AUSTRALIAN 152 BIRD WATCHER

AUSTRALIAN BIRD WATCHER 1995, 16, 152-160 Roosting Patterns of the Laughing Kookaburra Dacelo novaeguineae

by K. A. WOOD, 7 Eastern Avenue, Mangerton, N.S.W . 2500

Summary At Wollongong, New .SOuth Wales, a suburban roost of Laughing Kookaburras Dacelo novaeguineae was visited in four sampling periods, each of 47-65 nights, between October 1991 and September 1993. The roost-site was defined by 24 branches, 23 of which were in a row of five adjacent Spotted Gums Eucalyptus maculata. Some birds (maximum 9) were present on 178 of214 inspections (83%). There was an overall tendency for all Kookaburras at the roost to perch on the same branch, touching one another and facing the same direction. Roost-splitting on different branches occurred in two of the sampling periods and subsequently the number of birds in the group probably decreased. Data obtained on roost-splitting were incomplete because individual birds were not marked and only one roost-site was studied, but it is possible that this behaviour could indicate an unstable dominance hierarchy and may precede dispersal of some birds to other groups. Introduction Laughing Kookaburras Dacelo novaeguineae live in permanent territories as either mated pairs or family groups consisting of a mated pair and up to six auxiliary members (Parry 1973). These auxiliaries or helpers are the progeny of the breeding pair; they remain with their parents for several years and help incubate the eggs, brood and feed young of subsequent breeding seasons and defend the group territory. Within the family, the breeding male is dominant over all auxiliaries of the same sex and older male auxiliaries dominate younger males (Parry 1973). A similar dominance hierarchy exists among females. Eggs are laid in September or October. Between July and September, inter-group conflicts are most common as territory boundaries are adjusted in preparation for breeding. Some auxiliaries disperse from their natal territory at this time and become dominant breeding partners in other groups. In a two-year study near Melbourne, Parry (1970) found that family members roosted together in two or three regularly used trees throughout the year. In Sydney, Hindwood (1947) reported that five Kookaburras roosted on the same branch consistently from June to early October 1945. In the following weeks, he observed one bird incubating eggs and noted that the number of birds remaining on the usual roost branch was variable. Hindwood later concluded that the family group had split, leaving only the breeding pair and one helper to care for the young. In Tasmania, Prestedge (1993) reported six Kookaburras roosting together (touching) with four juveniles in the centre of the perched group and the parent birds bracketing them on opposite sides. Conversely, Shields (1994) saw a group of seven at a roost near Singleton, N.S.W., with the adult pair side by side on a eucalypt branch and the three helpers '10 metres away on a lower branch, close to two fledglings (still showing down)'. Mangerton is a 55-year-old suburb in Wollongong, N.S.W., with a mosaic of tree­ lined streets and natural public reserves. Kookaburras are common, perhaps because the landscape is dominated by 20-30 m eucalypts (see Blakers et al. 1984). Recent observations of four roosts near my residence revealed that the number of birds roosting at night was highly variable. Subsequently, I selected one particular roost and visited it as often as possible during four discrete periods between October 1991 and September 1993. The aim of the study was to determine (1) the fidelity of Kookaburras in the group to the particular roost, (2) whether a tree or branch or clump of trees was used and (3) whether sleeping birds clustered together or perched apart. I hoped to assess the stability of the dominance hierarchy within the group by regularly observing the spatial arrangement of roosting birds. VOL. 16 (4) DECEMBER 1995 Roosting of Laughing Kookaburra 153

m

g b p h ( d a e J f 12- rn l

(Sth) MANGERTON ROAD (Nth)

Figure 1. Spotted Gum roost trees and roost branches 'a' to 'x' used by Laughing Kookaburras during four sampling periods between October 1991 and September 1993.

Roost-site and methods The roost-site was a row of mature Spotted Gums Eucalyptus maculata growing in the front garden of a residential allotment (Figure 1). These trees were 23-27 m high [diameter at breast height (dbh) 22-54 em] with the crown foliage starting about 10 m above the ground. They were planted in a straight line so close to one another (spacing mostly about I. 6 m) that their foliage overlapped laterally. Based on numerous daytime searches for white faecal droppings on the ground and searches for roosting birds at night, I feel confident that no other trees within a radius of about 50 m were used as roosts. Counts of faecal dropping spots before and after the study revealed that a roosting Kookaburra defecates about 16 times a night (range 10-20, n=7). Unlike other workers, I was unable to find regurgitated pellets on the ground below any roost (see Eastman 1970). The site was visited mostly at night between 2000 h and 2200 h during four sampling periods as follows: ( 1) 52 of 69 nights between 3 October 1991 and 10 December 1991, (2) 50 of 51 nights between 14 May 1992 and 3 July 1992, (3) all47 nights between 30 October 1992 and 15 December 1992, and (4) all65 nights between 12 July 1993 and 14 September 1993. A further 55 night inspections were made mainly between these sampling periods to monitor changes in the roosting pattern. Observations were from a distance of 20-25 m with a hand-held spotlight and binoculars. On each AUSTRALIAN 154 WOOD BIRD WATCHER

v1s1t, I recorded the branches on which the ~ ~ N Kookaburras were perched, the approximate ~ compass directions that they faced and the estimated ~ spacings between adjacent birds. On seven =4> afternoons, I also looked for birds arriving at the ..... roost-site from before sunset until after dark . ! "" ...... s Civil twilight (sun 6 o below the horizon) is ::; ..... marked by the onset of street-lighting. ,gi .f!l"' Results 4> During the afternoon watches, all -~ ~ "' '"' N Kookaburras that were seen arrived and -a ~ \0 eventually settled in one of the trees ...... ~ depicted in Figure 1. Fourteen birds (70%) rl1= "' <:3< arrived during civil twilight, three arrived g before sunset (earliest 17 minutes before) ..... '"Q ~ and another three arrived just after civil = twilight (latest four minutes after). A wide ..... C) N g::"' variety of contact calls was given when Kookaburras arrived and when they moved ..... r- s:: about selecting a branch on which to fmally =~ ~ "''"' E settle. Most contact calls were given within ~ "tis:: a few seconds of landing on a branch but '"Q 4> ~- I was unable to determine whether the bird :g Cl:i which had just landed, or one of the others, 4> -'< made these utterances. The laughter song ..... 4> 4> -~ was heard three times: on 23 October 1991 ,-... :c ...... N by four birds four minutes after civil 4> E-o"' twilight, on 4 October 1992 by three birds ~ 0"1 (i; 16 minutes after sunset and on 12 September 1993 by two birds 13 minutes ~ ~ ~ $ after sunset. No additional Kookaburras

:~ ~ arrived after these laughter songs were 0 "" given. :=- ""-, 00 ~ A total of 24 roost branches was used, 4> .:"' lr) of which 23 were in five adjacent trees Col "' -.:t (Figure 1, Table 1). In each sampling = ..... period, the birds showed a distinct ,Q"' ~ -.:t - preference for roosting on one particular -"' 8 " branch. They used the 'e', 'd', 'r' and 's' - branches on 45, 41 , 36 and 23 nights "' -t:) respectively in periods 1, 2, 3 and 4 (Table ,Q-= 1). All branches used were between 15 and ...: <:I 8"' 24m high (83% between 15 and 20m). ~ Only three branches ('q', 'r' and 's') were - of dead timber. The preferred position on -="' live branches was on the central third, ~ where they were approximately horizontal toll ·a ..... N N ~ with estimated diameters from 25 to 50 0"1 0\ 0\ 0\ ...0 mm. On three occasions when winds were u .Q u

Oct.- Dec. 1991 May - Jul.1992 40.------, 40.------, n=52 n=50

Q) (.) c :;~ 20 20 8 0

0 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8

Oct.- Dec_ 1992 Jul- Sep. 1993 40 40.------~------~ n=47 n= 65

c Q) (.) 20 20 Q; c._

0 0+---~~-Y-~--,-,-,-.---~ 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 Number of roosting birds ·

Figure 2. Frequency of occurrence of various total nwnbers of Kookaburras at the roost­ site during four sampling periods between October 1991 and September 1993. moderate. On 12 September 1993, I watched two birds from the time they finally settled for the night in very light drizzle (1757 h) until they were asleep in light rain at 1827 h (civil twilight ended at 1806 h). The angle of their sleeping posture noticeably changed as the rain gradually became heavier, and they adopted a much more 'vertical' posture while perched in steady rain at 1827 h than during light drizzle at 1757 h. Before this study, the roosting positions of Kookaburras at the three other known roosts were 12 m above ground in a 20 m Blackbutt E. pilularis, 15 m high in a 22 m Blackbutt and 16 m high in a 24 m Bangalay E. botryoides. The number of Kookaburras at the roost-site varied nightly (Figure 2). Maximum and minimum numbers counted were: period 1 - 0-6 (median 4), period 2 - 0-8 (median 6), period 3 - 0-6 (median 3), period 4, 0-3 (median 1). On successive nights, it was usual to record different numbers of birds present, indicating that the territory held more than one roost-site. Overall, no birds were recorded on 36 of 214 inspections (17%). The longest runs of identical counts on successive nights in each period were: period 1 -six birds for: four nights, period 2- six birds for three nights, period 3 - five birds for three nights, and period 4 - zero birds for nine nights. Nine birds were observed once during the study, on 4 March 1993, when three faced north (two touching) on the 'n' branch, five faced south (some touching) on a branch about 2m above and a lone bird roosted some 7 m away on the 'q' branch. Fidelity to particular branches was estimated when a total of three or more birds was present and when a total of five or more was present (Table 2). Choice of these totals was arbitary. All Kookaburras roosted on the same branch for 63% of instances AUSTRALIAN 156 WOOD BIRD WATCHER

Table 2

Number of nights that all Kookaburras roosted on the same branch or otherwise (a) when three or more were present, and (b) when five or more were present. Oct.-Dec. 91 May-July 92 Oct.-Dec. 92 July-Sept. 93 Total (a) (b) (a) (b) (a) (b) (a) (b) (a) (b) All on same branch 27 18 31 25 10 1 1 0 69 44 On multiple branches 15 6 11 11 14 9 0 0 40 26 when three or more were present (pooled data, 69 of 109 nights) and coincidently for 63% of instances when five or more were present (pooled data, 44 of70 nights). When perched on the same branch, they preferred to roost side by side, touching (Table 3). Compact clustering in this manner was recorded in 67% of instances when three or more birds were on the same branch (pooled data, 70 of 105 nights, Table 3). A similar result was obtained when five or more birds were on the same branch (58%, 25 of 43 nights). If not touching one another, the alternative arrangements of all but one touching and in two compact groups with all touching in each sub­ group, were most common. When perched on the same branch in two sub-groups or with all but one bird touching, the estimated spacing between adjacent (non-contact) birds ranged from 25 mm to 360 mm (pooled data, mean=170 mm, n=26):Low scores in Tables 2 and 3 from July to September 1993 resulted from a paucity of birds at the roost during this period (three birds were seen only once, see Figure 2). Analysis of orientation data showed that there was an overall tendency for all birds roosting on the same branch to face the same direction (Table 3). When three or more were present, they all faced the same direction on 62 of the 105 instances recorded (59%). The overall tendency for five or more birds to face the same direction on the same branch was even greater (81%, 35 of 43 nights, Table 3). The only exception to such preference was in the period May to July 1992 when some of the three or more birds present faced opposite directions 25 times yet faced the same direction only 21 times (Table 3). From pooled data, all birds faced approximately north 25 times, south 20, east 11 and west 11 times respectively. When birds all faced the same direction, no particular compass point was preferred. Moreover, when some birds faced opposite directions, there was no preferred arrangement in their orientations with respect to one another.

Patterns within each sampling period The first sampling period was from October to December during the 1991 breeding season, but before detailed sampling began, searches for roosting birds were conducted every night from 9 to 19 September 1991. In this 11-night period, no birds were present once, one bird once, five birds thrice, six birds once, seven birds three times and eight birds twice. In the sampling period itself, a maximum of four birds was present until23 October, then five were consistently seen until 9 November (12 of 14 night inspections), after which six birds were often present (10 of 22 night inspections). Parry (1973) found that incubation and nestling periods were about 24 d and 36 d respectively and eggs were not laid before September. After allowing for the probability that one of the family was incubating eggs or brooding young between September and December 1991 , I was unable to determine how many birds were in the group from 9 September to 10 December or if group-splitting occurred, because numbers present at the roost fluctuated from four to eight without any obvious pattern. VOL. 16 (4) DECEMBER 1995 Roosting of Laughing Kookaburra 157

Table 3

Number of nights that Kookaburras huddled together touching or facing the same direction (a) when three or more were perched on the same branch, and (b) when five or more were perched on the same branch. Oct.-Dec. 91 May-July 92 Oct.-Dec. 92 July-Sept. 93 Total (a) (b) (a) (b) (a) (b) (a) (b) (a) (b) All huddled together touching 32 17 28 8 9 0 0 70 25 All but one huddled together touching 4 2 4 0 0 10 3 In two sub-groups with all touching in each group 15 12 5 0 0 21 14 Perched in some other arrangement 0 2 0 0 0 4 All facing the same direction 27 15 21 19 13 0 62 35 Some facing the opposite direction 11 4 25 3 7 0 0 43 8

In the second sampling period from May to July 1992, a maximum of eight birds was at the roost (see Figure 2), indicating that the group size was probably eight. However, pooling of data has masked a change in the roosting pattern. Whereas in the first 25 nights of this period, the group roosted always on the same branch (' d', Table 1) and mostly in two sub-groups (Table 3), the group split and roosted in two sub-groups about 10m apart on 11 of the remaining 25 nights (branches 'd' and 'r', Table 1). In the first and second halves of this sampling period, roost-splitting was evident on the same and different branches respectively. From October to December 1992, six Kookaburras were present only once (18 November, Figure 2). Allowing for one member to be incubating or brooding at night, these data suggest that the group size was either six or seven. One or two subordinates probably dispersed to another territory after the roost-splitting was noticed in period 2 (before breeding commenced). Moreover, from October to December 1992, group members split and roosted in two sub-groups on widely spaced branches ('i' and 'r', Table 1) more often than they roosted together as a cohesive unit on the same branch (14: 10 and 9:1, Table 2). That nine birds were seen at the roost in March 1993 suggests that two or three young may have been raised in the 1992 breeding season. In sampling period 4 (July to September 1993), the group size was only two or possibly three birds (Figure 2), suggesting that five or six group members from the 1992 breeding season dispersed to other groups, or split and formed a new group of their own before breeding began in 1993. Given that the roosting patterns observed in the other sampling periods were normal, it seemed unlikely that the five or six 'missing' birds were still part of this group but roosted for all65 consecutive nights at other roosts in the same territory as the studied roost. Within period 4, two birds were present on 25 nights, always on the same branch (touching each other on 23 nights). They were together on the 's' branch on 19 nights. AUSTRALIAN 158 WOOD BIRD WATCHER

Faecal droppings on Mangerton Road, Wollongong, N.S.W., under the · Laughing Kookaburra roost on 19 September 1991. Plate 35 Photo: K.A. Wood

Discussion Because birds were not marked and only one roost-site was studied in four sampling periods, conclusions about group size and group-splitting must be treated with caution. Deaths might have occurred without my knowledge, breeding was not investigated and subordinates from other groups might have been adopted in the group studied. Some roosting birds may have been present but were not detected and individual birds might have shuffled or moved about at night, particularly in orientation or proximity. I checked this aspect of behaviour only twice, and found that six birds were perched on the same branch with orientation and proximity arrangements at 1845 h on 15 September 1991 that differed from those at 2014 h on the same night, whereas four Kookaburras touched one another and all faced north at 1848 h and 2104 h on 23 October 1991. Eastman (1970) reported spotlighting a family of Laughing Kookaburras near Charters Towers, Queensland, several times one night and found that 'they frequently changed positions, head to tail, or tail to head, within the bundle' of birds on the branch. Notwithstanding these limitations, the observations reported provide additional insights into the roosting behaviour of the Laughing Kookaburra. Overall, it seems likely that some individuals were faithful to the roost-site during each of the sampling periods. The variation in numbers present accords with Parry's (1970) conclusion that two or three roosting trees are regularly used (presumably at different sites). It is not known for how long this roost was used before the study commenced. However, since completion of the study, numerous casual inspections for faecal spatter on the road below and 12 thorough inspections at night suggest that the site was used very infrequently by one (maybe two) birds until September 1994 and not at all from September 1994 to April 1995. Hindwood's (1947) report, which suggests that all five Kookaburras in the Sydney group roosted on the same limb for six months, may be misleading as Hindwood inspected the roost only 'on many occasions' , not every night. VOL. 16 (4) DECEMBER 1995 Roosting of Laughing Kookaburra 159

Roosting in tightly organised flocks is common among some landbirds in the Northern Hemisphere (Welty & Baptista 1988) and Australia (see Annels 1983, Kloot & Aston 1983), and is sometimes thought to represent heat-conserving behaviour. At the Mangerton most-site there was a tendency for all Kookaburras present to roost on the same branch, touching one another and facing the same direction. Such behaviour may be a thermoregulatory response but insufficient data were available to test for differences between clustering patterns at various ambient temperatures. The Mangerton Koookaburras certainly huddled together (touching) sometimes in summer and sometimes in winter. Alternatively, the nocturnal huddling observed may reflect the socially conditioned behaviour of this species when the group is stable, as among Australian passerines only the White-winged Chough Corcorax melanorhamphos is known to have helpers that assist with so many phases of breeding for a number of consecutive breeding seasons (see Rowley 1982). Roost-splitting can occur on the same branch, on different branches at a particular roost-site or by birds of the same group roosting at different roost-sites. The scale of spatial segregation at night might reflect the degree of instability within a group. In this study, all three types of roost-splitting were observed at one roost-site but it is not known if the split sub-groups always comprised the same family members, as individual birds were not identified by bands or markings. When birds roosted in two sub-groups on the same branch, the average spacing between adjacent (non­ contact) birds was only 170 mm (n=26). Birds roosting apart in sub-groups on the same branch at such small distances might indeed be in a stable hierarchy of dominance. Group-splitting is likely to be a regular occurrence among Kookaburras because particular helpers usually assist their parents for only one or two breeding seasons and are capable of breeding when one year old (Parry 1973, Morton & Parry 1974). Groups with more than eight birds are rare (Parry 1973, Reyer & Schmidl 1988). As the Mangerton group in 1991/92 was relatively large, it seemed probable that group-splitting was imminent. Roost-splitting on different branches may take place in stable groups but, conversely, it could be interpreted as instability in the dominance hierarchy. Such a pattern may precede dispersal of some birds to other groups. Reyer & Schmidl (1988) studied 16 groups of Kookaburras near Armidale, N.S.W. In two groups of five birds, all roosted together on some nights but split into two sub-groups of two and three on others. The spatial scale at which roost-splitting occurred was not mentioned. Reyer & Schmidl noted further that in groups of various size, all birds roosted together and intra-group attacks were rare if the dominance hierarchy was stable, whereas aggression was higher and roost-splitting occurred in unstable groups. In combination, the Wollongong and Armidale studies suggest that roost-splitting probably indicates an unstable hierarchy of dominance and may precede dispersal of some family members to other groups. That two Kookaburras roosted together on the same branch (mostly touching) on 25 nights in period 4 is consistent with this suggestion, as Kookaburra groups of two are most likely to be mated pairs in a stable dominance hierarchy. Furthermore, Prestedge's (1993) and Shields' (1994) observations, in Tasmania and New South Wales respectively, are not inconsistent with the suggestion, given that the Tasmanian group of parents and four juveniles was probably stable and roosted together whereas the Singleton group of parents, three helpers and two fledglings, may have been somewhat unstable and roosted apart. Wild birds instinctively choose well-protected roost-sites for sleeping (Arnlaner & Ball 1983), as did the Mangerton Kookaburras which roosted in branch positions where they were protected from weather and disturbance. The roosting height of 16-20 min 27 m eucalypts provided some canopy cover during rain and less movement of roost branches in strong winds than at higher positions. This height was also AUSTRALIAN 160 WOOD BIRD WATCHER sufficient to isolate the birds from human-related disturbance on the ground below yet offer protection from nocturnal avian predators flying above (see Fleay 1944). Unlike other localities, it could not be said that the roosting branch at any of the four roost-sites at Mangerton was 'usually in a cluster of leaves near the top' of the tree (Eastman 1970). Given that perching birds have flexor tendons that automatically clamp the digits around the branch when sleeping (King & McLelland 1984), a branch diameter of about 35 mm is perhaps the most comfortable for Kookaburras. In the study area, they roostedJilainly on the central third of live branches where the estimated diameter was 25 to 50 mm, a position which also minimises disturbance from nocturnal mammals. Common Brushtail Possums Trichosurus vulpecula are abundant in Mangerton (KAW unpubl. data) and were often seen in the trees at the studied roost.

Acknowledgements I wish to thank Ian Rowley, Brian Doherty, Max Edwards, Stephen Debus and an anonymous referee for their constructive comments on earlier versions of this report. References Amlaner, C. J. & Ball, N.J. (1983), 'A synthesis of sleep in wild birds', Belulviour 81, 85-119. Annels, R. (1983), 'Light intensity as a stimulating factor in clustering by Dusky Woodswal1ows Artamus cyanopterus', Aust. Bird Watcher 10, 69-72. Blakers, M., Davies, S. J. J. F. & Reilly, P. N. (1984), The Atlas of Australian Birds, Melbourne University Press, Melbourne. Eastman, W. (1970), The Life ofthe Kookaburra and other Kingfishers, Angus & Robertson, Sydney. Fleay, D. (1944), 'Watching the Powerful Owl' , Emu 44, 97-112. Hindwood, K. A. (1947), 'Nesting habits of the Kookaburra or Laughing Jackass (Dacelo gigas)', Emu 41, 117-130. King, A. S. & McLelland, J. (1984), Birds, Their Structure and Function, 2nd edn, Pitman Press, Bath. Kloot, T. & Aston, H.I. (1983), 'Night roosting of the Rainbow Bee-eater Merops ornatus', Aust. Bird Watcher 10, 104-105. Morton, S. R. & Parry, G. D. (1974), 'The auxiliary social system in Kookaburras: a reappraisal of its adaptive significance', Emu 14, 196-198. Parry, V. A. (1970), Kookaburras, Lansdowne Press, Melbourne. --(1973), 'The auxiliary social system and its effect on territory and breeding in Kookaburras', Emu 73, 81-100. Prestedge, G. (1993), 'Heads or tails', Bird Observers Assoc. Tas. Newsl. 23(6), 2-3. Reyer, H.-U. & Schmid!, D. (1988), 'Helpers have little to laugh about: group structure and vocalisation in the Laughing Kookaburra Dacelo novaeguineae', Emu 88, 150-160. Rowley, I. (1982), Bird Life, 2nd edn, Collins, Sydney. Shields, J. (1994), 'Laughing Kookaburra' , in Strahan, R. (Ed.), Cuckoos, Nightbirds & Kingfishers of Australia, Angus & Robertson, Sydney. Welty, J. C. & Baptista, L. (1988), The Life of Birds, 4th edn, Saunders College, New York . Received 6 December 1994 •