How directly-transmitted pathogens benefit from harming hosts is key to understanding virulence evolution. It is recognized that pathogens benefit from high within-host loads, often associated with virulence. However, high virulence may also directly augment spread of a given amount of pathogen, here termed “spreadability”. We used house finches and the conjunctival pathogen Mycoplasma gallisepticum to test whether two components of virulence– the severity of conjunctival inflammation and behavioral morbidity produced– predict pathogen spreadability. We applied ultraviolet powder around the conjunctiva of finches that were inoculated with pathogen treatments of distinct virulence and measured within-flock powder spread, our proxy for “spreadability”. When compared to uninfected controls, birds infected with a high-virulence, but not low-virulence, pathogen strain, spread significantly more powder to flockmates. Relative to controls, high-virulence treatment birds both had more severe conjunctival inflammation– which potentially facilitated powder shedding– and longer bouts on feeders, which serve as fomites. However, food peck rates and displacements with flockmates were lowest in the high-virulence treatment relative to controls, suggesting inflammatory rather than behavioral mechanisms likely drive augmented spreadability at high virulence. Our results suggest that inflammation associated with virulence can facilitate pathogen spread to conspecifics, potentially favoring virulence evolution in this system and others.

The dataset was collected using 45 wild-caught house finches that were housed in groups in an outdoor aviary. One bird per flock  (termed the "index" bird) was treated with one of three pathogen treatments: sham inoculation controls, infection with a low-virulence strain of the bacterial pathogen Mycoplasma gallisepticum, or infected with a high-virulence strain of Mycopasma gallisepticum. We have three types of data collection that resulted in three .csv files: first birds were captured at multiple timepoints to assess eye scores in response to treatment, take swabs for pathogen load, and to apply or collect data on UV powder spread as our metric of spreadability. Second, we collected behavioral data (foraging bout lengths at feeder ports and displacement interactions at feeder ports) via Radio Frequency Identification Devices that were present at each feeding port, and recorded the presence and duration of any bird at the feeder port via the PIT tags embedded in each birds color bands. Finally, we took video data and, by watching videos, collected information on the number of times that each bird stuck its head into the feeder port during the time it sat on the feeder port.

R (open-source) is the only program needed to process and our data files, which are all in .csv or .txt format.