To realize that a patient is in shock and then to provide relevant treatment is potentially lifesaving. It remains, therefore, perplexing that medical textbooks (and the Advanced Trauma Life Support guidelines; www.atls.dk) do not regularly provide an accurate description of the heart rate (HR) and blood pressure responses to haemorrhage (Secher & Bie, 1985). The HR and blood pressure responses to haemorrhage are crucial, since these variables remain the by far most used to assess the circulation in patients. A reduction in the central blood volume is followed by a moderate (most often to less than 100 beats min−1) increase in HR, while mean arterial pressure (MAP) is maintained. When a patient is bleeding, this condition can be described as preshock, but the cardiovascular manifestations are no more dramatic than those experienced when sitting or standing up (Van Lieshout et al. 2003). However, as described by Barcroft et al. (1944), a reduction of the central blood volume by approximately 30% elicits a Bezold–Jarish-like reflex with loss of (muscle) sympathetic activity, while vagal stimulation provokes an eventually serious slowing of HR similar to that seen during a vasovagal syncope. If the patient survives this stage of shock, a further blood loss provokes a new and larger (to 120 beats min−1) increase in HR with a transition to an irreversible state of shock (Secher et al. 1992). It was the observation of Little et al. (1989) that the HR response to haemorrhage does not always follow the above-mentioned path. The HR continues to increase when haemorrhage is provoked in an animal that is also exposed to pain associated with trauma or ischaemia, as presented in this issue of Experimental Physiology (Sawdon et al. 2009). It has been argued that the Bezold–Jarish-like response to a critically reduced central blood volume is attributed to a central nervous system δ-opioid mechanism (Evans et al. 1989), but in humans, naloxone used at a normal dose does not attenuate the response (Madsen et al. 1995). Sawdon et al. (2009) evaluated whether the persistently elevated HR response to haemorrhage is affected by naloxone in rats exposed to hindlimb ischaemia. The answer is negative, and naloxone did not affect the low MAP. Ischaemia provokes a marked increase in sympathetic activity, and experimentally, postexercise muscle ischaemia is used to provoke a maintained severe decrease in muscle oxygenation. In that situation, the prevailing MAP during exercise is maintained after exercise, or it increases, since muscle sympathetic activity is markedly enhanced (Mark et al. 1985). Thus, it is likely that the Bezold–Jarish-like reflex elicited in response to haemorrhage is overruled by sympathetic activation. In support of that idea, a vasovagal reaction can be terminated by muscle contractions, and the HR response to haemorrhage is enhanced by ileus (Jacobsen et al. 1993). Yet, it remains to be established whether the influence of sympathetic activation on the HR response to haemorrhage is due to an effect on the heart or the central blood volume or whether a central mechanism is involved. It also needs to be established why sympathetic activation is unable to influence MAP during severe haemorrhage when it can apparently combat a vasovagal syncope and does increase MAP during postexercise muscle ischaemia. Also, postexercise muscle ischaemia does not maintain an elevated HR. Clearly, there is a need to define when and why vagal activity is turned on and off versus the influence of sympathetic activation on HR. The study by Sawdon et al. (2009) calls for evaluation, for example, by using autonomic blockade, to elucidate the mechanisms that cause the persistent increase in HR when haemorrhage is associated with pain.