Access the full text.
Sign up today, get DeepDyve free for 14 days.
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
International Journal of Biodiversity Science, Ecosystem Services & Management, 2013 Vol. 9, No. 4, 298–310, http://dx.doi.org/10.1080/21513732.2012.709539 Drivers of species richness and abundance of butterﬂies in coffee–banana agroforests in Uganda a,b,c, M.B. Théodore Munyuli * Department of Agriculture, Biology and Environment, National Center for Research in Natural Sciences, CRSN-Lwiro, D.S. Bukavu, Kivu, Democratic Republic of Congo; Département de Nutrition et Diététiques, Centre de Recherche pour la Promotion de la Santé (CRPS), Institut Supérieur des Techniques Médicales, ISTM-Bukavu, Sud-Kivu, Democratic Republic of Congo; Department of Environmental and Natural Resource Economics, Faculty of Natural Resources and Environmental Sciences, Namasagali Campus, Busitema University, PO Box 236, Tororo, Eastern Uganda A study was conducted in 26 sites on agricultural landscapes in Central Uganda to collect baseline information about important drivers of butterﬂy richness and abundance. Data were collected for 1 year (2006) using line transects walk-and-counts, fruit-bait traps and handnets sampling methods. A total of 57,439 individuals belonging to 331 species were collected. Totals of 127, 131 and 299 species were recorded in transect counts, banana-bait and handnets, respectively. Of the 57,439 individuals registered, 75%, 19% and 6% were recorded in transect counts, handnet and banana-bait trap, respectively. Butterﬂy abundance and species richness were signiﬁcantly (p < 0.05) affected by climatic factors (rainfall, temperature) in previous years (2004 and 2005) and richness and abundance of wild nectaring plants. Butterﬂy species rich- ness (not the abundance) decreased with land-use intensity (p < 0.05) and was positively related to the cover of semi-natural habitats. Both butterﬂy species richness and abundance declined sharply with forest distance. Nearby forest remnants and high cover of semi-natural habitats are thus important for conservation of butterﬂies in coffee–banana agroforestry systems and farmers should be encouraged to protect such resources. Keywords: butterﬂy diversity; drivers; farmlands; Lepidoptera conservation; habitats protection; East Africa Introduction by crops. In their larval stage, butterﬂies feed on leaves of several wild plants found in the agricultural systems Butterﬂies have been recognized as a useful biodiversity and therefore release faeces that contain large amounts indicator group of tropical land-use systems because they of nutrients (Munyuli 2010). In addition, butterﬂies are are sensitive and react quickly to subtle changes in environ- food to birds and other predators and are hosts to sev- mental and habitat conditions (Kremen 1994; Libert 1994; eral parasitoids that suppress crop pests (Summerville Brown 1997;Larsen 2008; Özden et al. 2008; Pozo et al. et al. 2001; Cardoso et al. 2008, 2009). Consequently, 2008). Due to their short life cycle, narrow niches and rel- their conservation is essential to sustaining the produc- atively low mobility, they are more sensitive to land-cover tivity of natural and agricultural landscapes. Despite their and land-use changes than long-lived animals. Butterﬂies diversity, ubiquity and ecological importance, butterﬂies are relatively easy to capture, manipulate and identify remain relatively little studied, particularly with regard (Rogo and Odulaja 2001; Fitzherbert et al. 2006;Marin to their ecology, behaviour and functional role in farm- et al. 2009; Nyamweya and Gichuki 2010), which makes land habitats (Marchiori and Romanowski 2006; Stireman them important candidates for monitoring changes in habi- et al. 2009). In agricultural systems, butterﬂies are sus- tat, biodiversity and environmental conditions (Kremen pected to be important pollinators of wild and cultivated 1992;Howardetal. 2000; Cleary 2004; Cleary and Mooers crop species (Munyuli 2010). Roughly, 90% of butterﬂy 2004), including the impact of landscape and habitat man- species live in the tropics (Boriani et al. 2005; Bonebrake agement practices and disturbance regimes in terrestrial et al. 2010). However, knowledge of butterﬂies inhab- ecosystems (Stork et al. 2003; Öckinger and Smith 2008). iting farmland habitats is fairly good in Mediterranean Butterﬂy activities are closely controlled by weather regions compared with the sub-Saharan Africa. The rela- and many species are constrained by climate, mostly tive scarcity of data on tropical butterﬂy populations ham- occupying a small part of the range of their host plants pers the ability to effectively conserve them, particularly as (Fitzherbert et al. 2006; Nyamweya and Gichuki 2010). pollinating agents (Bonebrake et al. 2010) in agricultural Butterﬂies play a signiﬁcant ecological role in agricultural systems. landscapes. They perform essential ecosystem services Across sub-Saharan Africa, surveys of drivers of but- (Rogo and Odulaja 2001; Schmidt and Roland 2006), espe- terﬂy diversity in farmlands are rare in the literature. There cially in the recycling of nutrients (N, P, K) highly needed *Emails: email@example.com; firstname.lastname@example.org © 2013 Taylor & Francis International Journal of Biodiversity Science, Ecosystem Services & Management 299 are a few published papers quantitatively documenting the of agricultural intensity (Bolwig et al. 2006). Detailed diversity of butterﬂies for some agricultural regions in East vegetation, environmental and landscape characteristics of Africa. In Uganda, most of the works on butterﬂies have the 26 sites are presented in Munyuli (2011a). been carried out mainly in natural areas, forest ecosys- tems and in protected areas (Tumuhimbise et al. 1998; Howard et al. 2000; Molleman et al. 2006; Tushabe et al. Butterﬂy sampling 2006). There exist no published data describing extensively Field sampling of butterﬂies was conducted in each of the diversity of butterﬂies found in agricultural landscapes the 26 study sites. In each study site, an area of 1 km in Uganda in relationship to climatic, regional, landscape was selected and divided into ﬁve linear parallel tran- and local drivers. However, such information is impor- sects 1000 m long and 200 m apart. Butterﬂy data were tant for butterﬂy biodiversity conservation in the rural collected on one of the ﬁve transects per study site, and landscapes. the selected transect was surveyed during ﬁve study site While several studies (Chay-Hernández et al. 2006; visits in 2006. Butterﬂies were sampled using three com- Kivinen et al. 2008; Pickens and Root 2008; Dover and plementary methods (transects walk-and-counts, handnets Settele 2009) indicate that climatic, regional and landscape and fruit-bait traps) universally recommended and exten- factors are important drivers for butterﬂies in farmland sively used to survey and monitor butterﬂy populations areas, other studies indicate that local factors (availabi- and communities (Pollard and Yates 1993; Raguso 1993; lity of nectaring resources) are of foremost importance in DeVries and Walla 2001; Kitahara and Sei 2001; Yahner explaining the variation in species richness and total den- 2001;Fermonetal. 2002; Kitahara 2004;Barlowetal. sity of butterﬂies in farmland regions (Pöyry et al. 2009). 2007; Kitahara et al. 2008; van Swaay et al. 2008;Marín It is not clear which local, landscape, regional and cli- et al. 2009;Vu 2009) in terrestrial ecosystems (Lehmann matic factors are important predictors of butterﬂy species and Kioko 2005) in the tropics. These sampling methods richness and abundance in Central Uganda. have been applied in Uganda in previous studies (Owen The aim of this study was to explore the relation- 1971; Coe et al. 1999; Molleman et al. 2006; Akite 2008). ships between butterﬂy assemblages (species richness and Butterﬂy specimens were identiﬁed by consulting litera- abundance) and local, landscape, regional and climatic ture, nomenclature and coloured plates of butterﬂies and drivers. by using the reference collection available at Makerere University Zoology Museum. To avoid a mix up of the Materials and methods results, specimens collected by each of three sampling Study area and selection of study sites methods applied were identiﬁed separately. The taxonomic characteristics of butterﬂies were obtained from standard This study was conducted in 26 different study sites in guides, including (i) Butterﬂies of Kenya (Larsen 1996), the banana–coffee system of Lake Victoria Arc, covering (ii) ‘Butterﬂies of West Africa’ (Larsen 2006) and (iii) several districts of Central Uganda (Munyuli 2009a, ‘Butterﬂies of Uganda’ (Carder and Tindimubona 2002). 2009b, 2009c, 2011a, 2012; Munyuli et al. 2009). The Identiﬁcation of all voucher specimens was conﬁrmed Lake Victoria Arc is characterized by ferrisoils with by a butterﬂy taxonomist based at Makerere University high to medium fertility level and receives on average Zoology Museum. All voucher specimens collected during 1000–1800 mm of rains in a bimodal pattern (rainy the present research were deposited at Makerere University seasons: March–May, September–November; semi-dry Zoology Museum. to dry seasons: June–August, December–February) with 22–28 C and 60–75% of mean annual temperature and relative humidity, respectively (Munyuli 2011b). Rainfall Measurement of local, landscape, regional and climatic amounts and patterns are unpredictable. Several food and drivers cash crops are grown in small-scale monoculture and/or Local drivers polyculture ﬁelds that are integrated into the coffee–banana agroforest production systems where coffee and banana Local drivers (richness and abundance of ﬂoral nectar- are the main crops (Munyuli 2011c). Rural Central Uganda ing resources) of importance for butterﬂies were mea- is a mosaic landscape where ‘islands’ of patches of natural sured following approaches described by Munyuli (2010). habitats (forest fragments, forest reserves, wetlands, wood- These included the percent cover of ﬂowering plants lands) are found scattered within agricultural matrices (trees, shrubs, herbs and weeds), the number of nectaring dominated by linear and non-linear features of semi- plant species and the percentage cover of cultivated ﬂo- natural habitats (fallows, hedgerows, grasslands, woodlots, ral resources (proportion of cultivated pollinator-dependent rangelands). All study sites were situated at approximately and non-pollinator-dependent crops) per 1 km area. Data 790–1000 m above sea level. A total of 26 sites were on herbs were collected from 5 m × 5m(25m ) selected to represent a range of vegetation and habitat quadrats, while those on shrub and trees were collected 2 2 types of varying degrees of anthropogenic disturbance and from 10 m × 50 m (500 m ) quadrats. A total of 2025 m management intensities. Selected study sites were grouped quadrats and 20,500 m quadrats were established per using human population density as a surrogate measure transect per study site. 300 M.B.T. Munyuli Landscape drivers classiﬁes land use around Lake Victoria into four cat- egories: low, medium, high and very high. Land uses The area size of different land-use types in each 1 km site classiﬁed as low are areas where three-quarters of the land was determined using a global positioning system (GPS) are maintained uncultivated compared with one-quarter or using a tape measure for measurement of the dimen- that is under crop/animal production. Land uses classiﬁed sions of small ﬁelds (<50 × 50 m). Different land uses as medium are managed habitat types where there is an were grouped into types. Three landscape variables of almost equal amount of land cultivated and uncultivated, high ecological importance for butterﬂy community stud- whereas land uses classiﬁed as high are areas dominated ies in agricultural matrices were determined for each site: by crops or livestock and land uses classiﬁed as very high (i) the percentage of semi-natural habitat per square kilo- represent large monocultures (estates of tea, sugar, coffee, metre area; the term semi-natural habitats included all etc.). linear and non-linear semi-natural habitats such as fallows, hedgerows, ﬁeld margins, grasslands, roadsides, wood- lands, woodlots, track sides and stream edges. Because fallows play a particular role (as butterﬂy reservoirs and Climatic drivers as foraging/breeding habitats) in the maintenance of many Data on regional climatic factors were obtained from butterﬂy species in rural landscapes, the proportion of meteorological stations located in the study area cover (%) of young fallows per site was calculated along- (Kamenyamigo, Entebbe, Jinja and Kiige meteoro- side the proportion of cover of semi-natural habitats; (ii) logical stations). Using raw data from the stations, monthly the percentage of the land area cultivated; and (iii) the for- means for 10 years (1997–2006) for temperature and est distance, that is, the distance from a given study site rainfall (Figure 1) were calculated to detect trends in to the nearest potential pollinators’ source. Distances up to the rainfall patterns and temperature, since such oscilla- 100 m were measured with a tape measure and distances tions may affect the patterns (trends in occurrence and greater than 100 m were measured with GPS (Garmin activities) of butterﬂy species richness and abundance International, Olathe, KS, USA; corrected to ± 1 m accu- in rural landscapes. Later, other variables (of relevance racy with Pathﬁnder v 2.0). to the study of patterns of butterﬂy communities in farmlands) were calculated: (i) the overall mean rains (means/month/10 years); (ii) the overall daily mean Regional land-use intensity factors minimum and maximum temperatures (mean of 10 years); Different land-use categories were obtained from the (iii) the mean monthly rainfall (2004, 2005, 2006); (iv) the Makerere University Geographic Information Service that mean monthly maximum temperature (2004, 2005, 2006); 0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month of the year Overall monthly mean rainfall (rains) Overall daily mean maximum temperature (°C) Overall daily mean minimum temperature (°C) Figure 1. General pattern in temperature and rainfall data in the coffee–banana system, Central Uganda. Note: Data are means of ﬁve meteorological stations and means of 10 years: 1997–2006. Mean monthly rainfall (mm) Mean daily temperature International Journal of Biodiversity Science, Ecosystem Services & Management 301 and (v) the mean monthly minimum temperature (2004, Pierini (14.5%), Charaxini (12.7%), Mycalesini (10.9%) 2005, 2006). and Limenitidini (8.6%). The dominant (>5% of total indi- viduals recorded) butterﬂy species were Acraea acerata, Bicyclus saﬁtza, Catopsilia ﬂorella and Junonia sophia. Data analysis These dominant and widespread species could be found Data from the three sources (transect walk-and-counts, in all study sites. The most abundant genera were Acraea banana-bait traps and handnets) were pooled to obtain total (17.6%), Junonia (17.4%), Bicyclus (11.1%), Catopsilia butterﬂy biodiversity per transect per study site per sam- (7.9%), Eurema (7.7%), Belenois (6.0%) and Ypthima pling round. Lumping data together helped to minimize (5.1%), whereas the most abundant tribes included variance associated with individual sampling methods Junoniini (22.6%), Acraeini (22.4%), Mycalesini (14.2%), (Bossart et al. 2006). Butterﬂy abundance was calculated Pierini (10.2%), Polyommatini (7.4%) and Satyrini (5.2%). as the total number of individuals recorded per transect Around 91.2% (331 observed species against 360 expected per study site each sampling day, whereas butterﬂy species species) of the species present in the study area were richness was calculated as the total numbers of butterﬂy found. The results suggested that a complete inventory species recorded per transect per study site each sampling of the regional species pool was obtained. Fifteen species round. To determine the suite of explanatory variables most were identiﬁed as the most ecologically important species closely related to butterﬂy species richness and abundance (‘characteristic species’) of the coffee–banana system but- measurements, a correlation analysis was conducted. Thus, terﬂy assemblages, since they occurred with high con- correlation among dependent variables (butterﬂy species stancy value (>50%). These were C. ﬂorella, J. sophia, richness and abundance) and independent variables (cli- B. saﬁtza, A. acerata, Eurema hecabe, Ypthima albida, matic, local, landscape and regional variables) characteriz- Zizula hylax, Acraea ventura, Eurema brigitta, Neptidopsis ing the 26 study sites was tested using Pearson correlation ophione, Junonia eonone, Zizeeria knysna, Cupidopsis after transformation (log10 and arcsine square root trans- cissus, Junonia chorimene and Acrae uvui. They should formation) of the raw data (Munyuli 2009a, 2009b, 2009c; serve as reliable and efﬁcient indicators of change in Munyuli et al. 2009). Based on the correlation matrix conditions in agroecological environments in further sur- of all variables measured, only non-collinear independent veys. They may also be good candidates for monitoring variables that were signiﬁcantly (p < 0.05) related to the state of butterﬂies of the farmland habitats in East dependent variables were selected for further analyses in Africa. simple regression analyses. These also illustrated the trends and magnitude of the effects of independent factors on dependent variables. Scatter plots were used to illustrate Local drivers of butterﬂy richness and abundance the scale dependency of butterﬂies on the different drivers. The results of simple linear and non-linear regressions For all simple regression models obtained, the coefﬁcient indicated that mass ﬂowering wild plant resources (nec- of determinations (R ) were plotted to demonstrate the tar ﬂower abundance) were signiﬁcantly positively related level of inﬂuence of the type of drivers that correlate with to both species richness (Figure 2A) and abundance the response variable (abundance and species richness of (Figure 2B) of butterﬂies. Overall, mass ﬂowering plants butterﬂies). The value of the R was also used to iden- explained 48% and 62% of variations in butterﬂy species tify variables with the most predictive effects. Thus, for richness and abundance, respectively. Similarly, the num- all explanatory variables, the best regression model was ber of nectaring plant species (all weeds, herbs, shrubs, selected based on the highest cross-validated R . trees) was positively related to both species richness To analyse the effects of land-use categories on butter- (Figure 2C) and abundance (Figure 2D) of butterﬂies. ﬂy abundance and diversity, a general linear model analysis The number of nectaring plant species accounted for of variance was conducted with butterﬂy community vari- 38% and 22% of variations in butterﬂy species richness. ables (abundance, species richness) as the dependent vari- This result suggests that diverse ﬂowering plant species ables, and the regional land-use categorical variables (low, stimulated attraction of a species-rich ﬂower-visiting but- medium, high and very high) as ﬁxed factors. The least sig- terﬂy community in the farm landscapes. Unexpectedly, niﬁcant difference (LSD) tests were used as post hoc tests cultivated pollinator-dependent ﬂoral resources were not for multiple comparisons of means. related to butterﬂy species richness. Similarly, the rela- tionship between cultivated pollinator-dependent crops and butterﬂy abundance was not signiﬁcant (p > 0.05), and Results this indicated that some butterﬂy species could not visit Characteristics of butterﬂy communities recorded during animal-pollinated crop species. Apparently, butterﬂies do the surveys not beneﬁt much from available mass ﬂowering pollinator- In total, 331 species and 95 genera were registered com- dependent crops. However, cultivated non pollinator- prising a total of 57,439 individuals in 26 study sites. dependent ﬂoral resources were signiﬁcantly negatively Genera with the greatest number of species were Acraea related to butterﬂy species richness (Figure 2E). The per- (10.8%), Charaxes (8.9%), Bicyclus (7.6%) and Neptis centage of cultivated non-pollinator crops explained 21% (5.7%). The most rich tribes were Acraeni (15.5%), of the variation in butterﬂy species richness. Butterﬂy 302 M.B.T. Munyuli (A) (B) y = 0.5384x + 10.028 y = 20.254x – 2.2314 R = 0.484, n = 26, p < 0.001 R = 0.6246, n = 26, p < 0.001 0 0 020 40 60 5 10 15 20 25 30 35 40 45 50 Percentage cover of mass flowering plants Percentage cover of mass flowering plants (nectar flower abundance)/500 m (nectar flower abundance)/500 m (C) (D) y = 7.1129x + 266.43 y = 0.31x + 13.815 2 1200 2 50 R = 0.3863, n = 26, p < 0.001 R = 0.2225, n = 26, p < 0.001 0 0 020 40 60 80 0 20 40 60 80 Number of nectaring plant species Number of nectaring plant species (weeds/herbs/trees/shrubs)/site (weeds/herbs/trees/shrubs)/site (E) (F) y = –7.3297x + 612.17 y = –0.3587x + 29.66 R = 0.2172, n = 22, p < 0.05 R = 0.1023, n = 22, p = 0.63 5 0 5 15 25 35 45 5 15 25 35 45 Proportion (%) of cover of cultivated Proportion (%) of cover of cultivated non-pollinator dependent crops/site non-pollinator dependent crops/site Figure 2. Relationship between (i) nectar ﬂower abundance (percentage of mass ﬂowering plants), (ii) the abundance of cultivated ﬂoral resources (percentage of cultivated non-pollinator-dependent crops), (iii) the number of ﬂowering plant species (weeds, herbs, trees, shrubs) and the number of butterﬂy species (A, C, E) and butterﬂy abundance (B, D, F) per 2 ha transect/study site. Note: Pollinator-dependent crops included all annual, biannual and perennial non-entomophilous crops potentially offering nectar to butterﬂies. abundance was not signiﬁcantly (p > 0.05) related to richness. Similarly, the proportion of semi-natural habitats the proportion of pollinator-dependent crops, although in a 1 km area was signiﬁcantly positively related to there was a negative trend. These results indicate that butterﬂy species richness (Figure 3C) and not related to the occurrence of butterﬂies in the agricultural landscapes butterﬂy abundance (Figure 3D). Overall, the proportion of Central Uganda was largely inﬂuenced by wild ﬂo- of semi-natural habitats accounted for 23% of the varia- ral resources and by cultivated non-pollinator-dependent tion in butterﬂy species richness. These results indicated crops rather than by cultivated pollinator-dependent ﬂoral that the increase in amount of semi-natural habitats in the resources. landscape was likely to lead to increased species richness (not the abundance) of butterﬂies in agricultural matri- ces of Central Uganda. Forest distance was signiﬁcantly Landscape drivers of butterﬂy richness and abundance negatively related to both butterﬂy abundance (Figure 3F) and butterﬂy species richness (Figure 3E). Thus, the for- Cultivation intensity was signiﬁcantly negatively related to est distance explained 69% and 27% of variations in butterﬂy species richness (Figure 3A), but not to butterﬂy butterﬂy species richness and abundance, respectively abundance (Figure 3B). Proportion of cultivation inten- (Figure 3). sity explained 48% of the variation in butterﬂy species Butterfly species/site Butterfly species/site Butterfly species/site Butterfly individuals/site Butterfly individuals/site Butterfly individuals/site International Journal of Biodiversity Science, Ecosystem Services & Management 303 (A) y = –22.269x + 39.77 y = –520.7x + 869.43 (B) R = 0.1302, n = 26, p = 0.89 50 R = –0.2178, n = 26, p < 0.05 0.4 0.6 0.8 1 0.4 0.6 0.8 1 Proportion of cultivation intensity Proportion of cultivation intensity (C) (D) y = 6.4643x + 303.82 R = 0.1464, n = 26, p = 0.61 y = 0.2754x + 15.781 R = 0.2357, n = 26, p < 0.05 020 40 60 020 40 60 Percentage of semi-natural habitats/site Percentage of semi-natural habitats/site (E) (F) y = –0.0.129x + 34.663 y = –0.2136x + 652.38 R = 0.6969, n = 26, p < 0.001 R = 0.2732, n = 26, p < 0.001 0 500 1000 1500 2000 0 500 1000 1500 2000 Rainforest distance (m) Rainforest distance (m) Figure 3. Relationship between landscape drivers (cultivation intensity, distance to forest margins and percentage of semi-natural habitats) per square kilometre area and butterﬂy species richness (A, C, E) and abundance (B, D, F). Regional drivers of butterﬂy richness and abundance Climatic drivers of butterﬂy richness and abundance There were no signiﬁcant variations in butterﬂy abundance There were signiﬁcant correlations between climatic vari- among the four land-use intensity categories (p > 0.05, ables and butterﬂy community parameters (Table 1). Figure 4B), indicating that butterﬂy abundance was not Rainfall of the year 2004 was negatively correlated to very sensitive to the degree of agriculture intensiﬁca- both species richness (r = –0.47, p < 0.05, n = 26) and tion. However, there were signiﬁcant effects of the land- abundance (r = –0.49, p < 0.05, n = 26) of butter- use intensity on butterﬂy species richness (p < 0.05; ﬂies. Similarly, the mean rainfall of the year 2005 was Figure 4A) that were signiﬁcantly higher in low to medium signiﬁcantly negatively correlated to both species rich- land-use intensity categories than in high and very high ness (r = –0.48, p < 0.001, n = 26) and abundance land-use categories (Figure 4A). Study sites located in the (r = –0.53, p < 0.05, n = 26) of butterﬂies. However, low land-use category had almost double the number of butterﬂy abundance and species richness were not signif- species. icantly (p > 0.05) associated with the mean rainfall of the Butterfly species/site Butterfly species richness/site Butterfly species/site Butterfly individuals/site Butterfly individuals/site Butterfly individuals/site 304 M.B.T. Munyuli (A): GLM-ANOVA, F = 4.09, P = 0.019, DF = (3100) LSD(0.05) = 3.89 abundance and species richness. Both the number and cover of wild blooming plant species were more impor- 40 28.8 tant local drivers for butterﬂies than the abundance of 27.7 cultivated ﬂoral resources in the coffee–banana agrofor- 17.1 est systems in Central Uganda. It is likely that cultivated 13.4 ﬂoral resources (pollinator-dependent and non-pollinator- dependent crops) are not the primary food plants for the majority of butterﬂy species inhabiting farmlands proba- bly because locally grown plants are less rich in nectar, the main ﬂoral resource collected by most butterﬂies forag- ing in agricultural landscapes. A signiﬁcant strong positive Very high High Medium Low relationship between butterﬂy species richness and nectar- Regional land-use intensity gradient ing plant species richness was found in Central Uganda. (B): GLM-ANOVA, F = 1.64, P = 0.208, DF = (3100) 1200 This result is consistent with Simonson et al. (2001) who 592.5 found that butterﬂy species richness was positively corre- lated with total vascular plant species richness and native 564.2 plant species richness in Colorado (USA). Worldwide, pos- itive relationships between butterﬂy diversity and plant 344.3 a species richness are reported (e.g. Kumar et al. 2009): plant 289.6 species serve as important food plants for larvae, ﬂow- ers providing nectar resources for adult butterﬂies and the physical structure of plants create microhabitats providing shelter to numerous species of butterﬂies in rural land- scapes. Thus, high plant species diversity is accepted to Very high High Medium Low be an indicator of good habitat quality for generalist and Regional land-use intensity gradient specialist butterﬂy species (Kumar et al. 2009) inhabiting Figure 4. Effects of the regional land-use intensity categories farmland habitats. on the species richness (A) and abundance (B) of butterﬂies col- In this study, it was observed that butterﬂy species lected from farmlands of Central Uganda in 2006. richness was sensitive to land-use intensity whereas the Notes: Means (x ± SE) followed by different letters are signiﬁ- abundance was not, probably because very few adapted cant at p < 0.05 according to the LSD test. GLM, general linear farmland species can easily be represented by a high num- model; ANOVA, analysis of variance. ber of individuals in the farmlands while the majority (>70%) of species may be ‘rare species’ represented by year 2006 or the overall mean rainfall of 10 years (Table 1). one to two individuals only. While studying the effect In addition, the mean maximum temperature of 10 years of agricultural intensiﬁcation on the homogenization of was positively correlated to both species richness (r = 0.51, Lepidoptera community in Finland, it was observed that the p < 0.05, n = 26) and abundance (r = 0.48, p < 0.05, diversity of butterﬂy and day-active geometrid moth com- n = 26) of butterﬂies. In addition, the mean butterﬂy abun- munities found within 134 differently fragmented land- dance was positively correlated to both the maximum tem- scapes decreased with increasing agricultural intensity perature of the year 2004 (r = 0.51, p < 0.05, n = 26) and with the consequence: promotion of habitat generalists in the maximum temperature of the year 2005 (r = 0.46, landscapes with moderate to intensive agriculture com- p < 0.05, n = 26) (Table 1). These results indicated that pared to region with low agriculture intensity (Ekroos et al. current butterﬂy abundance and/or species richness were 2010). likely being affected by past events. In other words, but- The results from Central Uganda revealed that forest terﬂy communities found in farmlands of Central Uganda distance was the most signiﬁcant landscape driver fol- may be sensitive to climate change. Although rainfall is lowed by the proportion of semi-natural habitats. Both expected to affect positively the number and cover of the abundance and species richness of butterﬂies declined wild and cultivated ﬂowering plant species, no signiﬁcant linearly with increasing forest distance. This result is con- (p > 0.05) correlations were found between the abundance sistent with observations recorded in India where a study (and the diversity) of cultivated and wild ﬂoral resources on adult butterﬂy communities visiting coffee plantations and the mean rainfall of 10 years, or the mean rainfall of around a protected area in the Western Ghats found that the year 2004 or 2005 (Table 1). both butterﬂy species composition and species richness declined with increasing forest distance (Dolia et al. 2008) up to 2000 m. Distance of up to 1700 m from mature Discussion forests did not have any signiﬁcant effect on butterﬂy Local, landscape and regional drivers of butterﬂy species richness in a mosaic landscape in Central Sulawesi abundance and diversity in Indonesia (Veddeler et al. 2005), whereas in Central Among local drivers, the abundance and species richness Uganda, butterﬂy species richness declined sharply up of wild ﬂowering plants signiﬁcantly affected butterﬂy to a distance of 2000 m. The fact that forest distance Butterfly individuals/2 ha/site Butterfly species/2 ha/site International Journal of Biodiversity Science, Ecosystem Services & Management 305 Table 1. Cross-correlation matrix showing naïve multiple correlations of Lepidoptera (butterﬂy) variables with environmental, local and landscape variables. BC D E F H I J K L M N O P Q R S T U V W X ∗∗ ∗ ∗ ∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗ ∗∗ ∗∗ ∗∗ A 0.74 −0.52 −0.44 0.46 0.71 −0.68 −0.74 −0.70 −0.50 0.64 0.62 0.57 ∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗ ∗∗ ∗ ∗ B −−0.49 −0.57 0.77 0.66 −0.94 −0.95 −0.97 −0.46 0.69 0.49 0.44 ∗∗ ∗ ∗∗ ∗∗ ∗ C −−0.59 0.52 0.57 0.61 0.47 ∗∗ ∗ ∗∗ ∗∗ ∗∗ ∗ D − 0.57 0.52 0.55 0.69 0.78 −0.50 ∗∗ ∗∗ ∗∗ ∗ E −−0.78 −0.79 −0.72 0.52 ∗∗ ∗∗ ∗∗ ∗ ∗ F −−0.64 −0.72 −0.58 −0.46 0.45 ∗∗ ∗∗ ∗∗ ∗ ∗∗ H − 0.93 0.91 0.83 0.52 −0.58 ∗∗ ∗ ∗∗ ∗ ∗ I − 0.87 0.47 −0.67 −0.46 −0.45 ∗∗ ∗ J − −0.67 −0.49 ∗∗ ∗ ∗ ∗ ∗ K − 0.83 −0.50 0.51 −0.47 −0.48 ∗ ∗ ∗ ∗ ∗ L − 0.48 0.51 0.46 −0.49 −0.53 ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ M −0.84 0.84 −0.76 −0.96 −0.85 −0.75 −0.79 0.76 0.67 ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ N −−0.69 0.69 0.85 0.67 0.56 0.64 −0.55 −0.58 ∗∗ ∗∗ ∗∗ ∗ ∗∗ O −−0.56 0.64 0.58 0.43 −0.77 ∗∗ ∗∗ ∗∗ ∗ ∗ ∗∗ ∗∗ P −−0.93 −0.95 −0.93 −0.52 −0.47 0.94 0.93 ∗∗ ∗∗ ∗∗ ∗∗ Q − 0.90 0.89 −0.78 −0.98 ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ R − 0.93 0.61 0.63 −0.84 −0.84 ∗∗ ∗∗ ∗∗ ∗∗ S − 0.60 0.54 −0.97 −0.88 ∗∗ ∗∗ ∗∗ T − 0.83 0.97 −0.64 ∗∗ U − 0.87 ∗∗ V −−0.54 ∗∗ W − 0.80 X − ∗ ∗∗ Notes: Different levels of signiﬁcance of Spearman rank of correlation coefﬁcients: p < 0.05; p < 0.001, otherwise not signiﬁcant when no value is given. A, human population density; B, cultivation intensity; C, ﬂowering plant species richness; D, density of plant species (abundance of ﬂoral resources); E, cultivated with ﬂoral resources (pollinator-dependent crops); F, cultivated without resources (pollinator non-dependent crops); H, forest distance; I, percentage of semi-natural habitats/site; J, percentage of young fallows per site; K, mean butterﬂy species richness; L, mean butterﬂy abundance; M, overall mean rainfall (10 years); N, overall daily mean maximum temperature (mean 10 years); O, overall daily mean minimum temperature (mean 10 years); P, mean monthly rainfall (2006); Q, mean monthly maximum temperature (2004); R, mean monthly maximum temperature (2005); S, mean monthly maximum temperature (2006); T, mean monthly minimum temperature (2004); U, mean monthly minimum temperature (2005); V, mean monthly minimum temperature (2006); W, mean monthly rainfall (2004); and X, mean monthly rainfall (2005). 306 M.B.T. Munyuli was negatively related to butterﬂy species richness and Yates 1997; Roy and Sparks 2000; Watt and McFarlane abundance suggests that surrounding natural forests of 2002). Climatic variables may therefore be very impor- agricultural matrices remain an irreplaceable habitat for tant in determining emergency (van Asch and Visser 2007; several butterﬂy species belonging to different functional Saastamoinen and Hanski 2008), foraging, reproduction groups (e.g. breeding strategy) and life history traits in and breeding activities of many butterﬂy species in sub- Central Uganda. In addition, this observation suggests Saharan Africa. In fact, seasonal ﬂight activity, as well that the ability of different agricultural land uses (human- as species diversity and overall abundance vary greatly dominated landscapes) to support Lepidoptera communi- between sampling rounds (Munyuli 2010) in a study site. ties can be enhanced if tree plantations/crop ﬁelds are It is well established that temperature is a key factor established in proximity to surrounding primary forests increasing daily activities (ﬂight, foraging movements) (Hawes et al. 2009). Obviously, a contiguous distribution of butterﬂies, whereas rainfall affects indirectly butterﬂy of large secondary forest patches with many host trees is through enhancing the availability of nectaring resources. very important to conserve butterﬂy species in fragmented Most butterﬂy species need resources like host plants (for agricultural landscapes (Kobayashi et al. 2009). The obser- larvae), nectar plants (for adults) and sites for resting, vation made in Central Uganda about the effect of forest and both natural/semi-natural habitats and nectar plant fragments is consistent with the conclusion of Koh (2008) resources are also affected by rainfall pattern (Öckinger who clearly demonstrated that natural forests serve as an et al. 2009; Rossi and van Halder 2010). important population sources for butterﬂy species occur- In this study, the abundance and species occurrence of ring in nearby agricultural matrices. In terms of habitat butterﬂies were signiﬁcantly related to rainfall and mean use, butterﬂies were found more in forest margins, forest maximum temperature in previous years (2004, 2005), fallows, wetlands, swampy habitats, woodlots, woodlands, not the year of study (2006). The mean precipitation of grasslands and in adjacent ﬁelds (Munyuli 2011a) as com- the past 10 years was not signiﬁcantly associated with pared with ﬁelds that were located far from such natural any measure of richness and abundance of butterﬂies, but and semi-natural habitats. Forest is a primary factor dictat- the mean maximum temperature in the same period was. ing the occurrence of butterﬂies in agricultural landscapes The fact that butterﬂy community variables were related of Central Uganda. to mean maximum temperature and rainfall in previous In this study, both the abundance and diversity of but- years rather than current has also been found in tem- terﬂies increased with increasing amount of semi-natural perate regions (e.g. Pollard 1988). In relation to rainfall, habitats per square kilometre. Similarly, in southeastern results obtained from Uganda coincide at least partially Sweden (Jonason et al. 2010), butterﬂy species richness with results from Malaysia, where it was found (Brehm (not the abundance) was found to be positively related to 2005) that monthly rainfall variation correlated with moth increasing tree cover in the farm landscape. Apart from populations in previous months of the year, weather factors land use and farm management methods or farming prac- and plant phenology in the lowland dipterocarp forest. The tices (Weibull and Östman 2003), vegetation structure, intensity of tree ﬂowering in the previous months was an quality of the matrix surrounding an agricultural habi- important environmental factor that correlated positively tat (Binzenhöfer et al. 2008; Summerville et al. 2008), with the numbers of species and individuals of moths that diversity and types of habitats (Dessuy and de Morris emerged in any subsequent month (Intachat et al. 2001). 2007;Ngaietal. 2008; Kumar et al. 2009), landscape High rainfall in the previous 3 months led to an increase in heterogeneity and habitat connectivity (Davis et al. 2007) moth abundance in the following month (perhaps by stim- are important factors determining occurrence, movements, ulating an increase in fresh plant material), whereas high population dynamics, seasonality, persistence and long- rainfall and relative humidity thereafter served to decrease term survival of Lepidoptera faunal communities in the abundance, possibly by encouraging the spread and activity agricultural landscapes (Dennis 2003; Greza et al. 2004; of pathogens that impacted on early life-stage survivorship Chay-Hernández et al. 2006; Kivinen et al. 2008; Öckinger (Intachat et al. 2001). It was concluded that the diversity of and Smith 2008; Pickens and Root 2008; Stasek et al. geometrid moths correlated more with weather parameters 2008; Dover and Settele 2009; Brückmann et al. 2010). than with tree phenology (Intachat et al. 2001) in tropi- Generally, agricultural matrices that are more resembling a cal regions, and important weather parameters that inﬂu- nearby forest patch maintain higher butterﬂy diversity than enced moth abundance include monthly rainfall, relative matrices with lesser shade cover (Summerville et al. 2001; humidity and minimum temperature in previous months. Kitahara and Watanabe 2003; Weibull and Östman 2003; In addition, it is generally agreed that weather variables can Boriani et al. 2005;Avironetal. 2007; Ohwaki et al. 2007, have long-term effects on population ﬂuctuations of dif- 2008; Barlow et al. 2008; Bergman et al. 2008; van Halder ferent butterﬂy taxa (Sei-Woong 2003). Weather variables et al. 2008; Marín et al. 2009). can help in detecting potential butterﬂy population changes following changes (Sei-Woong 2003) in climatic factors (changes in temperature regimes, rainfall patterns, relative Climatic drivers of butterﬂy richness and abundance humidity) that are believed to affect negatively/positively It is generally accepted that climate factors regulate most species richness, abundance and distribution of butterﬂies insect species’ life cycles, including butterﬂies (Sparks and regionally as well as locally (Sei-Woong 2003; Kivinen International Journal of Biodiversity Science, Ecosystem Services & Management 307 et al. 2007). Although in Uganda the study was conducted Lepidoptera. It was also found that butterﬂy communi- on butterﬂies in agricultural systems, while in Malaysia ties were affected more by previous climatic events than the study was conducted on moths, the two studies were current ones. Landscape managers should advocate for able to show that Lepidoptera biodiversity can be related to the maintenance of high-quality habitat matrix in agricul- previous events (historical events) in weather factors. tural mosaic farm landscapes to protect native butterﬂy Practically, weather factors affect the production of faunal diversity and reduce their vulnerability to fur- ﬂoral resources in current and previous years. High pro- ther environmental/climatic changes in the region. This duction of ﬂowers in the previous wet months of the year could be supported by policies that enhance protection may guarantee the availability of food for different butter- and restoration of forest fragments and promotion of inter- ﬂy taxa (larval and adult stages), and this in turn may tend ventions that enable farmers to adopt butterﬂy friendly to increase larval survivorship and therefore the abundance conservation and farming practices, such as agroforestry of adult butterﬂies. In some cases, the abundance of adult systems with multipurpose tree species, to sustain butterﬂy butterﬂies can increase with high abundance and diversity communities and services in rural landscapes of Uganda. of ﬂowering plants in the current and not in previous years. To provide incentives to the conservation of butter- An increase in the number of host trees ﬂowering would ﬂy fauna in agricultural landscapes in Uganda, further also mean an increased number of oviposition sites for the research is needed to understand the behavioural ecology butterﬂies (particularly forest specialist species). of butterﬂies and to determine the pollination efﬁciency High rainfall in the months of the current year may (Martins and Johnson 2009) of different butterﬂy species deter butterﬂies from ﬂying, thus resulting in a low catch (as well as for other Lepidoptera taxa). Such studies will in traps. High rainfall in the previous months of the year help in highlighting the importance of butterﬂies (and may also decrease the survivorship of the larvae, resulting moths) as a group providing pollination services of high in lower adult butterﬂy catch in the current month. During economic importance to both wild and cultivated plants in very rainy periods, larvae are more likely to be washed off Uganda and in East and Central Africa. their host plants. Combination of a wetter period and a high relative humidity in the months before adult emergence Acknowledgements may also increase activities of disease microorganisms dur- I am very grateful to the Darwin Initiative (Defra, UK; project ing that period, resulting in higher risk of larval mortality reference: 14-032; project title: Conserving biodiversity in mod- from disease and therefore reduction in adult abundance ernized farm landscapes in Uganda) for funding this study. I am in the current month. Rainfall in the current month may also very grateful to project leaders Dr Juliet Vickery (RSPB- also encourage disease at the pupal stage and cause mor- Cambridge University, UK), Dr Phil Atkinson (British Trust for Ornithology, UK), Prof. Derek Pomeroy (Makerere University, tality if heavy during the vulnerable eclosion period. High Uganda), scientiﬁc supervisors Prof. Simon Potts (University rainfall, on the other hand, can encourage blooming of of Reading, UK) and Prof. Philip Nyeko (Makerere University, wild and cultivated plants in most tropical agroecosys- Uganda). I am also grateful to Dr Bwinja M and to Mr Maurice tems for the beneﬁt of butterﬂies (Munyuli 2010) and Mutabazi (research assistant) for his assistance in the ﬁeld. I am grateful to Akite Perpetra who helped during identiﬁcation of but- other Lepidoptera. Heavy blossoms in the previous months terﬂy to species level at Makerere University Zoology Museum. would encourage adult and larval survivorship and there- I am also grateful to anonymous reviewers of this journal for fore increase the number of emergence adult butterﬂies in constructive criticisms and comments on earlier versions of this the following months of the following year. article. In summary, if the annual and seasonal changes in rain- fall and temperature continue to oscillate and if rainfall continue diminishing while temperature continue rising in References Uganda (Munyuli 2010) this may, combined with already Akite P. 2008. Effects of anthropogenic disturbances on the existing negative impacts of environmental degradation, diversity and composition of the butterﬂy fauna of sites in the Sango Bay and Iriiri areas, Uganda: implications for lead to a high loss/decline in butterﬂy biodiversity while conservation. Afr J Ecol. 46(1):3–13. increasing the vulnerability of ecosystems relying on these Aviron ST, Kindlmanna P, Burel F. 2007. Conservation of keystone species to function properly. butterﬂy populations in dynamic landscapes: the role of farming practices and landscape mosaic. Ecol Model. 205(1–2):135–145. Conclusion Barlow J, Araujo IS, Overal WL, Gardner TA, da Silva Mendes F, Lake IR, Peres CA. 2008. Diversity and composition of Findings of this study indicated that the abundance and fruit-feeding butterﬂies in tropical Eucalyptus plantations. species diversity of butterﬂies in the coffee–banana agro- Biodivers Conserv. 17(5):1089–1104. forestry systems of Central Uganda were affected by local, Barlow J, Overal WL, Araujo IS, Gardner TA, Peres CA. 2007. The value of primary, secondary and plantation forests for landscape, regional and climate drivers. Plant diversity was fruit-feeding butterﬂies in the Brazilian Amazon. J Appl the most inﬂuential local factor, whereas forest distance Ecol. 44(5):1001–1012. and cultivation intensity were the most inﬂuential land- Bergman KO, Ask L, Askling J, Ignell H, Wahlman H, Milberg P. scape drivers. The presence of natural habitats (e.g. forest 2008. Importance of boreal grasslands in Sweden for butterﬂy reserves, wetlands) and semi-natural habitats in the vicinity diversity and effects of local and landscape habitat factors. of agricultural matrices was related to high diversity of Biodivers Conserv. 17(1):139–153. 308 M.B.T. Munyuli Binzenhöfer B, Biedermann R, Settele J, Schröder B. 2008. Dolia J, Devy MS, Aravind NA, Kumar A. 2008. Adult but- Connectivity compensates for low habitat quality and small terﬂy communities in coffee plantations around a protected patch size in the butterﬂy Cupido minimus.EcolRes. area in the Western Ghats, India. Animal Conserv. 11(1): 23(2):259–269. 26–34. Bolwig S, Pomeroy D, Tushabe H, Mushabe D. 2006. Crops, Dover J, Settele J. 2009. The inﬂuences of landscape structure trees, and birds: biodiversity change under agricultural inten- on butterﬂy distribution and movement: a review. J Insect siﬁcation in Uganda’s farmed landscapes. Danish J Geogr. Conserv. 13(1):3–27. 106(2):115–130. Ekroos J, Heliölä J, Kuussaari M. 2010. Homogenization of Bonebrake TC, Ponisio LC, Boggs CL, Ehrlich PR. 2010. More Lepidopteran communities in intensively cultivated agricul- than just indicators: a review of tropical butterﬂy ecology and tural landscapes. J Appl Ecol. 47(2):459–467. conservation. Biol Conserv. 143(9):1831–1841. Fermon H, Waltert M, Mühlenberg M. 2002. Movement and Boriani L, Burgio G, Marini M, Genghini M. 2005. Faunistic vertical stratiﬁcation of fruit-feeding butterﬂies in a man- study on butterﬂies collected in Northern Italy rural land- aged West African rainforest. J Insect Conserv. 7(2): scape. Bull Insectol. 58(1):49–56. 7–19. Bossart JL, Opuni-Frimpong E, Kuudaar S, Krumah S. 2006. Fitzherbert E, Gardner T, Davenport TRB, Caro T. 2006. Butterﬂy Richness, abundance, and complementarity of fruit-feeding species richness and abundance in the Katavi ecosystem of western Tanzania. Afric J Ecol. 44(3):353–362. butterﬂy species in relict sacred forests and forest reserves of Greza A, Zaviezo T, Reyes S. 2004. Short-term effects of habitat Ghana. Biodivers Conserv. 15(1):333–359. Brehm G. 2005. Diversity and community structure of geometrid fragmentation on the abundance and species richness of bee- moths of disturbed habitat in a montane area in the tles in experimental alfalfa micro-landscapes. Rev Chil Hist Ecuadorian Andes. J Res Lepid. 38(1999):1–14. Nat. 77(3):547–558. Brown JKS. 1997. Diversity, disturbance, and sustainable use Hawes J, Motta CDS, Overal WL, Barlow J, Gardner TA, Peres of neotropical forests: insects as indicators for conservation CA. 2009. Diversity and composition of Amazonian moths monitoring. J Insect Conserv. 1(1):25–42. in primary, secondary and plantation forests. J Trop Ecol. Brückmann SV, Krauss J, Steffan-Dewenter I. 2010. Butterﬂy 25(3):281–300. and plant specialists suffer from reduced connectivity in Howard PC, Davenport TRB, Kigenyi FW, Viskanic P, Baltzer fragmented landscapes. J Appl Ecol. 47(4):799–809. MC, Dickinson CJ, Lwanga J, Matthews RA, Mupada E. Carder N, Tindimubona L. 2002. Butterﬂies of Uganda. A ﬁeld 2000. Protected area planning in the tropics: Uganda’s guide to butterﬂies and silk moths on the collection of the national system of forest nature reserves. Conserv Biol. Uganda Society. Kampala (Uganda): The Uganda Society. 14(3):858–875. 89 p. Intachat J, Holloway JD, Staines H. 2001. Effects of weather Cardoso P, Henriques SS, Gaspar C, Crespo LC, Carvalho and phenology on the abundance and diversity of geometroid R, Schmidt JP, Sousa P, Szuts ˝ T. 2009. Species richness moths in a natural Malaysian tropical rain forest. J Trop Ecol. and composition assessment of spiders in a Mediterranean 17(3):411–429. Scrubland. J Insect Conserv. 13(1):45–55. Jonason D, Milberg P, Bergman KO. 2010. Monitoring of butter- Cardoso P, Scharff N, Gaspar C, Henriques SS, Carvalho ﬂies within a landscape context in south-eastern Sweden. J R, Castro PH, Schmidt JB, Silva I, Szüts T, Castro Nat Conserv. 18(1):22–33. AD, et al. 2008. Rapid biodiversity assessment of spiders Kitahara M. 2004. Butterﬂy community composition and con- (Araneae) using semi-quantitative sampling: a case study in servation in and around a primary woodland of Mount Fuji, a Mediterranean forest. Insect Conserv Diver. 1(2):71–84. central Japan. Biodivers Conserv. 13(5):917–942. Chay-Hernández DA, Delfín-González H, Parra-Tabla V. 2006. Kitahara M, Sei K. 2001. A comparison of the diversity and struc- ture of butterﬂy communities in semi-natural and human- Ichneumonoidea (Hymenoptera) community diversity in an modiﬁed grassland habitats at the foot of Mt. Fuji, central agricultural environment in the State of Yucatan, Mexico. Japan. Biodivers Conserv. 10(3):331–351. Environ Entomol. 35(5):1286–1297. Cleary DFR. 2004. Assessing the use of butterﬂies as indica- Kitahara M, Watanabe M. 2003. Diversity and rarity hotspots tors of logging in Borneo at three taxonomic levels. J Econ and conservation of butterﬂy communities in and around the Entomol. 97(2):429–364. Aokigahara woodland of Mount Fuji, central Japan. Ecol Res. Cleary DFR, Mooers A. 2004. Butterﬂy species richness and 18(5):503–522. community composition in forests affected by ENSO- Kitahara M, Yumoto M, Kobayashi T. 2008. Relationship of induced burning and habitat isolation in Borneo. J Trop Ecol. butterﬂy diversity with nectar plant species richness in and 20(4):359–367. around the Aokigahara primary woodland of Mount Fuji, Coe MJ, McWilliams NC, Stone GN, Packer MJ. 1999. Mkomazi: Central Japan. Biodivers Conserv. 17(11):2713–2734. the ecology, biodiversity and conservation of a Tanzanian Kivinen S, Luoto M, Heikkinen RK, Saarinen K, Ryttäri Savannah. London (UK): Royal geographical Society (with T. 2008. Threat spots and environmental determinants the Institute of British Geographers). p. 608. of red-listed plant, butterﬂy and bird species in boreal Davis JD, Debinski DM, Danielson BJ. 2007. Local and agricultural environments. Biodivers Conserv. 17(13): landscape effects on the butterﬂy community in frag- 3289–3305. mented Midwest USA prairie habitats. Landsc Ecol. 22(9): Kivinen S, Luoto M, Kuussaari M, Saarinen K. 2007. Effects 1341–1354. of land cover and climate on species richness of butter- Dennis P. 2003. Sensitivity of upland arthropod diversity to live- ﬂies in boreal agricultural landscapes. Agric Econ Environ. stock grazing, vegetation structure and landform. J Food, 122(4):453–460. Agric Environ. 1(2):301–307. Kobayashi T, Nakashizuka T, Kitahara M. 2009. Effects of Dessuy MB, de Morris ABB. 2007. Diversity of butterﬂies fragmentation of secondary broadleaf deciduous forests on (Lepidoptera, Papilionoidea and Hesperioidea) in fragments populations of the near-threatened butterﬂy, Sasakia charonda of decidual seasonal forests in Santa Maria, Rio Grande do (Lepidoptera, Nymphalidae), in Central Japan. Ecol Res. Sul State, Brazil. Rev Bras Zool. 24(1):108–120. 24(1):57–64. DeVries PJ, Walla TR. 2001. Species diversity and community Koh LP. 2008. Can oil palm plantations be made more hos- pitable for forest butterﬂies and birds? J Appl Ecol. structure in neotropical fruit feeding butterﬂies. Biol J Linn 45(4):1002–1009. Soc. 74(1):1–15. International Journal of Biodiversity Science, Ecosystem Services & Management 309 Kremen C. 1992. Assessing the indicator properties of species Munyuli TMB. 2012. Diversity of life-history traits, functional assemblages of natural areas monitoring. Ecol Appl. groups and indicator species of bee communities from farm- 2(2):203–217. lands of central Uganda. Jordan J Biol Sci. 5(1):1–14. Kremen C. 1994. Biological inventory using target taxa: a Munyuli TMB, Kyamanywa S, Luther GC. 2009. Effects of crop- case-study of the butterﬂies of Madagascar. Ecol Appl. ping system and insecticide application on the incidence of 4(3):407–422. arthropod parasitoids of cowpea insect pests in Uganda and Kumar S, Simonson SA, Stohlgren TJ. 2009. Effects of spa- Democratic Republic of the Congo. Tunisian J Plant Protec. tial heterogeneity on butterﬂy species richness in Rocky 4(1):76–90. Mountain National Park, CO, USA. Biodivers Conserv. Ngai JT, Kirby KR, Gilbert B, Starzomski BM, Pelletier AJD, 18(3):739–763. Conner JCR. 2008. The impact of land-use change on lar- Larsen TB. 1996. The butterﬂies of Kenya and their natural val insect communities: testing the role of habitat elements in history. Oxford (UK) : Oxford University Press. 453 p. conservation. Ecoscience. 15(2):160–168. Larsen TB. 2006. Butterﬂies of West Africa. Stockholm Nyamweya NH, Gichuki NN. 2010. Effects of plant structure on (Sweden): Apollo Publications. 612 p. butterﬂy diversity in Mt. Marsabit Forest–northern Kenya. Larsen TB. 2008. Forest butterﬂies in West Africa have Afric J Ecol. 48(2):304–312. resisted extinction ... so far (Lepidoptera: Papilionoidea and Öckinger E, Dannestam Å, Smith HG. 2009. The importance of fragmentation and habitat quality of urban grasslands for Hesperioidea). Biodivers Conserv. 17(12):2833–2847. butterﬂy diversity. Landsc Urban Plan. 93(1):31–37. Lehmann I, Kioko E. 2005. Lepidoptera diversity, ﬂoristic com- position and structure of three kaya forests on the south coast Öckinger E, Smith HG. 2008. Do corridors promote disper- of Kenya. J East Afr Nat Hist. 94(1):121–163. sal in grassland butterﬂies and other insects? Landsc Ecol. Libert M. 1994. Monitoring of numbers and species composi- 23(1):27–40. tion of the butterﬂy community of two forested hills near Ohwaki A, Nakamura K, Tanabe SI. 2007. Butterﬂy assem- Yaounde, Cameroun. Rev Ecol-Terre Vie. 25(2):151–175. blages in a traditional agricultural landscape: importance of Marchiori OM, Romanowski HP. 2006. Species composition secondary forests for conserving diversity, life history spe- and diel variation of a butterﬂy taxocene (Lepidoptera, cialists and endemics. Biodivers Conserv. 16(5):1521–1539. Papilionoidea and Hesperioidea) in a restinga forest at Itapua Ohwaki A, Tanabe S-I, Nakamura K. 2008. Effects of anthro- State Park, Rio Grande de Sul, Brazil. Rev Bras Zool. pogenic disturbances on the butterﬂy assemblage in an urban 23(2):443–454. green area: the changes from 1990 to 2005 in Kanazawa Marín L, León-Cortés JL, Stefanescu C. 2009. The effect of Castle Park, Japan. Ecol Res. 23(4):697–708. an agro-pasture landscape on diversity and migration pat- Owen DF. 1971. Tropical butterﬂies. The ecology and behavior terns of frugivorous butterﬂies in Chiapas, Mexico. Biodivers of butterﬂies in the tropics with special reference to African Conserv. 18(4):4919–4934. species. Oxford (UK): Oxford Clarendon Press. 228 p. Martins DJ, Johnson SD. 2009. Distance and quality of natural Özden Ö, Ciesla WM, Fuller WJ, Hodgson DJ. 2008. Butterﬂy habitat inﬂuence hawkmoth pollination of cultivated papaya. diversity in Mediterranean islands and in Pentadaktylos Inter J Tropic Insect Sci. 29(3):114–123. Pinus brutia forests of Cyprus. Biodivers Conserv. Molleman F, Kop A, Brakeﬁeld PM, Devries PJ, Zwaan BJ. 2006. 17(12):2821–2832. Vertical and temporal patterns of biodiversity of fruit feeding Pickens BA, Root KV. 2008. Factors affecting host-plant quality butterﬂies in a tropical forest in Uganda. Biodivers Conserv. and nectar use for the Karner blue butterﬂy: implications for 15(1):107–121. oak savanna restoration. Nat Areas J. 28(3):210–217. Munyuli TMB. 2009a. Is Pardosa pseudoannulata an effec- Pollard E. 1988. Temperature, rainfall and butterﬂy numbers. J tive predator agent of Aphis craccivora in Uganda and in Appl Ecol. 25(3):819–828. Democratic Republic of Congo? Tunisian J Plant Protec. Pollard E, Yates TG. 1993. Monitoring butterﬂies for ecology and 4(1):91–98. conservation. Vol. 1. London (UK): Chapman & Hall. Munyuli TMB. 2009b. Effects of native insect predators on pop- Pöyry J, Paukkunen J, Heliölä J, Kuussaari M. 2009. Relative ulation densities of Aphis craccivora and yields of Vigna contributions of local and regional factors to species rich- unguiculata and Arachis hypogeae grown under various crop- ness and total density of butterﬂies and moths in semi-natural ping systems, in Kivu province, eastern democratic republic grasslands. Oecologia. 160(3):577–587. of Congo. Tunisian J Plant Protec. 4(2):197–209. Pozo C, Luis-Martínez A, Lorente-Bousquets J, Salas-Suárez Munyuli TMB. 2009c. On-farm storages participatory evalu- N, Vargas-Fernández I, Warren AD. 2008. Seasonality and ation and validation of the capability of native botani- phenology of the butterﬂies (Lepidoptera: Papilionoidea and cals for control of bean bruchids (Acanthoscelides obtec- Hesperioidea) of Mexico’s Calakmul region. Fla Entomol. tus L., Coleoptera: Bruchidae) in South-Kivu province, 91(3):407–422. eastern of Democratic Republic of Congo. Tropicultura. Raguso RA. 1993. Preliminary checklist and ﬁeld observations of 27(3):174–183. the butterﬂies of the Maquipucuna Field Station (Pichincha Munyuli TMB. 2010. Pollinator biodiversity and economic Province, Ecuador). J Res Lepid. 32(3):135–161. value of pollination services in Uganda [PhD dissertation]. Rogo L, Odulaja A. 2001. Butterﬂies in two forest fragments at [Kampala (Uganda)]: Makerere University. 431 p. the Kenya coast. Afr J Ecol. 39(3):266–275. Munyuli TMB. 2011a. Pollinator biodiversity in Uganda and Rossi JP, van Halder I. 2010. Towards indicators of butterﬂy in Sub-Sahara Africa: landscape and habitat management biodiversity based on a multi-scale landscape description. strategies for its conservation. Inter J Biodivers Conserv. Ecol Indic. 10(2):452–458. 3(11):551–609. Roy DB, Sparks TM. 2000. Phenology of British butterﬂies and Munyuli TMB. 2011b. Factors governing ﬂower-visitation pat- climate change. Glob Change Biol. 6(4):407–416. terns and quality of pollination services delivered by social Saastamoinen M, Hanski I. 2008. Genotypic and environmen- and solitary bee species to coffee in central Uganda. Afric J tal effects on ﬂight activity and oviposition in the Glanville Ecol. 49(4):501–509. fritillary butterﬂy. Am Nat. 171(6):701–712. Munyuli TMB. 2011c. Farmers’ perceptions of pollinators’ Schmidt BC, Roland J. 2006. Moth diversity in a fragmented habi- importance in coffee production in Uganda. Agric Sci. tat: importance of functional groups and landscape scale in 2(3):318–333. the boreal forest. Ann Entomol Soc Am. 99(6):1110–1120. 310 M.B.T. Munyuli Sei-Woong C. 2003. The relationship between local distribution Tushabe H, Kalema J, Byaruhanga A, Asasira J, Ssegawa P, and abundance of butterﬂies and weather factors. Korean J Balmford A, Davenport T, Fjeldsa J, Friis I, Pain D, et al. Ecol. 26(4):199–202. 2006. A nationwide assessment of the biodiversity value Simonson SE, Opler PA, Stohlgren TJ, Chong GW. 2001. Rapid of Uganda’s important bird areas network. Conserv Biol. assessment of butterﬂy diversity in a montane landscape. 20(1):85–99. Biodivers Conserv. 10(8):1369–1386. van Asch M, Visser ME. 2007. Phenology of forest caterpillars Sparks TH, Yates TJ. 1997. The effect of spring temperature and their host trees: the importance of synchrony. Ann Rev on the appearance dates of British butterﬂies 1883–1993. Entomol. 52(2007):37–55. Ecography. 20(4):368–374. van Halder I, Barbaro L, Corcket E, Jactel H. 2008. Importance Stasek DJ, Bean C, Crist TO. 2008. Butterﬂy abundance and of semi-natural habitats for the conservation of butterﬂy movements among prairie patches: the roles of habitat qual- communities in landscapes dominated by pine plantations. ity, edge, and forest matrix permeability. Environ Entomol. Biodivers Conserv. 17(5):1149–1169. 37(4):897–906. van Swaay CAM, Nowicki P, Settele J, van Strien AJ. 2008. Stireman JO, Greeney HF, Dyer LA. 2009. Species richness Butterﬂy monitoring in Europe: methods, applications and and host associations of Lepidoptera-attacking Tachinidae perspectives. Biodivers Conserv. 17(14):3455–3469. in the northeast Ecuadorian Andes. J Insect Sci. 9(39): Veddeler D, Schultze C, Steffan-Dewenter I, Buchori D, 1–19. Tscharntke T. 2005. The contribution of tropical secondary forest fragments to the conservation of fruit-feeding but- Stork NE, Srivastava DS, Watt AD, Sen TB. 2003. Butterﬂy diversity and silvicultural practice in lowland rainforests of terﬂies: effects of isolation and age. Biodivers Conserv. Cameroon. Biodivers Conserv. 12(3):387–410. 14(14):3577–3592. Summerville KS, Dupont MM, Johnson AV, Krehbiel RL. Vu LV. 2009. Diversity and similarity of butterﬂy communities 2008. Spatial structure of forest Lepidopteran communi- in ﬁve different habitat types at Tam Dao National Park, ties in oak hickory forests of Indiana. Environ Entomol. Vietnam. J Zool. 277(1):15–22. 37(5):1224–1230. Watt AD, McFarlane AM. 2002. Will climate change have a Summerville KS, Metzler EH, Crist TO. 2001. Diversity of different impact on different trophic levels? Phenological Lepidoptera in Ohio forests at local and regional scales: development of winter moth Operophtera brumata and its how heterogeneous is the fauna? Ann Entomol Soc Am. host plants. Ecol Entomol. 27(2):254–256. 94(4):583–591. Weibull AC, Östman Ö. 2003. Species composition in agroe- Tumuhimbise G, Okwakol MJN, Kangwagye TN. 1998. cosystems: the effect of landscape, habitat, and farm man- Species diversity of swallowtail butterﬂies (Papilionidae: agement. Basic Appl Ecol. 4(4):349–361. Lepidoptera) in North Maramagambo Forest. Afr J Ecol. Yahner RH. 2001. Butterﬂy communities in residential landscapes 39(4):113–115. of central Pennsylvania. Northeast Nat. 8(1):113–118.
International Journal of Biodiversity Science, Ecosystem Services & Management – Taylor & Francis
Published: Dec 1, 2013
Keywords: butterfly diversity; drivers; farmlands; Lepidoptera conservation; habitats protection; East Africa
Access the full text.
Sign up today, get DeepDyve free for 14 days.