Seasonal cold hardiness of red raspberry (Rubus idaeus L.) canes was measured by freeze-induced electrolyte leakage test and visual rating of injury. Leakage data were transformed to percentage-adjusted injury values and related to lethal temperature by graphical interpolation and by the midpoint (T50) and inflection point (Tmax) estimates derived from three sigmoid (the logistic, Richards, and Gompertz) functions. Tmax estimates produced by Richards and Gompertz functions were corrected further using two different procedures. The 10 leakage-based hardiness indices, thus derived, were compared to lethal-temperature estimates based on visual rating. Graphical interpolation and Tmax of the logistic or T50 of the Gompertz function yielded lethal-temperature estimates closest to those obtained visually. Also, Tmax values of the Gompertz function were well correlated with visual hardiness indices. The Richards function yielded hardiness estimates deviating largely from visual rating. In addition, the Richards function displayed a considerable lack of fit in several data sets. The Gompertz function was preferred to the logistic one as it allows for asymmetry in leakage response. Percentage-adjusted injury data transformation facilitated curve-fitting and enabled calculation of T50 estimates.