Summary The number and distribution of the White Pelican Pelecanus onocrotalus roseus and the Pink‐backed Pelican P. rufescens were studied in the Ruwenzori National Park, Uganda, by means of aerial surveys carried out at approximately weekly intervals between May 1968 and December 1969. The White Pelican is migratory and was present in the Park at highest numbers in August 1968 and May 1969 when 2510 and 3927 respectively were recorded. It does not breed within the Park. The Pink‐backed Pelican is resident, with a population averaging about 1350 birds. Pelicans are distributed in discrete flocks on loafing grounds. Flocks of the White Pelican tend to be bigger than those of the Pink‐backed. Forty‐three loafing grounds were recognized in the Park, and it is shown that there is no difference between the species in the frequentation of these grounds. The areas favoured by both species are characterized by flat, open stretches close to water; steep slopes and heavily vegetated shore‐lines are avoided. The feeding behaviour was studied by direct observation and through cine‐photography. Only the White Pelican feeds communally in coordinated groups; this species also often feeds singly. The feeding sequence of both species is described, and is divided into four phases:searching, trapping, catching and swallowing. The searching phase lasts about six times longer in the Pink‐backed Pelican than in the White. The White Pelican feeds further from the shore than does the Pink‐backed. There is little difference between the species in the daily feeding pattern. Feeding takes place throughout the day but mostly from dawn to about 10.00 hrs and from 17.00 hrs until dusk. The rest of the day is spent mainly in preening, resting or bathing. The food of pelicans is exclusively fish. The White Pelican takes mainly large Tilupia, and Haplochromis and fish fry to a lesser extent. The Pink‐backed Pelican feeds largely on fish fry, but Tilapia and Haplochromis are frequently taken and by weight are more important than the fry. However, the Tilapiu are smaller than those taken by the White Pelican. Ecological separation between the two species is achieved through these differences in their feeding behaviour. It is estimated that the White Pelican takes 1201 g and the Pink‐backed Pelican 776 g of fish each day. The total amount of fish eaten by both species from Lake George during 1969 was calculated to be 591 709 kg. This is small (c. 3%) compared with the estimated total fish production of the lake but quite appreciable (12.7%) as a proportion of the total caught by man. It is concluded that the pelican populations are probably not important as a link in the chain whereby organic matter transferred from land to water by the Hippopotamus is returned to the land.

Summary Random sample counts were made from the air of kob, topi, warthog and waterbuck in the Rwenzori National Park, Uganda in 1971 and 1972. Estimates of the total populations varied widely and the means for each species were 12 301 (kob), 4932 (topi), 1530 (warthog) and 3563 (waterbuck). The kob and topi figures are considered as fair approximations to the true totals. The warthog total is thought to be much too low but the degree of inaccuracy could not be assessed. The waterbuck total is also considered to be too low and the application of a groundlair correction factor gives a total of 4454. The distribution of the animals over the park was unequal and the most heavily populated areas are identified for each species. There were significant differences between the species in their distribution over the various sectors of the park. A relationship between the rainfall and the number of animals counted was not established except in the case of the waterbuck which showed a positive correlation with rain falling in the month prior to the count. The densities determined from these surveys fell within the range of those found by other workers from ground counts. Résumé En 1971 et 1972, des comptages aériens, échantillons au hasard, se sont effectués au Parc national Rwenzori (Ouganda) des cobes de Buffon, damalisques, phacochéres et cobes defassa. Les estimations ont varié mais les moyennes pour chaque espéce ont été 12 301 cobes de Buffon, 4932 damalisques, 1503 phacochéres et 3563 cobes defassa. On croit que les nombres totals des cobes de Buffon et des damalisques sont corrects approximativement. Cependant le chiffre pour les phacochkres êst certainement trop bas mais on ne peut pas evaluer I'erreur. Le chiffre 4500 peut etre une juste estimation des damalisques. La distribution des animaux dans le parc n'ont pas été uniforme et les régions les plus préferées sont indiquées pour chaque espéce. Les differences entre la préférence des espèces divers pour les habitats favoris sont significatives. Une corrélation entre la quantité de la pluie et les nombres des animaux n'ont ktabli que pour les damalisques où on peut déterminer une correlation positive avec la pluie pendant le mois qui précède le recensement. Les densités déter‐minées dans ces dknombrenients aCriens se confornient assez bien avec les resullals des comptages terrestres des autres chercheurs.

SummaryInterspecific tree sharing and niche overlap was studied in six weaver bird species. The three species pairs which shared nest trees were P. cucullatus (35 colonies) and P. nigerrimus (34 colonies), M. rubricollis (71 colonies) and P. tricolor (8 colonies), M. scutatus (22 colonies) and M. nitens (12 colonies). In the first species pair high niche overlap values along all four dimensions (habitat, nest height, nest materials and diet) contradicted Pianka's (1974) niche overlap hypothesis that high overlap is the result of reduced competition. At an inter‐generic level his theory appeared applicable because high overlap values along one dimension followed low values along others. May & MacArthur (1972) predicted an inverse relationship between niche overlap and intensity of competition. Low alpha (a) values of nest materials (M. rubricollis‐P. tricolor) and nest height (M. scutatus‐M. nitens) fitted their explanation because competition must have caused complete distinctness in the two axes. Population assessment provided evidence that under shared colony conditions, only one species numerically dominated. Hence competition risk was reduced even though overlapping of certain niche dimensions allowed for the possibility of competition. The beaks of all six species are not specialized for diet although Ploceus spp. is granivorous and Malimbus spp. and P. tricolor are insectivorous. Within the malimbes segregation along this axis was incomplete while only partial niche segregation was evident between the two groups. Ecological separation of P. tricolor nests was considered to be the result of competition from P. cucullatus and P. nigerrimus.RésuméLe partage des ressources et le recouvrement des niches écologiques de certains tisserins des genres Ploceus et Malimbusà Ife, Nigéria. On a étudié le partage des arbres et le recouvrement des niches écologiques de six espèces de tisserins. Les trois paires d'espèces qui construisent leur nid dans les mêmes arbres sont P. cucullatus (35 colonies) et P. nigerrimus (34 colonies), M. rubricollis (71 colonies) et P. tricolor (8 colonies), ot M. scutatus (22 colonies) et M. nitens (12 colonies). Dans la premiere paire, un recouvrement important de la niche écologique à quatre points de vue (habitat, hauteur et matériau du nid, et régime alimentaire) contredit I'hypothèse de Pianka (1974) selon laquelle un recouvrement important est le résultat d'une compétition restreinte. Au niveau intergénérique, sa théorie semble applicable parce que de hautes valeurs de recouvrement dans un domaine coincident avec de basses valeurs dans d'autres. May et MacArthur (1972) prédisaient une relation inverse entre le recouvrement de la niche écologique et l'intensité de la compétition. Des valeurs alpha (a) basses pour le materiau du nid (M. rubricollis‐P. tricolor) et pour sa hauteur (M. scutatus‐M. nitens) s'accordent parfaitement à leurs explications parce que la compétition doit avoir causé une distinction complète dans ces deux domaines. L'évaluation de la population montre à l'évidence que, dans les conditions d'une colonie mixte, une seule espèce domine numérique‐ment. D'où le risque réduit de compétition même lorsque le recouvrement de certains aspects de la niche écologique aurait pu créer une situation de compétition. Les becs des six espèces ne sont pas spécialisés; cependant, Ploceus spp. est grani‐vore et Malimbus spp. plus P. tricolor sont insectivores. Entre les malimbes, la ségrégation alimentaire est incomplète, et I'on ne constate qu'une ségregation partielle de la niche écologique entre les deux groupes. La séparation écologique des nids de P. tricolor semble être le résultat de la compétition entre P. cucullurus et P. nigerrimus.

SummaryThe variations in the breeding seasons of the black‐headed village weaver and the chestnut‐and‐black weaver were analysed in fifteen of the 30 colonies sighted at Ife and one colony on Ilesha Road, about 14 km to the east. Both species commenced large‐scale breeding during the first peak in rainfall and established numerous unproductive satellite colonies. The changes in gonad weights indicated bimodal and discontinuous off‐season breeding in the black‐headed village weaver but breeding in the chestnut‐and‐black weaver was mainly unimodal, with two peaks observed in only one colony. The mean nest completion time and commencement of a new nest was 11 h (range = 9–15 h) and 6 days (range = 3–13 days), respectively whereas in the chestnut‐and‐black weaver the corresponding figures were 9 h (range = 6–11 h) and 3 days (range = 2–4 days). Mate attraction displays in both species were restricted during early stages of nest construction but were heightened on completion, when females were responsive. The mean bout lengths of undernest and branch displays in the chestnut‐and‐black weaver were significantly (P < 0.05) longer than in the black‐headed village weaver. The pair bond formation and motivation of males to build a new nest were investigated.RésuméOn analyse les variations des saisons de reproduction du tisserin gendarme et du tisserin noir de Vieillot dans 15 des 30 colonies connues à Ife et une colonie sur Ilesha Road, à environ 15 kilomètres vers l'est. Les deux espèces ont commencéà se reproduire en grand nombre lors du premier pic de la saison des pluies et ont établi de nombreuses colonies satellites improductives. Les changements du poids des gonades indiquaient une reproduction hors‐saison bimodale et discontinue chez le tisserin gendarme mais la reproduction du tisserin noir de Vieillot était surtout unimodale, avec deux pics de reproduction dans une colonie seulement. Le temps moyen pour terminer un nid et en commencer un autre était de 11 h. (de 9 à 15 h.) et 6 jours (de 3 à 13 jours) pour le premier alors que pour le tisserin noir de Vieillot les chiffres correspondants étaient de 9 h. (de 6 à 11 h.) et 3 jours (de 2 à 4 jours). Les manifestations pour l'accouplement étaient limitées au début de la construction du nid mais devenaient plus vives lorsque celui‐ci se terminait et que les femelles devenaient plus réceptives. La durée moyenne des manifestations sous le nid et sur les branches était significativement plus longue chez le tisserin noir de Vieillot (P <0,05) que chez le tisserin gendarme. On a étudié la formation du couple et la motivation des mâles à construire un nouveau nid.