Biogeographical and floristic predictors of the presence and abundance of mantled howlers (Alouatta palliata mexicana) in rainforest fragments at Los Tuxtlas Mexico.код для вставкиСкачать
American Journal of Primatology 67:209–222 (2005) RESEARCH ARTICLE Biogeographical and Floristic Predictors of the Presence and Abundance of Mantled Howlers (Alouatta palliata mexicana) in Rainforest Fragments at Los Tuxtlas, Mexico JURGI CRISTÓBAL-AZKARATE1*, JOAQUIM J. VEÀ1, NORBERTO ASENSIO1, and ERNESTO RODRÍGUEZ-LUNA2 1 Primats, Centre Especial de Recerca en Primats, Universitat de Barcelona, Barcelona, Spain 2 Parque de la Flora y Fauna Silvestre Tropical, Instituto de Neuroetologı´a, Universidad Veracruzana, Veracruz, México This research focuses on identifying the principal habitat characteristics that influence the presence and abundance of mantled howlers in forest fragments. We provide information on the demography of several fragmented Alouatta palliata mexicana subpopulations at Los Tuxtlas, Mexico, and relate this to the biogeographical and floristic characteristics of the forest fragments inhabited. The most important habitat characteristics related to the presence and abundance of howlers in the fragments were fragment size and floristic diversity. On the other hand, some evidence suggests that given the conditions under which howlers in our study area live (i.e., small and degraded fragments with high densities), secondary vegetation may be beneficial for the survival of the howlers. Finally, we discuss the possibility that the very low immature-tofemale ratio (IFR) in the groups, and the lack of juveniles found in many of the study groups may be due to high mortality rates in immatures. A reduction in food availability because of the high population densities of these groups may be responsible for this process. Am. J. Primatol. 67:209–222, 2005. r 2005 Wiley-Liss, Inc. Key words: Alouatta palliata; demographics; biogeographic characteristics; floristic characteristics; immature-to-female ratio; Los Tuxtlas Contract grant sponsor: Zientzi Politikarako Zuzendaritza, Basque Government; Contract grant sponsor: Ministerio de Ciencia y Tecnologı́a, Spain; Contract grant numbers: PB98-127; BS0200203340. *Correspondence to: Jurgi Cristóbal-Azkarate, Apartado Postal 199, C.P. 91001, Xalapa, Veracruz, México. E-mail: firstname.lastname@example.org Received 31 May 2004; revised 12 February 2005; revision accepted 10 March 2005 DOI 10.1002/ajp.20178 Published online in Wiley InterScience (www.interscience.wiley.com). r 2005 Wiley-Liss, Inc. 210 / Cristóbal-Azkarate et al. INTRODUCTION The Mexican mantled howler (Alouatta palliata mexicana), whose original distribution was from southeast Mexico (excluding the Yucatan Peninsula) to southern Guatemala, is at present restricted throughout this range to several isolated reserves and nonprotected areas, all of which are greatly fragmented [Rodrı́guez-Luna et al., 1996]. Hunting and the extensive deforestation of this region during the last 60 years [Collins, 1990; FAO, 2003] are responsible for the decline and isolation of A. p. mexicana populations [Rodrı́guez-Luna et al., 1996], and this subspecies is now classified as critically endangered [Cuarón et al., 2003]. Although the literature pertaining to the mantled howlers’ capacity to live in fragments is relatively abundant [Bicca-Marques, 2003; Clarke et al., 2002; Juan et al., 2000; Lovejoy et al., 1986; Schwarzkoft & Rylands, 1989], only a few studies have focused on the habitat characteristics (floristic and biogeographic) related to the presence and demography of A. palliata in a fragmented landscape, and contradictory results have been reported [DeGama-Blanchet & Fedigan, in press; Estrada & Coates-Estrada, 1996; Rodrı́guez-Toledo et al., 2003; Sorensen & Fedigan, 2000]. Such information is necessary for us to better understand how howler monkeys respond to habitat fragmentation, and to formulate effective management plans for the conservation of this species. According to the island biogeographical theory (IBT) [McArthur & Wilson, 1967], the number of species and individuals will diminish with the reduction of the area, and the degree of isolation of the fragments. Estrada and Coates-Estrada  found that the number of individuals of A. palliata inhabiting forest fragments in one area of the Los Tuxtlas region in Mexico was positively related to fragment size, and negatively to the isolation distance of the forest fragments. In contrast, Rodrı́guez-Toledo et al.  did not find any relationship between fragment size and number of individuals in the fragments in another area in the southern part of the same region. Furthermore, in a study in Costa Rica, DeGama-Blanchet and Fedigan [in press] found no relationship among forest fragment size, degree of isolation, and mantled howler density, which suggests that the large size of forest fragments in the study area (up to 95 km2) were responsible for the lack of relationship between the variables. Finally, in Costa Rica, Clarke et al.  observed that the construction of a canal that fragmented and reduced the habitat did not reduce the population size of A. palliata, although there was a reduction in group size. In general, the transformation of the original vegetation resulting from forest fragmentation has a negative impact on the number of mantled howlers. In a certain area of the Los Tuxtlas region, Estrada and Coates-Estrada  found a positive relationship between floristic diversity and the number of howlers in fragments, and associated the higher species diversity with large and undisturbed fragments. In another area of the same region, Rodrı́guez-Toledo et al.  found larger group sizes in undisturbed fragments. In Costa Rica, Sorensen and Fedigan  and DeGama-Blanchet and Fedigan [in press] found higher densities of A. palliata in older tropical dry forest than in younger regenerating forests. To enhance our understanding of how mantled howler monkeys respond to habitat fragmentation, we provide information in this paper on the demography of several fragmented A. p. mexicana subpopulations in the region of Los Tuxtlas, Mexico, and relate this to the biogeographical and floristic characteristics of the inhabited fragments. In addition, we discuss which habitat variables are most likely related to the presence and size of howler populations in the fragments. Howler Demography in Rainforest Fragment / 211 Finally, we analyze the demographic trends of these populations and discuss the possible causes of these trends. MATERIALS AND METHODS Study Site The study area (2,064,129 m, 2,049,182 m north and 272,772 m, 285,783 m east, zone 15, Universal Transform Mercator; elevation 0–400 m above sea level) is located in the northern part of the Los Tuxtlas Biosphere Reserve in southern Veracruz, Mexico. The vegetation type is known as high evergreen rainforest [Gómez-Pompa, 1973] with different degrees of transformation. The original habitat of this region, as well as many others in Central and South America, has been extensively transformed into pasture and agricultural landscapes [Dirzo & Garcı́a, 1992]. In our study area only 13% of the original 75,000 ha of rainforest remains today in fragments of different sizes. For a further description of the area, see Estrada and Coates-Estrada . Data Collection Census of howlers. We censused howlers from March to September 2000, and from March to December 2001, in forest fragments surrounding the San Martı́n Tuxtla Volcano in the northeastern portion of the Los Tuxtlas Biosphere Reserve. The censusing party consisted of Cristóbal-Azkarate and one trained assistant. Our census method involved the observers positioning themselves for several days (depending on the size of the fragment) in strategic spots (inside or outside the fragments) before sunrise, waiting for the monkeys to howl. This allowed us to determine the number and location of the groups in the fragments. In addition, we interviewed the local people who lived and worked in the proximity of the forest fragments about the presence or absence of howlers in the fragments. If after several days no howlers were heard in a fragment, and the local people reported that the fragment was not occupied by howlers, it was considered empty. The fragments were revisited on a opportunistic basis several times during the study period to confirm howler occupancy. Once we located the howlers, we recorded the number of adult males, adult females, juveniles, and infants following Carpenter’s  age classification, and revisited the group if doubt existed after the first day. To avoid data repetition by censusing the same individuals more than once, we identified group members by drawing the distinguishing color patterns on their feet and tails that are characteristic of this subspecies, and noting any other distinctive features. Since there is evidence that howlers in Los Tuxtlas may sometimes disperse between fragments [Estrada & Coates-Estrada, 1996], we considered the sum of all individuals living in a fragment as a subpopulation. We calculated several demographic indexes (Table I). The subpopulation size, population density, and number of groups were calculated for each subpopulation. The group size, sex ratio, juvenile-to-female ratio (JFR), infant-to-female ratio (IFFR), and immature-to-female ratio (IFR) were calculated for each group, and the mean values were calculated for each subpopulation. Biogeography of the forest fragments. We calculated the fragment size, the shortest distance to the nearest human settlement, and two isolation variables, the shortest distance to the nearest 212 / Cristóbal-Azkarate et al. TABLE I. Calculated Variables Demographic indexes Subopulation size: no. of individuals in each fragmenta Population density: no. of individuals per hectarea Number of groups: no. groups in the fragmenta Group size: no. individuals per groupb IFR: no. immatures (juveniles + infants) to no. adult female ratiob JRF: no. juveniles to no. adult female ratiob IFFR: no. infants to no. adult female ratiob Sex ratio: no. adult females to no. adult male ratiob Biogeographical variables Fragment size: (ha) SN: shortest distance to the nearest fragment (m) SNM: shortest distance to the nearest fragment with monkeys (m) DH: distance to the nearest human settlement (m) Floristic variables Tree species or family diversity: total no. of tree species or families Relative tree species or families diversity: no. of tree species or families to total no. of trees ratio Shannon index: average degree of uncertainty-complexity (N–4) Mean DBH: mean diameter at breast high (cm) Tree density: no. of trees per hectare; S5: the sum of the importance values of the five most consumed taxa by howler monkey at Los Tuxtlas according to Estrada  (0–300). a b These variables are calculated for each fragmented population. These variables are calculated for each group. fragment (SN) and the shortest distance to the nearest fragment with monkeys (SNM) (Table I) by analyzing 1999 aerial photographs of the study area (1:20,000 scale) with ArcView GIS software (version 3.1) and the Patch Analyst 2.2 extension [Rempel & Elkie, 1999]. The analysis of aerial photographs from previous years demonstrated that by 1976 the deforestation in our study area was extensive, but that the actual fragmentation of the forest took place between 1976 and 1986. Considering that all forest fragments studied were approximately the same age, this variable was not considered in the analysis. Census of the vegetation. For each fragment we censused a total area of 1,000 m2, distributed in 10 randomly located 50 2 m transects [Keel et al., 1993]. We registered all trees with a diameter at breast height (DBH) Z10 cm, noting the species, family, DBH, and position in the transect. We calculated several diversity indexes and floristic variables (Table I). One of the indexes we calculated to assess the floristic diversity of the forest fragments was the Shannon index. This index is used by ecologists to describe the average degree of uncertainty of predicting the species of an individual picked at random from the community. The uncertainty of occurrence increases both as the number of species increases and as the individuals are distributed more evenly among the species already present. The values ranged from 0 (indicating low community complexity) on up. The Shannon index was calculated with the use of a base e logarithm. Howler Demography in Rainforest Fragment / 213 According to Estrada , five taxa (Ficus spp., Nectandra ambigens, Poulsenia armata, Brosimum alicastrum, and Cecropia obtusifolia) account for 80% of the total feeding time of one troop of A. palliata living in continuous forest at Los Tuxtlas. The importance value refers to the relative contribution of a plant species to the entire community, as it incorporates into a single measure the relative density, the relative frequency, and the basal area of each species. Therefore, the sum of the importance values of the five most consumed taxa by howler monkeys at Los Tuxtlas (S5) is an approximation of the availability of preferred food species in the fragment, and its value ranges between 0 and 300. Data Analysis Statistics. We used six different statistical techniques for this study: 1) the Mann-Whitney U-test to determine which variables were related to the presence or absence of howler populations in the fragments; 2) a series of discriminant analyses to determine which group of variables would better predict the presence or absence of howler populations in a fragment; 3) stepwise multiple regressions to determine which group of variables would better predict the abundance of howlers in the fragments; 4) the Spearman-R correlation coefficient and 5) simple regressions to determine the relationship between the floristic, biogeographic, and demographic variables; and 6) Student’s t-test to determine whether the demographic characteristics of our study population were statistically different from those reported by Estrada and Coates-Estrada  for another location in Los Tuxtlas, and those reported by Chapman and Balcomb  for this species. The accepted significance level for all of the analyses was r.05. RESULTS Variables Related to the Presence of Howler Populations We identified 55 forest fragments within our study area (7,500 ha). Of these fragments, 21 were inhabited by at least one group of howlers (i.e., more than one male or one female), and 15 were censused intensively (Table II). Two small fragments were temporarily occupied by transient solitary males, but were considered ‘‘empty’’ for the analysis. We studied the biogeography of all 55 fragments and the floristic characteristics of 31 fragments (18 occupied by howlers (15 censused intensively) and 13 empty). The mean size of the inhabited fragments was significantly larger than that of the empty fragments (Table III), but we did not find any significant relationship between the degree of isolation (SN and SNM) and the presence or absence of howlers in the fragments. We did not find any significant relationship between the presence or absence of howler populations in the fragments and their proximity to human settlement. Compared with the empty forest fragments, the inhabited fragments exhibited a higher tree species and family diversity, Shannon index, and tree density (Table III). These four variables are strongly correlated with each other (Table IV). The group of variables that best predicted the presence or absence of howler populations in a fragment were the fragment size and Shannon index, with a 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Fragment 5 10 9 5 1 3 1 1 2 1 1 1 1 1 1 No. groups 5.4 10.5 6.44 8.4 12 4.3 5 4 7 6 2 5 6 3 3 Mean group size 0.12 0.71 0.58 0.67 1.66 2.46 0.54 0.55 0.43 0.34 0.29 1.38 2.85 4.35 3.07 Population density 29 108 63 43 12 13 5 4 14 6 2 5 6 3 3 No. individuals 15 32 27 17 4 3 2 2 5 2 1 2 1 1 1 No. males 10 48 24 18 7 9 2 2 6 2 1 3 4 1 1 No. females TABLE II. Demographic Characteristics of the Censused Howler Populations 1 15 5 2 0 0 0 0 3 2 0 0 0 0 0 No. juveniles 3 13 7 6 1 1 1 0 0 0 0 0 1 1 1 No. infants 4 28 12 8 1 1 1 0 3 2 0 0 1 1 1 No. immatures 0.67 1.50 0.89 1.06 1.75 3.00 1.00 1.00 1.20 1.00 1.00 1.50 4.00 1.00 1.00 Sex ratio 0.10 0.31 0.21 0.11 0 0 0 0 0.50 1.00 0 0 0 0 0 JFR 0.30 0.27 0.29 0.33 0.14 0.11 0.50 0 0 0 0 0 0.25 1.00 1.00 IFFR 0.40 0.58 0.50 0.44 0.14 0.11 0.50 0 0.50 1.00 0 0 0.25 1.00 1.00 IFR 214 / Cristóbal-Azkarate et al. Howler Demography in Rainforest Fragment / 215 TABLE III. Comparison Between the Habitat Characteristic of the Fragments Inhabited and Not inhabited by Howler Monkeys (Mann-Whitney U test) Inhabited Mean SD Fragment size 38.57 61.14 Tree species diversity 36.26 13.24 Shannon index 3.25 0.22 Tree density 632.51 223.66 Tree family diversity 22.95 5.07 Not inhabited Valid n Mean 21 19 19 19 19 SD Valid n U Z P 34 13 13 13 13 152.00 59.00 69.50 62.50 52.00 3.55 2.48 2.07 2.34 2.76 0.000 0.013 0.038 0.019 0.006 5.63 9.65 26 8.09 2.9 0.57 454.61 93.69 17.46 4.37 TABLE IV. Relationship Between Vegetation Variables Tree species diversity Tree species diversity Shannon index Tree density Tree families diversity Shannon index Tree density Tree families diversity rs p rs p rs p rs p – 0.95 0.72 0.87 – 0.00 0.00 0.00 0.95 – 0.62 0.83 0.00 – 0.00 0.00 0.72 0.62 – 0.57 0.00 0.00 – 0.00 0.87 0.83 0.57 – 0.00 0.00 0.00 – predictive value of 75% of the expected over the observed cases (l = 0.71, F (2, 29), Po0.025). Demographic Characteristics of the Fragmented Populations and Related Biogeographical and Floristic Variables We censused 316 howlers (104 resident adult males, 11 adult solitary males, 138 resident adult females, 28 juveniles, and 35 infants) in 43 groups living in 15 different forest fragments. The mean subpopulation size was 21.06 individuals (SD = 29.68, range = 2–108, n = 15), the mean number of groups in a fragment was 2.86 (SD = 3.04, range=1–10, n = 15), the mean group size was 7.09 (SD = 4.22, range = 2–18, n = 43), and the mean sex ratio was 1.47 (SD = 1.1, range = 0.5–6, n = 43). The mean IFR was 0.45 (SD = 0.43, range = 0–2, n = 43), the mean JFR was 0.18 (SD = 0.30, range = 0–1, n = 43), and the mean IFFR was 0.32 (SD = 0.41, range = 0–2, n = 43). Juveniles were present in only 40% of the populations and 34.8% of the groups, while 66.6% of the populations and 55.8% of the groups had infants. Population density correlated in a negative and significant manner with fragment size (rs = –0.69, P = 0.004, n = 15), and nonsignificantly with population size. None of the populations with a density higher than 0.71 had juveniles. Fragment size (R2 = 0.45, P = 0.006, n = 15) and tree density (R2 = 0.27, P = 0.045, n = 15) were significantly related to the number of howlers in the fragments, but the best set of habitat variables that explained the number of 216 / Cristóbal-Azkarate et al. howlers per fragment was the fragment size and the Shannon diversity index (r2 = 0.675, P = 0.001, n = 15). DISCUSSION Comparison With Previous Studies at Los Tuxtlas In a previous study, Estrada and Coates-Estrada  conducted a broad census of howler populations in 126 forest fragments in a different portion of Los Tuxtlas Biosphere Reserve. Our study area was located at the northern part of the reserve, and replicated only three of the fragments censused by the authors (Coates, personal communication). Compared to the results of our study, Estrada and Coates-Estrada  found much lower howler population densities (Table V) and population sizes (r15 individuals). It is interesting to note that 32% of their fragments were larger than 100 ha. Several factors may be related to these differences, as discussed below. The abundance of howlers is not uniform throughout the Los Tuxtlas Biosphere Reserve, as shown by the small population densities and population sizes observed in the southern part of the Reserve by Rodrı́guez-Toledo et al. . On the other hand, since the area was decreed a Biosphere Reserve in 1998, hunting and illegal trafficking of howlers has practically disappeared. This is consistent with our data demonstrating no relationship between the presence and abundance of howlers in the fragments and their proximity to human settlements. Interspecific competition could also account for these differences, since in contrast to some of the fragments studied by Estrada and Coates-Estrada , none of the howler populations censused by us shared their habitat with spider monkeys (Ateles geoffroyi). Finally, the differences between their study and ours could be due to the different censusing techniques used in each case. Whereas Estrada and Coates-Estrada  censused howlers by walking TABLE V. Comparison of the Demographic Characteristics of the Howler Populations Censused at Our Study area with the Data Presented for the Species by Estrada & CoatesEstrada  at Los Tuxtlas, and Chapman & Balcomb  at Different Sites Mean Population density (ind./ha) Our study at Los Tuxtlas Estrada & Coates-Estrada  Chapman & Balcomb a Group size Our study at Los Tuxtlas Estrada & Coates-Estrada  Chapman & Balcomb  Group immatures/female Our study at Los Tuxtlas Estrada & Coates-Estrada  Chapman & Balcomb  Group sex ratio Our study at Los Tuxtlas Estrada & Coates-Estrada  Chapman & Balcomb  a at Los Tuxtlas SD n Student t 1.33 1.28 15 0.067 0.11 60 0.52 .309 18 P d.f. 7.7 2.54 0.000 0.000 73 51 at Los Tuxtlas 7.09 7.03 14.62 4.22 3.34 4.51 43 60 39 0.08 7.81 0.9 101 0.000 80 at Los Tuxtlas 0.45 0.53 1.25 0.43 0.36 0.75 43 60 38 0.99 7.4 0.32 101 0.000 79 at Los Tuxtlas 1.47 1.85 2.34 1.11 0.74 0.63 43 60 38 209 4.22 0.039 101 0.000 79 Not considering Baldwin & Baldwin . Howler Demography in Rainforest Fragment / 217 through the fragments (40–2,500 m), we located the groups by their morning howls and spent several days in each fragment, which enabled us to detect more troops. This could explain why they did not detect more than one group per fragment. The mean group size and IFR of the groups do not differ significantly between the two studies; however, the mean sex ratio of the groups was larger in Estrada and Coates-Estrada’s  study (Table V). Variables Associated With the Presence and Abundance of Howlers in the Forest Fragments Biogeographical variables. Our results partially support the island biogeographical theory (IBT) [McArthur & Wilson, 1967]. We found a strong positive relationship between fragment size and the presence and abundance of howlers in the fragments, but no significant relationship was observed regarding the degree of isolation. This lack of relationship may be due to the general small degree of isolation found in the fragments in our study area (50% of the fragments are less than 100 m apart). Estrada and Coates-Estrada  observed howlers traveling distances of up to 200 m on the ground, although they emphasized that this behavior is very infrequent. Therefore, the short distances between fragments in our study area may facilitate the occasional movement of individuals, and may diminish the effects of isolation over howler populations. Floristic characteristics. Fragments that were occupied by howlers presented higher floristic diversity (i.e., total number of tree species and families, and Shannon index) compared to the empty fragments. Of these different floristic variables, the Shannon index in conjunction with fragment size was the variable that best predicted the presence and abundance of howlers in the fragments. Howlers prefer seasonal foods (young leaves, fruits, and flowers) to perennial foods [Juan et al., 2000; Milton, 1980; Rodrı́guez-Luna et al., 2003], since such foods are generally of a higher nutritional quality [Milton, 1980]. To meet their nutritional requirements, howlers feed from many different species. Most of these different species are used as leaf sources, both overall and daily. According to Milton , the pressure to diversify leaf sources is higher because young leaves have a lower content of ready energy and are usually less concentrated than fruit sources. Furthermore, leaf species may also be diversified to avoid an overload of toxic compounds, since leaves have a higher content of toxins. Therefore, the presence of high floristic diversity may allow howlers to obtain an optimal daily mix of food items (especially leaves) to fulfill their nutritional requirements and avoid toxic secondary compounds. Estrada and Coates-Estrada  also reported a positive relationship between floristic diversity and the number of howlers in the fragments, and related high species diversity to large and undisturbed forest fragments. Different studies support the idea that mature or undisturbed forests are better for howler monkeys [DeGamma-Blanchet & Fedigan, in press; Estrada & Coates-Estrada, 1996; Rodrı́guez-Toledo et al., 2003; Sorensen & Fedigan, 2000]. Nevertheless, this association was not clear in our study, since we found that high tree density (a characteristic of secondary or pioneer vegetation [Bongers et al., 1988]) was positively associated with the presence and abundance of howlers in the fragments, and the floristic diversity of the forest fragments. 218 / Cristóbal-Azkarate et al. In a previous study of howlers at Los Tuxtlas [Juan et al., 2000], their residence in small forest fragments was related to a more folivorous diet, presumably because of a reduction in fruit availability. Considering the socioecological situation in which many of our study groups live (i.e., small fragments (50% o10 ha) with small trees (mean DBH o30 cm in 75% of the fragments, and never Z40 cm), and high population densities (Table V)), we believe that leaves may comprise an important part of their diet. If this is the case, the presence of secondary vegetation in the fragments may be beneficial for howlers, for the following reasons: First, in comparison with mature forest, secondary vegetation is characterized by a relatively higher and less seasonal production of high-quality leaves (i.e., lower levels of chemical defenses and fiber, and higher values of protein, digestible nutrients, and energy) [Chiarello, 2003; Lovejoy et al., 1986]. Additionally, the pressures on howlers to diversify their food species are stronger when they consume leaves than when they consume fruit [Milton, 1980], and secondary vegetation appears to be positively related to floristic diversity. Therefore, secondary vegetation may have a positive effect on howler survival in fragments when fruit availability is reduced, as it offers a higher diversity and better quality of young leaves. We believe this is a very interesting possibility that deserves more attention, as it may change the way we judge the conservation potential that these increasingly abundant transformed habitats have for howler populations. Nevertheless, we consider that further studies are necessary, especially studies that will focus on the feeding habits of howlers in varying degrees of habitat transformation. Finally, other factors that may be associated with the presence and abundance of howlers in the fragments include poaching, which was practiced in the past [Estrada & Coates-Estrada, 1996], and the sampling effect, which means that the inclusion or exclusion of a group of howlers in a fragment at the time of isolation may be entirely due to chance, and independent of the suitability of the habitat [Rodrı́guez-Toledo et al., 2003; Schwarzkoft & Rylands, 1989]. Population Trends and Possible Causes The mean IFR in the study groups was very low (0.45, which is less than half the value reported for the species by Chapman and Balcomb  (Table V); only one group exhibited a higher value). It is noteworthy that there is no consensus on the aging criteria across studies. In fact, Chapman and Balcomb’s  estimate for the species IFR was based on studies that used different aging criteria. For this study we used Carpenter’s  aging criteria, which is the most conservative because it considers individuals of up to 50 months of age as immatures. In contrast, Milton  and Heltne et al.  considered individuals no older than 28 and 36 months, respectively, to be immatures. Therefore, the aging criteria used in this study tend to increase the IFR compared to other studies, which stresses the low value observed in our study groups. The IFR has been used as a measure of a stable, decreasing, or increasing population [Clarke et al., 2002; Heltne et al., 1975; Zucker & Clarke, 2003]. Zucker and Clarke  argued that group size constrains this variable, and that groups tend to show a cyclical pattern of increases and decreases in IFR. Therefore, the determination of the IFR from a single measurement could present an inaccurate view of the overall population dynamics. Nonetheless, 13 of our study groups (79%) in forest fragments exhibited below-average IFR values for the species [Chapman & Balcomb, 1998], which suggests that these low values Howler Demography in Rainforest Fragment / 219 may be due to the general process of demographic decline rather than a cyclical pattern of change. Different factors working independently or in synchrony, such as a high and early immature emigration rate, a reduced birth rate, and a high immature mortality rate, could be responsible for the low IFR found in our study populations. A. palliata is characterized by bisexual emigration of juveniles. Males and females typically emigrate at ~22 months and ~33 months of age, respectively [Glander, 1992], although this may occur much earlier (males at 1 year of age, and females at 2 years of age) [Clarke, 1990]. After they analyzed scars and mutilations, Cristóbal-Azkarate et al.  suggested that the migration patterns of howlers may be depressed in the Los Tuxtlas fragmented landscape. Supporting this hypothesis, our current study did not find any solitary females, and all 11 solitary males observed were adult individuals. On the other hand, the sex ratios of our study groups are very low (Table V), which is consistent with reduced rates of male emigration. These findings suggest that both females and males may be abandoning their natal groups later and at lower rates than in other localities, indicating that the low IFR value observed in our study groups is not principally due to high emigration rates. Milton  interpreted a high number of infants and a low number of juveniles in the groups in Panama (BCI) as an indication of high mortality rates in immatures, and not as a reproductive problem. In our study area, 55.9% of the groups and 66.6% of the fragments had at least one infant, whereas only 34.8% of the groups and 40% of the fragments had juveniles. On the other hand, the mean IFFR of the study groups was 0.32, while the mean JFR was 0.18. Therefore, considering that juvenile emigration may be reduced, we believe that the high immature mortality rates may be the principal cause of the low IFR and JFR values, as well as the reduced number of juveniles that we observed in our study groups. Nevertheless, we cannot dismiss a reduction in female reproductive rates as a possible cause of these reduced values. Infanticide by immigrating males or subordinate males attempting to ‘‘take over’’ from within the group is another possible cause of infant mortality [Clarke, 1983; Clarke et al., 1994]. However, although many studies of howlers at Los Tuxtlas have been conducted, this behavior has never been reported, and there have been no reports immatures showing signs of aggression. Therefore, we consider the possible effects of infanticide on the demography of these populations to be negligible. The regulatory effect that predators may have on the howler monkey populations at Los Tuxtlas is not known, but presumably their numbers are small in this area [Cristóbal-Azkarate et al., 2004]. Therefore, once again, we consider this effect on the demography of howler monkeys at Los Tuxtlas to be negligible. The population densities of our study fragments were very high–more than twice the value reported for this species by Chapman and Balcomb  (Table V). It is noteworthy that these high densities were significantly related to fragment size but not to the number of howlers. High population densities have been related to a reduction in food availability [Fowler, 1987], which has been argued to be one of the most important factors regulating howler population size [Milton, 1982; Otis et al., 1981]. The reduction in food availability has been associated with a reduction in group size [Chapman, 1989; Gaulin et al., 1980] (Table V) and an increase in immature mortality [Fowler, 1987], which could explain the small size and the very low IFR of the howler groups in our study area (Table V). In our fragmented subpopulation’s density distribution there is a gap between 0.71 and 1.37 ind./ha, and none of the groups that inhabited fragments with a population density equal to or larger than 0.71 included juveniles. These 220 / Cristóbal-Azkarate et al. results may suggest that surpassing a population density between these two values is reflected as an increase in immature mortality, which is most likely associated with a reduction in food availability. Finally, other causes of immature mortality, such as disease, cannot be dismissed. Finally, to reverse and avoid this process of population decline, we believe it is essential to first of all protect the remaining forest fragments from the lowintensity logging that still occurs in the region despite the creation of the Biosphere Reserve. Second, it is urgent to increase the amount of available habitat for howlers, especially by connecting forest fragments (both empty and inhabited). This could best be achieved by protecting and reforesting the vegetation along the many streams that cross the landscape in different fragments. These measures, in addition to increasing the amount of available habitat, would break the isolation and facilitate the dispersal of individuals and the genetic flux. CONCLUSIONS Fragment size and floristic diversity are the most important habitat characteristics associated with the presence and abundance of howlers in our study area. The relationship between habitat disturbance and howler demographics in our study area is not clear, but the evidence suggests that given the conditions under which the howlers in our study area live (i.e., small and degraded fragments with high population densities), secondary vegetation may play an important role in howler survival. The fragmented populations of howlers in our study may be in the process of a demographic decline. We believe the reason for this crisis may be an increase in immature mortality due to reduced food availability resulting from high population densities. To reverse this demographic trend, we believe it is imperative to protect the remaining forest fragments and increase the amount of available forest for the howlers by connecting the forest fragments. ACKNOWLEDGMENTS Jurgi Cristóbal Azkarate and Norberto Asensio were supported by a grant from the Zientzi Politikarako Zuzendaritza of the Basque Government. We thank Una Laffan for helping with the translation; Pulga, Lur, Hiru, Manolo, and Amets for their constant company and support; and R. Coates and two anonymous reviewers for their helpful comments and corrections regarding the manuscript. Finally, we thank Domingo Canales Espinosa, director of the Instituto de Neuroetologı́a, University of Veracruz, México, for his support. REFERENCES Baldwin JD, Baldwin JI. 1976. Primate populations in Chiriqui, Panama. In: Thorington RW, Heltne PN, editors. Neotropical primates. Field studies and conservation. Washington, DC: National Academy of Sciences. p 20–31. Bicca-Marques JC. 2003. How do howler monkeys cope with habitat fragmentation? In: Marsh LK, editor. Primates in fragments: ecology and conservation. New York: Kluwer Academics/Plenum Publishers. p 283–303. Howler Demography in Rainforest Fragment / 221 Bongers F, Popma J, Meavene J, Carabias J. 1988. Structure and floristic composition of the lowland rain forest of Los Tuxtlas. Vegetation 44:55–88. Carpenter CR. 1965. The howlers of Barro Colorado Island. In: DeVore I, editor. Primate behavior. New York: Rinehart and Winston. p 250–291. Chapman CA. 1989. Ecological constrains on group size in three species of neotropical primates. Folia Primatol 73:1–9. Chapman CA, Balcomb SR. 1998. Population characteristics of howlers: ecological conditions or group history. Int J Primatol 19:385–403. Chiarello AG. 2003. Primates in the Brazilian Atlantic Forest: the influence of forest fragmentation on survival. In: Marsh LK, editor. Primates in fragments: ecology and conservation. New York: Kluwer Academics/Plenum Publishers. p 99–121. Clarke MR. 1983. Infant-killing and infant disappearance following male takeovers in a groups of free-ranging howling monkeys (Alouatta palliata) in Costa Rica. Am J Primatol 5:241–247. Clarke MR. 1990. Behavioral development and socialization of infants in a freeranging group of howling monkeys. Folia Primatol 54:1–5. Clarke MR, Zucker EL, Glander KE. 1994. Group takeover by a natal male howling monkey (Alouatta palliata) and associated disappearance and injuries of immatures. Primates 35:435–442. Clarke MR, Crockett CM, Zucker EL, Zaldivar M. 2002. Mantled howler population of Hacienda La Pacifica, Costa Rica, between 1991 and 1998: effects of deforestation. Am J Primatol 56:155–163. Collins M. 1990. The last rain forest: a world conservation atlas. New York: Oxford University Press. Cristóbal-Azkarate J, Dias PMD, Veà JJ. 2004. Causes of interspecific aggression in Alouatta palliata mexicana: evidence from injuries, demography, and habitat. Int J Primatol 25:939–953. Cuarón AD, de Grammont PC, Cortés-Ortiz L, Wong G, Silva JCS. 2003. Alouatta palliata ssp. mexicana. In: IUCN 2004. 2004 IUCN red list of threatened species. owww. redlist.org4. Downloaded on 4 January 2005. DeGama-Blanchet H, Fedigan L. 2005. The effects of forest fragment age, isolation, size, habitat type, and water availability on monkey density in a tropical dry forest. In: Alejandro E, Garber P, Pavelka M, Luecke L, editors. New perspectives in the study of Mesoamerican primates: distribution, ecology, behavior and conservation. New York: Kluwer Academic Press (in press). Dirzo R, Garcı́a MC. 1992. Rates of deforestation in Los Tuxtlas, a neotropical area in southeast Mexico. Conserv Biol 6: 85–94. Estrada A. 1984. Resource use by howler monkeys (Alouatta palliata) in the rain forest of Los Tuxtlas, Veracruz, Mexico. Int J Primatol 5:105–131. Estrada A, Coates-Estrada R. 1996. Tropical rain forest fragmentation and wild populations of primates at Los Tuxtlas, Mexico. Int J Primatol 17:759–783. FAO. 2003. Estate of forestry in the Latin American and Caribbean region 2002. http:// www.fao.org/Regional/LAmerica/organos/ coflac/. Fowler CW. 1987. Density dependence as related to life history strategy. Ecology 62: 602–610. Gaulin SJC, Knight DH, Gaulin CK. 1980. Local variance in Alouatta group size and food availability on Barro Colorado Island. Biotropica 12:137–143. Glander KE. 1992. Dispersal patterns in Costa Rican mantled howling monkeys. Int J Primatol 13:415–436. Gómez-Pompa A. 1973. Ecology of the vegetation of Veracruz. In: Graham A, editor. Vegetation and vegetational history of northern Latin America. Amsterdam: Elsevier Scientific Publishing Company. p 89–148. Heltne PG, Turner DC, Scott NJ Jr. 1975. Comparison of census data on Alouatta palliata from Costa Rica and Panama. In: Thorington RW Jr, Heltne PG, editors. Neotropical primates: field studies and conservation. Washington, DC: National Academy of Science. p 10–19. Juan S, Estrada A, Coates-Estrada R. 2000. Contrastes y similitudes en el uso de recursos y patrón general de actividades en tropas de monos aulladores (Alouatta palliata) en fragmentos de selva de Los Tuxtlas, México. Neotrop Primates 8: 131–135. Keel S, Gentry AH, Spinzi L. 1993. Using vegetation analysis to facilitate the selection of conservation sites in eastern Paraguay. Conserv Biol 7:66–75. Lovejoy TE, Bierregaard RO Jr, Rylands AB, Malcolm JR, Quintela CE, Harper LH, Brown KS, Powell AH, Powell GVN, Schubart HOR, Hays MB. 1986. Edge and other effects of isolation on Amazon forest fragments. In: Soulé ME, editor. Conservation biology: the science of scarcity and diversity. Sunderland: Sinauer Associates. p 257–285. McArthur RH, Wilson EO. 1967. The theory of island biogeography. Princeton: Princeton University Press. 222 / Cristóbal-Azkarate et al. Milton K. 1980. The foraging strategy of howler monkeys: a study in primate economics. New York: Columbia University Press. Milton K. 1982. Dietary quality and demographic regulation in a howler monkey population. In: Leigh EG Jr, Rand AS, Windsor DM, editors. The ecology of a tropical forest: seasonal rhythms and long-term changes. Washington, DC: Smithsonian Institution Press. p 273–289. Otis JS, Froelich JW, Thorington RW Jr. 1981. Seasonal and age-related differential mortality by sex in mantled howler monkey, Alouatta palliata. Int J Primatol 2:197–205. Rempel RCA, Elkie P. 1999. Patch Analyst 2.2. Center for Northern Forest Ecosystem Research, URL http://flash.lakeheadu.ca/ ~rrempel/ecology/. Rodrı́guez-Luna E, Cortés-Ortiz L, Miller PH, Ellis S. 1996. Population and habitat viability assessment for the mantled howler monkey (Alouatta palliata mexicana). Apple Valley: IUCN/SSC Conservation Breeding Specialist Group. Rodrı́guez-Luna E, Domı́nguez-Domı́nguez LE, Morales Mávil J, Martı́nez-Morales M. 2003. Foraging strategy changes in Alouatta palliata mexicana troop released on an island. In: Marsh LK, editor. Primates in fragments: ecology and conservation. New York: Kluwer Academics/ Plenum Publishers. p 229–250. Rodrı́guez-Toledo EM, Mandujano S, Garcı́aOrduña F. 2003. Relationship between forest fragments and howler monkeys (Alouatta palliata mexicana). In: Marsh LK, editor. Primates in fragments: ecology and conservation. New York: Kluwer Academics/Plenum Publishers. p 79–97. Schwarzkoft L, Rylands AB. 1989. Primate species richness in relation to habitat structure in Amazonian rainforest fragments. Biol Conserv 48:1–12. Sorensen TC, Fedigan LM. 2000. Distribution of three monkey species along a gradient of regenerating tropical dry forest. Biol Conserv 92:227–240. Zucker EL, Clarke MR. 2003. Longitudinal assessment of immature-to-adult ratios in two groups of Costa Rican Alouatta palliata. Int J Primatol 24:87–101.