Introduction: Cross-country ski racing includes continual variations in intensity due to terrain and tactics {Mygind et al 1994; Norman et al 1989). Consequently, the recovery between periods with higher intensity might affect the outcome of the race. Both blood metabolites and respiratory variables are used for standard performance evaluations for endurance athletes, although there is a lack information if respiratory variables respond similar to blood lactate and acid/base values during prolonged variable exercise. Therefore, the aims with the present study were to 1) evaluate whether respiratory variables are associated with blood lactate and acid/base variables, 2) how/if these variables might predict physical performance and 3) whether a calculated heart rate-oxygen uptake (HR-VO2) relationship is valid during variable intensity exercise. Methods: 12 cross-country skiers classified as elite (E, n=6) and formerly well-trained (FWT, n=6) performed two roller ski tests. 1) An incremental test to establish maximal oxygen uptake (VO2max), maximum heart rate (HRmax) and lactate threshold (LT). Submaximal and maximal VO2 and HR during the incremental test were used for calculating the individual HR-VO2 linear relationship and 2) a 48-min long variable intensity protocol (VIP) at alternating exercise intensities, 90% (HI90) and 70% (MI70) of VO2max. Cardio-respiratory variables and venous blood samples were continuously collected throughout the VIP. Comparisons between E and FWT were performed using a two-tailed unpaired Student’s t-test and a ANCOVA analysis was used to determine which physiological variable best could prognosticate time to exhaustion (TTE). A simple linear regression was used to establish the relationship between HR and VO2. Results: Blood lactate concentrations [La] were higher and base excess [BE] lower for FWT from the first MI70 (P<0.05). FWT had augmented RER during all HI90 and an elevated VE/VO2 during the second and third HI90 in comparison to E (P<0.05). The expected HR were higher during the MI70 exercise intensities regardless of group affiliation (P<0.05). The blood [La] response predicted time to exhaustion earlier than respiratory variables (P<0.05). Discussion: Blood lactate and acid/base fluctuations were not reflected by RER and the ventilatory equivalents. Furthermore, blood lactate is to prefer, in comparison to ventilatory variables, to study performance related recovery processes during endurance exercise with variations in intensity. The expected HR-VO2 relationship was not valid during VIP.