Distribution and habitat use of Crested Larks Galerida cristata at Loiyengalani, Kenya Leon Bennun, Luca Borghesio, Luca Biddau, Edward Waiyaki and Luc Lens The Crested Lark Galerida cristata is a widespread and polytypic species (Cramp 1988). In Africa, the bird ranges widely to the north and south of the Sahara, along the Nile Valley, and — — the race Kenya, southeastern Ethiopia, and northern Somalia (Keith et al. 1992). In Kenya, Crested Larks are common in sandy semi-desert in the northwestern quarter of the country (Lewis & Pomeroy 1988). The species is considered mainly resident (Cramp 1988, Keith et al. 1992). In Europe, family units persist until autumn (Dolgushin et al. 1970) and movements away from the natal area involve wandering juveniles (Labitte 1957). Few data are available on Crested Lark densities and distribution patterns outside the breeding season. Scattered observations mainly involve single individuals or small flocks (Abs 1963, Kozlova 1975, Viulleumier 1979). Temporal aggregations and short-distance movements are assumed to be related to bad weather (Porter et al. 1969), abundant food supply (Cramp 1988) and/or water supply (Rustamov 1958). somaliensis in northern In this paper, we (i) introduce a new census technique, specially designed for standardized counts of ground-dwelling passerines in arid areas, and (ii) analyse non-breeding population densities and spatio-temporal distribution patterns of Crested Larks in relation to the characteristics of semi-desert habitat near Loiyengalani, Lake Turkana. Study area and methods Lake Turkana, 250 is km long and reaching the Ethiopian border along its northern end, the largest of the lakes in the region is April) and high, with 1992, Kenyan Rift Valley characterized by extreme dryness (<200 often subject to long periods of drought is mean maxima over 32°C (Hopson maximum area was some Hughes 1992). The (Hopson 1982). Temperatures are 1982); during our study in February be considered a semi-desert 3.2 (Pratt et al. 1966). in the southeastern part of the basin, close to the fishing village of Loiyengalani (2°46N, 36°43E). shoreline & daytime temperature ranged from 35 to 39°C. With these climatic features, the region can Our study (Hughes mm annual rainfall, peak March- km long, We and inland censused Crested Larks along a stretch of to some 3 km from the lake. Most of this area consisted of large sand plains with sparse, scattered vegetation and variable amounts of pebble and rock. Acacia tortilis is the dominant tree, with Doum palm Hyphaene compressa and thickets of Salvadora persica growing around springs and along dry riverbeds. Meadow-like Sporobolus spicatus grass covers a belt some hundreds of metres inland from the lake, petering out as one moves further away from the shore. Crested Lark populations were sampled between 14 January and 18 February 1992, using a specially-designed technique that we termed a circular-plot count (CPC). To perform a CPC, one observer stands stationary holding the end of a stretched string 56.4 m long (i.e., the radius of a 1-ha circle). The other walks around in a circle holding the opposite end. Birds are flushed or observed as the string passes over them, and all the birds seen inside the circle are recorded CPCs in 50 different plots. For each CPC date, time of day, percentage of terrain by the central observer. We carried out 78 the following parameters were recorded: covered by grass/pebbles, and distance from -69- Leon Bennun 70 human dwellings/lake (estimated to (HMSO, 1977) where appropriate). et al. the nearest 10 per cent, with the aid of a map Results Table 1 shows the results of a Poisson regression model, selecting the best predictors of the number of birds observed in the different CPCs. The parameters of the model maximum were estimated by goodness-of-fit of a in the likelihood, model including model were date (10 which provides a convenient index of a particular set of parameters. Variables included different sampling days), period of day (morning: 07:00- 11:00; mid-day: 11:00-15:00; evening: 15:00-19:00), percentage of terrain covered by grass, percentage of terrain covered by pebbles, distance from human dwellings and distance from the lake. program GLIM (NAG We used a stepwise backward procedure 1986) to select a final in the statistical model containing only significant parameters. In the final model, period of day and percentage of terrain covered by grass significantly affected the number of birds observed, with period being the strongest predictor. Table 1. Poisson regression model selecting parameters which best predict the number of CPC. All parameters are tested with a stepwise backward procedure, birds per (approximately equivalent to scaled deviance) when a calculating the change in parameter is excluded from the model. Parameter *, Adf P < 0.05 P Percentage pebbles 0.58 1 0.45 Distance from habitation 1.54 1 0.21 Distance from the lake 1.88 1 0.17 15.35 9 0.08 4.04 1 0.04* 13.89 2 0.001* Date Percentage grass cover Period of day = 0.35 (± 0.21) + 0.008 (± 0.003) 0.44 (± 0.26) period of day. Selected model: G(x.) grass cover - The percentage of grass cover was positively, though weakly, related to number of CPC (« = 78, r = 0.215, P = 0.059). To examine the effect of period of day on birds per number of birds per CPC in more detail, we selected 14 plots for which CPS were made in all three periods. The number of birds per CPC markedly decreased from the morning to mid-day, and increased again from mid-day to evening, leaving no statistical difference between morning and evening (Repeated measures F2 2,= 5.02,/' = 0.014; Fig. ANOVA: 1). The presence of human dwellings did not tend to affect the distribution pattern of Crested Larks. However, our CPC technique could not be applied within villages. We often observed Crested Larks gathering around shops or houses were food was available, and densities could have been high. Distribution and habitat use of Crested Larks 71 1. Mean number of Crested Larks per plot, estimated with the circular plot count technique, in 14 plots counted in each period of the day. The V -values shown correspond to single paired t-tests Figure morning mid-day Time The mean of evening day overall density of birds (individuals ha ') during the course of the day, estimated through CPCs, was 1.45. In the selected Poisson model, distance from the number of lake did not significantly affect the densities were recorded in a belt of 100 birds per although only during the morning period (Two-way 3.64, P = 0.031; CPC. However, higher m from the shore compared to further 'inland', ANOVA: factor Period, from Lake (<100 m versus >100 m), F^ .^^= F^^^= 2.68, P = 0.07). factor Distance 0.001; two-factor interaction, /^2 72 ~ UM^ P = Discussion Crested Larks are the commonest passerines around Lake Turkana. If both the overall density observed in the course of our study (1.45 individuals ha structure are similar to those an area of 2000 km^ (i.e., lake), the population found in other parts of the region, a belt of about 4 amounts to about km extended 300 000 all ') and the habitat we can estimate that individuals. This very tentative estimate could be improved by carrying out additional counts in other areas. technique would be appropriate for passerines in open habitats. and accurate CPCs It this, in around the shores of the The CPC and for censusing other ground-dwelling proved easy to perform, and appears to give consistent results. carried out at different distances movement between from the lake suggest a circadian pattern of the lakeshore and the surrounding areas. Our the birds were concentrated along the lakeshore during the early results suggest that morning hours. Some of these birds apparently scatter inland during the day. During the hottest hours, when CPC numbers were at a minimum, the birds may concentrate around dwellings, in the shade of rocky cliffs, or at favoured drinking sites around springs or standpipes. observed concentrations of birds census with the CPC technique. at all One We these types of sites, but they are difficult to lark, captured and ringed at a drinking-site, was Leon Bennun later observed feeding km away, 1 so et al. movements could be substantial. The afternoon appears to see the larks feeding in the open again as they begin a generalized movement back towards the lakeshore. Acknowledgements This study was carried out during an expedition organized by the Istituto Nazionale Biologia Museums Selvaggina, Bologna, Italy; the National Universita Pavia, Italy. Lombarde (Pavia), We of Kenya, and the Dipartimento Biologia Animale, thank Ferrino (Torino), Zeiss (Milano), Cassa Risparmio delle Province Agip (Kenya) and Olivetti (Kenya) for support, and James Wachira, Patrick field. We are grateful to Michel Louette for his Gichuki and Syingi Kathele for assistance in the comments on earlier drafts of this paper. References Abs, M. 1963. Vergleichende Untersuchungen an Haubenlerche {Galerida cristata L.) and Theklalerche {Galerida thekla A. E. Brehm). Bonner Zoologisches Beitretung 14: 1-128. Cramp, S. (ed.) 1988. The birds of the Western Palearctic. Vol 5. Oxford: Oxford University Press. Dementiev, G. p. & Gladkov, N. A. 1954. Pticy Sovetskogo sojuza. Moskva. DoLGUSHiN, I.A., KoRELOv, M.N., Kuz'mina, M.A., Gavrilov, E.I., Gavrin, V.F., Kovshar, A.F., BoRODiKHiN, I.F. & RoDioNov, E.F. 1970. Ptitsy Kazakhstana, 3. Alma-Ata. HMSO, 1977. Sheet 41, Kulal, series Y 633. [map] London: HMSO. HopsoN, A.J. 1982. Lake Turkana: A report of the findings of the Lake Turkana project, 19721975. Vol. 1. London: Overseas Development Administration. Hughes, R.H. & Hughes, J.S. 1992. A directory of Afiican wetlands. Gland, Nairobi and Cambridge: lUCN, UNEP and WCMC. KozLOVA, E.V. 1975. Ptitsy zonal' nghk stepey ipustyn Tsentral'noy Azii. Leningrad. Keith, S., Urban, E. & Fry, C.H. 1992. The birds of Africa, Vol 4. London: Academic Press. etude de la biologie de I'Alouette huppee au Pays Drouais (Eure-et-Loire). L'Oiseau R. F. O. 27: 143-149. Labitte, a. 1957. Contribution a Lewis, A. NAG, & Pomeroy, D. 1989. A 1' bird atlas of Kenya. Rotterdam: A. A. Balkema. 1986. Generalised Interactive Modelling System. Oxford: Numerical Algorithms Group. Porter, R.F., Squire, J.T. Report, 1966-7. & Vittey, A. (eds). 1969. Ornithological Society of Turkey Bird & Pratt, D.J., Greenway, P.J. Gwynne, M.D. 1966. A classification of East African rangeland with an appendix on terminology. Journal of Applied Ecology 3: 369-382. RusTAMOV, A.K. 1958. Ptitsy Turkmenistana 2. Ashkhabad. Viulleumier, F. 1979. Ornithological observations during the fall 1979 Nile cruise of the American Museum of Natural History. Unpublished MS held in the American Museum of Natural History. Leon Bennun, Edward Waiyaki and Luc Lens, Ornithology Department, National Museums of Kenya, Box 40658, Nairobi, Luca Borghesio, so. Re Umberto 42, Torino, Italy and Ornithology Department, National Museums of Kenya, Box 40658, Nairobi, and Luca Biddau, Dip. Biologia Animale, Univ. Torino, V. Accademia Albertina 17, Torino, Italy and Ornithology Department, National Scopus 19: 69-72, March 1997 View publication stats Museums of Kenya, Box 40658, Nairobi Received 10 September 1996