In the decades preceding and following the Second World War, knowledge about the human peripheral circulation increased rapidly. A number of factors contributed. One was the increasing interest in the field and the analytical techniques used initially by Sir Thomas Lewis at the Department of Medicine, University College London, and spread by some of his colleagues, G. W. Pickering, R. T. Grant and R. S. B. Pearson. Another was the increased use of venous occlusion plethysmography for non-invasive measurements of blood flow in human limbs (Lewis & Grant, 1925; Grant & Pearson, 1938; Barcroft & Edholm, 1943; Greenfield, 1954). Subjects suffered little discomfort with this technique so it was relatively easy to recruit students and colleagues as volunteers for experiments. Before the advent of ‘Research Ethical Committees’ in the 1960s, there were fewer constraints on working with human subjects. Experimenters, most of whom had medical qualifications, conformed to their own (usually high) ethical standards. They were familiar with clinical procedures and the risks involved but controversy did arise from time to time (Roddie, 1977). Such work might be more difficult to do today because of ethical, medico-legal and bureaucratic considerations but it does continue despite these problems (Joyner & Halliwill, 2000). It is important that it does so if we are to understand how new molecular information might or might not operate in whole humans. At the time the paper under review (Roddie et al. 1957a; available online with this issue of The Journal of Physiology) was written, it was known that when one limb of a human subject was immersed in warm water, the blood flow through the other limbs increased (Lewis & Pickering, 1931; Gibbon & Landis, 1932). Sympathetic nerves were known to mediate this response to ‘indirect heating’, since it was abolished if these nerves were blocked or cut (Grant & Pearson, 1938; Barcroft et al. 1943). There was evidence that the response was mediated by both release of vasoconstrictor tone (Barcroft & Hamilton, 1948) and activity in vasodilator nerves (Grant & Holling, 1938). The responses appeared to be confined to skin since they could not be demonstrated in underlying muscle (Barcroft et al. 1955; Edholm et al. 1956; Roddie et al. 1956) where the vasoconstrictor nerves (Barcroft et al. 1943) are involved in baroreceptor rather than thermoregulatory reflexes (Roddie et al. 1957b; Greenfield et al. 1966). Roddie et al. (1957a) looked at the relative contribution of vasoconstrictor and vasodilator nerves to the increase in blood flow in hand and forearm skin with indirect heating. Using venous occlusion plethysmography in subjects who were initially cooled, it was found that the forearm vasodilatation with indirect heating occurred in two phases. In the first phase there was a small increase in flow that started at the same time as the larger vasodilatation in the hand. In the second phase, some 5–10 min later, forearm flow increased markedly at about the time skin sweating commenced. Treating the forearm with intra-arterial atropine abolished sweating and reduced and delayed the second phase, suggesting that it had a cholinergic component. Atropine treatment did not affect the first phase of vasodilatation in the forearm or the vasodilatation in the hand. It was concluded that the mechanism of vasodilatation with indirect heating is different in hand and forearm skin. These differences are summarized in the idealized experiments shown in Fig. 1. In the hand, blood vessels are normally subjected to high levels of vasoconstrictor tone and the large increase in flow with heating can be explained entirely by release of this tone. In forearm skin, however, the vasodilatation that occurs with indirect heating in comfortably warm subjects is mediated mainly by vasodilator nerves and is associated with sweat gland activity. It is greater than can be explained by complete release of vasoconstrictor tone. Though forearm skin has a vasoconstrictor nerve supply, vasoconstrictor tone is minimal in subjects who are comfortably warm. Blocking the sympathetic nerves in such subjects does not increase flow (Barcroft et al. 1943). However, if the subjects are cooled, vasoconstrictor tone increases to decrease local flow and hence heat loss from the skin. Blocking sympathetic nerves blocks this reflex response (Blair et al. 1960). Figure 1 Schematic representation of changes in blood flow in normal and nerve-blocked hand and forearm during body cooling and heating. At CNB, vasomotor nerves to the hand and forearm skin were blocked with local anaesthetic solution. Reproduced with permission ... Later work has shown that the patterns of thermoregulatory response in the hand and forearm blood vessels occur in skin in other parts of the body (Roddie, 1983). In extremities, such as feet, ears, nose, lips etc., the pattern is similar to that in the hand - blood vessels are normally subjected to high levels of vasoconstrictor tone that is released when core temperature rises (Blair et al. 1960; Fox et al. 1962). However, skin in more proximal parts of the body such as the abdomen, chest, back, upper arms and thighs has minimal vasoconstrictor tone when the subjects are comfortably warm. The increase in skin flow in these areas with heating cannot be explained by release of vasoconstrictor tone. It is mediated mainly by vasodilator nerves and is associated with sweat gland activity (Blair et al. 1960, 1961; Fox et al. 1962). Though these nerves have been shown to be cholinergic (Kellogg et al. 1995), the neurotransmitter responsible for the vasodilatation remains elusive. However, the search for it continues today using variants of the analytical strategies employed in earlier studies. Nitric oxide may contribute but is not the prime agent (Dietz et al. 1994; Kellogg et al. 1998; Shastry et al. 2000). A current hypothesis is that an unknown substance, perhaps a vasoactive peptide co-released from the cholinergic nerve endings, is responsible (Kellogg et al. 1995).