Do feathered dinosaurs exist? Testing the hypothesis on neontological and paleontological evidence

The origin of birds and avian flight from within the archosaurian radiation has been among the most contentious issues in paleobiology. Although there is general agreement that birds are related to theropod dinosaurs at some level, debate centers on whether birds are derived directly from highly derived theropods, the current dogma, or from an earlier common ancestor lacking suites of derived anatomical characters. Recent discoveries from the Early Cretaceous of China have highlighted the debate, with claims of the discovery of all stages of feather evolution and ancestral birds (theropod dinosaurs), although the deposits are at least 25 million years younger than those containing the earliest known bird Archaeopteryx. In the first part of the study we examine the fossil evidence relating to alleged feather progenitors, commonly referred to as protofeathers, in these putative ancestors of birds. Our findings show no evidence for the existence of protofeathers and consequently no evidence in support of the follicular theory of the morphogenesis of the feather. Rather, based on histological studies of the integument of modern reptiles, which show complex patterns of the collagen fibers of the dermis, we conclude that "protofeathers" are probably the remains of collagenous fiber "meshworks" that reinforced the dinosaur integument. These "meshworks" of the skin frequently formed aberrant patterns resembling feathers as a consequence of decomposition. Our findings also draw support from new paleontological evidence. We describe integumental structures, very similar to "protofeathers," preserved within the rib area of a Psittacosaurus specimen from Nanjing, China, an ornithopod dinosaur unconnected with the ancestry of birds. These integumental structures show a strong resemblance to the collagenous fiber systems in the dermis of many animals. We also report the presence of scales in the forearm of the theropod ornithomimid (bird mimic) dinosaur, Pelecanimimus, from Spain. In the second part of the study we examine evidence relating to the most critical character thought to link birds to derived theropods, a tridactyl hand composed of digits 1-2-3. We maintain the evidence supports interpretation of bird wing digit identity as 2,3,4, which appears different from that in theropod dinosaurs. The phylogenetic significance of Chinese microraptors is also discussed, with respect to bird origins and flight origins. We suggest that a possible solution to the disparate data is that Aves plus bird-like maniraptoran theropods (e.g., microraptors and others) may be a separate clade, distinctive from the main lineage of Theropoda, a remnant of the early avian radiation, exhibiting all stages of flight and flightlessness.     

The early history of modern birds inferred from DNA sequences of nuclear and mitochondrial ribosomal genes

The traditional view of avian evolution places ratites and tinamous at the base of the phylogenetic tree of modern birds (Neornithes). In contrast, most recent molecular studies suggest that neognathous perching birds (Passeriformes) compose the oldest lineage of modern birds. Here, we report significant molecular support for the traditional view of neognath monophyly based on sequence analyses of nuclear and mitochondrial DNA (4.4 kb) from every modern avian order. Phylogenetic analyses further show that the ducks and gallinaceous birds are each other's closest relatives and together form the basal lineage of neognathous birds. To investigate why other molecular studies sampling fewer orders have reached different conclusions regarding neognath monophyly, we performed jackknife analyses on our mitochondrial data. Those analyses indicated taxon-sampling effects when basal galloanserine birds were included in combination with sparse taxon sampling. Our phylogenetic results suggest that the earliest neornithines were heavy-bodied, ground-dwelling, nonmarine birds. This inference, coupled with a fossil bias toward marine environments, provides a possible explanation for the large gap in the early fossil record of birds.     

Ultra-structural identification of interstitial cells of Cajal in the zebrafish Danio rerio

The interstitial cells of Cajal (ICCs) are important mediators of gastrointestinal (GI) motility because of their role as pacemakers in the GI tract. In addition to their function, ICCs are also structurally distinct cells most easily identified by their ultra-structural features and expression of the tyrosine kinase receptor c-KIT. ICCs have been described in mammals, rodents, birds, reptiles, and amphibians, but there are no reports at the ultra-structural level of ICCs within the GI tract of an organism from the teleost lineage. We describe the presence of cells in the muscularis of the zebrafish intestine; these cells have similar features to ICCs in other vertebrates. The ICC-like cells are associated with the muscularis, are more electron-dense than surrounding smooth muscle cells, possess long cytoplasmic processes and mitochondria, and are situated opposing enteric nervous structures. In addition, immunofluorescent and immunoelectron-microscopic studies with antibodies targeting the zebrafish ortholog of a putative ICC marker, c-KIT (kita), showed c-kit immunoreactivity in zebrafish ICCs. Taken together, these data represent the first ultra-structural characterization of cells in the muscularis of the zebrafish Danio rerio and suggest that ICC differentiation in vertebrate evolution dates back to the teleost lineage.     

Linking the molecular evolution of avian beta (β) keratins to the evolution of feathers

Feathers of today's birds are constructed of beta (β)-keratins, structural proteins of the epidermis that are found solely in reptiles and birds. Discoveries of "feathered dinosaurs" continue to stimulate interest in the evolutionary origin of feathers, but few studies have attempted to link the molecular evolution of their major structural proteins (β-keratins) to the appearance of feathers in the fossil record. Using molecular dating methods, we show that before the appearance of Anchiornis (～155 Million years ago (Ma)) the basal β-keratins of birds began diverging from their archosaurian ancestor ～216?Ma. However, the subfamily of feather β-keratins, as found in living birds, did not begin diverging until ～143?Ma. Thus, the pennaceous feathers on Anchiornis, while being constructed of avian β-keratins, most likely did not contain the feather β-keratins found in the feathers of modern birds. Our results demonstrate that the evolutionary origin of feathers does not coincide with the molecular evolution of the feather β-keratins found in modern birds. More likely, during the Late Jurassic, the epidermal structures that appeared on organisms in the lineage leading to birds, including early forms of feathers, were constructed of avian β-keratins other than those found in the feathers of modern birds. Recent biophysical studies of the β-keratins in feathers support the view that the appearance of the subfamily of feather β-keratins altered the biophysical nature of the feather establishing its role in powered flight.     

Birds perching on bushes: Networks to visualize conflicting phylogenetic signals during early avian radiation

Hybridization is increasingly seen as an important source of adaptive genetic variation and biotic diversity. Recent phylogenetic studies on the early evolution of birds suggest that the early diversification of neoavian orders perhaps involved a period of extensive hybridization or incomplete lineage sorting. Phylogenetic error, saturation, long-branch attraction, and convergence make it difficult to detect ancient hybridization events and differentiate them from incomplete lineage sorting using sequence data. We used recently published retroposon marker data to visualize the early radiation of Neoaves within a phylogenetic network approach, and found that the most basal neoavian taxa indeed show a complex pattern of reticulated relationships. Moreover, the reticulation levels of different parts of the network are consistent with the insertion pattern of the retroposon elements. The use of network-based analyses on homoplasy-free data shows true conflicting signals and the taxa involved that are not represented in trees.     

Vestigial skeletal structures in dinosaurs

The existence of vestigial structures is one of the main lines of evidence for macroevolution. Here I introduce a phylogenetic bracketing approach to the identification of vestigial structures and apply it to Dinosauria. According to this approach, a structure is considered vestigial if, in comparison with its homolog in at least three successive outgroups, it is reduced to one-third or less its size relative to adjacent structures and if at least distally it has lost the specialized morphology present in the three outgroups. This approach identifies fingers IV and V as vestigial in dinosaurs in general, II–V in sauropods, III in Tyrannosauridae and Caudipteryx , II and III in Shuvuuia and I and III in modern birds. The entire forelimb distal to the elbow is vestigial in Abelisauridae. Vestigial parts of the pelvic girdle and hindlimb include the pubic shaft in Iguanodontia and Ceratopsia, the entire pubis in Ankylosauria, the first metatarsal in derived Iguanodontia, the first metatarsal shaft in Theropoda and the fifth toe in dinosaurs in general. Derived Centrosaurinae and some Chasmosaurus exhibit vestigial supraorbital horns. Some centrosaurines have a vestigial nasal horn. The antorbital cavity is vestigial in Thyreophora, Iguanodontia and Ceratopsidae. I recommend that this information be exploited to increase public awareness of the evidence for macroevolution.     

Phylogeography and genetic structure of northern populations of the yellow warbler (Dendroica petechia)

Phylogeographic patterns of intraspecific variation can provide insights into the population-level processes responsible for speciation and yield information useful for conservation purposes. To examine phylogeography and population structure in a migratory passerine bird at both continental and regional geographical scales, we analysed 344 bp of mitochondrial DNA (mtDNA) control region sequence from 155 yellow warblers (Dendroica petechia) collected from seven locations across Canada and from Alaska. There is a major subdivision between eastern (Manitoba to Newfoundland) and western (Alaska and British Columbia) populations which appears to have developed during the recent Pleistocene. Some localities within these two regions also differ significantly in their genetic composition, suggesting further subdivision on a regional geographical scale. Eastern and western birds form distinct phylogeographic entities and the clustering of all western haplotypes with two eastern haplotypes suggests that the western haplotypes may be derived from an eastern lineage. Analyses based on coalescent models support this explanation for the origin of western haplotypes. These results are consistent with important features of Mengel's model of warbler diversification. From a conservation perspective they also suggest that individual populations of migrant birds may form demographically isolated management units on a smaller scale than previously appreciated.     

T cell nature and heterogeneity of recognition structures of human natural killer (NK) cells

Very low doses of trypsin (5 micrograms/ml) are sufficient to ablate NK cell activity. This finding was used to make several observations, and we have attempted to relate these observations to specific cell surface macromolecules. First, trypsinized effector cells no longer lysed seven different NK-susceptible targets, but the lysis of three additional targets was unaffected. These results suggest a heterogeneity of recognition potential that is inconsistent with the notion that there is only one class of NK "receptors" and one class of "target structures." Trypsin does not affect the conjugation of effector and target cells. Secondly, we have tried to identify those cell surface molecules that are affected by this low dose of enzyme. The examination of the 125I-labeled glycoprotein fraction NK-enriched cells showed that at least four molecules are cleaved, one of which may be in the T200 family. The examination of the [3H]galactose-labeled cell surface glycoproteins suggested in particular that some high m.w. glycoproteins were affected at the dose of trypsin that ablates NK function. Analysis of those molecules that we previously implicated in NK function, defined by monoclonal antibodies that block NK lysis, allowed us to rule out a role for the Tp 50 and Lp95-150 structures, while providing additional evidence of a role for the T200 glycoproteins in the trypsin-sensitive stage of cytolysis. Finally, closer examination of the electrophoretic mobilities and trypsin sensitivity of the T200 structures on highly enriched NK cells showed these structures to be indistinguishable from the T cell form of T200, yet quite distinct from the monocyte form. These results are therefore consistent with the possibility that NK cells are of the T rather than the monocyte lineage, and furthermore support a role for the T200 structure in the post-binding trypsin-sensitive stage of the NK cytolytic process.     

A separate circadian oscillator controls nocturnal migratory restlessness in the songbird Sylvia borin

When confined to a cage, migratory songbirds exhibit nocturnal migratory restlessness (also called Zugunruhe) during the spring and autumn migratory periods, even though these birds are exclusively diurnal during the remainder of the year. Zugunruhe, which has been demonstrated to be under the direct control of a circannual timer, is characterized by a stereotypic "wing-whirring" behavior while the bird is perched. To elucidate the role played by the circadian system in the regulation of Zugunruhe, the authors studied the activity of garden warblers (Sylvia borin), long-distance nocturnal migrants, under skeleton photoperiods of different lengths and under constant dim light. In 11.5D:1L:10.5D:1L skeleton photoperiods, the authors found that Zugunruhe free-ran in a substantial proportion of birds, while their normal daily activities (e.g., feeding and preening) remained synchronized to 24 h. Some birds expressing Zugunruhe under constant dim light continued to show 2 distinct bouts of activity: one corresponding to daily activities, the other to wing-whirring. In some cases, these 2 bouts crossed while free-running with different periods. Birds expressing Zugunruhe also had significantly longer free-running periods than birds that did not. The study data suggest that the seasonal appearance of Zugunruhe is the result of the interactions of at least 2 circadian oscillators and that it is the phase relationship of these 2 oscillators that determines when nocturnal migratory restlessness is expressed. Furthermore, these data are consistent with the previously proposed internal coincidence hypothesis as a model for the ontogeny of circannual rhythms.     

5 S and 5.8 S ribosomal RNA sequences and protist phylogenetics

More than 100 5 S 5.8 S rRNA sequences from protists, including fungi, are known. Through a combination of quantitative treeing and special consideration of "signature' nucleotide combinations, the most significant phylogenetic implications of these data are emphasized. Also, limitations of the data for phylogenetic inferences are discussed and other significant data are brought to bear on the inferences obtained. 5 S sequences from red algae are seen as the most isolated among eukaryotics. A 5 S sequence lineage consisting of oomycetes, euglenoids, most protozoa, most slime molds and perhaps dinoflagellates and mesozoa is defined. Such a lineage is not evident from 5.8 S rRNA or cytochrome c sequence data. 5 S sequences from Ascomycota and Basidiomycota are consistent with the proposal that each is derived from a mycelial form with a haploid yeast phase and simple septal pores, probably most resembling present Taphrinales. 5 S sequences from Chytridiomycota and Zygomycota are not clearly distinct from each other and suggest that a major lineage radiation occurred in the early history of each. Qualitative biochemical data clearly supports a dichotomy between an Ascomycota-Basidiomycota lineage and a Zygomycota-Chytridiomycota lineage.     

Did trypanosomatid parasites have photosynthetic ancestors?

Some molecular phylogenies of plastid-like genes suggest that chloroplasts (the structures responsible for photosynthesis in plants and algae) might have been secondarily lost in trypanosomatid parasites. Chloroplasts are present in some euglenids, which are closely related to trypanosomatids, and it has been argued that chloroplasts arose early in the diversification of the lineage Euglenozoa, to which trypanosomatids and euglenids belong (plastids-early hypothesis). This article reviews how euglenid ultrastructural systems are functionally integrated and phylogenetically correlated. I argue that chloroplast acquisition profoundly altered the structure of certain euglenids, and that the complete absence of these modifications in other euglenozoans is most consistent with their never having had a chloroplast. Ultrastructural evidence suggests that chloroplasts arose relatively recently within a specific subgroup of euglenids and that trypanosomatids are not secondarily non-photosynthetic (plastids-recent hypothesis).     

The evolution of endothermy in terrestrial vertebrates: Who? When? Why?

Avian and mammalian endothermy results from elevated rates of resting, or routine, metabolism and enables these animals to maintain high and stable body temperatures in the face of variable ambient temperatures. Endothermy is also associated with enhanced stamina and elevated capacity for aerobic metabolism during periods of prolonged activity. These attributes of birds and mammals have greatly contributed to their widespread distribution and ecological success. Unfortunately, since few anatomical/physiological attributes linked to endothermy are preserved in fossils, the origin of endothermy among the ancestors of mammals and birds has long remained obscure. Two recent approaches provide new insight into the metabolic physiology of extinct forms. One addresses chronic (resting) metabolic rates and emphasizes the presence of nasal respiratory turbinates in virtually all extant endotherms. These structures are associated with recovery of respiratory heat and moisture in animals with high resting metabolic rates. The fossil record of nonmammalian synapsids suggests that at least two Late Permian lineages possessed incipient respiratory turbinates. In contrast, these structures appear to have been absent in dinosaurs and nonornithurine birds. Instead, nasal morphology suggests that in the avian lineage, respiratory turbinates first appeared in Cretaceous ornithurines. The other approach addresses the capacity for maximal aerobic activity and examines lung structure and ventilatory mechanisms. There is no positive evidence to support the reconstruction of a derived, avian-like parabronchial lung/air sac system in dinosaurs or nonornithurine birds. Dinosaur lungs were likely heterogenous, multicameral septate lungs with conventional, tidal ventilation, although evidence from some theropods suggests that at least this group may have had a hepatic piston mechanism of supplementary lung ventilation. This suggests that dinosaurs and nonornithurine birds generally lacked the capacity for high, avian-like levels of sustained activity, although the aerobic capacity of theropods may have exceeded that of extant ectotherms. The avian parabronchial lung/air sac system appears to be an attribute limited to ornithurine birds.     

The Complete Mitochondrial Genome of Aix galericulata and Tadorna ferruginea: Bearings on Their Phylogenetic Position in the Anseriformes

Aix galericulata and Tadorna ferruginea are two Anatidae species representing different taxonomic groups of Anseriformes. We used a PCR-based method to determine the complete mtDNAs of both species, and estimated phylogenetic trees based on the complete mtDNA alignment of these and 14 other Anseriforme species, to clarify Anseriform phylogenetics. Phylogenetic trees were also estimated using a multiple sequence alignment of three mitochondrial genes (Cyt b, ND2, and COI) from 68 typical species in GenBank, to further clarify the phylogenetic relationships of several groups among the Anseriformes. The new mtDNAs are circular molecules, 16,651 bp (Aix galericulata) and 16,639 bp (Tadorna ferruginea) in length, containing the 37 typical genes, with an identical gene order and arrangement as those of other Anseriformes. Comparing the protein-coding genes among the mtDNAs of 16 Anseriforme species, ATG is generally the start codon, TAA is the most frequent stop codon, one of three, TAA, TAG, and T-, commonly observed. All tRNAs could be folded into canonical cloverleaf secondary structures except for tRNASer (AGY) and tRNALeu (CUN), which are missing the "DHU" arm.Phylogenetic relationships demonstrate that Aix galericula and Tadorna ferruginea are in the same group, the Tadorninae lineage, based on our analyses of complete mtDNAs and combined gene data. Molecular phylogenetic analysis suggests the 68 species of Anseriform birds be divided into three families: Anhimidae, Anatidae, and Anseranatidae. The results suggest Anatidae birds be divided into five subfamilies: Anatinae, Tadorninae, Anserinae, Oxyurinae, and Dendrocygninae. Oxyurinae and Dendrocygninae should not belong to Anserinae, but rather represent independent subfamilies. The Anatinae includes species from the tribes Mergini, Somaterini, Anatini, and Aythyini. The Anserinae includes species from the tribes Anserini and Cygnini.     

Evolutionary dynamics of intron size, genome size, and physiological correlates in archosaurs

It has been proposed that intron and genome sizes in birds are reduced in comparison with mammals because of the metabolic demands of flight. To test this hypothesis, we examined the sizes of 14 introns in a nonflying relative of birds, the American alligator (Alligator mississippiensis), and in 19 flighted and flightless birds in 12 taxonomic orders. Our results indicate that a substantial fraction (66%) of the reduction in intron size as well as in genome size had already occurred in nonflying archosaurs. Using phylogenetically independent contrasts, we found that the proposed inverse correlation of genome size and basal metabolic rate (BMR) is significant among amniotes and archosaurs, whereas intron and genome size variation within birds showed no significant correlation with BMR. We show statistically that the distribution of genome sizes in birds and mammals is underdispersed compared with the Brownian motion model and consistent with strong stabilizing selection; that genome size differences between vertebrate clades are overdispersed and punctuational; and that evolution of BMR and avian intron size is consistent with Brownian motion. These results suggest that the contrast between genome size/BMR and intron size/BMR correlations may be a consequence of different intensities of selection for these traits and that we should not expect changes in intron size to be significantly associated with metabolically costly behaviors such as flight.     

Local parasite lineage sharing in temperate grassland birds provides clues about potential origins of Galapagos avian Plasmodium

Oceanic archipelagos are vulnerable to natural introduction of parasites via migratory birds. Our aim was to characterize the geographic origins of two Plasmodium parasite lineages detected in the Galapagos Islands and in North American breeding bobolinks (Dolichonyx oryzivorus) that regularly stop in Galapagos during migration to their South American overwintering sites. We used samples from a grassland breeding bird assemblage in Nebraska, United States, and parasite DNA sequences from the Galapagos Islands, Ecuador, to compare to global data in a DNA sequence registry. Homologous DNA sequences from parasites detected in bobolinks and more sedentary birds (e.g., brown‐headed cowbirds Molothrus ater, and other co‐occurring bird species resident on the North American breeding grounds) were compared to those recovered in previous studies from global sites. One parasite lineage that matched between Galapagos birds and the migratory bobolink, Plasmodium lineage B, was the most common lineage detected in the global MalAvi database, matching 49 sequences from unique host/site combinations, 41 of which were of South American origin. We did not detect lineage B in brown‐headed cowbirds. The other Galapagos‐bobolink match, Plasmodium lineage C, was identical to two other sequences from birds sampled in California. We detected a close variant of lineage C in brown‐headed cowbirds. Taken together, this pattern suggests that bobolinks became infected with lineage B on the South American end of their migratory range, and with lineage C on the North American breeding grounds. Overall, we detected more parasite lineages in bobolinks than in cowbirds. Galapagos Plasmodium had similar host breadth compared to the non‐Galapagos haemosporidian lineages detected in bobolinks, brown‐headed cowbirds, and other grassland species. This study highlights the utility of global haemosporidian data in the context of migratory bird–parasite connectivity. It is possible that migratory bobolinks bring parasites to the Galapagos and that these parasites originate from different biogeographic regions representing both their breeding and overwintering sites.     

Temporal and Contextual Consistency of Leadership in Homing Pigeon Flocks

Organized flight of homing pigeons (Columba livia) was previously shown to rely on simple leadership rules between flock mates, yet the stability of this social structuring over time and across different contexts remains unclear. We quantified the repeatability of leadership-based flock structures within a flight and across multiple flights conducted with the same animals. We compared two contexts of flock composition: flocks of birds of the same age and flight experience; and, flocks of birds of different ages and flight experience. All flocks displayed consistent leadership-based structures over time, showing that individuals have stable roles in the navigational decisions of the flock. However, flocks of balanced age and flight experience exhibited reduced leadership stability, indicating that these factors promote flock structuring. Our study empirically demonstrates that leadership and followership are consistent behaviours in homing pigeon flocks, but such consistency is affected by the heterogeneity of individual flight experiences and/or age. Similar evidence from other species suggests leadership as an important mechanism for coordinated motion in small groups of animals with strong social bonds.     

Phylogeny, duplication, and intraspecific variation of Adh sequences in New World diploid cottons (Gossypium l., malvaceae)

The 13 "D-genome"cotton species are a monophyletic assemblage of morphologically diverse diploids that inhabit arid to semiarid regions in Mexico, with 1 disjunct species each in Peru and the Galapagos Islands and 1 species whose range extends northward into Arizona. While these species lack commercially significant fiber (i. e., cotton), they are important in that they represent one of the parental genomes of the cultivated tetraploid cottons. To assess phylogenetic relationships among these species, we sequenced and analyzed a region of a nuclear-encoded alcohol dehydrogenase gene (AdhA). Phylogenetic analysis resulted in a topology that is generally consistent with current taxonomic alignment of the species, although the phylogeny based on AdhA sequences conflicts with those inferred from cpDNA and ITS data sets, most notably in the position of the anomalous species Gossypium gossypioides. In one lineage, we detected both gene duplication and sequence polymorphisms that transcend species boundaries; sequences in this lineage formed a monophyletic clade, yet no taxon within the clade contained a monophyletic collection of sequences. Potential explanations for this latter phenomenon, including gene duplication, gene flow, and lineage sorting, are discussed.     

Phylogenetic reconstruction of parental-care systems in the ancestors of birds

Due to the controversy surrounding incipient avian parental care, ancestral parental care systems were reconstructed in a phylogeny including major extant amniote lineages. Using two different resolutions for the basal avian branches, transitions between the states no care, female care, biparental care and male care were inferred for the most basal branches of the tree. Uniparental female care was inferred for the lineage to birds and crocodiles. Using a phylogeny where ratites and tinamous branch off early and an ordered character-state assumption, a transition to biparental care was inferred for the ancestor of birds. This ancestor could be any organism along the lineage leading from the crocodile-bird split up to modern birds, not necessarily the original bird. We discuss the support for alternative avian phylogenies and the homology in parental care between crocodiles and birds. We suggest that the phylogenetic pattern should be used as a starting point for a more detailed analysis of parental care systems in birds and their relatives.     

Ornithophilous canopy species in the Atlantic rain forest of southeastern Brazil

ABSTRACT.   In tropical ecosystems, birds play a relevant role in plant reproduction. Although hummingbirds are regarded as the most important vertebrate pollinators in the Neotropics, the possible role of perching birds as pollinators has been neglected. From 2003 to 2005, we observed 68 species of plants visited by birds in an Atlantic rainforest in southeastern Brazil, including three canopy species: Spirotheca rivieri (Malvaceae, Bombacoidea), Schwartzia brasiliensis (Marcgraviaceae), and Psittacanthus dichrous (Loranthaceae). Flowers of these three species were visited by 15 different species of perching birds and by hummingbirds. The flowers of these three plants are colorful, ranging from deep red or purple to orange. Spirotheca rivieri blooms during the austral winter and Schwartzia brasiliensis during the summer. The flowers of these two species produce copious amounts of dilute nectar in easily accessible structures and both species appear to depend primarily on perching birds as pollinators, with hummingbirds being secondary or minor pollen vectors. The tubular, narrow flowers of P. dichrous are produced during the austral summer and are visited primarily by hummingbirds. Perching birds also visit the flowers, but destroy them. Our results suggest that previous estimates of the number of perching birds that feed on nectar may be too low and that flowers pollinated by perching birds may be more common in the canopy of Neotropical forests than previously thought.