Spatial Variation in Bird Community Composition in Relation to Topographic Gradient and Forest Heterogeneity in a Central Amazonian Rainforest

Table 3

Results of the multiple regression analyses performed to test the effect of the seven forest structure components on qualitative and quantitative bird composition. Analyses were performed on scores from the Principal Coordinate Analyses (PCoA).

Data (matrix)		Forest components	Qualitative (presence/absence)				Quantitative (abundance)
			Pillai-Trace	𝐹	DF	𝑃	Pillai-Trace	𝐹	DF	𝑃
Bird community	Model 1	Tree abundance	0.084	1.958	3; 64	0.129	0.085	1.986	3; 64	0.125
		Log abundance	0.049	1.109	3; 64	0.352	0.069	1.580	3; 64	0.203
		Snag abundance	0.023	0.500	3; 64	0.683	0.043	0.961	3; 64	0.417
		Leaf litter depth	0.111 (−)	2.667	3; 64	0.055	0.036	0.800	3; 64	0.498
		Distance to stream	0.118 (−)	2.861	3; 64	0.044	0.088	2.062	3; 64	0.114
	Model 2	Elevation	0.209 (+)	5.919	3; 67	0.001	0.079	1.928	3; 67	0.133
		Canopy openness	0.033	0.761	3; 67	0.520	0.015	0.351	3; 67	0.789

Because elevation and canopy openness were significantly correlated to other variables (see Table 1), they were analyzed in separate multiple linear regression models (see Section 3 and Figure 3). Three or two PCoA-axes were used in the models for quantitative and qualitative analysis, respectively (see Section 2). The negative and positive signals within parenthesis are just to indicate the directions of the relationships and are not related to Pillai-Trace values (see also Figure 4).