This is a working selective-breeding program for Rattus rattus, the black or roof rat, running eight years and many generations. As far as I am aware, it is the only domesticated colony of this species in existence.
The premise is straightforward. Take a species that has lived alongside humans for millennia as a commensal, but has never been intentionally domesticated. Then apply the same kind of selective breeding that produced fancy rats from wild Norway rats. Just faster, more deliberately, and with detailed records.
The result, after eight years, is a colony of animals that are physically still roof rats (same body, same athleticism, same instincts), but behaviorally something else. They are calm under handling. They approach humans willingly. They raise their litters reliably in conditions that would have caused our earliest generations to fail catastrophically. They are, in the limited but meaningful sense that domestication actually means, domesticated.
But none of this would exist without a single rat… Her name was Roofy, and this whole project traces back to her.
Roofy
We started in Southern California, in the late 2010s, with a handful of orphaned baby roof rats that local people had found and brought to us. We hand-raised them. Bottle feedings every few hours through the night, manual stimulation, heating pad, all of it. Most of those rescues grew into adults that were never going to be tame, no matter how many months of human contact they had. Despite being raised by hand from infancy, they remained fearful, sometimes aggressive, and were released rather than kept.
Roofy was different… She had been found as an abandoned pinkie: the only one left in her litter, badly dehydrated, very nearly gone. She survived the syringe-feedings and the bloating, and later she seemed to aspirate some milk and developed noisy breathing, so we gave her antibiotics. She came through all of it and eventually thrived: a healthy, active roof rat. She lived only about two years before dying suddenly of a stroke, where roof rats typically live three to four years. It seems that male roof rats often outlive females, but I am not sure why.
But from the start, she was set apart from the other rescues. While they fought over the exercise wheel, she sat back and watched them with what looked like a kind of calm intelligence. When the cage door opened, she was the only one of the rescues who would come out to take a nut from a human hand. She learned her own name and would come running when called, no matter where she was in the house. She was a remarkably people-oriented animal in a species whose wild members are anything but. She was charming enough, in those days, that she once filmed a scene for a television pilot!
She learned her own name and would come running when called, no matter where she was in the house.
Physically, she was a wild-type agouti roof rat. With two exceptions: a strikingly light, near-ivory belly that set her apart from her contemporaries, and a small patch of white hairs at the tip of her tail. Those two markings turned out, much later, to be visible signatures of the temperament traits the breeding program would spend years selecting for. Lighter belly color correlates with calmer disposition in this line. The white tail-tip marking (just a few hairs on Roofy, but expressed strongly in her son and grandchildren) has become one of the line’s signature markings, and tracks with reduced fear responses.
She was, in short… a black swan. Of the rats we rescued, only Roofy made a proper pet. Roof rats raised exactly the way she was raised, at the same time as her, did not become what she became. Without her, this project would have taken years longer, if it had worked at all.
She had four litters before she was adopted out as a pet at the end of her life. The entire current colony, every animal and every line, descends from her. Her mate was Stumpy. Her son Popeye carried her tail-tip marking forward and became one of the foundation males of the line. Popeye’s daughter Flopsy was a remarkable breeder in her own right; roughly half the colony descends specifically from her. One of our current favorites, a black male named Walter, is one of Flopsy’s descendants. The lineage runs unbroken back through them to Roofy.
What we discovered after her
Roofy was tame. Roofy was a good mother. Roofy did everything we needed a founder to do… What we did not understand, in those early years, was that what we were seeing in her was an exception. The realization came when we expanded the breeding program beyond her direct lineage and ran into problems she had never shown.
The biggest of those problems was maternal failure. Mother roof rats from wild lineage, when disturbed during pregnancy or with very young pups, will sometimes kill their litters. This is a behavior documented across many rodent species (hamsters being the familiar example, where it has never been bred out), and it is not specific to R. rattus. But it is the single most common reason that prospective breeders attempting to start their own colonies from wild stock have failed. Roofy herself never did this. She was a calm, cooperative mother through all four litters. The shock came later, with other females. We had to discover the problem before we could begin selecting against it. The current line rarely shows the behavior. So in that one respect, the colony has reached a place that hamsters never have.
We also noticed, only in retrospect, that Roofy’s tameness mostly skipped a generation. Her direct offspring (the first generation of pups born in the program) were less calm than she had been. It was her grandchildren, and the generations after them, where the trait reappeared and could be selected on reliably. The likely reason is that tameness in roof rats is inherited recessively, or at least requires a combination of alleles that has to come from both parents. Roofy carried whatever combination produced her exceptional disposition; her wild-caught mate did not. So the F1 offspring inherited only one half of what they needed and reverted toward wild-type behavior. The F2 generation (Roofy’s grandchildren) included the first animals to inherit the right combination from both sides, and from there, careful selection could begin to fix the trait. For anyone hoping to start a similar project from their own founder stock, this is the part that matters: you cannot take a single tame animal and breed a tame F1. You have to breed through a generation that may look like a regression, and then select hard from there. The patience required is real…
The project eventually relocated from Southern California to Florida, where it has continued.
What we have selected for
The breeding program tracks several axes at once. Some are scored quantitatively, some by trained observation, some by counting outcomes.
Health
Health is the foundation criterion. Without it, nothing else matters. The criteria applied to every breeding decision:
- No audible respiratory issues (sneezing, wheezing, noisy breathing). Respiratory and kidney problems are the biggest health issues we see. The rats are carriers of Mycoplasma, resistance to it seems to be genetically variable, and some of the lines that show it also happen to be very tame, so it is not a simple matter of not breeding them.
- Body weight appropriate for age and sex, with a slight bias toward larger animals. Weight loss correlates strongly with developing health problems.
- Normal mobility and activity levels. Animals that move less or move strangely don’t breed.
- Females that successfully carry pregnancies to term and raise litters to weaning age without significant losses.
- Males that respond appropriately to mating cues and produce viable litters.
Animals that meet these criteria become candidates for breeding. Animals that don’t, don’t.
Diet plays a role in health too. The antioxidant science behind part of what we feed them is on our ergothioneine page, including a short interactive game about its path through the body, Inner Space: The Sulfur Sentinel.
Temperament
This is where the visible signal of domestication lives. The traits we have selected for, generation after generation:
- Submitting to being scruffed and handled by humans without struggling, vocalizing, or biting
- Remaining calmly on a human’s body without attempting to flee, by jumping, scrambling, or running away
- Approaching humans with curiosity rather than fear; exploring calmly when introduced to new people or hands
- Accepting petting (around the face, head, ears, and body) with the relaxed eye-closure, postural softening, and bruxing characteristic of contented domestic rats
- Not initiating aggression toward other rats in shared housing
- Females willing to initiate mating prompts (without aggression toward males)
- Females willing and able to nurse and raise litters in an environment with normal household noise, smells, and disturbance, including being near other rat cages
A note on handling: we do not handle young rats until they are already running around their cage and approaching weaning age. The reason is that early handling can mask genetic tameness with learned acceptance. We want to know whether an animal is innately calm before we select it for breeding, not whether it can be socialized into calmness over weeks of human contact. Animals that show natural calmness with minimal prior handling are the ones we want for the next generation.
Coat genetics
Three coat-related genes have arisen naturally in the line and are tracked across breeding decisions:
- Black variant (MC1R). An incompletely dominant melanistic mutation. Homozygous animals (two mutant copies) show solid, uniform black. Heterozygous animals (one mutant copy) appear black overall but with the underlying agouti pattern still visible, particularly as lighter bellies and lighter markings on the sides and face. They may also “rust through” toward agouti more visibly with age. Black animals may run slightly heavier on average than agoutis from the same line, and may show slightly increased aggression (that last bit is less certain). We go deeper on this mutation, including how it interacts with the blonde dilution, on our black roof rat MC1R page. Whether our black rats carry the exact same mutation found in wild black rats is still an open question we are inviting researchers to help settle, tracked on our open research questions page.
- Blonde dilute (RAB38-null). A recessive mutation that produces a soft golden or blonde coat color in homozygotes. Interacts visibly with the MC1R allele, producing darker patches in the regions that would be darkest on a melanistic animal. Blonde animals are otherwise unremarkable in health and behavior. Their eyes remain dark but may show mild photophobia. We go deeper on this mutation, and on the open, unproven idea it raises about ergothioneine transport, on our ambroxol and the blonde rat page. There is also a short playable game about it, Inner Space: The Ambroxol Voyage.
- White tail-tip marking. A true depigmentation marking with a clean boundary at the tail tip. Inherited with incomplete dominance; presence of two copies of the underlying gene produces a larger white area. The marking correlates with subtle temperament effects (somewhat calmer, friendlier), but also with slightly reduced robustness: more prone to respiratory issues, slightly lower fertility. Roofy carried the marking in nascent form (just a few white hairs), and the trait expressed more strongly in her descendants.
A fourth coat-related trait, belly color (ranging from dark through gray to very light, though not pure white outside of Roofy herself), is not a true marking but appears to be a polygenic trait inherited with incomplete dominance. Lighter belly color correlates with calmer temperament, and with general lightening of coat color on the sides and face. Heterozygous MC1R animals tend to show lighter bellies.
The published research behind this section’s genetics is collected in our bibliography.
The colours and markings described above, in the line itself (click any photo to see it full size):
Size
Mature breeding-age weights run roughly 230 to 350 grams for males, and 150 to 270 grams for females that are not pregnant, with substantial individual variation. Animals significantly below these ranges are usually showing health problems, and don’t breed. We bias toward larger animals where everything else is equal, both because they tend to be healthier and because they have an easier time as breeding adults. We put the ceiling at 330 grams for a pregnant female and 400 grams for a male, and neither is usual. These are what we see in our own animals rather than a published range.
Behavioral problems addressed
The earliest generations had the same behavioral problems that have failed every other attempt to start a roof rat colony from wild stock. The major issues, in rough order of severity:
- Maternal failure. Mothers killing their litters under stress, as described above. The dominant cause of failure in prospective colonies of this species. Largely bred out of the current line.
- Male reproductive incompatibility with tameness. In early generations, very tame males sometimes failed to respond appropriately to mating cues; selection had to balance tameness against breeding function. Peter Cottontail (Popeye’s son and Flopsy’s brother) was a vivid example: exceptionally tame, beautifully marked, completely infertile. Less of an issue in later generations.
- Social aggression in shared housing. Unlike Norway rats, R. rattus are naturally more solitary and territorial. They do not form stable social hierarchies, and same-sex housing can result in serious fighting. This still occurs occasionally, particularly with unrelated males; selection has reduced but not eliminated it.
What has not changed
The animals are physically still roof rats. They still climb walls. They still jump distances that surprise people. They still squeeze through any opening larger than their skull. Their athleticism, their reflexes, their wild repertoire of movement and exploration: those have not been bred out of them, because there is no reason to.
Some of the specific things we have watched them do are collected on our Believe It or Not page.
What has changed is the social and emotional layer that runs on top of all that capability. The animals are still capable of escape. They just rarely want to.
Where we are now
The colony continues to advance, generation by generation. Some traits we are still actively selecting on; some have stabilized to the point that they require maintenance rather than progress. The work is far from finished. (There is no version of selective breeding that can ever be called finished.) But the foundation is laid, and the line is in a state where it can be passed to other keepers who will continue refining it from here.
It started with a litter of orphaned baby roof rats and one extraordinary female who came when she was called. It is not the kind of beginning anyone would have planned. It is the beginning we had, and the line traces back to it.
As far as we know, we are the only people breeding domesticated roof rats, so the whole line runs through our colony. That is why we keep a large population on purpose. A small group would inbreed quickly, and we would lose the health and vigor we worked so hard to build. So we keep enough rats, and we keep careful track of who is related to whom, to keep the line strong.
I am proud of that. These rats wouldn’t be here if not for our efforts.
This project only matters if the line outlasts one keeper. That is why I am looking for successor breeders to carry it forward.