NOTE
Physiology
Daily Rhythmicity of Body Temperature in the Dog
R. REFINETTI1) and G. PICCIONE2)
1)
Circadian Rhythm Laboratory, University of South Carolina, Walterboro, SC 29488, U.S.A. and 2)Dipartimento di Morfologia,
Biochimica, Fisiologia e Produzioni Animali, Facolta di Medicina Veterinaria, Universita degli Studi di Messina, 98168 Messina, Italy
(Received 6 February 2003/Accepted 8 April 2003)
ABSTRACT.
Research over the past 50 years has demonstrated the existence of circadian or daily rhythmicity in the body core temperature
of a large number of mammalian species. However, previous studies have failed to identify daily rhythmicity of body temperature in
dogs. We report here the successful recording of daily rhythms of rectal temperature in female Beagle dogs. The low robustness of the
rhythms (41% of maximal robustness) and the small range of excursion (0.5°C) are probably responsible for previous failures in detecting
rhythmicity in dogs.
KEY WORDS: canine, circadian rhythms, rectal temperature.
J. Vet. Med. Sci. 65(8): 935–937, 2003
Circadian rhythmicity is a pervasive property of mammalian physiology. A circadian pacemaker located in the
suprachiasmatic nuclei of the hypothalamus is responsible
for an oscillatory process that is expressed in practically
every function in the body [9, 19]. In the dog, daily rhythmicity has been demonstrated in a variety of physiological
functions, such as heart rate, blood pressure, respiratory
rate, bone metabolism, heat dissipation, and rest/activity [2,
3, 5, 7, 11]. In most cases, no attempt was made to demonstrate that these rhythms are truly endogenous—that is, that
they are not only nycthemeral rhythms but also circadian
rhythms. Except for transient feeding-related elevations, at
least three research groups have failed to identify even a
daily (nycthemeral) rhythm of body temperature in dogs
maintained under a light-dark cycle [4, 6, 8]. Here we report
the successful recording of daily rhythms of rectal temperature in Beagle dogs.
Dogs (Purebred Beagle) were obtained from the canine
breeding facility at the University of Messina (Italy). Seven
female dogs (mean body mass: 10 kg) were used. They
were housed in individual pens (140 × 200 cm) lined with
wood shavings and were fed Rieper dog food once a day (4
hr after lights-on) and water ad libitum. Lights were on for
12 hr each day (600 lux from 0600 to 1800 hr). Ambient
temperature was thermostatically maintained at 21 ± 2°C.
Rectal temperature (3 cm deep) was recorded every 2 hr
with an electronic thermometer (Model HI-92740, Hanna
Instruments, Bedfordshire, UK) for 8 consecutive days.
The first day of data was discarded, and the following 7
days were used for analysis of rhythmic parameters: rhythm
robustness, mean level, range of excursion, and acrophase
(i.e., the time associated with the daily peak). Rhythm
robustness refers to the regularity of the oscillatory process
(i.e., how strongly the oscillations deviate from mere random variation). The robustness of the daily rhythm was
determined by the chi square periodogram [18]. The periodogram’s QP statistic (for P=24 hr) was calculated for each
individual using the 7 days of data (84 data points). For data
sets of this size, a perfect wave (such as a mathematically
generated cosine wave) produces a QP value of 84. Thus QP
values were expressed as a percentage of 84, which represents a perfectly stable wave form. QP values above 23% of
maximal robustness are statistically significant at the 0.05
level, and values above 29% are significant at the 0.01 level.
Although the chi square periodogram procedure was originally developed for data sets collected in hourly intervals
[18], we have previously validated its use in data sets with
either higher [15] or lower [14] resolution.
The mean level of the body temperature rhythm was computed as the arithmetic mean of the 12 daily measurements
(2-hr bins) for each individual. Likewise, the range of
excursion was computed as the difference between the highest and the lowest temperature each day for each animal.
The acrophase of the rhythm was calculated by the fitting of
a family of cosine waves to each daily segment of the
rhythm. The time of day corresponding to the peak of the
best-fitting wave was taken as the acrophase [10].
All animals showed robust daily rhythmicity in body temperature with QP values above the 0.01 significance level.
Representative records for one animal are shown in Fig. 1.
In this animal, as in all other animals, body temperature
ascended very gradually during lights-on and reached the
daily summit at the time of lights-off each day.
The mean daily values for all animals are shown in Fig. 2.
Although body temperature started its daily rise immediately after feeding (arrow), the daily peak was not reached
until the time of lights-off, 8 hr later. Thus, although dietinduced thermogenesis (the thermal effect of feeding) may
have contributed to the initial rise in body temperature, it
cannot account for the full daily oscillation.
The mean robustness of the rhythm (± SEM) for the 7
dogs over the 7 days was 40.6 ± 1.8 % of maximal robustness, which is small when compared with that of a variety of
other mammalian species [12, 16] but still above the significance level (23% of maximal robustness). The mean
acrophase (computed separately for each dog) was 17.2 ±
0.4 hr, which indicates that the peak of the daily oscillation
of body temperature was achieved approximately 11 hr after
936
R. REFINETTI AND G. PICCIONE
Fig. 1. Representative records of the body temperature rhythm of a dog over 7 consecutive
days. The data points correspond to 2-hr bins. The white and black bars above the graph
indicate the duration of the light and dark phases of the light-dark cycle.
Fig. 2. Mean (± SEM) values of body temperature of 7 dogs over
7 days. The data points correspond to 2-hr bins. The white and
black bars above the graph indicate the duration of the light and
dark phases of the light-dark cycle. The arrow indicates the time
of feeding.
tion can induce physiological responses that affect body
temperature. Also, group-housed animals are likely to be
disturbed by interactions with neighbours, and individuallyhoused animals kept in a busy animal colony with frequent
visits by animal caretakers are likely to be disturbed as well.
Although monitoring of body temperature by telemetry can
eliminate disturbance caused by the experimenters themselves, it can not eliminate disturbance caused by other
extraneous variables. We were not involved in the three
previous studies that failed to identify a daily rhythm of
body temperature in dogs [4, 6, 8], but we believe that a
rhythm would have been found if extraneous variables had
been effectively eliminated. As we have indicated elsewhere, various studies in sheep and swine without proper
control of extraneous variables failed to identify a daily
rhythm of body temperature, whereas studies with proper
control of extraneous variables identified moderate to robust
rhythmicity [13].
REFERENCES
lights-on, or 1 hr before lights-off, which is consistent with
the mean curve of body temperature seen in Fig. 2.
The mean level of the body temperature rhythm was
39.13 ± 0.01°C. This value is about 1°C higher than that
found in the previous studies that failed to identify nonfeeding-induced daily rhythmicity [4, 6, 8]. However, the
average rectal temperature of the dog is considered to be
38.9°C [1], which is only 0.2°C lower than the value we
obtained. The mean range of excursion in our study was
0.51 ± 0.01°C.
These results allow the addition of dogs to the extensive
list of mammals that display a daily rhythm of body temperature [17]. Our dogs exhibited statistically significant daily
rhythmicity in a constant thermal environment, and the daily
rise in body temperature was not caused by feeding. The
low robustness of the rhythm (41% of maximal robustness)
and the small range of excursion (0.5°C) are probably
responsible for previous failures in detecting rhythmicity in
dogs. When a weak rhythm is studied, noise introduced by
extraneous variables can easily mask the rhythmicity.
Uncontrolled variations in ambient temperature or illumina-
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Andersson, B. E. and Jonasson, H. 1993. pp. 886–895. In:
Dukes'Physiology of Domestic Animals (Swenson, M. J. and
Reece, W. O. eds.), Cornell University Press, Ithaca, NY.
Ashkar, E. 1979. Am. J. Physiol. 236: R231–R236.
Besch, E. L. and Woods, J. E. 1977. Lab. Anim. Sci. 27: 54–59.
Hawking, F., Lobban, M. C., Gammage, K. and Worms, M. J.
1971. J. Interdiscipl. Cycle Res. 2: 455–473.
Liesegang, A., Reutter, R., Sassi, M. L., Risteli, J., Kraenzlin,
M., Riond, J. L. and Wanner, M. 1999. Am. J. Vet. Res. 60:
949–953.
Marvin, H. N. and Reese, W. G. 1986. Physiol. Behav. 36:
903–906.
Matsunaga, T., Harada, T., Mitsui, T., Inokuma, M., Hashomoto, M., Miyauchi, M., Murano, H. and Shibutani, Y. 2001.
Am. J. Vet. Res. 62: 37–42.
Miyazaki, H., Yoshida, M., Samura, K., Matsumoto, H., Ikemoto, F. and Tagawa, M. 2002. Exp. Anim. 51: 95–98.
Moore, R. Y. and Leak, R. K. 2001. pp. 141–179. In: Circadian
Clocks (Takahashi, J. S., Turek, F. W. and Moore, R. Y. eds.),
Kluwer, New York.
Nelson, W., Tong, Y. L., Lee, J. K. and Halberg, F. 1979.
Chronobiologia 6: 305–323.
TEMPERATURE RHYTHM IN DOG
11.
12.
13.
14.
15.
Nishino, S., Tafti, M., Sampathkumaran, R., Dement, W. C.
and Mignot, E. 1997. J. Sleep Res. 6: 120–127.
Piccione, G., Caola, G. and Refinetti, R. 2002. Biol. Rhythm
Res. 33: 113–119.
Piccione, G. and Refinetti, R. 2003. Front. Biosci. 8: 258–264.
Refinetti, R. 1991. Int. J. Biomed. Comput. 27: 125–132.
Refinetti, R. 1993. Physiol. Behav. 54: 869–875.
16.
17.
18.
19.
937
Refinetti, R. 1999. Am. J. Physiol. 277: R1493–R1500.
Refinetti, R. and Menaker, M. 1992. Physiol. Behav. 51: 613–
637.
Sokolove, P. G. and Bushell, W. N. 1978. J. Theor. Biol. 72:
131–160.
van Esseveldt, L. E., Lehman, M. N. and Boer, G. J. 2000.
Brain Res. Rev. 33: 34–77.