Abstract

Studies of cooperatively breeding birds and mammals generally concentrate on the effects that helpers have on the number of reproductive attempts females have per year or on the number and size of offspring that survive from hatching/weaning to independence. However, helpers may also influence breeding success before hatching or weaning. In the present study, we used an ultrasound imager to determine litter sizes close to birth, and multivariate statistics to investigate whether helpers influence female fecundity, offspring survival to weaning, and offspring size at weaning in cooperative meerkats, Suricata suricatta. We found that the number of helpers in a group was correlated with the number of litters that females delivered each year, probably because females in large groups gave birth earlier and had shorter interbirth intervals. In addition, although pup survival between birth and weaning was primarily influenced by maternal dominance status, helper number may also have a significant positive effect. By contrast, we found no evidence to suggest that helpers have a direct effect on either litter sizes at birth or pup weights at weaning, which were both significantly influenced by maternal weight at conception. However, because differences in maternal weight were associated with differences in helper number, helpers have the potential to influence maternal fecundity and offspring size within reproductive attempts indirectly. These results suggest that future studies may need to consider direct and indirect helper effects on female fecundity and investment before assessing helper effects on reproductive success in societies of cooperatively breeding vertebrates.

Helpers are often suggested to have two major effects on breeders in societies of cooperatively breeding birds and mammals: increasing breeder fecundity and increasing the growth and survival of breeders' offspring (Cockburn, 1998; Emlen, 1991; Hatchwell, 1999; Jennions and Macdonald, 1994). Multivariate analyses and removal experiments provide strong evidence in support of both of these effects (see Brown and Brown, 1981; Brown et al., 1982; Clutton-Brock et al., 2001c; Emlen and Wrege, 1991; Mumme, 1992; Solomon, 1991), although such findings are not universal (for review, see Cockburn, 1998; Emlen, 1991). However, measures of fecundity are typically restricted to the number of reproductive attempts performed by breeders each year, whereas measures of offspring growth and survival are generally restricted to involve nestling/postweaning periods only. In the present study, we use a portable ultrasound imager to investigate, for the first time in a free-living cooperative mammal, the effects of helper number on female fecundity within reproductive attempts and on offspring survival before weaning.

Helpers could have direct and/or indirect influences on the fecundity of breeding females within reproductive attempts. Helpers could influence maternal fecundity directly, if mothers invest differentially according to the number of individuals present to help in raising their offspring, or indirectly, if helpers influence the condition of mothers and if maternal investment is condition dependent. Studies of noncooperatively breeding birds or mammals show that both alternatives present viable possibilities. For instance, Petrie and Williams (1993) found that female peafowl (Pavo cristatus) directly alter the size of their clutch depending on their partner's attractiveness. Similarly, clutch size in birds can be associated, albeit generally weakly, with maternal condition (Perrins, 1971), whereas, in mammals, litter size is commonly related to maternal condition (Clutton-Brock, 1991; Hoogland, 1995; and references therein). Helpers may play a direct and/or indirect role in shaping maternal investment strategies in each offspring within a reproductive attempt, and again, findings from studies of noncooperative breeding birds or mammals provide evidence for both alternatives. For instance, Cunningham and Russell (2000) showed experimentally in mallard ducks (Anas platyrhynchos) that mate attractiveness can directly influence a female's investment in her eggs and, hence, the size of her offspring at hatching. Other studies suggest that in birds, female condition may at least explain some of the variation in the size of eggs within clutches (for review, see Christians, 2002), whereas in mammals, female condition commonly correlates with offspring condition at birth (Clutton-Brock, 1991; Hoogland, 1995; and references therein). Finally, in cooperative breeders, helpers could also play a more active role in influencing offspring condition and survival prehatching or preweaning because they may help to incubate eggs in birds (Zahavi, 1990) and guard or suckle preweaned offspring in mammals (Creel et al., 1991).

Meerkats are small (less than 1 kg), obligate, cooperatively breeding mongooses that live in year-round territories in arid regions of southern Africa. Groups comprise a dominant male and female breeder, which are parents of most of the litters born, and a variable number (2–30) of nonbreeding helpers of both sexes (Clutton-Brock et al., 2001b). Like many social carnivores, female meerkats give birth to their litter under ground, and therefore, accurate measures of litter sizes at birth are normally not possible. Litters remain in or at the burrow for their first month, during which time they are primarily dependent on their mother for milk, although some helpers may act as allo-sucklers (Scantlebury et al., 2002). Pups are protected by helpers while at the burrow (Clutton-Brock et al., 1998) and are then fed invertebrates and small vertebrates by helpers while they follow the group during their subsequent two months of dependence (Brotherton et al., 2001). Helpers have considerable effects on the growth and survival of weaned pups, independently of maternal investment (Clutton-Brock et al., 2001c; Russell et al., 2002), and allow dominant females to reduce investment in raising litters (Clutton-Brock et al., 1998, 2001a). Consequently, in meerkats, helpers have the potential to influence maternal investment strategies directly (because helpers influence breeding success) and indirectly (because helpers may affect maternal condition).

We address three specific questions pertaining to the potential direct and indirect helper effects on breeding success in meerkats by using a multivariate statistical approach. First, we investigate whether helper number affects two measures of fecundity: (1) the number of litters delivered by dominant females each year, and (2) using a portable ultrasound imager, the number of pups delivered by subordinate and dominant females in each attempt. Second, we investigate the effects of helper number on: (3) pup survival between ultrasound scanning and weaning, and (4) mean pup weights at weaning. Third, we investigate whether or not maternal weight at the conception of a current litter is associated with the number of individuals that helped to raise the previous litter. The present study provides the first investigation into helper effects on litter sizes close to birth and pup survival between birth and weaning in a free-living cooperative mammal.

METHODS

The present study was performed from December 1996–January 2001 in the South African Kalahari, 20 km south of the Botswana border close to Van Zyl's Rus (25° 58′ S, 20° 49′ E). During this period, we monitored 178 breeding events by 61 females in 13 groups. Details of the habitat and climate are given elsewhere (Brotherton et al., 2001; Russell et al., 2002). All individuals in this study were habituated to close observation (<1 m) and the majority (>90%) were weighable on each visit (±1 g) by using electronic balances. Groups were visited every 1–3 days. During these visits, data were collected on groups (number of individuals), females (age, weight, pregnancy, and dominance status), and litters (birth date, emergence date, first foraging date, size, weight, and sex ratio). Helper number, maternal age, and maternal weight were measured during the period when litters were being conceived, unless otherwise indicated. Helper number was defined as the number of individuals over 6 months of age, excluding the mother and father of the current litter; individuals under 6 months contribute little to pup care (Clutton-Brock et al., 2002). Helper number not only varied between groups (range, 3–25 individuals), but also within groups between different breeding events (range of differences in helper numbers within groups, 1–12). Dates of birth were determined (to within 3 days) by using visual observations and weight changes of pregnant females, whereas dominance status was determined by using behavioral observations (Clutton-Brock et al., 2001b). Conception date was defined as being 60–80 days before the birth of a litter; gestation in meerkats is approximately 70 days long (Clutton-Brock TH, unpublished data).

Statistical analyses

All analyses were initially conducted using residual maximum likelihood (REML) models (also known as restricted maximum likelihood models) or iterated reweighted residual (or restricted) maximum likelihood models (IRREML) in GENSTAT 5.4.1 (GENSTAT, 1993). REML and IRREML models are algorithms of general(ized) linear model (GLM) with normal and nonnormal error structures, respectively, which allow both fixed and random components to be fitted; random components in our models take into consideration repeated sampling within mothers and groups (Schall, 1991). The significance of random terms in a mixed model indicates whether or not the variation within individuals or groups is significantly less than the variation between individuals or groups. A significant random term will therefore suggest that inherent properties of that term (individual and/or groups) affect the response term in the analysis. When both random terms were found to have nonsignificant influences on the explanatory power of the model, we used GLM. The results arising from GLM did not differ qualitatively from their mixed model equivalents in which random terms were not significant. Using GLM is appropriate when repeated sampling has been shown to have a nonsignificant effect because one should always attempt to use the most parsimonious model.

Potential explanatory (fixed) terms considered were entered into each REML model, IRREML model, or GLM using the stepwise forward technique until the model only included those terms for which elimination would have significantly reduced the explanatory power of the model. For fixed effects, significant probability values are derived from having all significant terms fitted in the final model together, whereas those of nonsignificant fixed terms are obtained from having all significant terms in the model and each nonsignificant term fitted individually. Although the significance of all two-way interactions were tested in each model, none was significant.

Levene tests indicated that none of the dependent variables tested in REML models or the GLMs with normal error structures differed significantly from normality (p >.05), and thus, no transformations were required. All mean values are presented are ±1 SD.

Number of litters produced per season

The duration of a reproductive season is climate dependent, and in the present study, no (or very few) litters were born in May or June (winter). Therefore, for the purposes of this study, a reproductive season potentially began on 1 July and potentially ended on 30 April, and the numbers of litters produced by dominant females within a season were calculated for this period. Subordinate females were not included in this analysis because their conception is largely opportunistic, depending on age, weight, rainfall, and the presence of unrelated males (Clutton-Brock et al., 2001b).

The relationship between helper number and the number of litters produced by dominant females per season was investigated by using GLM on 51 births by 20 females over one to six seasons in 13 groups. Neither repeated sampling among mothers (n = 1–6) nor groups (n = 1–8) was found to constitute significant random components in a REML model (p >.5). In the GLM, helper number and female weight and age were fitted as fixed effects. Maternal age and weight were fitted to control for potential influences of maternal effects.

Litter size at birth

In most cooperatively breeding mammals, mothers give birth under ground. Consequently, litter sizes are generally measured in cooperative mammals when offspring first emerge at the surface. In the present study, we attempted to assess female fecundity more accurately by capturing, anesthetizing, and scanning pregnant females 5–30 days prepartum (mean = 16 days) by using an MLV concept model portable ultrasound imager. Although we use the term litter size at birth throughout this article, it is acknowledged that some reabsorption of fetuses is possible even during this late stage of pregnancy. However, we found no correlation between the timing of ultrasound scanning and the numbers of pups scanned (p >.9) (Table 1) or the timing of ultrasound scanning and litter sizes at weaning (p =.49) (Table 2). The terms associated with litter sizes at birth were investigated using a GLM on 35 litters, scanned from 20 females in nine groups. Helper number, maternal age, weight, and dominance status at conception, as well as the number of days before birth that litters were scanned, were all fitted as fixed effects. Repeated measures among mothers (n = 1–4) and groups (n = 2–9) did not significantly influence the variation explained in a REML model (p >.5).

Pup survival (birth-weaning)

Pups first emerge from their natal burrows at 11–30 days (mean = 18 days; this study), and begin to follow the group on foraging trips at 17–42 days (mean = 29 days) (Brotherton et al., 2001). Before following the group on foraging trips, pups receive all of their sustenance in the form of milk. Our aim here was to investigate whether or not helpers may influence pup survival (and weight, see below) before they begin to provide them with solid food. We therefore define weaning as the day immediately before pups go on their first forage and, hence, just before they begin to receive solid food.

Accurately assessing whether or not helpers have any influence on offspring before they being to receive solid food from the helpers is difficult because it is generally not possible to determine litter sizes at birth. To estimate mortality between birth and weaning, we compared the number of fetuses in the 35 ultrasound scans with the numbers of pups from those same 35 litters at weaning. The factors associated with pup survival between birth and weaning were investigated in a GLM with binomial error structure and logit link function in which litter size at weaning was fitted as the response variable and litter size at birth was fitted as the binomial denominator. Helper number, maternal status, age, and weight, as well as the number of days before birth that litters were scanned, were fitted as potential explanatory terms. Neither maternal identity nor group identity was a significant random component in an IRREML model (p >.5).

Litter sizes at weaning

Litter sizes at weaning were also investigated for a larger data set involving 169 litters delivered by 54 mothers in 10 groups. We investigate the factors that influence litter sizes at weaning in two separate analyses. First, we investigate what factors affect whether or not a mother has any pups at all at weaning. In this analysis, pups present at weaning (zero/one) were fitted as the response in a GLM with a binomial error structure and logit link function, and one was fitted as the binomial denominator. Neither maternal identity (repeated measures per mother, 1–13) nor group identity (repeated measures per group, 10–28) constituted significant random terms in an IRREML model. Second, of the 95 litters in which mothers did have at least one surviving pup, we investigate what factors influence the numbers of pups present at weaning. In this analysis, we fitted litter size as the response variable in a GLM with a normal error structure. Neither maternal identity (n = 28 mothers; repeated measures per mother, 1–12), nor group identity (n = 10; repeated measures per group, 3–17) was a significant random term in a REML model. In each analysis, helper number, maternal status, age, and weight were fitted as fixed effects.

Pup weights at weaning

The average weight of pups from litters was determined by weighing pups as close to weaning as possible, and always within a weak of weaning (definition of weaning as above). Factors influencing mean weights of pups at weaning were investigated in a GLM analysis on 37 litters born to 17 mothers from nine groups, with one to six repeated measures per female and one to eight repeated measures per group. However, repeated measures of mothers and groups were not significant random components in a REML model (p >.5). Fixed effects fitted were helper number, maternal age, weight and dominance status, litter age, size, and sex ratio. Maternal characteristics were measured at conception, whereas group and litter characteristics were measured at weighing.

Female weight is commonly a strong predictor of offspring condition in social noncooperatively breeding mammals (Clutton-Brock, 1991). A supplemental feeding experiment was conducted on pregnant females to directly investigate the role of maternal weight on investment. Seven dominant females were fed with approximately 12 g of egg twice daily for 48–83 days (mean, 71 days) during pregnancy. This type of feeding experiment has been shown to have significant effects on the weights and contributions to cooperation by helpers (Clutton-Brock et al., 2001c, 2002). We then compared the mean weight of offspring at weaning from experimental litters with those from the litter immediately before the experiment, for the same seven females. The age at which pups were weighed between experimental and control litters did not differ significantly (t11 = −1.10, p =.29), and although the age of mothers was obviously greater for experimental litters, maternal age has no influence on the weight of pups at weaning (see Results).

Helper number and maternal weight

If litter sizes or pup weights are influenced by maternal weight, and if maternal weight is correlated with helper number, then helper number may influence meerkat breeding success indirectly. Female weight at conception was calculated as the average morning preforaging weights of mothers 60–80 days before the birth of their litter, whereas helper number was the number of helpers that helped to raise the previous litter. We investigated the effect of helper number on maternal weight at conception by using a REML model of 169 conception weights from 60 mothers in 13 groups (one to 11 measures per mother and two to 25 measures per group). We used a REML model (maternal identity effect, p <.05) in which maternal weight was fitted as the response variable and maternal age, dominance status, helper number, interbirth interval, and season (hot/wet versus cold/dry) were fitted as fixed effects. Maternal identity was maintained as a random component.

RESULTS

Number of litters produced per season

The number of litters delivered by dominant females each season varied from one to four. This number was significantly influenced by the number of individuals that were present to help raise litters, with increases in helper number being positively associated with increases in the number of litters delivered per year (Figure 1). There are at least two potential mechanisms through which this effect could arise: (1) if females breed earlier and/or later in a season when in large groups, or (2) if females have shorter interbirth intervals in large groups. The month in which breeding began (i.e., first litter born) within a season ranged from July–January, and the month in which breeding ceased (i.e., last litter born) ranged from November–April (n = 51); interbirth intervals within breeding seasons ranged from 73–149 days (n = 70). The fact that gestation is around 70 days in meerkats, and interbirth interval can be as little as 73 days, suggests that female meerkats may reconceive within days of giving birth to their previous litter. Breeding began significantly earlier in the presence of more helpers (Kruskal-Wallis, H = 6.66, df = 2, p =.036), but there was no relationship between helper number and the month in which breeding ceased (H = 2.15, df = 2, p =.34). In addition, females had shorter interbirth intervals when helped by many individuals in the previous attempt (GLM: F1,57 = 8.17, p =.006).

Litter sizes at birth

Litter sizes in ultrasound scans ranged from three to six (mean = 4.1). Maternal weight at conception was the only factor to influence litter sizes at birth, with heavy females having significantly larger litters than those of lighter females. We found no significant effect of helper number, maternal age, or maternal status on litter sizes at birth (Table 1).

Pup survival (birth-weaning)

Comparison of litter sizes at birth with those at weaning suggested a preweaning mortality rate of 50%. Of the 35 litters scanned, 37% failed to produce any offspring to weaning (although all were known to give birth), whereas 29% had complete success. The average number of offspring dying in partially successful litters was 1.9 ± 1.0 (n = 12). Helper number showed a nonsignificant tendency to be associated with increased survival between birth and weaning, whereas female age and weight showed no significant influence (Table 2). Only maternal status significantly influenced pup survival between birth and weaning, with 68% of pups born to dominants surviving and only 28% of pups born to subordinates doing so. This difference was primarily caused by subordinate females having significantly greater complete breeding failure (i.e., no surviving pups) than that of dominant females (G test, G = 11.43, df = 1, p =.007), most probably as a consequence of infanticide (Clutton-Brock et al., 2001b).

Litter sizes at weaning

Of the 169 litters born during the study period, 73 (43%) failed to produce any offspring to weaning. No significant effect of helper number, maternal age, or maternal weight was found on the probability that at least one pup would survive from a given litter (Table 3). Again, maternal status was the primary factor to influence litter survival to weaning, with a significantly greater proportion of litters born to dominants producing surviving offspring than those born to subordinates (84% versus 23%).

Litter sizes at weaning in which at least one pup survived ranged from one to six (mean, 3.56). Although we were not able to detect a significant effect of maternal age or status on litter sizes at weaning, we did find significant effects of helper number and maternal weight (Table 3). The finding that maternal weight most significantly influences litter sizes at weaning suggests that litters in which at least one pup survived are more influenced by litter size at birth than by pup survival between birth and weaning.

Pup weights at weaning

Mean pup weights at weaning varied from 82–159 g (mean, 118 g). We found that the only factor that significantly influenced the average weight of pups at weaning (other than their age) was their mother's weight at conception (Table 4). Supplemental feeding of pregnant females suggested a causal link between maternal weight and the mean weight of emergent pups, for pups were heavier when their mothers had been supplementally fed compared with when they had not (paired t test: t7 = 2.43, p =.045).

Helper number and maternal weight

Maternal weight at conception ranged from 510 g to 956 g. Not surprisingly, maternal weight was strongly influenced by age (REML: χ2 = 52.90, df = 1, p <.0001). However, maternal weight was also significantly influenced by dominance status (χ2 = 4.19, df = 1, p =.041), the number of days since females last gave birth within a breeding season (χ2 = 10.24, df = 1, p =.001), and season (hot/wet > cold/dry, χ2 = 4.26, df = 1, p =.039). After correcting for these significant effects, maternal weight was found to be significantly influenced by the number of individuals that had helped to raise the previous litter, with those having been previously helped by few helpers being lighter at the conception of the following litter than those that had been helped by many helpers (χ2 = 6.35, df = 1, p =.012) (Figure 2).

DISCUSSION

We investigated the effects of helper number and maternal state on the numbers of litters produced by dominant females per year, the numbers of pups produced by females per litter, pup survival between birth and weaning, as well as pup weights at weaning. Although we found that the average number of helpers in a group was positively associated with the number of litters produced by dominant females per year, we found that maternal weight at conception was the only term to influence litter sizes at birth. Maternal weight was also the only term found to be associated with pup weights at weaning, whereas maternal dominance status was the primary predictor of pup survival between birth and weaning. However, helper number did have a significant effect on litter sizes at weaning. In addition, because maternal weight at conception was significantly influenced by the number of individuals that helped to raise the previous litter, the number of helpers was likely to have affected both the number of pups born, and their weights at weaning indirectly, though their influence on maternal condition.

Two possible explanations may account for the observed relationship between the number of helpers and the number of litters delivered by dominant females within each season. First, it could be that the relationship was simply a consequence of maternal or territory quality, if some mothers or territories were able to accommodate litters sooner and more often than others. Although it is difficult to rule this possibility out completely, we suggest that this explanation is unlikely to be true in our study. This is because a mixed model analysis revealed that neither maternal identity nor group identity contributed significantly to the variation in the number of litters produced by dominant females, suggesting that inherent properties of mothers or groups are unlikely to explain the results found. Alternatively, increases in helper number could cause increases in the number of litters produced by dominant females because, through “load-lightening” (Brown and Brown, 1981), many helpers may allow mothers to reach breeding condition earlier in the year and to have reduced interbirth intervals. This makes sense in meerkats because, as helper numbers increase, mothers are less likely to conduct any babysitting before weaning (Clutton-Brock et al., 1998) or to invest as much in pup-provisioning after weaning (Clutton-Brock et al., 2001a), investment in both of which are associated with significant weight loss (Russell et al., 2003).

Studies of other cooperative breeders have also suggested that helpers cause mothers to reduce their investment in maternal care (for review, see Crick, 1992; Emlen, 1991). Indeed, lightening of parental duties may be a primary way in which helpers of some species increase breeders' reproductive success and, hence, their own inclusive fitness (Hatchwell, 1999). Although the vast majority of such studies are correlational and are unable to differentiate cause and effect, experimental removal of helpers in some avian or mammalian systems confirm helpers to be effective in causing load-lightening (Hatchwell and Russell, 1996) as well as in reducing interbirth intervals (Powell and Fried, 1992; Solomon, 1991).

Although there is evidence to show that helpers influence fecundity by allowing mothers to reproduce more often within a season, whether or not they influence mothers to produce more or heavier offspring within an attempt is less well established. In noncooperatively breeding bird species, females are capable of directly increasing clutch size (Petrie and Williams, 1993) and egg size (Cunningham and Russell, 2000) in response to the quality of their mate. Similar findings may be expected for cooperatively breeding species in which helpers have a positive effect on offspring growth and survival. However, there are few data to show whether this is the case, although in dunnocks (Prunella modularis), polyandrously breeding females lay more eggs than do monogamously breeding females (Davies and Hatchwell, 1992). In wild cooperatively breeding mammals, access to newly born young is commonly difficult, and consequently, relationships between helper number and maternal investment are not known.

In the present study, we found a significant positive association between helper number and litter size at weaning. This kind of finding could be interpreted as providing evidence for direct maternal adjustment of litter size in response to helper number, because pups in large groups have faster rates of growth and generally greater survival between emergence and independence (Russell et al., 2002). However, evidence from ultrasound scanning of pregnant mothers suggested that this is unlikely, for litter sizes at birth were not influenced by helper number directly, but by a mother's age-related weight. Therefore, the relationship found between helper number and litter size at weaning is best explained by differential pup mortality in groups of different sizes, and our analysis of the factors affecting pup survival between birth and weaning support this suggestion. The likely reason for increased preweaning survival of pups in large groups is that such pups are more likely to be baby-sat by older, more experienced helpers (Clutton-Brock et al., 1998). Finally, maternal weight was also found to have a significant effect on litter sizes at weaning. In this case however, this is likely to be owing to the effect of maternal weight on litter sizes at birth, for there was no evidence that maternal weight influenced pup survival.

Maternal weight at conception was the only predictor of pup weights at weaning, and the causality of this effect was suggested in the supplemental feeding experiment. The fact that helper number did not have a significant effect on offspring weight at weaning was surprising given that allo-sucklers are more common in large groups (Scantlebury et al., 2002) and large groups may provide pups with greater thermodynamic benefits (Russell et al., 2002). However, Scantlebury et al. (2002) showed that mothers invest close to maximal amounts of energy in lactation and, in fact, reduce their investment in the presence of allo-sucklers. Therefore, allo-sucklers may not provide additive care to offspring in meerkats, but may instead serve to improve maternal condition. It also seems unlikely that helper number would directly influence pup weights at birth, although to rule this possibility out, fetus sizes must be estimated from ultrasound scans.

Although we found that maternal weight at conception (rather than helper number) influenced litter sizes at birth and pup weights at weaning, we did find that maternal conception weight was significantly influenced by the number of helpers that raised their previous litter. Mothers were lighter at the conception of their subsequent litter when their previous litter had been raised by few helpers, presumably because in such groups mothers contribute significantly more to babysitting (Clutton-Brock et al., 1998), pup-feeding (Clutton-Brock et al., 2001a), and possibly lactation (Scantlebury et al., 2002), although it may be caused, in part, by reduced foraging efficiency (Clutton-Brock et al., 1999). Thus, because in meerkats a mother's weight appears to influence her fecundity and investment, and because her weight is affected by the number of helpers she had in her group during the previous breeding attempt, helper number will affect both the number of offspring born and their weight at weaning indirectly. Furthermore, becasue the number of pups born and their weights at weaning are both significantly correlated with the number of pups present and their weights at independence (Russell et al., 2002), the indirect effects of helper number on female fecundity and investment will have long-term consequences (Clutton-Brock et al., 2001c, 2002; see also Solomon, 1994).

In conclusion, our results support the general contention that helpers influence female fecundity by allowing them to have more reproductive attempts per season. Moreover, however, we show that helpers may also be capable of influencing female fecundity and breeding success early within reproductive attempts. Two findings support this claim. First, helper number appeared to be associated with pup survival between birth and weaning. Second, helper number during the previous event was associated with maternal weight at their conception in the subsequent event, and maternal conception weight was positively associated with litter sizes at birth and mean pup weights at weaning. These results illustrate the importance of understanding helper effects preweaning (or prehatching), even if primary contributions by helpers appear to occur after this time. For instance, if helpers are associated with differences in maternal investment, either directly or indirectly, then helper effects may either be obscured or inflated depending on whether or not helpers are associated with increases or decreases in maternal investment. We therefore suggest that studies should be careful not to reject or to overinterpret the importance of helpers for reproductive success without controlling for helper effects on female fecundity and investment during early offspring rearing.

Figure 1

Relationship between the number of helpers and the number of litters produced by dominant females each year. Dominant females produced significantly more litters per year when in the presence of many helpers (GLM: F1,47 = 18.16, p <.001), explaining 26% of the overall variation. Analysis was conducted on 121 births, by 20 females over one to six seasons in 13 groups

Figure 1

Relationship between the number of helpers and the number of litters produced by dominant females each year. Dominant females produced significantly more litters per year when in the presence of many helpers (GLM: F1,47 = 18.16, p <.001), explaining 26% of the overall variation. Analysis was conducted on 121 births, by 20 females over one to six seasons in 13 groups

Figure 2

Relationship between the number of individuals helping to raise one litter and the weight of females during the conception of the litter immediately subsequent to this litter. Analysis was conducted on 169 conception weights of 60 mothers in 13 groups. Maternal weight was fitted to the analysis as a residual value to correct for the significant effects of maternal age on maternal weight

Figure 2

Relationship between the number of individuals helping to raise one litter and the weight of females during the conception of the litter immediately subsequent to this litter. Analysis was conducted on 169 conception weights of 60 mothers in 13 groups. Maternal weight was fitted to the analysis as a residual value to correct for the significant effects of maternal age on maternal weight

Table 1

General linear model of the factors associated with litter sizes at birth in meerkats.

Model term F df p 
Helper number 0.63 1,28 .43 
Maternal status 0.30 1,28 .85 
Maternal age 0.28 1,28 .60 
Maternal weight 11.30 1,28 .002 
Days before birth 0.01 1,28 .91 
Model term F df p 
Helper number 0.63 1,28 .43 
Maternal status 0.30 1,28 .85 
Maternal age 0.28 1,28 .60 
Maternal weight 11.30 1,28 .002 
Days before birth 0.01 1,28 .91 

Analysis was conducted on 35 litters scanned from 20 females in nine groups. Maternal weight accounted for 23% of the total variation.

Table 2

Generalized linear model of the factors associated offspring survival between birth and weaning in meerkats.

Model term F df p 
Helper number 2.76 1,32 .096 
Maternal status 17.61 1,32 <.001 
Maternal age 0.11 1,32 .74 
Maternal weight 0.02 1,28 .88 
Mean pup age 0.78 1,32 .38 
Days before birth 0.48 1,32 .49 
Model term F df p 
Helper number 2.76 1,32 .096 
Maternal status 17.61 1,32 <.001 
Maternal age 0.11 1,32 .74 
Maternal weight 0.02 1,28 .88 
Mean pup age 0.78 1,32 .38 
Days before birth 0.48 1,32 .49 

Analysis was conducted on the same 35 litters scanned in the ultrasound analysis (see Table 1). Maternal dominance status accounts for 16% of the variation.

Table 3

Generalized linear model and general linear model, respectively, of the factors associated with (top) the probability that at least one meerkat pup from a litter survived between birth and weaning, and (bottom) the size of litters at weaning in which at least one meerkat pup survived.

Model term F df p 
Helper number 0.03 1,157 .86 
Maternal status 67.86 1,167 <.0001 
Maternal age 0.01 1,164 .93 
Maternal weight 0.06 1,153 .81 
Helper number 2.02 1,84 .022 
Maternal status 1.65 1,84 .22 
Maternal age 1.63 1,84 .21 
Maternal weight 19.32 1,84 <.001 
Model term F df p 
Helper number 0.03 1,157 .86 
Maternal status 67.86 1,167 <.0001 
Maternal age 0.01 1,164 .93 
Maternal weight 0.06 1,153 .81 
Helper number 2.02 1,84 .022 
Maternal status 1.65 1,84 .22 
Maternal age 1.63 1,84 .21 
Maternal weight 19.32 1,84 <.001 

Analyses were conducted on a sample of 169 litters born to 54 females in 10 groups. The final models explained 23% and 19% of the variation, respectively.

Table 4

General linear model of the factors affecting mean pup weights at weaning.

Model term F df p 
Helper number 2.28 1,29 .14 
Maternal status 0.14 1,29 .71 
Maternal age 0.11 1,29 .75 
Maternal weight 7.50 1,29 .010 
Litter age 3.65 1,29 .041 
Litter size 0.92 1,29 .34 
Litter sex ratio 0.01 1,29 .94 
Model term F df p 
Helper number 2.28 1,29 .14 
Maternal status 0.14 1,29 .71 
Maternal age 0.11 1,29 .75 
Maternal weight 7.50 1,29 .010 
Litter age 3.65 1,29 .041 
Litter size 0.92 1,29 .34 
Litter sex ratio 0.01 1,29 .94 

Analysis was conducted on 37 litters born to 17 mothers from nine groups. Maternal weight accounted for 22% of the overall variation.

We thank Mr. and Mrs. H. Kotze and the Northern Cape Conservation Authority for allowing this work to be carried out on their land, Martin Haupt, Johan du Toit, and Elmarie Cronje at the University of Pretoria and Penny Roth at Cambridge for logistical help with the project, and more than 40 volunteers and students who contributed to data-collection. This paper has benefited from the help, advice, and/or comments of Tim Coulson, Giacomo Tavecchia, Virpi Lummaa, Jason Gilchrist, and two anonymous referees. This study was funded by the Natural Environmental Research Council and Biotechnology and Biological Sciences Research Council, for which we are most grateful.

REFERENCES

Brotherton PNM, Clutton-Brock TH, O'Riain MJ, Gaynor D, Sharpe L, Kansky R, McIlrath GM,
2001
. Offspring food allocation by parents and helpers in a cooperative mammal.
Behav Ecol
 
12
:
590
-599.
Brown JL, Brown ER,
1981
. Kin selection and individual fitness in babblers. In: Natural selection and social behavior (Alexander RD, Tinkle DW, eds). New York: Chiron Press; 244–256.
Brown JL, Brown ER, Brown SD, Dow DD,
1982
. Helpers: effects of experimental removal on reproductive success.
Science
 
215
:
421
-422.
Christians JK,
2002
. Avian egg size: variation within species and inflexibility within individuals.
Biol Rev
 
77
:
1
-26.
Clutton-Brock TH,
1991
. Parental care. Princeton, New Jersey: Princeton University Press.
Clutton-Brock TH, Brotherton PNM, O'Riain MJ, Griffin AS, Gaynor D, Kansky R, Sharpe L, McIlrath GM,
2001
. Contributions to cooperative rearing in meerkats, Suricata suricatta.
Anim Behav
 
61
:
705
-710.
Clutton-Brock TH, Brotherton PNM, Russell AF, O'Riain MJ, Gaynor D, Kansky R, Griffin AS, Manser MB, Sharpe L, McIlrath GM, Small T, Moss A, Monfort S,
2001
. Cooperation, conflict, and concession in meerkats groups.
Science
 
291
:
478
-481.
Clutton-Brock TH, Gaynor D, Kansky R, MacColl ADC, McIlrath GM, Chadwick P, Brotherton PNM, O'Riain MJ, Manser MB, Skinner JD,
1998
. Costs of cooperative behaviour in suricates Suricata suricatta.
Proc R Soc Lond B
 
265
:
185
-190.
Clutton-Brock TH, O'Riain JM, Brotherton PNM, Gaynor D, Kansky R, Griffin AS, Manser M,
1999
. Selfish sentinels in cooperative mammals.
Science
 
284
:
1640
-1644.
Clutton-Brock TH, Russell AF, Sharpe LL, Brotherton PNM, McIlrath GM, White S, Cameron EZ,
2001
. Effects of helpers on juvenile development and survival in meerkats.
Science
 
293
:
2446
-2449.
Clutton-Brock TH, Russell AF, Sharpe LL, Young AJ, Balmforth Z, McIlrath GM,
2002
. Evolution and development of sex differences in cooperative behavior in meerkats.
Science
 
297
:
253
-256.
Cockburn A,
1998
. Evolution of helping behavior in cooperatively breeding birds.
Annu Rev Ecol Syst
 
29
:
141
-177.
Creel SR, Monfort SL, Wildt DE, Waser PM,
1991
. Spontaneous lactation is an adaptive result of pseudopregnancy.
Nature
 
351
:
660
-662.
Crick HQP,
1992
. Load-lightening in cooperatively breeding birds.
Ibis
 
134
:
56
-61.
Cunningham EJA, Russell AF,
2000
. Egg investment is influenced by male attractiveness in mallards.
Nature
 
404
:
74
-77.
Davies NB, Hatchwell BJ,
1992
. The value of male parental care and its influence on reproductive allocation by male and female dunnocks.
J Anim Ecol
 
61
:
259
-272.
Emlen ST,
1991
. Evolution of cooperative breeding in birds and mammals. In: Behavioural ecology: an evolutionary approach (Krebs JR, Davies NB, eds). Oxford: Blackwell Scientific; 301–337.
Emlen ST, Wrege PH,
1991
. Breeding biology of white-fronted bee-eaters at Nakuru: the influence of helpers on breeder fitness.
J Anim Ecol
 
60
:
309
-326.
GENSTAT,,
1993
. GENSTAT 5: release 3 reference manual. Oxford: Clarendon Press.
Hatchwell BJ,
1999
. Investment strategies of breeders in avian cooperative breeding systems.
Am Nat
 
154
:
205
-219.
Hatchwell BJ, Russell AF,
1996
. Provisioning rules in cooperatively breeding long-tailed tits Aegithalos caudatus: an experimental study.
Proc R Soc Lond B
 
263
:
83
-88.
Hoogland JL,
1995
. The black-tailed prairie dog. Chicago: University of Chicago Press.
Jennions MD, MacDonald DW,
1994
. Cooperative breeding in mammals.
Trends Ecol Evol
 
9
:
89
-93.
Mumme RL,
1992
. Do helpers increase reproductive success: an experimental analysis in the Florida scrub jay.
Behav Ecol Sociobiol
 
31
:
319
-328.
Perrins CM,
1971
. British tits. London: Collins.
Petrie M, Williams A,
1993
. Peahens lay more eggs for peacocks with larger trains.
Proc R Soc Lond B
 
251
:
127
-131.
Powell RA, Fried JJ,
1992
. Helping by juvenile pine voles Microtus pinetorum, growth and survival of younger siblings, and the evolution of pine vole sociality.
Behav Ecol
 
3
:
325
-333.
Russell AF, Clutton-Brock TH, Brotherton PNM, Sharpe LL, McIlrath GM, Dalerum FD, Cameron EZ, Barnard JA,
2002
. Factors affecting pup growth and survival in cooperatively breeding meerkats Suricata suricatta.
J Anim Ecol
 
71
:
700
-709.
Russell AF, Sharpe LL, Brotherton PN, Clutton-Brock TH,
2003
. Cost minimization by helpers in cooperative vertebrates.
Proc Natl Acad Sci USA
 
100
:
3333
-3338.
Scantlebury M, Russell AF, McIlrath GM, Speakman JR, Clutton-Brock TH,
2002
. The energetics of lactation in cooperatively breeding meerkats Suricata suricatta. Proc R Soc Lond B 269:2147–2153.
Schall R,
1991
. Estimation in generalized linear models with random effects.
Biometrika
 
78
:
719
-727.
Solomon NG,
1991
. Current indirect fitness benefits associated with philopatry in juvenile prairie voles.
Behav Ecol Sociobiol
 
29
:
277
-282.
Solomon NG,
1994
. Effects of pre-weaning environment on subsequent reproduction in prairie voles, Microtus ochrogaster.
Anim Behav
 
48
:
331
-341.
Zahavi A,
1990
. The quest for social status in a cooperative breeder. In: Cooperative breeding in birds (Stacey PB, Koenig WD eds). Cambridge: Cambridge University Press; 102–130.