A new species of Rhytidognathus (Carabidae, Migadopini) from Argentina

Abstract The Migadopini are a small tribe of Carabidae with 47 species that occur in South America, Australia, and New Zealand, in the sub-Antarctic areas. In South America, most of the genera inhabit areas related to sub-Antartic Nothofagus forest except two monogeneric genera, the Ecuadorian genus Aquilex Moret and the Pampean genus Rhytidognathus Chaudoir. These two genera are geographically isolated from the remaining five South American genera. New material of Rhytidognathus from the northeast of Buenos Aires province and from Entre Ríos province permits establishing that the previous records of Rhytidognathus ovalis (Dejean) for Argentina were erroneous and that it belongs to a new species. Based on external morphological characters and from male and female genitalia we describe Rhytidognathus platensis as a new species. In this contribution we provide illustrations, keys, habitat characteristics and some biogeographic considerations on the distribution of Rhytidognathus.


introduction
The Migadopini are a small tribe of Carabidae, with 16 genera and 47 species. This tribe was considered related to the Holarctic tribes Elaphrini and Loricerini (Jeannel 1938), and Loricerini (Maddison et al. 1999). Ball and Erwin (1969) considered that the characters shared with Loricerini are convergent and do not show an ancestral relationship. The most modern classification considers the Migadopini as constituting the subfamily Migadopinae, together with the tribe Amarotypini (Johns 2010).
The species of Migadopini are distributed over fragments of the austral Gondwana, called Paleantarctic by Jeannel (1938). These species occur in southern South America (eight genera with con 17 species) (Roig-Juñent 2004), one monotypic genus in the Andean region of northern South America (Moret 1989), four genera with seven species in Australia (Baher 2009) and four genera with 19 species in New Zealand and circum-Antarctic islands (including a new genus and several new species not yet described) (Johns 2010). The only complete revision of the tribe is that by Jeannel (1938). Later, for South America, Straneo (1969), Négre (1972), and Baher (1997;1999) described new species or subspecies, Moret (1989) described a new genus and species and finally Roig-Juñent (2004) redescribed all the austral South American genera including male and female genitalia characters and developed a cladistic and biogeographic analysis of the genera. For Australia, Baher (2009) described a new genus with two species, and for New Zealand, Johns (2010) described 11 new species.
The number of species per genus is low. Of the 16 genera, eight are monospecific, four have two species and the most diverse in number of species is Taenarthrus Broun with 12 species (Johns 2010).
Migadopines constitute a characteristic element of the sub-Antarctic biota, and except some frequent species such as the South American Migadops latus (Guérin-Ménéville) the others are scarce in natural history collections, with just a few specimens of several species known. This is the case for the genus Rhythidognathus Chaudoir of which only 12 specimens are known: the holotype of R. ovalis (Dejean), nine more specimens from Uruguay, and two from Argentina. Of these last two specimens, one is lost, and we only have the account by Tremoleras (1931). Strange as well is the particular distribution of the genus Rhytidognathus, because it does not inhabit sub-Antarctic habitats, and its phylogenetically related genera are about 3000 km to the south.
Ecological studies conducted in the area of La Plata (Buenos Aires, Argentina) yielded the discovery of new specimens of Rhytidognathus, and particularly the capture of males allowed establishing that the previously cited species of Rhytidognathus from Argentina (Tremoleras 1931, Roig Juñent 2004 is not R. ovalis but instead a new species.
The objective of the present contribution is to describe this new species, including new data on its habitat, and discuss some biogeographic considerations.
Dissection methods, measurements, and the terminology used follow previous revisions of Migadopini (Jeannel 1938, Moret 1989, Roig-Juñent 2004, Johns 2010. Predictive species distribution models were built using the MAXENT program version 3.4.1 (Phillips et al. 2006), because MAXENT performed well with small sample sizes (Tognelli et al. 2009), which is the case of Rhytidognathus. Also because of the low number of known species localities, we performed the analysis at generic level.
Redescription. Habitus. Body shape rounded, depressed ( Fig. 1) Head. Labrum short, transverse, bilobate at anterior margin; clypeus with two subparallel lateral sulci slightly developed, projected at the base of the frons (Figs 2, 5); mentum and submentum not fused, mentum with four setae, two lateral to the tooth, and two at the base; mentum-tooth bifid; glossa with a central carina, with two apical setae; glossa with two setae, paraglossae rounded, not projected; galea biarticulate, distal article as long as anterior one; mandibles with several dorsal transverse sulci; last maxillary and labial palpomeres long and truncate at apex; antennomeres three times as long as wide; antennae long, reaching the base of the elytra (Fig. 5); antennomeres fusiform, pubescent from the fifth antennomere (Fig. 8).
Pterothorax: mesoepisternum with deep punctures (Fig. 9); metaepisternum with a row of punctures and two apical sulci (Fig. 9); elytra twice as wide as than pronotum, without shoulders (Figs 4, 7), with borders rounded, elytra increasing in width to the apex, the widest part on apical third (Fig. 1); elytral epipleura more than twice wider at base than at apex, decreasing in width from base to apex; scutellar stria complete; striae with punctures, deep on the basal third, shallower on the second third and on apical third imperceptible, striae well delimited and deep all along their length (Fig. 4); setae only on ninth interval, with six or seven setae. Apterous.
Abdominal sterna. Sterna III-V constituting more than two thirds of the length of abdomen; sulcus of separation of sterna III-IV and IV-V not reaching the center; female sterna VIII without apical sulcus, with two apical setae. Sternite III and IV with deep basal punctures.  Comparative notes. The genus Rhytidognathus shares with Pseudomigadops Jeannel the characteristic of having the elytral striae punctured and differs from it by having the articles of maxillary and labial palpi elongated and thin, as well as by having the mandibles carined dorsally. This last character is exclusive to the genus within the tribe.
Head. Head with deep punctures in front, eyes slightly protruding, sub-quadrangular. Maxillary palpi black or dark red.
Pterothorax: Elytra. Humeral angles rounded (Fig. 4); striae well impressed, and deeply foveate on basal third (Fig. 4), being less marked toward the apex; six to seven setae only in the ninth interval.
Intraspecific variation. Jeannel (1938) found some intraspecific variation in the intensity of basal punctures of the pronotum and also in the coloration of the legs. Diagnosis. Head with small punctures, on the borders; elytra black with interstria 8 reddish; labrum with the borders yellowish; interstriae flat; apex of median lobe sub-quadrangular.
Head. Head with small punctures in front; eyes slightly protruding, rounded (Fig.  8). Maxillary palpi black or dark red.
Prothorax. Wider than long, maximum width at middle (Fig. 5); dorsal surface with punctures on the base (Figs 1, 5), apex with small or no punctures. Lateral margins narrow, curved; central longitudinal sulcus slightly developed; posterior transverse foveae slightly impressed. Posterior angles rounded. Prosternum without punctures or one or two on the apex. Prosternal projections not marginate, with a small apical tubercle, sinuate dorsally (Figs 6, 9).
Etymology. The name of the new species is related to the area where it was collected, La Plata district, near the La Plata river in Buenos Aires Province, Argentina.
Taxonomic considerations. Tremoleras (1931) cited Rhytidognathus ovalis for Argentina. Tremoleras` specimen was held in his collection and now we can not find it. The description by Tremoleras (1931) does not allow a clear identification of this material. Roig-Juñent (2004) cited also Rhytidognathus ovalis for Entre Ríos province (Argentina), based on a female. In the present contribution, this female specimen is now considered as being R. platensis. Taking into account that R. platensis is distributed along the western shore of the La Plata river, we considered it more likely that Tremoleras` specimen belongs to the new species, R. platensis, and not to R. ovalis.
Habitat. The new material was collected in the locality of Lisandro Olmos (La Plata, Buenos Aires) at "La Nueva Era" farm (35°01'18"S, 58°02'07"W) (Fig. 20), devoted to horticultural production under organic management (Fig. 21). The area has elevations of about 30 m, with soils derived from the Buenos Aires belt corresponding to grassland soils. It is surrounded by horticultural crops grown under cover and in the open, primarily tomato, pepper, leafy vegetables, celery, eggplant and small plots of corn, among others. Cut flower production in greenhouse conditions is also important in this area.
Samples were collected by pitfall traps set up in a 2000 m 2 -area cultivated with lettuce (Lactuca sativa), onion (Allium cepa), radish (Raphanus sativus), rocket (Diplotaxis sp.), cabbage (Brassica oleracea) and different types of weeds. This habitat has no native vegetation. Probably Rhytidognathus platensis inhabits the patches of semi-natural vegetation surrounding the crops. It has been proven that carabids move between cultivated and uncultivated patches (Marshall andMoonen 2002, Magura 2002).
On the shores of La Plata river in Buenos Aires province we found two natural habitats. One habitat is close to the river and includes: a) cliffs, with small forest of Celtis tala and other arboreal species, b) riparian shallows extending between the cliffs and the river and constituting a low plain that gets flooded, similar to the marshes of the Paraná river delta. The soil is clay and salty, and the vegetation is characterized by halophytic steppe with dominance of low grasses such as Distichlis spicata. The second habitat, the Pampean plain, lies above the cliffs. This lowland has a temperate climate, with an even year-round precipitation regime, soil type is loam, and the plants that dominate the landscape are herbs that compose the extensive Pampean grassland, a steppe. The typical original plant community comprises species of the genera Stipa and Piptochaetium. This landscape is accompanied on different sites by low shrubs of several species of Bacharis.
Predictive models of distribution show that the genus Rhytidontahus is restricted to the coast and areas close to the La Plata river and the delta of the Paraná and Uruguay Rivers (Fig. 20), occupying shore habitats and the Pampean grassland near the shore. This Pampean plain has been strongly modified, allowing for great agricultural development with establishment of annual crops and pastures, leaving hardly any native vegetation in the region. The Pampean grassland and forest close to the La Plata river and to the high Paraná River differ in species and habitat conditions from the areas inhabited by nearly all sister groups of Rhytidognathus, the genera Lissopterus Waterhouse, Migadopidius Jeannel and Pseudomigadops. Migadopidius occupy temperate Nothofagus forests (Fig. 24, Table 1). Lissopterus and Pseudomigadops (Figs 22-23) occur in habitats closer to the shore, principally sub-Antarctic forest or moorlands (Figs 22-23, Table 1). The unique genus of the sister group inhabiting grassland is Pseudomigadops, in some part of Malvinas Islands. As we can see, Pseudomigadops inhabits coastal forest and grassland, like Rhytidognathus, but species composition in their habitats is far from being the same, as the former is of sub-Antarctic origin and the other of Neotropical origin (Morrone 2004). Climatic conditions are not the same either, and if we look at the variables that explain the predictive models of distribution of these four Migadopini genera, the most important variable is temperature (Table 1).

Biogeographic considerations
Because of its particular distribution pattern and its phylogenetic relationships with other tribes, the Migadopini have been used to explain some very different biogeographic views, such as an austral origin and separation by vicariance (Jeannel 1938, Brundin 1966 or a Holarctic origin, separate dispersal to the southern continents, extinction in tropical and subtropical regions (Darlington 1965). Beyond the different proposals regarding the origin of the tribe, everybody considers that its current restricted distribution is relictual (Jeannel 1938, Darlington 1965. Upon the advent of the theory of plates as applied to the continental drift, it was put forward that many groups with distribution patterns similar to those of migadopines be considered of austral origin, whose fragmentation led to their present distribution. By applying a Dispersal and Vicariance analysis, Roig-Juñent (2004) put both hypotheses to test and his conclusions concur with Jeannel's saying that the tribe has had an origin in the southern hemisphere and that its current distribution across the southern continents  as Jeannel proposed (1938). This shows that some clades would have originated before the fragmentation of some parts of Gondwana.
Regarding the present distribution of the Migadopini in South America, it is restricted to three disjunct areas. The first is in the Ecuadorian Andes, where the genus Aquilex occurs at about 4300 m elevation at Páramo (Moret 1989); the second is on the shores of the La Plata river where Rhytidognathus lives in Pampean grassland and riparian forest environments; and the third, which is the largest in surface area and coincides with the sub-Antarctic region in Chile and Argentina, includes all Nothofagus forests and sub-Antarctic regions up to Cape Horn. The latter is the area with highest number of Migadopini genera, and where most taxa show more phylogenetic affinity to other taxa from southern regions (New Zealand, Australia) than to those from the rest of the Neotropics. Although the present distribution of the Migadopini is largely restricted to the sub-Antarctic region in South America, it is likely that, at some point of the Cenozoic, the tribe may have had a broader distribution. The sub-Antarctic biota expanded to more northern areas and its later retraction left areas with relictual distributions. Such is the case of the Fray Jorge forests in Chile (30° 40´44" S, 71° 40´54" W) or the Araucaria forests in the south of Brazil and north of Argentina (26° 27¨S, 53° 37´W). This expansion might explain the presence of Rhytidognathus in the La Plata river because, being apterous and large-sized, this taxon has almost no capacity for dispersal. Moret (1989) considers the same situation for the genus Aquilex, which would have originated from its southern ancestors in the pulses of northward expansion of the sub-Antarctic biota during the Cenozoic.
Considering the particular distribution of Rhytidognathus, the biogeographic analysis carried out by Roig-Juñent (2004) shows that this genus would have been split by a vicariant event from its sister group (Lissopterus + Pseudomigadops + Migadopidius) which now inhabits the Magellanic region or the northern Nothofagus forests. Although the distance to the Magellanic region exceeds 3000 km and is 1000 km to the Nothofagus forest region, the possibility of a vicariant event is feasible because, as mentioned for the austral region of South America, its cold austral biota experienced expansions during the Cenozoic whereby the genus came to occupy areas more northern than the current ones (Romero 1986, Barrera andPalazzesi 2007). So the separation of Rhytidognathus may have been caused either by vicariance or by isolation upon the southward retraction of the austral biota. Numerous are the relictual taxa than can be found in the Pampean region and south of Brazil, such is the case among carabids of the tribe Broscini.
In analyzing the environmental features of each genus, we find that there could also have been environmental features involved in the split. Figures 21-24 show the potential distribution range of Rhytidognathus and that of its sister genera. For these four genera, we find three clearly separate areas, one is austral sub-Antarctic, another one comprises the cold-temperate forests, and the third one encompasses the Pampean steppe and riparian forests along the La Plata river. The Pampean region is the exception with respect to the other habitats where migadopines occur in South America, and to the remaining circum-Antarctic regions, because most are from cold-temperate or cold environments, such as the species of Loxomerus Chaudoir (Johnson 2010). Although the Pampean grassland is a temperate area, it has warm summers and the vegetation is Neotropical in origin, not austral.
In other cases, it has been put forward that there often is niche conservation, commonly observed in species of the same genus whose potential distributions show areas occupied by other species of the genus rather than by them. However, we see that a shift has occurred among these four genera regarding the environment occupied by some of them. We propose that the environment occupied by the ancestor of Rhytidognathus and the sister group could have been cold-temperate coastal or riparian habitats, either forest or grassland (present in Rhytidognathus and Pseudomigadops). An arid barrier formed during the Cenozoic between the Pampean and sub-Antarctic regions (Barreda and Palazzesi 2007), isolating Rhytidognathus, and the current species of this genus would have had to become adapted to this more temperate climate.