The intriguing difference between far-infrared photoconductivity spectroscopy and absorption spectroscopy in the measurement of the magnetoplasmon frequency in GaAs quantum wells reported by Holland et al. [Phys. Rev. Lett. 93, 186804 (2004)] remains unexplained to date. This Letter provides a consistent mechanism to solve this puzzle. The mechanism is based on the electron reservoir model for the integer quantum Hall effect in graphene [Phys. Lett. A 376, 616 (2012)]. We predict sharp kinks to appear in the magnetic induction dependence of the magnetoplasmon frequency at very low temperatures such as 14 mK in the same GaAs quantum well sample used by Holland et al.