Testing the Enemy Release Hypothesis on tall-statured grasses in South Africa, using Arundo donax, Phragmites australis, and Phragmites mauritianus as models.

The Enemy Release Hypothesis (ERH) predicts that introduced plant species can escape herbivory and therefore have a competitive advantage over native plants, which are exposed to both generalist and specialist natural enemies. In this study, the ERH was explored using the invasive alien species, Arundo donax and two native tall-statured grasses, the cosmopolitan Phragmites australis and African endemic Phragmites mauritianus in South Africa. It was predicted that A. donax would have reduced species richness of herbivores compared with the native Phragmites spp., that it would be devoid of specialist herbivores and would thus be experiencing enemy escape in the adventive range. The herbivore assemblages were determined from both field surveys and a literature review. The assumptions of the ERH were for the most part not met; 13 herbivores were found on A. donax compared with 17 on P. australis and 20 on P. mauritianus. Arundo donax had two specialist herbivores from its native range, and shared native herbivores with Phragmites spp. Although A. donax had reduced species richness and diversity compared with that found in the native distribution, it has partially re-acquired a herbivore assemblage which is similar to that found on analogous native species. This suggests that enemy release may not fully explain the invasive success of A. donax in South Africa.

Herbivory-induced mortality increases with radial growth in an invasive riparian phreatophyte.

BACKGROUND AND AIMS: Under equal conditions, plants that allocate a larger proportion of resources to growth must do so at the expense of investing fewer resources to storage. The critical balance between growth and storage leads to the hypothesis that in high-resource environments, plants that express high growth rates are more susceptible to episodic disturbance than plants that express lower growth rates. METHODS: This hypothesis was tested by measuring the radial growth, basal area increment (BAI) and carbon isotope ratios (δ(13)C) in tree-ring α-cellulose of 62 mature tamarisk trees (Tamarix spp.) occurring at three sites in the western USA (n = 31 live and 31 killed trees across all sites, respectively). All of the trees had been subjected to periods of complete foliage loss by episodic herbivory over three or more consecutive growing seasons by the tamarisk leaf beetle (Diorhabda carinulata), resulting in approx. 50 % mortality at each site. KEY RESULTS: Mean annual BAI (measured from annual ring widths) in the 10 years prior to the onset of herbivory was on average 45 % higher in killed trees compared with live trees (P < 0·0001). Killed trees that had higher growth rates also expressed higher (less negative) δ(13)C ratios compared with live trees. In fact, at one site near Moab, UT, the mean annual BAI was 100 % higher in killed trees despite having about a 0·5 ‰ higher δ(13)C relative to live trees (P = 0·0008). Patterns of δ(13)C suggest that the intrinsic water-use efficiency was higher in killed than surviving trees, possibly as a consequence of lower whole-canopy stomatal conductance relative to live trees. CONCLUSIONS: The results show that a likely trade-off occurs between radial growth and survival from foliage herbivory in Tamarix spp. that currently dominates riparian areas throughout the western USA and northern Mexico. Thus, herbivory by D. carinulata may reduce the overall net primary productivity of surviving Tamarix trees and may result in a reduction in genetic variability in this dominant invasive tree species if these allocation patterns are adaptive.