SHORT COMMUNICATIONS 
801 
Their use revealed stability of both temperature 
and RH, indicating little change in these values at 
Black Swift nests within 24-hr periods or 
throughout the breeding season. Nest tempera¬ 
tures from June through September at the colony 
at Box Canyon, Ouray. Colorado, averaged 
10.6 C lower than historic (1948-2006) ambient 
temperature readings (Western Regional Climate 
Center 2010) and RH averaged 45% higher than 
recent historic (2001-2011) ambient RH readings 
MesoWest 2012). Our results show a close 
relationship with the few data reported by 
previous researchers (Murphy 1951. Foerster 
1987. Marin 1997). 
Median temperatures throughout the breeding 
season at the California site were higher than at 
Colorado and New Mexico sites and likely 
represent elevation and latitude variations which 
exist among North American Black Swift breed¬ 
ing sites across North America. Dates for 
phenology events (onset of incubation, hatching. 
Hedging) occur earlier at lower elevation and 
lower latitude than at higher elevation and higher 
latitude sites as demonstrated by review of Black 
Swift phenology dates at locations throughout 
North America (Wiggins 2004). However, great 
variation in phenology can occur among nests at a 
single site (Hirshman et al. 2007). 
Eighteen percent of the data loggers were lost 
during (lie first years of the study. Some fell from 
their placement site, runoff from heavy rains 
dislodged some devices, some slipped into cracks 
in the rocks formed by freeze-thaw cycles making 
them impossible to retrieve, and corrosion pre¬ 
vented downloading of data. We resolved the loss 
"I data loggers by stabilizing deployed units with 
a thin wire running from the data logger to a nail 
temporarily placed in rock. Human tampering and 
loss of battery power were not identified as 
problems. 
The choice of colonies for data logger placement 
was not randomized. We chose colonies based on 
ease of access to nests or niches that represented a 
nest microclimate. This type of non-probability or 
convenience sampling, where sampling units are 
chosen based on tin arbitrary selection procedure 
such as accessibility, time or budget constraints, or 
study logistics, invariably introduces bias, and 
makes it difficult to develop a statistically valid 
estimate of surveillance parameters, in this case, 
temperature and humidity (Nusser et al. 2010). The 
somewhat broad elevation range, disparate loca¬ 
tions. and inclusion of one subterranean colony 
decreased this bias, and allowed the study to be 
conducted logistically. 
There has been much speculation whether 
Black Swift choice of nesting colonies near cool, 
shady, damp waterfalls is an attempt to provide 
optimal nest microclimate for incubating eggs and 
rearing young, or part of some other life history 
strategy (Knorr 1961. Marin 1997. Levad et al. 
2008). Minimal temperature and humidity fluctu¬ 
ations may slow the metabolism of nestlings and 
permit adults to leave them unattended for long 
periods during wide-ranging foraging flights 
(Levad et al. 2008); these suspected ecological 
requirements may in turn limit nesting sites for 
this species. 
Some authors extrapolated the habitat require¬ 
ments of other Apodidac swifts to possibly allow 
a reversible, temporary torpidity or partial poiki- 
lothermy in developing Black Swift chicks and 
adults (Koskimies 1948. 1950; Udvardy 1954) but 
true torpor has not been identified in Black Swifts 
(Lowther and Collins 2002). We have not 
observed torpor-like characteristics in adult or 
nestling Black Swifts (KMP, pers. obs.; CTC, 
pers. obs.). 
CONSERVATION IMPLICATIONS 
The Black Swift nest microclimate material 
presented, along with other physical ecological 
requirements (Levad et al. 2008). provides infor¬ 
mation for modeling geographical range shifts lor 
this species caused hy predicted climate change. 
Identification and evaluation ot these potential 
range shift areas in North America represent a vital 
first step in protection o( these areas. 
ACKNOWLEDGMENT'S 
The authors are grateful to the U.S. forest Service for 
purchase of data loggers that were used in the White Rivet 
National Forest and other forests in the Rocky Mountain 
Region. The Richard G. Levad Memorial Fund supported 
data logger placement at St. Charles Falls. We thank J. E. 
Bowers, J. G. Doerr. K. I. Giezentanner. R. W. Ghormley. 
S. E. and W. K. Hirshman, the Maleug family. R. S. Noll. 
K, D. Reid. R. E. Torretla. and J. G. Wargo for permits and 
permission to access waterfalls and for help deploying and 
retries ing data loggers. C. T. Collins. K S. foerster. Alicia 
Mendoza. T. VV. Pairick. and J. F Toolen assisted with 
logistics, placement, anil retrieval of data loggers at Lawler 
Falls. K. H. Knudsen supplied access to research library 
resources. M. M. Duffy and L. L, Jenks provided insightful 
informal reviews of the manuscript, and .1 A. Frambach 
assisted in mapping. C. T, Collins and an anonymous 
referee generously reviewed the manuscript and provided 
valuable comments. This paper is dedicated to the memory 
