Mammalian Biology 93 (2018) 1–4

Contents lists available at ScienceDirect

Mammalian Biology
journal homepage: www.elsevier.com/locate/mambio

Short communication

Suckling and allosuckling behavior in wild giraffe

(Giraffa camelopardalis tippelskirchi)
Miho Saito ∗ , Gen’ichi Idani
Wildlife Research Center of Kyoto University, 2-24 Tanaka-Sekiden-cho, Sakyo, Kyoto, 606-8203, Japan

a r t i c l e i n f o a b s t r a c t

Article history: Allosuckling has been reported in many mammalian species. In giraffe, allosuckling has been observed
Received 9 March 2018 in captivity, and the milk theft hypothesis with reciprocity is regarded as the likeliest explanation for
Accepted 10 July 2018 this behavior. However, reports of such behavior in the wild remain sparse to non-existent. Here, we
studied the suckling and allosuckling behaviors of three nursing giraffe and their offspring (two calves
Handled by Sabine Begall
and one juvenile) for 32 days in the Katavi National Park, Tanzania. In total, we observed 56 suckling
bouts, 96 suckling attempts, 5 allosuckling bouts, and 71 allosuckling attempts. We observed that the
Keywords:
Allosuckler
female decided when to nurse the offspring, as reported in previous studies; however, the suckling bouts
Milk theft of calves were terminated mainly by other individuals. On the other hand, all suckling bouts of juvenile
Weaning were terminated by the female. The milk theft hypothesis was supported since (1) the allosuckler always
joined a suckling pair and never succeeded when approaching a female by itself, (2) the female apparently
did not notice the non-filial offspring positioned behind the filial offspring, and (3) the female showed
active signs of rejection when she noticed the allosuckler. In addition, we found that juvenile close to
weaning showed the highest rate of allosuckling interactions. Therefore, we assume that the presence of
a weaning individual might drive the occurrence of allosuckling in giraffe in the wild.
© 2018 Deutsche Gesellschaft für Säugetierkunde. Published by Elsevier GmbH. All rights reserved.

Lactation incurs a high energy cost to females (Martin, 1984). milk evacuation, a female releases surplus milk that her own off-
However, females sometimes nurse a non-filial offspring, i.e., an spring did not consume (Roulin, 2002); social benefits, allonursing
allosuckler (Packer et al., 1992). This behavior is called allosuckling females gain stronger social bonds (Baldovino and Di Bitetti, 2008)
or allonursing, and it is considered as an extreme case of allo- and milk theft, the suckling by non-filial calves, usually with the
parental care in mammals (Riedman, 1982). Allosuckling has been filial calf rather than alone (Brandlová et al., 2013; Gloneková et al.,
reported in many ungulate species both in captivity and in the 2016, 2017; Packer et al., 1992) and with non-filial offspring suck-
wild; examples of the former include Plains zebra (Equus burchel- ling from the lateral position (Brandlová et al., 2013). In addition,
lii; Pluháček et al., 2011) and guanaco (Lama guanicoe; Zapata MacLeod and Lukas (2014) found that allonursing occurs more
et al., 2009a), and for the latter, Cape mountain zebra (E. zebra often when the cost of allonursing an additional offspring is low.
zebra; Penzhorn, 1984) and Guanaco (Zapata et al., 2009b). Sev- Giraffe (Giraffa camelopardalis) are reported to allosuckle when
eral hypotheses have been proposed to explain this phenomenon they are in captivity (Gloneková et al., 2016, 2017) and also in the
as either adaptive or non-adaptive for nursing females: kin selec- wild (Pratt and Anderson, 1979). In captivity, allosuckling occurred
tion, when the involved female and non-filial offspring share genes at the highest rate among mammals. The non-filial calves more
by common descent (Packer et al., 1992); parenting, so females can often allosuckled together with the filial ones than alone and tried
improve their maternal skills (Roulin, 2002); reciprocity, wherein to adopt positions where they may be harder to recognize and this
two involved females achieve a higher fitness when nursing each supported the “milk theft” hypothesis as well as the reciprocity
other’s offspring to a similar extent than when they do not share (Gloneková et al., 2016, 2017). In contrast, Pratt and Anderson
milk (Packer et al., 1992); misdirected parental care, the inadver- (1979) recorded only one case of allosuckling out of 37 suckling
tent transfer of milk to nonrelated offspring (Packer et al., 1992); attempts, and Gloneková et al. (2017) recorded no allosuckling
event while observing giraffe in a nature reserve for 22 days. These
results prompted them to conclude that female giraffe in the wild
∗ Corresponding author. nurse only filial offspring.
E-mail addresses: saito.miho.48u@st.kyoto-u.ac.jp (M. Saito), Here, we report in detail the occurrence of suckling and allo-
gidani@wrc.kyoto-u.ac.jp (G. Idani). suckling behavior in wild giraffe (G. c. tippelskirchi) and document

https://doi.org/10.1016/j.mambio.2018.07.005
1616-5047/© 2018 Deutsche Gesellschaft für Säugetierkunde. Published by Elsevier GmbH. All rights reserved.
2 M. Saito, G. Idani / Mammalian Biology 93 (2018) 1–4

the conditions of suckling which may influence the occurrence of Table 1

Observation hours for each dyad (upper row of each dyad) with the number of all
allosuckling. Even though our sample size was limited to three
suckling/allosuckling interactions, the number of successful suckling/allosuckling
female-offspring dyads, we also discuss the behavioral evidence for in brackets. The female–filial offspring dyads are indicated by a gray background,
the parenting, reciprocity, misdirected parental care, milk evacua- and successful allosuckling bouts were underlined.
tion, social benefits, and milk theft hypothesis.
Our study was conducted from October through November 2017
in Katavi National Park, Tanzania, in which miombo woodlands
form the major vegetation (Waltert et al., 2008). We collected data
from approximately 07:00–18:00. We conducted our study on foot
with an armed ranger. In the park where our study was conducted,
the living quarters of the park staff were located near our study area,
and the staff normally walked around on foot. Therefore, the giraffe
living in this area were more habituated to people than those living
outside of this area. Typically, the distance between the observer
(M.S.) and animals was 15–50 m, depending on the vegetation and suckled for at least 5 s. A suckling attempt (unsuccessful suck-
terrain. Since Leuthold (1979) and Pratt and Anderson (1979) noted ling/allosuckling) was defined as when the female starts to move
that nursing females often have their calf within 200-m, all individ- away before the offspring could reach the female’s nipple with its
uals within approximately this area were presumed to belong to the mouth or when she starts to move away within 5 s of initiating
same herd. The estimated age of all individuals were grouped as fol- the suckling by the offspring. An interaction event included both
lows: (1) Calf: <6 months. (2) Juvenile: 6–18 months. (3) Sub-adult: bouts and attempts. In our observations, we recorded the follow-
18 months to 4 years. (4) Adult: >4 years (Caister et al., 2003; Le ing data when they occurred: all suckling/allosuckling interactions,
Pendu and Ciofolo, 1999) based on giraffe body size, color, and the their success or failure, duration of a suckling bout, the initiator
shapes of the ossicones. We were able to identify each individual by (suckler or female; female approaches the offspring to within 3 m,
their unique pelage pattern. In total, 65 individuals (28 adult males, waits till the offspring comes close to her, and sniffs the offspring or
20 adult females, 10 sub-adult males, 3 sub-adult females, 2 juve- female approaches and sniffs the offspring), identity of the nursing
niles and 2 calves) were identified in 2017. Two female–calf dyads female, and the terminator of the suckling bout (suckler, female, or
(dyad 1 as F1 and C1, dyad 2 as F2 and C2) and one female–juvenile others).
dyad (dyad 3 as F3 and J3) were identified during our observations. The Steel–Dwass test was used for post hoc comparisons of
Female–filial offspring dyads are referred to using their respective suckling duration among the three offspring. This test was chosen
numerical identifier in the text and tables. In fact, we conducted our because Kruskal–Wallis test was significant for suckling duration
field study in 2010, 2011 and 2016, and these three females were among the three offspring (df = 2, p < 0.05). Ryan’s test for mul-
monitored since 2010. They had been observed with their offspring tiple comparisons was performed to assess the differences in (1)
in previous study periods; F1 was observed with a juvenile in 2016 the rate of suckling bouts per suckling interaction among the three
and F2 and F3 were each observed with a calf in 2010. These three offspring, (2) the number of allosuckling interactions per observa-
female-offspring dyads in 2017 often formed a nursery group, and tion time among the three offspring. Fisher’s exact test was used
no other female-offspring dyad was found in this nursery group. to examine the combination of two suckling conditions (bouts or
Sometimes another individual was found in the area where the attempts) in relation to the combination of the initiator. A G–test
nursery group was present, but after few hours they usually moved with the William’s correction was used to test for deviation from
away and got separated from the nursery group. However, there random occurrence of a terminator of suckling bouts among each
was an exception that F1 was followed by a particular male giraffe category. We report the mean ± SE in the text. All statistical calcula-
(named M1) for almost all the time, and they were observed in the tions were performed using the R statistical package (R Core Team,
same herd for 28 out of 29 observational days of F1. Their mating 2016), with a two-tailed significance level set at 0.05.
was observed on October 24 and November 7, respectively. With respect to suckling interactions, we observed 56 suckling
Since we also conducted another study from June through bouts and 96 suckling attempts. Suckling durations of the three off-
November 2016 and from May through September 2017, we could spring were on average as follows: C1, 18 ± 3 s (n = 37, min: 4 s, max:
estimate the birthdate of the three offspring in the dyads. To esti- 90 s); C2, 22 ± 5 s (n = 14, min: 5 s, max: 72 s); and J3, 40 ± 5 s (n = 5,
mate the offspring birthdate, we used the time between the last min: 26 s, max: 51 s), and C1 suckled for a significantly shorter dura-
date when a female was observed without her offspring and the tion when compared with J3 (Steel-Dwass test: C1 vs C2, t = 0.92, p =
first date she appeared with her offspring: for C1, 19–24 August; 0.63; C1 vs J3, t = 2.82, p < 0.05; C2 vs J3, t = 2.09, p = 0.09). Table 1
for C2, 5–20 July; for C3: 12–28 August 2016. Therefore, the esti- summarizes the frequency of suckling/allosuckling interactions per
mated ages of the three offspring when we conducted our study hour for each offspring. Rate of suckling bouts per suckling interac-
of suckling and allosuckling (i.e., October through November 2017) tion was significantly higher in C1 than in the other two (C1: 68%,
were as follows: C1, 2–3 months; C2, 3.5–4.5 months; J3, 14–15 C2: 23%, J3: 15%; Ryan’s test: C1 vs C2: ␣ = 0.03, p < 0.05; C1 vs J3:
months. Weaning may cease before reaching 1 year of age, but ␣ = 0.02, p < 0.05). The females initiated significantly more often
sometimes suckling may continue for up to 1.5–2 years (Dagg and suckling bouts than suckling attempts (Fig. 1a. Fisher’s exact test:
Foster, 1976; Langman, 1977). In addition, females are known to C1: p = 0.05; C2: p < 0.01; J3: p = 0.01). Additionally, even though
become pregnant while nursing (Nakamichi et al., 2015). the females approached their filial offspring, other individuals, i.e.
As Pratt and Anderson (1979) reported, giraffe calves and juve- M1 and J3, approached the suckling pair before the filial offspring
nile stay close to each other. Once located, we conducted a focal started to suckle, and the suckling failed. This was the main rea-
animal sampling of the offspring (Altmann, 1974). The duration of son why the suckling interactions initiated by females ended up as
observations (hours) of all dyads were shown in Table 1. We were an attempt at a high rate. The suckling bouts of calves were sig-
able to find the three offspring almost every research day (i.e., C1 nificantly more often terminated by others than by the females or
and J3 for 29 days, C2 for 28 out of 32 research days). The maximum the sucklers (Fig. 1b. G-test: C1: G = 45.1, df = 2, p < 0.01, C2: G = 8.6,
continuous observation time of a single session was approximately df = 2, p < 0.01). Female 1 was never observed to terminate a suck-
4 h. We defined a suckling bout (successful suckling/allosuckling) ling bout. Female 2 terminated 1 suckling bout after a duration of
as when the offspring took the female’s nipple into its mouth and 72 s. The suckling bouts of J3 were terminated only by female 3.
 M. Saito, G. Idani / Mammalian Biology 93 (2018) 1–4 3

the duration of suckling bouts of the calves were lower than that
of young individuals in previous studies (45 s–4 min in Langman,
1977; average of 66 s in Pratt and Anderson, 1979; and >50 s in
Nakamichi et al., 2015). This might be caused by the conditions
of termination. Both Pratt and Anderson (1979) and Nakamichi
et al. (2015) mentioned that female giraffe are very likely to con-
trol the nursing-related activities; e.g. initiation and termination.
Our results also showed that the initiator of suckling bouts was the
female, and so as long as the suckling bout went uninterrupted by
others, the female was the one to terminate it, much like female
3 did in our study. Nonetheless, in most of our cases, the suckling
bouts of calves were terminated by the other individual i.e. M1 and
J3. There was only one case wherein a suckling bout was termi-
nated by a nursing female (in dyad 2); the duration of the suckling
bout was significantly longer in this case than in those terminated
by others. Therefore, the termination of suckling bouts by others
might be the reason why the suckling duration of the calves was
shorter than that of the juvenile in our study as well as in previous
studies.
Allosuckling in captive giraffe occurs frequently (Gloneková
et al., 2016,2017. On the other hand, allosuckling has been rarely
documented in wild giraffe (Pratt and Anderson, 1979; Gloneková
et al., 2017). Here we recorded five allosuckling cases out of 76
allosuckling interactions in wild giraffe. As noted in previous stud-
ies (Pratt and Anderson, 1979; Gloneková et al., 2016, 2017), the
allosuckler always joined a suckling pair and never succeeds when
approaching a female by itself. This condition was also observed
in our study. Moreover, during the allosuckling bouts in this study,
Fig. 1. The frequency of initiator and terminator of each of the three offsprings’ the F1 apparently did not notice the non-filial offspring positioned
suckling sessions. (a) the initiator of suckling interactions. (SB = the suckling bouts, behind her filial offspring and female rejected allosuckler when she
SA = the suckling attempts, C1, SB: n = 37, SA: n = 18; C2, SB: n = 14, SA: n = 49; J3: SB: noticed its existence. Thus, these events appear to be cases of “milk
n = 5, SA: n = 29). (b) the terminator of suckling bouts (C1, n = 37; C2, n = 14; J3, n = 5).
theft” and are consistent with the view that females in the wild
rarely allow non-filial offspring to suckle, as documented in the
With respect to allosuckling interactions, we observed 5 allo- previous studies (Langman, 1977; Pratt and Anderson, 1979).
suckling bouts and 71 allosuckling attempts. The possible number In our study, most of the allosuckling interactions were observed
of dyads for allosuckling was six. Allosuckling bouts were observed in J3. Giraffe are weaned by the age of 1 year, but sometimes wean-
in female 1–C2 (once) and female 1–J3 (four times). Allosuckling ing may continue for up to 1.5–2 years (Dagg and Foster, 1976;
interactions were observed in all the offspring, but J3 showed the Langman, 1977). The estimated age of J3 was 14–15 months, so
highest number of attempts (C1: n = 5, C2: n = 2, J3: n = 69; Ryan’s she was close to weaning. It had the fewest suckling bouts but the
test: C1 vs J3: ␣ = 0.03, p < 0.01; C2 vs J3: ␣ = 0.02, p < 0.05; Table 1). highest number of allosuckling attempts. Therefore, this particular
The average duration of the allosuckling bouts was 7 ± 2 s (n = 5, juvenile may have tried to compensate for an insufficient supply of
min: 3 s, max: 16 s). A non-filial offspring individual was the initia- maternal milk by allosuckling from other lactating females. Hence,
tor of all such occurrences and it only joined after the filial calf had allosuckling in giraffe may occur simply as an incidental event by
begun to suckle. In two out of five successful allosuckling bouts, an individual that weans early.
the female apparently did not notice the allosuckler positioned on Since we did not conduct an mtDNA analysis to reveal the
the opposite side of her body from the filial offspring, or behind it, genetic relationships among the studied females, we could not test
because the female was bending her neck to sniff the filial offspring. the kin selection hypothesis. The other hypotheses that explain
After turning back her head, the female noticed the allosuckler allosuckling behavior seem not to apply to our observed cases. The
and displayed an active sign of rejection by moving her hind leg parenting hypothesis, which improves maternal skills, is inapplica-
and walking away. For three out of five cases of allosuckling we ble because all three females in our study had already experienced
could not identify whether female 1 noticed the existence of the nursing activities. With respect to the reciprocity hypothesis, in
allosuckler. Even though the allosuckler approached her from an which two females achieve a higher fitness by nursing each other
antiparallel position and its body was not hidden by her filial off- calves to the same extent, yet only female 1 was found to engage in
spring, she was standing still and her neck was in a straight position, allonursing. Misdirected parental care is also improbable because
even after the allosuckler started to put her nipple into its mouth when the allosuckler tried to allosuckle alone, the female always
(video 1). Allosuckling bouts were terminated by F1, which never rejected it. Social benefit is unlikely too, because no hierarchy
terminated the suckling bouts, or by others; for all cases, allosuck- exists among wild female giraffe and we never observed the three
lers were never observed to terminate these bouts. Individuals from females act aggressively toward each other. Finally, the milk evac-
other age-sex classes were never observed to participate in suck- uation hypothesis, in which a mother produces more milk than
ling. her offspring can consume, is not applicable. In our study, we
In general, the duration and frequency of nursing decreases did not observe a filial offspring terminating a sucking bout of its
with development (Clark and Odell, 1999), and this phenomenon own accord; therefore, the filial offspring may not have consumed
is observed in giraffe as well (Pratt and Anderson, 1979; Nakamichi enough milk.
et al., 2015). However, in our study, we found that even though the The frequency of allosuckling bouts in our study was low com-
suckling frequency was higher for the calves than for the juvenile, pared with that observed in captivity. Our study was conducted
the average suckling duration was the lowest for C1. In addition, during the beginning of a short rainy season; therefore, we assume
4 M. Saito, G. Idani / Mammalian Biology 93 (2018) 1–4

the quality/amount of food was still insufficient at this time. Baldovino, C.M., Di Bitetti, S.M., 2008. Allonursing in tufted capuchin monkeys
Allonursing occurs when the costs of allonurse become low, for (Cebus nigritus): milk or pacifier? Folia Primatol. 79, 79–92.
Brandlová, K., Bartoš, L., Haberová, T., 2013. Camel calves as opportunistic milk
example, under conditions where females have access to unlim- thefts? The first description of allosuckling in domestic bactrian camel
ited food (Packer et al., 1992; MacLeod and Lukas, 2014), which (Camelus bactrianus). PLoS One 8 (1), e53052, http://dx.doi.org/10.1371/
could explain the greater occurrence of allosuckling in captivity. journal.pone.0053052.
Caister, L., Shields, W.M., Gosser, A., 2003. Female tannin avoidance: a possible
Additionally, the visibility in captive settings is likely better than explanation for habitat and dietary segregation of giraffes (Giraffa
in our study area, where woodlands are dominated and the area Camelopardalis peralta) in Niger. Afr. J. Ecol. 41, 201–210.
of enclosure is limited. Therefore, in captivity, it might be easy Clark, S.T., Odell, D.K., 1999. Nursing papameters in captive killer whales (Orcinus
orca). Zoo Biol. 18, 373–384.
for an allosuckler to notice the ongoing nursing session and try
Dagg, A.I., Foster, J.B., 1976. The Giraffe: Its Biology, Behavior, and Ecology. Krieger
to participate in it compared with that in the wild environment. Publishing Company, Florida.
In conclusion, allosuckling in wild giraffe is an uncommon Gloneková, M., Brandlová, K., Pluháček, J., 2016. Stealing milk by young and
reciprocal mothers: high incidence of allonursing in giraffes, Giraffa
behavior that seems to be opportunistic and may be explained by
camelopardalis. Anim. Behav. 113, 113–123.
the “milk theft” hypothesis. Nevertheless, our dataset was based on Gloneková, M., Vymyslická, P.J., Žáčková, M., Brandlová, K., 2017. Giraffe nursing
only three pairs of female-offspring dyads and therefore additional behaviour reflects environmental conditions. Behaviour 154, 115–129.
data is required to support this interpretation and the suggested Langman, V.A., 1977. Cow-calf relationships in giraffe (Giraffa camelopardalis
giraffe). Z. Tierpsychol. 43, 264–286.
explanation. Le Pendu, Y., Ciofolo, I., 1999. Seasonal movements of giraffes in Niger. J. Trop. Ecol.
15, 341–353.
Acknowledgements Leuthold, M.B., 1979. Social organization and behaviour of giraffe in Tsavo East
National Park. Afr. J. Ecol. 17, 19–34.
MacLeod, K.L., Lukas, D., 2014. Revisiting non-offspring nursing: allonursing
The research was financially supported by Leading Gradu- evolves when the costs are low. Biol. Lett. 10, 20140378.
ate Program in Primatology and Wildlife Science (U04-JSPS) and Martin, P., 1984. The meaning of weaning. Anim. Behav. 32, 1257–1259.
Nakamichi, M., Murata, C., Eto, R., Takagi, N., Yamada, K., 2015. Daytime
Grant-in-Aid for JSPS Research Fellow (#17J09785). We thank mother-calf relationships in reticulated giraffes (Giraffa camelopardalis
the Tanzania Commission for Science and Technology (COSTECH), reticulate) at the Kyoto City Zoo. Zoo Biol. 34, 110–117.
the Tanzania Wildlife Research Institute (TAWIRI) and Tanzania Packer, C., Lewis, S., Pusey, A., 1992. A comparative analysis of non-offspring
nursing. Anim. Behav. 43, 265–281.
National Parks (TANAPA) for granting permissions. People of the Penzhorn, B.L., 1984. A long-term study of social organization and behaviour of
Katavi National Park, especially P. Karera, are acknowledged for Cape mountain zebra Equus zebra zebra. Ethology 64, 97–146.
their support during field work. We are grateful to F. Bercovitch and Pluháček, J., Bartošová, J., Bartoš, L., 2011. A case of adaption and allonursing in
captive plains zebra (Equus burchellii). Behav. Process. 86, 174–177.
two anonymous reviewers for valuable comments on this article.
Pratt, D.M., Anderson, V.H., 1979. Giraffe cow-calf relationships and social
development of the calf in the Serengeti. Z. Tierpsychol. 51, 233–251.
Appendix A. Supplementary data R Core Team, 2016. R: A Language and Environment for Statistical Computing. R
Foundation for Statistical Computing, Vienna, Austria.
Riedman, M.L., 1982. The evolution of alloparental care and adoption in mammals
Supplementary material related to this article can be found, in and birds. Q. Rev. Biol. 57, 405–435.
the online version, at doi:https://doi.org/10.1016/j.mambio.2018. Roulin, A., 2002. Why do lactating females nurse alien offspring? A review of
hypotheses and empirical evidence. Anim. Behav. 63, 201–208.
07.005.
Waltert, M., Meyer, B., Shanyangi, W.M., Balozi, J.J., Kitwara, O., Qolli, S., Krischke,
H., Muhlenberg, M., 2008. Foot surveys of large mammals in woodlands of
References western Tanzania. J. Wildl. Manage. 72, 603–610.
Zapata, B., Gonzalez, A.B., Ebensperger, A.L., 2009a. Allonursing in captive
guanacos, Lama guanicoe: milk theft or misdirected parental care? Ethology
Altmann, J., 1974. Observational study of behavior: sampling methods. Behaviour
115, 731–737.
49, 227–265.
Zapata, B., Gaete, G., Correa, A.L., Gonzalez, A.B., Ebensperger, A.L., 2009b. A case of
allosuckling in wild guanacos (Lama guanicoe). J. Ethol. 27, 295–297.