Variation of Loranthaceae impact on Vitellaria paradoxa C. F. Gaertn. fruit yield in contrasting habitats and implications for its conservation

Abstract

Introduction. Shea tree (V. paradoxa C.F. Gaertn.), a species endemic to the Sudanian savanna woodlands, is dominant in the parklands of West Africa where it is of great socioeconomic importance. However, shea tree has been reported in recent decades to be threatened by plant parasites, Loranthaceae. Our study aimed to assess possible variation of the impact of these parasites on shea tree fruit yield in two contrasting habitats. Materials and methods. We selected 41 weakly and 41 heavily infected shea tree individuals, of similar size, in a protected area as well as in its adjacent parklands. Shea tree traits such as diameter at breast height, canopy diameter, tree height, canopy height, number of fruit yielded, number of parasite stumps per tree and an impact index ratio were assessed on each shea tree individual. Two-way ANOVA was performed to compare parasite impact on shea tree fruit yield in relation to habitat. Hierarchical cluster, canonical discriminant and one-way ANOVA analyses were used to show quantitative traits that characterize shea tree groups from habitats. Results. Loranthaceae did not reduce fruit yield significantly either in the parklands or in the protected area. Quantitative traits tended to discriminate all pooled shea trees in relation to habitats. Shea tree individuals in parklands were characterized mostly by the highest value of number of infected stumps per tree and of the impact index ratio, suggesting that many shea tree individuals in parklands were sensitive to Loranthaceae impact on their fruit yield. Conclusion. These findings were helpful for implementing some shea tree conservation plans.

Full text

Original article
Fruits, vol. 68 (2) 109
Variation of Loranthaceae impact on Vitellaria paradoxa C. F. Gaertn. fruit
yield in contrasting habitats and implications for its conservation.
Abstract – Introduction. Shea tree (V. paradoxa C.F. Gaertn.), a species endemic to the Sudanian
savanna woodlands, is dominant in the parklands of West Africa where it is of great socioeconomic
importance. However, shea tree has been reported in recent decades to be threatened by plant
parasites, Loranthaceae. Our study aimed to assess possible variation of the impact of these parasites
on shea tree fruit yield in two contrasting habitats. Materials and methods. We selected 41 weakly
and 41 heavily infected shea tree individuals, of similar size, in a protected area as well as in its adjacent
parklands. Shea tree traits such as diameter at breast height, canopy diameter, tree height, canopy
height, number of fruit yielded, number of parasite stumps per tree and an impact index ratio were
assessed on each shea tree individual. Two-way ANOVA was performed to compare parasite impact
on shea tree fruit yield in relation to habitat. Hierarchical cluster, canonical discriminant and one-way
ANOVA analyses were used to show quantitative traits that characterize shea tree groups from habitats.
Results. Loranthaceae did not reduce fruit yield significantly either in the parklands or in the protected
area. Quantitative traits tended to discriminate all pooled shea trees in relation to habitats. Shea tree
individuals in parklands were characterized mostly by the highest value of number of infected stumps
per tree and of the impact index ratio, suggesting that many shea tree individuals in parklands were
sensitive to Loranthaceae impact on their fruit yield. Conclusion. These findings were helpful for
implementing some shea tree conservation plans.
Benin / Vitellaria paradoxa / fruits / yield / Loranthaceae / parasitism /
colonizing ability / habitats
Variation de l'impact des loranthacées sur le rendement en fruits de
Vitellaria paradoxa C.F. Gaertn. dans des habitats différentiés et implications
pour sa conservation.
Résumé –Introduction. Le karité (V. paradoxa CF Gaertn.), une espèce endémique des forêts de
la savane soudanienne, prédomine dans les parcs d’Afrique de l’Ouest où la plante a une grande
importance socio-économique. Cependant, au cours des dernières décennies, le karité s’est révélé
menacé par des loranthacées, parasites de végétaux. Notre étude a cherché à comparer l’incidence
de ces parasites sur le rendement en fruits d’arbres de karité dans deux habitats différentiés. Matériel
et méthodes. Nous avons sélectionné 41 individus de karité faiblement infectés et 41 individus
fortement infectés, de taille similaire, dans deux habitats : une zone de végétation protégée et des
parcelles adjacentes non protégées. Certaines caractéristiques des karités comme le diamètre des
troncs à hauteur de poitrine, le diamètre de la canopée, la hauteur des arbres, la hauteur de la canopée,
le nombre de fruits récoltés, le nombre de souches de parasites par arbre et un indice calculé évaluant
l’impact du parasite sur la production de l’arbre ont été mesurés sur chaque individu de karité
sélectionné. Une ANOVA à deux voies a été réalisée pour comparer l’impact des parasites sur le
rendement en fruits des karités selon l’habitat. Des analyses de classification hiérarchique, canonique
discriminante et ANOVA unidirectionnelle ont été utilisées pour mettre en évidence les caractères
quantitatifs caractérisant les groupes de karité dans les habitats. Résultats. L’impact des loranthacées
sur le rendement en fruits des arbres n’a pas été significativement différent dans les deux habitats
considérés. Les caractères quantitatifs ont eu tendance à discriminer des groupes d’arbres de karité
dans les deux habitats. Les individus de karité des parcelles non protégées ont été principalement
caractérisés par un plus grand nombre de souches de plantes parasites par arbre et par un indice
d’impact sur le rendement plus fort, ce qui suggère que, dans ces parcelles, de plus nombreux plants
de karité ont été sensibles à l’impact des loranthacées sur leur rendement en fruits. Conclusion. Ces
résultats ont été utiles pour mettre en œuvre des plans de conservation des arbres de karité.
Bénin / Vitellaria paradoxa / fruits / rendement / Loranthaceae / parasitisme /
aptitude à coloniser / habitat
1Lab. Appl. Ecol., Fac. Agron.
Sci., Univ. Abomey Calavi,
01 BP 526, Cotonou, Benin,
houehanout@yahoo.fr
2Mo. Bot. Garden,
P.O. Box 299, St. Louis,
MO 63166-0299, U.S.A.
Variation of Loranthaceae impact on Vitellaria paradoxa C. F.
Gaertn. fruit yield in contrasting habitats and implications
for its conservation
Thierry D. HOUEHANOU1*, Valentin KINDOMIHOU1, Tariq STEVART2, Brice TENTE1, Marcel HOUINATO1, Brice SINSIN1
* Correspondence and reprints
Received 18 January 2012
Accepted 3 April 2012
Fruits, 2013, vol. 68, p. 109–120
© 2013 Cirad/EDP Sciences
All rights reserved
DOI: 10.1051/fruits/2013057
www.fruits-journal.org
RESUMEN ESPAÑOL,p.120
Article published by EDP Sciences

110 Fruits, vol. 68 (2)
T.D. Houehanou et al.
1. Introduction
Vitellaria paradoxa C. F. Gaertn. (Sapota-
ceae),the shea tree, is endemic to wooded
Sudanian savannas of Africa, spreading from
Senegal to Uganda [1]. Local people enjoy
many products made from this species,
which is a dominant component of agricul-
tural systems in Africa [2,3]. Fruits of shea
tree are rich in nutrients. Its almonds are
transformed into butter, which is highly pre-
ferred by many rural people in Africa [4].
This butter is also used for chocolate man-
ufacturing and in the cosmetic industry [5,6].
Nevertheless, shea trees were recently
reported to be highly threatened by Loran-
thaceae, hemiparasitic plants [7]. According
to these authors, 95% of the shea tree pop-
ulations of Burkina Faso and Mali were
infected by Loranthaceae. These parasites
are strongly involved in the high mortality
rate of shea trees observed in these coun-
tries [8,7]. Moreover, referring to the impact
of Loranthaceae species on shea tree hosts,
these parasites start with a decrease in the
growth vigor of the parasitized branch and
gradually affect the seed yields [9]. There-
fore, shea tree conservation against Loran-
thaceae has become important and much
research [10–12] based on shea tree Loran-
thaceae has been undertaken.
There are different ways (for instance,
quantifying variation of quantitative traits,
studying genetic diversity, etc.) to define
and apply conservation strategies of tree
species. Thus, a good understanding of var-
iation of species traits is necessary for con-
serving and sustainably managing a tree
species [13–15]. Therefore, variation of shea
tree traits has been addressed in relation to
anthropogenic pressure [2,16,17] and geo-
graphical position [15,18]. It has been
shown that shea trees are more productive
in parklands in comparison with natural
savannas [16,17,19,20].
Regarding Loranthaceae impact on host
fruit production, some findings [21,22]
showed that host trees that carried a greater
load of parasites produced lower fruit
yields. As far as shea tree is concerned, it
was previously established [23] that infected
shea tree individuals carried greater parasite
load in parklands compared with those in
protected areas. Consequently, greater var-
iation of Loranthaceae impact on shea tree
fruit yield in parklands is expected com-
pared with protected areas. It has also been
shown that prevalence of shea tree infection
by Loranthaceae (i.e., shea tree infection
rate) was higher in parklands than in pro-
tected areas [23]. In that respect also, higher
prevalence of Loranthaceae impact on shea
tree fruit yield in parklands is expected com-
pared with protected areas.
Moreover, previous studies focusing on
shea tree Loranthaceae in West Africa were
related to Loranthaceae diversity and con-
trol [10,11], Loranthaceae distribution [10],
Loranthaceae impact on flowering and fruit-
ing behavior of infected shea tree branches
[12], and shea tree protection against Loran-
thaceae infection [23]. However, variation of
Loranthaceae impact on shea tree fruit yield
is seldom assessed in relation to anthropo-
genic pressure and should be helpful for
implementing in situ shea tree conservation
plans. Therefore, our study aimed to assess
such variation in addressing two hypothe-
ses: (1) Loranthaceae impact on shea tree
fruit yield is significantly higher in park-
lands than in protected areas, and
(2) prevalence of Loranthaceae impact on
shea tree fruit yield is higher in parklands
than in protected areas.
2. Materials and methods
2.1. Study species
Our study focused on Vitellaria paradoxa
C.F. Gaertn. (Sapotaceae) as the host tree of
Loranthaceae, plant parasites that live on
that host in the study area. Vitellaria para-
doxa is a tree species often reaching over
20 m in height and 50 cm diameter at breast
height. Within this species, two subspecies
have been distinguished, Vitellaria para-
doxa subsp. paradoxa and V. paradoxa
subsp. nilotica. Although it may sometimes
be difficult to differentiate between them,
most reports emphasized the occurrence of
V. paradoxa subsp. paradoxa in West Afri-
can countries, while the second subspecies

Loranthaceae impact on V. paradoxa fruit yield
Fruits, vol. 68 (2) 111
is known to occur in East Africa [15]. As far
as Loranthaceae are concerned, three spe-
cies occur in the study area. They are
Agelanthus dodoneifolius (DC.) Polh. &
Wiens, Tapinanthus globiferus (A. Rich.)
Van Tieghem and Tapinanthus ophiodes
(Sprague) Danser. Agelanthus dodoneifo-
lius is the most widespread on shea tree in
the study area.
2.2. Study area
The study was carried out in the Pendjari
Biosphere Reserve and in its adjacent
surrounding agroforestry parklands. The
reserve covers 4666.4 km2and is composed
of the Pendjari National Park (2660.4 km2),
the Pendjari hunting zone (1750 km2) and
the Konkombri hunting zone (251 km2). It
is located in the Sudanian zone of northern
Benin (10°40’–11°28’ N and 0°57’–2°10’ E)
in West Africa. In the protected area,
logging and agricultural activities are strictly
prohibited. Conversely, the surrounding
land areas are dominated by agroforestry
parklands that are built around selection of
desirable trees, with shea tree as a dominant
species. The climate is tropical with an aver-
age annual, unimodal rainfall of 1100 mm.
Monthly mean temperatures range from
19 °C to 34 °C. Annual potential evapotran-
spiration is about 1500 mm. Monthly mean
values of relative moisture range from 25%
to 85%. The rainy season begins in April,
followed by a dry season from November
to March [24]. The main soil type occurring
in the Pendjari Biosphere Reserve is tropical
ferruginous soil [25]. In this zone, local
people valorize shea tree fruits and nuts for
several purposes.
2.3. Sampling and data collection
During the shea tree ripening period, forty-
one weakly (mean number of infection
points equal to 1 and 3, respectively, in the
protected area and in the parklands) and
41 heavily (mean number of infection
points equal to 10 and 24, respectively, in
the protected area and in the parklands)
infected trees were selected in each habitat
(i.e., the protected area and surrounding
parklands). Overall, 164 shea tree individu-
als (82 trees in the reserve and 82 trees in
the surrounding parklands) were selected.
Tree sampling and data collection were car-
ried out from March to May, 2010. This
period corresponds yearly to the green fruit
phase of shea tree in the study area. Data
were collected in this shea tree stage of
immature green fruits in order to prevent
loss of fruit due to predation by wild ani-
mals in the reserve. During this stage, fruits
are not matured but are green [26,27]. The
rarity of healthy shea trees in the adjacent
parklands (due to the high infection rate of
shea tree in this area) [23] justified the
choice of weakly parasitized individuals in
the two investigated habitats instead of
healthy individuals.
Six traits were measured: diameter at
breast height, overall height of the shea
tree, crown diameter, crown height,
number of Loranthaceae stumps per shea
tree and number of fruit yielded per shea
tree. The number of fruit produced per shea
tree was evaluated by counting it. If the
number exceeded 300 fruits, twenty-five
percent of shea tree branches were sampled
and the result was extrapolated to the total
number of branches. In that case, the
branches were sampled by considering the
distribution of fruit on branches using find-
ings of previous studies [27,28]. Tree traits
such as diameter at breast height, overall
height of the tree, crown diameter and
crown height were similar between weakly
and heavily infected trees.
2.4. Data processing and analysis
Traits were compared statistically between
weakly and heavily infected trees in the pro-
tected area and in the parkland with Stu-
dent's independent ttest. The fruit yield and
Loranthaceae stump data were log-trans-
formed in order to normalize the distribu-
tion. To test variation of Loranthaceae
impact on shea tree fruit yield between the
protected area and parkland, a two-way
ANOVA was performed [29]. Student's ttest
was used to compare fruit yield of weakly
and heavily infected shea trees within each
habitat type (protected area and parklands)

112 Fruits, vol. 68 (2)
T.D. Houehanou et al.
and also to compare fruit yield of all weakly
and heavily infected shea trees.
To assess the difference in prevalence of
Loranthaceae impact on shea tree fruit yield,
an impact index ratio (IP) was constructed
as the ratio of the Loranthaceae stump
number (n) to fruit yield (N) per shea tree,
(IP = n/N). It was expected that the impact
index ratio (IP) had a value lower than 1.
Therefore, when for a shea tree individual,
the IP value is higher than 1, the more likely
it is that this shea tree individual is sensitive
to Loranthaceae impact on its fruit yield. All
shea tree individuals (164 total), with seven
traits (diameter at breast height; height of the
shea tree; crown diameter and crown
height; number of Loranthaceae stumps per
shea tree and number of fruit yielded per
shea tree; impact index ratio) were used
to perform hierarchical cluster analysis.
Groups of shea tree individuals with similar
traits were built using the Bray-Curtis dis-
tance measure and the group linkage
method of Flexible Beta in PC-ORD 5.0.
One-way ANOVA was used to compare shea
tree traits between shea tree discriminated
groups. Canonical Discriminant Analysis
(using the Mahalanobis distance) was per-
formed on groups and traits to reveal links
between the traits and plot distances
between groups. This analysis is a powerful
test to identify the discriminative traits
between entities that fall into groups [30].
Correlation among traits was performed
and tested using Pearson’s correlation. Since
there were significant strong correlations be-
tween diameter at breast height and height
of the shea tree, crown diameter and crown
height, we let height of the shea tree and
crown height when calculating the correla-
tion matrix of traits. Statistical analyses were
performed using SAS [31] and Minitab 14.
3. Results
3.1. Loranthaceae infection and shea
tree traits
Traits of shea trees were similar (p> 0.05)
between paired infected shea trees (weakly
and heavily) both in the protected area and
in the surrounding parklands (table I). All
things being equal, a significant difference
was expected between paired shea trees in
terms of fruit yield.
Table I.
Comparative shea tree traits in a protected area and in adjacent parklands in regard to Loranthaceae infection
(SE: standard error; p: probability of Student's ttest comparing the two paired sampled shea trees).
Area considered Tree diameter
(cm)
Tree height
(m)
Crown diameter
(cm)
Crown height
(m)
Protected area Weakly infected trees Mean 20.15 6.68 5.06 3.6
SE 0.72 0.23 0.2 0.18
Heavily infected trees Mean 19.80 7.02 5.10 3.98
SE 0.78 0.19 0.2 0.17
p0.7 0.2 0.8 0.1
Parklands Weakly infected trees Mean 23.87 6.97 6.50 4.12
SE 0.82 0.14 0.26 0.11
Heavily infected trees Mean 24.66 7.46 6.80 4.52
SE 0.82 0.22 0.24 0.20
p0.3 0.1 0.9 0.1

Loranthaceae impact on V. paradoxa fruit yield
Fruits, vol. 68 (2) 113
3.2. Variation of Loranthaceae
impact on shea tree fruit yield
in relation to habitat type
Loranthaceae impact on fruit yield did not
differ significantly between habitat types.
Indeed, the number of fruit yielded by
weakly infected shea trees did not differ
significantly from the heavily infected ones
in the protected area [(41.9 ± 8.4) fruits per
tree versus (64.41 ± 16.01) fruits per tree,
respectively; p= 0.2] or in the parklands
[(171.1 ± 39.3) fruits per tree versus
(159.5 ± 40.3) fruits per tree, respectively;
p= 0.8] (figure 1). The interaction between
habitat and shea tree status was also not
significant (p= 0.1). However, the yield of
fruit depended on the habitat. The number
of fruit was three times higher (p< 0.001)
in the parklands than in the protected area
[(165.32 ± 28.31) fruits per tree versus
(53.15 ± 9.07) fruits per tree]. Also, the fruit
number of all weakly infected shea trees
(106.51 fruits per tree) did not differ
significantly (p= 0.8) from that of all
heavily infected shea trees (111.97 fruits
per tree).
3.3. Variation of shea tree
quantitative traits in relation
to habitat type
The cluster analysis grouped the 164 indi-
viduals studied into five different groups
according to the seven assessed quantitative
traits (figure 2) with 60% of overall informa-
tion within each group. The composition
(i.e., rate of shea tree individuals of each
habitat belonging to each distinguished
group) of each shea tree group showed that
the G2 and G4 groups were composed
mostly of shea tree individuals from the
parklands, while shea trees from the pro-
tected area were mostly grouped in the G1,
G3 and G5 groups (figure 3). The canonical
discriminant analysis indicated globally that
the groups were significantly discriminated
(Wilks' Lambda = 0.18, p< 0.0001) by traits.
However, some traits (diameter at breast
height, crown height) showed similar mean
values among groups while others (crown
diameter, number of Loranthaceae stumps
per shea tree, number of fruit yielded per
shea tree, impact index ratio) showed the
opposite (table II).
The first two axes of canonical discrimi-
nant analysis explained 97% of the total var-
iation (p< 0.0001). Axis 1 of the canonical
discriminant analysis, that explained 79% of
total variation, was positively linked with
number of fruit yield (N), while axis 2
explained 19% of variation and was linked
with number of infected stumps (n) and the
impact index ratio (IP) (table III). Consider-
ing these results and the projection of shea
tree groups onto the two axes (figure not
shown), it can be concluded that shea trees
in the G4 group were characterized by the
highest value of fruit yield, while the G1
group showed the lowest one (table II).
Also, shea trees in the G2 group were char-
acterized by the highest value of number of
infected stumps (n) and of the impact index
ratio. This latter trait showed a value higher
than 1 in the G2 group and lower than 1 in
the other groups (table II).
Figure 1.
Comparative shea tree fruit
yield (mean ± standard error)
in regard to Loranthaceae
infection in parklands and
in a protected area.

114 Fruits, vol. 68 (2)
T.D. Houehanou et al.
3.4. Correlation among traits
When studying the correlation value among
traits, it can be noticed that the diameter at
breast height and the canopy diameter of
the shea tree were positively and signifi-
cantly correlated in the protected area and
in the parklands (table IV). The number of
infected stumps and the impact index ratio
were also positively and significantly corre-
lated in both habitats. Strong significant and
positive correlation between shea tree
diameter and number of infected stumps
was observed in the parklands, while the
protected area showed weak and non-sig-
nificant correlation.
4. Discussion and conclusion
4.1. Variation of Loranthaceae
impact on shea tree fruit yield
in relation to habitat type
According to the present study, weakly
infected shea trees did not show signifi-
cantly higher fruit yield compared with
heavily infected ones either in the parklands
or in the protected area. This suggests that
Loranthaceae do not significantly affect shea
Figure 2.
Diagram of hierarchical cluster analysis based on 164 shea tree individuals
and seven quantitative traits. Groups were built using the Bray-Curtis distance
measure and the group linkage method of Flexible Beta.
Figure 3.
Habitat composition of the five shea tree groups
defined by cluster analysis.

Loranthaceae impact on V. paradoxa fruit yield
Fruits, vol. 68 (2) 115
tree fruit yield either in parklands or in pro-
tected areas. Also, the interaction between
habitats (i.e., the parklands and protected
area) and shea tree status (i.e., weakly and
heavily infected shea trees) was not signif-
icant. These results support the rejection of
the first hypothesis, according to which
Loranthaceae were supposed to have a
higher impact on shea tree fruit yield in
parklands than in protected areas. If Loran-
thaceae impact on shea tree fruit yield was
not perceived significantly either in the
parklands or in the protected area, this
should be linked to the complexity of the
host-parasite relation [32]. It was also
revealed previously that the flowering and
fruiting behavior of infected shea tree
branches did not differ significantly from
healthy ones [12]. However, it was found
elsewhere [21,22] that host trees that carried
greater parasite load produced less fruit.
Thus, Loranthaceae impact on shea tree
needs to be investigated in other terms, such
as physiological impact. For instance, it
should be necessary to investigate the effect
of Loranthaceae on photosynthetic tissues
(i.e., water and carbohydrate nutrients) of
infected shea tree branches. Also, the impact
of Loranthaceae on the morphology and
composition (i.e., chlorophyll concentration
and biochemical composition) of shea tree
fruits may be investigated in the future in the
two contrasted habitats.
4.2. Variation of quantitative shea
tree traits according to habitat type
The hierarchical cluster analysis tended
to group shea tree individuals mostly on the
basis of the protected area (G1 and G3) and
parklands (G2 and G4). Additionally, the
canonical discriminant analysis indicated
that the fruit yield, the number of parasite
stumps and the impact index ratio discrim-
inated groups significantly. In a precedent
finding [16], shea trees in parklands produce
more than those in forests. This suggests that
shea tree fruit yield is a discriminant trait
Table II.
Mean value ± standard error of quantitative shea tree traits among groups built by hierarchical cluster analysis.
The numbers in brackets represent shea tree individual numbers that fall in each group.
Groups Diameter at breast
height (cm)
Tree height
(m)
Crown diameter
(m)
Crown height
(m)
Number of
infected stumps
Number of fruit
yielded
Impact
index ratio
G1 (45) 19.25 ± 0.54 4.58 ± 1.67 4.93 ± 0.10 3.89 ± 3.4 4.53 ± 0.77 6.8 ± 0.51 0.70 ± 0.20
G2 (13) 21.82 ± 1.15 5.19 ± 2.12 5.78 ± 0.15 4.59 ± 2.41 27.46 ± 2.43 18.1 ± 2.33 3.48 ± 1.82
G3 (62) 21.54 ± 0.67 5.30 ± 0.48 5.91± 0.12 4.78 ± 3.78 8.32 ± 1.06 50.0 ± 2.81 0.18 ± 0.02
G4 (20) 21.87 ± 1.02 5.68 ± 0.89 6.41 ± 0.17 5.12 ± 1.23 12.55 ± 2.77 533.7 ± 70.10 0.02 ± 0.00
G5 (23) 21.24 ± 1.02 5.51 ± 1.34 5.99 ± 0.13 4.97 ± 0.07 11.22 ± 2.75 154.6 ± 8.53 0.08 ± 0.02
Significance ns ns * ns ** ** **
ns: Not significant ; *: significant at p< 0.05; **: significant at p< 0.001.
Table III.
Correlation between shea tree traits and canonical discriminant
axes. Axis 1 (can1) is strongly correlated to the number of fruit
yielded and axis 2 (can2) to the number of infected stumps and
impact index ratio.
Shea tree traits Axis 1 (79%) Axis 2 (19%)
Diameter at breast height 0.117 0.159
Tree height 0.110 0.107
Crown diameter 0.238 0.129
Crown height 0.130 0.240
Number of fruit yielded 0.998 0.043
Number of infected stumps 0.120 0.906
Impact index ratio – 0.177 0.632

116 Fruits, vol. 68 (2)
T.D. Houehanou et al.
between natural areas (i.e., protected area)
and parklands. The canonical discriminant
analysis also indicated that G2 was com-
posed of shea trees with the highest value
of the number of infected stumps (n) and
impact index ratio. This suggests that park-
lands tend to have many shea tree individ-
uals with the highest parasite infection
degree. The impact index ratio was found
to be the highest in G2, with a value higher
than 1. Consequently, many infected shea
tree individuals in parklands carry a higher
value of the IP (i.e., a higher number of par-
asite stumps comparatively with a lower
number of fruit yielded). In that respect,
shea trees are more sensitive to Loran-
thaceae impact on their fruit yield in park-
lands than in protected areas. This suggests
the acceptance of the second hypothesis,
according to which prevalence of Loran-
thaceae impact on shea tree fruit yield is
higher in parklands than in protected areas.
Although Loranthaceae impact on shea tree
fruit yield has not been significantly proved
either in the protected area or in the sur-
rounding parklands, Loranthaceae are able
to induce that impact on many shea tree
individuals in parklands comparatively with
protected areas. This might be explained by
the fact that a greater number of shea tree
individuals carried greater parasite load in
parklands comparatively with protected
areas [23], as also shown by other authors
[33]onAlepis flavida (Loranthaceae). More-
over, shea tree individuals that carry a
greater parasite load should be more sensi-
tive to other environmental stresses (i.e.,
drought, light deficiency, soil nutrient defi-
ciency, etc.) and the combined effect of
Loranthaceae and environmental stress
should mostly induce an impact on shea tree
fruit yield.
The study also revealed that some shea
tree traits (diameter at breast height, tree
height and crown height) did not discrimi-
nate shea tree groups. Indeed, some groups
(G2 and G4) are composed mainly of shea
Table IV.
Correlation matrix between quantitative shea tree traits in a protected area and surrounding parklands.
• Protected area
Trait studied Diameter
at breast height
Crown diameter Number of fruit yielded
per shea tree
Number of Loranthaceae stumps
per shea tree
Crown diameter 0.501 * – – –
Number of fruit yielded
per shea tree
0.011 ns 0.063 ns – –
Number of Loranthaceae
stumps per shea tree
– 0.043 ns 0.087 ns 0.061 ns –
Impact index ratio – 0.122 ns – 0.048 ns – 0.146 ns 0.317 *
• Parklands
Trait studied Diameter
at breast height
Crown diameter Number of fruit yielded
per shea tree
Number of Loranthaceae stumps
per shea tree
Crown diameter 0.33 * – – –
Number of fruit yielded
per shea tree
0.119 ns 0.197 ns – –
Number of Loranthaceae
stumps per shea tree
0.261 * 0.154 ns 0.043 ns –
Impact index ratio 0.077 ns – 0.137 ns – 0.165 ns 0.214 *
ns: Not significant; *: significant at p< 0.05.

Loranthaceae impact on V. paradoxa fruit yield
Fruits, vol. 68 (2) 117
trees from the parklands. In West Africa, the
parklands have been built on selection of
desirable trees species and selected individ-
uals [34], and are continually managed by
humans. Therefore, this suggests that in a
given geographical environment, human
management is not able to induce pheno-
typic variation of those shea tree traits
because of shea tree gene flow. Shea tree
circumference in Ghana [15] and shea tree
crown size in Benin [35] discriminated shea
trees from different geographical positions.
Thus, it was expected that traits such as
diameter at breast height, tree height and
crown height are not able to discriminate
shea tree groups, because we assumed for
the study conditions that the environmental
effect (i.e., the climate and the soil condi-
tions are almost the same for all shea tree
sampled individuals) is minimized.
4.3. Correlation between traits used
Positive associations between the number
of infected stumps and the impact index
ratio (IP) were found in both habitats
(table IV). This means that when a shea tree
has a higher number of infected stumps, it
will be supposed to have a higher impact
index ratio (IP) (i.e., lower fruit yield than
number of infected stumps) and, conse-
quently, to be more sensitive to Loran-
thaceae impact on its fruit yield. Thus,
Loranthaceae are able to reduce shea tree
fruit yield when the parasite load increases.
The positive association between shea tree
diameter and number of infected stumps
observed in the parkland means that large
shea trees are supposed to carry a higher
parasite load (number of infected stumps)
in this habitat, whereas in the protected area
this is not the case.
4.4. Implications for potential shea
tree conservation
Our study offers opportunities for shea tree
in situ and ex situ conservation. As the shea
tree fruit yield is a desirable trait for farmers
[2], the in situ conservation approach in
parklands should aim to preserve potential
individuals suitable for high productivity
and, simultaneously, a low number of par-
asite stumps. This selection will be made
among the shea tree individuals that are
preserved by local people. Implementation
of the shea tree in situ conservation plan in
parklands may contribute to sustainable
use of shea tree in parklands and, conse-
quently, should decrease human pressure
on the protected populations. Indeed,
some protected areas exemplify degrada-
tion due to the use of plant resources [36].
However, in the protected area, the in situ
conservation of shea tree should consist to
preserve only large shea trees suitable for
high productivity.
The creation of a potential breeding shea
tree population is the common assumption
of ex situ conservation of plants. Shea trees
that have been selected to be preserved for
in situ conservation in each habitat can be
used as parents for selecting desired traits
linked to leaves and fruits of shea trees.
Indeed, there is a very low genetic relation
between traits related to shea tree and those
related to its leaves and fruits [15]. However,
prospective investigations based on Loran-
thaceae impact on shea tree fruit and leaf
morphology in relation to habitat will be
necessary for helping in leaf and fruit trait
selection.
Considering associations among traits,
the impact index ratio can be used as an
indicator to select shea trees for in situ con-
servation. Shea tree individuals that have a
value lower than 1 of the impact index ratio
will be desirable. Local people apply such
a selective plan to select and maintain shea
tree individuals in their farmed land. They
reported that, when they noted a high fruit
production potential for a shea tree individ-
ual, they cut off the parasites on it so as to
maintain a lower parasite load on the indi-
vidual concerned and, consequently, the
value of the impact index ratio was lower
than 1.
5. Conclusion
The current study offers information on var-
iation of Loranthaceae impact on shea tree
fruit yield in contrasting habitats. Our results

118 Fruits, vol. 68 (2)
T.D. Houehanou et al.
showed that Loranthaceae did not signifi-
cantly reduce shea tree fruit yield either in
the protected area or in the surrounding
parklands. The number of parasite stumps
and the defined impact index ratio were
revealed as significant discriminant traits of
shea tree groups between the two investi-
gated habitats. Shea tree individuals are
mostly sensitive (impact index ratio higher
than 1) to Loranthaceae impact on their fruit
yield in the parklands compared with those
in the protected area. These findings were
used for defining conservation plans for
shea trees. However, Loranthaceae impact
on shea tree needs to be investigated phys-
iologically in order to judge more com-
pletely Loranthaceae's effect on shea tree.
Acknowledgements
This research was conducted thanks to the
financial support provided by the BIOTA-
West III (Biodiversity Monitoring Transect
Analysis) project in West Africa. We thank
the donors of this project. We also thank
Prof. Romain Glèlè Kakaï and the anony-
mous reviewers for reading and comment-
ing on the manuscript and Odone Lifam for
his help with data collection.
References
[1] Hall J.B., Aebischer D.P., Tomlinson H.F.,
Osei-Amaning E., Hindle J.R., Vitellaria
paradoxa, A monograph, vol. 8, Univ. Wales,
Bangor, U.K., 1996.
[2] Lovett N.P., Haq N., Evidence for anthropic
selection of shea nut tree (Vitellaria para-
doxa), Agrofor. Syst. 48 (2000) 273–289.
[3] Teklehaimanot Z., Exploiting the potential of
indigenous agroforestry trees: Parkia biglo-
bosa and Vitellaria paradoxa in sub-Saharan
Africa, Agrof. Syst. 61 (2004) 207–220.
[4] Maranz S., Kpikpi W., Wiesman Z., Sauveur
A.D.S., Chapagain B., Nutritional values and
indigenous preferences for Shea fruits
(Vitellaria paradoxa C.F. Gaertn. F.) in African
agroforestry parklands, Econ. Bot. 58 (2004)
588–600.
[5] Plenderleith K., Brown N., Baillonella toxis-
perma: a state of knowledge study, Oxford
For. Inst., Dep. Plant Sci., Oxford Univ.,
Oxford, U.K., 2000.
[6] Boffa J.M.J., West African agroforestry
parklands: keys to conservation and sustai-
nable management, Unasylva 51 (2000)
11–17.
[7] Boussim I.J., Sallé G., Guinko S., Tapinan-
thus, parasite du karité au Burkina Faso, Bois
For. Trop. 238 (1993) 45–65.
[8] Sallé G., Boussim J., Raynal-Roques A.,
Brunck F., Le karité une richesse potentielle,
perspectives de recherche pour améliorer sa
production, Bois For. Trop. 228 (1991) 11–23.
[9] Boussim I.J., Les phanérogames parasites
du Burkina Faso: inventaire, taxonomie, éco-
logie et quelques aspects de leur biologie.
Cas particulier des Loranthaceae parasites
du karité, Thèse, Ouagadougou, Burkina
Faso, 2002, 285 p.
[10] Boussim I.J., Guinko S., Tuquet C., Sallé G.,
Loranthaceae of the agroforestry parklands
of Burkina Faso, Agrofor. Syst. 60 (2004)
39–49.
[11] Odebiyi J.A., Bada S.O., Omoloye A.A.,
AwodoyinR.O.,OniP.I.,Vertebrateandinsect
pests and hemi-parasitic plants of Parkia
biglobosa and Vitellaria paradoxa in Nigeria,
Agrofor. Syst. 60 (2004) 51–59.
[12] Lamien N., Boussim J.I., Nygard R.,
Ouédraogo J.S., Odén P.C., Guinko S.,
Loranthaceae impact on shea tree (Vitellaria
paradoxa C.F. Gaertn.) flowering and fruiting
behaviour in savanna area from Burkina Faso,
Environ. Exp. Bot. 55 (2006) 142–148.
[13] Brown A.H.D., Hardner C.M., Sampling the
gene pools of forest trees for ex situ conser-
vation, in: Young A., Boshier D., Boyle T.
(Eds.), Forest conservation genetics, prin-
ciples and practise, CSIRO Publ. CABI Publ.,
2000.
[14] El Kassaby Y., Effect of forest tree domesti-
cation on gene pools, in: Young A., Boshier
D., Boyle T. (Eds.), Forest conservation gene-
tics, principles and practise, CSIRO Publ.
CABI Publ., 2000.
[15] Sanou H., Picard N., Lovett P.N., Dembélé
M., Korbo A., Diarisso D., Bouvet J.-M., Phe-
notypic variation of agromorphological traits
of the shea tree, Vitellaria paradoxa CF
Gaertn., in Mali, Genet. Res. Crop Evol. 53
(2006) 145–161.
[16] Lamien N., Ouédraogo S.J., Diallo O.B.,
Guinko S., Productivité fruitière du karité
(Vitellaria paradoxa Gaertn. C. F., Sapota-
ceae) dans les parcs agroforestiers

Loranthaceae impact on V. paradoxa fruit yield
Fruits, vol. 68 (2) 119
traditionnels au Burkina Faso, Fruits 59
(2004) 423–429.
[17] Kelly B.A., Gourlet-Fleury S., Bouvet J.-M.,
Impact of agroforestry practices on the
flowering phenology of Vitellaria paradoxa in
parklands in southern Mali, Agrofor. Syst. 71
(2007) 67–75.
[18] Lovett P.N., Haq N., Diversity of the shea nut
tree (Vitellaria paradoxa C.F. Gaertn.) in
Ghana, Genet. Res. Crop Evol. 47 (2000)
293–304.
[19] Boffa J.M.J., Productivity and management
of agroforestry parklands in the Sudan zone
of Burkina Faso, West Africa, Purdue Univ.,
thesis, U.S.A., 1995, 101 p.
[20] Schreckenberg K., Forests, fields and
markets: a study of indigenous tree products
in woody savannas of the Bassila region,
Benin, Univ. Lond., thesis, U. K., 1996, 326 p.
[21] Sinha A., Bawa K.S., Harvesting techniques,
hemi-parasites and fruit production in two
non-timber forest trees species in South
India, For. Ecol. Manag. 168 (2002) 289–300.
[22] Ward D., Shrestha M.K., Musli I., Are invasive
Loranthaceae killing Ziziphus spina-christi?
Isr. J. Plant Sci. 54 (2006) 113–117.
[23] Houehanou T.D., Kindomihou V., Sinsin B.,
Effectiveness of conservation areas in
protecting shea trees against hemiparasitic
plants (Loranthaceae) in Benin, West Africa,
Plant Ecol. Evol. 144(3) (2011) 267–274.
[24] Sinsin B., Tehou A.C., Daouda I., Saidou A.,
Abundance and species richness of larger
mammals in Pendjari National Park in Benin,
Mammalia 66 (3) (2002) 369–380.
[25] Adomou C.A., Vegetation patterns and envi-
ronmental gradients in Benin. Implications for
biogeography and conservation, Wagening.
Univ., Wagening., thesis, Neth., 2005, 136 p.
[26] Seghieri J., Floret C.H., Pontanier R., Plant
phenology in relation to water availability:
herbaceous and woody species in the savan-
nas of northern Cameroon, J. Trop. Ecol. 11
(1995) 237–254.
[27] Okullo J.B.L, Hall J.B., Obua J., Leafing,
flowering and fruiting of Vitellaria paradoxa
subsp. nilotica in savannah parklands in
Uganda, Agrofor. Syst. 60 (2004) 77–91.
[28] Lamien N., Tigabu M., Guinko S., Oden P.C.,
Variations in dendrometric and fruiting cha-
racters of Vitellaria paradoxa populations and
multivariate models for estimation of fruit
yield, Agrofor. Syst. 69 (2007) 1–11.
[29] Zar J.H., Biostatistical analysis, 4th ed.,
Prentice Hall, N.J., U.S.A., 1999.
[30] Lowe A., Harris S., Ashton P., Ecological
genetics: design, analysis and application,
6th ed., Backwell, Lond., U.K., 2004.
[31] Anon., SAS Online Doc 9.1., SAS Inst. Inc.,
Cary, U.S.A., 2004.
[32] Tennakoon K.U., Pate J.S., Effects of
parasitism by Loranthaceae on structure and
functioning of branches of its host, Plant Cell
Environ. 19 (1996) 517–528.
[33] Bach C., Kelly D., Hazlett A.B., Forest edges
benefit adults, but not seedlings, of the
mistletoe Alepis flavida (Loranthaceae), J.
Ecol. 93 (2005) 79–86.
[34] Maranz S., Wiesman Z., Evidence for indige-
nous selection and distribution of the Shea
tree, Vitellaria paradoxa, and its potential
significance to prevailing parkland savanna
tree patterns in sub-Saharan Africa north of
the equator, J. Biogeogr. 30 (2003) 1505–
1516.
[35] Glèlè Kakaï R., Akpona T.J.D., Assogbadjo
A.E., Gaoué O.G., Chakeredz S., Gnanglè
P.C., Mensah G.A., Sinsin B., Ecological
adaptation of the shea butter tree (Vitellaria
paradoxa C.F. Gaertn.) along climatic gra-
dient in Benin, West Africa, Afr. J. Ecol. 49
(2011) 440–449.
[36] Mbayngone E., Thiombiano A., Dégradation
des aires protégées par l’exploitation des res-
sources végétales : cas de la réserve partielle
de faune de Pama, Burkina Faso (Afrique de
l’Ouest), Fruits 66 (2011) 187–202.

120 Fruits, vol. 68 (2)
T.D. Houehanou et al.
Variación del impacto de Loranthaceae en el rendimiento en frutos de
Vitellaria paradoxa C.F. Gaertn en los hábitats diferenciados, e implicaciones
para su conservación.
Resumen –Introducción. El karité (V. paradoxa CF Gaertn.), una especie endémica de los
bosques de la sabana de Sudán, predomina en los parques de África occidental, donde la
planta posee una gran importancia socio económica. Sin embargo, a lo largo de las últimas
décadas, el karité resultó estar amenazado por Loranthaceae, parásitos de vegetales. Nuestro
estudio pretendió comparar el índice de estos parásitos en el rendimiento en frutos de
árboles de karité, en dos hábitats diferenciados. Material y métodos. Seleccionamos 41 indi-
viduos de karité poco infectados y 41 individuos muy infectados, de tamaño similar, en dos
hábitats: una zona de vegetación protegida y parcelas adyacentes no protegidas. En cada indi-
viduo de karité seleccionado, se midieron ciertas características de los karités, tales como el
diámetro de los troncos a la altura del pecho, el diámetro de la canopea, la altura de los
árboles, la altura de la canopea, el número de frutos cosechados, el número de cepas de
parásitos por árbol y un índice calculado, evaluando el impacto del parásito en la producción
del árbol. Se realizó un análisis de la varianza, ANOVA, de dos vías para comparar el impacto
de los parásitos en el rendimiento en frutos de karité, según el hábitat. Se emplearon análisis
de clasificación jerárquica, canónica discriminante y ANOVA unidireccional para poner de
manifiesto los caracteres cuantitativos que caracterizan los grupos de karité en los hábitats.
Resultados. El impacto de Loranthaceae en el rendimiento en frutos de los árboles no fue
significativamente diferente en los dos hábitats considerados. Los caracteres cuantitativos ten-
dieron a discriminar a ciertos grupos de árboles de karité en ambos hábitats. Los individuos
de karité de las parcelas no protegidas se caracterizaron principalmente por un mayor
número de cepas de plantas parásito por árbol y por un índice de impacto en el rendimiento
mayor, lo que sugiere que, en estas parcelas, un mayor número de plantas de karité fueron
sensibles al impacto de Loranthaceae en su rendimiento en frutos. Conclusión. Estos resulta-
dos se emplearon para poner en marcha planes de conservación de los árboles de karité.
Benin / Vitellaria paradoxa / fruitas / rendimiento / Loranthaceae /
parasitismo / aptitud colonizadora / habitat