During mass rearing, adaptation of biological control agents to the rearing environment is a potential problem. Using the parasitoid wasp Trichogramma pretiosum, the performance of 26 highly inbred lines, five composite ‘populations’ (created from the inbred lines) and one insectary-reared population was compared using fertility life tables. Of the composite populations, three were created with maximal and identical genetic variation as a mixture of all 26 inbred lines, but these were then reared for a different number of generations (2, 6 or 17) before their performance was measured. The remaining two composite populations were created based on the performance of the individual inbred lines: one was a combination of two inbred lines with a high intrinsic rate of natural increase (rm), ‘high rm’; and the other was a combination of two lines with a ‘low rm’. High and low rm populations were reared for two generations prior to testing. Parameters measured were fertility, longevity and sex ratio. We found no difference between the maximally variable population reared for two generations and the ‘high rm’ population (rm = 0.285 and 0.282, respectively). ‘Low rm’ was the population with the lower performance (rm = 0.255). Genetically variable population reared for two generations for 48 h produced significantly more offspring than the populations reared for 6 and 17 generations. Hybrid population derived from the high-rm lines did significantly better than that derived from the low-rm lines. Low-performance populations become more male based than high performance at 48 h. The potential benefits to improve population's performance using inbred lines for mass rearing are discussed.