Paraná, Av. Beira
mar s/n, CP 50002, Pontal do Paraná, PR 83255-000, Brazil
2Curso
de Pós-Graduação
(Accepted June 30, 2005)
*To whom
correspondence and reprint requests should be addressed.
E-mail: garraffoni@ufpr.br
Abstract André Rinaldo Senna Garraffoni and
Maurício de Garcia Camargo (2005) First application of
morphometrics in a study of variations in uncinial shape present within the
Terebellidae (Polychaeta). Zoological Studies xx(x):
xx-xx. In this study,
morphometric analyses of the uncinial shape were used to differentiate
morphological groups within 4 Terebellidae subfamilies. Thirty species were examined, and 17
distances were measured and analyzed using non-metric multidimensional scaling
(n-MDS) and cluster analysis. The results derived from both
techniques were quite similar and clearly demonstrated that there are 2
different character states for the shape of the manubrium (one for the
Trichobranchinae and another for the other subfamilies), and that 3 different
character states define the overall uncini shape (one for Terebellinae and
Thelepodinae, another for Polycirrinae, and yet another for Trichobranchinae).
Key words: Coding, Terebelliformia, Terebellida,
Neurochaetae, Cladistics.
----------------------------------------------------------------------------------------------------------------
INTRODUCTION
The Terebelliformia (sensu Rouse and
Pleijel 2001) is a group of sedentary polychaetes comprising the families
Terebellidae, Ampharetidae, Alvinellidae, and Pectiinaridae. Within the taxon Terebelliformia, the family
Terebellidae contains more than 400 species grouped in approximately 60 genera
within 4 subfamilies: Polycirrinae, Thelepodinae,
Trichobranchinae (Rouse and Pleijel 2001, Garraffoni and Lana 2004), and
Terebellinae (Holthe 1986, Hutchings 2000, Rouse and Pleijel 2001).
The Terebellidae is characterized by
anterior grooved buccal tentacles, which are used to transport fine surface
particles to the mouth (Fauchald and Jumars 1979, Hutchings 2000, Rouse and
Pleijel 2001), and neurochaetae modified as uncini which have an anchor-like
function (Hutchings 2000, Rouse and Pleijel 2001), allowing them to anchor
their soft body within tubes in order to avoid the risk of predation (Holthe
1986, Woodin and Merz 1987).
Exceptions are found in the genera Hauchiella Hartman, 1959; Enoplobranchus
Webster, 1879; and Lysilla Malmgren, 1866, which lack neurochaetae, and
the genus Amaeana Levinsen, 1893, which has neurochaetal spines rather
than uncini. These 4 genera are included within the
terebellid subfamily Polycirrinae.
Analyses of neuropodial uncini in the families
Terebellidae, Ampharetidae, Alvinellidae, and Pectiinaridae have shown that
most morphological variation occurs in suprageneric taxa (Holthe 1986). In spite of these differences observed
in each family, however, it is possible to find some common patterns. All uncini of the Terebelliformia have a
main fang, an upper part or capitium (with numerous secondary teeth arranged in
different positions), and a lower part, where a subrostral process can be
present or absent (and a base called a manubrium) (Holthe 1986, Bartolomaeus
1995, Garraffoni and Lana 2004).
Morphometrics allow rigorous
quantitative analysis of variations in organism size from shape effects using
multivariate statistical methods (Kligenberg 2002). This method is able to distinguish the
effect on shape variations of different sizes of the organism or structure of
an organism being studied. The
most-common approach of morphometrics is to consider the landmark configuration
of morphological features (Kligenberg 2002). The landmarks can be interpreted as a
way to reduce the shape of an organism to a set of points that characterize
their general traits, which are assumed to be homologous across individuals and
populations (Rolf and Marcus 1993, David and Laurin 1998, Costa et al. 2004).
Morphometric studies using morphological
landmarks to define species or populations have not been widely used in
polychaetes. Some studies
(Ben-Eliahu 1987, Fauchald 1991) used morphometric studies in the broad sense
of the term, but in recent years, new studies on polychaetes using
morphometrics which consider landmark configurations have been published
(Maltagliati et al. 2001, Costa and Paiva submitted).
The main goal of this paper was to
use a morphometric approach to assess the different morphological classes of
uncini observed within the Terebellidae. This morphometric study represents
the 1st comprehensive study of the topic for Terebelliformia. As pointed out by Guerrero et al. (2003), the results of
statistical tests of character variations allow a consistent judgment of the
similarity among variants in order to establish character state identity.
----------------------------------------------------------------------------------------------------------------
MATERIALS AND METHODS
Operational
taxonomic units (OTUs)
Thirty species were selected within
Terebellidae: 3 species of Trichobranchinae, 4 species of Polycirrinae, 5
species of Thelepodinae, and 18 species of Terebellinae (Table 1).
Table 1. List of the Terebellidae species used in the present study
Species
|
Subfamily
|
Abbreviation
|
|
Polycirrus boholensis Grube, 1878 |
Polycirrinae |
Po_bo |
|
Polycirrus disjunctus Hutchings and Glasby, 1986 |
Polycirrinae |
Po_di |
|
Polycirrus medius Hessle, 1917 |
Polycirrinae |
Po_me |
|
Polycirrus
abrolhenis Garraffoni and Costa, 2003 |
Polycirrinae |
Po_abro |
|
Amphitrite pachyderma Hutchings and Glasby, 1986 |
Terebellinae |
Am_pa |
|
Arranooba booromia Hutchings and Glasby, 1986 |
Terebellinae |
Arr_bo |
Baffinia
biseriata Hutchings and Glasby, 1988
|
Terebellinae |
Ba_bi |
|
Lanassa ocellata Hutchings and Glasby, 1988 |
Terebellinae |
Lan_oc |
|
Lanice sinata Hutchings and Glasby, 1990 |
Terebellinae |
Lani_si |
|
Lanicides fascia Hutchings and Glasby, 1988 |
Terebellinae |
La_fa |
|
Lanicides tribranchiata Hutchings and Glasby, 1988 |
Terebellinae |
La_tr |
|
Loimia triloba Hutchings and Glasby, 1988 |
Terebellinae |
Lo_tr |
Longicarpus nodus
Hutchings, 1990
|
Terebellinae |
Lon_no |
|
Neolepra macrocercus Hutchings and
Glasby, 1986 |
Terebellinae |
Ne_ma |
|
Paraeupolymnia
uspiana Nogueira, 2003 |
Terebellinae |
Ni_sp1 |
|
Phisidia echuca Hutchings and
Glasby, 1986 |
Terebellinae |
Ph_ec |
Pista
sinusa Hutchings
and Glasby, 1986
|
Terebellinae |
Pi_si |
|
Pseudoproclea australis Hutchings and Glasby, 1990 |
Terebellinae |
Ps_au |
|
Reteterebella aloba Hutchings and
Glasby, 1986 |
Terebellinae |
Re_al |
|
Hutchingsiella
cowarrie (Hutchings, 1997) |
Terebellinae |
Sp_co |
|
Terebella muliarrus Hutchings, 1993 |
Terebellinae |
Te_mu |
|
Tyra owensi Hutchings, 1997 |
Terebellinae |
Ty_ow |
Euthelepus
setubalensis McIntosh, 1885
|
Thelepodinae
|
Eu_st |
|
Streblosoma comatus (Grube, 1859) |
Thelepodinae |
St_co |
|
Thelepus ambitus Glasby and Hutchings, 1987 |
Thelepodinae |
Th_am |
|
Pseudothelepus binara Hutchings, 1997 |
Thelepodinae |
Pst_bi |
|
Pseudostreblosoma serratum Hutchings and Murray, 1984 |
Thelepodinae |
Pss_se |
Octobranchus
antarcticus Monro, 1936
|
Trichobranchinae |
Oc_an |
|
Terebellides sepultura Garraffoni and Lana, 2003 |
Trichobranchinae |
Tere_sep |
|
Terebellides totae Elias and Bremec, 1999 |
Trichobranchinae |
Tere_to |
Measurements
In the present analysis, 17
distances were measured using 9 landmarks distributed in order to show
differences observed in the morphology of the thoracic uncini in the
Terebellidae (Fig. 1). We also used
4 points (1, 5, 7, and 13) that are not real landmarks because they do not fall
on the structure being analyzed.
However, these points are directly related to uncinial parts as they are
taken from projections of real landmarks.
These additional points therefore contributed to a better representation
of the morphological variability of the uncini, and no correlations were found
between these projections and the real landmarks they were based upon. Bookstein (1991) suggested that 3 main
kinds of landmarks may be used: juxtaposition, maximal curvature, and
externally constructed points, and also intermediary cases between types
2 and 3. We used an intermediary case between types 2
and 3, because of the nature of our data.
Fig. 1.
Morphometric measurements recorded for
Terebellidae uncini (modified from Hutchings and Glasby, 1988).
The distances were as follows:
distance 1 was the distance between points 2 and 6; distance 2, points 3 and 9;
distance 3, points 12 and 11; distance 4, points 11 and 10; distance 5, points
12 and 2; distance 6, points 12 and 3; distance 7, points 12 and 6; distance 8,
points 12 and 8; distance 9, points 2 and 1; distance 10, points 6 and 5;
distance 11, points 8 and 4; distance 12, points 8 and 7; distance 13, points 8
and 13; distance 14, points 2 and 3; distance 15, points 3 and 6; distance 16,
points 6 and 9; and distance 17, points 9 and 2. The uncini distances were measured
directly from the original illustrations of each species and were corrected for
the original scale. Only Polycirrus
abrolhensis, Nicolea sp. 1, Terebellides sepultura, and T.
totae were measured directly from drawings done by the authors. We also choose to include figures of
illustrated anterior thoracic uncini.
However, species for which the original description did not mention from
which segment the uncinus was dissected were not included in the analysis. The data matrix is available at the web
site (http://www.cem.ufpr.br/garraffoni/planilhauncini.xls).
These distances were chosen to avoid some
of the biases and weakness pointed out by Strauss and Bookstein (1982) in
traditional character sets, such as: 1) most characters tend to be aligned with
one of very few axes; 2) coverage of the form is highly uneven by region as
well as by orientation, being dense in some areas of the body and sparse in
other, and 3) many measurements extend over much of the body, so short
distances contain more-localized information than long ones.
In order to describe the different regions of the uncini, we used the
terminology proposed by Holthe (1986: 31, fig. 6).
Analysis
Cluster analysis and non-metric multidimensional scaling ordination
(n-MDS) were performed from a log-transformed matrix (OTUs x measurements)
using the software, Primer (Primer-E, Plymouth Marine Laboratory
The Primer framework (n-MDS + cluster
analysis + ANOSIM) is commonly used in ecological studies, but less often used
in morphometric studies. Actually,
n-MDS and cluster analysis are not new techniques, and they have been applied
to morphometric analyses elsewhere (Maltagliati et al. 2001, Chui et al. 2002,
Noireau et al. 2002, Costa et al. 2004). ANOSIM, on the other hand, was
introduced by the Primer framework (Clarke and Warwick 1994) and should be
interpreted in conjunction with the n-MDS.
ANOSIM simply calculates the probability of the random occurrence of the
observed groups.
----------------------------------------------------------------------------------------------------------------
RESULTS
The result of the n-MDS of the 30 Terebellidae
species (Fig. 2) shows a very low stress value ( minimum
stress, 0.04) proving high resolution (Clarke and Warwick 1994, Chui et al.
2002). The n-MDS analysis using 17
characters clearly separated the species into 3 clusters according to different
quadrants (Fig. 2). The
Trichobranchinae was clustered in the 1st quadrant, Polycirrinae in the 2nd
quadrant, and Thelepodinae and Terebellinae in the 3rd quadrant. The ANOSIM permutation test (Table 2)
confirmed this tendency, as every pair of groups was significantly separated at
the level of 5%, except for Terebellidae against Thelepodinae. Very similar classes as defined in the
n-MDS were also obtained by cutting the dendrogram at specific height points,
but some species from the same subfamily were not necessarily grouped together
(Fig. 3).
Fig. 2. Two-dimensional plot of uncini
character sets analyzed by non-metric multidimensional scaling. Dark gray circles represent Trichobranchinae species, light gray squares
represent Polycirrinae species, light gray triangles represent Terebellinae
species, and dark gray rhombuses represent Thelepodinae species.
Table
2. Results
of the ANOSIM (global R = 0.51, global significance = 0.2) using the
Terebellidae subfamilies. Abbreviations: Tere, Terebellinae; Tricho,
Trichobranchinae; The, Thelepodinae; Poly, Polycirrinae
|
Groups |
R statistic |
Significance level (p) |
Possible permutations |
|
Tere, Trico |
0.9 |
0.2 |
1540 |
|
Tere, Poly |
0.57 |
0.5 |
8855 |
|
Tere, The |
-0.038 |
54.5 |
8855 |
|
Trico, Poly |
1.0 |
2.9 |
35 |
|
Trico, The |
1.0 |
2.9 |
35 |
|
Poly, The |
1.0 |
2.9 |
35 |
Fig. 3. Dendrogram of the 30 Terebellidae
species obtained using Euclidean distances. For species abbreviations see table 1. Ter, Terebellinae; The, Thelepodinae; Poly, Polycirrinar; Tri,
Trichobranchinae.
The n-MDS and cluster analysis
revealed 2 distinct morphotypes within the 4 subfamilies. These 2 distinct morphotypes (or 2
distinct clusters in the dendrogram) were obtained by the morphological
differences in the size of the manubrium. Trichobranchinae has uncini with a
long-shafted manubrium, while Polycirrinae, Thelepodinae, and Terebellinae have
uncini with a short manubrium.
The dendrogram (a diagram in which
the similarities of 2 samples or groups are considered to have fused), in fig.
3, also shows that some species of the Terebellinae have an uncinial shape more
similar to species of Thelepodinae than to other Terebellinae species (Lanicides
fascia, Pista sinusa, and L. tribranchiata). This is because L. fascia, P.
sinusa, and L. tribranchiata have a developed chitinized shaft in the posterior part of the uncini, called
the posterior process. The other
species of the Terebellinae and Thelepodinae do not have this posterior
process.
----------------------------------------------------------------------------------------------------------------
DISCUSSION
Shape is a multidimensional component of
variation in the morphological form which is expected to have a high
information content regarding the evolutionary process responsible for the
observed diversity (Atchley and Hall 1991, Costa et al. 2004). One of
the main purposes of morphometric analysis is to establish and delimit the
nature of morphological variations as a 1st step to providing an assessment of
homology among different features, which can be used in subsequent studies to
improve one's understanding of the phylogenetic relationships among the taxa
(Reis 1988). Guerrero et al. (2003) also pointed out that
the similarity test of character states identified by morphometric analyses can
help the systematist judge whether some character variants are sufficiently
different to recognize them as a different character or character state.
Observing the morphological groups established
by n-MDS and the dendrogram, we can clearly evaluate the differences between
the posterior process and the long-shafted manubrium. Unfortunately, there is no information
on the ontogeny of the development of this posterior process and the
long-shafted manubrium in the literature to help us evaluate our results.
The 2 different morphotypes identified (in
n-MDS quadrant 1 vs. quadrants 2 and 3; and in the dendrogram, the 1st
dichotomy close to the base), separating the Trichobranchinae from the
Polycirrinae, Terebellinae, and Thelepodinae can be defined by the presence of
a long-shafted manubrium in the Trichobranchinae. Thus, these 2 different morphotypes,
short- and long-shafted manubrium, can be assumed to represent 2 character
states linked to the size of the manubrium.
The presence of a posterior process
in some Terebellinae species, namely Lanicides fascia, Pista sinusa,
and L. tribranchiata, is
also considered interesting as it potentially indicates homology with the
long-shafted manubrium. However,
our results showed that these structures are not homologous, and they are thus
treated as independent characters.
We suggest that the posterior process underwent different ontogenetic
development than that of the manubrium, and the former is only an extensive
development of the posterior part while the latter is a development of the
entire uncinial base.
Another important result from both
analyses is the identification of 3 different shape components showing
significant differences among the subfamilies. These 3 different shapes can also be
assumed to be 3 character states for the character of the overall uncinial
shape. In spite of each group having its own overall shape (Figs. 2, 3),
morphotypes of the Terebellinae and Thelepodinae are very similar, and it is
difficult to divide them into different states.
In summary, morphometry is a powerful tool
which allows an objective coding of morphologically quantitative traits into
character states, or different morphological patterns into specific levels to
understand microevolution and identify morphotypes. Our research provides a robust basis to
examine the overall performance of morphological variation in the Terebellidae
uncini and can help elucidate the major contribution of these characters to
delineating this family.
Terebellidae uncini is a rich source of taxonomic information that can
be used to help assess evolutionary relationships within the family.
Note added during proofing: After the time that this paper was
submitted, we began a new study using morphometrics on Terebellidae uncini in
order to complement this 1st approach.
In this new study, we dissected segments 7 and 16 of different
Terebellidae species, and took pictures of 3 different uncini on each
segment. Using measurement
software, we obtained all distances previously used in this paper and
additional measurements such as area and perimeter. Currently, we are still analyzing the
information using the same procedures that were applied here. In addition, an attempt is being made to
establish a mathematical model to predict different taxa (output) from uncini
measurements (input) using an artificial neural network (ANN) based on
multilayer perceptions.
Acknowledgments: We are much indebted to Elisa Costa, Cinthya
Santos, Silvio Nihei, Kai George, Pat Hutchings, and 2 anonymous reviewers for
their advice and comments.
----------------------------------------------------------------------------------------------------------------
REFERENCES
Atchley WR, Hall
BK. 1991. A model for development and evolution of
the complex morphological structures.
Biol. Rev. 66:
101-157.
Bartolomaeus T. 1995. Structure and formation of the uncini in
Pectinaria koreni, Pectinaria auricoma (Terebellida) and Spirorbis
spirobis (Sabellida): implications for annelid phylogeny and the position
of the Pogonophora. Zoomorphology 115:
161-177.
Chui YW, Chen HC, Lee SC,
Chen CA. 2002. Morphometric analysis of shell and
operculum variations in the viviparid snail Chipangopaludina chinensis
(Mollusca: Gastropoda) in
Clarke KD, RM
Costa
EM, Paiva PC. (submitted). A
morphometric analysis of six populations of
different species of Eunice (Annelida: Polychaeta). Zool. J. Linn. Soc.
Costa, LF, SF Reis, RAT Arantes, ACR
Alves, G Mutinari. 2004. Biological shape analysis by digital
curvature. Patterns
Recogn.] 37: 515-524.
David B, Laurin B. 1998. Geometric morphometrics approaches to
biodiversity. Océanis 24: 287-303.
Fauchald K, PA
Jumars. 1979. The diet of worms: a study of polychaete
feeding guilds. Oceanogr. Mar. Biol. Ann. Rev. 17: 193-284.
Fauchald K. 1991. A morphometric study of eunicid
polychaetes from
Garraffoni ARS, PC Lana. 2004. Cladistics analysis of subfamily
Trichobranchinae (Polychaeta: Terebellidae). J. Mar. Biol. Assoc. UK 87:
973-983.
Guerrero JA, E Luna, C Sánchez-Hernández. 2003. Morphometrics in the quantification of
character state identity for the assessment of primary homology: an analysis of
character variation of the genus Artibeus (Chiroptera:
Phyllostomidae). Biol. J. Linn. Soc. 80:
45-55.
Holthe T. 1986. Evolution, systematics
and distribution of the Polychaeta Terebellomorpha, with a catalogue of the
taxa and a bibliography. Gunneria 55: 1-236.
Hutchings P. 2000. Family Trichobranchidae. In PL Beesley, GLB Ross, CJ Glasby, eds. Polychaeta
and allies: the southern synthesis. Fauna of
Hutchings PA, CJ
Glasby. 1988. The Amphitritinae (Polychaeta:
terebellidae) from
Kligenberg CP. 2002. Morphometrics and the role of the
phenotype in studies of the evolution of developmental mechanisms. Gene 287: 3-10.
Maltagliati F, Camilli L, Lardicci C, Castelli
A. 2001. Evidence for morphological and genetic
divergence in Perinereis cultrifera (Polychaeta: Nereididae) from two
habitat types at
Noireau F, SM Santos, M Gumiel, JP Dujardin, MS
Soares, RU Carcavall, C Galvão, J Jurberg.
2002. Phylogenetic
relationships with the oliveirai complex (Hemiptera: Reduviidae:
Tritatominae). Inf. Gen. Evol. 2: 11-17.
Reis SF. 1988. Morfometria e estatística multivariada
em biologia evolutiva. Rev. Bras. Zool. 5: 571-580.
Rolf FJ, LF Marcus. 1993. A revolution in morphometrics. Trend. Ecol.
Evol. 8: 129-132.
Rouse GW, P.
Pleijel. 2001. Polychaetes.
Strauss RE, FL Bookstein. 1982. The truss: body form reconstructions in
morphometrics. Syst. Zool. 31: 113-135.
Woodin SA, RA Merz. 1987. Holding on by their hooks: anchors for
worms. Evolution 41: 427-432.