Evidence for the adaptive significance of circadian rhythms


*Correspondence: E-mail: rgreen@vms.huji.ac.il


Circadian (∼24 h) clock regulated biological rhythms have been identified in a wide range of organisms from prokaryotic unicellular cyanobacteria to higher mammals. These rhythms regulate an enormous variety of processes including gene expression, metabolic processes, activity and reproduction. Given the widespread occurrence of circadian systems it is not surprising that extensive efforts have been directed at understanding the adaptive significance of circadian rhythms. In this review we discuss the approaches and findings that have resulted. In studies on organisms in their natural environments, some species show adaptations in their circadian systems that correlate with living at different latitudes, such as clines in circadian clock properties. Additionally, some species show plasticity in their circadian systems suggested to match the demands of their physical and social environment. A number of experiments, both in the field and in the laboratory, have examined the effects of having a circadian system that does not resonate with the organism’s environment. We conclude that the results of these studies suggest that having a circadian system that matches the oscillating environment is adaptive.


What are circadian rhythms?

Circadian (∼24 h) rhythms are found in a wide range of organisms (Bell-Pedersen et al. 2005). They are generated by a system that is self-sustaining in the absence of environmental cues, temperature compensated and can be entrained (synchronized) to the organism’s environment (Dunlap 1999; Bell-Pedersen et al. 2005). It has been suggested that the circadian clock serves as an adaptation for a rotating world by creating an opportunity for an organism to anticipate the daily external environmental changes and prepare accordingly. Moreover, the circadian system is used by many organisms for measuring photoperiod to regulate annual events (Imaizumi & Kay 2006). Although hourglass mechanisms that track time in a linear way may also be used for both anticipation and photoperiod measurement (Saunders 2005), to be effective such a mechanism would require constant synchronization. Interestingly it has been proposed that an hourglass mechanism may be a heavily damped circadian system (Saunders 2005).

Until now circadian systems have not been identified in Archaebacteria, but they exist in cyanobacteria which are Eubacteria and appear to be almost ubiquitous in multicellular eukaryotes (Dunlap 1999). Moreover, amongst multicellular eukaryotes, it seems that circadian systems developed independently several times during evolution. Among the numerous processes that are regulated by the circadian system are cell division in cyanobacteria, melatonin levels in birds, activity/rest cycles in mammals, conidiation in Neurospora and stomata opening and leaf movements in plants (Bell-Pedersen et al. 2005; Yakir et al. 2007). Underpinning most of these and other rhythmic processes is the circadian regulation of a substantial percentage of the genome, for example around one third of the plant genome and up to a quarter of the Neurospora genome (Covington et al. 2008; Dong et al. 2008), evidence for the adaptive role of the circadian system. Increasingly, it is being recognized that the circadian system is also important for coordinating internal metabolic events (Green et al. 2008; Wijnen 2009).

During the past few years, the design of the circadian system has been studied at molecular, cellular and physiological levels in a range of model organisms including the cyanobacteria Synechococcus (Dong & Golden 2008), the fungus Neurospora crassa (Dunlap et al. 2007), the plant Arabidopsis thaliana (Harmer 2009), the fruit fly Drosophila melanogaster and the mouse Mus musculus (Weber 2009). The classic, although rather simplified, model of a circadian system shown in Fig. 1 comprises three basic components the input pathways, the oscillator and the output pathways. It has been suggested that the complexity of a trait such as circadian system is an indication of its adaptive importance (Futuyma 1998) and, on a molecular level, circadian systems are complex. The oscillators that generate the ∼24 h rhythms are comprised of positive/negative feedback loops of pacemaker elements (Bell-Pedersen et al. 2005). Oscillators can be set or entrained via input pathways by signals from the environment such as diurnal changes in light or temperature (Roenneberg et al. 2003). At the same time, built into the system is a strong temperature compensation capability that allows the oscillator to function with a period that is close to 24 h under wide range of ambient temperatures (Bell-Pedersen et al. 2005). In turn, the oscillator regulates output pathways to control the diverse variety of circadian processes, including those outlined above. Interestingly, although the components of eukaryotic circadian systems may differ, many important features of their circadian mechanisms, for example the positive/negative pacemaker element feedback loops of the oscillator, appear to be conserved (Paranjpe & Sharma 2005). The widespread distribution of circadian rhythms together with the convergent biology of their mechanism strongly suggests that circadian rhythms in different species are the result of a response to selection forces and therefore have a significant adaptive value. At the level of the whole organism, the organization of the circadian system differs. For example, in plants it appears that almost every cell has its own autonomous circadian system (Thain et al. 2002). By comparison mammals have peripheral oscillators in tissues such as the liver, that are synchronized by a small group of neurons in the hypothalamus of the brain called the suprachiasmatic nucleus (Green et al. 2008). Thus, what is known about the design of circadian systems suggests that they are matched to their task of generating stable, persistent and temperature-compensated ∼24 h rhythms that match the organism’s environment.

Figure 1.

 The circadian system. Environmental signals such as daily changes in light or abrupt changes in temperature conditions are transduced via input pathways to entrain the oscillator. The oscillator controls many different output processes depicted here as three idealized rhythms with different phases. In diurnally changing conditions, the period of the oscillator and thus of the circadian output (the entrained rhythms) is 24 h. Under constant conditions, in the absence of environmental cues, the oscillator continues to run with a period that is close to 24 hours to generate the free-running rhythms of output processes. In the output, box the grey and white boxes represent dark and light.

What is adaptation?

There is considerable debate amongst evolutionary biologists, and numerous reviews, about the definition of the term ‘adaptation’ (Reeve & Sherman 1993). In general, the choice of definition tends to depend on the research question being addressed and its biological context. In some cases, adaptation has been considered from an historical perspective and defined as a characteristic that evolved in response to a specific selective agent (Gould & Vrba 1982). However, such a definition requires documentation of the history of a trait to compare the current use of an adaptive trait with its original function, information that is often unavailable. An alternative, non-historical, view of adaptation focuses on the current effects of a trait on reproductive success; ‘An adaptation is a phenotypic variant that results in the highest fitness among a specified set of variants in a given environment’ (Reeve & Sherman 1993). Fitness is most commonly defined as the average contribution of a genotype to future generations and is thus a measurement of reproductive success (Johnson 2005). In this non-historical definition of adaptation, if a trait is fitter than the alternatives it can be assumed that it out-competed alternative traits in the population and that its presence is likely to be explained by selection.

The adaptive significance of circadian rhythms

Given the widespread occurrence of circadian systems it is not surprising that extensive efforts have been directed at understanding the adaptive significance of circadian rhythms. It is certainly possible that circadian rhythms may not be adaptive. Circadian rhythms may have evolved as a correlation with another adaptive trait. For example, circadian rhythms may have been adopted to synchronize internal metabolism, even though rhythms other than circadian ones may be used to synchronize internal metabolism and be adaptive as well. Alternatively, a circadian system, like the human appendix, may have been adaptive only for an organism’s ancestors and been retained when no longer necessary as it is not deleterious.

In this review, we show the different experimental approaches, being used to examine the adaptive significance of circadian rhythms. The experiments described cover a wide range of organisms and include studies on individual organisms as well as on complex intra- and inter-species interactions and have sometimes given apparently conflicting results. The review is divided into two parts: (i) Examining whether variations in the properties of circadian rhythms in organisms are correlated with variations in their environment and (ii) Experimental studies that involve manipulating either the phenotype of the circadian rhythms and/or the environment and assessing the resulting effects.

Correlations with the environment

In a number of studies, researchers have used correlations between the circadian rhythm properties of an organism and its physical and social environment as a basis for studying the adaptive significance of circadian rhythms. Our rotating Earth provides a wide range of niches with extreme variations in diurnal conditions. There are habitats on Earth where there are no daily variations in the environment, for example in caves or at extreme latitudes during the winter and the summer. By contrast, other niches present conditions of extreme environmental changes between day and night, for example in non-polar deserts where 30 °C daily changes of temperature are common. There are also the predictable changes in day-length at different latitudes as well as different temperature conditions. In addition, many organisms, such as honeybees, can be affected not only by their physical environment, but also by their social environment.

Organisms in varied environments

Properties of circadian rhythms in the same species living at different latitudes have been useful in examining the correlation of circadian rhythms with the environment. Results obtained from populations of Drosophila auraria collected at latitudes ranging from 34° to 43° N in Japan, show a correlation between the period of circadian rhythms and latitude of origin of the organism, with longer circadian period individuals found at higher latitudes (Pittendrigh & Takamura 1989).

A mechanistic explanation for the connection between latitude and circadian rhythms may be polymorphisms in one or more clock genes. The clock gene period (per) was one of the first circadian clock genes to be discovered and, together with timeless (tim), is an essential component of the circadian oscillator mechanism in Drosophila spp. (Hardin 2005). Polymorphisms in threonine–glycine (Thr-Gly) encoding repeats in the per gene have been found in populations of D. melanogaster. These polymorphisms are likely to be adaptations for the temperature compensation capacity of the circadian system (Zamorzaeva et al. 2005). Consistent with the idea that there maybe a link between circadian genes and latitudes, there is a significant latitude cline in the distribution of the Thr-Gly polymorphisms in Europe and North Africa (Costa et al. 1992; Sawyer et al. 1997). More recently, further evidence (O’ Malley & Banks 2008) for a latitudinal cline in polymorphisms was found when the circadian gene OtsClock1b was examined in Oncorhynchus tshawytscha (Chinook salmon), strengthening the suggestion that circadian rhythms have an adaptive value.

Studies on clines in polymorphisms in circadian genes do not, however, always provide such clear confirmation of the adaptive value of circadian rhythms. For example, clines in polymorphisms may be hard to prove unequivocally. A cline in Thr-Gly polymorphisms in per genes in Australian populations of D. melanogaster was recently reported (Sawyer et al. 2006; Kyriacou et al. 2007). Such a cline would be particularly interesting in view of the fact that D. melanogaster was only introduced to Australia around 100 years ago (Umina et al. 2005). However, this Australian D. melanogaster cline has been debated by other researchers (Weeks et al. 2006). Moreover, not all latitudinal clines in polymorphisms necessarily reflect adaptation. European populations of D. melanogaster have two common alleles of the tim gene, ls-tim that expresses both long and short forms of TIM protein and s-tim that only gives the short form of the protein. The ls-tim polymorphism, which probably arose in southern Italy round 8000–10 000 years ago, causes attenuated photosensitivity of the circadian clock. Flies with the ls-tim allele show higher rates of diapause that may enhance their ability to survive unfavourable seasonal conditions (Sandrelli et al. 2007; Tauber et al. 2007). There is a significant latitude cline in the distribution of the tim polymorphism with more ls-tim in southern Europe and more s-tim in northern Europe. It has been suggested that this distribution cline of tim polymorphisms, rather than reflecting an adaptive advantage of the polymorphism to certain latitudinal conditions, may be a result of the relatively new ls-tim allele still spreading under directional selection from its place of origin (Tauber et al. 2007).

Additional studies that suggest that the parameters of an organism’s circadian system are correlated with its latitude of origin have been carried out in plants. Arabidopsis thaliana is a small annual plant from the Brassicaceae family, widely distributed throughout the northern hemisphere. When circadian-regulated leaf movements were monitored in collection of 150 natural accessions of A. thaliana, the period length of their endogenous circadian systems was found to positively correlate with day-length at their latitude of origin (Michael et al. 2003). In A. thaliana, as in many plants, the circadian system has a key role in determining the timing of the switch from vegetative to reproductive growth to synchronize flowering with favourable conditions under different day lengths at different latitudes. In Glycine max (soybean) it was found that the circadian expression of the blue light receptor (CRY1) in the input pathway of the clock system is correlated with latitude and also with flowering time (Zhang et al. 2008). Thus, the correlation between the circadian system function and latitude might sometimes be a result of selection for the timing of flowering as well as for temperature compensation, as in D. melanogaster.

Organisms originating from locations at different altitudes have been used to examine the importance of circadian rhythms in organisms that live in conditions with similar day-lengths but different temperatures. For instance, heat tolerance has been examined in populations of the cactophilic fly, Drosophila buzzatii, from North-Western Argentina that originated from different altitudes (Sorensen & Loeschcke 2002). Heat tolerance in the flies was found to vary across the day. A difference was found in timing of the daily peak in heat-tolerance between D. buzzatii originating from the hot lowland and the flies originating from the cooler highlands. In the highland population of flies, heat tolerance both in the afternoon and early evening was significantly lower than in the lowland population. Although the authors of the study did not show differences in circadian rhythmicity between the different fly populations, the timing of the heat tolerance change was controlled by light phase suggesting that the circadian clock may be involved in environmental stress resistance in D. buzzatii. Interestingly, circadian regulation of the timing of resistance to temperature stress has also been demonstrated in plants (Rikin et al. 1993; Fowler et al. 2005). Moreover, genes involved in stress responses are significantly over-represented in the 30% of A. thaliana genes that are circadian regulated (Covington et al. 2008). These results suggest that circadian regulation of stress responses pathways may have an adaptive significance.

Organisms in constant environments

Caves are useful environments for studying conditions in which light, temperature and humidity remain constant throughout the day and night. Some cave-dwelling animals after living in constant darkness for several thousand generations have lost their pigmentation and show degenerate eyes. One such animal is the dull-white coloured Glyphiulus cavernicolus (a cave-dwelling millipede). Monitoring locomotor activity in individuals of G. cavernicolus demonstrated that after entrainment with light/dark cycles, between 56% and 66% of the millipedes showed some circadian rhythms in locomotor activity (Koilraj et al. 2000). These results suggest that circadian rhythms may be important even in constant conditions. The fact that up to half of the individual millipedes did not show circadian rhythms in locomotor activity might be explained by locomotor activity not being a consistently reliable marker for circadian rhythms. Thus, although many of the millipedes did not show circadian locomotor activity, they may still have had functional circadian oscillators. When circadian rhythms were studied in the pigmentation-less Astyanax mexicanus (blind cave fish) using the movement of cones and rods in the retina as a marker for rhythmicity, it was observed that all of the fish were rhythmic (Luis Espinasa 2006). Therefore, at least some cave-dwelling animals clearly retain rhythmicity. We can conclude from these studies, that either circadian rhythms are too beneficial for organisms to be allowed disappear, perhaps because they are important for internal synchronization of physiological processes, or that they are simply under a much slower process of extinction than other traits such as pigmentation or eyes (Jeffery 2005). There may be a lower selection against the circadian system than other adaptations to light/dark environments due to the low ‘cost’ of rhythms by comparison with the ‘cost’ of eyes and pigmentation.

Many species of subterranean mammals show adaptations to inhabiting a niche where daily fluctuations in temperature and humidity are minimal. Such species are of special interest for studying the adaptive significance of the mammalian circadian system. Heterocephalus glaber (naked mole rat) is a blind subterranean eusocial mammal that shows a reproductive division of labour. Heterocephalus glaber live in colonies averaging ∼60–70 individuals, each colony includes a single breeding female and up to three breeding males. Research has shown that individual H. glaber living in a colony generally do not show circadian rhythmicity of locomotor activity. The exception to this lack of locomotor rhythmicity was observed in individuals that were apparently ready to disperse above ground. However, in laboratory experiments on isolated individuals, ∼65% of the H. glaber were found to present robust circadian rhythms in locomotor activity in constant darkness following entrainment to light:dark cycles (Riccio & Goldman 2000). Thus, it seems that most of the H. glaber retained a functional circadian system but whether or not the animals exhibited circadian rhythmicity in their natural environment depended on their status. Further studies have demonstrated that other species of both solitary and social subterranean mole-rats show rhythmicity after entrainment to alternating light/dark cycles (Oosthuizen et al. 2003). Why have these subterranean animals retained functional circadian systems when they live in an almost constant physical environment? One possibility is that, as has been suggested for organisms that live in caves, circadian systems may be an important adaptation for coordinating internal metabolic processes or for regulating seasonal breeding. As cues for entraining their circadian systems they may be able to use small changes in, for example temperature or humidity, in their environment or social cues. Another possibility is that a functional circadian system is still important for the timing of the animals’ above-ground forays, even if they happen rarely, to minimize predation.

Organisms living in both constant and varied environments

Organisms that live in environments in which they experience alternating times of constant conditions and diurnally changing physical and social conditions are useful as they enable researchers to examine the plasticity of circadian rhythms in response to changes in the environment in the same organism and thus to understand whether circadian rhythms correlate with different conditions.

At high latitudes there is extreme variability in annual day-length from constant dark in winter to constant light in the summer. In a study to determine the effect of constant and diurnally changing conditions on circadian rhythms, locomotor activity was examined in two populations of reindeer, Rangifer tarandus platyrhynchus and R. t. tarandus (van Oort et al. 2005). The northern reindeer population, R. t. platyrhynchus, that live at a latitude of 78° N displayed distinct circadian locomotor activity during autumn and spring but attenuated circadian activity during the summer and winter periods when light conditions are constant. By contrast, the southern reindeer population, R. t. tarandus, that live at a latitude of 70° N showed stronger oscillations in circadian rhythms of activity for longer periods of the year (van Oort et al. 2007). The results suggest that the conditions of constant light during the summer at higher latitudes seem to favour weaker circadian regulation of activity in the reindeer. These results confirm an earlier study showing that daily oscillations of levels of the hormone melatonin in the high-arctic Lagopus mutus hyperboreus (Svalbard ptarmigan) are attenuated in the summer and winter months (Reierth et al. 1999). Such results demonstrate that circadian rhythms can be correlated with the environment. Thus, there is an adaptive advantage to having circadian activity if daily conditions are fluctuating. However, in constant environmental conditions, there may be no advantage to restricting activity to certain times of day. In general, it must be borne in mind that the loss of a particular circadian rhythm or rhythms does not necessarily mean that the circadian system itself is no longer functioning. For example, even though some reindeer show attenuated rhythms of activity at certain times of year, it is likely that their circadian oscillators are functioning normally to control other processes, such as metabolism, but have become partially uncoupled from the animal’s locomotor activity.

Animal societies with division of labour also provide an opportunity to study circadian rhythm plasticity as individuals from the same species with different tasks function under different social and environmental conditions. Researchers have used the eusocial Apis mellifera (honeybee), which presents an age-related division of labour, to study the correlation between age, division of labour and circadian rhythms. Apis mellifera larvae require constant care and young arrhythmic nurse bees work around the clock to provide it. By contrast, the older rhythmic foragers forage for nectar and pollen during the day and display sleep-like behaviour at night (Moore 2001). Foragers have a highly developed internal circadian clock that is used for sun compass navigation, dance communication and timing of visits to flowers for maximum nectar and pollen availability (Gould & Gould 1988). The differences in rhythmic behaviour between the young arrhythmic nurse bees and the older rhythmic forager bees were found to be associated with both increased levels and oscillations of per mRNA, a component of the A. mellifera circadian clock (Bloch et al. 2001). Moreover in A. mellifera, old foragers lost their robust circadian activity when forced to work as nurses demonstrating the enormous plasticity in circadian phenotype (Bloch & Robinson 2001). Thus, there is a strong correlation between the age that determines division of labour in the A. mellifera and their circadian phenotype.

In the eusocial Bombus terrestris (bumblebee), division of labour is based mainly on size rather than age as in the honeybee (Goulson et al. 2002). Bombus terrestris show a large variation in body size. The bigger bees become foragers and exhibit robust circadian rhythms, foraging during the day and resting during the night. By contrast, the smaller bees work as nurses around the clock and show attenuated circadian rhythms (Yerushalmi et al. 2006). When freshly enclosed bees were monitored individually, small bees did develop circadian rhythmicity but at a significantly later age than the larger bees (Yerushalmi et al. 2006). Large bees also show elevated levels of pigment dispersing factor, a neuromodulator in the insect circadian system (Weiss et al., 2009). Together with these difference in circadian properties, other features make the large bees better suited for foraging such as the organization and function of the visual and chemosensory systems in B. terrestris (Brockmann et al. 2003; Spaethe & Chittka 2003) and improved learning performance in B. impatiens (Worden et al. 2005). Therefore, it has been suggested that circadian rhythms are part of a group of adaptations that provide an advantage to forager bees (Yerushalmi et al. 2006).

Similarly, the circadian clocks of another social insect, Camponotus compressus (black common ant), are flexible, and appear to be task-dependent. Queen C. compressus exhibit robust circadian rhythms prior to mating. These rhythms may have a role in the timing of mating flights which are species-specific and may be important for creating reproductive isolation between species (Tauber et al. 2003). After mating, the queen looses her wings and starts to lay eggs to establish a new colony. During the period of egg laying, the queen shows arrhythmic locomotor activity. However, when the queen finishes egg laying activity she regains circadian locomotor activity (Sharma et al. 2004). Taken together, the research on ants and bees shows that the properties of an individual organism’s circadian rhythms may be correlated with environmental constraints imposed by the division of labour in the organism’s society.

Experimental approaches

Experimental approaches, including altering the environmental and/or the endogenous clock properties of the organism and assessing the effects have been valuable for testing the adaptive significance of circadian rhythms. The importance of the impact of the environment on circadian rhythms is well known in humans. Every time we fly across several time zones or work shifts, we experience the effects on our circadian rhythms of altering the environment, often with adverse effects on our health (Lack & Wright 2007). Most experiments in other organisms have examined adaptation by measuring indirect parameters of fitness such as longevity in flies, survival in chipmunks and size in plants. However, a few approaches have started to address the evolutionarily important question of whether circadian rhythms confer direct fitness.

Altering the environment

On Earth the days are 24 h long but in the laboratory it is possible to generate unusual day lengths (T-cycles). Measuring survival and longevity in altered T-cycles has been used as a technique to demonstrate the consequences of non-matching circadian rhythms. When wild-type D. melanogaster were reared under varied T-cycles or constant light, it was observed that survival rates were highest for both males and females grown under a T-cycle of 24 h (Pittendrigh & Minis 1972). This result suggests that survival is influenced by the match between the endogenous clock system and the environmental T-cycles. A similar result (von Saint Paul & Aschoff 1978) was observed with Phormia terraenovae (blowflies). The mismatch between the endogenous clock and the environment may reduce the efficiency of many physiological processes and thus shorten longevity.

Although adaptation has been defined as ‘a phenotypic variant that results in the highest fitness among a specified set of variants in a given environment’ (Reeve & Sherman 1993), the connection between circadian rhythms and fitness has only been found directly in a few experiments. Using altered environmental T-cycle conditions, Emerson et al. (2008) measured fitness directly in Wyeomyia smithii (pitcher plant mosquito). Larvae from natural populations of W. smithii were synchronized by growth under short-day conditions. Then, on the day they pupated, they were moved into conditions of 24-, 35- or 46-h T-cycles. The pupae, resulting adults and their eggs were maintained under the same conditions until the eggs hatched. To measure how the different T-cycle conditions affected fitness, the average fecundity per resulting female adult was determined. Fecundity was significantly lower for females from the 35 h T-cycle, the most remote from the ∼24 h circadian period. By contrast under the 46 h T-cycle that is closely related to a multiple of the circadian period, fecundity was higher. However, pupal survival, embryonic viability and adult longevity did not differ between the different treatments (Emerson et al. 2008). The results strongly suggest that having an endogenous pacemaker that matches the environment confers fitness in W. smithii.

Growth under constant conditions has been used to test whether circadian rhythms have an adaptive significance. When a D. melanogaster population was maintained for more than 700 generations under constant light conditions it was observed that there were no differences between the circadian rhythms of the original and final populations (Paranjpe et al. 2003). These results suggest that the environment did not affect the circadian clock and appear to be inconsistent with the observation that only 100 years after their introduction, there may be a cline in the distribution of polymorphisms of a key circadian gene in Australian D. melanogaster (Sawyer et al. 2006). However, the mechanism of natural selection requires both genetic variance and selection force (Futuyma 1998). Thus, if there is no genetic variance in the population or if the selection force is weak, it is possible that the rate of random mutations in the D. melanogaster population used by Paranjpe and co-workers would not have been high enough to show an effect after 700 generations.

Organisms are not only affected by their physical environment, but also by their social environment and scientists have examined the effects on circadian rhythms of altering social environment. In the hive, young honeybee nurses show attenuated circadian rhythms of both locomotor activity and per mRNA levels, even under light:dark regimes. However, when individual nurse bees were isolated in cages, they rapidly developed robust circadian rhythms (Shemesh et al. 2007). These experiments provide strong evidence that animals have evolved a remarkable degree of plasticity in their circadian systems that enable them to contend with their environmental conditions.

Altering the social environment of pre-social animals that live in simple societies can also affect circadian rhythms. Individual and group-living D. melanogaster were synchronized for 5 days to the same light/dark cycles. After 2 weeks in constant dark, flies living in a group still showed significantly more synchronization of their circadian-regulated locomotor activity than isolated flies (Levine et al. 2002). When flies with an altered phase or period were introduced to a host population, the synchronization of the host group was affected. The social influence on circadian rhythms is mediated mainly by chemosensory pathways, although auditory and tactile cues may also contribute (Levine et al. 2002). It has been suggested that such population synchronization of circadian rhythms may facilitate environment matching of individual flies or newborn larvae or be used for synchronizing mating (Tauber et al. 2003).

Work has also shown that intra-species interactions and constraints of competition and niche partition can affect the circadian system. Rattus norvegicus (Norway rats) are generally nocturnal. However in the presence of Vulpes vulpes (red fox), which is nocturnally active, R. norvegicus become diurnally active. Interestingly, rats in fox-proof enclosures reverted back to being nocturnally active (Fenn & MacDonald 1995). Similarly, Acomys russatus (golden spiny mice) show a diurnal activity pattern in the presence of Acomys cahirinus (common spiny mice) but a nocturnal activity pattern in their absence (Gutman & Dayan 2005; Levy et al. 2007). A combination of field and laboratory experiments showed that the different activity pattern observed in A. russatus was probably caused by changes downstream of the internal clock (Gutman & Dayan 2005; Levy et al. 2007). Together, these results provide evidence that the circadian system may be utilized to define the niche of a species and to enable better adaptation to the environment by avoiding predation and reducing competition.

Altering phenotype

Examining the effect of altering the endogenous circadian phenotypes of organisms has been a useful tool with which to study the adaptive significance of circadian rhythms. Many insects, including D. melanogaster, show daily rhythms in mating activity (Sakai & Ishida 2001). Using combinations of male and female, wild-type and arrhythmic mutants, it was found that females with mutations in circadian genes show arrhythmic mating behaviour even when paired with wild-type males. Thus the females are responsible for determining the mating rhythms (Sakai & Ishida 2001). Another closely related Drosophila species, D. simulans was found to have a courtship rhythm that was antiphasic to D. melanogaster suggesting that the female and species-specific circadian rhythms in Drosophila spp. mating patterns may be important for reproduction isolation (Sakai & Ishida 2001).

One important potential pitfall in experiments that utilize genetic mutants with altered phenotypes is that many genes, including circadian genes, have pleiotrophic effects. Beaver et al. (2002) found that D. melanogaster with mutations in the circadian genes Per, Tim, Clk and Cyc, showed dramatically reduced fertility in comparison with the wild-type. The quantity of sperm was reduced in males with mutated Per and Tim, and Per and Tim mutant females produced nearly 50% less oocytes and progeny (Beaver et al. 2002). However, in female D. melanogaster, Per and Tim in the ovary do not show circadian oscillations or respond to changes in light (Beaver et al. 2003). Moreover, other mutations in the circadian system that do not cause decreases in Per levels, have no effects on oocyte production. Therefore, the reduction in fertility observed in the Per, Tim, Clk and Cyc mutants may not be a result of a direct fitness advantage of having a functional circadian system but may indicate a pleiotropic effect of the clock genes in the reproductive system. We can conclude from these results that experiments in which phenotypes are altered genetically must be carefully interpreted to distinguish between the effects of the mutations on the circadian system and on other processes in the organism.

Many of the experiments described above have been carried out under laboratory conditions in the absence of selection pressures that organisms are normally subjected to in nature (Calisi & Bentley 2009). To overcome this problem, an ambitious project examined the effects of altering circadian phenotype on animals in their natural setting by introducing lesions in the suprachiasmatic nuclei of Tamias striatus (chipmunks) causing the chipmunks to loose their behavioural rhythmicity (DeCoursey et al. 2000). The lesioned T. striatus and control groups of surgically treated but non-lesioned and untreated animals were released to the wild and monitored using radio-collars. During the first 2 weeks of the experiment, T. striatus from both of the surgically treated groups were lost to predators at a higher rate than the untreated controls. However after this initial period, arrhythmic T. striatus were significantly more susceptible to predation, probably by weasels (DeCoursey et al. 2000). The authors suggest that night-time restlessness of the arrhythmic T. striatus resulted in elevated detection rates by weasels. Thus, it appears that functional circadian systems may provide the opportunity for an animal to restrict its activity to certain times of the day to improve its chances of survival.

Altering both the phenotype and the environment

Experiments in which both phenotype and environmental conditions are altered provide some of the strongest evidence for the adaptive significance of circadian rhythms. With its small size and large available collection of circadian mutants, A. thaliana has proven a useful organism with which to study the effects of altering both circadian phenotypes and the environment. When arrhythmic and wild-type A. thaliana were grown under extremely short-day conditions, it was observed that the arrhythmic plants were less viable than the wild-type plants (Green et al. 2002). Very short-day length conditions require maximum utilization of the availability of light for photosynthesis and it is possible that the absence of a functional circadian system may prevent the plant from anticipating the onset of light in the morning that enables them to maximize their potential for photosynthesis during the photoperiod (Green et al. 2002). In further experiments, it was shown that arrhythmic A. thaliana mutants fix less carbon than wild-type plants under T-cycles of 24 h. However, in a constant environment the clock may be disadvantageous; arrhythmic A. thaliana mutants have higher rates of CO2 assimilation than wild-type under constant light conditions (Dodd et al. 2005). Competition experiments were carried out with A. thaliana to examine the effect of having an endogenous clock that resonates with the environment. Short- and long-period A. thaliana mutants were sown densely under 20 and 28 h T-cycles. When parameters of chlorophyll content, leaf number, rosette diameter, and aerial biomass were measured, it was observed that the short-period mutant performed better under 20 h T-cycles whereas the long-period mutant performed better under 28 h T-cycles (Dodd et al. 2005). In addition, plants with circadian rhythms that did not resonate with the environment showed higher mortality rates (Dodd et al. 2005). Such experiments provide a good example of the adaptive significance of having a circadian system that resonates with the environment.

Competition experiments in cyanobacteria have supplied some of the most convincing evidence for the adaptive significance of circadian rhythms. During the course of natural selection, the selection force increases the frequency of the organisms in a population that are best adapted to the environment (Reeve & Sherman 1993). Although this is a long-term process, by applying a strong selection force, such as competition to a heterogeneous population it may be possible to compare the impact on the frequencies of different phenotypes in laboratory conditions. Pioneering research of Johnson and colleagues (Ouyang et al. 1998) made use of Synechococcus elongatus (a freshwater cyanobacterium) with different alleles of key circadian genes that resulted in free-run periods of 23 h (SP22), 30 h (P28) and 25 h (wild-type strains AMC149 and AMC343). The strains were grown under T-cycles of 22 and 30 h and the numbers of S. elongatus from each strain were counted. When the long- and short-period mutants were grown together in competition experiments for 27 days in short T-cycles, the short-period mutant performed better than the long period mutant. By contrast, in long T-cycles, the long period mutant outperformed the short period mutant. In competition assays between short-period and wild-type S. elongatus, the wild-type performed worse under 22 h T-cycles, but better under 30 h T-cycles. Moreover, in competition experiments between wild-type and arrhythmic S. elongatus, the wild-type grew better in 24 h T-cycle, but the arrhythmic strain out-competed the wild-type in constant light (Woelfle et al. 2004). When grown in isolation either under constant light or in different T-cycle conditions, the growth rates of the strains were indistinguishable. These results strongly suggest that S. elongatus with a circadian clock that resonates with their environment have an increased competitive advantage under diurnally changing conditions, but that under constant conditions there is less of advantage to having a circadian system and it may even be detrimental. Taken together, competition studies suggest that in both S. elongatus and A. thaliana, the existence of an oscillator itself is not adaptive if the oscillator does not produce rhythms that match the environment. However, in an oscillating environment having an oscillator that resonates with the environment is adaptive.

Conclusions and perspectives

Circadian rhythms are found in a wide variety of organisms. Indeed it is likely that all eukaryotes have circadian rhythms. Even in organisms for which rhythms have not yet been identified, the lack of demonstrable circadian rhythmicity does not rule out the possibility of there being a functional circadian oscillator driving as yet unidentified rhythms. It is, however, possible that there are some eukaryotic organisms that lack circadian systems – and it would be fascinating to identify them. Over the past two decades there have been tremendous advances in our understanding of the molecular mechanism of the circadian system in a number of model organisms (Dunlap et al. 2007; Dong & Golden 2008; Harmer 2009; Weber 2009) that may be useful in determining whether or not other organisms have functional circadian systems. As an example of such an approach, Avivi and coworkers identified per genes in H. glaber and showed that they oscillate with a period of around 24 h proving that, despite the fact that H. glaber often does not show circadian rhythms of locomotor activity (Riccio & Goldman 2000), it has a functional circadian system (Avivi et al. 2002).

As we have shown in this review, circadian rhythms have roles in such varied processes as minimizing predation (Fenn & MacDonald 1995), coordinating mating (Sakai & Ishida 2001; Tauber et al. 2003), internal synchronization (Green et al. 2008) and division of labour (Bloch et al. 2001; Sharma et al. 2004; Yerushalmi et al. 2006). Accordingly, the question of whether circadian rhythms have an adaptive significance has intrigued scientists over the years and prompted studies involving different experimental approaches.

Numerous studies have shown that circadian rhythms tend to match the environment of the organism. This matching can be seen on a number of levels. At the most basic level, the general design and complexity of circadian rhythms in every organism that has been studied suggest a trait that has been forced by natural selection to fit with the environment (Futuyma 1998). On a species and population level, differences can be seen in the circadian systems of organisms from different environments. For example, clines have been established over generations in the properties of the circadian systems and latitude of origin of A. thaliana and D. auraria (Pittendrigh & Takamura 1989; Michael et al. 2003). On the level of individual organisms, some animals, such as R. tarandus (van Oort et al. 2005), H. glaber (Riccio & Goldman 2000) and A. mellifera (Bloch & Robinson 2001) show plasticity in their circadian system and can switch from arrhythmic behaviour to rhythmic behaviour, or, in the case of R. norvegicus (Fenn & MacDonald 1995) and A. russatus (Gutman & Dayan 2005) from nocturnal to diurnal activity, according to the demands of their environment.

Studies that have examined the effects of having a circadian system that does not match the environment have also been useful. Experimental studies, like the ones carried out in S. elongates (Ouyang et al. 1998), W. smithii (Emerson et al. 2008) and A. thaliana (Dodd et al. 2005), have shown that having a circadian system that does not resonate with the organism’s environment is detrimental. However, having a functional circadian system in constant environmental conditions may not be a problem and may be important for synchronization of metabolic processes. Thus, A. mexicanus are still rhythmic despite living exclusively in caves under constant conditions for thousands of generations and loosing other traits associated with life outside caves, such as pigmentation (Luis Espinasa 2006). Taken all together, these studies provide mounting, but not yet overwhelming, evidence that having a circadian system that matches with the oscillating environment is adaptive.

What of the future? Certainly approaches like those that we have described in our review will continue to be used to ascertain the adaptive significance of circadian rhythms. However, new insights and technological advances should increasingly facilitate the design of novel experiments to test adaptiveness. For example, until now only a very few experiments have been able to directly demonstrate that having a functional, environment-matching circadian system can confer fitness, i.e. results in greater reproductive success. Using growth under different T-cycles, it has been demonstrated that having a circadian system that resonates with the environment improves fitness in both W. smithii (Emerson et al. 2008) and S. elongates (Ouyang et al. 1998; Woelfle et al. 2004). It will be interesting use variations of these experiments to examine fitness in other organisms. It may, for example, be feasible to extend the type of mini-evolution experiments carried out by Paranjpe et al. (2003) by applying a selection force to populations with genetic variability in circadian rhythm traits and examining distribution in subsequent generations to determine which are the predominant, and therefore adaptive, phenotypes.

Although many of the experiments that we have described have been conducted in the laboratory, it is becoming clear that laboratory and field experiments can give contrasting results (Calisi & Bentley 2009). Considering that in an organism’s natural environment there may be numerous physical and social factors that are absent under laboratory conditions, it is not surprising that laboratory and field results may differ. In the future more experiments are needed that will use combinations of molecular, physiological and behavioural studies to determine the adaptive significance of circadian rhythms in natural habitats.

Increasingly, understanding the basis for circadian adaptation on a molecular level will also become feasible. As we have described above, there are latitudinal clines in polymorphisms in Thr-Gly encoding repeats in the per gene of Drosophila spp. (Kyriacou et al. 2008) that may be adaptations for temperature compensation in the circadian system. By comparing correlations between two or more polymorphisms, within and between species, it can be possible to define those that have been selected and the form of the selection (Tauber & Kyriacou 2005). Using quantitative trait locus (QTL) analysis researchers are now also able to identify chromosomal regions that interact to cause phenotypic differences between populations or species. QTL analysis has already been successfully used to indicate genes that regulate temperature compensation in the circadian system of different accessions of A. thaliana (Edwards et al. 2005) and to identify genetic regions that may be important for circadian adaptation in N. crassa (Kim et al. 2007). Eventually, such experiments may allow us to not only to examine whether circadian rhythms have an adaptive significance but also uncover the genetics that control the adaptation.


The authors would like to thank Simon Barak, Ami Citri, Sebastian Kadener, Moshe Kiflawi, Ofer Steinitz, Esther Yakir and David Greenberg for their critical reading of the manuscript. Our apologies to the many researchers whose work was not cited due to the space limitation. This work was supported by BSF grant 0378415 and DIP grant 0307712.