842 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol 123, No. 4, December 2011 
consensus tree, and to calculate the Bayesian 
posterior probabilities to assess nodal support. 
Morphological identification of the Illinois and 
New Jersey specimens was based on comparisons 
with specimens in the bird collections of FMNH 
and MSB and through comparisons with pub¬ 
lished references. Wing chord measurements were 
taken with a standard wing rule and tail 
measurements were taken with a clear ruler. 
These measurements were compared to published 
measurements in Phillips (1986), West (1995), 
and Pyle (1997). 
RESULTS 
Both the Illinois and New Jersey Cave 
Swallows show a pale buffy throat, relatively 
pale cinnamon rump and forehead, and lack 
extensive rufous on the Hanks that distinguish P. 
f pallida from P. f fulva. However, the New 
Jersey specimen has several fresh, sheathed 
feathers growing in on the rump that are strikingly 
darker (chestnut) than the existing, pale orange- 
cinnamon rump feathers (photograph at http:// 
arctos.databa.se.museum/guid/MSB:Bird;29350). 
This coloration is due to wear and underscores the 
difficulty of subspecies identification by plumage 
alone. The Illinois specimen has a wing measure¬ 
ment of 104 mm and tail measurement of 44 mm 
The New Jersey specimen has highly asymmetri¬ 
cal wing measurements (right wing 107 mm, left 
wing 102 mm) and a tail measurement of 49 mm. 
The Illinois and New Jersey specimens were 
both females with 100% skull ossification and 
ovaries measuring 4 X 2 mm (finely granular), 
and . X 3 mm. respectively. Both are hatch-year 
birds, as indicated by suspended wing molt, with 
primaries 5-9 (IL) and primaries 4-9 (NJ) and the 
corresponding primary coverts relatively worn 
and pale, an d primaries 1-4 (IL) and primaries 1- 
3 (NJ) fresh and dark. 
TABLE 1. Uncorrected ^-distances for the Illinois. 
New Jersey, and New York Cave Swallow specimens 
highlighting (bold) the relatively lower uncorreclcd p- 
distances between the vagrant specimens and breeding P. f. 
pallida group individuals than between the vagrants and/'. 
f fulva group. 
n 
IL 
NJ 
NY 
IL 
i 
NJ 
i 
0.004 
NY 
i 
0.003 
0.005 
[P.f pallida]'' 
5 
0.002 
0.005 
0.004 
YUC 
2 
0.003 
0.007 
0.007 
TX3 
1 
0.001 
0.003 
0.002 
TX1/2 
2 
0.001 
0.005 
0.001 
[P. f. fulva]" 
7 
0.009 
0.011 
0.010 
CU 
I 
0.007 
0.009 
0.008 
FL1 
1 
0.012 
0.014 
0.013 
FL2 
I 
0.009 
0.011 
0.010 
PR7JA 
4 
0.008 
0.010 
0.009 
These lire Die averages of all pairwise comparisons between 9reaiire."m!' 
sampled by Kirchman el al. ( 2000 ) and each of the three vagrant •paifflen.'. 
(Valverde County, Texas to Cuba; Fig. 1). AH 
three vagrants, based on uncorrected eyt b p- 
distances, are genetically closer to previously 
published breeding P. f. pallida than to breeding 
P. /..fulva (Table I). 
The Bayesian analysis supports the monophyly 
of the P. f fulva group (Bayesian posterior 
probability = 0.97; Fig. 2). to the exclusion of 
the three vagrants. Bootstrap support for the F ./ 
fulva group is relatively weak (57%) because of 
the small number of informative characters among 
these recently diverged haplotypes. Neither an¬ 
alytical method provides strong statistical support 
for the monophyly of P. f. pallida. Thus, the exact 
position of the vagrant samples with respect to 
Caribbean and Texas/Mexico birds is unclear. 
The Illinois, New Jersey, and New York Cave 
Swallow haplotypes are each unique and different 
from one another and from the breeding speci¬ 
mens of P. f pallida from Tom Green Countv. 
lexas. The Illinois specimen is one base pair 
different from both Texas haplotypes (uncorrected 
^-distance of 0.1%; Table I). The New Jersey 
speomen is three to five base pairs different from 
die Texas hapJmyp^ (0 .3-O.5%), and the New 
from the C, T en " T !° f ° Ur base different 
shortest,iJK XaS / ap, ° types ( a| -0-4%). The 
honest number of steps between any P. /; pa „ ida 
and a member of the P. f fulva 
DISCUSSION 
Both the Illinois and New Jersey Cave 
Swallows, based on plumage characteristics, can 
be assigned to the P.f pallida group, although the 
potential for color changes due to pluraage-w^ 
makes this identification tentative. Wing and tail 
measurements of the Illinois specimen, as wed as 
wing measurements of the New Jersey specimen 
were in the area of overlap between the t* 10 
subspecies groups, but the tail measurement ol me 
specimen from New Jersey was outside the range 
of P.f fulva but within the range of P.fP a ^ il 
