Maternal effort and male quality in the bank
vole, Clethrionomys glareolus
Tuula A. Oksanen*, Rauno V. Alatalo, Taina J. Horne, Esa Koskela, Johanna Mappes
and Tapio Mappes
Department of Biological and Environmental Science, University of JyvÌskylÌ, PO Box 35, FIN- 40351 JyvÌskylÌ, Finland
Parental investment in reproduction is adjusted according to potential bene¢ts in terms of o¡spring
survival and/or mating success. If male quality a¡ects the reproductive success of a female, then females
mating with high-quality males should invest more in reproduction. Although the subject has been of
general interest, further experimental veri¢cation of the hypothesis is needed. We studied whether female
bank voles (Clethrionomys glareolus) adjusted their maternal e¡ort according to male quality, measured as
mating success. To enable the measurement of maternal e¡ort during nursing separately from male
genetic e¡ects the litters were cross-fostered. Further, the genetic background of male quality was
examined. Male quality did not correlate with litter size or o¡spring size at birth. O¡spring growth was
positively related to food consumption and milk production of mothers. However, these direct measure-
ments of maternal e¡ort were independent of male quality. Male mating success appeared to be
signi¢cantly heritable indicating that there are genetic bene¢ts. Still, females did not adjust maternal
e¡ort according to the genetic quality of their o¡spring. We suggest that female bank voles gain
signi¢cant genetic bene¢ts from mating with high-quality males whereas they cannot improve their
reproductive success by increasing maternal e¡ort.
Keywords: heritability of mating success; life history; male quality; mammals; maternal e¡ort;
o¡spring quality
several confounding e¡ects that may confuse the relation
1. INTRODUCTION
between male quality and female reproductive e¡ort.
Life-history theory predicts that the amount a female These e¡ects may arise particularly in a situation where
invests in reproduction will relate to the potential bene¢ts maternal care is dependent on paternal care (see, for
she gains from the current reproductive attempt example, Burley 1988; de Lope & MÖller 1993) or
(Williams 1966; Gadgil & Bossert 1970). In conditions because of an assortative mating pattern (see, for
where a female can expect a relatively large ¢tness incre- example, RintamÌki et al. 1998). Furthermore, if o¡spring
ment from reproduction, she should invest maximally in inherit the high quality of their fathers (for example, a
o¡spring production. Similarly, when the expected gain is higher growth rate), this may lead to the erroneous inter-
low a female should defer reproduction or invest less until pretation that females increase their reproductive e¡ort
the net gain would be better. when mated to high-quality males.
If male quality a¡ects the reproductive success (i.e. The aim of this study was to examine whether male
number and quality of o¡spring produced) of a female, quality a¡ected litter size and o¡spring size at birth or
then females mated to high-quality males should invest the maternal e¡ort during nursing (milk production and
more in reproduction. There are several studies showing food consumption). Further, the genetic background of
that females mated to preferred males produce more male quality was examined. To enable the measurement
viable o¡spring (reviewed in MÖller & Alatalo 1999). of maternal e¡ort separately from male genetic e¡ects the
However, only a few studies indicate higher maternal o¡spring were cross-fostered between females.
e¡ort in relation to male quality (see, for example, Burley
1988; de Lope & MÖller 1993; Petrie & Williams 1993;
2. MATERIAL AND METHODS
RintamÌki et al. 1998).
Our study species, the bank vole (Clethrionomys glareolus), The animals used in the experiment were second-generation
has a mating system in which males provide no material laboratory-born bank voles (Clethrionomys glareolus) originally
resources to the female or o¡spring. Females discriminate captured in Konnevesi, central Finland (62837' N, 26820' E).
between males according to their social status and prefer The animals were housed in standard mouse cages measuring
dominant mates (Ho¡meyer 1982; Horne & Ylo«nen 43 cm 26 cm 15 cm and maintained in a 16 L : 8 D photo-
1996). In this mating system it is possible to control for period. Wood shavings were used as bedding; water and food
were continuously available. All the females whose maternal
* e¡ort was measured had given birth to one litter before the
Author for correspondence (tuoksane@silmu.cc.jyu.¢).
Proc. R. Soc. Lond. B (1999) 266, 1495^1499 1495 & 1999 The Royal Society
Received 15 March 1999 Accepted 13 April 1999
, 1496 T. Oksanen and others Maternal e¡ort and male quality
experiment. Head width of all individuals was measured before within two days of each other were cross-fostered. Each pup in a
the experiment and used as a factor illustrating parent size. litter was replaced, resulting in a litter in which every pup came
from a di¡erent mother. Litter sizes remained the same and sex
(b) Male quality trials ratios within litters were kept, where possible, the same as in the
Male quality was de¢ned by a set of mating trials in which original litter. When the pups were 20 days old (weaning age)
each male competed against at least three (range 3^8, they were separated from the foster mother.
mean 4.64, s.d. 0.75) other randomly chosen males. A male
pair and a female in oestrus were released in an arena (e) Female food consumption and milk production
(1m 1m) with a Plexiglas cover. All females were injected Basal food consumption of individual females was de¢ned
supracollicularly with b-estradiol 3-benzoate (0.04 mg dissolved before the experiment (when the females were not pregnant or
in 0.1ml linseed oil) two days before the test to ensure their will- lactating) by weighing their food pellets during one week. During
ingness to mate. The female's area consisted of one-half of the lactation the food consumption was measured from parturition to
whole arena. The other half was divided between the two males. the o¡spring age of 15 days. From this age the pups start to feed
At the beginning of the test there was a wooden wall between on pellets beside the mother. Female milk production was
the males and a wire mesh between the female and the males. measured when the pups were 10 days old. The pups were sepa-
The wire mesh enabled opposite sexes to detect each other. rated from the foster mother for 3 h to let them consume the milk
During the ¢rst 10 min animals were allowed to familiarize in their stomachs. After that the foster mother was allowed to
themselves with the arena; then all the dividing walls were nurse the pups for 2 h. The pups were weighed before and after
removed. Observations were made until ejaculation occurred. If the nursing to determine the amount of milk they had received.
no mating behaviour had been seen after 20 min the test was Two measures of milk production were used in the analyses: total
repeated at least 24 h later. The quality of males was determined amount of milk produced (sum of mass increases in a litter) and
as a proportion of opponents against which the male copulated milk produced per pup (average mass increase of a pup in a
successfully with the female. In these trials it was possible to test litter).
male mating success in a situation where both female choice and
male ^ male competition were allowed to work. (f) Heritability of male quality
Because the random selection of male pairs in male quality To ¢nd out whether male mating success had a genetic back-
trials could have led to uneven distribution of opponent quality, ground, male quality trials were also performed (as described
these estimates were corrected with an equation, above) on male o¡spring after their maturation. The heritability
of male mating success was estimated by regressing the quality
Q FnF /FnF (1 ÿ S)ns , value of one randomly chosen son against the value of his father.
The heritability was estimated as twice the slope and the stan-
where Q is the corrected quality estimate of a male; F is the mean dard error was obtained by doubling the standard error of the
quality value of the opponents who failed to copulate; nF is the regression (Ro¡ 1997).
number of the unsuccessful opponents; S is the mean quality
value of the opponents who copulated successfully and nS is the
3. RESULTS
number of successful opponents (R. V. Alatalo, J. Ho«glund,
A. Lundberg and P. Rintama«ki, unpublished observations). (a) O¡spring size and number at birth
Paternal quality had no signi¢cant e¡ect on litter size
(c) Mating or the mean o¡spring size at birth (¢gure 1a,b; table 1).
A new group of females (di¡erent from those used in male- Hence, paternal quality does not seem to have any posi-
quality trials) was paired with males of known quality. A tive impact on o¡spring number or size. Supported by
randomly chosen male and a female were placed together in a the large sample sizes the e¡ects were, if anything, nega-
cage for a week. These females were not manipulated with any tive. O¡spring size was a¡ected by maternal size, which
hormones and the exact time of mating was dependent on their had a positive e¡ect on mean pup mass and head width
natural oestrus cycle, which is about four days. After a week the (table 1). Paternal size did not explain o¡spring mass or
male and female were separated and the mass of females was head width (table 1). Male quality did not correlate with
followed up to detect the pregnancy. If mating did not turn out male size measured as head width (r 0.17, n 85,
to be successful it was repeated after three weeks. After the ¢rst p 0.111).
parturition the mother was mated again in postpartum oestrus
with a randomly chosen male. This second mating was carried (b) Maternal e¡ort during nursing
out because being pregnant while lactating pups is a natural Female food consumption during lactation correlated
state for females in the wild (Bronson 1989). signi¢cantly with basal food consumption (r 0.54, n 43,
p50.001) and the number of pups (r 0.60, n 44,
(d) Pup measurements and cross-fostering p50.001). The trade-o¡ between litter size and mean mass
Pregnant females were checked twice a day to detect the of o¡spring in a litter at birth was also clear (r 70.55,
parturition. Immediately after birth the pups were weighed and n 75, p50.001). Because of these dependencies, female
their head width was measured with a microscope. At the ages basal food consumption and litter size were used as
of ten and 20 days the pups were weighed again and their head controlling factors in partial correlations (rP) below.
width was measured with a dial caliper. All masses were The amount of food females consumed during lactation
measured with electronic scales. Litter means of o¡spring size did not correlate with the quality of their mates (control-
were used in all the analyses. At birth the pups were also sexed ling for basal food consumption and litter size: rP 70.09,
and marked individually to identify them later. To randomize n 39, p 0.578). However, food consumption rate had a
the genetic in£uence of parents on o¡spring growth, litters born signi¢cant e¡ect on mean o¡spring mass at the ages of ten
Proc. R. Soc. Lond. B (1999)
