Indirect effects in an experimentally staged invasion by a major predator

We used a system of neighboring small islands with and without the lizard A nolis sagrei to stage a 7-yr experimental study of the effects of an invadi ng species. Lizard propagules were introduced to four islands, randomly sel ected from eight that lacked lizards naturally. Four other islands where li zards occurred naturally constituted a third treatment. We examined possibl e lizard indirect effects on the commonest shrub (via its arthropod herbivo res), two sizes of aerial arthropods, and hymenopteran parasitoids. Lizards reduced leaf damage; moreover, during the middle years of the experiment, damage was less on introduction islands than on islands having lizards natu rally. Lizards increased the number of small aerial arthropods but had no e ffect on large aerial arthropods; the full effect on small arthropods was o nly evident near the end of the experiment. Lizard introduction increased t he number of hymenopteran parasitoids at the end of the experiment; however , only islands having lizards naturally had significantly more parasitoids than no-lizard islands summed over the long term. In contrast to the indire ct effects, the mostly direct effect of lizards on spiders was Very strong; introduction transformed spider density to that on natural lizard islands relatively rapidly and monotonically. In addition to demonstrating how an,i ntroduced vertebrate predator can affect food webs even down to the produce r level, this study illustrates the greater strength and regularity of dire ct as opposed to indirect effects.