Characterization of Ambient PM10 Bioaerosols in a California Agricultural Town

Desert Research Institute, 2215 Raggio Parkway, Reno, Nevada 89512, USA The State Key Laboratory of Loess and Quaternary Geology, Institute of Earth Environment, Chinese Academy of Sciences, Xi’an, Shaanxi, 710075, China Graduate Faculty, University of Nevada, Reno, Nevada 89503, USA Department of Animal and Food Sciences, University of Delaware, 237 Townsend Hall, Newark, DE 19716, USA 5Department of Environmental and Occupational Health, University of Nevada, Las Vegas 89154, USA Corresponding author. Tel.: +1 775 674 7050; fax: +1 775 674 7009; email address: Judith.Chow@dri.edu

This study characterizes PM 10 (particles with aerodynamic diameters <~10 µm) bioaerosols at three neighborhoodscale sites (see Fig. 1) in the town of Corcoran on the edge of the Tulare dry lake in California's San Joaquin Valley (SJV).The SJV is a large agricultural area that produces cotton, oranges, grapes, almonds, milk, cattle, and poultry.The Corcoran area is dominated by production of Egyptian cotton and hosts major cotton processing facilities for the region.The Corcoran-Patterson (COP) site (CARB, 2015b) has measured PM 10 compliance in the SJV airshed from 1996 to present, and has shown many 24-hour averages higher than the California state limit (50 µg/m 3 ) (CARB, 2015a).This study intends to: 1) assess ambient bioaerosol Fig. 1.Locations of Corcoran (36°05′53′′N, 119°33′37′′W) in Central California and the three neighborhood-scale (within 1.7 km) sampling sites (insert).The town of Corcoran lies in the central San Joaquin Valley (SJV) with a population of ~15,000.The COP site was located in a local school yard surrounded by residential communities, the BAI site was near a cotton processing facility, and the GRA site was near grain elevators.Locations for previous bioaerosol studies in the SJV are also shown, including Modesto, Fresno, and Tulare County.concentrations on PM 10 samples in the vicinity of agricultural activities; 2) estimate the contributions of stable bioaerosol indicators, specifically fungal spores, pollen grains, and plant detritus to PM 10 mass and OC concentrations; and 3) examine the association between fungal spores and the three fungal biomarkers (i.e., (1→3)-β-D-glucan, arabitol, and mannitol).
Extensive efforts have been made to characterize SJV air pollution, especially for PM 10 and PM 2.5 (particles with aerodynamic diameters <~2.5 µm; Chow et al., 1993a;1996;2006;Chen et al., 2007).The PM 2.5 fraction is dominated by ammonium nitrate and carbonaceous aerosol during fall and winter.Sulfur dioxide emissions in the SJV are low (CARB, 2012), but oxides of nitrogen and ammonia emissions are high (Mansell and Roe, 2002), so lower temperatures and higher relative humidities during late fall and winter favor ammonium nitrate formation (Stockwell et al., 2000;Chow et al., 2005b;Lurmann et al., 2006;Chow et al., 2008).PM 2.5 carbonaceous aerosol in the SJV derives from engine exhaust, biomass burning, cooking, and conversion of organic gases to particles (Chow et al., 1992;Strader et al., 1999;Schauer and Cass, 2000;Chow et al., 2007b).
The PM 10 coarse fraction (PM 10-2.5 ) contains large contributions from fugitive dust and possibly bioaerosols.During citrus harvesting in Tulare County (see Fig. 1), Lee et al. (2004) reported average endotoxin and total bacteria and fungi levels of 293.2 endotoxin units (EU)/m 3 and 1.9 × 10 8 organisms/m 3 , respectively, based on the total suspended particles (TSP) dislodged from polycarbonatemembrane filters.They also measured 13,787 and 13,274 colony forming units (CFU)/m 3 for culturable bacteria and fungi, respectively, using Andersen two-stage bioaerosol samplers.Tager et al. (2010) found yearly average PM 10 endotoxin concentrations of 0.98-1.38EU/m 3 in Fresno (the largest city in the SJV, 80 km north of Corcoran) from 2001 to 2004, with higher concentrations during the dry season (i.e., May-October).Using the 16S rRNA clone library and Sanger sequencing, Ravva et al. (2011) observed phyla Firmicutes, Proteobacteria, and Bacterioidetes that dominated the bacterial community of TSP samples collected around two dairy farms near Modesto, CA (~140 km north of Corcoran).
None of the prior SJV studies examined contributions of bioaerosol to the mass loadings and carbonaceous fractions in PM 10 samples used to determine compliance with air quality standards.These prior studies focused on urban populations.Exposures in smaller communities scattered throughout the SJV may report higher concentrations owing to their proximity to crops and livestock.In this study, seven types of stable bioaerosol indicators were quantified on PM 10 samples using a combination of microscopy (i.e., counts of fungal spores, pollen grains, and plant detritus), anion exchange chromatography (i.e., arabitol and mannitol), and Limulus Ambebocyte Lysate (LAL) assays (i.e., endotoxin and (1→3)-β-D-glucan).
Pollen, spores, and plant parts derive from native, agricultural, and ornamental vegetation, and are the most widely recognized allergens (Bowers et al., 2013;Caillaud et al., 2014).Pollen count forecasts (Intermountain Allergy and Asthma, 2014;Pollen.com, 2014) are widely followed by people with allergies.In spite of the importance of bioaerosols to human health, minimal effort has been expended in simulating their emissions, transport, and human exposure (Raynor et al., 1983;Luo et al., 2006).Garfin et al. (2013) speculate that as the regional climate warms over the next several decades, earlier and longer spring bloom for many plant species may lead to enhanced pollen production.These substances have high ligno-cellulose contents and are stable over many years, as evidenced by their use in describing long-term climate change effects on vegetation (Rhode, 2003;Louderback and Rhode, 2009).
Endotoxin is a cell wall component of gram-negative bacteria, exposure to which may cause fever, shivering, pulmonary inflammation, non-allergenic asthma, airway obstruction, and lung function deterioration (Degobbi et al., 2011).High endotoxin levels have been measured near animal houses and other agricultural activities (Smid et al., 1992;Reynolds et al., 2002;Ko et al., 2010;Yang et al., 2013).Endotoxins are often associated with agricultural and house dusts, and they persist for long periods because the bacteria are no longer viable (Pearson et al., 1985;Maus et al., 2001).
(1→3)-β-D-glucan is a fungal cell wall component with health effects similar to endotoxins (Rylander et al., 1999;Douwes et al., 2000;Douwes et al., 2003), and it can be used as a surrogate for fungal exposure.(1→3)-β-D-glucan also attaches to fugitive dust and persists with time (Douwes et al., 1996), even being found on Asian dust aerosols that have transported over many days and long distances (He et al., 2013).
Arabitol and mannitol are sugar alcohols that are used as surrogates for fungi (Bauer et al., 2008a;Di Filippo et al., 2013).They have no known adverse health effects, are stable owing to their low vapor pressures, and are more efficiently measured than the microscopic identification and counting needed for the large variety of fungi.They are included in this study to determine how well they might be predictors of the fungi concentrations.

PM 10 Sampling and Chemical Analysis
Daily, 24-hr (midnight to midnight Pacific Standard Time) PM 10 samples were acquired during the cotton harvest season from 9 October, 2000, to 9 November, 2000, at the three sites in Fig. 1.Compliance PM 10 levels exceeded 50 µg/m 3 on five of the six every-sixth-day hivol filter samples acquired at COP during this period.Two PM 10 MiniVol samplers (Airmetrics Inc., Eugene, OR, USA) equipped with 47 mmdiameter Teflon-membrane (Teflo TM R2PJ047, Pall Corp., Ann Arbor, MI, USA) and quartz-fiber filters (Tissuquartz TM 2500 QAT-UP, Pall Corp., Ann Arbor, MI, USA) were operated at 5 L/min flow rates at each site with an inlet height of 3.3 m above ground level (Chow et al., 2005a).
Prior to sampling, quartz-fiber filters were prefired at 900°C for 4 hr while Teflon-membrane filters were equilibrated in a temperature (20-23°C) and relative humidity (30-40%) controlled environment for a minimum of four weeks prior to gravimetric analysis with a Mettler Toledo XP6 microbalance (sensitivity: ± 1 µg; Mettler Toledo Inc., Columbus, OH, USA).After sampling, filters were shipped and stored in airtight containers at < 4°C.Mass by gravimetry and elements by x-ray fluorescence (XRF) were measured on the Teflon-membrane filters soon after sampling (Watson et al., 1999).Portions of the quartz-fiber filters were analyzed for water-soluble ions by ion chromatography (Chow and Watson, 1999) and organic and elemental carbon (OC and EC) by thermal/optical reflectance (Chow et al., 1993b).Remaining filters were stored at <4°C in sealed Petri slides until they were re-analyzed for the bioaerosols.Since only stable markers were sought, as described above, this type and length of storage is not believed to cause a major bias in the concentrations.Chow et al. (2007a) demonstrate that (OC) concentrations can be reproduced on similar samples stored for many years.However, the results reported here represent a lower limit to the total bioaerosol content of Corcoran PM 10 .

Microscopic Analysis of Fungal Spores, Pollen Grains, and Plant Detritus
Microscopic analysis of particle morphology was performed with a Hitachi TM-1000 Tabletop SEM (Hitachi High-Technologies Corp., Tokyo, Japan).The Teflonmembrane filter was mounted on a metal disk specimen holder using conductive carbon tape (attached via the plastic support ring only).Carbon or gold coating was not needed due to the use of backscatter electrons and operation by charge compensation.Both standard (500-1000x) and high magnification (1000-5000x) were used to identify specific particle shapes and other characteristics.Fourteen to twentytwo images of randomly selected positions on the aerosol deposit were obtained for each filter, followed by image analysis using ImageJ software (Schneider et al., 2012;Wagner and Macher, 2012).Bioaerosols were identified based on their size, shape, and texture, and classified by category before being manually counted.The minimum detection limit (MDL) was estimated to be ~200 particles/m 3 .The precision of bioaerosol counts at the 90% confidence interval was ± 14% for fungal spores, ± 55% for pollen grains, and ± 22% for plant detritus.Pollen grain and plant detritus counts were low, so counting precisions were also low.

Anion-Exchange Chromatography Analyses of Arabitol and Mannitol
Quartz-fiber filter sections with deposit areas of 4.34 to 5.36 cm 2 (the remains of the filter after other analyses) were extracted in 5.0 mL of pyrogen-free water by sonication and gentle shaking for 60 min at room temperature.The extract was analyzed by high-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD; Dionex ICS-3000, Sunnyvale, CA, USA) (Iinuma et al., 2009), as detailed in the Supplemental Materials (S2).The MDLs and precisions were 19 ng/m 3 and ± 3.2% for arabitol and 18 ng/m 3 and ± 2.6% for mannitol, respectively.

PM 10 Mass and Chemical Composition
PM 10 mass, major elements, ions, carbon, and bioaerosol concentrations are summarized in Table 1.Average PM 10 mass ranged from 65.1 ± 1.0 µg/m 3 at the residential COP site to 82.1 ± 28.0 µg/m 3 at the cotton handling BAI site.The material balances in Fig. 2 show that geological material accounted for 35% (COP site) to 50% (BAI site) of measured PM 10 .Organic mass [OM = 1.4 × OC to account for unmeasured hydrogen, nitrogen, sulfur, and oxygen, see Watson (2002) for justification] and nitrate were the next largest components, accounting for 18-22% and 15-24% of PM 10 , respectively.Ammonium sulfate, EC, salt (available from Tulare dry lake bed, Chow et al., 2003), and trace elements constituted the remaining PM 10 mass.
Average bioaerosol concentrations were not that different among the sites.COP measured more than 80% of the concentrations at the BAI site for (1→3)-β-D-glucan, arabitol, mannitol, and pollens, and more than 60% for plant detritus.Similar large fractions were found for the comparison of COP with the GRA values, with pollen grains reduced to 63% and plant detritus to 50%.The largest differences were found for endotoxins, with COP measuring only 35% of the BAI average and 13% of the GRA average.This indicates that most of the bioaerosols are not dominated by the nearby source, but affect the broader Corcoran community.Average endotoxin concentrations of 1.7 ± 1.5 EU/m 3 at the COP site were similar to the 0.98-1.38EU/m 3 reported by Tager et al. (2010) for Fresno.Dungan et al. (2010) measured endotoxin near an open-feedlot dairy farm in southern Idaho and found that the average endotoxin concentration decreased by ~50% and ~86% at 200 m and 1,390 m downwind of the edge of the farm, respectively.Average endotoxin concentrations at the GRA site (13.0 ± 17.0 EU/m 3 ) were 2.7 and 7.6 times higher than concentrations at the BAI and COP sites, respectively.The highest concentration of 47.6 EU/m 3 (23 October, 2000, at the GRA site) was about half of the of 90 EU/m 3 exposure limit in the Netherlands (Dutch Expert Committee on Occupational Safety, 2010), the only country that defines such a limit.
Average (1→3)-β-D-glucan concentrations of 8.5-10.6 ng/m 3 were higher than those reported by Chen and Hildemann (2009) for indoor (0.1-8.9 ng/m 3 ) and outdoor (0.3-5.4 ng/m 3 ) environments in urban California and by Menetrez et al. (2009) at a wooded rural site (0.04 ± 0.017.4ng/m 3 ) in Orange County, NC.As shown in Table 2, average arabitol and mannitol concentrations of 170 and 132 ng/m 3 , respectively, in Corcoran were 1-2 orders of magnitude higher than those from past studies, except for a similar site near agricultural activities in central India (Nirmalkar et al., 2015).Fungal spore and pollen grain concentrations vary by over 3-4 orders of magnitude among studies listed in Table 3.With intense agricultural activities, counts of fungal spores (66,333 particles/m 3 ) and pollen grains (2,600 particles/m 3 ) at Corcoran were more than twice those found elsewhere.
Fungal spores and pollen grains are large particles in the range of 5-10 µm in diameters as shown in Table 4. Their concentrations, as well as the concentrations of fungal biomarkers including (1→3)-β-D-glucan, arabitol, and mannitol, were high at the three neighborhood sites.Table 4 shows that fungal spores were of spherical and prolate spheroid, with prolate spheroids accounting for > 90% of spore counts.The surface area and volume of pollen grains were 4 and 8-10 times larger than those of fungal spores, respectively.
Plant detritus particle counts were variable, ranging from 9,000 to 81,000 particles/m 3 or ~1-9.2 µg/m 3 at the three sites, with diverse morphology.Identification of plant detritus is uncertain due to their variable morphologies, absence of distinct texture, and surface attachment of fungal hyphae, fiber, and film.Examples of plant detritus, diatoms, bacteria (e.g., Streptococcus), slime molds (e.g., Myxomycete spores), and insect eggs from the SEM analysis are shown in Fig. 3.

Predictability of Bioaerosol Indicators from Each Other and from Major PM 10 Components
Correlation coefficients (r) indicate the extent to which different concentrations vary with each other, either because they are in the same particles, in different particles deriving from the same source, or affected by the same meteorology (Watson and Chow, 2015).A high correlation (r > ~0.85) indicates that one variable might be reliably predicted from   ).
e Mass concentration of fungal spores, pollen grains, and plant detritus are estimated based on their average volume per particle determined by scanning electron microscopy (SEM) analysis (See Table 4), assuming a density of 1 g/cm 3 .Highly variable morphologies of plant detritus particles (average volume per particle: 113.6 ± 318.7 µm 3 , N = 172) preclude an accurate assessment.the other, while moderate correlations (r = ~0.5 to 0.85) indicate some predictive ability, but with high uncertainty.
Low correlations (r < 0.5) are not useful for predictive purposes, even though statistical tests might show the relationships are significant.Table 5 demonstrates that correlations are low for most of the bioaerosols and for the more commonly measured PM 10 components.The highest correlation (r = 0.86) is between arabitol and mannitol.Similarly high arabitol/mannitol correlations are reported in other studies (Bauer et al., 2008a;Zhang et al., 2010;Burshtein et al., 2011;Liang et al., 2013).Arabitol and mannitol are only moderately correlated with fungal spores (r = 0.51 and 0.49, respectively).Bauer et al. (2008a) found slightly stronger relationships, with a fungal spore correlation with arabitol of r=0.56 and with mannitol of r = 0.62 OC is moderately correlated with (1→3)-β-D-glucan (r = 0.7) and plant detritus (r = 0.57).Plant detritus is moderately correlated with endotoxin (r = 0.59) and (1→3)-β-D-glucan (r = 0.60).None of the markers are associated with the geological minerals, probably indicative of its derivation from multiple SJV fugitive dust emitters (Ashbaugh et al., 2003;Carvacho et al., 2004).
Average arabitol and mannitol concentrations per fungal spore were 2.7 ± 1.0 pg/spore and 2.1 ± 0.9 pg/spore, 125% and 24% higher than the 1.2 ± 0.5 pg/spore and 1.7 ± 0.6 pg/spore reported by Bauer et al. (2008a), respectively.These ratios are expected to vary owing to differences in the fungal spore types, presence of arabitol and mannitol in other bioaerosols, and artifacts from different sampling and analysis methods.
(2013) examined size-segregated PM samples collected in urban/suburban Rome, Italy, and found different size distributions for arabitol and mannitol as compared to ergosterol.Arabitol and mannitol were correlated with levoglucosan and xylitol -two biomass burning markersin sub-micron size fractions.Both DiFilippo et al. (2013) and Burshtein et al. (2011) recommend using ergosterol as a surrogate for fungal counts.However, Yang et al. (2012) observe that ergosterol is susceptible to molecular degradation and instability.Lau et al. (2006) found that the ergosterol content per fungal spore varied with fungal species and their growth stage.With moderate correlations (0.49-0.51) found between fungal spore counts and these two sugar alcohols, this study is consistent with other findings that arabitol and mannitol are not reliable surrogates for fungi.Although fungi appear to be the dominant bioaerosol (Elbert et al., 2007;Winiwarter et al., 2009), concurrent release of pollen grains, plant detritus, diatoms, bacteria, and other sugar-alcohols during agricultural activities complicate predictability among different fungal biomarkers.

Contribution of Fungal Spores, Pollen Grains, and Plant Detritus to PM 10 mass and OC
Bioaerosol densities are difficult to measure; reported density values range from 0.9 to 1.5 g/cm 3 depending on the bioaerosol type (Burge, 1995;Cox and Wathes, 1995).An assumption of 1 g/cm 3 density (Johnson et al., 1999;Matthias-Maser and Jaenicke, 2000), 20% water in fresh mass, and 50% OC in dry mass for fungal spores (Bauer et al., 2008b;Wiedinmyer et al., 2009) is used to estimate their contributions to PM 10 mass and OC.As shown in Table 6, fungal spores, on average, accounted for 5.4-5.8% of PM 10 mass and 11.5-14.7% of PM 10 OC, with the highest ratios (9.9% of PM 10 mass and 27.1% of PM 10 OC)   140.1 ± 231.0 a N refers to the number of particles selected for size measurement by scanning electron microscopy (SEM).b The volume of prolate spheroid particles is calculated as: V = 4/3πa 2 b, where a is the minor axis and b is the major axis of a prolate spheroid; the surface area is calculated as: , where  found at the cotton handling BAI site on 24 October, 2000.This is consistent with fungal spore fractions of 0.4-64% and 4-22% in PM 10 and PM 2.5 OC, respectively, reported in previous studies (see Table 7).Using the conversion factor of 13 pg C/spore by Bauer et al. (2002b) would underestimate the fungal spore OC to total OC ratio by 46%.This is expected since fungal spore OC content should be size dependent, and larger fungal spores (58 µm 3 /spore) were found in this study (Table 4) compared to 34 µm 3 /spore by Bauer et al. (2002b).PM 10 contributions from pollen grains were lower than those from fungal spores, accounting for an average of 1.9-2.7% of PM 10 mass (4.2-5.4% OC) with a maximum of 6.3% PM 10 (11.6% OC) at the GRA site on 9 October, 2000.Plant detritus particles accounted for an average of 3.8-6.3%PM 10 mass (8.1-13.5% OC) with the maximum contribution of 11.5% PM 10 (24.1% OC) found at the cotton handling BAI site on 22 October, 2000.The sum of fungal spores, pollen grains, and plant detritus accounted for averages of 11.2-14.8%PM 10 mass (23.8-32.8%OC), ranging from 5.3% PM 10 mass (19.1% OC) at the BAI site to 24.1% PM 10 mass (30.1% OC) at the GRA site.

CONCLUSIONS
As abundances of bioaerosols in PM 10 near agricultural ( Johnson et al., 1999;Matthias-Maser and Jaenicke, 2000) and contain 20% of water in fresh mass and 50% of carbon content in dry mass (Bauer et al., 2008b;Wiedinmyer et al., 2009).The contributions of fungal spores, pollen grains, and plant detritus to PM 10 mass can be calculated as: N × V × ρ/PM 10 × 100%, where N is the number concentration of the bioaerosol, V is the average volume of bioaerosol particles (See Table 4), ρ is the particle density, and PM 10 is the PM 10 mass concentration.Their contributions to PM 10 OC can be calculated as: × C%]/OC × 100%, where H 2 O% is the water content in the bioaerosol's fresh mass, C% is the carbon content in its dry mass, and OC is the organic carbon concentration.
communities may contribute to excessive PM 10 levels, this study demonstrated the feasibility of using archived (< 4°C) filter samples to evaluate the contribution of the seven stable bioaerosol indicators.Except for the uncertainties of SEM analyses on Teflon-membrane filter samples, adequate minimum detection limits and reproducibility (~3-6%) were found.Daily bioaerosol concentrations varied over tenfold during the study period and were over twofold those reported in the literature for fungal spores (66,333 particles/m 3 ) and pollen grains (2600 particles/m 3 ) and among the highest for arabitol (170 ng/m 3 ) and mannitol (132 ng/m 3 ).Zones of influence from sources within the three-site network are ~2 km, except for endotoxin, demonstrating the neighborhoodscale influence of agricultural bioaerosols.Fungal spores were the dominant bioaerosol, accounting for 11.5-14.7% of PM 10 OC, followed by plant detritus (8.1-13.5% OC) and pollen grains (4.2-5.4% OC).Correlations between fungal spore counts and the three most commonly used biomarkers for fungi (i.e., (1→3)-β-D-glucan, arabitol, and mannitol) were low to moderate (0.36 < r < 0.51), casting doubt on the generality of their use as surrogates.Arabitol and mannitol were not correlated with (1→3)-β-D-glucan (r < 0.11), probably due to their variable sources.The specificity and applicability of these species as biomarkers for fungal exposure and their applications for estimating the count or mass of fungal is limited.Bioaerosols from different sources may differ in particle density, water content, and carbon content.To achieve a more accurate estimate of bioaerosol contributions to PM 10 , efforts are needed to determine the density and composition of dominant bioaerosol species.D. and LeMasters, G. (2006)

S1. Endotoxin and (1→3)-β-D-glucan Analyses
Endotoxin and (1→3)-β-D-glucan analyses used kinetic chromogenic limulus amebocyte lysate (LAL) assays with different activating factors: factor C for endotoxin (in Chromo-LAL assay) and factor G (in Glucatell assay) for (1→3)-β-D-glucan.The assays consist of a colorless substrate and a proenzyme extracted from amoebocyte cells in the blood of the horseshoe crab, Limulus polyphemus.The proenzyme is converted to an active enzyme with the presence of endotoxin or (1→3)-β-D-glucan.The enzyme then catalyzes the dissociation of the colorless substrate into a short peptide segment and a yellow organic compound (e.g., p-nitroaniline) that can be photometrically quantified.The speed of color development, measured by the time needed to attain a pre-specified optical density (i.e., onset time), is proportional to the concentration of endotoxin or (1→3)-β-D-glucan.
A Chromo-LAL endotoxin assay (Associates of Cape Cod Inc., East Falmouth, MA, USA) was run in duplicate on an incubating microplate reader (ELx808IU, BioTek Instrument Inc., Winooski, VT, USA) at 37 °C and absorbance wavelength (λ) of 405 nm, with an onset optical density (OD, i.e., absorbance) of 0.1 selected.A control endotoxin standard (Escherichia coli O113:H10; potency: 1 ng = 24 EU; Associates of Cape Cod Inc., East Falmouth, MA, USA) was diluted in series to 50, 5, 0.5, 0.05, and 0.005 EU mL -1 , three times each to establish a standard calibration curve.To eliminate the interference of (1→3)-β-D-glucan, the Chromo-LAL reagent was reconstituted with the Glucashield glucan-blocking buffer solution (Associates of Cape Cod Inc., East Falmouth, MA, USA).The possible inhibition or enhancement was tested by spiking a test sample with 10 µL of 5 EU mL -1 endotoxin standard, and a further dilution was conducted when the percent recovery of the spiked sample was > 200% or <50%.A sample was re-analyzed when the coefficient of variation (CV) between duplicates exceeded ±10%.
Glucatell (1→3)-β-D-glucan assay (Associates of Cape Cod Inc., East Falmouth, MA, USA) was run in duplicate on the incubating microplate reader at 37 °C and λ=405 nm, with an onset OD of 0.03 selected.A standard calibration curve was developed by diluting the stock standard (included in the Glucatell assay) to 100, 50, 25, 12.5, 6.25 and 3.125 pg mL -1 three times each.A similar quality assurance procedure, as that adopted for endotoxin analysis, was applied.
Both assays require the analyte concentration in filter extracts within the concentration range of calibration standards.Accordingly, the assays' minimum detection limits (MDLs) were calculated based on the lowest-concentration standard to be 0.046 EU m -3 and 0.029 ng m -3 for endotoxin and (1→3)-β-D-glucan, respectively.The precision of the assays was calculated by running replicates (>3 per microplate), following the method of Watson et al. (2001).The accuracy of the assays was primarily limited by the uncertainty in prepared calibration standards.Due to lack of a quality control standard, this uncertainty could not be reliably assessed.Thus, no estimation of the assays' accuracy was conducted.
Glassware and metal tools (e.g., tweezers and filter punchers) were baked at 250 °C for >4 hr prior to the experiment, and they were sterilized repeatedly on a micro-incinerator during the experiment to prevent cross-contamination among samples.Pipette tips, microplates and centrifuge tubes were certified by the suppliers to be pyrogen-free and were tested in the laboratory by running negative control samples.Laboratory and field blank samples were also tested, and they, as well as the negative control samples, contained endotoxin and (1→3)-β-Dglucan levels below the MDLs.Thus, no blank subtraction was conducted and all values reported were as-measured.
For both compounds, the MDL was calculated as three times the standard deviation of the lowest-concentration standard (0.05 µg mL -1 ).The test of field blank samples showed that their arabitol and mannitol concentrations were below the MDLs (19 ng m -3 for arabitol and 18 ng m -3 for mannitol).Thus, all values reported were as-measured and were not adjusted for field blanks.The precision of the analyses was calculated by running replicates (>3 per batch), following the method of Watson et al. (2001).It was estimated to be <±3.2% for arabitol and <±2.6% for mannitol.The accuracy of the analyses (<±1.5%) was assessed by differences between measured and actual concentrations of the quality control standard.
Similar to endotoxin and (1→3)-β-D-glucan analyses, glassware and metal tools were baked at 250 °C for >4 hr prior to the experiment.Pipette tips and vials were tested by running negative control samples and no arabitol or mannitol was detected.
Comparison of fungal spore and pollen grain concentrations in PM samples collected at different locations and seasons.

Fig. 3 .
Fig. 3. Example images of ambient bioaerosols detected from PM 10 Teflon-membrane filter samples collected in Corcoran, California, by scanning electron microscopy (SEM; see magnification on individual Figures) for: A -fungal spores (Type 1: prolate spheroid), B -fungal spores (Type 2: spherical), C -pollen grains, D -plant detritus, E -bacteria (Streptococcus), F -a diatom, G -a slime mold, and H -an insect egg.Pores on the pollen surface were caused by high-energy electron beams during SEM analysis (The opaque background is a result of Teflon-membrane filters; high magnification [1000-5000x] was used to identify specific particle shapes and other characteristics).

Table 5 .Fungala
Correlation coefficients (r) among bioaerosol indicators and major PM 10 components for all sites.r > 0.5 are in bold to indicate which components might be predicted to some extent by measuring another component.Contributions a of fungal spores, pollen grains, and plant detritus to PM 10 mass and organic carbon (OC).Fungal spores, pollen grains, and plant detritus were assumed to have a density of 1 g/cm 3

Table 2 .
Comparison of arabitol and mannitol concentrations in PM samples collected at different locations and seasons.

Table 4 .
Characteristics of fungal spores, pollen grains and plant detritus acquired from the PM 10 Teflon-membrane filter samples.

Table 7 .
Contribution of fungal spores to organic carbon in PM samples collected at different locations and seasons.