Abstract We examined the relationship between lacta-
tion performance and infant growth in a captive popula-
tion of common marmosets (Callithrix jacchus) that var-
ied in both maternal and litter size. Though common
marmosets display a typical primate pattern of dilute
milk and relatively slow infant growth rates (factors as-
sociated with low daily lactation investment and minimal
maternal size effects), we hypothesized that the marmo-
set?s small body size would make lactation investment
more sensitive to maternal size than is true for larger-
bodied primates. Smaller mothers rearing twins had low-
er milk fat, lower gross energy in milk samples collected
in mid to late lactation and lower nursing-bout frequen-
cies than did large mothers nursing twins. Lactation per-
formance and maternal behavior did not differ between
large and small mothers when rearing singletons, with a
single exception: small mothers had a lower gross energy
in mid-lactation milk samples. Relative growth rates in
twins but not singletons were affected by maternal size,
such that small mothers supported more growth per in-
fant when rearing singletons while large mothers sup-
ported more growth per infant when rearing twins.
Among the larger mothers, only, older mothers supported
somewhat, though not significantly, less growth per 
infant, regardless of litter size. Twin infants of small
mothers appeared to respond to below-optimal levels of
milk yield by initiating maternal carrying less often. The
relative energy intake of mothers was similar regardless
of litter or maternal size. Small mothers rearing twins
drew more heavily on reserves, reflected in a linear
weight loss during lactation; however, the reserves
drawn upon were inadequate to meet the lactation de-
mand, resulting in lower milk energy output. In addition,
small mothers rearing twins were more likely to be ill
and less likely to be fertile in the year following lactation
than were large mothers of twins or mothers of any size
rearing singletons.
Keywords Reproductive investment ? Lactation ? 
Marmoset ? Litter size ? Maternal size
Introduction
Numerous studies have addressed the relationship be-
tween intraspecific variation in maternal size and repro-
ductive investment in mammals. While some of these
studies found no relationships (Hansson 1992; Hare and
Murie 1992; Morris 1996; for reviews see Mattingly and
McClure 1982; Festa-Bianchet et al. 1998), others found
positive relationships between maternal size and mea-
sures of reproductive investment such as litter size, num-
ber of young weaned, and growth rates of young within a
given litter size (e.g., house mice: Knight et al. 1986;
cotton rats: Mattingly and McClure 1985; ground squir-
rels: Dobson and Michener 1995; Neuhaus 2000; big-
horn sheep: Festa-Bianchet et al. 1998; gray seals: 
Iverson et al. 1993). When positive relationships are
found, they are often only seen under certain circum-
stances, leading Mattingly and McClure (1985) to sug-
gest that ?body mass of the mother (or some associated
factor such as fat content) becomes increasingly impor-
tant as food availability decreases and reproductive costs
increase.? Therefore, body mass of the mother becomes
Communicated by P. Kappeler
S.D. Tardif (?) ? D.G. Layne
Department of Biological Sciences, Kent State University, Kent,
OH 44242-0001, USA
M. Power
American College of Obstetrics and Gynecology, 
Washington, DC 20024, USA
M. Power ? O.T. Oftedal
Department of Zoological Research, National Zoological Park,
Washington, DC 20008, USA
R.A. Power
Department of Physiology, College of Veterinary Medicine, 
University of Florida, Gainesville, FL 32610-0144, USA
Present address:
S.D. Tardif, Southwest Regional Primate Research Center, 
P.O. Box 760549, San Antonio, TX 78245-0549, USA
e-mail: stardif@sfbr.org, Fax: +1-210-2589883
Behav Ecol Sociobiol (2001) 51:17?25
DOI 10.1007/s002650100400
O R I G I N A L  A RT I C L E
Suzette D. Tardif ? Michael Power ? Olav T. Oftedal 
Rachel A. Power ? Donna G. Layne
Lactation, maternal behavior and infant growth 
in common marmoset monkeys (Callithrix jacchus): 
effects of maternal size and litter size
Received: 22 November 2000 / Revised: 2 July 2001 / Accepted: 4 July 2001 / Published online: 12 September 2001
? Springer-Verlag 2001
an important factor to consider in models of reproduction
allocation (Smith and Fretwell 1974; Reznick 1985;
Hochachka 1992).
In female mammals, lactation is the most energeti-
cally expensive aspect of reproductive investment 
(Gittleman and Thompson 1988). The extent to which
intraspecific variation in maternal size affects lactation
investment is likely to vary among mammalian taxa. 
Effects might be expected to be pronounced in small-
bodied species that generally have a high energetic in-
vestment in lactation relative to basal metabolic energy
needs because they support relatively fast growth in
large litters. In such cases, decreased energy availability
(e.g., food restriction) or increased energy demands (e.g.,
decreased temperature, increased litter size) can result in
the enhanced reproductive performance of larger relative
to smaller mothers (mice: Knight et al. 1986; cotton rats:
Mattingly and McClure 1982, 1985; ground squirrels:
Neuhaus 2000).
In larger-bodied species, maternal size has also been
linked to intraspecific variation in lactation investment
under conditions of decreased energy availability or in-
creased energy demands. However, in large-bodied spe-
cies, such effects are most marked in those taxonomic
groups that are capital breeders (Jonsson 1997), who 
pay the costs of lactation investment by drawing 
heavily from maternal energy reserves (domestic sheep:
Thompson 1983; gray seals: Iverson et al. 1993; bighorn
sheep: Festa-Bianchet et al. 1998) or in species produc-
ing large litters (e.g., house cat: Deag et al. 1987).
Primates tend to have a low daily energy investment
in lactation, producing low yields of dilute milk that sup-
port relatively slow growth rates of a single infant over
long periods of time (Oftedal and Iverson 1995). Oftedal
(1984) estimated relative milk energy yields at peak lac-
tation in primates to be significantly lower than those of
other mammals, including species producing many
young (e.g., pig, dog, rat) and large-bodied species pro-
ducing singletons (e.g., deer, sheep, horse). Such a pat-
tern suggests that effects of maternal size on lactation
might be expected to be minimal in primates. Data on
large-bodied anthropoid primates, such as baboons and
humans, support this expectation. Captive baboons, for
example, increase daily energy intake 27% during lacta-
tion (Roberts et al. 1985) ? compared to increases of
60?130% during lactation typical of rodents (Mattingly
and McClure 1982) ? and they experience only a 
modest weight loss of around 7% (Roberts et al. 1985;
Bercovitch 1987). Captive baboon mothers can maintain
normal milk yields while on a diet restricting them to
80% of expected energy, with increased use of energy re-
serves (weight loss) and decreased milk yields not ob-
served until 60% restriction (Roberts et al. 1985). In hu-
mans, reviews of numerous studies support the conten-
tion that maternal condition is either marginally related
or not related to lactation performance (Rasmussen
1992; McNamara 1995).
Common marmosets (Callithrix jacchus; family Calli-
trichidae) represent a particularly interesting species in
which to examine intraspecific variation in lactation in-
vestment given that marmosets are adapted to a life his-
tory that includes the typical primate pattern of dilute
milk and relatively slow infant growth rates combined
with a relatively small body size. Lee (1987) hypothe-
sized that, while nutritional effects upon reproduction
were limited in large-bodied anthropoid primates, small-
er-bodied primates ?would be more severely affected by
malnourishment.? Likewise, one might also expect a
more marked effect of maternal size on lactation invest-
ment in smaller-bodied anthropoids. Adult, feral marmo-
sets typically weigh 300?350 g; reproductive females in
one feral population had an average weight of 344.6 g
(Araujo et al. 2000). They usually produce twins (litter
size ranging from one to four in captivity) and can pro-
duce young and lactate as often as every 5.5 months.
This reproductive pattern gives marmosets a relatively
high reproductive investment (a maximum of approxi-
mately four young per year) compared to other primates
but a relatively low reproductive investment compared to
other mammals of a similar body size, such as rats or
squirrels. Lactation lasts for around 75 days; however,
weaning to solids begins at around day 30. Nievergelt
and Martin (1999) report that captive common marmoset
females increase daily energy intake during lactation by
100% and lose a modest amount of body weight (~8%)
when rearing twins, a pattern very similar to that ob-
served in similarly sized rodents.
We examined the relationship between lactation per-
formance and infant growth in a captive population of
common marmosets that varied in both maternal and lit-
ter size. Given the information available on these rela-
tionships in other mammals, we hypothesized the follow-
ing:
 Optimal reproductive output will be associated with
large females producing the most common litter size,
i.e., twins. Optimal lactation performance and maxi-
mal neonatal growth for a species generally occurs at
a modal litter size typical for feral representatives of
that species, with poorer growth and survival in 
both litters that are unusually large (Mattingly and
McClure 1982; Thompson 1983; Knight et al. 1986;
Konig et al. 1988; Leamy 1992; Arnbom et al. 1997)
or unusually small (Knight et al. 1986; Hammond 
et al. 1996). Given these findings, we propose that
marmosets will be adapted to produce a maximal in-
dividual growth rate in twins.
 Lactation performance, in terms of milk composition
and milk yields, will be related to maternal size at
larger but not necessarily at smaller litter sizes.
 Females will meet the energy demands of lactation
through both increased food intake and reliance on
stored energy. We therefore expect both increased
food intake and weight loss (Nievergelt and Martin
1999). The extent to which females rely on each may
depend upon the relative amount of reserve and the
relative investment. Small females might not be able
to meet the demand with just food intake as energy
18
assimilation is dependent upon body size (Hammond
et al. 1996). Therefore when nursing the largest litter
size (twins), small mothers are expected to have a
greater loss of reserves than large mothers, a lower
milk yield than large mothers, or both. This hypothe-
sis stems from the expectation that there will be limits
past which additional reserves will not be devoted to
reproduction. Studies of both small-bodied mammals
(hamsters and mice: Bronson and Marsteller 1985;
Schneider and Wade 1989; ground squirrels: Neuhaus
2000; rats: Crnic 1980; Mattingly and McClure 1985;
Rogowitz 1998) as well as large-bodied species 
(cattle: Oldham and Friggens 1989; elephant seals:
Arnbom et al. 1997; baboons: Roberts et al. 1985; hu-
mans: Dewey 1997) suggest that maternal energy re-
serves in mammals may be protected such that when
certain levels of body nutrient mobilization are ex-
ceeded, milk yields will decline (Steingrimsdotter 
et al. 1980; Roberts et al. 1985).
Methods
Study population
We collected data on lactation performance, maternal size and in-
fant growth for 36 litters produced by 15 different dams between
1995 and 1999. Dams and infants lived in family groups that in-
cluded the sire and zero to four older offspring. These litters were
divided into the following four categories for analyses:
a. Small maternal size/singleton: mothers whose average body
weight was less than the population average (smaller than 
average) rearing singletons
b. Small maternal size/twins: smaller-than-average mothers rear-
ing twins
c. Large maternal size/singletons: larger-than-average mothers
rearing singletons
d. Large maternal size/twins: larger-than-average mothers rearing
twins
Of the 12 singletons reared, 50% were the result of singleton de-
liveries and 50% were the result of twin or triplet deliveries in
which infants died or were removed within 3 days of birth. Infants
were removed if they were repeatedly found off carriers during the
first 3 days. Of the 24 twin litters reared, 83% were twin deliveries
and 17% were the result of triplet deliveries in which an infant
died or was removed within 3 days of birth. All infants remaining
after a removal were successfully reared to weaning. The percent-
age of viable infants that were male did not differ between smaller
(60.7%) and larger (56.2%) mothers.
Table 1 presents a basic description of the four maternal
size/litter size groups. The following parameters of lactation per-
formance, infant growth, and maternal size were assessed. Not all
parameters were measured in all lactation periods; Table 1 pro-
vides sample sizes for each variable.
Lactation performance
Milk samples were collected at days 21, 32, and 45 (?2 days). In
brief, dams were separated from their infants for 3?4 h prior to col-
lection, then sedated with ketamine hydrochloride (10?15 mg/kg)
and given 2 IU of oxytocin (intramuscularly) prior to manual milk
collection. The composition of milk samples was determined at the
Nutrition Laboratory of the Smithsonian National Zoological Park
using standard methods (Oftedal and Iverson 1995) that have been
used by this laboratory to assay other primate milks and have been
validated using both cow and human milk.
Milk energy output of dams during lactation was estimated us-
ing two methods. A labeled-water dilution technique was used to
estimate milk output per infant during the 10 days before infants
typically began to consume solid food, i.e., days 21?30. This was
assumed to represent a time period of highest lactation demand,
given that it represents the largest infant weight and weight gain
that are supported by milk alone. In summary, infants were given a
known dose of deuterium-labeled distilled water through gavage,
blood samples (0.05?0.10 ml) were collected at two to three time
points during the following week and assayed for deuterium-
labeled water concentration. Preliminary results indicated that, by
day 8 after gavage, the concentration of deuterium-labeled water
in samples was too small to measure reliably. The dilution of the
deuterium-labeled water, combined with the known water content
of milk samples collected on days 21 and 30, allowed for esti-
mates of milk output in g/day (Oftedal and Iverson 1987). The
milk composition estimates were then used to estimate gross ener-
gy output per day.
A second measure of milk energy output (gross energy in mid-
to late-lactation milk samples) was calculated as gross energy per
gram of milk (estimated from the protein, fat, and carbohydrate
content of samples from a given lactation period)?total weight of
milk samples collected at days 21, 30, and 45.
19
Table 1 Description of study population and data collected
Low-body-weight Low-body-weight High-body-weight High-body-weight
dam/singleton dam/twins dam/singleton dam/twins
Number of litters 6 11 6 13
Number of dams 4 7 5 6
Mean?SE dam weight (g) 337.2?4.3 330.6?5.5 375.0?5.5 381.8?5.2
Mean?SE dam age (years) 3.65?0.50 3.71?0.39 5.32?0.70 4.02?0.22
Sample sizes
Dam weights and weight change 6 11 6 13
Individual infant weights and growth rates 6 22 6 26
Litter growth rates 6 11 6 13
Dam body composition 3 4 2 4
Milk composition (% fat, protein, dry matter) 4 5 3 4
Milk energy output (per infant) 4 5 3 3
Milk energy in mid-lactation samples 4 6 4 3
Infant care behaviors (nursing, retrieval, harassment) 6 20 6 22
Parental food intake 4 7 4 3
Dam morbidity/fertility 6 11 6 13
Behaviors of the mother and infant(s) were documented during
30-min observation sessions occurring three to four times per
week from birth to day 60. During these sessions, all occurrences
of the following behaviors were recorded.
a. Nursing: defined by the position of the infant on the mother
such that the infant?s head was at the location of the nipple. On
some occasions, rooting and suckling behavior could be ob-
served, but on many occasions, the location of the mother?s
arm obscured the infant?s face. Therefore, the total time scored
as ?nursing? is probably an overestimate of the amount of time
the infants were actually suckling. However, all instances of
infants being located in this axial position do appear to be as-
sociated with some suckling. For this reason, the frequency of
nursing bouts (number of times the infant assumed this axial
position/number of observation sessions) was used as an esti-
mator of suckling frequency, rather than the total time spent in
the axial position
b. Retrieval: defined as actions by the mother that directed the in-
fant off another carrier and onto the mother (see Tardif et al.
1990). In this species, individuals other than the mother are re-
sponsible for the majority of infant transport. Therefore, the
frequency of retrieval of infants from other carriers provides an
estimate of maternal attraction to infants. The frequency of
maternal transport of infants initiated by mothers toward in-
fants was calculated from these data.
c. Harassment: defined as actions by the mother directed at dis-
lodging the infant from her, including biting, grabbing, and
rubbing the infant on branches or the sides of the cage. The
frequency of harassment of transported infants within 1 min of
transport initiation was calculated from these data.
Maternal size
Maternal size was assessed through weekly weighing of the dams.
Changes in maternal weight from days 1 to 80 postpartum were
determined by linear regression of maternal weight on day post-
partum. The slope of this regression (if P<0.15) was then used as
an estimator of average weight change across lactation. If the re-
gression had a P>0.15, the slope was estimated as 0. A probability
of P=0.15 was used in order to include weight changes that were
likely to be biologically significant but were not strictly linear
across the entire lactation period. In addition, total body water rel-
ative to body weight was determined at days 45 and 75 postpartum
through labeled-water dilution. Details on this technique are de-
scribed in Power et al. (2001). Comparison of body weights with
estimated total body water suggests that body weight is a good es-
timator of both relative lean and fat mass in this captive popula-
tion of marmosets (Power et al. 2001).
Food intake
Food intake in each group (dam+sire+weaning infants) was mea-
sured for a 48-h period at approximately 35?36 days postpartum,
by weighing all food presented and all food remaining. All food
weights were converted to dry weights and the amount of food
consumed per kilogram body weight of the pair was calculated ?
consumption by infants was assumed to be negligible relative to
adult consumption given that most infants did not begin to eat sol-
id food until day 30?35. Studies by Nievergelt and Martin (1999)
on common marmosets and Sanchez et al. (1999) on another calli-
trichid primate, the cotton-top tamarin, indicate that energy intake
by males does not increase during the mate?s lactation period (and
may in some cases decrease) while maternal energy intake is in-
creased. Therefore, differences between pairs in relative food con-
sumption can reasonably be conservatively interpreted as differ-
ences in maternal food intake.
Infant growth
Infant growth was assessed by repeated weighing of the infants.
Infants were typically weighed once a week, to the nearest 1.0 g,
using methods previously described (Tardif et al. 1998). The num-
ber of weight samples varied among subjects. Growth rates for
common marmosets are linear during the time period of interest
(days 0?42; Hearn and Lunn 1975; mean linear regression coeffi-
cient for 16 subjects in the present study with an average of eight
weights taken between days 1?42=0.952); therefore, growth rate
estimates were made simply by comparing weight at day 1 and
day 42 (?2 days). The relative growth rate was calculated as
[(W2?W1)/W1]/T, where W2=weight at end of period, W1=weight at
beginning of period, and T=number of days within period. The ab-
solute litter growth rate for the entire time period (1?42) was cal-
culated as (W2?W1)/T.
Subjects were fed a purified agar-based diet that provided
4.33 kcal/g dry weight and a canned commercial diet that provided
approximately 4.84 kcal/g dry weight. Some subjects received on-
ly the purified diet while others received both diets. Milk energy
outputs, body composition measures, and food intake estimates
were made on subjects receiving only the purified diet.
The relationships of lactation performance, maternal behavior,
maternal weight changes, and infant growth with the relative ma-
ternal size/litter size category of the dam were analyzed through a
series of analyses of variance, with maternal size/litter size catego-
ry as the primary factor. The effect of infant weight and parity on
each parameter was determined and, if either was an important
factor, the main effect of maternal size/litter size category was de-
termined in analyses of covariance, with infant weight at the be-
ginning of the time period in question (day 0 or day 21) or parity
as a covariate. If the ANOVA suggested a difference between ma-
ternal size/litter size groups (a probability of <0.05), post hoc ana-
lyses were performed (Scheffe?s test) and these results were com-
pared with appropriate within-dam differences, when there were
single dams who had been sampled across the categories of inter-
est. If parametric analyses were inappropriate because the data
were not normally distributed (weight change slope and some be-
havioral parameters), non-parametric tests were used. Sample siz-
es for all variables in all tests are provided in Table 1. All statisti-
cal analyses were performed in SPSS, version 9.0, and results
from all among-subject comparisons with a P<0.10 are described,
with the specific probability values given.
Results
The mean postpartum weight of dams (averaged over
days 0?21) was 357.7 g and this reflected non-pregnant
weights. The average difference between non-pregnant
weight and postpartum weight was ?2.06 g and there was
no difference between maternal size/litter size categories
in pre- versus postpartum weight change (F=0.387,
P=0.763). Smaller-than-average dams had a mean weight
of 332.9 g, while larger-than-average dams had a mean
weight of 379.6 g. The average weight of dams rearing
singletons was virtually identical to that of dams rearing
twins (356.1 vs 358.3 g; t=1.24, P=0.273). Of the 15 dams,
10 were represented in only one maternal size/litter size
category while 5 were represented in two to three differ-
ent categories; it is within-subject differences for these
dams that are compared to the results across all litters.
The average age of dams at delivery was 4.08 years.
Large dams rearing singletons were older, although not
significanty, than dams in the other groups (P=0.066; see
Table 1) and small dams rearing twins were of lower par-
ity that large dams rearing twins, but also not significant-
ly (Mann-Whitney=41.0, P=0.069).
20
Lactation performance
Figure 1 illustrates the mean values for milk fat (%) for
the four maternal size/litter size groups. Milk fat differed
between groups (F=3.78, P=0.038) and these differences
were not due to infant birth size or parity. Specifically,
small dams rearing twins produced milk that was lower
in fat than that from large dams rearing singletons
(P=0.009) or twins (P=0.017). It was also lower in fat
than milk from small dams rearing singletons, but this
result was not statistically significant (P=0.081). Results
within dams (n=2) were similar: small dams had milk fat
percentages that were roughly twice as high when rear-
ing singletons as when rearing twins (4.34 vs 2.22% and
3.53 vs 1.18%).
Gross milk energy output per infant (days 21?30) av-
eraged 10.00?0.82 (SE) kcal/day for singletons and
9.88?0.87 kcal/day for twins, with no difference between
litter sizes. Gross energy output per infant was correlated
with infant weight (r=0.736, P=0.005) and with relative
individual preweaning growth rates (r=0.751, P=0.004)
for twins but not for singletons (weight: r=0.314,
P=0.224; relative growth rate: r=?0.068, P=0.437).
When infant size was considered as a covariate, there
was no relationship between maternal size and gross
milk energy output per infant.
Gross energy in mid- to late-lactation milk samples
was positively correlated with relative litter growth rate
in twins (r=0.829, P=0.003) and was negatively correlat-
ed, though not significantly, with relative growth rate in
singletons (r=?0.541, P=0.083) during the preweaning
period. As illustrated in Fig. 2, gross energy in mid- to
late-lactation samples differed between maternal size/litter
size categories (F=4.682, P=0.02). Small dams had low-
er gross energy than did large dams, regardless of litter
size. Comparisons within dams suggest a differential re-
sponse to litter size, dependent upon the size of the
mother. For one small dam, the gross energy in milk
samples was 0.41 kcal when rearing a singleton but only
0.22 when rearing twins. In contrast, large dams who
reared both twins and singletons (n=2) had gross energy
in milk samples that was over twice as high when rearing
twins as when rearing singletons (1.26 vs 0.54 and 1.72
vs 0.83).
Nursing-bout frequency was related to maternal
size/litter size categories during both the preweaning
(0?21 days) and the weaning (21?42 days) period in an
identical fashion (Table 2). Nursing-bout frequencies for
singletons did not differ between small and large dams.
However, for twins, larger dams had a higher mean fre-
quency of nursing bouts (day 0?21, F=3.577, P=0.026;
day 21?42, F= 4.6065, P=0.010). This difference was
not related to infant weight or parity. Comparisons with-
in dams revealed similar results. The nursing-bout fre-
quencies during weaning of dams rearing twins (n=2)
were markedly higher when the dam was at a large body
size than when at small body size (1.52 bouts per obser-
vation vs 0.0, and 1.47 vs 0.59).
The lower nursing-bout frequency seen in small
mothers with twins could be due to behavioral differ-
ences in mothers or in infants. To examine which differ-
ence was more important, we compared the frequency of
mother-initiated infant transports as a measure of mater-
nal interest in infant contact and the frequency of infant-
initiated transport as a measure of infant interest in con-
tact. Maternal harassment of transported infants within
21
Fig. 1 Mean (?SE) percent milk fat in different dam categories
(lowwt-1 below-average-weight dams with singletons; lowwt-2 be-
low-average-weight dams with twins; highwt-1 above-average-
weight dams with singletons; highwt-2 above-average-weight
dams with twins). Categories with different letters differ at P<0.05
Fig. 2 Mean (?SE) gross energy in mid- to late-lactation milk
samples for different dam categories. See legend to Fig. 1 for defi-
nition of categories. Gross energy was calculated as (kcal/g of
milk)?(total g of milk collected on days 21, 30, and 45). Catego-
ries with different letters differ at P<0.05
1 min of transport initiation was also compared as a mea-
sure of maternal tolerance toward infant contact. The fre-
quency of mother-initiated retrievals per infant was high-
er in singleton than in twin groups [Kruskall-Wallis test
(K-W)=8.49, P=0.037], but did not differ between 
mothers of differing size. Frequency of maternal harass-
ment was lowest in groups with small mothers rearing
twins (K-W=6.27, P=0.099). Maternal size/litter size
groups did differ in the frequency of infant-initiated
transport, with fewer such transports in groups with
small mothers nursing twins than in any other group 
(K-W=11.08, P=0.011; Table 2). The same result was
seen in comparisons within dams. The frequency of in-
fant-initiated transports for twin litters (n=2) when dams
were small was less than half that of when dams were
large (0.30 vs 0.70 and 0.6 vs 1.4).
Infant growth
The average birth weights of singleton infants of large
mothers (34.0 g) was greater than that of twins of small
mothers (28.9 g; F=4.278, P=0.009); birth weights of
singletons of small mothers and twins of large mothers
were intermediate between these extremes. However, by
days 21 and 42, the differences in average weights were
between twins from small mothers (43.8 and 65.6 g at 21
and 42 days, respectively) and twins from large mothers
(51.2 and 83.45 g) at 21 days (F=4.097, P=0.011) and
42 days (F=8.587, P=0.0001), with weights of singletons
being intermediate between these. Because birth weights
differed, both absolute litter growth rates and relative lit-
ter growth rates were compared across categories ? the
differences among maternal size/litter size groups was
the same for relative and absolute growth rates. Figure 3
illustrates the relative litter growth rates supported by
dams of differing categories. Growth rates for singletons
did not differ between small and large mothers while lit-
ter growth rates for twins of small mothers were interme-
diate between the growth rates for singleton litters and
those for twins from large mothers. The growth rate for
singletons averaged 1.03 g/day. If individual growth rate
was unrelated to maternal size and litter size, the litter
growth rate for twins would be expected to be approxi-
mately double that for singleton litters, or around
2.06 g/day. However, twins reared by small mothers had
a lower average litter growth rate (1.75 g/day) while
twins reared by large mothers had a higher average
growth rate (2.49 g/day). One factor complicating this
comparison is that, for larger mothers, increasing age
was associated with slower infant growth (r= ?0.600,
P=0.007) and large mothers of singletons were older,
though not significantly, than large mothers of twins (see
Table 1). For smaller mothers, there was no relationship
between age and infant growth (r=?0.117, P=0.66).
Comparisons within dams were variable. Two dams
reared twins at both small and large body sizes. In one
case, the relative growth rate of the litter was higher
when the dam was larger (0.037 vs 0.064 g/g per day)
but in the other case, the relative growth rates did not ap-
pear to differ (0.048 vs 0.051).
Maternal size
Figure 4 illustrates the slope in weight change for dams
in each maternal size/litter size category. Small dams
22
Table 2 Behaviors of mothers and infants during lactation. Results with different letters designate within-row differerences at P<0.05
Low-body-weight Low-body-weight High-body-weight High-body-weight
dam/singleton dam/twins dam/singleton dam/twins
Mean litter nursing bout frequency per observation, 0.987a 0.988a 0.972a 1.36b
0?21 days
Mean litter nursing bout frequency per observation, 0.668a 0.708a 0.650a 1.27b
21?42 days
Mother-initiated infant transport per observation 0.304a 0.238b 0.745a 0.246b
(median)
Infant-initiated infant transport per observation 0.920a 0.503b 0.875a 1.02a
(median)
Carry with harassment per observation (median) 0.286a 0.205a 0.600a 0.356a
Fig. 3 Mean (?SE) relative growth rates for litters in different
dam categories. See legend to Fig. 1 for definition of categories.
Categories with different letters differ at P<0.05
rearing twins were more likely to have a linear weight
loss during lactation than were dams in any other catego-
ry (K-W=14.29, P=0.003). A comparison of weight
changes within dams revealed a similar pattern, with
dams rearing twins (n=2) losing more weight when they
were at a small body size than when at a large size
(?0.43 g/day vs ?0.25, and ?1.15 vs +0.08).
A posteriori analyses revealed that dams with a nega-
tive weight change slope, regardless of maternal size/lit-
ter size category, differed from other dams in three pa-
rameters. They were leaner relative to body weight (a
higher total body water/body weight ratio) than those not
losing weight (Spearman?s rho=?0.675, P=0.002), had 
a lower gross milk energy output for day 21?30 
(Spearman?s rho=0.588, P=0.008), and they had a lower
gross energy in mid- to late-lactation milk samples
(Spearman?s rho=0.603, P=0.022).
Daily food intake for breeding pairs on days 35?36
averaged 167.53 kcal/kg75 and did not differ between
maternal size/litter size categories (F=0.604, P=0.622).
Given the differences in lactation investment that ap-
peared across the four maternal size/litter categories, we
questioned, a posteriori, whether these differences affect-
ed maternal health or reproduction in the year following
the lactation in question. Dams were scored for health
problems (including chronic weight loss to emaciation,
gastrointestinal symptoms such a vomiting and diarrhea,
and other signs of infection) and whether they produced
at least one viable litter within the following year. Again,
small females rearing twins were clearly distinguished
from all other groups as having a high likelihood of mor-
bidity and a low likelihood of fertility during the year
following the lactation in question (Table 3).
Discussion
The results of this study indicate that maternal size af-
fects lactation performance in common marmosets, with
the effects being most evident in mothers rearing twins.
These effects were seen even in the relatively non-
demanding captive environment and differed from the
typical finding in large-bodied anthropoid primates of 
a limited relationship between maternal size/maternal 
energy reserves and lactation investment.
Milk composition, gross milk energy output, and
nursing frequency were all sensitive to the combination
of maternal size and lactation demand such that growth
in twins but not singletons was affected by maternal size.
This result is similar to that observed in rodents, for
which, within the normal range of litter sizes, larger
mothers are frequently capable of supporting higher
growth rates in larger litters (Mattingly and McClure
1982; Knight et al. 1986) while at smaller litter sizes, 
infant growth is limited (Knight et al. 1986; Hammond 
et al. 1996).
Smaller mothers rearing twins had lower milk fat,
lower gross energy in mid- to late-lactation milk sam-
ples, and lower nursing-bout frequencies than did large
mothers nursing twins, suggesting that small mothers
were investing less in twin litters both through less ener-
gy placed in a given volume of milk and by producing
less milk overall. These differences did not appear, how-
ever, to be driven by differences in maternal behavior:
small mothers of twins appeared to be just as attracted to
and tolerant of infants as all other mothers. However, the
twin infants of these small mothers responded behavior-
ally to suboptimal levels of milk yield by initiating car-
rying by the mother less often. The growth rates of twins
were correlated with lactation variables, so the lower to-
tal growth rates supported by smaller mothers was prob-
ably due in part to differences in lactation performance.
For singletons, maternal size was not associated with
lactation performance or maternal behavior with one ex-
ception: small dams had a lower gross energy in mid- to
23
Table 3 Morbidity and fertility of mothers following lactation
Low-body-weight Low-body-weight High-body-weight High-body-weight
dam/singleton dam/twins dam/singleton dam/twins
n 6 11 6 13
Percent fertility (?2=12.1, P=0.007) 100.0 36.4 66.7 92.3
Percent morbidity (?2=13.1, P=0.005) 0.0 63.6 16.7 7.6
Fig. 4 Weight change slope for dams in different categories. See
legend to Fig. 1 for definition of categories. Boxes represent the
interquartile range, the line across the box represents the median,
whiskers show the entire range
late-lactation milk samples than did large dams. The
growth rates of singleton infants were not related to any
lactation parameters and the growth rate of singletons
supported by small versus large mothers did not differ.
Marmosets appear to be adapted to support optimal
growth rates in twins. Large females could not gain, re-
productively, by producing singletons, because infant
growth appears to be limited by some combination of fac-
tors that prevent the infant from growing at a faster rate,
even though the dams appear to be capable of producing
more milk energy. Smaller females, however, might gain
by producing singletons, because their twins are likely to
be smaller and have slower growth rates that their single-
tons. The fact that small mothers rearing twins were high-
ly likely to be ill and infertile in the year following lacta-
tion, while small mothers rearing singletons were fertile
and showed no illness, suggests that the combined effects
of maternal and litter size might be an important factor in
reproduction allocation in this small primate. This result
is similar to that seen in house cats, in which maternal
stress (lack of coordination, bedraggled appearance, and
poor appetite) was most prevalent in small mothers nurs-
ing large litters (Deag et al. 1987).
The finding of better growth in twins of large 
mothers, compared to singletons, differs from that of oth-
er mammals that produce small litters of one to two off-
spring (e.g., sheep, cattle, seals, most primates), in which
mothers generally support higher growth rates in single-
tons than in twins, regardless of maternal size. One factor
that may complicate the comparison of singleton versus
twin growth of infants from larger mothers is maternal
age, because maternal age was related to infant growth
for larger mothers. Given the relatively long potential re-
productive life span of marmosets, further examination of
the interaction of age and maternal condition effects upon
reproductive investment would be valuable.
We calculated two estimates of gross energy output in
milk. One of these (using labeled-water dilution) was
relatively specific and estimated milk energy output dur-
ing a limited time period (days 21?30). Gross energy
output measured in this fashion was strongly related to
initial infant size, was correlated with preweaning
growth rates in twin infants only, and was not related to
maternal size. The second estimate, gross energy in 
mid- to late-lactation milk samples was less precise but
measured output over a longer time period. This measure
was positively related to growth in twin litters but not
singletons, and was related to maternal size, with small
mothers rearing twins having a lower gross milk energy
output indicator. Differences in the results between these
two methods suggest the possibility that differences in
milk energy output will only be seen when a longer peri-
od of lactation is examined. We assumed that the most
energetically demanding period of lactation would be
that immediately preceding the initiation of weaning,
when infants begin to receive some of their energy from
non-maternal sources. However, weaning in primates is a
particularly prolonged process. In marmosets, for exam-
ple, infants are consuming solids as well as milk for
around 60% of the total lactation period. Without knowl-
edge of the relative contribution of milk versus solid
food to infant energy intake during this period, reliably
assessing the relative energy expense for the mother or
the relative energy gains for the infants associated with
the different stages of primate lactation is not possible.
Though weight varied during lactation in all females,
small mothers of twins were the only maternal size/litter
size group that averaged a significant, linear weight loss
during lactation. In all other groups, females had either a
non-significant linear weight change or a significant 
linear weight gain during lactation. This result differs
from that of Nievergelt and Martin (1999) who found that
relatively large dams (average weight of 395 g) nursing
twins lost weight across the course of lactation; in our
study, the overall weight change for large dams nursing
twins averaged 0. There are a number of possible reasons
for the different results. The weight loss reported by 
Nievergelt and Martin (1999) averaged 8%, and overall
variation in weekly weights might make detection of such
a difference unlikely using a linear regression. The weight
losses observed in the low-maternal-weight/high-litter-
size dams averaged 12% and were not normally distribut-
ed. While direct comparison between these studies is not
possible due to methodological differences, the results of
the two studies, taken together with those of Sanchez et al.
(1999) on cotton-top tamarins, suggest that weight loss in
callitrichid mothers fed ad libitum is negligible to modest
(0?8%) in large mothers nursing twins or in mothers of
any size nursing singletons, while small mothers nursing
twins face a larger drain on energy reserves.
Because the food intake reported here represents the
intake of the mated pair, the specific energy values can-
not be compared with those of previous studies; howev-
er, the majority of the energetic demands of lactation
were apparently met by increased food intake, a result
similar to that of Nievergelt and Martin (1999). The food
intake of pairs rearing singletons and those rearing twins
did not differ. This result is perhaps not surprising given
the potential ?start-up? costs associated with lactation.
Hammond et al. (1996) report that the energetic costs of
lactation per pup in mice is greater in smaller litters be-
cause the start-up costs associated with lactation are sim-
ilar, regardless of litter size. So while the estimated food
consumption per day per gram pup produced was 0.24 g
for litters of two or greater, the consumption per gram of
pup per day for the first pup was double that (0.5?0.6 g)
(Hammond et al. 1996).
A similar relative intake in small and large mothers
suggests that intake during lactation may be near some
limit. Studies by Rogowitz (1998) on cotton rats and
Hammond et al. (1996) on mice suggest that increases in
energy assimilation reach some physiological limit at
higher litter sizes, such that energy assimilated per off-
spring begins to decline. If large mothers had relatively
lower basal metabolic requirements, they might be able
to devote more intake energy toward lactation.
A linear weight loss during lactation was associated
with lower gross energy output regardless of the number
24
of infants nursed. So while small mothers did draw more
heavily on reserves (as suggested by weight loss), the re-
serves drawn upon were inadequate to meet the lactation
demand, resulting in lower milk energy output. This re-
sult is similar to that found in other mammals (Mattingly
and McClure 1985; Roberts et al. 1985; Oldham and
Friggens 1989; Dewey 1997; Rogowitz 1998). It provides
further support for the proposal that, across a fairly wide
variety of mammals, maternal investment as reflected in
milk yields is sensitive to maternal energy balance.
Acknowledgements The authors thank Karen Bales, Cashell
Jaquish, and the staff of the Nutrition Laboratory, Department of
Zoological Research, National Zoological Park, for their help in
data collection, and Phyllis Lee and two anonymous reviewers for
their comments on a previous version of this manuscript. This re-
search was supported by NIH grant R01-RR02022 to S. Tardif.
References
Araujo, A, Arruda MF, Alencar AI, Alburquerque F, Nascimento
MC, Yamamoto ME (2000) Body weight of wild and captive
common marmosets (Callithrix jacchus). Int J Primatol 21:
317?324
Arnbom T, Fedak MA, Boyd IL (1997) Factors affecting maternal
expenditures in southern elephant seals during lactation. Ecol-
ogy 78:471?483
Bercovitch FB (1987) Female weight and reproductive condition
in a population of olive baboons (Papio anubis). Am J Prima-
tol 12:189?195
Bronson FH, Marsteller FA (1985) Effect of short-term food 
deprivation on reproduction in female mice. Biol Reprod
33:660?667
Crnic LS (1980) Models of infantile malnutrition in rats: effects
on maternal behavior. Dev Psychobiol 13:615?628
Deag JM, Lawrence CE, Manning A (1987) The consequences of
differences in litter size for the nursing cat and her kittens. 
J Zool (Lond) 213:153?179
Dewey KG (1997) Energy and protein requirements during lacta-
tion. Annu Rev Nutr17:19?36
Dobson FS, Michener GR (1995) Maternal traits and reproduction
in Richardson?s ground squirrels. Ecology 76:851?862
Festa-Bianchet M, Galliard J-M, Jorgensen JT (1998) Mass- and
density-dependent reproductive success and reproductive costs
in a capital breeder. Am Nat 152:367?379
Gittleman J, Thompson S (1988) Energy allocation in mammalian
reproduction. Am Zool 28:863?875
Hammond KA, Kent Lloyd KC, Diamond J (1996) Is mammary
output capacity limiting to lactational performance in mice? 
J Exp Biol 199:337?349
Hansson L (1992) Fitness and life history correlates of weight
variation in small mammals. Oikos 64:479?484
Hare JF, Murie JO (1992) Manipulation of litter size reveals no
cost of reproduction in Columbian ground squirrels. J Mam-
mal 73:449?454
Hearn JP, Lunn SF (1975) The reproductive biology of the 
marmoset monkey, Callithrix jacchus. In: Perkins FT, 
O?Donoghue PN (eds) Breeding simians for developmental bi-
ology. Laboratory Animals Ltd, London, pp 191?202
Hochachka W (1992) How much should reproduction cost? Behav
Ecol 3:42?52
Iverson SJ, Bowen WD, Boness DJ, Oftedal OT (1993) The effect
of maternal size and milk energy output on pup growth in gray
seals (Halichoerus grypus). Physiol Zool 66:61?88
Jonsson KI (1997) Capital and income breeding as alternative tac-
tics of resource use in reproduction. Oikos 78:57?66
Knight CH, Maltz E, Docherty AH (1986) Milk yield and compo-
sition in mice: effects of litter size and lactation number.
Comp Biochem Physiol 84A:127?133
Konig B, Riester J, Markl H (1988) Maternal care in house mice
(Mus musculus). II. The energy cost of lactation as a function
of litter size. J Zool (Lond) 216:195?210
Leamy L (1992) Morphometric studies in inbred and hybrid house
mice. VIII. Effects of litter size on brain size and body size.
Growth Dev Aging 56:35?43
Lee PC (1987) Nutrition, fertility and maternal investment in pri-
mates. J Zool (Lond) 213:409?422
Mattingly DK, McClure PA (1982) Energetics of reproduction in
large-littered cotton rats (Sigmodon hispisdus). Ecology 63:
181?195
Mattingly DK, McClure PA (1985) Energy allocation during lacta-
tion in cotton rats (Sigmodon hispidus) on a restricted diet.
Ecology 66:928?937
McNamara JP (1995) Role and regulation of metabolism in adi-
pose tissue during lactation. Nutr Biochem 6:120?129
Morris DW (1996) State-dependent life history and senescence of
white-footed mice. Ecoscience 3:1?6
Neuhaus P (2000) Weight comparisons and litter size manipula-
tion in Columbian ground squirrels (Spermophilus columbi-
anus) show evidence of costs of reproduction. Behav Ecol 
Sociobiol 48:75?83
Nievergelt CM, Martin RD (1999) Energy intake during reproduc-
tion in captive common marmosets (Callithrix jacchus).
Physiol Behav 65:849?854
Oftedal OT (1984) Milk composition, milk yield and energy out-
put at peak lactation: a comparative review. Symp Zool Soc
Lond 51:33?85
Oftedal OT, Iverson SJ (1987) Hydrogen isotope methodology for
measurement of milk intake and energetics in suckling young.
In: Huntley AC, Costa DP, Worthy GAJ, Castellini MA (eds)
Approaches to marine mammal energetics Soc Mar Mammal
Spec Publ 1
Oftedal OT, Iverson, SJ (1995) Comparative analysis of nonhu-
man milks. A. Phylogenetic variation in the gross composition
of milks. In: Jensen RG (ed) Handbook of milk composition.
Academic Press, San Diego, pp 749?788
Oldham JD, Friggens NC (1989) Sources of variability in lacta-
tional performance. Proc Nutr Soc 48:33?43
Power RA, Power ML, Layne DG, Jaquish CE, Oftedal OT, Tardif
SD (2001) Relations among measures of body composition in
the common marmoset. Comp Med 51:218?223
Rasmussen KM (1992) The influence of maternal nutrition on lac-
tation. Annu Rev Nutr 12:103?118
Reznick D (1985) Cost of reproduction: an evaluation of the em-
pirical evidence. Oikos 44:257?267
Roberts SB, Cole TJ, Coward WA (1985) Lactational performance
in relation to energy intake in the baboon. Am J Clin Nutr
41:1270?1276
Rogowitz G (1998) Limits to milk flow and energy allocation dur-
ing lactation of the hispid cotton rat (Sigmodon hispidus).
Physiol Zool 71:312?320
Sanchez S, Pelaez F, Gil-Burmann C, Kaumanns W (1999) Costs
of infant-carrying in the cotton-top tamarin (Saguinus oedi-
pus). Am J Primatol 48:99?111
Schneider JE, Wade GN (1989) Effects of maternal diet, body
weight and body composition on infanticide in Syrian ham-
sters. Physiol Behav 46:815?821
Smith CC, Fretwell SD (1974) The optimal balance between size
and number of offspring. Am Nat 108:499?506
Steingrimsdottir L, Greenwood M, Brasel J (1980) Effects of
pregnancy, lactation and a high-fat diet on adipose tissue in
Osborne-Mendel rats. J Nutr 110:600?609
Tardif SD, Carson RL, Gangaware BL (1990) Infant-care behaviors
of mothers and fathers in a communal-care primate, the cotton-
top tamarin (Saguinus oedipus). Am J Primatol 22:73?86
Tardif S, Jaquish C, Layne D, Bales K, Power M, Power R, 
Oftedal O (1998) Growth variation in common marmoset
monkeys (Callithrix jacchus) fed a purified diet: relation to
care-giving and weaning behaviors. Lab Anim Sci 48:264?269
Thompson GE (1983) The intake of milk by suckled, newborn
lambs and the effects of twinning and cold exposure. Br J Nutr
50:151?156
25