This is the one page on this site that reduces alarm rather than raising it, and it comes with two conditions.
Condition one: if your dog has already eaten something and you are frightened, this is the wrong page. Call your own veterinarian, your nearest 24-hour emergency animal hospital, the ASPCA Animal Poison Control Center at (888) 426-4435, or the Pet Poison Helpline at (855) 764-7661 (both US, 24/7, a fee may apply). Come back and read this afterwards.
Condition two: "not on a toxicity list" is not the same as "safe to feed". Nothing on this page is a recommendation to give your dog anything.
This page is general information from a dog parent who reads veterinary sources, not a veterinarian.
Start with what the reference source actually lists
The Merck Veterinary Manual's food hazards article, retrieved 2026-08-28 (reviewed October 2020, last updated May 2026), is the consumer-facing summary of the food toxicoses its toxicology section covers. The article focuses on a short list: avocado, bread dough, chocolate, macadamia nuts, raisins and grapes, and xylitol.
Six or seven items, depending on how you count raisins and grapes.
Now compare that to the length of the average "foods toxic to dogs" listicle, which routinely runs to twenty, thirty or fifty entries.
The gap between those two numbers is what this page is about, and it is worth understanding correctly rather than triumphantly, because the gap has three different explanations and only one of them means "the internet made it up".
The three reasons something is on the internet list but not the reference list
1. It is genuinely documented, just filed elsewhere
Onion and garlic are a real, documented toxicosis with a delayed presentation. Merck covers Allium toxicosis, and this site has several pages on it. It is not on the short food hazards list because the manual files it under its own heading.
Caffeine is named by Merck as one of the two primary toxic principles in chocolate. It is documented; it lives inside the chocolate entry.
Alcohol is covered under toxicoses from alcohols.
Salt has its own toxicosis page.
Moldy food produces tremorgenic mycotoxins, with a dedicated page of its own.
So a long list is not automatically a wrong list. A great deal of what appears on those listicles is real and simply cataloged differently. Absence from one article is not absence from the literature, and this is exactly the mistake a page like this could make and does not.
2. It is a mechanical or dietary problem rather than a toxicity
Several regulars on the internet list are not poisons at all. They are things that cause trouble for physical reasons.
Cooked bones are an obstruction and perforation risk. Not a toxin.
Fatty trimmings are a recognized pancreatitis trigger. Not a toxin.
Corn cobs are a classic obstruction. Not a toxin.
Large stones and pits are objects. Not toxins.
Rich or unfamiliar food in quantity causes gastrointestinal upset in dogs as it does in people.
These belong on a hazard list. They do not belong on a poison list, and conflating them is why the lists get so long and so undifferentiated that people stop reading them.
3. It is genuinely overstated
And this is the smallest category, which is itself the finding.
Dairy and milk. Many adult dogs digest lactose poorly, and poor digestion is not poisoning. The realistic outcome of a saucer of milk is loose stools, not an emergency.
Bread, plain and baked. The danger with dough is the live yeast, which is dead in baked bread. Plain baked bread is not the hazard raw dough is. What is in the bread is a separate question, and raisin bread is a completely different conversation.
Peanut butter. The peanut is not the problem. The problem is whether that jar contains xylitol, which is a label question rather than a food question, and it is the single most consequential thing on this page.
Apples and their seeds. The flesh is not a documented toxicosis. The seeds contain a cyanogenic compound and the quantity involved in a dog eating an apple is not what the alarm implies. The core is an obstruction question, which is category two rather than category three.
Tomatoes. The green parts of the plant are the part with a concern attached, not a ripe tomato in a salad.
Avocado, which is a genuinely split case and has its own page: Merck states that all parts can cause poisoning in animals, and the severity of that varies enormously by species, which is why the internet contains both "deadly" and "superfood" pages about it. The full picture is in separating the real avocado risk from the internet version.
Why any of this matters, practically
Because a list that cries wolf about apples gets ignored about xylitol.
That is the whole argument for this page. An owner who has been told that fifty foods are deadly, and who has watched their dog eat several of them without consequence, learns that the list is unreliable. That owner is then slower to call on the day it matters, and the day it matters is a xylitol day or a grape day or a raisin-in-a-scone day.
Accuracy is a safety feature. A short true list is more protective than a long alarming one.
What this page still will not do
It will not tell you any food is safe to feed your dog. Not on a toxicity list is not a nutritional recommendation, and this site does not give dietary advice.
It will not give you a quantity for anything. No thresholds, ever.
It will not tell you to skip a call. If your dog has eaten something and you are worried, the call is free of downside and the reassurance is worth more coming from someone who can ask questions. That applies to the items in category three as much as to anything else, because you may be wrong about what your dog actually ate, and because a normal-looking dog is not evidence, as covered in why a normal-looking dog can still be in serious trouble.
It will not stay accurate on its own. Formulations change, particularly for anything sugar-free, and a food that was fine last year may contain something different now. This page is dated for that reason.
FAQ
What foods are actually toxic to dogs?
Merck's food hazards article focuses on avocado, bread dough, chocolate, macadamia nuts, raisins and grapes, and xylitol. Others, including onion and garlic, alcohol, salt and moldy food, are documented elsewhere in the same manual rather than being absent from it.
Is milk bad for dogs?
Many adult dogs digest lactose poorly, which is a digestive issue rather than a poisoning. That is not a recommendation to give it to them.
Are apple seeds poisonous to dogs?
They contain a cyanogenic compound, and the alarm around a dog eating an apple exceeds what the quantity supports. The realistic risk from an apple core is obstruction rather than poisoning.
Is peanut butter safe for dogs?
The peanut is not the issue. Whether the jar contains xylitol is, and that is a label question. Read the full ingredient list every time, including on a jar you have used before.
If a food is not on the toxic list, can I feed it to my dog?
This page does not say that and does not mean it. Not being a documented toxicosis is not a nutritional recommendation, and diet is a conversation with your vet.
Sources read at the source on 2026-08-28 and named above: the Merck Veterinary Manual, Special Pet Topics, Poisoning, "Food Hazards" (reviewed October 2020, last updated May 2026), for the scope of the food hazards it covers and for its statement that all parts of the avocado can cause poisoning in animals; the Merck Veterinary Manual, Toxicology, Food Hazards, "Chocolate Toxicosis in Animals" (reviewed September 2024), for caffeine and theobromine as the primary toxic principles; the Merck Veterinary Manual, Toxicology, "Salt Toxicosis in Animals" (last updated September 2024) and "Toxicoses From Alcohols in Animals", for the existence of those separate entries; the Merck Veterinary Manual, Toxicology, Mycotoxicoses, "Tremorgenic Neuromycotoxicosis in Dogs" (full review November 2021, last updated September 2024), for mold-derived toxins being documented separately.
The three-category framework, and the argument that an over-long list erodes the trust that a short accurate one earns, are this site's own reasoning and are labeled as such. Where this page says something is not a documented toxicosis, that is a statement about the reference source's coverage and explicitly not a statement that a food is safe to feed.
This page is general information, not veterinary advice, and it does not diagnose or treat. If your dog has eaten something, the call comes first.