Content uploaded by Yvan Touitou
Author content
All content in this area was uploaded by Yvan Touitou on Mar 09, 2018
Content may be subject to copyright.
Life Sciences 68 (2001) 1607–1615
0024-3205/01/$ – see front matter © 2001 Elsevier Science Inc. All rights reserved.
PII: S0024-3205(01)00966-3
Ramadan fasting alters endocrine and neuroendocrine
circadian patterns. Meal–time as a synchronizer in humans?
André Bogdan
a
, Belal Bouchareb
b
, Yvan Touitou
a,
*
a
Laboratoire de Biochimie, Faculté de Médecine Pitié-Salpétrière 91 boulevard de l’Hôpital, 75013,
Paris, France
b
Laboratoire d’Hormonologie, Hôpital Militaire Universitaire, Oran, Algeria
Received 2 June 2000; accepted 17 August 2000
Abstract
Muslims must refrain from eating, drinking, smoking, and sexual relations from sunrise to sunset
during the month of Ramadan. Serum concentrations of melatonin, steroid hormones (cortisol, testos-
terone), pituitary hormones (prolactin, LH, FSH, GH, TSH) and thyroid hormones (free thyroxin and
free triiodothyronine) were documented around the clock at six 4-hourly intervals before Ramadan be-
gan and on the twenty-third day of Ramadan (daytime fasting). Time series were analysed with re-
peated measures ANOVA. Statistically significant differences were found in some variables: the noc-
turnal peak of melatonin was diminished and may have been delayed; there was a shift in the onset of
cortisol and testosterone secretion; the evening peak of prolactin was enhanced, FSH and GH rhythmic
patterns were affected little or not at all by Ramadan fasting and only the serum TSH rhythm was
blunted over the test time span. These data show that daytime fasting, modifications in sleep schedule
and psychological and social habits during Ramadan induce changes in the rhythmic pattern of a num-
ber of hormonal variables. © 2001 Elsevier Science Inc. All rights reserved.
Keywords:
Rhythms; Synchronizers; Hormones; Fasting; Ramadan
Introduction
One of the most important rules of Islam is that any healthy adult Muslim must refrain
from eating, drinking, smoking, and sexual relations from sunrise to sunset during the month
of Ramadan, the ninth month of the Muslim calendar. Since this is a lunar calendar, the tim-
ing of this month of fast changes each year and the duration of restricted food and beverage
intake can vary from between twelve to sixteen hours. Intake is restricted to the night hours
within a short span of time, which thus delays sleep and reduces its duration. Environmental
* Corresponding author: Tel.: [
1
33] 1 40 77 96 63; fax: [
1
33] 1 40 77 96 65.
E-mail address
: touitou@ccr.jussieu.fr (Y. Touitou)
1608
A. Bogdan et al. / Life Sciences 68 (2001) 1607–1615
factors such as the timing of the rest-activity cycle [1] and meals [2–4] play a part in the syn-
chronisation of individuals to the 24-h day (and are accordingly known as synchronisers).
They modulate or modify one or several of the parameters characterising the circadian
rhythm of a biologic variable [5,6]. We have previously shown that changes in the circa-
dian rhythms of nutrition-related biological variables [7] occur during Ramadan. The pur-
pose of the present study was to assess whether these inseparable modifications ie. long-lasting
changes in food intake, sleep schedule, and psychological and social habits could modify the
pattern of a set of hormonal variables including cortisol and melatonin, and therefore it was
on purpose that no attempt was made to unmask the data.
Methods
Subjects
Volunteers meeting the inclusion criteria participated in this study after giving their in-
formed written consent. The Ethics Committee of the Faculty of Medicine of Oran approved the
protocol. Ten healthy non-smoking male volunteers (medical doctors in the hospital where
the study was carried on) were included in the study after routine clinical and laboratory ex-
aminations. Their ages ranged from 32 to 40 years (mean
6
SD
5
34
6
3.7 yr.). None took
any medication either before or during the study; none had any chronic or acute somatic or
psychiatric disorder; and none had taken a transmeridian flight within two months of the study.
Protocol
The volunteers were studied twice over a 24-h span: one week before Ramadan (control:
end of December) and on the twenty-third day of Ramadan (Ramadan: end of January). On
both test days, cortisol, melatonin, testosterone, free thyroxin (FT4), free triiodothyronine
(FT3), thyroid-stimulating hormone (TSH), growth hormone (GH), prolactin (PRL), luteo-
tropic hormone (LH), and follicle stimulating hormone (FSH) were measured from each of 6
blood samples drawn from the antecubital vein through an indwelling catheter at the follow-
ing clock hours: 0815 h, 1215 h, 1615 h, 2015 h, 0015 h and 0415 h. Blood sampled during
the rest time was drawn in the illumination of a dim red light (
,
30 lux). Samples were al-
lowed to clot, and the serum centrifuged, divided into aliquots, and stored at
2
20
8
C until
analysed. The hormones studied were chosen because of the well-known interaction between
the pineal/melatonin and the hypothalamic-pituitary / adrenal / gonadal / thyroid axes.
Before Ramadan started, the subjects were synchronised to nocturnal rest from 0000 h
6
0100 h to 0700 h
6
0100 h and to diurnal activity. On test days they were awakened at 0415 h,
to have their blood drawn. Physical activity did not differ qualitatively or quantitatively dur-
ing Ramadan, compared with the control day (they had the same tasks and working hours be-
fore and during Ramadan). During Ramadan, except for the test day, the subjects slept unin-
terruptedly from 0200 h to 0800 h. Their average sleep time was thus 1 h shorter during
Ramadan than it had been during the control period. Since the purpose of the study was to
look for changes related to Ramadan, the lighting conditions (uncontrolled) were those habit-
ual for the season and activity of the subjects who remained indoors during test days (same
domestic intensity for all). All meals were quantitatively and qualitatively standardised by a
A. Bogdan et al. / Life Sciences 68 (2001) 1607–1615
1609
nutritionist and were eaten at fixed hours that fit the subjects’ usual schedules and Ramadan
customs.
Meal timing and composition before Ramadan was:
0800 h: bread, butter, coffee and milk (
<
400 kcal)
1200 h: meat, vegetables, bread and fruit (
<
1300 kcal)
1900 h: soup, noodles, chicken (
<
1000 kcal).
Meal timing and composition during Ramadan was:
1900 h: milk, dates, soup, meat, vegetables, bread, fruit, coffee and pastry (
<
2000 kcal)
0001 h: semolina, milk, fruit and pastry (
<
600 kcal).
Biochemical assays
Melatonin was assayed directly by a modified version of the RIA method of Fraser et al.
[8], with a [
125
I]-melatonin tracer, as previously described [9]. The assay sensitivity was 5–10
pg/mL, and the only significant cross-reactant was 6–hydroxymelatonin (0.1%). The intra-
and inter-assay coefficients of variation were respectively 9% and 8% for a concentration of
50 pg/mL, 10% and 16% for a concentration of 200 pg/mL, and 9% and 15% for a concentra-
tion of 1000 pg/mL. All other hormone variables were assayed with commercial immunoflu-
orometry (Delphia™, Wallac, Finland). The intra- and inter-assay coefficients of variation,
respectively, were as follows: for cortisol (350 nmol/L), 4.05% and 6.7%; for testosterone
(20 nmol/L), 4.46% and 8.06%; for FT4 (19 pmol/L), 3.7% and 10%; for FT3 (5 pmol/L),
16% and 7.1%; for TSH (0.70
m
UI/L), 2.48% and 3.5%; for LH (5 UI/L), 4.79% and 8.3%;
for FSH (6 UI/L), 3% and 5.1%; for GH (0.4 mU/L), 6% and 8%; for PRL (5 ng/mL), 5.6%
and 9%. Each assay was performed in one large series at the end of the protocol to minimise
inter-assay variations.
Statistical analysis
Time series were analysed with a repeated measures analysis of variance (ANOVA; 2
within, 0 between) with SuperAnova™ software (Abacus Concepts, Inc. Berkeley, CA,
USA) to test the time-related variations (effect of time in both experimental conditions), the
influence of Ramadan upon the 24-h mean concentrations (Ramadan vs. control) and the pos-
sible interaction of Ramadan upon the time-related variations (experimental day and time
interaction).
Results
ANOVA (Table 1) shows statistically significant time-related variations (Time effect) for
serum melatonin, cortisol, testosterone, PRL, FSH, GH, TSH and a statistically significant
decrease in the 24-h mean level of serum melatonin and FSH during Ramadan (Ramadan vs.
control). Experimental day and time interaction showed that during Ramadan, compared
with the control period, time-related variations of serum melatonin, cortisol, testosterone,
PRL and TSH changed significantly (Table 1). Finally, no significant effect at all could be
found for LH, FT3 or FT4.
1610
A. Bogdan et al. / Life Sciences 68 (2001) 1607–1615
Cortisol
Figure 1 shows that during Ramadan serum cortisol levels rose in the afternoon, while the
morning rise was apparently delayed. This finding accords with the significant experimental
day and time interaction (Table 1). Moreover, a higher morning peak and a sharper decline
were observed during Ramadan.
Melatonin
Figure 1 shows that during Ramadan the serum levels of this hormone had a flatter slope and a
smaller night peak (p
,
0.008; Student’s paired t-test). This peak was also delayed, although we
could not observe any change in the time melatonin secretion began with the sampling frequency
used here. This agrees with the significant experimental day and time interaction (Table 1).
Table 1
ANOVA for repeated measures (2 within, 0 between). Biological variables documented before (Control) and on
the twenty third day (Ramadan) of daytime fasting during the month of Ramadan.
Effect Degree of
freedom F-Value P-Value
Cortisol nmol/L Ramadan vs Control 1 1.702 NS
Time 5 27.847 0.0001
Experimental day and Time interaction 5 3.990 0.0044
Melatonin pg/mL Ramadan vs Control 1 23.377 0.0009
Time 5 20.364 0.0001
Experimental day and Time interaction 5 4.429 0.0023
Testosterone nmol/L Ramadan vs Control 1 0.225 NS
Time 5 32.824 0.0001
Experimental day and Time interaction 5 10.044 0.0001
Prolactin ng/mL Ramadan vs Control 1 0.466 NS
Time 5 15.938 0.0001
Experimental day and Time interaction 5 9.789 0.0001
FSH UI/L Ramadan vs Control 1 5.748 0.0401
Time 5 4.028 0.0042
Experimental day and Time interaction 5 0.251 NS
LH UI/L Ramadan vs Control 1 0.331 NS
Time 5 2.188 NS
Experimental day and Time interaction 5 1.789 NS
GH mU/L Ramadan vs Control 1 0.193 NS
Time 5 3.306 0.0136
Experimental day and Time interaction 5 1.995 NS
TSH
m
UI/L Ramadan vs Control 1 2.348 NS
Time 5 6.003 0.0002
Experimental day and Time interaction 5 6.623 0.0001
FT3 pmol/L Ramadan vs Control 1 0.305 NS
Time 5 0.719 NS
Experimental day and Time interaction 5 1.582 NS
FT4 pmol/L Ramadan vs Control 1 0.449 NS
Time 5 1.077 NS
Experimental day and Time interaction 5 1.062 NS
A. Bogdan et al. / Life Sciences 68 (2001) 1607–1615
1611
Testosterone
Figure 1 shows an obvious delay in the evening trough of serum testosterone levels (0000 h
instead of 2000 h). Although the slope of the night increase in its concentration seemed
unchanged, the decreased morning slope of the control day was replaced by a 0800 h–
Fig. 1. Patterns of serum cortisol, melatonin, testosterone, prolactin and TSH on the control day (before
Ramadan) and on the twenty-third day of Ramadan. Each time point is the mean and SEM of 10 subjects.
1612
A. Bogdan et al. / Life Sciences 68 (2001) 1607–1615
1600 h plateau. These changes accord with the significant experimental day and time in-
teraction (Table 1).
Prolactin
The major changes in serum prolactin levels (Fig. 1) discernible during Ramadan were the
increase at 2000 h and the replacement of the 0400 h–0800 h plateau by a 0800 h peak. No
obvious changes were observed in either the increase or decrease slopes.
TSH
The decreased midnight and increased afternoon values (Fig. 1) observed on the twenty-
third day of Ramadan were responsible for flattening the TSH circadian rhythm which agrees
with the significant experimental day and time interaction validated by the ANOVA (Table 1).
Discussion
Ramadan is the month during which Muslims must refrain from eating and drinking from
sunrise until sunset while maintaining their usual social and occupational activities. These
long-lasting modifications—daytime fasting accompanied by a delay and shortening of
night-time sleep and changes in behaviour and social habits—have been shown to result in a
phase delay of many biological rhythms [10–13]. Recently, we reported the alteration of the
circadian patterns of plasma gastrin, insulin, glucose, calcium and gastric pH in both healthy
and healed duodenal ulcer patients, and strikingly the presence of a number of these alter-
ations one month after the end of Ramadan [7] which enforces the relevance to the debate
about the strength of meal-time as a rhythm synchronizer in humans.
The purpose of our study was to consider possible modifications in hormonal variables not
directly linked to the timing of nutrient intake, i.e., melatonin, cortisol, testosterone, free thy-
roxin, free triiodothyronine, TSH, LH, FSH and GH. Because this study was carried out to
examine whether these characteristics changed in humans observing the rule of Ramadan, we
did not attempt to change the subjects’ environmental conditions or behavioural customs (in-
cluding eating habits). Light exposure characteristics were the same during control and Ra-
madan periods except for the duration (one hour longer during Ramadan). Our study shows
obvious modifications in the rhythmic patterns of some of the variables we studied, even
though the sleep schedule of the subjects studied here was much less shortened and delayed
than in another study of Ramadan-induced changes in subjects who stayed awake till dawn
[11]. We must underline that: a) the subjects had the same tasks and working hours before
and during Ramadan; b) they took no kind of medication either before or during Ramadan; c)
none of them complained of any discomfort (including sleep disorders); d) their sleep sched-
ule was shortened and delayed only slightly (one hour on average); e) therefore the major
change in their routine was the redistribution of meal timing.
The characteristics of the serum cortisol rhythm observed on the control day are concor-
dant with those reported in the literature [6,14,15]. Nonetheless, although the mid-afternoon
rise of cortisol secretion, following the mid-day meal, is a well-known and described phe-
nomenon, it did not appear clearly among our subjects on the control day. On the Ramadan
test day, the cortisol rhythm was overtly biphasic, with an evident rise in the serum concen-
A. Bogdan et al. / Life Sciences 68 (2001) 1607–1615
1613
tration starting at 1200 h and a plateau between 1600 h and 2000 h, i.e., at the time of the first
meal following the daytime fasting period. In addition, the morning cortisol peak was higher
and steeper during Ramadan than on the control day. Such a rise, which appears to anticipate
the time of feeding, has been described in food-restricted rats with access to food only during
the light span [16] ie. outside of the normal activity span which is also the case for the sub-
jects of the present study. Our data are consistent with findings in subjects on a five-day total
fast (water only): they experienced delayed maximum serum concentrations and increases in
the mass of their glucocorticoid secretory bursts [17]. The differences between this study and
ours are most likely related to the differences between total and daytime fasting. On the other
hand, probably due to the number of subjects participating in our study, we did not observe a
significant increase in the 24-h mean concentration of serum cortisol. Similarly, a trend to-
wards increased peak values of serum cortisol at 0800 h during Ramadan was suspected but
was not substantiated with Student’s paired t-test.
The serum melatonin rhythm observed on the control day also agreed with our previous
data [15,18]. The melatonin pattern remained circadian during Ramadan even though 24-h
mean concentration and amplitude decreased. The 0400 h nighttime peak concentration de-
creased significantly (p
,
0.008; Student’s paired t-test) on the twenty-third day of Ramadan.
The onset and offset times of melatonin secretion did not change from the pre-Ramadan to
the Ramadan test date. The flatter slope of the melatonin increase may be due to the pro-
longed exposure to artificial light during Ramadan. This resulted in a peak maximum during
Ramadan, rather than the plateau maximum of the control night, and it suggests that in a
group of subjects who went to bed earlier during the control period, the Ramadan schedule
might have caused a phase delay of the nighttime serum melatonin peak. Again, we must em-
phasise that since both experimental days occurred in winter, exposure to indoor lighting dur-
ing Ramadan lasted only an hour longer than during the control period. To our knowledge, no
other reports on melatonin circadian patterns during Ramadan fasting are available.
The findings of no time-of-day variations for LH, free triiodothyronine and free thyroxin
on the control day are consistent with those in the literature [6], and Ramadan did not change
the pattern of these hormones. This is contrary to the report by Fedail et al. [19] who found
an increase in serum thyroxin in a single sample taken before the main evening meal on the
first and last days of Ramadan. This difference may well be due to the circadian basis of our
study. Moreover, the duration of the daytime fasting span during Ramadan varies substan-
tially according to the season in which it occurs. We therefore cannot rule out the possibility
that this daytime fasting and accompanying changes in sleep schedule, and psychological and
social habits influences the serum concentrations of LH, FT4 and FT3 when this lunar month
occurs in summer rather than winter and thus substantially delays the timing of the fast-
breaking meal.
The well-known circadian rhythm of serum TSH was found on the control day, but its am-
plitude was flattened during Ramadan.
The control day rhythm for serum testosterone is in agreement with previous data [19].
Ramadan did not modify the 24-h mean concentration but delayed the onset of the increase.
Prolactin serum concentration was the only variable that exhibited an overt biphasic pat-
tern on both the control and Ramadan test days, without any difference for the 24-h mean
levels, but an increased evening peak. Although the profile of this hormone is known to be re-
1614
A. Bogdan et al. / Life Sciences 68 (2001) 1607–1615
lated to sleep pattern, it is unlikely that the observed changes were caused by the one-hour
sleep change in this study, especially since none of the subjects complained of any sleep
disorder.
Finally, while serum concentrations of FSH and GH underwent significant time-related
variations, ANOVA failed to show any experimental day and time interaction, but did observe
a statistically significant (although weak) decrease during Ramadan for the 24-h mean serum
concentration of FSH.
Overall, then, this study has shown different kinds of modification during Ramadan for
variables considered as markers of the circadian system: melatonin maintained its circadian
rhythmicity but with decreased amplitude and a trend towards a phase delay and cortisol
maintained its circadian rhythmicity but with a biphasic pattern. These data indicate that
these rhythms may be controlled by different oscillatory components, as other authors have
suggested [21,22]. Indeed, the ten variables here studied expressed different circadian sensi-
tivities to the altered synchronizers which would support the theory of a hierarchy of oscilla-
tors, individuality of oscillators as an alternative to a single master clock. Ramadan is in es-
sence a seasonal external factor as it changes its annual location but besides the possible
changes in the induced effects according to the duration of fasting time, one cannot rule out a
superimposed seasonal variation of the sensitivity of the target rhythms (phase shifting, am-
plitude, mean level) to Ramadan.
In conclusion, our study found that observance of Ramadan has an impact on a number of
major metabolic endocrine processes. The sleep schedule of the subjects studied here was
shortened and delayed only slightly (one hour on average), therefore the major change in
their routine was the redistribution of meal timing. Although none of the subjects participat-
ing in this experiment complained of any discomfort, this kind of study ought to be repeated
and extended to other variables and other seasons to assess the possible effects on public
health of this month of daytime fasting and modifications in sleep schedule, and psychologi-
cal and social habits that concerns a substantial fraction of the human population.
Acknowledgments
We wish to thank Pr A. Wirz-Justice (Basel, Switzerland) and Dr A. Reinberg (Paris,
France) for their useful comments, Miss Oussalah for her efficient help, and our friends and
colleagues of the Hôpital Militaire Universitaire of Oran who volunteered for the study.
References
1. Apfelbaum M, Reinberg A, Nillus P, Halberg F. Rythmes circadiens de l’alternance veille/sommeil pendant
l’isolement. La Presse Médicale 1969; 77: 879–82.
2. Zigmond MJ, Shoemaker WJ, Larin F, Wurtman RJ. Hepatic tyrosine transaminase rhythm: interaction of en-
vironmental lighting, food consumption and dietary protein content. Journal of Nutrition 1969; 98: 71–5.
3. Halberg F. Protection by timing treatment according to bodily rhythms. An analogy to protection by scrub-
bing before surgery. Chronobiologia 1974; 1 [Suppl. 1]: 27–72.
4. Nelson W, Scheving LE, Halberg F. Circadian rhythms in mice fed a single daily meal and different stages of
lighting regimen. Journal of Nutrition 1975; 105: 171–84.
A. Bogdan et al. / Life Sciences 68 (2001) 1607–1615
1615
5. Reinberg A. Chronobiologie et Chronothérapeutique; Heure Optimale d’Administration des Médicaments.
Paris: Flammarion Médecine-Sciences, 1991.
6. Touitou Y, Haus E. Biologic Rhythms in Clinical and Laboratory Medicine. Berlin, New York, London, Paris,
Tokyo, Hong Kong, Barcelona: Springer Verlag, 1994.
7. Iraki L, Bogdan A, Hakkou F, Abkari A, Touitou Y. Ramadan diet restrictions modify the circadian time struc-
ture in humans. A study on plasma gastrin, insulin, glucose and calcium and on gastric pH. Journal of Clinical
Endocrinology and Metabolism 1997; 82: 1261–73.
8. Fraser S, Cowen P, Francey C, Arendt J. Direct radioimmunoassay for melatonin in plasma. Clinical Chemis-
try 1993; 29: 396–7.
9. Zhao ZY, Touitou Y. Differences in isoproterenol-stimulated melatonin production by perifused rat pineal
glands: time-dependent effects and role of enantiomeric form. Acta Endocrinologica 1994; 54: 497–502.
10. Bensouda JD, Ouazzani W, Lachkar A, Harniani A, Belkhadir J. Variations du cycle du cortisol sous l’influ-
ence du je ne au cours du mois de Ramadan. Symposium International 22–23 Octobre 1991 (Abstract), Rabat-
Maroc. Bases Actuelles de l’Exploration Hormonale. SEPE/FRH, 1982.
11. Al-Hadramy MS, Zawawi TH, Abdelwahab SM. Altered cortisol levels in relation to Ramadan. European
Journal of Clinical Nutrition 1987; 42: 359–62.
12. Aadil-Benchekroun N. Influence du je ne du Ramadan sur la pharmacocinétique de la théophylline et le cy-
cle du cortisol. Thèse de Doctorat, Université Claude Bernard, Lyon I, 1993.
13. Haouari-Ourerro F, Haouari M, Aouidet A, Sfaxi A, Hedhili A, Nagati K. Ramadan et rythmes circadiens des
taux sériques de la bilirubine des enzymes hépatiques et des oligo-élements (Mg
2
1
, Cu
2
1
, Zn
2
1
) Proc. of l
er
Congrès International Ramadan et Santé (Abstract), Casablanca, Maroc, 1994.
14. Touitou Y, Sulon J, Bogdan A, Touitou C, Reinberg A, Beck H, Sodoyez JC, Demey-Ponsart E, Van Cauwen-
berge H. Adrenal circadian system in young and elderly human subjects: a comparative study. Journal of En-
docrinology 1982; 93: 201–10.
15. Touitou Y, Benoit O, Foret J, Aguirre A, Bogdan A, Clodoré M, Touitou C. Effects of a two hour early awak-
ening and bright light exposure on plasma patterns of cortisol, melatonin, prolactin and testosterone in man.
Acta Endocrinologica 1992; 127: 189–92.
16. Challet E, Pévet P, Vivien-Roels B, Malan A. Phase-advanced daily rhythms of melatonin, body temperature,
and locomotor activity in food-restricted rats fed during daytime. Journal of Biological Rhythms 1997; 12:
65–79.
17. Bergendahl M, Vance ML, Iranmanesh A, Thorner MO, Veldhuis JD. Fasting as a metabolic stress paradigm
selectively amplifies secretory burst mass and delays the time of maximal nyctohemeral cortisol concentra-
tions in healthy men. Journal of Clinical Endocrinology and Metabolism 1996; 82: 692–9.
18. Touitou Y, Fèvre, M Lagoguey M, Carayon A, Bogdan A, Reinberg A, Beck H, Cesselin F, Touitou C. Age-
and mental health-related circadian rhythms of plasma levels of melatonin, prolactin, luteinizing hormone
and follicle-stimulating hormone in man. Journal of Endocrinology 1981; 91: 467–75.
19. Fedail SS, Murphy D, Salih SY, Bolton CH, Harvey RF. Changes in certain blood constituents during Ra-
madan. American Journal of Clinical Nutrition 1982; 36: 350–3.
20. Reinberg A, Lagoguey M, Cesselin F, Touitou Y, Legrand JC, Delassalle A, Antréassian J, Lagoguey A. Cir-
cadian and circannual rhythms in plasma hormones and other variables of five young healthy human males.
Acta Endocrinologica 1978; 88: 417–27.
21. Illnerova H, Vanecek J. The evening rise in the rat pineal N-acetyltransferase activity under various photope-
riods. Neuroscience Letters 1983; 36: 279–84.
22. Elliott JA, Tamarkin L. Complex circadian regulation of pineal melatonin and wheel-running in Syrian ham-
sters. Journal of Comparative Physiology 1994; 174: 469–84.
uˆ
uˆ
