Assessing Habitat Fragmentation Effects on Primates: The Importance of Evaluating Questions at the Correct Scale

and/or keywords. This search revealed a total of 227 papers. We randomly selected 100 of these papers and identifi ed the ways in which habitat fragmentation effects are being assessed. Evaluating and Measuring Habitat Fragmentation Habitat may be broadly defi ned as “the range of environments suitable for a given species” (Hall et al. 1997 ). For primates this generally refers to broad vegetation types, such as tropical rain forest and tropical dry forest (Arroyo-Rodríguez and Mandujano 2009 ). Because native vegetation is important for many species, numerous researchers have equated “habitat” with native vegetation (Fischer and Lindenmayer 2007 ; Arroyo-Rodríguez and Mandujano 2009 ). Habitat fragmentation is a landscape-scale process in which continuous habitat is broken apart into smaller pieces (fragments) scattered within a matrix of nonhabitat. This implies the loss of habitat and its subdivision (fragmentation) into a variable number of fragments (Fahrig 1999 ; McGarigal and Cushman 2002 ; Fahrig 2003 ). However, habitat loss can occur without the subdivision of habitat (Fig. 2.2 ), and therefore, we advocate that it will be valuable for researchers to consider analyzing the independent effects of habitat loss and fragmentation to determine whether it is the overall loss of habitat or the separation of habitat into smaller pieces (hereafter termed “habitat fragmentation per se”; sensu Fahrig 1999 , 2003 ) that actually causes negative effects on primates. This can only be done through landscape-scale studies, that is, by using landscapes as the independent units of observation (McGarigal and Cushman 2002 ; Fahrig 2003 ; Arroyo-Rodríguez and Mandujano 2009 ). By using fragments as the unit of analysis (hereafter named “fragment-scale studies”), researchers cannot differentiate between the effects of the habitat loss and the breaking apart of habitat, as both processes can result in smaller and more isolated fragments (Fahrig 2003 ; Fig. 2.2 ). Most fragmentation measures (e.g., mean fragment isolation, total amount of edge, number of fragments) are strongly related in a nonlinear manner to the amount of habitat within a landscape, in such a way that below a certain threshold of habitat area, small changes in the extent of the habitat lead to big changes in these measures (Neel et al. 2004 ). For this reason, it is often diffi cult to determine the separate effects of habitat loss and fragmentation. For instance, studies with plants (ArroyoRodríguez et al. 2009 ) and animals (Andrén 1994 ; Pardini et al. 2010 ) suggest that species diversity in a fragment of a given size may vary in landscapes with different habitat amount. Actually, the effects of fragmentation per se are thought to be relatively more important below certain thresholds of habitat amount remaining in the landscape (Andrén 1994 ; Fahrig 1997 , 1998 ; With and King 2001 ). Below this threshold of habitat amount, the probability of persistence of populations drops signifi cantly. Given the crucial management implications that these thresholds have for primate conservation, we urgently need to analyze the response of primates under different scenarios of habitat loss and fragmentation. This cannot be done through fragment-scale studies; it requires studies at the landscape scale. 2 Assessing Habitat Fragmentation Effects on Primates...