The butterfly community of a managed West African rainforest: patterns of habitat specificity, diversity, stratification and movement.

Fruit trapping, standardised transect walks and point distance sampling were used to study the butterfly fauna of a managed, over-logged rainforest in south-eastern Côte d"Ivoire, West Africa. The study site within the 22,000 ha large Bossematié Forest Reserve included three different forest plots: a 15 ha tree plantation, a 30 ha thinned forest and a 30 ha regenerating control forest. The impact of the local forest management on the butterfly community was tested. The main emphasis was put on the fruit-feeding Nymphalidae, trapped with a standardised, fermenting fruit bait. 263 butterfly species are reported for the over-logged Bossematié Forest Reserve, and the estimated total species richness amounts to c. 500. The Bossematié"s butterfly community still mainly consists of true rainforest species, with only a small fraction of savannah species, and is very similar to that of the less disturbed, but equally logged Bia National Park. 28 Upper Guinean Endemics were found, as well as one nymphalid species (Bebearia dallastai) endemic to the Bossematié Region. Four butterfly species, characteristic of mature forests and abundant in the nearby Bia Reserve, were not found in the Bossematié Forest. Daily nymphalid capture rates were high compared to those known from other studies, however, the abundance of several range-restricted hostplant specialists feeding on understorey shrubs and climbers and common in other forest locations, was found to be relatively low. This implicates the (ongoing) local decline and potential loss in specialised forest species due to disturbance. The (nymphalid) species richness of a 3 km distant, 6 ha small forest remnant was only half of that found in a similar sized plot within the Bossematié"s Reserve: in this small isolated fragment, the grass-feeding species of the Satyrinae subfamily were dominating, several Nymphalinae species using understorey shrubs as hostplants, as well as Charaxinae species, were lacking. A total of 8,400 nymphalid individuals, representing 123 species, were baited during 996 trap days in a dry and two wet seasons on the 75 ha studied. There were pronounced variations in nymphalid abundance and diversity between years and seasons, with a shift in dominance from the Satyrinae in the dry to the Nymphalinae in both wet seasons. Alpha diversity was highest during wet season 1997, slightly lower during dry season 1996, and lowest during wet season 1996. Accumulated species richness seemed highest in, and comparable between, the dry season 1996 and the wet season 1997. Species similarity was highest between both wet seasons and lowest between different seasons in different years. Variation in nymphalid abundance and diversity was equally related to micro-scale variation in vegetation structure at each trap location (i.e. the four succession stages). Accumulated species richness and overall abundance seemed to decrease with habitat maturity, and differences were more pronounced during the dry season. In 33 nymphalid species, there was a significant relationship between forest maturity and capture frequency. Canopy flyers (especially among the Charaxinae), as well as understorey nymphaline and satyrine species with a wider geographical range found in various forest types (from wet evergreen to dry forests) were more abundant in gaps and early successional forest patches. In contrast, those nymphaline understorey species, centred on the wetter or moist forest zones and with a more restricted distribution, were more abundant in canopy-closed forest patches. In a mark-recapture study during 396 trap days in the dry season, 1,430 resp. 540 nymphalid individuals were marked resp. recaptured in total. In most species, females seemed to fly further than males. Average movement values ranged between 0 m and 433 m. Movement parameters were generally largest in species feeding on early successional larval hostplants characterteristic of gap and margin habitats and smaller in species characteristic of the darker forest understorey, feeding on climbers and smaller shrubs. Vertical stratification of the nymphalid community on the 75 ha study site, studied during 204 trap days of the wet season 1997, was pronounced at different taxonomic levels. Species richness and abundance decreased with increasing trap height. A defined understorey community could be separated from a distinct midstorey / canopy fruit-feeding butterfly assemblage. The increase in percentage of rare nymphalid species towards the canopy indicates that resources (rotting fruits) are more limited at higher strata than towards the forest floor. In several species, stratification patterns seemed to correlate with information on larval host-plant height or oviposition behaviour and sensitivity to light regime changes. Accumulated nymphalid species richness and diversity indices were lower in the control compared to the thinned forest, but lowest in the plantation. However, there was a significant, negative correlation between the species" preference for the control forest and the size of the species geographical range. Six out of the nine species with a lower capture frequency in the thinned forest showed preferences for mature succession stages and most of these were either Upper Guinean endemics or Guinea-Congolian-restricted species. The five species with higher capture frequencies in the thinned plot were all geographically more widespread and showed higher frequencies in younger forest succession stages. Ten species, mainly canopy specialists, ubiquitous species or species centred on the Guinea savannah / dry forests, were most abundant in understorey traps of the plantation. Seventeen more nymphalid species avoided the plantation forest. Thinning seems to affect hostplant and wet/moist forest specialists with smaller geographic ranges, thus risking loss of regional diversity. Patterns of habitat specificity and or diversity were largely consistent between sampling methods. Walk-and-capture revealed additional information in species which were not or less frequently baited but, as shown with point distance sampling data, differences in detectability between different forest vegetation structure might account for differences in encounter probabilities. Movement and stratification data confirm the (negative) impact of thinning on restricted range species of the forest understorey, and clarify trends found in higher strata species. In all species, movement parameters were largest in the plantation. Species feeding on early successional larval host-plants characteristic of gap and margin habitats showed shorter flights, whereas species feeding on climbers and smaller shrubs of the forest understorey showed larger movement in the thinned compared to the control forest. Higher strata species (e.g. within the Charaxinae) equally showed smaller movement in the control plot. Including samples taken at higher vertical strata, they also showed higher overall abundance in the control forest, although understorey sampling alone indicated the contrary. A comparison of nymphalid vertical distribution patterns between the thinned and the control forest indicated a clear disruption of stratification after thinning. Canopy species seemed to stay in the upper strata of the less disturbed control forest, and are more likely to fly down to the forest floor in the more open thinned forest. Their higher abundance in understorey traps of the early succession stages, and the thinned forest site, illustrates a sampling bias rather than a true habitat preference. Thinning might therefore not only affect the range-restricted understorey butterfly fauna, but also several more widespread species belonging to the forest canopy community.