4To be able to deal with uncertainty is of primary importance to all organisms. When 5 cues provide information about the state of the environment, organisms can use them 6 to respond flexibly. Thus information can provide fitness advantages. Without environ-7 mental cues, an organism can reduce the risks of environmental uncertainty by hedg-8 ing its bets across different scenarios. Risk mitigation is then possible by adopting a 9 life-history of bet-hedging, either randomly switching between phenotypes (diversifying 10 bet-hedging) or adopting intermediate phenotypes (conservative bet-hedging). Hence, un-11derstanding patterns of bet-hedging is necessary in order to quantify the fitness benefit of 12 environmental cues, since it provides a baseline fitness in the absence of informative cues.
13Quantifying fitness benefits in terms of mutual information reveals deep connections be-14 tween Darwinian evolution and information theory. However, physiological constraints 15 or complex ecological scenarios often lead to the number of environmental states to ex-16 ceed that of potential phenotypes, or a single intermediate phenotype is adopted, as in the 17 case of conservative bet-hedging. Incorporating these biological complexities, we gen-18 eralise the relationship between information theory and Darwinian fitness. Sophisticated 19 bet-hedging strategies -combining diversifying and conservative bet-hedging -can then 20 evolve. We show that, counterintuitively, environmental complexity can reduce, rather 21 than increase, the number of phenotypes that an organism can adopt. In conclusion, we 22 develop an information-theoretic extensible approach for investigating and quantifying 23 fitness in ecological studies. 24 1