Prostaglandins are released when the sympathetic nerves supplying the rat epididymal fat pad in vitro are stimulated (Shaw, 1966). It is not yet known whether this is a general phenomenon associated with adrenergic nerve stimulation but the experiments described in this paper show that prostaglandin E2 is released when the splenic nerve is stimulated. A preliminary account of this work has already been published (Davies, Horton & Withrington, 1967).In the investigation reported here we collected venous blood samples from the perfused dog's spleen in order to detect any prostaglandin release when the splenic nerve was stimulated. A large increase in prostaglandin E2 output was found. METHODSBlood-perfused dog spleen. An isolated spleen was perfused with blood from the femoral artery of a second (donor) dog using the technique described in the following paper (Davies & Withrington, 1968). The splenic venous blood was returned to the femoral vein of the donor except during the collection of a sample, when it was diverted out of a side arm. During the collection of a sample, blood from a reservoir was pumped into the donor. Electrodes were placed on the splenic nerve and the spleen was bathed in liquid paraffin at 370 C. The nerve was stimulated with 50 V pulses of 0.5 msec duration at 10/sec. Extraction procedures. The procedures are outlined in Fig. 1

Summary The responses of the smooth muscle of the capsule and blood vessels of the isolated, perfused human spleen to sympathetic nerve stimulation, adrenaline, noradrenaline, angiotensin, oxytocin, vasopressin, isoprenaline and acetylcholine have been investigated and compared with those of dog spleen. Stimulation of the postganglionic sympathetic nerves to the human spleen at frequencies of 3–10 Hz evoked graded vasoconstriction but very small changes in spleen volume. The injection of adrenaline and noradrenaline in doses of 0·25–25 μg to the human spleen produced graded increases in splenic vascular resistance with very small decreases in spleen volume. Administration of the α‐adrenoceptor blocking drug phenoxybenzamine completely abolished or considerably reduced the vascular responses of the human spleen to sympathetic nerve stimulation or the injection of noradrenaline. The vascular action of adrenaline was often reversed to elicit a vasodilatation after phenoxybenzamine suggesting the presence of β‐adrenoceptors in the vascular bed. This was confirmed by the administration of isoprenaline which induced a marked reduction in vascular resistance of the human spleen. The polypeptides angiotensin and vasopressin induced a marked vasoconstriction in the human spleen without changes in the spleen volume. These effects were uninfluenced by the administration of phenoxybenzamine. The polypeptide oxytocin caused a slight vasodilatation in the human spleen, an effect almost exactly mimicked by the preservative chlorobutanol. Preliminary experiments suggest that noradrenaline is the transmitter released by the postganglionic nerves to the human spleen. These results provide direct evidence that the normal human spleen, unlike that of the dog, does not have a reservoir function. It is suggested that contractions of the enlarged human spleen may occur in various pathological conditions.

1 The sympathetically-innervated arterial vascular bed of the dog's liver was perfused from a femoral artery. Arterial blood flow and perfusion pressure were measured continuously, and the hepatic arterial vascular resistance calculated. The preparation provided a means of assessing hepatic arterial vasodilatation quantitatively. 2 Isoprenaline, histamine, prostaglandin E2, glucagon and secretin were injected intra-arterially and all evoked dose-dependent vasodilatation of the hepatic arterial vascular bed. 3 The maximum reduction in the calculated hepatic arterial vascular resistance of 37-38% was the same for each of the five substances. 4 Comparisons on a weight basis revealed that prostaglandin E2 was the most potent, followed in potency order by secretin, isoprenaline, histamine and glucagon. 5Comparisons on a molar basis showed that secretin and prostaglandin E2 were intrinsically considerably more potent than isoprenaline, histamine or glucagon. 6 The onset of the vasodilatator responses to secretin, isoprenaline, histamine and prostaglandin E2 was rapid, and the duration of their actions was brief. 7 The onset of the vasodilator effects of glucagon was slow and its duration of action very prolonged. 8 The implications of these observations with respect to the physiological control of the hepatic arterial vascular bed of the dog are discussed.

I The hepatic artery of the anaesthetized dog was cannulated and perfused from a femoral artery, the blood flow and perfusion pressure being monitored continuously. The sympathetic periarterial nerves were divided. 2 Dose-dependent increases in hepatic arterial vascular resistance (HAVR) resulted from intraarterial injections of noradrenaline, angiotensin and vasopressin. 3 Single injections of glucagon (100 gg, i.a.) caused a transient significant fall in HAVR of 19.9 + 3.2%, and infusions of 25 tg/min of glucagon intra-arterially caused maintained reductions in HAVR of 16.9 + 4.2%. 4 After single injections of 100 gg glucagon intra-arterially the vasoconstrictor responses to noradrenaline, angiotensin, and vasopressin were reduced by about 85-95%. Recovery occurred in 8-10 minutes. 5 Intra-arterial infusions of glucagon, 2.5-50.0,g/min, reduced the effects of test doses of noradrenaline, angiotensin and vasopressin throughout the period of the infusions. 6 Dose-response curves to the constrictor agents were constructed before, during and after intraarterial infusions of 25 gg/min of glucagon. Glucagon caused a parallel shift of the curves for noradrenaline and angiotensin to the right, with no suppression of the maximum response. 7 Infusions of glucagon shifted the dose-response curve for vasopressin to the right, but, in contrast to noradrenaline and angiotensin, the shift was nonparallel and there was a suppression of the maximum response by about one-half. 8 A large dose of insulin, 10 iu, transiently reduced HAVR and caused a weak and very transient inhibition of the effect of test doses of noradrenaline. The characteristics of these effects were quite different from those of glucagon. 9 It is possible that the antagonism by glucagon of the vasoconstrictor responses of the hepatic arterial vasculature may be important in protecting this vascular bed from the effects of concomitantly released vasoconstrictor agents.

5. Intra-arterial injections of noradrenaline (0 1-50 jug) caused biphasic changes in hepatic arterial vascular resistance, and a rise in hepatic portal vascular resistance. Both hepatic vascular effects had a significantly shorter latency than any succeeding systemic cardiovascular effects.6. Intraportal injections of noradrenaline (0.1-50 ug) caused hepatic portal vasoconstriction, and a biphasic change in the hepatic arterial resistance. Both of these effects had a significantly shorter latency than any succeeding systemic effects.7. Intra-arterial injections of isoprenaline (041-10 jg) caused dose-dependent hepatic arterial vasodilatation but little change in portal vascular resistance. Intraportal isoprenaline caused little change in portal resistance but elicited dosedependent hepatic arterial vasodilatation.8. The time courses of the responses to intra-arterial and intraportal noradrenaline and isoprenaline indicate that the responses of the liver vascular bed which does not receive the direct injection were not due to recirculation ofthe vasoactive material.9. It is postulated that vasoactive material injected into one inflow circuit of the liver elicits changes in the vascular resistance of the other inflow circuit by an intrahepatic effect.