The Auk A QuarterlyJournal of Ornithology Vol. 110 No. 1 January 1993

The Auk 110:1-8 + frontispiece, 1993

RELIC OF A LOST WORLD: A NEW OF (TROCHILIDAE: HELIANGELUS) FROM "BOGOTA"

GARY R. GRAVES Departmentof VertebrateZoology, National Museum of NaturalHistory, SmithsonianInstitution, Washington, D.C. 20560, USA

AI•STRACT.--Anew speciesof , the BogotJSunangel (Heliangeluszusii), is describedfrom a unique specimenpurchased in 1909 in BogotJ,Colombia. Hellangeluszusii is intermediate in plumage between ungorgetedH. regalisand the typical gorgeted species in the . Now possiblyextinct becauseof ,H. zusiimay have inhabited cloud forest and forest edge in the Eastern Cordillera of the Colombian . Received10 April 1992, accepted6 July 1992.

FUELED BY the demands of a fashion-conscious specimenslabeled "Bogota,"Colombia, some of public, millions of destined for which were actually taken elsewhere.Many of the millinery trade in Europe and the United thesehave now been determined to be hybrids, Stateswere exported from South America be- genetic plumage variants, artifacts, or subspe- fore protection laws were passed in the ciesof previously describedspecies (e.g. Taylor early 20th century. The staggeringsize of these 1909, Simon 1921, Berlioz and Jouanin 1944, shipmentscan hardly be overemphasized.One Greenway 1978, Graves 1990).A tantalizing few London auction house alone sold 152,000 hum- (e.g. Isaacson'sPuffieg, Eriocnemisisaacsonii [Par- mingbirds between 1904 and 1911 (Doughty zudaki] 1845) may representvalid species,pos- 1975), a figure that undoubtedly exceedsthe sibly surviving in someornithologically unex- number of hummingbird specimens now plored locality or perhaps extinct. housed in the world's museums. In 1947, BrotherNic•foro Maria sent a spec- As expected,commerce in avian plumagehad tacular fork-tailed hummingbird purchasedin a profound effecton science.The golden age of Bogotain 1909to RudolpheMeyer de Schauen- hummingbird and systematicswas seeat the Academy of Natural Sciencesof Phil- attained much earlier than in other Neotropical adelphia (ANSP) for identification. After re- bird families (Gould 1861). More than a quarter ceiving the specimen, Meyer de Schauensee of all hummingbird specieswere describeddur- asked several of his colleaguesto examine the ing the 1840sand, by 1890, more than 91% had specimen (literature among the archives of been discovered(Table 1). By comparison,only ANSP). Although there was little agreement 73% of all antbird specieswere described by among the correspondents,their ideas consti- 1890. As the rate of discoverydeclined (1860- tute someimportant hypothesesconcerning the 1895), systematicsscoured imports for rarities identity and origin of the specimen. and described dozens of new trochilid taxa from JamesL. Peters (in litt., 10 April 1947) wrote, Frontispiece. Aerial pursuitof male BogotaSunangels (Heliangelus zusii sp. nov.) aboveflowering Brachyo- turnm•crodon, a common melastome in the EasternCordillera of the ColombianAndes. Painting by JonFjelds/i. 2 GARYR. GRAVES [Auk,Vol. 110

TABLE1. Number of hummingbird and antbird Finally, Alexander Wetmore (in litt., 5 May (Thamnophilidae and Formicariidae) speciesde- 1947) wrote, "I have never seen a bird like it .... scribed as new to scienceby decade(taxonomy of I am inclined to doubt any hybrid origin for Sibley and Monroe 1990). this bird, on the contraryI would supposethat Species it is a specimen of an unknown species.... If I were planning a description of this specimen I Decade Hummingbirds Antbirds would compareit closelywith Agelaiocercus[sic] 1750 13 0 to determine whether it should be placed in 1760 5 3 1770 0 1 that genusor named as a distinct genus." 1780 15 13 Meyer de Schauensee(1947:113) concluded, 1790 1 0 "Were it not for the fact that I have examined 1800 3 1 Metallura purpureicaudaand Zodaliathaumasta and 1810 15 9 1820 26 24 found them to be virtually identical in color- 1830 47 14 ation and pattern, but differing in the shape of 1840 87 15 the bill, I would not hesitate to describe Brother 1850 33 44 Nicfiforo's bird as a new speciesand perhaps 1860 21 33 1870 14 11 even a new genus.However, there are so many 1880 11 10 points of similarity between our specimen and 1890 9 8 Neolesbiathat without seeingthe type it seems 1900 3 19 the wisest course to identify it as a third ex- 1910 4 7 ample of Neolesbia.It should be noted that both 1920 2 12 1930 1 5 the type and the American Museum specimen 1940 2 2 have longer bills (19.25, 17 mm) than ours 1950 2 3 (15.5)." 1960 2 1 Recently,Hinkelmann et al. (199!) reaffirmed 1970 2 1 the identification of the ANSP specimen as an 1980 1 7 example of the problematic Nehrkorn's Sylph (Neolesbianehrkorni), which they judged to be a hybrid between the Long-tailed Sylph (Aglaio- "Assumingyour bird to be a hybrid, one of the cercuskingi) and the Fork-tailed Woodnymph parents is probably one of the Colombian pur- (Thaluraniafurcata).Here I presentevidence that ple-tailed forms of ;the shapeof the the hybrid origin of Brother Nic•foro's speci- bill, color and general shape of the tail and men can be rejected and that it does indeed luminous plaqueson forehead and throat could represent an undescribed species. quite conceivably result from a union between Aglaiocercusand some form of Heliangelus,but MATERIALS AND METHODS in sucha caseI should expectthe body plumage of the product to be green.... Now what • I compared the specimen (ANSP 159261) directly with all hummingbird taxain the ANSP and National crossedwith a C of predominately green col- Museum of Natural History, SmithsonianInstitution. oration (or vice versa) would produce a steel- Detailed notes and color transparenciesof the spec- blue offspring?Damned if I know. If Helianthea imen (Fig. 1) were comparedwith the trochilid col- []prunellei is involved I should expect lections of the Museum of Natural Sciences, Louisiana a longer billed result at least; Eriocnemisnigri- State University and the American Museum of Nat- vestiscan be dismissedas a geographic impos- ural History (AMNH), including a specimen with a sibility. SupposeNeolesbia to be distinct, should molting tail, provisionallyidentified asNeolesbia nehr- not the resultantproduct have at leasta slightly korni(AMNH 484177), the type of which cannot now decurved bill?... It's all very puzzling and I be found (Hinklemann et al. 1991). The specimenwas am afraid I haven't been any help." comparedwith the descriptionand colorplate of Neo- lesbianehrkorni (Berlepsch 1887), and color transpar- Meyer de Schauensee(in litt., 19 May 1947 enciesof a specimenin the Museum Heineanum Hal- to Alexander Wetmore) reported the observa- berstadt, Germany (courtesyof B. Nicolai), that was tions of John T. Zimmer of the American Mu- recently identified as N. nehrkorni(Hinkelmann et al. seum of Natural History, "Zimmer saw it and 1991). All color comparisonswere made under Ex- thought it could possiblybe a hybrid between amolites (Macbeth Corporation). Descriptions of Aglaiocercuskingi and Heliangelussquamigularis." structural colors are unusually subjective and actual •anuary 1993] New Speciesof Sunangel

Fig. 1. Dorsal.ventral. and lateral view of holotype of Heliangeluszusii. 4 GARYR. GRAVES [Auk, Vol. 110 color varies with angle of inspection. For this reason a white spot are found posterior to the eyes. I use general color descriptionsthroughout. The hindneck, back, and wing coverts are deep Measurementsof wing chord, bill length from an- bluish-black, slightly iridescent, turning deep terior edge of nostril, and rectrix length from point greenish-blueon the lower back and rump and of insertion of centralrectrices to the tip of the longest returning to bluish-black on the upper tail co- rectrix of each pair (from innermost to outermost), verts. Primaries and secondaries are dark brown were made with digital calipersto the nearest0.1 mm. with a purplish tint. Secondarieshave acumi- Measurementsof the Heineanum Halberstadt speci- men of Neolesbianehrkorni were provided by B. Ni- nate rather than broadly rounded tips. The colai. Diagnosticassumptions and methodsof hybrid deeply forked tail (fork 54% of length) is glit- diagnosis based on plumage characters and mor- tering dark purple above, duller and less iri- phology follow Graves(1990). descent on the lower surface. Basal portions of the rectrices, obscured in the folded tail, are the Heliangelus zusii sp. nov. same color as the exposed distal tips (unlike Bogot• Sunangel speciesof Aglaiocercusand Lesbia).Shafts of rec- trices are dark brown distally, fading to light Holotype.--Academy of Natural Sciences of brown basally.Rectrices are flat in crosssection. Philadelphia, No. 159261;adult (• ?);purchased The ventral plumage is duller than the dorsum. in Bogota,Colombia in 1909 by Brother Nic•- The chin is dull bluish-black, bordered poste- foro Maria. riorly by a brilliant golden-green gorget (same Diagnosis.--Heliangeluszusii is a dark bluish- color as frontlet). Gorget feathers have golden- black hummingbird with a deeply forked, dark green tips separated from the gray base by a purple tail, brilliant green frontlet and gorget, narrow violet band. Gorget and frontlet feath- and straight bill (seefrontispiece). It differs from ers are broadly rounded rather than narrow or H. regalisin having a frontlet and gorget, a pur- pointed. The gorget is bordered below and on plish rather bluish-black tail, and pale rather the sidesby lustrousbluish-black. When viewed than bluish-black undertail coverts. Other spe- head-on in direct light, the head and breast ap- cies of Heliangelushave green or bronzy-green pear black and contrastgreatly with the golden- body plumage and violet, purple, or orange gor- green gorget and frontlet. Feathersof the upper gets.Heliangelus zusii differs from Neolesbianehr- breast are bluish-black with narrow buffy-gray korni in having a gorget and frontlet, purple margins, becoming progressively paler on the rather than bluish-black tail, and a straight rath- lower breastand belly. The midline of the belly er than decurved bill. is medium gray but faint traces of bluish-black Genericrelationships.--The new speciescan be can be observed on barbules under magnifica- assigned to the genus Heliangelusby a combi- tion. Sidesand flanks are dark bluish-black (same nation of characters (Figs. 1 and 2): (1) short as back). There are no tracesof a pectoral band (15.1 mm), straight, unmodified, uniformly dark or subterminal white spots on breast feathers bill; (2) brilliant frontlet and gorget; (3) un- (present in some Heliangelushybrids; Graves modified remiges;(4) unmodified rectrices,flat 1990, Graves and Zusi 1990). Undertail coverts in crosssection; (5) lack of puffy tarsal plumes are pale creamy-white with a single lanceolate, (present in Eriocnemisspp.); and (6) nonbrilliant subterminal spot at the midline (brownish-gray undertail coverts (brilliant in most Eriocnemis with blue reflections). Tibial feathers are dull spp.). Heliangeluszusii appears to bridge the gap grayish-brown. The tarsi are now bare (former- between males of the ungorgeted, bluish-black ly feathered?). Soft-part colors in dried speci- H. regalisand the more typical members of the men: bill black; scuteson feet brown; foot pads genus with contrasting gorgets and predomi- yellowish-brown. nately green body plumage. Measurements (mm).--See Table 2 for standard Descriptionof holotype.--Taxidermy mount measurements. Greatest width of outermost rec- with glasseyes. Nasal feathers,lores, auriculars, trices, 9.2. Gorget length x width, ca. 14 x 10. and margins of crown are glittering steel-blue. Sex, age, and molt.--Although the sex of the A brilliant pale golden-green frontlet extends holotype is not known with certainty, all plum- from the forecrown posteriorly to an imaginary age characters resemble components of male line drawn between the centers of the orbits. plumage in Heliangelusspecies. The fully de- Feathers of the frontlet are adpressed as in H. veloped gorget and frontlet, unstriated maxil- exortis. Nasal feathers extend 0.85 mm distal to lary ramphothecum,and lack of molt in the the anterior edge of the nasalflanges. Traces of holotype indicate that it is an adult. January1993] NewSpecies ofSunangel 5

Fig. 2. Dorsal and lateral view of head and bill of the holotypeof Heliangeluszusii.

Distnl•ution.--Originand range unknown. See nominal origin of Heliangeluszusii and many discussion. other Andean species. Etymology.--I take great pleasure in naming this hummingbird for my friend and colleague, DISCUSSION Richard L. Zusi, in recognition of his contri- butions to the systematicsof hummingbirds. BecauseH. zusiiis representedby a unique The English common name commemoratesthe specimen,extraordinary care must be taken to 6 GARYR. GRAVES [Auk, Vol. 110

TABLE2. Measurements(mm) of holotypeof Helian- ation, has been documented in a number of geluszusii and Heineanum Halberstadtspecimen of trochiline species(Hartert 1922,Greenway 1978, Neolesbianehrkorni (courtesy of B. Nicolai). Bleiweiss 1985). Becauseof its unique size and Heliangelus Neolesbia shape(e.g. forked tail, straightbill, etc.),H. zusii Characters zusii nehrkorni could not represent a simple genetic plumage variant of another species. Wing 66.8 63.5 Bill 15.1 18.0 I also consideredthe possibilitythat H. zusii Rectrix 1 (innermost) 30.1 25.5 representsa well-differentiated allopatric form Rectrix 2 36.6 30.5 of the ungorgeted (H. regalis). Rectrix 3 45.7 37.5 Heliangeluszusii is larger in all dimensions (ca. Rectrix 4 56.2 50.5 Rectrix 5 (outermost) 64.8 66.5 25% in wing and tail length) than H. regalis. This difference exceedsthe magnitude of geo- graphic size variation found within sexes of other speciesof Heliangelus(Bleiweiss 1985; pets. rule out or reject alternate taxonomichypoth- observ.). Significantly, no Heliangelusspecies eses.Below I address the possibility that the (males)has "gorgeted" and "ungorgeted"pop- specimen represents:(1) Neolesbianehrkorni; (2) ulations or subspecies.Because the differences a genetic variant or undescribedsubspecies of between H. zusiiand H. regalisare both quali- some other species;(3) a hybrid; or (4) an arti- tative (seediagnosis) and quantitative, I believe fact. the two taxaare more genetically differentiated Neolesbianehrkorni?--As noted in the diag- than are many of the componenttaxa of Andean nosis,H. zusiidiffers from H. nehrkorniin having superspecies(Graves 1980, 1985, 1991). a brilliant gorget and frontlet. Could H. zusii Hybrid?--Determining whether a unique represent the mature male plumage of N. nehr- specimenrepresents a hybrid or a valid species korni?Several factors argue against this being can be very difficult depending on the circum- the case.Neolesbia nehrkorni (wing, 63.5, tail 67.5; stances(Graves 1990). Complicating this partic- Berlepsch1887) is approximately the samesize ular caseis the fact that the preciseorigin of H. as H. zusii (wing, 66.8, tail 64.8). Because tail zusii is unknown. For the purpose of hybrid length is correlated with age in males of long- diagnosis(Graves 1990), the speciespool must tailed hummingbirds(e.g. Aglaiocercus spp.), one include all trochiline hummingbird that occur would expectthe type specimensof N. nehrkorni in Colombia,about 120 species (Hilty and Brown and H. zusiito be roughly the sameage. Yet, N. 1986).However, the taskof determining wheth- nehrkornilacks a brilliant gorget and frontlet, er H. zusiiis a hybrid is simplified by its dis- charactersthat are acquired at maturity in He- tinctive characters:(1) deeply forked tail; (2) liangelusand are present in H. zusii. Signifi- straight bill; (3) brilliant gorget and frontlet; cantly, the bill of N. nehrkorniis slightly de- and (4) bluish-black body plumage. curved as opposed to straight in H. zusii In correspondence cited above, Peters and (Berlepsch 1887). Although intraspecific varia- Zimmer suggestedthat H. zusiimight represent tion in bill curvature in hummingbirds has not a hybrid of Aglaiocercussp. x Heliangelussp. been studied, I know of no example where bill Severalcombinations of speciesfrom thesegen- curvature varies with age. eracould have produced a hybrid with a deeply Heliangeluszusii and the Heineanum Halber- forked tail, straightbill, and brilliant gorgetand stadt specimen of N. nehrkornialso differ sig- frontlet. However, no speciesin either genus, nificantly in body proportions (Table 2). De- other than H. regalis,has bluish-blackbody spite having a longer wing, H. zusiihas a shorter plumage. Furthermore, H. zusiiexhibits no trac- bill and lessdeeply forked tail; the rectricesof es of the peculiar awn-shaped bill, tapered rec- H. zusiiare more evenly graduatedin length. trix tips, or bicoloredrectices of Aglaiocercusspp. In sum, H. nehrkorniand H. zusiido not rep- In conclusion, there appears to be no combi- resent different age classesof the same taxon. nation of species,considered two at a time, that Discussionof the systematicstatus of N. nehr- collectivelyexhibit the charactersof H. zusiiand korniwill appear elsewhere (Graves in prep.). bracket its morphological dimensions (Graves Geneticplumage variant or undescribedsubspe- 1990). cies?--Plumagecolor polymorphism and mel- Artifact?--Many artifacts, specimenscreated anism, uncorrelated with size and shape vari- from parts of two or more species,are known January1993] NewSpecies ofSunangel 7 among 19th century collections (pers. observ.). the Eastern Cordillera, where it is common These were created for the amusement of the (Snow and Snow 1980, Hilty and Brown 1986, preparators or to dupe eager collectors of nat- J. Fjeldst•in litt.). It, the BogotaSunangel, and ural-history specimens.The possibility that H. many other presently endangeredspecies are zusii is an artifact was rejected after a careful relics of a lost world that existed in the Andes examination of the feathers and skin under before the arrival of human civilization. magnification (7-30 x) and x-radiographs of the specimen.These proceduresrevealed no ev- ACKNOWLEDGMENTS idence of compositeconstruction of the speci- men. I thank R. C. Banks, L. C. Binford, R. E. Bleiweiss, Biogeography.--Ispeculate that the holotype C. Hinkelmann, B. L. Monroe, J. V. Reinsen, and R. of H. zusiioriginated from the EasternCordil- L. Zusi for comments on the manuscript, as well as lera of the Colombian Andes within a few hun- the curators and staff of the American Museum of dred kilometers of Bogota, or possibly in the Natural History, Louisiana State University Museum of Natural Science, and the National Museum of Nat- Central Cordillera. The fact that only a single ural History, Smithsonian Institution for assistance specimen is known suggeststhat the species in their collections.I am grateful to Jon Fjeldst•for had a relictual or restricted geographic distri- his beautiful and innovative frontispiece. Special bution when first collected. (Helian- thanks are due to M. B. Robbins and F. B. Gill for gelus) inhabit Andean cloud forest and shrub- loaning the unique specimenand providing accessto lands from 1,200 to 3,400 m elevation, a zone the archives of the Academy of Natural Sciencesof that has been largely deforestedin the Eastern Philadelphiaand to B. Nicolai for colortransparencies Cordillera for cultivation of coffee, maize, po- of Neolesbia.Research was partially supportedby grants tatoes, and other crops (Chapman 1917, Insti- from the Frank M. Chapman Memorial Fund of the tuto Geographico"Agustin Codazzi" 1982, Hil- American Museum and the Research Opportunities Fund of the Smithsonian Institution. ty 1985). Currently, only a single speciesof sunangel,the Amethyst-throatedSunangel (H. amethysticollis),occurs over most of the Eastern LITERATURE CITED Cordillera; two others, the Orange-throated (H. BERLEPSCH,H. VON. 1887. Beschreibungeines neuen mavors)and Tourmaline sunangels (H. exortis), Colibri und Bemerkungentiber eine Collection have restricted distributions (Hilty and Brown yon Vogelb•ilgenaus der Umgegendyon Sta.F• 1986). Although the likelihood that H. zusiistill de Bogotain Colombia. Zeitschrift ftir die ge- survives seems remote, efforts should be made saturateOrnithologie 4:177-187. to find it in remnantpatches of forest,including BERLIOZ,J. 1937. Three new casesof presumed nat- second growth, on both slopes of the Eastern ural hybridsamong Trochilidae. Ibis 1 (14th set.): Cordillera between 1,400 and 2,200 m. If extinct, 105-109. H. zusii would be the first such instance for a BERLIOZ,J,, AND C, JOUANIN, 1944, Liste de Trochil- id•s trouv•s dans les collections commerciales de hummingbird species in South America. The Bogota.Oiseau 14:126-155. only other speciesof hummingbird that is sus- BLEBVEISS,R. 1985. Iridescent polychromatismin a pected of being extinct is Chlorostilbonbracei, female hummingbird: Is it related to feeding the endemic emerald of New Providence Is- strategies?Auk 102:701-713. land, Bahamas (Graves and Olson 1987). CHAPMAN, F.M. 1917. The distribution of bird-life The effectsof commercialcollecting on 19th- in Colombia: A contribution to a biological sur- century hummingbird populationsin Colombia vey of South America. Bull. Am. Mus. Nat. Hist. are unknown, but nearly all species with his- 36:1-729. torically large geographicranges (> 30,000 km 2) DOUGHTY, R. W. 1975. Feather fashions and bird are common today in pristine habitats (Hilty preservation: A study in nature protection. Univ. and Brown 1986). This suggeststhat habitat de- California Press,Berkeley. GOULD,J. 1861. An introduction to the Trochilidae struction,rather than intensive collecting,is the or family of humming-. Published by the principal threat to most hummingbird species author, London. (see Hilty 1985). For example, the Black Inca GRAVES,G.R. 1980. A new speciesof metaltailhum- (Coeligenaprunellei) was one of commonestspe- mingbird from northern Peru. Wilson Bull. 92: cies in "Bogota" collections(Berlioz 1937). Re- 1-7. grettably, the speciesis known presently from GRAVES,G. R. 1985. Elevational correlatesof speci- only a few remnant patches of cloud forest in ation and intraspecificgeographic variation in 8 GARYR. GRAVES [Auk,Vol. 110

plumage in Andean forest birds. Auk 102:556- HINKELMANN, C., B. NICOLAI, AND R. W. DICKERMAN. 579. 1991. Notes on a hitherto unknown specimenof GRAVES,G.R. 1990. Systematicsof the "green-throat- Neolesbianehrkorni (Berlepsch, 1887; Trochilidae) ed sunangels"(Aves: Trichilidae): Valid taxa or with a discussionof the hybrid origin of this hybrids? Proc. Biol. Soc. Wash. 103:6-25. 'species.'Bull. Br. Ornithol. Club 111:190-199. GRAVES,G. R. 1991. Bergmann'srule near the equa- INSTITUTO GEOGRAPHICO "AGUSTIN CODAZZI." 1982. tor: Latitudinal clines in body size of an Andean Atlas regional Andino. passerinebird. Proc. Natl. Acad. Sci. USA 88:2322- MEYER DE SCHAUENSEE,R. 1947. New or little-known 2325. Colombian birds. Proc. Acad. Nat. Sci. Phila. 99: GRAVES, G. R., AND S. L. OLSON. 1987. Chlorostilbon 107-126. braceiLawrence, an extinct speciesof humming- SIBLEY,C. G., AND B. L. MONROE, JR. 1990. Distri- bird from New Providence Island, Bahamas. Auk bution and taxonomyof birds of the world. Yale 104:296-302. Univ. Press, New Haven, Connecticut. GRAVES,G. R., ANDR. L. ZIJSI. 1990. An intergeneric SIMON, E. 1921. Histoire naturelie des Trochilidae hybrid hummingbird (Heliodoxaleadbeateri x He- (synopsis et catalogue). Encyclopedia Roret, L. liangelusamethysticollis) from northern Colombia. Mulo, Paris. Condor 92:754-760. SNOW,D. W., AND B. K. SNOW. 1980. Relationships GREENwAY,J. C., JR. 1978. Type specimensof birds between hummingbirds and flowers in the An- in the AmericanMuseum of Natural History.Part des of Colombia. Bull. Br. Mus. Nat. Hist. (Zool.) 2. Bull. Am. Mus. Nat. Hist. 161:1-305. 38:105-139. HARTERT,E. 1922. Typesof birds in the Tring Mu- TAYLOR,W.P. 1909. An instanceof hybridization in seum.B. Typesin the general collection.Trochili. hummingbirds, with remarks on the weight of Novitat. Zool. 29:403-412. generic characters in the Trochilidae. Auk 26: HILTY,S.L. 1985. Distributional changesin the Co- 291-293. lombianavifauna: A preliminary blue list. Pages 1000-1012 in Neotropical ornithology (P. A. COLOR PLATE Buckley,M. S. Foster,E. S. Morton, R. S. Ridgely, and F. G. Buckley,Eds.). Ornithol. Monogr. No. Publicationof the color plate was supportedby the 36. Donald L. Bleitz Fund of The American Ornitholo- HILTY,S. L., ANDW. L. BROWN.1986. A guide to the gists' Union. birds of Colombia. Princeton Univ. Press, Prince- ton, New Jersey.