Climate change introduces numerous uncertainties over the livelihoods of farming communities that depend heavily on weather and climate. Rain-fed farmers in developing countries are among the most vulnerable communities. However, climate risks are not new to farmers. Coping with ‘natural variability’ of climate has been a constant challenge faced by farmers even though broad sweeping change in climate due to anthropogenic causes is a relatively new prospect. Some argue ‘climate change’ could be significantly different from ‘climatic variability’ known to and experienced by farmers. In spite of this it is widely accepted that understanding farmers’ behavior towards adapting to climatic variability could generate useful insights in facing the risk of climate change. In Sri Lanka, the village tank farming community in the dry zone is one of the most vulnerable communities thereby deserving the priority attention of policy makers. This study is based on information gathered in Anuradhapura district of Sri Lanka. It depends mainly on information from secondary sources supplemented by qualitative primary information. Analysis was guided by recently introduced behavioral economics concepts of decisions based on experience. Accordingly adaptation is viewed as a response to the climate perceptions of farmers' aided by judgments based on heuristics. Farmers' adaptation decisions can be explained on the basis of their perception of climate variability with two major components. Firstly, farmers perceive climatic variability as an average annual pattern with variable probabilities of seasonal distribution of precipitation. Farmers base their long-term adaptation responses on this perceived average pattern and many of the choices made by them in the existing farming system and resource management practices can be explained accordingly. The average pattern of variability is largely a shared perception and therefore enables the option of joint adaptation. The land allocation practice popularly known as ‘Bethma’ provides a fine example for this. Secondly, farmers also perceive feasibility of random shocks with variable probabilities across the average pattern. This gives rise to short-term responses in the farming system activities. Such responses seem to be more individually oriented and reflect the variations in individual perceptions of climate risks.