Free-roaming cats in the United States kill an estimated 1.4 to 3.7 billion birds and 6.9 to 20.7 billion mammals a year, according to the most-cited US estimate (Loss et al., 2013). Owned pets are only part of that total. The narrower question an owner can act on is what to put on the cat this week that measurably reduces what it brings home — and the three tested devices do not work on the same prey.
What a British cat actually brings home
The baseline comes from a Mammal Society survey of 986 cats in 618 British households, run from April to August 1997 (Woods et al., 2003). Those cats delivered 14,370 prey items. Among 696 cats with individual records, 91% brought home at least one item, and the back-transformed mean was 11.3 items in five months. Mammals were 69% of the catch, birds 24%. Scaled to roughly 9 million British cats, the authors estimated 92 million items in that window, including 27 million birds.
Two details matter more than the headline. The distribution is extremely skewed — one cat delivered 151 items, and a minority of cats does most of the killing. And only 33% of the 282 cats analysed in detail wore a bell, which is exactly why the bell question needed a controlled trial rather than another questionnaire.
Bells cut mammals by about half
Ruxton, Thomas and Wright (Journal of Zoology, 2002) ran that trial: 41 Lancashire cats that already brought prey home, randomized crossover, four weeks with a bell and four without inside an eight-week window. Mammals delivered fell from 167 to 82 — a drop of 51%. Delivery rates did not drift over the eight weeks, so within that period the cats did not simply learn to hunt silently.
The bird half of the result is the part owners rarely hear. In the Woods survey, wearing a bell predicted significantly fewer mammals brought home and had no significant effect on birds or reptiles. A bell warns a rodent; it does not beat a bird's eyesight. If birds are the concern, a bell is the wrong instrument, and the data say so without much ambiguity.
Colour is what a bird reacts to
Birds are tetrachromatic and resolve colour far better than cats do, so the Birdsbesafe collar cover turns the cat's neck into a bright band. Willson et al. (Global Ecology and Conservation, 2015) ran two 12-week seasons in New York State: cats wearing the cover brought home 0.72 birds a year against 5.56 without it, an 87% annual reduction built from 70% in fall and 95% in spring. Hall et al. (Applied Animal Behaviour Science, 2015) measured 47% to 54% fewer birds and lizards in Perth. The Australian figure is the more conservative one, and the fairer expectation outside songbird-heavy suburbs.
Device by device
| Device | Targets | Measured effect | Limits |
|---|---|---|---|
| Bell on a collar | Small mammals | 167 to 82 mammals in 8 weeks (−51%) | No bird effect found; 8 weeks only |
| Colour collar cover | Birds, some reptiles | −87% birds a year (US); −47% to −54% (Perth) | Bulkier; fit and colour matter |
| Cat bib | Birds and mammals | Large cuts in Australian field trials | Restricts the pounce; some cats refuse it |
| Locked in at dawn and dusk | Birds, reptiles | Fewer birds and mammals in surveys | Needs a routine, not a purchase |
Fitting it so the cat comes home
- Quick-release or breakaway buckle only. A band with no release point is the one design that has killed cats.
- Two fingers must still slide under the band once it is on; if they do not, it is too tight for a neck that turns and stretches.
- Keep the load small — one light bell or one cover, not a cluster of jingles plus three tags.
- Check the collar weekly and after every catch; a bent or gritty buckle will not release when it has to.
- Hold a spare fitted collar and a microchip with a current registry entry, so a slipped collar is never a lost cat.
The hours matter as much as the hardware
Songbirds feed hardest in the first three hours after dawn and again before dusk, roughly 5 to 9 am in a northern summer. Shutting the cat in for those windows removes the highest-value hunting hours at no cost. Pair it with five minutes of wand play before each meal, a window feeder about a metre from the glass, and a catio if the cat will not wear a collar at all.
What the surveys do not count
Every figure above counts what a cat brought home, which Woods and colleagues describe as a minimum, because cats eat, drop and abandon prey in the field. The US estimate is built from a different method and a different cat population, so the two should not be added together; the British number describes owned cats in one country over five months, the American one all free-roaming cats over a year. Neither tells you whether a local bird population is falling. That requires a local survey, which is why ringing schemes and garden bird counts still matter more than any single predation headline.
Where the evidence stops
Each study above ran for one or two seasons. Ruxton's cats still killed with bells on; a colour cover is a visual deterrent, not a fence; a cat that hunts at 4 am in a garden you cannot see defeats all three. Nothing here settles the population argument, and nothing replaces containment if the goal is zero. What the numbers do support is a ranking: colour for birds, a bell for small mammals, lock-in for both, and a catio when the cat fights the collar.
A seven-day trial you can run
- Day 1: fit the collar indoors for 10 minutes, with food.
- Days 2 and 3: supervised garden time only, 15 minutes each.
- Days 4 to 7: record every item brought home, by prey type and hour.
- Week 2: swap the device and repeat the same four days.
- Compare counts, not impressions. Two birds against five is a result; "seems quieter" is not.
If the cat refuses the collar
Some cats never accept one, and forcing it is a welfare problem rather than a conservation win. The fallback order is lock-in at dawn and dusk, then a catio, then a harness and lead, then a window feeder for the indoor hunt. All four beat a collar the cat spends the day trying to remove.
Frequently asked
Do bells stop cats killing birds?
Largely no. In the 1997 British survey of 986 cats, bells were associated with fewer mammals brought home but had no significant effect on birds. The controlled Lancashire trial (Ruxton et al., 2002) measured a 51% fall in mammals, from 167 to 82 items, and did not test birds. For birds, use a bright collar cover: 87% fewer birds a year in the US study, 47% to 54% in the Australian one.
Is a bell on a collar cruel to the cat?
There is no published evidence that a small bell damages hearing, and the Lancashire cats kept bringing prey home at the same rate across eight weeks, which argues against constant distress. Watch for real signs instead: freezing, ear flattening, or refusing to walk when the collar goes on. Use a quick-release buckle, keep the load to one light bell, and drop the bell entirely if the cat stays stressed.
My cat loses every collar — what else works?
Losing collars is normal with breakaway buckles, which are designed to release under load; buy spares and keep the microchip registry current. If the cat sheds collars weekly, stop fighting it: shut the cat in from roughly 5 to 9 am and again before dusk, add a catio or a harness for outdoor time, and put a window feeder about a metre from the glass so the indoor hunt has a target.