578 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 123, No. 3, September 2011 
quantitative estimates of gains. Standard linear 
regressions may not be appropriate due to the 
categorical nature of fat scores: but they have 
been used in other studies (e.g.. Dunn 2001,2002; 
Johnson and Winker 2008) and have heuristic and 
comparative value. We conducted both linear and 
ordinal regressions to examine whether using an 
ordinal regression would yield different results. 
We regressed fat scores on lime of capture using 
both methods. 
We compared (at the species level) estimated 
daily gains with an estimated distance of migra¬ 
tion yet to be accomplished to examine whether 
migration distance affected fattening strategy. We 
used linear regression to compare the slopes of the 
lines from the condition index versus time of day 
regressed against the distance to the middle of 
each species’ breeding range. Distances were 
estimated using range maps in the Birds of North 
America series (Payne 1992, Briskie 1994, Evans 
Ogden and Stutchhury 1994, Hall 1994. Van Horn 
and Donovan 1994. Cimprich and Moore 1995. 
Roth et al. 1996, Han tiers and Patton 1998, 
McDonald 1998, Lowther et al. 1999. Evans 
Mack and Wang 2000, Sedgwick 2000, Eckerle 
and Thompson 2001, Lowther et al. 2001) 
Migration strategies differ among our study 
species with some categorized as exclusively 
trans-gulf migrants, others following the coast 
northward, and a few using both routes; these 
same references were used to categorize each 
species. 
We estimated diurnal mass gains for each 
species by taking the slope of the regression line 
for trends in condition index versus time of day (if 
significantly different from zero), multiplying it 
by the average number of hours of bird activity 
(12.5 hrs from field notes), multiplying this value 
by the average wing chord of the individuals in 
the sample, and dividing by 100. Recaptures were 
excluded (Winker 1992b). Total 24-hr mass gains 
were estimated by subtracting nocturnal loss 
(estimated as 4.5% of average body mass from 
Mueller and Berger 1966) from the estimated 
average diurnal increase. We used a value of 
30.2 kJ energy/g fuel (Pennycuick 2003: Table I) 
for our flight capacity estimates, which assumes a 
fat to protein ratio of 0.95 for deposited fuel 
(Johnson and Winker 2008). 
We examined whether species carrying a higher 
proportion ol lat at first capture acquire less fuel 
than those with relatively low reserves by 
regressing the estimate of percent of mass gained 
in a 24-hr period against percent of a species' 
average mass greater than fat-free mass (after 
Dunning 1993) for the species that showed 
significant diel gains. We arcsine transformed 
the data and used linear regression due to the 
apparently linear nature of the data. 
Percentages of species recaptured after a night 
or more at the site were regressed against slopes 
of the condition index regression to examine 
whether a relationship was present between tins 
on the site and refueling. The number of hours and 
distance that the average captured individual m 
capable of flying were estimated using the 
average species-level diel gains, rates of energ} 
use during migration (Tucker 1974). and pub¬ 
lished values for the energetic content of fat and 
flight speed: 30.2 kJ/g (Pennycuick 2003) and 
40.7 km/hr (Nisbet et a). 1963). The latter 
estimate is based on Swainson's Thrush {Cathams 
ustulatus) and might vary among our stud) 
species (Alerstam ct al. 2007). We also calculated 
the proportion of the population capable of 
making the flight from the Sierra de Los Tuxtlas 
northward across the Gulf of Mexico in a single 
flight, using Galveston, Texas as the destination, 
RESULTS 
Swainson's Thrushes, Wood Thrushes (Hyloci- 
chla mustehna). Hooded Warbers. Magnolia 
Warblers (Dendraicct magnolia ). Kentucky War¬ 
blers (Oporurnis formosus). and Oven birds 
(Seiurus aurocapiUa) had significant positive 
slopes in condition index (Table 2, Fig. 2). No 
species had significant negative slopes. Diurnal 
condition slopes did not differ significantly from 
zero in six taxa: Empidonax, Gray-cheeked 
Thrushes (Cathams minimus). Gray Catbirds 
(Dumetella carolinensis). Worm-eating Wattles 
(Helmitheros vermivorum). Yellow-breasted 
Chats (Icteria virens). Painted Buntings {Passe 
ina ciris), and Indigo Buntings (P. cyanea). 
Both linear and ordinal regressions of fat scores 
on time of day indicated the same species had 
significant gains. These results corroborated the 
observed increases in condition index. Only the 
Ovenbird had significant diurnal condition gains 
but did not have corresponding diurnal increases 
in fat score. The Indigo Bunting did not have 
positive condition gains but had a significant 
increase in fat score (Table 2). A considerable 
percentage of Indigo Buntings (34%), Magnolia 
Warblers (34%), ''and Empidonax flycatchers 
(36%) had some body molt. This additional 
