Oecologia (1988) 76:353 363 Oecologia 9 Springer-Verlag1988 Effects of fire season on flowering of forbs and shrubs in longleaf pine forests William J. Platt*, Gregory W. Evans**, and Mary M. Davis*** Tall Timbers Research Station, Route 1, Box 678, Tallahassee, FL 32312, USA and Department of Biological Science, Florida State University, Tallahassee, FL 32306, USA Summary. Effects of variation in fire season on flowering of forbs and shrubs were studied experimentally in two longleaf pine forest habitats in northern Florida, USA. Large, replicated plots were burned at different times of the year, and flowering on each plot was measured over the twelve months following fire. While fire season had little effect on the number of species flowering during the year following fire, fires during the growing season de- creased average flowering duration per species and in- creased synchronization of peak flowering times within spe- cies relative to fires between growing seasons. Fires during the growing season also increased the dominance of fall flowering forbs and delayed peak fall flowering. Differences in flowering resulting from variation in fire season were related to seasonal changes in the morphology of clonal forbs, especially fall-flowering composites. Community lev- el differences in flowering phenologies indicated that timing of fire relative to environmental cues that induced flowering was important in determining flowering synchrony among species within the ground cover of longleaf pine forests. Differences in fire season produced qualitatively similar ef- fects on flowering phenologies in both habitats, indicating plant responses to variation in the timing of fires were not habitat specific. Key words: Flowering phenology- Fire season- Longleaf pine forests - Synchronization of flowering - Clonal growth Ground cover vegetation containing large numbers of flow- ering plants typically is present in habitats in which fires occur frequently. Fire affects flowering in these habitats (e.g. Daubenmire 1968; Vogl 1973; Rowley 1970; Dickin- son and Dodd 1976; Christensen 1981; Gill 1981b; Gun- derson et al. 1983; Whelan 1986). Within species, both the timing of flowering (Curtis and Partch 1950; Gill and In- gwerson 1976; Abrahamson 1984) and numbers of flower- ing stems produced (Burton 1944; Stone 1951 ; Kucera and Ehrenreich 1962) change following fire. Effects vary with * Current address and address for offprint requests." Department of Botany, Louisiana State University, Baton Rouge, LA 70803, USA ** Current addresses: Center for Prevention, Research and Bio- metry, Bowman Gray School of Medicine, Wake Forest Universi- ty, Winston-Salem, NC 27103, USA *** Center for Wetlands, University of Florida, Gainesville, FL 32611, USA season of fire (Gill 1981a; Henderson et al. 1983; Lovell et al. 1983; Snyder 1986). While there have been studies of flowering phenologies in fire-dominated habitats (e.g., Parrish and Bazzaz 1979; Anderson and Schelfhout 1980; Tepedino and Stanton 1980; Rabinowitz et al. 1981), the effects of fire upon flowering phenologies of coexisting spe- cies have received little attention. In this study, we experimentally manipulated fire season in longleaf pine forests of northern Florida, USA, and mea- sured characteristics of flowering by forbs and shrubs in the ground cover during the year following fire. We ad- dressed three questions. First, did season of" fire affect numbers of species that flowered and/or species-specific characteristics of flowering (i.e., numbers of flowering stems per species or the duration of flowering per species)? Sec- ond, were effects of fire season on flowering similar in dif- ferent habitats within longleaf pine forests? Third, did changing the season of fire alter characteristics of flowering by the flora in the ground cover of longleaf pine forests (i.e., the evenness and synchronization of flowering among species, or the timing of peak flowering)? Methods Study area. This study was conducted in second growth longleaf pine (Pinus palustris) forests in the St. Marks Na- tional Wildlife Refuge, Wakulla County, Florida. Old river bars and stream channels, once part of an old (Pleistocene) delta of the Ochlockonee River, constitute two distinct hab- itats that represent extremes of a local moisture gradient within the study area. Sandhills (old river bars) are xeric and are located on deep, well drained, layered sands. In contrast, flatwoods (old stream channels) are located on shallow layers of sand and clay above limerock. The water table is usually just below the soil surface, and lowest areas may be seasonally inundated. Within each habitat, the to- pography is generally flat, with slight slopes toward drain- ages. Although the entire refuge was timbered in the early 1900s, the presence of ground cover species typical of old- growth longleaf forests that have not been cleared and plowed (see Wells and Shunk 1931; Hebb 1957) indicates human disturbance of the soil has been minimal. Aristida stricta dominates the ground cover in both habitats. A number of other species of grasses, as well as numerous forbs and shrubs, are present in the ground cover of both sandhills and flatwoods. Vegetation is more dense in the latter habitat (Davis 1985). Virtually all species in the