MESH SIZE AND BIRD CAPTURE RATES IN MATO GROSSO DO SUL STATE, BRAZIL

Mist-nets alternating 36-mm and 61-mm mesh in woods and low vegetation of “cerrado” (Brazilian savanna) tested bird-capture efficiency relative to bird length and mass. Of 1,296 birds captured and 102 species, 785 (93 species) were with 36-m mesh and 511 (69 species) with 61-mm mesh. The 61-mm mesh improved capture rates only for some larger species; so, in general, 36-mm mesh mistnets are more appropriate for field work in “cerrado” areas.


INTRODUCTION
Mist nets have often been used for bird studies, and their results compared with other methods (e.g., Ralph et al., 1993;Rappole et al., 1998;Poulin et al., 2000). Rappole et al. (1998) have proposed a methodology that combines two procedures (mist nets and point counts) to provide a more accurate assessment of avian habitat use. According to Poulin et al. (2000), while point counts are less demanding in the field, the manipulation of netted birds allows documentation of various aspects of their biology. However, a lot of restrictions concerning mist-net use were listed by Remsen & Good (1996).
When mist nets are employed, the question arises as to which mesh size produces more captures per hour, since one reason why an individual cannot be sampled is its size; it can be too big or too small to get tangled in nets (MacArthur & MacArthur, 1974). Jenni et al. (1996) determined capture efficiency of 36-mm-mesh by observing birds flying towards the nets in various habitats. They concluded that different species (morphology and body mass) and climatic conditions (wind, shade) can provide different capture rates for 36-mm nets.
With this objective, Karr (1981) indicated that 36-mm-mesh nets were the most effective ones for a wide range of birds encountered in most terrestrial habitats, and these have been most frequently used (Karr, 1980;Blake & Loiselle, 1991;Loiselle & Blake, 1994). Pardieck & Waide (1992) observed however, that in most works involving mist-net captures, mesh size is not mentioned. From 25 studies on birds, using mist nets that the authors reviewed, in only four of them mesh size was specified. In more recent works, one can find a similar situation; as it has been exemplified by Young et al. (1998) for Costa Rican birds; Silkey et al. (1999) for Californian birds; Reinert et al. (1996), Marini et al. (1997), Oniki & Willis (1999) and Stratford & Stouffer (1999) (the four latter about Brazilian birds).
The goal of the present study was to compare bird capture efficiency of two mesh sizes (36 and 61 mm) in areas of cerrado and some other habitats of Mato Grosso do Sul State, Central Brazil.

MATERIALS AND METHODS
From ten to sixteen 36-and 61-mm-mesh mistnets were alternately spaced for 28 months (August 1994to December 1996 in linear transects. All nets had the same dimension (12 x 2 m) and were operated for the same length of time (usually from 5:30 am to 3:00 pm) in 14 locations. These study plots were located on large eucalyptus plantations outside the cities of Brasilândia (21º25'S; 52º03'W), with three study sites; Selvíria (20º36'S; 51º41'W), with two study sites; and Três Lagoas (20º75'S; 51º67'W), with nine study sites. Eucalyptus plantations belong to the paper manufacture and cellulose products industry, and large plantations of this exotic species coexist with fragments of native vegetation.
The fourteen study sites sampled five different vegetation types, with varying degrees of human disturbance: "cerrado" strictu sensu (five study sites -3,750 net-hours); "cerradão" -a closed-canopy tall cerrado (five study sites -5,300 net-hours); gallery forest (one study site -2,370 net-hours); degraded second-growth forest (one study site -710 net hours), and eucalyptus plantations with and without understory (two study sites -1,400 net-hours). Although sampling effort varied between study sites, both 36-mm and 61-mm mesh nets were always simultaneously used in equal numbers. Chi square tests were used to check statistical significance of different capture rates, and Yates' Correlation was used when there were only two categories in a distribution (Fowler & Cohen, 1995). All captured birds were banded with bands supplied by Cemave/Ibama.
For alternating 36-and 61-mm mesh nets, I compared captures relative to bird length (measured by means of a metal ruler) and body mass (measured with 50-, 100-, 300-, and 1000 g spring scales). The term "efficiency" is used here to indicate the mesh size that has captured the number of individuals of each species per unit of effort.
The smaller mesh nets captured most individuals and species (Table 1). Platyrhynchus mystaceus, one of the smallest species, was only captured with 36-mm mesh nets. However, for some larger species, the 61-mm mesh was more efficient. This is the case of Piaya cayana (only captured with these nets), Momotus momota, some doves (Leptotila verreauxi, L. rufaxilla and Claravis pretiosa), and Passeriformes (Taraba major and Automolus leucophthalmus).

Species
Mean body mass (g) Mean total length ( 1  TABLE 1  TABLE 1  TABLE 1  TABLE 1

ptures) and species with 36-and 61-mm mesh, es) and species with 36-and 61-mm mesh, es) and species with 36-and 61-mm mesh, es) and species with 36-and 61-mm mesh, es) and species with 36-and 61-mm mesh, with mean bod with mean bod with mean bod with mean bod
with mean body y y y y mass (g) and mean total length (mm). See Pir mass (g) and mean total length (mm). See Pir mass (g) and mean total length (mm). See Pir mass (g) and mean total length (mm). See Pir mass (g) and mean total length (mm   1 (  TABLE 1 (  TABLE 1 (  TABLE 1 (  TABLE 1 (C C C C Continued ontinued ontinued ontinued ontinued) ) ) ) ) The 36-mm mesh nets captured birds from 10 to 450 mm long (mean = 164 mm, s.d. = 64.7), while the 61-mm mesh captured birds from 83 to 490 mm long (mean = 199 mm, s.d. = 80.3). Both meshes were more efficient for birds between 100 and 149 mm long, and the 61-mm mesh captured more individuals over 250 mm long (Fig. 1).
Birds with body masses from 2.5 g to 265 g were captured with 36-mm mesh (mean = 24.5 g, s.d. = 25.5). With 61-mm mesh, body masses ranged between 4.5 g and 186 g (mean = 36 g, s.d. = 33.8). Both mesh sizes were effective for birds in the 10-20 g range, even though the 36-mm mesh captured more birds. For birds over 30 g, both meshes captured similar numbers and, in these cases, sometimes the 61-mm mesh was more efficient (Fig. 2).
Although I have found that bird length is related to mesh capture rates, Jenni et al. (1996) indicate that body mass and cranium size are more useful measures than the total length for determining the capture rates, since individuals' capture probability is directly related to head entanglement. Most birds were 10-20 g in three previous studies in southern Brazil (Reinert et al., 1996;Marini et al., 1997;Oniki & Willis, 1999).  The simultaneous use of two nets has the advantage of improving the abundance estimates for some larger species. In relation to species diversity, the isolated use of smaller mesh nets would produce almost the same results as the simultaneous use of two mesh sizes, at least in these "cerrados" and nearby areas.
By using only 36-mm mesh one can increase the capture rate, particularly when there is a greater density of small birds, which is almost always the case in tropical-forest understories. For bigger, heavier species, the 61-mm mesh would substantially improve capture rates.