In this part, I will talk about non-pharmaceutical interventions for zoonotic pathogens. After completing this part, you should be able: to describe potential non-pharmaceutical control measures for zoonotic pathogens; to describe interventions for controlling avian influenza transmission in poultry; and to describe interventions for controlling the risk of human infections with avian influenza. Zoonotic pathogens can pose a major risk to human health. In the past decade we have seen increasing concerns about chikungunya virus, West Nile virus, Ebola virus, MERS coronavirus, and various influenza viruses. In 2003, the SARS coronavirus emerged probably originally from similar coronaviruses found in Chinese horseshoe bats, via palm civets or other small mammals sold for consumption of their meat in markets in southern China. In 2009, a new influenza type A subtype H1N1 virus emerged from a reassortment of swine influenza viruses in North America, and then rapidly spread around the world. More recently, variant swine H3N2 viruses have caused generally mild human infections in North America, while avian H7N9 viruses have caused human infections and death in China. Meanwhile in the Middle East, the MERS coronavirus has caused human infections and deaths. Some emerging infections that remain largely zoonotic, such as H7N9 and MERS coronavirus, with low transmissibility between humans. In these cases, control measures have to focus on the animal side, and the animal human interface. Ideal interventions would reduce the prevalence of infections in the animal hosts, thus reducing rates of human infections. In addition, certain measures might be possible to reduce human exposure to the animals. Going back to 1997 in Hong Kong, a new highly pathogenic avian influenza H5N1 virus emerged and caused 18 cases of laboratory-confirmed human infections, of which six were fatal. When the outbreak was identified, the government rapidly intervened, closing all local live poultry markets, killing more than a million chickens territory-wide, and temporarily halting import of live chickens from mainland China. As a result of those drastic measures, the local H5N1 outbreak was curtailed, and some time later the sale of live poultry and chicken imports resumed. In the following years, Hong Kong implemented various routine control measures to reduce the prevalence of avian influenza viruses in local live poultry markets. In 2001, the government introduced monthly rest days, where once per month all live poultry remaining in the market are culled, and then a day is spent cleaning and disinfecting the market. In 2002, the sale of live quails was banned. In 2003, a second rest day was introduced every month. In 2004, a vaccine against H5N1 was introduced for all poultry. In 2008, there was a complete ban on holding live poultry overnight in live poultry markets, and all live poultry remaining at the end of the day were slaughtered. We examined the detection rates of H9N2 viruses in live poultry markets in Hong Kong as a marker for the circulation of avian influenza viruses and therefore the impact of the various control measures. We found that as the interventions became more and more stringent, circulation of H9N2 was lower and lower, and the most recent strategy of holding no poultry overnight has kept H9N2 detections at a very low level. We know that H9N2 is present in the poultry farms, and the key insight from this and other similar studies is that live poultry markets can act as amplifying locations, due to the high density of poultry, and the conditions in which the poultry are kept prior to sale. While we examined H9N2, which has low pathogenicity in poultry and is not thought to cause disease in humans, it is reasonable to believe that the same interventions would also be effective against other avian influenza viruses. More recently, an avian influenza H7N9 virus has caused laboratory-confirmed human infections in China since early 2013. A large epidemic occurred in the Yangtze River delta in eastern China in the spring of 2013, and was curtailed by the closure of hundreds of live poultry markets in Shanghai, Nanjing, Hangzhou and Huzhou. The decision to close markets was made because most cases of infection had occurred in urban residents, and most contacts with live poultry occurred in live poultry markets in these locations. While the timing of closures differed between cities, there was a considerable drop in incidence of laboratory-confirmed cases within a few days after the closures in each city. We used a statistical model to show that there was a 97 percent to 99 percent reduction in incidence of cases after the closures. As part of the analysis we were also able to estimate the incubation period to have an average of 3.3 days, based on how quickly the incidents dropped after the closures since some cases might have been infected before the closures but only developed symptoms afterwards. This estimate of the incubation period was very similar to a separate estimate we made based on assessment of prior exposures to live poultry among laboratory-confirmed cases, where the average incubation period was 3.1 days. These findings demonstrate the potential value of live poultry market control to reduce the risk of human infection with H7N9. While the H7N9 virus cannot spread easily between humans at present, the risk that the virus could adapt and develop greater transmissibility between humans is only likely to increase as more human infections occur. However, closure of live poultry markets has serious economic consequences, and more sustainable interventions such as regular rest days may be required for ongoing control of this and other avian influenza viruses.