Blake ami Loiselle • VARIATION IN GL YPHOR YNCHUS ABUNDANCE 
441 
0123456789 
Number observations per grid cell 
FIG. 5. Number of records per 50 X 50-m grid cell on 
each of two 100-ha study plots (400 grid cells/plot) at 
Tiputini Biodiversity Station, Ecuador. 
between plots (y/ = 8.28. df = 5, P = 0.142) when 
cells with no observations were omitted. Number 
of observations per grid cell had a clumped or 
clustered distribution on each plot (Indices of 
Dispersion = 1.94. 1.96. P < 0.001. on Harpia and 
Puma, respectively; corresponding Morisita's In¬ 
dex values were 1.49 and 1.67. respectively). 
Second-order analyses (using a 10-m distance 
increment and correcting for edges) indicated 
numbers of observations were clumped at distances 
up lo —70-80 and from •— 100 to 150 m from a 
point on both plots. Corrclograms showed a 
significant negative autocorrelation at 0-50 and 
300-350 m and a positive correlation at 200-250 
and 400-450 m on Harpia (Bonferroni corrected 
significance for total correlation - 0.068): signi¬ 
ficant negative correlations on Puma occurred at 
0-50 and 200-250 m and positive correlations at 
100-150 and 500-550 m (Bonferroni significance 
for total correlation = 0.009). 
DISCUSSION 
Glyphorynchus is one of the most widespread 
and abundant species in the forest understory at 
Tiputini Biodiversity Station (Blake 2007. Blake 
and Loiselle 2009) but shows considerable 
variation in abundance over time and space. 
Numbers of Glyphorynchus seen during annual 
counts indicated greater variation in abundance 
than did numbers of birds captured. The two 
methods also differed in that count data indicated 
somewhat greater variation on Puma than on 
Harpia but the reverse was true for captures. In 
neither case, however, was variation significantly 
different between plots. It is perhaps not surpris¬ 
ing that extent of temporal variation did not differ 
substantially between plots given that plots are in 
relalively close proximity. Glyphorynchus clearly 
varies in abundance from I year to the next but 
there was no consistent trend in numbers, either 
up or down. Thus, our results are similar to those 
of Stratford and Stouffer (2001) who reported no 
significant variation in capture rates over 9 years 
at their study site in Brazil. 
Variation in abundance across years might be 
related to changes in survival rate or reproductive 
success or to changes in behavior that might alter 
the likelihood of capture or observation. Annual 
survival of Glyphorynchus al our site is —0.6 
(Blake and Loiselle 2008) and a recent reanalysis 
of 12 years of data provided no indication of 
significant annual variation in survival iBlake and 
Loiselle. unpubl. data). We do not have data on 
reproductive success and are not able to tell if 
annual variation in abundance is related to 
changes in demographic parameters. Temporal 
patterns in local abundance may. in some cases, 
reflect habitat shifts in response to seasonal 
changes in resources (Beja et al. 2010) but it is 
not likely that Glyphorynchus undertakes signif¬ 
icant shifts in habitat (e.g., terra fume to varzea) 
given their relatively small home ranges and 
recapture distances. Changes in moisture may 
affect microclimate conditions and cause some 
species to shill their use of microhabitats, 
potentially affecting probability of capture (Kan* 
and Free mark 1983). Monthly rainfall has varied 
during the course of our study but there was no 
significant correlation between rainfall and num¬ 
bers of captures (based on an analysis of rainfall 
data from Yasuni Research Station). The relative¬ 
ly low level of variation in abundance across years 
(Table 1) suggests more random effects, such as 
daily changes in weather or behavior, might 
account for temporal variation in numbers of 
captures and observations. It is not likely temporal 
variation was the result of methodological issues 
as sampling methods were the same across years 
in our study. 
Glyphorynchus varies in abundance at several 
spatial scales; within plots, between plots, and 
over w ide geographic regions. All nets captured at 
least one individual (Fig. 4). but there were large 
differences in total captures per net, indicating 
Glyphorynchus activity was concentrated in 
specific areas of the plot. Captures per net 
provided no evidence of spatial autocorrelation 
(i.e., captures at I net were nor correlated with 
captures at nearby nets), suggesting activity levels 
