Recent efforts in controlling mosquito-borne diseases focus on biocontrol strategies that incapacitate pathogens inside mosquitoes by altering the mosquito’s 
microbiome. A case in point is the introduction of Wolbachia into natural mosquito 
populations in order to eliminate Dengue virus. However, whether this strategy can 
successfully control vector-borne diseases is debated; particularly, how artificial infection affects population dynamics of hosts remains unclear. Here, we show that natural Wolbachia infections are associated with unstable mosquito population 
dynamics by contrasting Wolbachia-infected versus uninfected cage populations of 
the Asian tiger mosquito (Aedes albopictus). By analyzing weekly data of adult 
mosquito abundances, we found that the variability of the infected populations is 
significantly higher than that of the uninfected. The elevated population variability is 
explained by increased instability in dynamics, as quantified by system nonlinearity (i.e., state-dependence). In addition, predictability of infected populations is substantially lower. A mathematical model analysis suggests that Wolbachia may alter mosquito population dynamics by modifying larval competition of hosts. These
results encourage examination for effects of artificial Wolbachia establishment on mosquito populations, because an enhancement of population variability with reduced predictability could pose challenges in management.