Fig. 24.The approximate length of stay of each known occupant of a den that was examined six times in the winter of 1955-1956 at the Damm Farm. Most of the occupants used the den as a temporary shelter and sought permanent hibernacula elsewhere. One turtle left the den for approximately two weeks and then returned to it for the rest of the winter. The temperature of the air outside the den (A) and the average body temperature of turtles in the den (B) are given at the bottom of the diagram for each date the den was examined. The symbol "J" represents a juvenal turtle.In general, body temperatures approximated the temperature of the soil around the turtle. Body temperatures tended to be slightly higher than soil temperatures in November and December but were slightly lower than soil temperatures in the months of February and March. The lowest body temperature recorded for any turtle thatsurvived a winter was 2.7 degrees, taken from an adult female on December 26, 1955. Body temperatures one to three degrees higher were common in the coldest part of the winter. Turtles in shallow hibernacula, like those observed in wooded areas at the Reservation, are probably subjected to freezing temperatures at least for short periods but I have no records of body temperatures this low, except where they were induced experimentally. Turtles exposed to temperatures of zero degrees or slightly lower would retain enough heat to survive without freezing for a period of several hours or even a day if well insulated. A temperature gradient exists within the body; cloacal temperatures, for example, differ from temperatures deep in the colon and temperatures in the dorsal and ventral parts of the body cavity (taken by manipulating the bulb of the thermometer while it was in the colon) differ from one another. Probably, therefore, some parts of some turtles—probably the top of the shell or the extremities—freeze in winter without causing the death of the turtle. Ewing (1939:91) found a female ofT. carolina, just emerging from hibernation, that had lost some scutes from its carapace; he found the missing scutes in the hibernaculum and attributed their loss to severe temperatures in the winter of 1933-34.The incidence of mortality due to freezing is unknown for most species of reptiles. The observations of Bailey (1948) on DeKay snakes (Storeria dekayi) and Legler and Fitch (1957) on collared lizards suggest that rates of mortality are high in dormant reptiles. Bailey (op. cit.) suggested that winter mortality might act as a natural check on snake populations. Neill (1948a:114) thought more box turtles (T. carolina) were killed in Georgia by cold weather in late autumn than "… by all other factors together," and that this winter mortality acted as an effective check on population levels. Neill reported that many turtles left their burrows in late autumn and began to forage; if the temperature dropped suddenly, the turtles became "… too torpid to dig" and froze.If ornate box turtles are occasionally caught in the open by a sudden cooling of air temperature, it would occur at a time of year when temperatures would approximate freezing but would drop not far below this level; laboratory and field records show that adults could probably survive these low temperatures overnight and warm up sufficiently on the following day to seek adequate shelter. Box turtles deepening their burrows in winter do so at body temperatures somewhat lower than 10 degrees (near the minimum temperature at which co-ordinated activity was observedin the laboratory); turtles found in the open in late October were known to burrow into the ground at body temperatures of approximately 15 degrees.Emergence from hibernation usually occurs in April but in some years a few turtles may emerge as early as the first week of March. Emergence is stimulated by temperature and humidity. Fitch (1956b:438) stated that emergence was delayed until "…the ground has been sufficiently moistened and until air temperatures have reached at least 26°." Box turtles at the Reservation emerged on April 21 in 1954 and from April 16 to 17 in 1955. William R. Brecheisen found recently emerged box turtles in Anderson County on April 2, 1955, and March 6, 1956.Turtles were found facing upward in their hibernacula in early March. As the temperature of the soil rises, they move slowly upward, usually following the route by which they entered. They remain just below the surface of the soil for a week or two before actually emerging; this final phase of emergence is probably hastened by spring rains that soften the soil. Activity may be sporadic after emergence if the weather is cold.A number of box turtles at the Reservation emerged in a cold rain in 1954 when the temperatures of the air and ground were 16 and 13 degrees, respectively, but remained inactive for several days afterward. In 1955 the air and ground temperatures were higher (28° and 17°, respectively) on the day of emergence and box turtles became active almost immediately.DIETPublished information on the food ofT. ornataconsists of a few miscellaneous observations. Cahn (1937:103) opened five stomachs that contained partly digested vegetable matter but no insects or other animal food: Ortenburger and Freeman (1930:187) noted that grasshoppers were a main part of the diet ofT. ornatain Oklahoma and that turtles displayed unsuspected agility in catching them. Those authors also saw turtles eating caterpillars and robber flies. Strecker (1908:79) stated that "The natural diet of this species consists of vegetable matter and earthworms." Norris and Zweifel (1950:3) observed the feeding habits of captiveT. o. luteola. Coyote melon (Cucurbita foetidissima) was eaten with reluctance but a collared lizard (Crotaphytus collaris) was quickly devoured. Tadpoles ofScaphiopus hammondiwere caught in a small pool and eaten. Adults of the same species were rejected after being caught; box turtles were seen wiping their mouthsafter rejecting adult toads. The authors suggested thatT. o. luteolais an important predator ofScaphiopus hammondi, since the two species occur together in many areas and the emergence of both is controlled to a large extent by rainfall. One individual ofluteolawas seen eating a dead box turtle on a road.Captive individuals ofT. ornata, observed in the present study, ate nearly every kind of animal and vegetable food given to them. Table scraps, consisting chiefly of greens, various fruits and vegetables, meat, and cooked potatoes, formed the main diet of turtles kept in outdoor cages.A number of persons have told me of ornate box turtles eating the succulent stems and leaves, and the fruits of various garden plants; similar incidents probably occur in areas of native vegetation. J. Knox Jones told me he saw an individual ofT. ornataeating a spiderwort (Tradescantiasp.) in Cherry County, Nebraska.Sight-records of foods eaten by box turtles at the Damm Farm (excluding the many records of individuals foraging in dung or eating mulberries) were for grasshoppers, caterpillars, and various kinds of carrion. Box turtles were often seen eating grasshoppers on roads in early morning; Sophia Damm told me of frequently seeing individuals catching grasshoppers in her garden. Ralph J. Donahue told me that on his farm in Bates County, Missouri, an individual ofT. ornatamade a circuit of the lawn each morning in summer and ate all the cicadas (Magicicada septendecim) found.Vertebrate remains found in the stomachs of box turtles seem to result chiefly from the ingestion of carrion. One box turtle ate a white egg (unidentified) that had fallen from a nest and another was seen with a blue down feather clinging to its mouth. Several colleagues have told me of box turtles eating small mammals caught in snap-traps and Marr (1944:489) reported a similar incident. J. Knox Jones told me he once found an ornate box turtle in the nest of a blue-winged teal in Cherry County, Nebraska; the three eggs in the nest had been broken. The only authentic record of an ornate box turtle preying on a vertebrate under natural conditions was one supplied by Ralph J. Donahue who saw an adult catch and eat one of a brood of bobwhite quail. In many areas where box turtles are abundant, it is the opinion of local residents that the turtles decimate populations of upland game birds by eating the eggs and young of these birds; these opinions result probably from rare encounters such as the one described by Donahue. I believe that box turtles at the Damm Farm were sometimes able to catch young frogs and tadpoles (chieflyRana catesbeianaandR. pipiens)at the margins of ponds. In autumn literally thousands of youngRanawere present in these places.Ornate box turtles ordinarily attempt to catch and, without further examination, to eat, small objects moving on the ground, but are more critical of stationary objects. Captive turtles, for example, would immediately chase and seize a grape that was pulled or rolled slowly across a floor but a stationary grape was examined and then smelled before it was eaten. Similar observations were made a number of times with living and dead insects in the field and in the laboratory. A turtle discovering an object that is of possible value as food, approaches it closely, turns the head from side to side (presumably using the eyes alternately to examine the object), and then, with head cocked at a slight angle, momentarily presses the nostrils against the object (Pl. 28, Fig. 4). If acceptable as food, the object is then swallowed whole or taken into the mouth with a series of bites; large insects are usually broken into several pieces in the process of being bitten and swallowed. Larger objects, such as dead vertebrates, are torn to pieces with the beak and forefeet before they are swallowed. Hatchlings, when fed for the first time, ignored inanimate foods but eagerly chased mealworms, catching them usually by the anterior end. The tendency of the young of certain species of turtles (especially captives) to be more carnivorous than adults is probably due to the association of movement with food; recognition of inanimate objects as food is presumably learned by older individuals.Mulberries (Morus rubra), when they are abundant, constitute all or an important part of the diet of ornate box turtles. On June 4, 1955, William R. Brecheisen and I drove along a road in Anderson County, Kansas, and stopped at each mulberry tree that we saw beside the road; we found at least one specimen ofT. ornataunder nearly every tree. Approximately twenty box turtles were collected in this manner in a little more than one hour. The heads and necks of most were stained dark-red from the fruit and, in some, nearly the entire shell was stained. Dissection of these turtles revealed that their stomachs were distended to two or three times normal size with mulberries; no other kinds of food were found in the stomachs. Some of the turtles voided purplish-black fluid from the cloaca when we handled them; the color of the fluid presumably resulted from mulberries.Several turtles were observed through binoculars as they foraged. Individuals snapped or lunged periodically at objects on the ground along the route of travel. Upon reaching an area where cow dungwas abundant, a turtle would move directly to a pile of dung and begin tearing it apart with the forelegs or burrowing into it. Turtles most often foraged in cow dung that had a superficial, dried crust. The invertebrate fauna of older dung was probably greater than that of fresh dung. Adult and larval insects were eaten, along with quantities of dung, as they were uncovered. Sometimes box turtles chased and caught larger insects that ran a foot or more away from the pile of dung; the turtles could cover the distance of one foot with three or four quick steps. Depressions made by box turtles in cow dung, as well as drier cow dung that had been more completely dissected, were regarded as characteristic "sign" ofT. ornataat the Damm Farm and in other areas studied (Pl. 26). Several persons have told me of box turtles "eating cow dung"; these reports, most of them made by competent observers, probably result from observations of box turtles ingesting cow dung incidentally, along with some unseen item of food.Contents of stomachs were analyzed. Scats and contents of lower digestive tracts, although obtained in large quantity, were unsuitable for analysis because of the fragmentary nature of the foods they contained. Relative amounts of various kinds of foods in stomachs were estimated; volume was determined by displacement of water or fine shot.Twenty-three stomachs of adults were selected at random (except for the fact that empty stomachs were discarded) from more than a hundred specimens collected in Douglas County, Kansas, in the period from June, 1954, to June, 1957; the sample included stomachs obtained in nearly all the months of the season of activity. Kinds of foods in stomachs did not differ significantly in regard to the sex of the turtles or to time of year. The stomach of each of two juveniles (included in Table 6) contained a greater variety of animal food than did the stomach of any adult, but no kind of animal was eaten by the juveniles exclusively.Each of the 23 stomachs contained animal matter and, in addition, all but two contained at least some plant material from dung, which constituted up to 20 per cent of total stomach contents.Insects were present in each of the 23 stomachs and constituted the bulk of the animal matter; beetles, caterpillars, and grasshoppers (ranked in descending order) were the kinds occurring most frequently and constituting the largest average percentages of total stomach-contents. Most of the beetles were scarabaeids and carabids; the bulk of the caterpillars were noctuids and arctiids. Grasshoppers, with one exception, were of a single species,Melanoplus differentialis. It is noteworthy that two of the kinds of insects frequently eaten (differential grasshoppers and noctuid caterpillars) are of economic importance in that they damage crops.Table 6.—Kinds of Animals Found in the Stomachs of 25Terrapene o. ornataof Both Sexes (23 adults, 2 juveniles) from Douglas County, Kansas. Frequency of Occurrence (number of stomachs in which found) is Given for Each Item Listed.Frequency of OccurrenceAdultsLarvaeTotalGastropodaHelisomasp1. . . .1Succiniasp1. . . .1Polygyrasp1. . . .1Retinellasp1. . . .1CrustaceaProcambaris gracilis1. . . .1Armadillidium vulgare4. . . .4Orthoptera (Locustidae)Locustinae (Melanoplus differentialis)13. . . .13Oedipodinae1. . . .1Lepidoptera (unspecified). . . .11Arctiidae. . . .99Noctuidae. . . .1010Pyralidae. . . .11Sphingidae. . . .11Diptera (Sarcophagidae). . . .11Coleoptera (unspecified)3. . . .3Cantharidae. . . .11Carabidae (unspecified)6. . . .6Carabidae(Eumolops colossus)1. . . .1Cerambycidae (Prionus fissicornis)1. . . .1Chrysomelidae (Diabotrica 12-punctata)1. . . .1Curculionidae (Calendra parvulus)3. . . .3Lampyridae (Photinus pyralis)2. . . .2Lampyridae(Photuris sp.)11Phengodidae. . . .11Scarabaeidae11. . . .11Hymenoptera (Formicidae)2. . . .2Phalangida1. . . .1Araneida (Epeira)1. . . .1Diplopoda1. . . .1Vertebrata (carrion). . . .. . . .4Table 7.—Occurrence of Insects, by Frequency and Volume, in Stomachs of 23Terrapene ornatafrom Douglas County, Kansas. Relative Volume is Based on Total Amount of Food Material Present, Excluding Stones and Vegetable Material Contained in Dung.Insects (all)OrthopteraLepidoptera (larvae)ColeopteraAveragevolumetric percentage88.628.726.932.5Range(volumetric percentage)trace to 1000 to 1000 to 1000 to 100Frequency of occurrence(percentage of totalstomachs in which found)100526574Snails, sowbugs, and the one individual of crayfish found in stomachs were kinds that could be expected to occur in moist grassland or in wooded stream courses. Mulberries were present in one stomach and fragments of bird's-nest fungi (Cyathus striatus) were present in another. Carrion consisted of remains of mammals and birds; the only identifiable items were bones of the eastern cottontail (Sylvilagus floridanus) and a chicken. Stones up to seven millimeters in diameter were found in many stomachs; stones constituted as much as half of total stomach-contents. Presumably the stones were accidentally swallowed when food was taken from the ground.The few adequate reports on dietary habits ofT. carolina(Allard, 1935:325-326; Carr, 1952:147, 150, 152, 153; Stickel, 1950:361; Surface, 1908:175-177) indicate that the species is omnivorous but that individuals tend to be herbivorous or carnivorous at certain times. Ornate box turtles resembleT. carolinain being opportunistic feeders but rely on insects as a staple part of the diet. In this respect the ornate box turtle seems to differ from all other kinds of box turtles in the United States and it is probably unique in its habitual utilization of dung communities as a source of food.POPULATIONSOrnate box turtles were probably more numerous on the Damm Farm than any other kinds of reptiles, excepting skinks (Eumeces fasciatusandE. obsoletus), and were by far the most conspicuous element of the reptilian fauna.The 194 box turtles that were marked at the Damm Farm were captured a total of 437 times. Seventy-nine (41 per cent) individuals were recaptured at least once, 49 (25 per cent) twice, 29 (15 per cent) three times, and 20 (10 per cent) were recaptured at least four times. Only three individuals were recaptured more than eight times. The greatest number of recaptures for a single individual, an old female, was 23.In all, 185 turtles (95 per cent of total recorded at Damm Farm) were captured on the pasture. Of these, 73 were in the northwest corner area, 44 in the house pond area, and 35 in the southern ravine area. The density of the population at the Damm Farm, considering the entire area, was .88 turtles per acre; for the woodland area alone, density was .41 turtles per acre and for the pasture alone, density was 1.49. Acreage and population density in the northwest corner, house pond, and southern ravine areas were respectively, 28 acres with 2.6 turtles per acre, 7 acres with 6.3 turtles per acre, and, 17 acres with 2.6 turtles per acre. The densities noted above for the wooded area and for the entire Damm Farm are low as a result of incomplete sampling in the wooded area. Estimates of population density for the subdivisions of the pasture seem more closely to approach the true population density in areas of favorable habitat.Fewer unmarked turtles were captured as the study progressed, but they were still being captured occasionally when field work was terminated. In order to estimate the number of turtles in the population at the Damm Farm the "Lincoln Index" (Lincoln, 1930) was used to compare the ratio of marked individuals to total number of individuals (17:56) in collections for June, 1956, to the ratio of marked individuals as of July 31, 1955 (87) to total individuals in the population; the result was 286.Fitch (1958:78) estimated the population ofT. ornatain one area of the Reservation (including woodland and ungrazed pasture) to be .076 turtles per acre. Stickel (1950:373) estimated the population of adultT. carolinato be four to five turtles per acre in favorable habitat at the Patuxent Research Refuge, Laurel, Maryland; juveniles comprised less than ten per cent of the population.Of the 194 turtles marked at the Damm Farm, 103 (53 per cent) were adult or subadult females, 61 (31 per cent) were maturemales, and 30 (16 per cent) were juveniles of undetermined sex. The ratio of males to females was then, 1.00 to 1.69, and the ratio of juveniles to adults was, 1.00 to 6.47. Eighteen of the 194 individuals were juveniles less than 90 millimeters in plastral length and only six had plastra less than 60 millimeters long (Fig. 25). The unbalanced ratio between males and females may result, in part, from sexual differences in habits. The studies of Carr (1952:9), Fitch (1954:140), Forbes (1940:132), Legler (1954:138), and Risley (1933:690), have shown, however, that unbalanced sex ratios, with females outnumbering males, are found in several species of reptiles, especially in turtles.Records for 540 adultT. ornatacollected at the Damm Farm, the Reservation, and on roads in eastern Kansas, show that females outnumber males just before and during the nesting season and again in late autumn (Fig. 26). The high incidence of females in May, June, and July, can be explained by their more extensive movements associated with nesting in these months. I have no explanation for the increased number of females captured in late autumn. In April and August, the only two months in which males were more abundant than females, the samples were small. The number of juveniles collected was too small to allow any trustworthy conclusions concerning their seasonal incidence; a few juveniles were taken in nearly all the periods in which adults were active.Risley (1933:690), studyingSternotherus odoratusin Michigan, found an over-all sex ratio of 1.0 male to 2.3 females; the percentage of females in collections ranged from 50 to 71 per cent in April and most of May and rose to 83 and 85 per cent in late May and mid-June, respectively.The infrequency with which hatchlings and small juveniles of ornate box turtles are observed is well known to naturalists. Several of my colleagues who are expert field observers and who have lived in areas where ornate box turtles are abundant, have never seen hatchlings; many other persons have seen only one or two. Rodeck (1949:33), noting the abundance of coleopterous insects in the scats of captives and the rarity of individuals of all age groups during dry periods in Colorado, commented, "It is possible that the young are even more subterranean than the adults. Perhaps they spend their early years in rodent or other burrows where there is a fairly abundant insect fauna. Increasing size might force them to the surface for feeding, with a daily return to a burrow for resting and protection."Fig. 25.Composition of the population ofT. o. ornataat the Damm Farm based on the 194 individuals marked there in the years 1954 to 1956. Individuals smaller than 100 mm. ordinarily could not be sexed accurately and are shown as open bars. Open bars in the groups larger than 100 mm. are for females, whereas solid bars are for males.My own experience in the field has shown that small examples ofT. ornataare not so rare as previous workers have believed. Small box turtles occupy the same microhabitat as do the adults and seem not to be more aquatic or subterranean in habits. Juveniles are found in burrows, in marshy areas, and in other sheltered places, but so are adults. Most of the juveniles that I found were in open situations where adults were abundant, sometimes within several inches of a place where an adult was feeding or basking. Nearly every one of the smaller turtles was discovered when I was closely scrutinizing some other object on the ground; sometimes juveniles were actually touched before being seen. Most juveniles were covered with cow dung or mud and blended so well with the substrate that they were detected only when they moved. It is likely that only a small number of the young box turtles present in an area is ever actually observed. Young are more vulnerable to predation and injury because of their small size, soft shells, and immovable plastra. They evidently rely, to a large extent, on inconspicuousness for protection.Fig. 26.The seasonal abundance of females ofT. o. ornatabased on 540 adults captured at the Damm Farm, the Reservation, and on roads in eastern Kansas, in the years 1954 to 1956. Records are grouped in periods of 30 days, the periods beginning with the dates shown at the bottoms of the bars. Juveniles are not considered. Numbers at the top of each bar indicate the size of the sample (both sexes) and give an approximate indication of relative seasonal abundance of adults, except for August, when little field work was done.MOVEMENTSThe only previous study of movements ofT. ornatais that of Fitch (1958:99-101). He recovered 14 markedT. ornataat the Reservation a total of 30 times, the period between recaptures varying from one to seven years. He reported that the average radius of home range was 274 feet (for an area of approximately 5.4 acres), excluding a single (presumably gravid) female that moved 1830 feet in 53 days.Although published information onT. ornatais scant, a considerable amount of information is available concerning its congener,T. carolina. The classic studies of Stickel (1950) on it constitute the most complete account of populations and movements for any reptile or amphibian, and probably, for any vertebrate. She found the average home range of adults to be 350 feet in diameter. Home ranges were not defended as territories and nearly all individuals were socially tolerant of one another. Movements (studied by means of a thread-trailing device) were characterized by frequenttravel over the same routes within the home range. Some turtles concentrated their activities in only one part of the home range, moving subsequently to another part, and some turtles had two ranges between which they traveled at varying intervals. Females ordinarily left their home ranges to nest.Other noteworthy, but less detailed, studies of populations ofT. Carolinaare those of Breder (1927) who found evidence of home range and homing behavior, and of Nichols (1939b) who, after observing a marked population on Long Island over a period of twenty years, found evidence of homing behavior and estimated normal home range to be approximately 250 yards in diameter. Numerous shorter papers such as those of Schneck (1886) and Medsger (1919) document the tendency ofT. carolinato remain in restricted areas over long periods.Important studies that indicate the presence of home range and homing behavior in other chelonians are those of Cagle (1944) onPseudemys scriptaandChrysemys picta, and of Woodbury and Hardy (1948) onGopherus agassizi. Grant (1936) and Bogert (1937) have also indicated that movements of individuals ofGopherus agassiziare restricted to limited areas.LocomotionOrnate box turtles moving forward over even terrain hold the plastron a quarter to a half inch above the ground and keep the head and neck lowered and extended. Each foreleg is brought forward and the humerus points nearly straight ahead when the foot touches the ground. Nearly all of the palmar surface is initially in contact with the ground but as the body is brought forward and the humerus swings outward, only the claws, and finally, only the two inner claws are in contact with the ground. Of the hind feet, the medial surfaces are the principal parts that touch the ground but some traction is derived from the hind claws at the beginning of each cycle of the hind leg. Under normal conditions, box turtles move slowly and pause to rest and examine their surroundings every few feet. When resting, the plastron is in contact with the ground, the legs relaxed, and the head and neck are extended upward. Some turtles seeking shelter from the heat of sunshine walk rapidly for a hundred feet or more without pausing.Turtles seen feeding under natural conditions displayed remarkable agility in making lunges, consisting of one or two short steps and a thrust of the head, at moving objects. Turtles kept in myhome were able, after being conditioned to hand-feeding, quickly to intercept a grape rolled slowly across a linoleum-covered floor.Frederick R. Gehlbach told me that, of several species of captive turtles observed by him,T. ornatacharacteristically walked with the plastron held well above the substrate, as didGopherus berlandieri, but thatT. carolina(specimens from the northeastern U. S.) dragged their shells as they walked. ApparentlyT. carolinain Kansas (currently referred to the subspeciestriunguis) differs somewhat in gait from populations in the eastern part of the range; several individuals ofT. carolinafrom Kansas that I observed in captivity, kept their plastra raised well above the smooth, hard substrate over which they walked.Box turtles at the Damm Farm were able easily to climb ravine banks that sloped at an angle of 45 degrees and, with some difficulty, could climb banks as steep as 65 degrees. Most individuals, however, were reluctant to walk directly downward on banks as steep as 45 degrees. Several individuals were seen to lose footing when climbing up or down a steep bank and to roll or slide to the bottom. Ordinarily,T. ornatais able to climb over a sheer surface as high as its shell is long, provided the surface is rough enough to give some traction to the foreclaws. The claws of first one, then the other forefoot are placed over the top of the barrier and then a hind foot, extended as far forward as possible, secures a hold as the turtle goes over the barrier.A number of observations on speed were made in the field where distance traveled and time elapsed were known approximately. Speeds ranged from 20 to 100 feet per hour in the course of foraging. Higher speeds (400 or more feet in one hour) were for turtles moving along pathways or seeking shelter. Gould (1957:346) observed somewhat faster speeds inT. carolina(192 feet per hour in cloudy weather and 348 feet per hour in sunny weather); he observed individuals that had been removed from their normal home ranges.Individuals ofT. ornatathat were placed in water swam moderately well but were clumsy in comparison to individuals of more aquatic emyids such asPseudemysandChrysemys. Box turtles were never observed to swim voluntarily, although they were frequently found in shallow water. On several occasions I confronted individuals at the edge of a pond so that the only unblocked route for their escape was through deeper water; nearly always these individuals attempted to crawl past me, to crawl away in shallowwater parallel to the shore, or to hide in soft mud at the edge of the water. Box turtles floated high in the water with the dorsal side upward and had little difficulty in righting themselves when turned over. The head and neck are extended and submerged when the turtle is swimming; forward progress is interrupted every few moments to elevate the head, presumably for purposes of breathing and orientation. The shell is never submerged. The swimming ofT. ornatais in general like that ofPseudemysorChrysemysthat have become dehydrated after long periods out of water and cannot submerge. These more aquatic turtles, however, quickly overcome their bouyancy, whereas examples ofT. ornata, even if left in water for several days, are unable to submerge. Clarke (1950) saw an ornate box turtle swim a 60-foot-wide stream in Osage County, Kansas; his description of swimming agrees with that given above.The meager swimming ability ofT. ornatais of apparent survival value under unusual conditions and enablesT. ornatato traverse bodies of water that would act as geographic barriers to completely terrestrial reptiles; however, swimming is a mode of locomotion seldom used under ordinary circumstances.Gehlbach (1956:366) and Norris and Zweifel (1950:2) observed individuals ofT. o. luteolaswimming in temporary rain pools and small ponds in New Mexico; the two authors last named saw an individual quickly enter a pond and dive beneath the water after being startled on the bank. Several of my colleagues, in conversation, have also reported seeingT. o. luteolain small bodies of water in the southwestern United States.Daily Cycle of ActivityThe daily cycle ofT. ornataconsists basically of periods of basking, foraging, and rest that vary in length depending upon environmental conditions. Turtles emerge from burrows, forms, and other places of concealment soon after dawn and ordinarily bask for at least a few minutes before beginning to forage; foraging is combined sometimes with basking, especially in open areas that are suitable for both kinds of activity. Foraging usually continues until shelter is sought sometime between mid-morning and noon. Turtles remain under cover (or continue to forage in shaded areas) until mid-afternoon or late afternoon when they again become active. They forage in both morning and afternoon. Study of travel records of a few of the turtles equipped with trailerssuggests that, under normal conditions, activity is slightly greater in forenoon than in afternoon, but that the converse is true of gravid females seeking nesting sites. Strecker (1908:79) reported that captiveT. ornata, after developing a feeding reflex, ate and retired until feeding time next day.As environmental temperatures rise in summer, the period of mid-day quiescence is lengthened. In the hottest part of the year, some turtles remain under cover for several days at a time. In periods of clear, cool weather at the beginning and end of the growing season, some turtles remain abroad and bask for most of the day.Examination of thread trails showed that activity of all individuals except nesting females was terminated at dusk. Breder (1927:236), Allard (1935:336), and Stickel (1950:358) reported a corresponding lack of nocturnal activity inT. carolina.Terrapene o. ornatain Kansas, andT. o. luteolain New Mexico (Norris and Zweifel, 1950:2)—unlike desert tortoises,Gopherus agassizi, which are active at night in hot weather (Woodbury and Hardy, 1948:186)—do not utilize the hours of darkness for foraging, even in the hottest part of the year.Seasonal Cycle of ActivityData obtained by mapping the movements of turtles that were equipped with trailing devices made it possible to compare distances traveled in the course of daily activities at different times of the year. Some of these data are expressed graphically inFigure 27. It should be noted that movement at all times in the season of activity was uneven; that is to say, an individual would move several hundred feet each day for a period of several days, and then, for an interval of one to several days, move only a few feet from one shelter to another, or not move at all. Such periods of rest could not be correlated definitely with environmental conditions; some individuals were inactive on days that were probably ideal (in terms of moderately warm temperatures and high humidity) for activity of box turtles. Analagous rest periods were noted inT. carolinaby Stickel (1950:358).Two males ofT. ornatathat had been removed by me from their normal home ranges traveled the longest average distance per day (429 feet). Gravid females in June traveled the next longest average distance per day (363 feet). The average distances traveled per day by non-gravid females in June (226 feet) and July(260 feet) and by males (within their known home ranges) in June (289 feet) were thought to approximate normal amount of movement under average environmental conditions. Average distance traveled per day by females in October (152 feet) was shortest because of frequent and extended rest periods. Nevertheless, in October actual distances traveled on days of activity tended to be longer than in any other month. A gravid female traveled farther in a single day than any other individual ofT. ornataobserved; she moved along a rock fence for approximately 700 feet, then left the study area and moved, in a nearly straight line, 1,200 feet across a cultivated field. Then the thread on her trailer was expended. The total distance moved, therefore, was at least 1,900 feet and probably more.Fig. 27.Average distances traveled per day by males and females at different times of the year, determined by mapping of thread trails at the Damm Farm. The diagram for "homing males" represents the distances traveled by two males removed from their normal home ranges to test homing ability. The data presented are for an aggregate of 136 days of trailing. Vertical and horizontal lines represent, respectively, the range and mean. Open and solid rectangles represent one standard deviation and two standard errors of the mean, respectively.An adult male at the Reservation traveled 2,240 feet in the 36-day period from October 16 to November 20, 1954, mostly on a woodedhillside. Eleven forms found along the route of the turtle's travels indicated that movement took place on roughly one out of three days in the elapsed period and demonstrated the sporadic nature of movements in autumn. The turtle remained active for an undetermined time after November 20.Home RangeData obtained from trailing and various methods of recapture at the Damm Farm indicated that each individual used only a small part of the total study area in the course of daily activities and tended to remain within a restricted area for a long time.The number of recaptures of no individual was great enough to permit application of refined calculations of size of home range as described by Odum and Kuenzler (1955). For individuals that were recaptured six or more times, or individuals for which adequate trailing records were available, the area enclosed by a line joining the peripheral points of capture was considered adequately representative of the home range of that individual, unless recaptures were all within a few feet of each other or lay in an approximately straight line. If less than six records of recapture were available, home range was estimated, in the manner described by Fitch (1958:73), by averaging the distance between successive points of recapture and letting this average represent the radius of home range; the actual area of home range was determined by the formula, π(R)2, for the area of a circle.Size of home ranges of males and females did not differ significantly and data for the two sexes were combined in the final analysis. The average radius of the home ranges of 44 adults (captured a total of 146 times) was 278 feet (extremes, 71 to 913) when computed by measuring the distance between successive captures; the average area of these home ranges was 5.6 acres. Data from 10 turtles that had been recaptured only once were combined with data from 34 turtles that had been recaptured more than once when it was found that the average size of home range in these two groups did not differ significantly. Data concerning the home ranges of eight of the 44 individuals were sufficient to permit actual measurement of home ranges with a planimeter; home ranges of these eight individuals had an average area of five acres (extremes, 1.2 to 10.2).A minimum home range could theoretically consist of the smallest area in which adequate food and shelter were available. Under favorable conditions a turtle could stay in an area ten to twenty feet in diameter. Although several such favorable small areasexisted on the Damm Farm, box turtles seldom stayed in one for more than a day or two. Seemingly, therefore, factors additional to food and shelter influence size of home range. At the Damm Farm these additional factors seemed to be: rock fences that acted as physical barriers; areas that were cultivated, barren, or otherwise unfavorable, acting as ecological barriers; and, cowpaths and ravines that offered relatively unobstructed routes along which box turtles tended to move.One subdivision of the main pasture, the northwest corner area, is an example of a relatively small natural area in which many individual box turtles had home ranges. This tract of 28 acres was roughly triangular and was bordered on two sides by rock fences that contained no gates or other passageways. On its third (southeastern) side the area sloped into a deep ravine. Habitat in this subdivision of the pasture (as well as in the other two subdivisions) was especially favorable for box turtles because of permanent water, rocky slopes, ravines, and several fruit trees. Box turtles usually foraged near the rock fences and the ravine (where dung was more abundant than in other parts of the area), and tended, as they foraged, to move parallel to these barriers. Turtles crossing the area eventually came either to one of the fences or the ravine. Therefore, most of the turtles in the northwest corner area eventually completed a circuit of the area. Turtles that came to the ravine tended to move along its bottom or sides. Several turtles were known to cross the ravine and to forage in the grassy area on its southeastern side. These turtles usually re-entered the ravine by way of smaller side-ravines. Of 22 box turtles known to have home ranges in the northwest corner area, only two individuals (both gravid females) were known to leave the area in the period in which observations were made.Two other subdivisions of the main pasture—the house pond area and the southern ravine area—although not so distinct as the northwest corner area in terms of limiting barriers, nevertheless constituted separate areas of favorable habitat, each of which contained a number of individual home ranges. Although the two areas were not far apart, but little movement was observed of turtles from one area to the other. The home range of only one turtle, an adult female, was known to include parts of both areas.Unbroken expanses of tall grass seem not to be optimum habitat. The crest of the hill at the Damm Farm (Pl. 17, Fig. 1) was an area of more or less homogeneous grassy habitat. Turtles were seldomfound on the crest of the hill although this area was as thoroughly searched for turtles as any other area. Known home ranges of nearly every individual observed were on either one of the sides of the hill but not on both sides.At several places on the border of the pasture, turtles were able to move freely into cultivated areas but seldom did so except for nesting. Trailing records show that most of the turtles that entered one of the cultivated areas returned again to the pasture.Ornate box turtles seem to find places of shelter by trial and error along regularly used routes of travel in their home ranges. The individuals that I studied never returned to the same forms, and seldom returned to the same natural burrows and dens. Probably foraging, basking, and watering sites are found also by trial and error.Stickel (1950:375) placed considerable importance on the occurrence of transient turtles in populations ofT. carolina; in estimating population density, she added to her study area a peripheral strip, half as wide as the average, estimated home range, to account for turtles that had home ranges only partly within the study area. The study area used by Stickel had no natural boundaries, as habitat conditions on all sides were essentially the same as those of the study area itself. The pasture at the Damm Farm, on the contrary, is a relatively isolated area of natural grassland, bordered by rock fences and cultivated fields. I believe that most of the box turtles found on the pasture were permanent residents there. Individual box turtles at the Damm Farm seemingly occupied but one home range and it did not change from year to year. Populations ofT. ornatain areas less isolated than the Damm Farm, like the populations ofT. carolinastudied by Stickel (loc. cit.), could be expected to have a higher percentage of transient individuals and individuals with multiple or changing home ranges. Henry S. Fitch told me that he considered most of the individuals ofT. ornatathat were captured only once at the Reservation were transients.Several females at the Damm Farm traveled long distances from their home ranges to nest but other females nested within their known or estimated home ranges. Seemingly a complex of environmental factors, including soil texture, weather, availability of water, and possibly the urge for random wandering in the breeding season, governs the distances traveled by gravid females and the ultimate selection of a satisfactory nesting site. Females, because of their more extensive travels in the nesting season, seem more likely thanmales to have multiple or changing home ranges. Males ofT. ornatadid not noticeably alter the extent or pattern of their movements in the breeding season. Hibernacula, unlike nesting sites, were within the known or estimated home ranges of all individuals studied.
Fig. 24.The approximate length of stay of each known occupant of a den that was examined six times in the winter of 1955-1956 at the Damm Farm. Most of the occupants used the den as a temporary shelter and sought permanent hibernacula elsewhere. One turtle left the den for approximately two weeks and then returned to it for the rest of the winter. The temperature of the air outside the den (A) and the average body temperature of turtles in the den (B) are given at the bottom of the diagram for each date the den was examined. The symbol "J" represents a juvenal turtle.
Fig. 24.The approximate length of stay of each known occupant of a den that was examined six times in the winter of 1955-1956 at the Damm Farm. Most of the occupants used the den as a temporary shelter and sought permanent hibernacula elsewhere. One turtle left the den for approximately two weeks and then returned to it for the rest of the winter. The temperature of the air outside the den (A) and the average body temperature of turtles in the den (B) are given at the bottom of the diagram for each date the den was examined. The symbol "J" represents a juvenal turtle.
In general, body temperatures approximated the temperature of the soil around the turtle. Body temperatures tended to be slightly higher than soil temperatures in November and December but were slightly lower than soil temperatures in the months of February and March. The lowest body temperature recorded for any turtle thatsurvived a winter was 2.7 degrees, taken from an adult female on December 26, 1955. Body temperatures one to three degrees higher were common in the coldest part of the winter. Turtles in shallow hibernacula, like those observed in wooded areas at the Reservation, are probably subjected to freezing temperatures at least for short periods but I have no records of body temperatures this low, except where they were induced experimentally. Turtles exposed to temperatures of zero degrees or slightly lower would retain enough heat to survive without freezing for a period of several hours or even a day if well insulated. A temperature gradient exists within the body; cloacal temperatures, for example, differ from temperatures deep in the colon and temperatures in the dorsal and ventral parts of the body cavity (taken by manipulating the bulb of the thermometer while it was in the colon) differ from one another. Probably, therefore, some parts of some turtles—probably the top of the shell or the extremities—freeze in winter without causing the death of the turtle. Ewing (1939:91) found a female ofT. carolina, just emerging from hibernation, that had lost some scutes from its carapace; he found the missing scutes in the hibernaculum and attributed their loss to severe temperatures in the winter of 1933-34.
The incidence of mortality due to freezing is unknown for most species of reptiles. The observations of Bailey (1948) on DeKay snakes (Storeria dekayi) and Legler and Fitch (1957) on collared lizards suggest that rates of mortality are high in dormant reptiles. Bailey (op. cit.) suggested that winter mortality might act as a natural check on snake populations. Neill (1948a:114) thought more box turtles (T. carolina) were killed in Georgia by cold weather in late autumn than "… by all other factors together," and that this winter mortality acted as an effective check on population levels. Neill reported that many turtles left their burrows in late autumn and began to forage; if the temperature dropped suddenly, the turtles became "… too torpid to dig" and froze.
If ornate box turtles are occasionally caught in the open by a sudden cooling of air temperature, it would occur at a time of year when temperatures would approximate freezing but would drop not far below this level; laboratory and field records show that adults could probably survive these low temperatures overnight and warm up sufficiently on the following day to seek adequate shelter. Box turtles deepening their burrows in winter do so at body temperatures somewhat lower than 10 degrees (near the minimum temperature at which co-ordinated activity was observedin the laboratory); turtles found in the open in late October were known to burrow into the ground at body temperatures of approximately 15 degrees.
Emergence from hibernation usually occurs in April but in some years a few turtles may emerge as early as the first week of March. Emergence is stimulated by temperature and humidity. Fitch (1956b:438) stated that emergence was delayed until "…the ground has been sufficiently moistened and until air temperatures have reached at least 26°." Box turtles at the Reservation emerged on April 21 in 1954 and from April 16 to 17 in 1955. William R. Brecheisen found recently emerged box turtles in Anderson County on April 2, 1955, and March 6, 1956.
Turtles were found facing upward in their hibernacula in early March. As the temperature of the soil rises, they move slowly upward, usually following the route by which they entered. They remain just below the surface of the soil for a week or two before actually emerging; this final phase of emergence is probably hastened by spring rains that soften the soil. Activity may be sporadic after emergence if the weather is cold.
A number of box turtles at the Reservation emerged in a cold rain in 1954 when the temperatures of the air and ground were 16 and 13 degrees, respectively, but remained inactive for several days afterward. In 1955 the air and ground temperatures were higher (28° and 17°, respectively) on the day of emergence and box turtles became active almost immediately.
DIET
Published information on the food ofT. ornataconsists of a few miscellaneous observations. Cahn (1937:103) opened five stomachs that contained partly digested vegetable matter but no insects or other animal food: Ortenburger and Freeman (1930:187) noted that grasshoppers were a main part of the diet ofT. ornatain Oklahoma and that turtles displayed unsuspected agility in catching them. Those authors also saw turtles eating caterpillars and robber flies. Strecker (1908:79) stated that "The natural diet of this species consists of vegetable matter and earthworms." Norris and Zweifel (1950:3) observed the feeding habits of captiveT. o. luteola. Coyote melon (Cucurbita foetidissima) was eaten with reluctance but a collared lizard (Crotaphytus collaris) was quickly devoured. Tadpoles ofScaphiopus hammondiwere caught in a small pool and eaten. Adults of the same species were rejected after being caught; box turtles were seen wiping their mouthsafter rejecting adult toads. The authors suggested thatT. o. luteolais an important predator ofScaphiopus hammondi, since the two species occur together in many areas and the emergence of both is controlled to a large extent by rainfall. One individual ofluteolawas seen eating a dead box turtle on a road.
Captive individuals ofT. ornata, observed in the present study, ate nearly every kind of animal and vegetable food given to them. Table scraps, consisting chiefly of greens, various fruits and vegetables, meat, and cooked potatoes, formed the main diet of turtles kept in outdoor cages.
A number of persons have told me of ornate box turtles eating the succulent stems and leaves, and the fruits of various garden plants; similar incidents probably occur in areas of native vegetation. J. Knox Jones told me he saw an individual ofT. ornataeating a spiderwort (Tradescantiasp.) in Cherry County, Nebraska.
Sight-records of foods eaten by box turtles at the Damm Farm (excluding the many records of individuals foraging in dung or eating mulberries) were for grasshoppers, caterpillars, and various kinds of carrion. Box turtles were often seen eating grasshoppers on roads in early morning; Sophia Damm told me of frequently seeing individuals catching grasshoppers in her garden. Ralph J. Donahue told me that on his farm in Bates County, Missouri, an individual ofT. ornatamade a circuit of the lawn each morning in summer and ate all the cicadas (Magicicada septendecim) found.
Vertebrate remains found in the stomachs of box turtles seem to result chiefly from the ingestion of carrion. One box turtle ate a white egg (unidentified) that had fallen from a nest and another was seen with a blue down feather clinging to its mouth. Several colleagues have told me of box turtles eating small mammals caught in snap-traps and Marr (1944:489) reported a similar incident. J. Knox Jones told me he once found an ornate box turtle in the nest of a blue-winged teal in Cherry County, Nebraska; the three eggs in the nest had been broken. The only authentic record of an ornate box turtle preying on a vertebrate under natural conditions was one supplied by Ralph J. Donahue who saw an adult catch and eat one of a brood of bobwhite quail. In many areas where box turtles are abundant, it is the opinion of local residents that the turtles decimate populations of upland game birds by eating the eggs and young of these birds; these opinions result probably from rare encounters such as the one described by Donahue. I believe that box turtles at the Damm Farm were sometimes able to catch young frogs and tadpoles (chieflyRana catesbeianaandR. pipiens)at the margins of ponds. In autumn literally thousands of youngRanawere present in these places.
Ornate box turtles ordinarily attempt to catch and, without further examination, to eat, small objects moving on the ground, but are more critical of stationary objects. Captive turtles, for example, would immediately chase and seize a grape that was pulled or rolled slowly across a floor but a stationary grape was examined and then smelled before it was eaten. Similar observations were made a number of times with living and dead insects in the field and in the laboratory. A turtle discovering an object that is of possible value as food, approaches it closely, turns the head from side to side (presumably using the eyes alternately to examine the object), and then, with head cocked at a slight angle, momentarily presses the nostrils against the object (Pl. 28, Fig. 4). If acceptable as food, the object is then swallowed whole or taken into the mouth with a series of bites; large insects are usually broken into several pieces in the process of being bitten and swallowed. Larger objects, such as dead vertebrates, are torn to pieces with the beak and forefeet before they are swallowed. Hatchlings, when fed for the first time, ignored inanimate foods but eagerly chased mealworms, catching them usually by the anterior end. The tendency of the young of certain species of turtles (especially captives) to be more carnivorous than adults is probably due to the association of movement with food; recognition of inanimate objects as food is presumably learned by older individuals.
Mulberries (Morus rubra), when they are abundant, constitute all or an important part of the diet of ornate box turtles. On June 4, 1955, William R. Brecheisen and I drove along a road in Anderson County, Kansas, and stopped at each mulberry tree that we saw beside the road; we found at least one specimen ofT. ornataunder nearly every tree. Approximately twenty box turtles were collected in this manner in a little more than one hour. The heads and necks of most were stained dark-red from the fruit and, in some, nearly the entire shell was stained. Dissection of these turtles revealed that their stomachs were distended to two or three times normal size with mulberries; no other kinds of food were found in the stomachs. Some of the turtles voided purplish-black fluid from the cloaca when we handled them; the color of the fluid presumably resulted from mulberries.
Several turtles were observed through binoculars as they foraged. Individuals snapped or lunged periodically at objects on the ground along the route of travel. Upon reaching an area where cow dungwas abundant, a turtle would move directly to a pile of dung and begin tearing it apart with the forelegs or burrowing into it. Turtles most often foraged in cow dung that had a superficial, dried crust. The invertebrate fauna of older dung was probably greater than that of fresh dung. Adult and larval insects were eaten, along with quantities of dung, as they were uncovered. Sometimes box turtles chased and caught larger insects that ran a foot or more away from the pile of dung; the turtles could cover the distance of one foot with three or four quick steps. Depressions made by box turtles in cow dung, as well as drier cow dung that had been more completely dissected, were regarded as characteristic "sign" ofT. ornataat the Damm Farm and in other areas studied (Pl. 26). Several persons have told me of box turtles "eating cow dung"; these reports, most of them made by competent observers, probably result from observations of box turtles ingesting cow dung incidentally, along with some unseen item of food.
Contents of stomachs were analyzed. Scats and contents of lower digestive tracts, although obtained in large quantity, were unsuitable for analysis because of the fragmentary nature of the foods they contained. Relative amounts of various kinds of foods in stomachs were estimated; volume was determined by displacement of water or fine shot.
Twenty-three stomachs of adults were selected at random (except for the fact that empty stomachs were discarded) from more than a hundred specimens collected in Douglas County, Kansas, in the period from June, 1954, to June, 1957; the sample included stomachs obtained in nearly all the months of the season of activity. Kinds of foods in stomachs did not differ significantly in regard to the sex of the turtles or to time of year. The stomach of each of two juveniles (included in Table 6) contained a greater variety of animal food than did the stomach of any adult, but no kind of animal was eaten by the juveniles exclusively.
Each of the 23 stomachs contained animal matter and, in addition, all but two contained at least some plant material from dung, which constituted up to 20 per cent of total stomach contents.
Insects were present in each of the 23 stomachs and constituted the bulk of the animal matter; beetles, caterpillars, and grasshoppers (ranked in descending order) were the kinds occurring most frequently and constituting the largest average percentages of total stomach-contents. Most of the beetles were scarabaeids and carabids; the bulk of the caterpillars were noctuids and arctiids. Grasshoppers, with one exception, were of a single species,Melanoplus differentialis. It is noteworthy that two of the kinds of insects frequently eaten (differential grasshoppers and noctuid caterpillars) are of economic importance in that they damage crops.
Table 6.—Kinds of Animals Found in the Stomachs of 25Terrapene o. ornataof Both Sexes (23 adults, 2 juveniles) from Douglas County, Kansas. Frequency of Occurrence (number of stomachs in which found) is Given for Each Item Listed.
Frequency of OccurrenceAdultsLarvaeTotalGastropodaHelisomasp1. . . .1Succiniasp1. . . .1Polygyrasp1. . . .1Retinellasp1. . . .1CrustaceaProcambaris gracilis1. . . .1Armadillidium vulgare4. . . .4Orthoptera (Locustidae)Locustinae (Melanoplus differentialis)13. . . .13Oedipodinae1. . . .1Lepidoptera (unspecified). . . .11Arctiidae. . . .99Noctuidae. . . .1010Pyralidae. . . .11Sphingidae. . . .11Diptera (Sarcophagidae). . . .11Coleoptera (unspecified)3. . . .3Cantharidae. . . .11Carabidae (unspecified)6. . . .6Carabidae(Eumolops colossus)1. . . .1Cerambycidae (Prionus fissicornis)1. . . .1Chrysomelidae (Diabotrica 12-punctata)1. . . .1Curculionidae (Calendra parvulus)3. . . .3Lampyridae (Photinus pyralis)2. . . .2Lampyridae(Photuris sp.)11Phengodidae. . . .11Scarabaeidae11. . . .11Hymenoptera (Formicidae)2. . . .2Phalangida1. . . .1Araneida (Epeira)1. . . .1Diplopoda1. . . .1Vertebrata (carrion). . . .. . . .4
Table 7.—Occurrence of Insects, by Frequency and Volume, in Stomachs of 23Terrapene ornatafrom Douglas County, Kansas. Relative Volume is Based on Total Amount of Food Material Present, Excluding Stones and Vegetable Material Contained in Dung.
Insects (all)OrthopteraLepidoptera (larvae)ColeopteraAveragevolumetric percentage88.628.726.932.5Range(volumetric percentage)trace to 1000 to 1000 to 1000 to 100Frequency of occurrence(percentage of totalstomachs in which found)100526574
Snails, sowbugs, and the one individual of crayfish found in stomachs were kinds that could be expected to occur in moist grassland or in wooded stream courses. Mulberries were present in one stomach and fragments of bird's-nest fungi (Cyathus striatus) were present in another. Carrion consisted of remains of mammals and birds; the only identifiable items were bones of the eastern cottontail (Sylvilagus floridanus) and a chicken. Stones up to seven millimeters in diameter were found in many stomachs; stones constituted as much as half of total stomach-contents. Presumably the stones were accidentally swallowed when food was taken from the ground.
The few adequate reports on dietary habits ofT. carolina(Allard, 1935:325-326; Carr, 1952:147, 150, 152, 153; Stickel, 1950:361; Surface, 1908:175-177) indicate that the species is omnivorous but that individuals tend to be herbivorous or carnivorous at certain times. Ornate box turtles resembleT. carolinain being opportunistic feeders but rely on insects as a staple part of the diet. In this respect the ornate box turtle seems to differ from all other kinds of box turtles in the United States and it is probably unique in its habitual utilization of dung communities as a source of food.
POPULATIONS
Ornate box turtles were probably more numerous on the Damm Farm than any other kinds of reptiles, excepting skinks (Eumeces fasciatusandE. obsoletus), and were by far the most conspicuous element of the reptilian fauna.
The 194 box turtles that were marked at the Damm Farm were captured a total of 437 times. Seventy-nine (41 per cent) individuals were recaptured at least once, 49 (25 per cent) twice, 29 (15 per cent) three times, and 20 (10 per cent) were recaptured at least four times. Only three individuals were recaptured more than eight times. The greatest number of recaptures for a single individual, an old female, was 23.
In all, 185 turtles (95 per cent of total recorded at Damm Farm) were captured on the pasture. Of these, 73 were in the northwest corner area, 44 in the house pond area, and 35 in the southern ravine area. The density of the population at the Damm Farm, considering the entire area, was .88 turtles per acre; for the woodland area alone, density was .41 turtles per acre and for the pasture alone, density was 1.49. Acreage and population density in the northwest corner, house pond, and southern ravine areas were respectively, 28 acres with 2.6 turtles per acre, 7 acres with 6.3 turtles per acre, and, 17 acres with 2.6 turtles per acre. The densities noted above for the wooded area and for the entire Damm Farm are low as a result of incomplete sampling in the wooded area. Estimates of population density for the subdivisions of the pasture seem more closely to approach the true population density in areas of favorable habitat.
Fewer unmarked turtles were captured as the study progressed, but they were still being captured occasionally when field work was terminated. In order to estimate the number of turtles in the population at the Damm Farm the "Lincoln Index" (Lincoln, 1930) was used to compare the ratio of marked individuals to total number of individuals (17:56) in collections for June, 1956, to the ratio of marked individuals as of July 31, 1955 (87) to total individuals in the population; the result was 286.
Fitch (1958:78) estimated the population ofT. ornatain one area of the Reservation (including woodland and ungrazed pasture) to be .076 turtles per acre. Stickel (1950:373) estimated the population of adultT. carolinato be four to five turtles per acre in favorable habitat at the Patuxent Research Refuge, Laurel, Maryland; juveniles comprised less than ten per cent of the population.
Of the 194 turtles marked at the Damm Farm, 103 (53 per cent) were adult or subadult females, 61 (31 per cent) were maturemales, and 30 (16 per cent) were juveniles of undetermined sex. The ratio of males to females was then, 1.00 to 1.69, and the ratio of juveniles to adults was, 1.00 to 6.47. Eighteen of the 194 individuals were juveniles less than 90 millimeters in plastral length and only six had plastra less than 60 millimeters long (Fig. 25). The unbalanced ratio between males and females may result, in part, from sexual differences in habits. The studies of Carr (1952:9), Fitch (1954:140), Forbes (1940:132), Legler (1954:138), and Risley (1933:690), have shown, however, that unbalanced sex ratios, with females outnumbering males, are found in several species of reptiles, especially in turtles.
Records for 540 adultT. ornatacollected at the Damm Farm, the Reservation, and on roads in eastern Kansas, show that females outnumber males just before and during the nesting season and again in late autumn (Fig. 26). The high incidence of females in May, June, and July, can be explained by their more extensive movements associated with nesting in these months. I have no explanation for the increased number of females captured in late autumn. In April and August, the only two months in which males were more abundant than females, the samples were small. The number of juveniles collected was too small to allow any trustworthy conclusions concerning their seasonal incidence; a few juveniles were taken in nearly all the periods in which adults were active.
Risley (1933:690), studyingSternotherus odoratusin Michigan, found an over-all sex ratio of 1.0 male to 2.3 females; the percentage of females in collections ranged from 50 to 71 per cent in April and most of May and rose to 83 and 85 per cent in late May and mid-June, respectively.
The infrequency with which hatchlings and small juveniles of ornate box turtles are observed is well known to naturalists. Several of my colleagues who are expert field observers and who have lived in areas where ornate box turtles are abundant, have never seen hatchlings; many other persons have seen only one or two. Rodeck (1949:33), noting the abundance of coleopterous insects in the scats of captives and the rarity of individuals of all age groups during dry periods in Colorado, commented, "It is possible that the young are even more subterranean than the adults. Perhaps they spend their early years in rodent or other burrows where there is a fairly abundant insect fauna. Increasing size might force them to the surface for feeding, with a daily return to a burrow for resting and protection."
Fig. 25.Composition of the population ofT. o. ornataat the Damm Farm based on the 194 individuals marked there in the years 1954 to 1956. Individuals smaller than 100 mm. ordinarily could not be sexed accurately and are shown as open bars. Open bars in the groups larger than 100 mm. are for females, whereas solid bars are for males.
Fig. 25.Composition of the population ofT. o. ornataat the Damm Farm based on the 194 individuals marked there in the years 1954 to 1956. Individuals smaller than 100 mm. ordinarily could not be sexed accurately and are shown as open bars. Open bars in the groups larger than 100 mm. are for females, whereas solid bars are for males.
My own experience in the field has shown that small examples ofT. ornataare not so rare as previous workers have believed. Small box turtles occupy the same microhabitat as do the adults and seem not to be more aquatic or subterranean in habits. Juveniles are found in burrows, in marshy areas, and in other sheltered places, but so are adults. Most of the juveniles that I found were in open situations where adults were abundant, sometimes within several inches of a place where an adult was feeding or basking. Nearly every one of the smaller turtles was discovered when I was closely scrutinizing some other object on the ground; sometimes juveniles were actually touched before being seen. Most juveniles were covered with cow dung or mud and blended so well with the substrate that they were detected only when they moved. It is likely that only a small number of the young box turtles present in an area is ever actually observed. Young are more vulnerable to predation and injury because of their small size, soft shells, and immovable plastra. They evidently rely, to a large extent, on inconspicuousness for protection.
Fig. 26.The seasonal abundance of females ofT. o. ornatabased on 540 adults captured at the Damm Farm, the Reservation, and on roads in eastern Kansas, in the years 1954 to 1956. Records are grouped in periods of 30 days, the periods beginning with the dates shown at the bottoms of the bars. Juveniles are not considered. Numbers at the top of each bar indicate the size of the sample (both sexes) and give an approximate indication of relative seasonal abundance of adults, except for August, when little field work was done.
Fig. 26.The seasonal abundance of females ofT. o. ornatabased on 540 adults captured at the Damm Farm, the Reservation, and on roads in eastern Kansas, in the years 1954 to 1956. Records are grouped in periods of 30 days, the periods beginning with the dates shown at the bottoms of the bars. Juveniles are not considered. Numbers at the top of each bar indicate the size of the sample (both sexes) and give an approximate indication of relative seasonal abundance of adults, except for August, when little field work was done.
MOVEMENTS
The only previous study of movements ofT. ornatais that of Fitch (1958:99-101). He recovered 14 markedT. ornataat the Reservation a total of 30 times, the period between recaptures varying from one to seven years. He reported that the average radius of home range was 274 feet (for an area of approximately 5.4 acres), excluding a single (presumably gravid) female that moved 1830 feet in 53 days.
Although published information onT. ornatais scant, a considerable amount of information is available concerning its congener,T. carolina. The classic studies of Stickel (1950) on it constitute the most complete account of populations and movements for any reptile or amphibian, and probably, for any vertebrate. She found the average home range of adults to be 350 feet in diameter. Home ranges were not defended as territories and nearly all individuals were socially tolerant of one another. Movements (studied by means of a thread-trailing device) were characterized by frequenttravel over the same routes within the home range. Some turtles concentrated their activities in only one part of the home range, moving subsequently to another part, and some turtles had two ranges between which they traveled at varying intervals. Females ordinarily left their home ranges to nest.
Other noteworthy, but less detailed, studies of populations ofT. Carolinaare those of Breder (1927) who found evidence of home range and homing behavior, and of Nichols (1939b) who, after observing a marked population on Long Island over a period of twenty years, found evidence of homing behavior and estimated normal home range to be approximately 250 yards in diameter. Numerous shorter papers such as those of Schneck (1886) and Medsger (1919) document the tendency ofT. carolinato remain in restricted areas over long periods.
Important studies that indicate the presence of home range and homing behavior in other chelonians are those of Cagle (1944) onPseudemys scriptaandChrysemys picta, and of Woodbury and Hardy (1948) onGopherus agassizi. Grant (1936) and Bogert (1937) have also indicated that movements of individuals ofGopherus agassiziare restricted to limited areas.
Locomotion
Ornate box turtles moving forward over even terrain hold the plastron a quarter to a half inch above the ground and keep the head and neck lowered and extended. Each foreleg is brought forward and the humerus points nearly straight ahead when the foot touches the ground. Nearly all of the palmar surface is initially in contact with the ground but as the body is brought forward and the humerus swings outward, only the claws, and finally, only the two inner claws are in contact with the ground. Of the hind feet, the medial surfaces are the principal parts that touch the ground but some traction is derived from the hind claws at the beginning of each cycle of the hind leg. Under normal conditions, box turtles move slowly and pause to rest and examine their surroundings every few feet. When resting, the plastron is in contact with the ground, the legs relaxed, and the head and neck are extended upward. Some turtles seeking shelter from the heat of sunshine walk rapidly for a hundred feet or more without pausing.
Turtles seen feeding under natural conditions displayed remarkable agility in making lunges, consisting of one or two short steps and a thrust of the head, at moving objects. Turtles kept in myhome were able, after being conditioned to hand-feeding, quickly to intercept a grape rolled slowly across a linoleum-covered floor.
Frederick R. Gehlbach told me that, of several species of captive turtles observed by him,T. ornatacharacteristically walked with the plastron held well above the substrate, as didGopherus berlandieri, but thatT. carolina(specimens from the northeastern U. S.) dragged their shells as they walked. ApparentlyT. carolinain Kansas (currently referred to the subspeciestriunguis) differs somewhat in gait from populations in the eastern part of the range; several individuals ofT. carolinafrom Kansas that I observed in captivity, kept their plastra raised well above the smooth, hard substrate over which they walked.
Box turtles at the Damm Farm were able easily to climb ravine banks that sloped at an angle of 45 degrees and, with some difficulty, could climb banks as steep as 65 degrees. Most individuals, however, were reluctant to walk directly downward on banks as steep as 45 degrees. Several individuals were seen to lose footing when climbing up or down a steep bank and to roll or slide to the bottom. Ordinarily,T. ornatais able to climb over a sheer surface as high as its shell is long, provided the surface is rough enough to give some traction to the foreclaws. The claws of first one, then the other forefoot are placed over the top of the barrier and then a hind foot, extended as far forward as possible, secures a hold as the turtle goes over the barrier.
A number of observations on speed were made in the field where distance traveled and time elapsed were known approximately. Speeds ranged from 20 to 100 feet per hour in the course of foraging. Higher speeds (400 or more feet in one hour) were for turtles moving along pathways or seeking shelter. Gould (1957:346) observed somewhat faster speeds inT. carolina(192 feet per hour in cloudy weather and 348 feet per hour in sunny weather); he observed individuals that had been removed from their normal home ranges.
Individuals ofT. ornatathat were placed in water swam moderately well but were clumsy in comparison to individuals of more aquatic emyids such asPseudemysandChrysemys. Box turtles were never observed to swim voluntarily, although they were frequently found in shallow water. On several occasions I confronted individuals at the edge of a pond so that the only unblocked route for their escape was through deeper water; nearly always these individuals attempted to crawl past me, to crawl away in shallowwater parallel to the shore, or to hide in soft mud at the edge of the water. Box turtles floated high in the water with the dorsal side upward and had little difficulty in righting themselves when turned over. The head and neck are extended and submerged when the turtle is swimming; forward progress is interrupted every few moments to elevate the head, presumably for purposes of breathing and orientation. The shell is never submerged. The swimming ofT. ornatais in general like that ofPseudemysorChrysemysthat have become dehydrated after long periods out of water and cannot submerge. These more aquatic turtles, however, quickly overcome their bouyancy, whereas examples ofT. ornata, even if left in water for several days, are unable to submerge. Clarke (1950) saw an ornate box turtle swim a 60-foot-wide stream in Osage County, Kansas; his description of swimming agrees with that given above.
The meager swimming ability ofT. ornatais of apparent survival value under unusual conditions and enablesT. ornatato traverse bodies of water that would act as geographic barriers to completely terrestrial reptiles; however, swimming is a mode of locomotion seldom used under ordinary circumstances.
Gehlbach (1956:366) and Norris and Zweifel (1950:2) observed individuals ofT. o. luteolaswimming in temporary rain pools and small ponds in New Mexico; the two authors last named saw an individual quickly enter a pond and dive beneath the water after being startled on the bank. Several of my colleagues, in conversation, have also reported seeingT. o. luteolain small bodies of water in the southwestern United States.
Daily Cycle of Activity
The daily cycle ofT. ornataconsists basically of periods of basking, foraging, and rest that vary in length depending upon environmental conditions. Turtles emerge from burrows, forms, and other places of concealment soon after dawn and ordinarily bask for at least a few minutes before beginning to forage; foraging is combined sometimes with basking, especially in open areas that are suitable for both kinds of activity. Foraging usually continues until shelter is sought sometime between mid-morning and noon. Turtles remain under cover (or continue to forage in shaded areas) until mid-afternoon or late afternoon when they again become active. They forage in both morning and afternoon. Study of travel records of a few of the turtles equipped with trailerssuggests that, under normal conditions, activity is slightly greater in forenoon than in afternoon, but that the converse is true of gravid females seeking nesting sites. Strecker (1908:79) reported that captiveT. ornata, after developing a feeding reflex, ate and retired until feeding time next day.
As environmental temperatures rise in summer, the period of mid-day quiescence is lengthened. In the hottest part of the year, some turtles remain under cover for several days at a time. In periods of clear, cool weather at the beginning and end of the growing season, some turtles remain abroad and bask for most of the day.
Examination of thread trails showed that activity of all individuals except nesting females was terminated at dusk. Breder (1927:236), Allard (1935:336), and Stickel (1950:358) reported a corresponding lack of nocturnal activity inT. carolina.Terrapene o. ornatain Kansas, andT. o. luteolain New Mexico (Norris and Zweifel, 1950:2)—unlike desert tortoises,Gopherus agassizi, which are active at night in hot weather (Woodbury and Hardy, 1948:186)—do not utilize the hours of darkness for foraging, even in the hottest part of the year.
Seasonal Cycle of Activity
Data obtained by mapping the movements of turtles that were equipped with trailing devices made it possible to compare distances traveled in the course of daily activities at different times of the year. Some of these data are expressed graphically inFigure 27. It should be noted that movement at all times in the season of activity was uneven; that is to say, an individual would move several hundred feet each day for a period of several days, and then, for an interval of one to several days, move only a few feet from one shelter to another, or not move at all. Such periods of rest could not be correlated definitely with environmental conditions; some individuals were inactive on days that were probably ideal (in terms of moderately warm temperatures and high humidity) for activity of box turtles. Analagous rest periods were noted inT. carolinaby Stickel (1950:358).
Two males ofT. ornatathat had been removed by me from their normal home ranges traveled the longest average distance per day (429 feet). Gravid females in June traveled the next longest average distance per day (363 feet). The average distances traveled per day by non-gravid females in June (226 feet) and July(260 feet) and by males (within their known home ranges) in June (289 feet) were thought to approximate normal amount of movement under average environmental conditions. Average distance traveled per day by females in October (152 feet) was shortest because of frequent and extended rest periods. Nevertheless, in October actual distances traveled on days of activity tended to be longer than in any other month. A gravid female traveled farther in a single day than any other individual ofT. ornataobserved; she moved along a rock fence for approximately 700 feet, then left the study area and moved, in a nearly straight line, 1,200 feet across a cultivated field. Then the thread on her trailer was expended. The total distance moved, therefore, was at least 1,900 feet and probably more.
Fig. 27.Average distances traveled per day by males and females at different times of the year, determined by mapping of thread trails at the Damm Farm. The diagram for "homing males" represents the distances traveled by two males removed from their normal home ranges to test homing ability. The data presented are for an aggregate of 136 days of trailing. Vertical and horizontal lines represent, respectively, the range and mean. Open and solid rectangles represent one standard deviation and two standard errors of the mean, respectively.
Fig. 27.Average distances traveled per day by males and females at different times of the year, determined by mapping of thread trails at the Damm Farm. The diagram for "homing males" represents the distances traveled by two males removed from their normal home ranges to test homing ability. The data presented are for an aggregate of 136 days of trailing. Vertical and horizontal lines represent, respectively, the range and mean. Open and solid rectangles represent one standard deviation and two standard errors of the mean, respectively.
An adult male at the Reservation traveled 2,240 feet in the 36-day period from October 16 to November 20, 1954, mostly on a woodedhillside. Eleven forms found along the route of the turtle's travels indicated that movement took place on roughly one out of three days in the elapsed period and demonstrated the sporadic nature of movements in autumn. The turtle remained active for an undetermined time after November 20.
Home Range
Data obtained from trailing and various methods of recapture at the Damm Farm indicated that each individual used only a small part of the total study area in the course of daily activities and tended to remain within a restricted area for a long time.
The number of recaptures of no individual was great enough to permit application of refined calculations of size of home range as described by Odum and Kuenzler (1955). For individuals that were recaptured six or more times, or individuals for which adequate trailing records were available, the area enclosed by a line joining the peripheral points of capture was considered adequately representative of the home range of that individual, unless recaptures were all within a few feet of each other or lay in an approximately straight line. If less than six records of recapture were available, home range was estimated, in the manner described by Fitch (1958:73), by averaging the distance between successive points of recapture and letting this average represent the radius of home range; the actual area of home range was determined by the formula, π(R)2, for the area of a circle.
Size of home ranges of males and females did not differ significantly and data for the two sexes were combined in the final analysis. The average radius of the home ranges of 44 adults (captured a total of 146 times) was 278 feet (extremes, 71 to 913) when computed by measuring the distance between successive captures; the average area of these home ranges was 5.6 acres. Data from 10 turtles that had been recaptured only once were combined with data from 34 turtles that had been recaptured more than once when it was found that the average size of home range in these two groups did not differ significantly. Data concerning the home ranges of eight of the 44 individuals were sufficient to permit actual measurement of home ranges with a planimeter; home ranges of these eight individuals had an average area of five acres (extremes, 1.2 to 10.2).
A minimum home range could theoretically consist of the smallest area in which adequate food and shelter were available. Under favorable conditions a turtle could stay in an area ten to twenty feet in diameter. Although several such favorable small areasexisted on the Damm Farm, box turtles seldom stayed in one for more than a day or two. Seemingly, therefore, factors additional to food and shelter influence size of home range. At the Damm Farm these additional factors seemed to be: rock fences that acted as physical barriers; areas that were cultivated, barren, or otherwise unfavorable, acting as ecological barriers; and, cowpaths and ravines that offered relatively unobstructed routes along which box turtles tended to move.
One subdivision of the main pasture, the northwest corner area, is an example of a relatively small natural area in which many individual box turtles had home ranges. This tract of 28 acres was roughly triangular and was bordered on two sides by rock fences that contained no gates or other passageways. On its third (southeastern) side the area sloped into a deep ravine. Habitat in this subdivision of the pasture (as well as in the other two subdivisions) was especially favorable for box turtles because of permanent water, rocky slopes, ravines, and several fruit trees. Box turtles usually foraged near the rock fences and the ravine (where dung was more abundant than in other parts of the area), and tended, as they foraged, to move parallel to these barriers. Turtles crossing the area eventually came either to one of the fences or the ravine. Therefore, most of the turtles in the northwest corner area eventually completed a circuit of the area. Turtles that came to the ravine tended to move along its bottom or sides. Several turtles were known to cross the ravine and to forage in the grassy area on its southeastern side. These turtles usually re-entered the ravine by way of smaller side-ravines. Of 22 box turtles known to have home ranges in the northwest corner area, only two individuals (both gravid females) were known to leave the area in the period in which observations were made.
Two other subdivisions of the main pasture—the house pond area and the southern ravine area—although not so distinct as the northwest corner area in terms of limiting barriers, nevertheless constituted separate areas of favorable habitat, each of which contained a number of individual home ranges. Although the two areas were not far apart, but little movement was observed of turtles from one area to the other. The home range of only one turtle, an adult female, was known to include parts of both areas.
Unbroken expanses of tall grass seem not to be optimum habitat. The crest of the hill at the Damm Farm (Pl. 17, Fig. 1) was an area of more or less homogeneous grassy habitat. Turtles were seldomfound on the crest of the hill although this area was as thoroughly searched for turtles as any other area. Known home ranges of nearly every individual observed were on either one of the sides of the hill but not on both sides.
At several places on the border of the pasture, turtles were able to move freely into cultivated areas but seldom did so except for nesting. Trailing records show that most of the turtles that entered one of the cultivated areas returned again to the pasture.
Ornate box turtles seem to find places of shelter by trial and error along regularly used routes of travel in their home ranges. The individuals that I studied never returned to the same forms, and seldom returned to the same natural burrows and dens. Probably foraging, basking, and watering sites are found also by trial and error.
Stickel (1950:375) placed considerable importance on the occurrence of transient turtles in populations ofT. carolina; in estimating population density, she added to her study area a peripheral strip, half as wide as the average, estimated home range, to account for turtles that had home ranges only partly within the study area. The study area used by Stickel had no natural boundaries, as habitat conditions on all sides were essentially the same as those of the study area itself. The pasture at the Damm Farm, on the contrary, is a relatively isolated area of natural grassland, bordered by rock fences and cultivated fields. I believe that most of the box turtles found on the pasture were permanent residents there. Individual box turtles at the Damm Farm seemingly occupied but one home range and it did not change from year to year. Populations ofT. ornatain areas less isolated than the Damm Farm, like the populations ofT. carolinastudied by Stickel (loc. cit.), could be expected to have a higher percentage of transient individuals and individuals with multiple or changing home ranges. Henry S. Fitch told me that he considered most of the individuals ofT. ornatathat were captured only once at the Reservation were transients.
Several females at the Damm Farm traveled long distances from their home ranges to nest but other females nested within their known or estimated home ranges. Seemingly a complex of environmental factors, including soil texture, weather, availability of water, and possibly the urge for random wandering in the breeding season, governs the distances traveled by gravid females and the ultimate selection of a satisfactory nesting site. Females, because of their more extensive travels in the nesting season, seem more likely thanmales to have multiple or changing home ranges. Males ofT. ornatadid not noticeably alter the extent or pattern of their movements in the breeding season. Hibernacula, unlike nesting sites, were within the known or estimated home ranges of all individuals studied.