z zyx zyxwvutsrqponm zyxwvutsrqpo h i s (1999) 141, 60-69 Breeding ecology of the Sulawesi Red=KnobbedHornbill Aceros cassidix MARGARET F. KINNAIRD* & TIMOTHY G. O'BRIEN Wildlife Conservation SocietK International Programs, 185th and Southern Blvd., Bronx, NY 10460, USA zyxwvutsrq D a t a are presented for a four-year study of the breeding biology of the Sulawesi Red-knobbed Hornbill Aceros cassidix. The breeding season normally began in mid-June and lasted 27-30 weeks. Initiation of nesting appeared t o be stimulated by the cessation of the rains and timed such that chicks emerged during a period of fruit abundance. Nesting period averaged 139 days and incubation was estimated at 35-40 days. Females remain sealed in the nest for an average of 108 days and nestlings fledged, on average, 28 days after the female emerged. Nesting densities were up t o 10.4/km', nesting success was high (up to 80°h) and repeated use of nests between years was common. Males delivered a lou.-protein diet of ripe fruits (89% of total diet) from 12 families and 52 species; invertebrates composed only 1 % of food items. Figs (Ficus spp.) were the primary diet item, accounting for 81% of fruit biomass. Males increased feeding visits throughout the study, but the biomass of fruit delivered declined shortly after t h e female emerged. Reduced feeding prior t o fledging may entice the nestling t o emerge. The long developmental period of Sulawesi Red-knobbed Hornbills may result, in part, from the lcm. protein content of the diet. Despite a 16% annual production, numbers in the study area have remained stable over the past 15 years. I t is suggested that high post-fledging mortality or dispersal to degraded areas outside the reserve maintains population numbers. Distinguishing between these mechanisms is important for understanding the dynamics of hornbill populations. Hornbill5 (familk Bucerotidae) are well k n o 1~1 for their ornate casques and unique nest-sealing beha\ lour Of the 5-1 recogniied hornbill species (Kenip 1995), all are restr:<ted to the Old World tropics The breeding biolog) ( ~ fseieral African hornbill species has been documented (Moreau 1937, Moredu & Moreau 1940, 1941, Kemp 1979 Kalina 1988, 1989) but few field stuitiv ha\ c been conducted on Asian hornbills With thr cx,izption ot i\orL in Thailand by Ponnswad et a1 (1983, 1 %7), f'oonswad (1993a, 1993b), Poonsvad and Tw11 (1994) and Tsuji et a1 (19871, studies of Asian hornbills h a l t not focused on breeding biolog) (Leighton 1986) or h a \ e been lery limited in sample u t ' (Kdnndii 1094 0 Rrien 1997) Asidn hornbills gcnerdlli occui a t lei\ densities and nest high i n the ( m o p \ , creating logisticdl problems in data i OlltY The Sulawesi Red-knobbed Hornbill Aceros cassidix is found only on the Indonesian island of Sulawesi and associated offshore islands (White & Bruce 1986). The species is one of the largest (mean weight = 2.5 kg) and most brightly coloured forest hornbills and is known to occur a t extraordinarily high densities in some areas of North Sulawesi (5 1 hornbills/km2; Kinnaird et al. 1996). Sulawesi Red-knobbed Hornbills are monogamous, non-territorial and are syinpatric with only one other hornbill, the North Sulawesi Tarictic Hornbill Penelopides exarhatus exarhatus. Red-knobbed hornbills subsist largely on fruit (Kinnaird et al. 1996, Suryadi et. al. 1994) and range over areas up to 58 kni' during one year (Suryadi et al. in press, Kinnaird unpubl. data). Kinnaird and O'Brien (1993) presented the first and preliminary data on the breeding biology of the Sulawesi Red-knobbed Hornbill. In this paper, data are presented for a four-year study of the breeding biology of the Sulawesi Red-knobbed Hornbill. Nesting chronology, reproductive success, diet and nest-site characteristics are described for a large sample zyxwvutsrq zyxwvutsrqpon zyxwvu zy tior I c 1999 British Ornithologists Union Breeding of Sulawesi Red-knobbed Hornbills zyx zy 61 zyxwvu of hornbill nests. Such data are important for the prediction of future breeding trends, population status and management of this rare, endemic hornbill. METHODS Study area This research was conducted in the TangkokoDuaSudara Nature Reserve (hereafter referred to as Tangkoko) on the northernmost tip of the Indonesian island of Sulawesi (1 "34'N, 125"14'E). The reserve encompasses approximately 8867 ha and is broadly classified as lowland tropical rainforest (IUCN 199I), although the reserve contains examples of beach and cloud forest. The reserve is dominated by three volcanoes: Tangkoko, the recent ash cone Batuangus and the twin peaks of DuaSudara that rise to 1350 m. The study was conducted within a 5 k m 2 area on the north slope of the Tangkoko volcano that is characterized by a mosaic of habitat types and disturbance regimes (Kinnaird et al. 1996). Rainfall a t Tangkoko averages 1700 mmlyear (1992-94) and falls primarily between November and May. chronology was analysed on a weekly basis for each year. The week during which a female began sealing was considered the first week of nesting; therefore, the same nest week may occur over different calendar weeks for various birds. Annual data were combined for all nests by week to estimate incubation period, length of female incarceration and number of days to fledging. Although nest initiation was not monitored for all nests in 1992, two pairs were observed from courtship to fledging of young and the data from these pairs were used to backdate chronology of nesting for pairs not observed from courtship. Breeding bird density was calculated using the number of active nests for 1994 and 1995, the years for which nest surveys were most complete. Observations were made using binoculars (1 0 x 40) or telescopes (~35-60) from hides built 10-50 m away from nest trees. Activities in and around the nest tree were recorded and the time of arrival, departure and sex of birds visiting the nest tree were noted. If a visiting bird delivered food to the nest, the time of initiation and completion of food delivery and the number of items delivered was recorded. Redknobbed Hornbills generally collected more than one food item before a load was delivered to the nest. Items stored in the throat pouch are regurgitated and passed singly into the nest allowing identification and counts of items delivered. Food items were categorized as fruit or insects. Fruit was classified by colour and species and were scored for length of the longest dimension (1: 210 mm, 2: 11-20 mm, 3: 21-30 mm or 4: >30 mm). During 1992 and 1993, dropped fruits and regurgitated seeds were collected for identification from burlap slurts attached to the base of nest trees. Voucher specimens were submitted to Herbarium Bogorense, West Java and most identifications were available by 1993. Estimates of fruit biomass delivered to the nest were calculated for 1993 and 1994 by multiplying the mean fruit wet mass (O'Brien unpubl. zy zyxwvut zyxwvuts zyxwvutsrqpon Nest observations Data were collected during four breeding seasons (Table 1). At intensively watched nests, nest chronology, breeding season diet and reproductive success were monitored. Nests were observed for three hours in the morning (06.00-09.00 hours) and three hours in the afternoon (13.00-16.00 hours) each week during 1992. During 1993 and 1994, nests were observed once each week for three hours, with morning and afternoon watches alternating weekly. The status of additional nests (termed non-intensive nests; Table 1) was checked a t one- to two-week intervals to improve estimates of nest chronology and nesting success. Nest Table 1. Numbers of Sulawesi Red-knobbed Hornbill nests, monitoring interval and total hours of observations over four breeding seasons. Year refers to the year during which breeding was initiated. Intensive: observed weekly on a systematic basis for nesting chronology, diet and reproductive success. Non-intensive: checked at least biweekly to supplement estimates of nest chronology and nesting success. Number of hours observed refers to sampling of intensive nests only. Number of nests Year Dates Intensive Non-intensive Hours of observation 1992 1993 1994 1995 20 Aug 92 to 10 Jan 93 25 Jun 93 to 21 Jan 94 17 Jun 94 to 21 Dec 95 15 Jun 95 to 30 Oct 95 10 21 20 5 18 41 60 561 1140 915 - - zyxw @ 1999 British Ornithologists' Union, Ibis, 141, 60-69 62 zyxwvutsrqp zyxwvutsrqponm M. Kinnaird & TG. O'Brien data) b> the number of item5 deli1ered for each species and then summing across species Fruits not identified during deli\er) (<1O'h of all food deli\eries for both vearsj \+tare assigned the dierage \vet \\eight of known fruits in their s i ~ eclass This method assumed that unidentified fruits represented a random sampling of fruits 1% ithin a gi\ en size class Insect4 were not considered i n food biomass calculations Because there \iere no signifit. ant differences betlveen morning and afteri i o o i i obscr\ ations in the number of items deli\ ered to the nest i n 1993 ( F = 0 84, ns) and time of day for 1 OW nplainrd less than 0 9 "of~the \ ariance in numhcr of items deli\ rred ( F ,,= 3 49, P < O US), morning and aftei noon dat,i 11ere combined for analysis 168 species in 22 plots, each 0.25 ha in area. An additional 2.1 k m transect was established specifically to monitor fruiting of 52 extra fig trees because figs lvere under-represented in the plots. For each fruiting tree, the number of fruits was estimated on an exponential scale, following Leighton (1 993). Monthly fruit biomass was calculated by multiplying a species' mean fruit wet mass by the combined fruit crop for all individuals of a given species, then summing across species and dividing by the total number of hectares sampled (5.5 ha). zyxwvutsr zyxwvu zyxwvutsrqp zyxwvuts zyxwvut Fruit availability I he a\ dilbilit\ of fruit resources was estimated each month from Lkiembei 1992 through December 1994 Vvv monitored 201 5 trees for fruit representing A total of 63 nest trees known to be used for at least once in the four years of the study were mapped to the nearest 0.25 ha. All but one nest tree were identified to species; 43 trees were measured for tree height, nest cavity height, diameter a t breast height (dbh) and direction of cavity opening. Height measurements \Yere made using a clinometer and, when necessary, adjusted for topography. Five nest trees were climbed to measure the internal dimensions of the cavities. To evaluate preference for particular tree species by Redknobbed Hornbills, all trees greater than or equal to the minimum dbh used for nesting (54 cm) were enumerated in 96 plots of 0.25 ha each. Selectivity was then evaluated by comparing the frequency distribution of tree species in the nest sample to that of the study area. zyxwvutsrq ' a 6 - I RESULTS L 8 - I zyxwvut : c 64 2 - 0 - E Nest tree characteristics 300 Nest chronology ~- .d J S N J M M J S N J M M J S N J M M J S 92 93 95 Month Figure 1. Number of Sulawesi Red-knobbed Hornbill nests from which chicks fledged (a), number of nests from which female emerged (b). number of nests initiated (c) and rainfall by week (d) for four breeding seasons. a 1999 British Ornithologists' Union The Sulawesi Red-knobbed Hornbill breeding season spanned 27-30 weeks (June to January). In 1992, 1993 and 1994, females began nesting in mid- to late-June with the last chicks fledging in late December to midJanuary (Fig. 1). In 1995, females did not initiate nesting until late July and early August. In the first three years, initiation of nesting appeared to be stimulated by a dramatic reduction in rainfall; continuation of heavy rainfall into the dry season of 1995 may have delayed nest initiation. During 1992 and 1993, chicks fledged during periods of high food availability (Fig. 2) and a similar pattern was obscrvcd in 1994. Although resource availability was not quantitatively measured at the time of fledging during 1994, an independent fig phenology (Suryadi et al. in press) showed a peak in fig biomass in January, suggesting that fig fruits, the most commonly eaten food, were abundant during zyxwvutsrqponm Ibis. 141, 60-69 zyx zyxwvutsrqpo zyxwvu Breeding of Sulawesi Red-knobbed Hornbills 1200 - +Chicks fledge - - 1 - Females exit -1- .I.-.Females enter 63 Table 2. Duration of female Red-knobbed Hornbills remaining in nest cavities, chicks remaining in nests afler the female emerged and overall nesting period by year. Results are mean f se. Yearly differences are not significant ( P > 0.05). - Female Initiation emergence of nesting Female to chick to chick in cavity (days) fledging (days) fledging (days) E 400 Year ~- 1992 ~~~ 9 3 f 18 ~~~~~ 32 f 16 133 NDJFMAMJ JASONDJFMAMJ J A S O N D 93 zy zyxwvutsrq 94 Month Figure 2. Dates (mean and range) when female Sulawesi Redknobbed Hornbills entered nest cavities, females exited cavities and chicks fledged with monthly estimates of fruit biomass for species. Monthly fruit biomass was fig (M) and non-fig (0) calculated by multiplying a species' mean fruit wet mass by the combined fruit crop for all individuals of a given species, then summing across species and dividing by the total number of hectares sampled (5.5 ha). fledging in1994. The length of the nesting period, measured as the number of days from initiation of sealing by a female to the fledging of the chick, did not vary greatly between years (Table 2); the overall average for 1992, 1993 and 1994 combined was 139 days. Females sealed or partially sealed themselves in nest cavities using their own faeces which was composed primarily of fig seeds and insect chitin. Males did not participate in nest sealing. Once females initiated sealing, they remained in the nest cavity and were fed by the male through a narrow slit in the seal for an average of 93-1 14 days, depending on year (Table 2). It was impossible to tell when eggs were laid, but we assumed egg-laying occurred just prior to sealing because nestlings were visible in most of the easily viewed nests within 40 days after the seal was complete, an interval that accords well with the incubation period from 1993 110k20 2 4 f 18 140 i 15 1994 1 1 4 f 10 28i11 144 f 12 zyxwvu Combined 108 f 10 28 f 9 141 k 10 captive pairs (32-35 days, J. Primm, Audubon Zoo, pers. comm; 34-36 days, E. Kowalczyk, Woodland Park Zoo, pers. comm.). Nestlings remained in the cavity an average of 28 days after the female emerged (all years combined; Table 2) and continued to be fed by both adults until fledging at approximately 100 days of age. Females never re-entered the nest cavity after emergence. Nesting success Active nest densities were high in 1993 and 1994, those years during which all nests in the study area were believed to have been located. In 1994, 52 pairs nested within the 5 k m 2 study area (10.4 breeding pairslkmz) and in 1995, 41 pairs nested within the study area (8.2 breeding pairs/km'). Hatching success, fledging success and overall nesting success were high in all three years (Table 3). Hatching success, measured as the proportion of sealed nests known to produce a chick, ranged from 89% in 1994 to 90Yo in 1992. Although Sulawesi Red-knobbed Hornbills are capable Table 3. Hatching, fledging and nesting success (defined in text) by year for Sulawesi Red-knobbed Hornbills in the TangkokoDuaSudara Nature Reserve. Data for 1994 include information from nests located in a 1 kmz area outside the main study area. zyxwvutsr Hatching success -~ Nesting success ~~ Year Nests Hatched 1992 10 1993 1994 Combined Fledging success Yo Nests Fledged % Nests Fledged Yo 9 90 9 8 88 10 8 80 39 35 90 35 31 89 39 31 80 61 54 89 54 48 89 61 48 79 110 98 89 98 87 89 110 87 80 zyxw zyx 01999 British Ornithologists' Union, his, 141, 60-69 zyxwvutsrqpo zyxwvutsr zyxwvutsrqpo M, Kinnaird & TG. O'Brien 64 Table 4. Percentage of Sulawesi Red-knobbed Hornbill nests reused in consecutive years for nests monitored up to four years. Number of years monitored UFigs U N o n - f i g fruits -Unknown =Insects zyxwvutsrqponm zyxwvutsrqp Reuse ("L) when nest active for 4y 3y 2y 1y Number of nests reused 50 33 8 8 12 62 27 12 26 ~ 4 3 zyxwvutsrqponm zyxwvutsr 2 65 38 23 I / of Ici\ing u ~ 'to three eggs (E Ko\valc~yk,Woodland Park Loo pers comm ) onl! one chick per nest \ \ a 5 t'r ohxm id t o fledge Kemp (1 995) reported for sc\ era1 Ixge hornbills that extra eggs rarely hatch or that hatLhling5 from second or third eggs die n i t h i i i the t i n t few da; s Because \\ e could not ascertain the number L)f eggs laid per nest, v e used one egg per nest t\ to estim,ite the proportion of nests that hatched eggs Fledging wcces>, measured as the percentage of nests I\ ith c h i c k s t h a t successfull, fledged 1oung, \vas approximatel) 88"1)in all y ears Nesting success, r i i r ~ i \ u r4t' as thcb percentage of initiated nests that \ucCtwfull\ t1edg:td 5 oung, ranged from 79"o in 1994 10 SO"! 1c302 i'1 Repex laking h e r nest failure \ \ a s ne\ er obser\ ed during die stud\ Repeated use of nests bet\\ een years, h o i \ t ~ t 1\ \ a s common (Table 4) Totals of 50"~ of the nests o l m w e d f n i four years and 62"11 of the nests { h s mc 11 t o r thrcr I e m Isere occupied coiitinuousl~, iis\urnin: that iiCst5 \\ere occupied b> the same pair i a c h \ t' i r 7his 'iswmption is supported b! nine identifirci birds that all returned to the same nest site for t h r c ~i o n ~ c c u tei ~4va1-s 1992 1993 Year 1994 Figure 3. Composition of the diet delivered to nests by male Sulawesi Red-knobbed Hornbills during 1992, 1993 and 1994. the most, both in terms of number of species and percentage of items delivered to the nest. Over three years, Ficus spp. accounted for 59-79"/0 (mean = 67.5%) of the diet (Fig. 3). The importance of figs to the hornbill diet is underscored by their contribution to the total fruit biomass delivered during nest observations in 1993 and 1994. Figs accounted for 81% (50.8 kg deli\-ered during observations) of the fruit biomass in 1993 and 81?h (68.4 kg delivered during observations) of the fruit biomass in 1994. Ficus spp. \\rere the top three diet species in year 2 and the top four diet species 1994. (Table 6). The mean visit rate to the nest and mean number of food items and fruit biomass delivered varied over the nesting period. The mean number of visits in 3 h increased weekly during 1993 and 1994 ( I , ? ~ ~=, 0.6'9, P < 0.001; Fig. 4a). Most of the variance in visits \sas due to weekly differences in visit rates rather than differences between years. After emerging from the cavity, females made fewer visits to t h e nest and delivered fewer items and total fruit biomass than males. After emergence, females contributed an average of 29% (24-35%, n = 40 nests) of visits to the nests for 1992-94 combined. Females gave an average of 14?'c1 of all items delivered and 14"/1 of total fruit biomass delivered to the nest after emergence for 1993 ( ) I = 15 nests). Similarly, in 1994 females delivered 18"k of the food items delivered and 15"h of the total zyxwvutsrqp Diet Fruits Ii m p iseJ t h c grcatcst proportion of the horn- hi11 di,t L h r i n g dll t h i w ncsting seasons (Fig 3) Ripe h i t s ioinprised an a c r a g e of 89"~ (86-94"o) of all items di 111 ered The majority of-ripe fruits deh\ ered t i p t oiiiprised an a\ erage of 6X"o (59-79"o) trurt\ & l i \ c n d Insects comprised, on a\ erage, 0111) I " (04 2f'o)of the dirt I'nidentihed ittms aleraged I O'1 i 2"o) of thc. dict and included fruits that could not bt. i 1 a w f i c J SI, fig or non-fig (5')o) and items that i t t*r( not \ i d d c during deli\ ery (5%) At lc.qst 52 fruit spccics, rcprcsenting 20 genera from I2 tamiiic.5, \I ei c deli\ ered over the threv nesting seamiis (Table 5) I'he plant famil} Moraceae contributed cre tigs. ot (I c] 1999 Br tish Ornithologists Union zyx zyxwvutsrqpo Ibis, 141, 60-69 Breeding of Sulawesi Red-knobbed Hornbills Table 5. Number of fruit species by family and genus delivered to Sulawesi Red-knobbed Hornbill nests during three breeding seasons. Number of species Family and genus 1992 1993 1994 Total number of species 1 1 1 0 1 0 1 1 1 1 1 2 1 3 1 4 1 0 0 1 2 1 2 2 1 0 1 1 0 1 1 1 2 2 2 2 1 1 1 1 1 1 0 1 2 0 0 1 11 16 16 2 1 1 1 2 1 1 1 2 1 1 1 2 3 1 3 1 3 36 2 0 37 1 2 39 2 4 52 65 Table 6. Biomass estimates for the top five food species delivered to Sulawesi Red-knobbed Hornbill nests during nest observations in 1993 and 1994. Mean fruit mass represents ripe fruit pulp only, with the exception of small-seeded figs which were weighed together with the pulp. Mean fruit wet mass (9) Species Anacardiaceae Koodersiodendron Buchanania Annonaceae Cananga Polyalthia Apocynaceae Alstonia Burseraceae Canariurn Euphorbiaceae Drypetes Gnetaceae Gneturn Lauraceae Cryptocarya Meliaceae Aglaia Chisocheton Dysoxylurn Melia Moraceae Ficus Myristicaceae Gyrnnocranthera Knerna Myristica Horsfieldia Myrtaceae Syzygiurn Verbenaceae Vitex Unknown species Specieslyear zyx zy 1993 Ficus altissirna Ficus benjarnina Ficus forstenii Cananga odorata Cryptocaria celebica 1994 Ficus benjarnina Ficus altissima Ficus sp. 83 Ficus sp. 89 Cananga odorata Biomass (kg) % in diet 5.36 1.80 2.32 2.12 9.40 14.56 14.25 5.71 2.02 1.82 27.4 26.7 10.7 4.0 3.6 1.80 5.36 7.80 1.13 2.12 18.17 13.92 7.55 4.57 2.94 28.8 22.1 12.0 7.3 4.7 zyxwvut zyxwv z 3- 19 cn 2- .-c -0 a, a, LL y = 1.08 + 0.24year P = 0.69 1 + 0.07week b y = 52.8 + 0.15week + 0.75week2 - 0.03week3 zyxwvut zyxwvutsrq zyxwvutsrqp zyxwvutsrq zyxw 175 h 2oo h F .- 150 a, P = 0.61 C zyxw . 125 fruit biomass delivered to the nests after emergence (n = 16 nests). Estimates of fruit biomass were strongly correlated with the number of items delivered to the nest in 1993 and 1994 (T,,,, = 0.87 and T,,, = 0.82 for 1993 and 1994, respectively, P < 0.001). Fruit biomass delivered to the nest increased rapidly after the estimated mean hatching date and declined steadily once the female emerged (Fig. 4b). A polynomial regression of week on fruit biomass resulted in a third-degree polynomial as the best fitting line to describe the rise and fall of deliveries over the nesting period (F4,41 = 15.9, P < 0.001). No significant patterns in diet composition over v $ 100 m g a u 0 0 LL 75 - 25 50 0 1 0 2 4 6 8 10 12 14 16 18 20 22 24 Week of breeding Figure 4. (a) The relationship between feeding visits per threehour observation period by Sulawesi Red-knobbed Hornbills coded 1 in regression) and 1994 and week of nesting for 1993 (0, (M,coded 2) and (b) the relationship between food biomass delivered per visit and nesting week for 1993 (0)and 1994 (M). @ 1999 British Ornithologists' Union, Ibis, 141, 60-69 66 zyxwvutsrqpon zyxwvutsrq zyxwvutsrqpo M. Kinnaird & TG. O'Brien zyxwvuts zyxwvutsr zyxwvutsrq Table 7. Measurements of 43 Sulawesi Red-knobbed Hornbill nest trees and five nest cavities. Measurement Diameter at breast height (cm) Mean f sd Range 116.9 i. 41.2 54-210 Tree height (m) 40.2 f 9.67 19.7-58 Cavity height above ground (m) 26.2 k 8.25 12-53 Internal cavity dimensions Height (cm) Width (cm) Depth (cm) 81.4 2 28.5 32.6 2 11.8 50.0 10.3 58-228 10-32 38-58 * zyxwvuts the nesting period were found. Dietary diversity (measured as the Shannon-Wiener Function, H'; Pielou 1975) and the percentage of figs in the diet displayed no consistent trends over the nesting period. Similarly, there were no significant patterns in fruit availability (fig or non-fig) and the diversity of fruit available in the environment during the breeding season, suggesting that variation in feeding visits and food biomass delivered to the nest (Fig. 4) were unaffected by patterns of fruit availability. Nest tree characteristics All hornbill nests were located in natural cavities of large (meandhh= 116.9 cm) living trees 12 to 53 m above the ground (Table 7). Hornbills do not choose cavities facing particular directions; cavity orientation was evenly distributed across the eight major cardinal directions ( ~ 2 , = 9.27, ns). The 63 nest trees were represented by 15 species, although 43% ( n = 27) of the nest trees were Palaquium amboinense. Based on species representation in the habitat of the top four canopy trees, nest tree species were not selected randomly ( ~ =224.4, ~ P < 0,001). The major contribution to the chi-squared statistic is due to the overrepresentation of Palaquium amboinense and an underrepresentation of Cananga odorata as nest trees. DISCUSSION Nesting chronology The hornbill family Bucerotidae is characterized by an incubation period closely correlated with body size which is followed by an unusually long nestling stage (42-90 days; Kemp 1995). Large hornbills are among the slowest breeding forest birds yet recorded (Kemp 1995). The estimated incubation period of 35-40 days 01999 British Ornithologists' Union, lbis, 141, 60-69 for Red-knobbed Hornbills falls within the range of the other hornbill species weighing more than 2 kg (37-46 days; Kemp 1995). Although the time females remain in the nest is highly variable (58-140 days; mean = 108 days), it encompasses the range reported for hornbills greater than 1 kg. The maximum recorded incarceration of 140 days for the Red-knobbed Hornbill however, is the longest reported for any hornbill. The Sulawesi Red-knobbed Hornbill initiates nesting with the cessation of rains and, when rains continue into the normal dry season, they delay nesting or fail to nest. Nest failure and nest abandonment have been reported for the Rufous-necked Hornbills Aceros nipalensis and Great Hornbills Buceros hicornis in Thailand when heavy rains extended into the normal dry season (P. Poonswad pers. comm.). Most authors agree that the onset of nesting in hornbills is timed to coincide with rainfall and/or an increase in the food supply (Kemp 1976, 1979, 1995, Leighton 1986, Poonswad et al. 1987), but the timing and relative importance of each may vary. For African Tockus spp. inhabiting highly seasonal savannah habitats, nest initiation is timed to coincide with the onset of rains which bring a flush of invertebrate prey and fruit (Kemp 1976). In Thailand, four species of hornbills initiate nesting during the cool period between the monsoon rains and the dry season, and excessive drought may result in reduced nesting (Poonswad et a1. 1987, Poonswad 1993). Leighton (1 986) studied seven sympatric hornbills in an aseasonal rain forest on Borneo and concluded that nesting was a supra-annual event timed to coincide with cyclical peaks in food supply. The peak in fruiting, however, apparently was triggered by a one-month drought suggesting that variation in rainfall may provide the initial cue to begin nesting (Leighton 1982). While lack of rainfall probably provides the proximate cue for breeding, food supply during and immediately after fledging may be the ultimate factor in the timing of breeding in Sulawesi Red-knobbed Hornbills and other Asian hornbills. The onset of rains in November appears to trigger a burst of flowering and lead to a pulse in fruiting a t the time of fledging in December and January (Kinnaird & O'Brien unpubl. data). Fruit resources, especially figs, peaked at the time of or shortly after, fledging in1 992-94 (Kinnaird et al. 1996, Suryadi at al. in press). Kemp (1979), Poonswad et al. (1987) and Kannan (1 994), working in different parts of Asia all report that food supplies are abundant a t the time of hornbill fledging. Among the hornbills that lack helpers, post-fledging parental care may be limited and chicks may require abundant food zy zyxwvu Breeding of Sulawesi Red-knobbed Hornbills resources as they begin to forage independently. Although older juveniles foraging with parents were seen, most Red-knobbed Hornbill juveniles were foraging independently of the parents within four months of fledging. Diet and breeding effort Red-knobbed Hornbills are among the most frugivorous hornbills (Kemp 1995) with 86-93% of the breeding season diet composed of ripe fruit. Nonbreeding season diets also are primarily ripe fruit (Suryadi et al. 1994). Diets composed largely of fruit are low in protein which may cause problems for maintenance and development (E. Dierenfeld pers. comm.). Developing chicks may require a diet with 15-30'1/0 protein (dry weight) for growth, whereas adults birds need 5-12% protein in the diet for maintenance (Robbins 1993). Birds on low protein diets but with enough carbohydrates and lipids for a positive energy balance require approximately 0.43 g N/kg" 7j/day for maintenance, but chicks may need three times as much (Robbins 1993). For this reason, frugivorous birds usually augment diets for developing chicks by feeding animal matter (invertebrates and vertebrates). The fact that Red-knobbed Hornbills augment chick diets with very little animal protein may help explain their extremely long developmental period. This may be true for other solitary nesting, frugivorous hornbills. The top five diet species used by Red-knobbed Hornbills (primarily figs; Table 6) average 7.3% crude protein (O'Brien & Kinnaird unpubl. data), a figure considered adequate for maintenance of adult body mass, but perhaps low for adequate growth (based on galliformes and small passerines; Robbins 1993). Figs are an excellent source of calcium, necessary for bone growth and feather development, and also phosphorus, necessary for all metabolic activities (O'Brien et al. 1998). Red-knobbed Hornbills may have evolved mechanisms to minimize metabolic nitrogen loss, or the favourable calcium-phosphorus balance in the diet allows hornbills to maximize assimilation of dietary protein. Anecdotal support for this idea comes from an examination of Redknobbed Hornbill and Great Hornbill nest 'doors' which are composed entirely of faecal material (Kinnaird & O'Brien 1993, Kannan 1994). Kanan analysed faecal composition of a nest door and reported a nitrogen content of only 40% that of chicken faecal matter, suggesting that hornbills are very efficient a t assimilating nitrogen. 67 Nesting density and nest site selection zyx Red-knobbed Hornbills nest at extraordinarily high densities (8.2-1 0.4 nests/km*), often in close proximity to one another (250 m). Small nesting territories, an abundant food supply (Suryadi et al. 1994, Kinnaird et al. 1996) and a surplus of suitable nesting sites probably account for the high concentration of nests in Tangkoko. Although not rigorously quantified, male birds appeared to defend territories of approximately 10 ha around the nest site (Kinnaird & O'Brien 1993) and often chased or called a t birds that flew over or moved close to the nest. Nest territory defence was not always successful, however, and a fruiting fig in a nesting territory often attracted many birds despite the most ardent attempts to defend the nesting territory. Leighton (1 986) reported that Rhyticeros (= Aceros undulatus and A m o s corrugatus) did not defend territories in Borneo, but Poonswad et al. (1987) reported that Wreathed Hornbills defend territories up to 100 m radius (3.14 ha) centred on the nest. Redknobbed Hornbills show little preference for nest site characteristics except for tree species. The preferred nest tree, Palaquium amboinense, has a hard wood that is useful for timber but is susceptible to heart rot which enhances cavity formation and makes it an ideal nest tree. Canopy-sized P. amboinense with dbh measurements greater than 54 cm (minimum size for a nest tree) occur at a density of 3.25 trees/ha (O'Brien & Kinnaird unpubl. data), indicating that a large number of trees are potentially available for nesting. Intraspecific competition for nest sites was never observed and in some years up to 2 1Yn of available nest cavities were unoccupied, further indicating that nest sites were not limiting. zyxwvu zyxwvutsrqpo Implications for hornbill population biology Several aspects of the breeding biology of the Sulawesi Red-knobbed Hornbill suggest that post-fledging mortality or juvenile dispersal is high. First, annual investment in breeding by Sulawesi Red-knobbed Hornbills is high, with pairs typically spending 50% or more of the year caring for young birds. Second, Sulawesi Red-knobbed Hornbills may nest consecutively for several years; annual nesting for three or more years is more common than in other species from Thailand or Borneo (Poonswad et al. 1987, Kemp 1995). Third, Sulawesi Red-knobbed Hornbills experience a low nest failure due to minimal risk of predation, no intra- or interspecific competition at the nest site and high fledging success (70-90%). This zyxw zyxw @ 1999 British Ornithologists' Union,Ibis, 141, 6 0 4 9 68 zyxwvutsrqp zyxwvutsrqpon M. Kinnard & TG. O’Brien investment should result in a high annual population growth rate b u t t h e population in Tangkoko has remained relativt’ly stable over t h e last 15 years (O’Brien & Kinnaird 1996).Because demographic data are lacking for Red-knobbed Hornbills (and all other hornbills, Kemp 1995), ~ v t ‘ c a n n o t differentiate hetween the effects of mortality and dispersal o n population stability. In a long-lived species like a hornbill, the juvenile stage is the most likely period at which mortality is high. T h e r e are few predators of Sulalvesi Redknobbed Hornbills other than humans and disease does not appear to b e a major mortality factor (M.F. Kinnaird unpubl. data). Starvation risk, however, may bc a significant source of mortality of juveniles. Between D e c e m b e r a n d February, Sulawesi Redknobbed Hornbills increase their daily travel distance by nearly 50% (1 0 . 2 km to 15.2 k m ; Suryadi et al. in press, M.F. Kinnaird unpubl. data), increasing the energetic demand on juvenile birds and the risk of starvation for inexperienced birds subsisting o n a fruit diet. For birds with low fat reserves, starvation can occur after only 1 t o 3 days of fasting (Robbins 1993). Alternatively, juvenile Sulawesi Red-knobbed Hornbills ma!. be dispersing to areas outside the reserve. If juveniles from Tangkoko successfully disperse, they will augment unprotected populations, underscoring t h e importance of the nature reserve. However, lowland forest area has decreased by over 60% throughout Sulaivesi over the last two decades (Whitten et al. 1987) and the remaining forest is highly fragmented with felv corridors between forest blocks. Such small forest patches may not contain sufficient densities of fruit trees on which Sulawesi Red-knobbed Hornbills rely. While it is difficult a t present to distinguish between t h e mechanisms of mortality and juvenile dispersal, t h e consequences are important in terms of reproductive success and long-term population viability for Red-knobbed Hornbills in North Sulawesi and should bc the focus of future studies. zy Kakauhe, Denand Kakauhe, Jukber Lamhaihang and Yopie Manderos are thanked for their many hours behind blinds and for scaling the mountain side to monitor nests. Finally, we thank Ellen Dierenfeld, Christine Shepard, Wendy Worth, Jamie Primm and, especially, Eric Kowalczyk for keeping us up to date kvith information about captive-bred hornbills. REFERENCES International Union for the Conservation of Nature. 1991. Atlas of Tropical Rainforests. International Union for the Conservation of Nature Special Publications. Gland: InternationalUnion for the Conservation of Nature. Kalina 5.1988. PhD thesis, Michigan State University. Kalina, J. 1989. Nest intruders, nest defense, and foraging behavior in the Black-and-whiteCasqued Hornbill Bycanistes subcylindricus. lbis 131: 567-571. Kannan, R. 1994. Unpublished PhD thesis, University of Arkansas. Kernp, A.C. 1976. Factors affecting the onset of breeding in African hornbills. In Proc. Int. Ornithol. Congr. 16: 248-257. Canberra: Australian Academy of Science. Kernp, A.C. 1979. A review of the hornbills: biology and radiations. Living Bird 17: 105-136 Kernp, A.C. 1995. The Hornbills. Oxford: Oxford University Press. Kinnaird, M. F. & O’Brien, T.G. 1993. Preliminary observation on the breeding biology of the endemic Sulawesi Red-knobbed Hornbill (Rhyficeros cassidix). Trop. Biodiv. 1: 107-1 12. Kinnaird, M.F., O’Brien, T.G. & Suryadi, S. 1996. Population fluctuation in Sulawesi Red-knobbed Hornbills:tracking figs in space and time. Auk 113: 431440. Leighton, M. 1982. Unpulished PhD thesis, University of California, Davis. Leighton, M. 1986. Hornbill social dispersion: variations on a monogamous theme. In Rubenstein, D.I. & Wrangham, R.W. (eds). Ecological Aspects of Social Evolution: 108-1 30 . Princeton, NJ: Princeton University Press. Leighton, M. 1993. Modeling dietary selectivity by Bornean Orangutans: evidence for integration of multiple criteria in fruit selection. Int. J. Primatol. 14: 257-313. Moreau, R.E. 1937. The comparative breeding biology of the African hornbills. Proc. Zool. SOC.Lond. 107A: 331-346. Moreau, R.E. & Moreau, W.M. 1940. Hornbill studies. lbis 14: 639-656. Moreau, R.E. & Moreau, W.M. 1941. Breeding biology of Silverycheeked Hornbills. Auk 58: 13-27. O’Brien, T.G. 1997. Behavioural ecology of the North Sulawesi Tarictic Hornbill Penelopides exarhatus exarhatus during the breeding season. lbis 139: 97-1 01. O’Brien, T.G. & Kinnnaird, M.F. 1996. Declining populations of mammals and birds of North Sulawesi. Oryx 130: 150-156. O’Brien, T.G., Kinnaird, M.F. , Dierendfeld, E.S., Conklin, N.L., Wrangharn, R.W. & Silver, S.C. 1998. What’s so special about figs? Nature 392: 668. Pulliarn, H.R. 1988. Sources, sinks, and population regulation. Am. #at. 132: 652-661. Pielou, E.C. 1975. Ecological Diversity New York: Wiley. Poonswad, P. 1993a. Aspects of the biology and ecology of some Asian hornbills. In Poonswad, P. & Kemp, A.C. (eds). Manual to the Conservation of Asian Hornbills: 76-97. Bangkok: Hornbill Project Thailand. zyxwvu zyxwvutsr zyxwvutsr zyxwvutsr This study xvas funded by the Wildlife Conservation Society and the National Geographic Society (grant no. 4912-92). We thank the Indonesian Institute of Sciences (LIPI) and the Directorati, General for Nature Preservation and Forest Protection (PHPA) for their sponsorship in Indonesia. We rspecially acknowledge Soetikno Wiryoatmodjo (PB/LIPI), Dedi Darnaedi (PWLIPI), Romon Palete (PMPNManado) and Graham Usher (IJSAID) for their continued support and assistance ivhile working in North Sulawesi. Suer Suryadi Assisted in data collection and brainstorming, and Pilai Poonswad dnd Andy Dobson made valuable comments on an earlier versions of the manuscript. Wisnu Halir, Nelman /i\ IPQP Rritich Arnjthnlnnictc’ I Ininn lhrc id4 & U O Breeding of Sulawesi Red-knobbed Hornbills Poonswad, P. 1993b. Unpublished PhD thesis, Osaka University, Japan. Poonswad, P. & Tsuji, A. 1994. Ranges of males of the Great Hornbill Buceros bicornis, Brown Hornbill Ptilolaemus tickelli and Wreathed Hornbill Rhyticeros undulatus in Khao Yai National Park, Thailand. lbis 136: 79-86. Poonswad, P., Tsuji, A. & Ngampongsai, C. 1983. A study of the breeding biology of hornbills (Bucerotidae) in Thailand. In Proc. Jean DelacourAFCB Symposium, Breeding Birds in Captivity: 239-265. Los Angelos, CA: International Foundation for the Conservation of Birds. Poonswad, P., Tsuji, A. & Ngarnpongsai, C. 1987. A comparative study on breeding biology of sympatric hornbill species (Bucerotidae) in Thailand with implications for breeding in captivity. In Proc. Jean Delacour/lFCB Symposium, Breeding Birds in Captivity; 250- 315. Los Angelos, CA: International Foundation for the Conservation of Birds. Robblns, C.T. 1993. Wildlife Feeding and Nutrition. San Diego, CA: Academic Press. Suryadi, S., Kinnaird, M.F., O’Brien, T.G., Supriatna, J. & Sornadikarta, S.1994. Food preferences of the Sulawesi Redknobbed Hornbill during the non-breeding season. Fop. Biodiv. 2: 377-383. zyx zy 69 Suryadi, S., Kinnaird, M.F. & O’Brien, T.G. in press. Home range and daily movements of the Sulawesi Red-knobbed Hornbill during the non-breeding season. In Poonswad, P. & Kemp, A.C. (eds). Proceedings of the Second lnternational Asian Hornbill Workshop, Bangkok, Thailand, 7996. Bangkok: Hornbill Research Foundation. Tsuji, A., Poonswad, P., & Jirawatkari, N. 1987. Application of radio tracking to study home ranging patterns of hornbills (Bucerotidae) in Thailand. In Proc. Jean Delacour/IFCB Symposium, Breeding Birds in Captivity: 316-351. Los Angelos, CA: International Foundation for the Conservation of Birds. White, D.M.N. & Bruce, M.C. 1986. The Birds of Wallacea (Sulawesi, the Moluccas and Lesser Sunda Islands, Indonesia): an Annofated Check-list. London: British Ornithologists’ Union. Whitten, A.J., Mustafa, M. & Henderson, H. 1987. Ecology of Sulawesi. Java, Indonesia: Gadjah Mada University Press. zyxwvutsrqponm zyxwvu zyxwvutsrqp zyx Received 5 December 1996; revision accepted 25 September 1997 01999 British Ornithologists’ Union, Ibis, 141, 60-69