ABSTRACT
BACKGROUND: Many species of birds are morphologically and physiologically adapted for migration. Migratory movements of birds can range from thousands of kilometers, such as when birds migrate from wintering to breeding sites in summer, to several kilometers, such as when birds migrate among habitats in a single mountain system. The main factor that influences bird migration is the seasonal fluctuation of food resources; climate, predation, competition for resources and endogenous programming are also important factors. Hummingbirds are highly dependent on nectar, so their migration is likely correlated with the blooming of plant species. The ecological implications of altitudinal migration in the mountains of North America as well as the latitudinal migration of Selasphorus rufus through Mexico are still poorly understood. To explore these issues, over three non-consecutive years, we evaluated interannual variation in the phenologies of a latitudinal migrant (S. rufus) and an altitudinal migrant (Amazilia beryllina) and their visited plants. METHODS: We assessed the relationship between two migratory hummingbirds and flower abundance in 20 fixed-radius plots (25 m radius). All available flowers were counted along transects (40 × 5 m) inside each fixed-radius plot. Sampling was performed every 10 days from November 12 through February 20 of 2010-2011, 2013-2014 and 2015-2016, resulting in a total of 11 samples of each plot per period. Phenological variation and the relationships among hummingbird abundance, flower abundance and vegetation type were evaluated using a generalized additive mixed model. RESULTS: S. rufus abundance was related to sampling time in the first and third periods; this relationship was not significant in the second period. A. beryllina abundance was related with the sampling time over all three periods. The abundance of S. rufus hummingbirds was significantly related to the number of Salvia iodantha flowers. The abundance of A. beryllina hummingbirds was related to the number of S. iodantha and Cestrum thyrsoideum flowers and the total number of flowers. We found a non-significant correlation between S. rufus and A. beryllina abundance and vegetation types. CONCLUSION: Contrary to expectations, the long-distance migration of S. rufus was not consistent over the sampling periods. The migration of S. rufus through the study region may be altered by changes in climate, as has occurred with other species of migratory birds. In the present study, the migration of S. rufus was correlated with the blooming of S. iodantha. In comparison, the altitudinal migrant A. beryllina responded to the availability of floral resources but was not associated with a particular plant. The migration of this latter species in the area probably depends on multiple factors, including climatic and demographic factors, but is particularly dependent on the supply of floral resources and competition for these resources.