Sir

Phenological data (observing seasonal plant and animal activity driven by environmental factors) have recently been used to suggest that the length of the growing season in Europe has increased over the past 30 years by perhaps a week in both spring and autumn1. These data could be interpreted somewhat differently, particularly those relating to the onset of autumn.

Clearly, the date of bud burst is a reliable indicator of the earliness of spring. To break quiescence, buds must accumulate a particular heat sum, in degree-days, above a certain threshold temperature. This is usually cited as 5 °C, presumably the point below which respiration is negligible. Climate changes occurring more than 1,000 years ago have been inferred from the recorded dates of such phenology-related events as the Kyoto cherry blossom festival2. Rarely, although perhaps increasingly frequently if the climate is getting warmer, there may be insufficient winter chilling to break rest completely, so that a late bud-burst date could mean a mild winter, not a cool spring.

Choosing a reliable phenological event with which to judge the arrival of autumn is more problematic, and dates of leaf colour change and leaf drop may be inappropriate. Various factors have been cited as causing these events, including cold, drought, photoperiod, clear weather and injury3. But leaves age4. Just as buds have a required heat accumulation for flushing, it seems that leaf life and senescence are also mediated by heat accumulation, at least in some species5. This is more likely in determinate species such as oaks, where all leaves are roughly of one age cohort5. A later end to the growing season, as judged by later leaf fall/drop dates, could be due to the fact that the leaves' heat sum is not met as early as in previous years because of later bud burst and/or cool summer temperatures, rather than a mild autumn. Incorporating summer temperatures into the model might usefully modify the conclusions about the date of the end of the growing season.

More reliable ways of assessing the effects of environment on phenology require more accurate estimates of phenological factors. The real start of spring is when mitosis starts in tissues adjacent to shoot apical meristems; the end of the growing season is when mitosis stops in the apical meristem. Unfortunately, these events are far more difficult to observe, and require destructive sampling.