Using Heatsum Models to Account for Geographic Variation in the Floral Phenology of Two Ericaceous Shrubs

The floral phenology of two ericaceous shrubs - Kalmia polifolia and Ledum groenlandicum - was monitored at eight peatbogs along an 800 km latitudinal transect in Ontario, Canada. Plants at different sites reached the same stage of floral development up to 3 weeks apart. The same was true when plants from four of the peatbogs were grown together in a transplant garden. Three alternative models were considered to interpret this phenological variation. The accuracy of the models was found to depend on the phenological event (i.e. flowerbuds appear versus flowering) and on the amount of inter-site variation apparent in the transplant garden. A model based on the assumption that the heatsum (i.e. cumulative degree-days) required for the phenological event was proportional to the total annual heatsum for the plant's site of origin was the most accurate of the three models tested in predicting the appearance of flower- buds but was least accurate in predicting flowering time. Flowering could be predicted more accurately using models based on the assumption that the required heatsum depended on a plant's latitude of origin (for L. groenlandicum) or that the required heatsum did not change with the plant's site of origin (for K. polifolia). As long as the appropriate type of heatsum model was used, it was possible to account for most of the variation in floral phenology recorded for the two species.