The leopard tortoise (Geochelone pardalis) is the largest of the southern African tortoise species and has a wide distribution range. However, there is a lack of ecological and physiological information about the species, especially arid and semi-arid regions. The Nama-Karoo, an arid region of South Africa, is subject to large fluctuations in rainfall, food availability and ambient temperatures (Ta). This study focused on the thermal behaviour, thermoregulatory, digestive and metabolic plasticity of the leopard tortoise within the Nama- Karoo biome. Seasonal changes in activity patterns and body temperature (Tb) were investigated in free ranging leopard tortoises in the Nama-Karoo. Leopard tortoises had unimodal daily activity patterns in winter, bimodal in summer, and there were daily and seasonal differences in the extent to which certain behaviours were practiced. Daily activity behaviours were executed at lower Tb and at lower Ta in winter compared to summer. In summer, core Tb of all tortoises oscillated on a daily basis well below maximum Ta, while core Tb of all tortoises in winter oscillated well above the daily Ta range. Tortoises were therefore able to maintain their Tb independently of Ta. Differences in Tb as measured from various positions on the tortoises body was investigated in relation to Ta. There was a strong seasonal and temporal influence on the relationship between various Tb’s, with the skin and external shell temperatures being more variable in response to fluctuating Ta’s compared with cloacal and core Tb. Cloacal temperatures were significantly different to other Tb measurements suggesting that it should be treated with circumspection as an exclusive measure of Tb. Heating and cooling rates of leopard tortoises were investigated in the field and under controlled laboratory conditions to determine if the tortoises maximise operational daily activity periods, and to determine the effect of behaviour and size on the rate of heat flux. In the laboratory, cooling rates were faster than heating rates in summer and winter for all size classes and decreased with increasing body mass. Leopard tortoises had significantly faster heating and cooling rates in winter than in summer. Free-ranging leopard tortoises had faster heating rates than cooling rates and their heat flux was largely independent of Ta. Heating and cooling rates were dependant on body mass and surface area-to-volume ratio of individuals. Under experimental conditions, tortoises physiologically adjusted their rate of heat flux, while free-ranging tortoises used physiological and behavioural mechanisms to minimise the risk of overheating, to aid thermal inertia and maximise operative activity time. Seasonal climatic cycles and fluctuating daily temperatures influence the oxygen consumption (VO2) of reptiles, however the result of these effects on metabolism in chelonians is poorly understood. The effect of seasonal and daily differences in Ta on VO2 was investigated. Leopard tortoises’ VO2 was slightly higher than reported for other chelonians. There were significant differences in tortoise VO2 at different Ta’s during the day and night and in different seasons. This metabolic plasticity is possibly an adaptive mechanism to cope with unpredictable environmental conditions. Unpredictable climatic conditions lead to unpredictable food and water availability. Little is known how tortoises adjust dietary parameters in response to food type and water availability, and if this affects body mass, energy and water balance. Therefore this study also considered whether leopard tortoises adjusted food transit rate, food intake and water loss to cope with a diet fluctuating in fibre and water content, and whether body mass, energy and water balance were maintained. Leopard tortoises fed a high fibre, low water content diet had lower food intake rates, longer food transit times, but lower daily energy assimilation compared with tortoises fed a low fibre, high water content diet. Tortoises fed a high fibre, low water content diet had lower urine osmolality, but similar total water loss to those fed a high fibre, low water content diet. The results indicate that tortoises can adjust digestive parameters according to diet composition and exercise some control over energy and water balance. It is concluded that leopard tortoises show a high degree of plasticity in their thermal behaviour and physiology which allows survival in an unpredictable environment, particularly where there are fluctuations in rainfall, food availability and Ta’s. Seasonal and daily variation in thermoregulation, metabolic rate and the uptake of energy allows the leopard tortoise to maximise the duration of operative temperature, to minimise energy loss and to use variable and unpredictable seasonal resources. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2007.

M.Sc. / Research into blood protozoans (haematozoans) infecting African tortoises is scanty with only a few records published, many during the early part of the last century. Little research had been done on the blood parasites of tortoises examined in this study namely, Kinixys lobatsiana, K. belliana belliana, K. natalensis, Geochelone pardalis pardalis, G. pardalis babcocki and Chersina angulata. The study therefore aimed to: 1) examine apicomplexan haematozoan parasites infecting several of South Africa’s indigenous tortoises and compare them with published species descriptions, especially from neighbouring Mozambique; 2) provide host details (identity, ectoparasites, host weight and gender, effects of blood parasites on host cells) and locality records in different seasons for described and new apicomplexan species; 3) describe new and recorded parasites using morphometrics and, if possible, ultrastructural characteristics 4) attempt apicomplexan DNA extraction, amplification and, if feasible, purification; and 5) establish a basis for future research as a result of the acquired knowledge. During the current study, 154 tortoises of six species in three genera, both captive and wild, and from four South African provinces (Gauteng, North West, Kwazulu-Natal and Western Cape) were sampled. Giemsa stained blood smears and use of image analysis enabled morphometric analysis of the apicomplexans and their effects on host cells, while some blood preserved in Karnovsky’s and Todd’s fixatives received detailed examination by transmission electron microscopy. Lastly, blood preserved in lysis buffer during collection, and with the highest parasitaemias, was subjected to parasite DNA extraction and amplification. Comparisons between a published account of apicomplexans recorded from K. b. belliana in Mozambique, and those found in the current study, identified two haemogregarine species. In the present research, Haemogregarina fitzsimonsi Dias, 1953 infected 2/27 (7%) wild North West K. lobatsiana, 2/3 (66%) captive Kwazulu-Natal K. natalensis, 7/14 (50%) captive Kwazulu- Natal K. b. belliana, 3/6 (50%) captive Kwazulu-Natal G. p. pardalis, 2/41 (5%) wild G. p. babcocki and 13/37 (35%) captive Gauteng G. pardalis. In addition, Haemogregarina parvula Dias, 1953, infected 2/14 (14%) captive K. b. belliana and 1/10 (10%) captive G. p. pardalis. An unknown species of haemogregarine, possibly also H. fitzsimonsi occurred in 6/16 (38%) Chersina angulata from the Western Cape. As well as haemogregarines, two haemoproteids were identified: Haemoproteus balazuci Dias, 1953 infected 2/27 (7%) wild North West K. lobatsiana, 2/2 (100%) captive Gauteng K. lobatsiana and 1/41 (2%) wild North West G. p. babcocki; Haemoproteus sp., a likely new species, was found in 1/3 (33%) captive K. natalensis. Infections with Haemogregarina and Haemoproteus were not concurrent in this study, but were found to occur concurrently in Dias (1953) findings, and only the two Haemogregarina spp. occurred together in captive Kwazulu-Natal G. p. pardalis tortoises, which do not occur naturally in the region. Haemogregarina fitzsimonsi did not appear region or host specific, since it infected 5/6 species of tortoises from all provinces sampled. Haemogregarina parvula apparently existed only in tortoises from Kwazulu-Natal. Furthermore, captive Gauteng female tortoises were found to have a higher rate of infection than males and heavier tortoises showed a lower intensity infection than lighter and younger tortoises. On average season appeared to have a slight affect on parasite prevalence, with a higher prevalence during the summer rather than the winter, possibly a result of the activity of the assumed vector, which may be the tick species Amblyomma marmoreum (found on G. pardalis) and/or Amblyomma hebraeum (found on C. angulata). For the new Haemoproteus sp., the small sample size meant that meaningful data on host-specificity and range was not gathered, but Hp. balazuci occurred in K. lobatsiana in the drier regions of the North West and Gauteng. Although DNA extraction was possible for H. fitzsimonsi, the technique requires further refinement and samples with greater parasitemias before it can be used with additional material, and sequencing can be attempted. Thus, new localities, hosts, host data and possible vectors (ticks) were recorded for the apicomplexan species identified by Dias (1953) and they were re-described using modern techniques. Also, possibly new Haemogregarina and Haemoproteus spp. were recorded, but their identity requires confirmation by DNA analysis. It is anticipated that these, and future results, will increase the knowledge of the ecology and biodiversity of apicomplexan haematozoans parasitising chelonian hosts in South Africa, with possible application to the conservation of these and other tortoise species around the world.

Particle size distribution of diet, feces, and change from diet to feces, as well as apparent digestibility (aDig, %) of selected nutrients, and novel fecal microbiome identification of mature female leopard tortoises (Stigmochelys pardalis, n = 16) fed exclusively one of three, nutritionally complete, pelleted diets were evaluated in a blind, complete randomized design study. Two diets included insoluble fiber (powdered cellulose) consisting of either 2.0 mm or 0.2 mm length. Insoluble fiber provides nutritional and physical benefits to both the animal host and the microorganisms that inhabit the gastrointestinal tract. Insoluble fiber length was used as a means of evaluating a physically effective fiber (peNDF) definition for hindgut-fermenting vertebrates. Numerical trends of each diet particle size distribution indicated a greater amount of particle recovery on the 2.0 mm sieve for the 2.0 mm diet, and a greater particle recovery on the 0.125 mm sieve for the 0.2 mm diet, both as expected based on the added fiber lengths. Fecal particle size distributions were not different between diets, however, distributions of the change in particle size from diet to feces were different between diets. Similar fecal particle size distributions across diets suggests both cellulose lengths are below the peNDF threshold of the leopard tortoise. Apparent digestibility (aDig, %) of dry matter (DM) and organic matter (OM) was not different based on diet, method, or a diet and method interaction; aDig (%) of neutral detergent fiber (aNDF) and sequential acid detergent fiber (sADF) was different based only on diet. These results suggest that while aDig (%) of OM did not change, the source of OM digestibility shifted, as both aNDF and sADF digestibility increased with the cellulose-added diets compared to the control diet. An increase in insoluble fiber digestibility suggests an "effectiveness" of the cellulose lengths. At both bacterial phyla and genera levels, fecal microbiomes were more similar between tortoises fed the cellulose-added diets versus the control diet, suggesting that the hindgut microbial communities adjusted in the hindgut of the tortoises fed the cellulose-added diets by shifting proportions of microbes, based on their role in the hindgut (i.e., cellulose digestion), to accommodate for the addition of cellulose in the two treatment diets. This may explain the similarity among fecal particle size distributions, and suggests that adaptability of the hindgut microbial communities should be considered when defining peNDF for hindgut-fermenting vertebrates.