Effects of chronic herbivory and historic land use on population structure of a forest perennial, Trillium catesbaei

Applied Vegetation Science, Dec, 2007 by Michael A. Jenkins, Christopher R. Webster, Janet H. Rock

In our study, the low T. catesbaei population densities at CC compared to the WOS reference site suggest that reproduction may not be sufficient to offset herbivory-induced mortality. Severe disturbances that leave few residual plants result in slow rates of population recovery because Trillium species have very short seed dispersal distances (Kalisz et al. 1999; Kahmen & Jules 2005), and post-disturbance seed production is inadequate to mitigate this limited dispersal. While both CC and WOS have experienced fairly severe anthropogenic disturbance in the past, the continued effects of deer herbivory in CC have caused a divergence in the rate of recovery for Trillium populations at the two sites. Herbivory at CC has likely maintained low densities of flowering T. catesbaei, resulting in low seed production and exacerbating the effects of limited dispersal distance (Knight 2003).

While site variables may explain some of the differences in T. catesbaei populations between CC and WOS, we took great care to avoid variability between sites. While generally southerly, aspects at our study sites did differ. Our sites at W OS were south to west facing, while our sites at CC were more south to southeast facing. In the southern Appalachians, vegetation communities on sheltered slope positions at low elevations, such as our sites at CC and WOS, are less influenced by differences in aspect (McNab 1992). Further, because T. catesbaei is most common in mesic to dry-mesic forests (Case & Case 1997), the more mesic aspects at CC are more favorable to T. catesbaei than the dryer more-westerly aspects at WOS.

According to observations made by the Park's first naturalist, Arthur Stupka, Trillium species other than T. catesbaei were plentiful in CC prior to the eruption of the deer population (Webster et al. 2005). Currently, T. catesbaei is the only Trillium species present insufficient numbers to be tallied as distinct populations. Trillium catesbaei appears to employ a stress tolerance life-history strategy (Grimes 1977) that allows it to avoid directly competing with other, more habitat sensitive, Trillium species in GSMNP Compared to these species, T. catesbaei is a habitat generalist that occurs in a variety of sites typically unfavorable to other Trillium species, including thickets of ericaceous shrubs and dry open woodlands (Case & Case 1997). In our study, the oldest individual T. catesbaei plant we sampled was 17 years old. Published ages for other Trillium species range from 30-40 years for T. grandiflorum (Hanzawa & Kalisz 1993; Vellend 2005) to 72 years for T. ovatum (Jules 1998). Generalist species, such as T. catesbaei, typically grow in less ideal, but more widespread, habitat and are generally smaller in stature than other species of the same genus. These traits may allow them to persist under intense herbivory when populations of larger more habitat-sensitive Trillium species disappear.

Several studies have examined the utility of Trillium species as indicators of deer browse intensity (Anderson 1994; Augustine & Frelich 1998; Augustine & Calestra 2003; Kirschbaum & Anacker 2005). Plant height and percent reproductive individuals within populations were found to be indicators of browsing intensity and deer abundance (Anderson 1994; Augustine & Frelich 1998; Augustine & deCalesta 2003). Augustine and deCalesta (2003) further recommended that population structure (frequency of small and large plants) be examined to assess browse intensity. In our study, the low densities of single-leaf and flowering (mature) plants observed at CC relative to those observed at WOS support this recommendation.