Hide menu
Figure 2. Species accumulation curves by traps (a) in a total of 30 traps, and by weeks (b) during 13 weeks, in gaps and understory. The estimated species richness in gaps and understory was calculated with the non-parametric estimator Chaos 2 (a). Each curve is presented with a 95% confidence interval.

A total of 1531 individuals distributed among 82 fruit-feeding Nymphalid species were captured during a 13 weeks study period (excluding 66 recaptures (4.3%) and 42 unidentified individuals (2.7%)). The number of individuals varied between 7 and 65 and the number of species between 5 and 22 for the gap traps and between 11 and 57 individuals and between 3 and 13 species for the understory sites. In total, 791 individuals were trapped in the gaps and 806 individuals in the understory. The species Panacea prola was n umerically dominant, representing 60.7% of the total number of individuals, both in gaps (52.3%) and understory (68.9%) habitat.

Figure 3. PCA ordination diagram of butterfly communities from traps in gaps (a), divided in small (< 300m2) and big gaps (> 300m2), and understory (b), linked by pair, and PCA ordination diagram of butterfly communities and environmental variables (c). Only the 22 species that contribute most to the model are shown. Eigenvalues of First Principal Component (PC1, x axis) and PC2 (y axis) are 17.5 and 10.7.

In total, 50 species were found in the understory and 71 in the gaps. Also the estimated richness (Chao 2 index) indicated a difference between gaps and understory and species richness was also higher (38%) in the gaps than in the understory (Fig.2a). The time-based accumulation curves showed a relatively steep increase with time, suggesting temporal differences in flight periods.


The PCA run to investigate the differences in butterfly assemblages between gaps and understory (Fig.3) revealed that there were overall larger differences between butterfly assemblages in gap traps than between understory traps. In many cases there were large differences even between the pair of traps close to each other in both gaps and understory. However, the butterfly assemblages from trap pairs from the understory control sites were in general more similar than the assemblages in the gap pairs (Fig 3a-b).

There was a clear difference in the butterfly species assemblages with gap and understory sites (Fig.3c). The majority of the species were associated with the gaps and with increasing light level. The species Adelpha jordani, Memphis glauce, Temenis laothoe, and Zaretis isidora showed the strongest association with gaps, showing a preference of higher amount of light, contrasting with Tigridia acesta and Nessea obrinius which were associated with the understory.

Table 1. pRDA scores of the log transformed data, with habitat as explanatory variable and pairs of traps as covariable. P-values were given by Monte-Carlo permutation tests with 9999 permutations (Trace=0.048, F-ratio=3.223, P-value=0.0001).

The pRDA, performed to investigate the degree of species’ association with gaps vs. understory, indicated that the species associated with the understory contained relatively few species, including Tigridia acesta, Catoblepia berecynthia and C. xanthus, and Nessea obrinius (Table 1). The “gap specialists” were more numerous with Pyrrhogyra crameri, Temenis laothoe, Archeoprepona demophon, Morpho helenor, Adelpha jordani, A. iphiclus, Memphis glauce and M. basiles as the most distinct ones (Table 1). We could also distinguish non-specialists, i.e. those with pRDA scores close to zero, e.g. Ps valentina.

Figure 4: Linear regression of chi2 distances between butterfly assemblages of Nymphaloidae in pairs of gap and understory trap and gap size. Gap size square root of m2 (y = 2.27 + 0.0451x; r = 0.3525, p = 0.0057; r2 = 0.1242).

The linear regression that investigates the dissimilarity (chi-squared difference) in butterfly assemblages, within pairs of gap/understory traps, showed a positive correlation with gap size (Fig 4). There were no significant differences between the understory traps (y=2,503 + 0,0258x; r = 0,2768; p = 0,3180; r2 = 0,0766).

Figure 5: PCA ordination diagram of butterfly communities (a) and environmental variables within 15 tree-fall gaps in Manu National Park, Peru. Only the 29 species that contribute most to the model are shown. Eigenvalues of PC1 (x axis) and PC2 (y axis) are 22.0 and 11.3

The second PCA, that considered only data from gaps, showed clear patterns of butterfly assemblages (Fig. 5). Many species seemed clearly associated with the environmental variables describing the vegetation structure of the gaps. The majority of the species were associated with the absence of vines, and the higher light levels in larger gaps (Fig. 5).

Only few species such as Taygetis angulosa, T. sosis, and T. laches, were associated with high vine density. Several species e.g. Nessea obrinius, was associated with small gaps, without low light, whereas Zaretis isidora, Adelpha jordani, and Memphis glauce, were associated with large gaps and higher amount of light.

Responsible for this page: Agneta Johansson
Last updated: 06/15/10