Available online at www.sciencedirect.

com

The Veterinary Journal

The Veterinary Journal 179 (2009) 219224 www.elsevier.com/locate/tvjl

Airborne transmission of foot-and-mouth disease in pigs: Evaluation and optimisation of instrumentation and techniques
Claudia M.F. Amaral Doel
a

a,*,1 ,

John Gloster

a,b,1

, Jean-Francois Valarcher

Institute for Animal Health, Pirbright Laboratory, Woking, Surrey GU24 0NF, UK b Met Oce, FitzRoy Road, Exeter EX1 3PB, UK c IVI Animal Health, Larkbacken, 740 20 Vange, Uppsala, Sweden Accepted 6 September 2007

Abstract Foot-and-mouth disease (FMD) can be transmitted in a variety of ways, one of which is through virus exhaled into the air by infected livestock. It is clear that where there is close contact there will be a range of possible mechanisms for the transmission of disease from animal to animal, including the airborne route if simple barriers between livestock exist. In transmission of FMD over longer distances, airborne transmission represents a signicant challenge to the veterinary services in that the mechanism is essentially uncontrollable if the primary source of the disease is not contained. In the event of an epidemic of FMD, such as the one experienced in the United Kingdom in 2001, it is important for disease control purposes to understand the contribution made to the overall spread of disease by aerosolised virus. This assessment is based on a combination of measurements made in the laboratory and through clinical observations in the eld. To date, laboratory measurements have used a number of instruments that were not specically designed for working with FMD virus or whose performance have not been fully compared and documented. This paper compares four samplers and describes the method by which samples are processed. Overall it is concluded that there is no optimum air sampling instrument which could be successfully employed for all situations but the work provides guidance to those wishing to make measurements in the future and establishes a baseline against which any new samplers can be compared. 2007 Elsevier Ltd. All rights reserved.
Keywords: Foot-and-mouth disease; Airborne; Air samplers; Spread

Introduction Pigs, cattle, sheep and other ruminants infected with foot-and-mouth disease virus (FMDV) emit aerosol particles into the atmosphere. These can then be inhaled or ingested by other animals which subsequently become infected. A comprehensive review of the pathogenesis of the disease is given by Alexandersen et al. (2003). Although airborne infection of FMD was rst hypothesised in the early 1900s (Penberthy, 1901; Bang, 1912) and there have been numerous results from laboratory experiments, no instrument capable of specically measuring FMDV
*

Corresponding author. Tel.: +44 1483 232 441; fax: +44 1483 232 448. E-mail address: claudia.doel@bbsrc.ac.uk (C.M.F. Amaral Doel). Joint rst authors.

has ever been specically designed and made generally available. Measurements have been taken with a range of instruments, mainly developed in the 1960s, and the samples analysed using a number of laboratory techniques. The samplers used have included the Cyclone, Litton, Porton, May and the SKC BioSampler (Errington and Powell, 1969; Donaldson et al., 1982; May and Harper, 1957; May, 1966). Whilst each instrument and sample analysis technique has been described in the literature, no comprehensive intercomparison has been made, nor have the instruments been optimised for FMDV sampling and the results documented. This paper addresses these shortcomings by comparing the Cyclone, Porton, May and SKC BioSampler under controlled and realistic laboratory conditions. The objective was to oer guidance to others

1090-0233/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.tvjl.2007.09.010

220

C.M.F. Amaral Doel et al. / The Veterinary Journal 179 (2009) 219224 allow for several samples to be taken and to ensure minimum discomfort to the pigs. The May sampler was operated at 55 L/min for 5, 10 or 20 min in the loosebox and for 5 min in the cabinet. The Cyclone sampler was operated for 5, 10 or 20 min at ow rates of 390, 570 and 750 L/min; due to the high airow it was only possible to sample air from the loosebox using the Cyclone. The sample collecting uid was MEM-HEPES with added antibiotics and BSA (Alexandersen and Donaldson, 2002), at pH 7.2. All samples were stored at 70 C before being assayed for both virus infectivity and viral RNA genome.

involved in sampling from FMD infected animals and to provide a baseline by which new instruments can be compared. For mathematical modelling purposes, our nal results took into account the duration of the air sampling period, the volume of sample collected and the ow rate of the air during sampling.
Materials and methods
Previous studies have shown that pigs, the most prolic of virus emitters, when infected with FMD can excrete virus in their breath for a period up to 45 days and maximum excretion occurs during the very early stages of acute disease (Alexandersen et al., 2003; Gloster et al., 2006). Consequently we decided to infect a group of pigs with FMDV, make a detailed clinical assessment of the animals on a daily basis and schedule aerosol measurements from 14 days post-infection (d.p.i.). Animal experimentation was carried out in accordance with the UK Animals (Scientic Procedures) Act 1986 and associated guidelines.

Virus isolation
The infectivity of the impinger uid collected from the air samples was assayed by inoculation of primary cultures of BTY cells (Snowdon, 1966). Tenfold dilution series of collecting uid samples were made and each dilution was inoculated onto three tubes. Titres were calculated by the Karber formula according to Lennette (1964). The concentration of virus per litre of air was determined by the end-point titration of virus in each collected sample multiplied by the volume of the collecting uid and divided by the ow rate of the sampler.

Animals
Six Large White cross Landrace pigs, 2030 kg in weight, were used in the experiment. They were housed in a biosecure isolation building and kept in a loosebox of 3.6 m 3.3 m 3.0 m for the duration of the experiment. Each pig was inoculated intradermally in the heel bulb of the left rear foot with approximately 105.0 TCID50 (BTY) in a dose volume of 0.5 mL (0 d.p.i.). The pigs were examined daily for clinical evidence of FMD and scored using a modied version of the clinical scoring framework published by Quan et al. (2004) with a range of score from 017. A numerical score is given when each animal is examined according to its temperature, the presence of clinical signs on each of the feet, snout, mouth, tongue, teats, and noting whether the pig is lame and/or there are any other characteristics (o food/water, behaviour changes, nasal discharge, secondary infection). Serum samples of all six pigs were taken on a daily basis, starting just before needle inoculation of virus (04 d.p.i.).

ELISA
An anigen ELISA (Hamblin et al., 1984) was used to conrm the specicity of the cytopathogenic eect observed in cell cultures inoculated with air samples. Serum samples were tested by an ELISA for antibodies to FMDV (Hamblin et al., 1986).

RNA extraction
QIAamp MinElute Virus Spin kit (Qiagen) was used for RNA extraction from air samples according to the manufactures instructions. Briey, RNA was extracted from 200 lL of the original samples and a nal volume of 18 lL was recovered at the end of the process. This material was used for RT-PCR. An automated extraction procedure was used for nucleic acid extraction of serum samples (Reid et al., 2003).

Virus
FMDV strain O UKG 34/01 was isolated from a pig at Cheale Meats Abattoir, Brentwood, Essex, UK, on the 20th February 2001. This was passaged once in pigs and the stock virus produced from foot lesion epithelium had a nal titre of 107.0 TCID50/mL in primary cultures of bovine thyroid (BTY) cells (Snowdon, 1966). Tissue culture infectious dose (TCID50) is the dose of agent (in this case FMDV) that will infect a susceptible cell in 50% of laboratory experiments.

Quantitative real-time reverse transcription PCR (RT-PCR)

These were performed as previously described, including the primers and probe used in Experiment C of Quan et al. (2004). The PCR was performed on a Strategene MxPro 3005P QPCR System. Results were analysed by means of MxPro QPCR software and genome copy numbers per reaction were also calculated according to Quan et al. (2004). The concentration of genome copy numbers per litre of air was determined by the quantity of virus genome per millilitre in each collected sample multiplied by the volume of the collecting uid and divided by the ow rate of the sampler.

Air sampling intercomparison

Air samples were taken with a Cyclone, Porton, May and the SKC BioSampler; a summary of the attributes of these is given at Table 1. Sampling times and ow rates, selected on past experience, were varied in order to determine the optimum sampling conguration. Air samples were taken daily from 14 d.p.i. in the loosebox and also on days 2, 3 and 4 p.i. from two pigs, randomly selected on each sampling day and placed in a 610-L cabinet (Gibson and Donaldson, 1986); the cabinet was thoroughly disinfected with FAM (Evans Vanodine International) between each animal sampling. Daily variations in virus production were anticipated so that instruments were compared sequentially, within minutes of each other. In the loosebox, the Porton and BioSampler were operated at 11 L/min for 5 and 10 min. In addition a 20 min sample using the BioSampler was also taken; a 20 min Porton sample was not possible due to the loss of sampling medium created by the high level of evaporation from the sampling chamber. In the cabinet, sampling time was restricted to 5 min to

Results Evidence of infection Typical FMD signs were registered in all pigs approximately 12 days after challenge. Generalised lesions were observed in all animals and, for humane reasons, pig number VL 38 was euthanised after sampling on 2 d.p.i., followed by VL 33, VL 34 and VL 35 on 3 d.p.i. The remaining pigs, VL 36 and VL 37, were kept until the end of the experiment at 4 d.p.i. Serum samples analysed by quantitative real-time RTPCR indicated that viraemia initiated at 1 d.p.i. (Fig. 1),

C.M.F. Amaral Doel et al. / The Veterinary Journal 179 (2009) 219224 Table 1 Characteristics of the air sampling instruments used in this study Porton Principle of operation Particle sizing Liquid impinger BioSampler Liquid impinger May Three stage impactor Cyclone Particles impacted on side of a glass chamber and washed by impinger uid

221

<18 lm

<18 lm

1st stage >6 lm 2nd stage 36 lm 3rd stage 30.8 lm 33 or 55

Not quantied for sampler used

Operational ow rate (L/min) Total ow (L) Sample volume for analysis (mL) Technique of FMDV detection Evaluation in the eld Strengths

11 or 12

11 or 12

390, 570 or 750

55, 60, 110, 120, 220 or 240 5

55, 60, 110, 120, 220 or 240 20

165, 275, 330, 550, 660, 1100 10 for each stage; total of 30 Cell culture and PCR

1950, 2850, 3750, 5700, 11400 Variable; impinger delivered at a rate of 1 mL/min Cell culture and PCR

Cell culture and PCR

Cell culture and PCR

Not for FMD Easy to use, clean and instrument highly reliable Simple device

Not for FMD Easy to use and clean

Not for FMD Mimics respiratory breathing process Provides data on size of particles containing virus

Not for FMD Very large volumes of air sampled

Reliability inferior to the Porton Re-aerosolisation of particles minimized due to introduction of three tangential nozzles

VL 33 VL 34 VL 35
12 10
Level s of Viraemia
log10 number of copies of virus/mL

VL 36 VL 37 VL 38

log10 number of copies of virus/mL

8 6 4 2 0

12 10 8 6 4 2 0 D0 D1 D2 D3 D4 Days post-inoculation VL 33 VL 34 VL 35 VL 36 VL 37 VL 38

In summary the results suggested that infection followed an expected pattern and that these conditions provided an excellent basis for the instrument intercomparison. Based on this evidence and past experience, FMDV aerosol was expected to take place from 14 d.p.i. Due to the severity of disease and the need to adhere to strict animal welfare criteria the usefulness of some of the samples from the air cabinet was a little reduced; some animals had to be euthanised, for humane reasons, before the end of the experiment. Air sampling intercomparison

D0

D1

D2

D3

D4

Days post-inoculation
Fig. 1. Viraemia curves of pigs at days 0 (D0), 1(D1), 2 (D2), 3 (D3) and 4 (4) post-inoculation. Animal identication numbers: VL 33, VL 34, VL 35, VL 36, VL 37 and VL 38.

reached peak levels at 2 d.p.i. and started to decline at around 3 d.p.i. No antibodies were detected in any of the pigs throughout the sampling period. It is known that antibodies against FMD are detectable in serum shortly after the cessation of viraemia and correlate well with the end of viral excretion (Sellers et al., 1977; Gibson and Donaldson, 1986).

May and BioSampler Table 2 shows the results obtained when, sampling from the loose box, the May and the BioSampler were compared. The May sampler was the only one to detect infectious particles on 1 d.p.i., whereas both samplers detected infectious particles on 2 d.p.i. All samples collected on 3 and 4 d.p.i. were negative for infectivity. In relation to TCID50 per litre of air, the May detected the highest infectious doses during a 5 min sampling regime on 1 d.p.i. and, on 2 d.p.i., sampling for 10 min gave a marginally higher number of TCID50 per litre of air than sampling for 5 or 20 min. When analysing the copy numbers of FMDV RNA extracted from the samples, our results indicated that the

222

C.M.F. Amaral Doel et al. / The Veterinary Journal 179 (2009) 219224

Table 2 Comparison between the May and the BioSampler Duration of sampling (min) Equipment Total number of litres of air Flow rate (L/min) D1 p.i. TCID50/ La 1.09 0 0.12 0 0.35 0 Copies/ Lb 2841.66 7249.35 145.89 1001.70 681.72 1031.83 D2 p.i. TCID50/ La 0.11 0.99 0.12 0 0.03 0 Copies/ Lb 3129.30 3407.65 1624.84 2938.03 353.25 671.25 D3 p.i. TCID50/ La 0 0 0 0 0 0 Copies/ Lb 0 0 0 0 251.95 0 D4 p.i. TCID50/ La 0 0 0 0 0 0 Copies/ Lb 0 0 92.13 919.04 0 0

5 10 20

May BioSampler May BioSampler May BioSampler

275 55 550 110 1100 240

55 11 55 11 55 11

Samples were taken in the loosebox during 5, 10 and 20 min. D: day, p.i.: post-infection. a TCID50/L: as isolated in BTY cells. b Copies/L: number of genome copies per litre of sample.

best sampler was the BioSampler on 1 d.p.i. when sampling for 5 min. As observed with infectivity, the 5 min sampling time generated a higher number of copies per litre of air on the majority of the collection days than all other sampling times examined here. On 3 d.p.i., the sample collected during 20 min gave the single positive result for that day. However, on 4 d.p.i., it was the 10 min sampling time which produced the only two positive samples for this day when collected by both the May and the BioSampler. Peak levels were recorded at 1 d.p.i. with samples collected with the BioSampler. May and Porton When the performances of the May and the Porton samplers were compared for sampling in the loose box, results suggested that the highest infectivity was recovered on 2 d.p.i. (Table 3). On 1 d.p.i. both samples collected by the May sampler gave positive results but the samples collected by the Porton sampler were both negative for infectivity. All samples from 2 d.p.i. were positive for infectivity whereas all samples collected during 3 and 4 d.p.i. with both the May and the Porton were negative for infectivity. The optimum sampling time seemed to be 5 min for both samplers. Detection of FMD viral RNA correlated with detection of infectivity for samples collected during 1 and 2 d.p.i. On 3 d.p.i., sampling for 5 min with both samplers generated
Table 3 Comparison between the May and the Porton samplers Duration of sampling (min) Equipment Total number of litres of air Flow rate (L/min) D1 p.i. TCID50/ La 0.35 0 0.05 0

the only two positive samples for viral RNA. However, on 4 d.p.i. only a single positive sample was observed; this was collected with the Porton sampler for a period of 10 min. Results obtained with samples collected from the sampling cabinet on 2, 3 and 4 d.p.i. were broadly similar to those obtained with the loose box and indicated that the May gave the highest TCID50 when its performance was compared with the BioSampler and the Porton (results not shown). Cyclone Table 4 summarises results obtained when the performance of the Cyclone sampler was evaluated by applying dierent ow rates per minute while collecting samples. The results show that, during 1 d.p.i., samples collected for 5 min at a ow rate of 570 (total of 2850 L) recovered the highest TCID50 as well as the highest copy numbers of genome per litre of sample than the samples collected for 10 or 20 min at this ow rate. No other ow rates were tested during 1 d.p.i. On 2 d.p.i., signicantly higher numbers of infectious particles and virus genome were observed from samples where a ow rate of 390 L/min was applied for 5 min. For 3 d.p.i., recovery of infectious particles was marginally greater at 3750 or 5700 L, where identical results were observed. However, in relation to virus genome, it was once again the 1950 ow rate that generated

D2 p.i. Copies/ Lb 779.36 0 293.81 0 TCID50/ La 15.51 9.51 4 7.49 Copies/ Lb 27426.32 8069.87 11,690 2815

D3 p.i. TCID50/ La 0 0 0 0 Copies/ Lb 1873.02 2,06,518 0 0

D4 p.i. TCID50/ La 0 0 0 0 Copies/ La 0 0 0 454.25

5 10

May Porton May Porton

275 55 550 110

55 11 55 11

Samples were taken in the loosebox during 5 and 10 min. D: day, p.i.: post-infection. a TCID50/L: as isolated in BTY cells. b Copies/L: number of genome copies per litre of sample.

C.M.F. Amaral Doel et al. / The Veterinary Journal 179 (2009) 219224 Table 4 Comparison amongst samples taken with the Cyclone sampler, in the loosebox, during 5, 10 or 20 min Total number of litres of air 5 min 1950 (390 L/min) TCID50/ La DPC1 DPC2 DPC3 DPC4 NC 0.45 0.08 0 Copies/ Lb NC 910.7 308.91 0 2850 (570 L/min) TCID50/ La 0.26 0.14 0.01 0 Copies/ Lb 437.66 303.55 59.94 0 3750 (750 L/min) TCID50/ La NC 0.14 0.04 0 Copies/ Lb NC 62.24 121.78 0 10 min 5700 (570 L/min) TCID50/ La 0.2 0.17 0.04 0 Copies/ Lb 404.71 728.18 63.97 0 20 min

223

11400 (570 L/min) TCID50/ La 0.09 0.07 0.02 0 Copies/ Lb 422.91 240.02 113.43 0

Flow rates employed were 390, 570 or 750 L/min. DPC: days post-challenge, NC: sample not collected. a TCID50/L: as isolated in BTY cells. b Copies/L: number of genome copies per litre of sample.

the greatest number of copies per litre. All samples collected on 4 d.p.i. were negative for both infectivity and viral RNA. Discussion In the event of an outbreak of disease, an accurate assessment of the risk of secondary infection, including airborne transmission, is required by those responsible for controlling and eradicating the disease. This assessment is aided by a detailed knowledge of how much virus is likely to have been produced by the infected livestock and this is underpinned by the quality of measurements made under laboratory conditions. Consequently, it is important to know which instruments are the best at detecting the virus, and their limitations. This is supported by the requirement to have optimised and documented standardised laboratory analysis techniques. In addition as new instrumentation becomes available a baseline against which they can be compared is believed to be essential. Based on the results from this study it is concluded that in general there is no optimum air sampling technique which could be successfully employed for all situations but that Cyclone type samplers are, perhaps the most ecient. This view is supported by Cox (1987) and Mouilleseaux (1990) who used the same type of instrument to measure other aerosols. All samples collected by the Cyclone sampler on 1, 2 and, in particular 3 d.p.i. were positive for infectivity suggesting that this is far the most ecient sampler in recovering infectious particles than any other sampler studied here. With regard to sample time, there appeared to be little merit in taking samples for longer than 5 min. However, whilst in an enclosed room this may be optimal, it must be remembered that under other circumstances, e.g. in a larger unit or even out of doors it may be advantageous to sample for much longer, thus increasing the possibility of detecting low concentrations of virus. On the other hand, there is a possibility that extended sampling may cause a decrease, or even loss, of infectivity and/or copy numbers of virus genome. The May sampler was the best in relation to recovery of infectious FMDV particles when compared with both the

Porton and the BioSampler. The May sampler has an additional characteristic which has not been evaluated during this experiment; it allows the aerosol sample to be divided into three size groups (>6 lm, 36 lm and <3 lm). Whilst this is an additional strength for laboratory measurements, it is also its weakness. Each size group is collected in a different chamber and, as a result, the May is much more difcult to clean between measurements. In addition, because it is a precision piece of glassware, it is complex and may not be suited to large scale commercial manufacture. However, for laboratory type use, the intercomparison has shown that it is a very good instrument. When viral release reached its peak on 2 d.p.i. the results indicated that the Porton sampler seemed to be a good choice of instrument to use (see Table 3). Samples collected by both the May and the Porton samplers gave very similar results and it is accepted that the Porton was far the easiest sampler to use and also to clean after each sampling. As with the Cyclone sampler, there appeared to be little merit in operating the Porton sampler for longer than 5 min under the specic circumstances used in the comparison. If the problem concerning evaporation could be reduced, possibly by using a more viscous medium, larger air sample volumes could be taken. When compared with the May, the BioSampler seemed to be less reliable for recovery of infectious virus particles; a single sample was positive for infectivity. However, it was noted that the BioSampler gave the highest number of copies of viral genome. Conclusions Based on this present study, the authors believe that the use of a combination of samplers would be not only ideal but also advisable when air sampling. The best results were not always obtained with the same sampler in one given set of conditions; this may have been due to natural uctuations in virus production and release from the pigs as the sampling procedure and laboratory analysis remained constant. Ultimately it would be a useful step forward if FMDV aerosol measurements could be made in the eld. However

224

C.M.F. Amaral Doel et al. / The Veterinary Journal 179 (2009) 219224 Gibson, C.F., Donaldson, A.I., 1986. Exposure of sheep to natural aerosols of foot-and-mouth disease virus. Research in Veterinary Science 41, 4149. Gloster, J., Williams, P., Doel, C., Esteves, I., Coe, H., Valarcher, J.F., 2006. Foot-and-mouth disease Quantication and size distribution of airborne particles emitted by healthy and infected pigs. The Veterinary Journal 174 (1), 4253. Hamblin, C., Armstrong, R.M., Hedger, R.S., 1984. A rapid enzymelinked immunosorbent assay (ELISA) for the detection of foot-andmouth disease virus in epithelial tissues. Veterinary Microbiology 9, 435443. Hamblin, C., Barnett, I.T., Hedger, R.S., 1986. A new enzyme-linked immunosorbent assay (ELISA) for the detection of antibodies against foot-and-mouth disease virus, I. Development and method of ELISA. Journal of Immunological Methods 93, 115121. Lennette, E.H., 1964. General principles for laboratory diagnosis of viral and rickettsial infections, in Diagnostic Procedures for Viral and Rickettsial Diseases. American Public Health Association, New York, pp. 4851. May, K.R., 1966. Multistage Liquid Impinger. Bacteriological Reviews 30, 559570. May, K.R., Harper, G.J., 1957. The eciency of various liquid impinger samplers in bacterial aerosols. British Journal Industrial Medicine 14, 287297. Mouilleseaux, A., 1990. Sampling methods for bioaerosols. Aerobiologia 6, 3235. Penberthy, J., 1901. Foot-and-mouth disease. Journal Comparative Pathology Therapeutics 24, 1629. Quan, M., Murphy, C.M., Zhang, Z., Alexandersen, S., 2004. Determinants of early foot-and-mouth disease virus dynamics in pigs. Journal Comparative Pathology 131, 294307. Reid, S.M., Grierson, S.S., Ferris, N.P., Hutchings, G.H., Alexandersen, S., 2003. Evaluation of automated RT-PCR to accelerate the laboratory diagnosis of foot-and-mouth disease virus. Journal of Virological Methods 107, 129139. Sellers, R.F., Herniman, K.A.J., Gumm, I.D., 1977. The airborne dispersal of foot-and-mouth disease virus from vaccinated and recovered pigs, cattle and sheep after exposure to infection. Research. Veterinary Science 23, 7075. Snowdon, W.A., 1966. Growth of foot-and-mouth disease virus in monolayer of cultures of calf thyroid cells. Nature 210, 10791080.

this will involve a number of signicant additional challenges to those experienced above. For example, relative humidity and temperature, which can inuence aerosol survival are rigorously controlled in our animal isolation units but are not so easily controlled in the eld (Donaldson, 1972). The development of a disposable sampler would also be an advantage to make eld measurements a possibility in the future. Acknowledgements The authors would like to thank the sta responsible for the animal isolation unit 8 at IAH for their careful management. We are grateful to David Paton for his constructive comments. This work was funded by the Department for Environment, Food and Rural Aairs (DEFRA), project SE2926. References
Alexandersen, S., Donaldson, A.I., 2002. Further studies to quantify the dose of natural aerosols of foot-and mouth disease virus for pigs. Epidemiology and Infection 128, 313323. Alexandersen, S., Zhang, Z., Donaldson, A.I., Garland, A.J., 2003. The pathogenesis and diagnosis of foot-and-mouth disease. Journal of Comparative Pathology 129, 136. Bang, B., 1912. Foot-and-mouth disease. Journal of Comparative Pathology 25, 115. Cox, C.S., 1987. The Aerobiological Pathway of Micro-organisms. John Wiley & Sons, Chichester. Donaldson, A.I., 1972. The inuence of relative humidity on the aerosol stability of dierent strains of foot-and-mouth disease virus suspended in saliva. Journal of General Virology 15, 2533. Donaldson, A.I., Ferris, N.P., Gloster, J., 1982. Air sampling of pigs infected with foot-and-mouth disease virus: comparison of Litton and Cyclone samplers. Research in Veterinary Science 33, 384385. Errington, F.P., Powell, E.O., 1969. Journal of Hygiene 67, 387399.