VOL. 10 (2) JUNE 1983 69

Light Intensity as a Stimulating Factor in Clustering by Dusky Artamus cyanopterus

By ROSS ANNELS, c/- A. R. Annels, 9 Duffield Street, Manjimup, Western 6258

Introduction The clustering habit of the Dusky Artamus cyanopterus has been recorded for many years. Various workers have attempted to explain the phenomenon. Among the reasons put forward have been: as a thermoregulatory function (Marshall 1957), shelter from weather (Cooper 1972), and as a social form of camouflage (Wood 1970). All the previously published records of clustering Dusky Woodswallows refer to them clustering either in the evening (Dove 1909, Chisholm 1929, Hindwood 1956) or in daylight during overcast or squally conditions (Sedgwick 1964, Cooper 1972, Wall 1976). Considering these observations it seems that the various reasons put forward apply in certain situations but do not appear to fit all recorded cases of clusters. Immelmann (1966) observed clusters of Woodswallows in the north of Australia on nights when the temperature did not drop below 30°C and this does not seem to support the thermoregulatory theory. Undoubtedly in some situations would gain some advantage from their mutual body heat. Wood (1970) suggests that clustering of Woodswallows is a form of social camouflage and, although this seems feasible at night when roosting birds could benefit from mutual camou­ flage against predators, it is difficult to envisage why, when daylight clusters are involved, birds should find any more need for social camou­ flage at the particular time of clustering than at any other time of the day. Shelter from weather (Cooper 1972) seems to apply for clusters during daylight but this does not explain why Dusky Woodswallows should cluster to roost on fine evenings. Apparently, then, some other factor is also involved that applies to the clustering of Dusky Woodswallows. Stokes and Hermes (1979) suggest a 'photosensitive endogenous rhythm' for Black-faced Wood­ Artamus cinereus where the dispersal of overnight clusters was closely related to a certain light intensity. Since all the published records of the clustering of the Dusky Woodswallows are in the evening or during overcast or squally conditions in daylight, I formed the hypothesis that the clustering of the Dusky Woodswallow is stimulated by a particular level of light intensity.

Method To test the hypothesis it was necessary to observe clusters, both in daylight and in the evening. Flocks of Dusky Woodswallows around the Manjimup area were observed every available evening for nearly two months until the roosting site was found. Close scrutiny of flocks of AUSTRALIAN 70 ANNELS WATCHER

Woodswallows during the day, regardless of weather conditions, led to the discovery of several daylight clusters on wet overcast days. Observations were made regarding time, number of birds, temperature, cloud cover, rainfall, strength of wind and light intensity at the time of clustering. The light intensity readings were taken, to the nearest calibration, on a hand-held Sekonic Studio Deluxe exposure meter. Results In all I observed nine clusters of which four were in daylight and the other five in the evenings. The details of each of these are included in Table I. It is worthy of note that the clusters were large, ranging from 47 to an estimated 200+ individuals. The largest numbers of birds were involved in daylight clusters. This differs from other published observa­ tions of daylight clusters; Wall (1976) records six birds, Cooper (1972) twenty birds and Sedgwick (1964) thirty birds.

Table 1 Environmental Conditions at Time of Cluster Formation

Day 12 26 26 26 26 27 29 30 3 Month in 1980 July July July July July July July July Aug.

Temperature •c 8 6 9 9 8.5 ... 11.5 13 14 Cloud cover (x/ 8) 6-7 8 7-8 7-8 dark 7 6 7 5 Wind strong none none none none light none none light breeze breeze breeze Rain Yes Yes Yes No No Yes No No No Light intensity foot-candles * * 120 120 * 120 120 120 120 Time of formation 1555 1020 1625 1643 1732 1605 1725 1735 1732 Estimated no. of birds in cluster 200+ 150- 150- 150- 150- * 47 200 * 200 200 200 200

Data una vailable.

The behaviour of the flocks of woodswallows studied varied from one occasion to another, however I have included the following observations from two occasions as these seemed fairly typical of behaviour. On 26 July the flock was observed mainly contact perching with indi­ vidual birds hawking for between the . At 1625 h a dark cloud passed overhead upon which the entire flock took to the air. The birds grouped closely and flew rapidly from to tree. Eventually selecting a Marri calophylla the birds flew directly to the near vertical trunk and shuffled together into a cluster (see plate 22). While in the air the birds were very vocal, using a whistling flock call, and they continued this calling after the formation of the cluster. At 1555 h on 12 July I observed a cluster of Dusky Woodswallows 24 metres up on the north-eastern face of a Blackbutt Eucalyptus patens. The day was dismal and overcast with a strong breeze blowing in gusts from the south-east and it was raining quite heavily at times. The cluster VOL. 10 (2) JUNE 1983 Clustering of Dusky Woodswallows 71

Cluster of more than 200 Dusky Woodswallows on a Marri Eucalyptus calophylla. Plate 22 Photo: Ross Annels involved 200 or more birds all individually gripping the bark. The majority of the birds had their heads facing up the tree but the birds around the edge of the oval-shaped cluster had their heads facing towards the middle of the mass. All that was visible of each individual bird was its tail and wing-tips as the rest was buried in the mass. All the time the birds kept up a soft whistling call. Occasionally a bird would land on the backs of its companions but it would soon leave the unsteady perch and grip the more solid trunk. As I watched the cluster it moved slowly down the tree as birds at the top would fly off and resettle at the bottom. Over a period of 35 minutes the cluster moved four metres down the tree into a more sheltered position under a branch surrounded by foliage. Discussion It seems fairly clear from the clusters studied that some reasons put forward for the clustering behaviour by other workers are not applicable in these situations. Thermoregulation, as suggested by Marshall (1957), even under the cool conditions experienced in the winter in south-western Australia, seems to be invalid as no trend appears to connect temperature and AUSTRALIAN 72 ANNELS BIRD WATCHER

clusters. On several occasions while the birds were under observation temperatures dropped lower than those recorded during clusters. How­ ever it cannot be denied that some benefit would be gained on some occasions from mutual warmth obtained during clustering. Protection from weather (Cooper 1972) cannot be ascribed to all situations, even those occurring in daylight, because in some cases there was no wind or rain. As with thermoregulation some advantage may be gained on some occasions from clustering together, sheltering from the weather. However the light intensity was constant at 120 foot-candles and even given that the observations were made over a limited period of time this points strongly to light intensity as the stimulating factor. Other evidence to support this theory can be found in accounts of daylight clustering. Wall (1976) notes, 'At about 1945 hours a light south­ westerly drizzle commenced and almost immediately six birds were seen to cluster . . . ' Such a drizzle could be associated with a dark rain cloud causing a drop in light intensity. Sedgwick (1964) records a cluster during a rain squall. He says, 'They remained bunched up for up to at least 15 minutes, then dispersed soon after the rain stopped.' The 'soon after' suggests that the birds may not have been seeking shelter from the rain, rather that they may have been waiting for the light intensity to increase. Cooper (1972) records a cluster on 'an overcast day' which points to a low light intensity. This evidence suggests that clustering of Dusky Woodswallows is stimulated by light intensity.

Summary It was postulated that light intensity was the stimulator involved in clustering of the Dusky Woodswallow Artamus cyanopterus. This was investigated by observation of clusters and a value of 120 foot-candles was found to be the intensity stimulating Dusky Woodswallows.

Acknowledgements I would like to thank Graeme Chapman, Ian Rowley, Dick and Molly Brown, Tony Annels, and Greg Strelein for their assistance.

References Chisholm, A. H . (1929), Birds and Green Places, pp.59-61 , London. Cooper, R . P. ( 1972), 'Daylight Clustering of the Dusky Woodswallow', Australian Bird Watcher vol. 4, 114. Dove, H. S. (1909), 'Woodswallows Clustering', Emu vol. 9, 93-95. Hindwood, H. A. (1956), 'Clustering of Woodswallows', Emu vol. 56, 165-186. Immelmann, K. (1966) , ' Observations on Woodswallows', Journal fiir Ornithologie vol. 107, 37-69 (translation) . Marshall, A. J. (1957) , 'On the Function of Clustering in Woodswallows', Emu vol. 57, 53-54. Sedgwick, L. E. (1964), 'Birds of the Stirling Ranges, Western Australia', Emu .val. 64, 1~ . Stokes, J. & Hermes, N . (1979), 'Cluster Roosting in the Black-faced Wood­ ', Emu vol. 79, 84-86. Wall, L. E. (1976), 'Dusky Woodswallows Clustering in Tasmania', Australian Bird Watcher vol. 6, 148-150. Wood, V. J. (1970) , 'Observations of Clustering of Little Woodswallows', Sunbird vol. 1, 59-64. •