290 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 124. No. 2. June 2012 
above ground of the nest upper rim varied from 30 
to 220 cm (110 ± 40 cm; n = 151). The height 
increased 7, 11, and 15 cm front the end of nest 
building to the end of the reproductive period for 
three nests. Nest height for the remaining nests 
decreased from I to 55 cm (15.4 ± 0.1 cm; n = 
30). Increase in nest height was the result of 
growth of the supporting vegetation, while the 
decrease was due to the natural bending of the 
vegetation. 
Nests were anchored to the supporting vegeta¬ 
tion using from one to I I attachment points (4.1 
± 2.3; n = 150) with use of dry (50%) and green 
vegetation. Most nests (;i = 151) were attached to 
two supporting points oriented in different 
directions (22% inclined and horizontal. 14% 
vertical and horizontal, 13% vertical and in¬ 
clined). but some were attached to supports in all 
directions (20%), or only to inclined (18%), 
horizontal (8%), or vertical (7%) vegetation. 
Nineteen plant species were used for nest 
attachment and more than half of the nests were 
supported by only one plant species, the most 
common being A. danaefolium and C. jaimiicen.se 
(Table 3). Up to four plant species were used as 
support for some nests with use of two species 
(32% ol Lhe nests) being more common than use 
of three or four species (~6 and 1 %, respectively) 
(Table 3). Most nests were attached to herbaceous 
plants, but 11 % were attached to herbaceous 
plants and a tree or shrub, and 11 % were attached 
only to trees or shrubs (Table 3). A nest attached 
to the fork of an additional tree species (Tibou- 
china trichopoda) was found in a marsh with 
distinct physiognomy. 
All nests examined had two oval-shaped white 
eggs (n = 163) with several irregular spots of two 
different shades of brown (N„ 0 : A 70 X M 80 and 
N8 o: A| 0 X M 60 ; n = 12) scattered throughout the 
egg, but concentrated at the middle or the largest 
pole (Fig. 1). Eggs measured from 18.0 to 
20.2 mm in length (18.9 ± 0.47 mm, n = 34), 
and 13.2 to 15.6 mm in width (14.0 ± 0.48 mm). 
Egg mass ranged from 1.3 to 2.0 g (1.8 ± 0 17 g- 
n = 33). 
DISCUSSION 
The reproductive season of the Marsh Antwren 
occurs from the end of the winter to the beginning 
of the summer in the Austral Hemisphere and, 
together with egg coloration and clutch size, are 
within the normal variation in Thamnophilidae 
(e.g., Velho 1932, Zimmer and Isler 2003). Sick 
(1985) proposed recognition of five broad nest 
types used by members of the Thamnophilidae 
Some nests of the Marsh Antwren do not fit in an) 
of the proposed categories, and we recognize a 
sixth nest type (open cup attached to vertical 
structures). Thus, some nests of the Marsh 
Antwren are typical open cup attached b) the 
rim at horizontal structures (31%) or open cup 
attached to vertical structures (15%). but most 
used both building techniques (50%); the remain¬ 
ing nests represented unusual sanations for the 
species. The great variation in height above 
ground for nests of the Marsh Antwren is probably 
related to the marsh habitat inhabited with tida 
variation that occasionally may reach low lying 
nests. 
The nest placement on vertical structures used 
by the Marsh Antwren is also found in other 
unrelated marsh-dwelling birds of the Tyrannise. 
Sylviidae, and Icteridae (Orians 1985. De la Pena 
1987, Skutch 1996. Narosky and Salvador 1998, 
Hansell 2000, Azpiroz 2003), as well as in 
grassland birds (Emberi/idae; De la Pena 1996, 
Di Giacomo 2005), The attachment of nests to 
several vertical structures is a solution to prevent 
the structure from slipping down in habitats with 
predominant vertical structure (Hansell 2000): the 
Marsh Antwren is the only thamnophilid known to 
use this strategy. 
ACKNOWLEDGMENTS 
BLR received a grant from Coordenayao de Aperfciyoa- 
mento de Pessoal de Nivel Superior (CAPESi and 
Fundayao de Amparu a Pesquisa do Estado de Sao Paul" 
<FAPESP: 04/13274-2). RB-t. was supported by gran's 
Irom C'onset ho Nacional de Desert vol vimento Cientificoc 
Tccnoldgico (CNPq/MCT: 132893/2009-6) and CAPES, 
and is currently supported by a grant from CNPqAtCT 
(141823/2011-9). DDS. MRP.’ and MAP arc supported h> 
grants from CNPq/MCT (135205/2009-3, 571334'JOtXs- ; 
and 503835/2008-0. respectively). MRB and EC 
supported by grants from the Programa de Reesiruiui K 
das t niversidades Fedcrais (REUNI). Additional vupp'n 
came from Fundayao O Boticario de Proteyao a Naiurt - 
(0682/20052 and 0740/20071). Idea Wild donated equip¬ 
ment. Y. S. Kumvoshi, O. A. Guimaraes. and F. S. Me> cr 
helped with terminology of herbaceous plants. Manv 
students helped during the fieldwork. M. S. Milana 
suggested and encouraged the realization of a Ph D by 
BLR. C. E. Braun. K. J. Zimmer, and an anonym** 5 
reviewer revised the text. 
LITERATURE CITED 
Altmann, J. 1974. Observational study of behavior- 
sampling methods. Behaviour 48:227-265. 
