Long-term seasonal dominance of the wasp Trihapsis polita Townes (Hymenoptera, Ichneumonidae) in the Brazilian Atlantic Forest

Abstract Background The temporal dynamics of insect populations in tropical environments is highly complex and poorly known. Long-term seasonality studies are scarce, and particularly so for ichneumonid wasps (Hymenoptera Ichneumonidae). This study represents an effort to elucidate aspects of seasonality and forest succession in the Brazilian Atlantic Forest. New information We report on the seasonal and successional dominance of the ichneumonid wasp Trihapsis polita (Cryptinae). A long-term survey of Cryptinae was carried out in a protected area of Brazilian Atlantic Forest, in primary, tall secondary and low secondary forest areas. Specimens were collected during rainy season (RS) and dry season (DS) between 2000 and 2008, with total sampling effort of 4,095 trap-days. A total of 8,385 specimens of Cryptinae were collected, of which 6,655 (79.4%) belonged to T. polita. The occurrence of T. polita species was heavily concentrated in the RS, with abundance 148× higher than during the DS. Seasonal fluctuation was also detected for Cryptinae as a whole, but was two orders of magnitude lower. Sampling efficiency also varied widely among areas, with the peak of abundance at the tall secondary forest. The dominance of T. polita in secondary vegetation might be of general interest, as this type of forest is currently on the rise, due to unprecedented levels of human pressure.


Introduction
The temporal dynamics of insect populations in rainforests is highly complex and poorly studied (Hamer et al. 2005;Wolda 1988). Long-term studies of insect seasonality in the tropics are scarce, largely due the difficulty in maintaining long-term sampling and sorting (Grimbacher and Stork 2009). For Ichneumonidae, one of the largest insect families and a dominant component of tropical faunas, data are particularly scarce, deriving mostly from species that are easy to collect and identify (Gauld 2006). For Cryptinae, its largest and least studied subfamily, with 408 genera and over 4,500 species, there are no detailed studies of seasonality for any Neotropical taxon. Cryptines occur in all terrestrial environments and are mostly ectoparasitoids of pupae and prepupae of a wide variety of hosts, including moths, butterflies, beetles, social and solitary wasps. Female Trihapsis polita Townes, habitus. Herein we report the seasonal distribution of a dominant cryptine species, Trihapsis polita Townes (Fig. 1). Trihapsis Townes is known from two Brazilian species, T. punctata Townes, and T. polita Townes, both described from a few specimens (Townes 1970). The genus seems to be restricted to the Brazilian Atlantic Forest, but is not frequently collected: a single specimen was obtained after an extensive collecting program in that region (Aguiar and Santos 2010). The hosts and other biological aspects of the genus are still unknown.

Sampling methods
Sampling description: Fieldwork was conducted over the course of eight years (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008). In each year, one field excursion was conducted during the dry season (DS, July-August), and one during the rainy season (RS, October-November). Three Malaise traps (Townes 1972) were set in each site, aligned and 25 meters apart from each other. The traps were left operating for 4-5 weeks; the content of each trap was collected every 7 days and treated as a sample. Samples from the dry season of 2006 could not be accessed and were not considered.
Total sampling effort was 4,095 trap-days, 1,827 in the DS and 2,268 in the RS. Sampling efficiency is expressed as the ratio between number of specimens and number of trapdays. The number of collected specimens and sampling efficiency for T. polita were compared to that of all other Cryptinae combined in order to contextualize its relative abundance and to compare the temporal dynamics of this particular species to the "background" of closely related species in the area. All specimens are deposited in the entomological collection of Universidade Federal do Espírito Santo (Brazil). The full dataset is available as supplementary material (Suppl. material 1).

Geographic coverage
Description: The study was conducted at Parque Estadual do Rio Doce, a 36,000 ha protected area in the Brazilian Atlantic Forest, state of Minas Gerais. Three sites within the park were sampled (

Column label Column description
Area Sampled area by successional state. LSF, low secondary forest; HSF, high secondary forest; PF, primary forest.

Dates-Start
Starting day for that sample, corresponding to the day when the Malaise trap sample corresponding was initially set.

Dates-End
Ending day for that sample, corresponding to the day when the Malaise trap sample was collected.

Results and Discussion
A total of 8,384 specimens of Cryptinae were collected, of which 6,657 (79.4%) belonged to T. polita, and 1,727 (20.6%) to all other Cryptinae combined ( Table 1). The overall sampling efficiency was 2.04. For T. polita alone, sampling efficiency was 1.62, and for all the other Cryptinae combined, 0.42. This quantitative superiority, however, was not uniform (Fig. 3). During the dry season (DS), T. polita represented only about 7% of the total Cryptinae collected. This situation was then drastically reversed in the rainy season (RS), when this ratio jumped to nearly 85%. The differences between dry and rainy seasons remain evident if the weekly totals are independently compared and are again clear if each week (Fig. 4)  For Cryptinae as a whole, sampling efficiency during the RS was 0.56, versus 0.25 in the DS, that is, 2.25× higher in the rainy season. The increase matches the observations of Kumagai (2002) for the Ichneumonidae in a preserved area in the Atlantic Forest, which tripled in number in the traps in the RS. This indicates that a two to threefold increase during the rainy season could be "normal" for Ichneumonidae as a whole. For T. polita, however, seasonal influence was higher by two orders of magnitude, with sampling efficiency rising from 0.02 in the DS to 2.92 in the RS -a 146× increase (Table 2).   The seasonal dominance of T. polita in the park was not a happenstance: it occurred repeatedly through the years (Fig. 4), as evidenced by its nearly uninterrupted numerical superiority in the samples (Table 3), and by its presence in 47.8% of the individual trap samples from the RS, against only 9.2% for the DS.  Table 2.
Comparison of sampling efficiency during the dry season (DS) vs. rainy season (RS) in the three sampled successional areas. RS/DS, how many times larger was the sampling efficiency in the rainy season in relation to the dry season. Sampling effort was identical for all areas, adding to 4,113 trap-days total, but different between seasons (DS=1,836, RS=2,277). %, percentage of specimens. There was a distinctive quantitative difference among the three successional areas for the occurrence of both T. polita and other Cryptinae (Table 1). All differences discussed next are statistically significant (see confidence interval bars in Fig. 3). For other Cryptinae, the primary forest (PF) area had the highest abundance (871 specimens, 50.4% of the total), closely followed by the tall secondary forest (TSF) (710; 41.1%) and low secondary forest (LSF) (146; 8.5%). For T. polita, however, the TSF supported a much higher number of specimens (4,691; 70.5%), followed by the PF (1,956; 29.4%). The LSF was clearly avoided by this species, with only 10 specimens collected there throughout the study (0.15%).
Parasitoid wasps in general are most abundant in the rainy season, as compared to dry periods (Gauld 1991 Kumagai and Graf 2000;Saaksjarvi et al. 2004) have never report any species as particularly dominating the collected samples. Considering that studies at the species level are scant for Ichneumonidae, the case for T. polita becomes of further interest, both because it is a rare study and because it signals that there is much still to be gained from studies on ichneumonid ecology.
The notorious success of T. polita in the tall secondary forest, as well as its distinct population dynamics in relation to other Cryptinae, suggest it may naturally act as a largescale parasitoid in intermediate successional stages. This might also reveal to be of more general interest, now or in the future: this type of forest is predicted to increase in the 21st century due to industrialization and urbanization (Thomlinson et al. 1996), and yet, "there is still a tremendous need to understand and further refine our knowledge of ecological processes involved in secondary succession" (Guariguata and Ostertag 2001).