1 We studied the regional differences both of reactivity to various stimulants and of neurogenic responses elicited by transmural stimulation in the longitudinal and circular muscles of the truncal portal vein, mesenteric vein, splenic vein and gastric vein of the dog portal tree. 2 Strong spontaneous activity appeared in the longitudinal muscle of the truncal portal vein (96% of preparations tested). Weak spontaneous activity sometimes appeared in the circular muscle of the truncal portal vein (41 %) and rarely in the longitudinal muscle of the mesenteric vein (12%). It did not appear in other segments. 3 The splenic vein and the gastric vein showed similar patterns in the relationship between resting tension and response to noradrenaline; that is, the responsiveness of either longitudinal or circular muscle of these two veins increased and then decreased almost parallel as resting tension increased and reached a maximum under the same resting tension. 4 The longitudinal muscle of both the truncal portal vein and the mesenteric vein was more responsive to noradrenaline, acetylcholine, histamine and KCI than the circular muscle; for example 2.02 and 1.44 times more responsive to noradrenaline, respectively. On the other hand, the longitudinal muscle of the splenic vein and the gastric vein responded less well than the circular muscle; for example 0.36 and 0.16 times as responsive to noradrenaline, respectively. 5 Acetylcholine and histamine caused marked contractions which were comparable to those elicited by noradrenaline in the longitudinal muscle of the truncal portal vein. Acetylcholine also elicited similar contractions in the longitudinal muscle of the mesenteric vein but the responses induced in preparations of other segments were small. 6 The longitudinal muscle of the truncal portal vein responded well even to low-frequency stimulation of 2 Hz, while the circular muscles of the truncal portal and splenic veins gave marked responses only to high-frequency stimulation of 10 or 20 Hz or more. These contractile responses were attenuated by phentolamine (10-6 M) or atropine (10-6 M). The longitudinal muscle of the splenic vein showed no significant response to stimuli of any frequency. 7 It seems that the portal tree receives not only adrenergic but also cholinergic innervation. In addition, the longitudinal muscle of the truncal portal vein may receive non-adrenergic, noncholinergic innervation as well. 8 The longitudinal muscle of the portal vein may be crucial to venous return in assisting movement of the blood it contains. If this is the case in man, then the regional differences in the portal tree demonstrated in this study may explain why varicose changes during portal hypertension occur preferentially in the oesophagogastric region and rarely in other regions, as blood stasis may occur more readily in the regions of the gastric and splenic veins where the longitudinal muscle is not very active.