- Email: [email protected]
Concussive methods of pre-slaughter stunning in sheep: assessment of brain function using cortical evoked responses
Research in Veterinary Science /986. 4/. 349-352
Concussive methods of pre-slaughter stunning in sheep: assessment of brain function using cortical evoked responses c. C.
DALY, N. G. GREGORY, S. B. WOTTON, P. E. WHITTINGTON, Agricultural and Food
Research Council, Food Research Institute, Langford, Bristol B818 7DY
A veraged cortical evoked responses were used to evaluate brain function in anaesthetised sheep. Effects on both evoked somatosensory responses (SERs) and visual evoked responses (VERs) were examined. Following capitlve bolt stunning, SERs and VERs were abolished instantaneously and did not reappear for the duration of the experiment. Similar results were found 'when animals were shot while conscious. It was concluded that captive bolt stunning in sheep produces an immediate, profound and long lasting brain failure and is therefore an effective preslaughter stunning method.
IN Britain, the captive bolt is the principal preslaughter stunning method used in cattle. It is also used in about 10 per cent of sheep. During a survey of sheep slaughtering practices, it was observed that slaughtermen were not always satisfied with the effectiveness of captive bolt stunning (Gregory and Wotton 1984). Four per cent of sheep subjected to this form of stunning received a second stun, even though the bolt had penetrated the head of the animal with the first application. Brain function following captive bolt stunning has principally been assessed in terms of effects on the waveform of the electroencephalogram (EEG); unconsciousness is thought to exist when cortical recordings show high amplitude, low frequency activity (Blackmore and Newhook 1983). However, this method is not without certain limitations. For example, the EEG has failed to provide unequivocal indication of the state of consciousness as produced by sleep (Jouvet 1967) or anaesthesia (Marshall et al 1965). Furthermore, West et al (1982) found that in rats subjected to non-penetrative concussion, recovery of posture and locomotion preceded recovery of EEG abnormalities. An alternative method of evaluating brain function following concussion is the measurement of cortical evoked potentials (Foltz and Schmidt 1956, Ommaya and Gennarelli 1974, Shaw and Cant 1984). Evoked potentials have been used as an index of extent of brain damage and prognosis for recovery. Letcher et al (1973) examined the effects of concussive blows on EEG activity and somatosensory evoked responses in
the chimpanzee; they noted a quantitative relationship between the intensity of the blow and the extent of depression of the evoked response. The early effects of concussion on the EEG did not show this relationship and Letcher et al (1973) concluded that the somatosensory evoked response may be a more reliable index of neurological impairment than spontaneous cortical activity. In this study, both visual evoked responses (VERs) and somatosensory responses (SERs) were used to evaluate brain function after captive bolt stunning in sheep. These techniques were used to determine whether this stunning method produces the immediate and long lasting unconsciousness necessary for humane stunning. Materials and methods
Experimental procedures Twenty mature ewes and wethers of mixed breeding were used in these expedments, They were anaesthetised and ventilated with nitrous oxide, oxygen and with between I· 5 and I' 8 per cent halothane (Gregory and Wotton 1983), except for five sheep which were shot while conscious. These latter animals were allowed to recover from anaesthesia following implantation of the electrodes for up to 24 hours before the experiment. Animals were suspended in hammocks throughout the experiments. For the anaesthetised animals, a rope was passed through the mouth to hold the head in a horizontal position for shooting. Electrodes were made from 5 mm silver rods insulated from their recording surfaces with Araldite. The recording surfaces were chlorided as described by Evans and Leal (1980).
Visual evoked responses. VERs were recorded according to the methods of Gregory and Wotton (1983). The negative electrode was located over the visual cortex midway between the coronal and lambda sutures. The positive electrode was placed 2,cm rostral to bregma. The earth electrode was placed
349
350
c. C. Daly, N.
G. Gregory, S. B. Wotton, P. E. Whittington
midway between the positive and negative electrodes or, when two earth electrodes were required, they were positioned equidistant between the positive and negative electrodes. All electrodes were situated 1 ern lateral to the midline. The eyelids of the anaesthetised animals were sutured open at the outset of the experiment. In the cases of animals shot while conscious, eyelids were held open manually after shooting. The visual stimulus consisted of light flashes (from a light source I m from the sheep's eyes and producing a luminous power at the eyes of 63 mW cm- Z) presented at a frequency of 2 flashes sec-I.
Somatosensory evoked responses. Electrode positions for the SERs were as described for VER recording except that the positive electrode was moved 2 cm rostral and 1 ern lateral to bregma. The stimuli were produced using two stainless steel wires passed subcutaneously in a loop midway along the forelimb on the side contralateral to the recording electrodes and separated by approximately 2 em, The proximal electrode was made positive and the stimulating voltage was increased until a cleat or leg twitch became apparent. Stimuli were presented at a frequency of 2 sec-I. Shooting with captive bolt occurred following a 530 stimulus pre-stun recording period and without interrupting the recording process. The captive bolt pistol employed in these experiments was a Short Cash Special (Accles & Shelvoke) using a one grain cartridge. The pistol was positioned over the temporal bone at the highest point of the head and I to 2 em lateral to the midline to avoid damaging the electrodes. Equipment The signals were transmitted from the electrodes to the recorder via screened cables (Filotex O' 9964). One of two recorders were used: an Elema Schonander Mingograf 34 recorder (0' 15 second time constant and 70 Hz upper frequency filter) ora Mingograph
10 recorder (0' I second time constant and 70 Hz upper frequency filter). A second earth electrode was used with the EEG 10 recorder to suppress AC artefact. Neural activity was stored on a frequency modulated tape recorder (Racal Store 7DS). The evoked responses were averaged with a Neurolog NL 750, using a 250 ms sweep duration. The start of averaging preceded the stimulus by 100 ms. The averaged evoked responses were analysed as described by Gregory and Wotton (1983) to arrive at their excursion distances which were taken as an overall measure of the size of the response. Spontaneous neural activity was quantified by capacitance discharge integration (Integrator GPA-IO; Narco BioSystems) over consecutive 15 second periods. The bolt trajectory through the brain was determined following each experiment to ensure that direct damage was not done to the primary visual and somatosensory pathways. In most cases the brain was fixed in formol saline and sliced by hand into sections approximately I cm thick. Sections were examined under 10 x magnification and anatomical regions were identified according to Yoshikawa (1968).
EEG
Results
Evoked responses VERs were recorded from eight sheep and SERs from seven. Evoked responses were lost immediately in all anaesthetised sheep stunned with the captive bolt. Fig I shows an example of the effect on the VER of one animal. Excursion distances averaged over the first 32 seconds after stun fell to less than 10 per cent of the pre-stun levels (Fig 2). Furthermore, both VERs and SERs were absent when averaging was carried out over the first four seconds following shooting, suggesting that the loss of this response was instantaneous and probably generalised in the cortex (data not shown). The evoked responses did not return within 320 seconds of administering the bolt, at which time the experiment was terminated.
32 FIG 1: Example of the effect of captive bolt shooting of an anaesthetised sheep on averaged visual evoked responses. Each trace is the average of 64 flashes, corresponding to a 32 second period. The negativegoing (downward) spike occurring 100 ms after the start of each trace is an event marker for the flash of light
Assessment of brain function in stunned sheep
351
thalamus and the rostral midbrain regions. In most cases, bone fragments were found imbedded in the brain. There was no evidence of direct damage to the neural pathways leading to the recorded visual cortex. In sheep shot while conscious, movement of the animal at the time of shooting resulted in the bolt trajectory crossing the midline in two animals. In one case, the bolt clearly damaged the lateral geniculate nucleus on the side ipsilateral to the recording electrodes; in this sheep, therefore, deafferentation will have contributed to the loss of the VERso
Discussion On conceptual grounds, it would be expected that a pre-slaughter stunning method would be highly effec2:5 tive from a humanitarian standpoint ifit produced an Time after captive bolt stunning instantaneous and prolonged loss of brain responsive(min) ness to external stimuli and also caused a rapid and FIG 2: Mean (± SEI percentage of pre-treatment excursion distances complete loss of spontaneous neural activity in the of both VERs and SERs (0) and of the integrated spontaneous cortical brain. The reasons for this proposition rest on the activity (.) following captive bolt shooting of anaesthetised sheep (n ~ 151 following points. First, since evoked responses occur in both the anaesthetised and conscious states, they represent the ability of the brain to respond at a rudiCaptive bolt stunning of unanaesthetised animals mentary level. These responses do not provide a was also examined, although in these experiments neurophysiological correlate of consciousness, but only VERs were recorded. In four of the five animals, conscious awareness of a stimulus will be dependent on, and occur subsequent to, the neural activity VERs were lost immediately. However, in the remaining animal, the VER persisted for 35 seconds but dis- represented by the evoked response; the abolition of appeared thereafter. VERs did not reappear in any of evoked responses would therefore reflect at least a these animals within 160 to 192 seconds of shooting. state of insensibility. For this reason, sensory evoked potentials may provide an appropriate method of determining whether the legal requirement of stunSpontaneous activity ning, namely the immediate onset of insensibility, is In general, the integrated spontaneous cortical met. • Justification for the second part of the proposition, activity recorded from the anaesthetised animals declined after shooting (Fig 2). However, there was namely the rapid induction of isoelectric or near isoconsiderable variation in the extent of the depression electric activity in the brain, depends on observations between individual animals. Four sheep retained of these states in the human. Such patients are invarimore than 50 per cent of pre-stun activity throughout ably in a moribund or vegetative state and their progthe post-stun period, whereas the activity in the nosis is poor. As such, spontaneous cortical activity is remaining animals fell to less than 10 per cent of the a useful guide in the diagnosis of brain death, pre-stun levels in a mean (±SE) time of 76± IS seconds although some authorities prefer to use evoked potentials as well (Greenberg et al 1981). Within the following shooting. Changes in integrated cortical activity of sheep shot context of pre-slaughter stunning procedures, loss of in the conscious state were less variable and became spontaneous activity would therefore increase the confidence of an effective stun. isoelectric on average in 53± 13 seconds. The use of evoked responses to determine the effecIn both the conscious and anaesthetised animals the captive bolt passed through the parietal bone lateral tiveness of captive bolt stunning depends upon to the midline. The immobility of the anaesthetised ensuring that the neural pathways transmitting the animals meant that the site of entry and trajectory of stimulus to the cortex are not directly transected by the bolt through the brain was more consistent than in the bolt and post mortem examination of the bolt the conscious animals. In the anaesthetised animals trajectory established that this did not occur. Also, it the bolt did not cross the midline and penetrated the is possible that the use of VERs could be contrafull length of the brain. The deep structures of the indicated by reports that the optic nerve can be brain which were affected were the mid and caudal damaged as it passes transcranially by concussive
o
o
352
c. C. Daly, N.
G. Gregory, S. B. Wotton, P. E. Whittington
blows (Brihaye 1980). However, previous work has shown that evoked activity recorded directly from the optic nerve is maintained following captive bolt stunning even in those animals where all evoked and spontaneous activity is lost in the cortex. The evidence therefore suggests that loss of evoked responses in the cortex following captive bolt stunning is not due to deafferentation but rather to an inability of the cortex and, or, the thalamus to respond to afferent stimuli. The use of evoked responses to assess brain function following captive bolt stunning provides strong justification for this method on humanitarian grounds. Its high success rate, as well as its irreversibility, means that insensibility can confidently be assumed to exist in sheep until death occurs through exsanguination. These results also show that the captive bolt appears to be more effective than electrical stunning in disturbing brain function (Gregory and Wotton 1985). References BLACKMORE. D. K. & NEWHOOK, J. C. (1983) Stunning of Animals for Slaughter. Ed G. Eikelenboom. Dordrecht, Maninus Nijhoff. pp 13-25
BRIHAYE, J. (1980) The Cranial Nerves. Eds M. Samii and P. J. Janella. Berlin, Springer Verlag. pp 116-124 FOLTZ, E. L. & SCHMIDT, R. P. (1956) Journal of Neurosurgery 13, 145-154 EVANS, T. D. & LEAL, J. R. (1980) Laboratory Practice 29, 846-847 . GREENBERG R. P., WARD, S. P., LUTZ, H., MILLER, J. D. & BECK D. P. (1981) Brain Failure and Resuscitation. Eds A. Grenvile and P. Savan, Edinburgh, Churchill Livingston. pp 67-90 GREGORY, N. G. & WOTTON, S. B. (1983) Research in Veterinary Science 34, 315-319 GREGORY, N. G. & WOTTON, S. B. (1984) British Veterinary Journal 140, 281-286 GREGORY, N. G. & WOTTON, S. B. (1985) British Veterinary Journal 141, 74-81 JOUVET, M. (1967) Physiological Review 47,117-177 LETCHER, F. S., CARRAO, P. G. & OMMAYA, A. K. (1973) Journal of Neurosurgery 39, 167-177 MARSHALL, M., LONGLEY, B. P. & STANTON, W. H. (1965) British Journal of Anaesthesia 37,845-857 OM MAYA, A. K. & GENNARELLI, T. (1974) Brain 97, 633-654 SHAW, N. A. & CANT, B. R. (1984) Australian Journal of Experimental Biology and Medicine 62,361-371 WEST, M., PARKINSON, D. & HAVLICEK, V. (1982) Electroencephalography and Clinical Neurophysiology 53, 192-200 YOSHIKAWA, T. (1968) Atlas of the Brain of Domestic Animals. University of Tokyo Press
Accepted October 25, 1985
Concussive methods of pre-slaughter stunning in sheep: assessment of brain function using cortical evoked responses c. C.
DALY, N. G. GREGORY, S. B. WOTTON, P. E. WHITTINGTON, Agricultural and Food
Research Council, Food Research Institute, Langford, Bristol B818 7DY
A veraged cortical evoked responses were used to evaluate brain function in anaesthetised sheep. Effects on both evoked somatosensory responses (SERs) and visual evoked responses (VERs) were examined. Following capitlve bolt stunning, SERs and VERs were abolished instantaneously and did not reappear for the duration of the experiment. Similar results were found 'when animals were shot while conscious. It was concluded that captive bolt stunning in sheep produces an immediate, profound and long lasting brain failure and is therefore an effective preslaughter stunning method.
IN Britain, the captive bolt is the principal preslaughter stunning method used in cattle. It is also used in about 10 per cent of sheep. During a survey of sheep slaughtering practices, it was observed that slaughtermen were not always satisfied with the effectiveness of captive bolt stunning (Gregory and Wotton 1984). Four per cent of sheep subjected to this form of stunning received a second stun, even though the bolt had penetrated the head of the animal with the first application. Brain function following captive bolt stunning has principally been assessed in terms of effects on the waveform of the electroencephalogram (EEG); unconsciousness is thought to exist when cortical recordings show high amplitude, low frequency activity (Blackmore and Newhook 1983). However, this method is not without certain limitations. For example, the EEG has failed to provide unequivocal indication of the state of consciousness as produced by sleep (Jouvet 1967) or anaesthesia (Marshall et al 1965). Furthermore, West et al (1982) found that in rats subjected to non-penetrative concussion, recovery of posture and locomotion preceded recovery of EEG abnormalities. An alternative method of evaluating brain function following concussion is the measurement of cortical evoked potentials (Foltz and Schmidt 1956, Ommaya and Gennarelli 1974, Shaw and Cant 1984). Evoked potentials have been used as an index of extent of brain damage and prognosis for recovery. Letcher et al (1973) examined the effects of concussive blows on EEG activity and somatosensory evoked responses in
the chimpanzee; they noted a quantitative relationship between the intensity of the blow and the extent of depression of the evoked response. The early effects of concussion on the EEG did not show this relationship and Letcher et al (1973) concluded that the somatosensory evoked response may be a more reliable index of neurological impairment than spontaneous cortical activity. In this study, both visual evoked responses (VERs) and somatosensory responses (SERs) were used to evaluate brain function after captive bolt stunning in sheep. These techniques were used to determine whether this stunning method produces the immediate and long lasting unconsciousness necessary for humane stunning. Materials and methods
Experimental procedures Twenty mature ewes and wethers of mixed breeding were used in these expedments, They were anaesthetised and ventilated with nitrous oxide, oxygen and with between I· 5 and I' 8 per cent halothane (Gregory and Wotton 1983), except for five sheep which were shot while conscious. These latter animals were allowed to recover from anaesthesia following implantation of the electrodes for up to 24 hours before the experiment. Animals were suspended in hammocks throughout the experiments. For the anaesthetised animals, a rope was passed through the mouth to hold the head in a horizontal position for shooting. Electrodes were made from 5 mm silver rods insulated from their recording surfaces with Araldite. The recording surfaces were chlorided as described by Evans and Leal (1980).
Visual evoked responses. VERs were recorded according to the methods of Gregory and Wotton (1983). The negative electrode was located over the visual cortex midway between the coronal and lambda sutures. The positive electrode was placed 2,cm rostral to bregma. The earth electrode was placed
349
350
c. C. Daly, N.
G. Gregory, S. B. Wotton, P. E. Whittington
midway between the positive and negative electrodes or, when two earth electrodes were required, they were positioned equidistant between the positive and negative electrodes. All electrodes were situated 1 ern lateral to the midline. The eyelids of the anaesthetised animals were sutured open at the outset of the experiment. In the cases of animals shot while conscious, eyelids were held open manually after shooting. The visual stimulus consisted of light flashes (from a light source I m from the sheep's eyes and producing a luminous power at the eyes of 63 mW cm- Z) presented at a frequency of 2 flashes sec-I.
Somatosensory evoked responses. Electrode positions for the SERs were as described for VER recording except that the positive electrode was moved 2 cm rostral and 1 ern lateral to bregma. The stimuli were produced using two stainless steel wires passed subcutaneously in a loop midway along the forelimb on the side contralateral to the recording electrodes and separated by approximately 2 em, The proximal electrode was made positive and the stimulating voltage was increased until a cleat or leg twitch became apparent. Stimuli were presented at a frequency of 2 sec-I. Shooting with captive bolt occurred following a 530 stimulus pre-stun recording period and without interrupting the recording process. The captive bolt pistol employed in these experiments was a Short Cash Special (Accles & Shelvoke) using a one grain cartridge. The pistol was positioned over the temporal bone at the highest point of the head and I to 2 em lateral to the midline to avoid damaging the electrodes. Equipment The signals were transmitted from the electrodes to the recorder via screened cables (Filotex O' 9964). One of two recorders were used: an Elema Schonander Mingograf 34 recorder (0' 15 second time constant and 70 Hz upper frequency filter) ora Mingograph
10 recorder (0' I second time constant and 70 Hz upper frequency filter). A second earth electrode was used with the EEG 10 recorder to suppress AC artefact. Neural activity was stored on a frequency modulated tape recorder (Racal Store 7DS). The evoked responses were averaged with a Neurolog NL 750, using a 250 ms sweep duration. The start of averaging preceded the stimulus by 100 ms. The averaged evoked responses were analysed as described by Gregory and Wotton (1983) to arrive at their excursion distances which were taken as an overall measure of the size of the response. Spontaneous neural activity was quantified by capacitance discharge integration (Integrator GPA-IO; Narco BioSystems) over consecutive 15 second periods. The bolt trajectory through the brain was determined following each experiment to ensure that direct damage was not done to the primary visual and somatosensory pathways. In most cases the brain was fixed in formol saline and sliced by hand into sections approximately I cm thick. Sections were examined under 10 x magnification and anatomical regions were identified according to Yoshikawa (1968).
EEG
Results
Evoked responses VERs were recorded from eight sheep and SERs from seven. Evoked responses were lost immediately in all anaesthetised sheep stunned with the captive bolt. Fig I shows an example of the effect on the VER of one animal. Excursion distances averaged over the first 32 seconds after stun fell to less than 10 per cent of the pre-stun levels (Fig 2). Furthermore, both VERs and SERs were absent when averaging was carried out over the first four seconds following shooting, suggesting that the loss of this response was instantaneous and probably generalised in the cortex (data not shown). The evoked responses did not return within 320 seconds of administering the bolt, at which time the experiment was terminated.
32 FIG 1: Example of the effect of captive bolt shooting of an anaesthetised sheep on averaged visual evoked responses. Each trace is the average of 64 flashes, corresponding to a 32 second period. The negativegoing (downward) spike occurring 100 ms after the start of each trace is an event marker for the flash of light
Assessment of brain function in stunned sheep
351
thalamus and the rostral midbrain regions. In most cases, bone fragments were found imbedded in the brain. There was no evidence of direct damage to the neural pathways leading to the recorded visual cortex. In sheep shot while conscious, movement of the animal at the time of shooting resulted in the bolt trajectory crossing the midline in two animals. In one case, the bolt clearly damaged the lateral geniculate nucleus on the side ipsilateral to the recording electrodes; in this sheep, therefore, deafferentation will have contributed to the loss of the VERso
Discussion On conceptual grounds, it would be expected that a pre-slaughter stunning method would be highly effec2:5 tive from a humanitarian standpoint ifit produced an Time after captive bolt stunning instantaneous and prolonged loss of brain responsive(min) ness to external stimuli and also caused a rapid and FIG 2: Mean (± SEI percentage of pre-treatment excursion distances complete loss of spontaneous neural activity in the of both VERs and SERs (0) and of the integrated spontaneous cortical brain. The reasons for this proposition rest on the activity (.) following captive bolt shooting of anaesthetised sheep (n ~ 151 following points. First, since evoked responses occur in both the anaesthetised and conscious states, they represent the ability of the brain to respond at a rudiCaptive bolt stunning of unanaesthetised animals mentary level. These responses do not provide a was also examined, although in these experiments neurophysiological correlate of consciousness, but only VERs were recorded. In four of the five animals, conscious awareness of a stimulus will be dependent on, and occur subsequent to, the neural activity VERs were lost immediately. However, in the remaining animal, the VER persisted for 35 seconds but dis- represented by the evoked response; the abolition of appeared thereafter. VERs did not reappear in any of evoked responses would therefore reflect at least a these animals within 160 to 192 seconds of shooting. state of insensibility. For this reason, sensory evoked potentials may provide an appropriate method of determining whether the legal requirement of stunSpontaneous activity ning, namely the immediate onset of insensibility, is In general, the integrated spontaneous cortical met. • Justification for the second part of the proposition, activity recorded from the anaesthetised animals declined after shooting (Fig 2). However, there was namely the rapid induction of isoelectric or near isoconsiderable variation in the extent of the depression electric activity in the brain, depends on observations between individual animals. Four sheep retained of these states in the human. Such patients are invarimore than 50 per cent of pre-stun activity throughout ably in a moribund or vegetative state and their progthe post-stun period, whereas the activity in the nosis is poor. As such, spontaneous cortical activity is remaining animals fell to less than 10 per cent of the a useful guide in the diagnosis of brain death, pre-stun levels in a mean (±SE) time of 76± IS seconds although some authorities prefer to use evoked potentials as well (Greenberg et al 1981). Within the following shooting. Changes in integrated cortical activity of sheep shot context of pre-slaughter stunning procedures, loss of in the conscious state were less variable and became spontaneous activity would therefore increase the confidence of an effective stun. isoelectric on average in 53± 13 seconds. The use of evoked responses to determine the effecIn both the conscious and anaesthetised animals the captive bolt passed through the parietal bone lateral tiveness of captive bolt stunning depends upon to the midline. The immobility of the anaesthetised ensuring that the neural pathways transmitting the animals meant that the site of entry and trajectory of stimulus to the cortex are not directly transected by the bolt through the brain was more consistent than in the bolt and post mortem examination of the bolt the conscious animals. In the anaesthetised animals trajectory established that this did not occur. Also, it the bolt did not cross the midline and penetrated the is possible that the use of VERs could be contrafull length of the brain. The deep structures of the indicated by reports that the optic nerve can be brain which were affected were the mid and caudal damaged as it passes transcranially by concussive
o
o
352
c. C. Daly, N.
G. Gregory, S. B. Wotton, P. E. Whittington
blows (Brihaye 1980). However, previous work has shown that evoked activity recorded directly from the optic nerve is maintained following captive bolt stunning even in those animals where all evoked and spontaneous activity is lost in the cortex. The evidence therefore suggests that loss of evoked responses in the cortex following captive bolt stunning is not due to deafferentation but rather to an inability of the cortex and, or, the thalamus to respond to afferent stimuli. The use of evoked responses to assess brain function following captive bolt stunning provides strong justification for this method on humanitarian grounds. Its high success rate, as well as its irreversibility, means that insensibility can confidently be assumed to exist in sheep until death occurs through exsanguination. These results also show that the captive bolt appears to be more effective than electrical stunning in disturbing brain function (Gregory and Wotton 1985). References BLACKMORE. D. K. & NEWHOOK, J. C. (1983) Stunning of Animals for Slaughter. Ed G. Eikelenboom. Dordrecht, Maninus Nijhoff. pp 13-25
BRIHAYE, J. (1980) The Cranial Nerves. Eds M. Samii and P. J. Janella. Berlin, Springer Verlag. pp 116-124 FOLTZ, E. L. & SCHMIDT, R. P. (1956) Journal of Neurosurgery 13, 145-154 EVANS, T. D. & LEAL, J. R. (1980) Laboratory Practice 29, 846-847 . GREENBERG R. P., WARD, S. P., LUTZ, H., MILLER, J. D. & BECK D. P. (1981) Brain Failure and Resuscitation. Eds A. Grenvile and P. Savan, Edinburgh, Churchill Livingston. pp 67-90 GREGORY, N. G. & WOTTON, S. B. (1983) Research in Veterinary Science 34, 315-319 GREGORY, N. G. & WOTTON, S. B. (1984) British Veterinary Journal 140, 281-286 GREGORY, N. G. & WOTTON, S. B. (1985) British Veterinary Journal 141, 74-81 JOUVET, M. (1967) Physiological Review 47,117-177 LETCHER, F. S., CARRAO, P. G. & OMMAYA, A. K. (1973) Journal of Neurosurgery 39, 167-177 MARSHALL, M., LONGLEY, B. P. & STANTON, W. H. (1965) British Journal of Anaesthesia 37,845-857 OM MAYA, A. K. & GENNARELLI, T. (1974) Brain 97, 633-654 SHAW, N. A. & CANT, B. R. (1984) Australian Journal of Experimental Biology and Medicine 62,361-371 WEST, M., PARKINSON, D. & HAVLICEK, V. (1982) Electroencephalography and Clinical Neurophysiology 53, 192-200 YOSHIKAWA, T. (1968) Atlas of the Brain of Domestic Animals. University of Tokyo Press
Accepted October 25, 1985