Are bluebirds good for farms, and are farms good for bluebirds?

Final Report for GS06-053

Project Type: Graduate Student
Funds awarded in 2006: $10,000.00
Projected End Date: 12/31/2009
Grant Recipient: University of Florida
Region: Southern
State: Florida
Graduate Student:
Major Professor:
Dr. Katie Sieving
Wildlife Ecology / UF
Expand All

Project Information


We investigated reproduction and pest consumption by Eastern Bluebirds (Sialia sialis) on farms and natural open lands. Findings suggest that farms provide suboptimal but useful habitat for breeding bluebirds. Farmland bluebirds laid more eggs and raised more clutches than natural land bluebirds in 2007, yet birds from all treatments produced similar numbers of nestlings. In 2008, farmland bluebirds produced nearly half the number of nestlings in the first clutch than bluebirds on natural lands. Therefore, reproductive success did not differ but reproductive effort was higher on farms. Preliminary results concerning pest consumption suggest that breeding pairs on reduced-impact farms consume 76 to174 arthropods per day when feeding nestlings; the most frequent prey types included grasshoppers, crickets, and larvae (Lepidoptera, Coleoptera).

Tables, figures or graphs mentioned in this report are on file in the Southern SARE office.

Contact Sue Blum at for a hard copy.


A variety of reasons have been discussed supporting efforts to increase birdlife on farms. Most farmers seem to appreciate birds (e.g., Jacobson et al. 2003), government and consumers are pushing for biodiversity-friendly produce, farmlands are increasingly replacing natural habitat, and many birds consume pest insects and could aid in biological control programs. However, scientists cannot responsibly encourage farmers to attract pest-eating songbirds to their lands without first answering two questions with more certainty. What are the realized economic benefits to farmers of encouraging wild birds to inhabit farmlands, and how suitable are farms as habitat for wild bird species? If pest-eating birds have negligible effects on pest populations or crop yield, or farms serve as ecological traps for birds, then the value to growers and society of encouraging bird-friendly agroecosystems should be vigorously questioned.

The purpose of this project is to address aspects of both questions in a farmland-bird system in North-central Florida. We provide information concerning the pest-eating potential of native insectivorous birds in organic (reduced-impact) vegetables, while also describing the reproductive success of these birds on a selection of organic and conventional farms and natural open lands. We were encouraged to conduct this work because organic farmers are facing new biodiversity guidelines from USDA/NOP stipulating that they not only protect, but actually enhance, native species on their farms (Wild Farm Alliance 2005). Therefore, in the trend toward ecological agriculture, organic producers must accommodate increasingly complex ecological interactions in systems that are already more complex than conventional systems. Ideally, management of ecological interactions and biodiversity could also favor production and economic viability, thereby enhancing the efficiency of growers’ operations. With attention to both the social and ecological aspects (and potential pitfalls) of bird-friendly farming, this work focused on providing information that can be used to assess farmland biodiversity-management options involving native bird predators of pests.

Agoecosystems, especially if they are organic or sustainably-managed, can support many native taxa, from soil invertebrates, beneficial foliage insects, bees and other pollinators, lizards, to bats, birds, and other vertebrates (Hole et al. 2005). Proven techniques to foster biodiversity, collectively termed “farmscaping for biodiversity,” are nearly as diverse as the species they attract (Bugg et al. 1998). However, growers dealing with ecological complexity require increasingly simple and low-cost techniques that serve multiple functions for enhancing overall sustainability of their operations (Ferron and Deguine 2005). A technique that attracts wildlife with neutral or negative roles in the agro-system is unlikely to be selected over a technique that brings in species playing primarily beneficial roles. In order to assess the relative value of different species that could be encouraged through habitat management (i.e., to determine which species may provide useful services such as pest control), growers need explicit information about what native species do in their systems.

Part of the underlying rationale for research reported here relates to the need for further assessment of how and when native insectivorous birds can help control invertebrate herbivore populations in agro-ecosystems. Strong (experiment-based) evidence is emerging in support of this idea from various systems, especially those dominated by woody plants. Experiments that exclude birds from access to foliage have documented that birds reduce pest numbers and increase growth of trees (Marquis and Whelan 1995), coffee plants (Greenberg et al. 2000), apple trees (Mols and Visser 2002), and even broccoli (Hooks et al. 2003). Bird suppression of pests and plant damage were measurable and/or significant in these studies. The increasing empirical evidence that birds can help control herbivorous insects is supported by a great deal of theoretical literature suggesting that because birds are generalist insectivores with high mobility, they should play strong positive roles in the mitigation of pest outbreaks (Holling 1988).

In addition to their potential as pest-controllers, birds are also highly desirable forms of “biodiversity” with wide appeal and familiarity to growers and the public (Jacobson et al. 2003). Fostering native birds of conservation importance on farms may therefore count as much or more in growers’ efforts to meet biodiversity guidelines than other taxa. Still, significant misconceptions concerning the relationship between birds and farms (in general) must be overcome. To most people, the goal of attracting birds into agricultural fields with row crops may seem foolish because, either, they have found farms to be wastelands for native birds (Peterjohn 2003) or that birds are serious crop destroyers (Avery et al. 1992). Many conservationists are convinced of the former, and many growers hold the latter perception of birds. However, birds have diverse relationships with farms, crops, and pests; dependent upon the agro-ecosystem and surrounding landscape features (Tavernier and Tolomeo 2002, Jones et al. 2005). Initial research in our lab conducted on vegetable farms in North-central Florida documented that many native birds were using organic and conventional operations, and were likely to be consuming significant numbers of important pests in vegetable fields — without causing crop damage (Jones et al. 2005). In a comparative survey, insect-feeding birds were particularly diverse and abundant in fields with mixed crops, hedges along field borders, and near wooded areas. In an experimental manipulation, pest-consuming bird species were significantly more abundant and spent significantly more time hunting for pest prey in fields where strips (intercrops) of sunflowers were grown among the vegetable crops (Jones and Sieving 2006). Thus, we began this current study with many indications that birds can participate positively in food production in vegetable cropping systems in North-central Florida, and that insectivorous birds could be readily attracted to these systems. But our concerns about whether farmlands were providing high quality habitat for native birds needed to be addressed.

Agricultural lands can be managed for both food and biodiversity production, and there is increasing pressure to do so. The US government requires organic farmers to protect biodiversity (Wild Farm Alliance 2005), and health/environmentally-conscious consumers are creating market demand for produce branded with biodiversity-friendly certifications (e.g., Smithsonian-certified bird-friendly coffee; Grankvist and Beal 2001). Finally, as global agricultural expansion is projected to replace upward of one billion hectares of natural habitat during the next fifty years, conservationists are increasingly eyeing farmlands as critical habitat to augment insufficient areas protected in reserves (Tilman et al. 2001, Fischer et al. 2008). The argument is that if agriculture can provide some use to many species and suitable habitat for a few, then it could serve as an ecological buffer around natural areas to boost their biodiversity-holding capacity (Phillips 2002).

Natural habitats and landscape mosaics that are fragmented by agriculture, however, often do not sustain healthy ecological communities. For example, various studies from different biomes document high rates of nest failure for many interior-forest obligate birds in fragments surrounded by agriculture (Rodewald and Yahner 2001a and b, Albrecht 2004, Knutson et al. 2004, Peak et al. 2004). Fragments suffer higher rates of exotic species invasion and nest predation (Tewksbury et al. 2006) and generally diminished species richness and evenness of native communities (Rodewald and Yahner 2001b). Though farmlands may create population sinks or ecological traps for forest-requiring species (e.g., Wood Thrush [Hylocichla mustelina], Fauth 2001, Zuria et al. 2007), current thinking is that they may provide suitable habitat for species that can tolerate, or that prefer, more open lands maintained for production of food or fiber. Indeed, a central focus in conservation research right now is to characterize the potential values of “wildlife-friendly farms” and farming landscapes more generally as biodiversity buffers (Daily et al. 2001). The most direct measure of habitat quality for wild species is to assess whether habitats support adequate reproductive success and production of viable recruits to breeding populations (Hall et al. 1997).

Avian reproductive success typically appears to be higher on organic and other reduced-impact (e.g., organically managed, sustainable) farms than on conventional farms (e.g., Bouvier et al. 2005, Britschgi et al. 2006, Hart et al. 2006 – but see Graham and Desgranges 1993). While this is not surprising, we still do not know in general how farmlands compare in quality to natural openlands ecosystems as wildlife habitat. The only way to do that is via comparisons of reproduction across the wildland-agriculture gradient of land-use (Hall et al. 1997). Unfortunately, the most common approach to assessing habitat quality has historically been via assessment of animal numbers (population density); where high densities have been taken as evidence of high quality habitat. In our own work, we have shown that bird densities can be enhanced on farms using attractive plantings (Jones and Sieving 2006) but this does not mean that birds attracted to fields are exposed to beneficial or benign conditions from their perspective. Ecological traps have been defined as attractive habitats that bring species into danger from higher predation, pollution, other mortality factors, or that provide substandard food or other critical resources leading to reproductive failure (Schlaepfer et al. 2002). Such traps could easily be created on farms if species are attracted to use on-farm habitats for foraging (e.g., Jones and Sieving 2006) or for reproduction (e.g., Mols and Vissar 2002) but suffer greater mortality or depressed reproduction as a result.

Project Objectives:

Our specific objectives were to:

1) Quantify foraging rates of Eastern Bluebirds on reduced-impact (e.g., organic, organically managed, sustainable) farm fields, including a summary of prey types and numbers taken per feeding visit to nest boxes.

2) Estimate pest availability in reduced-impact vegetable fields near bluebird nest-boxes.

3) Conduct nest-monitoring of bluebirds on reduced-impact farms, conventional farms, and natural open lands, allowing us to compare reproductive success among these land-management categories.


Click linked name(s) to expand
  • Kathryn Sieving


Materials and methods:

The Eastern Bluebird is a charismatic, primarily insectivorous, ground-foraging, secondary cavity-nester of open lands. Females lay multiple broods and are the only parent that incubates, though males feed incubating females. Because of the species’ penchant for using nest-boxes, bluebird reproduction is commonly studied and monitored by ornithologists and lay birders alike. The species was once uncommon in North America, but its populations have grown considerably due to community-based conservation efforts involving nest-box provisioning. However, it remains threatened in parts of its range (Gowaty and Plissner 1998). It is a logical choice as a study-species because 1) it is one of the three most common insectivorous species found on farmlands in North-central Florida (Jones et al. 2005), 2) it specializes on arthropods throughout the entire year, and 3) it takes readily to nest-boxes that can be strategically placed for management or research purposes (Gowaty and Plissner 1998).

We conducted this research in North-central Florida (Alachua and Putnam Counties) during the breeding seasons of 2007 and 2008. We erected nest-boxes for bluebirds on six conventional farms, eight reduced-impact farms, and at four natural open-land control areas (restored and semi-restored native grasslands) within the Ordway-Swisher Biological Station (> 3760 ha in area). Bluebird nest boxes have been provided and monitored in this protected area since the mid 1990s with sustained high levels of nest box occupancy and reproductive success (KES, unpubl. data). Most reduced-impact farms were USDA-certified organic operations or were managed in line with organic standards (but without certification). We asked growers about their use of synthetic pesticides before the beginning of field work. Three of eight reduced-impact farmers used synthetic pesticides (only one used insecticides), but they did so sparingly, only on certain crops at certain times. All participating conventional farmers applied insecticides and other pesticides on the entirety of their crops in multiple applications (for beans, corn, strawberries, tobacco, and/or melons, depending on the farm; we did not ask farmers the exact types or amounts of pesticides that they used). In order to control for predation, we mounted nest-boxes on narrow, metal poles (approximately 1.5 m high). We greased the poles when birds laid eggs and maintained these grease barriers throughout nesting (USDA Natural Resources Conservation Service 1999). If predators overcame this deterrence, then signs left in the grease or on the boxes allowed detection of predation events. We eliminated data for predated nests from statistical analyses as appropriate.

DATA COLLECTION: We originally sought to quantify pest consumption by monitoring bluebirds and other open-land insectivorous species with spotting scopes, but this proved inefficient for various reasons. We therefore purchased four motion-triggered all-weather “game” cameras and set them by nest-box entrances to record visits by adult bluebirds to provision their nestlings during the reproductive season. Cameras were rotated among different active nests on each of four farms and the natural area between March and July of 2008.

The digital cameras trigger when birds fly in front of them, and they store up to a week or more of photos (up to several thousand). Unfortunately, the cameras also trigger when trees or crops rustle in the wind, and with human traffic. Each camera generates several thousand photos over several days, and the process of separating out useful images and recording relevant data from each of them is slow. We are using volunteers (students) to separate the photos and record data from each one, and the process will take several months more. Once we have completed this task, we will quantify the number of parental visits per hour to each nest during a standardized sampling period per day. Cameras function best (i.e., reliably take photos when birds visit the nest) during the morning hours, before the cameras overheat in the Florida summer sun. We are therefore sampling photographs during morning hours only (up until 1100 hours). We will quantify the number of male visits, female visits, and total parental visits to the nest. We will assume that every visit to the nest represents one arthropod delivered to nestlings. If a photo was taken within 30s of another photo, then we will assume that the two photos represent one visit. If both parents are present in a photo, we will count that as two visits. When females are incubating eggs, we will only count male visits (as males provision incubating females). When nestlings are present, we will count both male and female visits.

Additionally, we observed foraging bluebirds for a few hours at each farm to estimate the percentage of foraging bouts that birds spent on versus off of the farms where their nests were located.

STATISTICAL ANALYSIS: We will calculate the average visits per hour per nest box monitored per day, then multiply the hourly mean by 12 (representing the approximately 12 hours of daylight when bluebirds are actively foraging; Gowaty and Plissner 1998) to calculate estimated daily numbers of visits in each day of sampling at each nest box. Then we can estimate/calculate the total amount of arthropods delivered to all active nests per farm by utilizing these daily means (which can change over the course of a nesting attempt – with more frequent visits as chicks grow older), the numbers of breeding pairs per field/farm, and the percentage of foraging bouts on versus off the farm fields. Previous researchers of Eastern Bluebird foraging behavior have estimated that bluebirds deliver 55% of the prey that they capture to the nest, leaving 45% for adult consumption (Goldman 1975). We will then multiply the estimated daily total amount of arthropods delivered to the nest by 145% to account for parental needs in the daily consumption of arthropods per family of bluebirds. We will present these descriptive statistics for every family of bluebirds involved in the study.

As we analyze photos, we are recording the number of each different type of prey item delivered to the nest. We will thereby compile a list of orders of prey items taken (e.g., grasshoppers/crickets, caterpillars, spiders) and the relative frequency that each type of prey item is delivered to the nest.

DATA COLLECTION: We sampled arthropod populations on an approximately weekly basis between February 17th (12 days before the first clutch of the season was laid) and July 17th. We conducted two types of surveys: “Grasshopper Walk” (GW) and “Walk-Brush” (WB) surveys. The purpose of GW surveys was to count larger, flying arthropods, especially grasshoppers and crickets. GW transects were positioned in microhabitat with the tallest herbaceous vegetation available during a given visit to a study site. We conducted GW surveys adjacent to fields with actively growing crops, in field edges or fallow fields with weeds. We conducted GW surveys by walking at 1 pace per second for 10 m while recording all arthropods that moved or were readily visible within approximately 1 m on either side of the observer, avoiding double-counting.

The purpose of WB surveys was to target smaller, ground- and leaf-dwelling arthropods. WB surveys were conducted between crop rows (principally vegetables such as kale, cabbage, broccoli, collard greens, etc.). We conducted the first 10 m of WB surveys in the same manner as GW surveys. During the second 10 m, we bent low, brushed vegetation with our hands, and looked under leaves and plants, in the middle of the vegetation layer, and on top of vegetation, within 1 m to either side of the transect line. Time to complete brushing varied depending on arthropod abundance (our aim was to conduct a thorough search), as the time it took to count and classify arthropods and record data increased as we encountered more arthropods.

We developed these sampling methods based on Gardiner et al. (2005) in order to sample prey that were frequently taken from foraging microhabitats commonly used by bluebirds nesting on the study sites (JJD unpublished data). Adult Orthoptera (i.e., grasshoppers and crickets) and Lepidoptera larvae (i.e., caterpillars) were the principal prey items identified in previous observations (though many smaller prey could not be identified), and bluebirds frequently foraged in crops, fallow fields, and field edges within 100m of the nesting site, but also flew further away (to unobservable locations off site; JJD unpublished data). Previous research confirms these patterns (Gowaty and Plissner 1998).

By conducting WB transects between crop rows and GW surveys in fallow vegetation surrounding farm fields (and in the most similar microhabitat available in natural control areas) we assessed representative food resources available to bluebirds. After bluebirds began nesting, we conducted prey surveys within 200 m of nests in each site in appropriate foraging habitats. To maintain independence of samples, observers avoided using any transect more than once. We selected transect locations subjectively on each visit to sites, by looking for those areas with vegetation that was most dense and/or actively growing (greenest), assuming that such vegetation would attract the kind of herbivorous prey that bluebirds target. We conducted at least two (up to 8) of each type of transect (GW and WB) at each visit to a site – the number of transects per site was proportional to the number of boxes clustered at each site (box number varied from 3 to 10 per site).

We only recorded arthropods that were greater than 0.5 cm in length. We classified arthropods into 6 size categories (0.5-1 cm, 1-2 cm, 2-3 cm, 4-5 cm, 5-6 cm, and 6 or more cm), and 13 identification groups (Orthoptera, Lepidoptera, spiders, flies, bees/wasps, ants, true bugs, adult non-ladybug beetles, Lepidoptera and beetle larvae, adult ladybug beetles, dragonflies, roaches, and unidentifiable arthropods). Five different observers conducted arthropod surveys. We monitored cloud cover, temperature, and wind speed during sample periods to ensure some degree of standardization of conditions. We conducted surveys between 0800 and 1630 hours, EST. We did not conduct surveys during rain events or when winds reached more than 15 km/hr. We conducted surveys in exposed (unshaded) microhabitats with full insulation and avoided the coldest periods during the day (early morning).

STATISTICAL ANALYSIS: We calculated the percentage of times each arthropod group was present during GW and WB surveys, and the mean abundance of each arthropod group per survey type. Once all photographic data are ready for analysis, we plan to compare the consumption of different arthropod groups by bluebirds to their relative abundance/availability as determined by arthropod surveys. In this manner, we will determine if bluebirds are targeting particular arthropod groups or if their consumption is dictated by what arthropod groups are most abundant.

DATA COLLECTION: In 2007, we monitored nest-boxes approximately every four days throughout the entire breeding season (February-August), when we recorded how many eggs and/or nestlings were in the nest, and the estimated age of nestlings (using the criteria of Gowaty and Plissner 1998). In 2008, we limited our land-management treatments to reduced-impact farms and natural control areas. We monitored nest-boxes approximately every four days during the early breeding season (when the first clutches were laid and hatched), allowing us to estimate egg and hatchling production for first clutches only. (In 2008, the price of fuel doubled and we were unable to complete a full season of field work with funds allocated. We extended the period in between visits up to three weeks for later clutches to accommodate, but this prevented us from estimating the number of eggs or hatchlings produced beyond the first clutch.) In 2008, we were, however, able to estimate the number of clutches produced with confidence, because the nesting cycle (from egg-laying to fledging) of Eastern Bluebirds lasts approximately 28 days and our visit frequency was sufficient to detect clutching events, and we calculated first-egg-date (i.e., the date that a bluebird female laid the first egg of her first clutch) from these data. Earlier first-egg-dates often correlate with higher clutch production over the breeding season in passerine birds, as birds have more time to lay more clutches (Elmberg et al. 2005, Nemeckova et al. 2008, Verhulst and Nilsson 2008).

In 2007, we measured nestling body condition and growth. We recorded a nestling’s weight and left-tarsus (the tarsus is the exposed portion of the leg) length at each visit. We recorded wing length just before fledging. We calculated a nestling Body Condition Index (BCI) and pre-fledging body-growth index (BGI) from these data. Higher values for both of these indices correspond to a greater chance of offspring survival (Jakob et al. 1996, Navara et al. 2005). We did not measure nestling body condition or growth in 2008.

STATISTICAL ANALYSIS: Analyses were conducted using general linear and generalized linear mixed models in SAS 9.1 (using α=0.05). In 2007, we compared first-egg-date, egg and nestling production over the entire breeding season, and nestling body condition and growth among reduced-impact farms, conventional farms, and natural control areas. In 2008, we limited comparisons to first-egg-date, overall clutch production, first-clutch egg production, first-clutch hatchling production, and first-clutch hatching success between reduced-impact farms and natural control areas.

Research results and discussion:

For this report we summarized photographic data taken from two breeding pairs of bluebirds feeding nestlings at one of the study-farms. One pair was photographed from April 26th-29th, and the other was photographed from July 4th-15th. Both nests were in the nestling-stage of the nesting cycle (i.e., eggs were hatched and parents were feeding nestlings). We estimated that the April family of bluebirds delivered 6-15 arthropods per hour to their nest, or consumed 104-261 arthropods per day when accounting for parental consumption (174 on average, with a standard error of 34.1; Figure 1 in Appendix; see Methods for details on estimation procedures). We estimated that the July family delivered 1-9 arthropods per hour to their nest, or consumed 17-156 arthropods per day when accounting for parental consumption (76 on average, with a standard error of 12.7; Figure 1 in Appendix). The only prey items captured in photos were crawlers (i.e., Lepidoptera and beetle larvae) and Orthoptera (i.e., grasshoppers and crickets).

The only prevalent groups present in GW surveys were Orthoptera (present in 91% of surveys), adult Lepidoptera (36%), bees (21%), unidentifiable arthropods (10%), flies (5%), and dragonflies (4%). Most arthropod groups were detected in at least 10% of WB surveys: Orthoptera were present in 69% of surveys, spiders in 55%, flies in 54%, bees in 45%, adult Lepidoptera in 36%, non-ladybug adult beetles in 31%, true bugs in 26%, unidentifiable arthropods in 26%, roaches in 17%, crawlers in 12%, ladybugs in 9%, ants in 9%, and dragonflies in 3%.

For GW surveys, the mean number of Orthoptera encountered per survey was 8.09. The mean number of adult Lepidoptera/survey was 0.82, bees/survey was 0.39, unidentifiable arthropods/survey was 0.15, flies/survey was 0.08, and dragonflies/survey was 0.05. For WB surveys, the mean number of Orthoptera encountered per survey was 8.01, crawlers/survey was 3.13, flies/survey was 1.92, spiders/survey was 1.69, bees/survey was 1.24, unidentifiable arthropods/survey was 0.88, adult Lepidoptera/survey was 0.83, true bugs/survey was 0.59, roaches/survey was 0.56, non-ladybug adult beetles/survey was 0.19, ladybugs/survey was 0.19, ants/survey was 0.11, and dragonflies/survey was 0.037.

When the photographic data are ready for analysis, we aim to compare bluebird consumption of arthropod groups by bluebirds to the relative abundance/availability of those groups as determined by arthropod surveys. We will thereby determine if bluebirds are targeting particular arthropod groups or if their feeding preferences simply correlate with the arthropod groups that are most abundant.

2007 RESULTS: First-egg-dates were earlier on reduced-impact and conventional farms than natural control areas (Figure 3). First-egg-date and land management were both related to the total production of eggs during the entire breeding season. Earlier first-egg-dates correlated with bluebirds that produced more eggs (Figure 3), and both farm treatments hosted birds that laid more eggs than in the natural control area (Figure 3). However, total nestling production over the breeding season was not different among land management treatments, despite the fact that farmland birds laid more eggs than birds from natural control areas.

Although not statistically significant, the production of 3 clutches by many farmland bluebirds versus 2 by those on natural sites is clearly a biologically significant difference in clutch production (Figure 3), as the production of third clutches requires a greater expenditure of time and energy for egg production and incubation. Many pairs of bluebirds went on to lay a third clutch on reduced-impact and conventional farms (Figure 3). On the contrary, only one pair of bluebirds in the natural control area produced a third clutch. There were no significant effects of land management on nestling body condition or growth.

2008 RESULTS: As previously mentioned, 2008 land management treatments were limited to reduced-impact farms and natural control areas. Confirming 2007 results, first-egg-dates were significantly earlier on farms than in natural control areas (Figure 3). However, first-egg-dates were generally earlier in 2008 than in 2007 (Figure 3). This may have led to natural-lands bluebirds producing a similar number of clutches as those on reduced-impact farms (Figure 3) simply because they nested early enough in 2008 to allow for third clutches. First-clutch egg production was similar for both land management treatments, but first-clutch hatchling production and hatching success were significantly lower on farms than natural control areas (Figure 3); this is in contrast to such comparisons for 2007, when reproductive success measures such as these were not different between farms and natural areas.

Participation Summary

Educational & Outreach Activities

Participation Summary

Education/outreach description:

DeLuca, J. J. 2008. Reproduction of Eastern Bluebirds (Sialia sialis) in relation to farmland management and food resources in North-central Florida (M.Sc. Thesis; hard copy sent to Southern SARE office). University of Florida, Gainesville, FL.

DeLuca, J.J., and K.E. Sieving. Manuscript in preparation. Reproduction of Eastern Bluebirds (Sialia sialis) in relation to farmland management in North-central Florida.

DeLuca, J.J., and K.E. Sieving. Manuscript in preparation. The pest-eating potential of Eastern Bluebirds (Sialia sialis) in North-central Florida.

Sieving, K.E., and J.J. DeLuca. Various extension publications in preparation.
(1) Pest-eating benefits of fostering bluebirds on vegetable farms.
(2) On-farm and near-farm habitat management to foster pest-consumption by native birds.
(3) Nest box management techniques for farmland birds that eat pests.

Planned (contingent on future funding):
A website about birds and farms, utilizing work done in the Sieving-lab as a guide to strategic assessment of the potential for birds to participate positively (with on- and off-farm management) in integrated pest control

Project Outcomes

Project outcomes:

Preliminary analysis of photographic data indicated that breeding bluebirds that are feeding young consume substantial numbers of insects per day, with average consumption levels estimated at 76-124 prey items per day. Preliminary analysis also indicated that crawlers (i.e., Lepidoptera and beetle larvae) and Orthoptera (i.e., grasshoppers and crickets) were the preferred prey items of bluebirds. This suggests that bluebirds may be eating grasshoppers and crickets in proportion to their presence and abundance in the surrounding farmland habitat (grasshoppers and crickets were by far the most commonly present and abundant group identified during GW and WB arthropod surveys). Although crawlers had a relatively high mean abundance of approximately 3 individuals per WB survey, they were only present in approximately 12% of surveys. In other words, crawlers were not often present, but when they were present, they occurred in very large numbers (suggesting that crawlers appear mostly during pest outbreaks). Our limited photographic analysis indicates that even though crawlers may not be especially present at any given strip of crop rows, they are a valued and targeted part of the bluebird diet. These preliminary data suggest that if our final analyses confirm the initial patterns, then farmers may indeed benefit in measurable ways by having bluebirds nesting on their farms. Initial patterns suggest bluebirds seem to prefer eating insect groups that tend to be among the most important of crop pests (i.e, Orthoptera and crawlers). We look forward to presenting the full analysis of these data in both scientific manuscripts and extension publications for communication to both ecologists and farmers.

In targeting farmland habitat for biodiversity conservation, policy-makers and natural resource management professionals must choose between two distinct types of farm management, “land-sparing” or “wildlife-friendly” operations. Land-sparing operations entail uniformly planted agricultural lands (i.e., monocultures) that are managed to increase yields through high-intensity inputs of fertilizer and pesticide, while maintaining separate natural reserves for biodiversity conservation. “Wildlife-friendly” operations integrate conservation and farming within more diverse landscapes over a larger area, without necessarily protecting separate reserves (Fischer et al. 2008).

Conservation biologists have recently expressed support for the latter method of land management (e.g., Green et al. 2005, Fischer et al. 2008). However, ecologists are only beginning to examine the potential of farmlands as productive habitat for breeding birds and other wildlife, and results remain inconclusive (e.g., Sekercioglu et al. 2007). Our research indicates that reduced-impact farms (which could just as well be considered “wildlife-friendly”) do not necessarily provide ideal habitat for farmland wildlife, in contrast to popular notions (Green et al. 2005, Hole et al. 2005, Fischer et al. 2008). Bluebirds on reduced-impact farms reproduced with much less efficiency (i.e., pairs laid more clutches and raised more broods but produced the same number of young) than bluebirds in natural control areas in 2007. Differences were also distinct in the subsequent year (2008); bluebirds produced approximately twice as many first-clutch hatchlings in natural control areas as on farms (a pattern that, if persistent throughout the entire breeding season, would result in lower production of young on farms). In striking contrast to most previous research (e.g., Bouvier et al. 2005, Britschgi et al. 2006, Hart et al. 2006), we did not detect substantial advantages of reduced-impact over conventional farms as avian breeding habitat, in terms of nestling production, condition, or growth.

While data presented here seem to indicate that farms did not provide breeding habitat as productive as natural areas, farmland bluebirds still produced substantial numbers of offspring. In 2007, farmland birds exhibited similar net reproductive success as bluebirds in natural areas, but signs from the first clutches of 2008 indicated that reproductive success was significantly lower on farms in that year. Another pattern that was different between the two years was that in 2007, farmland bluebirds worked harder to produce the same number of young (more clutches, same number of fledglings/season). Indications for 2008 were that they did not raise more clutches than birds on natural areas, but very likely produced fewer young over the season. Using these data, we will model the potential differences in reproductive success and overall reproductive effort in relation to annual variations in climate (e.g., with Population Viability Analysis). But what we can say at this point is that while farmlands may not be the best breeding habitat for bluebirds in every year, production of young by farmland birds was significant in the two years of our study. As we continue to lose natural habitat to development, farmlands could serve to augment bluebird production on increasingly limited natural areas. And in years when conditions are right, we believe they may be just as productive of bluebirds as the best natural areas. As we explore these relationships more fully, we expect to determine various management techniques that could enhance bluebird reproduction and foraging success on farms, thereby insuring a better partnership between growers and native insectivorous birds. More research is required to determine whether farmlands are on average (given large annual variations in rainfall and crop activities) source (good) or sink (bad) habitats for native insectivorous birds.

Economic Analysis

It costs very little to encourage the presence of insectivorous cavity-nesters on farms. Nest-boxes cost between $15 and $30 from a supplier, and even less if you construct them yourself (see for both suppliers and do-it-yourself instructions). Farmers may optionally grease poles or attach baffles and other predator-deterring devices to nests to encourage higher reproductive success; none of these items cost more than $15. Farmers may diversify their community of avian insectivores by attracting Purple Martin with colonial nest-boxes (see for details), and many other habitat management techniques to foster insectivorous birds are suggested by the literature (see Jones et al. 2005, Jones and Sieving 2006).

Farmer Adoption

Over half of the farmers involved in this research have personally erected nest-boxes suitable for Eastern Bluebirds on their properties. Most farmers involved have shown enthusiasm for the aesthetic beauty of bluebirds on their property, and many believe that they substantially help control pest insect populations.

With further encouragement, through workshops, websites, and extension publications, there is a lot more that can be done to both explore and implement better management of birds for pest-control and conservation on sustainably-managed farms within the wildland-urban gradient.


Areas needing additional study


Population source habitat occurs where high reproductive success results in a population surplus (i.e., excess reproduction prevents any deleterious effects of mortality on population size). Surplus individuals from source habitats emigrate to sink habitats, where reproduction and survivorship are lower than mortality (Brawn and Robinson 1996, Van Dyke 2003). If most bluebird fledglings survived to adulthood on farms, then farms in our study would have not been sink habitats. However, we did not monitor juvenile or adult survival in our study. Future research could measure reproductive output, survivorship, and mortality, thereby verifying if farms serve as source or sink habitat from one year to the next.

The assessment of whether or not farms serve as ecological traps also requires detailed population-level data. An ecological trap is defined as “an environment that has been altered suddenly by human activities, [where] an organism makes a maladaptive habitat choice based on formerly reliable environmental cues, despite the availability of higher quality habitat” (Schlaepfer et al. 2002). In order for farms to be considered ecological traps, bluebirds would have to fare worse on farms than alternative habitat (e.g., natural areas), while simultaneously preferring farms over alternative habitat. Farms certainly are environments that have been altered by human activities, and our results suggest that farms are suboptimal habitat in comparison to natural areas (see above). However, we cannot state that bluebirds make choices in favor of farmland habitat over natural habitat (or vice-versa) because we did not measure habitat preference. Future research could determine if farms act as ecological traps by monitoring bluebird reproductive success, survivorship, and mortality on replicate plots of natural open-lands adjacent to farmlands while quantifying competition levels among bluebirds in both of these treatments (thereby quantifying habitat preference).

Human disturbance is a potentially important factor in determining the suitability of farmland habitat for wildlife conservation, as it has been shown to inversely correlate with avian reproductive success (Yasue and Dearden 2006, Kight and Swaddle 2007). Farmer activity rarely subsides during the growing season – this seems especially true on reduced-impact farms, where a variety of crops are grown and harvested at different times, requiring near-constant activity by farmers (JJD personal observation). Such disturbance may adversely affect bluebird populations, as bluebird reproductive success has been shown to decrease with an increase in human disturbance (Kight and Swaddle 2007). The effects of human disturbance on farmland wildlife merits further research, as it will aid in the evaluation of the conservation potential of farms.

Spatial scale needs to be considered if research is to obtain comprehensive understanding of the complexities involved with sustaining wild birds and other vertebrates on agricultural lands (Hole et al. 2005). We conducted our research on few farms in North-central Florida, and all of our natural open-land sites were located within one large protected reserve (raising issues of spatial autocorrelation). Furthermore, the Eastern Bluebird is a highly mobile species that is fairly broad in its breeding habitat selection, including habitats not sampled here (e.g., orchards, graveyards, golf courses; Gowaty and Plissner 1998). Research conducted over a greater geographic range, with more replicates, and across a diversity of habitats would further strengthen our ability to assess the value of farms as habitat.

Finally, many native species adapted to open landscapes that inhabit farmlands are of greater conservation concern than bluebirds, and they deserve the attention of future research (e.g., Loggerhead Shrike [Lanius ludovicianus; Yosef 1996], Common Ground-Dove [Columbina passerina; Bowman 2002], Northern Bobwhite [Colinus virginianus; Brennan 1999]). A multi-species approach across a variety of open-land habitats would also improve ecosystem management. It would increase explanatory power while more accurately determining the landscape attributes and appropriate management schemes required for conserving birds of open lands (Lambeck 1997). Furthermore, agro-ecosystem designs that encompass biodiversity protection are among the most challenging and relevant problems in conservation biology (Daily et al. 2001; Hole et al. 2005).

Our research suggests that farms provide suboptimal breeding habitat in comparison to natural areas, yet farmland bluebirds still produced substantial numbers of offspring (see above). By measuring reproductive success alongside survivorship and habitat preference, future research could determine if farms act as habitat sources, sinks, or ecological traps (Brawn and Robinson 1996, Schlaepfer et al. 2002, Van Dyke 2003). Ecologists could further elaborate the effects of farmlands on reproduction, survivorship, and mortality by monitoring predation and human disturbance among farms and natural open lands. Lastly, future research could increase applicability and generality by focusing on open-land communities of priority conservation concern. Without such data, we cannot yet conclude that “wildlife-friendly” farms are truly wildlife-friendly. However, this research is the first to compare avian reproductive success among reduced-impact farms, conventional farms, and natural areas (to our knowledge; Hole et al. 2005, Scherr and McNeely 2008, Watson et al. 2008), and our results certainly illuminate fruitful paths for future research.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.