SYSTEMATICS OF THE SOUTHERN FORMS OF $ELASPHORUS (TROCHILIDAE) F. GARY STILES Escuelade Biologla,Universidad de CostaRica, CiudadUniversitaria, Costa Rica, Central America ABSTR•CT.--Basedupon evidence from morphology, behavior, and ecology, I propose that the taxonomyof the southernSelasphorus hummingbirds be set forth as follows: Selasphorusfiammula Salvin: VolcanoHummingbird S. f. fiammulaSalvin, 1864 (Volcfin Iraz6, VolcfinTurrialba, Costa Rica) S. f. torridusSalvin, 1870(Cordillera de Talamanca,Costa Rica-Panama) S. f. simoniCarriker, 1910(Volcgn Po•s, Volc•n Barba,Costa Rica) Selasphorusscintilla Gould, 1850:Scintillant Hummingbird (Cordillerade Tilarfin,Costa Rica south and east to Volc•n Chiriqul, Panama,at lower elevationsthan populationsof the preceding species). Selasphorusardens Salvin, 1870: Glow-throated Hummingbird (Serran•a de Tabasar•, Panama). S. scintilla and S. ardensmay comprisea superspecies.S. "underwoodii"is a hybrid be- tween S. scintillaand S. f. flammula. The breeding distributionsof fiammula,simoni, and torridusare entirely allopatric, but the birds may occurtogether in the nonbreedingseason. In particular,a pronouncedpost- breeding movementmay carry many torridusinto the breeding areasof fiammulaand even simoni. A possible evolutionary history of these forms is proposed in relation to post- Pleistocene climatic changes, ecological requirements, and probable populations sizes. Received7 June1982, accepted2 December1982. THE genus Selasphorus,as presentlydefined resident, although, as we shall see, they en- by mostauthors, includes 6 or 7 speciesof small gage in regular altitudinal movements. (2-31/2g) hummingbirdswith at leastsome ru- Due to the great similarity between some fous in the body plumage and tail; males have speciesand the variability of others,the genus orange,red, or purple gorgets.Geographically, Selasphorushas a long history of taxonomic the genus comprisestwo groups of species:a problems.Even in a relativelywell-studied area northern group, whose three members breed like California, Allen's and Rufous humming- in western North America, with one species birds were not recognized as distinct species extending south in the mountains to Guate- until 1877.The southerngroup presentsan even mala; and a southern group in the mountains more complexpicture, with the statusof tor- of Costa Rica and Panama. ridus and the affinities of simoni in particular Most of the North American populations of still being debated.In this paperI hope to clar- the northern speciesrufus (RufousHumming- ify the taxonomicand geographicrelationships bird), sasin (Allen's Hummingbird), and of the various southern forms of Selasphorus, platycercus (Broad-tailed Hummingbird) are based upon a detailed examination of their migratory, wintering mainly in western and plumagesand information on their breeding central Mexico; the population of the latter distribution, annual cycles, and altitudinal species breeding in Mexico and Guatemala, movements. however, is evidently resident.There is thus a gap of severalhundred kilometersbetween the TAXONOMIC HISTORY southernmostpopulation of platycercusand the nearest populations of the southern group, The nomenclatural history of the southern which is not bridged by migration. The south- Selasphorushummingbirds began with the de- ern forms scintilla(Scintillant Hummingbird), scription of the orange-gorgetedscintilla by simoni ("Cerise-throated Hummingbird"), Gould in 1850 (type locality Volcfin de Chiri- fiammula (Volcano Hummingbird), torridus qui). In 1864,the purple-gorgetedflammula was ("Heliotrope-throated Hummingbird"), and describedby Salvin from specimenscollected ardens(Glow-throated Hummingbird) are also by Arc• on VolcfinIrazti, CostaRica. Other Arc• 311 The Auk 100: 311-325. April 1983 312 F. GARYSTILES [Auk, Vol. 100 specimensfrom Volcan Chiriqui provided the suggestedthat it might be only subspecifically basisfor Salvin'sdescription of yet anotherpu- distinct from the latter (but he evidently did tative species, the steely-purple to greenish- not examine their morphologyin detail). He gorgeted torridus,in 1870. In the same paper alsostated that simoniand fiammula were "partly he describedthe red-gorgetedardens from a sympatric"but did not specifywhether or not Salvin and Godman specimentaken at Castillo, this referred to breeding distributions. Wet- Veraguas. Salvin finally bowed out of Selas- more (1968)restored torridus to subspecificsta- phorustaxonomy in 1897 with the description tus, consideringthat the VolcAnChiriqui birds, of a form with a reddish-orangegorget, under- while variable, were clearlydistinct from nom- woodii,from a bird collectedby Underwood on inate fiammula, which he found had not been Volcan Irazd. (For the original descriptions,see collectedby Berlioz on Volcan Chiriqui. Al- the citationsin Ridgway 1911.) though noting specificallya specimenof tor- The first critical analysis of the southern Se- ridus taken by Slud on Volcan Turrialba, how- lasphoruswas made by Carriker (1910),in con- ever, he failed to accountfor the presenceof sultation with the French hummingbird ex- both torridusand fiammulaon the Irazd-Tur- pert, Eugene Simon. Carriker strongly rialba massifand implied thatthe Chiriqui and suggestedthat underwoodiirepresented merely Irazd-Turrialba populations were virtual dis- an extreme variant of scintilla but did not ac- juncts,with birds of thefiammula-torridus com- tually reduce it to synonymy, as he had not plex being rare at best alongthe main Cordille- seen Salvin's type. (Subsequentauthors have ra de Talamanca (despite Slud's statement to considered underwoodiito be a synonym of the contrary). Wetmore did not comment on scintilla.)More important, following a sugges- the possible relationship of simonito ardens. tion by Simon,Carriker recognized that the red- It is worth noting that virtually all the taxo- gorgetedbirds that had been taken on Volcan nomic work on the southernSelasphorus hum- Barba in Costa Rica were different from ardens mingbirds to date has been based on color of Panama in the color of the tail and crissum characters,especially of the males'gorgets and and in bill length; accordingly,he named the central rectrices. Other aspects of plumage Costa Rican birds a new species, simoni. He morphologyhave gone almostentirely unap- was puzzledby the statusof fiammulaand tor- preciated, and measurements have been used ridus, stating that the only differencebetween haphazardly,without attentionto samplesizes. them was in the color of the males' gorgets. In addition, reliable age and sexcriteria, which Following Simon, he reducedtorridus to a sub- allow one to comparebirds of the samesex and speciesof S. fiammulabut noted that the two age classof differentforms, have not yet been had been taken together at severallocalities, derivedfrom identifyingbirds not in adult male which, in view of their supposedsubspecific plumage. The distributionsof the variousforms status, he found "rather difficult to explain." have previouslybeen delimited without regard He hypothesized that the torridusin question to distinguishing the breeding distribution were in realityjust worn or fadedfiammulaand from postbreeding altitudinal and local move- that true torriduswas found only in Chiriqui. ments; this, in turn, requiressome knowledge Ridgway (1911) accepted most of Carriker's of what months constitutethe breeding and conclusionsbut raised torridusto speciesrank molting seasons. based upon supposeddifferences in the color of the rectricesof femalefiammulaand torridus. METHODS The statusof torridusand fiammulawas in- vestigatedby Berlioz (1949), who collectedse- This study derives principally from an examina- ries of both forms. He ostensibly found such tion of museumspecimens, supplemented by field- variability in torriduson Volcan Chiriqui that work with all of the forms concernedexcept ardens. he consideredit to be a color phase or morph Measurementsof exposedculmen, wing chord, and tail lengthwere takenwith dial calipersof eachspec- of S. fiammula.This conclusionwas endorsed imen, and various morphologicaldetails were noted, by Slud (1964), who noted that torridusand especiallythe form and colorationof all the rectrices, fiammula"evidently coexistin Central Costa which I had found very helpful for establishingsex, Rica" but did not provide a detailed analysis. age, and speciescriteria among the northern group He asserted that simoni differed from ardens in of Selasphorus(Stiles 1972). I visited nearly all the colorcharacters only, not in measurements,and high mountainsand majormountain ranges of Costa April 1983] SelasphorusSystematics 313 Rica, where ! was able to observe numbers of simoni, Rectrices.--Patternsand shapesof typical ex- scintilla,fiammula, and torridusin the field. ! mist- amples of the rectrices of the adults of both netted at least some individuals of torridus, scintilla, sexesof all forms are diagrammed in Fig. 1; and simoniand weighed them to the nearest 0.1 or variations in the shape of certain rectricesare 0.05 g with Pesola spring balances.The form most intensively studied has been torridus(cf. Wolf et al. diagrammed in Fig. 2. The two most similar 1976), for which ! have been able to determine breed- forms are undoubtedly fiammula and torridus. ing and molting seasonsand to document some al- The amountof blackor duskyin the outer webs titudinal movements directly. This, in turn, has of the rectricesof adults is slightlymore exten- proved most helpful in evaluating seasonalityand sive on average, in the former,