Two methods of estimating populations of Tabanid larvae in salt-marsh sod were compared on Long Island, New York, in 1954-56. The first was bringing the larvae to the surface of a cleared area at low tide by spraying it twice at an interval of five minutes with a 1: 19 dilution of the New Jersey larvicide. This is a concentrate consisting of 66 per cent. kerosene or similar petroleum distillate, 0.07 per cent. pyrethrins, 0.5 per cent. sodium lauryl sulfate and 33.5 per cent. water [cf. R.A.E., B 23 205]. It was used at 1 U.S. gal. diluted emulsion per sq. yd. About 87 per cent. of the larvae coming to the surface in 60 minutes under its influence did so in the first 40 minutes. Limited tests indicated that O.5 oz. 10 per cent. γ BHC (lindane) in 1 U.S. gal. water [cf. 45 116] was about a; effective as the New Jersey larvicide, and a suspension of a wettable powder of pyrethrum and sulfoxide [1, 2-methylenedioxy-4-(2-(octylsulphinyl)propyl)ben-zene] somewhat more effective. The second method was cutting foot-square samples of sod and drying them in a Berlese funnel, so that larvae driven out in the drying process fell into a jar of marsh water. This water was examined every clay or two days. The samples were 1 1/2 ins. thick, as experiments in which plugs 8 1/2 ins. thick were cut into slices and these were examined separately showed that virtually all the larvae recovered were from the top 1 1/2 ins., even in winter.