1. Infra-red Colour Thermography was used to visualize skin temperatures over seven neonates nursed in a modified temperature controlled incubator. 2. Temperature distributions were recorded on cine film which was analysed to evaluate mean skin temperatures; these were subsequently compared with values obtained from multiple, weighted measurements from a thermocouple thermometer. In all cases, there was agreement to within +/- 0 . 6 degrees C. 3. Thermograms in thermo-neutral conditions (approximately 32 degrees C) showed a temperature distribution with the warmest skin overlying the central hot 'core' and temperatures falling towards the extremities. Temperature patterns in a cooler environment (approximately 28 degrees C) showed features due to the conductive or thermogenic nature of the structures underlying the skin. 4. Multiple skin temperature recordings made at two environmental temperatures were obtained from a further twelve infants. 5. Linear regression of skin temperature against rectal--environmental difference, performed for each of the measured sites, showed that the upper arm and, to a lesser extent, the upper thigh temperatures represented the mean value most closely. The upper abdomen and head were both warmer and less responsive to environmental change than the mean skin temperature. These findings were supported by the thermographic observations.

SUMMARY1. In thirty full-term infants in the first week of life, nursed in a constant volume, closed-circuit metabolism chamber in a neutral thermal environment (31-5-33.5 TC), measurements were made of oxygen consumption (To,) during periods of rapid eye movement (REM) sleep and non-rapid eye movement (NREM) sleep.2. The mean ro, during REM sleep was 5 97 ml. kg-'. min-. In NREM sleep the mean Po, was 5-72 ml. kg-'. min-'. This difference was significant (paired t test P < 0.05).3. When the direction of sleep state change was taken into account the difference in P0o between the two states was much less when REM sleep preceded NREM than when the change was in the opposite direction. In nineteen infants in whom the change was from REM to NREM the difference in ro, (6-18 and 6-03 ml. kg-'. min')was not significant (P > 0.05). The mean difference when the sleep state change was from NREM to REM was significant (P < 0.01), the values being 5-54 and 5-81 ml. kg-1. min' respectively.4. In the NREM state, a gradual diminution of Po, with time was consistently found. This was not the case in REM sleep.5. In twelve infants studied in a cool environment (29 + 0 5 00) Po, during REM sleep was 7*77, and during NREM sleep it was 6-58 ml. kg-'.min-', (P < 0 001). Thus even the maximum difference found in a neutral thermal environment of 6.6% was significantly increased to 14-9% (P < 0.01) with mild thermal stress.

SUMMARY The thermal balance of 13 term infants was measured in a closed-circuit metabolism chamber. Each was studied naked, then with a gamgee-lined hat, and finally with a 'cummerbund' made of a similar material and of similar dimensions. At 27°C the oxygen consumption of the 'hatted' babies was only 85 % and the total heat loss 75 % of the values measured with the infants naked. The cummerbund offered no detectable benefit. An additional 10 infants were studied while wearing a tubegauze hat at environmental temperatures of 28 5 (±0* 5)°C. This type of hat gave no measurable thermal protection. It is concluded that a substantial reduction of thermal stress in adverse environments can be achieved simply and cheaply by adequately covering the vault of the skull.

4. The jugular veins in the baby's neck are occluded by finger pressure and there follows an increase in skull volume, which is rapid at first, but which decreases exponentially as venous drainage diverts to non-occluded channels such as the vertebral venous plexus. At the instant of jugular occlusion the rate of skull volume increase is representative of the rate of flow in the jugular vessels prior to occlusion, and so provides an index of the relative magnitude of the intracranial blood flow. The method thus allows changes in intracranial blood flow to be followed. When occlusion is released cranial volume decreases, initially rapidly, but slowing exponentially as resting volume is regained.5. A theoretical model of the events occurring during the inflow and outflow phases has been developed, and a formula derived which allows an estimation to be made of the flow of blood through uncompressed channels. The measured value of jugular blood flow can then be augmented to an estimate of total intracranial flow.6. The mean cerebral blood flow of sixteen normal babies was estimated to be 40 ml. 100 g-1. min-' (S.D. = + 11t63).