Since the publication of the theory of island biogeography, ecologists have postulated that fragmentation of continuous habitat presents a prominent threat to species diversity. However, negative fragmentation effects may be artifacts; the result of species diversity declining with habitat loss, and habitat loss correlating positively with degree of fragmentation. In this study, we used butterfly assemblages on islands of Lake of the Woods, Ontario, Canada to decouple habitat fragmentation from habitat loss and test two competing hypotheses: (1) the island effect hypothesis, which predicts that decreasing fragment size and increasing fragment isolation reduces species diversity beyond the effects of habitat loss, and (2) the habitat amount hypothesis, which negates fragmentation effects and predicts that only total habitat area determines the diversity of species persisting on fragmented landscapes. Using eight independent size classes of islands (ranging from 0.1 to 8.0 ha) that varied in number of islands while holding total area constant, species diversity comparisons, species accumulation curves, and species-area relationship extrapolations demonstrated that smaller insular habitats contained at least as many butterfly species as continuous habitat. However, when highly mobile species occurring on islands without their larval food plants were excluded from analyses, island effects on potentially reproducing species became apparent. Similarily, generalized linear models suggested that effects of island isolation and vascular plant richness on insular butterfly richness were confounded by species of high mobility. We conclude that inter-fragment movements of highly mobile species may obscure important fragmentation effects on potentially reproducing populations, questioning support for the habitat amount hypothesis.

Aim: Investigate relationships between fragmentation and species diversity in the context of the theory of island biogeography, sample-area effect, and habitat diversity hypothesis.Location: Lake of the Woods, Canada. Taxon: Vascular plantsMethods: Vascular plant species diversity was inventoried on 30 islands, organized into two island sets. Each island set contained four size classes that varied in degree of fragmentation while controlling for the sample-area effect (small island set: 8 × 0.1-ha, 4 × 0.2-ha, 2 × 0.4-ha, and 1 × 0.8-ha islands; large island set: identical pattern utilizing 1.0-ha to 8.0-ha islands). Fragmentation effects were then examined using SLOSS-based analyses, addressing whether single large or several small islands contained more species/habitats: (a) direct comparisons of species and habitat richness across size classes; (b) extrapolations of species-area relationships; and (c) analyses of species and habitat accumulation curves. Multigroup path analysis was next used to quantify effects of habitat diversity, island area, and isolation on species richness for both island sets. Finally, pairwise and multiple-site dissimilarity was estimated for both species and habitats across 0.1ha and 1.0-ha islands to investigate whether: (a) variation in species composition was related to habitat composition; and (b) species dissimilarity increased with inter-island distance.Results: SLOSS-based analyses indicated that several small islands contained more species than single large islands in both island sets. This pattern was also observed for habitats, but only in the small islands set. Path analysis suggested that island area had significant direct and indirect (mediated by habitat diversity) effects on species richness. Habitat diversity and island isolation had significant positive and negative effects on species richness, respectively, independent of island area. Species and habitat dissimilarities were significantly related across 0.1-ha but not 1.0-ha islands, and showed no relationship to inter-island distance. Main conclusions:The overall positive relationship between fragmentation and species richness may be attributed to greater habitat diversity and increased species dissimilarity across smaller islands relative to larger islands. However, negative isolation effects indicate that landscape configuration is still an important conservation consideration. These results each align with different predictions of the theory of island biogeography, sample-area effect, and habitat diversity hypothesis, questioning the exclusivity of these theoretical frameworks. K E Y W O R D S habitat amount hypothesis, habitat diversity, habitat fragmentation, island biogeography, sample-area effect, SLOSS, small island effect, species-area relationship, structural equation model, vascular plants Author contributions: All authors collectively conceived of the study design; I.D.A. completed data collection and species identification with assistance from Z.G.M.; Z.G.M. completed the analysis and led the writing wi...