The nucleotide sequences of mitochondrial 12S rRNA (12S) and protein-coding cytochrome oxidase subunit I (COI) spanning 1,010 and 1,130 base pairs, respectively, are reported for seven owl species representing six genera: Otus megalotis (Philippine Scops-Owl), O. longicornis (Luzon Scops-Owl), Bubo virginianus (Great Horned Owl), Asio flammeus (Short-eared Owl), Ninox philippensis (Philippine Hawk-Owl), Tyto alba (Barn Owl; used as an outgroup), and the little known Mimizuku gurneyi (Giant Scops-Owl). Separate phylogenetic analyses of 12S, COI, and the combined 12S and COI data yielded congruent phylogenies, strongly supporting the phylogenetic affinity of Mimizuku with Otus, as opposed to Bubo. The conservative nature of mitochondrial DNA evolution in Mimizuku, in sharp contrast to its unique derived morphological attributes, suggests that Mimizuku is a recently evolved insular form of Otus that has undergone rapid morphological evolution towards gigantism, rather than being a small derivative of the much larger eagle-owls (Bubo) as has been suggested previously. Received 19 August 1996, accepted 5 March 1997.

Department of Biological Sciences, University of Cincinnati, Cincinnati, Ohio 45221, USA; 2 Department of Biology and Museum of Zoology, University of Michigan, Ann Arbor, Michigan 48109, USA; and 3 Frederick and Arney Geier Collections and Research Center, Museum of Natural History and Science, Cincinnati Museum Center, 1720 Gilbert Avenue, Cincinnati, Ohio 45202, USA THE GIANT SCOPS-OWL (Mimizuku gurneyi) is a monotypic species restricted to the Philip- pine islands of Mindanao, Dinagat, and Siar- gao. The evolutionary origin and phylogenetic placement of this species are unclear Mimizuku gurneyi was first described as Pseudoptynx gur- neyi by Tweeddale (1878), congeneric with P. philippensis (now Bubo philippensis ). Hachisuka (1934) removed gurneyi from Pseudoptynx and placed it in the monotypic genus Mimizuku. Many authors agree and recognize M. gurneyi as the sole member of this distinct genus (Peters 1940, dupont 1971, Clark et al. 1978, Eck and 4 E-mail: mirandhc@ucbeh.san.uc.edu Busse 1973, Marshall 1978, Dickinson et al. 1991), whereas others treat Mimizuku as a syn- onym of Otus (Delacour and Mayr 1946, Gross- man and Hamlet 1964, Burton 1973, Clements 1974, Gruson 1976). Amadon and Bull (1988: 304) state that Mimizuku "... is sometimes placed in Otus, but may be closer to Bubo ..." Its English name is no less controversial. Mimizuku gurneyi is widely known as the Giant Scops-Owl (Delacour and Mayr 1946, Burton 1973, Clark et al. 1978, Amadon and Bull 1988, Sibley and Monroe 1990), but more recently it has been referred to as the Lesser Eagle-Owl (Dickinson et al. 1991, Sibley and Monroe 1993) and the Mindanao Eagle-Owl (Inskipp et al. 1996). The uncertainty regarding the phyloge- netic placement of Mimizuku reflects insuffi- cient materials for comparative study and the lack of knowledge about its biology. During re- cent surveys of Mindanao conducted by the National Museum of the Philippines (NMP) and the Cincinnati Museum of Natural History and Science (CMNH), we obtained frozen tis- sue and a complete skeleton that have clarified whether this unique insular species represents a small eagle-owl or a large scops-owl. Reso- lution of this issue provides clues to the history and processes underlying avian species diver- sification in the Philippines and other archi- pelagos. METHODS We obtained DNA sequence characters for the en- tire mitochondrial (mt) 12S rRNA gene (977 base pairs [bp] long in M. gurneyi; 1,010 bp long in the alignment, including gaps) from two individuals of M. gurneyi and single representatives of Otus mega- lotis, O. longicornis, Bubo virginianus, Asio fiammeus, Ninox philippensis, and Tyto alba, the last being used for outgroup comparison. We also sequenced 1,130 bp of mitochondrial cytochrome oxidase subunit I (COI) from the same set of taxa except O. longicornis. All Otus, Ninox, and Mimizuku tissue samples were obtained during the 1992 to 1994 NMP / CMNH Phil- ippine Biodiversity Inventory surveys. Voucher spec- imens for individuals analyzed are: M. gurneyi (CMNH Nos. 36488 and 35735), O. longicornis (CMNH No. 36491), O. megalotis (CMNH No. 36490), and N. philippensis (CMNH No. 36492). GenBank ac- cession numbers for the sequences are: O. m. everetti (12S, U83754; COI, U83779), O. longicornis (12S, U84751), M. gurneyi (12S, U83756; COI, U83780), B. virginianus (12S, U83757; COl, U83777), A. fiammeus (12S, U83758; COl, U83781), N. philippensis (12S, U83760; COI, U83783), and T. alba (12S, U83749; COI U91604). DNA was extracted from muscle, liver, or heart tis- sue using standard techniques (Knight and Mindell 1993). Mitochondrial DNAs were amplified as a se- ries of overlapping segments using the following primer pairs. For 12S rRNA, we used L613/H887 (Mindell et al. 1991), L821 (5-GCCACACCCCCAC GGGTAC TCAGCAGT-3')/Hl194 (5'-TCGATTATA GAACAGGCTCCTCTAG-3'), L1091 (Kocher et al. 1989) /H1557 (Knight and Mindell 1994), and L1373 (5'-GAAATGGGCTACATTTTCT-3') / H1621 (5'-CTT IT/C][T / C]AGGTGTAAGCT[G / A][G / A]ATGCTT- 3'). For COI, we used L5844 (5'-CCTCTGTAAAAA GGACTACAGCC-3')/H6290 (5'-TTGCCAGCTAGT GGGGGTA-3'), L6195 (5'-AATAACATAAGCTTC TGACT-3') / H6762 (5'-GATGTAAAGTAGGCTC GGGTGTCTAC-3'), and either L6424 (5'-ACCG CCATCAACATAAAACCCCC-3') or L6704 with H7216 (Mindell et al. 1991). Letters L and H refer to light and heavy strands, and numbers correspond to approximate starting positions in the human mt- DNA sequence (Anderson et al. 1981). Bases for de- generate primer sites are shown in brackets. Double- stranded PCR amplifications were performed using Taq DNA polymerase and the reaction buffer (500mM KC1, 100 mM Tris-HC1, 1.0% Triton X-100) from Promega Corporation. Reaction mixtures were subjected to 35 cycles of denaturation at 94讈 for 30 s, annealing from 40 to 50讈 for 30 s, and extension at 72讈 for 1 min. After amplification, 15 mL of the double-stranded PCR product were separated in low melting-point agarose, cut from the gel, and purified using Magic PCR Preps (Promega). Both the heavy and light strands of all amplified products were se- quenced using the CircumVent Sequencing System (New England Biolabs) with direct incorporation of 35S-dATE Manual alignment of COI sequences is non-contro- versial, with no gaps being invoked. 12S rRNA align- ments, however, require numerous insertions and deletions. Although phylogenetic analyses can be sensitive to order of sequence input for alignments with numerous insertions and deletions (Mindell 1991), we focused on well-aligned regions only, and these did not change significantly in alignment bouts with alternative input orders. CLUSTAL W (Thomp- son et al. 1994) was used to generate the alignment using the default parameters. Conserved motifs and other information regarding secondary structure (Hickson et al. 1996, Mindell et al. 1997) also were used to improve and adjust the alignment by eye. Percent difference between sequences was initially calculated using p = zd / zt, where zd is the number of nucleotide differences between two sequences and z is the total number of nucleotides compared. Se- quence differences corrected for substitution-rate differences were also calculated using Kimura's (1980) two-parameter method in the DNAdist pro- gram of PHYLIP (Felsenstein 1995). Sites with inser- tions and deletions were excluded from sequence-di- vergence estimates. We conducted exhaustive parsimony analyses us- ing PAUP 3.1.1 (Swofford 1993) for the CO! and 12S rRNA data sets individually and combined. To ex- plore the effect of differential weighting of transi- tions and transversions, we used the following char- acter weighting schemes in alternative analyses: (1) all character changes equally weighted, (2) transver- sions (TVs) only, and (3) TVs:TIs (transitions) weighted 2:1. For the 12S rRNA alignments, we con- ducted separate analyses on all sites (1,010 aligned positions) and on well-aligned regions only (944 aligned positions), in which sequence regions with gaps were excluded to reduce homoplasy. We cal- culated support indices for each node within the most-parsimonious topology by finding the shortest TABLE 1. Percent differences that are presumed transitions (above diagonal) and percent sequence diver- gences (below diagonal) for pairwise comparisons of mitochondria112SrRNA sequences based on observed sequence differences (left of slash) and Kimura's (1980) two-parameter model (right of slash). I 2 3 4 5 6 7 I Mimizuku gurneyi -- 71.0 71.2 65.2 61.7 58.5 61.4 20tus megalotis 6.3/6.8 -- 72.4 64.8 64.1 63.3 63.0 30tus longicornis 11.2/11.1 8.0/8.7 -- 68.6 66.5 62.3 59.5 4 Bubo virginianus 16.4/16.1 14.1/17.1 16.1/20.0 -- 62.5 66.7 58.8 5 Asiofiammeus 17.7/21.8 16.1/19.6 16.2/19.6 15.1/18.1 -- 61.8 59.9 6 Ninoxphilippensis 16.6/20.0 15.5/18.3 16.0/19.2 16.8/20.1 16.8/20.2 -- 58.3 7 Tyto alba 17.3/20.7 16.6/19.8 15.5/18.5 16.9/19.9 16.9/21.7 16.5/19.5 -- tree in which the particular node is not found, and denoting on branches the number of additional steps required (Bremer 1988, Kallersj5 et al. 1992). In ad- dition, bootstrap resampling (Felsenstein 1985) was done with 2,000 replications using the branch-and- bound search option. RESULTS 12S RNA.--Of the 1,010 bp examined, 55 gaps were invoked. Within the regions without gaps, 325 positions were variable and 233 of those positions were phylogenetically infor- 12S rRNA 3-60 11-99 12-99 Otus longicornis I {-- Otus megalotis Mimizuku gurneyi 3-65 [ Bubo virginianus Asio flammeus Ninox philippensis Tyro alba FG. 1. The most-parsimonious phylogenetic tree demonstrating placement of Mimizuku gurneyi, based on all alternative analyses of the mitochondri- al 12S rRNA. An analysis using transversions only yielded the same topology as above, except Asiofiam- meus was basal, rather than sister, to Bubo virginianus. Tyto alba was designated as the outgroup. The first number above the branch is the support index (Bremmer 1988, Kallersj5 et al. 1992) denoting the number of additional steps needed to break the node, and the second number denotes bootstrap support based on 2,000 replications. These numerical calcu- lations are based on 12S rRNA well-aligned regions only. mative (23%). Presumed transitions for pairs of species ranged from 58.3 to 72.4% of all changes with an average of 63.9% across all comparisons (Table 1). Percent transitional dif- ferences were based on observed values. These relatively high percentages suggested that the 12S rRNA transition substitutions retained phylogenetic information, and that they were not yet saturated with change (Mindell and Ho- neycutt 1990, Mindell et al. 1991). Excluding sites with gaps (where alignments may be un- reliable), the mean observed pairwise percent divergence across all comparisons was 15.0%, ranging from 6.3 to 17.7%. Sequence diver- gence between O. megalotis and M. gurneyi was 6.3%, lower than between the two Otus species at 8.0%. Percent sequence divergences between M. gurneyi and the two Otus species were less than that found between Mimizuku and B. vir- ginianus at 16.4%. Transformed divergence measures using the Kimura 2-parameter model increased the average by 3.2%, but distances between M. gurneyi and O. megalotis, O. longi- cornis, and B. virginianus at 6.8%, 11.1%, and 16.1%, respectively, did not change drastically (see Table 1). As a precaution against mislabel- ing or contamination of samples, 500 bp were sequenced from a second Mimizuku individual, and no difference was found between the two individuals. Using Tyto alba as an outgroup, the parsi- mony analyses were unequivocal in support- ing an Otus/Mimizuku sister relationship (Fig. 1). Parsimony analyses of the 12S rRNA data set based on alternative character weighting approaches consistently yielded the same sister relationship. All of the most-parsimonious trees were identical, except for the transver- sion-only analysis for well-aligned regions in which A. fiammeus is basal rather than sister to TABLE 2. Mitochondrial COl percent transitional differences (above diagonal) based on first, second, and third codon positions (left of slash) and first and second codon positions only (right of slash). Values below diagonal are percent divergences for pairwise comparisons based on observed differences (left of slash) and Kimura's (1980) two-parameter model (right of slash). 1 2 3 4 5 6 I Mimizuku gurneyi -- 20tus megalotis 10.5/11.4 3 Bubo virginianus 15.9/18.2 4 Asio fiammeus 15.4 / 17.6 5 Ninox philippensis 17.2/20.0 6 Tyro alba 18.3/21.6 87.1/63.9 93.6/77.8 91.1/72.2 86.1/59.1 91.6/72.3 -- 91.8/67.9 90.0/60.7 83.8/47.2 91.0/62.5 14.7/16.7 -- 96.0/73.4 89.9/58.6 96.4/80.0 16.1/18.7 13.3/14.8 -- 89.2/62.5 95.8/76.5 17.1/20.1 16.3/18.4 15.6/17.9 -- 89.5/57.7 18.3/21.8 16.4/19.2 16.6/19.2 19.1/23.0 -- B. virginianus. All trees that were generated, based on all regions or well-aligned regions only, showed O. longicornis basal to the O. mega- lotis/M. gurneyi clade. The support index showed an additional 12 steps required to break the O. megalotis/M. gurneyi clade and 11 additional steps to break the entire O. longicor- nis / O. megalotis / Mimizuku clade. Bootstrap values for all branches were greater than 50%, with the O. megalotis and M. gurneyi clade showing a 99% bootstrap value. Cytochrome oxidase I.--COI showed 325 vari- able sites among five ingroup taxa. There were no indels (insertions/deletions) observed. For- ty-five changes (13.9%) were at the first codon positions, 28 (8.6%) were at the second codon positions, and 252 (77.4%) were at third codon positions, a ratio of roughly 1.5:1:9. Although the substitution rate difference between the first two codons and the third codon positions is relatively high, the observed patterns are consistent with studies of other protein-coding mitochondrial genes in birds (Kocher et al. 1989, Edwards and Wilson 1990, Lanyon and Hall 1994). High percent transition values for pairwise estimates were observed, ranging from 83.8% between O. megalotis and N. philip- pensis to 96.4% between T. alba and B. virgini- anus (Table 2). This was due mainly to the rapid nucleotide change at the third codon position. Substitution bias at the third codon position has been shown to contribute to phylogenetic misinformation due to saturation (Mindell and Thacker 1996). Percent transitional changes for both first and second codon positions were lower, ranging from 47.2% (Ninox and Otus) to 80.0% (Ninox and Bubo). Overall percent se- quence divergence based on the Kimura two- parameter model (Table 2) between strigiform species ranged from 11.4% (Otus and Mimizu- ku) to 23.0% (Tyto and Ninox). Sequence diver- gence between Mimizuku and Bubo was 18.2% a value higher than that between Otus and Mimizuku or Bubo and Asio. The COI gene also showed bias in base composition, with guanine deficiency in the second and third codon posi- tions, and relatively low thymine frequency at the third codon position, as observed in other mt protein-coding genes (Lanyon and Hall 1994). The mean COI percent base composition for the seven owl species was A: 25.4, C: 30.8, G: 16.6, and T: 24.7. Analyses with equal weighting for all char- acter changes, and with TVs:TIs weighted 2:1 (for first, second, and third codon positions combined) yielded a tree incongruent with the 12S topology in which Bubo is basal to a Ninox/ Asio clade. Transversion parsimony for all co- don positions differed, with Asio basal to a Bubo/Ninox clade. However, all topologies gen- erated under various weighting approaches based on the first, second, and third codon po- sitions supported the Otus/Mimizuku sister re- lationship. To reduce the amount of homoplasy due to multiple substitutions, third codon po- sitions were given an a priori weighting of zero. Exhaustive parsimony analysis of all characters in the first and second codon positions resulted in a single most-parsimonious tree that united Mimizuku with Otus (Fig. 2A). Parsimony anal- yses also were conducted using: (1) only the first codon position, (2) only the second codon position, and (3) transversions only at the first and second codon positions. Each of these sup- ported the sister relationship of Otus and Mim- izuku relative to Bubo. Transversion parsimony of codon position 1 and 2 only resulted in two shortest trees, both supporting the Otus/Mim- izuku clade, with Bubo basal to this branch. A consensus tree (Fig. 2A) yielded an unresolved Bubo/Asio/Ninox branch. Bremer and boot- strap support indices for the two resolved Phylogenetic Placement of Mimizuku gurneyi B 12S and COI combined 319 1-66 Otus megalotis Mimizuku gurneyi Bubo virginianus Asio flammeus Ninox philippensis Tyto alba j Hif65 2-62 FG. 2. (A) COl consensus of two shortest trees resulting from transversion parsimony analysis of all first and second codon position characters. (B) Single most-parsimonious tree found under all weighting schemes for 12S and COl combined, with and without exclusion of ambiguous alignment regions in 12S and the sat- urated third codon position in COl. The first numbers above branches represents the support index (Bremer 1988, Kallersj6 et al. 1992), and the second number denotes bootstrap support with 2,000 replications. branches were low (Fig. 2A), due in part to the small number of changes at the first and second codon positions. 12S rRNA and COI combined.--The combined data set comprised 2,140 nucleotide bases. Par- simony and transversion analyses of all data and all data excluding ambiguous 12S align- ment regions yield a single most-parsimonious tree (Fig. 2B) that was entirely congruent with the 12S gene tree (Fig. 1). This topology is ro- bust as indicated by the lack of sensitivity to al- ternative weighting approaches, and inclusion or exclusion of ambiguous alignment regions in 12S and the third codon position in COI. Skull morphology.--The recent availability of a M. gurneyi skeleton enabled gross comparison of its skull with those of O. megalotis and B. vir- ginianus. We evaluated the same characters used by Ford (1967): (1) medio-lateral slope of cranial side, (2) forehead size and shape, (3) cranial width relative to height, and (4) supra- orbital process shape. The O. megalotis skull ex- hibits a high and steep forehead, the supraor- bital process is long and pointed, and the cra- nium is high relative to width (Fig. 3). The Mimizuku skull also exhibits these features (Fig. 3). In contrast to Otus and Mimizuku, the Bubo skull has a low, thicker forehead and a small, knob-like supraorbital process, and the cranial height is low relative to the width (Fig. 3). The structural similarities of the major skull features in Otus and Mimizuku support the mo- lecular evidence presented here, although a de- tailed phylogenetic analysis of skeletal charac- ters is warranted. DISCUSSION We present evidence that Mimizuku is more closely related to Otus than to Bubo, assuming the monophyly of Bubo. All analytical ap- proaches for the 12S rRNA (i.e. all gene regions and the well-aligned regions only, all character changes weighted equally, transversions only, or transversions:transitions weighted 2:1) sup- port the single shortest tree presented in Fig- ure 1 with Mimizuku sister to Otus. This also is concordant with the COI analyses, the com- bined data set analyses, and the gross compar- isons of skull morphology (Fig. 3). Combining data from different regions of the mitochondri- al genome maximizes the amount of character evidence in parsimony analyses (Cracraft and Mindell 1989, Kluge 1989, deQueiroz et al. 1995). The apparent consistency of morpholog- ical and molecular evidence increases our con- fidence in the mitochondrial DNA analyses. Our evidence indicates that Mimizuku gur- neyi is a large scops-owl that is intermediate in size between scops-owls and eagle-owls. Ac- cordingly, this species should be known by the English name "Giant Scops-Owl." The plum- age pattern of the breast and abdomen differs from that of most scops-owls in being less cryptic, with plain black streaks along the shafts of the ventral feathers (see Frontispiece). Field observations by Miranda and Kennedy suggest that Mimizuku's call (a loud, single- note "whaaack") is not typical of scops-owl vo- calizations. Despite these derived characters, the mitochondrial 12S sequence divergence of FIG. 3. Lateral (A) and dorsal (B) view of skulls of (1) Otus rnegalotis (CMNH No. 36489), (2) Mimizuku gurneyi (CMNH No. 35735), and (3) Bubo virginianus (CMNH No. 34913). Dorsal view is shown 20% smaller than lateral view. M. gurneyi from O. megalotis is relatively low (6.8%) and is lower than the observed se- quence-divergence estimates between O. mega- lotis and O. longicornis (8.7%). The derived phe- notypic features of Mimizuku, in contrast to its conservative mitochondrial DNA evolution, suggest the occurrence of rapid differentiation. Discordant rates of morphological and molec- ular evolution also have been documented in other avian species (Shields and Wilson 1987; Avise and Zink 1988; Zink et al. 1991, 1995; Zink and Dittmann 1993) and other vertebrate lineages (King and Wilson 1975, Baverstock and Adams 1987, Schwaner and Sarre 1988, Radtkey 1996). This notion is supported by field investigations showing that speciation and evolutionary changes in morphology in birds can be rapid (e.g. Grant 1992). Mimizuku gurneyi is an uncommon resident of low and middle elevation forests. On Min- danao, its habitat overlaps with that of the more widespread Otus megalotis, which occurs in forested areas and open woodlands. It is possible to devise a scenario wherein Mimizuku represent a lineage of colonizers that under- went rapid morphological changes towards gi- gantism. The alternative scenario, that Mimi- zuku is more closely related to the larger eagle- owls, is that Mimizuku evolved smaller body size, perhaps in response to competition with the Philippine Eagle-Owl (B. philippen$is). Our data and phylogenetic analyses support the first scenario (toward gigantism) over the latter (toward size reduction). There is an active de- bate as to the ecological forces that drive shifts in body size among insular reptiles (Miles and Dunham 1996, Schoener 1970, Williams 1972, Roughgarden 1992). In insular forms, random genetic drift that accompanies a founder event can play a subtle but major role in rapid spe- ciation. Our study raises the issue of which criteria are appropriate and reasonable for delimiting genera. In the past, it was common to place a species in a separate genus if it exhibited a greater degree of morphological divergence than that observed among other related spe- cies. But, should a separate genus be recog- nized solely on the basis of apparent phenotyp- ic uniqueness, regardless of how recently the taxon has evolved in evolutionary time? Or, should genera reflect phylogeny and relative age regardless of degree of morphological dif- ferentiation? Taxonomists will encounter these questions with increasing frequency as greater emphasis is placed on phylogenetic perspec- tives and on historical contexts of biodiversity. ACKNOWLEDGMENTS We are grateful to numerous colleagues, particu- larly J. Brown and R. Fernandez, who helped obtain specimens of Mimizuku gurneyi. We owe special thanks to E C. Gonzales of the National Museum of the Philippines, and the many Philippine institu- tions and agencies that assisted during our field work. The NMP/CMNH Philippine Biodiversity In- ventory was supported by a grant from the John D. and Catherine T. MacArthur Foundation to P. C. Gon- zales and R. S. K. Molecular laboratory work was supported by National Science Foundation grants DEB-9019669 and 9496343 to D. E M. and a Univer- sity of Cincinnati Wieman/Wendell Summer Fellow- ship grant to H. C. M. B. Collier kindly prepared the photographs of the owl skulls. H. C. M. thanks T. Kane, B. Kinkle, S. Rogstad, and S. Pelikan for their advise and logistical support. We are indebted to J. Marks, D. Miles, M. Wink, R. Zink, and an anony- mous reviewer for their valuable comments on the manuscript. We also thank Mrs. Eugene Farny, and Joe and Jan Herron for financial support of Philip- pine research, and Outdoor Adventures for assis- tance with field equipment. This paper is Contribu- tion No. 15 of the NMP/CMNH Philippine Biodi- versity Inventory. LITERATURE CITED AMADON, D., AND J. BULL. 1988. Hawks and owls of the world: A distributional and taxonomic list (with the genus Otus by J. T. Marshall and B. E King). Proceedings of the Western Foundation of Vertebrate Zoology 3:295-357. ANDERSON, S., A. T. BANKIER, B. G. BARRELL, M. H. L. BRUIJN, A. R. COULSON, J. DROUIN, I. C. EPE- RON, D. P. NIERLICH, B. A. ROE, F. SANGER, P. H. SCHREIRER, A. J. H. SMITH, g. STADEN, AND I. G. YOUNG. 1981. Sequence and organization of the human mitochondrial genome. Nature 290:457- 465. AVISE, J. C., AND g. M. ZINK. 1988. Molecular genetic divergence between avian sibling species: King and Clapper rails, Long-billed and Short-billed dowitchers, Boat-tailed and Great-tailed grack- les, and Tufted and Black-crested titmice. Auk 105:516-528. BAVERSTOCK, P. g., AND M. ADAMS. 1987. Compar- ative rates of molecular, chromosomal and mor- phological evolution in some Australian verte- brates. Pages 175-188 in Rates of evolution (K. S. W. Campbell and M. E Day, Eds.). Allen and Un- win, London. BREMER, K. 1988. The limits of amino-acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42:795-803. BURTON, J. A. (Ed.). 1973. Owls of the world. E. E Dutton and Company, New York. CLARK, g. J., D.G. SMITH, AND L. H. KELSO. 1978. Working bibliography of owls of the world. Na- tional Wildlife Federation Scientific and Techni- cal Series No. 1, Washington, D.C. CLEMENTS, J. E 1974. Birds of the world: A checklist. Two Continents, New York. CRACRAFT, J., AND D. P. MINDELL. 1989. The early history of modern birds: A comparison of mo- lecular and morphological evidence. Pages 389- 403 in The hierarchy of life (B. Fernholm, K. Bre- mer, and H. J6rnvall, Eds.). Elsevier Science Pub- lishers, Amsterdam. DELACOUR, J., AND E. MAYR. 1946. Birds of the Phil- ippines. Macmillan, New York. DEQUEIROZ, g., M. J. DONOGHUE, AND J. KIM. 1995. Separate versus combined analysis of phyloge- netic evidence. Annual Review of Ecology and Systematics 26:657-681. DICKINSON, E. C., g. S. KENNEDY, AND K. C. PARKES. 1991. The birds of the Philippines: An annotated check-list. British Ornithologists' Union, Check- list No. 12. DUPONT, J. E. 1971. Philippine birds. Delaware Mu- seum of Natural History, Greenville. ECK, S., AND H. BUSSE. 1973. Eulen, die resentenund fossilen Formen Aves, Strigidae. Die Neue' Brehm-Bucherei. A. Ziemsen, Wittenberg Luth- erstadts, Germany. EDWARDS, S. V., AND A. C. WILSON. 1990. Phyloge- netically informative length polymorphism and sequence variability in mitochondrial DNA of Australian songbirds (Pomatostomus). Genetics 126:695-711. FELSENSTEIN, J. 1985. Confidence limits on phylog- enies: An approach using the bootstrap. Evolu- tion 39:783-791. FELSENSTEIN, J. 1995. PHYLIP (Phylogeny Inference Package), version 3.57c. Department of Genetics, University of Washington, Seattle. FORD, N. 1967. A systematic study of the owls based on comparative osteology. Ph.D. dissertation, University of Michigan, Ann Arbor. GRANT, P. R. 1992. Hybridization of bird species. Science 256:193-197. GROSSMAN, M. L., AND J. HAMLET. 1964. Birds of prey of the world. Bonanza Books, New York. GRUSON, G. S. 1976. Checklist of the world's birds. Quadrangle, New York. HACHISUKA, M. 1934. Birds of the Philippine Is- lands, vol.2. H. E & G. Witherby, London. HICKSON, R. E., C. SIMON, A. COOPER, G. S. SPICER, J. SULLIMAN, AND D. PENNY. 1996. Conserved se- quence motifs, alignment and secondary struc- ture for the third domain of animal 12S rRNA. Molecular Biology and Evolution 13:150-169. INSKIPP, T., N. LINDSEY, AND W. DUCKWORTH. 1996. An annotated checklist of the birds of the Ori- ental region. Chandlers Printer Limited, East Sussex, United Kingdom. KALLERSJO, M., J. S. FARRIS, A. G. KLUGE, AND C. BULT. 1992. Skewness and permutations. Cla- distics 8:275-287. KIMURA, M. 1980. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. Journal of Molecular Evolution 16:111-120. KING, M. C., AND A. C. WILSON. 1975. Evolution at two levels in humans and chimpanzees. Science 188:107-116. KLUGE, A.G. 1989. A concern for evidence and a phy- logenetic hypothesis for relationships among Epi- crates (Boidae, Serpentes). Systematic Zoology 38: 7-25. KNIGHT, A., AND D. P. MINDELL. 1993. Substitution bias, weighting of DNA sequence evolution, and the phylogenetic position of Fea's viper. System- atic Biology 42:18-31. KNIGHT, A., AND D. P. MINDELL. 1994. On the phy- logenetic relationship of Colubrinae, Elapidae, and Viperidae and the evolution of front-fanged venom systems in snakes. Copeia 1994:1-9. KOCHER, T. D., W. K. THOMAS, A. MEYER, S. V. ED- WARDS, S. P;t;BO, E X. VILLABLANCA, AND A. C. WILSON. 1989. Dynamics of mitochondrial DNA evolution in animals: Amplification and sequencing with conserved primers. Proceed- ings of the National Academy of Sciences USA 86:6196-6200. LANYON, S.C., AND J. G. HALL. 1994. Reexamination of barbet monophyly using mitochondrial-DNA sequence data. Auk 111:389-397. MARSHALL, J. t., JR. 1978. Systematics of smaller Asian night birds based on voice. Ornithological Monographs No. 25. MILES, D. B., AND A. E. DUNHAM. 1996. The paradox of the phylogeny: character displacement of analyses of body size in island Anolis. Evolution 50:594-603. MINDELL, D. P. 1991. Aligning DNA sequences: Ho- mology and phylogenetic weighting. Pages 73- 89 in Phylogenetic analysis of DNA sequence (M. M. Miyamoto and J. Cracraft, Eds.). Oxford University Press, Oxford. MINDELL, D. P., C. W. DICK, AND R. J. BAKER. 1991. Phylogenetic relationships among megabats, microbats, and primates. Proceedings of the Na- tional Academy of Sciences USA 88:10322- 10326. MINDELL, D. P., AND R. L. HONEYCUTT. 1990. Ribo- somal RNA in vertebrates: Evolution and phy- logenetic applications. Annual Review of Ecol- ogy and Systematics 21:541-566. MINDELL, D. P., M.D. SORENSON, C. J. HUDDLESTON, H. C. MIRANDA, JR., A. KNIGHT, S. J. SAWCHUK, AND t. YURI. 1997. Phylogenetic relationships among and within select avian orders based on mitochondrial DNA. Pages 211-245 in Avian molecular evolution and systematics (D. P. Min- dell, Ed.). Academic Press, New York. MINDELL, D. P., AND C. E. THACKER. 1996. Rates of molecular evolution: Phylogenetic issues and applications. Annual Review of Ecology and Systematics 27:279-303. PETERS, J. L. 1940. Checklist of birds of the world, vol. 4. Harvard University Press, Cambridge, Massachusetts. RADTKEY, R. R. 1996. Adaptive radiation of day- geckos (Phelsuma) in the Seychelles archipelago: A phylogenetic analysis. Evolution 50:604-623. ROUGHGARDEN, J. 1992. Comments on the paper by Losos: Character displacement versus taxon loop. Copeia 1992:288-295. SCHOENER, t. W. 1970. Size patterns in West Indian Anolis lizard. II. Correlations with the sizes of particular sympatric species-displacement and convergence. American Naturalist 104:155-174. SCHWANER, t. D., AND S. D. SARRE. 1988. Body size of tiger snakes in Southern Australia, with par- ticular reference to Notechisater serventyi. Journal of Herpetology 22:24-33. SHIELDS, G. E, AND A. C. WILSON. 1987. Subspecies of the Canada Goose (Branta canadensis) have distinct mitochondrial DNA's. Evolution 4:662- 666. SIBLEY, C. G., AND B. L. MONROE, JR. 1990. Distri- bution and taxonomy of birds of the world. Yale University Press, New Haven, Connecticut. SIBLEY, C. G., AND B. L. MONROE, JR. 1993. A sup- plement to distribution and taxonomy of birds of the world. Yale University Press, New Haven, Connecticut. SWOFFORD, D. L. 1993. PAUP. Phylogenetic analysis using parsimony, version 3.1.1. Illinois Natural History Survey, Champaign. THOMPSON, J. D., D. G. HIGGINS, AND t. J. GIBSON. 1994. Clustal W: Improving the sensitivity of progres- sive multiple sequence alignment through se- quence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22:4673-4680. TWEEDDALE, A. H. 1878. Contributions to the orni- thology of the Philippines, no. XI. On the collec- tions made by Mr. A. H. Everett at Zamboanga, in the island of Mindanao. Proceedings of the Zoological Society of London 1878:936-954. WILLIAMS, E. E. 1972. The origin of faunas. Evolu- tion of lizard congeners in a complex island fau- na: A trial analysis. Evolutionary Biology 6:47- 88. ZINK, R. M., AND D. L. DITTMANN. 1993. Gene flow, refugia, and evolution of geographic variation in the Song Sparrow (Melospiza melodia). Evolution 47:717-729. ZINK, R. M., D. L. DITTMANN, AND W. L. ROOTES. 1991. Mitochondrial DNA variation and the phylogeny of Zonotrichia. Auk 108:578-584. ZINK, R. M., S. ROHWER, A. V. ANDREEV, AND D. L. DITTMANN. 1995. Trans-Beringia comparisons of mitochondrial DNA differentiation in birds. Condor 97:639-649. Associate Editor: R. M. Zink