Phenological measurements of microsporogenesis in trees.

The value of two heat sum methods, one linear (degree days > 5 degrees C) and the other curvilinear (period units), were assessed together with calendar days as predictors of the duration of microsporogenesis in seven natural stands of Norway spruce (Picea abies (L.) Karst.) and eleven natural stands of Scots pine (Pinus sylvestris L.). Microsporogenesis was divided into two subperiods: March 19 to tetrads (i.e., the end of meiosis) and tetrads to anthesis. The total period from March 19 to anthesis was also assessed. The methods were compared on a calendar day basis. When annual deviations between the predicted (stand means) and the observed annual heat sums were converted to days, the period unit method outperformed the other methods for both subphases and for the total period of microsporogenesis. The degree day parameter was more variable but a better predictor of the duration of the initial phase up to tetrads and of the total period than the calendar day parameter, but the calendar day parameter more accurately predicted the duration of the subphase from tetrads to anthesis. The heat sum methods were better predictors of the duration of development of microsporogenesis in exceptionally cold or warm years than the calendar day method.