We estimated the weight of O. mauli to be c. 152 g (n = 1) through the Olmos et al. (1996) approach and c. 174
g (n = 2) through the Campbell and Marcus (1992) approach. The weight obtained from the femur length
regression for O. scops is 107±10 g (n = 25), the calculated range obtained with the same method (82–130 g) being
close to the observed range for this species (64–135 g; Dunning 2008), while the weight obtained from the femur
perimeter for O. scops ranges between 90 and 148 g (i.e., larger than the observed range). Although we should
assume that they are overestimates of the weight of O. mauli, due to the similar size of the pectoral girdle of both
species, our results seem to indicate that O. mauli probably had relatively weak powers of flight (Fig. 5).
The PCA analysis performed with leg bone proportions (data from Tables 3 and 4) captured two principal
components explaining 99.9% of total variance (Fig. 6). Of this total, 61.8% is explained by PC1 which shows a
high positive weighting for tarsometatarsus length and a high negative weighting for femur length. The PC2
explained 38.1% of variance that measured tibiotarsus length.
TABLE 2. Mean length (in mm) and proportions (%) of leg bones (F: femur; Tbt: tibiotarsus; Tmt: tarsometatarsus) to total leg
bone length (femur+tibiotarsus+tarsometatarsus) (Tot), and estimated weight (Em; in g) from mean femur length (using the
approach of Olmos et al. 1996) for extinct insular owl species: Otus mauli from Madeira (this paper); Athene trinacriae from
Sicily (Pavia & Mourer-Chauviré 2002); A. cretensis from Crete (Weesie 1982); A. angelis from Corsica (Mourer-Chauviré et
al. 1997); Mascarenotus sauzieri and M. murivorus from Mauritius and Rodrigues (Mourer-Chauviré et al. 1994); Grallistrix
orion, G. geleches and G. erdmani from Oahu, Molokai and Maui (Hawaii) (Olson & James 1991).