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

[FIGURE 2 OMITTED]

Plant height and age were significantly associated at both CC and the WOS reference site (P < 0.001; Fig. 3). The relationship was weaker in T. catesbaei populations at CC where age only explained 13% of the variation in plant height compared to 54% for populations at WOS. In general, larger plants were older individuals at both sites. At a given height, plants at CC were typically younger than those at WOS. Results of a multiple regression of plant height versus minimum age and site indicated that the slope of the relationship between height and age varied significantly between CC and WOS (P < 0.001). The intercepts of the respective curves, however, did not vary significantly between sites (P = 0.704).

According to binary logistic regression, smaller and younger individuals were significantly more likely to flower at Cades Cove than at Whiteoak Sink (Fig. 4, Table 2). Site was a significant predictor in both the height and minimum age models (P = 0.006 and 0.034, respectively). The site by age interaction term was also significant (P = 0.026); however, the site by height interaction was not significant (P = 0.276). The relationship between flowering likelihood and age is particularly striking; a 10 year-old plant at Cades Cove has a greater than 80% probability of flowering, compared to a 30% probability at Whiteoak Sink (Fig. 4). Few plants younger than 7 years of age flowered at either site.

[FIGURE 3 OMITTED]

Discussion

While reductions in plant height and flowering success observed in browsed populations may be a result of reduced growth in the short-term (Knight 2004), the long-term effect of herbivory on plant mortality has not been studied. In Cades Cove, chronic herbivory has not caused a population of stunted individuals distributed across all age classes, but has rather created a truncated age structure lacking mature individuals (>9 years minimum plant age), suggesting that mortality of older individuals has occurred. Further, our results show that the CC populations contained fewer single-leaf individuals than expected (Table 1), suggesting that recruitment has been impacted by herbivory. Although the contemporary deer population in the Cove is lower than historic levels, the lack of older individuals of T. catesbaei in our study suggests that population age structures have not recovered.

[FIGURE 4 OMITTED]

A loss of mature plants in a population of Trillium greatly alters recruitment since members of the genus typically require more than 10 years to reach the reproductive stage of development (Ohara 1989). In our study, CC contained an extremely low density of flowering individuals (40/ha) and most surviving plants were stunted (Fig. 3). While the presence of younger plants in CC indicates that some plants are able to escape herbivory and reproduce, single-leaf plant density is less than one percent of that at the Whiteoak Sink reference site. Some individual plants may be able to flower in refugia or at a younger age. However, episodic reproduction that results in younger populations of low density will likely result in poor persistence in the long-term (Endels et al. 2004). Individual species of Trillium have been reported to be non-clonal and thus only reproduce through seed (Wright & Barrett 1999; Ohara et. al. 2001; Kalisz et al. 2001). In our study, we found no evidence of vegetative reproduction in the 429 plants we excavated. Chronic browsing has significantly altered the stage structure of T. catesbaei populations in CC.