The Condor 105:428±438 q The Cooper Ornithological Society 2003 VARIATION IN MORPHOLOGY AND MATING SYSTEM AMONG ISLAND POPULATIONS OF GALA PAGOS HAWKS JENNIFER L. BOLLMER1,5,6,TANIA SANCHEZ2,MICHELLE DONAGHY CANNON3,7, DIDIER SANCHEZ2,BRIAN CANNON3,8,JAMES C. BEDNARZ3,TJITTE DE VRIES2, M. SUSANA STRUVE4,9 AND PATRICIA G. PARKER1,6 1Department of Evolution, Ecology, and Organismal Biology, The Ohio State University, 1735 Neil Avenue, Columbus, OH 43210 and Department of Biology, University of Missouri-St. Louis, 8001 Natural Bridge Road, St. Louis, MO 63121 2Departamento de BiologõÂa, PontifõÂcia Universidad CatoÂlica del Ecuador, Quito, Ecuador 3Department of Biological Sciences, Arkansas State University, P.O. Box 599, State University, AR 72467 4Charles Darwin Foundation, Inc., 407 N. Washington Street, Suite 105, Falls Church, VA 22046 Abstract. Interspeci®c variation in sexual size dimorphism has commonly been attributed to variation in social mating system, with dimorphism increasing as intrasexual competition for mates increases. In birds, overall body size has also been found to correlate positively with size dimorphism. In this study, we describe variation in morphology and mating system across six populations of the endemic GalaÂpagos Hawk (Buteo galapagoensis). GalaÂpagos Hawks exhibit cooperative polyandry, a mating system in which long-term social groups contain a single female and multiple males. Comparisons among islands revealed signi®cant differences in overall body size for both adults and immatures. Populations ranged from completely monogamous to completely polyandrous, with varying mean group sizes. Data did not support our prediction that sexual size dimorphism would increase with the degree of polyandry (number of males per group) or with body size; there was no correlation between mating system and sexual dimorphism. We did ®nd a signi®cant negative relation- ship between degree of polyandry and body size among islands, opposite of the pattern predicted. Key words: body size, Buteo galapagoensis, cooperative polyandry, GalaÂpagos Hawk, principal components analysis, sexual size dimorphism. VariacioÂn en MorfologõÂa y Sistema de Apareamiento entre Poblaciones de Buteo galapagoensis Resumen. VariacioÂn interespecõ®ca en dimor®smo sexual ha sido atribuõÂda comuÂnmente a variaciones del sistema social de apareamiento, de tal manera que el dimor®smo aumenta conforme aumenta la competencia intrasexual por parejas reproductivas. TambieÂn se ha encontrado que el tamanÄo corporal se correlaciona positivamente con el dimor®smo. En este estudio describimos la variacioÂn morfoloÂgica y el grado de poliandrõÂa de seis poblaciones de Buteo galapagoensis, una especie que exhibe un sistema de apareamiento denominado poliandrõÂa cooperativa. En este sistema los grupos de individuos reproductivos incluyen una sola hembra y muÂltiples machos. Se comprobo que existen diferencias signi®cativas en el tamanÄo del cuerpo de adultos y juveniles entre islas. Las poblaciones muestreadas variaron entre monoÂgamas y completamente poliaÂndricas, y el tamanÄo promedio de los grupos fue variable. Los datos no apoyaron las predicciones establecidas inicialmente pues el grado de dimor®smo sexual no aumento con el nivel de poliandrõÂa (nuÂmero de machos por grupo) ni con el tamanÄo corporal, ni hubo una correlacioÂn entre el sistema de apareamiento y el dimor®smo sexual. La relacioÂn entre el tamanÄo corporal y el sistema de apareamiento fue contraria a la que se predijo: hubo una correlacioÂn negativa signi®cativa entre el grado de poliandrõÂa y el tamanÄo corporal entre islas. Manuscript received 25 April 2002; accepted 20 March 2003. 5 E-mail: [email protected] 6 Present address: Department of Biology, University of Missouri-St. Louis, 8001 Natural Bridge Road, St. Louis, MO 63121. 7 Present address: Department of Forest Science, Oregon State University, 321 Richardson Hall, Corvallis, OR 97331-5752. 8 Present address: Oregon Department of Fish and Wildlife, Corvallis Research Lab, 28655 Highway 34, Corvallis, OR 97333. 9 Present address: CH2M Hill, 13921 Park Center Road, Herndon, VA 20171. [428] MORPHOLOGY AND MATING SYSTEM IN GALA PAGOS HAWKS 429 INTRODUCTION Morphological variation among species or pop- ulations is driven by a number of factors, one of which is sexual selection. Interspeci®c variation in sexual size dimorphism has commonly been attributed to variation in social mating system and differences in parental care (Darwin 1871, Andersson 1994). The high levels of intrasexual competition for mates characteristic of polyga- mous systems may result in large body size in the competitive sex. Increased sexual size di- morphism was found to be associated with in- creased polygyny in the New World blackbirds (Icterinae; Webster 1992) and increased polyg- FIGURE 1. Map of the GalaÂpagos archipelago show- yny and polyandry (reverse size dimorphism) ing islands where we acquired morphological mea- within the Charadrii (SzeÂkely et al. 2000). surements and group sizes for GalaÂpagos Hawks: Vol- Among 73 species, Owens and Hartley (1998) can Alcedo on Isabela, Santiago, Pinta, Marchena, found a signi®cant positive relationship between Santa Fe, and EspanÄola. The entire range of the Ga- laÂpagos Hawk is limited to nine islands: the six men- sexual size dimorphism in body mass and degree tioned above, Santa Cruz, PinzoÂn, and Fernandina. of polygamy, while in a larger analysis (n 5 1031 species), Dunn et al. (2001) found a similar relationship, with polygynous and lek species size dimorphism disappeared after controlling being more dimorphic than monogamous spe- for mating system, which Webster interpreted to cies. mean that body size affects dimorphism indi- Across taxa, species with greater size dimor- rectly through its effects on mating system. In phism also tend to be larger in overall body size. their study, Dunn et al. (2001) also found a pos- Webster (1992) reviewed 11 hypotheses that itive relationship between sexual size dimor- have been proposed to explain this relationship, phism and body size. all of which are based on males competing for In this study, we look at the relationship be- females in polygynous systems. Four of these tween morphology and mating system in the Ga- ideas are based on intrasexual competition as the laÂpagos Hawk (Buteo galapagoensis), which is cause of sexual size dimorphism. They predict endemic to the GalaÂpagos archipelago (Fig. 1). that the association with body size results from In addition to forming pairs, the GalaÂpagos polygyny being more likely among large-bodied Hawk exhibits cooperative polyandry, in which species, larger females being favored as better one female mates with up to eight males (usually competitors, genetic correlations increasing fe- two or three), and all aid in the care of a single male body size as male size increases, or the brood (Faaborg and Patterson 1981, DeLay et al. change in male body size being greater in larger 1996). Group members are unrelated (Faaborg species. Other hypotheses predict that while sex- et al. 1995) and highly territorial, defending ual selection drives dimorphism, larger species year-round, all-purpose territories (de Vries may have fewer ecological or energetic con- 1975). Observations of marked birds indicate straints on body size, or only the female is con- that group membership is stable across years, strained to a smaller body size. Still other hy- especially for males (Faaborg et al. 1980, Do- potheses argue that larger species may be more naghy Cannon 2001). dimorphic because of increased niche partition- Immatures are not territorial, and live in the ing, male predator defense, differences in ge- ¯oater population until they are at least three netic variation among the sexes, or allometric years old and able to secure a position in a ter- growth, with sexual selection having no effect. ritorial group. In addition to immatures, non- Among the New World blackbirds, Webster breeding adult-plumaged hawks also occur in (1992) found that sexual size dimorphism cor- the ¯oater population. On Santiago, adult male related with body size but less strongly so than ¯oaters are rarely seen; however, adult females with polygyny. The effect of body size on sexual account for a sizeable proportion of the ¯oaters 430 JENNIFER L. BOLLMER ET AL. (;17%; Donaghy Cannon 2001). Our observa- Hawks were caught on six islands (Fig. 1): Santa tions between 1998 and 2002 suggest that com- Fe (n 5 23), EspanÄola (19), Santiago (223), Vol- petition among females for breeding spots may can Alcedo on Isabela (91), Pinta (12), and Mar- be high: (1) there is higher turnover among ter- chena (25). The hawks were caught using two ritorial females than males (Donaghy Cannon methods: a bal-chatri trap (Berger and Mueller 2001); (2) four new territories carved out of area 1959), baited with live prey such as a rat (Rattus defended by existing groups were established by spp.), and a rope noose on a stick (Faaborg et ¯oater females and 1±3 males, never by males al. 1980). Each hawk was banded with either an alone (Donaghy Cannon 2001); and (3) we have aluminum or anodized color band bearing an al- observed ¯oater females ®ghting each other phanumeric code, or with both types of bands. (JLB, pers. obs.). The following morphological measurements Island populations tend to be small and iso- were taken: wing chord (un¯attened, to the near- lated, which fosters genetic and phenotypic dif- est mm), tail length (posterior base of uropygial ferentiation among them