ECOLOGY AND BEHAVIOR OF THE CRET AN AGRIMI FOLLOWING POPULATION REDUCTION Dissertation for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY THOMAS PAUL HUSBAND 1977 LIBRARY Michigan State Umversity This is to certify that the thesis entitled ECOLOGY AND BEHAVIOR OF THE CRETAN AGRIMI FOLLOWING POPULATION REDUCTION presented by THOMAS PAUL HUSBAND has been accepted towards fulfillment of the requirements for Ph.D. degree in Fisheries & Wildlife _/g7z j/ZZIEQ, Major professor Date February 3, 1977 0-7639 ABSTRACT ECOLOGY AND BEHAVIOR OF THE CRETAN AGRIMI FOLLOWING POPULATION REDUCTION B)’ Thomas Paul Husband A study of the last remaining agrimi (Capra aegagrus cretensis Lydekker) was conducted between May and November 1975 on Theodorou Island, Crete, Greece, following the removal of 24 agrimi in 1973 and 1974. The total population of 89 agrimi was captured and revealed a male:fema1e ratio of 1.28:1. The total agrimi biomass on the 68-ha island was 1915.0 kg. A mathematical model of population growth and mortality indicated for both sexes that the number of animals (24) trapped by the authorities and removed from the island since 1973 replaced the natural mortality which would have been expected to occur during the two-year period. A vegetation analysis of the island yielded density, frequency, and percentage cover data for each of 140 plant species present. The above-ground net primary production for the combined herbaceous species plus twig and leaf growth in shrubs equaled 1061 kg/ha/yr. The measured portion of the production available as forage was 618 kg/ha/yr. A chi-square matrix employed in conjunction with four of Thomas Paul Husband Jaccard's indexes of association identified several distinct plant groupings. Agrimi appeared to cease most daytime movement from June to September. During this hot dry period agrimi heavily sheltered beneath the shrubby cover of Pistacia lentiscus. Age— and sex-specific dominance was evidenced. Mating began in mid-October, at the beginning of the winter rains. Fighting between males for estrous females was rare, probably due to the previously established male hierarchy. A seawater consumption experiment and direct observations verified that agrimi partially fulfill their free-water requirements with seawater. Continued removals of agrimi are recommended to be carried out annually to prevent overpopulation and overgrazing. ECOLOGY AND BEHAVIOR OF THE CRETAN AGRIMI FOLLOWING POPULATION REDUCTION BY Thomas Paul Husband A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Fisheries and Wildlife 1977 W © Copyright by THOMAS PAUL HUSBAND 1977 ACKNOWLEDGMENTS I wish to thank my major professor, Dr. George Petrides and the members of my guidance committee, Drs. Duane Ullrey, Peter Murphy, and Leslie Gysel for their advice concerning field work and their editorial improvements of this thesis. In Greece, I thank the Deputy Director, Mr. D. Sideridin, for his complete cooperation. I especially appreciate the kindness of the regional forester at Chania, Mr. C. Kokalin. I am also very grateful for the friendship and long hours spent in the field by Mr. George Plastiras. I also wish to give a special thanks to Phillip B. Davis for keeping the home fires burning while I was overseas. I will never forget his letters of encouragement and his reliable assistance. The identification of plant specimens was carried out quickly and efficiently by the botanical staffs at the University of Athens and the Thessaloniki University. Other logistical assistance in Greece was provided by Dr. Nicolaos Papageorgiou. At Michigan State University, I was supported by assistant- ships through the Department of Fisheries and Wildlife in the College of Agriculture and Natural Resources. Funds for this study were furnished by a Grant-in-Aid of Research from the National Science Foundation. ii Finally, I am especially indebted to my wife, Pam, who has given so much of herself and given up so much. I am also indebted to my parents with special thanks to my father, who worked long hours at my side during the capture operations. iii LIST OF TABLES . . LIST OF FIGURES . INTRODUCTION STUDY AREA METHODS . . . . . . Agrimi Capture . . . Marking Individuals . Pellet-Group Count Mortality Data . . . Habitat Temperatures Seawater Consumption Vegetation Studies Agrimi Behavior . . Statistical Procedures RESULTS AND DISCUSSION Population Analysis . Sex and Age Composition of the Population . TABLE OF CONTENTS Natality . . . . . . Mortality and Removals Biomass of the Agrimi Population Pellet- -Group Count Vegetation Analysis . Wildlife Use of Shrub Cover . Samaria Gorge Vegetation Behavioral Analysis . . Activity Patterns . Thermoregulatory Behavior . Seawater Consumption . Flight Behavior . iv Page vi vii O\ OOQO 10 11 14 IS 16 16 16 18 18 23 27 27 41 42 43 43 45 45 50 Social Behavior . Male-Male Interactions Female-Male Interactions Female-Female Interactions Female-Kid Interactions . Mating Season . Courtship Behavior . Mating Season Dominance Behavior . Female-Kid Interactions During the Rut Sexual Territories MANAGEMENT RECOMMENDATIONS Carrying Capacity . . . Management Policy . SUMMARY . . LITERATURE CITED Page 52 52 55 57 59 6O ‘64 66 66 68 68 69 71 74 LIST OF TABLES Table 1. Calculation of the expected natural mortality over a two year period for the 1973 male agrimi population on Theodorou Island, Crete, Greece . . . . . . . . 2. Calculation of the expected natural mortality over a two year period for the 1973 female agrimi popu- lation on Theodorou Island, Crete, Greece . . . . . . 3. Production data for the male agrimi on the island of Theodorou, Crete, Greece, 1975 . . . . . . . . . . 4. Production data for the female agrimi on the island of Theodorou, Crete, Greece, 1975 . . . . . . 5. The density, frequency, cover, and above-ground net production of 140 plant species found on Theodorou Island, Crete, Greece, June, 1975 . . . . . . . 6. Percentage of moisture of some major forage species of agrimi, Theodorou Island, Crete, Greece, July 1975 O O I O O O O O O O I O O I O O O O 7. Relative frequency of aggression (head gesturing, lunging, chasing, and butting only) among different age classes of both sexes. The number of inter- actions won are arranged horizontally, the number lost, vertically . . . . . . . . . . . vi Page 21 22 25 26 28 51 54 LIST OF FIGURES Figure Page 1. Outline of Crete, Greece, showing two offshore islets and with topographic relief of Theodorou and Theodoropoula Islands (Greek Forest Service) . . . . . . 4 2. Schematic drawing of trap corral used to capture agrimi on Theodorou Island, Crete, Greece . . . . . . . . 7 3. Links of plastic chain used in marking agrimi in 1975 on Theodorou Island, Crete, Greece. Position 1 indicates where the release link was cut for large animals. The neck chains of smaller animals were cut at position 2 to allow release under less pressure . . . . . . . . . . . . . . . . . . . . . . . . 8 4. Species-area curve for the vegetation present in 1975 on Theodorou Island, Crete, Greece . . . . . . . . . . . 12 5. Age and sex composition of the living agrimi population on Theodorou Island, Crete, Greece, August, 1975 . . . . l7 6. Age and sex composition of the living agrimi population on Theodorou Island, Crete, Greece: August, 1973 (Papageorgiou, 1974); 1974, without mortality or removals included; August, 1975, plus mortality (crosses) and removals (black) which occurred since 1973 . . . . . . . . . . . . . . . . . . . . . . . . . . 20 7. Growth-weight curves for the 1975 agrimi population compared to those of 1973 (Papageorgiou, 1974). Theodorou Island, Crete, Greece . . . . . . . . . . 24 8. Species diagram partially showing the association of 35 of the most important plant species found on Theodorou Island, Crete, Greece, 1975. Correlations were determined with a chi-square matrix analysis. Numbers represent species ID numbers from Table 5 . . . . 40 vii Figure Page 9. Number of agrimi seen per hour of observation at trapsite in feeding areas, and near blinds. Nighttime values are from observation at the trap, June to September, Theodorou Island, Crete, Greece, 1975. Solid and open bars represent female and male portions of activity, respectively. Black dots represent mean air temperatures (°C). Numbers above columns indicate hours of obser- vation at that time . . . . . . . . . . . . . . . . . . . 44 10. Mean temperatures t (SE) C at three shaded-bulb stations located 50 cm above ground under Pistacia lentiscus shrubs and in nearby open areas, Theodorou Island, Crete, Greece, June to September, 1975 . . . . . . . . . . . . . . . . . . . 46 11. Weight changes in four captive agrimi during a sea- water consumption trial on Theodorou Island, Crete, Greece, in 1975. Solid lines denote the periods of fresh water availability, while dotted lines indicate an ad libitum seawater regimen . . . . . . . . . . . . . . . . . . . . . . . . . 49 viii INTRODUCTION The agrimi (Capra aegagrus cretensis Lydekker) is a subspecies of wild goat which occurs on Crete. It exists in reduced numbers in the White Mountains of Crete and, in order to preserve the agrimi, Theodorou Island, a tiny islet offshore Crete, Greece, was stocked from successful introductions in 1928 (1 pair), 1937 (1 pair), and 1945 (1 pair) (Schultze-Westrum, 1963). A complete history of the island's agrimi was reviewed by Papageorgiou (1974). Other wild goats, known variously as g, aegagrus Erxleben or g, hirggg Linnaeus (Ellerman and MOrrison-Scott, 1951; Harrison, 1968), are found in parts of Turkey, Iran and Afghanistan, Iraq, in the Causasus and Turkmenia areas of Russia, and in the Baluchistan and Sind provinces of Pakistan. Behavioral information about wild goats is sparse, with only a single study conducted on the courtship behavior of the Pakistan form (C, a, blithi_Lydekker) (Schaller and Laurie, 1974). Other pub- lished observations consist only of natural history notes (Danford, 1875; Wahby, 1931; Roberts, 1967) and hunting accounts (Stockley, 1928; Maydon, 1937). Schaller and Laurie (1974) have indicated the need for concise information on the behavior of this species. The food preferences of the agrimi on Theodorou Island were determined by Papageorgiou (1972) and, in 1973, various aspects of the population's energy relations were measured (Papageorgiou, 1974). With a total population of 97 animals, Papageorgiou found the vege— tative range to be seriously deteriorated as a result of overgrazing. Following his recommendations that population numbers be reduced, the Greek Forest Service, in late 1973 and 1974, removed 24 agrimi from the Theodorou population. These animals were translocated to various local parks throughout Greece. It was intended that the removals relieve the heavy grazing pressure and enable the island habitat to stabilize and perhaps improve. The present study was undertaken (1) to ascertain population responses of the agrimi to the reduction in density, (2) to measure the density, frequency, cover, and partial biomass of the island vegetation under reduced grazing pressure, and (3) to add to our basic knowledge of the species by documenting the behavior patterns of the agrimi. STUDY AREA The study area, Theodorou Island, is a 68-ha island which lies about 850 m from the northwest coast of Crete, Greece. Com- prised of rugged limestone (Figure l), the northern and western extremes of the island consist of steep cliffs, while the southern portions of the island's hilly topography exhibit less—severe slopes. The local climate is mediterranean, with cyclonic winter rains and hot dry summers. Data from the meteorological station at Chania, 8 km to the east, indicate a mean air temperature of 18.8 C with mean annual extremes of 15.1 C and 22.6 C. Absolute extremes of 1.8 C (January, 1968) and 41.4 C (July, 1960) have been recorded there. Five 50 C alcohol thermometers placed on the island burst, however, during the summer period. The higher temperature extremes on the island were probably due to the reflectivity of the rocky limestone outcroppings and to heat absorption by the considerable exposed soils there. Although there is no severe cold season, the winter is cool. The main growing season is spring, with the hot dry summer a rela- tively dormant period for plants. Rainfall on Chania averages 691 mm per year (National Meteorologic Service, Chania, Greece), mostly falling in the winter. .nouw>uom amouom xoouou mvaufimu adsomououoosh van :ouovooch mo mowaou aficmuprQOu cud: was muodmw ouocmmmo or» mewxonm .ouoouo .ououu mo unadu30u-.~ ouamflm a .: 6:22 35 U .38 I . E3 u .mtoEtaoEoo 390660... mm as me «a a. a 23:23» 5:!»— 32252:. 6:2». «3396662.... Only two brief showers fell on the island between May and September, 1975. Two or three months of complete summer drought are common in the area. The soil of Theodorou Island consists of a highly eroded terra- rossa profile of which only the clayey, plastic, bright red subsoil horizon remains. In many places small pockets of subsoil are held in cupped areas of the jagged limestone. Theodorou Island supports a shrub-forb plant community typical of the mediterranean sclerophyllous zone (see beyond). The verte- brate fauna includes the European hare (Lepus europeus), Norway rat (Rattus norvegicus), Eleonora falcon (Falco eleonorae), Greek partridge (Perdix graeca), raven (Corvus corax), and rock martin (Hirundo rupensis). Sightings of hares and rats were quite common throughout the island. These populations have apparently increased from the uncommon sightings of tracks and feces reported in 1971 by Papageorgiou (1972). On eight occasions approximately 20 to 30 hares were seen in the agrimi corral trap during night trapping operations. METHODS Agrimi Capture The entire agrimi population was trapped between 4 July and 4 September, 1975. This was after all kids had been born and before the mating season. The trap was a 40- x 60-m fenced corral with two remotely- controlled doors (Figure 2). The area within the corral was baited daily with hay and fresh water for six weeks prior to the capture attempts. Hay and water were also provided there daily during the trapping period. Animals caught in the corral were driven through a narrow funnel into a wooden trap box attached to the west end of the trap. There they were measured, weighed and tagged. Trapping was conducted on a 24-hour basis during 3- to 4-day sessions. These efforts were followed by 3-day rest periods during which the agrimi had free access to water and feed in the trap area. In daylight hours, the trap doors were closed merely when a sufficient number of untagged animals had gathered there. Nocturnal trapping, however, was staggered over all nighttime hours so that activity patterns during the night could be learned. At least three hours were allowed to elapse between trapping sessions so as to minimize the influence of human activity on agrimi activity near the trap area. .oomcuu .oumnu .vcmeH sonovoonh :o fiefinwm chaummo on vow: Haunoo menu mo wcwzmuv ofiumEonomtu.N oeswwm noon mmoou< :14 EC? xom # mane won< Ho:::m noon mmoou< , mouw>oa flawnw< uefiueum3 During trapping operations, any animals which displayed symptoms of hypoglycemia were given an injection of glucose and imme- diately released. A11 treated animals were later observed alive. Markingrlndividuals All agrimi on Theodorou Island were captured and marked so that they could be identified individually at a distance. A 9-cm- high NASCOreg nylon tag was attached to the neck by a white plastic chain collar. Females were marked by red and males by blue neck tags, both of which carried white numerals 5 cm in length. The numbers on the tags could be read with a 30X spotting scope at distances up to about 200 m. One link of the neck chain was cut in such a way (Figure 3) that upon applying about 20-25 kg pressure a snagged adult could effect its own release. The neck chains of small animals were set ./ I Figure 3.--Links of plastic chain used in marking agrimi in 1975 on Theodorou Island, Crete, Greece. Position 1 indicates where the release link was cut for large animals. The neck chains of smaller animals were cut at position 2 to allow release under less pressure. for release under less pressure (Figure 3). The white neck chains themselves were highly visible and allowed tagged animals to be recog- nized as such at great distances. As a backup system of identification, all agrimi also were tagged in one ear with yellow or red plastic NASCOreg ear tags. Their bodies also were given a distinctive mark with a permanent black NYASOLreg dye. Each animal could be identified by neck- and ear-tags and also by various recorded dye patterns. Unusual natural markings, such as distinctive color patterns or horn configurations, also assisted in the recognition of several individuals. Written descrip- tions of each animal were always carried by each observer in the field. Pellet-Group Count A pellet-group count (Smith, 1964; Rogers et a1., 1958) was conducted to test the hypothesis that this method could provide an accurate estimate of population numbers in the future. Four S-m x SO-m plots were randomly chosen and delineated. All agrimi pellet groups present at the time the plots were designated were sprayed with red or yellow paint. Thirty days later unpainted pellet groups in the plots were counted. Mortality Data Mortality data were collected by searching the island for skulls and other durable parts of dead animals. This search was conducted in September, 1975 after trapping was concluded. 10 Since a similar search had been conducted in 1973 (Papageorgiou, 1974), the remains found were for deaths during the intervening two years. Remains which were greatly deteriorated were assumed to have occurred earlier and were added to the mortality listed by Papageorgiou (1974) as having occurred before 1973. The ages of dead animals were determined by counts of the annular horn rings. Habitat Temperatures Three stations were established to assess the relative microclimates under Pistacia lentiscus shrubs, beneath which the agrimi sheltered on Theodorou Island. Each station was under a particular shrub seen to be used often by agrimi. At each location, one shaded-bulb thermometer was placed perpendicular to and 50 cm above the ground while a second instrument was similarly placed in an open area 2 m from the shrub. The three stations were located in the east, middle, and west sectors of the island. Temperatures were read at irregular times during the day and night from 28 June to 29 September, 1975 as indicators of microclimatic differences. Seawater Consumption An experiment was conducted to indicate whether agrimi would accept seawater and whether they could maintain weight when drinking salt water alone. Two male and two female agrimi were captured and housed in individual metabolism cages, each 95 x 80 x 140 cm in size. They were fed ad libitum a diet of air-dry hay and freshwater for the first 11 seven days of captivity. On the eighth day, seawater was offered ad libitum in place of the fresh water. They were kept on the sea- water regimen until it was rejected, after which they were given fresh water. Each animal was weighed to the nearest 0.5 kg every other day throughout the experiment. Though urine and water volumes were recorded and urine samples taken, these data were discarded when they were found to have been disturbed by the Forest Service guard. Vegetation Studies Island vegetation was evaluated in May and June to assess the effects of the reduced agrimi population. Areas devoid of vege- tation or inaccessible to agrimi because of extreme slopes were not sampled. Areas which were deemed usable by agrimi were divided into 200 blocks of SO x 50 m each, using a prismatic compass and range- finder. In each of the 200 blocks, a 1/2- x 2-m2 plot was located randomly in a similar manner. This size plot was adopted because its small size enabled accurate and convenient counting and yet was not so small as to cause large errors through edge effects. Plot size was also chosen for comparability with the previous vegetation analysis of the island (Papageorgiou, 1974). A species-area curve (Figure 4) (Cain, 1938) in conjunction with a plot of running means for major species (Kershaw, 1964) indi- cated the sample size for the small density quadrats was adequate. The sampling of biomass of abundant species was stOpped at the point at which additional sampling did not substantially affect 122 .oooouo .ououu .uemHmH :oaovooch co mnmfi a“ panacea eowumuomo> on» you o>u=o moumumowoonmuu.v ouamau ce— cv— am— Au E. 8.2 a: ram yaw— rov— seuoeds jo Jeqwn N 13 the mean biomass per twig or shoot. This was tested by calculating and plotting a cumulative mean during the quadrat analysis. Less abundant species were clipped in all plots where present. Line transect (Canfield, 1941) was used with a steel wire 2 m long and marked in l-mm graduations was stretched across the center of each plot about 5 cm above the ground. Ground cover was measured as the vertical projection of the crown or shoot area of each species transected by that line. Ground surface devoid of living vegetation was recorded as rock surface, plant litter, or bare ground. The density of each species found in a plot was determined by direct count of individual plants rooted within the plot. Shrub twigs within each plot were also counted. Herbaceous plants in each plot were categorized as grazed or ungrazed and clipped during early June when most species were fruiting and appeared to be at peak standing biomass. The above-ground net production for herbaceous plants was considered to be equal to the maximum standing biomass (Bray et al., 1959; Kilmer et al., 1960; Gilbert and Chamblee, 1959; Pieper, 1964; Koelling and Kucera, 1965; Blaisdell, 1958; Kelly et al., 1974). . Annual woody growth was not measured for shrub species. Only the annual peak standing biomass of stems and leaves was taken as a partial measurement of above-ground net production. Plant species were identified in the spring, when flowering parts were present. Plants which could not be identified in the field were numbered and later taken either to the University of 14 Athens or the University of Thessaloniki for further review by plant taxonomists. Agrimi Behavior Direct observations were made with a 30X spotting scope, binoculars, and the unaided eye at distances of 3.0 m to 1.5 km. Notes were recorded in the field and later transcribed to data forms. Some behaviors were recorded with a 35 mm, single-lens reflex camera, using a 75-205 mm zoom lens, sometimes in conjunction with a 3X teleconverter. Many observations were made from a small shrine located near the main trapping area and centrally sited on the south slope. Agrimi were also observed from a small blind at a bait station near the southwest shore. Although the wild agrimi were habituated to the baited areas, any sudden movement of the observer would cause precipitate-flight. Close-range observations could have been influenced by the presence of the observer, but similar behaviors to those seen at close range also were seen at long ranges (100 m to 1.5 km). Flight distances were measured by a range finder to the nearest meter when possible. Behavioral observations were made from 5 May until 20 November, 1975. During this period, there were over 249 hours of observations at the trapsite, and 156 hours at the secondary feeding station. An additional 446 hours of observation were spent on the open range and in three blinds located strategically on hilltops. 15 Statistical Procedures Differences between all observed values were considered to be statistically significant if they could have been due to chance at under the 0.05 level. Standard deviations (SD) are presented with appropriate means unless otherwise noted. To establish correlations between species, Jaccard's (1928) index of community similarity was applied. As an index of species association, this procedure compared conditions on sample plots with respect to the: 1. Frequency with which two species occurred together. 2. Biomasses present. 3. Correlation of the degree to which each species provided ground cover. 4. Correlation of the constancy with which each species occurred throughout the series. A chi-square matrix using 2 X 2 contingency tables for two- species correlations allowed comparisons between the number of plots occupied by a given species and the number expected under a hypothesis of randomness (Agnew, 1961; Mueller-Dombois and Ellenberg, 1974). An algorithm and computer program to aid in computations of the indexes of association and the chi-square matrix data was com- pleted using a CDC 6500 digital computer. RESULTS AND DISCUSSION Population Analysis Sex and Age Composition of the Population The capture of the complete population on Theodorou Island revealed 50 males and 39 females to be present in 1975 (Figure 5) where, in 1973, Papageorgiou (1974) found that the population con- sisted of 50 males and 47 females. Between these dates, 5 males and 10 females were removed in 1973 and 3 males and 6 females in 1974. It seemed that the unbalanced (1.28:1) sex ratio in 1975 could have been a result of their capture and removal by the Greek Forest Service. By categories, the 1975 population consisted of 12 kids, 5 yearlings, and 72 adults. In 1973, there were 14 kids, 13 yearlings, and 70 adults. The percentages of kids and adults remained about the same, however, the number of yearlings was much lower. The low number of yearlings in 1975 may have been related to the poor range conditions present in 1973 (Papageorgiou, 1974). While the population reduction in 1973 may have relieved the grazing pressure somewhat, it probably did not bring an immediate improvement in the availability of forage. The low number of first-year agrimi was probably the result of high kid mortality in 1974. 16 .mAmH .um=m=< .oooonu .ouonu .vcmHmH souowoosh :o :oflumasmom «afihwm wzw>wfi on» mo :ofiufimomaoo xom was om<--.m ousmwm m_mE_c< *0 LODEDZ GD — ‘ F N F 9 F on PGD P‘ N N D v u a bI 17 I—_ I moEEom 3:22 I C) (smelt) 96v 18 In contrast, the normal crop of kids in 1975 may well have been a response to the population reduction since some time had elapsed which would have permitted range improvement. The increased amount of forage available per agrimi may have also improved the con- dition of the agrimi which allowed higher natality. Natality Twelve kids were present and captured, 5 males and 7 females. The only unaccompanied kid observed was a male about 2 months old in late July. The 11 females, each with a kid, ranged in age from 3+ to 8+ years, with a mean of 5.6 i 1.8. No twins have been observed on Theodorou Island. There were 1.67 adults per kid, 2.40 kids per yearling, and 2.78 adult females per kid. The number of breeding-age females per kid was 2.0 in 1975 and similarly 1.78 in 1973 (Papageorgiou, 1974). Thus only about half of these females have offspring in any one year which survive through the summer months. Mortality and Removals A search of Theodorou Island on 28-30 August located the remains of 4 adults and 7 kids which had died since 1973. Since the fragile skulls of kids may break down rapidly or be consumed by rodents, it is probable that the 7 skulls found underestimated the extent of kid mortality. 19 Ten more heavily decayed skulls also were found of animals that had evidently died prior to 1973 and which had been missed during the previous searches of Papageorgiou (1974). No animals were lost to emigration due to the 850—m barrier of deep seawater which separated Theodorou Island from Crete. The 24 agrimi which had been removed plus the 4 adult agrimi, which had died since 1973, accounted completely for the adults enumerated in 1973 (Figure 6). A hypothetical model (Tables 1 and 2) of the agrimi population was constructed, based on age-specific mortality data obtained by Papageorgiou (1974) as amended by 1975 mortality findings. This model was used to estimate the projected effect of natural mortality on the 1973 population over a two-year period had there been no removals. The expected numbers of males and females surviving until 1975 from the 1973 population was about 28 and 33, respectively (Tables 1 and 2). Actual survival was 31 and 41 for the 1973 females and males, respectively, for those agrimi remaining on Theodorou Island. An undetermined number of those animals removed from the population and placed in parks from 1973 to 1975 also survived. The high number of surviving agrimi indicate that the expected natural mortality between 1973 and 1975 was replaced almost completely by the removals. In fact, the 41 males alive in 1975 were signifi- cantly greater than the 33 expected to survive. The greatest differences in male survival came in the older age groups. Mortality data prior to 1973 indicate that no male lived 20 .nnmu oucfim vouuauoo snag: axoadnv maa>oaon ecu amommouuu Nufimauhoa mafia .mnaa .um:w=< “wovsgocfl mHe>oaon no Nuwuuuuoe uaonuw3 .vnmfi qnvnma .:o«mnoowummmu mnm~ .um:m=< "oooouo .ououu .vcaHmH souovooge co :ofiuafismom «sfiuwm unm>fia on» we :ofiuwmomaoo xon new ou<--.o ouauam muaawc< mo hopes: OH w o v N o N v o m oa Na OH w o v N o N v o w ca NH an m e V N o N v o m CH Nu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I o .H .N .m .v .m we .5 .w .m rod .HH NH ma mofiuaom moan: mouaeom noun: mofimaou moan: v— mnmn chad whoa (5139A) 38v 21 Table l.--Ca1culation of the expected natural mortality over a two- year period for the 1973 male agrimi population on Theodorou Island, Crete, Greece. Expected Survival Rate No. Alive in Each Age Interval Age 1“ Age Interval* 1973 1974 1975 0-1 0.75 7 -- _ 1-2 0.95 7 5.25 — 2-3 0.90 2 6.66 5.00 3-4 0.95 6 1.80 6.00 4-5 0.94 4 5.67 1.70 5-6 0.94 3 3.76 5.33 6-7 0.93 1 2.81 3.53 7-8 0.93 1 0.93 2.62 8-9 0.77 4 0.93 0.87 9-10 0.40 11 3.08 0.71 10-11 0.25 3 4.40 1.23 11-12 0.00 1 0.76 1.11 12-13 0.00 -- 0.00 0.00 Totals 50 28.10 *Data from Papageorgiou (1974) amended by 1975 findings. 22 Table 2.--Calcu1ation of the expected natural mortality over a two- year period for the 1973 female agrimi population on Theodorou Island, Crete, Greece. Expected Survival Rate No. Alive in Each Age Interval Age 1“ Age Interva1* 1973 1974 1975 0-1 0.74 7 -- __ 1-2 0.95 6 5.19 -- 2-3 0.95 5 5.70 4.93 3-4 0.94 4 4.74 5.40 4-5 0.94 5 3.78 4.47 5-6 0.94 3 4.69 3.55 6-7 0.93 4 2.81 4.40 7-8 0.86 2 3.73 2.62 8-9 0.75 6 1.71 3.20 9-10 0.56 2 4.50 1.29 10-11 0.40 1 1.11' 2.50 11-12 0.00 1 0.40 0.44 12-13 0.00 0 0.00 0.00 Totals 46 32.80 *Data from Papageorgiou (1974) amended by 1975 findings. 23 more than 10 years of age. However in 1975, 8 males older than 10 years were captured, therefore, there was an increase in the longevity of the males. Biomass of the Agrimi Population Based on the living weights of animals captured, the total biomass of the 89 agrimi was 1915.0 kg. This is 260.8 kg less than the figure reported for 97 agrimi in 1973 (Papageorgiou, 1974). If a value of 1.8 kcal/g live weight is assumed for agrimi tissue (Papageorgiou, 1974), the energy contained in the agrimi standing crop on the 89-ha island translates to 4.94 kcal/m. The 1975 data (Figure 7) indicated a sizable reduction in the 0- to 7-year old age classes. Since nutritional needs tend to be highest in the very young of all species (Church and Pond, 1974), the reduced rates of growth in the younger 1975 animals indicates that there was a probable decline in nutrients available to these animals. Such a decline between 1973 and 1975 indicates that, despite the removal of some animals, the nutrient supply of range forage probably did not improve. The productivity of a model cohort of 1000 agrimi was cal- culated from the previously-collected mortality data of Papageorgiou (1974), amended by the more recent observations and integrated with growth-weight information from 1975. The total lifetime production of male and female cohorts was 25,660 kg (Table 3) and 17,018 kg (Table 4), respectively. Secondary production for the average male and female agrimi was 3.92 and 2.56 kg per year, respectively. 24 .ouoohu .ououu .veaHmH souovoonb .nvnma .aoflmuoommmwau mnmfi mo omonu ou voummsoo :ofiuwfisnom “sauna mnad osu How mo>h:u unawozunuzouuau.n ouamflu 3.33 225.2 Co 34. a _._I .._ m m i m m a m m m .. 1v @0353 2.2. a 8.2: 2.2 q .. 83522.2 . .2 M no.2: who? a m. C Q I..— W O O O a. n a m o m o o o o a .8 w. o o 4 Au HV 0 ‘ v'N m. a w 4 4 en“ m" . ) M. a q a a a w 0 Q 4 Q Q 4 M Q :8 25 .cuufin um pnwwoz cmo2.« .mamENCM mm mo :omumdsmoa Mmu0u on» Low «use :0 women. om.mmmmN om.ooomN ooo~ Name Lance mo.mNoH oo.o co.o oo.o m~.vm oo.o m~.em on o ma ov.~mo~ oN.NN- om.n- 5.6H- mN.o- wm.vm me.o- m~.vm om om NA mv.oNNv oo.No- ov.~mn oo.omu 0N.c- ww.vm ~m.o- No.vm ANA co Lu wo.~omo vw.mafi an.no m~.NmH mm.o on.vm mn.o m~.mm me dad ed on.nm~m o».mv~- om.nm- ov.m-- om.o- on.vm oo.o- ov.vm Sm com o oo.nmoH oo.vmu oo.mu oo.Hmu o~.o- o~.mm oN.o- oo.mm on mmv m oo.nwm oo.smn~ oo.mo oo.m~n~ om.~ cm.~m ow.v o~.mn on mac 5 m~.mnoN Nn.o~mfi nv.ooH mN.o~NH NH.“ mv.mN mm.N oo.on Hm mam o owwONON ma.mmo~ on.om wN.wv0N mo.~ om.oN mm.n mN.mN on coo m co.omo oo.HmHN om.oe o~.mc~N mo.~ NN.mN ~n.m nw.vN on one Q ma.onwa ws.aonm oo.~m wn.w~N~ oo.~ mo.oN mw.~ om.~N Hm coo m om.on om.men on.mo oo.m~mm o~.N vm.n~ wm.v mn.m~ on emu. N on.om- w~.mv-~ oo.m~m~ mv.omN~H m~.n NN.w NN.¢H mn.m~ mHN has H oo.ooo~ oo.owo~ acme.“ «awo.H coca o “me x Lac 18L 4 Has ”MIL Rule “mac 3.4 Away Aumwammwuwznmw away fizv + ASS Nuaaauuoz u z somuaaaaoa x ~6>uuu:H amxmwmu ecu x ow< x om< mo fimxmwwu .x ou< Amuaoxv mommog unuaoz momaeno usage: mcw>wg u a mo ucflonvwz coozuom :oozuom unacefiuom moozuom mo unaum may x om< noduaasmom munch :ofluaunmom fiance on x om< scum muswfioz usmfioz we we usuwo: memxa .oz u< o>wa< oz . momcazo ucmfioz eofluafisnom owcmnu pcwfioz omauo>< acosouueu omeuo>< m a u m < .mhmu .ouoouu .ououu .aouovoogp mo beefinw on» :o fisfiumw cams on» you dune :oMuonvoum--.n canny 26 ucuan um ungoz :mo:«« .mHmsHem am mo :oHumHsmoa Hauou ecu you «use co vommm. No.5mmoH um.NHonH oooH mNoo mHmuoH HN.onNH NH.oN me o mm.mamH m~.mmH NN.mm Ho.ow mo.o NH.ON AN.H w.cN «a no HH mn.wnmN m.mHN om.No oo.an om.o mo.mH oo.H mm.mH mNH an oH om.nmmN mN.ommHu mN.HwN- oo.moNH- mN.N- wN.ON om.v- mm.wH mNH NwN m vN.NwNm om.cmmH ev.on ow.mHoH om.H vo.HN mm.n no.nN omH Nov m om.vco No.mHm- mm.nH- no.Hom- mv.o- om.mH mm.o- mo.aH Hm mom N hm.mHo nm.mm NH.N 0H.mw no.o nw.mH «H.o em.mH Hm can 0 nn.mmHH nH.oNnH mo.no m.NoNH Ho.H mN.wH No.N ow.mH no mNo m Hm.nnm nw.wom no.5H N.Hmw nm.o HN.NH mH.H an.NH Hm wwo v 5N.Hom mm.Nmo 0N.vH so.woo . ov.o «H.0H mm.o no.0H Hm mHn m om.Hov om.VNNH ow.vN co.ooNH ow.o om.vH oo.H c~.mH Hm own N hm.cnoH mm.mmmHH mm.oHvH No.onoH no.0 mc.n Nc.mH oH.¢H mHN Hwh . H oo.omoH oo.oon «.wo.H .ewo.H oooH c a 3 A 3 u H 3 . p . E x E E x 3 as new. a: I a: -x .a x a n mev sz + Hag eoHuaHsuom x Hm>uoueH H v cam x ou< x ow< mo H v ax om< NuHHauuo: an quHuuuo: u z x mom< x mou< Hmumoxu momma; unuHox moweagu usuHo: ucH>HH u a mo ueHoavH: :oozuo eooauom ucHecHuom eoozuo , mo unaum oau x ow< 53388 :36... 83398.. ~33 3 x a? a2". 3 use sari «6 3 3.36: was .uz 2 632 .oz noucunu usuHor :oHuuHsmom mucuAU usuHo: omuuw>u ueoauuocH oueuo>< . m a u m < .mNmH .ouoouo .ououu .aonovoosu mo wedeH on» :o HaHuua oHdaom oz» pom even eoHuusvoumuu.e oHnaH 27 Differences in age-specific weights for 1973 and 1975 were not significantly different for either sex. Based again on the value of 1.8 kcal/g live weight, the average gross energy per year fixed by each agrimi in the hypothetical population was calculated to be about 5832 kcal. Pellet-Groungount A pellet-group count revealed that the 89 agrimi deficate at a rate of 13.02 i 1.80 pellet groups per day. Future pellet-group counts computed with this value could be used to estimate agrimi numbers. Vegetation Analysis Vegetation provides the agrimi with food, cover, and some moisture. Vegetation, or the lack of it, is also a major factor influencing erosion and ground temperatures. The vegetation of Theodorou Island comprised a mediterranean "disclimax" community, degraded by grazing and possibly fire, as evidenced by numerous charred remains of shrubs. One hundred and forty plant species (Table 5) were tallied. Only 120 of these, however, have yet been identified. The density, above-ground net production and extent of crown cover were determined for each of the 132 herbaceous species. Density, cover, and the net production of twig and leaf annual growth were measured for the 8 shrub species (Table 5). A large portion of the island was devoid of vegetation. Bare ground, rock surface and dead plant litter made up 12.70, 19.69, and 28 U cam.H cam.H «no.6 m.o~ coco OH mHaamH35 «geese mm6.~ mmo.~ o~o.o o.a omm m .am esHouHu «Ho.H «Ho.H moo.o m.o~ omAHH m .am mHaoHa swo.o HHo.o o.~ omwe A .am ssHumouaoo oHH.c ooo.o o.m comm o asuafiHmpm-s==HH escHHoaoum< swm.- mom.o o.o~ omHo m «EHUHHAE «HHfium ”mommMHu wad mnHom mmc.eom moo.¢ow mHA.mm Homofimuoommooe ”mmamzm 3onw can .no>oo .xoeoavohm .qumsov 0:9--.m oHnmb 29 mmv.o ma4.o ooo.o o.NH oONo mN .am oNNUHeoz owa.Nq NHN.o c.4H omNN NN NHm.c oeo.o o.HH comm NN .am msHeon Nvo.o Nvo.o 4oo.o o.NH ova SN aeHuHaaa moauapxam Nmm.o Nnm.o ooo.o o.NH coon mN moeHoHoHa aaHusuoamoua eeo.e 44o.4 ooo.o o.oH OONH 4N maHHomHusua mamauaam< NNH.o NNH.o coo.o o.m omHH mN .mm 65H6H eme.o vm4.o oNo.o m.mH oomN NN .am esoaxmnme NnH.o ooo.o o.mH comm HN .am msaoum owo.o owo.o ooo.o m.e omvH ON maHaa ssHmaHa oom.vH cow.eH NHH.o o.HN OONNH NH ssHHNHaouoHa esHuosme NNN.o woo.o o.NH omee NH asuaHHon asHHomHHa am4.o mm6.o Noo.o m.HN omNm NH «Nausea maoHHN6< mmo.o mmo.o ooo.o m.6 coNH 6H .am m9H6H oNH.o oNH.o Noo.o m.oH ooeN mH anamomaaau «HHusz Noo.o mHo.o m.v ooNH 4H «Heme «HHHH mmo.o amo.c ooo.o m.m com NH aouaHHoH asHHOH wwo.o mmo.o moo.o o.N cmNH NH espamuHaasm asHHH< wnm.o mmm.o coo.o o.m oomH HH msmoHHH>H=m mausHaHoum ammmw Emma 2H... may”. emu“. Hm. .eosaHuaou--.m 6Hnae 30 ooN.c ooN.o oNc.o m.Ho ooomm cc .cm mcaouc Hcm.c Hom.c Noo.o o.om ocoNH No acaccom aochmHHe cHo.m oHo.m NNo.o c.4o ocooN co xHHoHca «HoccHo> moo.H coo.H Noo.c o.NN ccoN mo oomNoech mHHNocco moo.H Nco.c o.Hm ocHNo co .om mosch oom.c occ.o m.cm cmoNH mo ochHH moocuocum com.c oom.c oco.c o.mc coocN No accuse eoHHcNHHH NcH.c NoH.c ooc.c m.o omc H4 chHHcmH «Hoocox omN.H omN.H moo.o o.NN cmco co oHHcHHHmccoH «Hoocom moo.N moo.N «HH.o c.co ochN om .om oocumcm mNN.H mNN.H oHc.c m.om ooNom om mcmccccch moHocomcoch ccm.o occ.o c.oo cmmmH Nm eoNHcocH aoHcHHHmoo omm.o oom.c occ.o m.NN omco om mmo.c mmc.c ooc.o o.oN ooHo mm eoHcoo ssHHNchH oNH.c oNH.o coc.o o.o occH cm .H mHm=6>Hc mHaocoo< NcH.H NmH.H cHo.o c.HH comm mm oHHcaHo> aoHHHz NoH.o NoH.c ccc.o c.NH omNm Nm mcHHH> oHchm HNo.o HNo.c cmo.c m.oN cmmHH Hm cocHHomooo mHHNoucHo< NoH.c NHo.o m.oH omNN cm .cm oHHoccmHHc .mmewemfim1mfiecmmm Ame. HH .ee ommuom poz .occoHocoo--.m oHoce 31 cHo.c cHo.c oco.c m.m cmo oN NccooHcHHu mooaoooooamz oco.c omm.o o.o o co HcHHHH mooH oco.c HcN.NH o.o c mo cocomm ammo oco.c ooo.oH o.o c No Hccu cacc com.N com.N oco.c o.Nm cocoN co coHoHocc mHHHcmc=< Nmm.o Nmm.o oco.c o.om omHHH mo cocHHcH mHHHHoch mom.N mom.N oco.c o.om cmmHm co mccHHHco coHHm Noc.H Nco.H Nco.o m.om ochN No .H ecococmmom eoHHooHHe moo.H moo.H oco.c o.om omNmN Ho .H «mecca mHHHHoH NNo.H NNc.H cHo.c m.Ho comNH cm .H coHooHo c>ch oom.NN oom.NN ooN.o m.Nc cmmooo om HoHccconHcec cmocooHa omN.NN omN.NN cmo.o o.oo oommom cm .om cocoomHo omm.o Hoc.o m.om ooNom Nm scocHHon-acHoHH ooeHHoHoom< omc.N cmo.N NHc.o m.Hm coNoo om cocHHommHHHoo cHucoHH HmN.m HmN.m oco.c o.Nc cmNHN mm oco.c oco.c oco.c c.Hm ooHo mm HoN.c HoN.c NHc.o c.co omoom Nm ccoonoo mHooHc Hmm.H Hmc.H HHo.o m.mo ooooc Hm cocoa» oHHoooHuouo moo.H moo.H oco.c m.co cmooc cm mHmoooooc mopoocuooxz .mmmom ammamw :mmm. mmmmw Home. .o .gge oumnom uoz .cooaHocco--.m oHcce 32 NcN.c NoN.o oco.c o.H cmcH cc .mmcc mHHochs coHcoNo ch.o «Nc.o oco.c o.H coN mm mHHHmooou mHHccocc mHH.c moN.H m.c cm No «Hosea moch HNN.o oco.c m.oH comN om ooouceaco chocmcoz mmc.c Hoc.o o.o omoH mo oooccHuuHo mcoucoocosmm ch.m oNN.o c.NN ocmNH cm mcHHHHa oomcoocce< ocN.o oco.c m.N comoN mm ocm.c oco.c o.o och No .om mosey coH.c oco.c m.m omc Hm moccmcuo monoccHHoco cmH.o cmH.o Hoc c c.m coco cm .om somooHaoo Nmo.o mNc.o m.c cmmm oN oNo.o oco.c c.m omoN cN coHHcmom cccHoc ocN.o oco.c o.o omHo NN oumHHcm cmeHHm oNc.o cHo.c m.HH ommN 6N .H aooocco soHHac mNH.H cHo.c m.cH omooH mN mooHcoHaou choomoH omH.o Nco.o m.oH chmH 4N moocch xc>m Nco.o Nco.o oco.c o.m cch mN .om mHacoo oco.c oco.c m.m omo NN cacao mooHHc camcoccamo oNo.c moc.c m.N omNN HN mHmoo>Ho chuouocm .mmmm fimmm 2mm... Home Home H.“ .cocoHcoou--.m oHooH 33 U cHo.c oco.c m.o om on mpmuaeeoo «Huonx cHo.c oco.c o.o om NoH .cm ogoococao cHo.c oco.c m.o om ooH .mm magooom ooH.c oco.c m.c om moH .om mH26cH=< cHo.c oco.c m.o om voH .mm monuomm mcc.o moo.o oco.c m.o om moH omccHsooHo coHoHomc cmH.c omH.o oco.c m.H omH NoH mocHoHUHm esHooommcHo moo.o oco.c m.N omm HoH omooapuo> caucHoooo CON.o oco.c m.o coH ooH .H szHucono aseHnqun< ooH.c oco.c o.H ooH oo «HHomHoooo «HoH> oco.c oco.c oco.c m.o om co .om «Hucocm oom.o oco.c m.N com No mHHmnmaHooH mHHxnoccmxza cmo.o oco.c m.o om co aouoouo ochcmccc ooH.c oco.c m.o ooH mm enumeHonmo nomuouou moo.o oco.c m.o ooH co .nm asHHomHuc oco.c oco.c m.c coo mo .om museum moo.o Hoc.o o.N och No .Hmoz mHHcHHHmcu ou>< moo.o oco.c m.o om Ho .m.m amouauoum «HaHouHoz cHo.c oco.c m.o omm co cH>c=oHc omcuech HHN\cc\NHo Hum\oc\mxo . . gamma 5%... 2mm... mmmm scam”. H... .coccHuoco--.m oHooe 34 oNc.o oco.c o.H ooN mNH Hoc.o Hoc.o oco.c o.o cm NNH mocHocom oooHHm cHo.c oco.c m.o cm oNH Hoc.o Hoc.o oco.c m.H ooN mNH Hoc.o oco.c o.o cm cNH Hoc.o oco.c m.H ooN mNH cHo.c cHo.c oco.c m.c cm NNH cHo.c omo.c m.c ooH HNH .om szHcoaxcuoHc mcc.o oco.c m.c cm oNH .om mmcoc< oNc.o mNo.o oco.c m.o omN oHH .om ooecHoNc cmo.o mmc.o oco.c o.H omN NHH .cm ccmcocmcue Nmo.c oco.c o.o omN oHH .om mHoouo oNc.o oco.c m.c ooH mHH .om cccomoa mmo.o oco.c o.N omm oHH .om cococoHo Nco.o Nco.o oco.c m.o cmm mHH .om «HchcNHoom cHo.c cHo.c oco.c m.c om NHH moHoHcHo chcccuHca ooH.c ooH.c oco.c o.H ocH HHH .om «mucosa ooH.c oco.c o.H cmH oHH cmcomom choHco oNc.o oco.c m.c ccH ooH .cm aoHoon Home fine”. 2mm... meme .mmmmc H... omenom uoz .coooHccco--.m oHocH 35 Hoc.o oco.c m.o ooH NvH Hoo.c ooc.o m.o om HvH Hoc.o oco.c m.o om ovH moo.o oco.c o.H coHN mmH Hoc.o oco.c m.o om me ONo.o oco.c c.H CON hmH Nco.o oco.c o.H CON omH Hoc.o oco.c m.o om mmH ooH.c ooH.c Nco.o o.H ooH va moo.o oco.c m.H omH mnH cHo.c oco.c m.c om NmH CON.o CON.o oco.c m.o ooH HnH oom.o occ.o m.o omv omH Nco.o Noo.c coo.o m.c cm mNH .mmmmm ammmww :mmc. mmmmw memes .H .mge ouuuom uoz .cocoHcoco--.m oHocH 36 .mono>HnNo: oaonH sonoooosh on» x: oHnMESmcoo :oHuosooum no: mo :oHuuom nacho .mnounm mo sazoum mooH com szo Hoseco mon mommoum woo mnHom mo mmoEoHn wcheoumnxoomn Hmunon mo Honaacv macaw mo Honssz u mnounw mucon mo Honasz n mommouw one mnuomo nmN.wHo HHo.HooH ooo.ooH mHou Noooscoum oqumooa 9H moHoomm omouom uoz .cocoHHoco--.m «Hoop 37 9.54 percent, respectively, of the island surface. Thus, about 42 percent of the agrimi habitat offered no useful vegetative forage or cover. The total above-ground parts of herbs plus the twig and stem annual growth of shrubs weighed 1061.01 kg/ha when air dried. With 2.17 plants per m2, Poterium spinosum, a small spiny shrub, had the greatest density. Where it occurred, the small aromatic shrub, Thymus capitatus, attained the second highest density at about 1.36 plants per m2. Pistacia lentiscus, a large evergreen shrub, and Poterium spinosum contributed the most cover at 17.24 and 15.14 percent of all ground cover, respectively. Pistacia lentiscus, Poterium spinosum, and Thymus capitatus comprised 57.5 percent of the measured biomass, with individual values totaling 315.09, 161.26, and 134.13 kg/ha, respectively. The portion of the measured production which consisted of species showing evidence of being eaten and which was available as edible forage for the island's herbivores totaled 618.73 kg/ha/year. This total includes forage from 59 species and is 44 percent higher than the value of 429.50 kg/ha/year reported by Papageorgiou (1974). The 1973 measurement, however, was based only on ten plant species. Papageorgiou (1974) considered forage production to equal the sum of the mean number of twigs or plants per plot times the mean dry weight per twig or plant. If this criterion is used, the forage production for the six major agrimi forage species (Pistacia lentiscus, Phlomis fruticosa, Colycotome villosa, Scilla maritima, Cistus incanus, 38 and Olea oleaster) rose from about 363 kg/ha/year in 1973 (Papageorgiou, 1974) to about 522 kg/ha/year in 1975. This substantial improvement in forage production was probably the result of the reduced grazing pres— sure by agrimi. A study was made to determine if the island's vegetation was a homogeneous mixture of species. Analysis revealed, however, that the island's plant species grew in some interesting patterns and groupings. Among the island's dominant shrub species (Table 5), Poterium spinosum and Thymus capitatus occurred together in a significant positive correlation, yet both showed significant negative correlations with the third major shrub, Pistacia lentiscus. The woody species, Cistus incanus, was negatively correlated with all three shrub species. The spiny shrub, Colycotome villosa, and Phlomis fruticosa were found not to be significantly correlated with any of the other shrub species, and only significantly correlated with five herbaceous species. All shrub species, except Colycotome villosa, showed significant relationships with the presence of and amount of surface of bare ground or rock. Phlomis fruticosa, however, was positively correlated with plant litter, a finding which agrees with the field observation that this species generally grows in less-eroded areas where some plant litter is present. The indexes of association for cover for Thymus capitatus and Poterium spinosum with nearby bare ground were very high at 0.590 and 0.717, respectively. The chi-square values comparing these two species with bare ground and rock surface were also significant. 39 These two species largely occurred in areas of high erosion and denudation, on the less-sloping central areas of the island. The presence of Euphorbia paralias was found to be signifi- cantly related to the presence of Pistacia lentiscus, and negatively related to the presence of Poterium spinosum. P, lentiscus did but E, paralias did not show a significant relationship with I, capitatus. Euphorbia paralias occurred almost exclusively on and was a major constituent of the vegetation on steeply-sloping areas on the periphery of the island. It showed a significantly negative relation- ship with the presence of nearby rock surface and bare ground. Pistacia lentiscus was found to be significantly related to bare ground surface and surface covered by plant litter. It also showed a significant negative relationship with the presence of nearby rock surface. There was high degree of association between many species pairs. To illustrate, 23 of the 35 highly-associated species were diagrammed (Figure 8). In this diagram, Valantia murulix (46) appears to be highly associated with all but two of the species con- sidered. This major grouping primarily occurred on the less-sloping portions of the island. The highly-abundant shrub, Poterium spinosum, was found to be a member of the group in Figure 8 and was significantly related with 16 other members. It may, therefore, be discerned that both shrubs and herbs may be included in one "ecological group" (Whittaker, 1967), growing together in the same habitat and showing closely similar distribution patterns. 40 64 62 65 \ 66 61 76 60 . \ \ . ‘..\ ‘- 79 59 . ‘ N ' , . . l - \. \ ’ ~ 53 , - .4 58 ' x ’ b \ _/ I r . 4s _ 7| . ‘ . I 57 I \ \ H- I 86 I ; ‘ 56 H — \ . ' 47 55 \ . 48 52 51 ._,.H 49 5° — P<0.005 -'- P< 0.010 ‘- P<0.050 Figure 8.—-Species diagram partially showing the association of 35 of the most important plant species found on Theodorou Island, Crete, Greece, 1975. Correlations were determined with a chi-square matrix analysis. Numbers represent species ID numbers from Table 5. 41 Wildlife Use of Shrub Cover About 150 plants of each of the six major shrub species on Theodorou Island were randomly chosen and evaluated for their use by wildlife as cover after Gysel and Lemmien (1955). Pistacia lentiscus shrubs exhibited an average height of 96 i 40 cm and width of 588 t 321 cm. The 32 percent of P, lentiscus shrubs that were utilized by agrimi, as evidenced by pellet groups and hair, had an average height of 140 i 25 cm and width of 888 i 265 cm. The dense crown of this evergreen shrub provided a cover under which agrimi could remain relatively cool (see beyond) during the hot dry season. Evidence of utilization by agrimi was also found under 44 percent of the Euphorbia paralias shrubs sampled. This species is deciduous and is leafless between late June and September, the hottest and driest season. Thus E, paralias does not provide leafy cover when it is needed most by the agrimi. The mean height and width of g, paralias used by agrimi as cover were 145 i 22 cm and 170 i 38 cm, respectively. About 63 percent of the Colycotome villosa shrubs sampled had hare fecal pellets under them. Rat pellets were also evident under about 10 percent of these shrubs. Poterium spinosum provided a dense leaf cover for small mam- mals, however, due to its characteristic inverted-cone shape it offered only marginal protective cover from ambient conditions and possibly predatory birds. 42 Thymus capitatus was heavily utilized by hares and rats with pellet groups of each animal under almost every shrub sampled. The structure of these shrubs was such that a 5- to lO-cm space existed between the ground and the dense crown, which provided a natural cavity for hares and rodents. Olea oleaster formed a dense thicket of branches which was impenetrable by small mammals. Samaria Gorge Vegetation A field trip was made to the agrimi's original range in the Samaria Gorge National Park in order to examine the vegetation of the region and possibly to observe agrimi. The narrow gorge itself is 17 km long and is located in the White Mountain range of Crete, about 60 km from Theodorou Island. The tree layer in the mountains around the gorge consisted of sparsely scattered junipers (Juniperus phoenicia), pines (Pinus brutia), and oaks (Quercus ilex). The principal shrubs were Quercus ilex, Satureia thybra, Phlomis lanata, Salvia sp., Verbasium spinosum, Olea oleaster, Colycotome villosa and Pistacia lentiscus. The most common herbs were Carlina sp., Brira media, Trifolium aureum, Dactylis glomerata, Scilla maritima, Asparaggs acutifolius, Andropogon hirtius, Cirsium sp., Cupressis sempervirons, Urospecurium picroides and Cistus cretica. The vegetation within the park boundaries did not appear to be heavily grazed in recent years, but areas immediately outside the park were heavily cropped by domestic sheep and goats. 43 All but one of the herbaceous plants and four of the shrubby plants found also occur on Theodorou Island. All of these species showed a lack of positive association with other plant species on Theodorou Island, and are not included in the dominant group of 35 species depicted in Figure 8. 0n Theodorou Island, most of these species are heavily grazed and appear to be decreasing in abundance. It seems likely that these species are remnants of a different flora which once occupied the island. Due to pressures exerted by local residents, the park officials of Samaria Gorge National Park have allowed several herds of domestic goats to graze in the park. According to park officials these herds totaled about 500 to 1000 animals. Although park rules forbid the grazing of domestic animals, the practice was passively accepted in 1975. This activity, if continued, will certainly detract from the available forage supply, and, at worst, may allow disastrous interbreeding between domestic goats and wild agrimi. Behavioral Analysis Activity Patterns Agrimi displayed distinctive seasonal and daily activity patterns. During the cool season, from October to May, agrimi on Theodorou Island were active both day and night, as evidenced by observations and trap data. In the hot dry summer months of June to September, however, they almost eliminated daytime movements (Figure 9). 44 .osHu pony uo :oHuo>uomoo mo mason ouooHocH massHoo o>ono muonaoz .Hoov monouonomaou HHo :ooe unomopmon muov xoon .xHo>HuoommoH .qu>Huoo mo mcoHunom oHoe use oHoaom anemonmon when ammo one oHHom .man .oooohu .ououu .ononH sonoooonh .uonfioumom op meow .mouu on» yo :oHuo>uomno sonm one mosHo> oEHuu:MHz .mveHHn moo: one .moono mcHooom :H ouHmmonu no :oHuo>Nomno mo noon you :oom HeHumo mo Honssz-u.m whamHm 3:55 .30 Ho 68:. :00: vu NN oN a— o. v. N— :— efd. . 5.. .IIIIIIIIJI ; m. rN \. a. woc—. % v. em: .mu r 1,6 m c He as. ON r n H... 1m “me. .a o 3 m1 1.: was. .SWm. S m. mm mm. m u. av .NH mum. m me. fiv— amg rm— 45 Following a peak just after sunset, there was a steady decrease in activity during the summer nights and a pronounced drop in movements at sunrise. Agrimi in summer largely spent the daylight hours in the shade of a rock or, more commonly, under the cover of a Pistacia lentiscus shrub (Figure 9). On at least ten occasions, agrimi under the shrub were seen to feed on forbs and grasses or on new shoots of the shrub itself. On five occasions, an agrimi was seen to spend the entire daytime under a single P, lentiscus shrub, exiting only at sunset when 'ambient temperatures had decreased substantially. Thermoregulatory Behavior Agrimi evidently used the dense shade of Pistacia lentiscus shrubs in an effort to keep cool. Daytime temperatures measured under three shrubs used by them were significantly lower and exhibited significantly less variation than did nearby ambient temperatures (Figure 10). An agrimi exposed to the high ambient temperatures at worst might incur a lethal heat load and at best would metabolize energy at a higher rate in order to maintain constant body temperature. Seawater Consumption Accelerated loss of body-moisture also seems likely to be an important factor influencing the activity pattern of the agrimi. Free water is almost nonexistent on the island except for small quantities made available from a 4- x 6-m cistern which collects ‘46 .mNoH .uonfiounow on 6:35 .ouoouo .ououu .ooonH souoooonh .moouo coma Apnoea cH one mnounm moomHucoH «HooumHa noon: ocoouu o>opo so on oouoooH mcoHuoum anonoooonm eons» yo 0 Hmmu A monououomaou :oo:--.oH ouomHm 3505 Non Ho 25... coo: onNMNHNmNazHHchHo—mHNm :2 D D P P I D b b P ’6! boo nu be In bv M N P— c L L l l 5 (as) : (0,)891n1medu191 ueew Ego EN 32o :30 c. D L 3:on on. : .1335. as z: 47 rainwater from the shrine roof in the winter. Water from the cistern is piped to troughs in the trap area on an irregular basis. During the 1975 study period, the water from the cistern was totally expended by mid-August. After this, agrimi apparently relied upon free-water in plants, metabolic water, and seawater. Historically, mariners often introduced goats to remote inhospitable islands as a potential source of fresh meat for ship— wrecked sailors. Though scientists and other travelers have been amazed that these animals have survived on small arid islands with no permanent sources of freshwater for drinking (Eibl-Bibesfeldt, 1961; Napier, 1929; Roughley, 1966; Williams and Rudge, 1969), the only evident explanation is that the goat, which is presumably the domesticated form of the agrimi, can utilize seawater. They have been seen drinking from the sea. This is true particularly on the Galapagos Islands (Bibl-Eibesfeldt, 1961; Nelson, 1968; Slevin, 1931), but also on Esk Island of the Palm group off the east coast of Australia (Dunson, 1974). Dunson believed that such behavior also could be associated with algal feeding, or with the imbibition of only token amounts of seawater by sodium-deficient animals. But goats may indeed have the capacity for renal salt excretion that is efficient enough to enable them to regularly drink seawater. Papageorgiou (personal communication) on at least two occasions saw agrimi on Theodorou drinking seawater. During the present study, this behavior was observed on 5 June, 15 July, and 23 August. On each occasion, an animal was seen to drink from a small pool of sea- water near the ocean. An inspection revealed that algae were not 48 present in these pools. It also seems unlikely that the food plants of these goats were sodium-deficient in view of their proximity to the sea. An experiment was conducted to establish whether or not caged agrimi would voluntarily consume seawater and if they could maintain weight balance drinking only seawater. All specimens lost weight during their first four days of captivity, though fresh water was available and was used. Their weights tended to hold steady, however, from the fifth through the seventh days (Figure 11). On the eighth day, when seawater was offered ad libitum instead of fresh water, all four agrimi readily accepted it, drinking quantities comparable to the amounts of fresh water normally taken. Over seven days, the mean fresh water consumption for the two females had been 1469 i 233 and 1104 i 180 cc/day, respectively, while over the ensuing 5 to 9 days, seawater consumption was slightly higher at 1504 i 17 and 1424 i 295 cc/day, respectively. Fresh water consumption by the two males averaged 1416 i 311 and 1669 i 356 cc/day, while sea- water consumption again was higher at 1707 i 84 and 1835 i 86 cc/day, respectively. This increase was probably due to the insatiability of the agrimi while on the seawater regimen. After 5 to 9 days (Figure 11), the agrimi refused both air dry forage (of only 2-3% moisture) and seawater. To avoid any pos- sibility of permanent damage, each animal then was returned to a diet of fresh water and dry hay for several days. They were then released and observed to be well. 49 .soEHmon nouozoom saanHH om no ouooHoeH mosHH oouuoo oHch .NHHHHnoHHo>o mono: amoum mo mooHuom on» ouonoo mocHH oHHom .man :H .oooouu .ououu .vaonH zonoooonh :o HoHuu :oHHQESmaoo Houozoom o wcHuoo HaHnwo o>Humoo Noom :H momcono ungozuu.HH oHSMHm Am>mnvarr chNoNNNcNHHHHHHNmmHmmwmcc \\ .2 .NH (6)1)1u619M SIBLUIUV U.I II." I II- I III. 50 All specimens lost considerable body weight while drinking seawater (Figure 11), but regained some of their original bulk after fresh water became available. No diarrhea was observed. Dunson (1974) determined that although ferral goats would voluntarily drink seawater, body weights could not be maintained where their foods contained only 9-13 percent moisture. Goats living on Esk Island (Palm group, east coast of Australia) in the dry season were found to maintain themselves well on food plants of 52-67 percent water content if seawater also was available. Major food plants on Theodorou Island ranged from about 41 to 73 percent moisture (Table 6). Thus it seems likely that agrimi on the island probably gain enough free water from food plants with occasional seawater consumption, to maintain a positive water balance in the dry season. The cistern on the island provides a supplemental source of fresh water, however, its presence may not be necessary to the agrimi's survival. Flight Behavior Agrimi on the open range stared at a source of alarm while in a frozen standing position with mane held somewhat erect. The ears usually were rotated in an effort to appraise any disturbing sounds. While in this posture, agrimi sometimes gave a deep snort. Disturbances precipitated varying responses which seemed to be related to the nearness of the disturbance. If a human was spotted approaching at 300 m or more, an agrimi would walk away, occasionally stopping to gaze back. If disturbed within a lOO-m range, agrimi tended to run. 51 Table 6.--Percentage of moisture of some major forage species of agrimi, Theodorou Island, Crete, Greece, July 1975. Species Approximate % Moisture of Grazed Plants Pistacia lentiscus 55.63 Colycotome villosa 52.15 Phlomis fruticosa 41.09 Poterium spinosum 48.62 Cistus incanus 72.62 In most cases, agrimi ran up steep hillsides in a zig-zag manner. Agrimi used both speed and climbing ability to elude a pursuer. The alarm and flight behavior of agrimi startled from under the cover of Pistacia lentiscus shrubs was quite different from that exhibited on the open range. Agrimi under cover had shorter flight distances. These varied according to circumstances but averaged 14 i 11 m for males and 48 i 30 m for females. Sheltered agrimi tended to sit tight until the flight distance was met. They then would bolt from their cryptic shrub cover and usually run on a straight course directly away from the intruder. The actions of a female with kid when frightened on the open range was similar to that of all females. The kid followed close behind her. When flight was effected from under a shrub, however, a mother agrimi would usually bolt and run at an angle away from the source of the disturbance. The kid usually remained quietly behind and only took flight when approached at a distance of about 3 m. The 52 adult's action is probably an adaptation to draw danger away from the more-vulnerable young animal. A similar response was noted in the pronghorn (Antilocapra america) by Autenrieth and Fichter (1975). Social Behavior Most social behaviors were observed during trapping sessions and at feeding stations. Although the agrimi may have been assembled unnaturally under these circumstances, their groupings would occur apart from baited areas during the mating season or at isolated water sources. Most of the year, agrimi stayed as solitary individuals. Animals of similar age and sex were observed feeding and moving together mainly during trapping and feeding Operations. Two to five agrimi were seen to travel together across most of the island's length (0.8 km) to the trap area, keeping about 5 m from each other. Several groups which frequented the trap area had relatively constant membership, but neither their place of origin nor their social organization upon departure could be determined. One daytime observation was of five males which had arrived at the trap together but departed individually in different directions and at different times. Other groups observed similarly may just come together by coincidence. .Male-Male Interactions Interactions observed in the field and at feeding stations between males indicated a linear hierarchy which was dependent on the 53 ages of the animals involved (Table 7). Older, and consequently larger, males always appeared to be dominant. Observations at the feeding stations and at trapsite indicated that the oldest males always dominated areas where food and water were concentrated. Within the trap corral when older males were present, many younger males and females entered to feed only after long hesitation. Even after they entered, subdominant animals fed along the perimeter fences and drank only when all of the more- dominant males were at least 5 m from water. A typical sequence of events occurred whenever a subdominant male encountered an older and more-dominant male at the feeding station. The latter animal usually initiated the interaction by vocalizing a "growl." If this did not induce the subdominant to move away, a head gesture and sometimes a short successful charge by the dominant animal usually followed. The dominant males' body position during such an interaction was usually standing erect on all four legs, either with the tail held horizontally back or in a lowered drooping position. His ears usually were erect. The dominant male occasionally initiated the interaction from a head-down feeding position. A subdominant male in the vicinity (5-10 m) of an older male exhibited a submissive posture to a degree which seemed to be dependent upon the net differences in their ages. Submissiveness usually was indicated by the ears held back and the tail between the legs. Subdominant males crouched their bodies only on the rare occasions when they approached (less than 2 m) an older male. 54 .oHoaom a macro» o>Hmmouwuo ons I m: .oHoa nonuooo whorou o>Hmmoummo oHos - :2 .oHoa o ouozou o>Hmmouwmo oHoEom 1 2m .oHoaom Hoguoco ouozou o>Hmmouwwo oHoaom I mm mm Hm N N m N n H NH 0 m nH n a v N H N m H mHouob m n H o H o H Huo N N o o o o o o o o N NIH m nH o o H N o H o o H o o v N N mun NH 0H o N c o o o c o .N N o N N o o v H H Nnm 0 OH 0 c c o o o o o n o o H H m o H o H o o muN oH mN H v H o o N o o w H c N N o o N o H o o o c o o HHum HH wH o H o o o o H o v n o o N o o o . H o o o o N o n o H o o nHuHH Hat? mo 238%? "H: o: E E a: i E mm a: .2.an n: "H: .2 E "a "H: z: x... cm "Hz .2 E "E a: H! E "E H383 6: H36... No .62 H36... H-o m-H mm Tm o-N HH-o mH-HH «ma .HHHooHuuo> .umoH Hoses: 0:» .NHHoueoNHno: concouno one :03 mcoHuoououeH mo hoses: one .moxom noon mo mommoHo ouo ucbuommHo macaw HNHco ucfiuuso woo .mchono .mchcsH .mcHunumow voonv :onmonumo mo Noooodoum o>HuoHozuu.N oHnoh 55 Males of the same age appeared to be mutually tolerant of each other most of the year. A distance of about 2 m usually was maintained between codominant males. When one codominant male approached another too closely, the latter usually made a low growl to warn the encroach- ing animal away. In the trap area and at the bait station, the older males dispersed rather evenly over the prime feeding area and effectively maintained a "dominance barrier" against less-dominant animals. Younger subdominants either fed along border areas or waited until the more-dominant animals had left. Female-Male Interactions Females were usually observed moving alone or with their kid on the open range. Females were only rarely observed with a group of males or with another female. Females which approached the trap area where males were feeding usually did so cautiously with bodies in a somewhat lowered posture and tails tucked firmly between their legs. Females only entered the trap area when all males were at least 7 m from the entrance. The females then darted through the entrance and fed in remote corners of the corral. Regardless of the age or relative size of the female, all were observed to be totally subdominant in encounters with any males at least one year of age. Males exhibited aggressiveness toward females with head and horn gestures. Such a gesture consisted of a stationary animal dipping its head and horns in the direction of another animal. In most cases, a gesture was enough to send the female out of the S6 immediate area. If the initial gestures failed, a step or short charge in the direction of the female usually induced the latter to depart. Actual aggressive contact between a male and a female was observed on only one occasion, when a female failed to heed the signals of a large male. The male charged, knocked the female down with his horns, and literally ran over her. After the interaction, the male went back to feeding. The female slowly arose and left the feeding area. The only display of aggression by a female toward a male was observed when an 8-year-old female did not yield to the aggressive gestures of a smaller 2-year-old male. The two animals rose on their back legs to a standing position and then fell toward each other, clashing their horns. The female subsequently turned and fled. The male pursued her for about 7 m and then returned to feeding. On an occasion in September, a male kid exhibited aggressive behavior toward a 7-year-old female by butting her with his head. The female responded by butting the kid. The kid then fled from the pursuing female to its nearby mother. Thereupon, the single female broke off pursuit and began feeding again. Trap data (Figure 9) indicate that the femalezmale ratio increases through the early morning hours. Many female agrimi probably avoid encounters with more-dominant males by feeding in the morning after the males have eaten and departed. 57 Female-Female Interactions Females on the Open range were mostly observed traveling alone or with their kid. At the feeding area and trapsite, females exhibited a social dominance hierarchy based on age similar to that shown among the males (Table 7). The status was maintained by aggressive actions in the form of head gestures, chases, or butting. An aggressive action only observed among females was a "stomping" gesture. A female would raise both her front feet about 10 cm above the ground and stomp them both down simultaneously. This behavior was usually accompanied by a loud snort. Stamping behavior also was exhibited by female agrimi in metabolism cages, as an aggressive action toward their caretaker. There was only one observation of actual butting between females. Both individuals were without kid and were approximately of the same age and weight. The bout consisted of two butts with the horns, after which one of the females rapidly left the feeding area. Female-Kid Interactions "The urge to follow" has been recognized in young birds (Hess, 1962; Hinde, 1961; 1970; Thorpe, 1961; Tinbergen et al., 1965; Van der Kloot, 1968). It has also been described by Portman (1961: 149) as occurring in ungulates immediately after birth. Young wilde- beests have been noted to walk and run closely beside their mothers as early as 4-5 minutes after birth (Talbot and Talbot, 1963:46). The distinction between approach and following in ungulates seems to be unclear and may depend on the activity of the mother 58 (Autenrieth and Fichter, 1975:30). In early summer (May-July), shortly after the birth season, an approach behavior was commonly observed. For example, when a female stopped moving, the kid would approach sometimes to attempt nursing. More often, it would just stop and look around until the mother moved on. By late fall (October- November), kid and mother would graze together, 5-10 m apart, with the mother setting the general direction of movement. This then appeared to be following behavior rather than the approach pattern observed earlier. It was not until 3 October, that two kids were seen running about 10 m ahead of their mothers as they approached the feeding station. Nursing was not observed on the open range but several (15) bouts took place at the trapsite. In every case, the kid approached its mother from one side or from between the front legs, and butted at the udder with its muzzle until a nipple was found. No tail wagging or kicking by the kid was observed during nursing. Nursing was usually brief, lasting only about 3-5 minutes. Kids of various ages at the trapsite usually consumed hay both prior to and after nursing. A kid with a dried umbilical cord still attached was observed to consume bait at trapsite in early July. No nursing was observed after 25 September. The agrimi mothers role during nursing was largely passive: she stood in a normal upright position and might have even continued feeding if food was in reach. There was no hunching of the back as in the pronghorn (Autenrieth and Fichter, 1975:58), nor was there any 59 sign of active participation as shown by the kudu cow which, according to Walther (1964:455), directs the calf to the udder by licking. Females were seen to stop a nursing bout by making a rapid lurch forward and then walking away from the kid, or by checking the kid with her head. No kids were observed feeding from females other than their mothers. When a kid approached another female, it was usually repelled by aggressive head gestures and brief charges. During the 15 bouts seen, no female allowed a kid other than her own to nurse. This was unlike the pronghorn which may allow several alien young to approach and feed (Autenrieth and Fichter, 1975). MatingLSeason The breeding season of ungulates, like that of most species, is usually timed so that the extra energy needed for reproduction is available and so that the young are born in a suitable time of the year. On Theodorou Island, the first substantial rains fell in the early October after a five-month dry period. Daytime temperatures dropped significantly at this time. Within two days of these first rains the island's vegetation became greener as a few species started vegetative growth. Fresh water became readily available to agrimi in numerous rock basins. In 1975, the agrimi mating season began in mid-October, 5 to 10 days after the first rains. Its beginning was signalled by anatomical, as well as physiological and behavioral changes. In males, an 8-10 cm high gray mane along the neck and back became prominent. 60 The breeding period of native ungulates varies with the climate and nutrition of the region (Fraser, 1968). Many times the time of mating is related to the rainy season (Talbot et al., 1965). Proper timing of kidding is essential for survival of the agrimi in its original range throughout the White Mountains of Crete. Mating there in late fall assures that kids will be dropped in early spring, after the harsh snowy winter and when chances of survival are greatest. It may be a factor, too, that the sexual drive of agrimi, like that of domestic buffalo (Hafez, 1953) and cattle (Hafez and Lindsay, 1965), also may be suppressed by the heat of summer. The main mating season of domestic goats on Crete is in August. The difference in peak mating season between domestic goats and agrimi may reduce the potential for cross-breeding. This danger remains, however, as exemplified by the obviously hybridized goat population on the island of Dias near Iraklion, Crete, where agrimi were introduced without the complete elimination of domestic stock. Courtship_Behavior At other times, males usually carried their tails in a hanging position. During the rut, however, the tails of adult males were held upright or stiffly horizontal. Fraser (1968) thinks that these positions may help to dissipate odors from the tail glands. The vertical tail position was common for adult males when in the presence of a female in heat. Young males near adult females, however, displayed erect tails only on a few occasions, probably due to their being subordinated by nearby older males. 61 On five occasions, male agrimi were observed to lower their hind quarters and to unsheath an erect penis. Three times this was in the presence of females. On two of these occasions, the male appeared to muzzle or lick the penis. In two other instances, the agrimi male urinated on the ground. No directional spraying of urine on any part of the body was observed, as reported for the wild goat of Pakistan by Schaller and Laurie (1974). During the first week of October, the male dominance over females was pronounced, but during the second week of the month there was a transition to courtship behavior. Then, the males followed females at a distance of 3-5 m. When a female stopped, usually the attendant male also halted and maintained a head-up standing position with body erect and tail vertical. Alternatively, the male approached the female with his ears back, and the mane usually bristled. The latter approach was observed only toward females and never toward males, as reported by Schaller and Laurie (1974) for the Pakistan wild goat. It has been suggested (Schaller and Laurie, 1974) that such displays "are a means of showing off, or trying to impress another animal, whether it be male or female." It is possible that such behavior helps to establish dominance between two males. However, it would seem unlikely that there is adaptive significance in agrimi males "showing off" or trying to "impress" other males when a social hierarchy, which is established by aggression, determines which animals mate. 62 After a male approached the female in an upright position or in the low-stretch, he usually sniffed and muzzled the females perineal region, often wagging his tail. During this behavior, the female usually moved away or stood motionless for a few seconds before walking a short distance. Females also wagged their tails. Usually the male followed the female and repeated his behavior. On four occasions the male did not follow, whence the female circled and passed near (0.5 m) the male, wagging her tail. In each such instance, this behavior rekindled the male's interest in her. These females always passed near the same male even when several other males were present. During a typical low-stretch approach by a male, the female would remain still. The male would raise his muzzle and curl his lips up in flehmen behavior. This posture was maintained for a few seconds or repeated several times. A male was often observed to direct this behavior toward a group of several females when they were congregated at the feeding station. Vaginal sniffing by the male was not seen to elicit urination in agrimi females as has been observed in the Pakistan wild goat (Schaller and Laurie, 1974) and domestic goat (Shank, 1972). Though recorded (Schaller and Laurie, 1974) for another wild goat population, kicking, twisting, and shoving were not observed in Theodorou Island agrimi. Schaller and Laurie (1974) observed females poking male wild goats with their horns during courtship. No agrimi females exhibited such actions toward males during the mating season. 63 In late October and early November, rut behavior became intense. Single males were observed to follow single females over the island for an entire day. Still later, sniffing and quiet following was then replaced by running "rape chases." Males were observed to stand alone or in small groups on the high ridges and hilltops of the island. When a female was seen, a chase would begin. 0n 8 November at 1504 hours, for example, a female was seen to leave the cover of a Pistacia lentiscus shrub. Immediately a male ran from a nearby ridge and began to chase her. The female tried to elude him by running in a circuitous manner at what appeared to be her maximum speed, and then darted into E, lentiscus shrubs. The male followed only centimeters behind the female. Often his head touched her rump. On several occasions, the female actually entered the shrubs, whence the male usually stopped and stood alertly outside for a few moments. Sometimes, he pursued the female into the shrub. The male appeared to mount the female briefly while both were in the shrub. The female then bolted and was further pursued by the male. This pattern was repeated and the chase was observed to con- tinue over 1.5 km until both animals were out of sight. Only one male agrimi was observed to chase a female at any given time, but displays and chases were made by more than one male with respect to individual females. Similar behavior noted by Hediger (1955) for the chamois (Rupicapra rupicapra) involved "brutal" chasing of females. Ibex males, on the other hand (Fraser, 1968), showed a mute and motionless 64 courting behavior in which, having climbed to the top of a look-out point, the male made typical rutting noises in the direction of the females. The only agrimi males seen to participate in courting and mating were in the 8-13 year age group. No following, chasing or mounts were attempted by younger males. With this age-specific pattern of mating, males must be at least six years old in order possibly to contribute their genes to the population. A single observation was made of a male attempting to mount a male kid. This lZ-year-old male followed the kid and attempted to mount him, but the kid immediately ended the interaction by running to his mother about 15 m away. On 7 November a large kid was observed attempting to mount his mother. This behavior was preceded and followed by the kid rubbing his body against the female‘s side.‘ Only 49 minutes later this same kid attempted to nurse its mother, but she avoided him by rapidly walking away. Mounting of this kind is known to occur also in very young lambs, pigs, calves and kids (Fraser, 1968). Collias (1956) found that both sexes of domestic goats exhibited full patterns of sexual behavior before they were fully grown. Mating Season Dominance Behavior During the rut, changes in dominance behavior patterns were noted at the feeding station. Older males, ordinarily tolerant of each other, contested for dominance during the mating season. Social status was established by aggressive displays and by actual combat. 65 Aggressive displays between older males consisted of head and horn gestures. Occasionally a brief charge would be successful, however, this usually precipitated actual combat. Butting might occur between two males standing side by side with their heads close. They would twist their torsos and lower their horns to butt each other. Fighting also was initiated between two males of similar size and stature when they came close to each other, particularly if an estrous female was nearby. During combat, both males would stand erect on their hind feet and fall toward each other, clashing their horns loudly. These violent clashes were repeated in rapid succession, until one of the males left the immediate area. Generally one or two con- tacts ended the conflict but, on one occasion, two ll-year-old males were observed to make contact five times in succession. Most (87%) combats occurred prior to mating season or during the first week of the rut. During the main rutting period, aggressive behavior in the presence of a female was at a minimum. This was like the Alpine ibex (Nievergelt, 1967), wherein males established rank by fighting before the rut and underwent a minimum of strife during the rut. Combat behavior observed in the agrimi was in contrast to that of the Pakistan wild goat, where dominance displays and actual combat occurred only during the rut (Schaller and Laurie, 1974). Why the agrimi and its close Pakistani relative should be dissimilar is not clear. 66 Female-Kid Interactions Duringthe Rut The rut appeared to be a time when agrimi mothers and kids ended the parent-child relation and became permanently separated. This was observed on two occasions when a female with a kid was chased by a male. The kids in both instances were unable to keep up with the adult agrimi and became separated from their mothers. During the rut, kids were observed on hilltops and ridges bloating for their mothers. In one instance a female, which was known to have become separated from her kid the previous day, bloated loudly as she wandered near the feeding station. During the rut, at least five females were seen to become separated from their kids. All of these kids were later seen grazing alone on the open range. Sexual Territories Signposting and marking to indicate sexual territory has been widely recognized in Cervidae (Graf, 1956; Hennig, 1962; Marshall, 1956). These animals often de-bark trees to give evidence of territorial claims and also may secrete various olfactory signs of identity. Large males were observed on six different occasions during the rut vigorously rubbing pine trees (Pinus brutia) and shrubs (Pistacia lentiscus and Euphorbia paralias) with their horns. Such actions broke limbs and removed bark. Throughout the island, many E, paralias shrubs were noted to have been de-barked. Since it did not occur near the ground, this de-barking was not the result of rodent activity and must be attributable to agrimi. 67 Whether territorial boundaries are ever enforced could not be determined with certainty. On several occasions during the rut, however, as many as five male agrimi were observed in close proximity to each other on ridge tops and displayed no evident aggressive attitudes. MANAGEMENT RECOMMENDATIONS CarryingCapacity Theodorou Island forage production of all plant species which showed signs of agrimi feeding was about 618 kg/ha/year or 250 kcal/ m2/year. If agrimi require about 773,106 kcal/year as determined by Papageorgiou (1974), it is possible that the island could have sup- ported a maximum of 162 average-sized animals in 1975. This value may be somewhat high, since some of the forage if not eaten at the most favorable time becomes unavailable due to drying and hardening of protective spines. It is also possible that some requirement other than energy may be in short supply, such as protein, which may vary in availability throughout the year. If the general rule for temperate-zone management is applicable (Stoddart et al., 1955), then 50 percent of both the vegetative forage and its reproductive parts may be removed without permanent damage to the range. On this basis, and if agrimi can maintain normal health on their present diets, then a maximum of 80 average-sized animals or about 1500 to 1700 kg of biomass could be maintained on the range indefinitely. Population levels should be maintained, however, that allow for range improvement to a fully-restored undisturbed condition. At 68 69 the proper level, the population would have a high level of repro- duction, good physical condition, and an increased average longevity. Management Policy Management policies directed toward relieving pressure on the island's forage resources in 1973 and 1974 has resulted in the removal mainly of breeding-age females. Since one male may mate with several females, males could be removed in larger numbers than females. Males should not be reduced too drastically, however, because their genetic variability may be important to the population. Post-reproductive females consume nutrients and make no further contribution to the population's gene pool. Females over age 9 might well be removed. Large-scale vegetative manipulation should only be carried out after thorough trials on small test areas. Consideration of the "ecological group" concept should be made if the plant habitat is to be manipulated. Plant species of the same ecological group (see page 39) probably have similar ecological requirements and may even be partially dependent on others in the group. In consequence, an attempt to remove an unwanted species such as Thymus capitatus or Poterium spinosum might create unfavorable conditions for the entire ecological group, including desirable grass and forb species. It must be remembered though, that no vegetative manipulation can be expected to be successful unless feeding pressure on the range by agrimi is alleviated on a sustained basis. If the agrimi on Theodorou are to have a secure future, affirmative action must be taken. Additional biologists should be 70 added to the local Forest Service staff to enable continuation of the program of habitat and population analysis. Specific duties of such personnel would be (1) to construct and maintain exclosures and maintain study transects so as continually to assess range trends and habitat manipulation techniques, (2) to conduct an annual population census and make necessary removals, (3) to conduct a rodent control program which will not endanger agrimi, (4) to build runoff aprons and collection tanks or otherwise to provide adequate fresh water for the agrimi on a year-around basis, and (5) to monitor environmental factors which might affect the population. Agrimi in excess of the island's carrying capacity could be translocated to various parks and zoos, thereby providing a symbol of Greece for display throughout the nation and elsewhere in the world. SUMMARY A study of the agrimi (Capra aegagrus cretensis Lydekker) was conducted between May and November 1975 on Theodorou Island, Crete, Greece, following the removal of 24 agrimi in 1973 and 1974. The study was undertaken (l) to ascertain population responses of the agrimi to the removals, (2) to measure the response of the island's vegetation to the reduced grazing pressure, and (3) to document the behavior patterns of the agrimi. The total population of 89 agrimi was captured and revealed a male:fema1e ratio of 1.28:1. This unbalanced ratio was the result of the capture and removal of more female than male agrimi. Based on living weights of the 89 agrimi captured, the total population biomass equaled 1915 kg. The 1975 data indicated a sizable reduction in the mean weights of the 0- to 7-year-old age classes. A mathematical model of population mortality indicated for both sexes that the number of animals trapped by the Greek Forest Service and removed from Theodorou Island since 1973 equaled the natural mortality which would have been expected to occur during the two-year period. The greatest differences in male survival came in the older age groups, where there was an increase in longevity. Mortality data prior to 1973 indicate that no male lived more than ten years, however in 1975, eight males older than ten years were captured. 71 72 A vegetation analysis of the island yielded density, frequency, and percentage cover data for each of the 140 plant species present. With 2.17 plants per m2, Poterium spinosum, a small spiny shrub, had the greatest density. Thymus capitatus, a small aromatic shrub, attained the second highest density at about 1.36 plants per m2. Pistacia lentiscus, a large evergreen shrub, and Poterium spinosum contributed the most to cover at 17.24 and 15.14 percent of all ground cover, respectively. About 42 percent of the agrimi habitat was devoid of all vegetation. The above-ground net primary production for the combined herbaceous species plus annual leaf and twig growth in shrubs equaled 1061 kg/ha/yr. The measured portion of the production avail- able as forage was 618 kg/ha/yr. The forage production for six major agrimi forage species rose from about 363 kg/ha/yr in 1973 to about 522 kg/ha/yr in 1975. This substantial improvement was probably the result of the reduced grazing pressure by agrimi. A chi-square matrix employed in conjunction with four of Jaccard's indexes of association identified several distinct plant groupings. The most interesting was a group of 35 highly-associated species, which included the island's most numerous plant, Poterium spinosum. Agrimi appeared to cease most daytime movement from June to September. During this hot dry period, agrimi heavily sheltered beneath the dense cover of Pistacia lentiscus to stay cool. Three temperature stations indicated that the daytime temperatures under 73 these shrubs were significantly less and were much less variable than were ambient temperatures. An age- and sex-specific dominance was evidenced with old males being the most dominant segment of the population. This hier- archy among males was established prior to the rut, probably to minimize strife during the rut. Mating began in mid-October, at the beginning of the winter rains. A sequence of mating behaviors were observed which increased in intensity throughout the rut. A seawater consumption experiment and direct observations verified that agrimi partially fulfill their free-water requirements with seawater. The suggested carrying capacity for the 68-ha island was about 80 average-sized animals or about 1500 to 1700 kg of biomass. Management policy for these agrimi should be directed toward relieving pressure on the island's forage resources through planned annual removals of agrimi. LITERATURE C ITED LITERATURE CITED Agnew, A. D. Q. 1961. The ecology of Juncus effusus L. in North Wales. J. Ecol. 49(1):83-102. Autenrieth, R. E., and E. Fichter. 1975. On the behavior and social- ization of pronghorn fawns. Wildl. Monogr. No. 42. 111 p. Blaisdell, J. P. 1958. Seasonal development and yield of native plants on the upper Snake River plains and their relation to certain climatic factors. U.S. Dept. Agr. Tech. Bull. 1190. 68 p. Bray, J. R.; D. 0. Lawrence; and L. C. Pearson. 1959. Primary pro- duction in some Minnesota terrestrial communities for 1957. Oikos 10:38-49. ‘ Cain, S. A. 1938. The species-area curve. Amer. Midl. Nat. 19(2): 578-581. Canfield, R. H. 1941. Application of the line interception method in sampling range vegetation. J. Forestry 39:388-394. Church, D. C., and W. G. Pond. 1974. Basic Animal Nutrition and Feeding. Albany Printing Co., Albany, Or. 300 p. Collias, N. E. 1956. The analysis of socialization in sheep and goats. Ecology 37(2):228-239. Danford, C. G. 1875. Notes on the wild goat Capra aegagrus. Proc. 2001. Soc. Lond. Pp. 458-468. Dunson, W. A. 1974. Some aspects of salt and water balance of feral goats from arid islands. Am. J. Physiol. 226(3):662-669. Eibl-Eibesfeldt, I. 1961. Galapagos-Noah's Ark of the Pacific. Doubleday, Garden City, N. Y. 192 p. Ellerman, J., and T. Morrison-Scott. 1951. Checklist of Palaearctic and Indian mammals, 1758 to 1946. British Museum, London. 810 p. Fraser, A. F. 1968. Reproductive Behavior in Ungulates. Academic Press, New York, N. Y. 202 p. 74 75 Gilbert, W. 8., and D. S. Chamblee. 1959. Effect of depth of water table on yield of ladino clover, orchard grass, and tall foscuo. Agron. J. 51:547-550. Graf, W. 1956. Territorialism in deer. J. Mammal. 37(2):165-170. Gysel, L. W., and W. Lemmien. 1955. The growth and wildlife use of planted shrubs and trees at the W. K. Kellogg multiple use forest. Quart. Bull. Mich. Ag. Exp. Sta., Michigan State Univ. 38(1):139-145. Hafez, E. S. E. 1953. Patterns in the mating behavior in the domestic buffalo. Acta. 2001. 34:233—239. Hafez, E. S. E., and D. R. Lindsay. 1965. Behavioral responses in farm animals and their relevance to research techniques. Anim. Breed. Abstr. 33:1—16. Harrison, D. L. 1968. The Mammals of Arabia. Ernst Benn, London. 192 p. Hediger, H. 1955. Studies of the Psychology and Behavior of Captive Animals in Zoos and Circuses. Butterworth, London. 430 p. Hennig, R. 1962. Uber das revierverhalten de rehbache. Z. Jadwiss. 8:61-81. Hess, E. H. 1968. Imprinting and the "critical period" concept. Ip_Roots of Behavior, E. L. Bliss (ed.). Harper and Bros., New York, N. Y. Pp. 254-263. Hinde, R. A. 1970. Animal Behavior (2nd ed.). McGraw-Hill Book Co., New York, N. Y. 876 p. Jaccard, P. 1928. Die statistisch-floristische Methode als Grundlago der Planzensoziologie. Ip_Abderhalden, Handb. biol. Arboitsmeth. 11:165-202. Kelly, J. M.; P. A. Opstrup; J. 8. Olson; S. I. Auerbach; and G. M. Van Dyne. 1969. Models of seasonal primary productivity in eastern Tennessee Festuca and Andro on ecosystems. ORNL- 4310 (Oak Ridge National Laboratory, Oak Ridge, Tenn.). 296 p. Kershaw, K. A. 1964. Quantitative and Dynamic Ecology. Edward Arnold Publishing Co., Ltd., London. 183 p. Kilmer, V. J.; O. L. Bonnet; V. F. Stahly; and D. R. Timmons. 1960. Yield and mineral composition of eight forage species grown at four levels of soil moisture. Agrnon. J. 52:282-285. 76 Koelling, M. R., and C. L. Kucera. 1965. Dry matter losses and mineral leaching in bluestem standing crop and litter. Ecology 46:529-532. Marshall, F. H. A. 1956. The breeding season. In Marshall's Physiology of Reproduction, A. S. Parkes (38.), vol. 1, part 1, Chap. 1, pp. 1-42. Longmans, Green, London. Maydon, H. 1937. Big Game of India. Phillip Allen, London. 445 p. Mueller-Dombois, D., and H. Ellenberg. 1974. Aims and Methods of Vegetation Ecology. John Wiley 8 Sons, Inc., New York, N. Y. 547 p. Napier, S. E. 1929. On the Barrier Reef. Angus G Robertson, Sydney. Pp. 38-39. Nelson, 8. 1968. Galapagos-Islands of Birds. Morrow, New York, N. Y. Pp. 239-240. , Nievergelt, B. 1967. Die zusammensetzung der gruppen beim alpenstein- bock. Z. Sfiugetierkundo 32:129-140. Papageorgiou, N. 1972. Food preferences and survival of the agrimi (Capra aegagpps cretica) on Crete. M.S. Thesis, Michigan State University. 49 p. 1974. Population energy relationships of the agrimi (Capra pggpgrus cretica) on Theodorou Island, Greece. Ph.D. Dissertation, Michigan State University. 83 p. Portman, A. 1961. Animals as Social Beings. The Viking Press, New York, N. Y. 249 p. Roberts, 0.; 0. Julandor; and W. L. Robinette. 1958. Pellet-group counts for deer census and range-use index. J. Wildl. Manage. 22(2):193-199. Roughley, T. C. 1966. Wonders of the Great Barrier Reef. World Publ., New York, N. Y. Pp. 265-266. Schaller, G. B., and A. Laurie. 1974. Courtship behavior of the wild goat. Z. Saugotierkunde 39(2):115-127. Schultzo-Westrum, T. 1963. Die wildziegen der agaischen Islen. Saug. Mitt., 4:145-182. Shank, C. 1972. Some aspects of social behavior in a population of feral goats (Capra hircus L.). 2. Tierpsychol. 30:488-528. 77 Slevin, J. R. 1931. Log of the schooner "Academy." On a voyage of scientific research to the Galapagos Islands 1905-1906. Occasional Papers Calif. Acad. Sci. 17:1-162. Smith, A. D. 1964. Defecation rates of mule deer. J. Wildl. Manage. 28(3):435-444. Stockley, C. 1928. Shikar. Constable and Co., London. Stoddart, L. A.; A. D. Smith; and T. W. Box. 1955. Range Management. McGraw—Hill, New York, N. Y. 532 p. Talbot, L. M., and M. H. Talbot. 1963. The wildobeest in western Masailand, East Africa. Wildl. Monogr. No. 12:1-88. Talbot, L. M.; W. J. A. Payne; H. P. Ledger; L. D. Verdcourt; and M. H. Talbot. 1965. The meat production potential of wild animals in Africa. Tech. Commun. Commonw. Bur. Anim. Breed. Genet. No. 16. Thorpe, W. H. 1951. The definition of terms used in animal behavior studies. Bull. Anim. Behav. No. 9:34-41. Tinbergen, N., and the editors of Life. 1965. Animal Behavior. Time-Life Books, Chicago, 111. 200 p. Van der Kloot, W. G. 1968. Behavior. Holt, Rinehart and Winston, Chicago, 111. 166 p. Whittaker, R. H. 1972. Gradient analysis of vegetation. Biol. Rev. 42:207-264. Wahby, A. 1931. Vie et moeurs dos Capra aegagpps (Pallas) dos Mts. Taurus (region d'Alaya). Archiv. Zool. Italiano 16:545-549. Walther, F. 1964. Horned animals-their fights and their weapons. "The Blue Book," No. 8. Pp. 1-5. Williams, G. R., and M. R. Rudge. 1969. A population study of feral goats (Capra hircus L.) from McCauloy Island, New Zealand. Proc. New Zealand Ecol. Soc. 16:17-28.