Frequency in the Genetic System of Marsupials

Heredity 60 (1988) 77—85 The Genetical Society of Great Britain Received 11 March 1987

An examination of the role of chiasma frequency in the genetic system of marsupials

P. J. Sharp and Genetics Department, University of Adelaide, D. L. Hayman Box 498, Adelaide 5001, Australia.

Chiasma frequencies have been collected from males of 33 species of marsupial, varying widely with respect to size, reproduction and development. Correlation analyses show that species with lower levels of recombination are smaller, have larger litters and develop more quickly. Two indices of recombination level were considered: Darlington's Recombination Index, and the Excess Chiasma frequency (chiasmata above one per bivalent). The EC is in general more strongly correlated than RI with aspects of life history. It is suggested that EC levels have evolved due to the effects of EC on recombination, but that chromosome number has evolved independently.

INTRODUCTION MATERIALS AND METHODS The species studied Presentknowledge of mammalian chromosome evolution is largely limited to comparisons Inorder to make the best use of what is known of between G-banded chromosomes from different the cytology of marsupials, and in particular the species. Such studies consider evidence relating to knowledge of the chromosomal relationships the changes in chromosome structure which distin- between the different species, the species examined guish one karyotype from another. They utilise the in this study have been arranged into two groups features of chromosome number and morphology (table 1). obtained from mitotic divisions. There are, by way Those in Group 1 are all non-macropodid of contrast, very few comparative studies of species. The group contains two subgroups. The meiosis in mammals. Consequently, little informa- first contains species which all possess a mor- tion is available about the variation in the number phologically similar chromosome complement and distribution of chiasmata present between with a common major G-band pattern (Rofe and related species. Hayman, 1985). The taxonomically diverse species Comparative chromosome evolution has been in this subgroup show an extensive range of well studied in Australian marsupial species and adaptations. The few chromosome differences considerable detail is known about the mitotic between species are limited to those which occur complements (Rofe and Hayman, 1985; Hayman within a chromosome, such as inversions. The et a!., 1987). As an extension of these studies other subgroup includes species which do not have meiosis has been examined in some 33 species. the common 2n =14complement. There is a range This paper reports on the results of these studies of chromosome number from 2n =16to 2n =24 on chiasma frequency and considers a possible in this subgroup, and the majority of species are interpretation of the data based upon aspects of from the family Petauridae. the biology of the different species. Species in Group 2 are all members of the family Macropodidae. These species have similar biological features, are closely related in evolution- *Presentaddress: Plant Breeding Institute, Mans Lane, ary terms, and many pairs of species are capable Trumpington, Cambridge CB2 2LQ, U.K. of forming viable though sterile hybrids. The group 78 P. J. SHARP AND D. L. HAYMAN

Table IThespeciesstudied Number Superfamily Family Species examined

Group 1 Dasyuroidea Dasyuridae Sminthopsis crassicaudata 17 Sminthopsis murina Dasyuroides byrnei 4 Dasyurus hallucatus Dasyurus viverrinus 3 Sarcophilusharrisii 3 Perameloidea Peramelidae Peramelesgunnii 4 Perameles nasula Isoodon obesulus 3 Isoodon macrourus Vombatoidea Phascolarctidae Phascolarcios cinereus Vombatidae Lasiorhinus latfrons 2 Vombatus ursinus Phalangeroidea Petauridae Petaurus breviceps 2 Pseudocheirus peregrinus 3 Phaangeridae Trichosurus vulpecula 69* Trichosurus arnhemensis Tarsipedoidea Tarsipedidae Tarsipes rostra tus

Group 2 Macropodidae Potorous tridactylus Setonix brachyurus 3 Lagorchestes conspicillatus 2 Bettongia pen icillata Bettongia lesueur 2 Thylogale billardierii Wallabia bicolor Macropus parma Macropus eugenii 3 Macropus agilis Macropus parryi Macropus rufogriseus Macropus robustus Macropus giganteus Macropus rufus

* Datareported fully in Sharp and Hayman (1985) exhibits a considerable range in both chromo- nificant differences in chiasma frequency between some number (2n =12—22)and morphology. individuals within a given species, so that it is G-banding studies show that each specific chromo- desirable to examine meiosis in a number of some complement is derivable from a single 2n = individuals in arriving at an estimate of chiasma 22 complementby simple fusion events (Rofe, frequency. Where possible this has been done. In 1978; Hayman et cii., 1987). one species Trichosurus vulpecula an extensive analysis of chiasma frequency variation in 69 males Therecombination data showed no effect of either age or season (Sharp and Hayman, 1985). Meioticdivisions were obtained from the testes of There is now convincing evidence that chias- healthy adult males using the technique described mata at diplotene indicate the sites of recombina- by Sharp (1982) which produces clear preparations tion (Tease and Jones, 1978; Kando and Kato, of diplotene cells (fig. 1). At least 20 such cells 1980). The position of a chiasma may affect the were analysed in almost all the individuals genetic significance of the actual recombination examined, and the average chiasma frequency was event. In particular, chiasma localisation markedly calculated for each species. There may be sig- restricts the amount of recombination that can CHIASMA FREQUENCY IN MARSUPIALS 79

Figure1 Diplotene meiocytes from male (a) Isoodon macrourus, and (b) Lasiorhinus 1atfrons. The XY sex bivalents are indicated. occur. No evidence of strongly localised distal or These include indices of the rate of development proximal chiasmata was seen in any of the species and of reproduction since these were considered studied, although this does not rule out the possi- as likely to be correlated with the frequency of bility that a degree of chiasma localisation has recombination. occurred in one or more species. Further, both The metrics employed involve within and between species with the same chromo- (a) Size: The physical size of a species is represen- some number and morphology, such as the 2n =14 ted by head-body length (exclusive of tail subgroup, the chiasmata were distributed propor- length since not all species have tails) and tionally to chromosome length (results not shown). weight. Some of the species are sexually The average chiasma frequency was used to dimorphic with respect to size and in these the calculate two measures of recombination for each values given are the means of the sexes. species, the recombination index (RI), and the (b) Development: Three indices of the rate of excess chiasma frequency (EC).TheRI is the sum development were used; length of gestation, of the haploid chromosome number and the time from birth to weaning, and the age of the average chiasma frequency (Darlington, 1939), females of the species at their first breeding. and estimates the number of independently Where there may be delayed implantation segregating units at meiosis. Its use allows com- (embryonic diapause), especially in parisons of the total amount of resulting recombi- macropodids, the data given are the lengths nation to be made between species with different of gestation in the absence of diapause. The chromosome numbers. The EC has been defined time from birth to weaning includes the length as the number of chiasmata in addition to those of pouch life plus the time the young continue necessary to ensure regular segregation (Burt and to suckle after leaving the pouch. Bell, 1987), and is calculated as the difference (c) Reproductive potential: The two indices used between the average chiasma frequency and the were the number of young weaned per mating, haploid chromosome number. Since there are no and numbers of litters per female in a year. In chiasmata between the sex chromosomes in male macropodids (Group 2) this measure is com- marsupials (Sharp 1982) we have used the haploid plicated by the possibility that at any one time number of autosomes in calculating both indices females may have two young; one "at foot" (table 3). and still suckling, one in the pouch as well as a delayed blastocyst. The value given for such The species is a maximum number of young biological data weaned in a year. The annual reproductive Datawere collected from the literature on some potential of a species can be regarded as the biological features of the species studied (table 2). number of young weaned per litter multiplied 80 P. J. SHARP AND D. L. HAYMAN

Table 2Biological features of the species studied

age Head-body Birth to No. of at first Multiple length Weight Gestation weaning young perbreeding reproduction Sociality Species* (mm) (gms) (days) (days) mating (months) in a year? (rank score)

Group 1 Sc 75 15 13 65 6 5 — Sm 85 20 125 65 6 5 / — Db 157 110 33 110 6 9 — Dh 210 525 21 150 6 12 x — Dv 355 1,010 21 150 6 12 X — Sh 610 7,000 31 210 4 24 x — Pg 305 650 12 60 25 3 1 — Pn 380 1,000 125 60 25 4 '1 — Jo 315 775 125 65 3 4 1 — Im 375 1,600 125 60 3 4 1 — Pc 750 8,950 35 360 1 24 x — LI 850 26,000 — 300 1 36 x — Vn 860 20,000 — 300 1 24 x — Pb 170 130 21 110 2 12 x — P 325 900 — 180 2 12 v1/X — Tv 450 2,800 18 180 1 12 .J/X — Ta 410 1,450 — 180 1 12 v — Tr 70 10 — 70 25 8 '1 — Group 2 Pt 360 1,100 38 120 1 12 I 13 Sb 470 3,450 27 — 1 12 x 75 Lc 425 3,500 — 150 1 12 .J/x 13 Bp 330 1,300 21 180 1 6 75 BI 280 1,000 21 116 1 5 ./ 75 Th 590 5,500 30 290 1 14 .J/x 105 Wb 725 15,000 35 450 1 16 x 13 Mp 490 4,500 35 315 1 12 X — Me 615 6,500 28 255 1 9 X 5 Ma 725 15,000 30 330 1 12 X 4 Mpy 827 13,500 36 450 1 18 X I Mrf 745 16,900 30 450 1 14 X 105 Mrb 960 23,300 34 — 1 24 X 75 Mg 1100 49,000 36 540 1 16 x 2 Mr 1070 46,300 33 335 1 18 X 3 * Theinitials only of each species are given, the complete names are in table 1 Data collected from Wood Jones (1923—5), Ride (1970), Collins (1973), Tyndale-Biscoe (1973), Kaufmann (1974), Stonehouse and Gilmore (1977), Rose (1978), Wells (1978), Hyett (1980), Gemmel (1982), Lee etaL,(1982), Morton (1982), Rose and McCartney (1982), Russell (1982) and Strahan (1983)

by the number of litters per year. No numerical incorporated in the analyses. Statistical tests were value could be given for this latter component made using Spearman's non-parametric rank cor- as the available data are too sketchy, but an relation coefficient (r).Thiswas used since the indication is given of whether multiple repro- data do not satisfy the requirements for applying duction in a year is common, occasional or Pearson's correlation coefficients, and although the absent. individual biological datum points may be revised (d) Sociality: Kaufmann (1974) reviewed what is by subsequent work it is unlikely that this would known of social organisation in macropodids. change the ranking of the species for any charac- His bar graphs (based on observations of the TER. The significance of the calculated r values size and permanence of groups) were used to was tested by calculating a t-test value, or provide rank scores of sociality. in cases where the sample size was small (less In addition, comparative measures of nuclear than 12), by reference to table 87 of Bishop DNA content and chromosome number were (1980). CHIASMA FREQUENCY IN MARSUPIALS 81

Table 3Karyotypic features and meiotic recombination

Relative Average cell Species 2n DNA content* chiasma frequency RI ECt

Group I Sc 14 864 1364 1964 764 Sm 14 — 13'50 1950 750 Db 14 864 1740 2340 1140 Dh 14 864 18•75 2475 12'75 Dv 14 — 1893 2493 1293 S/i 14 864 2083 2683 1484 Pg 14 — 1266 1866 666 Pn 14 109 1398 1998 798 Jo 14 116 1627 2227 1027 Im 14 114 1630 2230 1030 Pc 16 933 2340 3040 1640 LI 14 1036 2243 2843 1643 Vu 14 102 2220 2820 1620 Pb 22 — 1490 2490 490 Pp 20 154 1680 2580 780 Tv 20 102 1814 2714 914 Ta 20 — 1655 2555 755 Tr 24 — 2295 4495 1195 Group 2 Ptt 13d, 122 — 2095 2695 1495 Sb 22 938 2928 3928 1928 Lcl 15d, 169 903 1624 2224 824 Bp 22 — 2840 3840 1840 Bi 22 — 23•35 3335 1335 Th 22 845 2465 3465 1465 Wb1 lid, 109 977 1418 1918 918 Mp 16 986 2345 3045 1645 Me 16 982 2300 3000 1600 Ma 16 977 1775 2475 1075 Mpy 16 1246 1690 2390 990 Mrf 16 1161 2090 2790 1390 Mrb 16 1041 3170 3870 2470 Mg 16 100 3235 3935 2535 Mr 20 1039 3790 4690 2890 * FromHayman and Martin (1974) using Mg as their standard. t RI and EC are defined in the text. t These two species have an X-autosome translocation as a regular feature of their karyotype. The chiasmata in the autosomal arm are included in the data. § This species has both X- and Y-autosome interchanges, and the chiasmata in the autosomal arms are included in the data

RESULTS the macropodids (Group 2), all species produce only one pouch young at a time, so this parameter Thedata on nuclear DNA content, chiasma is not included, but an analysis of the data on frequency and the calculated indices of recombina- sociality is included. The non-macropodids tion are shown in table 3. The data relating to the (Group 1) showed a similar pattern of correlation life histories of the species are given in table 2. coefficients to the analysis considering all the Correlation coefficients between the levels of re- species, whereas in the macropodids (Group 2) no combination and the components of life histories significant correlations were found apart from the and karyotypic features were calculated and are predictable one of RI with 2n. The analysis for the summarised in Tables 4-7. two large groups of species with the same chromo- Table 4 presents the results of the analyses per- some number, the 2n =14non-macropodids and formed on all the species. The majority of the the 2n =16macropodids are given in Table 7. correlations are significant, with correlations with The majority of the correlations for the 2n =14 the EC being in general of higher numerical value. group are significant, but none are significantly Tables 5 and 6 present the same correlation different from zero for the 2n =16macro- coefficients for Groups I and 2. In the analysis of podids. 82 P. J. SHARP AND D. L. HAYMAN

Table4 Correlations between various features and the two indices of recombination

RI EC Feature N r, Significance r Significance Head-body length 33 +0-41 +0•60 Weight 33 +0-43 +0-62 Gestation time 27 +046 +0-60 Birth-+ weaning time 31 +057 No. of pouch young 33 * 0-30 ns age at 1st breeding 33 +033 ns +052 DNA content 24 —001 ns —007 ns 2n 33 +068 +021 ns

Footnotes to Tables 4-7 ns=not significantly different from zero, *P<005,**P<001, P<0-001 N =numberof species in comparison. It varies within the tables because not all information was available for each species (see table 3). The degrees of freedom used in testing the significance of the r value =N—2 =Spearman'srank correlation coefficient

Table 5 Correlations between various features and the two measures of recombination in the non-macropodids (Group 1)

RI EC Feature N Significance r Significance * * Head—body length 18 +053 +056 * Weight 18 +0-51 +058 * Gestation time 13 +080 +067 * ** Birth -*weaningtime 18 +082 "' +061 No. of pouch young 18 —052 * —008 ns age at 1st breeding 18 +0-80 'k +063 ** DNA content 12 —005 ns —024 ns 2n 18 +059 ** —006 ns

DISCUSSION that those species with higher levels of recombination tend to be larger, slower developing, and to The approach adopted in this study has been to produce fewer pouch young. test the null hypothesis that there is no association The two large groups of species give different between aspects of life history and recombination. results. The non-macropodids (Group 1, table 5) The analysis presented in table 4 disproves this which cover a range of chromosome numbers, null hypothesis. All the correlations between re- show the same overall pattern of significant corre- combination and the life history aspects con- lations as in table 4. The macropodids (Groups 2), sidered are significant. The correlations suggest a more closely integrated taxonomic group, have

Table 6 Correlations between various features and the two measures of recombination in the macropodids (Group 2)

RI EC Feature N Significance Significance Head-body length 15 +017 ns +030 ns Weight 15 +016 ns +0'28 ns Gestation time 14 —029 ns —002 ns Birth -*weaningtime 13 —0-01 ns +008 ns Sociality 14 —044 ns +042 ns 9 age at 1st breeding 15 +009 ns +020 ns DNA content 12 ns +0-28 ns 2n 15 +073 ** +046 ns CHIASMA FREQUENCY IN MARSUPIALS 83

Table7 Correlations between various features and chiasma frequency in the 2n= 14species and the 2n =16macropodids 2n=14 2n=16 Non-rnacropodids Macropodids Feature N r Significance N r, Significance Head-body length 12 +069 7 +032 ns Weight 12 +075 7 +043 ns Gestation time 10 +080 7 +021 ns Birth -weaningtime 12 +088 6 +004 ns No. of pouch young 12 —020 ns — — — 9 age at 1st breeding 12 +088 7 +017 ns DNA content 9 —0'13 ns 7 —014 ns Sociality — — — 6 —026 ns a substantial range of chromosome numbers, and factors concerning reproductive rate where almost an especially wide range of recombination indices, all macropodids have only one young per year. In but have a limited range of life history strategies this situation, any correlation between either ges- compared to non-macropodids. The significant tation time or time from birth to weaning and correlations evidenced in non-macropodids are recombination may be expected to be less strong absent from macropodids (table 6). This dif- or absent. ference may be related to the fact that all the Two indices of recombination were considered, macropodids give birth to one young at a time and Darlington's recombination index and the excess are not as diverse a group with respect to reproduc- chiasma frequency. It appears that the EC might tive characters. This, however, still does not be a more useful index to use in comparative explain the role played by the substantial range in studies where chromosome number varies. In com- both recombination indices in macropodids. paring the full range of species (table 4) the EC However, as most of the data for macropodids is more strongly correlated with the biological came from one individual per species the between features. Additionally, in all comparisons, the EC individual variation (Sharp and Hayman, 1985) and chromosome number are not correlated. It has might have produced a greater variation in the been suggested that chromosome number varies recombination indices in this group. due to other pressures such as social organisation In table 7 the correlations are shown for the and drift (Wilson et al., 1975; Bengtsson, 1980), two large groups of species of common chromo- which would appear to operate irrespective of their some number. Those species with the conserved subsequent effect on the level of recombination, complement of 2n =14enable correlations to be so that only the EC might be easily open to considered where the morphological aspects of the evolutionary modification due to its effect on re- chromosomal phenotype are kept constant. The combination level (Darlington, 1932; John and extreme diversity of adaptations present in this Lewis 1975). The other karyotypic feature con- group shows clear correlations with chiasma sidered, DNA content, was not correlated with frequency which varies over a two-fold range. Only either of the indices of recombination. the correlation of recombination level and the The only study comparable to the present number of pouch young is not significant. This report is that of Burt and Bell (1987). They collec- calculation does not take into account the annual ted chiasma frequency data of 24 wild mammals number of young reared, which would be more (including four marsupials) from the published highly correlated with recombination as the literature and found a strong positive relationship smaller species have multiple reproduction each of the RI and EC with both age to maturity and year (table 2). In contrast, the correlations between weight, and negative correlations with litter size. recombination level and the biological features In addition, they found that the correlations with considered in the 2n =16macropodids were not EC were stronger than those with RI, and that EC significant, even though the species have a two-fold was not correlated with 2n. The two studies there- range of chiasma frequency. However, the underly- fore show the same associations, and demonstrate ing trend in the macropodids and non- the biological significance of the EC. Burt and Bell macropodids may well be the same, since the signs (1987) also found that domesticated mammals of the calculated correlation coefficients are the have higher recombination levels than expected same (compare tables 5 and 6, and the 2n =14and on the basis of the correlations in the wild species 2n =16groups in table 7), with the exception of (as suggested by Hayman (1981) by analogy with 84 P. J. SHARP AND D. L. HAYMAN results from plant studies), demonstrating that the ponents of life history are themselves inter-corre- EC can be adjusted quickly in evolutionary terms. lated though no single "Index of Life History" The associations that have been found in mar- exists or seems possible. Grant (1975) suggested supials here, and by Burt and Bell (1987), are that many aspects of the genetic system of different reminiscent of those that have been reported in species might be related to the ecological concepts plants (Grant, 1958; Stebbins, 1958). Additionally, of r-andK-selection (MacArthur and Wilson, the marsupials with low levels of recombination 1967; Pianka, 1970). Certainly marsupial species resemble the organisms recognised by Mather which are large, develop slowly and have few (1953) as having an apparent compromise between pouch young (characteristics expected under releasing variation by rapid population turnover, K-selection) tend to have higher recombination and conserving variation by having lower rates of levels than those which are small, rapidly develop- meiotic recombination. ing and with a greater production of pouch young As Burt and Bell (1987) indicate, the associ- (characteristics expected under r-selection). ations found across species between recombina- A deficiency of this study, as with almost all tion levels and biological features may help other studies of chiasma frequency, is that data distinguish between two models of the evolution have been collected from one sex only. Differences of recombination, the "Red Queen" and "Tangled in chiasma frequency and distribution are known Bank", reviewed by Maynard Smith (1978), Bell to exist between the sexes in some eutherian mam- (1982) and Charlesworth (1985). The Red Queen mals (Polani, 1972) and lower vertebrates (Callan model, which suggests that recombination has been and Perry, 1977). However, it would be remarkable selected because the biotic environment of each if the situation in females of the species considered species is continually becoming more hostile, pre- here was the reverse of that described in the males, dicts that species with longer generations will have so that the level of recombination of the species higher levels of recombination due to the stronger as a whole and their life history components were selection for genetic diversity they face in com- not correlated. parison to species with shorter generations. The A novel difference has been reported in the species with longer generations (those with slower marsupial Sminthopsis crassicaudata, where re- development) in both the mammals studied by combination is very much lower in the female than Burt and Bell (1987), and in the marsupials dis- in the male (Bennett et a!., 1986), but it is not cussed here do have higher levels of recombina- known if this is a feature of other marsupials. tion, lending support to the Red Queen model. Interestingly, S. crassicaudata has one of the lowest The Tangled Bank model proposes that recombi- male recombination levels (table 2), so the nation has been favoured as it produces genetically extremely low level in the female of this species diverse sibs which are each more likely to survive might possibly be an extension of the situation in sib competition, and therefore predicts that species the male. with more intense sib competition will have higher Clearly, a definitive assessment of how useful recombination levels. Unfortunately, data on levels data on chiasma frequency will be in accommodat- of sib competition are not available. Burt and Bell ing the different strategies employed by mammals (1987) assumed that sib competition is positively will depend on further studies. It would be surpris- correlated with litter size, and concluded that the ing, however, if such a potentially important lack of positive correlation between recombination feature played an insignificant role in the genetic level and litter size was evidence against the systems of mammals. Tangled Bank model. However, it may be that the level of sib competition is not strongly correlated Acknowledgements We are greatly indebted to the following people who provided or allowed access to the animals studied; with litter size, but depends more on factors such Prof. J. H. Bennett, Dr W. Breed, and Dr M. J. Smith (University as the degree of dispersal of sibs and the level of of Adelaide); Dr R. Baker (Adelaide Zoo); Dr D. King and sociality of a species. Certainly, the Red Queen Dr A. Oliver (APB of WA); Prof A. Bradshaw and Prof. J. model does not predict the negative correlations Shield (University of Western Australia); Dr J Nelson (Monash between recombination levels and litter sizes University); Pastor I. Wittwer, Cleve; Dr P. Baverstock (IMVS, Adelaide); Dr G. Lyne (CSIRO, Division of Wildlife observed both by Burt and Bell (1987) and here Research); and the South Australian and Tasmanian National in the marsupials. Parks and Wildlife Services. Mrs D. Golding helped with animal The correlation analyses reported here were handling and husbandry. PJS was the r cipient of a Common- performed separately for each component of life wealth Postgraduate Research Award, and we also thank the history as if these were independent of each other. MV. Ingham Trust for financial assistance. A. Burt and G. Bell kindly provided a copy of their paper before publication. We Clearly, this is not the case. The various corn- thank Dr R. N. Jones and Prof B. John for discussiois. CHIASMA FREQUENCY IN MARSUPIALS 85

REFERENCES MACARTHUR, R. H. AND WILSON, E. o. 1967. The Theory of Island Biography. Prnceton University Press. MATHER, K. 1953. The genetical structure of populations. Symp. BENNETF, J.H.,I-lAYMAN,D. L. AND HOPE, P. M. 1986. Novel sex difference in linkage values and meiotic chromosome Soc. Exp. Biol., ? 66-95. behaviour in a marsupial. Nature, 233, 59-60. MAYNARD SMITH, J. 1978. The Evolution of Sex. Cambridge University Press. BELL, G. 1982.TheMasterpriece of Nature. Croom Helm, London. MORTON, S. R. 1982. Dasyurid marsupials of the Australian arid zone; An ecological review. In Archer, M. (ed.) Car- BISHOP, 0. N. 1981. Statistics for Biology:Apractical guide for the experimental biologist, 3rd edition. Longman, New nivorous Marsupials, Roy. Zoo!. Soc. NSW., Sydney. PIANK.A, E. R. 1970. On r- and K-selection. Amer. Natur., 104, York. 592-597. BURT, A. AND BELL, 0. 1987. Mammalian chiasma frequencies: POLANI, P. E. 1972. Centromere localization at meiosis and the A critical test of two theories of recombination. Nature, position of chiasmata in the male and female mouse. 326,803-805. H. 0. AND PERRY, P. E. 1977. Recombination in male Chromosoma, 36, 343-374. CALLAN, ROFE, R. H. 1978. G-banded chromosomes and the evolution and female meiocytes contrasted. PhiL Trans. R. Soc. Lond. of Macropodidae. Aust. Mammal., 2, 52-62. B. 277, 227—233. ROFE, R. H. AND HAYMAN, D. 1985. G-band evidence for a a. 1985. Recombination, genome size and CHARLESWORTH, conserved complement in the Marsupialia. Cytogenet. Cell chromosome number. In Cavalier-Smith, 1. (ed.) The Evo- Genet., 39, 40-50. lution of Genome Size, Wiley-Interscience, Chichester. ROSE, R. W. 1978. Reproduction and evolution in female COLLINS, L. R. 1973. Monotremes and Marsupials—A Reference for Zoological Institutions. Smithsonian Inst. Press, Wash- Macropodidae. Aust. Mammal., 2, 56-72. ROSE, R. W. AND MCCARTNEY, D. .j.1982.Growth of the ington. Red-bellied Pademelon, Thylogale billardierii, and age esti- DARLINGTON, C. D. 1932. Recent Advances in Cytology. J. and mation of pouch young, Aust. Wildl. Res. 9, 33-38. A. Churchill, London. RIDE, W. D. L. 1970 A Guide to the Native Mammals ofAustralia. DARLINGTON, C. D. 1939. The Evolution of Genetic Systems. Oxford University Press, Melbourne. Cambridge University Press. RUSSELL, E. M. 1982. Patterns of parental care and parental GEMMEL,0.T.1982.Breeding bandicoots in Brisbane (Isoodon investment in marsupials. Biol. Rev., 57, 423-486. macrourus; Marsupialia, Peramelidae), Aust. Mammal., 5, SHARP, P. J. 1982. Sex chromosome pairing during meiosis in 187—193. marsupials. Chromosoma, 86, 27-47. GRANT, V. 1958. The regulation of recombination in plants. SHARP, P. J. AND HAYMAN, D. L. 1985. Variation in chiasma Cold Spring Harbor Symp. Quant. Biology. 23, 337-363. frequency in male Trichosorus vulpecula (Marsupialia: GRANT, V. 1975. The Genetics of Flowering Plants. Columbia Mammalia). Genetica, 66, 145-150. University Press. STEBBINS, 0. L. 1958. Longevity, habitat and release of genetic HAYMAN, D. L. 1981. Components of genetic systems in mammals. In Atchley, W. R. and Woodruff, D. (eds.) Evolution variability in the higher plants. Cold Spring Harbour Symp. and Speciation: Essays in Honor of M. J. D. White, Cam- Quant. Biol., 23, 365-378. STONEHOUSE. B. AND GILMORE, D. 1977. The Biology of Mar- bridge University Press. supials. Macmillan, London. HAYMAN, D. L. AND MARTIN. P. G. 1974. Mammalia I: STRAHAN, R. 1983. Complete Book of Australian Mammals. Monotremata and Marsupialia. Vol.4: Chordata 4. In John, Angus and Robertson, and the Australian Museum, Mel- B. (ed.) Animal Cytogenetics, Borntraeger, Berlin. bourne. HAYMAN, 0. L., ROFE, R. H. AND SHARP, P. .t. 1987. Chromo- TEASE, C. AND JONES, 0. H. 1978. Analysis of exchanges in some Evolution in Marsupials. Chromosomes Today, Vol. differentially stained meiotic chromosomes of Locusta 9,91—102. migratoria after BrdU-substitution and FPG staining I. HYETr,i. 1980. Austrialian Mammals—A field guide for New Crossover exchanges in monochiasmate bivalents. South Wales, South Australia, Victoria and Tasmania. Chromosoma, 69, 163-178. Nelson, Melbourne. TYNDALE-BISCOE, H. 1973. Life of Marsupials, Edward Arnold, JOHN, B. AND LEWIS, K. R. 1975. Chromosome Hierarchy: An London. Introduction to the Biology of the Chromosome. Oxford WELLS, K. T. 1978. Field observations of the Hairy-Nosed Wom- University Press. bat, Lasiorhinus latfrons (Owen). Aust. Wildl. Res., 5, KANDO, N. AND KATO, H. 1980. Analysis of crossing over in 299—303. mouse meiotic cells by BrdU labelling technique. WILSON, A. C., BUSH, 0. L., CASE, S. M. AND KING, M. C. 1975. Chromosoma, 78, 113-122. Social structuring of mammalian populations and the rate K.AUFMANN, J. H. 1974. The ecology and evolution of social of chromosomal evolution. Proc. Nat. A cad. Sci. USA, 72, organization in the kangaroo family (Macropodidae). 506 1—5065. Amer. ZooL 14, 51—62. WOOD JONES, F. 1923-25. The Mammals of South Australia— LEE, A. K., WOOLLEY, P. AND BRAITHWAITE, R. W. 1982. Life Parts 1—111. Government Printer, Adelaide. history strategies of dasyurid marsupials. In Archer, M. (ed.) Carnivorous Marsupials, Roy. Zool. Soc. NSW, Sydney.