The Condor 101:299-310 0 The Cooper Ormthological Society 1999 TITS, WARBLERS, AND FINCHES: FOLIAGE-GLEANING BIRDS OF NEARCTIC AND PALEARCTIC BOREAL FORESTS? RUSSELL GREENBERG Smithsonian Migratory Center, National Zoological Park, Washington, DC 20008, e-mail: antbird@erols.com VLADIMIR PRAVOSUDOV~ Department of Zoology, Ohio State University, Columbus, OH 43210 JOHN STERLING Smithsonian Migratory Bird Center, National Zoological Park, Washington, DC 20008 ANNA KOZLENKO 4075 Monticello Blvd., Apt. 305, Cleveland Heights, OH 44121 VITALLY KONTORSHCHIKOV State Darwin 57, Vasilova, Moscow, Russia 117292 Abstract. We describe two major patterns of abundance of different types of canopy foliage-gleaning birds at four sites (northwestern Canada, eastern Canada, European Russia, and central Siberia) in boreal forest. First, because of the extreme numerical dominance and breadth of habitat use of the Chaffinch (Fringilla coelebs), the European Russian sites stand as outliers to the relationships between the number of species and the abundance of all foliage-gleaning birds, as well as to more detailed patterns of the composition of the foliage- gleaner guild. The Chaffinch is one of the few canopy foliage-gleaning birds in temperate forests that is abundant in agricultural habitats during the nonbreeding season, which prob- ably allowed a rapid expansion of its range and growth in abundance with the advent of farming in Europe. The second pattern emerges when the European Russian sites are ex- cluded from analyses: the proportional abundance and number of species of warblers or warbler-like birds were positively related to overall abundance of foliage gleaners. This suggests that warblers prosper when there is high resource productivity during the breeding season. A possible resolution between the patterns found in European Russia and the other boreal forest sites may be that Fringilla are functionally equivalent to warblers and have essentially replaced warblers to achieve their high abundance. The patterns of relative abun- dance give support to both the Resource Seasonality Hypothesis (warblers versus tits), and the Ecological Replacement Hypothesis (Fringilla versus warblers) as explanations for shifts in the numerical dominance of different types of foliage gleaners in different regions. Key words: boreal forests, community convergence, community structure, continental comparison, foliage-gleaning bird, Fringilla, warblers. INTRODUCTION The study of similar habitats in different conti- nents is a common approach to testing general- ities in the patterns of avian species assemblag- es. Boreal forests are an attractive focus for these types of studies because they form a con- tinuous band across North America and Eurasia of low diversity forests dominated by a few shared genera of canopy trees. Although one-to- one correspondence between forest types may be elusive, the range of habitats available to birds ? Received 24 February 1998. Accepted 6 January 1999. 2 Present address: Department of Biological Scienc- es, Purdue University, West Lafayette, IN 47907-1392. is quite comparable. Many of the dominant high- er taxa of songbirds, however, are only distantly related between the two continents (Haila and Jtivinen 1990). Variation in the taxonomic com- position of forest avifaunas has fueled the search for ecological equivalents between regions, as well as for biogeographic and ecological pro- cesses that have determined the composition of different assemblages (Morse 1989). The canopy foliage-gleaner guild is the most abundant and species-rich guild across all boreal forests. Various studies have compared the ecol- ogy of specific paired groups of foliage gleaners, particularly the Old World versus New World warblers (Cody 1974, Winkler and Leisler 1994). The problem with this approach is that [2991 300 RUSSELL GREENBERG ET AL. the flush of foliage insects in temperate forests often supports the breeding of a morphologically or ecologically diverse group of species. Species that forage on foliage arthropods are often war- bler or warbler-like birds, but also include tits, finches, and flycatchers. The relative impor- tance, both in terms of species richness and abundance, of these different types of foliage gleaners varies considerably between regions. The search for ecological equivalents between sites should, therefore, be expanded beyond spe- cific taxonomic pairings to include all foliage- gleaning species. Two hypotheses have been put forth to ex- plain intercontinental and regional differences in the dominance of different types of foliage gleaners, particularly with respect to the relative importance of tits versus warblers (Sylviinae or Parulinae). The Ecological Replacement Hy- pothesis posits that as a result of historical pat- terns of colonization and diversification, one type of foliage gleaner usurps resources limiting the ecological opportunities for other types. For example, New World assemblages support a large number of species of warbler or warbler- like birds, whereas Old World assemblages (par- ticularly those in southern or western Europe) often have a higher number of tit species (Haila and Jtivenin 1990). This has led some (Morse 1989) to speculate that tits replace wood war- blers to a large degree in Palearctic forests and to suggest that in southern Europe tits replace Old World warblers (Herrera 1978). The second hypothesis is more explicitly eco- logical. The Resource Seasonality Hypothesis posits that certain types of foliage gleaners are specialized on harvesting the bloom of summer arthropods and become more common where the breeding season resource peak is more pro- nounced. For example, Haila et al. (1987), high- lighting the negative correlation between tit and Phylloscopus abundance in European census data, suggested that the latter become more nu- merically dominant in more seasonal habitats. Related to this hypothesis, Wiens (1975) and Ra- benold (1978) proposed that the high abundance and diversity of wood warblers in Northeastern coniferous forests of North America is related to the large pulse of food during the breeding sea- son in these forests compared to Western and Southern forests. This paper is an initial step toward under- standing the relative importance of the major types of foliage gleaners in different boreal for- est habitats in light of these hypotheses and with an emphasis on comparing Palearctic and Ne- arctic sites. Considering survey data gathered from sites in different boreal forest regions, we address the following questions: (1) What is the pattern of variation in relative abundance of dif- ferent foliage-gleaner types in different regions of the boreal forest? (2) What is the relationship between species richness and abundance of fo- liage-gleaning birds and different types of fo- liage-gleaning birds? (3) Is there a predictable relationship between the abundance of different types of foliage-gleaning birds and the overall abundance of foliage-gleaning birds? (4) To what degree do the aforementioned relationships vary between different sites and continents in the boreal forest? METHODS STUDY SITES Fieldwork was conducted in successive sum- mers: central Siberia 1993, northwestern Canada 1994, European Russia 1995, and eastern Can- ada 1996. We selected four study sites organized into two comparable pairs (Fig. 1). The first pair is ?middle taiga? sites with highly continental climates: the southern MacKenzie District of the Northwest Territories (MacKenzie) and the mid- dle section of the Yenisey River between Tun- guska and Angara Rivers (Siberia). Both areas were centered on a major river system (Yenisey and Mackenzie are the fourth and seventh lon- gest rivers in the world, respectively) and had a similar range of latitudes (Siberia, 6O?lO? to 62?15?N; MacKenzie, 60?14? to 61?52?N). Re- search was conducted at three sites each in Northwest Territories and central Siberia. The Siberian sites (Mymoe, Vorogova, and Fomka) were a maximum of 295 km apart and the max- imum distance was 340 km for the MacKenzie sites (Fort Liard, Fort Simpson, and Fort Provi- dence). The second pair of study areas was lo- cated in less continental regions characterized as ?South Taiga.? Both the European Russian (Eu- rope) and eastern Canadian (Ontario) sites were far more disturbed and developed than the cen- tral Siberian and Mackenzie areas. Due to logis- tical constraints, the Russian fieldwork was con- ducted primarily within 40 km of the Kostrom- skaya Taiga Field Station located on the Unja River (58?14?N, 55?25?N). However, 10% of the BOREAL FOREST FOLIAGE-GLEANING BIRDS 301 FIGURE 1. Map depicting location of study sites in Canada and Russia. Sites marked with a square are in northern taiga and those with a circle are in southern taiga. survey points were conducted at the Zvenigorod Research Station 600 km southwest (near Mos- cow) with the addition of no species. The On- tario sites were primarily in the clay belt be- tween the cities of Gogama, Timmins, Cochran, and Hearst, as well as White River and Mara- thon with a maximum intersite distance of 490 km and a latitudinal range of 47?51?N to 4Y30?N. The lower latitudes of locations in On- tario reflect the overall southern distribution of the boreal forest in this region. HABITATS SAMPLED We sampled four types of mature forest habitat in Siberia: (1) riparian alder (Alnaster frutico- sus) and willow (Sulh spp.), (2) mixed flood- plain forest, with spruce-fir (Pica obovatu-Abi- es sibiricu) and birch (Beth spp.), (3) upland birch stands with some aspen (Populus tremulu), and (4) mixed upland taiga with Siberian pine (Pinus sibiricu), spruce, larch (Larix sibiricu), fir, and birch. We surveyed five habitat types in MacKenzie: (1) black spruce forest (Piceu mur- ianu), (2) mixed flood-plain forest, with white spruce (Piceu gluucu) and balsam poplar (Po- pulus bulsumiferu), (3) riparian, with balsam poplar, alder (Ahus tenuifoliu), and willow, (4) trembling aspen stands (Populus tremuloides), and (5) jack pine (Pinus bunksiunu) forests. In the European Russian site we sampled (1) mixed, upland second-growth taiga (approxi- mately 60 years old), with spruce (Piceu ubieu), aspen, linden (Tilliu cordutu), and birch (Beth pubescens), (2) birch-aspen forest, (3) European pine (Pinus sylvestris) forest, and (4) riparian elm-alder (Urnus-Ahus incunu) woods. In On- tario we sampled (1) aspen stands (aspen, birch [Betulu pupyriferu], and balsam poplar), (2) mixed upland taiga (white spruce, balsam fir [Abies bulsumeu], jack pine, birch, aspen, and poplar), (3) black spruce, and (4) jack pine stands. POINT COUNTS Because bird censusing techniques often differ between studies and regions, it is often difficult to make systematic comparisons. For the pur- poses of this analysis, we gathered field data em- ploying the same point-count methodology at all sites. Point counts (Hutto et al. 1986, Petit et al. 1995) consisted of a series of up to 15 5-min counts located 200 m apart. Censuses were con- 302 RUSSELL GREENBERG ET AL. ducted 30 min after sunrise for approximately 3 hr when wind speeds are less than 12 km hr?. The number and approximate distance of all birds were recorded as well as the mode of de- tection (song, call note, strictly visual). As much as possible, all points on a transect were located in a single habitat type at a minimum of 100 m from a habitat edge, which requires large tracts of habitat for surveying. Point counts were con- ducted from 5-25 June, which corresponds with the peak singing activity of migratory species. Resident species (tits) initiate breeding earlier than migrants and so were probably vocalizing less than migrants during our work. Although this may cause underestimation of tit popula- tions, we regard this as an equivalent bias for all sites. In addition, a comparison of our point count data with published census data from the same region showed a high correlation (r = 0.95) between the two data sets with respect to composition of different types of birds and no significant underestimation of tits (see Results). Because of the severe problems of differential detectability between species and habitats, we have relied upon the mean number of birds with- in 50 m of the point to assess foliage gleaner abundance (Petit et al. 1995). In order to stan- dardize the species list to assess overall species richness, we used the total number of species detected with an abundance greater than 0.02 in- dividuals per point, and standardized effort by randomly subsampling the data from 75 points. The summary of individual species by site is available at the Smithsonian Migratory Bird Center web site (http:Nwww.si.edu/smbc). In MacKenzie, we sampled mixed white spruce near Ft. Liard and Black Mesa (91 point counts, see below), Ft. Simpson (20), and Ft. Providence (27); black spruce near Ft. Liard (60), Ft. Simpson (32), and Trout Creek (17); riparian deciduous near Ft. Liard (55) and Ft. Simpson (35); aspen near Ft. Liard (43) and Ft. Simpson (32). In Siberia, we sampled upland taiga near Myrnoe (177), Volgova (45), and Fomka (58); floodplain taiga near Myrnoe (32), Volgova (64), and Fomka (57); birch near Vor- ogova (61) and Fomka (48); and riparian near Mymoe (43), Vorogova (48), and Fomka (57). In Ontario, we sampled upland mixed forest near Marathon (40), Hearst (57), Cochrane (18), White Lake (28), and Timmins (5); aspen-poplar near Timmins (91) and White River (9); and black spruce near Gogama (12), Timmins (1 l), Cochran (53), Marathon (6), and Obatanga (5). PUBLISHED CENSUS DATA Point counts provide a standardized rapid as- sessment, which allows direct comparison be- tween sites. We have examined published cen- suses based mainly on territorial mapping or more rarely on repeated fixed transects to assess the generality of some of the qualitative results of the analysis from our point count data. The censuses were selected to cover a broad habitat and geographic range, but are by no means ex- haustive. The data set (available from first au- thor) includes 21 censuses from eastern North America, 6 from western Canada, 28 from northern Europe (West of Urals), and 20 from Siberia. We calculated the mean value (or in the case of Breeding Bird Censuses from Canada, we used published means following Erskine [1977]) when more than one survey was con- ducted in the same habitat and site. CLASSIFICATION OF FOLIAGE-GLEANER TYPE We subdivided the canopy foliage-gleaning guild into four types, corresponding to differ- ences in morphology or, in the case of flycatch- ers, large differences in foraging behavior: (1) tits (species of Paridae or Aegithalos), (2) finch- es (Fringilla, Emberizids, Carduelinae), (3) war- blers (Parulinae, Vireonidae, and the probably polyphyletic [Sheldon and Gill 19961 Sylvinae including Regulus and Slyvia), and (4) flycatch- ers (Empidonax, Ficedula, Erithacus, Phoeni- curus). The last group includes species that for- age predominantly from leaves, twigs, and branches with a mean attack distance > 1 m, but not those such as Muscicapa and Contopus spp. which forage for aerial arthropods. Two uncom- mon species (Gray Jay Perisoreus canadensis, and Golden Oriole Oriolus oriolus) that were not easily classified were ignored in the analyses. We have included some species that are often not considered foliage-gleaning species. Based on our foraging observations at our study sites, these species probably do a substantial amount of their foraging off foliage or twigs. The non- traditional species include some that also forage on the ground (European Robin Erithacus ru- beculu, Dark-eyed Junco Junco hyemalis, Chip- ping Sparrow Spizella passerina, and Yellow- breasted Bunting Emberiza aureola). These spe- cies were commonly found in the canopy. How- BOREAL FOREST FOLIAGE-GLEANING BIRDS 303 ever, with ground-foraging birds it is often difficult to determine how much of their forag- ing time is spent on the ground, because there they are more difficult to observe. We also in- clude some Cardueline finches, which are well known to feed on seeds extensively, even when rearing young (Newton 1967). In the forest hab- itats we studied, siskins and bullfinches forage commonly for foliage arthropods during the height of the breeding season. These species were generally uncommon on our surveys and exclusion of these species does not greatly alter the major conclusions of this paper. RESULTS NUMERICAL DOMINANCE, ABUNDANCE, AND COMPOSITION OF FOLIAGE-GLEANING GUILDS We found significantly more foliage-gleaning in- dividuals per point in European Russia than oth- er sites, largely a result of the very high abun- dance of Chaffinches in all habitats (ANOVA based on mean per habitat, overall habitat effect F 3,,1 = 28.2, P = 0.006, Tukey?s post-hoc com- parison was significant only for Europe versus other sites). The upland Siberian sites had low densities compared to similar North American sites (Table 1). Jack pine forest on sandy glacial outwash is equivalent to European pine forests of Siberia, which were not surveyed and so were not considered in this comparison. The dominant species varied among habitats in the two North American and Siberian sites (Table l), with war- blers generally dominating the former and tits the latter. However, the Chaffinch was the dom- inant foliage-gleaner in all four habitats at the European Russian site. The average proportion of Chaffinches across habitats in European Rus- sia was 0.44. The highest values for the other sites were Willow Tit (Poecile montanus) in Si- beria (0.17), Tennessee Warbler (Vermivora per- egrina) in MacKenzie (0.19), and Yellow-rum- ped Warbler (Dendroica coronata) in Ontario (0.11). In terms of numbers of individuals, the rela- tive importance of different types of foliage gleaners varied among the different study sites. The proportion of finches varied significantly (Kruskal-Wallis H = 11.0, P = 0.01) with the highest proportion (56%, Table 1) in European Russia, also a consequence of the high Chaffinch abundance. Siberia had a moderate (26%), and the New World sites low (15% and 9%), pro- portions of finches. The New World sites had a significantly greater proportion of warblers (72%; H = 12.3, P = 0.006). We found no over- all significant difference in the proportion of tits and flycatchers across sites. However, except for the riparian vegetation, with its lack of cavities, central Siberia had a strikingly high proportion of tits (33%) compared to other regions. SPECIES RICHNESS VERSUS ABUNDANCE OF FOLIAGE GLEANERS AND FOLIAGE-GLEANER TYPES The correlation between species richness and abundance of foliage gleaners across habitats was significant and strong (r = 0.85, P < 0.001) with European Russia excluded, and somewhat weaker, but still significant, with the Russian habitats included (r = 0.73, P < 0.001, Fig. 2); the Russian sites are brought into conformity with other sites by removing the Chaffinch num- bers (9 = 0.83, P < 0.001). The relationship between species number and abundance across foliage-gleaner type (warbler, finch, etc.) also was strong for the pooled North American and Siberian sites (r = 0.96, P < O.OOl), but considerably weaker if the European Russian sites are included (r = 0.72, P < 0.001). Species richness and abundance of type were significantly correlated when considered sepa- rately for Siberia (r = 0.91), MacKenzie (0.84) and Ontario (0.97) (all P values < O.Ol), but not European Russia (0.54). ABUNDANCE OF FOLIAGE-GLEANER TYPES VERSUS OVERALL FOLIAGE-GLEANER ABUNDANCE It is likely that the variation in overall abun- dance of foliage-gleaning birds reflects differ- ences in food resources of different habitats. Based on this assumption, we have plotted the relative abundance of different foliage-gleaner types against total foliage-gleaning bird abun- dance (birds per point). We conducted ANCO- VA using site and the mean total number of fo- liage gleaners per point as independent variables and the mean number of individuals of different types as the dependent variable. The analysis was first run with the site by abundance inter- action term. When there was no significant in- teraction, the ANCOVA was then run without the interaction term to test for the effect of site and abundance (Wilkinson 1990). Because of the anomalously high abundance of Chaffinches BOREAL FOREST FOLIAGE-GLEANING BIRDS 305 . I . . . l Europe 0 Siberia 1 0 MacKenzie A Ontario I 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Mean foliage-gleaning birds per point per habitat FIGURE 2. The mean number of foliage gleaners per type per habitat versus the number of species per type of foliage-gleaners in 75 points in each habitat with abundance > 0.02 per point. at the European Russian site, we analyzed the data with all sites and with the European Rus- sian data excluded. When we eliminated the Eu- ropean Russian habitats from the analysis, with their overriding dominance of Chaffinches, we found a strong relationship between the abun- dance of warbler-like birds and foliage-gleaner abundance (Fig. 3A). The ANCOVA (overall r = 0.96) showed a large effect of bird abundance (F, 9 = 336.8, P < 0.001) and a smaller effect of site (F2,9 = 17.6, P < 0.001). When European Russia was included, the correlation between warbler abundance and overall foliage-gleaner abundance was considerably weaker, but still significant (Y = 0.69, P = 0.001). We found a weak or nonexistent relationship between the abundance of other types and over- all foliage-gleaner abundance (Fig. 3A). In terms of relative abundance (the proportion of foliage- gleaners of a given type) and with European Russia excluded, warblers showed a significant increase with increasing foliage-gleaner abun- dance (r = 0.81, P = 0.001) and tits showed a weaker, but significant, negative relationship (r = -0.58, P < 0.01). With European Russia in- cluded, only finches showed a significant cor- relation with abundance (r = 0.64, P = 0.006). When the relationship between warbler abun- dance and overall bird abundance (Fig. 3A) for all sites is examined, the high abundance of Chaffinches and the lower than expected abun- dance of warblers at the European sites suggest a possible functional relationship. Summing the abundance of Fringilla and warblers produces an extremely tight relationship between their 1 2 3 4 ? 14 . 12 - 1 2 3 4 Mean number of foliage-gleaning birds per point FIGURE 3. A. The mean number of total foliage gleaners per point versus the mean number of foliage gleaning birds per point for each habitat surveyed. Note that the four points at the far right of the graph for each type are European showing the reversal of the position of finches and warblers from the relationship of other sites. B. The total number of foliage gleaner species per type per habitat versus the mean number of foliage-gleaning birds per habitat. abundance and overall bird abundance (r = 0.99) similar to that found for warblers in the three non-European communities alone. The ANCOVA produced a strong effect of overall foliage-gleaner abundance (F,,,2 = 394.3, P < 0.001) and a weaker effect of site (F3,,2 = 15.5, P < 0.01). The number of species of warblers also was highly correlated with the abundance of foliage- gleaning birds (r2 = 0.85, P < 0.001) when the European site was excluded (Fig. 3B) and not significantly correlated when these data were in- cluded (Fig. 4, r = 0.39, P = 0.16). An AN- COVA on warbler species number (European Russia excluded) produced no significant inter- action between site and overall abundance. The 306 RUSSELL GREENBERG ET AL. 1.0 x09 * l Europe 0 Siberia go6 A 1 l Ontario 0 MacKenzie $ A 0 I I 0.05 0.15 0.25 0.35 0.45 0.55 0.65 1.0 . I 0.9 1 B A 0 ; 0.6 MacKenzie a lo 1, 0 Siberia Europe A Ontario I I 0.05 0.15 0.25 0.35 0.45 0.55 0.65 Mean proportion of birds per point FIGURE 4. A. Proportion of finches versus warblers for each habitat surveyed with point counts. B. Pro- portion of tits versus warblers for each habitat sur- veyed with point counts. ANCOVA without interactions produced a sig- nificant model (Y = 0.89, P < 0.001) with only foliage-gleaner abundance entering as a signifi- cant variable (F,,g = 20.8, P < 0.01). The di- versity of other types was not significantly re- lated to bird abundance or site. COMPLEMENTARITY OF ABUNDANCE OF DIFFERENT FOLIAGE-GLEANER TYPES A strong negative correlation (r = -0.79, P = 0.001, Fig. 4A) was found between the propor- tion of finches and warbler-like birds across dif- ferent habitats. New World communities are mostly represented at the high warbler/low finch portion of the graph, and European communities on the opposite end. Siberian communities show the greatest scatter because these habitats sup- port the greatest number of tits, which is not reflected in this relationship. On the other hand, the correlation between the proportion of war- bler-like birds and tits was suggestive, but not significant (Y = -0.48, P = 0.08, Fig. 4B). The suggestion of a relationship occurs primarily be- f 10 30 50 70 90 Mean percentage of individuals per type published FIGURE 5. Mean percentage of individuals in dif- ferent foliage-gleaner types for census plots from a re- gion (e.g., eastern North America, northern Europe, etc.) versus the mean percentage of individuals in dif- ferent foliage-gleaner types for point counts from the same region (this study). cause of the three Siberian habitats with high tit and low warbler abundance. When European Russian sites were excluded, the dominant neg- ative correlation was between warblers and tits (r = -0.73, P = 0.005) with a somewhat weak- er, yet significant negative correlation between warblers and finches (r = -0.61, P = 0.03). In the case of the analyses with and without Eu- ropean Russia, we found no significant relation- ship between tits and finches (Y = 0.10 and 0.28, respectively), nor did we find any significant re- lationships between the proportion of flycatchers and any other type. PATTERNS IN PUBLISHED CENSUS DATA Based on the comparison with the census plots or transects, the patterns of the point count data are robust (Fig. S), particularly considering that the geographic and ecological range of the pub- lished data were greater than for our study. The proportion of finches was highest (41.2%) in northern European and western Canadian (39.3%) communities and lowest in eastern North America (11 .O%). Warblers were propor- tionately most common in eastern North Amer- ica (75.7%) and western Canada (43.5%) and comparably low in the Palearctic regions (36%). Tits were relatively common in Siberia (33.8%) and uncommon in northern Europe and North America (4.1-10%). Flycatchers were relatively common in Europe (14%) and rare elsewhere (0.8-6.9%). The overall patterns of foliage- gleaning guild composition were surprisingly similar when plot and point count data were BOREAL FOREST FOLIAGE-GLEANING BIRDS 307 compared. The correlation between mean pro- portion of type in a point count region and plots from the comparable region (e.g., eastern North America vs. Ontario; northern Europe vs. Kos- troma, etc., with only spruce and pine habitats included for western Canada) had an r of 0.95 with a slope of 1.10 (Fig. 5). As in the boreal point-count data set, the pro- portion of warblers was related to the proportion of both finches (r = -0.61, P < 0.001) and tits (r = -0.58, P < O.OOl), values which are more similar to each other than those for the point count data (-0.79 and -0.48, respectively). When only Siberia and North America data were included, we found the major negative correla- tion between tits and warblers (r = -0.74, P < O.OOl), with a weaker relationship between finches and warblers (r = -0.57, P < 0.001). As in the point count data, the relationship be- tween tit and finch abundance was weak (r = 0.11 and 0.10 for analysis with and without northern Europe, respectively). None of the cor- relations between proportion of flycatchers and other types were significant. Chaffinch was the most common foliage- gleaning species in 19/25 of the northern Euro- pean communities, occurring commonly in broad-leafed and coniferous habitats alike, ac- counting for an average of 30% in northern Eu- rope (range 2-57%). Chaffinches were scarce only in montane sites in Norway, and far north sites (northern Finland and Pechora region of Russia). Probably the most consistently domi- nant New World species is Red-eyed Vireo (Vir- eo olivaceus). This species was the dominant species in only 5 of 20 censuses in eastern North America, comprising 13.6% of the foliage- gleaners in those habitats, and was almost absent from coniferous forests. DISCUSSION Given the clear overdominance of Chaffinches in most European communities, we separate dis- cussion of the distribution of foliage gleaning types into three issues: (1) what were the pat- terns of abundance of foliage-gleaner types in the non-European communities? (2) what ex- plains the numerical importance of Chaffinches in European forest bird communities? and (3) can the discrepancies between the European habitats and other regions be explained by the hypothesis that Fringilla functionally replace warblers? PATTERNS ACROSS NON-EUROPEAN SITES When the European site was excluded, the abun- dance and number of species of warblers showed a highly significant relationship with overall foliage gleaner abundance with only a small effect of site, whereas the abundance of finches and tits was unrelated to overall abun- dance of foliage gleaners. This observation sup- ports the Resource Seasonality Hypothesis for the importance of warblers in different boreal forest habitats. According to this hypothesis, the high proportion of tits in many Siberia habitats is primarily a result of the reduced summer re- sources and reduced warbler abundance. Tit populations are generally higher in Siberia, but the absolute number (overall, 0.26 per point) comprises a small portion of breeding foliage- gleaner assemblages (Fig. 4), and variation in the proportion of tits and warblers is largely de- termined by a large increase in abundance of warblers with increasing foliage-gleaner density. In contrast, the high proportion of tits in south- ern Europe (Herrera 1978, Haila and Jkvinen 1990) is associated with high absolute abun- dances and may be driven more by mild winter climates and greater winter food resources. We stress the Resource Seasonality Hypothe- sis for explaining variation in the proportion of warblers and tits in boreal forest. However, the consistently higher abundance of tits in Palearc- tic habitats is related to their higher regional spe- cies richness and may have a historical expla- nation. Still, tit abundance in Siberia is high even where only one common species of tit (Willow Tit) was found, which suggests that an ecological explanation is required as well. OVERDOMINANCE OF CHAFFINCHES IN EUROPEAN ASSEMBLAGES Many of the patterns discerned in the North American and Siberian data disappeared or were actually reversed when the European data were included. For example, the relationship between foraging-type abundance and species richness diverges strongly from those of the other sites. The proportion of warblers actually declines with foliage-gleaner abundance when the Euro- pean points were included in the analysis. All of these deviations can be attributed to the strong over-dominance of Chaffinches in all four hab- itats surveyed in European Russia. Except where it is replaced by the Brambling (Fringilla montifringilla) in the far north, the 308 RUSSELL GREENBERG ET AL. Chaffinch is a dominant foliage gleaning bird in all European forests. It comprises from 15-45% of the individual foliage gleaners, with the high- est proportions found in northern European for- ests-both coniferous and broad-leaved. No North American bird approaches this level of abundance in such a wide geographic range or variety of forest types. Within regions as diverse as southern France (Blonde1 et al. 1990) and northwest Russia (Morozov 1992), the Chaffinch has the broadest habitat amplitude of all forest birds. Based on MtDNA evidence, Chaffinches probably colonized Europe in the late Pleisto- cene, as recently as 50,000-95,000 years ago (Marshall and Baker, in press). The species con- tinues to spread into Siberia (Rogacheva 1992). Given the current abundance of this species in European farmlands during the nonbreeding sea- son (O?Conner and Shmbb 1986, Cramp and Perrins 1994), its spread into the largely forested areas of northern Europe was probably acceler- ated by the spread of agriculture. Such a recent spread into European farmlands has been sug- gested for the Dunnock (Prunella modularis; Murton 1971) and House Sparrow (Passer do- mesticus; Johnston and Klitz 1977). Thus far, we have presented a close relation- ship between both warbler abundance and rich- ness and overall foliage-gleaner abundance in non-European boreal forest, along with an anomalous pattern in the European habitats. These divergent observations can perhaps be re- solved if we consider that Fringilla are func- tionally similar to warblers in their resource use and that the expansion of Chaffinch in Europe has occurred at the expense of warbler abun- dance which would be consistent with the Eco- logical Replacement Hypothesis. This idea has some support: when we clustered foliage-glean- ing species based on foraging behavior and site, Chaffinches clustered tightly with certain Phyl- loscopus warblers, vireos, and Parulinae (Green- berg et al., unpubl. data). With this in mind, it is striking how close the relationship was be- tween warbler abundance and biomass and over- all foliage-gleaner abundance across all study sites if Fringilla are pooled with warblers (Fig. 3A). HISTORICAL DYNAMICS OF BOREAL FOREST ASSEMBLAGES A high abundance of Fringilla and a low abun- dance and diversity of warblers characterize Eu- ropean boreal forests. To a large degree, this pat- tern can be explained by historical factors. The center of origin of Phylloscopus, the most forest- adapted of the Old World warblers, was proba- bly in Asia. In European forests, Phylloscopus are represented by two clades that probably in- vaded in the Tertiary and have shown little eco- logical or morphological divergence (Richman 1996). It is likely that through selection or ex- tinction, the high loss of continuous forest cover during the Pleistocene and since human habita- tion has resulted in a depauperate forest-adapted warbler fauna (Monkkonen and Welsh 1994). For whatever reason, the presence of a depau- perate assemblage of migratory forest insecti- vores may have set the stage for the more re- cently expanding populations of Chaffinch to find suitable resources in European forests. POSSIBLE INTERACTIONS BETWEEN WARBLERS AND CHAFFINCHES If we are correct in our hypothesis that the Chaf- finch was a relatively recent invader of Euro- pean forests, then it suggests an interesting ques- tion: How did European forests support such a large increase in the density of foliage-gleaning birds? Two possibilities are that (1) assemblages were not close to a food-based carrying capacity and that increases in winter resources supplied by agriculture is a sufficient explanation for the expansion of Chaffinches, and/or (2) at least some of the increase occurred at the expense of the breeding population densities of warblers and other foliage-gleaning birds. Although competition is no longer generally accepted as important in structuring boreal forest communities (Monkkonen et al. 1997), the pos- sibility that competitive interactions may medi- ate the distribution and abundance of Fringilla and warblers should at least be considered. Frin- gilla tend to arrive earlier than warblers, being able to withstand periods of cold weather by shifting to foraging for seeds (Preobrazhen- skaya, pers. comm.), and hence may be able to dominate sites through prior access. In addition to prior residency, Fringilla are considerably larger than warblers (20-22 g vs. 7-11 g), which also would contribute to their behavioral domi- nance. Hogstad (1975) presented evidence for the competitive interaction of another Fringilla, the Brambling, and Willow Warblers (Phyllos- copus trochilus). Interestingly, he found that the Willow Warbler was dominant to the Brambling. BOREAL FOREST FOLIAGE-GLEANING BIRDS 309 However, other studies have not found evidence of such an interaction (Angell-Jacobsen 1980, Fonstad 1984). Given the biogeographic patterns described here, a close examination of Chaf- finch-warbler interactions seems warranted. CONCLUSIONS Abundance of foliage-gleaning birds is largely determined by abundance of migratory species of warblers or finches (predominantly Fringillu), and these groups increase in their importance with increasing abundance as well. Our results suggest that the warblers and finches are re- sponding to the relative amplitude of summer productivity and resource availability. Tits are generally more diverse and abundant in Palearc- tic communities, suggesting a role for a histori- cally longer association of this dominant group of resident insectivores in Old World forests. However, the variation in absolute abundance is small and contributes little to the decrease in proportion of migrant foliage-gleaning birds be- tween sites. Furthermore, the abundance of tits was highest in Palearctic communities with only one common tit (upland taiga in Siberia), which suggests that an ecological explanation may be required to explain high tit abundance in Old World sites as well. The northern boreal forests have developed recently (Haila and J%rvinen 1990), and the con- stantly changing mixes of trees provide highly variable habitat through the Pleistocene. How- ever, for the most part, recent phylogenetic an- alyses have shown considerable depth to the re- lationships in such groups as Dendroica (Ber- mingham et al. 1992), Phylloscopus and Regulus (Richman 1996), and Parus (Gill et al 1993). This suggests that current species or closely re- lated ancestral species have interacted in tem- perate forest since at least the late Tertiary. In this context, the apparent recent invasion of tem- perate and boreal forests by the Chaffinch stands out. The factors that allowed a less specialized insectivorous bird to achieve numerical abun- dance are unclear. Perhaps the high productivity of European forests coupled with a depauperate fauna of more specialized insectivores caused by high levels of forest loss during the Pleistocene (M6nkkiinen 1994, MBnkkiinen and Welsh 1994) set the stage for this invasion. ACKNOWLEDGMENTS We thank the Institute of Evolutionary Morphology and Animal Ecology of the Russian Academy of Sci- ences, in particular, Katya Preobzazhenskaya and Oleg Bursky, for providing housing and logistical support for work in Myrnoe and Kostromomskaya Field Sta- tions. Anwar Kerimov hosted us at the Zvenigorod field station. The following people assisted in gather- ing data for this project: Mark Amberle, Sam Droege, Todd Easterla, Jennifer Gammon, Gjon Hazard, Ri- chard Hoyer, and Olga Romanenko. Y. Ravkin provid- ed census data from the data base of the Institute of Animal Ecology, Russian Academy of Science, No- vosibirsk. Sam Droege and YrjG Haila provided useful comments on earlier drafts. A Pew Fellowship for Bio- diversity and the Environment provided funding to the senior author. LITERATURE CITED ANGELL-JACOBSEN, B. 1980. Overlap in feeding pattern between Willow Warbler Phylloscopus trochilus and Brambling Fringillu montifringilla in two for- est habitats in western Norway. Ornis Stand. 11: 146-154. BERMINGHAM, E., S. ROHWEK, S. FREEMAN, AND C. WOOD. 1992. Vicariance biogeography in the Pleistocene and speciation in North American wood warblers: a icst of Mengel?s model. Proc. Natl. Acad. Sci. 89:6624-6628. BLONDEL, J. 1990. Biogeography and history of forest bird faunas in the Mediterranean Zone, p. 95-109. In A. Keast [ed.], Biogeography and ecology of forest bird communities. SPB Academic Publish- ing, The Hague, Netherlands. CRAMP. S.. AND C. M. PERRINS. 1994. Handbook of birds of Europe, the Middle East and North Af- rica. Vol. VIII: Crows to finches. Oxford Univ. Press, Oxford. ERSKINE, A. J. 1977. Birds in boreal Canada. Can. Wildl. Serv. Rep. Ser. 41. FONSTAD, T 1984. Reduced territorial overlap between the Willow Warbler Phylloscopus trochilus and the Brambling Fringilla montifringillu in heath forests: competition of different habitat preferenc- es. Oikos 42:3 14-322. GILL, E B., A. MOSTROM, AND A. L. MACK. 1993. Spe- ciation in North American chickadees. I. Patterns of MtDNA genetic divergence. Evolution 47: 195- 212. HAILA, Y., AND 0. JARVINEN. 1990. Northern conifer forests and their bird species assemblages, p, 61- 97. In A. Keast [ed.], Biogeography and ecology of forest bird communities. SPB Academic Pub- lishing, The Hague, Netherlands. HAILA, Y., 0. J;~RVINEN, AND S. RAIVIO. 1987. Quan- titative versus qualitative distribution patterns of birds in the western Palearctic taiga. Annales Zoologici Fennici 24: 179-195. HERRERA, C. M. 1978. On the breeding distribution pattern of European migrant birds: MacArthur?s theme reexamined. Auk 95:496-509. HOGSTAD, 0. 1975. Interspecific relations between Willow Warblers (Phylloscopus trochilus) and Brambling (Fringilla ~ontifr~rzgillu) in subalpine forests. Norwegian J. Zool. 23:223-234. HUTTO, R. L., S. I%. PLETSCHET, AND P HENDRICKS. 1986. A fixed-radius point count method for non- 310 RUSSELL GREENBERG ET AL. breeding and breeding season use. Auk 103:593- 602. JOHNSTON, R. E, AND W. J. KLITZ. 1977. Variation and evolution in a granivorous bird: the House Spar- row, p. 15-5 1. In S. C. Kendcigh and J. Pinowski [eds.], Granivorous birds in ecosystems. Cam- bridge Univ. Press, Cambridge. MACARTHUR, R. 1959. On the breeding distribution pattern of North American migrant birds. Auk 76: 318-325. MARSHALL, H. D., AND A. J. BAKER. In press. Mito- chondrial DNA control-region sequences indicate a sequential range expansion and low current lev- els of gene flow among European populations of chaffinches (Fringilla coelebs). Mol. Ecol. M~NKK~NEN, M. 1994. Diversity patterns in Palearctic and Nearctic forest bird assemblages. J. Bioge- ography 21:183-195. M~NKK~NEN, M., I? HELLE, G. J. NIEMI, AND K. MONT- GOMERY. 1997. Heterospecific attraction affects community structure and migrant abundances in northern breeding communities. Can. J. Zool. 75: 2077-2083. M~NKKONEN, M., AND D. WELSH. 1994. A biogeo- graphic hypothesis on the effects of human caused landscape changes on the forest bird communities of Europe and North America. Annales Zoologici Fennici 3 I :61-70. MORSE, D. H. 1989. American warblers: an ecological and behavioral perspective. Harvard Univ. Press, Cambridge, MA. MURTON, R. K. 197 1. Man and birds. Collins, London. NEWTON, I. 1967. The adaptive radiation and feeding ecologv of some British finches. Ibis 109:33-108. O?CONNO~~R. J., AND M. SHRURR. 1986. Farming and birds. Cambridge Univ. Press, Cambridge. PETIT, D. R., L. J. PETIT, V. A. SAAB, AND T. E. MARTIN. 1995. Fixed-radius point counts in forests: factors influencing effectiveness and efficiency, p. 49-56. In C. J. Ralph, S. Droege, and J. Sauer [eds.], Monitoring bird population trends by point counts. Gen. Tech. Rep. No. PSW-GTR 149. USDA For. Serv. Albany, CA. RABENOLD, K. N. 1978. Foraging strategies, diversity, and seasonality in bird communities of Appala- chian spruce-fir forests. Ecol. Monogr. 48:397- 424. RICHMAN, A. D. 1996. Ecological diversification and communitv structure in the Old World leaf war- blers (genus Phylloscopus): a phylogenetic per- socctive. Evolution 50:2461-2470. ROGA&EVA, H. 1992. The birds of central Siberia. Hussum Druck- und Verlagsggesellschaft, Husum, Germany. SHELDON, E H., AND I? B. GILL. 1996. A reconsidera- tion of songbird phylogeny, with emphasis on the evolution of titmice and their Silvioid relatives. Syst. Biol. 45:473-495. WIENS, J. A. 1975. Avian communities, energetics, and functions in coniferous forest habitats, p. 226- 265. In Proceedings of the symposium on man- agement of forest and range habitats for nongame birds. Gcn. Tech. Rep. WO-1. U.S. For. Serv., Washington, DC. WILKINSON, L. 1990. The system for statistics. Systat Inc., Evanston, IL.