Two Rats in Seventy-Nine Days: What Cats Actually Do to an Urban Rat Colony

1. Introduction: the oldest answer

Ask anybody how to deal with rats and a cat will be among the first three suggestions. It has been the answer for several thousand years.

1.0b And every cat present did try

Stalking was universal across the animals observed, which makes this a finding about outcome rather than about interest.²

1.1 In a colony of well over a hundred

Described variously as around 120 to 150 animals.⁶

1.1b And the method costs nothing, which is its appeal

A cat requires no contract, no licence, no product and no return visit, which makes it the most attractive option on offer to anybody with a rat problem.

1.2 What this article argues

That the predation is negligible, that the behavioural effect is real, and that the two together produce a method which looks successful precisely because it is not working. Sections 13 to 15 are the case.

2. And nobody had counted

Which is the part that surprised us.

A co-author: until now, no one has provided good data on the number of city rats killed by cats.¹ The work is described as the first study to document interactions between feral cats and a wild rat colony.¹

2.0b And it is not for want of opportunity

Cats and rats have shared buildings continuously for millennia, in every city, in full view. The observation was available to anybody at any time and nobody made it.

2.1 A practice older than agriculture with no evidence base

Which places it alongside the light trap in our last technology article, where a method two thousand years old had never had its mechanism tested.

2.1b And the absence held across an enormous literature on cats

Their effect on birds has been estimated to the nearest hundred million, while the rat question had no number at all.⁵

2.2 Which is the shape of several findings in this journal

A practice everybody accepts, a mechanism everybody assumes, and a measurement that turns out never to have been taken.

3. How the study happened

The researchers were studying how scents influence rat behaviour, microchipping and filming the animals, when cats began showing up. We started asking the question of how can we get cats out of the site because they might interfere with rats, the lead author recalled.⁷

3.1 The cats arrived two days after Christmas

Into a building that had contained none at the start of the work.⁶

3.2 And the facility itself is worth picturing

A waste recycling plant where the dust falls so heavily that one researcher described it as almost looking like snow indoors.⁴

4. Which is a methodological accident

And the researcher is candid about it.

His reaction on the cats arriving was that this is just another headache, before deciding to lean in.⁴ Another account has him thinking: let us get rid of the cats so we can do our rat research.⁸

4.0b Which is worth pausing on

The researchers were lure-trapping rats with pheromone, and the cats came for the concentration of rats that produced. The study created the conditions it then observed.⁴

4.1 The cats were probably drawn by the study itself

The researchers suspect the pheromones and trapped rats attracted them.⁴

5. What that is worth

Our assessment, and it cuts both ways.

Nobody set out to test whether cats control rats, which means nobody designed the observation to produce a particular answer. That is a real strength and it is rare.

5.0b And nobody could have run it deliberately

Releasing cats into a rat colony to see what happens is not a study an ethics committee would approve, which is part of why the question stayed open for so long.

That is our suggestion rather than a reason any source gives.

5.0c And the researcher had no stake in the answer

He was studying something else and regarded the cats as an obstacle to it, which is about as free of motivated reasoning as a study gets.⁸

5.1 And a real weakness

An accidental study is not a controlled one. There is no comparison site without cats, no manipulation of cat density, and no measurement of the rat population before and after in a way that would isolate a predation effect.

5.1b And the cats were not controlled either

Five animals arriving on their own, with up to three present at any time, is a predator density nobody set and nobody can report as a treatment level.⁷

5.2 Which the design partly compensates for

Because the rats were already individually tagged for a different purpose, so the animals were known rather than estimated.²

6. The setup

Thirty-seven rats were captured, anaesthetised, tagged with radio frequency identification under the skin and released back into their colony, after which five feral cats took up residence at the plant.²

6.1 The tagging was done for another purpose entirely

Which is why individual animals could be followed at all, and is the single feature that makes the accidental study usable.⁶

6.2 And the building had no cats at the start

So the period before their arrival is a baseline nobody planned but which exists.⁶

7. What the cameras recorded

Motion-sensitive video cameras recorded cat behaviours in the presence of rats, including walking, running, stalking, chasing and feeding, producing 306 videos over almost three months. All the cats stalked rats, but over the entire period only two cats even dared to chase and attack a rat.²

7.0b The behaviours logged were ordinary ones

Walking, running, stalking, chasing and feeding, which is a coding scheme covering everything a cat might do rather than one built to find kills.²

7.1 Stalking is the behaviour that sustains the belief

A cat watching a rat looks exactly like a cat about to kill a rat, and every cat did the first thing while almost none did the second.²

8. The kill count

Two, in seventy-nine days.¹

Over that period the cats ambushed three of the facility's rats, killing two.⁵

8.1 Against a denominator that makes it starker

Over a hundred rats in the building, thirty-seven of them individually tagged, and two deaths.²⁶

8.2 And the cats were not short of opportunity

One account describes them putting a nose into the hole the rats used as their entrance.³

9. Which is reported three different ways

Across the coverage we read.

Two rat kills in 79 days in one account; three confirmed kills in another; killing or nearly killing only three rats in a third; and ambushed three, killed two in a fourth.¹⁴³⁵

9.1 The variation is in the verb as much as the number

Ambushed, attacked, confirmed kills, and killed or nearly killed describe four different events, and the figure attached shifts between two and three as the verb changes.⁵⁴³

10. And that matters for this journal

Less than usual, and for a reason worth stating.

Every version supports the same conclusion. Two or three kills in seventy-nine days against a colony of well over a hundred is negligible on any reading, so the drift here changes nothing substantive.

10.0b And this is the rare case where we can watch it happen

Eight accounts of one study published within weeks of each other, with the number already in three forms, which is the first step of a process our other articles have only been able to observe decades later.

10.1 But it is the same mechanism we have documented elsewhere

A precise figure loosening as it passes between reports, with ambushed becoming killed and killed becoming nearly killed.

10.1b And the direction of the drift is toward the dramatic

Nearly killed counted as killed makes the cats look more effective, not less, which is the opposite of what a reader might expect from coverage of a finding that cats do not work.

10.2 And we have not read the paper

So we cannot say which version is the paper's own, which §29 records.

11. The behavioural result

The rats did not move away. They mostly stayed out of sight and if they wandered into the open, they were a little more cautious.⁶

11.0b And the effect required no kills to produce it

Two deaths across the whole period, and a colony that changed how it lived.

11.1 Which is the opposite of what a control method needs

An animal that leaves is gone. An animal that hides is still there and is now harder to monitor.

That contrast is ours.

12. Why the rats did not leave

The lead researcher's explanation.

Like any prey, rats overestimate the risks of predation. In the presence of cats, they adjust their behavior to make themselves less apparent and spend more time in burrows.¹

12.0b And it explains why the effect is real without the predation being

The rats are not responding to a measured hazard. They are responding to the presence of a predator, which is the correct response whether or not that predator is any good.

12.0c And it is a costly response for the animal

Time in a burrow is time not feeding or breeding, so the behavioural change is a real imposition on the rats even though it does not reduce their number.

That reading is ours.

12.1 Overestimate is a precise claim

The rats behaved as though the danger were greater than two kills in seventy-nine days, which is rational for an animal that only needs to be wrong once.

That reading is ours.

13. What an observer would see

Following directly from §11.

Fewer rats in the open, fewer rats crossing a yard, fewer rats visible at all, with the same number of rats present.

13.1 And nothing about that observation is mistaken

Fewer rats really are visible. The error is entirely in what the observation is taken to mean.

14. Which is the article's central point

The method produces the appearance of working by a mechanism that is not working.

And the coverage says so: people see fewer rats and assume it's because of the cats.⁵

14.0b And the feedback loop is immediate

The behavioural change happens within days of the cats arriving, while any real population effect would take months to be visible, so the false signal arrives first and the true one never arrives.

14.1 So the belief is continuously reinforced

Somebody who puts cats on a site does observe an improvement. The observation is real and the inference from it is wrong.

14.1b And it survives the absence of any kill being seen

Nobody expects to witness a cat catching a rat, so the lack of evidence for predation is not read as evidence against it.

14.2 Which makes this belief unusually hard to dislodge

Most folk remedies this journal has examined fail visibly. This one succeeds visibly.

Both are ours.

15. And it is our sighting metric problem again

In a different setting.

Our article on waste containerization found municipal rodent programmes measured by resident sightings, and noted that an intervention acting on the circumstances that produce a sighting will move the metric whether or not it moves the population.

15.0b And the two cases share a structure worth naming

In both, the intervention acts on the thing being measured rather than on the thing being managed, and in both the measurement was chosen because it is the only one readily available.

15.1 Here the same confusion happens inside one person's head

Without any metric, reporting system or municipality involved. The observer and the measuring instrument are the same person.

16. Where the belief comes from

The researcher offers a specific origin.

A lot of people confuse rats with mice. So they've seen their cat bring back a mouse, and assume their cat is a rat killer. It's not.²

16.0b And the generalisation is one this trade encourages

Rodent control is sold as a single category covering both animals, and our own articles have had to keep separating them because the biology, the placement and the behaviour differ at almost every point.

16.1 Which is a generalisation across a large difference

Our own house mouse article described an animal weighing a fraction of a rat, with different behaviour, different range and different defences.

16.2 And it is reinforced by every cat that ever caught anything

A bird, a vole, a mouse on a doorstep all count as evidence for a general hunting competence, and the specific question of whether the animal will take a rat is never separately tested.

That observation is ours.

17. The size difference

Quantified in one account.

Adult rats at the facility averaged about 337 grams, described as more than ten times the weight of the prey a cat typically takes.³

17.0b Which is a different animal from the one the reputation was built on

A mouse of twenty or thirty grams presents no risk to a cat. A rat of three hundred and thirty does.³

17.1 And the prey fights back

One account describes the animal as relatively large, defensive and potentially dangerous.¹

17.1b And one account describes the rat holding its ground

Meeting a cat's gaze and continuing to move about apparently untroubled, which is not the behaviour of an animal that expects to be caught.⁸

17.2 Which explains the stalking without the attacking

Every cat watched. Two acted. A predator assessing a risky target and declining it is behaving sensibly.

Section 17.2 is ours.

18. The corroboration

Which matters, because the study is one site.

The results fit with earlier findings from alleys in another American city, where cats were similarly cagey around the bigger rodents, and a disease ecologist who studied rats there states: cats do occasionally catch rats, just not very often and not to the extent they reduce the size of the rat populations.³

18.0b And the word used is the same

Cagey, in the second account, describing cats keeping their distance rather than failing to catch what they attempted.³

18.1 A different city, a different researcher, the same finding

Which is the strongest thing in this article.³

18.2 And it predates the newer work by years

Which means the finding was available in the literature before anybody filmed it, and did not displace the belief then either.³

19. What the researcher says instead

The alternative, stated bluntly.

Predation, by cats or any other animal, will not reduce the rat population. Only cleaning up the garbage, the rats' food, will do that. With the addition that rats have a high reproductive rate, and if there is rubbish around, there will be rats no matter what is done.²

19.0b Which generalises past the specific method

The claim is not that cats are bad predators but that predation is the wrong lever, which would hold equally for a better predator.²

19.1 And the key to managing urban rodent populations is waste management

Not feral felines, in the same researcher's phrasing.⁵

19.2 And it removes the method from the argument entirely

Fewer rats follows from less food, on this account, without reference to what is or is not hunting them.⁵

20. Which our own articles have argued

From three different directions.

Our article on containerization found refuse management ranked first among municipal rodent strategies. Our article on the federal rat control programme found poisoning listed fifth of five measures. And our article on rodent fertility control found the same limit: a population fed by available food will replace whatever is removed.

20.0b And it is why the result should not have been surprising

A population limited by food and capable of rapid reproduction absorbs a mortality source and replaces it. Two deaths in seventy-nine days would be irrelevant even if it were two hundred.

20.0c And it is the finding this journal keeps arriving at

Across containerization, fertility control, federal programme history and now predation, the conclusion is the same and the method being assessed changes each time.

20.1 So this is the same finding with a different removal method

Predation rather than poisoning, and the conclusion is unchanged because the constraint is the food supply rather than the mortality source.

21. The method that makes cats hunt

Which deserves to be more widely known.

Hungrier cats get more committed to rat chasing, and some people who release cats to control a rat problem withhold food to encourage hunting, which the lead author calls a horrible thing to do, since it forces an animal to confront less-favoured and dangerous prey.³

21.0b And it is the only lever the practice offers

Nothing else about a released cat can be adjusted. You cannot train it, aim it or dose it, so hunger is the entire control input.

21.1 Which tells you the predation is not automatic

If a well-fed cat killed rats, nobody would need to starve one. The practice is an admission about the default behaviour.

That inference is ours.

22. And the other costs

A 2013 study is cited for the estimate that cats cause 2.4 billion bird deaths a year in one country, and a 2017 study described the species as among the most ubiquitous and environmentally damaging invasive predators on Earth, implicated in the extinction of at least 63 vertebrate species.⁵⁴ Faeces can spread a parasite with severe consequences when transmitted in pregnancy.⁵

22.0a And the costs fall on parties who did not choose them

Birds and other wildlife are not clients of anybody's rodent programme, which is the externality structure our economics articles describe in a different setting.¹

22.0b The asymmetry has an explanation

Bird predation is easy to study, since the evidence arrives on doorsteps and in stomach contents. Rat predation happens at night inside buildings nobody is watching.

That is our suggestion.

22.1 Which are documented far better than the predation was

The effect of cats on native wildlife has been quantified repeatedly while the number of rats they kill had never been counted, which is an odd shape for a literature to have.¹

23. Where cats are actually deployed

Named in one account.

Breweries, warehouses and waste facilities sometimes turn to cats for rat control.⁸

23.0b And those are sites with the most food and the most rats

Which §29 notes is the worst place for predation to matter, so the practice is concentrated where the method has least chance of working.

23.1 Which are exactly the sites our food premises articles concern

And where an animal defecating in a production area is a separate problem from whether it catches anything.

That observation is ours.

23.2 And the practice is described as commercial rather than casual

These are businesses turning to a method deliberately, not households acquiring a cat and noticing a benefit.⁸

24. What we would tell a client

Three things.

The predation is negligible. Section 8.¹

You will see fewer rats and there will not be fewer rats. Sections 11 to 14.⁶

And the food is the variable. Section 19.²

24.0b And one thing we would put in writing

That an apparent improvement after introducing cats should not be read as a population change, because §13 says the two are indistinguishable by eye.

24.1 And one question we would ask

Where the food is, since §19 makes that the variable and nothing else in the conversation matters until it is answered.²

25. And what we would not claim

Two things.

That cats never catch rats, which the sources explicitly allow. And that the behavioural effect is worthless, since a rat spending more time in a burrow is a rat doing less damage in the open, for as long as the cats remain.

25.0b And it would end when the cats do

A behavioural effect requires a present predator, so unlike an exclusion repair it delivers nothing durable.

25.1 Whether that is worth anything depends on the site

A warehouse cares about product contact. A householder cares about seeing one.

Section 25 is ours and no source makes the argument.

25.2 And we would not extend it to mice

Nothing in this study concerns them, and the researcher's own explanation in §16 assumes cats do catch them.²

26. A pest control company on the paper

Which we note as a matter of interest.

One of the co-authors is identified as being from a named exterminating company.¹

26.0b Though the quoted contribution is a careful one

The co-author's statement is about the absence of data and about a separate well-evidenced effect, rather than a claim that cats do not work, which would have been the commercially convenient thing to say.¹

26.1 A finding against a free alternative, co-authored by a paid one

Which is a conflict worth naming, and which does not by itself weaken a result corroborated from another city by an unrelated researcher.³

26.1b And the paper was published in an academic venue

As part of a themed issue on rodents, which is a different setting from a trade publication and carries different scrutiny.¹

26.2 And it is the same conflict we have

Section 27.

27. Our own position

The disclosure, and it is the most direct in this journal.

This article tells people that the free method does not work, and we sell the paid one. Every sentence in it should be read with that in mind.

27.0b And we would be glad to be wrong

A free method that worked would be better for every client we have than a paid one that does, and nothing in our interest changes what the cameras recorded.

27.1 What we can offer against it

That the evidence is not ours, that it was collected by accident by people studying something else, that it is corroborated independently, and that §19's conclusion points to waste management rather than to a service contract.

27.1b And we would rather publish this than not

A client who tries cats and then calls us has lost time. A client told at the outset what the evidence says can make a decision with it.

27.2 And §25 is where we argue against our own interest

By allowing that the behavioural effect may have some value.

28. The Manitoba position

28.1 The relevant species is the one studied

Norway rats are the rat of this city, as our article on the species replacement set out.

28.1b And the outdoor season here is short

A free-roaming cat in this climate spends a substantial part of the year in conditions that constrain hunting independently of anything about rats.

28.2 What we could not find

Any Canadian study of cat predation on urban rats, any local figure for free-roaming cat numbers, and any assessment of the bird predation estimates in a prairie context.

28.2b Which this article does not answer

Nothing here tests whether cats reduce mouse populations, and the study concerned rats throughout.¹

28.3 And the barn cat is a separate question

Rural buildings with mice rather than rats are the setting where the folk belief is least challenged by this evidence, since the confusion in §16 runs the other way.²

29. Limitations and open questions

We did not read the paper. Every figure, quotation and finding here comes from eight secondary accounts of one study, including a press release, and we could not check any of them against the publication.¹

That is the most important limitation and §9 shows why, since the headline number appears in at least three forms across those accounts and we cannot say which is the original.³⁴⁵

The sources disagree on the location. One places the facility in a different borough from the rest, which is a small error and an indication of how carefully the coverage was assembled.⁷

It is one site and one colony. A waste recycling facility is a food-rich environment where rats are unusually well fed and unusually numerous, which may be the worst possible place for predation to matter.²

And the cats were not selected. Five feral animals that wandered in are not the same as cats chosen, conditioned or kept for the purpose, which is what the practice in §23 actually involves.⁴

The behavioural finding is not quantified in what we read. More time in burrows and less apparent are qualitative descriptions, with no figures for movement, activity or visibility given anywhere in our sources.¹⁶

Sections 5, 10.1, 12.1, 13, 14, 15, 17.2, 20, 23.1 and 25 are our reasoning. The assessment of the accidental design, the argument that behavioural suppression produces a false appearance of control, the connection to our own work on sighting metrics and the concession about partial value are ours rather than sourced positions.

30. Conclusion

Cats have been the standard answer to rats for as long as people have kept grain, and a co-author of the 2018 study reports that until it was done, nobody had provided good data on the number of city rats killed by cats.¹ The data exist because researchers studying rat pheromones at a waste recycling facility, who had already microchipped thirty-seven animals from a colony of over a hundred, found five feral cats moving in and decided to film them rather than remove them.²⁷ Over 79 days and more than three hundred videos, every cat stalked rats and two chased and attacked. Three rats were ambushed and two were killed.²⁵¹ An earlier set of observations in another city found the same reluctance, with a researcher there concluding that cats occasionally catch rats but not often and not enough to reduce the population.³

The interesting result is not the kill count. It is what the rats did. They did not leave. They spent more time in burrows, stayed out of sight, and were more cautious when they did come into the open, which the lead researcher attributes to prey overestimating predation risk.⁶¹ The consequence is that a site with cats on it has fewer visible rats and the same number of actual ones. Somebody who introduces cats observes a genuine improvement and draws the wrong conclusion from it, which the coverage notes directly: people see fewer rats and assume the cats are responsible. This journal found the identical structure in municipal programmes measured by resident sightings. Here it operates inside one person's observation, with no metric involved at all, which is why this particular belief has survived several thousand years of testing by everybody who ever owned a cat.

The rest is context and it all points the same way. A rat at this site averaged around 337 grams, more than ten times a cat's usual prey, and is described as large, defensive and potentially dangerous, so a predator that watches and declines is behaving sensibly.³¹ The belief has a traceable origin in cats killing mice, which they do.² Making cats hunt rats reliably, where people try, sometimes means withholding their food, which the lead author calls a horrible thing to do.³ And the alternative he names is not a product but refuse: only cleaning up the garbage will reduce the population, because rats breed fast and food is the constraint.² We should say plainly that we sell the paid alternative to the free method this article criticises, that we have read none of the underlying paper, and that the conclusion we are endorsing recommends neither cats nor us.

References

1. Research news summary of the 2018 study, issued as a press release and carried by a science news aggregator. Institutional press material, which we flag as the publisher's own framing. Source for the description of the work as the first study to document interactions between feral cats and a wild rat colony; for the statement that researchers monitored the behaviour and movement of microchipped rats in the presence of cats living in the same area, showing the rats actively avoided the cats and recording only two rat kills in 79 days; for the quoted explanation that like any prey, rats overestimate the risks of predation and in the presence of cats adjust their behaviour to make themselves less apparent and spend more time in burrows; for the description of the rats as a relatively large, defensive and potentially dangerous prey commonly targeted by released cats; for the co-author's statement, attributed to a named exterminating company, that until now no one had provided good data on the number of city rats killed by cats while the data on the effect of cats on native wildlife have been very clear; for the note that the findings add to growing evidence that any benefit of using cats to control city rats is outweighed by the threat they pose to birds and other urban wildlife; and for the publication context as part of a special rodent issue of an ecology and evolution journal. https://www.sciencedaily.com/releases/2018/09/180927083342.htm
2. Newspaper report on the study, carried by a broadcaster's website. News material. Source for the account that the researchers began with another aim, tracking rodent activity at the plant and reactions to various scents; for the method of capturing and anaesthetising 37 rats, inserting radio frequency identification tags under the skin of each and releasing them into their colony; for the arrival of five feral cats taking up residence at the plant; for the use of motion-sensitive video cameras to record cat behaviours in the presence of rats including walking, running, stalking, chasing and feeding; for the figure of 306 videos over almost three months; for the finding that all the cats stalked rats but that over the entire period only two cats even dared to chase and attack a rat; for the researchers' conclusion that cats are ineffective in controlling urban rat populations despite their reputation; for the quoted explanation that many people confuse rats with mice, having seen a cat bring back a mouse and assumed it a rat killer; and for the statement that predation by cats or any other animal will not reduce the rat population, that only cleaning up the garbage will, and that rats have a high reproductive rate so rubbish means rats regardless. https://www.wral.com/cats-v-rats-in-new-york-the-rats-win/17917470/
3. Science news magazine report on the study. News material. Source for the figure that adult rats at the waste centre averaged about 337 grams, described as more than ten times the weight of a cat's typical prey; for the statement that over 79 days cameras showed cats killing or nearly killing only three rats and that such performance would not dent a rat population; for the account that researchers noticed five feral cats creeping around the microchipped rats and initially bristled at the intrusion; for the observation that the cats paid attention to the rats, sometimes putting a nose into the hole the rats used as their entrance; for the statement that hungrier cats get more committed to rat chasing and that some people who release cats to control rats withhold food to encourage hunting, with the lead author calling it a horrible thing to do to force an animal to confront less-favoured and dangerous prey; and for the corroboration from earlier findings in alleys in another American city, with a named disease ecologist stating that cats do occasionally catch rats, just not very often and not to the extent that they reduce rat population size. https://www.sciencenews.org/article/feral-cats-pathetic-controlling-new-york-city-rats
4. Feature article on the study published by an online magazine. News material. Source for the identification of the site as a waste management facility in a named Brooklyn neighbourhood and the study period as December 2017 to May 2018; for the statement that the surrounding feral cats landed just three confirmed kills in that time; for the description of the facility's dust falling so heavily that it almost looks like snowing indoors; for the researchers' suspicion that the pheromones and trapped rats attracted the cats and the lead author's initial reaction that this was just another headache before deciding to lean in; and for the citation of a 2017 study describing the species as one of the most ubiquitous and environmentally damaging invasive predators on Earth owing to its role in the extinction of at least 63 vertebrates worldwide. https://www.atlasobscura.com/articles/rats-and-cats
5. Magazine news item on the study published by a museum-affiliated publication. News material. Source for the statement that over the 79-day testing period local cats ambushed just three of the facility's roughly 150 rats, killing only two; for the researcher's position that the key to managing urban rodent populations is waste management rather than feral felines, with less street garbage meaning rats would moderate themselves; for the observation that people see fewer rats and attribute it to the cats; for the note that feline faeces spread a parasitic disease which can cause severe brain damage or death when transmitted from a pregnant mother to a fetus; and for the citation of a 2013 study suggesting cats are responsible for 2.4 billion bird deaths per year in one country. https://www.smithsonianmag.com/smart-news/cats-are-surprisingly-ineffective-keeping-urban-rat-populations-check-180970428/
6. Broadcaster science report on the study. News material. Source for the colony size given as around 120 to 150 rats in a semi-enclosed building at a waste recycling facility; for the statement that there were no cats in the building at the start of the study; for the description of the method of luring rats into traps with rat pheromone, anaesthetising them, taking measurements, implanting a microchip and releasing them; for the date of the cats' arrival two days after Christmas; for the finding that contrary to expectations the cats killed just two rats in nearly three months; and for the statement that the rats did not move away but mostly stayed out of sight and were a little more cautious when they wandered into the open. https://www.abc.net.au/news/science/2018-09-28/cats-dont-kill-new-york-city-rats-urban-ecology/10307924
7. Report on the study published by a professional wildlife society. Professional association material, which we flag because it names a different borough for the study site than the other sources. Source for the account that researchers set out to study how scents influence rat behaviour and were microchipping and filming the animals when cats began showing up; for the lead author's quoted recollection that they started asking how to get the cats out of the site because they might interfere with the rats, before realising they could take advantage of the uninvited guests to quantify what happens between the two species; for the statement that up to three cats were present; and for the figure of over 300 videos examined over nearly 80 days from motion capture cameras placed at the original scent study sites. https://wildlife.org/rats-feral-cats-fail-at-urban-rodent-control/
8. Magazine article on the study published by a popular science title. News material. Source for the statement that breweries, warehouses and waste facilities sometimes turn to cats for rat control; for the description of rats meeting a cat's gaze and continuing to move about apparently untroubled; and for the lead author's quoted reaction as a behavioural ecologist wanting to get rid of the cats so that the rat research could proceed. https://www.scientificamerican.com/article/cats-may-have-duped-us-about-being-great-rat-catchers/