animal biodiversity and conservation issue 28.2 (2005)

Anim

al Biodiversity and C

onservation, 28.2 2005

Form

erly

Mis

ce

l·là

nia

Zo

olò

gic

a

2005

AnimalBiodiversity Conservation28.2

and

Editor executiu / Editor ejecutivo / Executive Editor Joan Carles Senar

Secretària de redacció / Secretaria de redacción / Managing EditorMontserrat Ferrer

Consell assessor / Consejo asesor / Advisory BoardOleguer EscolàEulàlia GarciaAnna OmedesJosep PiquéFrancesc Uribe

Editors / Editores / EditorsPere Abelló Inst. de Ciències del Mar CMIMA–CSIC, Barcelona, SpainJavier Alba–Tercedor Univ. de Granada, Granada, SpainAntonio Barbadilla Univ. Autònoma de Barcelona, Bellaterra, SpainXavier Bellés Centre d' Investigació i Desenvolupament–CSIC, Barcelona, SpainJuan Carranza Univ. de Extremadura, Cáceres, SpainLuís Mª Carrascal Museo Nacional de Ciencias Naturales–CSIC, Madrid, SpainMichael J. Conroy Univ. of Georgia, Athens, USAAdolfo Cordero Univ. de Vigo, Vigo, SpainMario Díaz Univ. de Castilla–La Mancha, Toledo, SpainJosé Antonio Donazar Estación Biológica de Doñana–CSIC, Sevilla, SpainGary D. Grossman Univ. of Georgia, Athens, USADamià Jaume IMEDEA–CSIC, Univ. de les Illes Balears, SpainJordi Lleonart Inst. de Ciències del Mar CMIMA–CSIC, Barcelona, SpainJorge M. Lobo Museo Nacional de Ciencias Naturales–CSIC, Madrid, Spain Pablo J. López–González Univ de Sevilla, Sevilla, SpainJuan José Negro Estación Biológica de Doñana–CSIC, Sevilla, SpainVicente M. Ortuño Univ. de Alcalá de Henares, Alcalá de Henares, SpainMiquel Palmer IMEDEA–CSIC, Univ. de les Illes Balears, SpainFrancisco Palomares Estación Biológica de Doñana–CSIC, Sevilla, SpainFrancesc Piferrer Inst. de Ciències del Mar CMIMA–CSIC, Barcelona, SpainMontserrat Ramón Inst. de Ciències del Mar CMIMA –CSIC, Barcelona, SpainIgnacio Ribera Nacional de Ciencias Naturales–CSIC, Madrid, SpainPedro Rincón Museo Nacional de Ciencias Naturales–CSIC, Madrid, SpainAlfredo Salvador Museo Nacional de Ciencias Naturales–CSIC, Madrid, SpainJosé Luís Tellería Univ. Complutense de Madrid, Madrid, SpainFrancesc Uribe Museu de Ciències Naturals de Barcelona, Barcelona, Spain

Consell Editor / Consejo editor / Editorial BoardJosé A. Barrientos Univ. Autònoma de Barcelona, Bellaterra, SpainJean C. Beaucournu Univ. de Rennes, Rennes, FranceDavid M. Bird McGill Univ., Québec, CanadaMats Björklund Uppsala Univ., Uppsala, SwedenJean Bouillon Univ. Libre de Bruxelles, Brussels, BelgiumMiguel Delibes Estación Biológica de Doñana–CSIC, Sevilla, SpainDario J. Díaz Cosín Univ. Complutense de Madrid, Madrid, SpainAlain Dubois Museum national d’Histoire naturelle–CNRS, Paris, FranceJohn Fa Durrell Wildlife Conservation Trust, Jersey, United KingdomMarco Festa–Bianchet Univ. de Sherbrooke, Québec, CanadaRosa Flos Univ. Politècnica de Catalunya, Barcelona, SpainJosep Mª Gili Inst. de Ciències del Mar CMIMA–CSIC, Barcelona, SpainEdmund Gittenberger Rijksmuseum van Natuurlijke Historie, Leiden, The NetherlandsFernando Hiraldo Estación Biológica de Doñana–CSIC, Sevilla, SpainPatrick Lavelle Inst. Français de recherche scient. pour le develop. en cooperation, Bondy, FranceSantiago Mas–Coma Univ. de Valencia, Valencia, SpainJoaquín Mateu Estación Experimental de Zonas Áridas–CSIC, Almería, SpainNeil Metcalfe Univ. of Glasgow, Glasgow, United KingdomJacint Nadal Univ. de Barcelona, Barcelona, SpainStewart B. Peck Carleton Univ., Ottawa, Canada Eduard Petitpierre Univ. de les Illes Balears, Palma de Mallorca, SpainTaylor H. Ricketts Stanford Univ., Stanford, USAJoandomènec Ros Univ. de Barcelona, Barcelona, SpainValentín Sans–Coma Univ. de Málaga, Málaga, SpainTore Slagsvold Univ. of Oslo, Oslo, Norway

Secretaria de redacció / Secretaría de redacción / Editorial Office

Museu de Ciències Naturals Passeig Picasso s/n08003 Barcelona, SpainTel. +34–93–3196912Fax +34–93–3104999E–mail [email protected]

Animal Biodiversity and Conservation 28.2, 2005© 2005 Museu de Ciències Naturals, Institut de Cultura, Ajuntament de BarcelonaAutoedició: Montserrat FerrerFotomecànica i impressió: Sociedad Cooperativa Librería GeneralISSN: 1578–665XDipòsit legal: B–16.278–58

The journal is freely available online at: http://www.bcn.cat/ABC

"Le psettus rhomboidal (Psettusrhombeus,Cav. Nat.)" Le Règne Animal par Georges Cuvier; Paris: Fortin, Masson et Cie, Librairies; Pl. 42 Poissons

http://bcnweb13.bcn.es:81/NASApp/wprmuseuciencies/Museu.GeneradorPagines?idioma=3&seccio=11_1

101Animal Biodiversity and Conservation 28.2 (2005)

© 2005 Museu de Ciències NaturalsISSN: 1578–665X

Carrascal, L. M. & Palomino, D., 2005. Preferencias de hábitat, densidad y diversidad de las comunidadesde aves en Tenerife (Islas Canarias). Animal Biodiversity and Conservation, 28.2: 101–119.

AbstractSpecies–specific habitat preferences, density and species richness of bird communities in Teneriffe (CanaryIslands).— Bird distribution and abundance are described and analyzed in Teneriffe (Canary Islands). Inter–habitat differences in density, diversity and species richness are shown in table 1. Figure 2 shows the maindeterminants of bird species richness in Teneriffe, and tables 2 and 3 and figure 3 show the species–specificpatterns of spatial variation abundance (more detailed for Anthus berthelotii, Fringilla coelebs canariensis,Fringilla teydea, Parus caeruleus teneriffae, Phylloscopus canariensis, Regulus teneriffae, Serinus canariusand Turdus merula cabrerae). Deeply transformed environments due to human impact (urban habitats,agricultural mosaics, banana plantations) have high bird densities and species richness, even higher thanthose measured in native, unmodified habitats such as laurel forests or mature pinewoods. Urbanenvironments in Teneriffe are very permeable to native bird fauna, as they have been occupied by manywidespread endemic species/subspecies. Many of the endemic, well defined species or subspecies of islandbirds have high population densities within native, untransformed habitats. Density compensation and nicheexpansion is not a common phenomenon in the avifauna of Teneriffe. Nevertheless, all species/subspeciesbroadening the inter–habitat or altitudinal distribution are endemic of the Canary Islands.

Key words: Altitudinal distribution, Avifauna, Bird density, Habitat preferences, Island vs. mainland compari-sons, Teneriffe Island.

ResumenPreferencias de hábitat, densidad y diversidad de las comunidades de aves en Tenerife (Islas Canarias).—Mediante el empleo de transectos lineales, se describen las preferencias de hábitat, la distribuciónaltitudinal y la abundancia de la avifauna reproductora de Tenerife (Islas Canarias). Los hábitatsprofundamente transformados debido a la acción humana (e.g., áreas urbanas, mosaicos agrícolas,plantaciones de plátanos) tienen elevadas densidades y riquezas de especies, que llegan a ser tan altas omayores que las observadas en medios autóctonos no transformados como laurisilvas y pinares maduros.Muchas especies/subespecies taxonómicamente bien diferenciadas de las poblaciones continentales estándistribuidas mayoritaria o exclusivamente en hábitats autóctonos poco degradados. Las hipótesis de lacompensación de densidades y la expansión de nicho en poblaciones insulares no parecen cumplirse demodo generalizado en Tenerife. No obstante, todas las especies o subespecies que muestran una mayoramplitud de distribución en Tenerife son endémicas del archipiélago canario.

Palabras clave: Distribución altitudinal, Avifauna, Densidad de aves, Preferencias de hábitat, Tenerife.

(Received: 1 VI 04; Conditional acceptance: 16 IX 04; Final acceptance: 30 IX 04)

L. M. Carrascal & D. Palomino, Dept. Biodiversidad y Biología Evolutiva, Museo Nacional de CienciasNaturales–CSIC, c/ José Gutiérrez Abascal 2, 28006 Madrid, Spain.

Corresponding author: L. M. Carrascal. E–mail: [email protected]

Preferencias de hábitat, densidad ydiversidad de las comunidades de avesen Tenerife (Islas Canarias)

L. M. Carrascal & D. Palomino

mailto:[email protected]

102 Carrascal & Palomino

Introducción

En la actualidad los archipiélagos e islas oceánicosconstituyen escenarios de investigación ornitológicatan habituales en contextos de teoría ecológica comode biología de la conservación. Por una parte, suscondiciones de aislamiento geográfico a lo largo deenormes periodos de tiempo son idóneas para elexamen de hipótesis complejas y/o poco susceptiblesde experimentación en el campo de la ecologíaevolutiva. Algunos ejemplos de teorías y conceptosbiogeográficos basados en observaciones enecosistemas insulares serían el equilibrio dinámicode especies (MacArthur & Wilson, 1967), lacompensación de densidades poblacionales(MacArthur et al., 1972; Blondel et al., 1988), laexpansión/contracción de nichos ecológicos (Blondelet al., 1988; Martin, 1992; Prodon et al., 2002), o elciclo del taxón (Wilson, 1961; Ricklefs & Bermingham,1999). Por otra parte, los trabajos de corteconservacionista y aplicado sobre comunidadesinsulares de aves son también muy frecuentes, dadosel alto nivel de endemismos que acogen estosambientes (Johnson & Stattersfield, 1990; Stattersfieldet al., 1998) y su predisposición a padecer elevadosniveles de amenaza y tasas de extinción (Milberg &Tyrberg, 1993; Collar et al., 1994; Pimm et al., 1988;pero ver también Manne et al., 1999).A pesar del extenso conocimiento ornitológico delarchipiélago Canario (ver la extensa revisiónbibliográfica de Martín & Lorenzo, 2001), el númerode trabajos cuantitativos sobre la distribución,abundancia y biogeografía de su avifauna es aúnreducido. No obstante, los trabajos ya acumuladosabordan una notable diversidad de temas,incluyendo descripciones generales de distribucióny abundancia en el archipiélago (p.e., Báez, 1992;Fernández–Palacios & Andersson, 1993; Marshall& Baker, 1999; Carrascal & Palomino, 2002),estructura y composición de comunidades endeterminados hábitats (p.e., Suárez, 1984;Carrascal, 1987), procesos de selección de hábitaty uso del espacio en especies concretas (p.e.,Carrascal et al., 1992; Valido et al., 1994; Illera,2001), fenómenos de interacción planta–animal(p.e., Nogales et al., 2001; 2002) o paleontología(p.e., Rando et al., 1999; Rando, 2002).

La disponibilidad de información empírica sobrelos patrones de distribución y abundancia de lasaves es fundamental para análisis biogeográficosdetallados y para una adecuada gestión de labiodiversidad. Esta información puede servir paraestablecer de modo objetivo la rareza de los orga-nismos y de esta manera poder definir categorías deamenaza y Listas Rojas (ver por ejemplo los crite-rios y listas SPEC para aves europeas y UICN paratodo el planeta; Tucker & Heath, 1994; UICN, 2001).Para llevar a cabo esta tarea según los criterioscuantitativos, actuales es necesario enfrentarse apreguntas como ¿cuán abundante es cada especie?¿cuál es su valencia ecológica? y ¿en qué medidaestán creciendo o decreciendo sus poblaciones?Considerando estos hechos, este trabajo utiliza el

potencial que tienen los datos obtenidos a partir decensos extensivos de aves para examinar hipótesisbiogeográficas y macroecológicas, y cuantificar larareza actual de la avifauna de la isla de Tenerife,para de este modo proporcionar información quepueda contribuir a una mejor gestión de este recursonatural. Para ello, elaboramos modelos descriptoresde la distribución y abundancia de las especies deaves reproductoras en Tenerife, identificando paracada una de ellas los rasgos ambientales concretosque más les favorecen. Además, se examinan algu-nas hipótesis relativas a avifaunas insulares: expan-sión de nicho y compensación e incrementos dedensidad en poblaciones de especies insulares, yempobrecimiento diferencial de la avifauna en dis-tintos hábitats atendiendo a la estructura de la vege-tación (Blondel, 1979; Wiens, 1989; Brown &Lomolino, 1998).

Material y métodos

Área de estudio

Tenerife es una isla de 2.059 km2 localizada en elarchipiélago canario y distante 288 km de la costaafricana. Su gran extensión y la existencia de unamplio gradiente altitudinal (desde el nivel del marhasta el Teide a 3.718 m) determina una grandiversidad de condiciones climatológicas y forma-ciones vegetales diferentes (Anónimo, 1980;González et al., 1986). Los principales mediosautóctonos que pueden distinguirse son (1) tabaibalesy cardonales–tabaibales dominados por plantasmarcadamente xerófilas de porte arbustivo ysubarbóreo (Euphorbia spp., Plocama pendula, Kleinianeriifolia) localizados en el piso basal (0–500 m); (2)monteverde constituido por laurisilvas y diferentesetapas seriales de su degradación (brezales yfayales–brezales), localizado fundamentalmente enel norte de la isla entre los 500 y 1.200 m de altitudy dominado por diferentes especies de árboles yarbustos arborescentes (Erica arborea, Myrica faya,Persea indica, Ocotea foetens, Laurus azorica,Picconia excelsa e Ilex canariensis); (3) diversasformaciones de porte arbustivo distribuidas por enci-ma del cardonal–tabaibal hasta los 2.500 m dealtitud (jarales, codesares, retamares; dominadosprincipalmente por Spartocytisus spp., Chamaecytisusproliferus, Adenocarpus spp., Cistus spp. y Micromeriaspp.); (4) pinares de Pinus canariensis distribuidosdesde los 1.200 m en el norte de la isla y los 600 men el sur hasta los 2.100 m de altitud; (5) formacionesalpinas localizadas por encima de los 2.500 m ycaracterizadas por una escasísima cubierta vegetalrelegada a unos pocos caméfitos que se desarrollansobre malpaises y otros suelos volcánicos. A estosgrandes tipos de paisaje hay que añadir formacionesagropecuarias derivadas de las actividades humanasque se localizan desde el nivel del mar hasta los1.000 m principalmente (plataneras, mosaicos decultivo, pastizales) y áreas urbanas de distinto tama-ño y desarrollo urbanístico.

Animal Biodiversity and Conservation 28.2 (2005) 103

Método de censo

Durante Abril de 2002 y 2003 se efectuaron censos alo largo de toda la isla para cuantificar la distribucióny abundancia de las aves de Tenerife (fig. 1). Loscensos sólo incluyeron medios terrestres, habiéndosedescartado la línea de costa y las zonas húmedasartificiales (balsas, embalses). El método elegido fueel del transecto lineal, contabilizándose todas lasaves vistas u oídas a lo largo del trayecto. Se distin-guieron los contactos efectuados hasta una distanciade 25 m a cada lado del observador con el objeto deefectuar estimas de densidad. Sólo se censó en díassin viento ni precipitaciones, entre las 7:00–11:00 ylas 16:00–17:30 GMT. La velocidad media de progre-sión andando fue de 1–3 km/h. Para más detallesacerca de esta metodología consúltese Bibby et al.(2000). Debido a las horas de censo, las aves noctur-nas quedaron excluidas de los inventarios aunquefueran observadas (casos del Búho Chico–Asio otus,Lechuza Común–Tyto alba y Chocha Perdiz–Scolopaxrusticola). Los vencejos tampoco pudieron ser distin-guidos con toda seguridad durante los censos, por loque sólo se anotaron los individuos contados sinidentificarlos a nivel de especie (vencejos Común–Apus apus, Unicolor–A. unicolor y Pálido–A. pallidusen el área de estudio; Martín & Lorenzo, 2001) y nose incluyeron en los análisis de datos. Tampocofueron considerados en los análisis las avesmigradoras observadas sin constancia de reproduc-

ción segura y habitual en Tenerife. En el caso de laspalomas de laurisilva (Turqué–Columba bolli yRabiche–C. junoniae) y las tórtolas recientementeintroducidas (Turca–Streptopelia decaocto y de Ca-beza Rosa–S. roseogrisea) no siempre fue posibleidentificar al nivel de especie a todos los individuos.Sin embargo, las aves identificadas como "indetermi-nadas" fueron asignadas específicamente mantenien-do las proporciones observadas en aquellos indivi-duos identificados dentro de cada localidad (6 locali-dades en zonas de monteverde, y 7 localidades enáreas urbanas).

Los transectos fueron divididos en unidades de500 m. El punto medio de cada uno de ellos fuegeoreferenciado mediante un GPS Garmin 12(latitud, longitud y altitud) utilizando la funciónpromedio permaneciendo inmóvil durante 2 minutos.Los transectos se definieron en unidades de paisajey tipos de hábitat lo más homogéneos posible,mediante el estudio de mapas 1:25.000 y visitasprevias a las áreas de censo. En cada unidad detransecto se efectuaron tres estimas de la estructurade la vegetación (a 125 m, 250 m y 375 m dentrodel transecto de 500 m) que fueron promediadaspara caracterizarlo. Se midió la cobertura (enporcentaje) de herbáceas en el suelo, la coberturadel estrato arbustivo, la altura media del matorral,la cobertura del estrato arbóreo, la altura promediodel arbolado y la cobertura de suelo urbano. Seestablecieron 9 categorías de porcentajes (0, 1,

Fig. 1. Mapa de la localización en Tenerife de los 592 transectos lineales de 0,5 km. La latitud ylongitud se expresa en coordenadas (miles de metros) dentro del bloque 28R.

Fig. 1. Location of the 592 line transects of 0.5 km censused in Teneriffe Island. Latitude andlongitude in UTM coordinates (in thousands of metre) within the block 28R.

3170

3160

3150

3140

3130

3120

3110

3100

3090

Lat

itu

d (

UT

M)

310 320 330 340 350 360 370 380 390 400 Longitud (UTM)


2–5, 5–12, 13–25, 25–50, 50–75, 75–95, 95–100%)para establecer las coberturas. También secaracterizó cada transecto atendiendo a si incluíanbarrancos o suelo con uso agropecuario (diversostipos de cultivo o praderas de siega). Los grandestipos de bosques se codificaron atendiendo a sieran o no (si 1; no 0) pinares o monteverde, o a silos arbustos eran tabaibas (Euphorbia spp), brezos,codesos/escobones (Chamaecytisus proliferus,Adenocarpus spp.), o retamas del Teide(Spartocytisus supranubius). En total se efectuaron296 km de censo repartidos en 592 unidades decenso de 0,5 km. Para cada especie se proporcionóel porcentaje de unidades en que se observó(independientemente de la distancia a la que fuesecontactada; i.e., sin considerar bandas de censo).

Estas 592 unidades fueron agrupadas en 26 for-maciones vegetales – tipos de paisaje atendiendo asu localización geográfica (principalmente norte osur de la isla y posición altitudinal), proximidad yestructura de la vegetación (tabla 1). Cada una deellas fue representada por al menos 20 has decenso (8 unidades de 0,5 km). Debido a las carac-terísticas particulares de algunas muestras (mixtasentre distintas formaciones ambientales; p.e., tran-siciones pinar–fayal, laurisilvas degradadas,tabaibales–cultivos, brezales–huertos, áreas urba-nas muy dispersas) o escasez numérica total o porlocalidad (menos de 20 ha censadas), 103 unida-des de censo no fueron utilizadas en la tabla 1 aldefinir los 26 hábitats principales de Tenerife. Enellos se estableció la densidad de cada especie,excluyendo a las aves nocturnas, vencejos ymigrantes no reproductores. Estos valores de den-sidad deben considerarse medidas mínimas deabundancia, ya que el método del taxiado no detec-ta todos los individuos existentes, al oscilar gene-ralmente las detectabilidades (probabilidad de de-tectar un ave estando presente), según las espe-cies, entre un 33% y 80% (Bibby et al., 2000).

Las 592 unidades de censo también fueron agru-padas en seis bandas altitudinales cada 500 m. Encada una de ellas se calculó la abundancia relativade las especies expresada en aves/km, utilizandopara ello todas las aves vistas u oídas sin tener encuenta la banda principal de recuento de 25 m acada lado del observador.

Análisis de datos

Las comunidades de aves en cada uno de estos 26hábitats fue caracterizada por la densidad total deaves, la riqueza de especies y la diversidad. Lariqueza se midió mediante el número de especiescuya densidad era mayor de 0,5 aves/10 ha (S0,5).De este modo se evitó el efecto de la diferentesuperficie muestreada en distintas comunidades yla inclusión de especies accidentales o muy raras.La diversidad se estimó mediante el índice deShannon (H’ = – Spi · ln pi, donde pi es la propor-ción de la densidad de la especie i dentro de ladensidad total de aves). Las estimas de diversidadno se han visto influidas por la distinta superficie

muestreada en cada unidad ambiental, ya que H’ ysuperficie de censo (en logaritmo) no estánsignificativamente relacionados en la muestra delas 26 formaciones de la tabla 1 (r = 0,016,p = 0,939). Otro tanto ocurre al analizar el efectoque la superficie muestreada tiene sobre la estimade riqueza estandarizada a un mínimo de densidad(> 0,5 aves/10 ha; r = 0,008, n = 26, p = 0,970).

Con la abundancia relativa en las seis bandasde distribución se calculó la amplitud de distribu-ción altitudinal utilizando la siguiente fórmula:

pi2

Amplitud =6

donde pi es la proporción de la abundancia relativaen cada una de las seis bandas altitudinales. Esteíndice varía entre 0,17 y 1, de manera que a mayorvalor del índice se corresponde una mayor ampli-tud de distribución de la especie.

Los factores influyentes sobre la distribución delas aves se identificaron con árboles de regresión(De’Ath & Fabricius, 2000) aplicados al númerototal de aves observadas en cada una de lasunidades de censo de 0,5 km (i.e., sin considerarlas bandas principales de recuento de 25 m). Losárboles de regresión someten a la variable res-puesta (índice kilométrico de abundancia en estecaso; aves/0,5 km) a sucesivas divisionesdicotómicas para obtener grupos homogéneos demuestras. Tales divisiones se hacen según crite-rios determinados por las variables predictoras(las que describen las características de cadaunidad de censo). Mediante este procedimiento(1) se obvia la necesidad de establecer "a priori"patrones lineales homogéneos a todo el conjuntode datos (caso de la regresión múltiple), (2) seevita el ajuste "forzado" a distribuciones canóni-cas concretas a los cuales no tienen por quéajustarse los datos, y (3) se definen modelos deefectos jerarquizados que particionan la variabili-dad original en subconjuntos de datos en loscuales pueden estar operando de modo distintovariables predictoras diferentes (i.e., estima deinteracciones). Los árboles de regresión permitenenfrentarse con éxito a las complejidades inheren-tes de los datos en ecología, como son las rela-ciones no lineares entre las variables respuesta ylos predictores, o las interacciones entrepredictores, por lo que son muy adecuados paraexplorar los patrones de distribución y variaciónde la abundancia en aves.

Debido a la escasez de datos para algunasespecies (menos de 5 presencias en las 592 unida-des de censo), no fue posible obtener árboles deregresión en todos los casos. La complejidad delos árboles de regresión fue limitada atendiendo alas siguientes condiciones: árboles que redujesenla devianza significativamente, con un máximo deonce criterios de clasificación (i.e., ramificación)que definían doce puntas con, al menos, 5 unida-des de censo (los árboles incluyen principalmente

(3 ) 6

i = 1

–1


5–8 puntas, con unos 7–15 transectos de 0,5 kmpor punta). Se muestran y comentan los árboles deregresión de 8 especies / subespecies endémicasde las Islas Canarias, ampliamente extendidas yque ilustran los grandes modelos de distribución delas aves en Tenerife. Para el resto de las especies,los resultados de los árboles de regresión se sinte-tizan mediante la selección de los criterios quemaximizan la abundancia de cada especie enTenerife. Estos criterios definen las configuracionesdel paisaje que permiten el asentamiento de laspoblaciones más densas dentro de la isla.

La variación espacial de la riqueza de especiesen los 592 transectos lineales de 0,5 km también seanalizó utilizando los árboles de regresión, siguien-do los criterios expuestos en el párrafo anterior. Enesta ocasión, la riqueza se mide como el número deespecies observado en 0,5 km de transecto sinutilizar distancias límite de detección (i.e., se inclu-yen todas las especies observadas).

Resultados y discusión

Comunidades de aves

La tabla 1 muestra los valores de densidad, riquezay diversidad de la avifauna terrestre de la isla deTenerife en 26 formaciones ambientales.

La densidad de aves osciló entre 0 y 95 aves/10 ha.Las mayores densidades se midieron en zonasurbanas de gran extensión (e.g., ciudades, 75–100 aves/10 ha) y en una gran variedad de mediossituados en el área de influencia de los alisios en elnorte de la isla (pastizales, zonas agrícolas, breza-les, fayales–brezales y laurisilva, 60–70 aves/10 ha).Por el contrario, las menores densidades (< 10 aves/10 ha) se obtuvieron en zonas de alta montaña(> 2.200 m), en áreas bajas del sur de la isla conmuy poca cobertura vegetal (< 250 m y coberturade arbustos y herbáceas < 25%) y en pinares abier-tos altimontanos o jóvenes (> 1.650 m y < 11 m dealtura del arbolado).

La riqueza de especies con densidades mayoresde 0,5 aves/10 ha osciló entre 0 y 15 especies. Elmayor valor de riqueza se midió en las áreasagropecuarias del norte de la isla situadas entre500 y 1.000 m de altitud. También se obtuvieronelevados valores (10–12 especies) en una granvariedad de formaciones vegetales que incluyenlos tabaibales–cardonales del norte de la isla (es-pecialmente en los barrancos), otras zonas agríco-las (cultivos del sur y plataneras) y algunas zonasurbanas (tanto pueblos pequeños como ciudades).Las menores riquezas (< 5 especies) se observa-ron en áreas desarboladas de alta montaña(> 2.200 m), formaciones xéricas de las zonas ba-jas del sur de la isla (< 250 m), y pinares jóvenespoco densos. Un patrón similar se obtuvo para ladiversidad de aves (correlación entre riqueza ydiversidad: r = 0.823, n = 26, p < 0.001). El empo-brecimiento de los medios insulares es muy paten-te en los hábitats de alta montaña de Tenerife, que

no están representados por ninguna especie típica-mente alpina o cuyos máximos poblacionales seden allí. Este hecho contrasta fuertemente con lapresencia de varias especies alpinas en los gran-des macizos montañosos de la península ibérica(e.g., Lagópodo Alpino–Lagopus mutus, Bisbita Al-pino–Anthus spinoletta, Pechiazul–Luscinia svecica,Treparriscos–Tichodroma muraria, ChovaPiquigualda–Pyrrhocorax graculus, Gorrión Alpino–Montifringilla nivalis o Acentor Alpino–Prunellacollaris; Martí & Del Moral, 2003), algunos de loscuales tienen menor superficie por encima de2.000 m que Tenerife (e.g., Sierra Nevada, Cordi-llera Cantábrica).

A continuación se analizan los principales deter-minantes de la variación espacial de la riqueza deespecies en Tenerife. El 67,8% de la variabilidadobservada (0–14 especies/0,5 km) es explicada porel árbol de regresión de la figura 2 (χ2 = 3451,df = 12, p << 0.001). Los principales factores queinfluyen sobre la riqueza de especies son la situa-ción latitudinal y altitudinal dentro de la isla. En lamitad meridional de la isla (LAT<3.135; 3,5 espe-cies/0,5 km) la riqueza es menor que en la mitadseptentrional (6,9 especies/0,5 km). La altitud notiene un efecto lineal sobre la riqueza en el sur deTenerife, sino que alcanza su máximo entre 328 y1.100 m, siendo menor por encima de 1.656 m queen el piso basal (ALT<328 m; ver distintas ramifica-ciones para la variable ALT en las ramas de laizquierda del árbol de regresión; ver también losvalores medios de especies/0,5 km en la tabla 2).En el norte de la isla existe una asociación negativacon esta variable, de manera que hay más espe-cies por debajo de 1.270 m que por encima de estaaltitud. Los efectos de la altitud y la latitud sonmatizados por el desarrollo del estrato arbóreo, lapresencia de cultivos y la existencia de núcleosurbanos. El desarrollo del arbolado tiene un mar-cado efecto positivo sobre la riqueza de especiesen el sur de la isla (tanto considerando la cober-tura como la al tura de árboles; cr i ter iosCARB<22% y HARB<3,2 m). La existencia deáreas agrícolas incrementa ligeramente el núme-ro de especies en el norte de la isla (criterioAGR=0). La cobertura de suelo urbano tiene unefecto distinto según la localización dentro deTenerife. En el sur de la isla, y a altitudes meno-res de 328 m (LAT<3.135 y ALT<328), la riquezade especies es mayor en áreas urbanas densas(CURB>92%; 5 spp/0,5 km) que en el resto delos ambientes disponibles en esta área. Por elcontrario, en el norte de Tenerife el suelo urbano,aunque sea disperso (CURB>58%) disminuye elnúmero de especies.

Patrones específicos de distribución y abundancia

La tabla 2 muestra la variación de la abundanciarelativa de las aves a lo largo de un gradientealtitudinal, así como su amplitud de distribución.La tabla 3 ilustra la frecuencia de aparición de lasespecies en los 592 transectos efectuados, la den-


Bn Br FB

– – +

0.27 – –

– – –

– – –

– – –

0.08 – +

1.07 – –

– – –

– – –

1.07 0.44 2.67

– – 1.00

6.13 – –

– + –

– – –

– – –

0.27 11.56 11.00

+ – –

– 3.56 3.00

– – 0.67

0.53 – –

– – –

– – –

0.80 – –

– – –

7.73 6.67 3.67

– – –

– – –

Am Rm R Mo TCn TCs TCxs Txs

Accipiter nisus – – – – – – – –

Alectoris barbara – – – – 0.21 3.00 + –

Anthus berthelotii – 0.78 1.92 0.31 0.84 3.40 5.47 4.46

Bucanetes githagineus – – – – – – – 0.31

Burhinus oedicnemus – – – – – – 0.21 –

Buteo buteo – – – – 0.11 + – –

Carduelis cannabina – – – – 0.63 – – –

Carduelis carduelis – – – – – – – –

Carduelis chloris – – – – – – – –

Columba bollii – – – – – – – –

Columba junoniae – – – – – – – –

Columba livia – – – – 1.26 2.00 – 1.23

Corvus corax – + – – – – – –

Coturnix coturnix – – – – – – – –

Dendrocopos major – – – – – – – –

Erithacus rubecula – 0.11 0.32 – 3.16 – – –

Falco tinnunculus – + + – 0.21 + 0.42 +

Fringilla coelebs – – – – 0.42 – – –

Fringilla teydea – – – – – – – –

Gallinula chloropus – – – – – – – –

Lanius excubitor – 0.22 0.48 – – – + 0.77

Miliaria calandra – – – – – – – –

Motacilla cinerea – – – – – – – –

Myiopsitta monachus – – – – – – – –

Parus caeruleus – 0.11 – 3.38 2.74 1.60 0.21 –

Passer hispaniolensis – – – – – – – 0.46

Petronia petronia – – – – – – – –

Tabla 1. Densidades (aves/10 ha) de las aves en 26 medios diferentes en la isla de Tenerife. Enla parte inferior de la tabla se proporcionan las variables descriptoras de cada hábitat. Acontinuación se proporciona una breve descripción de los 26 medios censados (n, situación enel norte de la isla; s, situación en el sur de la isla): Am. Alta montaña; Rm. Retamar deSpartocitysus supranubius sobre malpais; R. Retamar de Spartocitysus supranubius; Mo.Matorrales occidentales (Cistus spp., Echium spp., Sonchus spp., 1.100–1.400 m); TCn y TCs.Tabaibales–cardonales situados en el norte y sur de la isla; TCxs. Tabaibales–cardonalesxéricos; Txs. Tabaibales xéricos; Bn. Barrancos cubiertos de tabaibales–cardonales y restos dearbolado termófilo; Br. Brezales; FB. Fayal–brezal; L. Laurisilva; Pn y Ps. Pinares de Pinuscanariensis situados en el norte y sur de la isla; Pjn. Pinares jóvenes de Pinus canariensis; Pm.Pinares maduros de Pinus canariensis; Pa. Pinares altitudinales de Pinus canariensis; Pah.Pastizales húmedos; Es. Campos de cultivo abandonados de porte estepárico en el sur de laisla; Pl. Plataneras; Cn y Cs. Mosaicos de cultivos situados en el norte y sur de la isla; Pbn yPbs. Pueblos situados en el norte y sur de la isla; Cun y Cus. Cascos urbanos extensos situadosen el norte y sur de la isla. S0,5. Número de especies con densidad mayor de 0,5 aves/10 ha.HAS. Hectáreas censadas; CHB. Cobertura de herbáceas; CMAT. Cobertura de arbustos; HMAT.Altura media de los arbustos; CARB. Cobertura del arbolado (árboles mayores de 3 m dealtura); HARB. Altura media del arbolado; AGR. Porcentaje de suelo dedicado a uso agrícola;CURB. Cobertura de suelo urbano.


Pah Es Pl Cn Cs Pbn Pbs Cun Cus

– – – – – – – – –

– 0.40 – 0.10 0.85 – – – –

9.07 8.60 4.44 3.12 3.15 – 1.33 – –

– – – – – – – – –

– + – – – – – – –

+ – + + – – – – –

7.47 – – 4.20 3.15 – 0.33 – –

– – 2.67 0.10 0.12 0.89 + – –

– – – 0.59 – 0.44 – 1.11 –

– – – – – – – – –

– – – + – – – – –

– 2.80 – – 3.88 0.44 10.67 44.89 6.00

+ – – – – – – – –

+ – – 0.59 – – – – –

– – – – – – – – –

– – – 3.80 – – – – –

+ 0.40 0.44 0.20 0.12 + + + +

– – – 1.37 – – – – –

– – – – – – – – –

– – – – – – – – –

– 0.60 – – – – – – –

10.40 – – 1.56 – – – – –

– – 1.33 0.49 1.58 1.78 1.33 0.89 –

– – – – – – – 0.44 –

0.53 – 2.22 0.88 2.79 2.22 2.33 1.33 0.20

– – 4.00 0.78 – – 7.33 13.78 39.00

15.73 – – – – – 4.00 – –

Table 1. Density (birds/10 ha) of bird species in 26 different habitats in Teneriffe Island. In the lower part ofthe table are shown the main characteristics of these habitats. Small letters with habitat names (n, locatedin the northern Teneriffe; s, located in the south of the island): Am. Poorly vegetated alpine habitats; Rm.Shrubland of Spartocitysus supranubius on volcanic outcrops and lava fields; R. Shrubland of Spartocitysussupranubius; Mo. Montane shrublands in the western part of the island (Cistus spp., Echium spp., Sonchusspp., 1,100–1,400 m); TCn and TCs. Scrublands of several Euphorbia species in north or south of Teneriffe;TCxs. Dry scrublands of several Euphorbia species in southern Teneriffe; Txs. Dry scrublands of severalEuphorbia species in southern Teneriffe lacking Euphorbia candelabrum; Bn. Deep gullies covered byEuphorbia shrubs and some thermophilic trees and shrubs; Br. Heathlands of Erica arborea; FB."Monteverde" mainly composed of tall heaths and trees of Myrica faya; L. Laurel forests; Pn and Ps.Pinewoods of Pinus canariensis in north or south of Teneriffe; Pjn. Young pinewoods of Pinus canariensis;Pm. Ancient pinewoods of Pinus canariensis; Pa. High altitude pinewoods of Pinus canariensis; Pah.Grasslands; Es. Abandoned agricultural fields, poorly vegetated, located in southern Teneriffe; Pl. Bananaplantations; Cn and Cs. Mosaic of agricultural fields devoted to several crops in north or south of Teneriffe;Pbn and Pbs. Small villages in north or south of Teneriffe; Cun y Cus. Large cities in north or south ofTeneriffe; Densidad. Density; S0,5. Number of species with densities higher than 0.5 birds/10 ha;Diversidad. Diversity; HAS. Hectares censused; CHB. Herbaceous layer cover; CMAT. Shrub cover; HMAT.Average shrub height; CARB.Tree layer cover (trees higher than 3 m); HARB. Average height of tree layer;AGR. Percentage cover of ground devoted to agricultural use; CURB. Urban cover.

L Pn Pjn Ps Pm Pa

– + – 0.53 + –

– – – – 0.44 –

– – – 0.27 – 0.80

– – – – – –

– – – – – –

– – – – – –

– – – – – –

– – – – – –

– – – – – –

5.08 – – – – –

1.85 – – – – –

– – – – – 0.27

– – – – – –

– – – – – –

– 0.63 0.29 0.27 1.78 –

6.46 4.42 – 2.67 1.78 –

– – + 0.27 0.89 +

4.31 – – – – –

– 1.26 0.29 6.93 4.44 1.33

– – – – – –

– – – – – 0.53

– – – – – –

– – – – – –

– – – – – –

5.54 5.68 3.14 7.73 8.44 0.80

– – – – – –

– – – – – –


Am Rm R Mo TCn TCs TCxs Txs

Phylloscopus canariensis – 0.44 2.88 10.15 8.00 3.40 0.84 –

Psittacula krameri – – – – – – – –

Regulus teneriffae – – – – 0.42 – – –

Scolopax rusticola – – – – – – – –

Serinus canarius – – – 2.77 1.68 1.00 – –

Streptopelia decaocto – – – – – 0.20 – –

Streptopelia roseogrisea – – – – – – – –

Streptopelia turtur – – – 0.62 + 0.40 0.42 –

Sylvia atricapilla – – – – 2.11 0.20 – –

Sylvia conspicillata – – 0.32 – – 0.60 0.42 0.62

Sylvia melanocephala – – – 1.85 2.11 2.40 – –

Turdus merula – 0.11 – – 2.95 – – –

Upupa epops – – – – – – – +

Densidad 0.0 1.8 5.9 19.1 26.8 18.2 8.0 7.8

S0,5 0 1 2 5 10 8 2 4

Diversidad 0.00 1.52 1.23 1.33 2.23 2.11 1.15 1.33

HAS 27.5 90.0 62.5 32.5 47.5 50.0 47.5 65.0

Altitud 2593 2202 2200 1256 338 445 220 45

CHERB 0.0 0.0 0.0 9.4 26.6 2.6 0.6 3.7

CMAT 0.6 9.9 46.8 45.4 61.2 40.6 23.2 13.3

HMAT 0.1 0.8 1.1 1.0 0.8 0.6 0.5 0.5

CARB 0.0 0.0 0.0 1.2 0.3 0.0 0.0 0.0

HARB 0.0 0.0 0.2 2.3 1.2 0.0 0.0 0.0

AGROPEC 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

CURB 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Bn Br FB

10.67 14.67 15.00

– – –

– 12.22 10.33

– + +

1.60 2.44 –

– – –

– – –

1.60 – –

3.73 1.56 –

– – –

2.13 – –

3.73 12.67 14.33

– – –

41.4 65.8 61.7

12 8 9

2.18 1.90 1.86

37.5 45.0 30.0

189 915 1006

21.0 1.7 1.7

63.2 40.8 39.9

0.9 2.3 2.6

2.1 54.3 63.8

3.6 5.0 9.3

0.0 0.0 0.0

1.5 0.0 0.0

Tabla 1. (Cont.)

sidad ecológica máxima medida en los 26 mediosdistinguidos, y las configuraciones del paisaje quedeterminan las máximas abundancias de las espe-cies de aves en Tenerife (obtenidas mediante árbo-les de regresión). La combinación de estos resulta-dos con los de la tabla 1 perfilan los principalespatrones de preferencias de hábitat, abundancia yamplitud de distribución de las especies en Tenerife.A continuación se resaltan los principales resulta-dos relativos a la rareza de las aves en la isla.

Las especies observadas en más del 33% detodos los transectos efectuados fueron el Anthusberthelotii (Bisbita Caminero), Turdus merula (Mir-lo Común), Phylloscopus canariensis (MosquiteroCanario), Parus caeruleus (Herrerillo Común) ySerinus canarius (Canario). Otras especies am-pliamente distribuidas fueron Falco tinnunculus

(Cernícalo Vulgar), Streptopelia turtur, Motacillacinerea (Lavandera Cascadeña), Erithacus rubecula(Petirrojo), Sylvia atricapilla (Curruca Capirotada),Sylvia melanocephala (Curruca Cabecinegra) ySylvia conspicillata (Curruca Tomillera). Columbalivia (Paloma Bravía) y Passer hispaniolensis (Go-rrión Moruno) alcanzaron densidades ecológicasmáximas muy elevadas (aprox. 40 aves/10 ha).Streptopelia decaocto, Turdus merula, Erithacusrubecula, Regulus teneriffae (Reyezuelo Canario),Phylloscopus canariensis, Petronia petronia (Go-rrión Chillón), Serinus canarius y Miliaria calandra(Triguero) fueron también localmente muy abun-dantes (10–25 aves/10 ha). Por el contrario, lasespecies autóctonas con menores frecuencias deaparición (FREC<2,5%) en los censos fueronCoturnix coturnix (Codorniz Común), Accipiter nisus


L Pn Pjn Ps Pm Pa

13.69 1.47 0.86 2.40 4.44 2.13

– – – – – –

17.08 13.05 0.29 2.67 4.00 –

+ + – – – –

0.92 0.84 1.14 1.07 4.89 –

– – – – – –

– – – – – –

0.31 0.21 0.29 – 0.44 –

0.15 – – – – –

– – – – – –

0.15 – – – – –

13.69 1.47 – 0.53 – –

– – – – – –

69.2 29.1 6.3 25.3 31.6 5.9

9 8 3 8 8 5

1.98 1.63 1.49 1.85 2.00 1.61

65.0 47.5 35.0 37.5 22.5 37.5

888 1268 1640 1640 1653 1987

1.2 9.5 2.1 2.7 0.6 0.0

13.4 28.8 12.0 11.4 26.0 9.2

1.5 1.8 0.9 1.6 1.3 1.1

85.5 72.3 35.3 65.1 38.7 21.9

9.9 17.0 11.0 14.9 18.6 8.2

0.0 0.0 0.0 0.0 0.0 0.0

0.0 0.0 0.0 0.0 0.0 0.0

Pan Es Pl Cn Cs Pbn Pbs Cun Cus

0.53 – 4.00 15.51 5.94 4.89 8.33 6.00 3.60

– – – – – – – – 2.00

– – – – – – – 0.44 –

– – + – – – – – –

22.93 – 7.56 15.61 1.21 1.33 0.33 4.67 –

– – 0.44 0.29 – 3.56 4.33 11.56 19.00

– – – – – – – 3.11 4.40

– – 0.89 0.88 0.24 0.44 1.33 1.33 –

– – 1.33 3.32 1.45 3.11 1.67 0.67 1.00

1.07 0.20 1.78 0.20 2.18 – 0.33 – –

– 0.20 0.44 2.15 0.61 0.44 0.33 – –

0.53 – 3.11 6.93 0.36 2.22 5.67 5.11 0.80

– – 0.44 – – – 0.33 – 0.40

68.3 13.2 35.1 62.6 27.6 21.8 50.0 95.3 76.4

9 3 11 15 11 8 11 12 8

1.68 1.09 2.39 2.26 2.33 2.20 2.28 1.78 1.44

37.5 50.0 22.5 102.5 82.5 22.5 30.0 45.0 50.0

722 86 87 744 727 356 547 358 38

91.3 15.7 15.5 68.5 42.1 2.6 12.9 4.8 0.0

5.3 9.4 16.7 13.0 28.8 11.0 16.9 1.9 6.4

0.6 0.6 1.3 1.0 0.8 0.9 0.5 0.2 0.5

0.2 0.0 46.6 2.1 1.1 2.8 3.3 4.7 1.1

0.5 0.0 3.0 3.4 2.4 5.7 3.8 6.6 6.3

0.0 0.0 100.0 100.0 93.9 0.0 8.3 0.0 0.0

0.0 0.0 1.8 1.8 6.8 81.4 56.8 84.4 92.5

(Gavilán Común), Columba junoniae, Dendrocoposmajor (Pico Picapinos), Petronia petronia ,Bucanetes githagineus (Camachuelo Trompetero)y Carduelis cannabina (Pardillo Común). De lacombinación de los valores observados de fre-cuencia de aparición, la abundancia ecológicamáxima y la restricción en la distribución a unospocos hábitats, se obtiene que las especies másraras de la avifauna terrestre de Tenerife (nointroducidas recientemente) fueron Coturnixcoturnix, Alectoris barbara (Perdiz Moruna),Burhinus oedicnemus (Alcaraván Común), Accipiternisus, Columba junoniae, Upupa epops (Abubilla),Dendrocopos major, Corvus corax (Cuervo), Miliariacalandra, Petronia petronia, Bucanetes githagineus,Carduelis carduelis (Jilguero) y Carduelis chloris(Verderón Común).

La figura 3 ilustra los resultados de los árbolesde regresión de ocho especies/subespeciesendémicas de Canarias, ampliamente extendidas yrepresentativas de los principales modelos dedistribución de las aves terrestres en Tenerife. Acontinuación se exponen muy sucintamente losprincipales resultados relativos a la variación desus abundancias.

Phylloscopus canariensis es una especie deamplísima distribución en Tenerife, cuya variaciónde abundancia se ve influida principalmente por laposición geográfica en la isla, siendo la latitud lavariable que más le influye. En la mitad meridionalde la isla (LAT<3.136) es bastante más escasa queen la mitad septentrional, donde es especialmenteabundante en Anaga (LAT>3.153). En este sectorde Tenerife es donde la especie alcanza sus mayores


abundancias, especialmente cuando la altura delestrato arbustivo (HMAT; mayoritariamente brezos)es mayor de 1,9 m. En la mitad meridional deTenerife su abundancia está influida positivamentepor el desarrollo del estrato arbustivo (CMAT>26),siendo siempre más densa en el extremo másoccidental de la isla (LONG<330 en dos ramifi-caciones).

El desarrollo del estrato arbóreo (en cobertura–CARB– y altura –HARB–) y la existencia debarrancos (BCO) son las variables que más influyensobre la variación de la abundancia de Paruscaeruleus, otra especie de muy amplia distribuciónen Tenerife. Alcanza sus mayores densidades enmedios arbolados muy desarrollados (CARB>28% yHARB>16,5 m; 7,8 aves/km), o en barrancos conmuy alta cobertura de arbustos (BCO=1 y CMAT>78%;9,1 aves/km) en ausencia de bosques. En mediosarbustivos no localizados en barrancos, la especiepuede llegar a ser medianamente abundante si lacobertura de matorral es muy alta (>78%).

Turdus merula es más abundante en aquellasáreas con presencia de brezos (BRZ=1) y ausenciade pinares (PIN=0), en especial en el sector másseptentrional de Tenerife (LAT>3157; i.e., brezales,fayales y laurisilvas de Anaga donde alcanza unpromedio de 15,1 aves/km). Si no hay brezos en el

estrato arbustivo (BRZ=0), la especie es muchomás abundante en el tercio norte de la isla(LAT>3135), donde alcanza sus mayores den-sidades en medios con cobertura de matorral(CMAT) mayor del 68%, situados a altitudes (ALT)superiores a 330 m (11,2 aves/km). Si no se cumplenlos requisitos anteriores de presencia de brezos enel estrato arbustivo y elevadas coberturas dematorral, la tercera configuración del paisaje dondela especie tiene elevadas densidades es en áreasagrícolas (AGR=1) situadas por encima de 486 mde altitud (7,7 aves/km). En los dos terciosmeridionales de Tanerife (LAT<3135) es muy escaso(promedio de 0,4 aves/km).

La abundancia de Serinus canarius es mayoren zonas agropecuarias (AGR=1) que en otrosambientes naturales de Tenerife (los índiceskilométricos de abundancia –IKA– de las ramasderechas de su árbol de regresión son mayoresque las situadas a la izquierda). Dentro de laszonas agrícolas, sus mayores densidades se hanobservado en el cuadrante noroccidental de laisla (LONG<351 y LAT>3.135) a alt i tudessuperiores a 683 m. En áreas sin uso agrícola suabundancia es mayor en la mitad oriental de laisla (LONG>340) si la cobertura de arbustos(CMAT) es superior al 24%.

Fig. 2. Árbol de regresión de la variación espacial de la riqueza de especies en Tenerife. Los criterioshacen referencia a las ramas de la izquierda (los valores son contrarios en las ramas derechas). Lalongitud de las ramas es proporcional a la devianza explicada por cada criterio (i.e., a mayor longitud,mayor variabilidad explicada). Los valores de las puntas expresan el número medio de especiesobservado en 0,5 km de transectos lineales. Para el significado de las siglas véase la tabla 3.

Fig. 2. Regression tree analyzing the spatial variation in species richness in Teneriffe. The split criteriarefer to the left side of each dichotomy. Branch lengths are proportional to deviance explained by eachsplit criterium. Figures in the tips of the tree ("leaves") measure the average number of bird speciesobserved per 0.5 km of line transect (without limiting the detection belt). See table 3 for acronyms.

LAT<3135

ALT<1656

ALT<328

CURB<92 ALT<1100

CARB<22

ALT<1270CURB<58

LONG<316

1.3 3.3

2.4 5.0 5.5 7.3 4.3

5.17.1 8.6

5.15.1

AGR=0


Tabla 2. Abundancia relativa (aves/km) de las aves de Tenerife a lo largo de un gradiente altitudinalde seis bandas a intervalos de 500 m: 0–500 m (A); 500–1.000 m (B); 1.000–1.500 m (C); 1.500–2.000 m (D); 2.000–2.500 m (E); > 2.500 (F); AA. Amplitud de distribución altitudinal (mínima 0,17;máxima 1); Transectos. Número de transectos de 0,5 km de longitud con los que se han calculadolas abundancias relativas; Altitud media. Altitud media de los transectos efectuados en cada bandaaltitudinal; Especies/0,5 km. Riqueza específica en cada banda altitudinal.

Table 2. Relative abundance (birds observed per km of line transect) of birds in Teneriffe Island acrossan altitudinal gradient: six belts at 500 m interval: 0–500 m (A); 500–1,000 m (B); 1,000–1,500 m (C);1,500–2,000 m (D); 2,000–2,500 m (E); > 2,500 (F); AA. Altitudinal breadth (minimum 0.17; maximum 1).Transectos. Number of transects of 0.5 km censused in each altitudinal belt; Altitud media. Averagealtitude of transects within each belt. Especies/0,5 km. Average species richness (spp/0.5 km of linetransect) within each altitudinal belt.

Bandas altitudinales

A B C D E F AAAlectoris barbara 0.38 0.08 0.28 0.03 0.00 0.00 0.42Anthus berthelotii 3.24 2.27 0.50 0.03 1.38 0.50 0.58Bucanetes githagineus 0.14 0.00 0.00 0.00 0.00 0.00 0.17Burhinus oedicnemus 0.03 0.00 0.00 0.00 0.00 0.00 0.17Buteo buteo 0.06 0.10 0.06 0.00 0.00 0.00 0.48Carduelis cannabina 0.26 2.30 0.25 0.00 0.00 0.00 0.24Carduelis carduelis 0.10 0.05 0.00 0.00 0.00 0.00 0.31Carduelis chloris 0.02 0.28 0.03 0.00 0.00 0.00 0.23Columba bollii 0.09 0.47 0.44 0.00 0.00 0.00 0.39Columba junoniae 0.00 0.13 0.25 0.00 0.00 0.00 0.30Columba livia 4.91 3.20 0.38 0.03 0.00 0.00 0.35Coturnix coturnix 0.00 0.19 0.00 0.00 0.00 0.00 0.17Dendrocopos major 0.00 0.00 0.09 0.48 0.00 0.00 0.23Erithacus rubecula 0.37 2.23 4.06 1.01 0.34 0.00 0.47Falco tinnunculus 0.39 0.30 0.06 0.24 0.10 0.25 0.80Fringilla coelebs 0.02 0.92 0.47 0.00 0.00 0.00 0.31Fringilla teydea 0.00 0.00 0.53 1.81 0.08 0.00 0.27Lanius excubitor 0.16 0.00 0.00 0.00 0.18 0.00 0.33Miliaria calandra 0.00 1.17 0.00 0.00 0.00 0.00 0.17Motacilla cinerea 0.32 0.42 0.09 0.00 0.00 0.00 0.40Parus caeruleus 1.70 2.30 3.63 3.57 0.05 0.00 0.62Passer hispaniolensis 3.18 0.79 0.00 0.00 0.00 0.00 0.24Petronia petronia 0.13 0.69 0.00 0.00 0.00 0.00 0.23Phylloscopus collybita 4.30 8.82 6.25 2.05 1.19 0.00 0.60Regulus teneriffae 0.06 2.24 3.50 1.01 0.00 0.00 0.42Serinus canarius 0.99 6.01 2.59 1.73 0.00 0.00 0.46Streptopelia decaocto 2.05 0.16 0.00 0.00 0.00 0.00 0.19Streptopelia roseogrisea 0.43 0.00 0.00 0.00 0.00 0.00 0.17Streptopelia turtur 0.48 0.59 0.47 0.43 0.00 0.00 0.66Sylvia atricapilla 1.25 1.43 0.16 0.00 0.00 0.00 0.37Sylvia conspicillata 0.41 0.45 0.03 0.00 0.08 0.00 0.41Sylvia melanocephala 0.69 0.66 0.34 0.00 0.00 0.00 0.46Turdus merula 1.28 5.84 3.97 0.24 0.05 0.00 0.42Upupa epops 0.05 0.02 0.00 0.00 0.00 0.00 0.28Altitud media 192 746 1216 1720 2187 2644Transectos 185 183 64 75 77 8Especies/0,5 km 4.8 6.7 5.0 3.4 1.2 0.2


Tabla 3. Patrones de distribución y abundancia de las especies de aves terrestres en Tenerife: Dmax.Densidad (aves/10 ha) maxima medida en Tenerife en los 26 hábitats de la tabla 1; Frec. Frecuencia deaparición de las especies en los transectos de 0,5 km de longitud (en %); D2%. Devianza explicado porlos árboles de regresión (en %); para cada especie también se proporcionan las características delespacio que describen los lugares donde son más abundantes, teniendo en cuenta los criterios de losárboles de regresión que conducen a las puntas con máxima densidad; IKA. Medida de abundanciamáxima de cada especie en las configuraciones del hábitat definidas por los árboles de regresión (avesobservadas/km de transecto; n. Número de muestras de 0,5 km con las cuales se obtiene el IKA). Trascada sigla se indica el criterio cuantitativo que define el hábitat (+, presencia de ese atributo; –, ausenciadel atributo): LAT, LONG. Latitud y longitud en coordenadas UTM (en kilómetros dentro del bloque 28R);ALT. Altitud (en m); CHB. Cobertura de herbáceas (en %); CMAT. Cobertura de arbustos (en %); HMAT.Altura media del matorral (en m); CARB. Cobertura de arbolado (en %); HARB. Altura media del arbolado(en m); AGR+. Suelo dedicado a la agricultura; LAUR+. Monteverde; PIN+. Pinares de Pinus canariensis;PIN–. Ausencia de pinares; BRZ+. Presencia de brezos en el estrato arbustivo; BRZ–. Ausencia debrezos; BCO+. Presencia de barrancos; CURB. Cobertura de suelo urbano (en %).

Table 3. Distribution and abundance of terrestrial birds in Teneriffe Island: Dmax. Maximum density(birds/10 ha) measured in 26 hábitats in Teneriffe (table 1); Frec. Frequency of occurrence of eachspecies in 592 transects of 0.5 km spread out in Teneriffe (in %); D2%. Deviance explained byregression trees analyzing the spatial variation of relative abundance of birds in line transects of 0.5km (in %); the criteria of tree regression analyses defining the spatial configuration of the areas wherethe species are more abundant is provided for each species; for some species it was not possibleto obtain regression models due to sample limitations; IKA. Average abundance of the species in theleaves defined by regression trees (birds per km of lineal transect; n. Number of transects of 0.5 kmwith which IKA’s are obtained). Quantitative criteria are provided after acronyms (+, presente of thathabitat attribute; – absence of the attribute): LAT, LONG. Latitude and longitude in UTM coordinates(in km with UTM block 28R); ALT. Altitude above the sea level (in m); CHB. Cover of the herbaceouslayer (in %); CMAT. Shrub cover (in %); HMAT. Average height of shrubs (in m); CARB. Cover ofthe tree layer (in %); HARB. Average height of the tree layer (in m); AGR+. Presence of groungdevoted to agriculture; LAUR+. "Monteverde" (laurel forest and/or woodlands mainly composed by tallheaths and Myrica faya trees); PIN+. Pinewood of Pinus canariensis; PIN–.Absence of pinewoods;BRZ+. Presence of Erica arborea in the shrub layer; BRZ–. Absence of Erica arborea in the shrublayer; BCO+. Deep gullies; CURB. Urban cover (in %).

Dmax Frec D2% IKA n

Accipiter nisus 0.53 0.5

Alectoris barbara 3.00 3.7 15.4 CHB>48, LAT<3124 6.9 7

Anthus berthelotii 9.07 39.9 47.6 HARB<4.2, LAT<3116, CHB>32 9.8 10

Bucanetes githagineus 0.31 0.8 22.0 LAT<3103, ALT<32 1.7 7

Burhinus oedicnemus 0.21 0.5

Buteo buteo 0.11 2.2 22.5 LONG>380, LAT<3158 1.4 7

Carduelis cannabina 7.47 14.5 54.3 CHB>42, AGR+, LAT>3140 9.5 11

Carduelis carduelis 2.67 0.8 13.7 AGR+, CARB>34 1.7 7

Carduelis chloris 1.11 2.4 24.6 LAT>3140, CHB>65 2.5 12

Columba bollii 5.08 5.1 57.0 LAUR+, LAT<3158, HARB>11.5 4.8 8

Columba junoniae 1.85 1.5 35.9 CARB>92, LAT<3136 3.2 10

Columba livia 44.89 23.5 48.5 CURB>4, LAT>3144, LONG<371 54.0 7

Corvus corax 0.05 0.3

Coturnix coturnix 0.59 1.5 27.1 CHB>68, LAT>3140 1.8 18

Dendrocopos major 1.78 2.4 44.6 PIN+, HARB>19.5 2.8 8

Erithacus rubecula 11.56 23.3 64.3 BRZ+, HMAT>1.7, CARB<72, LAT>3137 11.6 15

Falco tinnunculus 0.89 11.3 16.2 ALT<819, CURB<5.5, LAT<3102 1.8 8

Fringilla coelebs 4.31 9.0 31.8 BRZ+, PIN-, HARB>9.5 3.6 18


Anthus berthelotii es una especie generalista demedios deforestados, que se ve influida principalmentepor el desarrollo del arbolado (negativamente), laposición latitudinal en la isla (más abundante en elsur que en el norte) y el desarrollo del estrato herbáceo(positivamente). Es muy escasa (0,3 aves/km) enáreas arboladas (altura de árboles > 4,2 m). En losmedios con muy poco desarrollo del arbolado(< 4,2 m) es más abundante en el tercio meridionalde la isla (LAT<3116; ver localización en la figura 1),que en el norte. Su abundancia aumenta con eldesarrollo del estrato herbáceo, aunque requieremás cobertura de este estrato en el norte (68% aLAT>3116) que en el sur de Tenerife (32% cuandoLAT<3116). En las áreas con poco desarrollo delarbolado del norte de la isla, donde la cobertura deherbáceas es mayor del 68%, la especie esconsiderablemente más abundante en la mitadoccidental (LONG<345).

Regulus teneriffae es una especie forestalgeneralista cuya abundancia está supeditadaprincipalmente a la presencia de un estrato arbustivode brezos bien desarrollado. Si no existe brezo(BRZ=0), la especie sólo estará presente, con

relativamente bajas abundancias, si el estrato arbóreoestá muy desarrollado (HARB>16,5 m; 3,4 aves/km)o la cobertura de arbustos es elevada (CMAT>48%;1,4 aves/km). En hábitats con presencia de brezos(BRZ=1), Regulus teneriffae alcanza sus mayoresabundancias (>7,9 aves/km) en el tercio norte de laisla (LAT>3144), aunque su densidad decrece en lapenínsula de Anaga (LONG>375).

La densidad de Fringilla teydea (Pinzón Azul)alcanza su máximo en Tenerife (5,2 aves/km) enlos bosques con árboles de más de 13,5 m dealtura localizados a más de 1594 m de altitud en lamitad oriental de la isla (LONG>346). Estosambientes coinciden exclusivamente con los pinaresde Pinus canariensis del sector oriental de la co-rona forestal de Tenerife. Cuando el arbolado noestá muy desarrollado (HARB<13.5 m) la especiesestá virtualmente sólo presente en pinares (PIN=1),siendo considerablemente más abundante en losorientales (LONG>350; 2,5 aves/km) que en losoccidentales (promedio de 0,5 aves/km).

Fringilla coelebs (Pinzón Vulgar) es una especietípica de monteverde, cuya abundancia se ve influi-da principalmente por la presencia de brezos en el

Fringilla teydea 6.93 8.4 62.7 HARB>13.5, ALT>1594, LONG>346 5.2 12

Gallinula chloropus 0.53 0.2

Lanius excubitor 0.77 2.7 18.2 LAT<3102, HMAT>0.65 1.2 8

Miliaria calandra 10.40 5.1 58.9 CHB>88, CMAT<4, LAT>3137 12.3 7

Motacilla cinerea 1.78 9.0 35.7 CURB 1-52, CHB<51, CMAT<28, LON<352 3.2 10

Myiopsitta monachus 0.44 0.2

Parus caeruleus 8.44 42.1 41.9 CARB<28, BCO+, CMAT>78 9.1 11

Passer hispaniolensis 39.00 7.9 52.6 CURB>82, LAT<3118, CMAT<1 36.3 7

Petronia petronia 15.73 0.8 16.2 CHB>92.5, LAT<3136 16.3 7

Phylloscopus canar. 15.51 67.1 48.9 LAT>3153, HMAT>1.9 19.4 14

Psittacula krameri 2.00 0.5

Regulus teneriffae 17.08 16.6 75.2 BRZ+, LAT>3144, LONG<375 15.2 13

Serinus canarius 22.93 32.4 52.2 AGR+, LONG<351, LAT>3135, ALT>683 18.1 37

Streptopelia decaocto 19.00 6.9 60.0 CURB>94, ALT<35 23.2 8

Streptopelia roseogrisea 4.40 2.4 43.2 CURB>94, ALT<32 8.0 7

Streptopelia turtur 1.60 14.2 31.7 CHB>8, CARB>0.5, LAT>3151 3.4 14

Sylvia atricapilla 3.73 23.0 43.3 LAT>3136, CARB<32, HMAT>0.6 5.1 7HARB<6, ALT<180

Sylvia conspicillata 2.18 8.6 22.7 CHB>8, LAT<3137, LONG>358, HMAT>0.8 3.7 7

Sylvia melanocephala 2.40 14.4 31.6 ALT<1045, CMAT<29, AGR+, LONG>319 3.7 7CARB<0.5, CHB>88

Turdus merula 14.33 38.3 68.8 BRZ+, PIN-, LAT>3157 15.1 18

Upupa epops 0.44 1.0

Tabla 3. (Cont.)

Dmax Frec D2% IKA n


estrato arbustivo. Incluso existiendo brezos, la es-pecie está virtualmente ausente de pinares (BRZ=1y PIN=1; 0,0 aves/km). En el monteverde, su abun-dancia se asocia positivamente con el desarrollodel estrato arbóreo, tanto en altura (HARB>9,5 m;3,6 aves/km), como en cobertura si los árboles noson muy altos (CARB>55% cuando HARB<9,5 m;2,1 aves/km). En ausencia de brezos como plantadominante del estrato arbustivo, la especie sólo esmedianamente abundante en el norte de la penín-sula de Anaga (LAT>3.158; 1,7 aves/km) dondepuede ser observado en tabaibales–cardonales biendesarrollados (ver CULTn en tabla 1). También estápresente en algunas áreas agrícolas cuando éstastienen arbolado disperso desarrollado (AGR=1 yHARB>6,5 m; 1,5 aves/km).

Rareza, endemicidad e impacto humano

Varias especies autóctonas de la avifauna de Tenerifeson hoy marcadamente más escasas de lo quefueron hace 25–50 años (véase Martín, 1987; Martín& Lorenzo, 2001). Entre ellas se encuentran Upupaepops, Coturnix coturnix, Calandrella rufescens (Te-rrera Marismeña; no observada durante este estu-dio), Corvus corax, Petronia petronia, Cardueliscannabina, Carduelis carduelis y Miliaria calandra. Aellas hay que añadir los ya extintos Milvus milvus(Milano Real) y Neophron percnopterus (Alimoche).Por la información recopilada por diferentesornitólogos en la primera mitad del siglo XX (verrevisión de Martín & Lorenzo, 2001) todas ellasfueron abundantes, llegando a alcanzar elevadosefectivos poblacionales en zonas rurales y sus áreasagropecuarias colindantes, siendo escasas enhábitats naturales con poca influencia humana. Comoconsecuencia de los cambios en los usos tradiciona-les del suelo de los últimos 25–50 años (reducción oabandono de la agricultura de subsistencia, recupe-ración de la vegetación autóctona de tabaibales,matorrales de medianía y brezales, reducción de lacabaña ganadera de cabras, implantación de mono-cultivos industriales de plátano en áreas bajas cer-canas a la costa, urbanismo masivo) y el efecto delos plaguicidas usados en las décadas de 1950–1960, sus máximos efectivos poblacionales se hanreducido. Consistentemente con lo descrito en el

pasado (primeros 75 años del siglo XX; Martín &Lorenzo, 2001), en este estudio sus máximas densi-dades no se han medido en hábitats autóctonospoco degradados, sino en medios fuertementeimpactados por el hombre: mosaicos de cultivos,plataneras, pastizales y núcleos urbanos pequeños.Por otro lado, las últimas poblaciones de Calandrellarufescens de Tenerife (menos de 100 parejas) seasentaban en hábitats profundamente degradados(pastizales del aeropuerto de Los Rodeos, campo degolf, áreas de cultivo de tomate abandonados; Mar-tín & Lorenzo, 2001). La casi total destrucción por elhombre del hábitat original de estas especies en elpiso basal de la isla (zonas estepáricas o áreastermófilas de medianía) quizás las hizo desplazarsea unos hábitats secundarios muy alterados dondellegaron a ser abundantes y se encuentran hoy(áreas agrícolas). Por tanto, lo que observamos hoydía sería una disminución de sus tamañospoblacionales respecto a sus efectivos de hace unos100–50 años, como consecuencia de los cambiosen los usos del suelo (nuevas alteraciones y aban-dono de prácticas tradicionales; ver también Tucker& Heath, 1994 para la avifauna europea), pero muyposiblemente sus poblaciones ya venían mermadasdesde un pasado.

Esto no parece haber ocurrido con las especiesmás claramente endémicas cuyas preferencias dehábitat se establecen en medios autóctonos. Inclu-so especies estenoicas que en el pasado fueronconsideradas muy escasas, hoy día aumentan susefectivos paralelamente a la recuperación de sushábitats. Este es el caso de las palomas endémi-cas Columba bollii y Columba junoniae propias delaurisilva y de los ya desaparecidos bosquestermófilos (Martín & Lorenzo, 2001). También es elcaso de otras dos especies propias de pinaresautóctonos de Pinus canariensis: Dendrocopos majorcanariensis y Fringilla teydea teydea (Martín & Lo-renzo, 2001). Durante los últimos 20 años, la den-sidad de Dendrocopos major ha aumentadosustancialmente en numerosas zonas de pinar don-de antes no existía. Así, en pinares censados en1986 por Carrascal (1987) donde la especie noestaba presente (pinares por encima del valle de laOrotava), en la actualidad se han medido densida-des de 0,2–1 aves/10 ha. Las únicas especies

Fig. 3. Árboles de regresión para 8 especies / subespecies endémicas ampliamente distribuidas enTenerife. Los criterios hacen referencia a las ramas de la izquierda (los valores son contrarios en lasramas derechas). La longitud de las ramas es proporcional a la devianza explicada por cada criterio(i.e., a mayor longitud, mayor variabilidad explicada). Los valores de las puntas expresan laabundancia de las especies en aves/km. Para el significado de las siglas véase la tabla 3.

Fig. 3. Regression trees analyzing the spatial variation of abundance in 8 endemic species/subspecieswidely distributed in Teneriffe. The split criteria refer to the left side of each dichotomy. Branch lengthsare proportional to deviance explained by each split criterium. Figures in the tips of the trees (‘leaves’)measure the abundance expresed in birds per 1 km of line transect (without limiting the detection belt).See table 3 for acronyms.


Parus caeruleus

Serinus canarius

Anthus berthelotii

Fringilla coelebs

Phylloscopus canariensis

Turdus merula

Regulus teneriffae

Fringilla teydea

CARB<28

BCO=0

CMAT<78ALT<1974

ALT<174

CMAT<78

HARB<16.5

LONG<385COD=0

2.7 4.6 7.17.8

0.2 1.60.05

4.3 4.6 9.1

LAT<3136

LAT<3153CMAT<26

LONG<328 LONG<330 LONG<316 HMAT<1.9

AGR=0

CMAT<55

6.0 1.4 10.6 4.5 0.6

4.1 11.4

11.4

10.2 19.4

BRZ=0

LAT<3135

LAT<3157

PIN=0

CMAT<68AGR=0

LONG<317ALT<486

0.4

1.80.7

1.3 7.7

4.9 11.210.7 15.1 2.3

AGR=0

CMAT<24 LONG<340

LAT<3135

LONG<351

ALT>683

0.5 1.1 3.6

6.1

7.4 18.1

3.0

HARB<4.2

LAT<3116

CHB<32LONG<332

CHB<68

LONG<3452.3 5.4

9.8

1.46.8 2.7

0.3

BRZ=0

HARB<16.5 LAT<3144

HMAT<1.7 LONG<375

CARB<480.1 1.4 3.4

2.4 7.215.2 7.9

HARB<13.5

PIN=0

LONG<350LAT<3143

ALT<1594

LONG<346

2.0 5.2

0.6 2.60.5 2.5

0.01

BRZ=0

LAT<3158 PIN=0

HARB<9.5

CARB<55

AGR=0 HARB<6.50.0

0.2 1.5

1.7

1.1 2.1

0.0

3.6


endémicas de hábitats autóctonos que en la actua-lidad son mucho más escasas que en el pasado sonBurhinus oedicnemus distinctus y Bucanetesgithagineus amantum. Este declive se ha debidoprincipalmente a la alteración de hábitats semi-desérticos cercanos a la costa del sur de la isla(urbanismo y monocultivos industriales de plátano ytomate; Martín & Lorenzo, 2001).

A pesar de su carácter artificial, producto de ladegradación de formaciones vegetales autóctonas,las áreas agrícolas de Tenerife mantienen altos efec-tivos poblacionales de numerosas especies de avesen Tenerife. Así, Motacilla cinerea y Streptopelia turtur(Tórtola Europea) alcanzan en estos ambientes den-sidades tan altas como las observadas en formacio-nes autóctonas con casi nula influencia humana. Losescasos pastizales del norte de la isla sustentan laspoblaciones más densas de Carduelis cannabina y dealgunas otras especies bastante raras en Tenerife:Miliaria calandra, Petronia petronia y Carduelis chloris.Los mosaicos agrícolas de norte de la isla, mantienenáreas de praderas suficientemente extensas comopara albergar las últimas poblaciones densas deCoturnix coturnix. Las plataneras y áreas semi–urba-nas dispersas del norte de la isla acogen los mayoresefectivos de Carduelis carduelis.

Las zonas urbanas de Tenerife se caracterizanpor acoger una considerable riqueza y diversidad deespecies de aves. Así, los valores urbanos de estosparámetros son similares o incluso más altos quelos medidos en formaciones vegetales autóctonas(i.e., matorrales de montaña, tabaibales–cardonales,laurisilva y pinares maduros), y atípicos si los com-paramos con lo observado en áreas continentales,tanto del paleártico como del neártico (Melles et al.,2003; Marzluff et al., 2001; Clergeau et al., 1998).Este fenómeno se puede explicar atendiendo a lapermeabilidad que tienen las zonas urbanastinerfeñas para captar numerosas especies de laavifauna de su entorno. Así, las ciudades y pueblosde Tenerife acogen elevadas densidades de espe-cies que en Europa son raras en medios urbaniza-dos (e.g., Streptopelia turtur, Upupa epops, Motacillacinerea), o menos frecuentes que lo observado enTenerife (e.g. Turdus merula, Parus caeruleus,Phylloscopus canariensis, Sylvia atricapilla y Serinuscanarius; Fernández–Juricic, 2000; Jokimäki, 1999;Palomino y Carrascal, en preparación). Este hechoincrementa la variabilidad, y con ello la diversidad yriqueza, del grupo de especies más típicamenteurbanas (i.e., Columba livia, Passer hispaniolensis, ylas recientes colonizadoras durante los últimos 30años Streptopelia decaocto, S. roseogrisea, Carduelischloris y Sturnus vulgaris–Estornino Pinto).

Considerando estos patrones de distribución, pro-ponemos una hipótesis que vincula el carácter deendemicidad de las poblaciones de aves de Tenerifecon las preferencias de hábitat de las especies,atendiendo a la ocupación preferente de medios confuerte impacto humano frente a aquellos autóctonosno degradados. Esta hipótesis predice que es másprobable que hoy existan poblaciones tinerfeñas deespecies claramente diferenciadas (i.e., subespecies

o especies distintas) si su distribución y máximos deabundancia coinciden con formaciones vegetalesclimácicas (retamares altimontanos, matorralesmontanos, tabaibales–cardonales, pinares madurosy laurisilva) y no penetran medios antrópicos a noser que tengan elevadas amplitudes de hábitat. Porel contrario, sería poco probable que especies queson muy escasas o están ausentes de hábitatsautóctonos poco modificados y cuyas densidadesecológicas máximas están en medios producto de ladegradación ambiental (áreas agropecuarias y urba-nas) tuviesen poblaciones en Tenerife que hayansubespeciado respecto a las poblaciones continen-tales. La base teórica de este fenómeno debe en-contrarse en la hipótesis del "ciclo del taxón" (Ricklefs& Cox, 1978; Williamson, 1981; Ricklefs &Bermingham, 1999), que postula que las poblacio-nes insulares tienden a lo largo del tiempo a ladiferenciación especializándose en ambientes isle-ños concretos (ver no obstante Wiens, 1989 parauna controversia respecto a esta hipótesis). Esto sedebe a que la diferenciación taxonómica de pobla-ciones implica unas escalas de tiempo muy superio-res a las implicadas en la transformación históricade hábitats naturales por el hombre (últimos 2.000años, frente a las decenas de miles de años reque-ridos en procesos de especiación en aves –Klicka &Zink, 1997). Ejemplos del primer grupo de especies(endémicas distribuidas en medios autóctonos) se-rían Accipiter nisus granti, Burhinus oedicnemusdistinctus, Dendrocopos major canariensis, Erithacusrubecula superbus, Regulus [regulus] teneriffae,Fringilla coelebs canariensis, Phylloscopus [collybita]canariensis, Parus caeruleus teneriffae. La únicaexcepción a este patrón parece ser Motacilla cinereacanariensis (muy probablemente debido a la des-aparición de casi todos los cursos de agua naturalesen los barrancos debido a las canalizaciones actua-les; Martín & Lorenzo, 2001). Evidencias del segun-do grupo de especies serían aquellas cuyo estatustaxonómico de subespecies endémicas quedan des-cartados considerando recientes análisis de taxono-mía molecular (Calandrella rufescens rufescens,Coturnix coturnix confisa–Consejería de Política Te-rritorial y Medio ambiente, J. M. Naranjo, com.pers.). La validez y grado de generalización de estahipótesis podrá ser comprobado una vez que seandesarrollados estudios taxonómicos exhaustivos consólidas bases moleculares y morfométricas. La com-binación del conocimiento del estatus taxonómicode las poblaciones insulares, junto con sus preferen-cias de hábitat y ocupación de medios autóctonosvs. degradados por la acción humana servirá paradefinir prioridades de conservación a escala delarchipiélago (Dennis, 1997; Thomas et al., 1999;Gordon & Ornelas, 2000; Sangster, 2000).

Tenerife vs. península ibérica: compensación dedensidades y amplitud de distribución altitudinal

Debido a la menor cantidad de especies presentesen las islas que en el continente, varios autoreshan postulado que en las islas disminuye la presión


de depredación (i.e., menos depredadores presen-tes) y se reduce el solapamiento interespecífico delos nichos ecológicos, mitigándose considerable-mente el efecto de la competencia interespecífica.En este escenario de reducción de potencialesdepredadores y competidores, las poblaciones deespecies insulares aumentarían su abundancia res-pecto al continente, produciéndose el fenómenodenominado "compensación de densidad". No obs-tante, este incremento de abundancia también po-dría darse como consecuencia de que la dispersiónen las islas está limitada debido al efecto barreraimpuesto por el mar y la existencia de hábitatsdesfavorables dentro de las islas (Diamond, 1970;Yeaton, 1974; Emlen, 1979; Crowell, 1983; Wiens,1989). Este fenómeno ha sido observado en elsuroeste del Paleártico Occidental con la avifaunade Córcega (Blondel et al., 1988).

Utilizando los valores de densidades ecológicasmáximas obtenidos en este trabajo (tabla 3), y losobtenidos para la península ibérica (Martí & DelMoral, 2003; textos de cada especie y datos delapéndice 1) se puede comprobar si esta hipótesis secumple de modo generalizado con la avifauna deTenerife. Las únicas especies que alcanzandensidades máximas sustancialmente más elevadasen Tenerife que en la península ibérica (valoresexpresados en aves/10 ha) son: Anthus berthelotii(9,1 frente a 2,4 de Anthus campestris–BisbitaCampestre; ver Voelker, 1999 para una justificacióndel parentesco filogenético entre estas dos especies),Sylvia conspicillata (3,7 vs. 2.08), Petronia petronia(15,7 vs. 1,7) y Serinus canarius (22,9 vs. 13,3 deSerinus serinus–Verdecillo en la península). Lasdensidades ecológicas máximas en Tenerife de lasespecies/subespecies endémicas Lanius excubitor[meridionalis] koenigi (Alcaudón Real; 0,77 aves/10 ha), Regulus teneriffae (17,1), Phylloscopuscanariensis (15,5) y Carduelis cannabina meadewaldoi(7,5) son muy parecidas a las obtenidas por susequivalentes congenéricos en Iberia (Lanius excubitor[meridionalis] 0,74 aves/10 ha; Regulus regulus–Reyezuelo Sencillo 14,8 aves/10 ha, ver Sturmbaueret al., 1998 para una justificación del parentescofilogenético con Regulus teneriffae; Phylloscopusibericus–Mosquitero Ibérico 13,4 aves/10 ha;Carduelis cannabina: 7,7 aves/10 ha). Otras especieso subespecies tinerfeñas claramente diferenciadasde las formas continentales muestran densidadesconsiderablemente menores en Tenerife (valoresexpresados en aves/10 ha en Tenerife vs. penínsulaibérica): Dendrocopos major canariensis (1,78 vs.2,4 como media de los tres valores máximos medidosen Iberia), Motacilla cinerea canariensis (1,8 vs 3,0),Erithacus rubecula superbus (11,6 vs. 22,2), Turdusmerula cabrerae (14,3 vs 19,6), Sylvia melanocephala(2,4 vs. 15,3), Parus caeruleus teneriffae (8,4 vs.25,5), Fringilla coelebs canariensis y F. teydea (4,3 y6,9, respectivamente, vs. 23,2 de F. coelebs ibéricos).Esto mismo ocurre con especies de Tenerife noclaramente diferenciadas taxonómicamente de susequivalentes ibéricos: Sylvia atricapilla (3,7 vs. 12,2),Carduelis carduelis (2,7 vs. 7,8) y Carduelis chloris

(1,1 vs. 10,4). Por tanto, los datos no parecenapoyar de modo generalizado la hipótesis decompensación de densidades en la avifauna deTenerife.

Un paradigma clásico de la biogeografía insulares la expansión del nicho ecológico de las especiesen las islas respecto al continente, aunque existenevidencias tanto a favor como en contra de estahipótesis (véase Blondel, 1979; Wiens, 1989 y refe-rencias allí dadas). Para la isla de Córcega, Blondelet al. (1988) y Prodon et al. (2002) encuentranevidencias tanto apoyando como rechazando estahipótesis: incremento de la amplitud de hábitat enalgunas especies (p.e., Parus spp.), y contraccióngeneralizada en la amplitud de distribución altitudinal,aunque las subespecies endémicas manifiestan unincremento de ésta, expandiéndose por las zonaselevadas de la isla.

La ampliación del espectro de hábitats ocupadoses un hecho muy característico en unas pocasespecies (Parus caeruleus, Phylloscopus canariensis,Regulus teneriffae, Sylvia atricapilla, Serinus canarius;tabla 1), que ocupan distintos tipos de bosques,tabaibales, áreas agrícolas y zonas urbanas. Porotro lado, se produce una expansión de las preferen-cias de hábitat de las aves forestales hacia mediosestructuralmente más simples (caso de Paruscaeruleus, Phylloscopus canariensis y Sylviaatricapilla), aspecto que también se ha observado enla avifauna de Córcega (ver Blondel, 1979; Blondelet al., 1988 para una discusión de este tema). Noobstante, en el resto de las especies no se observatal incremento de su amplitud de hábitat en Tenerife(ver en Martí & Del Moral, 2003 los gráficos dedistribución entre hábitats de las especies de Passe-riformes en la península ibérica).

En Tenerife destacan los enormes rangosaltitudinales de algunas especies que se distribuyendesde el nivel del mar hasta 2.500 m de altitud (e.g.,Falco tinnunculus, Anthus berthelotii, Lanius excubitor[meridionalis], Erithacus rubecula, Turdus merula,Phylloscopus canariensis, Sylvia conspicillata, Paruscaeruleus; ver tabla 2 y Martín, 1987). Las amplitu-des de distribución altitudinal de estas especies sonmuy grandes y aparentemente mayores a las obser-vadas en la península ibérica (ver para comparaciónSánchez, 1991; Pleguezuelos, 1992; Martí & DelMoral, 2003). Todas estas especies que manifiestangrandes rangos de distribución son subespecies oespecies endémicas, algo parecido a lo obtenido porProdon et al. (2002) en Córcega. Otra especie quepresenta en Tenerife una gran distribución altitudinal,superior a la observada en la península esStreptopelia turtur (ver datos en la tabla 2 y enSánchez, 1991; Pleguezuelos, 1992). Sin embargo,hay otras especies y subespecies endémicas quemuestran una reducida dispersión altitudinal ya queestablecen sus preferencias de hábitat en formacio-nes ambientales muy localizadas (e.g, Burhinusoedicnemus y Bucanetes githagineus en tabaibalesxéricos; Columba bollii, Columba junoniae y Fringillacoelebs en laurisilvas; Dendrocopos major y Fringillateydea en pinares; Miliaria calandra y Coturnix


coturnix en praderas húmedas bajo influencia de losalisios). Otras especies distribuidas altitudinalmentede modo relativamente amplio en la península ibéri-ca, tienen en Tenerife un rango altitudinal bastanterestringido (e.g., Upupa epops, y las tres especiesdel género Carduelis; véase para comparaciónPleguezuelos, 1992). Por tanto, los resultados obte-nidos en Tenerife no parecen apoyar de modo claroy generalizado la hipótesis de expansión de distribu-ción en las poblaciones insulares. (véase Prodon etal., 2002 para un resultado similar obtenido con laavifauna de Córcega).

Agradecimientos

Este manuscrito se ha beneficiado de los comen-tarios de Alfredo Valido, Álvaro Ramírez y MarioDíaz.

Referencias

Anónimo, 1980. Atlas Básico de Canarias.Interinsular Canaria, Sta. Cruz de Tenerife.

Báez, M., 1992. Zoogeography and evolution of theavifauna of the Canary Islands. Natural HistoryMuseum of Los Angeles County Science Series,36: 425–431.

Bibby, C. J., Burgess, N. D., Hill, D. A. & Mustoe, S.H., 2000. Bird census techniques (2nd edition).Academic Press, London.

Blondel, J., 1979. Biogéographie et ecologie.Masson, Paris.

Blondel, J., Chessel, D. & Frochot, B., 1988. Birdspecies impoverishment, niche expansion, anddensity inflation in Mediterranean island habi-tats. Ecology, 69: 1899–1917.

Brown, J. H. & Lomolino, M. V., 1998. Biogeogra-phy. Sinauer, Sunderland.

Carrascal, L. M. & Palomino, D., 2002. Deter-minantes de la riqueza de especies de aves enlas islas Selvagem y Canarias. Ardeola, 49:211–221.

Carrascal, L. M., 1987. Relación entre avifauna yestructura de la vegetación en las repoblacionesde coníferas de Tenerife (Islas Canarias). Ardeola,34: 193–224.

Carrascal, L. M., Tellería, J. L. & Valido, A., 1992.Habitat distribution of Canary chaffinches amongislands: competitive exclusion or species–spe-cific habitat preferences? Journal of Biogeogra-phy, 19: 383–390.

Clergeau, P., Savard, J. P. L., Mennechez, G. &Falardeau, G., 1998. Bird abundance and diver-sity along an urban–rural gradient: a compara-tive study between two cities on different conti-nents. Condor, 100: 413–425.

Collar, N. J., Crosby, M. J. & Stattersfield, A. J.,1994. Birds to Watch 2. Birdlife International–Smithsonian Institution Press, Washington, D.C.

Crowell, K. L., 1983. Islands–insight or artefacts?Population dynamics and habitat utilization in

insular rodents. Oikos, 41: 442–454.De’Ath, G. & Fabricius, K. E., 2000. Classification

and regression trees: a powerful yet simple tech-nique for ecological data analysis. Ecology, 81:3178–3192.

Dennis, R. L. H., 1997. An inflated conservation loadfor European butterflies: increases in rarity andendemism accompany increases in species rich-ness. Journal of Insect Conservation, 1: 43–62.

Diamond, J. M., 1970. Ecological consequences ofisland colonisation by southwest Pacific birds, II.The effect of species diversity on total populationdensity. Proceedings of the National Academy ofSciences (USA), 67: 1715–1721.

Emlen, J. T., 1979. Land bird densities on BajaCalifornia Mexico islands. Auk, 96: 152–167.

Fernández–Juricic, E., 2000. Bird community com-position patterns in urban parks of Madrid: therole of age, size and isolation. Ecological Re-search, 15: 373–383.

Fernández–Palacios, J. M. & Andersson, C., 1993.Species composition and within archipelago co–occurrence patterns in the Canary Islands.Ecography, 16: 31–36.

González, M. N., Rodrigo, J. D. &. Suárez, C. S.,1986. Flora y vegetación del archipiélago canario.Edirca, Las Palmas de Gran Canaria.

Gordon, C. E. & Ornelas, J. F., 2000. Comparingendemism and habitat restriction in Mesoamericantropical deciduous forest birds: implications forbiodiversity conservation planning. Bird Conser-vation International, 10: 289–303.

Illera, J. C., 2001. Habitat selection by the CanaryIslands Stonechat (Saxicola dacotiae) (Meade–Waldo, 1889) in Fuerteventura Islands: a two–tierhabitat approach with implications for its conser-vation. Biological Conservation, 97: 339–345.

Johnson, T. H. & Stattersfield, A. J., 1990. A globalreview of island endemic birds. Ibis, 132: 167–180.

Jokimäki, J., 1999. Occurrence of breeding birdspecies in urban parks: effects of park structureand broad–scale variables. Urban Ecosystems,3: 21–34.

Klicka, J. & Zink, R. M., 1997. The importance ofrecent ice ages in speciation: a failed paradigm.Science, 277: 1666–1669.

MacArthur, R. H. & Wilson, E. O., 1967. The theoryof island biogeography. Monographs in Popula-tion Biology, 1.

MacArthur, R. H., Diamond, J. M. & Karr, J., 1972.Density compensation in island faunas. Ecology,53: 330–342.

Manne, L. L., Brooks, T. M., Pimm, S. L., 1999. Therelative risk of extinction of passerine birds oncontinents and islands. Nature, 399: 258–261.

Marshall, H. D. & Baker, A. J., 1999. Colonizationhistory of atlantic island common chaffinches(Fringilla coelebs) revealed by mitochondrial DNA.Mol. Phylogenet. Evol., 11: 201–212.

Martí, R. & Del Moral, J. C., 2003. Atlas de las avesreproductoras de España Dirección General dela Conservación de la Naturaleza–SociedadEspañola de Ornitología, Madrid.


Martín, A. & Lorenzo, J. A., 2001. Aves delarchipiélago canario. Francisco Lemus Editor, LaLaguna.

Martín, A., 1987. Atlas de las aves nidificantes de laisla de Tenerife. Instituto de Estudios Canarios,Sta. Cruz de Tenerife.

Martin, J–L., 1992. Niche expansion in an insularbird community: an autoecological perspective.Journal of Biogeography, 19: 375–381.

Marzluff, J. M., Bowman, R. & Donnelly, R., 2001.Avian conservation and ecology in an urbanizingworld. Kluwer Academic Publications, Boston.

Melles, S., Glenn, S. & Martin, K., 2003. Urban birddiversity and landscape complexity: species–en-vironment associations along a multiscale habi-tat gradient. Conservation Ecology, 7: 5. [online]URL: http://www.consecol.org/vol7/iss1/art5

Milberg, P. & Tyrberg, T., 1993. Naive birds andnoble savages–a review of man–caused prehis-toric extinctions of island birds. Ecography, 16:229–250.

Nogales, M., Medina, F. M., Quilis, V. & Gonzáles–Rodríguez, M., 2001. Ecological and biogeo-graphical implications of Yellow–Legged Gulls(Larus cachinnans Pallas) as seed dispersersof Rubia fruticosa Ait. (Rubiaceae) in the Ca-nary Islands. Journal of Biogeography, 28:1137–1145.

Nogales, M., Quilis, V., Medina, F. M., Mora, J. L. &Trigo, L. S., 2002. Are predatory birds effectivesecondary seed dispersers? Biological Journal ofthe Linnean Society, 75: 345–352.

Pimm, S. L., Jones, H. L. & Diamond, J. M., 1988.On the risk of extinction. American Naturalist,132: 757–785.

Pleguezuelos, J. M., 1992. Avifauna nidificante delas sierras Béticas orientales y depresión deGuadix, Baza y Granada. Monográfica Tierrasdel Sur, Univ. de Granada.

Prodon, R., Thibault, J–C. & Dejaifve, P–A., 2002.Expansion vs. compression of bird altitudinalranges on a Mediterranean island. Ecology, 83:1294–1306.

Rando, J. C., 2002. New data of fossil birds from ElHierro (Canary Islands): probable causes of ex-tinction and some biogeographical considera-tions. Ardeola, 49: 39–49.

Rando, J. C., López, M. & Seguí, B., 1999. A newspecies of extinct flightless Passerine (Embe-rizidae: Emberiza) from the Canary Islands. Con-dor, 101: 1–3.

Ricklefs, R. E. & Bermingham, E., 1999. Taxoncycles in the Lesser Antillean avifauna. Ostrich,70: 49–59.

Ricklefs, R. E., & Cox, G. W., 1978. Stage of taxon

cycle, habitat distribution, and population den-sity in the avifauna of the West Indies. AmericanNaturalist, 112: 875–895.

Sánchez, A., 1991. Estructura y estacionalidad delas comunidades de aves de la Sierra de Gredos.Ardeola, 38: 207–231.

Sangster, G., 2000. Taxonomic stability and avianextinctions. Conservation Biology, 14: 579–581.

Stattersfield, A. J., Crosby, M. J., Long, A. J. &Wege, D. G., 1998. Endemic Bird Areas of theWorld: Priorities for Biodiversity Conservation.Birdlife International–Smithsonian InstitutionPress, Washington, D.C.

Sturmbauer, C., Berger, B., Dallinger, R. & Föger,M., 1998. Mitochondrial phylogeny of the genusRegulus and implications of the evolution ofbreeding behaviour in sylvoid songbirds. Mo-lecular Phylogenetics and Evolution, 10: 144–149.

Suárez, F., 1984. Estructura y composición de lascomunidades de aves invernantes en las zonassemiáridas de Lanzarote y Fuerteventura (IslasCanarias). Ardeola, 30: 83–91.

Thomas, C. D., Glen, S. W. T., Lewis, O. T., Hill, J.K. & Blakeley, D. S., 1999. Population differen-tiation and conservation of endemic races: thebutterfly, Plebejus argus. Animal Conservation,2: 15–21.

Tucker, G.M., & Heath, M.F. 1994. Birds in Europe:their conservation status. Birdlife International(Birdlife Conservation Series No. 3), Cambridge,U.K.

UICN, 2001. IUCN Red List Categories: Version 3.1IUCN. Species Survival Commission (http://www.iucn.org/themes/ssc/redlists/ssc–rl–c.htm).

Valido, A., Tellería, J. L. & Carrascal, L. M., 1994.Between and within habitat distribution of theCanary Common Chaffinch (Fringilla coelebsombriosa): a test of the food abundance hypoth-esis. Ardeola, 41: 29–35.

Voelker, G., 1999. Molecular evolutionary relation-ships in the avian genus Anthus (Pipits:Motacillidae). Molecular Phylogenetics and Evo-lution, 11: 84–94.

Wiens, J. A., 1989. The ecology of bird communi-ties. Vol. 1. Cambridge Univ. Press, Cambridge.

Williamson, M., 1981. Island populations. OxfordUniv. Press, Oxford.

Wilson, E. O., 1961. The nature of the taxon cyclein the Melanesian ant fauna. American Natural-ist, 95: 169–193.

Yeaton, R. I., 1974. An ecological analysis ofchaparral and pine forest bird communities onSanta Cruz island and mainland California.Ecology, 55: 959–973.

http://www.consecol.org/vol7/iss1/art5

http://www.iucn.org/themes/ssc/redlists/ssc-rl-c.htm

http://www.iucn.org/themes/ssc/redlists/ssc-rl-c.htm


Secretària de Redacció / Secretaria de Redacción / Managing EditorMontserrat Ferrer

Consell Assessor / Consejo asesor / Advisory BoardOleguer EscolàEulàlia GarciaAnna OmedesJosep PiquéFrancesc Uribe

Editors / Editores / Editors Antonio Barbadilla Univ. Autònoma de Barcelona, Bellaterra, SpainXavier Bellés Centre d' Investigació i Desenvolupament CSIC, Barcelona, SpainJuan Carranza Univ. de Extremadura, Cáceres, SpainLuís Mª Carrascal Museo Nacional de Ciencias Naturales CSIC, Madrid, SpainAdolfo Cordero Univ. de Vigo, Vigo, SpainMario Díaz Univ. de Castilla–La Mancha, Toledo, SpainXavier Domingo Univ. Pompeu Fabra, Barcelona, SpainFrancisco Palomares Estación Biológica de Doñana, Sevilla, SpainFrancesc Piferrer Inst. de Ciències del Mar CSIC, Barcelona, SpainIgnacio Ribera The Natural History Museum, London, United KingdomAlfredo Salvador Museo Nacional de Ciencias Naturales, Madrid, SpainJosé Luís Tellería Univ. Complutense de Madrid, Madrid, SpainFrancesc Uribe Museu de Zoologia de Barcelona, Barcelona, Spain

Consell Editor / Consejo editor / Editorial BoardJosé A. Barrientos Univ. Autònoma de Barcelona, Bellaterra, SpainJean C. Beaucournu Univ. de Rennes, Rennes, FranceDavid M. Bird McGill Univ., Québec, CanadaMats Björklund Uppsala Univ., Uppsala, SwedenJean Bouillon Univ. Libre de Bruxelles, Brussels, BelgiumMiguel Delibes Estación Biológica de Doñana CSIC, Sevilla, SpainDario J. Díaz Cosín Univ. Complutense de Madrid, Madrid, SpainAlain Dubois Museum national d’Histoire naturelle CNRS, Paris, FranceJohn Fa Durrell Wildlife Conservation Trust, Trinity, United KingdomMarco Festa–Bianchet Univ. de Sherbrooke, Québec, CanadaRosa Flos Univ. Politècnica de Catalunya, Barcelona, SpainJosep Mª Gili Inst. de Ciències del Mar CMIMA–CSIC, Barcelona, SpainEdmund Gittenberger Rijksmuseum van Natuurlijke Historie, Leiden, The NetherlandsFernando Hiraldo Estación Biológica de Doñana CSIC, Sevilla, SpainPatrick Lavelle Inst. Français de recherche scient. pour le develop. en cooperation, Bondy, FranceSantiago Mas–Coma Univ. de Valencia, Valencia, SpainJoaquín Mateu Estación Experimental de Zonas Áridas CSIC, Almería, SpainNeil Metcalfe Univ. of Glasgow, Glasgow, United KingdomJacint Nadal Univ. de Barcelona, Barcelona, SpainStewart B. Peck Carleton Univ., Ottawa, Canada Eduard Petitpierre Univ. de les Illes Balears, Palma de Mallorca, SpainTaylor H. Ricketts Stanford Univ., Stanford, USAJoandomènec Ros Univ. de Barcelona, Barcelona, SpainValentín Sans–Coma Univ. de Málaga, Málaga, SpainTore Slagsvold Univ. of Oslo, Oslo, Norway

Secretaria de Redacció / Secretaría de Redacción / Editorial Office

Museu de ZoologiaPasseig Picasso s/n08003 Barcelona, SpainTel. +34–93–3196912Fax +34–93–3104999E–mail [email protected]

"La tortue greque" Oeuvres du Comte de Lacépède comprenant L'Histoire Naturelle des Quadrupèdes Ovipares, des Serpents, des Poissons et des Cétacés; Nouvelle édition avec planches coloriées dirigée par M. A. G. Desmarest; Brux-elles: Th. Lejeuné, Éditeur des oeuvres de Buffon, 1836. Pl. 7

Animal Biodiversity and Conservation 24.1, 2001© 2001 Museu de Zoologia, Institut de Cultura, Ajuntament de BarcelonaAutoedició: Montserrat FerrerFotomecànica i impressió: Sociedad Cooperativa Librería GeneralISSN: 1578–665XDipòsit legal: B–16.278–58



Waite, T. A., Vucetich, J., Saurer, T., Kroninger, M., Vaughn, E., Field, K. & Ibargüen, S., 2005. Minimizing extinctionrisk through genetic rescue. Animal Biodiversity and Conservation, 28.2: 121–130.

AbstractMinimizing extinction risk through genetic rescue.— According to the genetic rescue hypothesis, immigrants canimprove population persistence through their genetic contribution alone. We investigate the potential for such rescueusing small, inbred laboratory populations of the bean beetle (Callosobruchus maculatus). We ask how manymigrants per generation (MPG) are needed to minimize the genetic component of extinction risk. During Phase 1,population size was made to fluctuate between 6 and 60 (for 10 generations). During this phase, we manipulatedthe number of MPG, replacing 0, 1, 3, or 5 females every generation with immigrant females. During Phase 2, wesimply set an upper limit on population size (.10). Compared with the 0–MPG treatment, the other treatments wereequivalently effective at improving reproductive success and reducing extinction risk. A single MPG was sufficient forgenetic rescue, apparently because effective migration rate was inflated dramatically during generations whenpopulation size was small. An analysis of quasi–extinction suggests that replicate populations in the 1–MPGtreatment benefited from initial purging of inbreeding depression. Populations in this treatment performed so wellapparently because they received the dual benefit of purging followed by genetic infusion. Our results suggest theneed for further evaluation of alternative schemes for genetic rescue.

Key words: Extinction risk, Founder events, Genetic rescue, Inbreeding.

ResumenMinimización del riesgo de extinción mediante el rescate genético.— Según la hipótesis del rescate genético,los inmigrantes pueden mejorar la persistencia de una población mediante su contribución genética. Hemosinvestigado el potencial de un rescate de este tipo, utilizando pequeñas poblaciones endogámicas de laboratoriodel gorgojo del haba Callosobruchus maculatus. Nos preguntamos cuántos migrantes por generación (MPG)son necesarios para minimizar el componente genético del riesgo de extinción. Durante la Fase 1, se hizofluctuar el tamaño de la población entre 6 y 60 (durante 10 generaciones). En dicha fase manipulamos el númerode MPGs, reemplazando 0, 1, 3, o 5 hembras nativas por hembras inmigrantes en cada generación. Durante laFase 2, nos limitamos a poner un límite superior al tamaño de la población (.10). Comparados con eltratamiento de 0–MPG, los otros tratamientos resultaron ser igualmente efectivos en la mejora del éxitoreproductivo y la reducción del riesgo de extinción. Un único MPG era suficiente para el rescate genético,aparentemente debido a que la tasa de migración efectiva quedaba espectacularmente sobredimensionadadurante generaciones, cuando el tamaño de la población era pequeño. Un análisis de cuasi–extinción sugiereque las poblaciones replicadas durante el tratamiento 1–MPG se beneficiaron de un saneamiento inicial por ladisminución de la endogamia. Aparentemente, las poblaciones de este tratamiento se comportaron tan biendebido a que recibieron el doble beneficio del saneamiento seguido de la inyección genética. Nuestros resultadossugieren la necesidad de posteriores evaluaciones del rescate genético mediante esquemas alternativos.

Palabras clave: Riesgo de extinción, Acontecimientos de hundimiento, Rescate genético, Endogamia.

(Received: 19 VIII 04; Conditional acceptance: 8 X 04; Final acceptance: 9 XI 04)

T. A. Waite, T. Saurer, M. Kroninger, E. Vaughn, K. Field & S. Ibargüen, Dept. of Evolution, Ecology, andOrganismal Biology, Ohio State Univ., Columbus, Ohio 43210–1293, U.S.A.– J. Vucetich, School of ForestResources and Environmental Science, Michigan Technological Univ., Houghton, Michigan 49931, U.S.A.

Corresponding author: T. A. Waite. E–mail: [email protected]

Minimizing extinction riskthrough genetic rescue

T. A. Waite, J. Vucetich, T. Saurer, M. Kroninger,E. Vaughn, K. Field & S. Ibargüen


122 Waite et al.

plant (Silene alba) (Richards, 2000). Likewise, geneflow via immigration improved various fitness com-ponents in the self–incompatible mustard (Brassicacampestris) (Newman & Tallmon, 2001), and im-proved fitness and reduced extinction risk in thehouse fly (Musca domestica) (Bryant et al., 1999).Lastly, gene flow facilitated by an alien pollinator(African honeybee, Apis mellifera scutellata) is ap-parently responsible for improved reproductive out-put in an Amazonian tree (Dinizia excelsa[Fabaceae]) in pastures and forest remnants, wherenative pollinators are absent (Dick, 2001). Althoughsome earlier studies provided contradictory find-ings (references in Newman & Tallmon, 2001),these recent studies indicate that pollen– or immi-grant–mediated gene flow can dramatically improvefitness in small inbred populations.

Here, we describe an experiment that extendsthese recent findings. Using inbred laboratorypopulations of the bean beetle (Callosobruchusmaculatus), we manipulated the number of MPG byreplacing 0, 1, 3, or 5 females with immigrantfemales each generation. The experiment allowedus to evaluate: (1) whether even a single MPGcould lead to genetic rescue, and (2) how manymigrants are needed to minimize the genetic com-ponent of extinction risk.

Methods

Subjects

C. maculatus is an important pest species. Thebeetles used in our experiment were derived froma genetic strain from southern India and reared atOhio State University. Several features make thisspecies a suitable model organism (e.g., Vucetichet al., 2000): (1) it has a short generation time(4–6 weeks); (2) females oviposit on beans andoffspring emerge synchronously, with the adultstypically dying before the next generation emerges;and (3) because only one beetle typically emergesfrom each mung bean (Vigna radiata), carryingcapacity can be controlled simply by limiting thenumber of beans available.

Overview and rationale

The experiment was designed to quantify the requi-site number of MPG to minimize extinction risk insmall inbred populations. It comprised two phases.During Phase 1, population size was made to fluctu-ate between 6 and 60 individuals across 10 genera-tions. In each generation, the &ð:%ð sex ratio was5:1. During this phase, we manipulated the numberof MPG by replacing 0, 1, 3, or 5 females everygeneration with immigrant females from a largeoutbred population. At the end of this phase, wemeasured the reproductive fitness and founding suc-cess of each replicate population. During Phase 2,we limited N by simply providing 10 mung beans toeach replicate population for 10 generations. During

Introduction

Small isolated populations are subject to loss ofgenetic diversity through drift and inbreeding. De-spite a large body of findings implicating inbreedingas a contributor to extinction risk (reviewed byHedrick & Kalinowski, 2000), the strength of anycausal linkage between inbreeding and extinctionremains a point of contention. Until recently, therewas no direct evidence that genetic deteriorationcontributes to extinction of wild populations(Frankham & Ralls, 1998). Lacking such evidence,some workers have argued that stochastic demo-graphic and environmental events may typicallydrive small populations to the brink of extinctionbefore genetic deterioration poses a serious threat(Lande, 1988; Pimm et al., 1988; Caro & Laurenson,1994). Even so, there is widespread agreement thatloss of genetic diversity can lead to extinction.Support for this perspective comes from theoreticalstudies (Mills & Smouse, 1994; Lande, 1998;Tanaka, 2000; Finke & Jetschke, 1999; Fowler &Whitlock, 1999), laboratory experiments (Frankham,1995a; Bryant et al., 1999; Bijlsma et al., 2000;Reed & Bryant, 2000; Nieminen et al., 2001), fieldexperiments (e.g., Newman & Pilson, 1997), a land-mark study of a metapopulation in nature (Saccheriet al., 1998), and meta–analyses (Frankham, 1999).A recent review summarizes evidence, based onnew pedigree data and new data made possible bymolecular and analytical tools for estimating in-breeding, that inbreeding can adversely affect popu-lation performance (Keller & Waller 2002; see alsoGoudet & Keller 2002). Meanwhile, a flurry of re-cent experimental (e.g., Bryant et al., 1999; Reed &Bryant, 2000, 2001; Newman & Tallmon, 2001) andtheoretical studies (e.g., Fu et al., 1998; Bataillon &Kirkpatrick, 2000; Kirkpatrick & Jarne, 2000; Wang,2000; Whitlock, 2000; Linklater, 2003) have ex-plored ways to minimize the genetic component ofextinction risk.

What kind of genetic intervention, if any, isneeded? Ideally, genetic risks could be minimizedwithout intervention, simply by maintainingpopulations above minimum viable size (reviewedby Reed & Bryant, 2000; see also Lande, 1995;Lynch et al., 1995; Gilligan et al., 1997). However,when this approach is not feasible or has alreadyfailed, genetic diversity can be maintained or re-stored by facilitating gene flow via translocation ofindividuals or propagules (e.g., Madsen et al., 1999).Because the mere arrival of immigrants could fore-stall local extinction, to demonstrate unequivocallythat gene flow per se is beneficial, one must per-form experiments in which genetic diversity is intro-duced without a simultaneous increase in popula-tion size. Recent studies have sought to provideevidence for such genetic rescue (i.e., increase infitness due to gene flow) of recently fragmented ornewly colonized populations (reviewed byIngvarsson, 2001; see also Vila et al., 2003). Forexample, pollen–mediated gene flow improved fit-ness in small populations of a dioecious weedy


this phase, we documented extinctions. The ration-ale for each of these procedures is described below.

Most tests of the genetic rescue hypothesis havetreated the infusion of new genetic material di-chotomously, testing the effect of a single level ofinfusion versus no infusion. In an attempt to titratethe level of genetic infusion minimizing extinctionrisk, we assessed the effects of several levels onreproductive fitness and extinction risk. To avoidconfounding the results with effects attributable todemographic rescue, these migrants were replace-ments, not additions.

Due to fluctuations in population size (FPS)(Vucetich et al., 1997; Vucetich & Waite, 1999),skewed sex ratio, and extra–Poisson variation infecundity, most real populations exhibit Ne/N ratiosthat are less than unity (Frankham, 1995b). There-fore, during Phase 1, we manipulated the FPS andsex ratio to achieve Ne/N ratios typical of realpopulations. This approach resulted in an Ne/Nratio of -----0.2, which is close to the median ofsurveyed populations (Frankham, 1995b).

The most straightforward way to perform geneticrescue is to infuse a population with genetic mate-rial for a pulse (i.e., one, two, or a few generations).In Phase 2 of the experiment, we assess the re-sidual impact of genetic infusion. That is, we as-sess the effects of prior genetic management (im-posed during Phase 1) on extinction risk.

Detailed protocol

Preliminary stepsTo establish replicate inbred populations, we beganby orchestrating two successive full–sibling matings.Eighty female–male pairs, representing 80 uniquepairs of founders, were used. Offspring of thesepairs comprised the parental generation. The in-breeding coefficient, F, in these progeny was 0.375.This procedure served several purposes. First, wewere interested in investigating the effectiveness ofgenetic rescue of already–inbred populations. Sec-ond, we intended to purge the genetic load suchthat further purging would not confound our results.Finally, our prior work (unpubl. results) showed thatadditional full–sibling matings would push F be-yond the extinction quasi–threshold (Frankham,1995a).

To establish large outbred source populations ofpotential immigrants, we created five populationseach comprising -----5,000 individuals. ThroughPhase 1(described below), we housed these fivepopulations separately. Because the timing of emer-gence in the five source populations diverged overtime, this procedure was used to ensure a continu-ous supply of immigrants.

Phase 1We placed a female and a male with the same full–sibling parents into each of 80 petri dishes, eachcontaining 40 pristine (eggless) mung beans. Fol-lowing oviposition, we placed each of egg–ladenbean in a separate Eppendorf tube. As adults

emerged, we placed one male and a specifiednumber of females (5, 4, 2, or 0 in the 0–, 1–, 3–,and 5–MPG treatments, respectively) from the samereplicate population along with a complimentarynumber of immigrant females (0, 1, 3, or 5) intopetri dishes containing 250 pristine beans. Forexample, in each replicate in the 1–MPG treatment,one male was placed together with four femalesfrom the same replicate population along with oneimmigrant female. Thus, each of 80 dishes (20 rep-licates in each treatment) contained five femalesand one male in all odd–numbered generations. Onthe 21st day after the adults had been put together,any surviving adults were removed and each egg–laden bean was placed in a separate tube.

Upon emergence, we repeated the protocol, ex-cept that 10 males were put together with 50, 49,47, or 45 females (in the 0–, 1–, 3–, and 5–MPGtreatments, respectively) from the same replicatepopulation and with 0, 1, 3, or 5 immigrant females.Each replicate population thus comprised 60 adults(50 females and 10 males) in generation 2 (and alleven–numbered generations in Phase 1). We thenrepeated the above alternation between N = 6 inodd generations and N = 60 in even generationsthrough the 10th generation.

Throughout this phase, a pool of immigrantfemales was kept available by placing egg–ladenbeans from the source population singly into200 tubes every generation. By matching femaleimmigrants by date of 4th emergence in candidaterecipient populations, we ensured that female im-migrants were approximately the same age (i.e.,within 7 days) as most members of the recipientpopulation. Any potential female immigrant notassigned within two weeks following her emer-gence was excluded. Females satisfying the crite-ria for inclusion were transferred to appropriatepopulations according to the following rules. Egg–laden beans (one in each of # 250 tubes in eachreplicate) were monitored daily for onset of emer-gence. We designated the day of 4th emergence asDay 0. On Day 7, we determined whether at leastone male had emerged. If so and if the criterionnumbers of females and males had emerged, theywere combined in a petri dish with the specifiednumber of female immigrants. Mating was allowedto proceed. On Day 21, we transferred each egg–laden bean to a tube. If the criterion numbers offemales and males had not been reached by Day7, we placed the male(s) together with females(including immigrants) and added newly emergingindividuals daily. This process continued until thecriterion was met or until 3 consecutive dayspassed with no emergence. Then, 7 days after thelast individual was added or 2 weeks after initiallyputting beetles together (whichever was longer),we transferred each egg–laden bean to a tube.Finally, if no males had emerged by Day 7, wewaited until the first male emerged and then fol-lowed the just–described procedure. Somepopulations failed to reach the criterion numbersof adults, particularly in even–numbered genera-

124 Waite et al.

tions, when numerous populations failed to pro-duce 45 females. We refer to these failures asquasi–extinctions. We compare the incidence ofquasi–extinction between treatments and acrossgenerations during Phase 1, when very few trueextinctions took place.

Phase 2After generation 10, replicate populations acrossall treatments were treated uniformly. Every popu-lation was subjected to a constant carrying ca-pacity. No further immigration was orchestratedand no population variability was imposed. Thisphase lasted 10 generations. For each extant

population at the end of Phase 1, we placed allegg–laden beans (up to 210) in a large petri dishand then added 10 pristine beans. (Several repli-cates [10 in 0–MPG: 4 in 1–MPG , 2 in 3–MPG, 1in 5–MPG] had gone extinct during Phase 1;others were lost to human error [2 in 0–MPG, 1 in3–MPG].) Following oviposition, we discarded theoriginal beans and placed 10 pristine beans inthe dish with the 10 egg–laden beans. Followingthe next emergence and oviposition, we replacedthe 10 old beans with 10 pristine beans. Werepeated this process until extinction occurred oruntil the 20th generation (10th in Phase 2). Foreach population, time to extinction (in genera-

Fig. 1. Survivorship curves for small experimental populations of bean beetles (Calosobruchusmaculatus) in four migrant–per–generation (MPG) treatments, with 20 replicate populations pertreatment at the start of the experiment: A. Survivorship curves during Phase 1 of the experiment,when each population fluctuated between 6 individuals (1 male and 5 females) in odd generations and60 individuals (10 males and 50 females) in even generations; B. Survivorship curves during Phase 2of the experiment, when each population was subjected to an approximate carrying capacity of 10individuals during every generation.

Fig. 1. Curvas de supervivencia para pequeñas poblaciones experimentales del gorgojo del habaCalosobruchus maculatus, en tratamientos de cuatro migrantes por generación (MPG), con 20poblaciones replicadas en cada tratamiento al inicio del experimento: A. Curvas de supervivenciadurante la Fase 1 del experimento, cuando cada población fluctuaba entre 6 individuos (1 macho y 5hembras) en las generaciones impares y 60 individuos (10 machos y 50 hembras) en las generacionespares; B. Curvas de supervivencia durante la Fase 2 del experimento, cuando cada una de laspoblaciones estaba sujeta a una capacidad de carga aproximada de 10 individuos por generación.

1.0

0.8

0.6

0.4

0.2

0.0

1.0

0.8

0.6

0.4

0.2

0.0

Pro

po

rtio

n s

urv

ivin

gP

rop

ort

ion

su

rviv

ing

2 4 6 8 10

12 14 16 18 20 Generation

A

B


tions) was recorded. A population was judged tohave gone extinct if: (1) no oviposition took place,(2) no beetles emerged, or (3) beetles of only onesex emerged.

Fitness measurementTo estimate individual reproductive success at theend of Phase 1, we began by randomly selecting40 (of a possible 250) egg–laden beans from eachpopulation and placing two such beans in each of20 tubes. Next, we monitored emergence andtransferred five (whenever possible) female–malepairs to five petri dishes, each containing 100pristine beans. We then allowed mating and ovi-position to occur. Following emergence, we talliedthe offspring produced by each pair. We used thisquantity as our primary measure of fitness, but wealso took advantage of the fact that some pairsfailed to produce at least one adult offspring ofeach sex. We considered any such case to be afailed founding event. We thus compare both re-productive success and founding success acrossMPG treatments.

Data analysis

Survival analysis was performed using S–PLUS2000 (1999). Kaplan–Meier nonparametric survivalmodels were used to estimate mean time to extinc-tion in each MPG treatment. Cox proportional haz-ards models were used to evaluate the effect ofMPG treatment on risk of extinction. All pairwisecomparisons were performed. Nominal P–valuesare reported, with an indication of whether eachtest is significant at the experimentwise ±–level of0.05 (following Bonferroni correction). Other analy-ses were performed using SPSS (1999). Fisher’sexact tests were used to perform pairwise compari-sons of incidence of extinction and quasi–extinctionbetween MPG levels and between generations dur-ing Phase 1. One–way ANOVA was used to com-pare fitness (number of offspring produced perfemale–male pair) across MPG levels. Pairwisecomparisons were made using Tukey’s HSD method.Finally, we compared the incidence (arcsin squareroot transformed proportion) of successful founding(production of at least one offspring of each sex)across MPG levels. Because the normality testfailed (P < 0.001), we used Kruskal–Wallisnonparametric one–way "analysis of variance" onranks. Pairwise comparisons were made usingDunn’s method, with the critical ±–level set at 0.05.

Results

Survival analysis

Figure 1 shows the survival of replicate populations.No significant differences in incidence of extinctionemerged among treatments by the end of Phase 1(all Ps > 0.15, Fisher’s exact test), when fewextinctions occurred (i.e., 8 of 67 populations). In

Phase 2, mean persistence of populations in the 0–MPG treatment (6.4 generations) was substantiallyshorter than in every other treatment (1–MPG: 8.9,3–MPG: 8.7, and 5–MPG: 8.9), based on Kaplan–Meier survival analysis. Cox proportional hazardsanalysis revealed that extinction risk declined sig-nificantly with increasing MPG ($ = –0.23,exp[$] = 0.80 [95% CI: 0.68–0.94], P = 0.007),where exp($) quantifies the proportional effect of aunit increase in the experimental factor (MPG).Pairwise comparisons revealed significant effectsof increasing MPG from 0 to any other level (i.e., 1,3, or 5) (table 1). No other pairwise comparisonwas significant. Number of MPG was a significantpredictor of extinction risk only when the 0–MPGtreatment was included. The 1–, 3–, and 5–MPGtreatments appeared to reduce extinction risk withequivalent effectiveness.

Table 1. Results of pairwise Cox proportionalhazards comparisons (test of effect ofmanipulating number of migrants pergeneration). For each comparison, $ (= slope),exp($), and P are shown. Each of the firstthree comparisons is significant followingBonferroni adjustment for the number ofpairwise tests performed (i.e., the nominal P–value is less than 0.05/6). Any negative valueof $ corresponds with a value of exp($) < 1,which indicates the decrease in relative riskof extinction associated with a unit increasein the experimental factor MPG.

Tabla 1. Resultados de las comparaciones alazar de Cox de los riesgos proporcionales porparejas (test del efecto de la manipulación delnúmero de inmigrantes por generación). Paracada comparación se muestran $ (= pendiente),exp($), y P. Cada una de las tres primerascomparaciones es significativa según el ajustede Bonferroni para el número de test por parejasllevados a cabo (es decir, el valor nominal de Pes menor de 0.05/6). Cualquier valor negativode $ se corresponde con un valor de exp($) < 1,lo que indica un descenso del riesgo relativo deextinción asociado con un incremento unitariodel factor experimental MPG.

Between–treatmentcomparison $ exp($) P

0– vs 1–MPG –2.48 0.08 1.7×10–5

0– vs 3–MPG –0.76 0.47 3.6×10–6

0– vs 5–MPG –0.61 0.54 1.7×10–7

1– vs 3–MPG 0.10 1.10 0.65

1– vs 5–MPG –0.002 0.99 0.98

3– vs 5–MPG –0.12 0.88 0.53

126 Waite et al.

Quasi–extinction analysis

Figure 2 shows the incidence of quasi–extinction(defined as failure to produce at least 45 females ineven–numbered generations during Phase 1). Theincidence of quasi–extinction showed a pronouncedtemporal pattern in the 0–MPG treatment, decreas-ing from generation 2 to 4 (P = 0.003; Fisher’sexact test) and then increasing (generation 6 vs. 8:P = 0.009; generation 4 vs 8: P = 0.003) to theinitial level (generation 2 vs 10: P = 1.0). Thistemporal pattern of quasi–extinction in the 0–MPGtreatment suggests an initial purging of inbreedingdepression followed by onset of inbreeding depres-sion. An initial decline, from generation 2 to 4, wasdetectable in the 1–MPG treatment (Ps = 0.044 forcomparisons between generation 2 vs 4, 6, 8, and10), suggesting an initial purging of inbreedingdepression with no subsequent onset of inbreedingdepression by the end of Phase 1. Other pairwisecomparisons were nonsignificant (Ps = 1.0). Nosignificant between–generation differences in inci-dence of quasi–extinction emerged for either the 3–or 5–MPG treatment (Ps > 0.48), suggesting nei-ther an initial purging of inbreeding depression nora subsequent onset of inbreeding depression inthese treatments.

Within generations, several between–treatmentdifferences emerged. In generation 2, quasi–extinc-tion risk was reduced by the one–time immigrationof a single female (i.e., incidence of quasi–extinc-tion was lower in 1– than 0–MPG; P = 0.025).

Fig. 2. The incidence of quasi–extinction (proportion of populations failing to produce at least 45female offspring during even–numbered generations) in four migrant–per–generation treatments(MPG). Any extant population could experience quasi–extinction repeatedly.

Fig. 2. Incidencia de la cuasi–extinción (proporción de poblaciones que no consiguen producir al menos45 descendientes hembra durante las generaciones pares) en tratamientos de cuatro migrantes porgeneración (MPG). Cualquier población existente podría experimentar la cuasi–extinción repetidasveces.

0 MPG1 MPG3 MPG5 MPG

0.8

0.6

0.4

0.2

0.0 2 4 6 8 10

Generation

Pro

po

rtio

n q

uas

i–ex

tin

ctio

n

Quasi–extinction risk was further reduced by theimmigration of additional females (i.e., 1– vs 3–MPG: P = 0.008; 1– vs 5–MPG: P = 0.02). Inci-dence of quasi–extinction was minimized equiva-lently in the 3– and 5–MPG treatments. In genera-tion 4, only one comparison (0– vs 5–MPG) wasnominally significant. Within subsequent genera-tions, incidence of quasi–extinction was higher inthe 0–MPG treatment than in any other treatment(Ps < 0.001). All other pairwise comparisons werenonsignificant.

Fitness analysis

Figure 3A summarizes the results of the fitness testconducted at the end of Phase 1. Significant hetero-geneity emerged across treatments (F3,43.7 = 11.86,P < 0.001), but not among replicates (F19,37.9 = 1.33,P = 0.22). Tukey’s HSD tests revealed that thenumber of offspring produced per female–male pairwas significantly lower in the 0–MPG treatment thanin any other treatment (all Ps < 0.001). Although avisual inspection of Figure 3A suggests a weaktendency for female–male pairs in the 5–MPG treat-ment to produce more offspring (mean = 55.0) thanpairs in the 1–MPG (49.3) and 3–MPG treatments(48.6), neither of these comparisons was significant(1– vs 5–MPG: P = 0.20; 3– vs 5–MPG: P = 0.15)(nor was comparison between 1– and 3–MPG:P = 1.0).

Figure 3B shows the founding success (i.e.,proportion of pairs that produced at least one adult


Fig. 3. A. Fitness measurements (mean number of offspring produced per female–male pair) in fourmigrant–per–generation (MPG) treatments at the end of Phase 1 of the experiment; the thin line withineach box indicates the median, the thick line within each box indicates the mean, the box representsthe interquartile interval (25th to 75th percentile), and the whiskers show the 10th and 90th percentiles.B. The founding success (defined as proportion of female–male pairs that produced at least one adultoffspring of each sex) in four migrant–per–generation (MPG) treatments at the end of Phase 1 of theexperiment. Symbols indicate means and the error bars represent standard errors. For both analyses,the numbers of replicate populations were as follows: 0–MPG, 8; 1–MPG, 16; 3–MPG, 17; and 5–MPG,19 replicates.

Fig. 3. A. Mediciones de aptitud (número promedio de descendientes producidos por cada parmacho–hembra) en los tratamientos de cuatro migrantes por generación (MPG), al final de laFase 1 del experimento; la fina línea del interior de cada rectángulo indica la mediana, y la líneagruesa la media, mientras que los rectángulos representan los intervalos intercuartiles (lospercentiles 25 a 75); los extremos de las barras verticales muestran los percentiles 10 y 90. B. Éxitode fundación (definido como la proporción de parejas macho–hembra que produjeron al menos undescendiente adulto de cada sexo) en tratamientos de cuatro migrantes por generación (MPG) alfinal de la Fase 1 del experimento. Los símbolos indican las medias y las barras de error loserrores estándar. Para ambos análisis, los números de poblaciones replicadas fueron los siguien-tes: 0–MPG, 8; 1–MPG, 16; 3–MPG, 17; y 5–MPG, 19.

ing success was significantly lower in the 0–MPGtreatment than in any other treatment (qs > 3.54,Ps < 0.05). Other pairwise comparisons were non-significant.

offspring of each sex) of female–male pairs. Signifi-cant heterogeneity emerged among treatments(Kruskal–Wallis test: H3 = 24.75, P < 0.001). Dunn’smultiple–comparison procedure revealed that found-

80

60

40

20

0

1.0

0.8

0.6

0.4

0.2

0.0

0 1 3 5

0 1 3 5Migrants per generation

Fo

un

din

g s

ucc

ess

Rep

rod

uct

ive

succ

ess

A

B

128 Waite et al.

Discussion

Our survival analysis suggests that the introductionof a small number of migrants per generation (MPG)was sufficient for genetic rescue of small, inbredpopulations of the bean beetle (fig. 1, table 1).Compared with the control (0–MPG treatment), allof the other treatments (1–, 3–, and 5–MPG) im-proved population persistence. Because immigrantfemales were replacements rather than extras, thisfinding may be attributable to beneficial effects ofgene flow per se, perhaps including the masking offixed deleterious mutations. That is, our manipula-tion apparently led to genetic rescue, as distin-guished from demographic rescue.

In agreement with other studies, our findingssuggest that even a single MPG can lead to im-proved fitness (e.g., Newman & Tallmon, 2001; Vilaet al., 2003) and reduced extinction risk (e.g., Bryantet al., 1999). Moreover, our results suggest that theextent of genetic rescue was independent of numberof MPG, provided there was at least one MPG.Compared with the 0–MPG treatment, the othertreatments were equivalently effective both at im-proving fitness (fig. 3) and reducing extinction risk(table 1). This finding is superficially puzzling inlight of recent theoretical arguments that one actualmigrant per generation will often be inadequate(Mills & Allendorf, 1996), particularly when the re-cipient population fluctuates (Vucetich & Waite,2000). However, this theory does not apply here(Kalinowsky & Waples, 2002) because thepopulations did not merely fluctuate; they also fellto a very small size (in alternate generations).When this happens, the migration rate associatedwith a fixed number of migrants can be dramati-cally inflated (Vucetich & Waite, 2001). Because wedid not determine parentage, we cannot use genea-logical data to calculate the realized geneticallyeffective migration rate each generation. However,as a first approximation, we estimate that the aver-age migration rate, úm in the 1–, 3– and 5–MPGtreatments was 4.5, 14, and 23 in odd generations(and 0.5, 1.7, and 2.8 in even generations), whereú is mean population size and m is actual numberof immigrants divided by current size of the recipi-ent population. Thus, it appears that the 1–MPGtreatment performed so well because the rate ofgenetic infusion was adequate after all.

Our analysis of quasi–extinction (fig. 2) suggeststhat populations in the 1–MPG treatment mighthave benefited also from initial purging of the ge-netic load (e.g., Fu et al., 1998; Fu, 1999; Wang,2000; Reed & Bryant, 2001). Incidence of quasi–extinction in the 1–MPG treatment was high ini-tially, but then decreased dramatically and remainedlow, suggesting initial purging followed by fitness–enhancing gene flow. By contrast, incidence ofquasi–extinction in the 0–MPG treatment decreasedinitially but then increased, suggesting purging fol-lowed by onset of inbreeding depression in theabsence of gene flow. Further evidence that 1–MPG populations benefited, in part, from gene flow

is provided by the observation that quasi–extinctionrisk was reduced by the first introduction of a singlefemale (i.e., incidence of quasi–extinction was lowerin 1–MPG than 0–MPG treatment in generation 2;fig. 2) (see also Spielman & Frankham, 1992). Thisresult suggests that even a one–time immigrationby a single individual can make a sufficient geneticcontribution to provide a rescue effect (Ball et al.,2000; Vila et al., 2003). Our results also indicate,though, that quasi–extinction risk was further re-duced by additional immigrants (fig. 2). Thus, the1–MPG treatment did not perform as well as the 3–and 5–MPG treatments at this stage. Taken to-gether, these results suggest a duel benefit forpopulations in the 1–MPG treatment: initial purgingof inbreeding depression combined with subse-quent fitness–enhancing gene flow.

In summary, our results suggest that even asingle MPG may sometimes be useful for geneticmanagement of small, inbred populations. A singleactual MPG may sometimes be sufficient, particu-larly if the recipient population is small (Vucetich &Waite, 2001; see also Kalinowsky & Waples, 2002)and if inbreeding depression is purged initially (e.g.,Backus et al., 1995). The adequacy of one MPGcould be further enhanced if offspring of immi-grants exhibit heterosis (Ingvarsson & Whitlock,2000) or if immigrants are characterized by outbredvigor (Ball et al., 2000) and/or a mating advantage.Yet, it would be premature to promote the introduc-tion of just one MPG as a general practice. Addi-tional work should build upon experimental andtheoretical studies (cited in Introduction) that haveattempted to identify strategies for minimizing ex-tinction risk.

Acknowledgements

We thank D. Blazer, J. Vetter, D. Fletcher, S. Milne,E. Lawyer, and especially S. Reaser for help withthe experiment; G. Keeney for supplying the bee-tles; and D. Fowler and Robin Waples for com-ments on the ms. JAV was supported, in part, bythe U.S. National Science Foundation (DEB–9317401, DEB–9903671).

References

Backus, V. L., Bryant, E. H., Hughes, C. R. &Meffert, L. M., 1995. Effect of migration or in-breeding followed by selection on low–founder–number populations: implications for captivebreeding programs. Conservation Biology, 9:1216–1224.

Ball, S. J., Adams, M., Possingham, H. P. & Keller,M. A., 2000. The genetic contribution of singlemale immigrants to small, inbred populations: alaboratory study using Drosophila melanogaster.Heredity, 84: 677–684.

Bataillon, T. & Kirkpatrick, M., 2000. Inbreedingdepression due to mildly deleterious mutations


in finite populations: size does matter. GeneticalResearch, 75: 75–81.

Bijlsma, R., Bundgaard, J. & Boerema, A. C., 2000.Does inbreeding affect the extinction risk of smallpopulations?: predictions from Drosophila. Jour-nal of Evolutionary Biology, 13: 502–514.

Bryant, E. H., Vackus, V. L., Clark, M. E. & Reed, D.H., 1999. Experimental tests of captive breedingfor endangered species. Conservation Biology,13: 1487–1496.

Caro, T. M. & Laurenson, M. K., 1994. Ecologicaland genetic factors in conservation: a cautionarytale. Science, 263: 485–486.

Dick, C. W., 2001. Genetic rescue of remnanttropical trees by an alien pollinator. Proceedingsof the Royal Society of London B, 268: 2391–2396.

Finke, E. & Jetschke, G., 1999. How inbreeding andoutbreeding influence the risk of extinction: agenetically explicit model. MathematicalBiosciences, 156: 309–314.

Fowler, K. & Whitlock, M. C., 1999. The variance ininbreeding depression and the recovery of fit-ness in bottlenecked populations. Proceedingsof the Royal Society of London B, 266: 2061–2066.

Frankham, R., 1995a. Inbreeding and extinction: athreshold effect. Conservation Biology, 9: 792–799.

– 1995b. Effective population size/adult populationsize ratios in wildlife: a review. Genetical Re-search, 66: 95–107.

– 1999. Resolving conceptual issues in conserva-tion genetics: the roles of laboratory species andmeta–analyses. Hereditas, 130: 195–201.

Frankham, R. & Ralls, K., 1998. Inbreeding leadsto extinction. Nature, 392: 441–442.

Fu, Y. B., 1999. Patterns of the purging of deleteri-ous genes with synergistic interactions in differ-ent breeding schemes. Theoretical and AppliedGenetics, 98: 337–346.

Fu, Y. B., Namkoong, G. & Carlson, J. E., 1998.Comparison of breeding strategies for purginginbreeding depression via simulation. Conserva-tion Biology, 12: 856–864.

Gilligan, D. M., Woodworth, L. M., Montgomery, M.E., Briscoe, D. A. & Frankham, R., 1997. Ismutation accumulation a threat to the survival ofendangered populations. Conservation Biology,11: 1235–1241.

Goudet, J. & Keller, L., 2002. The correlation be-tween inbreeding and fitness: does allele sizematter? Trends in Ecology and Evolution, 17:201–202.

Hedrick, P. W. & Kalinowski, S. T., 2000. Inbreedingdepression in conservation biology. Annual Re-view of Ecology and Systematics, 31: 139–200.

Ingvarsson, P. K., 2001. Restoration of geneticvariation lost: the genetic rescue hypothesis.Trends in Ecology and Evolution, 16: 62–63.

Ingvarsson, P. K. & Whitlock, M. C., 2000. Hetero-sis increases the effective migration rate. Pro-ceedings of the Royal Society of London B, 267:1321–1326.

Kalinowski, S. T. & Waples, R. S., 2002. Relation-ship of effective to census size in fluctuatingpopulations. Conservation Biology, 16: 129–136.

Keller, L. F. & Waller, D. M., 2002. Inbreedingeffects in wild populations. Trends in Ecologyand Evolution, 17: 230–241.

Kirkpatrick, M. & Jarne, P., 2000. The effects of abottleneck on inbreeding depression and the ge-netic load. American Naturalist, 155: 154–167.

Lande, R., 1988. Genetics and demography inbiological conservation. Science, 241: 1455–1460.

– 1995. Mutation and conservation. ConservationBiology, 9: 782–791.

– 1998. Anthropogenic, ecological and genetic fac-tors in extinction and conservation. Researcheson Population Ecology, 40: 259–269.

Linklater, W. L., 2003. A novel application of theTrivers–Willard model to the problem of geneticrescue. Conservation Biology, 17: 906–909.

Lynch, M., Conery, J. & Burger, R., 1995. Mutationaccumulation and the extinction of smallpopulations. American Naturalist, 146: 489–518.

Madsen, T., Shine, R., Olsson, M. & Wittzell, H.,1999. Conservation biology: restoration of aninbred adder population. Nature, 402: 34–35.

Mills, L. S. & Allendorf, F. W., 1996. The one–migrant–per–generation rule in conservation andmanagement. Conservation Biology, 10: 1509–1518.

Mills, L. S. & Smouse, P., 1994. Demographicconsequences of inbreeding in remnantpopulations. American Naturalist, 144: 412–431.

Newman, D. & Pilson, D., 1997. Increased probabil-ity of extinction due to decreased genetic effectivepopulation size: experimental populations of Clarkiapulchella. Evolution, 51: 354–362.

Newman, D. & Tallmon, D. A., 2001. Experimentalevidence for beneficial fitness effects of geneflow in recently isolated populations. Conserva-tion Biology, 15: 1054–1063.

Nieminen, M., Singer, M. C., Fortelius, W., Schops,K. & Hanski, I., 2001. Experimental confirmationthat inbreeding depression increases extinctionrisk in butterfly populations. American Naturalist,157: 237–244.

Pimm, S. L., Jones, H. L. & Diamond, J., 1988. Onthe risk of extinction. American Naturalist, 132:757–785.

Reed, D. H. & Bryant, E. H., 2000. Experimentaltests of minimum viable population size. AnimalConservation, 3: 7–14.

– 2001. Fitness, genetic load and purging in ex-perimental populations of the housefly. Conser-vation Genetics, 2: 57–61.

Richards, C. M., 2000. Inbreeding depression andgenetic rescue in a plant metapopulation. Ameri-can Naturalist, 155: 383–394.

Saccheri, I., Kuussaari, M., Kankare, M., Vikman,P., Fortelius, W. & Hanski, I., 1998. Inbreedingand extinction in a butterfly metapopulation. Na-ture, 392: 491–494.

Spielman, D. & Frankham, R., 1992. Modeling prob-

130 Waite et al.

– 2000. Is one migrant per generation sufficient forthe genetic management of fluctuatingpopulations? Animal Conservation, 3: 261–266.

– 2001. Migration and inbreeding: the importanceof recipient population size for genetic manage-ment. Conservation Genetics, 2: 167–171.

Vucetich, J. A., Waite, T. A. & Nunney, L., 1997.Fluctuating population size and the ratio of effec-tive to census population size. Evolution, 51:2015–2019.

Vucetich, J. A., Waite, T. A., Qvarnemark, L. &Ibargüen, S., 2000. Population variability andextinction risk. Conservation Biology, 14: 1704–1714.

Wang, J. L., 2000. Effects of population structuresand selection strategies on the purging of in-breeding depression due to deleterious muta-tions. Genetical Research, 76: 75–86.

Whitlock, M. C., 2000. Fixation of new alleles andthe extinction of small populations: drift load,beneficial alleles, and sexual selection. Evolution,54: 1855–1.

lems in conservation genetics using captive Dro-sophila populations: improvement of reproduc-tive fitness due to immigration of one individualinto small partially inbred populations. Zoo Biol-ogy, 11: 343–351.

S–PLUS, 2000, Professional version 1.0, 1999.MathSoft Inc., Cambridge, Massachusetts, U.S.A.

SPSS, version 10.0., 1999. SPSS Inc., Chicago,Illinois, U.S.A.

Tanaka, Y., 2000. Extinction of populations by in-breeding depression under stochastic environ-ments. Population Ecology, 42: 55–62.

Vila, C., Sundqvist, A. K., Flagstad, O., Seddon,J., Bjornerfeldt, S., Kojola, I., Casulli, A.,Sand, H., Wabakken, P. & Ellegren, H., 2003.Rescue of a severely bottlenecked wolf (Canislupus) population by a single immigrant. Pro-ceedings of the Royal Society of London B,270: 91–97.

Vucetich, J. A. & Waite, T. A., 1999. Erosion ofheterozygosity in fluctuating populations. Con-servation Biology, 13: 860–868.



Martínez–Abraín, A., Oro, D., Belenguer, R., Ferrís, V. & Velasco, V., 2005. Long–term changes of speciesrichness in a breeding bird community of a small Mediterranean archipelago. Animal Biodiversity and Conserva-tion, 28.2: 131–136.

AbstractLong–term changes of species richness in a breeding bird community of a small Mediterranean archipelago.— Weanalyzed the pattern of species richness changes in a bird–breeding bird community on a small western Mediterra-nean archipelago (Columbretes Islands) over a 40–year period (1964–2003). The aim of this study was to qualitativelyaccount for the relative roles of local and regional factors in shaping the community. As expected, we found thatregional factors (at the metapopulation spatial scale) increased diversity whereas local factors (i.e. ecological)probably prevented further increases in diversity. We found that the archipelago gained four new species (two seabirdsand two falconids) during the study period, whereas no extinctions were recorded. The community seems partially orcompletely closed to some groups of species (e.g. small–sized birds such as passerines and storm–petrels), probablyowing to predatory exclusion by Eleonora falcons (Falco eleonorae). As newly arrived species have breedingcalendars that do not fully overlap with those of resident species, competition for space in a rather saturated area isprevented. Preservation of rare species which increase gamma (regional) diversity rather than alpha diversity withcommon species should be the main local conservation goal.

Key words: Colonization, Extinction, Diversity, Columbretes Archipelago, Conservation, Metapopulation.

ResumenCambios a largo plazo en la riqueza de especies en una comunidad de aves nidificantes en un pequeñoarchipiélago mediterráneo.— Este trabajo analiza los patrones de cambio en la riqueza de especies en unacomunidad de aves nidificantes de un pequeño archipiélago mediterráneo (las islas Columbretes, Castellón)durante un periodo de 40 años (1964–2003). El estudio pretende valorar cualitativamente la influencia relativa delos factores locales y regionales. Como se esperaba, se encontró que los factores regionales (a la escalaespacial de la metapoblación) aumentaron la diversidad, mientras que los factores ecológicos locales evitaronmayores incrementos. El archipiélago ganó cuatro especies durante el periodo de estudio (dos aves marinas ydos falcónidos), mientras que no se produjo ninguna extinción. La comunidad parece parcial o totalmentecerrada a ciertos grupos de especies, tales como las aves de pequeña talla (p.ej. Paseriformes y paíños)probablemente debido a la depredación excluyente por parte de los halcones de Eleonor (Falco eleanorae).Dado que las especies que son colonizadoras recientes tienen calendarios de cría que no se solapancompletamente con los de las especies residentes, se evita la competencia por el espacio de cría en un áreabastante saturada. La principal meta conservacionista debe ser la protección de las especies raras, queincrementan la diversidad gamma (regional), más que la diversidad alpha de las especies comunes.

Palabras clave: Colonización, Extinción, Diversidad, Columbretes, Conservación, Metapoblación.

(Received: 15 VI 04; Conditional acceptance: 17 XII 0; Final acceptance: 12 I 05)

A. Martínez–Abraín, D. Oro, Instituto Mediterráneo de Estudios Avanzados, IMEDEA (CSIC–UIB), c/ MiquelMarquès 21, 07190–Esporles, Mallorca, Spain.– R. Belenguer, V. Ferrís & R. Velasco, Reserva Natural de las IslasColumbretes, Conselleria de Territorio y Vivienda, Avda. Hermanos Bou 47, 12003–Castellón, Spain.

Corresponding author: A. Martinez–Abrain. E–mail: [email protected]

Long–term change of species richnessin a breeding bird community of asmall Mediterranean archipelago

A. Martínez–Abraín, D. Oro, R. Belenguer,V. Ferrís & R. Velasco


132 Martínez–Abraín et al.

breeding bird community includes only eight species(see table 1). Interestingly, there are no breedingpasserines on the island although they are veryabundant during migration (Giménez, 1987). Localseabird populations have been monitored annuallysince the early 80’s and annual records of breedingsuccess have been available since 1989. Data fromthe 60’s and 70’s comes from sporadic visits to theislands by expert ornithologists. The closest seabirdcolonies are located 80–100 km away, at the EbroDelta (Tarragona) and the island of Ibiza (BalearicIslands).

Information on the number and type of breedingspecies, as well as competitive interactions andenvironmental changes comes to a large extentfrom a literature review, including unpublished re-ports of the regional government (i.e. GeneralitatValenciana), over a 40–year period. Much of therecent literature on population dynamics howeverhas been generated by our group (Martínez–Abraínet al., 2001, 2002a, 2002b, 2003a, 2003b). Follow-ing the automation of the light house in 1975 overa decade elapsed without human habitation until apermanent crew of wardens (three per 15–day shift)was established on the islands.

Results

As seen in table 1 the "basal" community was madeup of two oceanic seabirds (Cory’s ShearwaterCalonectris diomedea diomedea and the EuropeanStorm–Petrel Hydrobates pelagicus), a raptor com-monly found in the Mediterranean islands (Eleonora’sFalcon Falco eleonorae) and the Yellow–legged GullLarus cachinnans, a gull species known to breed onthese islands at least since the 19th century (Salvator,1895; Bernis & Castroviejo, 1968). The first speciesto join this community was Audouin’s Gull Larusaudouinii, in 1974 (Pechuán, 1974, 1975; Gómez,1987). A second seabird species (European Shag,Phalacrocorax aristotelis) became established on theislands in 1991, although an isolated breeding at-tempt occurred in 1985 (see Martínez–Abraín et al.,2001). More recently, two new raptor species havecolonized the archipelago. A pair of peregrine fal-cons, Falco peregrinus, bred successfully in 2002and again in 2003 after several failed breeding at-tempts in the past, and a pair of European KestrelsFalco tinnunculus bred successfully during their firstattempt in 2003. Hence, during a 40–year period thearchipelago has experienced four colonization events,two of them by seabirds and two of them by birds ofprey. Seabird species can be considered as estab-lished breeders after several decades of continuousreproduction, whereas falconids cannot, owing totheir short breeding record to date. We have notincluded the presence in the colony of several breed-ing Cory’s Shearwaters from the Atlantic subspeciesCalonectris diomedea borealis (see Martínez–Abraínet al., 2002a) as a colonization event because theirtaxonomic identity (i.e. species or subspecies) isunder discussion.

Introduction

Much debate has taken place among ecologists dur-ing the last decades about the nature of communitiesand the regulation of their structure. Typically, a com-munity is now defined as an association of populationsof species in a certain area, with no fixed boundaries,and whose structure is shaped by the environment,by the interactions of the populations within the com-munity, and also by historical, regional and globalprocesses (Ricklefs & Latham, 1993). One of themain structural properties of communities is thenumber and types of species in the community (seeMagurran, 1988). This property is time–dependentand long–term records of species richness are thusnecessary to properly characterize the diversity of agiven system and to assess the role of both local andexternal factors in the pattern of change. The latteralso involve a spatial dimension and hence demo-graphic spatial processes of dispersal, typically emi-gration and immigration, which are not necessarilycorrelated with local factors. Examples of this arecolonization of empty patches or extinction of occu-pied patches, in the framework of the metapopulationtheory (e.g. Hanski, 1999).

We have extensively monitored the populationdynamics of most avian species breeding on a smallwestern Mediterranean archipelago but we have notstudied community properties as a whole to date.The aim of this paper is to describe and discusslong–term changes in species richness in a smallMediterranean island community, to derive informa-tion of conservation interest. This goal is especiallyrelevant from a regional perspective as many of thethousands of islands in the Mediterranean are verysmall and home to bird faunas which are among themost endangered in the world. (Rodríguez, 1982;Blondel & Aronson, 1999).

Material and methods

The Columbretes Islands are a small volcanic archi-pelago located close to the continental slope in thenorth–western Mediterranean, some 50 km from thecontinental coast (see fig. 1). The total area of theColumbretes archipelago is 19 ha, divided in fourmajor groups of islands. The largest island (GrossaIsland, with 13 ha. or 68% of the emerged land)holds most breeding birds. This island is horseshoe–shaped, has a maximum length of about 1,200 m, amaximum width of 220 m, a minimum width of 17 m,and a maximum height of 67 m, with steep cliffs allaround. As the archipelago is very small it is possi-ble to perform reliable surveys on breeding species.Vegetation mainly consists of a number of shrubsadapted to the arid conditions and high salinity, andannual plant species. Vegetation on the main island(Grossa Island) was deeply altered by humans dur-ing the construction of the light house in the mid–19th century when it was burnt down and pigs wereintroduced to deal with the abundant snakes (vi-pers?) on the island (Serrano, 1987). The present


Discussion

Regional factors increase alpha (local) diversity

Increases in diversity at Columbretes are betterexplained by factors external to the archipelago thanby changes in local conditions. For example, coloni-zation by Audouin’s gulls at the beginning of the 70’stook place before the islands were legally protected,when they were suffering a number of human–related disturbances. In addition, fishermen and light-house keepers had free access to the islands whichwere also used as targets for military exercises atthe time (Serrano, 1987). Audouin’s gull dynamics isknown to be very dependent on dispersal processeswithin metapopulations and the growth of this colonycannot be explained without attending to immigra-tion from the outside (Oro & Pradel, 2000; Oro &Ruxton, 2001; Oro et al., 2004). The growth of thecolony has recently been influenced by externalfactors and especially by the rescue effect of immi-grant individuals from the neighbour colony of theEbro Delta (Oro et al., 2004).

Colonization by shags coincided with a steepdecline of a large colony located in Majorca, owingto sand mining close to it that probably affected localfeeding grounds (Martínez–Abraín et al., 2001). Al-ternatively, since shag colonization occurred afterisland protection at the end of the eighties, it mightalso be related to a reduction in human disturbance.

Both Audouin’s gulls and shags are good exam-ples of the importance of emigration and immigra-tion processes in the colonization rates at themetapopulation level (Oro, 2003). At the sametime, the recent decline in Audouin’s gull numbersis also a consequence of dispersal to higher qualitysites, and extinction, in the absence of high immi-gration rates from the outside, is only buffered bythe high survival of old adult philopatric breeders(Cam et al., in press).

Colonization by falconids may respond to in-creasing numbers of these two species on thecontinent and major nearby islands during the lastdecades (g.o.b, 1997; Marti & Del Moral, 2003),and might be supported by eventually abundantand extended migratory flows of passerines duringthe spring. The magnitude of these flows is alsoindependent of local features.

Ecological factors prevent further increases in alphadiversity

Local factors (i.e. competitive or predatory exclu-sion) are known to reduce diversity, either by remov-ing species or preventing the invasion of new spe-cies. In Columbretes, predatory exclusion probablyplays a role in preventing the settlement of small–sized birds (e.g. small Passeriformes adapted to aridconditions such as warblers which are present dur-ing migration periods) owing to the large density ofbreeding Eleonora’s Falcons, that are known to preyupon all sorts of small–sized birds (Walter, 1979). Infact, during the study period falcon numbers in-

creased from ca. 20 pairs in 1964–1985 (Bernis &Castroviejo, 1968; Dolz & Díes, 1987) to the 40–45present pairs (Generalitat Valenciana, unpub. data).Falcons are present on the islands between Apriland November, excepting June when most birds flyto the mainland to feed on insects (own unpublishedobservations) and hence they are present on theislands during the breeding season of passerinesand storm–petrels. As early as the 19th century earlyvisitors to the archipelago suggested that Eleonora’sFalcons might have prevented colonization by smallbirds (see Salvator, 1895). Alternatively it could beargued that small species of long–distance migra-tory birds exhibit a great deal of biogeographicregionalism and are generally bad colonizers

Fig. 1. Map showing the location of theColumbretes Islands within the westernMediterranean and the major groups of islandswithin the archipelago.

Fig. 1. Mapa de localización de las islasColumbretes en el Mediterráneo occidental ylos principales grupos de islas de dicho archi-piélago.

Grossa Island

Ferrera

Foradada

Bergantín

1 km


(Boehning et al., 1998). It is also possible that thelow vegetation cover on some of the smaller islands,together with the small total surface area and theconsiderable distance to the continental coast, mightmake Columbretes an unsuitable target for non–migrant passerines. In addition, the small populationsof Passeriformes that the island could hold would bevery susceptible to local extinction owing to stochasticphenomena (Legendre et al., 1999). Eleonora’s Fal-con might also impose some pressure on the smallEuropean Storm–Petrel, which is scarce atColumbretes despite the seemingly favourable con-ditions of the islands. Although storm–petrels werenot found in the diet of Columbretes falcons by Dolz& Díes (1987), Columbretes wardens have foundremains of storm–petrels predated by falcons inseveral occasions (G. Urios, pers. comm.). In addi-tion, they have appeared as a component in theEleonora’s Falcon diet at other colonies such as theCabrera archipelago (Balearic Islands) where forag-ing activity at dusk on nocturnal insects, night–dwelling arthropods such as scorpions and even onbats, has been reported (Araujo et al., 1977; Suárez,2000). Additional support to our hypothesis comesfrom Bernis & Castroviejo (1968) who reported on astorm–petrel being attacked by Eleonora’s falconwhen it was released after ringing at Columbretes.Interestingly, the largest colonies of storm–petrelalong the eastern Iberian coast (e.g. Benidorm Is-land) occur on islands where Eleonora’s Falcons areabsent, despite the presence of suitable breedinghabitat.

Evidence of competitive exclusion also comesfrom the breeding calendar of newly arrived spe-cies. Interestingly, new species colonizing the

Columbretes Islands do not overlap in time withresident species, segregating over the breedingseason. Shags coincide with Eleonora’s falcons intheir preference for cliffs and crevices located out ofthe direct influence of solar radiation (Urios, 2003;own observations) and indeed some shag nests areplaced on ledges previously used by falcons. Theoptimal falcon nesting site is virtually occupied atthe moment (op. cit.), but there is no conflict withshags regarding nesting sites because the latterstart breeding in December–January, whereas fal-cons do so in July. Syntopic breeding of Audouin’sand Yellow–legged Gulls has been possible to alarge extent because the breeding calendar of theformer species is delayed about one month inrelation to that of the latter. Audouin’s Gulls there-fore occupy the space not taken by the Yellow–legged gull latter. Similarly, newly arrived Falcospecies (table 1) complete their reproduction well inadvance of Eleonora’s Falcons starting to buildtheir nests, thus avoiding conflict even though nest-ing–site preferences seem to coincide.

Island diversity and human activities

There is no evidence in the presently availableliterature on bird extinctions during the profoundenvironmental transformations associated with theconstruction of the light house and later humanoccupancy. Similarly, no extinction took place dur-ing the 40–year period considered by this study.This is at least partially explained by the life historytraits of most birds breeding at the study site: highadult survival, large generation times and relativelylow rates of population growth (negative or posi-

Table 1. Historical changes in the number of all breeding species recorded on the ColumbretesIslands (NW Mediterranean) between 1964–2003. A positive sign indicates that a species waspresent as a breeder and a negative sign indicates the opposite. The archipelago became legallyprotected in 1988.

Tabla 1. Cambios históricos en el número de todas las especies nidificantes observadas en las islasColumbretes (Mediterráneo nordoccidental), entre 1964 y 2003. Los signos positivos indican que laespecie estaba presente como nidificante, y los negativos, la situación opuesta. El archipiélagoempezó a estar bajo protección legal en 1988.

Species 1964–1973 1974–1983 1984–1993 1994–2003

Calonectris diomedea + + + +

Hydrobates pelagicus + + + +

Larus cachinnans + + + +

Falco eleonorae + + + +

Larus audouinii – + + +

Phalacrocorax aristotelis – – + +

Falco peregrinus – – – +

Falco tinnunculus – – – +


tive). However, human disturbance is an importantcomponent to understand patterns of species diver-sity, especially in the Mediterranean basin (Blondel& Aronson, 1999). In this region, population dy-namics and distribution patterns of many specieshave been shaped more by human activities thanby evolutionary determinants (op. cit.). Once directhuman impacts were removed at Columbretes byconservation laws and especially by effectively pro-tecting the site, only colonizations have occurred.However, human activities are changing rapidlyand could drive several species to local extinction.Indeed since 1991, Audouin’s Gulls have locallyfaced a steep decline owing to reduced food avail-ability during the chick–rearing period when a trawl-ing fishing moratorium was established, deprivinggulls from fishing discards, their main food sourcein the area since colonization of the archipelago(Oro et al., 2004). The declining colony is nowsuffering high levels of disturbance by Yellow–leg-ged Gulls, probably as a consequence of changedpredator/prey ratios (Oro et al., in press). Addition-ally, Cory’s Shearwaters seem to be following adecreasing trend at Columbretes because of adultmortality in long–line fishing gear (Belda & Sánchez,2001; Cooper et al., 2003).

Island biodiversity and conservation goals

Increases in diversity are not good per se. Pro-vided that our hypothesis is true, a decrease inbreeding numbers of Eleonora’s Falcons couldlead to the colonization of the islands by smallpasserines and probably to larger numbers ofbreeding storm–petrels, increasing local diversity.However, the preservation and growth of the colonyof this endemic Mediterranean raptor is a moredesirable conservation goal than the gain of com-mon species which are abundant elsewhere. Simi-larly, colonization by a common raptor species,such as the Common Kestrel, could result in in-creased predation of endemic lizards and beatlesreducing the diversity of other animal taxa. Localmanagement efforts should hence focus on pro-moting the persistence of rare species such asAudouin’s Gulls and Eleonora’s Falcons whichincrease gamma (regional) diversity.

Acknowledgements

This study is a contribution to the LIFE02NATURE/E/8608 for the conservation of Audouin’s Gull in theComunidad Valenciana, financed by the GeneralitatValenciana and the European Union. We are mostgrateful to all the wardens of the Columbretes Is-lands who have monitored bird populations since1988. We are also grateful to Juan Jiménez, JoséVicente Escobar and Josep Carda for promoting thescientific study of the Columbretes birds. R. E.Ricklefs, M. Giménez, J. L. Tella and an anonymousreviewer read early drafts of the manuscript andprovided substantial suggestions.

References

Araujo, J., Muñoz–Cobo, J. & Purroy, F. J., 1977.Las rapaces y aves marinas del archipiélago deCabrera. Naturalia Hispanica, 12. Madrid.

Belda, E. & Sánchez, A., 2001. Seabird mortality onlongline fisheries in the western Mediterranean:factors affecting bycatch and proposed mitigatingmeasures. Biological Conservation, 98: 357–363.

Bernis, F. & Castroviejo, J., 1968. Aves de las islasColumbretes en primavera. Ardeola, 12: 143–163.

Blondel, J. & Aronson, J., 1999. Biology and wildlifeof the Mediterranean region. Oxford Univ. Press,Oxford.

Boehning, G. K., Gonzalez–Guzman, L. I. & Brown,J. H., 1998. Constraints on dispersal and theevolution of the avifauna of the Northern Hemi-sphere. Evolutionary Ecology, 12: 767–783.

Cam, E., Oro, D., Pradel, R. & Jimenez, J., inpress. Assessment of hypotheses about disper-sal in a long–lived seabird using multistate cap-ture–recapture models. Journal of Animal Ecol-ogy, 73.

Cooper, J., Baccetti, N., Belda, E. J., Borg, J. J.,Oro, D., Papaconstantinou, C. & Sánchez, A.,2003. Seabird mortality from longline fishing inthe Mediterranean sea and Macaronesian wa-ters: a review and a way forward. Scientia Ma-rina, 67: 57–64.

Dolz, J. C. & Dies, N., 1987. El halcón de Eleonor(Falco eleonorae, Gené) en las islas Columbretes.In: Islas Columbretes: contribución al estudio desu medio natural: 241–262 (L. A. Matilla, J. L.Carretero & A. García–Carrascosa, Eds.).Generalitat Valenciana, Valencia.

Giménez, M., 1987. Notas sobre migración de avesen las islas Columbretes. In: Islas Columbretes:contribución al estudio de su medio natural: 205–214 (L. A. Matilla, J. L. Carretero & A. García–Carrascosa, Eds.). Generalitat Valenciana, Va-lencia.

G.O.B., 1997. Atles dels aucells nidificants deMallorca i Cabrera. G.O.B., Palma de Mallorca.

Gómez, J. A., 1987. Láridos nidificantes en lasIslas Columbretes: Gaviota Patigualda (Laruscachinnans) y la gaviota de Audouin (Larusaudouinii). In: Islas Columbretes: contribución alestudio de su medio natural: 215–222 (L. A.Matilla, J. L. Carretero & A. García–Carrascosa,Eds.). Generalitat Valenciana, Valencia.

Hanski, I., 1999. Metapopulation Ecology. OxfordUniv. Press, Oxford.

Legendre, S., Clobert, J., Møller A. P & Sorci, G.,1999. Demographic stochasticity and social mat-ing system in the process of extinction of smallpopulations: the case of passerines introducedto New Zealand. The American Naturalist, 153:449–463.

Magurran, A. E., 1988. Ecological diversity and itsmeasurement. Princeton Univ. Press, Princeton, NJ.

Marti, R. & Del Moral, J. C., 2003. Atlas de las avesreproductoras de España. Dirección General deConservación de la Naturaleza–Sociedad


Española de Ornitología, Madrid.Martínez–Abraín, A., González–Solís, J., Pedrocchi,

V., Genovart, M., Abella, J. C., Ruiz, X., Jiménez,J. & Oro., D., 2003a. Kleptoparasitism, distur-bance and predation of yellow–legged gulls onAudouin’s gulls in three colonies of the westernMediterranean. Scientia Marina, 67 (Suppl. 2):89–94.

Martínez–Abraín, A., Oro, D., Forero, M. G. &Conesa, D., 2003b. Modelling temporal and spa-tial colony–site dynamics in a long–lived seabird.Population Ecology, 45: 133–139.

Martínez–Abraín, A., Oro, D., Freís, V. & Belenguer,R., 2002b. Is growing tourist activity affectingthe distribution or number of breeding pairs in asmall colony of the Eleonora’s Falcon? AnimalBiodiversity and Conservation, 25.2: 47–51.

Martínez–Abraín, A., Oro, D. & Jiménez, J., 2001.The dynamics of a colonization event in the Euro-pean shag: the roles of immigration and demo-graphic stochasticity. Waterbirds, 24: 97–102.

Martínez–Abraín, A., Sánchez, A. & Oro, D., 2002a.Atlantic Cory’s shearwaters breeding in a colonyof Mediterranean Cory’s shearwaters. Waterbirds,25: 221–224.

Oro, D., 2003. Managing seabird metapopulationsin the Mediterranean: constraints and challenges.Scientia Marina, 67: 13–22.

Oro, D., Cam, E., Pradel, R. & Martínez–Abrain, A.,2004. Influence of food availability on demogra-phy and local population dynamics in a long–lived seabird. Proceedings of the Royal SocietyLondon, Series B, 271: 387–396.

Oro, D., Martínez–Abraín, A., Paracuellos, M., Nevado,J. C. & Genovart, M., in press. Influence of densitydependence on predator–prey seabird interactionsat large spatio–temporal scales. Proceedings ofthe Royal Society of London, Series B.

Oro, D. & Pradel, R., 2000. Determinants of local

recruitment in a growing colony of Audouin’sgull. Journal of Animal Ecology, 69: 119–132.

Oro, D. & Ruxton, G. D., 2001. The formation andgrowth of seabird colonies: Audouin’s gull as a casestudy. Journal of Animal Ecology, 70: 527–535.

Pechuán, L., 1974. La colonia de Larus audouiniien las islas Columbretes. Ardeola, 20: 358–359.

– 1975. Nidificación de Larus audouinii en las islasColumbretes. Ardeola, 21: 407–408.

Ricklefs, R. E. & Latham., R. E., 1993. Globalpatterns of diversity in mangrove floras. In: Spe-cies diversity in ecological communities: historicaland geographical perspectives: 215–219 (R. E.Ricklefs & D. Schluter, Eds.). Univ. of ChicagoPress, Chicago.

Rodríguez, J., 1982. Oceanografía del marMediterráneo. Editorial Pirámide, Madrid.

Salvator, L. von., 1895. Columbretes. Heinr. Mercyund Druck Verlag, Praga. (Traducción española:G. Urios & J. Nachtwey, 1990. Columbretes.Ayto. de Castellón, Castellón).

Suárez, M., 2000. Las rapaces nidificantes en elarchipiélago de Cabrera. In: Las aves del ParqueNacional marítimo–terrestre del archipiélago deCabrera (Islas Baleares, España): 233–252 (G.X. Pons, Ed.). G.O.B., Colecciones Técnicas delMinisterio de Medio Ambiente, Madrid.

Serrano, R., 1987. Historia de los asentamientoshumanos en Columbretes. In: Islas Columbretes:Contribución al estudio de su medio natural: 13–18(L. A. Matilla, J. L. Carretero & M. García–Carrascosa,Eds.). Generalitat Valenciana, Valencia.

Urios, G., 2003. Aplicación de los sistemas deinformación geográfica y modelos digitales delterreno a la cartografía ambiental en las islasColumbretes. Tesis doctoral, Univ. de València.

Walter, H., 1979. Eleonora’s Falcon. Adaptations toprey and habitat in a social raptor. The Univ. ofChicago Press, Chicago.



Komonen, A. & Kouki, J., 2005. Occurrence and abundance of fungus–dwelling beetles (Ciidae) in borealforests and clearcuts: habitat associations at two spatial scales. Animal Biodiversity and Conservation, 28.2:137–147.

AbstractOccurrence and abundance of fungus–dwelling beetles (Ciidae) in boreal forests and clearcuts: habitatassociations at two spatial scales.— Insect material (> 30,000 individuals) reared from the fruiting bodies ofwood–decaying Trametes fungi was compared between old–growth boreal forests and adjacent clearcuts inFinland. Sulcacis affinis and Cis hispidus occurred more frequently and were, on average, more abundantin the clearcuts. Interestingly, Octotemnus glabriculus and Cis boleti had a slightly higher frequency ofoccurrence in the forests, despite lower resource availability. The former also showed a higher averageabundance. On average, the cluster size of Trametes fruiting bodies occurring on woody debris was higherin the clearcuts than in the forests and had a positive effect on species occurrence and abundance in theseclusters. The independent effect of the macrohabitat (forest or clearcut) underscores the importance of themacrohabitat where specific resources occur, and this may override the positive effects of resourceavailability.

Key words: Forest landscape, Boreal forest, Coarse woody debris, Wood–decaying fungi, Trametes, Ciidae.

ResumenPresencia y abundancia de los escarabajos fungícolas (Ciidae) en los bosques y claros de tala boreales:asociaciones al hábitat según dos escalas espaciales.— Se compararon las cantidades de insectos(> 30,000 individuos) que se alimentan de los cuerpos fructíferos de los hongos desintegradores de lamadera Trametes en los bosques boreales maduros y los claros adyacentes en Finlandia. Sulcacis affinisy Cis hispidus aparecían con mayor frecuencia, y en promedio eran más abundantes en los claros. Llamala atención la frecuencia ligeramente mayor de Octotemnus glabriculus y Cis boleti en los bosques, a pesarde una menor disponibilidad de recursos. El primero también presentaba una abundancia promedio mayor.En promedio, el tamaño de las masas de cuerpos fructíferos de Trametes de los restos de árboles eramayor en los claros que en los bosques, y tenía un efecto positivo en la presencia y abundancia de especiesen dichas masas. El efecto independiente del macrohábitat (bosque o claro) subraya la importancia delmacrohábitat cuando los recursos específicos aparecen, pudiendo anular los efectos positivos de ladisponibilidad de recursos.

Palabras clave: Paisaje forestal, Bosque boreal, Restos gruesos de madera, Hongos desintegradores de lamadera, Trametes, Ciidae.

(Received: 20 IX 04; Conditional acceptance: 17 XI 04; Final acceptance: 12 I 05)

Atte Komonen & Jari Kouki, Fac. of Forest Sciences, P. O. Box 111, FI–80101, Univ. of Joensuu, Finland.

Corresponding author: A. Komonen. E–mail: [email protected]

Occurrence and abundance offungus–dwelling beetles (Ciidae) inboreal forests and clearcuts:habitat associations at twospatial scales

A. Komonen & J. Kouki


138 Komonen & Kouki

Introduction

Ecological research on habitat fragmentation hasoften metaphorically viewed suitable habitats as"islands" in a hostile "sea" (Haila, 2002). Be-cause there is a low degree of deforestation inthe boreal forest, landscapes rarely appear sim-ply as a black–and–white contrast between habi-tat and non–habitat, rather the many shades ofgrey ref lect d i fferent ia l habi tat sui tabi l i ty(Mönkkönen & Reunanen, 1999). Forestry prac-tices in boreal forests create highly dynamiclandscapes which remain forested while under-going spatial and temporal changes in structureand dynamics (Kouki et al., 2001; Schmiegelow& Mönkkönen, 2002). The structural heterogene-ity in the forest landscape is manifested in theamount and quality of living and dead wood(Siitonen, 2001), forest–stand age structure andnaturalness (Uotila et al., 2002), and in thedistribution of different stand types in the land-scape (Löfman & Kouki, 2003).

Dead wood is a key resource for thousands oforganisms in boreal forests (Esseen et al., 1997;Siitonen, 2001). Dead wood is scarce in man-aged forests and many dead–wood dependentorganisms are consequently absent or occur inlow numbers. This has contributed to the mis-conception that the majority of these specieswould require natural forests. However, manydead–wood dependent species can persist inclearcuts if critical resources are left in ad-equate densities and qualities in management

operations (Kaila et al., 1997; Jonsell et al.,2001; Jonsson et al., 2001; Martikainen, 2001;Sverdrup–Thygeson & Ims, 2002). However, theimportance of the macrohabitat and landscapecontext where these resources occur is poorlyunderstood, particularly for species that havehigh habitat specificity and limited powers ofdispersal, such as fungus–dwelling insects(Jonsell et al., 1999; Jonsson et al., 2001;Komonen et al., 2000). It is important to disen-tangle the large– and small–scale effects offorestry on habitat suitability to fully understandthe limiting factors for species occurrences andabundances (Mönkkönen & Reunanen, 1999).

Trametes fruiting bodies occur in a variety offorest environments where deciduous dead woodis available. Nevertheless, it is not known howTrametes–dwelling beetles respond to the differ-ent forest surroundings. There are generallymarked ecological differences among the fun-gus–dwelling insects in dispersal ability (Jonsellet al., 1999; Jonsson, 2003), habitat require-ments (Nilsson, 1997; Guevara et al., 2000a;Thunes et al., 2000; Jonsell et al., 2001) andhost–fungus specificity (Lawrence, 1973; Økland,1995; Fossli & Andersen, 1998; Guevara et al.,2000b; Komonen, 2001). In this paper, the oc-currence and abundance of four fungivorousbeetle species (Ciidae) co–occurring in Trametesfruiting bodies is investigated at two spatial scales.The small–scale effects of fungal–cluster sizeand the large–scale effects of macrohabitat (for-est–clearcut) are tested.

Table 1. Study site characteristics. Area of forest refers to the area (ha) with > 60 m3 fallen woodydebris (diameter at breast height ≥ 7 cm) ha–1: As. Area sampled; Af. Area of the forest; Ac. Areaof the clearcut, refers to the area (ha) logged at the same time and consists of interconnectedopenings rather than a single large opening; CWD. Number of stumps, snags, logs and branches(diameter ≥ 10 cm) of birch and aspen ha–1; Yl. Year of logging; MAl. Mean age at the time oflogging, refers to the mean age of trees belonging to the dominant canopy storey; Vl. Volume atlogging, refers to the volume of living spruce, pine and birch at the time of logging.

Tabla 1. Características del área de estudio. Área forestal se refiere al área (ha) con > 60 m3 deresiduos de madera caída (diámetro a la altura del tórax ≥ 7 cm) ha–1: As. Área muestreada; Af. Áreade bosque; Ac. Área de claros, se refiere al área (ha) talada al mismo tiempo, y consiste en variosclaros interconectados, más que en uno solo de mayor tamaño; CWD. Número de tocones, cepas,troncos y ramas (diámetro ≥ 10 cm) de abedules y álamos temblones ha–1; YI. Edad de talado; MAI.Edad media de talado, se refiere a la edad media de los árboles del piso dominante del dosel forestal;Vl. Volumen cuando la tala, se refiere al volumen de pinos, álamos y píceas en el momento de la tala.

Forest Clearcut

Study site As Af CWD A Yl MAl Vl CWD

1. Ruunavaara 4.0 37 50 51 1994 155 239 48

2. Pieni Hovinvaara 2.0 19 59 54 1990/1993 145/128 226/268 55

3. Haapahasianvaara 3.5 83 54 58 1993 127 197 111

4. Vankonvaara 4.5 64 52 22 1994 132 223 52


Information on Ciidae life history is limited. Thelife history is completed within the same piece offungus and several generations may occur beforethe adults emigrate to find a fresh piece of fungus(Entwistle, 1955). The adults are long–lived for aninsect (up to seven months; Klopfenstein, 1971)and during the summer all life–stages may befound at the same time. Three of the speciesincluded in this study [Sulcacis affinis (Gyllenhal),Octotemnus glabriculus (Gyllenhal) and Cis boleti(Scopoli)] lay eggs singly. The adults copulate atintervals during the oviposition period of the femaleand are not monogamous (Entwistle, 1955). Thelatter two species, in addition to Cis hispidus(Paykull) encountered in this study, are specialistsin Trametes (Fossli & Andersen, 1998; Guevara etal., 2000b; Selonen, 2004).

Study sites

The study area is located in North Karelia, easternFinland, in the middle boreal zone (63o 00’ – 30’ N,30o – 31o E). First, 12 spruce–dominated old–growth stands were visited. They were rich inaspen (Kouki et al., 2004) and thus assumed tohave a sufficiently high density of Trametes toprovide adequate sample size. Only four standswere adjacent to clearcuts and these were se-lected for the study: 1. Ruunavaara; 2. Pieni

Materials and methods

Study system

Four species of Trametes (formerly included inCoriolus Quél.) occur in the study region in east-ern Finland. Of these, Trametes ochracea (Pers.)Gilb. & Ryvarden is the most common speciesinhabiting dead deciduous trees, mainly aspen(Populus tremula L.) and birch (Betula spp.)(Niemelä, 2001). Over 95% of our samples wereT. ochracea (K. Junninen det.), but as the fruitingbodies of all Trametes species are very similar inphysical structure (Ryvarden & Gilbertson, 1993)it was impossible to identify some of the heavily–consumed samples with certainty. As far as it isdocumented, there are no great differences in theCiidae fauna associated with the different speciesof Trametes found in Fennoscandian boreal for-ests (Fossli & Andersen, 1998; Selonen, 2004).Trametes fruiting bodies are annual and typicallyoccur in a relatively early phase of decay succes-sion (3–7 yrs; Hintikka, 1993). They are commonin woody debris and stumps, and typically formclusters of fruiting bodies. Although the fruitingbodies of Trametes are annual, dead fruiting bod-ies can remain attached to wood for one to twoyears and become entirely consumed by insects,mainly larvae and adults of Ciidae.

Fig. 1. A. The mean ± SE fungal density in the study sites (density is measured as the number ofoccupied pieces of woody debris ha–1); B. The mean ± SE fungal–cluster weight in the study sites, notethat y–axis starts from 0.6: Black circles represent forests and open circles indicate clearcuts.

Fig. 1. A. Densidad fúngica media ± EE en las áreas de estudio (la densidad se mide como el númerode trozos de restos de madera ocupados ha–1); B. Peso medio ± EE de las masas fúngicas en lasáreas de estudio; obsérvese que el eje y parte de 0,6: Círculos negros representan los bosques;círculos vacíos, los claros.

1 2 3 4 1 2 3 4 Study sites Study sites

Fu

ng

al d

ensi

ty h

a–1 (

SE

)25

20

15

10

5

0

Mea

n l

og

. fu

ng

al w

eig

ht

(SE

)

1.4

1.2

1.0

0.8

0.6

A B

140 Komonen & Kouki

Hovinvaara; 3. Haapahasianvaara; 4. Vankonvaara(table 1). As not all stands had large enough (≥ 2 ha)adjacent clearcuts, two of the sampling quadrats(see below) in site 4, and one in site 2 were locatedfurther from the studied forest, yet within 1–kmdistance and adjacent to other patches of old–growth. All the clearcut areas had once been part ofthe larger old–growth forest but had been logged 7or 8 years earlier; part of the clearcut in site twohad been logged 11 years earlier (data from Forestand Park Service, Lieksa). This oldest clearcut areawas the only one with 2–3 m tall birch trees, theothers were fully open. The small difference be-tween the mean fungal density in site two resultsfrom this difference (fig. 1).

Sampling

Fungal samples were collected between 22 IX and6 X 01, the optimal time given the species phenol-ogy. One hectare study quadrats were randomlypositioned and marked in the field; in the forests,quadrats were at least 50 m from the forest–clearcutedge. An equal area was sampled in a given old–growth forest and the adjacent clearcut (table 1).The sampled areas varied in size among forest–clearcut pairs, because larger forest stands allowedlarger sampling coverage, thus increasing samplesize and statistical power. However, the area sam-pled does not consistently follow the area of forestsdue to the discrepancy between the forest area andthe area and shape of the adjacent clearcuts. Due tothe shape and size of the clearcut in site 4, it was

impossible to establish all the sample units as 1–haquadrats; instead we used 200 m x 50 m strips. Insites 3 and 4, two 50 m x 50 m quadrats were usedfor sampling the area of 0.5 ha.

In all study quadrats we examined all woodydebris (dbh ≥ 10 cm) of deciduous trees for Trametesfruiting bodies. As Trametes fruiting bodies oftenoccur in tight clusters, insect larvae could potentiallymove from one fruiting body to another. Thus, all thefruiting bodies in a cluster were considered onesample and carefully removed. If the fungal clusterswere located in separate parts of the same individualtree, for example in a stump and a trunk not at-tached to each other, these were considered sepa-rate samples. Samples were transferred to mesh–net covered plastic boxes and kept in outdoor condi-tions. On 8 II 02, all the samples were transferred toroom temperature and weighed after two weeks, aperiod which was considered adequate for excesswater to evaporate and make the weight of samplescollected under different daily weather conditionsmore comparable. The fungal cluster size was deter-mined by weighing, reflecting both the number andsize (g) of fruiting bodies. All the fruiting bodies werecarefully dissected and the adult insect individualswere removed and identified.

Statistical analyses

Generalized Linear Models (GLM; McCullagh &Nelder, 1989) were used to analyze the data. In allmodels, site was introduced as a random effectand management category (forest or clearcut) as

Table 2. Number of individuals and percent of samples occupied by ciid beetles in this study inforests and clearcuts.

Tabla 2. Número de individuos y porcentaje de muestras ocupadas por escarabajos cíidos en esteestudio, en los bosques y en los claros de tala.

Forest Clearcut

Species Indiv. % Indiv. %

Sulcacis affinis 80 9 18,125 87

Octotemnus glabriculus 4,096 80 3,529 72

Cis hispidus 508 53 3,719 87

Cis boleti 239 51 570 50

Sulcacis fronticornis 0 0 254 9

Cis comptus 1 1 94 9

Cis glabratus 13 5 19 4

Ennearthron laricinum 7 1 1 0.4

Orthocis alni 1 1 6 2

Cis lineatocribratus 0 0 4 1

Cis jacquemartii 1 1 2 1

Ennearthron cornutum 2 1 1 0.4


a fixed effect nested within site. Fungal–clusterweight was included as a continuous covariate.For the distribution of a species (i.e. the presenceor absence of a species in a cluster; a binary

response) a binomial error distribution and a logitlink–function were assumed. For the abundance ofa species (the average number of individuals percluster) an identity link–function and normally dis-

Table 3. GLM results on the fungal–clusteroccupancy for four ciid beetles. The changein deviance indicates the model improvementwhen a given term is included in the model.The full model was compared to a model withonly a constant term: F. Full model; Mng.Management nested within site; S. Site; W.Weight. Rd. Residual deviance

Tabla 3. Resultados GLM de la ocupación delas masas fúngicas por parte de cuatrocoleópteros cíidos. El cambio de la desvianzaindica la mejora del modelo cuando se incluyeen éste un término dado. El modelo completose comparó con un modelo con un únicotérmino constante: F. Modelo completo; Mng.Manejo anidado dentro del área; S. Área; W.Peso. Rd. Desvianza residual.

Change in deviance

Ciid species d.f. χ2 P

Sulcacis affinis

F 8 277.86 0.000

Mng 4 185.76 0.000

S 3 2.87 0.412

W 1 47.99 0.000

Rd 276 156.74 1.000

Octotemnus glabriculus

F 8 80.18 0.000

Mng 4 27.11 0.000

S 3 12.83 0.005

W 1 55.43 0.000

Rd 276 262.94 0.704

Cis hispidus

F 8 111.05 0.000

Mng 4 24.12 0.000

S 3 1.66 0.647

W 1 61.30 0.000

Rd 276 225.17 0.988

Cis boleti

F 8 99.49 0.000

Mng 4 12.57 0.014

S 3 8.11 0.044

W 1 91.53 0.000

Rd 276 315.30 0.052

Table 4. Parameter estimates ± SE give thechange in the log of the odds for a ciid speciesoccurring or not occurring in a fungal clustersampled from forests, holding fungal–clusterweight constant. Estimates are taken from theGLM in table 3: Mng. Management; W. Weight.(* Unstable estimate as this ciid species wasabsent from the forest in this site.)

Tabla 4. Las estimas de los parámetros ± EEnos dan el cambio en el logaritmo de losvalores predichos/observados ("odds") paracada especie de cíido, que aparece o no, enlas masas fúngicas muestreadas en losbosques, siendo constante el peso de dichasmasas fúngicas. Las estimas se tomaron delos valores GLM de la tabla 3: Mng. Control;W. Peso. (* Estima inestable cuando estaespecie de cíido estaba ausente del bosqueen ese lugar.)

Ciid species Estimate ± SE

Sulcacis affinis

Mng (site 1) –1.63 ± 0.35

Mng (site 2) –3.08 ± 0.79

Mng (site 3) –7.28 ± 26.6*

Mng (site 4) –2.10 ± 0.35

W 2.88 ± 0.48


Mng (site 1) 1.25 ± 0.30

Mng (site 2) 0.31 ± 0.45

Mng (site 3) 0.20 ± 0.30

Mng (site 4) 0.82 ± 0.32

W 2.38 ± 0.36

Cis hispidus

Mng (site 1) –0.68 ± 0.29

Mng (site 2) –1.34 ± 0.62

Mng (site 3) –0.35 ± 0.31

Mng (site 4) –0.85 ± 0.26

W 2.63 ± 0.38

Cis boleti

Mng (site 1) 0.79 ± 0.28

Mng (site 2) 0.30 ± 0.41

Mng (site 3) 0.54 ± 0.29

Mng (site 4) 0.17 ± 0.24

W 2.74 ± 0.34

142 Komonen & Kouki

Fig. 2. Percent of fungal clusters occupied by ciid species in the four study sites; black bars representforests and grey bars show clearcuts.

Fig. 2. Porcentaje de masas fúngicas ocupadas por especies de cíidos en las cuatro áreas estdiadas;las barras negras representan los bosques, y las grises los claros.

The average fungal–cluster weight was also con-sistently lower in the forest than in the clearcutsites (fig. 1B; change in deviance = 1.760,F4, 343 = 8.560, P < 0.000).

Patterns of beetle occupancy

A total of 32,193 insect individuals were removedfrom the sampled fruiting bodies. These included12 species of Ciidae, the four most common spe-cies of which (96% of all insect individuals) wereused in the analyses (table 2). Fungal–clusterweight had a positive effect of similar magnitudeon the probability of a cluster being occupied forall four species (tables 3, 4). Sulcacis affinis andCis hispidus had a very consistent pattern of oc-currence in al l sites, in comparison withOctotemnus glabriculus and Cis boleti, in that theywere more frequent in clearcuts (fig. 2, table 4).Interestingly, O. glabriculus and C. boleti were

tributed errors were assumed. The significance ofeach term was evaluated based on the increase indeviance when the term was dropped from the fullmodel containing all explanatory parameters. Fun-gal–cluster weight and the number of individualswere log10(x+1)–transformed in all analyses, un-less otherwise stated.

Results

Host availability

Altogether, 106 and 245 pieces of woody debrisoccupied by dead fruiting bodies of Trameteswere recorded and sampled from forests andclearcuts, respectively. The density of woody de-bris occupied by Trametes was consistently lowerin the forest sites (mean ha–1 ± SE = 7.34 ± 0.63)than in the clearcut sites (16.00 ± 2.97; fig. 1A).

100

80

60

40

20

0

100

80

60

40

20

0

100

80

60

40

20

0

100

80

60

40

20

0

% o

f cl

ust

ers

occ

up

ied

1 2 3 4 1 2 3 4


Sulcacis affinis Octotemnus glabriculus

Cis hispidus Cis boleti


more likely to be found from forests, as indicatedby signs of parameter estimates (table 4), despitelower fungal availability (fig. 1). For the two "forestspecies", the predicted probability of occurrenceincreased more slowly as a function of fungal–cluster weight in the clearcut clusters than in theforest clusters (fig. 3). The opposite pattern wasobserved for the two "clearcut species".

Patterns of beetle abundance

Sulcacis affinis and C. hispidus were more abundanton average in the clearcut than in the forest clusters,and O. glabriculus was more abundant in the forestclusters, after controlling for fungal weight (table 5;fig. 4). Cis boleti did not show significant differencein abundance between a forest and a clearcut. It wasfurther tested whether there were interspecific differ-

ences in average abundances in forest and clearcutclusters. In the forests, O. glabriculus had a largermean population size in the clusters than S. affinis,C. hispidus and C. boleti (F3, 200 = 35.273, r2 = 0.346,P < 0.05; Dunnett’s C for pairwise comparisons). Inthe clearcuts, S. affinis had a significantly largerpopulation size than the other three species and C.boleti population size was significantly smaller(F3, 722 = 112.343, r2 = 0.318, P < 0.05; Dunnett’s Cfor pairwise comparisons).

Discussion

Findings from this study demonstrate that themacrohabitat where specific resources occur is im-portant for ciid beetles in Trametes fruiting bodies (c.f.Thunes & Willassen, 1997; Jonsell et al., 2001;

Fig. 3. The predicted probability of occurrence of the four ciid species in the fungal clusters as afunction of cluster weight (untransformed), based on a logistic regression model. Figures give theestimated odds ratios and 95% CIs; black dots represent forests, grey dots clearcuts.

Fig. 3. Probabilidad predicha de la presencia de las cuatro especies de cíidos en las masas fúngicas,como función del peso de dichas masas (no transformado), basada en un modelo logístico deregresión. Las cifras se refieren a la estima de la razón entre los valores predichos y observados("odds ratio") y los CIs 95%; los círculos negros representan los bosques, y los grises los claros.

1.0

0.8

0.6

0.4

0.2

0.0

1.0

0.8

0.6

0.4

0.2

0.0

1.0

0.8

0.6

0.4

0.2

0.0

1.0

0.8

0.6

0.4

0.2

0.0

0 20 40 60 80 100 120 0 20 40 60 80 100 120

0 20 40 60 80 100 120 0 20 40 60 80 100 120Fungal weight (g) Fungal weight (g)

Pre

dic

ted

pro

bab

ility

of

occ

urr

ence

Sulcacis affinis 1.107 (1.053–1.165) 1.064 (1.026–1.103)

Octotemnus glabriculus 1.052 (1.028–1.077) 1.133 (1.022–1.255)

Cis hispidus 1.111 (1.055–1.171) 1.121 (1.053–1.193)

Cis boleti 1.048 (1.029–1.067) 1.139 (1.065–1.218)

144 Komonen & Kouki

Table 5. GLM results on the mean abundance of ciid beetles in fungal clusters. The change indeviance indicates the model improvement when a given term is included in the model. Theinteraction term between management category and weight was non–significant for each species(F < 2.63, P > 0.1), and thus it was excluded from the final models: η2. Proportion of explainedvariance; Mng. Management nested within site; W. Weight; Rd. Residual deviance; * d.f. is only3 because one site did not include any occupied fungal clusters).

Tabla 5. Resultados GLM de la abundancia media de coleópteros cíidos en las masas fúngicas. Elcambio de la desvianza indica la mejora del modelo cuando se incluye en éste un término dado. Eltérmino de interacción entre la categoría de gestión y el peso para cada especie (F < 2,63, P > 0,1)no fue significativo, por lo que se excluyó de los modelos finales: η2. Proporción de varianzaexplicada; Mng. Manejo anidado dentro del área; W. Peso; Rd. Desvianza residual; * d.f. es sólo de3 debido a que uno de los lugares no incluía ninguna masa fúngica ocupada).

Species

Source d.f. Change in deviance F P η2

Sulcacis affinis

Mng 3* 3.775 17.113 0.000 0.193

Site 3 3.064 0.376 0.779 0.287

W 1 27.736 125.723 0.000 0.370

Rd 214 0.221


Mng 4 3.405 16.415 0.000 0.207

Site 3 0.263 0.080 0.968 0.056

W 1 14.665 70.686 0.000 0.219

Residual deviance 252 0.207

Cis hispidus

Mng 4 0.552 4.525 0.001 0.065

Site 3 0.383 0.705 0.597 0.344

W 1 16.675 136.666 0.000 0.344


Cis boleti

Mng 4 0.128 1.568 0.185 0.036

Site 3 0.207 1.634 0.311 0.539

W 1 2.443 29.815 0.000 0.151


Jonsson et al., 2001). Most interestingly, someciids seem to prefer forests over clearcuts despitelower fungal density and smaller size of fungalclusters in the former. In the absence of experi-ments, however, we cannot make explicit causalinferences. We therefore discuss two potential eco-logical factors that could contribute to the differ-ence observed in beetle species occurrence andabundance between forests and clearcuts.

Small–scale effects of fungal cluster

For all ciid species of the present study, fungal–cluster weight contributed positively to the prob-

ability of a sample being occupied, both in theforest and clearcut (see also Midtgaard et al.,1998). The higher frequency of occurrence of ciidscan be explained by the greater probability ofdetecting larger fungal clusters (actively or bychance only), as well as by a longer expectedpersistence time of the local beetle population inlarger clusters. There is evidence that both walk-ing and flying ciids are attracted to the volatilecompounds of their host fungus rather than find-ing the fungi accidentally (Jonsell & Nordlander,1995; Guevara et al., 2000a, 2000b). If largerfungal clusters emit greater amounts of volatilecompounds, then they would also attract more


specific ecological requirements. In larger clustersthere is also more potential for resource partitioning,which could facilitate coexistence by reducinginterspecific competition between ciid species (Guevaraet al., 2000a). Of the species recorded in this study,only S. affinis is known to inhabit —in great num-bers— fruiting bodies of fungal species other thanTrametes, namely Pycnoporus cinnabarinus (Jacq.:Fr.) P. Karsten (Økland, 1995). The fruiting bodies ofthis fungus occur exclusively in warm, open areas(Niemelä, 2001) and were also abundant in theclearcuts of this study. This supplementary host fun-gus could contribute to the very high frequency ofoccurrence of S. affinis in the clearcuts.

Larger fungal clusters inherently support larger ciidpopulations (Midtgaard et al., 1998), which in turncould affect local occurrence patterns via increasednumber of dispersing individuals. Evidence on thedispersal ability of ciid beetles is scarce and indirect,making it impossible to assess the dispersal rate

Fig. 4. The mean ± SE population size per fungal cluster occupied by a given ciid species in the studysites; black circles represent forests and open circles clearcuts.

Fig. 4. Tamaño medio ± EE de la población por masa fúngica ocupada por una especie dada de cíidoen las áreas de estudio; los círculos negros representan los bosques, y los vacíos los claros.

ciids, partly explaining the positive effect of fungalweight on the probability of occurrence.

In this study it is shown that "forest species" havea higher probability of occurrence in a fungal clusterof a given size in forests (optimal macrohabitat) vsclearcuts (suboptimal macrohabitat), and "clearcutspecies" show an opposite pattern. These resultssuggest that species may compensate for adverseenvironmental conditions in the suboptimalmacrohabitat by utilizing larger fungal clusters. As themeasure of fungal cluster weight used here incorpo-rates both the number and the weight of fruitingbodies, it is difficult to say much about the possiblemechanisms that cause the observed difference inspecies incidence. The observed pattern may resultfrom a higher colonization rate or from a longer ciidpopulation persistence time in larger fungal clusters.Besides, larger clusters probably have greater varia-tion in the quality of fruiting bodies, thus increasingthe likelihood that a given beetle species meets its

Sulcacis affinis Octotemnus glabriculus

Cis hispidus Cis boleti

2.0

1.5

1.0

0.5

0.0

2.0

1.5

1.0

0.5

0.0

1.4

1.2

1.0

0.8

0.6

0.4

0.80

0.75

0.70

0.65

0.60

0.55

0.50

0.45

0.40

1 2 3 4 1 2 3 4


Mea

n l

og

. p

op

ula

tio

n s

ize

(SE

)M

ean

lo

g.

po

pu

lati

on

siz

e (S

E)

146 Komonen & Kouki

between fungal clusters. Jonsell et al. (1999) demon-strated that some ciid species were absent fromfruiting bodies placed out in forest fragments 350–2,000 m from the natural forest. Similarly, Rukke(2000) found out that the incidence of ciid species infruiting bodies was negatively affected by increasedisolation at the scale of individual trees, 15 to over500 m depending on the species. Given the highdensity of Trametes fruiting bodies it seems unlikelythere would have been great difficulties for ciids inmoving between clusters in clearcuts. However, re-stricted movement between clusters in forests andbetween adjacent forests and clearcuts is possible.

Large scale effects of macrohabitat

The forest management category had an independ-ent effect on species frequency of occurrence,even less so than fungal–cluster weight. Similarly,the management category contributed positivelyto the abundance of S. affinis and C. hispidus inthe clearcuts and O. glabriculus in the forests. Inthis study, it is difficult to explicitly distinguishbetween the effect of fungal cluster density andmanagement category on the frequency of occur-rence of the ciid beetles in the clusters, as thedensity was consistently higher in the clearcuts.The apparent inconsistent response of the two"forest species" to the management category maybe due to opposing forces of microclimate andresource availability (Jonsell et al., 2001; Jonssonet al., 2001). However, a larger sample of forestand clearcut sites is needed to clarify whethersuch trade–off exists, or if the inconsistency re-sults from inadequate sampling.

Microclimatic conditions are one of the most im-portant abiotic differences between forests andclearcuts. Therefore, varying interspecific responsesof ciid species to forest management may resultfrom different microclimatic optima. Other studieshave documented that increased sun–exposure anddryness of fruiting bodies increase the probability ofoccurrence of some fungus–dwelling insect species(Midtgaard et al., 1998; Rukke & Midtgaard, 1998),whilst other species occur more frequently in moistfruiting bodies or in shady conditions (Økland, 1996;Jonsell et al., 2001; Thunes et al., 2000). Sverdrup–Thygeson & Ims (2002) showed that among thebeetle species in dead aspen there are clear prefer-ences concerning the degree of sun–exposure. Theirwindow–trap material also included O. glabriculusand, as in the present study, the species preferredshady conditions being more abundant in traps onshady aspen logs. Fossli & Andersen (1998) col-lected Trametes fruiting bodies from forests and,again, S. affinis was very rare. In Germany, S. affinisoccurs readily in clearcuts, whereas O. glabriculusand C. boleti manage well in more shady conditions(Reibnitz, 1999). Microclimate can affect the abun-dance of ciids directly by speeding up the individualdevelopment and indirectly by affecting the quality offruiting bodies. However, there are no studies linkingpopulation growth to the quality of fruiting bodies.

Conclusions

Many dead–wood dependent organisms can suc-cessfully occur in clearcuts if critical resources areleft in adequate densities and qualities (Jonsell etal., 2001; Jonsson et al., 2001; Martikainen, 2001;Sverdrup–Thygeson & Ims, 2002). Prior to the ex-tensive fire suppression in Finnish forests in the1900s, many of these species may have favoredopen areas created by forest fires. Nevertheless,sweeping generalizations about species responsesshould be avoided even for common species, asdemonstrated here. Despite higher Trametes den-sity, clearcuts are more ephemeral environmentsfor Trametes–dwelling insects in comparison withold–growth forests. The rationale is that most woodydebris becomes unsuitable for Trametes over thecourse of years (3–7 years; Hintikka, 1993), afterwhich it takes decades before new woody debris isavailable. Retaining green and dead deciduous treesin clear–cutting makes woody debris available forTrametes and many other more demanding speciesafter the logging residues have become unsuitable.

Acknowledgements

We thank the Forest and Park Service (Lieksa) forproviding data on the study areas. Tomas Roslinand Mats Jonsell kindly commented on the earlierversion of the manuscript. This study was fundedby the Academy of Finland, Centre of ExcellenceProgramme (# 64308).

References

Entwistle, H. M., 1955. The biology and morphol-ogy of the fungus beetles of the family Ciidaeand their parasites. MSc thesis, Univ. of London.

Esseen, P.–A., Ehnström, B., Ericson, L. & Sjöberg,K., 1997. Boreal forests. Ecol. Bull., 46: 16–47.

Fossli, T.–E. & Andersen, J., 1998. Host preferenceof Ciidae (Coleoptera) on tree–inhabiting fungi innorthern Norway. Entomol. Fennica, 9: 65–78.

Guevara, R., Hutcheson, K. A., Mee, A. C., Rayner,A. D. M. & Reynolds, S. E., 2000a. Resourcepartitioning of the host fungus Coriolus versicolorby two ciid beetles: the role of odour compoundsand host ageing. Oikos, 91: 184–194.

Guevara, R., Rayner, A. D. M. & Reynolds, S. E.,2000b. Orientation of specialist and generalistfungivorous ciid beetles to host and non–hostodours. Physiol. Entomol., 25: 288–295.

Haila, Y., 2002. A conceptual genealogy of frag-mentation research: from island biogeography tolandscape ecology. Ecol. Appl., 12: 321–334.

Hintikka, V., 1993. Occurrence of edible fungi andother macromycetes on tree stumps over a six-teen–year period. Acta Bot. Fenn., 149: 11–17.

Jonsell, M. & Nordlander, G., 1995. Field attractionof Coleoptera to odours of the wood–decayingpolypores Fomitopsis pinicola and Fomes


Mönkkönen, M. & Reunanen, P., 1999. On criticalthresholds in landscape connectivity: a manage-ment perspective. Oikos, 84: 302–305.

Niemelä, T., 2001. Guide to the polypores of Fin-land. 13th ed. Botanical Bulletins of the Univ. ofHelsinki 179.

Nilsson, T., 1997. Survival and habitat preferencesof adult Bolitophagus reticulatus. Ecol. Entomol.,22: 82–89.

Økland, B., 1995. Insect fauna compared betweensix polypore species in a southern Norwegianspruce forest. Fauna Norv. Ser. B, 42: 21–26.

– 1996. Unlogged forests: important sites for pre-serving the diversity of mycetophilids (Diptera:Sciaroidea). Biol. Conserv., 76: 297–310.

Reibnitz, J., 1999. Verbreitung und Lebensräumeder Baumschwammfresser Südwestdeutsch-lands (Coleoptera: Cisidae). Mitt. Ent. Stuttgart,34: 1–76.

Rukke, B. A., 2000. Effects of habitat fragmenta-tion: increased isolation and reduced habitat sizereduces the incidence of dead wood fungi bee-tles in a fragmented forest landscape. Ecography,23: 492–502.

Rukke, B. A. & Midtgaard, F., 1998. The importanceof scale and spatial variables for the fungivorousbeetle Bolitophagus reticulatus (Coleoptera,Tenebrionidae) in a fragmented forest landscape.Ecography, 21: 561–572.

Ryvarden, L. & Gilbertson, R. L., 1993. EuropeanPolypores 1. Abortiporus–Lindtneria. SynopsisFungorum 6. Fungiflora A/S, Oslo.

Schmiegelow, F. K. A. & Mönkkönen, M., 2002.Habitat loss and fragmentation in dynamic land-scapes: avian perspectives from the boreal for-est. Ecol. Appl., 12: 375–389.

Selonen, V. A. O., 2004. Polypores and associatedbeetles in a forest–clearcut ecotone. MSc thesis,Univ. of Jyväskylä.

Siitonen, J., 2001. Forest management, coarsewoody debris and saproxylic organisms:Fennoscandian boreal forests as an example.Ecol. Bull., 49: 11–41.

Sverdrup–Thygeson, A. & Ims, R. A., 2002. Theeffect of forest clearcutting in Norway on thecommunity of saproxylic beetles on aspen. Biol.Conserv., 106: 347–357.

Thunes, K. H. & Willassen, E., 1997. Species compo-sition of beetles (Coleoptera) in the bracket fungiPiptoporus betulinus and Fomes fomentarius(Aphyllophorales: Polyporaceae): an explorative ap-proach with canonical correspondence analysis. J.Nat. Hist., 31: 471–486.

Thunes, K. H., Midtgaard, F. & Gjerde, I., 2000.Diversity of coleoptera of the bracket fungusFomitopsis pinicola in a Norwegian spruce forest.Biodivers. Conserv., 9: 833–852.

Uotila, A., Kouki, J., Kontkanen, H. & Pulkkinen, P.,2002. Assessing the naturalness of boreal forestsin eastern Fennoscandia. For. Ecol. Manage.,161: 257–277.

fomentarius. Ann. Zool. Fenn., 32: 391–402.Jonsell, M., Nordlander, G. & Ehnström, B., 2001.

Substrate associations of insects breeding infruiting bodies of wood–decaying fungi. Ecol.Bull., 49: 173–194.

Jonsell, M., Nordlander, G. & Jonsson, M., 1999.Colonization patterns of insects breeding in wood–decaying fungi. J. Insect Conserv., 3: 145–161.

Jonsson, M., 2003. Colonization ability of thethreatened tenebrionid beetle Oplocephalahaemorrhoidalis and its common relativeBolitophagus reticulatus. Ecol. Entomol., 28:159–167.

Jonsson, M., Jonsell, M. & Nordlander, G., 2001.Priorities in conservation biology: a comparisonbetween two polypore–inhabiting beetles. Ecol.Bull., 49: 195–204.

Kaila, L., Martikainen, P. & Punttila, P., 1997. Deadtrees left in clear–cuts benefit saproxylicColeoptera adapted to natural disturbances inboreal forest. Biodivers. Conserv., 6: 1–18.

Klopfenstein, P. C., 1971. The ecology, behavior,and life cycle of the mycetophilous beetle,Hadreule blaisdelli (Casey) (Insecta: Coleoptera:Ciidae). Thesis, Bowling Green State Univ., Ohio.

Komonen, A., 2001. Structure of insect communi-ties inhabiting two old–growth forest specialistbracket fungi. Ecol. Entomol., 26: 63–75.

Komonen, A., Penttilä, R., Lindgren, M. & Hanski,I., 2000. Forest fragmentation truncates a foodchain based on an old–growth forest bracketfungus. Oikos, 90: 119–126.

Kouki, J., Löfman, S., Martikainen, P., Rouvinen, S.& Uotila, A., 2001. Forest fragmentation inFennoscandia: linking habitat requirements ofwood–associated threatened species to land-scape and habitat changes. Scand. J. ForestRes. Suppl., 3: 27–37.

Kouki, J., Arnold, K. & Martikainen, P., 2004. Long–term persistence of aspen –a key host for manythreatened species– is endangered in old–growthconservation areas in Finland. Journal for Na-ture Conservation, 12: 41–52.

Lawrence, J. F., 1973. Host preference in Ciidbeetles (Coleoptera: Ciidae) inhabiting the fruit-ing bodies of basidiomycetes in North America.Bull. Mus. Comp. Zool., 145: 163–212.

Löfman, S. & Kouki, J., 2003. Scale and dynamicsof transforming forest landscape. For. Ecol. Man-age., 175: 247–252.

Martikainen, P., 2001. Conservation of threatenedsaproxylic beetles: significance of retained as-pen Populus tremula on clearcut areas. Ecol.Bull., 49: 205–218.

McCullagh, P. & Nelder, J. A., 1989. Generalizedlinear models. 2nd ed. Chapman & Hall, New York.

Midtgaard, F., Rukke, B. A. & Sverdrup–Thygeson,A., 1998. Habitat use of the fungivorous beetleBolitophagus reticulatus (Coleoptera: Tene-brionidae): effects of basidiocarp size, humidityand competitors. Eur. J. Entomol., 95: 559–570.



Simón Benito, J. C., Espantaleón, D. & García–Barros, E., 2005. Stachorutes cabagnerensis n. sp.,Collembola (Neanuridae) from Central Spain, and a preliminary approach to phylogeny of genus. AnimalBiodiversity and Conservation, 28.2: 149–157.

AbstractStachorutes cabagnerensis n. sp., Collembola (Neanuridae) from Central Spain, and a preliminary approachto phylogeny of genus.— A new species of the genus Stachorutes, Stachorutes cabagnerensis n. sp., fromcentral Spain is described. It is characterized by the presence of 6+6 eyes in the head, retinaculum 2+2teeth, dentes with 5 hairs, and the absence of mucron. A phylogenetic analysis of this genus was attempted.Potential synapomorphies supporting the monophyly of Stachorutes are presented. One member of thegenus (the Nearctic S. navajellus) appears as a basal form, phylogenetically distant from the remaining (OldWorld) species. There is evidence for a monophyletic infrageneric clade with the species S. dematteisi, S.jizuensis and S. sphagnophilus. However, more information is required for further phylogenetic resolution.

Key words: Collembola, Stachorutes, Spain, Phylogeny.

ResumenStachorutes cabagnerensis sp. n., Collembola (Neanuridae) de la región central de España, y unaaproximación preliminar a la filogenia del género.— Se describe una nueva especie del genero Stachorutesde la region central de España. Stachorutes cabagnerensis nov. sp. se caracteriza por la presencia de 6+6ojos en la cabeza, retinaculum con 2+2 dientes y 5 sedas en cada rama del dentes; la furca carece demucrón. Se ha efectuado un análisis filogenético. Las sinapomorfias potenciales establecen la monofilia delgénero. Una especie del mismo, S. navajellus, aparece como forma basal, filogenéticamente distante delresto de especies (Viejo Mundo). Se podría establecer un clado infragenérico con las especies S.dematteisi, S. jizuensis y S. sphagnophilus. Sin embargo, se precisa de mayor información para poderconfirmarlo.

Palabras clave: Colembolos, Stachorutes, España, Filogenia.

(Received: 13 VIII 04; Conditional acceptance: 18 XI 04; Final acceptance: 13 I 05)

José Carlos Simón Benito, David Espantaleón & Enrique Garcia–Barros, Unidad de Zoología, Depto. deBiología, Fac. de Ciencias, Univ. Autónoma de Madrid, Cantoblanco 28049, Madrid, Spain.

Corresponding author: J. C. Simón Benito. E–mail: [email protected]

Stachorutes cabagnerensis n. sp.,Collembola (Neanuridae)from Central Spain,and a preliminary approach tophylogeny of genus

J. C. Simón Benito, D. Espantaleón &E. García–Barros


150 Simón Benito et al.

Introduction

The genus Stachorutes was established by Dallai(Dallai, 1973) from Italian specimens from the prov-ince of Canzo, according to the type speciesStachorutes dematteisi. Dallai characterized it forits small Antenna IV with simple sensillae, jaw withtwo teeth, simple maxillae, 2+2 eyes disposed as inthe genus Micranurida, postantennal organ present,claw without empodial appendage, tenent hairsabsent, reticulum and furca present, and withoutanal spines. Deharveng & Lienhard (1983) estab-lished two new species taken from the EasternFrench Pyrenees and the Swiss Alps, and re–de-fined the genus. A further re–definition was laterproposed by Jordana et al. (1997) in their mono-graph on Iberian Collembola. Thibaud & Palacios–Vargas (2000) added new characters, which re-sulted in an even sharper definition of the genus. Asummary of the morphological features of potentialuse for defining Stachorutes is given in table 1.

At present, Stachorutes Dallai consists of 16 spe-cies distributed all around the world. Eight speciesoccur in Europe: S. cabagnerensis n. sp.; S.dematteisi Dallai, 1973; S. longirostris Deharveng &Lienhard, 1983; S. ruseki Kovac,1999; S. scheraeDeharveng & Lienhard, 1983, S. sphagnophilusSlawaska, 1996; S. tatricus Smolis & Skarzynski2001 (Smolis & Skarzynski, 2001), S. valdeaibarensisArbea & Jordana, 1991. One species is known fromAfrica (S. dallai Weiner and Najt, 1998), three fromNorth America: S. escobarae (Palacios–Vargas,1990), S. maya Thibaud & Palacios–Vargas, 2000and S. navajellus Fjellberg, 1994 and four in Asia: S.ashrafi (Yosii, 1966), S. jizuensis Tamura and Zhao,1997; S. tieni Pomorski & Smolis, 1999 and S.triocelatus Pomorski & Smolis, 1999. The primarypurpose of the present study was to describe a newtaxon, S. cabagnerensis n. sp. However, since noattempt has currently been directed either at testingthe monophyly of the genus, or at resolving itsinternal relationships, a preliminary approach to thesequestions was attempted. This task was complicatedby the generally low degree of phylogenetic resolu-tion of the group. However, given the small numberof species, and the lack of information on the major-ity of them, any insight into the problem may help infacilitating taxa and character selection for furtherphylogenetic studies on pseudachorutine springtails.


The new species of Stachorutes

Soil samples were taken following the standardmethod designed for the research project"Bioasses". An area of 1 km2 was chosen in eachselected land unit and a sample was taken every200 m. A total of 16 units was prospected, eachconsisting of six spots representing different stagesin development of the vegetation. In Spain, theNational Park of Cabañeros was chosen and di-

vided into six different units. In each of them,samples were collected according to the methoddesigned. Individuals belonging to the genusStachorutes were obtained in wooded units only.These locations were: Unit 1. Natural wood withvegetation of holly oaks Quercus ilex ilex L., oaksQuercus pyrenaica Wild. and cork oaks Quercussuber L., with undergrowth of cistus Cistus ladaniferL., heather Erica australis L. and Erica arborea L.,arbutus Arbutus unedo L., located in the provinceof Ciudad Real in Navas de Estenas, UTM:30SVJ5754. Samples 106–H, fallen leaves fromholly oak, with heather and moss, nine specimens.Unit 2. Reforestation pinewood, Pinus pinasterAiton, also in the province of Ciudad Real, inHorcajo de los Montes, UTM: 30SVJ6973. Sample204–H2, fallen pine leaves after second year, threespecimens. Sample 207–H, fallen pine leaves, twospecimens. All samples were taken during themonths of April and May (2001), a period charac-terised by a seasonal maximum in the numbers ofboth individuals and species (Simón, unpubl. data).

Cladistic analysis

The data matrix was analysed using the programHenning 86 (Farris, 1988; options h*, mh* andbb*), treating all multistate characters as unordered.As an alternative exploratory option, the succes-sive weighting approach (Farris, 1969, 1989) wasattempted with the same programme. All otheranalyses, as well as character state optimisation,were completed through Winclada (Nixon, 2002)and Nona (Goloboff, 1993) (strict, majority con-sensuses, as well as bootstrap and jacknife testswith 100 replicates). Since the number of speciescurrently included in Stachorutes is low, an effortwas made to include all of them in the analyses.This resulted in two problems related to the char-acter state coding (see character list below): First,only partial information from S. ashrafi was avail-able, and as many as 5 of the 13 characters werecoded as unknown. Second, character 12 wasfound to be variable among individuals of S.longirostris from Pyrenean samples. Given theinability of the cladistic packages used in thisstudy to make a different treatment of unknownvs. non comparable character states, this specieswas entered as four different taxa (a,b,c,d). Webelieve that these decisions may be acceptablegiven the prospective nature of our approach.

Description

Stachorutes cabagnerensis n. sp.

Length of 0.66 mm in adult/s and 0.44 mm inyoung, dark blue in the adult/s, and light blue in thejuveniles. Dorsal setae reduced, subequal, withsensillae longer and thicker than in normal hairs.Integument with thickly grained surface.

Reduced mouth parts, jaw with 4 teeth, maxilla


styli form with two teeth,one ending in two apicalteeth and the other in a hook with 5–6 teeth alongthe apical area. Labium without hair L, its relationwith the length of the nail is 3 ( fig. 2).

Ocular spot with 6 eyes, 3 anterior (A, B, C) and3 posterior (E, F, G), the H and D are lacking (fig. 4).

Oval post antennal organ with 13 to 15 vesiclessimilar in size, disposed in one line, twice longerthan the nearest corneola and with approximatelythe same diameter (fig. 4).

Antenna quite thick, the relation with the headdiameter is 0.54 in adults and 0.58 in juveniles.Antenna I with 7 hairs, II with 12, III and IV are joineddorsally; segment III shows about 17 hairs and asensorial organ formed by two microsensilla (si) asa war club and long thick lateral sensilla: (sgd) and(sgv) and an extra ventral microsensilla (sa). An-tenna IV with normal hairs, straight, some small,without sensorial hair–brush in the ventral area.With 6 olfactory sensilla, 4 in the dorsal area (S1 toS4) forming a rhomb, and 2 in the ventro–apicalarea (S7, S8). Furthermore, there is a dorsal exter-nal microsensilla (m) and a very small distal organ.An apical tri–lobed vesicle is located at the apex ofthe antenna (fig. 5).

Reduced dorsal chaetotaxy (fig. 1), the positionof the most internal sensilla S is: 3, 3/4, 4, 4, 4, 2hairs, from thorax II to abdomen V. The formula ofthe dorsal inner hairs is 1, 3, 3/3, 3, 3, 2. Headwithout hair a0, the hairs d0 odd. Pronotum with 3+3hairs. Mesonotum with a2. Tibiotarsi I, II, III with 19,19, 18 hairs disposed in two whorls, the apical with11 hairs and the basal with 8 hairs, except in thethird pair of legs which shows 7 hairs, withouttenent hairs (fig. 3). Ventral tube with 4+4 hairs,2+2 basal and 2+2 apical.

Claw without teeth and empodial appendage.Tenaculum with 2+2 teeth. Dens without mucron,

and five hairs, manubrium with 7–9+7–9 hairs(fig.6).

Genital orifice in the male with 18 hairs, in thefemale with 8 hairs in the anterior margin of thegenital orifice, plus two central ones.

DiscussionThe new species shows 6+6 eyes like two otherspecies of this genus, ruseki Kovac, 1999 fromSlovakia and ashrafi (Yosii, 1966) from Nepal.This latter species may however not belong tothis genus because of the number of olfactory

Table 1. Anatomical features of potential use in the definition of the genus Stachorutes, as statedby different authors (+ feature present; – feature absent): A. Dallai, 1973; B. Deharveng & Lienhard,1983; C. Jordana et al., 1997; D. Thibaud & Palacios–Vargas, 2000. Note that some of thesecharacters display no variation within Stachorutes, and were not coded for phylogenetic analyses.

Tabla 1. Carácteres anatómicos de uso potencial en la definición del género Stachorutes, segúndistintos autores (+ caracter presente; – caracter ausente): A. Dallai, 1973; B. Deharveng & Lienhard,1983; C. Jordana et al., 1997; D. Thibaud & Palacios–Vargas, 2000. Nótese que algunos de estoscaracteres no presentan variación alguna dentro del género Stachourutes, y no fueron utilizados paralos análisis filogenéticos.

Anatomical features A B C D

Antenna IV with 5–6 sensilla cylindrical and one microsensilla + + +

Antenna IV with sensilla in flame–shape and one microsensilla +

Antenna IV without hair–brush + +

Antenna IV and III join + +

Postantennal organ with vesicles simple never moruliform +

Less than 8+8 eyes + + + +

Maxilla styliform + + + +

Empodial appendage – – – –

Tenent hairs – – – –

Furca reduced + + + +

Tenaculum + + + +

Mucron – +/– + +/–

Chaetotaxy dorsal reduced + – +

Macrochaetae – –

Anal spines – – –


hairs. Stachorutes cabagnerensis differs from bothin the number of teeth of the tenaculum (2 com-pared to 3), and in the structure of the dentes (5hairs against 6), and in the lack of mucron.

Phylogenetic analysis

A matrix of 13 characters (8 binary and 5 multistate)was prepared (table 2). The state 0 corresponds tothe plesiomorphic character state. Rooting wasdone with reference to Pseudachorutes parvulusBörner, 1901.

Results

Henning analysis produced 626 trees with a lengthof 50 steps (CI = 50, RI = 63). All characters except2, 6 and 8 showed a certain degree of homoplasy,(table 3).

The strict consensus of these cladograms isshown in fig. 7 (tree length, 76; CI, 32; RI, 25).On the assumption that Stachorutes represents amonophyletic assemblage, its monophyly can besupported by the synapomorphies 0:1 and 4:2(that is, state 1 of character 0, and state 2 of

Figs. 1–6. 1. Dorsal chaetotaxy; 2. Labium; 3.Tibiotarsus; 4. Eyes and postantennal organ; 5. IV andIII antennal segment; 6. Furca.

Figs. 1–6. 1. Quetotaxia dorsal; 2. Labio; 3. Tibiotarso; 4. Ojos y órgano postantenal; 5. Segmentoantenal III y IV; 6. Furca.

1 2 3

4

5

6

d0

P3

p3

P4

P4

P2


Table 2. Matrix of species x character state. Pseudachorutes parvulus was used as outgroup: 0. Jaw(number of teeth: 0 = 4 teeth; 1 = 3 teeth; and 2 = 2 teeth); 1. Olfactory hairs of the Antenna IV (0= tubular; 1 = in the shape of a sparkling flame); 2. PAO (shape of the postantenna organ: 0 = circular;1 = elliptic); 3. Number of vesicles in the postantennal organ (0 = less than 10; 1 = 11 or more); 4.Number of eyes (0 = 8 eyes; 1 = 6 eyes; 2 = 5 eyes; 3 = 3 eyes; 4 = 2 eyes; 5 = 1 eye; 6 = 0 eyes);5. Hair a0 in the head (0 = with this hair; 1 = without it); 6. Hair d0 in the head (0 = odd or withoutit; 1 = pair); 7. Number of hairs in the pronotum (0 = 3+3 hairs; 1 = 2+2 hairs; 2 = 4+4 hairs); 8. Haira2 in the mesonotum (0 = with hair; 1 = without hair); 9. Number of hairs in the tibiotarsum I, II, III(0 = 19, 19, 18; 1 = another condition); 10. Number of teeth in the retinaculum (0 = 3+3 teeth; 1 =2+2 teeth); 11. Number of hairs in the dentes (0 = 6 hairs; 1 = 5 hairs; 2 = 4 hairs; 3 = 3 hairs; 4 =2 hairs; 5 = 1 hairs); 12. Mucron (0 = mucron separated from the dentes; 1 = mucrodens; 2 = absent).? Indicates that the character state is unknown.

Tabla 2. Matriz de especies x estados del carácter estudiado. Se utilizó a Pseudachorutes parvuluscomo grupo externo: 0. Mandíbula (número de dientes: 0 = 4 dientes; 1 = 3 dientes; and 2 = 2 dientes);1. Sedas olfactorias de la antena IV (0 = tubular; 1 = en forma de llama de bujía); 2. PAO (forma delórgano postantenal: 0 = circular; 1 = elíptica); 3. Numero de vesículas del órgano postantenal (0 = menosde 10; 1 = 11 o más); 4. Número de ojos (0 = 8 ojos; 1 = 6 ojos; 2 = 5 ojos; 3 = 3 ojos; 4 = 2 ojos; 5= 1 ojo; 6 = 0 ojos); 5. Seda a0 de la cabeza (0 = con esta seda; 1 = sin ella); 6. Seda d0 de la cabeza(0 = impar o sin seda; 1 = par); 7. Número de sedas del pronoto (0 = 3+3 sedas; 1 = 2+2 sedas; 2 =4+4 sedas); 8. Seda a2 del mesonoto (0 = con seda; 1 = sin seda); 9. Número de sedas del tibiotarsoI, II, III (0 = 19, 19, 18; 1 = otra condición); 10. Número de dientes del retináculo (0 = 3+3 dientes; 1 =2+2 dientes); 11. Número de sedas del dentes (0 = 6 sedas; 1 = 5 sedas; 2 = 4 sedas; 3 = 3 sedas; 4= 2 sedas; 5 = 1 sedas); 12. Mucrón (0 = mucrón separado del dentes; 1 = mucrodens; 2 = ausente).? indican que se desconoce el estado del carácter citado.

Species 0 1 2 3 4 5 6 7 8 9 10 11 12 Distribution

Pseudachorutes parvulus 0 0 0 0 0 0 0 0 0 0 0 0 0 Cosmopolitan

Stachortes ashrafi 1 0 1 1 1 ? ? ? ? ? 0 0 1 Nepal

S. cabagnerensis n. sp. 0 0 1 1 1 1 0 0 0 0 1 1 2 Spain

S. dallai 2 0 0 0 4 1 1 0 0 0 1 1 1 Tanzania

S. dematteisi 2 0 0 0 4 1 0 0 0 ? 1 4 2 Italy

S. escobarae 2 0 1 0 6 1 0 0 0 0 1 3 1 Mexico

S. jizuensis 2 1 0 0 5 1 0 0 1 1 1 5 2 China

S. longirostris (a) 1 0 1 0 2 1 0 1 1 0 0 3 1 France (Pyrenees)

S. longirostris (b) 1 0 1 0 2 1 0 1 1 0 0 2 1 France (Pyrenees)

S. longirostris (c) 1 0 1 0 2 1 0 1 1 0 0 1 1 France (Pyrenees)

S. longirostris (d) 1 0 1 0 2 1 0 1 1 0 0 0 1 France (Pyrenees)

S. maya 0 0 1 1 4 1 1 2 0 1 0 2 1 Mexico

S. navajellus 1 0 0 0 2 0 0 0 0 ? 1 0 0 USA and Canada

S. ruseki 1 0 1 0 1 1 0 0 1 0 0 0 0 Slovakia

S. scherae 1 0 1 0 2 1 0 0 0 0 0 0 1 Switzerland

S. sphagnophilus 0 1 0 0 4 1 0 0 1 1 0 4 2 Poland

S. tatricusa 2 0 1 0 4 1 0 0 1 0 0 0 1 Poland

S. tatricusb 2 0 1 1 5 1 0 0 1 0 0 0 1 Poland

S. tieni 2 0 0 0 2 1 1 1 1 1 0 2 1 Vietnam

S. triocellatus 1 0 0 0 3 1 1 1 1 1 1 3 1 Vietnam

S. valdeaibarensis ? 0 0 0 2 1 0 0 0 0 0 1 1 Spain

character 4). It is true, however, that the defini-tion of the consensus tree was low, and that themonophyly of this and other pseudachorutine gen-

era awaits further and more thorough reassess-ment based on a taxonomically wider sample. Allthe species except S. navajellus seem to belong to


one clade supported by the synapomorphies 2:1,5:1, 8:1 y 12:1, while navajellus shows theplesiomorphic state for most of these characters.Two clades stand out from the main group: theone formed by (S. dematteisi + (S. jizuensis + S.sphagnophilus)), characterized by 11:4, and (S.triocellatus + S. tieni), which can only be definedthrough homoplasies.

Bootstrap and jacknife analyses rendered thesame results (not presented in detail). Due to thelow resolving power of the data, speculations maybe ventured on the basis of three alternative proce-dures: setting all multistate characters as ordered(with analysis proceeding as beforehand), succes-sive weighting (multistate characters unordered),and majority consensus based on the set of

Figs. 7–10. 7. Unordered multistate characters, strict consensus (L = 76, Cl = 32, RI = 25, derived from626 trees with L = 50, Cl = 50, RI = 63). Each circle represents one change in one character (filled =homoplay–free apomorphies, empty = convergences or reversals). The number above each circle isthe number of characters, the one below it is the state of that character at that node. 8. Strictconsensus tree, multistate characters ordered; consensus (L = 88, Cl = 28, RI = 33, derived from 235trees with L = 60, Cl = 41, RI = 63). 9. Strict consensus, based on rescaled consistency index (L = 72,Cl = 34, RI = 30, derived from 11 trees with L = 177, Cl = 63, RI = 74). 10. Majority rule consensus,based on more than 600 trees.

Figs. 7–10. 7. Caracteres multiestados no ordenados, consenso estricto (L = 76, CI = 32, RI = 25,derivados de 626 árboles con L = 50, CI = 50, RI = 63). Cada círculo representa un cambio en uncarácter (lleno = apomorfias sin homoplasia, vacíos = convergencias o inversiones). El número que sehalla sobre cada círculo es el número del carácter, y el de debajo el estado de dicho carácter en elnodo. 8. Árbol de consenso estricto, caracteres multiestados ordenados. Consenso (L = 88, CI = 28,RI = 33, derivados de 235 árboles con L = 60, CI = 41, RI = 63). 9. Consenso estricto, basado en uníndice de consistencia re–escalado (L = 72, CI = 34, RI = 30, derivados de 11 árboles con L = 177,CI = 63, RI = 74). 10. Consenso de la mayoría, basado en más de 600 árboles.

Table 3. Performances of the characters 1–13 in the initial parsimony analysis (above; charactersunordered, without weight), and in the strict consensus derived from these trees (below). The figuresgiven are the number of steps in the tree, consistency index (CI) and retention index (RI).

Tabla 3. Comportamiento de los caracteres 1–13 en el análisis de parsimonia inicial (arriba; caracteresno ordenados, sin peso), y en el consenso estricto derivado de estos árboles (abajo). Las cifras sonel número de escalones en el árbol, el índice de consistencia (CI) y el índice de retención (RI).

Characters

1 2 3 4 5 6 7 8 9 10 11 12 13

Best fits (all trees)

Steps 6 11 3 2 8 1 2 2 2 2 4 8 3

CI 33 100 33 50 75 100 50 100 50 50 25 62 66

RI 55 100 75 66 71 100 66 100 87 75 50 62 80

Worst fits (all trees)

Steps 8 1 4 4 9 1 3 2 4 3 5 9 4

CI 25 100 25 25 66 100 33 100 25 33 20 55 50

RI 33 100 62 0 57 100 33 100 62 50 33 50 60

Number of steps, CI and RI in the consensus tree

Steps 10 1 5 4 12 1 3 6 8 3 7 12 4

CI 20 100 20 25 50 100 33 33 12 33 14 41 50

RI 11 100 50 0 14 100 33 20 12 50 0 12 60


pseudac

navajel

escobar

longia

maya

longib

dallai

longic

valdeai

cabagno

tatrica

tatricb

longici

scherae

ashrafi

ruseki

triocel

tieni

dematte

jizuen

sphagno

10

10 4 8 10 11

2 6 0 1 37 11

1 30 3 4 6 7 8 9 11

0 1 4 1 2 0 1 27 11

1 20 2 4 6 8 10 11

2 0 4 1 0 1 17 11

1 12 8 11

0 0 10 3 4 8 10 11 12

0 1 1 0 1 1 20 4

2 10 3 4

2 1 57

18

03 4

1 14 12

1 0

2 6 7 9

0 1 1 1

0 2 4 11 12

2 0 4 4 2

4 10 11

3 1 30 11

2 2 8 10

0 1 4 10 11

5 1 50

1 9

1 1

2 6 8 12

1 1 1 1

0 4

1 2

pseudac

navajel

escobar

longia

maya

longib

dallai

longic

valdeai

cabagno

longid

scherae

triocel

tieni

dematte

jizuens

sphagn

tatrica

ruseki

tatricb

ashrafi

10

1

0 4

1 2

2 5 11 12

1 1 1 1

0 4 10 11

2 6 1 37 6 11

1 1 30 3 4 6 7 9 11

0 1 4 1 2 1 27 9 11

1 1 20 2 4 6 10

2 0 4 1 17 8

1 12

00 3 4 10 12

0 1 1 1 27 8 11

1 1 011

0

2 5 7 8 9 11

0 1 1 1 1 2

0 2 4 9 10 1112

2 0 4 1 1 4 2

0

2

4 8 11

1 1 0

4 10 11

3 1 30

2

1 8

1 1

4 11

5 50 10

0 0

12

0

8

1

0 4

2 5

pseudac

navajel

valdeai

scherae

tatrica

tatricb

ashrafi

ruseki

escobar

dallai

cabagno

dematte

jizuens

sphagn

longia

longib

longic

longid

maya

triocel

tieni

100

84

84

84

100

100

100

10060

7

pseudac

navajel

triocel

longia

maya

longib

tieni

longic

valdeai

cabagno

tatrica

tatricb

longid

scherae

ashrafi

ruseki

dallai

dematte

escobar

jizuens

sphagn

2 6 8 11 12

1 1 1 1 1

0 4

1 2

10

12 4 6 7 8 10 11

0 3 1 1 1 1 37 11

1 30 3 4 6 7 8 9 11

0 1 4 1 2 0 1 27 11

1 20 2 6 7 9 11

2 0 1 1 1 27

12 8

0 00 3 4 8 10 12

0 1 1 0 1 20 4 11

2 1 00 3 4 11

2 1 5 07 11

1 08 11

0 03 4 11

1 1 04 11 12

1 0 0

0 4 8 10

2 4 0 1

2 6

0 1 4

6

2 11 12

0 4 2

11

3

1 8 9

1 1 1

4 11

5 50 10

0 0

8

9 10


cladograms obtained in the first trial. Treating themultistate characters as ordered resulted in 235cladograms (L = 60, CI = 41, RI = 63); the consen-sus tree L = 88, CI = 28, RI = 33 is shown in fig. 8.In contrast with the former results, S. dallai and S.escobarae were associated to the clade formed by(S. dematteisi + (S. jizuensis + S. sphagnophillus)),although no synapomorphy free of homoplasy wasfound to define this group.

Further possible tree structures were suggestedby the results of successive weighting (rescaled CI)and majority consensus. None of these trees rep-resent a maximum parsimony solution. However,they are interesting to the extent that either theaddition of new characters, or the recoding of char-acters where some states are non–applicable orunknown, might not support the present most par-simonious solution. These results suggested a po-tential link between S. escobarae, S. dallai and S.cabagnensis, and the clade formed by (S. dematteisi+ (S. jizuensis + S. sphagnophilus)) (fig. 7), as wellas a degree of relatedness between S. longirostrisand S. maya with the clade (S. triocellatus + S.tieni) (fig. 9). Finally, although not strongly sup-ported and based on homoplastic features, somerelationship between the species (S. tatricus + S.ashrafi + S. ruseki) could be determined (fig. 10).

Conclusions

Although the overall resolution of the cladograms islow, a few points can be highlighted: (1) from aparsimonious point of view, the anatomical infor-mation currently available does not permit a de-tailed, highly resolved, phylogenetic hypothesis. Thiscould only be solved by adding new characters; (2)some partial conclusions may be of interest forfurther studies:

1. The monophyly of the group can be provision-ally supported on the basis of two synapomorphies(fig. 7, characters 0 and 4: jaw with 3 teeth, and fiveocelli present). This interpretation requires furtherassessment as there is no out–group taxon topolarise character state changes at the basal node(i.e., external to Stachorutes + Pseudachorutes).Characters 2, 5, 8 and 12 support the monophyly ofan infrageneric clade including all the other mem-bers considered in this study except navajellus.

2. The position of navajellus is peculiar for itapparently belongs to an isolated basal group or,itmay have lost some features of the rest of thegroup due to reversal in characters 2, 5 , 8 and 12.

3. Within the ingroup, the resolution of the con-sensus is low. Only one clade with three species(dematteisi, jizuensis y sphagnophilus) can be de-fined with some accuracy, and even so, on thebasis of one single nonhomoplasic synapomorphy(character 11, state 4). There is some evidence forthe existence of two or three additional clades, butthis is supported by homoplastic features only.

The combination of the inferred phylogeneticinformation (e.g., fig. 10) and the known geographic

distributions of the species dealt with here results instrikingly broad geographic ranges at the supra–specific clades, in contrast with rather local specificdistributions. Thus, for instance, the clade compris-ing S. tatricus, S. ashrafi and S. ruseki could beclassified as of wide Palaearctic distribution (orEurasian, e.g. Cox, 2001; Morrone, 2002), rangingfrom Western Europe to the Himalayas (Yosii, 1966;Smolis & Skarzynski, 2001). The clade comprisingS. longirostris and S. tieni includes species of theWestern Palaearctic, Nearctic, and Oriental regions.The three species represented in the best sup-ported clade (S. dematteisi + S. jizuensis + S.sphagnophilus) were described from Central Eu-rope, Italy, and the Yunnan region in South–West-ern China. The closest relatives of these three taxainclude one East African and one Nearctic member(Dallai, 1973; Palacios–Vargas, 1990), together withthe new species S. cabagnerensis for Spain. Ahighly conservative interpretation of such patternsis recommended. Moreover, the authors’ feeling isthat either (a) very little insight on the phylogeny ofthe group has actually been gained, (b) the distribu-tions of these springtail species are still very poorlyknown, or (c) an important number of relatedcollembolan species in each of the main geographicareas mentioned have not yet been described. It isquite likely that all three hypotheses are equallypertinent.

Acknowledgements

This work was conducted with financial supportfrom the European Union for the Research Project"Biodiversity assessment tools" (Grant Nº 0028,2000).

References

Cox, C. B., 2001. The biogeographic regions recon-sidered. Journal of Biogeography, 28: 511–523.

Dallai, R., 1973. Ricerche sui Collemboli. XVI.Stachorutes dematteisi n. gen., s. sp., Micranuridaintermedia n. sp. e considerazioni sul genereMicranurida. Redia, 54: 3–31.

Deharveng, L. & Lienhard, C., 1983. Deux nouvellesespèces du genre Stachorutes Dallai, 1973(Collembola). Revue suisse de Zoologie, 90:929–934.

Farris, J. S., 1969. A successive approximationsapproach to character weighing. Systematic Zo-ology, 18: 374–385.

– 1988. Henning version 1.5, Reference guide.Published by the autor. Admiral Street, PortJefferson Station, New York.

– 1989. The retention index and the rescaled con-sistency index. Cladistics, 5: 417–419.

Fjellberg, A., 1984. Collembola from the ColoradoFront Range U.S.A. Arctic and Alpine Research,16: 193–209.

Jordana, R., Arbea, J. I., Simón, C. & Luciáñez, M.


J., 1997. Collembola, Poduromorpha. In: FaunaIbérica, V. 8: Museo Nacional de Ciencias Natu-rales, Madrid.

Goloboff, P. A., 1993. Nona, version 2.0 for Win-dows. Inst. Miguel Lillo. Miguel Lillo 205, 4000S. M. Tucumán, Argentina.

Kovac, L., 1999. Stachorutes ruseki sp. n.(Collembola, Neanuridae) from Slovakia. Biologia,Bratislava, 54: 35–138.

Morrone, J. J., 2002. Biogeographic regions undertrack and cladistic scrutiny. Journal of Biogeog-raphy, 29: 149–152.

Nixon, K. C., 2002. WinClada version 1.00.08. Pub-lished by the autor. Ithaca, New York.

Palacios–Vargas, J. G., 1990. Nuevos Collemboladel estado de Chichuahua, México. FoliaEntomológica Méxicana, 79: 5–32.

Pomorski, R. J. & Smolis, A., 2000. Two newspecies of Stachorutes Dallai, 1973 from NorthVietnam (Collembola, Neanuridae). Annaleszoologici, 49: 151–156.

Smolis, A. & Skarzynski, D., 2001. A new species ofthe genus Stachorutes Dallai, 1973 from Poland(Collembola: Neanuridae). Genus, 12: 407–410.

Slawska, M., 1996. Stachorutes sphagnophilousn. sp. from Northern Poland (Collembola:Neanuridae). Genus, 7: 325–329.

Tamura, H. & Zhao, L., 1997. Two new species ofthe family Pseudachorutidae from Mt. Jizu,western Yunnan, southwest China (Insecta:Collembola). Natural History Bulletin Ibaraki Uni-versity, 1: 45–50.

Thibaud, J. M. & Palacios–Vargas, J. G., 2000.Remarks on Stachorutes (Collembola: Pseuda–chorutidae) with a new Mexican species. FoliaEntomologica Mexicana, 109: 107–112.

Weiner, W. M. & Najt, J., 1998. Collembola(Entognatha) from East Africa. European Jour-nal Entomology, 95: 217–237.

Yosii, R., 1966. Collemboles of Himalaya. Journalof the College of Arts and Sciences, Chiba Univ.,4: 461–531.



Martin, C. S., Jeffers, J. & Godley, B. J., 2005. The status of marine turtles in Montserrat (EasternCaribbean). Animal Biodiversity and Conservation, 28.2: 159–168.

AbstractThe status of marine turtles in Montserrat (Eastern Caribbean).— The status of marine turtles in Montserrat(Eastern Caribbean) is reviewed following five years of monitoring (1999–2003). The mean number of nestsrecorded during the annual nesting season (June–October) was 53 (± 24.9 SD; range: 13–43). Inaccordance with earlier reports, the nesting of hawksbill (Eretmochelys imbricata) and green (Cheloniamydas) turtles was confirmed on several beaches around the island. Only non–nesting emergences weredocumented for loggerhead turtles (Caretta caretta) and there was no evidence of nesting by leatherbackturtles (Dermochelys coriacea); however, it is possible that additional survey effort would reveal low densitynesting by these species. Officially reported turtle capture data for 1993–2003 suggest that a mean of 0.9turtle per year (± 1.2 SD; range: 0–4) were landed island–wide, with all harvest having occurred during theannual open season (1 October to 31 May). Informed observers believe that the harvest is significantlyunder–reported and that fishermen avoid declaring their catch by butchering turtles at sea (both during andoutside the open season). Of concern is the fact that breeding adults are potentially included in the harvest,and that the open season partially coincides with the breeding season. The present study has shown thatalthough Montserrat is not a major nesting site for sea turtles, it remains important on a regional basis forthe Eastern Caribbean.

Key words: Caribbean, Eretmochelys imbricata, Hawksbill sea turtle, Chelonia mydas, Green sea turtle,Conservation.

ResumenEstatus de las tortugas marinas en Montserrat (Caribe oriental).— Se ha estudiado la situación de lastortugas marinas en Montserrat (Caribe oriental) mediante un seguimiento de cinco años (1999–2003). Elnúmero medio de nidos registrados durante la estación anual de nidificación (junio–octubre) fue de 53(± 24.9 SD; rango: 13–143). En concordancia con informes anteriores, se confirmó la nidificacón de lastortugas carey (Eretmochelys imbricata) y verde (Chelonia mydas) en varias playas alrededor de la isla.En la tortuga boba (Caretta caretta) sólo se registraron salidas sin nidificación, y no se encontraronpruebas de que la tortuga laúd (Dermochelys coriacea) nidificase; sin embargo, es posible que ulterioresestudios pongan de manifiesto una baja densidad de nidificación de esta especie. Los datos oficiales decapturas de tortugas (1993–2003) sugieren que en toda la isla llegaban a tierra una media de 0.9 tortugasanuales (± 1.2 SD; rango: 0–4), produciéndose todas las capturas cuando se había levantado la veda.Observadores bien documentados creen que las cifras de recolección están significativamente falseadasa la baja, y que los pescadores evitan declarar sus capturas sacrificando las tortugas en el mar (con laveda abierta o cerrada). Es preocupante que en esta caza puedan incluirse tortugas que crían, y que elperíodo de captura permitida coincide en parte con la estación reproductora. Este estudio demuestra queaunque Montserrat no es un lugar principal de nidificación de las tortugas marinas, sigue siendoimportante a escala regional en el Caribe oriental.

Palabras clave: Caribe, Eretmochelys imbricata, Tortuga carey, Chelonia mydas, Tortuga verde,Conservación.

The status of marine turtles inMontserrat (Eastern Caribbean)

C. S. Martin, J. Jeffers & B. J. Godley

160 Martin et al.

(Received: 14 X 04; Conditional acceptance: 14 II 05; Final acceptance: 25 IV 05)

Corinne S. Martin, Dept. of Geographical and Life Sciences, Canterbury Christ Church Univ., CanterburyCT1 1QU, U.K.– John Jeffers, Dept. of Fisheries, Ministry of Agriculture, Government of Montserrat,Brades, Montserrat, West Indies.– Brendan J. Godley, Marine Turtle Research Group, Centre for Ecologyand Conservation, Univ. of Exeter in Cornwall, Tremough Campus, Penryn TR10 9EZ, U.K.

Corresponding author B. J. Godley. E–mail: [email protected]

mailto: [email protected]


sand dominates at Rendez–vous beach, the north-ern most beach on the island’s western side (Anony-mous, 1993). The volcanic origin of the island wasdramatically exposed in 1995 when the SoufrièreHills’ volcano located in the southern part of theisland began exhibiting signs of volcanic activity.Since then, there has been an ongoing volcaniccrisis, with an evacuation of the southern part of theisland (including the Capital, Plymouth), a safety"exclusion zone" that covers almost two thirds of theisland (fig. 1; Gell & Watson, 2000) and widespreadhuman emigration and economic disruption.

Nesting populations

Day–time monitoring of marine turtle nesting

The Fisheries Department of Montserrat’s Ministryof Agriculture has been coordinating the monitoringof beaches for turtle activities (including nesting,hatchling emergences, and nest excavations) since1999. Although ad–hoc day–time beach monitoringhas been carried out by dedicated island residentswho regularly check local beaches for turtleemergences and nests, the bulk of the monitoringeffort has been carried out by the Fisheries Depart-ment (J. J.). Monitoring frequency of nesting beacheshas been uneven, being especially patchy (i.e. afew times a year) on the beaches located in theexclusion zone (fig. 1). Safe, accessible beacheswere walked and checked for turtle tracks and nestson a fairly regular basis (i.e. up to twice a week atthe peak of the nesting season).

Beach monitoring datasheets were completed(by J. J.) each time a beach was visited, even if nonesting activity had taken place; other island resi-dents did so only when they detected nesting activ-ity. As a result, the number of beach monitoringsheets filled during a given period of time was onlyloosely indicative of the monitoring effort. Nests(N), i.e. adult emergences thought to have resultedin the deposition of a clutch of eggs, were individu-ally counted. Non–nesting emergences (NNE) werenot counted individually but, instead, their presenceor absence on any given survey day was recorded.No distinction was made among species based ontrack morphology, as in many cases the nature ofthe beach, the type of substratum, the age of thetracks, and the relative inexperience of some therecorders precluded reliable species identification.

All island beaches were monitored a minimum ofonce a week for one month (mid–August / mid–September) in 2003. Although one month of com-prehensive survey was insufficient to accuratelyassess, on an annual basis, the extent of the spatialbias caused by uneven monitoring effort, it wasthought sufficient for detecting any major underes-timation of nesting activity for beaches relativelyless monitored during the five year dataset (1999–2003). Due to the relatively low nesting activity,monitoring beaches a minimum of once a weekwas sufficient to detect all activities occurring dur-ing the preceding week. More frequent monitoring

Introduction

Four species of sea turtles have been reported asnesting in Montserrat (Eastern Caribbean). Earlystudies suggested that the green (Chelonia mydas)and hawksbill (Eretmochelys imbricata) turtlesnested in small numbers, whilst loggerhead(Caretta caretta) and leatherback (Dermochelyscoriacea) turtle nests were only occasionally en-countered (Meylan, 1983; John, 1984; Groombridge& Luxmore, 1989). A recent review of hawksbillturtle nesting in the Caribbean region (Meylan,1999) reported that nesting in Montserrat is "inci-dental" although this result was based on recon-naissance of beaches and interviews cited inMeylan (1983). Meylan (1983) concluded that nest-ing levels were low, presumably because of con-stant human activity on the island’s beaches (whichwere widely used for boat storage and recreationalpurposes).

Both adult and juvenile hawksbill and greenturtles are found in Montserrat’s inshore waters(Meylan, 1983; John, 1984). Montserrat’s TurtleOrdinance (1951) states that turtles can be cap-tured, sold and bought during an annual openseason (1 October to 31 May). Although there areno quota or species restrictions, harvested turtlesmust weigh at least 20 lbs (ca 9.1 kg), and thereare no restriction on the maximum size of har-vested turtles. For several years now, the island’sfisheries authorities have been attempting to raiseawareness about biodiversity conservation and tur-tle stock management issues among the island’slocal fishermen. During these conversations, localfishermen are often verbally encouraged by thefisheries authorities to report any sea turtle catch tothem (along their fish catches). It is not known,however, what proportion of fishermen actually re-port their turtle catches to the authorities.

We present a five–year marine turtle monitoringdataset gathered with limited resources to elucidatespatial and temporal patterns of marine turtle nest-ing in Montserrat. The first estimates of sea turtlenest numbers for Montserrat are provided. In addi-tion, available turtle capture data are presented,offering preliminary insights into the local marineturtle fishery.


Study site

The Caribbean Island of Montserrat (62° 12’ W,16° 45’ N) is part of the Leeward Islands of theLesser Antilles. It is 104 km² in area and situatedapproximately 35 km southwest of Antigua and 60 kmnorthwest of Guadeloupe (fig. 1; Blankenship, 1990).Apart from Trant’s and Farm beaches (east coast),all of Montserrat’s sandy beaches are located on thewestern side of the island (fig. 1). The island is ofvolcanic origin and all but one of its sandy beachesconsist of black volcanic sand; white calcareous

162 Martin et al.

(i.e. more than once a week) facilitated speciesidentification based on track morphology (followingPritchard & Mortimer, 1999). Beaches were eitherwalked or checked from a distance with binoculars(e.g. from a helicopter/boat). Special permissionwas granted from the authorities to access andwalk some of the beaches of the exclusion zone (atthe time Trant’s, Farm, Fox’s, Bransby Point, HotWater Pond). In these surveys, individual non–nesting emergences and nests were counted.

Night–time monitoring of marine turtle nesting

In 2002 and 2003, logistics permitting, beacheswere monitored at night for the presence of nestingturtles. When possible, nesting turtles were meas-ured (Curved Carapace Length, CCL) and taggedsubcutaneously with Passive Integrated Transponder(PIT) tags.

Fishery harvest data

Records of turtle harvests were obtained fromMontserrat’s Fisheries Department in the form of alist detailing the month and year (1993–2003) ofcapture, the turtle species (if known), and the weightof the animal (in lb, if measured). The list had beencompiled, over the years, by officers working at theisland’s main harbours (Plymouth then Carr’s Bay).No other information is available, hence it is notknown what percentage of the turtle catch theserepresent or if certain forms of fishing are over orunder represented.

Results

Nesting populations

Day–time monitoring of marine turtle nesting

For the five year dataset (1999–2003), data origi-nating from a total of 453 beach monitoring formswere analysed. The mean annual number of nestswas 53 (± 24.9 SD, range: 13–143). Records ofnon–nesting emergences (NNE) and numbers ofnests (N) followed patterns similar to the monitor-ing effort (as defined by the number of completedbeach monitoring datasheets) (fig. 2A). As couldbe predicted, the seasonality of nesting closelyfollows the seasonality of the monitoring effort(fig. 2B). The inventory of completed beach moni-toring datasheets reveals that relatively little sur-vey effort was expended annually during the fivemonths between January and May, inclusive, and,given the seasonal pattern of nesting of theleatherback and loggerhead in the region, may inpart explain the absence of documentation of the-ses species. Nevertheless, the collected informa-tion revealed that nesting activities followed astrong seasonal pattern, with 97% of activities(non–nesting emergences and nests) recordedbetween the months of June and October, clearlypeaking in September (fig. 2B).

During the monitoring period (1999–2003), Wood-lands beach demonstrated the greatest nesting in-tensity of all beaches, but was also the most moni-tored beach on the island (fig. 3A). The three other

Montserrat0 1 2 4 km

Caribbean Sea

Montserrat

South America

Trant's Bay

Farm Bay

BlackburneAirport

(abandoned)

Exclusion Zone(closed)

Soufriere HillsVolcano

O'Garro's Estate

N

DaytimeEntryZone

Plymouth(abandoned)

Kinsale

Germans Bay

Sugar Bay

Fox Bay

Illes BayOld Road Bay

Lime Kiln Bay

Woodlands Bay

Bunkum Bay

Soldier Ghaut BayCarr's Bay

Little Bay

Rendez–vous Bay

Bransby Point

Hot Water Pond

Fig. 1. The island of Montserrat in the Eastern Caribbean, showing nesting sites and the ExclusionZone.

Fig. 1. Isla de Montserrat en el Caribe oriental, mostrando los lugares de nidificación y la Zona deExclusión.


key nesting beaches appeared to be Rendez–vous,Fox’s Beach/Bransby Point and Old Road/Iles Baybeaches (fig. 3A). Preliminary results for the 2004season indicated that Fox’s Bay was no longer aprime nesting site, while beaches near Plymouth(Hot Water Pond, Sugar Bay, Kinsale) showed in-creased nesting activities.

During the study period (1999–2003), 594 nestingattempts (including 263 successful nests) were docu-mented (table 1). During the more intensive moni-toring period between mid–August and mid–Sep-tember 2003, a total of 79 nesting attempts, includ-ing 19 successful nests were recorded (table 2).There were 21 non–nesting emergences and sixnests from green turtles, and 17 non–nestingemergences and three nests from hawksbill turtles.Because of their relatively large widths, four asym-metrical tracks observed on Trant’s beach wereattributed to loggerhead turtle(s), despite no nestbeing observed. The spatial distributions of non–nesting emergences and nests for mid–August/mid–September 2003 (fig. 3B) showed patterns simi-lar to those shown when all data are pooled for1999–2003 (fig. 3A). The numbers of non–nestingemergences for mid–August/mid–September 2003were highly correlated with the total numbers ofrecorded non–nesting emergences for the period1999 to 2002 (Spearmans rank correlationRs = 0.84; P < 0.01). This relationship in spatialpattern was also detected between the numbers ofnest for mid–August/mid–September 2003 and thetotal number of nests for the period 1999 to 2002(Rs = 0.57; P < 0.05).

Night–time monitoring of marine turtle nesting

In 2002 and 2003, a total of 28 individual nestingturtles were measured: 16 green turtles (12 in 2002,four in 2003; mean CCL (cm) = 106.9 ± 6.3 SD,range: 103–118) and 11 hawksbill turtles (nine in2002, two in 2003; mean CCL (cm) = 87.8 ± 6.8 SD;range: 79–103). A total of nine hawksbill (eight in2002, one in 2003) and 13 green turtles (11 in 2002,two in 2003) were PIT tagged. All were tagged onWoodlands beach, with the exception of threehawksbill turtles tagged on Carr’s Bay (two in 2002,one in 2003). In 2002, two green turtles were re–sighted on Woodlands beach, 11 and 12 days re-spectively, after having been PIT tagged on thatbeach. These data were supplemented by one sight-ing (by a member of the public) of a loggerheadturtle nesting on Woodlands beach in August 2002and hatchling leatherback turtles being discoveredand filmed on the same beach in the mid 1990’s(J. J., unpublished data).

Fishery harvest data

For the period 1993 to 2003, the harvest of 10 tur-tles was declared to the Fisheries Department(fig. 4), hence a mean of 0.9 harvest per year(± 1.2 SD; range: 0–4). All captures took placeduring the open season (October to May). Onegreen turtle (9.1 kg) and seven hawksbill turtles(13.6 kg, 18.1 kg, 29.5 kg, 45.4 kg, 45.4 kg,63.1 kg, 90.9 kg; mean mass (kg) = 43.7 ± 26.9 SD)were declared to the authorities. There were two

Fig. 2. The total numbers of completed beech monitoring sheets, records of non–nesting emergences(NNE), and records of nesting emergences (N) by year and cumulatively (A) and by month (B) for theperiod 1999 to 2003: J. January; F. February; Mr. March; Ap. April; My. May; J. June; Jl. July; Ag.August; S. September; O. October; N. November; D. December.

Fig. 2. Cifras totales de las hojas de control de las playas, registros de salidas sin nidificación (NNE)y registros de las salidas con nidificación (N) por: A. Año y acumulativamente; B. Mes para el período1999–2003. (For abbreviations of fig. 2B see above.)

A B

200

150

100

50

0

200

150

100

50

01999 2000 2001 2002 2003 J F Mr Ap My Jn Jl Ag S O N D

Year Month of the year

Beech monitoring sheets NNE N Beech monitoring sheets NNE N

Fre

qu

ency

Fre

qu

ency

164 Martin et al.

declared captures for which the species was notrecorded. All reported landings were of turtles thatmet the legal minimum size criteria (> 20 lbs, or9.1 kg).

Using a published regression equation be-tween mass and CCL for hawksbill turtles(Log10 (mass) = 2.8966 * Log10 (CCL) – 3.8534, withmass in kg and CCL in cm, Limpus et al., 1983),the masses of nesting hawksbill turtles that weremeasured in Montserrat were estimated to rangefrom 43.9 to 94.8 kg. When compared to the massesof harvested turtles, it appeared that four out of theseven harvested hawksbill turtles declared to au-thorities could have been adults. These four poten-tially adult turtles were captured during the monthsof February (N = 1 turtle), October (N = 1 turtle) andNovember (N = 2 turtles).

Discussion

Although marine turtle monitoring had been ongo-ing since preliminary studies in the early 1980’s(Meylan, 1983; John, 1984), almost all relevantdata were lost, along with many government records,in the volcanic flows that engulfed Plymouth in

1997. Monitoring efforts documented by this study(1999–2003) were intermittent and uneven, mean-ing that caution is warranted in making any recom-mendation regarding population status. There are,however, a few key points that can be extractedfrom the existing data. Green and hawksbill turtlesnest in modest yet regionally important numbers forthe Eastern Caribbean, probably every year.Leatherback and loggerhead turtles may also nest,but at lower densities. The lack of documentedleatherback nesting may be attributed to a com-paratively low level of monitoring during peak nest-ing months (April–June), however it is unlikely thatnesting of this species is more frequent than occa-sional. The data are in concord with the widerliterature which suggests that green, hawksbill andleatherback turtles (and loggerheads to a muchlesser extent) are the most common species ofnesting sea turtles in the Lesser Antilles (e.g. Carret al., 1982; Meylan, 1983, 1999; Eckert et al.,1992; Eckert & Honebrink, 1992; Fuller et al., 1992;Sybesma, 1992; D’Auvergne & Eckert, 1993; Scott& Horrocks, 1993; Richardson et al., 1999; Cheva-lier & Lartiges, 2001).

The magnitude of nesting data recorded wasclosely correlated with survey frequency in time

Fig. 3. The total numbers of completed beach monitoring sheets, records of non–nesting emergences(NNE) and numbers of nests (N), per beach (A) during the years 1999–2003 and (B) for the periodmid–August to mid–September 2003. Beach codes: Rv. Rendez–vous; Lt. Little Bay; Cr. Carr’s Bay;SG. Soldier Ghaut; Bn. Bunkum Bay; Wl. Woodslands Beach; LK. Lime Kiln Bay; OrI. Old Road/IlesBay; FB. Fox’s Bay/Bransby Point; HSK. Hot Water Pond/Sugar/Kinsale; GG. German’s/O’Garro’s;TF. Trant’s/Farm Bay); * One hawskbill turtle nest; ** Two hawksbill turtle nests. (For other abbreviationssee figure 2.)

Fig. 3. Cifras totales de las hojas de control de las playas, registros de las salidas sin nidificación (NNE)y números de nidos (N) por playa (A) durante los años1999–2003 y (B) para el período de mediadosde agosto–mediados de septiembre del 2003. Códigos de las playas: Rv. Rendez–vous; Lt. Little Bay;Cr. Carr’s Bay; SG. Soldier Ghaut; Bn. Bunkum Bay; Wl. Woodlands Beach; LK. Lime Kiln Bay; OrI. OldRoad/Iles Bay; FB. Fox’s Bay/Bransby Point; HSK: Hot Water Pond/Sugar/Kinsale; GG. German’s/O’Garro’s; TF. Trant’s/Farm Bay; * Un nido de tortuga carey; ** Dos nidos de tortuga carey. (Para lasotras abreviaturas ver la figura 2.)

200

150

100

50

0

Fre

qu

ency

Rv Lt Cr SG Bn Wl LK OrI FB HSK GG TF Rv Lt Cr SG Bn Wl LK OrI FB HSK GG TF Beach code Beach code

24

18

12

8

0

NNE N

ExclusionZone

Fre

qu

ency

Beach monitoring sheets NNE N

A B

ExclusionZone


Table 1. Breakdown of the records of non–nesting emergences (NNE), and the numbers of nests (N),per beach and year, for the period 1999–2003.

Tabla 1. Detalle de los registros de salidas del mar sin nidificación (NNE), y número de nidos (N), porplaya y por año, para el período 1999–2003.

1999 2000 2001 2002 2003

NNE N NNE N NNE N NNE N NNE N

Rendez–vous 7 6 1 3 4 3 23 25 22 34

Little 0 0 0 0 0 0 0 0 1 0

Carr’s 1 0 0 0 0 0 2 3 2 0

Soldier Ghaut 1 1 0 0 1 0 1 0 0 0

Bunkum 2 0 0 0 0 0 11 2 5 0

Woodlands 4 4 1 0 4 4 93 70 36 21

Lime Kiln 0 0 0 0 1 0 10 16 7 0

Old Road/Iles 9 4 1 4 9 3 11 7 2 0

Fox’s/Bransby Point 20 5 4 5 7 3 10 15 5 10

Hot Water Pond/Sugar/Kinsale 3 4 2 1 0 0 4 5 0 0

German’s/O’Garro’s 3 0 0 0 0 0 0 0 0 0

Trant’s/Farm 0 0 0 0 0 0 0 0 1 5

Total 50 24 9 13 26 13 165 143 81 70

Table 2. Breakdown of the numbers of non–nesting emergences (NNE), and the numbers of nests(N), per beach and by species, for the period mid–August to mid–September (2003).

Tabla 2. Detalle de los registros de salidas del mar sin nidificación (NNE), y número de nidos (N), porplaya y por especie, para el período mediados de agosto–mediados de septiembre (2003).

Green Hawksbill Loggerhead Undetermined

NNE N NNE N NNE N NNE N

Rendez–vous 0 1 0 2 0 0 7 4

Little 0 0 0 0 0 0 0 0

Carr’s 0 0 0 0 0 0 0 0

Soldier Ghaut 0 0 0 0 0 0 0 0

Bunkum 1 0 3 0 0 0 2 0

Woodlands 14 4 5 0 0 0 3 4

Lime Kiln 0 0 3 0 0 0 0 0

Old Road/Iles 0 0 4 1 0 0 1 0

Fox’s/Bransby Point 2 1 0 0 0 0 3 2

Hot Water Pond/Sugar/Kinsale 2 0 2 0 0 0 0 0

German’s/O’Garro’s 0 0 0 0 0 0 0 0

Trant’s/Farm 2 0 0 0 4 0 2 0

Total 21 6 17 3 4 0 18 10

166 Martin et al.

Fig. 4. The temporal distribution of reports of turtle captures (1993–2003; N = 10 turtles). The closedseason is highlighted. (For abbreviations see figure 2.)

Fig. 4. Distribución temporal de los registros de capturas de tortugas (1993–2003; N = 10 tortugas).Se ha destacado la estación de veda. (Para las abreviaturas ver la figura 2.)

and space. It is likely that recorders more frequentlycarried out surveys at times and locations when theprobability of recording turtle nesting activity wasmore likely. Although this may have resulted inspatial and temporal biases in the dataset, theseasonality of the Montserrat nesting season asdescribed by the data set is plausible, peaking fromJune to October, if we assume that hawksbill andgreen turtles are the dominant nesting species. Al-though because of the nature of the data, it was notpossible to discriminate between the seasonality ofthe different species, the temporal distribution of thedata are consistent with seasonality of nesting re-ported for hawksbill and green turtles (Fuller et al.,1992; Hirth, 1997) and hawksbill turtles (Eckert &Honebrink, 1992; Corliss et al., 1989; Scott &Horrocks, 1993) in the Eastern Caribbean region.Additionally, data collected during the period of in-tensive monitoring in 2003 generated a spatial distri-bution of nesting broadly similar with that of the datagathered in previous years. Notwithstanding, it islikely that comprehensive (e.g. once weekly, year–around) island–wide surveys would reveal more com-plex patterns of habitat use by gravid females.

The key nesting beaches for green and hawksbillturtles in Montserrat appeared to be Woodlands(so far unreported in the literature), Rendez–vous,Fox’s/Bransby Point and Old Road/Iles Beaches.Even though green turtles left tracks on many of theisland’s beaches, actual nesting by this specieswas only confirmed for Rendez–vous, Woodlandsand Fox’s/Bransby Point beaches. Based on inter-views with island residents and beach reconnais-sance, Meylan (1983) reported that green turtlesmight also be nesting at Little and Iles beaches.

Actual nesting by hawksbill turtles was solely con-firmed in the present study for Rendez–vous andOld Road/Iles Beaches, although Meylan (1983)also quotes Carr’s, Little and Soldier Ghaut beachesas nesting sites for this species. On Trant’s beach,tracks possibly left by loggerhead turtles were re-ported, in agreement with the belief that loggerheadturtles occasionally nest on the island (John, 1984).

It is thought that the turtle fishery has declinedsignificantly in magnitude since the extensive emi-gration from the island in recent years. Only tenturtles were declared to the fishing authorities forthe period 1993 to 2003. Popular accounts sug-gest that it is likely that this low total is the resultof significant under–reporting. Fishermen are saidto avoid declaring their catch to the authorities bybutchering turtle carcasses at sea both in andoutside the open season. Of great concern, asevidenced by the temporal distribution of declaredturtle capture records and the fact that potentialbreeding adults are being captured, is that theopen season for the turtle fishery overlaps partiallywith the nesting season. Consequently, in plannedregulations, it has been suggested that the closedseason be defined as 1 March (the beginning ofleatherback nesting season in the central EasternCaribbean) to 1 December. Other suggestedchanges in the regulation include the prohibition ofcatching turtles on land and an increase of theminimum weight of harvested turtles from 20pounds (9.07 kg) to 50 pounds (22.68 kg). How-ever, a recent report to the UK Government (Godleyet al., 2004) recommended that legislation befurther revised to "ensure a permanent and com-plete prohibit ion of harvest of any large,

3

2

1

0 J F Mr Ap My Jn Jl Ag S O N D

Month

Closedseason

Hawksbill Green Undetermined

Turt

le c

aptu

res


reproductively valuable turtles by instigating a maxi-mum size limit". It was suggested that this thresh-old should be based upon further research into thefishery and turtle stocks and that a curved cara-pace length threshold was developed. A shift fromweight–based to size–based limits enables a fish-erman to more easily determine the legality of thecatch while still at sea.

The harvest information suggests that a wide sizerange of green and hawksbill turtles could be presentyear round in Montserrat’s waters. Relatively little isknown of the current state of Montserrat’s marineand coastal habitats with regards to suitability asmarine turtle foraging areas. The area of the coastalshelf is relatively small (140 km2) and only general-ized distributions of primary substrata types areavailable (Meylan, 1983; Anonymous, 1993). Before1995, coral communities (foraging habitats forhawksbill turtles) were found in small patches inter-spersed with sand and sediment on the north,south and west coasts (Gell & Watson, 2000). Theharmful consequences of sediments on coral reefcommunities and associated organisms are welldocumented (e.g. Rogers, 1990). In Montserrat,volcanic sediments are thought to have had asevere impact on reef growth, particularly those inthe east and southwest of the island (Gell & Watson,2000). Direct deposits of ash and waterbornesediments have led to some coral bleaching anddisintegration of large sponges. Some reef areas,however, are thought to be recovering (Wolfe Krebs,pers. comm. 2003). In recent times, three mainseagrass beds (foraging habitats for green turtles)were known: the largest, 750 ha, being located atthe northern tip of the island and the other two onthe east and west coasts (Gell & Watson, 2000). Itis thought that seagrass beds suffered considerabledamage during Hurricane Hugo in 1989, althoughthe effect on the spatial distribution of foraginghabitat for green turtles is not known.

Montserrat presents a relatively narrow coastalshelf, dropping off rapidly to nearly 200 m only650 m from the shoreline along the southern half ofthe island, whilst in the north, northeast and west,the shelf slopes more gently (the 200 m contour isapproximately 5 km offshore, Gell & Watson, 2000).The result is a high energy, erosion prone coastline,with generally intermittent beaches (Anonymous,1993). For this reason, the quality of Montserrat’sbeaches with regards to sea turtle nesting appearsto be naturally poor. Although only assessed qualita-tively to date, beach erosion destroys incubatingeggs and periodically prevents gravid turtles fromnesting. Additional factors of concern are linked tothe volcanic eruptions and include ash deposits andbeach mining. Occasional ash deposits cover nest-ing beaches, rendering them less suitable or whollyunsuitable for nesting until they are cleared by heavystorms. For Montserrat’s rebuilding after the cata-strophic eruptions of 1997, extraction of beachsediments, largely of volcanic origin, are common-place. Such extraction has ceased at Isle’s Bays (in2003) but is ongoing at Trant’s Bay. It is important

that the integrity of Trant’s Bay be maintained andthat ongoing sea turtle monitoring, preferably on amore frequent basis, include the relocation of clutchesfrom high risk to lower risk beach areas. Nest preda-tion by feral pigs and feral/domestic dogs has alsobeen recorded (J. J. and B. J. G. pers. obs.), but theactual levels are yet to be quantified.

The present study has drawn a more accuratepicture of the status of marine turtles in Montserrat.Further studies involving species identification withincreased survey effort will more fully elucidate thestatus of nesting populations. Of high priority formarine turtle conservation are a revision of theregulatory framework to feature a more restrictedharvest season (and one that does not coincidewith the turtle breeding season), maximum ratherthan minimum size limits, new measures to en-courage fishermen to report their turtle catches,the full protection of nesting adults, their eggs andyoung, the careful management of beach sedi-ment extraction, and the control of feral pigs andferal/domestic dogs.

Acknowledgements

The authors would like to thank the staff ofMontserrat Fisheries Department, Montserrat Gov-ernors Office, Montserrat Ministry of Agriculture,Montserrat National Trust, Montserrat Volcano Ob-servatory, Royal Society for the Protection of Birds,Sea Wolf Diving School, and the following individu-als: Crystal & Dean Archer, Mrs Hilda Blake, Helen& Gerard Cooper, Bo Dalsgaard, Mr & Mrs Darby,Alfred Edwards, Lexvern Fenton, Anne–Marie &David Graham, Gerard Gray, Linda Green, JohnKeller, Mr & Mrs Krebs, Melissa O’Garro, GeoffPatton, Joe Philips, Sarah Sweeney , Mr & MrsWalker. Much of the fieldwork for this study wascarried out as part of the Turtles in the CaribbeanOverseas Territories (TCOT) project funded byDEFRA and the FCO’s UK Overseas TerritoriesEnvironment Fund. BG is a NERC Research Fel-low. Time to support final manuscript preparationwas provided through funding by the OverseasTerritories Environment Programme (OTEP) for theTurtles in the UK Overseas Territories (TUKOT).The manuscript considerably benefited from thecomments of Catherine Bell, Annette Broderick,,Claude Gerald, Matthew Godfrey and Kartik Shankerand two reviewers.

References

Anonymous, 1993. Environmental Profile, An As-sessment of the Critical Environmental IssuesFacing Montserrat with an Action Agenda for theFuture. United Nation Development Program(UNDP), Project No. MOT/92/002/A/01/99

Blankenship, J. R., 1990. The wildlife of Montserrat(including an annotated bird list for the island).Montserrat National Trust, Montserrat, West Indies.

168 Martin et al.

Groombridge, B. & Luxmoore, R., 1989. The greenturtle and hawksbill (Reptilia: Cheloniidae) worldstatus, exploitation and trade. Lausanne, Swit-zerland.

Hirth, H. F., 1997. Synopsis of the biological data onthe Green turtle Chelonia mydas (Linnaeus 1758).Biological Report 97(1), Fish and Wildlife Serv-ices, U.S. Department of Interior.

John, C. T., 1984. The national report for thecountry of Montserrat. In: Proceedings of theWestern Atlantic Turtle Symposium: 3.332–3.328, volume 3, appendix 7, the national re-ports (P. Bacon, F. Berry, K. Bjorndal, H. Hirth,L. Ogren & M. Weber, Eds.). Univ. of MiamiPress, Miami.

Limpus, C. J., Miller, J. D., Baker, V. & McLachlan,E., 1983. The hawksbill turtle, Eretmochelysimbricata (L.), in the North–Eastern Australia:the Campbell Island Rookery. Australian WildlifeResearch, 10: 185–197.

Meylan, A. B., 1983. Marine turtles of the LeewardIslands, Lesser Antilles. Atoll Research Bulletin,278: 1–43.

– 1999. Status of the hawksbill turtle (Eretmochelysimbricata) in the Caribbean region. ChelonianConservation & Biology, 3: 177.

Pritchard, P. C. H. & Mortimer, J. A., 1999. Tax-onomy, External Morphology, and Species Iden-tification. In: Research and Management Tech-niques for the Conservation of Sea Turtles:21–38 (K. L. Eckert, K. A. Bjornda, F. A. Abreu–Grobois & M. Donnelly, Eds.). IUCN/SSC Ma-rine Turtle Specialist Group Publication No. 4.

Richardson, J. I., Bell, R. & Richardson, T. H.,1999. Population Ecology and Demographic Im-plications Drawn From an 11–Year Study ofNesting Hawksbil l Turtles, Eretmochelysimbricata, at Jumby Bay, Long Island, Antigua,West Indies. Chelonian Conservation and Biol-ogy, 3: 244–250.

Rogers, C., 1990. Responses of coral reefs andreef organisms to sedimentation. Marine Ecol-ogy Progress Series, 62: 185–202.

Scott, N. & Horrocks, J. A., 1993. WIDECAST Seaturtle Recovery Action Plan for St. Vincent andthe Grenadines, CEP Technical Report No. 27.In: UNEP Caribbean Environment Programme:1–80 (K. L. Eckert, Ed.). Kingston, Jamaica.

Sybesma, J., 1992. WIDECAST Sea turtle RecoveryAction Plan for the Netherlands Antilles, CEPTechnical Report No 11. In: UNEP CaribbeanEnvironment Programme: 1–63 (K. L. Eckert,Ed.). Kingston, Jamaica.

Carr, A., Meylan, A. B., Mortimer, J., Bjorndal, K. A.& Carr, T., 1982. Survey of sea turtle populationsand habitats in the Western Atlantic. NOAA Tech-nical Memorandum NMFS–SEFC 91, U.S. De-partment of Commerce.

Chevalier, J. & Lartiges, A., 2001. Les TortuesMarine des Antilles, Etude Bibliographique. Of-fice National de la Chasse et de la FauneSauvage, CNERA Faune d’Outre Mer.

Corliss, L. A., Richardson, J. I., Ryder, C. & Bell,R., 1989. The hawksbills of Jumby Bay, Antigua,West Indies, In: Proceedings of the Ninth AnnualWorkshop on Sea Turtle Conservation and Biol-ogy: 33–35 (S. A. Eckert, K. L. Eckert, T. H.Richardson, Eds.). NOAA Tech. Memo. NMFS–SEFC–232. U. S. Department of Commerce.

D’Auvergne, C. & Eckert, K. L. 1993. WIDECASTSea turtle Recovery Action Plan for St Lucia,CEP Technical Report n°26. In: UNEP CaribbeanEnvironment Programme: 1–70 (K. L. Eckert,Ed.). Kingston, Jamaica.

Eckert, K. L. & Honebrink, T. D., 1992. WIDECASTSea turtle Recovery Action Plan for St Kitts andNevis, CEP Technical Report n°17. In: UNEPCaribbean Environment 292 Programme: 1–92(K. L. Eckert, Ed.). Kingston, Jamaica.

Eckert, K. L., Overing, J. A. & Lettsome, B. B.,1992. WIDECAST Sea turtle Recovery ActionPlan for British Virgin Islands, CEP TechnicalReport n°15. In: UNEP Caribbean EnvironmentProgramme: 1–116 (K. L. Eckert, Ed.). Kingston,Jamaica.

Fuller, J. E., Eckert, K. L. & Richardson, J. I., 1992.WIDECAST Sea turtle Recovery Action Plan forAntigua and Barbuda, CEP Technical Report n°16.In: UNEP Caribbean Environment Programme: 1–88 (K. L. Eckert, Ed.). Kingston, Jamaica.

Gell, F. & Watson, M., 2000. UK Overseas Territo-ries in the Northeast Caribbean: Anguilla, BritishVirgin Islands, Montserrat. In: Sea at the Millen-nium: an Environmental Evaluation: 615–626 (C.Sheppard, Ed.). Pergamon, Elsevier Science Ltd.,United Kingdom.

Godley, B. J., Broderick, A. C., Campbell, L. M.,Ranger, S., Richardson, P. B., 2004. An Assess-ment of the Status and Exploitation of MarineTurtles in the UK Overseas Territories in theWider Caribbean. Final Project Report to theDepartment of Environment, Food and RuralAffairs and the Foreign and Commonwealth Of-fice: 1–253.Available online at: http://www.seaturtle.org/mtrg/projects/tcot/

http://www.seaturtle.org/mtrg/projects/tcot/

http://www.seaturtle.org/mtrg/projects/tcot/



Hilaluddin, Kaul, R. & Ghose, D., 2005. Conservation implications of wild animal biomass extractions inNortheast India. Animal Biodiversity and Conservation, 28.2: 169–179.

AbstractConservation implications of wild animal biomass extractions in Northeast India.— We investigated thepatterns of wild meat extraction and consumption by indigenous communities in Northeast India. Ourrespondents hunted at least 134 species of wild animals over the previous year in the villages surveyed andcontinued to harvest and use wild meat as their cash income increased. These indigenous communities ofNortheast India showed an average of 32 to 59% dependency on the forestry sector. Wild meat contributedsignificantly (up to 25%) to their economies, suggesting previous assessments of dependence on theforestry sector should be reviewed. All sections of the society exploited wild meat equally. As educationseems to play a role in reducing wild meat extractions, increased awareness in conservation of naturalresources should be promoted .

Key words: Wild meat consumption, Wild meat trade, Dependency, Northeast India.

ResumenRepercusiones en la conservación debidas a las extracciones de biomasa animal salvaje en el nordeste dela India.— Investigamos los patrones de extracción y consumo de carne de caza por parte de lascomunidades indígenas del nordeste de la India. En la aldea estudiada, los sujetos interrogados habíancazado al menos 134 especies de animales salvajes durante el año anterior, y continuaron cazando yutilizando la carne de caza cuando sus ingresos aumentaron. Estas comunidades indígenas del nordestede la India dependían en promedio del 32 al 59% del sector forestal. La carne de caza contribuíasignificativamente (hasta un 25%) a sus economías, lo que sugiere que deberían revisarse las evaluacionesprevias sobre la dependencia del sector forestal. Todas las capas sociales explotaban la carne de caza deigual forma. Dado que parece que la educación juega un papel significativo en la reducción de lasextracciones de carne de caza, debería promoverse una mayor concienciación de la conservación de losrecursos naturales.

Palabras clave: Consumo de carne de caza, Comercio de carne de caza, Dependencia, Nordeste de laIndia.

(Received: 11 VIII 04; Conditional acceptance: 25 I 05; Final acceptance: 3 V 05)

Hilaluddin, Rahul Kaul & Dipankar Ghose, World Pheasant Association, South Asia Field Office, J–7/21, DLFPhase II, Gurgaon–122 002.

Corresponding author: Hilaluddin. E–mail. [email protected]

Conservation implications ofwild animal biomass extractions inNortheast IndiaHilaluddin, R. Kaul & D. Ghose


170 Hilaluddin et al.

Introduction

In Northeast India, people hunt wild animals forseveral reasons and therefore rural people are heav-ily dependent on wild meat (Hilaluddin, 2005a,2005b). However, game may often be over–huntedand may have caused local extinctions of severalspecies such as the Green peafowl Pavo muticus inSoutheast Asia (McGowan et al., 1998). The prob-lem can only be tackled by looking at the widereconomic and institutional context within which hunt-ing occurs, from household economies to trade (Ab-ernethy et al., 2003). However, quantitative data onthe amount of wild meat harvest, its consumptionand trade in Southeast Asia in general and North-east India in particular, are lacking. Therefore, thereis an urgent need to quantify the intensity of wildmeat extractions and assess impacts of such extrac-tions on wild animal populations.

There has also been little work to determine thecontribution of the forestry sector to the life of localpeople (Bahuguna, 1993). Economic benefits ac-cruing to local economies from the forests haveseldom been estimated (Bahuguna, 2000) and inmost cases are incomplete. The economic value ofanimal biomass may have been significant but itwas often ignored in earlier assessments whichmainly pertained to timber and non–timber forestproducts (fuelwood, fodder, fruits, seeds, medicinalderivatives of plants, etc.). People’s dependence onwild meat, in particular, remains unknown despiteharvesting of roughly 23,500 tonnes annually inSarawak (Bennett, 2002), 67,000–1,64,000 tonnesin the Brazilian Amazon (Robinson & Redford, 1991;Peres, 2000) and 1 million to 3.4 million tonnes inCentral Africa (Wilkie & Carpenter, 1999; Fa et al.,2002).

There is also a need to assess the benefitsderived from the wild meat in order to demonstratethe tangible contribution of the forestry in generaland wild meat in particular to the society. This isalso essential to understand the significance of wildmeat in the local economy —both for cash andsubsistence needs— and local cultural beliefs (Ab-ernethy et al., 2003).

The economic theory of "Income and Consump-tion" (Kuznets, 1955), which is now used world–widein most natural resources conservation action plans,suggests that consumptions of a commodity go upwith an increase in household income if it has nosubstitutes or is considered superior to substitutes.Otherwise, the use of goods falls with rising income,showing inverted "U shape" patterns. Kuznets’ modelof consumption may not be universally applicable toall goods, however, even if they are inferior, especiallyin regions of the world where people have developeda taste for a few specific goods for reasons other thaneconomic. His model may thus vary across the natureof goods and areas. Therefore, there is a need toinvestigate the validity of Kuznets (1955) model inconsumption of important forest products such aswild meat before its incorporation into a conservationaction plan.

We undertook a survey in Northeast India toassess whether the extraction of wild meat byAngami, Apatani, Mizo and Nishi communitieswas a conservation problem in the region. Spe-cifically, we sought to determine whether con-sumption of wild meat was linked to people’sincome. In order to answer this question westudied the prevalence of wild meat extractionand consumption, the species hunted and differ-ences in hunting patterns of indigenous commu-nities, the linkage between wild animal huntingand trade, the role of wild meat in local economy,and the impact of education, age and professionof a person on wild meat extraction. We alsocollected information on other forest productsharvested by a household in order to calculateincome of that household from the forestry. Theamounts of all forest products extracted by ahousehold are quantified and their quantities areconverted into monetary values based on theirprevalent spot prices for estimating a householdincome from forestry (Malhotra et al., 1991; Hedgeet al., 1996). According to Bahuguna (1993,2000), the income of a household must be calcu-lated by summing incomes of that householdfrom all sources viz. agriculture (labour and crops),forestry (forest products and forest managementactivities) and other employment opportunities(self and government employment).

Methods

The survey included three methods: A general vil-lage level survey, a household level survey, andfinally a market survey. Animal extraction data werecollected by way of a detailed set of questionnairesand were not independently measured amounts.

The qualitative and quantitative information bothat village/hamlet and household level on the animalextraction patterns was gathered following a combi-nation of PRA (Sankaran et al., 2000) and RRA(Sethi & Hilaluddin, 2001) methods. We collectedinformation on the animal species and theirnumber(s) killed by a household during the previ-ous year. The respondents were shown pictures ofanimal species for the purpose.

A total of 25 villages were surveyed, represent-ing four communities (Angamis 6; Apatanis 5; Nishis8; Mizos 6). The villages were from the interior andexterior forest blocks among the settlements of thestudied communities, thereby covering most of theirhabitation ranges.

Generally, one interview with a group of villagerswas conducted at the village/hamlet level. Duringthis interview we sought wide–ranging informationabout the resource use patterns (those interested inthe questionnaire and the list of species hunted withtheir numbers will be sent the information uponrequest to the author) Such interactions were usuallya good introduction to the purpose of our surveys,and subsequent data collection at the village levelbecame easier (Hilaluddin et al., in press b).


After the village level focal group interviews, wewere able to focus on individuals involved in somelevel of forest produce gathering. A total of 134household level interviews (Angamis 33; Apatanis33; Nishis 30; Mizos 38) were conducted in Aizwal,Kohima and Lower Subansiri districts of NortheastIndia (fig. 1).

Following these interviews, we divided familiyunits into hunting or non–hunting households. Wefurther classified them into business, farming andservice communities. We selected respondents forhousehold level interviews following random sam-pling techniques (Sutherland, 1996). Sampling ef-forts covered at least five percent for each cat-egory of households (hunting and non–hunting) ineach surveyed village and town for collecting dataon range of animals extracted by the household.We also gathered information on agriculture cropsand wild plants (timber, firewood, fodder, bamboo,medicinal plants and other NTFPs) products, withtheir prevailing spot price, gathered by the house-hold during the previous year. In addition, a house-hold’s income from other avenues (agriculture la-bour, forestry labour and other employment oppor-tunities) was also quantified. We also collecteddata on age, education status and size of therespondent’s family. Educational level was as-sessed from the number of school years (1–15)he/she had passed from a recognized institution.

The respondents in the household level inter-views were mainly selected randomly but some-times on the advice of our guide who hailed fromthe village. If both a man and a woman from thehousehold were present, we interviewed the manbecause only male members, within the indig-enous communities studied, hunt wild animals.

We also conducted wild meat trade surveys for aperiod of 15 days each in the local markets ofKohima city (Nagaland) and Hapoli town (ArunachalPradesh). The main purpose of this survey was toestablish whether there was trade of wild meat inurban centers and also whether these marketsconnected to the remote areas of our survey sites.We recorded species being sold in the markets withtheir numbers and price.

Data on the intensity of hunting within a villagewas calculated from the estimated number of ani-mals killed by each household/annum for eachspecies. Crude wildmeat amount extracted by ahousehold for each species was calculated usingthe average body weight of adult individuals. Meanbody masses of animals were taken from the litera-ture (Prater, 1971; Ali & Ripley, 1987) with theexception on fishes. Information on the quantitiesof extraction of fishes and other forest products bya household were directly gathered in per unitmeasurement in the field.

We calculated a household’s income from for-estry by converting quantities of wild animal andplant species extracted by that household into mon-etary values based on their prevalent spot prices.We also included the income of that householdfrom forest management activities such as forestrylabour, nursery, and forest watch and ward activi-ties. The gross annual incomes of households werecalculated by summing their incomes from variousincome sectors viz. agriculture (crops and labour),forestry (plants, animals and employment throughforest management activities) and other employ-ment opportunities (self and government employ-ment).

We investigated the relationship between wildmeat extraction and consumption rates of Angami,Apatani, Mizo and Nishi communities using Inde-pendent sample t test because these communitiesharvest wild meat both for self–use and for sale.The impact of socio–economic variables, specifi-cally age, educational status, and incomes derivedfrom cash avenues on wildmeat extraction rates,were calculated using Pearson’s correlation coeffi-cient. The differences in mean values of wild meatextracted by people in different occupations wereinvestigated using the Kruskal–Wallis test.

The monetary significance of wildmeat extrac-tions to local economies of Angami, Apatani, Mizoand Nishi communities were examined using OneWay ANOVA and therefore the null hypothesis "thevariations in mean values of dependency sourceswere statistically non–significant" was tested. Thiswas used to infer whether dependencies of a house-hold on each source contributed significantly to thelocal economy. We also investigated the impact ofcash income from agriculture, forestry (other thanwildmeat) and other employment opportunities onwildmeat consumption using Pearson’s Productcorrelation. This was used to examine the impact ofcash income on wildmeat use by a household. Wecompared gross annual incomes and dependenciesof surveyed households on various income sources

N

India

2 41

3

Fig. 1. Locations of surveyed villlages inNortheast India: 1. Angami; 2. Apatani; 3.Mizo; 4. Nishi.

Fig. 1. Localización de las aldeas del nordestede la India estudiadas: 1. Angami; 2. Apatani;3. Mizo; 4. Nishi.


across their respective villages using Kruskal–Wallis.All the data, wherever appropriate, were normalizedand statistical procedures were applied followingSokal & Rohlf (1995).

Results

Socio–economic profile of respondents

Out of 134 respondents, Angami and Apatani indig-enous communities each represented 25%, 22%were from the Nishi community and the rest wereMizo. Most of our respondents were literate. Theirage, education status and family size are presentedin table 1.

Wild meat survey

Our respondents in the villages surveyed extractedat least 137 wild animal species, including 50 mam-mals and one reptile, during the previous year.Apatani household extracted on average of 282 kgwild meat annually (table 2), mainly from mammals(85%). Birds formed 5.98% of Apatani’s extraction.The most relevant group among birds wasgalliformes. Other animals (mainly fish) representeda significant part of the Apatani diet and constituted8.4% of the extracted meat by weight. An Angamihousehold extracted about 457 kg of wild meatannually, which was mainly (89%) mammals. Birdstoo formed a substantial component (9.4%). A Mizohousehold extracted a mean of approximately 278 kgwild meat annually, of which 89% came from mam-mals. Birds constituted 3.4% and the majority weregalliformes. Other animals formed 7.8% of theMizos’ total wild meat extracted. Amongst Nishis,mammals constituted 69% of the total wild meatextracted (average approximately 545 kg) annually.Other animals formed a substantial component(17.5%) and birds formed about 13.4%.

Wild meat market survey

We observed a total of 773 dead animals (233mammals and 540 birds) in the markets of Kohimaover 15 full days of observation and recorded 53wild animal species (15 mammals and 38 birds).Similarly, we examined a total of 601 dead wildanimals (418 mammals and 183 birds) in the mar-kets of Hapoli, and recorded 19 wild animal species(10 mammals and 9 birds). A total of 118.62 kg ofwild meat (80.39% from mammals and the restfrom birds) was available at Hapoli and 154.33 kgof wild meat (73.92% from mammals and the restfrom birds) at Kohima. All animal meat came fromadjoining rural areas.

Wild meat and socio—economic variables

We investigated the relationship between wild meatextraction and socio–economic variables (table 3).A significant relationship emerged only amongst

Angami and Mizo communities. Angamis with ahigher income from sources other than wild meattended to harvest more wild meat. The extraction ofwild meat amongst Mizos declined the higher theeducation level. Extraction of wild meat showed nostatistically difference in relation to occupation(Kruskal–Wallis, n.s.)

An analysis using Pearson correlation coefficientwas performed to determine the effect of income onwild meat consumption. With the exception of Mizocommunity (fig. 1), significant positive correlationswere observed between gross cash income andamount of wild meat consumed by Angami, Apataniand Nishi communities (fig. 2).

Incomes and dependencies

We estimated average annual gross incomes of theindigenous communities included in the study (ta-ble 4). Incomes were interpreted as accruals on thebasis of cash values of the forest and agriculture–based goods obtained by a household in addition toincomes from other employment opportunities (e.g.self–employment i.e. business, and governmentemployment i.e. state and federal governmentfunded employment in various public departments).

Bulk of average income (approximately 25%) toa Nishi household is derived from wild meat, whichis conspicuously higher than their incomes fromagriculture and other employment vocations (selfand government employment). Similarly, Angami,Apatani and Mizo households derived average 14–16% incomes from wild meat.

Gross annual incomes of the study communi-ties from various income sources (crops, agricul-tural work, wild plant products, forest manage-ment activities, wild meat, self–employment andgovernment employment) and now they state morepossibi l i t ies) showed significant differences(Angami: F6 224 = 6.47, P < 0.001; Apatani:F6 224 = 11.6, P < 0.001; Mizo: F6 259 = 2.71, P < 0.01;Nishi: F6 203 = 3.92, P < 0.001, One way ANOVA).

Similarly, these sources of income varied signifi-cantly among the four communities (Angami:F6 224 = 16.4, P < 0.001; Apatani F6 224 = 18.3,P < 0.001; Mizo F6 259 = 7.66, P < 0.001; NishiF6 203 = 15.53, P < 0.001, One way ANOVA). How-ever, incomes and dependencies of these commu-nities on various sources across their respectivevillages did not show significant variations (Kruskal–Wallis test, n.s.).

Discussion

A large number of mammals and birds are huntedin Northeast India (Hilaluddin, 2005a, 2005b) andmany of these are of concern to conservation(Birdlife International, 2000; IUCN, 2003). In thevillages surveyed, the hunted animals included 20species considered as threatened on the Red DataList (IUCN, 2003); four Endangered (Acerosnipalensis, Bubalus bubalus, Elephas maximus and


Table 1. Socio–economic profile of the respondents' communities.

Tabla 1. Perfil socioeconómico de los sujetos de las comunidades interrogados.

Age (years) Education (class) Family size % Literacy rate

Tribe Mean S.E. Mean S.E. Mean S.E. Literate Illiterate

Angami (N = 33) 49 3 7 1 6 1 84.6 15.4

Apatani (N = 33) 45 3 6 1 6 1 57.6 42.4

Nishi (N = 30) 42 3 5 1 8 1 53.3 46.7

Mizo (N = 38) 51 2 7 1 7 1 94.7 5.3

Table 2. Wild meat extraction patterns (in kg/household/annum, mean ± CI) of sampled indigenouscommunities in Northeast India. (Data at 95% confidence level.)

Tabla 2. Patrones de extracción de carne de caza (en kg/familia/año, media ± CI) de las comunidadesindígenas muestreadas del nordeste de la India. (Datos con un nivel de confianza del 95%.)

Mode of Income Angami (N = 33) Apatani (N = 33) Nishi (N = 30) Mizo (N = 38)

Wildmeat 651.7 ± 349.1 282.1 ± 138.4 545.9 ± 186.3 277.7 ± 140.8457.5 ± 211.8 239.1 ± 92.0 496.2 ± 151.5 188.8 ± 71.7

Mammals 564.3 ± 291.8 241.4 ± 105.5 377.3 ± 132.0 246.5 ± 132.8408.2 ± 192.7 208.5 ± 84.9 346.2 ± 109.6 172.0 ± 70.7

Herbivores 9.91 ± 201.98 184.2 ± 88.92 277.66 ± 117.6 206.5 ± 110.11322.64 ± 147.97 164.56 ± 69.65 255.76 ± 92.8 151.68 ± 66.01

Carnivores 164.42 ± 95.68 57.17 ± 23.76 99.64 ± 35.9 40.02 ± 24.8285.53 ± 50.53 43.93 ± 21.91 90.49 ± 31.08 20.30 ± 9.15

Birds 51.5 ± 31.1 16.9 ± 9.1 73.1 ± 48.1 9.5 ± 6.043.2 ± 22.9 16.0 ± 7.8 54.5 ± 19.4 6.2 ± 2.7

Galliformes 23.38 ± 18.9 11.9 ± 8.74 29.73 ± 13. 34 4.83 ± 4.4215.62 ± 8.14 11.19 ± 7.32 26.66 ± 10.43 3.00 ± 1.44

Other birds 28.16 ± 14.72 4.86 ± 2.43 43.35 ± 43.03 4.67 ± 2.2627.60 ± 19.53 4.81 ± 2.06 27.8 ± 14.51 3.17 ± 1.68

Other animals 35.9± 42.7 23.8 ± 23.6 95.5 ± 74.7 21.7 ± 12.36.1 ± 5.7 14.6 ± 13.3 95.5 ± 74.7 10.6 ± 3.2

Table 3. Pearson’s correlation coefficients between socio–economic factors and wild meat harvestof the sampled indigenous communities of Northeast India: A. Age; E. Education; I. Income; *Denotes level of significance (P < 0.05).

Tabla 3. Coeficientes de correlación de Pearson entre los factores socioeconómicos y la extracción decarne de caza de las comunidades indígenas muestreadas del nordeste de la India: A. Edad; E.Educación; I. Ingresos; * Indica nivel de significación (P < 0,05).

Angami (N = 33) Apatani (N = 33) Nishi (N = 30) Mizo (N = 38)

Product A E I A E I A E I A E I

Wild meat –0.29 0.04 0.44* 0.11 –0.27 0.26 –0.23 –0.07 –0.09 0.08 –0.40* –0.07


Panthera tigris), eight Vulnerable (Capricornissumatraensis, Macaca assamensis, Manispantadactyla, Neofelis nebulosa, Panthera pardus,Presbytes pileatus, Selenarctos thibetanus andTragopan blythii) and rest Lower Risk: near threat-ened (Columba punicea, Cuon alpinus, Felisbengalensis, F. viverrina, Hylopetes alboniger,Nemorhaedus goral, Nycticebus coucang andPrionodon pericolor).

In India, under the Wild Life Protection Act1972, it is illegal to kill any wild life (Anon, 2003).Our interactions with respondents revealed thatalmost half were aware of this law and the penal-ties for violation . We therefore feel that in somecases our respondents may have revealed lowerfigures of animals than those actually hunted andthe conservation problem may be graver thanreported here.

The loss of species to hunting warrants urgentattention in Northeast India because forests hereare already much reduced in area and are increas-ingly fragmented as a result of shifting cultivation(FSI, 2003). This implies that populations of spe-cies endemic to this habitat type are not only at riskof loss of habitat and populations becoming iso-lated from each other, but also from easier accessto hunters. The loss of relatively small numbers ofindividuals, especially species that are included inthe Red List, may have a disproportionate impacton small and isolated populations.

Another important issue that warrants attention isthe wild meat consumption patterns of the surveyedindigenous communities. Our findings (fig. 2) arecontrary to the Kuznets model (1955) of income andconsumption of goods. Our models indicate thathouseholds in Northeast India continue to use wild

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.00 1 2 3 4 5

0 1 2 3 4 US $ income

6

5

4

3

2

1

0

Fig 2. Relationship between wildmeat consumption (kg/year) and incomes: A. Mizo community; B.Nishi community. (All values normalized.)

Fig. 2. Relación entre el consumo de carne de caza (kg/año) y los ingresos: A. Comunidad Mizo; B.Comunidad Nishi. (Valores normalizados.)

Wild

mea

t co

nsu

mp

tio

nW

ildm

eat

con

sum

pti

on

A

B

r = 0.22N = 38P = n.s.

r = 0.51N = 30P < 0.01


meat and may even increase their consumption ofwild meat as their income increases. The modelsshow that a rise in income may have the unexpectedand undesirable side effect of promoting consump-tion–driven increases in hunting pressure. Given theopen access nature of wild meat and its demand inthe region, a rise in income levels could easilyenhance the demand for wild meat and conse-quently induce over–harvesting of the species bothin the short and the long term. The long term is notrelevant for the species that are most threatened byhunting, for which extinction within a decade is a realpossibility (Nelleman & Newton, 2002). Therefore,strong intervention is required where there is a needto reduce hunting levels.

It is essential to understand not only the impactof hunting on wild populations but also the reasonwhy certain species are hunted (Kaul et al., 2004).

Firstly, hunting has a religious and cultural signifi-cance to many communities in Northeast India(Hilaluddin, 2005a). For example, the religious ritu-als of the Apatani community include generousofferings of smoked Funambulus palmarus, F.pennanti, Hylopetes alboniger and Dremomyslokriah. The Apatani community also sacrificesMacaca assamensis to propitiate their deity duringtheir annual spring festival, "Morum". The festival’sfeasting includes a voluminous amount of Muntiacusmuntjak and Sus scrofa meat.

Barbets, specifically Megaliama virens, are oftenserved to entertain special family guests. Nishipriests decorate their headgear with Selenarctosthibetanus skins and a pair of hornbill tail feathers.Furthermroe, Nishis prize the skin of Presbytespileatus for making sheaths for their traditionaldaggers, "Davs". Other community members adorn

Fig. 2. Relationship between wildmeat consumption (kg/year) and incomes: C. Angami community; D.Apatani community. (All values normalized.)

Fig. 2. Relación entre el consumo de carne de caza (kg/año) y los ingresos: C. Comunidad Angami;B. Comunidad Apatani. (Valores normalizados.)

0 1 2 3

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 US $ income

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0

5

4

3

2

1

0

–1

Wild

mea

t co

nsu

mp

tio

nW

ildm

eat

con

sum

pti

on

C

D

r = 0.73N = 33P < 0.001

r = 0.97N = 33P < 0.001


Table 4. Gross annual incomes (in US $) of sampled indigenous communities of Northeast Indiafrom wild meat and other income avenues. (Data at 95% confidence level.)

Tabla 4. Ingresos anuales brutos (en $ americanos) de las comunidades indígenas muestreadas delnordeste de la India, provinentes de la carne de caza y otras fuentes. (Datos con un nivel de confianzadel 95%.)


Mode of Income Mean CI Mean CI Mean CI Mean CI

Agriculture 2,164.9 656.6 11,505.7 396.9 853.5 333.9 1,434.8 1,052.7

Agriculture crop 2,105.9 656.8 11,459.7 401.3 754.5 316.3 1,361.4 1,052.6

Agriculture labor 59.0 50.5 445.9 29.7 98.9 73.1 73.4 95.9

Forestry 2,365.6 902.6 11,343.0 300.8 2,149.6 775.4 1,216.8 559.2

Plant 1,088.6 330.7 711.1 172.3 1,241.4 739.3 545.1 463.9

Timber 736.3 319.9 172.1 113.1 194.4 168.2 66.8 129.8

Gross NTFPs 352.2 136.1 539.0 91.8 1,047.9 688.2 478.4 452.8

Bamboo 68.4 98.4 176.1 57.8 20.5 9.07 293.4 257.7

Fuelwood 231.9 114.2 324.7 66.7 487.4 161.2 183.9 65.1

Other NTFPs 51.8 30.3 38.1 20.9 539.1 538.4 1.1 0.9

Forest management 0 0 136.5 155.1 91.2 136.4 109.9 109.0

Animal 1,277.1 847.5 495.4 219.8 817.0 275.9 561.7 294.9

Mammals 875.3 539.9 368.7 169.3 517.2 173.5 420.0 245.9

Birds 324.2 249.1 86.9 50.1 156.6 64.6 46.8 30.9

Other animals 77.7 88.6 39.8 45.6 143.2 124.3 94.9 65.5

Other employment opportunity 1,636.7 965.2 1,219.5 472.4 962.4 615.8 2,171.5 1,193.6

Government employment 927.2 388.4 833.8 367.8 538.8 313.2 835.9 357.1

Self employment 709.5 779.5 385.6 366.5 423.6 465.4 1,335.5 1,066.5

Gross Income 6,163.91,679.8 4,068.2 679.9 3,965.51,230.3 4,823.1 2,016.1

their caps with hornbill beaks, specifically Acerosnipalensis, and a pair of Dicrurus paradiseus and/orD. remifer tail feathers.

Several species are also popular among localsfor their role in traditional medicines in local beliefs(Hilaluddin, 2005b). Amongst Mizos, flesh of Macacaassamensis is associated with relieving delivery painsand is also believed to aid the development of theinfant while inside the mother’s womb; bats aresupposed to cure asthama; the gall bladder ofSelenarctos thibetanus heals jaundice; and the liverof Hylobates hoolock kills malarial parasites. Angamisconsume Upapa epops to alleviate male impotency.

Secondly, it appears that the primary objective isto secure an animal for consumption or sale. Theopportunity cost for the extraction of a wild animalwhich is relatively more common than others shouldbe less than that of less common ones. Thus, themost abundant wild animals are expected to beharvested more intensively than the less abundantones. However, the opportunity cost also depends

on body size of the target quarry, and therefore thequantity of meat rather than the quality generallydictates direct preferences. Unfortunately, abun-dance estimates for most animal species are lack-ing for Northeast India in general and our studyarea in particular, making it difficult to determinewhether offtake is adversely affecting wildpopulations (Hilaluddin et al., in press a). Thisrequires investigation.

Thirdly, wild meat in our surveyed areas is alsoharvested for trade. It appears that families living incomparatively remote areas have poor access tomarkets and where substitutes are not available,people mainly rely on wild meat for protein. How-ever, those who have migrated to cities and townsfor a better living have not lost the "taste" for wildmeat. In such areas, wild meat constitutes a "supe-rior good" and people pay 1 to 5 times the domesti-cated animal meat. The markets in the towns seemto be fed directly from the remote areas wherepeople may kill wild animals mainly to cater for the


demand of urban areas and thereby ensuring asupply of wild meat even if city dwellers do not havethe time or opportunity to hunt regularly.

Therefore, wild animal hunting in our study areaclearly demonstrates a direct link between level ofharvest and economic growth of those involved inwildlife trade. Thus, there is likely to be an increasein the wild meat extraction intensities of commer-cial hunters with an increase in urban populations.This increased appetite for goods should furtherstress the need to exploit animal resources in re-mote forest areas. One such example is found inthe link between demand for tropical hard timber inthe international market and over–exploitation offorests in Southeast Asia and the Amazon basin(Kolk, 1996; Dauvergne, 1997; Barker, 1998).

The majority of our respondents belonged to aneconomic stratum well above the poverty line (an-nual income above Indian Rupees 11,000/house-hold or 244.44 US $). Taking the other incomesources into account (fodder and shifting cultiva-tion), these figures would further add up to a signifi-cant annual income.

Amongst Nishis, a large proportion of ruralincome is derived from the forestry. Amongst Mizos,the bulk of income is derived from other employ-ment opportunities (self and government employ-ment) and the agriculture sectors followed by theforestry. Such patterns are contrary to the generaleconomy and employment of rural India which islargely agriculture based (Sethi & Hilaluddin, 2001).However, the rural economy of Angamis seems toconform the general agriculture based economicpattern of rural India. The rural economy of theApatanis shows equal dependence on agricultureand forestry sectors. Amongst the Nishis, wild meatoccupies an important place in village economy.Such an economic pattern is similar to the ruraleconomy of Ghana where wild meat makes a sig-nificant contribution to both the household foodsupply and as cash income (Dei, 1989). Our re-spondents seemed to be highly dependent uponwild meat for both their kind and cash values.

Our estimated annual incomes and dependen-cies of Angami, Apatani, Mizo and Nishi communi-ties on the forestry are not directly comparable with

Table 5. % Dependencies of sampled indigenous communities of Northeast India on wild meat andother income sources. (Data of 95% confidence level.)

Tabla 5. Dependencias de la carne de caza y de otras fuentes de ingresos de las comunidadesindígenas muestreadas del nordeste de la India. (Datos con un nivel de confianza del 95%.)


Mode of Income Mean CI Mean CI Mean CI Mean CI

Agriculture 39.4 9.3 37.0 5.8 22.98 5.8 29.8 9.0

Agriculture crop 38.0 9.2 35.2 5.4 20.04 5.6 28.0 8.8

Agriculture labor 1.3 1.3 1.7 1.2 2.94 2.0 1.8 2.2

Forestry 37.7 7.2 37.4 7.1 59.07 7.6 32.4 8.1

Plant 23.5 7.9 19.2 3.5 31.66 5.8 12.4 3.9

Timber 16.9 7.6 3.8 1.9 4.91 2.8 1.3 2.4

Gross NTFPs 6.6 2.5 15.4 2.9 26.74 5.6 11.2 3.5

Bamboo 0.7 0.6 4.7 1.7 0.63 0.2 3.3 2.4

Fuelwood 4.8 2.2 9.5 2.0 16.34 4.4 7.8 2.7

Other NTFPs 1.0 0.6 1.2 0.7 9.78 3.7 0.1 0.1

Forest management 0 0 3.5 3.6 2.21 3.7 3.5 4.5

Animal 14.1 5.8 14.7 5.8 25.2 7.4 16.4 6.4

Mammals 10.3 4.4 10.7 4.5 15.79 4.6 11.9 5.5

Birds 3.1 1.4 2.9 1.6 5.81 2.7 1.1 0.5

Other animals 0.7 0.6 1.1 1.1 3.6 2.8 3.4 2.4

Other employment opportunity 23.1 8.5 25.6 8.7 17.95 7.2 37.8 9.6

Government employment 16.4 7.9 18.4 7.7 12.43 5.3 21.0 9.3

Self employment 6.7 4.9 7.2 5.8 5.52 6.1 16.8 6.6


income levels and dependencies on the forestrysector reported earlier in India or elsewhere in theworld. This is because in their estimates, previousworkers (e.g. Malhotra et al., 1991; Bahuguna,1993, 2000; Hedge et al., 1996; Sethi & Hilaluddin,2001) have overlooked incomes derived from wildmeat contributing significantly to local economies.We thus feel previous estimates are incompleteand that previous appraisals should be revised

Our analysis on occupation status vis–à–viswildmeat extraction suggests that all sections of thesociety: be they custodian of the law or farmer orbusinessman, remove animal biomass equally. It alsoseems that an increase in education among the Mizodecreased the amount of wild meat extraction. With ahigher level of education, people have access tobetter jobs, and this in turn presumably leaves themwith little time to hunt. However, in certain communi-ties such as the Angami, increased cash incomesfrom vocations other than wild meat resulted in higherextraction of wild meat. It is likely that an improve-ment in financial status of a household also increasesthe desire to consume more. Therefore, policies link-ing poverty alleviation programs with the conserva-tion of natural resources should be drafted with ut-most care. Policies linking extraction of wild meat toalleviate poverty with conservation of natural resourcesrequire major review.

Acknowledgements

Our sincere thanks are due to Mr. James Goodhart,who provided financial assistance for fieldwork.Drs. Claudia Ruttee, John Carroll, Michael Conroy,Peter Garson, Philip McGowan, Francesc Uribe,Francisca Castro, Ghazala Shabuddin and IndraniChowdhary commented on draft manuscript andmade several useful suggestions. We are gratefulfor their efforts and concern! We also thank ourrespondents for their tremendous hospitality duringfieldwork and also for sharing their views openlyduring interviews.

References

Abernethy, K., Baker, M., Banett, L. E., Boclmer,R., Brashanes, J., Cowlishaw, G., Elkan, P.,Eves, H., Fa, J. E., Milner–Gulland, E. J., Peres,C., Roberts, C, Robinson, J. G., Rowcliffe, M. &Wilkie, D., 2003. Wild meat: the bigger picture.Trends in Ecology and Evolution, 18: 351–357.

Ali, S. & Ripley, S. D., 1987. A Handbook of Birds ofIndia and Pakistan: compact edition. Oxford Uni-versity Press. New Delhi.

Anon, 2003. Indian Wildlife Protection Act 1972 (asamended 2002). Natraj Publishers. DehraDun,Uttranchal, India.

Bahuguna, V. K., 1993. Forestry in Ecodevelopment.Indian Institute of Forest Management. Bhopal.

– 2000. Forests in the economy of rural poor: anestimation of the dependency level. Ambio, 29:

126–129.Barker, C., 1998. Forest resource scarcity and social

conflicts in Indonesia. Environment, 40: 4–37.Bennett, E. L., 2002. Is there a link between wild

meat and food security? Conservation Biology,16: 590–592.

Birdlife International, 2000. Threatened Birds of theWorld. Clark, Birdlife International, Cambridge,U.K.

Dauvergene, P., 1997. Shadows in the Forest: Ja-pan and the Politics of timber in Southeast Asia.MIT Press. Boston, U.S.A.

Dei, G. J. S., 1989. Hunting and gathering in aGhanaian rain forest community. Ecology, Foodand Nutrition, 22: 225–243.

Fa, J. E., Peres, C. A. & Meeuwig, J., 2002.Bushmeat exploitation in tropical forests: an In-ternational comparison. Conservation Biology, 16:232–237.

FSI, 2003. The State of Forest 2003. Forest Surveyof India. DehraDun, Uttranchal, India.

Hedge, R., Suryaprakash, S., Achoth, L. & Bawa K.S., 1996. Extraction of non–timber forest prod-ucts in the forests of Biligiri Rangan Hills, India:contribution to rural income. Economic Botany,50: 243–251.

Hilaluddin, 2005a. Illicit staple. Down to Earth, 13:50–52.

– 2005b. Wild meat boom and wildlife bust inNortheast. Kashmir Times dated 22 January.

Hilaluddin, Kaul, R. & Ghose, D. (In press a).Galliformes extraction and use by indigenouspeople of Northeast India. In: proceedings of 3rd

International Galliformes Symposium (R. A. Fuller& S. Browne Eds.). World Pheasant Associa-tion–International, U.K.

Hilaluddin, Kaul, R., Pradhan, S. & Taylor, J. (Inpress b). Conservation significance of wildmeatexploitation and use in the north Bengal Himalaya,India. Conservation & Society.

IUCN, 2003. 2003 Red List of Threatened Species.IUCN, Gland, Switzerland.

Kaul, R., Hilaluddin, Jandrotia, J. S. & McGowan,P. J. K., 2004. Hunting of large mammals andpheasants in the western Indian Himalaya. Oryx,9: 426–431.

Kolk, A., 1996. Forests in International Environmen-tal Politics: International Organizations, NGOsand the Brazilian Amazon. International Utrecht.

Kuznets, S., 1955. Economic growth and incomeinequity. American Economic Review, 445: 1–28.

Malhotra, K. C., Deb, D., Datta, M., Vasulu, T. S.,Yadav, G. & Adhikari, M., 1991. Role of non–timber forest produces in village economy: ahousehold survey in Jamboi Range, Midnapore,West Bengal. Institute of Socio–biological Re-search and Development, Calcutta, India, Un-published Report. 192 pp.

McGowan, P. J. K., Duckworth, W., Xianji, W.,Balen, S. W., Xiaojun, Y., Kahn, M., Yatim, H.,Thanga, L., Setiawan, I. & Kaul, R., 1998. Areview of the status of Green peafowl Pavomuticus and recommendations for future action.


Bird Conservation International, 8: 331–341.Nelleman, C. & Newton, A., 2002. The Great Apes:

The Road Ahead. UNEP, Gland, Switzerland.Peres, C. A., 2000. Effects of subsistence hunting

on vertebrate community structure in Amazonianforests. Conservation Biology, 14: 240–253.

Prater, S. H., 1971. The Book of Indian Animals.Oxford Univ. Press, New Delhi.

Robinson, J. G. & Redford, K. H., 1991. NewTropical Wildlife Use and Conservation. ChicagoUniversity Press, Chicago.

Sankaran, V., Hilaluddin, R. K. & Bandhopadhyay,P., 2000. The process of participatory mappingfor biodiversity conservation. In: SettingBiodiversity Priorities for India: 34–56 (S. Singh,A. R. K. Shastri, R. Mehta & V. Uppal, Eds.).WWF–India, New Delhi.

Sethi, P. & Hilaluddin, 2001. Structuring financialempowerment for localized development withJoint Forest Management (JFM): examples fromMadhya Pradesh, India. Sustainable Develop-ment, 9: 87–102.

Sokal, R. R. & Rohlf, F. J., 1995. Biometry: Princi-pals and Practice of Statistics in Biological Re-search. W.H. Freeman and Company. New York,U.S.A.

Sutherland, W. J., 1996. Why census? In Ecologi-cal census techniques: a handbook: 1–9 (W. J.Sutherland, Eds.). Cambridge Univ. Press. Cam-bridge, U.K.

Wilkie, D. S. & Carpenter, J. F., 1999. Bushmeathunting in the Congo Basin: an assessment ofimpacts and options for mitigations. BiodiversityConservation, 8, 929–955.



Örstan, A., Pearce, T. A. & Welter–Schultes, F., 2005. Land snail diversity in a threatened limestone districtnear Istanbul, Turkey. Animal Biodiversity and Conservation, 28.2: 181–188.

AbstractLand snail diversity in a threatened limestone district near Istanbul,Turkey.— The limestone meadowslocated to the north–northwest of Istanbul, Turkey, are in danger of being overrun by the rapidly expandingcity. Past surveys showed that these habitats harbor rare plant species, including endemics to Turkey. Tofurther evaluate the conservation value of these habitats, especially in terms of the often neglectedinvertebrates, one limestone area to the north of Küçükçekmece Lake and surrounding Sazlidere Dam wassurveyed for land snails. Our findings strengthen the case for the protection of these unique habitats.Twenty–four species of land snails were identified in the survey area. Of these, 21 are native to Turkey,including three whose type location is Istanbul. In addition, two species that are at or near the limits of theirranges are considered to represent peripheral populations that may be especially worth conserving.Although the area surrounding Sazlidere Dam is under protection, the other limestone habitats are severelythreatened by ongoing development.

Key words: Biodiversity, Conservation, Istanbul, Pulmonata, Prosobranchia.

ResumenDiversidad de los caracoles terrestres en una zona caliza amenazada cercana a Estambul, Turquía.— Laspraderas calcáreas situadas al NNO de Estambul están en peligro de ser rápidamente invadidas por laciudad en expansión. Estudios anteriores demostraron que estos hábitats albergan especies vegetalesraras, incluyendo algunos endemismos turcos. Con objeto de seguir evaluando el valor conservativo dedichos hábitats, en especial en cuanto a los invertebrados, a menudo ignorados, se han estudiado loscaracoles terrestres de una zona calcárea al norte del lago Küçükçekmece y alrededor de la presaSazlidere. Nuestros descubrimientos enfatizan la necesidad de una política de protección de estos hábitatsúnicos. En el área estudiada se identificaron 24 especies de caracoles terrestres. De ellas, 21 son nativasde Turquía, incluyendo tres cuya localización tipo es Estambul. Además, se considera que dos especies quese hallan en o cerca de los límites de su zona de distribución representan poblaciones periféricasespecialmente merecedoras de conservación. A pesar de que la zona que rodea a la presa Sazlidere estáprotegida, el resto de los hábitats calcáreos está muy amenazado por el creciente desarrollo.

Palabras clave: Biodiversidad, Conservación, Estambul, Pulmonata, Prosobranchia.

(Received: 4 V 05; Conditional acceptance: 21VI 05; Final acceptance: 8 VII 05)

Aydin Örstan, Timothy A. Pearce, Section of Mollusks, Carnegie Museum of Natural History, 4400Forbes Ave., Pittsburgh, PA, 15213, U.S.A.– Francisco Welter–Schultes, Zoologisches Inst., BerlinerStr. 28, D–37073, Goettingen, Germany.

Land snail diversityin a threatened limestone districtnear Istanbul, Turkey

A. Örstan, T. A. Pearce & F. Welter–Schultes

182 Örstan et al.

hills surrounding Sazlidere Dam (fig. 1). The areawas not forested, but consisted of meadows andgrassy hills with limestone outcrops. The surveywas conducted on two days (26 VI and 8 VII 2004).Twelve stations, scattered along an approximately12.5–km long transect, were designated in the field(fig. 1), but during the analysis of the results threeof the stations (C4, C5 and C6) that were locatedwithin less than 100 m of each other were treatedas one. The UTM coordinates (for zone 35) andelevation of each station were measured with aGPS receiver with an accuracy of about 10 m.

The following list gives the description and coor-dinates of each station (fig. 1): C1. Limestone cliffabove road to Ôamlar Village; UTM E646269 m,N4548620 m; elevation 20 m. C2. Northeast bankof Sazli Creek; UTM E645861 m, N4548704 m;elevation 0 m; C3. Rocky hill northeast of SazliCreek; UTM E645575 m, N4550187 m; elevation70 m; C4–C6. Rocky hill south of road to SazlidereDam; UTM E645578 m, N4550370 m; elevation55 m; C7. Limestone rocks near unpaved road,north of Ôamlar Village; UTM E646246 m,N4554293 m; elevation 100 m; C8. Limestone rockson hillside above Sazlidere Dam lake; UTME645082 m, N4553385 m; elevation 30 m; C9.Limestone rocks along grassy field; UTME643946 m, N4553427 m; elevation 25 m; C10.Limestone rocks on hillside above unpaved road,north of old limestone quarry. UTM E641977 m,N4555331 m; elevation 25 m; C11. Grassy fieldalong road to Hadimköy, west of roadway acrossSazlidere Dam lake; UTM E638674 m, N4557811 m;elevation 25 m; C12. Limestone rocks on hillsidefacing a residential district, north of KüçükçekmeceLake; UTM E646015 m, N4547613 m; elevation 65 m.

Surface collections were done at each station bytwo or three people. In addition, soil samples weretaken from six stations, sieved and sorted for smallshells. The identifications of Oxyloma elegans,Monacha ocellata, Monacha solidior, Xerolenta obvia,Xeropicta krynickii and Cernuella virgata were con-firmed by dissection. Sixty–one lots (537 speci-mens), including at least one lot of every land snailspecies found in the study area (excluding Eobaniavermiculata), have been deposited with the CarnegieMuseum of Natural History, Pittsburgh, PA, U.S.A.(CM 70300–70357, 70762–70764). Additional lotsare in the collection of the first author. Referencesamples of Albinaria caerulea were obtained fromthe Field Museum of Natural History (FM), Chi-cago, U.S.A.

Results

We found 24 species of land snails in the survey area,representing 12 families (table 1). In addition, slugshells were collected at stations C1 and C9, but thesecould be identified only to the family level (Reuse,1983). A live, dormant Eobania vermiculata was seenat station C1 but not taken. Fourteen species (58% oftotal) were found only as empty shells (table 1) and

Introduction

Unprotected wildlife habitats located near expandingresidential or industrial centers are subject to rapidand permanent destruction. Especially in developingcountries, unique habitats near growing cities maybe destroyed before they are properly surveyed andmeasures implemented for their protection. One casein point is the city of Istanbul, whose population hasgrown from about 800,000 in 1927 to 10 million in2000 (Istanbul Metropolitan Municipality, 2005). Priorto the 20th century, the then much smaller city ofIstanbul was surrounded by villages separated fromeach other by more or less degraded, but neverthe-less uninhabited and undeveloped land that includedagricultural fields and orchards (Anonymous, 1844).Such areas must have provided habitats for at leastsome of the native wildlife. However, since the be-ginning of the 20th century, the rapidly expandingIstanbul has absorbed most of the villages, turningthem into districts within the city and, in the process,has all but decimated the wildlife habitats.

Özhatay et al. (2003) recently brought attention toseveral threatened unique habitats (designated asImportant Plant Areas) surrounding Istanbul that arerich in rare and endemic plant species. One of theseImportant Plant Areas is the limestone meadows tothe north–northwest of Istanbul. Özhatay et al. (2003)singled out three remaining fragments of these mead-ows in the region extending west from the vicinity ofthe town of GaziosmanpaÕa north of Istanbul to thenorth of the Küçükçekmece Lake and identified 19rare plant taxa, including seven that are endemic toTurkey, growing on these meadows. The westernmostof these fragments is located along Sazli Creek(Sazli Dere) that empties into Küçükçekmece Lake(fig. 1). To supply drinking water for Istanbul, SazlidereDam was recently built over this creek, partly flood-ing the creek’s broad valley.

Istanbul and its environs are the type locations ofabout 10 species of land snails that were describedmostly in the 19th century (Schütt, 2001). Unfortu-nately, because of the ongoing loss of land to devel-opment, it has now become difficult to find habitatsin and around the city that still maintain the originalnative land snail fauna. Our attention was, therefore,attracted to the Sazli Creek area not only because ofthe presence of limestone, which generally supportsa high diversity of land snails, but also because ofthe threatened status of this habitat type in theIstanbul area. Furthermore, the area had not beenproperly surveyed for land snails. Therefore, to fur-ther evaluate the conservation value of the limestonemeadows, especially in terms of the diversity of theirfauna of native land snails, we conducted a landsnail survey of the Sazlidere Dam area.

Materials and methods

The study area, located west of the city of Istanbul,extended from the north of Küçükçekmece Lake upthe broad valley of Sazli Creek to the low limestone


two species (8% of total) only as live specimens. Weconsider 21 of the land snail species to be native tothe survey area. Cochlicella acuta, Cernuella virgataand Eobania vermiculata are non–native species thathave been introduced to Turkey.

Additional notes on some of the species

Pomatias elegansThe earliest record of P. elegans from around Istan-bul dates to the 19th century (Sturany, 1894). Thisspecies strictly requires calcareous substrates (Boy-cott, 1934) and we found it quite abundantly atsome of our stations.

Oxyloma elegansTwo individuals were found crawling on plants grow-ing in Sazli Creek at station C2. Oxyloma elegans isfound throughout Europe (Hecker, 1965; Kerney &Cameron, 1979) and has been recorded in Turkeybefore (Schütt, 2001). However, our record of thisspecies appears to be the first for the Istanbul area.Oxyloma elegans cannot always be reliably sepa-rated from O. sarsii (Esmark, 1886) by shell char-

acteristics (Kerney & Cameron, 1979). Therefore,the previous records of O. elegans and O. sarsiifrom Turkey that were not confirmed by dissectionmay not be reliable.

Pupilla cf. sterriiThe Pupilla shells from stations C1 and C12 werefinely striated and their apertures had three teeth: aparietal, a palatal and a deep columellar that wasvisible when the aperture was turned slightly side-ways (fig. 2). The sample of nine adult shells fromC1 had a mean length of 2.83 mm and a meandiameter of 1.57 mm. To identify these specimens,we considered two species: P. triplicata (Studer,1820) and P. sterrii. Although P. sterrii is stated tohave usually two teeth, a parietal and a palatal(Kerney & Cameron, 1979; Falkner, 1990), we iden-tified our specimens tentatively as P. sterrii ratherthan P. triplicata, because the microsculpture of ourshells agreed better with that of P. sterrii and thedimensions of our specimens were closer to thoseof P. sterrii and slightly above the ranges, especiallyfor diameters, of those of P. triplicata (Germain,1930; Kerney & Cameron, 1979).

Fig. 1. The survey site showing the locations of the collection stations. The inset shows the locationof the survey site in relation to the metropolitan Istanbul (shaded area). The background satellitephotograph (file name: ISS008–E–21753.jpg, taken on 16 IV 2004) was downloaded from http://eol.jsc.nasa.gov (image courtesy of the Image Analysis Laboratory, NASA Johnson Space Center).

Fig. 1. Lugar del estudio mostrando la situación de las estaciones de recolección. El recuadro muestra lalocalización del lugar de estudio en relación con el área metropolitana de Estambul (área sombreada). Lafotografía de satélite del fondo (nombre del archivo: ISS008–E–21753.jpg, tomada el 16 IV 2004) se bajóde http://eol.jsc.nasa.gov (por cortesía del Image Analysis Laboratory, NASA Johnson Space Center).

Sea of Marmara

N

5 km

KüçükçekmeceLake

SazliCreek

SazlidereDam

Istanbul

Sea of Marmara

Istanbul

Surveyarea

Black Sea

http://eol.jsc.nasa.gov

184 Örstan et al.

Multidentula ovularisThis species was not previously recorded from theprovince of Istanbul (Forcart, 1940; Schütt, 2001).We found it in abundance at our station C3 andless abundantly at three other stations.

Albinaria caeruleaThis strictly calciphilic species was abundant atseveral of our stations. Several subspecies of A.caerulea are distributed along the coastal regionsof southwestern Turkey (Örstan, 2001), on the Greekislands near mainland Turkey (Nordsieck, 1977;Zilch, 1977) and in Attiki, Greece (Giokas & Mylonas,2002). In addition, Schütt (2001) gave a record of

A. caerulea from Çatalca, about 18 km west of oursurvey area. We compared our specimens withsamples of A. caerulea caerulea from the island ofChios (FM 206645), A. caerulea milleri (Pfeiffer,1850) from the island of Delos (FM 206815), and A.caerulea maculata (Rossmassler, 1836) and A.caerulea calcarea (Boettger, 1878) from the vicinityof Ephesus, Turkey (Örstan, private collection).Using these comparisons, we determined that ourspecimens were conchologically closest to the nomi-nal subspecies. We note that our specimens werealso identical to the Field Museum lot of A. caeruleacaerulea (FM 161499) with the collection locationgiven broadly as "Thracien, Istanbul".

Table 1. The land snail species collected in the survey area and the stations where each was found:* Species that were found only as empty shells. (For information on stations see the text, Materialand methods, and fig. 1.)

Tabla 1. Especies de caracoles terrestres recogidas en la zona de estudio y estaciones donde sehallaron. Los asteriscos indican las especies de las que sólo se encontraron conchas vacías. (Parainformación sobre las estaciones ver el texto, Material and methods, y la fig. 1.)

Snail species Stations

Pomatias elegans* (Müller, 1774) C3, C4–C6, C7, C8, C9, C12

Oxyloma elegans (Risso, 1826) C2

Truncatellina cylindrica* (Férussac, 1807) C4–C6

Granopupa granum* (Draparnaud, 1801) C1, C4–C6, C8, C12

Pupilla cf. sterrii* (Voith, 1838) C1, C12

Pleurodiscus balmei* (Potiez and Michaud, 1838) C1, C9

Chondrus tournefortianus* (Férussac, 1821) C3, C8

Multidentula ovularis* (Olivier, 1801) C1, C3, C4–C6, C8

Mastus carneolus (Mousson, 1863) C1, C3, C4–C6, C7, C8, C9, C10, C11, C12

Oxychilus hydatinus* (Rossmässler, 1838) C1, C3, C4–C6

Cecilioides acicula* (Müller, 1774) C4–C6, C10

Albinaria caerulea (Deshayes, 1835) C1, C3, C4–C6, C7, C8, C12

Bulgarica denticulata* (Olivier, 1801) C1, C3, C4–C6, C8, C9, C12

Cochlicella acuta* (Müller, 1774) C1, C7, C10

Trochoidea pyramidata (Draparnaud, 1805) C1, C3, C4–C6, C7, C8, C9, C10, C12

Monacha claustralis* (Mousson, 1859) C10, C11

Monacha ocellata (Roth, 1839) C1, C8, C9, C11

Monacha solidior (Mousson, 1863) C3, C7, C8, C9

Xerolenta obvia (Menke, 1828) C1, C3, C4–C6, C8, C9, C10, C12

Xeropicta krynickii (Krynicki, 1833) C1, C2, C3, C4–C6, C7, C8, C9, C10, C11, C12

Cernuella virgata (Da Costa, 1778) C1, C2, C3, C7, C8, C9, C10, C11, C12

Eobania vermiculata (Müller, 1774) C1

Helix lucorum* Linnaeus, 1758 C1, C8, C9

Helix pomacella* Mousson, 1854 C4–C6, C11

Limacidae C1, C9


Discussion

We are unaware of any previously published surveyof the land snails of the Sazlidere Dam area. How-ever, we found published records for four species ofland snails from the vicinity: C. acuta fromKüçükçekmece (Sturany, 1902), M. carneolus fromYeÕilköy (previously San Stefano) and Florya, dis-tricts southeast of Küçükçekmece Lake (Sturany,1902; Gittenberger, 1967) and O. hydatinus and A.caerulea from Çatalca west of our survey area(Riedel, 1995; Schütt, 2001). We also found thesefour species in our survey (table 1). Sturany’s 1902record of C. acuta from Küçükçekmece indicatesthat this introduced species has been in the area formore than 100 years.

In the survey area we saw A. caerulea aestivat-ing attached to limestone rocks. There is evidencethat Albinaria species that aestivate on rocks haveoccasionally been transported to areas outside oftheir native ranges by humans on rocks intendedfor buildings or other purposes (Welter–Schultes,1998). Therefore, we considered the possibility thatA. caerulea was introduced to our survey area onrocks that were brought from elsewhere, perhapssouthwestern Turkey. However, because of the rela-tively slow dispersal rate (~2 m/year) of Albinariaspecies (Schilthuizen & Lombaerts, 1994), uninten-tional introductions by humans usually result inlocalized distributions of the introduced species

(Welter–Schultes, 1998). In comparison, the dis-tance between the two farthest stations where wefound A. caerulea, about 6.7 km, indicates that thedistribution of the species in our survey area wasnot localized. We also note that Schütt’s (2001)record of A. caerulea from Çatalca, west of oursurvey area, suggests that the range of this speciesextends over an even wider area. Moreover, there isno evidence that large quantities of calcareousrocks were transported into the Sazli Creek area inthe past (such materials are rarely used in modernbuildings); the limestone quarry near station C10that was in operation until recently indicates thatlimestone was actually exported from the area.Therefore, these arguments lead us to concludethat A. caerulea is native to the survey area.

In addition to the previously published recordslisted above, Chalcolithic fossils of Helix pomatiaLinnaeus, 1758 (from a layer ~6880 radiocarbonyears B.P.) were reported from the YarimburgazCave within our survey area (Meriç et al., 1991).However, because H. pomatia is not an extantspecies in Turkey (Schütt, 2001; Yildirim et al.,2004), we suggest that the specimens Meriç et al.(1991) reported as H. pomatia are probably theconchologically similar H. lucorum, which we foundat stations C1, C8 and C9. Nevertheless, we notethat the present day range of H. pomatia extendsfrom northern Europe through the Balkan Peninsuladown to Macedonia (Falkner, 1990) and that duringthe Chalcolithic period the species may have livedas far south as our survey area or may have beentaken there by humans.

We found only empty shells of 14 species (58%of total) and only live specimens of two species (8%of total). These results, obtained during the dryseason in the Istanbul area, are comparable to theresults Rundell & Cowie (2003) obtained in Hawai-ian dry forests, where 40 to 47% of species werecollected dead only and 0 to 7% live only. AsRundell & Cowie (2003) pointed out, if a surveyproduces only empty shells of a species, this resultmay imply that the species is either very rare orextinct at that location. However, since most snailspecies hide deep in the soil or under rocks duringthe dry season, we believe that we would havefound live specimens of most, if not all, of therecorded species if we had collected during a rainyperiod, or if we had searched more intensely. Weconsider our results as constituting a baseline andwe believe that only by conducting follow–up sur-veys of the area in the future will it be possible toaccurately monitor any changes in the extant landsnail fauna.

Özhatay et al. (2003) based their arguments forthe conservation value of the last remaining lime-stone meadows in the Istanbul area on their botani-cal richness. The results of our survey add fouradditional justifications for the protection of thesehabitats. First, the majority of the land snail speciesfound in the survey area (21 out of 24) are native toTurkey. We believe that the protection of the fewremaining undeveloped areas in and around Istan-

Fig. 2. A specimen of Pupilla cf. sterrii(2.7 x 1.5 mm) from station C1. Arrows pointat the teeth.

Fig. 2. Ejemplar de Pupilla cf. sterri i(2,7 x 1,5 mm) de la estación C1. Las flechasseñalan los dientes.

186 Örstan et al.

bul is urgently necessary to provide habitats for thesnail species that are native to the region.

Second, Istanbul is the type location of three ofthe species that we found: M. carneolus, M. ocellataand H. pomacella. For taxonomical reasons, wethink that it is very important to protect the typelocations of animal and plant taxa. If in the futurespecimens are needed for critical taxonomical com-parisons, for example for genetic or anatomicalanalyses, the most suitable place to get speci-mens that are likely to be the same species as theone that was described from the type locationwould be the type location itself. Although theSazlidere Dam area was not specified as the typelocation of any of these species, we believe thatthe area is close enough to Istanbul to be consid-ered within the type location. Therefore, consider-ing that the original type locations within Istanbulare likely to have been destroyed by now, theprotection of the nearby areas with the samespecies would ensure the survival of thesetaxonomically important populations.

Third, some of the species we found, for exam-ple, P. elegans and A. caerulea, are strict calciphilesand would not survive on noncalcerous substrates.The strict dependence of these species on calcare-ous rocks and soil underlines the need to protecttheir limestone habitats.

Fourth, in our survey area P. elegans and A.caerulea may be at or near the limits of theirdistribution ranges. The range of P. elegans ex-tends from England and an isolated spot in western

Fig. 3. Residential development that is gradually overrunning the limestone hills. The photograph wastaken on 26 VI 2004 from station C6 (the rocks in the foreground) looking west across the broad valleyof Sazli Creek.

Fig. 3. Desarrollo urbano que gradualmente va invadiendo las colinas calizas. La fotografia se tomó el26 VI 2004 desde la estación C6 (rocas del primer plano) mirando hacia el oeste a través del anchovalle del riachuelo Sazli.

Ireland to northwestern Turkey (Kerney & Cameron,1979; Örstan, 2005). The Istanbul area may be ator close to the southeastern limit of the range ofthis species (Örstan, 2005). As for A. caerulea, allof its other known populations are from southwest-ern Turkey (Örstan, 2001), the adjacent Greek is-lands (Nordsieck, 1977; Zilch, 1977) and southernGreece (Giokas & Mylonas, 2002), so the Istanbularea certainly represents the northernmost limit ofits range. Since peripheral populations are oftengenetically and morphologically divergent from cen-tral populations, and since genetically divergentpopulations are valuable as potential sites of futurespeciation events (Mayr, 1970; Lesica & Allendorf,1995), peripheral populations are important candi-dates for conservation (Lesica & Allendorf, 1995).At least until further studies have been carried outto evaluate the degree of genetic divergence fromcentral populations of the peripheral land snailcolonies around Istanbul, the habitats of peripheralland snail colonies should be protected.

Özhatay et al. (2003) classified the conservationneeds of these limestone meadows as "very ur-gent". Our observations during the survey supportthe determination of Özhatay et al. (2003) that thelimestone meadows north–northwest of Istanbulare under imminent threat from the expansion ofthe nearby residential neighborhoods. For example,our station C1 was on a cliff below a denselypopulated hilltop, while C12 was less than 100 mfrom recently constructed apartment buildings. Theovertaking of the latter station by further develop-


ment is probably only a matter of time. A photo-graph taken from station C6 (fig. 3) shows theextent of encroaching development and illustratesthe general threat these habitats are facing.

On the other hand, the presence of a dam withinour survey area that was built to supply water forthe ever–growing population of Istanbul paradoxi-cally offers some protection to the surrounding land(Özhatay et al., 2003). The regulations of the mu-nicipal agency that administers the dam, IstanbulSu ve Kanalizasyon Idaresi (ISKI; Istanbul Waterand Sewer Administration), prohibit all develop-ment, other than those associated with water puri-fication, within 1000 m of water reservoirs if thewater collection basin in question extends at leastthat far (ISKI, 2004). Therefore, if these regulationsare enforced as intended, the presently undevel-oped land within 1,000 m of the dam lake may beconsidered to be under protection for the timebeing. However, the same regulations do allow forresidential buildings outside the 1,000–m limit, whichmeans that our stations C1, C2, C3 and C12 andthe surrounding areas located up to about 4 kmaway from Sazlidere Dam may be lost in the futureunless protected.

Turkey currently has a number of national parksand various types of nature conservation areas(Kaya & Raynal, 2001; Guclu & Karahan, 2004).The protection provided by such areas to all wild-life notwithstanding, the establishment of parksand conservation areas is usually justified in termsof the protection they will offer to mostly largemammals, birds and plants (Yilmaz, 1998; Kaya &Raynal, 2001; Guclu & Karahan, 2004), while theconservation needs of invertebrates are almostnever taken into consideration. The land snailfaunas in many countries are increasingly beingthreatened with extinction (Lydeard et al., 2004).Turkey has a rich land snail fauna with manyendemic species (Schütt, 2001) that, in our opin-ion, deserve no less protection than any otheranimal or plant group. We hope that our resultsregarding these threatened limestone meadowswill bring attention to the conservation needs ofthe native land snails in particular and inverte-brates in general.

Acknowledgements

We thank Teri Varnali for driving us to and aroundthe survey area, Bernhard Hausdorf for his helpwith some of the identifications and Zeki Yildirim forcomments that improved the manuscript.

References

Anonymous, 1844. Map of Constantinople,Stambool. In: Maps of the Society for the Diffu-sion of Useful Knowledge, vol. 1. Chapman andHall, London.Available from http://www.davidrumsey.com

Boycott, A. E., 1934. The habitats of land Molluscain Britain. Journal of Ecology, 22: 1–38.

Falkner, G., 1990. Binnenmollusken. In: Weichtiere.Europäische Meeres– und Binnenmollusken (R.Fechter & G. Falkner, Eds.). Mosaik, München.

Forcart, L., 1940. Monographie der türkischenEnidae. Verhandlungen der NaturforschendenGesellschaft, 51: 106–263.

Germain, L., 1930. Mollusques Terrestres etFluviatiles. Faune de France, vol. 21. Lechevalier,Paris.

Giokas, S. & Mylonas, M., 2002. Spatial distribu-tion, density and life history in four Albinariaspecies (Gastropoda, Pulmonata, Clausiliidae).Malacologia, 44: 33–46.

Gittenberger, E., 1967. Die Enidae (Gastropoda,Pulmonata) gesammelt von der niederländischenbiologischen Expedition in die Türkei in 1959.Zoologische Mededelingen, 42: 125–141.

Guclu, K. & Karahan, F., 2004. A review: the historyof conservation programs and development ofnational parks concept in Turkey. Biodiversityand Conservation, 13: 1373–1390.

Hecker, U., 1965. Zur Kenntnis der mitteleuro-paeischen Bernsteinschnecken (Succineidae).I. Archive für Molluskenkunde, 94: 1–45.

Istanbul Metropolitan Municipality, 2005. Thepopulation of Istanbul according to the censusyears.Available from http://www.ibb.gov.tr/IBB/DocLib/pdf/bilgihizmetleri/yayinlar/istatistikler/demografi/t211.pdf

Istanbul Su ve Kanalizasyon Idaresi (ISKI), 2004.Içmesuyu Havzalari Koruma ve Kontrol Yönet-meligi [Drinking Water Basins Protection andControl Regulations].Available from http://www.iski.gov.tr/dosya/yonetmelikler/ICHKOYNT.pdf

Kaya, Z. & Raynal, D. J., 2001. Biodiversity andconservation of Turkish forests. Biological Con-servation, 97: 131–141.

Kerney, M. P. & Cameron, R. A. D., 1979. A FieldGuide to the Land Snails of Britain and North–west Europe. Collins, London.

Lesica, P. & Allendorf, F. W., 1995. When areperipheral populations valuable for conserva-tion? Conservation Biology, 9: 753–760.

Lydeard, C., Cowie, R. H., Ponder, W. F., Bogan, A.E., Bouchet, P., Clark, S. A., Cummings, K. S.,Frest, T. J., Gargominy, O., Herbert, D. G.,Hershler, R., Perez, K. E., Roth, B., Seddon, M.,Strong, E. E. & Thompson, F. G., 2004. Theglobal decline of nonmarine mollusks. BioScience,54: 321–330.

Mayr, E., 1970. Populations, Species, and Evolu-tion. Harvard Univ. Press, Cambridge.

Meriç, E., Sakinç, M., Özdogan, M. & Açkurt, F.,1991. Mollusc shells found at the YarimburgazCave. Journal of Islamic Academy of Sciences,4: 6–9.

Nordsieck, H., 1977. Taxonomische Revision desGenus Albinaria. Archive für Molluskenkunde,107: 285–307.

http://www.davidrumsey.com

http://www.ibb.gov.tr/IBB/DocLib/pdf/bilgihizmetleri/yayinlar/istatistikler/demografi/t211.pdf



http://www.iski.gov.tr/dosya/yonetmelikler/ICHKOYNT.pdf

http://www.iski.gov.tr/dosya/yonetmelikler/ICHKOYNT.pdf

188 Örstan et al.

Schütt, H., 2001. Die türkischen Landschnecken.Acta Biologica Benrodis, Supplementband 4:1–549.

Sturany, R., 1894. Zur Molluskenfauna dereuropäischen Türkei. Annalen des K. K.Naturhistorischen Hofmuseums Wien, 9:369–390.

– 1902. Beitrag zur Kenntnis der kleinasiatischenMolluskenfauna. Sitzungsberichte derMathematisch–Naturwissenschaftlichen Classeder Kaiserlichen Akademie der WissenschaftenWien, 111: 123–140.

Welter–Schultes, F., 1998. Human–dispersed landsnails in Crete, with special reference to Albinaria(Gastropoda: Clausiliidae). Biologica Gallo–hellenica, 24: 83–106.

Yildirim, M. Z., Kebapçi, Ü. & GümüÕ, B. A., 2004.Edible Snails (Terrestrial) of Turkey. Turkish Jour-nal of Zoology, 28: 329–335.

Yilmaz, K. T., 1998. Ecological diversity of theEastern Mediterranean region of Turkey and itsconservation. Biodiversity and Conservation, 7:87–96.

Zilch, A., 1977. Die Typen und Typoide des Natur–Museums Senckenberg, 57: Mollusca: Clausiliidae.Archive für Molluskenkunde, 107: 309–363.

Örstan, A., 2001. A preliminary survey of Albinariapopulations around Kusadasi Bay, Turkey. Triton,No. 4: 42–44.

– 2005. The status of Pomatias elegans in Istan-bul, Turkey. Tentacle, No. 13: 8–9.Available from http://www.hawaii.edu/cowielab/Tentacle.htm

Özhatay, N., Byfield, A. & Atay, S., 2003. Türkiye’ninÖnemli Bitki Alanlari [Important Plant Areas inTurkey]. WWF Türkiye, Istanbul.

Reuse, C., 1983. On the taxonomic significance ofthe internal shell in the identification of Euro-pean slugs of the families Limacidae andMilacidae (Gastropoda, Pulmonata). BiologischJaarboek Dodonaea, 51: 180–200.

Riedel, A., 1995. Zonitidae sensu lato (Gastropoda,Stylommatophora) der türkei. Übersicht der Arten.Fragmenta Faunistica, 38: 1–86.

Rundell, R. J. & Cowie, R. H., 2003. Preservation ofspecies diversity and abundances in Pacific is-land land snail death assemblages. Journal ofConchology, 38: 155–163.

Schilthuizen, M. & Lombaerts, M., 1994. Populationstructure and levels of gene flow in the Mediter-ranean land snail Albinaria corrugata (Pulmonata:Clausiliidae). Evolution, 48: 577– 586.

http://www.hawaii.edu/cowielab/Tentacle.htm

http://www.hawaii.edu/cowielab/Tentacle.htm



Pradel, R., Gimenez, O. & Lebreton, J.–D., 2005. Principles and interest of GOF tests for multistatecapture–recapture models. Animal Biodiversity and Conservation, 28.2: 189–204.

AbstractPrinciples and interest of GOF tests for multistate capture–recapture models.— Optimal goodness–of–fitprocedures for multistate models are new. Drawing a parallel with the corresponding single–state proce-dures, we present their singularities and show how the overall test can be decomposed into interpretablecomponents. All theoretical developments are illustrated with an application to the now classical study ofmovements of Canada geese between wintering sites. Through this application, we exemplify how theinterpretable components give insight into the data, leading eventually to the choice of an appropriategeneral model but also sometimes to the invalidation of the multistate models as a whole. The method forcomputing a corrective overdispersion factor is then mentioned. We also take the opportunity to try todemystify some statistical notions like that of Minimal Sufficient Statistics by introducing them intuitively. Weconclude that these tests should be considered an important part of the analysis itself, contributing in waysthat the parametric modelling cannot always do to the understanding of the data.

Key words: Memory, Transients, Trap–dependence, Test WBWA, Contingency tables partitioning, Test M.

ResumenPrincipios e interés de los test Bondad de Ajuste (GOF) para los modelos de captura–recapturamultiestado.— Los procedimientos óptimos de bondad de ajuste, aplicados a los modelos multiestado,son nuevos. Trazando un paralelismo con los correspondientes procesos de uniestado, presentamos susparticularidades y mostramos como el test general puede descomponerse en componentes susceptiblesde ser interpretados. Todos los desarrollos teóricos están ilustrados con una aplicación del ya clásicoestudio de los desplazamientos de la barnacla canadiense entre sus lugares de invernada. Mediante estaaplicación, presentamos un ejemplo de cómo los componentes susceptibles de ser interpretados nosproporcionan una idea de los datos que nos pueden llevar a la elección de un modelo general apropiado,pero también a veces a la invalidación de los modelos de multiestados en su conjunto. Se mencionaentonces el método para calcular un factor de corrección de la sobredispersión. Aprovechamos estaocasión para intentar también desmitificar algunas nociones estadísticas, como las Estadísticas SuficientesMínimas, introduciéndolas intuitivamente. La conclusión es que estas pruebas deberían considerarse unaparte importante del propio análisis, contribuyendo a la comprensión de los datos, de un modo que elmodelaje paramétrico no siempre consigue.

Palabras clave: Memoria, Transeúntes, Dependencia del trampeo, Test WBWA, Partición de tablas decontingencia, Test M.

(Received: 8 VII 05; Final acceptance: 19 VII 05)

Roger Pradel, Olivier Gimenez & Jean–Dominique Lebreton, CEFE, CNRS, 1919 Route de Mende,34293 Montpellier cedex 5, France.

Principles and interestof GOF tests for multistatecapture–recapture models

R. Pradel, O. Gimenez & J.–D. Lebreton

190 Pradel et al.

1989)). Today, bootstrap procedures may be away around distributional problems. However, an-other weakness of the omnibus approach of com-paring expected and observed numbers is that itlacks power against specific alternatives and thatit is not informative when it rejects. Specializedtests have been built to address frequent causesof departure. Examples are the Leslie–Carotherstest of equal catchability (Carothers, 1971), theBrownie–Robson test of marking–induced deaths(Robson, 1969; Brownie & Robson, 1983), whichhas later been shown to test also for the presenceof transients (Pradel et al., 1997), and, in thecontext of multistate models, a test of memory(Pradel et al., 2003). However, the relationshipsbetween the particular tests will remain unknownuntil a careful study of the likelihood is carried out.Only such a study can provide the basis for asound partitioning of the information.

A major step in this direction was the develop-ment of optimal goodness–of–fit procedures forthe CJS model (Pollock et al., 1985). The globaltest, organized into several interpretable compo-nents and based on adequately pooled tables, wasimplemented in the RELEASE programme(Burnham et al., 1987). Since then, several spe-cialized tests have been shown to be componentsof this general test (the test for the presence oftransients (Pradel et al., 1997): that for trap–dependence (Pradel, 1993)) and a slightly differ-ent version of the general test is now proposed inprogram U–CARE (Choquet et al., 2005). Re-cently, Pradel et al. (2003) have developed asimilar approach for the multistate model calledJMV (Brownie et al., 1993), a model which gener-alizes the Arnason–Schwarz model by allowingencounter probabilities to vary by site occupied atthe previous occasion. The AS model, regarded asthe reference model for multistate capture–recap-ture, has not yet received a specific treatment.

The purpose of this paper is to review the princi-ples on which the goodness–of–fit tests for CJSand JMV are based, underlining their similaritiesand differences, and to examine how alternatives ofinterest can be embedded within the general tests.This paper is intended for the biologist with someexperience of capture–recapture analysis but nodeep statistical training. Thus, we assume that thereader knows what the CJS and the AS models are.On the other hand, we have tried to use everydaywords in place of statistical terms. For instance, weseek to introduce notions like minimal sufficientstatistics from a practical angle. Most of the paperis illustrated with one example, that of the Canadagoose data originally analyzed by Hestbeck et al.(1991). We proceed by steps. First, we present anddiscuss the features of the goodness–of–fit test ofthe simpler CJS model and specialized tests em-bedded within it. For the specialized tests, we ex-amine some statistics particularly suitable to ad-dress the alternatives of interest. The second sec-tion presents and discusses the goodness–of–fittest of the JMV model, drawing a parallel —as far

Introduction

Multistate capture–recapture models are very ap-pealing for studying a variety of biological ques-tions such as dispersal where states are geo-graphical sites (Hestbeck et al., 1991), trade–offbetween reproductive status and survival wherestates are breeder vs. non–breeder (Nichols et al.,1994), rate of accession to reproduction wherestates are experienced vs. inexperienced breeders(Pradel & Lebreton, 1999), etc. Furthermore, dif-ferent types of demographic information, such aslive recaptures and recoveries of dead individualsby the public can be analyzed simultaneouslyusing adequate multistate models (Lebreton et al.,1999). A general review of the biological relevanceof multistate capture–recapture models can befound in Lebreton & Pradel (2002). In multistatecapture–recapture models (Arnason, 1972, 1973;Hestbeck et al., 1991), marked individuals canmove among a finite number of states, or die,between discrete time occasions. Survivors aredetected ("encountered") in each state, not ex-haustively at each occasion. Based on parameterswhich are the transition, survival and encounterprobabilities, the probability of an individual en-counter history —conditional on the date and stateof first encounter, marking and release— can becalculated in a way similar to that used for theclassical one–state Cormack–Jolly–Seber (CJS)model. Under the assumption of independencebetween individuals, the likelihood for a particulardata set is then obtained as the product of theprobabilities for each individual encounter history.

The rationale of model selection, based on theAIC, assumes that the set of models consideredencompasses a model that fits the data (Burnham& Anderson, 1998). If not, the deviance will tend tobe inflated, favoring the incorrect selection ofoverparametrized models and thus leading to erro-neous biological conclusions. Moreover, the preci-sion of the final estimates will also be biased ifsome lack–of–fit or overdispersion is ignored. Theconsequences of lack–of–fit are thus too deleteri-ous to be ignored. Yet, difficulties with goodness–of–fit issues have been recurrent in the applicationof capture–recapture methodology. In a survey ofthe literature, Begon (1983) concluded that fewerthan 11% of the applications of the Jolly–Sebermodel addressed in a quantitative way or dis-cussed the assumptions inherent in the model.This state of fact was the consequence of theabsence at that time of any general goodness–of–fit procedure. The simplest approach, which con-sists of comparing observed and expected num-bers of animals with a particular encounter history,was hampered by the large number of encounterhistories (in the one–site case, with 10 occasions,there are more than 1,000 different encounterhistories), and as a consequence by the very lowexpected numbers (the resulting sparseness makes² distributions for quadratic X² statistics or for the

deviance quite inadequate (McCullagh & Nelder,


as is possible— with the goodness–of–fit test of theCJS model. Finally, the last section is devoted toproposals for the improvement of the present situa-tion and tries to identify future directions of research.The material presented in this paper has been imple-mented in program U–CARE, and is freely availableat http://ftp.cefe.cnrs.fr/biom/Soft–CR/.

Likelihood–based goodness–of–fit test forthe single–site CJS model

A perfect segregation of information between"estimation of parameters" and "test of assumptions"

For the sake of illustration, let us consider theobservations of 28,849 Canada Geese (Brantacanadensis) banded with individually–coded neckbands and re–observed at three locations: the mid–Atlantic (New York, Pennsylvania, New Jersey), theChesapeake (Delaware, Maryland, Virginia), and theCarolinas (North and South Carolina) (Hestbeck etal., 1991). Ignoring the locations for the moment, thedata can be summarized in what is called an m–array (table 1). At the beginning of each row is thenumber of geese released on each occasion, fol-lowed by the numbers of them reencountered for the

first time on each subsequent occasion. The m–array is an interesting summary because it turns outthat any set of encounter histories that produces thesame m–array yields the same maximum likelihoodestimates (MLE) of survival and encounter probabili-ties under the Cormack–Jolly–Seber (CJS) model.For this reason, the m–array is said to be a sufficientstatistic for the CJS model. Actually, even the mar-gins of the m–array, i.e. the total number ofreencounters per occasion mj’s and the numbersever seen again among those released at everyoccasion ri, are sufficient (Burnham et al., 1987).This is in fact the maximum reduction possible andthese margins are thus logically called minimal suffi-cient statistics (MSS). Table 2 presents a differentm–array with the same margins as the CanadaGeese data set. Thus, this m–array leads to thesame MLE’s under the CJS model. However, of twodata sets that lead to the same estimates one mayrespect the model assumptions while the other maynot. For instance, of the 3,494 individual geesereleased at occasion 1, we know for sure that309 + 159 + 64 + 42 = 574, which had not been en-countered at occasion 2 but were encountered later,were still alive at occasion 3. At the same time,1,941 + 734 + 345 + 154 = 3,174 of the 7,098 geesereleased at occasion 2 were also alive. The twogroups had experienced distinct encounter historiesup to occasion 3 but under the assumptions of theCJS model, this should be irrelevant as regards theirfuture; for instance, each of them should have anequal chance of being encountered at occasion 3.This may be tested using a contingency table:

Seen at 3 Seen later

Last seen at 1 309 265

Last seen at 2 1,941 1,233

Table1. m–array for the Canada goose data(Hestbeck et al., 1991) pooled over sites: i.Occasion of release; Ri. Number released ati; mij. Number reencountered at j among thosereleased at i; ri. Number ever reencounteredamong those released at i; mj. Total numberreencountered at occasion j.

Tabla 1. Serie m para los datos de la barnaclacanadiense (Hestbeck et al., 1991) una vezreunidos los de diversas localidades: i. Ocasiónde liberación; Ri. Número de liberaciones en laocasión i; mij. Número de reencuentrossiguientes en la ocasión j con una liberacióndada en la ocasión i; ri. Número que se hanvuelto a ver con una liberación dada en laocasión i; mj. Número de reencontrados en laocasión j.

mij’s

i Ri 2 3 4 5 6 ri

1 3,494 1,138 309 159 64 42 1,722

2 7,098 1,941 734 345 154 3,174

3 7,603 2,180 740 329 3,249

4 6,804 1,905 702 2,607

5 5,170 1,472

mj 1,138 2,250 3,073 3,054 2,699

Table 2. A fake m–array with the same marginsas that of the Canada goose data. (Forabbreviations see table 1.)

Tabla 2. Una serie m simulada, con los mismosmárgenes que la de los datos de la barnaclacanadiense. (Para las abreviaturas ver tabla 1.)

mij's

i Ri 2 3 4 5 6 ri

1 3,494 1,138 9 459 64 42 1,722

2 7,098 2,241 34 745 154 3,174

3 7,603 2,580 40 629 3,249

4 6,804 2,205 402 2,607

5 5,170 1,472

mj 1,138 2,250 3,073 3,054 2,699

http://ftp.cefe.cnrs.fr/biom/Soft�CR/

192 Pradel et al.

As it turns out, this table is slightly unbalancedbut far less than the corresponding table from thefake m–array:

Seen at 3 Seen later

Last seen at 1 9 565

Last seen at 2 2,241 933

Thus, while knowing the MSS suffices to esti-mate the parameters, details of the data i.e. theencounter histories, are needed to test the modelassumptions. One may wonder, on the other hand,whether something can be learned from the MSSabout the respect of model assumptions.

The answer to this question depends on themodel. In general, there is indeed something tolearn from the examination of the MSS but not inthe case of the CJS model. The CJS model hasindeed a peculiarity: its number of MSS is exactlyequal to its number of parameters. For instance,the Canada goose study spans 6 years; thus thereare 5 m’s and 5 r’s, and hence a total of 10margins. However, the sum of the m’s and that ofthe r’s are both equal to the total number of animalsin the data set. Therefore, there are only 9 minimalsufficient statistics (one margin can be spared).The CJS model with 6 occasions has 5 survival and5 encounter parameters, 10 parameters in total; butagain, at the last time step, only the product of thelast survival by the last encounter is estimable, andhence there are only 9 true parameters in total. Itcan be shown that every time that the number ofindividual statistics making up the MSS is exactlyequal to the number of parameters in the model —as is true of the CJS model— there is nothing tolearn from the MSS with respect to model assump-tions. The likelihood can always be factorized intotwo terms: one, the probability of the encounterhistories given the MSS, and the other, the prob-ability of the MSS given the parameters.

Pr (data; )= Pr (data / MSS) Pr (MSS; ) (1)

In the case of the CJS model, (1) correspondsto a perfect separation of the information.Pr (data / MSS) serves to check the model as-sumptions, and Pr (MSS; ) serves solely toestimate the parameters. The construction of anoptimal goodness–of–fit test is thus based on thesole first part, Pr (data / MSS).

The CJS model makes several assumptions.Based on the encounter histories of otherwisesimilar individuals, not all are verifiable: for in-stance, the assumption that the marked animalsare representative or that the band codes are notmisread. In fact, based on the study of the part Pr(data / MSS) of the likelihood, it can be shown thatthe verifiable assumptions come down to essen-tially one thing: all animals present at any giventime are assumed to behave the same. Pollock etal. (1985) have further shown that this, in turn, canbe divided into two (conditionally) independent main

points to be checked: 1) all animals released to-gether have the same expected future whatevertheir past encounter history and 2) all animals aliveat the same date that will be seen again do notdiffer in the timing of their reencounters whetherthey are currently encountered or not. The firstpoint leads to what is known as TEST 3; the secondto TEST 2, which is also known as the Jolly–Balsertest (Balser, 1984). This is actually not the only wayof breaking down the general test (Pollock et al. dopropose another form of their goodness–of–fit test)but it is the most commonly used and the one wewill consider. Starting from this decomposition, itbecomes possible to see how tests of specifichypotheses articulate with the general test andamong themselves. This has not been done sys-tematically and to our knowledge, the Leslie–Carothers test of unequal catchability (Carothers,1971) for instance has never been related to theoptimal GOF test of CJS. There are already at leasttwo specific tests which have been fully incorpo-rated into the GOF test of the CJS model and towhich alternative models have been attached. Weexamine them now in turn.

A test of transience

TEST 3 theoretically compares, at each occasion,the future history of encountered individuals withrespect to their previous encounter history. In practi-cal implementations, the comparison is limited tonewly marked and previously marked individuals.That these two categories should have similar ex-pectations implies an equal chance of being seenagain. It is thus possible to distinguish two steps inTEST 3: first, the check that newly and already–marked animals have an equal chance of being seenagain and then, for those seen again, the check thatthe spread of next reencounters over time is similarin the two categories. (This corresponds in practiceto the partitioning of contingency tables, a veryclassical statistical technique.) The firstsubcomponent (table 3) has been suggested manytimes and has been known since at least 1969(Robson, 1969). It has received an interpretation asa test for an effect of marking on immediate survivali.e. in the period immediately following release(Brownie & Robson, 1983). It has also been shownto be the adequate test to detect the presence oftransients, animals that are passing through thestudy site en route to other locations (Pradel et al.1997). This test is called the Brownie–Robson test orTEST 3.SR. An interesting point is that it is the testof comparison of CJS —or ( t, pt) in the notation ofLebreton et al. (1992)— with the more general modelthat provides for 2 age–classes in survival ( a2*t, pt).As a consequence, a GOF test for model ( a2*t, pt) isreadily available from the GOF test of CJS byignoring subcomponent TEST 3.SR.

If the alternative of interest is the presence oftransients, the direction of departure is predictable.In this case a directional test is appropriate. Onesuch test can be computed by taking the square


root of the Pearson chi–square statistics and givingit a conventional sign (see table 3).

For the Canada geese (table 4), the overall testis highly significant ( 2(4) = 54.24; P < 10–10). Amore specific and thus more powerful overall testof transience can be based on the statistic

∑=

=p

iiz

pz

1

1

which is standardized normal N(0,1) under H0 (notransience or age effect) and will tend to be positiveunder H1. Here z = 6.766 is highly significant.

A test of short–term trap–dependence

The other test of a specific alternative that has beenfully incorporated in the general GOF test of CJS isdirected at detecting immediate trap–dependenceon encounter probability, meaning that animals thatare encountered at occasion i have a different,higher (in case of trap–happiness) or lower (in caseof trap–shyness), probability of encounter than therest of the population at the next occasion i+1(table 5). The tables built in Section "A perfectsegregation of information between 'estima-tion of parameters' and 'test of assumptions'" areexamples of this test. As mentioned earlier, TEST 2compares the future of animals alive at the sameoccasion which are then encountered or not en-countered. Just as TEST 3.SR was obtained as asubcomponent of TEST 3, the test of trap–depend-ence, called 2.CT, is obtained as a subcomponentof TEST 2. (The complement of TEST 2.CT inTEST 2 investigates the timing of next encounters

Table 3. TEST 3.SR. This subcomponent ofthe CJS goodness–of–fit test is also a specifictest of transience. The signs indicate theexpected difference between observed andexpected values if there are transients caughtin the samples: Sl. Seen later; Nsa. Neverseen again; Nsb. Never seen before; Sb. Seenbefore.

Tabla 3. TEST 3.SR. Este subcomponente deltest de bondad de ajuste CJS es también untest específico de divagancia. Los signosindican la diferencia esperada entre los valoresobservados y esperados, si en las muestrasexisten transeúntes: Sl. Visto posteriormente;Nsa. No se ha vuelto a ver; Nsb. Nunca vistocon antelación; Sb. Visto con antelación.

Sl Nsa

Nsb ("new" or "newly marked") – +

Sb ("old" or "already marked") + –

Table 4. Results of TEST 3.SR for the Canadagoose data. The test can be calculated ateach of the 4 intermediate occasions. Thetable gives the Pearson chi–square statistics(X2) and the corresponding P–value (P) aswell as the signed square root (z) of thePearson chi–square statistic which is normallyN(0,1) distributed. z is positive when there isan excess of never seen again among thenewly marked.

Tabla 4. Resultados del TEST 3.SR para losdatos de la barnacla canadiense. Este testpuede calcularse en cada una de las cuatroocasiones intermedias. La tabla proporcionala ji–cuadrado de Pearson (X2) y sucorrespondiente valor P (P), así como la raízcuadrada provista de signo (z) de la ji–cuadrado de Pearson, que presenta unadistribución normal N(0,1). z es positiva cuandoexiste un exceso de individuos nunca vistosantes entre los que acaban de ser marcados.

Occasion z X2 P

2 1.11 1.24 0.27

3 5.16 26.58 0.00

4 3.79 14.33 0.00

5 3.48 12.09 0.00

Table 5. TEST 2.CT. This subcomponent ofthe CJS goodness–of–fit test is also a specifictest of immediate trap–dependence. The signsindicate the expected difference betweenobserved and expected values in case of trap–happiness. They should be reversed for trap–shyness: E i+1. Encountered at i+1; El.Encountered later; NEi. Not encountered at i;Ei. Encountered at i.

Tabla 5. TEST 2.CT. Este subcomponente deltest de bondad de ajuste CJS es también untest específico de la dependencia del trampeoinmediato. Los signos indican la diferenciaesperada entre los valores observados yesperados en el caso de animales habituadosa la trampa. Deberían ser contrarios en el casode individuos no habituados a la trampa: Ei+1.Encontrado en i+1; El. Encontrado más tarde;NEi. No encontrado en i; Ei. Encontrado en i.

Ei+1 El

NEi – +

Ei + –

(for p components)

194 Pradel et al.

Table 6. Results of TEST 2.CT for the Canadagoose data. This test can be calculated at eachintermediate occasions but the penultimate.The table gives the Pearson chi–squarestatistics (X2) and the corresponding P–value(P) as well as the signed square root (z) of thePearson chi–square statistics which is normallyN(0,1) distributed. z is negative when there isan excess of "encountered at i+1" among the"encountered at i".

Tabla 6. Resultados del TEST 2.CT para losdatos de la barnacla canadiense. Este testpuede calcularse en cada una de las ocasionesintermedias exceptuando la penúltima. La tablaproporciona la ji–cuadrado de Pearson (X2) ysu correspondiente valor P (P), así como laraíz cuadrada provista de signo (z) de la ji–cuadrado de Pearson, que presenta unadistribución normal N(0,1). z es negativa cuandoexiste un exceso de individuos "encontradosen i+1" entre los "encontrados en i".

Component z 2 P

2 -3.29 10.86 0.00

3 -5.02 25.16 0.00

4 -3.13 9.80 0.00

of the animals not encountered at i+1.) The alterna-tive model to CJS here is the generalization, noted( t,pt*m), allowing for a different encounter probabil-ity of animals just released. A GOF test for thismodel can be obtained by leaving out subcomponent2.CT from the GOF test of CJS.

Unlike for transients, the direction of departurecan be in any direction with trap–dependence. Yet,we expect the effect to be consistent over occa-sions. Thus the signed z statistic remains usefulwhen combining the TESTs 2.CT of the differentoccasions: the evidence for a trap effect accumu-lates with tables repeatedly unbalanced in the samedirection (table 6).

For the geese, there is overwhelming evidencethat encounter probability is much higher for agoose encountered at the previous occasion. Boththe omnibus chi–squared statistics and the direc-tional test are highly significant (X2 = 45.8212,P < 10–9; z = –6.6061, P < 10–10). However, wehave up to now ignored the site of observation. If,as is likely, the effort of observation is unequal andthe geese tend to be faithful to the same site fromyear to year, a goose that frequents the site withhigh observation pressure will tend to be reobservedconsistently more often, leading to a spurious trapeffect. To get around this problem, we now turn ourattention toward multisite (also multistate) models,more specifically, the JMV model.

Likelihood–based goodness–of–fit test forthe multistate model JMV

In multisite protocols, the individuals are sampledover K occasions and s sites. In the example of theCanada goose, there are 3 main areas in the Atlanticflyway, which we will now distinguish. The data canagain be summarized in a multisite or multistate m–array (Brownie et al., 1993) (table 7), a generaliza-tion of the m–array for one–site data. The compari-son of table 7 with table 1 should make clear how themultistate m–array is built. Therefore, we introducehere another approach to the m–array. Each en-counter history can be split into several pieces, fromthe first release to the next reencounter, from thesubsequent release to the next reencounter and soon until the end of the study period. For instance, thecapture history 302300 over 6 occasions may beseen as made of the three pieces: 302000, 002300and 000300. Each time that the individual isreencountered (the first two pieces), it is treated as ifremoved from the data set; this insures that only oneindividual remains present at the same time in thedata set. Each piece is then treated as if comingfrom an independent individual. The m–array is es-sentially the tally of these pieces arranged by rowsaccording to the occasion and state of release, andby columns according to the occasion and state ofnext reencounter (plus a "never–reencountered" col-umn). Obviously, for a model that assumes that thefate of an individual is not affected by its pastcapture history, the information retained is sufficient.However, because of the loss of information accom-panying the construction of the m–array, some as-sumptions can no longer be checked; for instance,whether some individuals are encountered signifi-cantly more often than others. This explains why theobjectives of checking the model assumptions andthat of estimating the parameters tend to use thecomplementary part of the total information.

An imperfect segregation of information between"estimation" and "test of assumptions"

The basic assumptions inherent in the JMV model aresimilar to those of the CJS model except that differ-ences between individuals in the different states arenow acknowledged. Again, the fate of the individualsthat are in the same state at the same time does notdepend on their past. A consequence is that themultistate m–array is a sufficient statistic. Moreover, itcan be shown that, unlike the one–site m–array, themultistate m–array is minimally sufficient. Now, thenumber of sufficient statistics, i.e. the number of inde-pendent cells in the multistate m–array, is K (K – 1) s² / 2. This is greater than the number of identifiableparameters as soon as K > 3: there are indeed (K – 1) s² transition probabilities, plus (K – 1) s2 encounterprobabilities minus s2 because the encounter probabili-ties of the last occasion are not estimable separatelyfrom the last transitions; a total of (2K – 3) s2 trueparameters. The JMV model does not therefore havethe nice properties of the CJS model. For instance, it is


no longer possible to factorize the likelihood into a termused solely for parameter estimation and another forassessing the goodness of fit; the term Pr(MSS; ) offormula (1) retains some information about the respectof model assumptions and has to be examined whenassessing the fit of the JMV model. There is, however,an analogy with the CJS model which still holds. Theverifiable assumptions come down again to one verysimilar thing: all animals present at any given time atthe same site behave the same. And this is againequivalent to the verification of two (conditionally) inde-pendent points: 1) animals released together have thesame expected future whatever their past encounterhistory and 2) animals present at the same site at thesame date that are eventually reencountered do notdiffer in the timing of their reencounters whether theyare currently encountered or not. Thus, apart for theprecision of a common site, the exact same two maincomponents are retrieved.

Past encounter history should not matter (TEST 3G)

The first main component of the GOF test of theJMV model, called TEST 3G, examines the effect ofthe past capture history on the future of animalscaptured and released at the same time on thesame site (Pradel et al., 2003). It is thus theequivalent of TEST 3 of which it is a generalization.Again, there are many possible past capture histo-

ries and the practical implementation of this test asfound in program U–CARE version 2.0 (Choquet etal., 2003) considers only a limited number of situa-tions: the newly caught animals are on the first rowwhile the previously caught ones are dispatchedover the subsequent rows according to their site ofmost recent encounter (see table 8); the columnscorrespond to the particulars (time and site) of thenext encounter if any.

As can be seen in table 8, even with a largedata set like that of the Canada geese, empty cellseasily occur and some sort of pooling is needed.The results in table 9 were obtained with U–CAREversion 2.0 which has an automatic pooling algo-rithm built in. They show that the Canada geesecaught together differ strongly depending on theirpast (over all TEST 3G: X2(103) = 749.27; P < 10–14).A close examination of the individual contingencytables like that of table 8, especially the compari-son of expected and observed numbers in eachcell, might prove useful in understanding the rea-sons for the departure. However, the breaking upof TEST 3G into meaningful subcomponents is abetter option.

A generalized test of transienceA first subcomponent can be built to test for thepresence of transients in each sample defined by asite and an occasion (table 10). This straightfor-

Table 7. Multisite m–array for the Canada goose data. Sites are North Atlantic (1), the Chesapeake(2) and the Carolinas (3). Only the first 2 and the last occasions of release are shown: mij

r s. Numberof next reencounter at occasion j in site s given release at occasion i in site r; i. Release occasion;r. Release sites; Ri

r. Number released.

Tabla 7. Serie m multilocalidad para los datos de la barnacla canadiense. Las localidades son elAtlántico Norte (1), la región de Chesapeake (2) y las dos Carolinas (3). Sólo se muestran las dosprimeras y la última ocasión de liberación: mij

r s. Número del siguiente reencuentro en la ocasión j enla localidad s dada la liberación en la ocasión i en la localidad r; i. Ocasión de liberación; r. Localidadesde liberación; Ri

r. Número de liberaciones.

mijrs

2 3 4 5 6

i r Rir 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

1 1 785 239 53 0 36 18 0 13 6 0 6 5 1 5 2 0

1 2 2,086 85 615 6 36 158 2 22 92 3 7 32 2 3 22 0

1 3 623 24 49 67 11 30 18 3 10 10 0 8 3 2 5 3

2 1 2,082 491 134 0 149 71 3 51 42 3 21 13 0

2 2 3,918 159 869 15 63 335 10 41 164 3 18 74 2

2 3 1,698 14 101 158 8 47 48 7 16 18 1 14 11

… … … … … …

5 1 1,291 271 99 2

5 2 2,887 137 654 18

5 3 992 18 105 16

196 Pradel et al.

Table 8. Component 3G (2,1) of the JMV goodness–of–fit test applied to the Canada goose data.This component is based on the animals caught at occasion 2 on site 1. They are dispatchedaccording to the site of most recent encounter in rows and the particulars (time and location) of thenext encounter in columns. The "–" sign is used for animals that are caught for the first time (firstrow) or that will never be encountered again (last column).

Tabla 8. Componente 3G (2,1) del test de bondad de ajuste JMV aplicado a los datos de labarnacla canadiense. Este componente se basa en los animales capturados en la ocasión 2 en lalocalidad 1. Se han distribuido en filas según el lugar de encuentro más reciente y en columnassegún las circunstancias (tiempo y localidad) del siguiente hallazgo. El signo "–" se utiliza para losanimales capturados por primera vez (primera fila) o que nunca serán vueltos a encontrar (últimacolumna).

Time (j) and location (v) of first reencounter

Location j = 3 4 5 6 –

at time 1 v = 1 2 3 1 2 3 1 2 3 1 2 3 –

– 390 124 0 122 64 3 46 35 3 18 9 0 920

1 75 3 0 21 4 0 5 2 0 1 0 0 128

2 19 6 0 4 3 0 0 2 0 1 3 0 47

3 7 1 0 2 0 0 0 3 0 1 1 0 9

Table 9. Results of TEST 3G for the Canadagoose data. The table gives the Pearson chi–square statistic ( 2) and the correspondingP–value (P) as well as the number of degreesof freedom after pooling (df): Oc. Occasion;S. Site.

Tabla 9. Resultados del TEST 3G para losdatos de la barnacla canadiense. Las tablaspresentan la ji–cuadrado de Pearson (X2) ysu correspondiente valor P (P), así como elnúmero de grados de libertad (df) tras lareunión: Oc. Ocasión; S. Localidad.

X2 P df Oc S

40.85 0.000 14 2 1

6.73 0.566 8 2 2

17.57 0.007 6 2 3

115.35 0.000 12 3 1

72.64 0.000 15 3 2

57.64 0.000 7 3 3

89.19 0.000 8 4 1

94.49 0.000 12 4 2

50.57 0.000 5 4 3

62.33 0.000 6 5 1

53.75 0.000 6 5 2

88.16 0.000 4 5 3

Table 10. A suitable partitioning of TEST3G isolates TEST 3G.SR. Thissubcomponent is exactly similar to TEST3.SR but involves a further stratification bysite (and not only by date). It is also aspecific test of transience. The signs indicatethe expected difference between observedand expected values when transients arecaught in the samples: Sl. Seen later, Nsa.Never seen again; Nsb. Never seen before;Sb. Seen before.

Tabla 10. Una partición adecuada del TEST 3Gaísla al TEST 3G.SR. Este subcomponentetiene una similitud exacta con el TEST 3.SR,pero presenta una mayor estratificaciónrespecto a la localidad (y no sólo a la fecha).También constituye un test específico de ladivagancia. Los signos indican la diferenciaesperada entre los valores observados yesperados cuando en las muestras se capturantranseúntes: Sl. Visto posteriormente; Nsa. Nose ha vuelto a ver; Nsb. Nunca visto conantelación; Sb. Visto con antelación.

Sl Nsa

Nsb ("new" or "newly marked") – +

Sb ("old" or "already marked") + –


ward generalization of TEST 3.SR, called TEST3G.SR, provides a GOF test of a generalized JMVmodel, a model with two age classes on survival.This generalization of the classical JMV model isdenoted Fa2*from*t, from*to*t, pfrom*to*t in the notation ofChoquet et al. (2004): F is survival, transition andp encounter probability ; from is the site of depar-ture, to the site of arrival. The goodness–of–fit testof this model is obtained by leaving outsubcomponent 3G.SR when calculating the overallGOF test.

Applied to the Canada geese, TEST 3G.SR re-veals an interesting feature (table 11). Althoughglobally significant (X2(12) = 117.753; P < 10–13),the test is not significant when restricted to site 2only (X2(4) = 3.19; P = 0.53). Thus, there seem tobe no transients in the central Chesapeake region!A directional z statistic could be calculated in thesame manner as with TEST 3.SR.

A test of memoryAnimals may make decisions of movement basedon the knowledge of previously visited sites.Hestbeck et al. (1992) have identified this phenom-enon in the election of wintering sites by Canadageese. This "memory effect", which is probablycommon in many long–lived species, is a violationof the assumption of the JMV model of the sort thatTEST 3G examines: it leads to different behaviourfor animals belonging to the same sample depend-ing on which sites they had visited previously. Thismemory effect is detectable by the specific test of

memory, called WBWA, proposed by Pradel et al.(2003). We will show in the next section that TESTWBWA presented in table 12 is a subcomponent ofTEST 3G.

Applied to the Canada geese (table 13), TESTWBWA confirms the very strong role of memory inthe movements of these birds (X2(20) = 472.86;P < 10–14): the overdispersion factor calculated forthis test alone is 472.86/20 = 23.6, much higherthan that of the overall TEST 3G (7.27) or that ofthe test for transience (9.81).

In order to more specifically target the depar-tures which are expected under the memory effectalong the diagonal, an alternative statistic to thePearson Chi–square can be used. One such possi-bility is Cohen’s kappa (Cohen, 1960), which has astandardized normal N(0,1) distribution. The indi-vidual kappa tests can be combined in the samemanner as the z tests of section 1.2 to get anoverall test of memory. They are added and theirsum is divided by the square root of the number ofcomponents p (Gimenez, 2003):

∑=

p

iip 1

1 κ

Table 11. Results of TEST 3G.SR for theCanada goose data: Oc. Occasion; S. Site.

Tabla 11. Resultados del TEST 3G.SR paralos datos de la barnacla canadiense: Oc.Ocasión; S. Localidad.

X2 P Oc S

0.004 0.95 2 1

0.000 0.99 2 2

8.130 0.00 2 3

11.394 0.00 3 1

2.708 0.10 3 2

33.459 0.00 3 3

10.608 0.00 4 1

0.353 0.55 4 2

10.168 0.00 4 3

11.013 0.00 5 1

0.129 0.72 5 2

29.785 0.00 5 3

Table 12. TEST WBWA, this subcomponentof TEST 3G tests for a memory effect. Itcompares the site of the most recentobservation in row ("Where Before") to thesite of the next observation in column ("WhereAfter"). The "+" signs indicate where theobserved values should exceed the expectedvalues when animals tend to return topreviously visited sites: Nss. Next seen onsite; Lss. Last seen on site.

Tabla 12. TEST WBWA, este subcomponenteTEST 3G comprueba el efecto de la memoria.Compara la localidad de la observación másreciente en la fila ("Where Before"), con lalocalidad de la siguiente observación en lacolumna (“Where After”). El signo "+" indicadónde los valores observados deberíanexceder a los valores esperados, cuando losanimales tienden a volver a las localidadespreviamente visitadas: Nss. Siguienteavistamiento en la localidad; Lss. Últimoavistamiento en la localidad.

Nss 1 Nss 2 … Nss s

Lss 1 +

Lss 2 +

… +

Lss s +

198 Pradel et al.

This directional test of memory confirms thestrong positive correlation between the previousand the next sites of observation of the Canadageese ( = 16.92; P < 10–13). There is as yet nosimple alternative model associated to TEST WBWA(like the 2–age model associated to TEST 3G.SR),but the model that accounts for the location at i–1in the transitions (Brownie et al., 1993) will prob-ably treat most of the "memory effect". Unfortu-nately, this model cannot be fitted in the frameworkof multistate models for the full data as it belongsto a more general family of capture–recapture mod-els (Pradel, 2005).

The full decomposition of TEST 3GTEST 3G.SR and TEST WBWA are two independ-ent subcomponents of TEST 3G but they do notmake up for TEST 3G alone. We illustrate herehow the original table of TEST 3G is partitioned toisolate these specific tests with the example of theCanada geese encountered at occasion 2 and site1 (table 8). This procedure of partitioning is verygeneral with contingency tables (Everitt, 1977). Ina first step, table 8 is replaced with two tables.This step consists in setting aside the previously

captured geese (first 3 rows) in a separate tableand then confronting them all pooled togetheragainst the newly captured geese in a secondtable.

(see 2 contingency tables of step 1 below)

Then, within each one of these two new tables,the never–seen–again geese (last column) areset aside leading to four new tables, one of whichis the component of TEST 3.SR relative to occa-sion 2 and site 1. In this step, the timing of firstreencounters is compared among the differentrows in a first table, and then the total ofreencounters is compared with the number ofnever–seen–again animals among the same rowsin a second table.


Eventually, the first of the four previous tables,which summarizes the first reencounters of thepreviously encountered individuals, is replacedwith four tables: one contains the reencountersmade at site 1, one those made at site 2, onethose made at site 3, and the last one contraststhe number of reencounters at each site depend-ing on the site of most recent encounter (inrows). This last one is the component of TESTWBWA relative to occasion 2 and site 1. Of the 7tables obtained at this stage (the last three tablesof step 2 remain unchanged in the last step), onlytwo belong to the specific tests described in theprevious sections.


Table 13. Results of TEST WBWA for theCanada goose data. The table gives thePearson chi–square statistic (X2) and thecorresponding P–value (P) as well as thenumber of degrees of freedom after pooling(df): Oc. Occasion; S. Site.

Tabla 13. Resultados del TEST WBWA paralos datos de la barnacla canadiense. Lastablas presentan la ji–cuadrado de Pearson(X2) y su correspondiente valor P (P), asícomo el número de grados de libertad (df)tras la reunión: Oc. Ocasión; S. Localidad.

X2 P df Oc S

19.59 0.000 2 2 1

37.87 0.000 2 2 2

4.49 0.034 1 2 3

80.59 0.000 1 3 1

98.76 0.000 4 3 2

0.81 0.369 1 3 3

27.71 0.000 1 4 1

53.69 0.000 2 4 2

25.29 0.000 1 4 3

43.66 0.000 1 5 1

50.93 0.000 2 5 2

29.48 0.000 2 5 3

Table 14. Directional test of memory appliedto the Canada goose data. This test isdistributed as N(0,1) and looks at a consistentexcess (or lack) on the diagonal.

Tabla 14. Test direccional de memoria aplicadoa los datos de la barnacla canadiense. Estetest se distribuye como N(0,1) y demuestraun exceso (o una falta) consistente en ladiagonal.

P P

3.87 0.00

4.33 0.00

1.81 0.04

8.59 0.00

5.98 0.00

1.01 0.21

4.33 0.00

7.22 0.00

4.22 0.00

6.36 0.00

6.47 0.00

4.44 0.00


Step1. Two contingency tables.

Paso 1. Dos tablas de contingencia.

75 3 0 21 4 0 5 2 0 1 0 0 128

19 6 0 4 3 0 0 2 0 1 3 0 47

7 1 0 2 0 0 0 3 0 1 1 0 9

390 124 0 122 64 3 46 35 3 18 9 0 920

101 10 0 27 7 0 5 7 0 3 4 0 184

Step 2. Four contingency tables.

Paso 2. Cuatro tablas de contingencia.

75 3 0 21 4 0 5 2 0 1 0 0

19 6 0 4 3 0 0 2 0 1 3 0

7 1 0 2 0 0 0 3 0 1 1 0

111 128

38 47

15 9

390 124 0 122 64 3 46 35 3 18 9 0

101 10 0 27 7 0 5 7 0 3 4 0

(TEST 3G.SR)

814 920

164 184

The remaining tables constitute together TEST3G.Sm of U–CARE version 2.2. To summarize,TEST 3G is made up of 3 subcomponents: TEST3G.SR, which tests specifically for transients; TESTWBWA, which aims at detecting a memory effect,and the complementary composite TEST 3G.Sm.

To be caught or not should have no effect (TEST M)

The second main component of the JMV modelGOF test, called TEST M, contrasts the animals

not caught at a given occasion —yet known to bealive— to those caught at the same occasion.Again, the JMV assumptions imply that thereshould be no difference between two animalswhen one is caught and the other is not. How-ever, the exact location of the animals not en-countered remains unknown. This is the mostfar–reaching difference with the one–site contextand the reason why the multistate JMV modelhas not all the nice properties of the single–siteCJS model (see Section "An imperfect segrega-

200 Pradel et al.

Step 3. Seven contingency tables.

Paso 3. Siete tablas de contingencia.

75 21 5 1

19 4 0 1

7 2 0 1

3 4 2 0

6 3 2 3

1 0 3 1

0 0 0 0

0 0 0 0

0 0 0 0

(TEST WBWA)

102 9 0

24 14 0

10 5 0

111 128

38 47

15 9

(TEST 3G.SR)

814 920

164 184

390 124 0 122 64 3 46 35 3 18 9 0

101 10 0 27 7 0 5 7 0 3 4 0

Table 15. Component of TEST M relative to date 2 for the Canada geese. The first three rowscorrespond to the geese not observed at occasion 2 that were released at sites 1, 2 and 3respectively at date 1. The last three rows are for the geese observed at date 2 on the three sitesin the same order. The columns correspond to the particulars (site within date) of the nextreencounter.

Tabla 15. Componente del TEST M, relativa a la fecha 2, para la barnacla canadiense. Las tresprimeras filas corresponden a los gansos no observados en la ocasión 2, que fueron liberados enla fecha 1 en las localidades 1, 2 y 3, respectivamente. Las tres últimas filas corresponden a losgansos observados en la fecha 2 en las tres localidades en el mismo orden. Las columnascorresponden a los datos (localidad en una fecha determinada) del siguiente reencuentro.

Time (j) and location (v) of first reencounter

j = 3 4 5 6

v = 1 2 3 1 2 3 1 2 3 1 2 3

last seen at site

1 36 18 0 13 6 0 6 5 1 5 2 0

2 36 158 2 22 92 3 7 32 2 3 22 0

3 11 30 18 3 10 10 0 8 3 2 5 3

currently seen at site

1 491 134 0 149 71 3 51 42 3 21 13 0

2 159 869 15 63 335 10 41 164 3 18 74 2

3 14 101 158 8 47 48 7 16 18 1 14 11


Table 16. Results of TEST M for the Canadageese. This test cannot be computed at thefirst occasion or less than 2 occasions beforethe end of the study: Oc. Occasion.

Tabla 16. Resultados del TEST M para labarnacla canadiense. Este test no puedecalcularse a la primera ocasión, o a menos dedos ocasiones antes del final del estudio.

X2 P df Oc

24.119 0.044 14 2

36.037 0.007 18 3

23.098 0.006 9 4

Table 17. TEST M.ITEC: this subcomponentof TEST M tests for trap–dependence. In caseof local trap–happiness, observed numbersshould exceed expected numbers where thereare "++" signs. If trap–dependence is presenteven if the animal has moved to another site(or more presumably state), the excesseswill show also where there are "+" signs:Sns1. Seen next occasion at site 1; Sns2.Seen next occasion at site 2; Sls1. Sen laterat site 1; Sls2. Seen later at site 2; Lss1.Last seen at site 1; Lss2. Last seen at site2. Cs1. Currently seen at site 1; Cs2.Currently seen at site 2.

Tabla 17. TEST M.ITEC: este subcomponentede los TEST M está destinado a demostrar ladependencia a la trampa. En el caso de unahabituación a la trampa local, las cifrasobservadas deberían exceder a las esperadasen los lugares en los que existen signos "++".Si la dependencia a la trampa existe inclusosi el animal se ha desplazado a otro lugar (omás probablemente a otro estado), losexcesos se pondrán también de manifiestodonde existan signos "+": Sns1. Visto lasiguiente ocasión en la localidad 1; Sns2.Visto la siguiente ocasión en la localidad 2;Sls1. Visto más tarde en la localidad 1; Sls2.Visto más tarde en la localidad 2; Lss1. Vistola última vez en la localidad 1; Lss2. Visto laúltima vez en la localidad 2. Cs1. Vistohabiatualmente en la localidad 1; Cs2. Visohabitualmente en la localidad 2.

Sns1 Sns2 … Sls1 Sls2 ...

Lss1

Lss2

…

Cs1 ++ + +

Cs2 + ++ +

… + + ++

tion of information between 'estimation' and 'testof assumptions'"). As regards the tests, becauseof the uncertain location of the not–encounteredindividuals, homogeneity tests of contingency ta-bles are now replaced with more complex tests ofmixtures.

Let us consider the table retained in U–CAREversion 2.0 for the component of TEST M relativeto date 2 for the Canada geese (table 15). Thefirst three rows correspond to the geese that werelast released at date 1 at sites 1, 2 and 3 respec-tively; the last three rows to the geese currentlyreleased at the same sites. The columns corre-spond to the timing and place of the nextreencounter. In this table, the first three rows do notplay the same role as the last ones; they should beapproximate linear combinations of the last ones.The rationale for this is as follows: the animals notobserved at date 2 may have moved since theywere last released; hence, their current location canbe any one of the three sites. These animals arethus a mixture of animals in the different sites inunknown proportions. In accordance with the modelassumptions, those at site 1 (resp. at site 2 and 3),i.e. on rows 1 (resp. 2 and 3), should behave likethose caught and released at site 1 (resp. at site 2and 3), i.e. on rows 4 (resp. 5 and 6).

The results concerning the Canada geese (ta-ble 16) are significant (overall test: 2(41) = 83.254;P < 10–3), although not as strong as those fromTEST 3G and its subcomponents. It is difficult toknow the reason for departure simply by examininga complex table like table 15, even if the expectednumbers were given. A suitable partitioning is againthe key to a better understanding.

A test of short–term trap–dependenceDrawing a parallel with the CJS GOF test, a testfor trap–dependence can be considered. The im-mediate question that arises is what trap–de-pendence means when there are several sites.Once an animal has been caught, is it expected

to change its behaviour the next time only if itremains at the same site, or should it changeeven if it moves to a different site? Presumably,the first option is more reasonable. However,when dealing with states instead of sites, thesecond option may be better: the animal will befaced with the same trap whatever its state. Thetable for testing for an immediate trap responseis the same in both cases (table 17). It is theregion of the table where departure is expectedtath differs. In the second case, the whole lowerleft quarter of the table should exhibit high (resp.low) numbers observed in case of trap–happi-

202 Pradel et al.

ness (resp. trap–shyness); in the first case, onlythe diagonal in the same quarter is expected tobe affected. The table corresponding to date 3 forthe Canada geese is given in table 18.

There is evidence of local trap–happiness in thegeese with a systematic excess on the diagonal ofthe lower left quarter of the table (table 18) and aglobally highly significant test (X2(27) = 68.177;P < 10–4) (table 19).

Discussion

Goodness–of–fit testing is not the most popular partof a capture–recapture analysis, probably because itis neither automatic nor very appealing. Although,some automatic procedures, such as the bootstrapprocedure built in MARK (White, 2001), are avail-able, they have their limitations (White, 2002) andabove all do not suggest what may be wrong whena model is rejected. On the other hand, optimalgoodness–of–fit procedures exist only for a very

limited number of models, and have long beenentirely missing for the multistate models. However,we believe that such procedures can be made moreuser–friendly and interpretable than they currentlyare, and that they have a great potential in helpingunderstand capture–recapture data. There is cer-tainly a lot of work yet to be done in this direction,but we have tried to show in this paper that there isalready a lot to be learned from them. We haveshown in particular that the goodness–of–fit multistatetest of the JMV model as proposed by Pradel et al.(2003) can be partitioned in subcomponents directlyrelated to some frequent violations of the assump-tions (transience, trap–dependence, memory). Some

Table 19. Results of TEST M.ITEC for theCanada geese: Oc. Occasion.

Tabla 19. Resultados del TEST M.ITEC parala barnacla canadiense: Oc. Ocasión.

X2 P df Oc

14.242 0.114 9 2

30.837 0.000 9 3

23.098 0.006 9 4

Table 20. Overdispersion factor calculated fordifferent components (straight police) andsubcomponents (italics) of the JMV and CJSgoodness–of–fit tests.

Tabla 20. Factor de sobredispersión calculadopara diferentes components (redondilla) ysubcomponents (cursiva) de los test de bondadde ajuste JMV (primeras dos columnas) yCJS (últimas dos columnas).

JMV CJS

3G 7.25

3GSR 9.8 13.5 3SR

WBWA 23.64

M 2.03

M.ITEC 2.53 15 2.CT

Table 18. Component of TEST M.ITEC for the Canada geese relative to date 3. There is evidence oflocal trap–happiness with the number in bold greater than expected: Sdi–sj. Seen at day i and site j.

Tabla 18. Componente del TEST M.ITEC para la barnacla canadiense, relativa a la fecha 3. Existenpruebas de una adicción a la trampa local, siendo las cifras en negrita mayores de lo esperado: Sdi–sj. Visto el día i en la localidad j.

Sd4–s1 Sd4–s2 Sd4–s3 Sd>4–s1 Sd>4–s2 Sd>4–s3

Last seen at site 1 162 77 3 83 62 4

Last seen at site 2 85 427 13 69 292 7

Last seen at site 3 11 57 58 10 43 35

Currently at site 1 564 200 8 202 162 7

Currently at site 2 125 1,017 36 82 471 20

Currently at site 3 7 45 178 12 48 65


of these effects can be treated more (transience) orless (trap–dependence) easily by fitting appropriatemultistate models; others (memory) call for entirelydifferent models. The fit of the alternative model tobe used in case of transience, namely the general-ized JMV model with two age classes on survival, isitself exactly testable by deducting TEST 3.SR fromthe overall goodness–of–fit test of JMV. Thus, this isa new model for which an optimal goodness–of–fittest is available. It would be interesting to examinewhether a model with an optimal goodness–of–fittest could be identified as well in case of trap–dependence. The memory effect is a more difficultchallenge. If this effect is strong, like for the Canadageese, all multistate models are invalidated. How-ever, if it is weak, it can be kept out of the structuralpart of the model provided an overdispersion factor(the ratio of the X2 statistic to its number of degreesof freedom) calculated from the goodness–of–fit testsis used in the analysis. An overdispersion factor canbe used more generally whenever there is no obvi-ous structural explanation for a lack of fit. Suitablypartitioned goodness–of–fit tests are thus a moregeneral tool than initially apparent for a correctassessment of the situation.

Beyond their purely technical usage, partitionedgoodness–of–fit tests can serve to unveil somebiological information. For instance, the intensity oftransit is likely related to dispersion (Perret et al.,2003; Cam et al., 2004); heterogeneity of capture(a test of which has yet to be incorporated within ageneral goodness–of–fit test) may be a reflection ofthe intensity of social structuration; the role ofmemory helps understand how the organism ap-prehends its environment. This potential has yet tobe fully exploited. The analysis of the Canadagoose data set that we have used throughout thispaper yields examples of the insight gained fromthe different components and subcomponents ofthe goodness–of–fit tests. A simple way to rank therelative strength of different effects is to calculatean overdispersion factor per components orsubcomponents of the CJS and the JMV good-ness–of–fit tests (table 20).

A first remark is that the correspondingsubcomponents for transients, 3.SR and 3G.SR,and particularly trap–dependence, 2.CT and M.ITEC,present higher overdispersion coefficients in theone–site than in the multisite context. Obviously,taking account of the location has removed part ofthe heterogeneity. This is not surprising as encoun-ter probabilities tend to be higher at some sites andat the same time the geese exhibit a high fidelity totheir wintering sites; hence, the same individualgeese tend to be consistently reencountered. Theexamination of the subcomponents of the multisiteTEST 3G reveals in turn that memory is by far themost important cause of departure confirming theneed for specific generalized models (Hestbeck etal., 1991; Brownie et al., 1993). Going through theoccasion– and site–specific tables, we have alsogained along the way new insights into the data:transit seems to affect only the peripheral sites and

trap dependence is more precisely local trap happi-ness. All this information has been obtained withoutfitting a single model so that, at the onset ofmodelling, we know for instance that a model withtransients on the two peripheral sites is appropri-ate. The risk of overfitting, which must be kept inmind, is limited here by the consistency of theeffects through several occasions. Another safe-guard is provided by the use of even more special-ized tests more precisely targeting the alternative ofinterest. The z–score tests of transience and theCohen’s kappa test of memory are two examples,but more can be developed, in particular for thedetection of trap–dependence.

Although the Arnason–Schwarz model (AS) isgenerally considered as the reference for multistateanalyses, we have not examined it specifically. Thisis because there is currently no specific goodness–of–fit test for it. The best approach is to treat the ASmodel as a particularization of the JMV model. Afterassessing the fit of JMV, JMV and AS can be fittedusing program M–SURGE (Choquet et al., 2004)and the two models compared with the AIC criterion(possibly modified to incorporate an overdispersionfactor). However, there is no more a priori reason tofit the AS model than any other multistate model. Atpresent, the most urgent need is the study of thestatistical properties of the new tests, notably thetests of mixture. For instance, in presence of sparsedata, there is no equivalent to the Fisher’s exacttest. Another very promising extension is the use ofthe multistate tests with recovery data. This ispossible because recoveries can be presented asmultistate data with two states: ‘alive’ and ‘dead’(Lebreton et al., 1999). However, as the state ‘dead’is absorbing, the tests have first to be modifiedaccordingly. There are more generally various po-tential original applications of the non–parametrictests presented in this paper (see for instanceGauthier et al. (2001) for seasonal trap–depend-ence). We believe that these tests should no longerbe considered only as the necessary routine firststep of a capture–recapture analysis but also as animportant part of the analysis itself, contributing inways that the parametric modelling cannot alwaysdo to the understanding of the data.

References

Arnason, A. N., 1972. Parameter estimates frommark–recapture experiments on two populationssubject to migration and death. Researches onPopulation Ecology, 13: 97–113.

– 1973. The estimation of population size, migra-tion rates and survival in a stratified population.Researches on Population Ecology, 15: 1–8.

Balser, J. P., 1984. confidence interval estimationand tests for temporary outmigration in tag–recapture studies. Doctoral dissertation, CornellUniv., Ithaca, New York.

Begon, M., 1983. Abuses of mathematical tech-niques in ecology: applications of Jolly’s cap-

204 Pradel et al.

ture–recapture method. Oikos, 40: 155–158.Brownie, C., Hines, J. E., Nichols, J. D., Pollock, K.

H. & Hestbeck, J. B., 1993. Capture–recapturestudies for multiple strata including non–Markovian transitions. Biometrics, 49: 1173–1187.

Brownie, C. & Robson, D. S., 1983. Estimation oftime–specific survival rates from tag–resightingsamples: a generalization of the Jolly–Sebermodel. Biometrics, 39: 437–453.

Burnham, K. P. & Anderson, D. R., 1998. Modelselection and inference: a practical information–theoretic approach. Springer–Verlag, New York.

Burnham, K. P., Anderson, D. R., White, G. C.,Brownie, C. & Pollock, K. H., 1987. Design andanalysis methods for fish survival experimentsbased on release–recapture. American FisheriesSociety, Bethesda, Maryland.

Cam, E., Oro, D., Pradel, R. & Jimenez, J., 2004.Assessment of hypotheses about dispersal in along–lived seabird using multistate capture–re-capture models. Journal of Animal Ecology, 73:723–736.

Carothers, A. D., 1971. An examination and exten-sion of Leslie’s test of equal catchability. Biomet-rics, 27: 615–630.

Choquet, R., Reboulet, A.–M., Pradel, R., Gimenez,O. & Lebreton, J.–D., 2003. U–CARE (Utilities –CApture–REcapture). CEFE, Montpellier, France.

– 2004. M–SURGE: new software specificallydesigned for multistate capture–recapture mod-els. Animal Biodiversity and Conservation, 27.1:207–215.

Choquet, R., Reboulet, A.–M., Lebreton, J.–D.,Gimenez, O. & Pradel, R., 2005. U–CARE 2.2User’s Manual. CEFE, Montpellier, France.

Cohen, J., 1960. A coefficient of agreement fornominal scales. Educational and psychologicalmeasurement, 20: 37–46.

Everitt, B. S., 1977. The analysis of contingencytables. Chapman & Hall, Londres.

Gauthier, G., Pradel, R., Menu, S. & Lebreton, J.–D., 2001. Seasonal survival of greater snowgeese and effect of hunting under dependence insighting probabilities. Ecology, 82: 3105–3119.

Gimenez ,O., 2003. Estimation et tests d’adéquationpour les modèles de capture–recapture multiétats.Ph. D. thesis. Univ. Montpellier II, Montpellier.

Hestbeck, J. B., Nichols, J. D. & Hines, J. E., 1992.The relationship between annual survival andmigration distance in mallards: an examinationof the time–allocation hypothesis for the evolu-tion of migration. Canadian Journal of Zoology,70: 2021–2027.

Hestbeck, J. B., Nichols, J. D. & Malecki, R. A.,1991. Estimates of movement and site fidelityusing mark–resight data of wintering canada

geese. Ecology, 72: 523–533.Lebreton, J.–D., Alméras, T. & Pradel, R., 1999.

Competing events, mixture of information andmultistrata recapture models. Bird Study, 46:39–46.

Lebreton, J.–D., Burnham, K. P., Clobert, J. &Anderson, D. R., 1992. Modeling survival andtesting biological hypotheses using marked ani-mals: A unified approach with case studies. Eco-logical Monographs, 62: 67–118.

Lebreton, J.–D. & Pradel, R., 2002. Multistate re-capture models: modelling incomplete individualhistories. Journal of Applied Statistics, 29: 353–369.

McCullagh, P. & Nelder, J. A., 1989. Generalizedlinear models, second edition. Chapman andHall, New York, USA.

Nichols, J. D., Hines, J. E., Pollock, K. H., Hinz, R.L. & Link, W. A., 1994. Estimating breedingproportions and testing hypotheses about costsof reproduction with capture–recapture data.Ecology, 75: 2052–2065.

Perret, N., Pradel, R., Miaud, C., Grolet, O. & Joly,P., 2003. Transience, dispersal, and survival ratesin newt patchy populations. Journal of AnimalEcology, 72: 567–575.

Pollock, K. H., Hines, J. E. & Nichols, J. D., 1985.Goodness–of–fit tests for open capture–recap-ture models. Biometrics, 41: 399–410.

Pradel, R., 1993. Flexibility in Survival analysisfrom recapture data: Handling trap–dependence.Pages 29–37 in Lebreton & North, editors. Markedindividuals in the study of bird population.Birkhaüser Verlag, Basel, Switzerland.

Pradel, R., Hines, J. E., Lebreton, J.–D. & Nichols, J.D., 1997. Capture–recapture survival models tak-ing account of transients. Biometrics, 53: 60–72.

Pradel, R. & Lebreton, J.–D., 1999. Comparison ofdifferent approaches to the study of local recruit-ment of breeders. Bird Study, 46: 74–81.

Pradel, R., Wintrebert, C. M. A. & Gimenez, O.,2003. A proposal for a goodness–of–fit test tothe Arnason–Schwarz multisite capture–recap-ture model. Biometrics, 59: 43–53.

Robson, D. S., 1969. Mark–recapture methods ofpopulation estimation. BU–168, Cornell Univ.

White, G. C., 2001. Advanced features of ProgramMark. Pages 368–377 in Field, Warren & Sievert,editors. Wildlife, Land, and People: Priorities forthe 21st Century. Proceedings of the SecondInternational Wildlife Management Congress. TheWildlife Society, Bethesda, Maryland, USA.

– 2002. Discussion comments on: the use ofauxiliary variables in capture–recapture modelling.An overview. Journal of Applied Statistics, 29:103–106.

Animal Biodiversity and Conservation 28.2 (2005) I

ISSN: 1578–665X © 2005 Museu de Ciències Naturals

Animal Biodiversity and Conservation

Animal Biodiversity and Conservation (abans Miscel·lània Zoològica) és una revista inter disciplinària publicada, des de 1958, pel Museu de Zoologia de Barcelona. Inclou articles d'inves tigació empírica i teòrica en totes les àrees de la zoologia (sistemàtica, taxo nomia, morfo logia, biogeografia, ecologia, etologia, fisiologia i genètica) procedents de totes les regions del món amb especial énfasis als estudis que d'una manera o altre tinguin relevància en la biología de la conservació. La revista no publica catàlegs, llistes d'espècies o cites puntuals. Els estudis realitzats amb espècies rares o protegides poden no ser acceptats tret que els autors disposin dels permisos corresponents. Cada volum anual consta de dos fascicles.

Animal Biodiversity and Conservation es troba registrada en la majoria de les bases de dades més importants i està disponible gratuitament a internet a http://www.bcn.es/ABC, de manera que permet una difusió mundial dels seus articles.

Tots els manuscrits són revisats per l'editor executiu, un editor i dos revisors independents, triats d'una llista internacional, a fi de garantir–ne la qualitat. El procés de revisió és ràpid i constructiu. La publicació dels treballs acceptats es fa normalment dintre dels 12 mesos posteriors a la recepció.

Una vegada hagin estat acceptats passaran a ser propietat de la revista. Aquesta es reserva els drets d’autor, i cap part dels treballs no podrà ser reproduïda sense citar–ne la procedència.

Normes de publicació

Els treballs s'enviaran preferentment de forma electrònica ([email protected]). El format preferit és un document Rich Text Format (RTF) o DOC que inclogui les figures (TIF). Si s'opta per la versió impresa, s'han d'enviar quatre còpies del treball juntament amb una còpia en disquet a la Secretaria de Redacció. Cal incloure, juntament amb l'article, una carta on es faci constar que el treball està basat en investigacions originals no publicades anterior ment i que està sotmès a Animal Biodiversity and Conservation en exclusiva. A la carta també ha de constar, per a aquells treballs en que calgui manipular animals, que els autors disposen dels permisos necessaris i que compleixen la normativa de protecció animal vigent. També es poden suggerir possibles assessors.

Quan l'article sigui acceptat, els autors hauran d'enviar a la Redacció una còpia impresa de la versió final acompanyada d'un disquet indicant el programa utilitzat (preferiblement en Word). Les proves d'impremta enviades a l'autor per a la correcció, seran retornades al Consell Editor en el termini de 10 dies. Aniran a càrrec dels autors les despeses degudes a modificacions substancials introduïdes per ells en el text original acceptat.

El primer autor rebrà 50 separates del treball sense càrrec a més d'una separata electrònica en format PDF.

Manuscrits

Els treballs seran presentats en format DIN A –4 (30 línies de 70 espais cada una) a doble espai i amb totes les pàgines numerades. Els manus crits han de ser complets, amb taules i figures. No s'han d'enviar les figures originals fins que l'article no hagi estat acceptat.

El text es podrà redactar en anglès, castellà o català. Se suggereix als autors que enviïn els seus treballs en anglès. La revista els ofereix, sense cap càrrec, un servei de correcció per part d'una persona especialitzada en revistes científiques. En tots els casos, els textos hauran de ser redactats correctament i amb un llenguatge clar i concís. La redacció del text serà impersonal, i s'evitarà sempre la primera persona.

Els caràcters cursius s’empraran per als noms científics de gèneres i d’espècies i per als neologismes intraduïbles; les cites textuals, independentment de la llengua, seran consignades en lletra rodona i entre cometes i els noms d’autor que segueixin un tàxon aniran en rodona.

Quan se citi una espècie per primera vegada en el text, es ressenyarà, sempre que sigui possible, el seu nom comú.

Els topònims s’escriuran o bé en la forma original o bé en la llengua en què estigui escrit el treball, seguint sempre el mateix criteri.

Els nombres de l’u al nou, sempre que estiguin en el text, s’escriuran amb lletres, excepte quan precedeixin una unitat de mesura. Els nombres més grans s'escriuran amb xifres excepte quan comencin una frase.

Les dates s’indicaran de la forma següent: 28 VI 99; 28, 30 VI 99 (dies 28 i 30); 28–30 VI 99 (dies 28 a 30).

S’evitaran sempre les notes a peu de pàgina.

Format dels articles

Títol. El títol serà concís, però suficientment indicador del contingut. Els títols amb desig nacions de sèries numèriques (I, II, III, etc.) seran acceptats previ acord amb l'editor.Nom de l’autor o els autors.Abstract en anglès que no ultrapassi les 12 línies mecanografiades (860 espais) i que mostri l’essència del manuscrit (introducció, material, mètodes, resultats i discussió). S'evitaran les especulacions i les cites bibliogràfiques. Estarà encapçalat pel títol del treball en cursiva.Key words en anglès (sis com a màxim), que orientin sobre el contingut del treball en ordre d’importància.Resumen en castellà, traducció de l'Abstract. De la traducció se'n farà càrrec la revista per a aquells autors que no siguin castellano parlants. Palabras clave en castellà.Adreça postal de l’autor o autors.(Títol, Nom, Abstract, Key words, Resumen, Palabras clave i Adreça postal, conformaran la primera pàgina.)

http://bcnweb13.bcn.es/NASApp/wprmuseuciencies/Museu.GeneradorPagines?idioma=1&seccio=3_1


II

Introducción. S'hi donarà una idea dels antecedents del tema tractat, així com dels objectius del treball.Material y métodos. Inclourà la informació pertinent de les espècies estudiades, aparells emprats, mètodes d’estudi i d’anàlisi de les dades i zona d’estudi.Resultados. En aquesta secció es presentaran únicament les dades obtingudes que no hagin estat publicades prèviament.Discusión. Es discutiran els resultats i es compararan amb treballs relacionats. Els sug geriments de recerques futures es podran incloure al final d’aquest apartat.Agradecimientos (optatiu).Referencias. Cada treball haurà d’anar acompanyat de les referències bibliogràfiques citades en el text. Les referències han de presentar–se segons els models següents (mètode Harvard):* Articles de revista:Conroy, M. J. & Noon, B. R., 1996. Mapping of spe

cies richness for conservation of biological diversity: conceptual and methodological issues. Ecological Applications, 6: 763–773.

* Llibres o altres publicacions no periòdiques:Seber, G. A. F., 1982. The estimation of animal abun-

dance. C. Griffin & Company, London. * Treballs de contribució en llibres:Macdonald, D. W. & Johnson, D. P., 2001. Dispersal

in theory and practice: consequences for conservation biology. In: Dispersal: 358–372 (T. J. Clober, E. Danchin, A. A. Dhondt & J. D. Nichols, Eds.). Oxford University Press, Oxford.

* Tesis doctorals:Merilä, J., 1996. Genetic and quantitative trait vari

ation in natural bird populations. Tesis doctoral, Uppsala University.

* Els treballs en premsa només han d’ésser citats si han estat acceptats per a la publicació:Ripoll, M. (in press). The relevance of population

studies to conservation biology: a review. Anim. Biodivers. Conserv.

La relació de referències bibliogràfiques d’un treball

serà establerta i s’ordenarà alfabè ticament per autors i cronològicament per a un mateix autor, afegint les lletres a, b, c,... als treballs del mateix any. En el text, s’indi caran en la forma usual: “...segons Wemmer (1998)... ”, “...ha estat definit per Robinson & Redford (1991)...”, “...les prospeccions realitzades (Begon et al., 1999)...” Taules. Les taules es numeraran 1, 2, 3, etc. i han de ser sempre ressenyades en el text. Les taules grans seran més estretes i llargues que amples i curtes ja que s'han d'encaixar en l'amplada de la caixa de la revista. Figures. Tota classe d’il·lustracions (gràfics, figures o fotografies) entraran amb el nom de figura i es numeraran 1, 2, 3, etc. i han de ser sempre ressenyades en el text. Es podran incloure fotografies si són imprescindibles. Si les fotografies són en color, el cost de la seva publicació anirà a càrrec dels autors. La mida màxima de les figures és de 15,5 cm d'amplada per 24 cm d'alçada. S'evitaran les figures tridimensionals. Tant els mapes com els dibuixos han d'incloure l'escala. Els ombreigs preferibles són blanc, negre o trama. S'evitaran els punteigs ja que no es repro dueixen bé. Peus de figura i capçaleres de taula. Els peus de figura i les capçaleres de taula seran clars, concisos i bilingües en la llengua de l’article i en anglès.

Els títols dels apartats generals de l’article (Introducción, Material y métodos, Resultados, Discusión, Conclusiones, Agradecimientos y Referencias) no aniran numerats. No es poden utilitzar més de tres nivells de títols.

Els autors procuraran que els seus treballs originals no passin de 20 pàgines (incloent–hi figures i taules).

Si a l'article es descriuen nous tàxons, caldrà que els tipus estiguin dipositats en una insti tució pública.

Es recomana als autors la consulta de fascicles recents de la revista per tenir en compte les seves normes.

Animal Biodiversity and Conservation 28.2 (2005) III



Animal Biodiversity and Conservation (antes Miscel·lània Zoològica) es una revista interdisciplinar, publicada desde 1958 por el Museo de Zoología de Barcelona. Incluye artículos de investigación empírica y teórica en todas las áreas de la zoología (sistemática, taxo nomía, morfología, biogeografía, ecología, etología, fisiología y genética) procedentes de todas las regiones del mundo, con especial énfasis en los estudios que de una manera u otra tengan relevancia en la biología de la conservación. La revista no publica catálogos, listas de especies sin más o citas puntuales. Los estudios realizados con especies raras o protegidas pueden no ser aceptados a no ser que los autores dispongan de los permisos correspondientes. Cada volumen anual consta de dos fascículos.

Animal Biodiversity and Conservation está registrada en todas las bases de datos importantes y además está disponible gratuitamente en internet en http://www.bcn.es/ABC, lo que permite una difusión mundial de sus artículos.

Todos los manuscritos son revisados por el editor ejecutivo, un editor y dos revisores independientes, elegidos de una lista internacional, a fin de garantizar su calidad. El proceso de revisión es rápido y constructivo, y se realiza vía correo electrónico siempre que es posible. La publicación de los trabajos aceptados se realiza con la mayor rapidez posible, normalmente dentro de los 12 meses siguientes a la recepción del trabajo.

Una vez aceptado, el trabajo pasará a ser propiedad de la revista. Ésta se reserva los derechos de autor, y ninguna parte del trabajo podrá ser reproducida sin citar su procedencia.

Normas de publicación

Los trabajos se enviarán preferentemente de forma electrónica ([email protected]). El formato preferido es un documento Rich Text Format (RTF) o DOC, que incluya las figuras (TIF). Si se opta por la versión impresa, deberán remitirse cuatro copias juntamente con una copia en disquete a la Secretaría de Redacción. Debe incluirse, con el artículo, una carta donde conste que el trabajo versa sobre inves tigaciones originales no publi cadas an te rior mente y que se somete en exclusiva a Animal Biodiversity and Conservation. En dicha carta también debe constar, para trabajos donde sea necesaria la manipulación de animales, que los autores disponen de los permisos necesarios y que han cumplido la normativa de protección animal vigente. Los autores pueden enviar también sugerencias para asesores.

Cuando el trabajo sea aceptado los autores deberán enviar a la Redacción una copia impresa de la versión final junto con un disquete del manuscrito preparado con un pro cesador de textos e indicando el programa utilizado (preferiblemente Word). Las pruebas de imprenta enviadas a los autores deberán

remitirse corregidas al Consejo Editor en el plazo máximo de 10 días. Los gastos debidos a modificaciones sustanciales en las pruebas de im pren ta, introducidas por los autores, irán a cargo de los mismos.

El primer autor recibirá 50 separatas del trabajo sin cargo alguno y una copia electrónica en formato PDF.

Manuscritos

Los trabajos se presentarán en formato DIN A–4 (30 líneas de 70 espacios cada una) a doble espacio y con las páginas numeradas. Los manuscritos deben estar completos, con tablas y figuras. No enviar las figuras originales hasta que el artículo haya sido aceptado.

El texto podrá redactarse en inglés, castellano o catalán. Se sugiere a los autores que envíen sus trabajos en inglés. La revista ofre ce, sin cargo ninguno, un servicio de corrección por parte de una persona especializada en revistas científicas. En cualquier caso debe presentarse siempre de forma correcta y con un lenguaje claro y conciso. La redacción del texto deberá ser impersonal, evitán dose siempre la primera persona.

Los caracteres en cursiva se utilizarán para los nombres científicos de géneros y especies y para los neologismos que no tengan traducción; las citas textuales, independientemente de la lengua en que estén, irán en letra redonda y entre comillas; el nombre del autor que sigue a un taxón se escribirá también en redonda.

Al citar por primera vez una especie en el trabajo, deberá especificarse siempre que sea posible su nombre común.

Los topónimos se escribirán bien en su forma original o bien en la lengua en que esté redactado el trabajo, siguiendo el mismo criterio a lo largo de todo el artículo.

Los números del uno al nueve se escribirán con letras, a excepción de cuando precedan una unidad de medida. Los números mayores de nueve se escribirán con cifras excepto al empezar una frase.

Las fechas se indicarán de la siguiente forma: 28 VI 99; 28, 30 VI 99 (días 28 y 30); 28–30 VI 99 (días 28 al 30).

Se evitarán siempre las notas a pie de página.

Formato de los artículos

Título. El título será conciso pero suficientemente explicativo del contenido del trabajo. Los títulos con designaciones de series numéricas (I, II, III, etc.) serán aceptados excepcionalmente previo consentimiento del editor.Nombre del autor o autores.Abstract en inglés de 12 líneas mecanografiadas (860 espacios como máximo) y que exprese la esencia del manuscrito (introducción, material, métodos, resultados y discusión). Se evitarán las especulaciones y las citas bibliográficas. Irá encabezado por el título del trabajo en cursiva.Key words en inglés (un máximo de seis) que especifiquen el contenido del trabajo por orden de importancia.



IV

Resumen en castellano, traducción del abstract. Su traducción puede ser solicitada a la revista en el caso de autores que no sean castellano hablan tes. Palabras clave en castellano.Dirección postal del autor o autores.(Título, Nombre, Abstract, Key words, Resumen, Palabras clave y Dirección postal conformarán la primera página.)

Introducción. En ella se dará una idea de los antecedentes del tema tratado, así como de los objetivos del trabajo.Material y métodos. Incluirá la información referente a las especies estudiadas, aparatos utilizados, metodología de estudio y análisis de los datos y zona de estudio.Resultados. En esta sección se presentarán únicamente los datos obtenidos que no hayan sido publicados previamente.Discusión. Se discutirán los resultados y se compararán con otros trabajos relacionados. Las sugerencias sobre investigaciones futuras se podrán incluir al final de este apartado.Agradecimientos (optativo).Referencias. Cada trabajo irá acompañado de una bibliografía que incluirá únicamente las publicaciones citadas en el texto. Las referencias deben presentarse según los modelos siguientes (método Harvard):* Artículos de revista:Conroy, M. J. & Noon, B. R., 1996. Mapping of spe

cies richness for conservation of biological diversity: conceptual and methodological issues. Ecological Applications, 6: 763–773

* Libros y otras publicaciones no periódicas:Seber, G. A. F., 1982. The estimation of animal abun-

dance. C. Griffin & Company, London. * Trabajos de contribución en libros:Macdonald, D. W. & Johnson, D. P., 2001. Dispersal


* Tesis doctorales:Merilä, J., 1996. Genetic and quantitative trait vari

ation in natural bird populations. Tesis doctoral, Uppsala University.

* Los trabajos en prensa sólo se citarán si han sido aceptados para su publicación:Ripoll, M. (in press). The relevance of population

studies to conservation biology: a review. Anim. Biodivers. Conserv.

Las referencias se ordenarán alfabética men te por autores, cronológicamen te para un mismo autor y con las letras a, b, c,... para los tra bajos de un mismo autor y año. En el texto las referencias bibliográficas se indicarán en la forma usual: "...según Wemmer (1998)...", "...ha sido definido por Robinson & Redford (1991)...", "...las prospecciones realizadas (Begon et al., 1999)..." Tablas. Las tablas se numerarán 1, 2, 3, etc. y se reseñarán todas en el texto. Las tablas grandes deben ser más estrechas y largas que anchas y cortas ya que deben ajustarse a la caja de la revista.Figuras. Toda clase de ilustraciones (gráficas, figuras o fotografías) se considerarán figuras, se numerarán 1, 2, 3, etc. y se citarán todas en el texto. Pueden incluirse fotografías si son imprescindibles. Si las fotografías son en color, el coste de su publicación irá a cargo de los autores. El tamaño máximo de las figuras es de 15,5 cm de ancho y 24 cm de alto. Deben evitarse las figuras tridimen sionales. Tanto los mapas como los dibujos deben incluir la escala. Los sombreados preferibles son blanco, negro o trama. Deben evitarse los punteados ya que no se reproducen bien.Pies de figura y cabeceras de tabla. Los pies de figura y cabeceras de tabla serán claros, concisos y bilingües en castellano e inglés.

Los títulos de los apartados generales del artículo (Introducción, Material y métodos, Resultados, Discusión, Agradecimientos y Referencias) no se numerarán. No utilizar más de tres niveles de títulos.

Los autores procurarán que sus trabajos originales no excedan las 20 páginas incluidas figuras y tablas.

Si en el artículo se describen nuevos taxones, es imprescindible que los tipos estén depositados en alguna institución pública.

Se recomienda a los autores la consulta de fascículos recientes de la revista para seguir sus directrices.

Animal Biodiversity and Conservation 28.2 (2005) V



Animal Biodiversity and Conservation (formerly Miscel·lània Zoològica) is an inter dis ci pli nary journal which has been published by the Zoological Museum of Bar celona since 1958. It includes empirical and theoretical research in all aspects of Zoology (Systematics, Taxo nomy, Morphology, Bio geography, Ecology, Etho logy, Physio logy and Genetics) from all over the world with special emphasis on studies that stress the relevance of the study of Conservation Biology. The journal does not publish catalogues, lists of species (with no other relevance) or punctual records. Studies about rare or protected species will not be accepted unless the authors have been granted all the relevant permits. Each annual volume consists of two issues.

Animal Biodiversity and Conservation is registered in all principal data bases and is freely available online at http://www.bcn.es/ABC, thus assuring world–wide access to articles published therein.

All manuscripts are screened by the Executive Editor, an Editor and two independent reviewers in order to guarantee the quality of the papers. The process of review is rapid and constructive. Once accepted, papers are published as soon as practicable, usually within 12 months of initial submission.

Upon acceptance, manuscripts become the property of the journal, which reserves copyright, and no published material may be reproduced without quoting its origin.

Information for authors

Electronic submission of papers is encouraged ([email protected]). The preferred format is a document Rich Text Format (RTF) or DOC, including figures (TIF). In the case of sending a printed version, four copies should be sent together with a copy on a computer disc to the Editorial Office. A cover letter stating that the article reports on original research not published elsewhere and that it has been submitted exclusively for consi deration in Animal Biodivers ity and Conservation is also necessary. When animal manipulation has been necessary, the cover letter should also especify that the authors follow current norms on the protection of animal species and that they have obtained all relevant permissions. Authors may suggest referees for their papers.

Once an article has been accepted, authors should send a printed copy of the final version together with a disc. Please identify software (preferably Word). Proofs sent to the authors for correction should be returned to the Editorial Board within 10 days. Expenses due to any substantial alterations of the proofs will be charged to the authors.

The first author will receive 50 reprints free of charge and an electronic version of the article in PDF format.

Manuscripts

Manuscripts must be presented on A–4 format page (30 lines of 70 spaces each) with double spacing. Number all pages. Manuscripts should be complete with figures and tables. Do not send original figures until the paper has been accepted.

The text may be written in English, Spanish or Catalan. Authors are encouraged to send their contributions in English. The journal provides a FREE service of correction by a professional translator specialized in scientific publications. Care should be taken in using correct wording and the text should be written concisely and clearly. Wording should be impersonal, avoiding the use of the first person.

Italics must be used for scientific names of genera and species as well as untrans latable neologisms. Quotations in whatever language used must be typed in ordinary print between quotation marks. The name of the author following a taxon should also be written in small print.

The common name of the species should be written in capital letters. When referring to a species for the first time in the text, both common and scientific names must be given when possible.

Place names may appear either in their original form or in the language of the manuscript, but care should be taken to use the same criteria throughout the text.

Numbers one to nine should be written in full in the text except when preceding a measure. Higher numbers should be written in numerals except at the beginning of a sentence.

Dates must appear as follows: 28 VI 99, 28,30 VI 99 (days 28th and 30th), 28–30 VI 99 (days 28th to 30th).

Footnotes should not be used.

Formatting of articles

Title. The title must be concise but as infor mative as possible. Part numbers (I, II, III, etc.) should be avoided and will be subject to the Editor’s consent.Name of author or authors.Abstract in English, no longer than 12 type written lines (840 spaces), covering the con tents of the article (introduction, material, methods, results and discussion). Speculation and literature citation must be avoided. Abstract should begin with the title in italics.Key words in English (no more than six) should express the precise contents of the manuscript in order of importance.Resumen in Spanish, translation of the Abstract.Summaries of articles by non –Spanish speaking authors will be trans lated by the journal on request. Palabras clave in Spanish.Address of the author or authors.(Title, Name, Abstract, Key words, Resumen, Palabras clave and Address should constitute the first page.)

Introduction. The introduction should in clude the historical background of the sub ject as well as the aims of the paper.



VI

Material and methods. This section should provide relevant information on the species studied, materials, methods for collecting and analysing data and the study area.Results. Report only previously unpublished results from the present study.Discussion. The results and their comparison with related studies should be discussed. Sug gestions for future research may be given at the end of this section.Acknowledgements (optional).References. All manuscripts must include a bibliography of the publications cited in the text. References should be presented as in the following examples (Harvard method):* Journal articles:Conroy, M. J. & Noon, B. R., 1996. Mapping of spe

cies richness for conservation of biological diversity: conceptual and methodological issues. Ecological Applications, 6: 763–773.

* Books or other non–periodical publications:Seber, G. A. F., 1982. The estimation of animal abun-

dance. C. Griffin & Company, London.* Contributions or chapters of books:Macdonald, D. W. & Johnson, D. P., 2001. Dispersal


* Ph. D. Thesis:Merilä, J., 1996. Genetic and quantitative trait vari

ation in natural bird populations. Ph. D. Thesis, Uppsala University.

* Works in press should only be cited if they have been accepted for publication:Ripoll, M. (in press). The relevance of population

studies to conservation biology: a review. Anim. Biodivers. Conserv. References must be set out in alphabetical and

chronological order for each author, adding the letters a, b, c,... to papers of the same year. Biblio graphic citations in the text must appear in the usual way: "...according to Wemmer (1998)...", "...has been defined by Robinson & Redford (1991)...", "...the pros pec tions that have been carried out (Begon et al., 1999)..." Tables. Tables must be numbered in Arabic numerals with reference in the text. Large tables should be narrow (across the page) and long (down the page) rather than wide and short, so that they can be fitted into the column width of the journal.Figures. All illustrations (graphs, drawings or photographs) must be termed as figures, numbered consecutively in Arabic numerals (1, 2, 3, etc.) and with re ference in the text. Glossy print photographs, if essential, may be included. Colour photographs may be published but its publication will be charged to authors. Maximum size of figures is 15.5 cm width and 24 cm height. Figures will not be tridimen sional. Both maps and drawings must include scale. The preferred shadings are white, black and bold hatching. Avoid stippling, which does not reproduce well. Legends of tables and figures. Legends of tables and figures must be clear, concise, and written both in English and Spanish.

Main headings (Introduction, Material and methods, Results, Discussion, Acknowled ge ments and References) should not be number ed. Do not use more than three levels of headings.

Manuscripts should not exceed 20 pages including figures and tables.

If the article describes new taxa, type material must be deposited in a public institution.

Authors are advised to consult recent issues of the journal and follow its conventions.

VIIAnimal Biodiversity and Conservation 28.2 (2005)

Animal Biodiversity and Conservation Subscription Form

Please enter our subscription to Animal Biodiversity and Conservation 66 e Spain 69 e Europe 76 e rest of world Single use subscription: 21 e Spain 24 e Europe 31 e rest of world Please despatch my issues by air mail (supplement of 6 e for outside Europe)

Please send me the Instructions to authors

Name

Institution

Address

Payment method

International Cheque payable toInstitut de Cultura de Barcelona and drawn against a Spanish bank.

Send cheque by postal mail to: Lluïsa Arroyo Dept. of Scientific Publications Museu de Ciències Naturals de la Ciutadella Psg. Picasso s/n. 08003 Barcelona, Spain Bank Transfer to Caixa d’Estalvis i Pensions de Barcelona ("La Caixa") IBAN: ES03 2100 3000 1422 0170 1046 Swicht code / bic code: CAIX ES BB

Send this order form by postal mail to: Lluïsa Arroyo Dept. of Scientific Publications Museu de Ciències Naturals de la Ciutadella Psg. Picasso s/n 08003 Barcelona, Spain

VIII Animal Biodiversity and Conservation 28.2 (2005)

Welcometo the electronic version of


http://www.bcn.es/ABC

Animal Biodiversity and Conservation joins the recent worldwideOpen Access Initiative

of providing a permanent online version free of charge and access barriers.

This is the result of the growing point of view that open access to research is essential

for efficient and rapid scientific communication.

ABC alert, a free alerting service, provides e–mail information on the latest issue.

To sign on for this service, please send an e–mail to:[email protected]

Recommend

this electronic access

to your library!

http://www.www.museuzoologia.bcn.es/servis/ABCag.htm


Animal Biodiversity and Conservation 28.2 (2005) IX

Arxius de Miscel·lània Zoològica vol. 2 (2004)2004 Museu de Ciències Naturals de la CiutadellaISSN: 1698–0476

Web: http://www.bcn.es/arxiusMZ

Índex/Índice/ContentsFuentes, M. V., Sáez, S., Trelis, M., Muñoz–Antoli, C. & Esteban, J. G., 2004. The helminth community of Apodemus sylvaticus (Rodentia, Muridae) in the Sierra de Gredos (Spain): Arxius de Miscel·lània Zoològica, 2: 1–6.

Abstract The helminth community of Apodemus sylvaticus (Rodentia, Muridae) in the Sierra de Gredos (Spain).— The Spanish mountain range of Gredos was included in the studies conducted on the Iberian peninsula to inves-tigate helminth fauna of small mammals. The helminth community of the wood mouse, Apodemus sylvaticus (Rodentia, Muridae), was analysed. Qualitatively, 13 helminth species were detected: Plagiorchis sp. I and Plagiorchis sp. II (Trematoda); Taenia parva larvae, T. martis larvae, T. taeniaeformis larvae, Rodentolepis straminea and R. fraterna (Cestoda); and Trichuris muris, Heligmosomoides polygyrus, Syphacia stroma, S. frederici, Aspiculuris tetraptera and Rictularia proni (Nematoda). Quantitatively, the highest prevalence (65.0%) and the mean abundance (36.9%) of H. polygyrus stand out. In comparison with the other mountain ranges studied, analysis of the global results demonstrates that the helminth fauna of the host species studied is diverse despite the adverse climatic conditions. This could be related to both the particular ecological characteristics and the appropriate state of preservation of this ecosystem.

Key words: Helminths, Apodemus sylvaticus, Rodentia, Muridae, Sierra de Gredos, Spain.

Bros, V., 2004. Mol·luscs terrestres i d’aigua dolça de la serra de Collserola (Barcelona, NE península Ibèrica). Arxius de Miscel·lània Zoològica, 2: 7–44.

Abstract Land and freshwater molluscs of the Collserola mountains (Barcelona, NE Iberian peninsula).— A malaco-logical survey was made taking into account all the 91 1 x 1 km UTM grid squares that cover the area of the Collserola mountains, Barcelona (north-eastern Iberian peninsula). From 108 sampled localities, 1,261 records of molluscs were obtained. A total of 73 species were identified: seven freshwater species, 11 slugs, and 55 land molluscs. The first discussions on the results and a draft of the initial conclusions are shown. Information on dominant gastropod habitat preference in the ecosystems of Collserola is also provided.

Key words: Mollusca, Continental molluscs, Collserola Park, Faunistics, Monitoring, Bioindicators.

All works are licensed under aCreative Commons Attribution –NonCommercial 3.0 License

http://bcnweb13.bcn.es:81/NASApp/wprmuseuciencies/Museu.GeneradorPagines?idioma=3&seccio=11_2

http://creativecommons.org/licenses/by-nc/2.5/

Animal Biodiversity and Conservation 28.2 (2005) XI

Agradecimiento a los asesoresOur grateful thanks to the referees

El Editor ejecutivo, los Editores, el Consejo editor y el Consejo asesor quieren agradecer a todos los aseso-res su incalculable ayuda en la revisión de los artículos sometidos a Animal Biodiversity and Conservation durante el período 2003 a 2005:

The Executive Editor, the Editors, the Editorial Board and the Advisory Board wish to thank all the referees for their invaluable help in reviewing articles submitted to Animal Biodiversity and Conserva-tion for the period 2003 to 2005:

Allen, C.Alonso, J. C. Arnason, N.Atlegrim, O. Baillie, S.Bairlein, F. Baker, A. S.Balke, M. Bareth, C.Bellés, X.Bewster–Wingard, G. L.Biström, O.Borges, S.Brooks, S. J. Brown, C. R. Burnham, K. P. Cam, E. Cameron, R. A. D.Cárdenas Talaverón, A. M.Carrascal, L. M.Carroll, J. P.Casals, F.Castro, F.Cobo, F.Conroy, M. J.Cooch, E.Cooper, W. E.Cuervo, J. J.Dávalos, L. M.Dhondt, A.Díaz, M.Doherty, P.Eckert, K. L.Estrada, A. Fagan, W. Ferguson L. M.Fernández–Haeger, J. Fernández–Juricic, E.

Ferrer, M. Fiedler, K.Francis, C. M.Fresneda, X.García–Barros, E.Gómez, B. J. González, J.Grim, E. Grossman, G. D. Hines, J.Hopkins, B. Horrocks, H.Illera Cobo, J. C.Jordana, R. Juan, C. Kitahara, M.Klompen, H.Konopacka, A. Krell, F.–T.Laayouni, H. Lebreton, J.–D. Lee, D. C.Lhonoré, J.Machado, A.MacCall, A. MacKenzie, D.Manfrín, L.Maranhao, P. Marco, A. Márquez, F.Martikainen, P.Martin, B. Mendes, L. F.Metcalfe, N. Micó Balaguer, E. Miller, K. B.Mínguez, E.Mischis, C. C. de

Montreuil, O. Munguira, L. M. Negre, B. Nichols, J. D.Nogales, M.Nomakuchi, S.Novoa, F.Ortuño Hernández, V. Palmer Vidal, M. Pérez–Enciso, M. Pérez–Tris, J. Pollock, K. H.Pons, J. Pons, P.Prat Baella, F.Pretus, J. M. Real, R. Rosser, A.Santos Maroño, M. Schmitz García, M. F. Schwarz, C. J.Seoaane Pinilla, J.Serrano Marino, J.Simón Benito, J. C.Teisaire, E. S.Tella, J. L. Templado, J.Thibaud, J. M. Thomson, D. L.Valido, A.Van Veller, M. G. P.Varga, Z.Watanabe, M. White, G. C.Wiklund, C. Yildirim, Z.Zardoya, R.Zhang, Z. G.

Les cites o els abstracts dels articles d’Animal Biodiversity and Conservation es resenyen a /Las citas o los abstracts de los artículos de Animal Biodiversity and Conservation se mencionan en /Animal Biodiversity and Conservation is cited or abstracted in:

Abstracts of Entomology, Agrindex, Animal Behaviour Abstracts, Anthropos, Aquatic Sciences and Fisheries Ab-stracts, Behavioural Biology Abstracts, Biological Abstracts, Biological and Agricultural Abstracts, Current Primate References, Directory of Open Acces Journals (DOAJ), Ecological Abstracts, Ecology Abstracts, Entomology Abstracts, Environmental Abstracts, Environmental Periodical Bibliography, Genetic Abstracts, Geographical Abstracts, índex de Sumaris Electrònics del Consorci de Biblioteques de Catalunya, Índice Español de Ciencia y Tecnología, International Abstracts of Biological Sciences, International Bibliography of Periodical Literature, International Developmental Abstracts, Marine Sciences Contents Tables, Oceanic Abstracts, Recent Ornitho-logical Literature, Red de Revistas Científicas Españolas (REVICIEN), Referatirnyi Zhurnal, Science Abstracts, Serials Directory, Ulrich’s International Periodical Directory, Zoological Records.

Índex / Índice / Contents

Animal Biodiversity and Conservation 28.2 (2005)ISSN 1578–665X

101–119Carrascal, L. M. & Palomino, D.Preferencias de hábitat, densidad y diversi-dad de las comunidades de aves en Tenerife (Islas Canarias) 121–130 Waite, T. A., Vucetich, J., Saurer, T., Kroninger, M., Vaughn, E., Field, K. & Ibargüen, S.Minimizing extinction risk through genetic rescue

131–136Martínez–Abraín, A., Oro, D., Belenguer, R., Ferrís, V. & Velasco, V.Long–term changes in species richness in a small Mediterranean archipelago bird–breeding community

137–147Komonen, A. & Kouki, J.Occurrence and abundance of fungus–dwelling beetles (Ciidae) in boreal forests and clearcuts: habitat associations at two spatial scales

149–157Simón Benito, J. C., Espantaleón, D. & García–Barros, E. Stachorutes cabagnerensis n. sp., Collem-bola (Neanuridae) from Central Spain, and a preliminary approach to phylogeny of genus

159–168Martin, C. S. Jeffers, J., Godley, B. J.The status of marine turtles in Montserrat (Eastern Caribbean)

169–179Hilaluddin, Kaul, R. & Ghose, D.Conservation implications of wild animal biomass extractions in Northeast India

181–188Örstan, A., Pearce, T. A., Welter–Schultes, F.Land snail diversity in a threatened limes-tone district near Istanbul, Turkey

189–204Pradel, R., Gimenez, O. & Lebreton, J.–D.Principles and interest of GOF tests for multistates capture–recapture models

205–206A la memoria de Xavier Domingo–Roura (1964–2005)

207–208In memoriam: Xavier Domingo –Roure (1964–2005)

IXAbstracts del volumen 2 (2004) de Arxius de Miscel·lània ZoològicaAbstracts of volume 2 (2004) of Arxius de Miscel·lània Zoològica

animal biodiversity and conservation issue 28.2 (2005)

Documents

autnoma de barcelona

ciencias naturalescsic

mar cmimacsic

carrascal museo nacional

lobo museo nacional

lpezgonzlez univ de

of glasgow

of georgia