BiulqictliJourrial oj"the Limean Suciep (2000); 71: 117 135. M'ith 3 figures
doi: 10.1006/1~~1.2000.045~,
available online at http://www.idealibrar);.com on
0 E hL"
The evolution of reproductive strategies in
dasyurid marsupials: implications of molecular
phylogeny
CAREY KRAJEWSKI*
Department of <oology and Centerfor Systematic Biology, Southern Illinois Uniuer&,
Carbondale, Illinois, 62901-6501, U.S.A.; Department o f Genetics, La Trobe Uniuersip,
Bundoora, Ectoria, 3083, Australia
PATRICIA A. WOOLLEY
Department ~f.i7ooLogy,LA Trobe University, Bundoora, Ectoria, 3083, Australia
MICHAEL WESTERMAN
Department of Genetics, La Trobe Universig, Bundoora, Ectoria, 3083, Australia
Receioed 27 September 1999; accepted jor publication 25 April 2000
Dasyurid marsupials show a remarkable diversity of reproductive patterns ranging from
aseasonal polyoestry to restricted annual breeding in which males synchronously die after a
brief mating season. Previous studies have categorized dasyurid reproduction into six strategies,
defined on the basis of five life-history characters. We provide an up-to-date summary of
reproductive traits in dasyurid species and examine the evolution of these characters on a
phylogeny for the family recently obtained from DNA sequence data. Our results suggest
that reproductive evolution in modern dasyurids is characterized by a basal separation of
subfamily lineages employing Strategy I1 (monoestrous females, restricted breeding season,
1 1 months to maturity; Dasyurinae) and Stratecgy\' (polyoestrous females, extended breeding
season, 8-1 1 months to maturity; Sminthopsinae).Strategies I (male die-off) and I11 (facultative
polyoestry) appear to have arisen several times from Strategy I1 or V ancestors, and Strategy
IV appears to haire arisen within Sminthopsis from a Strategy V ancestor. Strategy VI (aseasonal
breeding) has arisen independently in each of the four major dasyurid lineages (tribes), and
is highly (but not perfectly) correlated with New Guinean endemism. This scenario is not
strongly affected if rcproductive characters are optimized on an alternative phylogeny more
consistent with morphology-based opinions on species relationships. When evaluated in light
of current habitat associations and geogaphic distributions, the reproductive data suggest
that the hfiocene diversification of modern dasyurids may have been correlated with the
invasion of dry forest or woodland habitats.
0 2000 The Linnean Societ) of London
N)DITIONAL KEY M'ORDS:-Iife
history
~
reproductive biology
~
mammals.
* Corresponding author, at Southern Illinois University. E-mail: [email protected]
0024-4066/00/1
10417+ 19 $35.00/0
-117
0 2000 ' I h r 1,innean Society of London
I x c , \\~oollcy8: Braithwaite (1 9811) classified the reproductive patterns of dasyuritl
rri;iiwpials into six 'strategies' based on the states offi\.e lifehistory characters: oestrous
pattrru. nuniber of seasons per male. timing and duration of Iireeding season, aiitl age'
at stwial inaturit) j'l'ahle 1). Lee et a/. (1 982) characterized the strategies of30 dasyuricl
spc('ics for \\-hich data were available, and subsequent authors have enlarged this data
has(. t o include other estahlishcd or newly described species (e.g.Fox 8r \\Tliitford. 1982:
1)ickmaii Kr RraithLvaitc, 1992; \Iroolley~1994; Friend el al., 1997; Crowthcr. Dickmaii
& I,! i i m i , 1 !)99). Despite such extensive study, details of the reproductivc biolog?. arc
u i i k i i o n m (or at least unpublished) for many dasyurids. ,LUthough Lec et a/. ( 1992) and
I .cc' B Ck)ckbui-n( 1985)discussed thc adaptive \.slue of dasyurid reproducti\.c pattcrns.
I I O I)rt>\.iotisauthors have analyscd the evolution of these patterns from a phylogeiic~ic.
pcr\pccti\.c. In this study, LVC pro\~iclcsuch an analysis based on a rccciit h>.potlic.siso f
ph)-logcnctic. rvlationships amorig dasyurids based on DN:! scqueiiccs (Krajc\\+i.
\\'roe Sr \\'estermail, 2000).
1)asyiir~idsexhibit three patterns of oestrous cycling. In sonic spccics, such a s
nieinl)crs of .htechiniis, females are monoestrous and raise a single litter per )'car. Iii
others, such as most dunnarts (Sminthopsis), females are polyoestrous and attempt
two litters per year. A third pattern, facultative polyoestry, in which females may
attempt a second litter (if, for example, they did not mate early in the breeding
season or lost their first litter), is found in a few species (e.g. Dasycercus cristicauda).
Dasyurid males are relatively short-lived and rarely mate in more than two
successive years. In some species (e.g. members of Phascogale), male mortality occurs
synchronously within a population immediately after the mating season (‘annual’ or
‘die-off males). Lee & Cockburn (1985) described the physiological mechanism of
this pattern, noting that males experience prolonged elevation of corticosteriods
during and after their brief mating season. Hypothalamic feedback mechanisms that
lower these hormone levels in other mammals are apparently lacking in die-off
species, with the result that males eventually suffer fatal, stress-induced organ failure.
In most species, however, male survivorship is determined by proximate ecological
factors rather than physiological ones. Although Lee et al. (1982) classified as ‘annual’
species in which most males fail to survive beyond one year, but which lack the dieoff physiology, we follow Woolley & Valente (1 986) and Woolley (1990b)in describing
such species as ‘perennial’.
Breeding is seasonal in most dasyurids (the breeding season encompasses the
period during which mating occurs), but species that occur in New Guinea (e.g.
members of Murexia) and two northern Australian subspecies (Planigale maculutu
sinualis, Sminthopsis viyiniae uiTiniae) breed aseasonally throughout the year. Among
seasonal breeders, many species have a relatively brief (e.g. one month) mating
period each year (‘restricted breeding’ e.g. members of Pseudantechinus). Others (e.g.
many Sminthopsis) produce litters over several months (‘extended breeding’).
Most dasyurids with restricted breeding reach sexual maturity in approximately 1 1
months, with the notable exception of the Tasmanian devil (Sarcophilusharrisii) in which
breeding is delayed until the second year. Species with extended breeding seasons tend
to reach maturity more quickly, either by six months (c.g. many species of Sminthopsis)
or 8-1 1 months (e.g. species of Planigale). Kitchener, Cooper & Bradley (1986: 23) note
that males of Ningaui “appear to mature sexually at some time between 6 and 11
months, probably generally at about 8 months”. Age at maturity is also 8-1 1 months
for the aseasonal species that have been studied to date (e.g. Phascolosorex).
These life-history parameters define the reproductive strategies displayed by
dasyurid species (Table 1). Strategy I consists of monoestrous females and die-off
males that have restricted breeding seasons and mature at 11 months. Strategy I1
is identical to StrateLgyI except that males may survive beyond the end of the mating
season. Strategy I11 is identical to Strategy I1 except that females are facultatively
polyoestrous. In Strategy IV species, females are polyoestrous, males are perennial,
breeding seasons are extended, and maturity is attained in six months or less.
StrateLgyV is identical to strategy IV except that maturity occurs at 8-1 1 months.
Strategy VI is identical to strategy V except that breeding is aseasonal. As defined
by Lee et al. (1982)and modified here, Strategies 1-111 and IV-VI appear to represent
distinct, higher-level groups of reproductive patterns.
Adaptive value of reproductive strategies
Braithwaite & Lee (1979) and Lee et al. (1982) argued that monoestry is an
adaptive response to variable food supplies in habitats occupied by Strategy 1-111
4 21
C. KlWJEiVSKIE7:lI..
species. More specifically in habitats where the insect food of most dasyurids
undergoes strong seasonal pulses, it is adaptive for females to time their repr0ductiL.e
cycle such that the energetically demanding period of late lactation coincides with
the peak of insect availability. Lee & Cockburn (1985) speculated further that
temporal constraints on reproduction may be imposed by the fixed duration of
gestation arid pouch life in smaller dasyurids (e.g. Antechinus) in the Strategy ILIII
group, such that matching late lactation to peak food-abundance entails a very
i i a i ~ o wtinie-window for copulation and the onset of pregnancy (usually in mid- late
Lviiitcr). In contrast to female constraints, males of such species must balance the
cost of copulation in their first year against their probability of survival to breed in
tlic future. For species such as Antechinus, low overwinter survival makes high
invcstment in first-year reproduction optimal. Moroever, the brief duration of female
ocstrous requires that male mating behaviour be precisely timed. In Strate&? I
species, such investment takes the form of a brief but intense period of copulatory
iictix-ity that increases an individual male’s chance of mating, but results in death
immediately afterwards. Lee et al. (1982) emphasize that Strategy I ill only he
optimal if scasonal food pulses are predictable (i.e. stable from year to year). In
coiitrast, males of Strategy II--111 species tend to have a higher probability of survival
t o thcir second year. An optimal strategy for such males is to invest less than lOO?i,
of thcir reproductive capital in first-year mating, as manifested by lack of a die-off:
Stratqgy I11 invol\.es only a minor modification of the Strategy I1 cycle, in which
females failing to produce a litter early in the season may make a second attempt
latcr. Oh\iously, StrateLgyI11 is only feasible with perennial males.
S t r a t e s IV-VI females are polyoestrous, a condition that appears more adaptive
in habitats rvhere food is abundant throughout the year or in which there arc
multiple peaks of availability. Strategy IV and V species may employ polyoestry as
an insurancc against reproductive failure, possibly because they occur in habitats
rvith less predictable seasonal fluctuations in resource abundance (Lee et al., 1982).
Intcrcstingly, Strate<gyIV VI species display an acceleration in time-to-maturity.
\.\ith Strategy IV the acceleration is extreme (6 months or less), entails a reduction
in maternal investment per offspring (or at least a reduced duration of investment),
and allo~vsan increase in the frequency of female reproductive effort. Acceleration
dramatic in Strategy V-VI species (8-1 1 months). It is not yet clear what
proximate selection pressures might favor Strategy IV versus Strategy V, though
1,cr Pt a/. (1 982) note that Strategy IV may be more characteristic of species colonizing
cplicmeral habitats. Strate\gy VI, however, appears to be an adaptive response t o
)-ear-round aldability of food in aseasonal (often rainforest) habitats.
Eijolution of reproductive strategies
Iletnilcd theorizing on the adaptive value of reproductive strategies has not been
‘ic ( onipanicd by any substantial studies of historical patterns in the e\dution of
these traits. Lee rt a(. (1982) noted that the distribution of strategies among species
s h o ~ed
i little correlation with a phenogram of species relationships based on dental
c h L t r ~ iten
(
(,Ircher, 1976), but in fact the dental phenogram bears little similarit1 to
\ulxequent estimates of dasyurid phylogeny. Even prior to such estimates, how?\ rr.
it \ ~ n \clear that some strategies hale originated more than once over the course of
dn\)urid c\olution. The best example is Strategy I, which appears in species ol
Antechinus, Phascogale and Dasykaluta. Although initially classified as a member of
Antechinus by Ride (1964), all subsequent studies of Dasykaluta have confirmed that
it is a member of the Tribe Dasyurini (sensu Kirsch, Lapointe & Springer, 1997).
Moreover, Krajewski et al. (1 996) presented evidence that Phascogale and Antechinuy
are not sisters within Tribe Phascogalini (sensu Kirsch et al., 1997), necessitating
either a parallel origination or reversion of Strategy I within the tribe. Even more
provocative was the finding of Krajewski et al. (1996) and Armstrong, Krajewski &
Westerman (1 998) that species of Murexia form the sister-group to Antechinus, a pairing
that forces the derivation Strategies I and VI from a single common ancestor.
Further evidence of evolutionary lability comes from the discovery by Dickman &
Braithwaite ( 1992) that specific populations of Parantechinur apicalis and D a s y u m
hallucatus may in some years display male die-off, whereas previously studied
populations of both species showed perennial males typical of Strategy 11. (It is
significant that the testicular failure characteristic of other Strategy I species has not
yet been demonstrated for the die-off populations of Parantechinus and Dasyurus). Even
before such variation among clades was known, Braithwaite (1 979) noted that the
adaptive explanation ~ v e above
n
for Strategy I did not seem to apply to all habitats
occupied by Antechinus rtuartii (Jensu luto) across its range; he speculated that Strategy
I represents an irreversible adaptive ‘trap’, effectively precluding it from being
ancestral to any other reproductive pattern.
Dasyurid phylogey
Krajewski et al. (2000) provided a comprehensive phylogenetic analysis of DNA
sequences from 62 of the 68 recognized species of dasyurids, as well as the
dasyuromorphian families Myrmecobiidae and Thylacinidae. The data set for these
analyses consisted of complete sequences of two mitochondria1 (cytochrome 6, 12s
rRNA) and one nuclear (protamine P 1) locus, together comprising approximately
2.7 kilobases. The composition and arrangement of major groups on the sequence
tree is consistent with the phylogenetic classification of dasyurids proposed by Kirsch
et al. (1 997), in which sister-tribes Dasyurini and Phascogalini comprise the subfamily
Dasyurinae, and sister-tribes Sminthopsini and Planigalini comprise the subfamily
Sminthopsinae. Dasyurini includes the New Guinean genera Phascolosorex and ,Neephascogale (Archer’s [ 1984bl ‘phascolosoricines’) as the sister-group to a
Dasyums Surcophilus clade. This arrangement is well-supported by sequence data,
but is difficult to reconcile with morphology (Krajewski et al., 2000). Phascogalini
includes Murexia (placed by Archer [ 1982, 1984133 in a separate subfamily ‘Muricinae’),
resolved as sister to Antechinus. Armstrong et al. (1 998) demonstrated that Murexia
includes all endemic New Guinean species formerly assigned to Antechinus, as well
as Murexia longicaudata and Murexia rothschildi. Within Sminthopsini, Blacket et al.
(1999) supplemented the three-gene data set with partial sequences of the mitochondrial control region to resolve Antechinomys as sister to a Sminthopsis Ningaui
clade, but could not resolve the monophyly of Sminthopsis. The molecular phylogen)
represents only ‘modern’ dasyurids; the fossil subfamily Barinyainae, considered by
Wroe (1999) as sister to all the modern genera, is not included.
The DNA-sequence phylogeny is noteworthy for nodes lacking resolution as well
as those that are strongly supported. Specifically, basal radiations within Dasyurini
and Sminthopsini remain as unresolved polytomies in trees recovered from the
+
+
scqucnces. Kra-jewski et a/. (1 997) argued that the dasyurin polytomy could represent
ii \~irtuallysimultaneous origination of eight lineages in a burst of cladogenesis.
Blacker et (11. (1 999) favoured a similar explanation for sminthopsin lineages, but
tioted tlir gcncral agreement between molecular and morphological data on the
composition of those lineages. After calibrating sequence divergences against thr.
prcsumed separation of dasyurids and thylacinids some 25 million years ago (hlya).
K r a j m d i c/ 01. (2000) argued that basal divergences within each of the four dasyurid
trilic‘s occurred some 1 1-16 Mya in thr mid-Miocene. This conclusion is significant
fbr understanding the evolution of reproductive strategies in that continent-wide
c.limatic clkcts on late Tertiary habitats were experienced by all lineages of modern
das).urids.
C:omplete or partial information on the lifc-histoq Lwiables tabulated by Lee
P/
d.11982j arc now available for 42 dasyurid species arid subspecies, as Lvell as the
ihl>.rmecohiidac)and recently extinct thylacine (Thylacinidae). ‘I’he DN;\
ph~~logciiy
for these taxa is shown in Figure 1, and the data matrix of their
rr.Iiroductive characters is given in Table 2. In constructing the phylogeny, we
rctaiiied polytomies from the maximumparsimony tree of Krajewski e/ 01. (2000).
t)ut iiicludcd all bifurcating nodes for which bootstrap support was 2 70?0.I’ve also
rt.taiiied a monophyletic Sminthopszs, though this node was not resohwi by D5.A
sequence data. The potential effects of topologcal inaccuracies are discussed beloiz .
Rcpi-oductive characters were optimized under the parsimony criterion on the
rnolccular trce using the MacClade 3.05 software package (Maddison 8r hladdison.
1992)> subjcct to a number of assumptions. All five characters were treated as
inclcpendcnt, unordered, and unpolarized. Although some pairs of characters arc
ncithcr logically nor hiologicallj- independent (c.g. species with ‘restricted breeding’
must also be ‘seasonal’), in no case did recoiistructioiis yield ancestors with coinIiiitations of traits not found in living species. Only ‘oestrous pattern’ and ‘age-atmaturity‘ display multiple (3 4) states, and in neither case was there prior cvidencc
of’constraints preventing direct e\.olutionary transitions among them. .An exception
to the latter point is Braithwaite’s (1979) idea that male die-off‘ is irrc\usihlc, a
possitiilit~.that we eiduate in light of our results.
Optimizations iwrc performed with and without the outgroup-states information
piwicied ti?- Tliylminus and 3Qrmecohiu.r. The modern ‘i?yLucinus ynocephalus M
rcIati\.cly derive-d member of a historically diverse family (Muirhead & M’roe, 1998).
and thus may not accurately reflect the ancestral thylacinid repr0ductk.c pattern.
.[lie li\?iug numbat constitutes a monotypic family with no pre-Pleistocene fossil
ord (I\rchcr, 1984b), making inferences about ancestral myrmecobiid rcproduction
sptwilati\~t.at best. ’The influence of controversial topological arrangements on thc
molrcular tree was investigated by reoptimizing reproductive characters on a
ph>hgeny more consistent with current opinions based on morphological evoluton:
I h w ” o , \ o r e . ~ +.Yeo/,hascogale as the sister-group to other dasyurins; (2) Archer‘s
1 ! N 2 ; ‘Parantechini’ genera (Pumntechinus, Pseudnntechinus, and Da.$alutu) as monopliylt,tic aiid forming the sister-group to other dasyurins except ‘phascolosoricines‘:
1 3 I Ihi.!rcercm.c and Da.yuroidc,r as sisters; (4) ilntechinus (sensu stricto) and Phascogal~ as
sisters: ( 5 ) ~-lntechinomqls,rather than . \?n<qaui, as sister to Sminthopsis.
iiuiiiliat
M. melas VI
Ps. ningbingI I
Dasykalufa I
Pa. apicalis II
DasyuroidesV
Ps. macdonnellensisI1
M,P.R,S,ll
c
M.?.R.S,11
-
.
D
,
M,P.R,S,ll
M,P,R.S,I1
6
M,?,R.S,lI
P.A,E,s.~
F
P,A~~E
;,11
A
PA.E,S,~
P.A,E.S.8
MureMa V I
Phascogale I
PI. ingmmiV
PI. gilesi V
PI. tenuimstrisV
PI. m. maarlata V
PI. m. sinualisv
P,A,E.S,G
PAES.8
P,A,E,S.B
S. griseoventer II
S. longicaudataV
s. O0ldea
S. youngsoni ?
1
AntechinomysV
Thylednus ?
MynneoobiusII
Figure 1. Ancestral states of reproductive characters inferred by parsimony optimization on the
phylogcny of dasyurid species from Krajewski et al. (2000). Reproductive strategirs are shown as
Roman numerals after species or genus names (see Table 1 for strate'gy definitions). Keconstructed
traits are shown on each branch in the following order: oestrous pattern (PI =monoestrous. P =
polyoestrous, B =facultative polyoestrous); number of seasons pcr male (A = annual, P =perennial);
duration of breeding (E = extended, R = restrictrd); seasonality of brccdiiig (S =seasonal, '4 = aseasonal):
age at sexual maturity (6 = 6 months, 8 = 8-1 1 months, 1 1 = 1 1 months). Qucstion marks (?) iIidic,ate
equivocal states. Node labels identify suprageneric groupings: A = Dasyuridae, B = Dasyurinae, C =
Sminthopsinae, D = Dasyurini, E = Phascogalini, F = Planigalini. G = Sminthopsini. Only taxa for
which some reproductivc data arc availablr are included. Only genus names are given for monotypic
genera and genera in which all specirs have the same reproductive s t r a t r s . Genus abbrcviations:
D.=Dasyurus, M. =hQoictis, h:=,.Lingaui, Pa. = Parantpchinus, Ph. = Pharrolosowx, A. = Pseudantechinus, PI. =
Plankale, S.= Sminthopsis.
Davcemu
My. m h
My. wallmei
Pa. apicalis
Pa. bilami
Ps. niacdonnelkmis
Ps. niigbing
Ps. woollque
Ps. m i m n b
Dasykaluta
Phascog. tapoatafa
Pharcog. calura
A. agilis
'4. b e l h
A. jauipes
A . godmani
A. lea
A. minimu
A. stuartii
A . swainsonii
hlu. habbema
Mu. longicawiata
Mu. m l a n u m
Mu. naso
Mu. rothchildi
VI
VI
VI
VI
VI
I
I
I
1
I
I
I
I
I
I
I
?
I1
I1
I1
I1
I1
?
VI
111
V
?
?
111
I1
VI
?
I1
111
111
VI
D.hallucah
D.uiumnus
D. maculntur
D. albopunctatlu
D. spartacw
D. geofii
Sarcophilus
Phascol. dorsalis
Phascol. dorim
fleophascogale
Dasyuroides
Strategy
Species
mono
mono
mono
mono
mono
?
mono
mono
mono
mono
mono
mono
mono
mono
mono
mono
mono
POlY
POlY
POlY
POlY
POlY
Y'Y
POlY
POlY
fac. poly
Y'Y
mono
fac. poly
fac. poly
?
POIY
fac. poly
mono
Oestrous
Srasons
per male
?
extended
restricted
extended
?
restricted
restricted
restricted
restricted
restricted
?
restricted
restricted
restricted
restricted
restricted
restricted
restricted
restricted
restricted
restricted
restricted
extended
extended
extended
extended
extended
3
restricted
restricted
restricted
extended
?
restricted
restricted
extended
Duration
of breeding
seasonal
seasonal
seasonal
aseasonal
?
seasonal
seasonal
aseasonal
?
?
seasonal
seasonal
aseasonal
?
seasonal
seasonal
seasonal
seasonal
seasonal
?
seasonal
seasonal
seasonal
seasonal
seasonal
seasonal
seasonal
seasonal
seasonal
seasonal
seasonal
aseasonal
aseasonal
aseasonal
aseasonal
aseasonal
Seasonality
TABLE
2. Dasyuromorphian reproductive d a t a matrix
?
8-1 1
8-1 1
8-1 I
II
8-1 1
11
48
8-1 1
?
?
8-1 1
11
8-1 1
?
11
11
11
11
11
?
11
11
11
11
11
11
11
11
11
11
>
11
11
11
?
Months to
maturity
References*
.P
N
.P
>
11
J
W)
IV
CVI)
?
V
V
(111)
11
>
J
>
IV
rv
?
?
?
?
V
V
V
V
IV
IV
IV
>
v1
V
V
V
\?
VI
>1
rxtcnded
rxtrnded
J
mono
>I
>I
?
?
yb
mnno
>I
2
7
>1
>I
POb
POlY
Y1Y
POlY
>1
?
3
fac. p d y
>I
?
>I
>I
>I
?
?
exteiided
extendcd
extended
PxLrnded
cxtcnded
extended
extended
?
restricted
restrictd
J
extendrd
extendrd
extended
?
restricted
3
i
3
>I
>I
>I
>I
21
>
extended
extrndcd
extended
extended
cxtendcd
extended
extended
cxtendcd
extended
extended
rxtrndcd
extended
extended
>I
>I
3
>I
>I
>I
>I
>I
mono
YIY
POIY
POlY
?
?
>
POIY
YIY
POI?
POlY
PolY
poly
POlY
>
POlY
Ply
P(jlY
poly
POlY
POlY
aseasonal
orial
seasonal
aseasonal
scasoiial
awasnnal
seasonal
7
seasonal
seasonal
seasonal
seasonal
?
3
?
?
?
seasonal
seasonal
seasonal
seasonal
>
srasonal
wasonal
seasonal
seasonal
wasonal
seasonal
seasonal
seasonal
?
aseasonal
seasonal
seasonal
6
6
8 11
6
6
6
8 I1
8 I1
8 11
8-1 I
8-1 I
8-1 I
8-1 I
6
6
8-1 1
>
11
>
J
?
?
I1
1
I1
1
?
?
?
?
>
3
>
>
3839
40,41
42
3
29,303 I
32
33
34
35,36
36,37
28
6,16,26
6,27
25
24
3
1
20,2 1,43
1,23
l.l6,22,23
1,18
43
I Y,43
1
1
1,16
1,17
1,17
3
3,6
* Referrncrs: 1 = Lee P / a/. ( 1 982): 2 = Dickman & Rraithwaitr ( I 992); 3 = Woolley ( 1 991); 4 = Soderquist 8r Serciia (1990); 5 =]onrs in Strahan (1 995: 84); 6= I’i\ LVoolley
(unpublishrd data); 7 = (;ihsini & C:ok ( I 992); 8 = LI’oolley 1199 I c); 9 =Woolley ( 1 995); 10 =Woolley ( I 99 I b); 1 I = Woolley (1988): 12 = \Voollcy (1 991a); I3 = Rradlry in Strahan
(1995: 103); 14=l)ickman in Strahat~(1995: 101); 15=Uwycr (1077); 16=PA \Voollry & C . Elliott (unpublished data); 17=Kead (I984); I8=M’oolley (1984); l0=I)unlop ill
Strahan ( I 995: 1 18); 20 = Bos & Carthew (1 999); 2 I = Kitchrner in Strahan (1995: 120); 22 =Woolley ( I 99Oa); 23 = Woolley ( 1 9901,); 24 =Woinarski & Van Dyck in Strahan (199.5:
127); 25 = Friend rl al. ( I 997); 26 = Woolley in S
an ( 1 9%: 135); 27 =Morris & hIrKenzir in Strahan ( I 995: 137); 28 = Crowthcr Pt ul. ( I 999); 29 =Read ut al. ( 1 983); 30 = \ V o o l l q
& Ahern (1983);3 I =L\Ti)oky 8r C;ilfilliln ( I 990)
= W o n l l q 8r Va1t:iite (1986)
=r\slin (1!)8?);34=Pearsorl in Strahan (199.5: 15.5);35=7hplin (1980);36=\Voollcy in Str;ih;in
( 1 995: 156); 37 = hlorton, i\mistrong & Rraithwaite ( I 987); 38 = McKenzie & Cole in Strahan ( I 995: 158); 39 = Dickman et nl. ( I 999); 40 =Smith ( I 982); 41 = Dixon (1989): 42 =
Friend in Strahan ( I 9%: i 6 1); 43 = Kirchener u/ a/. ( I 986).
.2~nnecobii~s
Thylacinus
S. hzrtipe.r
S. leuropuc
S. lon~qcaudata
,S.ooldea
S. fiuimmophila
S. a ciigznzue
S. u. nitela
S. 11. ru&enir
S.y o n n p n i
S.d o u p h i
S gilberti
S.granulipes
S.griseowntm
S crassirnudata
S. marroura
S. inurinu
S. aithni
S. archeri
.C bindi
S. butleri
S.dolichui-a
.My o n n e a e
Mu. uilhelmina
A. m. maculata
A. m. sinualir
P1. ingrami
PI. gilesi
PI. tenuimrtrir
PI. nouaeguineae
Antechinomys
,h<d e i
,V timealcyi
5
tr:
LJ
z
s7
z>
5
2
r
s
2
‘A
3
h
li
rn
rr
i
2
E
.c
2
2
3
23
&Z
?=
=;
n
z
n
2
r
Because Braithrvaite & Lcc (1979) and Lee e t a / . (1982) attached ni?jor significaiicc
t o t tic influence of environmental characteristics on the evolution of reproducti\.c
\tr;ttegies, \ve tabulated the hasic habitat associations and climatic zoiics for e ~
das)-urid species (Table 3). Habitat descriptions were obtained froni E’larinery ( 1090’1
and Strahaii i 1995)and placed into one or more categories (rainforest, rvct scleroph?dl
tiircsi. dry sclerophyll forest, woodland, shrubland, heathland, mallee, and grassland)
l ~ a s c dO I I the coiidcnscd version of Specht’s (198 1) habitat classification pro\.ided ti!.
.-!rc,ticr & Fox (19841, to rvhich we added the catcgory ‘desert’. Each specks [ \ - a s
A()assigiicd t o one or more of the hiogeographic regions (rainforest, north niotisooii.
iiontropical eastern, Tasmania, southryest, Eremaean) described by Johnson 8r Briggs
1975) and prescntcd in Archcr 8r Fox (1984). to which lve added ’Neri. Guirira‘.
\Vc sclcctcd this classification of arcas because it reflects general zoogcographic
tli\ isions. rainfall patterns, and annual tcmperaturc regimes (C‘l‘hite, 1 W4\. Arra
itssociations 1 w - e optimized as unordered states on the molecular phylogeiiy. I n i t in
this c’ase anccstral rcconstructions \\’ere not interpreted as featurcs of anccstral
ypwics. Ikithcr. they suggest potential trends in the evolution of cii\ironmrntal
a l T i i i i h iniplicd by dasyurid phylogeny.
Ancestral itafec on ihc molmlar trce
\\’hcn character statcs of ,\yrmecohius and 7hJvlarinus are included. all reproductivv
clxiractc-rs in the ancestor of modcrii Dasyuridae are unambiguously reconstructed
Fiyi 1 . 2). ‘l’liis ancestor was a morioestrous species rvith perennial nialcs. seasonall!rrstrictcd mating, and an age-at-maturity of 1 1 months (Strate,gy 11). ?‘hc sanic
Ytaics art- fouiid in the ancestors of Dasyuririae a i d Dasyurini, and all hut on(’ ol‘
r l i c m arc uiiarnbiguouslv assigned to the Phascogaliiii ancestor as ~vcll(annual and
i)crcnnial rnalcs are equally parsimonious for tlie latter), consistcnt Lvitli Strateqirs
I o r 11. Siniilarl>-.the ancestors of Sininthopsinae, Sminthopsini. and Planigalini arc
iiniimhiguously rccoristructed as Straw9 V. Under this scenario, the basal separation
of das).urine and sminthopsinc lineages corresponds to the derivation of pol>-ocsrr)..
; i n erteridcd hreediiig season, arid accelerated delrelopmcnt in sminthopsiiics, rvhilc
da~).urincsretained fcatures of the common ancestor.
‘l’he cffcct of excluding outgroup taxa is to reduce the number of stater; that can
I)(, unanihiguously reconstructed (results riot slioum). The only characters that c;w
d fix the dasyurid ancestor are perennial males and seasonal breeding?;.
t nionoestry and polyoestry, as \\.ell as 1 1 or 8 -1 1 months to maturit!-.
itre cqually parsimonious at this node, tlie reconstructed ancestor is consistent \\-it11
ctitlicr Strateg? I1 or V. Dasyurinc and dasyurin ancestors retain these statcs and
also show restricted tlrccding and 1 1 months to maturity (Strate,? TI); tlit. num1x.r
tsoiis per male remains ambiguous in the phascogaliri ancestor (Stratcgir’s I
( ) r I1 1. Sminthopsine, sminthopsin, and planigaliii ancestors, horvc\.cr, are again
i.cc.on~ti.uc.tct1
unambiguously as Strategy V. The sceiiario for tiasal divcrgcncr in
t h i \ anal>.sis is siniilar to that obtained with outgroups, except that it is unclear
\\4icthcr the dasyurid ancestor resenibled a dasyurine or a sminthopsinc. Under
cithcr scenario? tribal separatioiis rvithin subfamilies do riot seem t o have riitailcd
i i i w l i rcproducti\.r e\dution.
h
TULE3. Ilasvurid habitats and biogrographic rcgions. Habitats and distributions from Flanncry ( I 990;
and Strahan (199.5); catcgories modified from Archer & Fox ( 1 984)
Species
Habitats
Biogeographic regions*
Dry sclerophyll. woodland, inallrr, desert
Dr). sclerophyll. woodland, grasvland
\Vet & di? sclrrophyll~raiiil'orest. woodland. heathland
SM', Errmacan
\Vet & d n sclerophyll. shrubland, heathland
Rainforest
Grasslaiid
llry sclerophyll, Lzoodland
Raiiili )rest
Rainforrat
Kainforrar
Dcscrt
Drscrt
Grasslmd
Rainforrst
Heathland
Ilr). sclerophyll
Desert
\Soodland
\Voodland
M'oodland
Dry sclerophyll, shrubland
Dry sclerophyll, woodland
Dy sclcrophyll, woodland
Rainforest, wet & dry sclerophyll. woodland
Rainforest
Rainforest
Htathland, grassland
Rainforest, wrt 8r d n sclrrophyll
Rainforrst, wet sclcrophyll, hcathland
I V e t & d r ) sclerophyll
RainforcsL
Rainforest
Rainforest
Rainforr\t
Rainforest
Rainforrst
Grasslands. woodlands
Grasslands, woodlands
Grasslands. rvoodlands, rnallee
IVoodland. grassland
Rainforrst, N e t sclcropliyll, grassland
Desert, gmssland. shrubland
Grassland
Grassland, shrubland, mallcc
<;rassland, mallce, hrathland
Mallrc
LVoodland
ISoodland
\2'oodland
IVoodland. qhruhland, grassland. drscrt
IVoodlaiid, chruhlarid. hcathland, grdcsland, malllee
Grassland
Hcathland. m a k e , woodland
Shrubland, m a k e
M'oodland. hrathland
Eremaean, moiisonn
NTB, Tas, moricooti,
raincorest
W E , -fits
N(;
pic
'l'as
Nc:
KG
NG
Ercmaean
Erernaean
Eremaean
NG
SIV
RIonsoon
Eremaran
Mollsoon
Rlurisoon
Ereinaraii
Ercmaean
S\V; NTE, inonsomi
Slonsoon
Rainforest. S\,V, N'I'E
Rainforrst
Rainforest
NTE
NTE
NTE, ?'as, rainforest
NTE
Die;
n.G
Nc;
N(;
IVc;
NG
Err maean
hlonsoon
Ercmaean
NC;
Rainforest, monsoon;
NTE
Ervrnaean
Ercmaean
Eremaran
Erc ni iw an
NTE
hlonsoon. KG
Monsoon
hfonsoon
SLV; Eremaenn. NTE
Err mnean
.\lonsoo11
SM', Eremaean
SM'
SLS
ionhnurd
C:. KIUJE\\'SKI 87,U..
+?:I
I:XRI.E 3. Ilasyurid habitats and biogeographic rcgions. Habitats and distributions from Flaiinrr) ( 1 990
nnd Strahan (1995); categories modified from Archer 8r Fox (1981) -rontinirrd
\Voodlaiid, shrubland, hcatliland, grassland
\Voodlaiid. dr) sclcrophyll, heathland. grassland.
raiiifurmt
Grassland
Shruliland, \voodlaiid, grassland
\\oodland, di? scl(vipli) 11, ticathlaiid
\\oodland, mallre, grassland
Grassland. scrutiland, \wodl;tnd
\\oodlancl. g ~ i t ~ ~ l i i i i d
Desert
Dry ~ l e r o p h y l &
l ~oodland
Dr) sclrrophyll & wuodland
* S ( ;= Sc\\
(;iiinra. N'IX = non-tropical east. Sly = southwest. Tas = Tasmania.
Dasyurini
Phascogalini
Planigalini
Sminthopsini
Imono
perennial
restricted
seasonal
restricted
seasonal
perennial
extended
seasonal
perennial
extended
seasonal
II
11 months
8-11 rnos.
perennial
seasonal
2 . Summar)- of rcconstructcd statcs (in boxes) and reproductive stratrgirs (iicxt t o hosw oi
tribnl. rubfamilial, and modern familial ancestors (filled circles) 011 thc molrcular phylogeny of
I-iyurc 1. Reconstructions include outgroup information. Reproductive traits are sho\i 11 in rach b o x
iii thc folio\\ ing ordrr: oestrous pattrrn (moiio = monoestrous, poly = polyoetr(lusJ, iiumtxx of srasori<
pcr inalc. scasonalit); of Ix-ecding, agr at srxual maturity. Question inarks (I?) indicatr equivocal s t a t n .
1;iyur.c
cl,i5> rid
Regardless of outgroup effects, reconstructed ancestors within trihes suggest that
traits associated kvith Strategies 111, IV, and V1 appeared relatively late in dasyurid
r\.olution. and that they arose repeatedly from Strategy I1 or S t r a w 9 1' ancrstors.
Strate<q I11 (facultative polyoestry) arose twice within Dasyurini and once Lvithin
Sminthopsini. Strate<gyIV originated within Sminthopsis, where maturation time is
rcduced from & l l to 6 months. Stratqgy VI (aseasonal tireeding) appears iiidependenti)- at least six times and in all four tribes on the molecular tree, espcciall).
iii lineages endemic to New Guinea. These strategies thus seem to reprrseiit adapti\.c
'fine tuning' of more general, ancestral patterns.
Interpreting the distribution of Strate<gyI is more problematic. &[ale die-off is not
unambiguously reconstructed as the ancestral state for any sister-pair of dasyurid
11,
v
Ficgure 3. Summary of rcconstructcd reproductive stratqies of major dasyurid anccstors 0x1 a morr
traditional phylogeny (sce text for description). Reconstructions include outgroup information. Inferrcd
ancestral strategies are based n rtconstructions of indi\.idual reproductivr characters as in Figs 1 2.
~
genera. This is consitent with Braithwaite’s (1979) suggestion that Strate<gyI is
irreversible, and with the argument of Krajewski et al. (1996) that the ancestral
phascogalin did not employ Strategy I. O n the other hand, the reconstructions are
inconsistent with the latter authors’ contention that the phascogalin ancestor employed Strategy VI. Rather, Phascogalini appears to present instances of extreme
parallelism in reproductivc evolution. If the ancestor w7as a Strategy I1 species, then
male die-off arose independently in Phascogale and Antechinus (as well as Da?ykaluta).
More spectacularly, however, is the derivation of Strategy VI in Murexza via changes
in virtually all of the reproductive traits considered here. Under this scenario, no
modern phascogalin possesses the reproductive strategy of its tribal ancestor; rather,
these closely related genera represent completely opposite specializations in the
repertoire of dasyurid reproductive behaviour.
Ancestral states on a morphological tree
Parsimony optimizations on the morphology tree (Fig. 3) again show sminthopsine,
sminthopsin, and planigalin ancestors with Strategy V, and Strategy IV derived
within Sminthopsis. The effect of morphologcal groupings on dasyurine ancestors,
however, is primarily to create equivocal reconstructions, with all b u t one of thc
major ancestors showing only perennial males and seasonal breeding (Strategies I1
or V; states entailing strategies I11 or IV are less parsimonious). Indeed, the only
node reconstructed more clearly for the traditional groupings is the ancestor of
‘Phascogalinae’ which, not surprisingly, displays Strategy I. This reduces but does
not eliminate the extent of parallelism in the evolution of male die-off (it also arose
in Dasykaluta, a dasyurin). In contrast, the number of independent origins of Strateqies
VI is unaffected by the alternative phylogeny.
Although the topologcal differences between Figures 1-2 and 3 do not necessarily
cntail strong conflict over the states of tribal or subfamilial ancestors, oiie rcarrangcnicnt of the tiranchs in Figure 1 \vould havc a major impact on these reconstructions.
Sprcificall\,, relocation of A<yoicti.s to a position between ‘phascolosoricines‘and otlicr
dasyurins on Figure 3 would place two Strate<q VI cladcs as sequential branchcs
fi-om a coninion ancestor. As a consequence, the dasyurid anccstor is reconstructed
as Strateq. \’, and states of the dasyurine, dasyurin, and phascogaliri ancestors arc
miisistent ~vithStrategies V or VI. This scenario would imply that the reproducti\.c
clifti.rcnces hettvccn smiiithopsincs and dasyurincs resulted from latcr rvolution
\ \ i thin tribcs rather than from the di\w-gence of subfamily ancestors as suggrstcd
I)! tht, molecular tree. hloreovcr, an ancestral Strategy VI would suggest that this
i i ( n\ predoniinantly Ncw Guinean reproductive pattern was also found in hIiocenc
c1as)urincs. Such a basal position for ;Ilyuictis and ‘phascolosoricines‘is of considerahlc
itilcrcst in that it is consistent with recent morphocladistic analyscs of dasyurins is.
\\-ror. pcrx. c(iinni.), and because the molecular placcment of ‘phascolosoriciiies‘
I-ctnains prokinatic for morphologists. This example illustrates the importance of
niorc’ thorough resolution of phvlogcnetic rclationsliips within Dasyuridac.
Habitat,s a i d reproductirv strategies
Habitat afhiities of dasyurid species included in this study are shotvn in Table 3 .
.,Istriking feature of this distrihution is that, with two exceptions (Pi.rn. mar.ufata and
,S, /~umpus),all dasyurids that occur in rainforests or wet sclerophJ-I1forests lielong
to Ilasyurinae. For that matter, only three sminthopsines (S. feucopus, 5‘. nzurim. arid
P/. 111. macukatn) occur in any type of forest habitat, in contrast to the many forestd\velling dasyurines. It thus appears that the two basal cladcs of modern dasyurids
:w characterized by different habitat associations, a distinction which ma)- also have
applied to their Miocene ancestors. Krajewski et al. (2000) noted that Early Llioccnc
rlirnatcs in Australia were relatively warmer and wetter than thoso of the 1,atc
Oligoccnc, resulting in an expansion of \vet forest habitats from Oligocene refugia.
Nevertheless, areas of drier woodland (that were more extensive in the Oligocenc.1
persisted into the Early hliocenc. Dasyurine and sminthopsine ancestors may h2n.c
tiif‘crrd in their affinities for these habitat types. with the former occupying forest?
atid the latter occupying woodlands.
If Krajchj.ski el al. (2000) are correct in suggesting that thc ancestor of modern
tlas!w-ids arosc in the latest Oligocene, this species would hare existed at a time
\\-hen aseasonal: wet-forests habitats bvere contracting and being replaced by morc
scasoiial, dr\;-forest assemblages (14Iiite, 1994).Like modern numbats and thylacincs,
this anccstor may have occurred primari1)- in dry-sclerophyll or woodland habitats.
In contrast. the fossil Raricya which M’roe (1999) described as sister to motlcrn
das)-urids, \$‘as recovered from an Early Miocene rainforest fiuna. ‘l’hylacinids\vcrc
also common in Oligocene~~h/liocerie
rainforest environments (hluirhcad 8r \2’roc,
1998). Inderd, M’roc (1996) and Krajewski et al. (2000) noted that the progressi\-c
dt.clinc in thylacinid (and bandicoot) diversity throughout the hriocene \vas paralleled
1)). tlir risc and diversification of modern dasyurids. It is tempting to speculatc t h a t
niorlcrn das) urids ma); have gaincd an early evolutionary foothold \vhcn their
ancestor i n \ d e d dry-forest or woodland habitats of the Late Oligocene. Rcexpansion of wet-forcst environments in the Early Miocene may then have pro\.idcd
separate \niucs for the isolation of dasyurine and sminthopsine lineages.
E\’OLUl‘ION OF REPROUUC’I‘IYE STRA‘TEGIES IN DASYURID hlARSUPIALS
131
Preference of ancestral dasyurines and sminthopsinses for forest and nonforest
habitats, respectively, would be correlated with the adoption of different reproductive
strategies. Ancestral sminthopsines that lived in woodlands or other dry, nonforest
habitats would conform to the adaptive scenario for Strategy V described by Lee et
al. (1982: 9), in which polyoestry in particular is expected to occur “where resources
are seasonal but may be temporarily in short supply as in temperate grasslands and
habitats in arid Australia”. In contrast, forest-dwelling ancestral dasyurines would
have occupied habitats “where resources are predictable and highly seasonal, but
also where adult survival is enhanced” (Lee et al., 1982: 8), a scenario consistent
with their use of Strategy 11.
There is less correlation between general habitats and reproductive strategies
below the subfamily level. Strategy IV originated within Sminthopsis, but dunnart
species occur in a wide variety of nonforest habitats, as do other members of
Sminthopsini and Planigalini. Similarly diverse habitats characterize dasyurines,
especially within the phascogalin genus Antechinus (all species of which employ
Strategy I). Interestingly, Dasyurini may be partitioned into two groups with different
habitat affinities: ( I ) a crown group composed of forest-dwelling Dayurus, Sarcophilus,
‘phascolosoricine’, and perhaps MToictis species; and (2) a basal set of arid-adapted
lineages (Dayuroides, Daycercus, Parantechinus, Pseudantechinus and Dusykaluta). These
groups, however, have reproductive strategy I1 in common, but strategies I, 111, and
VI occur in various individual lineages of arid- and forest-adapted species. It thus
appears that ‘fine-tuned’ strateies I, 111,IV, and VI arose in response to environmental
factors other than general habitat type.
Biogeography and reproductive strategies
Optimization of biogeographic regions on the molecular phylogeny (not shown)
yielded only one unambiguously reconstructed state at nodes deeper than terminal
sister-pairs: dasyurin, sminthopsin, planigalin, dasyurine, sminthopsine, and dasyurid
ancestors all display the ‘Eremaean’ state. Again, this reconstruction cannot indicate
that Oligo-Miocene dasyurids were found in arid areas of central Australia, because
such habitats did not come into existence until later in the Tertiary (Archer, 1984a).
What it does reveal, however, is that basal lineages in each of the four tribes are
comprised partly (e.g. Phascogale cahra in Phascogalini) or wholly (e.g. Antechinomys
and Ningauz in Sminthopsini) of dry-country taxa. In conjunction with the habitat
associations described above, this finding reinforces the idea that the rise of modern
dasyurids was related to their invasion of drier habitatdareas in the Oligo-Miocene.
It also suggests that extant, wet-forest inhabitants (e.g. New Guinean species) are
not ecologically primitive (Archer, 1984a), but rather are secondarily specialized to
habitats frequented by their extinct relatives (i.e. Barinya and thylacinids).
Unfortunately, the biogeographic reconstruction provides limited insights on
ancestral reproductive strategies. Temperature/rainfall associations of modern dasyurids show no strong phylogenetic partitioning among major lineages. Although
there are no Tasmanian or southwestern planigales, and only one Eremaean
phascogalin, all other biogeographic associations are found in each of the four
tribes. Lack of a close relationship between dasyurid phylogeny and traditional
zoogeographic regions is itself strikng, but even if we ignore phylogenetic relationships
there appears to be little correlation between biogeography and reproduction. Table
I
\
\I
1
,
ti
I
1
il
i)
4 i\ ii taI)ulation of the numbers of dasyurid species ivith knotvn reproducti\-c
yti.;it(:gich that occur in each biogeographic region. Ferv cells in this matrix arc
( L n i p t > ' : Strates. 1 species occur in all six Australiaii regions, and Stratcgics IT 1'
o c ( ~ i rin h u r o r five of the six.
notable exception to this pattern is the stroiig
insociation I)ct\vecn Strate% VI and New Guinea. Howcver, the 12 species t h a t
siioit. this association include rcprcsentatiws from all four tribcs, and iiidicatc no
niajor ph>-logcneticeffect. Like Stratqq I, Straten VI provides a dramatic instance
of'piirallcl reproducti\.c cvolution ainong dasyurids. It is kvorth noting again, lionww.
tliat interpretation of Strate,gy VI as a recent specialization is contingent t o sonic
('xtciit o i l tlic molecular placement of 'phascolosoricines' arid ,\I~JO~C~~.S
a s dcri\.td
liiic,agcs I\ itliin Dasyurini, a point that is contested by morphologists and rccluii-cs
l i i rtlier tehtiiig with molecular data.
'l'lic fbregoing scenario of rcpr0ductk.e eidution in dasyurids must be qualified
sc\.eral important respects. First, reconstruction of ancestral life history traits ivitli
p.ir~imoii!. ciitails substanti\Te assumptions about the evolutionary proccss. Chiel
ainong thcsc is that character e\.olution itself has not twcn wildly uriparsimonio~is.
;I claim for ivhich \vc have no evidence. Unfortunately, it is not yet compL~tation~tll>
fi~asildcto implcmen t maximum likelihood mcxthods for ancestral state reconstruction
o i l a data set of this size, though such an approach will be important for qu;intif>.ing
t h e influcncc of lx-ancli lengths on character evolution (Pagcl, 1999). Second. tvc
awirricd that species for which reproductive data arc unavailable (c.g. many
, ~ / ~ ~ ~ ~ ~ivill
~ / ~not
~ ~ ha\.e
/ ~ , ~niajor
~ , s ) cffkcts on reconstructions. Third, we assumed that
;iiicvstors can be characterized strictly in terms of charactcr-states disp1a)ecl 1))
rxtaiit species, and again there is no way to substantiate such a claim. Fourth. \I-c
;rbsunicd that outgroup information from recent T/~ylucinusand ,l!l~nii~robiztsspccic.\
dc K Y not distort reconstructions mithin Dasyuridae, though wc have no data on thy
i . ( , i ) i - o [ ~ L i ~ . t i \ . ~heha\+xr of extinct thylacinids or myrmecobiids. Finall>->
\I-cassumed
that the niolecular phylogeny is correct. \Ye attempted to evaluate the rohustness
ol' our r t w l t s t o the latter t\vo assumptions, but the fact remains that inferred
p,irtcrns of' character e\.olution are contingent on a specific phylogenctic h>-po~hcsis
t l i a t includes only modern tam. Despite the limitations these assumptions place on ( ) u r
iii
E\’OLU‘TION OF REPRODUCTIVE S’I’R4lEGIES IN DASYURID hlARSUPIALS
131
conclusions, those conclusions (however tentative) represent the first comprehensive
hypothesis of reproductive evolution in Dasyuridae.
Perhaps the most significant finding to emerge from our study is that reproductive
evolution in dasyurids is characterized by both phylogenetic constraint and lability.
The well-supported cladistic grouping of modern dasyurids into the subfamilies and
tribes of Figure 1 confirms, for example, that extreme reproductive strategies I (male
die-off) and VI (aseasonal breeding) have indeed arisen more than once within the
family. O n the other hand, it seems that basal dasyurine and sminthopsine clades
are characterized by the more general strategies I1 and V, respectively, and
that these strategies may represent ancestral adaptations to forest and nonforest
environments of the Early Miocene. The diversity of strategies exhibited by extant
dasyurids, however, appears to have arisen during the great tribal radiations of the
Middle Miocene. Within these radiations, the influence of general habitat and
climatic associations on dasyurid reproduction does not appear to have been strong
(except for the correlation between New Guinean environments and Strategy VI).
Lack of published data on the seasonal availability of dasyurid food prevents us
from evaluating the contention of Lee et al. (1982) and Lee & Cockburn (1985) that
this factor is a major determinant of reproductive patterns. Further elucidation of
the evolutionary forces influencing dasyurid reproduction will require a more fully
resolved phylogeny and more detailed information on habitat and life-history
variables for all species.
ACKNOWLED(>F,hlENTS
We thank M.J. Blacket and M. G. Fain for helpful discussions on this manuscript,
and C. R. Dickman for access to unpublished information on Sminthopsis youngsoni.
This research was supported by funds from the U.S. National Science Foundation
(grant DEB-941 9907 to C.K.), the Australian Research Council, and La Trobe
University.
REFERENCES
Archer M. 1976. The dasyurid dentition and its relationship to that of didelphids, thylacinids,
borhyaenids (Marsupicarnivora) and peramelids (Peramelina: Marsupialia). Australian Journal of
<oology Supplemental Series 39: 1-31.
Archer M. 1982. Review of the dasyurid (Marsupialia) fossil record, integration of data bearing on
phylogenetic interpretation and suprageneric classification. In: Archer M, ed. Carnivorous hfarsupia1.r.
hlossman: Royal Zoological Society of New South Wales, 397-443.
Archer M. 1984a. Evolution of arid Australia and its consequences for vertebrates. In: Archer M,
Clayton G, eds. Vertebrate zoogeography and Evolution in Australasia. Carlisle: Hesperian Press, 97- 108.
Archer M. 1984b. The Australian marsupial radiation. In: Archer M, Clayton G, eds. Vertebrate
<oogeography and Evolution in AustralaJia. Carlisle: Hesperian Press, 633-708.
Archer M, Fox B. 1984. Background to vertebrate zoogeography in Australia. In: Archer M, Clayton
G, eds. Vertebrate zoogeography and Ecolution in Australasia. Carlisle: Hesperian Press, 1-16.
Armstrong LA, Krajewski C, Westerman M. 1998. Phylogeny of the dasyurid marsupial genus
Antechinus based on cytochromr-b, 12S-rRNA, and protamine-P1 genes. Journal of MammaloQ 79:
1379-1 389.
Aslin HJ. 1983. Reproduction in Sminthopsis ooldea (Marsupialia: Dasyuridae). Australian Mammalogy 6:
93-95.
434
C. KRAJEWSKI ETAL.
Blacket MJ, Krajewski C, Labrinidis A, Cambron BC, Cooper S, Westerman M. 1999.
Systematic relationships within the dasyurid marsupial tribe Sminthopsini a multigene approach.
Molecular Phylogenetics and Evolution 12: 140- 155.
Bos D, Carthew S. 1999. Population dynamics and movement of the southern ningaui (Ningaui
yvonnae) in the middleback ranges of South Australia. Abstract. Australian Mammal Sock9 Meeting,
5-8 July 1999, University of Western Sydney, Hawkesbury, New South Wales.
Braithwaite RW. 1979. Social dominance and habitat utilization in Antechinus stuartii (Marsupialia).
Australian Journal of <oolvgy 27: 517-528.
Braithwaite RW, Lee AK. 1979. A mammalian example of semelparity. American Naturalist 113:
151-155.
Crowther MS, Dickman CR, Lynam AJ. 1999. Sminthopsti griseouenter boullangeremi (Marsupialia:
Dasyuridae), a new subspecies in the S. murina complex from Boullanger Island, Western Australia.
Australian Journal o f ~ o o l 47:
v ~ 2 15-243.
Dickman CR, Braithwaite RW. 1992. Postmating mortality of males in the dasyurid marsupials,
Daryunu and Parantechinus. Journal of Mammalogy 73: 143-1 47.
Dickman CR, Haythornthwaite AS, McNaught G, Mahon PS, Tarnayo B. 1999. Population
dynamics of three species of dasyurid marsupials in arid central Australia. Abstract. Australian
Mammal Sock9 Meeting, 5-8 July 1999, University of Western Sydney, Hawkesbury, New South
Wales.
Dixon JM. 1989. Thylacinidae. In: Walton DW, Richardson BJ, eds. Fauna ofAustralia, Volume IB,
Mammalia. Canberra: Australian Government Publishing Service, 549-559.
Dwyer PD. 1977. Notes on Antechinus and Cacartetus (Marsupialia) in the New Guinea Highlands.
Proceedings ofthe Royal Socieb o f Queensland 88: 69-73.
Flannery T. 1990. Mammals ofNm Guinea. Carina: Robert Brown & Associates.
Fox BJ, Whitford D. 1982. Polyoestry in a predictable coastal environment: reproduction, growth
and development in Sminthopsir murina (Dasyuridae, Marsupialia). In: Archer M, ed. Carnivorous
Marsupials. Mossman: Royal Zoological Society of New South Wales, 39-48.
Friend GR, Johnson BW, Mitchell DS, Smith GT. 1997. Breeding, population dynamics and
habitat relationships of Sminthopsis dolichura in semi-arid shrublands of Western Australia. M.ildl@
Research 24: 245-262.
Gibson DF, Cole JR. 1992. Aspects of the ecology of the mulgara, Dasycmus cnsticauda (Marsupialia:
Dasyuridae) in the Northern Territory. Australian Mammalogy 15: 105-1 12.
Johnson LAS, Briggs BG. 1975. O n the Proteaceae - the evolution and classification of a southern
family. BotanicalJournal ofthe Linnean Sock9 70: 83-1 82.
Kirsch JAW, LaPointe FJ, Springer MS. 1997. DNA-hybridisation studies of marsupials and their
implications for metatherian classification. Australian Journal of ~ o o l 45:
o ~ 2 1 1-280.
Kitchener DJ, Cooper N, Bradley A. 1986. Reproduction in male Ningaui (Marsupialia: Dasyuridae).
Australian W d l @ Research 13: 13-25.
Krajewski C, Buckley L, Woolley PA, Westerman M. 1996. Phylogenetic analysis of cytochrome
b sequences in the dasyurid marsupial subfamily phascogalinae: systematics and the evolution of
reproductive strategies. Journal ofMammalian Evolution 3: 8 1-91.
Krajewski C, Wroe S, Westerman M. 2000. Molecular evidence for the pattern and timing of
cladogenesis in dasyurid marsupials. <oolo&al Journal of the Linnean Socieg 130: (in press).
Krajewski C, Young J, Buckley L, Woolley PA, Westerman M. 1997. Reconstructing the
evolutionary radiation of dasyurine marsupials with cytochrome b, 12s rRNA, and protamine gene
trees. Journal of Mammalian Euolutwn 4 2 17-236.
Lee AK, Cockburn A. 1985. Evolutionaly Ecology o f Marsupials. Cambridge: Cambridge University
Press.
Lee AK, Woolley P, Braithwaite RW. 1982. Life history strategies of dasyurid marsupials. In:
Archer M, ed. Carnivorous Marsupials. Mossman: Royal Zoological Society of New South Wales,
1-1 1.
Maddison WP, Maddison DR. 1992. MacClade: Anahsir of Plylogmy and Charach Evolution, Vusion
3.0. Sunderland Sinauer Associates.
Morton SR, Armstrong MD, Braithwaite RW. 1987. The breeding season of SminthopsZ;r vip'niae
in the Northern Territory. Australian Mammalogy 1 0 41-42.
Muirhead J, Wroe S. 1998. A new genus and species Badjcinus turnbulli (Thylacinidae: Marsupialia),
from the Late Oligocene of Riversleigh, northern Australia, and an investigation of thylacinid
phylogeny. Vitebrate PaleontoloD 18: 6 12-626.
-
I<\'OI,UTIC>N O F Rl<PROl)LC'I'l\T Sl'ILYI'F,GIES IN I):\SYURID II.4KSUI'IAI.S
43.5
Page1 M. 1999. The maximum likelihood approach to reconstructing ancestral charactcr statcs of
discrete characters on phylogenies. $stmatic B i o l o ~48: 6 12-622.
Read DG. 1984. Reproduction and Iirccding of Plan&ulegileti and P tmuiro.\tric [llassupiala: Dasyuridac).
Australian '2lammalogv 7: 161 1 7 3.
Read DG, Fox BJ, Whitford D. 1983. Notes on breeding in Smint//opsz.s (Marsupiala: U
Aiutralian .IlammalnQ 6: 82-92.
Ride WDL. 1964. dntrchinu mamnndar, a nen species of dasyurid marsupial from thc Pilbara District
of \Vestern Australia. with remarks on thc classification of Antechinus. I.V?steslpm duslralian ,,Vaturalz.\/
9: 38-65,
Smith M. 1982. Revicw of the thylacinc (hlarsupialia, Thylacinidae). In: Archer h l , cd. Camirwou\
,Lfarsupials. Mossman: Royal Zoological Socirv of Ncw South M'alcs, 237-253.
Soderquist TR, Serena M. 1990. Occurrencc and outcome of polyocstry in wild western quolls,
Dasyuiw genffrozi (hfarsupialia: Dasyuridae). ilu.rtralian M a m m a l n ~13: 205-208.
Specht RL. 1981. Major vegetation formations in Australia. Pages 163-298. In: Kcast A. ed. Ecnlngical
Biogeograplly ofAustralia. The Hague: Junk.
Strahan R. 1995. 772e A21arnmalsofAustralia. Chatswood: The Australian hIuseum and Keed Books.
Taplin ME. 1980. Some obsenations on thc reproductive liiolocq of Sminthnpszs r,ilgzniae (Tarragon)
(hlarsupialia: Dasyuridae). Australian zoologist 20: 407 41 8.
White ME. 1994. A j e r the Grerning: T h e Browning ofAustralia. New South CVales: Kangaroo Press.
Woolley PA. 1984. Reproduction in r1ntechinomy.c lanker ('spenceri' form) (klarsupialia: Dasyuridae): field
and laboratory obsenations. z4ustralian Mldlije Re.rearch 11: 48 1-489.
Woolley PA. 1988. Reproduction in the Ningbing Antechinus (Marsupialia: Uasyuridae): field and
laboratory obscnations. ilustralian Wildlzji Research 15: 149-1 56.
Woolley PA. 1990a. Reproduction in Sminth0psi.s macroura (Marsupialia: Dasyuridac) I. Thc female.
Australian Journal of <oologv 38: 187-205.
Woolley PA. 1990b. Reproduction i n Sminthopsis macroura (klarsupialia: Dasyuridac) 11. T h r male.
Australian Journal of zonlogv 38: 207-2 1 7.
Woolley PA. 1991a. Reproduction in DaJykaluta mamondae (hfarsupialia: Dasyuridae): ficld and
laborator); observations. Australian ,&ma1 of<oologv 39: 549-568.
Woolley PA. 1991b. Reproduction in Psrudantrchinus mac.donnrllens/s (hfarsupialia: Dasyuridae): firltl
and laboratory obsrnations. tlildlije Re.erarch 18: 13 25.
Woolley PA. 1991c. Rcproducti\re beha\ior of captive Boullanger Island Dibblrrs, Pal-antrchinus apicali.\
(Marsupialia: Dasyuridac). Wildlife ReJeaich 18: 157- 163.
Woolley PA. 1994. Thc dasyurid marsupials of New Guinea: use of museum specimens to assrss
seasonality of brccding. Scirncr in .Lbw Guznea 20: 49-56.
Woolley PA. 1995. Observations on reproduction in captive Paranlec.hinu.\ bilami (Marsupialia: Das)uridac). Azlstralian AfammaloQ 18: 83-85.
Woolley PA, Ahern LD. 1983. Obscnations on the ecolog and rcprductive bioIo<qof .linzntho/~\ic
1eucopu.s (Marsupialia: Dasyuridae). ProcrediTgs qf thr Royal SorieQ of Pktoria 95: 169 180.
Woolley PA, Gilfillan SL. 1990. Confirmation of polyoestry in captive white-footed dunnarts.
Sminthopsis leucopus. Australian l\farnmalogy 14: 137-1 38.
Woolley PA, Valente A. 1986. Reproduction in Sminthopsis lnngimudata (hlarsupialia: Dasyuridac):
laboratory observations. .lusti-alian Mildlzji Research 13: 7-12.
Wroe S. 1996. LLfurbacinus gad@uli (Thylacinidae: Marsupialia), a vcry plesiomorphic thylacinid from
the Miocene of Ri\w-sleigh, northwestern Queensland, and thr problem of paraphyly for thc
Dasyuridae (hlarsupialia). Journal oJ PalenntoloQ 70: 1032-1 044.
Wroe S. 1999. The geologically oldest dasyurid, from Miocenr deposits of Ri\rersleigh. north-\vest
Queensland. Paleontnlou 42: 501-527.
~