Metacaligus latus n. sp. (Copepoda, Caligidae) is described from specimens found parasitic in the oral and gill cavities of the cutlassfish, Trichiurus lepturus Linnaeus, caught from the Strait of Taiwan and landed at Dong-Shih Fishing Port in Chiayi County, Taiwan. It is close to M. uruguayensis (Thomsen, 1949), but can be distinguished from the latter by the possession of shorter caudal rami in both sexes and wider cephalothorax and genital complex in the male. Cladistic (phylogenetic) analysis of the Family Caligidae, based on 23 selected morphological characters, revealed that Metacaligus Thomsen, 1949 is distantly related to Caligus Müller, 1758. It is a valid genus and occurs in sister-group relationship with Apogonia Cressey and Cressey, 1990.
Two species of Metacaligus Thomsen, 1949 are known parasitic on the marinefishes of East Asia: M. hilsae (Shen, 1957) from China and M. uruguayensis Thomsen, 1949 from Taiwan (Lin and Ho, 2000). While the former is confined to China (Shen, 1957) the latter has been reported from Uruguay (Thomsen, 1949), India (Pillai, 1963), Sri Lanka (Kirtisinghe, 1964), Venezuela (Ho and Bashirullah, 1977) and Brazil (Boxshall and Montú, 1997). In this paper we report a new species of this genus recovered from the cutlassfish, Trichiurus lepturus Linnaeus, caught in the Straight of Taiwan.
In his work on the copepod parasites of marine fishes of Uruguay, Thomsen (1949) proposed the subgenus Metacaligus to accommodate a new species of Caligus found parasitic on the cutlassfish from Uruguay. It was called “Caligus (Metacaligus) uruguayensis n. sp.” Furthermore, he included in this subgenus three other species of Caligus, namely, C. afurcatus Wilson, 1913; C. enormis Wilson, 1913; and C. rufus Wilson, 1908. Twenty-eight years later, on the occasion of the discovery of C. (Metacaligus) uruguayensis from the same species of host caught off Venezuela, Ho and Bashirullah (1977) reevaluated Thomsen's (1949) concept of Metacaligus and elevated it to the generic rank. Moreover, after examining the type-material of the three species of Caligus included in the Metacaligus by Thomsen (1949), Ho and Bashirullah (1977) excluded C. afurcatus and C. enormis, and included Caligus hilsae Shen, 1957, in the genus Metacaligus.
The rationale for Ho and Bashirullah (1977) to elevate Metacaligus to the generic rank was the possession in common of a combination of three character states by M. uruguayensis, M. rufus, and M. hilsae. Those three common morphological properties are (1) absence of sternal furca, (2) absence of accessory processes on two middle terminal spines of second exopodal segment of leg 1, and (3) possession of only one outer spine (instead of two) on the terminal exopodal segment of leg 2. Since none of the three states of character are unique to Metacaligus and each of them is respectively found in some species of Caligus, Dojiri (1983) was skeptical about the recognition of Metacaligus as a genus distinct from Caligus. Therefore, in this report, in addition to providing the description of a new species of Metacaligus, a cladistic analysis based on the morphological characters of the Family Caligidae is attempted. It is believed this analysis shall provide means to explore phylo-genetic relationships within the family and to examine if the establishment of Metacaligus as a genus separate from Caligus is justifiable.
MATERIALS AND METHODS
The cutlassfish (Trichiurus lepturus Linnaeus) caught in the Strait of Taiwan and landed at Dong-Shih Fishing Port in Chiayi County was purchased and transferred in an icebox to National Chiayi University where the laboratory examination of the copepod parasites was carried out. The copepod parasites removed from the fish were preserved in 70% ethanol. In studying the parasites, the specimens were cleared in 85% lactic acid for 1 to 2 hours before making dissection in a drop of lactic acid. The hanging-drop method devised by Humes and Gooding (1964) was employed to examine the dissected parts and appendages under the compound microscope. All drawings were made with the aid of a camera lucida. Measurements of the body and body parts were taken from 10 randomly selected specimens. The measurements given in the description are the average with its range shown in the parentheses that follow.
In this paper only qualitative characters are used in the cladistic analysis. Also, autapomorphic characters are excluded because they do not aid in a search for the wanted phylogenetic relationships, except for help in boosting consistency index of the generated cladogram(s). The type-species of each valid caligid genus is selected to extract the wanted homologous characters for the OTU. Based on Benz's (1993) cladogram for the siphonostomatoids parasitic on the vertebrates, Euryphorus Milne Edwards, 1840 and Trebius Krøyer, 1838 were utilized as the outgroup to polarize the transformation of the identified homologous character states of the Caligidae. Moreover, in coding multistate characters, when a transformation series containing a single basal bifurcation (dichotomous transformation) was encountered, we followed the method of “internal rooting” proposed by O'Grady and Deets (1987). In this case, the state of character that received a code of “0” is not plesiomorphic.
The computer program HENNIG86 Version 1.5 (Farris, 1988) was used to analyze the phylogenetic relationships between the genera of the Caligidae. The commands “mhennig*” and “bb*” were employed to produce multiple, shortest trees through performance of extended branch swapping.
261 ♀ ♀, 105 ♂ ♂, and 20 chalimus larvae attached to adults of both sexes were recovered from oral and gill cavities of cutlassfish, Trichiurus lepturus Linnaeus, caught in Strait of Taiwan and landed at Dong-Shih Fishing Port on following dates in 2001: 1 ♀and 1 ♂fromJ.-S. Ho and C.-L. Lin, 19 ♂ ♂and 3 larvae from 27 fishes on 27 April; and 201 ♀ ♀, 85 ♂ ♂, and 17 larvae from 62 fishes on 22 June. One holotype female, 1 allotype male, and 60 paratypes (30 ♀ ♀and 30 ♂ ♂) were selected from the last collection and deposited in the National Museum of Natural History, Smithsonian Institution in Washington, D.C.
Body (Fig. 1A) 3.29 (2.42–3.72) mm long, excluding setae on caudal rami. Cephalothoracic shield subcircular, 1.13 (0.82–1.30) mm long and 1.01 (0.78–1.32) mm wide, excluding marginal hyaline membranes. Fourth pediger 2.21 times wider than long. Genital complex globular, 1.13 (0.86–1.40) mm long and 1.11 (0.82–1.42) mm wide. Abdomen stocky and 1-segmented, 0.56 (0.40–0.82) mm long and 0.23 (0.18–0.38) mm wide. Caudal ramus (Fig. 3C) 1.63 (0.13×0.08 mm) times longer than wide; with 3 short, subterminal setae and 3 long, terminal setae in addition to a setule-bearing papilla on dorsal surface. Egg sac longer than body, 3.9 mm and containing 35 eggs.
Antennule (Fig. 1B) 2-segmented; proximal segment with 27 setae on anterodistal surface, distal segment with 1 subterminal seta on posterior margin and 11 setae plus 2 aesthetascs on distal margin. Antenna (Fig. 1C) 3-segmented; proximal segment smallest and unarmed; second segment rectangular and also unarmed; distal segment stocky claw bearing 1 seta in proximal region and another one close to middle region. Postantennal process (Fig. 1C) small and bluntly pointed, carrying 2 basal papillae with each bearing 4 setules; another similar papilla located nearby on sternum. Mandible (Fig. 1D) 4-segmented; with 12 teeth on medial margin of distal blade. Maxillule (Fig. 1C) comprising short, stout, dentiform process and prominent basal papilla with 3 naked setae. Maxilla (Fig. 1E) 2-segmented; proximal segment (lacertus) large and unarmed; slender, distal segment (brachium) carrying a large, subterminal, hyaline membrane on outer edge; both subterminal canna and terminal calamus fringed with fine spinules. Maxilliped (Fig. 2A) 3-segmented; proximal segment (corpus) largest but unarmed; middle (shaft) and distal (claw) segments fused to form subchelum, with small, blunt knob in basal region. Sternal furca absent.
Armature on rami of legs 1–4 as follows (Roman numeral indicating spines and Arabic numeral, setae):
Leg 1 (Fig. 2B) protopod with long, plumose outer seta and another small, plumose inner seta, in addition to a papilla bearing 2 long setules near outer margin of coxa; vestigial endopod small conical protrusion; first segment of exopod with row of spinules on posterior edge and short spiniform seta at outer distal corner; middle 2 of 4 terminal elements on last segment of exopod without accessory process but pinnate bilaterally; large pecten at base of each distal 3 spines; 3 plumose setae on posterior margin about half length of terminal element 4. Leg 2 (Fig. 2C) coxa small, with large, plumose, inner seta on posterior edge and small papilla tipped with a short seta on ventral surface; basis with short, naked outer seta; posterior margin of basis with long, narrow membrane and long marginal setule near midway of this membrane; outer spines on exopod (Fig. 2D) differentially constructed, with largest one on proximal segment and smallest one on middle segment. Leg 3 (Fig. 3A) protopod (apron) with small, plumose, outer seta and long, plumose inner seta in addition to an outer and a posterior marginal membrane; one short and one long setule with each at tip of a papilla just inner to baseline of posterior marginal membrane; 2 patches of denticles on ventral surface; velum well developed and fringed with marginal setules. Leg 4 (Fig. 2E) protopod with small, pinnate seta at outerdistal corner; pecten at base of innermost outer spine on exopod (Fig. 2F). Leg 5 (Fig. 3B) represented by 2 papillae on posterolateral margin of genital complex, one tipped with a pinnate seta and another one, 2 pinnate setae.
Body (Fig. 3D) short and broad, 2.31 (1.72–2.70) mm long, excluding setae on caudal rami. Cehalothoracic shield subcircular, 1.27 (0.96–1.48) mm long and 1.22 (1.00–1.56) mm wide, excluding hyaline marginal membranes. Fourth pediger 4.44 times wider than long. Genital complex shaped as a jar, 0.51 (0.36–0.58) mm long and 0.55 (0.44–0.60) mm wide. Abdomen 2-segmented; proximal segment short, 0.09 (0.06–0.12) mm long and 0.30 (0.22–0.38) mm wide and anal segment about as long as wide, 0.26 (0.20–0.34) × 0.24 (0.18–0.28) mm. Caudal ramus 1.29 (0.09×0.07 mm) times longer than wide and armed as in female. Antennule (Fig. 4A) generally as in female, but proximal segment armed with 29 setae. Antenna (Fig. 4B) with corrugated, adhesion pad on mediodistal surface of middle segment; terminal segment with lamellate cuticular tip and 2 basal setae with one on each side. Postantennary process (Fig. 4B) stronger than in female, with prominent terminal hook. Dentiform process of maxillule (Fig. 4C) bearing subterminal tine. Maxilliped (Fig. 3E) with small medial seta in middle region of subchelum. Leg 5 as in female but leg 6 represented by a papilla bearing 2 setae located in distal region of ventral ridge on genital complex (Fig. 4D).
The species name latus is a Latin for “broad” or “wide”. It alludes to the possession of broad cephalothoracic shield and genital complex in both sexes of this species.
Three species of Metacaligus are currently known, they are: M. rufus (Wilson, 1908), M. uruguayensis (Thomsen, 1949) and M. hilsae (Shen, 1957). The last species differs from all other species of Metacaligus, including the present new species, in the possession of a long, 4-segmented abdomen, which is longer than its genital complex (Shen, 1957). The first species differs from the present species in having the middle segment of the exopod of leg 2 equipped with a large outer spine, which is as large as the outer spine on the basal segment of the same ramus, and the outermost spine at the tip of leg 1 exopod longer than the segment carrying it (Wilson, 1908; Ho and Bashirullah, 1977). Thus, M. uruguayensis bears the closest resemblance to the present new species. Nevertheless, M. latus is distinguishable from M. uruguayensis in the possession of the following four characters: (1) broader cephalothoracic shield and genital complex in both sexes, (2) shorter caudal ramus in both sexes, (3) papillae on the postantennary process tipped with 4 setules in both sexes, and (4) the anal somite in male distinctly longer than the proximal abdominal somite.
Both M. uruguayensis and M. latus are found in the gill cavities of the cutlassfish taken from the Strait of Taiwan. However, in our past two years of examination of the cutlassfishes landed at Dong-Shih Fishing Market, these two congeners have never been found co-existing on a fish host. Another intriguing point to mention is about the occurrence of the chalimus larvae of M. latus. On 27 April and 22 June 2001, chalimus larvae in various stages of development (from chalimus II to chalimus IV) were recovered attached randomly to the cephalothorax, genital complex or abdomen of the adult of both sexes. This is unusual in that the chalimus larval stages are always found attaching directly to the fish host. Of the 20 chalimus larvae, 8 were found attached to the female and 12, to the male.
VALID CALIGID GENERA
In their work on the parasitic copepods of Brazilian fishes, Boxshall & Montú (1997) provided a key to 29 genera of the family Caligidae. It is the latest published work showing the recognized genera of the Caligidae. However, it is worthy to point out that while Dojiri (1983) considered Caligopsis and Diphyllogaster insufficiently known and relegated Heniochophilus to a synonym of Anuretes, these three genera were recognized valid by Boxshall & Montú (1997). Further difference between those two works is the treatment of Midias Wilson, 1911. This genus was not found in the key given by Boxshall & Muntú (1997) but considered valid by Dojiri (1983).
According to Dojiri's (1983) redescription of Midias lobodes, based on Wilson's (1911) type material kept at the Smithsonian Institution, this parasite of the great barracuda, Sphyraena barracuda (Walbaum), is not attributable to Caligus as considered by Kabata (1979) and followed by Boxshall and Montú (1997). Thus, in following Dojiriís (1983) opinion, Midias is treated in this paper as a valid genus of the Caligidae.
In their preliminary review of the 19 species of Anuretes Ho and Lin (2000) concurred with Dojiri's (1983) treatment to relegate Heniochophilus to a junior synonym of Anuretes. Our examination of the original works of Caligopsis by Markewitsch (1940) and Diphyllogaster by Brian (1899) showed that both genera had not been fully characterized for making an evaluation of their taxonomic status. Thus, both Caligopsis and Diphyllogaster are considered incertae sedis and, therefore, 28 genera of caligid copepods are recognized in this paper. They are listed in Table 1.
The valid caligid genera.
Inasmuch as all caligids are known of their females and the male is unknown in seven genera, the homologous characters utilized in the following analysis are confined to the adult females. Furthermore, in search for homologous characters in a genus comprising several species with different states of character, the ones possessed by the type-species are selected disregarding their status in the order of the character transformation. For instance, in the genus Caligus where there are more than 200 species with many of them bearing 3-segmented exopod on leg 4, the state of 2-segmented ramus is selected for this character even though 3-segmented state is more primitive. This is simply because the type-species of this genus, C. curtus Müller, 1785, has a 2-segmented exopod on its leg 4.
The characters selected for performing the analysis of phylogenetic relationships between the 28 caligid genera (Table 1) are listed in Appendix A with the code for each state indicated in the parentheses that follow. All together 36 apomorphic states were recognized and the coding of those 36 apomorphic states are given in Appendix B. It is from this list an imput file was created to conduct the cladistic analysis.
Ten of the 23 selected characters exhibit multiple states of character transformation. The series of transformation were coded in an ordered fashion that reflects a predetermined direction of character evolution (see Appendix A). Inasmuch as an analysis with ordered character coding has been criticized as running the risk of predetermining the topology of the resultant cladogram (O'Grady and Deets, 1987; Dojiri and Deets, 1988; Ho, 1991; Ho, 1994), both ordered and unordered analysis were performed to search for the best resolved cladogram(s) of the Caligidae.
Analysis with ordered coding
In employing the computer program HENNIG86 to perform the cladistic analysis, the algorithm “mhennig*” uses the available data, such as the matrix in Appendix B, to construct initial cladogram(s) and saves all shortest trees (cladograms) after running a limited branch-swapping. Five such trees were produced in this analysis. Then, with the application of “bb*” command the computer uses those five initial trees to perform extended branch-swapping to produce multiple shortest trees. As a result of this combined algorithms, a total of 668 equally parsimonious trees were generated, with a length of 96 steps, a consistency index (CI) of 40 and a retention index (RI) of 55. Close examination of these 668 trees revealed that there are seven groups of cladograms distinguished by the number of nodes that they carry.
As shown in Table 2, the number of nodes in those seven groups of trees runs from 20 to 26, with 1/3 (226/668 trees, 33.83 %) of the obtained cladograms showing 23 nodes on the tree. The highest number of nodes on a tree is 26 and only three of the 668 trees show the best resolution like this. It is called the best resolved because only one place (node) on these trees shows a polytomy where there are three branches (clades). There are ten least resolved trees with 20 nodes on the cladogram, which displays five places (nodes) with 17 clades (branches) occurring in polytomy, three nodes with three clades and two nodes with four clades (see Table 2).
Number of trees in each of the seven groups of cladograms (identified with the number of nodes that they carry) obtained in the analysis with ordered coding.
In the output file of the analysis performed by employing the algorithms “mhennig*” and “bb*” with the ordered coding, the three best-resolved trees were designated as Trees 27, 50 and 280. They are reproduced in Fig. 5, with Tree 50 shown in its entirety. As to Trees 27 and 280 only the parts different from Tree 50 are shown. The difference between Tree 50 and Tree 27 is the displacement of Lepeophtheirus, moving one step (node) closer to the stem, and between Tree 50 and Tree 280, the displacement of Sinocaligus, jumping from Node 48 to Node 47. It should be emphasized that the topology of the 18 genera (Pseudolepeophtheirus, Pseudocaligus, Belizia, Apogonia, Metacaligus, Anchicaligus, Sciaenophilus, Ecuetus Caligodes, Caritus, Hermilius, Abasia, Dartevellia, Arrama, Kabataella, Anuretes, Mappates, and Pseudanuretes) on Clade (Node) 46 of Tree 50 is exactly the same between these three best-resolved trees.
Analysis with unordered coding
Unlike the previous analysis, application of the algorithm “mhennig*” produced four initial, shortest trees (cladograms), and, with the execution of “bb*” it generated 3,165 equally parsimonious trees. The length of these trees is 87, with a consistency index (CI) of 44 and a retention index (RI) of 55. Although the tree length and CI are improved over the previous analysis with ordered coding, the value of RI remained unchanged and the overall resolution of the obtained trees is not as high.
Nine groups of cladograms are recognized with the number of nodes on a giving tree varying from 16 to 25. In other words, the best resolved tree (with 25 nodes) is not as good as those produced in the previous analysis and the least resolved tree (with 16 nodes) is worse than those found in the previous analysis. While there are 11 best resolved trees (Trees 439, 454, 493, 496, 498, 3044, 3046, 3054, 3056, 3059 and 3061) there is but one least-resolved tree (Tree 1180). Those 11 best-resolved trees can be divided into two groups with one carrying a polytomy comprising four branches (Trees 493, 496, 3056 and 3059) and the other, two polytomies with each carrying 3 branches (Trees 439, 454, 498, 3044, 3046, 3054 and 3061). One tree from each of these two groups was reproduced in Fig. 6. Tree 496 is shown with its entirety and Tree 3061, only the parts that are different. The least resolved tree, Tree 1180, has five polytomies, with two nodes containing three branches and another three nodes, five branches.
It is apparent from the above two sections that analysis of caligid phylogeny with ordered coding generated better-resolved trees, producing three cladograms with each carrying 26 nodes. Having only one node with three branches, any of these three trees are next best to the completely resolved tree where every node has dichotomy. Since Tree 50 occupies the central role with only one difference from each of the two other 26-node trees, it is adopted as a working hypothesis for the phylogeny of the Caligidae.
Examination of the transformation of each character on Tree 50 (see Appendix A) shows that Character 10 (outgrowth on the abdomen) changed at Node 49 from the plesiomorphy (“0”, smooth and cylindrical) to an apomorphy (“1”, bearing lateral aliform structures), and at Node 46 it reverted to the original plesiomorphy. It is this character reversal (homoplasy) of the abdominal outgrowth that serves as a synapomorphy to unite those 18 genera on Clade 46 to a monophyletic clade.
The absence of sternal furca (apomorphy of Character 13) is a synapomorphy at Node 43 (Table 3) and that places the afore-mentioned 18 genera in a monophyly excluding Pseudocaligus and Pseudolepeophtheirus. It is interesting to point out that the sternal furca is present in the species of both Anuretes and Hermilius, but these two genera are shown on the three trees in Fig. 5 as the descendents of an ancestor lacking the sternal furca.
Metacaligus and Sinocaligus are the only two genera of caligids known to have two outer spines on the terminal segment of the exopod of leg 2 (Character 18 coded with “0”), but they are not in a sister group relationship through the sharing of this unique apomorphy. In other words, according to the adopted caligid phylogeny, the evolution of the loss of one outer spine on the terminal segment of leg 2 exopod had occurred independently in two lineages, one lead to the development of Metacaligus and another one, the genesis of Sinocaligus. The same phenomenon is found on the character state of 16(2) (absence of 3 medial plumose setae on terminal segment of leg 1 exopod). This derived character is present in only three of the 28 recognized caligid genera: Apogonia, Kabataella and Parechetus. But, the result of analysis did not recognize it as a synapomorphy of the three genera.
From Tree 50 in Fig. 5 it is apparent that Metacaligus and Apogonia are sister groups on Clade 41, which is in turn on a monophyletic clade comprising the above-mentioned18 genera on Clade 46. Although these 18 genera are united by a synapomorphy of bearing a smooth, cylindrical abdomen (Character 10 coded with “0”), which is also the state of character found in Caligus, cladistic analysis based on the principle of parsimony did not identify the latter as a member of this 18-genera monophyletic clade. In other words, Metacaligus is only remotely affiliated with Caligus on the caligid phylogeny. Accordingly, treatment of Metacaligus as a genus of Caligidae separate from Caligus is justified.
The sister-taxa relationship between Apogonia and Metacaligus is not only found in the results of analysis with ordered coding (see Fig. 5) but also in the analysis with unordered coding (see Fig. 6). So far as we are aware, only one species of Apogonia, A. stocki Cressey and Cressey, is known. It is a parasite of the dwarf cardinalfish (Apogon mosavi Dale) occurring in the Caribbean Sea (Cressey and Cressey, 1990). It is intriguing to point out that a species of Metacaligus, M. uruguayensis (Thomsen), is also known from the Caribbean Sea (Ho and Basirullah, 1977).
We would like to thank Ching-Kuo Chang of Dong-Shih Fishing Port for his kindness and cooperation in making the necessary arrangements for us to purchase the fishes landed at the fishing port. Our appreciation is also due to Ping-Ju Wu and Wei-Shen Tzeng of the National Chiayi University for their assistance in transportation and examination of fishes for copepod parasites. The field and the laboratory works of this project was made possible through the grants (NSC 87-2313-B-021-012 and NSC 87-2313-B-021-18) from the National Science Council of Taiwan to Ching-Long Lin. Completion of this manuscript was aided by another grant from the Paramitas Foundation to Ju-shey Ho.
Characters and character states used in the claldistic analysis of Caligidae. Numbers in parentheses denote the numerical coding of the character states. Coding in Characters 2, 8, 9, and 18 employed “internal rooting” proposed by O'Grady and Deets (1987). Euryphorus and Trebius were utilized as the outgroup in polarization of the character states transformation.
Data matirix of 23 characters and their states in 28 genera of Caligidae as used in the cladistic (phylogenetic) analysis. The question mark “?” indicates an unknown state. Dut to the application of “internal rooting” (O'Grady and Deets, 1987) those characters coded with “1” in the outgroup is treated as plesiomorphic and “0” in the ingroup, apomorphic.