Hearing in Dogs: Range, Deafness and What the Measurements Establish

1.1 The Range

The reference measurement remains a study of four dogs — a Chihuahua, a Poodle, a Dachshund and a Saint Bernard — trained to respond to the presence or absence of a tone. Their hearing range was determined as roughly 63 Hz to 47 kHz at 60 dB SPL (Heffner, 1983).

The commonly quoted "67 Hz to 45,000 Hz" is a rounding of that result, and it describes where detection becomes possible at a fairly loud presentation level, not where hearing is good.

1.2 Where Hearing Is Actually Best

The more useful figure is rarely quoted. High sensitivity — frequencies detectable below 10 dB SPL — ran from about 4 kHz to 16 kHz, with the best sensitivity around 8 kHz (Heffner, 1983).

That band matters practically: it covers most of the acoustic energy in a human voice's consonants, in a whistle, and in the higher components of a squeak. It is where a dog hears best, not where it hears at all.

1.3 The Comparison With Humans

In the low range, roughly 125 to 500 Hz, canine and human thresholds are similar. Above that, dogs detect quieter sounds — in the 500 to 8,000 Hz range, by something in the order of 13 to 19 decibels, with the gap widening at higher frequencies.

Above about 20 kHz humans hear nothing and dogs continue for another two octaves. That is the genuinely large difference, and it concerns a range in which almost nothing in ordinary human communication occurs.

1.4 What "Better Hearing" Would Have to Mean

The claim that dogs hear better than humans conflates several different capacities, and separating them changes the answer each time.

Range — dogs extend roughly two octaves higher, which is the largest and clearest difference. Sensitivity — dogs detect quieter sounds across most of the shared range, by a substantial margin at higher frequencies. Frequency discrimination — how finely two tones can be told apart; here the human system is generally better, and canine data are sparse. Localisation — humans localise more precisely, which follows from the wider spacing of our ears and our better low-frequency resolution.

So dogs hear further up and more faintly; humans discriminate and localise better. "Four times better" describes none of these.

1.5 What Dogs Do Not Hear Better

Below about 60 Hz, humans hear slightly lower than dogs — down to roughly 20 Hz as pitch. The popular claim that dogs perceive infrasound has no support in the audiometric data.

Nor is canine hearing "four times better" in any measurable sense. That figure appears to be a garbled version of the frequency ratio, and it is not a statement about sensitivity, discrimination or anything else that could be quantified.

2. Where the Numbers Come From

2.1 The Method

Behavioural audiometry works by training the animal to signal detection. A tone is presented at varying frequencies and intensities, the animal responds or does not, and the threshold is the intensity at which detection becomes reliable.

It is slow and it produces the most direct answer available, because the animal itself reports what it perceives (in contrast to the inference problem elsewhere in behaviour research).

2.2 Four Dogs

The sample is the limitation. Four animals, one per breed, tested individually — which is normal for psychophysics and thin for a figure that has been generalised to the species for four decades.

2.3 The Finding That Contradicts the Folk Version

Heffner's question was whether body size predicts high-frequency hearing, since across species the ability to hear high frequencies is inversely related to the distance between the ears.

Within dogs, it does not. Across breeds spanning a twofold difference in functional head size, high-frequency hearing varied only from 41 to 47 kHz — and members of both the smallest and the largest breeds had the best high-frequency hearing (Heffner, 1983).

The low-frequency claim fares no better. The dog that heard lowest was the Poodle; the Saint Bernard came last.

2.4 What That Establishes

High-frequency hearing in dogs appears to be a species character rather than an individual or breed one. The widespread claim that large dogs hear low sounds better and small dogs high sounds better is not supported by the study it is usually attributed to.

2.5 A Modern Replication, With a Surprise

A recent study applied a staircase psychophysical method to determine thresholds at three frequencies, testing five dogs at each. Thresholds were 19.5 dB SPL at 0.5 kHz, 14.5 dB SPL at 4 kHz — both consistent with earlier work — and 8.5 dB SPL at 20 kHz, markedly lower than expected.

The authors raise a possible explanation worth noting: dogs produce vocalisations above 20 kHz, and selective pressure linked to intraspecific communication in social canids could account for unexpected sensitivity in that band.

3. What Hearing Is For

3.1 Localisation

Detecting a sound and locating it are different tasks. Localisation depends on comparing arrival time and intensity between the two ears, which is why the distance between them matters and why the loss of one ear changes the problem fundamentally.

3.2 What One-Sided Hearing Costs

The consequence of losing an ear follows directly from the mechanism. Detection is largely preserved — a dog with one working ear hears the doorbell — while the comparison that produces direction is gone.

What that looks like in practice is a dog that responds to its name and then searches, orients to the wrong side, or fails to find a thrown object it clearly heard land. It is a specific deficit, and it is invisible unless someone is looking for it.

3.3 Hearing Human Speech

The band in which dogs hear best overlaps substantially with human speech, and canine responsiveness to spoken cues is not in doubt. What the acoustic data add is a constraint worth knowing.

Most of the information distinguishing one word from another sits in the higher-frequency consonants rather than the vowels — which is the part of the signal that attenuates fastest over distance and is most easily masked by background noise. A cue that works at three metres in a quiet room is not the same signal at thirty metres in wind (with the gesture and signal literature covered separately).

Intonation carries in a different band and travels further, which is part of why tone often reaches a dog when the word does not.

3.4 Why the Ultrasonic Range Exists

The standard functional account is prey-related: small rodents produce and communicate in ultrasonic frequencies, and a predator able to hear them has an advantage. The staircase study's suggestion adds a second possibility — that canine vocal production above 20 kHz makes the band socially relevant as well.

Both are plausible functional accounts rather than demonstrated selection histories (the same standard that applies to other adaptive explanations).

3.5 Hearing and Emotion

Sound reaches emotional processing quickly, and startle responses to sudden noise occur before any evaluation of the source (with the arousal consequences documented). That pathway is the reason acoustic events feature so heavily in canine anxiety (with noise sensitivity the most commonly reported anxiety trait).

3.6 What Ear Position Does Not Tell You

Ear carriage in most breeds is a fixed morphological trait under breed-standard selection, not an index of hearing ability (with conformation carrying little behavioural information generally). A drop-eared dog is not hearing-impaired by conformation, and a prick-eared dog is not thereby more alert (as the domestication literature illustrates for morphological inference generally).

Ear movement, by contrast, is informative — and it is one of the more reliable behavioural indicators in stress research (with ear base position the strongest single indicator in one video-coded study).

4. How Hearing Develops

4.1 Puppies Are Born Deaf

The ear canals are closed at birth in dogs and typically open around two weeks of age, which means the first fortnight of life is spent without hearing. The developmental work at Jackson Laboratory placed the startle response to sound at roughly nineteen to twenty days, shortly after the canals open (Scott & Fuller, 1965).

That timing is not incidental to anything else in this article. It puts the onset of hearing at the start of the socialisation period, which is why acoustic experience during those weeks matters and why a puppy that never hears a household in its first months meets one later without any reference for it (with the developmental window covered separately).

4.2 Why This Matters for Deafness

The two facts sit awkwardly together. Hearing begins around two weeks; congenital sensorineural degeneration occurs across the same early weeks. A puppy affected bilaterally may hear briefly and then lose it, which is why the condition is described as congenital rather than as a failure of development.

For breeders the practical consequence is that a litter cannot be screened before the pathway has matured, and the window between maturation and placement is narrow.

4.3 What Early Acoustic Experience Does

The general principle is well established across species: sensory systems calibrate to the input they receive during development (with the same logic applying to other early experience). Sound-exposure protocols in breeding programmes rest on that principle rather than on canine outcome data, which do not exist in a form that would settle the question.

That is a reasonable inference from developmental neuroscience, stated as one.

5. Congenital Deafness

5.1 The Mechanism

Canine congenital sensorineural deafness is primarily caused by degeneration of the stria vascularis in the cochlea. The structure requires melanocytes to function, and where pigment cells are absent from the inner ear, the blood supply to the sensory epithelium fails and the hair cells die.

The deafness is not caused by whiteness itself; both traits arise from disrupted pigment cell development in specific tissues. That mechanism applies to particular pigmentation loci rather than to white coat colour in general, which has several genetic routes.

5.2 It Happens Early and Is Permanent

The degeneration occurs in the first weeks of life, after which the loss is complete and irreversible. A puppy is not born deaf in the usual sense — it loses the hearing it had, before most breeders would notice.

5.3 Two Genetic Routes

Two pigmentation loci carry most of the risk: piebald, on chromosome 20, which underlies the pattern in Dalmatians, Bull Terriers, English Setters and Australian Cattle Dogs; and merle, on chromosome 10.

They are separate mechanisms producing the same outcome through the same tissue.

5.4 What the Genetics Have Not Resolved

Association studies have identified regions linked to deafness in piebald breeds, and the results have been inconsistent between studies of the same breed — including the notable absence of MITF, a strong candidate on mechanistic grounds, from any positive finding.

The reading offered in that literature is that the analyses are underpowered, the trait is genuinely complex, or phenotyping differences are obscuring the signal. No single-gene test for congenital deafness exists in any breed.

6. How Often It Occurs

6.1 Dalmatians

The best-characterised population. In a survey of 900 Dalmatians, 72.0 percent had normal responses, 21.0 percent had unilateral absence and 7.0 percent bilateral absence or clinical deafness (Holliday, Cunningham & Strain, 1992).

A parallel study of 1,031 Dalmatians across three geographically separate areas found 8.1 percent bilateral and 21.6 percent unilateral, an overall 29.7 percent incidence (Strain et al., 1992).

6.2 The Sex Difference

Among the 900, females showed 24.0 percent unilateral and 8.2 percent bilateral abnormality against 17.8 percent and 5.7 percent in males (Holliday et al., 1992). A separate Dalmatian study found females 4.4 percent more affected, a difference that did not reach significance.

The sex effect is therefore suggested rather than settled.

6.3 Merle

Across 153 dogs genotyped for the merle allele, deafness prevalence overall was 4.6 percent unilateral and 4.6 percent bilateral. The split by genotype is the important part: dogs heterozygous for merle showed 2.7 percent unilateral and 0.9 percent bilateral, while homozygotes showed 10 percent unilateral and 15 percent bilateral (Strain et al., 2009).

The distinction is genotypic, not phenotypic. "Double merle" is used in practice as a description of appearance, and appearance is an unreliable guide — which is why the study genotyped from buccal cells rather than scoring coats.

No significant association was found with eye colour or sex — a finding that contradicts the widespread assumption that blue eyes indicate deafness risk in merle dogs (with the merle genetics covered in full separately).

6.4 Why Older Merle Figures Are Much Higher

Figures of 54.6 percent in homozygous and 36.8 percent in heterozygous merle dogs circulate widely. They come from a study limited to a small established population of one breed, and have been extrapolated to all merle-carrying breeds.

A UK study of 2,303 Border Collie puppies reported 2.8 percent overall — 2.3 percent unilateral and 0.5 percent bilateral. Breed matters more than the merle allele alone.

6.5 Why the Numbers Differ Between Studies

Three Dalmatian datasets give overall figures of 28 percent, 29.7 percent and 17.8 percent. The spread is not measurement error, and the reasons are worth separating.

Sampling. Dogs brought for testing are not a random sample of the breed, and the direction of the bias is unclear — breeders screening whole litters produce different figures from owners presenting a suspect puppy.

Time. The lowest figure comes from the dataset spanning 1992 to 2019, and part of that difference is a genuine decline rather than a methodological artefact (Lewis et al., 2020).

Definition. Unilateral and bilateral cases are pooled in some summaries and separated in others, and the two have very different practical significance (a definitional problem with direct consequences).

6.6 Unilateral Cases Outnumber Bilateral

Across every Dalmatian dataset the ratio runs the same way — 21.0 against 7.0 percent, 21.6 against 8.1 percent, 13.4 against 4.4 percent. Roughly three unilaterally affected dogs for every bilateral one.

That ratio is the reason the detection problem in section 8 matters as much as it does: the larger group is the one that looks normal.

6.7 The Finding That Should Encourage Breeders

Screening data from 8,955 Dalmatian puppies tested between 1992 and 2019 produced an overall prevalence of 17.8 percent — 13.4 percent unilateral, 4.4 percent bilateral — with heritability around 0.3 and a significant improving trend over the period (Lewis, Freeman & De Risio, 2020).

The authors attribute the decline to selection: breeders avoiding the roughly four to five percent of animals with the highest genetic risk. This is one of the few places in canine welfare where a documented breeding intervention produced a measured improvement.

7. The Pigment Associations

7.1 What Correlates With Risk

In the Dalmatian screening data, genetic correlations were large: +0.6 between deafness and blue irises, and −0.86 with a pigmented head patch (Lewis et al., 2020).

A patch — a solid area of colour present at birth — is strongly protective. Blue irises are associated with elevated risk.

7.2 Why the Merle Data Differ

The merle study found no significant association with eye colour (Strain et al., 2009). That is not a contradiction: different pigmentation loci, different breeds, different mechanisms of pigment cell distribution.

It does mean the eye-colour heuristic cannot be transferred between breeds, which is how most of the confusion arises.

7.3 What This Does Not License

Pigmentation correlates with risk at population level and does not diagnose an individual. A dog with blue eyes may hear normally; a dog with a patch may not. The correlation informs breeding decisions, not the assessment of the animal in front of you (a level-of-analysis distinction that recurs throughout).

8. Detecting It

8.1 Why This Section Carries the Practical Weight

Everything above describes a condition that is permanent, untreatable and reasonably common in specific breeds. What follows is the part that changes outcomes, because the difference between a dog that is identified and one that is not is entirely a matter of whether anyone looked.

8.2 Why Observation Fails

This is the finding with the most practical weight in the whole literature. Among the 900 Dalmatians, dogs with unilaterally absent responses were not clinically deaf and appeared to depend on their normal ear for auditory-cued behaviour — and they were often misidentified as normal by uninformed observers (Holliday et al., 1992).

A dog hearing on one side responds to its name, startles at noise, and reacts to the door. What it cannot do is localise, which shows up as looking in the wrong direction, missing a recall from behind, or failing to find a thrown object.

8.3 What BAER Measures

The brainstem auditory evoked response records electrical activity in the auditory pathway following a click stimulus. Electrodes are placed subdermally, a click is delivered to one ear while the other is masked with white noise, and several hundred to a thousand responses are averaged to extract the waveform.

Each ear is tested separately, which is what makes unilateral deafness detectable at all. A dog is classified as bilaterally hearing, unilaterally deaf, or bilaterally deaf (Strain et al., 2009).

8.4 What It Does Not Measure

BAER establishes whether the pathway conducts a response to a loud click. It is not an audiogram: it does not describe the frequency range, the thresholds, or the quality of what the dog hears.

A dog can pass BAER and have significant hearing impairment at particular frequencies. The test answers a categorical question, and it answers it reliably (unlike several of the behavioural assessments used elsewhere).

8.5 What the Test Is Like

Practically, it is brief and minimally invasive. Fine subdermal electrodes are placed at the vertex and near each ear, foam inserts or headphones deliver the clicks, and the recording takes minutes per ear.

Most adult dogs tolerate it without sedation; puppies frequently sleep through it. Sedation is used where a dog cannot settle, and it does not affect the response.

8.6 When to Test

Testing is usually performed from about five weeks, once the auditory pathway has matured — which places it inside the socialisation period (with the developmental timing set out separately). For breeds carrying piebald or merle, screening before placement is the only way to identify unilateral cases — and identifying them is what makes selection against deafness possible, as the Dalmatian trend shows.

9. Losing It Later

9.1 A Different Category

Congenital deafness is present from the first weeks and does not change. Acquired loss develops in an animal that has heard normally, which makes it harder to notice and easier to misread — there is no moment at which the dog obviously stopped hearing.

9.2 Age-Related Loss

Hearing declines with age in dogs as in humans, typically affecting high frequencies first and progressing gradually. The clinical picture is well described; systematic longitudinal data on canine presbycusis are considerably thinner than the confidence with which it is discussed.

9.3 How It Presents

Almost never as an owner noticing deafness. It presents as a dog that has become less obedient, sleeps through arrivals, is startled when approached from behind, or barks at nothing.

The behavioural reading is available and usually taken. The sensory one requires someone to consider it (the same interpretive gap documented for other conditions).

9.4 The Differential That Matters

An older dog showing disorientation, altered sleep, changed interaction and apparent unresponsiveness may be losing its hearing, developing cognitive dysfunction, or both. The two overlap substantially in presentation and are distinguished by different assessments.

9.5 What Can and Cannot Be Done

There is no treatment for sensorineural hearing loss in dogs, congenital or acquired. Hearing aids have been trialled and are not in routine use — the fitting problem is substantial and tolerance is limited.

What changes outcomes is management: moving cues to modalities the dog still has, adjusting safety arrangements, and recognising the loss early enough that the behavioural consequences are not misattributed.

9.6 Noise Exposure

Occupational and environmental noise damages hearing in every species examined. By analogy with occupational exposure limits in humans, working dogs in aircraft, vehicle and firearms environments — and dogs housed in kennels with sustained high sound levels — encounter intensities that would trigger hearing protection requirements in a human workplace.

Direct measurement of noise-induced hearing loss in dogs is sparse, and the mechanism is not species-specific. This is an extrapolation, stated as one (with kennel environments carrying documented physiological costs).

10. What This Means in Practice

10.1 What the Evidence Supports Doing

Four things follow from the material above, and they are ordered by how much evidence stands behind them.

Test, do not observe — directly supported by the misidentification finding. Screen whole litters in at-risk breeds — supported by the population decline in Dalmatians. Read the localisation signs — supported by the description of how unilaterally deaf dogs behave. Consider hearing before temperament in an older dog that has become unresponsive — clinical reasoning rather than a study finding, and stated as such.

10.2 Test Rather Than Observe

For a puppy from a piebald or merle breed, observation cannot rule out unilateral deafness — the finding that owners and breeders routinely misidentify these dogs as normal is directly documented (Holliday et al., 1992).

10.3 Reading the Signs of One-Sided Loss

Consistently looking the wrong way for a sound. Poor recall when called from behind or from the side. Difficulty finding a thrown toy by sound. Sleeping unusually deeply and startling when touched.

None of these is diagnostic on its own, and each is worth acting on — the test settles it in minutes, and nothing else will.

10.4 Working With a Deaf Dog

Visual and tactile cues replace acoustic ones, and vibration is often more reliable than either for gaining attention at a distance. The training is not different in kind — the same learning processes apply, with a different signal modality (as the learning literature sets out).

The safety consideration is real and specific: a dog that cannot hear a vehicle needs management rather than trust.

10.5 What Not to Assume

That a dog ignoring a cue is being disobedient. That a startle response proves hearing is intact — it proves one ear works. And that a hearing test at eight weeks covers later loss, which it does not.

10.6 For Breeders

The Dalmatian data show what is achievable. Screening every puppy in a litter — not only the ones that seem affected — identifies the unilateral cases that carry the genetic risk forward invisibly, and selection against them produced a measurable population decline across nearly three decades (Lewis et al., 2020).

The pigment correlations give a rough prior rather than a decision rule: a pigmented head patch was strongly protective, blue irises associated with elevated risk (Lewis et al., 2020), and neither substitutes for testing the individual.

10.7 Where Hearing Meets Behaviour

Two of this library's most-used articles connect here directly. A dog with reduced hearing may become more reactive to what it can still detect, and a dog that startles when approached has a reason that is not temperament (with the reading error documented).

Conversely, sound sensitivity is not a hearing deficit. A noise-fearful dog usually hears perfectly well (with the fear mechanism described separately).

11. Where the Popular Claims Come From

11.1 The Pattern

Several of the confident statements in circulation share a structure: a real finding, generalised past what it supports, then repeated until the qualifier disappears.

11.2 "Dogs Hear Four Times Better"

There is no measurement this could report. It appears to be a mangled version of the frequency ratio — dogs reach roughly 45 kHz against a human 20 kHz — converted into a claim about acuity, which is a different property entirely.

11.3 "Big Dogs Hear Low Frequencies Better"

This has a plausible mechanism behind it, which is why it survives: across species, larger interaural distance does correlate with lower high-frequency limits. Within dogs it was tested directly and was not found, and the Saint Bernard was the worst low-frequency performer in the sample (Heffner, 1983).

11.4 "Blue Eyes Mean Deafness"

This one is partly right and breed-specific. In Dalmatians the genetic correlation between blue irises and congenital deafness was +0.6 (Lewis et al., 2020) — substantial. In merle dogs, no significant association with eye colour was found (Strain et al., 2009).

Transferring the heuristic between breeds is where it becomes wrong, and it is transferred constantly.

11.5 "White Dogs Are Deaf"

The colour is not the cause; both follow from the same absence of melanocytes, and the association holds only for the specific loci and tissue involved. Plenty of white dogs hear normally, and the pigmentation patterns that carry risk are identifiable (with the merle case documented in detail).

11.6 Why These Persist

Each has a kernel of something true, each is easier to state than the qualified version, and none is costly to believe — which removes the pressure that would otherwise correct them (a pattern documented across canine practice).

12. Summary at a Glance

The range comes from four dogs — Roughly 63 Hz to 47 kHz at 60 dB SPL, measured in a Chihuahua, Poodle, Dachshund and Saint Bernard (Heffner, 1983).

Best hearing is 4 to 16 kHz — With peak sensitivity around 8 kHz, detectable below 10 dB SPL (Heffner, 1983).

Size does not predict high-frequency hearing — Across a twofold range in head size, high-frequency limits varied only from 41 to 47 kHz, with the smallest and largest breeds both performing best (Heffner, 1983).

A modern replication found unexpected 20 kHz sensitivity — 8.5 dB SPL, markedly lower than expected, possibly related to canine vocal production above 20 kHz.

Deafness follows pigment through the cochlea — Absence of melanocytes causes degeneration of the stria vascularis in the first weeks of life; the loss is permanent.

Dalmatian prevalence is high and falling — 17.8 percent overall across 8,955 puppies screened 1992–2019, with heritability around 0.3 and a significant improving trend attributed to selection (Lewis et al., 2020).

Homozygous merle carries the higher risk — 10 percent unilateral and 15 percent bilateral in the studied population, against 2.7 and 0.9 percent in heterozygotes (Strain et al., 2009). The distinction is genotypic, and appearance is an unreliable proxy.

Unilateral cases outnumber bilateral by roughly three to one — Consistently across Dalmatian datasets: 21.0 against 7.0, 21.6 against 8.1, 13.4 against 4.4 percent.

Puppies are born deaf — Ear canals open around two weeks, with the startle response to sound appearing at roughly nineteen to twenty days (Scott & Fuller, 1965).

Unilateral deafness is routinely missed — Dogs with one absent response are not clinically deaf and were often misidentified as normal by uninformed observers (Holliday et al., 1992).

13. Research Gaps and Critical Appraisal

The audiogram rests on four animals. Heffner (1983) is careful psychophysics with an n of four, one per breed. Everything quoted about canine hearing range derives from it, and the modern replication tested five dogs per frequency at three frequencies.

Sensitivity data are thinner than range data. The frequently quoted endpoints describe detection at 60 dB. How well dogs discriminate frequency, localise, or hear in noise is far less well characterised.

Prevalence figures vary by an order of magnitude between studies. Merle deafness has been reported at 9.2 percent overall in one genotyped sample and above 50 percent in homozygotes in another, with breed and sampling explaining much of the difference.

No genetic test exists. Association studies in piebald breeds have produced inconsistent loci between studies of the same breed, and MITF has not been positively identified despite being a strong mechanistic candidate.

BAER answers a categorical question. It establishes conduction, not audibility across frequencies, and a dog that passes may still have functionally significant impairment.

Presbycusis is under-studied. Age-related hearing loss in dogs is clinically familiar and lacks the longitudinal data available for humans.

Frequency discrimination is barely characterised. Detection thresholds are known; how finely dogs distinguish one frequency from another, and how that degrades in noise, is not.

No canine data on early acoustic experience. Sound-exposure protocols in breeding rest on general developmental principles rather than on outcome studies in dogs.

Noise-induced loss is largely extrapolated. Working and kennelled dogs are exposed to intensities that would trigger protection requirements in human workplaces, and direct canine measurement is sparse.

14. Conclusion

Canine hearing is described with more confidence than its evidence base supports, and the specific failures are instructive. The range everyone quotes comes from four dogs tested in 1983 — good psychophysics, small sample, and generalised to the species ever since. The claim that big dogs hear low frequencies better is not merely unsupported; it is contradicted by the study it is attributed to, in which the Poodle heard lowest and the Saint Bernard heard worst. What that work did establish is more useful anyway: hearing is best between 4 and 16 kHz, and high-frequency ability appears to be a species character rather than something that varies with breed or head size. On deafness the evidence is stronger and the practical implications sharper. Congenital sensorineural deafness follows pigmentation through a specific mechanism — melanocyte absence causing cochlear degeneration in the first weeks of life — with piebald and merle as the two main routes and prevalence varying widely by breed. The finding that deserves the most attention is that dogs deaf in one ear are not clinically deaf, respond normally to most things, and were routinely misidentified as hearing by observers who did not know what to look for. They can only be found by testing each ear separately. The ratio matters as much as the prevalence: unilateral cases outnumber bilateral ones roughly three to one across every Dalmatian dataset, which means the larger affected group is the one that passes casual observation. And the Dalmatian screening data supply something this field rarely produces: evidence that a breeding intervention worked, with prevalence and genetic risk both declining across nearly three decades because breeders selected against it (unlike the conformational problems where selection has run the other way).

Key Insights (Takeaways)

• The famous numbers come from four dogs, and one popular claim is contradicted by them. The canine hearing range of roughly 63 Hz to 47 kHz derives from a study of a Chihuahua, Poodle, Dachshund and Saint Bernard (Heffner, 1983). Across that twofold range in head size, high-frequency hearing varied only from 41 to 47 kHz, and the Poodle — not the Saint Bernard — heard lowest. Size does not predict hearing in dogs.

• Where dogs hear best is more useful than where they hear at all. High sensitivity runs from about 4 to 16 kHz with a peak near 8 kHz (Heffner, 1983) — the band containing most of the acoustic information in a spoken cue or a whistle. The quoted endpoints describe detection at a loud presentation level.

• Deafness follows pigment through a specific cochlear mechanism. Melanocyte absence causes degeneration of the stria vascularis in the first weeks of life, producing permanent loss. Piebald and merle are the two main routes, and no genetic test exists — association studies in the same breed have produced inconsistent loci.

• Unilateral deafness is routinely mistaken for normal hearing. Dogs with one absent BAER response are not clinically deaf, depend on their working ear, and were often misidentified as normal by uninformed observers (Holliday et al., 1992). Each ear has to be tested separately, and in Dalmatians unilateral cases outnumber bilateral ones by roughly three to one.

• Dogs hear further up and more faintly; humans discriminate and localise better. "Better hearing" conflates range, sensitivity, frequency discrimination and localisation, and the four give different answers. Dogs extend roughly two octaves higher and detect quieter sounds across most of the shared range; humans distinguish similar frequencies more finely and localise more precisely.

• Selection against deafness demonstrably worked. Across 8,955 Dalmatian puppies screened between 1992 and 2019, prevalence fell significantly, with heritability around 0.3 and genetic correlations of +0.6 with blue irises and −0.86 with a pigmented head patch (Lewis et al., 2020). The improvement is attributed to breeders avoiding the highest-risk four to five percent.

References

Heffner, H. E. (1983). Hearing in large and small dogs: Absolute thresholds and size of the tympanic membrane. Behavioral Neuroscience, 97(2), 310–318. https://doi.org/10.1037/0735-7044.97.2.310

Holliday, T. A., Cunningham, J. G., & Strain, G. M. (1992). Unilateral and bilateral brainstem auditory-evoked response abnormalities in 900 Dalmatian dogs. Journal of Veterinary Internal Medicine, 6(3), 166–174. https://doi.org/10.1111/j.1939-1676.1992.tb00332.x

Lewis, T., Freeman, J., & De Risio, L. (2020). Decline in prevalence of congenital sensorineural deafness in Dalmatian dogs in the United Kingdom. Journal of Veterinary Internal Medicine, 34(4), 1524–1531. https://doi.org/10.1111/jvim.15776

Scott, J. P., & Fuller, J. L. (1965). Genetics and the Social Behavior of the Dog. University of Chicago Press.

Strain, G. M., Kearney, M. T., Gignac, I. J., Levesque, D. C., Nelson, H. J., Tedford, B. L., & Remsen, L. G. (1992). Brainstem auditory-evoked potential assessment of congenital deafness in Dalmatians: Associations with phenotypic markers. Journal of Veterinary Internal Medicine, 6(3), 175–182. https://doi.org/10.1111/j.1939-1676.1992.tb00333.x

Strain, G. M., Clark, L. A., Wahl, J. M., Turner, A. E., & Murphy, K. E. (2009). Prevalence of deafness in dogs heterozygous or homozygous for the merle allele. Journal of Veterinary Internal Medicine, 23(2), 282–286. https://doi.org/10.1111/j.1939-1676.2008.0257.x