UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN BIOLOGY 1 -; r.-v < I FIELDIANA Zoology Published by Field Museum of Natural History /J^-^-l VOLUME 61, NUMBER 1 The Co-evolution of Social Behavior AND Cranial Morphology in Sheep and Goats (BOVIDAE, CAPRINI) WILLIA3M M. SCHAFFER AMD CHARLES A. REED SHKUBftA^KQejHe ^{«'V£RSITY0^ MAY 12, 1972 m 18l97i filOLOSy LIBRARY ^Oi 8URRIU HMM FIELDIANA Zoology Published by Field Museum of Natural History VOLUME 61, NUMBER 1 The Co-evolution of Social Behavior AND Cranial Morphology in Sheep and Goats (BOVIDAE, CaPRINI) WILLIAM M. SCHAFFER Department of Biology, Princeton University AND CHARLES A. REED Department of Anthropology University of Illinois at Chicago Circle and Field Museum of Natural History Chicago MAY 12, 1972 PUBLICATION 1146 \ Patricia M. Williams Managing Editor, Scientific Publications Library of Congress Catalog Card Number: 75-179168 PRINTED IN THE UNITED STATES OF AMERICA BY FIELD MUSEUM PRESS TABLE OF CONTENTS PAGE Introduction 1 Acknowledgements 1 General considerations 3 Behavior 11 Comparative cranial anatomy 16 Horns 16 Shape of the skull 22 Sinuses 26 Discussion 31 Growth, delayed reproduction, and natural selection 31 Prevention of brain damage 46 Fighting style and the shape of the horns 50 Summary and conclusions 56 Addendum 58 References 59 Plates 63 INTRODUCTION The present study was suggested by the experiences of one of us (Reed) while sorting and identifying broken bones of numerous ani- mals excavated from Neolithic sites in southwestern Asia. Identifi- cation of parts of skulls of small bovids, all presumably sheep and goats, was particularly difficult, a problem which led to our bisecting the skulls of wild and domestic sheep (Ovis) and goats (Capra) for detailed comparisons. The differences observed in cranial shapes and proportions suggested basic differences in the uses of horns and skulls, and thus of general behavior. Preliminary investigation of all genera of the bovid tribe Caprini (goats and sheep) showed the validity of the general assumption (Reed and Schaffer, 1966), and the present more detailed analysis of the correlated evolution of cephalic mor- phology and use of the horns and head followed. Acknowledgements This study was supported by NSF Grant no. GE 6141 to under- graduates in the Department of Biology at Yale University, and in part by pre-doctoral fellowships in ecology from Princeton University and the National Science Foundation. Most of the skulls studied were obtained from the Yale Peabody Museum and the American Museum of Natural History. We are especially grateful to Dr. Richard Van Gelder of the American Museum for permission to sec- tion some of his skulls. Additional observations were made on speci- mens at Field Museum of Natural History and the Museum of Comparative Zoology, Harvard University. The help of John Howard, formerly staff photographer at the Peabody Museum, is gratefully acknowledged. Discussions with Dr. J. W. Justusson of the Department of Applied Sciences at Yale University, and with Dr. Henry Horn of Princeton University were of the utmost value in formulating many of the ideas in a meaningful sense. The encourage- ment given by wives, parents, and friends was also essential to the completion of the project. Finally, it is a pleasure to acknowledge the inspiration provided by Dr. G. Evelyn Hutchinson, whom we were both privileged to know while at Yale. GENERAL CONSIDERATIONS The tribe Caprini is composed of five genera: Capra, the goats and ibexes; Hemitragus, the tahrs; Ammotragus, the ami, aoudad, or barbary sheep ; Pseudois, the bharal or blue sheep ; and Ovis, the true sheep (mouflons, argahs, and bighorn). Table 1 lists the taxa con- sidered in this study of Caprini and related Rupicaprini along with common names and regions of occurrence. The plates include pho- tographs of the intact skulls of representative species. The tribe Caprini is essentially Old World in its distribution and probably took its origin in Asia (Sushkin, 1925; McCann, 1956) from a group not unlike the so-called goat-antelopes (the rupicaprines Nemorhaedus and Capricornis) that live in this part of the world today. All of these animals are upland forms dwelling in mountainous areas, al- though some Ovis live on rolling hills and adjacent plains. As in other upland species, their fossil record is meager and their evolu- tionary history poorly understood (Simpson, 1945). Presumably, the sheep-goat line diverged from the true antelope during the middle Miocene (Pilgrim, 1939, 1947), but the details of this branching are left to conjecture. Antelopes whose horns are somewhat sheep-like have been found in the upper Miocene of Mongolia and in the Euro- pean Pliocene (Pilgrim, 1934, 1939), but these are undoubtedly far from the main-line of Caprine evolution. Essentially modern sheep first appear in the upper Pliocene along with Sivacapra, a primitive relative of Hemitragus (Lydekker, 1898; Pilgrim, 1939, 1947), Hemi- tragus, Capra, and Ovis have all been found in Pleistocene deposits in Europe (Kurt^n, 1968), and Hemitragus is also known from the Siwalik fauna of India (Colbert, 1935). There are numerous features which distinguish the Caprini from their less specialized allies, the Rupicaprini (Reed and Schaffer, 1966). These characters include the following: 1. Increased size of the horns (both absolute and relative to the size of the skull); 2. Hypertrophy of the frontal bones and a concomitant relative re- duction of the parietals (see fig. 9 for bones of skull) ; 3. Heightening of the skull as measured from the plane of the vertex to the foramen magnum; 4. Expansion of the frontal and cornual sinuses and the development of bony septa within these; 5. Bending of the basicra- Table 1. — Populations of Caprini and Rupicaprini discussed in this paper. CAPRINI Genus Species Common name Capra la. hircus aegagrus Pisang, Bezoar lb. hircus hircus Domestic goat 2. falconeri Markhor 3a. ibex caucasica Caucasian tur 3b. ibex sibirica Siberian ibex 3c. ibex nubiana Nubian ibex Hemitragus 1. jayakari 2. hylocrius 3. jemlahicus Ammotragus 1. lervia Pseudois 1. nayaur Ovis 1. canadensis 2. ammon 3. musimon 4. aries Arabian tahr Nilgiri tahr Himalayan tahr Barbary Sheep, Arui, Aoudad Blue Sheep, Bharal American Big- horn, Bighorn Argali Mouflon Domestic sheep Geographical distribution Aegean Islands and moun- tains of southwestern Asia excluding Syria, the Levant, and all of Arabia except Oman World-wide (introduced) Afghanistan and western Pakistan, and adjacent mountainous areas Caucasus Mountains Mountains of central Asia from Himalayas north into Siberia Ethiopia, the Red Sea Hills of Sudan and Egypt, Sinai, and some parts of western Arabia Oman (southeastern Arabia) Nilgiri Hills and adjacent ranges, southern India Himalaya Mts., from Kash- mir to Sikkim; introduced, New Zealand Hills and mountains of northern Africa, from Rio de Oro to Egypt and the Sudan Mountains of central Asia, from Himalayas to Inner Mongolia Western North America, eastern Siberia Central Asia, from northern Himalayas to Mongolia and Siberia and west into Russian Turkestan Sardinia and Corsica; introduced in Europe Introduced world-wide I SCHAFFER & REED: CAPRINI Table 1. — Populations of Caprini and Rupicaprini discussed in this paper (Cont.) RUPICAPRINI Genus Species Common name Geographical distribution Rupicapra 1. rupicapra Chamois European mountains from Spain to Caucases and eastern Asia Minor Nemorhaedtis 1. goral Goral Southeastern Siberia, Man- churia, China, Korea, Tibet, Burma, Assam, Nepal, Punjab to Kashmir Capricornis 1. sumatrensis Serow Kansu and southern China, and west to Burma, Nepal and Punjab; south into Indo- China, Malaya, Sumatra Oreamnos 1. americanus Rocky Moun- tain Goat Northwestern North America For a full survey of the species and subspecies belonging to the Caprini and Rupicaprini see Ellerman and Morrison-Scott, 1951. nium in the region of basioccipital-basisphenoid fusion with the re- sulting ventrad rotation of the posterior elements of the braincase (fig. 11). In the acquisition of these characters, Capra has lagged, whereas Ammotragus, Pseudois, Hemitragus jemlahicus, and larger sheep display them to the extreme. Since this study is to a large extent concerned with the complex of cranial characters associated with the horns, it is helpful to con- sider these structures in a general sense before proceeding to the indi- vidual genera and species. The bovid horn originates from the frontal bone and consists of a bony core surmounted by a corneous sheath. The horn core may or may not contain an air-filled sinus, but among the Caprini and their allies well-delineated sinuses exist in both frontal bones and horn cores. The sinuses of each side connect with each other and ultimately with the nasal passage and hence with the outside air. The basic pattern is illustrated in Figure 1. The figure shows a typical Rupicaprine, Oreamnos americanus, in which the outer table of the frontal bones has been removed. The skull of a Nubian ibex, Capra ibex nuhiana was prepared in a similar fashion (fig. 2). From the figures, it is apparent that each frontal bone contains its own system of air cavities which are separated from those of the other frontal by a mid-sagittal septum. Each frontal bone is thus divided into dorsal and ventral tables, separated by the inter- polation of the sinuses. Medially, the two tables are connected by the mid-sagittal septum; laterally, they come together to form the dorsal rim of the orbit. Sections of the skull in the sagittal plane FIELDIANA: ZOOLOGY, VOLUME 61 OR Fig. 1. Oreamnos americamis (Rupicaprini). The dorsal layer of the frontal bone and the frontal surface of the right horn core have been removed. The left horn core has been removed entirely. C — cornual sinus; HC — horn core; L — lat- eral branch of the frontal sinus; M — medial branch of the frontal sinus; OR — orbit. (fig. 3) reveal the extent of the sinuses and the separation of the parts of the frontal bone. Each frontal sinus (left and right) is further subdivided into lateral and medial compartments by a septum that extends from the supraorbital canal posteriad. Anterior to the canal, the two compartments are united and the single sinus extends to the anterior edge of the frontal bone. The sinus of the horn core con- nects with the lateral compartment of the frontal sinus. In Rupi- caprini, the sinuses are small and uncomplicated by the presence of additional internal septa; the cornual sinus is confined to the base of the horn core (fig. 1). In Caprini, the sinuses are relatively more extensive, and in many species there are numerous septa, mostly dorso- ventral, which have complex inter-relationships (see photo- graphs, especially IX, XXIV, and XXVII); in addition, the lateral compartments of the frontal sinuses extend mediad behind the Fig. 2. Capra ibex nubiana (Caprini). Prepared in the same way as Figure 1. F — frontal surface of the horn core; LT — lateral surface of the horn core; MD — medial surface of the horn core; N — nasal bone; NU — nuchal surface of the horn core; P — parietal bones; PM — premaxillary bone; S — supraoccipital bone. Other letters as in Figure 1. o M TJ a ai S 0) OJ o > t, OJ 3 2 — • 03 C OS Oi OS OS Oi Ph Ol OS o^ o> a> 05 05 ""J OS 05 Oi C^ o t f tr- ,_, « o "S £ n ee •^ CO CO CO 1-H C Q) O xn 1—1 (M 1 X O CO CO in O CO .s|l I-H rH T— ( (M CO ^ "Z, lO lO lO iCi in - in lO kO in ■^ T ta to p ^ s ° 3 CL, eo o 00 in as in OS ■* Ii|e f «o 00 m N •— ( (N in 00 •^ 1-H 2i S >» c E S.ti 1 S 3 gcg liio - 2J J3 < (£> "^ (M CO . Z t— 1 in lO lO (N in 1-H in 1 X J3 *^ 'I' +^ 3 1-H •• bo t"* . CO C<3 CO CO CO CO in in 1-H 3 c« +^ <1> CO D. - W Ki Tl< U3 «> 05 CO «< c E Si S o . £ 8 2 ti 1 X •^ t> CO CO t> 'S' t-H lill =" : "53 12; t> t> t- ■>* ■ S-. -u OJ '^ jyi .^ +3 3 to 3 -Q •-' rt (= C tn JJ ■4^ fl r? •> >i CO i 1 O J3 s hJ H^ 1J > fe PlH o o o •o S^ Fig. 3. Parasagittal sections (immediately to one side of the midsagittal sep- tum) of several species. The sinuses are shown in black, a. Nesotragus moschatus (a primitive bovid without sinuses) ; b. cfCapra hircus hircus; c. d'Pseudois nayaur; d. (f Hemitragus jemlahicus; e. Ovis canadensis — top, male; bottom, female; f. Am- motragus lervia — top, male; bottom female. 10 FIELDIANA: ZOOLOGY, VOLUME 61 medial compartments (fig. 2, Capra) . This complexity is associated with an expansion of the bases of the horn cores. The cornual si- nuses may also be more extensive and in some species extend to the very tips of the horn cores. Sexual dimorphism with regard to the development of horns, sinuses, and associated structures is apparent in all Caprini and is considerably more pronounced than in rupicaprines. As an example, a series of male and female skulls of Nemorhaedus goral were com- pared for six cranial characters (table 2) . Only two of the measure- ments revealed significant differences between the sexes (P (t) >0.95). A similar comparison in a series of Ammotragus lervia revealed sig- nificant differences in five of the six measurements. BEHAVIOR All Caprini live in bands or herds and it would appear that this social situation has provided the basis upon which more complex behavioral relationships have been built. In all species studied, the males engage in some sort of intraspecific fighting, which typically involves combat with the horns. Such competition is an ancient trait in vertebrates, and is to be found in bony fish and a variety of tetrapods. Among the more specialized ruminants, the head, antlers, and horns have come to be used both as weapons and as parts of a stylized threat pattern (Geist, 1966b). Giraffes engage in contests in which long, swinging blows are delivered with the head (Innes, 1958) ; deer lock antlers and push and shove (Darling, 1937) ; musk- oxen charge and ram as sheep do (Teal, 1970); while the various antelope fight in a variety of ways (Walther, 1958, 1964; Estes, 1967). But whereas antelope fight primarily by locking horns and "wrestling" with their necks, the tendency to deliver sharp blows with the head and horns is characteristic of the Ovibovini, Rupi- caprini, and the Caprini. Among Rupicaprini, the most vigorous form of combat appears to be body-butting, blows delivered with the horns to an opponent's flanks. Such behavior has been reported in Oreamnos (Geist, 1964), and Rupicapra (Couturier, 1938) and was undoubtedly antecedent to the more forceful head-to-head butting observed in male Caprini. Among the latter, although behavioral data are almost totally lacking for Hemitragus and Pseudois,^ an interesting story unfolds. The butting behavior of two young adult Caucasian turs {Capra ihex caucasica) was observed by one of us (Reed) at the Catskill Game Farm in New York. These males butted vigorously, horn to horn, and sometimes horn to forehead. Typically, one stood erect on his hind legs, took two or three steps forward, and then with neck down-arched crashed down upon his opponent. The male receiving 1 Fragmentary descriptions of the fighting behavior of Hemitragus hylocrius (Hutton, 1947) and Pseudois nayaur (Lydekker, 1898) suggest that the former may engage in a type of butting behavior similar to that of Ammotragus, whereas the latter may ram more like true sheep. 11 12 FIELDIANA: ZOOLOGY, VOLUME 61 the blow remained quadrupedal, moving his head to insure that con- tact was made properly, with horns crossed as typical for Capra. As in many such combats reported for other caprines, the whole thing appeared to be a cooperative affair which has led Reed to com- pare such "battles" with the joustings of medieval knights. In this instance, the male that initiated the contests was subdominant at the feed trough. Despite his seeming aggressiveness, he slowly withdrew up a precipitous slope. Walther (1961) has observed the same sort of behavior in other Caucasian turs and in the markhor, C. falconeri. He reports that in his experience both animals stand bipedally and stresses that the opponents are usually not oriented exactly head to head. As a result, the horns are crossed at the moment of impact. Similar observations have been reported for C ibex ibex (Couturier, 1962), for C. ibex sibirica (in very anecdotal fashion by Grzimek, 1966), and for C. hircus hircus (CoUias, 1956; Haefez and Scott, 1962; Scott, 1960). In the case of Ammotragus, a different sort of fighting has been reported (Katz, 1949; Haas, 1959). Typically, the contestants ap- proach each other quadrupedally with heads lowered in the custo- mary threat posture. When they are almost together, they bring the heads down even further, thus directing the bases of the horns forward; this necessitates taking a few steps backward to make room for such a maneuver. Immediately after, they run toward each other and collide, the blow being delivered with the bases of the horns. Katz indicates that impact is horn to horn and that the heads are slightly tilted so that the horns are partly crossed. On the other hand, Haas reports that most of the blows are of the horn-to-forehead variety, so that the horns interdigitate. (Such interdigitation was observed by Reed about 10 per cent of the time in the Caucasian turs he observed.) No doubt both patterns occur. In addition, Ammo- tragus indulge in other forms of combat which include head-to-head pushing, horn hooking, and wrestling with the neck and horns. Combative behavior in Ovis has been most thoroughly studied in the North American form 0. canadensis in the Death Valley National Monument by Welles and Welles (1961). Their descriptions are mostly in agreement with other accounts (Murie, 1944; Geist, 1966a, 1969), but present more details. In the most spectacular form of combat, the opponents walk away from each other, whirl simultane- ously, and rush toward each other at high speed. During the charge, which may cover as much as 30 ft., the animals run bipedally; the Welles estimate the closing velocity to range between 50 and 70 miles SCHAFFER & REED: CAPRINI 13 per hour. Within the course of a few hours, a single pair may engage in several dozen such charges and rammings and show no signs of injury. Welles and Welles stress that the contesting males co-operate to insure exact horn-to-horn contact, although interdigitation (Geist, 1969) and other miscalculations are known to occur. Normally, a ram will not attack if his opponent is off balance or otherwise unpre- pared for battle. Indeed, the rare injuries that have been observed appear to result from the contestants having misjudged their ap- proach. Walther (1961) has observed similar fighting behavior in 0. ammon poll. Domestic rams charge at a lesser velocity and re- main quadrupedal (Scott, 1960; Grubb, personal communication; Schaffer, personal observation). The conclusion emerges from the foregoing discussion that the force of impact increases from Capra to Ammotragus to Ovis (Schaf- fer, 1968). Furthermore, as Geist (1966b) has indicated, strict "ramming" has been observed only in Ammotragus and Ovis. How- ever, the heaviness of the bone and the complexity of the internal struts of the horn core (pi. XIX) and skull roof of Pseudois nayaur, coupled with observable damage to the horns of at least one specimen (pi. XVII), indicates to us that males of this population also use the horns in ramming, and in this aspect of their behavior are probably more like sheep than like goats. In the male of Hemitragus, the horn core (fig. 16c) and skull roof (fig. 3d) are also heavily braced internally, more so than in goats and ibexes, and we assume that the horns and skulls have evolved by natural selection, as in other Caprini, in correlation with the stresses to which the structures have been subjected. However, for the males of Hemitragus, no such head-to-head or horn-to-horn combats have been witnessed by Caughley {personal communication), who has observed for several years the Himalayan tahrs introduced into New Zealand. Even so, the fact that the total cranial structure of this particular species of tahr, H.jemlahicus, is of the configuration we have called "advanced" (p. 23), coupled with the other characters of horn cores and skull roofs mentioned above, leads us to believe that the males of Hemi- tragus also use their heads and horns in intraspecific combat, for the establishment of individual dominance, quite as other Caprini do. Only further study can substantiate or invalidate our assumption. In contrast to the "ramming" of Ovis, and presumably of Pseudois and Ammotragus, the combats of Capra might better be described as "clashing," since there is no charge. Actually, in comparison to human activities, the male combats of the Caprini can best be com- 14 FIELDIANA: ZOOLOGY, VOLUME 61 pared to jousting, in which opponents of nearly equal skill, weapons, and ability at defense compete in patterns of behavior common to both, and understood and expected by both. The contests are no less serious for all of that. A final point, first observed by Geist (1968a) in bighorn sheep, concerns the age of first entry into the rut. Whereas both females and males are sexually mature by the end of the second year, males do not actually participate in rutting activity until their sixth, and often not until their eighth year. A shorter delay in the onset of re- productive activity has also been observed in Hemitragus jemlahicus imported into New Zealand (Caughley, personal communication). In this species, the males do not breed until the fourth or fifth year, although they become mature by the third year. Unfortunately, the Caprini have not been sufficiently studied to assert that this is a general characteristic of the group. Delayed male reproductive activity has been observed in other species in which dominance hier- archies exist among males (Sadlier, 1968), and we suspect that this situation is also true for most Caprini. We should also point out that harem formation, as observed in pinnipeds and deer, is not a necessary prerequisite either for the development of delayed reproductive behavior or for the evolution of sexual dimorphism with regard to the structure of the horns, skull, etc. (see Discussion). We shall, in the following sections, consider in greater detail the cranial anatomy of the Caprini and the relation it bears to the be- havioral patterns just described. Materials and Methods: The following measurements were taken on the skulls and horns of more than 100 skulls of caprines and rupicaprines: I. On the skull: 1. FL — Length of the frontal bones, as measured along the curve of the roof of the skull at the midline from the fronto-nasal suture to the fronto-parietal suture. 2. PL — Length of the parietal bones, as measured along the curve of the roof of the skull at the midline from the fronto-parietal suture to the parieto-occipital suture. 3. VL — Ventral length of the skull, as measured directly between the anterior edges of the maxillaries and the anterior edge of the foramen magnum. SCHAFFER & REED: CAPRINI 15 II. On the horns: 1. CHC — The basal circumference of the horn core. 2. FN — The fronto-nuchal (antero-posterior) diameter of the horn core,^ measured at the base. 3. LM — The latero-medial diameter of the horn core,^ measured at the base. 4. OLHC — The length of the horn core, along its frontal curve. 5. OLHSh — The length of the horn sheath, measured along the frontal surface. 6. RAD — The radius of the horn core, the radius of the hypo- thetical circle made by the first complete whorl of the horn, if the horn were to grow to that extent. In addition, permission was obtained from the American and Peabody Museums to section a total of 17 skulls. Sagittal sections of the horn cores and parasagittal sections of the skulls, barely to the right or left of the mid-sagittal septum, were prepared. On some specimens, the roof of the skull was also removed. Two angles were measured on the sectioned skulls (fig. 11) and the radius of the horn core (RAD) was determined, measured from the approximate center of the curvature to the midline of the sectioned horn cores. See Figure 1 for horn-surfaces. COMPARATIVE CRANIAL ANATOMY To facilitate consideration of the comparative cranial morphol- ogy, we have divided the material into three sections — horns, skull shape, and sinuses — to be followed by a discussion of the adaptive significances of some of the major features. Although no attempt has been made to provide detailed descriptions of the various species, we believe that we have indicated the main trends along with the obvi- ous exceptions. The reader is again referred to the photographs of the intact skulls. Horns Measurements are given in Table 3 for the horns of 11 species of Caprini and for Nemorhaedus goral and Nesotragus moschatus, the latter an extremely primitive antelope from East Africa. In all Caprini, the horns of the males are longer and larger in cross-section than are those of females of the same species. Primi- tively, the horns were nearly straight, quite short, circular in cross- section and had their origin from the frontal bones slightly anterior to the posterior rim of the orbit. In addition, they did not rise above the facial plane. Thus, if a straight line were drawn along the top of the skull from the anterior edges of the nasal bones through the vertex of the skull and continued beyond, no part of the horn cores would lie above it. All of these characters are to be observed in the rupica- prine, Nemorhaedus goral. Among Caprini, however, most of these primitive characters have been modified. In general, the horn core is longer, more greatly curved along its main (proximo-distal) axis, elliptical — as opposed to circular — in cross-section, and in all but two species {Hemitragus jemlahicus and Ammotragus lervia), angled up- ward out of the facial plane. The radius of curvature is greatest in Capra (26-36.6 cm.), and this fact, and the extreme length of the horns and their upward angulation, is correlated with the displace- ment of the horn tips behind the skull. As in other Caprini, the horns (with rare exceptions) diverge from each other laterally, so that each describes a spiral in three dimensions (Thompson, 1945). In most caprines, only part of the first whorl is actually formed; this situation 16 (M CO CO C<1 ooo CO ^ LO oico -^ t-Ci ^ CO OOi ;d CO OS t-H Q <3 fl « Ih a 03 o -n c c3 ^ ,^ C/J CI w ^ J o o a 3 p:; OOOi CO to CO -^co>J^t-o 00,-it-oo T-ioo 00 CO Oi I CD t- T-IOO -* 00 «D cooocot- co'-O 00 t> c^iasai m-^i— 1-^ CDt> »0 o COt-fNl 1—1 1—1 (>J 1—1 1—1 1—1 C<1 CO rH (M CO "^ { plH I IC '^ CO I ^ !M t> 00 t> lO (M 1— I U3 1—1 CO CO -* j ^ ^ S S C 2 S S g CO cc s =2 S^ S w -^ s m cj ^l^oof^ odoo !i:t:rj -;: ^ cSdd JJ^ o :3 0) C m OS O +J h -^ (1) ' — ! g-* < OS £H fe w 17 18 FIELDIANA: ZOOLOGY, VOLUME 61 0 OLHC LM OLHC CHC _J_ y ■ H O Fig. 4. Ratios of length of horn core (OLHC) to width of horn core (LM) and circumference of horn core (CHC) for the males of five genera of Caprini. C — Capra; A — Ammotragus; H — Hemitragus; P — Pseudois; 0 — Oris. results either from a low spiral angle (large radius), Capra and Pseud- ois; short horns, Hemitragus; or a combination of a moderately high spiral angle and moderately long horns, Ammotragus. Among spe- cies of the genus Ovis, however, long horns and high spiral angles produce the obviously coiled structure typical of the group. In some older rams, the horns may complete more than 13^ turns. In most Caprini, the horn bases are located above the posterior half of the orbit. In Ammotragus lervia and Hemitragus jemlahicus, however, the bases have been displaced behind the eye socket. The horn of Pseudois is unique in that the primitive frontal surface has i been rotated mediad. As a result, the horns diverge from each other i at an angle approaching 90°, and consequently only the tips extend behind the skull. j Relative to the basal circumference, the horns of Capra are longer ^ than are those of the remaining genera (fig. 4) ; also, they are longer relative to the width of the frontal surface of the horn at its bases (latero-medial width). As a result, the horns of the various goats appear to be less "massive" than are those of other Caprini, even in F SCHAFFER & REED: CAPRINI 19 C HC 50 o •• cp 50 VL Fig. 5. Circumference of the horn core (CHC) plotted against length of skull (VL) for Rupicaprini (both sexes, small solid circles) and Caprini (large circles — solid, males; hollow, females). Log log plot. The slopes of the regression lines are: Rupicaprini, 1.3; female Caprini, 1.6; male Caprini, 2.0; see Table 7. cases where they are absolutely heavier and larger in cross-section. In some populations (C. hircus aegagrus, C. h. hircus, and C. fal- coneri), the frontal surface has been further narrowed to form a median keel, but even in species with squared frontal surfaces (the ibexes), the frontal surface is not at all comparable in breadth (either absolutely or relative to the horn length) to that of the "sheep-like" Caprini — Ammotragus, Pseudois, and Ovis. The horns of Hemitragus present a special case; though extremely narrow, they are so short that the ratio of length of horn core to circumference approaches unity. To a great extent (table 7), the size of the horns can be predicted from the ventral length of the skull. Figures 5-7 are log log scatter 20 FIELDIANA: ZOOLOGY, VOLUME 61 OLHC 50 oCQ)(? ':o- o 5 50 VL Fig. 6. Length horn core (OLHC) plotted against length of skull (VL) as in Figure 5. The slopes of the regression lines are: Rupicaprini, 0.8; female Caprini, 0.9; male Caprini, L4; see Table 7. plots of the various horn measurements (CHC, OLHC, and OLHSh) against skull length (VL). Rupicaprines (male and female), female Caprini, and male Caprini are indicated by different symbols. The slopes of the regression lines are calculated in each case (table 7). In the case of each pair of measurements, the regression line was steepest for male Caprini, and least steep for the lumped Rupica- prini. Analysis by "t" test, however, indicates the differences in slope to be insignificant at the 0.05 level. Differences in the relative proportions of the horns and skull for the males of the different OLHSh ^ o o o • VL 50 Fig. 7. Length of horn sheath (OLHSh) plotted against length of skull (VL), as in Figures 5 and 6. The slopes of the regression lines are: Rupicaprini, 0.7; female Caprini, 1,5; male Caprini, 1.6; see Table 7. 21 22 FIELDIANA: ZOOLOGY, VOLUME 61 genera and for the combined Rupicaprine males are shown in Figure 8. Relative to the length of the skull, Ammotragus, Pseudois, and Ovis have wider horns; Hemitragus has shorter horn cores and sheaths. Shape of the skull The following measurements for the skulls of 14 species of Cap- rini and Rupicaprini and for Nesotragus moschatus, are given in Table 4: 1. VL — Ventral length of the skull. 2. FL — Length of the frontal bones. 3. PL — Length of the parietal bones. In addition, two angles were measured on skulls sectioned in the parasagittal plane (table 5). The angles are illustrated in Figure 11. Angle 1 gives the slope of the basioccipital bone relative to the plane of the palate; angle 2, the slope of the postero-dorsal wall of the skull. Fig. 8. Logarithms of horn measurements (X) divided by the logarithm of length of skull (VL) in male Caprini and Rupicaprini (R). SCHAFFER & REED: CAPRINI 23 Table 4. — Measurements on skulls of Nesotragus, Rupicaprini, and Caprini, N VL FL PL r M F M F M F M F Nesotragus moschatus 3 1 8.2 8.6 3.2 3.3 3.3 3.2 Nemorhaedus goral Capricornis sumatrensis Oreamnos americanus Rupicapra rupicapra 7 8 7 5 5 4 3 14.5 21.3 22.0 14.1 14.6 21.0 19.5 7.4 11.0 9.0 8.0 7.4 10.1 8.2 3.5 5.1 4.4 3.9 3.2 5.5 4.0 Capra hircus aegagrus C. hircus hircus C. ibex sibirica 4 3 4 3 2 2 16.9 17.3 20.7 13.1 14.9 17.8 10.0 10.7 12.5 7.4 7.5 9.0 4.0 3.4 4.8 3.1 3.3 4.7 Hemitragus hylocrius H. jemlahicus 2 4 2 1 17.8 18.3 16.2 16.0 12.4 15.8 10.4 12.8 4.2 2.3 4.3 3.0 Ammotragus lervia 5 5 21.7 19.0 18.8 13.5 2.1 2.3 Pseudois nayaur 3 2 15.9 15.4 21.2 9.0 2.3 3.0 Ovis musimon 0. canadensis 0. ammon 2 4 4 2 2 3 15.8 20.2 24.1 13.5 17.4 20.7 10.8 16.3 19.4 8.2 10.3 12.0 3.3 4.8 3.7 2.8 4.3 3.8 Key: All measurements in cm,; N — number of specimens examined; M — males; F — females. The positions and shapes of the various bones of the Caprine skull are presented in Figures 9, 10 (external), and 11 (internal). Among Caprini, there are two basic cranial shapes. These we will designate primitive and advanced. As the names imply, the first configuration is similar to that observed in Rupicaprini and other unspecialized Bovidae (fig. 3). Presumably, the ancestors of the sheep-goat line had skulls of this type. The characteristic feature is the fact that the postero-dorsal wall of the braincase projects behind the bases of the horn cores. Associated with this condition are the following characters: 1. Angle 2 is less than 50°, and 2. the parietal bones are at least 20 per cent as long as the ventral length of the skull (VL). Such a primitive configuration of the skull is observable in all species of Capra, in Hemitragus jayakari and H. hylocrius, and in Ovis musimon. The advanced skull is distinguished by the place- ment of the foramen magnum beneath the horn bases. As a result, the back of the skull becomes more nearly vertical. Angle 2 is greater than 70°, and the ratio of parietal length (PL) to ventral length of the skull (VL) is reduced to about 15 per cent. Hemitragus jemlahicus, Ammotragus lervia, Pseudois nayaur, and Ovis ammon 24 FIELDIANA: ZOOLOGY, VOLUME 61 have skulls that may be classified as advanced, while Ovis canadensis is intermediate in skull shape between the two categories. Thus, of the five genera, one (Capra) has retained the ancestral pattern, two Fig. 9. Lateral view of skull of female O. mtisimon. F — frontal bone; J — jugal; L — lacrimal; M — maxilla; N — nasal; OC — occipital condyle; OR — orbit; P — parietal; PM — premaxilla; SO — supraoccipital; SQ — squamosal. {Ammotragus and Pseudois) have evolved the advanced shape, and the remaining two (Hemitragus and Ovis) contain species with both configurations (fig. 3). Obviously, considerable independent and parallel evolution has occurred in several of the Caprine genera, but study of these factors will not be undertaken here. These comments on skull shape apply only to males. Among females the skull shape tends either to be primitive — in species with primitive males — or in- termediate between the male configuration and the primitive pattern — in species with advanced males (fig. 3 ; table 5) . Although at least four species have advanced skulls, the means by which this configuration has been obtained differ. In Ammo- tragus, Hemitragus jemlahicus, and Ovis canadensis (to the extent that 0. canadensis is advanced), the hraincase has actually been ro- tated ventrad. This is best seen by observing the inclination of the basioccipital relative to the plane of the palate (angle 1). Among the primitive species, this angle is low (12° in Ovis rrlusimon; 24°-29° in the different species of Capra) ; in the species that have rotated braincases the angle is higher (36°-41°). In Pseudois and Ovis am- mon, on the other hand, the advanced configuration has been ob- tained by posterior growth of the frontal bones. Angle 1 has remained low (24° in Pseudois; 15° in 0. ammon), while the length of the frontal SF Fig. 10. Top view of skull of female O. musimon. SF — supraorbital foramen; other bones labelled as in Figure 9. Fig. 11. Parasagittal section of Capra hircus hircus illustrating the bones of the basicranium and the angles measured. BO — basioccipital; BS — basisphenoid; L — lateral branch of the frontal sinus; M — medial branch of the frontal sinus; 0 — supraoccipital bone; P — parietal bone; PP— plane of the palate; PS — pre- sphenoid; 1 — angle 1; 2 — angle 2. Both angles are taken relative to the plane of the palate. 25 26 FIELDIANA: ZOOLOGY, VOLUME 61 Table 5. — Measurements of Angles 1 and 2, skulls of Caprini. Species Shape Angle FL FL PL Primitive A 10.0 PL 2.5 VL Capra hircus aegagrus 1 24 2' 27 .24 C. hircus hircus Primitive 25 40 10.7 3.2 .20 C. ibex sibirica Primitive 29 43 12.5 2.6 .24 Ovis musimon Primitive 12 48 10.8 3.2 .21 Hemitragus hylocrius Primitive — — 12.4 3.0 .24 Ovis canadensis In termed.; rot. 36 75 16.3 3.4 .24 Hemitragtis jemlahicus Advanced; rot. 40 74 15.8 7.0 .13 Ammotragu^ lervia Advanced; rot. 41 84 18.8 9.0 .10 Pseudois nayaur Advanced; back 24 92 21.2 9.2 .14 Ovis ammon Advanced; back 15 78 19.4 5.2 .15 Capra hircus hircus (F) Primitive 25 48 7.5 2.3 .22 Ovis canadensis (F) Intermed.j ; rot. 36 72 10.3 2.4 .25 Ammotragus lervia (F) In termed.; : rot. 13 56 13.5 5.9 .12 Key: Angle measurements in degrees, FL in cm.; F connotes female; rot. tation of the braincase; back — backward growth of the frontal bones. bones has increased greatly. In all species with advanced skulls, angle 2 is greater than in primitive species (74°-92° vs. 27°-48°) ; the frontal bones are absolutely longer; the ratio f^ is greater; and the ratio 1^ is smaller. These considerations are summarized in Table 5. In species with advanced skulls, little correlation exists between parietal length and skull length. Among species with primitive skulls such a correlation does exist (figs. 12, 13). Among all species, there is a fairly good correlation between frontal length and ventral length (fig. 14). Regression analysis yields slopes of 0.6 for Rupicaprini, 1.1 for females of Caprini, and 1.6 for males of Caprini excluding Pseudois (table 7; see Discussion). Ratios of |^ y^ for each genus and for the Rupicaprini lumped are given in Figure 15. Sinuses The fundamental plan of the caprine sinus system was discussed in the section on General Considerations. Figures 1 and 2 depict the sinuses as viewed from above the skull with the external tables of the frontal bones and frontal surface of the horn core removed ; Figure 3 depicts the skulls of several species in para-sagittal section revealing the extent of the frontal sinus in each. Figure 16 presents horn cores sectioned in the sagittal plane. Additional examples of sec- tioned skulls and horn cores are reproduced in the photographs. SCHAFFER & REED: CAPRINI 27 Table 6. — Data on skulls and cranial sinuses, Rupicaprini and Caprini. Species Skull ] Frontal sinus Cornual A sinus Extent Septa Extent Septa Sex M F M F M F M F M F Oreamnos americanus P P N « U * N * U * Rupicapra rupicapra P P N * U * N ♦ U * Capra hircus aegagrus P P N * u * E * U * C. hircus hircus P P N N t u t t X u C. ibex sibirica P P N * t * t * u * C. ibex nubiana P P * * * * t * u ♦ Ovis musimon P P t * X * t + X + 0. canadensis I I E t c t E E c u 0. ammon A A t * c * t * c * Hemitragus jemlahicus A I E 4i c * E * u ♦ Ammotragus lervia A I E t c X E E C X Pseudois nayaur A I E * c « t * c * Key: Skull — skull shape, see Table 5; Extent — extent of sinus; * — specimens not available for section; H — females hornless; P — primitive; A — advanced; I — intermediate; N — ^not extensive; f — moderately extensive; E — extensive; U — uncomplicated; X — moderately complex; C — complex. The extent of the frontal sinuses is correlated with the shape of the skull. In species with advanced skulls, the sinuses overlie most, if not all, of the brain; in species with primitive skulls, the sinuses cover only the anterior third of the brain. Since the sinuses pene- trate approximately to the posterior edge of the frontal bones in all species, the position of the fronto-parietal suture is the prime de- terminant of the proportion of brain overlaid by sinus. Unlike the frontal sinuses, the cornual sinuses vary from species to species in what appears to be an unpredictable fashion (table 6) . For example, in Capra hircus aegagrus, they extend to the very tips of the horn cores, while in the domestic goats (G. h. hircus) observed by us, they are confined to the horn core's basal third. A similar situation is to be found in the genus Ovis, where the cornual sinuses are extensive in 0. canadensis, but relatively short in 0. musimon and 0. ammon. Within the sinuses are the bony septa. Primitively, these are few in number and are almost totally absent in Rupicapra and Ore- amnos. Each frontal sinus does, however, contain the strip of bone that divides the sinus into lateral and medial compartments (see figs. 1, 2). In species with primitive skulls, this septum may be the 28 FIELD lANA: ZOOLOGY, VOLUME 61 only one in each frontal sinus. Other septa, if present, are few in number and irregularly placed; there is a tendency to form rings of bone that extend like hollow pillars between the ventral and dorsal tables of the frontal bones, but this is very slight. In species with advanced skulls, this tendency is more pronounced. Removal of the dorsal portion of the frontals reveals a pattern like a honey comb, which is most complex in 0. canadensis and Pseudois nayaur, but is also apparent in Ammotragus lervia, Hemitragus jemlahicus, and Ovis ammon. Viewed in para-sagittal section, the septa appear to be dorso-ventral struts. The tendency to interlock and form columnar structures can be appreciated only by removal of the outer table of the frontal bones (see pis. XIV, XIX, XXV, XXVII, XXVIII, XXX, XXXI). For each species, the complexity of the cornual septa usually par- allels that of those within the frontal sinus. Thus, in many Capra, cornual septa are almost entirely lacking. Those that do exist are PL ^ r 50 VL Fig. 12. ' Parietal length (PL) plotted against length of skull (VL) for Rupi- caprini and primitive Caprini together. Log log plot; correlation coefficient r=0.86, P>. 999. Solid circles males; hollow circles females. SCHAFFER & REED: CAPRINI 29 PL «o 0( 50 VL Fig. 13. Parietal length (PL) vs. length of skull (VL) for advanced Caprini, as in Figure 12. r=0.18; P<< 0.9. short processes originating from the interior side of the frontal sur- face and extending proximally toward the center of the core.^ More complex patterns exist in other genera, and these seem to have been produced by the posterior extension of the septa into the horn base where at least some attach to the septum dividing the lateral and medial compartments of the frontal sinus. Primitively, the cornual septa appear to have been nearly flat structures originating from the inside of the frontal surface and joining the lateral and medial sur- faces. In advanced Caprini the cornual septa form hollow tubes (figs. 16, 17). As with the septa in the frontal sinuses, this condition is best observed in Ovis canadensis and Pseudois nayaur. Even in these, however, the underlying pattern is unchanged as revealed in an adolescent 0. canadensis (pi. XXIV). No matter what the 1 In the alpine ibex {Capra ibex ibex), however, Couturier (1962) has observed intracornual septal structure of approximately the same complexity which we have observed in Ovis musimon. 30 FIELDIANA: ZOOLOGY, VOLUME 61 degree of septal complexity, all cavities of the sinus system are con- fluent with each other and with the nasal passages. In primitive species, the frontal sinuses of the females are similar to those of the males; in advanced species the frontal sinuses of the females are less extensive than those of the males. This sexual di- morphism agrees with differences in skull shape between the sexes and the correlation of the extent of the sinus with cranial shape. In all species, the septa within cornual and frontal sinuses of the females are less numerous and complex than in males of the same species. However, the extent of the cornual sinus does not vary in any regular fashion with sex. All of these considerations are summarized in Table 6. DISCUSSION The major conclusions of the preceding section can be summarized simply: 1. Caprini (male and female) have horns and frontal bones that are relatively larger than are those of the more primitive Rupi- caprini. 2. Among the various caprine genera and species, differ- ences exist with regard to the size and shape of the horns, the shape of the skull, and the extent of the sinuses. In the next section we shall discuss the significance of these statements. Growth, delayed reproduction, and natural selection Log log plots of various cranial and cornual measurements (FL, CHC, OLHC, and OLHSh) against skull length (VL) have been presented for Rupicaprini (males and females lumped), female Ca- prini, and male Caprini (figs. 5-7, 14) . The slopes of the least square regression lines are listed in Table 7. For each cranial or cornual measurement, the slope is greatest for male Caprini, and least for Rupicaprini, The slope of such a regression line is often a good esti- mate of relative growth rates (Huxley, 1932; Simpson, 1944), sug- gesting that the large horns of male Caprini are in part the result of increased growth rates. However, with the exception of frontal length plotted against skull length, the differences between male Caprini and Rupicaprini are not significant at the 0.05 level as indi- cated by the "t" test (table 8). This situation, we believe, is the result of small sample size and of our having lumped populations which have undergone divergent evolution into the same analysis. Only when the separate species are studied independently will it be possible meaningfully to compare growth rates in male Caprini with those in Rupicaprini. Such comparisons, of course, will require a far greater number of specimens than were available to us. Despite ambiguities concerning growth rates, it is clear that the horns of Caprini are relatively and absolutely larger than are those of Rupicaprini, and that male Caprini have larger horns than females (table 9). Additionally, male Caprini may be divided into two groups, primitive and advanced, on the basis of horn size relative to skull length (table 9) and skull shape. 31 Table 7. — Cranial and cornual proportions, Rupicaprini and Caprini. Measurements Group N b r P(r) FL vs. VL Rupicaprini 7 0.6 0.82 .025 Caprini females 10 1.1 0.75 .005 Caprini males 9 1.6 0.84 .0025 CHC vs. VL Rupicaprini 7 1.3 0.88 .005 Caprini females 9 1.6 0.84 .0025 Caprini males 9 2.0 0 . 91 . 0005 OLHC vs. VL Rupicaprini 7 0.8 0.74 .05 Caprini females 9 0.9 0.31 — Caprini males 9 1.4 0.52 .10 OLHSh vs. VL Rupicaprini 7 0.7 0.64 .10 Caprini females 6 1.5 0 . 74 .05 Caprini males 9 1.6 0.55 .10 Key: Measurements of male P. nayaur were omitted because of the small size of the sample (1 adult male). N — number of specimens; b — slope of the regres- sion line; r — correlation coefficient for the scatter; P(r) — probability that ob- served correlation is due to chance. Table 8. — Comparison of the regression slopes from Table 7; male Caprini compared with Rupicaprini; "t" test one-sided. Only FL vs. VL sig- nificant at the 0.05 level, d.f. — degrees of freedom. Measurement t d.f. P(t) FL vs. VL 2.3 12 < .025 CHC vs. VL 1.5 12 < .10 OLHC vs. VL 0.62 12 < .30 OLHSh vs. VL 0.89 12 < .20 log FLi logVL log CHC» logVL log OLHC logVL log OLHSh' logVL 0.76 0.74 0.76 0.69 0.80 0.75 1.05 1.00 0.84 0.79 0.97 0.81 1.06 0.88 1.16 1.17 0.97 0.85 1.10 0.84 1.22 0.83 1.44 1.11 Table 9. — Proportions of characters of skulls and horns of male and female Rupicaprini and Caprini. Group Rupicaprini M. Rupicaprini F. Prim. Cap. M. Prim. Cap. F. Adv. Cap. M. Adv. Cap. F. ^ Caprini primitive for shape of skull are the following: all Capra; H. hylocrius, and O. musimon. The remaining Caprini are advanced. 2 Caprini primitive for horn circumference were all Capra and all Hemitragus.^ 2 Caprini primitive for lengths of horn core and sheath were all Hemitragusfi a We believe that Hemitragus is not truly primitive for its characters of horn j core and sheath, but is probably instead specialized in ways not as yet understood. The differences ween Rupicaprini and primitive males, between primitive males and advanced males, and between Rupicaprini and advanced males are all significant at the 0.05 level as indicated by "t" test. 32 SCHAFFER & REED4 CAPRINI 33" FL 50 5 50 VL Fig. 14. Frontal length (FL) plotted against length of skull (VL) for Rupi- caprini (both sexes, small solid circles), female Caprini (large, hollow circles), and male Caprini (large, solid circles). Log log plot. The slopes of the regression lines are: Rupicaprini — 0.6; female Caprini — 1.1; male Caprini — 1.6; see Table 7. Increased horn size in males is evidently associated with compe- tition for mates. Among females, there is no such competition and the morphology is little changed from the ancestral Rupicaprine con- dition. Indeed, female mouflons have lost the horns entirely and those of female Pseudois are extremely small. Geist (1966a, 1968a) has observed that among North American mountain sheep increased horn size is correlated with increased mating success inasmuch as the larger (and older) males do most of the mating. There are two rea- sons for this: first, the females are more inclined to permit mounting by large-horned males, and, second, large-horned males attempt a far 34 FIELDIANA: ZOOLOGY, VOLUME 61 logFL logVL Fig. 15. Ratios of logarithm of frontal length (FL) to length of skull (VL) for Rupicaprini and various genera of Caprini. The height of the solid bar gives the values for females; that of the hollow bar, the ratio for males. Note that the amount of sexual dimorphism is generally greater in Caprini than in the Rupi- caprini. greater number of copulations. Indeed, young, small-horned rams are so very easily intimidated by the threat displays of males with larger horns that their participation in the rut is minimal. With increasing age and horn size, male bighorn seemingly grow more courageous and by the sixth to eighth year attain behavioral ma- turity. At this age they can be classified as truly "adult," and par- ticipate fully in mating activity. Since male bighorn are sexually mature by the end of the second year (Geist, 1968a), adolescence lasts four to six years. As in human populations, adolescence can be defined as the time between puberty and the attainment of adult morphology and behavioral patterns. The evolutionary basis of such a social system is grounded in the potential for bloodshed and mortal damage inherent in male compe- tition for mates. The fact that students of the animals' behavior report few fatalities resulting from combat should not cause us to lose sight of the fact that the present situation exists as a consequence SCHAFFER & REED: CAPRINI 35 of the process of evolution. Natural selection has "chosen" the modifications of behavior which maximize each individual's fitness. In the case of the Caprini, these modifications have masked the po- tential for mortal combat since subordinate males give way before the threat displays of rivals with larger horns. Fisher (1958) has observed that an animal's fitness, as measured by the number of its offspring, behaves in much the same manner as interest compounded on an investment. Clearly, delaying the date at which a man begins to deposit money in the bank reduces his profit. In a similar fashion, postponing the age of first reproduction — and this is the result of being intimidated by older males — reduces reproductive fitness. To the present authors, it is inconceivable that behavioral patterns re- sulting in delayed onset of reproductive activities would evolve in the absence of some advantage accruing to the individuals that be- haved in this fashion. Such a benefit cannot be discovered if one assumes that intraspecific combat is intrinsically harmless. Much more plausible is the supposition that combat between males of dis- proportionate size would result in injury or death to the smaller con- A C D Fig. 16. Sagittal sections of horn cores; sinuses in black. All males. A, Ore- amnos americanus (Rupicaprini) ; B. Capra ibex nubiana; C. Hemitragus jemlahicus', D. Ammotragus lervia. 36 FIELDIANA: ZOOLOGY, VOLUME 61 testant. However, such contests are avoided by the behavioral mechanisms discussed above. These considerations can be used to construct a model that predicts optimal reproductive behavior under Fig. 17. Schematized drawings of primitive and advanced septal configura- tions in longitudinal sections of horn cores of Caprini. Left — primitive; right — advanced. a variety of environmental conditions. By "optimal," we mean the behavioral pattern that maximizes the individual's fitness. Here we consider r, the malthusian parameter of Fisher, as the most appro- priate measure of fitness. We also assume that, among members of a given species, behavioral differences will have a genetic basis and, fur- ther, that different behavioral phenotypes will have different fitnesses. Consider a parameter k which reflects that rate at which a male matures behaviorally. A high value of k means that the animal at- tempts to compete with older males and mount females at an early age; a low value of k indicates that the animal does not first challenge the dominant males until he is much older. Evidently, k is a measure of a young male's ability to stand his ground in the face of threat displays by older, larger-horned rivals. Indeed, k is inversely pro- portional to the length of adolescence. In any population, there will be a distribution of k values among the various animals. Given a genetic basis for these values, it is reasonable to suppose that the modal value of k observed in the population corresponds closely to the optimal value. Any evolutionary innovation, behavioral or mor- phological, will have a cost and a profit (Gadgil and Bossert, 1970). The fitness of the modification can be expressed by the equation Fitness= Profit -Cost (1) SCHAFFER & REED: CAPRINI 37 In the case of selection for optimal k, we write Fitness=P(k)-C(k) (2) where P(k) is the profit and is a function of k and C(k) is the cost and is also a function of k. Since fitness is also a function of k, we seek to determine the value of k that maximizes the expression P(k) — C (k) . We have observed, as have also Cole (1954) and Gadgil and Bossert (1970), that fitness decreases when the age of first repro- duction is increased. This decrease will be our cost function C(k). At k=0, C(k) is infinite because the organism never reproduces; at k= 00, C(k) = 0 because the animal reproduces as soon as he is physi- ologically capable. C(k) will thus be a decreasing function asymp- totially approaching zero as k approaches ^ (fig. 18). If there is no Ar K 00 Fig. 18. Cost, C(k), and profit, P(k), functions, plotted against k. k is pro- portional to the rate at which a male matures behaviorally. Fitness (Ar) is maximized when P(k) — C(k) = maximum. k is the value of k that maximizes r. See text for further explanation. 38 FIELDIANA: ZOOLOGY, VOLUME 61 AGE Fig. 19. Survivorship curves in harsh (H) and benign (B) environment. I X is the probability that an animal will live to age x. profit for delaying the first reproduction, equation 2 will be maxi- mized at k=oo. This is the case for female Caprini which do not compete for mates, and it is not surprising that female mountain sheep begin reproducing shortly after the onset of puberty. In males, however, there is a profit function. Increasing k lowers the age of first reproduction and increases the possibilities of fatal con- tests between younger rams and the larger males. Decreasing k re- duces the possibility of such encounters. Since males essentially cease growth by the eighth year, reducing k below the value that in- dicates first reproduction at this age does not further increase P(k). Accordingly, when plotted against k, P(k) will be finite and essen- tially constant for low values of k and then decline to 0 as k ap- SCHAFFER & REED: CAPRINI 39 Ar Fig. 20. k is obtained for harsh (H) and benign (B) environments. In the harsh environment r is greater (kn) than in the benign one (ke). proaches oo (fig. 18). Plotting profit and cost functions together, we obtain Figure 18. The value of k that maximizes the difference between P(k) and C(k) is denoted k. This is the optimal value of k. Notice that if P(k) = 0 for all k, k = oo. This, we observed, was the predicted value of k for females. We now inquire into the effect of varying environmental harsh- ness on k. The term "environmental harshness" is defined in terms of a population's survivorship curve (Deevy, 1947) . Obviously, an environment that is harsh for one species may be perfectly hospitable for another. Again, an environment adjudged harsh by one scientist may not be so judged by another. We shall take the shape of the survivorship curve as a measurement of harshness. A harsh environ- ment is one in which the survivorship curve falls off steeply; in a O O 00 I o o I "^ 00<£> r-lOO*. tH O I T-l 1-1 IM (N J3 be O Ui ■> C! (1> 73 C >, tc 'o a -s JO o E XI < < O lO o w <^ ■b IX <^ U o :2; i eo 0<-l i-HrH "tlOOO •<4 toco O ^ i>5 1^ .^li.i a JS iceoiMo M OWTtTt s B bfl a 3 O c a CO OIM(M(M 0) o 40 SCHAFFER & REED: CAPRINI 41 benign environment, the curve falls off more gently (fig. 19) . As an index of shape of the survivorship curve we will use the statistic x" H=l max. where X is the expected lifespan of an individual and X^j^x. is the maximum life expectancy. If all individuals survive to their maxi- mum age, H=0. This is the least harsh environment conceivable; if nearly all the individuals die just after birth, H=l, the harsh- est possible environment. H will clearly be affected not only by physical factors — average temperature, rainfall, etc. — but also by biological factors — competition, predation, parasitism, etc. — and, in addition, by factors peculiar to the physiology and behavior of the species in question. In Figure 20, the effects of increasing H on P(k) and C(k) are shown. If males have little probability of living to an old age regardless of their reproductive behavior, P(k)^the increase in survivorship resulting from avoiding the rut — will be reduced be- cause there will be fewer older males to injure aggressive adolescents. The cost function, C(k), on the other hand, will be increased by re- ducing k, for the very reason that delay will more likely mean that the individual will not reproduce at all. As a result, k will be greater in harsh environments than in benign ones. Stated another way, we expect the age of first reproduction of males to decrease with increas- ing environmental harshness. Survivorship data for large mammals are difficult to obtain. How- ever, for Caprini, life tables of reasonable accuracy are available for Ovis canadensis (Deevy, 1947; Murie, 1944); for Hemitragus jemla- hicus imported into New Zealand (Caughley, 1966) ; for feral Soay sheep {Ovis aries) on the island of Hirta, St. Kilda, on the western fringe of the Outer Hebrides, Scotland (Boyd et al., 1964) ; and for Ovis musimon (Pfeffer, 1967). In addition. Couturier (1949) gives life table data for the rupicaprine Rupicapra rupicapra. In Table 10, we have recorded the following data for these populations: H, X, Xniax.» A (the age at first reproduction), and A/X^iax. (the relative age of first reproduction). In Figure 21, we have plotted H against A/Xnjax. for both males and females. In accordance with our pre- diction, the relative age of first reproduction among males declines with increasing environmental harshness. No such pattern is visible among the females. Changes in male reproductive behavior are most dramatically illustrated by the Soay sheep. In this popula- 42 FIELDIANA: ZOOLOGY, VOLUME 61 9 H • males o females o o o ° 25 50 A/X^ Fig. 21. Environmental harshness, H, plotted against the relative age of first reproduction, A/Xm (Xm=Xmax.)f for different populations of Caprini and Rupicaprini. Among males, increasing environmental harshness is associated with an earlier age of first reproduction. The correlation coefficient, r=.98; P(r)<.005 (one-sided). No such trend is apparent among females. tion, the male lambs attempt to participate in the rut, both running with the mature rams and attempting copulations (Boyd et al., 1964; Grubb, personal communication). Indeed, Grubb (personal com- munication) reports that these lambs are subjected to quite vicious attacks by the older males. Clearly, they are not so easily intimi- dated by the threats and attacks of the adults as are the young American bighorn studied by Geist (1966a). The low survivorship rates of male Soay sheep are due both to the limited grazing area available on the island and the presence of an internal parasite which periodically decimates the population. Most susceptible to the para- site are the yearling rams. Thus, selection for attempted copulation by lambs is quite understandable. Contrast this situation with that of the mouflons studied by Pfeffer (1967). Although the two species are quite closely related, survivorship is higher in the males of wild mouflons, where H=.67, than in Soay males (H=.87). Associated with increased survivorship, is an increase in the relative age of first reproduction (A/Xnjax.= -40 vs. .05 for Soay males), a correlation which our theory predicted. SCHAFFER & REED: CAPRINI 43 In the tahr the situation is less clear. Published studies of be- havior are entirely lacking, and Caughley's estimate of the modal age of male reproduction must be viewed with some caution. Also, the only available survivorship data are for females, and we must admit to the possibility that these do not closely agree with the male data, which have never been collected. Nonetheless, we work with what we have and the agreement between the predictions of the theory and the observations on the populations is encouraging. Further confirmation of the theory is provided by Geist's obser- vations (1964) of the rupicaprine Oreamnos americanus, the Rocky Mountain "goat." In this species, the horns are but slightly curved and end in sharp tips. This fact makes fighting between rival males a potentially lethal proposition for both contenders no matter what the size differential between them. The usual result of the rare fights that do occur is the sustaining of extremely damaging puncture wounds. As a result, it does not "pay" an adult male to fight even a juvenile as the chances of injury are so great. Predictably, there is little or no fighting between males. Indeed, Geist has observed that adult males are subdominant to yearlings and females. Since older males do not attempt to exclude their younger rivals from the rut, the profit function P(k) is 0 for all k. One then expects the males to reproduce as soon as physiologically possible, which indeed they do, as both males and females first reproduce at 23/^ years. Growth is the opposite side of the coin we have been discussing. Energy devoted to increased horn size must be diverted from some other function. The usual correlates are decreased survivorship or fecundity. This diversion of energy is the cost function and will be selected for only if there is an associated profit — in this case increased effectiveness in combat and display. As horn size increases, the profit function levels off and becomes constant when the horns are suffi- ciently large to enable their bearer to defeat or intimidate most of his rivals. The cost function, on the other hand, continues to increase at all horn sizes. Thus, an optimal horn size is determined. The profit and cost functions associated with increased horn size are affected by environmental harshness in much the same way that the profit and cost associated with delayed reproduction are affected. In a harsh environment, the profit is diminished — because there will be fewer older, large-horned males with whom to compete. Simi- larly, because life expectancy is diminished, the cost of any given diversion of energy will be increased. As a result, we predict smaller 44 FIELDIANA: ZOOLOGY, VOLUME 61 •9 H 7 0 HORN SIZE Fig. 22. Environmental harshness (H) plotted against relative horn size for different populations of Caprini and Rupicaprini. Among males, horn size de- creases with increasing harshness. r=.95; P(r)<. 01 (one-sided). Among females, no such trend is apparent. horns in a harsh environment. As a measure of horn size, we use the statistic H.S., which is defined by the relation O.L.H.Sh. + C.H.C. H.S.= l/2 ■ ( V.L. ) Values of H.S. are included in Table 10. In Figure 22, we have plotted H.S. against H, our measure of environmental harshness, for both males and females of different Caprini and Rupicaprini. In accordance with the prediction, increased environmental harshness is associated with smaller horns in the males. Among females, there is no such pattern. This situation is to be expected since the females do not use the horns to fight with each other for mates. Indeed, the functions of the horns in female Caprini remain unclear to us. Asso- ciated with the relatively larger horns of male Caprini is increased sexual dimorphism with regards to both horns and skulls. In gen- eral, Caprini are more highly dimorphic than Rupicaprini (table 11). SCHAFFER & REED: CAPRINI 45 Table 11. — Values of sexual dimorphism, Rupicaprini and Caprini. Species VL FL CHC OLHC OLHSh Nemorhaedus goral —.01 0 .23 .14 .06 Capricornis sumatrensis .01 .08 .09 .16 .17 Oreamnos americamis .11 .09 .28 .19 .02 Average Rupicaprini .04 .04 .20 .16 .08 Capra hircus aegagrus .23 .26 .56 .73 .72 C.hircushircus .14 .30 .52 .57 .62 C.ibexsibirica .14 .28 .50 .67 J Hemitragus hylocrius .09 .16 .30 .30 .22 H.jemlahicus .13 .19 .36 .44 J Ammotragus lervia .12 .28 .42 .41 .45 Pseudois nayaur .03 .58 .70 .80 % Ovis musimon .15 .24 1.00* 1.00* 1.00* O. canadensis .14 .37 .64 .69 .73 O.ammon .14 .38 .62 .70 .60 Soay Sheep** .04 .13 .34 .72 Average Caprini .12 .29 .54 .64 Key: The index of dimorphism for each measurement, X, is given by the Xm Xf formula = index of sexual dimorphism. * Females hornless. X Females with hornsheaths not available. ** Hornless females excluded. This, we feel, reflects both increased survivorship and a greater tend- ency to joust strenuously with the horns. Male Rocky Mountain goats, as we have previously noted, do not fight at all — hence the similarity of horn structure between males and females. Of the other Rupicaprini, behavioral and life historical data exist only for Rupi- capra rupicapra (Couturier, 1938, 1949). The males of this species do fight with each other and use their horns for this purpose. How- ever, fighting is less intense than in Caprini and survivorship is con- siderably lower (H=.83). The relatively small size of the horns in the males and the early age of first reproduction agree with these observations. In closing this section, we would like to observe that the harsh- ness or benevolence of an environment is a function not only of ex- trinsic factors — the physical and biological environment — but also of the morphology and behavior of the animals in question. In addition to cranial specialization, the Caprini have evolved remark- able talents for climbing, jumping, and running. These, coupled with their relatively large size and preference for mountainous habitats, 46 FIELDIANA: ZOOLOGY, VOLUME 61 have undoubtedly reduced predation pressures on healthy adults. In essence, it is possible that becoming adapted to live in the moun- tains with grace, enabled the Caprini to increase their survivorship and thus convert a harsh world into a benevolent one. Such a con- version would then pre-adapt the animals for evolutionary paths that led to delayed reproductive patterns in males and the growth of splendid horns. In our view, then, delayed reproductive behavior can evolve without the development of the type of harem system such as is observed in sea lions, deer, and some antelope. These social systems do, of course, also involve an unwillingness on the part of young males to challenge the dominant males. Unfortu- nately, the conditions that led to the evolution of a caprine type social system as opposed to a harem system with its concomitant territoreality remain unclear to us. Prevention of brain damage Thus far, we have dealt with generalities: growth and its adap- tive significance in the context of a species' behavior and ecology. We now turn to some specifics which, while perhaps more mundane to the evolutionist, are of the utmost importance to a ram contem- plating intraspecific combat. The first of these topics concerns the prevention of damage to the brain. Caprini appear to be protected against two, and possibly all three, of the following causes of head injury: 1. Inbending of the bone surrounding the brain; 2. angular acceleration of the skull; 3. linear acceleration of the skull. We shall consider each of these potential sources of injury and the means by which their realization is prevented in the Caprini. The most obvious result of impact to the calvarium is fracture. However, actual breaking of bones is not necessary to produce con- cussion and even death, since local, reversible inbending can generate transient increases in pressure throughout the brain cavity. Because the skull is much more rigid than is the spinal tube,^ increased intra- cranial pressure will cause fluid to flow out of the braincase and into the spinal tube. Such flow generates shear stresses in the region of the cranio-spinal junction (Gurdjian and Lissner, 1961; Edberg et al., 1963) which result in lesions in this area. In addition, the snap-back of the bone will produce local negative pressures that may be accom- panied by cavitation and destruction of nerve tissue and blood ves- 1 The dura mater which encloses the spinal contents is free to expand between the vertebrae if the fluid pressure within the tube is increased. SCHAFFER & REED: CAPRINI 47 sels. Finally, impact of sufficiently great force will momentarily compress the entire skull, generating shear stresses in the interior of the brain and causing destruction of tissue in this region as well. In all Caprini observed, at least some of the blows land on the forehead, which means that impact occurs on the frontal bones. But these have been divided into two layers of bone with an intervening air sinus. Compression of the outer layer will not necessarily result in compression of the inner layer which surrounds the brain. Geist (1966a) has pointed out that frontal sinuses are common in many animals that butt with the head. The protection against deforma- tion of the calvarium afforded by such sinuses would appear to be their raison d'etre. In species such as Ovis canadensis in which the force of impact is great, the outer layer of the frontal bones is but- tressed by the septa within the sinus. Thick, regularly shaped septa would, of course, transmit deformation of the outer layer to the brain- case. It is significant that in all Caprini the septa are thin and shaped irregularly. Unterharnscheidt (personal communication) has sug- gested that the septa act as springs to cushion the blow by under- going slight deformation themselves. Such a mechanism is quite reasonable, but can only be tested by experiments on freshly-killed material which was unavailable to us. Rotation of the skull about the cranio-spinal junction is a second possible source of injury. Holbourne (1945) demonstrated that in man such rotation will produce shear stresses within the brain mass. Unterharnscheidt and Higgens (1969) have pointed out that angular acceleration will also cause the brain to rotate relative to the inner wall of the braincase, causing superficial lesions and the tearing of nerves and blood vessels. Pudenz and Sheldon (1946) observed such rotation in monkeys whose calvaria had been replaced with Incite domes. In a previous publication, one of us (Schaffer, 1968) pointed out that the development of the neck muscles in the various Caprine species is proportional to the magnitude of the torque resultant from impact. The implication is that Caprines resist rotation of the skull by contracting the neck muscles during impact. While the develop- ment of the neck muscles is important in this regard, attention also needs to be given to the shape of the skull. In goats, the skull is primitive. The brain protrudes behind the bases of the horns. Since impact occurs far out on the horns (fig. 23), a torque will be gen- erated that will cause rotation about the occiput in a clockwise di- rection. Accordingly, the neck muscles and their areas of insertion of the skull are quite large. If the skull shape were advanced, i.e., if the foramen magnum were beneath the horn bases, the torque CAPRA Fig. 23. Impact in Capra. The impact vector Fi is displaced from the sup- port vector of the neck (Fs), by the lever arm X. Notice that if the skull configu- ration were of the advanced type, the condyles would underUe the horn bases and X would be larger. (After Schaffer, 1968.) 48 SCHAFFER & REED: CAPRINI 49 would be much greater in magnitude. The retention of the primitive skull pattern in goats is, we believe, correlated with the fact that im- pact occurs far out along the horns. Alteration of the skull shape in the direction of the more sheep-like Caprini would needlessly lengthen the lever arm, hence requiring the presence of even larger neck mus- cles to counter the increased rotary torque. We should also observe that crossing relatively narrow horns is an effective way of receiving a blow only if the force of impact is relatively low. In advanced Caprini, on the other hand, the force of impact is presumed to be much higher than in the goats (Schaffer, 1968). As a result, the horns are much broader and in Ovis and Ammotragus impact occurs near the bases of the horns. In this case, impact would tend to ro- tate the skull counterclockwise if the skull shape were still primitive. Of course, it is not (fig. 24), and we would suggest that the adaptive significance of the advanced skull configuration lies in the fact that it brings the supporting column of the neck in line with the vector of impact, thus reducing the magnitude of the rotary torque. Place- ment of the vertebral column in line with the force of impact also provides more efficient support for the skull. This is of particular importance in the larger sheep in which the force of impact may be as much as three to four times that in goats of corresponding size. Thus, the point of impact, which is a function of the placement and spiral angle of the horns, is correlated not only with the development of the neck muscles, but also with the shape of the skull and the magnitude of the force of impact.' A third possible source of damage to the brain is linear decelera- tion. For a braincase whose diameter in the direction of impact is 5-8 cm,, declerations causing 350-500 G will produce a complete vacuum in the region of the brain opposite the point of impact (Sellier and Unterharnscheidt, 1963), Among the larger sheep, the deceleration caused by impact may indeed approach this figure. The change in velocity may be as much as 1400 cm. /sec. (=30 mph) but the time interval during which this occurs has not been determined. If the horn and septa act as cushioning devices, one would expect the deceleration time to be rather long (say several milliseconds) in which case the maximum tolerable limit would not be exceeded. On the other hand, it may be that male sheep do sustain brain damage 1 An apparent exception to these remarks is Pseudois, in which impact occurs about half way along the horns. These, it will be recalled, diverge from each other at an angle approaching 90° so that the points of impact and the support vector of the neck are roughly coplanar. As in Ovis and Ammotragus, a primitive skull shape would result on impact in the production of counter-clockwise torques about the occiput. 50 FIELDIANA: ZOOLOGY, VOLUME 61 during combat which in the long run proves injurious. The resolu- tion of this problem, however, will come only after experimentation far beyond the scope of this study. We do emphasize that in species of Caprini that do not charge their opponents at high speed — the goats, Ammotragus, and probably Hemitragus — the change in veloc- ity will be insufficient for linear deceleration alone to cause brain damage, cushioning or not. Fighting style and the shape of the horns In the section on behavior, we observed that in Ovis and Ammo- tragus (and probably also in Pseudois) impact is frontal and usually horn to horn. In goats, on the other hand, the horns are crossed and the animals' heads are tilted at the moment of impact, producing rotational torques in the transverse and sagittal planes (Schaffer, 1968). Such torques are not generated when impact is frontally directed and it is not surprising that in species in which the males actually charge each other (thus increasing the magnitude of the im- pact force) the approach is as directly head-on as the two antagonists can co-operatively produce. In such cases, obviously, it is important that the contestants not slip past each other and violently wrench their necks. Correlated with horn-to-horn butting are broad frontal surfaces (fig. 4; table 9), and, in the cases of Ovis and Ammotragus, impact near the bases of the horns where they are broadest. In Capra, on the other hand, the horns are crossed on impact. In this case, broad horns are superfluous. Predictably, the horns of Capra are relatively and absolutely more narrow than those of Ammotragus, Pseudois, and Ovis. In addition to the effects of interspecific differences in fighting style, one can also consider the effects of butting behavior in a more general sense. For example, the horn core, like any other structure, can be expected to be designed efficiently. Like a beam attached at one end and loaded at the other, efficient design means constructing the horn core in such a way that the ratio of the compressive and ten- sile stresses equals the ratio of the compressive and tensile strengths. Bone is stronger in compression than in tension (Evans, 1957). Ac- cordingly, we expect the horn core to be shaped in such a way that the maximum compressive stresses exceed the maximum tensile stresses. Figure 25 illustrates a beam anchored at one end and loaded at the other (L). The fibers near the top of the beam are stretched; those near the bottom compressed. Within the beam is a layer of fibers in the neutral axis (n), in which the fibers are neither stretched nor compressed. In a straight beam the neutral axis cor- OVIS Fig. 24. Impact and support vectors in Ovis. Impact is to the bases of the horns and the impact and support vectors are nearly co-Hnear. Retention of the primitive skull pattern would place the condyles behind the point of impact and would generate a counter-clockwise torque. (After Schaffer, 1968.) 51 52 FIELDIANA: ZOOLOGY, VOLUME 61 L Fig. 25, Compressive and tensile stresses in a loaded cantilever beam. The loading (L) is at the end of the beam. Within the beam, bending stresses of ten- sion (T) and compression (C) are generated. Along the neutral axis (n), the fibers are neither stretched nor compressed. The magnitude of the stresses increases with distance from the neutral axis. responds to the line joining the centroids of successive cross-sections. In a curved beam (fig. 26), the neutral axis lies below the centroidal axis. The distance — denoted by "e" in the figure — is a function of the shape of the beam's cross-section and the upper and lower radii of curvature. The magnitude of stresses induced by loading in- creases with distance from the neutral axis. As a result the maximum stresses occur along the edges of the beam. Also, the position of the centroid affects the relative magnitude of the compressive and tensile stresses. For a straight beam, the ratio of compressive to tensile stress is given by the equation <7C «rt hi h2 (1) where h^ is the distance from the centroid to the bottom of the beam and h2 the distance to the top of the beam (fig. 26). For a curved beam, expression (1) is complicated by the effect of the differing radii SCHAFFER & REED: CAPRINI 53 I i '2 Fig. 26. Cross-section of a curved cantilever beam. C is the centroid, which is located at the center of mass. Because the beam is curved (down in this illus- tration), the neutral axis is displaced below the centroid a distance e. hi is the distance from the centroid to the bottom edge of the beam; h2, the distance to the top edge. The radius of the top edge is given by xz; that of the bottom edge by n. of curvature of the top and bottom edges of the beam, the ratio of stresses is given by the expression ffc hi — e ri at In this case, (2) h2+e xz where ri is the inner radius of the beam and x<2, the outer radius (Timo- shenko and Young, 1968) ; e is the distance between the centroidal axis and the neutral axis. In an unpublished study, one of us (Schaffer) determined the position of the centroids of basal cross sections of 12 caprine horn cores (table 12). As a result, it was possible to measure hi and h2 54 FIELDIANA: ZOOLOGY, VOLUME 61 Table 12. — Ratios of maximum compressive to maximum tensile stresses in the horn cores of Caprini. I Genus N ri rj hi h^ e at Capra 3 28.0 35.2 2.9 2.6 0.1 1.27 Hemitragus 1 12.9 18.3 4.9 4.5 0 1.55 Ammotragus 2 13.1 18.6 3.1 2.5 0.2 1.61 Ovis 6 10.5 19.8 4.9 4.3 0.6 1.70 Average 1.56 Std. dev. 0.17 Key: N — number of specimens; other measiu-ements are defined in Figures 25, 26. * The ratio of stresses was computed for each specimen and these averaged for each genus. The ratios of the stresses for all 12 specimens were computed and averaged; the standard deviation was also based on the ratios for the individual specimens. as well as the radii of curvature; e was calculated from the shape of the cross-section. These data were then used to compute the ratio of stresses in each specimen (table 12) . The results were in striking agreement with the ratio of ultimate strengths in compression and tension calculated for fresh bone (Evans, 1957). The ratio of the stresses was 1.56 (std. dev.=0.17); that of the strengths, 1.57 (std. dev,= 0.31) . This result must be viewed with caution, however, since the analysis is only valid if loading (in this case impact during but- ting) is quasi-static; that is the time during which the horn is loaded, i.e., the length of time that the opponents' horns are in contact, greatly exceeds the time it takes the shock produced by impact to travel down the horn to its base and back to the point of contact. In addition, it is assumed that relative to the stresses produced by bending, shear stress is negligible. Preliminary investigations indi- cate that this is probably true for Capra and Ammotragus. In Ovis, however, the force of impact is greater than in the other genera, and the point of impact closer to the base. Actual measurements of im- pact times and forces will be necessary to determine the applicability of the computations to this genus. The fact that the stresses placed upon the horn by ramming so greatly exceed all others suggests additional studies which make use of the methods of structural design. The optimal stiffness of the horn, the resistance to shearing, and the means by which the energy of impact is dissipated are but three. All of these, however, require more detailed knowledge of the nature of impact than is now available. We shall close this section with a single observation concerning the orientation of the septa within the horn core. Reference to Fig- SCHAFFER & REED: CAPRINI 55 ures 16 and 17 shows that most of the septa lie along the lines of principal stress, which because of the curved shape of the core are mostly lines of compression. Such orientation gives maximum strength while utilizing a minimum of material. It should be noted that this sort of arrangement of bone has also been observed in the human femur and calcaneum (Thompson, 1945). Its presence in the horn cores of Caprini is therefore not surprising inasmuch as these structures are subjected to far greater loading forces. SUMMARY AND CONCLUSIONS A study of the agonistic behavior and cranial anatomy in the sheep and goats (bovid tribe Caprini — Capra, Hemitragus, Ammo- tragus, Pseudois, and Ovis) reveals the following patterns character- istic of the group: 1. An emphasis on head-to-head butting or ram- ming as a major form of intrapsecific competition between adult males as a prelude to mating; 2. Lack of harem formation, but ex- clusion of sexually competent juvenile and adolescent males from the rut by the intimidative activities of mature males in populations with high survivorship rates; 3. Great increase in the size of horns and horn cores in the males; 4. Concomitant increase in the size and com- plexity of the frontal and cornual sinuses and an increase in the complexity of the septa within these cavities; 5. An associated altera- tion in the shape and relative proportions of the skull, the changes being particularly profound in those species in which the force of impact is great and directed to the bases of the horns. The evidence indicates that the different characters, behavioral and morphological, evolved together as an adaptive complex. Fe- males do not compete for mates by butting and ramming each other, and do not have large horns, nor do they exhibit the delayed repro- ductive patterns characteristic of the males. Additionally in females, the shape of the skull and the extent and complexity of the sinuses and septa have remained primitive. By contrast, the living Rupi- caprini, presumed to resemble closely the antecedents of the sheep and goats, are characterized by body butting (a pattern observed in juvenile sheep, but only infrequently in adults — Geist, 1968b), and little or no increase in size of the horns, sinuses, and septa. The shape of the skull is similar to that of caprine females and to that of the most primitive male Caprini. Little is known concerning the age at which Rupicaprine males first attempt copulation with estrous females, except in Rupicapra rupicapra where the first reproduction is delayed less than in mountain sheep and tahr. This agrees with the fact that survivorship in Chamois is lower than in the aforemen- tioned Caprini. In Rocky Mountain goats (Oreamnos americanus) fighting be- tween males of disproportionate or equal size has been selected 56 SCHAFFER & REED: CAPRINI 57 against because of the injurious effects to both contestants produced by penetration of the sharp horns into the body. Predictably, the males do not compete with each other during the rut nor do they delay their first entry into the rut beyond the age at which they first become sexually mature. Male bighorn, on the other hand, become sexually mature at about the same age as the mountain goats (two to three years), but delay their first entry into the rut until the sixth to eighth year. This fact, we believe, is associated with the observa- tion that male sheep will fight with each other in the manner de- scribed earlier in this paper. Such agonistic behavior is possible, we think, because an older male can defeat a younger rival without risk- ing injury. In general, male Caprini are distinguished by curving horns that permit head-to-head butting and ramming that will not result in injury to the larger animal. The evolution of a curved horn is very likely the adaptation upon which subsequent changes in horn size and reproductive behavior are based. Within the Caprini, the configuration of the skull has remained primitive in those species in which the force of impact is low. In contrast, Ammotragus, the larger sheep and presumably Pseudois ram each other with much greater force and have evolved a new skull shape in which the foramen magnum is rotated to a point below the site of impact. The brain is protected from damage by several adaptations which include increased extent of the cornual and frontal sinuses and the septa within, and again by the placement of the sup- porting column of the neck vertebrae in line with the impact vector. In goats, rotation of the skull about the occiput at the second of im- pact is prevented by contraction of the neck muscles which are rela- tively larger than in species with advanced skull shapes. The correlated morphological and behavioral differences between the living genera of Caprini, and particularly those between Capra and Ovis, lead us to reject the suggestion by Payne (1968) that sheep may be a late Pleistocene or early Recent derivative of goats. Our interpretation is, instead, that the evolution of the functional com- plex of male agonistic behavior and cranial morphology, so funda- mentally different in goats and sheep, has been progressing for several million years. The necessity for postulating a relatively long period of gradual evolutionary change is supported by Pilgrim's (1947) statement that fossil sheep have been recovered from the upper Pliocene and by the evidence summarized by Kurt^n (1968) that Capra, Hemitragus, and Ovis have all been present in Europe at least 58 FIELDIANA: ZOOLOGY, VOLUME 61 for much of the Pleistocene. We do not, in addition, find convincing Payne's suggestion that all living Caprini form a coenospecies, with potential gene-flow possible throughout all populations of the tribe. On the other hand, we are hopeful that the evidence presented in this paper on the correlation between male behavior and cephalic morphology in different Caprini will provide paleontologists with in- sights into the behavior of populations of Caprini represented by fossils. The evolution of Caprini has been distinguished by the evolution of the ability to live safely in mountainous areas by virtue of large size and increased agility. These factors, we feel, have contributed to the convex survivorship curves that appear to distinguish the larger species. Increased survivorship, we have pointed out, in con- cert with a curved horn that permits non-mortal combat, makes adaptive the growth of larger horns and the evolution of delayed reproductive patterns among males. ADDENDUM While the present article was in page proof, a fine book on sheep by Geist (1971) appeared, clarifying several factors of behavior, particularly of Bighorn and Stone's sheep in western Canada. He emphasizes that larger-horned males are dominant to smaller-horned, that the most intensive combat ("jousting") is between strange males with horns of equal size, that the winner gains dominance, and that the more dominant males (invariably those younger adults with the largest horns) irregardless of age are relatively more successful in the rut and thus sire the majority of the next year's lambs. Thus selec- tion against small-horned rams is intensive, but selection for all of the characters of head and horn whereby jousting rams resist con- cussion is also intense, as we had concluded before seeing Geist's book. REFERENCES Boyd, J. M., J. M. Doney, R. G. Gunn, and P. A. Jewell 1964. The Soay sheep on the island of Hirta, St. Kilda: A study of a feral pop- ulation. Proc. Zool, Soc, London, 147, pp. 127-163. Caughley, G. 1966. Mortality patterns in mammals. Ecology, 47, pp. 906-917. Colbert, E. H. 1935. Siwalik mammals in the American Museum of Natural History. Trans. Amer. Phil. Soc, n.s., 26, pp. x+401. Cole, L. C. 1954. The population consequences of Hfe history phenomena. Quart. Rev. Biol., 29, pp. 103-137. COLLIAS, N. E. 1956. The analysis of socialization in sheep and goats. Ecology, 37, pp. 228- 239. Couturier, M. A.-J. 1938. Le chamois: Rupicapra rupicapra (L.). B. Arthaud, Grenoble. xiii4- 802 pp. 1949. Sur les traces de mes 500 chamois de France. B. Arthaud, Grenoble. 276 pp. 1962. Les bouquetins des Alpes. Privately published, Grenoble, xiii +1,565 pp. Cowan, I. McT. 1940. Distribution and variation in the native sheep of North America. Amer. Midi. Nat., 24, pp. 505-593. Darling, F, F. 1937. A herd of red deer: A study in animal behavior. Oxford University Press, London, x+215 pp. Deevey, E. S. 1947. Life tables for natural populations of animals. Quart. Rev. Biol., 22, pp. 283-314. Edberg, S., J. Reiker, and A. Angrist 1963. Study of impact pressure and acceleration in plastic skull models. Lab. Invest., 12, pp. 1305-1311. Ellerman, J. R. and T. C. S. Morrison-Scott 1951. Checklist of Palaearctic and Indian mammals 1758-1946. Printed by Order of the Trustees of the British Museum, London, pp. ix+810. EsTES, R. D. 1967. The comparative behavior of Grant's and Thompson's gazelles. J. Mam- mal., 20, pp. 189-209. 59 60 FIELDIANA: ZOOLOGY, VOLUME 61 Evans, F. G. 1957. Stress and strain in bones: Their relation to fractures and osteogenesis. Charles C Thomas, Springfield, 111. x+245 pp. Fisher, R. A. 1958. The genetical theory of natural selection. 2nd ed., rev. Dover Publi- cations, New York. 291 pp. Gadgil, M. and W. H. Bossert 1970. Life historical consequences of natural selection. Amer. Nat., 104, pp. 1-24. Geist, V. 1964. On the rutting behavior of the mountain goat. J. Mammal., 45, pp. 551-568. 1966a. The evolutionary significance of mountain sheep horns. Evolution, 20, pp. 558-566. 1966b. The evolution of hornlike organs. Behavior, 27, pp. 175-214. 1968a. On delayed social and physical maturation in mountain sheep. Canad. J. Zool., 46, pp. 899-904. 1968b. On the interrelation of external appearance, social behavior and social structure of mountain sheep. Z. Tierpsychol., 25, pp. 199-215. 1969. Ovis canadensis (Bovidae). Fighting behavior between males. Wiss. Film, AII(5), pp. 525-535. 1971. Mountain sheep: a study in behavior and evolution. The University of Chicago Press, Chicago, xv + 383 pp. Grzimek, B. 1966. Wild animal, white man: Some wildlife in Europe, Soviet Russia, and North America. Hill and Wang, New York. 360 pp. Gurdjian, E. S. and H. R. Lissner 1961. Mechanism of concussion. In F. G. Evans, Biomechanical studies of the musculo-skeletal system. Charles C Thomas, Springfield, 111., pp. 198-208. Haas, G. 1959. Untersuchungen liber angeborene Verhaltensweisen bei Mahnenspringern {Ammotragus lenia Pallas). Z. Tierpsychol., 16, pp. 218-242. Hafez, E. S. E. and J. P. Scott 1962. The behavior of sheep and goats. In E. S. E. Hafez, The behavior of domestic animals. Ballier, Tindall & Cox, London, pp. 297-333. HOLBOURN, A. H. S. 1945. The mechanism of brain injuries. Brit. Med. Bull., 3, pp. 147-149. HUTTON, A. F. 1947. The Nilgiri tahr (Hemitragus hylocrius): A duel between males. J. Bom- bay Nat. Hist. Soc, 47, pp. 374-376. Huxley, J. S. 1932. Problems of relative growth. Dial Press, New York. xix-|-276 pp. Innes, a. C. 1958. The behavior of the giraffe. Proc. Zool. Soc. London, 131, pp. 245-278. Katz, I. 1949. Behavioral interactions in a herd of Barbary sheep. Zoologica, 34, pp. 9-18. SCHAFFER & REED: CAPRINI 61 KURTEN, B, 1968. Pleistocene mammals of Europe. Aldine Publishing Company, Chicago. vii+317 pp. Lydekker, R. 1898 . Wild oxen, sheep & goats of all lands. Rowland Ward, London. xiv + 318 pp. McCann, L. J. 1956. Ecology of the mountain sheep. Amer. Midi. Nat., 56, pp. 297-323. MURIE, A. 1944. The wolves of Mt. McKinley. Fauna of the National Parks of the U. S., Fauna Ser. no. 5, pp. xix+238. Payne, S. 1968. The origins of domestic sheep and goats: A reconsideration in the light of fossil evidence. Proc. Prehist. Soc, 34, pp. 368-384. Pfeffer, p. 1967. Le Mouflon de Course (Ovis ammon musimon Schreber, 1782): Position systematique, ecologie, et ethologie comparees. Mammalia, 31, pp. 1-262. Pilgrim, G. 1934. Two new species of sheep-like antelope from the Miocene of Mongolia. Amer. Mus. Nov., 716, pp. 1-29. 1939. The fossil Bovidae of India. Palaeontol. Indica (n.s.), 26, pp. iii-|-356. 1947. The evolution of the buffaloes, oxen, sheep and goats. J. Linn. Soc. London, 41, pp. 272-286. PuDENz, R. H. and C. H. Sheldon 1946. The lucite calvarium — a method for direct observation of the brain. II. Cranial trauma during brain movement. J. Neurosurg., 41, pp. 272-286. Reed, C. A. and W. Schaffer 1966. Evolutionary implications of cranial morphology in the sheep and goats (Caprini, Simpson 1945). Amer. Zool., 6, p. 565. Sadlier, R. M. F. S. 1969. Ecology of reproduction in wild and domestic mammals. Methuen, London. xii+324pp. Schaffer, W. M. 1968. Intraspecific combat and the evolution of the Caprini. Evolution, 22, pp. 817-825. Scott, J. P. 1960. Comparative social psychology. In R. H. Waters, D. A. Rethlingshafer, and W. E. Caldwell, Principles of Comparative Psychology, McGraw-Hill Book Co., New York, pp. 250-288. Sellier, K. and F. Unterharnscheidt 1963. Mechanik und Pathomorphologie der Hirnschaden nach stumpfer Gewalteinwirkung auf den Schadel. Hefte zur Unfallheilkunde, 76, pp. xi-t-140. Simpson, G. G. 1944. Tempo and mode in evolution. Columbia University Press, New York. xviii+237 pp. 1945. Principles of classification and classification of mammals. Bull. Amer. Mus. Nat. Hist., 85, pp. 1-350. 62 FIELDIANA: ZOOLOGY, VOLUME 61 SUSHKIN, P. R. 1925. The wild sheep of the Old World and their distribution. J. Mammal., 6, pp. 145-157. Teal, J. L., Jr. 1970. Domesticating the wild and woolly musk ox. National Geog., 137, no. 6, pp. 862-879. Thompson, D'A. W. 1945. On growth and form (A new edition.) Cambridge, at the University Press, x + 1116 pp. TiMOSHENKO, S. and D. H. Young 1968. Elements of strength of materials. 5th ed. D. Van Nostrand Co., Princeton, N. J. x +377 pp. Unterharnscheidt, F. and L. S. Higgens 1969. Pathomorphology of experimental head injury due to rotational accel- eration. Acta Neuropath., 12, pp. 200-204, Walther, F. 1958. Zum Kampf und Paarungsverhalten einiger Antilopen. Z. Tierpsychol., 15, pp. 340-380. 1961. Einige verhaltens Beobachtungen am Bergwild des Georg von Ope/ Frei- gehege. Georg v. Opel — Freigehege f. Tierforschung E.V., Jahrb., 1960-1961, pp. 53-89. 1964. Verhaltensstudien an der Gattung Tragelaphus De Blainville, 1916, in Gefangenschaft. Z. Tierpsychol., 21. pp. 393-467. Welles, R. E. and F. B. Welles 1961. The bighorn of Death Valley. Fauna of the National Parks of the U. S., Fauna Ser. no. 6, pp. xv+242. PLATES Plate I. d'Nesotragus moschatus, a primitive bovid skull. i Plate II. d^Rupicapra rupicapra, sectioned horn core showing only slight penetration by the frontal sinus. 64 Oh 65 Plate IV. &Capra hircus aegagrus, skull with horn sheaths showing damage due to fighting. 66 ILL Plate V. c^Capra hircus aegagrus, skull with horn sheaths removed. 67 Plate VI. c^Capra ibex sibirica, young 68 Plate VII. cfCapra ibex sibirica, young, with horn sheaths removed. 69 \ ^* / Plate VIII. d'Capra ibex nubiana, frontlet and horns of adult. 70 Plate IX. d'Capra ibex sibirica; parasagittal section of skull, showing frontal sinus and septa, position of base of horn, and primitive configuration of skull. (In this specimen, the right horn was abnormal, having grown in a tight spiral.) 71 Plate X. d'Hemitragus jemlahicus, adult. 'iP Plate XI. d'Hemitragus hylocrius, adult. 72 Plate XII. d^Hemitragus hylocrius, adult, skull with horn sheaths removed. Plate XIII. d^ Ammotragus lervia, adult. 73 Plate XIV. d'Ammotragus lervia, longitudinal section of horn core in frontal plane. y l^^iioiii' Plate XV. 9 Ammotragits lervia, adult. Plate XVI. 9 Ammotragus lervia, skull with horn sheaths removed. 74 Plate XVII. d'Pseudois nayaur, adult, showing areas of injury to the horns. Plate XVIII. (fPseudois nayaur, skull with horn sheaths removed. 75 Plate XXI. cf Om musimon, adult. 78 Plate XXII. dOvis canadensis, 5+ years. Plate XXIII. 9 Ovis canadensis, adult. 79 Plate XXI.