SINCE Robert Hooke1 suggested that the siphuncle (which he took to be the gut) of Nautilus produced gas to force liquid from the shell chambers and so buoy up the animal, there have been many speculations on its function. These speculations required that the siphuncular tube was watertight2, or air-tight3–5, or permeable to gas6–8, or permeable to liquid9,10 or permeable to both gas and liquid11. The only experimental information, however, on the transmission properties of the siphuncular tube is that it is permeable to water10. The experiments described here show that the wall of the siphuncular tube is permeable to sea water (with a linear relationship between flow rate and pressure), impermeable to air and able to withstand pressures equivalent to well above 450 m of water.