Diffuse competition can be reversed: a case ... - Wiley Online Library

Diffuse competition can be reversed: a case history with birds in Hawaii LEONARD A. FREED1, 2

AND

REBECCA L. CANN2

1 Department of Biology, University of Hawaii at Manoa, Honolulu, Hawaii 96822 USA Department of Cell and Molecular Biology, University of Hawaii at Manoa, Honolulu, Hawaii 96822 USA

Citation: Freed, L. A., and R. L. Cann. 2014. Diffuse competition can be reversed: a case history with birds in Hawaii. Ecosphere 5(11):147. http://dx.doi.org/10.1890/ES14-00289.1

Abstract. Diffuse competition, based on niche overlap of an invader with most native species in a community, is considered a mechanism preventing its successful establishment in the community. However, native species must themselves have extensive niche overlap, implying that diffuse competition can be reversed if some external force increases numbers of the introduced species while native community numbers remain unchanged. Reverse diffuse competition, defined as an invader that outcompetes most of the native community, can then lower numbers of native species. We illustrate these principles by documenting diffuse competition between the introduced Japanese White-eye (Zosterops japonicus) and eight species of Hawaiian forest birds at Hakalau Forest National Wildlife Refuge followed by reverse diffuse competition. Competition was based on extensive overlap of multiple foraging substrates with native species. The external force which reversed diffuse competition was a restoration area contiguous with much of the old-growth forest where diffuse competition previously occurred. Propagule pressure from that area increased white-eye numbers in the old-growth forest, without initial change in numbers of native birds, resulting in greater than sixfold increase in adult white-eyes on our study site. This increase resulted in stunted skeletal growth of young birds, and diverse changes in molt of both young and adults of native birds and white-eyes. While native birds declined from stunted growth, white-eyes increased adult survival. Changes in condition of native birds in time and space matched associated changes in white-eye density. Normalizing natural selection of lower mass and shorter bills resulted in loss of almost half of native birds in a 3373 ha area of the refuge, adjacent to the restoration area, and 10% of native birds in a lower 1998 ha area. Competition with female native birds was initially stronger than with males as indicated by more male-biased adult sex ratios in five species. Structure of the forest, small size of arthropods available, greater use of the understory, and lower minimum air temperature resulted in successful white-eye competition with the entire native community. Reverse diffuse competition is a process in community ecology and a way an additional species can be packed into a community. Key words: adult sex ratio change; demographic change; diffuse competition, Hakalau Forest National Wildlife Refuge; Hawaiian honeycreepers (Drepanidinae); introduced species; Island of Hawaii; Japanese White-eye; population declines; propagule pressure; reverse diffuse competition; Zosterops japonicus. Received 11 September 2014; accepted 12 September 2014; published 25 November 2014. Corresponding Editor: D. P. C. Peters. Copyright: Ó 2014 Freed and Cann. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. http://creativecommons.org/licenses/by/3.0/ E-mail: [email protected]

INTRODUCTION

account for the number of species in a community, their relative abundances, and the resistance

A major challenge for community ecology is to v www.esajournals.org

to invasion. This entails identifying the regional 1

November 2014 v Volume 5(11) v Article 147

FREED AND CANN

sources of potential member species (Ricklefs 1987, Holt 1993), and documenting the niche overlap and partitioning of resources of the final community (Hutchinson 1959, MacArthur and Levins 1967, Schoener 1974, Roughgarden 1976). Colonizing species have always existed, but with human transport of exotic species to different locations (Mooney and Hobbs 2000, McKinney and Lockwood 2001, Olden 2006, Blackburn et al. 2009), native communities in those locations are increasingly subjected to invasions, in some cases with disastrous consequences (Simberloff 1981, 1996, Ricciardi et al. 1998, Freed and Cann 2012a). These invaders can profitably be viewed as potential experimental additions to native communities. They provide an opportunity to examine what processes determine community structure (Shea and Chesson 2002, Olden et al. 2004, Sax et al. 2007). Robert MacArthur (1972) proposed diffuse competition (DC) as the means by which an established community could resist invasion by an additional species. He reasoned that if most species present had some niche overlap with the potential invader, and thus competed weakly with it, the sum of this minor competition would have a substantial negative effect on the invader, preventing its increase in numbers. There are differences of opinion surrounding the functional role of biodiversity in ecology (Naeem et al. 1995, Tilman 1999, Naeem 2002). Diffuse competition could be a potential stabilizing force by keeping introduced species from successfully invading established communities or keep them at low numbers. Diffuse competition implies that native species in the community also compete among themselves. Pianka (1974) showed that the extent of niche overlap, and thus DC, increased with community size. Case (1990) simulated DC and showed that tightly interacting native communities of larger size could prevent a potential invader, initially at low density, from increasing in numbers. He modeled tight interaction as moderately strong competition coefficients among the native species, thus implying a high degree of niche overlap. He showed that properties of the core community were more important than those of the potential invader, which was randomly drawn from the same community, but started at low density as if it were an invader. v www.esajournals.org

Niche overlap produced an ‘‘activation threshold’’ that enabled the community to resist invasion. Diffuse competition has been documented in the context of native species competing among themselves in diverse communities (Keddy 2001). Examples include fungi (Holmer and Stenlid 1996, Fryar et al. 2005), plants (Wilson and Keddy 1986, Burger and Louda 1995, Dyer and Rice 1997, McLellan et al. 1997, Callaway and Pennings 2000), ants (Davidson 1980, 1985), lizards (Pianka 1980), birds (Terborgh and Weske 1975, Bock et al. 1992), and mammals (Heske et al. 1994). However, these studies were not designed to show that such competition would prevent establishment or increase of an invasive species. Diffuse competition is given short attention in modern textbooks on community ecology, where stronger examples of competition are emphasized (Morin 2011, Mittelbach 2012). Nevertheless, diverse taxonomic examples of DC suggest that it is widespread in established ecological communities. Potential invaders to native communities are most commonly introduced species. Exotic species frequently have lag times before they increased and these are poorly understood (Crooks and Soule 1999, Crooks 2005). There are ecological issues about what causes the lag and what then enables the increase. Hypotheses range from demographic factors including the Allee effect (Taylor and Hastings 2005, Tobin et al. 2007) to evolutionary changes in the introduced species (Sakai et al. 2001, Kolbe et al. 2004, Hufbauer and Torchin 2007). Diffuse competition has been proposed to account for lag times (Freed and Cann 2012a), but has not been formally studied to do so. Here we document DC as the cause of the lag time of an introduced bird in a community of Hawaiian birds. We also deal with an aspect of DC that has been largely unexplored. If DC is responsible for the lag, we hypothesize that the same DC that can resist an invader at low density can switch to the newcomer negatively affecting most of the native community at higher density. We designate this switch in competitive effects as reverse diffuse competition (RDC). Reverse diffuse competition has its basis in the same niche overlap required for DC. During the initial phase of the increase in density of the invader, the same niche 2


FREED AND CANN Table 1. Conditions that lead from diffuse competition to reverse diffuse competition. Condition

Evidence

1. Niche overlap of invader with native community leads to diffuse competition 2. Increase of numbers of invader leads to more birds per area 3. Evidence of competition of the invader with the native community 4. Invader has greater niche overlap with each native species than those species have with each other

Niche overlap; invader remains at low numbers for many years; density dependent differences in condition No change in native community but change in number of invader Community-wide changes in condition and survival associated with density of invader Statistical comparison of niche overlap, invader maintains its numbers and does not lose fitness as do native species

overlap that formerly established DC now establishes RDC. It logically follows that the invader at higher density will compete with the native community. The higher density of the invader may involve propagule pressure stemming from the dispersal of the invader from an off-site location (Simberloff 2009), some form of ecological release such as reduced predation or pathogen pressure (Keane and Crawley 2002), or even climate shifts that favor the invader (Hellmann et al. 2008). Reverse diffuse competition involves packing an additional species into a community, thus addressing issues in niche-based competition theory (Pianka 1976), specialization and generalization (Futuyma 2001), and relative abundance of species and stability (May 1976). Overall carrying capacity of a habitat may be reduced, since more food resources will be consumed than before by the enlarged community. Reverse diffuse competition may result in lowered food availability to specialized species and limited options available to generalized species. The previous weak competition among the native species may become stronger, with suddenly appearing or renewed indirect competitive effects (Wootton 1994).

before, such as a land use change which results in a larger population size. Propagule pressure from a different area is the documented basis of many successful invasions (Lockwood et al. 2005, Colautti et al. 2006, Simberloff 2009), and can overwhelm ‘‘biotic resistance’’ of a native community (von Holle and Simberloff 2005, Hollebone and Hay 2007). The increase of the invader will make food limiting, if some other factor previously was the source of limitation, or increase the extent of food limitation. This should produce changes in somatic condition of individuals in the community, including the invader. Third, changes in condition result in reduction of fitness of native species and their decline in the enlarged community. The invader may or may not decline depending on propagule pressure or demographic changes. There must be a community-wide change in condition and decline for RDC to be realized. The link between condition and decline is the outcome of competition and is missing in some animal competition studies. Experimental removal of one competitor that results in an increase of the other is usually considered sufficient (Keddy 2001, Morin 2011, Dhondt 2012, Middlebach 2012). Fourth, the invader must have greater niche overlap with each native species than those species have with each other. This is the basis for influencing many native species with increased numbers of a single species. The invader thus must be an extreme generalist and its increase in numbers must be sustained. This results in greater competition between the invader and each native species than the native species have with each other, which is the reversal of DC. A simple habitat may also play a role, because such a habitat must have greater niche overlap among species than a highly complex habitat, and provide fewer opportuni-

Conditions under which RDC may be expected to occur Four conditions must be met sequentially for DC to become reversed (Table 1). First, the potential invader, at low density, must overlap multiple foraging substrates with most or all of the native community. Reverse diffuse competition must be preceded by DC. Second, the invader must increase in numbers before any change in condition or decline of native species. This is usually facilitated by an ecological opportunity which did not exist v www.esajournals.org

3


FREED AND CANN

Fig. 1. Japanese White-eye density during the Hawaii Forest Bird Survey in 1977 (Scott et al. 1986). The H indicates the approximate location of Hakalau Forest National Wildlife Refuge. The southern half of the refuge, with highest density of endangered species, was located in a pocket of low white-eye density. The darker area surrounding the pocket had 401–800 white-eyes/km2. Inside the pocket density varied from 101–200 (stippled areas) to 201–400 (light grey areas) white-eyes/km2.

common bird in the state by the time of the Hawaii Forest Bird Survey conducted in the late 1970s (Scott et al. 1986). Biologists were concerned that it was competing with native birds because it consumes arthropods, nectar, and fruit in the same locations as native birds (Ralph and Noon 1988). Mountainspring and Scott (1985) documented competition between the white-eye and several species of native Hawaiian birds, using negative correlation in densities across study sites, adjusted for habitat quality. However, they did not document RDC, which requires documentation of DC beforehand. Our study occurred at Hakalau Forest National Wildlife Refuge on the Island of Hawaii. During the Hawaii Forest Bird Survey, most of the land that eventually became the refuge existed largely in a pocket of low white-eye density (Fig. 1). This pocket was surrounded by higher white-eye density to the north, east, and south. To the west was pastureland. The survey documented that two endangered Hawaiian

ties for niche shifts to specialized resources. An increase of one competitor is predicted to result in the decrease of the other for two species based on asymmetrical lowering of carrying capacity (Keddy 2001). These conditions for RDC lower carrying capacity of the habitat for the native community, incorporating the same logic. Carrying capacity for the invader may also be lowered as it lowers such capacity for native species, except if the increased number is based primarily on propagule pressure. It is the greater niche overlap of the invader with most native species than native species have with each other that enables the transition from DC to RDC.

Introduction to empirical study Here we use the Japanese White-eye (Zosterops japonicus), an invasive bird in the long-established community of Hawaiian forest birds, to document both DC and RDC. The white-eye was introduced to Hawaii in 1929 to control insects (Caum 1933). It quickly became the most v www.esajournals.org

4


FREED AND CANN

Fig. 2. Map of main portion of Hakalau Forest National Wildlife Refuge. (A) Aerial photograph showing restoration area above open forest area above closed forest area. (B) Outline of the areas. Solid circles indicate study sites with elevations in m represented by four digits. The three in the southern end of the refuge include the long-term study site (1987–2006) at 1900-m elevation and two sites operated during 2004–2006 at 1770-m and 1650-m elevation. The isolated study site closer to the north is at 1700-m elevation used for density dependent condition in 1995. The 1585-m site was used during 2004–2006.

forest birds were much rarer in the northern portion of the refuge located outside of the pocket where the white-eye existed at higher density. We showed that rarity of the two endangered birds outside of the pocket persisted (Freed and Cann 2013a). We here document competition between the white-eye and Hawaiian birds from DC to RCD in real time. Diffuse competition maintained the pocket of low white-eye density for 22 years. Then we reveal the transition between DC and RDC using a banded community of birds that was followed over 20 years. We have previously documented the increase of white-eyes and decline of the community of native birds in the refuge (Freed and Cann 2012a), but not in the context of DC and RDC. Here we combine new analyses with former ones to specifically deal with the four conditions of RDC.

v www.esajournals.org

METHODS System and study sites Hakalau Forest National Wildlife Refuge, formed in 1985, is located on the windward slope of Mauna Kea. The refuge contains an oldgrowth ohia-lehua/koa (Metrosideros polymorpha/ Acacia koa) forest (Freed 2001), where 87.7% of woody plants were ohia-lehua (Freed et al. 2008a), so the forest approximates a monoculture. Several other woody plants are smaller and rarer (Cheirodendron, Styphelia, Myrsine, Ilex, Coprosma, Rubus, and Vaccinium). Hakalau contains six species of Hawaiian honeycreepers (Drepanidinae), two other native passerine birds, and a native hawk (Buteo solitarius). Three of the honeycreeper species are endangered and the hawk is now considered threatened. Created from a former cattle ranch, the refuge consists of three areas (Fig. 2). The 1314-ha restoration area, started in 1989, consisted mainly of introduced pasture grasses. That area was 5


FREED AND CANN

above the old-growth forest with native and introduced birds detected during the survey in 1977 (Scott et al. 1986). Immediately below the pasture is the 3373-ha open forest area with a matrix of introduced grasses and native understory. Below that is a 1998-ha closed forest area. Cattle ranching was restricted in the closed forest area so most of the understory consists of native species. We established study sites, designated by elevation, in both forest areas (Fig. 2). Most sites were in the southern end of the refuge where endangered species had highest abundance (Freed and Cann 2013a). The closed forest area presents many physical contrasts to the open forest area. Less light reaches the understory and there is less lateral canopy on individual trees. In addition, the closed forest is wetter, caused by tradewinds moving up Mauna Kea and unloading rain first and heaviest at mid-elevations (Carlquist 1980, Ziegler 2002). The closed forest area thus has soil that is more saturated with water than the open forest area. Otherwise, it has the same two dominant canopy trees as the open forest. The same avian community existed in both forest areas. Insectivorous Hawaiian honeycreepers (Drepanidinae) include the Hawaii Akepa (Loxops coccineus coccineus), Hawaii Creeper (Oreomystis mana), Akiapolaau (Hemignathus monroi ), and Hawaii Amakihi (Hemignathus virens virens). Nectarivorous honeycreepers include the Iiwi (Vestiaria coccinea) and Apapane (Himatione sanguinea), but ohia-lehua nectar lacks essential amino acids so they must also consume arthropods to acquire protein. The Hawaii Elepaio (Chasiempis sandwichensis sandwichensis) is a monarchine flycatcher. The Omao (Myadestes obscurus) is a native thrush that is omnivorous, consuming fruit as well as gleaning arthropods on foliage. The Japanese White-eye co-occurred with these native species in both forest areas when first surveyed in 1977 (Scott et al. 1986). Ohia-lehua leaves are small, ranging in length from 1 to 8 cm (Wagner et al. 1990), but those of large trees at Hakalau are 2–4 cm in length, observed during collection of canopy samples for study of arthropods. Arthropods are characteristically small with an average of 0.4 individuals/g of dry foliage (Freed et al. 2007). Spiders are numerically abundant in samples and have highest biomass, especially in September when v www.esajournals.org

most spiderlings appear (Fretz 2000). Caterpillars are second in abundance and biomass (Fretz 2000). Their small size is based on extensive radiation of microlepidoptera in Hawaii (Zimmerman 1978). Arthropods show strong seasonality, with peak abundance from October to January, followed by significant decline extending to July (Freed et al. 2007). Most bird species breed during the decline. Molt of adults formerly began in June and most ended in October (Freed and Cann 2012b). Endangered passerines with extended fledgling periods feed fledglings while molting (Hart and Freed 2003, Freed et al. 2007).

Statistical analysis Most statistical analyses were performed using S-plus version 8.2. The niche overlap calculations and the percentage similarity comparison of community structure before and after the decline used programs in Ecological Methodology 6.1. Splus and R code for estimating breakpoints in piecewise regression is documented in Supplementary Information. Survey data for all years, 1977 and 1987–2007, the latter collected annually in March, were obtained from Camp et al. (2009), who used variable circular plot methodology (Buckland et al. 1993).

Diffuse competition With other researchers, before we hypothesized RDC, we matched foraging substrates in the forest for native birds and the white-eye to estimate niche overlap (Freed et al. 2008a). We have never observed agonistic interactions between the white-eye and native birds so competition would be simply exploitative. We then conducted a regression analysis of the initial white-eye density during 1977 and 1987 through 1999 on year, to determine if the bird changed density over the 22-year time period. All native species and the white-eye have biparental care, so we document the distribution of female and male bill lengths together for all species to indicate the extent of morphological similarity, and test the distribution with a linear and an exponential model. This enabled us to compare ratios of bill lengths of native species, who evolved gleaning arthropods on ohia-lehua foliage, with the invasive white-eye. To determine if the white-eye was affected by the same types of food as native birds, we use a 6


FREED AND CANN

food shortage during 1995, which had the lowest arthropod food in ohia-lehua foliage during 1994–1997 (Freed et al. 2007). The reduced food supply was likely caused by a refuge-wide irruption of yellow-jacket wasps (Paravespula pensylvaticus), another introduced species in Hawaii (Gambino et al. 1987). Wilson et al. (2009) show that in tropical Hawaii, the wasps can shift to large perennial colonies from small annual colonies and consume an astonishingly wide range of arthropods. We observed the wasps on ohia-lehua and koa foliage and heard them in flight. We used two sites where birds existed at different density to assess densitydependent changes in condition. We compared the number of individuals of each species at the 1900- and 1700-m sites (Fig. 2) captured in mist-nets in years 1994–1997. Mistnets were operated in an 18 ha area at each site. The product of net-hours and number of nets was approximately equal between sites over the same time period. We thus divided the counts of individual birds of each species by this product, and compared this estimate of density between the sites with a paired t-test. If density dependence in condition occurred during 1995, birds at the denser 1900-m site would be expected to have poorer condition. We used prevalence of extended and early molt, during January through May 1995, an outlier year for non-normal molt at the 1900-m site (Freed and Cann 2012b). Molt was assessed by inspecting wing, tail, and body feathers of captured birds. Prevalence of non-normal molt was compared between the sites using logistic regression with species and site as factors along with the interaction. We also measured furcular fat levels as empty, trace, bottom-filled, partially full, full, and overflowing. Defining low fat as empty through bottom-filled during these months, we compared prevalence of low fat between the sites using a similar logistic regression. In addition, we compared non-normal molt between sites in 1994 with a test of proportions, to determine if there were consistent differences between the sites. We also compared the number of hatch-year white-eyes captured versus same age native birds during 1995 in both sites. In the 1900-m site, we correlated hatch-year birds captured during 1987–1994 to put 1995 in perspective. v www.esajournals.org

Timing and extent of white-eye increase The white-eye increased first in the restoration area, where the population was growing exponentially, then in the open forest area, and finally in the closed forest area (Freed and Cann 2012a). Based on this timing, the proximity of the open forest area to the restoration area, and the decrease of white-eye juvenile survival from 0.27 to 0 in our 1900-m site (Freed et al. 2008a), the increased white-eye numbers in that site had to be from propagules. The closest source of propagules was from the restoration area. The increase of the white-eye in the entire open forest area, including the northern section where the white-eye existed at high density, was stepwise, inferred through a randomization test from years 1987–1999 and 2000–2007 (Freed and Cann 2012a). Here we conduct several randomization tests using years 1998 to 2002 as the hypothesized change in density to determine if the white-eye increased before native birds declined. The first of two or more consecutive years of a significant increase was considered the start of the increase. The white-eye population in the closed forest area began growing several years after the stepwise increase in the open forest area (Freed and Cann 2012a). We then investigated if white-eyes continued to increase in our open forest area study sites by greater captures of white-eyes in relation to mistnet hours operated. For the 1900-m site, we estimated the overall capture rate from 1987 to 1999 and compared rates calculated for years 2000 to 2006 using a t-test with the mean being the 1987–1999 capture rate. We used two additional sites in the open forest area (Fig. 2; 1770 and 1585 m) to determine if the capture rates increased after 2000. We asked if the highest capture rates observed in 2006 in two sites and in 2005 in one site could be expected by chance using a combinatorial test. We used our long-term 1900-m site to determine how many adult white-eyes were present each year from 1987 through 2006. By knowing the number of net-hours operated each year, we could identify years with similar numbers of mist-hours. The years were 1988 and 2002 with approximately 4000 net hours. This enabled us to determine the extent of increase with mist-net hours held constant. We explore the possibility that changes oc7


FREED AND CANN

curred in adult white-eye survival. A KaplanMeier survival analysis was conducted to different years on initially captured adults. Years 1987– 1998 were one cohort (n ¼ 164) and years 2000– 2005 were the second cohort (n ¼ 349). We eliminated birds initially captured during 1999 because survival would only be during the white-eye increase. Because higher survival could be based on greater sampling effort, we compared recaptures during consecutive years for each cohort. If higher recapture was based on net-hours, there should have been a higher probability of being recaptured during consecutive years. We compared recaptures of 33 birds before the increase and 69 birds during the increase for consecutive years of recapture with a test of proportions.

shivering when they were captured before and during the white-eye increase. High elevation Hawaii amakihi have 318C as the lower bound of the thermal neutral zone (MacMillen 1974). Mean air temperature at the refuge weather station at 1941 m elevation is 12.38C (Freed and Cann 2012b), implying that substantial ingested food was converted to heat to remain in the neutral zone (Moyes and Schulte 2008). Insufficient food would require reserves to be used and metabolically expensive shivering would be necessary to maintain body temperature. During 2003 and 2004, we used an intensive care unit supplied with chemical heat for 10 minutes for shivering birds to determine if the supplemental heat stopped the shivering.

Decline of native birds Changes in condition of birds

We had previously modeled the decline of native birds in both study areas from 2000 to 2007 (Freed and Cann 2012a), and of the Hawaii Akepa and other birds in the 1900-m site beginning in 2006 (Freed et al. 2008a). Here we deal with declines in the two areas with an estimated breakpoint from piecewise regression of survey data for each species in each area, and with community-wide declines by summing the densities of native species each year from 1987 to 2007. We also more fully document the decline in the 1900-m site. The estimated breakpoint from piecewise regression for each species in the open forest area used densities from years 1997 to 2007. This was because the estimates of density from years 1987 to 1994 had a negative relation between the estimate and its coefficient of variation, unlike later years, most likely because the people conducting the survey became more experienced in later years (Freed and Cann 2010). We modeled survival of each native species using mark/recapture methodology from 1987 to 1994 and 1995 to 1999 on the 1900-m site and showed no change in annual survival over those years (Freed and Cann 2010). This means the high community-wide estimate in 1987 and low estimate in 1988 (Fig. 3) are inconsistent with the constant adult survival and low reproductive rate (2-egg clutches) of these birds. The year 1996 also had a community-wide spike in density, which is why we use 1997–2007. For the closed forest area, surveys were first conducted in 1999.

We reanalyzed some of our data from the 1900-m site (Freed and Cann 2009 and Freed et al. 2008b) which showed significant differences in condition between 1987 and 1999 and after 1999 to determine if the major changes occurred during 2002 and later with increased exposure to white-eyes at higher density. For stunted bill growth, which used normalized data for each sex and each second-year native species (Freed and Cann 2009), we subtracted the below 0 mean scores from the above 0 mean scores for each year between 2002 and 2005, and compared 2002–2003 with 2004–2005 with a t-test. We also determined if a linear trend was evident by regressing the scores on year. We repeated this analysis for white-eyes. This test determined if stunted bill growth was worse during the latter set of years and if the native species and whiteeye had the same pattern. Different patterns were expected because white-eyes hatched and developed in the restoration area and dispersed into the old-growth forest while native birds hatched in the old-growth forest. For broken feathers, which were rare, we evaluate the number of cases in 2002 with those in 2000 and 2001. Extended and early molt became more prevalent beginning in 2002 compared with the two earlier years (Freed and Cann 2012b). We also determined change in prevalence of extended and early molt of white-eyes with tests of proportions. In addition, we counted the number of birds v www.esajournals.org

8


FREED AND CANN

Fig. 3. Annual sums of density of individual native species to portray community-wide surveys in the open forest area (open circles) and closed forest area (solid circles). The apparent increase in the community through 2000 was not indicated by increased annual adult survival of native species in our 1900-m site (Freed and Cann 2010), but declines in the open area during 2003 to 2007 are well documented in this site. This figure indicates that declines are widespread in the open forest area. Using the same data in the figure, but for individual species, Camp et al. (2010, 2014) assert that no declines have occurred in any native species at Hakalau during 1987–2007. They fail to recognize 1988 as an outlier low value that had no demographic or observational support during 1987 or 1989. The similar low value in 2007 was well supported by observation and demographic change before and after.

in each species in years 2003 through 2006 and adjusted the number of birds in 2006 by the maximum number of birds captured August through December within the three previous years. We then compared the mean number of birds during 2003 to 2005 with the adjusted number of birds in 2006 and calculated percentage decline or increase, evaluating change with a paired t-test. We document here the decline of five more species: Iiwi, Hawaii Amakihi, Apapane, Omao, and the Hawaii Elepaio, using declining recapture rates during 2002–2004 combined with a set of control years during 1989–1991 for those species. For the control years, we use a cut-off date of 1993 to determine if declining recapture rates simply occurred because the study ended in 2006. We consider recapture a success and use logistic regression with species as a factor and year as trend, along with the interaction. We also determine change in adult sex ratio of these species because of the drastic change in the

The piecewise regression was formed without an intercept because for most species the estimated intercept was an outlier. If some species in an area did not significantly decline, we used an overall test of P-values among all species in the area to assess significance (Winer 1971). We determine when the mean breakpoint for each area occurred relative to the white-eye increase. We perform a polynomial regression on the community-wide decline during 2002 to 2007. We show below that 2002 was the year before the mean breakpoint for species in both areas. We deal extensively with the declines on our long-term 1900-m study site because of changes in the condition of the birds (Freed and Cann 2009) and direct observations of the declines (Freed et al. 2008a). We calculated the number of adult birds (native and white-eye) each year during 2003 though 2005, where extensive banding occurred each month. During 2006, we banded during January through July on the 1900m site. We calculated the mean number of birds v www.esajournals.org

9


FREED AND CANN

endangered Hawaii Creeper from near equality to 75% males (Freed and Cann 2013b). Sex was established by cloacal protuberance in males, brood patches without protuberance in females, and by CHD molecular testing (Griffiths et al. 1998). We calculate adult sex ratios in years 1999– 2002 to establish the adult sex ratio before the declines, and in 2005 during the decline, and use logistic regression to analyze the difference in adult sex ratios. Year was a two-level factor (1999–2002 and 2005). The model included species, year, and the interaction. We performed a Pradel analysis of survival and recruitment in Program Mark (White and Burnham 1999), for each species (except the rare Akiapolaau) in the 1900-m site, and used the best supported model that estimated survival and recruitment each year from 2002 to 2006 to determine if recruitment made up for lower survival in 2006. Detection of survival was supplemented by resighting color-banded individuals of the akepa, creeper, and Iiwi.

We also performed a percentage similarity analysis of the community of native birds and white-eye in the open forest area using mean densities during 2000–2001 during the early white-eye increase and during 2006–2007 when declines occurred. In addition to providing percent similarity in the community at two points in time, the analysis also provided change in relative abundance of each species.

Greater niche overlap of the white-eye with each native species and niche shifts We estimated niche overlap using the Pianka symmetric measure (1973) for foraging substrates. Then we compared niche overlap of the white-eye and each native species with the mean of that native species with the other native species, with a paired t-test. The first model used all 18 substrates as if they had the same abundance, with proportional use of sites in relation to the total for each species. The second model was more detailed and eliminated substrates that were just seasonally available, such as flowers. We had shown that 87.7% of 3210 native trees, 10 cm or greater in diameter at breast height, were ohia-lehua and 2.9% were koa (Freed et al. 2008a). We compared the number of birds captured in matched nets that blocked two ohia-lehua or one ohia-lehua and one koa tree with a paired t-test (no nets blocked two koa trees) to determine if the latter captured fewer birds. We estimated for ohialehua that foliage, bark, and leaf buds constituted 75%, 15%, and 10% of the tree available to birds. For koa, we estimated that the finely dissected foliage, bark, wood, and lichen-covered branches constituted 65%, 15%, 15%, and 5% of the tree. We captured all species on our study sites in aerial mist-nets positioned at different heights. The 9.7% other plants were used to accommodate flycatching, frugivory, and occasional foliage gleaning, which we estimated represented 52%, 31%, and 17% of these substrates. Then we assigned proportional use of these foraging substrates by observations and data in Freed et al. (2008a), VanderWerf (1993), and Fretz (2000). Table 2 shows the resource availability and the estimated use by each species. The third model simulated greater utilization of the least used foraging substrates, incorporating the same resource levels and use as in the

Sustained exposure to the white-eye and community change Bender et al. (1984) identified sustained exposure to a perturbation as a press experiment. Sustained increases in density would have a cumulative effect on the native community leading to the declines. We calculated increased exposure to white-eyes by: first, calculating the median density of white-eyes in the open forest area during 1987–1999 and in the closed forest area during 1999–2001 (both before the increase of the introduced bird (Freed and Cann 2012a); second, subtracting the median from each year of density during 2000–2007 (open area) and 2002– 2007 (closed area); and third, summing the differences sequentially by year in each area. We used linear regression of cumulative increase in each area on year to estimate the extent of increased exposure to white-eyes over the previous median. To identify changes in composition of the community, we calculated the proportion of white-eyes in the combined native and whiteeye community each year. We began with the year before the increase, 1999 for the open forest area and 2001 for the closed forest area (Freed and Cann 2012a), and regressed the arcsinetransformed proportion on year. v www.esajournals.org

10


FREED AND CANN Table 2. Foraging substrates and use by species. Resource

Relative availability

JAWE

AKEP

HCRE

AKIP

HAAM

IIWI

APAP

ELEP

OMAO

Ohia foliage Ohia bark Ohia leaf buds Koa foliage Koa bark Koa wood Koa lichens Aerial insects Fruit Other plants

0.66 0.13 0.09 0.02 0.004 0.004 0.001 0.05 0.03 0.02

0.70 0.10 0 0.05 0 0 0.05 0 0 0.20

0.70 0.05 0.20 0.05 0 0 0 0 0 0

0.05 0.90 0 0.05 0 0 0 0 0 0

0.10 0 0 0 0.25 0.15 0.45 0 0 0.05

0.70 0.05 0 0.05 0 0 0.05 0 0 0.15

0.80 0.10 0 0 0.10 0 0 0 0 0

0.70 0.10 0 0.05 0.05 0 0.05 0 0 0.05

0.50 0.05 0 0.05 0 0 0 0.20 0 0.20

0.50 0.10 0 0.05 0.10 0 0 0 0.25 0

Note: Abbreviations are: JAWE, Japanese White-eye; AKEP, Hawaii Akepa; HCRE, Hawaii Creeper; AKIP, Akiapolaau; HAAM, Hawaii Amakihi; IIWI, Iiwi; APAP, Apapane; ELEP, Hawaii Elepaio; OMAO, Omao.

second analysis. We subtracted 0.10 from the most used substrate of each species and added it to the lowest used substrate or divided it evenly among the lowest used substrates. We investigated the potential for niche shifts in the white-eye and native birds using pole nets, which captured birds using the understory, and followed individual white-eyes which were initially captured in aerial nets and determined if any subsequent recaptures occurred in polenets. Table 3 shows net hours before and during the increase. We generated expected values from the difference in pole-based net hours operated before and during the increase and compared the observed white-eyes recaptured during both time periods with a chi-square goodness of fit test. We next compared captures of native birds and white-eyes in pole-nets for each bird initially captured during 1987–1999 and 2000–2006. Proportions of native birds and white-eyes were compared between time periods. We also compared white-eye and native bird captures in pole nets by month. Differential use of the understory by month may give birds an alternative source of food when canopy arthropods have declined from greater consumption from more birds, and used a contingency table analysis to identify differences in use of the understory.

We also documented an explosive increase in chewing lice in the birds (Freed et al. 2008b). All three of these alternatives require the host to expend energy fighting the infection or to replace heat that is lost through degraded plumage. We evaluate the role of pox virus by comparing prevalence of active sores in the 1900 and 1650-m sites during the white-eye increase using a test of proportions. We evaluate the role of chewing lice in the stunted bill growth of secondyear birds during 2004–2005. We compare the prevalence of lice in birds that had positive normalized bill length with birds that had negative normalized bill length with a test of Table 3. Mist-net hours before and during white-eye increase. Year 1987–1999 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Subtotal 2000–2006 2000 2001 2002 2003 2004 2005 2006 Subtotal

Alternative factors to changes in condition and the declines of native species We documented an increase in avian malaria at 1900 m, even in resident species, from introduced vector (Culex quinquefasciatus) movement (Freed et al. 2005, Freed and Cann 2013c). Avian poxvirus has been found on all species of native birds and is also transmitted by the same vector. v www.esajournals.org

11

Aerial net hours

Pole net hours

Total mist net hours

1019 3818 2745 2465 3276 674 1375 1613 642 648 310 204 1580 20,369

340 296

7 1333

1459 4114 2745 2465 3276 674 1473 1948 702 806 348 204 1587 21,702

1234 1123 3937 4330 5323 3877 722 20,546

72 292 349 657 851 41 2262

1234 1195 4229 4679 5980 4708 763 22,808

98 335 60 158 39


FREED AND CANN

feeding when predators appear. Changing air temperature has caused decline in birds (Moller et al. 2010). There was no change in mean air temperature from 1993 to 2006 (Freed and Cann 2012b), a period covering five El Nino events. However, the steepest decline of native birds occurred between 2006 and 2007, shown below. Lower ambient temperatures outside the thermal neutral zone of endotherms require more energy to maintain body temperature. From air temperatures measured every hour from the refuge weather station at 1941 m, we use mean daily temperature and daily maximum and minimum temperatures from 2002 to 2012. We calculate the mean temperature for each year and use a t-test to determine if cooling began in 2007. Then we determine if temperature changes resulted from lower mean maximum and/or lower mean minimum air temperatures by regressing each of those temperatures on year. We use multimodel inference (Burnham and Anderson 2002) to determine the role of competition and lower air temperature in the declines of native species in the open forest area. Competition models include cumulative increase in white-eyes over the 1987–1999 median beginning in 2002 through 2007, and discounted competition, where cumulative increase before the current year is divided by two, exponentially reducing previous increases and emphasizing the most recent increase. Air temperature models are based on years beginning in April and ending the next March, when surveys of birds are conducted. These models include the lowest temperature recorded during the year, the proportion of winter days (January, February, March) below the respective median temperatures for each month, and the proportion of all days during the year that were lower than the grand median of the respective months. We use an AICc difference less than two to consider models as having adequate support.

Fig. 4. Use of foraging substrates at Hakalau Forest National Wildlife Refuge by native passerine birds and the Japanese White-eye, indicating substantial overlap. From Freed et al. (2008a).

proportions. Malaria will be considered in the discussion. Predators can cause change in condition, and the refuge contains introduced mammalian predators (Lindsey et al. 2009), and an introduced avian predator and a native hawk (Klavitter 2009). The effect of predators on condition could result if vigilance and feeding nestlings are mutually exclusive activities and if more predators occurred during 2004–2005. When the Hawaiian hawk is present, the birds stop singing and hide in foliage. The number of predators was not estimated. However, if more predators existed in 2004 and 2005, and more vigilance occurred, fewer birds should have been captured those years. Total captures of birds during 2003, 2004, and 2005 were compared under the assumption that stunted growth and non-normal molt result from birds becoming vigilant and not v www.esajournals.org

RESULTS Diffuse competition Every native species of Hawaiian passerine at Hakalau overlaps multiple foraging substrates with the Japanese White-eye (Fig. 4). Bill lengths indicate mild sexual dimorphism in all species that increases with length (Fig. 5). The distribu12


FREED AND CANN

Fig. 5. Bill length of males and females of native birds and the Japanese White-eye. Samples were from birds captured during years 1990 to 1999 before bill length became stunted beginning in 2000 (Freed and Cann 2009). Females are represented in white bars and males in black bars.

tion, based on ordered female bill length, is much better supported by a linear than an exponential model (DAICc ¼ 66.8). Overall mean bill length ratio between species, after averaging between the sexes, is 1.14. However, the Hawaii Akepa, Hawai Amakihi, Apapane, and Iiwi, native species that forage mainly for arthropods on ohia-lehua foliage, have a mean ratio of 1.33. The white-eye has a mean bill length ratio less than 1.25 with half of the native species in the community, suggesting a potential for DC and later for RDC. Density of the white-eye in the open forest area of the refuge between 1977 and 1999 did not change (slope ¼0.08, P ¼ 0.32, R 2 ¼ 0.08; Fig. 6). Thus the population of this introduced bird was maintained at low density for 22 years after initially surveyed. Density-dependent differences in condition of birds occurred between the 1900 and 1700 m sites during the food shortage of 1995. The 1700 m site had fewer birds for seven of the eight native species in a similar area of forest and fewer white-eyes (paired t8 ¼ 3.4, P ¼ 0.01; Fig. 7). Birds at the denser 1900-m site had higher prevalence of low fat (logistic regression, species, P , 0.0001; v www.esajournals.org

site, P , 0.0001; interaction, P ¼ 0.67; Fig. 8A) and non-normal molt (species, P ¼ 0.0004, site, P ¼ 0.04; interaction, P ¼ 0.05; Fig. 8B). Poorer condition was indicated in all honeycreeper species and the white-eye in the denser 1900-m site. During 1994, there was no significant difference in non-normal molt between the sites (1/79 vs 21/335, test of proportions, X12 ¼ 2.3, P ¼ 0.13). No hatch-year white-eyes were captured in either site during 1995, in contrast to 66 and 68 native birds at 1900- and 1700-m sites, respectively. The number of hatch-year white-eyes was correlated with hatch-year native birds at the 1900-m site during 1987–1994 (Pearson r ¼ 0.78, t6 ¼ 3.1, P ¼ 0.02) and were captured each of those years. Thus the white-eye was apparently more limited in numbers by the same resources as were native birds in 1995, consistent with DC.

The white-eye increased in numbers before the native birds changed condition Randomization tests indicated no increase in the open forest area in 1998 or 1999 (P . 0.05). The increase began in 2000 (P , 0.05). In the open forest area, there was no increase or decline 13


FREED AND CANN

Fig. 6. Density of the Japanese White-eye in the open forest area between 1977 and 1999, indicating no change in density.

present in the open forest area beginning in 2000. White-eyes increased through 2006 on our study sites in the open forest area (Fig. 10), in contrast to the stepwise increase in the entire open forest area. Years 2000 to 2006 had higher

in native species between 1997–1999 and 2000– 2002 (mean change ¼ 0.44 birds/ha; paired t7 ¼ 0.63, P ¼ 0.55). Only the white-eye increased by 1.53 birds/ha during the same time period (t4 ¼ 2.89, P ¼ 0.04; Fig. 9). More birds were thus

Fig. 7. Number of individual birds captured (native birds and Japanese White-eyes) during 1994–1997 at 1700m site (open bars) and 1900-m site (filled bars), standardized by product of number of nets times mist-net hours.


14


FREED AND CANN

Fig. 8. Differences in prevalence of low fat (A) and non-normal molt (B) of native birds and Japanese Whiteeyes during months of January through June 1995 at 1700 m (open bars) and 1900 m (filled bars). These months during 1995 had a food shortage (Freed et al. 2007), with an anomalous spike of non-normal molt at 1900 m (Freed and Cann 2012b).

white-eye capture rates than the average for years 1987 to 1999 in the 1900-m site. (t6 ¼ 2.57, P ¼ 0.04). Year 2006 had the highest mist-net capture rate in the 1900- and 1770-m sites, close to the high during 2005 in the 1545-m site. There were 7 times 3 times 3 ways (63) for the highest

capture rate to occur at random in each of the sites. Only three ways had 2006 with highest capture rates or similar to the 2005 rate (P ¼ 0.048). Of note is the fact that the 1900-m and 1545-m study sites spanned the restoration area (Fig. 2).

Fig. 9. Comparison of changes in density of native birds and the Japanese White-eye between 1977–1999 and 2000–2002, indicating more birds present in the open forest area.


15


FREED AND CANN

Fig. 10. Increase in capture rates of Japanese White-eyes in mist-nets (individual birds per mist-net hour) in three study sites in the open forest area. (A) Long-term 1900-m study site. Left-most bar represents long-term capture rate during 1987–1999. (B) Shorter-term sites at 1770 and 1585 m. White-eyes increased after 2000 in all three study sites.

Cann 2009), and diverse changes in molt (Freed and Cann 2012b). Broken feathers (illustrated in Appendix A) occurred before 2000 but increased in prevalence thereafter. Of the 11 cases documented during 2000 to 2002, 9 occurred during 2002. Stunted growth of second-year native birds was minor during 2002–2003 (positive difference between birds above and below the mean), but negative differences occurred during 2004–2005 (t2 ¼ 6.5, P ¼ 0.023), and there was no linear trend (F1,2 ¼ 3.09, P ¼ 0.22; Fig. 12). In contrast, there was no difference between the set of years in the white-eye (t2 ¼ 2.27, P ¼ 0.15), but a linear trend existed (F1,2 ¼ 21.8, P ¼ 0.04; Fig. 12). Non-normal molt (extended molt, early molt) increased over background levels in 2000–2001 beginning in 2002, the same year asymmetric molt of primary fight feathers began (Freed and Cann 2012b). The white-eye had increased prevalence of extended molt in the 1900-m site (0.19 of 89 vs 0.59 of 298; X12 ¼ 42.2, P , 0.0001), but not the higher prevalence of early molt (0.18 of 28 vs 0.13 of 86; X12 ¼ 0.3, P ¼ 0.71). Sixteen birds were captured shivering, all during 2001–2004, and not earlier. Thirteen occurred in 2003–2004, well after a malaria

The dip in capture rate in 2004 (Fig. 10) will become important. Based on total captures of individual white-eyes per net hour over years 2002–2005, year 2004 had lower rates (1.49 vs 2.35 birds per 100 net hours; t2 ¼ 7.8, P ¼ 0.016). This suggests an interruption in propagule pressure, with 65 initially captured birds in 2004 versus a mean of 80 such birds from contiguous years. The extent of increase in our 1900-m site was realized by counting the number of adult whiteeyes each year. The year 1988 had mist-netting comparable to year 2002 (4114 and 4229 net hours (Table 3). As indicated in Fig. 11, the number of adult white-eyes persisted at high numbers for four years through 2005. There was a 6.6-fold increase in adult white-eyes between 1988 and those years. Based on the similarity of captures per net hour (Fig. 10), this increase occurred in the open forest sites close to the restoration area. Changes in condition of native birds and the white-eye occurred mainly during the 6.6-fold increase. All species of native birds measured encountered the same changes in broken feathers (Freed et al. 2008b), stunted growth (Freed and v www.esajournals.org

16


FREED AND CANN

Fig. 11. Number of adult white-eyes in the 1900-m site from 1987 to 2006. The year 1988 and 2002 had comparable net hours. The change in condition of birds was from 2002 through 2006.

epizootic in 2001 ending in 2002 (Freed et al. 2005). Shivering species included Hawaii Akepa (5), Hawaii Amakihi (2), Iiwi (3), Apapane (5), and a single white-eye. All shivering birds

captured while banding during 2003–2004 were warmed with an artificial heat source 10 minutes and ceased shivering when removed from the unit.

Fig. 12. Change in stunted bill length of native birds (circles) and white-eye (squares). The change is based on normalized score for each individual based on deviation from the mean score for that species and sex. Points are the mean difference between positive and negative normalized scores, summed across native species and also the Japanese White-eye.


17


FREED AND CANN Table 4. Estimated breakpoints in piecewise regression followed by second slope. Species Hawaii Akepa

Area

Open Closed Hawaii Creeper Open Closed Akiapolaau Open Closed Hawaii Amakihi Open Closed Iiwi Open Closed Apapane Open Closed Hawaii Elepaio Open Closed Omao Open Closed

Breakpoint Second slope 2004 2005 2005 2002 2003 2002 2006 2002 1998 2001 2005 2004 2001 2004 2006 2004

0.44 0.39 0.70 0.28 0.03 0.05 19.78 0.85 3.57 1.37 6.10 1.13 0.31 0.18 1.97 0.17

competition. Mean estimated breakpoints occurred in year 2003 in the closed forest area and between 2003 and 2004 in the open forest area. This was after the white-eye increase and the change in condition in 2002. There was no difference in distribution of breakpoints between areas (t14 ¼ 0.59, P ¼ 0.57). All native species in the open forest area had significant negative slopes after the breakpoint. All native species in the closed forest area had negative slopes, less negative than species in the open forest area, and not all slopes were significant. However, the overall test of P-values was highly significant (X162 ¼ 42.86, P ¼ 0.0003), indicating communitywide decline in the closed forest area as well. In contrast, white-eyes did not decline during 2003–2005 and had higher adult survival from initial capture during 2000–2006, compared with 1987–1998 (Kaplan-Meier test for difference in survival, X12 ¼ 8.5, P ¼ 0.004; Fig. 13). A bird banded by us in 2005 was recaptured by others in the same site in 2013, a record lifespan of at least eight years. Higher survival at each age occurred despite the potential survival period being almost three times as long (20 years) before 2000 compared with 7 years from 2000 to 2006.

P 0.005 0.09 0.03 0.35 0.01 0.04 0.009 0.0002 0.0001 0.06 0.008 0.27 0.0003 0.59 0.03 0.21

Native bird declines followed changes in condition The estimates of the breakpoint in piecewise regression of native bird densities are shown in Table 4. Only one species, the Iiwi, indicated a decline in the open forest area before the whiteeye increase. This was associated with a spike in white-eye density in 1996 (Fig. 6), and this species may been very susceptible to white-eye

Fig. 13. Changes in Japanese White-eye survival to different years from initial capture in the 1900-m site. The dashed line represents birds initially captured 1987–1998. The solid line represents birds initially captured 2000– 2005. The higher survival of the latter to each year occurred at higher density and was right censored from termination of our study.


18


FREED AND CANN

Four of five native species declined on our 1900-m study site. This was indicated by decreasing recapture rates without regard to sex during 2002–2004 (logistic regression, species, P , 0.0001, year, P ¼ 0.005, species by year interaction, P ¼ 0.11; Fig. 14A). Consistent with the piecewise regression, the major decrease in recapture rates occurred in 2004. The control years during 1989–1991 for these same species indicated no trend in recapture rate (logistic regression, species, P , 0.0001, year, P ¼ 0.57, interaction, P ¼ 0.83). Three of the five species had more male-biased adult sex ratios during 2005 than during 1999–2002 (logistic regression, species, P ¼ 0.26, year, P ¼ 0.01, species by year interaction, P ¼ 0.37; Fig. 14B), indicating that females declined faster than males. Between 2005 and 2006 there was a steep decline in birds, including the white-eye, at the 1900-m site (average 47%, paired t7 ¼ 5.41, P ¼ 0.001; Table 6). This preceded by a year the major decline in the open forest area between 2006 and 2007 (Fig. 3), indicating heterogeneity of decline in the area. The polynomial model for the entire open forest area had an R 2 ¼ 0.89, with significant linear decline (17.4 birds/ha/yr, P ¼ 0.03) and

Table 5. Pradel model estimates of survival and recruitment during 2005–2006. Species Hawaii Akepa Hawaii Creeper Hawaii Amakihi Iiwi Apapane Omao Hawaii Elepaio

Survival 95% C.I. Recruitment 95% C.I. (U) (U) (f ) (f ) 0.20 0.15 0.09 0.09 0.16 0.18 0.11

0.09, 0.04, 0.05, 0.04, 0.05, 0.05, 0.01,

0.40 0.40 0.17 0.19 0.40 0.45 0.53

0.04 0.06 0.17 0.24 0.30 0.12 0.06

0.002, 0.41 0.01, 0.42 0.10, 0.27 0.14, 0.36 0.13, 0.54 0.02, 0.44 0.002, 0.65

The percentage of birds recaptured during consecutive years was similar (72.7% vs 75.6%) before and during the white-eye increase (X12 ¼ 0.002, P ¼ 0.97), eliminating sampling effort as the cause of higher survival. Lower survival of native species was not compensated by higher recruitment (Table 5). The three most common species (Iiwi, Apapane, Hawaii Amakihi, with two broods per year) had higher recruitment than survival, but not sufficient to compensate for lower survival. The other species (Hawaii Akepa, Hawaii Creeper, Hawaii Elepaio, and Omao, with one brood per year) had lower recruitment than low survival.

Fig. 14. Declines of the five most common native birds on the 1900-m site. Decreasing recapture rate during 2002–2004 (A) Change in adult sex ratio from combined 1999 and 2001 to 2005 (B), indicating higher female mortality.


19


FREED AND CANN Table. 6. Declines of birds at the 1900-m site between 2005 and 2006. Species

Mean no. birds 2003-2005

Birds in 2006

Adjusted 2006

Percentage change

Hawaii Akepa Hawaii Creeper Akiapolaau Hawaii Amakihi Iiwi Apapane Hawaii Elepaio Omao Japanese White-eye

62 35.33 2.33 156.67 167 108 21.33 17.67 108

27 7 0 40 36 52 5 6 27

36 13 1 63 60 116 6 10 43

42 63 43 60 64 þ7 71 44 40

marginally significant quadratic decline (13.1 Change in community structure birds/ha/yr2, P ¼ 0.06). While a linear decline in The white-eyes expanded their role in the the closed forest area was not significant (P ¼ community in both forest areas. Between 2002 0.96), the two lowest values during 2004–2007 and 2007, the white-eye proportion increased were lower than the two lowest during 2001– from 4.4% to almost 10% of the community in the 2003, without overlap, consistent with the nega- open forest area (regression, slope ¼ 0.0047, F1,9 ¼ tive slopes of native species in the closed forest 6.4, P ¼ 0.04; Fig. 16A). The white-eye proportion area (Table 4). The major collapse in the open in the closed forest area increased more graduforest area and lesser declines in the closed forest ally from 4.8% to 7.8% over the 8-yr period (slope area were associated with more than double the ¼ 0.0033, F1,7 ¼ 8.34, P ¼ 0.023; Fig. 16B). cumulative increase in white-eyes over the The percentage similarity of community strucprevious medians (Fig. 15). ture before and during the decline was 95.65, reflecting the fact that no species went extinct and all but one species declined. Table 7 shows

Fig. 15. Cumulative exposure to increased number of Japanese White-eyes in open forest area (open circles and closed forest area (solid circles), calculated as increases over the baseline. Baseline for initial exposure in open forest area was 2.56 (median from 1987–1999, not shown). Baseline for initial exposure in closed forest area was 2.29 (median from 1999–2001). The last two years of the median in the closed forest are indicated by 0 values to compare the cumulative exposure of the two areas.


20


FREED AND CANN

Fig. 16. Proportional increased representation of Japanese White-eyes in the native community from 2000 to 2007 in the open forest area (A) and closed forest area (B).

the Hawaiian honeycreepers were represented less in the community while the Hawaii Elepaio, Omao, and Japanese White-eye rose in relative proportion.

species than the native species had with each other. This was consistent in all 3 models: model 1 with all 18 foraging substrates evenly used (paired t7 ¼ 4.6, P ¼ 0.003); model 2 with proportional resources and use by species as in Table 2 (paired t7 ¼ 2.6, P ¼ 0.038; Fig.17A); model 3 where 0.10 was subtracted from the highest used substrate and added to the least used substrates for each species (paired t7 ¼ 6.24, P ¼ 0.0004; Fig. 17B). There were major changes in use by individual white-eyes of the understory during their increase. White-eyes, initially captured in aerial

Greater niche overlap of the white-eye with native birds Nets that blocked two ohia-lehua trees captured more birds than nets that blocked a koa tree on one side (523 vs 172; Welch-modified t3.69 ¼ 4.5, P ¼ 0.01), indicating that koa trees were less used. Fig. 17 and Table 2 suggest that the whiteeye had greater niche overlap with each native

Table 7. Changes in community structure in the open forest area between 2000–2001 and 2006–2007, as density decreased from 56.4 to 34.3 birds/ha. Density

Proportion of community

Species

2001–2002

2006–2007

2000–2001

2006–2007

Hawaii Akepa Hawaii Creeper Akiapolaau Hawaii Amakihi Iiiwi Apapane Hawaii Elepaio Omao Japanese White-eye

1.63 1.22 0.16 12.58 20.41 10.68 3.14 2.01 4.31

0.77 0.75 0.08 7.25 11.85 6.36 2.50 1.75 2.98

0.029 0.022 0.003 0.228 0.362 0.189 0.056 0.036 0.076

0.022 0.022 0.002 0.221 0.346 0.185 0.073 0.051 0.087


21


FREED AND CANN

Fig. 17. Foraging substrate overlap between Japanese White-eyes and native species (open bars) and each native species with the mean overlap with the rest of the native community (filled bars). Model based on resource availability and use in Table 2 (A), and model based on subtraction of 0.10 from most used foraging substrate and addition of 0.10 to least used foraging substrate (B).

nets, were recaptured in pole nets both before and during the increase. A 66% increase (15 to 25) was expected from greater pole net hours. The observed 244% increase (9 to 31) (X12 ¼ 3.844, P ¼ 0.049) indicated that more white-eyes were captured than expected in pole nets. We infer that the white-eye used the understory more than native birds. There were 287 whiteeyes and 3683 native birds captured during 1987– 1999, and 398 compared with 3008 native birds captured during 2000–2006. The white-eye was captured in pole-nets more often than native birds both before and during the increase (before: X12 ¼ 28.47, P , 0.0001, during X12 ¼ 171.96, P , 0.0001; Fig. 18). Use of the understory by native birds also increased (X12 ¼ 22.98, P , 0.0001; Fig. 18), but did not equal the increased exploitation by white-eyes. Total captures of white-eyes and native birds in pole nets differed by month (X112 ¼ 50.98, P , 0.0001, even when collapsing months for sample sizes less than 5, X82 ¼ 50.56, P , 0.0001; Fig. 19). Modal white-eye captures occurred during June while native birds were captured more during March and April, when they were observed to v www.esajournals.org

collect dry grass as nesting material during the breeding season (Fig. 19).

Competition combined with climate cooling Years 2007–2012 had mean air temperature 0.698C lower than years 2002–2006 (t9 ¼ 5.3, P ¼ 0.0005; Fig. 20A). There was no trend during 2002–2006 (F1,3 ¼ 0.06, slope ¼ 0.018C, P ¼ 0.82). Year 2007 had lower temperature than any year during 2002–2006 as a consequence of mean annual minimum temperature decreasing during 2007–2012 (F1,4 ¼ 8.2, P ¼ 0.046 ) without significant change in mean annual maximum temperature (F1,4 ¼ 2.9, P ¼ 0.17; Fig. 20B). AIC models indicated a role for both competition and climate cooling. Declines supported by these factors varied among native species (Table 8). The three endangered honeycreepers had support only from the cumulative competition model. Two other species had dual support from the cumulative competition and minimum temperature models. Two species had support from the discounted competition model and temperature models. The Hawaii Elepaio did not have support from either competition model. 22


FREED AND CANN

Fig. 18. Native bird (white bars) and Japanese White-eye (black bars) captured in pole-based mist nets before and during the increase in white-eyes.

sores (test of proportions, X12 ¼ 5.0, P ¼ 0.025), suggesting that this mosquito-borne disease did not play a role. The prevalence of chewing lice did not differ in 41 second-year birds with stunted bills (61%) and 107 birds with normal bills (56%) (test of proportion, X12 ¼ 0.13, P ¼ 0.72). More birds

Alternatives to RDC Avian pox virus was not responsible for the native bird decline because active sores were more prevalent at the 1650-m site than the 1900m site during 2004–2005. Twenty-three of 296 individuals (0.078) at 1650 m but only 102 of 2228 individuals (0.046) at the 1900-m site had active

Fig. 19. Distribution of total captures of native birds (white bars) and Japanese White-eyes (black bars) in polebased mist-nets by month. These data include recaptures of individual birds.


23


FREED AND CANN

Fig. 20. Climate cooling at Hakalau Forest National Wildlife Refuge. Mean air temperature (A). Mean maximum and minimum temperatures (B), showing that the change in mean temperature was caused by lower minimum temperatures.

were captured during 2004 and 2005 (1330, 1380), when stunted growth was more serious, than during 2003 (1198) when stunted growth was less serious, suggesting no change in vigilance to predators. Other alternatives will be considered in the discussion.

was present at low density in the old-growth forest with the native bird community for at least 22 years before it started increasing in density in 2000. Declines of native species started after 2002. Here we provide additional evidence of DC, evaluate whether DC became reversed when the white-eye became more abundant, and identify RDC as a process in community ecology.

DISCUSSION

Total evidence of DC

The eight species of native passerine birds at Hakalau previously represented the most intact native bird community left in the main Hawaiian Islands (Scott et al. 1986), and reflect the resource partitioning that evolved over the last half million years, the age of the Island of Hawaii (Freed 1999, Lerner et al. 2011). The white-eye

The white-eye consumes the same food as native birds based on similar changes in condition with native birds during the food shortage in 1995. This indicates that the multiple foraging substrates shared with each native species represent niche overlap assumed by DC. That

Table 8. AICc differences from best supported competition or temperature model. Model

AKEP

HCRE

AKIP

HAAM

APAP

IIWI

OMAO

ELEP

Cumulative competition Discounted competition Minimum temperature Proportion winter days low temp. Proportion annual days low temp.

0 5.41 6.64 6.79 5.05

0 4.32 2.23 4.06 3.89

0 3.25 3.20 2.75 3.23

1.28 3.34 0 4.21 3.42

1.92 4.78 0 4.54 5.22

3.03 0 1.26 3.05 3.91

2.90 1.86 1.83 0 0.44

7.00 5.91 0 7.65 7.10

Note: See Table 2 for an explanation of bird species abbreviations.


24


FREED AND CANN

same year, Hart (2001) compared Hawaii Akepa fledglings to adults in mixed species flocks in both sites, documenting density-dependent reproductive success. Some hatch-year birds were expected in 1995 by the correlation with hatchyear native birds established during 1987–1994. Since no hatch-year white-eyes were captured at either site in 1995, the year of the yellow-jacket irruption, we infer that the white-eye fared worse than most native birds. The white-eye population could barely maintain itself during 1987–1999 in the 1900-m site with estimated adult survival of 0.56 and juvenile survival of 0.27 (Freed et al. 2008a). Estimates of k, the natural rate of population change, ranged from 0.80 to 0.97, depending on estimates of breeding success (Freed et al. 2008a). Some immigration may have been required from surrounding areas (Fig. 1) even to maintain the white-eye at low density on our study site. The restoration area was not involved as the source because trees were planted beginning in 1989 and white-eyes persisted at low density in that area for years after that (Freed and Cann 2012a). These data support DC as the mechanism limiting white-eye numbers for 22 years in the pocket where native birds had highest density. The introduced bird may have persisted at low density because it used food resources in the understory more than native species before the increase in density.

study site (Freed et al. 2008a). This lack of recruitment was not associated with juveniles having lower mass, the usual reason for lower juvenile survival in birds (Medeiros and Freed 2009). It is likely that juveniles produced and captured in the 1900-m site were displaced by adults at higher density. Juveniles may have taken a path of least resistance which would have been the closed forest area with lowest white-eye density during 2000 to 2003 (Freed and Cann 2012a). This could account for both the delayed increase in that area (Freed and Cann 2012a) and the lesser changes in condition and lesser mortality of native birds since they were less exposed to cumulative increases in the white-eye. No hatch-year white-eyes were captured in the 1650-m site in the closed forest during the two years of banding during 2004–2005. Thus whiteeyes from the restoration area displaced juveniles from the open forest sites which dispersed to the closed forest sites. This dynamic probably occurred throughout most of the open forest area contiguous with the restoration area.

White-eye density underlies the changes of condition of native birds Every change in condition of native birds occurred during 2002–2006, when the white-eye had achieved its higher density on our 1900-m study site. All species of native birds had broken wing and tail feathers, with 82% of cases in 2002. Broken feathers result from fault bars developed during molt, when food-limited individuals may miss one or more days of depositing melanin in the growing feather (King and Murphy 1984, Negro et al. 1994; Appendix A). Also, juveniles of all species of native birds in the 1900-m site had stunted growth (shorter tarsi, shorter bills, lower mass), which are caused by food shortages. The greatest change occurred during 2003–2004. During 2004–2005, juvenile native birds in the 1900- and 1770-m study sites had stunted growth, while those in the 1650-m site, with one-fifth the white-eye captures in mist-nets, had more normal growth (Freed and Cann 2009). The change in condition is the consequence of RDC because juveniles of all native species except the rare Akiapolaau (no juveniles captured) had stunted growth with higher white-eye density. The difference in growth of native species between sites makes stunted growth an outcome

Dynamics of the secondary invasion The density of white-eyes on our study sites in the open forest area was greater than that in the whole open forest area. There was continuing increase in the portion of the open forest area contiguous with the restoration area, bracketed by our 1900- and 1585-m study sites, and a 6.6fold increase in numbers on the 1900-m site starting in 2002 and persisting through 2005. This implies that the increased number of white-eyes in our study sites beginning in 2000 were derived from the restoration area where the white-eye population was growing exponentially (Freed and Cann 2012a). The increase of white-eyes in our study sites was based on propagules. White-eye fledglings and young juveniles were captured during the increase, there was no recruitment of locally produced juvenile white-eyes on our long-term v www.esajournals.org

25


FREED AND CANN

of competition. The white-eye also had stunted bill growth, expected from niche overlap with native species. However, it had a different trajectory compared with stunted bill growth of native species between 2002 and 2005. The increase in whiteeyes in old-growth forest was from white-eyes raised in the restoration area, whereas native species were raised in the old-growth forest. The white-eye population was growing exponentially in the restoration area (Freed and Cann 2012a), which may account for the linear worsening of stunted bill growth. In contrast, it took several years of exposure to the 6.6-fold increase of white-eyes before stunted bill growth of native species became worse. Beginning in 2002, there was higher prevalence of non-normal molt in all native species. Most native species had extended and early molt, asymmetric molt of primary flight feathers, and molt-breeding overlap in females (Freed and Cann 2012b). These changes in molt can be induced in birds maintained in the laboratory by withholding food (King and Murphy 1985, Murphy et al. 1988, Swaddle and Witter 1994). Moreover, beginning in 2003, between 10% and 30% of individuals of all native species had major fault bars on wing or tail feathers (Freed et al. 2008b). The timing of stunted growth followed by changes in molt suggests that birds were severely food limited throughout the year, not just the breeding season. Annual and spatial variation in non-normal molt are also associated with white-eye densities. In 2006, highest prevalence was in the 1650-m site in the closed forest area, where in previous years it was highest in the 1900- and 1770-m open forest sites. This was associated with a decline of white-eye density in the open forest area but a continuing increase in the closed forest area (Freed and Cann 2012b). In addition, prevalence of non-normal molt in native species was lower during 2004 than during contiguous years (Freed and Cann 2012b), associated with significantly fewer white-eyes present in 2004. Non-normal molt is another outcome of RDC because the entire native community was effected. So is asymmetric molt of primary flight feathers as we documented from 2002 to 2005 (Freed and Cann 2012b). The year 2003 had highest prevalence of asymmetric molt, the year v www.esajournals.org

that white-eye numbers were greatest on our 1900-m study site.

Alternative factors to RDC It is extremely difficult to identify alternative factors to account for the spatial and temporal variation in condition of native birds. Yellowjacket wasps, chewing lice (Phthiraptera), avian malaria epizootics, poxvirus, mammalian and avian predators, and parasitoid hymenoptera escaped from biological control are biotic factors that could affect all species. Here we deal with each of these. Mosquito-borne disease can increase food requirements to fight the infection. Malaria is considered a long-term problem for the birds at upper elevation (Atkinson and LaPointe 2009). However, malaria is rare (Freed et al. 2005, Freed and Cann 2013c), less than 5% in resident species from 1994 to 2002, a year before most native species began declining. This is because air temperature is too low for development of the parasite in the mosquito vector, did not increase between 2002 and 2006, and became cooler in 2007–2012. Several studies have attempted to detect mosquito eggs and larvae in streams at Hakalau (LaPointe 2000, Woodworth et al. 2001) and in oviposition buckets (Freed et al. 2005). Malaria could not have resulted in the increased non-normal molt in 2002 because no mosquito eggs or larvae were detected in the oviposition buckets operated that year and blood samples of birds tested mainly negative except for Hawaii Elepaio (Freed et al. 2005, Freed and Cann 2013c). Yet there was the 6.6-fold increase in white-eyes that year. Pox virus was more prevalent in our closed forest site, so could not have caused stunted growth which was more prevalent in our open forest study sites with five times the capture rate of white-eyes (Freed and Cann 2009). Yellow-jacket wasps accounted for the community-wide change in molt and fat in 1995, and our study may be the first to document competition between these wasps and native birds. However, afterwards wasps were monitored with baited traps, and nests were destroyed. The wasps could have existed only ephemerally, yet native birds suffered stunted growth and non-normal molt throughout the year from 2004 to 2005. Moreover, the wasps prefer nesting in 26


FREED AND CANN

more open areas and cannot account for more non-normal molt in the 1650-m site in 2006 when white-eyes increased. We documented densitydependence in condition of native birds and white-eyes in 1995 from these wasps which goes counter to results in 2002–2005. During those years, community wide densities were similar in the open forest area and closed forest area, yet stunted growth and non-normal molt were greater in the open forest area with more whiteeyes. Peck et al. (2008) identified introduced parasitoid wasps at Hakalau that escaped from biological control. These wasps parasitize native caterpillars and thus reduce food available to both native spiders and native birds. However, these wasps were more abundant at lower elevations, where birds had more normal growth and lesser declines with fewer white-eyes. Thus they could have had only a minor effect on the worse condition of growth and declines of birds in the open forest area. Diverse introduced predators exist on the refuge: three species of rats (Rattus rattus, R. exulans, R. novegicus), the small Indian mongoose (Herpestes auropunctatus), feral cats (Felis catus), and the barn owl (Tyto alba) as well as the native hawk. The nocturnal predators could not have caused stunted growth because parental passerines do not forage at night. Nests are usually located high in trees in terminal twigs too delicate for these predators to normally reach, so females could continue to brood nestlings with little disturbance. Diurnal predators potentially could have caused the stunted growth and nonnormal molt because vigilance and food acquisition are mutually exclusive activities. However, more birds were captured during 2004 and 2005 than during 2003, inconsistent with greater vigilance those years compared with 2003. Avian chewing lice include Ischnocera, which chew feathers, and Amblycera, which can also consume blood. However, chewing lice could not have caused the change in condition of native birds because non-normal molt increased in year 2002, with the 6.6-fold increase in white-eyes, but the lice increased in 2003 (Freed et al. 2008b). Prevalence of lice increased annually from 2003 to 2005, so the lowering of non-normal molt in 2004 occurred independently of lice. In addition, second year birds with negative relative bill v www.esajournals.org

lengths did not have significantly higher prevalence of chewing lice than birds with positive relative bill lengths. However, lice had one effect. Native birds with lice had higher prevalence of major fault bars in their wing or tail feathers than birds without lice (Freed et al. 2008b). Therefore, lice would have amplified slightly the apparent strength of RDC from white-eyes. Effects of lice on hosts in other systems have been documented mainly in nestlings and in the absence of another factor causing food limitation (Price et al. 2003). The only study that showed effects of ectoparasites on survival of adult hosts, from a fumigation experiment, had swallow bugs (Oeciacus), and bird fleas (Ceratophyllus) as well as Ischnocera and Amblycera chewing lice (Brown et al. 1995). In summary, white-eye density accounts for variation in condition of native birds in space and time. No other biotic factors can account for that variation. Appendix B documents competition between white-eyes and individual native species.

Declines of native species Stunted growth led in part to the declines. Juveniles with lower mass had lower juvenile survival (Freed and Cann 2009), typical for malnourished birds (Martin 1987, Medeiros and Freed 2009). Second-year birds with stunted bill growth had lower second-year survival, and even older adults with stunted bills, that survived their second year, were less likely to be recaptured than adults with normal size bills (Freed and Cann 2009). There was no evolutionary change in native birds because the mass of surviving juveniles was the same before and during the white-eye increase, as was the bill length of surviving adults (Freed and Cann 2009). Stunted growth thus led to extensive normalizing selection with the loss of suboptimal individuals throughout most of the native community (Freed and Cann 2009). This may be the strongest example of such selection in nature. The lesser declines in the closed forest area (Freed and Cann 2012a) were based on more normal growth with fewer white-eyes (Freed and Cann 2009). Competition during 2002–2005 had a greater effect on females than on males in five species, four documented here and the endangered 27


FREED AND CANN

Hawaii Creeper (Freed and Cann 2013b). In four species there was a substantial increase in moltbreeding overlap that rarely or never occurred before (Freed and Cann 2012b). While all species of native birds have biparental care of nestlings and fledglings, female Hawaiian honeycreepers engage in that care after producing a clutch of two eggs and incubating them alone, supplemented by mate feedings. The high mortality of females may have been a cost of reproduction during food limitation, as identified for the creeper (Freed and Cann 2013b). However, this overlap did not occur in the Apapane, which instead had the greatest change in extended molt in November. Limiting similarity (Hutchinson 1959, MacArthur and Levins 1967) may also have played a role because females of all species except the Hawaii Akepa are closer in bill length to white-eyes than males, particularly white-eye males which are the more abundant sex (Freed and Cann 2013b). The akepa bill is almost identical in length to the white-eye bill in both sexes, which is perhaps the reason why both sexes declined equivalently (Freed et al. 2008a). We directly observed the decline of native birds on our 1900- and 1770-m study sites during 2006–2008 (Freed et al. 2008a). At the 1900-m site in 2006 even common species were rare and only about half the number of the usually long-lived Hawaii Akepa survived from 2005. During 2008, at the 1770 m site, a 2-hr inspection during fine weather in March, the height of the mating season, did not detect a single native bird. This never happened before in 20 years of research. These observations were consistent with the great drop in density in the open forest area between 2006 and 2007. In 2006, female akepa dropped one syllable in their two-syllable begging call to males, and in two cases continued begging after being fed by their mate; in 2008, fledgling Hawaii Creeper dropped one or two syllables in their normal three-syllable begging call to parents. In 2008, half of the family groups of creepers lost their fledglings well before the termination of parental care. These observations suggest that the white-eye may have recovered in the 1900-m site after declining in 2006 (Table 6). It is also possible that the lowered carrying capacity continued from arthropod community dynamics resulting from RDC. Hawaii Akepa numbered 31 birds in the v www.esajournals.org

1900-m site in 2013 (E. Paxton, personal communication), indicating that this endangered bird has not recovered over the decline in 2006, with 36 birds on site. Far from being a cryptic member of the community, this species has a distinctive call and song, group displays, and males are flamboyantly bright orange (Lepson and Freed 1995).

Changes in white-eye demography and behavior that contributed to RDC Increased white-eye adult survival to all ages after initial capture indicates an adaptive response to higher density, and represents the first case of inverse density-dependence in an increasing population (Courchamp et al. 1999). Higher juvenile and adult survival occurs with better nestling condition (Lindstrom 1999, Medeiros and Freed 2009, Saino et al. 2012), but extended survival 4 to 8 years in the future is unprecedented. It is important to determine if the higher survival is based on nestling condition in the restoration area or natural selection favoring higher survival at higher density. Higher whiteeye adult survival occurred before the decline of native birds, and indicates asymmetric competition, which is implied by RDC because of the greater niche overlap of the white-eye with each native species. Greater capture in pole nets during the increase than native birds indicates that whiteeyes exploited the understory much more than native birds, especially in June when arthropod abundance in the ohia-lehua foliage is at a seasonal low point (Freed et al. 2007). However, during 2004–2005 there was no difference in captures of native species and white-eyes in pole nets between the 1900-m and 1650-m sites by species (Freed et al. 2008a). This suggests that the understory is not an empty niche exploited exclusively by the white-eye. Rather, the bird used the understory more when arthropods in the canopy were limiting, as part of its RDC on native birds. This extreme generalization in the white-eye may be traced to the family Zosteropidae, which is renowned for range expansion and niche diversification (Moyle et al. 2009). During range expansion on a South Pacific island, a congener, Zosterops lateralis, had greater niche breadth at the population level because individuals varied in their personal niche breath (Scott et al. 2003). 28


FREED AND CANN

The niche breadth of the Japanese White-eye at Hakalau is so great that individuals could also vary in their personal niche breadth, as occurs among diverse taxa (Bolnick et al. 2003). This may be the adaptation that underlies the increased survival under higher density with stunted bills (Freed and Cann 2009). The whiteeye persisted despite stunted bills becoming more so each year. Individual white-eyes could have foraged in substrates most suited to their individual bill length, and for resources in the understory underutilized by the native bird community. Relative to the white-eye, the entire native community is specialized, which is a condition of RDC.

except for females when breeding. Native birds perch in open sunny areas during the day and probably have torpor at night in dense foliage to minimize heat loss. The shivering detected only during the white-eye increase, corrected in the intensive care units, thus indicates insufficient food during stressful periods with which to generate sufficient heat through normal metabolic pathways. In an environment where air temperature is far below the lower critical value of the thermal neutral zone, the worst problem for a small food-limited endotherm is for air temperatures to become even lower. Competition from an introduced bird and lower air temperatures represents a novel synergy among those previously identified in the historical decline of Hawaiian forest birds (Pimm 1996). While great attention has been paid to climate warming globally (Parmesan and Yohe 2003, Parmesan 2006, Rosenzweig et al. 2008, Moller et al. 2010) and in Hawaii (Loope and Giambelluca 1998, Giambelluca et al. 2008), little attention has been directed at climate cooling for birds (Bicudo et al. 2010, Moller et al. 2010). However, Giambelluca et al. (2008) documented climate cooling in Hawaii during the mid-1990s when native birds still had high survival (Freed and Cann 2010). Lower air temperatures require more food to replace the additional heat that is lost, but competition from white-eyes makes less food available. The Japanese white-eye ranges from sea level to 2900-m elevation in Vietnam (Robson 2005), which may be the reason it did not suffer as much from climate cooling in Hawaii.

Air temperature and competition Lower mean air temperature, documented from 2007 to 2012, is an abiotic factor that affected most native species. The 1900-m site is close to the 2250-m base of the inversion layer in nearby Hilo (Cao et al. 2007), which frequently descends at night, bringing cooler air above the layer to 1900 m. The morning after such descent we can look down from 1900 m to the tops of the clouds. During the months of October through March, frost appears on the ground in open areas during such descent. This movement is the likely reason there was no change in temperature through 2006 and why climate cooling occurred during 2007–2012, thus minimizing the likelihood of night-biting mosquitoes causing undetected malarial epizootics. As the AIC modeling indicates (Table 8), air temperature is not an alternative but complements competition with white-eyes for four species. The three endangered species declined from cumulative competition without contributions from low air temperatures. The mean high air temperature of 13.68C during August (1993–2006) (Freed and Cann 2012b) is far below the lower critical value 318C of the thermal neutral zone. This requires physiological adaptation to exist at upper elevation. Tropical biologists have noticed that native Hawaiian forest birds forage all day long, unlike the early morning and late afternoon flurry of activity typical of other tropical birds. Most of the energy ingested thus provides heat to maintain body temperature. These birds typically have low fat levels ranging from trace to partially full, v www.esajournals.org

Summary of the evidence for RDC Diffuse competition was documented for 22 years (Condition 1 based on similarity in foraging substrates). The white-eye increased from propagule pressure, creating a higher overall density of birds in the forest (Condition 2). The entire native community had stunted growth and diverse changes in molt. The stunted growth led to normalizing selection on the 1900m site and declines throughout the open forest area (Condition 3). The white-eye had greater niche overlap with each native species than a native species had with the rest of the native community (Condition 4). This study highlights the mechanism of competition which is food 29


FREED AND CANN

limitation caused by increases of white-eyes, because all problems occurred during 2002– 2006 when white-eyes had increased 6.6-fold. The food limitation is caused by pure exploitative competition and the decline of the white-eye and native birds in 2006 is consistent with competition theory where carrying capacity is diminished (Keddy 2001). By 2007, white-eyes represented 1 out of 10 individuals in the open forest area and the relative abundance of species in the community had changed.

using the resources. A tradeoff promoting coexistence could also develop if the better competitor were more subject to predation, climatic variation, or disease. Tilman (2007) predicted that a species in a community without tradeoffs would outcompete the entire community. The Japanese white-eye appears to be such a species because it did not suffer survival changes with stunted bill growth as did the native species (this study, Freed and Cann 2009). Rather it increased adult survival with shorter bills. It also apparently does not suffer high mortality to avian malaria by high density at lower elevations where malaria is more prevalent (van Riper et al. 1986). The white-eye occurs at highest density at low elevations (Scott et al. 1986). Meszena et al. (2006), in a theoretical study of competitive exclusion and limiting similarity, advance the concept of impact and sensitivity niches for each population. Impact niches have differential impact on population regulating variables. Sensitivity niches have differential sensitivity towards regulating variables. The greater niche overlap of the white-eye with each species of native bird than native birds have with each other may make these two niches nearly the same. This is indicated by higher white-eye density leading to stunted growth and higher mortality of native species. According to the theory of Meszena et al. (2006), the asymmetric competition makes coexistence of the white-eye at high density and native species impossible.

How important is RDC as a process in community ecology? There are many examples of a single species causing a decline in multiple native species. These include weedy plants (Radosevich et al. 1997, Booth et al. 2003, Forseth and Innis 2004), zebra mussels (Ricciardi et al. 1998), ants (Porter and Savignano 1990, Human and Gordon 1996, Holway 1999), fish (Schoenherr 1981, Herbold and Moyle 1986, Ogutu-Ohwayo 1990, Ross 1991), and birds (Mountainspring and Scott 1985, Tindall et al. 2007, Grarock et al. 2012). If the declines in native species were based on competition for food or other resources, this would be consistent with RDC. Reverse diffuse competition can include interference as well as exploitative competition (e.g., Holway 1999, Tindall et al. 2007). However, some of these examples incorporate biotic processes in addition to competition, including allelopathy, predation, and biofouling. It is important to isolate the effects of competition from these other factors as we did here. Reverse diffuse competition was actually modeled by Case (1990) in a simulation study of DC. By drawing competition coefficients of the core (native) community and the invader from the same distribution, with the invader starting at very low density, he found in some simulations that the invader replaced one or more native species. Case did not compare competition coefficients, but surely there were instances where the invader had higher competition coefficients than most native species, similar to the white-eye/native birds niche overlap. Tilman (2007) developed a mechanistic model of competition for plants that may apply to RDC. He showed that species could coexist sharing two resources if they had a tradeoff in efficiency v www.esajournals.org

CONCLUSION Reverse diffuse competition integrates invasion biology with community ecology by combining niche overlap and propagule pressure, and making the intensity of competition indistinguishable from its importance (Welden and Slauson 1986). This is the first case where an introduced bird has its greatest effect in a natural habitat from competitive superiority (Sol et al. 2012). Replacement of native species makes the bird a driver of ecological change (Didham et al. 2006). The RDC is affecting specialized and generalized native birds equivalently, unlike other systems where specialized species decline first (Devictor et al. 2008). The effects on the birds almost certainly extend to the arthropods they use as prey, thus affecting overall biodiversity 30


FREED AND CANN Hawaiian honeycreepers. Journal of Avian Medicine and Surgery 23:53–63. Bender, E. A., T. J. Case, and M. E. Gilpin. 1984. Perturbation experiments in community ecology: theory and practice. Ecology 65:1–13. Bicudo, J. E. P. W., W. A. Buttemer, M. A. Chapell, J. T. Pearson, and C. Bech. 2010. Ecological and environmental physiology of birds. Oxford University Press, Oxford, UK. Blackburn, T. M., J. L. Lockwood, and P. Cassey. 2009. Avian invasions: the ecology and evolution of exotic birds. Oxford University Press, Oxford, UK. Bock, C. E., A. J. Cruz, M. C. Grant, C. S. Aid, and T. R. Strong. 1992. Field experimental evidence for diffuse competition among southwestern riparian birds. American Naturalist 140:815–828. Bolnick, D. I., R. Svanba¨ck, J. A. Fordyce, L. H. Yang, J. M. Davis, C. D. Hulsey, and M. L. Forister. 2003. The ecology of individuals: incidence and implications of individual specialization. American Naturalist 161:1–28. Booth, B. D., S. D. Murphy, and C. J. Swanton. 2003. Weed ecology in natural and agricultural systems. CABI, Wallingford, UK. Brown, C. R., M. B. Brown, and B. Rannala. 1995. Ectoparasites reduce long-term survival of their avian host. Proceedings of the Royal Society B 262:313–319. Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 1993. Distance sampling: estimating abundance of biological populations. Chapman and Hall, London, UK. Burger, J. C., and S. M. Louda. 1995. Interaction of diffuse competition and insect herbivory in limiting brittle prickly pear cactus, Opuntia fragilis (Cactaceae). American Journal of Botany 82:1558–1566. Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference. Second edition. Springer, New York, New York, USA. Callaway, R. M., and S. C. Pennings. 2000. Facilitation may buffer competitive effects: indirect and diffuse interactions among salt marsh plants. American Naturalist 156:416–424. Camp, R. J., T. K. Pratt, P. M. Gorresen, J. J. Jeffrey, and B. L. Woodworth. 2009. Passerine bird trends at Hakalau Forest National Wildlife Refuge, Hawai’i. Hawaii Cooperative Studies Unit Technical Report HCSU-011. University of Hawaii at Hilo, Hawaii, USA. Camp, R. J., T. K. Pratt, P. M. Gorresen, J. J. Jeffrey, and B. L. Woodworth. 2010. Population trends of forest birds at Hakalau Forest National Wildlife Refuge, Hawaii. Condor 112:196–212. Camp, R. J., T. K. Pratt, P. M. Gorresen, B. L. Woodworth, and J. J. Jeffrey. 2014. Hawaiian forest bird trends: using log-linear models to assess long-term trends is supported by model diagnostics and

(Davis 2003). The timing of the increases with the white-eye population growth in the restoration area makes this an extreme example of unintended negative consequences associated with restoration. Hakalau Forest National Wildlife Refuge needs to revise the recently established 15-year conservation plan to include management of white-eyes at the refuge. This paper indicates that such management will prevent further declines of native Hawaiian birds and recover the community in the open forest area and even in the closed forest area. We have been encouraging the refuge to control white-eyes since 2004, although no action has been taken. The only recourse for native birds without such management is to disperse to lower elevations with lower densities of white-eyes. However, malaria will become increasingly a problem (Freed and Cann 2013c). As recently as 2013, endangered species had been seen at lower elevations than detected during the Hawaiian Forest Bird Survey in 1977. Such movement represents an ecological trap between food limitation at upper elevation and avian malaria at lower elevation (Schlaepfer et al. 2002, 2005). There may be insufficient time for these species to evolve adaptations allowing them to persist outside old-growth forest in which they evolved.

ACKNOWLEDGMENTS We are grateful to the following people for discussion of ideas and improving the manuscript: Thomas Smith, Floyd Reed, Mark Hixon, Ottar Bjornstadt, Amber Wright, and anonymous reviewers. We thank Eben Paxton for communicating the recapture of the 8yr old white-eye and the number of Hawaii Akepa in the 1900-m site in 2013. We appreciate funding from the John D. and Catherine T. MacArthur Foundation World Environment and Resources Program (8900287), the National Center for Environmental Research (Science to Achieve Results, Environmental Protection Agency R82-9093). The MacArthur Foundation, National Science Foundation (DBI96-02547), and the University of Hawaii at Manoa provided funding to build Hakalau Forest Biological Field Station that facilitated our research.

LITERATURE CITED Atkinson, C. T., and D. A. LaPointe. 2009. Introduced avian diseases, climate change, and the future of


31


FREED AND CANN assumptions (reply to Freed and Cann 2013). Condor 116:97–101. Cao, G., T. W. Giambelluca, D. E. Stevens, and T. A. Schroeder. 2007. Inversion variability in the Hawaiian trade wind regime. Journal of Climate 20:1145–1160. Carlquist, S. 1980. Hawaii, a natural history. Second edition. Pacific Tropical Botanical Garden, Lawai, Hawaii, USA. Case, T. J. 1990. Invasion resistance arises in strongly interacting species-rich model competition communities. Proceedings of the National Academy of Science USA 87:9610–9614. Caum, E. L. 1933. The exotic birds of Hawaii. Occasional Papers of the B.P. Bishop Museum 10:1–55. Colautti, R. I., I. A. Grigorovich, and H. J. MacIsaac. 2006. Propagule pressure: a null model for biological invasions. Biological Invasions 8:1023–1037. Courchamp, F., T. Clutton-Brock, and B. Grenfell. 1999. Inverse density dependence and the Allee effect. Trends in Ecology and Evolution 14:405–410. Crooks, J. A. 2005. Lag times and exotic species: the ecology and management of biological invasions in slow-motion. Ecoscience 12:316–329. Crooks, J. A., and M. E. Soule. 1999. Lag times in population explosions of invasive species: causes and implications. Pages 103–125 in O. T. Sandlund, P. J. Schei, and A. V. Viken, editors. Invasive species and biodiversity management. Kluwer, Dordrecht, The Netherlands. Davidson, D. W. 1980. Some consequences of diffuse competition in a desert seed-eating ant. American Naturalist 116:92–105. Davidson, D. W. 1985. An experimental study of diffuse competition in harvester ants. American Naturalist 125:500–506. Davis, M. 2003. Biotic globalization: does competition from introduced species threaten biodiversity. BioScience 53:481–489. Devictor, V., R. Julliard, and F. Jiguet. 2008. Distribution of specialist and generalist species along spatial gradients of habitat disturbance and fragmentation. Oikos 117:507–514. Dhondt, A. A. 2012. Interspecific competition in birds. Oxford University Press, Oxford, UK. Didham, R. K., J. M. Tylianakis, M. A. Hutchinson, R. M. Ewers, and N. J. Gemmell. 2006. Are invasive species the drivers of ecological change? Trends in Ecology and Evolution 20:470–474. Dyer, A. R., and K. J. Rice. 1997. Intraspecific and diffuse competition: the response of Nassela pulchra in a California grassland. Ecological Applications 7:484–492. Forseth, I. N., and A. F. Innis. 2004. Kudzu (Pueraria montana): history, physiology, and ecology combine to make a major ecosystem threat. Critical Reviews


in Plant Science 23:401–413. Freed, L. A. 1999. Evolution and extinction of Hawaiian honeycreepers: a comparative study. Pages 137–162 in L. Landweber and A. Dobson, editors. Genetics and the extinction of species. Princeton University Press, Princeton, New Jersey, USA. Freed, L. A. 2001. Significance of old-growth forest to the Hawaii akepa. Studies in Avian Biology 22:173– 184. Freed, L. A., and R. L. Cann. 2009. Negative effects of an introduced bird species on growth and survival in a native bird community. Current Biology 19:1736–1740. Freed, L. A., and R. L. Cann. 2010. Misleading trend analysis and decline of Hawaiian forest birds. Condor 112:213–221. Freed, L. A., and R. L. Cann. 2012a. Increase of an introduced bird competitor in old-growth forest associated with restoration. NeoBiota 13:43–60. Freed, L. A., and R. L. Cann. 2012b. Changes in timing, duration and symmetry of molt are associated with extensive decline of Hawaiian forest birds. PLoS One 7:e29834. Freed, L. A., and R. L. Cann. 2013a. More misleading trend analysis of Hawaiian forest birds. Condor 115:442–447. Freed, L. A., and R. L. Cann. 2013b. Females lead population collapse of the endangered Hawaii creeper. PLoS One 8:e67914. Freed, L. A., and R. L. Cann. 2013c. Vector movement underlies avian malaria at upper elevation in Hawaii: implications for transmission of human malaria. Parasitology Research 112:3887–3895. Freed, L. A., R. L. Cann, and G. R. Bodner. 2008a. Incipient extinction of a major population of the Hawaii akepa owing to introduced species. Evolutionary Ecology Research 10:931–965. Freed, L. A., R. L. Cann, and K. L. Diller. 2009. Sexual dimorphism and the evolution of seasonal varation in sex allocation in the Hawaii akepa. Evolutionary Ecology Research 11:731–757. Freed, L. A., R. L. Cann, M. L. Goff, W. A. Kuntz, and G. R. Bodner. 2005. Increase in avian malaria at upper elevation in Hawaii. Condor 107:753–764. Freed, L. A., J. S. Fretz, and M. C. Medeiros. 2007. Adaptation in the Hawaii akepa to breed and moult during a seasonal food decline. Evolutionary Ecology Research 9:157–167. Freed, L. A., M. C. Medeiros, and G. R. Bodner. 2008b. Explosive increase in ectoparasites in Hawaiian forest birds. Journal of Parasitology 94:1009–1021. Fretz, J. S. 2000. Relationship of canopy arthropod prey to distribution and life history of the Hawaii akepa. Dissertation. University of Hawaii at Manoa, Honolulu, Hawaii, USA. Fryar, S. C., W. Booth, J. Davies, I. J. Hodgkiss, and

32


FREED AND CANN K. D. Hyde. 2005. Evidence of in situ competition between fungi in freshwater. Fungal Diversity 18:59–71. Futuyma, D. J. 2001. Ecological specialization and generalization. Pages 177–189 in C. W. Fox, D. A. Roff, and D. J. Fairbairn, editors. Evolutionary ecology: concepts and case studies. Oxford University Press, Oxford, UK. Gambino, P., A. C. Medeiros, and L. L. Loope. 1987. Introduced vespids Paravespula pensylvanica prey on Maui’s endemic arthropod fauna. Journal of Tropical Ecology 3:169–170. Giambelluca, T. W., H. F. Diaz, and M. S. A. Luke. 2008. Secular temperature changes in Hawaii. Geophysical Research Letters 35:L12702. Grarock, K., C. R. Tidemann, J. Wood, and D. B. Lindenmayer. 2012. Is it benign or is it a pariah? Empirical evidence for the impact of the common myna (Acridotheres tristis) on Australian birds. PLoS One 7:e40622. Griffiths, R., M. C. Donble, K. Orr, and R. J. G. Dawson. 1998. A DNA test to sex most birds. Molecular Ecology 7:1071–1075. Hart, P. J. 2001. Demographic comparisons between high and low density populations of Hawaii Akepa. Studies in Avian Biology 22:185–193. Hart, P. J., and L. A. Freed. 2003. Structure and dynamics of interspecific flocks of Hawaiian birds. Auk 120:82–95. Hellmann, J. J., J. E. Byers, B. G. Bierwasen, and J. S. Dukes. 2008. Five potential consequences of climate change for invasive species. Conservation Biology 22:534–543. Herbold, B., and P. B. Moyle. 1986. Introduced species and vacant niches. American Naturalist 128:751– 760. Heske, E. J., J. H. Brown, and S. Mistry. 1994. Longterm experimental study of a Chihuahuan desert rodent community: 13 years of competition. Ecology 75:438–445. Hollebone, A. L., and M. E. Hay. 2007. Propagule pressure of an invasive crab overwhelms native biotic resistance. Marine Ecology Progress Series 342:191–196. Holmer, L., and J. Stenlid. 1996. Diffuse competition for heterogeneous substrate in soil among six species of wood-decomposing basidomycetes. Oecologia 106:531–538. Holt, R. D. 1993. Ecology at the mesoscale: the influence of regional processes on local communities. Pages 77–88 in R. E. Ricklefs and D. Schluter, editors. Species diversity in ecological communities. University of Chicago Press, Chicago, Illinois, USA. Holway, D. A. 1999. Competitive mechanisms underlying the displacement of native ants by the invasive Argentine ant. Ecology 80:238–251.


Hufbauer, R. A., and M. E. Torchin. 2007. Integrating ecological and evolutionary theory of biological invasions. Pages 79–96 in W. Nentwig, editor. Biological invasions. Springer, Berlin, Germany. Human, K. G., and D. M. Gordon. 1996. Exploitation and interference competition between the invasive Argentine ant, Linepithema humile, and native ant species. Oecologia 105:405–412. Hutchinson, G. E. 1959. Homage to Santa Rosalia, or why are there so many kinds of animals. American Naturalist 93:145–159. Keane, R. M., and M. J. Crawley. 2002. Exotic plant invasives and the enemy release hypothesis. Trends in Ecology and Evolution 17:164–170. Keddy, P. A. 2001. Competition. Second edition. Kluwer, Dordrecht, The Netherlands. King, J. R., and M. E. Murphy. 1984. Fault bars in the feathers of white-crowned sparrows: dietary deficiency or stress of capture and handling? Auk 101:168–179. King, J. R., and M. E. Murphy. 1985. Periods of nutritional stress in the annual cycles of endotherms: fact or fiction. American Zoologist 25:955– 964. Klavitter, J. L. 2009. The ecology and conservation of Hawaiian raptors. Pages 293–311 in T. K. Pratt, C. T. Atkinson, P. C. Banko, J. D. Jacobi, and B. L. Woodworth, editors. Conservation biology of Hawaiian forest birds. Yale University Press, New Haven, Connecticut, USA. Kolbe, J. J., R. E. Glor, L. R. Schettino, A. C. Lara, A. Larson, and J. B. Losos. 2004. Genetic variation increases during biological invasion by a Cuban lizard. Nature 431:177–181. LaPointe, D. A. 2000. Avian malaria in Hawaii: the distribution, ecology, and vector potential of forestdwelling mosquitoes. Dissertation. University of Hawaii at Manoa, Honolulu, Hawaii, USA. Lepson, J. K., and L. A. Freed. 1995. Variation in male plumage and behavior of the Hawaii akepa. Auk 112:402–414. Lerner, H. R. L., M. Meyer, H. F. James, M. Hofreiter, and R. C. Fleischer. 2011. Multilocus resolution of phylogeny and timescale in the extant adaptive radiation of Hawaiian honeycreepers. Current Biology 21:1838–1844. Lindsey, G. D., S. C. Hess, E. W. I. Campell, and R. T. Sugihara. 2009. Small mammals as predators and competitors. Pages 274–292 in T. K. Pratt, C. T. Atkinson, P. C. Banko, J. D. Jacobi, and B. L. Woodworth, editors. Conservation biology of Hawaiian forest birds. Yale University Press, New Haven, Connecticut, USA. Lindstrom, J. 1999. Early development and fitness in birds and mammals. Trends in Ecology & Evolution 14:343–348. Lockwood, J. L., P. Cassey, and T. M. Blackburn. 2005.

33


FREED AND CANN The role of propagule pressure in explaining species invasions. Trends in Ecology and Evolution 20:223–228. Loope, L. L., and T. W. Giambelluca. 1998. Vulnerabiity of island tropical montane cloud forests to climate change, with special reference to east Maui, Hawaii. Climatic Change 39:503–517. MacArthur, R. H. 1972. Geographical ecology: patterns in the distribution of species. Harper and Row, New York, New York, USA. MacArthur, R. H., and R. Levins. 1967. The limiting similarity, convergence, and divergence of coexisting species. American Naturalist 101:377–385. MacMillen, R. E. 1974. Bioenergetics of Hawaiian honeycreepers: the amakihi (Loxops virens) and the anianiau (Loxops parva). Condor 76:62–69. Martin, T. E. 1987. Food as a limit on breeding birds: a life-history perspective. Annual Review of Ecology, Evolution, and Systematics 18:453–487. May, R. M. 1976. Patterns in multi-species communities. Pages 142–162 in R. M. May, editor. Theoretical ecology: principles and applications. W.B. Saunders, Philadelphia, Pennsylvania, USA. McKinney, M. L., and J. L. Lockwood. 2001. Biotic homogenization: a sequential and selective process. Pages 1–17 in M. L. McKinney and J. L. Lockwood, editors. Biotic homogenization. Kluwer Academic, New York, New York, USA. McLellan, A. J., R. Law, and A. H. Fitter. 1997. Response of calcareous grassland plant species to diffuse competition: results from a removal experiment. Journal of Ecology 85:479–490. Medeiros, M. C., and L. A. Freed. 2009. A fledglingmass threshold greatly affects juvenile survival in the Hawaii akepa. Auk 126:319–325. Meszena, G., M. Gyllenberg, L. Pasztor, and J. A. J. Metz. 2006. Competitive exclusion and limiting similarity: a unified theory. Theoretical Population Biology 69:68–87. Mittelbach, G. G. 2012. Community ecology. Sinauer, Sunderland, Massachusetts, USA. Moller, A. P., W. Fieldler, and P. Berthold, editors. 2010. Effects of climate change on birds. Oxford University Press, Oxford, UK. Mooney, H. A., and R. J. Hobbs. 2000. Invasive species in a changing world. Island Press, Washington, D.C., USA. Morin, P. J. 2011. Community ecology. Second edition. Wiley-Blackwell, West Suffix, UK. Mountainspring, S., and J. M. Scott. 1985. Interspecific competition among Hawaiian forest birds. Ecological Monographs 55:219–239. Moyes, C. D., and P. M. Schulte. 2008. Principles of animal physiology. Second edition. Benjamin Cummings, San Francisco, California, USA. Moyle, R. G., C. E. Filardi, C. E. Smith, and J. Diamond. 2009. Explosive Pleistocene diversification and


hemispheric expansion of a ‘‘great speciator.’’ Proceedings of the National Academy of Science USA 106:1863–1868. Murphy, M. E., J. R. King, and J. Lu. 1988. Malnutrition during the postnuptial molt of white-crowned sparrows: feather growth and quality. Canadian Journal of Zoology 66:1403–1413. Naeem, S. 2002. Ecosystem consequences of biodiversity loss: the evoluton of a paradigm. Ecology 83:1537–1552. Naeem, S., L. J. Thompson, S. P. Lawler, J. H. Lawton, and R. M. Woodfin. 1995. Empirical evidence that declining species diversity may alter the performance of terrestrial ecosystems. Philosophical Transactions of the Royal Society B 347:249–262. Negro, J. J., K. L. Bildstein, and D. M. Bird. 1994. Effects of food deprivation and handling stress on fault-bar formation in nestling American kestrels. Ardea 82:263–267. Ogutu-Ohwayo, R. 1990. The decline of the native fishes of lakes Victoria and Kyoga (East Africa) and the impact of introduced species, especially the Nile perch, Lates niloticus, and the Nile tilapia, Oreochromis niloticus. Environmental Biology of Fishes 27:81–96. Olden, J. D. 2006. Biotic homogenization: a new research agenda for conservation biogeography. Journal of Biogeography 37:2027–2037. Olden, J. D., N. LeRoy Poff, M. R. Douglas, M. E. Douglas, and K. D. Fausch. 2004. Ecological and evolutionary consequences of biotic homogenization. Trends in Ecology and Evolution 19:18–24. Parmesan, C. 2006. Ecological and evolutionary responses to climate change. Annual Review of Ecology, Evolution, and Systematics 37:637–669. Parmesan, C., and G. Yohe. 2003. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37–42. Peck, R. W., P. C. Banko, M. Schwarzfeld, M. Euaparadorn, and K. W. Brinck. 2008. Alien dominance of the parasitoid wasp community along an elevation gradient on Hawai’i Island. Biological Invasions 10:1441–1455. Perkins, R. C. L. 1913. Introduction to fauna Hawaiiensis. Cambridge University Press, Cambridge, UK. Pianka, E. R. 1973. The structure of lizard communities. Annual Review of Ecology and Systematics 4:53–74. Pianka, E. R. 1974. Niche overlap and diffuse competition. Proceedings of the National Academy of Science USA 71:2141–2145. Pianka, E. R. 1976. Competition and niche theory. Pages 114–141 in R. M. May, editor. Theoretical ecology: principles and applications. W.B. Saunders, Philadelphia, Pennsylvania, USA. Pianka, E. R. 1980. Guild structure in desert lizards.

34


FREED AND CANN Oikos 35:194–201. Pimm, S. L. 1996. Lessons from a kill. Biodiversity and Conservation 5:1059–1067. Porter, S. D., and D. A. Savignano. 1990. Invasion of polygyne fire ants decimates native ants and disrupts arthropod community. Ecology 71:2095– 2016. Price, R. D., R. A. Hellenthal, R. L. Palma, K. P. Johnson, and D. H. Clayton. 2003. The chewing lice: world checklist and biological overview. Illinois Natural History Survey Special Publication 24. Radosevich, S., J. Holt, and C. Ghersa. 1997. Weed ecology. Second edition. John Wiley & Sons, New York, New York, USA. Ralph, C. J., and B. R. Noon. 1988. Foraging interactions of small Hawaiian forest birds. Acta XIX Congresssus Internationalis Ornithologici 1986 2:1992–1996. Ricciardi, A., R. J. Neves, and J. B. Rasmussen. 1998. Impending extinctions of North American freshwater mussels (Unionoidea) following the zebra mussel (Dreissena polymorpha) invasion. Journal of Animal Ecology 67:613–619. Ricklefs, R. E. 1987. Community diversity: relative roles of local and regional processes. Science 235:167–171. Robson, C. 2005. Birds of Southeast Asia. Princeton University Press, Princeton, New Jersey, USA. Rosenzweig, C., D. Karoly, M. Vicarrelli, P. Neofotis, Q. Wu, G. Casassa, A. Menzel, T. L. Root, N. Estrella, B. Seguin, P. Tryjanowski, C. Liu, S. Rawlins, and A. Imeson. 2008. Attributing physical and biological impacts to anthropogenic climate change. Nature 453:353–358. Ross, S. T. 1991. Mechanisms structuring stream fish assemblages: Are there lessons from introduced species? Environmental Biology of Fishes 30:359– 368. Roughgarden, J. 1976. Resource partitioning among competing species—a coevolutionary approach. Theoretical Population Biology 9:388–424. Saino, N., M. Romano, R. Ambrosini, D. Rubolini, G. Boncoraglio, M. Caprioli, and A. Romano. 2012. Longevity and lifetime reproductive success of barn swallow offspring are predicted by hatching date and phenotypic quality. Journal of Animal Ecology 81:1004–1012. Sakai, A. K., F. W. Allendorf, J. S. Holt, D. M. Lodge, J. Molofsky, K. A. With, S. Baughman, R. J. Cabin, J. E. Cohen, N. C. Ellstrand, D. E. MaCauley, P. O’Neil, I. M. Parker, J. N. Thompson, and S. G. Weller. 2001. The population biology of invasive species. Annual Review of Ecology, Evolution, and Systematics 32:305–335. Sax, D. F., J. J. Stachowicz, J. H. Brown, J. F. Bruno, M. N. Dawson, S. D. Gaines, R. K. Grosberg, et al.


2007. Ecological and evolutionary insights from species invasions. Trends in Ecology and Evolution 22:465–471. Schlaepfer, M. A., M. C. Runge, and P. W. Sherman. 2002. Ecological and evolutionary traps. Trends in Ecology and Evolution 17:474–480. Schlaepfer, M. A., P. W. Sherman, B. Blossey, and M. C. Runge. 2005. Introduced species as evolutionary traps. Ecology Letters 8:241–246. Schoener, T. W. 1974. Resource partitioning in ecological communities. Science 185:27–39. Schoenherr, A. A. 1981. The role of competition in the replacement of native fishes by introduced species. Pages 173–204 in R. J. Naiman and D. L. Soltz, editors. Fishes in North American deserts. John Wiley and Sons, New York, New York, USA. Scott, J. M., S. Mountainspring, F. L. Ramsey, and C. B. Kepler. 1986. Forest bird communities of the Hawaiian Islands: their dynamics, ecology, and conservation. Studies in Avian Biology 9:1–431. Scott, S. N., S. M. Clegg, S. P. Blomberg, J. Kikkawa, and I. P. F. Owens. 2003. Morphological shifts in island-dwelling birds: the roles of generalist foraging and niche expansion. Evolution 57:2147–2156. Shea, K., and P. Chesson. 2002. Community ecology theory as a framework for biological invasions. Trends in Ecology and Evolution 17:171–176. Simberloff, D. 1981. Community effects of introduced species. Pages 53–81 in M. H. Nitecki, editor. Biotic crises in ecological and evolutionary time. Academic Press, New York, New York, USA. Simberloff, D. 1996. Impacts of introduced species in the United States. Consequences 2:1–13. Simberloff, D. 2009. The role of propagule pressure in biological invasions. Annual Review of Ecology, Evolution, and Systematics 40:81–102. Sol, D., I. Bartomeus, and A. S. Griffin. 2012. The paradox of invasion in birds: competitive superiority or ecological opportunism? Oecologica 169:553–564. Swaddle, J. P., and M. S. Witter. 1994. Food, feathers and fluctuating asymmetries. Proceedings of the Royal Society B 255:147–152. Taylor, C. M., and A. Hastings. 2005. Allee effects in biological invasions. Ecology Letters 8:895–908. Terborgh, J., and J. S. Weske. 1975. The role of competition in the distribution of Andean birds. Ecology 56:562–576. Tilman, D. 1999. The ecological consequences of changes in diversity: a search for general principles. Ecology 80:1455–1474. Tilman, D. 2007. Interspecific competition and multispecies coexistence. Pages 84–97 in R. M. May and M. McLean, editors. Theoretical ecology: principles and applications. Third edition. Oxford University Press, Oxford, UK. Tindall, S. D., C. J. Ralph, and M. N. Clout. 2007.

35


FREED AND CANN Changes in bird abundance following common myna control on a New Zealand island. Pacific Conservation Biology 13:202–212. Tobin, P. C., S. L. Whitmire, D. M. Johnson, O. N. Bjornstad, and A. M. Liebhold. 2007. Invasion speed is affected by geographical variation in the strength of allee effects. Ecology Letters 10:36–43. VanderWerf, E. A. 1993. Scales of habitat selection by foraging elepaio in undisturbed and human altered habitat. Condor 95:980–989. van Riper, III, C., S. G. van Riper, M. L. Goff, and M. Laird. 1986. The epizootiology and ecological significance of malaria in Hawaiian land birds. Ecological Monographs 56:327–344. von Holle, B., and D. Simberloff. 2005. Ecological resistance to biological invasion overwhelmed by propagule pressure. Ecology 86:3212–3218. Wagner, W. L., D. R. Herbst, and S. H. Sohmer. 1990. Manual of the flowering plants of Hawai’i. Volume 1. University of Hawaii Press, Honolulu, Hawaii, USA. Welden, C. W., and W. L. Slauson. 1986. The intensity of competition and its importance: an overlooked distinction and some implications. Quarterly Review of Biology 61:23–44. White, G. C., and K. P. Burnham. 1999. Program MARK: survival rate estimation from both live and dead encounters. Bird Study 46 (Supple-


ment):S120–S139. Wilson, E. E., L. M. Mullen, and D. A. Holway. 2009. Life history plasticity magnifies the ecological effects of a social wasp invasion. Proceedings of the National Academy of Sciences USA 106:12809– 12813. Wilson, S. D., and P. A. Keddy. 1986. Species competitive ability and position along a natural stress/disturbance gradient. Ecology 67:1236–1242. Winer, B. J. 1971. Statistical issues of experimental design. Second edition. McGraw-Hill, New York, New York, USA. Woodworth, B. L., J. T. Nelson, E. J. Tweed, S. G. Fancy, M. P. Moore, E. B. Cohen, and M. S. Collins. 2001. Breeding productivity and survival of the endangered Hawaii creeper in wet forest refuge on Mauna Kea, Hawaii. Studies in Avian Biology 22:164–172. Wootton, J. T. 1994. The nature and consequences of indirect effects in ecological communities. Annual Review of Ecology, Evolution, and Systematics 25:443–466. Ziegler, A. C. 2002. Hawaiian natural history, ecology, and evolution. University of Hawaii Press, Honolulu, Hawaii, USA. Zimmerman, E. C. 1978. Insects of Hawaii. Volume 9. Microlepidoptera, Part I. University of Hawaii Press, Honolulu, Hawaii, USA.

36


FREED AND CANN

SUPPLEMENTAL MATERIAL APPENDIX A

Fig. A1. Iiwi illustrating broken feathers and fault bars.

APPENDIX B Competition between Japanese White-eyes and each native species

(see Table 4). Birds are considered individually as species. All species have a 2-egg clutch except for the Akiapolaau (Hemignathus monroi ) which can have a single egg clutch. Hawaii Creeper (Oreomystis mana).—Females of this endangered species, with a single clutch, had greater bill stunting than males and their similarity in bill length with male white-eyes, which tripled in abundance, was reduced from 1.16 to 1.13 (Freed and Cann 2013b), illustrating a role for the theory of limiting similarity (Hutch-

Here we provide evidence that the introduced Japanese White-eye (Zosterops japonicus) competes for food with each species of native passerine bird at Hakalau Forest National Wildlife Refuge on the Island of Hawaii. These are the reasons, beyond overall stunted growth (Freed and Cann 2009) and overall changes in molt (Freed and Cann 2012b), why every native species has negative slopes after the breakpoint v www.esajournals.org

37


FREED AND CANN

inson 1959, MacArthur and Levins 1967). Creeper females, captured after 2002, died after breeding with lower fat and with molt-breeding overlap, but males did not die. This indicates a cost of reproduction from competition with the white-eye, augmented by sexual conflict because male creepers had higher fat. As a result, the adult sex ratio of this endangered bird changed from close to parity to 75% males within five years (Freed and Cann 2013b). This extremely male-biased adult sex ratio occurred in the 1900, 1770, and 1650-m study sites, the latter in the closed forest area. The competition is consistent with that documented over space by Mountainspring and Scott (1985). Indirect effects are a part of competition (Wootton 1994), and there is some evidence of this from our observations. We never saw Hawaii Amakihi foraging for arthropods on branches of ohia-lehua trees, the primary creeper foraging niche, before the white-eye increase. But on several occasions this behavior was noticed during the increase. Competition between the white-eye and amakihi could have generated greater competition between the amakihi and the creeper as an indirect effect, in addition to the direct effect of competition between the whiteeye and creeper (Freed and Cann 2013b). Hawaii Akepa (Loxops coccineus coccineus).—This endangered bird became non-viable during the white-eye increase as k, the natural rate of increase, become significantly less than 1, and half of adults known to be alive in 2005 did not survive to 2006 (Freed et al. 2008a). The change from viable to non-viable population was based in part on lower nesting success and dismantling of two life history adaptations. This endangered bird, with a single clutch, breeds during a deeper seasonal decline in food than other species. Their nestlings have an overgrowth adaptation, weighing 125% as much as adults. They lose mass as fledglings during steeper declines in food while their parents, which are still feeding them, molt primary flight feathers (Freed et al. 2007). Medeiros and Freed (2009) showed that fledglings with high mass survive just as well as their parents, an extraordinary phenomenon in avian life history, which is the selective pressure that generated the adaptation. Competition with white-eyes reduced nestling and fledgling mass, resulting in lower juvenile survival (Freed et al. v www.esajournals.org

2008a). In addition, the akepa has a seasonal variation in sex allocation adaptation. Females nesting during March and April produce mainly sons, while those nesting during May and June produce mainly daughters (Freed et al. 2009). More food exists during March and April (Freed et al. 2007) so sons are better able to grow normal length bills through September; the bill is subject to stabilizing selection in males (Freed et al. 2009). Males have two years of delayed plumage maturation during which they are avoided by older and brighter females (Lepson and Freed 1995). Thus the sex allocation adaptation is to ensure survival of sons for a minimum of three years to when they become attractive to all females. This adaptation has also been dismantled by white-eye competition. We documented that the white-eye capture rate was greater over the four month period during 2000–2005, and was higher during the months of May and June when mainly daughters are being produced compared to March and April during those years (Freed et al. 2009). This was because white-eye captures included both adults and fledglings, whereas during March and April just white-eye adults were captured. The consequence is that the sex ratio of young akepa before 2000 was 57% females, but after 1999 the sex ratio was 13% females (Freed et al. 2009). Years of this malebiased recruitment resulted in a more malebiased adult sex ratio throughout our study sites in the southern portion of the refuge, including the 1650-m site in the closed forest area. Given the lower food available during May and June, combined with increased white-eye captures, the lower nesting success was likely due to failed nests during May and June. The competition between the white-eye and akepa may involve different life history stages of the same insect prey. The akepa has asymmetric jaws that cross, enabling it to pry open ohialehua leaf buds that have microlepidopteran Carposina larvae within. The larvae are not accessible to the white-eye. However, Perkins (1913) observed two native species, the Hawaii Creeper and Hawaii Elepaio, consume the adult moths. If these two species could capture the adult moths, particularly the creeper which does not fly-catch, then white-eyes would have no 38


FREED AND CANN

trouble consuming them. The white-eye overlaps all other foraging substrates of the akepa, and can compete for Carposina larvae with the akepa by consuming the adult moths. The endangered Hawaii Akepa, like the Hawaii Creeper, illustrates the role of the theory of limiting similarity (Hutchinson 1959, MacArthur and Levins 1967) in competition. The ratio of the white-eye bill to the akepa (1.03) is even more similar than the 1.13 with the creeper. It may be the primary reason why these two endangered species have remained vanishingly rare in the northern parts of the refuge. The rarity is no longer a distributional anomaly (Scott et al. 1986). It is a consequence of high white-eye density. The competition between the akepa and whiteeye is not consistent with Mountainspring and Scott (1985) who identified a positive correlation in density with the two species among sites. However, this correlation could be produced by high abundance in good sites and low abundance in lesser quality sites. Our negative correlation is based on a single study site in real time. The dynamic negative correlation between the two species on transects 1–4, spanning a distance of two km, is consistent with our calculation of k and the dismantling of life history adaptations (Freed et al. 2008a). Akiapolaau.—Some native species have foraging specializations that the white-eye is unable to employ, but competition is still likely. The Akiapolaau excavates wood in koa trees to obtain burrowing larvae. While this is a unique behavior, it is also the most energetically expensive foraging behavior of the bird. The white-eye overlaps all other foraging substrates of the Akiapolaau. By competing with the Akiapolaau on these other foraging substrates, the endangered bird has only its most expensive foraging behavior safe from competition. This species is at a disadvantage compared to its former situation where it could easily supplement its diet with other widely available resources. These three endangered species declined solely from cumulative competition from white-eyes (Table 8 in main text). Hawaii Amakihi (Hemignathus virens virens), Apapane (Himatione sanguinea), and Iiwi (Vestiaria coccinea).—These birds forage mainly on ohialehua foliage, the most common foraging subv www.esajournals.org

strate in the forest. A study of duration of foraging bouts with these three species and the white-eye showed that the duration of white-eye bouts was indistinguishable from those of the Apapane and amakihi, and longer than those of the Iiwi (Freed et al. 2008a). All species were captured in aerial nets blocking different heights of the canopy. The white-eye moves more methodically through the foliage than the other species, looking at the upper surface and lower surface of leaves and petioles. This behavior likely results in more prey removal per unit foraging time than the native species. The whiteeye, with the shortest bill in this set of species has been observed in all vertical positions on the tree. It also forages in other substrates and has much higher captures in pole nets than the common native species. Even if the white-eye is most efficient with small prey, it can consume most prey sizes because most prey are still of small size (Freed et al. 2007). As more are removed from ohialehua foliage by the greater number of birds, there would be less overall prey to support the common native species and the additional whiteeye. The significant change in adult sex ratio indicates that females bore a higher cost of reproduction associated with an increase in molt-breeding overlap in the Iiwi and amakihi and strong increase in extended molt in the Apapane (Freed and Cann 2012b). The decline of these common honeycreepers is in contrast to other studies where specialized community members decline before generalized species (Devictor et al. 2008). However, habitat alteration and fragmentation, as the environmental changes, are very different from the change in numbers of a competitor. Those physical changes do not cause this severe food limitation, but rather impact behaviors reflecting dispersal or establishment of territories. Omao (Myadestes obscurus).—This thrush is omnivorous and consumes fruit which is available during much of the year. The white-eye also consumes fruit, particularly small fruit from Cheirodendron trees and soft fruit from the Hawaiian raspberry. However, most fruit is not limiting as much fruit dries on the plant. The white-eye competes with the Omao for insects and spiders in ohia-lehua foliage, just as it does with the Iiwi, Apapane, and amakihi. The Omao 39


FREED AND CANN

diversity of substrates other than ohia-lehua (VanderWerf 1993), and this may be the reason why the white-eye competes least with this native species.

had the greatest change in adult sex ratio of all species, from a higher cost of reproduction associated with the greatest increase in moltbreeding overlap (Freed and Cann 2012b). Hawaii Elepaio (Chasiempis sandwichensis bryani ).—The white-eye does not fly-catch as does this monarchine flycatcher. However, the whiteeye can consume moth and fly larvae, thereby reducing the number of adults available to the elepaio for hawking. The elepaio also uses a


SUPPLEMENT R code for piecewise regression Ecological Archives http://dx.doi.org/10.1890/ES14-00289.1.sm.

40
