Genetic variation in resistance to 16 species of herbivorous insects was studied in 18 clones of Solidago altissima growing in an old field near Ithaca, New York, USA. Resistance to each insect, defined as the abundance of a species attacking a particular host genotype relative to other genotypes, was measured in both the natural stand and in two experimental gardens. The heritability of resistance was estimated by parent-offspring regression and sibcorrelation. The primary result was that clones differed in resistance to 15 of 16 insect species. The resistance of genotypes to these insect species remained relatively constant over the four years of the study. However, for only 10 of these resistances were the heritability estimates significantly different from zero. Thus the common assumption of plant-insect studies - that phenotypic variation in insect abundance is closely correlated with underlying genetic variation - is only conditionally true. There is heritable variation in resistance to many insects, but not all. The insects for which we observed heritable variation in plant resistance represent five different orders and several functional groups, including leaf chewers, phloem and xylem feeders, and gall formers. There was no apparent pattern between the degree of heritability of plant resistance and the destructiveness, feeding method, breadth of host range, or taxonomic group of the insects. The lack of marked heritable variation in resistance to some insects may be the result of (a) reduced variation caused by strong selection during prolonged or repeated insect outbreaks, and (b) genotype-environment interactions that obscure differences among genotypes.

Most plant species are attacked by insects with widely different feeding styles and life histories. As a first step toward understanding how plants adapt to such diverse pressures, we investigated the population responses of 17 insect species to the genotypic variation exhibited by clones of goldenrods, Solidago altissima, that were drawn from the same population and grown in common gardens. A cluster analysis on the genetic correlations between the insects' responses produced four groups, termed "herbivore suites," that are attracted and repelled by similar plant genotypes (and their underlying sets of traits). There was no obvious correspondence of an insect's phylogenetic affinity, host range, or feeding style with its membership in a particular suite. Put another way, insects belonging to the same family or guild respond differently to genetically variable traits in populations of this native, perennial plant. We found no negative genetic correlations among suites and few among pairs of species, which suggests that trade—offs between the plant's resistance traits are relatively unimportant. The existence of suites of herbivores, whose members exhibit correlated patterns of abundance on different host genotypes, has several interesting consequences. First, the genetic variability of the plant population will be reflected in the diversity and composition of the insect community. Second, selection caused by one herbivore species can result in increased resistances to other, rarer species whose responses are correlated with those of the damaging species. Third, selection by suites of herbivores may be more intense and less variable than selection by individual insect species, especially when many species are rare. By recognizing that resistance can evolve as a function of selection by entire suites of species we can develop better estimates of selection and perhaps achieve a greater understanding of the process of diffuse coevolution. The linkage of interactions within suites suggests some unexplored possibilities for explaining the rarity of many insect species.

No abstract

We studied clone structure and degree of genotypic mixing of Solidago altissima L. (Asteraceae) clones in four old fields near Ithaca, New York. The fields differed in time from agricultural abandonment and were approximately 1, 5, 20, and 35 years old. In the three older fields, three 0.75 m2 plots were excavated intact and rhizome connections among ramets were mapped. In the youngest field 30 ramets were dug up singly. The genotype of all ramets was determined using electrophoresis of four polymorphic enzyme systems. Fields differed in the number and dispersion of genotypes within plots, and the degree of connection among ramets in the same clone. The one‐year‐old field was composed of single ramet genotypes which had probably established from seed the previous year. The five‐year‐old field contained many small contiguous clones of S. altissima with highly interconnected ramets. In the oldest two fields clones were highly intermixed and ramets of the same genotype were not extensively interconnected. These results demonstrate that clones of S. altissima display considerable phenotypic variability between fields and patterns of clone development may differ. The causes of this variability remain to be identified. We suggest that either selection for different genotypes or changing habitat conditions during succession may lead to changes in clone form.