The neotropical species of Mesocyclops (Copepoda, Cyclopoida): an upgraded identification key and comments on selected taxa

In this work we present an analysis of upgraded characters used in the taxonomy of cyclopine copepods to provide a new key for the identification of the members of the freshwater genus Mesocyclops recorded in the neotropical region. This analysis, which included revision of type and museum specimens, resulted also in the clarification of the taxonomical status of two nominal species not previously revised, M. varius Dussart, 1987 and M. venezolanus Dussart, 1987. These two species are recognized herein as junior synonyms of M. brasilianus Kiefer, 1933. Considering the addition of new records (introduced Afro‐Asian or Asian forms), the designation of M. araucanus Campos et al., 1974 to species rank, and the conflict of determining the taxonomical status of M. annulatus diversus, the number of taxa recognized in the region includes 20 species and one variety. A general morphological analysis of the Old and New World species allowed us to contrast and evaluate some of the differential characters of these two groups. We propose several characters that can be considered as potentially valuable to separate reliably both female and male specimens of the neotropical species. The distribution of selected species or species groups of Mesocyclops is also revised in light of this new frame.


Introduction
The world distribution of the freshwater copepod genus Mesocyclops G. O. Sars, 1914 is mainly tropical and subtropical (van de Velde 1984;Hołyń ska et al. 2003). With more than 70 nominal species, the genus is among the most diverse of the family Cyclopidae (Boxshall and Halsey 2004). Careful morphological analysis of species from Europe, Asia, and Africa (Kiefer 1981;van de Velde 1984;Hołyń ska 2000b) has allowed the identification of useful features for establishing the limits between species.
Around 20 species of Mesocyclops have been recorded in tropical America (Gutiérrez-Aguirre and Suárez-Morales 2001b). The morphology and distribution of some of these have been analysed in detail (Reid 1993;Reid and Reed 1994;Fiers et al. 1996;Reid and Moreno 1999;Hołyń ska et al. 2003). However, the specific determination of many neotropical forms has become a difficult task because they belong to taxonomically complex groups. An example of such a trend is represented by the M. meridianusbrasilianus complex [including M. pseudomeridianus, M. meridionalis, and M. varius (Hołyń ska 1994(Hołyń ska , 2000a], known mainly from South and Central America (Collado et al. 1984;Dussart 1987) and containing several taxa with an uncertain taxonomical status.
The taxonomical problems of many species may be related to different factors such as the unavailability or loss of type specimens, low sampling efforts, and lack of detailed, upgraded descriptions (Dahms and Fernando 1995). Most of the published descriptions of species of Mesocyclops do not include a morphological analysis of males; only female characters are generally used in keys and species diagnoses (Dahms and Fernando 1995). There is an urgent need to increase efforts in order to provide a complete, detailed morphological analysis of taxonomically complex groups. Consistent, robust characters will allow a reliable separation of females and males of tropical American species.
The most recent identification keys that include neotropical Mesocyclops were published by Reid (1985), Dussart (1987), and Reid and Pinto-Coelho (1994). However, some of the morphological features used in these works have a relatively low resolution to separate many of these species derived from the current development of the genus taxonomy. This work presents an analysis of new morphological characters for the identification of the neotropical species of Mesocyclops based on (1) the analysis of both sexes of each species and (2) a set of the current, upgraded morphological data used in the taxonomy of the genus. Based on this information, the taxonomical status of some species belonging to the M. brasilianus-varius complex is resolved, thus proposing a new synonymy for these nominal taxa.

Methods
While preparing this key for the identification of the neotropical species of Mesocyclops, we analysed and compared the morphological characters that have been proved to be useful to separate species from Europe, Africa, and Australasia, i.e. shape of hyaline membrane of last antennular segment, features of seminal receptacle, ornamentation of thoracic limbs, among others (see Hołyń ska 2000a, b). Additionally, other complementary features were evaluated as an aid to ease the identification of the neotropical Mesocyclops, such as the ornamentation of maxillular coxa, anal somite, intercoxal sclerites, etc., in both females and males.
The morphological structure of several species was examined by light microscopy. This process included observations of type and museum specimens deposited in the collections of different institutions: the Muséum National d'Histoire Naturelle, Paris (MNHN), the National Museum of Natural History, Smithsonian Institution, Washington, DC (USNM), the Institut voor Systematiek et Populatiebiologie, Amsterdam (ZMA), the Staatliches Museum fü r Naturkunde, Karlsruhe (SMNK), and the Instituto Nacional de Pesquisas da Amazô nia (INPA) (see Table I). The morphological terminology used herein follows that proposed by van de Velde (1984), Huys and Boxshall (1991), and Hołyń ska (1994).
The type specimens of M. ogunnus, M. chaci, M. yutsil, and M. araucanus were not observed, but they were included in the key following the descriptions and illustrated features published in van de Velde (1984), Fiers et al. (1996), and Pilati and Menu-Marque (2002). Mesocyclops annulatus diversus Herbst, 1962 was excluded from the key because discriminating characters for this species are not defined, i.e. the ornamentation of the frontal surface of the antennal basis, the presence or absence of spines next to the exopodal seta. Also, the depository institution of this material is unknown to us; in fact, the type specimens are probably lost (M. Hołyń ska, personal communication). The taxonomical status of species belonging to problematic groups was analysed in detail. A general discussion of the geographical distribution of the neotropical species is also presented herein. Kiefer, 1933 (Figures 1, 2) M. varius Dussart, 1987, junior synonym. M. venezolanus Dussart, 1987, junior synonym. ? M. venezolanus: Reid and Reid 1994 The bad condition of the type specimens of M. brasilianus prevented a detailed analysis; however, we based our criteria on the observations of 10 adult females from Itacoatiara, Amazonas, Brazil (03u10.8079S, 58u14.6319W, west from Manaus, established by Kiefer 1936 as the type locality). There is only one female labelled as M. varius in the world, from Taxisco, Guatemala (see Table I). Dussart (1987) recognized the resemblance between this specimen and M. brasilianus. Hołyń ska et al. (2003) stated that its taxonomical status remains questionable; however, they included this species as part of the M. meridianus-brasilianus complex. Our analysis indicates that the morphological features used by Dussart (1987) to separate these species are not consistent: the presumed absence of an acute angle on the outer margin of the first endopodal segment of the first swimming leg has proved to be erroneous ( Figure 1A-C). The same is true with respect to the assumed absence of ornamentation on the coxa and inner margin of the fourth leg basis ( Figure 1D-F). The holotype of M. varius bears a spermatophore fixed to its genital aperture ( Figure 1G), but no differences were observed on the structure of the seminal receptacle between M. varius and M. brasilianus ( Figure 1G, H). Another character used to separate M. varius was the length ratio of two caudal setae: dorsal seta/lateral seta [relatively longer in comparison with M. brasilianus (Dussart 1987)]. We found that the proportional length observed in the holotype specimen of M. varius shows a lower figure than that determined from the original description of M. brasilianus (1.2 versus 1.3). In addition, the comparison of the seta length proportions ( Figure 2B, E) in different populations of M. brasilianus observed here (Table I), through an exploration of frequencies, shows a wide variability, even in the same population ( Figure 3); therefore, we consider that this proportion is a weak feature for separating species. Dussart (1987) described another species, M. venezolanus; this was also recognized as being similar to M. brasilianus. The description was based on specimens collected in Lake Valencia, Venezuela. The differences by which M. venezolanus was separated from M. brasilianus were: caudal rami without spines on distal-lateral corner; abdominal somites with serrated hyaline fringe; dorsal seta on caudal rami shorter than outer seta; fifth thoracic somite with dispersed hair-like setae; seminal receptacle with two ''bumps'', followed by two curved horns. All these features are present in M. brasilianus too (Figures 1B, C, E, F, H, I, 2B, C, E, F), including the lack of spines on the distal-lateral corner of the caudal rami: no adult female of M. brasilianus examined here, even the specimens identified by Kiefer, showed those spines. The type specimen is useless for taxonomical analysis, and now we speculate that the spines shown in the illustrations, but not mentioned in the original description of M. brasilianus, indicate that this specimen is an immature CV female (Gutiérrez-Aguirre and Suárez-Morales 2003). Another specimen collected in Managuiri, Amazonas, Brazil, and identified by Kiefer (see Table I), does not have these spines. Therefore, the spine presence/absence pattern that would separate M. brasilianus and M. venezolanus is not a usable character. Therefore, we conclude that both M. varius and M. venezolanus are junior synonyms of M. brasilianus and should be considered as that in future taxonomic accounts of neotropical Mesocyclops. However, the identity of the specimens recorded by Reid and Reed (1994) as M. venezolanus from the Yukon Territory remains unverified, taking into account the differences of the hairornament observed by Hołyń ska et al. (2003) in comparison with M. brasilianus.
Mesocyclops longisetus and related forms. The differences in length/width ratio of the caudal rami, and the third endopodal segment of the fourth leg, as well as the shape of the lateral arms of the seminal receptacle, have been used to separate M. longisetus s. str. and M. longisetus var. curvatus (Dussart 1987;Reid and Pinto-Coelho 1994). The species of Mesocyclops are defined by differences in the ornamentation of cephalic or thoracic appendages, or the seminal receptacle shape; therefore, the status in these taxa remain without changes. However, features such as the presence of spines on the base of the lateralmost terminal caudal seta, a higher length/width ratio of caudal rami, and the structure of the third endopodal segment of the fourth leg, were important to distinguish M. araucanus as a separate form with species rank, different from M. longisetus s.str. (Pilati and Menu-Marque 2002). These forms have a latitudinal distribution between 51uS and 10uN in the Americas: M. longisetus s.str. has been recorded in Mexico, the Guadeloupe Islands, Venezuela, and Guatemala, whereas M. longisetus var. curvatus occurs in Mexico, Panama, Brazil, Guadeloupe Islands, Honduras, southern USA, and Canada. Mesocyclops araucanus has a latitudinal distribution between 39u369S and 51u359S (Pilati and Menu-Marque 2002). We verified its presence in Las Chulta Lagoon, Argentina (see Table I).
Mesocyclops annulatus diversus. The subspecies was described by Herbst (1962) from specimens collected in Rio Urindéua, Brazil. However, the original description shows morphological differences between the strict form (recorded in Argentina), and the subspecific form: the length ratio of the apical terminal spines on the third endopodal segment of the fourth leg is considerably longer in M annulatus diversus than in M. annulatus; the seminal receptacle of the subspecies is more similar to M. paludosus Lindberg, 1956, a species restricted to East Africa, with wide arms, instead of the thin arms present in the strict form. Furthermore, M. annulatus has hair-like setae on the ventrolateral surface of the fifth pediger versus a naked condition in M. annulatus diversus. It is suggested that the specimen analysed by Herbst (1962) is not a subspecies of M. annulatus, but probably a different taxon at a species level. The depository institution of this material is unknown to us; in fact, it is probably lost (M. Hołyń ska, personal communication). We speculate that a revision of specimens from the type locality would result in the change of its status to become a species. On the other hand, M. annulatus appears to be restricted to South America with records in Argentina, Bolivia, Chile, Paraguay, Peru, and Uruguay (Gutiérrez-Aguirre and Suárez-Morales 2001b).
Mesocyclops ellipticus-M. reidae. The record of M. ellipticus from caves in Yucatan, Mexico by Yeatman (1977) was later assigned to M. reidae by Petkovski (1986) and Reid (1993). However, analysis of Yeatman's illustrations clearly shows an intriguing mixture of features of both M. ellipticus and M. reidae in these cave-dwelling specimens. For instance, presence of two rows of hair-like setae on the fourth intercoxal sclerite, hair-like setae on the inner margin of the caudal rami, as depicted in the original description of M. ellipticus (see Kiefer 1936), and a group of spines next to the insertion of the exopodal seta on the antennal basis (relevant features of M. ellipticus). The oval seminal receptacle, and the inner caudal seta short, no more than 1.5 times the length of lateral seta are important features in M. reidae. The original material, collected by Yeatman from Grutas Xtacumbilxunam, Campeche, Mexico, is lost (H. C. Yeatman, personal communication). New collection efforts in the same area are likely to produce additional specimens in order to determine the status of the Yucatan record of M. ellipticus. The diversification patterns of the genus in the Yucatan Peninsula, with closely related species in geographically restricted areas (Suárez-Morales et al. 2004) suggest that Yeatman's specimens could represent a new, endemic taxon. On the other hand, the record by Herbst (1962) in Amazonas is most probably not referable to M. ellipticus because the analysed specimen shows features not present in the species: the seta on the inner basis of P1 is very delicate, the hyaline membrane on antennal segment 17 is completely serrated, without a distal notch, and the fourth intercoxal sclerite lacks cuticular projections and hair-like setae. In fact, the specimen examined by Herbst (Dussart and Frutos 1985), M. chaci, M. yutsil, in caves and cenotes from Yucatan, Mexico , and M. intermedius (hypogean waters from Bonaire) (Pesce 1985). Mesocyclops paranaensis appears to be restricted to South America, with valid records in Argentina (Dussart and Frutos 1985) and probably in Brazil, and Paraguay (Lowndes 1934) (see Hołyń ska et al. 2003 Species with wide distribution. The distributional patterns of other well-known species of Mesocyclops were also revised. For instance, M. reidae has been recorded in North and South America, as well as in the Caribbean region (Colombia, Cuba, Haiti, Jamaica), Cayman Islands, Mississippi, and Mexico (Petkovski 1986;Reid 1993). Mesocyclops edax is considered to be the most widely distributed species of this genus in the Nearctic region, but with additional neotropical records in Central America, Antillas (Reid and Moreno 1999), and south-east Mexico .

Morphological remarks
Most Australasian species of Mesocyclops show a peculiar combination of features in the trunk limbs: (1) presence of large spines on the medial expansion of the basis, first leg (+); (2) fourth intercoxal sclerite with pointed projections on distal margin (+); and (3)  pseudomeridianus. Therefore, in the New World Mesocyclops, there is a tendency to reduce the spines on the first leg basis, and a persistence of a seta on the inner margin of the first leg basis; of course, a phylogenetic analysis is urgently required in order to elucidate the evolutionary change of the morphological features of the genus.
In the same sense, the presence of group d (transversal row of tiny spines) on the antennal basis, caudal surface, together with the shape of the channel connected to the copulatory pore (resembling a ''comma''), are features present only in species that probably originated in the Old World (M. aspericornis, M. ogunnus, M. pescei, and M. thermocyclopoides). Contrastingly, this group of spines is absent and the channel is wide and bent in most neotropical species.
Another unique feature observed only in neotropical species (except by the Australian M. darwini Dussart and Fernando, 1988), is the division of the hyaline membrane at the insertion of the medial seta (on females, last antennular segment). This character is present in M. brasilianus, M. meridianus, M. meridionalis, and M. pseudomeridianus. The importance of the maxillular ornamentation for distinguishing species was noted only in reference to the palp of the African species M. ogunnus. In this study, we found that when it is present, the ornamentation of the praecoxal surface has taxonomic value, similar to that attributed to the ornamentation of maxillar coxa, fourth leg coxa or shape of the seminal receptacle, because this structure (among others) is an important feature for distinguishing M. brasilianus, M. evadomingoi, and M. meridianus.
On the other hand, because many morphological differences were found among the examined species, the following features can be considered as potentially valuable to separate the neotropical species: (1) presence/absence of ornamentation on antennular segments 1, 4, 5, and 6 in the female; (2) presence/absence of ornamentation on antennular segments 15 and 16 in the male; (3) ornamentation on the antennal basis; (4) presence/ absence and type of ornamentation on the maxillular praecoxa, posterior surface; (5) ornamentation on both coxa and basis of first and fourth swimming legs; (6) shape of seminal receptacle; (7) ornamentation on abdominal somites; (8) presence/absence of ornamentation on the furca; and (9) presence/absence of ornamentation on the anal somite (posterior margin, and ventral or dorsal surfaces). These features were valuable for the identification of both female and male specimens.