Occurrence and Characterization of Steroid Growth Promoters

Article pubs.acs.org/est

Occurrence and Characterization of Steroid Growth Promoters Associated with Particulate Matter Originating from Beef Cattle Feedyards Brett R. Blackwell,† Kimberly J. Wooten,† Michael D. Buser,§ Bradley J. Johnson,‡ George P. Cobb,∥ and Philip N. Smith*,† †

Texas Tech University, Department of Environmental Toxicology, Lubbock, Texas 79409, United States Texas Tech University, Department of Animal and Food Sciences, Lubbock, Texas 79409, United States § Oklahoma State University, Department of Biosystems and Agricultural Engineering, Stillwater, Oklahoma 74078-6016, United States ∥ Baylor University, Department of Environmental Science, Waco, Texas 76706, United States ‡

S Supporting Information *

ABSTRACT: Studies of steroid growth promoters from beef cattle feedyards have previously focused on effluent or surface runoff as the primary route of transport from animal feeding operations. There is potential for steroid transport via fugitive airborne particulate matter (PM) from cattle feedyards; therefore, the objective of this study was to characterize the occurrence and concentration of steroid growth promoters in PM from feedyards. Air sampling was conducted at commercial feedyards (n = 5) across the Southern Great Plains from 2010 to 2012. Total suspended particulates (TSP), PM10, and PM2.5 were collected for particle size analysis and steroid growth promoter analysis. Particle size distributions were generated from TSP samples only, while steroid analysis was conducted on extracts of PM samples using liquid chromatography mass spectrometry. Of seven targeted steroids, 17αestradiol and estrone were the most commonly detected, identified in over 94% of samples at median concentrations of 20.6 and 10.8 ng/g, respectively. Melengestrol acetate and 17α-trenbolone were detected in 31% and 39% of all PM samples at median concentrations of 1.3 and 1.9 ng/g, respectively. Results demonstrate PM is a viable route of steroid transportation and may be a significant contributor to environmental steroid hormone loading from cattle feedyards.

■

steroid growth promoters in a laboratory setting.4,5 The lipophilic nature of steroid hormones can facilitate exposure to aquatic organisms in impacted aquatic systems, and adverse effects in fish6−12 and amphibians13,14 have been observed following laboratory exposure to steroid growth promoters, even at low, ng/L concentrations. Due to the greater exposure potential of steroid growth promoters to aquatic organisms, most research on steroid transport has focused on transport via effluent or surface runoff;15−21 however, recent research has documented steroid transport via airborne particulate matter (PM) emanating from beef cattle feedyards.22,23 The top six cattle producing states (Nebraska, Texas, Kansas, Iowa, Colorado, and California) account for 75% of U.S. beef cattle on feed.2 Within these top cattle producing states, cattle

INTRODUCTION

Beef cattle are an important agricultural and economic resource in the United States, with an estimated economic value of $88 billion in 2013.1 As of January 2015, there were approximately 90 million cattle and calves within the United States and 13.1 million of those cattle were being fed at feedyards.2 Modern beef production relies on use of steroid growth promoters to increase productivity and profitability.3 Three commonly used steroid growth promoters are trenbolone acetate (TBA), estradiol (17βE2), and melengestrol acetate (MGA). TBA and 17βE2 are administered to cattle via slow-release steroidal implants for steers and heifers, while MGA is administered only to heifers as a feed additive.3 Steroids administered for growth promotion can be excreted as parent compounds or converted to less biologically active metabolites. After excretion, steroids can enter the environment and function as endocrine disrupting compounds following exposure among nontarget receptors. Adverse effects in terrestrial birds have been observed after exposure to high concentrations of © 2015 American Chemical Society

Received: Revised: Accepted: Published: 8796

April 13, 2015 June 5, 2015 June 22, 2015 June 22, 2015 DOI: 10.1021/acs.est.5b01881 Environ. Sci. Technol. 2015, 49, 8796−8803

Article

Environmental Science & Technology

Table 1. Feedyard Site Size (Reported Capacity) and Sample Descriptive Statistics from TSP Samples Collected at Each Feedyard from 2010−2012a location Feedyard A Feedyard B Feedyard C Feedyard D Feedyard E grand mean

cattle capacity 35 000 56 000 40 000 50 000 6000

filters (n)

TSP concentration (μg/m3)

mass median diameter (MMD)

geometric standard deviation (GSD)

PM2.5 (%)b

PM10 (%)b

67 89 91 92 95

1100 ± 1252 1508 ± 1375c 2738 ± 2773d 1017 ± 854c 549 ± 617e 1403 ± 1763

31.0 ± 10.2 31.6 ± 15.3c 30.9 ± 14.4c 29.7 ± 20.7c 38.3 ± 17.9d 32.5 ± 16.6

2.99 ± 0.51 2.85 ± 0.37c,d 2.98 ± 0.60c,d 2.80 ± 0.52c 3.05 ± 0.66d 2.93 ± 0.55

1.65 ± 0.71 1.68 ± 0.83c,d 1.64 ± 0.56c 2.07 ± 0.62d 1.62 ± 0.88c 1.74 ± 0.75

18.0 ± 6.6c,d 19.1 ± 7.2c 20.8 ± 9.0c 20.8 ± 7.9d 15.9 ± 7.1d 19.0 ± 7.9

c

c

c,d

c

All particle size distribution values are given for aerodynamic equivalent diameter (AED). Values presented as mean ± SD. bValues presented as percentage of observed TSP concentration. c,d,eValues followed by different letters denote significant differences (p < 0.05) among feedyards.

a

the interior of the feedyard where changing wind directions would have less prominent effects on PM source. In total, 8 TSP, 4 PM10, and 4 PM2.5 samplers were deployed at each feedyard at a height of 2 m for each sampling campaign. All samplers utilized Zefluor 47 mm filters for sample collection. The airflow rate for the low-volume samplers (16.67 lpm) was monitored and controlled by a system developed by the USDA-ARS Cotton Production and Processing Research Unit in Lubbock, TX. To collect adequate mass for steroid analysis, sampling time for TSP, PM10, and PM2.5 samplers was 48, 72, and 144 h, respectively. Temperature and relative humidity (RH%) were recorded using a Measurement Specialties HRM2500LF sensor module (Hampton, VA). Temperature and RH% were compiled as mean values for each sample collection period. Average daily wind speed and direction data from local meteorological stations near each feedyard (West Texas Mesonet; Oklahoma Mesonet) were collected, with distances from feedyards to nearest meteorological stations ranging 4.8−72 km. Chemicals and Reagents. Steroid standards 17α-trenbolone (17αTb), 17β-trenbolone (17βTb), estrone (E1), 17βestradiol (17βE2), and 17β-Estradiol-d5 were obtained from Cerilliant (Round Rock, TX). Trendione (TbO) was purchased from Steraloids (Newport, RI). Melengestrol acetate-d3 and 17βtrenbolone-d3 were obtained from RIVM (Bilthoven, Netherlands). Melengestrol acetate (MGA), 17α-estradiol (17αE2), and Florisil cartridges were purchased from Sigma (St. Louis, MO). LC−MS grade solvents were obtained from Fisher (Pittsburgh, PA). Regenerated cellulose syringe filters were obtained from Phenomenex (Torrance, CA). Zefluor 47 mm filters and Nanosep 0.45 μm hydrophilic polypropylene membrane centrifugal filters were from Pall Life Sciences (Port Washington, NY). PM Characterization. Gravimetric analysis was conducted on all filters prior to deployment in the field. Filters were conditioned in an environmental chamber (21 ± 2 °C; 35 ± 5% RH) for 48 h prior to weighing. Samples were weighed in the environmental chamber on a Mettler MX-5 microbalance (Mettler-Toledo Inc., Columbus, OH) after being passed through an antistatic device (Mettler-Toledo Inc.). Following collection of PM, filters were transported over dry ice to the laboratory. Filters were reconditioned in an environmental chamber for 24 h, weighed, and mass of PM calculated. Mass of PM collected from TSP filters and sampled air volume were used to generate PM concentration in air (μg/m3). One TSP sample from each sampling period was utilized for particle size distribution (PSD), while one TSP, all PM10 and all PM2.5 were used in steroid analysis. Equivalent spherical diameter (ESD) PSD was determined on select TSP filters using a Beckman Coulter LS230 Laser Diffraction Particle Size Analyzer

feeding operations are predominantly located in semiarid to arid regions, averaging less than 510 mm of precipitation annually. Dry conditions are conducive to PM production at feedyard surfaces and subsequent transport via wind; therefore, in semiarid and arid regions, PM could represent a significant route of steroid transport. Moreover, climatic trends and projections of climate models suggest the arid regions of the Southwest U.S. may continue to experience dry, worsening conditions throughout the century. If regions of the U.S. most densely populated by beef cattle feedyards do continue to trend toward semiarid and arid climates, PM production and transport will become even more important to consider for human and ecological exposure to veterinary medicines and chemicals. The objective of this study was to characterize fugitive PM emissions from beef cattle feedyards and determine steroid occurrence and concentration in airborne PM. Herein, PM concentration and particle size distributions along with occurrence and concentration of seven steroid hormones associated with PM are reported. No previous study has characterized the distribution of steroid hormone concentrations in PM originating from beef cattle feedyards, and data generated in this study will provide a better understanding of steroid transport into the environment. Aerial transport of nonvolatile chemical compounds is rarely considered, and this study underscores the importance of airborne transport from cattle feedyards in semiarid and arid environments. PM emissions could represent a viable pathway of transport for not only steroids, but also other commonly used veterinary compounds such as antibiotics24 or β-adrenergic agonists.

■

MATERIALS AND METHODS Site Descriptions and Sample Collection. PM sample collection was conducted in cooperation with Texas Cattle Feeders Association (TCFA) at commercial feedyards on the Southern Great Plains, U.S. (n = 5). Sites were selected based upon feedyard size, proximity to other feedyards, and spatial layout. Capacity at four of five feedyards ranged from 35 000 to 56 000 head, while one was 6000 head (Table 1). Feedyards were enclosed by either pastureland or cropland and were not located adjacent to other animal feeding operations. Eight separate sampling campaigns were conducted at each feedyard from November 2010 through March 2012. Each campaign consisted of collocating two USDA-ARS designed low-volume total suspended particulate (TSP), one Thermo Scientific (East Greenbush, NY) low-volume ambient PM10, and one Thermo Scientific low-volume ambient very sharp cut cyclone PM2.5 at four sites per feedyard. Sampling sites were located within feedyard boundaries in unoccupied pens or other open spaces within the feedyard. If possible, these sites were located toward 8797

DOI: 10.1021/acs.est.5b01881 Environ. Sci. Technol. 2015, 49, 8796−8803

Article

Environmental Science & Technology

Figure 1. Particulate matter (PM) concentrations by size fraction at five commercial cattle feedyards during sampling campaigns from 2010 to 2012. Lowercase letters denote significant differences (p < 0.05) among feedyards for each PM fraction. Note the varying y-axis scale for each size fraction.

melengestrol acetate-d3 varied independently. The consistently lowered response in high mass samples is indicative of matrix suppression, which can be accounted for by the use of appropriate standards. Additionally, to determine potential steroid degradation during sample collection and processing, concentrations of steroids in spiked filters were determined following 24 h sampling and 24 h processing procedures (details in the SIOn-Filter Steroid Degradation). Statistical Analysis. Statistical analyses were performed using R statistical software.26 For all tests, α was set at 0.05 to determine statistical significance. Meteorological conditions and steroid concentrations in PM among or within feedyards were analyzed by one-way or two-way analysis of variance (ANOVA). Further posthoc analysis was conducted using Tukey’s test. Pearson’s chi-squared test was used to assess differences in steroid detection frequencies among feedyards and filter types. Pearson’s product-moment correlation was used to assess relationships between PM concentrations and meteorological conditions or concentrations of steroids in PM. For statistical analysis of steroid concentrations, nondetections were assigned a value of one-half the limit of quantitation (0.25 ng/filter) for calculation of steroid concentration, thereby, a distribution of steroid concentration was created for nondetections ranging from 0.06−2.5 ng/g PM.

(Beckman Coulter, Inc., Brea, CA). Aerodynamic equivalent diameter (AED) was determined assuming particle density of 2.65 (g/cm3) and dynamic shape factor of 1.40. Steroid Hormone Analysis. Following mass determination, filters for steroid analysis were stored at −80 °C until extraction. Steroid hormones were extracted from filters and analyzed as described previously,23 with a slight modification to separate isomers of trenbolone and estradiol.25 Briefly, filters were extracted using liquid−solid extraction, Florisil cleanup, and analyzed by liquid chromatography tandem mass spectrometry using atmospheric pressure chemical ionization (LC−APCI− MS) with isotope dilution. Filters were spiked with 20 ng of 17βtrenbolone-d3, 17β-estradiol-d5, and melengestrol acetate-d3 prior to extraction. Final sample extracts were reconstituted in 100 μL of mobile phase. Instrument operation, data acquisition and processing were performed using Xcalibur 2.1 software (Thermo Fisher Scientific, San Jose, CA). Quality Assurance. Reagent blanks, method blanks, and laboratory spikes (5 ng/filter) were analyzed with each sample batch to ensure accurate identification of target compounds. No steroids were observed in reagent or method blanks above detection limits. Determined laboratory spike concentration averaged 89−106% with a RSD < 13.6% for all compounds. Internal standard 17β-trenbolone-d3 was used for quantification of all trenbolone metabolites, 17β-estradiol-d5 was used for quantification of all estrogens, and melengestrol acetate-d3 was used for MGA quantification. As internal standards were added prior to extraction, standard response can be influenced by both extraction efficiency and matrix effects. Response of internal standards was tracked for all samples and used to develop a quality control chart of internal standard response (Supporting Information, SI, Figure S1). Responses falling outside of the mean ±2SD were flagged and further examined to determine data quality. QC rejection rate based upon response of all internal standards was 16.5%. PM samples varied greatly in mass, ranging from 0.135−392 mg, with a mean ± SD of 42.3 ± 60.4 mg. High mass samples consistently resulted in lowered internal standard response; thus, filters with a mass one SD above the mean were not excluded based solely on low internal standard response. Conversely, low mass filters under 10 mg were excluded from steroid analysis as these low mass samples were expected to fall below limits of detection. Results of internal standard response highlight the importance of using appropriate labeled standards, as the response of 17β-trenbolone-d3, 17β-estradiol-d5, and

■

RESULTS AND DISCUSSION PM Characterization. Due to occasional sampler malfunctions and power outages, useable samples were not always obtained from all sites at each feedyard. TSP concentrations ranged from 21−9380 μg/m3 with an overall mean of 1403 ± 1763 μg/m 3 across all five feedyards (Table 1). PM concentrations varied significantly (p < 0.001) among feedyards (Table 1; Figure 1). A total of 434 TSP filters from all feedyards were collected for PSD analysis. AED mass median diameter (MMD) had an overall mean of 32.5 μm, but varied significantly among feedyards (Table 1). PM10 comprised a mean of 19.0% of TSP while PM2.5 comprised 1.74% of TSP. The relative percentage of both PM10 (p < 0.001) and PM2.5 (p < 0.001) was significantly different among feedyards (Table 1). TSP concentrations determined in this study were within the range of concentrations reported elsewhere. Concentrations in the current study were higher than those reported by Sweeten et al., which ranged from 84−1762 μg/m3 at three Texas feedyards;27 however, Sweeten et al. collected PM samples 8798

DOI: 10.1021/acs.est.5b01881 Environ. Sci. Technol. 2015, 49, 8796−8803

Article

Environmental Science & Technology

Figure 2. Occurrence of steroid growth promoters in airborne particulate matter (PM) among PM size fractions (A) and among five commercial cattle feedyards (B) collected from 2010 to 2012. Lowercase letters denote significant differences (p < 0.05) among sample types or feedyards for each compound.

30−100 ft downwind from feedyards. PM concentrations are generally expected to decrease with increasing distance from feedyards. TSP concentrations reported here were lower than those reported by Elam et al.28 (14 200 ± 11 815 μg/m3; TSP collected 2 h following sundown at 10 feedyards). An increase in PM concentration occurs at cattle feedyards close to sundown resulting from increased cattle activity and more stable atmospheric conditions as daytime temperatures drop.29−31 Concentrations reported by Elam et al.28 were collected during this period of high cattle activity and were expected to be higher than the current study. PM concentrations from the current study may have also been elevated due to an extended period of drought which spanned the 16 month sampling campaigns. During this period, meteorological stations nearest to feedyards received an average of less than 240 mm of total rainfall. PM emissions have been negatively correlated with increasing pen surface moisture content,27,32−34 thus extreme drought conditions likely contributed to elevated TSP concentrations. Log-transformed TSP concentration was correlated with mean temperature (r = 0.602, p < 0.001, SI Figure S2), RH% (r = −0.466, p < 0.001, SI Figure S2), and wind speed (r = 0.104, p = 0.029, SI Figure S2) at the time of PM collection. All three parameters had significant effect upon PM concentration prediction (p < 0.001). PM concentrations were significantly different among feedyards (Table 1; Figure 1), yet there were no significant differences in temperature (p > 0.8) or wind speed (p > 0.4) among feedyards. RH% was significantly different between only Feedyard A and Feedyard D; however, TSP concentrations were not significantly different between these two feedyards. Other factors for which data were not available may have

influenced TSP concentration among feedyards. Increased pen surface manure density has been associated with lower PM10 emissions.32,33 Pen surface conditions would be determined primarily from individual feedyard manure management practices, which likely varied across feedyards. Particle size distribution of TSP collected from feedyards varied among the five feedyards (Table 1). The four high capacity feedyards had similar MMD, but the single, low capacity feedyard (Feedyard E) had significantly larger MMD indicating PM was composed of larger particles. The MMD of TSP from all feedyards is larger than reported for other studies at cattle feedyards; however, methods of sample collection and PSD analysis vary across studies. Sweeten et al. reported MMD of 9.54 ± 1.45 μm from samples collected 30−100 feet outside feedyard boundaries.31 Guo et al. reported a geometric mean diameter 7− 18 μm but also collected samples outside of feedyard boundaries.33 In the present study, PM was collected from within feedyard boundaries and would be composed of a greater proportion of large, coarse particles. PM sampled outside feedyard boundaries may be composed of fewer large particles due to rapid settling and deposition back onto feedyard surfaces, which can explain the observed differences among studies. The percentage of PM composed of PM10 and PM2.5 fractions also varied significantly among feedyards (Table 1). Similar to TSP concentration and MMD, Feedyard E had the lowest fraction of TSP present as PM10, significantly lower than three of four other feedyards. This parallels MMD discussed above, as Feedyard E PM was composed of significantly larger particles. Unlike PM10, the fraction of TSP present as PM2.5 showed no direct relationship with feedyard size. Observed percentage of 8799

DOI: 10.1021/acs.est.5b01881 Environ. Sci. Technol. 2015, 49, 8796−8803

Article

Environmental Science & Technology

Figure 3. Concentration (ng/g) of steroid growth promoters in airborne particulate matter (PM) among PM size fractions (A) and among five commercial cattle feedyards (B) collected from 2010 to 2012. Lowercase letters denote significant differences (p < 0.05) among sample types or feedyards for each compound.

Concentrations of steroid growth promoters in PM were in the low ng/g range for all compounds (Figure 3, SI Table S2). The median concentration of the predominant TBA metabolite 17αTb was 1.90 ng/g and for TbO was 0.79 ng/g PM. Concentrations of both 17αTb and TbO varied significantly among PM size fractions, and TbO concentrations varied significantly among feedyards (Figure 3). Estrogens were present at the highest concentrations of all measured steroids, and median concentration of 17αE2 was 20.61 ng/g PM. The stereoisomer 17βE2 was detected less frequently and at a median concentration of 0.96 ng/g PM. E1 was detected at the highest frequency with a median concentration of 10.84 ng/g PM. Concentrations of all three estrogens were significantly different among feedyards, but only 17αE2 and 17βE2 concentrations varied significantly among PM size fractions (Figure 3). MGA was detected at relatively low concentrations, with a median concentration of 1.27 ng/g PM. Concentrations of MGA varied significantly among feedyards, but were similar among PM size fractions (Figure 3). It should be noted that based upon degradation experiments, steroids, especially trenbolone metabolites and melengestol acetate, could potentially be degrading during the sample collection process. The longer collection times for PM10 and PM2.5 increases this potential degradation in those PM size fractions (see details in the SIOn-Filter Steroid Degradation). Using only PM samples with complementary detections of steroids, concentrations of several steroids were positively correlated. Concentration of 17αTb was positively correlated with 17αE2 (n = 178, r = 0.31, p < 0.001, SI Figure S3) and MGA (n = 85, r = 0.70, p < 0.001, SI Figure S3); concentration of

PM10 and PM2.5 fractions of TSP were lower than studies where PM was collected outside of feedyard boundaries. Sweeten et al. reported PM10 and PM2.5 as 33.2% and 5%, respectively, of TSP31 while Guo et al. reported 41% and 10%, respectively, of TSP.33 This may again be directly attributable to sampling location, as the present study collected PM within feedyard boundaries. The differences in respirable fractions of PM collected within feedyard boundaries compared to PM collected outside boundaries highlights the importance of small PM fractions on potential PM transport. The relatively large size of most feedyard derived PM prohibits transport over long distances but will increase deposition in areas adjacent to feedyards. PM10 and PM2.5 fractions have potential to be transported much greater distances and are associated with a myriad of adverse human health effects.35,36 Steroid Occurrence. A total of 456 (289 TSP, 134 PM10, 33 PM2.5) samples were collected from 5 feedyards and analyzed for growth promoters. Detection frequencies of individual steroids among sample types and feedyards are summarized in Figure 2 and SI Table S1. TBA metabolites 17αTb and TbO were detected in 39.3% and 21.3% of filters, respectively, while 17βTb was detected in only 0.4% of filters. As such, 17βTb is not addressed further in the following results. Estrogens were detected at the highest frequency of any compounds with 17αE2, 17βE2, and E1 detected in 94.1%, 24.6%, and 98.5% of filters, respectively. The progestin MGA was detected in 31.8% of filters. Detection frequency of 17αTb, 17αE2, and MGA was significantly different among PM size fractions (Figure 2). Detection frequency for all compounds was significantly different among feedyards (Figure 2). 8800

DOI: 10.1021/acs.est.5b01881 Environ. Sci. Technol. 2015, 49, 8796−8803

Article

Environmental Science & Technology

Table 2. Emission Factors of Steroid Hormones Estimated Using PM Emission Factors from Bonifacio et al.42 and Median Steroid Concentration Observed for Each Sample Type (See SI Table S2)a

a

PM fraction

PM emission factor (g/hd/d)

trendione

17α-trenbolone

estrone

17β-estradiol

17α-estradiol

melengestrol acetate

TSP PM10 PM2.5

57 21 11

42 17 15

121 31 19

611 225 128

49 22 19

1126 440 292

72 26 16

Values given for steroids are presented as ng/hd/d.

17αE2 was positively correlated with E1 (n = 427, r = 0.46, p < 0.001, SI Figure S3). No other significant correlations were observed among steroid concentrations. TBA and 17βE2 are commonly coadministered to cattle via steroidal implants,3 thus the correlation between 17αTb and 17αE2 is not unexpected. E1 is a secondary excreted estrogen in cattle25 and the primary biotransformation product of 17αE2,37,38 and accordingly a strong relationship was observed between these two steroids. MGA is administered in feed to heifers only; therefore, its use may vary independent of TBA and 17βE2 in steroid implants. Although 17αTb and MGA were concurrently detected in fewer samples than other correlated steroids, the strong relationship may suggest steroid implants are frequently coadministered to heifers receiving MGA. Data on feedyard steer to heifer ratios, implant strategies, or implantation schedules were not available, and each of these factors could potentially alter steroid concentrations in PM. Concentrations of steroids were further compared with meteorological conditions to determine potential influences of seasonal variation on steroid concentration. Analysis was performed using only 17αE2 and E1 to avoid potential artifacts caused by the high percentage of nondetections in other analytes. Mean temperature was positively correlated with 17αE2 (r = 0.526, p < 0.0001, SI Figure S4) and E1 (r = 0.487, p < 0.0001, SI Figure S4) concentration while RH% was negatively correlated with 17αE2 (r = −0.222 p < 0.0001, SI Figure S4) and E1 (r = −0.188, p < 0.0001, SI Figure S4) concentration. These correlations suggest hot, dry conditions will increase steroid concentrations in PM in addition to increasing airborne PM concentration, as previously discussed. A study by Khan and Lee39 demonstrated a decrease in trenbolone metabolite degradation in agricultural soils under high temperatures with low soil moisture. The results observed for 17αE2 and E1 in PM follow a similar trend, and degradation rates may partly explain seasonal variability of steroid concentrations. Another consideration with changing meteorological conditions is PM source. As discussed regarding airborne PM concentrations, PM emissions are negatively correlated with pen surface moisture content,27,32−34 thus it is possible that with increasing temperature and decreasing RH%, a greater proportion of total collected PM was derived from pen surfaces. Initial research on steroid transport via PM by Blackwell et al.22 reported frequent detections of 17αTb and TbO at concentrations up to 38.8 ng/g PM. Mean concentrations of 17αTb and TbO in the current study are slightly lower than those reported by Blackwell et al.;22 however, the current study employed different PM collection methods. The previous study collected TSP samples near sundown when PM concentrations are greatest and when most airborne PM arises from pen surfaces. The current study collected TSP samples for a duration of 48 h or longer and throughout daytime periods characterized by increased intrafeedyard road traffic and reduced cattle activity. These conditions likely contributed to a greater proportion of PM originating from outside cattle pens, and therefore lower

concentrations of steroids than were observed in the previous study. Studies on the origin of feedyard PM by Huang et al.40 and Lange et al.41 have identified pen surfaces as the primary contributor to PM10 emissions, comprising an average of 78%40 and 41%41 of PM10, respectively. Therefore, PM steroid concentrations should be similar or less than those from pen surfaces as a result of dilution with PM from other sources. Other studies have examined steroids from fresh manure, pen surfaces, and in runoff from surfaces. Concentrations of 17αTb in PM are lower than fresh manure from implanted cattle, reported at 21−68 ng/g dry weight,21,25 but consistent with pen surface concentrations reported by Webster et al.,21 who documented 17αTb at 0.6−11.8 ng/g in surface soils of pens housing implanted steers. Median concentrations of 17αE2 in PM were higher than in pen surface soils as reported by Mansell et al.20 (