FEATURE ARTICLES

The Condor 106:449±456 ᭧ The Cooper Ornithological Society 2004

CONSERVATION PRIORITIES FOR RESPLENDENT BASED ON ANALYSIS OF MITOCHONDRIAL DNA CONTROL- REGION SEQUENCES

SOFIA SOLOÂ RZANO1,3,ALLAN J. BAKER2 AND KEN OYAMA1 1Centro de Investigaciones en Ecosistemas, UNAM, Antigua carretera a Patzcuaro 8701, San Jose de la Huerta, Morelia, 58190, Michoacan, Mexico 2Centre for Biodiversity and Conservation Biology, Royal Ontario Museum, 100 Queen's Park, Toronto, ON M5S 2C6, Canada

Abstract. The Resplendent ( mocinno) is a threatened classi®ed into two putative subspecies (P. m. mocinno and P. m. costaricensis) and distrib- uted in cloud forests of seven countries in Mesoamerica. Because the are rare, tissue samples are dif®cult to obtain, but we analyzed genetic diversity in 25 quetzals from ®ve countries based on 255 bp of domain I of the control region of mitochondrial DNA. Eight haplotypes were detected. Nucleotide diversity for Mexico (P. m. mocinno: 0.0021) and Panama (P. m. costaricensis: 0.0026) were low, and did not differ from the values estimated for other birds species irrespective of whether they were endangered. A haplotype tree rooted with the Pavonine Quetzal (P. pavoninus) recovered two reciprocally monophyletic clades corresponding to each subspecies, so we propose that each subspecies be considered as an evolutionarily signi®cant unit for conservation planning. A minimum spanning network showed the number of genetic differences separating haplotypes within subspecies was small relative to the number of substitutions among them, indicating strong population subdivision (FST ϭ 0.37). In spite of the limited sampling we propose that in conservation practice Mexico±Guatemala, Nicaragua, El Salvador, and Panama be considered preliminarily as independent conservation management units since they each have unique haplotypes. Ad- ditionally, these countries should construct international agreements to protect the natural vegetation corridors among cloud forests of Mesoamerica and to curtail the illegal trade of quetzals. Key words: evolutionarily signi®cant units, Mesoamerica, mitochondrial DNA, manage- ment units, Pharomachrus mocinno.

Prioridades de ConservacioÂn para el Quetzal Basadas en el AnaÂlisis de la RegioÂn Control del ADN Mitocondrial Resumen. El quetzal (Pharomachrus mocinno) es una especie de ave amenazada clasi- ®cada en dos subespecies (P. m. mocinno y P. m. costaricensis) distribuidas en los bosques de niebla de siete paõÂses de MesoameÂrica. Debido a que eÂsta es un ave rara, las muestras de tejido son difõÂciles de obtener, pero pudimos analizar la diversidad geneÂtica en 255 pb del dominio I de la regioÂn control del ADN mitocondrial en 25 quetzales procedentes de cinco paõÂses. Se encontraron ocho haplotipos. La diversidad nucleotõÂdica para MeÂxico (P. m. mocinno: 0.0021) y PanamaÂ(P. m. costaricensis: 0.0026) fue baja, pero no di®ere de la estimada para otras especies de aves amenazadas o no amenazadas. El aÂrbol de haplotipos enraizado con P. pavoninus mostro dos clados recõÂprocamente mono®leÂticos, correspon-

Manuscript received 10 September 2003; accepted 15 April 2004. 3 Present address: Apartado Postal 21-041, C.P. 04021, MeÂxico D. F., Mexico. E-mail: slso®[email protected]. unam.mx

[449] 450 SOFIA SOLOÂ RZANO ET AL.

diendo cada uno a cada subespecie, por lo que proponemos que para planes de conservacioÂn cada subespecie sea considerada como unidad evolutiva signi®cativa independiente. Una red de distancias mõÂnimas mostro que el nuÂmero de diferencias geneÂticas que separa a los haplotipos dentro de las subespecies fue pequenÄo con respecto al nuÂmero de sustituciones que existe entre ellas, indicando una fuerte divisioÂn poblacional (FST ϭ 0.37). Considerando nuestro muestreo limitado proponemos que para ®nes de conservacioÂn praÂcticos MeÂxico± Guatemala, Nicaragua, El Salvador y Panama sean considerados preliminarmente como unidades de manejo independientes ya que eÂstos presentan haplotipos uÂnicos no compartidos entre localidades. AdemaÂs, estos paõÂses deberõÂan ®rmar acuerdos internacionales para pro- teger los corredores de vegetacioÂn naturales entre los bosques de niebla de MesoameÂrica y tratar de reducir el comercio ilegal de los quetzales.

INTRODUCTION reproductive behavior (e.g., courtship, chick The worldwide distribution of the Resplendent rearing), breeding season (dry season), diet, and Quetzal (Pharomachrus mocinno) ranges from vocalizations appear not to differ between sub- southern Mexico across Guatemala, Honduras, species (SS, pers. obs.). El Salvador, Nicaragua, Costa Rica, and Pana- Although the subspecies concept is arti®cial, ma. However, the distribution of quetzals is not it has practical advantages in conservation be- continuous through these countries, but is deter- cause it allows us to identify conservation pri- mined by the patchiness of cloud forests (SoloÂr- orities within species (Ryder 1986). The con- zano et al. 2003). These forests, ranging from cepts of ``evolutionarily signi®cant units'' and 1300 to 3000 m elevation, represent the breeding ``management units'' provide useful tools in ge- habitats for this species (Stotz et al. 1996). netic conservation (Ryder 1986, Moritz 1994a, Presently, the is classi- 1994b, Frankham et al. 2002). The intent of evo- ®ed under the IUCN Red List in the category of lutionarily signi®cant units is to preserve evo- ``Lower Risk Near Threatened'' (IUCN 2002), lutionary heritage across taxa assuming a histor- but this classi®cation should be revaluated con- ical origin for the genetic differences between sidering that the species is limited to remnant them. This concept emphasizes the phylogeo- patches of evergreen cloud forests, most popu- graphic pattern of genetic variation rather than lations are geographically isolated and small, degree of genetic divergence or the quantity of and breeding habitats are being lost at high an- genetic diversity pooled in a certain population. nual rates (Hanson 1982, SoloÂrzano et al. 2003, Management units represent populations with P. Bubb, unpubl. data). These habitats are them- signi®cant divergence in allele frequencies, but selves conservation priorities because they har- which have not achieved reciprocal monophyly bor high biological and ecological diversity as required for evolutionarily signi®cant units (Hamilton et al. 1995, Nadkarni and Wheel- (Moritz 1994a). The main difference between wright 2000, Bubb 2001). evolutionarily signi®cant units and management Based on morphological characters, two sub- units is that evolutionarily signi®cant units are species of Resplendent Quetzals have been rec- concerned with historical population structure ognized: P. m. mocinno, distributed from south- and serve long-term conservation planning, ern Mexico to northern Nicaragua, and P. m. whereas management units focus on the current costaricensis, found in Costa Rica and western population structure and are useful to establish Panama (Sibley and Monroe 1990). These sub- short-term monitoring programs based on ge- species apparently are isolated geographically netic discontinuity among populations or unique by Lake Nicaragua (8324 km2) and the absence haplotypes harbored in individual populations of the breeding habits in regions adjacent to this (Moritz 1994a, 1994b). The criteria to de®ne lake (Fig. 1). The age of this barrier is not well these concepts have changed from the initial established but probably arose when the Pana- proposals (cf. Ryder 1986, Moritz 1994a, manian Isthmus formed 3 million years ago 1994b), and are still debated (Fraser and Ber- (Keigwin 1982). Although preliminary partial natchez 2001). genetic analysis based on a subunit of ND6, In the present study, our goal is to identify tRNAGlu and control region of mtDNA of 10 in- conservation units for quetzals in Mesoamerica dividuals suggested genetic differentiation of based on mtDNA haplotypes. We discuss our re- subspecies (SoloÂrzano et al., unpubl. data), the sults in the context of genetic conservation rath- QUETZAL CONSERVATION BASED ON MTDNA 451

FIGURE 1. Current distribution (symbols) of Resplendent Quetzals in geographically isolated remnant cloud forests (constructed from Hanson 1981, Sibley and Monroe 1990, SoloÂrzano et al. 2003). The past distribution is indicated by bold lines (modi®ed from Johnsgard 2000). Triangles correspond to P. mocinno mocinno and circles to P. m. costaricensis. Sampled localities: Mexico: NM: Northern Mountains, ET: El Triunfo biosphere reserve, FC: Finca Santa Cruz; Guatemala: BQ: Biotopo Quetzal, SM: Sierra de las Minas biosphere reserve; El Salvador: PM: Parque Montecristo; Nicaragua: PA: Parque Arenal; Panama: BG: Bajo Grande. One sample was not located because its origin is uncertain (ZOOMAT Zoo donation). The outgroup sample PP (P. pavoninus) is from (inset).

er than a structural analysis of the control region emy of Natural Sciences of Philadelphia, Penn- variation and composition, which will be pub- sylvania. lished elsewhere (SoloÂrzano et al., unpubl. data). DNA ISOLATION, PCR AMPLIFICATION, AND METHODS SEQUENCING We conducted ®eldwork in eight localities (Fig. Total genomic DNA was isolated from blood 1), collecting blood samples during the breeding and tissue samples using standard proteinase-K season to guarantee the origin of the samples. and SDS treatment, followed by extraction with Additionally, the ZOOMAT Zoo in Chiapas, phenol-chloroform and a ®nal ethanol precipita- Mexico, made two blood samples of P. m. mo- tion (Sambrook et al. 1989). We ampli®ed 255 cinno available. Although the origin of one of bp from the hypervariable domain I of the these birds is uncertain because the adult male mtDNA control region using primers ND62LQTF was con®scated, the trader reported it came from and H438 (SoloÂrzano et al., unpubl. data). PCR Guatemala. The other one is from Finca Santa ampli®cation reactions contained 0.1 ␮M of each

Cruz, Mexico (Fig. 1). All 25 samples were primer, 100 ␮M of each dNTP, 2.0 mM MgCl2, stored at ambient temperature in buffer solution 10 mM Tris-HCl pH 8.3, 50 mM KCl, 0.5 units (Hillis et al. 1996) before being transported to TaqPol (Boehringer, Ingelheim, Germany) and 1 the laboratory and stored at Ϫ70ЊC. ␮L of total genomic DNA. The samples were The two tissue samples for the outgroup, the sequenced with 33P-ATP using a Thermoseque- Pavonine Quetzal (P. pavoninus), originated in nase kit (Amersham Biosciences, Chalfont St. Ecuador (Fig. 1) and were donated by the Acad- Giles, UK) according to the manufacturer's in- 452 SOFIA SOLOÂ RZANO ET AL.

TABLE 1. Haplotypes identi®ed from the hypervariable domain of the control region of mtDNA in two subspecies and 25 individuals of the Resplendent Quetzal sampled across its range in Mesoamerica. Refer to Figure 1 for sampling locations. ``Zoo'' refers to a quetzal at ZOOMAT con®scated from an unknown location. Position is read vertically; dots indicate similarity with haplotype G.

Sampling location Position North- Finca Sierra Parque 11111 Haplo- ern El Tri- Santa Biotopo de las Monte- Parque Bajo 12334622347 type Mtns. unfo Cruz Quetzal Minas cristo Arenal Grande Zoo 53024145072 Gb 1 CTCACAGATCA Aa 131 1 1 ....TGAGCAT Ba 5 1 2 ...GTGAGCAT Ca 1 .CT.TGAGCAT Da 1 ..T.T.AGCAT Eb 6 ...... A. Fb 1 T...... A. Hb 1 .C...... A. a Identi®ed in P. m. mocinno. b Identi®ed in P. m. costaricensis.

structions. The sequencing reactions were elec- and the transition:transversion ratio (R) were es- trophoresed for 2 hr on a denaturing (6M urea) timated for the 25 samples (Kumar et al. 2001). 6% polyacrylamide gel at a constant 70 W. The Correspondence between genetic variation gel was vacuum-dried at 80ЊC for 1 hr and then and geographic distribution was tested with a exposed to X-ray ®lm for 48 hr. Since P. mo- Mantel test (Liedloff 1999) by computing cor- cinno has been reported to be heteroplasmic for relations between pairwise genetic distances the control region (SoloÂrzano et al., unpubl. (Kimura 2-parameter) and geographic separation data) we used only a portion of the sequence (km; georeferenced to UTM units with Arc- with no observable double peaks. When the am- View-GIS; ESRI 1996) under assumptions of an pli®ed product of the outgroup (P. pavoninus) isolation by distance model. We constructed a was sequenced, double peaks appeared in do- neighbor-joining (NJ) tree rooted with P. pavo- main I but the ¯anking NADH6 gene showed ninus to estimate the phylogenetic relationships only a single sequence. To resolve this problem among haplotypes of both subspecies. Tree ro- the PCR products from the outgroup were bustness at the nodes was estimated by bootstrap cloned into pGEM௡-T Easy Vector System ac- analysis (Kumar et al. 2001). cording to the manufacturers recommendations We estimated the degree of genetic differen-

(Promega, Madison, Wisconsin). The clones tiation between subspecies (FST), and derived a were sequenced with BigDye௡ Terminator V3.1 minimum spanning network from the genetic cycle sequencing kit (Applied Biosystems, Fos- distances estimated in Arlequin 2.0 (Schneider ter City, California) using an ABI Prism 310 Ge- et al. 2000). netic Analyzer. Five cloned products were se- quenced, and after verifying that each had the RESULTS same control-region sequence they were used in GENETIC DIVERSITY AND NUCLEOTIDE subsequent genetic analyses. COMPOSITION The sequences of 255 bp of mitochondrial DNA GENETIC ANALYSIS control region from 25 specimens of P. mocinno After aligning sequences with Clustal X representing two subspecies (P. m. mocinno and (Thompson et al. 1997), overall haplotypic (h) P. m. costaricensis) revealed eight haplotypes and nucleotide diversity per site (␲) were cal- (GenBank accession numbers AY607565± culated (Rozas and Rozas 1999) for the 25 sam- AY607572) based on 12 segregating sites, 10 of ples as a group (regardless of their origin), as which were transitions and two transversions well as individually for Mexico (8 samples) and (Table 1). The homologous segment obtained for Panama (9 samples). Nucleotide composition P. pavoninus differed from P. mocinno at 52 QUETZAL CONSERVATION BASED ON MTDNA 453

FIGURE 2. Genealogical relationships among control-region haplotypes of Resplendent Quetzals. The neigh- bor-joining tree was constructed using the Kimura 2-parameter model of substitution, and was rooted with a sequence from Pavonine Quetzal. Bootstrap values shown at the nodes were computed with 1000 replicates of the data. Localities of the haplotypes are in parentheses. sites. Average nucleotide composition of the 25 DISTRIBUTION OF GENETIC DIVERSITY Resplendent Quetzal sequences exhibited a high The correlation between geographical and ge- proportion of adenine residues (36%), followed netic distances was consistent with a model of by cytosine (25%), thymine (23%), and guanine isolation by distance (r2 ϭ 0.81, P ϭ 0.003). (16%). Nucleotide diversity for the P. mocinno However, the largest genetic distance (0.0406) samples as a group was ␲ϭ0.0171 Ϯ 0.0015, was between haplotypes F from Panama and C with a haplotype diversity (h) of 0.8180 Ϯ from Nicaragua. 0.0470. However, as expected these parameters The single population sampled in Panama were lower when Mexico (n ϭ 8, ␲ϭ0.0021 (Fig. 1) had four private (unique) haplotypes (E, Ϯ 0.0005, h ϭ 0.5360 Ϯ 0.1230), and Panama F, G, and H), whereas the three populations sam- (n ϭ 9, ␲ϭ0.0026 Ϯ 0.0009, h ϭ 0.5830 Ϯ pled in Mexico had two haplotypes (A and B) 0.1830) were treated separately. that were shared with birds sampled from Bio- topo Quetzal and Sierra de las Minas, Guatemala (Table 1). Haplotypes C and D, from Nicaragua and El Salvador respectively, were unique to each of these populations. The NJ tree showed a clear split between the putative subspecies P. m. mocinno (86% bootstrap support) and P. m. costaricensis (97% bootstrap support), and a shallow separation for haplotypes from different populations within the P. m. mocinno range (Fig. 2). The minimum spanning network (Fig. 3) il- FIGURE 3. Minimum spanning network showing mutational differences among the eight haplotypes lustrated the small number of genetic differences found in Resplendent Quetzals representing the sub- separating haplotypes within subspecies, and the species P. m. mocinno (triangles) and P. m. costari- larger number of substitutions among them. Ge- censis (circles). The size of each symbol is propor- netic distances among haplotypes in different tional to haplotype frequency, and the lengths of the subspecies were 3±10 times higher than between branches are proportional to the genetic distance be- tween haplotypes. The solid and dotted crossmarks haplotypes within each subspecies. Taking hap- represent the number of transitions and transversions, lotype frequencies into account this translates to respectively, between haplotypes. an FST value of 0.37 among subspecies. 454 SOFIA SOLOÂ RZANO ET AL.

DISCUSSION graphic distance. This suggests that gene ¯ow Overall nucleotide diversity (0.0171) in the Re- among populations occurs primarily among ad- splendent Quetzal in Mesoamerica was similar jacent populations and is constrained as distance increases. The clear geographic differentiation to that estimated for some endangered species between subspecies implies that they are largely (e.g., cheetahs [Acinonyx jubatus], ␲ϭ0.0130, isolated genetically, but this needs to be con- Freeman et al. 2001), but higher than for others ®rmed in future studies by sampling more ex- (e.g., Florida Grasshopper Sparrow [Ammodra- tensively in their zone of contact. Radio-telem- mus savannarum ¯oridanus], ␲ϭ0.0060± etry studies (Powell and Bjork 1994, 1995) 0.0130, Bulgin et al. 2003). However, as the val- showed that an individual quetzal can traverse ues obtained for Mexico (P. m. mocinno) and 50 km during its seasonal migrations and can Panama (P. m. costaricensis) were similar to visit other breeding areas and lower forests, but those recorded in other bird species irrespective all individuals tagged returned to their past of whether they were endangered or not (Mar- breeding locality. However, all of these tagged shall and Baker 1997, Fry and Zink 1998, Kvist individuals were caught as reproductive adults, et al. 1999, Griswold and Baker 2002), it is clear so the extent of natal philopatry is unknown. that a species endangerment level cannot be de- Regarding the rapid destruction and fragmenta- rived simply from measures of genetic diversity. tion of breeding habitats it is clear that conser- The most obvious reason for this is that this es- vation programs for quetzals and their habitats timator is a product of the evolutionary history must consider this distance as a minimum ac- of lineages studied (Nei 1987) and probably ceptable to avoid increased genetic fragmenta- does not re¯ect recent ecological processes. In tion of populations. In the present study, dis- Resplendent Quetzals, the value of ␲, like the tances less than 50 km occurred between El minimum spanning network, illustrates the small Triunfo and Finca Santa Cruz (32 km) and Sierra number of genetic differences separating haplo- de las Minas and Biotopo Quetzal (34 km), and types within subspecies, and the larger number this is consistent with the observation that pop- of substitutions among them probably re¯ects ulations there shared haplotypes. However, pop- genetic diversity related to historically larger ulations sampled in Guatemala and Mexico, sep- population size that declined recently due to loss arated by more than 300 km, also shared hap- of habitat. Although data are sparse about the lotypes. Unless long-distance migration occurs, historical and recent changes of quetzal distri- this suggests that recent destruction of quetzal bution in Mesoamerica, between 1970 and 2000 habitats that provided connectivity between 2 about 661 km of cloud forests were destroyed these two areas has only recently caused geo- in southern Mexico (SoloÂrzano et al. 2003), graphical separation and the genetic effects have causing the loss of 80% of the Chiapas nesting not yet become apparent. Another possible ex- habitat. Moreover, the current abundance of planation is that the sequences we utilized do not quetzals is very low (0.5±1.5 individuals per ha, provide suf®cient resolution to track microgeo- SoloÂrzano et al. 2003). Considering that frag- graphic structuring of populations, and thus lon- mentation and destruction of forest is a common ger sequences of the complete control region problem in Mesoamerica (Bubb 2001) it is likely may be needed in future studies. that local extinctions and a general decline in In the present study, we obtained very few abundance has occurred. Analyses over genera- blood samples mostly because of the rarity of tional time scales or comparisons with a histor- quetzals and the high rates of both nest poaching ical genetic diversity baseline would be required and nest predation. These factors limited our to distinguish trends in loss of genetic variability analyses, and the identi®cation of genetic pri- and its association with an increased risk of ex- orities consequently suffered. However, since tinction. While this is not yet possible due to these are the ®rst genetic results for quetzals we lack of data, the present study can act as such a make the following preliminary recommenda- baseline for future periodic monitoring of ge- tions. Because the mtDNA control-region se- netic parameters and assessment of extinction quences show that P. m. mocinno and P. m. cos- risks for the species. taricensis exhibit reciprocal monophyly, each Our results indicate that there is geographical subspecies should be managed as two indepen- structuring of populations with increasing geo- dent evolutionarily signi®cant units (Moritz QUETZAL CONSERVATION BASED ON MTDNA 455

1994b). Therefore, Mexico, Guatemala, Hondu- and Nicaragua. We thank O. Haddrath, ROM, L. MaÂr- ras, El Salvador, and Nicaragua should join ef- quez-Valdelamar, IB-UNAM, and M. Renteria and R. Salas, IE-UNAM for laboratory assistance. M. Castillo forts in the task of conservation of P. m. mocin- and D. Diaz, LAIGE, Ecosur, Chiapas, drafted Figure no regardless of political borders. The same is 1. The comments of A. Abreu-Grobois, J. L. Osorno, required of Costa Rica and Panama to preserve two anonymous reviewers, and Editor D. S. Dobkin P. m. costaricensis. improved our manuscript. Personnel of El Triunfo Bio- With the small sample sizes available at pre- sphere Reserve, Sierra de las Minas, Biotopo Quetzal, CECON, Defensores de la Naturaleza, Parque Nacio- sent it is dif®cult to de®ne management units, nal Arenal, and Parque Nacional Montecristo kindly which require genetic distinctiveness of lineag- provided the research permissions. Cocibolca Foun- es, though not at the level of reciprocal mono- dation, K. Aparicio, G. LoÂpez, A. Oliveras, G. Montes, phyly. The shallow branches within each of the W. RodrõÂguez, and Universidad Nacional de El Sal- vador helped SS obtain CITES and sampling permis- subspecies assemblages of haplotypes suggests sions. SEMARNAT, MeÂxico; MARENA, Nicaragua; that different populations of P. mocinno have Secretaria de Agricultura y Ganaderia, El Salvador; been fragmented relatively recently, as haplo- and CONAP, Guatemala, authorized the sampling. types in either assemblage differ by only 1±2 ZOOMAT Zoo, Chiapas, donated two quetzal blood substitutions. More extensive sampling is re- samples, and the Academy of Natural Sciences of Phil- adelphia provided the tissue sample of P. pavoninus quired to determine if the localities (countries) (5732 and 5485). Molecular research on P. mocinno with unique haplotypes represent discrete genet- was funded by an NSERC grant to AJB. SS received ic units. In spite of the very limited sampling grant support from WWF-Corredor BioloÂgico Mesoa- within P. mocinno, populations in Mexico±Gua- mericano (PT37), Idea Wild, CONACyT (93860), Pos- temala, El Salvador, Nicaragua, and Panama grado C. BiomeÂdicas, PAEP (202343) UNAM, and the Chapman Fund from the American Museum of Natural might be preliminarily separate management History, New York. units, as we found haplotypes were not shared among these countries. Therefore, concerted in- LITERATURE CITED ternational efforts may be much better for the long-term conservation of a widely distributed BUBB, P. 2001. Desarrollo de una base de datos para los bosques nublados del NeotroÂpico, p. 51±62. In species that spans international borders. M. Kappelle and A. D. Brown [EDS.], Bosques nu- The present study con®rms the need for sig- blados del NeotroÂpico. Instituto Nacional de Bio- ni®cant increases in efforts to increase the diversidad, Santo Domingo de Heredia, Costa chance of long-term survival of quetzals in Cos- Rica. ta Rica and Panama (for P. m. costaricensis) and BULGIN, N. L., H. L. GIBBS,P.VICKERY, AND A. J. BAKER. 2003. Ancestral polymorphism in genetic Nicaragua, El Salvador, and Guatemala±Mexico markers obscure detection of evolutionarily dis- (for P. m. mocinno). Furthermore, as this study tinct populations in the endangered Florida Grass- shows that quetzal populations in each of these hopper Sparrow (Ammodramus savannarum ¯o- countries have low but apparently unique genet- ridanus). Molecular Ecology 12:831±844. ic diversity they should also be preserved under ESRI. 1996. Using ArcView-GIS. Environmental Sys- tems Research Institute Inc., Redlands, CA. national and international agreements. This con- FRANKHAM, R., J. D. BALLOU, AND D. A. BRISCOE. servation task should also include the monitor- 2002. Introduction to conservation genetics. Cam- ing, protection, and restoration of quetzals and bridge University Press, Cambridge, UK. their breeding and migration habitats. Finally, FRASER,D.J.,AND L. BERNATCHEZ. 2001. Adaptive the activities of wild-bird dealers who buy dead evolutionary conservation: towards a uni®ed con- cept for de®ning conservation units. Molecular and living quetzals for sale to private collec- Ecology 10:2741±2752. tions, aviculturists, and zoos, must be curbed FREEMAN, A. R., D. E. MACHUGH,S.MCKEOWN,C. signi®cantly or eliminated because of the threat WALZER,D.J.MCCONNELL, AND D. G. BRADLEY. they pose to the quetzal's survival. This illegal 2001. Sequence variation in the mitochondrial DNA control region of wild African cheetahs trade has already impacted wild populations and (Acinonyx jubatus). Heredity 86:355±362. has exacerbated the decline in numbers of this FRY,A.J.,AND R. M. ZINK. 1998. Geographical anal- rare species (Johnsgard 2000). ysis of nucleotide diversity and Song Sparrow (Aves: Emberizidae) population history. Molecu- ACKNOWLEDGMENTS lar Ecology 7:1303±1313. GRISWOLD, C. K., AND A. J. BAKER. 2002. Time to the We thank M. Grosselet and B. Granados for ®eld as- most recent common ancestor and divergence sistance. M. PeÂrez, O. Chassin, C. Romero and J. M. times of populations of Common Chaf®nches Zolotoff drove long distances in Mexico, Guatemala, (Fringilla coelebs) in Europe and North Africa: 456 SOFIA SOLOÂ RZANO ET AL.

insights into Pleistocene refugia and current levels NEI, M. 1987. Molecular evolutionary genetics. Co- of migration. Evolution 56:143±153. lumbia University Press, New York. HAMILTON, L. S., J. O. JUVIK, AND F. N. SCATENA. 1995. POWELL,V.N.G.,AND R. D. BJORK. 1994. Implications Tropical montane cloud forests. Springer-Verlag, of altitudinal migration for conservation strategies New York. to protect tropical biodiversity: a case study of the HANSON, D. A. 1982. Distribution of the quetzals in quetzal Pharomachrus mocinno at Monteverde, Honduras. Auk 99:385. Costa Rica. Bird Conservation International 4: HILLIS, D. M., C. MORITZ, AND B. K. MABLE. 1996. 243±255. Molecular systematics. 2nd ed. Sinauer Associates POWELL,V.N.G.,AND R. D. BJORK. 1995. Implications Inc., Sunderland, MA. of intratropical migration reserve design: a case IUCN [ONLINE]. 2002. 2002 IUCN Red List of Threat- study using Pharomachrus mocinno. Conserva- ened Species. Ͻwww.redlist.orgϾ (27 April tion Biology 9:354±362. 2003). ROZAS, J., AND R. ROZAS. 1999. DnaSP version 3. An JOHNSGARD, P. A. 2000. and quetzals of the integrated program for molecular population ge- world. Smithsonian Institution Press, Washington, netics and molecular evolution analysis. Bioinfor- DC. matics 15:174±175. KEIGWIN, L. D. 1982. Isotopic paleoceanography of the RYDER, O. A. 1986. Species conservation and system- Caribbean and east Paci®c: role of Panama uplift atics: the dilemma of subspecies. Trends in Ecol- on late neogene time. Science 217:350±353. ogy & Evolution 1:9±10. KUMAR, S., K. TAMURA,I.B.JAKOBSEN, AND M. NEI. SAMBROOK, J., E. F. FRITSCH, AND T. M ANIATIS. 1989. 2001. MEGA2: molecular evolutionary genetics Molecular cloning. A laboratory manual. 2nd ed. analysis software. Arizona State University, Tuc- Cold Spring Harbor Laboratory Press, Cold son, AZ. Spring Harbor, New York. KVIST, L., M. RUOKONEN,J.LUMME, AND M. ORELL. SCHNEIDER, S., D. ROESSLI, AND L. EXCOFFIER. 2000. 1999. Different population structures in northern Arlequin: a software for population genetics data and southern populations of the European Blue Tit analysis. Version 2.0. Genetics and Biometry Lab- (Parus caeruleus). Journal of Evolutionary Biol- oratory, Department of Anthropology, University ogy 12:798±805. of Geneva, Geneva. LIEDLOFF, A. 1999. Mantel Version 2.0. Mantel non- SIBLEY, C., AND B. L. MONROE JR. 1990. Distribution parametric test calculator. Queensland University and of birds of the world. Yale Univer- of Technology, Queensland, Australia. sity Press, New Haven, CT. MARSHALL, H. D., AND A. J. BAKER. 1997. Structural SOLOÂ RZANO, S., M. A. CASTILLO-SANTIAGO,D.A.NA- conservation and variation in the mitochondrial VARRETE-GUTIEÂ RREZ, AND K. OYAMA. 2003. Im- control region of fringilline ®nches (Fringilla pacts of the loss of Neotropical highland forests spp.) and the Green®nch (Carduelis chloris). Mo- on the species distribution: a case study using Re- lecular Biology and Evolution 14:173±184. splendent Quetzal an endangered bird species. Bi- MORITZ, C. 1994a. Applications of mitochondrial DNA ological Conservation 114:341±349. analysis in conservation: a critical review. Molec- STOTZ, D. F., J. W. FITZPATRICK,T.A.PARKER III, AND ular Ecology 3:401±411. D. K. MOSCOVITS. 1996. Neotropical birds. Ecol- MORITZ, C. 1994b. De®ning `evolutionarily signi®cant ogy and conservation. University of Chicago units' for conservation. Trends in Ecology & Evo- Press, Chicago. lution 9:373±375. THOMPSON, J. D., T. J. GIBSON,F.PLEWNIAK,F.JEAN- NADKARNI, N. M., AND N. T. WHEELWRIGHT. 2000. MOUGIN, AND D. G. HIGGINS. 1997. The Clustal X Monteverde. Ecology and conservation of a trop- windows interface: ¯exible strategies for multiple ical cloud forest. Oxford University Press, Oxford, sequence alignment aided by quality analysis UK. tools. Nucleic Acids Research 24:4876±4882.