Biogeographical endemism and habitat restriction are two easily measured quantities that can be used as indicators of species' ecological restrictions. We analysed and compared these two types of information from available literature sources in an attempt to identify all bird species critically dependent on tropical deciduous forests of western Mexico and Central America. Based on patterns of biogeographical restriction, we identified 42 endemics, 33 disjunct endemics, 59 corridor species and 3 seasonal endemics associated with tropical deciduous forest (TDF) in this region. Based on patterns of habitat use in these species, we classified them as 50 tropical deciduous forest-restricted species and 82 apparent habitat generalist species. No habitat use information was available within the TDF belt for five of the biogeographically restricted species. We found a high proportion of apparent habitat generalists (60%) among biogeographically restricted species. We discuss three specific scenarios in which species critically dependent on TDF may nonetheless appear generalized in their patterns of habitat use. These species are termed “cryptic habitat specialists.” The proportion of apparent habitat generalists is highly variable among biogeographical regions and habitat types. Thus, global biodiversity conservation prioritizations that rely on habitat restriction as an indicator of ecological sensitivity may underestimate conservation needs within bioregions that contain high proportions of cryptic habitat specialists.

It is known that White Eared-pheasants Crossoptilon crossoptilon drouyni interbreed widely with Tibetan Eared-pheasants C. harmani at the boundary of their ranges. A new hybrid zone has been found recently in eastern Tibet, far away from the boundary of the parental species' ranges. Based on ecological observations of eared-pheasants and the geographical history and pattern of modern glaciers, we have attributed the complex morphological variation of eared-pheasants and the high biodiversity of the eastern Qinghai-Tibet plateau to its varied geography.

In a recent study, O'Brien et al. (1998a) sought to provide a better understanding of the distribution and status of the Bornean Peacock-pheasant Polyplectron schleiermacheri in Central Kalimantan, Indonesia, based on an analysis of questionnaires and semi-structured interviews with local inhabitants. This was the first scientific contribution concerning this species after the launch of the Conservation Action Plan for Pheasants (McGowan and Garson 1995). Although the authors of this study acknowledged and discussed the limitations of the questionnaire approach and some of the anomalies in the replies they received, we still feel that the questions posed in this survey may have been phrased in such a way as to invite slight biases in the answers, and that some of the analyses of the replies may have created further biases. We therefore feel it may be helpful to review these uncertain areas in order to help improve questionnaire surveys in future fieldwork. Two of us (I.S. and A.P.S.) helped conduct the surveys in question, and all of us have experience of galliforms in Kalimantan.

We occasionally refer to additional information concerning the survey which was presented in another publication (O'Brien et al. 1998b).

We thank Sözer et al. (2000) for paying such close attention to our paper and taking the time to express their opinion so extensively. Sözer et al. express the opinion that our questionnaire and treatment of the data were biased. They objected to our request for assistance that accompanied the questionnaire, the use of pictures of unequal size, the wording of most questions, and what they considered a flawed semi-structured interview. Finally, they question the ability of the Kalimantan people to comprehend the interview process and respond accurately. They conclude a long discussion of the semi-structured interview with the comment “the data presented by O'Brien et al. (1999) are not incontrovertible”. The observations of Sözer et al., however, are simply opinions and are not backed by data or supporting arguments. Our opinions differ, and we do not find their arguments compelling. We have not presented our study as incontrovertible, and have in fact discussed the limitations of our survey at length in the manuscript.

China is a large country, occupying 9,600,000 km2. It lies between longitudes 73°40′E and 135° 05′E, a distance of some 5,200 km. Mountainous in the west and flat in the east, the land surface of China slopes downward from west to east in a three-step staircase and can be divided into moist, semi-moist, semi-arid and arid regions from the east coast to the inland north-west respectively. From north to south, China covers frigid, temperate, subtropical and tropical zones. All of this variation in geographical and climate conditions combine to give it a high diversity of fauna and flora.

The country is rich in Galliformes. Sixty-one species of two families (Tetraonidae and Phasianidae) are found in the country (Cheng 1994), which is about 22% of the 277 Galliformes species of the world (Sibley and Monroe 1990). In recent decades, as a result of big changes in the environments and the increasing human activities, the ranges of Chinese Galliformes have decreased and the population densities have declined (Zheng and Zhang 1993). The status of the country's Galliformes was assessed in the mid 1990s as part of two global reviews: through the compilation of IUCN/World Pheasant Association (WPA) action plans for pheasants (McGowan and Garson 1995) and partridges (McGowan et al., 1995), and through the revision of BirdLife International's list of threatened bird species (Collar et al. 1994). Both assessments considered that a high proportion of the world's threatened Galliformes occurred in China.

In the Cerrado of Brazil, the largest Neotropical savanna region, open habitats are suffering an intense process of conversion to pasture and agricultural land. This study evaluated the responses of birds to the thinning of tree and shrub cover in plots of cerrado sensu stricto, the dominant habitat in the region, and examined whether disturbed plots supported bird communities of open natural habitats with similar structure. Birds were censused by point counts in five disturbed sites and in seven preserved patches, with a total of 304 samples of 20 minutes in 68 points. Cleared plots were not suitable for species strongly dependent on trees and shrubs, such as frugivores and most insectivores. Their abundance was associated with the complexity of vegetation structure in these plots. Shrubby pastures held about 120% more species than the traditional cleared ones. Bird communities of pastures showed large differences in relation to those of grasslands with similar structure. Communities of pastures were dominated by a seedeater, the Blue-black Grassquit Volatinia jacarina, and most species were habitat generalist or oportunistic predators, such as raptors and insectivores. Obligate grassland species, like the Grass Wren Cistothorus platensis, Cock-tailed Tyrant Alectrurus tricolor and Black-masked Finch Coryphaspiza melanotis, as well as other abundant species in grasslands, such as Sharp-tailed Grass-tyrant Culicivora caudacuta and Capped Seedeater Sporophila bouvreuil do not adapt to pastures. Grassland grasses probably do not recover spontaneously in abandoned pastures and in other agricultural lands. Thus, conservation of grassland birds will strongly depend on a urgent programme of identification and protection of large areas of Cerrado dominated by campo limpo and campo sujo. These species would also benefit from the maintenance of a band of grassland along gallery forests.

Since the mid-1980s, exhausted pastures in Ecuador have been increasingly abandoned, allowing forest regeneration. At approximately 2,200 m in the Tandayapa valley I surveyed four abandoned pastures to evaluate their use by birds. Each former pasture represented a different age of vegetation maturity. The number of bird species recorded in each successively older abandoned pasture increased but only half the number of species recorded in the undisturbed forest site was recorded in the most mature pasture. However, at least four restricted-range bird species were recorded in a single pasture. As pastures rapidly convert to secondary forest, more bird species and rarer bird species use them, even in highly disturbed areas where surrounding pristine forest constitutes less than 10% of local forest cover. Conservation efforts should then be directed toward them.

We investigated the decline and current status of the Black Kite Milvus m. migrans Boddaert 1783 and the endemic and distinct Red Kite (Cape Verde Kite or Cape Verde Red Kite) Milvus milvus fasciicauda Hartert 1914 on the Cape Verde archipelago. In 1996 and 1997, the first census of both kites on the entire archipelago (Hille 1998) revealed that fewer than 10 individuals of each species still existed and that extinction of all kites on Cape Verde was imminent. As expected from this first survey, the tiny island populations of kites further declined and only two individual Red Kites and one Black Kite were found in 1999. Both populations are expected to become extinct very soon. The presumed causes of extinction such as indirect poisoning, lack of food in a changing landscape, and human persecution are discussed.

Previously known from two specimens, Kaempfer's Tody-tyrant Hemitriccus kaempferi is a globally threatened species endemic to the southern Brazil Atlantic Forest. The first field observations of this little-known taxon were made in the 1990s and our data therefore provide the first published information on its natural history, including breeding biology, behaviour and vocalizations, as well as a new locality. As very little information has appeared in the literature concerning the identification of this species, we present notes on its plumage and habits. Remarks concerning the species's conservation prospects and the importance of the remaining forest in the vicinity of the type-locality for other threatened bird species are made.