SHORT COMMUNICATIONS 
841 
Carolina of P./ pallida and an autumn specimen 
from Missouri in 1977 that was identified by 
measurements and plumage as P.ffulva (Easterla 
2008). There are also recent spring records from 
eastern North America (c.g.. Massachusetts 
[SzantvvlOlO]: and Ontario [Wormington 2010]). 
A Caw Swallow collided with a window at 
McCormick.Place (41 51.308' N. 87 36.770' W 
± 8 ml on 10 November 2008 along the Chicago. 
Illinois lakefront and was found dead by V1HH and 
deposited at the Field Museum of Natural History 
iFMNH). The bird was prepared as a study skin 
:FMNH 461103) and a tissue sample was taken 
MCP08-625), Multiple Cave Swallows were 
found dead at the Cape May Congress Hall, Cape 
May, New Jersey (38 55.856' N, 74' 55.469' W 
r 100 ra) on 27 November 2007 (O'Brien 2007). 
One of these birds was deposited at the University 
of New Mexico's Museum of Southwestern 
Biology (MSB 29350), prepared as a study skin 
with partial skeleton (ABJ 2455). and a tissue 
sample was taken (NKl70651). Our objectives, 
based on these specimens, were to (1) identify 
the specimens to subspecies based on morphol¬ 
ogy' and genetics, and (2) consider the identifica¬ 
tions in terms of expanding populations and 
vagrancy. 
METHODS 
We sequenced a portion of the mtDNA 
cytochrome b (cyt b) gene of each specimen to 
independently assess their identification and 
geographic origins with respect to the mtDNA 
data set previously published by Kirchman et al. 
12000). We also included data recently published 
lDor et al. 2010) from a vagrant Cave Swallow 
Mlvaged in New York and. to increase our sample 
s ize of breeding birds, we sequenced one 
additional Cave Swallow from Valverde County. 
Texas (MSB 18680). 
W'e extracted total genomic DNA from two 
'agram (FMNH 461103 and MSB 29350) and one 
breeding (MSB 18680) Cave Swallow using the 
DNeasy tissue extraction kit tQiagen, Valencia. 
CA. USA) following the manufacturer’s proto¬ 
cols. We used primers L.14841 (all 3 samples; 
Kocher et al. 1989). HI6065 (FMNH sample; 
Helm-Bychowski and Cracrafi 1993). and H4a 
'MSB samples; Harshman 1996) to amplify and 
directly sequence a portion of the mtDNA cyt b 
gene. We followed Patel ct al. (2011) for thermal 
cycling, visualization, and sequencing protocols 
for the FMNH sample. Cytochrome b for the MSB 
samples was amplified in 15 pi reactions using 2 pi 
of the DNA extract and the following reagents: 
0.15 pi of Taq Gold polymerase (ABI. Mountain 
View, CA. USA). 200 pM of each dNTP, 1.5 mM 
MgCl 2 . and 0.5 pM of each primer. Eppendorf 
Mastercycler (Eppendorf, Hamburg. Germany) 
thermal-cyclers were used to conduct the follow¬ 
ing PCR protocol: 95 lor 8 min. (95 for 30 sec, 
50 for 30 sec. 72 for 60 sec) X 35 cycles, and 
72 for 10 min. PCR products were visualized on 
1% agarose gel and cleaned using ExoSAP-IT 
(USB. Cleveland. OH, USA). Sequencing reac¬ 
tions with external primers used BigDye 3.1 
chemistry (Applied Biosystems. Foster City. CA. 
USA) and were visualized using an ABI 3130 
automated sequencer. We assembled the sequenc¬ 
es and inspected chromatograms manually using 
Sequencher 4.7 (at MSB) and 4.10.1 (at FMNH. 
Gene Codes Corp., Ann Arbor. MI. USA). 
We aligned 921 bp of ihese sequences (Gen- 
hank accession #s JN227534—JN227536) with 
sequences deposited m Genbank by Kirchman et 
al. (2000) (accession #s AF182379-182391) and 
Dor et al. (2010) (accession # GU460285) using 
Sequencher 4.10.1 (Gene Codes Corp.. Ann 
Arbor. MI. USA). Sequences from Kirchman et 
al. (2000) were taken from breeding colonies 
throughout the Cave Swallow s breeding range, 
the Dor et al. (2010) specimen was an autumn 
vagrant found dead on 19 November 2005 in 
Tompkins County. New York (Cornell University 
Museum of Vertebrates 51713). 
We generated a 95% statistical parsimony 
haplotype network using TCS Version 1.21 
(Clement et al. 2000). We used PAUP* (Version 
4.0b 10; Swofford 2003) to construct a maximum 
parsimony tree using a heuristic search with TBR 
branch swapping and 100 random addition 
replicates. Support for nodes was estimated by 
1.000 bootstrap replicates with one random 
addition per replicate. PAUP* was also used to 
calculate uncorrected ^-distances. We conducted a 
Bayesian analysis using MrBayes 3 (Ronquist and 
Huelsenbeck 2003) and used a general-time- 
reversible model of sequence evolution incorpo¬ 
rating parameters for invariable sites and gamma 
rate heterogeneity. All parameters were estimated 
as part of the analysis and we conducted two 
parallel runs, each with four Markov chains and 
for 5 million generations. We sampled the Markov 
chains every 500 generations and used these 
10.000 parameter point estimates minus the burn- 
in (500 generations) to create a 50% majority rule 
