Professional Documents
Culture Documents
Brill: BRILL Is Collaborating With JSTOR To Digitize, Preserve and Extend Access To Behaviour
Brill: BRILL Is Collaborating With JSTOR To Digitize, Preserve and Extend Access To Behaviour
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
BRILL is collaborating with JSTOR to digitize, preserve and extend access to Behaviour.
http://www.jstor.org
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
VALERIUS GEIST 1)
(Zoology Department, University of British Columbia, Vancouver, B.C., Canada)
(With i6 Figures)
(Rec. I5-XII-I965)
INTRODUCTION
In the evolution of mammals horn-like organs have appeared a number
of times. We find them assuming a great variety of forms; small, skin
covered pedicles in the Okapi (Okapia), large horn curls in sheep (Ovis),
or many-pointed, bizarre bony paddles in moose (Alces). Horns originated
on different areas of the head and grew from several tissues. We find horns
on animals as distantly related as Cervus, a deer, and Mylogaulus, a miocene
rodent. Also, we find horns evolving independently a number of times in
closely related groups, such as in the ruminants and the rhinos. It is inherent
in our concept of the selective force in evolution that these organs, whatever
their shape, size or position should have been an important, adaptive feature
of their owner. Yet the evolution of horns and antlers has been a puzzling
problem. COLBERT(I955, P. 398) writes: "on the fact of it, we would think
that one good pair of horns would enable a species of antelope to protect
itself and spread widely. Yet in Africa there are literally dozens of antelope
species with an astonishing variety of horns. There are straight horned
Oryx, the twisted-horned elands, the spiral-horned Kudus, the curve-horned
sable antelopes, the recurve-horned wildebeest and so on. And many of these
antelopes seemingly live in the same environment. What, therefore, are the
advantages of a Kudu's horns over an eland's horns, or vice-versa? These
are questions still to be answered."
Since the time of COLBERT'S(I955) writing, behaviour studies on ungulates have thrown new light on the function of horns. In particular WALi) Acknowledgement: I am most grateful to Dr IAN McTAGGART-COWAN
and Dr
H. D. FISHERfor critical reading of the manuscript.
Behaviour XXVII
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
12
T76
VALERIUS
GEIST
.,,/
Ar
K~~~~~~~~~~~~
f'
Fig. i. Hornis evolved in great variety of form and size. Despite this divTersitythere
are somnemain functionaltypes towards which quite distantly relatcd nmammals
converged.
Such types are (I) (a) Eobasileus and (b) Protoceras. (II) (c) Tetraceros and (d)
Giraffokeryx. (TIII) (e) Arsinotheriumf,(f) Brongtopsand (g) a dicerathere. (TV) (h)
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
EVOLIUTION
OF HORN-LIKE
ORGANS
I77
Cerv7ss elaphus and (k) Tragelaphubs.(V) (i) Muntiacus, (1) Antilocagpraand (m)
Capreolus. The inverse relationshipofEtusks and antlers is illustrated in the series (h),
While Protoceras (b) stood at the beginning of protoceratid
(i) and (j) HIydropo3tes.
evolultion,Synthetoceras (n) was the specialised last genlus from the pliocene.
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
178
VALERIUS GEIST
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
EVOLUTION
OF HORN-LIKE
ORGANS
I 79
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I80
VALERIUS GEIST
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I8I
SOME HISTORICAL
AND
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I82
VALERIUSGEIST
specific fighting. e.g. Aepyceros, Cobus, Redunca. These forms have sexual
dimorphism, or horns in one sex only.
3. Horns which are never used as weapons, but have ceremonial function
only. e.g. Giraffa.
Let us turn to the most commonly stated function of horns, that is, as
weapons. This view is implicit in COLBERT'Sstatement cited earlier and
SIMPSON'S (I949) discussion on horn types of antelopes. There are several
dominant assumptions in these views. Not only that horns function as
weapons, but that they are similarly employed and have been guided in
their evolution by much the same forces. Their diversity is regarded as a
multiple solution to the same problem, which is to ward off enemies and to
dominate over conspecifics (SIMPSON, 1949). Horns are evidently conceived as stabbing weapons that evolved to inflict wounds on adversaries.
Clearly, had this been the evolutionary goal, then "one good pair of horns",
as COLBERTSOfittingly states, would have been all we could have expected.
It is interesting that both COLBERTand SIMPSON think that the country
in which bovids live is somehow relevant to horn evolution. This is understandable if one assumes that horns evolved primarily in response to
carnivores, so that horn bearers in one region evolved similar horns against
similar carnivores. Both COLBERTand SIMPSON noted however that great
diversity of horn types is found among antelopes inhabiting the same region.
The same observation can be made on Eurasian and North American
mountain bovids. No greater difference in horn size and shape could be
conceived than between the Caprinae and Rupicaprinae. Chamois and ibex,
rocky mountain goat and big horn sheep live frequently in the same habitat.
But not the slightest indication of convergence of horns can be detected.
For example, let us take Oreamnos americanus and Ovis dalli. Both are of
similar body size and may be found living side by side, although habitat
optima differ. The combined weight of the horns of large male mountain
goats is between Ioo and 120 grams. The horns from an old ram sheep weigh
about 6,900 grams, approximately 8 %oof his live weight.
The cervids offer rich material for arguments against the influence of
predators on horn evolution. Antlers are found in males only and are
functional for a very short part of the year. Male cervids lose their antlers
in winter at a time when they are harrassed most by predators. The study of
MECH (I963) showed that wolves harrass about I2 moose for every one they
kill. Without antlers the harrassed moose were capable opponents. Most deer
are well known for their ability to defend themselves with their forelegs,
rather than with antlers. Giraffes (BACKHAUS,I96I) also use their hind or
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I83
front legs against enemies. For many small horn-bearing bovids like sheep or
goat, flight is a better protection from predators than fighting. There is no
evidence that wild sheep can fend off an attack by wolves, one of their
major predators. Furthermore, observations on Stone's sheep (Ovis dalli
stonei) and mountain goat (Oreamnos) revealed that these species are but
rarely found in regions of their habitat where a quick escape to precipitous
terrain is not possible. Only in those species which regularly meet the same
predator and do not flee, is there any hope of detecting an influence of
predators on horn evolution. The up-curving horn tip of musk oxen may
be an example. These bovids form a phalanx against sled dogs and presumably wolves, and are reported to hook at sled dogs with their horns
(PEDERSEN, I958). Species such as moose, though devoid of antlers during
most of the winter, may stand and fight harrassing wolves. In these the
rather terrifying defensive threat (GEIST, I963) may be a mechanism
primarily evolved as a predator deterrent. It is noteworthy that nothing
comparable to the defensive threat of moose has been found in any other
ruminant studied. Yet, this very threat type is common among carnivores.
There is little argument that horns can function as weapons and inflict
wounds on opponents. Yet this statement is a far cry from the generalization
that horns evolved as weapons. From the above discussion it appears unlikely
that horns evolved in response to predators, although in a very few cases
predators may have had some slight effect. Interspecific engagements appear
hence unlikely as an impetus to horn evolution. Intraspecific contacts therefore remain as the most probable functional influences.
THE EVOLUTION
OF SMALL HORNS
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
184
VALERIUS
GEIST
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I85
SCHLOETH, I96I;
GEIST,
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
i86
VALERIUS GEIST
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
EVOLUTION
OF HORN-LIKE
ORGANS
I87
phus), a short horned Indian antelope. Although the nilgau is fully capable
of butting, its intense threat display and neck fighting are probably a
means to reduce the primitive fighting form. It is noteworthy that nilgau
bulls face the opponent in a kneeling position while butting, and hereby
reduce the target area of the body which can be reached by the opponent's
horns.
It is postulated therefore that the first stages of horn evolution were
represented by small horns. These were associated with a broadside
intimidation display and with such defense mechanisms as dermal shields,
inhibition to strike, withdrawal or reduction of target area, early or quick
establishment of dominance order, or emphasis on non-damaging forms of
fighting in intraspecific engagements. These are perfectly functional solutions
as the presence of these forms today indicates. There is however no impetus
towards the evolution of larger horns. Large horns in association with
the suggested defense mechanisms have no conceivable advantage over
smaller horns. Hence the evolution towards larger horns is unlikely.
THE EVOLUTION
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
i88
VALERIUS GEIST
whereas Oreamnos roars but rarely. Horning the ground and pawing or
wallowing are part of Bos and Bison agonistic behaviour. Oreamnos does
not wallow, while horning and pawing are infrequently seen in agonistic
interactions. In general, however, the actual displays of these three bovids
have many features in common.
During fighting, the target area for horn blows is identical in these three
genera. MCHUGH (I958) states that buffalos attempt to horn the flank of
the opponent. FULLER (1960) states that buffalo-bulls during fighting
attempt to deliver blows into the opponent's side SCHLOETH i96ia) writes
that in Bos, fighting partners are attempting not to lose frontal contact
until they can push the opponent into a position which allows them to
deliver a horn thrust into the body. McHUGH presents considerable evidence
showing that casualties from fights die of body wounds. These wounds are
located predominantly on the lower part of the body, and resulted from
horn punctures or tears. He recorded bulls with intestines protruding from
punctured bellies; several buffalos died from punctured and collapsed
lungs; some had punctured hind or front legs and were lame. On butchered
bulls McHUGH recorded that 23% of all bulls over four years old had one
to three fractured ribs. GEIST (I964) on the basis of field observations
on Oreamnos and the autopsy of two wounded males, showed that the
ventral body region is the target area for horn blows in fighting mountain
goat. As in buffalos, horn punctures and blows were located on the
abdomen, chest, as well as on the inside and outside of rear and front legs.
Oreamnos fights by swinging the head sideways; and the same type of horn
jab was observed in buffaloes by MCHUGH (1958).
Oreamnos, Bos and Bison hence show many similarities, yet their manner
of fighting differs conspicuously. This difference is clearly understandable, if the head to head engagements of Bos or Bisont serve as a
defense mechanism against attack on the vulnerable body. Whereas the head
evolved structures to absorb the shock of a frontal encounter, the outcurving
horns hold and control the opponent's head and prevent the flank attack.
In the evolution of horns and fighting behaviour, Bos and Bison should
be regarded but one step beyond the Oreamnos stage.
Not only the bovids have made a change from the lateral to the frontal
engagement. The Suidae have done the same. Sus is in many ways
comparable to the Oreamnos stage. Boars fight laterally, catching the
opponent's blow from the tusked snout on the dermal shield covering
shoulders and sides. As in Oreamnos, fights are damaging, as blood may be
found on the ground after some engagements. Boars shot during the rut
frequently show bleeding scars (SNETHLAGE, I957).
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I89
The African giant forest hog (Hylochoerus) and the wart hog (Phacochoerus) are comparable to the Bos-Bison stage. Boars fight by running
head on into each other from some distance. Then follows head to head
pushing, while the opponents attempt to slash each other in the sides with
their lower tusks (MOHR, I960). If successful the lower tusks can cut
through the thick hide of Hylochoerus, and inflict deep wounds (D6NHOFF,
I942, from MOHR,I960). Hylochoerusand Phacochoeruspossess large upper
tusks which protrude laterally and curve up -very much like Bison horns
do. These tusks could not be primarily used for rooting - the popular
functional and evolutionary explanation for the wart hog tusks. The giant
forest hog, with tusks similar in shape to wart hog tusks apparently does
not root. Its nasal disk is covered by hair, which presumably would be worn
off in rooting. EWER (1958) was the first to recognize the large upper tusks
as organs analogous to ruminant horns. As in Bos or Bison, the curved upper
tusks of Hylochoerus and Phacochoerus help to hold and control the
opponent's head preventing the latter from using his sharp lower tusks
effectively. It may be that the spectacular tusks of Babirusa have a similar
function. In Phacochoerus the size of the lateral, dermal shield has apparently been reduced.
The analogy of horn-like organs should be extended to include the tusks
of elephants. KUHME (196I, I963) describes the engagements of captive
Loxodonta bulls. The animals move at each other with raised trunks and,
placing the bases of their trunks together, attempting to push each other
back. The tusks, if long enough, lie parallel along the opponent's head
(Fig. 2). If the weaker of the two opponents turns, it is regularly caught
in its side by the tusks of the stronger (KUHME, I96I). Hence, one of the
functions of the tusks must be to hold the opponent's head, and prevent latter
from bringing his tusks into action. Furthermore, tusks probably allow
effective pushing with the trunk bases, because they do not allow the
opponents' heads to slide past one another.
It is postulated that in the second phase of horn evolution, horns have
the function of a guard, as well as that of a bruising and piercing organ.
They help block an opponent's attack and make it ineffective. However,
a head to head contact of opponents easily leads to pushing and shoving.
This sets the evolutionary impetus toward the next stage of horn evolution.
THE EVOLUTION
WALTHER
OF COMPLEX HORNS
African antelope, proposed that the major function of horns was to guide
the opponents' heads together during the clash, then bind the heads to-
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
[90
VALERIUS GEIST
gether in the following pushing and tugging match. The spirals, bumps and
ridges on the horns of various antelope and gazelles serve to hold the horns
together during the match, allowing the opponents to develop full strength
wrestling engagements. The shape of the antelope horns and mode of
fighting isi such that the horns are quite harmless to the conspecific
opponent. Although to human observers such horns may look immensely
dangerous (i.e. the oryx, WALTHER, I962).
Elaboration of horns from the simple Bos-Bison stage hence served
primarily to make pushing and wrestling matches more effective. There is
little danger to opponents from such fights, hence they may occur quite
frequently. This was noted for the impala (Aepyceros) by DASMANN&
MOSSMAN (I962),
who used elk (Cervus) and male deer (Odocoileus)
as standards for comparison.
I would suggest that WALTHER'Sideas about antelope horn function
apply equally to cervids. That is, cervid antlers should be regarded
primarily as structures binding opponents together during pushing and
wrestling matches. However, they must equally be regarded as guards
against the opponent's attack on his adversary's flank.
There is a surprising concensus of opinion among students of cervid
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
ENVOLUTION OF HORN-LIKE
ORGANS
I9I
biology, that males attempt to gore the body of their opponent. HENNIG
(I962) writes that roe-bucks (Capreolus) attempt to strike the opponent
ventrally, but also to catch the opponent's blow. RIECK (I956) writes that
roe-bucks, engaging from a lateral display, try to grip the opponent
ventrally with their backwardly directed antler points (see Fig. 1 m). He
cites an observation of a roe-buck slashing into an opponent's side when
it fell. KIDDIE (I962) in his extensive observations on Sika deer (Cervus
nippon) states (p. 23): "... the object of the fight appears to be to cripple
the opponent by goring him in the ribs or flanks." Sika stags fight fiercely;
the antlers take the weight of the charge. KIDDIE once shot one of two
fighting stags, whereupon the other gored his fallen opponent in the ribs
and stomach. KIDDIE also observed a fight between a Sika stag and a
small red-deer stag. He states: "The red stag could push the sika stag
where he wished, but was too slow to avoid the charge of the opponent,
which was attacking his ribs and flanks whenever he could. The red stag
seemed to be ushing and attempting to overthrow his opponent, while the
sika stag maneuvered and appeared to be waiting for a chance to deliver
a swift upthrust with the antlers. The sika stag bore severe wounds about
the neck and head, and the red stag was punctured in the stomach, with
some severe slashes on the shoulder that were bleeding badly. Now and then
the red stag attempted to run, but the sika stag was quick enough to thrust his
antlers at his opponent's unguarded flank, causing the red stag to turn and
fight once more." LINKE (1957) in his discussion of antler function in red
deer, also thought of them as organs binding opponents together during fights
and hence allowing a matching of strength, but he did not think this to be
their primary function. He also states that red deer attempts to gore
opponents into their side.
In deer, as in bovids, the body appears to be the primary target of
attack. In forms with elaborate horns however, the head itself becomes the
target in at least some engagements. WALTHER (I962) suggests that with
increased horn complexity went a greater tendency in bovids toward
frontal rather than flank attack. I concur with this view. However, I
suggest that in species with complex horns most attacks are caught by the
opponent with his horns (Fig. 3). Also, among social ruminants many
small engagements occur, which lack the characteristics of serious fights.
In such attacks the horns are indeed aimed at the opponent's head. Ovis
dalli stonei illustrates this well. During fights around the oestrus ewe, rams
forget their ceremonial head-on clashes. The defending dominant ram
attempts to crash into his opponent's body. Whirling around one another
rams in vicious fights hit out sideways and deliver blows on the opponents
Behaviour XXVII
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I3
I92
VALERIUS GEIST
sides, shoulders, haunches, head, virtually any part of the body (Fig. 4).
These observations imply that during serious fights the body remains the
principal target area. The same is implied by BUECHNER'S (I963) observations on the Uganda kob (Adenota), (p. 6i); "'A third male may run in
to break up a fight by butting his horns into the side or hind quarters of
the combatants, sometimes inflicting severe wounds." It is apparent therefore that some bovids retain the orientation toward body attack while
directing most of their aggressive behaviour toward the head. Yet in some
species (i.e. Gazella tompsoni) only the head to head engagement may be
present (WALTHER, in print).
The above discussion suggests that cervid antlers and bovid horns evolved
convergently and function as weapons, secondly as guards and thirdly as
organs binding opponents together during pushing and wrestling contests.
Horns and antlers are elaborate. In the Bovidae however some groups
followed another road of horn evolution. We may term them the "rammers". The Bos-Bison stage serves as an equally convenient jump off point
for these as for the "wrestlers". Horns increased in size, though not complexity, and served not to bind the opponents together, but to deliver a
heavy blow. The sheep and goat, as well as the musk ox (Ovibos), are
excellent examples. But not only a heavy skull need be evolved to absorb
the shock of heavy bodies in head-on collision. MCHUGH(I958) points out
that the cushion of long hair on the head of Bison serves the same function.
He states that a bullet from a 30-o6 rifle fired from 30 feet into the head
of a bull, bounced back and was recovered a twisted piece of lead. In
principle the manner of battle in the "rammers" is the same. Conspecific
opponents clash head-on (Fig. 3). In particular fighting methods differ.
A conceivable evolutionary sequence begins with the Bos-Bison stage, where
opponents do butt heavily head on followed by some pushing. The next
stage is illustrated by Ammotragus or Ovis musimon. Opponents collide
head-on after a rush (BRUHIN, I953). The final stage has been reached by
Ovis canadensis, 0. dalli or 0. ammon poli (WALTHER, I96I), where
opponents rise and charge bipedally, then jump into the clash (Fig. 3).
THE EVOLUTION
OF HORNS AS DISPLAY
GUIDING ORGAN
OR ATTENTION
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
EVOLUTION
otes
hyar
epaied
OF HORN-LIKE
s()ilsrtedthe
remo
()ilutatdth
Fneig.
aneior
5.istnctsome
d.Istnction.gener
Iener ovi
ruiats,hespavy
Ofi ruiathe
ame-rworgns
dispavy
ORGANS
fome.'
ome.A
I93
oralhirrig
osl
ar
ig
are
dithiued
solerdisla
ditriued sover
thelaborgans
body,ans
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I94
VALERIUS
GEIST
Fig. 6. A trend towards cephalisationof display organs is shown among some species of
Suidae. Sus scrofa shows a relatively undifferentiated head. SWuverrucosus displays not
only larger tusks but highly differentiated face warts. Phacochoerus shows the largest
tusks among living species of Suidae and is probablymore differentiated anteriorly than
any other suids (after MOHR, i960).
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I95
parallel an increase in tusk size and in the large facial warts so prominent on
some members (Fig. 6). It can be anticipated that with a change in display
organs goes a change in display behaviour.
It was mentioned that many bovids, giraffids, some cervids and camelids,
whose behaviour has been studied, display laterally. In short horned bovids
like Oreamnos and Boselaphus, the whole of the animal's body is made
conspicuous during display. Let us look at the nilgau (Boselaphus), whose
display has been so meticulously described by WALTHER (I958). The bull
stretches head and neck stiffly forward and up. He appears extremely
tense. The steps are slow and stiff. The ears point back. The tail is depressed, but periodically is jerked up violently. Upon getting close the bull
assumes a full lateral stance to the opponent. To be noted are the colour
patterns on the bull's face, his dark mane, the conspicuous neck beard and
the white rings around his legs, supplemented by the tail flash. Any artist
would recognize on that bull the very principles of attention fixation and
guidance that he himself uses in painting. Small contrasting spots, sharply
projecting features, quick movements, localised jerks, quivers or flashesall these capture the attention of the viewer, and in the nilgau's case should
rotate the attention around the whole of the bull's body. The discovery of
edge receptors in vertebrate retinas (GRUSSER et al, I964), as well as
corticular cells in mammals that respond best to line or edge stimuli, or
moving edges (HUBEL, I963) supply the physiological basis for effectiveness
of attention-guiding adaptations.
If we look at the display in the Bison bull, we see him approaching his
opponent at a slow, stiff walk, while he may swing his head from side to side
(MCHUGH, I958; FULLER, I960). Due to the hair growth pattern on the
bull, most attention should be on the head and shoulder regions.
We recall the suggestion that horns functioned in display. Let us
contrast the display of small horned and large horned cervids. Little Capreolus
displays laterally to an opponent while raising his head very high (RIECK,
1956; HENNIG, I962). Odocoileus hemionus males display laterally to each
other while circling (COWAN & GEIST, I96I). From the lateral position
Capreolus and Odocoileus males suddenly clash and lock antlers. Both genera
are comparatively small-horned, excepting the largest of Odocoileus males,
and lack anterior differentiation as exemplified by Cervus (a mane), Rangifer (a mane) and A Ices (a skin outgrowth below the lower jaw, the bell). Nor
is there any report of the use of their antlers in display behaviour. However,
adult bull moose were on several occasions observed to stretch up the neck,
pull in the chin, tipping the antlers forward, thus displaying the light-colored,
huge palms, and to move slowly and stiffly at the opponent. This posture
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I96
VALERIUS GEIST
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I97
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
198
VALERIUS
GEIST
To point 2.
Rams were classified on the basis of horn size into four categories. Rams
whose horns formed a complete curl or more made up Class IV; rams
with horns less than full curl but with at least 94 curl heads were put into
Class III; rams with less than /4 curl but more than '4 curl made up Class
II; rams with '4 curl or less were placed into Class I (Fig. io). Classes
arranged in order I to IV showed a progressive increase in horn size and age.
An increase in body size was noticeable up to Class, III; a change in facial
features due to increase in thickness of horn bases, up to Class III. Class I
represented rams 17 to 25 months of age, Class II 2Y2 to 5 year old, Class
III 5 to 9 years old, Class IV 7 to 14 years old. Usually rams reached the full
curl stage at age 8 to 9 on the study area. The greatest change from Class I
to IV is shown by the horns. These data refer to Ovis dalli stonei.
ioo%
7.5
25
+3
Fig.
ii.
t2
+1
-1
-2
gradients of rams. Here 0 means same size class; about 70% of all approachesto the
samesize classweresexualapproaches.
+ i and- i meanthe next largersize class,and
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
7,~~~~~~~~~~~~~~~~~~,A
U1.~~~~~~~~~~
aYaS~~~~~~~~~~~~~~~~~~~~
Fig. 3. The main fighting form of the "rammers"is the head on clash, here i
by Ovis caniadensiscatiadensis. This photograph illustrates several other poi
attack was initiated by the smaller ram on the left. The larger is seen in a turn j
to catching the opponent'sblow with his horns. The attack was hence aimed at
ram's body, not head. It is also evident that Ovis canadensis jumps into the cl
hindlegs of the ram on the left are in mid air. (All pictures of sheep taken in M
in December) showing rams in the initial stages of moult).
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
It
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
NE~~~~~~~~~~~~~~~~~~~~~~~~~~~~~7
'~~~~~~~~~~~
~ ~
.
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
:, .:
t
'*
ij(llC(:
i ,:::
W^:
re..
su
::: P
jr ci:I
*:
-+
**.- .X i^ M
,- ;:s
?
...
.,00,.
..
- VJ
'-*
j "."
***'.
.-*.
'j
,L*
,,
-IldCC
1S
LY
**
**14)6
lqC:
'j*:?
*v .
"*
**** ..
*;.r*i.
''":""; I'tiu t" l
3
L;
".r
;"
g'
aE
iJ...
SkCL-
''*iC
*rr,
'*
:"C
*
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
Si,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
.
_--J.
- ''t
~~~~~~~~~~~A
?N?.?
Fig. io. (a) A two,year old ram, Class I, (b) a three year old Class II ram, (c)
7 year old Class III ram, (d) a Class IV ram, well over io years of age. Thes
are of correct relative size. Note, although body size increases up to Class III a
the rna'or-change in the appa-nt characteristicsof rams is brought about by horn
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
EVOLUTION
O1
HORN-LIKE
ORGANS
I99
my studies all three approaches have led to the same result but only the
third method will be discussed here in detail.
Rams do not differentiate in their behaviour between males and females.
Rather, they differentiate between sheep larger or smaller than they are.
All smaller sheep, irrespective of their sex are treated like females, whereas
all larger sheep are confronted with agonistic behaviour. For instance, a
large full curled ram will treat a sub-dominant full curl like an ewe, while the
latter reacts with agonistic patterns. The behaviour performed towards a
smaller sheep will be termed "sexual", although for present purposes this
term serves as no more than a nametag! Preceding an interaction,
a ram already shows during the approach whether it has recognized the size
of its opponent. Towards smaller sheep the initiator adopts usually the low
stretch (Fig. 7). This is frequently followed by a sexual pattern. Towards
larger rams the initiator usually shows a horn-threat, followed by an agonistic
pattern (Fig. 8). Neither sexual nor agonistic treatment will be discussed
here further but will be treated in detail in a later paper. The important
point is the following: Using the first two behaviour patterns shown by a
ram initiating an engagement, we can classify his approach as sexual or
aggressive.
We derive the percent frequencies of sexual approaches shown by each
class towards all others. Next we add the percent frequencies for the following situations: all classes approaching the same class; all classes
approaching the next smallest class, the second smallest, the third smallest
class; all classes approaching the next largest class, the second larger and
third larger class. The frequencies are averaged for each situation. The result
is shown in Fig. ii. Next we graph the difference between the sexual
approach frequencies for one situation and the next following. This is shown
in Fig. 12. The interpretation of this graph is that a ram is most certain of
his actions towards larger or smaller rams, but least certain towards rams of
his own size. Horn size is the major difference between classes and it seems
valid to conclude that rams are indeed differentiating horn size classes.
To point 3.
It must be shown that rams enjoy advantages solely due to their horn size.
Table I shows that smaller horned rams have a greater tendency to withdraw
from larger horned rams. However, these data would not fulfill our requirements. It could be argued that the superior fighting ability of larger rams
accounted for the more frequent withdrawals of the smaller. After all, rams
in constant association with each other do set up a dominance order. Table I
may show no more than the result of the dominance order. Clearly, if horn
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
VALERIUS
200
GEIST
>~50-
S,53020 -
10
I
gXI
+2
+3
+1
-1
-2
-3
2
larger
sizeclasses
smaller
Fig. I2. Abscissa and ordinate same as Fig. i i. Here the difference in the percent
frequency of sexual approaches is graphed. The graph is interpreted as showing that
rams are more uncertain of how to behave towards equal size opponents than towards
smaller or larger rams.
lol
s40-
-e
X30'
z
H
20-
10-
+ 3
+ 2
+ 1
-'1
-2
-3
Fig. I3. Abscissa as in Fig. ii and 12. Ordinate shows the numberof behaviourpatterns
of sexual (black) and agonistic (white) nature receivedby a strange ram entering a band.
It can be seen that he receives most patterns from rams of the same size class, and least
by much larger or much smaller rams.
size carries some advantage, it must show itself before a dominance order
based on actual physical interactions is established. The large horns of full
or three quarter curled rams should derive for their bearer a priori some
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
TABLE
20I
I36
73
134
or
movedaway
No. of withdrawalsin which a ram
movedaway, ratherthan turnedaway
only from his opponent.
30 (22%)
(I3%)
37 (28%)
2I
(57%)
3I (43%)
30 (97%)
in Fig. I3.
The quantitative data tells but a fraction of the story. In fact, it appears
from the graphs that some rams have not entirely recognised the attributes
of the new companion. For instance, recorded in the data is the following:
A strange (full curl) Class IV ram appeared in one of my ram bands under
observation. A Class III ram approached and displayed dominantly (using
sexual or courtship patterns!) to the full curl. This lasted but a few moments,
for the full curl whirled and rose into the bipedal threat display. The Class
III ram ran away. The dominance order was settled without a fight. Yet
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
VALERIUS
202
GEIST
on what basis had the Class III judged his opponent? Both were 7 years old,
and of very nearly the same body size.
Not included in the data are observations on one very large full curl and
a small 4 year old Class II ram that crossed to one of my observation bands.
In the following seven days but one single engagement was observed between
that full curl and any member of the band. That engagement was with a
large Class III ram who acted a priori agonistically rather than sexually,
indicating that he had submitted. None of the large rams from the observation
band ever approached this large stranger. The little 4 year old Class II ram
remained close to the large full curl for several days. The very day he was
observed without his large companion, two 5 year old Class II rams had
engaged him. This engagement was partially filmed.
It must next be shown that large-horned rams are more likely to reproduce
than smaller horned rams.
3.0-
All mounts
.0
2.0-
1.0
IV
d
I
III
Hornsizeclasses
To point 4.
During the sheep rut of I96I and I962, Stone's rams were observed for
46 ram-hours in competition around estrous ewes. The distribution of
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
EVOLUTION
OF HORN-LIKE
ORGANS
203
ft'W
Fig. I5. The evolution of horns in bovids. The primitive horns are short, sharp. The
skull is thin and fragile; engagements are lateral and broadside display prominent
(Oreamnos, GEIST, I964). Next the head to head combat evolved as a defense mechanism.
The skull is heavy, horns curve out to hold opponent.Broadside display prominent (Bos,
SCHLOETH,I96I). From the primitive frontal engagement two combat forms evolved,
the "wrestlers" illustrated by Antelope (WALTHER,I96Ib). Combat form a frontal pushing
204
VALERIUS GEIST
mounting frequencies for various ram-classes are shown in Fig. I4 (a) and
(b). (b) is a record of all mountings, be they interfered with or not. In (b)
mountings are included which could not have led to copulations due to two
factors, one the withdrawal of the ewe, two the interference by the old, full
curled rams. The data shows clearly that the full curls perform most of the
copulations.
Not included in the graphs, but very apparent in the field, is the observation that small rams are able to chase away and steal the ewe only after the
guarding, large ram has copulated frequently and appears exhausted. Or,
the large ram has stopped guarding, feeds and rests, while the next largest
ram commences guarding and copulating with the ewe. This is usually a large
Class III ram.
To point 5.
If large horn size carries evolutionary significance, large horns must
carry dominance status irrespective of the individual's age. It is very readily
apparent in bands of rams that among the full curls for instance there may
be a sizeable number of younger rams. In the study area for Stone's sheep the
largest and most dominant rams were in fact all 7 and 8 year old full curls.
Hence dominance is not entirely age dependent. We saw that horn size
correlated closest with dominance rank. Hence, horns carry dominance
status irrespective of the individual's age.
To point 6.
If the large-horned dominant loses his horns, he suffers a reduction in
dominance also. This proposition could not be investigated. Sheep carry
large, permanent horns and unlike deer do not lose them. In cervids the
above proposition could readily be investigated. Observations by BRUHIN
(I953) on fallow deer (Dama) show that the loss of horns can lead to loss
of dominance. A similar observation was made on moose (GEIST, I963).
COLLIAS (1956) observed that in his herd of goat (Capra hircus), individuals
with horns regularly dominated individuals without horns. Loss of horns in
the Oryx need not lead to a loss in dominance (WALTHER, I958). There is
however, no evidence, direct or circumstantial, that oryx horns have display
function. On the contrary, the indirect evidence points to these horns as
binding organs only.
On the basis of the reported date I conclude that sheep horns assumed
among other functions, the function of display organs. Hence horns have
evolved from a function of weapons only, to assume, in various genera, the
additional functions of guard, binding organ, or display organs (Fig. i6).
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
EVOLUTION
OF HORN-LIKE
ORGANS
205
k.
~ ~
Fig. i6. A male of the Uganda kob (Adenota). In this species no conspicuoushorn size
gradients as in Ovis exists. Males of the Uganda kob are territorial during mating and
show a lech type of breeding (see BUECHNER I96I; I63).
for instance, living in temperate climates. Males live off stored body energy
during the rut. The less energy expended during the rut, the more is left
for the strenuous winter following. I postulate that in species with graded
horn sizes males learn to associate large horns with superior strength during
frequent, minor agonistic engagements outside the rutting season. Males
fight extensively only with equal-size opponents, hence saving energy by not
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
206
VALERIUS GEIST
fighting all males. The saved energy can be used during the following winter.
For sheep there is substantial evidence that increased energy expenditure
during the rut is associated with increased winter-mortality (GEIST, unpublished). Frequent minor agonistic interactions between males prior to the
rut were reported by SCHLOETH (i96ia) for red-deer, and GEIST (I963)
for moose. Stone's or bighorn rams engage daily in minor interactions. The
hypothesis outlined above indicates the evolutionary "raison d'etre" for
graded horn sizes, but it does not explain why other species feature adult
males that differ little from each other.
If not asocial, agonistic animals live in dominance hierarchies. These
establish priority of actions and possession. They allow individuals to live in
a predictable social environment, as each can anticipate specific actions from
dominants and subdominants. They allow increased exercise and selection
for memory. Dominance hierarchies allow stability in groups of animals,
which ensures the least energy expenditure towards social needs, and a
maximum of energy devoted to growth, reproduction and individual wellbeing.
Dominance hierarchies can exist in groups of individuals that differ
little from each other. They depend on the knowledge of each individual
about the agonistic potentials of his companions. The society of most
primates appears to be of this type. Among ruminants cattle may dwell
in such societies. Both SCHLOETH (I96I) and MCHUGH (I958) report groups
of individuals uniting against one companion - a good indication of closed
societies. These individuals live in very predictable social environments as
long as the troupe remains together and recruitment is from the inside. There
is however, no predictability of social relationships if a number of strange,
similar individuals are united. Such predictability is preserved however, if
some sign system exists which parallels fighting potential, and hence dominance order. In sheep (and probably large cervids) horn size parallels
dominance order. Hence predictability of social relationships is preserved
even if strange individuals meet. In species with a rank system based on
horn or body size one can expect open societies, in which troupes periodically
assemble from strangers. This is what happens to bands of Stone's sheep.
However, rank systems can be based on criteria other than horn or body size.
Recent studies on African antelope and gazelles established the existence
of territoriality in Alcelaphus (BACKHAUS, I958), Adenota (BUECHNER,
I96I), Aepyceros (DASMANN & MOSSMANN, I962), Gazella thomsoni
(WALTHER, in print), and temporary or periodically shifting territories in
Gorgon (TALBOT & TALBOT, I963). In all these species adult males are
very similar in appearance. The same is true of the territorial South
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
207
American camelid, the vicunia (KOFORD, I957). In all these social, territorial artiodactyls, territory holders make up but part of the total male
population, while non-territorial males form associations on communal
areas.
Territory-guarding males are very conspicuous on their territorial grounds.
Gorgon bulls assume a very erect head posture and a rocking gait while
circling the female herds. Non-territorial bulls do not assume these striking
behaviour patterns (TALBOT & TALBOT, I963). Adenota males show
ritualised display towards approaching rivals (BUECHNER,I963). Gazella
thomsoni marks its territorial borders extensively with glandular secretion
(WALTHER,in print). The status of territory ownership is hence signalled
to conspecifics. But territory ownership is associated with a high dominance
rank. Hence, in these species, where males are similar in appearance, territory, rather than horn or body size signals a predictable agonistic
potential.
The extensive olfactory marking observable in territorial forms should
be thought of as an olfactory intimidation mechanism, as was independently
suggested by WALTHER(I96I), and GEIST (I964). This mechanism would
still be functioning in the absence of the owner. Olfactory, visual and
auditory signs produce to the conspecific an unmistakable situation, which of
course, allows predictable social relationships. Without this, an animal would
not live in a predictable environment, which after all, its adaptations
anticipate. It is suggested that territoriality in ruminants evolved as a
mechanism ensuring predictability of social relationships. Territoriality and
horn size gradients should be thought of as compensating. Both are rank
systems which parallel fighting potential and hence dominance order. Sexual
as well as age dimorphisms may equally function as rank systems. One of
the main significances of rank indicators is probably that they allow the
existence of "open"societies, where strange conspecifics can integrate with
a minimum of combat, damage and energy loss.
Yet why should there be displays and display organs, dominance hierarchies and rank systems? What evolutionary reason perpetuates these
universal phenomena?
The mechanisms in question protect and retain two utterly essential
attributes of life - self assertion and aggression. By the former I mean the
individual's energetic and egoistic attempts to satisfy its appetite relating
to food, reproduction and habitat selection. Aggression arises from self
assertion when one individual attempts to displace another during conflict
of mutual interests. Aggression is, adaptive. It ensures the more aggressive
having priority in satisfying its basic appetite; dominance hierarchies
Behaviour XXVII
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
I4
208
VALERIUS
GEIST
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
209
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
210
\TALERIUS
GEIST
the species genes. As the genome of the species changes so does its hornshape since different genes would indirectly influence the horngrowth
characteristics differently. Individuals of any one species would carry similar
hornshapes, since they would have similar gene compliments. This view
explains hornshape, but it is all but untestable.
HUXLEY'S(193I) view of heterogenic growth serves to explain relative
hornsize differences between small and large forms. Unfortunately, this
explanation is faced by many exceptions. The interested may for instance
compare the horn and bodysizes among the various genera of the closely
related Caprinae and Rupicaprinae or ponder the hornsize differences between
Cervus unicolor and Cervus canadensis, or between Bibos and Bubalus,
or between Odocoileus and Hippocamelus. If in a species evolutionary
history hornsize increased with bodysize according to the laws of heterogonic growth, it appears to me more likely due to the adaptive value of
large horns than to the accident of heterogonic growth.
SUMMARY
Hornlike organs evolved independentlyin a number of mammalian families. Though
these organs assumed great diversity they did evolve into several general functional
types. A short review of the structure and development of hornlike organs is given.
Some views on horn function and evolution are critically discussed.
The evolution of hornlike organs is visualised as follows:
In primitive large mammals the head blow became effective as a fighting form due
to increased mass and inertia of the heads. Some forms grasped this potential. Combats were carried out from the broadside while opponents delivered head blows on
each others body. Skull protuberancesnow became adaptive. Concurrently, defensive
mechanisms evolve, decreasing the effectiveness of these protuberances. Foremost
among them is a thick, heavy hide or specialised dermal shield. These adaptive syndromes gave no impetus towards larger and more complex horns.
This impetus arose with the appearanceof a new method of defense -catching the
opponent's blows with the horned head. This leads to the evolution of heavy skulls
and horns capable of catching and holding the opponent's head. The target area of
attack remained the body. Frontal engagements resulted from the opponents' attempts
to control each others horned head.
It is shown that bovids and suids followed similar evolutionary roads in their mode
of combat.The tusks of the suidae fulfill the same functionand were subjectto
similar selection as the horns of short horned bovids. Thus Sus and Oreamnos,and
Bos and Phacochoerusare entirely similar in their mode of combat, hornlike organs
and defense mechanisms.
The primitive frontal engagement gave rise to two different modes of combat,
ramming and wrestling. The "wrestlers" evolved complex horns, whose function is to
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
EVOLUTION
OF HORN-LIKE
2II
ORGANS
American wild sheep. Display centers about one heavy, very large horns of males.
Next it is verified that the rams horns are display organs. It is shown, the horn
display is the prerogative of the dominant; sheep can differentiate horn size classes;
large horned males enjoy reproductiveadvantage; horn size carries a priori dominance
status.
Hence horns evolved to function as weapons inflicting damage; as defense organs
shielding their owner; as binding organs allowing opponents a secure lock in battle;
as display organs having an a priori intimidating effect on certain conspecifics.
Some species of ruminants possess hornlike organs which change progressively in
size with the age of their owner. It is postulated that, as in North American wild
sheep, horn size parallels dominance order. Conspecifics learn to associate lhorn size
and agonistic potential of opponents. Hence horn size signals to conspecifics predictable
social relationships. In large cervids from temperate and cold climates this system
must save stored energy during the rut, which males need for the severe winter.
Graded horn sizes allow wild sheep to live in an "open" society in which strange
individualsmay meet and fit into the dominancehierarchy with a minimum of combat.
Graded horn sizes hence serve as rank indicators. It is next made plausible that
territories in ruminantsfunction in conjunctionwith displaying males as rank indicators.
Outside the territories males mix freely. Hence graded horn sizes and territoriality may
enable "open" societies, by acting as communicationmechanisms and creating a predictable, social environment.
A hypothesis is presented explaining the selective forces for the evolution of display.
It predicts that, antagonistic selection forces shape display and display-organs, that a
potential exists for rapid spread in a population of differences in behaviour and
morphology, that intimidation mechanisms evolved for all sensory pathways, that external appearanceaf animals changes more repidly than their primary adaptations.
LITERATURE
BACKHAUS,D. (1959). Beobachtungen uiber das Freileben von Lewel-Kuhantilopen
(Alcelaphus buselaphuslelwel) und Gelegenheitsbeobachtungenan Senner Pferdeantilopen (Hippotragus equinus bakeri). - Z. Saugetierkunde24, p. I-34.
(I96I). Beobachtungen an Giraffen in Zoologischen Garten und freier Wildbahn.
Ecology 37,
P. 228-239.
COWAN,I. McT. & GEIST,V. (i96i). Aggressive behavior in deer of the genus
Odocoileus.- J. Mammal. 42, P. 522-526.
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
VALERIUS
2I2
GEIST
Press, London.
of Impala in Southern
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
EVOLUTION
OF HORN-LIKE
ORGANS
2I3
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions
2I4
ORGANS
This content downloaded from 137.149.200.5 on Thu, 09 Apr 2015 09:57:34 UTC
All use subject to JSTOR Terms and Conditions