In general, studies on plant phenotypic plasticity concentrate on plant responses to different levels of a single environmental factor. Under natural conditions, however, multiple environmental factors often vary simultaneously. I studied the consequences for lifetime fitness caused by single treatments or treatment combinations by investigating patterns of phenotypic plasticity within and between generations. The parental plants (three genotypes of the annual plant Arabidopsis thaliana) received zero, one or two stress treatments at an early life‐stage. The treatments included wounding, shading, chilling, and their pairwise combinations. In the second generation, offspring of treated plants received either the parental or no treatment. Offspring of non‐treated plants were reared under all treatment conditions. Plants responded strongly to the treatments, especially through delayed reproduction, which positively affected lifetime fitness. Notably, treatment combinations triggered stronger plastic responses on average. Because the delay in reproduction was offset by a delay in senescence, the treatments resulted in a fitness gain instead of a loss. However, under adverse environmental conditions, this delay represents a potential fitness cost, especially when the time for reproduction is limited. The treatments ‘wounding’ and ‘shading’ triggered parental effects that increased fitness only in plants that themselves received the treatment. Untreated offspring of wounded or shaded parents performed like control plants. Also, these parental effects were not accompanied by potential fitness costs, such as delayed reproduction. Chilling triggered genotype‐specific parental effects that increased or reduced fitness. Of the treatment combinations only ‘wounding’ and ‘shading’ resulted in genotype‐specific parental effects that increased or reduced fitness independently of offspring treatment. These results suggest that the response of annual plants to treatment combinations triggers predominantly within‐generation plastic responses that include potential fitness costs, which cannot be inferred from studies that manipulate environmental factors individually. Therefore, single treatment studies likely underestimate the costs of plasticity in natural environments.