674 
THE WILSON JOURNAL OF ORNITHOLOGY • VoL 123, No. 4. December 2011 
TABLE 4. Measurements (mm) from trunk skeletons of Wake Island Rail (Gallirallus wakensis). Those from the three 
complete skeletons of G. wakensis (Table 2) are also included. Range with mean in parentheses ± standard deviation. 
Measurements taken as described in Table 2. Shaft width of femur is at midpoint. 
Measurement 
G. wakt 
ivis males 
Humerus proximal width 
Coracoid length 
Sternum length 
Sternum depth 
Pelvis length 
Pelvis width 
Femur length 
Femur proximal width 
Femur shaft width 
Femur distal width 
6.6- 7.4 (7.0 ± 0.25) n = 11 
15.7- 17.7(17.1 ± 0.53) n = 18 
23.2- 27.3 (25.fi ± 1.88) n = 19 
5.4-8.0 (7.0 ± 0.64) n = 19 
29.0-32.9 (31.3 ± 1.02) n = 19 
16.8- 18.0(17.4 ± 0.39) n = 19 
36.3- 39.1 (37.9 ± 0.71) n = 19 
6.6- 7.5 (7.1 ± 0.23) n = 19 
2.9-3.3 (3.0 ± 0.10) n = 19 
6.7- 7.6 (7.0 ± 0.20) n = 18 
G. »-akrnsis females 
6.1- 7.2 (6.8 ± 0.50) n = 4 
16.2-16.6 (16.4 ± 0.13) n = 10 
23.0-27.0(25.2 ± 1.09)// = 10 
6.3- 7.7 (7.1 ± 0.48) n = 10 
25.8- 31.5 (30.1 ± 1.58) n = 10 
15.9- 18.0 (16.5 ± 0.37) n = 10 
35.6-36.7 (36.1 ± 0.41) n = 10 
6.4- 6.9 (6.8 ±0.l5)n= 10 
2.6-3.0 (2.8 ± 0.13) n = 10 
6 . 1 - 6.8 ( 6.6 ± 0 . 21 ) n = 10 
what appeared to be new primaries. Abrasion 
against vegetation is probably the cause of this 
degradation of the primaries: apparently males 
were more territorial than females and spent more 
time chasing other males through the scrub. 
Males of G. wakensis are distinctly larger than 
females (Tables 3. 4: Figs. 6. 7), as typical of 
many species of rails. Analyses of coefficients of 
variation, however, show that neither males nor 
females are more variable than the other (S. .). 
Parry, pers. comm.). The size difference does not 
hold for the wing and pectoral elements, however, 
which are nearly identical between males and 
females. There may have been selection for 
relatively larger wings in females, which presum¬ 
ably had at least as much responsibility for care of 
young as males, because the wings were used in 
threat postures against potential predators of eggs 
and young such as hermit crabs and rats. 
Additional Measurements .—Tail measurements 
were not included (Table 3) because tails of the 
Tanager series of G. wakensis are extremely 
abraded or completely molted, making measure¬ 
ments almost useless. Tails of six relatively intact 
specimens ranged from 37.0 to 40.2 mm (mean = 
38.3). Tail lengths in two less worn males in other 
collections were 42.5 and 43.0 mm and in the 
unsexed lectotype. 36.3 ram. 
The following measurements arc from the three 
complete skeletons and are included for potential 
comparison with new species of Gallirallus that 
may be found in the fossil record of Pacific 
islands. The first two are males and the third is a 
female. Mandible: symphysis length 7.1, 7.0. and 
6.1 mm. Humerus: shaft width at midpoint 1.9, 
1.9, and 1.7 mm; distal width 5.1,5.3, and 4.8 mm. 
Carpometaearpus: proximal depth 4.3, 4.6, and 
3.9 mm. Tibiotarsus: minimum shaft width 2.5, 
2.5, and 2.4 mm; distal width 5.8,5.7, and 5.2 mm. 
Tarsometatarsus: proximal width 6.3, 6.1. 5.7; 
shaft width at midpoint 2.8. 2.9. and 2.5 mm; 
distal width 6.4, 6.4, and 6.0 mm. 
Molt, —Only two of the 10 AMNH Owston 
specimens taken at an unknown time of year have 
reasonably unworn primaries. Most of the rest 
have the primaries broken and are variably 
bleached and worn. The tail is partly or com¬ 
pletely absent in four specimens. Wetmore's 
measurements of this series included the tail, 
and the four specimens that we considered to have 
the tail worn or absent have the smallest 
measurements; perhaps Wetmore was measuring 
coverts, rather than rectrices. The birds in the 
Tanager series, except for the very abraded 
remiges, are mainly in fairly fresh plumage but 
more faded than those collected in June by 
Spencer and Garlough. Some individuals have 
very worn feathers on the back of the head and the 
lower back, suggesting the mantle and forecrown 
were renewed first. One female is distinct in 
hav ing the foreparts in fresh plumage (forward of 
the pectoral band ventrally. a little farther back 
dorsally) but the wings and the rest of the back, 
breast, and belly are extremely worn and faded, 
interspersed with a few new feathers on the flanks 
and belly. The rump feathers also appear new. 
This entire series appears to be in tail molt or to 
have the rectrices worn away. Lyons remarked on 
birds with sandy or straw colored primaries, 
which is how they appear when worn and very 
faded. He noted only one of 60 birds in this 
condition on 9 June but that the number had 
considerably increased by 25 August (Lyons 
1939). A specimen in completely fresh plumage 
