Abstract. BACKGROUND:Isokinetic eccentric cycling is increasingly being utilised to examine the effect of chronic eccentric muscle training however little is known about how individuals familiarise to such a unique training modality. OBJECTIVE: To examine longitudinal variation in power output and lower limb muscle activation during familiarisation to maximal recumbent isokinetic eccentric cycling. METHODS: Twelve male volunteers, unfamiliar with eccentric cycling, completed four trials, separated by 7-10 days, each comprising 6 × 10 s maximal isokinetic eccentric efforts between 20-120 rpm. Peak power and average power output (PO), and surface electromyography (sEMG) of the rectus femoris (RF), vastus lateralis (VL), biceps femoris (BF), and medial gastrocnemius (MG) were recorded throughout. Systematic error across repeated trials was assessed using one-way ANOVA, and random error quantified using coefficient of variation (CV, %). RESULTS: Average PO at 60 rpm and RF activation at 20 rpm increased from trial 1-2 (p < 0.05), with no other systematic error between trials at any cadence. Across all cadences, the CV for peak PO (∼13%), average PO (∼10%), VL activation (∼13%) and RF activation (∼19%) was moderate and plateaued after one familiarisation (i.e. T2-T3). However, for BF (∼24%) and MG (∼22%) activation reliability was generally poor. For the majority of variables the reliability was best at 60 rpm. CONCLUSIONS: Therefore, with one familiarisation, 60 rpm is recommended to achieve moderate between-session reliability in the measurement of power output and lower limb muscle activation during recumbent, eccentric cycling.