Sie sind auf Seite 1von 8

Inhibition of Escherichia coli in cultivated cattle manure1

Z. G. Weinberg,*2 G. Szakacs, Y. Chen,* R. Pinto,* S. Bernstein,* B. Konya, and S. Sela (Saldinger)*

*Department of Food Quality and Safety, Agricultural Research Organization,
The Volcani Center, Bet Dagan 50250, Israel. Department of Applied Biotechnology
and Food Science, Budapest University of Technology and Economics, Gellert ter 4, 1111 Budapest, Hungary

ABSTRACT: A common practice on Israeli dairy barns successive days to eliminate from the manure antagonistic
comprises daily cultivation of the manure. Cultivation is microorganisms. Each cultivationsterilization treatment
a mechanical process used to break up and till the manure was performed in triplicate jars. Following sterilization, E.
bedding and it results in a drier and aerated bedding and coli numbers in the cultivated and noncultivated manure
cleaner cows, which consequently reduces the incidence of were comparable, while in the nonsterilized manure the
mastitis. Cultivation was associated with a shorter survival numbers were lower in the cultivated compared with the
of Escherichia coli in cultivated manure as compared with noncultivated manure. Several fungi isolated from the cul-
noncultivated manure. The objective of the current study tivated manure samples displayed inhibition effect on the
was to elucidate the mechanism responsible for the shorter tagged E. coli. Antagonistic fungi were also isolated from
survival duration of E. coli in the cultivated manure. We large-scale cultivated manure samples collected on several
hypothesized that microorganisms that are antagonistic to dairy farms in Israel. These findings support the notion
E. coli, developing in the cultivated manure, are respon- that manure cultivation might facilitate the development of
sible for this phenomenon. A cow manure derived E. coli microorganisms that are antagonistic to E. coli, thus con-
strain expressing the green fluorescence protein and antibi- tributing to the general hygiene of the cattle. Identifying
otic resistance markers was used to inoculate cow manure the mechanisms by which the antagonistic fungi affect
in 1.5-L jars. Manure treatments included cultivated and the survival of E. coli in manure could be exploited for
noncultivated manure. Half the jars of each cultivation improvement of the animal health and for limiting the
treatment were autoclave sterilized at 121C for 1 h on 3 transmission of zoonotic pathogens to food and water.

Key words: cattle manure, cultivation, Escherichia coli, fungi

2014 American Society of Animal Science. All rights reserved. J. Anim. Sci. 2014.92:23362341

INTRODUCTION feeding alley to a storage basin and used along with the
drier manure for composting. According to a new man-
Cattle may be a source of pathogens that might agement practice, the wet slurry is added daily to the
infect cattle and contaminate the food supply. A sig- solid manure at the barn followed by daily cultivation to
nificant number of recent foodborne disease outbreaks a depth of 50 to 60 cm (Friedman, 2008). Cultivation is
were attributed to Escherichia coli O157:H7 and implemented by a duck feet cultivator (Fig. 1) drawn
Salmonella spp., which are commonly found in bovine by a tractor, the shovels of which break up, loosen, and
manure and slurry (Pell, 1997; Semenov et al., 2009). turn the bedding. Recently, the feeding alleys were re-
In Israel the animals in all dairy barns are loose placed by feeding tubs, which are placed in the center
with no stalls. The wet slurry is removed daily from the of the barn. In this way the slurry is constantly mixed
with the manure, which continues to be cultivated daily.
1Contribution from the Agricultural Research Organization, The
This management saves the need to shovel the slurry
Volcani Center, Bet Dagan, Israel, number 668/13-E, 2013 series. from the feeding alleys by a front load bucket tractor
This study was supported by the Israeli Milk Board. and pour it over the roofed bedding. It was suggested
2Corresponding author:
that daily cultivation of the manure results in overall
Received October 22, 2013. aerated and drier bedding, which favors the welfare of
Accepted February 23, 2014.
Downloaded from at ProQuest on June 9, 2014
Escherichia coli in cultivated cattle manure 2337

Figure 2. The garden cultivator used in the experiments. See online ver-
sion for figure in color.

sterilized, noncultivated, and noncultivatedsterilized.

There were triplicate jars per treatment. The efficiency
Figure 1. Duck-feet cultivator. Curtsy Yoram Kalgrad, Kibbutz Harduf,
of the sterilization process was tested by total aerobic
Israel. See online version for figure in color.
counts on plate count agar (Scharlau Microbiology,
Barcelona, Spain) and on actinomycete isolation agar
the animals and lowers the incidence of mastitis. This (Difco, Becton, Dickinson and Company, Sparks, MD).
notion was supported by Menis et al. (2011) who found All the jars were then inoculated with a cow manure de-
drier bedding, cleaner cows, lower incidence of masti- rived E. coli strain ZP1 tagged with GFP and antibiotic
tis, and lower numbers of various pathogenic bacteria resistant markers (Weinberg et al., 2011). Inoculation
in cultivated manure as compared with noncultivated was performed by emptying the contents of each jar into
manure on the same dairy farm of kibbutz Netzer-Sireni, a tub and adding 20 mL of the E. coli suspension fol-
Israel. This type of management has gained much popu- lowed by thorough mixing. The inoculated manure was
larity on dairy farms all over Israel. then returned to the jars.
Weinberg et al. (2011) used an E. coli strain that was The inoculation rate (cfu/g) was determined from
isolated from cow manure and tagged with green fluo- the number of the E. coli in the growth suspension in
rescence protein (GFP) and antibiotic resistance mark- Luria-Bertani broth. The number of the tagged E. coli in
ers to study the effect of cultivation on the fate of E. coli. each jar was determined on Days 0, 1, 2, 3, 4, 7, and 9
The cultivated manure had a higher DM content and after inoculation. A composite 4-g sample from each jar
contained lower numbers of the tagged E. coli, which was resuspended with 36 mL of sterile saline solution
could be detected for shorter duration time as compared in 50-mL sterile conical polypropylene tubes equipped
with the noncultivated manure. There might be several with a lid (Miniplast, Ein-Shemer, Israel). The bacteria
explanations for the observed effect on E. coli. The ob- were extracted by vigorous vortexing for 1 min followed
jective of the current study was to test the hypothesis by serial 10-fold dilutions in saline. Samples were plated
that microorganisms that are antagonistic to E. coli de- on Tryptic soy agar (Difco) supplemented with kanamy-
velop in the cultivated manure and limit its survival. cin and streptomycin at 50 and 100 g/mL, which en-
abled visualization of GFP fluorescence under UV light.
MATERIALS AND METHODS All plates were incubated at 39C for 24 h.
Experiment 2 (2013). A similar experiment was per-
formed that served as a repetition of Exp. 1. In this ex-
Experimental periment, 30 mL of the tagged E. coli suspension was
Experiment 1 (2012). Untreated dairy cow manure added to 750 g manure in each jar. In this experiment ad-
from the Volcani Center dairy research farm (Bet Dagan, ditional 2 treatments were added: cultivated manure from
Israel) was brought to the laboratory. Part of it was cul- the Harduf dairy farm and sterilization of this manure as
tivated daily for 1 mo with a small manual garden cul- described above. Each treatment comprised 3 jars.
tivator (Fig. 2) in a 40 by 30 cm plastic tub. Then, the In both experiments, no colonies were found on
noncultivated and cultivated manure were placed each plate count agar or on actinomycete isolation agar after
in six 1.5-L glass jars, 1 kg per jar, which were covered sterilization of the manure.
with aluminum foil. Three jars from each manure type
were autoclave sterilized 3 times at 121C for 1 h on 3 Additional Analyses
successive days. This sterilization process was applied
to eliminate from the manure also heat-resistant spores, The following analyses were performed on manure
such as those of Actinomycetes and Bacillus spp. Four from each of the triplicate jars at the end of Exp. 1 and
manure treatments were included: cultivated, cultivated 2. Dry matter content of the manure was determined by

Downloaded from at ProQuest on June 9, 2014

2338 Weinberg et al.

oven drying at 105C for 24 h. The pH value was deter- meat extract, 3.0 g/L yeast extract, 4.0 g/L peptone from
mined with a pH meter (Mettler-Toledo, Schwerzenbach, casein, 6.0 g/L peptone from meat, 1.0 g/L glucose, and
Switzerland). Total volatile base nitrogen (TVBN) was 18.0 g/L agar, pH 6.5 (before sterilization); and medium
determined only in Exp. 2 on Day 3. Total volatile base ATCC-5 (sporulation agar), 1.0 g/L yeast extract, 1.0
nitrogen was determined by the Kjeldahl method with- g/L beef extract, 2.0 g/L tryptose, 0.05 g/L FeSO4, 10.0
out the digestion step, according to Anderson (2008). g/L glucose, and 19.0 g/L agar, pH 7.2 (before steriliza-
tion). The isolated actinomycetes or fungi were inocu-
Isolation of Microorganisms with Inhibitory Activity lated as streaks on GR-1 and ATCC-5 agar media in petri
against Escherichia coli by Isolated Cultivated Manure plates. After 2 d of incubation at 30C, they were cross-
streaked with overnight broth cultures of 7 test bacte-
Mesophilic actinomycetes and filamentous fungi ria (Fig. 5). These included Bacillus subtilis American
have been isolated from dried samples of the cultivated Type Culture Collection (ATCC) 7972, Bacillus subti-
cattle manure by conventional isolation methods in pe- lis ATCC 14593, E. coli ZP1 GFP tagged, Micrococcus
tri plates at 30C. These microorganisms were isolated luteus ATCC 10240, Pseudomonas fluorescens ATCC
both from cultivated manure of laboratory Exp. 2 and 13525, Pseudomonas fluorescens Loc (local isolate in
from cultivated manure of 3 commercial dairy farms in Hungary), and Micrococcus luteus (Sarcina subflava)
Israel. For isolation of actinomycetes, medium ATCC- ATCC 7468. Bacillus subtilis and M. luteus strains are
5 (sporulation agar) and Bacto actinomycete isolation gram positive while E. coli and P. fluorescens are gram
agar (Difco Manual, 10th ed., 1984, p. 66, Olson, per- negative. Except for the manure derived E. coli strain,
sonal communication) were used. Fungi were isolated the strains were originated from ATCC (American Type
on potato dextrose agar (PDA) medium supplemented Culture Collection, Manassas, VA) or Loc (local isolate
after sterilization with 100 g/mL doxycycline hyclate in Hungary). The plates were then further incubated at
(Sigma-Aldrich, St. Louis, MO) to suppress bacterial 37C for additional 1 d. Growth inhibition was observed
growth. Primary fungal colonies were further purified as a clearing zone near the central streaks of actinomy-
by streaking on PDA medium supplemented with 0.5 g/L cetes or fungal isolates.
Triton X-100 (colony growth restricting agent; Sigma-
Aldrich) to isolate single colonies. Pure actinomycetes Statistical Analysis
and fungal cultures were freeze-dried for long-term
storage and deposited at the Technical University of The experiments were performed in a randomized
Budapest culture collection ( or block design to determine significant differences be- tween the applied treatments within a sampling date of
From manure samples cultivated in Exp. 2, alto- each experiment. Such design is also consistent with our
gether 16 different actinomycetes and 12 fungal strains previous work (Weinberg et al., 2011). Therefore, 1-way
have been isolated. Additional strains were obtained and ANOVA was applied to the results of each experiment
freeze-dried from the large-scale cultivated cattle ma- and day of sampling separately in the General Linear
nure materials collected on the Harduf (Lower Galilee, Model of SAS (SAS Inst. Inc., Cary, NC). All manure
northern Israel) and Netzer-Sireni (central coastal plain) treatments were included in the analysis. Significant
kibbutz dairy farms. From the Harduf farm samples 8 differences between means of the dependent variables
actinomycetes and 27 fungi were isolated. Sampling at (log10 cfu/g and pH, DM content, and TVBN) were
the Netzer-Sireni farm resulted in 7 actinomycetes and identified by using the Tukeys studentized range test in
16 fungal isolates. All these microorganisms were tested which P < 0.05 was declared as significant. In addition, a
in a later step against the tagged E. coli and other bacte- 2-way ANOVA was also performed to determine the sig-
ria for antagonistic activity. nificance of the effects of cultivation, sterilization, and
Inhibition Assay. Agar cross-streak method was their interactions, which were included in the statistical
used for testing the antimicrobial activity of the isolated model.
actinomycetes and fungi against E. coli and other bac-
teria . The isolates were streaked centrally on Grove- RESULTS
Randall number 1 (GR-1) and ATCC-5 sporulation
agar media in petri plates. These 2 media were selected Table 1 gives the results of the chemical analysis of
because production of antibiotics depends on the quan- the manure of the various treatments at the end of Exp.
tity of carbon source, for example, glucose. In medium 1 and 2. This analysis was performed to follow changes
GR-1 the glucose concentration is low while in medium in the manure incurred by both cultivation and steriliza-
ATCC-5 the concentration of sugar is high. The com- tion treatments, which could have affected the survival
position of these 2 media is as follows: GR-1, 1.5 g/L of the added E. coli strain. In Exp. 1, DM content varied

Downloaded from at ProQuest on June 9, 2014

Escherichia coli in cultivated cattle manure 2339

Table 1. Analytical results of the cattle manure at the treatment. The 2-way ANOVA revealed that on Days 7
end of Exp. 1 and 2 and 9 the noncultivated manure had significantly higher
Experiment Treatment1 pH DM, g/kg TVBN2 numbers of the tagged E. coli than the cultivated manure.
1 CULT 9.4 587ab The interactions cultivation sterilization were nonsig-
CULT-STE 9.3 575b nificant at all sampling times.
NOCULT 9.4 639a Similar results were obtained in Exp. 2 (Fig. 4). On
NOCULT-STE 9.4 637ab Days 1 through 4 the noncultivated manure tended to
SEM 0.07 19.5
have higher E. coli counts than the cultivated manure;
2 CULT 9.0ab 838a 13.9a
on Days 1 through 7 the sterilized manure had higher
CULT-STE 8.6b 758a 15.7a
counts than the nonsterilized manure. The interactions
NOCULT 9.1a 497b 3.0b
NOCULT-STE 8.7ab 487b 14.3a
cultivation sterilization were nonsignificant at all sam-
HARDUF 9.2a 594b 1.5b pling times. Also the cultivatedsterilized manure from
HARDUF-STE 8.9ab 564b 16.8a Harduf had significantly higher E. coli counts compared
SEM 0.17 5.5 3.14 with the nonsterilized manure.
a,bFor each experiment, means within the same column followed by differ- The majority of fungi isolated from the manure were
ent superscripts differ significantly (P < 0.05). Aspergillus spp. and of actinomycetes were Streptomyces
1CULT = cultivated; NOCULT = noncultivated; STE = sterilized; spp. All isolates (actinomycetes and fungi) were tested
HARDUF = cultivated manure from Kibbutz harduf. first against the tagged E. coli on GR-1 and ATCC-5
2TVBN = total volatile base nitrogen. media. Strains with anti-E. coli activity were retested
against the 7 bacterial species indicated above. The re-
between 575 and 639 g/kg and pH values were around sults show that none of the 31 actinomycetes isolates
9.4. In Exp. 2, the DM content varied between 487 (for (laboratory and farm isolates) had inhibitory activity
noncultivated manure) and 838 g/kg (for cultivated ma- against E. coli ZP1 strain. However, 10 out of 55 fungal
nure) and the pH values varied between 8.6 and 9.2. In isolates inhibited the growth of E. coli and they were
Exp. 2, cultivation tended to increase DM content and subsequently tested against 7 other bacterial species.
sterilization tended to decrease pH values. The effect of Four selected fungi with activity against the different
sterilization on TVBN values was inconsistent among bacteria are shown in Fig. 5.
cultivation treatments.
Figure 3 gives the log10 cfu per gram of the added DISCUSSION
E. coli in the various manure treatments of Exp. 1. The
counts on Day 0 agreed with the calculated inoculation Cattle manure and slurry are accumulated at large
rate mentioned above. The numbers of the tagged E. coli quantities in rural areas and may potentially contami-
tended to decrease with time, however, to a smaller ex- nate soil and groundwater. Pathogens that exist in the
tent in the noncultivated sterilized manure. From Day 1 manure bedding may cause mastitis, which results in re-
and on the sterilized manure contained significantly (P< duction in milk yield and quality (Barkema et al., 2009;
0.05) higher counts of the tagged E. coli as compared Burnevich et al., 2003; Vangroenweghe et al., 2005).
with the nonsterilized manure of the same cultivation

Figure 4. Change in green fluorescence protein Escherichia coli during in-

Figure 3. Change in green fluorescence protein Escherichia coli during in- cubation in cattle manure (Exp. 2). CULT = cultivated; NOCULT = not cultivated;
cubation in cattle manure (Exp. 1). Cult = cultivated; nocult = not cultivated; ste STE = sterilized; HARDUF = cultivated manure from Kibbutz harduf.. a,bWithin a
= sterilized. a,bWithin a day, columns accompanied by different letters are signifi- day, columns accompanied by different letters are significantly different (P< 0.05).
cantly different (P < 0.05). See online version for figure in color. See online version for figure in color.

Downloaded from at ProQuest on June 9, 2014

2340 Weinberg et al.

Figure 5. Inhibition of growth of 7 test bacteria by selected 4 antagonistic fungi from cultivated cattle manure on medium Grove-Randall number 1 and
medium ATCC-5. Test bacteria are a) Bacillus subtilis American Type Culture Collection (ATCC) 7972, b) Bacillus subtilis ATCC 14593, c) Escherichia coli
green fluorescence protein tagged, d) Micrococcus luteus ATCC 10240, e) Pseudomonas fluorescens ATCC 13525, f) Pseudomonas fluorescens Loc (local isolate
in Hungary); and g) Micrococcus luteus (= Sarcina subflava) ATCC 7468. Med. = medium. See online version for figure in color.

Escherichia coli is 1 of several etiological agents caus- for the survival of E. coli. In the current study we have
ing bovine mastitis (e.g., Bradley and Green, 2000; tested the hypothesis that the conditions in the culti-
Blum and Leithner, 2013). vated manure favor the development of microorganisms
The Harduf practice developed several years ago in that are antagonistic to E. coli. Such microorganisms
Israel comprises daily deep cultivation of the manure could be fungi and actinomycetes since both groups are
bedding along with the wet slurry. Klaas et al. (2010) well known for their capacity to produce antimicrobial
and Menis et al. (2011) showed that such management substances (Berdy, 2005; Demain and Sanchez, 2009;
results in drier manure, keeps the cows cleaner, decreas- Monaghan and Barrett, 2006; Pell, 1997). The idea was
es the microbial load, and decreases the incidence of that sterilization would eliminate from the manure all
mastitis. In a previous study (Weinberg et al., 2011) we microorganisms including those that inhibit E. coli, and
tested the effect of cultivation on the survival of a ma- therefore, the survival of the added E. coli would be
nure-derived E. coli strain in a model system. Cultivation comparable in both cultivated and noncultivated ma-
of cattle manure resulted in drier and aerated manure nure. Although sterilization resulted in some changes in
as compared with noncultivated manure. Concurrently, the manure composition (Table 1), in both Exp. 1 and
the viable numbers of the inoculated E. coli decreased 2 the numbers of the tagged E. coli were significantly
much faster in cultivated than in the noncultivated ma- higher in sterilized manure as compared with nonsteril-
nure. This result was attributed to the dryness of the cul- ized manure. This included also the sterilized cultivated
tivated manure, which could have been less favorable manure, in spite of its high DM content (Table 1). These

Downloaded from at ProQuest on June 9, 2014

Escherichia coli in cultivated cattle manure 2341
findings are concurrent with the hypothesis that the sur-
vival of E. coli in the manure is affected by the presence The autochthonous microbial flora in cattle manure
of antagonistic microorganisms. consists of microorganisms, which affect the survival of
The chemical analyses (Table 1) ware performed E. coli in the manure environment. It is possible that the
to follow possible changes in the manure incurred by shorter survival duration of E. coli in cultivated manure
both cultivation and sterilization, which could have also as compared with noncultivated manure is related to the
affected the survival of the inoculated E. coli strain. The establishment of microbial populations with anti-E. coli
relatively high TVBN value in the laboratory-cultivated activity. Further studies are required to support this con-
versus noncultivated manure might be related to increased clusion in commercial cow barns.
aerobic microbial activity in the former. It is possible that
sterilization has also resulted in nitrogen release in all Literature Cited
samples as a result to the heat treatment. The low TVBN Anderson, A. K. 2008. Biogenic and volatile amine-related qualities
concentration in the sample from Harduf may be due to of three popular fish species sold at Kuwait fish market. Food
unknown intrinsic factors. Further studies are needed to Chem. 107:761767.
negate possible nonbiological effects of the sterilization Barkema, H. W., J. M. Green, A. J. Bradley, and R. N. Zadoks. 2009.
process on the survival of E. coli. The E. coli counts were Invited review: The role of contagious disease in udder health. J.
Dairy Sci. 92:47174729.
similar or slightly higher in noncultivated manure than in Berdy, J. 2005. Bioactive microbial metabolites. A personal view. J.
cultivated manure, which is generally in agreement with Antibiot. 58:126.
our previous findings (Weinberg et al., 2011). Further Blum, S. E., and G. Leithner. 2013. Genotyping and virulence factors
research is required to confirm the effect of cultivation assessment of bovine mastitis Escherichia coli. Vet. Microbiol.
on mastitis-related strains of E. coli and other pathogens, 163:305312.
Bradley, A., and M. Green. 2000. A study of the incidence and sig-
such as Staphylococcus aureus and streptococci.
nificance of intramammary enterobacterial infections acquired
Two fungal strains isolated from the laboratory culti- during the dry period. J. Dairy Sci. 83:19571965.
vated manure exhibited activity against the tagged E. coli Burnevich, C., V. Van Merris, J. Mehrzad, A. Diez-Fraile, and L.
strain. Additional 8 fungal strains have been found in the Duchateau. 2003. Severity of E. coli mastitis is mainly deter-
large-scale (farm) cultivated samples. Interestingly, none mined by cow factors. Vet. Res. 34:521564.
of the actinomycetes isolated from the laboratory and farm Demain, A. L., and S. Sanchez. 2009. Microbial drug discovery: 80
years of progress. J. Antibiot. 62:516.
samples were active against the tagged E. coli. This was Friedman, S. 2008. The economical potential of cattle manure. (In
somewhat surprising because actinomycetes are known Hebrew.) Meshek Habakar Vehachalav 332:2224.
to produce antibacterial compounds against gram-nega- Klaas, I. C., B. Bjerg, S. Friedman, and D. Bar. 2010. Cultivated
tive bacteria. The reason might be that the GFP-tagged barns for dairy cows. Dansk Veterinaertidsskrift 93:2029.
E. coli carries genes for streptomycin and kanamycin re- Menis, D., A. Schwimmer, Z. Feldheim, Z. Weinberg, and S.
Friedman. 2011. The effect of manure daily deep cultivation in
sistance. Another reason might be that almost all these
cowsheds on udder health and milk quality. In: IDF World Dairy
isolated strains belong to the group of white-sporulating Summit, Parma, Italy, October 1519. (Poster).
actinomycetes, and therefore they represent only a nar- Monaghan, R. L., and J. F. Barrett. 2006. Antibacterial drug discov-
row spectrum of the actinomycetes group. ery Then, now and the genomics future. Biochem. Pharmacol.
Since both in the laboratory and farm-scale experi- 71:901909.
ments only some fungi proved to be antagonistic against Pell, A. N. 1997. Manure and microbes: Public and animal health
problem. J. Dairy Sci. 80:26732681.
E. coli, the DM content of the manure during cultivation Semenov, A. V., L. Van Overbeek, and A. H. V. Van Bruggen. 2009.
may be involved. It should be optimal for fungal growth Percolation and survival of Escherichia coli and Salmonella en-
in large-scale technology. Another important parameter terica serotype Typhimurium in soil amended with contaminated
is the occasional mixing of the material to provide suf- dairy manure or slurry. Appl. Environ. Microbiol. 75:32063215.
ficient amount of air for metabolism, while anaerobic or Vangroenweghe, F., L. Duchateau, P. Boutet, P. Lekeux, P. Rainard,
M. J. Paape, and C. Burnevich. 2005. Effect of Carpofen treat-
semi-anaerobic conditions should be avoided.
ment following experimentally induced Escherichia coli masti-
Further studies are needed to elucidate the role of tis in primiparous cows. J. Dairy Sci. 88:23612376.
antagonistic fungi in inhibiting E. coli and potentially Weinberg, Z., Y. Chen, P. Khanal, R. Pinto, and S. Sela. 2011. The
other bacterial pathogens in cultivated cattle manure. In effect of cattle manure cultivation on moisture content and sur-
case such a role is found, it would be possible to develop vival of Escherichia coli. J. Anim. Sci. 89:874881.
inoculants for cultivated cattle manure comprising such
fungi, which could inhibit E. coli and perhaps other det-
rimental bacteria at the barn.

Downloaded from at ProQuest on June 9, 2014

References This article cites 14 articles, 2 of which you can access for free at:

Downloaded from at ProQuest on June 9, 2014

Reproduced with permission of the copyright owner. Further reproduction prohibited without