This study examines the effects of TMS of the right PPC on the latency of saccades and vergence alone or combined and the role of experimental design. Two designs were used: pure blocks with exclusively no-TMS or TMS trials; mixed blocks in which no-TMS and TMS trials were interleaved; a control study with TMS of the primary motor cortex (pure blocks) was also conducted and showed no effects on latencies. In contrast, in the experiment with TMS of the PPC latencies for TMS trials increased relative to no-TMS trials for almost all eye movements (isolated saccades, convergence, divergence, and for saccade and divergence components of combined eye movements). However, such increase was significant for pure blocks only. In mixed blocks no difference between TMS and no-TMS was found mainly because the latency of no-TMS trials increased relative to corresponding latencies in pure blocks. A second study centered on isolated convergence and divergence confirmed the interaction between block-design and TMS effects, and showed significant TMS/no-TMS differences only for the pure design and for a design in which the rate of TMS trials was high (75%). Again, the absence of difference was due to increase of latency for no-TMS trials in mixed blocks with low rates of TMS trials (50% or 25%), but also to decreased effects for the TMS trials themselves. We conclude that latency of all eye movements, saccades and vergence is highly influenced by the context. Such a contextual factor is the number of TMS versus no-TMS trials within a block; low numbers of TMS trials (50% or less) increases baseline latencies. The design of mixed blocks with 50% or less of TMS trials should not be recommended as it underestimates the direct effects of TMS on cortical processing. In fact, the majority of TMS studies on eye movements do use paradigms with high rates of TMS trials (75% or more). Our study confirms the validity of such paradigms.