Summary
The effect of phytophagous beetles on plant succession in an old—field community in southeastern Ontario was investigated over a 5—yr period. An experimental comparison of insecticide treated and untreated plots indicates that grazing by Trirhabda spp. on the goldenrod Solidago canadensis resulted in a reduction in precent cover of this species (from 40—70% to °1%), and a significant increase in percent cover of a number of earlier stage successional species.
Methodology
The study site is an abandoned hayfield which is 1-2 ha in size, surrounded by second-growth deciduous woodland, and located =50 km north of Kingston, Otario. The amount of Solidago canadensis present in-dicated that the field had been last cut at least 4-5 yr previously. Six 15 x 6.7 m experimental plots were staked out in the field in 1975. The plots were located so that road or woodland edge effects were minimized and vegetation and soil similarity between plots were maximized.
From 1976 to 1980, vegetation in the plots was sam-pled in June or July and again in late August or early September, using a technique discussed by Phillips (1959). In each plot, a 50 x 50 em square was random-ly placed five times, and an estimate of the percent ground cover of each species present in the quadrat was recorded. Plant species which were present in the plot but not found in any of the quadrats were also recorded. Cover estimation was used mainly because it is a nondestructive sampling method rapid enough to be used frequently for large sample sizes. It is a generally accepted method (Kershaw 1964, Ohmann eta!. 1980) despite its obvious inaccuracy.
Beginning in 1976, the insect populations in the study site were sampled approximately every 2 wk from mid-May to late August or early September by walking methodically through each plot and making 50 sweeps through the vegetation with a net measuring 30 em in diameter. All sampling was done on dry, bright, usu-ally sunny days, between 1000 and 1500. Insect samples were stored in 70% ethanol. The largest individual collection of larvae in any year was used to compare the Trirhabda population in the research plots from year to year. This was the most reliable measure, first because the larvae are less mobile than the adults, and second because they consume a much greater proportion of the goldenrod foliage.