Farmland conservation in West Africa: how do hedgerow characteristics affect bird species richness?

Abstract

Capsule The value of farmland for birds at a site in Nigeria, West Africa, was highly dependent upon the habitat structure, with wider and more continuous hedges supporting more bird species.

Aim To understand the value of hedges in Nigerian farmland to bird species richness.

Methods We used line transects to survey birds and vegetation on 166 systematically selected hedgerows on fallow, cultivated and harvested fields in Jos East, Nigeria.

Results Wider hedges supported higher species richness. Plant species composition was a weak predictor of bird species richness compared to hedge width. There was no observed difference in distribution of bird richness across field types, suggesting that the value of hedgerows for biodiversity is comparable, regardless of farming method. Continuous hedgerows (hedgerows with vegetation gap < 1.5 m) held higher species richness compared with discontinuous (hedgerows with vegetation gap > 1.5 m) hedges.

Conclusion This result suggests that hedge size and structure, rather than plant species composition, is an important predictor of avian species richness in hedgerows surrounding farmland. Therefore, the value of farmland for birds in this part of Africa is highly dependent upon the habitat structure, with wider and more continuous hedges supporting more bird species.

Tropical deforestation represents the greatest single threat to global biodiversity because tropical forest may support about 70% of the planet's plant and animal species (e.g. Mayaux et al. 2005, Chomitz 2007). Although the rate of deforestation has slowed down since the 1990s, Africa loses 0.6% of its forest annually, with the greatest loss occurring in heavily forested countries (Kelatwang & Garzuglia 2006, UNEP 2008). Throughout sub-Saharan West Africa, habitat loss has been rapid through human clearance for fuel wood, grazing and conversion to intensive agriculture as a consequence of population growth (e.g. Grimmett 1987, Chomitz 2007, Wilson et al. 2007, Norris et al. 2010, Riggio et al. 2012). These practices result in a reduction in the natural habitat available for biodiversity and also a negative impact on biodiversity due to the effect of habitat fragmentation (Dunn 2004, Manu et al. 2007, Norris et al. 2010). Much effort has gone into designating protected areas in Africa with the hope of saving these areas of crucial habitat (UNEP 2008). As a result, conservation effort has been directed to protected areas with little or no attention given to the potential of the replacement anthropogenic habitats such as farmland to conserve species.

Loss of biodiversity in agricultural environments is considered a high conservation priority in Europe (e.g. Tucker & Evans 1997, Bradbury et al. 2000, Butler et al. 2009, Wilson et al. 2009, Wretenberg 2010) and the previous UK government set a public service agreement target to reverse the decline in farmland bird populations by 2020 (Robinson & Sutherland 2002, Butler et al. 2007a,b), although this has since been abolished. The USA has also developed a programme that encourages farming in a more holistic, environmentally sound manner in order to enhance biodiversity and provide habitat for wildlife (Sullivan et al. 2004, Stamps et al. 2008). However, the value of farmland habitats to bird species is rarely assessed in Africa (but see Wilson et al. 1997, Mangnall & Crowe 2003, Bolwig et al. 2006, Mulwa et al. 2012), despite much native habitat being destroyed to make way for farmland, especially in Nigeria. With this pressure on land in Africa, particularly in the light of a growing human population and many people living below the poverty line, optimizing replacement farmland habitat to promote productivity while still maintaining biodiversity is a key problem.

Natural hedgerows usually arise as a result of clearing natural continuous habitats for agricultural encroachment, industrial development and unsustainable logging practices. Hedges have been shown to be important to animal and plant species diversity in Europe (Parish et al. 1994, Sparks et al. 1996, Hinsley & Bellamy 2000, Wehling & Diekmann 2009, Pereira & Rodriquez 2010, Vickery & Arlettaz 2012). In lowland-farming landscapes in Britain, hedgerows comprise one of the most important surviving elements of semi-natural habitat for birds (Hinsley & Bellamy 2000). Agricultural field boundaries such as hedgerows in Europe, Asia and North America have beneficial impacts on adjoining cultivated land; however, many farmlands are cleared of such habitats to increase space for cultivation. Similarly, natural hedges are destroyed because many farmers view hedges as a threat to their crops since they attract birds (Jobin et al. 2001). In Great Britain, increased use of machinery, and resulting demand for large fields, resulted in removal of 50% of hedgerow stock (Robinson & Sutherland 2002).

One major variable that accounts for bird diversity in any habitat is vegetation structure, which is further enhanced by the plant species composition (Wilson et al. 1997, Hinsley & Bellamy 2000, Jobin et al. 2001, Mulwa et al. 2012). Effects of habitat fragmentation, patch size and connectivity on species diversity, abundance and distribution have also been considered in many studies (Saunders et al. 1991; Wiens 1997; Manu et al. 2007).

The aim of this study was to investigate the effects of hedgerow structure and plant species composition on bird species richness in farmlands in Jos-East, Nigeria. The objectives were to determine the effect of hedge size, hedge continuity, and field condition, number of trees in the hedge, inter-hedge distance and plant composition on bird species richness. As a starting point and given the relative unstudied nature of farmed landscape and impact of agricultural practices (e.g. cattle grazing, hedgerow removal or thinning) on bird populations in farmland habitats in this region, this paper describes the species composition and bird community structure of hedgerows in our study region and evaluates the effects of hedgerow characteristics on overall bird species richness.

METHODS

Surveys were carried out on agricultural fields with hedgerows 15 km east of Jos, north-central Nigeria (N9°52ʹ E8°58ʹ). The Jos landscape comprises of high plains with scattered rock outcrops ranging from 1220 to 1450 m asl and a number of granite hill ranges that rise to 1781 m. The vegetation is scattered bush and grass, grazed by cattle, sheep and goats, with riparian forest fragments and extensive cultivation (Hulme 2007). Commonly cultivated crops include Guinea Corn Sorghum bicolor, Millet Pennisetum glaucum, Acha Digitaria exilis, Maize Zea mays, Rice Oryza sativa, Irish potato Solanum tuberosum, Mango Mangifera indica, Cucumber Cucumis sativus and Lettuce Lactuca sativa. Mixed cropping is the general farming practice in this area where some fields have trees and/or hedgerows. Other fields are open without trees or hedgerows, while some plots are left fallow some years. On average, intensive rain-fed farming begins late April to early May and lasts until early September. Few fields are cultivated during the dry season (late September to early April). Bird and vegetation data were collected from hedgerows in three field types: fallow (n = 72), cultivated (n = 21) and harvested fields (n = 73). A hedge was considered to be on a given field type if fields on both sides of the hedge were in the same condition, i.e. both sides of the hedge were cultivated. One hundred and sixty-six hedgerows each measuring 100 m (hereafter referred to as ‘hedgerow plot’) were systematically selected (Watt 1997), with hedgerow plots spaced at intervals of at least 100 m to ensure that bird counts were reasonably independent. No two plots were sampled from the same field. The entire area covered was approximately 100 km².

Bird data

Bird surveys were carried out from May to June 2008 at the beginning of the wet season. Hedgerows were surveyed between 0600 and 1000 hours each day. Line transects (Bibby et al. 2000) were used to census birds. Bird counts were all completed by one observer (HU). During each visit, a transect was walked along the length of the hedgerow plot and the total number of birds detected by sight or sound in the hedgerow or within 15 m on the adjacent field was recorded. Each start and end point of each hedgerow plot was marked with a GPS (GPS Garmin eTrex®, version 3.10).

Hedge data

Each hedgerow plot was divided into four 25-m sections so as to obtain estimates of key features of the habitat. At each section, hedgerow height and width (horizontal spread of vegetation) were measured and recorded. Similarly foliage density index was determined using a 50 × 50 cm black and white chequered board with 10-cm subdivisions, recording the distance at which 50% of board was obscured by vegetation (Bibby et al. 2000). Within each section, a 2 × 2 m quadrat was placed centrally and ground cover (taken as amount of dry litter on the ground) and bare ground were estimated by eye to the nearest 5% (Bullock 2006). Plant species composition was determined by recording presence/absence and identified using Arbonnier (2004). The number of trees in each hedgerow plot (trees with circumference at breast height over 30 cm) was recorded. Hedgerow continuity was determined by observing whether a hedgerow plot was continuous (hedgerows with vegetation gap = 0 or < 1.5 m) or discontinuous (hedgerows with vegetation gap > 1.5 m). Distance of the nearest hedge to a focal hedgerow was estimated by eye to the nearest 5 m. The number of shrubs (bushes) and trees within each adjacent field, and the field status (cultivated, harvested and fallow) were recorded for each field.

Data analysis

R version 2.15.0 was used for statistical analyses. Means are given with ± one standard error.

For each 100 m hedge, bird species richness was calculated as the total number of species recorded. We summarized bird species according to guilds by considering the major food resource consumed (based on species food habits described in Fry et al. 1988, Borrow & Demey 2010). This resulted in six functional groups. The total number of species in each feeding guild was summed per hedge. The presence/absence of each species and guild for each hedgerow plot was determined and proportion frequency of occurrence over 166 hedgerow plots was calculated (Table 1).

Table 1. Frequency of occurrence of bird species in hedges. Frequency is the number of occurrences of a species or guild over 166 hedgerow plots; proportion is the proportion of hedgerow plots where the species or guild was recorded.

Guild	Species	Frequency	Proportion	No. of individuals recorded
Carnivore		4	0.02
	Black-shouldered Kite Elanus caeruleus	1	0.01	1
	Lizzard Buzzard Kaupifalco monogramicus	1	0.01	1
	Black Kite Milvus migrans	2	0.01	3
	Greyish Eagle-owl Bubo cinerascens	1	0.01	1
Frugivore		98	0.59
	African Grey Hornbill Tockus nasutus	5	0.03	6
	Bearded Barbet Lybius dubius	3	0.02	3
	Vieillot's Barbet Lybius vieilloti	3	0.02	6
	Speckled Mousebird Colius striatus	37	0.22	59
	Adamawa Turtle Dove Streptopelia hypopyrrha	2	0.01	3
	Black-billed Wood Dove Turtur abyssinicus	28	0.17	35
	Bruce's Green Pigeon Treron waalia	1	0.01	2
	Laughing Dove Streptopelia senegalensis	31	0.19	57
	Red-eyed Dove Streptopelia semitorquata	2	0.01	3
	Vinaceous Dove Streptopelia vinacea	7	0.04	8
	Violet Turaco Musophaga violacea	2	0.01	5
	Western Grey Plantain-eater Crinifer piscator	9	0.05	20
	Senegal Parrot Poicephalus senegalus	5	0.03	8
Granivore		149	0.90
	Crested Lark Galerida cristata	7	0.04	12
	Sun Lark Galerida modesta	7	0.04	14
	Cinnamon-breasted Rock Bunting Emberiza tahapisi	4	0.02	4
	Black-bellied Firefinch Lagonosticta rara	4	0.02	6
	Black-rumped Waxbill Estrilda troglodytes	6	0.04	23
	Bronze Mannikin Spermestes cucullata	8	0.05	27
	Lavender Waxbill Estrilda caerulescens	13	0.08	27
	Orange-cheeked Waxbill Estrilda melpoda	15	0.09	60
	Red-billed Firefinch Lagonostica senegala	32	0.19	89
	Red-cheeked Cordon-bleu Uraeginthus bengalus	89	0.54	236
	Rock Firefinch Lagonosticta sanguinodorsalis	13	0.08	20
	Yellow-fronted Canary Serinus mozambicus	1	0.01	2
	Bush Petronia Petronia dentata	1	0.01	2
	Northern Grey-headed Sparrow Passer griseus	15	0.09	38
	Black-winged Bishop Euplectes hordeaceus	19	0.11	26
	Chestnut-crowned Sparrow Weaver Plocepasser superciliosus	7	0.04	14
	Heuglin's Masked Weaver Ploceus heuglini	2	0.01	5
	Little Weaver Ploceus luteolus	21	0.13	30
	Speckled-fronted Weaver Sporopipes frontalis	21	0.13	57
	Northern Red Bishop Euplectes franciscanus	3	0.02	163
	Village Weaver Ploceus cucullatus	32	0.19	67
	Vitelline Masked Weaver Ploceus vitellinus	17	0.10	30
	Pin-tailed Whydah Vidua macroura	9	0.05	48
Insectivore		153	0.92
	African Pygmy Kingfisher Ceyx pictus	3	0.02	2
	Red-shouldered Cuckoo-Shrike Campephaga phoenicea	6	0.04	7
	Yellow-fronted Tinkerbird Pogoniulus chrysoconus	22	0.13	30
	Long-tailed Nightjar Caprimulgus climacurus	3	0.02	4
	Abyssinian Roller Coracias abyssinicus	1	0.01	2
	African Cuckoo Cuculus gularis	1	0.01	2
	Didric Cuckoo Chrysoccocyx caprius	5	0.03	5
	Jacobin's Cuckoo Oxylophus jacobinus	3	0.02	3
	Senegal Coucal Centropus senegalensis	24	0.14	32
	Grey-headed Kingfisher Halcyon leucocephala	1	0.01	1
	Yellow-billed Shrike Corvinella corvina	10	0.06	20
	Black-crowned Tchagra Tchagra senegalus	27	0.16	35
	Northern Puffback Dryoscopus gambensis	2	0.01	3
	Sulphur-breasted Bush-shrike Malaconotus sulfureopectus	8	0.05	9
	Tropical Boubou Laniarius aethipicus	1	0.01	2
	Yellow-crowned Gonolek Laniarius barbarus	46	0.28	73
	European Bee-eater Merops apiaster	3	0.02	6
	Red-throated Bee-eater Merops bullocki	9	0.05	14
	African Blue Flycatcher Elminia longicauda	1	0.01	2
	Yellow-throated Longclaw Macronyx croceus	5	0.03	10
	Northern-black Flycatcher Melaenornis edolioides	43	0.26	23
	Pale Flycatcher Melaenornis (Empidornis) pallidus	10	0.06	14
	Whinchat Saxicola rubetra	2	0.01	2
	Green wood Hoopoe Phoeniculus purpureus	3	0.02	3
	Cardinal Woodpecker Dendropicus fuscescens	1	0.01	1
	Grey Woodpecker Dendropicus goertae	6	0.04	7
	Common Wattle-eye Platysteira cyanea	1	0.01	1
	Senegal Batis Batis senegalensis	5	0.03	10
	Common Bulbul Pyconotus barbatus	91	0.55	187
	Yellow-throated Leaflove Chlorocichla flavicollis	9	0.05	19
	Purple-glossy Starling Lamprotornis purpureus	15	0.09	37
	Violet-backed Starling Cinnyricinclus leucogaster	9	0.05	11
	African Moustached warbler Melocichla mentalis	1	0.01	1
	Croaking Cisticola Cisticola natalensis	14	0.08	16
	Rock-loving Cisticola Cisticola aberrans	8	0.05	10
	Dorst's Cisticola Cisticola dorsti	3	0.02	4
	Grey-backed Camaroptera Camaroptera brachyura	51	0.31	82
	Northern Crombec Sylvietta brachyura	16	0.10	3
	Senegal Eremomela Eremomela pusilla	7	0.04	12
	Tawny-flanked Prinia Prinia subflava	51	0.31	100
	Brown Babbler Turdoides plebejus	5	0.03	22
	African Thrush Turdus pelios	52	0.31	84
	Familiar Chat Cercomela familiaris	4	0.02	5
	Cliff Chat Myrmecocichla cinnamomeiventris	6	0.04	9
	Northern Anteater Chat Myrmecocichla aethiops	3	0.02	4
	Snowy-crowned Robin-Chat Cossypha niveicapilla	22	0.13	34
	Yellow White-eye Zosterops senegalensis	7	0.04	13
	Greater Honeyguide Indicator indicator	1	0.01	1
Nectarivore		107	0.64
	Copper Sunbird Cinnyris cupreus	2	0.01	2
	Green-headed Sunbird Cyanomitra verticalis	5	0.03	7
	Scarlet-chested Sunbird Chalcomitra senegalensis	87	0.52	153
	Variable Sunbird Cinnyris venustus	57	0.34	84
Omnivore		13	0.07
	Double-spurred Francolin Francolinus bicalcaratus	10	0.06	20
	Stone Patridge Ptilopachus petrosus	3	0.02	12
	Yellow-billed Oxpecker Buphagus africanus	1	0.01	1

Shrub density for each 100 m hedgerow was calculated using MacArthur foliage density index, K:

K = ln2/D

where D is the distance from the observer to the chequered board (McArthur & McArthur 1961, Bibby et al. 2000).

To control for non-orthogonality in our model, Spearman's correlation was used to determine significant associations between habitat variables: hedge height, groundcover and foliage density index showed strong and significant correlation with hedge width (coefficient of correlation, r _s > 0.50, P < 0.05). Consequently we omitted these three predictor variables from our model while retaining hedge width (we considered hedge width to be more biologically important). We fitted Generalized Linear Models to the data using Poisson error and log link, with bird species richness as the response variable; explanatory variables incorporated in the model were hedge width, hedge continuity, number of trees on hedge, number of plant species in the hedge, distance to nearest hedge and field condition. Models for all possible parameter subsets were compared in terms of parsimony and prediction through the calculation of AIC (Burnham & Anderson 2002). These analyses were performed with the MuMIn package in R which returns all possible models in order of importance and plausibility with the lower AIC values representing better models. For the purpose of effect size, we standardized covariates (i.e. converted as mean = 0, sd = 1).

RESULTS

As a group, seed- and insect-eating birds occurred in at least 90% of surveyed hedges, with higher densities in wider and continuous hedges; frugivores and nectarivores occurred in more than 50% of surveyed hedgerow plots; carnivores (raptors) occurred in 2% of surveyed hedges (Table 1). Ninety-five species belonging to 39 families were recorded from 166 hedgerow plots, with Estrildidae, Sylviididae, Ploceidae, Columbididae, Pycnonotidae and Nectariniidae occurring in at least 50% of surveyed hedges. As species, Common Bulbul Pycnonotus barbatus, Red-cheeked Cordon-Bleu Uraeginthus angolensis and Scarlet-chested Sunbird Chalcomitra senegalensis were recorded in more than half of surveyed hedges; Variable Sunbird Cinnyris venustus, African Thrush Turdus pelios, Grey-backed Camaroptera Camaroptera brachyura and Tawny-flanked Prinia Prinia subflava were recorded in at least 30% of hedges (Table 1). Fifty-six plant species were identified in the hedgerow plots.

Bird–habitat relations

The model selection procedure showed that there were five top models (Table 2). Discontinuous hedges (negative effect) and hedge width (standardized coefficient, sc = 0.111) were included in all five best models (Fig. 1). Distance to nearest hedge (sc = 0.057) and number of trees in the hedge (sc = 0.045) were included in four and three of the five best models, respectively, and both had positive effects on species richness. The mean (± se) distance between hedges was 63.1 (± 2.8) m (n = 152). Within farms the mean number of shrub and trees were 4.35 ± 0.36 (n = 165) and 3.13 ± 0.26 (n = 165), respectively.

Table 2. Top Generalized Linear Models (GLMs) of variables predicting bird species richness on hedgerows on West African farmlands. Estimates of variables not included in the models are indicated by blank spaces. ∆ _i is difference in AIC values of each model relative to the best model. Cultivated field and continuous hedge are reference categories. The table reports all models with ∆ _i  < 5. *significant predictors in GLM.

AIC	∆ i	Explanatory variables
		Intercept	Fallow fields	Harvested fields	Hedge width* (m)	Discontinuous hedges*	No. trees on hedge	No. of plant species	Distance to nearest hedge (m)
791.447	0.000	1.724			0.062	–0.176	0.022		0.002
791.813	0.366	1.777			0.062	–0.180			0.002
792.154	0.707	1.632			0.055	–0.174	0.025	0.017	0.002
793.098	1.651	1.715			0.057	–0.179		0.012	0.002
793.348	1.900	1.809			0.065	–0.166	0.022
793.819	2.372	1.864			0.065	–0.171
794.300	2.853	1.729			0.059	–0.165	0.025	0.015
794.809	3.361	1.770	–0.019	–0.060	0.061	–0.176	0.022		0.002
795.130	3.682	1.828	–0.028	–0.067	0.061	–0.180			0.002
795.285	3.838	1.812			0.061	–0.170		0.010
795.545	4.097	1.661	0.002	–0.045	0.053	–0.175	0.025	0.017	0.002
796.048	4.601	1.883	–0.071	–0.102	0.065	–0.164	0.022
796.369	4.922	1.941	–0.080	–0.109	0.066	–0.168
Standard error		0.168	0.096	0.093	0.018	0.062	0.014	0.015	0.001

Figure 1. Relationship between bird species richness and hedge width for different hedge structures. The predicted lines are based on the best model: solid line (continuous hedge) and dashed line (discontinuous hedge).

Although hedgerow plots adjacent to cultivated fields had the highest (mean ± se) bird species richness (1.91 ± 0.09) compared with fallow (1.86 ± 0.04) and harvested fields (1.69 ± 0.06), these differences were not important in predicting bird richness.

DISCUSSION

The results presented here show that hedgerows in West Africa seem to play similar role to those in Europe. Bird species richness showed a positive relationship with hedgerow width. Parish et al. (1994) also found a similar relationship with hedge width, indicating though that taller hedges were more important for bird species richness. Similarly, larger hedgerows (greater width) contain more species, presumably because they meet the varied requirements of foraging, roosting, nesting and protection or refuge from predators (Sparks et al. 1996, Hinsley & Bellamy 2000, Whittingham & Evans 2004). In Britain, Hinsley et al. (1999) found that bird species richness and abundance showed a better relationship with hedge width than with hedge height; a narrow hedge will be generally unfavourable to most birds probably due to a combination of lack of resources and exposure to weather and predators.

Plant species composition was a weaker predictor of bird species richness in hedgerows compared to hedge width. This result contradicts the prediction that higher plant species diversity within hedges should result in greater potential food resources and hence more birds (Hinsley & Bellamy 2000). It suggests, therefore, that habitat structure may be more important and probably indicates that hedges on fields serve more to provide cover or protection for birds (Hinsley & Bellamy 2000) and nests (Sparks et al. 1996, Perkins et al. 2002) and also as connection between habitats (Fahrig & Merriam 1994). The botanic composition of a hedge may become less important if the width of the hedge is narrow. Besides, most of the species recorded here were not specialists. However, we note that in this study, the use of 2 × 2 m quadrats in four equidistant sections to sample plant species in a length of 100 m may mean that some species may have been missed. Similarly, the effect of the presence of individual species – for example, plants with a high production of fruit – might be greater than the number of species per se, so plant composition cannot be ruled out entirely as an important factor.

Bird species richness was not affected negatively by distance to the nearest hedge, possibly because the maximum hedge isolation distance was insufficient in this study to act as much of a constraint. Similarly, most of the farms contained trees or bushes that probably served as connections between the hedges, thereby further precluding any negative effect of hedge isolation (Fahrig & Merriam 1994).

Bird species richness increased with number of trees in the hedge. This result is consistent with other studies (e.g. Parish et al. 1994, Hinsley et al. 2000); increasing number of trees in hedges probably indicates an increase in niche opportunities and habitat types which affects bird abundance positively. However, it should be noted that the UK studies referred to involve much smaller total numbers of species (20–30), with the same core species likely to be found in all survey plots compared with an average richness of 8 species per plot with a total pool of 95 for this study. This implies that there must be many replacement species within guilds playing the same roles in different plots. In a review of value of hedgerows to birds in lowland farming landscape in Britain, Hinsley et al. (2000) discussed that the presence of well-grown, dead or decaying trees is beneficial to many species, providing nest-holes, foraging sites and perches. Tall hedges with many trees are probably attractive to many woodland bird species because of the resemblance of this type of structure to woodland edge. Wilson et al. (2007) reported that number of bird species, total number of birds and the diversity of birds decreased significantly with a decrease in tree density, whereas there was no similar decline on adjacent farmland habitat that had already been deforested.

Field condition, i.e. whether a field was cultivated, harvested or fallow, did not significantly influence bird species richness, suggesting that the value of hedgerows for biodiversity is comparable, regardless of farming type employed. However, hedgerows adjacent to cultivated and fallow fields had a higher number of bird species compared to harvested fields. This might reflect higher food availability and a complex habitat structure around fallow fields. Fallow fields would also be favoured especially by granivorous species, in this case weavers, that seek grains and weed seed (Bradbury et al. 2000) as well as insects because of no artificial chemical input and low disturbance. Chamberlain et al. (1999), in a study of the effects of habitat type and management on the abundance of breeding skylarks in Britain, showed that set aside (equivalent to fallow) was consistently associated with high densities of birds. The cultivated habitat had a semblance of continuous vegetation which graded into the hedge itself (see Krebs et al. 1991) to increase its spread and created habitat for more species and may have provided some level of cover from predators.

Continuity in hedgerow vegetation was important because continuous hedges held higher bird species compared with discontinuous hedges. This pattern is probably an indication that a continuous matrix provides more space, reduces predation risk and creates corridors for movement along the hedge to adjoining suitable habitats. Ninety-five percent of the bird species encountered in this study were edge species (cf. Manu 2002, Borrow & Demey 2004) and they respond differently to differences in hedge continuity or fragmentation, suggesting that detrimental edge effects have a disproportionately greater influence in small, discontinuous habitat patches (Tattersall et al. 2002). A discontinuous hedge will function less as a corridor in terms of aiding birds moving between habitats. Jobin et al. (2001) also found a negative association between the proportion of gaps in margins and bird communities. It has been suggested generally that farm vegetation structure and predation risk may influence the demography of farmland bird populations (Whittingham & Evans 2004).

CONCLUSIONS

Our results suggest that hedge size rather than plant species composition is an important predictor of avian species richness in hedgerows surrounding farmlands. Similarly, discontinuity in hedge vegetation greatly affects bird species richness. This suggests that, as in Europe, the value of farmland for birds in African farmland is highly dependent upon the habitat structure (Sparks et al. 1996, Robinson & Sutherland 1999, Hinsley & Bellamy 2000, Whittingham & Evans 2004, Wretenberg et al. 2010, Mulwa et al. 2012), with wider and more continuous hedges supporting more bird species. Therefore, sustainable conservation efforts should be aimed at promoting fields with more edge habitat, so that farmland can play a dual role in supporting biodiversity and man, particularly if the conservation goal is to maintain high local species richness (see also Mulwa et al. 2012).

However, as Africa expands and intensifies its agriculture (Wilson et al. 2007, Norris et al. 2010, Deininger & Byerlee 2011), with local farmers increasingly having access to bulldozers to clear and tractors to prepare fields, hedgerow removal, as well as shrubs and small bushes within fields, is likely to increase. Such removal is one of the drivers of biodiversity decline in European farming systems (Robinson & Sutherland 2002, Newton 2004, Donald et al. 2001, Whittingham et al. 2006, Butler et al. 2007a,b, Wilson et al. 2009, Norris et al. 2010) and so Africa may experience a similar trend in loss of biodiversity, but little of this has been documented to date. This is one of the few studies in Africa, particularly in West Africa, that has investigated aspects of agricultural practices that are perceived to cause declines in biodiversity. It is therefore important to expand research in understanding the impact of man-managed habitat to biodiversity in this region because it appears that agricultural landscape will continue to increase, especially in developing countries as world food demand is expected to more than double its present rate in the next 40 years (Tilman et al. 2002, Green et al. 2005, Godfray et al. 2010).

ACKNOWLEDGEMENTS

This work was carried out during an MSc to Hope Usieta funded by the A.P. Leventis Conservation Foundation. Afan Ajan provided assistance in the field. We thank Katherine Jones and Tatsuya Amano who advised on statistical analysis and William J. Sutherland for helpful comments and suggestions to earlier versions of this paper. Gavin Siriwardena and a second anonymous reviewer provided useful comments and suggestions. This is publication 60 from the AP Leventis ornithological Research Institute.