The purpose of this study was to evaluate whether greater and sustainable stroke volume (SV) responses may be obtained by exercise intensities corresponding to peak SV (SVpeak) vs. maximal O-2 consumption ((V)over dot O-2max), and short vs. long intervals (SI vs. LI). Nine moderate-to well-trained male athletes competing at regional level specialists of cyclist, track and field volunteered to take part in the study ((V)over dotO(2max): 59.7 +/- 7.4 mL . min(-1) . kg(-1)). Following familiarisation sessions, (V)over dotO(2max) was determined, and then SVpeak was evaluated using exercise intensities at 40%-100% of (V)over dotO(2max) by nitrous-oxide rebreathing (N2ORB) method. Then each separate participant exercised wattages corresponding to individual (V)over dotO(2max) and SVpeak during both SI (SI(V)over dotO(2max) and SISVpeak) and LI (LI(V)over dotO(2max) and LISVpeak) workouts on a cycle ergometer. Main results showed that both SI(V)over dotO(2max) and SISVpeak yielded greater SV responses than LI(V)over dotO(2max) and LISVpeak (p = 0.05). Mean SV responses were greater in LISVpeak than in LI(V)over dotO(2max) (p = 0.05), but there was no statistical difference between SISVpeak and SI(V)over dotO(2max). However, there was significantly less physiological stress based on VO2, respiratory exchange ratio, heart rate and rate of perceived exhaustion in (SVpeak) than in (V)over dotO(2max) intensities (p = 0.05). Moreover, SV responses at exercise phases increased in the early stages and remain stable until the end of SI(V)over dotO(2max) and SISVpeak workouts (p > 0.05), while they were gradually decreasing in LI(V)over dotO(2max) and LISVpeak sessions (p = 0.05). In conclusion, if the aim of a training session is to improve SVpeak with less physiological stress, SISVpeak seems a better alternative than other modalities tested in the present study.