Early birds, late owls, and the ecological role of intra-population chronotype variation

Abstract

While the molecular mechanisms underlying variation in chronotypes within populations have been studied extensively, the ultimate selective forces governing it are poorly understood. The proximate cause is variation in clock genes and protein expression, which produces variation in tau (period length of the circadian clock), with early individuals having shorter tau. We studied within-population variation in foraging activity times of two Acomys species in the field. This variation manifested in a regular and consistent sequence of individual foraging activity that is positively and strongly correlated with variation in tau. Thus, variation in circadian clock period length (tau) appears to be the mechanism underlying the regular pattern of intraspecific temporal partitioning. Late chronotypes also spent more time torpid than earlier ones, suggesting an energetic cost to this strategy and possible tradeoffs. We suggest that variation in tau is an adaptive mechanism to reduce competition between individuals within a population.

Full text

1
1
Early birds, late owls, and the ecological role of intra-population chronotype variation 2
Ella Royzrakh-Pasternak, Tamar Dayan, Ofir Levy, Noga Kronfeld-Schor* 3
* Corresponding author 4
5
Ella Royzrakh-Pasternak: pasternak@mail.tau.ac.il, School of Zoology, Tel Aviv University, Tel 6
Aviv, Israel. 7
Ofir Levy, levyofir@tauex.tau.ac.il, School of Zoology, Tel Aviv University, Tel Aviv, Israel. 8
Tamar Dayan, dayant@tauex.tau.ac.il, School of Zoology, Tel Aviv University, Tel Aviv, Israel, 9
and The Steinhardt Museum of Natural History, Tel Aviv University, Tel Aviv, Israel. 10
Noga Kronfeld-Schor, nogaks@tauex.tau.ac.il, School of Zoology, Tel Aviv University, Tel Aviv, 11
Israel. E-mail: Phone: 972-528325314 12
13
Abstract 14
While the molecular mechanisms underlying variation in chronotypes within populations have been 15
studied extensively, the ultimate selective forces governing it are poorly understood. The proximate 16
cause is variation in clock genes and protein expression, which produces variation in tau (period 17
length of the circadian clock), with early individuals having shorter tau. We studied within-18
population variation in foraging activity times of two Acomys species in the field. This variation 19
manifested in a regular and consistent sequence of individual foraging activity that is positively and 20
strongly correlated with variation in tau. Thus, variation in circadian clock period length (tau) 21
appears to be the mechanism underlying the regular pattern of intraspecific temporal partitioning. 22
Late chronotypes also spent more time torpid than earlier ones, suggesting an energetic cost to this 23
strategy and possible tradeoffs. We suggest that variation in tau is an adaptive mechanism to reduce 24
competition between individuals within a population. 25
26
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

2
Introduction 1
The diel activity patterns of animals have attracted significant scientific interest focusing on their 2
evolution, ecological significance, regulating mechanisms, and impacts of anthropogenic activities 3
(1-5). The evolution of diel rhythms is driven by the 24 h light-dark cycle and its derivatives, 4
including both environmental conditions and ecological interactions. In the past two decades a 5
growing volume of research has addressed interspecific niche partitioning along the diel axis, with a 6
multidisciplinary approach that involves community ecology, evolutionary biology, and biological 7
rhythms research (2, 4). This research ranged from macroevolutionary patterns to studies of 8
behavioral plasticity in response to human activity and artificial illumination and their bearing on 9
species coexistence (e.g., (5, 6)). 10
Concurrently biological rhythms research has developed dramatically and while a significant part of 11
it focuses on humans and model organisms, the past decade has seen a growing literature on animal 12
biological rhythms. The 24 h light and dark cycle is a strong and predictable environmental pattern; 13
daily rhythms in behavior and physiology are governed by a biological clock that is synchronized to 14
this universal cue. When biological clocks synchronize or entrain to the environment (usually to the 15
light/dark cycles), they not only show the same period as the zeitgeber (cue) cycle (24 h), but also 16
establish a stable phase relationship with the zeitgeber, called the phase of entrainment (7, 8). 17
In spite of millions of years of natural selection to a strong universal cue, a significant variation in 18
the period length of the internal biological clock occurs, resulting from variation in clock genes and 19
protein expression (9, 10); consequently different individuals synchronize differently to the 20
light/dark cycle, resulting in variability in their phase of entrainment relative to the environmental 21
cue or zeitgeber (e.g., sunrise), called chronotypes . Chronotype describes the manifestation of 22
daily rhythms, which are the outcome of an interaction between the endogenous circadian clock and 23
environmental conditions (7, 11). They correlate with the free-running period of the circadian clock 24
(the endogenous circadian clock period under constant environmental conditions, called tau), which 25
has a significant genetic component, with early chronotypes displaying a shorter tau than late 26
chronotypes (9, 12-14). Here we investigate the role that this genetic and phenotypic variation may 27
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

3
have in the coexistence of individuals within a population and hence the selective forces that may 1
be at play maintaining this variation. 2
Individual variation in daily rhythms under free-living conditions has been the focus of intensive 3
research in humans during the last decades, with earlier individuals described as larks and late 4
individuals as owls (7, 9, 13). Variability in chronotypes was also described in different animal 5
species, including rodents, birds, fishes, insects and others, but these experiments were conducted 6
almost exclusively under controlled laboratory conditions; very few studies attempted to address 7
variation of chronotypes in natural populations of species other than humans. Even fewer studies 8
(14-16) focused on the selective forces driving this variation which are yet to be understood (8). 9
Several studies in recent years have demonstrated intraspecific patterns of temporal partitioning that 10
implicated intraspecific competition (17, 18). We studied intraspecific variation in chronotypes in 11
two coexisting wild rodent species, its adaptive significance and proximate underlying mechanisms, 12
and the selective forces driving it. 13
In an earlier study we found regular intraspecific patterns of foraging activity among golden spiny 14
mice (Acomys russatus), with specific individuals tending to arrive and forage early (early 15
chronotype) and others to arrive and forage later (late chronotype) during the day (19). Persistence 16
of intraspecific variation in chronotypes in natural populations suggests that a specific chronotype 17
entails benefits or costs which depend on environmental conditions, and have fitness consequences. 18
Under our field experimental setup, where food was supplemented and replenished every morning, 19
we found that early individuals spent less time torpid (controlled reduction of body temperature (20, 20
21)) than late individuals (19). As use of torpor was related to a decrease in reproductive success in 21
this species (22), it appears that early chronotypes had a selective advantage in this study system. 22
Intra-population morphological variation has been studied empirically with a theoretical basis that 23
related it to intra-specific competition, with variation related to resource breadth and availability 24
(23). In the current study we hypothesized that in a similar manner, foraging order is determined by 25
chronotype, which in turn is determined by tau, and that intra-population variation in tau, as a 26
proximate mechanism driving variation in chronotype (and hence foraging sequence), will be 27
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

4
likewise selected for to reduce resource use overlap and thus reduce intraspecific competition at the 1
diel niche axis. To test this hypothesis, we measured the order of arrival to a foraging patch 2
(foraging sequence) and use of torpor in golden and common spiny mice (A. cahirinus) under semi-3
natural conditions, and the free-running period lengths (tau) of the foraging individuals under 4
controlled laboratory conditions. 5
Results 6
The experiment yielded a total of 550,000 foraging logs and 300,000 body temperature records. For 7
one common spiny mouse (from enclosure 3) there were no body temperature data, and three 8
golden spiny mice (two from enclosure 3 and one from enclosure 2) were not recaptured and 9
consequently we could not measure their torpor use or period length; these individuals were 10
excluded from the data analysis. 11
Common spiny mice 12
Individuals from the same enclosure entered the foraging patches in a constant order during the 13
experimental period (Fig 1A). The free running period length (tau) of common spiny mice ranged 14
between 23.7 h to 24.3 h (Fig. 2A). 15
Our statistical models suggested that individuals that had shorter period length  arrived earlier to 16
the foraging patches (Estimate= 0.6741, SD= 0.2353; t-value= 2.865; DF=16; p-value=0.00418, 17
∆AIC=4.2451, N=18 ; Fig.3A). 18
Most of the common spiny mouse individuals (82.35%) used torpor for fewer than 60 minutes 19
during the day on average and (71.42% out of them used torpor for fewer than 20 minutes). The rest 20
of the individuals (17.64%) used torpor for fewer than 80 minutes a day on average. 21
The period length did not affect the daily torpor usage of the common spiny mice (Estimate= 22
-0.3880, SD= 0.2829; t-value= -1.372; p-value= 0.17; DF=15, ∆AIC=-0.16, N=18; Fig.4A), 23
probably due the overall low torpor usage of the common spiny mice. In both tests, we used 24
Generalized Linear Mixed Effect Models (GLMM), as indicated in the Methods section, page 12. 25
26
27
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

5
Golden spiny mice
1
Individuals from the same enclosure entered the foraging patches in a constant order during the
2
experimental period (Fig 1B). Furthermore, the frequency distribution of the period length of the
3
golden spiny mice individuals is dispersed between 23.2 h to 25.4 h (Fig. 2B). In enclosure # 2
4
where four golden spiny mice were placed, two pairs of individuals overlapped almost entirely in
5
tau, and the four individuals were fairly close in tau as well (Fig 2B). Hence no correlations could
6
be discovered and this enclosure was excluded from the following analyses.
7
Individuals that had shorter period length (τ) arrived earlier to the foraging patches (Estimate=
8
0.3036, SD= 0.1591; t-value= 1.908; p-value= 0.0564; DF=13; ∆AIC=1.1721, N=15; Fig.3B). The
9
period length had a significant positive effect on the daily torpor use of golden spiny mice
10
(Estimate= 86.55, SD= 19.91; t-value= 4.346; DF=12; p-value<0.001; ∆AIC=5.68, N=15; Fig.4B).
11
The daily time spent torpid increased by 2.4 minutes with every change of 1 minute in the period
12
length. In both tests, we used Generalized Linear Mixed Effect Models (GLMM), as indicated in
13
the Methods section, page 12.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
A
B
Figure 1: The mean order of entry (±SE) of common (A, n=18 individuals, 4 enclosures (replicates)) and
golden (B, n=19 individuals,
4
enclosures (replicates)) spiny mice in each enclosure (1-4). Each point
represents a different individual. One common spiny and three golden spiny mice were excluded from
the data analysis, due to the reasons mentioned in the Results, p. 4). Source files of all the data used for
the analysis are available in the Figures 1a, 2a, 3a first visit and tau A. cahirinus and Figures 1b, 2b, 3b
first visit and tau A. russatus source data files.
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
A
B
Figure 2: Distribution of free running period length (tau
) in common (A, n=18
individuals
,
4
enclosures (replicates)) and golden (B, n=1 individuals, 4 enclosures (replicates)) spiny mice
individuals in each enclosure separately. One common spiny and three golden spiny mice were
excluded from the data analysis, due to the reasons mentioned in the Results, p. 4). Source files of
all the data used for the analysis are available in the Figures 1a, 2a, 3a first visit and tau A. cahirinus
and Figures 1b, 2b, 3b first visit and tau A. russatus source data files.
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Figure 3: The correlation between individual mean daily order of entries (±SD) and period length (h) in
common (A, n=18 individuals,
4
enclosures (replicates), p-value=0.00418,
GLMM) and golden (B,
n=15 individuals, 3 enclosures (replicates), p-value=0.0564,
GLMM) spiny mice. The four golden
spiny mice in enclosure # 2 were excluded from the data analysis due to the reasons mentioned in the
Methods, p.5. Source files of all the data used for the analysis are available in the Figures 1a, 2a, 3a
first visit and tau A. cahirinus and Figures 1b, 2b, 3b first visit and tau A. russatus source data files.
B
A
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

8
1
Figure 4: The correlation between the individual mean daily torpor time (minutes; ±SD) and the period
length (h) in common (A, n=18 individuals, 4 enclosures (replicates), p-value=0.17, GLMM) and golden
(B, n=15 individuals, 3 enclosures (replicates), p-value<0.001, GLMM) spiny mice. The four golden
spiny mice in enclosure # 2 were excluded from the data analysis due to the reasons mentioned in the
Methods, p.5. Source files of all the data used for the analysis are available in the Figures 4a first visit
and torpor A. cahirinus and Figure 4b first visit and torpor A. russatus source data files.
A
B
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

9
Discussion 1
We found a constant order of arrival to the food patch in both A. russatus and A. cahirinus: 2
individuals of both species tended to arrive to the food patch in the same order along the 3
experimental period. In seven of the eight population in the enclosures, we also found that the order 4
of arrival is positively correlated with the period of the internal biological clock (tau), measured 5
under controlled conditions in the laboratory. In the A. russatus population in enclosure #2 we 6
found no significant positive correlation, and hence removed it from further analysis. Finally, in A. 7
russatus foraging sequence correlated with torpor use, an energy conserving mechanism. Research 8
suggests that tau has a significant genetic component (9, 12-14), so it appears that the order of 9
arrival to forage in a food patch, returns in energy acquisition, and the consequent energy 10
conserving strategy required, are strongly influenced by the genetic makeup of the individual. 11
We note that the entire pool of mice studied constitutes random sampling by trapping the local free-12
living spiny mouse populations, while the individuals populating specific enclosures are merely a 13
randomly assigned sub-sample of this pool. The analysis of all studied individuals within species 14
yielded clear correlations. However, as the numbers of individuals within species populating the 15
enclosures are limited (4-6), they cannot be taken to fully represent tau in the native population. 16
Even with this limitation, a clear pattern emerges in seven of eight populations. We thus suggest 17
that in the field, intra-population variation in tau, as a proximate mechanism for driving variation in 18
chronotype, is selected for to reduce resource use overlap or intraspecific aggression (see below) 19
and thus to reduce intraspecific competition at the diel niche axis. 20
While use of time as a niche axis for reducing interspecific competition has received considerable 21
attention during the last decade (reviewed elsewhere (3, 4)), it was seldom addressed in relation to 22
intraspecific interactions. We have previously shown that activity during different parts of the diel 23
cycle facilitates coexistence between the two spiny mouse species in this system (2, 24). Being 24
active at different times exposes the two omnivorous species to different prey worlds and thus 25
reduces resource use overlap (24, 25). It is possible that in the same manner, individuals within a 26
population of each species separate foraging activity time along the day or nighttime axis. 27
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

10
However, it is also possible that being active at different times reduces direct, aggressive 1
interactions between individuals of the same species: in a previous study, we found that only rarely 2
did two individuals forage concurrently in a foraging patch and in the infrequent cases when this 3
occurred, 90% of the individuals left within 2 seconds (19). 4
The question whether the deviation of the internal clock period from 24 hours has any specific 5
function or causal meaning has attracted attention of scientists for years. Traditionally, scientists 6
considered the deviation of tau from 24 hours a simple biological imperfection which is corrected 7
daily by a Zeitgeber (light exposure), and therefore does not lead to serious functional deficits. 8
Accordingly, most studies of this variation focused on its proximate casual mechanism, trying to 9
identify the molecular mechanisms generating the circadian phenotypic diversity, and not on its 10
functional, adaptive significance. However, since the deviation of tau from 24 hours defines the 11
chronotype, one could expect it to be under selection pressure (11). In 1981, Pittendrigh wrote that 12
the deviation is "a strategy, and not a tolerated approximation", arguing that it contributes to the 13
stabilization of the 24 h rhythms (26). Daan and Beersma (2002), modeled the contribution of the 14
deviation to the stability of the rhythm and concluded that there is no strong theoretical basis for 15
this argument. However, the occurrence of clines in tau along latitudinal gradients suggest the role 1 6
of natural selection (27, 28). We suggest that this variation reduces intraspecific competition, 17
whether consumptive or aggressive and hence may well be selected for. In fact, variation in tau 18
may reflect a tradeoff between individual dominance in aggressive encounters and foraging 19
efficiency. 20
Intraspecific morphological variation in the trophic apparatus of different animals has been 21
interpreted as a means for reducing intraspecific resource competition (23, 29). The niche variation 22
hypothesis, first coined by Van Valen (30), suggests that when interspecific selection is relaxed, 23
intraspecific variation increases to reduce intraspecific competition (23). We suggest that the 24
variation in tau that appears in these two species likewise reflects the selective role of intraspecific 25
competition. 26
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

11
In the current study A. cahirinus very seldom used torpor, while A. russatus entered torpor 1
frequently, and the time spent torpid was correlated not only with individual order of arrival (19), 2
but also with tau. This means that there was a genetic component in the proximate factors that 3
influenced individual chronotype, and that, in turn, determined the ability of the mice to acquire 4
energy and affected their use of torpor as means of reducing energy expenditure. Since previous 5
research suggests that use of torpor in these species results in reduced reproductive activity (22), 6
under the current experimental conditions, where food is supplemented at the beginning of the day 7
and the beginning of the night, earlier chronotypes are expected to have a clear selective advantage. 8
Indeed, previous researchers have suggested that intraspecific temporal partitioning can allow 9
dominant individuals to forage at times of higher rewards, as documented in bellbirds (31) and 10
brown trout (18). 11
However, under natural conditions with no food supplemented, if the selective advantage depends 12
on food availability, optimal foraging times may vary during the day. Since both species consume 13
mainly arthropods (especially during summer) whose availability changes along the day (25), and 14
since environmental conditions influencing energy expenditure vary along the day and between 15
days and seasons, it is possible that tradeoffs occur between the different selective pressures and 16
that different chronotypes will have a selective advantage during different times. 17
The source of variation in chronotypes of spiny mice is currently unknown. In humans, genetic 18
influences (heritability) appear to account for a significant proportion. For example, 56% of the 19
variability in food timing, particularly breakfast (32), and 68% of sleep timing on free days 20
(33). An even higher heritability was found in great tits (Parus major), with h2 of 0.86 ± 0.24 21
(14). Additionally, chronotypes vary with age and sex: in humans chronotype delays from the age 22
of 10 to 20 years (adolescence), and then advances with age, and on average, females have later 23
chronotypes than males until the average age of menopause (34). A similar change in chronotype 24
with age was also documented in Rhesus Monkeys (Macaca mulatta) (35), and an opposite trend, 25
with older individuals having earlier chronotypes was found in in rodents (e.g.,(36)). Spiny mice 26
may live for 2-3 years in natural conditions, so it is entirely possible that their tau changes with age. 27
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

12
We know not the ages of the individual spiny mice studied, nor the age structure of the population. 1
While it would be interesting to speculate that the significant intrapopulation variation in tau 2
reflects both selective pressures at the individual level and age-driven individual variation, further 3
research is required to address this question. 4
In humans, extreme chronotypes are usually interpreted as syndromes (e.g., delayed and advanced 5
sleep syndrome), although some studies suggest that variability in chronotypes may have adaptive 6
significance under natural conditions, which became maladaptive in the modern world. For 7
example, in the Hadza tribe of hunter-gatherers of Tanzania, a variability in chronotype resulted in 8
one or more individuals awake during 99.8% of the time. Having at least one person awake, will 9
reduce the hazards of sleep to the whole group. When group size drops, active sentinel behavior is 10
adopted (37). Thus intraspecific variation in tau may have evolutionary significance for anti-11
predator vigilance and for coexistence between competing individuals. Possibly, it has other roles 12
that should be further explored. 13
In sum, our study suggests that the significant variation in tau that is frequently found among 14
individuals is not a mere ‘biological imperfection’ but has evolutionary significance and is selected 15
for; our study system demonstrates its role in reducing intraspecific competition between 16
individuals within a population. 17
Methods 18
Study design 19
The research took place on July- October 2016, under semi natural conditions at four 1000m2 field 20
enclosures located in the Dead Sea (for details see (19)). We captured 19 Common spiny mice and 21
22 Golden spiny mice in the area around the enclosures, using Sherman live traps. Based on 22
previous studies in from our laboratory in these enclosures, and according to the spiny mice 23
population density in nature, and specifically in the Ein Gedi area (19, 22, 38, 39), each enclosure 24
was populated randomly with 8-11 individuals, 4-6 individuals of each species, with a sex ratio that 25
is close to 1. The four enclosures, and the two species served as biological replicates (2 species, 4 26
enclosures, 4-6 individuals of each species in each enclosure: total of 8 populations). We did not 27
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

13
have technical replications since there was no experimental manipulation on the spiny mice 1
populations in the enclosures. 2
When populating the enclosures, each spiny mouse captured was populated in the enclosure that 3
had the least individuals from the captured individual species and sex, until all enclosures were 4
populated. We had no information regarding period length (which was measured only at the end of 5
the experiment) when the enclosures were populated, thus in this regard the individuals were 6
populated completely randomly in the four enclosures. 7
Mice were implanted with PIT tags (Destron- Fearing, South St. Paul, MN) subcutaneously and 8
iButtons (DS1922L-F5 thermochron; San Jose, CA, USA) in the abdominal cavity. Body and 9
ambient temperatures were measured continuously (every 20 minutes) over four months, using 10
iButtons. The ambient temperature was measured by placing two iButtons under the boulders of 11
each terrace in each enclosure. Foraging activity and order of arrival to auto-monitored food 12
patches (see below) were measured for 14 days in each month under half-moon conditions (7 13
consecutive days twice a month). The order of entry of each individual was measured for 14 days a 14
month, for four months (43 days in total), and torpor was measured each day, for 100 days in total 15
(considered as repeated measures for each individual, as explained in the Statistical analysis, p.13). 16
Monitoring foraging time 17
We used auto-monitored foraging patches, comprising a plastic tray (25 cm diameter), in which 2L 18
of local soil were mixed with 3 g of broken sunflower seeds (3g seeds per individual). In each 19
enclosure, we placed two round antennas (20 cm diameter) beneath two foraging trays and each 20
antenna was connected to a transceiver (2001F-ISO; Biomark Ltd, Boise, ID), which logged the tag 21
ID and time (in seconds) of each mouse that entered the patch. Solar panels and a generator 22
provided power supply to the transceivers and antennas. The foraging trays were replenished at 23
dawn and at dusk, the soil was sieved and all the remaining seeds were removed. 24
Monitoring the Period length (of Spiny mice 25
Mice were trapped on October 2016 and transferred to the laboratory at Tel Aviv University, where 26
they were individually housed in 38X24X13 cm plastic cages in isolated sound proof rooms, with 27
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

14
an ambient temperature of 29oC, under constant dark conditions (DD) with food (standard rodent 1
chew) and water ad libitum for 10-14 days. Activity was continuously recorded using infrared 2
detectors (Intrusion detector model MH10; Crow group, Kiriyat- Teufa, Israel) connected to a 3
personal computer. Data were collected by a PC at 6-min intervals using software designed for this 4
purpose (ICPC, Netanya, Israel). The period length of each individual was measured once. 5
Data and statistical analysis 6
Foraging sequence: Forging sequence in each foraging patch in each enclosure was determined 7
daily according to the first time an individual arrived to the patch. We calculated the mean order of 8
arrival to the food patches for each individual across days. 9
Use of torpor: torpor temperature threshold was defined according to Willis (40). For each day, we 10
calculated individual total time torpid. Torpor bout duration was determined as time between the 11
beginning of the torpor bout until body temperature returned above the torpor threshold. For each 12
individual, we calculated the mean time spent torpid across days. 13
Determining Period length : Actograms were generated using ClockLab (Actimetrics, 14
Wilmette, IL, USA) and chi-square periodograms were used to calculate the periods of free running 15
period length of each individual. 16
Statistical analysis: For each species, we fitted a statistical model to explore how period length  17
of each individual as a continuous covariate has affected its (1) daily order of entry to the food 18
patches and (2) the amount of time it spent torpid. 19
All statistical models were fitted using Generalized Linear Mixed Effect Models (GLMM) with a 20
Gamma distribution and the identity of each individual and enclosure as nested random factors. 21
Since our data did not fit the test assumptions of regular linear models (such as normality and 22
homogeneity of variances), we fitted our statistical models using Generalized Linear Mixed Effect 23
Models (GLMM) with a Gamma distribution, which is flexible and does not assume 24
homogeneity. To account for the repeated measures of individuals within different enclosures, we 25
included their identity as nested random factors. 26
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

15
Based on AIC (Akaike Information Criteria) criteria, we used the identity link function for all 1
models, except model 1 for golden spiny mice, where we used the log link function. We also used 2
AIC (Akaike Information Criteria) (41) criteria, to select the important fixed and random effects in 3
each model. For all tests, we used the glmer function (lme4 package (42)) in the R software (R Core 4
Team, 2018).). To improve the stability of the statistical models, we run the models with 5
standardized values of the period length (minutes), and de-standardized the effect sized back to the 6
units of minutes. 7
For each statistical model, we report the slope of the period length, the significance of the effect of 8
the period length as the change in AIC compared to the model without period length as a dependent 9
covariate, and the p-value of the effect sizes as given by the model. Since we used the log link 10
function in model 1 for the golden spiny mice, we report the estimated change in the order of entry 11
for every unit increase in the period length as , where k is the estimated slope63. 12
Acknowledgements 13
Funding: This research was supported by THE ISRAEL SCIENCE FOUNDATION (grant No. 14
934/12 and 2129/20) to N.K-S. 15
Competing interests: 16
The authors declare no competing interests. 17
References 18
1. Daily GC, Ehrlich PR. Nocturnality and species survival. Proceedings of the National 19
Academy of Sciences. 1996;93(21):11709-12. 20
2. Kronfeld-Schor N, Dayan T. Partitioning of time as an ecological resource. Annual Review 21
of Ecology Evolution and Systematics. 2003;34:153-81. 22
3. Hut R, Kronfeld-Schor, N, van der Vinne, V, De la Iglesia, H. In search of a temporal niche: 23
environmental factors. Progress in Brain Research. 2012;199:281-304. 24
4. Kronfeld-Schor N, Visser ME, Salis L, van Gils JA. Chronobiology of interspecific 25
interactions in a changing world. Philosophical transactions of the Royal Society of London Series 26
B, Biological sciences. 2017;372(1734). 27
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

16
5. Gaynor KM, Hojnowski CE, Carter NH, Brashares JS. The influence of human disturbance 1
on wildlife nocturnality. Science. 2018;360(6394):1232-5. 2
6. Maor R, Dayan T, Ferguson-Gow H, Jones KE. Temporal niche expansion in mammals 3
from a nocturnal ancestor after dinosaur extinction. Nature ecology & evolution. 2017;1(12):1889-4
95. 5
7. Roenneberg T, Hut R, Daan S, Merrow M. Entrainment Concepts Revisited. Journal of 6
Biological Rhythms. 2011;25(5):329-39. 7
8. Helm B, Visser ME, Schwartz W, Kronfeld-Schor N, Gerkema M, Piersma T, et al. Two 8
sides of a coin: ecological and chronobiological perspectives of timing in the wild. Philosophical 9
Transactions of the Royal Society B: Biological Sciences. 2017;372(1734):20160246. 10
9. Brown SA, Kunz D, Dumas A, Westermark PO, Vanselow K, Tilmann-Wahnschaffe A, et 11
al. Molecular insights into human daily behavior. Proceedings of the National Academy of 12
Sciences. 2008;105(5):1602-7. 13
10. Lane JM, Vlasac I, Anderson SG, Kyle SD, Dixon WG, Bechtold DA, et al. Genome-wide 14
association analysis identifies novel loci for chronotype in 100,420 individuals from the UK 15
Biobank. Nature communications. 2016;7:10889. 16
11. Daan S, Beersma DGM. Circadian Frequency and Its Variability. In: Kumar V, editor. 17
Biological Rhythms. Berlin, Heidelberg: Springer Berlin Heidelberg; 2002. p. 24-37. 18
12. Aschoff J, Wever R. Circadian period and phase-angle difference in chaffinches (Fringilla 19
coelebs L.). Comparative biochemistry and physiology. 1966;18(2):397-404. 20
13. Duffy JF, Rimmer DW, Czeisler CA. Association of intrinsic circadian period with 21
morningness–eveningness, usual wake time, and circadian phase. Behavioral neuroscience. 22
2001;115(4):895. 23
14. Helm B, Visser ME. Heritable circadian period length in a wild bird population. Proceedings 24
of the Royal Society of London B: Biological Sciences. 2010;277(1698):3335-42. 25
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

17
15. Dominoni DM, Helm B, Lehmann M, Dowse HB, Partecke J. Clocks for the city: circadian 1
differences between forest and city songbirds. Proceedings of the Royal Society B: Biological 2
Sciences. 2013;280(1763):20130593. 3
16. Graham JL, Cook NJ, Needham KB, Hau M, Greives TJ. Early to rise, early to breed: a role 4
for daily rhythms in seasonal reproduction. Behav Ecol. 2017;28(5):1266-71. 5
17. Price CJ, Tonn WM, Paszkowski CA. Intraspecific patterns of resource use by fathead 6
minnows in a small boreal lake. Canadian Journal of Zoology. 1991;69(8):2109-15. 7
18. Alanärä A, Burns MD, Metcalfe NB. Intraspecific resource partitioning in brown trout: the 8
temporal distribution of foraging is determined by social rank. Journal of Animal Ecology. 9
2001;70(6):980-6. 10
19. Levy O, Dayan T, Rotics S, Kronfeld-Schor N. Foraging sequence, energy intake and 11
torpor: an individual-based field study of energy balancing in desert golden spiny mice. Ecology 12
Letters. 2012;15(11):1240-8. 13
20. Geiser F, Ruf T. Hibernation versus daily torpor in mammals and birds - physiological 14
variables and clasification of torpor patterns. Physiological Zoology. 1995;68(6):935-66. 15
21. Kronfeld-Schor N, Dayan T. Thermal Ecology, Environments, Communities, and Global 16
Change: Energy Intake and Expenditure in Endotherms. Annual Review of Ecology, Evolution, and 17
Systematics, Vol 44. 2013;44:461-80. 18
22. Katz N, Dayan T, Kronfeld-Schor N. Fitness effects of interspecific competition between 19
two species of desert rodents. Zoology. 2018;128:62-8. 20
23. Dayan T, Simberloff D. Ecological and community-wide character displacement: the next 21
generation. Ecology Letters. 2005;8(8):875-94. 22
24. Kronfeld-Schor N, Dayan T. The dietary basis for temporal partitioning: food habits of 23
coexisting Acomys species. Oecologia. 1999;121(1):123-8. 24
25. Vonshak M, Dayan T, Kronfeld-Schor N. Arthropods as a prey resource: Patterns of diel, 25
seasonal, and spatial availability. Journal of Arid Environments. 2009;73(4-5):458-62. 26
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

18
26. Pittendrigh CS. Circadian systems: entrainment. Biological rhythms: Springer; 1981. p. 95-1
124. 2
27. Randler C, Rahafar A. Latitude affects Morningness-Eveningness: evidence for the 3
environment hypothesis based on a systematic review. Scientific Reports. 2017;7(1):39976. 4
28. Kyriacou CP, Peixoto AA, Sandrelli F, Costa R, Tauber E. Clines in clock genes: fine-5
tuning circadian rhythms to the environment. Trends in Genetics. 2008;24(3):124-32. 6
29. Hutchinson G, MacArthur RH. A theoretical ecological model of size distributions among 7
species of animals. The American Naturalist. 1959;93(869):117-25. 8
30. Van Valen L. The study of morphological integration. Evolution. 1965;19(3):347-9. 9
31. Craig JL, Douglas ME. Temporal partitioning of a nectar resource in relation to competitive 10
asymetries. Animal Behaviour. 1984;32(MAY):624-5. 11
32. Lopez-Minguez J, Dashti HS, Madrid-Valero JJ, Madrid JA, Saxena R, Scheer F, et al. 12
Heritability of the timing of food intake. Clinical Nutrition. 2019;38(2):767-73. 13
33. Inderkum AP, Tarokh L. High heritability of adolescent sleep-wake behavior on free, but 14
not school days: a long-term twin study. Sleep. 2018;41(3). 15
34. Roenneberg T, Kuehnle T, Juda M, Kantermann T, Allebrandt K, Gordijn M, et al. 16
Epidemiology of the human circadian clock. Sleep medicine reviews. 2007;11(6):429-38. 17
35. Golub MS, Takeuchi PT, Hoban-Higgins TM. Nutrition and circadian activity offset in 18
adolescent rhesus monkeys. Adolescent sleep patterns. 2002:50-68. 19
36. Hagenauer MH, King AF, Possidente B, McGinnis MY, Lumia AR, Peckham EM, et al. 20
Changes in circadian rhythms during puberty in Rattus norvegicus: developmental time course and 21
gonadal dependency. Hormones and behavior. 2011;60(1):46-57. 22
37. Samson DR, Crittenden AN, Mabulla IA, Mabulla AZ, Nunn CL. Chronotype variation 23
drives night-time sentinel-like behaviour in hunter–gatherers. Proc R Soc B. 24
2017;284(1858):20170967. 25
38. Rotics S, Dayan T, Kronfeld-Schor N. The effect of elevated predation risk on intraspecific 26
interactions of common spiny mice. Israel Journal of Ecology & Evolution. 2011;56(1):100-. 27
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint

19
39. Rotics S, Dayan T, Levy O, Kronfeld-Schor N. Light Masking in the Field: An Experiment 1
with Nocturnal and Diurnal Spiny Mice Under Semi-natural Field Conditions. Chronobiology 2
International. 2011;28(1):70-5. 3
40. Willis CKR. An energy-based body temperature threshold between torpor and 4
normothermia for small mammals. Physiological and Biochemical Zoology. 2007;80(6):643-51. 5
41. Burnham KP, Anderson DR. Practical use of the information-theoretic approach. Model 6
Selection and Inference: Springer; 1998. p. 75-117. 7
42. Bates D. LME4: linear mixed-effects models using S4 classes. Version 0.99875-9. 2008. 8
9
.CC-BY 4.0 International license(which was not certified by peer review) is the author/funder. It is made available under a The copyright holder for this preprintthis version posted August 10, 2020. . https://doi.org/10.1101/2020.08.09.243261doi: bioRxiv preprint