Nest observations
I recorded 43 blue-capped cordon-bleu nests (41 in Mbeya and two in Iringa) and 15 red-cheeked cordon-bleu nests (seven in Mbeya and eight in Iringa). I defined that cordon-bleus were using the nests when bringing or lining nest materials (Movie S1) or staying inside the nest (Fig. 2b). When I found a cordon-bleu nest, I got as close as I could to the nest and recorded its location (longitude and latitude) by BirdLasser, a mobile application to record the observation of birds with the GPS function of the iPhone 7 (Lee & Nel 2020). There are reports that the GPS accuracy error of the iPhone is in the 7-13 m range (Merry and Bettinger 2019). It may not be suitable for detecting the exact nest locations, but I concluded that it is practical enough to grasp the rough distribution of nesting sites for this study.
I checked the following conditions of the nests: whether the nest owners are blue-capped or red-cheeked cordon-bleus (Fig. 1), if a wasp nest was found nearby on the same tree, and if cordon-bleus built the nest themselves or used weaver nest (Fig. 2). Using these data, I examined species differences in the nesting strategy between blue-capped and red-cheeked cordon-bleus.
I also collected the plant types of nesting sites and the height of the nest from the ground to test if there are any position preferences for nesting in two cordon-bleus. Both cordon-bleus usually nested onAcacia spp. or Acacia -like trees (Faidherbia albida) , those are thorny trees with fern-like leaves and are dominant in their habitat. They also nested on shrubs or broadleaf trees as described in the literature (Goodwin 1982). I recorded the types of nested plants into the following three categories: (1)Acacia (-like) trees, (2) shrubs, and (3) broadleaf trees. I also measured the height above the ground of 26 blue-capped nests and 13 red-cheeked nests. Missing values in the nest height are due to that I decided to measure the height from the middle of fieldwork. Therefore it is not related to any external constraints (e.g., too high) around the nest site for the measuring.
Additionally, I sampled a total of five cordon-bleu nests that were no longer used by cordon-bleus to observe the structures and components of their nests. I considered the nests abandoned because I observed neither nest-building nor incubating behavior for more than two weeks after I recorded the locations of the nests. I sampled two own-built nests of blue-capped cordon-bleus, one weaver nest used by blue-capped cordon-bleus, and two own-built nests of red-cheeked cordon-bleus (Fig. 5). I took the photos and observed the appearances. I also checked the nest components by arbitrarily pulling out some of the nest materials (Fig. 5).