BEHAVIOR OF THE SPARROW AND TWO SPECIES OF SEED-FINCHES

ALAN LILL

The behaviorof many of the predominantlygranivorous neotropical speciesconventionally grouped under the familial headingFringil- lidae in field guidesis poorly known. Documentingit may help promote a better understandingof the phylogenyand systematicsof thesebirds and clarify the evolutionand adaptivesignificance of their socialsystems. It thus seemsworthwhile to report some brief observationsI made on the territorial behaviorand feedinghabits of GrasslandSparrows (Myo- spiza humeralis), Large-billed Seed-Finches(Oryzoborus crassirostris), and Lesser Seed-Finches(O. angolensis)during a short field trip to in July 1971. The behavior of these speciesseems not to have been reported in any detail. These observationswere an extensionof a socioecologicalinvestigation of the "finches"of Trinidad and Tobago to be publishedelsewhere. I use "finch" and "Fringillid" as convenient terms for specieslisted under the family heading Fringillidae in Meyer de Schauensee(1970a). I spent 35 hours observingthese specieswith the aid of 10 x 50 binocularsat four locationsin Guyana.

GRASSLAND SPARROW The GrasslandSparrow is known from both tropical and subtropical zones of South America. Meyer de Schauensee(1970a) lists it as a species inhabiting grassland, Hayerschmidt (1968) records it from open sandy savannawith low grassesand scatteredbushes in Surinam, and Snyder (1966) notesits in Guyana as grasslandand savanna. The specieswas common on the campus of the University of George- town, a flat, rather damp tract at sea level comprisinggrassland with some sedges(notably Cyperus articulatus L.), which was apparently mowedat intervals.The only relatedspecies commonly observed there was the Chestnut-belliedSeedeater (Sporophila castaneiventris). Grass- land Sparrows were also abundant near the Manari River on the Rupununi Savannasof the interior where they occurred in exactly the same habitat as Hayerschmidt (ibid.) lists for Surinam. Cohabiting re- lated species here were the Wedge-tailed Grass-Finch (Emberizoides herbicola) and the GrasslandYellow-Finch (Sicalis luteola), but both thesespecies mainly occupiedcompletely open, bushlesssavanna adjacent to grasslandsthe sparrowsoccupied. I spent 15 hours watching the sparrowsat these two locations,mostly in the early morning and late afternoon.

35 The Auk 91: 35-43. January 1974 36 AL^•: L•LL [Auk, Vol. 91

In both areas Grassland Sparrows fed exclusively on the ground. They walked, or more rarely hoppedalong, stabbingdownwards to take seedsof mown or growing short-stemmedgrasses and sedges,and also stretchedup to seize seedsgrowing a little above head height. At the Georgetownsite they did not exploit the tall sedge,Cyperus articulatus, by perchingdirectly on the racemeto feed, as Sporophilacastaneiventris frequently did. This is a common feeding technique of the lighter Sporophila species,but may be difficult or impossiblefor the heavier and stockiersparrows to perform. On the university campusI recorded sparrowseating the seedsof the grassesEchinochloa colonu•n, Eriochloa ramosa, and Axonopus attenuatus, and also of the sedge Eleocharis mutata. They may occasionallyfeed on seedson taller stems by an essentiallyground-feeding technique, for I watchedone male jump up and try to pull a taller seed-bearingstem of Eriochloa down to the ground. Severalpredominantly ground-feeding neotropical finch species, notably yellow-finchesof the genus Sicalis, commonlypull tall grass stems down to the ground where they anchor them with the feet while feeding on the raceme. Twice I saw individuals take arthropods, one a small moth and the other an unidentifiedinsect, which they caught after short running chases. Chestnut-belliedSeedeaters also fed on the three identified sparrow food-plants,but their main food on the uni- versity campuswas Cyperus. The GrasslandSparrow seems to be adapted for the exploitationof of short grassesand sedgesseeding close to groundlevel. Its bill, like thoseof someother predominantlyground- feedingneotropical Fringillids, such as the Blue-blackGrassquit (Vola- tinia jacarina) and certain speciesof Sicalis, is more pointed and less conicalthan thoseof the largelygranivorous seedeaters. This couldbe an adaptation enabling the speciesboth to take more arthropod food and to probe efficiently amonglow-seeding grasses and for fallen seeds. Myospiza males were territorial at both study sites, occupyingterri- tories whose diameters I estimated very roughly to be 30-50 yards, which they defendedagainst conspecific males, and on which they sang, rested, and fed extensively. At the Georgetownsite they occasionally seemedto feed outsideof their territoriesalso. Typical songsof Grass- land Sparrowsat Georgetownwere tripartite, falling in pitch on the last note, but some included two additional terminal syllables. Most males at a locality sang a similar song pattern, but song was audibly different at the two locations,which are about 260 miles apart. The syllabic structure of the song at Manari can be rendered"zaa-zi-zee- zaa-zee-zaa," rising on the first three and falling on the last three syllables;it had a plaintive, whistlingquality. Males sang from a num- ber of repeatedlyused, prominent song posts, such as earth clods,bushes, January 1974] Guyana Sparrows 37 termite nests, and telephonewires from 1-40 feet above the ground, and also on the ground while foraging. Crest-raisingduring song was usual and autopreeningcommon between songsof a bout. Most males seemedto be paired and spent much time foraging with single indi- viduals, presumedto be females. Four agonisticencounters between adjacentterritory ownersinvolved (a) silent supplanting;(b) supplant- ing succeededby a chasein which one or both uttered chattering "zeek-zeek-zeeks"; (c) fighting with associatedchattering calls; and (d) silent fighting, after which the territory owner gaped with opened mandibles and emitted several "tsip" calls. On another occasion a memberof a presumedbonded pair gave this latter call after supplanting its partner. Interactions between presumed territorial pair members included a dancingdisplay observedtwice, in which a territorial male ran to and fro in front of the presumedfemale, facing her and shaking or flicking his lowered primaries. During one of these displays the female also danced a little and adopted a gaping posture with extendedneck and openedmandibles. The dance was apparently a courtshipdisplay. Gap- ing posturesoccurred also at other times in responseto approachby a pair memberand elicited retreat. This type of gaping is a threat display characteristic of several speciesof Sporophila seedeatersand grassquits (genera Volatinia and Tiaris). Other agonistic interactions between membersof presumedheterosexual pairs were fighting and supplanting- plus-chasingwith the pursuer chattering or uttering "tsip-tsip-tsip" calls. Followinga possiblelow aerial sexualchase, pair membersgave a different, multisyllabic,bubbling call upon landing. Other vocalizations noted were a grating "kee-kee-kee"and a sharp, monosyllabic"chip." The former was givenby one memberof a presumedpair perchingclose together on their territory and also by a male foraging alone on his territory. The "chip" was emitted by solitary birds, integrated pairs, and a family group; it resembledalarm "chips" of certain grassquitsand yellow-finchesand may well have been elicited by my presence. The "tsip" call was also given by flying birds. While these observationson vocal communicationin Grassland Sparrows are merely preliminary, it was evident that the specieshas a sizeable vocal repertoire and uses vocal communicationin agonistic,sexual, and other social contexts. No group larger than a three-bird family unit was seen, and the birds' behavior at dusk indicated that, unlike many other neotropicalFringillids that gather in large communalroosts even during the breedingseason, the Manari GrasslandSparrows roosted solitarily or in pairs or family groupson their territories. 38 ALANL•LL [Auk, Vol. 91

LARGE-BILLED SEED-FINCI-I Snyder (1966) describesthe habitat of the Large-billed Seed-Finch in Guyanaas forestedge and clearings.I found it in It6 palm (Mauritia /lexuosa)swamps adjoining low]and forests much disturbed by Amerindian cultivation along the Nappi River in the Rupununi savannasof the interior. Standingwater was 28 inches deep at that time in the rainy season,and the dorhinant emergent sedgeswere S½leriapratensis and Fuirena umbellata, with the hydrophyllicgrass Panicurn zizanioides also fairly common. Related cohabitingspecies were the Lesser Seed-Finch, Blue-black Grassquit,and Gray, Lined, and Ruddy-breastedSeedeaters (Sporophilaintermedia, S. lineola, and S. minuta). I judged the birds to be O. crassirostrisand not O. maximiliani by the shiny, smooth bill, bluish horn in color (cf. Meyer de Schauensee1970b), but I took no specimensand did not comparethe two speciesin the field. Both occur together at Annai only 40 miles away (Meyer de Schauensee1970b). Sight identifications can only be regarded as tentative in distinguishing these two recently separatedsibling species. I spent about 5 hours watchingthe specieson two mornings. Two or three males held adjacent swamp territories they defended, sang and fed on, though song fre.quencieswere not high in mid-July. A territory defenseencounter between two males involved chasingand supplanting. Singingoccurred at songposts high in the Moriche palms, also in emergentsaplings 2 or 3 feet above the water during solitary feeding. Snyder (pers. comm.) describesthe song as comprisedof six notesrising in pitch. I saw one cock perform a song flight display with a high trajectory over his territory; the songemitted during it contained terminal trills and warbles not heard in song given statically. The dis- play closely resembled that of many Sporophila seedeaters. Males seemed to be associatedwith single presumed females on their terri- tories and at times fed in an integrated manner with them. I saw one presumedhen supplanther mate foragingclose by and noticeda possible sexual chase. Two adjacent, territorial, presumedpairs jointly mobbed a large anaconda (Eunectes murinus), a speciesknown to prey on such swamp birds as the Wattled Jaqana (Jacana jacana) (Haverschmidt 1970). The two territorial males intensively studied held overlapping territorieswith LesserSeed-Finches and Gray and Ruddy-breastedSeed- eaters,and the only interspecificaggression noted was the chasingof a territorial male O. crassirostrisby a female-plumagedSporophila minuta, possiblyan immaturemale. Solitarily feedingmales emitted a variety of simple calls that I tran- scribedas "chint-chint," "chwit," "chweeoo,""choo," and "peep." Pre- sumedfemales feeding alone also gave the "chwit" call interspersedwith January 1974] Guyana Sparrows 39 longer"wit-wit-wit-oo" vocalizations.I was impressedby the resemblance of the "chint," "peep,"and "wit" calls to vocalizationsgiven by Gray and Variable (S. americana) Seedeatersin similar nonsocial,but also in social,contexts. The monosyllabicvocalizations may possiblyserve to maintain contactwith the mate out of sight elsewhereon the ter- ritory.

LESSER All the LesserSeed-Finches I saw in the Lethem area of the Rupununi savannas had maroon-chestnutlower breast and belly . This speciesis called the Chestnut-belliedSeed-Finch by those who regard a predominantlyblack form foundwest of the Andesas a distinctspecies, O. ]unereus (Snyder 1966). Possibleintermediates between the two have been reported from (Meyer de Schauensee1970a). Meyer de Schauenseelists the species'habitat as woodlandedge, wasteland, and shrubbery;I foundit at Nappi in the swampand abandonedforest cultivation clearings and also in a partly grassy, partly cultivated clearingin, but near the edgeof, forest along the MocomocoRiver at the foot of the Kanuku Mountains. The species utilized the forest fringe as well as the adjacentopen areas. In addition to the cohabiting specieslisted for the Nappi study site above, Sicalisluteola was present in large numbersat Mocomoco. I spent 15 hours on three mornings and oneafternoon watching this species. Territorial behavior closely resembled that seen in O. crassirostris. Males sangin low bushes,high trees,and while feedingon their terri- tories,erecting the throat feathersduring singingand interspersingthe songswith boutsof autopreening.The song,which Snyder (1966) renders "hwee-hwee-hwee-HWEE-hwee-hwee,"was often augmented by long terminal trills and warbles, exactly as in some statically emitted songs and all songflight displaysof someSporophila seedeaters. Similar trilled songswere emitted by a solitary cock during a territorial song flight displayand by a male chasinga female-plumagedbird acrosshis territory. Someseedeaters, such as Sporophilaamericana also give extended,trilled songduring sexual chasing (Lill MS). ThoughI heard this type of song only during songflight display in O. crassirostris,it may not constitute a real differencebetween the two congenericspecies, as sexual chasing and singingfrequencies were low in O. crassirostrisduring my observa- tions. Males seemedto be paired with singlefemales, but twice I saw feeding associationsof a territorial male and two female-plumagedbirds on stands of Scleria. Males joined mixed speciesfringillid flocks feeding on their territories, and I noted some interspecific aggressionwith 40 An^• Ln•n [Auk, Vol. 91

S. minuta in such aggregations,but interspecificterritoriality towards related specieswas lacking. O. angolensismales and female-plumaged birds gave monosyllabic"chint" calls like those of O. crassirostriswhen feeding both solitarily and in pairs. A presumedhen feeding alone and membersof an integratedfeeding pair also gave a multisyllabicchattering call. At Nappi seed-finchesfed predominantlyor exclusivelyon the sedge Scleria pratensis. If food resourcesimpose a limit on population density in neotropicalfringillids, one would expect to find some divergencein diet selectionand/or feeding behavior among coexisting species. Brief observationslike mine might well not reveal any such differences, especiallybecause they were made when severalof the coexistingspecies were apparently breeding, a time when one would expect food resources to be optimal. It is interestingtherefore that during my limited observa- tions at Nappi none of the seedeatersfed on Scleria, but two (S. in- termedia and minuta), took Panicurn zizanioidesseeds, a rare element in the diet of only the smaller of the two seed-finches(O. angolensis), and one (S. minuta) took Fuirena umbellata seedsnot eaten by the seed-finches.It is also interesting to compare the feeding behavior pat- terns employedby the two seed-finchspecies on Scleria. Table 1 classi- fies the feedingmethods they usedand estimatestheir frequency. The LesserSeed-Finch had a broader repertoireof feedingbehavior pat- terns (12 as opposedto 4) than the Large-billed Seed-Finch, including some in which adjacent supporting vegetation was not used. Whereas O. crassirostrisused exclusivelyadjacent woody plants as support while feeding, O. angolensisalso used Scleria and employed feeding postures demandinggreat agility more commonly than did crassirostris.While someof thesedifferences may prove to be lessmarked on more extensive observation,they may also relate to the two speciesrelative sizes; Haverschmidt (1968) gives wing chord ranges of 66-67 mm for crassi- rostrisand 56-59 mm for angolensisin Surinam. If the observeddietary differencesbetween seed-finchesand seed- eaters at Nappi is a real one, it could be related to genericdifferences in morphologyand feeding behavior patterns and/or food preferences. Seedeatershave smaller bills, but I doubt if that precluded their feed- ing on Scleria seeds,which are significantly shorter but not wider than Fuirena seeds; the mean lengths of six collectedseeds of these species were 2.6 mm (Scleria) and 5.8 mm (Fuirena) (P < 0.001) and the mean widths 2.0 and 1.8 mm (P > 0.05, t-test, two-tailed). It is possible that Scleria seeds are harder to husk than those of the seedeater food plants, and that the greater weight of the seed-finches,especially O. crassirostris,(see Haverschmidt1968 for figures) restrictstheir ability January 1974] GuyanaSparrows 41

ddddd ddd dd 42 AL^• Ln•L [Auk, Vol. 91 to feed on less sturdy and more procumbentfood plants, which the seedeaters,with their ability to feed directly at the raceme and use acrobatic feeding postures,can exploit. The reality and causationof the observeddietary differenceand the competitiverelationships between these two groupswill emergeonly after a long-term, detailed comparison of their feedingecology.

Discussion Although these brief observationsmerely scratch the surface of the behavior and ecology of Grassland Sparrows and seed-finches,a few interesting facts emerged from them. They indicated territorial breed- ing systemssimilar to those of many other neotropical (and temperate) fringillids. Males defended territories apparently containing a sub- stantial percentageof their food requisitesat that seasonagainst con- specificmales. These areas were probably nesting territories, for most males occupyingthem seemedto be paired or, in Myospiza humeralis, associatedwith family units. The social behavior of the seed-finches closely resembledthat of Sporophila seedeaters,particularly in the lack of agonisticand courtshipdisplays other than sexual chasingand song flights and in the form and context of vocal communications.The behavior of these two genera suggeststhat they are probably closely related. Contrastingly the GrasslandSparrow had an apparent court- ship display but, perhaps because of the greater risk of predation in its more open habitat, lacked an aerial, territory advertisementdisplay. Three types of behaviorreducing interspecific'competition and facilitat- ing coexistenceamong these and related fincheswere tentatively identi- fied. Some coexistingspecies had slightly differing habitat preferences (Myospiza humeralis, Emberizoides herbicola, and Sicalis luteola in the Rupununi savannas). Other cohabitingspecies were partially ecologi- cally isolated at the time of observationby dietary differences(Myo- spiza humeralis and Sporophila castaneiventris at Georgetown; seed- finches and seedeatersat Nappi), and the two speciesof seed-finches differed in their methods of feeding on the same food plant and thus in the portionsof standsof the plant they could exploit. Differencesin feeding ecologyseemed in all casesto be related to morphologicaldif- ferencesthat permitted the use of different feedingmethods. Ecological isolation through differencesin feeding ecology is particularly common among fringillids (Lack 1971). The morphologicaladaptation of the GrasslandSparrow to its ground feeding niche in short grass savanna evidently enabledit to exploit man-madeshort grasshabitats of recent origin suchas the Georgetownuniversity campus. January 1974] Guyana Sparrows 43

AC K1N'OWLEDGi•fE1N'TS

I wish to thank the government of Guyana (Ministry of Home Affairs) for permitting me to make this field trip. Eugene Eisenmann and Dorothy Snyder offered valuable comments on the manuscript. The work was supported by grant GB 8313 from the National Science Foundation and by the New York Zoological Society.

Su•¾ Brief observationson the social organization, habitat, and feeding ecologyof the GrasslandSparrow, Large-billed Seed-Finch,and Lesser Seed-Finchin Guyana during July 1971 showedall three speciester- ritorial, with malessinging on and defendingprobable nesting territories. The form and context of sexual and agonistic interactionsand vocal communicationare described. Grassland Sparrows occupied open, short grass habitats and were ground feeders on the seeds of short stemmed grassesand on insects. The seed finches occupied forest fringe habitats and fed on the seeds of tall sedgesand grassesby perching on the raceme or using adjacent supporting vegetation. The feeding ecology and behavior of the two seed-finch species and cohabiting Sporophila seedeaterswas compared; I tentatively concluded that differences in diet and feeding behavior patterns may facilitate coexistenceamong thesespecies.

HAVn•SCaMInT, F. 1968. Birds of Surinam. Edinburgh• Oliver and Boyd. HAVmSCaMInT, F. 1970. Wattled Jacana caught by an anaconda. Condor 72: 364. LxcK, D. 1971. Ecological isolation in birds. Cambridge, Massachusetts, Harvard Univ. Press. MF.¾ER•)E SCaX•JE•SF.F..R. 1970a. A guide to the birds of South America. Nar- berth, Pennsylvania, Livingston Publ. Co. MF.¾F.RnE SC•J•JF.•SF.F.,R. 1970b. A review of the South American Finch Oryzo- borus crassirostris. Not. Nat. 428: 1-5. S•,z•R, D. E. 1966. The birds of Guyana. Salem, Massachusetts,Peabody Mus. New York Zoological Society, New York, New York 10460 and Rockefeller University, New York, New York 10021. Present address: Department of Zoology, Monash University, Clayton• Victoria 3168, Australia. Accepted 12 March 1973.