Nine established cigarette smokers were each studied four times, smoking two identical cigarettes on each occasion. After an acclimatization study, they smoked one of three types of cigarettes, either their usual brand or one of two types of special low tar cigarettes. These latter both had tar yields of about 8 mg with nicotine yields of 0.55 (LN) and 0.90 (MN) mg respectively. The test order was randomized between individuals and before using the special cigarettes the subjects were given a pack to accustomize themselves. While smoking each cigarette, magnetic tape recordings were made of puff rate, ventilation measured by respiratory inductive plethysmography (RIP) and nasal airflow measured with a modified oxygen cannula. The data were then processed digitally off-line. Salivary nicotine and alveolar carbon monoxide levels were measured before and after smoking each cigarette, and the cigarette butt was analysed for nicotine. While smoking behaviour varied considerably between the various subjects only small differences were seen between the different cigarette types in puff volume and duration and shape of the puff profile. Some changes in smoking behaviour occurred during the course of smoking a single cigarette. Ventilatory patterns showed consistent inter-subject differences but there were no apparent variations due to the various cigarettes. Most subjects puffed during an expiration with the buccal cavity closed off, and then took a slower, deeper inspiration breathing through the mouth. Others, however, took puffs at any point in the respiratory cycle. The different nicotine yields of the cigarettes produced marked changes in the butt and salivary nicotine measurements, but neither these, nor the changes in alveolar carbon monoxide, were closely related to ventilatory measurements. Possible explanations for these discrepancies are discussed.