STRESS AND ECOSYSTEMS 67 



ability can have different effects on different organisms and that 

 constancy and contingency may be as important as elements of 

 adaptive strategies as they are as environmental constraiints on 

 evolution. 



Since predictability implies anticipation, predictable environ- 

 ments can be anticipated, but unpredictable environments cannot. 

 The abihty to anticipate rests on the system subjected to the 

 environment; thus the relationship between the system and its 

 environment becomes crucial in defining the degree of predictability 

 or unpredictability an environment represents to a system. Slobodkin 

 and Sanders (1969) said that predictabiUty v^as partly dependent on 

 the organism and was not necessarily an environmental property. We 

 would expect that, all other conditions being equal, the smaller and 

 simpler the system, the greater its susceptibility to short-term 

 environmental variability. However, miniaturization may be a suc- 

 cessful evolutionary response to a long history of wide fluctuations. 

 Regardless of strategy, if we assume that survival implies adaptability 

 and that adaptability enhances anticipation of the environment, we 

 would call the same environment an unpredictable one for the 

 system that perishes and a predictable one for the system that 

 survives. 



Environments have also been described as constant, inconstant, 

 cycUc, and randomly fluctuating or as being in any intermediate 

 state. Stearns (1978) showed that it is incorrect to think in terms of 

 just two contrasting situations (predictable or unpredictable). He 

 found that even within one type of ecosystem (reservoirs) analysis of 

 four variables separated 19 of them into distinct classes of variation 

 that formed a diverse mosaic of environments. Stearns concluded 

 that different environments could rank differently for any measure 

 of fluctuation or stability for each significant environmental 

 variable. Whittaker (1975) pointed out several possible characteristics 

 of fluctuations that need consideration — quahty and duration, 

 relative amplitude of a regular fluctuation, relative irregularity, and 

 duration of the pattern in evolutionary time. Obviously the number 

 of points along a continuum of possible intensities of stress 

 associated with these types of environmental situations is significant 

 to any analysis of stress and ecosystems. 



The classification of an environment in any category depends on 

 our abihty to analyze and understand data and on the completeness 

 of the data sets. Complex time-series analyses are required to detect 

 long-term cyclic events on data sets that before analysis appear to be 

 random and unpredictable. The long-term cycles, although perhaps 

 not important to organisms, may be significant to populations or 



