346 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol 124. No. 2. June 2012 
to numerous individuals as well as to an Australian 
list-server (BirdingAus-L) on 19 September. All 
replies were tabularized into a data set. A follow-up 
e-mail posted to the same list-servers on 25 
September resulted in additional information 
which expanded the data set. A literature search 
occurred during mid September 2007-mid Novem¬ 
ber 2011 with most references from before 2008. 
Only cases of birds trapped in a single spider 
web were included tor consistency of compari¬ 
sons. Reports of birds found encumbered with silk 
strands attached to the body were not included 
because the circumstances prior to the observation 
were undocumented. The bird may have simply 
picked up strands of an abandoned, broken spider 
web during flight. 
Wing chord data were collected from study 
specimens at the Houston Museum of Natural 
Science (HMNS), the Field Museum (FMNH), 
and the American Museum of Natural History 
(AMNH). Measurements were collected using 
standard dial calipers and measured to the nearest 
mm. Mass (g) data were gleaned from specimen 
tags at HMNS and FMNH, and from Dunning 
(2008) lor AMNH specimens where mass data 
were lacking front tags. 
Birds could be trapped in a web strictly due to 
entanglement in the sticky threading. Those 
wrapped in silk by the spider did not survive 
and this was noted when the data were compiled. 
Additional bird fates examined included whether 
entrapped birds were: (1) able to free themselves 
or released with human intervention (and if they 
subsequently survived), (2) killed or consumed by 
a spider, or (3) found dead in the web. Chi-square 
tests were used to test probabilities (P < 0.05) 
with the exception of a Pearson product-moment 
correlation used to test the relationship between 
avian mass and wing chord. 
RESULTS 
Patterns of Diversity and Size of Trapped Birds.— 
Sixty-nine cases of birds representing 54 species in 
23 families were reported trapped in spider webs 
(Table I). The Ruby-throated Hummingbird {Archi¬ 
lochus colubris) represented the species with the 
most cases of entrapment (n = 6). Hummingbirds 
(1 rochilidae) represented the most diverse family 
with nine species, as well as the family with the most 
cases of entrapment (// = 20), 
Avian mass (s = n g) was significantl 
comiated wing chord (X = 61 mm) (r i 
•7851, P < 0.0001, „ = 67). Sixty ( 87 %) of the 
69 cases of entrapped birds have a mass <15 g 
and 61 (88%) have a wing chord <90 mm 
(Fig. I). The Little Hermit (Phaethomis lony 
uemareus) and Vervain Hummingbird (Mellisugu 
minima) represent the smallest species with a 
mass of 2 g and wing chord of 37 mm. whereas 
the largest species was the Laughing Dove 
(Streptopelia senegalensis ) with mass of 80 g 
and wing chord of 138 mm. 
Patterns of Bird Fates.— Sixty-six cases of 
avian fates were reported. Sixty-two included 
information detailing whether birds were wrapped 
in spider silk: 18 cases (29%) involved birds 
wrapped in silk and 44 (71 %) did not {X : = 10.08. 
P < 0.001, n = 62). Both wrapped (range = 0- 
34 g) and unwrapped (range = 2-80 g) birds have 
a mean mass of 11 g and differences were not 
significant, as were differences in wing chord for 
wrapped (58 mm, range = 41-96 mm) versus 
unwrapped (64 mm, range = 37-138 mm) birds. 
I'welvc (66%) birds wrapped in silk were found 
dead in the web and six (33%) were assisted by 
human intervention (5 released unharmed. I died). 
Light (18%) of the birds not wrapped in silk freed 
them,selves and three (7%) were dead. Thirty-three 
(75%) were assisted by human intervention (31 
released unharmed, 2 died). Only eight birds were 
able to free themselves without intervention, whereas 
the number surviving due to human intervention was 
31 (A* = 12.42. P < 0.0003, n = 39). 
Twenty-two (73%) birds in the natural and 
unbiased cases died as a result of spider web 
entrapment. This excluded biased situations 
where humans intervened and released trapped 
birds unharmed but added four additional cases 
with no data on silk wrapping to bring the total 
number of cases of entrapped birds to 30. All 
eight eases of birds that freed themselves from 
webs were not wrapped in silk: those wrapped in 
silk invariably died unless freed by a human 
observer. Mean mass of the eight birds that were 
able to tree themselves was 11 g (range = 5—34g 
and wing chord was 66 mm (range = 56- 
129 mm), whereas the 22 that died had a mean 
mass of 9 g (range = 2-34 g) and wing chord ol 
53 mm (range = 37-96 mm). The mass and wing 
chord of birds that freed themselves without 
intervention was slightly greater than those that 
died, blit the results are not statistically signT' 
cant lor both measurements. 
Patterns of Spider Taxonomy ami Pervasio - 
ness. All of the 46 records that included a 
description of the spider associated with the web 
