Sleep and waking in the three-toed sloth, Bradypus tridactylus

Sleep and waking in the three-toed sloth, Bradypus tridactylus

Camp. Bkx*htw. P~IwoI. Vol. ?6A. No. 2, pp 345-355. 1983 0300-9629W $3.00 + 0.00 Printed in Great Biitain ( 1981Pergaman Press Ltd SLEEP AND WAKI...

1MB Sizes 0 Downloads 208 Views

Camp. Bkx*htw. P~IwoI. Vol. ?6A. No. 2, pp 345-355. 1983

0300-9629W $3.00 + 0.00

Printed in Great Biitain

( 1981Pergaman Press Ltd









GOMES [email protected] Department of Physiology and Pharmacology, Federal University of Pemambuco, Recife. Brazil and ZDepartment of- Biology. University of Houston, Houston, TX 77004, LJ.S.A. (Receiwd

I5 December


Abstract-l. The sleep-waking cycles of the three-toed sloth, &u&us triduct+s, were studied by direct observation and po~ygrapk~c recording and the results were correlated. 2. Three states of waking behavior were recognized, “awake-exploring”, “awake-alert” and “awakefixating”. 3. EEG recordings were classified as two waking states. A, which was seen during both “awakeexploring” . _ and “awake-alert“ behavior and ALwhich had some of the characteristics of drowsy behavior of other mammals and often acconlp~ni~d the “awake-~xatin~” behvaior. 4. EEG patterns of sleep were divided into five types: two in light sleep, two in deep sleep and one in paradoxical sleep. 5. Animals spent about 30”/, of the time in various degrees of waking behavior and about 709; asleep. 6. Waking EEGs of various types were seen about 345, of the time, about IO”;,of that during behavioral sleep. 7. EEG patterns of light sleep constituted about 56U,;,of the records. about h?;, of that during some form of waking behavior. 8. EEG patterns of deep sleep constituted about IO?, of the records; all of this came during behavioral sleep. 9. The total duration of sleep and of paradoxical sleep did not depart materially from the expected pattern based on phylogenetic position. 10. The pattern of sleep is polycyclic but in the laboratory situation animals were sleeping most soundly between 6a.m. and noon.

INTRODUCTION There are many attributes of the sloths of Central and South America which have interested naturalists and physiologists alike; for the latter, at least, the extreme slowness of movement is most intriguing. What are the physiological causes and what arc the results in terms of day-to-day functioning for an animal with locomotor speeds of 0.25-0.35 km/hr (Britton and Kline, 1939) and sleep apparently lasting 20 out of 24 hr (Beebe, 1926)? A number of workers including Enger and Bullock (1965) have carefully examined the physiology of sloths, system by system, in an attempt to determine which systems if any are deficient or better-termed specialized, because, from an evolutionary point of

*This work was supported by Conselho Naeional de Pesquisas (CNPq) and FINEP. TPresent address: Departamento de ReabilitaCao, Universidade Federal de Pernambuco CEP 50000 Recife, Pernambuco, Brazil. $Present address: Departamento de ~europsiquiatria, Universidade Federal de Pernambuco CEP SOOOOReeife, Pernambuco, Brazil. Correspondence should be addressed to: Dr Sara E. Huggins, Department of Biology, University of Houston, Houston, TX 77004. U.S.A.

view, these are not primitive animals with primitive simpIicjt~es but animals of extreme specialization. For scientists interested in neurophysiology the phenomena of extreme muscular slowness and apparently long sleep-waking ratio are very compelling features for study. Of the two genera of sloths extant, the genus Bru&rms is the slower moving and more quiescent in behavior. According to observers (Beebe, 1926; Britton, Kline and Silvette, 1938; Tirler, 1966) species of the three-toed genus are at rest, apparently asleep, for 20 out of 24 h. The true nature of this long inactive period, however, has not been ascertained for this genus although Holmes and Goffart (1946) have recorded the electroencephalograms of the two-toed genus, C~o~~epus, including periods when the animals were presumed to be sleeping. Animals of the genus Brudypus have been said to be crepuscular (Goffart, 1971) feeding actively morning and evening. Some observers have stated that they are completely inactive at night (Beebe, 1926; Tirler, 1966) and others considered them to be truly nocturnal (Lundy, 1952; Kreig, 1961). Although the popular impression of sloths is that they spend all their days hanging by four limbs upside-down, feeding, moving and sfeeping in this position, naturalists well acquainted with their habits observe that they spend the major portion of their time seated in the forks of trees with one or both pairs 345



apparatus, a Gilson polygraph, model PR 5 (Gilson Medical Electronics Inc.. Middleton, Wisconsin), were in a separate room, the door between being curtained with only a 4 cm’ opening for viewing the subject. Recordings of 90 s out of each 15 min period were made over each 24 h period. All of these recordings were made in the months of February, nature in terms of electroencephalographic activity. March and April, that is late summer and early fall at this latitude. Animals of group III were two males and two females MATERIALS AND METERS weighing between 2.6 and 4.9 kg. The preparation was the same as for group II except that an additional set of EFG Twenty-four sloths, Bradypus rridactylus, juveniles and electrodes having one needle over the vertex and one over adults, weighing between 2.1 and 4.7 kg were used for this study. Both sexes were included in the study groups but were the right temporal region were added, a transducer for not divided on this basis, specifically. The animals were respiration consisting of a thermocouple in the right nostril caught in the forests of the State of Pernambuco, Brazil and was used, and that the placement of the ECG electrodes were kept free in a large room in the vivarium of the approximated the conformation of lead 1 rather than that Department of Physiology and Pharmacology of the Fed- of lead II. Several other methods of picking up respiration eral University of Pemambuco. There they had access at all were tried with very little success. The recording was done times to water and fresh leaves of the ymbahyba tree on an eight channel Grass Polygraph (model SP p4. Grass (c’c~cropiu sp.). Before a week elapsed all animals were Medical Instruments, Quincy, Massachusetts). Except for feeding regularly and most of them were quite tame. the obviously necessary intervals for paper changing and For the purposes of this study 24 animals were used as some brief periods of disruptive animal movement, recordfollows: group t-12 animals which were used for pre- ings were continuous over 24 hr periods. This group of liminary studies i.e. to develop appropriate electrodes, to records were made in the months of September and Octodetermine good recording conditions, to note behavior in ber. To evaluate these records the epoch of 40sec was captivity, etc.; group II--eight animals which were used to adopted and the calculations were done by an HP 1000 study the correlation between observed behavior and the computer (Hewllet Packard do Brasil). A program was recorded EEG and to determine the circadian rhythm of developed to received 2160 data bits from seven stages of the species and group III-four animals which were used EEG record detailed below. to quantitate the various patterns of the EEG record and The frequencies of respiration and heart beat were related to correlate the respiratory and heart rates with these to the different stages of sleep and wakefulness calculated by patterns. the method of Snyder et al. i 19641.The average of different Because long-term recording from the preparations was epochs was calculated by periods of Zmin of each phase one of the goals of these experiments and because ~nditions taken at intervals of at least IOmin. Behavior patterns were divided by a modi~cation of the made control of infection difficult, intracranial electrodes were not successful. Silver disc electrodes of varying sizes scheme used by Barratt (1965) for evaluating activity of the applied to the shaved scalp with Redox cream gave signals opossum: (A) “awake-exploring”, the animal was alert. too weak to be analyzed but finally a satisfactory needle moving head and neck, blinking its eyes frequently and electrode to be fixed in the scalp was developed. These were sniffing, Barratt’s “waking, alert. searching”; (B) “awakemade from 2.5 cm of stainless steel orthodontic wire, No. 70 alert”, the head was up. the eyes open and showing occa(Labordental Ltd), sharpened and polished, and insulated sional blinks, Barratt’s “waking, alert”; (C) “awake(Varnish-GE 9564) except for a 2mm tip. These needles fixating”, head up, eyes open but showing a kind of tonic were soldered to 3 m of flexible connecting wire. immobility, really Barratt’s “waking, alert, fixation” and The eight sloths of group II included three males and five “tonic immobility” combined; and (D) “behavioral sleep”. females weighing from 2.1 to 4.5 kg. Preparation included the animal sitting with one or both pairs of limbs around an shaving the head and neck, infiltrating the skin of selected upright limb, head down on the chest. eyes closed, Barrett’s “sleep”. sites with 29; xylocaine without a vas~onstrictor (Astra Quemica, Ltd), waiting 5 min, and then implanting the S.C. The el~troen~phalographic recording was classified by electrodes as follows: the first pair over the frontal bones; an adaptation of the criteria of Dement and Kleitman (I 957) the second pair at the external angles of the eyes for and thus divided as follows: A,-“Awake-alert”, low amplielectro-occulograms; the third pair over the parietal bone at tude, rapid frequency waves in the beta range; EOG and the level of the parieto-occipital fissure; and the fourth pair EMG showing very active patterns. A,-“Awake-relaxed”. low amplitude, fast activity alternating with higher ampli: in the suboccipital region beneath the paravertebral mustude, slower activity in the alpha range; EOG showing culature for the electromyogram. A single electrode between the scapulae served as a ground. Pairs of electrodes were undulating movement; EMG showing moderate to little activity. S,---“Light sleep”, a mixture of freauencies but placed symmetrically and a minimum distance of 1 cm was kept between all electrodes. In the earlier animals a radiowith activity in the alpha-and theta ranges predbminant but graphic check of positions was made. Silver plate electrodes also with up to 207: of the activity of high amnlitude and were placed in shaved patches on the dorsal surface of the a frequency-in the delta range; EOG show& no movemellt: trunk to give an electrocardiog~m approxi~dting lead II. EMG showing either a very low amplitude trace or isoAll lead wires were formed into a flexible cable, A sleeveiess electric. S,--“Light sleep”. a mixture of frequencies, with jacket served to hold trunk electrodes in place and a cap of delta activity less than 2076, and including spindles and adhesive tape stabilized the head electrodes, The animal sharp waves in the vertex; EOG showing no eye movement: being studied was free to move for some distance but was EMG showing little or no activity. +“Deep sleep”, EEG provided with a tree limb having a suitable fork in which characterized by high voltage activity in the delta range to sit. the food leaves provided were attached to this limb, constituting up to SOP;,of the record and with prominent and the animal usually remained in one place for long spindles; EQG showing no eye activity; EMG showing little periods of time. or no movement. S,--“Deep sleep”, EEG record showing The laboratory temperature ranged from 25 to 29 C with more than SO/;;high voltage activity in the delta range and a mean of 28°C. In the daytime light came from windows no spindles; EOG showing no eye activity; EMG record in the room; at night illumination was kept as low as was isoelectric. S, or paradoxical sleep (PS), EEG of low voltage, consistent with necessary observation, a 15 W light bulb fast activity in the beta range; EOG showing bursts of being ali that was used. The observer and the recording activity; EMG remaining isoelectric,

oflimbs wrapped around the trunk and with the head down upon the chest (Beebe, 1926; Britton, 1941). The present study was undertaken to answer the questions of whether the long periods of inactivity do indeed represent sleep and if not what then is their

Sleep and waking

in the three-toed



Fig. 1. Recording of sloth “awake-exploring”. A,. This animal was eating, taking typical, tiny, almost continuous bites. All recordings are from the same individual. No. 2 of group III.

RESULTS Figures 1-9 show typical polygraphic recordings giving examples of sloth records classified by each of the above stated criteria. Figures 1 and 2 are variations of the A, state showing considerable muscular activity and, of course are variations of the behavioral “awake, exploring” condition. Figure I records the physiological parameters of an animal chewing actively. Because of the interference of the jaw muscles’ activity there is very little to be noted but the rapidity of these movements was, in itself interesting. The small bursts of activity are produced by the tiny, nibbling bites taken as the animal eats in a tight spiral around the edge of a large, palmate leaf blade persisting until nothing remains but the petiole. The




rate at which these characteristic little bites were taken was about 114 per min and demonstrated that there was at least one activity at which the sloth proceeded rapidly. All the figures shown were taken from the record of the small female, No. 2 of group III. The heart and respiratory rates were high in comparison with other individuals in the group, being about 138 beats/min and 11-12 breaths/min respectively while chewing in this instance. This, however, was at least 30”/, higher than for the same animal in a quiet waking state and demonstrates that eating was done at a great expenditure of energy. Figure 2 shows mostly low voltage fast activity but also contains bursts of sinusoidal waves in the 35 Hz range and with a voltage greater than the background activity. These were most obvious in the frontal



electrode record and were presumed to be olfactory spindles. The respiratory record showed shallow, very frequent movements suggesting sniffing, an activity producing olfactory spindles in many animal EEGs. The record also showed slow undulations in the EOG which would be expected when the eyes were searching the environment, these were also picked up in the frontal EEG record. Figure 3 shows an alert (A,) stage but without sufficient movement to interfere with the EEG trace. The heart rate was still atypically fast, i.e. 108 per min, but had slowed materially from that recorded while the animal was eating. The A> state is illustrated in Fig. 4. The EEG shows a predominance of low voltage, beta range activity but also some slower activity in the alpha range about 407; of the time. This was not, however, in clear, conspicuous. high amplitude trains of typical alpha activity. There were also some bursts of higher amplitude sinusoidal activity in the 35 Hz range, apparently occurring at the time of inspiration and, there-

fore, considered to be olfactory spindles. Activity in the nuchal muscles was very low, having diminished gradually over the foregoing range of activities. The heart rate was. however, very fast, back up to about 126 beatsjmin. Figure S shows what was considered to be the earliest stage of sleep, S,, with mixed activity in the alpha and theta ranges predominating but also showing about 20% activity of a delta frequency and some “olfactory spindles“. The heart rate was slower, 108 beats/min, and quite regular. Although the sample does not demonstrate it, the respiration was also regular. The EOG and EMG records were essentially isoelectric but the former lead picked up EEG activity. The first stage to show sleep spindles, S:, is shown in Fig. 6. There was still about 20% activity in the alpha frequency range and there were some high amplitude, sharp waves in the vertex trace. The EOG record is again showing EEG activity. The heart rate remained at about 108 beats/min and regular.

Sleep and waking

Fig. 5. Recording

in the three-toed sloth

of sloth showing

The animal in the S, stage as shown in Fig. 7 had an increase in amplitude of activity and much of it was of theta and delta frequencies, with the latter constituting between 20 and 50% of the activity. There were sleep spindles and sharp vertex waves. Respiration became deeper and the heart rate faster. The EMG was isoelectric and continued so through the deeper sleep stages. The EOG lead continued to pick up EEG activity. Figure 8, classified as S,, consisted of more than 50% activity in the delta range and there were no sleep spindles or sharp waves from the vertex lead. The rates of heart and respiration were approx. as those in stage S,. The recording, Fig. 9, considered to represent paradoxical sleep showed a predominance of EEG activity in the low voltage, fast range. The eyes appeared to be active and it was considered that the artifacts from this rhythmic activity appeared to be picked up in the EEG traces. Heart and respiratory rates were both irregular and that of the heart, at


light sleep, S,

least, was slower than in previous sleep records; the rates were approx. 90 beats/min and 6 breathsimin. The EMG continued to be isoelectric although there are artifacts of EOC activity included. Table 1 shows the percentage of time spent by the sloths in each behavioral state and the time in each EEG classification and the relationship between the two. The means included in this table are for 10 animals, the eight of group II and the two more typical animals of group III, Nos. 1 and 4. The sloths showed one or another type of waking behavior during 30.67: of the observations made and behavioral sleep during 69.4%. Awake-exploring was the commonest of the observed waking behaviors and awake-alert was next. The awake-fixating behavior was relatively rare. According to the EEG records animals were awake about 34% of the time, in light sleep about 56% of the time, and in one or another of the types of deep sleep only about 10% of the time. The interesting point here was that the EEG record showed light sleep when the observer reported that





Recording of sloth showing deep sleep, S,.

Sleep and waking in the three-toed I. The relationship


EEG patterns Awake (A,, AZ) Light sleep (S,, S,) Deep sleep (S,, S,,



B. Awakealert N %

C. Awakefixating N “/

166 2

59 43

5 19

17.29 0.20

6.14 4.48


and electroencephalograms

A. Awakeexploring N %



Totals-behavioral patterns N-Number



D. Behavioral sleep N %

0.52 1.97

96 471








10 49.06

Totals-EEG patterns N 9/ 326 535

33.95 55.72









of observations, % of total.

the animal was showing a waking behavior and behavioral sleep sometimes was accompanied by a waking EEG pattern. The EEG patterns of deep sleep, however, were never seen when waking behavior was reported. Figure 10 and 11 show, respectively, the percentage of time spent in each behavioral state as related to the time of day and the percentage of time in each EEG state also related to time of day. Under conditions in this laboratory, sleep behavior predominated in the morning hours, 6 a.m.-12 noon, but was also encountered at times throughout the whole 24 hr period. Active waking time was fairly evenly distributed during the 18 hr from 12 noon to 6 a.m. The awake-

alert behavior too was distributed quite evenly over the afternoon and night periods. The awake-fixating behavior which constituted only a very small proportion of activity was seen chiefly around midday when the animals were waking from their long morning sleep period. The distribution of EEG activities correlated fairly well with the distribution of behaviors but not completely so. Sleep patterns occupied most of the morning period, ending a bit earlier than behavioral sleep and at the time when awake-fixating behavior most often occurred. Waking types of EEG patterns predominated in the afternoon and until midnight, but thereafter sleep and waking patterns were fairly

- = Behavloro,

___ = Awoke-exploring + = Awoke-flxatlng











12 Hours

Fig. 10. Circadian





rhythm in sloth behavior, expressed as percentage of total observations. made at I5 min intervals over a 24 hr period. Means of 10 animals.



= Awake = Light




S, E Sz :

Fig. 11. Circadian rhythm in EEG patterns of sloths expressed as percentage of total recording Recordings made for 90 set out of each I5 min period. Means of 10 animals.






evenly mixed. Deep sleep patterns were more common between midnight and noon than in afternoon or evening and proceeded from light sleep stages, not directly from waking EEGs. Table 2 shows the record for the four animals of group III with their various activities presented individually. From this it may be seen that the individual sloths varied considerably in the amount of time given to each EEG stage, that under certain circumstances deep sleep was totally lacking, and that there were no obvious differences based upon sex. The mean amount of waking EEG for this group was much more than was shown by group II, 56 vs 34?;,, in spite of the fact that No. 2 showed less than average waking time. As with the other group, group III showed a decided predominance of A, activity over drowsy A? activity, and lighter sleep patterns also predominated over those of deep sleep. Animal No. 3 failed to show any deep sleep. Even including animal No. 3, this group spent a greater proportion of sleep time in the deep stages, S?, S, and PS, than did group 11 animals. 22 vs 16’;,,, but the percentage of the total 24 hr day was about the same. Omitting animal No. 3, group III animals spent about l5?; of the day in PS, the mean duration of paradoxical sleep episodes was just under 8 min, the latency to PS was about 2 hr, and the intervals between episdoes averaged approx. 38min. Table 3 shows the heart and respiratory rates as they were related to the EEG states of three animals of group III. The unusually high heart rates of animal No. 2, previously noted, are very obvious and probably influenced the mean of this small sample unduly. In general heart rates diminish over the continuum from waking to paradoxical sleep. The differences between the means of waking and paradoxical sleep and between light sleep and paradoxical sleep are significant at the P < 0.05 level. There is also an apparent diminution in respiratory rates over the same continuum but the difference of the means is significant only for the rate during light sleep vs deep sleep. P < 0.05. DISCUSSION

The EEG patterns of the three-toed sloth were quite well developed and comparable to those of other orders of Eutherian mammals. A major difference lay in the lack of a well developed, relatively pure, high amplitude alpha rhythm such as is seen in primates, although there was low voltage activity in this frequency range mixed with other frequencies during A1 and S,. This is a deficiency shared by most mammals (Lippold, 1973). There was also a lack of true K-complexes although vertex sharp waves were prominent in S, and S, but Kcomplexes are also not frequently seen in lower mammals. The sleep spindles seen in the sloth were of a lower frequency (67 c/s) than is common, e.g. man 10-14 (Hill and Parr, 1963) cats 8-16 (Sterman et cd., 1965) guinea pigs 13-l 5 (Pellet, 1966) opossums 8-11 (Van Twyver and Allison, 1970) but only slightly lower than the 7-9 c/s seen by Van Twyver (1969) in several species of rodents. Originally it seemed possible that the relatively lower body temperature of sloths, 32 C,

Sleep and waking in the three-toed sloth


mights account for this slightly lower frequency of the sleep spindles, this in light of the fact that Putkonen and Sarajas (1968) had demonstrated that the sinusoidal olfactory spindles of rabbits varied in frequency directly with body temperature. But later this explanation appeared less probable when it was found that the frequency of the olfactory spindles of sloths, 35 Hz, was in the same general range as that of other mammals, e.g. opossums 354Oc/s (Van Twyver and Allison, 1970), dogs and monkeys [email protected]/s (Domino and Ueki, 1960), man 25-39 c/s (Sem-Jacobsen et ul., 1956; Gault and Leaton, 1963). The presence of low voltage, fast EEG activity in the complete absence of electrical activity in the nuchal muscles taken together with irregularity of heart and respiratory rates seemed to be clear indication of paradoxical sleep (PS), and the finding of some rhythmic eye movement seemed to justify also terming it REM sleep. It was to be expected that sloths would show PS because it had been seen earlier by Affani and his co-workers (Affani, 1972) in two other species of Edentates, Priodontes giganteus and Chaetophractus ~illosus, both types of armadillo. The mean time spent by the three-toed sloth in behavioral sleep, 69.40/i (Table 1) indicates either that more casual observers have tended to accord to sloths somewhat more inactivity than is warranted or that behavior in this regard varies considerably with the environment; probably the latter. The EEG records denoted sleep time as 66% of the 24 hr day. This was near enough to the behavioral observations to indicate that the observers were substantially accurate. Table 2 indicates that there was a very substantial individual variation with regard to recorded sleep; this r?ay have had to do with age and/or size, with the smaller and probably younger individual sleeping more, as studies of the ontogeny of sleep in other animals have shown (Jouvet et al., 1961; Roffwarg, et al., 1966). It is also possible that the reactions of individuals to changed environment and the circumstances of captivity differ substantially. The largest of group III animals slept materially less than the mean and failed to demonstrate any form of deep sleep. The mean time spent by group I animals in EEG sleep, 66”/,, is more than that spent by the adults of many groups of mammals, especially those usually considered to be phylogenetically “higher”, e.g. cats 58% (Sterman et a/., 1965), whales 22% (Shurley et al., 1969), various domestic ungulates 12-3.5:/, (Ruckebusch, 1972). It was about the same amount as has been seen in various species of rodent, including five species studied by Van Twyver (1969), 5&600/, and it was decidedly less than was seen by Affani in another Edentate, the giant armadillo, 75.4% or than was seen in opossums, 80.8% (Van Twyver and Allison, 1970). The lack of absolute agreement between observed behavior and the EEG record on specific occasions was interesting. A few times waking behavior was accompanied by the EEG of light sleep but more significant was the fact that animals apparently sleeping had waking EEGs about 10% of the time, indicating possibly that simulation of sleep may be a defence mechanism. Norton ef al. (1964) found that naive opossums showed EEGs indistinguishable from waking EEGs during periods of feigned sleep occur-



ing when the animal was under attack of trained dogs. There was never an EEG denoting deep sleep in an animal with waking behavior. Paradoxical sleep occupied about 11% of the total sleep time (Table 2) or, if the one animal that failed to show any form of deep sleep is omitted, a mean of 157;. This percentage was of the same general order as was seen by Pegram et al. (1970) in the Rhesus monkey and by Van Twyver (1969) in the Chinchilla. It was definitely less than has been seen in Marsupials, e.g., by Van Twyver and Allison (1970) in opossums, 29. I:‘,, or by AFFani (1972) in L. crossicauduta, 33.7%. It was somewhat less than was seen in some species of rodents, e.g. rats 19.5, hamsters 23.4 or squirrels 24.7 (Van Twyver, 1969). On the other hand it was more than has been seen in most primates, e.g. baboons 9% (Pegram et al., 1970). Thus it would appear that sloths, although relatively inactive, are not excessively “sleepy” nor that they are extremely specialized in their sleep patterns but rather it appears that they fit rather neatly into their usually assigned phylogenetic position as far as total sleep time is concerned. The factors governing PS seem to be less related to phylogenetic position but the sloth does not present a major deviation from the general pattern of inverse relationship between the amount of PS and phylogenetic rank. The mean duration of episodes of PS, approx. 8 min in the sloth, was longer than that seen in rodents by Van Twyver (1969). The range of mean durations for the three sloths (Table 2), 66l2min, was in the range seen by Michel et al. (1961) in rats. by Weitzman et (11.(1970) in macaques, by Shurley er al. (1969) in whales, and by Van Twyver and Allison (1970) in opossums. The mean interval, 37.8 min, between episodes of PS was longer than was seen by Van Twyver (1969) in rodents or than he and Allison found in opossums but slightly less than Weitzman et al. (1965) found in a species of monkey. The mean latency of 122 min to PS was long compared to that of the Rhesus monkey (Kripke et ul., 1968) and, of course, very long compared with animals like rodents which have very short sleep cycles. Figure 10 indicates that these sloths when in captivity were not, strictly speaking, crepuscular but were active throughout the afternoon and night, with the period of greatest activity between noon and 6 p.m., giving some credence to the Indian saying quoted by Goffart (I 971) that sloths travel “when the wind blows”, that is after 10 a.m. in the tropics. This could be because the wind brings the scent of fresh food leaves. In the period from 10 a.m. to 1 p.m. there was the highest level of the “awake-fixation” behavior and also this was the time when behavioral and EEG criteria were in least agreement. After 1 p.m. there was a high proportion of alert and exploring behavior. Actual waking behavior of any kind was seldom seen between 6 and 10 a.m. but patterns of both light and deep sleep were found oB and on throughout the whole 24 h day. This pattern of morning sleep and afternoon wakefulness seems somewhat unusual among animals. Thus in this species activity vs inactivity as related to light and dark may not have evolved as a defence against predators as has been postulated for some species. On the other hand as a matter of personal experience in

this laboratory in another series of experiments, not involving EEG recording, the sloths were in the laboratory with the investigators during the entire day, the animals were free to move about as they pleased, but they appeared to be asleep during the entire day, including the afternoons. The only obvious exception came when fresh food leaves were brought in and placed near the “sleeping” individuals. It is unfortunate that no EEG recording could have been done at this time but the difference in behavior has been interpreted here as an indication that the recording conditions for the EEG study were more nearly normal or natural and that the results might approach what could be expected of free ranging animals in the forests (unpublished data). As was pointed out earlier. one of the animals of group III, that is No. 2, had a very high heart rate, one seen by most previous investigators only in animals under considerable stress, e.g. Enger and Bullock (1965) give a rate of 114 beats/min for an animal “resisting”. The deviation seen here could have been the result of small size or of immaturity either of which in other species can lead to higher heart rates than are normal for adults of the species. However, Oliveira et ul. (1980) studied a group of small sloths of the same general size as No. 2, and found a mean heart rate of only 93 beats/min under rather stressful experimental conditions. Also Duarte et al. (1982). working with seated free animals including small individuals (weight range 2.334.7 kg), found a mean heart rate of only 76.4 beatsimin. Thus it appears that animal No. 2 in this study may have been an exceptional individual, possibly younger than size would indicate. Anitnals Nos. I and 4 had mean heart rates comparable to those found by Duarte et ul. All three individuals showed a tendency toward gradual reduction in heart rate in the continuum from A, through the several stages of sleep to PS. The rate was quite irregular during PS. for example in animal No. 2 at the time when the mean rate was 90 beats/mm the beat to beat rate varied from 83 to 108 bedts’min in a 2 min period. A degree of relative bradycardia has been found in other “lower” mammals during PS, e.g. it was found by Michel et ul. (1961) in rats. This is incontrast to the usual finding in man and probably in other primates where there is a tendency toward an increase in heart rate during PS. The irregularity, on the other hand, seems to be nearly universal, although Allison and Van Twyver (1970) found that moles showed a lower rate of heart activity during PS but a tendency toward greater rather than less regularity. No analysis for the heart rate specifically related to burst of eye movement was attempted in this study. The respiratory rates of the sloths also tended to decrease as sleep became deeper although the step by step change was significant only between light and deep sleep. In PS respiration was also irregular, a condition seen by Allison and Van Twyver ( 1970) in tnoles and which has also been noted in other animals. Because of the method of recording respiration no statement can be made concerning the depth of the breathing, In man the heart and respiratory rates tend to diminish generally over the whole night, somewhat independent of specific sleep stages but with periodic

Sleep and waking in the three-toed sloth increases in relative rate accompanied by irregularity during periods of REM sleep. This continued overall

decline in rates would not be expected in an animal having a polycyclic sleep pattern but might be observed in the sloth if only the longer morning sleep periods

were analyzed


REFERENCES Affani J. (1972) Observations on the sleep of some South American marsupials and edentates. In Sleep (Edited by Chase K. H.). Brain Information Service, Brain Research Institute, Los Angeles. Allison T. and Van Twyver H. (1970) Sleep in the moles, Scalopus aquaticus and Condylura cristata. Expl. Neurol. 27, 564578. Barrat E. S. (1965) EEG correlates of tonic immobility in the opossum (Didelphis uirginiana). Electroenceph. clin. Neurophysiol. 18, 709-71


Kripke D. F., Reite M. L., Tegram G. V., Stephens L. M. and Lewis 0. F. (1968) Nocturnal sleep in rhesus monkeys. Electroenceph. clin. Neurophysiol. 24, 582586.

Lippold 0. (1973) The Origin of the Alpha Rhythm. Churchill Livingstone, Edinburgh. Lundy W. E. (1952) The upside-down animal. Nor. ffist. 61, 114-119. Michel F., Klein M., Jouvet D., and Valatx J. L. (1961) Etude polygraphique du sommeil chez le rat. C. r. Seanc. Sot. Biol. 155, 2389-2392. Norton A. C., Beran A. V. and Misrahy G. A. (1964) Eie~troencephalograph during ‘feigned’ sleep in the opossum. Nuiure, Lond. 204, 162-163. Oliveira L. H. A., da Costa, C. P. and Huggins S. E. (1980) Cardiac mass, blood temperature and ventricular fibrillation: A study of the comparative physiology of the three-toed sloth and the domestic cat, Comp. Biochem. Physiol. 67A, 483-490.


Pegram V., Bert J., Naquet R. and Rhodes J. (1969) The Beebe W. (1926) The three-toed sloth Brudypus czieu~~iger ontogeny of EEG sleep patterns in the baboon. Psyehocuculiiger Wagler. Zoologira 7, l-67. physiol. 6, 228. Britton S. W. (1941) Form and function in the sloth. Q. Reo. Pellet J. (1966) Etude electropolygraphique et comBiol. 16, 13-34. portementale des &tats de veilie et de sommeil chez le Britton S. W. and Kline R. F. (1939) On deslothing the cobaye (Cavia porcellus). C. r. Seunc. Sot. Biol. 169, sloth. Science 90, 1617. 14761482. Britton S. W., Kline R. F. and Silvette H. (1938) BloodPutkonen P. T. S. and Sarajas H. S. S. (1968) Olfactory chemical and other conditions in normal and adrenalactivity vs brain temperature in hypothermic rabbits. ectomized sloths. Am. J, Physiol. 123, 701-704. Brain Rer. Osaka 10, 466469. Dement W. and Kleitman N. (1957) Cyclic variations in Roffwarg H. P., Muzia J. N. and Dement W. C. (3966) EEG during sleep and their relation to eye movements, Ontogenetic development of the human sleep-dream body motility and dreaming. Electroenceph. clin. Neurocycle. Science. 152, 604619. ohvsiol. 9. 673-690. Ruckebusch Y. (1972) Comparative aspects of sleep and wakefulness in farm animals. In The Sleeping Brain Domino E. F. and Ueki S. (1960) An analysis of the (Edited by Chase M. H.). Brain Information Service. electrical bursts phenomenon in some rhinencephalic structures of the dog and monkey. E~ectrae~eepll. clin. Brain Research Institute, Los Angeles. &m-Jacobsen C. W.. Petersen M. C.. Dodge H. W. rr ~europ~lysiol. 12, 635-648. Jacks Q. D., Lazarte J. A. and Holman-C. B. (19%; Duarte D. P. F., Da Costa D. P. and Huggins S. E. (1982) Electric activity of the olfactory bulb in man. Am. J. med. The effects of posture on blood pressure and heart rate in Sci. 232, 243-251. the three-toed sloth, Bradypus tridactylus. Comp. Biothem. Physiol. 73A, 697-702. Shurley J. T., Serafetinides E. A., Brooks R. E., Elsner Enger P. S. and Bullock T. H. (1965)Physiological basis of R. and Denny D. W. (1969) Sleep in cetaceans: 1. The pilot whale, Globicephala seammon. P.~~chophysiol. 6. slothfulness in the sloth. Hvalr~d. Sk. 48, 143-106. Gault F. P. and Leaton R. N. (1963) Electrical activity of 230. Snyder F., Hobson J. A., Morrison D. F. and Goldfrank F. the olfactory system. Electroenceph. chn. hteurophysiol. (1964) Changes in respiration, heart rate. and systolic IS, 299-304. blood pressure in human sleep. J. appl. Physiol. 19, G&art M. (1971) Function and Form in the Sloth. Pergamon 417422. Press, Oxford. Sterman M. B., Knauss T., Lehmann D. and Clemente C. Hill D. and Parr G., Eds. (1963) Electroencephafography. D. (I965) Circadian sleep and waking patterns in the Macmillan Co., New York. laboratory cat. Elertroen~eph. c/in. ~europhysioi. 19, Holmes 0. and Goffart M. (1966) Appendix. In Meulders 509-S 17. M., Gybels J., Bergman S. J., Gerebtzoff M. A. and Tirler H. (1966) A SIoth in the Family. Harvill Press. Gofhut M. (1966) Sensory projections of somatic, auditory and visual origin to the cerebral cortex of the London. sloth (Choloepus hofmanni Peters). J. Camp. Neurol. 126, Van Twyver H. (1969) Sleep patterns of five rodent species. 535-545. Physiol. Behav. 4, 901-905. Jouvet D., Valatx J.-L. and Jouvet M. (1961) Etude polyVan Twyver H. and Allison T. (1970) Sleep in the opossum, graphique du sommeil du chaton. C. r. Seanc. Sot. Biol. DidelphB marsupiulis. Electroenceph. chn. Neurophysiol. 155, 23892392.

Kreig H. (1961) Das Verhalten der Faultiere (Bradypodidae). In Handbuch der Zoologie 10, 20-23 (Edited by De Gruyter, W.). Berlin.

29, 181-189.

Weitzman E. D., Kripke D. F., Pollak C. and Dominguez J. (1965) In sleep of Mucuca mulatta. Archs. Neural. 12, 463467.