Cond_106_311.567_579.tp

᭧ The Cooper Ornithological Society 2004 FACTORS RELATED TO FECAL ESTROGENS AND FECAL BRIAN E. WASHBURN1,3, DOUGLAS J. TEMPEL2, JOSHUA J. MILLSPAUGH1,4, R. J. GUTIE´RREZ2 1Department of Fisheries and Wildlife Sciences, University of Missouri, 302 A-BNR Building, 2Department of Fisheries, Wildlife, and Conservation Biology, University of Minnesota, St. Paul, MN 55108 We estimated concentrations of fecal reproductive steroid metabolites in free- ranging California Spotted Owls (Strix occidentalis occidentalis) during the breeding season.
We collected fresh fecal samples (n ϭ 142) from 65 individual owls in the Sierra Nevadaduring April–August of 2001. We developed and validated radioimmunoassay proceduresto quantify fecal estrogen metabolites and fecal testosterone metabolites. We used an infor-mation-theoretic approach to identify factors that might influence fecal estrogen (E), fecaltestosterone (T), and fecal estrogen:testosterone ratio (E:T ratio) levels during the owl’sbreeding season. We hypothesized that factors related to sampling procedures, owl charac-teristics (sex, reproductive status), and habitat might influence fecal reproductive steroidlevels. Our analyses suggested that sampling factors and owl characteristics, but not habitatvariables, were related to fecal reproductive steroid levels in Spotted Owls. Our most sup-ported models explained Ͻ30% of the observed variation. Fecal testosterone levels werehigher in male Spotted Owls than females, whereas E:T ratios were higher in femalescompared to males. High fecal estrogens were correlated with high fecal glucocorticoids innonbreeding Spotted Owls, whereas fecal estrogens and fecal glucocorticoids were not re-lated in breeding birds. Sampling influenced fecal reproductive steroid measures, and biasfrom small-mass fecal samples might partially explain these relationships. Noninvasive mea-surements of fecal reproductive steroids might be useful for sex determination and repro-ductive assessment of free-ranging Spotted Owls. However, more research is needed tounderstand the variability we observed in sex steroids before this technique can be effectivein conservation studies.
California Spotted Owl, estrogens, glucocorticoids, noninvasive, testoster- one, physiology, Strix occidentalis occidentalis. Factores Asociados con los Estro´genos Fecales y la Testosterona Fecal en Estimamos las concentraciones de metabolitos esteroides reproductivos en individuos silvestres de la especie Strix occidentalis occidentalis durante la e´poca repro-ductiva. Colectamos muestras fecales frescas (n ϭ 142) pertenecientes a 65 lechuzas en laSierra Nevada entre abril y agosto de 2001. Desarrollamos y validamos un procedimientode radioinmunoensayo para cuantificar metabolitos de estro´geno fecales y metabolitos detestosterona fecales. Utilizamos un enfoque informativo-teo´rico para identificar los factoresque podrı´an influenciar los niveles de estro´genos fecales (E), testosterona fecal (T) y elcociente entre estro´genos y testosterona (cociente E:T) durante la e´poca reproductiva de laslechuzas. Hipotetizamos que factores relacionados con los procedimientos de muestreo, ca-racterı´sticas de la lechuza (sexo, estado reproductivo) y el ha´bitat podrı´an influenciar losniveles de esteroides reproductivos en las fecas. Nuestros ana´lisis sugieren que los factoresasociados al muestreo y las caracterı´sticas de la lechuza se correlacionaron con los nivelesde esteroides reproductivos en las fecas, pero e´stos no se relacionaron con las variables deha´bitat. Nuestro modelo ma´s robusto explico´ Ͻ30% de la variacio´n observada. Los nivelesde testosterona fecal fueron mayores en los machos que en las hembras, mientras que elcuociente E:T fue mayor en las hembras que en los machos. En lechuzas no reproductivas,los niveles altos de estro´genos fecales se correlacionaron con niveles altos de glucocorti-coides fecales, mientras que en individuos reproductivos los estro´genos fecales y los glu- Manuscript received 12 May 2003; accepted 27 March 2004.
3 Present address: USDA, Wildlife Services, National Wildlife Research Center, 6100 Columbus Avenue, 4 Corresponding author. E-mail: millspaughj@missouri.edu corticoides fecales no se correlacionaron. Estas relaciones pueden ser explicadas en partepor la influencia del muestreo sobre las medidas de esteroides reproductivos fecales y porel sesgo causado por muestras fecales muy livianas. Los niveles de esteroides reproductivosfecales medidos con te´cnicas no invasivas pueden ser u´tiles para la determinacio´n de sexosy la evaluacio´n del estado reproductivo de individuos silvestres de la especie S. occidentalisoccidentalis. Sin embargo, antes de que esta te´cnica pueda ser efectiva en estudios deconservacio´n, se requiere de ma´s investigacio´n para entender la variabilidad que observamosen los niveles de esteroides sexuales.
Reproductive steroid hormone levels provide important information regarding the reproduc- The Northern and Mexican Spotted Owls (Strix tive status of animals. However, longitudinal occidentalis caurina and S. o. lucida) are listed studies of reproductive function in free-ranging in the United States as threatened subspecies, Spotted Owls using traditional techniques (i.e., whereas the California subspecies (S. o. occi- blood sampling) would require repeatedly sub- dentalis) is not (USDI 1990, 1993, 2003). Be- jecting owls to the stress of capture, restraint, cause the conservation status of Spotted Owls and blood collection (Le Maho et al. 1992). In has great economic ramifications, there has been contrast, noninvasive monitoring of reproductive extensive research on its biology (Gutie´rrez et steroids using fecal hormone metabolites is ad- al. 1995, Franklin et al. 2004). The biological vantageous because sample collection does not status of the California Spotted Owl is of interest require animal capture, samples can be collected because forest management practices and habitat without disturbing study animals, and numerous conditions within its range are diverse (Verner samples can be collected from individuals (Ber- et al. 1992). In particular, Sierra Nevada forests, covitz et al. 1982, Kofuji et al. 1993). Our ob- where most California Spotted Owls reside, have jectives were: (1) to validate radioimmunoassay been managed using a wide variety of silvicul- (RIA) procedures to quantify estrogen and tes- tural practices, with much less emphasis on tosterone metabolites in Spotted Owl feces and clearcutting than in the range of the Northern (2) to evaluate factors that might explain varia- Spotted Owl. Also, the California Spotted Owl tion in fecal estrogen (E), fecal testosterone (T), is generally well distributed within its limited and estrogen:testosterone ratio (E:T ratio) levels in free-ranging California Spotted Owls during Central to current Spotted Owl research has been a desire to understand the mechanisms thatinfluence its population dynamics and habitat se- lection (Franklin et al. 2000, 2004). Because ex-periments with this species are difficult to exe- cute (Noon and Franklin 2002), alternative ap- Our 925-km2 study area was located in the proaches are being used to elucidate basic eco- north-central Sierra Nevada between George- town and South Lake Tahoe, California (39ЊN, meta-analysis (Burnham et al. 1996), population 120ЊW). During 1986–2001, we surveyed a core genetics (Barrowclough et al. 1999), and the ef- area of 355 km2 each year for all Spotted Owls fects of disturbance using noninvasive tech- plus selected territories within an adjacent 570 niques (Wasser et al. 1997, Tempel and Gutie´r- km2 (Seamans et al. 2001). Public (USDA Forest rez 2004). A field protocol exists for assessing Service) and private land comprised 63% and reproductive status in owls (Franklin et al.
1996), but a negative assessment using this pro- Elevation on the study area ranged from 366 tocol does not always mean that the bird did not m to 2257 m. Vegetation was typical Sierran attempt to nest. One noninvasive technique that mixed-conifer forest (Ku¨chler 1977, Rundel et might provide useful insight into the reproduc- al. 1977). From 600 m to 1500 m the forests tive biology of the Spotted Owl is the evaluation were dominated by ponderosa pine (Pinus pon- of fecal reproductive hormone levels (i.e., tes- derosa) on more xeric sites and white fir (Abies tosterone and estrogen) as a measure of physi- concolor) on more mesic sites. Above 1500 m ological response to social and environmental there was a transition zone dominated by red fir (Abies magnifica). Other common tree species that occurred within the study area included sug- lowed the ICN protocol for the total estrogens ar pine (Pinus lambertiana), Douglas-fir (Pseu- I125 RIA. We measured testosterone metabolites dotsuga menziesii), incense cedar (Calocedrus in Spotted Owl feces using a commercially decurrens), canyon live oak (Quercus chrysole- available testosterone I125 double-antibody radio- pis), California black oak (Quercus kelloggii), immunoassay kit (Cat. #DSL-4100, Diagnostic and Pacific dogwood (Cornus nuttallii).
Systems Laboratories, Webster, Texas). We fol-lowed the manufacturer’s method for the testos- terone I125 RIA, except that we halved the vol- We collected fecal samples from April to August ume of all reagents. We conducted standard val- 2001, with early April marking the onset of idations including parallelism and recovery of Spotted Owl nesting on the study area. We lo- exogenous sex steroid validation assays (Jeff- cated owls using standard techniques (Forsman coate 1981, Grotjan and Keel 1996, O’Fegan 1983). We placed a clean polyethylene sheet be- 2000) on six pooled fecal extracts to confirm neath a roosting owl (when possible) to facilitate these assays were accurately and precisely mea- sample collection. We removed the fecal portion suring estrogen and testosterone metabolites in of excreta from the urine portion (Wasser et al.
Spotted Owl feces. Pooled fecal extracts (ex- 1997) and placed the fecal samples into cryo- pected low, medium, and high levels) were com- genic vials. Fecal samples were then placed into posed of fecal extracts from three males and liquid nitrogen or placed on ice until they could three females. We added exogenous testosterone be transferred to liquid nitrogen. We collected to the medium and high pool fecal extracts to fecal samples from adults and did not use cecal obtain testosterone values under higher dilution discharge samples (Tempel and Gutie´rrez 2004).
levels. Interassay variation was calculated from Most individual owls on the study area were col- the three owl samples. Intra-assay variation was or banded, and unmarked birds were captured calculated by averaging the CVs of replicate and color banded at the first opportunity. We tubes from 20 randomly chosen samples. We used these color bands to assign fecal samples used tests for equal slopes (parallelism) to de- to individuals. The sex of all birds was deter- termine if log-transformed curves of serially di- mined by vocal characteristics or nesting behav- luted pool fecal extracts were parallel to log- ior (Forsman 1983, Franklin et al. 1996).
transformed total estrogen and testosterone stan-dard curves (Neter et al. 1990). We used SAS (SAS Institute 1985) to perform all statistical We placed frozen fecal samples in a lyophilizer (Freeze-dry Specialties, Inc., Osseo, Minnesota)for 24 hr. Once samples were freeze-dried, we FACTORS RELATED TO REPRODUCTIVESTEROID LEVELS ground them, sifted them through a stainlesssteel mesh to remove large particles, and mixed Based on the literature, we developed a priori them thoroughly. We weighed each fecal sample verbal hypotheses regarding the correlation of to the nearest 0.001 g. We extracted glucocor- various factors with fecal estrogen, fecal testos- ticoids from feces using a modification of terone, and estrogen:testosterone (E:T) ratio lev- Schwarzenberger et al. (1991). We placed dried els in California Spotted Owls during the breed- feces (ϳ0.1 g) in a test tube with 2.0 mL of 90% ing season. First, we hypothesized that sample methanol and vortexed at high speed in a mul- collection or storage factors might influence fe- titube vortexer for 30 min. Samples were then cal reproductive steroid concentrations (sam- centrifuged at 500 g for 20 min, and the super- pling hypothesis). Second, we hypothesized that natant was saved and stored atϪ84ЊC.
characteristics of individual owls and the timingof sample collection might influence fecal repro- ductive steroid levels (biological hypothesis).
Last, we hypothesized that habitat characteristics We quantified estrogen metabolite levels in might be related to fecal reproductive steroid Spotted Owl fecal extracts using a commercially levels (habitat hypothesis). For each hypothesis, available total estrogens I125 double-antibody ra- we developed a set of a priori models. We used dioimmunoassay kit (Cat. #07–140202, ICN several covariates to develop the set of a priori Biomedicals, Costa Mesa, California). We fol- candidate models representing multiple hypoth- eses of covariate effects on estrogen metabolites, between territory centers (Peery et al. 1999, testosterone metabolites, and E:T ratios in owl Franklin et al. 2000). A territory center was the location of a nest site, or for non-nesting owls, the average location of roost sites recorded at Sampling covariates. Sampling covariates in- least 1 week apart. Most roost sites within a ter- cluded the initial storage method for collected ritory were near each other within the same for- samples (SSTOR) and the mass of an individual est stand. Mean nearest-neighbor distance was fecal sample (SMASS). Storage methods includ- estimated for the year of highest owl density on ed placing vials containing feces into liquid ni- trogen within 6 hr of collection or storage of We defined owl habitat as all mature conifer vials on ice for up to 36 hr before storage in forest, irrespective of canopy cover, and medi- liquid nitrogen. Sample mass was the mass of um-sized conifer forest with canopy cover the fecal sample (to the nearest 0.001 g) after Ն70% because California Spotted Owls nest, freeze-drying and sieving. Tempel and Gutie´rrez roost, and forage in these habitats (Gutie´rrez et (2004) found that fecal glucocorticoids were al. 1992). Mature forest consisted of stands hav- higher in Spotted Owl feces initially stored on ing dominant trees Ն61 cm diameter at breast ice and in samples of small mass. Thus, we pre- height (dbh), and mid-seral forest had dominant dicted fecal estrogens and fecal testosterone trees ranging from 30 to 60 cm dbh. We esti- would be higher in samples initially stored on mated habitat variables as described in Tempel ice and samples having smaller mass (Appendix Franklin et al. (2000) found that territory-lev- Biological covariates. Biological covariates el habitat characteristics (e.g., the amount of included sex of the bird (SEX), reproductive core owl habitat within a Spotted Owl’s territo- life-history stage (RLHS) of the bird, fecal glu- ry) were important predictors of reproductive cocorticoid level (CORT), and sample collection output in Northern Spotted Owls. Thus, we hy- date (DATE). We determined the sex of an in- pothesized that the amount of core area in the dividual owl from its unique color band or by owl’s territory or the number of discrete habitat the pitch of its call (Forsman et al. 1984). We patches in the owl’s territory influenced repro- hypothesized that male Spotted Owls would ductive steroid levels in California Spotted have higher fecal testosterone levels than fe- Owls. Specifically, we predicted fecal estrogen males, whereas females would have higher fecal and fecal testosterone concentrations would be estrogen and E:T ratios than males (Appendix higher in owls having more core area and fewer B). We estimated the reproductive life-history habitat patches in their territory (Appendix B).
stage of each individual owl using the methods Model selection. We used information-theo- described in Franklin et al. (1996). Each indi-vidual owl was classified as nonbreeding or retic model selection (Burnham and Anderson breeding (i.e., actively nesting or fledging young). We expected breeding owls to have MIXED in SAS) where individual owls were the higher fecal reproductive steroid levels than sampling unit, the date of fecal sample collec- nonbreeding owls. Glucocorticoid metabolite tion was a repeated effect, and all other covaria- concentrations were estimated as described in tes were fixed effects (Littell et al. 1996). Po- Tempel and Gutie´rrez (2004). Since elevated tential nonindependence among fecal samples corticosterone levels might suppress reproduc- from the same individual was accounted for by tion in birds (Wingfield 1988, Wingfield et al.
the repeated-measures analysis. To account for 1998), we hypothesized that owls with high fe- possible heterogeneous sampling variances cal glucocorticoid levels would have lower re- among individuals, we first used restricted max- imum-likelihood estimation to select the appro- Habitat covariates. Habitat covariates includ- priate covariance structure for the global model ed the amount of core area in the owl’s territory using PROC MIXED (Littell et al. 1996); the (CORE) and the number of discrete habitat sample variance structures were ranked using patches in the owl’s territory (PATCH). We de- AIC (a small-sample-size correction for Akai- fined an owl territory as a circle with a radius ke’s Information Criterion). The diagonal matrix of one-half the mean nearest-neighbor distance grouped by sex was the best fit for estrogens and Likelihood-ratio tests for goodness-of-fit of subglobal models estimating fecal estrogens, fecal tes- tosterone, and fecal estrogens:testosterone ratios in California Spotted Owls in the north-central Sierra Nevada,California, April–August 2001. Explanations of covariate abbreviations appear in Appendix A. A priori modelsthat contained covariates from subglobal models that were not significant were not included in the final model-selection subset.
a The Sampling subglobal model contained SMASS, SSTOR, and SMASS*SSTOR.
b The SEX*RLHS subglobal model contained SEX, RLHS, and SEX*RLHS.
c The CORT*RLHS subglobal model contained CORT, RLHS, and CORT*RLHS.
d The CORT*DATE subglobal model contained CORT, DATE, (DATE)2, and CORT*DATE.
e The habitat subglobal model contained CORE, PATCH, and CORE*PATCH.
testosterone, whereas the Gaussian spatial struc- samples were diluted with steroid dilutent to 1: ture grouped by sex was the best fit for E:T.
64 (e.g., 50 ␮L of fecal extract into 1.55 mL of We used a two-stage approach to model fitting steroid dilutent) prior to assay. Intra-assay vari- (Franklin et al. 2000, Seamans et al. 2001). First, ation was 2.4%, calculated from 20 randomly we compared models containing only sampling chosen samples, and interassay variation for five covariates. We used AIC to select the sampling assays was 11.2%. The manufacturer’s reported model with the most support for each of the cross-reactivity of ICN total estrogens antisera three hormone response variables. Second, we was 100% with estrone and estradiol 17-␤, 9% compared models containing biological and hab- with estriol, 7% with estradiol 17-␣, and Ͻ1% itat covariates. To each of the a priori biological for other steroids. Assay sensitivity was 100 pg and habitat models, we also included the cova- riate(s) from the best-fit sampling model. We as- Fecal testosterone. The DSL testosterone I125 sessed goodness-of-fit using likelihood-ratio RIA reliably quantified testosterone metabolites tests for a series of subglobal models and re- in Spotted Owl feces. Serial dilutions (1:2 up to moved those models which did not have signif- 1:128) of low, medium, and high pool fecal ex- icant fit from consideration (Table 1). We con- tracts yielded displacement curves that were par- sidered test results significant for P Ͻ 0.05. For allel to the testosterone standard curve (Fig. 1; the final evaluation, a priori models were ranked all P Ͼ 0.3). Mean recovery of exogenous tes- using AIC (Burnham and Anderson 1998).
tosterone (range 0.5–10 ng mLϪ1; levels chosen to correspond with expected fecal testosterone levels from actual samples) added to low and high pool fecal extracts was 94.1 Ϯ 2.9% (n ϭ Fecal estrogens. The I125 RIA reliably quantified 18). Acceptable recovery of exogenous analyte estrogen metabolites in Spotted Owl feces. Se- (within 90–110%) and demonstration of paral- rial dilutions (1:2 up to 1:128) of low, medium, lelism suggested no sample matrix effects (Jeff- and high pool fecal extracts yielded displace- coate 1981, Grotjan and Keel 1996, O’Fegan ment curves parallel to the estrogens standard 2000). Extracts from fecal samples were not di- curve (Fig. 1; all P Ͼ 0.2). Mean recovery of luted in assay dilutent prior to assay. Intra-assay exogenous total estrogens (range 10–50 pg variation, calculated from 20 randomly chosen mLϪ1) added to low and high pool fecal extracts samples, was 2.2%. Interassay variation for the was 108.2 Ϯ 1.6% (n ϭ 18). Acceptable recov- five assays was 6.7%. The manufacturer’s re- ery of exogenous analyte (within 90–110%) and ported cross-reactivity of DSL testosterone an- demonstration of parallelism suggested no sam- tisera was 100% with testosterone, 6.6% with ple matrix effects (Jeffcoate 1981, Grotjan and 5␣-dihydrotestosterone, 2.2% with 5-andro- Keel 1996, O’Fegan 2000). Extracts from fecal stane-3␤, 17␤-diol, 1.8% with 11-oxotestoste- For each of the three response variables (es- trogen, testosterone, E:T), the sampling subglo-bal model adequately fit the data (Table 1).
Among the four sampling models, the modelESTR thus was included with second-stage models(Table 2). The model TEST tosterone data and thus was included with sec-ond-stage models (Table 3). Similarly, for the E:T data, the model E:T ported and was included with second-stage mod-els (Table 4).
SEX*RLHS subglobal model adequately fit thedata, whereas the habitat subglobal model didnot CORT*DATE subglobal models adequately fitthe estrogens and testosterone data, but did notfit the E:T data (Table 1).
For fecal estrogens, the model ESTRCORT*RLHS was the best supported, with more than doublethe AIC weight of the nearest competing model 25% of the total observed variation. The nearestcompeting model, ESTR was also well supported and was within two ⌬AIC units of the best model. Both models in- dicated that fecal estrogen levels were positively gens and (B) fecal testosterone for California SpottedOwls. Curves of percent binding of I125 tracer (%B related to fecal glucocorticoid levels in non- BoϪ1) versus serially diluted (log-transformed doses of breeding Spotted Owls and unrelated to fecal 1:2 up to 1:128) low pool (n ϭ 2), medium pool (n ϭ glucocorticoids in breeding owls. The second 2), and high pool (n ϭ2) fecal extracts from male and model indicated that fecal estrogen levels were female free-ranging California Spotted Owls were par- negatively related to sample mass and initial allel (test of equal slopes, all P Ͼ 0.3) to total estrogen(log-transformed doses of 5–200 pg mLϪ1) or testos- storage of the samples in liquid nitrogen.
terone standard curves (log-transformed doses of 0.1– 25 ng mLϪ1). Diamonds: total estrogen or testosterone standard curve points, circles: serially diluted low fecal ue, and its AIC weight was more than three extracts, triangles: serially diluted medium pool fecal extracts, and stars: serially diluted high pool fecal ex- times higher than the other models (Table 3).
This model explained 29% of the observed var-iation. The nearest competing models wereTEST rone, and Ͻ1% for other steroids. The sensitivity which were also supported and within four ⌬AIC units of the best model. These three mod- els indicated that male California Spotted Owls had higher fecal testosterone levels than females and that fecal testosterone levels decreased as During the 2001 breeding season, we collected the mass of fecal samples increased. Addition- and assayed 142 fecal samples from 65 individ- ally, two of the models suggested that breeding ual California Spotted Owls (32 females and 33 owls had higher fecal testosterone levels.
males) for fecal estrogen and fecal testosterone For E:T, the most supported models were E: metabolites. We collected an average of 2.2 Ϯ 0.1 (range 1–6) fecal samples per individual which explained 27% of the total observed var- Ranking of a priori models estimating fecal estrogens in California Spotted Owls in the north- central Sierra Nevada, California, April–August 2001. Models were ranked based on Akaike’s InformationCriterion adjusted for small sample size (AICc). ⌬AICc is the difference in AICc between a model and the best-approximating model. AICc weights sum to 1 and indicate the relative likelihood of the current model. Seeappendices for explanation of covariate abbreviations and model hypotheses.
a Maximized log-likelihood value.
b Number of model parameters.
c The lowest AICc score was 1125.4 for the sampling models and 1097.5 for the overall models.
d Covariates from this model were then included in the overall models.
iation. Both E:T models indicated that E:T ratios life-history stage of individual owls was corre- were higher in female California Spotted Owls (Fig. 2). The nearest competing model was E:T within two ⌬AIC units of the best model (Table Our findings indicated the ICN I125 double-anti- 4). All three models indicated that the ratio of body total estrogens RIA and the DSL I125 dou- E:T was higher in female Spotted Owls com- ble-antibody testosterone RIA used in this study pared to males (Fig. 2) and that E:T ratios were were effective for quantifying immunoreactive negatively related to sample mass. Additionally, estrogen and testosterone metabolites, respec- two of the models suggested the reproductive tively, in feces from male and female Spotted Ranking of a priori models estimating fecal testosterone in California Spotted Owls in the north- central Sierra Nevada, California, April–August 2001. Models were ranked based on Akaike’s InformationCriterion adjusted for small sample size (AICc). ⌬AICc is the difference in AICc between a model and the best-approximating model. AICc weights sum to 1 and indicate the relative likelihood of the current model. Seeappendices for explanation of covariate abbreviations and model hypotheses.
a Maximized log-likelihood value.
b Number of model parameters.
c The lowest AICc score was 971.3 for the sampling models and 954.6 for the overall models.
d Covariates from this model were then included in the overall models.
Ranking of a priori models estimating fecal estrogen:testosterone ratios (E:T) in California Spotted Owls in the north-central Sierra Nevada, California, April–August 2001. Models were ranked based on Akaike’sInformation Criterion adjusted for small sample size (AICc). ⌬AICc is the difference in AICc between a modeland the best-approximating model. AICc weights sum to 1 and indicate the relative likelihood of the currentmodel. See appendices for explanation of covariate abbreviations and model hypotheses.
a Maximized log-likelihood value.
b Number of model parameters.
c The lowest AICc score was 559.6 for the sampling models and 529.7 for the overall models.
d Covariates from this model were then included in the overall models.
Owls. The performance characteristics (e.g., par- ing fecal samples from Spotted Owls of known allelism, recovery of exogenous analyte, accu- sex will be needed to verify the usefulness of racy, precision) of these assays verified that they this technique and to determine the potential er- were accurate, precise, and had an appropriate range of sensitivity. Monfort et al. (1997) suc- Sampling factors and characteristics of indi- cessfully used the same total estrogens and tes- vidual owls influenced reproductive steroid me- tosterone assays to quantify fecal estrogen and tabolite levels in California Spotted Owl feces testosterone metabolites, respectively, in African during the breeding season. The amount of feces wild dogs (Lycaon pictus). Our findings corrob- available for analysis influenced the estrogen orate other studies suggesting the reproductive and testosterone measurements in Spotted Owls.
condition of captive (Bishop and Hall 1991, Ber- Consequently, the E:T ratios also were influ- covitz et al. 1982, Lee et al. 1995) and free- enced by fecal sample mass. Tempel and Gu- tie´rrez (2004) found a similar relationship be- Rounce 1995) birds can be monitored using fe- tween fecal mass and fecal glucocorticoid levels cal reproductive steroid metabolite assays. Ad- in California Spotted Owls. Samples of small ditionally, studies of domestic geese (Anser an- mass (e.g., 0.02 g dry weight) might have been ser; Hirschenhauser et al. 2000) and domestic biased high. Although the reason for this bias fowl (Gallus domesticus; Cockrem and Rounce was unknown, we suspected the extraction effi- 1994) have demonstrated that fecal levels of re- ciency of steroid metabolites from the fecal ma- productive steroid metabolites were related to terial may have been higher with very small fe- cal samples. Researchers should be aware that Previous studies have demonstrated that fecal small samples might provide spurious results.
reproductive steroid analyses can be used to de- Thus, we suggest that in future studies of fecal termine the sex of individuals from monomor- reproductive steroids in Spotted Owls, research- phic bird species (Bercovitz et al. 1978, Stavy ers decide a priori whether to exclude very et al. 1979). The E:T ratio has been shown to small (e.g., Ͻ0.02 g dry weight) fecal samples be particularly useful. Our findings suggest E:T from analysis. Additionally, the uric acid portion ratios from California Spotted Owl feces might of the samples, which might vary among sam- be used to determine the sex of the owl. An E: ples, could have influenced our results.
T ratio of 2.9 or higher indicated the owl was a Fecal samples with high levels of fecal estro- female, whereas an E:T ratio of 1.6 or less in- gen metabolites also had high levels of fecal glu- dicated the owl was a male. Further studies us- cocorticoids. This finding was unexpected, as coids (such as metabolizing energy reserves) un-related to stress responses.
While we observed that variability in sex ste- roids was related to sampling and sex of thebirds, our models explained relatively little var-iation. Our results suggested a more complex re-lationship among Spotted Owl sex steroids andfeatures of the environment than could be ex-plained by our a priori models. As with otherfecal steroid techniques (e.g., glucocorticoid as-says), it will be important to understand the bi-otic and abiotic factors influencing these hor-mones before the real utility of these techniquescan be realized. Basic studies that examine in-dividual variation and between-year variation inreproductive steroids among birds in differentphysical conditions and life-history stages willbe needed to understand variability in fecal re-productive steroid metabolites. Further work isneeded to determine the influence of fecal sam-ple collection and processing on fecal reproduc-tive steroid measurements. In addition, the in-terrelationships between acute and chronic stressresponses and reproductive hormone levels needto be thoroughly examined. These issues will ne-cessitate the study of birds under controlled con-ditions where environmental conditions (e.g.,food availability) and sample handling are al-tered.
In summary, we validated radioimmunoassay procedures that quantify estrogen and testoster-one metabolites in Spotted Owl feces. Nonin-vasive monitoring of reproductive steroid me- Estrogen:testosterone ratios of California tabolite levels in feces, used in combination with Spotted Owls in the central Sierra Nevada, California, demographic information (e.g., reproductive in- during April–August 2001. Bars represent 95% CI. (A)adult males and females, (B) breeding and nonbreeding dices), may provide an effective tool for exam- owls, and (C) male and female breeding (filled circles) ining the effects of various naturally occurring and nonbreeding (unfilled circles) owls.
and anthropogenic influences on the reproduc-tive aspects of Spotted Owl populations. Futureresearch examining how fecal reproductive ste- chronically high corticosterone levels have been roid levels may vary between male and female shown to suppress reproduction in birds (Wing- Spotted Owls during early stages of the breeding field 1988, Wingfield et al. 1998). Sampling fac- cycle (e.g., courtship), during the nonbreeding tors (e.g., sample mass) also influence fecal glu- season, in juvenile birds, and in other owl pop- cocorticoid and fecal reproductive steroid con- ulations inhabiting areas of differing habitat centrations (Tempel and Gutie´rrez 2004; this quality are necessary. Additional research is study). We suspected one or more factors that needed to assess fundamental issues related to elevated fecal glucocorticoid levels also elevated fecal reproductive steroid metabolites, such as fecal estrogen levels in the same individuals.
sampling issues and understanding normal pro- That is, birds in breeding status might have a files in birds of various conditions and life his- concomitant increase in stress hormone secre- tory stages, before the technique can be applied tion related to permissive actions of glucocorti- for collecting and analyzing demographic data onthe Northern Spotted Owl. Studies in Avian Bi- Financial and logistical support for this project was pro- vided by Region 5, U.S. Forest Service (contract FRANKLIN, A. B., D. R. ANDERSON, R. J. GUTIE´RREZ, #USDA-FS/53-9158-02-EC06 to RJG), Region 1, U.S.
AND K. P. BURNHAM. 2000. Climate, habitat qual- Fish and Wildlife Service (service order #101811M663 ity, and fitness in Northern Spotted Owl popula- to RJG), a University of Missouri (MU) Life Science tions in northwestern California. Ecological Mission Enhancement Postdoctoral Fellowship, a MU Research Board Grant, the MU Department of Fish- FRANKLIN, A. B., R. J. GUTIE´RREZ, J. D. NICHOLS, M.
eries and Wildlife Sciences, and the University of Min- E. SEAMANS, G. C. WHITE, G. S. ZIMMERMAN, J.
nesota. Fecal hormone metabolite assays were con- E. HINES, T. E. MUNTON, W. S. LAHAYE, J. A.
ducted in the Wildlife Physiology Laboratory in the BLAKESLEY, G. N. STEGER, B. R. NOON, D. W. H.
MU Department of Fisheries and Wildlife Sciences.
SHAW, J. J. KEANE, T. L. MCDONALD, AND S. BRIT- We thank M. Larson for assistance with the statistical TING. 2004. Population dynamics of the California analyses. G. Zimmerman provided insightful com- Spotted Owl (Strix occidentalis occidentalis): a ments that greatly improved the manuscript. This re- meta-analysis. Ornithological Monograph 54.
search was approved by the University of Minnesota GROTJAN, H. E., AND B. A. KEEL. 1996. Data interpre- Animal Care and Use Committee (Animal Subjects tation and quality control, p. 51–93. In E. P. Dia- mandis and T. K. Christopoulos [EDS.], Immuno-assay. Academic Press, New York.
GUTIE´RREZ, R. J., J. VERNER, K. S. MCKELVEY, B. R.
NOON, G. N. STEGER, D. R. CALL, W. S. LAHAYE, BARROWCLOUGH, G. F., R. J. GUTIE´RREZ, AND J. G.
B. B. BINGHAM, AND J. S. SENSE. 1992. Habitat GROTH. 1999. Phylogeography of Spotted Owl relations of the California Spotted Owl, p. 79–98.
(Strix occidentalis) populations based on mito- In J. Verner, K. S. McKelvey, B. R. Noon, R. J.
chondrial DNA sequences: gene flow, genetic Gutie´rrez, G. I. Gould, Jr., and T. W. Beck [EDS.], structure, and a novel biogeographic pattern. Evo- The California Spotted Owl: a technical assess- ment of its current status. USDA Forest Service BERCOVITZ, A. B., J. COLLINS, P. PRICE, AND D. TUTTLE.
General Technical Report PSW-GTR-133.
1982. Noninvasive assessment of seasonal hor- GUTIE´RREZ, R. J., A. B. FRANKLIN, AND W. S. LAHAYE.
mone profile in captive Bald Eagles (Haliaeetus 1995. Spotted Owl (Strix occidentalis), p. 1–28.
leucocephalus). Zoo Biology 1:111–117.
In A. Poole and F. Gill [EDS.], The birds of North BERCOVITZ, A. B., N. M. CZEKALA, AND B. L. LASLEY.
America, No. 179. The Academy of Natural Sci- 1978. A new method of sex determination in ences, Philadelphia, PA, and The American Or- monomorphic birds. Journal of Zoo and Animal nithologists’ Union, Washington, DC.
HIRSCHENHAUSER, K., E. MO¨STL, P. PE´CZELY, B. WALL- BISHOP, C. M., AND M. R. HALL. 1991. Non-invasive NER, J. DITTAMI, AND K. KOTRSCHAL. 2000. Sea- monitoring of avian reproduction by simplified sonal relationships between plasma and fecal tes- faecal steroid analysis. Journal of Zoology (Lon- tosterone in response to GnRH in domestic gan- ders. General and Comparative Endocrinology BURNHAM, K. P., D. R. ANDERSON, AND G. C. WHITE.
1996. Meta-analysis of vital rates of the Northern JEFFCOATE, S. L. 1981. Efficiency and effectiveness in Spotted Owl. Studies in Avian Biology 17:92– the endocrine laboratory. Academic Press, San Di- BURNHAM, K. P., AND D. R. ANDERSON. 1998. Model KOFUJI, H., M. KANDA, AND T. OISHI. 1993. Breeding selection and inference: a practical information- cycles and fecal gonadal steroids in the Brown theoretic approach. Springer-Verlag, New York.
Dipper (Cinclus pallasii). General and Compara- COCKREM, J. F., AND J. R. ROUNCE. 1994. Faecal mea- surements of oestradiol and testosterone allow the KU¨CHLER, A. W. 1977. The map of natural vegetation non-invasive estimation of plasma steroid concen- of California, p. 908–938. In M. G. Barbour and trations in the domestic fowl. British Poultry Sci- J. Major [EDS.], Terrestrial vegetation of Califor- COCKREM, J. F., AND J. R. ROUNCE. 1995. Non-invasive LEE, J. V., C. S. WHALING, B. L. LASLEY, AND P. MAR- assessment of the annual gonadal cycle in free- LER. 1995. Validation of an enzyme immunoassay living Kakapo (Strigops habroptilus) using fecal for measurement of excreted estrogen and testos- steroid measurements. Auk 112:253–257.
terone metabolites in the White-crowned Sparrow FORSMAN, E. D. 1983. Methods and materials for lo- (Zonotrichia leucophrys oriantha). Zoo Biology cating and studying Spotted Owls. USDA Forest Service General Technical Report PNW-GTR- LE MAHO, Y., H. KARMANN, D. BRIOT, AND Y. HAN- DRICH. 1992. Stress in birds due to routine han- FORSMAN, E. D., E. C. MESLOW, AND H. M. WIGHT.
dling and a technique to avoid it. American Jour- 1984. Distribution and biology of the Spotted Owl LITTELL, R. C., G. A. MILLIKEN, W. W. STROUP, AND FRANKLIN, A. B., D. R. ANDERSON, E. D. FORSMAN, K.
R. D. WOLFINGER. 1996. SAS system for mixed P. BURNHAM, AND F. F. WAGNER. 1996. Methods models. SAS Institute Inc., Cary, NC.
MONFORT, S. L., S. K. WASSER, K. L. MASHBURN, M.
TEMPEL, D. A., AND R. J. GUTIE´RREZ. 2004. Factors BURKE, B. A. BREWER, AND S. R. CREEL. 1997.
related to fecal corticosterone levels in California Steroid metabolism and validation of noninvasive Spotted Owls: implications for assessing chronic monitoring in the African wild dog (Lycaon pic- stress. Conservation Biology 18:538–547.
tus). Zoo Biology 16:533–548.
U.S. DEPARTMENT OF THE INTERIOR. 1990. Endangered NETER, J., W. WASSERMAN, AND M. H. KUTNER. 1990.
and threatened wildlife and plants; determination Applied linear statistical models. 3rd ed. Irwin of threatened status for the Northern Spotted Owl.
NOON, B. R., AND A. B. FRANKLIN. 2002. Scientific U.S. DEPARTMENT OF THE INTERIOR. 1993. Endangered research and the Spotted Owl (Strix occidentalis): and threatened wildlife and plants: final rule to list opportunities for major contributions to avian the Mexican Spotted Owl as a threatened species.
population ecology. Auk 119:311–320.
O’FEGAN, P. O. 2000. Validation, p. 211–238. In J. P.
EPARTMENT OF THE INTERIOR. 2003. 12-month finding for a petition to list the California Spotted Owl (Strix occidentalis occidentalis). Federal EERY, M. Z., R. J. GUTIE´RREZ, AND M. E. SEAMANS.
1999. Habitat composition and configuration around Mexican Spotted Owl nest and roost sites VERNER, J., K. S. MCKELVEY, B. R. NOON, R. J. GU- in the Tularosa Mountains, New Mexico. Journal ´ RREZ, G. I. GOULD JR., AND T. W. BECK. 1992.
Assessment of the current status of the California RUNDEL, P. W., D. J. PARSONS, AND D. T. GORDON.
Spotted Owl, with recommendations for manage- 1977. Montane and subalpine vegetation of the ment, p. 3–26. In J. Verner, K. S. McKelvey, B.
Sierra Nevada and Cascade ranges, p. 559–599.
R. Noon, R. J. Gutie´rrez, G. I. Gould, Jr., and T.
In M. G. Barbour and J. Major [EDS.], Terrestrial W. Beck [EDS.],The California Spotted Owl: a vegetation of California. J. Wiley and Sons, New technical assessment of its current status. USDA Forest Service General Technical Report PSW- SAS INSTITUTE. 1985. SAS user’s guide: statistics. Ver- sion 5. SAS Institute Inc., Cary, NC.
WASSER, S. K., K. BEVIS, G. KING, AND E. HANSON.
SCHWARZENBERGER, F., E. MO¨STL, E. BAMBERG, J. PAM- 1997. Noninvasive physiological measures of dis- MER, AND O. SCHMEHLIK. 1991. Concentrations of turbance in the Northern Spotted Owl. Conser- progestagens and oestrogens in the faeces of preg- nant Lipizzan, trotter and thoroughbred mares.
WINGFIELD, J. C. 1988. Changes in reproductive func- Journal of Reproduction and Fertility Supplement tion of free-living birds in direct response to en- vironmental perturbations, p. 121–148. In J. H.
EAMANS, M. E., R. J. GUTIE´RREZ, C. A. MOEN, AND Stetson [ED.], Processing of environmental infor- the central Sierra Nevada. Journal of Wildlife mation in vertebrates. Springer-Verlag, Berlin.
WINGFIELD, J. C., C. BREUNER, J. JACOBS, S. LYNN, M.
STAVY, M., D. GILBERT, AND R. D. MARTIN. 1979. Rou- RAMENOFSKY, AND R. RICHARDSON. 1998. Ecolog- tine determination of sex in monomorphic species ical basis of hormone-behavior interactions: the using fecal steroid analysis. International Zoo emergency life history stage. American Zoologist Covariates used in a priori models to estimate fecal estrogens, fecal testosterone, and fecal estrogens:testosterone ratios in California Spotted Owls in the north-central Sierra Nevada, California, April–August 2001.
Storage method (same initially stored in liquid N2 or on ice) Mass of dried sample analyzed (to the nearest 0.001 g) Reproductive life-history stage (breeding or nonbreeding) Fecal glucocortiocoid metabolite level (ng gϪ1) estimated from the same sample Amount of interior Spotted Owl habitat (Ͼ100 m from an edge) within the Number of discrete patches of Spotted Owl habitat within the owl’s territory Description, representation, and predicted directions of a priori models to estimate fecal estro- gens, fecal testosterone, and fecal estrogens:testosterone ratios in California Spotted Owls in the north-centralSierra Nevada, California, April–August 2001. Each model was evaluated for each of the three response variables(estrogens, testosterone, and E:T).
reproductive life-histo-ry stage of the bird are related to repro-ductive life-historystage of the bird the number of patcheswithin an owl’s terri-tory the amount of corearea within an owl’sterritory and the amount ofcore area within anowl’s territory are related to theamount of core areawithin its territory a Expected direction of coefficients, given that the hypothesized model is the most supported.

Source: http://fwcb.cfans.umn.edu/research/owls/lit%20folder/Washburn_et_al_2004.pdf