TED2004
Sheila Patek: The shrimp with a kick!
Filmed:
Readability: 4
1,659,037 views
Biologist Sheila Patek talks about her work measuring the feeding strike of the mantis shrimp, one of the fastest movements in the animal world, using video cameras recording at 20,000 frames per second.
Sheila Patek - Biologist, biomechanics researcher
Biologist Sheila Patek is addicted to speed -- animal speed. She's measured the fastest animal movements in the world, made by snail-smashing mantis shrimp and the snapping mandibles of trap-jaw ants. Full bio
Biologist Sheila Patek is addicted to speed -- animal speed. She's measured the fastest animal movements in the world, made by snail-smashing mantis shrimp and the snapping mandibles of trap-jaw ants. Full bio
Double-click the English transcript below to play the video.
00:25
If you'd like to learn how to play the lobster, we have some here.
0
0
3000
00:28
And that's not a joke, we really do.
1
3000
2000
00:30
So come up afterwards and I'll show you how to play a lobster.
2
5000
3000
00:33
So, actually, I started working on what's called the mantis shrimp
3
8000
4000
00:37
a few years ago because they make sound.
4
12000
3000
00:40
This is a recording I made of a mantis shrimp
5
15000
2000
00:42
that's found off the coast of California.
6
17000
2000
00:55
And while that's an absolutely fascinating sound,
7
30000
3000
00:58
it actually turns out to be a very difficult project.
8
33000
3000
01:01
And while I was struggling to figure out how and why mantis shrimp,
9
36000
5000
01:06
or stomatopods, make sound, I started to think about their appendages.
10
41000
4000
01:10
And mantis shrimp are called "mantis shrimp" after the praying mantises,
11
45000
3000
01:13
which also have a fast feeding appendage. And I started to think,
12
48000
4000
01:17
well, maybe it will be interesting, while listening to their sounds,
13
52000
3000
01:20
to figure out how these animals generate very fast feeding strikes.
14
55000
3000
01:23
And so today I'll talk about the extreme stomatopod strike,
15
58000
4000
01:27
work that I've done in collaboration with Wyatt Korff and Roy Caldwell.
16
62000
3000
01:30
So, mantis shrimp come in two varieties:
17
65000
3000
01:33
there are spearers and smashers.
18
68000
2000
01:35
And this is a spearing mantis shrimp, or stomatopod.
19
70000
3000
01:38
And he lives in the sand, and he catches things that go by overhead.
20
73000
5000
01:43
So, a quick strike like that. And if we slow it down a bit,
21
78000
5000
01:48
this is the mantis shrimp -- the same species --
22
83000
2000
01:50
recorded at 1,000 frames a second,
23
85000
2000
01:52
played back at 15 frames per second.
24
87000
2000
01:54
And you can see it's just a really spectacular extension of the limbs,
25
89000
6000
02:00
exploding upward to actually just catch
26
95000
3000
02:03
a dead piece of shrimp that I had offered it.
27
98000
2000
02:05
Now, the other type of mantis shrimp is the smasher stomatopod,
28
100000
5000
02:10
and these guys open up snails for a living.
29
105000
3000
02:13
And so this guy gets the snail all set up and gives it a good whack.
30
108000
5000
02:18
(Laughter)
31
113000
1000
02:19
So, I'll play it one more time.
32
114000
2000
02:21
He wiggles it in place, tugs it with his nose, and smash.
33
116000
4000
02:25
And a few smashes later, the snail is broken open, and he's got a good dinner.
34
120000
7000
02:32
So, the smasher raptorial appendage can stab with a point at the end,
35
127000
4000
02:36
or it can smash with the heel.
36
131000
2000
02:38
And today I'll talk about the smashing type of strike.
37
133000
3000
02:41
And so the first question that came to mind was,
38
136000
2000
02:43
well, how fast does this limb move?
39
138000
3000
02:46
Because it's moving pretty darn fast on that video.
40
141000
3000
02:49
And I immediately came upon a problem.
41
144000
3000
02:52
Every single high-speed video system in the biology department
42
147000
3000
02:55
at Berkeley wasn't fast enough to catch this movement.
43
150000
4000
02:59
We simply couldn't capture it on video.
44
154000
2000
03:01
And so this had me stymied for quite a long period of time.
45
156000
3000
03:04
And then a BBC crew came cruising through the biology department,
46
159000
3000
03:07
looking for a story to do about new technologies in biology.
47
162000
5000
03:12
And so we struck up a deal.
48
167000
2000
03:14
I said, "Well, if you guys rent the high-speed video system
49
169000
2000
03:16
that could capture these movements,
50
171000
2000
03:18
you guys can film us collecting the data."
51
173000
3000
03:21
And believe it or not, they went for it. (Laughter)
52
176000
2000
03:23
So we got this incredible video system. It's very new technology --
53
178000
4000
03:27
it just came out about a year ago --
54
182000
2000
03:29
that allows you to film at extremely high speeds in low light.
55
184000
5000
03:34
And low light is a critical issue with filming animals,
56
189000
2000
03:36
because if it's too high, you fry them. (Laughter)
57
191000
3000
03:39
So this is a mantis shrimp. There are the eyes up here,
58
194000
5000
03:44
and there's that raptorial appendage, and there's the heel.
59
199000
3000
03:47
And that thing's going to swing around and smash the snail.
60
202000
3000
03:50
And the snail's wired to a stick,
61
205000
1000
03:51
so he's a little bit easier to set up the shot. And -- yeah.
62
206000
4000
03:55
(Laughter)
63
210000
2000
03:57
I hope there aren't any snail rights activists around here.
64
212000
3000
04:00
(Laughter)
65
215000
2000
04:02
So this was filmed at 5,000 frames per second,
66
217000
5000
04:07
and I'm playing it back at 15. And so this is slowed down 333 times.
67
222000
5000
04:12
And as you'll notice, it's still pretty gosh darn fast
68
227000
3000
04:15
slowed down 333 times. It's an incredibly powerful movement.
69
230000
4000
04:19
The whole limb extends out. The body flexes backwards --
70
234000
3000
04:22
just a spectacular movement.
71
237000
3000
04:25
And so what we did is, we took a look at these videos,
72
240000
2000
04:27
and we measured how fast the limb was moving
73
242000
2000
04:29
to get back to that original question.
74
244000
2000
04:31
And we were in for our first surprise.
75
246000
3000
04:34
So what we calculated was that the limbs were moving
76
249000
3000
04:37
at the peak speed ranging from 10 meters per second
77
252000
2000
04:39
all the way up to 23 meters per second.
78
254000
2000
04:41
And for those of you who prefer miles per hour,
79
256000
2000
04:43
that's over 45 miles per hour in water. And this is really darn fast.
80
258000
5000
04:48
In fact, it's so fast we were able to add a new point
81
263000
4000
04:52
on the extreme animal movement spectrum.
82
267000
3000
04:55
And mantis shrimp are officially the fastest measured feeding strike
83
270000
3000
04:58
of any animal system. So our first surprise.
84
273000
4000
05:02
(Applause)
85
277000
1000
05:03
So that was really cool and very unexpected.
86
278000
3000
05:06
So, you might be wondering, well, how do they do it?
87
281000
3000
05:09
And actually, this work was done in the 1960s
88
284000
3000
05:12
by a famous biologist named Malcolm Burrows.
89
287000
2000
05:14
And what he showed in mantis shrimp is that they use
90
289000
3000
05:17
what's called a "catch mechanism," or "click mechanism."
91
292000
3000
05:20
And what this basically consists of is a large muscle
92
295000
4000
05:24
that takes a good long time to contract,
93
299000
2000
05:26
and a latch that prevents anything from moving.
94
301000
3000
05:29
So the muscle contracts, and nothing happens.
95
304000
2000
05:31
And once the muscle's contracted completely, everything's stored up --
96
306000
3000
05:34
the latch flies upward, and you've got the movement.
97
309000
4000
05:38
And that's basically what's called a "power amplification system."
98
313000
3000
05:41
It takes a long time for the muscle to contract,
99
316000
2000
05:43
and a very short time for the limb to fly out.
100
318000
2000
05:45
And so I thought that this was sort of the end of the story.
101
320000
3000
05:48
This was how mantis shrimps make these very fast strikes.
102
323000
4000
05:52
But then I took a trip to the National Museum of Natural History.
103
327000
4000
05:56
And if any of you ever have a chance,
104
331000
2000
05:58
backstage of the National Museum of Natural History
105
333000
2000
06:00
is one of the world's best collections of preserved mantis shrimp. And what --
106
335000
4000
06:04
(Laughter)
107
339000
1000
06:05
this is serious business for me.
108
340000
2000
06:07
(Laughter)
109
342000
1000
06:08
So, this -- what I saw, on every single mantis shrimp limb,
110
343000
5000
06:13
whether it's a spearer or a smasher,
111
348000
2000
06:15
is a beautiful saddle-shaped structure
112
350000
2000
06:17
right on the top surface of the limb. And you can see it right here.
113
352000
4000
06:21
It just looks like a saddle you'd put on a horse.
114
356000
2000
06:23
It's a very beautiful structure.
115
358000
2000
06:25
And it's surrounded by membranous areas. And those membranous areas
116
360000
5000
06:30
suggested to me that maybe this is some kind of dynamically flexible structure.
117
365000
4000
06:34
And this really sort of had me scratching my head for a while.
118
369000
3000
06:37
And then we did a series of calculations, and what we were able to show
119
372000
4000
06:41
is that these mantis shrimp have to have a spring.
120
376000
4000
06:45
There needs to be some kind of spring-loaded mechanism
121
380000
3000
06:48
in order to generate the amount of force that we observe,
122
383000
2000
06:50
and the speed that we observe, and the output of the system.
123
385000
3000
06:53
So we thought, OK, this must be a spring --
124
388000
3000
06:56
the saddle could very well be a spring.
125
391000
2000
06:58
And we went back to those high-speed videos again,
126
393000
2000
07:00
and we could actually visualize the saddle compressing and extending.
127
395000
6000
07:06
And I'll just do that one more time.
128
401000
3000
07:09
And then if you take a look at the video --
129
404000
2000
07:11
it's a little bit hard to see -- it's outlined in yellow.
130
406000
2000
07:13
The saddle is outlined in yellow. You can actually see it
131
408000
2000
07:15
extending over the course of the strike, and actually hyperextending.
132
410000
4000
07:19
So, we've had very solid evidence showing
133
414000
2000
07:21
that that saddle-shaped structure actually compresses and extends,
134
416000
4000
07:25
and does, in fact, function as a spring.
135
420000
2000
07:27
The saddle-shaped structure is also known as a "hyperbolic paraboloid surface,"
136
422000
5000
07:32
or an "anticlastic surface."
137
427000
2000
07:34
And this is very well known to engineers and architects,
138
429000
2000
07:36
because it's a very strong surface in compression.
139
431000
3000
07:39
It has curves in two directions,
140
434000
2000
07:41
one curve upward and opposite transverse curve down the other,
141
436000
3000
07:44
so any kind of perturbation spreads the forces
142
439000
3000
07:47
over the surface of this type of shape.
143
442000
3000
07:50
So it's very well known to engineers, not as well known to biologists.
144
445000
4000
07:54
It's also known to quite a few people who make jewelry,
145
449000
4000
07:58
because it requires very little material
146
453000
3000
08:01
to build this type of surface, and it's very strong.
147
456000
3000
08:04
So if you're going to build a thin gold structure,
148
459000
2000
08:06
it's very nice to have it in a shape that's strong.
149
461000
2000
08:08
Now, it's also known to architects. One of the most famous architects
150
463000
5000
08:13
is Eduardo Catalano, who popularized this structure.
151
468000
3000
08:16
And what's shown here is a saddle-shaped roof that he built
152
471000
3000
08:19
that's 87 and a half feet spanwise.
153
474000
4000
08:23
It's two and a half inches thick, and supported at two points.
154
478000
3000
08:26
And one of the reasons why he designed roofs this way is because it's --
155
481000
5000
08:31
he found it fascinating that you could build such a strong structure
156
486000
4000
08:35
that's made of so few materials and can be supported by so few points.
157
490000
4000
08:39
And all of these are the same principles that apply
158
494000
4000
08:43
to the saddle-shaped spring in stomatopods.
159
498000
2000
08:45
In biological systems it's important not to have a whole lot
160
500000
3000
08:48
of extra material requirements for building it.
161
503000
3000
08:51
So, very interesting parallels between the biological
162
506000
3000
08:54
and the engineering worlds. And interestingly, this turns out --
163
509000
4000
08:58
the stomatopod saddle turns out to be the first
164
513000
2000
09:00
described biological hyperbolic paraboloid spring.
165
515000
3000
09:03
That's a bit long, but it is sort of interesting.
166
518000
3000
09:06
So the next and final question was, well, how much force
167
521000
3000
09:09
does a mantis shrimp produce if they're able to break open snails?
168
524000
4000
09:13
And so I wired up what's called a load cell.
169
528000
2000
09:15
A load cell measures forces, and this is actually
170
530000
2000
09:17
a piezoelectronic load cell that has a little crystal in it.
171
532000
3000
09:20
And when this crystal is squeezed, the electrical properties change
172
535000
4000
09:24
and it -- which -- in proportion to the forces that go in.
173
539000
2000
09:26
So these animals are wonderfully aggressive,
174
541000
3000
09:29
and are really hungry all the time. And so all I had to do
175
544000
3000
09:32
was actually put a little shrimp paste on the front of the load cell,
176
547000
3000
09:35
and they'd smash away at it.
177
550000
2000
09:37
And so this is just a regular video of the animal
178
552000
4000
09:41
just smashing the heck out of this load cell.
179
556000
3000
09:44
And we were able to get some force measurements out.
180
559000
3000
09:47
And again, we were in for a surprise.
181
562000
2000
09:49
I purchased a 100-pound load cell, thinking,
182
564000
2000
09:51
no animal could produce more than 100 pounds at this size of an animal.
183
566000
4000
09:55
And what do you know? They immediately overloaded the load cell.
184
570000
2000
09:57
So these are actually some old data
185
572000
2000
09:59
where I had to find the smallest animals in the lab,
186
574000
2000
10:01
and we were able to measure forces of well over 100 pounds
187
576000
3000
10:04
generated by an animal about this big.
188
579000
3000
10:07
And actually, just last week I got a 300-pound load cell
189
582000
2000
10:09
up and running, and I've clocked these animals generating
190
584000
3000
10:12
well over 200 pounds of force.
191
587000
2000
10:14
And again, I think this will be a world record.
192
589000
3000
10:17
I have to do a little bit more background reading,
193
592000
2000
10:19
but I think this will be the largest amount of force produced
194
594000
3000
10:22
by an animal of a given -- per body mass. So, really incredible forces.
195
597000
5000
10:27
And again, that brings us back to the importance of that spring
196
602000
3000
10:30
in storing up and releasing so much energy in this system.
197
605000
4000
10:34
But that was not the end of the story.
198
609000
2000
10:36
Now, things -- I'm making this sound very easy, this is actually a lot of work.
199
611000
3000
10:39
And I got all these force measurements,
200
614000
2000
10:41
and then I went and looked at the force output of the system.
201
616000
4000
10:45
And this is just very simple -- time is on the X-axis
202
620000
3000
10:48
and the force is on the Y-axis. And you can see two peaks.
203
623000
3000
10:51
And that was what really got me puzzled.
204
626000
4000
10:55
The first peak, obviously, is the limb hitting the load cell.
205
630000
3000
10:58
But there's a really large second peak half a millisecond later,
206
633000
6000
11:04
and I didn't know what that was.
207
639000
2000
11:06
So now, you'd expect a second peak for other reasons,
208
641000
3000
11:09
but not half a millisecond later.
209
644000
2000
11:11
Again, going back to those high-speed videos,
210
646000
2000
11:13
there's a pretty good hint of what might be going on.
211
648000
4000
11:17
Here's that same orientation that we saw earlier.
212
652000
2000
11:19
There's that raptorial appendage -- there's the heel,
213
654000
3000
11:22
and it's going to swing around and hit the load cell.
214
657000
3000
11:25
And what I'd like you to do in this shot is keep your eye on this,
215
660000
3000
11:28
on the surface of the load cell, as the limb comes flying through.
216
663000
5000
11:33
And I hope what you are able to see is actually a flash of light.
217
668000
5000
11:38
Audience: Wow.
218
673000
2000
11:40
Sheila Patek: And so if we just take that one frame, what you can actually see there
219
675000
4000
11:44
at the end of that yellow arrow is a vapor bubble.
220
679000
3000
11:47
And what that is, is cavitation.
221
682000
2000
11:49
And cavitation is an extremely potent fluid dynamic phenomenon
222
684000
4000
11:53
which occurs when you have areas of water
223
688000
3000
11:56
moving at extremely different speeds.
224
691000
2000
11:58
And when this happens, it can cause areas of very low pressure,
225
693000
4000
12:02
which results in the water literally vaporizing.
226
697000
3000
12:05
And when that vapor bubble collapses, it emits sound, light and heat,
227
700000
4000
12:09
and it's a very destructive process.
228
704000
2000
12:11
And so here it is in the stomatopod. And again, this is a situation
229
706000
5000
12:16
where engineers are very familiar with this phenomenon,
230
711000
3000
12:19
because it destroys boat propellers.
231
714000
2000
12:21
People have been struggling for years to try and design
232
716000
3000
12:24
a very fast rotating boat propeller that doesn't cavitate
233
719000
4000
12:28
and literally wear away the metal and put holes in it,
234
723000
2000
12:30
just like these pictures show.
235
725000
2000
12:32
So this is a potent force in fluid systems, and just to sort of take it one step further,
236
727000
9000
12:41
I'm going to show you the mantis shrimp approaching the snail.
237
736000
3000
12:44
This is taken at 20,000 frames per second, and I have to give
238
739000
4000
12:48
full credit to the BBC cameraman, Tim Green, for setting this shot up,
239
743000
4000
12:52
because I could never have done this in a million years --
240
747000
3000
12:55
one of the benefits of working with professional cameramen.
241
750000
3000
12:58
You can see it coming in, and an incredible flash of light,
242
753000
4000
13:02
and all this cavitation spreading over the surface of the snail.
243
757000
4000
13:06
So really, just an amazing image,
244
761000
3000
13:09
slowed down extremely, to extremely slow speeds.
245
764000
4000
13:13
And again, we can see it in slightly different form there,
246
768000
3000
13:16
with the bubble forming and collapsing between those two surfaces.
247
771000
4000
13:20
In fact, you might have even seen some cavitation going up the edge of the limb.
248
775000
5000
13:25
So to solve this quandary of the two force peaks:
249
780000
3000
13:28
what I think was going on is: that first impact is actually
250
783000
2000
13:30
the limb hitting the load cell, and the second impact is actually
251
785000
3000
13:33
the collapse of the cavitation bubble.
252
788000
2000
13:35
And these animals may very well be making use of
253
790000
3000
13:38
not only the force and the energy stored with that specialized spring,
254
793000
4000
13:42
but the extremes of the fluid dynamics. And they might actually be
255
797000
4000
13:46
making use of fluid dynamics as a second force for breaking the snail.
256
801000
4000
13:50
So, really fascinating double whammy, so to speak, from these animals.
257
805000
6000
13:56
So, one question I often get after this talk --
258
811000
2000
13:58
so I figured I'd answer it now -- is, well, what happens to the animal?
259
813000
3000
14:01
Because obviously, if it's breaking snails,
260
816000
3000
14:04
the poor limb must be disintegrating. And indeed it does.
261
819000
3000
14:07
That's the smashing part of the heel on both these images,
262
822000
3000
14:10
and it gets worn away. In fact, I've seen them wear away
263
825000
2000
14:12
their heel all the way to the flesh.
264
827000
2000
14:14
But one of the convenient things about being an arthropod
265
829000
3000
14:17
is that you have to molt. And every three months or so
266
832000
3000
14:20
these animals molt, and they build a new limb and it's no problem.
267
835000
5000
14:25
Very, very convenient solution to that particular problem.
268
840000
4000
14:29
So, I'd like to end on sort of a wacky note.
269
844000
5000
14:34
(Laughter)
270
849000
3000
14:37
Maybe this is all wacky to folks like you, I don't know.
271
852000
4000
14:41
(Laughter)
272
856000
1000
14:42
So, the saddles -- that saddle-shaped spring --
273
857000
3000
14:45
has actually been well known to biologists for a long time,
274
860000
4000
14:49
not as a spring but as a visual signal.
275
864000
4000
14:53
And there's actually a spectacular colored dot
276
868000
2000
14:55
in the center of the saddles of many species of stomatopods.
277
870000
6000
15:01
And this is quite interesting, to find evolutionary origins
278
876000
3000
15:04
of visual signals on what's really, in all species, their spring.
279
879000
6000
15:10
And I think one explanation for this could be
280
885000
2000
15:12
going back to the molting phenomenon.
281
887000
2000
15:14
So these animals go into a molting period where they're
282
889000
3000
15:17
unable to strike -- their bodies become very soft.
283
892000
3000
15:20
And they're literally unable to strike or they will self-destruct.
284
895000
3000
15:23
This is for real. And what they do is, up until that time period
285
898000
7000
15:30
when they can't strike, they become really obnoxious and awful,
286
905000
3000
15:33
and they strike everything in sight; it doesn't matter who or what.
287
908000
4000
15:37
And the second they get into that time point when they can't strike any more,
288
912000
4000
15:41
they just signal. They wave their legs around.
289
916000
3000
15:44
And it's one of the classic examples in animal behavior of bluffing.
290
919000
4000
15:48
It's a well-established fact of these animals
291
923000
2000
15:50
that they actually bluff. They can't actually strike, but they pretend to.
292
925000
4000
15:54
And so I'm very curious about whether those colored dots
293
929000
2000
15:56
in the center of the saddles are conveying some kind of information
294
931000
4000
16:00
about their ability to strike, or their strike force,
295
935000
3000
16:03
and something about the time period in the molting cycle.
296
938000
3000
16:06
So sort of an interesting strange fact to find a visual structure
297
941000
5000
16:11
right in the middle of their spring.
298
946000
3000
16:14
So to conclude, I mostly want to acknowledge my two collaborators,
299
949000
5000
16:19
Wyatt Korff and Roy Caldwell, who worked closely with me on this.
300
954000
3000
16:22
And also the Miller Institute for Basic Research in Science,
301
957000
3000
16:25
which gave me three years of funding to just do science all the time,
302
960000
4000
16:29
and for that I'm very grateful. Thank you very much.
303
964000
3000
16:32
(Applause)
304
967000
1000
ABOUT THE SPEAKER
Sheila Patek - Biologist, biomechanics researcherBiologist Sheila Patek is addicted to speed -- animal speed. She's measured the fastest animal movements in the world, made by snail-smashing mantis shrimp and the snapping mandibles of trap-jaw ants.
Why you should listen
Sheila Patek, a UC Berkeley biologist, made a name for herself by measuring the hyperfast movements of snail-smashing mantis shrimp heels and bug-snapping ant jaws, using high-speed video cameras recording at up to 20,000 frames per second. In 2004, she and her team showed that peacock mantis shrimp had the fastest feeding strike in the animal kingdom, and that they are the only known animal to store energy in a hyperbolic paraboloid, a super-strong Pringles-shaped structure more often found in modern architecture.
Then in 2006, she and her team announced that trap-jaw ants had stolen the title of fastest striker from the mantis shrimp, when their research measured the ants' snapping jaws at an awesome 78 to 145 miles per hour, accelerating at 100,000 times the force of gravity. Patek's previous research focused on the sounds made by spiny lobsters, discovering that they used a mechanism similar to a violin. In 2004, she was named one of Popular Science magazine's "Brilliant 10." The Patek Lab at University of Massachusetts Amherst, where Patek is the principal investigator, continues to explore evolutionary questions through the integration of physiology, biomechanics, evolutionary analysis and animal behavior.
Sheila Patek | Speaker | TED.com