Our research in this topical area has been oriented largely toward defining the kinds of habitats used by different members of the forest avifauna. Secondly, our research examines the kinds of habitats that are created by prevalent human activities, to answer the question: do these land uses provide habitat capable of supporting various members of the forest-adapted fauna?

The faunal groups we have studied in this regard are forest raptors, resident forest-dwelling non-raptorial birds (a large segment of the bird community), migrant, North American-breeding songbirds, and to a lesser extent, bats.

We examined bird habitat affinities in natural forest types, bird use of different ages of second-growth resulting from shifting agriculture, bat community composition in forest fragments in the farming landscape, and the effects of selective mahogany logging on forest birds.

In the typical slash-and-burn farming landscape, how many forest-dwelling bird species are able to persist? Of the total bat fauna of the intact forest, what species are able to use forest fragments remaining in the farming landscape?

What policy should conservationists take toward slash-and-burn farming? Is it possible to make this farming style more compatible with conservation goals? Or is it necessary to move away from this style of farming in order to save tropical forests?

What are the human dimensions of this question? How can the lives of poor farmers be improved while at the same time slowing the rate of deforestation caused by farming?

The above are some of the questions that informed our choice of research topics.

One aspect of our research was a large set of point counts we conducted in order to compare the raptor assemblage of Tikal's pristine forest with that of the slash-and-burn farming landscape nearby. This farming landscape retained half of its mature forest cover, while half the land surface had been converted to small, hand-worked corn fields, fallow fields of regenerating second-growth, cattle pastures, and remnant patches of tropical forest, often degraded by occasional fire and various human uses.

Based on these point counts, we characterized 29 of Tikal's raptor species with respect to the habitats of their most frequent occurrence along this gradient from pristine to partly deforested landscape. Fifteen species proved to be strongly associated with mature forest, seven species were common in both the farming landscape and in pristine forest, four were ubiquitous but most abundant in the farming landscape, and three were detected only in the farming landscape. Hence about half the species appeared quite reliant on mature forest, a quarter appeared capable of using partly deforested landscapes, and a quarter were common in the partly deforested landscape.

Most of our detailed studies of raptor species' biology were conducted in Tikal's mature forests. However, for two species we compared their ecology in the primary forest with that in the slash-and-burn farming landscape.

Margaret Parker (1997) studied the Laughing Falcon (Herpetotheres cachinnans) and Theresa Panasci (1995) studied the Roadside Hawk (Buteo magnirostris), comparing these species' ecology in primary forest and the farming landscape.

While Laughing Falcons nested in both habitats, they had higher nesting densities and apparently smaller home ranges in the farming landscape than in primary forest. These falcons fed exclusively on snakes in the primary forest, but broadened their diet in the farming landscape to include lizards, a few small mammals, and at one eyrie, even fish. Whether the falcons fared better or worse in terms of population viability in the farming landscape compared to the forest remains unknown.

Roadside Hawks nested closer together and reached 20% higher nesting densities in the farming landscape than in primary forest, but were also common in primary forest (Panasci 1995, in press, Panasci and Whitacre 2000). In the primary forest, these hawks were very selective with regard to nesting habitat. They nested in swales, in trees that protruded over the relatively low-canopied forest that dominated in these topographic positions. Such emergent trees were presumably less subject to nest predation by climbing predators than were trees joined to the forest canopy.

Incidence of non-nesting by territorial Roadside Hawk pairs was higher (50%) in mature forest than in the farming landscape (20%). Nest success was 0.17 in mature forest and 0.30 in the farming landscape, and productivity (fledglings per territorial pair) was 0.08 in mature forest and 0.32 in the farming landscape. Prey delivery rates did not differ between the two habitats (Panasci, in press).

Why Roadside Hawks in the farming landscape performed better in these regards than hawks in the primary forest is unknown. This outcome may be related, however, to the more species-rich raptor community in primary forest relative to the farming landscape. Greater raptor species richness may exert more controlling influence on Roadside Hawks in the forest, either through nest-site competition, dietary overlap, or both, relative to the situation in the farming landscape, where the raptor community is much impoverished in species richness (Panasci and Whitacre, in press).

As both of these species are well known to thrive in human-modified landscapes, the above research does not address what might be regarded as the "burning questions" for maintenance of intact raptor communities in human-modified Neotropical forests--the fate of the large, space-demanding species such as the Harpy Eagle (Harpia harpyja) and Crested Eagle (Morphnus guianensis), and of species believed to be mature forest obligates, whatever their body size and spatial needs. Detailed ecological studies of more forest-restricted species in landscapes exhibiting varying degrees of deforestation and fragmentation would be valuable.

We conducted a large set of point counts in the farming landscape near Tikal, and in a variety of natural forest types (Whitacre et al. 1995). This enabled us to assign 90 bird species to eight tentative "habitat response groupings." Species ranged from those that appeared strictly associated with tall, mature, upland forest, through those occurring in many forest types and in older second-growth, to those commonest in young second-growth and other disturbed habitats. In addition, some species were found mainly in mature forest, but in low, "scrub swamp" forest rather than in tall, upland forest. All in all, the analysis suggested that roughly 40-50 percent of the forest bird species at Tikal are closely associated with mature forest. We predict that many of these species will decline seriously or disappear from landscapes subject to heavy deforestation.

From the standpoint of agricultural and conservation policy, the following question is relevant: can conservation gains be made by lengthening the fallow period? That is, if farmers could leave an abandoned corn field for 10 or 15 years before again farming it--rather than 3-7 years as is now the norm--would that provide desirable results for conservation?

We approached this question by examining 46 bird species which our data showed to be essentially absent from the youngest (2-4 m tall) second-growth we studied. For these 46 species, we examined whether they appeared in good numbers in older second-growth, and if so, in what age of second-growth they did so.

Eight species first appeared in significant numbers in 4-6 m tall second-growth, eight species did so in 6-10 m second-growth, and six species did so in second-growth taller than 10 m (30 year-old second-growth). Seventeen of the 46 species (37%) did not reach appreciable numbers even in 30-year old second-growth, and an additional seven species (52% total) showed only a minor increase in abundance in older stands of second-growth.

These results, obtained over a few-month period, admittedly do not provide a thorough analysis of the year-round or multi-year patterns of bird response to habitats in our study area. Nonetheless, these results provide an initial indication that second-growth must reach a substantial age (10 to 30 years) before it begins to have significant value for many forest-dependent bird species, and many such species do not make much use of even 30 year-old second-growth.

Based on the above, we made the following policy conclusion. In terms of forest bird conservation, little is liable to be achieved by a modest lengthening of fallow periods; such a modest lengthening does not make these habitats more usable by many forest-reliant species. By the same token, demand for farmland in many tropical nations today--due to intense human population pressure, as well as other factors--is such that it is very unlikely that fallow periods could be lengthened at any rate. Hence, while we initially considered it possible that one viable strategy toward shifting cultivation would be to search for ways to speed succession of fallowed fields, we have concluded that this strategy is unlikely to benefit many mature forest-dependent species, and is unlikely to be practical in any event. This chain of logic led, in part, to our conclusion that the best policy toward shifting cultivation is this: to "sacrifice" some land essentially permanently to crop production, but utilizing methods that will make it unnecessary to continually rotate and cut down additional forest. By intensifying production on some subset of the land already involved in the rotational farming cycle, we should be able to both spare the need to cut additional mature forest, and allow a substantial amount of existing second-growth to recover old-growth characteristics in time.

Many small songbird species that nest in the U.S. and Canada migrate to Latin American and the Caribbean each year, sometimes spending more of the year in these tropical homes than they do on their breeding grounds. This annual migration of millions of tiny birds, often across the Gulf of México or in a long arc over the Atlantic, is truly one of the wonders of the natural world, but it is a threatened phenomenon.

Several of these migrant species appear to be declining in number. While reasons for these declines are not fully understood, they probably involve factors on both the breeding and wintering grounds.

In the north, these birds often face increasingly fragmented forest habitats in which they are exposed to high rates of nest predation by generalist predators such as Blue Jays (Cyanocitta cristata), raccoons (Procyon lotor), and skunks (Mephitis mephitis), as well as high rates of nest parasitism by the Brown-headed Cowbird (Molothrus ater), which penetrates into the forest fragments from the surrounding open country, laying its eggs in the nests of many host species, thereby depressing the reproductive rate of these hosts.

On their tropical wintering grounds, some of these species are probably being adversely affected by habitat modification by farming and other human activities. While many migrants thrive during the winter in young successional forest, plantations, and even urban shade trees, others are highly reliant on mature forest.

Prior to our project, little research had been conducted on the habitat needs of wintering migrant songbirds in our project area, the extensive forests of Petén, Guatemala. Because these forests are so extensive, they provide a winter home to literally millions of migrant songbirds.

Our studies of wintering migrants at Tikal focused on assessing the extent to which several migrant species used different habitats. We approached this topic mainly by systematic mist-netting, and used capture rates per unit sampling effort as an index of abundance in these habitats. We sampled the range of mature forest types occurring at Tikal--from tall, complex, upland forest on well-drained sites, to low, densely-understoried scrub swamp or bajo forest in low-lying areas. We also sampled all ages of successional vegetation we could find in the slash-and-burn farming landscape. This ranged from low, brushy second-growth only three years of age (since it was last farmed) to taller second-growth 30 years of age.

We analyzed capture data for nine species of migrants. Seven of the nine were captured at least as often in young second-growth as they were in mature forest. These were the Indigo Bunting (Passerina cyanea), Hooded Warbler (Wilsonia citrina), Ovenbird (Seiurus aurocapillus), Yellow-breasted Chat (Icteria virens), Gray Catbird (Dumetella carolinensis), Worm-eating Warbler (Helmitheros vermivorus), and White-eyed Vireo (Vireo griseus: Whitacre et al. 1993, 1995).

Two species, the Wood Thrush (Catharus mustelina) and Kentucky Warbler (Oporornis formosus), were captured most often in mature forest. This result agrees with results from elsewhere that indicate that these two species (along with some other migrant species) depend largely on mature forest while on their tropical wintering grounds. As a result of this substantial reliance on mature forest, we hypothesize that tropical deforestation may have played a role in apparent recent population declines of these two species.

We also found high densities of several migrant species in the low-canopied scrub swamp forests, and have drawn attention to the importance of these forest types to migrant birds and to conservation objectives in general.

In another facet of our research, we conducted intensive banding (including color-banding) on two 25-ha study plots in two mature forest types. In the end, we focused on one of these study plots, in order to determine density, winter survivorship, and year-to-year site fidelity of Kentucky Warblers and Wood Thrushes--among the two most forest-dependent migrant species wintering abundantly at Tikal. Data are not yet fully analyzed. However, it was clear that both species held winter territories in the forest understory and often returned to the same territory or a neighboring territory in consecutive winters.

Kentucky Warblers occurred at a density of 0.96 individuals per hectare (Madrid et al. 1995). This density was nearly twice the highest estimate obtained by researchers in Panama (Mabey and Morton 1992), and one third the density inferred from the territory size (0.3 ha) given by other researchers in southern Veracruz, México (Rappole and Warner 1980). The intermediate density we found might be expected if winter density decreased with distance from the breeding range, as was suggested by Mabey and Morton (1992).

In our mature upland forest study plot, Kentucky Warblers appeared to fill all or nearly all available space. Mist-net capture rates in bajo forest were yet higher than those in upland forest. While it is not safe to directly extrapolate from capture rates to population density (among other things, the vertical distribution of vegetation leads to different capture rates in vegetation types that differ in structure, regardless of population density), it seems likely that many of the natural forest types comprising the 1.8 million hectares of forest in the Maya Biosphere Reserve may have Kentucky Warbler densities similar to those we measured. Hence we conservatively estimate that the Maya Biosphere Reserve provides a winter home to at least a million Kentucky Warblers.

Since densities of Wood Thrushes on our main study plot appeared similar to those of Kentucky Warblers, we likewise estimate that a million or more Wood Thrushes winter annually in the Maya Biosphere Reserve.

These results underscore the direct importance of the Maya Biosphere Reserve to U.S. citizens. If the millions of hectares of forest of this protected area were to succumb to slash-and-burn farming, conversion to cattle ranches, or other existing pressures, there is a very high likelihood that this would translate directly into fewer Kentucky Warblers, Wood Thrushes, and other migrant birds that nest in the eastern U.S. and Canada, and that grace our springtime woodlands with the beauty and mystery of their song and their fleeting movements.