A common outcome of population models incorporating trophic interactions is that population behavior is transient; however, due to the rarity of long-term records of population abundance, the prevalence of transient behavior in natural systems is unclear.  We investigate changes in the periodicity of gypsy moth populations through time by applying wavelet analysis to 85-yr time series (the longest ever used in a formal analysis of gypsy-moth periodicity) of forest defoliation in each of 5 states in the northeastern United States.  We also reassess the relative strength of the 4-5 and 8-10 year cycles of gypsy moth populations in light of a recent correction of a long-standing bias in the computation of wavelet power coefficients.

Results/Conclusions

Our results suggest that gypsy moth population behavior shifted through time on up to four occasions during the twentieth century; strongly periodic outbreaks were observed between ca. 1940-1960 and ca. 1975-1995, whereas aperiodic dynamics were observed in the intervening years.  Despite correcting for the inherent bias of wavelet analysis for producing higher wavelet power coefficients with increasing period length, our results support earlier studies concluding that the gypsy moth’s 8-10 year cycle is generally dominant to the 4-5 year cycle.