Haddock, cod and whiting were fed on a number of food organisms and the rateof elimination of these from the stomachs was investigated by three methods. These were: (a) the “single meal” technique, in which pre-starved fish were given a single meal and then killed and examined after a given interval of time, (b) the “regular feeding” technique. For this method, fish were given constant meals at regular intervals for a period of about 5 days. By then, it was assumed that the stomach content weight would be fluctuating about some equilibrium value, and that this value could be estimated directly by killing the fish at the middle of a feeding interval, (c) for one food species (Ophiopholis), the rate of decline was observed directly by catching haddock that were known to contain recently eaten Ophiopholis and killing them after they had been kept in deck tanks for various periods. Consistently different results were obtained from the first two methods, the single meal technique giving lower rates of digestion than the regular feeding technique. This may have been due to the effect of starving the fish for several days prior to its receiving the experimental meal in the case of the “single meal” experiments. It was noted also that method (c) was not strictly comparable with either of methods (a) and (b) but was included as it provided the only data available for Ophiopholis. It was concluded that the history of the fish immediately prior to each experiment was important and should be described in all digestion experiments. The main food species used was saithe (Pollachius virens), but experiments were also conducted using Centronotus, sprats, three-bearded rockling, sandeels, Nephrops, Carcinus, Nereis, Nephthys, shell-less Mytilus and Ophiopholis. A comparison of rates of elimination of comparable meals by groups of haddock of different sizes showed that the rate of elimination was dependent, amongst other factors, on the size of the fish. The relationship was found to be where r is the rate of elimination from the stomach in g/unit time, L is the length of the fish (cm) and W is the weight of the fish (g). The relationship r ∝ L1·4 has been used to adjust all experimental results to relate to an arbitrary standard fish length of 40 cm, using a correction factor C = (40/L)1·4 = 175 L−1·4. Investigation of the effect of temperature showed that rate of elimination was relatedto temperature by the equation where To is the observed temperature. Various ways of analysing the effect of meal size on the rate of elimination were examined. The most consistent result was obtained by plotting the logarithms of the rates of elimination (g/h) against the logarithms of the mean stomach content weights, which gave a linear relationship with a slope of 0·46 for each of the main food species (saithe, Crangon, Nereis and Nephthys) using either the single or the regular feeding methods and for experiments conducted at both 6°C and 12°C. This result can be summarised by the equation where x is the weight of food in the stomach and Q' is a coefficient equivalent to the rate of elimination when the stomach content weight is 1 g. This can be related to a standard fish 40 cm in length at an arbitrary temperature Te, by rewriting in the form where L is the length of fish in cm and Q is the rate of elimination of 1 g of food from the stomach of a 40 cm fish at an arbitrary temperature (Te). Using this result, a theoretical equation was derived where M is the initial size of a meal and x is the weight of the meal after t hours in the stomach. This equation states that values of M0.54 − x0.54, adjusted for a standard fish length of 40 cm, when plotted against time in the stomach, should fall on a straight line with a slope equal to 0.54 Qt100.035(To − Te). Observation confirmed that this equation could be applied satisfactorily to the experimental results. Estimates of Q for a range of experimental results are also given. Estimates were also made of the time that elapses, after a meal has been ingested, before it actually begins to decrease in weight. These tended to be longer for crustaceans than for fish and appeared to be temperature dependent. It also appeared probable that these times were longer for large crustaceans than for small ones. Comparisons were made of the rates of elimination of food in haddock, cod and whiting stomachs. It appeared that under comparable conditions of fish size, meal size, temperature and experimental method, the rates of elimination of food in these three species were effectively the same.