This paper was presented at the Postgraduate Student Association Symposium in August, 1993.

Student : Susan Attard, BAppSc(UCQ) currently enrolled as an Honours student in the Faculty of Applied Science, Department of Biology

Supervisor : Stephen McKillup, BSc(Hons) Ph.D.(Adel) Veterinary Adviser : Gillian Swarbrick, BVSc(Qld)

 
Bandicoot obviously much distressed by capture

The reproductive ecology of free-living populations of the northern brown bandicoot, Isoodon macrourus, has only been studied at a few locations. This species exhibits either seasonal or year-round breeding. Photoperiod, rainfall and temperature have been suggested as likely cues controlling reproductive patterns. The species has not been studied in Central Queensland, nor does the climate here resemble that of other study sites.

Four populations were chosen : one unwatered site, one mulched site with trickle irrigation, one with adjacent sprinkling and one heavily watered and mulched. This allowed testing for two separate effects, rainfall and soil moisture. Sprinkler output was measured with a rain gauge. Soil samples were dried to a constant weight to calculate soil moisture. Each of the four sites were trapped for three nights each month, using rigid wire cages. Adult bandicoots were marked, weighed, their body measurements taken, and reproductive condition noted. Pouch young head lengths were taken and their developmental stage recorded.

Preliminary analysis indicates that females at the unwatered or lightly watered sites weigh less, are less likely to breed, produce smaller litters and are less successful in weaning young. At the heavily watered and mulched site the females are relatively heavier, continue to breed through winter, produce larger litters and are more likely to successfully wean their offspring. Lighter females at this site produce larger litters than females of similar weight at other sites. This site apparently contains only a small established population, which is isolated from other groups. The destination of offspring raised here and their adult size will be points of future research interest.

Introduction

This study was done in response to the lack of information on the reproductive ecology of Isoodon macrourus, the northern short-nosed bandicoot, in Central Queensland. This solitary, crepuscular or nocturnal marsupial is semi-fossorial, feeding on soil-litter invertebrates and small fruit. Males typically weigh about 1 - 1.5 kg [maximum recorded 3.1 kg], and females about 0.6 - 1.2 kg [maximum recorded 1.7 kg], depending on their age, the season of the year and the locality. Their length including tail is about 57 cm in males and about 48 cm females. The range covers the coastal areas from the Kimberleys through the monsoonal tropics of the Northern Territory, and from Cape York to the Hawkesbury River in New South Wales.

Breeding in this species appears to be seasonal in the northern and southern extremities of the range while in the central parts breeding is year round with some evidence of seasonality. Near Newcastle seasonal breeding occurred from July to April and this correlated strongly with bandicoot body weight increases and less strongly with rainfall patterns. Most rainfall in this region occurs in autumn / mid winter but because breeding can continue despite lack of rain and fail early despite good rains, Gordon (1971) suggested rainfall was a modifying influence through food source. Variation in commencement of breeding both temporally and spatially was presented as evidence against photoperiod being a primary cue. Termination of breeding correlated with poor rainfall and subsequent food shortage.

Gordon (1974) trapped near Brisbane and found no seasonality in breeding activity with one female reported to have produced six litters in 13 months, an average of one litter every 56 days. No information on possible breeding cues was offered.

However also near Brisbane Gemmell (1982) reported a lack of wild bandicoot breeding from April until June. Year-round breeding was reported in captive colonies kept in outside enclosures but births peaked from August to October and were at their lowest from March until June (Gemmell, 1988). There was little definition of seasonality in breeding. Gemmell (1990, and Barnes 1984) has suggested either temperature or photoperiod as the primary cue initiating breeding in the Brisbane region.

Finally Hall (1983) also trapped around Brisbane and reported year-round breeding with a peak number of litters produced in August after a low in June. Hall found no statistical evidence for periodicity in breeding activity and no cues were suggested.

Kemper et al (1990) on the Mitchell Plateau found births occurred from September to April with a peak in January. No breeding cues were suggested in this paper. Friend (1990) near Darwin found I. macrourus births occurred from late August until mid April with a peak between October and March. Both these studies were done in the monsoonal wet-dry tropics where about 80% of total rain falls from November or December through until March. Friend suggested rainfall as a major determinant of seasonal breeding in these regions. Since daylength varies little in the tropics and temperatures are high throughout the year, neither of these factors are likely cues to trigger breeding. However rainfall is highly seasonal and this affects the density of invertebrates, the bandicoots primary food. This in turn influences bandicoot body weights and body weight is a major determinant in mammalian breeding (Widdowson, 1981).

Data are needed on animals living in natural surroundings around Rockhampton. There is no information on the ecology or reproductive biology of local populations of Isoodon macrourus in the literature and the climate in Central Queensland does not resemble that of the previous study sites. Additionally evidence is strong for rainfall as the major determinant of bandicoot breeding due to its effect on soil-litter invertebrate numbers. Bandicoots which are living at artificially watered sites may exhibit different reproductive patterns from those in natural settings.

The foci of this study are to monitor wild populations and to compare populations at sites which receive "rainfall only" with those which receive "rainfall and sprinkler" input. Is there seasonality of wild bandicoot breeding in Central Queensland? Can a comparison of "rainfall only" versus "rainfall and sprinkler" watered populations indicate cues used to initiate breeding in the central part of the range? Is there a significant difference between the populations regarding reproductive patterns?

Methods and Materials

Animal Experimentation Ethics Committee approval has been gained for all procedures used in this study. The data has been collected by a trap / mark / release / recapture method over 3 nights per site at 4 sites per month using treadle or hook activated Sherman traps baited with rolled oats and peanut butter. The four sites studied were as follows. One site received only natural rainfall. A second site received rainfall and some sprinkler input, but the sprinkled area was small and only partly mulched. The third site, small in area and not sprinkled, was trickle irrigated and thickly mulched. A fourth site not only received rainfall, but was heavily and widely sprinkled and thickly mulched. The third site was used as a control for the effect of soil moisture but without the cue of rainfall.

Rainfall at all sites was deemed to be that recorded by the Bureau of Meteorology in Rockhampton. Sprinkler augmentation was measured by a rain gauge. Readings were taken from two locations at each site each month and then averaged, where possible. With reduced sprinkling in colder months, this was not always possible and one reading was then used to calculate the month's record. Soil moisture content was analysed by taking triplicate samples from all sites each month, weighing the fresh sample, drying it to a constant weight, then calculating the percentage weight lost. The samples were taken from diverse locations in each site.

Initially all the animals were to be anaesthetised for examination but concerns were raised as to both the necessity for and the dangers involved in this procedure. Some of the larger males have been anaesthetised using a simple vaporiser, in which compressed Oxygen is passed through an air flow regulator and vaporises the anaesthetising agent, contained in a glass bottle. This then passes to a muzzle where the vaporised anaesthetic is inhaled by the animal. This method has not been used on most occasions.

Identification of all adults has been by ear marking with a leather punch following a standard code. Measurements of all adult individuals have been recorded to compare with earlier work (ie weight, pes, manus, ear, head and tail lengths). Head lengths of pouch young have also been recorded to allow estimation of their age, from which the birth date and hence the month in which their parents bred, can be estimated. Hall (1990) found that head length was a reliable indicator of the age of pouch young and the equation presented in that paper was used in this study. Since Hall (1990) also presented a table of developmental stages in the growth of pouch young of I. macrourus which were stated to occur with 'small variations in timing' (p. 124), this table was used to generate a checklist of characteristics for additional estimation of age.

Results

The following data are for five months of trapping at each of four sites. The sites were : 1 the UCQ Ecological Reserve, 2 the Limestone Creek Environmental Reserve, 3 the UCQ Biology Compound and 4 the Kershaw Gardens. It is recognised that this study was pseudoreplicated but this was unavoidable due to time and trap limitations. Also the small sample size obtained prevents statistical analysis of the results. Weights graphed were monthly mean weights where a female's weight varied between captures.

In the UCQ Ecological Reserve a total of four females and six males were trapped, but one of these females and three males were subsequently trapped more frequently in the Biology Compound. One female was carrying a litter about one week old in late March but these were not present (and had apparently been ejected) when trapped a month later. None of the three remaining females bred. Tracks of an adult and at least one juvenile observed in late April may indicate that a female produced a litter at this site in late February. Graph 1 shows that these females were all below 1.2 kilograms and their weight remained constant or decreased throughout the study.

At the Limestone Creek ER, (which included a small lightly sprinkled area) two females and four males were trapped. Neither of the females had pouch young, or pigmented nipples indicative of a suckling litter. However they were not caught frequently and this may reflect the availability of an alternative feeding area. Infrequent trapping may prevent detection of a litter. Graph 2 gives the weights of these female; none exceeded 1.1 kilograms and again weights remained relatively constant for the period of the study.

At the Biology Compound, (which was trickle irrigated) two females and six males were trapped although some males were only been recorded once at this site. One female moved into this area from the first site and did not produce a litter here until late July when four pouch young were found. The second female produced a litter of 3 in early May but ejected that litter in late May when the young were developmentally retarded. This female produced a litter of one male in late June which grew successfully until five weeks of age. The fate of this offspring is unknown. Graph 3 shows the weights of females at this site, these ranged from 1.15 to 1.7 kilograms. Females at this site are heavier than at the previous two sites, and gained weight around mid July. The second female continued to gain weight during the second lactation, in contrast with the previous litter.

In the Kershaw Gardens, (which was heavily watered and mulched) four females and four males were trapped. All females trapped were either carrying a litter or showed signs of suckling a nested litter. These females were trapped frequently and development of their young followed closely. Graph 4, documenting their body weights, clearly shows the much heavier weight of Number 6 throughout the earlier part of the year, and the subsequent decrease in weight over winter. This female weighed 2.0 kilograms in April and this is the heaviest wild-caught female of this species ever recorded. The other females weighed less throughout the year, and were similar to those from the trickle irrigated site, but bred regularly throughout the study period in contrast to females of other sites. Females at this site produced between 3 and 5 offspring per litter in eight litters. Four litters were known to be raised to within a week of pouch independence.

Rainfall data are shown in Graph 5. Although incomplete, this demonstrates differences in total precipitation amongst sites. Sprinkling rates decreased over the colder months. These figures can only be taken as a guide due to uneven application from sprinklers and 'rain shadows' caused by trees. Both sites would therefore have an heterogeneous pattern of rainfall and the readings vary depending on the location of the rain gauge.

Soil-litter moisture data are shown in Graph 6. The UCQ Ecological Reserve has a high clay content and there was some difficulty in drying this to a constant weight. This water-holding ability of the soil means most of the soil moisture at this site is physiologically unavailable. At the Limestone Creek Environmental Reserve samples were taken from both watered and dry locations but the sandy soil holds little water. Similarly the Biology compound samples were taken from watered and dry locations since animals foraged in both areas. The Kershaw Gardens samples were from heavily watered areas with abundant organic material.

Discussion

The differences in breeding patterns between populations in this study cannot be explained by temperature or photoperiod cues, since all sites were within seven (7) kilometres of each other and intersite variations in these factors were negligible.

Gemmell (1993) suggested that ovulation in I. macrourus is inhibited by melatonin, levels of which are elevated at night. Therefore breeding commences as the length of the night decreases. This primary photo-periodic cue would be modified by extrinsic factors, such as availability of food. A nutritionally stressed population might be expected to have a longer non-breeding season since ovulation could be inhibited by undernourishment as well as high melatonin levels. However a population with adequate nourishment would still be expected to have a non-breeding season due to the inhibitory effects of elevated melatonin. My study does not support suppression of ovulation by melatonin in the populations studied in Rockhampton, since females at two sites bred in the shorter days of winter.

Temperature is unlikely to be a significant inter-site influence in this study. Three sites were on open ground and the females foraging in the UCQ Biology Compound were probably nesting in or close to the UCQ Ecological Reserve. The Kershaw Gardens site was more sheltered but the animals were nesting in an adjacent gully where cold air could pool. The minimum temperatures in the Rockhampton region over the period of study were above average. If temperature were a primary factor in preventing females breeding, then most or all females would be expected to respond in the same way. If the temperatures were inhibitory, then few or no females would breed. Temperature can be an important factor since Bronson (1985) emphasises that thermoregulatory costs while foraging in mild temperatures can offset any energy gains made by small mammals. A female may then be only maintaining or losing weight and so reproductive attempts will be abandoned or not attempted. Any effect of temperature in this study would be as a modifier, such that heavier females may continue breeding while lighter animals would be unable to do so.

Female body weight is a recognised factor in reproductive condition (Perry, 1971). There is evidence to show that ovulation does not occur in humans unless there is sufficient energy available to maintain a subsequent pregnancy (Hogarth, 1978). Females of many mammalian species do not commence breeding until they have attained a species-specific body weight (Perry, 1971). In Friend's study a critical minimum weight of 550 grams was found, below which no female bred. In this study a critical minimum of approximately 1.0 kilogram was found, and below this weight no female indicated a recent attempt at breeding, while some females above this weight also did not breed.

Isoodon macrourus is reported to feed primarily on insects and other invertebrates but will also eat plant material (Gordon, 1974, 1983). Since it is established that insect abundance (Wolda, 1978) and litter arthropod populations (Levings and Windsor, 1982) in the tropics vary seasonally with numbers positively correlated with rainfall, higher rainfall and soil moisture levels would be expected to result in greater densities of the bandicoots' preferred food. This would result in heavier body weights which in turn would lead to greater success in breeding. The results of this study corroborate the suggestion of Friend that 'food availability ... has an important influence on breeding' (1990, p. 363).

Since body weight depends on food supply, nutrition is important to mammalian breeding. The higher critical weight found in this study can be related to climatic predictability. In the tropics, commencement of the monsoon is usually followed by several months of reliable rain (Friend, 1990, pp. 363-4). Hence any reproductive attempt will usually be supported by an ongoing supply of food. Selection will favour females which commence breeding in the earliest rainfalls although they are without body fat stores. In a less predictable climate with irregular and short term rainfall, females cannot be assured of a steady food supply. In this environment, selection will favour females which only attempt reproduction when they carry adequate energy stores to maintain their own condition and provide for the demands of their offspring.

The female bandicoots of the Biology Compound, already heavier than the critical minimum, gained weight and bred after natural rainfall, in early June and in early July, which augmented the trickle irrigation. This is interesting because few conical digging holes were observed at this site, suggesting that most of the food intake was of litter- or surface-dwelling insects. If the abundance of insects were positively correlated with rainfall, an increase would be expected after rainfall, leading to an increase in body weight of bandicoots. This did not occur in the UCQ Ecological Reserve or the Limestone Creek Environmental Reserve, where the female weights were below the critical minimum necessary for reproduction. The rainfall and resultant invertebrates were insufficient to overcome these low weights.

The relationship between body weight and reproductive success is evident in the female bandicoots from the Kershaw Gardens. Seasonal breeding in a captive population in Brisbane (Gemmell, 1988) where nearby wild populations exhibited year-round breeding (Gordon, 1974; Hall, 1983) suggests some inadequacy of the diet supplied to the captive animals. If the captive population were nutritionally stressed, cooler temperatures in late autumn / winter could be enough to prevent females breeding. This would give the appearance of seasonal breeding which may then have been attributed solely to the stimulus of short daylength.

Conclusions

Continual breeding was only found at one of the sites in this study. At the heavily watered and mulched site, females bred throughout the study period, raised young to late lactation and their body weights were on average heavier than those at other sites. At the trickle irrigated site, breeding occurred infrequently and only one young was raised until late lactation. These females were lighter throughout the study, but gained weight slightly in July. At the remaining two sites, only one litter was known to be produced and this was known to be lost before weaning. At these two sites the average female weights were lighter than at the two thickly mulched sites. These weights changed little throughout the study, suggesting that the animals are existing on minimal nutrition.

Since body weight results from food abundance, breeding in this species is ultimately dependent on rainfall, as demonstrated by the results of this study. The precise pattern of breeding is determined by local factors such as season of rainfall, soil type and surface covering.

Gemmell's finding that captive populations exhibited seasonal breeding while nearby wild populations did not suggests a dietary inadequacy of the captive animals.

Future research will address several questions raised by the results of this study. The UCQ Ecological Reserve should be monitored closely since there is presently little or no recruitment and it appears to be a population sink. The Kershaw Gardens population appears to be expanding into new areas. The source of this population may have been only a few individuals, since the 1990-91 flood is likely to have removed any previous population. Recolonisation by a steady process down Moore's Creek is unlikely since there has been no indication of bandicoots above the present site until very recently. The genetic makeup and relationships of the present population need to be ascertained and to this end DNA analysis of tissue samples needs to be done. Accurate identification of pouch young and ongoing monitoring of the population will be necessary. The recruitment rate, destinations and adult size of the pouch young produced at this site need to be recorded.

Acknowledgements

I would like to thank the following people:

my Honours supervisor, Dr. Stephen McKillup for direction in project design and editorial comments,

Dr. Lesley Warner for the use of her traps,

Gillian Swarbrick for veterinary advice and assistance,

Graeme Turner for supply of some anaesthetic equipment,

CIG Ltd. for special use of a Medical Oxy regulator,

Rockhampton City Council for permission to trap in the Kershaw Gardens,

Mr. Kevin Quinn and other staff of the Kershaw Gardens for their advice and assistance,

postgraduate students Donna Rayner, Peter Wright and Margaret Winter for their fellowship and support,

and my husband Ron Attard for his assistance.

Bibliography

Barnes, A., Gemmell, R. T., 1984, 'Correlations between breeding activity in the marsupial bandicoots and some environmental variables' Aust. J. Zool. 32:219-26.

Bronson, F. H., 1985, 'Mammalian reproduction: an ecological perspective' Biology of Reproduction 32, 1-26.

Friend, G. R., 1990, 'Breeding and population dynamics of Isoodon macrourus (Marsupialia: Peramelidae): studies from the wet-dry tropics of northern Australia' pp. 357-65 in Bandicoots and Bilbies, Seebeck, J. H., Brown, P. R., Wallis, R. L. and Kemper, C. M. (eds.), Surrey Beatty and Sons, Sydney, Australia.

Gemmell, R. T., 1982, 'Breeding bandicoots in Brisbane (Isoodon macrourus; Marsupialia, Peramelidae)' Aust. Mammal. 5:187-193.

Gemmell, R. T., 1988, 'Breeding season and litter size of the bandicoot, Isoodon macrourus (Marsupialia: Peramelidae), in captivity' Aust. Mammal. 12:77-79.

Gemmell, R. T., 1990, 'The initiation of the breeding season of the northern brown bandicoot Isoodon macrourus in captivity' pp. 205-212 in Bandicoots and Bilbies, Seebeck, J. H., Brown, P. R., Wallis, R. L. and Kemper, C. M. (eds.), Surrey Beatty and Sons, Sydney, Australia.

Gemmell, R. T., 1993, 'Control of seasonal reproduction in Australian marsupials - effects of biotic factors' in press.

Gordon, G., 1971, A study of island populations of the short-nosed bandicoot, Isoodon macrourus (Gould), Ph.D. Thesis, University of New South Wales, Kensington, New South Wales.

Gordon, G., 1974, 'Movements and activity of the short-nosed bandicoot Isoodon macrourus Gould (Marsupialia)' Mammalia 38(3):405-431.

Gordon, G., 1983, Northern brown bandicoot Isoodon macrourus, pp. 96-96 in :The Australian Museum complete book of Australian mammals, ed. Strahan, R., Angus and Robertson, Australia.

Hall, L. S., 1983, 'Observations on body weights and breeding of the Northern Brown Bandicoot, Isoodon macrourus, trapped in south-east Queensland' Aust. Wildl. Res. 10:467-76.

Hall, L. S., 1990, 'Growth and a description of the development of external features of pouch young of captive Isoodon macrourus' pp. 123-33 in Bandicoots and Bilbies, Seebeck, J. H., Brown, P. R., Wallis, R. L. and Kemper, C. M. (eds.), Surrey Beatty and Sons, Sydney, Australia.

Hogarth, P. J., 1978, Biology of Reproduction, Blackie and Son Ltd., Glasgow, Scotland.

Kemper, C., Kitchener, D. J., Humphreys, W. F., How, R. A., Schmitt, L. H., Bradley, A., 1990, 'The biology of the Northern Brown Bandicoot, Isoodon macrourus (Marsupialia : Peramelidae) at Mitchell Plateau, Western Australia' Aust. J. Zool. 37:627-44.

Levings, S. C., Windsor, D. M., 1982, 'Seasonal and annual variations in litter arthropod populations.' pp. 355-87 in: The ecology of a tropical forest - seasonal rhythms and long term changes, eds. Leigh, E. G., Rand, A. S. and Windsor, D. M., Smithsonian Institution Press, Washington, D. C..

Perry, J. S., 1971, The ovarian cycle of mammals, Oliver and Boyd, Edinburgh, Scotland.

Widdowson, E. M., 1981, The role of nutrition in mammalian reproduction, pp. 145-59 in: Environmental factors in mammalian reproduction, eds. Gilmore, D., Cook, B., Macmillan Publishing, London.

Wolda, H., 1978, 'Seasonal fluctuations in rainfall, food and abundance of tropical insects.' Journal of Animal Ecology 47:369-81.

Email Susan at Creekholme

Return to

home page