Volume and composition of the body fluid compartments are kept within narrow limits. This is done by changes in intake and excretion of water and salt. I found that changes in food intake and salt excretion play an important role in body fluid regulation during water deprivation and investigated the mechanisms of these changes. Urine volume changed little in rats deprived of water for 10 hours. The main osmoregulatory response during water deprivation was a loss of solute from the body. Food intake fell by 43%, which reduced the load of solute to the tissues and allowed absorption of water already in the gut. Water deprived rats also excreted more sodium, potassium and chloride. Thus, although rats lost 8% of their body water during 10 hours of water deprivation, plasma tonicity rose by only 2%. On rehydration, when no food was present rats rapidly reduced excretion of sodium, potassium and chloride. Rats allowed water and food drank more and excreted more solute. These changes contribute to restoration of the body fluid compartments. Brain infusions caused changes in electrolyte excretion that were similar to those seen during water deprivation and rehydration. Electrolyte excretion increased during infusion (1 $\mu$L/min for 2 h) of cerebrospinal fluid (CSF) with 300 mM NaCl in the lateral ventricle. Infusion of low-sodium CSF reduced electrolyte excretion in water deprived rats, but had little effect in rats that were not water deprived. The time course of the changes after rehydration and brain infusions was the same, and the same solutes were involved. This suggests the mechanisms are similar. Low-sodium CSF made isotonic with mannitol had the same effect as hypotonic low-sodium CSF. The sensor probably monitors brain interstitial fluid, not CSF. Food intake fell within 1 hour of water deprivation. Meals were smaller, but meal frequency did not change. Dehydration anorexia is caused by a sensor located in the gut, portal circulation or liver, because infusion of water in the stomach, jejunum or cecum (10 mL/6 h) restored food intake in rats not allowed to drink, but intravenous infusions had no effect. Intravenous infusions did not alter urinary water loss and did not alter food intake in rats allowed to drink.

Mass, solute (chloride), isotope (deuterium) and thermal balances were completed at Perry Lakes, two semi-permanent 'water table' lakes near Perth, Western Australia. All balance components except groundwater discharge/recharge were measured independently. These difficult to measure groundwater components of lake-aquifer interaction were estimated by integrating mass, solute and chloride data in sequential 4 day balances spanning two years. Before urbanisation, such wetlands functioned predominantly as flow-through lakes. Now, large winter storm water inputs (and summer artificial level maintenance pumped locally from groundwater) dominate. In East Lake these inputs together comprise 42% of the annual water budget; groundwater discharge is reduced to just 2%. Even under flow-through conditions, these 'non natural' inputs are so large East Lake always tends towards a recharge state and commonly becomes a local groundwater mound. Flow-through is established in both lakes over winter. Initially each lake functions separately however as winter progresses shared capture and release zones are established. Maintenance of lake levels in early summer forces East Lake back to recharge status. Sediment heat flux (Qse) is significant in these very shallow lakes. Over summer Qse was negative, with a net movement of heat from the water into the sediments which act as a seasonal heat sink. In winter Qse was positive and stored summer heat was returned to the water column. This flux at times exceeded 40 W m-2. Evaporation was determined independently by floating pan, leaving Qse as the thermal balance residual. Ignoring Qse, annual evaporation determined by thermal balance was over estimated by 7%. Over and under estimates of individual 12 day balance period evaporation exceeded 50%. Monthly Class A (Perth airport) pan coefficients varied from 0.54 (January) to 0.86 (September). Ten empirical equations for evaporation were calibrated and compared with the East Lake floating pan. Best performer was the Makkink which tracked the floating pan closely throughout all seasons. Poorest were the Penman, DeBruin-Keijman, Priestly-Taylor and Brutsaert-Stricker which grossly over estimated late winter evaporation. Transpiration from Typha orientalis, estimated using hydrograph techniques was 43% of open water evaporation in summer and 28% annually. Temperature controlled evaporation pans (tracking lake temperature) experimentally determined the local deuterium content of lake evaporate ƒÔE, required for isotopic balances. Techniques employing pans evaporated to dryness and pans evaporated at constant volume were run in tandem continuously for two years. This study singularly integrates mass, solute and isotope balances thereby allowing groundwater components to be accurately quantified. The isotope balances are unique, being the only such balances incorporating experimentally derived local deuterium values of lake evaporate. This study represents the only thermal balance, the only accurate determination of pan-lake coefficients and the first calibration of commonly used empirical evaporation equations for Swan Coastal Plain wetlands. Groundwater levels in the western suburbs of Perth have declined over 40 years and a disproportionate larger decline now seriously threatens Perry Lakes. Modelling suggests regional groundwater extraction exceeds recharge. Wetland managers can no longer maintain East Lake via local groundwater extraction. Artificial recharge using imported surface and waste water are possible future management options.