Killer Whales Are Manufacturing Grooming Tools, Mostly With Family

What the Drone Caught

Start with the geography, because it explains why anyone was flying a drone over these particular whales in the first place. The southern resident killer whale population is three matrilineal pods, J, K and L, that spend much of the year in the inland waters between Washington State and British Columbia: the Salish Sea. Unlike the transient killer whales that share the same coastline and hunt seals, southern residents eat almost nothing but fish, overwhelmingly Chinook salmon, and they have been counted individually by name and photograph since the 1970s. NOAA Fisheries' most recent census, from July 2025, puts the whole population at 74 animals [14]. It is one of the most closely watched groups of wild mammals on Earth, which is part of why a routine-looking behavior went unremarked for so long: everyone was watching the surface, and this happens mostly underwater, at the waterline, in a pose that from a boat looks like two whales lying against each other and not doing much of anything.

From an aircraft looking straight down, it looks like something else. Between 10 April and 27 July 2024, Weiss and colleagues flew a small remotely piloted aircraft over the three pods and, across that period, filmed 30 separate bouts of two whales manipulating a length of kelp between their bodies [1]. The species involved is bull kelp, Nereocystis luetkeana, the kind that grows as a single long stipe topped by a gas-filled float and a crown of blades, anchored to the rocky bottom by a holdfast that looks a little like a fist of roots. Solitary whales draping loose kelp over themselves, sometimes called "kelping," has been reported before, in several whale species, and a 2023 review treats it as a fairly ordinary and widespread piece of cetacean behavior [4]. What Weiss and colleagues were filming was not that. It was two animals, a shared object, and, on close inspection, a manufacturing step that solitary kelping does not require.

The Word Doing All the Work

Every write-up of this paper, including the press coverage circulating this year, leads with "tool use." That undersells it. Killer whales already have a tool-use record: pods that beach themselves briefly to snatch seals off shingle, a hunting technique passed down within families, are routinely described that way. What makes this paper different is the word manufacture, and the tool-use literature treats that as a genuinely higher bar. The standard reference the paper itself cites, Shumaker, Walkup and Beck's synthesis of the field, splits the topic into tool use — picking up something already lying around and applying it to a job — and tool manufacture, which additionally requires that the user structurally alter the object first, so that it does the job better than it would have unmodified [2]. A sea otter smashing a clam on a rock it found is tool use. A crow bending a wire into a hook is manufacture. Wild cetaceans, despite decades of observation, had no confirmed case of the second kind at all. A 2013 review of tool use across aquatic animals is the source the paper itself cites for exactly that gap: cetacean tool use is largely restricted to foraging contexts, with no prior report of a cetacean fashioning a tool by modifying an object [1, 3]. Not because these animals lack ingenuity, but because "modify an object" is a phrase built for hands, and nothing in the water has a hand.

Rows one through three are the paper's direct observations, and we will walk through the footage behind each of them in the next two sections. Row four is where the authors are more careful than the headlines that followed them. Their own sentence reads: allokelping "would be the first case of tool manufacture ... in marine mammals," but that clause is explicitly conditioned on an earlier one — "if a functional benefit of allokelping can be confirmed" [1]. The manufacturing step is not in doubt. Whether the finished tool does anything useful, as opposed to being a socially learned habit with no measurable payoff, is the part the authors themselves flag as still open. We think that hedge is worth repeating exactly because so few summaries of this paper repeated it.

How You Make One

The mechanics, as described in the paper's methods: a whale approaches an intact, still-attached kelp stalk, grips it with its teeth near the narrow end of the stipe close to the holdfast, and then uses a combination of the kelp's own drag in the water, its own body motion, and its teeth to shear off a short length [1]. No limbs are involved, because odontocetes do not have any that could hold or bend an object; the entire manufacturing step is mouth, water resistance and momentum.

The paper is careful about how many times this specific moment was actually captured on camera, as opposed to inferred. Whales were seen acquiring a stipe segment for allokelping on 8 occasions total. Of those 8, 7 were unambiguous detach events off an intact, rooted stalk [1]. That is a small number set against 30 total grooming bouts, and it says something honest about the limits of aerial video: manufacture happens fast, near the surface or just below it, and most of the 30 bouts simply began with the tool already in hand, out of camera view of the moment it was made. What the 8 clear cases establish is the method. What they do not establish is that every one of the 30 bouts used a freshly bitten piece rather than, say, a kelp segment recovered from an earlier bout. The authors do not claim otherwise.

What Two Whales Do With It

Once a length of stipe exists, the behavior itself is almost startlingly simple to describe and, per the video, visually unmistakable from above: one whale positions the kelp between its own body and a partner's, and the pair rolls it back and forth along their skin, in extended contact, for as long as several minutes at a time [1]. Weiss has described the material itself in terms anyone who has handled bull kelp will recognise — firm but flexible, with a slick outer surface, closer to a length of garden hose than to anything leaf-like.

The paper reports the timing precisely enough to plot. Across the documented bouts, the whole interaction — from the kelp being positioned to the pair separating — ran a median of 1 minute 19 seconds, with an interquartile range spanning roughly 81 seconds and a longest observed bout of 12 minutes 4 seconds. Sustained skin-to-skin contact within a bout, a narrower measure than the bout as a whole, ran a median of 32 seconds, interquartile range about 38 seconds, longest single contact 186 seconds [1].

Ruling Out a Simpler Explanation

Anything two young mammals do with a shared floating object invites an obvious alternative reading: they are playing with it. The authors considered this directly and argue against it on two grounds. First, none of the behavioral markers of object play in cetaceans showed up alongside allokelping — no play chases, no surface-active leaping, no "keep-away" games with the kelp piece, all of which are documented elsewhere in cetacean play more broadly [1, 5]. Second, and more tellingly, the behavior looked the same every time. Its "highly stereotyped" form, repeated the same way across very different pairs, combined with the fact that it turned up across every age-sex class rather than clustering in juveniles the way object play usually does, is what the authors point to as evidence this is not simply kelp being treated as a toy [1].

Where the Cameras Were, and For How Long

The sampling effort was lopsided, and the paper says so plainly. Almost all of it fell on J pod: 9 hours of usable drone footage across 11 days, during which 19 of J pod's 25 individuals were seen allokelping, spanning all six of the pod's matrilines. K pod got 1 hour across 2 days. L pod got 21 minutes on a single day [1]. Individuals named in the paper's own figures include J41, J19, J51, J56 and J57 — a spread across at least two generations within J pod, consistent with the age range the authors report participating overall.

Doing Our Own Division

Nothing in this section is the paper's arithmetic. It is ours, built entirely from the effort and count numbers in the previous section, and it is here because the authors did not report a rate and we were curious what one would look like.

Add the three pods' footage together: 9 hours plus 1 hour plus 21 minutes is 10 hours 21 minutes, or 10.35 hours as a decimal. Divide the 30 recorded bouts by that total: 30 ÷ 10.35 ≈ 2.90 bouts per hour of drone footage, or one new bout starting, on average, roughly every 20.7 minutes that a camera was up. That is a rate of observation, not a rate of behavior — whales are not allokelping only while a drone happens to be overhead, and the true frequency could be much higher or somewhat lower. What the number does support is the paper's own informal impression, that this is not a rare curiosity glimpsed once, but something the crew could expect to catch again most times they flew.

A second small division, from the role-switching data discussed next: 28 re-initiations across 13 bouts works out to about 2.15 re-initiations per bout, among the bouts where more than one initiation occurred at all. Whatever allokelping is for, it is not typically a single one-sided gesture; on average it involves the pair re-engaging with each other more than twice before the bout ends.

Who Grooms Whom

This is the section that turns a charming video clip into a social-behavior finding. Restricting the analysis mostly to J pod, where the sampling was deep enough to support it — 27 of the 30 total bouts, involving 25 individually identified whales — the authors modelled the odds that two whales would be seen allokelping together as a function of how closely related they are and how close in age they are, in a Bayesian framework [1].

Two effects came back strong. Maternal kinship was positively associated with partnering up, with a reported coefficient of 0.91 ± 0.21 and a posterior probability that the effect is positive of greater than 0.99. Age similarity was almost as strong, at a posterior probability of 0.98. A third, weaker signal turned up around skin condition: whales showing visible moulting or dead skin were somewhat more likely to be involved, posterior probability 0.93 — language the paper itself calls "tentative." A fourth candidate effect, whether the odds of allokelping simply decline with a whale's own age, found no support at all: posterior probability of a decline, 0.53, indistinguishable from a coin flip [1].

Put in words rather than probabilities: two whales are far more likely to be seen making and sharing a kelp tool if they are mother and calf, or siblings, or otherwise closely related, and separately more likely if they are close in age — a pattern that lines up with earlier work on this same population showing that age and sex shape which killer whales interact socially, even when raw physical association does not [11]. Skin condition nudges the odds up a little further, but not conclusively. How old a whale is, on its own, predicts almost nothing.

Taking Turns

One more behavioral detail rounds out the social picture. Within a bout, whichever whale initiates contact is not fixed. Among bouts where more than one initiation was recorded, the initiating role switched between the two partners in 66% of them [1]. Combined with the reciprocal rolling motion itself, that switching is difficult to square with one whale simply using the other as a stationary scratching post; it reads more like a genuinely shared, turn-taking activity, which is also the shape most primate and human bonding rituals involving mutual touch tend to take [10].

A Guess at Why, Beyond Bonding

The functional question — row four of the manufacture checklist — is the one nobody has closed. Two hypotheses run through the discussion, and they are not mutually exclusive. The social-bonding reading leans on a large existing literature showing that mutual touch, across primates generally, tends to be a mechanism for maintaining relationships rather than a side effect of having them [10]; the kinship and age-similarity results above fit that story well, since those are exactly the relationships animals in most social species invest the most in maintaining.

The skin-care reading takes the moulting correlate seriously and looks for cetacean precedent. It has some. Bowhead whales in the eastern Canadian Arctic engage in "rock-nosing," rubbing their heads and bodies against large boulders in shallow water, and the leading explanation for that behavior is that it helps manage moulting skin [6]. Captive bottlenose dolphin mothers and calves have been filmed rubbing flippers against each other in a pattern the observing researchers interpreted as body-surface care rather than play [8]. Neither of those is tool manufacture — a rock is not modified, and a flipper is not detached from anything — but both support the idea that cetaceans do use mutual contact for skin maintenance, independent of any social payoff. Allokelping may simply be southern residents doing the same job with a tool instead of a boulder, and getting a bonding benefit as well, rather than instead.

The Company This Behavior Keeps

It is worth placing allokelping next to the small set of comparisons that make it interesting rather than merely cute. The most famous case of cetacean tool use is sponge-carrying in Shark Bay, Australia, where some bottlenose dolphins wear a marine sponge over their rostrum while foraging over sharp rubble, apparently to protect their beak; the behavior is passed down mostly from mother to daughter [12, 13]. Notably, the paper that first described it was titled as a question — "a foraging specialization involving tool use?" — because even its own discoverers were cautious about the label [12]. Sponging clears the tool-use bar easily. It does not clear manufacture, because the sponge is selected and worn, not structurally altered.

The closer parallel, and one the paper itself reaches for, sits outside cetaceans entirely: chimpanzees have been filmed passing a stripped plant stem to a grooming partner for dental cleaning, a behavior that is both a manufactured tool and, like allokelping, directed at another individual rather than used solely on oneself [9]. That combination — manufacture plus a social, other-directed target — is genuinely rare across the whole animal kingdom, and it is the specific combination that makes the authors' second claim, that this may be the first tool use in a wild animal benefiting two individuals at once, worth taking seriously even though they phrase it, again, as "may be" rather than as settled fact [1].

What Could Take It Away

The paper closes on a note that has nothing to do with cognition and everything to do with water temperature. Bull kelp is sensitive to warming, and its decline in parts of the southern residents' range is already documented independent of this study. If that decline continues, the authors write, it "may pose a threat to the persistence of this behavior" [1]. They draw the comparison to a better-known case: the northern resident killer whale population, a separate group living further north, has its own well-documented cultural practice of rubbing against smooth pebble beaches, a behavior specific enough to particular sites that Canadian conservation policy already protects those beaches on its account [7]. The paper's suggestion is that allokelping, if it turns out to matter as much to southern residents as beach-rubbing does to northern residents, may eventually warrant the same kind of protection extended to a habitat feature rather than to the animals directly.

That is a striking place for a two-page report to end up: a description of thirty video clips, filed as a note rather than a full paper, concluding with a recommendation about kelp forest management for an endangered population that already has entire agencies devoted to counting it one animal at a time. We do not think the paper oversells that connection. We do think it is the sentence most coverage of this story skipped past on the way to the words "tool use," and it is the one we would ask a reader to sit with longest.

References

1. 1.

   Weiss, M. N., John, R. E., Caro-Ruiz, A. M., Ellifrit, D. K., Grimes, C. A., Redmond, T. A., Vargas, A. A. & Croft, D. P. (2025). Manufacture and use of allogrooming tools by wild killer whales. Current Biology 35(12), R599–R600. doi:10.1016/j.cub.2025.04.021

2. 2.

   Shumaker, R. W., Walkup, K. R. & Beck, B. B. (2011). Animal Tool Behavior: The Use and Manufacture of Tools by Animals. Johns Hopkins University Press.

3. 3.

   Mann, J. & Patterson, E. M. (2013). Tool use by aquatic animals. Philosophical Transactions of the Royal Society B 368, 20120424. doi:10.1098/rstb.2012.0424

4. 4.

   Meynecke, J.-O. & Kela, H. (2023). What's at play: Humpback whale interaction with seaweed is a global phenomenon. Journal of Marine Science and Engineering 11, 1802. doi:10.3390/jmse11091802

5. 5.

   Hill, H. M., Dietrich, S. & Cappiello, B. (2017). Learning to play: A review and theoretical investigation of the developmental mechanisms and functions of cetacean play. Learning & Behavior 45, 335–354. doi:10.3758/s13420-017-0291-0

6. 6.

   Fortune, S. M. E., Koski, W. R., Higdon, J. W., Trites, A. W., Baumgartner, M. F. & Ferguson, S. H. (2017). Evidence of molting and the function of "rock-nosing" behavior in bowhead whales in the eastern Canadian Arctic. PLOS ONE 12, e0186156. doi:10.1371/journal.pone.0186156

7. 7.

   Ford, J. K. B. (1989). Acoustic behaviour of resident killer whales (Orcinus orca) off Vancouver Island, British Columbia. Canadian Journal of Zoology 67, 727–745. doi:10.1139/z89-105

8. 8.

   Sakai, M., Katsumata, H. & Kohshima, S. (2022). Observations of flipper rubbing in mother–calf pairs of captive bottlenose dolphins (Tursiops truncatus) suggest a body-surface care function. Journal of Ethology 40, 257–264. doi:10.1007/s10164-022-00754-1

9. 9.

   McGrew, W. C. & Tutin, C. E. G. (1973). Chimpanzee tool use in dental grooming. Nature 241, 477–478. doi:10.1038/241477a0

10. 10.

    Dunbar, R. I. M. (2010). The social role of touch in humans and primates: Behavioural function and neurobiological mechanisms. Neuroscience & Biobehavioral Reviews 34, 260–268. doi:10.1016/j.neubiorev.2008.07.001

11. 11.

    Weiss, M. N., Franks, D. W., Giles, D. A., Youngstrom, S., Wasser, S. K., Balcomb, K. C., Ellifrit, D. K., Domenici, P., Cant, M. A., Ellis, S., Nielsen, M. L. K., Grimes, C. & Croft, D. P. (2021). Age and sex influence social interactions, but not associations, within a killer whale pod. Proceedings of the Royal Society B 288, 20210617. doi:10.1098/rspb.2021.0617

12. 12.

    Smolker, R., Richards, A., Connor, R., Mann, J. & Berggren, P. (1997). Sponge carrying by dolphins (Delphinidae, Tursiops sp.): A foraging specialization involving tool use? Ethology 103, 454–465. doi:10.1111/j.1439-0310.1997.tb00160.x

13. 13.

    Krützen, M., Mann, J., Heithaus, M. R., Connor, R. C., Bejder, L. & Sherwin, W. B. (2005). Cultural transmission of tool use in bottlenose dolphins. Proceedings of the National Academy of Sciences 102, 8939–8943. doi:10.1073/pnas.0500232102

14. 14.

    National Marine Fisheries Service, NOAA Fisheries (2025). Southern Resident killer whale species page. Population census, July 2025: 74 individuals.