Comparative surface-to-hand and fingertip-to-mouth transfer efficiency of gram-positive bacteria, gram-negative bacteria, and phage
* Correspondence to: Patricia Rusin, Department of Soil, Water and Environmental, Building 38, University of Arizona, Tucson, Arizona, USA (e-mail: firstname.lastname@example.org).
Aims: To determine the transfer efficiency of micro-organisms from fomites to hands and the subsequent transfer from the fingertip to the lip.
Methods and Results: Volunteers hands were sampled after the normal usage of fomites seeded with a pooled culture of a Gram-positive bacterium (Micrococcus luteus), a Gram-negative bacterium (Serratia rubidea) and phage PRD-1 (Period A). Activities included wringing out a dishcloth/sponge, turning on/off a kitchen faucet, cutting up a carrot, making hamburger patties, holding a phone receiver, and removing laundry from the washing machine. Transfer efficiencies were 38·47% to 65·80% and 27·59% to 40·03% for the phone receiver and faucet, respectively. Transfer efficiencies from porous fomites were <0·01%. In most cases, M.luteus was transferred most efficiently, followed by phage PRD-1 and S. rubidea. When the volunteers' fingertips were inoculated with the pooled organisms and held to the lip area (Period B), transfer rates of 40·99%, 33·97%, and 33·90% occurred with M. luteus, S. rubidea, and PRD-1, respectively.
Conclusions: The highest bacteral transfer rates from fomites to the hands were seen with the hard, non-porous surfaces. Even with low transfer rates, the numbers of bacteria transferred to the hands were still high (up to 106 cells). Transfer of bacteria from the fingertip to the lip is similar to that observed from hard surfaces to hands.
Significance and Impact of the Study: Infectious doses of pathogens may be transferred to the mouth after handling an everyday contaminated household object.
The role of fomites in the transmission of disease remains a controversial subject. Some epidemiological studies have suggested that contaminated surfaces may play a role in the spread of respiratory viruses (Hendley et al. 1973; Reed 1975; Hall et al. 1980) and laboratory studies have supported this hypothesis. Other studies have implicated environmental surfaces in the transmission of bacteria (Manning et al. 2001; Ekanem et al. 1983; Bures et al. 2000; Manning et al. 2001) However, the role of environmental surfaces in the transmission of disease remains an issue of scientific debate and fundamental information concerning the microbial transfer rates from environmental surfaces to the hands and from the hands to the mouth remains scarce.
Separate studies have shown that microorganisms are more efficiently transferred from nonporous than from porous surfaces. Transfer of Escherichia coli from a laminate surface to fingers was 40% up to two hours after the contamination event (Scott and Bloomfield 1990a) while the transfer efficiency of E. coli from a damp cloth to the human hand was only 0·47% (Mackintosh and Hoffman 1984). Bean et al. (1982) showed that viral transfer from porous surfaces was poorer than from stainless steel although percent transmission was not described. The efficacy of transfer of rhinovirus from a donor's fingertips to a recipient's hands via door knobs was as high as 22% (Pancic et al. 1980). Studies do not provide direct comparisons of the transmission rates from surfaces to hands of bacteria and viruses. This information may be important to determine which types of pathogens will be most affected by environmental sanitation practices.
More quantitative information regarding the transmission of viruses and bacteria from the fingertip to the lip is also needed. This information is vital to understand the possible role of environmental surfaces and cross-contamination in the transmission of disease. It is also important to establish a basis for a risk assessment approach in the domestic environment. Many infections are thought to arise within the home. Although many of these infections are not life-threatening, they do result in significant health care costs (Bloomfield 2001). This study is the first to directly compare Gram-positive bacteria, Gram-negative bacteria, and viral transfer efficiency from porous and nonporous fomites to the human hand and from the fingerpad to the lip.
Materials and methods
The study consisted of two evaluation periods: A and B. In Evaluation Period A, subjects' hands were sampled following contact with one of eight common surfaces that were inoculated with a pool of three microorganisms comprising Serratia rubidea American Type Culture Collection (ATCC, Rockville, MD) 11634, Micrococcus luteus ATCC 533, and PDR-1 phage (J. Ito, University of Arizona). In Evaluation Period B, subjects' lower lips were sampled after they had been touched with a fingertip that had been inoculated with a pool of the same three microorganisms.
Subjects were healthy adults, aged 18–65 years. All subjects signed informed consents as approved by the University of Arizona Human Subjects Committee.
Control wash and disinfection procedures
Prior to all study activities of Evaluation Periods A and B, the following control wash was performed. Hands were squirted with 70% ethanol for 10 s, subjects rubbed the alcohol thoroughly over their hands and wrists for 15 s, and then hands were washed with 2 ml of liquid Ivory (Procter and Gamble, Cincinnati, OH) for 30 s, rinsed for 15 s, and dried on paper towels. Prior to sampling the lower lip, the area was wiped for approximately 10 s with an alcohol swab/wipe.
After all study sampling involving the prepared inoculum of bacteria and phage, the following disinfection procedure was performed. Subjects' hands were squirted with 70% alcohol for 10 s, the alcohol was rubbed over hand and wrist surfaces for 15 s, and then hands were rinsed under running tap water for 15 s and dried with paper towels. Subjects then conducted an Ivory soap-and-water wash for at least 30 s, followed by a 30-s wash with Hibiclens® (4% chlorhexidine gluconate; AstraZeneca, Wilmington, DE). Subjects' lower lips (Evaluation Period B) were disinfected by twice wiping the area with an alcohol wipe (10 s each) followed by swabbing with Hibistat® (0·5% chlorhexidine gluconate; AstraZeneca, Wilmington, DE) solution.
Preparation of inoculum
In each case, the inoculum was a pooled culture of S. rubidea ATCC 11634, M. luteus ATCC 533, and PDR-1 phage. These organisms were chosen because they were low risk to the subjects and the environment. The bacteria were pigmented so they could be differentiated from the normal flora. The pooled inoculum used on/in the fomites consisted of approximately 108 CFU or PFU ml-1 of both types of bacteria and the phage. The pooled inoculum used on thefingertip consisted of approximately 106 CFU and PFU ml-1. The bacteria and phage in the pooled inoculum were enumerated daily before use, using the spread-plate method for bacteria and the agar-overlay method for phage as described in the Enumeration and Incubation section below. The volumes of inoculum used in/on the various fomites and on the fingertips are described in the Microbial Transfer and Sampling Procedures section.
The inoculum was prepared as follows. A frozen aliquot of S. rubidea was transferred to Tryptic Soy broth (TSB, Difco, Detroit, MI) incubated for 18 ± 2 h at 35 ± 2°C, and streaked for isolation onto Tryptic Soy Agar (TSA, Difco). An isolated colony was selected, transferred to a tube of TSB, and incubated for 18 ± 2 h at 35 ± 2°C for use in the study. The same procedure was used in the preparation of the M. luteus culture. The PDR-1 phage stock was prepared by adding the Salmonella typhimurium LT2 (host bacterium) to TSB and incubating for 12–18 h at 35 ± 2°C to bring the culture to a stationary phase. The culture was brought to log phase by inoculating 1 ml of the stationary phase culture into 100 ml TSB and incubating for 2–3 h at 37°C on a rotating shaker table (150–180 rev min-1). Phage suspension (0·1 ml) followed by log phase host culture (1 ml) was then added to top agar tubes melted and maintained at 48°C. The inoculated top agar tube was mixed and poured over a TSA plate, the solidified agar overlay was inverted, and overlay plates were incubated at 37°C for 24 h. After plaques were confluent, TSB (5 ml) was added to each plate and maintained at room temperature for 2 h to allow the phage to diffuse through the solution. The TSB was aspirated and centrifuged lightly, after which the solution was filtered using 0·45 µm GN-6 filters (Gelman Sciences Inc., Ann Arbor, MI), and the phage stock culture was stored at 4°C. Stock phage cultures were titrated 24 h before use. At each study start, the three cultures were pooled in TSB to achieve the concentrations described above.
Microbial transfer and sampling procedures
Evaluation period A. Transfer of microorganisms to hands from inoculated porous surfaces. Sampling was performed after contact with six fomites/surfaces, which had been inoculated with a pool of Serratia rubidea, Micrococcus luteus, and PRD-1 coliphage. (approximately 108 CFU or PFU ml-1 of each organism). The fomites were a sponge, a dishcloth, laundry, a carrot, and raw ground beef. Preparation and inoculation of the fomites/surfaces and subject contact procedures were as follows: (1) Prior to inoculation, sponges (Scotch Brite, 3M, Mt. Paul, MN) were boiled in Letheen broth (Difco, Detroit, MI) for at least 1 h on three consecutive days. Each sponge was saturated with 100 ml of pooled inoculum. Subjects were asked to wring out the sponge for 10 s after which the hands were allowed to dry for 1 m before sampling; (2) cotton dishcloths (30 × 30 cm, Dayton Hudson, Minneapolis, MN) were rinsed in distilled water and autoclaved. Each dishcloth was inoculated with 50 ml of pooled organisms. Subjects were asked to wring out the dishcloth for 10 s after which subjects' hands were allowed to air dry for 1 min before sampling; (3) Each load of laundry consisted of 10 swatches (100 cm2, either 100% cotton or 50 : 50 cotton/polyester blend, Hancock Fabrics, Tucson, AZ). Swatches were weighed dry, saturated with a known quantity of inoculum (100 ml for the cotton blend and 200 ml for the 100% cotton swatches), placed in the final spin cycle of the washer, removed, reweighed, and the remaining inoculum counts determined. Each subject transferred a load of laundry to the dryer. Subjects' hands were allowed to air dry for one min before sampling; (4) Carrots were purchased at a grocery store (Kroger, Cincinnati, OH), cut into four segments and gamma-irradiated at 25 kGy/min for 12 h, and kept frozen until the day of use to maintain sterility. On the day of the test, carrots were thawed and then dipped into the pooled inoculum. Subjects were asked to cut the carrot into pieces, after which the hand used to hold the carrot in place was allowed to air dry for 1 min before sampling. Seeded carrots were assayed for density of microorganisms by immersing the carrot into 40 ml of sterile physiological saline, vortexing at high speed for 60 s, and enumerating the indicator organisms as described in the Enumeration and Incubation section; (5) Ground beef was purchased at a grocery store (Kroger, Cincinnati, OH), gamma-irradiated at 25 kGy/min for 12 h, and kept frozen until the day of use to maintain sterility. The beef was divided into 450-gram parcels and placed into a freezer bag. A 23-ml aliquot of the pooled organisms in TSB was added to each parcel. The parcel of ground beef was kneaded for 10 min Subjects were asked to prepare four hamburger patties from one pound (450 g) of inoculated ground beef after which subjects' hands were allowed to air dry for 1 min. The ground beef was assayed for the inoculum level at the start and end of the study.
Transfer of microorganisms to hands from inoculated nonporous surfaces. A hard nonporous surface will not absorb a known amount of inoculum such a porous surface will. It was particularly difficult to apply a known volume of pooled microorganisms to the round stainless steel faucet handle due to runoff of the inoculum. Therefore, the hard surfaces tested were inoculated by dipping the fomite into a pool of the test organisms and allowing the inoculum to air dry before handling. Sampling was performed after contact with two surfaces that had been inoculated with a pool of S.rubidea, M. luteus, and PRD-1 coliphage: a phone receiver and a single-lever kitchen faucet handle. (1) The area of the phone receiver to be handled by the subject was marked. The phone receiver was disinfected with 70% ethanol and allowed to dry, then dipped into the pooled inoculum and allowed to air dry. Subjects were asked to hold the receiver for 30 s as if answering the phone. The hand was allowed to dry for 1 min and then sampled. The area of the receiver handled by the participant was also sampled using a Dacron swab (Becton Dickinson, Sparks, MD) and the density of indicator organisms determined; (2) A single-lever faucet handle was disinfected with 70% ethanol before and between each study use. The faucet handle was dipped into a solution of pooled test organisms in TSB and allowed to air dry. Each subject turned the handle on and off twice. The hand used on the faucet was allowed to air dry for 1 min and then sampled using a Dacron swab (Becton Dickinson, Sparks, MD). The residual seeded microorganisms on the faucet handle were also enumerated using a Dacron swab. The density of each test organism in the inoculum was determined in each case.
Evaluation period B. Transfer of microorganisms from fingertip to lip. Sampling of each subject's lower lip was performed after 10-s contact with a fingertip that had been inoculated with the pooled three-organism inoculum (approximately 106 CFU or PFU ml-1) described above. A total of 5 µl of inoculum was applied to the assigned finger and allowed to air dry for 30 s. The subject then placed the fingertip to the middle of the lower lip for 10 s. The fingertip and area of contact on the lip were sampled using Dacron swabs.
Bacterial sampling procedure
One or both hands (per description) were sampled following contact with the fomites, and the fingertip and lip area were sampled in the hand-to-lip transfer. Each sampling procedure was done with a Dacron swab moistened in 3 ml of Letheen broth with the excess pressed out. Following sampling, the swab(s) were returned to a tube of Letheen broth and placed on ice for microbial enumerations. For hand sampling, the swab was rubbed over the entire ventral surface of the hand (including ventral surfaces of the thumb and fingers) twice, using opposite directions of movement. When two hands were sampled, the two hands were sampled separately and the swabs from both hands were placed into a single tube of Letheen broth. The fingertip was sampled by rubbing the swab over the inoculated area of the fingertip, and the lip was sampled by rubbing the swab over the area of the lip contacted by the fingertip.
Enumeration and incubation
For enumeration of S. rubidea and M. luteus, samples were plated and counted using the spread plate technique. Samples were serially diluted to countable numbers and plated onto duplicate plates of TSA. For phage analysis, the overlay technique was used. Dilutions of phage suspension (0·1 ml) followed by log phase host culture (1 ml) was added to melted top agar tubes, the inoculated top agar tubes were mixed and poured over a TSA plate, the solidified agar overlay was inverted, and overlay plates were incubated at 37°C for 24 h. Plated samples were incubated aerobically for 18–24 h at 35 ± 2°C, after which colonies typical of M.luteus (lemon yellow) and S. rubidea (red) were enumerated as were numbers of plaques of PRD-1. Numbers reported are the average of the counts from the plates in the range of ≥25 to ≤250 CFU/PFU.
All microbial numerical results were converted to base 10 logarithms. The mean counts recovered from the hands/lip areas and fomites/fingertips were determined and these mean counts were then used to evaluate transfer efficiency using SAS version 6·12.
Fomite-to-hand transfer (Evaluation Period A)
Table 1 summarizes the fomite-to-hand transfer results. The Gram-positive bacterium, M. luteus, was transferred more efficiently than the virus or phage, PRD-1, and the Gram-negative bacterium, S. rubidea, from all but two cases. The phage was the most efficiently transferred organism from the carrot and the phone receiver. The lowest transfer rates were consistently observed for S. rubidea.
Table 1. Results from fomite-to-hand transfer (Evaluation Period A) *
| Dishcloth||10·44||6·90|| 0·04|
| Sponge|| 9·58||5·98|| 0·03|
| Faucet|| 6·13||5·59||40·03|
| Carrot|| 9·05||6·31|| 0·21|
| Hamburger|| 9·79||5·70|| 0·06|
| Phone receiver|| 6·60||6·19||41·81|
| Laundry – 100% cotton|| 9·73||6·17|| 0·13|
| Laundry – 50 : 50 cotton/polyester|| 9·39||5·99|| 0·06|
| Dishcloth|| 9·85||5·95|| 0·03|
| Sponge||10·44||6·46|| 0·02|
| Faucet|| 5·83||4·70||33·47|
| Carrot|| 7·97||5·43|| 0·35|
| Hamburger|| 8·77||3·93|| 0·01|
| Phone receiver|| 4·92||4·68||65·80|
| Laundry – 100% cotton|| 8·73||3·63||<0·01 (0·005)|
| Laundry – 50 : 50 cotton/polyester|| 8·34||2·71||<0·01 (0·0005)|
| Dishcloth||10·34||5·42||<0·01 (0·0045)|
| Sponge||11·06||6·50||<0·01 (0·0037)|
| Faucet|| 6·08||5·22||27·59|
| Carrot|| 8·97||5·85|| 0·12|
| Hamburger|| 9·91||5·12||<0·01 (0·002)|
| Phone receiver|| 6·31||5·75||38·47|
| Laundry – 100% cotton|| 9·79||4·40||<0·01 (0·003)|
| Laundry – 50 : 50 cotton/polyester|| 9·01||3·64||<0·01 (0·0009)|
All three organisms were most efficiently transferred to the hands from the phone receiver, the faucet handle and the carrot, in descending order. Transfer rates were always higher from the dishcloth than from the sponge. Lower transfer rates were consistently observed from the 50 : 50 cotton/polyester laundry swatches than from the 100% cotton swatches. Although percent transfer was higher from porous surfaces than from nonporous surfaces, the levels of contamination of the hands were often very high after handling porous fomites such as the dishcloth or the sponge.
Fingertip-to-lip transfer (Evaluation Period B)
The fingertip-to-lip transfer results are summarized in Table 2. As observed for surface-to-hand transfer, M. luteus had the highest percent transfer from the fingertip to the lower lip (40·99%). However, the transfer efficiency of S.rubidea was slightly higher than that observed for phage PRD-1.
Table 2. Transfer efficiency of bacteria and phage from hand to mouth *(Evaluation Period B)
The possible role of fomites in the transmission of disease requires further evaluation. In many nosocomial infections, the route of transmission is not documented and environmental surfaces are often not tested (Spender et al. 1986; Cone et al. 1988). The role of fomites in the domestic setting is even more difficult to assess. The number of homes that would be required to statistically evaluate the role of cross-contamination and cross-infection would be prohibitive (Bloomfield 2001). However, a first step that can be taken to assess the risk of transmission from contaminated surfaces is to evaluate the transfer efficiency rates of different types of bacteria and viruses from surface-to-surface. This is the first study to describe the transfer of two types of bacteria and a virus from a variety of common household fomites to the hands.
The present study suggests that Gram-positive bacteria are transmitted most readily from environmental surfaces followed by viruses and Gram-negative bacteria. In a study by Scott and Bloomfield (1990a), Staphylococcus aureus and Escherichia coli were transferred from a laminate surface to the fingertip at similar rates. Therefore, more research is needed to determine whether Gram-positive bacteria are, indeed, transferred more readily than Gram-negative bacteria or whether transmission rates from fomites are organism specific.
Different survival rates may have influenced some of these results. Survival rates of bacteria have been shown to differ considerably (Perez et al. 1990; Snelling et al. 1991; Falsey and Walsh 1993) and are even strain specific (Noskin et al. 1995; Neely and Maley 2000). Viral survival is also highly variable (Hall et al. 1980; Mbithi et al. 1992; Adler 1996). Although inocula were not allowed to dry for prolonged periods, differential survival abilities may have influenced some of the results here although that is unlikely. Bacteria have been shown to survive and even grow in damp objects such as contaminated cloths and ground beef (Scott and Bloomfield 1990b; Dickson and Olson 2001). In addition, the seeded fomites in this study were handled by volunteers within minutes of inoculation. Therefore differential survival or growth rates probably do not account for the different transfer rates observed between Gram-positive and Gram-negative bacteria from the objects described here.
In this study, transfer efficiency from nonporous surfaces was calculated differently than from porous surfaces. This was a consequence of the nature of the two types of surfaces. Porous surfaces can be inoculated with, and hold, a known volume of pooled bacteria and phage. Most of a measured volume of inoculum will run off of hard, smooth, curved, nonporous surfaces. Hence transfer efficiency rates must be calculated in a different manner for these two types of surfaces. Transfer rates from hard, nonporous surfaces were more efficient than from porous surfaces. A porous surface, such as a sponge, offers many deep recesses in which bacteria and viruses reside becoming less accessible to the human hand. A hard smooth surface does not offer crevices or passages in which microorganisms may hide, hence higher transmission. However, high levels of hand contamination occurred in spite of poor transfer rates from some of the porous fomites. After squeezing out a sponge or a dishcloth, the subjects' hands were highly contaminated.
These results suggest that commonly handled objects that are microbially contaminated can serve as reservoirs of bacteria and viruses that can easily transfer to the hands through direct contact, which in turn can be easily transferred to the lip. Because of the seeding of these fomites, concentrations of organisms are possibly higher than what one would find in the household. However, previous studies by the authors show that concentrations of coliforms (which include opportunistic pathogens) are sometimes quite high in the common household (Rusin et al. 1998). These authors found that the water recovered from the common kitchen sponge contained 6·51 (log10) coliforms per ml, reflecting very high numbers (approximately 3·2 × 108 cells) in the sponge itself. In addition, early work (Davis et al. 1968) reported counts of E. coli in domestic dishcloths of 107. Based on a 0·0037% transfer efficiency, 11 840 coliforms would be transferred to the hand. Assuming 3·2% (379) of these bacteria are distributed on the fingertip, then 34% or 129 cells would be transferred into the mouth. This means that if some members of Enterobacteriaceae, such as Shigella or E. coli O157:H7, were in the dishcloth or sponge in high numbers, infectious doses could easily be transferred to the lip or mouth as low numbers of these pathogens may cause disease (Dupont et al. 1989; Boyce et al. 1995). It should also be remembered that the risk of infection via a contaminated dishcloth is heightened by multiple uses of the cloth during the day.
The high transfer rates seen in this study suggest that a telephone receiver could also easily serve to transmit disease. Large numbers of Salmonella may be excreted in the stool of an infected person (up to 1010Salmonella per gram of faeces (Feachem et al. 1983). Hence, if only 0·001 g of residual stool were transferred from an infected person's contaminated hand to a telephone receiver, the next user would have 107 104 Salmonella cells on the fingertip. If this were placed in the mouth, the person would receive a dose of 36 383 cells that could easily result in disease (Blaser and Newman 1982). The transmission of pathogenic bacteria via a telephone receiver may be enhanced by multiple uses of the telephone and by the fact that bacteria may survive for hours on hard surfaces, especially when dried in natural excretions (Scott and Bloomfield 1990a; Snelling et al. 1991).
In comparison to bacteria, very little information is available in the literature regarding the concentrations of pathogenic human viruses in the domestic environment. However, we do know that viruses can survive (remain infectious) for hours to days on a hard surface (Ansari et al. 1988; Brady et al. 1990), that low numbers of infectious units have been shown to cause disease (Douglas 1970; Ward et al. 1986), and that large numbers are often found in human excretions (Gwaltney 2000; Rusin et al. 2000). Hence, if a virus such as the rotavirus or Norwalk agent were on the surface of a fomite such as a telephone receiver, infectious doses could easily be transferred to persons handling the fomite under ordinary circumstances. As an example, if a telephone receiver were contaminated with a low concentration of rotavirus agent (e.g. 10 000 infectious particles), 6580 of these would be transferred to the hand during normal use of the telephone with 211 of them found on the fingertip. Our results with phage PRD-1 show that 72 infectious particles could be ingested by the host, which could result in disease since the infectious dose has been shown to be as low as 1 PFU (Ward et al. 1986).
Likewise, the faucet could also transmit the rhinovirus from person to person. Gwaltney et al. (1978) found that hand-to-hand transmission of this virus was a more efficacious route of infection than the aerosol route and the virus was often detected on the hands of volunteers. The 50% infectious dose (ID50) for this virus is less than one 50% tissue culture infectious dose (TCID50) (Douglas 1970). The transmission of rhinovirus from contaminated surfaces has resulted in disease, as has been shown by Gwaltney and Hendley (1982). These authors showed that 50% of recipients developed rhinovirus infection after exposure to virus-contaminated coffee cup handles and 56% of volunteers became infected after exposure to contaminated plastic tiles.
Food products may also bring pathogens into the home. Cross contamination of kitchen surfaces by contaminated meat products has been demonstrated (Dewit et al. 1979; Deboer and Hahne 1990). Indeed, a variety of raw meats, including raw ground beef (Daise et al. 1986; Gill et al. 1996), have been shown to be colonized by high numbers of bacteria (Shaw et al. 1987; Gill and Jones 1996; Ramos and Lyon 2000). Little information is available assessing bacterial levels on raw vegetables entering the home.
It is difficult to assess the risk of laundry as a fomite. Certainly, the percent of seeded organisms transferred to the hand was very small. However, in a recent study (Larson and Gomez-Durate 2000), the use of a commercial laundromat and the practice of not using bleach in the laundry were the only two factors that correlated strongly with disease transmission within the households of inner city populations.
More quantitative information needs to be gathered to evaluate the role of contaminated inanimate surfaces in the domestic setting for disease transmission. In order to perform a risk assessment in the domestic environment for bacterial and viral pathogens we need more information regarding: (1) occurrence and levels of potential pathogens in homes (2) transfer efficiencies of various microorganisms from surfaces to hands (3) survival times of pathogens in natural secretions found in foods and in/on inanimate objects (4) levels of pathogens secreted by humans, and (5)infectious doses to perform risk assessments in the household.
This work was supported by Procter and Gamble, Cincinnati, OH. The authors would like to thank K. Wiandt, W.Billhimer, J. Philippo and B. Keswick for their assistance in preparing this manuscript.