New epidemic diseases have an evolutionary advantage if they are of "intermediate" severity, research shows.
Scientists tested the theory that pathogens (disease-causing organisms) that inflict intermediate levels of harm on their host are the most evolutionarily successful.
The study, by the University of Exeter, Arizona State University and Auburn University, found that natural selection favours pathogens of intermediate virulence (how much harm a pathogen causes) at the point the disease emerges in a new host species.
This occurs because virulence and transmission are linked, with virulence arising because pathogens need to exploit hosts to persist, replicate and transmit.
While too-low virulence will be detrimental for pathogens if they cannot transmit, virulence that is too high will also be a disadvantage if infection kills hosts so fast that the pathogen does not have time to transmit.
Over time, pathogens that show intermediate levels of virulence should therefore have an evolutionary advantage.
"For a long time, conventional wisdom held that new diseases evolved to become harmless," said Dr Camille Bonneaud, of the Centre for Ecology and Conservation on Exeter's Penryn Campus in Cornwall.
"Although theoretical developments in evolutionary biology in the 1980s showed that this was not necessarily the case, such belief still holds firm, even today.
"Our study focussed on the 'virulence-transmission trade-off' hypothesis, which allows us to make predictions about pathogen evolution.
"Experimental evidence for this theory is rare, but we were able to test it by using more than 50 variants of the infectious bacterial pathogen Mycoplasma gallisepticum, which infects house finches."
In the study, house finches from populations that had never encountered the disease were exposed to one of the different variants, simulating conditions at epidemic outbreak.
"We found that variants that were more virulent transmitted faster, but that variants of intermediate virulence were the most evolutionarily successful," Dr Bonneaud said.
"Our results therefore provide support for using the virulence-transmission trade-off hypothesis as a framework for understanding and predicting emerging pathogen evolution."
Counter to commonly held beliefs, however, variants of the pathogen that replicated faster during infection and achieved higher densities did not transmit better or faster than those that achieved lower densities.
"This tells us that transmission is not always a numbers game and that we cannot use pathogen numbers as a proxy for their success."
The paper, published in the journal Evolution Letters, is entitled: "Experimental evidence for stabilising selection on virulence in a bacterial pathogen."