On the taxonomy of Latonigena auricomis ( Araneae , Gnaphosidae ) , with notes of geographical distribution and natural history

The male of Latonigena auricomis Simon, 1893 is described for the first time and the female is redescribed. New records are provided for Argentina, Brazil and Uruguay. Notes on the natural history and a potential distribution model of the species are presented in the Neotropical Region.

Gnaphosidae is one of the most speciose spider families, with 2,128 species distributed along the world (Platnick, 2012).Most taxonomic studies have been developed in Palearctic, Nearctic, Oriental and Australasian representatives meanwhile many Neotropical groups are still understudied.In this region, previous genera revisions and other taxonomic contributions were made mainly by Norman Platnick and collaborators (see Murphy, 2007;Platnick, 2012).The genus Latonigena was proposed by Simon (1893a) for a single species, L. auricomis Simon, 1893 based on a female and two immatures from Buenos Aires, Argentina (Simon, 1893b).Later a second species was described from South Africa, Latonigena africanus Tucker, 1923.The genus and the species were described using somatic characters as ocular disposition, body coloration and the morphology of carapace, chelicerae, sternum and coxae.The original description of L. auricomis did not include figures of the genital features so the taxonomic status of the genus and the species were so far not clear.The generic name Latonigena is feminine, but in Roewer's catalogue (Roewer, 1954) the epithet auricomis was changed to the masculine auricomus.Although Bonnet (1957) corrected this mistake in his Bibliographia Araneorum, later Murphy (2007) and Platnick (2012) used Roewer's nomination.Murphy (2007) included L. auricomis in the Herpyllus group by the following characters: medium size spiders, dorsal pattern in the abdomen with dark and whitish coloration and the presence of a dorsal scutum in anterior abdomen of males.Considering that L. africanus has not the dorsal pattern in the abdomen, this author included it in the Echemus group, suggesting that this species is not congeneric with L. auricomis.Moreover a recent study proposed seven new South American species of Latonigena and provided a diagnosis for the genus, contributing to define its taxonomic limits and formally transferred L. africana to Trichothyse Tucker, 1923 (Ott et al., 2012).The study of collections from Argentina and Uruguay and recent field trips in both countries allowed us to obtain several specimens of both sexes of L. auricomis.Therefore the main objectives of this paper are: the redescription and illustration of L. auricomis, with the first description of the male, the citation of new records with an analysis of its geographic distribution and the report of data about the natural history of the species.Murphy (1987) and Ott et al. (2012).

Specimens
A species distribution model was elaborated using Maxent 3.3.1,a maximum entropy based program (Phillips et al., 2006).The occurrence data were obtained from the records of the collections studied and in two cases, the specimens were identified in the field, sexed but not collected (see list below the section "Other material examined").The climatic variables were obtained from the database WorldClim 30 sec, with a resolution of 1 km (Hijmans et al., 2005a), that comprises nineteen different bioclimatic variables related to temperature and precipitation.They are widely used in researches related with the ecophysiological tolerances of a species in the habitat (Graham & Hijmans, 2006;Dias et al., 2011).To determine whether climatic variables have a greater impact on the model we use a Jackniffe analysis to measure the weight of each variable.For developing the model we choose the "equal training sensitivity plus specificity" threshold rule, with the following options: 2,500 maximum interaction, random test percentage of 25%, logistic output formatted, remove duplicates from the same grid cell and auto features (Dias et al., 2011).The total area modeled was situated between the following coordinates: latitude -56.4939 to -21.9056 S, longitude -75.8 to -41,6340 W. The model was evaluated using the Maxent function Receiver Operating Characteristic (ROC) (Hanley & McNeil, 1982).The resulting map of the model was elaborated with DIVA-GIS 7.0 (Hijmans et al., 2005b).Taxonomic note.A female (MNHN -AR9782) from "Banda Oriental del Uruguay", currently República Oriental del Uruguay was considered by Murphy (2007) as the holotype but in the original description it was indicated Buenos Aires, Argentina as the type locality.The body length and the somatic characters of this female agree with the Simon's description.The collector Friedrich Wilhelm Karl Berg, a German naturalist who lived and died in Buenos Aires, also stayed in Uruguay for two years.The species is present in both countries so the actual confirmation of the type locality is difficult to determine.

Latonigena auricomis Simon, 1893
Diagnosis.Females of L. auricomis resemble to those of L. beni Ott et al., 2012 andL. lami Ott et al., 2012 by the shape of the epigynum (without anterior hoods), but differ by having a more elongated and narrow median septum (Fig. 5).Males are very similar to those of Latonigena taim Ott et al., 2012(see Ott et al. 2012, figs 53-58) by the ventrally curved dorsal branch of the RTA with a basal hump, however, the bulb of L. auricomis is larger, the dorsal branch of the RTA is longer, and the basal hump is less prominent (but see variation below) (Figs 7-18).
Variation.Dorsal coloration.Some specimens present a black area dividing the second transversal light band in two sectors.Measurements, females (5 specimens): carapace long, 3.3-3.7;carapace width, 2.0-2.5;femur I length, 1.7-2.2.Measurements, males (5 specimens): carapace long, 1.9-2.4;carapace width, 1.2-2.0;femur I length, 1.6-2.4.We found variations in some characters of the palp morphology.Among Argentine specimens, the specimen from Campo de Mayo (MACN-Ar 28737) have the RTA more separated (distally diverging) from the bulb, and the basal hump of the RTA well developed, as in L. taim (Figs 7,11,15); the male from General Roca (MACN-Ar On the taxonomy of Latonigena auricomis (Araneae, Gnaphosidae)... 18364), the Southernmost record, in Northern Patagonia, shows the RTA closer to the bulb, which have more elongated proportions than other specimens (Figs 8,12,16).The male from Burzaco (MACN-Ar 18365) appears as intermediate between the above mentioned Argentine males (Figs 9,13,17).In Uruguayan representatives, the males have the RTA close to the bulb (Figs 10, 14, 18, FCE 2875), except one specimen from Paysandú (FCE 2889) where this structure is more separated.These findings show that the RTA morphology of L. taim could fit among the range of variability observed across the Argentine and Uruguayan males.Given that we have not found significant differences among all the females examined in the studied area, and that L. taim was described on base of a single male, the finding of females of L. taim will clarify the taxonomic status of this species.On the taxonomy of Latonigena auricomis (Araneae, Gnaphosidae)... Natural history.In natural areas this species can be found in rocky hills, under stones, usually near the nests of Camponotus mus Roger, 1863 (Hymenoptera, Formicidae) to which the spiders are strikingly similar, and inside the nests of Anumbius annumbi (Vieillot, 1817) (Passeriformes, Furnariidae).Several records of L. auricomis were obtained in urban areas from Argentina, Brazil and Uruguay (under wall crevices inside or outside houses, in gardens, and under bark of Eucalyptus in urban parks or plantations).This fact suggests that this species seems to be secondarily adapted to synanthropic conditions.
Geographic distribution.The distribution model indicates that the most suitable area for the species is situated on southern Uruguay and province of Buenos Aires, Argentina, but the distribution of the species could comprise almost all the Uruguayan territory, and parts of the states of Rio Grande do Sul and Santa Catarina in Brazil and the provinces of Buenos Aires, Entre Ríos, Santa Fe, Córdoba, and Corrientes in Argentina (Fig. 19).The presence was set over the 0.196 logistic threshold and the algorithm converged after 720 iterations.Others thresholds were selected and the same performance of the model was obtained.The AUC test for training data was 0.960 that indicate a very good prediction of the model.The Jacknife analysis showed that the Precipitation of Driest Month (38.7%),Isothermality (20.5%) and Temperature Seasonility (18.2%) were the most important climatic variables for the model contribution.Furthermore, the climatic variable that decreases the gain the most when it is omitted was Precipitation of driest month which therefore appears to have the most information that isn't present in the other variables.Most of the records were registered in modified environments such agroecosystems and urban areas.In Argentina, five points were situated outside the suitable area: Santa Rosa (La Pampa), General Roca (Río Negro), Calamuchita (Córdoba), Alta Gracia (Córdoba) and Bahía Blanca (Buenos Aires).

DISCUSSION
The distribution of this species appears to be mostly influenced by one pattern of precipitation and two of temperature.The precipitation of the driest month is an index that indicates the month with the lowest cumulative precipitation total (O'Donnel & Ignizio, 2012).The high value of this variable in the percentage contribution of the model suggests that the potential range of L. auricomis is more influenced by extreme precipitation conditions.Isothermality is a quantification of how large is the monthly diurnal temperature oscillation in comparison to the year oscillation (Xu & Hutchinson, 2011).Temperature seasonality indicates the temperature variation along the year calculated on the basis of the standard deviation of the monthly mean temperatures (O'Donnel & Ignizio, 2012).The influence of Isothermality and the Temperature Seasonality in the potential distributional range of L. auricomis indicates that this species is also sensible to temperature oscillations.This fact could explain that the best suitable conditions are represented in the model by a relative small area that comprises part of Argentina, Uruguay and Brazil.Therefore the species is mostly distributed in the province of Pampa, which is characterized by savannas with temperate grasslands and shrublands (Morrone, 2001).Only one record was reported by the province of Monte (General Roca).The synanthropic condition of L. auricomis is supported by the fact that most records were associated with human activities.The five records outside the suitable area could suggest that they could respond to peripheral populations of the species or probably others variables, not considered in the model, could be influencing in the distribution of this species (Rubio et. al, 2010).As was indicated, urban habitats bring good conditions to the establishment of some species than natural areas (Simó et al., 2011).Although L. auricomis was registered in native habitats, our results indicated that this species is beneficiated by urban areas in association with other species, and this would explain its presence out of predicted areas by the model.Future studies could be focused in the study of the relationships of L. auricomis with nest of birds or ants and the influence of the human activities in the distribution of the species.