Twenty-fourth on the American Kennel Club’s 2025 registration list, and the smallest purebred entry in our 268-breed dataset. It is also the breed where we have to tell you most often that a number you will meet elsewhere is built on almost nothing.
A ceiling, and no floor
The American standard is a single sentence on size: “Weight - A well balanced little dog not to exceed 6 pounds.” The list of disqualifications then repeats it as a rule: “Any dog over 6 pounds in weight.”
So the standard is a limit, not a range. It names no minimum. It gives no height either, only a shape: “The body is off-square; hence, slightly longer when measured from point of shoulder to point of buttocks than height at the withers.”
That matters for reading the panel at the top of this page.
| What you are looking at | Where it came from |
|---|---|
| The 6 lb top of the range | the standard, and a show disqualification |
| The bottom of the range | our source chart, not the standard |
| The height figures | not in the standard at all |
| Males shown same as females | the standard never splits the sexes |
Real dogs run heavier than the show limit. UK veterinary records for 11,647 Chihuahuas give a median adult bodyweight of 3.4 kg (about 7.5 lb), with the middle half between 2.7 and 4.3 kg and the whole range from 0.8 to 9.8 kg. A pet Chihuahua that a ring would disqualify is closer to typical than to unusual.
Keep that study in mind. Those 11,647 dogs came out of 336,865 under care at 438 UK clinics in 2016, and the breed’s share of newborn puppies climbed from 1.01% in 2005 to 5.35% in 2016. It is a young population that grew fast, and that bends several numbers later on this page.
Where our own curve sits on a four-pound dog
Feed our model an adult weight of 4.5 lb and half of it arrives at about 14 weeks. The rest of the shape:
| Age | Share of adult weight | On a 4.5 lb adult |
|---|---|---|
| 8 weeks | 31% | 1.4 lb |
| 12 weeks | 44% | 2.0 lb |
| 16 weeks | 56% | 2.5 lb |
| 26 weeks | 78% | 3.5 lb |
| 52 weeks | 97% | 4.4 lb |
Now the part we would rather you knew. Our fifteen-point validation set is mostly built from measured dogs, but four points take their adult weight from a breed standard instead, and this is one of the four. We say so on our methodology page. It is not a small caveat: the worst point in the set, at +20.0%, is another of those four.
This one behaves well. Given a 2.2 lb puppy at 13 weeks, the model returns 4.6 lb against a published chart’s 4.75 lb, an error of −3.2%. That is the tightest of the four and better than the set’s overall 6.9% mean absolute error. But a good result on a weak input is still a weak input, and the outside check on that page covers dogs of 27.5 to 47.5 kg. Nothing audits us down here.
What has actually been measured about the temperament
Start with the ceiling on every breed claim. Surveying owners of 18,385 dogs and sequencing 2,155 of them, a genomics team put breed at roughly 9% of the behavioural difference between one dog and another. That paper is sharper still about the trait people attach to this breed: for agonistic threshold (how easily a dog is provoked by something frightening or uncomfortable), “breed is almost uninformative”.
Now the awkward part, which we will not skate over. Where snappiness has been counted, this breed is near the top.
The most-cited aggression survey, Duffy and colleagues, recruited 11 breeds through breed clubs and 33 breeds online, eight of them in both samples. The Chihuahua is in the online 33, with 56 dogs. Its row in the table counting dogs that scored the maximum for snapping, biting or trying to bite:
| Aimed at | Chihuahuas | Share |
|---|---|---|
| Strangers | 9 of 56 | 16.1% |
| Their own household | 3 of 56 | 5.4% |
| Unfamiliar dogs | 10 of 56 | 17.9% |
| Another dog in the same home | 2 of 42 | 4.8% |
Read that last denominator. Only households with a second dog can answer it, so it is 42, not 56. The authors’ summary is blunt: breeds with the greatest share of dogs showing serious aggression toward humans “included Dachshunds, Chihuahuas and Jack Russell Terriers”.
A Finnish survey of 9,270 purebred dogs across 22 breeds got there by a different road, and this breed is one of the 22. Its authors report that “Lagotto Romagnolo, Chihuahua, German Shepherd Dog, and Miniature Schnauzer had a significantly higher odds for aggressive behaviour than Golden Retriever and Labrador Retriever”. Do not build a prevalence out of it: those 9,270 were sifted from 13,715 by keeping the clearest cases and non-cases.
UK clinical notes agree. Aggression was recorded for 238 of 5,660 Chihuahuas in one year (4.2%, interval 3.7 to 4.8) at a median age of 3.5 years. Males ran higher than females, 4.9% against 3.5%, at P = 0.008. It is the fifth most common thing a vet wrote down about the breed.
The half of that finding about the owners
The Finnish paper also notes that owners of small dogs “play with and obedience train their dogs less frequently than owners of large dogs”. An Austrian survey of 1,276 owners split at 20 kg measured that directly. Owners of the smaller dogs were less consistent and did less training and play, each at P < 0.001. Within that group, consistency and regular training tracked with better obedience, while punishment went with more aggression and excitability.
None of that proves the owner causes it. It does mean size and handling arrive together in every survey, and only one of them is yours to change.
The first year
Weight is done between six and nine months, as the panel says, so growth management is not the job here. Two other things are.
The baby teeth may not leave. Retained deciduous dentition, a milk tooth still in place once the adult tooth is through, was recorded in 324 of 5,660 dogs, 5.7%, at a median age of 1.6 years. Third most common finding in the breed, and it is a mouth built too small for its own teeth. Raise it before the neuter, which is otherwise an anaesthetic you pay for twice.
Adolescence still happens. The clearest work on it followed 285 puppies (retrievers, German Shepherds and crosses) with owner questionnaires at 5, 8 and 12 months. A separate, smaller arm put 93 retrievers through a sit command at 5 and 8 months only. In that group of 93, the odds of the command being ignored at 8 months ran 2.14 times higher than at 5 months when the dog’s own carer asked, on an interval of 1.46 to 3.11. A stranger asking got better compliance, at odds of 0.40.
A British cohort puts a number on the peak. Undesirable behaviour of some kind was reported for 513 of 1,219 dogs at twelve months: 42.1%, the highest of its three surveys, against samples of 1,718 at six months and 684 at eighteen. Neither study held this breed, so take the timing and leave the rate.
What to raise at a check-up
Two kinds of evidence follow and they are not interchangeable. The stronger kind is primary-care records: every Chihuahua a set of clinics saw, sick or well. The weaker kind is the screening registry, where owners choose whether to send a result in. One study measured that choice instead of guessing at it. Among 93 dogs X-rayed for certification only 49 sets of films were sent, and 92% of those looked normal against 50% of the ones that stayed in the drawer, a submission rate 8.2 times higher for a good-looking hip. Every registry figure below is a floor.
The hip file is too small to say anything
174. That is how many Chihuahua hip X-rays the registry holds. Of them, 5.7% were graded dysplastic, placing the breed 176th of the 219 breeds ranked.
Do not quote that. Set 174 beside a Labrador Retriever’s 324,363 evaluations, or beside this breed’s own 5,471 kneecap exams. This is not a breed with good hips on the evidence. It is a breed whose breeders do not X-ray hips, and ten abnormal films is a fact about paperwork rather than anatomy. We say so plainly because the 5.7% circulates as reassurance, and on 174 dogs it is compatible with a great deal.
Kneecaps, where the file is real
A luxating patella is a kneecap slipping out of the groove it should run in. Here the registry has volume: 5,471 dogs examined, 6.4% graded abnormal, 22nd on that report and immediately behind the French Bulldog at 6.6% of 8,926.
This one is not read off an X-ray. The exam is done by hand from twelve months, with the dog awake (“chemical restraint is not recommended”) and graded 1 to 4. A grade 1 kneecap can be pushed out and drops straight back, and a grade 1 dog may never limp.
The population data is better still. Across 210,824 dogs at 119 English practices, 1.30% had been diagnosed with patellar luxation, on an interval of 1.21 to 1.39. Two columns of that paper get confused constantly, so here they are side by side:
| Breed | Dogs in the study | Share affected | Odds vs crossbreds |
|---|---|---|---|
| Pomeranian | 1,040 | 6.5% (3.9–9.2) | 6.5 (4.0–10.7) |
| Yorkshire Terrier | 6,980 | 5.4% (4.4–6.3) | 5.5 (4.3–7.1) |
| Chihuahua | 4,619 | 4.9% (3.8–6.0) | 5.9 (4.4–7.9) |
| French Bulldog | 1,280 | 4.0% (2.1–5.8) | 5.4 (3.1–9.3) |
The Chihuahua is third on raw share and second on odds. It leads neither. The columns disagree because the odds are adjusted for age, sex and bodyweight and the share is not. What the authors themselves wrote is that “The Pomeranian, Chihuahua and French Bulldog were particularly predisposed to patellar luxation.”
In the breed’s own clinical records, patellar luxation was diagnosed in 224 of 5,660 dogs in one year (4.0%) at a median age of 2.9 years. Watch for a skipped step: one back leg carried for a stride, then forgotten.
Legg-Calvé-Perthes: small file, real disease
This is the ball of the hip joint dying because its blood supply fails. The registry sees it in “the miniature and toy breeds between the ages of 4 months to a year”, and reads it off the same hip film used for dysplasia. A dog over twelve months with normal hips is normal for this too.
201 Chihuahuas have been screened and 1.5% came back abnormal, three dogs. That is 5th on the report. Rank alone sounds alarming, so anchor it: the Yorkshire Terrier is 1st with 3.0% of 574 and the Poodle 12th with 0.8% of 3,391. Few breeds submit at all, so fifth is fifth among a handful.
The practical version is age. A limp in a toy puppy between four months and a year is worth an X-ray, not a fortnight of rest.
The molera, and what it does not tell you
The soft spot is in the standard, which describes “A well rounded ‘apple dome’ skull, with or without molera”: permitted, not required, not a fault either way.
How common is it? A CT study of 50 Chihuahuas found 46 of them (92%) had at least one persistent fontanelle, 138 gaps across 118 skull sutures in all. Of the 138 gaps, 72 (52%) sat on top of the head, 49 were at the back and 17 at the sides. Counted by suture rather than by gap, 57 of the 118, 48%, were on top. Either way, feeling for the one soft spot people check tells you little about how many the dog has: the average was 2.8 gaps a dog, and 15% of the affected dogs had none at all at the classic bregmatic position.
Now the question everybody asks. The honest answer is that it is unresolved.
The breed club states the reassuring side directly: a domed head with a molera present “does not predispose the Chihuahua to this condition”. It cites work it says found no link between a fontanelle and hydrocephalus. We could not locate that study.
The measured side is messier. Part II of the same CT project concluded that persistent fontanelles “are associated with small size, SM, ventriculomegaly, and CCJ overcrowding”. SM is syringomyelia, fluid-filled cavities in the spinal cord. CCJ overcrowding means the skull base is too tight for what must pass through it.
Three reasons that still is not “a molera predicts hydrocephalus”.
Enlarged ventricles are not the disease. Ventriculomegaly means the fluid spaces in the brain measure large on a scan. Among 67 dogs imaged in one study, 25 had exactly that and no symptoms, against 22 with clinically relevant internal hydrocephalus and 20 normal brains. Telling the groups apart took a list of further imaging signs; a wide ventricle alone did not do it.
Fifty dogs is fifty dogs, referred to a neurology service. That is not a sample of pet Chihuahuas.
Something upstream may drive both. A 2025 study of another 50 Chihuahuas found more fontanelle area on shorter, taller, wider skulls, and the more extreme skull on the lighter dogs: “lower bodyweight is associated with more extreme brachycephaly and extreme brachycephaly is associated with PFs”. If skull shape drives both, the molera is a marker rather than a cause.
So a molera is not a diagnosis and not a reason to panic. It is also not a licence to ignore a puppy that circles, presses its head into a wall, seems blind or stops learning. Those signs earn a vet visit whatever the skull feels like.
Eyes: a clean screening record and one emergency
4,125 Chihuahuas have had eye exams and 2.6% were abnormal, 190th of the 219 breeds ranked, near the clean end. A real sample pointing a good direction, with the usual caveat about who submits.
The exposure is mechanical, not degenerative. A large eye in a shallow socket can be shoved forward past the lids, which then shut behind it. That is proptosis, a same-hour emergency. In a series of 100 dogs treated for it, “The most commonly reported breeds were Chihuahua (20), Shih tzu (19), mixed breed dogs (17)”. Seventy-seven came from bite wounds, and 54 of the hundred lost the eye. So the lesson is about other dogs rather than about eyes.
Teeth: the commonest entry in the file
Periodontal disease was the most recorded disorder in the breed: 762 of 5,660 dogs in one year, 13.5%, interval 12.6 to 14.4, median age 5.2 years. Add the 5.7% with retained baby teeth and the mouth is where this breed’s record is heaviest.
Set that against all dogs, and read the whole comparison rather than the flattering half. In 22,333 UK dogs followed for a year, 2,797 were diagnosed with dental disease, 12.5%. On that line the breed barely stands out. But the same analysis holds more rows:
| Comparison in that study | Figure |
|---|---|
| Dogs under 10 kg vs dogs of 30 to under 40 kg | 3.07 times the risk |
| Dogs of 12 years or more vs dogs aged 2 to under 4 | 3.91 times the risk |
| Flat-faced breeds vs the rest | 1.25 times the risk |
| Breeds at increased risk | 18 of them, led by the Toy Poodle at 3.97 |
The first row is the one that matters here, and the one usually left out. Being under 10 kg roughly triples the risk. And the breed’s 13.5% came from a population whose median age was 2.8 years, a young dog’s figure for a disease that compounds with age.
An owner-reported survey pushes the same way. Waltham researchers questioned 12,753 owners in the United States and Canada and found periodontal disease in 50.5% of dogs, with toy and small breeds above large ones. That is four times the UK figure because owner reports are a softer measure: a different question, not a contradiction. The Chihuahua is not one of the five breeds it names as raised, which is worth saying and does not rescue it from the size row.
Start home cleaning while the puppy still thinks it is a game.
One more entry from the same records belongs beside the teeth. Obesity was the second most recorded disorder (333 of 5,660 dogs, 5.9%) at a median age of 4.4 years. On a 3 kg frame, half a pound is a sixth of the dog, so our guide to spotting an overweight puppy matters more here than on a Labrador.
The windpipe
Tracheal collapse is the airway’s cartilage rings softening until it flattens as the dog breathes. Surgeons name this breed among those affected, alongside “Yorkshire terriers, Pomeranians, Poodles”, and describe the sound as a “harsh dry cough that sounds like a goose honking”. About 70% show some improvement on medical management alone.
There is no population prevalence here, only counts of dogs already referred. In a Seoul hospital series of 110 small dogs seen from 2022 to 2024, 8 were Chihuahuas, 7.2%, well behind the Maltese at 34. Two findings there travel better than the breed counts: mean age was 11.37 years, and X-rays confirmed the collapse in only 64.5% of cases. A clear chest film does not rule it out.
Use a harness rather than a collar from the first walk.
The heart, where the registry answers the wrong question
The registry’s Chihuahua cardiac file looks excellent: 1,967 dogs screened for congenital heart disease, 0.2% abnormal. Four dogs.
It is also the wrong question. That database covers malformations a puppy is born with, and the registry says plainly that a clear exam does not rule out disease that develops later. The later kind is what kills Chihuahuas. Heart murmurs were noted in 115 of 5,660 dogs at a median age of 8.9 years, and among the 80 deaths with a cause written down, heart disease was the largest group: 15 dogs, 18.8%, ahead of lower respiratory tract disorders at 13 and traumatic injury at 11.
So a congenital cardiac certificate says something real about the puppy’s plumbing at birth and nothing about its ninth year. Once the dog is middle-aged, ask for the chest to be listened to yearly.
Merle: a registry row that must not be read as a breed trait
Our screening extract contains a line that looks spectacular: 74 Chihuahuas tested for the merle gene, 74 carriers (100%), ranking the breed first.
Publish that as a breed characteristic and you have made an error worth understanding, because its shape is everywhere in registry data.
Merle is a coat pattern, not a disorder, and the DNA test costs money. Nobody tests a solid black Chihuahua with no merle in its pedigree. What gets tested is dogs that look merle, dogs bred from merles, and dogs suspected of the version you cannot see, the cryptic merle, where a shorter stretch of repeated DNA gives “no Merle pattern, no change to coat color” in a dog that can still pass it on. So those 74 are a sample of suspicions, not of the breed, and a test that confirms suspicions comes back positive.
The rule generalises: before quoting a registry percentage, ask who ended up in the denominator. Hip figures come from dogs whose breeders X-ray hips. Merle figures come from dogs somebody already thought were merle. Neither denominator is the breed.
What is known about merle comes from elsewhere. The Langevin study behind the cryptic merle above covered 181 dogs across 14 breeds without a single Chihuahua. Two copies of the gene are the risk: a double merle “may possess a wide range of auditory and ophthalmologic defects”. Stopping two merles being bred together is why the test exists. Looking at a merle puppy, ask what the other parent was and whether it was tested.
How long they live
The published figures are low, and why they are low matters as much as they do.
A UK life table built from 30,563 dead dogs gives life expectancy at birth of 7.91 years for this breed, from 453 Chihuahua deaths, on an interval of 7.62 to 8.19. Dogs overall came out at 11.23. The clinical study above found a median age at death of 8.2 years across its 102 deaths, against 12.0 for all breeds, and split by sex: females 10.2 years, males 6.9, unlikely to be chance at P = 0.005.
Three things explain most of that before any theory about the breed is needed.
The population is young and growing. A breed whose puppy share rose fivefold in eleven years is packed with young dogs, and they supply deaths at young ages to a table with too few old dogs to balance them. The life-table authors make this argument themselves about another fast-growing breed.
The spread is enormous. The same paper reports that for Chihuahuas, remaining life expectancy fell to 1.5 years only at age 15 to 16: “the highest of all the breeds, indicating a high variation of lifespans among Chihuahuas”. Plenty reach sixteen. The average is dragged down from the other end.
Accidents kill small dogs. Traumatic injury was 11 of the 80 recorded causes of death, and most of the proptosis cases above arrived in another dog’s mouth. This is the item you can act on.
A Japanese pet-cemetery dataset is widely quoted as giving 11.8 years here, which fits the picture better. Its abstract opened for us and its breed table did not on the routes we tried. Other routes into that table exist (our Pomeranian page quotes it), but we are not printing a Chihuahua figure we could not open ourselves.
What we could not stand behind
Any month-by-month Chihuahua weight chart. Adults here run from under 1 kg to nearly 10 kg. One chart cannot describe both, and the ones in circulation cite nothing.
The study behind the club’s molera position. The club refers to work finding no relationship between a fontanelle and hydrocephalus. We could not find it, so we cannot tell you its size or its method. Our reading of what is published is that the link is unproven in either direction, and “unproven” is not “disproven”.
A hydrocephalus rate for the breed. It is not among the commonly recorded conditions in 5,660 Chihuahuas, and the only mention of it in that paper is a note that a textbook lists it among 21 reported predispositions. A textbook’s list is not a measurement.
The registry’s hip result. 5.7% of 174 is on this page because leaving it out would be worse, not because it establishes anything.
Any population figure for tracheal collapse here. Everything we found counts dogs already referred to a specialist.
Our own adult weight for the breed. It rests on a standard that states only a maximum, and that maximum sits below the median weight of real dogs in the largest dataset we have. Our data flags the row medium confidence for that reason.