FIG. 2. Number of individual Rufous Fantails captured on Saipan showing centers at P I (bar stippling) and primaries 
2-4 (bar hatching), and proportion of individuals with centers at primary 2-4 (line) by month during February to 
August 2009-2010. 
and gender (P = 0.038) remained significant or 
nearly so, but the effects of wing chord (P = 
0.170) and mass (P = 0.828) were not significant 
when analyzed using multiple ANOVA (/A47). 
The within-month proportion of individuals 
captured in February to August that exhibited a 
center among primaries 2-4 varied from 20 to 
47% with little evident seasonality (Fig. 2). Six ol 
eight birds captured during separate prcbasic molt 
episodes had the typical sequence from P I during 
both episodes, one DCB female had a center at P 1 
in 2009 and a center at P 4 in 2010. and one SCB 
female had a center at P I in 2009 and a center at 
P 3 in 2010; these two individuals represented 
exceptions to the age-specific pattern. 
DISCUSSION 
Multiple molt centers or series among primaries 
are poorly documented in passerines and our 
results indicating a center at primaries 2. 3, or 4 
for some Rufous Fantails provides the first 
evidence for multiple molt scries in the Rhipidur- 
idae (Higgins et al. 2006). The only North 
American passerine with an atypical primary molt 
sequence may be the American Dipper, where 
primaries 2-6 may drop simultaneously before P 
1. an adaptation enabling rapid molt and resulting 
in brief flightlessncss (Sullivan 1965). This should 
be confirmed, however, because the well-studied 
White-throated Dipper (Cinclus cinclus ) report¬ 
edly molts rapidly, but in typical sequence from P 
I (Cramp 1988). 
The distribution of molt centers among the 
primary 2-4 group of Rufous Fantails with only 
3% initiating molt at P 2 and the remainder 
showing a normal distribution, suggests this group 
exhibits a different mechanism affecting replace¬ 
ment sequence from the primary-1 group, rather 
than the center simply drifting among P 1 and P 4. 
The mechanism affecting center resulted in an 
average initiation point just distal to P 3 among 
our sample of individuals with medial molt 
centers. Some individuals appeared to have 
molted primaries 3 and 4 at or near the same 
time, suggesting the molt center may not be fixed 
at an individual primary, but can occur between 
two primaries or along a defined area along the 
alar tract. Fluidity in both molt commencement 
point and directionality has not been fully 
documented to our knowledge, although it appears 
to occur in Savi’s Warblers (Locustellu lusci- 
"iodes ; Jenni and Winkler 1994). Passerines with 
eccentric preformative molts among primaries can 
vary their center of initiation but distal sequence 
is maintained (Pyle 1997). 
Mechanisms affecting molt strategies in birds 
remain poorly known due to complex interactions 
between physiological, environmental, and genet- 
