PS 10-122 - Defend or grow? The prickly case of suburban forest shrubs.

Abstract: Some plants exhibit defenses to protect themselves from herbivory by increasing growth of mechanical structures like prickles. However, these plants may experience a trade-off between higher defense production and growth. In suburban forests, overabundant deer and competition from a plant community that typically includes abundant nonindigenous, invasive plant species can impose significant stresses on plants that experience this trade-off, as they must be well defended from deer but also need sufficient growth in order to compete. Here, we focused on the indigenous, prickle-defended species Rubus allegheniensis (common blackberry), for which previous work showed greater prickle production per unit of stem growth (prickle density) under greater ambient deer pressure, and in deer-access plots versus fenced plots. These findings prompted a fuller investigation of prickle defense by testing a system-wide hypothesis through structural equation modeling (SEM), with a focus on how forest age, photosynthetically active radiation (PAR), and stress from deer and herb layer competitors directly and indirectly influence the prickle density, height, and proportion cover of blackberries. Data were obtained from 60 plots across four suburban forests in Mercer County, NJ, USA, where fenced and unfenced 16 m² plots have been part of a white-tailed deer exclosure experiment since 2012. We used piecewiseSEM in R to test our system-wide hypothesis, refining it to reach a fitted model containing only important paths. The fitted model (Fisher’s C=24, df=26, P=0.58) explained some of the variation in blackberry height, with positive effects of forest age, light, and competitors, but no effect of deer pressure. There also were positive effects on prickle density from competitors, PAR, and deer pressure, but the amount of variation in prickle density explained by these drivers was low. Forest age had relatively large effects on competitor cover, light, and blackberry height, which suggests that these variables differed largely because of their location, as forest age was measured on the forest level. The multivariate SEM approach allowed us to put our previous univariate results about deer effects on defense into a fuller ecological context. However, it revealed that the ecological factors we expected would increase prickle defense at the expense of stem growth still leave us with an incomplete explanation. This leads us to hypothesize that variation in prickle density may be largely due to genetic factors, rather than the large ecological stresses imposed on plants in a suburban forest environment filled with invasive plants and overabundant deer.