In post-war America, one of the country’s most influential philanthropic organisations put serious money behind an unusual scientific question: could intelligence be selectively bred? The answer was supposed to emerge from dogs. Starting in 1945, over 500 dogs were observed over about 13 years at a special experimental building built for this purpose in Maine.
Cocker spaniels, beagles, basenjis, Shetland sheepdogs, and wire-haired fox terriers underwent intensive scrutiny and behavioural experiments from the first days of their lives.
The project was eventually nicknamed the “School for Dogs.” Its original ambition, however, went far beyond understanding puppy behaviour.
Backed by Rockefeller Foundation funding worth $632,000 at the time — equivalent to roughly $9 million today according to the BBC's account — researchers were interested in whether traits such as intelligence, temperament and trainability could be traced strongly enough to heredity to make selective breeding possible.
What emerged was considerably more complicated.
The research did not produce the highly intelligent “super-dog” that some of its backers had imagined. Instead, it produced evidence that challenged the idea that a single superior combination of inherited traits could simply be bred into existence.
Rockefeller’s big bet on canine intelligence
This was an especially complex time in the history of genetics.
It was a period when all the brutality associated with Nazi race theories and the holocaust left the credibility of the science of eugenics in disrepute. Yet the underlying question of how heredity influenced behaviour continued to attract scientific interest.
Based on the BBC’s report, one of the people who showed interest in this area was Alan Gregg, who at the time was a senior member of the Rockefeller Foundation. Gregg was specifically concerned with the aspects of human behaviour related to social pathology and psychiatry.
Dogs offered an attractive alternative research subject.
Unlike most laboratory animals, dogs already displayed extraordinary variation. One Great Dane and one Pomeranian can be entirely different from each other even though they are in the same species. Scientists thought that maybe their physical differences extended to their behaviour too.
The Rockefeller Foundation provided the funding for an unusually ambitious study. Researchers John Paul Scott and John Fuller took on the project, assembling a population of dogs representing several breeds and establishing a facility where their development could be followed systematically.
The scale was impressive for that time period. The puppies were studied from birth, when researchers recorded their developmental stages, and then subjected them to various tests aimed at determining their level of intelligence and other abilities.
The most fundamental question proved to be quite straightforward: if the genes significantly determine intelligence, would it be possible to select proper dogs and breed the next more intelligent generation?
The results would make that question much harder to answer.
500 puppies produced a problem for the ‘super-dog’ idea
These tests were carried out on the dogs in various ways, such as learning and problem solving.
The notion that a certain breed would stand out as being the best did not occur.
Although there were some breeds which showed excellence in certain tasks, this did not mean that the best overall breed was found because some breeds could do well in one task and yet poorly in another.
That distinction became important.
This study did not present a simple hierarchy of intelligence in dogs but instead showed that some traits might be differently spread between the breeds or even particular animals.
The same tendency could be observed in regard to temperament, studied by the scientists.
Thus, fox terriers might exhibit aggressive behaviour in some situations, which was explained by their hunting past. But behaviour was not fixed simply by breed. The surroundings and circumstances in which a dog found itself could alter its response.
As the BBC's account of the research notes, Scott and Fuller ultimately argued against the notion that selective breeding could produce an all-purpose “super-dog” possessing every desirable quality.
That finding also cut against a much broader assumption: that desirable human characteristics could be concentrated through selective breeding in the same straightforward way.
The researchers came to see variation itself as useful.
However, a social species need not be advantaged by all its members acting the same way. The same trait that makes one dog valuable in one circumstance can be quite useless to another dog in a different environment. The variety of behaviour can be what gives the group an edge over the population where everyone is behaving the same way.
In other words, the test aimed at proving dog superiority ended up showing how important behavioural variety can be.
The surprising discovery about puppy socialisation
The most enduring part of the research may have had little to do with intelligence.
As the project developed, Scott and Fuller became increasingly interested in what happened when puppies were deprived of normal social experiences.
In one extreme case, the puppies were raised in isolation. They were allowed to roam in big fenced outdoor spaces and were supplied with food and water but had no human contact and no dog contact either.
The researchers differed in terms of how long the puppies were isolated.
It turned out that there is a critical period for socialisation, which is around four to 12 weeks old. The puppies that received proper social interaction within that period could learn to be socially active. Otherwise, they found it harder to adapt.
This discovery brought some focus on the significance of early experiences rather than innate behaviour.
Scott later described isolated puppies as becoming increasingly difficult to socialise as they grew older. A few of the animals proved to be untamable and had to be put down, something that would pose serious ethical problems by today’s standards.
Nonetheless, the research did prove influential in the developing field of canine psychology.
Today, the concept of the impact of early experiences on how the puppies will interact with humans, other dogs, and new surroundings is a critical topic for puppy socialisation. The BBC has recognised the research done by Scott and Fuller as part of the history of canine psychology.
Why Rockefeller’s experiment still matters today
The original Rockefeller-backed project eventually wound down after the foundation declined to continue funding it. The anticipated super-dog had never appeared.
But the research did not disappear with the facility.
Scott and Fuller's work helped establish dogs as serious subjects for behavioural and cognitive research. Their experiments also demonstrated why the relationship between genes and behaviour could not easily be reduced to a simple formula.
Breed mattered in some circumstances. Individual variation mattered. Environment mattered. And, crucially, when an experience happened could matter too.
That makes the project particularly interesting today.
Dog research in today’s world does not require such a massive construction and the breeding of several hundred puppies in a lab.
The magnitude of the experiment conducted in Maine will probably not be repeated by anyone, primarily due to the fact that some of the techniques used could not be considered acceptable from the standpoint of current norms of animal welfare.
But there is still one lesson from the experiment worth remembering: striving to determine the index of a dog’s intelligence may lead us astray from its various other capabilities.
The Rockefeller Foundation had helped finance a search for a dog that could demonstrate the power of inherited intelligence. Instead, the project helped expose the limitations of that idea.
The “super-dog” never arrived. The bigger discovery was that there may be no single recipe for a superior dog — and that what happens during puppyhood can matter enormously.
