Recibido: October 7 el año 2016 | Aceptado: 5 de Junio 2017

DOI: 10.1111 / mec.14230

ARTÍCULO ORIGINAL

la genética del paisaje comparativos de dos murciélagos

frugívoros en un corredor biológico se someten a la
intensificación agrícola

Katherine A. Cleary1,2 | Lisette P. Waits1 | Bryan Finegan3

1
Departamento de Pesca y Ciencias de la
Vida Silvestre, Universidad de Idaho, Abstracto
Moscow, ID, EE.UU.
La intensificación agrícola en un entorno tropical plantea una nueva amenaza para la
2
Departamento de Biología,
Universidad de Clarkson, Potsdam,
capacidad de los corredores biológicos para mantener la conectividad funcional para las
Nueva York, EE.UU. especies nativas. Utilizamos un enfoque de la genética del paisaje para evaluar los
3Bosques, biodiversidad y cambio climático
impactos de la expansión de las plantaciones de piña en dos murciélagos frugívoros
Programa, CATIE, Turrialba, Cartago,
Costa generalizadas y abundantes en un corredor biológico en Costa Rica. Nuestra hipótesis
Rica
es que cuanto más grande, más Artibeus jamaicensis móvil será menos afectado por la
Correspondencia piña que el más pequeño castanea Carollia. En 2012 y 2013, tomamos muestras de 735
Katherine A. Cleary, Departamento de
Biología, Universidad de Clarkson, murciélagos en 26 parches de bosque remanente rodeados de diferentes propor-ciones
Potsdam, Nueva York, EE.UU.. E-mail: de bosques, pastizales, cultivos y piña. Se utilizó 10 loci de microsatélites para A.
katecleary98@gmail.com
jamaicensis y 16 loci de microsatélites para C. castanea para estimar la diversidad
información sobre financiación genética y el flujo de genes. Los análisis de correspondencia canónica indicar que tipo
Fundación Nacional de Ciencia, Grant /
Premio de cobertura terrestre que rodea los parches no tiene impacto en la diversidad genética
Número: 0903479, 1313824; Fulbright de
EE.UU.
de A. jamaicensis. Sin embargo, para C. castanea, ambos forestales y piña porcentuales
Programa de Estudiantes; conservación de parches que rodean explican una proporción significativa de la variación en la diversidad
los murciélagos
Internacional; Sociedad Americana de genética. Menor costo transecta anal-yses (LCTA) y G pairwise "st sugieren que para A.
mastozoólogos jamaicensis, piña es más permeable al flujo de genes de lo esperado, mientras que
como se esperaba, bosque es la cobertura de la tierra más permeable para el flujo de
genes de C . castanea. Para ambas especies, LCTA indican que el desarrollo puede
desempeñar un papel en la inhibición del flujo de genes. El presente estudio responde a
la llamada para la investigación genética paisaje se centró en la tierra paisajes de
tropicales y agrícolas,

PALABRAS CLAVE
intensificación de la agricultura, los murciélagos, corredor biológico, genética paisaje,
método de transectos de menor costo, paisajes agrícolas tropicales

1 | INTRODUCCIÓN
corredores biológicos son una estrategia de conservación ampliamente
utilizada en los paisajes modificados por el hombre en todo el mundo
(Gilbert-Norton, Wilson, Ste-Vens, y Barba, 2010). El objetivo de la
mayoría de los corredores es proteger la conectividad func-cional, o la
capacidad de las especies nativas para moverse entre parches de
hábitat remanente, y para sobrevivir y reproducirse con éxito incluso en
zonas aisladas (Belisle, 2005; Kadoya, 2009). En muchas regiones tem-
Perate, los estudios han encontrado que los corredores biológicos
existentes
han logrado este objetivo y están facilitando con éxito el movimiento
y dando como resultado el flujo de genes para una variedad de
especies a través de diversos paisajes de la tierra (Gilbert-Norton et
al, 2010;. Mech y Hallett, 2001; Sharma et al, 2013).. Sin embargo,
poco se sabe acerca de si biológicos Corri-dors en los trópicos son
un éxito similar (Brodie et al., 2015). A pesar del hecho de que los
trópicos contienen más del 50% de la biodiversidad del mundo
(Mace, Masundire, y Baillie, 2005) y los paisajes tropicales tienen
tasas desproporcionadamente altas de conversión de los bosques y
la pérdida de especies (Hansen et al., 2013), hasta la fecha no han
sido muy
 |
Molecular Ecology. 2017; 26: 4603-4.617. wileyonlinelibrary.com/journal/mec © 2017 John Wiley & Sons Ltd 4603
4604 | CLEARY ET AL.
20 km o menos (Meyer, Kalko, y Kerth, 2009; Ripperger, Tschapka,
Kalko, Rodríguez-Herrera, y Mayer, 2013) y han proporcionado evi-
algunas pruebas empíricas de la eficacia corredor en las regiones
dencia que esta estructura genética está relacionada con la cantidad de
tropicales (pero véase Gilbert-Norton et al, 2010;. Hidalgo-Mihart,
hábitat adecuado en la matriz agrícola (Ripperger et al., 2013). Si bien
Contreras-Moreno, Jesús de la Cruz, y Juárez-López, 2017;
no hay estudios hasta la fecha han probado explícitamente el impacto de
Khazan, 2014).
la agricultura intensiva planta-ción sobre la estructura genética de los
En los últimos años, la viabilidad de corredores biológicos en
murciélagos neotropicales, pruebas
regiones tropicales ha sido amenazado por una nueva tendencia
uso de la tierra: la rápida expansión de la agricultura intensiva de
plantaciones (Brannstrom, 2009; Per-Fecto y Vandermeer, 2008;
máquina de afeitar y otros, 2015).. A través de los trópicos globales,
la matriz agrícola entre parches de bosque remanente está en la
transición de una mezcla diversa de pastos y cultivos de los
mercados locales a los monocultivos intensivos de cultivos de
exportación como el aceite de palma, soja y piña (Arvor, Meirelles,
Dubreuil, Begue, y Shimabukuro, 2012;. Fagan et al, 2013; Wilcove
y Koh, 2010), y esta tendencia se prevé que continúe (Harvey,
Alpizar, Chacon, y Madrigal, 2005; Wright, Lago, y Dolman, 2012).
Este tipo de agricultura simplifica y homogeniza la matriz agrícola,
reduce la cobertura total del árbol (Hobinger, Schindler, Seaman,
Wrbka, y Weissenhofer, 2012; . Máquina de afeitar et al, 2015) y
aumenta insumos agroquímicos (Harvey et al, 2005;.. Tscharntke et
al 2012). Los pocos estudios existentes de movimiento de los
animales y el flujo de genes a través de intensivo espectáculo
agricultura planta-ción que tiene bajo potencial para proporcionar
conectividad hábitat: Vaughan, Ramírez, Herrera, y Guries (2007)
no encontraron movimiento de la radio de cuello de dos o tres
dedos perezosos a través de plantaciones de piña, y Blair y Melnick
(2012) mostraron que las plantaciones de palma de aceite
comerciales representaban una barrera para el flujo de genes entre
poblaciones de monos ardilla.
Los murciélagos son un grupo focal excelente para probar el impacto
de la agricultura intensiva en la plantación de corredores biológicos,
especialmente en el Neotrópico. En esta región, los murciélagos son
abundantes y generalizadas y sirven como dispersores de semillas y
polinizadores primaria para cientos de especies de plantas nativas
(Fleming, Geiselman, y Kress, 2009; Lobova, Geiselman,
& Mori, 2009). Debido a que los murciélagos son volant, a menudo se
supone que tienen un alto potencial de dispersión y mantendrá la
conectividad funcional independientemente de la pérdida de hábitat, la
fragmentación y el cambio de uso del suelo, pero no se sabe si esto es
cierto en los paisajes que experimentan la intensificación agrícola
(Razgour et al., 2014;. Struebig et al, 2011).
A pesar de que la población y el paisaje genética enfoques
proporcionan un marco útil para evaluar la conectividad funcional, muy
pocos estudios han puesto en práctica un enfoque genético para evaluar
la conectividad para los murciélagos neotropicales. De éstos, dos
encontraron ningún efecto detectable de la pérdida y fragmentación del
hábitat: murciélagos frugívoros en paisajes agrícolas culturales
fragmentados en Argentina, Paraguay y Costa Rica mostraron la
diversidad genética similar en fragmentos de bosque y bosque continuo,
y ninguna estructura genética significativa en las escalas de 20- 120
kilometros (McCul-loch et al, 2013;. Ripperger, Tschapka, Kalko,
Rodríguez-Herrera, y Mayer, 2014). Por el contrario, otros dos estudios
han demostrado que las poblaciones de iso-RELAClONADAS de
murciélagos frugívoros en los paisajes modificados por el hombre-puede
desarrollar la estructura genética significativa incluso a escalas multa de
 Vitousek, y Strain, 1990 ; Laksa, 1997). Más de la mitad de la dieta de C.
castanea se compone de las especies de Piper-aceae tolerantes a la
a partir de estudios no genéticos indica que este tipo de agricultura
sombra de esta familia (Thies y Kalko, 2004). A diferencia de,
puede impedir el movimiento (Cormier, 2014; Freudmann, Mollik,
Tschapka, y Schulze, 2015; Phommexay, Satasook, Bates, Pearch,
y Bumrungsri, 2011) y alterar la composición global de encaje
(Cleary, espera , y Finegan, 2016;. Freudmann et al, 2015).
En este estudio, utilizamos un enfoque de la genética del paisaje
para evaluar los efectos del cambio de uso del suelo y la intensificación
de la agricultura en un corredor biológico de Costa Rica en la
conectividad funcional para dos murciélagos abundantes y
generalizadas. La historia del uso del suelo en el área de estudio refleja
un patrón común en los trópicos: la deforestación inicial para establecer
pastos y cultivos mercado local creado un complejo paisaje de bosque
remanente incrustado en una matriz agrícola diversa, y en los últimos
años, esta matriz diversa está siendo rápidamente sustituido con
monocultivos de piña a gran escala (Fagan et al, 2013;.. Morse et al,
2009). Nuestro estudio es uno de los únicos ejemplos de pruebas de la
eficacia de un corredor biológico utilizando un enfoque genético (Christie
y Knowles, 2015) y favorece la investigación corredor biológico mediante
la evaluación de la conectividad funcional para múltiples especies
simultáneamente (Beier, Spencer, Baldwin, y McRae, 2011;. Brodie et al,
2015). Además, este estudio aborda varias áreas reconocidas como
prioridades de investigación en genética del paisaje. En primer lugar, la
mayoría de los estudios de genética del paisaje hasta la fecha se han
centrado principalmente en especies de clima templado (Cleary, Waits, y
Sanfiorenzo, 2015; Manel y Holderegger, 2013; Storfer, Murphy, Lanza,
Holderegger, y Waits, 2010) a pesar de que tropical regiones son puntos
calientes de biodiversidad (Myers, Mittermeier, Mittermeier, da Fonseca,
y Kent, 2000). En segundo lugar, los estudios de genética paisaje
comparativos de múltiples especies en el mismo paisaje siguen siendo
escasos, pero pueden proporcionar información valiosa sobre
respuestas comunes a las características del paisaje (Balkenhol,
Cushman, Storfer, y Waits, 2015; Stor-fer et al., 2010). En tercer lugar,
aunque las evaluaciones actuales de uso de tierras a nivel mundial
indican que más del 40% de la superficie libre de hielo del planeta se ha
convertido a la agricultura, los paisajes agrícolas están insuficientemente
representadas en la genética del paisaje estudios (Ellis, Klein Goldewijk,
Siebert, Lightman, y Ramankutty, 2010; Storfer et al., 2010).
Nos centramos en dos especies de murciélagos frugívoros con
movilidad diferencial: El palo de castaño de cola corta, (Carollia
castanea) una pequeña frugivore con baja movilidad, y el murciélago de
la fruta de Jamaica (Artibeus jamaicensis), una frugivore móvil más
grande y más (Meyer & Kalko, 2008). Esperamos grado de movilidad a
estar directamente relacionado con el impacto del cambio de uso del
suelo en la conectividad funcional para estas especies, ya que estudios
previos han demostrado que las especies de murciélagos menos
móviles como C. castanea, con áreas de distribución pequeñas, anchas
alas cortas y menores proporciones de el tamaño del cuerpo a la zona
de ala, son más vulnerables a la fragmentación. (Burns y Broders, 2014;
Meyer, Fründ, € Lizano, y Kalko, 2008), y muestran estructura superior
genética en paisajes fragmentados (Meyer et al, 2009; Struebig et al .,
2011). También esperamos que la dieta influye en la forma en que estas
especies se ven afectadas por el cambio del uso del suelo. Carollia
castanea es un especialista en plantas de la familia Piperaceae, que
contiene tanto intolerante a la sombra de especies de sucesión
temprana y especies-finales de sucesión tolerantes a la sombra que se
encuentran en el sotobosque de bosque primario (Den-lento, Schultz,
CLEARY ET AL.
A. jamaicensis es un generalista sabe que se alimentan de más de
100 SPE-cies de plantas, pero su dieta está dominada por los higos
(Ficus; Handley, Wil-hijo, y Gardner, 1991; López y Vaughan, 2007;
Ortega y Castro- Arellano, 2001). higueras producen grandes
cosechas de fruta, pero maduran asyn-chronously, la creación de
recursos alimenticios distribuidos irregularmente que A. ja-
maicensis viaja largas distancias para localizar (Morrison, 1978).
Sobre la base de estos rasgos de historia de vida, predecimos que
C. castanea mostrará mayor interrupción en la conectividad funcional a
través del paisaje de San Juan-La Selva (SJLS) que A. jamaicensis.
Como resultado, el más pequeño, menos vagile C. castanea exhibirá
inferior diversidad genética y superior estructura genética que el más
móvil, generalista A. ja-maicensis. También calculamos estructura
genética separado por sexo para ambas especies y la hipótesis de que
como la mayoría de mamíferos exhiben dispersión de sexo masculino
sesgada (Greenwood, 1980), la estructura será más alto para las
mujeres en ambas especies. Para evaluar el impacto de los diferentes
tipos de cobertura del suelo sobre la diversidad genética, probamos si el
porcentaje de cobertura de la tierra alrededor de los sitios de muestreo
explica la variación en la diversidad genética. Predecimos que el
porcentaje de bosques y la piña que rodea a un sitio explicará una
proporción significativa de la variación de la diversidad genética y que
estos efectos será más fuerte para C. castanea que para A. jamaicensis.
Para evaluar el efecto de diferentes tipos de cobertura de la tierra sobre
el flujo de genes, calculamos diferenciación genética en general y la
distancia genética pairwise para cada especie y la hipótesis de que C.
castanea mostrará mayor diferenciación general y más pares de sitios
que se desvían de panmixia de A. jamaicensis. Por último, para
identificar el tipo de cobertura terrestre más permeable al flujo de genes
de cada especie, usamos un enfoque transecta menos costo y de
máxima probabilidad efectos pop-ulación (MLPE) modelos introducido
por Van Strien, Keller, y Holderegger (2012). Prevemos que tanto C.
castanea y A. ja-maicensis,
Figura 1 Mapa de la zona de estudio y
sitios de muestreo. Los sitios de muestreo
están etiquetados para facilitar la
interpretación de los resultados. El mapa
inserto muestra el contexto más amplio
del área de estudio, incluyendo toda la
extensión del corredor biológico San
Juan-La Selva y las áreas protegidas
| 4605
2 | MATERIALES Y MÉTODOS
2.1 | Sitio de estudio y muestreo de campo
Este estudio se realizó en el SJLS corredor biológico y áreas sur-
redondeo (Figura 1; 10 ° 220-10 ° 430N, 83 ° 600-84 ° 200W). La
región SJLS encuentra en la zona Holdridge vida de los bosques
tropicales húmedas (Holdridge, Grenke, Haheway, Liang, y Tosi,
1975), a una altura de 0-400 m. temperatura media anual es 25 ° C
y precipitación media anual es de aproximadamente 4,000 mm
(Sesnie, Finegan, Gessler, y Ramos, 2009), con una estación seca
generalmente dura desde enero hasta principios de abril. Los suelos
son ácidos (pH ~ 4,5) y el apoyo de un bosque domi-nados por
Pentaclethra macroloba y palmas (Sesnie et al., 2009).
Durante los últimos 50 años, la región SJLS ha experimentado un
rápido cambio de uso de la tierra. Debido a la tala para la agricultura de
subsistencia y ganaderas pastos, cubierta de bosque primario se redujo
de 75,6% en 1960 a 48,5% en 1996 (Schelhas y Sánchez-Azofeifa,
2006). Para proteger la cubierta forestal restante, Ley Forestal de 1996
de Costa Rica prohibió la tala de bosques primarios (Morse et al., 2009).
En 2001, se estableció el corredor biológico 246, SJLS 608 hectáreas,
con el objetivo de proteger la conectividad del hábitat entre la Reserva
Biológica Indio Maíz en Nicaragua y el Parque Nacional Braulio Carrillo
en Costa Rica, al tiempo que permite la producción agrícola (Villate,
Canet- Desanti, Chassot, y Monge-Arias, 2009). plantaciones de piña de
manejo intensivo llegaron a la región SJLS en la década de 1990,
extendiéndose rápidamente y reemplazar extensa, pastos degradados
en la matriz (Fagan et al, 2013;. Schelhas & Sanchez-Azofeifa, 2006). La
propagación de la piña ha llevado a un aumento de puestos de trabajo
en la región SJLS, que ha impulsado el crecimiento de población y un
expan-sión asociada de las áreas urbanas y carreteras (Schelhas y
Sánchez-Azofeifa, 2006; máquina de afeitar y otros, 2015).. El paisaje
actual se caracteriza por una mezcla
regionales que se conecta el corredor
[Color de la figura se pueden ver en
wileyonlinelibrary.com]
4606 | CLEARY ET AL.
2001). Los estudios que informan abundancia absoluta de ambas
especies en esta región muestran que la abundancia es
de edad madura y bosque secundario (43,8% del paisaje; aquí
generalmente más alta en C. castanea (Cleary, Waits, y Finegan,
después referido como “bosque”), pastos (39,7%), la piña y el suelo
2016; López y Vaughan, 2007; Rex, Kelm, Wiesner, Kunz, y Voigt,
desnudo en la preparación para la siembra de piña (6,5%), cultivos
2008 ), lo que sugiere que C. castanea puede tener un tamaño
en pequeña escala y las plantaciones de árboles (5,9%), y el
popu-mento efectivo mayor que A. jamaicensis en la zona de
desarrollo, que incluye las áreas urbanas y carreteras (~ 1,7%)
estudio. vágiles especie es un generalista de alimentación conocido
(Shaver et al., 2015).
para dispersar las semillas de al menos 91 plantas neotropicales
Para capturar la gama de posibles respuestas de la comunidad
(López y Vaughan, 2007), y se ha observado que volar hasta 8 km
de murciélagos frugívoros en la región SJLS a esta historia del
entre las perchas de día y zonas de alimentación (Morrison, 1978).
cambio de uso del suelo, se seleccionaron dos especies focales en
tiempo de generación de A. jamaicensis es aproxi madamente-1
los extremos opuestos del espectro de tamaño y vagilidad. Carollia
año (Ortega y Castro Arellano, 2001). Los estudios que informan
castanea es una pequeña frugivore que pesa 11-16 g y tiene un
abundancia absoluta de ambas especies en esta región muestran
tamaño rango de hogar promedio de <7 ha (Bonaccorso et al,
que la abundancia es generalmente más alta en C. castanea
2006;. Reid, 2009). Este murciélago se produce en los bosques
(Cleary, Waits, y Finegan, 2016; López y Vaughan, 2007; Rex,
húmedos de tierras bajas desde Honduras a Bolivia, y es un
Kelm, Wiesner, Kunz, y Voigt, 2008 ), lo que sugiere que C.
especialista en alimentación con una fuerte dependencia de las
castanea puede tener un tamaño popu-mento efectivo mayor que A.
plantas de la familia Piperaceae, especialmente las que se
jamaicensis en la zona de estudio. vágiles especie es un generalista
encuentran en el sotobosque de bosque maduro (Bonaccorso,
de alimentación conocido para dispersar las semillas de al menos
1979; Thies y Kalko, 2004). No hay información directa es
91 plantas neotropicales (López y Vaughan, 2007), y se ha
disponible de C. castanea tiempo de generación, pero el C.
observado que volar hasta 8 km entre las perchas de día y zonas de
perspicillata estrechamente relacionado tenga un tiempo de
alimentación (Morrison, 1978). tiempo de generación de A.
generación de aproximadamente 1 año (Cloutier y Thomas, 1992).
jamaicensis es aproxi madamente-1 año (Ortega y Castro Arellano,
jamaicensis Artibeus es una gran frugivore pesaje 29-51 g,
2001). Los estudios que informan abundancia absoluta de ambas
encontrado en los bosques de tierras bajas húmedas de México a
especies en esta región muestran que la abundancia es
Ecuador (Reid, 2009). Esta especie grande, vágiles es un
generalmente más alta en C. castanea (Cleary, Waits, y Finegan,
generalista alimentación conocido para dispersar las semillas de al
2016; López y Vaughan, 2007; Rex, Kelm, Wiesner, Kunz, y Voigt,
menos 91 plantas neotropicales (López y Vaughan, 2007), y se ha
2008 ), lo que sugiere que C. castanea puede tener un tamaño
observado que volar hasta 8 km entre las perchas de día y zonas de
popu-mento efectivo mayor que A. jamaicensis en la zona de
alimentación (Morrison, 1978). tiempo de generación de A.
estudio. jamaicensis es aproxi madamente-1 año (Ortega y Castro
jamaicensis es aproxi madamente-1 año (Ortega y Castro Arellano,
Arellano, 2001). Los estudios que informan abundancia absoluta de
2001). Los estudios que informan abundancia absoluta de ambas
ambas especies en esta región muestran que la abundancia es
especies en esta región muestran que la abundancia es
generalmente más alta en C. castanea (Cleary, Waits, y Finegan,
generalmente más alta en C. castanea (Cleary, Waits, y Finegan,
2016; López y Vaughan, 2007; Rex, Kelm, Wiesner, Kunz, y Voigt,
2016; López y Vaughan, 2007; Rex, Kelm, Wiesner, Kunz, y Voigt,
2008 ), lo que sugiere que C. castanea puede tener un tamaño
2008 ), lo que sugiere que C. castanea puede tener un tamaño
popu-mento efectivo mayor que A. jamaicensis en la zona de
popu-mento efectivo mayor que A. jamaicensis en la zona de
estudio. jamaicensis es aproxi madamente-1 año (Ortega y Castro
estudio. encontrado en los bosques húmedos de tierras bajas de
Arellano, 2001). Los estudios que informan abundancia absoluta de
México a Ecuador (Reid, 2009). Esta especie grande, vágiles es un
ambas especies en esta región muestran que la abundancia es
generalista alimentación conocido para dispersar las semillas de al
generalmente más alta en C. castanea (Cleary, Waits, y Finegan,
menos 91 plantas neotropicales (López y Vaughan, 2007), y se ha
2016; López y Vaughan, 2007; Rex, Kelm, Wiesner, Kunz, y Voigt,
observado que volar hasta 8 km entre las perchas de día y zonas de
2008 ), lo que sugiere que C. castanea puede tener un tamaño
alimentación (Morrison, 1978). tiempo de generación de A.
popu-mento efectivo mayor que A. jamaicensis en la zona de
jamaicensis es aproxi madamente-1 año (Ortega y Castro Arellano,
estudio.
2001). Los estudios que informan abundancia absoluta de ambas
especies en esta región muestran que la abundancia es Para seleccionar los sitios de muestreo, se utilizaron mapas de

generalmente más alta en C. castanea (Cleary, Waits, y Finegan, cobertura terrestre a partir de 1986, 1996, 2001 y 2011 para

2016; López y Vaughan, 2007; Rex, Kelm, Wiesner, Kunz, y Voigt, identificar los parches de bosque de edad madura dentro de la

2008 ), lo que sugiere que C. castanea puede tener un tamaño frontera del corredor SJLS y en la región más fragmentada a lo

popu-mento efectivo mayor que A. jamaicensis en la zona de largo de la frontera occidental de la SJLS. Se utilizaron los

estudio. encontrado en los bosques húmedos de tierras bajas de siguientes criterios para la selección de parche: 15-60 ha de

México a Ecuador (Reid, 2009). Esta especie grande, vágiles es un tamaño, separadas por al menos 1,5 km, aproximadamente el

generalista alimentación conocido para dispersar las semillas de al mismo grado de aislamiento de otro bosque para últimos 25 años y

menos 91 plantas neotropicales (López y Vaughan, 2007), y se ha rodeado de una gama de porcentajes de los tipos de cobertura de

observado que volar hasta 8 km entre las perchas de día y zonas de cuatro tierra pertinentes para intensificación de la agricultura:

alimentación (Morrison, 1978). tiempo de generación de A. bosque, pasto, piña y desarrollo (Figura 1). Probamos el conjunto

jamaicensis es aproxi madamente-1 año (Ortega y Castro Arellano, final de 26 sitios 3-6 veces de enero a mayo de 2012 y 2013. Col-
cionado una muestra de tejido de 0,5 cm de diámetro de la
uropatagio de cada C. castanea y A. jamaicensis capturado en
redes de niebla y muestras almacenadas en tampón de lisis.
2.2 | genotipificación de microsatélites y
estadísticas de resumen
Se extrajo ADN genómico usando un kit QIAGEN Tissue DNeasy
modificados para uso con tampón de lisis. Las PCR se realizaron
utilizando el kit Qiagen múltiple con conjuntos multiplexados de 4-10
cebadores directos marcados (condiciones de la PCR en la Tabla S1).
individuos Artibeus jamaicensis
fueron genotipados en 10 loci de microsatélites, de los cuales cinco
eran específicos de especie (AjA151, AjA40, AjA47, AjA74, AjA80;
Ortega, Maldonado, Arita, Wilkinson, y Fleischer, 2002) y cinco
fueron transversal amplificaron a partir de A. literatus (AL579 ,
AL700, AL821, AL850, AL854; McCulloch
& Stevens, 2011). Carollia individuos castanea se genotipo en 16 loci de
microsatélites, de los cuales 10 eran específicos de la especie (CC7,
CC10, CC13, CC18, CC19, CC23, CC25, CC26, CC27, CC29; Cleary,
Waits,
& Hohenlohe, 2016), cinco fueron transversal amplificó a partir Carollia
Brevi-cauda (AAGG1, AAGG7, AAGG98, AAGG117, AAGG119;
Bardele-ben, Campbell, Lara, y Moore, 2007), y uno era transversal
amplificó a partir de A. literatus (AL700; McCulloch y Stevens, 2011).
Alelos fue evaluado utilizandoGenemapper3.7 (Applied Biosystems). Para
comprobar si hay errores, el 50% de las muestras se realizaron dos
veces. tasas de errores de genotipado se calcularon utilizando
elRConGenR paquete (Lonsinger y Waits, 2015). Todos los loci se
ensayaron para determinar significativo desequilibrio de ligamiento y-
desviaciones, de Hardy-Weinberg usandoGENEPOP 4,2 (Raymond
& Rousset, 1995).
2.3 | estructura genética de la población
Para cada especie en cada sitio de muestreo, se identificaron alelos
privados y se calculó la heterocigosidad esperada utilizando
imparcial Genalex6.5 (Peakall y Smouse, 2012) y la riqueza de alelos
usando HP-Rare (Kali-nowski, 2005). Para comparar la
diferenciación entre las especies en general, se calculó G global "en
stGenalex6.5 utilizando sólo sitios muestreados para ambas especies
(Figura 1, Tabla 1). A continuación, para cada especie se calculó la
distancia genética entre pares todos los sitios utilizando G "en
stGenalex6.5. Para evaluar si machos y hembras dentro de las
especies exhiben estructura de la población di-ferente, global y G
dos a dos "c fueron también calcu-lated para hombres y mujeres por
separado, utilizando sólo los sitios con un mínimo de cinco
individuos del sexo respectiva (Tabla S2). Signifi-cance de pairwise
G "st se ensayó mediante el cálculo de las probabilidades de todos
los valores con 999 permutaciones.
2.4 | Influencia del paisaje en la
diversidad genética
Para probar si la diversidad genética en los sitios de muestreo se ha
INFLU-mentado por la ocupación del suelo circundante, lo primero
que utilizó el mapa de cobertura de la tierra en 2011 ARCGIS10.1 para
crear buffers circulares centradas en cada sitio, en tres escalas
prevé que sea relevante para C. castanea y A. jamacen-sis: 500,
1000 y 2000 m. La escala más pequeña de 500 m CORRE-ponde al
eje largo promedio de rango de hogar de C. castanea (Bonaccorso
et al., 2006), mientras que la mayor escala de 2,000 m tiene el
potencial de contener el rango de hogar promedio de A. jamacensis
( Ber-nardo y Fenton, 2003; Morrison, 1978). A continuación, se
utilizóFRAGSTATS(V.4, McGarigal, Cushman, y Ene, 2012) para
calcular el porcentaje de bosques, pastos, piña y desarrollo dentro
de las memorias intermedias. Se utilizó el análisis de
correspondencia canónica (CCA) para probar si estas cuatro
variables de cobertura terrestre explican una proporción significativa
de la varia-ción de la diversidad genética de cada especie. CCA
permite la IDENTIFICACIÓN-ción de las relaciones entre los
patrones de diversidad genética y múltiples variables ambientales, y
pruebas que las variables con-homenaje de manera significativa a
la explicación de la variación de la diversidad genética
CLEARY ET AL. | 4607

TABLA 1 Tamaño de la muestra (N), año muestreado (años), no sesgada heterocigosidad esperada (UHE), la riqueza de alelos (AR) y el
número de alelos privados (PA) de Artibeus jamaicensis y Castanea Carollia en todos los sitios de muestreo

castanea Carollia jamaicensis Artibeus

Arkansa Pensilv Arkansa Pensilva
norte Año ühe s ania norte Año ühe s nia

BT4 7 2013 0,695 4.63 0 17 2013 0,742 5.08 0

bt5 8 2013 0,766 5.19 0 22 2013 0.763 5.06 0

BT9 20 2013 0.698 4.52 1 12 2013 0.714 4.96 1

BTLL 11 2012 0.732 5.25 0 5 2012 0.687 4.3 0

ER1 22 2012 0.740 4.99 2 27 2012 0.751 5.07 1

ER14 14 2012/2013 0.714 4.92 1 17 2012/2013 0.760 4.99 0

ER15 12 2012 0.732 4.78 1 13 2012 0.747 4.99 0

ER19 16 2012/2013 0.701 4.63 1 10 2012/2013 0.726 4.98 0

ER20 13 2012 0.728 5.06 0 12 2012 0.732 4.93 0

ER23 9 2013 0.719 4.83 0 23 2013 0.754 5.05 1

ER24 20 2013 0.742 4.95 1 18 2013 0.739 4.99 0

ER27 21 2013 0.728 4.98 0 7 2013 0.710 4.64 0

ER36 11 2013 0.692 4.77 3 24 2013 0.747 4.97 0

ER40 10 2013 0.717 4.84 0 8 2013 0.793 5.14 0

P1 9 2012 0.726 4.97 0 9 2012 0.765 4.97 0

P10 18 2013 0.726 5.05 5 11 2013 0.761 4.98 1

P11 11 2013 0.721 5.06 0 10 2013 0.738 5.1 0

P12 0 2013 – – – 27 2013 0.752 4.97 0

P2 8 2013 0.729 5.05 0 12 2013 0.716 4.79 0

P3 11 2012/2013 0.713 4.62 2 29 2012/2013 0.750 4.92 0

P6 2 2013 – – – 12 2013 0.734 4.88 0

P9 13 2013 0.711 4.71 0 24 2013 0.704 4.67 0

PD 26 2012 0.698 4.7 1 0 2012 – – –

PPA 24 2012/2013 0.722 4.92 1 16 2012/2013 0.758 4.85 0

SV1 15 2013 0.730 4.9 0 21 2013 0.726 4.87 0

SV2 20 2013 0.722 4.81 0 0 2013 – – –

Avg/total 0.721 4.88 19 0.740 4.92 4

Sites with fewer than five individuals sampled were not used in analyses for that species and are indicated by a “–”.

(Angers, Magnan, Plante, & Bernatchez, 1999; Storfer et al., 2007).
Three metrics of genetic diversity were used as the response vari-
able in separate CCAs: allelic richness, unbiased expected
heterozy-gosity and observed heterozygosity. Response matrices
were expressed as relative allelic richness or relative allele
frequencies per locus, per site (Angers et al., 1999). Permutation
tests were used to determine the significance of the overall model
(including all four land cover variables), and each individual land
cover variable (Oksa-nen et al., 2015). To assess directionality of
relationships, we exam-ined bivariate scatterplots of each land cover
variable and each metric of genetic diversity. All analyses were
performed using pack-age VEGAN in the R statistical package
(Oksanen et al., 2015, R Core Development Team).
2.5 | Influence of landscape on genetic distance
To evaluate how land use change in the SJLS region has affected
gene flow for the focal bat species, we used a least-cost transect
analysis (LCTA) approach developed by Van Strien et al. (2012). One of
the most persistent challenges in landscape genetics is objectively
estimating resistance distances between sampling sites (Spear,
Balkenhol, Fortin, McRae, & Scribner, 2010). A common approach is to
first create a resistance surface (RS) where each land cover type is
assigned a resistance value based on expert opinion, model opti-
mization or ecological field data (Hall & Beissinger, 2014; Spear et al.,
2010), and then define resistance distance as the sum of the resistance
values in a straight-line or least-cost path (LCP) between sites. These
methods lack objectivity because they require the researcher to make a
priori decisions about how different landscape features will affect gene
flow for a given species. A potential solu-tion is offered by transect-based
approaches, which quantify the percentage of each land cover type
within different buffer widths around straight lines connecting sites and
correlate these values with genetic distances (e.g., Goldberg & Waits,
2010; Pavlacky, Goldizen, Prentis, Nicholls, & Lowe, 2009). Rather than
assigning resistance values a priori to different land cover types,
transect-based
4608 | CLEARY ET AL.
approaches allow researchers to directly test whether each land
cover type facilitates or inhibits gene flow (Spear et al., 2010). How-
ever, transect approaches based on straight-line distances assume
that dispersal occurs along a straight-line path, an unlikely hypothe-
sis for many systems. The LCTA approach retains the objectivity of
transect-based approaches, and addresses the problem of the
straight-line path by replacing straight-line distances between sites
with LCPs and then calculating the percentage of each land cover
type inside transect widths around the LCP. These percentages,
combined with total transect length, can be used as explanatory
vari-ables in models of the effect of the landscape on gene flow.
The first step in the LCTA is to create a set of binary RSs where
each land cover type of interest has the opportunity to be consid-ered
optimal dispersal habitat. All analyses were conducted in ARCGIS 10.1
(ESRI) unless otherwise noted. For each land cover type (forest, pasture,
pineapple, development), we created three RSs using the 2011 land
cover map with 5-m resolution. In these three RSs, the land cover type
being considered optimal dispersal habitat was given the lowest
resistance value of 1, while all other land cover types were given a
progressively higher resistance value (23, 26, 29; Table S3). Separate
LCP analyses were conducted for each of the 12 resulting RSs, using
the ARCGIS landscape genetics toolbox (Ethering-ton, 2011). To represent
a scenario of isolation by Euclidean distance (IBD), we created a set of
LCPs that were simply straight-line dis-tances between all pairs of
sampling sites. LCPs for the 12 RSs and the IBD scenario were buffered
at three different transect widths (100, 400 and 800 m), as bats respond
to land use change at multi-ple scales (Cleary, Waits, & Finegan, 2016).
The smallest scale of 100 m was chosen because given that home range
size of C. cas-tanea is approximately 7 ha, it is unlikely that land use
inside a tran-sect narrower than 100 m would affect long-term gene flow
of this species, much less of the larger A. jamaicensis. The largest scale
of 800 m was the maximum scale at which heterogeneity in land cover
proportions between LCPs was retained; at larger scales, the land cover
proportions in most LCPs approached average proportions across the
entire study area. For each transect, we used Geospatial Modelling
Environment (v 0.7.4.0, Beyer, 2015) to calculate the total length of each
LCP and the percentage of each of the four land cover types within the
three transect widths. The final product of these analyses was 36 LCP
data sets (4 land cover types 9 3 resis-tance values 9 3 transect widths),
plus three LCP data sets for the IBD scenario (IBD least-cost path 9 3
transect widths) (Table S3). Each LCP data set contained five
explanatory variables for each pair of sites: the total length of the LCP,
and the percentage of pasture, forest, pineapple and development within
the transect.
Next, we modelled the relationship between these explanatory
variables and the response variable of genetic distance using the MLPE
method of Clarke, Rothery, and Raybould (2002). Unlike other linear
mixed-effect models, MLPE models incorporate a covariance structure
specifically tailored for the dependency between pairwise values in
genetic distance data sets (Van Strien et al., 2012). For each species,
we applied the MLPE approach to model the relation-ship between
explanatory variables and pairwise genetic distance (G″
st). Preliminary tests indicated the land cover variables showed high
levels of collinearity, which would make interpretation of multivari-
able models difficult. Therefore, we restricted our model sets to con-
tain only univariate models. We used Akaike’s information criterion
(AIC), which has been shown to be effective for REML mixed model
selection (Gurka, 2006), to select the top model from each of the 39
LCP data sets and then compared these models and reported the
top models separated from all other models by a minimum of seven
AIC, a cut-off which Richards (2005) showed allows for a 95% prob-
ability that the best model is retained. We also reported the next
model after the seven AIC difference cut-off for comparison.
3 | RESULTS
3.1 | Microsatellite genotyping and summary
statistics
We collected sufficient genetic samples of C. castanea and A. ja-
maicensis from 24 of the 26 sampling sites (Table 1). After removing
samples with incomplete or inconsistent genotypes, we assembled
complete, reliable genotypes for 349 C. castanea individuals and
386 A. jamaicensis individuals. Due to this screening, error rates in
our final genotypes were very low: for C. castanea, the false allele
rate was 0.38% and the allelic dropout rate was 0.58%, and for A.
ja-maicensis, the false allele and allelic dropout rates were 0.24%
and 0.69%, respectively. Tests for deviations from Hardy–Weinberg
equi-librium indicated that for C. castanea loci, 24 of 384 (6.25%)
Hardy– Weinberg tests were significant at a = .05, with no patterns
across loci or sites. No pair of loci was consistently in linkage
disequilibrium across more than three sites. For A. jamaicensis loci,
two of 240 (0.83%) Hardy–Weinberg tests were significant at a =
.05, with no patterns across loci or sites. No pair of loci was
consistently in link-age disequilibrium across more than four sites.
All loci were retained for the full analysis.
3.2 | Population genetic structure
Genetic diversity was high across the study area for both species.
Average unbiased heterozygosity across all sites was 0.721 for C.
castanea and 0.740 for A. jamaicensis, and average allelic richness
was 4.88 for C. castanea and 4.92 for A. jamaicensis (Table 1). The
total number of private alleles across all sites differed dramatically
between species: C. castanea had 19 while A. jamaicensis had only
four.
Overall, levels of genetic differentiation across the study area
were low but significant for both C. castanea (G″st = 0.025, p = .012)
and A. jamacensis (G″st = 0.022, p = .044). At the level of the
sampling site, significant pairwise G″st values were found for 47
pairs (17.1%) of C. castanea populations and 44 pairs (15.8%) of A.
jamaicensis populations (Table 2). For C. castanea, more than half
of the significant pairwise G″st values were from two patches: BT.9
and ER.36 (Fig. S1a). These two patches were very isolated from
continuous forest and surrounded by considerable amounts of
CLEARY ET AL. | 4609

T A B L E 2 Population differentiation in Carollia castanea and Artibeus jamaicensis

C. castanea A. jamaicensis
All Males Females All Males Females

Global G″st (p value) 0.025 (0.012)* 0.035 (0.006)* 0.014 (0.166) 0.022 (0.044)* 0.016 (0.197) 0.021 (0.139)
Percentage of population pairs with significant G″st 17.1 14.3 3.5 15.8 10.4 15.0

For each species, we report global G″st values and the percentage of pairs of populations with significant G″st values at a = 0.05, total and separately
by sex. The values in italics are the p values. Global G”st values significant at a = .05 are indicated by a "*".

T A B L E 3 Summary statistics for canonical correspondence analyses of genetic diversity (AR, uHe, Ho) for Carollia castanea, constrained
by the percentage of forest, pasture, pineapple and development in a circular buffer of radii 500, 1,000 and 2,000 m

Carollia castanea
Axis 1 Axis 2 p value model p value variable Explained variation (%)

500 m
AR (0.003) (0.002) .011* 26.14

% Forest 0.859 0.072 .003*

% Development 0.068 0.842 .048*
uHe (0.001) (0.0003) .214 18.96
– –
Ho (0.002) (0.001) .068 22.06

% Pineapple 0.026 0.778 .029*

1,000 m
AR (0.003) (0.001) .038* 23.27

% Forest 0.660 0.540 .022*

% Pineapple 0.674 0.445 .029*
uHe (0.0005) (0.0003) .852 12.54
– –
Ho (0.002) (0.0001) .346 15.66
– –
2,000 m
AR (0.003) (0.001) .031* 23.72

% Forest 0.612 0.624 .034*

% Pineapple 0.778 0.398 .023*
uHe (0.0004) (0.0003) .792 13.33
– –
Ho (0.002) (0.0009) .678 15.22
– –

For each model, the eigenvalues of the first two axes are in parentheses. The values in italics are the p values. Only land cover variables significant at a
= .05 are shown; for each significant land cover variable, p values are shown and the axis loadings for the first two canonical axes are given. p values
significant at a = .05 are indicated with a “*”. No significant relationships at any scale were found for A. jamaicensis (Table S5).

pineapple (Figure 1). In comparison, no patches separated only by
continuous forest showed significant G″st. For A. jamaicensis, more
than half of the significant pairwise G″st values were from two dif-
ferent patches: ER.19 and ER.23 (Fig. S1b). These two patches are
close together, but are not more isolated than average from continu-
ous forest (Figure 1).
Estimates of global and pairwise G″st separately by sex indicated
higher structure in C. castanea males than females: global G″st for
females was not significant (G″st = 0.014, p = .166) and significant
G″st values were only found for 6 of 171 (3.5%) pairs of populations,
while global G″st for males was highly significant (G″st = 0.035, p =
.006) and significant G″st values were found for 31 of 210 (14.3%)
pairs of populations. In contrast, there was no significant dif-ference
in genetic structure between sexes in A. jamaicensis: global G″st
was not significant for males (G″st = 0.016, p = .197) or females
(G″st = 0.021, p = .139), and significant G″st values were found for
males in 16 of 153 (10.4%) pairs of populations, and for females in
23 of 153 (15.0%) pairs of populations.
4610 | CLEARY ET AL.

3.3 | Influence of landscape on genetic diversity
For C. castanea, the CCA revealed that the land cover variables
explained a low but significant proportion of the variation in allelic
richness at all three scales, but did not explain a significant propor-
tion of the variation in observed heterozygosity or unbiased
expected heterozygosity at any scale (Table 3). We focus on results
from the analysis at 1,000 m, as the results at the 500- and 2,000-m
scales show similar patterns and explain similar amounts of vari-
ability in allelic richness (26.1%, 23.2% and 23.7%, respectively;
Table 3). At the 1,000-m scale, the significant land cover variables
were percentage land in forest (p = .022) and percentage land in
pineapple (p = .029). Bivariate scatterplots indicated that the rela-
tionship between forest and allelic richness is positive, while the
relationship between pineapple and allelic richness is negative
(Table S4). The first canonical axis explained 58.57% of the con-
strained variability in the allelic richness data, and the biplot scores
indicate this axis was strongly negatively correlated with forest and
positively correlated with pineapple (Table 3). The second canonical
axis explained 21.4% of the constrained variability, and was domi-
nated by a negative correlation with forest and a positive correla-tion
with pineapple and development. Although it was not significant at
the 1,000-m or 2,000-m scale, development showed a marginally
significant relationship with allelic richness at the 500-m scale (p =
.046; Table 3), and bivariate scatterplots indicated this relationship is
negative (Table S4). For A. jamaicensis, the CCA showed that the
selected land cover variables did not explain a sig-nificant proportion
of the variation in any of the genetic diversity metrics, at any scale
(Table S5).
3.4 | Influence of landscape on genetic distance
The model representing the IBD scenario was never selected as top
model for either species, suggesting that genetic differentiation in
these species cannot be determined by distance alone and that land
cover may play a role in determining dispersal pathways for C.
castanea and A. jamaicensis. The top two models explaining genetic
differentiation in C. castanea differed from each other by only one
AIC, but were different from all other models by more than seven
AIC (Table 4, Richards, 2005). Both top models
identified forest as the most likely dispersal habitat for this species,
with a transect width of 100 m and with the intermediate (26) and
highest (29) resistance values, respectively. For A. jamaicensis, the
first top model was different from the second top model by more
than seven AIC (Table 4), and identified pineapple as most likely
dispersal habitat, with the highest resistance value (29) and a tran-
sect width of 100 m. The AIC values of the second and third mod-els
did not differ by more than seven AIC, and these models also
identified pineapple as the most likely dispersal habitat. For both
species, the land cover variable in both top models was develop-
ment, which had a negative impact on gene flow (positive correla-
tion with G″st) in all models. The top 10 models for each species are
available in Table S6.
4 | DISCUSSION
Our findings suggest that despite agricultural intensification, the
SJLS biological corridor is effectively facilitating functional
connectivity for A. jamaicensis, but historical habitat loss and
fragmentation and the recent expansion of pineapple have begun to
disrupt movement and resulting gene flow across the SJLS region
for C. castanea. This study represents one of the few empirical tests
of the efficacy of a biologi-cal corridor in the tropics, and one of the
few tests of connectivity of biological corridors in any ecosystem to
implement a genetic approach (Christie & Knowles, 2015; Gilbert-
Norton et al., 2010; Waits, Cushman, & Spear, 2016). The different
responses of these two frugivorous bats to land use change show
the importance of considering multiple species when designing and
managing biological corridors. This study demonstrates that
landscape genetics approaches can serve as valuable tools for
understanding how to restore and protect functional connectivity in
tropical agricultural landscapes, which are underrepresented in
landscape genetics (Cleary et al., 2015; Storfer et al., 2010).
4.1 | Population genetic structure in an intensifying
agricultural landscape
Estimates of genetic diversity and genetic differentiation provide
some support for our hypothesis that C. castanea has been more

T A B L E 4 Results from maximum-likelihood population effects models with genetic distance (G″st) as the response variable and results
from the least-cost transect analyses as the explanatory variables
Species Genetic distance Dispersal habitat Resistance level Transect width (m) Variables AIC DAIC

Carollia castanea G″st Forest 26 100 Development ( ) 1298.84 0.00

Forest 29 100 Development ( ) 1298.29 0.55
Forest 29 400 Development ( ) 1288.46 10.38
Artibeus jamaicensis G″st Pineapple 29 100 Development ( ) 1165.73 0.00
Pineapple 23 100 Development ( ) 1157.80 7.93
Pineapple 29 800 Development ( ) 1157.43 8.30
For the least-cost transect analyses, resistance surfaces were created with each land use type (forest, pasture, pineapple and development) as the dis-
persal habitat, with three levels of resistance (23, 26, 29) and three transect widths (100, 400, 800 m). Shown are the top three models (based on AIC)
for each species, and the variables in each model with direction of correlation. Additional models for each species are shown in Table S6.
CLEARY ET AL.
strongly impacted by recent land use change than A. jamaicensis,
although the differences in structure between the two species are
marginal. Genetic diversity across the SJLS landscape is slightly
lower in C. castanea than in A. jamaicensis, and C. castanea also
shows mar-ginally higher genetic structure and two more pairs of
sampling sites with significant genetic differentiation than A.
jamaicensis. In addi-tion, C. castanea populations have almost five
times as many private alleles as A. jamaicensis populations, which
may indicate a reduction in gene flow and genetic drift causing loss
of shared alleles between C. castanea populations (Slatkin, 1985).
The few previous studies of population genetic structure in A. ja-
maicensis have sampled fewer populations and at smaller scales
than this study, but also found low or no significant genetic
differentia-tion across populations (Ortega, Maldonado, Wilkinson,
Arita, & Fleischer, 2003; Vazquez-Domınguez, Mendoza-Martınez,
Orozco-Lugo, & Cuaron, 2013). McCulloch et al. (2013) found that a
closely related species of only slightly larger size, Artibeus literatus,
showed no significant genetic structure across distances of more
than 120 km in a region of Paraguay characterized by high forest
frag-mentation. This suggests that A. jamaicensis may be able to
maintain genetic connectivity over a scale much larger than the
SJLS biologi-cal corridor, providing the landscape contains sufficient
land use types that can provide roosting and foraging habitat.
Genetic responses of C. castanea to land use change are not
well documented, but Ripperger et al. (2014) found no evidence of
genetic structure in C. castanea in the same landscape as this
study. However, Ripperger et al. (2014) sampled on a finer spatial
scale (~300 km2) and only in the southeastern part of the SJLS, a
region characterized by lower forest fragmentation and less
pineapple than the broader extent (~900 km2) over which our study
sampled. The fact that C. castanea shows significant genetic
structure in our study but not in Ripperger et al. (2014) may be due
to our sampling design capturing the response of C. castanea
populations over a larger scale and to higher levels of forest
fragmentation and agricultural intensifi-cation. In addition, Ripperger
et al. (2014) used mitochondrial DNA markers, which have a slower
mutation rate than nuclear DNA microsatellite markers, and are thus
less capable of detecting responses to recent landscape change
(Cleary et al., 2015; Wang, 2010).
Comparing genetic structure between these two types of mark-ers in
C. castanea in the SJLS region leads to an interesting insight about sex-
biased dispersal in this species: higher genetic structure in
microsatellites than in mitochondrial DNA suggests female-biased
dispersal. This is contrary to our prediction that C. castanea, like the
majority of mammals (Greenwood, 1980), would show male-biased
dispersal patterns. Although comparing data between two studies can be
unreliable, additional support for female-biased dispersal is provided by
data within each study. Separate analyses of genetic structure evaluating
G″st for male and female C. castanea in this study showed significant
global genetic structure in males, but not in females. Ripperger et al.
(2014) also found significant genetic struc-ture in C. castanea males, but
not in females. Female-biased dispersal is rare in mammals
(Greenwood, 1980), but evidence for it has been
| 4611
documented in several other Neotropical bat species (Nagy, Gunther,€
Knornschild,€ & Mayer, 2013; Nagy, Heckel, Voigt, & Mayer, 2007),
including two species in the same family as C. castanea
(Phyllostomus hastatus; McCracken & Bradbury, 1981; Uroderma
bilobatum, Sagot, Phillips, Baker, & Stevens, 2016).
4.2 | Influence of landscape on functional
connectivity
Landscape genetic analyses indicated that historical loss of forest
cover and the more recent expansion of pineapple plantations into
the SJLS biological corridor are driving moderate loss of functional
connectivity for C. castanea, but no effects were detected for A. ja-
maicensis. Tests of pairwise population differentiation in C. castanea
revealed that 30 of the 47 total significant pairwise G″st values for C.
castanea populations were accounted for by the two most iso-lated
sampling sites, located outside the limits of the SJLS biological
corridor and surrounded by the highest percentage of pineapple
(ER.36 and BT.9, Figure 1, Fig. S1). This pattern indicates reduced
gene flow between these sites and the rest of the study area. The
negative impact of pineapple plantations on C. castanea is also
shown by the CCA and the bivariate scatterplots, which found a sig-
nificant and negative relationship between allelic richness and the
percentage of pineapple surrounding sampling sites.
Although significant relationships between C. castanea allelic
richness and land cover were detected at all three scales, a similar
response in C. castanea heterozygosity was only detected at the
500-m scale (Table 4). Allelic richness is known to respond more
quickly to reduced gene flow than heterozygosity, as rare alleles
may be quickly lost from an isolated population while heterozygosity
can remain high even with low allelic richness (Allendorf & Luikart,
2007). Therefore, the decline in allelic richness but not in heterozy-
gosity may be attributed to the fact that pineapple has only been in
the SJLS region for approximately 20 years, or 20 generations for C.
castanea (Cloutier & Thomas, 1992). Empirical and simulation
landscape genetics studies have shown that there is a temporal lag
between initial habitat loss and fragmentation and the appearance of
detectable responses in genetic diversity and differentiation (Ander-
son et al., 2010; Landguth et al., 2010; Spear & Storfer, 2008). The
significant and negative relationship between allelic richness and
pineapple may be the first detectable response of C. castanea to
agricultural intensification; if functional connectivity is not protected,
then in the future a similar pattern is likely to appear more strongly in
heterozygosity as well.
While pineapple plantations may impede gene flow for C. cas-tanea,
forest cover, as expected, appears to facilitate it. None of the sites
separated only by continuous forest showed significant pairwise G″st
values, and the CCA and bivariate scatterplots found a signifi-cant and
positive relationship between forest and allelic richness at all three
scales. In addition, the LCTA identified forest cover as the preferred
dispersal habitat for this species. This dependence on for-est cover may
be because approximately half the diet of C. castanea is comprised of
fruits of shade-tolerant Piperaceae species
4612 | CLEARY ET AL.
preferentially occurring in interior forest understory (Laksa, 1997; Thies &
Kalko, 2004). Roosting habits of C. castanea are also depen-dent on
forest; this species is not known to roost in man-made structures or
agriculture, but rather appears to roost exclusively in cavities among tree
roots in riparian zones (Bonaccorso et al., 2006; Thies, Kalko, &
Schnitzler, 2006). Previous studies have found that C. castanea
abundance is lower in modified agricultural areas than in interior forest
(Cortes-Delgado & Perez-Torres, 2011; Medina, Har-vey, Merlo, Vilchez,
& Hernandez, 2007), and is especially low in intensive agriculture like
full-sun coffee and oil palm plantations (Freudmann et al., 2015; Numa,
Verdu, & Sanchez-Palomino, 2005).
The negative impact of pineapple on functional connectivity for
C. castanea is likely directly related to the low amount of forest
cover present in this land use. A recent analysis of spatial character-
istics of land cover in the SJLS region found that pineapple has
lower total tree cover than all other agricultural land cover types
except banana (Shaver et al., 2015). This analysis also showed that
the conversion of pasture to pineapple increases the distance
between remnant forest patches, because small-scale forest ele-
ments are usually eliminated to maximize production efficiency
(Sha-ver et al., 2015). Dispersing C. castanea individuals likely
prefer to move through land cover types with more forest elements,
spaced more closely together, such as pasture with scattered trees
and live fences.
The larger, more mobile A. jamaicensis showed a less marked
response to landscape composition. In CCA, the four land cover
vari-ables did not explain a significant proportion of the variation in
genetic diversity, and results from the LCTA suggest that this
species can likely move through pineapple plantations during
dispersal. Addi-tional evidence that pineapple does not represent an
impediment to A. jamaicensis is that although 44 of the 276 pairwise
tests for popu-lation differentiation were significant, significant tests
were not more evident between sites surrounded by a high
percentage of pineapple than they were for sites surrounded by
primarily forest (Fig. S2). The majority (65%) of significant tests were
actually between sites ER.19 and ER.23 (Figure 1), but none of the
tested landscape variables could explain this pattern. There may be
undetected past or current processes that have reduced connectivity
of these sites, such as a lack of concentrated food resources leading
to low levels of migra-tion into the site.
The diet of A. jamaicensis likely contributes to its ability to main-tain
functional connectivity across the SJLS region. This species is a
generalist, consuming fruits of more than 100 different plant species
(Lopez & Vaughan, 2007; Ortega & Castro-Arellano, 2001), including
many early-successional species (e.g., Cecropia) species used for live
fences (e.g., Spondias) and local market crops commonly planted in the
SJLS region, such as avocado, mango and cashew (Ortega & Cas-tro-
Arellano, 2001). This means that A. jamaicensis is not dependent on
foraging in interior forest, and can also exploit food resources found in
degraded forest, live fences and local market crops. This diet flexibility
may be why previous studies have found A. jamaicen-sis at similar or
higher abundances in modified agricultural areas as in forest (Estrada,
Coates-Estrada, & Meritt, 1993; Faria, 2006), even
when the agricultural matrix is comprised of intensive plantation
agriculture like full-sun coffee and oil palm (Freudmann et al., 2015;
Numa et al., 2005).
Although A. jamaicensis feeds on a wide variety of plant species,
two studies suggest almost 50% of its diet is figs (Ficus; Handley et
al., 1991; Lopez & Vaughan, 2007), and during this study, we cap-
tured many individuals close to fruiting Ficus. This strong reliance on
figs may help explain why contrary to our hypotheses, results of the
LCTA analysis suggest that A. jamaicensis is likely able to move
through large-scale, monoculture pineapple plantations, despite the
fact that it is not known to feed on pineapple (Ortega & Castro-Are-
llano, 2001) and pineapple plantations contain few small-scale forest
elements that could offer food resources or roosting habitat (Hand-
ley et al., 1991; Shaver et al., 2015). Unlike Piper plants, which pro-
duce a relatively small number of fruits over a prolonged period, fig
trees reproduce with a “big bang” strategy, where all fruits on a tree
ripen within a short amount of time (Thies & Kalko, 2004). The high
wing loading and aspect ratio of A. jamaicensis’ wings make it well
adapted for flying long distances to track these patchily distributed
food resources (Norberg & Rayner, 1987), so extensive, open
pineapple plantations may serve as highly traversable flyways for A.
jamaicensis to reach fruiting fig trees in nearby forest patches. In
fact, radiotelemetry studies of A. jamaicensis have found that this
species will travel 13 times farther to reach fruiting fig trees in a
fragmented agricultural landscape than it does in an intact forested
landscape (Morrison, 1978). Although ecologically, foraging and dis-
persal are different processes, previous studies have found that bat
species which move long distances to track patchily distributed food
resources also tend to have higher gene flow in fragmented land-
scapes (McCulloch et al., 2013; Meyer et al., 2009).
One similarity in how the landscape affects functional connectiv-ity
for C. castanea and A. jamaicensis was predicted by our hypothe-ses:
the LCTA identified infrastructure development as an impediment to
gene flow for both species. In C. castanea, this finding was supported by
the CCA, where development emerged as a mar-ginally significant land
cover variable explaining variation in allelic richness, albeit only at the
smallest scale of 500 m. However, the CCA showed no similar pattern
for A. jamaicensis, so it remains uncertain whether this species is
affected by development in the same way as the smaller, less vagile C.
castanea. One possible expla-nation is that long-term dispersal patterns
of A. jamaicensis are affected by development only at larger scales than
those evaluated in the CCA, such as the whole study area scale of the
LCTA.
In our study area, the majority of land categorized as develop-ment is
represented by roads (Figure 1). As bats are volant, it is often assumed
that roads will not impede their movement. However, several studies
have documented negative effects of roads on tem-perate zone bats,
including interruption of foraging movements (Zurcher, Sparks, &
Bennett, 2010), avoidance of roads and sur-rounding areas due to noise
and light pollution (Berthinussen & Altringham, 2012) and direct
mortalities due to collision with vehi-cles (Lesinski, Sikora, & Olzewski,
2010), and these effects are likely similar for Neotropical bats. The
expansion of pineapple would likely
CLEARY ET AL.
exacerbate the threat posed by development; profitable pineapple
production requires efficient market access, which is facilitated by
the construction of new roads in and around plantations (Pitacuar-
Meneses, 2010).
4.3 | Implications for conservation
The dual goals of the SJLS biological corridor are to protect func-tional
connectivity for native species while allowing for sustainable agricultural
production (Villate et al., 2009). This study shows that the current
landscape of the SJLS region is succeeding in this goal for A.
jamaicensis populations, and mostly succeeding for C. castanea
populations within the limits of the corridor. However, loss of forest cover
and expansion of pineapple outside of the SJLS boundaries where
pineapple is more prevalent are beginning to strain the ability of C.
castanea to maintain functional connectivity. Other bat species of the
same or smaller size and similar life history traits as C. cas-tanea may
also be experiencing loss of functional connectivity in these areas. For
example, the two other species of Carollia found in the SJLS, C.
perspicillata and C. brevicauda, have similar home range sizes and diets
as C. castanea (Bonaccorso et al., 2006; Lopez & Vaughan, 2007).
Together, these three species comprise more than 30% of bats captured
in mist nets in other studies in the region (Rex et al., 2008; Cleary, Waits,
& Finegan, 2016). Loss of functional con-nectivity for this considerable
segment of local bat assemblages will mean a corresponding reduction
in gene flow of the many plant spe-cies dependent on these species for
seed dispersal, and some studies have already documented genetic
structure in bat-dispersed canopy and understory tree species in
fragmented agricultural landscapes (Aldrich, Hamrick, Chavarriaga, &
Kochert, 1998; Hanson, Brunsfeld, Finegan, & Waits, 2008; Sanfiorenzo,
2016). These Carollia species primarily disperse pioneer plant species
(e.g., Cecropia, Piperaceae; Lopez & Vaughan, 2007), so reduced
movement of these species could disproportionately affect genetic
diversity of plant communi-ties in regenerating forests. The importance of
bat-mediated seed dispersal for both pioneer and mature forest plant
species increases in fragmented agricultural landscapes like the SJLS,
where many large- and medium-sized mammalian seed dispersers such
as tapir (Tapirus bardii) and paca (Cuniculus paca) may be locally
extirpated due to hunting pressure and habitat loss (Melo, Rodrıguez-
Herrera, Chazdon, Medellın, & Ceballos, 2009; Wright et al., 2007).
Our findings suggest that increasing forest cover inside pineap-
ple plantations would help restore and protect functional connec-
tivity and resulting gene flow for C. castanea and similar bat species
in the SJLS. Interior forest is probably the habitat most favourable to
C. castanea, but forest edge may also provide valu-able food
resources as it often contains high densities of early-suc-cessional
plants on which all three Carollia species feed (Lopez & Vaughan,
2007). Therefore, even small, linear forest elements with a high
edge-to-interior ratio could provide movement pathways for C.
castanea, and retaining this type of forest cover within planta-tions
requires minimal changes in cultivation methods. Profit losses
resulting from these changes could be offset by adding applicable
| 4613
financial incentives to Costa Rica’s existing 1996 Forest Law; for
example, the minimum number of hectares of forest needed to
receive financial compensation for forest protection could be
reduced so that even very small, linear forest elements would qual-
ify (Shaver et al., 2015). Reducing total distance between remnant
forest patches by retaining small forest elements would also
increase functional connectivity for other species in the SJLS bio-
logical corridor for which forest cover is positively related to gene
flow, including the Chestnut-backed antbird (Myrmeciza exsul, Wolt-
mann, Kreiser, & Sherry, 2012), Thomas’s fruit-eating bat (Derma-
nura watsoni, Ripperger et al., 2013) and the litter frog (Craugastor
bransfordii, Nowakowski, DeWoody, Fagan, Willoughby, & Donnelly,
2015). Another recommendation stemming from our findings would
be to minimize the growth of road networks, specifically paved roads
which are likely to have a higher impact than the more com-mon dirt
roads due to increased traffic and higher levels of noise and light
pollution.
If implemented, these management actions will also protect func-
tional connectivity for the mutualistic plant species which depend on
C. castanea and other small, frugivorous bat species for seed
disper-sal. In addition, future work could apply the methodologies
used here to other tropical bat species with even more limited
mobility than C. castanea, such as understory insectivorous species,
which show stronger negative responses to agricultural
intensification than most frugivorous bat species (Cleary, Waits, &
Finegan, 2016). Over-all, our findings provide insight into the impact
of agricultural intensi-fication on the effectiveness of biological
corridors in human-modified tropical landscapes, and emphasize the
importance of an integrated view of connectivity for forest organisms
when designing and managing biological corridors.
ACKNOWLEDGEMENTS
We acknowledge support from the National Science Foundation
under IGERT grant Award No. 0903479 and CNH grant Award No.
1313824, the Fulbright U.S. Student Program, Bat Conservation
International, and American Society of Mammalogists. Any opinions,
findings and conclusions or recommendations expressed in this
material are those of the author(s) and do not necessarily reflect the
views of the National Science Foundation. We thank H. Lara and D.
Schelly for invaluable field assistance, Jennifer Adams for laboratory
assistance, the landowners of Sarapiquı for access and SINAC and
CONAGEBIO for permits.
DATAACCESSIBILITY
Previously published C.castanea microsatellite primer sequences are
archived in NCBI GenBank, Accession nos.: KX060618, KX060619,
KX060620, KX060621, KX060622, KX060623, KX060624, KX060625,
KX060626, KX060627, KX060628, KX060629, KX060630, KX060631.
Empirical microsatellite genotypes used for analyses and coordinates of
sampling sites are available from the Dryad Digital Repository:
https://doi.org/10.5061/dryad.nv92q.
4614 | CLEARY ET AL.
AUTHORCONTRIBUTIONS
All authors collaborated to design the study. K.A.C. secured funding,
obtained permits, collected field data, performed genetic and statistical
analyses and wrote the manuscript. L.P.W. secured funding, oversaw
field data collection and laboratory and genetic data analyses and edi-ted
the manuscript. B.F. secured funding, oversaw field data collection, gave
conceptual advice on analyses and edited the manuscript.
REFERENCES
Aldrich, P. R., Hamrick, J. L., Chavarriaga, P., & Kochert, G. (1998).
Microsatellite analysis of demographic genetic structure in frag-
mented populations of the tropical tree Symphonia globulifera.
Molec-ular Ecology, 7, 933–944.
Allendorf, F., & Luikart, G. (2007). Conservation and the genetics of
popula-tions (664 pp.). Malden, MA: Wiley-Blackwell.
Anderson, C. D., Epperson, B. K., Fortin, M. J., Holderegger, R., James,
P. M. A., Rosenberg, M. S., . . . Spear, S. (2010). Considering spatial
and temporal scale in landscape-genetic studies of gene flow.
Molecular Ecology, 19, 3565–3575.
Angers, B., Magnan, P., Plante, M., & Bernatchez, L. (1999). Canonical
correspondence analysis for estimating spatial and environmental
effects on microsatellite gene diversity in brook charr (Salvelinus
fonti-nalis). Molecular Ecology, 8, 1043–1053.
Arvor, D., Meirelles, M., Dubreuil, V., Begue, A., & Shimabukuro, Y. E.
(2012). Analyzing the agricultural transition in Mato Grosso, Brazil,
using satellite-derived indices. Applied Geography, 32, 702–713.
Balkenhol, N., Cushman, S., Storfer, A., & Waits, L. (2015). Landscape
genet-ics, concepts, methods, applications (288 pp.). West Sussex,
UK: Wiley-Blackwell.
Bardeleben, C. P., Campbell, M., Lara, A., & Moore, R. L. (2007). Isola-
tion of polymorphic tetranucleotide microsatellite markers for the silky
short-tailed bat Carollia brevicauda. Molecular Ecology Notes, 7, 63–
65.
Beier, P., Spencer, W., Baldwin, R. F., & Mcrae, B. H. (2011). Toward
best practices for developing regional connectivity maps.
Conservation Biol-ogy, 48, 667–676.
Belisle, M. (2005). Measuring landscape connectivity: The challenge of
behavioral landscape ecology. Ecology, 86, 1988–1995.
Bernard, E., & Fenton, M. B. (2003). Bat mobility and roosts in a frag-
mented landscape in central. Biological Conservation, 134, 332–343.
Berthinussen, A., & Altringham, J. (2012). The effect of a major road on
bat activity and diversity. Journal of Applied Ecology, 49, 82–89.
Beyer, H. L. (2015). Geospatial Modelling Environment (Version 0.7.4.0)
(software). http://www.spatialecology.com/gme.
Blair, M. E., & Melnick, D. J. (2012). Scale-dependent effects of a hetero-
geneous landscape on genetic differentiation in the Central American
squirrel monkey (Saimiri oerstedii). PLoS One, 7, e43027. http://doi.
org/10.1371/journal.pone.0043027
Bonaccorso, F. J. (1979). Foraging and reproductive ecology in a
Panama-nian bat community. Bulletin of Florida Museum of
Biological Science, 24, 359–408.
Bonaccorso, F. J., Winkelmann, J. R., Shin, D., Agrawal, C. I., Aslami,
N., Bonney, C., . . . Kunz, T. H. (2006). Evidence for exploitative
competi-tion, comparative foraging behavior and roosting ecology of
short-tailed fruit bats (Phyllostomidae). Biotropica, 39, 249–256.
Brannstrom, C. (2009). South America’s neoliberal agricultural frontiers:
Places of environmental sacrifice or conservation opportunity?
Ambio, 38, 141–149.
Brodie, J. F., Giordano, A. J., Dickson, B., Hebblewhite, M., Bernard, H.,
Mohd-Azlan, J., . . . Anderson, J. (2015). Evaluating multispecies
land-scape connectivity in a threatened tropical mammal community.
Con-servation Biology, 29, 122–132.
Burns, L. E., & Broders, H. G. (2014). Correlates of dispersal extent pre-
dict the degree of population genetic structuring in bats.
Conservation Genetics, 15, 1371–1379.
Christie, M. R., & Knowles, L. L. (2015). Habitat corridors facilitate
genetic resilience irrespective of species dispersal abilities or
population sizes. Evolutionary Applications, 8, 454–463.
Clarke, R. T., Rothery, P., & Raybould, A. F. (2002). Confidence limits
for regression relationships between distance matrices: Estimating
gene flow with distance. Journal of Agricultural, Biological and
Environmental Statistics, 7, 361–372.
Cleary, K. A., Waits, L., & Finegan, B. (2016). Agricultural intensification
alters bat assemblage composition and abundance in a dynamic
Neotropical landscape. Biotropica, 48, 667–676.
Cleary, K. A., Waits, L. P., & Hohenlohe, P. (2016). Development and char-
acterization of fourteen novel microsatellite markers for the chestnut short-
tailed fruit bat (Carollia castanea), and cross-amplification to related
species. PeerJ, 4, e2465. https://doi.org/10.7717/peerj.2465
Cleary, K., Waits, L., & Sanfiorenzo, A. (2015). Genetics as a tool for bio-
diversity conservation, examples from Central America. In F. Huett-
man (Ed.), Central American biodiversity—Conservation, ecology,
and a sustainable future (pp. 573–602). New York, NY: Springer.
Cloutier, D., & Thomas, D. (1992). Carollia perspicillata. Mammalian
Spe-cies, 417, 1–9.
Cormier, A. C. A. (2014). Species diversity and activity of insectivorous
bats in three habitats in La Virgen de Sarapiquı, Costa Rica. Revista
De Biologia Tropical, 62, 939–946.
Cortes-Delgado, N., & Perez-Torres, J. (2011). Habitat edge context and
the distribution of phyllostomid bats in the Andean forest and
anthropogenic matrix in the Central Andes of Colombia. Biodiversity
and Conservation, 20, 987–999.
Denslow, J. S., Schultz, J. C., Vitousek, P. M., & Strain, B. R. (1990).
Growth responses of tropical shrubs to treefall gap environments.
Ecology, 71, 165–179.
Ellis, E. C., Klein Goldewijk, K., Siebert, S., Lightman, D., & Ramankutty,
N. (2010). Anthropogenic transformation of the biomes, 1700 to
2000. Global Ecology and Biogeography, 19, 589–606.
Estrada, A., Coates-Estrada, R., & Meritt, D. (1993). Bat species
richness and abundance in tropical rain forest fragments and in
agricultural habitats at Los Tuxtlas, Mexico. Ecography, 16, 309–318.
Etherington, T. R. (2011). Python based GIS tools for landscape
genetics: Visualising genetic relatedness and measuring landscape
connectivity. Methods in Ecology and Evolution, 21, 52–55.
Fagan, M. E., DeFries, R. S., Sesnie, S. E., Arroyo, J. P., Walker, W.,
Soto, C., . . . Sanchun, A. (2013). Land cover dynamics following a
defor-estation ban in northern Costa Rica. Environmental Research
Letters, 8, 034017. http://doi.org/10.1088/1748-9326/8/3/034017
Faria, D. (2006). Phyllostomid bats of a fragmented landscape in the
north-eastern Atlantic forest, Brazil. Journal of Tropical Ecology, 22,
531–542.
Fleming, T. H., Geiselman, C., & Kress, W. J. (2009). The evolution of
bat pollination: A phylogenetic perspective. Annals of Botany, 104,
1017– 1043.
Freudmann, A., Mollik, P., Tschapka, M., & Schulze, C. H. (2015).
Impacts of oil palm agriculture on phyllostomid bat assemblages.
Biodiversity and Conservation, 24, 3583–3599.
Gilbert-Norton, L., Wilson, R., Stevens, J. R., & Beard, K. H. (2010). A
meta-analytic review of corridor effectiveness. Conservation Biology,
24, 660–668.
Goldberg, C. S., & Waits, L. P. (2010). Comparative landscape genetics
of two pond-breeding amphibian species in a highly modified agricul-
tural landscape. Molecular Ecology, 19, 3650–3663.
Greenwood, P. J. (1980). Mating systems, philopatry and dispersal in
birds and mammals. Animal Behavior, 28, 1140–1162.
Gurka, M. J. (2006). Selecting the best linear mixed model under REML.
The American Statistician, 60, 19–26.
CLEARY ET AL.
Hale, M. L., Burg, T. M., & Steeves, T. E. (2012). Sampling for micro-
satel-lite-based population genetic studies, 25–30 individuals per
popula-tion is enough to accurately estimate allele frequencies. PLoS
ONE, 7, e45170.
Hall, L., & Beissinger, S. R. (2014). A practical toolbox for design and
analysis of landscape genetics studies. Landscape Ecology, 29,
1487– 1504.
Handley, C. O., Wilson, D. E., & Gardner, A. L. (1991). Demography and
natural history of the common fruit bat, Artibeus jamaicensis, on
Barro Colorado Island, Panama. Smithsonian Contributions to
Zoology, 511, 1–173.
Hansen, M. C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. A.,
Tyukavina, A., . . . Townshend, J. R. G. (2013). High-resolution global
maps of 21st-century forest cover change. Science, 342, 850–853.
Hanson, T. R., Brunsfeld, S. J., Finegan, B., & Waits, L. P. (2008). Pollen
dispersal and genetic structure of the tropical tree Dipteryx panamen-
sis in a fragmented Costa Rican landscape. Molecular Ecology, 17,
2060–2073.
Harvey, C., Alpizar, F., Chacon, M., & Madrigal, R. (2005). Assessing
link-ages between agriculture and biodiversity in Central America,
histori-cal overview and future perspectives. Mesoamerican and
Caribbean Region, Conservation Science Program, The Nature
Conservancy, San Jose, Costa Rica.
Hidalgo-Mihart, M. G., Contreras-Moreno, F. M., Jesus de la Cruz, A., &
Juarez-Lopez, R. (2017). Validation of the Calakmul–Laguna de
Termi-nos corridor for jaguars Panthera onca in south-eastern
Mexico. Oryx, 1–8. https://doi.org/10.1017/S0030605316001083
Hobinger, T., Schindler, S., Seaman, B. S., Wrbka, T., & Weissenhofer,
A. (2012). Impact of oil palm plantations on the structure of the agro-
forestry mosaic of La Gamba, southern Costa Rica, potential implica-
tions for biodiversity. Agroforestry Systems, 85, 367–381.
Holdridge, L. R., Grenke, W. G., Haheway, W. H., Liang, T., & Tosi, J. J.
A. (1975). Forest environments in tropical life zones. New York, NY:
Perg-amon Press.
Kadoya, T. (2009). Assessing functional connectivity using empirical data.
Population Ecology, 51, 5–15.
Kalinowski, S. T. (2005). HP-RARE, a computer program for performing rar-
efaction on measures of allelic diversity. Molecular Ecology Notes, 5,
187–189.
Khazan, E. S. (2014). Tests of biological corridor efficacy for
conservation of a Neotropical giant damselfly. Biological
Conservation, 177, 117– 125.
Laksa, M. (1997). Structure of understory shrub assemblages in adjacent
secondary and old growth tropical wet forests, Costa Rica. Biotropica,
29, 29–37.
Landguth, E. L., Cushman, S. A., Schwartz, M. K., McKelvey, K. S., Mur-
phy, M., & Luikart, G. (2010). Quantifying the lag time to detect barri-
ers in landscape genetics. Molecular Ecology, 19, 4179–4191.
Lesinski, G., Sikora, A., & Olzewski, A. (2010). Bat casualties on a road
crossing a mosaic landscape. European Journal of Wildlife Research,
57, 217–223.
Lobova, T. A., Geiselman, S. K., & Mori, S. A. (2009). Seed dispersal by
bats in the Neotropics (359 pp.). New York, NY: New York Botanical
Garden.
Lonsinger, R. C., & Waits, L. P. (2015). ConGenR, rapid determination of
consensus genotypes and estimates of genotyping errors from repli-cated
genetic samples. Conservation Genetics Resources, 7, 841–843.
Lopez, J. E., & Vaughan, C. (2007). Food niche overlap among
neotropical frugivorous bats in Costa Rica. Revista de Biologıa
Tropical, 55, 301– 313.
Mace, G., Masundire, H., & Baillie, J. (2005). Ecosystems and human
well-being: Current state and trends: Findings of the Condition and
Trends Working Group. In R. Hassan, R. Scholes, & N. Ash (Eds.),
Mil-lennium ecosystem assessment (pp. 2–20). Washington, DC:
Island Press.
| 4615
Manel, S., & Holderegger, R. (2013). Ten years of landscape genetics.
Trends in Ecology & Evolution, 28, 614–621.
McCracken, G. F., & Bradbury, J. W. (1981). Social organisation and kin-
ship in the polygynous bat Phyllostomus hastatus. Behavioral
Ecology and Sociobiology, 8, 11–34.
McCulloch, E., & Stevens, R. (2011). Rapid development and screening
of microsatellite loci for Artibeus lituratus and their utility for six
related species within Phyllostomidae. Molecular Ecology
Resources, 11, 903– 913.
McCulloch, E. S., Tello, S., Whitehead, A., Rolon~-Mendoza, C. M. J.,
Mal-donado-Rodrıguez, M. C. D., & Stevens, R. D. (2013).
Fragmentation of Atlantic Forest has not affected gene flow of a
widespread seed-dispersing bat. Molecular Ecology, 22, 4619–4633.
McGarigal, K., Cushman, S. A., & Ene, E. (2012). FRAGSTATS v4: Spatial
pat-tern analysis program for categorical and continuous maps.
Computer software program produced by the authors at the
University of Mas-sachusetts, Amherst. Retrieved from
http://www.umass.edu/landeco/ research/fragstats/fragstats.html
Mech, S. G., & Hallett, J. G. (2001). Evaluating the effectiveness of corri-
dors: A genetic approach. Conservation Biology, 15, 467–474.
Medina, A., Harvey, C. A., Merlo, D. S., Vilchez, S., & Hernandez, B.
(2007). Bat diversity and movement in an agricultural landscape in
Matiguas, Nicaragua. Biotropica, 39, 120–128.
Melo, F. P. L., Rodrıguez-Herrera, B., Chazdon, R. L., Medellın, R. A., &
Ceballos, G. G. (2009). Small tent-roosting bats promote dispersal of
large-seeded plants in a Neotropical forest. Biotropica, 41, 737–743.
Meyer, C. F. J., Frund,€ J., Lizano, W. P., & Kalko, E. K. V. (2008).
Ecologi-cal correlates of vulnerability to fragmentation in Neotropical
bats. Journal of Applied Ecology, 45, 381–391.
Meyer, C. F. J., & Kalko, E. K. V. (2008). Bat assemblages on Neotrop-
ical land-bridge islands: Nested subsets and null model analyses of
species co-occurrence patterns. Diversity and Distributions, 14, 644–
654.
Meyer, C. F. J., Kalko, E. K. V., & Kerth, G. (2009). Small-scale frag-
mentation effects on local genetic diversity in two phyllostomid bats
with different dispersal abilities in Panama. Biotropica, 41, 95– 102.
Morrison, D. (1978). Influence of habitat on the foraging distances of the
fruit bat, Artibeus jamaicensis. Journal of Mammalogy, 59, 622– 624.
Morse, W. C., Schedlbauer, J. L., Sesnie, S. E., Finegan, B., Harvey, C.
A., Hollenhorst, S. J., . . . Wulfhorst, J. D. (2009). Consequences of
envi-ronmental service payments for forest retention and recruitment
in a Costa Rican biological corridor. Ecology and Society, 14, 23–36.
Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A., &
Kent, J. (2000). Biodiversity hotspots for conservation priorities. Nat-
ure, 403, 853–858.
Nagy, M., Gunther,€ L., Knornschild,€ M., & Mayer, F. (2013). Female-
biased dispersal in a bat with a female-defence mating strategy.
Molecular Ecology, 22, 1733–1745.
Nagy, M., Heckel, G., Voigt, C. C., & Mayer, F. (2007). Female-biased
dis-persal and patrilocal kin groups in a mammal with resource-
defence polygyny. Proceedings of the Royal Society B: Biological
Sciences, 274, 3019–3025.
Norberg, U. M., & Rayner, J. M. V. (1987). Ecological morphology and
flight in bats (Mammalia; Chiroptera), wing adaptations, flight perfor-
mance, foraging strategy and echolocation. Philosophical
Transactions of the Royal Society of London B, 316, 335–427.
Nowakowski, A. J., DeWoody, J. A., Fagan, M. E., Willoughby, J. R., &
Donnelly, M. A. (2015). Mechanistic insights into landscape genetic
structure of two tropical amphibians using field-derived resistance
surfaces. Molecular Ecology, 24, 580–595.
Numa, C., Verdu, J. R., & Sanchez-Palomino, P. (2005). Phyllostomid
bat diversity in a variegated coffee landscape. Biological
Conservation, 122, 151–158.
4616 | CLEARY ET AL.
Oksanen, J. F., Blanchet, G., Kindt, R., Legendre, P., Minchin, P. R.,
O’Hara, R. B., . . . Wagner, H. (2015). VEGAN: Community Ecology
Package. R package version 2.3-1.
Ortega, J., & Castro-Arellano, I. (2001). Artibeus jamaicensis.
Mammalian Species, 662, 1–9.
Ortega, J., Maldonado, J. E., Arita, H. T., Wilkinson, G. S., & Fleischer,
R. C. (2002). Characterization of microsatellite loci in the Jamaican
fruit-eating bat Artibeus jamaicensis and cross-species amplification.
Molec-ular Ecology Notes, 2, 462–464.
Ortega, J., Maldonado, J. E., Wilkinson, G. S., Arita, H. T., & Fleischer,
R. C. (2003). Male dominance, paternity, and relatedness in the
Jamai-can fruit-eating bat (Artibeus jamaicensis). Molecular Ecology,
12, 2409–2415.
Pavlacky, D. C., Goldizen, A. W., Prentis, P. J., Nicholls, J. A., & Lowe, A. J.
(2009). A landscape genetics approach for quantifying the relative influ-
ence of historic and contemporary habitat heterogeneity on the genetic
connectivity of a rainforest bird. Molecular Ecology, 18, 2945–2960.
Peakall, R., & Smouse, P. E. (2012). GENALEX 6.5, genetic analysis in Excel.
Population genetic software for teaching and research – an update.
Bioinformatics, 28, 2537–2539.
Perfecto, I., & Vandermeer, J. (2008). Biodiversity conservation in
tropical agroecosystems: A new conservation paradigm. Annals of
the New York Academy of Sciences, 1134, 173–200.
Phommexay, P., Satasook, C., Bates, P., Pearch, M., & Bumrungsri, S.
(2011). The impact of rubber plantations on the diversity and activity
of understorey insectivorous bats in southern Thailand. Biodiversity
and Conservation, 20, 1441–1456.
Pitacuar-Meneses, M. L. (2010). Analisis economico para entender la
forma en que se asignan los bosques y otros usos de la tierra en el
Corredor Biologico San Juan La Selva, Costa Rica. MSc thesis,
CATIE, Turrialba, 85 p.
Raymond, M., & Rousset, F. (1995). GENEPOP (version 1.2), population
genetics software for exact tests and ecumenicism. Journal of Hered-
ity, 86, 248–249.
Razgour, O., Rebelo, H., Puechmaille, S. J., Juste, J., Ibanez,~ C., Kiefer, A.,
. . . Jones, G. (2014). Scale-dependent effects of landscape variables
on gene flow and population structure in bats. Diversity and Distribu-
tions, 20, 1173–1185.
Reid, F. (2009). Mammals of Central America and Southeast Mexico
(346 pp.). New York, NY: Oxford University Press.
Rex, K., Kelm, D. H., Wiesner, K., Kunz, T. H., & Voigt, C. C. (2008). Spe-cies
richness and structure of three Neotropical bat assemblages. Bio-logical
Journal of the Linnean Society, 94, 617–629.
Richards, S. A. (2005). Testing ecological theory using the information-
theoretic approach: Examples and cautionary results. Ecology, 86,
2805–2814.
Ripperger, S. P., Tschapka, M., Kalko, E. K. V., Rodriguez-Herrera, B., &
Mayer, F. (2013). Life in a mosaic landscape: Anthropogenic habitat
fragmentation affects genetic population structure in a frugivorous
bat species. Conservation Genetics, 14, 76–89.
Ripperger, S. P., Tschapka, M., Kalko, E. K. V., Rodrıguez-Herrera, B., &
Mayer, F. (2014). Resisting habitat fragmentation: High genetic con-
nectivity among populations of the frugivorous bat Carollia castanea
in an agricultural landscape. Agriculture, Ecosystems and
Environment, 185, 9–15.
Sagot, M., Phillips, C. D., Baker, R. J., & Stevens, R. D. (2016). Human-
modified habitats change patterns of population genetic structure and
group relatedness in Peter’s tent-roosting bats. Ecology and Evo-
lution, 6, 6050–6063.
Sanfiorenzo, A. (2016). Forest fragmentation influences on the genetic struc-
ture and diversity of understory tree Symphonia globulifera: An interdisci-
plinary approach. Dissertation, University of Idaho, Moscow, ID.
Schelhas, J., & Sanchez-Azofeifa, G. A. (2006). Post-frontier forest
change adjacent to Braulio Carrillo National Park, Costa Rica.
Human Ecology, 34, 408–431.
Sesnie, S., Finegan, B., Gessler, P., & Ramos, Z. (2009). Landscape-
scale environmental and floristic variation in Costa Rican old-growth
rain forest remnants. Biotropica, 4, 16–26.
Sharma, S., Dutta, T., Maldonado, J. E., Wood, T. H., Panwar, H. S., &
Sei-densticker, J. (2013). Forest corridors maintain historical gene
flow in a tiger metapopulation in the highlands of central India.
Proceedings of the Royal Society B: Biological Sciences, 280,
https://doi.org/10. 1098/rspb.2013.1506.
Shaver, I., Chain-Guadarrama, A., Cleary, K. A., Sanfiorenzo, A.,
Santiago-Garcıa, R. J., Finegan, B., . . . Waits, L. P. (2015). Coupled
social and ecological outcomes of agricultural intensification in Costa
Rica and the future of biodiversity conservation in tropical agricultural
regions. Global Environmental Change, 32, 74–86.
Slatkin, M. (1985). Rare alleles as indicators of gene flow. Evolution, 39,
53–65.
Spear, S. F., Balkenhol, N., Fortin, M. J., McRae, B. H., & Scribner, K.
(2010). Use of resistance surfaces for landscape genetic studies:
Con-siderations for parameterization and analysis. Molecular
Ecology, 19, 3576–3591.
Spear, S. F., & Storfer, A. (2008). Landscape genetic structure of coastal
tailed frogs (Ascaphus truei) in protected vs. managed forests.
Molecu-lar Ecology, 17, 4642–4656.
Storfer, A., Murphy, M. A., Evans, J. S., Goldberg, C. S., Robinson, S.,
Spear, S. F., . . . Waits, L. P. (2007). Putting the “landscape” in land-
scape genetics. Heredity, 98, 128–142.
Storfer, A., Murphy, M. A., Spear, S. F., Holderegger, R., & Waits, L. P.
(2010). Landscape genetics: Where are we now? Molecular Ecology,
19, 3496–3514.
Struebig, M. J., Kingston, T., Petit, E. J., Le Comber, S. C., Zubaid, A.,
Mohd-Adnan, A., & Rossiter, S. J. (2011). Parallel declines in species
and genetic diversity in tropical forest fragments. Ecology Letters, 14,
582–590.
Thies, W., & Kalko, E. K. V. (2004). Phenology of neotropical pepper
plants (Piperaceae) and their association with their main dispersers,
two short-tailed fruit bats, Carollia perspicillata and C. castanea
(Phyl-lostomidae). Oikos, 104, 362–376.
Thies, W., Kalko, E. K. V., & Schnitzler, H. (2006). Influence of environ-ment
and resource availability on activity patterns of Carollia castanea
(Phyllostomidae) in Panama. Journal of Mammalogy, 87, 331–338.
Tscharntke, T., Clough, Y., Wanger, T. C., Jackson, L., Motzke, I.,
Perfecto, I., . . . Whitbread, A. (2012). Global food security,
biodiversity conser-vation and the future of agricultural intensification.
Biological Conser-vation, 151, 53–59.
Van Strien, M. J., Keller, D., & Holderegger, R. (2012). A new analytical
approach to landscape genetic modelling: Least-cost transect analysis
and linear mixed models. Molecular Ecology, 21, 4010–4023.
Vaughan, C., Ramırez, O., Herrera, G., & Guries, R. (2007). Spatial
ecology and conservation of two sloth species in a cacao landscape
in limon, Costa Rica. Biodiversity and Conservation, 16, 2293–2310.
Vazquez-Domınguez, E., Mendoza-Martınez, A., Orozco-Lugo, L., &
Cuaron, A. D. (2013). High dispersal and generalist habits of the bat
Artibeus jamaicensis on Cozumel Island, Mexico, an assessment
using molecular genetics. Acta Chiropterologica, 15, 411–421.
Villate, R., Canet-Desanti, L., Chassot, O., & Monge-Arias, G. (2009). El
Corredor Biologico San Juan-La Selva, una estrategia exitosa de
conser-vacion (93 pp.). San Jose, Costa Rica: National Biological
Corridor Program of Costa Rica.
Waits, L. P., Cushman, S., & Spear, S. (2016). Applications of landscape
genetics to connectivity research in terrestrial animals. In N. Balken-
hol, S. Cushman, A. Storfer, & L. P. Waits (Eds.), Landscape
genetics: Concepts, methods, applications (pp. 199–214). West
Sussex, UK: Wiley-Blackwell.
Wang, I. J. (2010). Recognizing the temporal distinctions between land-
scape genetics and phylogeography. Molecular Ecology, 19, 2605–
2608.
CLEARY ET AL. | 4617

Wilcove, D. S., & Koh, L. P. (2010). Addressing the threats to biodiversity SUPPORTINGINFORMATION
from oil-palm agriculture. Biodiversity and Conservation, 19, 999–
1007. Additional Supporting Information may be found online in the sup-
Woltmann, S., Kreiser, B. R., & Sherry, T. W. (2012). Fine-scale genetic porting information tab for this article.
population structure of an understory rainforest bird in Costa Rica.
Conservation Genetics, 13, 925–935.
Wright, S. J., Hernandez, A., Condit, R., Wright, S. J., Hernandez, A., &
Condit, R. (2007). The bushmeat harvest alters seedling banks by How to cite this article: Cleary KA, Waits LP, Finegan B.
favoring lianas, large seeds, and seeds dispersed by bats, birds, and Comparative landscape genetics of two frugivorous bats in a
wind. Biotropica, 39, 363–371.
biological corridor undergoing agricultural intensification.
Wright, H. L., Lake, I. R., & Dolman, P. M. (2012). Agriculture—a key ele-
ment for conservation in the developing world. Conservation Letters, Mol Ecol. 2017;26:4603–4617. https://doi.org/10.1111/
5, 11–19. mec.14230
Zurcher, A. A., Sparks, D., & Bennett, V. (2010). Why did the bat not
cross the road? Acta Chripterologica, 12, 337–340.