3.3 Focal taxa of tracking studies
Across studies, 165 species were tracked in four taxonomic groups: 84
bird species, 75 mammal species, four reptile species and two fish
species. Almost half of the mammals (48%) studied were bat or flying
fox species. Mammals were the most well studied group (49.4% of
studies), followed by birds (45.7%), reptiles (3.7%) and fish (1.2%).
Geographically, birds were the target of most tracking studies in Europe
(61.1%), North America (68.4%) and Oceania (87.5%), whereas mammals
made up most studies in tropical regions across South America (52.9%),
Central America (61.5%), Africa (64%) and Asia (75%) (Figure 2). Only
48 species were studied more than once, including 23 birds, 23 mammals
and 2 reptiles. The most common frugivores studied were the Seba’s
short-tailed bat (Carollia perspicillata ), tracked on six
occasions, followed by the Eurasian Jay (Garrulus glandarius ),
African bush elephant (Loxodonta africana ), Asian elephant
(Elephas maximus ) and Little yellow-shouldered bat
(Sturnira lilium ), all tracked on four occasions.
A diverse range of species were studied, which is shown by large ranges
in body mass (7.6 – 44000g and 9.4 – 4750000g, for birds and mammals
respectively). Body mass proved significant in predicting an increased
probability of GPS tags being used in frugivore studies; with every one
unit increase in body mass (g), there is an estimated 0.000012 increase
in the log-odds of GPS tags being used (estimate ± std error = 0.000012
± 0.0000059, p = 0.0384, df = 195). However, there was no effect of taxa
(estimate ± std error = 0.5365 ± 0.3577, p = 0.1337, df = 195) or an
interaction between taxa and body mass (estimate ± std error = -0.000038
± 0.000038, p = 0.321, df = 194) on the outcome of GPS tags being used.
For mammals, there was a significant effect of study year on the
probability of GPS tags being used; with every succeeding year, there
was an estimated expected 0.385 increase in the log-odds of GPS tags
being used (estimate ± std error = 0.3849 ± 0.09692, p = 0.000072, df =
89; Figure 3). However, body mass had no detectable effect (estimate ±
std error = -0.000079 ± 0.000055, p = 0.147, df = 89), nor was there any
detectable interaction between year and body mass (estimate ± std error
= 0.0000027 ± 0.0000018, p = 0.125, df = 89). For birds, there was no
detectable effect of year on the outcome of GPS tags being used
(estimate ± std error = -0.03877 ± 0.1432, p = 0.78666, df = 101).
However, body mass (estimate ± std error = -0.03935 ± 0.01125, p =
0.00072, df = 101), and an interaction between body mass and year
(estimate ± std error = 0.00146 ± 0.000412, p = 0.0.000408, df = 101;
Figure 3) did demonstrate significant in predicting the outcome of GPS
tags being used.