...

78. Lilly, John C. 1962. "Vocal Behavior of the Bottlenose Dolphin

by user

on
Category:

dating

1

views

Report

Comments

Transcript

78. Lilly, John C. 1962. "Vocal Behavior of the Bottlenose Dolphin
78. Lilly, John C. 1962."VocalBehaviorof the BottlenoseDolphin."Proc. Am.
Philos.
Sec. Vol.
106. P. 520-529
rd
VOCAL BEHAVIOR
OF THE
JOHN
Director, Communication
BOTTLENOSE
DOLPHIN
C. LILLY, M.D.
Research Institute,
(REad ilpril
il/IiaIni, Florida
27, 1962)
IN TIIEliteratureof ancientGreeceand Rome, to slowdownthe highspeedand high-pitchedprothere are manyreferencesto dolphins. It may be ductionsof the dolphinsatlctthus to recognizethat
appropriate to raise the question of the relation-
they call produce "humanoid" sounds.
ships that may have existedbetweenman and The ancientdolphinreferredto by the above
tlolphinin ancienttimes. Of the manydifferent writersmayhavebeenthe commondolphin,Del-
liintlsof relationshipthat may have existed, one phin2sdelphis,of the Mediterranean;however,ex-
of the mostintriguinginvolvesthe vocalbehavior amination
of the speciesof dolphinspresentin the
of the dolphin. The representation of the dolphin Meditesranean aild the Black Sea shows that the
as a messengerof the god Apollo,as an objectof Greeksmay have been acquaintedwith the same
worshipfor varioussects,as a subjectfor many speciesthat we are workingwithtoday.
stories of friendships with boys and rescuers of
men, as a subject for at least one of Aesop's fables,
The sul,ject
of this research
is the bottlenose
dolphin which has the proper name of Tzwsiops
representations
on coins, statues,and mosaics, frlLlcatzls
(Montagu)_
(fig. i). He is a shoalshow a surprising
concentration
of interest
of a
water dolphin, extensively distributed around the
special sort by the Minoans, the Greeks, and the
shores, bays, and sea near the coasts of the south-
Romans.
easternUnitedStates,Illexico,and the Caribbean
Aristophanes
(448 to 380 B.c.) wrote in The
I;l·ogs, "(the dolphin)
races here and oracles
Sea, and also is found far from this region, for
example near the southern coast of England clur-
there." Aristotle(354 to 322 -s.c.) wrote in i,,gthesummer,theMediterranean
Sea,etc. This
I-lisro?·its Alzilrlaliunz,
the dolphin when taken out of the water,
particular species of dolphin lives in small groups
gives
consisting of a bull and two to four cows with their
squeaksand moans in the air ...
has a voice land
sounds), for it is
for this creature offspring. Each group in shallow water seems to
call therefore utter vocal or vowel
have a certain territory through which they cruise
furnished with a lung and a wind
and hunt certain
pipe: but its tongue is not loose, nor has it lips, so as
kinds
of fish for food.
to give utterance to an articulate sound (or a sound
This is the animal that appears in the circus
of a vowelanda consonantin combination).
GaiusPliniusSecundus(the Elder) (A.D.23-79)
wrote "pro voce ,oemitushumane simiiis" (for a
voice [the dolphin] has a moaningor a wailing
similarto that of the human).
acts in the various marineaquaria in the United
St"tes. Sometimesthis animal is called "porpoise";it is not, scientifically
speaking,a porpoise.
The porpoisesare smaller,have shorter·jnws,and
have differentlyshapetl teeth and usually small
~a~~y scholars have labeled these and similar I'Ody sizes. Both this animal and the true por-
ancientwritingsas farfetched,mythical,legendary,
imaginative,and apocryphal. Such deductionsof
these scholarsin regard to the dolphinshouldbe
questionedill the light of the findingspresented
I)oisesare mammals,membersof the whalefamily,
tl'e Cetacea. (The gatne fish whichis commonly
c"llecla dolphinwas nameda dolphinseveralcentu'ies ago apparentlybecauseof its habit of rush-
in this paper,
ing along the surface of the sea, spurting water
In brief it looks as though the ancients knew "pw"rd on his blunt forehead, giving the appearmore about these atlimals than ally of the subse- ""ce similar to the blowingof the proper dolphins.)
q"ent scholars. ~Ve have succeeded in training
Aristotle knew the differencesbetweenthe mamdolphins to produce sounds which resemble spoken
English. We have so taught them with only one
piece of apparatus other than that available to
the ancient Greeks. The modern apparatus is
the magnetic tape recorder. This machine aided
lls in obtaining our first insights by allowing us
mal porpoise and the mammal dolphin. He said,
"some people are of the opinion that the porpoise
of the Eusine is related to the dolphin." He was
probably speaking of Plzocenlln, the porpoise of the
Black Sea and Dclplzinlls, the common dolphin (or
even T~~rsiops).
FROCEEDINGSOIZTHE AMERICAN PHILOSOPHICAL SOCIETY, VOL. 106, NO. 6, DECEMBER, 1962
Reprint
Flinted
in U.S.A.
.ils=
=
"
=
=C:-';
:
-~Rs~C:2·-/:
1
S x ~~~' ~·p
=.~
=~
T-
2'''=~:
1~3~~~~
rCL^~
=
"
=. ; =
= "c,%=;
;=·=
EIl
r:~~
Yi:?3_
i?~~;;~;
B% ~.3-t·
tu ~:
ro n
'=
*~`;g~^
rr
-na:7
N
T3 ~ -
=rn-p:,
F), ~h 2·
`o,
c
-
·,
-
""
"
·n
oc-'~,12.
-'~ = = ~^-sa
I-8
S~a
'i=
P
=-o
~
N
=·= R
2~~~~3~
h ~~~-=.
-=
^3 ,J=·
"-;~=
-
-
3"
=
rr ~
n=
e=
I ?3
Z~
2."
5
ii
'
·nn=
·~ - · -t
~ ",
3
=
=~=·=
=
n
^
^=~
c
N ri ~ J
i\=~~Y
o
C
r?
=
=
-~-·
=_
==
·
_--
_
cj
C
x?=
=-79$
5~
r,
CD
~
C
W
~
~~--
CO
I.
·
=
· W" r
C
,=_
'+
s
Lr
%
·I·
u/
Cn^
"
,r:
=
-:·e
·
3
~I
3`r,
=
cr.
z
·j
· 75
=C~
~Si
3'=
r:
~-=
r,
o
=
=
·<
c.
-rr s`r r.
" :1
=,4~"~
:: "
I:/I
I)
%
'3
"
=
C
=`
5
-r
=' ''
-i~
" r^
"
'=
r.=
m F'~
7~
L· 'Ir`
;='
I'%
==-==
rC
=
·r
"
'Er
r
1,
-
·
, 12,
-~
~C
=
,z
% ~%
"
-r
.I
7
'C
fL·~
'·
C
·r
IC
·- 1~
-t r i r. "
r
-t
-
r
=
%
<
"
--·--
=
n"rr
c
12.
=
~if't
·
" ~-~ ~ " r:
- P;r.=
3
=
" ·I--
n~;rf
n: ;
~·-
a "
"C
~~--t
C,
ii
iii
i
T sZi n - = =
~ ~T
~ = ~ =,=3
= 71 '1
:C%Z'' ". r.=
-"r
r: = r;·5
-.·T ~ " ~ ~ = -= r1 'r = ^ s : =~~= ,·- ii --= ,
C.
~= C~
1,
".
=I
F:-
:L ,S
C
_61
2
"T.
=
Cr3n-·
3\ 03--·~J'
m
c.
c='-: C "-·i,
wdr
%3""
L ^-=%
7:
==·
=
==~='
~
? ; c. -t ,
= ';;
Z~-ri-o";c
rr
=.-=
j
'<
"P,ci
o
o
T]
=
;·
" 3 1~=
=~lc.
'"
ws
r- = C~c.
p
r ;. 5s ~
r
7e
-3
=
3
-n,
=
· =-ilat
c,
i:
"
Z 5 J~ " r· .
"3
r
O
=iZ ~·:5;j o _ if ~ =~5*~
~ ~ I ;~"~
5!c, i, =?-~I
i~o ~ ;~
, , I:
T~-,
1=-_ --' r 'r=;,
=~; ,2=
E"
;
29 S
n3
L 53=-
= ? re t, i _
~-'~-3
i;-L~~~""~=E
----: =
3- ·c~
I
3
c ' · 1I~.
; "
_
~Si-<
" S ~ =;~-G:;. T~T~~
~~P r:'
i
-1
"
3 -,?la,
=_q__
1
rz
-r
=
~
1
~-12-I?i~5?·i=~=·
-j"
·~:
on$
1,,,
~= " = ~ p
= T5`Ta,
F.
C~n r~
s
~"~=
-r:
s
3. =t7rC
=
-h
R3
=
=
~~j~"·c. n ~4~ 3
c: 4r: 3
"
3
=
rr.
r; o
T~T~-
~
3~"K:
,,
e
MC·h
T~?ZIjS;~Z
-
3
';o'·-=~P
·I . 3
=
=
;-iS
i' ~ =2.~ =
R
~
3_ ~·-"
~-= ==
=
s
i~i~
-
--I--z
c
i. ·I I
~
..
"e
522
KC
32
····~
24
II'
~c,uiitl~ can l,e ~1~~ilc,rt ,is tllrre t~r ill~~r c·ztreiiicl~
~Y/~~P"S-T~"--~·a;P
l~iLII
16
8
Inutl click~ ~\·l~icli
;Ire I,rcil,;il,l·tJic ~c,-c;~llctl
j;i\c·l~,llge~lit,r;l ie~~ tellths c,i;l sc·c·c,n~l
c,r c·\;r·I1a ie~\·
O
0. 1 sec
I,trll:s,anti n:h~ntllcl· are I,rc,l~lrlgc~l
c,vc·rloll:i·er
;i,i
·.!i!,~!liiiii!iillir~;iil:~~~;.:e:l;;;;
::.;-i
i!...
:·:;I:sji~:~
I,eric,tlsof timr c,r ~\hen ge~eral;Inin~al~are emitiliiiiiiiilili!ii$li~ii~~aestll:~ii!lj;i~?l!~i;iiuiiUill;ii:
i:i;l;lili~iiii~d~Jitlll(HI/IJlt~~
ting tllt·nli the.;lrc· calle(l me\\:in:i·sI,r even ;~,clil-
;li~ll!!l!ia$iii:iililiili;iish~;~I
lii~a~bk~i~ljl;ll!llI!I1IIIIIIIJII?U
i~~gs ( tig. 4).
'fhe s(,-called j;l\~-clal,~ ;Ire literall\ estreinely
4iia!i:lr~~iiisi,kli'.;iviiiiii~lillll~
short I,arks.
111 ;111c,f clnr cxl,erience
~·ith the
arlinl;lls closing tlieir jaw-s ver?l ral,itll!. the sountl
~~I(;.
3. I~PPI.'·:
Suulltlsl,ectrograph
analysisof a whistle irt,lll.the jaws is ;~ ver~ 1~,~-irecluency "thutl."
allti collcurrcllt
train
of clicks
(sonic
and supeisonic
coruponcnts~. The frecluencies of the sounds emittetl
in tllia ~ccluct~cc·cstend from about 6 to at least 64
I;c\·i~cc. This iigure shows the fundamental, first,
ha\e
I,ccnclnphasizetl
in the sonograph.Without
ZUCIIellll~hil~iSit call be shown that most of the
rnrrg?· of the ~·histlelies between6 and 20 key/sec,
in the fulldanlental
~ilnilar presentation
anti
first
and
second
I
iii ·ii
I":
:;I
The sclua\\·l< consists of a series of clicks whose
rel,etitioli r;lte i~ varietl by the allinlal fron~ 400 per
~····LL··lh
clicks
The anitnals are
up to the order
of 1,300 per
~c·corltl
alltl do~·n to 1 per minute. This figure also
~il~,i\;
theI~Rl~dlj
Oifrecjuencies
whose
intensities
are
ullder the control of the animal. Apparently they
nlodulate irequencies selectively by changing the size
if their
air cavities
in their
nose.
iiI
i4iil(ji
u
i
I;ii
·i:i·-·
i/
Itil jII:I
ii
cluency
antitimescalesare the sameas in figrure
3 up-
···
i
'''i;
I.~~·ri·~·,·: \ ~ollic sl,ectrogram
I,f a sin~ultanc·ous
~c~ua~·l;;lntl whistle from a t,ottlenose dolphin. Fre-
···I····I··
II
wh~_~
of other sequences shows peaks
~t·cotld tlo~·n to 2-30 per second.
; i
harmonics.
forthc-cliclisc~~,rrespontling
to thoseforthewhistles.
per.
iiiili~
r ~ii!l
iii i
''i:!I
anti i;ec~,lltl Il;lrnlonics,
some of the third, anti a hit
The high frecluencies
i,i tile il,ul·tll up to 33 key/sec.
In the vicinity
of
the ~·hi.stle record the control of the frequency bands
i:, partic~~lal-lv obvious. (Lilly, John C., and Alice ~I.
Ilillcr. SL.~L.IIL.C
133: 3·165, 1689-93, May 26, 1961.)
;-
::~
iii'
iiiliII!1!iRiF
~ltjill:i
i
#hlstlc~
I:(ik
:_ii
iIii'
-20
-16
~I
:
ii
:i
-
-1
Ir
a
0.1
sec
FIc. 4. Record of a scluawk emitted under\\'ater during
stimulation by a humarl being. The s(Iua\\-l; \~a~ I,receded anti follo\~etl by a train of clicks at a lo~ rate
of repetition (e3 to 40 clicks per second~. T\\·o
\\.histles occurred during this s(]uawk, one (top trace)
our animals tends to click only on the left sitle and
\\·histle onl~ (,t~ the right side and call do so on
I,othsides;imultaneously
or separately(fig.Sj.
4 sonic spectrographic analysis of the clicks3
showsseparatetlorninant
frecluencies
for eachclick
(fig. ,3).
Tn the case of the whistlel one carl see
the iunrlalnental x F, and at least 2 x F, 3 x F
~ultl;~~least 3 x I;. The animals can protluce the
first. secc,n(l;In(l uI, to the seventh harmonic of
the I~,wcl~tirec]uency emittetl.
1X r; is usually,
thr,ughncl~necessarily,corltinuousanti showsthe
c]c,l~~jll;tnt"Trccluency nlotlulatetl" nature of the
I~,west irecluencv grr,uping·s.
at 0.7 and one (middle trace) at 0.05 second after the
~,eginniIlgof the squawk. The freclucllcv calibration
r the tinle
tin kilocycles per second) is given, \\·lth
scale(realtime),onthebottom
tract. ~ record
of
frequencies
versustimeon the sonograph
of this same
tape
shows
alloftheseirequencies,
plusothersupto
at least 64 key/sec. Clicks of repetition rates of 110
to 400 per second occurred, in this particular squawl;,
,l,,,t 0.82 secondafter the heRrinnitla
of the sclua~l;
(beginning of middle trace!. In other squa\\·ksl
clicksat rates up to 800per secondhavebeen sus-
tainctl for as nluch as 0.5 second. A loud squawk
was heard from the open hlowhole itl air sitnultan-
eouslywithenlission
of this scluawli
underwater.
(L.illy,Jo~~nC., at~tlAlice M. Miller. S~·ic·~~l.e
133:
3465, 168?-1~~3,Alay 36, 1961.i
52,3
JOHN C. I,II,LY
Since their teeth interdifi·itateso well and go into
IHIIR
sot't tissue rather than collitlingentl on with one
another tjlat there seemsto I,e no ~:avthat the jaws
II'IIOC.
AMER.
PHIL.
SOC.
InalR
I~ - i:-·--_-
:Ilr~;
thenlselve~or the teeth coultlgive a "clap" or a
0 ~"~~l~~ii~i~E~:·4·irie3~Si~i~ii~i~
%: i:~ ~~*8
snal'.
\\~e have t'omntl, however, ~·hen an animal elnit~ a ver\l short sharp series of loutl cliclcs
~
( fig. ,,). that he tends to open anti close his mouth
(This gesture acts as a threat.) ~Ye
have seen ott~eranimals move rapitlly away irom
::,L~-~i~"-~~""""ag~~:o·
·:':::;::::: -:Orll_E~fl(CR:-·:_::_
::;::UNOEAHITER
.
S11C11
a ,oesturing ··-···
Inrl rr~rll;7;no?~,;~,,·,1LLIIIL
n~~rl..·n
·v~LLllrlllS
LLIIIIIILII
~VC ~..r
VLLI-
sel\es 1,~~11
our hand or arm or leg rapidly away
front
such;ln
aninlal.
Such
a O.PCtllTP rnt~~ll;nPrl
b'~"L'~
C~""'-""~-~L
11_8
is.s
x:
~itll the loutl short click train prol>ably,aaverise
to the mistalien notion of a "jaw clap."
Severalauthors ha~e tlescriheed
a specialI·rintl
-96XID
Y I
I;Ic. 5. Sonic spectrogram of sounds produced in air
and under~ateralternately
by a bottlenose
dolphin.
ol~clicl·rin~\\·hich ~~singthe analogue of the creak-
The upper sonogramis recordedwith a narrow band
ing rusty hing·ehas been dul,l,ecl"creakings."
This seemsusuallyto he associatedwithfoodfintl-
45cyclespersecondfilter. Thebottomsonogram
is
recorded
with
a
wide
band
of
300
cycles
per
second
filter. The first grroup of sounds produced are re-
ing anti iootl recognition nl7clis thought to be evitlence of the activity of these animals escellent
ported by listeners to resemble those of a stringed
sonar" operations, descrihetl so well by Scheville,
f;'""p OfSoundsresemblebarlcs. The wide-bandrec-
Norris. I(ellogg, and others.7·S~"~'0~1'~"~ 1"
Some
of our recent
studies
throw
musicalinstrumentresemblinga hanjo. The second
ortl showsthat the soundsconsistof short sharp
cloul,t on the
pulses of sound in which the repetition rate is varied
by the animal. The narrow hand shows that the
necessary anti sufficient sollic cl·enkings as the
pitch ill the first group of sounds is 740 per second,
source of the sonar pulses. In our esperience
i" the secondgroupis 1,050cDs,in the fourthgroup
thel-e
areitltl·nsoi~ic
plllscs
cluite
separately
ei?litted Bto~s~i~rts
atI,2so
cpsatldfnlls
to9GO
cpsduring
the
irom the so~~icpitlscs."
Ii one listens with a radio
This record is 1.2 seconds long.
The modulation
receiverconnectedto a hydrophoneat about 100
andselection
of variousbandsof frecluencies
by the
liilocycles every so often one call hear a stream of
animal is well illustrated by the first, second, and
I"'lsesI,eingen7ittetlstraightaheadfrom the gi\rell
thi'd
sounds.II1the thirdsoundthe animalinhasconselected the fifth partial alone for enhancement
animal.
Such ultrasounds
can I,e dissociated
with
any sollic output whatsoever. The ultrasonic
p';lsescrunbe loclced
in or not locl<ed
in withsonic
I"'lses.
When the sonic pulses are associated
Mliththe ultrasonic,they are apparentlyl,eiIlg
u~etl to comnlunicate
the sonar information
to the
;Schevill,M-.,and B. T,a~·rence,
High-irecluency
audi-
tory response of a bottlenosetl porpoise (Tl~,.siops tr2liacntlts), Jozlr. E1L-pfl.%ool. 124: 147, 1953.
trast
to
the
first
sound
in
which
the
animal
has
selectively
dropped
outthefirstpartialandenhanced
the third,fourth,fifth,sixthandseventh.In the
Seco"dsound the second through the sixth partials
a'e
enhaIlced
as the firstpartialis dropped
out
towards the end of the emission. Each one of the
barlcs on the original
record call be shown to be a
slnall Ilumber of clicks ranging from 5 to 12 (see
test).
"Schevill,W., and B. Lawrence,
Foodfindingby a ~lea'byanimalI,y meansoi the lowercommunica-
captive porpoise,B,c~i,iol·n
53: i, 1956.
tion (soIIIC) hand of frecluencies.
!'I~ellogg,
W. N.,R. I(ohler,
II. N. Norris,Porpoise Thusa moreup-to-elate
catalogue
ofsoundsprosounds as sonal- signals, Scie,2ce 117: 239, 1953.
"'I(ellogg, \li. hi., R. I(ohler, H. ?T. Norris, Echo
r-anging in the I,orpoise, Scielnce 128: 982, 1958.
tluced
by dolphins
would
inclutle
at least
five
classes: (1) whistles, (2) sonic clickst (3) ultra'1I(ellogg, W. S., R. I(ohler, H. N. Norris, Auditory Sonic clicl<s, (4j rapitl click trains with various
percel,tionoi submergedobjects by porpoises,Jollr. characteristics. anti (5) a class of sounds ("hu-
~lcotlsf.
S`oc.
rill2C)..
31:1,Igjg~
1" McRlide, A., quoted hy W. Schevill,
Evitlence for
manoicl")whichwe tliscussat morelengthin the
echolocationby cetaceans, Drcp-Sea Rcscnrcla3: 153, latter part of this paper.
1956.
'" Norris, I(enneth S., John II. Prescott, Paul V. AsaDorian anti Paul Perkins, AI1 experimental demonstration
of echo-location I,ehaviour in the porpoise, ?'lcisiops trtul-
cntzls(i\iIontagu),Biol. 82111.
120(2): 163-176,April,
1961.
NATUIIALI,Y OCCURRING VOCAL EXCI-IANGES
EETWEEN DOLPHINS
I)uring the course of our stntlies on the bottle-
nose tlolphin, we found tlefinitive evidence that
VOL.
106,SO.6, l~b2]
VOCAL BEHAVIOR OF THE DOLI'HIN
F-
I
·1;11 ~""~"'''"~'~~`
"""'
524
ing or a wailingdepenclingr
uponits duration,show
''"
each to be a fast series of clicks.
(Some of these
are asymmetrical cliclcst i.e., there are short pulses
_r
not I,alancecl out by other short p"l"es on the other
side of the base line (fig. 2). Ill other words, the
" '
emitter in the dolphin's head is a nonlinear biological tlevice, not a linear one.!
12 I1
I.IIC
Such sountls seem common only during intense
Frc. 6. A graphic record of a vocal eschange between
two tlolphins. (ToD trace in each pair) emissions of
emotional escitement on the part of the animal,
either sexually arousetl, an=ry, or in similar in-
the female(F): (bottomtracein eachpair) emis- te"se states. (One can elicit such sounds (uiith
sions of the Inale (M). The upper pair of traces
sho~s a click-and-whistleexchange; the lower pair,
trainingj under nonelnotional conditions.)
These sounds are used naturally in what one
a continuation
might call cu7lotionnl csclznllgcs."
of the same
record
without
the cliclcs.
(Dots bet~een pairs of traces) seconds of elapsed
time; elapsed time for the whole record, 15 seconds.
is irritated
with
the
behavior
If one dolphin
of another
one
or
of
For reproduction,the peaks of the clicks of the fe-
a human ol,server, he et~~its such a rapid series
male were marlced with black dashes; the tips of
of clicks at great intensity
and at the same time
those of the male, with black dots. Whistlesare makes gestures such as rapid head movements,
I~umbered in sec]ueIlce for each animal.
turbances
in the base
line
are,
in most
Other dis-
cases,
water
either vertically
or horizontally,
with the mouth
Iloises.(T~illy,
JohI1C.,andAliceM. Miller,Scic~lce, either opell or closed. Such movements and
134 (3493): 1573-1576,December 8, 1961.)
sounds signify in no uncertain terms that the animal is emotionally upset. For esample, if an ob-
they tlo eschang·esomeof the abovesoundsin serverputs a leg into the tank and the dolphin
appropriatefashions."
doesnot want him to enter the tank at that point,
either silent, in the case of the whistles, or there is
mal I,egins to I,ang on the leg with the sitle of its
Eac2l anilnal waits until the other animal is
such a sound Inay be elnittecl just before the ani-
an opportunityto alternatewithit~the train in the jaw in rapid oscillatingmovements. They will
case of cliclis (fig. C,). One call hear clicleex- treat one anotherin similarfashionthoughsomechanges goillg on bet~ieen two animals with little
c,verlap. A close study of the overlap shows that
r.
~1
they alternate their clicks during the period of
overlap. The whistles are very "politely" eschangetl escept for one case called a "duet." Each
animal whi·tles simultaneously; they match fre-
,~-r//pL·l·4
Mp.--·CI·.l~·-·g·Uaa~·U~
7
qu"n'ies so \\lellthat one can hear beats between
the two emissions. (This seems to be analogous
to the rather irritating hal,it that solne people have
of sayinga wordsimultatleously
withone's self.)
Sonic spectrographic
analysis
/··
~histles
(fig. 7).
The
We
have
iountl
:'"
F
2
3
·*"O"O·"a""Bg"L"·"**·6Q""""*"scsrp~~
of this kind of
fundamental
that
some
~b·
--16 kcps
fre-
cl"ency is usually continuous. The first, second,
thirtl, fourth, anti higher harn7onics are usually
discontinuous.
.'··
· ~"4~ . ~ .. ~P"C·~s~
eschangreshows some of the real complesities of
these
0.1 sec.
8
of the
~
-0
FIc.. 7. Sonogramsof a portion of an exchangebetween
t~;o tlolphins. (M)
Emissiot,s of the n,ale;
(r;),
of
harmonics of these whistles can still I>e detected
as high as 150 I<ilocycles. The harmonics appear
to the numbering
of the amplitude
and tlisappear in comples ways.'l
The
(f)
(2f and3f) Inaybe seellon thesesonograllls.Addi-
Each
tlolphin's
voice differs
very IT7UChirom
th' female. El-nissionsare numbered to correspond
fundamental
and
the
first
trace in fig. 6.
two
overtones
tional sonogratns with twice this frequency scale
each other voice For esample, son~e anir7lals
fill in I,etween emissions with low frequency
meIltal
~histlin= (somewhat the way some people say
enhancement
of the higherfrecluencies
ill recording
"aah" I,et~·een words).
Analysis of the sountls called a bark or a mew-
show that energy itl the 3rtl, 4th and 5th overtones
tlecreases rapidly as con~pared to that of the fundaand
the
first
two
overtones.
There
was
some
the soIlograms. (Lilly, John C., and hlice M. Miller,
Srir,lce134 (3493): 1973-1876,
DecemberS, 1961.)
525
JOHN C. LILLY
times even more violently than they treat the
[PKOC.
AMER.
PHIL.
SOC..
ALTERNATING
AIRANDUNDERWATER
EMISSIONS
humans.Afterseveral
weeksincaptivity
dolphins
apparently
learn that humans
souncisen~ittetlunder water very easily, and they
begin to espress their state by emitting such
sountls in air above water aimed at the particular
human involved (fig. 5).
THE
~skCps
do not hear these
PRODUCTION
OF SOUNDS BY DOLPHINS
IN CONTACT WITH MAN
·:~
t
w
o
· a*l~dp·..pp·.~bl~~~·a
Certain sountls of the tlolphins tend to be emit-
~~
0·5fec·
tedin air aftertheyhavebeenin closecontactwith
::1·-:~
_8kcps
man for several weeks." At first they tend to
emit their naturally occurring soullds described
above and merely transfer them irom untler water
into air. The first sounds heard itl air emitted by
a newlycaptured
dolphin
areusuallywhistles
and
-::
·r
Z;
,u,~i·
·'"
clicks.
'
In the case of baby dolphins, there is a stream
of I:,ubbles associated with their production of
whistles under water.
o
Ii the head is moved into
the air one can hear the whistles very faintly in
air being emitted at the exit to the blowhole slit.
This is the extent of the known naturally occur-
ringairbornesoundsof the dolphin.
As the ani~nals
remainin captivity
for longer
ar7tllonger periotls, they open their blowholesand
emit estremely loud clicks and whistles (fig. 8).
For a personstandingfiveor ten feetawayfrom
such an anitnal, the intensity of the sound call be
so high that it is definitelyuncomfortable.The
clicl<ssotlntl like very loud hand clapping by a
i:,
~~~~~~Alo
1
FIc. 8. A dolphin emitting whistles anti clicks sirnultaneously alternating between air and underwater
emissioIls. This anilnal is whistliIlg with her right
nasal passage and sacs and clicking with her left
"asal passageand sacs. At the beginningof the
recording
therighthalfofAtthe
blowhole
is openas
0.12 second after the
she begins her whistle.
hesil,lling
ofthewhistle
(upper
alldiuwer
trace)-she closesthe right half of her blowholeshuttingoff
the eT"ission
of souIldin air (lowertrace) and en-
h""ci"gtheunderwater
whistle
(uppertrace).The
whistle rises in frecluency from the iIlitial value of
3.5 kc to 6.3 Itc followed by a silent break.
The
whistlestartsagainat 7.8kc,andfallsto 1 kc (see
below). Just beforethe silence,underwaterclicks
human: tile whistles sound as loud as the maxi-
a'e Started (upper trace) which are not detectedby
mum intensity of those by an expert human
the beginning of the whistle).
whistler,I,utare of a higherfrequency."
the air microphone (lower trace, 0.6 second after
After the emission of
15clicksuIlderwater
theleftblowhole
is·opened
and
After several weelcsof such noises one begitls
to notice a changingpattern of the airborne sounds
clicksin air are detectable(lowertrace,oIlesecond
"fte' the beginningof the whistle). The next eight
sions,greaterrichnessof selectionof frequencies
whistle.At 1.38seconds
fromthebeginning
ofthe
to morecomplessoundsinvolving
longeremis-
CliCkS
aredetectable
illboththeunderwater
andair
channels. She stopped clicking before the end of the
anti harmc,nics. In our experience such changes
occur if and only if people have been talkirngto the
animals tlirectly and very loudly individually (fig.
5). Slowly but surely these sounds become more
whistlethe right half of the blowholeis onceagain
O"e"edand the whistleappearsin air (lowertrace).
trace.
and rnore like those of human speechO (figs. 9
anti 10).
In theoriginalrecording
thewhistlecanbetraced
downto 1 kilocycle
1.42seconds
afterthe beginning
II1 air the whistle can be seen to continue beyond the
4 kilocycles at which it disappears in the underwater
When I first discovered this effect in 1957 it
Oithe whistle. The abovewhistlerecordingis con-
was in a seriesof experimentsdesignedfor totally
Sid"edto be twowhistleemissions
separated
by a
different purposes than studying the vocaliza-
Sho't silence. This pair closelyresemblesthe two
tiofls.'ilThe effectsoccurredduringa studyon
SI'eci"l
whistles
of the bottlenose
dolphin's
distress
1" Lilly, John C., and Alice M. Miller, Operant conditioIling-of the bottlenose dolphin with electrical stimu-
lation of the brain, Jozlr. Colltp. G Physiol. Psychnl.
55: i, 73-79,February,1962.
call. Only the fundamental irequencies are shown in
this display. The first and second harmonics are
12-30 decibels below the intensity of the fundamental.
The air record shows many room echos betweenthe
ceilingand the water.
V~L.106,h.O.6, 1')621
o
tlOCAL
stbp it
BEHAVIOR
o k'l 8
z
m
o
OF THE
:::
.~-
It was many months
:
: j ~-`-lrr----l-~
.~;;u~·;-*esuY-·~S---
----
Human .
I,efore we were able to
believe this evitlence. From November, 1957,
until ;\/Iay, 195S, we stutlied the accumulated iniorn7ation from three animals; in May a preliminary announcement of our fitldings was Inade.l"
b:-
z
526
-Bkcps Of"'essages,we did not cluitetrust this evidence.
-~a~--
..
DOLT'HIN
:::::(:
· ··-·
These findings were met with total disbelief on the
I'"'t Of others who had worketl with the dolphins.
O
: ·:.,Do/p~in
(Itl some cluartersthis disbeliefesists even to-
tlay.)
However,
we also found there were many
-8 kcps peoplewhoseinterestwas ilnmetliately
aroused
o
Z
a
..
anti many of our "loyal skeptics" encouraged us
to continue.
Since then Mie have acculnulated
;.,'·
m ;·~i:;
much corrol,orative evidence, not only bearing out
s
t
:nu~Dac·
i
o
these early fintlings, but extending the observations
into new areas of esperience.
The first copies of the human voice by the ciol-
~IC.9. Part of a vocalexchangebetweena humanand I)hi" tin 1957 and 195S) were at a relatively low
a tlolphin.
narrow
The
As in previous figures the upper trace is
band
whole
and
trace
says "stop it."
makes
a sountl
tl,e
There
which
saying 'Lokay."
lower
trace
is 2.4 seconds
is
wide
long.
The
band.
ba
human
bal
I
-4.8 kcps
is a pause, and the dolphin
closely
resembles
The fundamental
a small
child
pitch of the dol-
4
phin's "o" starts at 500 cps arld rises to 800 just beiore
the sound oi the "k."
The pitch of the terminating
1LE"
SOUlld
starts at 950 cl,s and falls to 800 cps. It
is to be noticetl however that the dolphin enhances
4S,i.
-~;~-.'-·ra~4*e~a~w'~·
variouspartialsselectively
so that duringthe "o"
the first two partials contain the high energy and
during
the "E" sound
and fourth partials
secoIld through
the
energy
starts
in the
O
third
at
and broadens out to include the
the eighth
partial.
nuring
_4.8 kcps
the rather
noisy "1<'1sound the maximal energy emitted is intermediate I,etween that of the "o" sound and that of
R-
;
the((E"SOUIld.
Theoverallmaximal
energybaIltllies
bet~;een i kilocycle and 6. II1 contrast to the dolphin,
the Ilun~an's masimal
energy is between 200 cps and
3.50 kcps.
This record illustrates
that a dolphin in
:,
:ri:
ui:i
m
~~~6-2.4
eschanges with a human can emit sounds resembling
a wortl
ether
than
that
spoken
by the human.
~~
.- ::::;~~~~_ii~:!:·:~
· :!~:~;:,~
~; :Humor
the brain oi the first animal; we need not go into
the tletails i,i the other experiment here.'"
Dur-
FIc.
10.
A humaI1
"bye-bye"
dolphin
exchanae.
and the dolphin
The
says
human
something
says
which
ing the course of these esperiments under certain
resembles"ba-byee"I,eforethe humanhas finished
contlitions,
the last portion of her emission. This record is 1.2
r:-he animal was emitting
very peculiar
sountls that we had not heard from any other
secondslong. The frequencyrange shownis 4.8
tlolphin.
Later in retrospective
analysis of the
tapes on which we were recording our information,
we heard unmistakable resemblances to the human
voice in these emissions from this dolphin.
Further atlalysis at~d further study gave us a totally
kilocycles which is the hand used by humans to traIlsmit tneaning in speech.
The lowest frequency reached
unespected and surprising correlation between
fromabout1,000
to 8,000.Thisrecordillustrates
the
I'"'tS Of what I had said in tlictating information
to the tape anti the subsecluent emissions by the
by the dolphin is 200 cycles towards
the end of his
emission. The selective enhancement of the partials
above 1,000 cycles is illustrated in the upper trace.
II1 general
tile human
articulates
with
energies
from
400 cycles to 3,000, whereas the dolphin uses a band
extremely rapid responsesof a cooperatingdolphin.
la Lilly, Sohn C., Sorne considerations
regarding
basic
tlolPhin. Because of the well-known,facile nature
,,,,,l,,,,i,,,.i
positiveantinegativetypesof motivations,
of the human ear which call supply missing parts
Al,lcr. Joltr. Ps3'cl2infr3115: 498-504,December,1958.
527
JOHN C. LILLU
[I't~OC.
~MI~R.
I'HIL.SOC.
amPlitutle. The most recent ones on the part of
She hit the side of the tanleand said "stop it." He
our three current animals are sometimes painfully
squirted her again, she hit the side of the tank
loutl for the human observer. The tlevelopment again anti said loudly, "stop it." At about the
of these sounds by a given dolphin is as follows:
fourth "stop it" said by Alice, Elvar started
Our longestobservationshave I,een of an ani- emittiIlgveryshort sharp soundsin air very loudly
mal I,y the name of Elvar who joined us on the
fifth of July, 1960. Quite early in captivity Elvar
with his blowhole open. On playback of the resuiting tape at normal speed, this sounds like a
was in intraspeciessolitudebut in constantdaily very high pitched "wee." Immediatelyafter the
contact with human observers from that day to
fifth "stop it" by Alice he said something which
this (25 months). Instructionsto the humanob- at normaltape speedis recognizableas a two-part
servers are to talk to him using words appropriate sound, very high pitched and very short. By
to tl~etasks at hand. Itl fact, they are asked to slowing the tape down by a factor of two, one can
sI)ealeloudlyto him ill order to penetratethe sur- hear a very definite"stop it" producedby Elvar;
face of the water with the sound. For a long time there are many echoes in the room because of the
each person in the laboratory was quite skeptical high intensity. From the be,oinningit was found
of the r-esultsto be obtained by this method. Each that if he is cooperating he follows the human's
one felt a little foolish in attempting to elicit wortlsvery rapidly. For esamplehis "bye-bye"coI'iesof hun7anwordsfroma smallwhale. How- likesoundfollowsimmediatelyafter her "bye-bye"
ever, by persistent efforts, in Septenlber, 1961, we
(fig. 10).
were able to determine that Elvar was quite cap-
From that episode onward, we began to elicit
able of the production of sounds like those of human speech. We-began to encourage him to shape
many more words from Elvar and iound the following order of events.
tlefiniteand distinctivewords (figs. 9 and 10).
When Elvar emits a sound in responseto a
During his first year in captivity he had gradu-
sountl repeated by a human again and again, in
ally tlevelopecl air-borne sounds analogous to those
general he tends to breale it down into its com-
that we use in our speech including the vowel ponent parts, and emit one of these components
sounds, plosives, hissingrs,and similar noises that after each production by the human. A typical
we emPloyin communicationwith one another, sessionis that of October23 in whichAlicesays
However
he was emitting
these
sounds
in fre-
"more Elvar"
and Elvar comes back with "more
q"encies well above those which the nonnal atlult
human nlale or female-emits. They were more
co~np"'"hl"to those emitted by a very small child
as it begins to emit these sounds in a very high-
var." a very high-pitched shortened "morelvar"
all run together and then finally in a clear highpitched nonalcl Dude voice with a delphinic accent says "more Elvar," in a clearly hutnanlike
I)itched ialsetto.
understandable tnanner.l0
Spectrographic
sonic analyses
showedti~atthe lowestfrequenciesthat he was emI'loying at that tin7e were of the order of 1,000 to
From the productionsof Septemberthrough
those of Octol,er he had lowered his lowest fre-
2,000cyclesper second (fig. 5). Resemblances cl~~ency
so that one couldsee from spectrographic
to h~uman
speechwere heard by slowingthe tapes a~~alyses
that he hall movetlhis lowestpartial well
tlownI,y a factorof two or four. By September, tlown in the region which is producedby illiss
1961,El\rarhad accomplished
the first task which Miller,i.e.lfrom about 1,000cycleshe had moved
we set him, i.e. emitting secluencesof sounds bear-
tlown to about 350 cycles.
ing a high pitchetl resemblance to the sounds that
h~uman I,eings employ in their speech activities.
Any time that he wishes he call raise his pitch
al7d the frecluellcies emittetl back up into a proper
He had not pet formed any words which we could
tlelphinic region at two to four times the nonnal
recognize;thesesoundsweremorelilcethe I-,al,- ,,pperhumanfrequencyregionand decreasethe
blil7brr
(,I a bal,y before the wortls are accluiretl.
time during which he puts out the emission
(with
In Septemberwe decidedthat he was readyfor the proper tlelphinicduration). Thus one finds
step 2, the formation of understandable words. A
typical esperience is that of R/IissAlice n/Iiller on
ill the I,rotluctions of October, Nove~nber, and
December estremely short and high-pitchetl
the tenth of September.
sountls n~ised in with those that resemble human
room and started
of his tank.
In
spealting
his
usual
She entered the tank
to Elvar
fashion
at the edge
he
filled
his
mouth full of water and squirted it all over her,
1"Lilly,John C., and Alicebl. Miller,Productionof
I~umanoid
speech
script, 1962.
sounds
by the bottlenose
dolphin,
Manu-
VOL.
106,
NO.
~,l~h21
VOCALBEHAVIOROF 7'HE DOLPHIN
528
wortl~at Ilc,rmalspeetl. Severalof these very g·ootlcopyof my wortl"water"emphasizin~
nly
hil~h-l'itchetll
shortsoundscanI,eshownto hevery 'r." In contrastto iVIisshliller'smelotlious
"r"
short high-pitchetl
colnpressetl
reproductions
of sountl,my "r" is a veryrattley,I,uzzyone. He
human\\-c,rtls
I,yplayingthemI,aclesloweddo\\-nl copietlthis"r" \\-itha loudfast sequence
of cliclis.
~ome\\-hatin the fashion of the 1957 results and
I-Ie brought his lo~-est pitch do\rll to about 200
the "ilop it" of Septeml,er.
c!-clestluringthis rendition,a new low for his
In ~e\lerale.·;l,erimentswith difierent human ol,- lowest frecluencies."'
~ervers atltl tlifferent human voicesl it is shown
that Elvar tentls to esamine each new h~umanvoice
\I'e are still worl<ing on stage 21 i.e., perfecting
intlivitlual -~ortls in English, with three animals,
at~tlattenll,ts to reprotlucethe novel characteristics
of that \oice compared to the previc,us one. A
typical esample is my voice ~orl<ing with him on
the secc,ntlof Deceml,er.1961. I ran him through
a \-ocal,ular~tl~atr2licehall I,een helping him per-
Elvar. Chee Chee, and Sissy. Slowly I,ut surely
\\e are Illoving over to stage 3 in which we are
insisting on the animals using the words in the apI,ropriate situatiolls and contests. M'e are also
esploring the ~,eginninSsof stage 1, the ability of
~ect n·hich inclutletl such words as "speak, up,
the animal to abstract iniormation and to transmit
loutler, morel etc." I-Tekne~ the word ilsquirt" it app'op'iately I,y means of wortls. The analyses
I repeatetl the wortl "sq"irt" in artier of these latter stutlies are not far enough along to
to intluce him to scluirt water an(l at the same
report at the present time.
time to say "scluirt." I-Ie went thro~~f:-h
the varic-,usI)arts oi the \ord "scluirt" anti finally said it
T,, s~unmary,the hottlenose tlolphin is al,le to
,,,,it under water a series of sounds in his natural
clearl!- eno~~gh for that particular tlay.
state ~~hich ruin irom \\rhistles and cliclis to huzz-
I then
\\-ent on tc, \\-orli with him with a new wortl ;,g,~ mewi~~gs,
anti I,arl<ings,anti several unclassii;water" in the phrase "squirt water.""'
fietl series of sountls.
In captivity, the naturally
Elvar practicetl ''wa" separately irom "ter" and occurring sounds continue and can he showI1to he
then fillallv came out in the clear ~·ith a fairly usetl in formal polite alternating eschanges bet~een the intlivitluals of pairs of anilnals. The
~:histles anti clicl<s apparently are used to e?c-
BODY
LENGTH
INFEET8 INCHES
2
~ i. 6·9'~1T 6' 9'~13·6~9·81
s·91
TURSIOPS
TRUNCATUS
cn
B
change
iniormation
ofa nonennotional
nature.The
barlis are usetl i~~emotional situations
to influence
other allinlalsanti humans.
Tn adtlitic,n to the naturally
1800
occurring
under-
\\ater sounds, air-borne sounds are etnittetl in
great profusion after a period of captivity. Dol-
u
1600
Ihhinscan I,e, by appropriatetechniclues,induced
to iormsountlsresemblingr
thoseof hulnnaln
speech.
XTtera periotlof emittingsuchhumanoidsounds,
~ 1400
z
the animals can I,e intlucetl to I,eg-into form rela-
ti\-elv clear words very loudly or very softly in air.
~ 1200
Theseresultsillustrate
thattheverylargeI,rain
of T1~1·siops
tl·lllcaf~is (20 to 40 per cent larger
1000 1
iOO
'
150 200
250 300 350
80DYWEIGHT
IN POUNDS
Frc. 11. Relations het\~;een brain weight, body length,
thal7 that of the average
human)
(fig. 11) may
have ~ithi~~its co~nplesstructure speech capahilities, ii not realized, at least potential, sinlilar to
those of the human.'"
and I,otly \\·cigi~t of the bottlenose dolphin. The Ilorrnal human braiII ~'eights lie in the lower portion of
this graph. The critical threshold brain weight for
the develoI,inent of speech iI1 the hun~an child lies
just helo~this graph(800-1,000
grams). The
SUMMARY
It can be shownthat the l:,ottlenose
tlolphin
smallestbottlenosedolphinbrain foundby us was (T~~vsiopstl·llllcaf~~s
i~Iontagu) elnits several clif1,150granis. Presumahly the tlolphin's brain may
ierent classes of coInples sounds.
Sonle of these
I,elargecnougllanticolnplesenoughfor the develop- sountls are encountered in the natural state, others
Incnt of speech pl.ocesse~ analog.ous to those of the
human.iLilly,JollnC.,andAliceM.Miller,
JD2IY.
are acquiretlduringlongperiotlsof captivityin
Cn~lp.~- I'l1~siol.rs?lclaol.,Fehruary,1962.)
close contact with scientific investigators. Some
529
30HN
C. LII,L?I
[PROC.AMER.PFIIL.SOC.
of these sounds are emitted under water without
mimetic activities
loss of air, others are emitted in air with the open
and clean-cut.
blowhole above the water.
The naturally occurring underwater sounds consist of several classes inclutling whistles. clicks.
barli~, creakings, etc. All of these sountls are
relatively hig2~ pitched compared to those of the
hm7lan voice and estend from 2 kcps to approsimately 150 kcps. It can be demonstrated
that,
when a dolphin opens its blowhole anti emits these
sounds, the characteristic
irecluencies heard are
lo~ler than those heard simultaneously
under
animals of using such "words"
in progress.
\vater.
means
cause
.-i dolphin can I,e ititluced by various
to emit
of their
another
class
resemblance
of sountls
to the
sounds
in air.
Be-
of human
speech we have named these sounds "humanoid
emissions."
rlnalysis I,y means of the' sound
sPectrograph and oscillographic methods demonstrates that these soullds are basically white
noise-his~ings anti/or higrh-pitched I,uzzing~s, motlulatetl in selective frequency I,ands by the efforts
are at times surprisingly
clear
Tests of the capability of these
appropriately
are
~CKSOWLEDG~IENTS
The work clescribetl in this paper has been supported by the National Institute of Pvlental Health
of the National Institutes of Health; the Coyle
Foundation;
the Office of Science of the Department of Defense; the Office of Naval Research;
the ?\'ational Institute of Neurological Diseases
and Blindness
of the National
Institutes
of Health;
the Air Force Office of Scientific
Research;
the
National Science Foundation
tions
to the
Communication
I am indebted
Communication
and private contril,uResearch
to the staff,
Research
Institute.
the Trustees
Institute
and
to
of the
several
intlividual scientists for their helpful criticism, inclutling Franklin Cooper, J. C. Steinbergr, Kel7neth
of the animal. Some of these emissions appear
N· Stevens, Arthur House, Frank S. EssaDian,
to be attempts on the part of the animal to repro-
Orr E. Reynolcis, David Brown, James Bosma,
c!uce words
and Anatol Holt.
spoken
by the investigators.
Such
Fly UP