Effects of a cold wave on an Amazonian avifauna in the upper Paraguay drainage , Western Mato Grosso , and suggestions on Oscine-Suboscine relationships

The avüauna of the forested upper reaches of the Sopotuba River. in the upper Paraguay drainage just south of the Serra dos Parecís, proved a rnixture of central-Amazonian birds from just east of the Madeira Rtver and of birds of dry forests (Xingú-Tapajóz) or lower Andean forests peripheral to the southern Amazonian region . A cold wave that passed during the study period caused a major drop in insect activity and in the activity of inseci-eating birds and hummingbirds. Fruit-eating birds and "omnivores" (insect and fruit eaters) became relatively conspicuC'Ius. It is suggested that occasio· nal "clisasters" of the type of a cold wave inhibit specialized insectivores at this and other margins of Amazonian forests, and encourage hJgher percentages of omnivores . The previously noted tendency for oscines to replace suboscines at various types of neotropical forest rnargins ls reinterpreted as prlmarily due to declines of furnarioid suboscines (mostly forest-livmg insectivores> and their replacement by omnivf)rous tyrannoid suboscines anel by omnivorous oscines. The conventional theory that furnarioid suboscines are prirnitive and that they and other neotropical suboscines are being pushed into forests by recently immigrating oscines leads to two ecological anomalies: virtual absence of omnlvorous passerines in the carly Cenozoic an:i absence of seed-eat.ing passerines for much of the late Cenozoic . As alternate theory, that of separation of forest-edge suboscines by continental dritr followed by evolut.ion of forest-dwelling furnarioid suboscines in the New World, with open-country oscines either separated by the drift or entering South America early in the Ceno.zoic, removes both anomalies . In western Mato Grosso, a narrow península of Amazoni:an forest follows the Guaporé River southward and eastward on the south side of the Serra dos Parecís, a c2mpo or cerrado-covered plateau of some 500 meters'elevation. This forest península once re.ached east into the drainage of the upper Paraguay River, at least to the town of Barra ( • ) Universidade Estadual de Campinas, SP. ACTA AMAZONICA fl(3) : 379-394 Edwin O'Neill Willis (*) de Bugres. East of Barra de BugrE.s, the remarmng vegetation is mostly dry forest grading into cerrado and pantanal. From Barra de Bugres westward. the foresr is rapidly

Since birds of the Paraguay-drained eastern part of this forest had only been studied briefly by Natterer (in Pelzeln, 1867) and by Naumburg (1930) and coworkers, I welcomed an opportunity to look at birds there on an expedition of the Zoology Department of the Universidade Estadual de Campinas (UNICAMP) from 11 to 21 July, 1975.During our vísit, a major cold wave or friagem swept north, bringing frost and snow to souther~ Brazil and ending as floods in northeastern Br&,•il.This offered an opportunity to look at the effects of a friagem at the southern margin of the Amazonian forest, and contrast observations with previously reported minar effects of cold waves well to the north (such as Bates, 1962: 330).
Here I gfve the species observed, and by means of daily counts of individuais and species eating various foods illustrate certain effects of the cold wave which, magnified in reglons of greater altitude or farther south, could lead to such oft-reported avifaunal phenomena as replacement of suboscine by oscine Passeriformes at various kinds of borders of neotropical forests (Siud, 1960;Novaes, 1973).

LOCALES AND CHRONOLOGY
Few patches of forest remain along the road from Barra de Bugres 70 km west to the plateau town of Tangará da Serra.July 11 and 12, the UNICAMP expedition visited forest remnants at about 150 m elevation and 30 km W of Barra de Bugres.plus continuous zones of seasonally flooded woodlands (várzea) along the small Branco Rivcr just south.July 13 and 14, after crossi ng a zone o f dry wood!ands and of cerrado on a spur ("Serra dos Cinquenta ") of the Serra at and east of Tangará, we visited large patches of recently logged forest on the east side of the Sopotuba Ri ver (spelled "Sepotuba" and oi'her variants by Pelzeln, 1867, andNaumburg, 1930) some 20 km W of Tangará.
July 15 to 17 we worked in sandier but extensive low woodlands some 30 to 40 km NW of Tangará and beyond the Sopotuba River, on the road from the Maracanã River to the "Fazenda do Português".Near the sandstone cliffs of the Serra, some areas are very sandy, for instance, one area of Jow scrub and scrub-woodland supported singing Rhytipterna immunda, a bird that near Manaus occurs in structurally similar sandy "campinas" (Lisbôa, 1975) and in Surinam in sand-ridge scrubland (Haverschmidt, 1975).
July 18 we investigated forest remnants 20 km SW of Tangará, near Fazenda Tapirapoã on the Sopotuba River.July 19-20 we visited tall forests being cleared on Fazenda Arapitanga, about 30 km NE of town on the Sopotubinho River.Most of the forests on these days, as on July 13-14, were on deep clay riverin soils and at 200-350 m elevation.July 21 we returned to the wooded patches 30 km W of Barra de Bugres.

'I'HE "FRIAGEM"
July is the middle of the long dry winter in western Mato Grosso, Skies are normally cloudless, night temperaturas cool, and day temperaturas high.July 11-16 temperaturas became steadily warmer, reaching about 35° C at midday.At 15:00 on July 16, undulating dark clouds suddenly covered the southern sky and approached rapidly.At 16:00 wave after wave of Whiteeyed Parakeets (Aratinga leucophthalma) carne high 011er, headed northward with constant gabble.I estimated that some 500 birds crossed within my sight from a narrow road through low forest.The clouds were Jess black as they carne overhead 380-in waves of dark and light gray, and instead of rain there was a steady rush of wind twisting the treetops.Temperaturas dropped very slowly, even with the constant wind and with the undulating canopy of cloud$ racing northward.The trees near ou r camp Qroaned and there was a !ight drizzle much of the night, but the morning of the 17th was only moderately cool.At dawn, birds were normally noisy.However, the wi!"ld and clo:.Jds continued to hurry north, and temperatures barely rose beyond the low night levei.Not one of the stingless bees and "African" bees, a constant nuisance in the sandy forest the preceding two days, appeared the whole day.Two colonies of army ants (Eciton burchelli) under observation failed to move the whole day, and individuais responded as if in slow moticn even when di5turbed.Even when the sun appeared about 13:00 hardly a butterfly moved along the road.
Clear skies continued after the front passed.By 16:00 temperaturas were dropping rapidly.Night temperaturas apparently reachf'd below 3° C (temperatura recorded in Cuiabá, at 150 m elevation), with light frost reported from the tips of leaves of coffee sprouts.The following morning was very cold, anti forest birds silent.Grndually temperatures rose.but the day was still cold and insect-free.
July 19 and 20 day temperaturas in the forest continued decidedly cool, much like those of a subtropical forest.A colony of army ants started swarming only just before noon on July 19, and the ants moved 6 m per hour rather than the usual 15 or so.July 20 the ants started about 09:30, and traveled about 1 O m per hour.Butterfly and other insect activity seemed fairly high after noon on July 19 and after about midmorning on July 20, but on the second day many butterflies were dying on the road.

THE AVIFAUNA, .'\ND EFFECTS OF THE COLO WAVE
In the Appendix, species of birds observed are listed as well as numbers of individuais seen or heard in the indicated numbers of hours afield each day.A birêl marked (?) is probably the spccies indicated, but could be close relative of it.The list follows Meyer de Schauensee (1970) even when !ater revisers or my own observations suggest a different taxonomy.Types of food are indicated for each family as follows: A, nectar and insect feeders: B, fruit or seed eaters; C, fruit and insect eaters; D. insectivores; E, carnivores other than insectivores.When a species differs from the majority of its family, the fcod type is indicated after its name .Food types are from Haverschmidt (1968) and other authors, plus my own observations .Table 1 totais counts of species and individuais of food types A through D for the days July 13-20, when we were working in fairly comparable areas close to Tangará da Serra.More detailed analyses, such as timed census runs in the same area, were not possible both because I did not know that a cold wave would pass and because we had to move about for the research of others in the group.
Table 1 indicates that on the two coldest days, July 17 and 18, the number of specie& observed was about one-half of normal and the number of individuais about one-third of normal.(The waves of parakeets rai&ed the total unduly on July 16; probably 50 or less of the 500 would have been seen on an crdinary day, and 50 i s the number used here for certa in comparisons).Specifically, after the normal!ynoisy dawn of July 17 (indicating that cold itself did not dampen calling and song until the birds had to look for food).forest birds became silent and hard to locate.Parrots and pigeons continued relatively noisy, but even th::>y were rather silent July 18.On July 19 and 20, the amount of calling increased to leveis possibly but not clearly below the usual forest levei.
Table 2 separates the birds by food types A through D .Some of the u insectivores" may also eat some fruits, as diets of most species are not well known.Hummingbirds (Type A) most strongly showed effects of the cold wave.None was seen on the 17th, one on the 18th, and one (a Heliomaster furcifer very weak of flight in a sunlit treetop) on the 19th .Fair numbers were active at well-lighted treetop sites and in some areas of undergrowth on the 20th.Hummingbirds are known to enter torpor at night and during cold periods, and it is possible that some individuais did so during much of the cold wave.
Fruit-eating birds remained active and fairly conspicuous decreasing in absolute numbers seen but rising slightly in percentage of the avifauna, from the normal 10 or 15 percent of species to about 20 percent on the 18th, and from the normal 25 percent of the individuais seen to 40 percent on that day and the next.
"Omnivorous" birds, eating fruit and insects, were less frequently observec!during the cold wave but remained at about the usual percentages of the avifauna -40 percent of the species and slightly over 35 percent of the individuais.These species continued active and fairly conspicuous, but flocked to fruiting trees instead of hunting insects (except in sunlit areas at the warmest hours of afternoon.when they stayed close to the fruit trees in their hunts).lnsectivorous birds were quiet and inconspicuous, though the few seen seemed to forage actively.Percentages decreased from a normal species percentage of 40 to 31 on the 18th, and from a normal 35 or so percent The analyzable avifauna (species other than widely distributed ones without dis• tinctive subspecies) of the forests near Tangará da Serra is most closely related to the avifauna of forests east of the rio Madeira, but shows some elements from other regions, such as lower Andean forests and dry forests south and east of the Tapajóz River.The relationships are most clearly shown by birds that regularly follow army ants, fcr most occur only or mainly east of the Madeira: Rhegmatorhina hoffmannsi, an anl-follow• ing species of the east bank of the Madeira, is replaced by related species west of the Madeira and on both sides of the lower Tapajóz (Willis, 1969).Oendroco/aptes hoffmannsi ranges from the Madeira to the Xingú.and is replaced by different subspecies of the related O. picumnus to north, south, northeast, and west .(Local distribution 'lf birds of this group is complex.For instance, Naumburg, 1930, reports that near Tangará the related D. platyrostris wedges in from the east, presumably in gallery woodlands on the open planalto of the Serra, and O. picumnus pal/escens wedges up from the south, in pantanal or semiopen woodlands).Hyle><etastes perroti uniformis is widespread between the Madeira and the Xingú, but is replaced by H. stresemanni westward.
The Dendrocincla fuliginosa observed here were probably of the subspecies atrothorax from between the Madeira and Tapajóz Rivers, for they were streaked and had the unusual call characteristic of this subspecies; this subspccies is already known from the Guaporé River (I do not, however.know O. f. trumaii from the upper Tapajóz-Xingú region, and suspect it may not be distinguishable in the field).The Oendrocincla merula seemed of the subspecies o/ivascens from the Madeira to the Tapajóz, not of bartletti or remota from west of the Madeira nor of castanoptera from east of the Tapajóz .Ph/egopsis nigromaculata seemed of the large-spotted and white-sided form bowmani, which ranges from the Madeira to the Xingú.The-;e species or distinctive subspecies suggest entrance of the avifauna east

382-
ot the Madeira, either up the Guaporé or across the Serra dos Parecís trom the Gi-Paraná River (Machados) or western heaclwaters ot the Tapajóz.Tangará da Serra matks the southwestern limit of the known ranges of most of these species, but many may possibly be tound to circle the upper Guaporé into eastern Santa Cruz, Bolívia .I did not see Dendrocolaptes concolor, an ant-following bird reported from the Madeira to the Xingú and replaced by subspecies of D. certhia to east ant west, but concolor has been reported from the upper Guaporé (Pelzeln, 1867) and probably occurs near Tangará.
In contrast to the above, Hylophyfax poecilonota griseiventris ot Tangará is a form that occurs both east and west of the Madeirêa.Pyriglena /euconota maura, which I encountered only in sandy low woodland near the Serra, is the southern representative of a species that crosses central Brazil in dry torests and extends northewestward in subtropical forests along the Andes.
Certain species of the region ot Tangará da Serra seem species of the gallery woodlands and dry-forested belt extending from east of the Tapajóz River acmss into eastern Bolívia, even though the belt seems interrupted by such large areas of open vegetation as the top of the Serra dos Parecis.Ara maracana, Ara nobilis, Brachyga/ba lugubris, Captto dayi, Herpsilochmus longirostris, Neope/ma pal/escens, Odontorchilus cinereus, and possibly Hylophilus pectoralis are here re-Effects or a cold wavc .
corded at or near their southwestern limits in this belt, and do not occur on the Andes or in humid forests on the lower Madeira.
The analyzable avifauna of the torests near Tangará thus seems a mixture of species of humid Amazonian forests, entering primarily trom the north along the east side of the Madeira, and southern species ot upper tropicC~I or dry forests fringing the central part of Amazonia .Many species widespread In humid central Amazonia 'ieemed absent, at least in this briet survey: Tinamus guttatus, Tinamus major, Pionites /eucogaster, Gypopsitta vulturina, Thamnomanes saturninus, etc.The vegetation is not the tall and complex forest ot central Amazonia, but a torest of moderate (rarely low) height and moderately complex structure; and the climate is marked by a long dry "winter • .Probably the mixture of centrai-•Amazonian and peripherai-Amazonian fore~t birds is what one could expect from the subequatorial latitude (nearly 15° S). climate and vegetation.
Studies of butterflies have also indicated a mixture of Peruvian-Bolivian (upper tropical), Xingú-Tapajóz (dry forest).and central Amazonian forms, but without strong restriction to Amazonian forms from east ot the Madeira; the Madeira seems less important as a ba~rier for butterflies than for birds (Brown, 1972: 191 and pers. comm.).Buttertlies show some endemic forms, suggesting to Brown a former forest "refuge" in the region during dry climatic periods, but the avifauna seems a mixture-as of recolonization after a dry epoch.Haffer (1969Haffer ( , 1974) ) does not indicate that the region was an important refuge or center of speciation for birds.lt is possible.however, that further west along the Serra dos Parecís are better torests and better avifaunas-certainly the butterfly fauna is richer (K. S. Brown.pers.comm.).

DISCUSSION
In the context of the "peripheral" (to the Amazonian "center ") nature ot this avi fauna and forest, the obgervations of birds during the cold wave <;uggest some phenomena which, amplified even more peripherally by more frequent or stronger cold, might lead to replacement of suboscine Passeriform birds by oscines in those more peripheral zones.lnsectlvorous birds became less conspicuoun during the cold wave, while frugivorous and omnivorous ones remained relatively consp!cuous.This suggests that insectivorous birds had difficulties finding prey, and had no time or extra energy for singing or the like; and that frugivores and omnivores moved to fruiting trees where food rema ined available despite the cold.The decreased insect activity was obvious even to an ornithologist, and was confirmed by entomologists in our party (K. S. Brown, Jr.; Mohamed Habib).Fruit, by contrast.comes from larga and resistant organisms, which do not become unavailable in such brief fluctuations as a cold wave (or do so only after the wave has passed.Observations in central São Paulo state and in lguaçú National Park after the cold wave indicated that many trees lost leaves, and that a few were killed).
In terms of avifaunas, the difficulties of insectivores in extreme weather are significant because a large group of suboscines, the furnarioid suboscines (families Formicariidae, Furnariidae, Dendrocolaptidae), are essentially insectivorous (Appendix).Tyranno:d suboscines (Cotingidae.Pipridae, Tyrannidae) and oscines (Hirundinidae and following families in the appendix), by contrast, are mostly omnivores, rummage in protected places for food, or can migrate In unfavorable periods.In general.furnarioid subosclnes dominate in tall Amazonian forests and tyrannoids and oscines take over in any slightly lower vegetation peripherally (Table 3 compares passerines of a forest near Manaus with forest~ near Tangarã and with a forest even further south in São Paulo) --whether the low vegetation is due to seasonal flooding (várzeas).droughts (caatingas, desert scrubs), cold (subtropical forests).humans (capoeira; Novaes, 1973).or other reasons.The Cotingidae and Pipridae.which are primarily fruit eaters, also occur in good numbers in or near the edges of tall Amazonian forests, as do many frUit-eating nonpasseriform birds (Tinamidae, Cracidae, Psittacidae, etc.).lt seems that, under the relatively constant condltions of central Amazonia, specialists either on insects (fumarioid suboscines) or fruits (birds in the last sentence) are favored.Slightly further peripherally, occasional cold waves or other • disasters" increasingly make specialized insectivores vulnerable, and they are replaced by generalized omn!vores that can swi1ch back and forth between insects and fruit.
The contrast between central and peripheral Amazonian distributions of furnarioid suboscines and other passeriforms thus is likely to have at least one ecological basis, that ot occasional disasters which cauc;e difftculties for insectivores.Table 4 compares numbers of species of insectivores, omnivores, and frugivores near Manaus (3° S).Tangará (15° S), and in São Paulo (23° S), showing the lower percentage of species of pure insectivores near Tangará.Some individual insectivores may be killed outright, at least in regions where cold waves or other disasters are more frequent and stronger than at Tanga-  1974) .
Fruit and insect eaters, however.af~ more likely to survive brief cold waves, or can use fruits for adults while feeding insects to nestlings, thus increasing the number of young they can produce in short nesting seasons .
Fruit eaters also have the advantage of greater mobility, allow1ng movements to better sites, than do the generally territorial or semiterritorial insectivores .However, fruit eaters have to turn to insects, or move about, in "ha!•d times ft for fruits (Morton, 1973).
One finds a rise in insectivorous speci€s rather than a continued decrease as one goes even further from tall Amazonian forests, ini.:o the " ring '' of habit~ls beyond the "peritropical ring '' .(The data from São Paulo in Tables 3  and 4 provide an example) .Furnariidae and Parulidae, plus a number of insectivorous Tyrannidae, occur out in wide open or semiope:n zones or in temperate-zone forests.The explanation here, as far as my own observations go, seems not so much that fruits are rare (being replaced by seeds) but that many of these bird species are migratory, eat insects in enclosed places, or have other ways of avoiding bad periods, which are so common as to be predictable .Closer to ta li Amazonian forests, these species are probabiy at a disadvantage in direct competition with insectivores which specialize and stay on territories (Willis, 1966). in the long periods between d1sasters or hard times .
Perhaps the constantly favorable " tropical rain forest" has specialized birds, the next unpredictably bad o r semiopen "ring " of habitats has generalists, and the cc nstantly unfavorable and more open second "ring" of habitats has speci alists again?local lists of the type here, from a wide variety of places and habitats, might show if thi~ is indeed a rule of wide applicability.
The ornithologist must ask the additional question: why there is a correlation between major groups of passerines and their food types?Was there no selection for fruit-eating furnarioids , which could thus have occupied niches now occupied by tyrannoids and oscines?Why is there no furr.arioidor tyrannoid that eats seeds?The oscines that eat seeds mostly live in lower or more open habitats of the " second ring ", and hence h ave not been considered above except in the list of birds of open areas in the region of Tangará da Serra; but they and the suboscines that do not eat seeds add odd distortions to the generally accepted or " conventional" view of evolution of neotropical passerines.
The conventional view is that neotropical suboscines (especially t he " an: estral " furnarioids) are primitive forms that have been retreating into tropical forests ever since the Pliocene closure of the Panamani an land bridge allowed oscines to flood south (Amadon.1973).Some authors (Siud , 1960;Willis, 1966) have indicated doubts about a theory that requires oscines not to fly over water cmd that requires extensive flow of species from species-poor habitats (open ones; Central and North America) to speci es-rich ones (forested: South America).but the oddities of food types raise further doubts .In ecology, major niches are rarely left unfilled for long; but the conventional theory seems to require unfilled (by birds) fruit-eating niches prior to tyrannoid appearance, plus ur.filled seed-eating niches until the Pliocene.I would like '!"o suggest an alternative possibility for evolution of major groups of neotropical passerines: that suboscines that t ook to eating insects in closed neotropical habitats became furnarioids, related birds that earlier took to fruit-insect use in lower vegetation (in both New and Old World) became various suboscines, including tyrannoids, while those early passerines that took to various types of feeding (including seed eating) i n open areas became oscines.To use open habitats, a complex song and hence complex syringeal musculature are advantageous (Morton, 1975).lt is possible that, in intermediate habitats, the variably simple to intermediate syringes of tyrannoids are sufficient; and that in closed habitats.the peculiar "tracheophone" syringes of furnarioids produce greater volume even i f less modulation.Migrations, pointed or nine-primaried wings, and many other open--country adaptations of oscines are not necessarlly advances over furnarioides; just necessary given an open-country environment for oscines.
The alternative view would suggest that oscines were prcsent long ago in the New World, and that their competition prevented development of seed-eating suboscines there as in the Old World.Separation of oscines by continental drift in late Cretaceous or entry into South America by waif dispersai by the Oligocene are possible, sínce Primates and rodents somehow got to South America by the Oligocene (Simpson, 1969), and since Brodkorb ( 1960) suggests that seeds led to blrd evolution in the M iocene.New World seedeaters (Fringillidae sensu lato, including Emberizinae and Cardinalinae) could have reinvaded the Old World, explaining why there are two main groups of seedeating Oscines (Fringillidae; and Ploceidae of the Old World).The nearest relatives of Fringillidae are nine--primaried tanagers and other New-World birds, and I think it more likely that Fringillidae have moved recently into the Old World than

386-
that they have moved recently into the New.Reversing the direction of fringillid (and hence nine-primaried oscine) movement almost eliminates the supposed recent movement of oscines into the neotropics.Wrens, mimids, thrushes, and jays are small groups that may have moved in after closure of land bridge~ in the Pliocene, as suggested by Amadon.
I agree with Keast ( 1972) that there was probably not much movement intt' South America prior to the Pliocene, thus leaving nine-primaried oscines and suboscines (especially tyrannoic!s)free to radiate; but suggest that one entrance early led to the huge nine-primaried radiation.
Cracraft (1973) follows Keast and earlier authors (Mayr, 1964) in suggesting that nine--primaried oscine::; arose in tropical North America, or that the most primitive ones (vireos, warblers) did so, even though most groups are South American.Considering that (a) South America is likely to have been southeast of North America throughout the Cenozoic; (b) nine-primaried passerines are adapted for long-distance flights (Averill, 1925) it seems likely that (c) some migrating nine--primaried birds would have flown over water to or from South America and radiated southward almost as quickly as they did northward.Thus, evolution of nine-primaried oscines in North America until the Pliocene , followed by their flooding southward, still seems unlikely.A northward direction of dispersai, from South to North America, is less likely because fewer southern forms migrate well northward.However, survival of certa in primitive migratory nine-primaried oscines mainly in North America could have resulted from relative lack of competition from subosclnes, rather than from radlation there .Cox (1968) has suggested that these migrants are mainly ones that were interspecifically outcompeted toward the equator; if so, the drift of continents raises anew the • southern ancestral home" theory in an ecological version for these migrants, with the twist that they would have been avoiding a highly developed and competitive avifauna by migrating northward across the Panamá gap.lt does not seem likely, under e!ther the conventional or alternate views .that nine-•primaried and other oscines originated separately in Old and New Worlds.lt is true that Sibley (1970) suggests wrens show resemblances to furnariids in egg-white proteins.lt also is interesting that the clambering behavior and plush crown of the oscine Catablyrhynchus diadematus suggest the furnariid Metopothryx aurantiacus, while the clambering furna1 iid Xenerpestes minlosi looks like a leaf-clinging Parulid.However.Sibley indicates resemblarce of egg-white proteins between nine-primaried oscines and severa I Old-World groups.These and thP.similarities of oscine syringes in general (Ames.1971) throw doubt on separate origin of nine-primaried oscines from local suboscines; it seems more likely that oscines had a common origin and were separated by continental drift or immigrated early to South America.
Under both thc conventlonal and alternative view.suboscines have been unsuccessful in the Old World .Whether they were ever common t here is not clear; but the presence of Eurylaimidae in Africa and Asia and of Philepittidae in Madagascar indicates that suboscines evolvcd before continental drift.The small size and historical vagarias of Old-World tropical forests could have given the advantage to forest-edge oscines over forest-living suboscines.The absence of furnarioíds from the Old World , h.Jwever, índicates that this group may well have evolved in the New World after Africa separated from South America.perhaps from tyrannoid ancestors (Sibley, 1970, doubts that Old-World suboscines are tyrannoids; if he is right, tyrannoids and furnarioids could both postdate drift).
The ancient or pre-continental drift nature of oscines is possible but uncertain under both the alternative and conventlonal views.Oscines occur on Madagascar, in Australia , and in New Zealand.three areas much more isolated than is South America.Some could have wandered from nearby continents, for open country oscines are usually good fliers and cross water gaps easily; but orhers in Effects or a cold wave .. .each region seem primitiva (as the lyrebirds of Australia and the Acanthisittidae of New Zealand).suggesting derivativas of oscines present before continental drift.Absence of suboscines in New Zealand and Australia coul d be due to their extinction or failure to reach adversely open environments there.rather than an indication that subosclnes arose after separation of Australia.
Oscines seem unlikely to have descended from birds presently called "suboscines •• which have a bony stapes differwt from that of other Aves and from Reptilia (Ft::duccia, 1974,1976)."Suboscines", with thelr variable syringes (Ames.1971) could easily have descended from primitiva oscines, before or as the latter developed their complex syringes.Possibly, to avoid confusion, the "suboscínes" with the modified stapes should in future works be given a noncommittal name like "paraoscínes ". re<>erving the name "suboscines" for fossil early passerines of simple syringes and primitiva stapes.
A CKNOWLEDGEMENTS appreciate the support of the Universidade Estadual de Campinas for the expedition and the aid of other members of the group.especially Keith S. Brown.Jr .. W . Be!lson.K. Brown.D. Pearson.E. Eisenmann and Yoshika Oniki helped with the manuscript.

TABLE 2
Percentages of Species and Individuais of Forest Birds by Food Type o .When 450 "extra" porokeets excludod from total.