Seasonal and annual abundance of Ephuta wasp (Hymenoptera: Mutillidae) in Panama

Introduction: Ephuta Say is a widespread New World genus of mutillid wasp whose ecology is poorly understood. Objective: The objective of this study was to determine how Ephuta species abundance varies annually and seasonally driven by weather conditions and hosts. Methods: Ephuta specimens, located in the “old forest” at Barro Colorado Island, were examined from weekly samples (2001-2006) of ten Malaise traps. The monthly abundance of each Ephuta species was compared with monthly average humidity, solar radiation, temperature and rainfall. Results: Nine species and ten morphospecies were identified. Most specimens were collected from March to June. April was the month reported with the greatest abundance. Of the four abiotic variables measured, only average monthly temperature was correlated with Ephuta abundance. Months with the highest number of Ephuta specimens were correlated with suspected host abundance, as indicated by the monthly abundance of Pompilidae (Hymenoptera) captured during the year 2007. Conclusions: We concluded that Ephuta display strong seasonal variation in abundance, with the peaks occurring during the end of the dry season and beginning of the rainy season, which correlate broadly with temperature and the abundance of their pompilid hosts.


Study site:
The study site was the field station of the Smithsonian Tropical Research Institute (STRI) in Barro Colorado Island (BCI), located in Gatun Lake, Panama Canal, (9°09′17″ N & 79°50′53″ W). The island has a territorial expanse of 54 km 2 , an elevation approximately 137 m above sea level and is covered by lowland tropical forest (Cambra et al., 2018a). The climate is humid and warm, with an annual rainfall of 2 600 mm and average temperatures ranging from 25 to 30 °C (Corro, 2014).
Sampling: Mutillid and pompilid specimens were retrieved from the weekly samples of 10 Malaise traps (Townes, 1972). These were located along the western boundary of the 50-hectare "old forest" (> 400 years old) permanent plot on the central plateau of BCI and were placed at approximately 100 m intervals. Mutillidae specimens were sampled for six continuous years (2001)(2002)(2003)(2004)(2005)(2006), while Pompilidae specimens were only sampled during 2007. The mutillid specimens were identified to species (using the study of the types), or recognized as undescribed morphospecies, by Roberto Cambra. All material examined in this study is deposited in the Museo de Invertebrados G. B. Fairchild, University of Panama (MIUP).

Environmental variables:
We used meteorological data recorded annually for BCI through the bioinformatics website of the Tropical Research Institute Smithsonian (http:// biogeodb.Stri.si.edu/physical_monitoring/) and from Windsor (1990). The parameters compared were the monthly averages of humidity (%), solar radiation (MJ/m 2 /day), temperature (°C) and rainfall (mm).

Data processing and analyses:
Collection data (trap, month and year) for all specimens were recorded in Excel spreadsheets. A Kruskall-Wallis analysis was used to test for differences in monthly abundance. The seasonality of species abundance was tested with circular statistics by the Rayleigh test of uniformity. Shannon (H) and Simpson (1-D) diversity indices, Equitability (J), and Dominance (D) were calculated by month and sampling years. All environmental parameters were correlated with the abundance of individuals using Spearman's rank-order correlation. Finally, graphs of host (Pompilidae) abundance were constructed from data generated by Corro (2014). All statistical analyses were performed using the programs Oriana 4 (Kovach, 2011), Past 3 and R (R Core Team, 2015).

Composition of the Ephuta community:
Nine species (males) plus ten morphospecies (six males, four females) of Ephuta were recovered from the examined samples. During the sampling period (2001)(2002)(2003)(2004)(2005)(2006), species richness remained at 17 species, except for 2001 (16) and 2006 (13) ( Table 1). Species that were not collected in certain years had overall lower frequencies with fewer than ten individuals in total, those being: E. abadia (Cresson, 1902) (6), E. singularis (Spinola, 1841) (4) and four morphospecies (3-10 individuals). A significant difference in species richness was observed between sampling months (P < 0.05), varying from nine to 18 species. The months from January to May had the greatest species richness (14-18), while July to December (9-13) had the lowest (Table 2). We found evidence of seasonal in the distribution of species abundance   (Fig. 1A). The months with the greatest abundance were April (269 specimens) and May (243 specimens), comprising 44.7 % of all individuals in total. More than 100 individuals were collected in both March and June, with the remaining months contributing fewer than 75 specimens each (Fig. 1B). The most abundant species were E. triangularis (Cameron) (344 specimens or 30.0 %) and E. panama Schuster (108 or 9.4 %). Ephuta triangularis was the most abundant species across all the years of this study, as well as in each of the sampling months. The monthly diversity indices were similar: 1-D (0.8307-0.8949) and H (2.055-2.463); as well as the dominance and equitability: D (0.1051-0.1693) and J (0.7906-0.9065).

Seasonal variation in the abundance of
Ephuta, its relationship with abiotic factors and the presence of hosts: Species abundance did not vary significantly with monthly average humidity, solar radiation or rainfall (P > 0.05). There was, however, a positive relationship between average monthly temperature and the number of individuals collected (P < 0.05) in all species combined. Abundance was greatest in months with the highest average temperature (26 °C) (February, March, April, May and June) (Fig. 1B). Although average monthly rainfall did not correlate with Ephuta abundance, the highest abundances occurred in months recording less than 150 mm of rainfall. Monthly variation in Ephuta abundance throughout the study was similar to that observed in their putative Pompilidae hosts for the year 2007. The months of April, May and June stand out as the most abundant for both wasp families (Fig. 1C). Each of the 19 species of Ephuta studied have different abundances, but they present the same general pattern of abundance per sampling month (Fig. 1D). This last figure (Fig. 1D) shows that the seasonal occurrence of each Ephuta species follows a similar pattern of abundance and richness throughout the months of collection. In certain months, some species cannot be visualized in the graph because some species do not appear in certain months, some lines overlap and four species have consistently low abundance.

DISCUSSION
The March to June period of high Ephuta abundance coincided with the greatest species richness, which indicates the existence of a shared pattern of seasonality in all Ephuta species on BCI. This was especially pronounced in the most abundant species (Fig. 1B). Diversity indices suggest low differences between the species, with each displaying low dominance and homogeneous equitability. The similar effect of seasonal pattern on species richness and abundance suggests that these indices vary equitably among all Ephuta species on the study site.
The seasonal activity pattern of Ephuta resembles that of other Mutillidae, especially Dasymutilla Ashmead, which has a similar pattern over the same years of collection on BCI (Cambra et al., 2018a). Additionally, the seasonal activity of Pompilidae species on BCI (Corro, 2014) coincides closely with that found in Ephuta species, possibly reflecting the ecological link between putative host and parasitoid (Aranda & Graciolli, 2016;Cambra et al., 2017). It should be noted that in a single year of Pompilidae sampling, 910 specimens were collected. On the other hand, for Ephuta, the maximum collected was 281, a proportion largen than 3.2:1 for the ratio of Pompilidae to Ephuta specimens. Unfortunately, we do not have comparative data on specimens of Pompilidae and Ephuta from the same year. However, the collection methodology and locations are identical.
The synchrony visualized between Pompilidae-Ephuta (Fig. 1C) has been corroborated using a correlation analysis. Without taking into account that our correlation provided with high statistical significance, we do not include it here because the Ephuta data were collected from 5 different years than the Pompilidae records. Since there is not a valid comparison available, presenting a correlation analysis with statistical significance would be misleading. Our sole goal with this data was to visualize their relationship as a preliminary hypothesis that would be worth analyzing at a later date with additional data.
Similar to the results of Corro (2014) for Pompilidae, we found that Ephuta abundance was significantly correlated with average monthly temperature. The latter had no apparent relation with average monthly rainfall, and no significant correlation was observed between average monthly rainfall and average monthly temperature over the six years of our study. When these last two factors were correlated over a longer period , however, a significant negative relationship was observed between average monthly rainfall and temperature. This suggests that rainfall may lower temperatures and thus indirectly plays a role in the seasonal activity of these wasps. When calculating the average monthly rainfall over 20 years , the dry season in BCI presents an average monthly rainfall of 200 mm. This could explain why, in this study, no relationship was observed between the average monthly rainfall and Ephuta abundance since the monthly rainfall average in the whole study period (2001)(2002)(2003)(2004)(2005)(2006) was 200 mm, except for November. Wolda (1988) argued there existed a variety of seasonal patterns in insect species of tropical forests and thought that these patterns were primarily related to temperature. Our results support this hypothesis, since greater abundance of Ephuta occurs when temperatures are higher than normal, and conversely, lower abundances occur when temperatures are lower than normal. Wolda (1989) also mentioned that activity peaks in seasonal insects are related to the start and end of the dry season in BCI. In this respect, our results are only partly congruous with Wolda's (1989) hypothesis because the last month of the dry season, April, showed the highest abundance of Ephuta on BCI. No similar peak, however, was found at the beginning of the dry season. Studies with other mutillid genera and in other habitats would be useful for further comparing these hypotheses. Thereby, our data shows that even though the sampling of Pompilidae and Ephuta on BCI was not conducted throughout the same period, their seasonal abundance appears to be synchronous which further reinforces their close host/parasite ecological relationship. Additionally, we have found that abundance for both groups seems to vary according with the average monthly temperature which in turn is a likely result of average monthly rainfall.
Ethical statement: authors declare that they all agree with this publication and made significant contributions; that there is no conflict of interest of any kind; and that we followed all pertinent ethical and legal procedures and requirements. All financial sources are fully and clearly stated in the acknowledgements section. A signed document has been filed in the journal archives.

ACKNOWLEDGMENTS
We thank Dumas Gálvez and the Central American Master's Program in Entomology at the University of Panama for the comments that helped improve this manuscript. The longterm malaise trap sampling on BCI was instituted by John Pickering (University of Georgia, USA) and was supported by funding from the "Environmental Sciences Program" of the Smithsonian Institution.