Evaluating the diversity of multiple taxa is fundamental to understand community assembly and to assess the integrity and functionality of tropical secondary forests. In this study, we analyzed the natural regeneration of tropical dry forests (TDFs) in three regions of Brazil using Hill-Simpson diversity, abundance and β-diversity of trees and five groups of insects (herbivores, fruit-feeding butterflies, ants, culicid mosquitoes and dung beetles). Sampling was conducted in 39 0.1 ha plots using a chronosequence approach (13 plots in early, intermediate and old-growth forests). We evaluated the contribution of three different levels to γ-diversity: α (within plots), β¹ (among plots) and β² (among successional stages), and further determined the relative importance of turnover (species replacement) and nestedness (differences in species number among sites) to β². Our results showed that, unexpectedly, the Hill-Simpson diversity was consistently higher in early than old-growth stages for all regions, but varied more widely in the intermediate stages. For each group separately, the same trend was observed for butterflies, ants, dung beetles and herbivores and did not differ among stages for mosquitoes. Successional differences in abundance were only detected for trees (increasing along the gradient) and for mosquitoes (decreasing). According to our expectations, the additive partitioning analysis showed that β2-diversity contributed more to γ-diversity than β1-diversity, when all taxa were considered together and for most of them separately (except for butterflies and dung beetles). Most of the β²-diversity was due to species turnover, but this contribution varied among groups and regions, with the highest turnover for herbivores and the lowest for dung beetles. Our results suggest that the Hill-Simpson diversity and changes in species composition (as given by β²-diversity) are better indicators of forest natural regeneration than raw species richness, corroborating previous studies with plants and animals.

Methods - Tree Sampling

From January to May 2012, 39 plots of 50 x 20 m were delimited (9 in Serra do Cipó, 15 in North MG and 15 in Patos), with 13 plots in each successional stage. Whenever possible, plots from different stages were interspersed, with a minimum distance of 200 m among plots (Figure S1). The floristic survey was carried out only once in the same period. In each plot, all living trees with diameter at breast height (DBH = 1.30 m from the ground) ≥ 5 cm were identified and measured. Height was visually estimated with the aid of a 2 m stick as a reference. These inventoried individuals were marked and numbered with aluminum tags and identified according to the Brazilian Flora project (Brazil Flora Group 2021). Voucher specimens were deposited in the Herbarium of the Federal University of Pernambuco (UFPE) and in the Herbarium Montes Claros (MCMG) at the State University of Montes Claros (UNIMONTES).

Methods - Fauna Sampling

Herbivorous insects

Herbivorous insects associated with the TDF canopy were sampled according to Macedo-Reis et al. (2019). We used Malaise/Window-type traps based on the model proposed by Basset (1988), in which the upper and lower parts contained bottles with 70% alcohol. One trap was uplifted in the center of each plot until reaching the canopy, where they remained for five days (120 hours) in each of the four samplings. The captured herbivores were classified into one of the following three guilds, according to their mouthparts: chewers, suckers and xylophages (Macedo-Reis et al. 2019). The insects were separated into morphospecies, identified to the lowest possible taxonomic level and deposited in the entomological collection of the Laboratory of Insect Ecology at the Federal University of Minas Gerais (UFMG).

Butterflies

Frugivorous butterflies were sampled according to Beirão et al. (2017). Four Van Someren-Rydon traps (VSR; Rydon 1964) were used in each plot. On the first day of sampling, the traps were placed with their base about 1 m above the ground and baited with fermented banana and sugarcane juice. Captured butterflies were removed daily and baits were replaced during three consecutive days. Captured individuals were marked with permanent ink when possible and information such as date, location and species identification were recorded (Beirão et al. 2017). When identification in the field was not possible, individuals were stored in envelopes with all the information for later identification in the laboratory, using field guides by DeVries (1987) and Uehara-Prado et al. (2004). The specimens were deposited in the reference collection of the Laboratory of Insect Ecology at the Federal University of Minas Gerais in Belo Horizonte, Minas Gerais, Brazil.

Ants

Ant sampling was carried out at five points in each plot (one at each corner and one at the center) according to Silva et al. (2017). Unbaited pitfalls were used with 200 ml of a solution composed of water and detergent. For epigeic ants, the traps were buried at ground level. To capture arboreal ants, pitfalls were installed at a height of 1.30 m in trees with circumference at breast height (CBH) ≥ 15 cm. Thus, 10 pitfall traps were placed in each plot. The traps remained in the field for 48 hours and the captured ants were taken to the laboratory for identification at the lowest possible taxonomic level according to Baccaro et al. (2015). The specimens were deposited in the reference collection of the Laboratory of Insect Ecology at the Federal University of Minas Gerais.

Culicid mosquitoes

Mosquitos from the Culicidae family were sampled using two types of trap. At dusk, a Shannon light trap was exposed for two hours and two CDC light traps were exposed for 12 hours in each plot. Additionally, active sampling was carried out for the capture of mosquitoes with daytime activity, in which a manual aspirator was used to collect all the mosquitoes that landed on the researcher before they started the blood meal. This type of sampling was performed once in each plot for a period of 45 minutes (Santos et al. 2015). The specimens were identified using dichotomous keys by Lane (1953), Faran (1980), Consoli and Oliveira (1994) and Forattini (2002) and deposited in the entomological collection of the Laboratory of Ecology and Biological Control of Insects at the State University of Montes Claros.

Dung beetles

Beetles from the Scarabaeinae subfamily were sampled using five pitfall traps baited with 50g of fresh human feces per plot (one at each and one at the center). The pitfalls were buried with the top opening leveled with the ground and contained 250 ml of a liquid mixture of water, detergent and salt to preserve the captured beetles. The bait was placed inside the pitfall and a cover was placed to prevent leaves and rainwater from entering. After 48 hours, the traps were removed and the captured scarab beetles were stored in alcohol and identification in the laboratory at the lowest possible taxonomic level according to Vaz-de-Mello et al. (2011). The specimens collected in Patos were deposited in the Entomological Collection of the Federal University of Pernambuco while those collected in North MG and Serra do Cipó were deposited in the Entomological Collection of the Federal University of Cuiabá, Mato Grosso, Brazil.

References