Time Perception in Dogs: Rhythm, Duration and Absence

1. What the Question Actually Is

1.1 Several Different Questions Hide Inside "Sense of Time"

At least four separable capacities are routinely compressed into the phrase "sense of time": whether a dog's physiology follows a daily rhythm; whether duration itself can control behavior; whether the dog retains information about when an event occurred; and whether an hours-long absence is registered in a way that changes behavior or welfare.

These questions require different methods and they do not have the same evidential status. Daily rhythmicity is well established. Seconds-range interval timing has been demonstrated experimentally in several canine studies. Dogs can also retain temporal information about past events. Direct experimental work on hours-long owner absence, however, remains exceptionally limited.

1.2 What We Cannot Ask Directly

We cannot ask a dog whether five minutes felt short or long. The subjective experience of duration is therefore not directly reportable in the way it is in a verbal human participant. That does not mean subjective timing cannot be studied at all: psychophysical tasks can estimate points of subjective equality and discrimination functions from behavior.

The distinction matters. A bisection task can show that a dog treats one interval as more like a learned "short" duration and another as more like a learned "long" duration. It does not tell us what waiting feels like to the dog (why measured behavior is not an inner state).

1.3 Four Research Traditions, Not One

The relevant literature comes from chronobiology, operant psychophysics, memory research and welfare/separation research. The methods range from collar actigraphy and lever or feeder choices to odor-memory tests, home video and salivary cortisol.

Results from one tradition should not be used as proof for another. Showing that a dog discriminates a two-second from an eight-second signal does not establish that it measures a six-hour absence by the same mechanism. Showing a daily activity rhythm does not establish a conscious representation of clock time.

1.4 Why the Question Matters in Practice

Households rarely ask about time perception for theoretical reasons. The practical questions are usually whether a dog "knows" someone is late, whether a longer working day is harder on the dog, or whether a dog can be left for a particular number of hours.

Those are partly welfare questions, not simply timing questions. The useful scientific task is therefore to separate what has actually been measured from what remains an inference, and to identify which observations can be made directly in the individual dog.

2. The Biological Clock

2.1 Dogs Show Daily Rest-Activity Rhythms

Dogs show strong day-night patterns in locomotor activity. Zanghi and colleagues monitored 48 beagles in three age groups under a 12-hour light/dark schedule and found clear diurnal activity rhythms as well as age-related differences (Zanghi et al., 2012).

This establishes rhythmic organization of activity. It does not require the dog to estimate elapsed time from a starting event or to represent a clock time in the human sense.

2.2 What the Feeding Study Actually Manipulated

Zanghi et al. (2012) is often paraphrased as showing that dogs become active at the exact time at which they are usually fed. That is not what the experiment tested. The study compared feeding frequency: once-daily versus twice-daily feeding.

Twice-daily feeding changed several activity measures, including nighttime activity and activity shortly before the light phase. The result supports the broader point that feeding regimen can influence rest-activity organization. It does not show that moving a meal from one clock time to another will necessarily move a dog's activity peak by the same amount.

2.3 Entrainment Is Not Clock Reading

Daily rhythms can be synchronized by recurring environmental and physiological cues. Light-dark cycles, feeding schedules, household activity and other regular events can all covary with time of day.

A dog that becomes active around a recurring event may therefore look impressively punctual without representing "07:00" as a clock time. An entrained biological system can produce accurate-looking anticipation.

2.4 What Actigraphy Measures

Collar-worn accelerometers record movement continuously and are useful for studying activity and rest across days. Zanghi et al. (2013) also used actigraphy to characterize behavioral sleep in dogs of different ages and feeding regimens.

Actigraphy is useful precisely because it is unobtrusive, but it is not electroencephalography. Low movement can be classified as behavioral sleep according to validated rules, yet the method does not provide sleep architecture and cannot establish physiological sleep states with the precision of EEG-based recording (how canine sleep is measured).

2.5 Rhythm Research Is Relevant Without Being a Time-Perception Test

Rest-activity organization matters in aging and welfare research because fragmentation, altered nighttime activity and changes in behavioral sleep can accompany clinical change. That makes rhythm useful, but for a different question from duration perception (what changes in canine cognitive aging).

3. Interval Timing: Dogs Can Use Duration as Information

3.1 The First Canine Bisection Evidence

Domeniconi and Machado (2017) extended a temporal bisection task to five dogs. The dogs learned different responses after a 1-second versus a 4-second tone and were then tested with intermediate durations. The proportion of "long" choices increased as duration increased, and the point of subjective equality was close to the geometric mean of the trained intervals.

The sample was tiny, but the experiment provided direct evidence that duration itself could control canine choice behavior.

3.2 Fixed-Interval Timing and Overshadowing

Macpherson and Roberts (2017) used both fixed-interval and bisection procedures. In one experiment dogs learned a response pattern around a 30-second interval. In another they discriminated 2-second and 8-second compound signals and were tested at intermediate durations.

The study also showed an overshadowing effect: when auditory and visual cues were presented together, the dogs' timing was controlled more strongly by the auditory component. This is useful evidence that interval timing exists in dogs, but also that what is timed depends on which cue controls behavior.

3.3 A Broader Range of Durations

Cliff et al. (2019) tested canine timing across four duration pairs: 0.5 versus 2 seconds, 1 versus 4, 2 versus 8 and 4 versus 16 seconds. Dogs again showed orderly temporal bisection, with psychophysical functions consistent with an interval-timing process.

Across these studies the defensible conclusion is no longer merely that interval timing "has been tested once." Dogs have repeatedly shown seconds-range duration discrimination in controlled tasks, although the individual studies remain small and highly trained.

3.4 What These Experiments Establish

They establish a capacity to bring behavior under the control of elapsed duration. They do not establish that dogs count minutes or hours during everyday owner absence, and they do not show that the same mechanism explains anticipation of daily routines.

The jump from seconds in a psychophysical task to hours in a household remains experimentally unbridged.

3.5 Timing and Training

These experiments are conceptually relevant to training because delays are durations the animal can discriminate. They support the general idea that timing differences can become behaviorally meaningful.

They do not directly test marker timing, delayed reinforcement or the size of an effective reinforcement window. Advice about immediate markers and consequences rests primarily on learning theory and reinforcement research rather than on canine temporal-bisection studies themselves (how reinforcement timing functions in training).

3.5a What the Overshadowing Result Adds

The finding that a tone controlled timing while a light presented alongside it did not is more than a technical detail. It shows that when several stimuli mark the start of an interval, they do not contribute equally: one can dominate and the other can end up carrying no control at all.

For applied work that is a caution about compound signals. A handler who marks the start of a duration with a word, a hand movement and a step forward has no guarantee that the dog is timing from the element the handler has in mind (how contingencies and signals are arranged).

3.5b What the Laboratory Setting Selects For

Timing tasks require a dog to learn a discrimination, hold two responses available and attend to a stimulus whose only relevant property is its duration. Dogs that do not acquire the discrimination contribute no data, so every sample is selected for animals that solved the task.

That is standard in psychophysics and worth remembering when the result is quoted as "dogs can tell time". What these experiments establish is a capacity under trained conditions, not a description of what an untrained dog does in a kitchen (why test samples rarely represent the population).

3.6 Circadian Timing and Interval Timing Are Different Systems

A circadian system cycles continuously on an approximately 24-hour scale and is synchronized by recurring cues. An interval-timing task begins with an event and requires behavior to vary with the duration that follows.

Both can make a dog appear to "know the time," but they are not interchangeable mechanisms.

4. Hours Alone: What Direct Studies Actually Show

4.1 Rehn and Keeling: 0.5, 2 and 4 Hours

Rehn and Keeling (2011) studied 12 privately owned dogs with no history of separation-related behavior problems. Each dog was recorded on three occasions when left alone at home for 0.5, 2 and 4 hours. Video began before departure and continued for 10 minutes after the owner returned; heart-rate measures were also collected.

This is a valuable repeated-measures design because each dog experienced all three durations. Its major limitation is equally obvious: twelve dogs are not a population-wide answer to how dogs experience being left alone.

4.2 No Significant Difference Is Not Proof of Equality

At equivalent intervals during the separations, the researchers found no significant treatment differences in the measured behaviors. The dogs were generally inactive for much of the absence.

That does not mean a dog left for four hours "behaved exactly like" one left for half an hour, and it does not demonstrate equivalence. With such a small sample, failure to detect a difference can reflect limited statistical power as well as genuinely small effects.

4.3 Differences Appeared at the Reunion

Several differences emerged when the owner returned. After the longer separations the dogs showed more physical activity, attentiveness, tail wagging and interaction-related behavior. Lip licking and body shaking also increased in the reunion period.

These behaviors show that separation duration affected the transition at reunion. They do not map cleanly onto a single emotional label. Lip licking and body shaking, for example, are not specific measures of "stress," "conflict" or "happiness" and should be interpreted together with the rest of the behavioral sequence.

4.4 Owner Behavior Was Not Standardized Away

The study recorded 10 minutes after the owner returned, but owners were instructed to use their normal routines when leaving and returning. That means reunion behavior includes the dog's response to the duration of separation and the interaction pattern that followed the owner's return.

This does not invalidate the result. It does mean that greeting intensity cannot be treated as a pure readout of what the dog experienced during the preceding hours.

4.5 A Second Direct Duration Study: Approximately 2, 4 and 6 Hours

Silbermann and Gansloßer (2023) added a second, even smaller direct test. Six single-household dogs without known separation-related problems, all already accustomed to at least six hours of separation, were recorded during separations averaging approximately 2:05, 4:08 and 6:10 hours.

The dogs were mostly inactive, especially lying resting or lying alert, and vocalization was almost absent. Across the tested duration classes, behavior did not change significantly. Salivary cortisol was assessed around the 2- and 6-hour conditions and likewise did not show a significant duration effect.

4.6 What the Two Direct Studies Allow Us to Say

Together, Rehn and Keeling (2011) and Silbermann and Gansloßer (2023) show that direct experimental evidence on hours-long separation exists, but is exceptionally sparse. The samples were 12 dogs and 6 dogs, and both studies selected dogs without established separation-related problems.

That is nowhere near enough to derive a universal "safe" number of hours, nor to conclude that longer durations are behaviorally irrelevant. The studies are best read as descriptions of small samples of dogs that were already able to remain alone.

4.6a Inactivity Is an Ambiguous Measure

Both direct studies found dogs largely inactive while alone, which is easy to read as contentment and does not establish it. Lying still is compatible with resting, with sleeping and with a dog that has stopped acting because nothing it does changes the situation (why stillness is hard to interpret).

That is the reason both studies added a physiological measure alongside the ethogram. Heart rate, heart rate variability and salivary cortisol are not perfect either, but they fail in different ways than behavioral scoring does, which is what makes the combination worth having.

4.7 Behavior and Physiology Complement Each Other

Adding heart rate, heart-rate variability or cortisol can strengthen a study because physiology may change when overt behavior does not. But physiological measures are not transparent emotion meters either. Cortisol varies with time of day, activity, anticipation and many other factors.

The strongest interpretation comes from converging behavioral, physiological and contextual information rather than from treating any one variable as a direct measure of welfare.

5. The Smell-as-a-Clock Hypothesis

5.1 The Popular Idea

A widely repeated explanation proposes that the owner's odor gradually decreases in the home and that the dog can use the remaining odor concentration as an index of elapsed time. It is intuitively attractive because dogs have extraordinary olfactory abilities (how canine olfaction works).

5.1a Why the Idea Fits Canine Biology

The hypothesis is attractive because it connects two things that are separately well supported: that olfaction is the dog's dominant perceptual channel, and that volatile compounds dissipate at rates that are in principle informative (how canine olfaction works).

A mechanism that is plausible on both counts still has to be tested, and the gap between plausible and demonstrated is where most of the confident writing about this topic sits.

5.2 Its Evidential Status

The crucial experiment has not been done. There is no controlled canine study in which scent concentration is manipulated independently of elapsed owner absence and the dog's timing-related response is measured.

The scent account is therefore a plausible hypothesis, not an established finding.

5.3 Many Variables Change With Time

As an absence progresses, odor is not the only thing that changes. Light, temperature, traffic, neighboring activity, hunger, bladder state, sleep-wake cycles and internal physiological signals can all covary with elapsed time.

Current canine evidence does not tell us which of these cues, alone or in combination, carries the most information about an owner's return. There is no basis for ranking ordinary household cues as stronger or weaker than scent decay without testing them against each other.

5.4 What a Better Test Would Require

A rigorous experiment would need to separate elapsed time from odor information. One approach would hold duration constant while experimentally changing the owner's scent profile; another would hold scent information as constant as possible while varying duration. Suitable control conditions would also be needed because adding or removing odor may itself change behavior.

Until such experiments exist, "dogs smell how long you have been gone" should not be presented as a discovered mechanism.

6. Memory for What, Where and When

6.1 Episodic-Like Memory for Observed Events

Fugazza, Pogány and Miklósi (2016) used unexpected recall tests after incidental encoding. Dogs reproduced actions that they had observed without knowing in advance that those actions would later need to be recalled. The authors interpreted this as evidence for episodic-like memory.

The important methodological point is the unexpected test: successful performance cannot be reduced simply to deliberately rehearsing material for a known later request.

6.2 Dogs Can Retain "When" Information

Lo and Roberts (2019) tested dogs with spatially and temporally unique odor sequences. The dogs showed memory for combinations of what-when, where-when and what-where-when.

This is directly relevant to claims that canine memory research has only tested "what" happened. Dogs can retain temporal information about past events. The study did not establish whether all components were bound into a single integrated episodic memory or retrieved as separate kinds of information, and it does not show that dogs monitor an owner's absence in the same way.

6.3 Memory of the Dog's Own Actions

Fugazza et al. (2020) showed that dogs could unexpectedly repeat their own previous actions after delays ranging from seconds to one hour. Performance declined as the retention interval increased, a pattern consistent with memory decay.

Again, this is evidence about memory across time, not evidence of a conscious stopwatch. Temporal information can influence memory without implying human-like autobiographical time experience.

6.3a Why Memory Research Sits Next to This Question

What-where-when work and interval timing answer different questions. A dog that chooses the box visited earlier is using stored information about order; a dog that responds after 30 seconds is measuring an ongoing interval. Both get filed under time, and the two capacities could in principle be present or absent independently.

The link to the household question is indirect in both cases. Retaining information about an earlier event is a precondition for noticing that time has passed since it, without being evidence that the passage itself was measured (what canine cognition research establishes).

6.4 Waiting Is Not a Pure Timing Test

A dog that remains in a stay, waits at a door or leaves food untouched until released is showing behavioral inhibition under learned stimulus control. Temporal learning may contribute, especially when release consistently follows a particular duration, but the behavior does not isolate an interval-timing mechanism.

The clean conclusion is therefore not "waiting is not timing" but: waiting by itself is not evidence that interval timing is what controls the response (what is known about canine inhibitory control).

7. Routine, Prediction and Temporal Cues

7.1 A Routine Provides Many Predictors

Daily household sequences contain alarm clocks, light changes, clothing, keys, vehicles, neighbors, feeding events and recurring movement patterns. A dog can learn these relationships without representing a human clock time.

If a dog regularly waits near the door before someone arrives, several hypotheses are possible: daily rhythmicity, external sounds, learned household sequences or other cues correlated with the person's routine. Current evidence does not justify assuming one is generally more likely than the others.

7.2 Predictability Is a Learning and Welfare Question

Predictability can reduce uncertainty when the predicted event is tolerable or desirable. When the predicted event is aversive, the same predictive sequence can instead produce anticipatory arousal or distress.

That principle comes from broader learning and welfare research, not from the canine time-perception experiments reviewed here. It should therefore be treated as a practical framework rather than as a result of temporal-bisection studies (how predictive cues can be changed).

7.2a What Predictability Buys and What It Costs

Predictable conditions are generally associated with better welfare in captive animals, and predictability is closely tied to control over what happens (what the work on controllability shows). A dog that can anticipate the sequence of a day has less to resolve moment by moment.

The same mechanism has a cost where the predicted event is aversive. A reliable pre-departure sequence makes the departure foreseeable too, which is why in separation-related cases those cues are usually part of the problem rather than a neutral warm-up (how predictive cues are worked with).

7.3 Flexible Routine Is Different From Chaos

Reliable access to sleep, food, activity, elimination and social contact does not require every event to occur at one exact minute. For many households it is useful to keep important needs predictable while allowing harmless variation in their timing.

This is a management principle rather than a demonstrated treatment derived from the timing literature.

8. Separation-Related Behavior Is a Separate Clinical Literature

8.1 Dogs With Separation Problems Cannot Be Inferred From Non-SRB Samples

Both direct duration experiments deliberately studied dogs that were already able to remain alone without known separation-related problems. Their results therefore cannot set a duration threshold for dogs with separation-related distress.

In clinical separation cases, behavior can change rapidly after departure and the relevant question is often whether the dog remains below a distress threshold at all, not whether four hours differs from six.

8.1a Why the Distinction Is Not Academic

A dog without separation-related problems and a dog with them differ in what the absence consists of, not merely in degree. In the first case the animal rests through a period of nothing happening; in the second it is in a state that behavior and physiology both register, often within the first minutes.

Applying duration findings from settled dogs to distressed ones therefore answers a question nobody asked. The relevant threshold in a clinical case is set by what the dog does after the door closes, not by a figure derived from animals for whom closing the door is uneventful.

8.2 Changes in Routine Can Matter

Harvey et al. (2022) followed 1,807 UK pet dogs across changes in owner routines during the COVID-19 period. The study did not experimentally test time perception, but it did show that changes in time left alone were associated with the risk of newly reported separation-related behaviors.

That belongs in this article because it prevents a category error: the behavioral consequences of changing an alone-time routine are important even when the mechanism is not "perception of elapsed time."

8.3 Departure, Absence and Reunion Are Three Observation Windows

No single phase is automatically the most informative. Pre-departure behavior may reveal responses to predictive cues. Video during the absence can show vocalization, pacing, orientation to exits, rest, activity and whether those patterns change over time. Reunion behavior adds information about the transition when the owner returns.

Rehn and Keeling (2011) is especially useful because it shows that effects can appear at reunion even when significant differences were not detected during equivalent portions of the absence. That is a reason to observe both, not to replace absence video with greeting behavior.

8.4 What a Household Record Can and Cannot Do

Repeated video of representative absences of different lengths can answer individual questions that population studies cannot: whether this dog settles, whether activity changes later in the absence, whether vocalization occurs, and whether reunion behavior differs after longer versus shorter absences.

There is no validated two-week recording period and no fixed number of observations that guarantees an answer. The record should instead sample enough normal household situations to show whether a stable pattern exists.

8.5 What Can Be Changed When an Absence Cannot Be Shortened

Where the hours are fixed by work, the adjustable variables are not temporal ones: whether the dog has emptied itself and moved beforehand, whether the resting place is quiet and comfortable, whether anything breaks the absence, and what the first minute after the return looks like.

None of that follows from timing research, and it does not need to. It follows from the ordinary requirements of a resting animal, and it is worth stating because households often search for a permissible number of hours instead (what is known about separation-related behavior).

9. Situations in Practice

9.1 The Dog That Appears at the Door Before Someone Arrives

This is evidence that something predicts the arrival, not evidence that the dog reads clock time. Possible cues include daily rhythmicity, building or traffic sounds, learned household events and cues that people themselves do not notice.

A practical test is to vary arrival time on multiple occasions while recording the dog. If door-waiting remains tied to the old time, rhythmic or environmental cues become more plausible; if it shifts with other predictors, those deserve attention.

9.2 The Early Riser

Feeding regimen is one candidate among several, because feeding frequency has experimentally altered activity patterns in beagles (Zanghi et al., 2012). Light exposure, household noise, elimination needs, previous reinforcement and age-related sleep changes may also contribute.

The evidence does not support a rule that feeding time should always be the first variable changed.

9.3 Extending an Absence From Four to Six Hours

No existing study can tell a household whether that change is acceptable for its individual dog. Silbermann and Gansloßer (2023) included six dogs accustomed to approximately six hours alone and found no significant duration effect across their 2-, 4- and 6-hour conditions, but that highly selected sample cannot define a general welfare limit.

The useful answer comes from the dog's own behavior, health, elimination needs, previous alone-time history and, where relevant, clinical assessment.

9.4 A Household Changes Working Hours

A sudden schedule change alters many correlated cues at once: departure sequence, light exposure, activity, feeding pattern, social contact and duration alone. It is therefore unnecessary to invoke a single "internal clock" to explain temporary restlessness.

Gradual change may be practically useful for some dogs, but Zanghi et al. (2012) should not be cited as experimental proof that shifting the clock time of a meal will shift the dog's entire rhythm in a predictable number of steps.

9.5 A Dog Rests During the Absence but Explodes at Reunion

Video showing prolonged resting is reassuring information about what the dog was doing, but it does not turn the preceding hours into a proven welfare-neutral period. Likewise, intense reunion behavior does not by itself prove hidden distress during the absence.

The two phases should be assessed separately. Reunion arousal can have its own learning history, while absence behavior provides direct information about what happened when the owner was away.

9.6 What a Household Can Record

A short log covers what the literature cannot: departure and return times, what the dog did in the first ten minutes after the door closed, when it settled, and what the first minute after the return looked like. A camera pointed at the resting place produces all of it without anyone changing their routine.

Two weeks of that answers the household-level version of the question — does this dog settle, how long does it take, and does the reunion look different after a longer absence — which is the resolution at which the question can currently be answered at all (why measurement precedes interpretation).

10. Summary at a Glance

Dogs can discriminate short durations. Temporal bisection and fixed-interval studies published in 2017 and 2019 show orderly seconds-range timing in trained dogs (Domeniconi & Machado, 2017; Macpherson & Roberts, 2017; Cliff et al., 2019).

Daily rhythm is a separate phenomenon. Dogs show strong rest-activity rhythms, and a study in 48 beagles found effects of age and once- versus twice-daily feeding on activity organization (Zanghi et al., 2012).

Feeding frequency is not the same as feeding time. Zanghi et al. did not experimentally move one fixed meal from one clock time to another, so the study should not be cited as proof that changing meal time shifts the whole rhythm.

Hours-long absence has barely been studied directly. Rehn and Keeling studied 12 dogs across 0.5, 2 and 4 hours; Silbermann and Gansloßer later studied 6 highly selected dogs across approximately 2, 4 and 6 hours.

No significant difference is not proof of equivalence. Small samples limit the size of effects those studies could detect.

Reunion behavior can change with separation duration. Rehn and Keeling detected several differences after the owner returned, but normal owner routines were retained, so the greeting is not a pure measure of the preceding absence.

Dogs can remember temporal information. What-when and what-where-when memory has been demonstrated in odor tasks, and dogs can unexpectedly recall their own previous actions after delays (Lo & Roberts, 2019; Fugazza et al., 2020).

The scent-clock idea remains untested. No controlled canine experiment has isolated owner-scent decay from elapsed time.

Separation-related behavior is not answered by time-perception studies. Changes in alone-time routine can be associated with SRB risk, but that is a welfare and behavioral question rather than proof of one specific timing mechanism (Harvey et al., 2022).

Reunion signals were not all positive — Lip licking and body shaking at the return also increased with the length of the separation in the 12-dog study, which complicates reading the greeting as simple pleasure (what those signals do and do not show).

Rhythm measures double as aging markers — Fragmented rest and increased nighttime activity are used as clinical indicators rather than as statements about time perception (what changes in aging dogs).

11. Research Gaps and Critical Appraisal

Direct hours-long duration studies are exceptionally few. The main direct experiments involve only 12 and 6 dogs, respectively, and both samples excluded established separation-related problems (Rehn & Keeling, 2011; Silbermann & Gansloßer, 2023).

Seconds and hours are not experimentally connected. Multiple studies demonstrate seconds-range interval timing, but no canine study has shown that the same mechanism tracks an owner's hours-long absence.

The olfactory account remains untested. Scent decay is a plausible cue, but no study has separated it from elapsed duration and competing household cues.

Home cues are heavily confounded. Light, temperature, traffic, neighboring routines, sleep-wake cycles, hunger and elimination needs all change with time and have not been experimentally separated in an owner-absence paradigm.

Dogs with clinical separation-related problems need their own duration research. Findings from dogs already comfortable being alone cannot be assumed to generalize to dogs showing distress.

Reunion measures are noisy. Owner behavior on return is part of the behavioral sequence and should be recorded rather than assumed irrelevant.

What-where-when memory does not settle subjective time experience. Dogs can retain temporal information without that result demonstrating human-like autobiographical time consciousness.

Age and cognitive decline remain underexplored in interval timing. Age effects are well documented for activity rhythms, but canine psychophysical timing across development and cognitive aging has received little direct study.

Individual differences remain poorly characterized. Existing timing studies are small and selected for dogs able to complete demanding trained tasks. We do not yet know how strongly timing performance varies with sensory factors, training history, temperament or breed.

No universal alone-time threshold follows from this literature. A welfare limit requires more than a time-perception experiment and will depend on the individual dog's behavior, health, history and environment.

Individual differences are unexamined. Whether some dogs track duration more accurately than others, and whether that relates to age, training history or temperament, has not been studied in any of these paradigms.

12. Conclusion

Dogs are not "timeless," and they do not need a human-like clock to organize their day. They show daily rest-activity rhythms, can discriminate seconds-range durations in controlled psychophysical tasks, and can retain temporal information about past events. Those are separate findings supported by different methods.

The household question is much less settled. Two direct studies have varied hours-long owner absence, but together they contain only eighteen dogs, selected for being able to remain alone without established separation-related problems. Rehn and Keeling (2011) found no significant behavioral differences during equivalent portions of 0.5-, 2- and 4-hour absences but did find several differences at reunion. Silbermann and Gansloßer (2023) found no significant behavioral or cortisol differences across approximately 2-, 4- and 6-hour separations in six dogs already accustomed to such durations. Neither study defines a universal safe number of hours.

The popular odor-decay explanation remains a hypothesis. For practice, the most defensible approach is therefore not to argue about whether a dog "knows" the clock. Record departure, absence and reunion; assess the individual dog's behavior across representative situations; distinguish clinical separation-related behavior from normal resting; and avoid turning small experimental samples into universal household rules.

Key Insights (Takeaways)

• "Sense of time" covers different capacities: circadian rhythm, interval timing, temporal memory and responses to hours-long absence.

• Dogs have repeatedly shown seconds-range interval timing in controlled tasks, not just in one experiment.

• Dogs can retain what-when and what-where-when information, but that does not prove a human-like conscious experience of elapsed time.

• Zanghi et al. (2012) tested once- versus twice-daily feeding, not simple shifts of a fixed meal from one clock time to another.

• Only two very small direct studies have compared hours-long owner absences, and both studied dogs without established separation-related problems.

• Reunion behavior is useful data, but it should be interpreted alongside what happened during the absence and the owner's own return behavior.

• The idea that dogs use the decay of an owner's scent as a clock remains untested.

• No universal number of hours that a dog can safely be left alone can be derived from this literature.

References

Cliff, J. H., Jackson, S. M. K., McEwan, J. S. A., & Bizo, L. A. (2019). Weber's law and the scalar property of timing: A test of canine timing. Animals, 9(10), 801. https://doi.org/10.3390/ani9100801

Domeniconi, C., & Machado, A. (2017). Temporal bisection task with dogs: An exploratory study. Psychology & Neuroscience, 10(1), 101–108. https://doi.org/10.1037/pne0000080

Fugazza, C., Pogány, Á., & Miklósi, Á. (2016). Recall of others' actions after incidental encoding reveals episodic-like memory in dogs. Current Biology, 26(23), 3209–3213. https://doi.org/10.1016/j.cub.2016.09.057

Fugazza, C., Pongrácz, P., Pogány, Á., Lenkei, R., & Miklósi, Á. (2020). Mental representation and episodic-like memory of own actions in dogs. Scientific Reports, 10, 10449. https://doi.org/10.1038/s41598-020-67302-0

Harvey, N. D., Christley, R. M., Giragosian, K., Mead, R., Murray, J. K., Samet, L., Upjohn, M. M., & Casey, R. A. (2022). Impact of changes in time left alone on separation-related behaviour in UK pet dogs. Animals, 12(4), 482. https://doi.org/10.3390/ani12040482

Lo, K. H., & Roberts, W. A. (2019). Dogs (Canis familiaris) use odor cues to show episodic-like memory for what, where, and when. Journal of Comparative Psychology, 133(4), 428–441. https://doi.org/10.1037/com0000174

Macpherson, K., & Roberts, W. A. (2017). On the clock: Interval timing and overshadowing in domestic dogs (Canis familiaris). Journal of Comparative Psychology, 131(4), 348–361. https://doi.org/10.1037/com0000083

Rehn, T., & Keeling, L. J. (2011). The effect of time left alone at home on dog welfare. Applied Animal Behaviour Science, 129(2–4), 129–135. https://doi.org/10.1016/j.applanim.2010.11.015

Silbermann, J., & Gansloßer, U. (2023). Factors influencing isolation behavior of dogs: A holder-based questionnaire and behavioral and saliva cortisol responses during separation. Animals, 13(23), 3735. https://doi.org/10.3390/ani13233735

Zanghi, B. M., Kerr, W., de Rivera, C., Araujo, J. A., & Milgram, N. W. (2012). Effect of age and feeding schedule on diurnal rest/activity rhythms in dogs. Journal of Veterinary Behavior, 7(6), 339–347. https://doi.org/10.1016/j.jveb.2012.01.004

Zanghi, B. M., Kerr, W., Gierer, J., de Rivera, C., Araujo, J. A., & Milgram, N. W. (2013). Characterizing behavioral sleep using actigraphy in adult dogs of various ages fed once or twice daily. Journal of Veterinary Behavior, 8(4), 195–203. https://doi.org/10.1016/j.jveb.2012.10.007