Improved methods of analyzing length-weight relations in the fast-growing salp Thaha dernocratica enable future calculation of productivity from population data and show that carbon density in relation to wet weight for sizes of salps commonly available as food is similar to that of Crustacea. A power relation was developed: Carbon Weight (@gC) = 1.62 X Length (r2 = 0.87, n = 39). Measurements were taken over a wider range of sizes than before (including 0.4 mm embryos). No significant difference between solitary and aggregate stages could be detected. Dry weight was 8.04 % of wet weight (SE = 1.01), ash-free dry weight was 46.5 O/o of dry weight (SE = 1.1), and carbon weight was 38.7 % of ash-free dry weight (SE = 2.8). Carbon was found to be a better reference measurement than either wet or dry weight. The C: N ratio was 3.81 : 1 (SE = 0.06). Tissue shrinkage after preservation was significantly different in buffered formalin than in Steedman's solution. Formulae are given for calculating live length from preserved length up to 2 yr after preservation. Previous estimates of wet weight, dry weight and ash in salps were improved by removlng internal seawater and salt and driving off most of the 'bound' water; these improvements are also applicable to other gelatinous organisms. Salp weight increases as length squared rather than cubed, because tissue density declines with growth between 3 and 10 mm. Salp tissue density is much closer to that of other marine invertebrates, such as molluscs and crustaceans, than was previously reported in the literature. Contrary to expectations, the ratio of dry and ash-free dry weight to wet weight is high during the most rapid growth period, from 2 to 7 mm. Whlle the average density of T. democratica in relahon to wet weight is now seen to be comparable to that of molluscs and chaetognaths (not, as previously thought, to that of medusae and siphonophores), the density of 4 mm individuals reaches that of the arthropods. Coupled with high growth rates, this result makes the salp likely to be a much more important source of food for fish than was previously realized.

Experiments were carried out in wind tunnels, in a test tank, and at sea, to investigate the interaction between the water column and a towed plankton net. Flow patterns about stationary nets held in a stream of air were observed and photographed. The velocity and turbulence of the stream near the net, and the drag upon the net, were measured. Experimental tows were made in the laboratory and in the field to test the effect of various factors on "filtration efficiency". This interaction between the discharge and the flow around the net leads to enhanced filtration. The effect is accentuated by a flared mouth which increases the velocity differential. The effect is reduced by encasing the net, which separates the interacting streams and probably causes the stream from the net surface to lose momentum. Filtration efficiency is independent of towing velocity, except for a gradual decline at velocities less than 1-2 kt. Filtration efficiency is also independent of filtering area in nets of usual length, but declines sharply in nets shorter than twice their mouth diameter. There is a zone of turbulence and reduced velocity in the centre of a net with bridles. At approximately one-quarter of a radius from the ring, the stream is flowing at mean mouth velocity. This seems to be the most suitable site for a flowmeter.

A newly developed vertical free fall net uses a weighted ring to propel it downward through the water at a standard speed. It is closed after a preselected time by a strangling rope secured aboard ship. This design eliminates the three‐point bridle and the long wire in front of a vertically hauled net which cause avoidance reactions in active zooplankton. The depth and volume filtered are unaffected by drift. The net does not require an oceanographic winch for its operation, only a warping drum for retrieval after closure, and therefore can also be used on unspecialized vessels. These properties recommend it is a standard survey sampler.

Three plankton nets, similar in all respects but size, were hauled vertically through the water column (100 m) at speeds ranging from 0.4 to 2.4 m/sec. Several components of the catch were counted, and the counts compared to test the effect of hauling speed. With increased hauling speed, the number of juvenile and adult euphausiids decreased (beyond 0.6-0.8 m/sec) ; the number of small organisms increased; and the condition of the catch deteriorated (beyond 1.5 m/sec). It is suggested that the mesh selectivity of a net depends in part upon the number of contacts between organism and mesh in transit from mouth to cod-end. This could explain why fewer small organisms were taken at low hauling speeds than at high hauling speeds.

The Australian and American models of the sampler were tested at water velocities up to 3.5 kt. A calibration tank designed to simulate normal flow conditions through the flowmeter was used to calibrate flowmeter rate against flow rate measured at a 90� V-notch exit. This calibration was used to determine the discharge through samplers towed for a measured distance in a water channel. Thus the filtration characteristics of the samplers were examined. The filtration coefficient for the American sampler without a net averaged 88 %, compared with 92% for the Australian sampler. Australian nets reduced the filtration coefficient of the Australian sampler by 2-15 %. The American net had little effect on the filtration coefficient of the American sampler. The filtration coefficient varied with mesh aperture and clogging but was largely independent of towing velocity. Other experiments suggested that net length was less important, and the flow of water around the net more important, than previously thought. Above approximately 1 kt, the calibration value was nearly independent of velocity and hardly altered with net attachment no matter how fine the mesh.