Variability in yields of tar, nicotine and CO as a function of smoking regime was assessed by a re-examination of previous data from 115 brands of Canadian cigarettes smoked under 5 regimes. Here average standard deviations for tar yields increased from about 1 to 3.4 mg/cig when the smoking regime was changed from ISO to Canadian 'intense' (puff volume, 55 mL; interval, 30 s; duration, 2 s). This increase was due largely to 1) the necessity for a decrease in the number of cigarettes smoked per pad to avoid pad break through and 2) a 10-fold increase in the variability of water determinations which has other consequences. Under 'intense' smoking conditions, it was found that the percentage of water in TPM collected using a linear 20 port smoking machine was much greater than that determined using a rotary machine. This could account for the reported loss of repeatability of the 20 port linear under 'intense' smoking conditions. With respect to other analytes, an examination of yields under 'intense' conditions clearly demonstrated that the Canadian 'intense' regime does not "approximate the maximum exposure level to which an ordinary smoker could reasonably be expected to be subject". For example, the concentration of 4-aminobiphenyl in mainstream smoke decreased by about 25% when conditions were changed from ISO to 'intense' while some compounds remained constant and others increased. This is also true of biological response where mutagenicity decreases but cytotoxicity increases under the Canadian 'intense' regime. This type of observation may be useful since compounds whose concentration tracks biological response are clearly candidates for further study.