Conservation of the yellow-shouldered blackbird Agelaius xanthomus, an endagered West Indian species

Biological Conservation 55 (1991)119-138

Conservation of the Yellow-shouldered Blackbird Agelaius xanthomu~ an Endangered West Indian Species

James W. Wiley* Patuxent Wildlife Research Center, US Fish and Wildlife Service, Laurel, Maryland 20708, USA

William Post:~ & Alexander Cruz Environmental, Population and Organismic Biology Department, University of Colorado, Boulder, Colorado 80309, USA (Received 7 January 1989; revised version received 5 March 1990; accepted 13 March 1990)

ABSTRACT The yellow-shouldered blackbird Agelaius xanthomus, endemic to Puerto Rico and Mona Island, is endangered, mainly because of brood parasitism by the shiny cowbird Molothrus bonariensis, which reached Puerto Rico at least 30 years ago. The yellow-shouldered blackbird populations have since declined, about 770-1200 remaining (470-900 on Mona Island) by 1982-1986 compared to a population of about 2400 in 1975. Nearly all nests of blackbirds in most of its habitats are parasitized by cowbirds. This significantly reduces nesting success, but blackbirds have evolved no defenses against brood parasitism. Removal of cowbirds from the yellow-shouldered blackbird nesting grounds, modeled after similar programmes for the brownheaded cowbird Molothrus ater on Kirtland's warbler Dendroica kirtlandii nesting areas, significantly increased blackbird production. Blackbirds readily accept nest boxes, and breeding populations can be established in otherwise unusable sites and can be concentrated in mangrove habitats, where *Correspondence address: Southwest Research Group, 2140 Eastman Ave, Suite 100, Ventura, California 93003, USA. ~:Present address: Charleston Museum, 360 Meeting Street, Charleston, South Carolina 29403, USA. 119 Biol. Conserv.0006-3207/90/$03-50 © 1990 Elsevier Science Publishers Ltd, England. Printed in Great Britain

120

James W. Wiley, William Post, Alexander Cruz

they are more easily protected by cowbird removal. Furthermore, yellowshouldered blaekbird pairs using cavities, including nest boxes, fledged more young per nest than pairs using open nests.

INTRODUCTION The endemic yellow-shouldered blackbird Agelaius xanthomus (Fig. 1) was formerly widespread and abundant in Puerto Rico and Mona Island. However, since about 1940-50, its numbers have declined precipitously and it is now mainly confined to three locations: Mona Island and mangrove habitats in southwestern and eastern Puerto Rico (Post & Wiley, 1976). This blackbird is considered endangered by Federal and Commonwealth of Puerto Rico conservation agencies (Federal Register, 1976; Raffaele et al., 1977). Many factors have contributed to the blackbird's decline, including disease (Post, 1981a), loss of feeding and nesting habitats, and nest predation by the pearly-eyed thrasher Margarops fuscatus and introduced mammals (e.g. roof rat Rattus rattus and Norway rat R. norvegicus) (Post & Wiley, 1976). However, the major cause of the decline has been the extensive parasitism of nests by the shiny cowbird Molothrus bonariensis (Post & Wiley, 1977a; Wiley, 1985).

Fig. 1. Adultyellow-shoulderedblackbird at nest in black mangrove.

Conservation of yellow-shoulderedblackbird

121

Within the 20th century, the shiny cowbird extended its range from South America through the West Indies as far north as Cuba and then to the Florida Keys (Post & Wiley, 1977b; Cruz et al., 1985; Smith & Sprunt, 1987). It is a brood parasite and has become an important factor in depressing reproductive success of several species of native birds (Cruz et al., 1985; Wiley, 1985; P6rez-Rivera, 1986). Cowbird parasitism significantly reduces blackbird nesting success, but blackbirds have evolved no defenses against brood parasitism (Wiley, 1982). In 1975, we began experimenting with conservation techniques to reduce cowbird parasitism and improve yellow-shouldered blackbird breeding success and productivity. A portable trap suitable for the mangrove environment was designed to test the effectiveness of cowbird control on local bird populations. Concurrently, we developed methods of improving blackbird habitat by provisioning areas otherwise unusable for nesting, but adjacent to suitable nesting habitat, with nest boxes. Here we report on these conservation efforts and their results.

STUDY AREAS We conducted our work at two coastal zone study sites: Boquer6n Commonwealth Forest in southwestern Puerto Rico and Roosevelt Roads Naval Station in easternmost Puerto Rico. The Boquer6n Forest site is about 1 km wide and extends 15 km from La Parguera to Pita Haya, a village on the eastern shore of Bahia Sucia. The area is mainly mangrove (black mangrove Avicennia germinans, white mangrove Laguncularia racemosa, button mangrove Conocarpus erectus, and red mangrove Rhizophora mangle) forests and salt fiats, which border arid scrub and savanna. The Roosevelt Roads site (3260 ha, of which approximately 814 ha or 2 5 o are mangrove forest) also has well-developed mangrove forests, but less extensive salt fiats than the Boquer6n Forest. Mangrove species composition is similar to that at the Boquer6n Forest site. More detailed descriptions are available (Post & Wiley, 1976; Wiley & Wiley, 1979; Post, 1981b; Wiley, 1985).

METHODS In 1973-75 and 1981-82, we counted yellow-shouldered blackbirds, shiny cowbirds, and Greater Antillean grackles Quiscalus niger as they entered offshore mangrove cay roosts in the evening in Boquer6n Forest (Post & Wiley, 1977a, b; Post & Post, 1987). In both periods, counts were made from

122

James IV. Wiley, William Post, Alexander Cruz

a boat from which most flight lines into the roost could be seen. One or two persons counted the birds, while another recorded. Because icterids do not roost on offshore cays at Roosevelt Roads, we determined blackbird population size there by counting nesting pairs from 1975 to 1986. We tested 16 nest-box designs to determine which were acceptable to blackbirds and whether any deterred use by cowbirds. Boxes were constructed of 9-5-mm thick exterior grade plywood and were fitted with access doors to the nesting chamber for inspection of contents and for maintenance. We placed nest boxes in yellow-shouldered blackbird nesting areas from 1975 to 1977 at Roosevelt Roads (n = 51) and from 1977 to 1982 at Boquer6n Forest (n = 189). At Roosevelt Roads in 1975-77, we placed boxes in a grid pattern at 30-m intervals. Thereafter, boxes were irregularly spaced, dependent on what we judged to be suitable locations at Boquer6n Forest (Fig. 2). In addition to boxes, we provided modified natural cavities (sections of hollow black mangrove trunks fitted with floors and roofs) of dimensions similar to those natural cavities used by nesting yellow-shouldered blackbirds; 5 were established in the Boquer6n Forest study area in 1978 and 12 at Roosevelt Roads from 1977 to 1986. As with the nest boxes, the

Fig. 2. Placingnest box in open salt pan adjacent to yellow-shoulderedblackbird nesting habitat at Boquer6n Forest, Puerto Rico.

Conservation of yellow-shouldered blackbird

123

provided cavities were placed at heights of 1-2-1"9 m on posts or tree trunks. Boxes and cavities were considered 'occupied' if birds built nests within. Annual occupation rates are the number of pairs nesting in sites as a fraction of all available sites, including newly provided sites and sites surviving from former years. We monitored modified cavities, natural cavities, and boxes throughout the year. Sites were considered 'available' until they fell, rotted, or lost a top or bottom panel. We caught cowbirds in portable, walk-in decoy traps (Fig. 3). The traps, miniature versions of larger decoy traps used for cowbird control in Michigan (Shake & Mattsson, 1975), consisted of four wooden frame panels (each panel 1.2 x 1.8 m) and a funnel top, all covered with 12.7-mm mesh hardware cloth. One panel was fitted with a door, through which the operator could enter. Cowbirds could be caught by driving them into small gathering cages placed on the outside of the trap in one of the upper corners. Details of nest-box designs and decoy traps are available from the authors. Traps were placed on dry ground or, if over water, on elevated (I-m) wooden platforms and were baited with cracked maize and water. We removed cowbirds at least every other day, leaving 5-10 decoys in the trap. Decoys were changed at least once a week. During development of cowbird capture methods at Roosevelt Roads,

Fig. 3. Youth Conservation Corps (YCC) volunteers placing portable walk-in traps for shiny cowbirds in yellow-shoulderedblackbird nesting habitat at Boquerbn Forest, Puerto Rico.

James W. Wiley, William Post, Alexander Cruz

124

1975-82, we monitored 3-10 traps in blackbird nesting habitat and cowbird foraging sites. We also used mist nets to capture cowbirds (n = 2800) at a Roosevelt Roads horse stable, where they congregated to feed and for social activities. Nets were most effective during 0530-0830 h, when large flocks of cowbirds flew to the stables from their overnight communal roost. From the stable, the birds dispersed to other, natural feeding and social areas. Each cowbird was marked with a unique combination of coloured leg bands and, usually, a US Fish and Wildlife Service metal band. During 1975-82, all cowbirds were released after marking at Roosevelt Roads. We conducted cowbird removal experiments at Boquer6n Forest in JuneJuly 1980 and at Roosevelt Roads in June-July 1983 (i.e. during the middle of the blackbird's nesting season). In the 1980 experiments, 12 traps were evenly spaced (~ = 0.9 km apart) in openings along an 11-km strip of mangrove forest and salt pans, adjacent to yellow-shouldered blackbird nesting areas. We performed four experimental manipulations, using three traps for each: (A) all cowbirds captured were banded and released; (B) all cowbirds were destroyed; (C) female cowbirds were destroyed, male cowbirds were banded and released; and (D) male cowbirds were destroyed, females were banded and released. Each experimental manipulation was randomly assigned to the traps. Blackbird nesting success in the vicinity (~ surrounding area of nest monitoring for the 12 traps = 25 ha; R = 12-40 ha) of each trap was the assay for each treatment. For the June-July 1983 trapping effort at Roosevelt Roads, seven traps were set in two blackbird nesting areas that were approximately 1.5 km apart. The two traps in the 'MS' ('Main Swamp') study area were 150 m apart and the five traps in the 'OC' ('Officer's Club') area were positioned an average of 140 + 23 (SE) m from the nearest trap in a triangular pattern that TABLE 1 Winter (December to February) Roost Counts of Blackbirds in Boquer6n Forest, Puerto Rico

Locality

Yellow-shouldered blackbird

Shiny cowbird

1974-75

1981-82

1974-75 1981-82

Pita Haya La Parguera Bahia Montalva Bahia Sucia Buey Boquer6n

1050 156 284 147 9 17

165 30 14 52 0 5

11 45 897 203 0 10

Total

1 663

266

1 166

Greater Antillean grackle 1974-75

1981-82

157 279 396 539 0 16

321 406 1 538 1 513 250 1 400

716 659 1 369 943 800 928

1 387

5428

5415

Conservation of yellow-shouldered blackbird

125

encompassed approximately 6 ha. All trapped cowbirds were destroyed in 1983. We inspected nests at 2-3 day intervals. Parasitism rates, productivity, and nest success are presented for blackbird pairs nesting in sites where cowbird populations were not manipulated or for the period before cowbird removal, unless otherwise noted. Also, Boquer6n Forest data are for mainland Puerto Rico nests only. Successful nests (from the standpoint of nest predation or desertion) were those fledging at least one host or cowbird chick. Statistical procedures follow Zar (1974). Significance was set at the 0.05 level.

RESULTS A N D DISCUSSION Yellow-shouldered blackbird population size Between 1974-75 and 1981-82, the yellow-shouldered blackbird population in southwestern Puerto Rico declined by about 80%, whereas the population of Greater Antillean grackles, a species not affected by shiny cowbird parasitism (Wiley, 1985), did not change (Table 1). During the same interval, shiny cowbird populations increased by about 20%. We estimated the southwestern Puerto Rico population of yellow-shouldered blackbirds at about 300 in 1982. At Roosevelt Roads, we estimated about 200 yellow-shouldered blackbirds existed in 1975-76 (Post & Wiley, 1976). By 1982, the population had declined to six pairs (97% decline); in 1985 and 1986, only two nesting pairs were known. The blackbird population on Mona Island apparently has not been affected by cowbird parasitism; E. Hernandez-Prieto and A. Cruz (unpublished data) estimated 467 to 908 individuals in 1987-88. Thus, the total world population of the yellow-shouldered blackbird (1982-86) stands at about 771-1212 birds. Provision of nest sites Yellow-shouldered blackbirds use open sites and natural cavities (primarily black mangrove) for nesting. Of the 202 nests we inspected at Boquerrn Forest from 1975 to 1982, 39% (n = 77) were open nests and 61% (n = 118) were in natural or provided cavities and boxes. Natural cavities .were used by 10% (n = 20) of the nesting blackbird pairs (Table 2; Fig. 4). The mean annual occupation rate of natural cavities by blackbirds at Boquerrn Forest was 30+ 14.2% ( R = 16"7-55-6% per year; 13/41; n = 6 years, 1977-82).

126

James W. Wiley, William Post, Alexander Cruz

TABLE 2 Use of Nest Boxes, Modified Natural Cavities, and Natural Cavities at Boquer6n Forest, Southwestern Puerto Rico, 1977-82 Sites available b

Natural cavities: 41 Modified natural cavities: 20 Nest boxes: 467

Number of sites used ~ Honeybees

Percentage of sites used

1

0

73-2

4

4

1

60.0

49

46

22

67.0

YsBb

PRF

CM CG-D Rats Wasps

13 [20 c]

7

0

2

7

2

1

0

0

103

25

8

0

Names of species: YsBb, yellow-shouldered blackbird; PRF, Puerto Rican flycatcher; CM, Caribbean martin; CG-D, common ground-dove Columbina passerina. b Includes newly provisioned sites (n = 189) and sites surviving from former years. Includes yellow-shouldered blackbird nests for 1975, when cavity occupation by other species and total available cavities were not assessed.

Although apparently suitable cavities (resulting from man-caused die-offs of mangroves) were more abundant at Roosevelt Roads than at Boquer6n, only 4 (4%) of 114 blackbird nests we observed there from 1974 to 1986 were in natural cavities (cavity occupation rate at Roosevelt Roads v e r s u s Boquer6n Forest: Z2=64.82, d.f.= 1, p=0.0001; Table 3). The yellowshouldered blackbird occupation rate of natural cavities in eastern Puerto Rico was 3 %. Natural cavities had an observed mean life of 4.6 + 0-3 years (n = 42 cavities monitored for longevity). TABLE 3 Occupation of Nest Boxes, Modified Natural Cavities, and Natural Cavities at Roosevelt Roads, Eastern Puerto Rico, 1975-86 Sites available b

Natural cavities: 156 Modified natural cavities: 44 Nest boxes: 248

Number of sites used ~

Percentage of sites YsBb ZD CG-D PeT GAG Rats Wasps Honeybees used

4

0

36

14

14

0

0

22

57"7

1

0

6

15

0

5

0

0

61-4

0

1

16

11

7

32

69

9

58-5

a Names of species: YsBb, yellow-shouldered blackbird; ZD, zenaida dove Zenaida aurita; CG-D, common ground-dove; PeT, pearly-eyed thrasher; GAG, Greater Antillean grackle. b Includes newly provisioned sites (n = 51) and sites surviving from former years.

Conservation of yellow-shouldered blackbird

127

Fig. 4. Natural cavity used for nest site by yellow-shoulderedblackbirds, Boquer6n Forest, Puerto Rico. We provided blackbird nesting areas with several modified natural cavities to test their acceptance. These provisioned cavities were of approximately the same volume as natural cavities used by blackbirds. Of the 202 yellow-shouldered blackbird nests studied at Boquer6n Forest, 1% (n = 2) were in our provided cavities (Table 2). The mean annual occupation rate of provided cavities by blackbirds at Boquer6n Forest was 1 0 _ 13.7 %

128

James w. Wiley, William Post, Alexander Cruz

(r = 0-25 % annually; 2/20 sites available; n = 5 years, 1977-82; Table 2). At Roosevelt Roads, blackbirds used 1 (2%) provided cavity of 44 available (Table 3); 1% of all yellow-shouldered blackbird nests were in these sites. Provided cavities had an average life of 4.2+0.3 years (n = 15 modified natural cavities monitored for longevity). As suggested by Post & Wiley (1976), yellow-shouldered blackbirds accept artificial nesting boxes. We found that blackbirds will use a variety of sizes and shapes of boxes. In fact, blackbirds accepted every box design (n = 16) we provided. At Boquer6n Forest, 51% of the yellow-shouldered blackbird nests we monitored (n = 103/202) were built in nest boxes. Blackbirds showed an average annual occupation rate of 2 4 + 9 . 4 % ( n = 6 years, 1977-82; range = 18-40% annually) of boxes provided at Boquer6n Forest (Table 2). No boxes were accepted by blackbirds at Roosevelt Roads (Table 3). Nest boxes had a mean life of 2.7+0.1 years ( n = 189 boxes monitored through life of structure). The simplest and most economical nest-box design is one similar to an eastern bluebird Sialia sialis box, but with an oblong entrance about 8 x 5 cm (Fig. 5). Some of the boxes we provided approximated the volume

Fig. 5. Yellow-shoulderedblackbird feeding chick in nest box, Boquer6n Forest, Puerto

Rico.

Conservation of yellow-shouldered blackbird

129

of natural cavities, whereas others were larger than the 'average' natural cavity used by yellow-shouldered blackbirds. F r o m 1977 to 1982, we tested various designs of nest boxes, such as ones with long entrance necks, to determine if shiny cowbirds could be excluded. However, cowbirds parasitized all box styles that were used by the blackbirds. It is unlikely that it will be possible to devise a nest box that blackbirds will accept, but which shiny cowbirds will not enter, partly because the two species are approximately the same size (Wiley, 1982). For example, in Argentina, shiny cowbirds even parasitize the rufous hornero Furnarius rufus, which builds a mud nest with a long, spiraled entrance chamber (Hudson, 1920; Mason & Rothstein, 1986). Parasitism rate, nest success, and productivity In the Boquer6n Forest, 95% (191/202) of yellow-shouldered blackbird nests were parasitized from 1973 to 1982. All nests in natural cavities (n = 20) and provided cavities (n = 2) were parasitized, whereas 96% (n = 99/103) of nests in boxes and 91% (n=70/77) of open nests were parasitized (all p > 0.05; no difference in parasitism rate among nest types; Fisher exact probability test). At Roosevelt Roads, 95 % (n = 100/105; 1975-82) of blackbird nests (99 open, 1 natural cavity) were parasitized. All blackbird nests in natural cavities (n = 3) and the single provided cavity were parasitized after cowbird control was begun (1983). Even though parasitized, artificial cavity nests have the potential advantage of producing more host young than do open nests that are parasitized because the incidence of predation is lower than at open sites (predation rate at open nests = 31.9%, at closed nests = 18.3%; g 2 = 6.54, d.f. = 1, p=0-011; Post & Wiley, 1977a). Therefore, even though artificial boxes might be parasitized, blackbird production can be improved. At Boquer6n Forest, nest success of yellow-shouldered blackbirds from 1973 to 1982 was 51% (102/202). Nest success at open nests (36 %; n = 28/77, 70 parasitized) was lower than at natural cavities (75%, n = 15/20; all parasitized; Yates' X2= 8.10, d.f. = 1; p<0.05), provided cavities (100%, n = 2; all parasitized), and boxes (55 %, n = 57/103; 99 parasitized; ~2 = 6"37; p < 0.05). At Roosevelt Roads, 35 % (37/105; all but one open sites) of blackbirds were successful from 1975 to 1982 (before cowbird control began). Nest success of the few nests built in natural (n = 4) and provided cavities (n = 1) was 25% and 100% (1982-86, cowbird control performed 1983-86). At Boquer6n Forest, 0-38+0"05 ( n = 2 0 2 nests; 1973-82) yellowshouldered blackbirds fledged per nest. Clutch sizes (calculated before egg

130

James W. Wiley, William Post, Alexander Cruz

removal by cowbirds) did not vary among nest types (~ open nests = 2.98 + 0.62 eggs, n = 202; ~ nest boxes = 2"78 -I- 0.71, n --- 103; ~ cavity nests = 2.82 _ 0.80, n = 2 2 ; X2=8.11, d.f.=6, p=0.230). Pairs using open nests fledged fewer blackbird chicks (0.13 +0.05; n = 7 7 ) than did pairs using boxes (~ = 0.59 __+0.08 chicks/nest; n = 103; t = 5.05, p < 0.0001; Fisher-Behrens ttest) and natural cavities (0.40 ___0-1 t chicks; n = 20; t = - 2"51,p < 0"02). The two provided cavities used by blackbirds fledged 0"50+0"50 blackbird chicks/nest. Furthermore, fewer cowbirds fledged from parasitized boxes than from open nests in Boquer6n Forest: for open nests ~ = 1-56 + 0.14 (n = 70) and for boxes ~ = 0.72 + 0-08 (n = 99; Fisher-Behrens t = - 5.29, p < 0.001). An additional advantage of artificial nest sites is that they can be placed in sites away from trees and fitted with predator-guards. R o o f rats are the main predators of hole-nesting passerine birds in Puerto Rico. In 1977, we placed nest boxes and modified cavities on trunks of trees or on wooden posts without anti-rat devices, and rats used these sites as nests and shelters. We do not know what effect the provision of nest sites might have had in increasing rat populations in the blackbird breeding areas. In 1978, we attached 30-cmlong galvanized and aluminum flashing to the support posts o f some boxes. Results showed fewer rats occupying these sites; rats occupied 30% of the boxes in 1977, but an average of only 6% in 1978-81. Several additional boxes were fitted with flashing in 1979, and in 1980 nests were installed on polyvinyl chloride (PVC) pipe. In 1982, we put boxes on wooden posts, around which we placed 10-cm PVC pipe. The pipes extended from ground or water level to about 1"5 m. None of the 30 boxes that had bird nests, all of which were protected in this manner, were attacked by rats. The metal rat guards on nest boxes that had been placed in the blackbird nesting colonies in 1978-80 had disintegrated by 1982, when an inspection revealed that about 70% of 80 boxes had rat nesting material. In any case, no nesting boxes should be placed unless they are protected with rat guards. Some natural cavities (e.g. in trees distant from other trees and thus not providing corridors for rats) can also be fitted with rat guards. Another important feature ofartifical nesting cavities is that the boxes can be used to create new nest sites in areas adjacent to traditional nesting habitat, such as salt fiats that have no trees. There are m a n y such salt fiats in the mangroves, and blackbirds can be induced to colonize these sites by provisioning suitable nest boxes. Furthermore, new colonies can be protected through cowbird control (see below) and the boxes easily monitored. In addition to providing nest sites for blackbirds, the boxes and provided cavities were also used by other bird species and arthropods (Tables 2 and 3). In addition to the bluebird-type boxes, Caribbean martins Progne

Conservation of yellow-shouldered blackbird

131

dominicensis and Puerto Rican flycatchers Myiarchus antillarum used large boxes, similar to eastern screech-owl Otus asio boxes, that we placed in the Boquer6n study area in 1980. The blackbirds used few of these boxes (9% occupancy; 18 nests in 205 nest-box years), but in a conservation scheme involving artificial nests, it will be important to continue providing these larger boxes as a means of reducing competition for nest sites (see Prescott, 1982; Snyder et al., 1987). We found a major problem with the use of nest boxes. The microclimate of the box seems to be particularly favorable to nest mites. Two species of mites, Ornithonyssus bursa and Androlaelaps casalis, occur in both natural and artifical nest cavities in southwestern Puerto Rico (Post, 1981b). We found no difference in mite infestation rates among natural cavities, modified cavities, and nest boxes (X2=2.066, d.f.=2, p=0.356). The infestation by mites might lead to premature desertion of the nest by young birds, as Klaas (1970) found for the eastern phoebe Sayornisphoebe. In 1982, we also noted four instances of adult blackbirds deserting well-grown young that were in cavities. All the deserted nests were infested with mites. In contrast, four open nests that we examined during the same period that had well-grown young had no apparent mite infestation. In the future, it will be necessary to treat cavities with an insecticide, such as 2% Sevin powder. However, caution should be exercised in treating nest boxes with potentially hazardous chemicals (Jackson, 1985). In addition, other arthropods, such as scorpions, wasps and bees, should be destroyed, as they reduce the attractiveness of the nest boxes to blackbirds (Tables 2 and 3).

Concentration of the yellow-shouldered blackbird population A recommended conservation objective for the yellow-shouldered blackbird is to keep the populations in areas that can be readily protected and surveyed. The Boquer6n Commonwealth Forest, where the blackbirds are now concentrated, is ideally suited. It is protected by the Department of Natural Resources (DNR), but due to understaffing the D N R is unable to adequately patrol the landward side of the forest. This suboptimal vigilance allowed some yellow-shouldered habitat disturbance by humans. However, since 1977, the D N R has patrolled the navigable sections of the forest and has succeeded in stopping the destruction of mangroves and the invasion of cays by squatters and houseboats. This accomplishment is particularly important in the case of the offshore cays, because they provide the only blackbird habitat that is still relatively cowbird-free. In 1962-77, many rhesus macaques Macaca mulatta were held in the two large islands, La Cueva and E1 Guayac/m, that form part of Boquer6n Commonwealth Forest. These animals were provisioned ad libitum with

132

James W. Wiley, William Post, Alexander Cruz

large quantities (up to 200 kg/day) of monkey chow and dog food, placed at 5-10 widely separated feeders. By 1974-75, at least 1000 yellow-shouldered blackbirds and an undetermined number of shiny cowbirds were using these feeders daily. This was also the period when an estimated 250 pairs of yellow-shouldered blackbirds were nesting in the mangrove adjacent to the two islands. In 1981, the monkeys were taken off the islands, and food provisioning was ended. We do not know what effect this sudden reduction in food availability might have had on the dispersion of yellow-shouldered blackbirds. It is possible that the current reduction in the blackbird population in southwestern Puerto Rico is based in part on dispersal, although we know of no other significant breeding populations other than those on Mona and at Ceiba. The interior population at San Germhn, which had consisted of 10-12 pairs in 1975, could not be found in 1982. A possible method of assuring that many blackbirds stay in the mangrove z o n e would be to provide food to birds in the vicinities of the nesting colonies. This could be done with a variety of foods; the blackbirds eat sorghum, rice, bananas, papayas, and commercial animal food. Any feeding stations would have to be protected from rats. Also, as rats carry food away from feeding sites, it would be necessary to reduce rat populations in the vicinities of any feeding stations. This could be accomplished with rodenticides. Feeding stations could be operated by the same workers that monitor cowbird traps. Cowbird control Another, and possibly the most effective, conservation tool for the recovery of the yellow-shouldered blackbird is cowbird control. This technique is modeled after that employed to control brown-headed cowbirds Molothrus ater on Kirtland's warbler Dendroica kirtlandii and least Bell's vireo Fireo bellii pusillus nesting grounds (Shake & Mattsson, 1975; Mayfield, 1977; Kelly & DeCapita, 1982; Beezley & Rieger, 1987; Rothstein et al., 1987). In 1980, we tested the technique in the Boquer6n Commonwealth Forest. A total of 1692 shiny cowbirds was trapped in 12 traps during 11 days, 14-31 July (13 birds/trap day). In the A (control) trap vicinities, 92% of 12 yellow-shouldered blackbird nests were parasitized. In contrast, around the B traps, where all cowbirds were killed, 45% of blackbird nests were parasitized (Table4). The parasitism rate at blackbird experimental areas where only males were removed (D) was 67 %. The lowest incidence of parasitism (30%) was found in the experimental class C, where all females were destroyed. The differences among treatments A, B and C are significant (Table 4). Compared with areas where cowbirds were not removed, blackbird clutch

Conservation of yellow-shouldered blackbird

133

0

~5

.?

6

~5

e.,

ct~

6

e., 0 e~

.~ ~ ¢.~

0 0 r~

0 L~ "r9

0 e~ 0 . ~,

ex~

o

0

~"

•~

V

e~

e.

e~

e.a

0

0

~.~ ~.~ ..~ •

e. 0

0

E

..0 0

.=_ .,~ "d

g=

134

James W. Wiley, William Post, Alexander Cruz

size was significantly larger in some cowbird removal areas. Shiny cowbirds affect yellow-shouldered blackbird clutch size by destroying and removing host eggs. Blackbirds also desert nests in which cowbirds have laid eggs (Wiley, 1982). Near the control (A) traps, blackbird clutch size was 2"58 ___0.29 (n = 12), and the nests contained an average of 3.36 ___0.19 (n = 12) cowbird eggs. In contrast, in the vicinity of the C traps (females removed), blackbird clutch size was 3.00 ___0-15 (n = 10;p < 0.005; Mann-Whitney test), and the nests had an average of only 0-70 + 0.37 (n = 10) cowbird eggs. More important, nest success of blackbird pairs in experimental area B was higher than that of pairs in the control group (p = 0-05, Fisher exact probability test). Of the total captures of cowbirds, 77% were repeats, which indicates a resident population. After the first 3 days of trapping, most new individuals were juveniles, perhaps transients. In the D trap vicinities (male removal), the mean recapture rate of males marked in other experimental areas was 9%. Elimination of males in an area apparently allowed foreign males to move into the activity spaces occupied by females. In contrast, in the female elimination (C) areas, no marked females moved into the experimental areas, and few new females were caught. The latter result might mean that either males continued to defend their activity spaces against intruders of both sexes or no additional females were available for recolonization. Removal of only male cowbirds from the nesting areas did little to improve yellow-shouldered blackbird nesting performance. The best results came from removing either all the cowbirds or removing all the females. It will be necessary to make more detailed studies of the territorial behavior of individual male cowbirds, to determine whether they indeed keep strange females out of their activity spaces. If this does not prove to be the case, it will be preferable to remove all cowbirds, as males are known to destroy eggs (Hoy & Ottow, 1964). At Roosevelt Roads in June-July 1983, we captured 325 cowbirds in 149 trap days for an average of 2.2 cowbirds per trap day. Whereas in 1981 and 1982, no yellow-shouldered blackbird pairs fledged blackbird young from this study area, in 1983, a minimum of 3 blackbird pairs fledged their own chicks, perhaps partly as a result of the cowbird removal. Jon Heisterberg and Fernando Nufiez intensified the trapping effort at Roosevelt Roads after 1983, through a cooperative programme among the US Fish and Wildlife Service, US Navy, and US Department of Agriculture. Nest success of yellow-shouldered blackbirds has shown improvement after cowbird control was instituted. Blackbird nest success in 1975-82 was 35 % (37/105), but increased to 71% (10/14; Fisher exact probability test with Tocher = 6"77, p < 0-05) during 1983-86. Of the cowbirds trapped in yellow-shouldered blackbird nest areas at

Conservation of yellow-shouldered blackbird

135

Roosevelt Roads, 91% were marked individuals that regularly used the horse stable for staging and feeding before moving into the mangroves. Thus, we recommend a cowbird trapping strategy that includes removal of birds at feeding and social staging areas, such as livestock feed lots, as well as trapping in nest areas to remove transient cowbirds or those originating from other roosts that do not use the staging areas. Cowbird control through trapping can be accomplished with little risk to yellow-shouldered blackbirds. Blackbirds were not attracted to the cowbird traps; we caught 6 blackbirds (all but one were juveniles) in 7112 trap days (0.0008/trap day). However, traps must be inspected at least daily and continuously well-supplied with shade, food, and water to minimize the chance of bird deaths.

CONCLUSIONS Although programmes for the control of the brown-headed cowbird on the breeding grounds of the endangered Kirtland's warbler have been operated for more than 15 years, the warbler population remains at only about 200 pairs and one may question whether the trapping of thousands of cowbirds has had any real benefit to the recovery of the warbler. Studies of the effectiveness of brown-headed cowbird removal in other host populations have also demonstrated limited success (Beezley & Rieger, 1987; Rothstein et aL, 1987). In the Sierra Nevada, attempts to experimentally trap-out a population of brown-headed cowbird resulted in only partial success because cowbird foraging habitat was widespread and dispersed (Rothstein et al., 1987). Similarly, shiny cowbird foraging habitat is widespread and dispersed around the nesting habitat of yellow-shouldered blackbirds in Puerto Rico. Despite the limitations of trapping, perhaps the best available strategy for reducing the impact of shiny cowbird parasitism on yellow-shouldered blackbirds is to remove the cowbirds from the blackbird's nesting areas. There is no possibility that the parasite can be eliminated from entire islands, because it is now so abundant and uses such a variety of hosts. However, by intensively trapping cowbirds in and around blackbird nesting colonies, we believe the cowbird removal programme has excellent potential for improving yellow-shouldered blackbird and other host species' reproductive success. The success of the cowbird removal programme should be evaluated by the number of breeding pairs of blackbirds, and not just the percent parasitism and blackbird productivity per nest. As stressed by Rothstein et al. (1987) the control programme should be continued for several years to determine its efficacy. Also, cowbird removal should be performed in areas

136

James I4/. Wiley, William Post, Alexander Cruz

adjacent to the hosts' breeding habitat. Although trapping might not be the most effective means of controlling cowbird parasitism on native avifauna, it is adequate at least as a stop-gap measure until less labor-intensive means can be devised. The conservation techniques that we have outlined can be undertaken at relatively little cost in time and money. It is important that the local government be involved in this work. In 1983, cowbird control programmes were begun at Boquer6n Forest by the Commonwealth of Puerto Rico Department of Natural Resources and at Roosevelt Roads by the US Fish and Wildlife Service. The recovery of the yellow-shouldered blackbird is a worthy goal--of all the threatened species in the Caribbean, it possibly has the best chance of recovery. Success with this species could serve as an example of what can be accomplished in the Caribbean with an aggressive conservation programme.

ACKNOWLEDGEMENTS This work was supported by the US Fish and Wildlife Service; National Science Foundation Grant PRM-8112194 to the University of Colorado, Alexander Cruz, principal investigator; New York Zoological Society; University of Miami; University of Colorado Grant-in-Aid; US Forest Service Institute of Tropical Forestry; and the Caribbean Islands National Wildlife Refuge. We thank Sean Furniss, Caribbean Islands National Wildlife Refuge Manager, for allowing us to use Young American Conservation Corps (YACC) and Youth Conservation Corps (YCC) summer employees and US Fish and Wildlife Service personnel to build nest boxes and cowbird traps and to help monitor traps in the Boquer6n Forest study area. We are grateful to Lt Cmdr GeoffCullison and Mr Felix Mestey, US Navy, for their support and encouragement during this programme. We also thank the Department of Natural Resources, Commonwealth of Puerto Rico, for allowing us to undertake investigations in the Commonwealth Forest of Boquer6n. For their assistance in the field and laboratory, we thank Jos6 Antonio Diaz Algarin, Ruthy Andujar, Wayne J. and Angela Arendt, Cheryl Belitsky, John Blankenship, Julio Cardona, Jos6 Col6n, Scott Corbett, Carlos A. Delannoy, Joseph diTomaso, Sean Furniss, Robert Johnson, Elizabeth Litovich, Tammie K. Nakamura, Evelyn and Fernando Nufiez, Guillermo Ortiz Malare, Eric C. Phoebus, Colin A. Post, Lucy Ramos, Nydia Rivera, Rafael Hernandez S~nchez, Eduardo Santana C., Dwight Smith, John Taapken, Augustine Valido, Beth Wiley, and Joseph M. Wunderle, Jr. We are grateful to Chandler S. Robbins, Sandra MacPherson,

Conservation of yellow-shouldered blackbird

137

David Ehrenfeld, Stephen Rothstein, and an a n o n y m o u s reviewer for their help in improving the manuscript.

REFERENCES Beezley, J. A. & Rieger, J. P. (1987). Least Bell's vireo management by cowbird trapping. Western Birds, 18, 55-61. Cruz, A., Manolis, T. & Wiley, J. W. (1985). The shiny cowbird: a brood parasite expanding its range in the Caribbean region. In Neotropical Ornithology, ed. P.A. Buckley, M.S. Foster, E.S. Morton, R.S. Ridgely & F.G. Buckley. A.O.U. Ornithol. Monogr., No. 36, pp. 607-20. Federal Register (1976). 14 FR 51021; 19 Nov. 1976. Hoy, G. & Ottow, J. (1964). Biological and oological studies of the molothrine cowbirds (Icteridae) of Argentina. Auk, 81, 186-203. Hudson, W. H. (1920). Birds of La Plata. E. P. Dutton, New York. Jackson, J. A. 0985). On the control of parasitism in nest boxes and the use of pesticides near birds. Sialia, 7, 17-25. Kelly, S. T. & DeCapita, M. E. (1982). Cowbird control and its effect on Kirtland's warbler reproductive success. Wilson Bull., 94, 363-5. Klaas, E. E. (1970). A population study of the eastern phoebe, Sayornis phoebe, and its social relationships with the brown-headed cowbird, Molothrus ater. PhD dissertation, University of Kansas, Manhattan. Mason, P. & Rothstein, S. I. {1986). Coevolution and avian brood parasitism: cowbird eggs show evolutionary response to host discrimination. Evolution, 40, 1207-14. Mayfield, H. (1977). Brood parasitism--reducing interactions between Kirtland's warblers and brown-headed cowbirds. In Endangered Birds, ed. S. A. Temple. University of Wisconsin Press, Madison, pp. 85-91. P6rez-Rivera, R. A. (1986). Parasitism by the shiny cowbird in the interior parts of Puerto Rico. J. Field Ornithol., 57, 99-104. Post, W. (1981a). The prevalance of some ectoparasites, diseases, and abnormalities in the yellow-shouldered blackbird. J. Field Ornithok, 52, 16-22. Post, W. (1981b). Biology of the yellow-shouldered blackbird--Agelaius on a tropical island. Bull. Florida St. Mus., Biol. Sci., 26, 125-202. Post, W. & Post, K. W. (1987). Roosting behavior of the yellow-shouldered blackbird. Florida FieldNaturalist, 15, 93-105. Post, W. & Wiley, J. W. (1976). The yellow-shouldered blackbird--present and future. Amer. Birds, 30, 13-20. Post, W. & Wiley, J. W. (1977a). Reproductive interactions of the shiny cowbird and the yellow-shouldered blackbird. Condor, 79, 176-84. Post, W. & Wiley, J. W. (1977b). The shiny cowbird in the West Indies. Condor, 79, 119-21.

Prescott, H. W. (1982). Using paired nesting boxes to reduce swallow-bluebird competition. Sialia, 4, 3-7. Raffaele, H. A., Velez, M. J., Cotte, R., Wheelan, J. I., Keil, E. R. & Cumpiano, W. (1977). Animales taros yen peligro de extinci6n en Puerto Rico. Department of Natural Resources, San Juan, Puerto Rico.

138

James W. Wiley, William Post, Alexander Cruz

Rothstein, S. I., Verner, J., Stevens, E. & Ritter, L. V. (1987). Behavioral differences among sex and age classes of the brown-headed cowbird and their relation to the efficacy of a control program. Wilson Bull., 99, 322-37. Shake, W. F. & Mattsson, J. P. (1975). Three years of cowbird control: an effort to save the Kirtland's warbler. Jack-Pine Warbler, 53, 48-53. Smith, P. W. & Sprunt, A., IV (1987). The shiny cowbird reaches the United States. Amer. Birds, 41,370-1. Snyder, N. F. R., Wiley, J. W. & Kepler, C. B. (1987). The Parrots of Luquillo: Natural History and Conservation of the Puerto Rican Parrot. Western Foundation of Vertebrate Zoology, Los Angeles. Wiley, J. W. (1982). Ecology of avian brood parasitism at an early interfacing of host and parasite populations. PhD dissertation, University of Miami, Coral Gables, Florida. Wiley, J. W. (1985). Shiny cowbird parasitism in two avian communities in Puerto Rico. Condor, 87, 165-76. Wiley, J. W. & Wiley, B. N. (1979). The biology of the white-crowned pigeon. Wildl. Monogr., No. 64. Zar, J. H. (1974). Biostatistical Analysis. Prentice-Hall, New Jersey.