1
Sex differences in parental care
2
Hanna Kokko, Michael D Jennions
Research School of Biology,
The Australian National University,
Canberra, ACT 0200,
Australia
4
6
8
WORD COUNT: 10 033
10
References = 88
Figures = 4
12
2
14
16
18
6.1 Introduction
Have a look at the images in Fig. 6.1. Although you might not be able to identify the species,
can you tell which individual is the male, and which is the female? If you always picked the
animal on the left as the male you were correct. Why is it that we can often tell the sexes
apart, even in a species we have never seen before?
<Fig 6.1 about here>
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
In general, the assumption that females are the main care-givers is biologically justified. In
about 90% of mammals, males inseminate females, fertilize eggs and thereafter provide no
parental care. Likewise, in arthropods there is a very strong bias towards female-only parental
care. Biparental care is rare, and male-only care has only evolved in eight of the many
lineages in this enormous taxonomic group (Tallamy 2000) (although new discoveries of
male-only care continue to be made; e.g. Requena et al. 2010). In reptiles parental care is
uncommon: eggs are usual buried and abandoned. However, when egg guarding occurs it is
almost always by the female, with biparental care having evolved from female-only care in
some crocodilians (De Fraipont et al. 1996; Reynolds et al. 2002). In birds, the sex bias in
parental care is weaker but still present. There is female-only care in 8% of species,
biparental care in about 81% of species, and biparental care with input from helpers occurs in
another 9% of cooperatively breeding species (Cockburn 2006). Even when both parents care
there is, however, still a propensity for the female to spend more time building the nest and
incubating eggs (Schwagmeyer et al. 1999; Møller & Cuervo 2000).
Before we become too confident in our abilities to discriminate between the sexes based on
which sex provides care, it is worth looking at Fig 6.2. How confident are you about picking
the male in these species? If you are less certain than you were before your trepidation is
justified. Amphibians show a fantastic variety of forms of parental care and sexual division of
parenting. In some species females lay trophic eggs to feed tadpoles, brood frogs inside their
stomachs or carry developing frogs in fleshy capsules on their backs. Similarly, males
sometimes guard schools of tadpoles, build canals to move tadpoles between temporary
pools, or carry them to new site (Wells 2007). There are even species where males carry
froglets on their backs (Bickford 2002). Biparental care is rare in anurans but male-only and
female-only care have evolved equally often from a state of no care (Beck 1998; Reynolds et
al. 2002; Summers et al. 2006). The caring frog parent in Fig 6.2 is therefore equally likely to
be a male or female.
<Fig 6.2 about here>
In teleost fish, the sex of the carer is more predictable, but this is because male-only care is
found in nine times as many genera as female-only care (Gittleman 1981; Gross & Sargent
1985; Reynolds et al. 2002). Excluding livebearers, in fish taxa where phylogenies have been
analysed for evolutionary transitions, female-only care is a more recently derived state,
arising almost exclusively from biparental care (e.g. cichlids: Goodwin et al. 1998; Klett &
Meyer 2002; Gonzalez-Voyer et al. 2008; ray-finned fish: Mank et al. 2005). Finally, while
birds tend to show a female bias in parental care, shorebirds often show ‘sex role reversal’
with males caring for eggs and nestlings and females being larger, more aggressive and more
brightly coloured than males (Thomas et al. 2007). Interestingly, if you had based your
criteria for sex identification on the assumption that males invest more in courtship and mate
attraction you might have been on safer ground. Even in frogs with male-only care it is still
3
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
almost always the case that only males call to attract females. Likewise, in fish with maleonly care, if there is sexually dimorphism, males still tend to be more brightly coloured than
females.
Many non-biologists would probably have misclassified examples in figure 6.2 by assuming
that females are always the primary carer-givers, a product of naïvely extrapolating from the
still common norm in human societies. Even so, and despite variation in care roles in major
taxa, among related species, and even within species (Trillmich 2010), our brief taxonomic
review reveals a strong trend towards female-biased care in nature. This is true whether we
consider the absolute number of species where female care predominates (given the sheer
number of invertebrates), or the number of independent evolutionary transitions towards
greater female than male care. In this chapter we will explain why females tend to provide
more care, and what, on the other hand, maintains such fantastic diversity in outcomes.
6.2 Why does an individual’s sex predict its behaviour?
The predictable relationship between an individual’s sex and its breeding behaviour is most
parsimoniously explained if there is something inherent to being a male or a female that
applies across all taxa. We can immediately eliminate the mechanism of sex determination as
a common denominator. The method of sex determination varies greatly among species:
some have sex chromosomes, others are haplodiploid or exhibit temperature-dependent sex
determination. The mechanism can even differ among populations of the same species (Pen et
al. 2010). Moreover, even in species with sex chromosomes, females can be the
heterogametic sex (e.g. ZW in birds and butterflies) or the homogametic sex (e.g. XX in
placental mammals), and platypuses have a set of ten sex chromosomes.
Given this diversity, how do we decide to which sex an individual belongs? Why can we state
with confidence that pregnant individuals are always male in seahorses even though the exact
mechanism of sex determination is unknown (Tony Wilson, pers. comm.)? The answer is
simple: biologists have agreed on a convention that individuals producing the smaller of two
types of gametes — sperm or pollen— are called males, and those producing larger gametes
— eggs or ovules — are called females. This classification works unambiguously for a far
larger number of species than any other systematic sex difference because, when gamete size
varies, it usually does so with strong bimodality (i.e. anisogamy). In species where different
mating types are not distinguishable based on gamete size the terminology of two sexes —
males and females — is abandoned. We then talk of sexual reproduction between isogamous
mating strains, usually labelled + and –.
Armed with this information of what ‘being a male or a female’ means, we can now rephrase
our earlier question about female-biased care: why is producing a small gamete associated
with not providing additional parental care once gametes fuse to form a zygote? To answer
this question it is important to understand the selective pressures that create anisogamy with a
bimodal distribution of gamete sizes. In turns out that the reasons why a parent might be
sparing in provisioning each gamete with resources have something in common with the
reasons why the same parent might be less willing to subsequently invest in providing
additional parental care. The resources in the gamete that are available to the zygote after
fertilization represent the minimal level of parental care provided by a parent. A study of the
evolution of gamete size is therefore simultaneously a study of the evolution of parental care.
4
6.3 The first sex differences in parental care: the evolution of anisogamy
102
104
106
108
110
112
114
116
118
120
122
124
126
128
130
132
There are several explanations for the evolution of anisogamy (see the excellent review of
Lessells et al. 2009). One argument is based on elevating offspring fitness by reducing
intracellular conflict. This favours uniparental inheritance of cellular organelles. Such
uniparental inheritance does not, however, always occur through females. For example,
although mitochondria are usually maternally inherited they are transmitted via sperm and
pollen in several taxa (Cummins 2009). Here we focus on the more standard explanation that
is also more closely linked to the male-female definition and the economics of parental
investment. It is based on the economics of gamete production: reducing the investment in
each gamete allows the parent to produce more gametes on the same budget.
This increase in gamete production is an example of the more general principle of a trade-off
between offspring size and number (Lack 1947, Smith and Fretwell 1974, Lloyd 1987). From
a parental perspective the optimal offspring size depends on maximizing the number of
surviving offspring given a trade-of between offspring survival (which is assumed to be sizedependent) and parental fecundity. For example, if a mother in an egg-laying species could
produce one very large offspring that survived to sexual maturity with a probability of 70%,
or two medium sized offspring that each had a 40% survival probability, or three small
offspring who have a 10% survival probability, the expected number of surviving offspring is
0.7, 0.8 and 0.3 respectively. Selection therefore favours the intermediate clutch size of two.
In general, optimality models for offspring size assume that only the mother determines their
size. Sperm is assumed to make a negligible material contribution to offspring size, i.e.,
sperm are assumed to be tiny. When we turn our attention to the evolution of anisogamy,
such an a priori assumption is clearly inappropriate. The number-size trade-off therefore
involves additional considerations when the gamete size produced by either parent is free to
evolve.
A gamete first has to survive until it finds a compatible gamete to form a zygote. Thereafter,
zygote survival will depend on the combined effect of both gametes on offspring size. In a
deviation from the simple optimization of the size-number trade-off, small gametes are not
necessarily destined to give rise to offspring with poor survival prospects since they might
fuse with large gametes. In effect, one of the mating types, which by definition we end up
labelling as ‘male’, is selected to produce small gametes (sperm) to ‘parasitize’ the
provisioning of the other mating type.
142
Consider a scenario where there is isogamy with + and – mating types. Initially mutants of
one mating type that produce smaller that average gametes (a proto-male) will suffer a
decline in offspring survival because this slightly reduces zygote size, but this is more than
compensated for by the proto-males’ higher fertilization success, simply because they can
produce more gametes. In turn, being fertilised by a smaller gamete selects for increased
gamete size in the other mating type (proto-females) to ‘compensate’ for the decline in zygote
size that begun with the evolution of sperm. Game theory models that make plausible
assumptions about the size-number trade-off and the effects of gamete and zygote size on
survival show that the resultant disruptive selection can readily drives the evolution of
anisogamy (Parker et a. 1972, Bulmer & Parker 2002).
144
These classic models for the evolution of anisogamy do not explicitly model one
consequence of altering gamete size: a gamete might never become part of a zygote. If one
134
136
138
140
5
146
148
150
152
154
156
158
160
162
164
166
168
170
172
174
176
178
180
182
184
186
188
sex makes many more gametes than the other then most of these gametes are doomed to
failure. This simple fact invokes the Fisher condition (Houston & McNamara 2005), named
after R.A. Fisher who insightfully noted that since each offspring in a sexually reproducing
diploid species has one genetic mother and one father, the total reproductive output of the two
sexes must be identical at the population level. Fisher used this simple fact to explain the
ubiquity of a 1:1 offspring sex ratio: if one sex is overproduced in the population then the
other must have higher per capita reproductive success. The only stable outcome occurs when
there is no such advantage to be gained, which is reached when parents invest equally into
sons and daughters. If producing an offspring of either sex costs the same then the offspring
sex ratio will be 1:1.
If genes from one individual of each sex are required to create a zygote, then the Fisher
condition also means that the number of successful sperm must equal the number of
successful eggs. In the case of gametes we never observe a 1:1 ratio of eggs to sperm. The
size-number trade-off means that the sex with small gametes is producing a vast oversupply:
most of these sperm are wasted. At first sight this makes the idea of ‘gametic parasitism’ by
males seem less lucrative. If most sperm will not find an egg to parasitize, why does this not
select for males to produce fewer, larger gametes? The answer lies in the trade-off between
the likelihood of his gametes encountering other gametes to form a zygote, and increasing the
survival of a zygote. When we produce a model for gamete size evolution that tracks the
success of individual gametes and explicitly derives survival over time, isogamy can be
maintained when gametes of both sexes are ‘reasonably likely’ to become a zygote (Lehtonen
& Kokko 2011). Anisogamy evolves when there is a deviation from this ‘reasonable
likelihood’ due to either a high encounter rate such that more numerous type of gametes
compete locally for the less numerous type, or to a low encounter rate so that both gamete
types struggle to find each other.
Why do such vastly different scenarios both lead to anisogamy? The answer can be visualised
in terms of the strength of selection on proto-males to increase the number of their own
gametes that will fuse with compatible gametes, versus the strength of selection to elevate the
resultant survival of the zygote (i.e. increased parental care in the form of larger gametes).
The combination of these two processes determines the net reproductive success of protomales.
In a high encounter scenario (e.g., a broadcast spawner in a small body of water with many
spawning parents) any numerical imbalance between smaller and larger gametes implies a
situation akin to sperm functioning like tickets in a raffle where a limited number of prizes
are guaranteed (the eggs are there). The only way to do better than other participants is to
have more tickets. Admittedly this is an unusual raffle: each participant can be visualised as
being given a piece of paper which he can cut up to make as many tickets as he wants.
Succeeding in this stage of competition — the sperm competition stage — is a better
predictor of male reproductive success than ensuring that the resulting zygote has a greater
chance of survival. Increased zygote survival would still be beneficial, but it is difficult for a
male to elevate this component of fitness economically. If a male is to increase the amount of
parental provisioning he has to put extra energy into every sperm because the male cannot
predict in advance which of his ‘tickets’, if any, will win a prize. Most of this parental care
would be wasted because few of his sperm will locate an egg. Later in this chapter we will
return to the problem of determining whether or not male care will yield dividends, as it is
6
190
equally relevant for considering post-zygotic male parental care when paternity is uncertain
as it is for pre-zygotic provisioning.
192
Why does the opposing scenario of low gamete density also select for anisogamy? The need
for a male to increase the likelihood that one of his gamete locates an egg before a rival’s
gametes is now replaced by the task of simply locating an egg. The advantage of producing
numerous sperm now comes from raising the likelihood that at least some sperm will find an
egg. The analogy with a raffle with a guaranteed prize is no longer valid (some prizes are
never found); perhaps a better analogy is putting up many reward notices around your
neighbourhood in the hope that one will be read by the person who has found your missing
cat. Females are no longer victims of parasitism: they too benefit from males producing
numerous, albeit small, sperm because some of their eggs would otherwise fail to be
fertilized. It is an obvious question to ask why eggs do not evolve to be small too, because the
largest absolute number of fertilizations might well occur if both sexes evolved very small
gametes. This solution is not feasible, however, as the resulting small zygotes would have
very low survival prospects. Given that both gamete type cannot simultaneously evolve to be
very numerous and very small, there is even the potential for selection for bigger eggs to
provide a larger target for sperm (Levitan 2006).
194
196
198
200
202
204
206
208
210
212
In summary, the most basic form of parental care is the provisioning of gametes that
determines initial zygote size. Male parental care is reduced due to sexual selection to locate
unfertilized eggs. Returns on parental investment remain low if difficulties in finding eggs to
fertilize mean that most of the investment is wasted. This can happen either because
unfertilized eggs are rare (low density), or because of the presence of rivals who compete to
fertilise eggs.
6.4 What happened next? Sex roles in post-zygotic parental care
214
216
218
220
222
224
226
228
230
232
The previous scenarios are usually envisaged as eggs and sperm being released into an
aquatic medium so that the zygote forms away from parents. If so, anisogamy means that the
female is by definition the primary care provider. This book would not exist, however, if the
story ended there. Post-zygotic care of offspring has evolved numerous times in both external
and internal fertilizers. Minimally, however, its evolution requires that the male and/or
female parent can locate zygotes formed from their gametes. In internal fertilizers, the zygote
is initially surrounded by parental tissue (e.g. the brood pouch of a male seahorse after the
female has inserted her eggs, or the reproductive tract of a female guppy after a male has
inseminated her). Whichever sex provides this tissue is in a better position to continue to
provide care. In this sense it is unsurprising that, say, female mammals provide more postzygotic care than males. Given gestation and live birth females have the option to provide
additional energy resources to their offspring through a placenta (for placental evolution in
fish see Pollux et al. 2009) or post-birth provisioning in the form of lactation. If the males
remain nearby, however, there is no obvious barrier to prevent him from evolving the ability
to lactate (Hosken & Kunz 2009, Kunz & Hosken 2009). Males have the same basic
physiological and morphological trait required for lactation as females (Wynne-Edwards &
Reburn 2008). Similarly, when fertilization is external and eggs are laid onto the substrate
both parents are present when zygotes form.
Seahorses are a reminder that ‘pregnancy’, or more broadly ‘baby-carrying’, are not a priori
tasks exclusively performed by females. Either sex could take sole responsibility for carrying
7
234
236
238
240
242
244
246
248
250
252
254
256
258
260
262
264
266
268
270
272
274
276
278
and/or guarding fertilized eggs, especially when there is short interval between fertilization
and zygote availability (e.g. in external fertilizers). Nothing in principle prevents the male
from doing this: male frogs transport tadpoles to water; in Belastomatid water bugs females
cement fertilised eggs onto the male’s back (Inada et al. 2011); and male-only guarding of fry
occurs in numerous fish (Mank et al. 2005). One approach to explain current differences in
male and female care is therefore to focus on environmental or social factors that influence
the likelihood each parent will encounter its own offspring, while treating existing sex
differences (e.g. female gestation) as constraints. For example, post-zygotic male care might
be related to whether or not there is sufficient food for a female to be able to remain on a
male’s territory during pregnancy. This kind of explanation is often instructive to account for
special cases such as sex role reversal in taxa where females usually provide the bulk of.
These explanations invoke already existing constraints (sex biases), however, so they are
insufficient for explaining the general trend for females to provide more post-zygotic care
than males. We are primarily interested in working from first principles, rather than invoking
ecological and phylogenetic contingencies. We want to determine why selection in most taxa
favours the large gamete producer remaining with zygotes, rather than taking this as given. In
other words, gestation in mammals seems to predispose females to be the sole care-givers,
but this only raises the question why females rather than males evolved the ability to gestate.
Ancestral sexual asymmetries in parental care can constrain much of the subsequent variation
we observe among extant species, but they in turn must also be explained. So, leaving aside
whether one sex is more likely to be physically present when zygotes form, what else might
explain a bias towards greater female care?
An obvious point to make is that any explanation for why females care more than males must
take into account the potential for sexual conflict over caring (Lessells 2006; Johnstone &
Hinde 2006; Olson et al. 2008). Unless there is lifelong monogamy each parent would
probably do better if it could convince the other parent to bear the full costs of caring. Indeed,
many behavioural studies investigate species where one sex compensates by caring more
when the other sex cares less, so that survival of the young is only compromised to some
degree. The short-term stability of biparental care is largely based on the assumption that any
such compensation is only partial, as complete compensation would favour immediate
desertion by one parent. In general this assumption seems valid (Harrison et al. 2009). There
is often an emphasis on symmetry when parents are ‘negotiating’ levels of care: individuals
of either sex benefit from caring less or, in extreme cases, deserting the brood first so that the
other parent has to care alone (Szentirmai et al. 2007). In a longer term evolutionary
perspective (which might no longer involve a negotiation over biparental care if one sex does
not provide care), this raises the question of why it is so often the male who deserts and
leaves the female with the brood.
Most empirical studies of parental care are based on a comparison of the costs/benefits of
caring versus deserting, making the assumption that these can be detected because both sexes
adaptively adjust their behaviour to maximise their fitness. This allows us to infer how
specific factors affect these costs/benefits for each sex (e.g. Jennions & Polakow 2001), while
taking into account that how the other sex responds affects the cost of providing less care. In
an evolutionary context, however, we must ask how changes in the mean behaviour of each
sex affect the fitness of the other, creating selection pressure for changes in each sexes
behaviour until an evolutionary stable state is reached where individuals of neither sex can
improve their fitness by behaving differently.
8
280
282
284
286
288
290
292
294
296
298
300
302
304
306
308
310
312
314
316
318
320
322
In the following three sections we provide a non-mathematical account of recent models that
account for the evolution of female-biased care, as well as for variations around this theme.
Our account is mainly based on the arguments of Queller (1997) and the subsequent
modelling and extension of these ideas by Kokko & Jennions (2008). We start by explaining
what factors select against male care. Second, we discuss factors that limit the fitness gains
males can achieve by caring less. These limitations help explaining why male care or
biparental care can evolve while an overall female bias towards greater care-giving still
persists.
6.5 Uncertainty about parentage reduces male care
Earlier we stated that anisogamy results in sperm vastly outnumbering eggs. There are two
assumptions behind this statement. First, that there are not many more female than male
adults currently trying to breed. Second, that the total budget each sex invests into gamete
production is of a similar size. If, for example, each male only produced a few sperm or there
were very few males releasing sperm per receptive female then sperm would not outnumber
eggs. Given that the selection driving anisogamy results in an enormous size difference
between eggs and sperm this would, however, require very large deviations of the breeding
sex ratio towards females and/or towards greater female than male reproductive investment
before sperm become rarer than eggs. Thus, the numerical asymmetry is fairly robust: sperm
compete to locate eggs before their rivals, rather than eggs participating in a similar
competition for sperm. Of course, the numerical abundance of sperm does not make sperm
competition inevitable. For example, there are insect species where females only seem to
mate once (Arnqvist 1998) and males still produce small sperm. There is, however, some
evidence that there is then natural selection for males to produce fewer and/or larger sperm
(e.g. Simmons & García-González 2008). Also, as already discussed, very low adult densities
in a broadcast spawner have the comparable effect of eliminating sperm competition (Levitan
2010).
Exceptions aside, given the numerical abundance of sperm, it is likely that sperm from
several males are present near the same egg shortly before fertilization. This creates
uncertainty as to which male has sired a zygote. This statement is certainly true when
fertilization is external and it remains true with internal fertilization if females also mate
multiply (which is almost always the case; Slatyer et al. 2011). The female situation is
different. Eggs are rare, so when a female remains in the vicinity of her eggs there is typically
little likelihood that she will mistake the eggs of another female for her own. Once again,
there are exceptions. Egg dumping occurs in some birds (Petrie & Møller 1991) and insects
(Tallamy 2005) and, in extreme cases, eggs that a female might consider her own are actually
those of another species (e.g. cuckoos). When there is communal spawning by females the
mixing of eggs also makes it difficult or impossible for females to know which zygotes are
their offspring. Still, since sperm are more numerous than eggs, there is a sexual asymmetry
in that male are more often in situations where they are uncertain about a zygote’s parentage
than are females. Although we will stick to convention and talk about how ‘uncertain
parentage’ (usually uncertain paternity) lowers the benefits of caring, it is actually the
certainty of the decline in mean relatedness to putative offspring that matters. For example, if
sperm from three males is released when a single female spawns it must follow that the
average reproductive success of these males is a third of a clutch.
9
324
326
328
330
332
334
336
338
340
Across species, or between the sexes, higher uncertainty of parentage decreases the benefit of
caring. Unfortunately, the literature on how confidence of paternity relates to male care is
built around questions concerning phenotypic plasticity of individuals within a species (i.e.
adaptive shifts in the extent of male care over his lifetime). It is easy to become confused and
question the validity of our statement. Before dealing with some of the sources of confusion,
we would like to reassure the reader by presenting a few simple questions. If the only benefit
of caring is to increase the survival of one’s own offspring, should a male who has no
paternity in a brood provide care? (Answer: No). Is the maximum benefit gained from caring
linearly related to the proportion of offspring in a brood that are your own? (Answer: Yes).
So, if species vary in the mean uncertainty of paternity (certainty of siring fewer offspring in
each brood), and nothing else, is the maximum benefit of care lower for males when
uncertainty is greater? (Answer: Yes). So, if females have a lower uncertainty of parentage
than males, is the maximum benefit of caring greater for females? (Answer: Yes). All that is
then required is to realise that any function relating caring to its benefits must asymptote at a
maximum: at the extreme, caring cannot ever increase offspring survival beyond 100%. If
each sex provides the same quality of care, the lower maximum benefit of caring for males
must mean that the marginal benefits from additional care decline, at some point at least,
more rapidly for males than females.
342
So, why might one be confused about the relationship between the certainty of paternity and
male care? There are at least three sources of confusion, which we deal with below.
344
6.5.1. Why exactly does paternity matter?
346
348
350
352
354
356
358
360
362
364
366
368
It is easy to fall into the trap of arguing that paternity should not matter when lifetime fitness
depends on balancing current and future reproduction. If paternity is likely to be higher in a
future breeding attempt, then selection might favour ‘saving’ energy to invest more into one’s
own survival and future care. If a male consistently has the same likelihood of paternity,
however, his current reproduction has been argued to be devalued by the same amount as his
future reproduction, implying that across species average paternity will not matter when
determining the mean level of male care (Maynard Smith 1977). The flaw in this argument is
that it neglects the fact that males gain offspring not only in broods for which they care, but
also by siring extra-pair young (Queller 1997; Houston & McNamara 2002). The flipside of
lost paternity is that extra-pair mating opportunities must exist elsewhere. Because these
opportunities are part of a male’s future reproductive success (but are not included in the
fitness gains from his current brood), low paternity, even if it is consistent across broods,
does not have equivalent effects on a male’s current and future reproduction (Westneat &
Sherman 1993, Queller 1997, Houston & McNamara 2002
Support for a relationship between lower than average paternity and a reduction in parental
care (or even greater ‘anti-care’ in the form of offspring cannibalism) can be found in withinspecies studies on phenotypically plastic male care (e.g. Neff 2003; Gray et al. 2007). It is
worth noting that tests for a negative effect of paternity on male care should not look at the
correlation across males. For example, unattractive males might have consistently low
paternity but, if they have no ability to gain extra-pair matings, their best bet might be to
invest more in care than do attractive males. Ideally one requires a comparison of changes in
care by a male when his current share of paternity differs from his future expectations (which
might be the population average, or an expectation based on the male’s own life history
circumstances). This is ideally achieved by experimentally manipulating a male’s perceived
share of paternity, which could also require a change in actual paternity if males can assess
10
370
372
374
376
paternity directly (e.g. Neff 2003). Given that we are primarily interested in the evolution of
care patterns given selection on the mean levels of investment by each sex, it is also
important to pay attention to phylogenetic studies. In birds these show that male parental care
is lower when extra-pair paternity is high (Møller & Birkhead 1993; Møller & Cuervo 2000).
This is an encouraging sign that paternity has a strong influence on male care, because mean
paternity can depend on other traits that might also affect male care, so that the reverse
relationship could arise (Houston & McNamara 2002).
6.5.2. Behavioural and evolutionary timescales are not equivalent
378
380
382
384
386
388
390
392
394
396
398
400
402
404
406
408
410
412
A common source of confusion is the tendency to ignore differences in how selection can act
over behavioural and evolutionary timescales. The average paternity in a population
influences male care decisions because male possess ‘evolutionary knowledge’ about their
expected paternity, based on strong selection in the past to provide a level of care that
maximizes fitness. Simultaneously, a male’s likelihood of being the parent of a given
offspring can differ from the average male in the current generation (or even his own lifetime
average). This is why it is dangerous to test the selection imposed by factors of interest (e.g.
level of paternity) by assuming that behavioural or other phenotypic responses will mimic the
predicted evolutionary response to selection. More specifically, how do we interpret failure to
detect a behavioural response to low paternity? It could mean that the focal factor does not
select for the predicted traits (i.e., lower paternity does not select for reduced care), or that
selection for adaptive phenotypic plasticity has been insufficient, perhaps because the
relevant cues or the ability to detect these cues do not exist to elicit an adaptive behavioural
response. These are not equivalent explanations. The former case would force us to question
whether our theoretical predictions are valid, the latter merely raises issues about constraints
on selection.
Failure to distinguish between how selection acts over behavioural and evolutionary
timescales is surprisingly common. It can easily lead to claims that female will pursue mating
strategies that promote the evolution of desirable male traits in the other sex. For example,
Morton et al. (2010) argued that, “genetic monogamy [...] may be a female tactic that reduces
the likelihood of males evolving counter-adaptations to female desertion”. Although it seems
self-evident that individual selection cannot favour traits that reduce undesirable evolutionary
outcome for other females who exist in the future (except as a fortuitous by-product of other
selective processes), this type of reasoning is sufficiently common that it is worth drawing
attention to it (the weakness of a similar argument is discussed by Kokko & Jennions 2010).
The existence of cues to detect deviations from prevailing parentage certainty is particularly
pertinent when considering how loss of paternity generates sexual conflict over care (Kokko
1999, Mauck et al. 1999). To illustrate the importance of behavioural and evolutionary
timescales, and the presence and accuracy of cues that allow for adaptive behavioural
responses, consider a species with biparental care where females sometimes engage in extrapair matings. If a female could signal to her mate that his share of paternity is above the
current norm in the population (i.e. the average over recent evolutionary time) she might
induce him to provide more care (Shellman-Reeve & Reeve 2000). Such a signalling system
is, however, unlikely to be stable. All females would benefit by signalling to their mate that
he had above average paternity. Instead selection is more likely to favour females who
conceal information about paternity, even though males would benefit from an honest signal.
11
414
416
418
420
422
424
426
428
If males do not detect deviations from mean paternity that occur over behavioural time then
within each generation females do not pay the cost for mating multiply that would arise if
males adaptively reduced their care (Kokko 1999). Selection on male parenting decisions will
therefore only depend on evolutionary knowledge of the long-term average paternity in the
population. This means that very small net benefits are enough to favour extra-pair mating
(Slatyer et al. 2011), even though they will make paternity decline over evolutionary time.
Despite the absence of within-generation selection on females not to mate multiply, mean
male care will still decrease over evolutionary time because of evolutionary knowledge of the
decline in paternity. If female become sufficiently promiscuous (or actively bias paternity
towards extra-pair males, e.g. Pryke et al. 2010), a possible evolutionary outcome is that
females end up as the sole caregivers (Kokko 1999) (Fig 6.3). This could result in a decline in
mean female breeding success representing a ‘tragedy of the commons’ (Rankin et al. 2007).
Females evolve to lose a valuable resource (male parental assistance) because no female
within any generation pays the price for her actions.
<Figure 6.3 about here>
6.5.3. Traits that protect paternity can coevolve with care
430
432
434
436
438
440
442
444
446
448
450
452
454
It can be challenging to establish the casual link between paternity and parental care if care
coevolves with traits that protect a male’s paternity (Kvarnemo 2006). What if there are male
traits that function to safeguard paternity that also enhance offspring survival? Or, what if
sexual selection for traits that increase male care has the by-product of increasing paternity?
If such traits exist, selection will lead to male care being greater than expected if we only
considered responses to changes in paternity arising from factors uncorrelated with male care
itself.
Protecting paternity involves better defence of unfertilized eggs (e.g. pre-copulatory mate
guarding), while enhancing offspring survival is mostly dependent on process that occur after
zygotes have formed. Thus, arguing that care can evolve as a correlated response to paternity
protection requires that the relevant traits are able to bridge this temporal gap. Kvarnemo
(2006) lists a range of intriguing possibilities as to how this might happen. For example,
nuptial gifts or courtship feeding can increase sperm uptake by females (increasingly a male’s
likely share of paternity) while also providing more nutrients for offspring. Similarly, in fish,
a well built nest might improve a male’s ability to gain paternity as well as making offspring
better defended. Current male care could also affect paternity in future breeding events. For
example, males might repeatedly breed with the same female and her propensity to engage in
extra-pair matings might be lower when the male has provided more care during the previous
breeding event.
A mechanistic explanation for a link between a male’s ability to gain paternity and to provide
care is that there are ‘constraints’, possibly mediated by hormonal trade-offs (McGlothlin et
al. 2007), that lead to behavioural consistency in different contexts (e.g. Royle et al. 2010).
For example, males that are more attentive at mate guarding might have a greater propensity
to aggressively defend offspring against predators. The ideas of Kvarnemo (2006) appear to
be worthy of further study, not least because any factor that keeps parents near putative
young is a prerequisite for the future provision of care.
12
456
458
460
462
464
466
468
470
472
474
476
478
480
482
484
486
488
490
492
494
496
498
500
6.6 Don’t bother caring till the going gets tough: the OSR
Although the primary sex ratio is usually near 1:1 due to the Fisher condition generating
negative frequency-dependent selection on the returns per offspring of each sex, there is no
selection to maintain sex ratio equality after parental care has ended (West 2009). Dramatic
differences in mortality between the sexes are possible. How then does the adult sex ratio
(ASR) affect parental care by each sex? This is an important topic as recent theoretical work
reveals some counterintuitive outcomes. A persistent theme in the literature has been the
claim that whichever sex experiences difficulties in finding mates will benefit from investing
more heavily into mating effort (Trivers 1972; Dawkins 1989; Clutton-Brock & Parker
1992). This has been interpreted to mean that, by implication, this sex will not gain equally
much benefit by providing more parental care.
The difficulty of finding mates is usually expressed as the operational sex ratio (OSR): the
ratio of sexually available males to females (Emlen & Oring 1977). The OSR takes the ASR
as its baseline, but modifies it by removing all individuals who are currently unavailable as
mates because they are caring for young and/or replenishing resources required to mate and
breed again. Individuals who are ready to mate are said to be spending ‘time in’ the mating
pool, the rest are spending ‘time out’ (Clutton-Brock & Parker 1992, Parker & Simmons
1996). There is therefore no guarantee of a linear relationship between the OSR and ASR
because sex divergence in parental care is a major determinant of the OSR in a way that is not
for the ASR (We note, however, that if there are sexual differences in caring and ‘time in’
and ‘time out’ activities have different mortality rates this will result in sex differences in
mortalities than affect the ASR. We will return to this issue in Section 6.7.).
The notion that difficulty in acquiring mates favours greater mating effort and less parental
care suggests that any initial bias in the OSR will be self-reinforcing over evolutionary time
until the rarer sex in the OSR is the sole provider of care. According to this logic, selfreinforcement happens because if sex A cares slightly more than sex B this increases its ‘time
out’ so that A becomes rarer in the mating pool. Sex B is then assumed to respond to the
intensified competition for matings by investing in traits that increase the likelihood of
mating. This diverts resources away from caring in sex B which, in turn, further selects for
greater care by sex A. This further increases the bias towards B in the OSR until eventually
sex A provides no care. If there is any initial bias favouring slightly greater care by one sex,
then this sex will eventually be the sole caregiver.
There is an obvious empirical problem with any account that makes strongly divergent care
roles appear inevitable: in reality, biparental care exists. Sometimes such cases can be
explained by the benefits of synergy. Care models predict, rather unsurprisingly, that
biparental care is more likely if two parents perform parental duties more efficiently (i.e.
greater offspring survival per unit of care) than either parent caring alone (Kokko &
Johnstone 2002). The real problem, however, is that the self-reinforcing process relies on an
effect of the OSR that, when explicitly modelled, appears problematic. Numerous models
have shown that an overabundance of males selects for more male care (Yamamura & Tsuji
1993, McNamara et al. 2000, Houston & McNamara 2002, Kokko & Jennions 2008). Similar
conclusions have been reached for other male traits that increase fitness by remaining with
the current mate. For example, selection for mate guarding a female to protect paternity,
rather than seeking out other females, becomes stronger in more male-biased populations
(Carroll & Corneli 1995, Fromhage et al. 2005, 2008). In hindsight, it is obvious that a male
13
502
504
506
508
510
512
514
516
518
520
522
524
526
528
530
532
534
536
538
540
is more likely to desert a current female or his young when there are more, not fewer,
reproductive opportunities available elsewhere, i.e. when the OSR is female-biased.
It is now worth recalling the insights gained from considering gamete size evolution. Gamete
size can show self-reinforcing divergence due to the greater abundance of sperm than eggs.
Why then does male abundance (in the OSR) not inevitably lead to a similar divergence
where the OSR becomes ever more male-biased? Instead, models predict that the difficulties
of securing additional matings make the male more likely to care. The solution depends
crucially on the seemingly trivial fact that a male is only in a position to provide care once he
has mated and his potential offspring exist. We noted earlier that anisogamy evolves under
sperm competition because “the male cannot predict in advance which of his ‘tickets’, if any,
will win the prize”. Contrast this with the evolution of post-zygotic male care, and assume
that the level of paternity is p (between 0 and 1). We can now ask how the expected future
number of reproductive opportunities with other females will influence whether the male
should continue to care to improve offspring survival, or desert and seek a new mate. A male
in a male-biased population experiences a trade-off between investing in offspring that
already exist and investing in gaining future ones that might never exist. Caring for existing
offspring offer an assured fitness return that remain unchanged (albeit discounted by p given
uncertain parentage) as future mating opportunities become scarcer, while the return from
future offspring is weighted by the likelihood of securing a mating. Securing a mating is
clearly more difficult when the number of competitors per female increases. For the same
paternity certainty p, a male will gain relatively more by improving the survival of his
existing young if securing additional matings becomes more difficult. This is why a malebiased OSR selects for male care. Some of his current paternal investment is wasted if
parentage p is less than 100%, but this is less waste than that experienced at the gamete
production stage. The abundance of sperm and intense competition for eggs favours reduced
pre-zygotic male care because a male cannot direct parental care towards the few sperm that
will end up with above average success in the same way that he can direct paternal care
towards already existing offspring.
In sum, a numerical bias in gametes towards sperm tends to create sperm competition, reduce
the per-sperm investment and lead to anisogamy (i.e. low pre-zygotic male care). Sperm
competition, by reducing the certainty of paternity, also selects against post-zygotic male
care. Intense pre-copulatory competition for mates (a male bias in the OSR), however,
favours greater post-zygotic male parental care. The same argument applies for females when
the OSR is female-biased. The Fisher condition tends to equalise the level of post-zygotic
male and female care, and it can explain why biparental care can exist even if synergistic
benefits of care are absent. If one sex cares for longer (or provides more resources so that it
takes longer to recover) it spends more time out of the mating pool. The opposite sex
therefore experiences difficulties in securing a mate when it returns to the mating pool. This,
in turn, selects for it to do something else that raises fitness (e.g. provide more care). This
negative frequency-dependent selection offers a plausible explanation why sex roles do not
always diverge to the extent that only one sex cares.
542
6.7 Orwell was right, not all animals are equal: sexual selection and the
adult sex ratio
544
If uncertainty of parentage and biased operational sex ratios were the only factors driving
parental care, we would predict that female multiple mating or group spawning select for
14
546
548
550
552
554
556
558
560
562
564
566
568
570
572
574
576
578
580
582
584
586
588
590
reduce male care, and that this is countered by frequency-dependent selection due to the
negative effect on male mating opportunities. The process is not self-reinforcing. Uncertainty
of parentage can easily promote a modest sex bias in levels of care (e.g. Fig. 3c in Kokko &
Jennions 2008), but the Fisher condition makes it difficult to explain uniparental care. This is
analogous to factors such as differences in production costs of each sex that select for a
biased primary sex ratio, but the inevitable frequency-dependent selection on the sex ratio
(again due to the Fisher condition) always prevents one sex from being completely absent.
How, then, do we account for uniparental care being so common in nature? Aside from very
low paternity prospects for males (Fig 6.4), what other explanations exist?
One complication missing from the preceding explanations is that not all individuals of a
given sex in the mating pool experience the same level of competition. The Fisher condition
dictates that the more numerous sex in the mating pool must, on average, have a slower
mating rate than the rarer sex. Even so, some individuals of this sex could still have a reliably
high mating rate. Sexual selection acts on traits that increase success when competing for
mates (or, more accurately, for access to gametes of the opposite sex). Individuals with traits
that are favoured by sexual selection will spend less than the average time for their sex in the
mating pool. Crucially, only individuals that mate are ever in the position where their
propensity to provide care is exposed to selection. Only they, having mated, have young that
they can apportion care to. Should these successful individuals care or desert?
If sexual selection is strong, mating per se indicates that an individual bears traits that will
also lead to above average mating success in the future. These individuals are less strongly
affected by the Fisherian frequency-dependent selection that inevitably lowers the mating rate
of the sex that is more common in the mating pool. Consequently, they will sooner reach the
point where they can gain more by returning to the mating pool than they can gain by
continuing to provide care.
Mating systems are usually defined by a combination of the sex-bias in parental care, the
intensity of sexual selection on each sex, and the level of multiple mating/group spawning.
These factors are all related to each other because, as we have shown, sexual selection and
uncertainty of parentage influence the sex-bias in parental care. If sexual selection is
sufficiently strong, it can seemingly counter frequency-dependent selection due to the Fisher
condition. Mated individuals of the sex that is common in the mating pool can benefit from
providing no care and deserting their young to try and mate again, if they possess traits that
make their mating success repeatable. Depending on the relative strength of sexual selection,
the system can shift from biparental care with a mild sex bias to totally uniparental care (see
Kokko & Jennions 2008 for numerical examples). The Fisher condition never actually
disappears, but with strong enough sexual selection, it no longer constrains the mated
individuals of the more numerous (less caring) sex in the mating pool to have a lower mating
rate than that experienced by the less numerous (more caring) sex.
The central role of sexual selection in determining parental care us makes it important that we
can measure the strength of sexual selection, see which individuals it favours, and contrast
this with the average difficulty of gaining matings experienced by an individual of that sex.
The OSR is directly related to the latter average, but less clearly to the former. It is tempting
to assume that sexual selection is always stronger on the sex that is more common in the
mating pool (Emlen & Oring 1977), but this is false. When there is a biased OSR the mean
mating success per time unit is low for the more common sex, but it does not follow that we
can equally easily assert that mating success in this sex will depend strongly on any particular
15
592
594
596
598
600
602
604
606
608
610
612
614
616
618
620
622
624
626
628
630
632
634
636
trait. Ultimately a causal link between a trait and the ability to acquire fertilization defines
sexual selection. This strict definition is essential for an individual’s mating success to be
repeatable which, as we have seen above, is needed to explaining patterns of care that deviate
from a small amount of variation around biparental care. Without traits that repeatable
increase mating success, a mated individual cannot assume it will have high mating success
after deserting its current young and avoid the full force of Fisherian frequency-dependence
which otherwise leads to biparental care.
It might often be true, but is unwarranted to simply assume that the bias in the OSR predicts
the strength of sexual selection (Fitze & Le Galliard 2008, Kokko & Jennions 2008, Klug et
al. 2010). For example, sexual selection has been shown to either intensify or diminish as the
absolute density of individuals increases (Kokko & Rankin 2006). If density-dependence has
different effects on how the mating success of each sex depends on specific traits then, for a
given population density, a change in the OSR could increase or decrease sexual selection on
each sex (Klug et al. 2010). The relationship between the OSR and sexual selection has to be
determined empirically but to date, very few studies have quantified this relationship (Fitze &
Le Galliard 2008). In particular, there are surprisingly few experimental studies that
manipulate the OSR and document sexual selection on focal traits (or relevant proxies; see
Weir et al. 2011). Studies documenting changes in variance in mating success are a poor
substitute because selection acts on traits and variance can be high due to purely chance
events (Sutherland 1985).
Understanding how the OSR related to the strength of sexual selection is crucial to account
for sex-biased parental care. If sexual selection intensifies with an increased bias in the OSR,
then mild initial biases in care become self-reinforcing. Successful members of the more
common sex in the mating pool will comprise an ever smaller subset of that sex. This means
that they are unaffected by the increasing bias in the OSR that lowers the mating rate for the
average individual of their sex. They can keep on deserting their young and maintain high
fitness by seeking out more matings. If, however, sexual selection saturates or even declines
at a high OSR (e.g. because it is impossible for a male to monopolize matings when there are
many competitors, Klug et al. 2010) then the subset of successful males becomes larger. The
per capita decline in mean mating rate with an increasingly biased OSR affects more of these
males, thereby selecting for greater male care.
In our view, one of the most interesting feedbacks between sexual selection, sex ratios and
parental care could operates through the relative mortality costs of caring and competing. The
ASR sets the baseline for the OSR. The OSR is the ASR corrected for the difference in care
by each sex that determines their ‘time out’ of the mating pool. Any factor that influences the
ASR will therefore shift the balance of the forces that determine the level of male and female
care. But what if sex differences in care themselves affect the ASR? Parental care and
competition for matings are often dangerous activities (Liker & Székely 2005) and/or require
initial investment in traits that can later affect mortality (Moore & Wilson 2002). If caring is
a more life threatening activity than an alternate activity such as mate searching, then
whichever sex cares more will become rarer in the population. This, all else being equal, will
— perhaps counterintuitively — select for the opposite sex to care more: being a member of
the overrepresented sex means that future mating opportunities are scarce. If, however,
seeking out mates is a more dangerous activity (costs of sexual competition are often severe)
then the sex that spends less time caring and more time in the dangerous activities of the
16
638
640
642
644
646
648
650
652
mating pool will become rarer in the ASR. All else being equal, surviving members of this
sex then have a relatively greater number of mating opportunities.
The outcome of a scenario where the more caring sex is rare is reminiscent of patterns seen in
many bird species: they tend to have a male-biased ASR and biparental care — nicely fitting
with the idea that the Fisher condition makes males less optimistic about securing matings
elsewhere when males are common (while extra-pair paternity can conceivably explain the
slightly greater female investment per brood). Intriguingly, bird species that lack male care
tend to have a female-biased ASR (Donald 2007), which resembles the typical situation in
mammals. In most mammals male care does not occur and there is often high male mortality
(usually attributed to male investment in traits that will increase mating success) (Wilson &
Moore 2002) so that the ASR is female-biased. In such a setting, males — especially those
who have already proven their ability to gain matings by siring young — can be relatively
optimistic about their future mating success, and are unlikely to benefit as much by caring for
young. Even though the lack of male care might, depending on the ASR, lead to a strongly
male-biased OSR, the Fisher condition reminds us that each baby produced has been sired by
a living male, which makes the mating optimism of males appropriate.
6.8 Conclusions
654
656
658
660
662
664
666
668
670
672
674
676
678
680
We have sketched a general outline of why females care more than males but, as every
evolutionary biologist knows, contingencies of history can result in taxon-specific traits that
sometimes seem to rewrite the rules. When observed outcomes disagree with the predictions
of theoretical model it is, however, often prudent to check whether the model’s assumptions
hold in the study system rather than automatically assuming the model is fatally flawed. An
interesting case study is the common occurrence of male-only care in many fish. We have
assumed that there is trade-off between caring and re-mating. This is a valid assumption in
many taxa. In fish, however, this does not seem to hold because parental care involves
activities such as nest defence and egg fanning that can be performed almost equally easily
for a few or many offspring (Stiver & Alonzo 2009). Males are therefore willing to receive
additional eggs from new mates. Females appear to be willing to lay eggs with already mated
males because the presence of eggs is a sign that the male is a good parent and/or is attractive
to other females (so might have attractive sons) and/or because the per capita risk of
predation or even cannibalism declines when a male guards more eggs. Little or no increase
in parental costs when caring for many zygotes can also explain some seemingly counterintuitive results. For example, despite our general argument that uncertainty of paternity will
lower care, male care increases with lower paternity in the ocellated wrasse Symphodus
ocellatus (Alonzo & Heckman 2010). This makes sense, however, if the male is still caring
for an absolutely greater number of offspring that he has sired, and the costs of caring are not
rising in direct proportion to their number of young.
Even so, there are still many taxa where it is a challenge to work out why only one sex cares,
and why it is the sex that it is. For example, many frogs seem to have rather similar breeding
systems and comparable ecological requirements so why within the same genus do only
males provides care in some species and females in others? Likewise, mouthbrooding in fish
would seem to place limits of mating rates, so why does male-only care still occur? Can these
differences be explained by ‘tipping points’ within the framework we have provided? For
example, that small ecological changes result in large shifts in the strength of sexual
17
682
684
686
688
690
692
694
696
698
700
702
704
706
708
710
712
714
716
718
720
722
724
726
selection, the degree of multiple mating or the adult sex ratio. Or do they depend on taxonspecific traits that might not be not readily applied across taxa. For example, that females
prefer males providing care as in fish (but see Tallamy 2000 for evidence of a similar
preference in insects). Or are these outstanding general factors that are missing from the
current theoretical models?
One of the exciting areas we see is the attempt to empirically test the model we have outlined.
This is a challenging task, however, because – as we have noted – we are making predictions
about evolutionary responses and these do not necessarily have to be mimicked when making
comparisons between parents in different environments (i.e. adaptive plasticity). This places
limitations on the extent to which behavioural studies of extant variation in care can be used
to test models. Cross-species studies are more powerful and there have been a few attempts to
conduct phylogenetic tests to see whether transitions between different major patterns of
parental care can be predicted based on the level of sexual selection (e.g. Gonzalez-Voyer et
al. 2008; Olsson et al 2009). Again, however, because phylogenetic studies report
correlations we need to analyses them cautiously as species often differ in several respects.
Ideally, we would like to run experimental evolution studies akin to recent ones in which
lines are assigned to monogamy and polyandry treatments and trait evolution is then
monitored (e.g. Simmons & Garcia-Gonzalez 2008), but with a focus on parental care. One
could, for example, establish selection lines in which the adult sex ratio is made either
heavily male-biased or female-biased each generation. The obvious problem is, however,
identifying suitable species with a sufficiently short generation time and readily measurable
care. More pragmatic, perhaps, is to rely on ‘natural experiments’ by investigating
differences among populations in key factors (e.g. the adult sex ratio or level of paternity)
and then testing whether the predicted pattern of parental care occurs. Although we can never
be sure that the factors in question are driving the differences this does at least provide one
way forward. As an interesting example, we recently noted a dataset that focused on variance
in male and female mating success in different human populations (Brown et al. 2009).
Populations were classified as polygynous, serially monogamous or monogamous. The level
of male care per female is presumably lowest in the former and highest in the later. If the
mean reproductive success of males and females is assumed to be correctly estimated then,
given the Fisher condition, the inverse of this ratio is a measure of the ASR. There was a
significant difference in the ASR between the three social systems (Fig 6.4).
<Figure 6.4 about here>
Finally, we end with a strange piece of natural history that reminds us that nature always
throws up challenges when we try to come up with grand theories unified by common
themes. We have made much of the Fisher condition – one mother and one father. Even this
‘fact’ is, however, violated in at least one diploid sexually reproducing organism – and a
primate no less. Marmosets Callithrix kuhlii live in polyandrous groups with two males
caring for offspring. Fascinatingly, they are sometimes chimeric with siblings exchanging
cells in utero. As a result, one offspring has two fathers and, more importantly, from an
evolutionary perspective, the germ line tissue is also chimeric so that either father’s genes
might be transmitted to the next generation (Ross et al. 2007). Amazingly, the more chimeric
offspring were tissues that males could detect, the more they were carried by males,
presumably because offspring scent profiles match both fathers. This is the kind of wonderful
quirk of natural history that reminds us how easily the limits of our own imaginations can
sometimes constrain our ability to solve problems. That said, we are still confident that
18
728
paternity, sexual selection and adult sex ratios will continue to predict parents of parental
care.
730
References
732
Alonzo, S.H. and Heckman, K.L. (2010). The unexpected but understandable dynamics of
mating, paternity and paternal care in the ocellated wrasse. Proceedings of the Royal
Society, Lond. Series B 277, 115-122.
734
Arnqvist, G. (1998). Comparative evidence for the evolution of genitalia by sexual selection.
Nature 393, 784-786.
736
Beck, C.W. (1998). Mode of fertilization and parental care in anurans. Animal Behaviour 55,
439-449.
738
Bickford, D. (2002). Animal behaviour - Male parenting of New Guinea froglets. Nature 418,
601-602.
740
Brown G.R., Laland, K.N. and Borgerhoff-Mulder M. (2010). Bateman’s principle and
human sex roles. Trends in Ecology and Evolution 24, 297-304.
742
Bulmer, M.G. and Parker, G.A. (2002). The evolution of anisogamy: a game-theoretic
approach. Proceedings of the Royal Society, Lond. Series B 269, 2381-2388.
744
746
Carroll, S.P. and Corneli, P.S. (1985). Divergence in male mating tactics between two
populations of the soapberry bug: II. Genetic change and the evolution of a plastic
reaction norm in a variable social environment. Behavioural Ecology 6, 46-56.
748
Clutton-Brock, T.H. and Parker, G.A. (1992). Potential reproductive rates and the operation
of sexual selection. Quarterly Review of Biology 67, 437-456.
750
Cockburn, A. (2006). Prevalence of different modes of parental care in birds. Proceedings of
the Royal Society, Lond. Series B 273, 1375-1383.
752
Cummins, J. (2009). Sperm motility and energetics. In T.R. Birkhead, D.J. Hosken, D.J. and
S. Pitnick, eds. Sperm biology: an evolutionary perspective, pp .XX-XX. Academic
Press, New York.
754
Dawkins, R. (1989). The Selfish Gene, 2nd ed. Oxford University Press, Oxford.
756
De Fraipont, M., Clobert, J. and Barbault, R. (1996). The evolution of oviparity with egg
guarding and viviparity in lizards and snakes: a phylogenetic analysis. Evolution 50,
391-400.
758
Donald, P.F. (2007). Adult sex ratios in wild bird populations. Ibis 149, 671-692.
760
Emlen, S.T. and Oring, L.W. (1977). Ecology, sexual selection, and the evolution of mating
systems. Science 197, 215-223.
762
Fitze, P.S. and Le Galliard, J.-F. (2008). Operational sex ratio, sexual conflict and the
intensity of sexual selection. Ecology Letters 11, 432-439.
19
764
Fromhage, L., Elgar, M.A. and Schneider, J.M. (2005). Faithful without care: the evolution of
monogyny. Evolution 59, 1400-1405.
766
Fromhage, L., McNamara, J.M. and Houston, A.I. (2008). A model for the evolutionary
maintenance of monogyny in spiders. Journal of Theoretical Biology 250, 524-531.
768
Gittleman, J.L. (1981). The phylogeny of parental care in fishes. Animal Behaviour 29, 936941.
770
Gonzalez-Voyer, A., Fitzpatrick, J.L. and Kolm, N. (2008). Sexual selection determines
parental care patterns in cichlid fishes. Evolution 62, 2015-2026.
772
774
Goodwin, N.B., Balshine-Earn, S. and Reynolds, J.D. (1998). Evolutionary transitions in
parental care in cichlid fish. Proceedings of the Royal Society, Lond. Series B 265,
2265-2272.
776
Gray, S.M., Dill, L.M. and McKinnon, J.S. (2007). Cuckoldry incites cannibalism: male fish
turn to cannibalism when perceived certainty of paternity decreases. . American
Naturalist 169, 258-263.
778
Gross, M.R. and Sargent, R.C. (1985). The evolution of male and female parental care in
fishes. American Zoologist 25, 807-822.
780
Harrison, F., Barta, Z., Cuthill, I. and Székely T. (2009). How is sexual conflict over parental
care resolved? A meta-analysis. Journal of Evolutionary Biology 22, 1800-1812.
782
Hosken, D.J. and Kunz, T.H. (2009). But is it male lactation or not? Trends in Ecology and
Evolution 24, 355.
784
786
Houston, A.I. and McNamara, J.M. (2002). A self-consistent approach to paternity and
parental effort. Philosophical Transactions of the Royal Society of London Series B
357, 351-362.
788
Inada, K., Kitade, O. and Morino H. (2011). Paternity analysis in an egg-carrying aquatic
insect Appasus major (Hemiptera: Belostomatidae) using microsatellite DNA
markers. Entomological Science 14, 43-48.
790
Jennions, M.D. and Polakow, D.A. (2001). The effect of partial brood loss on male desertion
in a cichlid fish: an experimental test. Behavioural Ecology 12, 84-92.
792
Johnstone, R. A. and Hinde, C.A. (2006). Negotiation over offspring care: how should
parents respond to each other's efforts? Behavioural Ecology 17, 818-827.
794
Klett, V. and Meyer, A. (2002). What, if anything, is a Tilapia? Mitochondrial ND2
phylogeny of Tilapiines and the evolution of parental care systems in the African
cichlid fishes. Molecular Biology and Evolution 19, 865-883.
796
Klug, H., Heuschele, J., Jennions, M.D. and Kokko, H. (2010). The mismeasurement of
sexual selection. Journal of Evolutionary Biology 23, 447-262.
798
Kokko, H. (1999). Cuckoldry and the stability of biparental care. Ecology Letters 2, 247-255.
20
800
Kokko, H. and Jennions, M.D. (2008). Parental investment, sexual selection and sex ratios.
Journal of Evolutionary Biology 21, 919-948.
802
Kokko, H. and Jennions, M.D. (2010). Behavioural ecology: ways to raise tadpoles. Nature
464, 990-991.
804
806
808
Kokko, H. and Johnstone, R. A. (2002). Why is mutual mate choice not the norm?
Operational sex ratios, sex roles, and the evolution of sexually dimorphic and
monomorphic signalling. Philosophical Transactions of the Royal Society of
London Series B 357, 319-330.
Kokko, H. and Rankin, D.J. (2006). Lonely hearts or sex in the city? Density-dependent
effects in mating systems. Philosophical Transactions of the Royal Society of
London Series B 361, 319-334.
810
Kunz, T.H. and Hosken, D.J. (2009). Male lactation: why, why not and is it care Trends in
Ecology and Evolution 24, 80-85.
812
Kvarnemo, C. (2006). Evolution and maintenance of male care: is increased paternity a
neglected benefit of care? Behavioural Ecology 17, 144-148.
814
Lack, D. (1947). The significance of clutch size. Ibis 89, 302-352.
816
Lehtonen, J. and Kokko, H. (2011). Two roads to two sexes: unifying gamete competition
and gamete limitation in a single model of anisogamy evolution. Behavioural
Ecology and Sociobiology 65, 445-459.
818
Lessells, C.M. (2006). The evolutionary outcome of sexual conflict. Philosophical
Transactions of the Royal Society of London Series B 361, 301-317.
820
Lessells, C.M., Snook, R.R. and Hosken, D.J. (2009). The evolutionary origin and
maintenance of sperm: selection for a small, motile gamete mating type. In T.R.
Birkhead, D.J. Hosken, D.J. and S. Pitnick, eds. Sperm biology: an evolutionary
perspective, pp .43-67. Academic Press, New York.
822
824
826
828
Levitan, D.R. (2006). The relationship between egg size and fertilization success in
broadcast-spawning marine invertebrates. Integrative and Comparative Biology 46,
298-311.
Levitan, D.R. (2010). Sexual selection in external fertilizers. D.F. Westneat and C.W. Fox,
eds. Evolutionary Behavioral Ecology, pp. 365-378. Oxford University Press,
Oxford.
830
Liker, A. and Székely, T. (2005). Mortality costs of sexual selection and parental care in
natural populations of birds. Evolution 59, 890–897.
832
Lloyd, D.G. (1987). Selection of offspring size at independence and other size-versus-number
strategies. American Naturalist 129, 800-817.
834
Mank, J.E., Promislow, D.E.L. and Avise, J.C. (2005). Phylogenetic perspectives in the
evolution of parental care in ray-finned fishes. Evolution 59, 1570-1578.
21
836
Maynard Smith, J. (1977). Parental investment: a prospective analysis. Animal Behaviour 25,
1–9.
838
840
McGlothlin, J.W., Jawor, J.M. and Ketterson, E.D. (2007). Natural variation in a
testosterone-mediated trade-off between mating effort and parental effort. American
Naturalist 170, 864-875.
842
McNamara, J.M., Székely, T., Webb, J.N. and Houston, A.I. (2000). A dynamic gametheoretic model of parental care. Journal of Theoretical Biology 205, 605-623.
844
Møller, A.P. and Birkhead, T.R. (1993). Certainty of paternity covaries with paternal care in
birds. Behavioural Ecology and Sociobiology 33, 261-268
846
Møller, A.P. and Cuervo, J.J. (2000). The evolution of paternity and paternal care in birds.
Behavioural Ecology 11, 472–485.
848
Moore, S.L. and Wilson, K. (2002). Parasites as a viability cost of sexual selection in natural
populations of mammals. Science 297, 2015-2018.
850
Morton, E.S., Stutchbury, B.J.M. and Chiver, I. (2010). Parental conflict and brood desertion
by females in blue-headed vireos. Behavioural Ecology and Sociobiology 64, 947954.
852
Neff, B.D. (2003). Decisions about parental care in response to perceived paternity. Nature
422, 716-719.
854
Olson, V.A., Liker, A., Freckleton, R.P. and Székely, T. (2008). Parental conflict in birds:
comparative analyses of offspring development, ecology and mating opportunities.
Proceedings of the Royal Society, Lond. Series B 275, 301-307.
856
858
Olson, V.A., Webb, T.J., Freckleton, R.P. and Székely, T. (2009). Are parental care tradeoffs in shorebirds driven by parental investment or sexual selection? Journal of
Evolutionary Biology 22, 672-682.
860
Parker, G.A. (1982). Why are there so many tiny sperm - sperm competition and the
maintenance of 2 sexes. Journal of Theoretical Biology 96, 281-294.
862
Parker, G.A. and Simmons, L.W. (1996). Parental investment and the control of sexual
selection: predicting the direction of sexual competition. Proceedings of the Royal
Society, Lond. Series B 263, 315-321.
864
866
Parker, G.A., Baker, R.R. and Smith, V.G.F. (1972). The origin and evolution of gamete
dimorphism and the male-female phenomenon. Journal of Theoretical Biology 36,
181-198.
868
Pen, I., Uller, T., Feldmeyer, B., Harts, A., While, G.M. and Wapstra, E. (2010). Climatedriven population divergence in sex-determining systems. Nature 468, 436-438.
870
Petrie, M. and Møller A.P. (1991). Laying eggs in others nests - intraspecific brood
parasitism in birds. Trends in Ecology and Evolution 6, 315-320.
22
872
874
876
Pollux, B.J.A., Pires, M.N., Banet, I.A. and Reznick, D.N. (2009). Evolution of placentas in
the fish family Poeciliidae: an empirical study of macroevolution. Ann. Rev. Ecol.
Syst. 40, 271-289.
Pryke, S.R., Rollins, L.A. and Griffith, S.C. (2010). Females use multiple mating and
genetically loaded sperm competition to target compatible genes. Science 329, 964967.
878
Queller, D.C. (1997). Why do females care more than males? Proceedings of the Royal
Society, Lond. Series B 264, 1555-1557.
880
Rankin, D.J., Bargum, K. and Kokko, H. (2007). The tragedy of the commons in evolutionary
biology. Trends in Ecology and Evolution 22, 643-651.
882
Requena, G.S., Nazareth, T.M., Schwertner, C.F., and Machado, G. (2010). First cases of
exclusive paternal care in stink bugs (Hemiptera: Pentatomidae). Zoologia 27, 10181021.
884
886
888
Reynolds, J.D., Goodwin, N.B. and Freckleton, R.P. (2002). Evolutionary transitions in
parental care and live bearing in vertebrates. Philosophical Transactions of the
Royal Society of London Series B 357, 269–281.
890
Ross, C.N., French, J.A. and Orti, G. (2007). Germ-line chimerism and paternal care in
marmosets (Callithrix kuhlii). Proceedings of the National Academy of Sciences,
U.S.A. 104, 6278-6282.
892
Royle, N.J., Schuett, W. and Dall, S.R.X. (2010). Behavioral consistency and the resolution
of sexual conflict over parental investment. Behavioural Ecology 21, 1125-1130.
894
896
898
900
902
Schwagmeyer, P.L., St Clair, R.C., Moodie, J.D., Lamey, T.C., Schnell, G.D. and Moodie,
M.N. (1999). Species differences in male parental care in birds: a reexamination of
correlates with paternity. Auk 116, 487-503.
Shellman-Reeve, J. and Reeve, H.K. (2000). Extra-pair paternity as the results of
reproductive transactions between paired mates. Proceedings of the Royal Society,
Lond. Series B 267, 2543-2546.
Simmons, L.W. and Garcia-Gonzalez, F. (2008). Evolutionary reduction in testes size and
competitive fertilization success in response to the experimental removal of sexual
selection in dung beetles. Evolution 62, 2580-2591.
904
Slatyer, R., Mautz, B., Backwell, P.R,Y. and Jennions, M.D. (2011). Estimating genetic
benefits of polyandry from experimental studies: a meta-analysis. Biological
Reviews of the Cambridge Philosophical Society (in press)
906
Smith, C.C. and Fretwell, S.D. (1974). Optimal balance between size and number of
offspring. American Naturalist 108, 499-506.
908
Stiver, K.A. and Alonzo, S.H. (2009). Parental and mating effort: is there necessarily a tradeoff? Ethology 115, 1101-1126.
23
910
Summers, K., McKeon, C.S., Heying, H. (2006). The evolution of parental care and egg size:
a comparative analysis in frogs. Proceedings of the Royal Society, Lond. Series B
273, 687-692.
912
Sutherland, W.J. (1985). Chance can produce a sex difference in variance in mating success
and explain Bateman’s data. Animal Behaviour 33, 1349-1352.
914
916
Szentirmai, I., Székely, T. and Komdeur, J. (2007). Sexual conflict over care: antagonistic
effects of clutch desertion on reproductive success of male and female penduline
tits. Journal of Evolutionary Biology 20, 1739-1744.
918
Tallamy, D.W. (2000). Sexual selection and the evolution of exclusive paternal care in
arthropods. Animal Behaviour 60, 559-567.
Tallamy, D.W. (2005). Egg dumping in insects. Annual Reviews of Entomology 50, 347-370.
920
Thomas, G.H., Szekely, T. and Reynolds, J.D. (2007). Sexual conflict and the evolution of
breeding systems in shorebirds. Advances in the Study of Behaviour 37, 279-342.
922
Trillmich, F. (2010). Parental care, adjustments to conflict and cooperation. In P. Kappeler,
ed. Animal Behaviour: evolution and mechanisms, pp 267-298. Springer, Berlin.
924
Trivers, R. (1972). Parental investment and sexual selection. In B. Campbell, ed. Sexual
selection and the descent of man 1871-1971, pp 139-179. Aldine Press, Chicago.
926
Weir L.K., Grant, J.W.A. and Hutchings, J.A. (2011). The influence of operational sex ratio
on the intensity of competition for mates. American Naturalist 177, 167-176.
928
Wells, K.D. (2007). The Ecology and Behaviour of Amphibians. University of Chicago Press,
Chicago.
930
West, S.A. (2009). Sex Allocation. Princeton University Press, Princeton.
932
Wynne-Edwards, K.E. and Reburn, C.J. (2008). Behavioral endocrinology of mammalian
fatherhood. Trends in Ecology and Evolution 15, 464-468.
934
Yamamura, N. and Tsuji, N. (1993). Parental care as a game. Journal of Evolutionary
Biology 6, 103-127.
24
936
938
940
IMAGES ARE JUST TO GIVE GENERAL IDEA.
942
Figure 6.1 Which is the male and which is the female? (Figure will comprise need a male
and female bird, mammal, reptile and invertebrate)
25
944
946
948
950
952
FIGURE TO BE MADE
954
Figure 6.2. Which is the male and which is the female? (Figure will comprise a male and
956
female fish (mouth brooding cichlid?) poison dart frog and shorebird/wader).
26
958
960
962
964
966
968
970
972
974
976
Figure 6.3 Predictions of evolutionarily stable care effort by females (circles) and males
(squares) when females can either mate with their social mate (’faithful female’) or multiply
(’unfaithful female’), and males can to some extend detect female mating behaviour but this
is either relatively inaccurate (A) or relatively accurate (B). Evolutionary changes occur
along the x axis, behavioural responses along the y axis. Males possess evolutionary
knowledge of average female behaviour, and multiple mating by females selects against male
care (which always declines with the proportion of multiply mating females). Whether
multiple mating is selected against or not in females, however, depends on the fitness
difference between unfaithful and faithful females (vertical differences between circles in the
female fitness graph). Males paired to unfaithful females tend to be suspicious but some of
them remain unsuspicious, and males paired to faithful females tend to be unsuspicious but
some of them become erroneously suspicious. When the accuracy of cues is low (case A),
errors in such judgments are common. Consequently males are not selected to put much
weight on behavioural evidence, and the care effort between the different types of males
remains largely similar at each point in time. The mating system fails to penalize unfaithful
females sufficiently to prevent multiple mating from spreading. The horizontal arrows,
depicting female evolution over time, show that the proportion of multiply mating females
27
978
980
982
approaches 1, and male care is entirely lost from the system. Selection works against female
multiple mating only if females of each generation lose a large amount of male care by being
unfaithful. This occurs in the middle region of (B), indicated by the arrow pointing left, but
not elsewhere. Even (B) will evolve towards more multiple mating if the population starts out
from a point where multiple mating is, for any reason, already very common (rightmost
arrow). Figure is modified from Kokko (1999).
28
984
986
988
990
992
Figure 6.4. Estimated adult sex ratio (ASR) based on the ratio of mean reproductive success
(number of live births) for females and males for populations assigned to three mating system
types. The Fisher condition applies so males and female must have the same total
reproductive output. Hence any difference in mean reproductive success has to be due to a
biased ASR. The data is from Brown et al (2009) who compiled data from 18 countries. Note
that when Brown et al. classified a country as having two mating systems (polygyny and one
other) we classified it as polygyny. Sample sizes are 6,3 and 9.