Saliva has been collected from conscious rabbits in response to feeding pellets and carrots and amylase concentrations and flow-rates were measured. Saliva was collected from a polythene tube permanently inserted into the main duct, which, in most cases, kept patent for about a fortnight without any obvious decrease in the secretory capacity of the glands. 2. With pellets the flow-rate varied between 50 and 1250 microliter./min, and the corresponding amylase concentration was relatively constant around 250 units/ml. When carrots were fed, the flow-rates were about threefold lower, but the amylase concentration was raised to a mean value around 1000 units/ml. In spite of the differences in flow-rates, the two kinds of food promoted the same maximum output of amylase, and the output for both kinds of food promoted the same maximum output of amylase, and the output for both kinds of food was found to increase with the flow-rate. 3. The amylase concentration in the saliva decreased after pre- or post-ganglionic sympathetic denervation, reducing the output of amylase by about 50%. However, the amylase concentration was further lowered by beta-adrenoceptor block, which decreased the output by an additional 25%, suggesting that circulating catecholamines contributed to the secretion of amylase. 4. The fluid secretion in response to pellets and carrots was mainly dependent on parasympathetic activity, and for both kinds of food the range of flow-rates was unaltered by sympathectomy or beta-adrenoceptor block. However, at flow-rates below 50 microliter./min, where 25% of all samples with carrots were obtained, sympathetic activity may have contributed significantly to the fluid secretion. 5. In experiments on anaesthetized rabbits, frequency-response curves for amylase and fluid secretion in response to parasympathetic and sympathetic activation were obtained. A comparison between these observations and those obtained in the conscious animals during feeding suggests a parasympathetic activity mainly between 1 and 5 Hz and a sympathetic around 1 Hz. 6. It is concluded that both parasympathetic and sympathetic secretory nerves are reflexly activated during feeding, and that the normal secretion during a meal is dependent on an interplay between the nerves. The results suggest that at least two different afferent nervous pathways are involved in the control of the secretory nerves.